file
stringlengths
27
28
text
stringlengths
19
1.51M
PMC006xxxxxx/PMC6193568.txt
==== Front Cureus Cureus 2168-8184 Cureus 2168-8184 Cureus Palo Alto (CA) 30349765 10.7759/cureus.3158 Cardiology Endocrinology/Diabetes/Metabolism Internal Medicine Effects of Body Mass Index, Glycemic Control, and Hypoglycemic Drugs on Serum Uric Acid Levels in Type 2 Diabetic Patients Muacevic Alexander Adler John R Hussain Azhar 1 Latiwesh Omar B 2 Ali Farwa 3 Younis M.Y. G 4 Alammari Jamal A 5 1 Medicine, Xavier University School of Medicine, Oranjestad, ABW 2 Medical Laboratory, Higher Institute of Medical Professions, Benghazi, LBY 3 Medicine, American University of Antigua College of Medicine, New York, USA 4 Assistant Professor and Head of the Department of Biochemistry, University of Benghazi, Faculty of Medicine, Benghazi, LBY 5 Public Health, Higher Institute of Comprehensive Vocations, Gamins, LBY Azhar Hussain [email protected] 17 8 2018 8 2018 10 8 e31589 8 2018 17 8 2018 Copyright © 2018, Hussain et al. 2018 Hussain et al. https://creativecommons.org/licenses/by/3.0/ This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. This article is available from https://www.cureus.com/articles/14197-effects-of-body-mass-index-glycemic-control-and-hypoglycemic-drugs-on-serum-uric-acid-levels-in-type-2-diabetic-patients Background Plasma uric acid has been shown to be associated with an increased risk of hypertension, cardiovascular disease, chronic kidney disease, insulin resistance, and metabolic syndrome. Conflicting data regarding plasma uric acid levels in type 2 diabetes mellitus and their role in the development and progression of diabetic complications have been observed by many studies. The present study aimed to evaluate plasma uric acid levels in type 2 diabetic patients and to determine the effects of hypoglycemic drugs and pharmacologic insulin on plasma uric acid concentration. Subjects and methods The study included 162 type 2 diabetic patients divided into three groups (insulin taking group (N=58), glibenclamide taking group (N=40), and metformin taking group (N=64), and 47 normal healthy controls. A questionnaire that included variables such as age, sex, duration of disease, and body mass index (BMI) were answered by all the participants. Blood samples were collected and estimated for serum uric acid (SUA), fasting blood sugar (FBS), and glycated hemoglobin (HbA1c) using standard methods and the data were statistically analyzed. Results Diabetic patients showed a significant increase in serum uric acid, fasting blood sugar, glycated hemoglobin, and body mass index when compared to control subjects. The serum uric acid levels of metformin and glibenclamide taking groups were significantly higher than the control group. The difference of serum uric concentration between the insulin taking group and both the control and metformin groups was statistically non-significant. On the other hand, obese diabetics showed a significantly higher serum uric acid than overweight and lean diabetics. Furthermore, serum uric acid had a significant strong positive correlation with body mass index. Conclusion Type 2 diabetes mellitus (T2DM) is associated with high serum uric acid levels. Hypoglycemic drugs and pharmacologic insulin do not have a large impact on SUA concentration, but obesity seems to be the primary determinant of SUA levels in T2DM patients. The condition of diabetes may have a direct effect on the oxidation of the purine nucleotides resulting in the increased uric acid (UA) levels. In addition, hyperinsulinemia could lead to hyperuricemia by increasing the rate of xanthine oxidase synthesis. There is a strong relationship between T2DM and obesity with high uric acid levels. serum uric acid type 2 diabetes mellitus cardiovascular diseases hypoglycemic drugs insulin ==== Body pmcIntroduction Uric acid (UA) is the final oxidation product of purine catabolism [1]. Uric acid can act as a pro-oxidant, particularly at increased concentrations and may thus be a marker of oxidative stress [2]. Thus, it is unclear whether increased concentrations of UA in diseases associated with oxidative stress, such as atherosclerotic coronary heart disease (CHD), stroke, and peripheral arterial occlusive disease, are a protective response or a primary cause [3]. Type 2 diabetes mellitus (T2DM) is a risk factor for nephrolithiasis and has been associated with UA stones [4]. It has been suggested that patients with UA stones, especially if overweight, should be screened for T2DM or metabolic syndrome [5]. The association between high uric acid levels and insulin resistance is not fully understood [3,4]. Hyperuricemia in T2DM is usually the result of underexcretion of urate. Reaven et al. attributed the presence of hyperuricemia in metabolic syndrome to a secondary response to hyperinsulinemia. The association has been attributed to the effects of insulin on proximal tubular urate transport of the kidney [6]. Insulin can also enhance renal tubular sodium reabsorption [7], which in turn can reduce renal excretion of UA. Hyperinsulinemia could lead to hyperuricemia by increasing the rate of xanthine oxidase synthesis, an enzyme involved in UA production [4]. Because endothelial nitric oxide synthase deficiency results in the features of insulin resistance and metabolic syndrome [8], and because UA has been shown to reduce nitric oxide bioavailability [9], consequently, hyperuricemia may have a key role in the pathogenesis of insulin resistance and thus of T2DM. Moreover, many studies have shown the implication of UA and its metabolites in the development and progression of diabetic complications such as peripheral neuropathy [10], retinopathy [11], nephropathy [12], and cardiovascular diseases [13]. T2DM is a major health problem in the Libyan population and accounts for a high mortality rate; hence this study was conducted to evaluate the association of serum UA levels with obesity and glycemic control in T2DM patients and to find out the effects of hypoglycemic drugs and pharmacologic insulin on serum UA levels. This study was presented at the 2nd Libyan Conference on Chemistry and Its Applications, LCCA-2 on September 9, 2017. Materials and methods A total of 162 patients with T2DM were recruited from the Benghazi Center for Diagnosis and Treatment of Diabetes. They are divided into the following three groups according to diabetes treatment: (insulin taking group (N=58), glibenclamide taking group (N=40), and metformin taking group (N=64). Forty-seven apparently healthy age and sex-matched individuals were selected from our family members and relatives to serve as controls. Informed consent was obtained from all the participants before the study, and approval was obtained from the Ethics Review Board of the University of Benghazi. All patients were diagnosed with T2DM based on the American Diabetes Association criteria 2006, i.e., A1c ≥ 6.5%, or fasting plasma glucose level ≥ 126 mg/dL, or 2-h plasma glucose ≥ 200 mg/dl during an oral glucose tolerance test. Clinical information and medical history were obtained through a questionnaire that included variables such as age, sex, date of the diagnosis, physical activity, following a diet, and any health problems, or prescribed drugs. The height and weight were measured, and obesity was defined as body mass index (BMI) of ≥ 30 kg/m2, where BMI was calculated by dividing the weight in kilograms over height in meters squared. All patients presented stable metabolic conditions. Patients presenting any disease that could affect their metabolic status and the parameters studied, such as nephrotic syndrome, acute or chronic renal failure, hepatitis or other liver diseases, cardiovascular diseases, arthritis, acute or chronic inflammatory conditions, gout, and cerebrovascular diseases, were excluded from the study. Patients with a history of smoking or alcohol intake were also excluded. Pregnant and lactating women were excluded. The history of medication was recorded and patients taking any drugs that could affect serum UA levels were also excluded. The control group consisted of healthy subjects who were not suffering from an acute infection or metabolic or psychological disorder. They were non-smokers and non-overweight. They had no history of acquired or inherited hyperuricemia or diabetes mellitus. Venous blood samples were drawn from all the participants after at least 10 hours of fasting. Blood was collected in ethylenediaminetetraacetic acid (EDTA) and plain tubes, and sera were separated from plain tubes and stored at – 20oC until the assays were performed. The whole blood was stored at 4 – 8oC and analyzed for HbA1c within a week using a fully automated Cobas Integra 400 plus (Roche, Germany). Sera were analyzed manually for fasting blood sugar, and UA were analyzed by commercial kits supplied by Linear Chemicals SL, Spain, using Photometer 4040v5+ Robert Riele GmbH & Co, Germany. The data were statistically analyzed using Statistical Package for the Social Sciences (SPSS 17, IBM Corporation). Analysis of variance test (ANOVA) and independent samples T-test were used to determine the variance between different subject groups. Pearson's correlation analysis was done to evaluate the degree of association between different clinical and biochemical parameters. Descriptive characteristics of the study participants were calculated as the mean ± standard deviation (SD), and a P value < 0.05 was considered as statistically significant. Results The mean age and standard deviation of the diabetic group was 51.2 ± 10.9, and the male:female ratio was 4:5. The age range was 18-80 years with the duration of disease ranging from 1-30 years. The mean age and SD of the healthy control subjects was 49.4 ± 12.6, and the male:female ratio was 12:13. The age range was 35-81 years. As shown in Table 1, body mass index (BMI), fasting blood sugar (FBS), glycosylated hemoglobin (HbA1c), and serum uric acid (SUA) levels were significantly higher in the diabetic group than in the normal control group (p< 0.05). Fasting blood sugar was significantly lower in the control group when compared to each of the diabetic groups (p< 0.05). The difference in FBS was also statistically significant between the metformin taking group and both the insulin and the glibenclamide taking groups (p< 0.05). No significant difference has been found between the insulin group and the glibenclamide group. HbA1c concentration was statistically higher in all diabetic groups when compared to that in the control group (p< 0.05). Moreover, the difference in mean HbA1c between different diabetic groups was statistically significant (p< 0.05). The body mass index was significantly higher in both metformin and glibenclamide groups when compared to either insulin or control groups. On the contrary, significant difference in BMI has not been observed between the metformin group and the glibenclamide group or between the insulin group and the control group. Furthermore, similar to BMI, SUA levels in the metformin and glibenclamide groups were significantly higher than in the insulin and control groups. Similarly, there was a non-significant difference in UA levels between the metformin and the glibenclamide groups, as well as between the insulin and the control groups. Table 1 Mean ± standard deviation (SD) of body mass index (BMI) and biochemical characteristics of type 2 diabetic patients and normal control groups. Multiple Comparisons  (Post Hoc analysis)   Diabetic and control groups   Parameters Metformin (N= 64 ) Glibenclamide (N= 40 ) Insulin (N= 58) Control (N= 47 ) P value   Fasting Blood Sugar (mg/dl) 161.9 ± 58 214.7 ± 91.3 193.6 ± 80 92.7 ± 12 0.00   Metformin - 0.001 0.007 0.00   Glibenclamide 0.001 - 0.17 0.00   Insulin 0.007 0.17 - 0.00   Control 0.00 0.00 0.00 -   Glycated Hemoglobin (HbA1c ) (%) 7.7 ± 1.8 9.5 ± 2.4 8.4 ± 2.2 5 ± 0.4 0.00   Metformin -   0.036 0.00   Glibenclamide 0.00 - 0.017 0.00   Insulin 0.036 0.017 - 0.00   Control 0.00 0.00 0.00 -   Body Mass Index (kg/ m2) 29 ± 6.1 29.4 ± 4.7 26 ± 5.6 24.4 ± 0.5 0.002 Metformin - 0.72 0.004 0.004   Glibenclamide 0.72 - 0.012 0.009   Insulin 0.004 0.012 - 0.82   Control 0.004 0.009 0.82 -   Serum Uric Acid (mg/dl) 5.8 ± 2.7 6.4 ± 2 4.9 ± 1.8 4.7 ± 0.9 0.001 Metformin - 0.19 0.02 0.007   Glibenclamide 0.19 - 0.003 0.001   Insulin 0.02 0.003 - 0.6   Control 0.007 0.001 0.6 -   Dividing diabetic patients according to the BMI cut-off value of 30 kg/m2 revealed a significantly higher SUA level in obese diabetics in comparison to lean and overweight diabetics (mean SUA of 6.52 ± 2.6 in obese diabetics versus mean SUA of 5.09 ± 2 in lean and overweight diabetics, p= 0.00). In addition, by dividing diabetic patients according to the HbA1c cut-off value of 7.5%, we found a significantly higher SUA concentration in patients with HbA1c > 7.5% when compared to those with HbA1c ≤ 7.5% (mean SUA of 6.05 ± 2.5 in diabetics with HbA1c > 7.5% versus mean SUA of 4.9 ± 2.1 in diabetics with HbA1c ≤ 7.5%, p= 0.006). In the type 2 diabetic group, Pearson’s correlation analysis revealed significantly positive correlations between SUA levels and both body mass index (p= 0.001, r= 0.279) (Figure 1), and HbA1c concentration (p= 0.017, r= 0.197) (Figure 2). On the other hand, no significant correlation has been found between SUA and FBS (p= 0.12, r= 0.126), age (p= 0.5, r= - 0.4), or duration of disease (p= 0.59, r= - 0.05). Figure 1 Correlation between serum uric acid (SUA) and body mass index (BMI). Figure 2 Correlation between serum uric acid (SUA) and glycated hemoglobin (HbA1c). Discussion In the present case-control study, SUA levels were significantly higher in diabetic patients than normal healthy controls. This finding is in line with data published in previous studies in which high SUA level has been associated with T2DM [14-17]. Persons diagnosed with T2DM have shown very high UA levels in their blood compared to people suffering from diseases such as gout. This indicates that the condition of diabetes may have effects on the oxidation of purine nucleotides. However, the actual relationship between the two is not fully understood due to the complications of metabolic syndrome [17]. Hyperuricemia in T2DM is usually the result of underexcretion of urate as a secondary response to hyperinsulinemia [6,7]. In addition, hyperinsulinemia could lead to hyperuricemia by increasing the rate of xanthine oxidase synthesis, an enzyme involved in UA production [4]. Some studies showed a non-significant difference in UA levels between diabetic patients and controls [18], while other studies showed a significant lower SUA concentration in T2DM patients when compared to normal controls [19]. In the present study, SUA showed significant positive correlation with HbA1c in diabetics. Choi et al. [20] in their study of hemoglobin A1c, fasting glucose, serum C-peptide, and insulin resistance in relation to SUA levels observed that SUA levels and the frequency of hyperuricemia increased with moderately increasing levels of HbA1c and fasting plasma glucose (FPG) and then decreased with further increasing levels of HbA1c (bell-shaped relation). A biological mechanism underlying the bell-shaped relation between blood glucose levels and SUA levels is thought to be due to the uricosuric effect of glycosuria, which occurs when the blood glucose level is greater than 180 mg/dl [21]. We found a significant positive correlation between SUA and BMI in the type 2 diabetic group, and this observation is in agreement with the results of other studies [20-21]. This might be explained with the presence of increased intracellular adenosine (uric acid precursor), a derivative of higher adenosine monophosphate (AMP) concentrations due to increased synthesis of fatty acyl-CoA in peripheral tissues [22], and experiments on mice showed a high xanthine oxidase activity in adipose tissues [23]. In the present study, body mass index was significantly higher in both the metformin and glibenclamide groups when compared to either insulin or control groups. In contrast to our findings, a study by Barskova VG et al., to evaluate results of metformin (MF) therapy during one year of UA metabolism and the clinical course of gout with insulin resistance, revealed the hypouricemic effect of MF and hypothesized that MF reduces the production of UA in patients with gout due to the inhibition of synthesis of free fatty acids [24]. Moreover, metformin use in T2DM improves the sensitivity of peripheral tissues to insulin, which results in a reduction of circulating insulin levels [25], thus decreasing the effect of hyperinsulinemia in reducing the excretion of uric acid. The effect of pharmacologic insulin on SUA has been studied by Lindsey A et al. and Ter Maaten JC et al, and they found a significant increase in serum uric acid levels in diabetic and healthy individuals treated with exogenous insulin, and they referred their finding to insulin’s effects on renal handling of urate [26, 27]. After the treatment with metformin, an increased level of SUA has been observed in type 2 diabetic patients. In comparison to patients treated with rosiglitazone, there is a non-significant difference in SUA levels [28]. A study conducted in Iraq by Ismail NS revealed a non-significant difference in SUA level between both glibenclamide and metformin groups and concluded that glibenclamide and/or metformin had no significant effect on the SUA level in patients with T2DM [29]. In a study by Luque-Ramírez M et al., SUA levels were measured in 40 polycystic ovary syndrome (PCOS) patients and 40 normal healthy women matched for BMI and obesity grade and were followed up for 24 weeks in 34 PCOS patients who were randomized to an oral contraceptive (Diane Diario) or metformin (850 mg twice daily). They found a non-significant difference in SUA between the PCOS group and normal women. When they divided PCOS and normal women as a whole according to BMI, it was revealed that obese women showed higher UA concentrations than lean and overweight women, and that BMI is the main determinant of SUA levels in PCOS patients [30]. This observation supports our finding that metformin and glibenclamide groups both had significantly higher BMI and SUA levels than either insulin group or control group and that SUA level was significantly higher in obese diabetics than lean and overweight diabetics. Conclusions T2DM is associated with high serum uric acid levels. Hypoglycemic drugs and pharmacologic insulin do not have a large impact on SUA concentration, but obesity seems to be the primary determinant of SUA levels in T2DM patients. The condition of diabetes may have a direct effect on the oxidation of the purine nucleotides resulting in increased UA levels. In addition, hyperinsulinemia could lead to hyperuricemia by increasing the rate of xanthine oxidase synthesis. There is a strong relationship between T2DM and obesity with high uric acid levels. Human Ethics Animal Ethics Consent was obtained or waived by all participants in this study. Ethics Review Board of the University of Benghazi issued approval N/A. Informed consent was obtained from all the participants before the study, and approval was obtained from the Ethics Review Board of the University of Benghazi. Animal subjects: All authors have confirmed that this study did not involve animal subjects or tissue. The authors have declared that no competing interests exist. ==== Refs References 1 Towards the physiological function of uric acid Free Radic Biol Med Becker BF 615 631 14 1993 8325534 2 Uric acid and oxidative stress: relative impact on cardiovascular risk Nutr Metab Cardiovasc Dis Strazzullo P Puig JG 409 414 17 2007 17643880 3 Uric acid, type 2 diabetes, and cardiovascular diseases: fueling the common soil hypothesis? Cli Chem Koenig W Meisinger C 231 223 54 2008 4 Urine composition in type 2 diabetes: predisposition to uric acid nephrolithiasis Clin J Am Soc Nephrol Cameron MA Maalouf MN Adams-Huet B Moe WO Sakhaee K 1422 1428 17 2006 5 Type 2 diabetes increases the risk for uric acid stones Clin J Am Soc Nephrol Daudon M Traxer O Conort P Lacour B Jungers P 2026 2033 17 2006 6 The kidney: an unwilling accomplice in syndrome X Am J Kid Dis Reaven GM 928 931 30 1997 9398143 7 Decreases in serum uric acid by amelioration of insulin resistance in overweight hypertensive patients: effect of a low-energy diet and an insulin-sensitizing agent Am J Hypertens Tsunoda S Kamide K Minami J Kawano Y 697 701 15 2002 12160192 8 Insulin stimulation of glucose uptake in skeletal muscles and adipose tissues in vivo is NO dependent Am J Physiol Endocrinol Metab Roy D Perreault M Marette A 692 699 274 199 9 Hyperuricemia induces endothelial dysfunction Kidney Int Khosla UM Zharikov S Finch JL 1739 1742 67 2005 15840020 10 Association between serum uric acid level and microalbuminuria to chronic vascular complications in Thai patients with type 2 diabetes J Diabetes Complications Chuengsamarn S Rattanamongkolgul S Jirawatnotai S 124 129 28 2014 24412514 11 Diabetic retinopathy is associated with visceral fat accumulation in Japanese type 2 diabetes mellitus patients Metab Clin Exp Anan F Masaki T Ito Y 314 319 59 2010 20004426 12 Diabetic hypouricemia as an indicator of clinical nephropathy Am J Nephrol Shichiri M Iwamoto H Marumo F 115 122 10 1990 2190467 13 The association of serum uric acid levels with outcomes following percutaneous coronary intervention J Interv Cardiol Spoon DB Lerman A Rule AD Prasad A Lennon RJ Holmes DR Rihal CS 277 283 23 2010 20636849 14 Correlation of the serum insulin and the serum uric acid levels with the glycated haemoglobin levels in the patients of type 2 diabetes mellitus J Clin Diagn Res Gill A Kukreja S Malhotra N 1295 1297 7 2013 23998049 15 To study serum uric acid in type 2 diabetes mellitus patient IOSR J Dent Med Sci 8 2018 Khare S Vishandasani JK Kansal A 5 14 2015 https://pdfs.semanticscholar.org/23d6/1f2f9b6a6e42c1047260d3037131734f9335.pdf 16 Relation between serum uric acid and non insulin dependent diabetes mellitus (NIDDM) Natl J Integr Res Med 8 2018 Nayak MS Shah RM 72 75 4 2013 http://www.scopemed.org/?mno=47448 17 To study serum uric acid and urine microalbumin in type-2 diabetes mellitus Int J Med Sci Suryawanshi KS Jagtap PE Belwalkar GJ Dhonde S Nagane NS Joshi VS 24 29 2 2015 http://www.internationaljournalssrg.org/IJMS/2015/Volume2-Issue3/IJMS-V2I3P104.pdf 18 A study of serum uric acid in diabetes mellitus and prediabetes in a South Indian tertiary care hospital Nitte Uni J Heal Sci 8 2018 Rao MS Sahayo BJ 18 23 2 2012 http://nitte.edu.in/journal/juneSplit/Nitte%20University%20Journal%20June%202012_18_23.pdf 19 Hypouricemia and hyperuricemia in type 2 diabetes: two different phenotypes Eur J Clin Invest Bo S Cavallo‐Perin P Gentile L Repetti E Pagano G 318 321 31 2001 11298778 20 Haemoglobin A1c, fasting glucose, serum C-peptide and insulin resistance in relation to serum uric acid levels—the Third National Health and Nutrition Examination Survey Rheumatology Choi HK Ford ES 713 717 47 2008 18390895 21 Serum uric acid, serum glucose and diabetes: relationships in a population study Post Grad Med J Cook DG Shaper AG Thelle DS Whitehead TP 1001 1006 62 1986 22 Uric acid is associated with features of insulin resistance syndrome in obese children at prepubertal stage Nutr Hosp 8 2018 Gil-Campos M Aguilera CM Cañete R Gil Ay 607 613 24 2009 http://scielo.isciii.es/scielo.php?script=sci_arttext&pid=S0212-16112009000500013&lng=en&nrm=iso&tlng=en#back 19893872 23 Uric acid secretion from adipose tissue and its increase in obesity J Biol Chem 8 2018 Tsushima Y Nishizawa H Tochino Y 27138 27149 288 2013 http://www.jbc.org/content/288/38/27138 23913681 24 Effect of metformin on the clinical course of gout and insulin resistance. [Article in Russian] Klin Med Barskova VG Eliseev MS Kudaeva FM Aleksandrova EN Volkov AV Nasonova VA Nasonov EL 41 46 87 2008 https://www.ncbi.nlm.nih.gov/pubmed/19705791 25 Metformin New Engl J Med Bailey CJ Turner RC 574 579 334 1996 8569826 26 The effect of initiating pharmacologic insulin on serum uric acid levels in patients with diabetes: a matched cohort analysis Semin Arthritis Rheum MacFarlane LA Liu CC Solomon DH 592 596 44 2015 25455681 27 Renal handling of urate and sodium during acute physiological hyperinsulinaemia in healthy subjects Cli Sci Ter Maaten JC Voorburg A Heine RJ Ter Wee PM Donker AJM Gans ROB 51 58 92 1997 28 Metabolic effects of rosiglitazone and metformin in Greek patients with recently diagnosed type 2 diabetes Int J Experi Cli Pathophys Drug Res 8 2018 Iliadis F Kadoglou NP Hatzitolios A Karamouzis M Alevizos M Karamitsos D 1107 1114 21 2007 http://iv.iiarjournals.org/content/21/6/1107.long 29 Effects of glibenclamide and metformin on serum uric acid level in patients with type 2 diabetes mellitus Iraq J Pharm 8 2018 Ismail NS 11 2011 https://www.iasj.net/iasj?func=fulltext&aId=49595 30 Serum uric acid concentration as non-classic cardiovascular risk factor in women with polycystic ovary syndrome: effect of treatment with ethinyl-estradiol plus cyproterone acetate versus metformin Hum Reprod Luque-Ramírez M Álvarez-Blasco F Uriol-Rivera MG Escobar-Morreale HF 1594 1601 23 2008 18375410
PMC007xxxxxx/PMC7012834.txt
==== Front Nat Commun Nat Commun Nature Communications 2041-1723 Nature Publishing Group UK London 32047164 14658 10.1038/s41467-020-14658-6 Article Atomic structures of anthrax toxin protective antigen channels bound to partially unfolded lethal and edema factors http://orcid.org/0000-0003-1539-6768 Hardenbrook Nathan J. 1 http://orcid.org/0000-0002-0561-1981 Liu Shiheng 23 http://orcid.org/0000-0002-3034-2808 Zhou Kang 23 http://orcid.org/0000-0002-2344-9631 Ghosal Koyel 1 http://orcid.org/0000-0002-8373-4717 Zhou Z. Hong [email protected] 23 http://orcid.org/0000-0002-4911-5824 Krantz Bryan A. [email protected] 1 1 grid.411024.2 0000 0001 2175 4264 Department of Microbial Pathogenesis, University of Maryland, Baltimore, Baltimore, MD 21201 USA 2 grid.19006.3e 0000 0000 9632 6718 Department of Microbiology, Immunology and Molecular Genetics, University of California, Los Angeles, CA 90095 USA 3 grid.509979.b 0000 0004 7666 6191 California NanoSystems Institute, University of California, Los Angeles, CA 90095 USA 11 2 2020 11 2 2020 2020 11 8408 7 2019 15 1 2020 © The Author(s) 2020, corrected publication 2023 https://creativecommons.org/licenses/by/4.0/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Following assembly, the anthrax protective antigen (PA) forms an oligomeric translocon that unfolds and translocates either its lethal factor (LF) or edema factor (EF) into the host cell. Here, we report the cryo-EM structures of heptameric PA channels with partially unfolded LF and EF at 4.6 and 3.1-Å resolution, respectively. The first α helix and β strand of LF and EF unfold and dock into a deep amphipathic cleft, called the α clamp, which resides at the interface of two PA monomers. The α-clamp-helix interactions exhibit structural plasticity when comparing the structures of lethal and edema toxins. EF undergoes a largescale conformational rearrangement when forming the complex with the channel. A critical loop in the PA binding interface is displaced for about 4 Å, leading to the weakening of the binding interface prior to translocation. These structures provide key insights into the molecular mechanisms of translocation-coupled protein unfolding and translocation. The bacterial translocase channel anthrax toxin is composed of the protective antigen (PA) that forms a translocase channel and two cytotoxic enzymes: lethal factor (LF) and edema factor (EF), which go through the PA channel to enter host cells. Here the authors provide mechanistic insights into LF and EF translocation by determining the cryo-EM structures of the anthrax toxin transmembrane channel protein in complex with LF and EF. Subject terms Cryoelectron microscopy Pathogens issue-copyright-statement© Springer Nature Limited 2020 ==== Body pmcIntroduction Protein translocation is an essential process in cells. Nearly one half of all proteins are translocated across a membrane to perform their respective functions1. This process often requires dedicated protein translocation machineries, generally referred to as translocons, to catalyze the unfolding and translocation of proteins1. In their native state, most proteins are thermodynamically stable. Therefore, translocons require energy in various forms, such as a proton gradient2, hydrolysis of ATP3, or membrane potential3,4, to drive the translocation of their substrates. This process utilizes polypeptide clamps, or catalytic active sites that are responsible for promoting translocation of the protein. In many types of unfoldases5, translocases6, and secretion channels7, these polypeptide clamps engage the polypeptide chain nonspecifically as it is unfolded and translocated. However, in the absence of high-resolution structures of translocons engaged in translocation of an unfolded protein substrate, the biophysical mechanisms involved in protein unfolding and translocation through translocons remain poorly understood. Anthrax toxin8 is well suited for the study of protein translocation. The toxin functions as a binary A2B toxin, with enzymatic A factors, lethal factor (LF, 91 kDa) and edema factor (EF, 89 kDa), and a cell binding B factor, protective antigen (PA, 83 kDa). Anthrax lethal factor (LF) is a 776-amino acid protein consisting of four protein domains; domain 1 is a PA-binding domain (PABD), domain 2 a VIP2-like domain, domain 3 is a helical bundle, and domain 4 is the catalytic center domain (CCD)9. LF has been shown to be a protease, which targets the mitogen activated protein kinase (MAPK) pathway, specifically by cleaving MAPK kinases10–12. Edema factor is also a four domain protein with a PA-binding domain (PABD), two adenylate cyclase domains (ACD), and a helical domain (HD)13. As an adenylate cyclase, EF requires calmodulin (CaM) for its activity upon translocation to the host cytosol14. EF has a catalytic rate of ~2000 molecules per second, resulting in high levels of cyclic adenosine monophosphate (cAMP) that activates protein kinase A (PKA) signaling pathways15,16. Currently, the only protein structure of EF containing all four domains is a CaM-bound crystal structure. Thus while these proteins are both translocated by PA into the cytosol, they perform very different functions. The PA undergoes furin cleavage to form a ring-shaped homooligomeric pre−channel, either a heptamer or an octamer17. The pre-channel PA can then bind to LF or EF, forming lethal toxin (LT) or edema toxin (ET), respectively. The toxin is then endocytosed into an acidic compartment inside the cell. Within the endosomal compartment, the acidic environment induces a conformational change in the PA, resulting in the formation of a β-barrel channel that can insert itself into the endosomal membrane18. A proton gradient forms between the endosome and the cytosol to drive the translocation process2. The enzymatic factors, LF and EF, bound to the PA channel are destabilized by the acidic environment within the endosome and then unfold and translocate through the channel19. Atomic structures of the anthrax toxin PA pre-channel and channel have been determined by X-ray crystallography17,20,21and cryo-EM18, respectively, revealing structural features supporting protein unfolding and translocation. The overall structure of the PA channel has a mushroom-shaped architecture, similar to bacterial α-hemolysin22. The PA channel contains three polypeptide clamp sites23: the α clamp20, the ϕ clamp24, and the charge clamp25. The α clamps, found on the topmost surface of PA, are clefts formed between two PA subunits that binds to α helices nonspecifically20. Through the α clamps and other more specific binding sites, the PA heptamers or octamers can bind three or four LF and/or EF, respectively. In addition to the α clamp, there is a binding interface formed between PA and the C-terminus of the PABD of LF (LFN) in this region. PA residues K213 and K214 have been shown previously to be important in the binding of LFN to PA26. PA K213 was shown to interacts with D187 in LFN in this binding interface26. This was later shown structurally, with K213 and K214 in PA forming salt bridges with LFN residues D187 and D184, respectively20. A charge reversal in either of these PA residues was shown to drastically inhibit binding of LFN26. Directly below the α clamp within the center of the channel is the ϕ clamp. The 2.9-Å resolution cryo-EM structure of the PA pore reveals that the ϕ clamp forms a constricted 6-Å bottleneck of Phe427 residues18. The α clamp and ϕ clamp appeared to behave in an allosteric manner that the peptide binding at the α-clamp site are required for allosteric gating of the ϕ clamp to a clamped state27. Below the ϕ clamp is a charge clamp formed by the transmembrane β-barrel25. As the partially protonated polypeptide chain moves from the lower pH in the endosome toward the higher pH within the cytosol, it passes the negatively charged acidic residues of the clamp. Here the chain becomes deprotonated, and thereafter cannot retro-translocate back through the channel. The inner diameter of the channel spans a range of diameters as low as 20 Å, wide enough to accommodate α helix in the translocating peptide, but not large enough to fit folded domains of the enzymatic factors. Many questions remain with respect to the translocation of substrates through the PA channel. How do the catalytic domains of the substrate proteins interact with the channel? Are there binding sites beyond the α clamp that stabilize partially unfolded substrate? Are there changes within the channel structure when bound to substrate? A lack of structural information on the PA channel bound to substrate has made it difficult to address these questions. Here we report the cryo-EM structures of the PA channel bound to LF and EF. These high-resolution structures of the PA with partially unfolded protein factors reveal conformational changes occurring within the enzymatic factors upon binding to the PA channel, providing key insight on the mechanism of proton-driven protein translocation. Results Overall structures of PA channel in complex with LF and EF Conversion of the PA pre-channel to the channel by in vitro acidification treatment leads to rapid and irreversible aggregation due to exposure of the hydrophobic transmembrane β-barrel structure. Attempts to prevent aggregation by screening detergents were mostly unsuccessful. We next tried to apply low-pH treatment of PA pre-channels directly on carbon-coated grids as done before17, but were only able to obtain limited number of dispersed particles of PA channel without aggregation. To overcome these issues, we used lipid nanodiscs28 to assemble water-soluble complexes containing the PA channel bound to LF and EF29,30. Each complex was assembled on nickel affinity resin using His-tags in the enzyme substrates, and eluted with imidazole. The resulting complexes of PA bound by its cytotoxic substrates inserted into lipid nanodiscs provide soluble samples that take random orientation allowing for single-particle cryo-EM analysis (Supplementary Fig. 1). The available space on the heptameric PA channel for EF binding can only accommodate up to three EF molecules due to steric hindrance. Indeed, 2D classification of cryo-EM images of PA channel bound with EF showed that the EF binding varies in different classes (Supplementary Fig. 1a, b), suggesting that the space is not fully occupied. Therefore, we used a symmetry expansion method in Relion for 3D classification and were able to resolve the asymmetrically attached EF (Methods, Supplementary Fig. 2). Remarkably, the same symmetry expansion method also worked for cryo-EM images of PA channel with LF bound even though the 2D classification failed to classify the asymmetrically bound LF in the PA-LF complex (Methods, Supplementary Figs. 1c, d and 3). In total, we determined four structures: one for the LF in complex with the heptameric PA channel and three for the EF(s) in complex with the heptameric PA channel, at an average resolution of 4.6 Å and 3.2–3.4 Å, respectively (Supplementary Figs. 2 and 3), based on the “gold-standard” Fourier shell correlation (FSC) 0.143 cutoff criterion31,32. The resulting maps revealed a “flower-on-a-stem” heptameric channel with 27-Å wide β barrel, consistent with the channel conformation. In all our structures, the conformation of the PA channel remains largely unchanged from the previously determined structure of the PA channel without substrate bound (PDB 3J9C18) (Figs. 1a and 2a). Atomic models of LF and EF were built into cryo-EM density maps. Only one LF is visible in the LF binding complex (refer to as PA7-LF), while there are three configurations of the EF-bound structures: one EF and two isoforms of two EF in the EF binding complexes (referred to as PA7-EF; PA7-1,3-EF; PA7-1,4-EF) (Supplementary Fig. 2). Regardless of this, we could clearly observe that an amino-terminal helix of both LF and EF binds to the α clamp of the heptameric PA channel in all complex structures; meanwhile, the rest of the amino-terminal domains of both LF and EF are well-ordered (Supplementary Fig. 4). These structures reveal how the enzymatic factors bind to the PA channel to form a complex and how the subunits in the complex interact with one another in preparation for the translocation process.Fig. 1 Structure of LF-bound PA7 channel (PA7-LF). a Two orthogonal views of the overall PA7-LF structure. b Structure comparison of substrate-binding α clamp between PA7-LF channel and PA8-(LFN)4 pre-channel. Fig. 2 Structure of EF-bound PA7 channel (PA7-EF). a Overall structure of PA7-EF shown as ribbon. b Zoom-in view (view 1) of the PA7 α-clamp site showing its detailed interactions with α1 of EF. The cryo-EM density is shown as semi-transparent gray. c Rotated view 1 showing structure comparison of the substrate-binding α-clamp between PA7-EF channel (color) and PA8-(LFN)4 pre-channel (gray), except that the density is not shown for clarity. d Zoom-in view (view 2) showing the details of the PABD domain of EF binding to PAN and PAC. e Rotated view 2 showing the superposition of PA-bound EF (purple for EF, orange for PA) and PA-bound LF (gray for both PA and LF), except that the density is not shown for clarity. Hydrogen bonds are shown as dashed lines. α-clamp site from pre-channel to channel complex In our PA7-LF channel structure, the LFN (the amino terminal domain of LF) binds two neighboring PA subunits, one denoted as PAN, which binds the N-terminus of LFN, and the other as PAC, which binds the C-terminus of LFN (Fig. 1a). The cryo-EM density reveals a helix of LFN bound in the α-clamp site (Supplementary Fig. 4), indicating that this site continues to engage the enzymatic factors in the PA channel. This helix of LFN in PA7 α-clamp appears to bind within this site in much the same manner as in the PA8-(LFN)4 pre-channel structure (PDB 3KWV20) (Fig. 1b). However, at 4.6-Å resolution, it is not possible to determine whether hydrogen bonds form within the α-clamp site between LFN β1 and PAN β13. Upon alignment with 3KWV20, nonetheless, β13 in the PA channel and in the PA8-(LFN)4 pre-channel has a highly similar conformation, with the first α-helix (LFN α1) aligning well between the two structures. This alignment indicates that LFN α1 binds within the α-clamp site similarly in both the PA channel and the pre-channel. The catalytic domain of LF is invisible in our EM structure, suggesting that it is flexible. The amino terminal domain of EF (EFN) and LFN share similar structures upon binding to the PA channel (Figs. 2a and 1a). In the crystal structures free of PA binding, LF α1 is an ordered α-helix9 but the homologous region in EF is flexible and disordered33. Upon PA binding, these disordered residues of EF (residues 20–30) refold into an α-helix (EFN α1) and bind within the α-clamp site (Fig. 2b). β1 (Leu33 to Lys35) of EFN forms parallel β-sheet with β13 (Leu203 to Pro205) of PAN (Fig. 2b). The hydrogen bonds between the two β-strands are analogous to those found in the PA8-(LFN)4 pre-channel structure (PDB 3KWV20), confirming predictions that the amino terminus of EF binds in a similar way as LF20. Plasticity of helix binding within the α-clamp site While α1 and β1 in LF and EF bind to the α clamp of PA analogously with β1 forming hydrogen bonds with PAN, their α1 helices dock within the α clamp differently, indicating that there is structural plasticity of α-helix binding within this α-clamp site (Fig. 2c). LF α1 is angled downward towards the pore in the α-clamp site, while the amino-terminal end of EF α1 is elevated ~2.9 Å as measured using the carbonyl groups on LF Glu34 and EF Glu24. This elevation in EF α1 appears to be caused by a change in the orientation of PAN Phe464. The phenyl ring in PAN Phe464 is positioned outwards toward the bound EF α1. This change in the orientation of Phe464 appears to restrict EF α1 in its elevated conformation in the α-clamp site. Overall, this structural plasticity makes sense, given previous work determining that the α clamp in PA binds α helices repeatedly and nonspecifically during translocation of its substrates27. Interface destabilization may play role in translocation A hydrophobic interface is formed in the carboxy-terminal subdomain of EFN with PA between EF residues Val223, Leu226, Tyr227 and PAC residues Phe202, Pro205, Ile207, Ile210 (Fig. 2d). Like those in the PA8-(LFN)4 pre-channel complex crystal structure previously determined20, PAC Ile210 and EF Tyr227 are well packed in this hydrophobic interface, allowing the phenol hydroxyl to form hydrogen bonds within the interface with PAC residues His211 and Asp195 (Fig. 2d). Despite the above similarity, it is worth noting that, upon undergoing the conformational change from pre-channel to channel, the substrate appears to have moved up (Fig. 2e), away from the binding interface with PAC compared to the pre-channel structure20. This conformational change has resulted in a loss of salt bridges that had previously formed upon binding of the LF to the pre-channel between residues PAC Lys213 and Lys214 and LF Asp187 and Asp184, respectively, where the distance between PAC Lys213 and LF Asp187 increases from 3.5 Å in the pre-channel structure to 4.6 Å in the PA7-EF channel structure; the distance between PAC Lys214 and LF Asp184 increases from 2.8 Å in the pre-channel structure to 4.3 Å in the PA7-EF channel structure (Fig. 2e). The loss of these salt bridges in the binding interface of PA7-LF should destabilize the binding interface, preparing the substrate for subsequent dissociation and unfolding prior to its translocation. Indeed, when we mutate EF residues Asp171 and Asp174 to alanine, we see no change in binding affinity compared to wild type using planar bilayer electrophysiology (Supplementary Fig. 6). This result indicates that the salt bridges are weakened significantly once PA reaches the channel state. Thus, the LF/EF binding interface with PA can be maintained in a higher affinity mode when PA is in the pre-channel conformation and complex assembly is more important; but when PA converts into the channel state, the affinity of the LF/EF binding interface is destabilized, allowing for more rapid dissociation and unfolding of LF/EF during translocation. EF domains reorganize upon binding the PA channel Unlike LF, all the domains of EF are well resolved in our channel complex structures (Fig. 3a, b). The corresponding amino acid sequence for the different domains with respective α helices and β sheets is shown in Fig. 3c. In all our structures of the EF-bound channel (PA7-EF, PA7-1,3-EF, and PA7-1,4-EF) presented here, EF undergoes a significant conformation change compared to its calmodulin (CaM)-bound structure33. In the previous CaM-bound EF structure (PDB: 1XFY33), CaM stabilizes the CA and CB ACD and the HD of EF, and there are no significant interactions among these domains of EF (Fig. 3e). While in the PA7-bound EF structures, the HD domain contributes to a new conformation by bridging the PABD and ACD. Further analysis indicated that the folding pattern within the three domains only changes slightly from CaM-bound EF to PA7-bound EF, but the three domains are reorganized in PA7-bound EF (Fig. 3e and Supplementary Movie 1).Fig. 3 Structural comparison of EF between its PA7-bound and CaM-bound forms. a Domain architecture of EF with individual domains colored and the boundary residues numbered. b Structure of PA7-EF with EF shown as ribbon and PA7 as surface colored by protomers. The three domains of EF—PABD, ACD, HD—are colored as in a. c Sequence and secondary structures of the PA7-bound EF. d Close-up view at the interactions among PABD, ACD and HD domains in PA7-bound EF. The structural elements involved in domain interactions are highlighted and hydrogen bonds are shown as dashed lines. e Superposition of EF structures in its PA7-bound and CaM-bound (PDB: 1XFY) forms. The two EFs are aligned by the PABD domain for clarity. Three domains of CaM-bound EF—PABD, ACD, HD—are colored in green, light blue and cyan, respectively. Domain reorganizations are marked by arrows. In more detail, on one side of the HD, residues near α29 and α30 of HD interact with those near α2 and β1 of the PABD. Hydrogen bonds are formed between inter-domain residues, one from HD and the other from PABD, such as Gln746-Asn40, Lys767-Gln50, Asn737-Ile71, Asn737-Phe73, and Glu739-Phe73 (Fig. 3d). On the other side of HD, a loop between α26 and α27 interacts with residues near α22 and α24 of ACD, mainly through inter-domain hydrogen bonds (Fig. 3d). With the extensive interactions mentioned above, HD moves toward and binds PABD eventually. Notably, the refolding of N-terminal residues (Lys20 to Thr42) of PABD (Fig. 3e), which is a consequence of PA7 binding, yields the space that enables the interactions between HD and PABD (Fig. 3d, pink arrow in Fig. 3e, and Supplementary Movie 1), leading to a 60° swing of ACD (yellow arrow in Fig. 3e), which mounts α22 and α24 of ACD on the loops near α25, α26, and α27 of HD (Fig. 3d). Discussion Here we report a total of four cryo-EM structures of heptameric PA channel bound with toxin substrates: three for the complex with EF at resolutions ranging from 3.2 to 3.4 Å and one for the complex with LF at 4.6 Å resolution. Our results reveal that upon the binding of the substrate to the PA channel, conformational changes occur in the enzymatic substrates LF and EF. When full-length LF binds to the PA channel, its catalytic domains exhibit significant flexibility, and thus only the PA binding domain, LFN, is visible in the cryo-EM density map; by contrast, the PA-binding and catalytic domains are visible in the crystal structure of LF9. In the case of EF, its domains reorganize, compared to the EF structure bound to CaM33. This CaM-bound structure is the only other full-length structure of EF available for comparison, but highlights the conformational changes EF undergoes during its lifetime. It is interesting that EF binds to PA in a different way than it binds to CaM, with the helical domain stabilizing its PABC and ACD. This reorganization involves refolding of PABD residues (Lys20 to Thr42), a 70° swing of HD toward PABD and mounting of ACD to HD. The reorganized conformation of EF is stabilized by the formation of hydrogen bonds. We suggest that this reorganization of the domains plays a role in the ordered translocation of the EF through the channel. Previously, Feld et al. showed in detail that different substrates could bind to the α clamp20, indicating nonspecific binding at the α-clamp site. Our results also show the α clamp engages different α helices, either from EF or LF. Interestingly, PA’s Phe464, a residue lining the α clamp, changes conformation to accommodate different residues in these helices. These results demonstrate plasticity within the α-clamp site, which allows for the binding of different helical substrates. When bound to the pre-channel, LFN forms numerous stabilizing interactions on its amino and carboxyl terminal subdomains. Upon conversion to the channel conformation, the carboxyl terminal subdomain of LFN destabilizes its interface with PA. This destabilization occurs while the complex is exposed to the acidic pH of the endosomal compartment. This interface destabilization, paired with the acidic environment, most likely plays an important role in allowing the bound substrate to unfold and translocate through the channel more efficiently. Our four high-resolution structures of PA channel with LF and EF—representing the structures of the complex in the channel conformation—provide further insights into the mechanism of how substrate proteins are translocated across membranes by the PA channel. In our current model (Fig. 4), the enzymatic factors bind to PA pre-channels, before the cell undergoes endocytosis. The PA prechannel undergoes a conformational change within the endosomal compartment, forming the channel state. This conformational change results in an alteration of the binding interaction between the channel and its substrate enzymes, thus destabilizing the interaction. This destabilization, accompanied by partial protonation of the polypeptides, allows the proton gradient to drive translocation of the bound substrate through the channel. As the polypeptide is translocated through the channel, it is engaged by the α-clamp repeatedly and non-specifically27. During much of the translocation process, the polypeptide is accommodated by the channel in its secondary structure. It is engaged as an α helix while binding within the α clamp. As it moves down and is bound in the ϕ-clamp site, the α-clamp engages the polypeptide again. When the α clamp re-engages with the polypeptide, it causes an allosteric change in the ϕ clamp24. This change in the ϕ clamp applies force to the α-helix, changing its conformation to extended chain and driving it past the charge clamp site. Once past the ϕ clamp, the polypeptide is deprotonated within the anionic charge clamp. This prevents retro-translocation of the polypeptide chain back toward the endosome. At this point the polypeptide can begin to reform its secondary structure. Once exiting the channel, the translocating polypeptide refolds into its tertiary structure and can perform its enzymatic effects on the host cytosol. In the case of EF, this involves binding CaM and taking on its CaM-bound domain organization33.Fig. 4 Mechanism of EF translocation. Illustration of the anthrax toxin channel translocation steps with EF. Initially, EF binds to the PA pre-channel, and the N-terminal α helix of the PABD of EF docks into the α clamp, yielding the space for domain reorganization of EF. After the PA pre-channel changes to the channel state at low pH, the destabilization of the interface between the PABD of EF and the PA channel allows the N-terminal α helix to translocate down to the ϕ-clamp site. In parallel, the α clamp engages the EF polypeptide again, causing an allosteric change in the ϕ clamp. The change in the ϕ clamp applies force to the α helix, changing its conformation to extended chain and driving it past the charge clamp site located near the top of the β barrel. The cycle repeats on the next section of EF polypeptide. Recently, structures of the substrate-engaged SecY protein translocon have been determined using X-ray crystallography and cryoEM34,35. The SecY system is one of the few other protein translocation systems where structural information is available. Like PA, within the SecY channel there is a hydrophobic pore ring that interacts with the translocating polypeptide. In addition, a polypeptide clamp has been identified in SecA which would position the translocating polypeptide right above the SecY pore36. The recent structure of the clamp bound to the translocating substrate indicates that it engages with the polypeptide in a sequence-independent manner by inducing short β strand conformations in the polypeptide35. This action would allow a broad range of polypeptides to be bound and translocated by the SecA. Hence this clamp is like the α clamp in PA, which also engages multiple sequences. This similarity suggests that perhaps there are universally shared phenomenon amongst different translocons, in which substrate is engaged sequence-independently based on secondary structure. In general, these two translocons allow different polypeptide segments to be engaged repeatedly and nonspecifically as they translocate through their respective channels. Methods Protein expression and purification Heptameric PA oligomer (PA7) was prepared as described17. Briefly, PA83 was expressed in Escherichia coli BL21(DE3) using a pET22b plasmid directing expression to the periplasm. PA83 was extracted from the periplasm and further purified using Q-Sepharose anion-exchange chromatography in 20 mM Tris-chloride, pH 8.0, and eluted with a gradient of 20 mM Tris-chloride, pH 8.0 with 1 M NaCl. PA83 was then treated with trypsin (1:1000 wt/wt trypsin:PA) for 30 min at room temperature to form PA63. The trypsin was inhibited with soybean trypsin inhibitor at 1:100 dilution (wt/wt soybean trypsin inhibitor:PA). The trysinized PA was subjected to anion-exchange chromatography to isolate the oligomerized PA7. The trypsinized PA was applied to the anion exchange column in 20 mM Tris-chloride, pH 8.0, and the oligomerized PA7 was eluted from the anion exchange column using a gradient of 20 mM Tris-chloride, 1 M sodium chloride, pH 8.0. Recombinant WT LF and WT EF and EF point mutants, containing an amino-terminal six-histidine His-tag (His6) were overexpressed in Escherichia coli BL21(DE3) from pET15b constructs and purified from the cytosol using His6 affinity chromatography. Cytoplasmic lysates of His6-LF and His6-EF were made by treatment with hen egg white lysozyme for 30 min at room temperature. The lysates were briefly sonicated at 4 °C (for 2 min) to shear genomic DNA and reduce sample turbidity. His6-LF and His6-EF lysates were applied to immobilized nickel affinity chromatography column in 20 mM Tris-chloride, 35 mM imidazole, 1 M sodium chloride pH 8.0, and His6-LF and His6-EF were eluted using a gradient of 20 mM Tris-chloride, 500 mM imidazole, 1 M sodium chloride pH 8.0. Affinity-purified His6-LF and His6-EF were then subjected to S200 gel filtration chromatography in 20 mM Tris-chloride, 150 mM sodium chloride, pH 8.0. EF point mutants were made using the Quik-Change mutagenesis kit (Stratagene) according to the manufacturers procedure with the primer designs listed in Supplementary Table 1. PA-LF and PA-EF complex assembly His6-LF or His6-EF were mixed with PA7 pre-channel at a ratio of 5:1 (LF/EF:PA7) and allowed to assemble on ice for 1 h. The PA7 pre-channel in complex with His6-LF and His6-EF was then purified over S400 gel filtration in 20 mM Tris-chloride pH 8.0, 150 mM sodium chloride. Nanodisc insertion The His6 tag was removed from membrane scaffold protein 1D1 (MSP1D1)28. pMSP1D1 was a gift from Stephen Sligar (Addgene plasmid #20061). In all, 300 µL wet volume Ni-NTA Superflow resin (Qiagen) was added to an 800-µL centrifuge column (Pierce) twice with 50 mM sodium chloride, 50 mM Tris-chloride pH 7.5 (Buffer A). In all, 300 µL of 1 µM of our PA complex and 300 µL of 2 M urea were added to the resin, for a final urea concentration of 1 M. This mix was collected and incubated at 37 °C for 5 min to induce conversion from the pre-channel to channel conformation29. The mix was then collected and added back into a centrifuge column, and the resin (now bound to complex) was washed twice with 500 µL Buffer A. A mixture containing MSP1D1 and palmitoyloleoyl phosphocholine (POPC) was made by first evaporating chloroform off of POPC, then adding MSP1D1 and sodium cholate in Buffer A. The final concentration contained 4 µM MSP1D1, 400 µM POPC, and 25 mM sodium cholate in Buffer A. In all, 500 µL of a MSP1D1-(POPC) mix was added to the dry resin bound with PA complex30. This resin slurry was then collected and dialyzed in Slide-A-Lyzer cassette (10 kDa molecular weight cut-off) (Thermo Scientific) in excess Buffer A for 8–12 h at a time, with two buffer changes. The Ni-NTA was then collected after dialysis. The resin was washed twice with 500 µL Buffer A. The resin was then washed with 500 µL of 50 mM NaCl, 50 mM imidazole, 50 mM Tris pH 7.5 to elute any remaining proteins bound non-specifically. The nanodisc complex was then eluted in 50 mM sodium chloride, 300 mM imidazole, 50 mM Tris-chloride pH 7.5. This eluted sample was then dialyzed into Buffer A and concentrated to 0.274 mg ml−1 (PA channel in complex with LF) and 0.498 mg ml−1 (PA channel in complex with EF) Concentration was estimated by a Nanodrop spectrophotometer. Cryo-EM sample preparation and imaging For cryo-EM sample optimization, an aliquot of 2.5 μl of sample was applied onto a glow-discharged holey carbon copper grid (300 mesh, QUANTIFOIL® R 2/1). The grid was blotted and flash-frozen in liquid ethane with an FEI Mark IV Vitrobot. An FEI TF20 cryo-EM instrument was used to screen grids. Cryo-EM grids with optimal particle distribution and ice thickness were obtained by varying the gas source (air or H2/O2), time for glow discharge, the volume of applied samples, chamber temperature/humidity, blotting time/force. For the PA channel in complex with LF, our best grids were obtained using H2/O2 for glow discharge and with the Vitrobot sample chamber set at 12 °C temperature and 100% humidity. For the PA channel in complex with EF, our best grids were obtained using air for glow discharge and with the Vitrobot sample chamber set at 16 °C temperature and 100% humidity. Optimized cryo-EM grids were loaded into an FEI Titan Krios electron microscope with a Gatan Imaging Filter (GIF) Quantum LS device and a post-GIF K2 Summit direct electron detector. The microscope was operated at 300 kV with the GIF energy-filtering slit width set at 20 eV. Movies were acquired with Leginon37 by electron counting in super-resolution mode at a pixel size of 0.535 Å per pixel. A total number of 45 frames were acquired in 9 s for each movie, giving a total dose of ~60 e−/Å2/movie. Drift correction for movie frames Frames in each movie were aligned for drift correction with the graphics processing unit (GPU)-accelerated program MotionCor238. The first frame was skipped during drift correction due to concern of more severe drift/charging of this frame. Two averaged micrographs, one with dose weighting and the other without dose weighting, were generated for each movie after drift correction. The averaged micrographs have a calibrated pixel size of 1.07 Å on the specimen scale. The averaged micrographs without dose weighting were used only for defocus determination and the averaged micrographs with dose weighting were used for all other steps of image processing. Structure determination for PA channel in complex with EF For the PA channel in complex with EF, the defocus value of each averaged micrograph was determined by CTFFIND439 generating values ranging from −1.5 to −3 μm. Initially, a total of 1,481,285 particles were automatically picked from 6811 averaged images without reference using Gautomatch (http://www.mrc-lmb.cam.ac.uk/kzhang). The particles were boxed out in dimensions of 256 × 256 square pixels square before further processing by the GPU accelerated RELION2.1. Several iterations of reference-free 2D classification were subsequently performed to remove bad particles (i.e., classes with fuzzy or un-interpretable features), yielding 725,251 good particles. The reported map of the heptameric anthrax toxin PA channel18 (EMD-6224) was low-pass filtered to 60 Å to serve as an initial model for 3D classification. After one round of 3D classification with C7 symmetry, only the classes showing feature corresponding to the intact PA7 channel were kept, which contained 486,169 particles. We re-centered those particles and removed duplications based on the unique index of each particle given by RELION32. The resulting 486,169 particles were applied one round of auto-refinement by RELION, yielding a map with an average resolution of 3.0 Å. Next, we expanded C7 symmetry to C1, yielding 3,403,183 (486,169 × 7) particles. These particles were submitted to further classification (skip align) with 29 classes. A cylinder mask was created only for the EF binding region (Supplementary Fig. 2) and applied for the focus classification. Among these 29 classes, four different types of density maps were identified. Four classes have no clear density of EF (PA7), 14 classes show clear density of only one EF binds to the PA7 channel (PA7-EF), six classes with density of two EF (PA7-1,3-EF), and four classes with density of two EF which were located further away from each other (PA7-1,4-EF) (Supplementary Fig. 2). Subsequently, we merged the particles from classes belonging to PA7-EF, PA7-1,3-EF, PA7-1,4-EF, respectively. After removing duplications based on the unique particle names given by RELION, we got 333,455 particles for PA7-EF (68.8% of all particles), 72,864 particles for PA7-1,3-EF (15.0% of all particles) and 73,784 particles for PA7-1,4-EF (15.1% of all particles). The unique particles of each dataset (PA7-EF, PA7-1,3-EF, PA7-1,4-EF) resulting from the focused classification were subjected to a final step of 3D auto-refinement with C1 symmetry. The two half maps of each dataset from this auto-refinement step were subjected to RELION’s standard post-processing procedure. The final maps of PA7-EF, PA7-1,3-EF, PA7-1,4-EF achieved an average resolution of 3.2, 3.4, and 3.4 Å, respectively, based on RELION’s gold-standard FSC (see below). Structure determination for PA channel in complex with LF For the PA channel in complex with LF, the defocus value of each averaged micrograph was determined by CTFFIND439 generating values ranging from −1.5 to −3 μm. Initially, a total of 616,153 particles were automatically picked from 2502 averaged images without reference using Gautomatch (http://www.mrc-lmb.cam.ac.uk/kzhang). The particles were boxed out in dimensions of 320 × 320 square pixels square and binned to 160 × 160 square pixels (pixel size of 2.14 Å) before further processing by the GPU accelerated RELION2.1. Several iterations of reference-free 2D classification were subsequently performed to remove bad particles (i.e., classes with fuzzy or un-interpretable features), yielding 204,395 good particles. The reported map of the heptameric anthrax toxin PA channel18 (EMD-6224) was low-pass filtered to 60 Å to serve as an initial model for 3D classification. After one round of 3D classification with C7 symmetry, only the classes showing feature corresponding to the intact PA7 channel were kept, which contained 194,849 particles. We re-centered those particles and removed duplications based on the unique index of each particle given by RELION. The resulting 194,775 particles were un-binned to 320 × 320 square pixels (pixel size of 1.07 Å) and applied one round of auto-refinement by RELION, yielding a map with an average resolution of 3.4 Å. The C7 symmetry was then expanded to C1, giving 1,363,425 (194,775 × 7) particles for further classification. A cylinder mask was created only for the LF-binding region (Supplementary Fig. 3) and applied for the focus classification with seven classes. Six of the seven classes show clear density for only one LF bound to the PA7 channel (PA7-LF) (Supplementary Fig. 3). We next merged the good particles from the six classes and removed duplications based on the unique particle names given by RELION. The 63,807 un-binned, unique particles (10.4% of all particles) resulting from the focused classification were subjected to a final step of 3D auto-refinement with C1 symmetry. The two half maps from this auto-refinement step were subjected to RELION’s standard post-processing procedure. The final map of the PA7-LF complex has an average resolution of 4.6 Å based on RELION’s gold-standard FSC. We also got a 3D auto-refinement result (3.6 Å) with C7 symmetry using this dataset, which helped the model building process (see model building below). Resolution assessment All resolutions reported above are based on the “gold-standard” FSC 0.143 criterion40. FSC curves were calculated using soft spherical masks and high-resolution noise substitution was used to correct for convolution effects of the masks on the FSC curves41. Prior to visualization, all maps were sharpened by applying a negative B-factor, which was estimated using automated procedures20. Local resolution was estimated using ResMap42. The overall quality of the maps for the PA channel in complex with EF and LF is presented in Supplementary Figs. 2 and 3, respectively. Data collection and reconstruction statistics are presented in Supplementary Table 2. Model building and refinement Atomic model building was accomplished in an iterative process involving Coot43, Chimera44, and Phenix45. For the PA7-LF complex, the structure of anthrax toxin PA channel heptamer18 (PDB ID: 3J9C) was fitted into cryo-EM map (4.6 Å, C1 symmetry) by using the ‘fit in map’ routine in Chimera. The atomic model building of PA7 channel was facilitated by using the 3.6 Å cryo-EM map in C7 symmetry (63,807 particles, Supplementary Fig. 2). Next, the crystal structure of LF9 (PDB ID: 1J7N) was fitted in to the cryo-EM map (4.6 Å, C1 symmetry) to create a full atomic model for PA7-LF. Finally, the structure was manually adjusted using Coot and refined using Phenix in real space with secondary structure and geometry restraints. For the PA channel in complex with EF, we have three different types of density maps—PA7-EF, PA7-1,3-EF, and PA7-1,4-EF. Owing to the higher resolution and single EF binding in PA7-EF, we firstly carried out model building on this density map. The structure of PA7 channel18 (PDB ID: 3J9C), was fitted into the cryo-EM map of PA7-EF as initial model by using the ‘fit in map’ routine in Chimera. This fit revealed the extra density corresponding to EF. However, further docking showed the density of EF in cryo-EM map has significant differences with respect to the crystal structure of EF33 (PDB ID: 1XFX). The full-length EF consists of four domains, the (PABD), two catalytic core domains CA and CB forming the ACD, and the HD. The domains in the cryo-EM map have a different arrangement, however. Thus, we fit the domains into the density separately to create an initial atomic model for PA7-EF, which was refined by “real-space refinement” in Phenix. We then manually adjusted the main chain and side chains to match the cryo-EM density map with Coot. This process of real space refinement and manual adjustment steps was repeated until the peptide backbone and side chain conformations were optimized. Secondary structure and geometry restraints were used during the refinement. Refinement statistics of the PA channel in complex with LF and EF are summarized in Supplementary Table 2. These models were also evaluated based on MolProbity scores46 and Ramachandran plots (Supplementary Table 2). Representative densities for the proteins are shown in Supplementary Fig. 4. Planar lipid bilayer electrophysiology apparatus Planar lipid bilayer currents were recorded using an Axopatch 200B amplifier and a Digidata 1440 A acquisition system (Molecular Devices Corp., Sunnyvale, CA)17,47. Ensemble recordings were recorded at 200 Hz and filtered at 100 Hz using PCLAMP10 software. The membrane potential difference is defined as Δψ ≡ ψcis − ψtrans (ψtrans ≡ 0 V). Ensemble binding analysis using electrophysiology A prior method20 was used to monitor EF binding to PA channels at symmetrical pH and a Δψ of 0 mV by means of an applied potassium chloride gradient. Membranes were painted on a 100 μm aperture of a 1-mL, white-Delrin cup with 3% (wt/vol) 1,2-diphytanoyl-sn-glycerol-3-phosphocholine (Avanti Polar Lipids, Alabaster, AL) in n-decane (Sigma-Aldrich, St. Louis, MO); and the cis chamber was bathed in 10 mM potassium phosphate, 100 mM potassium chloride, pH 7. During the setup, the trans chamber was bathed in 10 mM potassium phosphate, pH 7. PA channels were inserted by adding 20 nmol of PA7 to the cis chamber at pH 7. PA currents reached ~5 nA. Upon stabilization of the ensemble current, the cis chamber was perfused to exchange fresh 10 mM phosphate, 100 mM KCl at pH 7. EF and mutants thereof were added in small increments to the cis side of the membrane, allowing for binding to reach equilibrium as indicated by the observed decrease in current which reached a steady-state plateau. Fraction of closed channels (θobs) versus [P] plots (where P denotes free EF) were fit to a simple single-site model, θobs = 1/(1 + KD/[P]), to obtain an equilibrium dissociation constant, KD. Three to four independent measurements of KD for each EF mutant and wild type were made and averages and standard deviations were computed in Microcal ORIGIN9 software. Reporting summary Further information on research design is available in the Nature Research Reporting Summary linked to this article. Supplementary information Supplementary Information Peer Review File Description of Additional Supplementary Files Supplementary Movie 1 Reporting Summary Source data Source Data Supplementary information Supplementary information is available for this paper at 10.1038/s41467-020-14658-6. Acknowledgements We thank J. Jiang for their suggestions about sample preparation and data processing, Y. Cui for assistance in cryo-EM and suggestions about data processing, I. Atanasov and W. Hui for assistance in cryo-EM. This work was supported in part grants from the National Science Foundation (NSF, under grant no. DMR-1548924) and by grants from the National Institutes of Health (R01GM071940/AI094386/DE025567 to Z.H.Z. and R21AI124020 to B.K.) and the Training Program in Integrative Membrane Biology at the University of Maryland, Baltimore (T32GM008181). We acknowledge the use of resources in the Electron Imaging Center for Nanomachines supported by UCLA and grants from the NIH (S10RR23057, S10OD018111, and U24GM116792) and NSF (DBI-1338135). K.Z. acknowledges support from the China Scholarship Council. Author contributions Z.H.Z. and B.K. conceived the project; N.J.H. engineered and isolated samples; S.L. and K.Z. evaluated the samples, performed electron microscopy, processed the data, built atomic models, and prepared figures; N.J.H and K.G. performed equilibrium binding electrophysiology experiments; all authors wrote the paper. Data availability The cryo-EM maps have been deposited in the Electron Microscopy Data Bank under accession numbers EMD-20459, EMD-20955, EMD-20957, and EMD-20958. The atomic structure coordinates have been deposited in the Protein Data Bank under the accession numbers 6PSN [10.2210/pdb6PSN/pdb], 6UZB [10.2210/pdb6UZB/pdb], 6UZD [10.2210/pdb6UZD/pdb], and 6UZE [10.2210/pdb6UZE/pdb]. The source data underlying Supplementary Fig. 6 are provided as a Source Data file. Other data can be obtained from the corresponding authors upon reasonable request. Competing interests The authors declare no competing interests. Peer review information Nature Communications thanks Doryen Bubeck, Rodney Tweten and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. These authors contributed equally: Nathan J. Hardenbrook, Shiheng Liu, Kang Zhou. Change history 7/13/2023 In the original version of this article, the given and family names of Z. Hong Zhou were incorrectly structured. The name was displayed correctly in all versions at the time of publication. The original article has been corrected. ==== Refs References 1. Wickner W Schekman R Protein translocation across biological membranes Science 2005 310 1452 1456 10.1126/science.1113752 16322447 2. Krantz BA Finkelstein A Collier RJ Protein translocation through the anthrax toxin transmembrane pore is driven by a proton gradient J. Mol. Biol. 2006 355 968 979 10.1016/j.jmb.2005.11.030 16343527 3. Matouschek A Protein unfolding—an important process in vivo? Curr. Opin. Struct. Biol. 2003 13 98 109 10.1016/S0959-440X(03)00010-1 12581666 4. Zhang S Udho E Wu Z Collier RJ Finkelstein A Protein translocation through anthrax toxin channels formed in planar lipid bilayers Biophys. J. 2004 87 3842 3849 10.1529/biophysj.104.050864 15377524 5. Martin A Baker TA Sauer RT Pore loops of the AAA+ ClpX machine grip substrates to drive translocation and unfolding Nat. Struct. Mol. Biol. 2008 15 1147 1151 10.1038/nsmb.1503 18931677 6. Goyal P Structural and mechanistic insights into the bacterial amyloid secretion channel CsgG Nature 2014 516 250 253 10.1038/nature13768 25219853 7. Berg BVanDen X-ray structure of a protein-conducting channel Nature 2004 427 36 44 10.1038/nature02218 14661030 8. Young JAT Collier RJ Anthrax toxin: receptor binding, internalization, pore formation, and translocation Annu. Rev. Biochem. 2007 76 243 265 10.1146/annurev.biochem.75.103004.142728 17335404 9. Pannifer AD Crystal structure of the anthrax lethal factor Nature 2001 414 229 233 10.1038/n35101998 11700563 10. Duesbery NS Proteolytic inactivation of MAP-kinase-kinase by anthrax lethal factor Science 1998 280 734 737 10.1126/science.280.5364.734 9563949 11. Vitale G Bernardi L Napolitani G Mock M Montecucco C Susceptibility of mitogen-activated protein kinase kinase family members to proteolysis by anthrax lethal factor Biochem. J. 2000 352 739 745 10.1042/bj3520739 11104681 12. Vitale G Anthrax lethal factor cleaves the N-terminus of MAPKKs and induces tyrosine/threonine phosphorylation of MAPKs in cultured macrophages Biochem. Biophys. Res. Commun. 1998 248 706 711 10.1006/bbrc.1998.9040 9703991 13. Drum CL Structural basis for the activation of anthrax adenylyl cyclase exotoxin by calmodulin Nature 2002 415 396 402 10.1038/415396a 11807546 14. Ulmer TS Calcium dependence of the interaction between calmodulin and anthrax edema factor J. Biol. Chem. 2003 278 29261 29266 10.1074/jbc.M302837200 12724328 15. Tang WJ Guo Q The adenylyl cyclase activity of anthrax edema factor Mol. Asp. Med. 2009 30 423 430 10.1016/j.mam.2009.06.001 16. Moayeri M Leppla SH Vrentas C Pomerantsev AP Liu S Anthrax Pathogenesis Annu. Rev. Microbiol. 2015 69 185 208 10.1146/annurev-micro-091014-104523 26195305 17. Kintzer AF The protective antigen component of anthrax toxin forms functional octameric complexes J. Mol. Biol. 2009 392 614 629 10.1016/j.jmb.2009.07.037 19627991 18. Jiang J Pentelute BL Collier RJ Zhou ZH Atomic structure of anthrax protective antigen pore elucidates toxin translocation Nature 2015 521 545 549 10.1038/nature14247 25778700 19. Krantz BA Trivedi AD Cunningham K Christensen KA Collier RJ Acid-induced unfolding of the amino-terminal domains of the lethal and edema factors of anthrax toxin J. Mol. Biol. 2004 344 739 756 10.1016/j.jmb.2004.09.067 15533442 20. Feld GK Structural basis for the unfolding of anthrax lethal factor by protective antigen oligomers Nat. Struct. Mol. Biol. 2010 17 1383 1390 10.1038/nsmb.1923 21037566 21. Lacy DB Wigelsworth DJ Melnyk, R. a, Harrison, S. C. & Collier, R. J. Structure of heptameric protective antigen bound to an anthrax toxin receptor: a role for receptor in pH-dependent pore formation Proc. Natl Acad. Sci. USA 2004 101 13147 13151 10.1073/pnas.0405405101 15326297 22. Song L Structure of staphylococcal α-hemolysin, a heptameric transmembrane pore Science 1996 274 1859 1866 10.1126/science.274.5294.1859 8943190 23. Thoren KL Krantz BA The unfolding story of anthrax toxin translocation Mol. Microbiol. 2011 80 588 595 10.1111/j.1365-2958.2011.07614.x 21443527 24. Krantz BA A phenylalanine clamp catalyzes protein translocation through the anthrax toxin pore Science 2005 309 777 781 10.1126/science.1113380 16051798 25. Wynia-Smith SL Brown MJ Chirichella G Kemalyan G Krantz BA Electrostatic ratchet in the protective antigen channel promotes anthrax toxin translocation J. Biol. Chem. 2012 287 43753 43764 10.1074/jbc.M112.419598 23115233 26. Lacy DB A model of anthrax toxin lethal factor bound to protective antigen Proc. Natl Acad. Sci. USA 2005 102 16409 16414 10.1073/pnas.0508259102 16251269 27. Das D Krantz BA Peptide- and proton-driven allosteric clamps catalyze anthrax toxin translocation across membranes Proc. Natl Acad. Sci. USA 2016 113 9611 9616 10.1073/pnas.1600624113 27506790 28. Grinkova YV Lazarides AA Sligar SG Directed self-assembly of monodisperse phospholipid bilayer nanodiscs with controlled size J. Am. Chem. Soc. 2004 126 3477 3487 10.1021/ja0393574 15025475 29. Akkaladevi N Assembly of anthrax toxin pore: Lethal-factor complexes into lipid nanodiscs Protein Sci. 2013 22 492 501 10.1002/pro.2231 23389868 30. Katayama H Three-dimensional structure of the anthrax toxin pore inserted into lipid nanodiscs and lipid vesicles Proc. Natl Acad. Sci. USA 2010 107 3453 3457 10.1073/pnas.1000100107 20142512 31. Rosenthal PB Henderson R Optimal determination of particle orientation, absolute hand, and contrast loss in single-particle electron cryomicroscopy J. Mol. Biol. 2003 333 721 745 10.1016/j.jmb.2003.07.013 14568533 32. Scheres SHW RELION: implementation of a Bayesian approach to cryo-EM structure determination J. Struct. Biol. 2012 180 519 530 10.1016/j.jsb.2012.09.006 23000701 33. Shen Y Zhukovskaya NL Guo Q Florián J Tang WJ Calcium-independent calmodulin binding and two-metal-ion catalytic mechanism of anthrax edema factor EMBO J. 2005 24 929 941 10.1038/sj.emboj.7600574 15719022 34. Li L Crystal structure of a substrate-engaged SecY protein-translocation channel Nature 2016 531 395 399 10.1038/nature17163 26950603 35. Ma C Structure of the substrate-engaged SecA-SecY protein translocation machine Nat. Commun. 2019 10 1 9 10.1038/s41467-019-10918-2 30602773 36. Zimmer J Nam Y Rapoport TA Structure of a complex of the ATPase SecA and the protein-translocation channel Nature 2008 455 936 943 10.1038/nature07335 18923516 37. Carragher B Leginon: an automated system for acquisition of images from vitreous ice specimens J. Struct. Biol. 2000 132 33 45 10.1006/jsbi.2000.4314 11121305 38. Zheng SQ MotionCor2: anisotropic correction of beam-induced motion for improved cryo-electron microscopy Nat. Methods 2017 14 331 332 10.1038/nmeth.4193 28250466 39. Rohou A Grigorieff N CTFFIND4: fast and accurate defocus estimation from electron micrographs J. Struct. Biol. 2015 192 216 221 10.1016/j.jsb.2015.08.008 26278980 40. Scheres SHW Chen S Prevention of overfitting in cryo-EM structure determination Nat. Methods 2012 9 853 854 10.1038/nmeth.2115 22842542 41. Chen S High-resolution noise substitution to measure overfitting and validate resolution in 3D structure determination by single particle electron cryomicroscopy Ultramicroscopy 2013 135 24 35 10.1016/j.ultramic.2013.06.004 23872039 42. Kucukelbir A Sigworth FJ Tagare HD Quantifying the local resolution of cryo-EM density maps Nat. Methods 2014 11 63 65 10.1038/nmeth.2727 24213166 43. Emsley P Cowtan K Coot: model-building tools for molecular graphics Acta Crystallogr. Sect. D. Biol. Crystallogr. 2004 60 2126 2132 10.1107/S0907444904019158 15572765 44. Pettersen EF UCSF chimera—a visualization system for exploratory research and analysis J. Comput. Chem. 2004 25 1605 1612 10.1002/jcc.20084 15264254 45. Adams, P. D. et al. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr. Sect. D Biol. Crystallogr. 66, 213–221 (2010). 46. Chen VB MolProbity: all-atom structure validation for macromolecular crystallography Acta Crystallogr. Sect. D. Biol. Crystallogr. 2010 66 12 21 10.1107/S0907444909042073 20057044 47. Thoren KL Worden EJ Yassif JM Krantz BA Lethal factor unfolding is the most force-dependent step of anthrax toxin translocation Proc. Natl Acad. Sci. USA 2009 106 21555 21560 10.1073/pnas.0905880106 19926859
PMC007xxxxxx/PMC7265647.txt
==== Front Mol Cancer Mol Cancer Molecular Cancer 1476-4598 BioMed Central London 32487167 1215 10.1186/s12943-020-01215-4 Research Circular RNA circCORO1C promotes laryngeal squamous cell carcinoma progression by modulating the let-7c-5p/PBX3 axis Wu Yongyan 12345 Zhang Yuliang 12 Zheng Xiwang 12 Dai Fengsheng 13 Lu Yan 6 Dai Li 13 Niu Min 12 Guo Huina 12 Li Wenqi 13 Xue Xuting 12 Bo Yunfeng 7 Guo Yujia 12 Qin Jiangbo 8 Qin Yixiao 13 Liu Hongliang 129 Zhang Yu 410 Yang Tao 5 Li Li 9 Zhang Linshi 11 Hou Rui 12 Wen Shuxin 13 An Changming [email protected] 14 Li Huizheng [email protected] 15 Xu Wei [email protected] 161718 http://orcid.org/0000-0001-7836-2851 Gao Wei [email protected] 12349 1 grid.263452.4 0000 0004 1798 4018 Shanxi Key Laboratory of Otorhinolaryngology Head and Neck Cancer, Shanxi Medical University, Taiyuan, 030001 Shanxi People’s Republic of China 2 grid.452461.0 0000 0004 1762 8478 Shanxi Province Clinical Medical Research Center for Precision Medicine of Head and Neck Cancer, The First Hospital of Shanxi Medical University, Taiyuan, 030001 Shanxi People’s Republic of China 3 grid.452461.0 0000 0004 1762 8478 Department of Otolaryngology Head & Neck Surgery, The First Hospital of Shanxi Medical University, Taiyuan, 030001 Shanxi People’s Republic of China 4 grid.263452.4 0000 0004 1798 4018 Key Laboratory of Cellular Physiology, Ministry of Education, Shanxi Medical University, Taiyuan, 030001 Shanxi People’s Republic of China 5 grid.263452.4 0000 0004 1798 4018 Department of Biochemistry & Molecular Biology, Shanxi Medical University, Taiyuan, 030001 Shanxi People’s Republic of China 6 grid.454145.5 0000 0000 9860 0426 Department of Otolaryngology Head & Neck Surgery, The First Hospital, Jinzhou Medical University, Jinzhou, 121001 Liaoning People’s Republic of China 7 grid.263452.4 0000 0004 1798 4018 Department of Pathology, Shanxi Cancer Hospital, Shanxi Medical University, Taiyuan, 030013 Shanxi People’s Republic of China 8 grid.254020.1 0000 0004 1798 4253 Department of Otolaryngology Head & Neck Surgery, Heping Hospital Affiliated to Changzhi Medical College, Changzhi, 046000 Shanxi People’s Republic of China 9 grid.263452.4 0000 0004 1798 4018 Department of Cell Biology and Genetics, Basic Medical School of Shanxi Medical University, Taiyuan, 030001 Shanxi People’s Republic of China 10 grid.263452.4 0000 0004 1798 4018 Department of Physiology, Shanxi Medical University, Taiyuan, 030001 Shanxi People’s Republic of China 11 grid.13402.34 0000 0004 1759 700X Department of Hepatobiliary and Pancreatic Surgery, The Second Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou 310009 Zhejiang, People’s Republic of China 12 grid.1012.2 0000 0004 1936 7910 Harry Perkins Institute of Medical Research, QEII Medical Centre and Centre for Medical Research, the University of Western Australia, PO Box 7214, 6 Verdun Street, Nedlands, Perth, Western Australia 6009 Australia 13 grid.263488.3 0000 0001 0472 9649 General Hospital, Shenzhen University, Shenzhen, 518055 Guangdong People’s Republic of China 14 grid.506261.6 0000 0001 0706 7839 Department of Head and Neck Surgery, Cancer Hospital, National Cancer Center, Chinese Academy of Medical Sciences & Peking Union Medical College, Beijing, 100021 People’s Republic of China 15 grid.411971.b 0000 0000 9558 1426 Department of Otolaryngology Head & Neck Surgery, Dalian Municipal Friendship Hospital, Dalian Medical University, Dalian, 116100 Liaoning People’s Republic of China 16 grid.27255.37 0000 0004 1761 1174 Shandong Provincial ENT Hospital Affiliated to Shandong University, Jinan, 250022 Shandong People’s Republic of China 17 Shandong Provincial Institute of Otolaryngology, Jinan, 250022 Shandong People’s Republic of China 18 grid.27255.37 0000 0004 1761 1174 Key Laboratory of Otolaryngology, Ministry of Health, Shandong University, Jinan, 250022 Shandong People’s Republic of China 2 6 2020 2 6 2020 2020 19 9920 1 2020 11 5 2020 © The Author(s) 2020 https://creativecommons.org/licenses/by/4.0/ Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Background Laryngeal squamous cell carcinoma (LSCC) is a common malignant tumor of the head and neck. LSCC patients have seriously impaired vocal, respiratory, and swallowing functions with poor prognosis. Circular RNA (circRNA) has attracted great attention in cancer research. However, the expression patterns and roles of circRNAs in LSCC remain largely unknown. Methods RNA sequencing was performed on 57 pairs of LSCC and matched adjacent normal mucosa tissues to construct circRNA, miRNA, and mRNA expression profiles. RT-PCR, qPCR, Sanger sequencing, and FISH were undertaken to study the expression, localization, and clinical significance of circCORO1C in LSCC tissues and cells. The functions of circCORO1C in LSCC were investigated by RNAi-mediated knockdown, proliferation analysis, EdU staining, colony formation assay, Transwell assay, and apoptosis analysis. The regulatory mechanisms among circCORO1C, let-7c-5p, and PBX3 were investigated by luciferase assay, RNA immunoprecipitation, western blotting, and immunohistochemistry. Results circCORO1C was highly expressed in LSCC tissues and cells, and this high expression was closely associated with the malignant progression and poor prognosis of LSCC. Knockdown of circCORO1C inhibited the proliferation, migration, invasion, and in vivo tumorigenesis of LSCC cells. Mechanistic studies revealed that circCORO1C competitively bound to let-7c-5p and prevented it from decreasing the level of PBX3, which promoted the epithelial–mesenchymal transition and finally facilitated the malignant progression of LSCC. Conclusions circCORO1C has an oncogenic role in LSCC progression and may serve as a novel target for LSCC therapy. circCORO1C expression has the potential to serve as a novel diagnostic and prognostic biomarker for LSCC detection. Keywords circCORO1C Let-7c-5p PBX3 Laryngeal squamous cell carcinoma Epithelial–mesenchymal transition Metastasis http://dx.doi.org/10.13039/501100001809 National Natural Science Foundation of China 81872210 81802793 81802948 Wu Yongyan Liu Hongliang Gao Wei http://dx.doi.org/10.13039/501100010031 Postdoctoral Research Foundation of China 2016M591412 2017M610174 Wu Yongyan Gao Wei The Excellent talent science and technology innovation project of Shanxi Province201605D211029 201705D211018 201805D211007 Wu Yongyan Wen Shuxin Gao Wei Youth Science and Technology Research Fund of Shanxi Province201901D211486 201901D211490 Wu Yongyan Guo Huina Shanxi Province Scientific and Technological Achievements Transformation Guidance Foundation201604D131002 201604D132040 201804D131043 Wen Shuxin Gao Wei Youth Foundation of The First Hospital Affiliated with Shanxi Medical UniversityYQ1503 Gao Wei Youth Top Talent Program Fund of Shanxi Province20180610 Wu Yongyan Fund of Shanxi “1331” Projectissue-copyright-statement© BioMed Central Ltd., part of Springer Nature 2020 ==== Body pmcBackground Laryngeal squamous cell carcinoma (LSCC) is a common malignant tumor of the head and neck originating from the laryngeal mucosal epithelium. LSCC accounts for approximately 2.4% of systemic malignancies worldwide each year; in 2018, around 95,000 people died of laryngeal cancer [1]. The onset of LSCC is occult, and approximately 60% of patients are in the advanced stages when diagnosed (clinical stages III and IV) [2]. The proneness of LSCC to local invasion and cervical lymph node metastasis seriously interferes with patient survival rates [3]. Surgery remains the main treatment approach for LSCC [4]. Unfortunately, LSCC is one of the few tumors with a decreasing survival rate in recent years, and its 5-year survival rate has declined from 66 to 63% over the past 40 years [5], which is mainly attributed to its unclear mechanism of occurrence and progression. Therefore, it is urgent to reveal the pathogenesis of LSCC, identify biomarkers for its diagnosis, and investigate effective new therapeutic targets. Circular RNA (circRNAs) is a recently identified non-coding RNA that has become the latest hotspot in cancer research. The circRNA molecule has a closed loop structure that is not affected by exonucleases and is not easily degraded. circRNAs also have features of high conservation and abundance [6]. Hence, circRNAs have unique advantages as biomarkers for disease diagnosis and prognosis. Recent studies have shown that circRNA molecules are rich in miRNA binding sites and can specifically bind miRNAs, thereby eliminating the inhibitory effect of miRNAs on target genes and upregulating the expression level of target genes, that is, functioning as competing endogenous RNA (ceRNA) [7]. circRNAs also bind to RNA binding proteins and may translate proteins to exert their functions [8, 9]. circRNAs have critical regulatory effects in the occurrence and development of a variety of cancers, affecting cell cycle, apoptosis, metabolism, invasion, and metastasis [10]. circAGFG1 upregulates CCNE1 expression and promotes the proliferation, migration, and invasion of breast cancer cells [11]. circPPP1R12A-encoded protein circPPP1R12A-73aa promotes tumor growth and metastasis of colon cancer [12]. However, to date, little is known about the expression, functions, and regulatory mechanisms of circRNAs in LSCC. Pre-B-cell leukemia homeobox transcription factor 3 (PBX3) is a member of the evolutionarily conserved three-amino acid-loop-extension (TALE) homeodomain transcription factor family. A recent study revealed that PBX3 is a critical regulatory protein of the epithelial–mesenchymal transition (EMT) network in colorectal cancer [13]. Dysregulation of PBX3 expression has been observed in many cancer types, such as prostate, gastric, cervical, and liver cancer [14–17]. Nonetheless, the expression and function of PBX3 in LSCC are still unknown. In this study, we performed RNA sequencing of 57 pairs LSCC tissues and matched adjacent normal mucosa (ANM) tissues and identified abnormally upregulated expression of circCORO1C in LSCC tissues. Furthermore, the expression of circCORO1C was strongly associated with the clinical features and prognosis of LSCC patients. We found that circCORO1C could bind to let-7c-5p and prevent it from decreasing the level of PBX3, which promoted EMT and stimulated the proliferation, migration, and invasion of LSCC cells in vitro and in vivo. Methods LSCC patient tissue A total of 164 pairs of LSCC tissues and matched ANM tissues (taken 1–3 cm from the edge of cancer tissues) were obtained from patients undergoing surgery at the Department of Otolaryngology Head and Neck Surgery, The First Hospital of Shanxi Medical University, from January 2013 to January 2017. None of the patients received chemotherapy or radiotherapy before surgery. The tissue samples were diagnosed independently by two experienced clinical pathologists. The histological types of LSCC were determined according the World Health Organization (WHO) system, and TNM (Tumor, Node, Metastasis) stage was defined according to the criteria of the American Joint Committee on Cancer (AJCC, 8th edition). Fresh specimens were immediately frozen in liquid nitrogen. Among the 164 pairs of tissue samples, 57 paired LSCC (Additional file 1: Table S1) and ANM tissues were used for RNA sequencing, and 107 paired samples for qPCR analysis (Additional file 1: Table S2). Cell lines and cell culture Human LSCC cell line FD-LSC-1 (a gift from Professor Liang Zhou [18]) was cultured in BEGM™ Bronchial Epithelial Cell Growth Medium (Lonza, Walkersville, MD, USA) supplemented with 10% FBS (Biological Industries, CT, USA). Human LSCC cell line TU-177 purchased from Bioleaf Biotech Corporation (Shanghai, China) was maintained in DMEM supplemented with 10% FBS. Human HEK293T and MRC-5 cell lines were purchased from the China Center for Type Culture Collection (CCTCC). HEK293T cells were cultured in DMEM with 10% FBS. MRC-5 cells were cultured in MEM with 10% FBS. Human oral keratinocytes (HOK) purchased from ScienCell Research Laboratories (Carlsbad, CA) were cultured in DMEM with 10% FBS. All cells were cultured at 37 °C with 5% CO2. Cell lines were tested for mycoplasma contamination using the TransDetect PCR Mycoplasma Detection Kit (TransGen Biotech, Beijing, China). RNA and genomic DNA (gDNA) extraction Total RNA was extracted from tissues or cells using Trizol reagent (Invitrogen, Waltham, MA) following the manufacturer’s instructions. The nuclear and cytoplasmic fractions were extracted using a PARIS kit (ThermoFisher Scientific, Waltham, MA). gDNA was extracted using a genomic DNA isolation kit (TIANGEN Biotech (Beijing) Co., Ltd., Beijing, China). RNA sequencing analysis The RNA integrity of 57 pairs of LSCC/matched ANM tissues was examined with a Bioanalyzer 2100 (Agilent, Santa Clara, CA). High-quality RNA (RIN > 7) samples were subjected to library construction, and then each library was sequenced on an Illumina HiSeq 4000 (circRNA and mRNA) and Illumina HiSeq 2000 (miRNA) following the standard procedures by Novogene (Beijing, China). RNA sequencing data were deposited at GEO and are accessible via accession numbers GSE127165 and GSE133632. Differentially expressed circRNAs, miRNAs, and mRNAs were screened as reported [19] (Additional file 1: Table S3–5). RT-PCR and quantitative real-time PCR (qPCR) For PCR of mRNA and circRNA, RNA was reverse-transcribed using a HiScript II 1st Strand cDNA Synthesis Kit (Vazyme, Nanjing, China). For qPCR of miRNA, cDNA was synthesized using an All-in-One™ miRNA First-Strand cDNA Synthesis Kit (GeneCopoeia, Rockville, MD). qPCR was performed using ChamQ SYBR qPCR Master Mix (Vazyme, Nanjing, China) on an ABI Stepone Plus system. The relative expression levels were calculated using the 2 (−△△CT) method. The circRNA and mRNA levels were normalized by 18 s rRNA. The miRNA level was normalized against U6 small nuclear RNA. Primer sequences are listed in Additional file 1: Table S6. RNase R treatment Total RNA (2 μg) was incubated for 10 min at 37 °C with or without 3 U/μg RNase R (Geneseed Biotech Co., Ltd., Guangzhou, China), followed by RNA purification using a RNeasy MinElute Cleanup kit (Qiagen, Hilden, Germany) and analyzed by RT-PCR. Agarose gel electrophoresis PCR products were separated by 2% agarose gel electrophoresis with TAE buffer using a 100 bp DNA ladder (TransGen Biotech, Beijing, China). The bands were photographed under an Azure C600 imager (Azure Biosystems, Dublin, CA). Fluorescence in situ hybridization (FISH) Cy3-labeled circCORO1C probes (5′- AGAGCAATTGGTTCCTGCATATTTTTCTGGCAATCTCACATTTGTTAACATC -3′) were synthesized by Sangon Biotech (Shanghai, China). FISH was performed using a FISH kit (RiboBio, Guangzhou, China) according to the manufacturer’s instructions. Nuclei were stained with DAPI. Images were acquired on a Leica TCS SP8 confocal laser scanning microscope (Leica Microsystems Inc., Buffalo Grove, IL). Plasmid construction and cell transfection The PBX3 overexpression plasmid was generated by inserting PBX3 CDS sequence into the p3 × FLAG-CMV-10 vector (Sigma-Aldrich, St. Louis, MO). shRNA lentiviral plasmid targeting circCORO1C (psh-circCORO1C) was constructed by inserting annealed shRNA template DNA sequence into the pLKO.1 vector. For luciferase reporter plasmids, the sequences of circCORO1C, wild type, and let-7c-5p binding site mutant PBX3 3′ UTR were cloned into the psiCHECK-2 vector (Promega, Madison, WI). Cells were transfected using Lipofectamine 3000 (Invitrogen) according to the manufacturer’s instructions. siRNAs, miRNA mimics, and inhibitor siRNAs targeting circCORO1C (si-circCORO1C #1: 5′-AGAUUGCCAGAAAAAUAUGCA-3′; si-circCORO1C #2: 5′-UUGCCAGAAAAAUAUGCAGGA-3′), negative control siRNAs (si-NC), stable oligonucleotides (modified by cholesterol, 2′-OMe and phosphorothioate), let-7c-5p mimics, and NC were synthesized by GenePharma (Shanghai, China). miRNA inhibitor is small, chemically modified single-stranded RNA molecule that can competitively bind to and inhibit the function of specific endogenous mature miRNA. 2′-OMe-modified let-7c-5p inhibitor and NC inhibitor were synthesized by GenePharma. Generation of circCORO1C knockdown cells To generate FD-LSC-1 cells with stable knockdown of circCORO1C, lentiviruses were produced in HEK293T cells by cotransfection with psh-circCORO1C and packaging plasmids GAG and VSVG. Virus supernatant was harvested 48 h after transfection, mixed with polybrene (8 μg/ml), and added to FD-LSC-1 cells. After 48 h incubation, 2 μg/ml puromycin (Santa Cruz Biotechnology, Dallas, TX) was added for 1 week to screen for stable cell clones. CCK8 assay After 24 h transfection, cells were digested and seeded into 96-well plates (3 × 103/well). At 0, 24, 48, 72, and 96 h after seeding, each well was replaced with 100 μL fresh complete medium and 10 μL TransDetect CCK (TransGen Biotech, Beijing, China) followed by incubation at 37 °C with 5% CO2 for 1 h. The absorbance of the solution was measured at 450 nm using a Spectra Max i3x Multifunctional microplate detection system (Molecular Devices, San Jose, CA). 5-Ethynyl-2′-deoxyuridine (EdU) staining Cells were incubated with DMEM medium containing 50 μM EdU (RiboBio) at 37 °C with 5% CO2 for 2 h. Cells were washed twice with PBS, fixed with 50 μL 4% paraformaldehyde for 30 min, neutralized with 50 μL 2 mg/mL glycine solution and permeabilized by adding 100 μL 0.5% Triton X-100. After washing with PBS, 100 μL 1 × Apollo dye was added to each well, then cells were incubated at room temperature for 30 min. Next, 100 μL Hoechst 33342 was added and incubated for another 30 min. Images were captured and analyzed on an ImageXpress high-content screening system (Molecular Devices). Colony formation assay Transfected cells were seeded at a density of 600 cells/well into a 35-mm dish and then cultured for 10 days. Cells were washed with PBS once and colonies were fixed with 4% paraformaldehyde for 20 min and stained with 0.1% crystal violet solution for 10 min at room temperature, followed by image capture. Transwell migration and invasion assays After 24 h transfection, cells were digested, washed twice with PBS and resuspended in serum-free DMEM. Transwell chambers for invasion assay were precoated with Matrigel (BD Biosciences, San Jose, CA). Serum-free DMEM (200 μL) containing cells (4 × 104 cells/well for migration assay, 1 × 105 cells/well for invasion assay) was added to the upper chamber. Then 500 μL DMEM medium supplemented with 20% FBS was added to the lower chamber. After 24 h, cells in the upper chamber were removed with cotton swabs and the lower side of the chamber was gently washed twice with PBS, fixed with 4% paraformaldehyde for 20 min, and stained with 0.1% crystal violet for 10 min, and then images were captured by microscope. Apoptosis analysis Apoptosis was determined using a Dead Cell Apoptosis kit (ThermoFisher Scientific). Briefly, cells were digested with EDTA-free trypsin and washed with ice-cold PBS, followed by a 15-min incubation with Alexa Fluor 488 annexin V and PI, then cells were analyzed by a NovoCyte flow cytometer (ACEA Biosciences, Hangzhou, China). Prediction of RNA interaction Target gene prediction of let-7c-5p was performed using the ENCORI online program with strict stringency (http://starbase.sysu.edu.cn/index.php). The interaction between circCORO1C and miRNA was predicted by seedVicious v1.0 and RegRNA 2.0 (https://seedvicious.essex.ac.uk/predict.html, http://regrna2.mbc.nctu.edu.tw/index.html). TCGA data analysis Transcriptome sequencing data and clinical features of head and neck squamous cell carcinoma (HNSCC) were downloaded from The Cancer Genome Atlas (TCGA) HNSCC cohort (https://portal.gdc.cancer.gov/projects/TCGA-HNSC), followed by expression analysis of PBX3 and let-7c-5p with normalized FPKM and RPM values. RNA immunoprecipitation (RIP) RIP experiments were performed with a Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore, Billerica, MA) according to the manufacturer’s instructions. Briefly, 1 × 107 cells were collected and resuspended in 300 μL RIPA lysis buffer containing protease inhibitor cocktail and RNase inhibitors. The cell lysates (200 μL) were incubated with 5 μg AGO2 antibody (#2897; CST, Danvers, MA) or rabbit IgG and protein A/G magnetic beads at 4 °C overnight with rotation. Immunoprecipitated RNA was purified using a RNeasy MinElute Cleanup kit (Qiagen). The enrichment of circCORO1C was evaluated by qPCR. Luciferase reporter assay HEK293T cells were cotransfected with luciferase reporter plasmid and let-7c-5p mimics or NC mimics for 48 h. The luciferase activity was measured using a dual luciferase reporter assay system (Promega) on a Spectra Max i3x Multifunctional microplate detection system (Molecular Devices). The luciferase values were normalized and then the relative luciferase activity was calculated. Western blotting Total protein was extracted with RIPA buffer containing protease inhibitor cocktail (ThermoFisher Scientific). The protein concentration was determined using a Coomassie (Bradford) Protein Assay Kit (ThermoFisher Scientific). Equal amounts of total protein (30 μg) were separated by 12% SDS-PAGE and transferred onto PVDF membranes (Millipore, Billerica, MA), followed by blocking with 5% skim milk. The membranes were incubated with antibodies against PBX3 (#12571-1-AP; Proteintech, Wuhan, China), E-cadherin (#3195S; CST, Danvers, MA), N-cadherin (#13116S; CST), Vimentin (#5741S; CST), Slug (#sc-166476; Santa Cruz Biotechnology), or GAPDH (#HC301-02; TransGen Biotech) overnight at 4 °C. Then membranes were washed three times with TBST followed by secondary antibody incubation for 2 h at room temperature. Bands were detected by a chemiluminescence imaging system (SageCreation Science, Beijing, China) with Western Bright ECL HRP substrate (Advansta Inc., San Jose, CA). Xenograft tumorigenesis SPF-grade male BALB/C nude mice (6-8 weeks) were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (Beijing, China) and housed under SPF conditions (TECNIPLAST S.p.A., Italy). A total of 2 × 106 FD-LSC-1 cells were suspended in 200 μL serum-free DMEM and subcutaneously injected into the right flank of each mouse. The volumes of tumors were measured from 7 days after injection. Tumor volume was calculated as follows: V (volume) = (length × width2)/2. After 25 days, the mice were killed and the tumors were dissected, weighed, and processed for histological analysis. Immunohistochemical (IHC) staining IHC staining was performed as previously described [3]. In brief, tissues were fixed in 4% (v/v) formaldehyde in PBS, embedded in paraffin, and cut into 3-μm sections. Sections underwent dewaxing, re-hydration, antigen retrieval, and blocking, and then were incubated with antibodies against PBX3, Ki67, E-cadherin, N-cadherin, and Vimentin overnight at 4 °C in a moist chamber, and washed three times with PBST. Sections were incubated with HRP-conjugated secondary antibody (CST) for 15 min at room temperature, washed three times with PBST, and then stained with DAB and hematoxylin. Next, sections were dehydrated and mounted with coverslips. Statistical analysis Statistical analysis was performed using GraphPad Prism 7.0 software (La Jolla, USA). Comparisons between two groups were performed using the two-tailed Student’s t-test. Correlations were analyzed by Pearson’s correlation. Kaplan-Meier survival curve and log-rank test were employed to depict the overall survival probability of LSCC patients with different expression levels of circCORO1C. Results are presented as mean ± standard deviation (SD). P values of < 0.05 were considered statistically significant. Results circCORO1C is frequently upregulated in LSCC and is associated with malignant progression and poor prognosis We performed RNA sequencing in 57 pairs of LSCC and matched ANM tissues. Differential expression screening showed that 410 circRNAs were upregulated in LSCC tissues (Additional file 1: Table S3), in which 18 circRNAs were detected in all sequenced tissues (Fig. 1a). Expression of the 18 circRNAs was verified by RT-PCR and Sanger sequencing, and 12 circRNAs were validated successfully (Additional file 2: Figure S1a and b). Next, we screened circRNAs that affect LSCC proliferation by siRNA-mediated knockdown and high-content screening. We found that knockdown of circRNA hg19_circ_0008714 significantly inhibited LSCC cell proliferation (Additional file 2: Figure S1c). Hence, we focused on this circRNA in this study. Sequence analysis revealed that hg19_circ_0008714 was formed by back-splicing of exons 7 and 8 of the Coronin-like actin-binding protein 1C gene (CORO1C) and was therefore named circCORO1C (Fig. 1b). RT-PCR and Sanger sequencing were performed to verify the expression and head-to-tail splicing of circCORO1C in LSCC (Fig. 1b). Moreover, we compared the expression levels of circCORO1C in LSCC cells and normal control cell lines by qPCR. The results showed that the expression levels of circCORO1C in LSCC cell lines FD-LSC-1 and TU-177 were significantly higher than those in normal control cell lines HEK293T, HOK, and MRC-5 (Fig. 1c). Fig. 1 circCORO1C is upregulated in LSCC tissues and is associated with poor prognosis. a RNA sequencing of 57 pairs of LSCC and matched adjacent normal mucosal (ANM) tissues to screen differentially expressed circRNAs. Heatmap showing circRNAs expressed in all tissues and those upregulated in LSCC tissues. b Schematic illustration showed the circularization of CORO1C exons 7 and 8 to form circCORO1C. The back-splicing junction of circCORO1C was verified by RT-PCR and Sanger sequencing. c circCORO1C expression levels in human LSCC cell lines (FD-LSC-1, TU-177) and normal cell lines (HOK, HEK293T, MRC-5) were determined by qPCR. d circCORO1C expression in FD-LSC-1 and TU-177 cells verified by RT-PCR. Agarose gel electrophoresis showed that divergent primers amplified circCORO1C in cDNA but not genomic DNA (gDNA). GAPDH served as a negative control. e Validation of circCORO1C stability by RNase R treatment and RT-PCR analysis. f qPCR analysis of the abundance of circCORO1C and linear CORO1C in FD-LSC-1 and TU-177 cells treated with actinomycin D at the indicated times. g Expression levels of circCORO1C in an additional 107 paired LSCC tissues were determined by qPCR. h–j Correlation analysis between circCORO1C expression levels and clinicopathological parameters of LSCC. Expression level of circCORO1C was significantly associated with T stage (h), N stage (i), and clinical stage (j). k Kaplan-Meier analysis of the correlation between circCORO1C expression and overall survival of 96 patients with LSCC. l circCORO1C abundance in nuclear and cytoplasmic fractions of FD-LSC-1 and TU-177 cells was evaluated by qPCR. 18S RNA acted as a positive control of RNA distributed in the cytoplasm, and U6 RNA acted as a positive control of RNA distributed in the nucleus. m Localization of circCORO1C in FD-LSC-1 and TU-177 cells was detected by FISH. Nuclei were stained with DAPI (blue) and circCORO1C probes were labeled with Cy3 (red). The error bars (c, f and l) represent SD of three independent experiments. **P < 0.001 Head-to-tail splicing sequences may be produced by mechanisms other than the formation of circRNA, such as trans-splicing and genomic rearrangements [20]. To rule out the possibility of the latter two, convergent primers for linear and special divergent primers for circular RNA were designed. The convergent primers could be amplified when the template contained CORO1C linear mRNA or genomic DNA (gDNA), while the divergent primers could only be specifically amplified in the presence of circCORO1C. cDNA and gDNA from FD-LSC-1 and TU-177 cells, respectively, were used as PCR templates. Nucleic acid electrophoresis results indicated that the divergent primers could amplify circCORO1C only in cDNA, but no products were detected in the gDNA (Fig. 1d). High stability is a crucial feature of circRNA. To confirm the stability of circCORO1C, RNase R was used to pretreat the RNA; the results demonstrated that linear CORO1C mRNA was significantly reduced after RNase R treatment, while circCORO1C was resistant to RNase R (Fig. 1e). Moreover, we further compared the half-life of the circular form and linear CORO1C through treatment with RNA transcription inhibitor actinomycin D and qPCR assay, and found that circCORO1C had a significantly longer half-life than the linear CORO1C (Fig. 1f). These data confirmed essential features of circCORO1C. To investigate the correlation between circCORO1C levels and LSCC, we detected the expression of circCORO1C in 107 pairs of LSCC and ANM tissues by qPCR. The relative abundance of circCORO1C in LSCC tissues was significantly higher than that in ANM tissues (Fig. 1g). Moreover, the expression level of circCORO1C was significantly correlated with T stage, N stage, and clinical stage. Patients in the advanced stage and those with cervical lymph node metastasis had high expression levels of circCORO1C (Fig. 1h–j). Importantly, Kaplan-Meier analysis revealed that LSCC patients with high circCORO1C levels had poor overall survival (Fig. 1k). Next, we investigated the location of circCORO1C in cells by nuclear and cytoplasmic RNA extraction and qPCR. The results showed that circCORO1C was mainly localized to the cytoplasm (Fig. 1l). FISH further confirmed that circCORO1C was mainly localized to the cytoplasm (Fig. 1m). These results indicated that circCORO1C upregulation is common in LSCC and may have important functions in the progression of LSCC. circCORO1C promotes the proliferation, migration, and invasion of LSCC cells To investigate the functions of circCORO1C in LSCC cells, we designed and synthesized two siRNAs that specifically targeted the back-splicing region of circCORO1C. After LSCC cell lines FD-LSC-1 and TU-177 were transfected with siRNA, qPCR was performed to evaluate the knockdown efficiency. The results showed that both siRNAs could significantly reduce circCORO1C, while the level of linear CORO1C was not significantly changed, and siRNA #1 had the highest knockdown efficiency (Fig. 2a). We also tested cell viability by CCK-8 assay, and found that knockdown of circCORO1C significantly inhibited the viability of LSCC cells (Fig. 2b). EdU staining experiments confirmed that knockdown of circCORO1C inhibited the proliferation of LSCC cells (Fig. 2c). Colony formation experiments showed that knockdown of circCORO1C significantly inhibited the colony formation of FD-LSC-1 and TU-177 cells (Fig. 2d). Fig. 2 circCORO1C promotes the proliferation, migration, and invasion of LSCC cells. a circCORO1C siRNA was transfected into FD-LSC-1 and TU-177 cells, and the expression levels of cicCORO1C and CORO1C were detected by qPCR. b FD-LSC-1 and TU-177 cells were transfected with circCORO1C siRNAs. Cell proliferation capacity was detected at the indicated time points by CCK8 assays. c FD-LSC-1 and TU-177 cells were transfected with circCORO1C siRNAs, and the changes in cell proliferation were determined by EdU staining. d circCORO1C knockdown inhibited colony formation of both FD-LSC-1 and TU-177 cells. e & f Knockdown of circCORO1C inhibited the migration and invasion of FD-LSC-1 (e) and TU-177 (f) cells as determined by Transwell migration and invasion assays. g FD-LSC-1 and TU-177 cells were transfected with circCORO1C siRNAs. Cells were stained with Annexin V-FITC and PI, and the percentage of apoptotic cells was detected by flow cytometry. Data are presented as the mean ± SD of three independent experiments. *P < 0.05; **P < 0.001 Furthermore, we investigated the effects of circCORO1C on the migration and invasion of LSCC cells by Transwell assays. Knockdown of circCORO1C significantly decreased cell migration and invasion (Fig. 2e and f). We further investigated the effect of circCORO1C on apoptosis, and found that knockdown of circCORO1C promoted apoptosis in LSCC cells (Fig. 2g). Taken together, these findings suggested that circCORO1C has an oncogenic role in LSCC, and si-circCORO1C #1 had the strongest effect on the cell functions, which was consistent with the knockdown efficiency. Thus, in subsequent studies, we performed experiments using si-circCORO1C #1. circCORO1C acts as a miRNA sponge of let-7c-5p in LSCC cells Studies have showed that circRNAs can function as miRNA sponges to competitively bind to miRNA, thus abrogating the inhibitory effect of miRNA on downstream target genes. Since circCORO1C is distributed in the cytoplasm, we studied whether it could function as a miRNA sponge in LSCC cells. We used the online tools RegRNA and seedVicious to predict the circCORO1C-binding miRNA (Additional file 1: Table S7), and intersected the data with miRNAs that were found to be downregulated in RNA sequencing of our 57 pairs of LSCC and ANM tissues. Notably, let-7c-5p was the only one common miRNA in these three datasets (Fig. 3a). Fig. 3 circCORO1C acted as a sponge for miRNA let-7c-5p in LSCC cells. a Combined analysis of bioinformatics prediction and LSCC tissue RNA sequencing data to screen for circCORO1C-binding miRNAs. b RIP assays were performed using AGO2 antibody in FD-LSC-1 and HOK cells, then the enrichment of circCORO1C was detected by qPCR. c HEK293T cells were co-transfected with let-7c-5p mimics and wild-type or mutant circCORO1C luciferase reporter vector, and luciferase reporter activity was detected. d Correlation analysis of circCORO1C and let-7c-5p RNA levels in 20 pairs of LSCC tissues. Expression of circCORO1C and let-7c-5p in 20 cases of LSCC and matched ANM tissues was determined by qPCR, and the relative expression of circCORO1C and let-7c-5p was normalized to ANM. e Expression levels of circCORO1C and let-7c-5p in FD-LSC-1 and TU-177 cells transfected with circCORO1C siRNAs were evaluated by qPCR. Data are presented as the means ± SD of three independent experiments. *P < 0.05; **P < 0.001 As one of the critical components of RNA-induced silencing complex (RISC), Argonaute 2 (AGO2) is the major protein that mediates the interaction between circRNA and target miRNAs [21]. To demonstrate that circCORO1C functions as a miRNA sponge, we performed RIP assay using AGO2 antibody in FD-LSC-1 and HOK cells. The results showed that circCORO1C was pulled down with AGO2 and was less enriched in HOK cells compared with FD-LSC-1 cells, which consistent with the differential circCORO1C levels of these two cell lines (Fig. 3b). In addition, the results of RIP in let-7c-5p-transfected FD-LSC-1 cells indicated that circCORO1C was specifically enriched by AGO2 antibody (Additional file 2: Figure S2). Moreover, we constructed luciferase reporter vectors for wild-type circCORO1C and let-7c-5p binding site mutant circCORO1C and co-transfected HEK293T cells with let-7c-5p mimics or NC mimics. The results showed that the luciferase activity in the wild-type group co-transfected with let-7c-5p mimics was significantly reduced, while the luciferase activity in the binding site mutant group was not significantly changed (Fig. 3c). Furthermore, the expression correlation between circCORO1C and let-7c-5p was analyzed by qPCR in 20 pairs of LSCC tissues, revealing that let-7c-5p levels were negatively correlated with circCORO1C levels in LSCC tissues (Fig. 3d). In addition, we found that expression of let-7c-5p in FD-LSC-1 and TU-177 cells was increased significantly after circCORO1C knockdown (Fig. 3e). Collectively, these results indicated that circCORO1C functions as a miRNA sponge to directly interact with let-7c-5p in LSCC cells. let-7c-5p is downregulated in LSCC tissues and inhibits the malignant phenotype of LSCC cells RNA sequencing data indicated that let-7c-5p expression levels in LSCC tissues were significantly lower than those in ANM tissues (Fig. 4a). Analysis of TCGA data confirmed that let-7c-5p expression was downregulated in HNSCC and LSCC (Fig. 4b and c), indicating that let-7c-5p may have important roles in LSCC. Therefore, we investigated the functions of let-7c-5p in LSCC cells. FD-LSC-1 and TU-177 cells were transfected with let-7c-5p mimics or negative control mimics (NC mimics), then the transfection efficiency was verified by qPCR, which revealed that let-7c-5p expression was elevated (Fig. 4d). CCK8 assay and EdU staining indicated that overexpression of let-7c-5p inhibited LSCC cell proliferation (Fig. 4e and f). Colony formation experiments found that the colony formation ability of LSCC cells overexpressing let-7c-5p was significantly decreased (Fig. 4g), while Transwell assay showed that the migration and invasion of LSCC cells were significantly attenuated after overexpression of let-7c-5p (Fig. 4h and i). Apoptosis assay showed that let-7c-5p overexpression promoted the apoptosis of LSCC cells (Fig. 4j). Overall, these data demonstrated that let-7c-5p inhibits the proliferation, migration, and invasion of LSCC cells and promotes their apoptosis. Fig. 4 let-7c-5p inhibited the proliferation, migration, and invasion of LSCC cells. a Expression of let-7c-5p in 57 LSCC tissues and matched ANM tissues was analyzed using RNA sequencing data. b & c Analysis of let-7c-5p expression in HNSCC (b) and LSCC (c) tissues using transcriptomic sequencing data from TCGA database. d FD-LSC-1 and TU-177 cells were transfected with let-7c-5p mimics or NC mimics, then let-7c-5p expression was determined by qPCR. e FD-LSC-1 and TU-177 cells were transfected with let-7c-5p mimics or NC mimics. Cell proliferation was detected by CCK8 assay. f FD-LSC-1 and TU-177 cells were transfected with let-7c-5p mimics or NC mimics, and cell proliferation was assessed by EdU staining. g Proliferative capacity of FD-LSC-1 and TU-177 cells transfected with let-7c-5p mimics or NC mimics was evaluated by colony formation assay. h & i. Effect of let-7c-5p on the migration and invasion of FD-LSC-1 (h) and TU-177 (i) cells was assessed by Transwell migration and invasion assays. j FD-LSC-1 and TU-177 cells were transfected with let-7c-5p mimics or NC mimics. Cells were stained with Annexin V-FITC and PI, and the percentage of apoptotic cells was detected by flow cytometry. Data are presented as the mean ± SD of three independent experiments. *P < 0.05; **P < 0.001 let-7c-5p reversed the tumor-promoting effects of circCORO1C in LSCC cells To identify whether circCORO1C promoted LSCC cell proliferation, migration, and invasion by interacting with let-7c-5p, we conducted a rescue experiment. FD-LSC-1 and TU-177 cells were cotransfected with si-circCORO1C and let-7c-5p inhibitor (Fig. 5a), and then CCK-8, EdU staining, and colony formation assays were conducted. The results showed that when let-7c-5p function was inhibited with miRNA inhibitor, the proliferation and colony formation ability of LSCC cells was significantly enhanced (Fig. 5b–d). Notably, transfection with let-7c-5p inhibitor could reverse the decreased cell viability caused by si-circCORO1C (Fig. 5b–d). Transwell assays showed that let-7c-5p inhibition reversed the reduction in the migration and invasion of FD-LSC-1 and TU-177 cells caused by circCORO1C knockdown (Fig. 5e). Compared with cells transfected with si-circCORO1C alone, co-transfection of LSCC cells with si-circCORO1C and let-7c-5p inhibitor significantly reduced apoptosis (Fig. 5f). Taken together, these results indicated that circCORO1C promoted the malignant progression of LSCC cells mainly by abolishing the anti-tumor effect of let-7c-5p. Fig. 5 let-7c-5p reversed the tumor-promoting effect of circCORO1C in LSCC cells. a FD-LSC-1 and TU-177 cells were transfected with si-circCORO1C or co-transfected with si-circCORO1C and let-7c-5p inhibitor. CircCORO1C and let-7c-5p expression was detected by qPCR. b FD-LSC-1 and TU-177 cells were transfected with si-circCORO1C or co-transfected with si-circCORO1C and let-7c-5p inhibitor. Cell proliferation was determined by CCK8 assay. c Effects of si-circCORO1C and let-7c-5p inhibitor on the proliferation of FD-LSC-1 and TU-177 cells were evaluated by EdU staining. d Colony formation assays were performed to evaluate the proliferative ability of FD-LSC-1 and TU-177 cells transfected with si-circCORO1C or co-transfected with si-circCORO1C and let-7c-5p inhibitor. e Effects of si-circCORO1C and let-7c-5p inhibitor on the migration and invasion of FD-LSC-1 and TU-177 cells were evaluated by Transwell migration and invasion assays. f FD-LSC-1 and TU-177 cells were transfected with si-circCORO1C or co-transfected with si-circCORO1C and let-7c-5p inhibitor. Cells were stained with Annexin V-FITC and PI, and the percentage of apoptotic cells was detected by flow cytometry. Data are presented as the means ± SD of three independent experiments. *P < 0.05; **P < 0.001 PBX3 is a direct target of let-7c-5p and functions as driver gene in LSCC According to ceRNA theory, circCORO1C is positively correlated with the expression of downstream target genes, while the target gene is negatively correlated with let-7c-5p expression. We predicted the possible let-7c-5p target genes by miRanda, PicTar, PITA, and TargetScan, and 257 genes intersected by these four programs were obtained (Fig. 6a; Additional file 1: Table S8). Then we intersected these 257 genes with the mRNAs that were found to be upregulated in LSCC tissues upon RNA sequencing, and 51 intersected genes were obtained (Fig. 6b; Additional file 1: Table S9). Next, we analyzed the expression correlation of circCORO1C and let-7c-5p with the 51 genes using RNA sequencing data of 57 pairs of LSCC samples. Pearson correlation analysis indicated that circCORO1C was positively correlated with PBX3, while let-7c-5p was negatively correlated with PBX3 in LSCC and ANM tissues (Fig. 6c and d). RNA sequencing data showed that PBX3 was upregulated in 73.7% (42/57) of LSCC tissues (Fig. 6ec and d). Moreover, analysis of the transcriptomic data of TCGA database found that PBX3 was upregulated in both HNSCC and LSCC (Fig. 6f). In addition, overexpression of let-7c-5p significantly decreased the expression of PBX3 mRNA and protein (Fig. 6g), while downregulation of let-7c-5p remarkably increased it in FD-LSC-1 and TU-177 cells (Fig. 6h). Fig. 6 PBX3 is a direct target gene of let-7c-5p, which acted as an oncogene in LSCC cells. a Venn analysis of the target genes of let-7c-5p predicted by miRanda, PicTar, PITA, and TargetScan. b Integrated analysis of bioinformatics-predicted target genes and RNA sequencing data of 57 pairs of LSCC tissues was performed to screen for let-7c-5p target genes. c & d Correlation analysis between circCORO1C (c) or let-7c-5p (d) and PBX3 expression using RNA sequencing data of 57 pairs of LSCC tissues and matched ANM tissues. e PBX3 expression in RNA sequencing data of 57 pairs of LSCC tissues and matched ANM tissues. The expression levels of PBX3 in each LSCC tissue were normalized to corresponding matched ANM tissue. f Analysis of PBX3 expression in HNSCC and LSCC tissues using transcriptome sequencing data from TCGA database. g & h FD-LSC-1 and TU-177 cells were transfected with let-7c-5p mimics (g), let-7c-5p inhibitor (h) or NC, and PBX3 expression was detected by qPCR and western blotting. i HEK293T cells were co-transfected with let-7c-5p mimics and wild-type or mutant PBX3 3′ UTR reporter plasmids, and luciferase reporter assays were performed to evaluate the effect of let-7c-5p on luciferase activity. j FD-LSC-1 and TU-177 cells were transfected with let-7c-5p mimics or co-transfected with let-7c-5p mimics and PBX3 overexpression plasmids, and CCK8 assay was performed to detect cell proliferation. k & l FD-LSC-1 (k) and TU-177 (l) cells were transfected with let-7c-5p mimics or co-transfected with let-7c-5p mimics and PBX3 overexpression plasmids. Changes in cell migration and invasion capacity were evaluated by Transwell assays. Data are presented as the means ± SD of three independent experiments. *P < 0.05; **P < 0.001 To demonstrate that let-7c-5p interacts directly with the 3′ UTR of PBX3, we constructed wild-type (WT) PBX3 3′ UTR and let-7c-5p binding-site mutant (Mut) luciferase reporter plasmids. The wild-type and mutant reporter vectors were co-transfected with let-7c-5p mimics in cells. Luciferase reporter assays showed that let-7c-5p mimics significantly decrease the luciferase activity of WT, while the luciferase activity of the Mut group was not significantly changed (Fig. 6i), indicating that let-7c-5p suppresses PBX3 expression by directly binding to the 3′ UTR of PBX3 mRNA. Subsequently, we investigated the functions of PBX3 in LSCC cells. Overexpression of PBX3 promoted the proliferation, migration, and invasion of LSCC cells (Fig. 6j–l). Notably, we observed that overexpression of PBX3 counteracted the inhibitory effects of let-7c-5p on LSCC cell proliferation, migration, and invasion (Fig. 6j–l). Collectively, these findings suggested that PBX3 is a driver gene and a direct target of let-7c-5p in LSCC. circCORO1C facilitates the malignant progression of LSCC cells by targeting PBX3 To investigate whether circCORO1C promoted the malignant progression of LSCC cells by regulating the downstream target gene PBX3, we simultaneously transfected PBX3 overexpression plasmid and si-circCORO1C into FD-LSC-1 and TU-177 cells (Fig. 7a) and detected changes in the cell phenotypes. CCK-8 assay and EdU staining were performed, and the results showed that overexpression of PBX3 could inhibit the decrease in cell proliferation caused by circCORO1C knockdown (Fig. 7b and c). Consistently, overexpression of PBX3 rescued the decreased colony formation ability by circCORO1C knockdown (Fig. 7d). Furthermore, Transwell assay showed that overexpression of PBX3 could reverse the decline in cell migration and invasion ability caused by circCORO1C knockdown (Fig. 7e). We also detected protein changes in EMT marker genes by western blotting. circCORO1C knockdown enhanced the expression of E-cadherin while inhibiting the expression of N-cadherin, Vimentin, and Slug (Fig. 7f), and overexpression of PBX3 could reverse the regulatory effects of circCORO1C on these EMT markers. Compared with the si-circCORO1C group, the expression of E-cadherin was reduced, and the expression of N-cadherin, Vimentin, and Slug were increased in the group cotransfected with si-circCORO1C and PBX3 overexpression plasmid (Fig. 7f). These findings indicated that circCORO1C promoted the proliferation, migration, and invasion phenotype of LSCC cells by specifically upregulating the expression of the target gene PBX3 and affecting the EMT process at the same time. Fig. 7 CircCORO1C contributed to the malignant phenotype of LSCC cells through regulating the expression of PBX3. a FD-LSC-1 and TU-177 cells were transfected with si-circCORO1C or co-transfected with si-circCORO1C and PBX3 overexpression plasmids. PBX3 expression was detected by qPCR. b & c FD-LSC-1 and TU-177 cells were transfected with si-circCORO1C or co-transfected with si-circCORO1C and PBX3 overexpression plasmids. Cell proliferation was determined by CCK8 assay (b) and EdU staining (c). d FD-LSC-1 and TU-177 cells were transfected with si-circCORO1C or co-transfected with si-circCORO1C and PBX3 overexpression plasmids. Cell proliferation ability was evaluated by colony formation assay. e Effects of circCORO1C knockdown and overexpression of PBX3 on the migration and invasion of FD-LSC-1 and TU-177 cells were determined by Transwell assays. f E-cadherin, N-cadherin, Vimentin, and Slug expression in FD-LSC-1 and TU-177 cells with knockdown of circCORO1C and overexpression of PBX3 were detected by western blotting. Data are presented as the means ± SD of three independent experiments. *P < 0.05; **P < 0.001 circCORO1C enhances the growth of xenograft tumors of LSCC cells in vivo To investigate the regulatory effect of circCORO1C on LSCC under in vivo conditions, we constructed a shRNA lentiviral plasmid targeting circCORO1C and screened FD-LSC-1 cells following stable knockdown of circCORO1C (sh-circCORO1C). Next, we constructed xenograft tumor models of nude mice by subcutaneously injecting stably transfected FD-LSC-1 cells. The xenograft tumors formed by circCORO1C-deficient LSCC cells had a significantly smaller volume than those of the control group (sh-NC) (Fig. 8a), and the tumor weight was also significantly lower than the sh-NC group (Fig. 8b). The total RNA of xenograft tumors was extracted, and qPCR was used to detect the expression of circCORO1C, let-7c-5p, and PBX3. The results confirmed decreased circCORO1C and PBX3 expression, while let-7c-5p was increased in tumors with circCORO1C knockdown (Fig. 8c). Furthermore, hematoxylin and eosin (H&E) staining showed that knockdown of circCORO1C remarkably reduced the number of lesions (Fig. 8d). IHC staining demonstrated that the expression of PBX3 and proliferation marker Ki67 was decreased in sh-circCORO1C xenograft tumors (Fig. 8e). In addition, the changes in EMT marker E-cadherin, N-cadherin, Vimentin, and Slug expression were determined by IHC. The results revealed that knockdown of circCORO1C attenuated the mesenchymal phenotype (Fig. 8e). These data confirmed that circCORO1C promoted the malignant progression of LSCC in vivo. Fig. 8 CircCORO1C promoted the tumor growth of LSCC cells in vivo. a Nude mice were subcutaneously injected with negative control (sh-NC) and shRNA-circCORO1C stably transfected FD-LSC-1 cells. After 25 days, tumors were dissected and imaged (left). Starting from day 7 after injection, the tumor volume was measured every 3 days, and the tumor growth curve was plotted (right). b Tumor weight was calculated on the day the mice were killed. Data represents mean ± SD (n = 6 each group). c Expression levels of circCORO1C, let-7c-5p, and PBX3 in xenograft tumors were determined by qPCR. d H&E staining revealed the structure of xenograft tumors derived from sh-NC and sh-circCORO1C LSCC cells. Scale bar, 200 μm. e Changes in PBX3, Ki67, E-cadherin, N-cadherin, and Vimentin expression in xenograft tumors were detected by IHC staining. Scale bar, 20 μm. f Schematic illustration of the regulation of LSCC malignant progression by the circCORO1C–let-7c-5p–PBX3 axis. *P < 0.05; **P < 0.001 Discussion Studies have shown that circRNA has important regulatory effects in a variety of biological processes, especially in the occurrence, development, and metastasis of various malignant tumors [22–24]. circRNA expression profiling revealed a series of differentially expressed circRNAs in LSCC tissues [25, 26], suggesting that circRNA may have important roles in the occurrence and progression of LSCC. In this study, we performed large-scale RNA sequencing of LSCC and matched ANM tissues, and established the circRNA, miRNA, and mRNA expression profiles of LSCC tissues. We identified and verified that circCORO1C was highly expressed in LSCC tissues and cells, and its expression levels were correlated with clinicopathological parameters and LSCC patient survival. Loss-of-function experiments demonstrated that circCORO1C promoted the proliferation, migration, and invasion of LSCC cells and inhibited their apoptosis. Mechanistic studies showed that circCORO1C bound to let-7c-5p and attenuated the inhibition of let-7c-5p on the target gene PBX3, leading to PBX3 accumulation and enhancing the proliferation, migration, and invasion of LSCC cells. The CORO1C-encoded WD repeat protein family member regulates actin-dependent processes through F-actin assembly [27]. Studies have shown that CORO1C promotes the metastases of breast cancer and lung squamous cell carcinoma [28, 29]. Cheng et al. reported that CORO1C is highly expressed in gastric cancer tissues, and in vitro experiments demonstrated that CORO1C promotes the proliferation, migration, and invasion of gastric cancer cells [30]. However, it is unclear whether circRNA is formed by CORO1C, and the roles of CORO1C-formed circRNA in disease or normal physiological processes have not yet been reported. In this study, RNA sequencing data analysis and experiments demonstrated that circCORO1C, which was highly expressed in LSCC tissues, was composed of exons 7 and 8 of CORO1C. Treatment with actinomycin D showed that the half-life of circCORO1C was significantly longer than that of linear CORO1C RNA. RNase R has 3′ to 5′ exoribonuclease activity that digests all linear RNAs except circular RNA structures [31]. When treated with RNase R, there is no significant change in circCORO1C level, proving that it has high stability as previously reported circRNA [32, 33]. Importantly, we found that the high expression of circCORO1C was positively correlated with advanced T stage, cervical lymph node metastasis, and clinical stage of LSCC, as well as poor prognosis in patients with LSCC, suggesting that circCORO1C may exert an important regulatory effect on the occurrence and development of LSCC. RNA sequencing and bioinformatics analysis indicated that circRNA has an important regulatory effect in the occurrence and development of head and neck tumors [34]. Experimental studies further demonstrated that circHIPK3 promotes cell proliferation and invasion in nasopharyngeal carcinoma [35], while Hsa_circ_0005379 inhibits the cell migration, invasion, proliferation, and in vivo tumorigenesis of oral squamous cell carcinoma [36]. CircRNAs CDR1as and hsa_circ_0023028 promote the proliferation, migration, and invasion of LSCC cells [37, 38]. Moreover, our previous studies found that the circRNA hg19_circ_0005033, which is highly expressed in LSCC stem cells, promotes proliferation, migration, invasion, and chemotherapy resistance [19]. There are very few LSCC cell lines available, among which FD-LSC-1 and TU-177 are well-characterized [18, 39]. Our data showed that expression of circCORO1C in FD-LSC-1 and TU-177 cells was higher than that in normal control cell lines. Therefore, we used these two cell lines to investigate the role of circCORO1C in LSCC cells. Consistent results showed that knockdown of circCORO1C inhibited cell proliferation, migration, invasion, and promoted apoptosis of LSCC, indicating that circCORO1C acts as an important oncogene to promote the malignant progression of LSCC. Transcripts with the same miRNA binding site, such as circRNA, mRNA, and lncRNA, regulate the expression of each other by competitively binding miRNAs. These molecules form a complex and precise post-transcriptional regulatory network, namely the ceRNA network [40]. As an important member of the ceRNA network, circRNA is involved in the formation of the circRNA–miRNA–mRNA axis, which has regulatory functions in a variety of diseases and is the most reported mechanism of action of circRNA [41–43]. In this study, we found that circCORO1C was localized to the cytoplasm, suggesting that it functions as a ceRNA [44]. let-7c-5p has been demonstrated to have anti-tumor effects in malignant tumors including non-small cell lung cancer and liver cancer [45, 46]. The combined bioinformatics prediction and transcriptomic analysis showed that let-7c-5p may bind to circCORO1C. We further demonstrated that let-7c-5p expression levels in LSCC were significantly lower than that in adjacent normal tissues, and overexpression of let-7c-5p inhibited cell proliferation, migration, and invasion in LSCC. The luciferase reporter assay and AGO2 RIP experiments demonstrated that let-7c-5p bound to circCORO1C, while rescue experiments revealed that inhibition of let-7c-5p reversed the inhibitory effect of knockdown of circCORO1C on LSCC malignant phenotypes. These findings indicated that circCORO1C sponged let-7c-5p to exert tumor-promoting functions in LSCC cells. PBX3 is highly expressed in a variety of cancer tissues, such as prostate and cervical cancer [14, 16]. Han et al. demonstrated that PBX3 expression is a critical determinant for maintaining the characteristics of tumor-initiating cells in hepatocellular carcinoma [17]. In this study, we found that PBX3 expression was upregulated in LSCC tissues, and functional studies indicated that PBX3 promoted cell proliferation, migration, and invasion in LSCC. Our data revealed that PBX3 was a direct target of let-7c-5p, and circCORO1C competitively bound to let-7c-5p and relieved the inhibitory effect of let-7c-5p on PBX3 expression, thereby upregulating PBX3 expression. We further confirmed that circCORO1C promoted the malignant progression of LSCC cells by upregulating PBX3. EMT is the basis of tumor cell migration and invasion [47, 48], and PBX3 is an essential regulator of the EMT signaling network [13]. We observed that changes in the expression levels of circCORO1C or PBX3 affected the expression of EMT markers, indicating that the circCORO1C–let-7c-5p–PBX3 axis promoted the migration and invasion of LSCC cells by regulating EMT. Finally, we demonstrated that knockdown of circCORO1C inhibited the growth of LSCC cell xenograft tumors through preclinical models and verified the regulatory relationship of the circCORO1C–let-7c-5p–PBX3 axis in vivo. In future, exploring the upstream regulator of circCORO1C and developing non-invasive circCORO1C detection methods in LSCC and other HNSCC types will be of great significance in promoting clinical translation. Conclusions In summary, our data revealed that circCORO1C competitively binds let-7c-5p to eliminate its inhibitory effect on PBX3, thereby promoting LSCC cell proliferation, migration, and invasion (Fig. 8f). High expression of circCORO1C is an important marker of poor prognosis for LSCC. These findings provide new insights into the occurrence and progression of LSCC and indicate the potential of circCORO1C as a biomarker and therapeutic target for LSCC. Supplementary information Additional file 1: Table S1. Clinical features of 57 LSCC samples for RNA sequencing. Table S2. Clinical features of 107 LSCC samples for qPCR validation. Table S3. Differentially expressed circRNAs in LSCC tissues. Table S4. Differentially expressed miRNAs in LSCC tissues. Table S5. Differentially expressed mRNAs in LSCC tissues. Table S6. Primer sequences for RT-PCR and qPCR analysis. Table S7. Prediction of circCORO1C and miRNA interaction by seedVicious. Table S8. let-7c-5p target gene prediction by ENCORI. Table S9. Intersection of predicted let-7c-5p targets and upregulated mRNAs in LSCC tissues Additional file 2: Figure S1. RNA sequencing and high-content screening reveals that circCORO1C affects the proliferation of LSCC cells. a Flowchart showing the steps for identifying functional circRNAs in LSCC. b Validation of circRNA expression in LSCC tissues by RT-PCR and Sanger sequencing. c High-content screening of circRNAs that affect the proliferation of LSCC cells. GFP-labeled FD-LSC-1 cells were transfected with siRNAs targeting the indicated circRNA. After 24 h transfection, cells were seeded into 96-well plates, and the cell number was counted at the indicated time points. Representative images (left) and fold change in cell count (right) are shown. Data are presented as the means ± SD of three independent experiments. *P < 0.05. Figure S2. FD-LSC-1 cells were transfected with let-7c-5p mimics or NC mimics for 48 h, then RIP assay was performed using AGO2 antibody and circCORO1C levels were measured by qPCR. **P < 0.001. Abbreviations LSCC Laryngeal squamous cell carcinoma circRNA Circular RNA ceRNA Competing endogenous RNA EMT Epithelial–mesenchymal transition ANM Adjacent normal mucosa qPCR Quantitative real-time PCR FISH Fluorescence in situ hybridization EdU 5-Ethynyl-2′-deoxyuridine RIP RNA immunoprecipitation 3′ UTR 3′-untranslated region HNSCC Head and neck squamous cell carcinoma Supplementary information Supplementary information accompanies this paper at 10.1186/s12943-020-01215-4. We thank Prof. Tao Bai from the Department of Pathology, The First Hospital of Shanxi Medical University for pathological analysis of LSCC sections. Authors’ contributions WG, WX, HZL, and CMA conceived the study and participated in the study design. YYW, YLZ, XWZ, FSD, YL, MN and HNG performed cell culture, colony formation, western blots, RIP, and flow cytometry assays. YYW, XWZ, LD, WQL, XTX, YFB and RH performed bioinformatics analysis, IHC, and pathological diagnosis. JBQ, YXQ, HLL, YZ, TY, and LL performed high-content screening, EdU staining and luciferase reporter assays. YYW, YLZ, XWZ, FSD, YL, LSZ and HNG contributed FISH and qPCR experiments. YYW, WQL and YLZ performed xenograft experiments. YL, YFB and SXW contributed to the clinical samples collection, follow-up and clinical data analysis. YYW, YLZ and XWZ performed primer design and plasmid construction. YYW, YJG, WG, WX, HZL, and CMA analyzed the data, organized figures and wrote the manuscript. All authors read and approved the final manuscript. Funding This work was supported by the National Natural Science Foundation of China (grants: 81872210, 81802793, and 81802948), Postdoctoral Research Foundation of China (grants: 2016 M591412 and 2017 M610174), The Excellent talent science and technology innovation project of Shanxi Province (grants: 201605D211029, 201705D211018, and 201805D211007), Youth Science and Technology Research Fund of Shanxi Province (grants: 201901D211486, 201901D211490), Shanxi Province Scientific and Technological Achievements Transformation Guidance Foundation (grants: 201604D131002, 201604D132040, and 201804D131043), Youth Foundation of The First Hospital Affiliated with Shanxi Medical University (grant: YQ1503), Youth Top Talent Program Fund of Shanxi Province, Fund of Shanxi “1331” Project. Availability of data and materials RNA sequencing raw data and normalized results were deposited at GEO database (GSE127165, GSE133632). All data that support the findings of this study are available from the corresponding authors upon reasonable request. Ethics approval and consent to participate The clinical samples were obtained with the consent of patients and approved by The Medical Ethics Committee of The First Hospital of Shanxi Medical University. Informed consent per institutional guidelines was obtained from all patients who agreed to participate in this study. Animal experiments were conducted according to the Health Guide for the Care and Use of Laboratory Animals approved by the Animal Experimental Research Ethics Committee of Shanxi Medical University. Consent for publication Not applicable. Competing interests The authors declare that they have no competing interests. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Yongyan Wu, Yuliang Zhang, Xiwang Zheng, Fengsheng Dai and Yan Lu contributed equally to this work. Change history 7/12/2023 A Correction to this paper has been published: 10.1186/s12943-023-01819-6 Change history 3/15/2023 A Correction to this paper has been published: 10.1186/s12943-023-01756-4 ==== Refs References 1. Bray F Ferlay J Soerjomataram I Siegel RL Torre LA Jemal A Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries CA Cancer J Clin 2018 68 394 424 10.3322/caac.21492 30207593 2. Steuer CE El-Deiry M Parks JR Higgins KA Saba NF An update on larynx cancer CA Cancer J Clin 2017 67 31 50 10.3322/caac.21386 27898173 3. Gao W Zhang C Li W Li H Sang J Zhao Q Promoter methylation-regulated miR-145-5p inhibits laryngeal squamous cell carcinoma progression by targeting FSCN1 Mol Ther 2019 27 365 379 10.1016/j.ymthe.2018.09.018 30341010 4. Ang KK Zhang Q Rosenthal DI Nguyen-Tan PF Sherman EJ Weber RS Randomized phase III trial of concurrent accelerated radiation plus cisplatin with or without cetuximab for stage III to IV head and neck carcinoma: RTOG 0522 J Clin Oncol 2014 32 2940 2950 10.1200/JCO.2013.53.5633 25154822 5. Siegel RL Miller KD Jemal A Cancer statistics, 2016 CA Cancer J Clin 2016 66 7 30 10.3322/caac.21332 26742998 6. Kristensen LS Andersen MS Stagsted L Ebbesen KK Hansen TB Kjems J The biogenesis, biology and characterization of circular RNAs Nat Rev Genet 2019 20 675 691 10.1038/s41576-019-0158-7 31395983 7. Zhao X Cai Y Xu J Circular RNAs: biogenesis, mechanism, and function in human cancers Int J Mol Sci 2019 20 3926 10.3390/ijms20163926 31412535 8. Du WW Zhang C Yang W Yong T Awan FM Yang BB Identifying and characterizing circRNA-protein interaction Theranostics. 2017 7 4183 4191 10.7150/thno.21299 29158818 9. Yang Y Gao X Zhang M Novel role of FBXW7 circular RNA in repressing Glioma tumorigenesis J Natl Cancer Inst 2018 110 304 315 10.1093/jnci/djx166 28903484 10. Bach DH Lee SK Sood AK Circular RNAs in Cancer Mol Ther Nucleic Acids 2019 16 118 129 10.1016/j.omtn.2019.02.005 30861414 11. Yang R Xing L Zheng X Sun Y Wang X Chen J The circRNA circAGFG1 acts as a sponge of miR-195-5p to promote triple-negative breast cancer progression through regulating CCNE1 expression Mol Cancer 2019 18 4 10.1186/s12943-018-0933-7 30621700 12. Zheng X Chen L Zhou Y Wang Q Zheng Z Xu B Wu C Zhou Q Hu W Wu C Jiang J A novel protein encoded by a circular RNA circPPP1R12A promotes tumor pathogenesis and metastasis of colon cancer via hippo-YAP signaling Mol Cancer 2019 18 47 10.1186/s12943-019-1010-6 30925892 13. Lamprecht S Kaller M Schmidt EM Blaj C Schiergens TS Engel J Jung A Hermeking H Grünewald TGP Kirchner T Horst D PBX3 is part of an EMT regulatory network and indicates poor outcome in colorectal Cancer Clin Cancer Res 2018 24 1974 1986 10.1158/1078-0432.CCR-17-2572 29391352 14. Ramberg H Grytli HH Nygård S Wang W Ögren O Zhao S Løvf M Katz B Skotheim RI Bjartell A Eri LM Berge V Svindland A Taskén KA PBX3 is a putative biomarker of aggressive prostate cancer Int J Cancer 2016 139 1810 1820 10.1002/ijc.30220 27273830 15. Li B Zhang S Shen H Li C MicroRNA-144-3p suppresses gastric cancer progression by inhibiting epithelial-to-mesenchymal transition through targeting PBX3 Biochem Biophys Res Commun 2017 484 241 247 10.1016/j.bbrc.2017.01.084 28111340 16. Li H Sun G Liu C Wang J Jing R Wang J Zhao X Xu X Yang Y PBX3 is associated with proliferation and poor prognosis in patients with cervical cancer Onco Targets Ther 2017 10 5685 5694 10.2147/OTT.S150139 29225475 17. Han H Du Y Zhao W Li S Chen D Zhang J Liu J Suo Z Bian X Xing B Zhang Z PBX3 is targeted by multiple miRNAs and is essential for liver tumour-initiating cells Nat Commun 2015 6 8271 10.1038/ncomms9271 26420065 18. Wu CP Zhou L Gong HL Du HD Tian J Sun S Li JY Establishment and characterization of a novel HPV-negative laryngeal squamous cell carcinoma cell line, FD-LSC-1, with missense and nonsense mutations of TP53 in the DNA-binding domain Cancer Lett 2014 342 92 103 10.1016/j.canlet.2013.08.041 24001612 19. Wu Y Zhang Y Niu M Shi Y Liu H Yang D Li F Lu Y Bo Y Zhang R Li Z Luo H Cui J Sang J Xiang C Gao W Wen S Whole-Transcriptome analysis of CD133+CD144+ Cancer stem cells derived from human laryngeal squamous cell carcinoma cells Cell Physiol Biochem 2018 47 1696 1710 10.1159/000490992 29949786 20. Jeck WR Sharpless NE Detecting and characterizing circular RNAs Nat Biotechnol 2014 32 453 461 10.1038/nbt.2890 24811520 21. Li Y Chen B Huang S Identification of circRNAs for miRNA targets by Argonaute2 RNA Immunoprecipitation and luciferase screening assays Methods Mol Biol 1724 2018 209 218 22. Yin Y Long J He Q Li Y Liao Y He P Zhu W Emerging roles of circRNA in formation and progression of cancer J Cancer 2019 10 5015 5021 10.7150/jca.30828 31602252 23. Arnaiz E Sole C Manterola L Iparraguirre L Otaegui D Lawrie CH CircRNAs and cancer: biomarkers and master regulators Semin Cancer Biol 2019 58 90 99 10.1016/j.semcancer.2018.12.002 30550956 24. Hsiao KY Lin YC Gupta SK Chang N Yen L Sun HS Tsai SJ Noncoding effects of circular RNA CCDC66 promote Colon Cancer growth and metastasis Cancer Res 2017 77 2339 2350 10.1158/0008-5472.CAN-16-1883 28249903 25. Fan Y Xia X Zhu Y Diao W Zhu X Gao Z Chen X Circular RNA expression profile in laryngeal squamous cell carcinoma revealed by microarray Cell Physiol Biochem 2018 50 342 352 10.1159/000494010 30282067 26. Lu C Shi X Wang AY Tao Y Wang Z Huang C Qiao Y Hu H Liu L RNA-Seq profiling of circular RNAs in human laryngeal squamous cell carcinomas Mol Cancer 2018 17 86 10.1186/s12943-018-0833-x 29716593 27. Williamson RC Cowell CA Hammond CL Bergen DJ Roper JA Feng Y Rendall TC Race PR Bass MD Coronin-1C and RCC2 guide mesenchymal migration by trafficking Rac1 and controlling GEF exposure J Cell Sci 2014 127 4292 4307 25074804 28. Lim JP Shyamasundar S Gunaratne J Scully OJ Matsumoto K Bay BH YBX1 gene silencing inhibits migratory and invasive potential via CORO1C in breast cancer in vitro BMC Cancer 2017 17 201 10.1186/s12885-017-3187-7 28302118 29. Mataki H Enokida H Chiyomaru T Mizuno K Matsushita R Goto Y Nishikawa R Higashimoto I Samukawa T Nakagawa M Inoue H Seki N Downregulation of the microRNA-1/133a cluster enhances cancer cell migration and invasion in lung-squamous cell carcinoma via regulation of Coronin1C J Hum Genet 2015 60 53 61 10.1038/jhg.2014.111 25518741 30. Cheng X Wang X Wu Z Tan S Zhu T Ding K CORO1C expression is associated with poor survival rates in gastric cancer and promotes metastasis in vitro FEBS Open Bio 2019 9 1097 1108 10.1002/2211-5463.12639 30974047 31. Xiao MS Wilusz JE An improved method for circular RNA purification using RNase R that efficiently removes linear RNAs containing G-quadruplexes or structured 3′ ends Nucleic Acids Res 2019 47 8755 8769 10.1093/nar/gkz576 31269210 32. Zhang Z Yang T Xiao J Circular RNAs: promising biomarkers for human diseases EBioMedicine. 2018 34 267 274 10.1016/j.ebiom.2018.07.036 30078734 33. D'Ambra E Capauto D Morlando M Exploring the regulatory role of circular RNAs in neurodegenerative disorders Int J Mol Sci 2019 20 5477 10.3390/ijms20215477 31689888 34. Wang WL Yang Z Zhang YJ Lu P Ni YK Sun CF Liu FY Competing endogenous RNA analysis reveals the regulatory potency of circRNA_036186 in HNSCC Int J Oncol 2018 53 1529 1543 30066847 35. Ke Z Xie F Zheng C Chen D CircHIPK3 promotes proliferation and invasion in nasopharyngeal carcinoma by abrogating miR-4288-induced ELF3 inhibition J Cell Physiol 2019 234 1699 1706 10.1002/jcp.27041 30070690 36. Su W Wang Y Wang F Sun S Li M Shen Y Yang H Hsa_circ_0005379 regulates malignant behavior of oral squamous cell carcinoma through the EGFR pathway BMC Cancer 2019 19 400 10.1186/s12885-019-5593-5 31035951 37. Zhang J Hu H Zhao Y Zhao Y CDR1as is overexpressed in laryngeal squamous cell carcinoma to promote the tumour's progression via miR-7 signals Cell Prolif 2018 51 e12521 10.1111/cpr.12521 30182381 38. Chen X Su X Zhu C Zhou J Knockdown of hsa_circ_0023028 inhibits cell proliferation, migration, and invasion in laryngeal cancer by sponging miR-194-5p Biosci Rep 2019 39 BSR20190177 10.1042/BSR20190177 31123169 39. Liu TJ Zhang WW Taylor DL Roth JA Goepfert H Clayman GL Growth suppression of human head and neck cancer cells by the introduction of a wild-type p53 gene via a recombinant adenovirus Cancer Res 1994 54 3662 3667 8033080 40. Tay Y Rinn J Pandolfi PP The multilayered complexity of ceRNA crosstalk and competition Nature 2014 505 344 352 10.1038/nature12986 24429633 41. Bi W Huang J Nie C Liu B He G Han J Pang R Ding Z Xu J Zhang J CircRNA circRNA_102171 promotes papillary thyroid cancer progression through modulating CTNNBIP1-dependent activation of β-catenin pathway J Exp Clin Cancer Res 2018 37 275 10.1186/s13046-018-0936-7 30424816 42. Li H, Xu JD, Fang XH, Zhu JN, Yang J, Pan R, Yuan SJ, Zeng N, Yang ZZ, Yang H, Wang XP, Duan JZ, Wang S, Luo JF, Wu SL, Shan ZX. Circular RNA circRNA_000203 aggravates cardiac hypertrophy via suppressing miR26b-5p and miR-140-3p binding to Gata4. Cardiovasc Res. 2019. 10.1093/cvr/cvz215. 43. Sang Y Chen B Song X Li Y Liang Y Han D Zhang N Zhang H Liu Y Chen T Li C Wang L Zhao W Yang Q circRNA_0025202 regulates Tamoxifen sensitivity and tumor progression via regulating the miR-182-5p/FOXO3a Axis in breast Cancer Mol Ther 2019 27 1638 1652 10.1016/j.ymthe.2019.05.011 31153828 44. Liu F Zhang H Xie F Tao D Xiao X Huang C Wang M Gu C Zhang X Jiang G Hsa_circ_0001361 promotes bladder cancer invasion and metastasis through miR-491-5p/MMP9 axis Oncogene 2020 39 1696 1709 10.1038/s41388-019-1092-z 31705065 45. Zhao B Han H Chen J Zhang Z Li S Fang F Zheng Q Ma Y Zhang J Wu N Yang Y MicroRNA let-7c inhibits migration and invasion of human non-small cell lung cancer by targeting ITGB3 and MAP 4K3 Cancer Lett 2014 342 43 51 10.1016/j.canlet.2013.08.030 23981581 46. Huang Y Xiang B Liu Y Wang Y Kan H LncRNA CDKN2B-AS1 promotes tumor growth and metastasis of human hepatocellular carcinoma by targeting let-7c-5p/NAP1L1 axis Cancer Lett 2018 437 56 66 10.1016/j.canlet.2018.08.024 30165194 47. Mittal V Epithelial Mesenchymal transition in tumor metastasis Annu Rev Pathol 2018 13 395 412 10.1146/annurev-pathol-020117-043854 29414248 48. Ribatti D Epithelial-mesenchymal transition in morphogenesis, cancer progression and angiogenesis Exp Cell Res 2017 353 1 5 10.1016/j.yexcr.2017.02.041 28257786
PMC007xxxxxx/PMC7943858.txt
==== Front Front Oncol Front Oncol Front. Oncol. Frontiers in Oncology 2234-943X Frontiers Media S.A. 33718183 10.3389/fonc.2021.621992 Oncology Original Research Advanced NSCLC Patients With EGFR T790M Harboring TP53 R273C or KRAS G12V Cannot Benefit From Osimertinib Based on a Clinical Multicentre Study by Tissue and Liquid Biopsy Fu Yulong 1 2 † Wang Anqi 1 2 † Zhou Jieqi 1 2 Feng Wei 1 Shi Minhua 3 Xu Xiao 4 Zhao Hongqing 5 Cai Liming 6 Feng Jian 7 Lv Xuedong 8 Zhang Xiaodong 9 Xu Wenjing 10 Zhang Zhengrong 11 Ma Guoer 12 Wang Jian 13 Zhou Tong 14 Zhao Dahai 15 Fang Haohui 16 Liu Zeyi 1 2 17 * Huang Jian-an 1 2 17 * 1 Department of Pulmonary and Critical Care Medicine, The First Affiliated Hospital of Soochow University, Suzhou, China 2 Suzhou Key Laboratory for Respiratory Diseases, Suzhou, China 3 Department of Respiratory Medicine, The Second Affiliated Hospital of Soochow University, Suzhou, China 4 Department of Respiratory Medicine, Affiliated Suzhou Hospital of Nanjing Medical University, Suzhou, China 5 Department of Respirology, Nanjing Medical University Affiliated Wuxi Second Hospital, Wuxi, China 6 Department of Respiratory Medicine, Affiliated Hospital of Jiangnan University, Wuxi, China 7 Department of Respiratory Medicine, Affiliated Hospital of Nantong University, Nantong, China 8 Department of Respiratory Medicine, The Second Affiliated Hospital of Nantong University, Nantong, China 9 Department of Medical Oncology, Nantong Tumor Hospital, Nantong, China 10 Departments of Respiratory Medicine, Northern Jiangsu People’s Hospital, Clinical Medical College of Yangzhou University, Yangzhou, China 11 Department of Respiratory Medicine, First People’s Hospital of Yangzhou City, Yangzhou, China 12 Department of Respiratory Medicine, Affiliated Hospital of Jiangsu University, Zhenjiang, China 13 Department of Respiratory Medicine, Zhenjiang First People’s Hospital, Zhenjiang, China 14 Department of Oncology, Changzhou Cancer Hospital Affiliated to Soochow University, Changzhou, China 15 Department of Respiratory and Critical Care Medicine, The Second Hospital of Anhui Medical University, Hefei, China 16 Department of Respiratory Medicine, Anhui Chest Hospital, Hefei, China 17 Institute of Respiratory Diseases, Soochow University, Suzhou, China Edited by: Shiv K. Gupta, Mayo Clinic, United States Reviewed by: Sonia Jain, Mayo Clinic, United States; Xue Wu, Geneseeq Technology Inc., Canada *Correspondence: Zeyi Liu, [email protected]; Jian-an Huang, [email protected] This article was submitted to Cancer Molecular Targets and Therapeutics, a section of the journal Frontiers in Oncology †These authors have contributed equally to this work 24 2 2021 2021 11 62199227 10 2020 11 1 2021 Copyright © 2021 Fu, Wang, Zhou, Feng, Shi, Xu, Zhao, Cai, Feng, Lv, Zhang, Xu, Zhang, Ma, Wang, Zhou, Zhao, Fang, Liu and Huang 2021 Fu, Wang, Zhou, Feng, Shi, Xu, Zhao, Cai, Feng, Lv, Zhang, Xu, Zhang, Ma, Wang, Zhou, Zhao, Fang, Liu and Huang https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Background Non-small cell lung cancer (NSCLC) patients treated with first-generation epidermal growth factor receptor-tyrosine kinase inhibitor (EGFR-TKI) almost always acquire resistance, and the development of novel techniques analyzing circulating tumor DNA (ctDNA) have made it possible for liquid biopsy to detect genetic alterations from limited amount of DNA with less invasiveness. While a large amount of patients with EGFR exon 21 p.Thr790 Met (T790M) benefited from osimertinib treatment, acquired resistance to osimertinb has subsequently become a growing challenge. Methods We performed tissue and liquid rebiopsy on 50 patients with EGFR-mutant NSCLC who acquired resistance to first-generation EGFR-TKIs. Plasma samples underwent droplet digital PCR (ddPCR) and next-generation sequencing (NGS) examinations. Corresponding tissue samples underwent NGS and Cobas® EGFR Mutation Test v2 (Cobas) examinations. Results Of the 50 patients evaluated, the mutation detection rates of liquid biopsy group and tissue biopsy group demonstrated no significant differences (41/48, 85.4% vs. 44/48, 91.7%; OR=0.53, 95% CI=0.15 to 1.95). Overall concordance, defined as the proportion of patients for whom at least one identical genomic alteration was identified in both tissue and plasma, was 78.3% (36/46, 95% CI=0.39 to 2.69). Moreover, our results showed that almost half of the patients (46%, 23/50) resistant to first-generation EGFR-TKI harbored p.Thr790 Met (T790M) mutation. 82.6% (19/23) of the T790M positive patients were analyzed by liquid biopsy and 60.9% (14/23) by tumor tissue sequencing. Meanwhile, a wide range of uncommon mutations was detected, and novel mechanisms of osimertinib resistance were discovered. In addition, 16.7% (2/12) of the T790M positive patients with either TP53 R237C or KRAS G12V failed to benefit from the subsequent osimertinib treatment. Conclusion Our results emphasized that liquid biopsy is applicable to analyze the drug resistance mechanisms of NSCLC patients treated with EGFR-TKIs. Moreover, we discovered two uncommon mutations, TP53 R273C and KRAS G12V, which attenuates the effectiveness of osimertinib. non-small cell lung cancer liquid biopsy gene sequencing epidermal growth factor receptor-tyrosine kinase inhibitor (EGFR-TKI) drug resistance ==== Body pmcIntroduction With advances in genomic research and targeted therapy, the therapeutic strategies and prognosis of patients with NSCLC, especially EGFR-mutated lung adenocarcinoma, have been greatly improved (1). Even though EGFR-targeted therapy has become the standard treatment in advanced EGFR-positive NSCLC, most patients eventually develop acquired resistance to first-generation EGFR-TKIs after 12–24 months of treatment (2–4). Among many reasons leading to acquired resistance, the T790M mutation in an exon of the EGFR gene accounts for more than half of the cases, followed by other mechanisms, including human epidermal growth factor receptor-2 (HER2) amplification, phenotypic transformation, and mesenchymal epithelial transition proto-oncogene (MET) amplification (5–10). Hence, genome sequencing is required to determine the presence of the T790M mutation before the application of the third-generation EGFR-TKI osimertinib, which targets T790M (11, 12). Moreover, there are still patients with the T790M mutation who do not respond or partially respond to osimertinib, suggesting that additional resistance mechanisms may decrease the efficacy of osimertinib (13–15). A repeatable assay to comprehensively detect and monitor somatic alterations with high sensitivity during EGFR-TKI treatment is therefore required. The emergence of novel techniques has allowed the discovery of multiple genetic mutations in malignant tumors (16). Tissue-based analyses have been widely used to detect driver or resistance mutations in NSCLC, providing clinicians with valuable therapeutic insights. However, it also has evident limitations considering the invasiveness of tumor biopsy, and the insufficient size of specimens obtained might neglect the spatial and temporal tumor heterogeneity (17–19). Therefore, liquid biopsy caught our attention. Liquid biopsy is a diagnostic method that can repeatedly collect ctDNAs from multiple tumor deposits with minimal invasiveness (20, 21). ctDNAs harbor tumor-specific sequence alterations from tumor cells derived from primary tumors and metastatic lesions, allowing a more comprehensive analysis of the tumor genome than tissue DNA from a single site (22–24). While tissue biopsy is still irreplaceable in diagnosing tumor malignancy, we intend to investigate whether liquid biopsy could be conducted to detect EGFR-TKI resistance mechanisms in NSCLC patients and hopefully save patients from undue suffering. Despite the increasing utilization of plasma analyses in guiding clinical decisions, few studies have evaluated the efficiency of different approaches. NGS has been recognized as an efficient and reliable approach to detect multiple genetic mutations simultaneously (25, 26). On the other hand, ddPCR features extremely high sensitivity with a relatively narrow detection range (27, 28). Such advantages and limitations can be influential for the determination of subsequent treatment. In the present study, we used ddPCR and NGS sequencing to analyze plasma samples, and NGS and Cobas sequencing to analyze tissue samples to identify somatic mutations in EGFR-positive NSCLC patients during disease progression. We found a similar detection rate and a high correspondence between liquid biopsy and conventional tissue biopsy, while ddPCR and NGS each have advantages. Meanwhile, our results showed a wide spectrum of gene mutations, and two uncommon EGFR mutations were identified related to osimertinib resistance. Herein, our study emphasized the utility of liquid biopsy and might uncover two novel molecular mechanism of resistance. Materials and Methods Study Population and Inclusion Criteria This was a multi-centre study performed at the Department of Respiratory and Critical Care Medicine in 15 hospitals located in Jiangsu Province and Anhui Province in East China. The inclusion criteria were as follows (1): Patients histologically diagnosed with stage IIIB or IV lung adenocarcinoma. (2) EGFR mutations that were previously detected and treated with first-generation EGFR-TKIs, including gefitinib, erlotinib, and icotinib. (3) Presently graded progressive disease (PD) after first-line EGFR-TKI treatment. (4) Availability of rebiopsied tumor tissue or liquid biopsy samples for analysis. A total of 50 patients with advanced NSCLC were recruited from 01 February 2018 to 01 June 2019. Twenty-seven (54%) were male, and 23 (46%) were female, with a median age of 65 years (range, 44–84 years). Two patients had plasma samples only, and two patients underwent tissue rebiopsy only, while all the other patients’ tissue and paired plasma samples were collected within three days, and detections were finished within seven days. This study was approved by the institutional review board of the First Affiliated Hospital of Soochow University (No. 2018013). All patients enrolled in this study signed an informed consent form. Specimen Collection and DNA Extraction The tumor tissues and/or paired peripheral blood samples (8 ml) from each patient were collected within the scope of routine diagnostic procedures and were available in all patients. Four to eight sections of fresh tissue immersed in formalin (5-mm thickness) or formalin-fixed, paraffin-embedded (FFPE) diagnostic tumor tissue samples were used for tissue DNA extraction. The DNA from tumor tissues was isolated from the sample using the QIAamp DNA Mini Kit (Qiagen, Hilden, Germany), and the DNA from FFPE slides was isolated by using a QIAamp DNA FFPE Tissue Kit (Qiagen, Hilden, Germany). Both procedures were performed following the specific manufacturer’s instructions. Peripheral blood samples were collected in cell-free DNA BCT® tubes (Streck, La Vista, Nebraska, U.S.) and processed within 6 h. The samples were centrifuged at 3,000 rpm for 10 min at 4°C and then centrifuged for a second time at 16,000 g for 10 min at 4°C. Plasma was then separated and stored at −80°C until cfDNA extraction. Circulating cell-free DNA was extracted from at least 4 ml of plasma using a QIAamp Circulating Nucleic Acid kit (Qiagen, Hilden, Germany) according to the manufacturer’s protocol. All the procedures for the molecular analysis were performed following the specific manufacturer’s instructions. Mutations were then detected using the eluted DNA samples. Mutation Detection In a subset of cases (n = 50), ddPCR was carried out for hotspot mutations in EGFR using the QX100 Droplet Digital PCR System (Bio-Rad, Hercules, CA) at Shanghai Yuanqi Bio-pharmaceutical Company. Positivity was defined as the fractional abundance of ≥0.044% for the plasma samples. NGS was carried out using the Ion AmpliSeqTM Cancer Hotspot Panel v2 on the Ion Proton sequencer (ThermoFisher Scientific) at Shanghai Singlera Genomics. NGS data were analyzed using Torrent Suite software (version 5.0.2). Results that above the detection limitation of 0.1% were considered positive. The real-time PCR Cobas® EGFR Mutation Test Kit (Roche Diagnostics) uses a pool of primers allowing for a target size from 85 to 155 bp, which was divided into three different mixes for each sample and control. The Cobas assay was run using the Cobas z480 thermocycler (Roche Diagnostics) at Shanghai Di-an Diagnosis Technology Company and all the positive results were included. Statistical Analysis The chi-square test was used to calculate the concordance between mutation detection rates of plasma and tissue samples. The Pearson correlation coefficient was used to assess the correlation of variant frequencies between the assays. Multivariable logistic regression was performed to evaluate the association between clinical characteristics and ctDNA detection rate. The cut-off date used for therapy outcome analysis was 31 October 2019. Data were analyzed by SPSS Statistics version 25.0 (IBM Corporation, NY, USA) and SPSS 17.0 software (SPSS, Chicago, IL). All P values are two-sided, and confidence intervals are at the 95% level, with statistical significance defined as P < 0.05. Results Mutation Detection in Oncogenes After Resistance We used four different approaches to detect mutations among the 50 patients with advanced EGFR-mutant lung cancer who acquired resistance to EGFR-TKIs, including ddPCR and NGS assays performed in plasma samples and NGS and Cobas assays performed in tissue samples. A total of 50 patients from 15 hospitals, including 27 males and 23 females, were recruited for our investigation. The clinical characteristics of the patients are summarized in Table 1 . All patients were initially diagnosed with lung adenocarcinoma with an EGFR mutation and received first-generation EGFR-TKI treatment. According to the National Comprehensive Cancer Network (NCCN) guidelines for NSCLC, it is recommended to perform routine mutation tests for EGFR, BRAF, ERBB2, and MET as well as rearrangements in ALK, ROS1, and RET in all patients diagnosed with advanced NSCLC. Here, we performed the NGS panel containing 12 genes covering all the mutations mentioned above, and the ddPCR and Cobas assays each had their respective deficiencies, which are described in Table 2 . Table 1 Clinical characteristics of enrolled NSCLC patients with EGFR mutations (n = 50). Clinical characteristics Number of patients (%) Gender Man 27 (54) Women 23 (46) Age, median (range) 65 (44–84) Histological type Adenocarcinoma 50 (100) Cancer stage IIIb 12 (24) IV 38 (76) EGFR-activating mutation Exon 19 deletion 19 (38) Exon 21 L858R mutation 18 (36) Exon 18 G719X mutation 1 (2) Unknown 12 (24) Previous EGFR-TKI therapy Gefitinib 28 (56) Erlotinib 6 (12) Icotinib 16 (32) Sample type upon PD Plasma only 2 (4) Tissue only 2 (4) Plasma and tissue 46 (92) Table 2 Detectable mutations by assay. Gene Exon Mutation Plasma Tissue ddPCR NGS-plasma NGS-tissue Cobas® EGFR 18 G719X √ √ √ √ 19 19-del √ √ √ √ 20 T790M √ √ √ √ 20 Insert √ √ √ 21 L858R √ √ √ √ Others √ √ ALK √ √ ALK Fusion 20 √ √ BRAF √ √ ERBB2 √ √ FGFR1 √ √ MET √ √ NRAS √ √ KRAS √ √ PIK3CA √ √ RET Fusion 10 √ √ 11 √ √ 12 √ √ ROS1 Fusion 32 √ √ 33 √ √ 34 √ √ 35 √ √ 36 √ √ TP53 √ √ In the present study, 48 screened patients provided matched tumor tissue and plasma samples. Among the genotyping results obtained from 96 samples in 50 patients, a total of 48 ddPCR results and 46 NGS results were obtained from 48 plasma samples (2 plasma samples received no NGS results due to quality failure), while 48 NGS results and Cobas results were obtained from tumor tissue samples. A total of 46 out of 50 (92%) patients underwent both liquid biopsy and tissue rebiopsy for multimethod genotype sequencing after tumor progression. The distribution of the gene mutations classified by patients and alterations is shown in Figure 1 . Figure 1 Summary of putative resistance mechanisms to first-generation EGFR-TKIs identified in 50 patients. In total, somatic mutations were found in 96% (48/50) of the patients. NGS and ddPCR detected somatic mutations in 82.6% (38/46) and 70.8% (34/48) of the plasma samples, while NGS and Cobas detected gene alterations in 93.6% (44/47) and 69.4% (25/36) of the tissue samples, respectively. Thus, the mutation detection rate was slightly lower in samples of plasma than in tissue with no significant difference (41/48, 85.4% vs. 44/48, 91.7%; OR = 0.53, 95% CI = 0.15 to 1.95, P = 0.336), and NGS was the most sensitive among all the approaches. The overall concordance between plasma and tissue sequencing results, defined as the proportion of patients who carried at least one matching result between each of the two methods in tissue and plasma examining, was 78.3% (36/46, 95% CI = 0.39 to 2.69). Multivariable logistic regression on ctDNA detection rate and clinical factors such as age, sex, and disease stage did not identify any statistically significant variables that predicted the detection of plasma ctDNA. Resistant Mechanisms to First-Generation EGFR-TKI Treatment As demonstrated in Figure 1 , TP53 was the most frequently mutated gene (33/50, 66%), followed by EGFR T790M (23/50, 46%). Sensitive EGFR mutations, including ex19del, L858R, G719X, and L816Q alterations, were found in 82% (41/50) of the patients. In addition, resistance mechanisms to EGFR-TKIs are known to be heterogeneous, and multiple aberrations may be present simultaneously, which is consistent with our findings (18, 29, 30). Ninety percent (45/50) of the patients bore more than one genomic alteration. During first-line treatment, 28 of the patients (56%) received gefitinib, 6 of them (12%) received erlotinib, and the remaining 16 (32%) were treated with icotinib. Hence the nine most frequently detected resistant mutations that might have clinical significance were sorted by different subsets of first-generation TKIs in Figure 2 , and the pattern of resistance mechanisms to gefitinib, erlotinib, and icotinib appeared to differ. Figure 2 Cases of potential resistance mechanisms detected in patients sorted by different EGFR-TKIs. The prevalence of the TP53 mutation was still the highest and was observed in 46% of the patients (23 of 50), especially among patients treated with gefitinib and erlotinib. Interestingly, the T790M mutation, known to be the most common mechanism of acquired resistance to gefitinib, was first identified in the current study. It was also detected in 46% of all patients (23/50) and 43% of patients treated with gefitinib (12/28). EGFR amplification (16%, 8/50) and MET amplification (12%, 6/50) were also observed. Apart from these differences, there were no significant differences in the prevalence of other resistance mechanisms between the three groups. Despite the limited number of samples analyzed, these results might reflect the heterogeneity of TKI resistance. Detection Rate of the T790M Mutation We next investigated the detection rate of T790M, the gene alteration most related to second-line therapy after EGFR-TKI resistance, among four different methods. The results are shown in the form of a Venn diagram in Figure 3A . Among all the T790M-positive patients, 19 of 23 were detected in plasma samples, while 14 of 23 were detected in tissue samples. The concordance between plasma and tissue NGS genotyping for the T790M mutation was 50% (9/18), and the detection rate of T790M was the same in NGS-plasma samples and in NGS-tissue samples (60.9%). Furthermore, ddPCR and NGS detected EGFR T790M mutations in 78.3% and 60.9% of plasma samples respectively, and the detection concordance between NGS-plasma and ddPCR is 68.4% (13/19). We further determined that T790M abundance detected by ddPCR was significantly lower than NGS in Figure 3B . Combined with the information in Table 2 , we concluded that NGS is a more comprehensive approach with a relatively satisfactory detection rate. On the other hand, ddPCR is a more sensitive test, especially regarding the T790M mutation, despite its comparatively small examination range. Figure 3 (A) Comparison of T790M detection among four different approaches. (B) Comparison of the T790M abundance detected by ddPCR and NGS in plasma samples. (*P < 0.05). Potential Resistance Mechanisms to Osimertinib Treatment According to the follow-up investigation, 9 of the 23 T790M-positive patients withdrawn from EGFR-TKIs treatment and chose chemotherapy of palliative treatment due to the relatively high expense of osimertinib. The other two patients were ruled out because their T790M mutation was detected by only one approach that could lead to low credibility. Thus, we analyzed 12 cases that received third-generation EGFR-TKI osimertinib treatment, as shown in Table 3 . Among the 12 patients, 10 showed stable disease or reduced tumor size by computerized tomography (CT) scanning, while 2 of them exhibited disease progression. Moreover, most of the patients (9/12, 75.0%) showed more than one genetic mutation, including T790M combined with a TP53 point mutation (2/12, 16.7%), a PIK3CA point mutation (3/12, 25.0%), or even a triple mutation like EGFR L858R/MET/PIK3CA. Based on the assumption that coexisting mutations may be related to the mechanism underlying the development of EGFR-TKI resistance, we further investigated two patients, No. 17 and No. 45, who progressed on osimertinib and found several seldom studied point mutations. Table 3 Response of patients with T790M mutation to osimertinib. ID Resistance mutation PFS to first-generation EGFR-TKIs (month) Response to therapy 03 EGFR T790ML, PIK3CA R88QL 30 Effective 04 EGFR T790ML,T, PIK3CA E545KL,T 12 Effective 08 EGFR T790ML,T, TP53 C124Ter T 14 Effective 11 EGFR T790ML, EGFR AmpL, ERBB2 Amp T 13 Effective 15 EGFR T790ML 4 Effective 17 EGFR T790ML,T, EGFR Amp T, TP53 R273CL 35 Progression 18 EGFR T790ML,T, PIK3CA Amp T 35 Effective 25 EGFR T790ML,T 52 Effective 30 EGFR T790ML,T, EGFR Amp T, TP53 D281G T 12 Effective 33 EGFR T790ML,T, TP53 c.559+1G>A T, TP53 R249SL 19 Effective 41 EGFR T790M T 7 Effective 45 EGFR T790M T, KRAS G12VL, TP53 G244D T 5 Progression L, detected by liquid biopsy; T, detected by tissue biopsy, PFS, progression-free survival. Patient No. 17, a 75-year-old female, never smoker, was diagnosed with right lung adenocarcinoma positive for EGFR ex19del. The patient was initially treated with first-generation EGFR-TKI gefitinib since EGFR exon 19 deletion was detected upon diagnosis biopsy. However, disease progressed 32 months after occurrence, as shown in Figure 4. A rebiopsy revealed EGFR T790M, TP53 R273C mutation, and EGFR amplification, with persistence of EGFR ex19del. Hence, second-line treatment with osimertinib was carried out. Unfortunately, disease progression occurred 3 months later, treatment with osimertinib was discontinued. Considering that other patients harboring EGFR T790M and EGFR amplification have shown sensitivity to osimertinib treatment (31), we identified the TP53 R273C mutation to promote resistance to osimertinib. Figure 4 Illustration of the genotype data and treatment received by patient No. 17 along with representative CT scans at the time points indicate. Patient No. 45, a 72-year-old male, presented with uncontrollable cough, after which he was diagnosed with an EGFR-mutated ex19del right lung adenocarcinoma. First, chemotherapy with cisplatin and pemetrexed was used for a total of 7 months until mediastinal lymph node enlargement was observed. Gefitinib was then administered, being optimally tolerated at the beginning. However, after 4 months, gefitinib was discontinued because of disease progression, as shown in Figure 5. The subsequent rebiopsy results confirmed ex19del and revealed the emergence of EGFR T790M, KRAS G12V, and TP53 G244D mutations. The patient then commenced a new treatment with osimertinib. However, his condition rapidly declined, and platinum-based combination chemotherapy was initiated. Since there is currently no report about TP53 G244D mutation-induced TKI resistance, it is suggested that the KRAS G12V mutation may prompt resistance to third-generation EGFR-TKIs. Figure 5 Illustration of the genotype data and treatment received by patient No. 45 along with representative CT scans at the time points indicated. Discussion Acquired resistance to EGFR-TKIs has become an important issue for NSCLC patients with EGFR activating mutations and is mostly due to secondary mutations such as T790M (5, 6). Given that former studies have shown intra-patient heterogeneity of resistant mechanisms (8, 9, 32), rebiopsies were requested at disease progression in the detection of putative resistant mutations. However, traditional bronchoscopy tissue biopsy has limitations, such as invasiveness. More importantly, tissue biopsy can only analyze one single tumor deposit at a time and requires a sufficient amount of tissue to perform all the assays (19), which would likely underestimate the temporal and spatial heterogeneity in resistance mutations (17, 18). Hence, a novel method for detecting EGFR-TKI-resistant mutations with high sensitivity and accuracy is urgently required for clinical practice. Here, we aimed to determine the most suitable rebiopsy assay for patients with advanced NSCLC resistant to first-line EGFR-TKIs. Liquid biopsy caught our attention in that it can simultaneously capture and detect mutations in multiple tumor deposits with minimum invasiveness. It can also detect different types of gene alterations, including single nucleotide variants, copy number variants, deletions, and fusion genes. Thus, analysis of ctDNA can be conducted repeatedly throughout the whole course of treatment to monitor genomic alteration (22, 33–35). In the present study, we compared the efficiency of four different methods of genomic sequencing for the first time. We used NGS and Cobas sequencing for tissue samples and detected somatic mutations in 91.7% (44/48) of samples. NGS and ddPCR were used for plasma samples, and the respective detection rate was 85.4% (41/48). The concordance between tissue and plasma genotyping was 78.3% (36/46), indicating that there is no significant difference between their detection rates. Considering the advantages mentioned before, our findings suggest that liquid biopsy is more applicable to EGFR-TKI resistance detection than tissue biopsy. Subsequently, we sought to identify the difference between two assays of liquid biopsy, ddPCR and NGS. In performing ctDNA analysis, we observed a high frequency of molecular heterogeneity in resistance mechanisms following treatment of NSCLC patients with first-generation EGFR-TKIs. NGS has the advantage of comprehensively detecting multiple gene alterations (36). It is capable of testing thousands of genes or even the whole genome with a relatively low DNA input (26, 37). On the other hand, ddPCR showed the highest sensitivity in detecting certain gene mutations that guide second-line treatment, such as T790M (28). Its detection rate of T790M is higher than that of NGS (78.3% vs. 60.9%), and ddPCR can identify T790M at a significantly lower abundance than NGS. In summary, NGS can provide a comprehensive view of the genome, while ddPCR is applicable to detecting specific mutations guiding clinical practice. After we identified 12 patients with the T790M mutation by different assays, osimertinib treatment was initiated. Previous studies have confirmed the effectiveness of the third-generation TKI osimertinib for patients with acquired T790M mutations upon disease progression (38–40). However, the duration of response varies even for patients who initially respond to osimertinib, since heterogeneity has long been an obstacle in targeted treatment (30). Recent researched have shown spatial temporal heterogeneity among lung cancer, and the emergence of concurrent mutations such as MET amplification or other driver oncogene before or during pharmacologic intervention can also induce osimertinib resistance (18, 29, 41). Our study here revealed two patients who demonstrated no response to osimertinib treatment. Given that multiple gene alterations were detected in our study and that prior studies have observed intrapatient heterogeneity of resistance mechanisms in NSCLC patients with EGFR mutations, we hypothesized that the mutations coexisting with T790M may impact the efficacy of treatment with third-generation EGFR-TKIs. Moreover, concurrent TP53 alterations have been considered to be related to a lower likelihood of responding to EGFR-TKIs and a shorter PFS (11, 42, 43), and TP53 R273 has been confirmed able to reduce cell apoptosis which might be the mechanism underlying R273C induced osimertinib resistance (44, 45). Studies have suggested that KRAS mutations are associated with decreased responsiveness to EGFR-TKIS in NSCLC (16, 46), probably because KRAS mutation can lead to the activation of downstream RAF-MEK-ERK pathway (47). KRAS G12V that we concerned was found more likely to induce EGFR-TKI resistance by regulating MEK/ERK than PI3K/AKT pathway (48). Thus, we identified two candidate mutations in patient No. 17 and No. 45, TP53 R273C and KRAS G12V, that may have antagonistic effects on the efficacy of osimertinib. Collectively, these findings underlined that genomic results of liquid biopsy can be indicative of the clinical response to second-line EGFR-TKI treatment, such as osimertinib. There are still limitations to our retrospective study. First, it included a relatively small number of patients. Second, analysis of plasma DNA at baseline was not performed. Finally, the efficacy of subsequent treatment with osimertinib according to genetic profile was not monitored in our study. Our study showed that liquid biopsy is sufficient to identify somatic mutations and resistance mechanisms at disease progression during treatment with first-generation EGFR-TKIs compared to tissue biopsy. Different approaches of ctDNA analysis have their respective advantages and limitations. Hence, a combination of multiple approaches can offer patients the best therapeutic options. Data Availability Statement The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding authors. Ethics Statement The studies involving human participants were reviewed and approved by the First Affiliated Hospital of Soochow University. The patients/participants provided their written informed consent to participate in this study. Author Contributions J-AH and ZL designed the experiments. YF, AW, and JZ analyzed the data and wrote the manuscript. WF, MS, XX, HZ, LC, JF, XL, XZ, WX, ZZ, GM, JW, TZ, DZ, and HF provided tissue and/or plasma samples and clinical data. All authors contributed to the article and approved the submitted version. Funding This work was supported by the Suzhou Key Laboratory for Respiratory Medicine (grant number SZS201617), the Clinical Medical Center of Suzhou (grant number Szzx201502), the Jiangsu Provincial Key Medical Discipline (grant number ZDXKB2016007), and the Clinical Key Specialty Project of China. Conflict of Interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Supplementary Material The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fonc.2021.621992/full#supplementary-material Click here for additional data file. Click here for additional data file. ==== Refs References 1 Mok TS Wu YL Thongprasert S Yang CH Chu DT Saijo N . Gefitinib or carboplatin-paclitaxel in pulmonary adenocarcinoma. N Engl J Med (2009) 361 (10 ):947–57. doi: 10.1056/NEJMoa0810699 2 Rosell R Carcereny E Gervais R Vergnenegre A Massuti B Felip E . Erlotinib versus standard chemotherapy as first-line treatment for European patients with advanced EGFR mutation-positive non-small-cell lung cancer (EURTAC): a multicentre, open-label, randomised phase 3 trial. Lancet Oncol (2012) 13 (3 ):239–46. doi: 10.1016/S1470-2045(11)70393-X 3 Sequist LV Yang JC-H Yamamoto N O’Byrne K Hirsh V Mok T . Phase III Study of Afatinib or Cisplatin Plus Pemetrexed in Patients With Metastatic Lung Adenocarcinoma With EGFR Mutations. J Clin Oncol (2013) 31 (27 ):3327–34. doi: 10.1200/JCO.2012.44.2806 4 Maemondo M Inoue A Kobayashi K Sugawara S Oizumi S Isobe H . Gefitinib or chemotherapy for non-small-cell lung cancer with mutated EGFR. N Engl J Med (2010) 362 (25 ):2380–8. doi: 10.1056/NEJMoa0909530 5 Yu HA Arcila ME Rekhtman N Sima CS Zakowski MF Pao W . Analysis of Tumor Specimens at the Time of Acquired Resistance to EGFR-TKI Therapy in 155 Patients with <em<EGFR</em<-Mutant Lung Cancers. Clin Cancer Res (2013) 19 (8 ):2240. doi: 10.1158/1078-0432.CCR-12-2246 23470965 6 Pao W Miller VA Politi KA Riely GJ Somwar R Zakowski MF . Acquired resistance of lung adenocarcinomas to gefitinib or erlotinib is associated with a second mutation in the EGFR kinase domain. PloS Med (2005) 2 (3 ):e73–3. doi: 10.1371/journal.pmed.0020073 7 Kobayashi S Boggon TJ Dayaram T Jänne PA Kocher O Meyerson M . EGFR mutation and resistance of non-small-cell lung cancer to gefitinib. N Engl J Med (2005) 352 (8 ):786–92. doi: 10.1056/NEJMoa044238 8 Sequist LV Waltman BA Dias-Santagata D Digumarthy S Turke AB Fidias P . Genotypic and histological evolution of lung cancers acquiring resistance to EGFR inhibitors. Sci Transl Med (2011) 3 (75 ):75ra26. doi: 10.1126/scitranslmed.3002003 9 Suda K Murakami I Katayama T Tomizawa K Osada H Sekido Y . Reciprocal and Complementary Role of MET Amplification and EGFR T790M Mutation in Acquired Resistance to Kinase Inhibitors in Lung Cancer. Clin Cancer Res (2010) 16 (22 ):5489. doi: 10.1158/1078-0432.CCR-10-1371 21062933 10 Facchinetti F Proto C Minari R Garassino M Tiseo M . Mechanisms of Resistance to Target Therapies in Non-small Cell Lung Cancer. In: Mandalà M Romano E , editors. Mechanisms of Drug Resistance in Cancer Therapy. Cham: Springer International Publishing (2018). p. 63–89. 11 Cross DAE Ashton SE Ghiorghiu S Eberlein C Nebhan CA Spitzler PJ . AZD9291, an Irreversible EGFR TKI, Overcomes T790M-Mediated Resistance to EGFR Inhibitors in Lung Cancer. Cancer Discov (2014) 4 (9 ):1046. doi: 10.1158/2159-8290.CD-14-0337 24893891 12 Jänne PA Yang JC Kim DW Planchard D Ohe Y Ramalingam SS . AZD9291 in EGFR inhibitor-resistant non-small-cell lung cancer. N Engl J Med (2015) 372 (18 ):1689–99. doi: 10.1056/NEJMoa1411817 13 Thress KS Paweletz CP Felip E Cho BC Stetson D Dougherty B . Acquired EGFR C797S mutation mediates resistance to AZD9291 in non-small cell lung cancer harboring EGFR T790M. Nat Med (2015) 21 (6 ):560–2. doi: 10.1038/nm.3854 14 Piotrowska Z Niederst MJ Karlovich CA Wakelee HA Neal JW Mino-Kenudson M . Heterogeneity Underlies the Emergence of EGFRT790 Wild-Type Clones Following Treatment of T790M-Positive Cancers with a Third-Generation EGFR Inhibitor. Cancer Discov (2015) 5 (7 ):713–22. doi: 10.1158/2159-8290.CD-15-0399 15 Planchard D Loriot Y André F Gobert A Auger N Lacroix L . EGFR-independent mechanisms of acquired resistance to AZD9291 in EGFR T790M-positive NSCLC patients. Ann Oncol (2015) 26 (10 ):2073–8. doi: 10.1093/annonc/mdv319 16 Chabon JJ Simmons AD Lovejoy AF Esfahani MS Newman AM Haringsma HJ . Circulating tumour DNA profiling reveals heterogeneity of EGFR inhibitor resistance mechanisms in lung cancer patients. Nat Commun (2016) 7 (1 ):11815. doi: 10.1038/ncomms11815 27283993 17 Zhang J Fujimoto J Zhang J Wedge DC Song X Zhang J . Intratumor heterogeneity in localized lung adenocarcinomas delineated by multiregion sequencing. Science (New York NY) (2014) 346 (6206 ):256–9. doi: 10.1126/science.1256930 18 de Bruin EC McGranahan N Mitter R Salm M Wedge DC Yates L . Spatial and temporal diversity in genomic instability processes defines lung cancer evolution. Science (New York NY) (2014) 346 (6206 ):251–6. doi: 10.1126/science.1253462 19 Overman MJ Modak J Kopetz S Murthy R Yao JC Hicks ME . Use of research biopsies in clinical trials: are risks and benefits adequately discussed? J Clin Oncol (2013) 31 (1 ):17–22. doi: 10.1200/JCO.2012.43.1718 23129736 20 Jovelet C Ileana E Le Deley M-C Motté N Rosellini S Romero A . Circulating Cell-Free Tumor DNA Analysis of 50 Genes by Next-Generation Sequencing in the Prospective MOSCATO Trial. Clin Cancer Res (2016) 22 (12 ):2960. doi: 10.1158/1078-0432.CCR-15-2470 26758560 21 Crowley E Di Nicolantonio F Loupakis F Bardelli A . Liquid biopsy: monitoring cancer-genetics in the blood. Nat Rev Clin Oncol (2013) 10 (8 ):472–84. doi: 10.1038/nrclinonc.2013.110 22 Qiu M Wang J Xu Y Ding X Li M Jiang F . Circulating tumor DNA is effective for the detection of EGFR mutation in non-small cell lung cancer: a meta-analysis. Cancer Epidemiol Biomarkers Prev (2015) 24 (1 ):206–12. doi: 10.1158/1055-9965.epi-14-0895 23 Schwarzenbach H Hoon DSB Pantel K . Cell-free nucleic acids as biomarkers in cancer patients. Nat Rev Cancer (2011) 11 (6 ):426–37. doi: 10.1038/nrc3066 24 Aung KL Board RE Ellison G Donald E Ward T Clack G . Current status and future potential of somatic mutation testing from circulating free DNA in patients with solid tumours. Hugo J (2010) 4 (1-4 ):11–21. doi: 10.1007/s11568-011-9149-2 22132062 25 Tuononen K Mäki-Nevala S Sarhadi VK Wirtanen A Rönty M Salmenkivi K . Comparison of Targeted Next-Generation Sequencing (NGS) and Real-Time PCR in the Detection of EGFR, KRAS, and BRAF Mutations on Formalin-Fixed, Paraffin-Embedded Tumor Material of Non-Small Cell Lung Carcinoma—Superiority of NGS. Genes Chromosomes Cancer (2013) 52 (5 ):503–11. doi: 10.1002/gcc.22047 26 Meldrum C Doyle MA Tothill RW . Next-generation sequencing for cancer diagnostics: a practical perspective. Clin Biochem Rev (2011) 32 (4 ):177–95. 27 Zhu G Ye X Dong Z Lu YC Sun Y Liu Y . Highly Sensitive Droplet Digital PCR Method for Detection of EGFR-Activating Mutations in Plasma Cell-Free DNA from Patients with Advanced Non-Small Cell Lung Cancer. J Mol Diagn (2015) 17 (3 ):265–72. doi: 10.1016/j.jmoldx.2015.01.004 28 Xu Q Zhu Y Bai Y Wei X Zheng X Mao M . Detection of epidermal growth factor receptor mutation in lung cancer by droplet digital polymerase chain reaction. Onco Targets Ther (2015) 8 :1533–41. doi: 10.2147/OTT.S84938 29 Suda K Murakami I Obata K Sakai K Fujino T Koga T . Spatial heterogeneity of acquired resistance mechanisms to 1st/2nd generation EGFR tyrosine kinase inhibitors in lung cancer. Lung Cancer (2020) 148 :100–4. doi: 10.1016/j.lungcan.2020.08.010 30 Suda K Murakami I Sakai K Tomizawa K Mizuuchi H Sato K . Heterogeneity in resistance mechanisms causes shorter duration of epidermal growth factor receptor kinase inhibitor treatment in lung cancer. Lung Cancer (2016) 91 :36–40. doi: 10.1016/j.lungcan.2015.11.016 26711932 31 Shan L Wang Z Guo L Sun H Qiu T Ling Y . Concurrence of EGFR amplification and sensitizing mutations indicate a better survival benefit from EGFR-TKI therapy in lung adenocarcinoma patients. Lung Cancer (2015) 89 (3 ):337–42. doi: 10.1016/j.lungcan.2015.06.008 32 Ohashi K Sequist LV Arcila ME Moran T Chmielecki J Lin YL . Lung cancers with acquired resistance to EGFR inhibitors occasionally harbor BRAF gene mutations but lack mutations in KRAS, NRAS, or MEK1. Proc Natl Acad Sci U S A (2012) 109 (31 ):E2127–2133. doi: 10.1073/pnas.1203530109 33 Thress KS Brant R Carr TH Dearden S Jenkins S Brown H . EGFR mutation detection in ctDNA from NSCLC patient plasma: A cross-platform comparison of leading technologies to support the clinical development of AZD9291. Lung Cancer (2015) 90 (3 ):509–15. doi: 10.1016/j.lungcan.2015.10.004 34 Punnoose EA Atwal S Liu W Raja R Fine BM Hughes BG . Evaluation of circulating tumor cells and circulating tumor DNA in non-small cell lung cancer: association with clinical endpoints in a phase II clinical trial of pertuzumab and erlotinib. Clin Cancer Res (2012) 18 (8 ):2391–401. doi: 10.1158/1078-0432.ccr-11-3148 35 Sorber L Zwaenepoel K Deschoolmeester V Van Schil PE Van Meerbeeck J Lardon F . Circulating cell-free nucleic acids and platelets as a liquid biopsy in the provision of personalized therapy for lung cancer patients. Lung Cancer (2017) 107 :100–7. doi: 10.1016/j.lungcan.2016.04.026 36 Sunami K Takahashi H Tsuchihara K Takeda M Suzuki T Naito Y . Clinical practice guidance for next-generation sequencing in cancer diagnosis and treatment (Edition 1.0). Cancer Sci (2018) 109 (9 ):2980–5. doi: 10.1111/cas.13730 37 Luthra R Chen H Roy-Chowdhuri S Singh RR . Next-Generation Sequencing in Clinical Molecular Diagnostics of Cancer: Advantages and Challenges. Cancers (2015) 7 (4 ):2023–36. doi: 10.3390/cancers7040874 38 Goss G Tsai CM Shepherd FA Bazhenova L Lee JS Chang GC . Osimertinib for pretreated EGFR Thr790Met-positive advanced non-small-cell lung cancer (AURA2): a multicentre, open-label, single-arm, phase 2 study. Lancet Oncol (2016) 17 (12 ):1643–52. doi: 10.1016/s1470-2045(16)30508-3 39 Yang JC Ahn MJ Kim DW Ramalingam SS Sequist LV Su WC . Osimertinib in Pretreated T790M-Positive Advanced Non-Small-Cell Lung Cancer: AURA Study Phase II Extension Component. J Clin Oncol (2017) 35 (12 ):1288–96. doi: 10.1200/jco.2016.70.3223 40 Mok TS Wu YL Ahn MJ Garassino MC Kim HR Ramalingam SS . Osimertinib or Platinum-Pemetrexed in EGFR T790M-Positive Lung Cancer. N Engl J Med (2017) 376 (7 ):629–40. doi: 10.1056/NEJMoa1612674 41 Nishiyama A Takeuchi S Adachi Y Otani S Tanimoto A Sasaki M . MET amplification results in heterogeneous responses to osimertinib in EGFR-mutant lung cancer treated with erlotinib. Cancer Sci (2020) 111 (10 ):3813–23. doi: 10.1111/cas.14593 42 Labbé C Cabanero M Korpanty GJ Tomasini P Doherty MK Mascaux C . Prognostic and predictive effects of TP53 co-mutation in patients with EGFR-mutated non-small cell lung cancer (NSCLC). Lung Cancer (2017) 111 :23–9. doi: 10.1016/j.lungcan.2017.06.014 43 VanderLaan PA Rangachari D Mockus SM Spotlow V Reddi HV Malcolm J . Mutations in TP53, PIK3CA, PTEN and other genes in EGFR mutated lung cancers: Correlation with clinical outcomes. Lung Cancer (Amsterdam Netherlands) (2017) 106 :17–21. doi: 10.1016/j.lungcan.2017.01.011 28285689 44 Li J Yang L Gaur S Zhang K Wu X Yuan YC . Mutants TP53 p.R273H and p.R273C but not p.R273G enhance cancer cell malignancy. Hum Mutat (2014) 35 (5 ):575–84. doi: 10.1002/humu.22528 45 Rho JK Choi YJ Ryoo BY Na II Yang SH Kim CH . p53 enhances gefitinib-induced growth inhibition and apoptosis by regulation of Fas in non-small cell lung cancer. Cancer Res (2007) 67 (3 ):1163–9. doi: 10.1158/0008-5472.CAN-06-2037 46 Pao W Wang TY Riely GJ Miller VA Pan Q Ladanyi M . KRAS mutations and primary resistance of lung adenocarcinomas to gefitinib or erlotinib. PloS Med (2005) 2 (1 ):e17–7. doi: 10.1371/journal.pmed.0020017 47 Stinchcombe TE Johnson GL . MEK inhibition in non-small cell lung cancer. Lung Cancer (2014) 86 (2 ):121–5. doi: 10.1016/j.lungcan.2014.09.005 48 Ihle NT Byers LA Kim ES Saintigny P Lee JJ Blumenschein GR . Effect of KRAS oncogene substitutions on protein behavior: implications for signaling and clinical outcome. J Natl Cancer Inst (2012) 104 (3 ):228–39. doi: 10.1093/jnci/djr523
PMC008xxxxxx/PMC8206684.txt
==== Front Proc Biol Sci Proc Biol Sci RSPB royprsb Proceedings of the Royal Society B: Biological Sciences 0962-8452 1471-2954 The Royal Society 34130501 10.1098/rspb.2021.0228 rspb20210228 10011460Behaviour Research Articles Host personality predicts cuckoo egg rejection in Daurian redstarts Phoenicurus auroreus Host personality predicts cuckoo egg rejection in Daurian redstarts Phoenicurus auroreus http://orcid.org/0000-0002-0459-0429 Zhang Jinggang Conceptualization Data curation Formal analysis Funding acquisition Investigation Methodology Project administration Resources Software Validation Visualization Writing-original draft Writing-review & editing 1 2 Santema Peter Conceptualization Formal analysis Methodology Software Visualization Writing-review & editing 2 http://orcid.org/0000-0002-0029-1747 Li Jianqiang Conceptualization Formal analysis Methodology Software Visualization Writing-review & editing 3 Yang Lixing Data curation Investigation Resources Writing-review & editing 1 http://orcid.org/0000-0002-2355-567X Deng Wenhong Conceptualization Funding acquisition Investigation Methodology Project administration Resources Supervision Validation Writing-review & editing [email protected] 1 http://orcid.org/0000-0002-7505-5458 Kempenaers Bart Conceptualization Formal analysis Methodology Project administration Supervision Validation Writing-review & editing 2 1 Ministry of Education Key Laboratory for Biodiversity Sciences and Ecological Engineering, College of Life Sciences, Beijing Normal University, Beijing, People's Republic of China 2 Department of Behavioural Ecology and Evolutionary Genetics, Max Planck Institute for Ornithology, Seewiesen, Germany 3 School of Ecology and Nature Conservation, Beijing Forestry University, Beijing, People's Republic of China Electronic supplementary material is available online at https://doi.org/10.6084/m9.figshare.c.5448679. 30 6 2021 June 30, 2021 16 6 2021 June 16, 2021 16 6 2021 June 16, 2021 288 1953 2021022828 1 2021 January 28, 2021 19 5 2021 May 19, 2021 © 2021 The Authors. 2021 https://creativecommons.org/licenses/by/4.0/ Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited. In species that are subject to brood parasitism, individuals often vary in their responses to parasitic eggs, with some rejecting the eggs while others do not. While some factors, such as host age (breeding experience), the degree of egg matching and the level of perceived risk of brood parasitism have been shown to influence host decisions, much of the variation remains unexplained. The host personality hypothesis suggests that personality traits of the host influence its response to parasitic eggs, but few studies have tested this. We investigated the relationship between two personality traits (exploration and neophobia) and a physiological trait (breathing rate) of the host, and egg-rejection behaviour in a population of Daurian redstarts Phoenicurus auroreus in northeast China. We first show that exploratory behaviour and the response to a novel object are repeatable for individual females and strongly covary, indicating distinct personality types. We then show that fast-exploring and less neophobic hosts were more likely to reject parasitic eggs than slow-exploring and more neophobic hosts. Variation in breathing rate—a measure of the stress-response—did not affect rejection behaviour. Our results demonstrate that host personality, along the bold-shy continuum, predicts the responses to parasitic eggs in Daurian redstarts, with bold hosts being more likely to reject parasitic eggs. personality , breathing rate , cuckoo parasitism , egg rejection , Daurian redstart National Natural Science Foundation of China http://dx.doi.org/10.13039/501100001809 31372219 31672297 China Scholarship Council http://dx.doi.org/10.13039/501100004543 201906040159 cover-dateJune 30, 2021 ==== Body pmc1. Introduction Obligate avian brood parasites lay their eggs into the nests of other species and thus transfer the costs of parental care to their hosts [1]. Consequently, hosts have evolved a variety of defences to reduce the incidence of parasitism and/or to minimize the negative fitness consequences of successful parasitism [2–9]. One widespread anti-parasite defence is the rejection of parasitic eggs from the nest [5,6,10]. Despite its effectiveness, hosts show variation in levels of egg rejection both within and among populations [8,11]. Understanding why some hosts reject brood parasitic eggs while others do not, remains challenging. When deciding whether to accept or reject a potential parasitic egg, hosts have to balance the risk of mistakenly rejecting their own eggs against the cost of accepting a parasitic egg [12]. Empirical studies have shown several factors that are associated with between- and within- individual variation in egg-rejection behaviour in a host population. First, age and experience of the individual may play a role. For example, great reed warblers Acrocephalus arundinaceus that were older or had previous experience with being parasitized were more likely to reject a parasitic egg than young breeders or first-time hosts [13]. Second, rejection often depends on traits of the parasitic egg. For example, experiments using artificial cuckoo eggs in ashy-throated parrotbills Paradoxornis alphonsianus, a frequent host of the common cuckoo Cuculus canorus (hereafter, cuckoo), showed that hosts were less likely to reject the parasitic egg when it was more similar to its own eggs [14]. Other studies showed that the rejection behaviour of the host depended on the perceived risk of brood parasitism [15]. For example, hosts were more likely to reject a parasitic egg when they observed a cuckoo near their nest [16,17]. Yet, a considerable amount of variation in egg-rejection behaviour among individuals typically remains unexplained. Avilés & Parejo [18] proposed the host personality hypothesis, suggesting that the response to parasitic eggs depends on personality traits of the host. This may be adaptive if the risk of being parasitized differs between hosts with different personality. However, hitherto to our knowledge, no empirical evidence exists to support this hypothesis. Animal personalities are defined as consistent inter-individual variation in behavioural traits, such as activity, aggressiveness, boldness, neophobia and exploratory behaviour [19,20]. Different personality traits are often correlated with each other, such that suites of covarying traits form behavioural syndromes [21,22]. For instance, aggressive individuals also tend to be proactive, bold, risk-taking, less neophobic and fast-exploring [23–25]. A growing body of evidence suggests that personality traits can influence individuals in many aspects of their life history [20,26,27]. For example, exploratory behaviour has been related to individual survival [28], natal dispersal [29], extra-pair mating patterns [30], nest defence [31] and territory defence [32]. Empirical tests of a relationship between personality traits and egg-rejection behaviour in brood parasitized hosts remain scarce. The only direct test comes from a study on great reed warblers, in which the relationship between host aggressiveness and egg rejection was investigated, but no relationship was detected [33]. Several indirect lines of evidence suggest that personality traits could affect host egg-rejection behaviour. For example, egg rejection is based on a learning mechanism [34], whereby good learners would be better at discriminating parasitic eggs. Empirical evidence suggests that individual variation in discrimination learning is connected to individual variation in exploratory behaviour. For instance, in black-capped chickadees Poecile atricapillus, fast explorers learned acoustic cues more quickly [35], and similar results have been found in great tits Parus major and common starlings Sturnus vulgaris [36,37]. These results lead to the prediction that fast explorers would be better at rejecting parasitic eggs. On the other hand, slow-exploring black-capped chickadees learned to reverse previously learned natural category rules more quickly than fast explorers, suggesting that slow explorers may be more sensitive to environmental stimuli [38]. In this case, slow explorers are predicted to be better at egg-rejection behaviour. Moreover, since egg-rejection behaviour comes with potential costs (e.g. recognition error and revenge by the parasite [12,39]), shy hosts may be more likely to tolerate or accept the brood parasitism to avoid these costs, while bold individuals may be more likely to take the risk and reject the parasitic egg. Because they are generally more aggressive, bold hosts may also be more effective than shy individuals at driving brood parasites away from their nest [31]. Being more successful at earlier lines of host defence may decrease selection on later lines of defence [40,41]. Thus, bold individuals may be better at keeping brood parasites away from their nest, but less efficient at discriminating and ejecting the parasitic egg, as they may have less opportunity to learn to reject a parasitic egg [18], but see [33]. On the other hand, shy individuals may be less frequently parasitized, for example, because they are less active and thus less conspicuous than bold individuals [42,43], and they may therefore also experience less selection to be good egg rejecters [18]. We report on an experimental study to test the host personality hypothesis using the Daurian redstart Phoenicurus auroreus as a model species. Daurian redstarts are a common cuckoo host and individuals of this species vary in their responses to parasitic eggs [44]. Males never reject a parasitic egg, but about half of the females do, while the other half accept such an egg [44]. Thus, the Daurian redstart provides an ideal system in which to investigate variation in egg-rejection behaviour [44]. The species also shows an egg colour polymorphism, with some females laying blue and others pink eggs, whereby the latter are more distinct from the blue cuckoo eggs [44,45]. The egg colour polymorphism is often interpreted as an adaptation against brood parasitism [14,46]. In our study population, redstarts laying blue eggs appear to suffer higher risk of parasitism than hosts laying pink eggs, although unbiased information on parasitization rate of the latter hosts are lacking as they may have ejected the parasitic egg before we detect it [44]. We previously found that egg-rejection behaviour in Daurian redstarts varied with host clutch colour and with the risk of parasitism: females laying pink eggs were more likely to reject foreign eggs than individuals laying blue eggs, and hosts experiencing a higher risk of being parasitized (cuckoo presence, see below) had higher egg-rejection rates [44]. However, a lot of variation in egg rejection among females remains unexplained. Therefore, we explored whether host personality traits affect egg-rejection behaviour. We first tested whether Daurian redstarts showed consistent inter-individual variation in two personality traits (exploration and neophobia) and in a physiological trait (breathing rate) across time. Second, we show how the two personality traits and breathing rate are correlated. Lastly, we investigated how these traits influence the response of the host to a parasitic egg. 2. Methods (a) Study system and general procedures We studied a population of Daurian redstarts in the village of Shuangyu in Jilin, northeast China (43°37′19″ N and 126°09′54″ E) in 2019 and 2020. The study site is about 50 ha, and contains 170 nest-boxes. In our study site, females start laying from mid-April onwards and typically produce at least two clutches within one breeding season (clutch size (mean ± s.d.): 6.4 ± 0.6 (n = 99) and 5.5 ± 1.0 (n = 163) in the first and second clutch, respectively). Cuckoos arrive at the breeding grounds around mid-May (13 May in 2019), when most hosts have nestlings or are in the late incubation stage of their first clutch. Thus, in this population of Daurian redstarts, the risk of cuckoo parasitism varies within each breeding season from zero in the first clutch to a high risk in subsequent clutches (for detailed information, see [44]). Cuckoo eggs in Daurian redstart nests are pale blue with or without thin brown lines; they mimic the blue morph of host eggs, but are paler and bigger [44]. During each breeding season, we searched for natural nests every day and checked nest-boxes every week. When cuckoos were present at the study site, we checked active nests (natural or in a nest box) every 1–2 days to assess whether it contained a cuckoo egg. We followed a total of 577 redstart nests, 370 in 2019 and 207 in 2020. Of these, 67 were naturally parasitized by a cuckoo egg, 43 in 2019 and 24 in 2020. (b) Experimental procedure To assess egg-rejection behaviour of Daurian redstart females, we performed a brood parasitism experiment using a real cuckoo egg or a model egg that mimics a real cuckoo egg. We manufactured model cuckoo eggs using clay and painted them with acrylic colours. Mass and size of the model cuckoo eggs were similar to real cuckoo eggs [44]. We performed a total of 97 trials, 54 in 2019 and 43 in 2020. For each trial, we introduced a model cuckoo egg into the focal nest, during either the late-laying phase of the host, i.e. when the nest contained three eggs, or during early incubation, i.e. within 3 days after clutch completion. After artificially parasitizing a nest or after finding a naturally parasitized nest, we checked it daily for 6 days to decide the fate of the parasitic egg. We considered the experimental egg ‘accepted' when it was still present in an active nest 6 days after it was introduced and ‘rejected' when it disappeared while the nest was active and the host clutch was not reduced. Out of 67 nests that were naturally parasitized, we had personality data for 18 host females (eight in 2019 and 10 in 2020). Only these 18 nests were therefore included in the analysis. We considered the nest to be deserted when the parasitic egg was still present, but the host had abandoned the nest within 6 days [47,48]. Our previous work showed that nest desertion rates did not differ between experimentally parasitized and non-manipulated control nests [47]. We therefore assumed that nest desertion in this study was not a consequence of the (artificial) parasitism and excluded deserted nests from further analysis. (c) Personality assays (i) Exploratory behaviour We tested exploratory behaviour using the novel-cage approach described in Kluen et al. [49] and validated in other passerines [50,51]. The exploration cage was adapted from a wooden box (L 60 × W 40 × H 80 cm), fitted with six perches (25 cm) and one mesh side, and connected to a small metal compartment (L 20 ×W 20 × H 20 cm) (electronic supplementary material, figure S1). During the 2019 and 2020 breeding season, when nestlings were 4 days old (hatching date = day 0), we caught adults by one of four methods (i.e. mist-net, tuck net, spring net traps or bird glue; see the electronic supplementary material). After banding, each individual was kept in the small compartment for 10 min of acclimatization. Then, the bird was released into the exploration cage through the connecting door without handling, and its behaviour recorded with a video camera, placed 3 m in front of the exploration cage, for 2 min. We scored exploratory behaviour as the number of hops within a location plus the number of movements (flights or hops) between different locations, including two floor sections and six sections within the cage area (scores ranged from 2 to 151 [51,52]). (ii) Neophobia Neophobia is commonly measured as the reluctance of individuals to return to a known resource in the presence of a novel object [53–56]. In this study, we measured the female's latency to return to her nest (nest-box or natural nest) in the presence of the novel object during the incubation stage. During the 2020 breeding season, we performed novel object and control tests between 9 and 11 days after the start of incubation. We conducted experiments only in the afternoon (15.00–18.00), and only when the female was on the nest. First, we induced the female to leave by tapping the nest-box or the foundation of the natural nest. Then, we placed a yellow or red ping-pong ball on top of the nest-box or 10–15 cm above the natural nest and a video camera 5 m from the focal nest [55,56]. We then recorded the nest (box) for 60 min and measured return latency (in minutes). To confirm that the observed responses were caused by the novel object rather than by human disturbance, we conducted control trials, following the same procedure but placing no novel object near the nest. For 43 females that had undergone the artificial brood parasitism experiment, we conducted a first novel object test (yellow ball) and a control test. For 27 of these females, we performed a second novel object test using a red ball to determine the repeatability of the neophobia response. We performed the tests (control, yellow ball and red ball) on three consecutive days. To avoid order effects, we performed the first novel object (yellow ball) and control tests in a randomly determined order (days 9 and 10 of incubation). The second novel object test (red ball) was performed on the third day (day 11). (iii) Breathing rate Breathing rate is a physiological trait that has been proposed as an indicator of the stress response in songbirds [57]. Immediately after capture, we measured breathing rate by counting the number of breast movements within 30 s, following [58]. (d) Statistical analyses As defined, a meaningful personality trait should be individually repeatable [22]. We therefore tested the adjusted repeatability of all three measured traits using linear mixed-effects models (LMMs) with the trait measure as the dependent variable and bird identification as the random effect, using the R package rptR [59,60]. Following recommendations of Nakagawa & Schielzeth [59], we retained individuals with only one measure in the models. For breathing rate and exploratory behaviour (number of movements), we included sex, capture method, the date on which the test was performed (day of the year), test sequence, and the interval between two tests (in days) as fixed effects. For the neophobia response, we included return latency (during novel objects trials) as the dependent variable, with object type (yellow ball/red ball), nest type and baseline return latency as fixed effects. To avoid possible model overfitting, we further ran LMMs to detect what variables explained the significant variation in the personality trait. We then recalculated the repeatability of the three traits again, only including variables that explained significant variation in the personality trait. This approach gave qualitatively the same results (see the electronic supplementary material). We log-transformed data of return latency to meet the normality assumption. To determine whether the return latency during the novel object treatment was a neophobia response rather than a response to human disturbance, we used a Wilcoxon matched-pairs test to compare the return latency between the control and the first novel object trial (yellow ping-pong ball). We calculated a Pearson correlation coefficient between breathing rate and exploratory behaviour (number of movements), and a Spearman's rank correlation coefficient between return latency and either exploratory behaviour or breathing rate. We used generalized linear models (GLMs) with a binomial error structure to examine whether the hosts' personality traits explained their response towards a parasitic egg. In all models, we included the response to the parasitic egg (rejected/accepted) as the dependent variable, and the three measured traits, clutch colour (blue/pink) and cuckoo egg (real/model) as explanatory variables. When testing the effects of breathing rate and exploratory behaviour on egg rejection, we also included cuckoo status (present/absent) as an explanatory variable. In 2020, we started the fieldwork only in late May (because of the COVID-19 pandemic), when cuckoos had already arrived at the study site. Thus, the novel object experiments were all conducted during cuckoo presence. Therefore, we did not include cuckoo status as a fixed effect when examining the effect of return latency on hosts' response to the parasitic egg. We used an information-theoretic approach to establish a candidate set of all possible models, and selected the best-fit model by comparing the corrected Akaike's information criterion (AICc). First, we selected a subset of models with the δ AIC value (ΔAICc) lower than 2 [61,62]. Then, we chose the most parsimonious model (i.e. the one with the smallest number of parameters) from this subset [63]. We calculated the total explanatory power of the model using Nagelkerke's R2 (R package fmsb) [64], and the explanatory power of the parameters retained in the final model was assessed using hierarchical partitioning (R package hier.part) [65]. We also evaluated multicollinearity using the all variance inflation factor (VIF) in the final models. All VIF values were lower than 2, indicating weak correlation between the explanatory variables [66]. Some individuals were subjected twice to the breathing rate measurement, exploration test and parasitic egg experiments. When calculating correlation coefficients and running GLMs, we only used the data of the first measurement or test for these individuals. All statistical analyses were conducted in R 3.4.2 [67]. 3. Results (a) Personality traits Females took significantly longer to return to the nests during the first novel object trials than during the control trials (V = 129, p < 0.001, n = 43), indicating that females showed a neophobia response to the novel object. The return latencies were repeatable across the novel object trials (r = 0.78 (0.63, 0.90), p < 0.0001, n = 70 observations on 43 females), indicating that the neophobia response is a consistent personality trait. Both breathing rate and exploratory behaviour (number of movements) were repeatable across years in Daurian redstarts (breathing rate: r = 0.53 (0.39, 0.78), p < 0.001, n = 343 observations on 307 birds; exploratory behaviour: r = 0.31 (0.10, 0.68), p = 0.03, n = 343 observation on 305 birds). There was a negative relationship between exploratory behaviour and return latency (rs = −0.54, p < 0.001, n = 43; figure 1), i.e. fast-exploring females returned faster to their nests when a novel object was present than slow explorers. However, breathing rate was not significantly related to either return latency (rs = 0.02, p = 0.90, n = 43), or to exploratory behaviour (rp = −0.04, p = 0.47, n = 301). Figure 1. The relationship between exploratory behaviour (number of movements in the novel cage) and neophobia response (return latency). The grey shading indicates the 95% confidence intervals. Only data from the first novel object trial are shown (n = 43 females). (b) Personality and egg rejection Of the 18 redstart nests that were naturally parasitized by cuckoos in our population, five females rejected the egg, 11 females accepted the egg and two nests were predated during the experiment. Out of 97 nests that were artificially parasitized, 50 females rejected the egg, 45 females accepted the egg and two nests were predated during the experiment. The models including the neophobia response, exploratory behaviour and stress response had an R2 value (total variation in egg rejection explained) of 0.30, 0.42 and 0.35, respectively. The neophobia response (return latency) was the only significant predictor in the model (p = 0.005; table 1). Exploratory behaviour was also an important predictor, making up 25.77% of the explained variance (p = 0.012; table 1). The stress response (breathing rate) did not predict egg-rejection behaviour, and was excluded from the model (table 1). The neophobia response and exploratory behaviour are thus important and significant predictors of egg-rejection behaviour in Daurian redstarts, with females that are less neophobic and more explorative being more likely to reject the parasitic egg (figure 2). Table 1. Generalized linear models predicting the probability that a female Daurian redstarts rejected a parasitic egg. (The presented models are the most parsimonious models with a δ AICc lower than 2 (electronic supplementary material, table S2). The full models included the personality trait of interest, egg colour (blue or pink), cuckoo egg type (real or model) and cuckoo status (present or absent). For each fixed effect, the reference category is indicated in parentheses). I (%) is the proportion of the total variance explained by the models. VIF indicates the variance inflation factor for each predictor.) personality trait fixed effect estimate 95% CI Z p-value I (%) VIF neophobia response intercept 5.44 2.22–9.49 2.99 0.003 return latency −1.90 −3.71 to −0.78 −2.47 0.005 100 – exploratory behaviour intercept −2.99 −4.60 to −1.62 −3.97 <0.001 exploration score 0.02 0.01–0.04 2.50 0.012 25.77 1.06 clutch colour (blue) 1.96 0.97–3.06 3.71 <0.001 43.79 1.14 cuckoo egg type (model) −1.64 −3.36 to −0.18 −2.07 0.039 13.55 1.07 cuckoo status (absent) 1.40 0.35–2.54 2.54 0.011 16.89 1.22 stress response intercept −1.71 −2.77 to −0.81 −3.46 <0.001 clutch colour (blue) 1.95 1.01–3.01 3.87 <0.001 55.11 1.13 cuckoo egg type (model) −1.70 −3.37 to −0.30 −2.23 0.026 17.38 1.07 cuckoo status (absent) 1.72 0.73–2.83 3.26 0.001 27.52 1.20 Figure 2. Relationships between the probability of rejecting the parasitic egg and (a) return latency and (b) exploration score (number of movements in novel cage) of the host. The grey shading indicates the 95% confidence intervals from the GLM. Tick marks indicate raw data points. To test the robustness of our finding that egg-rejection behaviour was related to personality type, we additionally performed a Wilcoxon signed-rank test to compare the return latency and exploration score, and a Student's t-test to compare the breathing rate, between females that rejected the egg and females that accepted the egg. Females that rejected the egg indeed had a shorter return latency (mean ± s.d.: 9.6 ± 5.4 min, n = 26) than females that accepted the egg (16.8 ± 8.1 min, n = 17; V = 357, p < 0.001). Females that rejected the egg also had a higher exploration score (80.6 ± 34.7, n = 55) than females that accepted the egg (59.3 ± 22.7, n = 56; V = 936.5, p < 0.001). However, females that rejected the egg had a similar breathing rate (171.7 ± 32.8 min−1, n = 54) as females that accepted the egg (176.0 ± 28.1 min−1, n = 56; t = 0.74, p = 0.46). 4. Discussion This study shows that personality traits of a common host predict the host's response to a brood parasitic egg. Specifically, we show that fast-exploring and less neophobic hosts were more likely to reject parasitic eggs than slow-exploring and more neophobic hosts. We did not find a significant effect of breathing rate, a measure of the stress response of an individual, on host egg-rejection behaviour. We further demonstrate that object neophobia and exploration are repeatable across tests or across years in Daurian redstarts, and thus represent personality traits. Breathing rate was also highly repeatable across years, suggesting that it is a reliable indicator of the stress response in this species. Besides host personality, we found that host clutch colour (reflecting the degree of similarity between host and foreign eggs), the type of parasitic egg and cuckoo presence (perceived risk of brood parasitism) had significant effects on the female's egg-rejection decision (table 1), consistent with our previous findings [44]. Together, these variables explained 30–42% of variation in egg-rejection behaviour. This implies that there is still unexplained variation in egg-rejection behaviour among Daurian redstarts that warrants further investigation. (a) Repeatability of personality traits Exploratory behaviour is a commonly studied personality trait in both captive and wild bird populations [24,36]. In this study, we show that exploration of a novel cage was individually repeatable over years, also in Daurian redstarts. The repeatability value was similar to that reported in other species [28,51]. Most studies on neophobia as a personality trait were conducted with captive birds, but a few studies used individuals of free-living populations. [55,56,68]. Here, we show that the estimate of neophobia (return latencies between the first and second novel object trials) were repeatable in Daurian redstarts. Therefore, our study provides clear evidence supporting that neophobia is a personality trait in this natural population. (b) Correlations between personality traits We detected a negative correlation between exploratory behaviour and neophobia in Daurian redstarts, similar to patterns found in other species [55,56]. Exploratory behaviour, neophobia and boldness are inter-related, and form a bold-shy continuum [28,69], which reflects the best-studied personality axis in non-human animals [70,71]. Empirical evidence suggests that bold individuals tend to be proactive, fast-exploring and risk-taking (less neophobic), while shy individuals are reactive, slow-exploring and risk-averse (more neophobic) [70,72]. We did not find significant relationships between the stress response (breathing rate) and either exploratory behaviour or neophobia. This result is in line with a study on great tits, where no significant correlations were detected between breathing rate and exploration in both city and forest populations [73]. (c) Host personality predicts the response to a parasitic egg According to the host personality hypothesis, personality traits of the host may influence their anti-parasite defence behaviour [18]. The first supportive evidence came from a study on the great reed warbler, showing that more aggressive females (during handling after being caught) had higher levels of nest defence (aggression towards cuckoos), but no relationship between host aggressiveness and egg rejection behaviour was detected [33]. In this study, we show that fast-exploring and less neophobic female hosts were more likely to reject a parasitic egg than slow-exploring and more neophobic females. A study on great reed warblers showed that females which devoted more time to clutch inspection ejected experimental eggs more quickly than hosts inspecting their parasitized clutches only briefly [74], which may at least indirectly suggest a positive relationship between host exploratory behaviour and egg rejection. According to the bold-shy continuum, our results suggest that bold hosts are more likely to reject parasitic eggs than shy individuals, which contrasts with the prediction from Aviles & Parejo [18]. One adaptive explanation regarding the existence of the bold-shy continuum is that bold individuals tend to maintain high productivity but at a potential cost to their survival, whereas shy individuals do the opposite [26,69,75]. Studies exploring the fitness consequences of variation in personality largely support this hypothesis, showing that bold individuals often outperform shy ones in terms of reproductive success, but also have reduced survival [26,27,69,76]. In the context of brood parasitism, a trade-off might also exist if bold hosts tend to reject the parasitic egg but at a potential cost stemming from recognition error or punishment by the cuckoo, whereas shy hosts accept (or tolerate) the brood parasitism but suffer reduced reproductive success if the cuckoo egg hatches. Moreover, bold individuals may be more active and hence may be more likely to attract the attention of a cuckoo to their nest, leading to a higher risk of parasitism than in shy (passive) hosts [42]. Higher levels of parasitism would then favour bold hosts to become egg rejecters [18]. Egg-rejection behaviour may be further mediated or directly regulated by physiological mechanisms [77]. For example, a study on American robins Turdus migratorius showed that hosts with higher levels of corticosterone, a hormone linked to the stress response, were more likely to reject a parasitic egg [78]. Another study showed that decreasing the levels of prolactin facilitated Eurasian blackbirds Turdus merula to reject foreign eggs [79]. The shy-bold continuum may also reflect variation in a range of physiological traits [69,80], including the stress response [81]. For example, bold (proactive) birds generally have relatively low corticosterone responses whereas shy (reactive) individuals have relatively high corticosterone responses [82,83], but see [84]. However, we did not find a relationship between the probability of egg rejection and breathing rate, a physiological trait related to the stress response [57]. In this study, all personality assays were conducted after the egg-rejection experiments (see Methods). Therefore, hosts that perceived that they were parasitized (and rejected the egg) may have become bolder as a consequence of the treatment. However, this reversal of causation seems less likely, because exploratory behaviour is consistent across years, despite the fact that individuals vary in their experience with artificial or natural brood parasitism. In conclusion, Daurian redstart females showed strong covariation between exploration and neophobia in the wild, and host personality traits predicted the response to cuckoo parasitism. Specifically, bold hosts (fast-exploring and less neophobic) were more likely to reject a parasitic egg than shy females (slow-exploring and more neophobic). This implies that a cuckoo would have lower success (fitness) when parasitizing a bold host. In the coevolutionary arms race, selection should therefore favour cuckoos that lay their eggs in the nest of a shy host. Whether cuckoos pay attention to the personality of the host or whether bold hosts are more likely to attract a cuckoo warrants further study. Supplementary Material Click here for additional data file. Acknowledgments We are grateful to Eunbi Kwon for her helpful comments with data analysis. We thank Yongji Forestry Bureau, Jilin, China, for their permission and cooperation. We also thank Mengjia Wu, Silvia Quilici, Ziyin Xiong, Rui Wang, Jie Deng, Guobin Zheng and Yangbo Kou for assistance with fieldwork. Ethics The experiments comply with the current laws of China, where they were performed. Fieldwork was carried out with permission from Yongji Forestry Bureau, Jilin, China. Experimental procedures were conducted under license from the Animal Management Committee at the College of Life Sciences, Beijing Normal University (permit no. CLS-EAW-2018-001). Data accessibility Data available from the Dryad Digital Repository: https://doi.org/10.5061/dryad.xwdbrv1cs [85]. Authors' contributions J.Z.: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, software, validation, visualization, writing—original draft, writing—review and editing; P.S.: conceptualization, formal analysis, methodology, software, visualization, writing—review and editing; J.L.: conceptualization, formal analysis, methodology, software, visualization, writing—review and editing; L.Y.: data curation, investigation, resources, writing—review and editing; W.D.: conceptualization, funding acquisition, investigation, methodology, project administration, resources, supervision, validation, writing—review and editing; B.K.: conceptualization, formal analysis, methodology, project administration, supervision, validation, writing—review and editing. All authors gave final approval for publication and agreed to be held accountable for the work performed therein. Competing interests We declare we have no competing interests. Funding This study was supported by the National Natural Science Foundation of China grant nos (31372219 and 31672297 to W.D.), and the China Scholarship Council grant no. (201906040159 to J.Z.). ==== Refs References 1. Davies NB. 2000 Cuckoos, cowbirds and other cheats. London, UK: T & AD Poyser. 2. Møller AP, Diaz M, Liang W. 2016 Brood parasitism and proximity to human habitation. Behav. Ecol. 27 , 1314-1319. (10.1093/beheco/arw049) 3. Tolvanen J, Forsman JT, Thomson RL. 2017 Reducing cuckoo parasitism risk via informed habitat choices. Auk 134 , 553-563. (10.1642/Auk-17-30.1) 4. Trnka A, Grim T. 2013 Color plumage polymorphism and predator mimicry in brood parasites. Front. Zool. 10 , 25-25. (10.1186/1742-9994-10-25)23663311 5. Spottiswoode CN, Stevens M. 2010 Visual modeling shows that avian host parents use multiple visual cues in rejecting parasitic eggs. Proc. Natl Acad. Sci. USA 107 , 8672-8676. (10.1073/pnas.0910486107)20421497 6. Yang C, Wang L, Cheng S, Hsu Y, Stokke BG, Røskaft E, Moksnes A, Liang W, Møller AP. 2015 Deficiency in egg rejection in a host species as a response to the absence of brood parasitism. Behav. Ecol. 26 , 406-415. (10.1093/beheco/aru202) 7. Feeney WE, Welbergen JA, Langmore NE. 2014 Advances in the study of coevolution between avian brood parasites and their hosts. Annu. Rev. Ecol. Evol. 45 , 227-246. (10.1146/annurev-ecolsys-120213-091603) 8. Soler M. 2014 Long-term coevolution between avian brood parasites and their hosts. Biol. Rev. 89 , 688-704. (10.1111/brv.12075)24330159 9. Sato NJ, Tokue K, Noske RA, Mikami OK, Ueda K. 2010 Evicting cuckoo nestlings from the nest: a new anti-parasitism behaviour. Biol. Lett. 6 , 67-69. (10.1098/rsbl.2009.0540)19776068 10. Davies NB, Brooke MD. 1989 An experimental-study of co-evolution between the cuckoo, Cuculus Canorus, and its hosts. I. Host egg discrimination. J. Anim. Ecol. 58 , 207-224. (10.2307/4995) 11. Moksnes A, Roskaft E, Braa AT. 1991 Rejection behavior by common cuckoo hosts towards artificial brood parasite eggs. Auk 108 , 348-354. (10.1093/auk/108.2.348) 12. Davies NB, Brooke MDL, Kacelnik A. 1996 Recognition errors and probability of parasitism determine whether reed warblers should accept or reject mimetic cuckoo eggs. Proc. R. Soc. B 263 , 925-931. (10.1098/rspb.1996.0137) 13. Moskát C, Bán M, Hauber ME. 2014 Naïve hosts of avian brood parasites accept foreign eggs, whereas older hosts fine-tune foreign egg discrimination during laying. Front. Zool. 11 , 45. (10.1186/1742-9994-11-45)25024736 14. Yang C et al. 2010 Coevolution in action: disruptive selection on egg colour in an avian brood parasite and its host. PLoS ONE 5 , e10816. (10.1371/journal.pone.0010816)20520815 15. Soler M, Martín-Vivaldi M, Fernandezmorante J. 2012 Conditional response by hosts to parasitic eggs: the extreme case of the rufous-tailed scrub robin. Anim. Behav. 84 , 421-426. (10.1016/j.anbehav.2012.05.016) 16. Moksnes A, Roskaft E, Hagen LG, Honza M, Mork C, Olsen PH. 2000 Common cuckoo Cuculus canorus and host behaviour at reed warbler Acrocephalus scirpaceus nests. Ibis 142 , 247-258. (10.1111/j.1474-919X.2000.tb04864.x) 17. Moksnes A, Roskaft E, Korsnes L. 1993 Rejection of cuckoo Cuculus canorus eggs by meadow pipits Anthus pratensis. Behav. Ecol. 4 , 120-127. (10.1093/beheco/4.2.120) 18. Aviles JM, Parejo D. 2011 Host personalities and the evolution of behavioural adaptations in brood parasitic-host systems. Anim. Behav. 82 , 613-618. (10.1016/j.anbehav.2011.07.025) 19. Dall SRX, Houston AI, McNamara JM. 2004 The behavioural ecology of personality: consistent individual differences from an adaptive perspective. Ecol. Lett. 7 , 734-739. (10.1111/j.1461-0248.2004.00618.x) 20. Dingemanse NJ, Both C, Drent PJ, Tinbergen JM. 2004 Fitness consequences of avian personalities in a fluctuating environment. Proc. R. Soc. B 271 , 847-852. (10.1098/rspb.2004.2680) 21. Sih A, Bell A, Johnson JC. 2004 Behavioral syndromes: an ecological and evolutionary overview. Trends Ecol. Evol. 19 , 372-378. (10.1016/j.tree.2004.04.009)16701288 22. Sih A, Bell AM. 2008 Insights for behavioral ecology from behavioral syndromes. Adv. Stud. Behav. 38 , 227-281. (10.1016/S0065-3454(08)00005-3) 23. Verbeek MEM, Boon A, Drent PJ. 1996 Exploration, aggressive behavior and dominance in pair-wise confrontations of juvenile male great tits. Behaviour 133 , 945-963. (10.1163/156853996(00314) 24. Van Oers K, Naguib M. 2013 Avian personality. Chicago, IL: University of Chicago. 25. Thys B, Pinxten R, Raap T, De Meester G, Rivera-Gutierrez HF, Eens M. 2017 The female perspective of personality in a wild songbird: repeatable aggressiveness relates to exploration behaviour. Sci. Rep. 7 , 7656. (10.1038/s41598-017-08001-1)28794486 26. Biro PA, Stamps JA. 2008 Are animal personality traits linked to life-history productivity? Trends Ecol. Evol. 23 , 361-368. (10.1016/j.tree.2008.04.003)18501468 27. Smith BR, Blumstein DT. 2008 Fitness consequences of personality: a meta-analysis. Behav. Ecol. 19 , 448-455. (10.1093/beheco/arm144) 28. Hall ML, van Asten T, Katsis AC, Dingemanse NJ, Magrath MJL, Mulder RA. 2015 Animal personality and pace-of-life syndromes: do fast-exploring fairy-wrens die young? Front. Ecol. Evol. 3 , 28. (10.3389/fevo.2015.00028) 29. Dingemanse NJ, Both C, van Noordwijk AJ, Rutten AL, Drent PJ. 2003 Natal dispersal and personalities in great tits (Parus major). Proc. R. Soc. B 270 , 741-747. (10.1098/rspb.2002.2300) 30. Van Oers K, Drent PJ, Dingemanse NJ, Kempenaers B. 2008 Personality is associated with extrapair paternity in great tits, Parus major. Anim. Behav. 76 , 555-563. (10.1016/j.anbehav.2008.03.011) 31. Hollander FA, Van Overveld T, Tokka I, Matthysen E. 2008 Personality and nest defence in the great tit Parus major. Ethology 114 , 405-412. (10.1111/j.1439-0310.2008.01488.x) 32. Amy M, Sprau P, de Goede P, Naguib M. 2010 Effects of personality on territory defence in communication networks: a playback experiment with radio-tagged great tits. Proc. R. Soc. B 277 , 3685-3692. (10.1098/rspb.2010.0598) 33. Trnka A, Grim T. 2014 Testing for correlations between behaviours in a cuckoo host: why do host defences not covary? Anim. Behav. 92 , 185-193. (10.1016/j.anbehav.2014.04.006) 34. Lotem A, Nakamura H, Zahavi A. 1995 Constraints on egg discrimination and cuckoo host coevolution. Anim. Behav. 49 , 1185-1209. (10.1006/anbe.1995.0152) 35. Guillette LM, Reddon AR, Hurd PL, Sturdy CB. 2009 Exploration of a novel space is associated with individual differences in learning speed in black-capped chickadees, Poecile atricapillus. Behav. Process. 82 , 265-270. (10.1016/j.beproc.2009.07.005) 36. Verbeek MEM, Drent PJ, Wiepkema PR. 1994 Consistent individual-differences in early exploratory-behavior of male great tits. Anim. Behav. 48 , 1113-1121. (10.1006/anbe.1994.1344) 37. Boogert NJ, Reader SM, Laland KN. 2006 The relation between social rank, neophobia and individual learning in starlings. Anim. Behav. 72 , 1229-1239. (10.1016/j.anbehav.2006.02.021) 38. Guillette LM, Reddon AR, Hoeschele M, Sturdy CB. 2011 Sometimes slower is better: slow-exploring birds are more sensitive to changes in a vocal discrimination task. Proc. R. Soc. B. 278 , 767-773. (10.1098/rspb.2010.1669) 39. Soler M, Pérez-Contreras T, Soler JJ. 2017 Brood parasites as predators: farming and Mafia strategies. In Avian brood parasitism behaviour, ecology, evolution and coevolution (ed. M Soler), pp. 271-286. Cham, Switzerland: Springer. 40. Britton NF, Planque R, Franks NR. 2007 Evolution of defence portfolios in exploiter-victim systems. B. Math. Biol. 69 , 957-988. (10.1007/s11538-006-9178-5) 41. Grim T. 2006 The evolution of nestling discrimination by hosts of parasitic birds: why is rejection so rare? Evol. Ecol. Res. 8 , 785-802. 42. Banks AJ, Martin TE. 2001 Host activity and the risk of nest parasitism by brown-headed cowbirds. Behav. Ecol. 12 , 31-40. (10.1093/oxfordjournals.beheco.a000375) 43. Gill SA, Grieef PM, Staib LM, Sealy SG. 1997 Does nest defence deter or facilitate cowbird parasitism? A test of the nesting-cue hypothesis. Ethology 103 , 56-71. (10.1111/j.1439-0310.1997.tb00007.x) 44. Zhang J, Møller AP, Yan D, Li J, Deng W. 2021 Egg rejection changes with seasonal variation in risk of cuckoo parasitism in Daurian redstarts Phoenicurus auroreus. Anim. Behav. 175 , 193-200. (10.1016/j.anbehav.2021.03.007) 45. Yang C, Li Z, Zhang Y, Wang H, Liang W, Møller AP. 2016 Egg polymorphism and egg discrimination in the Daurian redstart Phoenicurus auroreus, a host of the common cuckoo Cuculus canorus. Ornithol. Sci. 15 , 127-132. (10.2326/osj.15.127) 46. Yang C et al. 2016 Plaintive cuckoos do not select tailorbird hosts that match the phenotypes of their own eggs. Behav. Ecol. 27 , 835-841. (10.1093/beheco/arv226) 47. Samas P, Rutila J, Grim T. 2016 The common redstart as a suitable model to study cuckoo-host coevolution in a unique ecological context. BMC Evol. Biol. 16 , 255. (10.1186/s12862-016-0835-5)27887566 48. Thomson RL, Tolvanen J, Forsman JT. 2016 Cuckoo parasitism in a cavity nesting host: near absent egg-rejection in a northern redstart population under heavy apparent (but low effective) brood parasitism. J. Avian Biol. 47 , 363-370. (10.1111/jav.00915) 49. Kluen E, Kuhn S, Kempenaers B, Brommer JE. 2012 A simple cage test captures intrinsic differences in aspects of personality across individuals in a passerine bird. Anim. Behav. 84 , 279-287. (10.1016/j.anbehav.2012.04.022) 50. Herborn KA, Macleod R, Miles WTS, Schofield ANB, Alexander L, Arnold KE. 2010 Personality in captivity reflects personality in the wild. Anim. Behav. 79 , 835-843. (10.1016/j.anbehav.2009.12.026) 51. Stuber EF, Araya-Ajoy YG, Mathot KJ, Mutzel A, Nicolaus M, Wijmenga JJ, Mueller JC, Dingemanse NJ. 2013 Slow explorers take less risk: a problem of sampling bias in ecological studies. Behav. Ecol. 24 , 1092-1098. (10.1093/beheco/art035) 52. Dingemanse NJ, Both C, Drent PJ, Van Oers K, Van Noordwijk AJ. 2002 Repeatability and heritability of exploratory behaviour in great tits from the wild. Anim. Behav. 64 , 929-938. (10.1006/anbe.2002.2006) 53. Webster SJ, Lefebvre L. 2001 Problem solving and neophobia in a columbiform-passeriform assemblage in Barbados. Anim. Behav. 62 , 23-32. (10.1006/anbe.2000.1725) 54. Greggor AL, Clayton NS, Fulford AJC, Thornton A. 2016 Street smart: faster approach towards litter in urban areas by highly neophobic corvids and less fearful birds. Anim. Behav. 117 , 123-133. (10.1016/j.anbehav.2016.03.029)27429456 55. Cole EF, Quinn JL. 2014 Shy birds play it safe: personality in captivity predicts risk responsiveness during reproduction in the wild. Biol. Lett. 10 , 20140178. (10.1098/rsbl.2014.0178)24829251 56. Grunst AS, Grunst ML, Pinxten R, Eens M. 2019 Personality and plasticity in neophobia levels vary with anthropogenic disturbance but not toxic metal exposure in urban great tits: urban disturbance, metal pollution and neophobia. Sci. Total Environ. 656 , 997-1009. (10.1016/j.scitotenv.2018.11.383)30625686 57. Carere C, van Oers K. 2004 Shy and bold great tits Parus major: body temperature and breath rate in response to handling stress. Physiol. Behav. 82 , 905-912. (10.1016/j.physbeh.2004.07.009)15451657 58. Torne-Noguera A, Pagani-Nunez E, Senar JC. 2014 Great tit Parus major breath rate in response to handling stress: urban and forest birds differ. J. Ornithol. 155 , 315-318. (10.1007/s10336-013-1025-5) 59. Nakagawa S, Schielzeth H. 2010 Repeatability for Gaussian and non-Gaussian data: a practical guide for biologists. Biol. Rev. 85 , 935-956. (10.1111/j.1469-185X.2010.00141.x)20569253 60. Stoffel MA, Nakagawa S, Schielzeth H. 2017 rptR: repeatability estimation and variance decomposition by generalized linear mixed-effects models. Methods Ecol. Evol. 8 , 1639-1644. (10.1111/2041-210x.12797) 61. Wagenmakers E, Farrell S. 2004 AIC model selection using Akaike weights. Psychon. Bull. Rev. 11 , 192-196. (10.3758/BF03206482)15117008 62. Burnham KPKP, Anderson DRDR. 2002 Model selection and multimodel inference: a practical information-theoretic approach, 2nd edn. New York, NY: Springer. 63. Arnold TW. 2010 Uninformative parameters and model selection using Akaike's information criterion. J. Wildl. Manage. 74 , 1175-1178. (10.1111/j.1937-2817.2010.tb01236.x) 64. Nakazawa M., Nakazawa MM. 2019 Package ‘fmsb’. See https://cran.r-project.org/web/packages/fmsb/fmsb.pdf. 65. Walsh C, Nally RM. 2013 hier.part: hierarchical partitioning. See https://cran.r-project.org/web/packages/hier.part/index.html. 66. Zuur AF, Ieno EN, Elphick CS. 2010 A protocol for data exploration to avoid common statistical problems. Methods Ecol. Evol. 1 , 3-14. (10.1111/j.2041-210X.2009.00001.x) 67. R Core Team. 2017 R: a language and environment for statistical computing (version 3.4. 2). Vienna, Austria: R Foundation for Statistical Computing. 68. Krams IA, Vrublevska J, Sepp T, Abolins-Abols M, Rantala MJ, Mierauskas P, Krama T. 2014 Sex-specific associations between nest defence, exploration and breathing rate in breeding pied flycatchers. Ethology 120 , 492-501. (10.1111/eth.12222) 69. Réale D, Garant D, Humphries MM, Bergeron P, Careau V, Montiglio PO. 2010 Personality and the emergence of the pace-of-life syndrome concept at the population level. Phil. Trans. R. Soc. B. 365 , 4051-4063. (10.1098/rstb.2010.0208)21078657 70. Koolhaas JM, Korte SM, De Boer SF, Van Der Vegt BJ, Van Reenen CG, Hopster H, De Jong IC, Ruis MAW, Blokhuis HJ. 1999 Coping styles in animals: current status in behavior and stress-physiology. Neurosci. Biobehav. R. 23 , 925-935. (10.1016/S0149-7634(99)00026-3) 71. Groothuis TGG, Carere C. 2005 Avian personalities: characterization and epigenesis. Neurosci. Biobehav. R. 29 , 137-150. (10.1016/j.neubiorev.2004.06.010) 72. Coppens CM, de Boer SF, Koolhaas JM. 2010 Coping styles and behavioural flexibility: towards underlying mechanisms. Phil. Trans. R. Soc. B 365 , 4021-4028. (10.1098/rstb.2010.0217)21078654 73. Charmantier A, Demeyrier V, Lambrechts M, Perret S, Gregoire A. 2017 Urbanization is associated with divergence in pace-of-life in great tits. Front. Ecol. Evol. 5 , 53. (10.3389/fevo.2017.00053) 74. Požgayová M, Procházka P, Polačiková L, Honza M. 2011 Closer clutch inspection-quicker egg ejection: timing of host responses toward parasitic eggs. Behav. Ecol. 22 , 46-51. (10.1093/beheco/arq163) 75. Wolf M, van Doorn GS, Leimar O, Weissing FJ. 2007 Life-history trade-offs favour the evolution of animal personalities. Nature 447 , 581-584. (10.1038/nature05835)17538618 76. Quinn JL, Cole EF, Bates J, Payne RW, Cresswell W. 2012 Personality predicts individual responsiveness to the risks of starvation and predation. Proc. R. Soc. B 279 , 1919-1926. (10.1098/rspb.2011.2227) 77. Abolins-Abols M, Hauber ME. 2018 Host defences against avian brood parasitism: an endocrine perspective. Proc. R. Soc. B 285 , 20180980. (10.1098/rspb.2018.0980) 78. Abolins-Abols M, Hauber ME. 2020 Endocrine regulation of egg rejection in an avian brood parasite host. Biol. lett. 16 , 20200225. (10.1098/rsbl.2020.0225)32574535 79. Ruiz-Raya F, Ibanez-Alamo JD, Parenteau C, Chastel O, Soler M. 2020 Prolactin mediates behavioural rejection responses to avian brood parasitism. bioRxiv. (10.1101/2020.10.28.358994) 80. Hennemann WW. 1983 Relationship among body-mass, metabolic-rate and the intrinsic rate of natural increase in mammals. Oecologia 56 , 104-108. (10.1007/Bf00378224)28310776 81. Koolhaas JM, de Boer SF, Coppens CM, Buwalda B. 2010 Neuroendocrinology of coping styles: towards understanding the biology of individual variation. Front. Neuroendocrin. 31 , 307-321. (10.1016/j.yfrne.2010.04.001) 82. Cockrem JF. 2007 Stress, corticosterone responses and avian personalities. J. Ornithol. 148 , S169-S178. (10.1007/s10336-007-0175-8) 83. Cockrem JF. 2013 Corticosterone responses and personality in birds: individual variation and the ability to cope with environmental changes due to climate change. Gen. Comp. Endocr. 190 , 156-163. (10.1016/j.ygcen.2013.02.021)23467071 84. Medina-García A, Jawor JM, Wright TF. 2017 Cognition, personality, and stress in budgerigars, Melopsittacus undulatus. Behav. Ecol. 28 , 1504-1516. (10.1093/beheco/arx116)29622930 85. Zhang J, Santema P, Li J, Yang L, Deng W, Kempenaers B. 2021 Data from: Host personality predicts cuckoo egg rejection in Daurian redstarts Phoenicurus auroreus. Dryad Digital Repository. ( 10.5061/dryad.xwdbrv1cs )
PMC008xxxxxx/PMC8206686.txt
==== Front Proc Biol Sci Proc Biol Sci RSPB royprsb Proceedings of the Royal Society B: Biological Sciences 0962-8452 1471-2954 The Royal Society 34130506 10.1098/rspb.2021.0908 rspb20210908 10016069181Ecology Research Articles DNA metabarcoding reveals trophic niche diversity of micro and mesozooplankton species DNA Metabarcoding reveals trophic niche diversity of micro and mesozooplankton species http://orcid.org/0000-0001-8910-6183 Novotny Andreas Conceptualization Data curation Formal analysis Investigation Methodology Project administration Software Validation Visualization Writing-original draft Writing-review & editing [email protected] http://orcid.org/0000-0002-7822-3197 Zamora-Terol Sara Conceptualization Data curation Investigation Methodology Project administration Validation Writing-review & editing † Winder Monika Conceptualization Funding acquisition Project administration Resources Supervision Writing-review & editing Department of Ecology, Environment, and Plant Science, Stockholm University, Svante Arrhenius Väg 20A, 106 91 Stockholm, Sweden † Present address: Institute of Marine Research, Bergen, Norway. Electronic supplementary material is available online at https://doi.org/10.6084/m9.figshare.c.5448706. 30 6 2021 June 30, 2021 16 6 2021 June 16, 2021 16 6 2021 June 16, 2021 288 1953 2021090818 4 2021 April 18, 2021 17 5 2021 May 17, 2021 © 2021 The Authors. 2021 https://creativecommons.org/licenses/by/4.0/ Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited. Alternative pathways of energy transfer guarantee the functionality and productivity in marine food webs that experience strong seasonality. Nevertheless, the complexity of zooplankton interactions is rarely considered in trophic studies because of the lack of detailed information about feeding interactions in nature. In this study, we used DNA metabarcoding to highlight the diversity of trophic niches in a wide range of micro- and mesozooplankton, including ciliates, rotifers, cladocerans, copepods and their prey, by sequencing 16- and 18S rRNA genes. Our study demonstrates that the zooplankton trophic niche partitioning goes beyond both phylogeny and size and reinforces the importance of diversity in resource use for stabilizing food web efficiency by allowing for several different pathways of energy transfer. We further highlight that small, rarely studied zooplankton (rotifers and ciliates) fill an important role in the Baltic Sea pelagic primary production pathways and the potential of ciliates, rotifers and crustaceans in the utilization of filamentous and picocyanobacteria within the pelagic food web. The approach used in this study is a suitable entry point to ecosystem-wide food web modelling considering species-specific resource use of key consumers. zooplankton , food web , trophic niche diversity , metabarcoding , rotifer Swedish Research Council 2016-04685 cover-dateJune 30, 2021 ==== Body pmc1. Introduction The ability for ecosystems to maintain functionality and productivity under annual and seasonal variation in primary production relies on energy transfer pathways sustained by a network of diverse primary consumers [1,2]. In marine food webs, functionally diverse assemblages of planktonic bacteria, protists and metazoans regulate the flow of energy from primary producers to higher trophic levels [3–5]. While crustacean zooplankton (e.g. copepods and cladocerans) constitute the primary link between phytoplankton and planktivorous fish [6], microzooplankton are main grazers of primary production at times when the biomass of phytoplankton is low or inedible [3,7]. In order to estimate the resilience of marine ecosystems, a mechanistic understanding of resource use by the primary consumers is needed [8]. However, in most food web studies, the trophic niche is based on size or phylogeny due to a lack of detailed information about feeding interactions in nature. Consequently, the entire niche diversity of the zooplankton community is not accurately considered [9,10]. Variation in temporal abundance, feeding traits, size, phenotypic plasticity, growth rate and predation resistance contribute to the total diversity of zooplankton functional groups in marine food webs [11]. While most trophic studies have clustered zooplankton into broad phylogenetic groups [9], recent studies show that incorporating traits, particularly size, has consequences for interpreting food web dynamics and productivity [12–14]. For example, the rotifer phylum contains members of different size classes [15], as well as organisms with various feeding behaviours including filter feeders [16], selective feeders [17–19] and in some cases even carnivores [20]. Similarly, copepods and cladocerans can perform different feeding strategies including, among others, feeding-current and ambush feeding [21], thereby using a wide spectrum of resources. Consequently, trophic niche partitioning at a high phylogenetic level, such as class or family, will underestimate the trophic niche diversity in zooplankton guilds. Traditionally, method limitations have made it difficult to assess trophic niche differences between zooplankton species. Experiments to estimate the grazing impact of zooplankton are time-consuming and elaborate, and the amount of biological material required for biogeochemical tracer studies on plankton communities [22] often exceeds what is feasible to sort out from diverse zooplankton samples. Consequently, few species are often included in these studies, limiting the ability to describe key niche differences between them. Similarly, the diet spectrum in plankton experimental studies is often limited to a small number of a priori hypothetical prey species and may not display the food web's full complexity with enough resolution. The challenges and limitations of studying feeding traits of zooplankton species have further created a biased knowledge towards larger organisms in the food web that are more frequently studied [23]. Since most of the zooplankton feeding studies are conducted at different sites and times of the year, often using different methods, the comparison of existing information on zooplankton trophic niches is laborious. DNA metabarcoding of selected organisms has proven to be a useful tool for resolving trophic interactions [24,25] and is increasingly being used for studying trophic interactions of zooplankton [26–28]. With the high sensitivity of the polymerase chain reaction, metabarcoding requires very little biological material and is a non-a priori method with high taxonomic resolution. Metabarcoding allows for a food web-oriented approach as several zooplankton species can be investigated simultaneously [29], thereby providing detailed insights on trophic interactions and better linking the trophic niche diversity with energy flow. Detailed knowledge about trophic interactions can be of particular importance for coastal ecosystems that experience a shift in phytoplankton community with an increase of cyanobacteria due to climate warming and eutrophication [30–32]. The Baltic Sea is a prime example, illustrating that nitrogen fixed by cyanobacteria supports the productivity of upper trophic levels [33,34]. Yet, as filamentous cyanobacteria are often considered unpalatable for copepods [35], the mechanism of trophic incorporation is not fully understood. While rotifers and microzooplankton are abundant in the Baltic Sea (figure 1), their potential trophic link with cyanobacteria is seldom considered. Without knowledge about trophic partitioning between micro and mesozooplankton, the possible fate and sinks of cyanobacterial production in the plankton food web remain largely unknown. Figure 1. (a) Abundance of zooplankton and (b) biovolume of phytoplankton at Landsort Deep in the Baltic Sea. Interpolated daily means over the years 2006–2018. The data are available at the Swedish national archive for oceanographic data: https://sharkweb.smhi.se/. Samples are taken weekly to bi-weekly during the spring and summer period and monthly during winter. The points indicate (a) taxa abundance and (b) biovolume at the date of sampling during this study. (Online version in colour.) In this study, we aimed to investigate the trophic niches of functionally diverse groups of zooplankton spanning both size and phylum. By sequencing 18S rRNA and 16S rRNA genes, we analysed zooplankton-associated prey of selected individuals of different size classes, including a ciliate, rotifers, copepods and cladocerans. The study was done at an offshore station in the Baltic Sea, where ciliates, rotifers and crustaceans at times dominate the zooplankton community. Our results show that the trophic niche diversity extends beyond broad phylogenetic groups and size classes and that small, rarely studied zooplankton fill an important role in the pathways of the coastal pelagic primary production. 2. Methods (a) Sampling Zooplankton and water samples were collected at Landsort Deep monitoring station BY31 (58′35 N18′14 E) located in the eastern Baltic Sea proper, an offshore station at the deepest location of the Baltic Sea with 495 m depth. This frequently monitored offshore station experiences strong seasonal changes in biotic and abiotic properties (see electronic supplementary material, figures S1 and S2). To capture the seasonality of zooplankton (figure 1), samples were collected on 19 June and 15 August 2017, and on 16 March 2018, synchronized with the Swedish national pelagic monitoring programme [36]. Water samples, used for validation and to describe the potential zooplankton prey present in the water, were collected with 10 l Niskin bottles with 5 m depth intervals above the thermocline (0–30 m depth). The depths were mixed by adding an equal volume of water from the Niskin bottles; 1–3 l was sequentially filtered onto 25 mm diameter filters with 20 µm (nylon), 2 µm and 0.2 µm (polycarbonate) pore size. Filters were stored frozen at −80°C until further analysis. Zooplankton samples were collected with three vertical hauls from 0–30 m, 30–60 m and 60–100 m using a 90 µm WP2 closing plankton net (Hydrobios, Kiel, Germany). Ciliates were sampled with a 55 µm hand-towed plankton net in the upper 10 m layer (Hydrobios, Kiel, Germany). The zooplankton and ciliate samples were immediately preserved in 95% ethanol. (b) Zooplankton sorting and DNA metabarcoding Individuals of abundant zooplankton species were identified under a stereomicroscope and selected from depth layers where they were most abundant (electronic supplementary material, table S1 and figure S2). This includes the rotifers Synchaeta baltica, Synchaeta monopus and Keratella spp., the cladocerans Evadne nordmanni and Bosmina spp., and the copepods Temora longicornis, Acartia spp., Pseudocalanus spp. and Centropages hamatus. All individual rotifers were rinsed five times in ethanol; crustaceans were rinsed five times in miliQ water and after that soaked for 30 s in a 1% bleach solution to remove contamination of external DNA. Five to 12 individuals from each species were randomly pooled into one sample tube and stored in 180 µl ALT lysis buffer (Qiagen, Hilden, Germany). A representative of the protozooplankton community, the ciliate Helicostomella, was transferred from the zooplankton samples onto a PET membrane-coated glass slide (Zeiss, Oberkochen, Germany) and covered with resin-based liquid cover glass (Zeiss). Single cells of Helicostomella were collected using a laser capture microdissection microscope (Zeiss) and 10–15 individuals per sample pooled into 10 µl ALT lysis buffer (Qiagen). All of the sorted zooplankton samples were prepared in at least five replicates that were treated separately in all downstream analyses. In the DNA metabarcoding analysis, we amplified a 500 bp long fragment of the V3–V4 region of the 16S rRNA gene (16S) using universal primers 341F and 805R targeting both bacterial and plastidial 16S of phototrophic eukaryotes [37,38], and a 400 bp long fragment of the V4 region of the 18S rRNA gene (18S) using the primers 528F and 706R [39]. The amplicon libraries were sequenced on MiSeq (MSC 2.5.0.5/RTA 1.18.54) pair-end set-up (2 × 300 bp, v. 3, Illumina, San Diego, California). DNA sequences and associated metadata were uploaded to the European Nucleotide Archive (ENA) under accession no. PRJEB39191. 16S sequences were assigned to a custom-made database combining the SILVA 16S reference database [40] with the PhytoREF database [41] to achieve an adequate taxonomic resolution for both prokaryotes and photoautotrophic eukaryotes. 18S sequences were assigned to the Protist Ribosomal Reference database [42]. Details of sample processing, sequencing and bioinformatic analysis can be found in the electronic supplementary material. (c) Data analysis and visualization Data filtering and statistical analysis were facilitated by the Phyloseq R package [43]. All sequences originating from the respective zooplankton consumer species in each sample were removed before data visualization. Heterogeneous sequencing depth was controlled for using subsampling (rarefaction) and subsequent conversion to relative abundance. Non-metric multidimensional scaling plots were based on Bray–Curtis distances and calculated with the ‘metaMDS’ function in the Vegan R package [44]. We used Bray–Curtis similarity index to assess diet overlap between samples (1-Bray–Curtis distance). Differences in the proportion of the specific diet of consumers were modelled with β regression using the ‘betareg’ function in R. We used the ‘simper’ function in the Vegan package to decide which prey species contributed most to the differences in diet overlap. Figures were made using the ggplot2 R package [45]. The most important prevalent taxa (determined as taxa occupying at least 0.1 per cent of the sequences in at least 70 per cent of the samples in each sample group) were visualized in bipartite networks made in the Circlize R package [46]. All data used for the statistical analysis and plotting together with the R scripts to generate the figures were uploaded to the Dryad Digital Repository [47]. 3. Results (a) Diversity of biotic associations The Illumina sequencing effort produced over 37 million sequence reads that passed quality control. The 16S rRNA gene (16S) that targets bacteria and photoautotrophic eukaryotes (plastids), generated 1492 unique ribosomal sequence variants (RSVs) of which 988 were found in the bulk water samples and 996 found in the selective zooplankton samples. The 18S rRNA gene (18S) that targets all eukaryotes generated 3267 RSVs, of which 2258 were in the bulk water samples and 1394 found in the zooplankton samples. We found a broad range of organisms associated with the zooplankton organisms, including heterotrophic and autotrophic bacteria, phytoplankton, protozoans and metazoans. We found that, on average, 85% of the 16S sequence reads associated with the zooplankton samples were proteobacteria, which varied between zooplankton species and season (electronic supplementary material, appendix S1 and figure S4). Among photoautotrophic taxa (cyanobacteria and plastic-containing eukaryotes), associations of zooplankton consumer samples were dominated by cyanobacteria, green algae (Chlorophyta), diatoms (Bacillariophyta) and dinoflagellates (Dinophyceae) (figure 2a). Based on the 18S reads, the zooplankton species were associated with a diversity of eukaryotic organisms, comprising both photoautotrophic and heterotrophic plankton and a diversity of potential symbiotic or parasitic organisms with oomycetes and dinoflagellates (figure 2b). Figure 2. Relative abundance of sequence counts per family of (a) 16S rRNA gene reads (photoautotrophic organisms only) and (b) 18S rRNA, for different zooplankton consumer species and months in the Baltic Sea. The bars represent unique biological replicates. (Online version in colour.) (b) Trophic niche diversity in spring During the spring months, from March to June, the rotifer Synchaeta baltica was the dominating zooplankton species in the Baltic Sea proper, accompanied by less abundant copepod species (figure 1a). The main primary producers were bloom-forming dinoflagellates and diatoms, but also the mixotrophic ciliate Myrionecta (figure 1b). In March, at the beginning of spring bloom, diet overlap between the zooplankton species was relatively low, according to the 16S reads. The rotifer S. baltica had a diet overlap between 0.1 and 0.17 with the copepod groups, while the highest overlap in diet was found between the copepods Temora and Pseudocalanus (0.53) (figure 3b; electronic supplementary material, figure S5). The rotifer S. baltica was mainly associated with the bloom-forming dinoflagellate Peridiniella (occupying on average 76% of the 16S reads) (figure 3a). The copepods Temora and Pseudocalanus were associated with fewer sequences of Peridiniella compared to the rotifer (on average 6% of 16S reads, d.f. = 3, z = 16, p < 0.001), but instead associated with various groups of small phytoplankton and picocyanobacteria. In March, Acartia was almost exclusively associated with filamentous cyanobacteria (figure 3a). The 18S sequences supported the association between S. baltica and Peridiniella but also revealed associations with the ciliate Myrionecta. The 18S sequences further revealed associations between all zooplankton species and diatoms (figure 2b). Figure 3. (a) Zooplankton consumer species (upper) with their most prevalent prey families (lower) based on 16S rRNA gene reads. The thickness of the bars is proportional to relative rRNA read abundance. (b) Non-metric multidimensional scaling plot of Bray–Curtis distances between zooplankton samples (represented by coloured points) based on their prey (16S rRNA reads). The prey families responsible for the largest percentage of dissimilarity between any pair of zooplankton species are represented as numbers. Synchaeta baltica reached its peak abundance in the Baltic Sea towards the end of the spring, in June, coordinated with the decline of dinoflagellates (Miozoa) (figure 1). Diet overlap between zooplankton species became more apparent but did not cluster according to phylogenetic affiliation. In June, S. baltica had an equally high diet overlap with the copepod Centropages (0.48) and the cladoceran Evadne (0.39), compared to the sister species S. monopus (0.40) (figure 3b; electronic supplementary material, appendix S1 and figure S3). Similarly, the copepod Acartia had a higher diet overlap with S. monopus (0.53) than with the other copepods (overlap of 0.11 with Temora) (electronic supplementary material, figure S5). At the end of spring, cyanobacteria became more apparent in the diet of the rotifers, indicating a transition from a spring to a summer prey community (figure 3a). (c) Trophic niche diversity in summer In August, the abundance and diversity of crustacean zooplankton increased in the Baltic Sea, and Keratella was the most abundant rotifer. The rotifer Synchaeta baltica was still present but with low abundance (figure 1a). The primary production was characterized by extensive blooms of filamentous cyanobacteria (figure 1b). In August, a large part of the variation in zooplankton diet read abundance could be explained by the abundance of filamentous cyanobacteria (Nostocaceae) and picocyanobacteria (Cyanobiaceae) (figure 3a). The highest diet overlap was found between the heterotrophic ciliate Helicostomella and the rotifer Keratella in August (0.75), as they were mostly associated with filamentous cyanobacteria (occupying 93% and 74% of 16S reads, respectively) (figure 3; electronic supplementary material, figure S5). Synchaeta baltica together with the cladoceran Bosmina and the copepod Acartia were associated with a lower proportion of filamentous cyanobacteria than the Keratella and Helicostomella (on average 39%, d.f. = 4, z = 5.7, p < 0.001), but with a larger proportion of picocyanobacteria (50%, d.f. = 5, z = 5.7, p < 0.001) as well as diverse small phytoplankton. Thus, the diet overlap between the two rotifer species Keratella and S. baltica was lower (0.42) than the overlap both between Keratella and the heterotrophic ciliate Helicostomella (0.75) and between S. baltica and the copepod Acartia (0.54). The copepods Temora and Centropages were associated with a low proportion of filamentous cyanobacteria (8%) and were almost exclusively associated with a higher relative proportion of picocyanobacteria (80%) compared with Acartia, Bosmina and S. baltica (d.f. = 5, z = 4.6, p < 0.001). Consequently, the copepod Acartia had a higher diet overlap with the cladoceran Bosmina (0.70) than with the other copepods (e.g. Temora, 0.52). Finally, Pseudocalanus, clustering alone, was associated with a significant proportion of unclassified organisms (up to 31% of 16S reads) (figure 3a). The 18S sequences revealed various groups of heterotrophic flagellates associated with S. baltica, Keratella and Helicostomella. Small phytoplankton (chlorophytes and eustigmatophytes), heterotrophic protozoans of different phyla, as well as metazoans dominated the 18S sequences of the cladocerans and copepods in summer (figure 2b). 4. Discussion (a) Niche diversity and overlap In order to resolve the trophic niche diversity of zooplankton, we analysed the physical associations of several micro and mesozooplankton species using 18S and 16S rRNA gene sequencing of selected zooplankton. The trophic position and role of zooplankton in food webs are often derived from taxonomic groupings and size estimation [12,48]. However, our results highlight that clustering zooplankton by size or phylogeny does not capture the true differences in diet niche and leads to an underestimation of the trophic niche diversity of primary consumers in the pelagic food web. Rotifers in the Baltic Sea are often grouped with microzooplankton and are referred to as obligate filter feeders [49,50]. Despite this, we can not find support for a high diet overlap or clustering between rotifer species (figure 3b). Instead, our study indicates that Synchaeta baltica (approx. 350 µm) has a trophic niche more similar to cladocerans and copepods than to the other rotifers, S. monopus, Keratella and the ciliate Helicostomella. The diet of S. baltica in spring included bloom-forming phytoplankton taxa, including the dinoflagellate Peridiniella (Peridiniaceae, 20–35 µm) and the mixotrophic ciliate Myrionecta (Cyclotrichia, 45–55 µm) (figure 3), findings that are in line with previous studies that have observed predation on large phytoplankton and protozoa up to 50 µm by Synchaeta [16–19]. In contrast with Synchaeta, the smaller sized rotifer Keratella peaks in abundance during the summer (figure 1) and was mainly associated with larger filamentous cyanobacteria (Nostocaceae). Keratella revealed a higher diet overlap with the tintinnid ciliate Helicostomella than with S. baltica (figure 3). The size of the Keratella (150 µm) and Helicostomella (100 µm) compared to cyanobacteria filaments that often exceed 1 mm suggests that these consumers do not feed directly on filamentous cyanobacteria, but rather on degraded filaments. This is supported by experiments indicating a filter-feeding behaviour of Keratella [15] that prefers partially degraded food (detritus) over living cells [51]. The same can be expected for Helicostomella, although few studies have investigated the selectivity of this ciliate [52–54]. The detritivorous feeding niche of Keratella and Helicostomella suggested here is further supported by a relatively high proportion of associated crustacean DNA (figure 2b), which for similar reasons is unlikely to be preyed upon directly and is likely ingested in the form of particulate organic matter. We suggest that filamentous cyanobacteria likely contribute to a pool of organic matter that is both available and attractive for detritivorous rotifers and ciliates. The results point to the importance of grouping zooplankton according to function rather than taxa, with some rotifers occupying the function of selective feeders similar to copepods, a distinction already proposed in a study by Arndt in 1993 [15]. Similar to rotifers, there was no clear clustering within copepod and cladoceran species based on their prey composition. While Temora and Centropages shared high diet overlap, mostly associated with picocyanobacteria, Acartia had a diet more similar to the cladocerans and the rotifer Synchaeta baltica during the summer months, relying on a large diversity of resources, including both filamentous and picocyanobacteria (figure 3). On the other side, the copepod Pseudocalanus occupies its own niche, feeding in all seasons on various unclassified organisms. The different feeding niches of copepods may reflect their vertical distribution (electronic supplementary material, figure S2), as Temora and Centropages are more abundant at 30–60 m depth compared to Acartia that dominates in the upper 30 m. By contrast, Pseudocalanus extends deeper in the water column compared to the other zooplankton [55,56]. The association of Pseudocalanus with unidentified taxa might indicate that the components of its natural diet are not well represented in the taxonomic databases. The differences in trophic niches of copepods are supported by a more intense seasonal sample analysis of crustacean zooplankton described in the previous study [29] and more extensive sampling several locations in the Baltic Sea proper (Baptiste Serandour 2021, pers. comm). (b) Food web implications The trophic niche partitioning of zooplankton has implications in the food web dynamics in spring and summer. As the decline of dinoflagellates in spring coincides with the peak of Synchaeta (figure 1), grazing by rotifers should be considered a potential cause of the phytoplankton spring bloom decline in addition to nutrient limitation in the upper water column [57]. This is further supported by a study from the Mediterranean Sea, where Synchaeta was estimated to consume up to 80% of the daily production of a dinoflagellate bloom [58]. Given that cladocerans and copepods are temporally decoupled from the spring bloom (figure 1), Synchaeta is likely a major pathway of energy transfer to higher trophic levels in the Baltic Sea and possibly in other ecosystems where this species is abundant [58]. Pico and filamentous cyanobacteria are favoured under climate warming and eutrophication and are increasing in both marine and freshwater systems [30,59]. Due to their significance as food resources in the summer community of the Baltic Sea [33,49,60], understanding the pathways of cyanobacteria incorporation into food webs is important. The trophic role of filamentous cyanobacteria is widely debated as they are the main source of nitrogen fixation in the Baltic Sea, which is suggested to be nitrogen limited in summer [61]. Several studies suggest little or no grazing on filamentous cyanobacteria by zooplankton [34,62,63], but our study shows that the microzooplankton Helicostomella and Keratella, but also to some extent Bosmina and Acartia grazed on filamentous cyanobacteria, either living or degraded. Zooplankton feeding on filamentous cyanobacteria may also act as important vectors of diazotroph nitrogen availability to upper trophic levels by stimulating the microbial food web. This is supported by an experiment by Arndt [15] showing that the presence of filter-feeding Keratella stimulates the growth of both heterotrophic flagellates and bacteria. Arndt proposed that Keratella, through its feeding, enhances leaking of dissolved matter from the algae [4], thereby supporting the increased biomass of both bacteria and protozoa. Tracer studies suggest that diazotroph nitrogen is mainly incorporated in the Baltic Sea food web by passive leaking of nitrogen compounds by filamentous cyanobacteria [64,65] that stimulate the production of heterotrophic bacteria and picocyanobacteria [33,34,66]. Our study confirms that picocyanobacteria are a key resource for the larger zooplankton species, primarily Temora and Centropages, but also, to some extent, Acartia, Bosmina and Synchaeta (figure 4a), and are as such indirectly supported by filamentous cyanobacteria. In addition to passive leaking, feeding by copepods, cladocerans, rotifers and ciliates on filamentous cyanobacteria actively enhances leaking of diazotroph nitrogen. This constitutes an alternative pathway of cyanobacteria incorporation into the pelagic food web and enhances the support for copepods that rely on the microbial food web (figure 4b). Figure 4. Illustration of dominant zooplankton consumers and alternative energy transfer pathways in the pelagic food web. (a) Size range of selected zooplankton species aligned with the relative read abundance of associated pico- and filamentous cyanobacteria. (b) Black arrows illustrate alternative pathways of energy transfer from primary producers to zooplankton consumers. Ingestion of picocyanobacteria can be directly (1) or via primary consumers (2). Filamentous cyanobacteria can be ingested alive (3) or in the state of decay (4). Shaded arrows denote putative pathways of nitrogen fixed by filamentous cyanobacteria (5) that may enter the food web either via consumption (3, 4) but may also stimulate the production of picocyanobacteria (6). Grazers of filamentous cyanobacteria may enhance the release of dissolved inorganic nitrogen (7), in contrast with previously suggested passive leaking (8). (Online version in colour.) While DNA metabarcoding has become more frequently used over the last decade, few studies have until now exploited the potential of investigating the feeding niche diversity of the entire zooplankton community spanning several phyla and size classes. We show that the method used here is applicable for several metazoan taxa and potentially protozooplankton, using the ciliate Helicostomella as an example. DNA metabarcoding could be relevant in future investigations to unveil the role of rarer species and better comprehend the ecosystem function. Our approach shows the advantage of investigating the prey composition of diverse species in natural systems with DNA metabarcoding that reveals the entire food spectrum. By putting weight on the relative comparison between zooplankton species, we could capture key differences in zooplankton resource. While we can discuss possible ecosystem effects of diverse zooplankton feeding, the metabarcoding data is inevitably proportional. Thus, the data do not reveal information about feeding rates or biomasses. Despite this, metabarcoding has the potential to serve as an important complement to food web models that implement population biomasses and metabolic energy demands [67], by bringing details of species-specific feeding interactions to the model. Our results highlight a large variation in resource use between groups of zooplankton that may stabilize energy transfer in food webs by pathways of energy flow that are rarely described, particularly during seasons when primary producers include pico and filamentous cyanobacteria. The presence of multitrophic species with the ability to prey on different food web components may contribute to ecosystem resilience. We emphasize the importance of understanding the trophic niche diversity of key zooplankton taxa to generate an accurate understanding of ecosystem functioning. Food web models based on size or phylogeny may not capture the important role of individual species and may not be detailed enough to predict energy pathways of plankton food webs and thus the vulnerability of ecosystems to environmental change. Combined with estimates of prey biomass and predator feeding rates, the approach used in this study is a suitable entry point to food web modelling and ecosystem network analysis. Supplementary Material Click here for additional data file. Acknowledgements We would like to direct our gratitude to the pelagic monitoring group at Stockholm University, especially Stefan Svensson and Jakob Walve for hosting us during the sampling effort. We also acknowledge support from the National Genomics infrastructure in Stockholm and Uppsala, funded by Science for Life Laboratory, the Knut and Alice Wallenberg Foundation and the Swedish Research Council, and SNIC/Uppsala Multidisciplinary Centre for Advanced Computational Science for assistance with massively parallel sequencing and access to the UPPMAX computational infrastructure. Data accessibility DNA sequences and associated metadata were uploaded to the ENA under accession no. PRJEB39191 (https://www.ebi.ac.uk/ena/browser/view/PRJEB39191). RSV data files used for statistical analysis, together with the R code to generate the figures are available from the Dryad Digital Repository: https://doi.org/10.5061/dryad.gb5mkkwpw [47]. Authors' contributions A.N.: conceptualization, data curation, formal analysis, investigation, methodology, project administration, software, validation, visualization, writing-original draft, writing-review and editing; S.Z.-T: conceptualization, data curation, investigation, methodology, project administration, validation, writing-review and editing; M.W.: conceptualization, funding acquisition, project administration, resources, supervision, writing-review and editing. All authors gave final approval for publication and agreed to be held accountable for the work performed therein. Competing interests We have no competing interests. Funding The research was funded by the Swedish Research Council project no. 2016-04685. ==== Refs References 1. Cadotte MW, Carscadden K, Mirotchnick N, Cadotte MW, Carscadden K, Mirotchnick N. 2011 Beyond species: functional diversity and the maintenance of ecological processes and services. J. Appl. Ecol. 48 , 1079-1087. (10.1111/j.1365-2664.2011.02048.x) 2. Naeem S, Duffy JE, Zavaleta E. 2012 The functions of biological diversity in an age of extinction. Science 336 , 1401-1406. (10.1126/science.1215855)22700920 3. Calbet A, Landry MR. 2004 Phytoplankton growth, microzooplankton grazing, and carbon cycling in marine systems. Limnol. Oceanogr. 49 , 51-57. (10.4319/lo.2004.49.1.0051) 4. Steinberg DK, Landry MR. 2017 Zooplankton and the ocean carbon cycle. Ann. Rev. Mar. Sci. 9 , 413-444. (10.1146/annurev-marine-010814-015924) 5. Mitra A, Davis C. 2010 Defining the ‘to’ in end-to-end models. Prog. Oceanogr. 84 , 39-42. (10.1016/j.pocean.2009.09.004) 6. Cushing DH. 1990 Plankton production and year-class strength in fish populations: an update of the match/mismatch hypothesis. Adv. Mar. Biol. 26 , 249-293. (10.1016/S0065-2881(08)60202-3) 7. Sherr EB, Sherr BF. 2002 Significance of predation by protists in aquatic microbial food webs. Antonie Van Leeuwenhoek 81 , 293-308.12448728 8. Bindoff NL et al . 2019 Changing ocean, marine ecosystems, and dependent communities. In Special report on the ocean and cryosphere in a changing climate (eds H-O Pörtner et al .), pp. 447-588. New York, NY: IPCC. 9. Mitra A et al. 2014 Bridging the gap between marine biogeochemical and fisheries sciences; configuring the zooplankton link. Prog. Oceanogr. 129 , 176-199. (10.1016/j.pocean.2014.04.025) 10. Heneghan RF, Everett JD, Blanchard JL, Richardson AJ. 2016 Zooplankton are not fish: improving zooplankton realism in size-spectrum models mediates energy transfer in food webs. Front. Mar. Sci. 3 , 1-15. (10.3389/fmars.2016.00201) 11. Petchey OL, Gaston KJ. 2006 Functional diversity: back to basics and looking forward. Ecol. Lett. 9 , 741-758. (10.1111/j.1461-0248.2006.00924.x)16706917 12. Boyce DG, Frank KT, Leggett WC. 2015 From mice to elephants: overturning the ‘one size fits all’ paradigm in marine plankton food chains. Ecol. Lett. 18 , 504-515. (10.1111/ele.12434)25919397 13. Sommer U, Stibor H. 2002 Copepoda – Cladocera – Tunicata: the role of three major mesozooplankton groups in pelagic food webs. Ecol. Res. 17 , 161-174. (10.1046/j.1440-1703.2002.00476.x) 14. Stibor H et al. 2004 Copepods act as a switch between alternative trophic cascades in marine pelagic food webs. Ecol. Lett. 7 , 321-328. (10.1111/j.1461-0248.2004.00580.x) 15. Arndt H. 1993 Rotifers as predators on components of the microbial web (bacteria, heterotrophic flagellates, ciliates) – a review. In Hydrobiologia, pp. 231-246. Dordrecht, The Netherlands: Springer Netherlands. (10.1007/BF00025844) 16. Pourriot R. 1977 Food and feeding habits of Rotifera. Hydrobiol. Beih. Ergebn. Limnol. 8 , 243-260. 17. Bogdan KG, Gilbert JJ, Starkweather PL. 1980 In situ clearance rates of planktonic rotifers. In Hydrobiologica, pp. 73-77. Dordrecht, The Netherlands: Springer Netherlands. (10.1007/978-94-009-9209-2_14) 18. Bogdan KG, Gilbert JJ. 1982 Seasonal patterns of feeding by natural populations of Keratella, Polyarthra, and Bosmina: clearance rates, selectivities, and contributions to community grazing1. Limnol. Oceanogr. 27 , 918-934. (10.4319/lo.1982.27.5.0918) 19. Gilbert JJ, Jack JD. 1993 Rotifers as predators on small ciliates. In Rotifer symposium VI, pp. 247-253. Dordrecht, The Netherlands: Springer Netherlands. (10.1007/978-94-011-1606-0_32) 20. Gilbert JJ. 1980 Observations on the susceptibility of some protists and rotifers to predation by asplanchna girodi. Hydrobiologia 73 , 87-91. (10.1007/BF00019431) 21. Kiørboe T. 2011 How zooplankton feed: mechanisms, traits and trade-offs. Biol. Rev. 86 , 311-339. (10.1111/j.1469-185X.2010.00148.x)20682007 22. Post DM. 2002 Using stable isotopes to estimate trophic position: models, methods, and assumptions. Ecology 83 , 703. (10.2307/3071875) 23. Gutiérrez-Rodríguez A, Décima M, Popp BN, Landry MR. 2014 Isotopic invisibility of protozoan trophic steps in marine food webs. Limnol. Oceanogr. 59 , 1590-1598. (10.4319/lo.2014.59.5.1590) 24. Pompanon F, Deagle BE, Symondson WOCC, Brown DS, Jarman SN, Taberlet P. 2012 Who is eating what: diet assessment using next generation sequencing. Mol. Ecol. 21 , 1931-1950. (10.1111/j.1365-294X.2011.05403.x)22171763 25. Roslin T, Majaneva S 2016 The use of DNA barcodes in food web construction—terrestrial and aquatic ecologists unite!. Genome 59 , 603-628. (10.1139/gen-2015-0229)27484156 26. Ray JL et al. 2016 Metabarcoding and metabolome analyses of copepod grazing reveal feeding preference and linkage to metabolite classes in dynamic microbial plankton communities. Mol. Ecol. 25 , 5585-5602. (10.1111/mec.13844)27662431 27. Yi X, Huang Y, Zhuang Y, Chen H, Yang F, Wang W, Xu D, Liu G, Zhang H. 2017 In situ diet of the copepod Calanus sinicus in coastal waters of the South Yellow Sea and the Bohai Sea. Acta Oceanol. Sin. 36 , 68-79. (10.1007/s13131-017-0974-6) 28. Craig C, Kimmerer WJ, Cohen CS. 2014 A DNA-based method for investigating feeding by copepod nauplii. J. Plankton Res. 36 , 271-275. (10.1093/plankt/fbt104) 29. Zamora-Terol S, Novotny A, Winder M. 2020 Reconstructing marine plankton food web interactions using DNA metabarcoding. Mol. Ecol. 29 , 3380-3395. (10.1111/mec.15555)32681684 30. Paerl HW, Huisman J. 2008 Blooms like it hot. Am. Assoc. Adv. Sci. 320 , 57-58. (10.1210/jcem-10-10-1361) 31. Cloern JE et al. 2016 Human activities and climate variability drive fast-paced change across the world's estuarine-coastal ecosystems. Glob. Chang. Biol. 22 , 513-529. (10.1111/gcb.13059)26242490 32. Schmidt K et al. 2020 Increasing picocyanobacteria success in shelf waters contributes to long-term food web degradation. Glob. Chang. Biol. 26 , 5574-5587. (10.1111/gcb.15161)32506810 33. Karlson AML et al. 2015 Nitrogen fixation by cyanobacteria stimulates production in Baltic food webs. Ambio 44 , 413-426. (10.1007/s13280-015-0660-x)26022324 34. Loick-Wilde N, Fernández-Urruzola I, Eglite E, Liskow I, Nausch M, Schulz-Bull D, Wodarg D, Wasmund N, Mohrholz V. 2019 Stratification, nitrogen fixation, and cyanobacterial bloom stage regulate the planktonic food web structure. Glob. Chang. Biol. 25 , 794-810. (10.1111/gcb.14546)30628151 35. Engstrom J. 2000 Feeding interactions of the copepods Eurytemora affinis and Acartia bifilosa with the cyanobacteria Nodularia sp. J. Plankton Res. 22 , 1403-1409. (10.1093/plankt/22.7.1403) 36. Naturvårdsverket. 2009 Beskrivning av delprogram Fria vattenmassan. See https://www.havochvatten.se/download/18.64f5b3211343cffddb280005451/1348912810782/beskrivning-delprogram-fria-vattenmassan.pdf (accessed on 3 April 2018). 37. Herlemann DPR, Labrenz M, Jürgens K, Bertilsson S, Waniek JJ, Andersson AF. 2011 Transitions in bacterial communities along the 2000 km salinity gradient of the Baltic Sea. ISME J. 5 , 1571-1579. (10.1038/ismej.2011.41)21472016 38. Hu YOO, Karlson B, Charvet S, Andersson AF. 2016 Diversity of pico- to mesoplankton along the 2000 km salinity gradient of the Baltic Sea. Front. Microbiol. 7 , 679. (10.3389/fmicb.2016.00679)27242706 39. Ho TW, Hwang J-S, Cheung MK, Kwan HS, Wong CK. 2017 DNA-based study of the diet of the marine calanoid copepod Calanus sinicus. J. Exp. Mar. Bio. Ecol. 494 , 1-9. (10.1016/j.jembe.2017.04.004) 40. Pruesse E, Quast C, Knittel K, Fuchs BM, Ludwig W, Peplies J, Glockner FO. 2007 SILVA: a comprehensive online resource for quality checked and aligned ribosomal RNA sequence data compatible with ARB. Nucleic Acids Res. 35 , 7188-7196. (10.1093/nar/gkm864)17947321 41. Decelle J et al. 2015 PhytoREF: a reference database of the plastidial 16S rRNA gene of photosynthetic eukaryotes with curated taxonomy. Mol. Ecol. Resour. 15 , 1435-1445. (10.1111/1755-0998.12401)25740460 42. Guillou L et al. 2013 The Protist Ribosomal Reference database (PR2): a catalog of unicellular eukaryote small sub-unit rRNA sequences with curated taxonomy. Nucleic Acids Res. 41 , D597-D604. (10.1093/nar/gks1160)23193267 43. McMurdie PJ, Holmes S. 2013 phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PLoS ONE 8 , e61217. (10.1371/journal.pone.0061217)23630581 44. Oksanen J, Kindt R, Legendre P, O'Hara B, Simpson GL, Solymos PM, Stevens MHH, Wagner H. 2007 The vegan package. Commun. Ecol. Packag. 10 , 613-637. 45. Wickham H. 2016 Ggplot2: elegant graphics for data analysis. Berlin, Germany: Springer. 46. Gu Z, Gu L, Eils R, Schlesner M, Brors B. 2014 circlize implements and enhances circular visualization in R. Bioinformatics 30 , 2811-2812. (10.1093/bioinformatics/btu393)24930139 47. Novotny A, Zamora-Terol S, Winder M. 2021 Data from: DNA metabarcoding reveals trophic niche diversity of micro and mesozooplankton species. Dryad Digital Repository. (10.5061/dryad.gb5mkkwpw) 48. Sommer F, Stibor H, Sommer U, Velimirov B. 2000 Grazing by mesozooplankton from Kiel Bight, Baltic Sea, on different sized algae and natural seston size fractions. Mar. Ecol. Prog. Ser. 199 , 43-53. (10.3354/meps199043) 49. Motwani NH, Gorokhova E. 2013 Mesozooplankton grazing on picocyanobacteria in the baltic sea as inferred from molecular diet analysis. PLoS ONE 8 , e79230. (10.1371/journal.pone.0079230)24260175 50. Grinienė E, Šulčius S, Kuosa H. 2016 Size-selective microzooplankton grazing on the phytoplankton in the Curonian Lagoon (SE Baltic Sea). Oceanologia 58 , 292-301. (10.1016/j.oceano.2016.05.002) 51. Starkweather PL, Bogdan KG. 1980 Detrital feeding in natural zooplankton communities: discrimination between live and dead algal foods. Hydrobiologia 73 , 83-85. (10.1007/BF00019430) 52. Mironova E, Telesh I, Skarlato S. 2012 Diversity and seasonality in structure of ciliate communities in the Neva Estuary (Baltic Sea). J. Plankton Res. 34 , 208-220. (10.1093/plankt/fbr095) 53. Aberle N, Lengfellner K, Sommer U. 2007 Spring bloom succession, grazing impact and herbivore selectivity of ciliate communities in response to winter warming. Oecologia 150 , 668-681. (10.1007/s00442-006-0540-y)16964503 54. Capriulo GM. 1982 Feeding of field collected tintinnid micro-zooplankton on natural food. Mar. Biol. 71 , 73-86. (10.1007/BF00396994) 55. Renz J, Hirche HJ. 2006 Life cycle of Pseudocalanus acuspes Giesbrecht (Copepoda, Calanoida) in the Central Baltic Sea: I. Seasonal and spatial distribution. Mar. Biol. 148 , 567-580. (10.1007/s00227-005-0103-5) 56. Schulz J et al. 2012 Spatial and temporal habitat partitioning by zooplankton in the Bornholm Basin (central Baltic Sea). Prog. Oceanogr. 107 , 3-30. (10.1016/j.pocean.2012.07.002) 57. Tamminen T, Andersen T. 2007 Seasonal phytoplankton nutrient limitation patterns as revealed by bioassays over Baltic Sea gradients of salinity and eutrophication. Mar. Ecol. Prog. Ser. 340 , 121-138. (10.3354/meps340121) 58. Calbet A, Vaqué D, Felipe J, Vila M, Alcaraz M, Estrada M. 2003 Relative grazing impact of microzooplankton and mesozooplankton. Mar. Ecol. Prog. Ser. 259 , 303-309. (10.3354/meps259303) 59. Winder M, Sommer U. 2012 Phytoplankton response to a changing climate. Hydrobiologia 698 , 5-16. (10.1007/s10750-012-1149-2) 60. Motwani NH, Duberg J, Svedén JB, Gorokhova E. 2018 Grazing on cyanobacteria and transfer of diazotrophic nitrogen to zooplankton in the Baltic Sea. Limnol. Oceanogr. 63 , 672-686. (10.1002/lno.10659) 61. Granéli E, Wallström K, Larsson U, Granéli W, Elmgren R. 1990 Nutrient limitation of primary production in the Baltic Sea area. Ambio 19 , 142-151. 62. Loick-Wilde N, Dutz J, Miltner A, Gehre M, Montoya JP, Voss M. 2012 Incorporation of nitrogen from N2 fixation into amino acids of zooplankton. Limnol. Oceanogr. 57 , 199-210. (10.4319/lo.2012.57.1.0199) 63. Wannicke N, Korth F, Liskow I, Voss M. 2013 Incorporation of diazotrophic fixed N2 by mesozooplankton: case studies in the southern Baltic Sea. J. Mar. Syst. 117–118 , 1-13. (10.1016/j.jmarsys.2013.03.005) 64. Ploug H, Musat N, Adam B, Moraru CL, Lavik G, Vagner T, Bergman B, Kuypers MMM. 2010 Carbon and nitrogen fluxes associated with the cyanobacterium Aphanizomenon sp. in the Baltic Sea. ISME J. 4 , 1215-1223. (10.1038/ismej.2010.53)20428225 65. Ploug H, Adam B, Musat N, Kalvelage T, Lavik G, Wolf-Gladrow D, Kuypers MMM. 2011 Carbon, nitrogen and O 2 fluxes associated with the cyanobacterium Nodularia spumigena in the Baltic Sea. ISME J. 5 , 1549-1558. (10.1038/ismej.2011.20)21390075 66. Eglite E, Graeve M, Dutz J, Wodarg D, Liskow I, Schulz-Bull D, Loick-Wilde N. 2019 Metabolism and foraging strategies of mid-latitude mesozooplankton during cyanobacterial blooms as revealed by fatty acids, amino acids, and their stable carbon isotopes. Ecol. Evol. 9 , 9916-9934. (10.1002/ece3.5533)31534704 67. Kortsch S et al. 2021 Disentangling temporal food web dynamics facilitates understanding of ecosystem functioning. J. Anim. Ecol. 90 , 1-12. (10.1111/1365-2656.13447)
PMC008xxxxxx/PMC8206688.txt
==== Front Proc Biol Sci Proc Biol Sci RSPB royprsb Proceedings of the Royal Society B: Biological Sciences 0962-8452 1471-2954 The Royal Society 34130502 10.1098/rspb.2021.0823 rspb20210823 1001145870Evolution Research Articles Climate variability and parent nesting strategies influence gas exchange across avian eggshells Climate variability and parent nesting strategies influence gas exchange across avian eggshells http://orcid.org/0000-0002-8509-3677 Attard Marie R. G. Conceptualization Data curation Formal analysis Investigation Methodology Project administration Validation Visualization Writing-original draft Writing-review & editing [email protected] 1 2 http://orcid.org/0000-0002-2438-2352 Portugal Steven J. Conceptualization Formal analysis Funding acquisition Investigation Methodology Project administration Resources Supervision Writing-review & editing 1 3 1 Department of Biological Sciences, School of Life and Environmental Sciences, Royal Holloway University of London, Egham, Surrey TW20 0EX, UK 2 School of Engineering and Innovation, Open University, Milton Keynes MK7 6AA, UK 3 The Natural History Museum, Tring, Herts HP23 6AP, UK Electronic supplementary material is available online at https://doi.org/10.6084/m9.figshare.c.5448697. 30 6 2021 June 30, 2021 16 6 2021 June 16, 2021 16 6 2021 June 16, 2021 288 1953 202108237 4 2021 April 7, 2021 14 5 2021 May 14, 2021 © 2021 The Authors. 2021 https://creativecommons.org/licenses/by/4.0/ Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited. Embryo survival in birds depends on a controlled transfer of water vapour and respiratory gases through the eggshell, and this exchange is critically sensitive to the surrounding physical environment. As birds breed in most terrestrial habitats worldwide, we proposed that variation in eggshell conductance has evolved to optimize embryonic development under different breeding conditions. This is the first study to take a broad-scale macro-ecological view of avian eggshell conductance, encompassing all key avian taxonomic groups, to assess how life history and climate influence the evolution of this trait. Using whole eggs spanning a wide phylogenetic diversity of birds, we determine that body mass, temperature seasonality and whether both parents attend the nest are the main determinants of eggshell conductance. Birds breeding at high latitudes, where seasonal temperature fluctuations are greatest, will benefit from lower eggshell conductance to combat temporary periods of suspended embryo growth and prevent dehydration during prolonged incubation. The nest microclimate is more consistent in species where parents take turns incubating their clutch, resulting in lower eggshell conductance. This study highlights the remarkable functional qualities of eggshells and their importance for embryo survival in extreme climates. avian eggshells , climate , life history , nest , temperature seasonality , water vapour conductance Leverhulme Trust http://dx.doi.org/10.13039/501100000275 RPG-2018-332 cover-dateJune 30, 2021 ==== Body pmc1. Introduction Adaptive diversification across species typically occurs amidst an array of distinct ecological niches and environments and is a key driver in the development of novel functional traits to enhance the fitness of an organism [1]. The evolution of a new trait may provide the adaptive potential to exploit a resource that was not previously possible, or interact with its environment in a new way without a specific change in the external environment [2]. Close asso­ciation between certain traits and a species environment and life history can therefore point to probable causes of trait divergence [3]. Traits can evolve rapidly over several generations or slowly over millions of years in accordance with environmental rates of change [4]. Species persistence during abrupt climate change will, therefore, depend on their ability to rapidly respond and adapt to novel environmental conditions [5]. Individual species will either move to more favourable conditions, tolerate or adapt to their changed environment, or go extinct [6]. Understanding the evolutionary history and diversification of functional traits closely linked to reproductive success will help predict how species will react to these new environmental pressures. Foremost, the survival of any species is reliant on having a viable embryo. One crucial step in understanding avian responses to environmental differences over evolutionary time is a better appreciation of factors shaping avian incubation and their subsequent influence on the embryo [7]. Birds have evolved multiple functional traits to improve offspring survival in the nest: arguably one of the most important is the eggshell. Most bird species lose 10–20% of their fresh egg mass over the incubation period through the passive diffusion of water vapour through the eggshell to the ambient air [8]. Eggs that lose too much water during incubation frequently do not hatch due to desiccation [9], while embryos that do not lose enough water from the egg experience respiratory problems or drown [10]. Maintaining a controlled loss of water from inside the egg to the external environment while allowing sufficient exchange of respiratory gases is therefore essential for normal embryo development and hatching. Birds are highly diversified and widely distributed, occupying every continent on Earth and every terrestrial habitat within it [11]. Some birds breed in extremely inhospitable environments, such as cold and dry regions [12], deserts [13], moist wetlands [14] and high altitudes [15]. Among these are ground-nesting birds in alpine or Arctic/Antarctic regions that must cope with unpredictable wind, precipitation and snow conditions, with ambient temperatures fluctuating from below freezing to over 45°C [16]. Avian embryos in such cold regions will freeze to death if left unattended by their parents [17]. Desert birds that breed in the Sahara, Arabian and Kalahari regions face extreme physiological challenges to conserve water and avoid dehydration for the eggs, adults and hatchlings [18]. In contrast with dry, xeric environments, eggs exposed to high precipitation are prone to rain-induced suffocation [19]. The major challenges for birds breeding in high-altitude regions like the Himalayas is the low barometric pressure and high solar radiation, which can result in desiccation of egg contents and overheating of the embryo [15]. Species living at such environmental extremes must adapt behaviourally or physiologically at each stage of their breeding cycle if they are to produce viable offspring [20]. Water vapour conductance through the avian eggshell, herein referred to as conductance or GH2O, is influenced by the properties of the eggshell (e.g. pore length, functional pore area and eggshell cuticle) and humidity and gas composition of the surrounding environment. Species that incubate their eggs buried [21], in dry [22] or wet environments [23], or at high altitudes [15] have particularly unusual vapour pressure gradients, yet are still able to maintain water loss within acceptable limits. GH2O may be optimized to suit particular environments through changes in nest-site preferences, eggshell structures and incubation behaviours [14], making eggs and their species-specific conductance ideal model systems for understanding how trait selection varies over time during diversification. Predicting GH2O of a species is not straight forward, as multiple ecological factors must be taken into account. For example, brood-chambers of burrow-nesting birds are often permanently saturated with water vapour, resulting in a low water vapour pressure (favouring enhanced conductance) and longer incubation periods (favouring reduced conductance) [24]. Inter-species differences in GH2O thus can only be untangled by considering the contribution of multiple life-history traits and the phylogenetic history of the lineage. A study across 141 non-passerine species detected differences in GH2O between nest types and parental incubation behaviours [25], emphasizing the importance of maintaining a suitable nest microclimate for optimum egg-water loss. However, it is unknown whether a similar relationship between conductance and nesting behaviour is expected in the passerines, which comprise over 6000 species and represent almost 60% of all living birds [26]. Moreover, previous studies have typically focused on either (i) one group of birds (e.g. gulls), with the goal to look for micro-adaptations between closely related species [27], or (ii) eggs of ‘extreme nesters’ such as desert-nesting Bedouin fowl (Gallus domesticus) [28] and grey gulls (Larus modestus) [29], water-nesting grebes and divers [30] and marsh-nesting black terns (Chlidonias niger) [31]. The role of life-history and environmental factors in the evolution of avian eggshell conductance thus requires a large-scale comparative analysis encompassing all key taxonomic groups. Our aim was to evaluate how climate and life history influence GH2O across a wide taxonomic distribution of birds spanning 28 avian orders, after accounting for the effects of adult body mass and phylogeny. Previous comparative analyses of eggshell conductance have not corrected for allometric effects of body mass [25], which can hide potentially important adaptive information relating to the environment and nesting behaviour of the species. Based on previous findings, we predicted GH2O would be primarily explained by body mass. By contrast, we predict that mass-independent conductance (RGH2O) would be primarily associated with traits known to affect nest humidity, including climate, nest location and type. 2. Material and methods (a) Egg samples and preparation In total, 365 bird species were included in this study. Conductance of whole emptied eggs at the Natural History Museum, Tring (NHM, UK) was established using the standard protocol of measuring the decrease in egg mass as a result of water loss over consecutive days, in eggs kept in constant moisture-free conditions [32]. GH2O measured using whole eggshells is preferable to eggshell fragments as shell thickness and porosity varies between different regions of an egg [33]; therefore, we only used values from whole eggs in this study. Eggs were prepared by gently cleaning the surface, filling the egg with water then sealing the blow hole (see electronic supplementary material). Eggs were placed in an acrylic desiccator cabinet (ThermoFisher Scientific, Nalgene, catalogue number: 5317-0070) inside a constant temperature thermocabinet (Porkka, Hertfordshire, UK) at 30 ± 1°C. Temperature was monitored via a logtag analyser every 10 min (Loggershop, Bournemouth, Dorset, UK). Self-indicating silica gel (Merck, Honenbrunn, Germany, catalogue number: 101 969) were placed in the desiccator to remove all moisture. Any loss in egg mass was entirely due to the diffusion of water vapour via the shell pores [34]. The first 24 h can give unexpectedly high mass loss values as the outer shell surface dries out [35]. Therefore, the eggs were left 24 h before being weighed to the nearest 0.1 mg (Sartorius, Göttingen, Germany), then were returned to the desiccator. Eggs were weighed at the same time of day on 3 successive days to give two values of 24 h mass loss (MH2O). Species GH2O was then calculated, as described in the electronic supplementary material. Species mean GH2O values of whole eggs reported in the literature (n = 188) were incorporated if specimens had been measured under constant conditions (temperature and humidity) and followed protocols used in the present study. GH2O measures from whole fresh eggs (unemptied or water-filled) and museum (water-filled) eggs were combined as GH2O does not differ significantly between these treatments [36]. Mean GH2O values reported in the literature were corrected to standard barometric pressure (1 ATM) at 30°C (see electronic supplementary material). (b) Life-history and ecological data We collated data on 18 key life-history traits that have previously been hypothesized to play a role in the evolution of avian conductance in addition to climate variables (table 1). These data were extracted from multiple sources detailed in the FigShare repository (doi:10.6084/m9.figshare.12490559). Major sources are detailed in section (e) of electronic supplementary material. Only 13 predictors were included in the analysis due to collinearity (see electronic supplementary material). The phylogenetic generalized least-squares (PGLS) method was used to test the evolutionary association between whole eggshell GH2O life-history traits, within a phylogenetic context [37]. In this procedure, closely related species are assumed to have more similar traits because of their shared ancestry and consequently will produce more similar residuals from the least-squares regression line. By taking into account the expected covariance structure of these residuals, modified slope and intercept estimates are generated that account for interspecific autocorrelation due to phylogeny. Table 1. Putative predictions and definitions for 13 possible explanations for variation in water vapour conductance (GH2O) in birds. predictor hypothesis definition body mass as adult body mass is correlated to egg mass, heavier birds will have higher GH2O due to greater egg surface area mean body mass (g) of adult birds clutch size evaporation from multiple eggs will create a nest atmosphere of greater humidity and reduced water vapour transfer, so GH2O should be higher for species with larger clutches number of eggs per brood, measured as geometric mean of the typical minimum and maximum clutch size calcium content eggshells of calcium-poor species are expected to be thinner, less dense and more porous than calcium-rich species, and thus facilitate higher GH2O (1) calcium-rich: species that ingest mollusc shells, fish, shellfish, calcareous grit, calcareous ash or bones (2) calcium-poor: species with primarily insectivorous or granivorous diet egg maculation maculated eggs are expected to have lower GH2O than immaculate eggs to reduce the risk of desiccation (1) immaculate: no spotting or markings on eggshell surface (2) maculation: maculation present on eggshell surface nest type fully enclosed nests have less air movement than semi-enclosed and exposed nests, facilitating greater GH2O (1) exposed: nest is open above and has no side walls (no nest, scrape, saucer, platform, heap) (2) semi-enclosed: nest is partially open and has side walls (cup, bowl, pendant, sphere, dome, pouch) (3) enclosed: nest is entirely enclosed (cavity, burrow, crevice) nest location nests above ground have lower risk of flooding or water accumulation, therefore will have lower GH2O (1) ground: nest location in or on the ground, or floating on water (2) tree: nest located in tree, bush, shrub, wall, cave roof, or attached to reed (3) cliff: nest located on cliff nest lining incorporation of nest lining will better insulate the egg, therefore will have higher GH2O (1) lined: nest lining is always or sometimes present (2) not lined: nest lining is absent habitat among open nesting species, more direct sunlight reaches eggs in open habitats and experience greater air movement around the nest than closed habitats; open-nesting species in open habitats will have lower GH2O than in closed habitats (1) open: breeds in desert, grassland, open water, open moorland, low shrubs, rocky habitats, seashores and cities (2) semi-open: breeds in open shrubland and bushland, scattered bushes, parkland, forest edge (3) dense: breeds in forest with a closed canopy, or in the lower vegetation strata of dense thickets, shrubland, mangroves or marshland incubating parent nest vapour pressure will decrease when the parent leaves the nest uncovered, which is more likely to occur if incubation is not shared between parents, resulting in lower GH2O (1) shared: contact incubation of eggs by two adults (2) not shared: contact incubation of eggs by single adult mode of development higher GH2O may contribute to improving the use of nutritional support by the embryo of precocial species by removing excess water, thus resulting in increased development at hatching (1) altricial: newly born young are relatively immobile, naked, and usually require care and feeding by the parents (2) precocial: newly born young are relatively mobile, covered in feathers, and independent parental care the eggs of species that provide biparental care are expected to have higher GH2O as nest humidity and temperature can be better maintained when both parents assist (1) uniparental: the brood is provisioned and/or defended by one adult (2) biparental: the brood is provisioned and/or defended by two adults parental contact the wet incubating parent returning to the nest will increase the nest's humidity, thus are excepted to have higher GH2O (1) wet plumage: adults returned habitually to the nest with wet plumage; this included species that feed on freshwater or marine prey or use nests built on water (2) dry plumage: adults did not return habitually to the nest with wet plumage temperature seasonality eggs incubated in environments with highly variable temperature will experience lower GH2O as high temperature seasonality occurs in cooler environments average temperature seasonality (BIO4) of breeding/resident range, based on WorldClim v1 data precipitation seasonality eggs incubated in environments with highly variable precipitation will experience higher GH2O to combat temporary periods of excessive rain average precipitation seasonality (BIO15) of breeding/resident range, based on WorldClim v1 data Prior to updated avian phylogenies based on genomic DNA, near-passerines was a term given to tree-dwelling birds (within the conventional non-passerines) that were traditionally believed to be related to Passeriformes due to ecological similarities. In this study Pterocliformes (sandgrouse), Columbiformes (pigeons), Cuculiformes (cuckoos), Caprimulgiformes (nightjars) and Apodiformes (swifts, hummingbirds) were defined as near-passerines. All passerines and near-passerines are land birds and have altricial and nidicolous (stay within the nest) chicks, while non-passerine chicks vary in their mode of development and include water and land birds [38]. Sandgrouse are an exception as they have precocial young and are not tree-dwelling [39]. In respect to nest architecture, most passerines build open-cup nests, though some build more elaborate dome structures with roofs [40]. Dome nests, however, are more common among passerines than non-passerines and are particularly frequent among very small passerines [41]. Although these groups are no longer recognized as near-passerines, this definition was used here to distinguish between ecologically profound differences among birds. Avian phylogenetic trees were constructed online (http://www.birdtree.org) from the complete avian phylogeny of Jetz et al. [42] and used the primary backbone tree of Hackett et al. [43]. Ten thousand trees were constructed and statistical analyses were performed in the program R, v. 3.6.1 (R Software, Vienna, Austria, http://www.R-project.org). All quantitative variables (except absolute median latitude, annual temperature and temperature range) were log10-transformed prior to phylogenetic analysis to reduce skewness [44]. As body mass affects all aspects of animal biology and ecology [45], our initial set of phylogenetic analysis account for adult body mass by including this variable as a predictor of log(GH2O). We repeated our phylogenetic analysis using mass-corrected GH2O as the response variable, herein called relative GH2O (RGH2O), thereby removing adult body mass as a predictor. RGH2O values were computed as residuals from a PGLS regression of log(GH2O) on log(body mass) (slope = 0.53 ± 0.03 s.e.; intercept = −0.69 ± 0.12 s.e.; λ = 0.68; electronic supplementary material, figure S1). Using this second series of models, we can ask how well one or more life-history traits results in higher or lower GH2O than is expected for a given body mass of the adult bird. Phylogenetic signal in GH2O and RGH2O was measured by Pagel's lambda (λ) [46] using the phylosig function in the package ‘phytools’ [47] to determine to what extent related species were more likely to share similar conductance values than species drawn randomly from a tree. The phylosig function was used to test the hypothesis that Pagel's λ is different from 0. To test the alternative hypothesis (that Pagel's λ is less than 1), we computed the difference in the log-likelihood ratio of the lambda model (phylosig function) and Brownian motion model (brownie.lite function), then compared it to a chi-squared (χ2) distribution with 1 degree of freedom. PGLS models were fitted using the phylolm function in the package ‘phylolm’ [48]. We ran the full model containing all traits as predictor variables, then used the pdredge function from the package MuMIn [49] to fit all possible model combinations with a maximum of five predictors following protocols by Powney et al. [50], in addition to a null model comprising only the intercept. The best subset of models had an AICc (Akaike's information criterion adjusted for low sample size) within two of the model with the lowest AICc [51]. Conditional model averaging was then used to identify parameter estimates and importance for each trait present in at least one of the subset models [52]. 3. Results In total, we used over 2533 eggs from 364 species to assess diversification in conductance across the avian phylogeny. These species span across 85 families and represent 28 of the 49 extant avian orders. Overall, bird species in Australia, North America and South America had higher log(GH2O) and RGH2O than species in Africa, Europe and Asia (electronic supplementary material, figure S2). GH2O was highest for large flightless birds (e.g. ostriches (Struthio camelus) (106.99 mg day−1 Torr−1)), nightjars (Caprimulgiformes 0.55 ± 0.19 mg day−1 Torr−1) and songbirds (Passeriformes 0.74 ± 0.05 mg day−1 Torr−1). GH2O was also high for aquatic birds (e.g. common loons (Gavia immer) 98.82 mg day−1 Torr−1), kiwis (Southern brown kiwi (Apteryx australis) 26.22 mg day−1 Torr−1) and penguins (Sphenisciformes 22.66 ± 5.45 mg day−1 Torr−1). Viewing total phylogenetic variation in this trait (figures 1a and 2a) revealed that log(GH2O) and RGH2O were typically lower in passerines and near-passerines, than non-passerines (figure 1b). Figure 1. Relationship between conductance of whole eggs and ecological variables for 364 bird species. (a) Phylogenetic tree from which water vapour conductance (GH2O) data were obtained. The bar plot around the phylogeny represents the only significant predictors of log(GH2O) in conditionally averaged models. Conditional model averaging was used to obtain a single average model when more than one PGLS model was best ranked (i.e. more than one model with ΔAICc < 2 from the top-ranked model). Branch colours show the diversification in log(GH2O) across the phylogeny and ancestral trait estimates. GH2O is plotted as a function of (b) avian group and (c) adult body mass (g) within each of the three avian groups. Silhouette illustrations came from PhyloPic (http://phylopic.org), contributed by various authors under public domain licence. (Online version in colour.) Figure 2. Relationship between conductance of whole eggs and ecological variables for 364 bird species. (a) Phylogenetic tree of residual water vapour conductance (RGH2O). Bar plots and rings around the phylogeny represent significant predictors of RGH2O in conditionally averaged models. Conditional model averaging was used to obtain a single average model when more than one PGLS model was best ranked (i.e. more than one model with ΔAICc < 2 from the top-ranked model). Branch colours show the diversification in RGH2O across the phylogeny and ancestral trait estimates. RGH2O is plotted as a function of (b) temperature seasonality within each avian group and (c) whether both parents incubate the eggs. In the hybrid box plot, species RGH2O are shown as filled circles, vertical lines indicate the median, box shows the interquartile range (IQR) and the whiskers are 1.5 × IQR (distribution is shown as histograms). p-values are given in asterisks, with *less than 0.05, **less than 0.01 and ***less than 0.001. Silhouette illustrations came from PhyloPic (http://phylopic.org), contributed by various authors under public domain license. (Online version in colour.) (a) Phylogenetic correlation Phylogenetic signal for log(GH2O) and RGH2O (table 2) was significantly different from 0 (i.e. no phylogenetic signal) (p < 0.001) and 1 (i.e. the Brownian explanation) (p < 0.001), meaning that while there is an effect of phylogeny on conductance, it is influenced by evolutionary processes that are weaker than would be seen with a Brownian motion model of trait evolution. Phylogenetic signal was high for Log(GH2O) (λ = 0.96), showing that closely related species exhibit similar eggshell conductance prior to accounting for differences in body mass, and this biological similarity decreases as the evolutionary distance between species increases. Phylogenetic signal was intermediate for RGH2O (λ = 0.55), suggesting that phylogeny and other selective pressures (e.g. those associated with species life history or climate) are important in determining eggshell conductance, after accounting for differences in species body mass. Table 2. Estimates of phylogenetic signal in eggshell water vapour conductance (GH2O) in all birds. Phylogenetic signal was analysed separately for log10-transformed GH2O (log(GH2O)) and residual water vapour conductance (RGH2O). The p-value tests the null hypothesis of no phylogenetic signal (λ = 0) and Brownian motion model (λ = 1) of evolution. response variable Pagel's λ log-likelihood log-likelihood for λ = 0 log-likelihood for λ = 1 p for λ = 0 p for λ = 1 log(GH2O) 0.96 −74.39 590.76 −125.64 <0.001 <0.001 RGH2O 0.55 27.20 258.50 −92.27 <0.001 <0.001 (b) Life history and climate influence conductance across birds Adult body mass and temperature seasonality were the strongest predictors of log(GH2O) across all birds based on conditionally averaged models (electronic supplementary material, table S1). Log(GH2O) was significantly higher among heavier species (z = 18.40, p < 0.001; figures 1c and 3a) since initial egg mass increases with adult body mass (n = 251, r2 = 0.89, p < 0.001 [52]). Log(GH2O) was negatively associated with increased temperature seasonality across all birds (z = 2.13, p = 0.03; figures 2b and 3b). Temperature seasonality is defined here as the amount of temperature variation over a given year (or averaged years) based on the standard deviation of monthly temperature averages [53]. There was also a weaker yet significant effect of dietary calcium, nest location, mode of development, shared incubation and parental contact among top-ranked models (electronic supplementary material, table S2). Figure 3. Partitioning of variation of water vapour conductance (GH2O­) among avian orders. Dots (one per species) (n = 364) show the distribution of log(GH2O­) as a function of (a) adult body mass (g) and (b) residual water vapour conductance (RGH2O­) as a function of temperature seasonality. The minimum convex hull is plotted for all species within a subset of avian orders. Silhouette illustrations came from PhyloPic (http://phylopic.org), contributed by various authors under public domain license. (Online version in colour.) Temperature seasonality (z = 2.20, p = 0.03) and whether contact incubation was shared among parents (z = 2.22, p = 0.03) were significant in conditionally averaged models after accounting for adult body mass (RGH2O) (electronic supplementary material, table S3; figure 2). RGH2O overall decreased with temperature seasonality (figures 2b and 3b). RGH2O was higher in species where both parents incubate the clutch (figure 2c). Dietary calcium, mode of development, nest location and parental contact showed weaker but significant correlations with RGH2O among top-ranked models (electronic supplementary material, table S4). RGH2O was higher in species with calcium-rich diets, precocial young, parents that return to the nest with wet plumage and ground nesters compared to tree nesters (electronic supplementary material, figure S6). Based on conditionally averaged models for GH2O and RGH2O, eggshell conductance across birds is primarily influenced by adult body mass, temperature seasonality and parent incubation strategies. 4. Discussion This study focused on one performance trait—conductance—of modern avian eggshells to better understand how birds have achieved high ecological diversity. We identified the importance of phylogeny, physiology (body mass and mode of development), behaviour (diet, parental incubation strategies and nest location) and climate in the evolution of this trait. This study is the first to identify a broad-scale reduction in eggshell conductance where temperature seasonality increases. Regions with greater temperature seasonality experience a greater range in temperatures over the course of a year and correlate with an organisms' temperature tolerance breadth [54]. Increased temperature seasonality occurs further from the equator and is associated with a decline in annual temperature, precipitation and day length [55]. A comparative study on 139 bird species found that adults inhabiting low and seasonally variable temperatures had lower basal metabolic rate after removing the effects of body mass [56]. In the light of this, it appears possible that eggshells are already preparing the embryo for adulthood, with respect to their environment and breeding biology. Amniotic embryos adjust their metabolic activity and active cell division in response to varying environmental conditions, and by doing so, alter their period of development [57]. Reproductive strategies to prolong the egg state are most diverse in reptiles and less varied in birds and mammals that provide more parental care [58]. Even so, the low metabolic rate expected for embryos incubated in highly seasonal climates would favour a reduction in conductance to prolong their incubation period. Broad-scale geographical trends in RGH2O identified here may be the result of long-term evolutionary responses or short-term physiological modifications [59]. Evolutionary adaptation would involve changes in GH2O over (rather than within) generations when natural selection acts on genetic variants while acclimatization would involve reversible changes to GH2O that can happen gradually (greater than 1 day) in response to the recent environment [60]. Intraspecific variation in GH2O has been reported across altitude [61,62] and humidity [63,64] gradients of multiple species, but the timeframe in which GH2O diversification has taken place is unknown. Some studies propose that rapid evolution of eggshell structure from exposure to novel environments is unlikely [65,66] and is instead compensated by behavioural modifications of the parents. Other studies find that incubation behaviour does not significantly modulate conductance [67], so adaptive responses must be accomplished by changes in eggshell structure [63]. Birds are seemingly capable of short-term and instantaneous physiological adjustments in shell structure in response to environmental variation. Pigeons (Columba livia) bred for several years within an environmentally controlled room experienced approximately 30% lower GH2O than predicted when exposed to high temperature and low humidity over a short period [68]. Similarly, domestic chickens (Gallus domesticus) bred at high elevation for multiple generations produced eggshells with a 30% higher GH2O within two months of being translocated to low altitudes [69]. In other species, GH2O did not change when individuals were transferred to higher altitudes [70] or were exposed to natural seasonal changes in humidity [71], suggesting there is variation in the plasticity of a species response. Identifying the speed of the response in eggshell parameters to novel environments across multiple species will be very informative in determining climate change effects on bird species and their breeding. We found that conductance across birds was also dependent on nest location, whether parents alternate nest attendance, and whether the parent returns to the nest with wet plumage, corroborating previous studies [25,72]. Shared incubation between two parents allows one of them to be relieved from incubation to feed while the other incubates the egg, thus allowing the eggs to be covered at all times [73]. Clutches that are incubated by both parents encounter less variation in egg temperature than clutches that are incubated by a single parent [74] and thus, are expected to have higher eggshell conductance. Water added to the nest by parents can be many orders of magnitude higher than water lost by the eggs [75]. Consequently, RGH2O is significantly higher in species where parents return the nest with wet plumage [25]. Eggs laid on the ground, in a burrow, mound or on floating vegetation are subject to higher humidity than arboreal nesters, leading to eggshell adaptations that promote water loss. Common loons (Gavia immer), for example, had the highest RGH2O of the species investigated. This may be attributed to their high eggshell porosity [76] in response to building nests on or near the water where transpiration of water is high, and nest materials can be wet [64]. Nest location and whether parents return to the nest with wet or dry plumage was significant in most top models where these predictors were included, but this effect was weak compared to life-history traits retained in conditionally averaged models. Combined, our results demonstrate that different behavioural strategies used by parents to alter nest humidity have contributed to the evolution of conductance among birds. Variation in the incubation period across the altricial-precocial spectrum reflects a trade-off between embryo growth rate and degree of maturity when hatched. Precocial species take up to two times longer to incubate an egg of the same size as altricial species, but are far more developed when they hatch [77]. For eggs of the same mass, precocial species incur a higher total energy cost than altricial species because the embryo is larger for a longer period during incubation [78]. Consequently, eggs of species with fast (precocial) growing offspring had significantly higher RGH2O than those of species with slow (altricial) growing offspring based on top-ranked models. As higher conductance enables greater gas exchange, this may optimize embryo access to high energy content in precocial eggs [79], thus resulting in a more developed chick at birth. RGH2O in passerines was found here to be particularly low, likely because they have altricial young, whereas non-passerines consist of precocial and altricial species. Supplementary Material Click here for additional data file. Acknowledgements We are grateful to Douglas Russell at The Natural History Museum Tring for his generous assistance in working with the eggshell collection, and for useful discussions. We thank Craig White for useful discussions and providing code for phylogenetic comparative analysis, and Stephanie McClelland, Jennifer Cantlay and Jack Thirkell for their comments on early drafts. Data accessibility Data are publicly available in the FigShare repository, including specimen and species-specific water vapour conductance, life histories and sources used in this study (doi:10.6084/m9.figshare.12490559) [80]. Tables for all PGLS analyses and sources for figure illustrations are available in the electronic supplementary material. Authors' contributions M.R.G.A.: conceptualization, data curation, formal analysis, investigation, methodology, project administration, validation, visualization, writing-original draft, writing-review and editing; S.J.P.: conceptualization, formal analysis, funding acquisition, investigation, methodology, project administration, resources, supervision, writing-review and editing All authors gave final approval for publication and agreed to be held accountable for the work performed therein. Competing interests We declare we have no competing interests. Funding This project and M.R.G.A. was funded by a Research Project (grant no. RPG-2018-332) from the Leverhulme Trust, awarded to S.J.P. ==== Refs References 1. Losos JB. 2010 Adaptive radiation, ecological opportunity, and evolutionary determinism. Am. Nat. 175 , 623-639. (10.1086/652433)20412015 2. Yoder JB et al . 2010 Ecological opportunity and the origin of adaptive radiations. J. Evol. Biol. 23 , 1581-1596. (10.1111/j.1420-9101.2010.02029.x)20561138 3. Losos JB, Schoener TW, Langerhans RB, Spiller DA. 2006 Rapid temporal reversal in predator-driven natural selection. Science 314 , 1111. (10.1126/science.1133584)17110568 4. Ho W, Zhang J. 2018 Evolutionary adaptations to new environments generally reverse plastic phenotypic changes. Nat. Commun. 9 , 350. (10.1038/s41467-017-02724-5)29367589 5. Hendry AP, Gotanda KM, Svensson EI. 2017 Human influences on evolution, and the ecological and societal consequences. Phil. Trans. R. Soc. B 372 , 20160028. (10.1098/rstb.2016.0028)27920373 6. Hoffmann AA, Hercus MJ. 2000 Environmental stress as an evolutionary force. Bioscience 50 , 217-226. (10.1641/0006-3568(2000)050[0217:ESAAEF]2.3.CO;2) 7. Durant SE, Willson JD, Carroll RB. 2019 Parental effects and climate change: will avian incubation behavior shield embryos from increasing environmental temperatures? Integr. Comp. Biol. 59 , 1068-1080. (10.1093/icb/icz083)31168619 8. Ar A, Rahn H. 1980 Water in the avian egg overall budget of incubation. Integr. Comp. Biol. 20 , 373-384. (10.1093/icb/20.2.373) 9. Carey C. 1986 Tolerance of variation in eggshell conductance, water loss, and water content by red-winged blackbird embryos. Physiol. Zool. 59 , 109-122. (10.1086/physzool.59.1.30156096) 10. Wangensteen OD, Rahn H. 1970 Respiratory gas exchange by the avian embryo. Respir. Physiol. 11 , 31-45. (10.1016/0034-5687(70)90100-3)5491114 11. Konishi M, Emlen ST, Ricklefs RE, Wingfield JC. 1989 Contributions of bird studies to biology. Science 246 , 465-472. (10.1126/science.2683069)2683069 12. Le Maho Y. 1977 The emperor penguin: a strategy to live and breed in the cold. Am. Sci. 65 , 680-693. 13. Carey C. 2002 Incubation in extreme environments. In Avian incubation: behaviour, environment, and evolution (ed. DC Deeming), pp. 238-253. Oxford, UK: Oxford University Press. 14. Zicus MC, Rave DP, Riggs MR, Zicus MC, Rave DP, Riggs MR. 2003 Mass loss from mallard eggs incubated in nest structures. Wildl. Soc. Bull. 31 , 270-278. 15. Carey C. 1980 Adaptation of the avian egg to high altitude. Am. Zool. 20 , 449-459. (10.1093/icb/20.2.449) 16. Martin K, Wiebe KL. 2004 Coping mechanisms of alpine and arctic breeding birds: extreme weather and limitations to reproductive resilience. Integr. Comp. Biol. 44 , 177-185. (10.1093/icb/44.2.177)21680497 17. Spellerberg IF. 1969 Incubation temperatures and thermoregulation in the McCormick Skua. Condor 71 , 59-67. (10.2307/1366049) 18. McKechnie AE, Wolf BO. 2010 Climate change increases the likelihood of catastrophic avian mortality events during extreme heat waves. Biol. Lett. 6 , 253-256. (10.1098/rsbl.2009.0702)19793742 19. Öberg M, Arlt D, Pärt T, Laugen AT, Eggers S, Low M. 2015 Rainfall during parental care reduces reproductive and survival components of fitness in a passerine bird. Ecol. Evol. 5 , 345-356. (10.1002/ece3.1345)25691962 20. Cloudsley-Thompson JL. 1988 Adaptations to Extreme Environments. In Evolution and adaptation of terrestrial arthropods (ed. JL Cloudsley-Thompson), pp. 80-98. Berlin, Germany: Springer. 21. Seymour RS, Ackerman RA. 1980 Adaptations to underground nesting in birds and reptiles. Am. Zool. 20 , 437-447. (10.1093/icb/20.2.437) 22. Rahn H, Hammel HT. 1982 Incubation water loss, shell conductance, and pore dimensions in Adelie penguin eggs. Polar Biol. 1 , 91-97. (10.1007/BF00263805) 23. Adkerman RA, Platter-Rieger M. 1979 Water loss by pied-billed grebe (Podilymbus podiceps) eggs. Arner. Zool. 19 , 921. 24. Lill A, Fell PJ. 2007 Microclimate of nesting burrows of the Rainbow Bee-eater. Emu 107 , 108-114. (10.1071/MU06046) 25. Portugal SJ, Maurer G, Thomas GH, Hauber ME, Grim T, Cassey P. 2014 Nesting behaviour influences species-specific gas exchange across avian eggshells. J. Exp. Biol. 217 , 3326-3332. (10.1242/jeb.103291)25232199 26. Ericson PG, Klopfstein S, Irestedt M, Nguyen JM, Nylander JA. 2014 Dating the diversification of the major lineages of Passeriformes (Aves). BMC Evol. Biol. 14 , 8. (10.1186/1471-2148-14-8)24422673 27. Rahn H, Dawson WR. 1979 Incubation water loss in eggs of Heermann's and western gulls. Physiol. Zool. 52 , 451-460. (10.1086/physzool.52.4.30155936) 28. Arad Z, Marder J. 1982 Egg-shell water vapour conductance of the domestic fowl: comparison between two breeds and their crosses. Br. Poult. Sci. 23 , 325-328. (10.1080/00071688208447964) 29. Guerra C, Aguilar R, Fitzpatrick L. 1988 Water vapor conductance in Gray gulls (Larus modestus) eggs: adaptation to desert nesting. Colon. Waterbirds 11 , 107-109. (10.2307/1521176) 30. Sotherland P, Ashen M, Shuman R, Tracy C. 1984 The water balance of bird eggs incubated in water. Physiol. Zool. 57 , 338-348. (10.1086/physzool.57.3.30163723) 31. Davis TA, Ackerman RA. 1985 Adaptations of black tern (Chlidonias niger) eggs for water loss in a moist nest. Auk 102 , 640-643. (10.1093/auk/102.3.640) 32. Portugal SJ, Hauber ME, Maurer G, Stokke BG, Grim T, Cassey P. 2014 Rapid development of brood-parasitic cuckoo embryos cannot be explained by increased gas exchange through the eggshell. J. Zool. 293 , 219-226. (10.1111/jzo.12144) 33. Rokitka MA, Rahn H. 1987 Regional differences in shell conductance and pore density of avian eggs. Respir. Physiol. 68 , 371-376. (10.1016/S0034-5687(87)80021-X)3616182 34. Booth DT, Seymour RS. 1987 Effect of eggshell thinning on water vapor conductance of malleefowl eggs. Condor 89 , 453-459. (10.2307/1368635) 35. Bamelis FR, De Ketelaere B, Mertens K, Kemps BJ, Decuypere EM, De Baerdemaeker JG. 2008 Measuring the conductance of eggshells using the acoustic resonance technique and optical transmission spectra. Comput. Electron. Agric. 62 , 35-40. (10.1016/j.compag.2007.08.009) 36. Portugal SJ, Maurer G, Cassey P. 2010 Eggshell permeability: a standard technique for determining interspecific rates of water vapor conductance. Physiol. Biochem. Zool. 83 , 1023-1031. (10.1086/656287)20939733 37. Symonds MRE, Blomberg SP. 2014 A primer on phylogenetic generalised least squares. In Modern phylogenetic comparative methods and their application in evolutionary biology (ed. L Garamszegi), pp. 105-130. Berlin, Germany: Springer. 38. Carnaby T. 2010 Beat about the bush, birds. Johannesburg, South Africa: Jacana Media. 39. Temrin H, Tullberg BS. 1995 A phylogenetic analysis of the evolution of avian mating systems in relation to altricial and precocial young. Behav. Ecol. 6 , 296-307. (10.1093/beheco/6.3.296) 40. Price JJ, Griffith SC. 2017 Open cup nests evolved from roofed nests in the early passerines. Proc. R. Soc. B 284 , 20162708. (10.1098/rspb.2016.2708) 41. Collias NE. 1997 On the origin and evolution of nest building by passerine birds. Condor 99 , 253-270. (10.2307/1369932) 42. Jetz W, Thomas GH, Joy JB, Hartmann K, Mooers AO. 2012 The global diversity of birds in space and time. Nature 491 , 444-448. (10.1038/nature11631)23123857 43. Hackett SJ et al . 2008 A phylogenomic study of birds reveals their evolutionary history. Science 320 , 1763-1768. (10.1126/science.1157704)18583609 44. Feng C, Wang H, Lu N, Chen T, He H, Lu Y, Tu XM. 2014 Log-transformation and its implications for data analysis. Shanghai Arch. psychiatry 26 , 105-109. (10.3969/j.issn.1002-0829.2014.02.009)25092958 45. Brown JH. 1995 Macroecology. Chicago, IL: University of Chicago Press. 46. Pagel M. 1999 Inferring the historical patterns of biological evolution. Nature 401 , 877-884. (10.1038/44766)10553904 47. Revell LJ. 2012 phytools: an R package for phylogenetic comparative biology (and other things). Methods Ecol. Evol. 3 , 217-223. (10.1111/j.2041-210X.2011.00169.x) 48. Ho LST, Ané C, Lachlan R, Tarpinian K, Feldman R, Yu Q, Ho MLST. 2014 Phylolm: phylogenetic linear regression. R package version 2.1. 49. Barton K, Barton MK. 2019 Package ‘MuMIn’. R package version 1. 50. Powney GD, Rapacciuolo G, Preston CD, Purvis A, Roy DB. 2014 A phylogenetically-informed trait-based analysis of range change in the vascular plant flora of Britain. Biodivers. Conserv. 23 , 171-185. (10.1007/s10531-013-0590-5) 51. Angert AL, Crozier LG, Rissler LJ, Gilman SE, Tewksbury JJ, Chunco AJ. 2011 Do species' traits predict recent shifts at expanding range edges? Ecol. Lett. 14 , 677-689. (10.1111/j.1461-0248.2011.01620.x)21535340 52. Deeming DC. 2007 Effects of phylogeny and hatchling maturity on allometric relationships between female body mass and the mass and composition of bird eggs. Avian Poult. Biol. Rev. 18 , 21-37. (10.3184/147020607X245039) 53. O'Donnell MS, Ignizio DA. 2012 Bioclimatic predictors for supporting ecological applications in the conterminous United States. US Geol. Surv. Data Ser. 691 , 4-9. (10.3133/ds691) 54. Sunday J et al . 2019 Thermal tolerance patterns across latitude and elevation. Phil. Trans. R. Soc. B 374 , 20190036. (10.1098/rstb.2019.0036)31203755 55. De Frenne P et al . 2013 Latitudinal gradients as natural laboratories to infer species' responses to temperature. J. Ecol. 101 , 784-795. (10.1111/1365-2745.12074) 56. White CR, Blackburn TM, Martin GR, Butler PJ. 2007 Basal metabolic rate of birds is associated with habitat temperature and precipitation, not primary productivity. Proc. R. Soc. B 274 , 287-293. (10.1098/rspb.2006.3727) 57. Rafferty AR, Reina RD. 2012 Arrested embryonic development: a review of strategies to delay hatching in egg-laying reptiles. Proc. R. Soc. B 279 , 2299-2308. (10.1098/rspb.2012.0100) 58. Maurer G, Portugal SJ, Cassey P. 2011 Review: an embryo's eye view of avian eggshell pigmentation. J. Avian Biol. 42 , 494-504. (10.1111/j.1600-048X.2011.05368.x) 59. Walsberg GE, Voss-Roberts KA. 1983 Incubation in desert-nesting doves: mechanisms for egg cooling. Physiol. Zool. 56 , 88-93. (10.1086/physzool.56.1.30159969) 60. Llewelyn J, Macdonald SL, Moritz C, Martins F, Hatcher A, Phillips BL. 2018 Adjusting to climate: acclimation, adaptation and developmental plasticity in physiological traits of a tropical rainforest lizard. Integr. Zool. 13 , 411-427. (10.1111/1749-4877.12309)29316349 61. Carey C, Garber SD, Thompson EL, James FC. 1983 Avian reproduction over an altitudinal gradient. II. Physical characteristics and water loss of eggs. Physiol. Zool. 56 , 340-352. (10.1086/physzool.56.3.30152599) 62. Sotherland PR, Packard GC, Taigen TL, Thomas J. 1980 An altitudinal cline in conductance of cliff swallow (Petrochelidon pyrrhonota) eggs to water vapor. Auk 97 , 177-185. (10.1093/auk/97.1.177) 63. Stein LR, Badyaev AV. 2011 Evolution of eggshell structure during rapid range expansion in a passerine bird. Funct. Ecol. 25 , 1215-1222. (10.1111/j.1365-2435.2011.01887.x) 64. Davis A, Platter-Reiger MF, Ackerman RA. 1984 Incubation water loss by pied-billed grebe eggs: adaptation to a hot, wet nest. Physiol. Zool. 57 , 384-391. (10.1086/physzool.57.4.30163340) 65. Simkiss K. 1980 Eggshell porosity and the water metabolism of the chick embryo. J. Zool. 192 , 1-8. (10.1111/j.1469-7998.1980.tb04213.x) 66. Board R. 1982 Properties of avian eggshells and their adaptive value. Biol. Rev. 57 , 1-28. (10.1111/j.1469-185X.1982.tb00362.x) 67. Walsberg GE. 1983 A test for regulation of nest humidity in two bird species. Physiol. Zool. 56 , 231-235. (10.1086/physzool.56.2.30156054) 68. Arad Z, Gavrieli-Levin I, Marder J. 1988 Adaptation of the pigeon egg to incubation in dry hot environments. Physiol. Zool. 61 , 293-300. (10.1086/physzool.61.4.30161246) 69. Rahn H, Ledoux T, Paganelli CV, Smith AH. 1982 Changes in eggshell conductance after transfer of hens from an altitude of 3800 to 1200 m. J. Appl. Physiol. 53 , 1429-1431. (10.1152/jappl.1982.53.6.1429)7153140 70. Carey C, Hoyt DF, Bucher TL, Larson DL. 1984 Eggshell conductances of avian eggs at different altitudes. In Respiration and metabolism of embryonic vertebrates (ed. RS Seymour), pp. 259-270. Berlin, Germany: Springer. 71. Walsberg GE. 1985 A test for regulation of egg dehydration by control of shell conductance in Mourning Doves. Physiol. Zool. 58 , 473-477. 72. Vleck CM, Vleck D, Rahn H, Paganelli CV. 1983 Nest microclimate, water-vapor conductance, and water loss in heron and tern eggs. Auk 100 , 76-83. (10.1093/auk/100.1.76) 73. Seddon P. 1989 Patterns of nest relief during incubation, and incubation period variability in the yellow-eyed penguin (Megadyptes antipodes). New Zeal. J. Zool. 16 , 393-400. (10.1080/03014223.1989.10422905) 74. Reneerkens J, Grond K, Schekkerman H, Tulp I, Piersma T. 2011 Do uniparental sanderlings Calidris alba increase egg heat input to compensate for low nest attentiveness? PLoS ONE 6 , 1-9. (10.1371/journal.pone.0016834) 75. Grant GS. 1982 Avian incubation: egg temperature, nest humidity, and behavioral thermoregulation in a hot environment. Washington, DC: American Ornithologists’ Union. 76. Tullett SG, Board RG. 1977 Determinants of avian eggshell porosity. J. Zool. 183 , 203-211. (10.1111/j.1469-7998.1977.tb04182.x) 77. Ricklefs RE, Austin SH, Robinson WD. 2017 The adaptive significance of variation in avian incubation periods. Auk 134 , 542-550. (10.1642/AUK-16-171.1) 78. Hoyt DF. 1987 A new model of avian embryonic metabolism. J. Exp. Zool. Suppl 1 , 127-138.3598486 79. Sotherland PR, Rahn H. 1987 On the composition of bird eggs. Condor 89 , 48-65. (10.2307/1368759) 80. Attard MRG, Portugal SJ. 2021 Climate variability and parent nesting strategies influence gas exchange across avian eggshells. FigShare.
PMC008xxxxxx/PMC8206691.txt
==== Front Proc Biol Sci Proc Biol Sci RSPB royprsb Proceedings of the Royal Society B: Biological Sciences 0962-8452 1471-2954 The Royal Society 34130499 10.1098/rspb.2021.0616 rspb20210616 10017060Evolution Research Articles Juvenile ecology drives adult morphology in two insect orders Juvenile ecology drives adult morphology in two insect orders http://orcid.org/0000-0003-2776-6172 Rühr Peter T. Conceptualization Data curation Formal analysis Investigation Methodology Project administration Visualization Writing-original draft Writing-review & editing ruehr@uni-bonnde 1 http://orcid.org/0000-0001-7390-1318 van de Kamp Thomas Investigation Methodology Resources Software Writing-review & editing 2 3 Faragó Tomáš Investigation Methodology Resources Software Writing-review & editing 3 http://orcid.org/0000-0002-6744-6811 Hammel Jörg U. Investigation Methodology Resources Software Writing-review & editing 4 Wilde Fabian Investigation Methodology Resources Software Writing-review & editing 4 Borisova Elena Investigation Methodology Resources Software Writing-review & editing 5 http://orcid.org/0000-0003-1760-6669 Edel Carina Investigation Writing-review & editing 1 http://orcid.org/0000-0002-5264-6319 Frenzel Melina Investigation Writing-review & editing 1 Baumbach Tilo Resources Writing-review & editing 2 3 http://orcid.org/0000-0003-4385-6039 Blanke Alexander Conceptualization Formal analysis Funding acquisition Investigation Project administration Resources Supervision Writing-review & editing 1 6 1 Institute of Evolutionary Biology and Animal Ecology, University of Bonn, An der Immenburg 1, 53121 Bonn, Germany 2 Institute for Photon Science and Synchrotron Radiation (IPS), Karlsruhe Institute of Technology (KIT), Hermann-von-Helmholtz-Platz 1, 76344 Eggenstein-Lepoldshafen, Germany 3 Laboratory for Applications of Synchrotron Radiation (LAS), Karlsruhe Institute of Technology (KIT), Kaiserstr. 12, 76131 Karlsruhe, Germany 4 Institute of Materials Physics, Helmholtz-Zentrum Hereon, Max-Planck-Straße 1, 21502 Geesthacht, Germany 5 Swiss Light Source, Paul Scherrer Institute, Forschungsstrasse 111, 5232 Villigen, Switzerland 6 Medical and Biological Engineering Research Group, School of Engineering and Computer Science, University of Hull, Hull HU6 7RX, UK Electronic supplementary material is available online at https://doi.org/10.6084/m9.figshare.c.5452532. 30 6 2021 June 30, 2021 16 6 2021 June 16, 2021 16 6 2021 June 16, 2021 288 1953 2021061612 3 2021 March 12, 2021 24 5 2021 May 24, 2021 © 2021 The Authors. 2021 https://creativecommons.org/licenses/by/4.0/ Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited. Most animals undergo ecological niche shifts between distinct life phases, but such shifts can result in adaptive conflicts of phenotypic traits. Metamorphosis can reduce these conflicts by breaking up trait correlations, allowing each life phase to independently adapt to its ecological niche. This process is called adaptive decoupling. It is, however, yet unknown to what extent adaptive decoupling is realized on a macroevolutionary scale in hemimetabolous insects and if the degree of adaptive decoupling is correlated with the strength of ontogenetic niche shifts. It is also unclear whether the degree of adaptive decoupling is correlated with phenotypic disparity. Here, we quantify nymphal and adult trait correlations in 219 species across the whole phylogeny of earwigs and stoneflies to test whether juvenile and adult traits are decoupled from each other. We demonstrate that adult head morphology is largely driven by nymphal ecology, and that adult head shape disparity has increased with stronger ontogenetic niche shifts in some stonefly lineages. Our findings implicate that the hemimetabolan metamorphosis in earwigs and stoneflies does not allow for high degrees of adaptive decoupling, and that high phenotypic disparity can even be realized when the evolution of distinct life phases is coupled. adaptive decoupling , metamorphosis , complex life cycles , ontogenetic niche shifts , disparity , geometric morphometrics Karlsruhe Institute of Technology http://dx.doi.org/10.13039/100009133 BL 1355/4-1 Helmholtz-Zentrum Geesthacht I-20170190 I-20170896 I-20190019 H2020 European Research Council http://dx.doi.org/10.13039/100010663 754290 Deutsche Forschungsgemeinschaft http://dx.doi.org/10.13039/501100001659 BL 1355/4-1 Paul Scherrer Institut http://dx.doi.org/10.13039/501100004219 20171469 cover-dateJune 30, 2021 ==== Body pmc1. Background Species ecology can change dramatically during development [1,2], a process called ‘ontogenetic niche shift' [3]. If phenotypic traits are coupled between life phases, ontogenetic niche shifts may result in adaptive conflicts, because coupled traits cannot evolve independently according to divergent, life-phase-specific needs [2,4,5]. Metamorphosis, a process of rapid change in morphology, physiology and behaviour [6,7], can break up trait correlations between life phases [5,8–12] and therefore allows for an adaptive decoupling of traits [4,13,14]. Even though around 80% of animals, including all winged insect lineages (Pterygota), show complex life cycles with metamorphic periods [13], the macroevolutionary relationship between ontogenetic niche shifts and adaptive decoupling has only been studied in Echinodermata [9,10,15], ray-finned fishes [16], frogs [17–19] and salamanders [20]. In these groups, it has been shown that adaptive decoupling via metamorphosis seems to largely allow independent phenotypic evolution in distinct life-history phases. In insects, however, no macroevolutionary assessment of adaptive decoupling has been undertaken, so it remains unknown to what extent the nymphal life phase influences the phenotypic evolution of the adult phase. Winged insects show two types of metamorphosis: hemimetaboly and holometaboly. Holometabolous insects possess a pupal stage between the adult and the juvenile life phase in which strong morphological remodelling takes place (‘complete metamorphosis' [7,11]). While some traits may still be linked [21–24], general morphology, feeding behaviour, and modes of locomotion of holometabolous juveniles and their adults are usually strongly divergent [4,7]. The separation of life phases by complete metamorphosis has been hypothesized as a key innovation driving insect diversification [25,26]. Hemimetabolous insects, which gradually develop towards adulthood without a pupal stage, can also show strong ontogenetic niche shifts between the juvenile and the adult phases. For example, dragonflies (Odonata), mayflies (Ephemeroptera) and stoneflies (Plecoptera) shift from aquatic to terrestrial habitats, and some stonefly species develop from predatory nymphs to non-feeding adults. Yet, considerable morphological changes during this step mostly affect wings, genitals and gills, while the general body plans of nymphs and their adults remain relatively similar (hemimetabolan metamorphosis [7]). Here, we used three-dimensional geometric morphometrics, a de novo generated database on ecological traits and multivariate statistics to investigate macroevolutionary correlations between nymphal ecology and adult shape. We hypothesize that (i) the single moult between the juvenile and the adult life phase might not be able to allow for high degrees of adaptive decoupling, (ii) adaptive decoupling is stronger in taxa with stronger ontogenetic niche shifts, and (iii) the independent evolution of life phases in taxa with strong adaptive decoupling results in diversification into more ecological niches and thus a higher shape disparity. A statistically rigorous approach required selecting two morphologically similar lineages with different ecologies, which are as closely related as possible to allow a meaningful correction of potentially biasing phylogenetic signal [27–29]. We chose to study the two closely related [30,31] hemimetabolous insect orders earwigs (Polyneoptera: Dermaptera, approx. 2000 described species [32]) and stoneflies (Polyneoptera: Plecoptera, approx. 3400 species [32]). Earwigs and stoneflies both possess prognathous biting–chewing mouthparts, similar antennal and eye positions and sizes, a low performance flight apparatus, and exhibit a similar size range with body lengths mostly between 10 and 30 mm [33,34]. However, earwigs and stoneflies differ in the degree of ontogenetic niche shift: earwig nymphs and adults are both terrestrial (figure 1) and feed on similar food sources [33], while stoneflies, with few secondarily evolved exceptions (e.g. [35,36]), are the only polyneopterans that show an amphibiotic lifestyle with aquatic nymphs and terrestrial adults (figure 1) [34]. They transition from an aquatic to a terrestrial environment over the course of the final moult, often accompanied by a shift in feeding mode [37]. Figure 1. Overview of microhabitat occupation by nymphal and adult earwigs and stoneflies. Insect pictograms indicate different taxa and life stages. Habitat preference colour-coded. (Online version in colour.) 2. Material and methods (a) Taxon sampling We studied 219 species (electronic supplementary material, table S10), 144 earwigs and 75 stoneflies, covering all extant families, 80.3% of extant subfamilies and 32.28% of extant genera [38,39]. Specimens were loaned from the Natural History Museum (NHMUK) (London, UK), Museum für Naturkunde (MfN) (Berlin, Germany), Zoologisches Forschungsmuseum Alexander Koenig (ZFMK) (Bonn, Germany), Muséum national d'Histoire naturelle (MNHM) (Paris, France), Naturhistorisches Museum (NHMV) (Vienna, Austria), Zoologische Staatssammlung München (ZSM) (Munich, Germany) and several private collectors (see Acknowledgements). To exclude the use of possibly shrunken museum material, almost all Plecoptera samples analysed in this study were loaned as alcohol specimens and dried at the critical point (EM CPD300, Leica Microsystems GmbH, Wetzlar, Germany) prior to scanning. Six air-dried specimens from the Paris and Berlin collections, in which no shrinking could be detected, were used as well. (b) Tomography scanning and data processing Head shape was investigated using synchrotron radiation micro-computed tomography (SR-µCT). This allowed for a high spatial resolution, high tissue contrast and rapid image acquisition. 177 specimens were scanned at the imaging cluster of the KIT Light Source (Karlsruhe Institute of Technology (KIT), Karlsruhe, Germany), 39 specimens at the IBL-P05 imaging beamline [40–42] (operated by the Helmholtz-Zentrum Hereon at PETRA III, Deutsches Elektronen Synchrotron (DESY), Hamburg, Germany), and one specimen at the TOMCAT beamline [43] (Swiss Light Source (SLS), Paul-Scherrer-Institute (PSI), Villigen, Switzerland). Two larger specimens were scanned with a commercially available µCT-machine (phoenix nanotom, General Electric, Boston, MA) operated by Hereon. Rotated, three-dimensional regions of interest (ROIs) of the insect heads were manually extracted from the virtual image stacks and downsampled to less than 300 MB using a custom macro for FIJI [44] available online (github.com/Peter-T-Ruehr/stack_cropping). A second Fiji macro (github.com/Peter-T-Ruehr/checkpoint_converter) was used to convert the image stacks to ‘Checkpoint’ files (*.ckpt) including their associated *.tif stacks to skip the manual import within ‘Checkpoint’ v. 6 (Stratovan Corporation, Davis, CA). Downsampled tomography scans are available at Zenodo (doi:10.5281/zenodo.4280412). (c) Head shape quantification with three-dimensional geometric morphometrics Shape was quantified by using 3D geometric morphometrics (41 landmarks per species). Ten homologous type 1 landmarks, six type 2 landmarks [45] and 25 curve sliding semilandmarks (electronic supplementary material, figure S1a, tables S6 and S12) were placed on each 3D head model in Checkpoint to capture the head shape diversity of our sample. All type 2 landmarks and semilandmarks lie along the midsagittal plane of the head. Mandible shape was characterized by six type 1 landmarks (electronic supplementary material, figure S1b, tables S6 and S12). To skip manual landmark export in Checkpoint, a custom script (github.com/Peter-T-Ruehr/checkpoint_importer_for_R) was used to import the landmarks of all species into the programming environment ‘R' v. 4.0.5 [46] directly from the Checkpoint files. Generalized procrustes analyses (GPAs) were performed using ‘gpagen' in ‘geomorph' v. 3.3.1 [47] to remove the effects of non-shape variation from the dataset [48,49]. Sliding of the semilandmarks during the GPA was based on minimizing bending energy. Procrustes distance outliers were identified for each superfamily using the ‘plotOutliers' function in geomorph and their landmarks were double checked. Head shape variation was subsequently investigated via phylogenetic PCAs (pPCAs [28,50]) and visualized with the ‘phylomorphospace' function in ‘phytools' v. 0.6.99 [51]. (d) Database on ecological and morphological traits We established a novel literature database to link life stage-specific ecological data with adult shape variation by screening 1950 literature sources, of which 960 were informative (listed in electronic supplementary material, table S11), for information on the following traits: ‘microhabitat' and ‘feeding habits' (for both nymphs and adults), and ‘hydrodynamic pressure' (for nymphs). For each of these characters, we defined several character states (electronic supplementary material, table S7). Following Wilman et al. [52], who introduced a standardized interpretation of ecological and morphological wording, the often non-quantitative expert descriptions of traits were translated into semiquantitative information about the relative importance of this trait within its category in scores from 0 to 100% in 10% intervals. The standardized literature screening followed the same protocol for each species: a Google Scholar search with the full species name in quotes was queried in ‘Publish or Perish’ v. 7 [53]. The first 20 publications, sorted by ‘rank', were checked for ecological information. In many cases, secondary literature based on the reference list of a given publication was searched. Additionally, we searched in all publications listed in the respective ‘species files' [38,39]. Literature data on species name synonyms, which were automatically retrieved from the Global Biodiversity Information Facility website (https://www.gbif.org) via the ‘rgbif' package [54], was also searched for in the same way. Literature screening was finished on 3 June 2020, resulting in 3380 ecological data entries (electronic supplementary material, tables S1–S5). General statements about the ecology of taxon levels higher than species (such as genus, subfamily or family) were not added to the databank because even congeners may differ in their ecology. Early general contentions on non-feeding Plecoptera [55–59] were not taken into account because they have been widely disproven in later studies (see electronic supplementary material, tables S2 and S3). If information on different nymphal instars was available, only data on the last instar was taken, because we were interested in the effect of only the last (metamorphic) moult on adaptive decoupling. (e) Phylogenetic supertree generation We digitized a Bayesian inference tree of Dermaptera based on five loci (18S rDNA, 28S rDNA, COI, Histone 3, and Tubulin Alpha I [60]) and a Bayesian inference tree of Plecoptera based on mitogenomes [61] using the ‘phylo.tracer' function of the ‘physketch' package v. 0.1 [62]. We manually added missing taxa to these phylogenies either by substituting closely related species or by using ‘bind.tip' in phytools. Subsequently, we used the ‘chronos' function in ‘ape' v. 5.3 [63] to fit chronograms based on the branch lengths of the original phylogenies and the median node ages of the most recent common ancestors (MRCAs), according to Misof et al. [30], of Dermaptera (79.39 Ma) and Plecoptera (167.41 Ma). These phylogenies were then combined to a supertree using the median node age of the MRCA (302.05 Ma) [30] of Dermaptera and Plecoptera. Numerical imprecisions were eliminated by the ‘force.ultrametric' function in phytools. The resulting ultrametric supertree (electronic supplementary material, figure S5) was pruned to only contain the taxa present in our analysis. (f) Models of shape evolution Some of the functions used in this study require that the mode of shape evolution follows a Brownian motion model. To check this, we ran ‘fitContinuous’ in ‘geiger’ v. 2.0.6.2 [64] and tested which of the following models most closely describes the shape evolution in our dataset: ‘BM’, Brownian motion [65]; ‘OU’, Ornstein-Uhlenbeck [66]; ‘EB’, early burst [67], also known as ‘ACDC’ (accelerated/decelerated) [68]; ‘lambda’, phylogeny predicts covariance of shape among species [69]; and ‘white noise’, non-phylogenetic white-noise [64]. Using ‘aic.w’ in phytools, we compared the sample-size corrected Akaike information criterion (AICc) of the fitted models (electronic supplementary material, table S8) and found that the Brownian motion model had the best fit (AICc = −827.6, AICc-weight = 0.526). We thus concluded that the assumption of a Brownian motion model of shape evolution is sufficiently met by our dataset. (g) Allometric and phylogenetic signal We analysed the effect of size on head shape in earwigs and stoneflies by performing a regression of the Procrustes-aligned shape data against log-centroid size using ‘procD.lm' in geomorph with 10 000 permutations. We tested if earwigs and stoneflies share common allometries by comparing a linear model with the null hypothesis of unique allometric slopes (shape ∼ log(centroid size) × order) versus a linear model with the null hypothesis of a common allometric slope (shape ∼ log(centroid size) + order). The same test was repeated for models with unique or common allometric slopes of superfamilies. All model fits were compared with an ANOVA. In downstream analyses, we accounted for the allometric effect using the residuals of a regression of shape on centroid size. The multivariate nature of the results of the principal coordinate analyses (PCoAs) of the ecological data (see below) did not allow for separate comparisons of allometric slopes of ecological groups. Phylogenetic signal in the Procrustes-aligned shape data was evaluated by calculating the Kmult statistic [70], a multivariate generalization of the K statistic [68], using ‘physignal' in geomorph with 10 000 iterations. (h) Integration of ecology and shape To test our hypotheses that adult head and mandible shape in Dermaptera and Plecoptera covaries to varying degrees with ecological characteristics of either nymphs and/or adults, we calculated the degree of integration of the Procrustes shape data with the extracted multivariate ecological traits. We first calculated the Bray-Curtis dissimilarity index [71] of all species using ‘vegdist’ in ‘vegan’ v. 2.5-6 [72]. We ran PCoAs on these dissimilarity matrices using ‘pcoa’ in ape (electronic supplementary material, figure S4). The integration between shape and the PCoA vectors of the multivariate traits was separately calculated for each trait using ‘phylo.integration’ in ‘geomorph’ with 10 000 iterations (table 1; electronic supplementary material, table S9). This function identifies integration of multivariate traits while accounting for the phylogenetic non-independence of taxa by using an evolutionary covariance matrix under a Brownian motion model of evolution in the partial least squares (PLS) assessment of trait covariation [73]. Additionally, two-block PLS analyses were run to test for the integration of ecology and shape without considering phylogeny using ‘two.b.pls’ in geomorph (electronic supplementary material, table S9). Since we had to filter our data before the analyses according to species coverage for the ecological data, species numbers were different in every analysis for each ecological character. Additionally, the number of Procrustes coordinates in the shape data of head capsules versus mandibles on the one hand, and the number of PCos in the ecological data on the other hand, varied considerably. Both factors influence the results of the PLS correlation coefficient, because this coefficient is dependent on the number of specimens and trait characters [74]. In order to be able to compare the explanatory values of the phylogenetic PLS analyses, we used ‘compare.pls’ in geomorph which calculates the effect sizes as z-scores. All above described analyses were carried out for the whole dataset (Dermaptera & Plecoptera) as well as for dermapteran and plecopteran subsets independently (electronic supplementary material, table S9). Table 1. Multivariate integrations of head shape (left) and mandible shape (right) of earwigs and stoneflies with ecological covariates expressed as effect sizes (z-scores) of phylogenetic partial least square results. Non-significant interactions left blank. See electronic supplementary material, table S9 for more statistical details and all test results. hydrodyn. p., hydrodynamic pressure; microh., microhabitat. All further downstream analyses were performed (i) not accounting for allometry or phylogeny, (ii) accounting for allometry only, (iii) accounting for phylogeny only and (iv) accounting for both allometry and phylogeny (electronic supplementary material, table S9). We only report the results of the phylogeny-corrected analyses in the main text. (i) Morphological disparity Differences in morphospace occupation between Dermaptera and Plecoptera were estimated by running separate Procrustes-alignments of the shape data on the order-level. The morphological disparity for each of the order-subsets was calculated with ‘morphol.disparity' in geomorph. Additionally, we compared the adult head shape disparity of perloidean stoneflies versus all other stoneflies and earwigs and non-perloidean stoneflies versus earwigs. 3. Results and discussion (a) Adult head shape evolution is allometrically and phylogenetically structured Head size (measured as log (centroid size)) has a significant but weak influence on head shape (Procrustes ANOVA, R² = 0.044, p = 1e−4, n = 219). The ANOVA comparing the log-transformed linear relationships of head size and shape in earwigs and stoneflies yielded a statistically significant difference of allometric slopes of the two orders (ANOVA, R² = 0.026, p = 1e−4, n = 219). However, the low explanatory value of the model indicates low biological meaningfulness of this slope heterogeneity. ANOVA analyses on the allometric slopes of superfamily level taxonomic groups yielded similar results (R² = 0.054, p = 4e−4). The inspection of the prediction lines [75] and regression scores [76] of the allometric analyses (electronic supplementary material, figure S3) showed that allometric slopes do not diverge greatly from each other in the different taxonomic groups. Given these results, we treated the whole dataset as if all taxa had a common allometric slope when correcting for allometric effects. The dataset also contains significant phylogenetic signal (Kmult = 0.22, p = 1e−4, n = 219). Kmult values below 1 indicate that the head shapes of closely related species are less similar to each other than expected under a Brownian motion model of evolution and could be explained by adaptive components in their evolution that do not follow the underlying phylogeny [68,70]. The results of the analyses with and without allometric corrections differ only slightly in their explanatory values and z-scores but not in the general pattern of influences of ecological factors on adult head and mandible shape. Not correcting for phylogeny, however, resulted in much higher correlations of all ecological factors with shape (electronic supplementary material, table S9), indicating that closely related taxa generally share more common ecologies. (b) Convergent evolution into unique ecological niches The first principal component (PC1) axis of an allometry- and phylogeny-corrected PC-analysis of head shape accounts for 20.68% of the variation and, generally, separates short headed taxa (left side, figure 2) from taxa with more elongated head shapes (right side, figure 2). PC2 (18.74%) separates earwigs from stoneflies. Within earwigs, the phylomorphospace spanned by the first three PCs mainly separates the families Hemimeridae and Arixeniidae from all other lineages. These two families share a phylomorphospace region at the lower ends of PC1, PC2 and PC3 (14.5%). They do not form a monophyletic clade [60], but convergently evolved epizoic lifestyles and live, at least partly, in the fur of mammals [77–80]. All other earwig (super)families cluster near the centre of PCs1–3 (figure 2), a relative shape homogeneity which could be explained by the uniform feeding habits of these groups (see electronic supplementary material, tables S2 and S3 for taxon specific feeding habit extractions). Figure 2. Head capsule phylomorphospace of adult earwigs and stoneflies. Principal component (PC) 1 and PC2 (upper part) and PC1 and PC3 (lower part) account for approximately 54% of the shape variation. Point colours represent superfamily level memberships, smoothed convex hulls show order-level memberships. Schematic drawings of the dorsal (including tentorial structures in grey) and lateral head shapes are based on µCT scans. Drawings are shown for selected species (black bordered points) at the edges of the phylomorphospace spanned by the first three PCs. Schematics not to scale. See electronic supplementary material, figure S3 for an overview of the first 8 PCs. (Online version in colour.) Within the stonefly morphospace, PC1 mainly describes differences between the species with predatory nymphs (most of them belonging to Perloidea and Eusthenioidea) and those with detritivoric nymphs (figures 2 and 3). Both Eusthenioidea and Perloidea occupy a similar phylomorphospace region at the upper region of PC1 and lower region of PC2, despite the fact that they are geographically and phylogenetically separated: Eusthenioidea, belonging to the suborder Antarctoperlaria, are restricted to the Southern Hemisphere, while Perloidea, belonging to the suborder Arctoperlaria, are, with a few secondary exceptions, restricted to the Northern Hemisphere [81]. Many lineages of Perloidea [82] and some lineages of Eusthenioidea [83,84] are able to fully develop their eggs within the last nymphal instar already. This results in a drastically reduced time to oviposition of a few days compared to many days or weeks in other species [82], and possibly lowered selection pressures on adult head shapes that are related to a regular uptake of nutrient-rich food [2]. Figure 3. Morpho-ecological landscape illustrating the relationship of adult head shape with the feeding ecologies of nymphs (upper tiles) and adults (lower tiles). All tiles are overlaid with the phylomorphospace explained by the first two principal components of adult head shape in earwigs (dark green) and stoneflies (bright green). PC1 of nymphal food preference (a) mainly represents detritivoric (blue) versus predatory (red) habits, while PCo1 of adult feeding ecology (b) mainly represents herbivorous (blue) versus predatory (red) habits (see electronic supplementary material, figure S4 for PCoA biplots). Tiles (c–k) show density estimations of species with mainly non-feeding, (c,d), predatory (e,f), herbivorous (g,h) and (j,k) detritivoric nymphs and adults, respectively, to illustrate shifts of feeding habits across metamorphosis. Notable shifts in feeding habits are indicated by dashed arrows for clarity. (Online version in colour.) (c) Stronger ontogenetic niche shifts may result in lower adaptive decoupling in hemimetabolan insects Multivariate, phylogenetically corrected integration tests of the Procrustes-aligned shape data against the results of the principal coordinate analyses (PCoA) of ecological covariates show that the head shape of adult stoneflies is most strongly influenced by the feeding habits of their nymphs (R² = 0.79, z = 3.91, p = 1e−4, n = 46), and not by feeding habits of the adults themselves (p = 0.41, n = 37; table 1; electronic supplementary material, table S9). Nymphal feeding habits are also the only significant covariate of adult mandible shape in our analysis (R² = 0.73, z = 3.38, p = 1e−4, n = 47). This high correlation of nymphal ecology and adult morphology indicates a low degree of adaptive decoupling: the last moult between the nymphal and adult stage in stoneflies does not seem to facilitate a disruption of trait coupling. Despite the strong ontogenetic niche shift resulting from the amphibiotic life style of stoneflies (figure 1; electronic supplementary material, tables S4 and S5), adult head and mandible morphology could not evolve independently from nymphal ecology. In addition to the habitat shift, changes in feeding ecology also occur across stonefly metamorphosis: predatory stonefly nymphs mostly metamorphose into liquid-feeding or non-feeding adults, while detritivoric stonefly nymphs mostly become herbivorous adults (figure 3; electronic supplementary material, tables S2 and S3). Selection pressures that act on the nymphal stage of stoneflies therefore seem to outweigh those acting on their adult stage, so that adult shape evolution is mainly driven by nymphal selection pressures. Earwigs, on the other hand, show a relatively high correlation of both nymphal and adult feeding preference with adult head shape (R² = 0.73, z = 2.89, p = 9e−4, n = 35; table 1). They do not undergo major ontogenetic niche shifts when reaching adulthood, so their shape evolution may be equally adapted to the largely congruent selection pressures of both life phases. However, the degree of adaptive decoupling that could possibly be realized in earwigs may be as limited as in stoneflies since they also enter the adult phase with a single moult. Low degrees of adaptive decoupling in earwigs could therefore be either adaptively beneficial due to overlapping selection pressures, or the result of the limited possibility of metamorphic change across the final moult, or a combination of both effects. (d) Stronger ontogenetic niche shifts are accompanied by higher adult shape disparity If patterns of variation are decoupled across metamorphosis, ecologically divergent life phases can effectively evolve towards unique selection pressures. Our data show that adult head shape is, however, not adaptively decoupled from nymphal ecology in earwigs and stoneflies. Instead, nymphal ecology drives adult head shape evolution (table 1). We hypothesized that this trait correlation would constrain the diversification of adult head shape, because head shape would not be free to evolve towards the unique selection pressures of the adult stage. Especially for stoneflies with strong ontogenetic niche shifts but high trait correlations, we expected that adult head shapes show a low shape disparity. Contrary to our expectations, however, adult head shape disparity in the amphibiotic stoneflies is significantly higher than in the fully terrestrial earwigs (Procrustes variance = 0.018 versus 0.013; p = 3e−3). More detailed analyses revealed that highest levels of adult head shape disparity within stoneflies are concentrated at the superfamily Perloidea: when this group is excluded from the analysis, disparity within earwigs and stoneflies does not significantly differ from each other (0.013 versus 0.016; p = 0.26). Indeed, Perloidea alone show a significantly higher head shape disparity (0.026) than non-perloidean stoneflies (p = 2e−3) and earwigs (p = 3e−4). This observation could be explained by the fact that perloidean stoneflies do not rely on frequent feeding of hard food in the adult stage in order to sustain egg development, because nutrients have been already stored by the predatory nymphs [37,82]. The adaptive importance of feeding-related head structures in adult Perloidea might therefore be lowered, and the relative weight of selection pressures on the nymphal stage may be increased. This could have facilitated the evolution into new ecological niches in Perloidea, because adaptive conflicts between the life phases are reduced, possibly resulting in the observed higher head shape disparity of this group. Consequently, low degrees of adaptive decoupling can still facilitate increased phenotypic disparity. 4. Conclusion About 80% of all animals show a complex life cycle with distinct life phases. Such life phases are characterized by phenotypic adaptations to their phase-specific ecological niches. To avoid adaptive conflicts, traits can be decoupled between life phases, and metamorphosis is thought to aid in the breakup of trait correlations. We showed that adult head shape evolution in earwigs and stoneflies, two closely related hemimetabolous insect taxa, is not decoupled from juvenile ecology despite sometimes strong ontogenetic niche shifts. We therefore conclude that the hemimetabolan metamorphosis in earwigs and stoneflies does not facilitate a disruption of trait couplings, resulting in a constrained phenotypic evolution of the adult phase. Additionally, stronger food-related ontogenetic niche shifts within stoneflies have resulted in higher shape disparity in the adults of some stonefly families, possibly because of the liberation of the adult stage from food-related functions. Supplementary Material Click here for additional data file. Acknowledgements B. Price (NHMUK), H. Zettel, D. Zimmermann (both NHMV), J. Deckert (MfN), R. Peters (ZFMK) and L. Hendrich (ZSM) kindly organized access to and loan of specimens. O. Béthoux, A. Caires, S. Mtow, P. Pessacq and P. Zwick generously provided specimens from private collections. We acknowledge the KIT light source for provision of instruments at their beamlines and we would like to thank the Institute for Beam Physics and Technology (IBPT) for the operation of the storage ring, the Karlsruhe Research Accelerator (KARA). M. Dabringhaus, A. Huenebeck, U. Kolesnikov, M. Prelle, L. Scheele, I. Javed and C. Voss are thanked for their help with µCT data processing. The infrastructural support of the Büschges group (University of Cologne) during early parts of the study is highly appreciated. We are grateful for helpful comments by X. Belles (CSIC-Universitat Pompeu Fabra), J. Collet (Centre national de la recherche scientifique), S. Fellous (Université de Montpellier) and J. W. Truman (University of Washington) on an earlier version of this manuscript. We thank two anonymous reviewers for their helpful comments. Data accessibility All 219 cropped and down-sampled µCT scans used in this study have been deposited at Zenodo.org (https://doi.org/10.5281/zenodo.4280412). Full datasets are available upon request from the corresponding author. The R code used in this study is publicly available on GitHub (https://github.com/Peter-T-Ruehr/Adaptive_decoupling_insects). The data are provided in electronic supplementary material [85]. Authors' contributions P.T.R.: conceptualization, data curation, formal analysis, investigation, methodology, project administration, visualization, writing-original draft, writing-review and editing; T.v.d.K.: investigation, methodology, resources, software, writing-review and editing; T.F.: investigation, methodology, resources, software, writing-review and editing; J.U.H.: investigation, methodology, resources, software, writing-review and editing; F.W.: investigation, methodology, resources, software, writing-review and editing; E.B.: investigation, methodology, resources, software, writing-review and editing; C.E.: investigation, writing-review and editing; M.F.: investigation, writing-review and editing; T.B.: resources, writing-review and editing; A.B.: conceptualization, formal analysis, funding acquisition, investigation, project administration, resources, supervision, writing-review and editing. All authors gave final approval for publication and agreed to be held accountable for the work performed therein. Competing interests We declare we have no competing interests. Funding P.T.R., M.F. and A.B. were supported by the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation program (grant agreement no. 754290, ‘Mech-Evo-Insect’). A.B. and C.E. were supported by the Deutsche Forschungsgemeinschaft (DFG) under the Individual Research Grants program (grant agreement no. BL 1355/4-1). µCT-scanning was funded by the following, facility-specific grants awarded to A.B. Hereon at DESY: I-20170190, I-20170896, I-20190019, SLS: 20171469. ==== Refs References 1. Ebenman B. 1987 Niche differences between age classes and intraspecific competition in age-structured populations. J. Theor. Biol. 124 , 25-33. (10.1016/S0022-5193(87)80249-7) 2. Ebenman B. 1992 Evolution in organisms that change their niches during the life cycle. Am. Nat. 139 , 990-1021. (10.1086/285370) 3. Werner EE, Gilliam JF. 1984 The ontogenetic niche and species interactions in size-structured populations. Annu. Rev. Ecol. Syst. 15 , 393-425. (10.1146/annurev.es.15.110184.002141) 4. Moran NA. 1994 Adaptation and constraint in the complex life cycles of animals. Annu. Rev. Ecol. Syst. 25 , 573-600. (10.1146/annurev.es.25.110194.003041) 5. Collet J, Fellous S. 2019 Do traits separated by metamorphosis evolve independently? Proc. R. Soc. B 286 , 20190445. (10.1098/rspb.2019.0445) 6. Bishop CD et al . 2006 What is metamorphosis? Integr. Comp. Biol. 46 , 655-661. (10.1093/icb/icl004)21672776 7. Belles X. 2020 Insect metamorphosis: from natural history to regulation of development and evolution. London, UK: Academic Press. 8. Wald G. 1981 Metamorphosis: an overview. In Metamorphosis: a problem in developmental biology (eds LI Gilbert, E Frieden), pp. 1-39. Boston, MA: Berlin, Germany: Springer US. 9. Smith AB, Littlewood DTJ, Wray GA, Harvey PH, Leigh Brown AJ, Smith JM. 1995 Comparing patterns of evolution: larval and adult life history stages and ribosomal RNA of post-Palaeozoic. Phil. Trans. R. Soc. B 349 , 11-18. (10.1098/rstb.1995.0085) 10. Wray GA. 1995 Evolution of larvae and developmental modes. In Ecology of marine invertebrate larvae (ed LR McEdward), pp. 413-448. Boca Raton, FL: CRC Press. 11. Rolff J, Johnston PR, Reynolds S. 2019 Complete metamorphosis of insects. Phil. Trans. R. Soc. B 374 , 20190063. (10.1098/rstb.2019.0063)31438816 12. Truman JW. 2019 The evolution of insect metamorphosis. Curr. Biol. 29 , R1252-R1268. (10.1016/j.cub.2019.10.009)31794762 13. Wilbur HM. 1980 Complex life cycles. Annu. Rev. Ecol. Syst. 11 , 67-93. (10.1146/annurev.es.11.110180.000435) 14. Werner EE. 1988 Size, scaling, and the evolution of complex life cycles. In Size-structured populations (eds B Ebenman, L Persson), pp. 60-81. Berlin, Heidelberg: Springer. 15. Strathmann RR, Eernisse DJ. 1994 What molecular phylogenies tell us about the evolution of larval forms. Am. Zool. 34 , 502-512. (10.1093/icb/34.4.502) 16. Katz HR, Hale ME. 2016 A large-scale pattern of ontogenetic shape change in ray-finned fishes. PLoS ONE 11 , e0150841. (10.1371/journal.pone.0150841)26943126 17. Sherratt E, Vidal-García M, Anstis M, Keogh JS. 2017 Adult frogs and tadpoles have different macroevolutionary patterns across the Australian continent. Nat. Ecol. Evol. 1 , 1385-1391. (10.1038/s41559-017-0268-6)29046549 18. Wollenberg Valero KC et al . 2017 Transcriptomic and macroevolutionary evidence for phenotypic uncoupling between frog life history phases. Nat. Commun. 8 , 15213. (10.1038/ncomms15213)28504275 19. Phung TX, Nascimento JCS, Novarro AJ, Wiens JJ. 2020 Correlated and decoupled evolution of adult and larval body size in frogs. Proc. R. Soc. B 287 , 20201474. (10.1098/rspb.2020.1474) 20. Fabre A-C et al . 2020 Metamorphosis shapes cranial diversity and rate of evolution in salamanders. Nat. Ecol. Evol. 4 , 1129-1140. (10.1038/s41559-020-1225-3)32572219 21. Pereira HS, Sokolowski MB. 1993 Mutations in the larval foraging gene affect adult locomotory behavior after feeding in Drosophila melanogaster. Proc. Natl Acad. Sci. USA 90 , 5044-5046. (10.1073/pnas.90.11.5044)8506349 22. Osborne KA, Robichon A, Burgess E, Butland S, Shaw RA, Coulthard A, Pereira HS, Greenspan RJ, Sokolowski MB. 1997 Natural behavior polymorphism due to a cgmp-dependent protein kinase of Drosophila. Science 277 , 834-836. (10.1126/science.277.5327.834)9242616 23. Fellous S, Lazzaro BP. 2011 Potential for evolutionary coupling and decoupling of larval and adult immune gene expression. Mol. Ecol. 20 , 1558-1567. (10.1111/j.1365-294X.2011.05006.x)21299661 24. Edelsparre AH, Vesterberg A, Lim JH, Anwari M, Fitzpatrick MJ. 2014 Alleles underlying larval foraging behaviour influence adult dispersal in nature. Ecol. Lett. 17 , 333-339. (10.1111/ele.12234)24386971 25. Yang AS. 2001 Modularity, evolvability, and adaptive radiations: a comparison of the hemi- and holometabolous insects. Evol. Dev. 3 , 59-72. (10.1046/j.1525-142x.2001.003002059.x)11341675 26. Rainford JL, Hofreiter M, Nicholson DB, Mayhew PJ. 2014 Phylogenetic distribution of extant richness suggests metamorphosis is a key innovation driving diversification in insects. PLoS ONE 9 , e109085. (10.1371/journal.pone.0109085)25275450 27. Steel EA, Kennedy MC, Cunningham PG, Stanovick JS. 2013 Applied statistics in ecology: common pitfalls and simple solutions. Ecosphere 4 , art115. (10.1890/ES13-00160.1) 28. Revell LJ. 2009 Size-correction and principal components for interspecific comparative studies. Evolution 63 , 3258-3268. (10.1111/j.1558-5646.2009.00804.x)19663993 29. Polly PD, Lawing AM, Fabre A-C, Goswami A. 2013 Phylogenetic principal components analysis and geometric morphometrics. Hystrix Ital. J. Mammal. 24 , 33-41. (10.4404/hystrix-24.1-6383) 30. Misof B et al . 2014 Phylogenomics resolves the timing and pattern of insect evolution. Science 346 , 763-767. (10.1126/science.1257570)25378627 31. Wipfler B et al . 2019 Evolutionary history of Polyneoptera and its implications for our understanding of early winged insects. Proc. Natl Acad. Sci. USA 116 , 3024-3029. (10.1073/pnas.1817794116)30642969 32. Zhang Z-Q. 2011 Phylum Arthropoda von Siebold, 1848 In: Zhang, Z.-Q. (Ed.) Animal biodiversity: an outline of higher-level classification and survey of taxonomic richness. Zootaxa 3148 , 99-103. (10.11646/zootaxa.3148.1.14) 33. Günther K, Herter K. 1974 Dermaptera (Ohrwürmer). In Handbuch der zoologie. IV.: Arthropoda. 2. Hälfte: insecta (eds JG Helmcke, D Starck, H Wermuth), pp. 1-158. Berlin, Germany: Walter de Gruyter & Co. 34. Zwick P. 1980 Plecoptera (Steinfliegen). In Handbuch der zoologie. IV.: Arthropoda. 2. Hälfte: insecta (eds JG Helmcke, D Starck, H Wermuth), pp. 1-158. Berlin, Germany: Walter de Gruyter & Co. 35. Illies J. 1960 Die erste auch im Larvenstadium terrestrische Plecoptere (zugleich ein Beitrag zur Kenntnis der Unterfamilie Andiperlinae). Mitteilungen Schweiz. Entomol. Ges. 33 , 161-168. 36. McLellan ID. 1979 New Zealand terrestrial stoneflies and some ideas on speciation. Gewässer Abwässer 64 , 56-59. 37. Tierno de Figueroa JM, López-Rodríguez MJ. 2019 Trophic ecology of Plecoptera (Insecta): a review. Eur. Zool. J. 86 , 79-102. (10.1080/24750263.2019.1592251) 38. Hopkins H, Maehr MD, Haas F, Deem LS. 2020 Dermaptera. See http://Dermaptera.SpeciesFile.org. 39. DeWalt RE, Maehr MD, Neu-Becker U, Stueber G. 2020 Plecoptera. See http://Plecoptera.SpeciesFile.org. 40. Lytaev P et al . 2014 Characterization of the CCD and CMOS cameras for grating-based phase-contrast tomography. Developments in X-ray Tomography 9 , 921218. (10.1117/12.2061389) 41. Moosmann J, Ershov A, Weinhardt V, Baumbach T, Prasad MS, LaBonne C, Xiao X, Kashef J, Hofmann R. 2014 Time-lapse X-ray phase-contrast microtomography for in vivo imaging and analysis of morphogenesis. Nat. Protoc. 9 , 294-304. (10.1038/nprot.2014.033)24407356 42. Wilde F et al . 2016 Micro-CT at the imaging beamline P05 at PETRA III. AIP Conf. Proc. 1741 , 030035. (10.1063/1.4952858) 43. Stampanoni M et al . 2006 Trends in synchrotron-based tomographic imaging: the SLS experience. SPIE Proc. 6318 , 63180M. (10.1117/12.679497) 44. Schindelin J et al . 2012 Fiji: an open-source platform for biological-image analysis. Nat. Methods 9 , 676-682. (10.1038/nmeth.2019)22743772 45. Bookstein FL. 1991 Morphometric tools for landmark data: geometry and biology. Cambridge, UK: Cambridge University Press. 46. R Development Core Team R. 2021 R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. See http://www.R-project.org. 47. Adams DC, Collyer ML, Kaliontzopoulou A. 2020 Geomorph: software for geometric morphometric analyses. See https://cran.r-project.org/package=geomorph. 48. Gower JC. 1975 Generalized procrustes analysis. Psychometrika 40 , 33-51. (10.1007/BF02291478) 49. Rohlf FJ, Slice D. 1990 Extensions of the procrustes method for the optimal superimposition of landmarks. Syst. Zool. 39 , 40-59. (10.2307/2992207) 50. Adams DC, Collyer ML. 2018 Multivariate phylogenetic comparative methods: evaluations, comparisons, and recommendations. Syst. Biol. 67 , 14-31. (10.1093/sysbio/syx055)28633306 51. Revell LJ. 2012 phytools: an R package for phylogenetic comparative biology (and other things). Methods Ecol. Evol. 3 , 217-223. (10.1111/j.2041-210X.2011.00169.x) 52. Wilman H, Belmaker J, Simpson J, Rosa Cdl, Rivadeneira MM, Jetz W. 2014 EltonTraits 1.0: Species-level foraging attributes of the world's birds and mammals. Ecology 95 , 2027-2027. (10.1890/13-1917.1) 53. Harzing AW. 2007 Publish or Perish. See https://harzing.com/resources/publish-or-perish. 54. Chamberlain S, Barve V, McGlinn D, Oldoni D, Geffert L, Ram K. 2020 rgbif: Interface to the Global Biodiversity Information Facility API. See https://CRAN.R-project.org/package=rgbif. 55. Neeracher F. 1910 Die Insektenfauna des Rheins und seiner Zuflüsse bei Basel: Faunistik, Biologie, Systematik. Revue Suisse Zool. 18 , 497-590. (10.5962/bhl.part.16672) 56. Smith JB. 1910 Order Plecoptera. In Annual report of the New Jersey state museum including a report of the insects of New Jersey, pp. 39-41. Trenton, NJ: MacCrellish & Quigley. 57. Rousseau E. 1921 Les larves et nymphes aquatiques des insectes d'Europe: morphologie, biologie, systématique. Brussels, Belgium: J. Lebègue. 58. Mertens H. 1923 Biologische und morphologische Untersuchungen an Plekopteren. Arch. Naturgeschichte. 89 , 1-38. (10.5962/bhl.part.28225) 59. Kühtreiber J. 1934 Die Plekopterenfauna Nordtirols. Berichte Naturwissenschaftlich-Med. Ver. Innsbr. 44 , 1-219. 60. Naegle MA, Mugleston JD, Bybee SM, Whiting MF. 2016 Reassessing the phylogenetic position of the epizoic earwigs (Insecta: Dermaptera). Mol. Phylogenet. Evol. 100 , 382-390. (10.1016/j.ympev.2016.03.012)27033951 61. Ding S, Li W, Wang Y, Cameron SL, Murányi D, Yang D. 2019 The phylogeny and evolutionary timescale of stoneflies (Insecta: Plecoptera) inferred from mitochondrial genomes. Mol. Phylogenet. Evol. 135 , 123-135. (10.1016/j.ympev.2019.03.005)30876966 62. Revell LJ. 2017 physketch: drawing phylogenetic objects free-hand. See http://github.com/liamrevell/physketch. 63. Paradis E, Schliep K. 2019 ape 5.0: an environment for modern phylogenetics and evolutionary analyses in R. Bioinformatics 35 , 526-528. (10.1093/bioinformatics/bty633) 64. Harmon LJ, Weir JT, Brock CD, Glor RE, Challenger W. 2008 GEIGER: investigating evolutionary radiations. Bioinformatics 24 , 129-131. (10.1093/bioinformatics/btm538)18006550 65. Felsenstein J. 1973 Maximum-likelihood estimation of evolutionary trees from continuous characters. Am. J. Hum. Genet. 25 , 471-492.4741844 66. Butler MA, King AA. 2004 Phylogenetic comparative analysis: a modeling approach for adaptive evolution. Am. Nat. 164 , 683-695. (10.1086/426002)29641928 67. Harmon LJ et al . 2010 Early bursts of body size and shape evolution are rare in comparative data. Evolution 64 , 2385-2396. (10.1111/j.1558-5646.2010.01025.x)20455932 68. Blomberg SP, Garland T, Ives AR. 2003 Testing for phylogenetic signal in comparative data: behavioral traits are more labile. Evolution 57 , 717-745. (10.1111/j.0014-3820.2003.tb00285.x)12778543 69. Pagel M. 1999 Inferring the historical patterns of biological evolution. Nature 401 , 877-884. (10.1038/44766)10553904 70. Adams DC. 2014 A generalized K statistic for estimating phylogenetic signal from shape and other high-dimensional multivariate data. Syst. Biol. 63 , 685-697. (10.1093/sysbio/syu030)24789073 71. Bray JR, Curtis JT. 1957 An ordination of the upland forest communities of Southern Wisconsin. Ecol. Monogr. 27 , 325-349. (10.2307/1942268) 72. Oksanen J et al . 2019 vegan: community ecology package. See https://CRAN.R-project.org/package=vegan. 73. Adams DC, Felice RN. 2014 Assessing trait covariation and morphological integration on phylogenies using evolutionary covariance matrices. PLoS ONE 9 , e94335. (10.1371/journal.pone.0094335)24728003 74. Adams DC, Collyer ML. 2016 On the comparison of the strength of morphological integration across morphometric datasets. Evolution 70 , 2623-2631. (10.1111/evo.13045)27592864 75. Adams DC, Nistri A. 2010 Ontogenetic convergence and evolution of foot morphology in European cave salamanders (Family: Plethodontidae). BMC Evol. Biol. 10 , 216. (10.1186/1471-2148-10-216)20637087 76. Drake AG, Klingenberg CP. 2008 The pace of morphological change: historical transformation of skull shape in St Bernard dogs. Proc. R. Soc. B 275 , 71-76. (10.1098/rspb.2007.1169) 77. Vosseler J. 1906 Einiges über Hemimerus und sein Wirtstier. Zool. Anz. 31 , 436-450. 78. Popham EJ. 1962 The anatomy related to the feeding habits of Arixenia and Hemimerus (Dermaptera). Proc. Zool. Soc. Lond. 139 , 429-450. (10.1111/j.1469-7998.1962.tb01838.x) 79. Burr M, Jordan K. 1912 On Arixenia Burr, a suborder of Dermaptera. In 2nd Int. Congress of Entomology, pp. 398-421. Aylesbury, UK: Hazell, Watson & Viney. 80. Marshall AG. 1977 Interrelationships between Arixenia esau (Dermaptera) and molossid bats and their ectoparasites in Malaysia. Ecol. Entomol. 2 , 285-291. (10.1111/j.1365-2311.1977.tb00893.x) 81. Zwick P. 2000 Phylogenetic system and zoogeography of the Plecoptera. Annu. Rev. Entomol. 45 , 709-746. (10.1146/annurev.ento.45.1.709)10761594 82. Hynes HBN. 1976 Biology of Plecoptera. Annu. Rev. Entomol. 21 , 135-153. (10.1146/annurev.en.21.010176.001031) 83. Helson GAH. 1934 The bionomics and anatomy of Stenoperla prasina (Newman). Trans. Proc. N. Z. Inst. 64 , 214-248. 84. Zwick P. 1979 Revision of the stonefly family Eustheniidae (Plecoptera), with emphasis on the fauna of the Australian region. Aquatic Insects 1 , 17-50. (10.1080/01650427909360975) 85. Rühr PT et al . 2021 Juvenile ecology drives adult morphology in two insect orders. FigShare.
PMC008xxxxxx/PMC8206706.txt
==== Front Proc Biol Sci Proc Biol Sci RSPB royprsb Proceedings of the Royal Society B: Biological Sciences 0962-8452 1471-2954 The Royal Society 34130504 10.1098/rspb.2021.0940 rspb20210940 10014470200Evolution Research Articles The evolution of trait correlations constrains phenotypic adaptation to high CO2 in a eukaryotic alga The evolution of trait correlations constrains phenotypic adaptation to high CO2 in a eukaryotic alga http://orcid.org/0000-0002-7408-9461 Walworth Nathan G. Conceptualization Data curation Formal analysis Investigation Methodology Project administration Software Validation Visualization Writing-original draft Writing-review & editing 1 Hinners Jana Conceptualization Investigation Methodology Writing-review & editing 2 Argyle Phoebe A. Investigation Methodology Writing-review & editing 3 http://orcid.org/0000-0002-7819-5851 Leles Suzana G. Data curation Formal analysis Investigation Methodology Visualization Writing-review & editing 1 Doblin Martina A. Conceptualization Funding acquisition Investigation Methodology Writing-review & editing 3 http://orcid.org/0000-0003-3856-4285 Collins Sinéad Conceptualization Data curation Formal analysis Funding acquisition Investigation Methodology Project administration Resources Supervision Validation Writing-review & editing 2 http://orcid.org/0000-0002-4963-0535 Levine Naomi M. Conceptualization Data curation Formal analysis Funding acquisition Investigation Methodology Project administration Resources Software Supervision Validation Writing-original draft Writing-review & editing [email protected] 1 1 Department of Biological Sciences, University of Southern California, Los Angeles, CA 90089-0371, USA 2 Institute of Evolutionary Biology, University of Edinburgh, Edinburgh EH9 3FL, UK 3 Climate Change Cluster, University of Technology Sydney, Sydney, New South Wales 2007, Australia Electronic supplementary material is available online at https://doi.org/10.6084/m9.figshare.c.5448710. 30 6 2021 June 30, 2021 16 6 2021 June 16, 2021 16 6 2021 June 16, 2021 288 1953 2021094022 4 2021 April 22, 2021 17 5 2021 May 17, 2021 © 2021 The Authors. 2021 https://creativecommons.org/licenses/by/4.0/ Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited. Microbes form the base of food webs and drive biogeochemical cycling. Predicting the effects of microbial evolution on global elemental cycles remains a significant challenge due to the sheer number of interacting environmental and trait combinations. Here, we present an approach for integrating multivariate trait data into a predictive model of trait evolution. We investigated the outcome of thousands of possible adaptive walks parameterized using empirical evolution data from the alga Chlamydomonas exposed to high CO2. We found that the direction of historical bias (existing trait correlations) influenced both the rate of adaptation and the evolved phenotypes (trait combinations). Critically, we use fitness landscapes derived directly from empirical trait values to capture known evolutionary phenomena. This work demonstrates that ecological models need to represent both changes in traits and changes in the correlation between traits in order to accurately capture phytoplankton evolution and predict future shifts in elemental cycling. microbial evolution , trait correlations , trait adaptation , phytoplankton , biogeochemistry , principal component analyses Simons Foundation http://dx.doi.org/10.13039/100000893 509727 Gordon and Betty Moore Foundation http://dx.doi.org/10.13039/100000936 MMI 7397 cover-dateJune 30, 2021 ==== Body pmc1. Introduction Microbes play a critical role in regulating biogeochemistry and the global climate. In recent years, there has been a significant increase in global change studies examining the role of microbial evolution in shaping future biogeochemical cycles. This work has helped to more explicitly integrate the fields of evolution and microbial ecology, resulting in both long-term experimental evolution studies with ecologically important microbes and, to a limited extent, the incorporation of adaptation into ecological and ocean circulation models [1–13]. These studies are just the first step in tackling the immensely complex challenge of microbial evolution and its influence on global biogeochemistry. We still have only a limited understanding of how microbial communities will respond to multi-stressor and fluctuating environmental change. Additionally, the sheer number of interacting environmental and trait combinations exceeds our experimental ability to robustly quantify these responses [14,15]. Hence, experimental and theoretical methods to reduce dimensionality and extract broad evolutionary patterns across traits and taxa are critical for creating a framework that can both help guide experiments and make more accurate future predictions [5]. Seminal research in quantitative genetics has investigated the impact of trait variation, genotypic variability, inheritance, epistasis and environmental variability on adaptive walks using multivariate and eigenvector methods on theoretical populations experiencing environmental change [16–20]. These studies have broadly found that an evolving population may be able to access only a subset of phenotypes. Other theoretical approaches emphasized the role of de novo mutations in a fitness landscape without accounting for standing genetic variation [21,22]. While these studies have generated evolutionary theory [22,23] relevant for marine microbes, the gulf between evolution literature and biological oceanography has widened such that current state-of-the-art evolutionary models are not suited to interface with biogeochemical ODE models [24]. Empirical data have also demonstrated that biological systems produce certain phenotypic variants more readily than others in response to a perturbation (mutation or environmental change) due to the inherent structure, composition and evolutionary history of a population [25,26]. These findings contrast with the long-held assumption of isotropic (i.e. equal) variation [27] and have revealed instead that only a limited part of multivariate phenotypic space (i.e. only certain phenotypes) can be accessed [18,20,23,25–28]. Figure 1 shows an illustrative example of how the accessibility of high-fitness phenotypes (i.e. peaks in a fitness landscape) differs depending on the starting location (ancestral phenotype) and the initial trajectory, which is dictated by a population's collective genetic architecture (figure 1, magenta and black circles and paths). In summary, a growing body of the literature has shown that genetic architecture influences how traits and trait correlations are impacted by environmental shifts and that these shifts produce non-random distributions of phenotypes [29–31]. These studies have substantial consequences for understanding future shifts in marine phytoplankton function and thus carbon cycling and global climate [32,33]. However, there have been few attempts to investigate the implications of this phenomenon for the evolution of trait and trait correlations of photosynthetic microbes [19,20]. Figure 1. Comparison of adaptive walks for two different phenotypes in a rugged fitness landscape with four high-fitness peaks. Two example starting phenotypes are represented as circles (magenta and black). The x- and y-axis represent dimensions in fitness space (e.g. different traits). The phenotypes start with low fitness (z-axis) and through trait and trait correlation changes move to higher fitness. The adaptive walk is governed by historical bias, or different initial trait architecture, that impacts the movement of the population within the landscape. As the adaptive walk proceeds, the population moves to the top of one of the fitness peaks. While there are several paths available to each starting phenotype (represented by magenta and black arrows), due to historical bias (trait correlation constraints), some paths can be inaccessible (denoted by the grey and purple arrows). Note that depending on historical bias and the phenotypic starting location, some high-fitness peaks are either more difficult to access or completely inaccessible. We introduce a framework for understanding how evolutionary trajectories of phenotypes (suites of traits) in microbial populations are impacted by the evolutionary relationships between traits (historical bias) and the evolution of trait relationships. Here, we define bias as standing trait correlations (i.e. relationships) in a population that are heritable and can impact fitness such that, over time, these correlations can constrain the direction of evolution [16]. Critically, this framework is based on measurable trait values and so does not require the quantification of the fitness impacts of mutations. This is important as most marine microbes are not genetically accessible and mapping the fitness effects of individual genetic variants is resource intensive. Our trait correlation evolution (TRACE) model is a first step towards investigating how correlated metabolic traits with clear biogeochemical significance may impact elemental cycling under environmental change (e.g. ocean acidification). Specifically, TRACE generates a trait-based fitness landscape based on empirical trait data thus facilitating future integration into global biogeochemical models in which marine microbes are represented using trait-based functional groups. Our results indicate that populations harbouring trait correlations oriented in (i.e. consistent with) the direction of selection may experience accelerated rates of adaptation. We suggest that incorporating these dynamics into biogeochemical models will be important for accurately predicting the impact of microbial adaptation on rates of biogeochemical cycling in the ocean. 2. Material and methods (a) Trait-based fitness landscape We generated a multivariate trait-based fitness landscape (trait-scape) using principal component analyses (PCA). Specifically, a trait-scape was created using four independent ecologically relevant traits (growth rate, respiration, cell size and daughter cell production) from five genotypes of high-CO2-adapted Chlamydomonas reinhardtii [3]. PCA was conducted on standardized evolved traits resulting in a total of 86% of variance explained on two axes, 54% and 32% on axes PC1 and PC2, respectively (figure 2a). To select a start and endpoint for the adaptive walk, ancestral populations were projected onto the evolved PC axes. A single genotype was selected for the modelling exercise where the observed ancestral trait values defined the start point of the adaptive walk (tan circle in figure 2b; electronic supplementary material, file S1) and the corresponding evolved population trait values defined the evolutionary endpoint (red circle in figure 2b; electronic supplementary material, file S1). Additional simulations were conducted in which the start point was varied. As the specific traits themselves were not relevant for this study, we will hereafter refer to them as traits 1–4. We refer the reader to [3] for an in-depth discussion of the evolution experiment. Figure 2. Principal component analysis (PCA) of ancestral and evolved trait values and the corresponding trait correlations. (a) Ancestral PCA calculated from the values of four ancestral traits across five genotypes where each point represents an independent biological population (i.e. culture) coloured by genotype. Percentages associated with PC1 and PC2 denote the amount of variance explained by each PC axis. (b) Evolved PCA plot calculated from the evolved values of the same four traits as in (a) across five genotypes. Filled circles represent the independent populations of the evolved genotypes. Open circles represent the corresponding populations of the ancestral genotypes (a) projected onto the evolved PC axes. The tan and red filled circles denote the start and end coordinate of the model, respectively. (c) Table of all six possible trait combinations and their values in their ancestral and evolved genotypes. An ‘X’ indicates a non-statistically significant trait correlation (p > 0.05). (b) Trait correlation evolution model dynamics The TRACE model framework simulates the adaptive walk of a microbial population across a trait landscape (trait-scape) towards a high-fitness area. TRACE was adapted from an individual based Fisher model of adaptation [1,34,35]. Each generation, each individual in the population experienced either a change in traits or changes in traits and trait correlations. Changes in trait values moved these individuals across the trait-scape while trait correlations constrained the direction of movement. Selection was imposed based on distance to the evolutionary endpoint in the trait-scape (described below), such that the population evolved towards the high-fitness region of the trait-scape. In essence, this framework selected for individuals with the smallest overall difference across all trait values from the empirically observed high-fitness phenotype. The weighting of the traits was derived from the observed evolved phenotypes evaluated using PCA, such that traits that were not observed to play an important role in fitness in the high-CO2 environment had low weight. It is important to note that the model did not directly select for trait correlations, but that specific correlations emerged in the population if they provided a fitness advantage in terms of trait dynamics. In the default model simulations (referred to as 90/10), in each generation 90% of individuals were randomly chosen to experience a random change in a trait value (while maintaining all existing trait correlations) while 10% experienced both a trait and trait correlation change. These changes were drawn from a Gaussian distribution (mean = 0 and standard deviation = 0.05) such that small changes were common and large changes were rare. For each individual, the randomly chosen trait change was added to the existing trait value. Following this first trait change, the remaining three trait values were updated using the trait correlations for that individual in that time step. For example, if trait 1 was initially changed, then traits 2, 3 and 4 would subsequently be updated by multiplying the new trait 1 value by the three trait correlations (1v2, 1v3 and 1v4). The remaining 10% of the population experienced both a trait and a trait correlation change. For each individual, one of the six trait correlations was randomly selected to change. Similar to the trait change, a random value was drawn from a Gaussian distribution with a mean of 0 and standard deviation of 0.05 and added to the existing correlation value. Next, one of the two traits associated with that correlation was randomly chosen and a trait change was selected in the same manner as above. Finally, the second trait tied to the correlation was updated using the new correlation and trait value (the other modelled trait values were not updated in this generation). These trait changes moved the individuals within the trait-scape. Selection was imposed using the Euclidean distance of each individual to the high-fitness area (zi). Specifically, the fitness of each for each individual (wi) was calculated as [1,34]2.1 wi=e−zi2/2. Individuals were then randomly sampled with replacement weighted by wi such that high-fitness individuals were more likely to persist to the next generation, as in [1,34]. (c) Model simulations The model trait-scape and the high-fitness area (red circle in figure 2b) were defined based on empirical data from the Chlamydomonas long-term evolution study [3]. The model was initialized with a population of 1000 individuals with either (i) all individuals containing the same trait values corresponding to the empirically observed ancestral trait values (tan circle in figure 2b), (ii) all individuals containing the same trait values corresponding to an alternative starting location or (iii) a mixed population with multiple different starting trait values based on the empirical data. To explore the impact of historical bias, trait correlations in the starting populations were initialized in three different ways: mixed mode, ancestral mode and evolved mode (described below). The modes differed only in the initial conditions not in model dynamics. Each model run was conducted for 2000 generations with 100 replicates each. All model parameters are given in electronic supplementary material, table S1. Previous work by us and others have demonstrated that adaptive outcomes using this framework are robust across a wide range of population sizes (electronic supplementary material, information S1) [1,34]. Several sensitivity studies were conducted to test model dynamics (see electronic supplementary material, information S2). (i) Mixed mode (no bias) simulations To test all possible adaptive walks between the ancestral start point and the evolutionary endpoint, simulations were conducted with no historical bias. Specifically, each individual was assigned correlation values randomly chosen from a standard uniform distribution over the interval (−1,1). Hence, every individual started with the same four trait values but random correlation values. See electronic supplementary material, information S3 for additional model detail. (ii) Ancestral mode simulations To test the effects of systematically adding ancestral bias, four ancestral sub-modes were conducted: A1, A2, A3 and A4. For sub-mode A1, random correlation values were generated as above for five of the six trait correlations, and one empirical ancestral correlation was used for all individuals. For A2, A3 and A4, all steps were the same except that two, three and four empirical ancestral correlations were added back, respectively. For simplicity, we chose to sequentially add in ancestral correlations based on the empirically calculated significant trait correlations from most significant to least significant (R2 = −0.89 to 0.54; figure 2c). (iii) Evolved mode simulations The same procedure described above for the ancestral mode was conducted for the evolved mode, but instead empirical evolved correlations were systematically added (modes E1–E4). (d) Hierarchical clustering Hierarchical clustering with multiscale bootstrap resampling (1000 replicates) on mean trait correlation values was conducted with R package pvclust [36] using Euclidean distance and the average (UPGMA) method. Principal component analysis using mean correlation values was conducted with R package vegan [37], and pvclust clusters with approximately unbiased (AU) p-values < 75% were projected onto the PC coordinate plane as convex hulls. 3. Results (a) Multi-dimensional trait evolution To understand the constraints on phenotype evolution, we must consider how multi-dimensional traits are altered by selection. Previous empirical work has shown that both trait values and the correlations between traits are altered as a population adapts to a new environment. For example, when five genotypes of the green alga C. reinhardtii were selected under high CO2 [3], all quantified traits changed to varying degrees depending on the genotype and the correlations between many traits changed with some traits becoming correlated (e.g. 1v2) while others becoming uncorrelated (e.g. 2v4; figure 2c). To understand how trait movement within the trait-scape can be constrained by historical bias (correlations between traits in the ancestral population), we developed a statistical model (TRACE) of multi-trait adaptation and investigated probabilities of different emergent evolutionary outcomes. We began with a ‘null hypothesis’ model in which there was no historical bias (mixed mode simulation) and then systematically added in bias to determine the impact on population level adaptation. An example of model dynamics from a single run in mixed mode is shown in figure 3a where a representative population consisting of a thousand individuals moved over time from the ancestral starting phenotype to the evolved high-fitness area (figure 3a). This resulted in an overall increase in fitness of the population over time (figure 3b). The underlying dynamics of the model (changes in trait values and trait correlation changes) for three representative traits are shown in figure 3c. Figure 3. Representative adaptive walk in the evolved trait-scape of a population of 1000 individuals. (a) Density plots of an adaptive walk of a single population for a single run (n = 1000 individuals) starting at the tan dot and ending at the red dot. Each plot represents a different point in time (i.e. generation) in the adaptive walk with the colour representing the density of individuals in a given area. (b) Fitness plot of the population across the entire adaptive walk with the coloured line and grey region representing the mean and standard deviation, respectively. Both the y-axis and colour indicate fitness. (c) Trait versus trait plots representing the same adaptive walk where lower fitness denotes the start of the walk and higher fitness denotes the end. As in (b), each point represents the mean standardized trait value of all individuals at a specific generation, or step. Though TRACE is a novel modelling framework for understanding trait adaptation, the model captures well known dynamics of adaptions. For example, fitness effects produced from changes at the beginning of the walk were significantly greater than at the end of the walk consistent with previous experimental and theoretical work [1,34,38–41] (figure 3). Although some individuals reached a maximum possible fitness of 1 (i.e. the evolutionary end coordinate), the mean population fitness consistently remained below 1 (figure 3b). This is due to the fact that the model is simultaneously optimizing multiple traits and their correlations, which inherently introduces small but significant amounts of persistent phenotypic variation. In addition, while the average movement of the population was fairly linear in PC space (figure 3a), the trajectory of trait changes was not linear (figure 3c). (b) Accessibility of cryptic phenotypes Four distinct population types (i.e. traits + trait correlations for the final population) emerged across the replicate runs (1000 individuals × 100 replicate runs = 100 000 individuals total). We term these population types (Pop-MA, Pop-MB, Pop-MC and Pop-MD) ‘cryptic phenotypes’ as they had statistically similar end mean fitness values and occurred in the same region of the trait-scape but had distinct trait correlations and, to some extent, distinct trait values (figure 4; electronic supplementary material, figure S1). In other words, these cryptic phenotypes represent four distinct evolutionary outcomes of different trait correlations + trait values that all converged on the single evolutionary end coordinate in the evolved trait-scape. For some correlations such as trait 1 versus trait 2 (1v2), little to no overlap was observed across each of the four population types (figure 4, row 5), while for others, several population types shared the same trait correlations. For example, individuals in Pop-MA and Pop-MD shared the same 1v3 correlation (figure 4, row 2, columns 1 and 4) but had completely different relationships for 1v4 (figure 4, row 3, columns 1 and 4). The observance of emergent, cryptic phenotypes with distinct underlying trait combinations are qualitatively in line with experimental evolution studies that observed convergent phenotypes derived from a mix of parallel and divergent mutational and transcriptional changes across replicate populations adapting to the same environment [7,42–45]. Figure 4. Four distinct, emergent population types (cryptic phenotypes) from model runs seeded with no bias. Each row displays one of the six possible trait correlations (2v4, 1v3, 1v4, 2v3, 1v2 and 3v4) with the distribution of the emergent trait correlation values for all individuals in all replicate runs (n = 100 000) shown in grey. Highlighted in colour in each subplot are the trait correlation values for the individuals belonging to each of the emergent population types (columns). Each population type has a clearly defined set of trait correlation values. For example, the 2v4 mean correlation for Pop-MA was 0.66 ± 0.22 while the 2v4 mean correlation for pop-MD was −0.28 ± 0.07. Pop-Ma and Pop-MD were the most accessible population types and so the trait correlation values associated with these population types (cryptic phenotypes) had the largest frequency (y-axis). When the model was run without the influence of trait correlations, only one phenotype emerged, as expected, but in contrast to the simulations where trait correlations were included (electronic supplementary material, figure S2 and information S2). Here individuals were unconstrained by bias and so were able to quickly move directly to the high-fitness area. This demonstrates that trait correlational constraints can produce different evolutionary strategies (i.e. emergent, cryptic phenotypes) and, if constraints are not present, individuals are able to explore phenotypic space more freely and arrive at the high-fitness phenotype more rapidly. The emergence of multiple high-fitness phenotypes (e.g. figure 5a) occupying a single high-fitness area in multivariate space demonstrates that, by using PCA for the trait-scape, our model captures a rugged trait-based fitness landscape with multiple high-fitness peaks (e.g. figure 1). Figure 5. Representative trait, population and fitness dynamics in the TRACE model. (a) Trait versus trait plot denoting the four distinct population types (i.e. cryptic phenotypes) that emerged from 100 replicate model runs in mixed mode (i.e. no bias) with default model dynamics (90/10). Each hollow point represents the final trait values of a given individual in the last generation (2000th generation) coloured by fitness. Coloured lines represent the average trait values at each generation for each population with the black point denoting the final generation. (b) Population dynamics of the four emergent populations showing the number of replicates (out of 100) that resulted in specific populations (size of circle) along with each population's rate of adaptation (colour of circles). (c) The left plot displays the fitness of each population over time while the right displays boxplots representing the distribution of the final fitness values across all individuals of all replicate runs (n = 100 000). Black lines in the boxplots denote the median with the edges denoting the 25th and 75th percentiles. Not all cryptic phenotypes were equally accessible by the model populations. Here, we define accessibility as the fraction of replicates that converged on an emergent population type. When the model was run without bias, Pop-MA was the most accessible with 55% of replicates converging on this population type while Pop-MD was the second most accessible with 33% (figure 5b). Pop-MA also exhibited the most variance in trait values within the population (i.e. broadest peak when plotted in more traditional pairwise trait space; e.g. Figure 5a), indicating a relatively larger range of trait values conferring high-fitness when associated with Pop-MA's trait correlations. The most accessible population type, Pop-MA, also had the fastest rate of fitness gain (figure 5b,c). Although Pop-MA and Pop-MB exhibited similar rates of adaptation (figure 5c, left plot), Pop-MB was not nearly as accessible with only 6% of the replicates converging on this population type (figure 5b). Instead, Pop-MD with a slower adaptive rate was the second most accessible population type (figure 5b,c). Pop-MA and Pop-MD trait correlations were more similar overall than those of Pop-MB. To examine the impact of ancestral trait values on the accessibility of different cryptic phenotypes, we ran the mixed-mode model using (i) a single population at a different starting location (i.e. trait values) in the trait-scape that was equidistant to the high–fitness area and (ii) four different subpopulations (i.e. four different starting locations in the trait-scape). These model runs converged on the same cryptic phenotypes observed with the empirical starting location (Pop-MA, Pop-MB and Pop-MD for run 1 and all four cryptic phenotypes for run 2). However, shifting the starting location did alter the accessibility of the cryptic phenotypes (electronic supplementary material, figures S3 and S4). These runs indicate that the high-fitness phenotypes were conserved, and that starting an adaptive walk from another location influenced the accessibility of certain cryptic phenotypes thereby biasing evolutionary outcomes. The fact that no new populations emerged further supports the ability of this framework to capture the known phenomenon that there are a limited number of accessible phenotypes [27]. (c) Adding historical bias Movement of the populations within the trait-scape was impacted by the historical bias of the ancestral population (i.e. trait correlations). In the mixed-mode model, the population contained a large diversity of trait correlations among individuals with analogous trait values. The robustness of the resulting cryptic phenotypes across model runs indicates that certain trait correlations confirm a fitness advantage within the empirically defined trait-scape and thus were selected for. To assess how different types of bias (ancestral versus evolved correlations) impacted the accessibility of the cryptic phenotypes and rate of adaptation, we conducted a suite of simulations where bias was systematically added (sub-modes A1–A4 and sub-modes E1–E4). For both ancestral and evolved modes, systematically adding more bias (e.g. going from A1 to A4) changed the accessibility of the high-fitness phenotypes across replicate runs (electronic supplementary material, figure S5). However, the type of bias (e.g. ancestral versus evolved correlations) had a differential impact on cryptic phenotype accessibility. Bias from the ancestral correlations was typically maladaptive and resulted in fewer accessible cryptic phenotypes and slower adaptive rates (electronic supplementary material, figure S5a). On the other hand, introducing bias derived from the observed evolved trait relationships (i.e. consistent with the trait-scape) generally resulted in faster adaptive rates and greater overall accessibility to the cryptic phenotypes (electronic supplementary material, figure S5b). These results are consistent with prior empirical and theoretical observations in developmental and quantitative genetic studies where bias (e.g. trait correlations) accelerated adaptive evolution if existing biological orientation aligned with the direction of selection but constrained adaptation if it limited variability in the direction of selection [16,17,27]. Specifically, depending on a starting population's bias, different phenotypes are more probable than others with some being generally inaccessible as found in other studies [27,28]. Here, we demonstrate these dynamics using a novel framework for modelling multivariate adaption in phytoplankton based on easy to quantify empirical trait data. (d) Meta-analysis of phenotypes across different modelled modes We assessed the similarity between emergent cryptic phenotypes across all model simulations (9 modes with 100 replicates each) and showed five high-confidence cryptic phenotypes. Specifically, we grouped similar population types using hierarchical clustering with multiscale bootstrap resampling (1000 replicates) on mean trait correlation values at the 2000th generation. We also included the empirical data from the ancestral and evolved populations in this analysis. Hierarchical clustering revealed five high-confidence clusters (I–V) harbouring 93% of the phenotypes (n = 26 of 28) with AU p-values greater than 75 (figure 6a). Two population types, Pop-MB and Pop-EB-E1, clustered with II and IV, respectively, albeit with less confidence relative to the high-confidence clusters. The empirical ancestral population did not fall within any of the high-confidence clusters, which is expected as the ancestral population was not well adapted in the evolved trait-scape. By contrast, the empirical evolved population clustered with high-confidence in cluster V. This empirically observed population type (cluster V) only emerged in the model simulation when evolved bias was present in the starting population and was found by 55% of the replicates from sub-modes E1, E3 and E4. Figure 6. Hierarchical clustering and principal component analysis of mean trait correlation values calculated across all emergent population types from mixed, ancestral and evolved mode model simulations. (a) Hierarchical clustering with multiscale bootstrap resampling (1000 replicates) on trait correlation values from the emergent population types (e.g. figure 4) across all model runs (mixed mode, A1–A4 and E1–E4) along with empirical ancestral and evolved correlation values. AU p-values >75 are labelled at the nodes. We identified five overarching clusters with high-confidence AU p-values (colours), which contained even higher confidence sub-clusters. (b) Principal component analyses with trait correlation values as in (a) with the five clusters projected onto the coordinate plane as convex hulls. Percentages on x- and y-axis denote the per cent of explained variance along each axis. Vectors C1–C6 denote correlations 1–6 as defined in figure 2c. The clustering observed through the hierarchical analysis also emerged through a PC analysis of the population trait correlations. Specifically, we observed three general regions of convergence in PC space among the population types, as clusters II, III and IV collapsed into a small region of the lower left quadrant in the PCA plot (figure 6b). Importantly, these convergent regions emerged from thousands of possible trait and correlation values across varying degrees of bias. They provide valuable insight into probable combinations of high-CO2 adaptive trait correlations along a reduced set of biological axes. 4. Discussion We need to bridge the gap between evolutionary models and trait-based ecosystems models (ODE models) in order to better predict how marine microbes will adapt to shifts in the environment [11,46]. This work takes a critical first step in developing a framework (TRACE) which uses empirically derived multivariate trait-based landscapes to provide insight into the interaction between historical bias (trait correlations) and evolved phenotypes for marine phytoplankton. Critically, TRACE is derived from and provides predictions of easily quantifiable traits—such as those commonly measured by biological oceanographers. Using data from an experimental evolution study with a model green alga, we found that a limited set of integrated phenotypes underlie thousands of possible trait correlational scenarios and that only certain phenotypes were accessible depending on the amount and type of bias. By leveraging empirical ancestral trait correlations and the observed changes in these correlations as a result of adaptation to high CO2, we were able to simulate adaptive walks with endpoints anchored in real evolutionary outcomes. This study provides a roadmap for future integration of evolutionary theory with biological oceanography. While we focus on a case study using an experimental evolution study to generate the trait-scape, our framework can be used to generate a trait-scape for any given environment A using trait data from an in situ population (assuming that the in situ population is well adapted to environment A). This trait-scape could then be used with the TRACE framework to develop hypotheses as to how a new population from a different environment B might adapt upon exposure to environment A. This insight can be gained with easily quantifiable trait measurements and without requiring genetic manipulations, which are currently not possible for most marine microbes. Our work demonstrates that ecological models need to represent both changes in traits (already existing in some ecological models) and changes in the correlation between traits in order to accurately capture phytoplankton evolution. The TRACE framework could be combined with an ODE based ecosystem model to predict marine microbial adaptation more accurately. While the development of a fully integrated TRACE + ODE model is beyond the scope of this paper, we demonstrate that such an integration would result in a substantially different set of phenotypes and impact community dynamics (electronic supplementary material, information S4 and figures S6 and S7). Such an integration provides the added benefit of removing the need to define the evolutionary endpoint, as fitness in an ODE model can be dynamically estimated based on prognostically calculated growth and mortality rates. This study demonstrates that shifts in trait correlations are fundamental for understanding the evolved phenotype and provides a novel framework for linking easily quantifiable trait measurements to predictions of evolved phenotypes for marine phytoplankton. This is particularly exciting because we can create trait-scapes from field data and use TRACE to understand how invading microbial populations may be able to adapt (i.e. create new phenotypes). Importantly, this approach can also help inform future experimental designs aimed at testing the probability of adaptive outcomes across multivariate environments through the analysis of a select set of traits. We are at a critical juncture where we need ecosystem and biogeochemical models to incorporate evolutionary dynamics in order to robustly predict future shifts in ecosystem dynamics [33]. Due to the seemingly infinite amount of possible interacting biological and environmental variables to test, evolutionary and mathematical tools that allow us to efficiently combine experiments with modelling will be critical to help predict microbial population responses to future global change scenarios through the lens of evolutionary phenomena. Supplementary Material Click here for additional data file. Data accessibility The model code is available at https://github.com/LevineLab and a version of the manuscript is available from the biology preprint server bioRxiv: https://www.biorxiv.org/content/10.1101/2020.08.04.237230 [47]. The data are provided in electronic supplementary material [48]. Authors' contributions N.G.W.: conceptualization, data curation, formal analysis, investigation, methodology, project administration, software, validation, visualization, writing-original draft, writing-review and editing; J.H.: conceptualization, investigation, methodology, writing-review and editing; P.A.A.: investigation, methodology, writing-review and editing; S.G.L.: data curation, formal analysis, investigation, methodology, visualization, writing-review and editing; M.A.D.: conceptualization, funding acquisition, investigation, methodology, writing-review and editing; S.C.: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, supervision, validation, writing-review and editing; N.M.L.: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, software, supervision, validation, writing-original draft, writing-review and editing. All authors gave final approval for publication and agreed to be held accountable for the work performed therein. Competing interests We declare that we have no competing interests. Funding This work was supported by the Moore Foundation grant MMI 7397 (to N.M.L., S.C., M.A.D.) and by the Simons Foundation grant 509727 (to N.M.L.). ==== Refs References 1. Walworth NG, Zakem EJ, Dunne JP, Collins S, Levine NM. 2020 Microbial evolutionary strategies in a dynamic ocean. Proc. Natl Acad. Sci. USA 117 , 5943-5948. (10.1073/pnas.1919332117)32123112 2. Hinners J, Hense I, Kremp A. 2019 Modelling phytoplankton adaptation to global warming based on resurrection experiments. Ecol. Modell. 400 , 27-33. (10.1016/j.ecolmodel.2019.03.006) 3. Lindberg RT, Collins S. 2020 Quality–quantity trade-offs drive functional trait evolution in a model microalgal ‘climate change winner’. Ecol. Lett. 23 , 780-790. (10.1111/ele.13478)32067351 4. Walworth NG, Fu F-X, Lee MD, Cai X, Saito MA, Webb EA, Hutchins DA. 2018 Nutrient-colimited Trichodesmium as a nitrogen source or sink in a Future Ocean. Appl. Environ. Microbiol. 84 , e02137-17-14. (10.1128/AEM.02137-17)29180365 5. Brennan GL, Colegrave N, Collins S. 2017 Evolutionary consequences of multidriver environmental change in an aquatic primary producer. Proc. Natl Acad. Sci. USA 114 , 9930-9935. (10.1073/pnas.1703375114)28847969 6. Hutchins DA, Fu F. 2017 Microorganisms and ocean global change. Nat. Microbiol. 2 , 1-11. (10.1038/nmicrobiol.2017.58) 7. Walworth NG, Lee MD, Fu F-X, Hutchins DA, Webb EA. 2016 Molecular and physiological evidence of genetic assimilation to high CO2 in the marine nitrogen fixer Trichodesmium. Proc. Natl Acad. Sci. USA 113 , E7367-E7374. (10.1073/pnas.1605202113)27830646 8. Hutchins DA, Walworth NG, Webb EA, Saito MA, Moran D, McIlvin MR, Gale J, Fu F-X. 2015 Irreversibly increased nitrogen fixation in Trichodesmium experimentally adapted to elevated carbon dioxide. Nat. Commun. 6 , 1-7. (10.1038/ncomms9155) 9. Hellweger FL, van Sebille E, Fredrick ND. 2014 Biogeographic patterns in ocean microbes emerge in a neutral agent-based model. Science 345 , 1346-1349. (10.1126/science.1254421)25214628 10. Schaum C-E, Buckling A, Smirnoff N, Studholme DJ, Yvon-Durocher G. 2018 Environmental fluctuations accelerate molecular evolution of thermal tolerance in a marine diatom. Nat. Commun. 9 , 1-14. (10.1038/s41467-018-03906-5)29317637 11. Ward BA, Collins S, Dutkiewicz S, Gibbs SJ, Bown P, Ridgwell A, Sauterey B, Wilson JD, Oschlies A. 2019 Considering the role of adaptive evolution in models of the ocean and climate system. J. Adv. Model. Earth Syst. 11 , 3343-3361. (10.31223/osf.io/srdh3)32025278 12. Schluter L, Lohbeck KT, Gröger JP, Riebesell U, Reusch TBH. 2016 Long-term dynamics of adaptive evolution in a globally important phytoplankton species to ocean acidification. Sci. Adv. 2 , e1501660. (10.1126/sciadv.1501660)27419227 13. Beckmann A, Schaum C-E, Hense I. 2019 Phytoplankton adaptation in ecosystem models. J. Theor. Biol. 468 , 60-71. (10.1016/j.jtbi.2019.01.041)30796940 14. Boyd PW, Cornwall CE, Davison A, Doney SC, Fourquez M, Hurd CL, Lima ID, McMinn A. 2016 Biological responses to environmental heterogeneity under future ocean conditions. Glob. Change Biol. 22 , 2633-2650. (10.1111/gcb.13287) 15. Boyd PW et al. 2018 Experimental strategies to assess the biological ramifications of multiple drivers of global ocean change-a review. Glob. Change Biol. 24 , 2239-2261. (10.1111/gcb.14102) 16. Aguirre JD, Hine E, McGuigan K, Blows MW. 2014 Comparing G: multivariate analysis of genetic variation in multiple populations. Heredity 112 , 21-29. (10.1038/hdy.2013.12)23486079 17. Agrawal AA. 2019 A scale-dependent framework for trade-offs, syndromes, and specialization in organismal biology. Ecology 101 , e02924. (10.1002/ecy.2924) 18. Barton S, Jenkins J, Buckling A, Schaum C-E, Smirnoff N, Raven JA, Durocher GY. 2020 Evolutionary temperature compensation of carbon fixation in marine phytoplankton. Ecol. Lett. 23 , 722-733. (10.1111/ele.13469)32059265 19. Brandenburg KM, Wohlrab S, John U, Kremp A, Jerney J, Krock B, Van de Waal DB. 2018 Intraspecific trait variation and trade-offs within and across populations of a toxic dinoflagellate. Ecol. Lett. 21 , 1561-1571. (10.1111/ele.13138)30117252 20. Malcom JW, Hernandez KM, Likos R, Wayne T, Leibold MA, Juenger TE. 2014 Extensive cross-environment fitness variation lies along few axes of genetic variation in the model alga, Chlamydomonas reinhardtii. New Phytol. 205 , 841-851. (10.1111/nph.13063)25264298 21. Anderson RW. 1995 Learning and evolution: a quantitative genetics approach. J. Theor. Biol. 175 , 89-101. (10.1006/jtbi.1995.0123)7564394 22. Tenaillon O. 2014 The utility of Fisher's geometric model in evolutionary genetics. Annu. Rev. Ecol. Evol. Syst. 45 , 179-201. (10.1146/annurev-ecolsys-120213-091846)26740803 23. Martin G, Lenormand T. 2015 The fitness effect of mutations across environments: Fisher's geometrical model with multiple optima. Evolution 69 , 1433-1447. (10.1111/evo.12671)25908434 24. Ward BA, Collins S, Dutkiewicz S, Gibbs S, Bown P, Ridgwell A, Sauterey B, Wilson JD, Oschlies A. 2019 Considering the role of adaptive evolution in models of the ocean and climate system. J. Adv. Model. Earth Syst. 11 , 3343-3361. (10.1029/2018MS001452)32025278 25. Parter M, Kashtan N, Alon U. 2008 Facilitated variation: how evolution learns from past environments to generalize to new environments. PLoS Comput. Biol. 4 , e1000206. (10.1371/journal.pcbi.1000206)18989390 26. Kashtan N, Noor E, Alon U. 2007 Varying environments can speed up evolution. Proc. Natl Acad. Sci. USA 104 , 13 711-13 716. (10.1073/pnas.0611630104) 27. Uller T, Moczek AP, Watson RA, Brakefield PM, Laland KN. 2018 Developmental bias and evolution: a regulatory network perspective. Genetics 209 , 949-966. (10.1534/genetics.118.300995)30049818 28. Elhanan Borenstein DCK. 2008 An end to endless forms: epistasis, phenotype distribution bias, and nonuniform evolution. PLoS Comput. Biol. 4 , e1000202. (10.1371/journal.pcbi.1000202)18949026 29. Gomulkiewicz R, Houle D. 2009 Demographic and genetic constraints on evolution. Am. Nat. 174 , E218-E229. (10.1086/645086)19821744 30. Braendle C, Baer CF, Félix M-A. 2010 Bias and evolution of the mutationally accessible phenotypic space in a developmental system. PLoS Genet. 6 , e1000877. (10.1371/journal.pgen.1000877)20300655 31. Houle D, Bolstad GH, van der Linde K, Hansen TF. 2017 Mutation predicts 40 million years of fly wing evolution. Nature 548 , 447-450. (10.1038/nature23473)28792935 32. Monroe JG, Markman DW, Beck WS, Felton AJ, Vahsen ML, Pressler Y. 2018 Ecoevolutionary dynamics of carbon cycling in the Anthropocene. Trends Ecol. Evol. 33 , 213-225. (10.1016/j.tree.2017.12.006)29398103 33. Baltar F et al. 2019 Towards integrating evolution, metabolism, and climate change studies of marine ecosystems. Trends Ecol. Evol. 34 , 1022-1033. (10.1016/j.tree.2019.07.003)31350054 34. Kronholm I, Collins S. 2015 Epigenetic mutations can both help and hinder adaptive evolution. Mol. Ecol. 25 , 1856-1868. (10.1111/mec.13296)26139359 35. Fisher RA. 1930 The genetical theory of natural selection. Oxford, UK: Oxford University Press. 36. Suzuki R, Shimodaira H. 2006 Pvclust: an R package for assessing the uncertainty in hierarchical clustering. Bioinformatics 22 , 1540-1542. (10.1093/bioinformatics/btl117)16595560 37. Oksanen J, Blanchet FG, Kindt R, Legendre P. 2013 vegan: community ecology package. R package version 2.0-4. 38. Lenski RE. 2017 Convergence and divergence in a long-term experiment with bacteria. Am. Nat. 190 , S57-S68. (10.1086/691209)28731830 39. Avrani S, Bolotin E, Katz S, Hershberg R. 2017 Rapid genetic adaptation during the first four months of survival under resource exhaustion. Mol. Biol. Evol. 34 , 1758-1769. (10.1093/molbev/msx118)28369614 40. Cooper VS, Lenski RE. 2000 The population genetics of ecological specialization in evolving Escherichia coli populations. Nature 407 , 736-739. (10.1038/35037572)11048718 41. Elena SF, Lenski RE. 2003 Microbial genetics: evolution experiments with microorganisms: the dynamics and genetic bases of adaptation. Nat. Rev. Genet. 4 , 457-469. (10.1038/nrg1088)12776215 42. Jerison ER, Ba ANN, Desai MM, Kryazhimskiy S. 2020 Chance and necessity in the pleiotropic consequences of adaptation for budding yeast. Nat. Ecol. Evol. 4 , 601-611. (10.1038/s41559-020-1128-3)32152531 43. Travisano M, Vasi F, Lenski RE. 1995 Long-term experimental evolution in Escherichia coli. III. Variation among replicate populations in correlated responses to novel environments. Evolution 49 , 189-200. (10.1111/j.1558-5646.1995.tb05970.x)28593661 44. Nakatsu CH, Korona R, Lenski RE, de Bruijn FJ, Marsh TL, Forney LJ. 1998 Parallel and divergent genotypic evolution in experimental populations of Ralstonia sp. J. Bacteriol. 180 , 4325-4331. (10.1128/JB.180.17.4325-4331.1998)9721265 45. Fong SS. 2005 Parallel adaptive evolution cultures of Escherichia coli lead to convergent growth phenotypes with different gene expression states. Genome Res. 15 , 1365-1372. (10.1101/gr.3832305)16204189 46. Walworth NG, Hinners J, Argyle PA, Leles SG, Doblin MA, Collins S, Levine NM. 2020 The evolution of trait correlations constrains phenotypic adaptation to high CO2 in a eukaryotic alga. bioRxiv. (10.1101/2020.08.04.237230) 47. Walworth NG, Hinners J, Argyle PA, Leles SG, Doblin MA, Collins S, Levine NM. 2021 The evolution of trait correlations constrains phenotypic adaptation to high CO2 in a eukaryotic alga. bioRxiv. (10.1101/2020.08.04.237230) 48. Walworth NG, Hinners J, Argyle PA, Leles SG, Doblin MA, Collins S, Levine NM. 2021 The evolution of trait correlations constrains phenotypic adaptation to high CO2 in a eukaryotic alga. FigShare.
PMC008xxxxxx/PMC8207559.txt
==== Front Proc Biol Sci Proc Biol Sci RSPB royprsb Proceedings of the Royal Society B: Biological Sciences 0962-8452 1471-2954 The Royal Society 34132112 10.1098/rspb.2021.0320 rspb20210320 10017014460Palaeobiology Research Articles Florivory of Early Cretaceous flowers by functionally diverse insects: implications for early angiosperm pollination Florivory of Early Cretaceous flowers by functionally diverse insects: implications for early angiosperm pollination http://orcid.org/0000-0002-6940-1473 Xiao Lifang Conceptualization Data curation Formal analysis Investigation Methodology Supervision Validation Visualization Writing-original draft Writing-review & editing 1 2 http://orcid.org/0000-0002-4838-5099 Labandeira Conrad Conceptualization Funding acquisition Methodology Project administration Supervision Validation Visualization Writing-original draft Writing-review & editing [email protected] 1 2 3 http://orcid.org/0000-0001-7226-6703 Dilcher David Conceptualization Data curation Investigation Resources Supervision Validation Visualization Writing-original draft Writing-review & editing 4 http://orcid.org/0000-0001-8660-0901 Ren Dong Conceptualization Data curation Funding acquisition Project administration Resources Supervision Validation Visualization Writing-original draft Writing-review & editing [email protected] 1 1 College of Life Science and Academy for Multidisciplinary Studies, Capital Normal University, Beijing, People's Republic of China 2 Department of Paleobiology, National Museum of Natural History, Smithsonian Institution, Tenth Street and Constitution Avenue, Washington, DC, USA 3 Department of Entomology, University of Maryland, College Park, MD, USA 4 Department of Geology, Indiana University, 1001 Tenth Street, Bloomington, IN, USA Electronic supplementary material is available online at https://doi.org/10.6084/m9.figshare.c.5448682. 30 6 2021 June 30, 2021 16 6 2021 June 16, 2021 16 6 2021 June 16, 2021 288 1953 202103208 2 2021 Feburary 8, 2021 20 5 2021 May 20, 2021 © 2021 The Authors. 2021 https://creativecommons.org/licenses/by/4.0/ Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited. Florivory (flower consumption) occurs worldwide in modern angiosperms, associated with pollen and nectar consumption. However, florivory remains unrecorded from fossil flowers since their Early Cretaceous appearance. We test hypotheses that earliest angiosperms were pollinated by a diverse insect fauna by evaluating 7858 plants from eight localities of the latest Albian Dakota Formation from midcontinental North America, in which 645 specimens (8.2%) were flowers or inflorescences. Well-preserved specimens were categorized into 32 morphotypes, nine of which displayed 207 instances of damage from 11 insect damage types (DTs) by four functional-feeding groups of hole feeding, margin feeding, surface feeding and piercing-and-sucking. We assessed the same DTs inflicted by known florivores on modern flowers that also are their pollinators, and associated insect mouthpart types causing such damage. The diverse, Dakota florivore–pollinator community showed a local pattern at Braun's Ranch of flower morphotypes 4 and 5 having piercing-and-sucking as dominant and margin feeding as minor interactions, whereas Dakotanthus cordiformis at Rose Creek I and II had an opposite pattern. We found no evidence for nectar robbing. These data support the rapid emergence of early angiosperms of florivore and associated pollinator guilds expressed at both the local and regional community levels. florivory , fossil plant damage , functional-feeding group , generalist , nectar robbing , pollination Project of High-level Teachers in Beijing Municipal Universities in the Period of 13th Five-year Plan http://dx.doi.org/10.13039/501100012435 Grant IDHT20180518 National Science Foundation of United States DEB 10720 DEB 75-02268 DEB 75-19849 DEB 77-04846 EAR 79-00898 National Natural Science Foundation of China http://dx.doi.org/10.13039/501100001809 Grants 31730087 Grants 32020103006 cover-dateJune 30, 2021 ==== Body pmc1. Introduction Flowers are the most successful plant reproductive structures ever to evolve on land [1] and angiosperms (flowering plants) presently are the most abundant and diverse clade of vascular plants [2], currently consisting of over 369 000 described species [1]. This lineage probably originated very early during the Cretaceous, with robust molecular phylogenies placing the origins of the clade at 139.35–136 Ma [3]. This timing is consistent with the earliest documented appearance of angiosperm pollen around 136 Ma [4] and earliest known intact flowers at 125 Ma, from the Early Cretaceous of northeastern China [2,5]. The best documented and earliest known bisexual flower, the ‘Rose Creek Flower’, from the Early Cretaceous Dakota Formation of the United States, examined in this report, is approximately 103 million years (103 Ma) in age [6,7]. Despite the long and sporadic record of fossil flowers and given the abundance and diversity of Dakota Formation flowers, the time is propitious for examination of insect florivory (electronic supplementary material, text S1), which is the consumption of flowers prior to seed coat formation [8], and associated pollination in probably the earliest angiosperm deposit conducive to such an assessment. Such an evaluation could provide a better understanding of the role that mutualisms and antagonisms of angiosperms, effected by their insect pollinators, had on their joint diversification [9]. Here, three hypotheses are posed that we seek to test in our study of Dakota Formation flowers and inflorescences. The first hypothesis is ‘general features of insect florivory and related pollination, as measured by damage patterns on Dakota flowers, are very similar or the same as those made by modern florivores’. The second hypothesis is ‘major taxonomic groups of insect florivores and pollinators from the Dakota Formation are very similar or the same as those of today’. The third, more focused, hypothesis is ‘nectar robbing was present on Dakota flowers’. By addressing these three hypotheses, we place this study in a broader context of Cretaceous angiosperm pollination and provide a glimpse into early angiosperm florivory and associated pollination about 30 million years after the earliest appearance of angiosperms. 2. Material and methods (a) The Dakota formation All compression or impression specimens of flowers and other reproductive material (heretofore termed flower specimens) described in this report were collected from the Dakota Formation of midcontinental United States at eight localities, each bearing a flora of mid-Cretaceous (latest Albian to earliest Cenomanian) age [7], equivalent to approximately 104–97 Ma [10], depending on the locality (electronic supplementary material, figure S1 and table S1). Flower specimens were collected during the 1970s and 1980s by D. Dilcher and colleagues and are stored at the Florida Museum of Natural History, in Gainesville, Florida. The localities are distributed along a younger to older north–northeast to south–southwest transect [11] (electronic supplementary material, figure S1). The localities were deposited along the eastern coast of the North American Mid Cretaceous Seaway that extended from the Gulf of Mexico to the Canadian Arctic [12,13]. Lithological facies within the Dakota Formation are dominated by shales and sandstones of the upper Janssen Clay Member and subjacent claystones of the lower Terra Cotta Clay Member [11,13]. Dakota depositional environments are represented by brackish estuaries, freshwater swamps, low energy channels, floodplain ponds and ox-bow lakes (electronic supplementary material, table S1). These biotic environments consisted of humid forests and woodland [11,13]. The Rose Creek I locality, where most of the flowers were deposited, represented a mangal-like marsh similar to extant communities in southeast Asia [14]. (b) Dakota formation floras Dakota Formation floras occur in a northeastern to southeastern trend at Courtland, Minnesota; Pleasant Dale, Nebraska (NE); Rose Creek I and II, NE; Braun's Ranch, Kansas (KS); Linnenberger Brothers' Ranch, KS; Acme Brick Quarry, KS; and Hoisington, KS [11] (electronic supplementary material, figure S1). Known vascular plant species diversity throughout the Dakota localities consists of 134 species. An earlier conservative estimate is 150–200 angiosperm species/morphotypes present in the Dakota Formation, as there is less than 25% species overlap between any two localities [6,7,11,15–18] (electronic supplementary material, table S1). Previous occasional descriptions of flowers, inflorescences, infructescences and fruits are known from several studies [7,19,20], yet affiliations of vegetative taxa with reproductive taxa remain largely unknown, although a few mostly wind-pollinated flowers, are associated with vegetative material that have been described [21]. Except for one locality, localities with higher abundance and diversity of foliage also appear to have a higher diversity and abundance of floral morphotypes (electronic supplementary material, figure S2 and table S1). (In this report, we use the term, floral, to refer to a flower, and not a bulk flora involving principally foliage.) The only described and best-known flower from the Dakota Formation is Dakotanthus cordiformis [7]. This flower has five, crescent-shaped, nectariferous pads that occur at the base of the gynoecium, each of which is aligned with a sepal. Dakotanthus, the most abundant morphotype in our dataset, is a member of the Rosidae 1 clade [7] and apparently very similar to a modern taxon with a lobed nectary disc. Other Dakota flower morphotypes show poor development or apparent absence of nectaries or nectary-like structures. However, leaf taxa occurring in the same localities as the unaffiliated Dakota flowers and infructescences have been assigned to extant families within Austrobaileyales, Chloranthales, Canellales, Magnoliales, Laurales and Rosidae 1 [15], which share a common pattern of fluid rewards for pollinating insects [22]. This pattern consists of: (i) staminoidal appendages (sterile stamens) that produce at their base glandular secretions of nectar-like fluids, mucilage, or ‘viscous substances’; (ii) nectariferous glands at the base or tips of fertile stamens; (iii) stigmas that secrete nectar-like substances, usually at their tips; (iv) nectar secreting, parenchymatous tissue on the adaxial surfaces of petals or sepals; and (v) large, substantive glands at the base of stamens that would qualify as true nectaries [22]. From these observations, it is highly likely that Dakota flower morphotypes produced nectar or other secretory, nectar-like fluids that attracted insect florivores and pollinators. (c) Identification of insect-damaged flowers and possible culprits The collection of florivory data is analogous to data for foliage or other vegetative organs and follows the same system of evaluating plant–insect associations [23], extensively used in fossil herbivory studies [24]. This system uses the functional-feeding group (FFG)–damage type (DT) system in which the overarching unit of herbivory is the FFG, examples of which are hole feeding, margin feeding, surface feeding and piercing-and-sucking for Dakota flower damage. Each FFG encompasses several or more DTs, which are the basic units of damage for fossil herbivory studies. A DT may be used in three ways. First, a DT may be used in terms of DT richness, referring to the kinds of DTs present; or as DT occurrences, as in the individual instances of damage of on a leaf; or as a formal name, such as DT405, which is a defined, specific mode of margin feeding damage. Details of photodocumentation and statistical methods are given in the electronic supplementary material, text S2. For Dakota plants, previous assessments of herbivory involved almost entirely mining damage on leaves [25–27]. However, Dakota plants, similar to amber deposits [28,29], provide considerable indirect evidence for flower–insect associations in the fossil record. (d) Distinguishing florivores and pollinators Insect visitors to flowers are of two fundamental groups, florivores and pollinators [30]. Not all florivores are pollinators and not all pollinators are florivores, and the relationships between these two ecological guilds are complex [8]. Florivores typically leave damage on flowers, overwhelmingly on petals [31], often resulting in negative interactions [32]. However, some florivore interactions are neutral or even positive [31,33], as petals occasionally contain nutritive or highly scented tissues designed for consumption by florivores as pollinators [34,35]. Florivory can be a form of predation if plant embryonic tissues are destroyed before the opening of the flower, or if there is the consumption of immature pollen, features that do not appear present in bowl-shaped Dakota flowers, as the damage is overwhelmingly on inner petal surfaces. Consequently, florivores such as Orthoptera (katydids), Hemiptera (aphids, bugs), Thysanoptera (thrips), Coleoptera (beetles) and Hymenoptera (sawflies, wasps, bees) with mandibulate, stylate or similarly modified mouthparts [36], provide good proxy data for the broad spectrum of pollinator interactions on flowers [37] (table 1). However, a substantial component such as most adult Diptera [51] and Lepidoptera are nondamagers, as they do not leave damage on flowers. Table 1. The potential florivory and pollination insect taxa from fossil and modern evidence. 3. Results (a) Total and insect-damaged flower morphotype abundance and diversity The eight localities of the Dakota Formation consist of approximately 7858 total plant specimens, which yielded 645 (8.2%) flower specimens that were assessed for insect damage, some of which were photographically documented (electronic supplementary material, text S2). Flower specimens previously were identified to morphotype by Dilcher and Manchester [10,36], and Xiao, but mostly by the latter. The plant specimens were categorized into 32 flower morphotypes, one of which was Dakotanthus cordiformis (figure 1; electronic supplementary material, figures S2, S3, S6 and appendix 1). The 32 morphotypes consisted of Dakotanthus cordiformis [6,7], 14 flower morphotypes, eight inflorescence–infructescence morphotypes, five reproductive morphotypes, two flower–seed–fruit morphotypes and two Braun Ranch flower morphotypes. Unidentifiable specimens and poor preserved morphotypes, not assigned to one of the 32 morphotypes, were Acme unidentified inflorescence–infructescence, unidentified flower and unidentified stamen, which amounted to one, eight and nine specimens, respectively, attributable to a very limited local sample size or poor preservation. Based on the diversity and abundance of floral morphotypes (electronic supplementary material, figures S2 and S3), our estimate of the log normal fit is 182 species. We also obtained a Fisher's α value of 7.94 (electronic supplementary material, appendix S2). Figure 1. The florivore assemblage on Dakotanthus cordiformis (a–l), displaying petal DTs from all four FFGs of hole feeding, margin feeding, surface feeding and piercing-and-sucking. Specimen UF-12941 at (a) shows DT13 margin feeding cusps on the distal edge of the petal, enlarged in (b); smaller versions of DT12 margin feeding and DT46 punctures enlarged in (c). From the same specimen are several, V-shaped, margin feeding, DT405 notches along the petal edge, enlarged in (d), and further enlarged, including a DT46 ovate puncture, at (e). Specimen UF-3522 with several stamens at (f) shows a series of DT405 edge notches, enlarged in (g). Specimen UF-5612 displays DT12 margin feeding at the bottom of the petal at (h), enlarged at (i). Dakotanthus cordiformis (UF-5773) at (j) are DT02 hole-feeding damage, enlarged in (k), and, together with small dark, ovate punctures of DT46, enlarged in (l). Note well-developed reaction rim surrounding DT01. Scale bars: white, 5 mm; black, 1 mm; empty, 0.5 mm. (Online version in colour.) (b) General patterns Insect damage was present on 109 of the 645 examined flower and related specimens, for a specimen-based florivory rate of 17.2%. This damage was represented by four FFGs, 11 DTs [23] and 207 individual DT occurrences (electronic supplementary material, table S2). Some DTs occurred multiple times on the same specimen. One or more DTs on a specimen was present on nine of the 32 flower morphotypes (28.1%) (electronic supplementary material, table S2). The four FFGs present were hole feeding, margin feeding, surface feeding and piercing-and-sucking (see the electronic supplementary material, tables S2, S3 and text S3 for additional DT occurrence details). (c) Assessing florivore host specificity by flower morphotype The distribution of DTs on plant hosts revealed three levels of host specificity [23]. Borrowing from studies of fossil herbivory as an example [24], host specificity is categorized as specialized damage if three or more occurrences of the same DT are present on the same host morphotype or on a very closely related host; damage is of intermediate specificity if the distribution of three or more occurrences of the same DT are present on more distantly related hosts; and generalized damage if three or more occurrences of the same DT are present on unrelated hosts [23]. For Dakota folivory data, because the phylogenetic relationships among flower morphotypes are unknown, terms expressing host specificity are referenced to the distribution of DTs on the flower species and morphotypes (electronic supplementary material, table S2). The three examples of specialist damage are small hole-feeding DT01 on flower morphotype 4 that hosts 12 of 13 (92.3%) of all occurrences; circular holes between 1 and 5 mm in diameter on Dakotanthus that hosts all eight (100%) of occurrences; and notched margin feeding of DT405 along the petal edges on Dakotanthus, which hosts 66 of 67 (98.5%) of all occurrences. The single example of damage of intermediate specificity is DT12, evidenced by too few DT distributions across three flower morphotypes. Seven examples of generalized damage are present. They are single, random, piercing-and-sucking damage assigned to DT46 on Dakotanthus and flower morphotypes 6, 9 and 10; clustered piercing-and-sucking assigned to DT402 on Dakotanthus and flower morphotypes 1, 5 and 8; and DT13, DT29, DT48, DT138 and DT383 that defaults to generalized specificity, each having only one or two occurrences on Dakotanthus and flower morphotypes 4, 5 and 7. This pattern of host specificity indicates three examples of specialized damage, one of intermediate specificity damage, and seven of generalized damage (electronic supplementary material, table S2). (d) Assessing the geographical distribution of florivory by locality Of the eight localities examined, flower morphotypes from three localities—Rose Creek I, Rose Creek II and Braun's Ranch—showed evidence of florivory. The combined Rose Creek I and II localities exhibited three flower morphotypes with 117 DT occurrences, whereas the Braun's Ranch locality showed a higher diversity of seven flower morphotypes and 90 DT occurrences (electronic supplementary material, table S2). The number of florivorized to total flower morphotypes at each locality (electronic supplementary material, table S2)—three of 10 (30%) at Rose Creek I and II, and seven of 13 (53.8%) at Braun's Ranch—are distinctly significant subsets of the number of available hosts at each locality. At Rose Creek, Dakotanthus overwhelmingly was the dominant flower morphotype present, which displayed a rich spectrum of damage, with three of the four FFGs and seven of the 11 DTs represented (electronic supplementary material, table S2). By comparison, Braun's Ranch showed two florivorized flower morphotypes with a less rich spectrum of FFGs and DTs. Flower morphotype 4 had three of four FFGs and five of 11 DTs present. Similarly, flower morphotype 5 displayed three of four FFGs and six of 11 DTs. These latter flower morphotypes from the Rose Creek and Braun's Ranch localities exhibited a similar distribution of FFGs and DTs. (e) Assessing functional-feeding group and damage type on flower morphotypes (i) Hole feeding Hole feeding on flowers of the Dakota Formation mostly is single, small and circular perforations of the entire petal thickness that are ovate or circular in shape (electronic supplementary material, text S4). Reaction rims are variably developed and occasionally associated with necroses of adjacent petal tissue. Damage type DT01 consists of holes 1 mm or less in diameter and is associated with flower morphotype 4 (electronic supplementary material, figure S5H–J). DT02 consists of holes between 1 and 5 mm in diameter and occurs on Dakotanthus (figure 1j–k). Dakota hole feeding consists of 21 perforations of DT01 and DT02 that represent 10.1% of all DT occurrences among flower morphotypes. Of all hole-feeding occurrences, 38.1% was present on Dakotanthus cordiformis from the Rose Creek I and II localities and 57.1% on flower morphotype 4 from the Braun's Ranch locality. It was noted that 71.4% of hole feeding was present on the lower half, rather than the upper half of the petals. (ii) Margin feeding Margin feeding on Dakota Formation flowers is represented by DTs DT12, DT13 and DT405. DT12 and DT13 consist of cuspate to U-shaped excisions, typically several mm in chord length, occurring along the edges of petals and sepals (electronic supplementary material, text S4). The cut edge, in addition to a bordering rim of dark reaction tissue, occasionally displays micromorphological features such as protruding veinal stringers, necrotic tissue flaps and cuspules within the overall cut edge, analogous to damage on foliage. DT12 occurred along the petal side edges on Dakotanthus (figure 1c,h,i), flower morphotype 4 (electronic supplementary material, figure S5E,F), flower morphotype 5 (electronic supplementary material, figure S4A) and flower morphotype 17 (electronic supplementary material, figure S6F). DT13 was present on the tips of petals of Dakotanthus (figure 1a,b) and flower morphotype 5 (electronic supplementary material, figure S4A). Careful examination of DT13 was required to determine if the damage was present, to eliminate confusion with a retuse or apically embayed margin. DT405 is a newly described DT (electronic supplementary material, text S5) and previously has not been recorded in the fossil record. (iii) Surface feeding The only example of a surface feeding FFG on a Dakota flower morphotype was DT DT29 (not illustrated) occurring on flower morphotype 7. DT29 is highly variable in size and shape, featuring polylobate to ovate patches of surface-fed petal tissue with distinct development of a reaction rim resulting from abrasion, scraping or delamination of a surface tissue layer (electronic supplementary material, text S4). (iv) Piercing-and-sucking Piercing-and-sucking damage of Dakota floral morphotypes is represented by the five DTs of DT46, DT48, DT138, DT383 and DT402. These DTs consist of various patterns of punctures that penetrate or slice into shallow to deep floral tissues (electronic supplementary material, text S4). DT46 and DT48 are single, randomly dispersed punctures less than 1 mm in diameter, present on petals or other flower elements. DT46 is a circular, concave mark with a crater-like rim and occurs on Dakotanthus (figure 1c,e,j,l), flower morphotype 1 (electronic supplementary material, figure S5A,B), and flower morphotype 5 (electronic supplementary material, figure S5C,D,M–O). By contrast, rare DT48 (not illustrated) is an elliptical puncture, with either a cratered rim or a convex central boss. DT138 are linear rows of punctures that occur on flower morphotype 5 (electronic supplementary material, figure S4D,E). DT383 are compact circular to polylobate clusters of punctures that probably accessed deeper tissues (electronic supplementary material, table S5). DT402, a newly described DT (electronic supplementary material, text S5), represents typically elongate, compact clusters of punctures in shallow tissues that occur especially along petal or sepal edges. DT402 occurs on flower morphotype 4 (electronic supplementary material, figure S5F,G and S5K–M,O,P) and flower morphotype 5 (electronic supplementary material, figure S4B,C,F–I). No preferential occurrences of piercing-and-sucking DTs were noted for the five DTs occurring on the upper versus lower halves of the petals. (f) Matching insect mouthpart classes with feeding damage Associations were established between the pattern of Dakota insect damage with the relevant insect mouthpart class borne by an insect that would have produced that damage [36,52,53]. Such relationships, based on modern data [36] (electronic supplementary material, table S3), were made to better constrain the identities of potential florivores and pollinators (electronic supplementary material, text S6). However, Dakota damage caused by adult ectognathate, larval ectognathate and sericterate mouthpart classes typically could not be separated from each other. These three mandibulate (chewing) mouthpart classes account for 37.7% of all DT occurrences. Damage caused by the maxillolabiate (a complex apparatus for nectar-extraction) and rhynchophorate (perforating) mouthpart classes that result in hole-feeding damage cannot be distinguished from each other, but these two mouthpart classes that create holes on leaves account for 10.1% of all DT occurrences. Damage caused by the segmented beak mouthpart class, responsible for puncturing deeper tissues with stylate mouthparts, involve piercing-and-sucking feeding and account for 31.4% of all DT occurrences. Damage attributable to mouthcone mouthparts modified for punch and sucking of shallow, epidermal tissues account for 20.8% of DT occurrences. These seven mouthpart classes, responsible for four major feeding styles, indicate that florivory was dominated by edge feeders and tissue penetrating piercer and suckers, and less so by hole feeders and shallowly penetrating piercer and suckers, reconstructed in figure 2. Figure 2. A reconstruction of the insect pollinator community on Dakotanthus cordiformis [7] based on patterns of florivory. This scene is from the Rose Creek locality of the Early Cretaceous (late Albian) Dakota Formation of Southwestern Nebraska, USA. Painted by Xiaoran Zuo. (Online version in colour.) (g) Analyses of the damage For the five most prevalent flower morphotypes, the two metrics expressing the abundance (DT occurrences) and percentage of DTs present are a near-exact match of each other (electronic supplementary material, table S4). This near duplication is shown in (i) the percentage contribution of each morphotype to the total, (ii) the morphotype rank order, and (iii) the cumulative totals. Much less similar is the third metric of the percentage of specimens that are florivorized, which departs from the two other metrics in exhibiting greater differences in the contribution of each morphotype to the total, a different morphotype rank order after the first two most florivorized morphotypes of Dakotanthus and flower morphotype 4, and a higher cumulative total of 97.7 versus the first two of 86.6 and 85.3. A plot of the percentage of florivory of the three dominant FFGs of hole feeding, margin feeding and piercing-and-sucking was made for the major florivorized flower morphotypes of Dakotanthus, flower morphotype 4 and flower morphotype 5 for the three Dakota localities of Rose Creek I, Rose Creek II and Braun’ Ranch (figures 3 and 4). This analysis revealed three patterns. First, the florivore–pollinator communities of flower morphotypes 4 and 5 at Braun's Ranch locality are dominated by the piercing-and-sucking FFG, whereas the hole feeding and margin feeding FFGs played a minor role at both localities. By contrast, the Dakotanthus florivore faunas are dominated by the margin feeding FFG at the Rose Creek I and II localities, whereas the hole feeding and piercing-and-sucking FFGs have minor roles at both localities, although piercing-and-sucking appears to be subdominant at the Rose Creek II locality. Second, whereas margin feeding or piercing-and-sucking may have played a dominant role, depending on locality, hole feeding always had a minor role in the florivory spectrum across all localities. Third, with the exception of hole feeding, the florivore communities at Braun's Ranch versus the Rose Creek I and II localities, were largely feeding inversions of each other, suggesting heterogeneity in the pollinator assemblages by locality. Figure 3. Percentage representation of folivory for the five most insect-damaged flower morphotypes. Figure 4. Per cent representation of florivore FFGs for flower morphotypes and sites. HF, hole feeding; MF, margin feeding; SF, surface feeding; PS, piercing-and-sucking. (Online version in colour.) 4. Discussion (a) How similar is Dakota florivory, folivores and pollinators to modern counterparts? The presence of four FFGs subsuming 11 DTs in Dakota Formation plants (figures 1 and 2; electronic supplementary material, figures S4–S6) provides a very modern cast to the documented florivory (electronic supplementary material, table S3, text S4 and figure S7). In terms of the distinctiveness of the insect-mediated damage, the mouthpart classes responsible for the damage (electronic supplementary material, text S6), the proportional distribution of the damage on the flower morphotypes (electronic supplementary material, table S4), and their near-identical comparison to damage produced by known modern taxa [8] affirms the hypothesis that Dakota florivory is indistinguishable from its modern equivalent. This similarity suggests a similar pollinator community, although evidence for fluid feeding, nondamaging adult taxa is circumstantial. The richness of Dakota florivory and suggested associated pollinators provides, to our knowledge, the first extensive evidence for a community of insect visitors on the earliest, well-documented bowl-shaped flowers (electronic supplementary material, table S5, figure S7 and text S7). The florivore component of Dakota insect pollinators is established based on (i) distinctive DTs, (ii) extant lineages that were present during Dakota time from fossil occurrences, and (iii) evidence of relevant fossil occurrences or presence of closely related clades (phylogenetic bracketing) (electronic supplementary material, texts S6 and S7). The florivore assemblage consisted of a variety of insects with mandibulate and piercing-and-sucking mouthparts that produced recognizable damage patterns on flowers (electronic supplementary material, text S6). This pattern affirms the hypothesis that major taxonomic groups of insect florivores were very similar to their extant counterparts (electronic supplementary material, table S6 and text S8). Probably associated with this community of insect damagers of floral tissue, the indirect evidence indicates presence of lineages of fluid feeding, adult and nondamager taxa that left no trace on Dakota flowers. Based on the Dakota florivory data and modern studies, the core pollinators were heteropterans (especially pentatomorphs), thrips, polyphagan beetles (principally scarab and leaf beetles, and weevils), and bees. A subordinate component of early-diverging moths, sawflies, several major lineages of nematoceran and brachyceran flies, and perhaps parasitoid wasps probably were present (electronic supplementary material, table S6 and text S7). (b) Is there evidence for nectar robbing? The data on hole feeding across the flower morphotype hosts is intriguing. Hole feeding consisting of 23 perforations of DTs DT01 and DT02 represent 10.1% of all DT occurrences among Dakota flower morphotypes. Of hole-feeding occurrences, 95.2% were present on Dakotanthus from the Rose Creek I locality and on flower morphotype 4 from the Braun's Ranch locality. There is a distinct preference for the lower half (71.4%) rather than the upper half of the petal. All holes were circular or nearly so, and no evidence of slits or tears was observed as holes on the petals. Although these data suggest that nectar robbing was present, nectar robbing would be ineffectual if flowers are open, bowl shaped, and with rewards such as nectar readily available to insect visitors [54]. All modern nectar robbing occurs with tubular or similar flowers that that are highly enclosed and have access through a narrowly throated corolla [54]. The floral morphology of Dakota flowers, exemplified by Dakotanthus, is inconsistent with nectar robbing, and holes are circular and rounded in contrast with modern nectar robbers that construct slits or holes with jagged outlines [55,56]. Consequently, hypothesis 3, that nectar robbing was present on Dakota flowers, is rejected. Given this outcome, a more productive search for the earliest nectar robbing would be among early occurrences of tubular or otherwise enclosed flowers in the younger Late Cretaceous [2]. (c) A brief historical perspective on Cretaceous angiosperm pollination The early fossil history of angiosperm pollination [2] is illustrated by the bowl and similarly shaped floras in the 21 localities listed in geochronological order that provide data on inferred insect pollinator lineages and their functional-feeding group membership (electronic supplementary material, table S5 and text S7). This list shows that the Dakota insect pollinator fauna probably had taxonomic similarities to that of extant basal angiosperms (electronic supplementary material, table S6; texts S7 and S8). The next, four-million year-younger assemblage of pollinating insects originates from the very geographically, ecologically and taphonomically different locality of Myanmar amber, which shows a distinct pollinator fauna (electronic supplementary material, table S5 and figure S8). Subsequent, Late Cretaceous floras originate from a variety of localities that reveal the expansion of dicot angiosperms and an associated pollinator fauna. 5. Conclusion This study provides a new approach for the study of pollination in the fossil record. By examining exquisitely preserved insect damage on Dakota flowers, the three hypotheses have been tested that were initially proposed in this study. The first hypothesis—florivore damage patterns on Dakota flowers are similar to those of today—is supported. One proviso to this conclusion is that evidence for pollinating insects on Dakota flowers result from insect immatures and adults capable of leaving detectable damage, principally those with mandibulate and stylate mouthparts or their modifications, but excludes adult fluid feeding insects whose mouthparts are incapable of damaging tissues. Nevertheless, the indirect evidence of phylogenetic bracketing indicates the existence of these nondamaging, fluid feeding and pollinating clades. The second hypothesis—major taxonomic groups Dakota insect florivores are similar to those of today—is supported. Again, a caveat is that major insect taxa lacking the ability to inflict damage on flowers remain undetected. Circumstantial evidence based on fossil occurrences and phylogenetic bracketing indicates that typical nectar-feeding taxa were present. Last, the third hypothesis—nectar robbing was present—is rejected. The reason for rejection is not that a great preponderance of hole feeding was located at petal bases, but rather the open, bowl shape of Dakota flowers would obviate the need for nectar robbing. While addressing these three hypotheses extends understanding of florivory and pollination to the Dakota Formation, important gaps in knowledge of early angiosperm pollination remain. Highly relevant deposits from this time interval (electronic supplementary material, table 5, text S7 and figure S7) should be further explored. Supplementary Material Click here for additional data file. Acknowledgements We especially thank two reviewers who give us important comments. We thank Gussie Maccracken, Terry Lott, Finnegan Marsh, Steven Manchester, Hongshan Wang, Nareerat Boonchai and Xiaodan Lin who individually provided support in the Labandeira and Manchester laboratories. Sandra Schachat is thanked for producing figures 3 and 4. Gussie Maccracken provided a critique of a preliminary draft of the manuscript and offered needed overall advice. We are especially grateful to Steven Manchester for sponsoring the senior author at the Florida Museum of Natural History, in Gainesville for two months and their dedicated staff for general assistance. We thank Huayan Chen who provided photos of modern florivores and Torsten Dikow who identified the fly in electronic supplementary material, figure S6. We thank Liang Chen who helped us in figure construction. The Fossilworks database is acknowledged. This is contribution 378 to the Evolution of Terrestrial Ecosystems consortium at the National Museum of Natural History, in Washington, D.C. Data accessibility The data supporting the analyses of this article consist of (i) table 1 in this article; (ii) more extensive, linked electronic supplementary material that includes raw and related data among the three fossil localities as well as modern data; (iii) appendix 1 that provides detailed descriptions of flower morphotypes/species; and (iv) appendix 2 that provides R code files and Excel data files in a zip file. The data are provided in the electronic supplementary material. Authors' contributions L.X.: conceptualization, data curation, formal analysis, investigation, methodology, supervision, validation, visualization, writing—original draft, writing—review and editing; C.L.: conceptualization, funding acquisition, methodology, project administration, supervision, validation, visualization, writing—original draft, writing—review and editing; D.D.: conceptualization, data curation, investigation, resources, supervision, validation, visualization, writing—original draft, writing—review and editing; D.R.: conceptualization, data curation, funding acquisition, project administration, resources, supervision, validation, visualization, writing—original draft, writing—review and editing. All authors gave final approval for publication and agreed to be held accountable for the work performed therein. Competing interests We declare that we have no competing interests. Funding This research was supported by the National Natural Science Foundation of China (grant nos. 31730087, 32020103006 and 41688103). Fieldwork was supported in part by National Science Foundation of United States grant nos. DEB 75-02268, DEB 75-19849, DEB 77-04846, DEB 10720 and EAR 79-00898 to D.D. ==== Refs References 1. Sauquet H, von Balthazar M, Magallón S. 2017 The ancestral flower of angiosperms and its early diversification. Nat. Comm. 8 , 16047. (10.1038/ncomms16047) 2. Friis EM, Pedersen KR, Crane PR. 2006 Cretaceous angiosperm flowers: innovation and evolution in plant reproduction. Palaeogeogr. Palaeoclim. Palaeoecol. 232 , 251-293. (10.1016/j.palaeo.2005.07.006) 3. Magallón S, Gómez-Acevedo S, Sánchez-Reyes LL, Hernández-Hernández T. 2015 A metacalibrated time-tree documents the early rise of flowering plant phylogenetic diversity. New Phytol. 207 , 437-453. (10.1111/nph.13264)25615647 4. Crane PR, Friis EM, Pedersen KR. 1995 The origin and early diversification of angiosperms. Nature 374 , 27-33. (10.1038/374027a0) 5. Sun G, Zheng S, Sun C, Sun Y, Dilcher DL, Miao Y. 2002 Androecium of Archaefructus, the Late Jurassic angiosperms from western Liaoning, China. J. Geosci. Res. Northeast Asia 5 , 1-6. (in Chinese and English). 6. Basinger J, Dilcher DL. 1984 Ancient bisexual flowers. Science 224 , 511-513. (10.1126/science.224.4648.511)17753776 7. Manchester SR, Dilcher DL, Judd WS, Corder B, Basinger JF. 2018 Early Eudicot flower and fruit: Dakotanthus gen. nov. from the Cretaceous Dakota Formation of Kansas and Nebraska, USA. Acta Palaeobot. 58 , 27-40. (10.2478/acpa-2018-0006) 8. McCall AC, Irwin RE. 2006 Florivory: the intersection of pollination and herbivory. Ecol. Lett. 9 , 1351-1365. (10.1111/j.1461-0248.2006.00975.x)17118009 9. Pellmyr O. 1992 Evolution of insect pollination and angiosperm diversification. Trends Ecol. Evol. 7 , 46-49. (10.1016/0169-5347(92)90105-K)21235949 10. Walker JD, Geissman JW, Bowring SA, Babcock LE. 2013 The Geological Society of America geologic time scale. Geol. Soc. Am. Bull. 125 , 259-272. (10.1130/B30712.1) 11. Wang HS. 2002 Diversity of angiosperm leaf megafossils from the Dakota Formation (Cenomanian, Cretaceous), north Western Interior, USA. Dissertation, Department of Geological Sciences, University of Florida, Gainesville, FL, USA. 12. Ludvigson GA, Witzke BJ, Joeckel RM, Ravn RL, Phillips PL, Gonzalez LA, Brenner RL. 2010 New insights on the sequence stratigraphic architecture of the Dakota Formation in Kansas–Nebraska–Iowa from a decade of sponsored research activity. Curr. Res. Earth Sci. Bull. 258 , 1-36. 13. Retallack GJ, Dilcher DL. 2012 Outcrop versus core and geophysical log interpretation of mid-Cretaceous paleosols from the Dakota Formation of Kansas. Palaeogeogr. Palaeoclim. Palaeoecol. 329–330 , 47-63. (10.1016/j.palaeo.2012.02.017) 14. Richards PW. 1952 The tropical rain forest: an ecological study. Cambridge, UK: Cambridge University Press. 15. Upchurch GR, Dilcher DL. 1990 Cenomanian angiosperm leaf megafossils, Dakota Formation, Rose Creek locality, Jefferson County, southeastern Nebraska. U.S. Geol. Surv. Bull. 1915 , 1-55, pls 1–31 16. Wang HS, Dilcher DL. 2006 Early Cretaceous angiosperm leaves from the Dakota Formation, Braun Ranch locality, Kansas, USA. Palaeontographica Abt. B 273 , 101-137. (10.1127/palb/273/2006/101) 17. Wang HS, Dilcher DL. 2009 Late Cretaceous angiosperm leaves from the Courtland clay pit, Minnesota, USA. Palaeontographica Abt. B 281 , 143-177. (10.1127/palb/281/2009/143) 18. Wang HS, Dilcher DL. 2018 Early Cretaceous angiosperm leaves from the Dakota Formation, Hoisington III locality, Kansas, USA. Palaeontol. Elect. 21.3.34A , 1-49. (10.26879/841) 19. Lesquereux L. 1892 The flora of the Dakota Group. U.S. Geol. Surv. Monogr. 17 , 1-400. 20. Dilcher DL, Crepet W, Becker CD, Reynolds HC. 1976 Reproductive and vegetative morphology of a Cretaceous angiosperm. Science 191 , 854-856. (10.1126/science.191.4229.854)17730999 21. Dilcher DL, Crane PR. 1984 Archaeanthus: an early angiosperm from the Cenomanian of the western interior of North America. Ann. Missouri Bot. Gard. 71 , 351-383. (10.2307/2399030) 22. Erber C. 2014 Nectar secretion and nectaries in basal angiosperms, magnoliids and non-core eudicots and a comparison with core eudicots. Pl. Div. Evol. 131 , 63-143. (10.1127/1869-6155/2014/0131-0075) 23. Labandeira CC, Wilf P, Johnson KR, Marsh F. 2007 Guide to insect (and other) damage types on compressed plant fossils (version 3.0 – spring 2007). Washington, DC: Smithsonian Institution. see http://paleobiology.si.edu/pdfs/insectDamageGuide3.01.pdf. 24. Xu Q, Jin H, Labandeira CC. 2018 Williamson Drive: herbivory on a north-central Texas flora of latest Pennsylvanian age shows discrete component community structure, early expansion of piercing and sucking, and plant counterdefenses. Rev. Palaeobot. Palynol. 251 , 28-72. (10.1016/j.revpalbo.2018.01.002) 25. Labandeira CC, Dilcher DL, Davis DR, Wagner DL. 1994 Ninety-seven million years of angiosperm–insect association: paleobiological insights into the meaning of coevolution. Proc. Natl Acad. Sci. USA 91 , 12 278-12 282. (10.1073/pnas.91.25.12278) 26. Labandeira CC. 1998 The role of insects in Late Jurassic to middle Cretaceous ecosystems. In Lower and middle Cretaceous terrestrial ecosystems, vol. 14 (eds SG Lucas, JI Kirkland, JW Estep), pp. 105-124. Albuquerque, NM: New Mexico Museum of Nature and Science. 27. Labandeira CC. 2007 Assessing the fossil record of plant–insect associations: ichnodata versus body-fossil data. In Sediment–organism interactions: a multifaceted ichnology, vol. 88 (eds RG Bromley, LA Buatois G Mángano, JF Genise, RN Melchor), pp. 9-26. Tulsa, OK: Society for Sedimentary Geology. 28. Peñalver E, Labandeira CC, Barrón E, Delclòs X, Nel P, Nel A, Tafforeau P, Soriano C. 2012 Thrips pollination of Mesozoic gymnosperms. Proc. Natl Acad. Sci. USA 109 , 8623-8628. (10.1073/pnas.1120499109)22615414 29. Peris D, Labandeira CC, Barrón E, Delclòs X, Rust J, Wang B. 2020 Generalist pollen-feeding beetles during the mid-Cretaceous. Iscience 23 , 100913. (10.1016/j.isci.2020.100913)32191877 30. de Jaeger ML, Ellis AG. 2014 Floral polymorphism and the fitness implications of attracting pollinating and florivorous insects. Ann. Bot. 113 , 213-222. (10.1093/aob/mct189)24052554 31. Nunes CEP, Peñaflor MFGV, Bento JMS, Salvador MJ, Sazima M. 2016 The dilemma of being a fragrant flower: the major floral volatile attracts pollinators and florivores in the euglossine-pollinated orchid Dichaea pendula. Oecologia 182 , 933-946. (10.1007/s00442-016-3703-5)27538674 32. Cardel YJ, Koptur S. 2010 Effects of florivory on the pollination of flowers: an experimental field study with a perennial plant. Int. J. Plant Sci. 171 , 283-292. (10.1086/650154) 33. Althoff DM, Seagraves KA, Pellmyr O. 2005 Community context of an obligate mutualism: pollinator and florivore effects on Yucca filamentosa. Ecology 86 , 905-913. (10.1890/04-1454) 34. Theis N. 2006 Fragrance of Canada thistle (Cirsium arvense) attracts both floral herbivores and pollinators. J. Chem. Ecol. 32 , 917-927. (10.1007/s10886-006-9051-x)16739013 35. Kessler D, Diezel C, Clark DG, Colquhoun TA, Baldwin IT. 2013 Petunia flowers solve the defence/apparency dilemma of pollinator attraction by deploying complex floral blends. Ecol. Lett. 16 , 299-306. (10.1111/ele.12038)23173705 36. Labandeira CC. 2019 The fossil record of insect mouthparts: innovation, functional convergence, and associations with other organisms. In Insect mouthparts – form, function, development and performance, vol. 5 . Zool. Monogr. pp. 567-671. Cham, Switzerland: Springer. 37. Maldonado M, Sakuragui CM, Trigo JR, Rodrigues D. 2015 The selective florivory of Erioscelis emarginata matches its role as a pollinator of Philodendron. Entomol. Exp. Appl. 156 , 290-300. (10.1111/eea.12332) 38. Song H, Amédegnato C, Cigliano MM, Desutter-Grandcolas L, Heads SW, Huang DY, Otte D, Whiting MF. 2015 300 million years of diversification: elucidating the patterns of orthopteran evolution based on comprehensive taxon and gene sampling. Cladistics 31 , 621-651. (10.1111/cla.12116)34753270 39. Zhang SM, Mound L, Feng JN. 2019 Morphological phylogeny of Thripidae (Thysanoptera: Terebrantia). Invertebr. Syst. 33 , 671-696. 40. Szwedo J, Nel A. 2011 The oldest aphid insect from the Middle Triassic of the Vosges, France. Acta Palaeobot. Pol. 56 , 757-766. (10.4202/app.2010.0034) 41. Burdfield-Steel ER, Shuker DM 2014 The evolutionary ecology of the Lygaeidae. Evol. Ecol. 4 , 2278-2301. 42. Jung SH, Lee SW. 2012 Molecular phylogeny of the plant bugs (Heteroptera: Miridae) and the evolution of feeding habits. Cladistics 28 , 50-79. (10.1111/j.1096-0031.2011.00365.x)34861758 43. Yao YZ, Ren D, Rider DA, Cai WZ. 2012 Phylogeny of the Infraorder Pentatomorpha based on fossil and extant morphology, with description of a new fossil family from China. PLoS ONE 7 , e37289. (10.1371/journal.pone.0037289)22655038 44. McKenna DD, Farrell BD. 2009 Beetles (Coleoptera). In The timetree of life (eds SB Hedges, S Kumar), pp. 278-289. New York: NY: Oxford University Press. 45. Wang M, Li LF, Shih CK, Gao TP, Ren D. 2019 Hymenoptera – sawflies and wasps. In Rhythms of insect evolution: evidence from the Jurassic and Cretaceous in northern China (eds D Ren, CK Shih, TP Gao, YJ Wang, YZ Yao), pp. 429-496. Chichester, UK: Wiley-Blackwell. 46. Cardinal S, Danforth BN. 2013 Bees diversified in the age of eudicots. Proc. R. Soc. B 280 , 20122686. (10.1098/rspb.2012.2686) 47. Mitter C, Davis DR, Cummings MP. 2017 Phylogeny and evolution of Lepidoptera. Annu. Rev. Entomol. 62 , 265-283. (10.1146/annurev-ento-031616-035125)27860521 48. Han Y, Ye XU, Feng CP, Zhang KY, Shih CK, Ren D. 2019 Diptera – true flies with two wings. In Rhythms of insect evolution: evidence from the Jurassic and Cretaceous in northern China (eds D Ren, CK Shih, TP Gao, YJ Wang, YZ Yao), pp. 496-553. Chichester, UK: Wiley-Blackwell. 49. Proctor M, Yeo P, Lack A. 1996 The natural history of pollination. London, UK: HarperCollins. 50. Faegri K, Van der Pijl L. 1980 Principles of pollination ecology, 3rd Revised edn. Oxford, UK: Pergamon. 51. Wolda H, Sabrosky CW. 1986 Insect visitors to two forms of Aristolochia pilosa in Las Cumbres, Panama. Biotropica 18 , 295-299. (10.2307/2388572) 52. Kirk WDJ. 1984 Ecological studies on Thrips imaginis Bagnall (Thysanoptera) in flowers of Echium plantagineum L. in Australia. Austral. J. Ecol. 9 , 9-18. 53. Iannuzzi R, Labandeira CC. 2008 The oldest record of external foliage feeding and the expansion of insect folivory on land. Ann. Entomol. Soc. Am. 101 , 79-94. (10.1603/0013-8746(2008)101[79:TOROEF]2.0.CO;2) 54. Irwin RE, Bronstein JL, Manson JS, Richardson L. 2010 Nectar robbing: ecological and evolutionary perspectives. Annu. Rev. Ecol. Evol. Syst. 41 , 271-292. (10.1146/annurev.ecolsys.110308.120330) 55. Clement SL. 1992 On the function of pea flower feeding by Bruchus pisorum. Entomol. Exp. Appl. 63 , 115-121. (10.1111/j.1570-7458.1992.tb01566.x) 56. Eliyahu D, McCall AC. 2015 Florivory and nectar-robbing perforations in flowers of pointleaf manzanita Arctostaphylos pungens (Ericaceae) and their effects on plant reproductive success. Arthro. Pl. Interact. 9 , 613-622. (10.1007/s11829-015-9399-3)
PMC008xxxxxx/PMC8220259.txt
==== Front Proc Biol Sci Proc Biol Sci RSPB royprsb Proceedings of the Royal Society B: Biological Sciences 0962-8452 1471-2954 The Royal Society 34157877 10.1098/rspb.2021.0782 rspb20210782 10016070198Global Change and Conservation Research Articles Divergent regional evolutionary histories of a devastating global amphibian pathogen Divergent regional evolutionary histories of a devastating global amphibian pathogen http://orcid.org/0000-0002-1329-8702 Rothstein Andrew P. Conceptualization Data curation Formal analysis Investigation Methodology Visualization Writing-original draft Writing-review & editing [email protected] 1 2 Byrne Allison Q. Conceptualization Data curation Formal analysis Investigation Methodology Visualization Writing-original draft Writing-review & editing 1 2 3 Knapp Roland A. Conceptualization Data curation Funding acquisition Investigation Methodology Project administration Supervision Visualization Writing-original draft Writing-review & editing 4 5 Briggs Cheryl J. Conceptualization Data curation Funding acquisition Investigation Project administration Supervision Writing-original draft Writing-review & editing 5 6 Voyles Jamie Conceptualization Data curation Funding acquisition Investigation Project administration Supervision Writing-original draft Writing-review & editing 7 Richards-Zawacki Corinne L. 8 Rosenblum Erica Bree Conceptualization Data curation Funding acquisition Investigation Methodology Project administration Supervision Visualization Writing-original draft Writing-review & editing 1 2 1 Department of Environmental Science, Policy, and Management, University of California Berkeley, Berkeley, CA, USA 2 Museum of Vertebrate Zoology, University of California Berkeley, Berkeley, CA, USA 3 Center for Conservation Genomics, Smithsonian Conservation Biology Institute, National Zoological Park, Washington, DC, USA 4 Sierra Nevada Aquatic Research Laboratory, University of California, Mammoth Lakes, CA, USA 5 Earth Research Institute, University of California, Santa Barbara, CA, USA 6 Department of Ecology, Evolution, and Marine Biology, University of California, Santa Barbara, CA, USA 7 Department of Biology, University of Nevada, Reno, NV, USA 8 Department of Biological Sciences, University of Pittsburgh, Pittsburgh, PA, USA Electronic supplementary material is available online at https://doi.org/10.6084/m9.figshare.c.5459508. 30 6 2021 June 30, 2021 23 6 2021 June 23, 2021 23 6 2021 June 23, 2021 288 1953 202107826 4 2021 April 6, 2021 28 5 2021 May 28, 2021 © 2021 The Authors. 2021 https://creativecommons.org/licenses/by/4.0/ Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited. Emerging infectious diseases are a pressing threat to global biological diversity. Increased incidence and severity of novel pathogens underscores the need for methodological advances to understand pathogen emergence and spread. Here, we use genetic epidemiology to test, and challenge, key hypotheses about a devastating zoonotic disease impacting amphibians globally. Using an amplicon-based sequencing method and non-invasive samples we retrospectively explore the history of the fungal pathogen Batrachochytrium dendrobatidis (Bd) in two emblematic amphibian systems: the Sierra Nevada of California and Central Panama. The hypothesis in both regions is the hypervirulent Global Panzootic Lineage of Bd (BdGPL) was recently introduced and spread rapidly in a wave-like pattern. Our data challenge this hypothesis by demonstrating similar epizootic signatures can have radically different underlying evolutionary histories. In Central Panama, our genetic data confirm a recent and rapid pathogen spread. However, BdGPL in the Sierra Nevada has remarkable spatial structuring, high genetic diversity and a relatively older history inferred from time-dated phylogenies. Thus, this deadly pathogen lineage may have a longer history in some regions than assumed, providing insights into its origin and spread. Overall, our results highlight the importance of integrating observed wildlife die-offs with genetic data to more accurately reconstruct pathogen outbreaks. wildlife disease , Batrachochytrium dendrobatidis , genetic epidemiology , population genetics , amphibians Division of Environmental Biology http://dx.doi.org/10.13039/100000155 1457695 1551488 1557190 166311 National Park Service http://dx.doi.org/10.13039/100007516 DOD SERDP cover-dateJune 30, 2021 ==== Body pmc1. Introduction Globalization has contributed to a surge in the incidence, severity and spread of emerging infectious diseases (e.g. [1,2]). Emerging diseases of wildlife are particularly important to global biological diversity as they can cause devastating population declines and exacerbate other threats such as habitat loss, overharvesting, invasive species and climate change [3–6]. Recent advances in the study of disease emergence and spread integrate epidemiological and genetic data to test theoretical predictions about the ecological history of the pathogen given the underlying evolutionary signal [7,8]. However, most applications of this approach have been for quickly evolving pathogens (i.e. RNA viruses) and those that directly impact human health. There have been a handful of studies applying methodological advances in genetic epidemiology to emerging wildlife diseases (see recent reviews [9,10]), but such frameworks are still largely underutilized. Amphibians are declining worldwide [11,12]. One of the major drivers of amphibian declines is the global spread of the disease chytridiomycosis, caused by the fungal pathogen Batrachochytrium dendrobatidis (Bd) [13]. Bd infects the keratinized skin cells of susceptible host species, disrupts vital amphibian skin functions and can cause mortality [14]. In some cases, Bd infections can spread quickly across individuals, populations, and species leading to pizootic outbreaks and population and community collapses [15,16]. Since the earliest observations of Bd-related die-offs in the late 1990s, Bd has emerged as a global threat to amphibian biodiversity and now impacts amphibians on every continent where they are present [12]. Bd has a complex evolutionary history with multiple lineages found in different parts of the world. Phylogenetically, Bd is characterized by several early branching lineages endemic to different regions (BdCAPE, BdASIA1, BdBrazil/ASIA2 and BdASIA3) and one more recently derived hypervirulent panzootic lineage (BdGPL). BdGPL has been linked to declines of amphibian communities around the world and is the only Bd lineage with a truly global distribution [6,17,18]. Whole-genome data have been important for revealing the dynamics of BdGPL spread [6,19]. BdGPL typically exhibits little phylogenetic or spatial genetic structure (with the exception of two subclades BdGPL-1 and BdGPL-2) [20,21], suggesting that this lineage spread rapidly around the world [18,19]. Moreover, compared to other Bd lineages, BdGPL genomes have fewer pairwise genetic differences among them and highly variable genetic diversity values [6]. Observations of minimal pairwise genetic differences are consistent with rapid BdGPL spatial radiation, and variability in genetic diversity suggests episodes of population size fluctuation. However, we still lack a connection between our understanding of Bd evolutionary history at a global scale and regional Bd emergence and spread. Two of the most emblematic BdGPL-related declines occurred in the montane amphibian communities of the Sierra Nevada of California and Central Panama. In the Sierra Nevada of California, mountain yellow-legged frogs (Rana sierrae/muscosa), were historically one of the most abundant vertebrates [22]. Over the last century, these frogs vanished from more than 90% of their historic range, and Bd (along with invasive fish) was a significant factor in their decline [23]. Available information suggests that Bd has been spreading across the Sierra Nevada since at least the 1960s [24,25] and has caused epizootics and subsequent extirpations in hundreds of populations [16,26,27]. Some populations that experienced Bd-related declines are beginning to rebound, but remaining naive populations are still at risk for Bd epizootics [28]. Similarly, in Central America, amphibian population declines were first observed in the late 1980s [29,30]. As Bd spread southeast into Central Panama starting in the early 2000s [15], many susceptible amphibian host species declined—or even disappeared completely—across the region [15,31,32]. Although some species seem to be recovering [32], Bd-related declines have fundamentally reshaped these tropical communities [31,33,34]. From an epizoological perspective, amphibian declines in the Sierra Nevada and Central Panama appear quite similar. In both regions, initial detection of Bd was followed by devastating outbreaks and host mortality. Patterns of decline in both the Sierra Nevada and Central Panama also appear to provide evidence of a ‘wave’-like spread of Bd across the landscape [16,35]. Pathogen prevalence and population decline data in both systems suggest that new infections appear in a predictable spatial direction and that Bd outbreaks move a predictable distance each year [15,16,35]. Coupled with a global phylogenetic view of Bd, the prevailing hypothesis suggests that BdGPL is a recent invasive pathogen in these two regions [36]. However, epizoological data based on observed outbreaks and host outcomes may or may not reflect the true history of Bd arrival and spread. The Sierra Nevada and Central Panama differ dramatically in climate, habitat, and amphibian community composition. Ecological and environmental factors impact the physiological limits of Bd, transmission dynamics across the landscape and host immunity; extensive evidence suggests these factors influence prevalence and host disease outcomes [20,37–39]. Stark differences between regional environments likely contribute to different Bd dynamics in the Sierra Nevada and Panama, making their apparent similarities in disease outcomes all the more intriguing. Although it is often assumed that Bd arrived recently and spread in a wave-like fashion in both regions, it is possible that different evolutionary histories of Bd underlie these observed patterns. Molecular data can reveal nuances of a pathogen's history that cannot be obtained by field observations alone. Genetic and genomic approaches have previously been used to investigate the evolutionary history of Bd at regional and global scales [6,18,19,40]. However, most studies of the evolutionary history of Bd in emblematic systems like the Sierra Nevada and Central Panama have relied on a small number of Bd isolates for any one region. Live and pure Bd cultures have been the source of high-quality DNA for genomic sequencing (e.g. [6,19,41]) but are inherently challenging to obtain, isolate and maintain. Low sample sizes and poor spatial coverage has made it difficult to test fine-scale hypotheses about Bd emergence and spread. However, advances in sequencing technology now allow for leveraging fine-scale sampling of frog skin swabs, previously used to determine Bd presence/absence and load, to robustly characterize Bd genotypes across relevant spatial scales [42]. Thus, we can now test whether patterns of Bd emergence that appear similar across systems result from shared underlying processes. We used fine-scale genetic sampling to investigate assumptions about the history of BdGPL in the Sierra Nevada and Central Panama. Using non-invasive skin swabs collected across similar spatial and temporal scales, we targeted hundreds of loci across the Bd genome to examine the hypothesis of recent Bd emergence and unidirectional epizootic spread in these two emblematic systems. Our work provides an in-depth understanding of pathogen evolutionary dynamics in natural systems and an example for how researchers should not expect infectious diseases to emerge and spread similarly across the globe. 2. Material and methods (a) Sampling and genotyping We used skin swab DNA samples collected from the Sierra Nevada and Central Panama across similar timescales (2011–2017) and across equivalent spatial scales (approx. 130 km across Euclidean distance between furthest two sites) (figure 1a). Sites are defined as collections of lakes and streams that cluster together geographically within a region. We sampled 10 sites from both the Sierra Nevada (nsamples = 130; nspecies = 2) and Central Panama (nsamples = 80; nspecies = 17). Sierra Nevada samples comprised skin swabs from two sister species of frogs (R. sierrae/muscosa) [23] and Central Panama samples comprised skin swabs from 16 different frog species. Additionally, we included 120 previously published sequenced samples from a global Bd dataset, comprised of samples across 59 frog species from six continents (see electronic supplementary material). The global samples were all previously positioned within the BdGPL clade based on a comprehensive assessment of hundreds of Bd samples [17] and serve to add a global context to levels of genetic structure and diversity observed in the Sierra Nevada and Central Panama. Figure 1. Study system map and principal component analysis of within region genotypes. (a) Map of sites sampled in the study in the Sierra Nevada and Central Panama. (b) PCA within Sierra Nevada samples, coloured by the major site. Samples cluster by site, suggesting strong genetic structuring across the Sierra Nevada. (c) PCA within Central Panama samples, coloured by the site. Compared to samples from the Sierra Nevada, Central Panama samples exhibit a dramatically different pattern, i.e. panmixis, despite a similar spatial and temporal scale of sampling. Colours in (b,c) correspond to geographical locations in (a). (Online version in colour.) We sequenced 240 regions (each 150–200 bp long) of the Bd genome from the Sierra Nevada and Central Panama skin swab samples. We first conducted a pre-amplification step (which improves performance for amplicon sequencing) and then used these pre-amplified products in a microfluidic PCR approach (see electronic supplementary materials for DNA extraction, preparation and PCR conditions) [42]. Pre-amplified products were loaded into a Fluidigm Access Array IFC, individually barcoded, then pooled for sequencing on ¼ of an Illumina MiSeq lane with 2 × 300 bp paired-end reads at the University of Idaho IBEST Genomics Resources Core. From raw sequence reads, we used the dbcAmplicons software (https://github.com/msettles/dbcAmplicons) to trim adapter, primer sequences and merged to continuous reads. We de-multiplexed and filtered sequences using the reduce_amplicons.R script within the dbcAmplicons repository into two sequence types: ambiguities and raw fastq for each sample. Ambiguities sequence files used IUPAC ambiguity codes to identify multiple alleles. Raw fastq files are all sequences for each sample. Ambiguity sequences were used for phylogenetic analyses and the fastq by the sample was used for alignment, variant calling and filtering VCF for downstream analyses. Specific parameters used for alignment, variant calling and variant filtering can be found in the electronic supplementary material. Post filtering our raw 4534 variants, we recovered 2268 variable sites across 235 amplicons. (b) Genetic diversity We applied PCA to examine genetic clustering and structuring among all samples. We estimated PCs using adegenet [43] and visualized in R (v. 3.6.1). We calculated summary diversity statistics using ANGSD [44]. For this analysis, diversity statistics were calculated based on genotype likelihoods, which is distinct from the variant calling approach by Freebayes described above. Given that sample sizes can impact diversity metrics, we randomly subsampled our the Sierra Nevada and global BdGPL samples to equal the number of Central Panama samples (n = 80). Additionally, 49 amplicons (previously developed as Central Panama-specific markers) were removed from the filtered 235 amplicons, leaving 186 amplicons for diversity statistics. Using filtered BAMs from our variant calls, we generated a folded site frequency spectrum given an unknown ancestral state. After estimating the site frequency spectrum for each region, we calculated per-site Watterson's θ and π for the Sierra Nevada (mean amplicon depth = 85.3; s.d. = 78.9; range = 7.0–464.8), Central Panama (mean amplicon depth = 207.8; s.d. = 206.2, range = 15.4–909.6) and global BdGPL (mean amplicon depth = 323.7; s.d. = 383.1; range = 5.5–1342.5) samples. We tested for significant differences in mean Watterson's θ and π and using analysis of variance followed by Tukey's HSD in R (v. 3.6.1), given that we had multiple pairwise comparisons of our global BdGPL reference, Sierra Nevada and Central Panama samples. (c) Phylodynamics Using ambiguity sequences by sample, we created a phylogeny including Sierra Nevada, Central Panama and our global BdGPL reference panel. We removed amplicons that had no data and included samples that had least 20 amplicons. We trimmed loci that had greater than 5 bp difference between the minimum and maximum sequence length to control for improper alignments near large indels. A final list of 206 loci were individually aligned using the MUSCLE package in R (v. 3.6.1, [45]) and concatenated (28 688 bp in length). We also included previously published sequences of UM142 BdBrazil as an outgroup [17]. Using this concatenated alignment, we built a phylogeny using IQ-Tree with 1000 ultrafast bootstrap replicates and chose the best model from AIC scores using ‘model-finder’ (GTR + F + I + G4) [46]. We assessed temporal signal in our phylogeny using TempEst (v. 1.5.3) and a date randomization test with TipDatingBeast (v. 1.1) (details available in electronic supplementary material, Methods) [47,48]. After confirming the temporal signal, we inferred time-measured phylogenies, with our concatenated alignment and recorded sampling years, using both BEAST2 [49] and Nextstrain [50]. We used time-measure parameters from a previously published whole-genome phylogenies for Bd to ensure comparability [6]. Briefly, for BEAST2 we used a GTR substitution model with estimated mutation rates 7.29 × 10−7 (lower; 3.41 × 10−7, upper; 1.14 × 10−6) and extended Bayesian skyline plot as demographic parameter [6]. Using this model, we ran a chain which drew samples every 3000 MCMC steps from a total of 575 000 000 steps, after a discarded burn-in of 57 500 000 steps. Convergence of distribution and effective sample size greater than 150 were checked through Tracer (v. 1.7.1) [51]. Our best-supported tree was estimated using maximum clade credibility through TreeAnnotator (v. 2.6) [49] and was visualized using FigTree (v. 1.4.4) (https://github.com/rambaut/figtree). Comparative methods for Nextstrain can be found in the electronic supplementary material. It is important to note that we used BEAST2 and Nextstrain as analytical frameworks to compare patterns between the Sierra Nevada and Central Panama but not to infer exact introduction dates. Applying the same evolutionary models across two geographical regions provides a powerful comparative tool and allows us to infer relative evolutionary rates and introduction timings. However, we interpret specific dates with caution given that patterns of Bd genome evolution may violate a number of model assumptions (e.g. variation across the genome in recombination and mutation rates, variation in chromosomal copy numbers, the potential for both meiotic and mitotic recombination) [19,41], and because our sampling dates do not necessarily correspond to first introduction dates. Given that any violation of basic model assumptions would be shared across study regions, comparisons between the Sierra Nevada and Central Panama can be used to draw conclusions about the relative invasion history in these regions. 3. Results (a) Bd from the Sierra Nevada shows greater population structure than Bd from Central Panama When comparing within regions, we found significant genetic clustering across the Sierra Nevada (figure 1b) but no genetic clustering across Central Panama (figure 1c). Samples collected from the same site in the Sierra Nevada clustered together, regardless of collection year or species (electronic supplementary material, figure S1). Starting with Unicorn Ponds at the north, samples generally follow a pattern of isolation by distance. LeConte Divide and Conness Pond are somewhat anomalous, however, because they overlap in PC space but are geographically separated by approximately 80 km (figure 1b). By contrast, Central Panama genotypes exhibited panmictic patterns, regardless of locality, collection year or species, indicating no genetic structuring across a similar spatio-temporal scale (figure 1c, electronic supplementary material, figure S2). (b) Bd from the Sierra Nevada shows greater variation and diversity than Bd from Central Panama We confirmed that Bd from the Sierra Nevada and Central Panama belong to the global BdGPL lineage. Both Sierra Nevada and Central Panama samples clustered with a panel of global samples that were previously identified as BdGPL [17]. Interestingly, the Sierra Nevada and Central Panama samples clustered separately from each other in PC space when compared to global BdGPL samples (figure 2a). Additionally, we found that overall genetic diversity was significantly higher in the Sierra Nevada as compared to Central Panama [Tukey HSD, p < 0.0001] (figure 2b,c). Remarkably, we also found that Sierra Nevada Bd samples have comparable and, in the case of Watterson's θ, higher diversity than the set of global BdGPL samples [Tukey HSD, p < 0.0001]. When comparing Central Panama and the Sierra Nevada using individual sites with similar samples sizes, we found that the majority of Sierra Nevada sites had higher mean diversity compared to Central Panama sites (both Watterson's θ and π) [Tukey HSD, p < 0.001], except in the lowest sample size pairing (N = 5) where El Valle S. had significantly higher mean diversity than LeConte Divide (electronic supplementary material, figure S3). Figure 2. Genetic differentiation and diversity among the Sierra Nevada, Central Panama, and global BdGPL samples. (a) PCA based on BdGPL genotypes from the Sierra Nevada (n = 130), Central Panama (n = 80) and global reference panel (n = 120). Colours indicate samples from each region. The global reference panel included samples from dozens of frog species across all continents with BdGPL. Samples from the Sierra Nevada and Central Panama are almost entirely separated in PC space with the Sierra Nevada samples showing greater genetic variation than Central Panama samples. (b) Distribution of mean genetic diversity (Watterson's θ) for all variable sites based on region. Samples from the Sierra Nevada and global panels were randomly subsampled to match Central Panama sample size (all regions n = 80). Mean genetic diversity was significantly higher for Sierra Nevada samples compared to Central Panama samples and to the global BdGPL panel [Tukey HSD, p < 0.0001]. (c) Distribution of mean nucleotide diversity (π) for all variable sites based on the region using the same samples as (b). Mean nucleotide diversity was significantly lower for Central Panama samples compared to Sierra Nevada samples and the global BdGPL panel [Tukey HSD, both p < 0.0001]. Each box plot shows the median (horizontal line), first and third quartiles (bottom and top of box, ‘hinges’), lowest and highest values within inter-quartile range of the lower and upper hinges (vertical lines) and outliers (points). (Online version in colour.) (c) Bd in the Sierra Nevada is inferred to be older than Bd in Central Panama Using a time-dated phylogenetic approach that included previously published global BdGPL samples for reference [17], we found branches from Sierra Nevada samples were comparatively older than those in Central Panama (figure 3, electronic supplementary material, figure S4). As discussed in the Material and methods, we do not assume the specific inferred dates are accurate given the likelihood that dynamics of Bd genome evolution violate several model assumptions. Our root-to-tip regression showed somewhat low temporal signal in our data (R2 = 0.02) (electronic supplementary material, figure S5). However, our date randomization tests showed no overlap between real and randomized datasets, indicating a sufficient level of the temporal signal (electronic supplementary material, figure S6). While these results may appear contradictory, root-to-tip regression is a conservative approach assuming a strict molecular clock [47], while date randomization provides a more statistically robust method of comparison [48]. Therefore, as discussed in the Methods section above, our time-dated approaches are appropriate for inferring relative invasion histories across regions rather than proposing specific divergence or invasion dates. Figure 3. BEAST2 timed dated phylogeny among the Sierra Nevada, Central Panama, and global BdGPL samples. Branch tips are colour coded by region. The tree is rooted by an outgroup from a more basal Bd lineage (BdBrazil isolate UM142). Sierra Nevada samples are found across the tree, in multiple clusters, and with longer branch lengths than Central Panama samples suggesting a longer history of Bd in this region. (Online version in colour.) For BEAST2, the time to most recent common ancestor (tMRCA) for Sierra Nevada samples was estimated to be 474 years from present day (95%HPD 510–393 years from present day) and estimated tMRCA in Central Panama was 277 years from present day (95%HPD 389–60 years from present day) (figure 3). For Nextstrain, tMRCA for Sierra Nevada samples was estimated as 1407 years from present day (95% CI 4498–1151) and tMRCA for Central Panama was estimated as 666 years from present day (95% CI 1914–534) (electronic supplementary material, figure S4); dynamic Nextstrain visualizations are available at: https://nextstrain.org/community/andrew-rothstein/bd-gpl/auspice/viz. Therefore, even without ascribing weight to specific inferred dates, Bd in the Sierra Nevada appears to be much older than Bd in Panama. Confidence intervals for the inferred tMRCA do not overlap between regions with either analysis. The BEAST2 and Nexstrain time-dated phylogenetic approaches also corroborated PCA results (figure 1). Sierra Nevada samples largely clustered by the site while Central Panama samples had little to no structure based on-site location. (electronic supplementary material, figure S4) Finally, phylogenetic trees show an expected split within BdGPL, supported by high BEAST posterior node values (electronic supplementary material, figure S7). This correspond to a previously reported split separating BdGPL into two subclades: BdGPL-1 and BdGPL-2 [20,21]. Only GPL-2 is represented in Panama samples while GPL-1 and GPL-2 are both found in the Sierra Nevada samples. 4. Discussion Bd has caused mass amphibian declines in many regions of the world and has been an example of the devasting effects of emerging wildlife diseases [12,15,16,35,52–54]. However, assessments of Bd emergence and spread have yet to incorporate genetically informed epizoology to examine pathogen dynamics at fine spatial scales. Our study used comparative population genetics to examine the genetic signatures of BdGPL across two emblematic regions with disease-related amphibian declines. The alpine lakes of the Sierra Nevada and the tropical forests of Central Panama have dramatically different climate, habitat and host communities. However, they have been described as having similar histories of recent Bd emergence and spread. We tested the assumption that BdGPL was recently introduced to these two regions and swept through each in a unidirectional epizootic wave. We found dramatic differences in Bd evolutionary history across regions, with an unexpectedly deep history of Bd in the Sierra Nevada. Here, we explore differences across regions providing a new perspective on these important historic declines. As wildlife disease rapidly continue to spread across the world, our framework is broadly applicable to interrogating observed patterns of pathogen emergence and spread to uncover important evolutionary pathogen histories. (a) How do patterns of pathogen genetic variation differ across regions? (i) BdGPL in Central Panama is genetically similar and spatially unstructured Our results from Central Panama support the hypothesis of a recent introduction, with Bd in this region lacking any spatial structure. All Bd genotypes from the Central Panama group tightly together, are generally distinct from Bd collected in the Sierra Nevada, and are all part of the GPL-2 subclade. This pattern supports previous studies reporting a single fast-moving outbreak of Bd through Central Panama [35]. Our samples from Central Panama were collected approximately 8 years after observed outbreaks (between 2012 and 2016), and the observed lack of genetic structure indicates that Bd did not diverge on a site-specific basis over this time period. Our findings support other recent studies showing a lack of genetic, phenotypic and functional shifts in Central Panama Bd across similar temporal scales [32]. BdGPL appears to have arrived in Panama much more recently than in the Sierra Nevada, maintained low levels of genetic diversity, and, over the last two decades, currently has no detectable genetic sub-structure. (ii) BdGPL in Sierra Nevada is genetically diverse and spatially structured We observed a dramatically different pattern in the Sierra Nevada, where we found high levels of genetic variation between sampling sites and spatial structuring of Bd genotypes. Although Bd samples were collected across a similar spatial and temporal scale as those from Panama, our genetic data indicate that BdGPL has likely had a much longer historical presence in the Sierra Nevada than it has in Panama. This conclusion is supported by multiple lines of evidence. First, Sierra Nevada Bd contains more genetic variation and diversity than Central Panama (figure 2a). Measures of nucleotide diversity (π), are higher in Sierra Nevada Bd samples compared to Central Panama and Sierra Nevada Bd genetic diversity (Watterson's θ) is significantly higher than the entire global panel of BdGPL samples (figure 2b). This result is consistent with previous evidence that BdGPL in the Sierra Nevada has higher levels of genetic diversity than BdGPL from Arizona, Mexico, or Central Panama [55]. Second, we also observed a surprising pattern of spatially structured genetic diversity for BdGPL in the Sierra Nevada. Sierra Nevada BdGPL genotypes typically cluster by site and segregate by geographical distance in PC space and in the phylogeny (figures 1b and 3b). Much of the observed genetic structure in the Sierra Nevada is consistent with a pattern of isolation by distance, suggesting a much longer history of Bd on the landscape. Third, even the exceptions to the pattern of isolation by distance suggests a deeper and more complex history of Bd in the Sierra Nevada. Samples from LeConte Divide and Conness Pond are genetically distinct from all other samples in the Sierra Nevada and cluster in PC space (figure 1b). These samples belong to a separate, early branching clade referred to as GPL-1 (figure 3). The presence of both BdGPL-1 and BdGPL-2 subclades could represent multiple independent introductions or much deeper in situ divergence, possibilities we revisit below. (b) What do regional differences suggest about BdGPL origin and invasion history? (i) BdGPL in Sierra Nevada likely predates the most recently observed wave of declines One key factor that could contribute to radically different patterns of Bd genetic variation between Central Panama and the Sierra Nevada is invasion history (the timing and number of introductions). Our Nextstrain and BEAST2 analyses infer that Bd from the Sierra Nevada is older than Bd from Central Panama (figure 3, electronic supplementary material, figure S4). While our inference indicates that BdGPL has been in the Sierra Nevada longer than Central Panama, it is difficult to assert specific invasion dates. As discussed in the Material and methods section, patterns of Bd genome evolution may violate a number of model assumptions. Although our analyses used a species-specific mutation rate inferred from Bd whole-genome analyses [6] our assay targets regions of the Bd genome that are most informative for discriminating among Bd lineages [42] and therefore may not evolve with a shared background mutation rate. Even without specific introduction dates, studies using histology and qPCR to test for Bd presence in museum specimens have often shown Bd presence prior to field-observed die-offs [24,56,57], which could indicate older introduction timings than previously assumed. As such, Bd presence has been detected in samples as far back as 1932 in the Sierra Nevada [24] and 1964 in Costa Rica (adjacent to Panama) [56]. Moreover, field observations suggest that Bd may be present in the environment well before an outbreak is observed. In some lakes, Bd is present at almost undetectably low prevalence and load for years before Bd loads spike and die-offs occur [16,28,58]. In some systems, Bd can even be detected from eDNA surveys before die-offs occur [59]. Such dynamics challenge our a priori expectations that Bd die-offs occur immediately after the pathogen first arrives in an area. In some systems, such as the Sierra Nevada and parts of Costa Rica [24,56,57], it is possible that Bd had a more wide-spread presence earlier than perceived. Whether there actually were earlier Bd-caused die-offs remains an open question. Increased surveillance of Bd before and during early outbreaks is needed to decouple initial pathogen invasion from observed pathogen-induced declines. (ii) The Sierra Nevada is a potential source for BdGPL High levels of genetic variation, deep spatial genetic structure and the presence of both subclades of BdGPL in the Sierra Nevada suggest a longer evolutionary history of Bd in the region than previously appreciated. The presence of both BdGPL-1 and BdGPL-2 could represent multiple asynchronous invasions of BdGPL, a hypothesis raised by another recent spatial–temporal study of Bd presence in the Sierra Nevada [24]. An alternative explanation is that California is a potential source of Bd that has spread to other regions. As sampling resolution improves, it is possible that we will find other regions of the world with highly diverse and spatially structured BdGPL populations. However, it is also worth continuing to challenge our assumptions about the origin and spread of this lineage. While the most basal lineage of Bd is from Asia [6], the origin of BdGPL remains highly uncertain. Although we often assume that BdGPL presence results from recent invasions, the region from which BdGPL originated would be expected to have general characteristics similar to what we observe in the Sierra Nevada (i.e. relatively high genetic diversity and deep spatial structure). No such region other than the Sierra Nevada has yet been identified. Global sampling with greater spatial and temporal resolution will be needed to ultimately determine the origins of this highly virulent Bd lineage. (c) How do biotic and abiotic factors influence observed Bd genetic variation? (i) Differences in topography, host life history, community structure and climate also likely contribute to divergent patterns of pathogen genetic structure across regions Biotic and abiotic factors also likely influence patterns of Bd genetic variation in a consistent direction, with increased opportunity for pathogen mixing in Central Panama relative to the Sierra Nevada. Central Panama is home to a diverse amphibian assemblage, with dozens of sympatric species that use a variety of microhabitats and have different reproductive modes [15,31]. A diverse host community in Panama with year-round activity and some direct-developing species (i.e. those without an aquatic larval phase) could provide more opportunities for Bd spread [31,60]. Central Panama contains landscape features that may be barriers to dispersal for some amphibian species [61], but interconnected stream networks still allow for fairly high connectivity among sites. By contrast, in the Sierra Nevada, our samples are from the only common—and highly susceptible—amphibian species in the alpine lake habitats (R. sierrae/muscosa) [62]. Rana sierrae/muscosa have high site fidelity, limited overland movements, spend the majority of each year under ice and inhabit disjunct alpine lakes separated by high mountain passes [63]. These features all impede connectivity among host populations and provide fewer opportunities for Bd dispersal [64]. Therefore, landscape and host factors consistently provide decreased opportunities for Bd gene flow in the Sierra Nevada, which is reflected in the greater pathogen spatial structure in this region. In addition, Central Panama is significantly warmer and wetter than the Sierra Nevada. Temperature differences are particularly important because warmer temperatures (to a point) can lead to faster pathogen growth, an increased number of generations per year and greater opportunity for rapid evolutionary change. Slower Bd growth, generation time and evolutionary rates in the Sierra Nevada compared to Central Panama, make the patterns of higher genetic diversity and strong spatial genetic structure in the Sierra Nevada all the more interesting. (b) How can pathogen genetic data help inform wildlife disease mitigation efforts? Ultimately, integrating genetic, spatial and epizootic data within an evolutionary framework is a powerful way to understand the dynamics of emerging diseases of wildlife. Typically, studies of wildlife disease dynamics rely on a priori assumptions about pathogen introductions (i.e. based on earliest infection known from wild populations or museum records). However, our results, using Bd as an example for global pathogens, clearly demonstrates that outbreaks with similar epizoological signatures can still have radically different underlying pathogen histories. In our study, two regions with similar observed epizoological patterns in the field exhibit dramatically different pathogen evolutionary histories. In fact, one of the regions—the Sierra Nevada—has considerable pathogen diversity and genetic structure. Supporting evidence suggests that Bd in this region may persist in populations of highly susceptible host species at very low levels over many years without causing epizootics, opening the possibility that the pathogen has a much longer evolutionary history than previously appreciated. When we treat all population declines as the same, we overlook important nuances that could assist on-the-ground recovery and mitigation efforts. For example, if we incorporate Bd genotype data into choices of donor frog populations when planning translocations and reintroductions, we can mitigate human-induced mixing of Bd genotypes. Such actions could be an important component for species recovery efforts. By combining genetic and epizoological data, we can better understand differences in pathogen invasion history across regions and support more effective policies for biodiversity conservation and management. Supplementary Material Click here for additional data file. Acknowledgements We thank the Sierra Nevada Aquatic Research Laboratory field crew, Danny Boiano from the National Park Service, Matt Robak, Angie Estrada, Renwei Chen and Mary Toothman for assisting in field collection and pathogen qPCR. Ethics All sample collections were authorized by research permits provided by NPS, USFWS and the Institutional Animal Care and Use Committees of UC Berkeley, UC Santa Barbara, University of Nevada, Reno and the University of Pittsburgh. Data accessibility Processed data including BEAST2 tree XML, VCF file, genetic diversity values, tests for temporal signal and code for figures: https://figshare.com/s/0b3bcabff81fae2fb8e8. Raw sequences of new samples deposited in NCBI SRA BioProject ID PRJNA686993: https://www.ncbi.nlm.nih.gov/bioproject/PRJNA686993. Nextstrain code: https://github.com/andrew-rothstein/bd-gpl.git. The data are provided in the electronic supplementary material [65]. Authors' contributions A.P.R.: conceptualization, data curation, formal analysis, investigation, methodology, visualization, writing—original draft, writing—review and editing; A.Q.B.: conceptualization, data curation, formal analysis, investigation, methodology, visualization, writing—original draft, writing—review and editing; R.A.K.: conceptualization, data curation, funding acquisition, investigation, methodology, project administration, supervision, visualization, writing—original draft, writing—review and editing; C.J.B.: conceptualization, data curation, funding acquisition, investigation, project administration, supervision, writing—original draft, writing—review and editing; J.V.: conceptualization, data curation, funding acquisition, investigation, project administration, supervision, writing—original draft, writing—review and editing; C.L.R.: conceptualization, data curation, funding acquisition, investigation, project administration, supervision, writing—original draft, writing—review and editing; E.B.R.: conceptualization, data curation, funding acquisition, investigation, methodology, project administration, supervision, visualization, writing—original draft, writing—review and editing. All authors gave final approval for publication and agreed to be held accountable for the work performed therein. Competing interests We declare we have no competing interests. Funding This work was supported by DoD SERDP contract RC-2638 (to E.B.R., C.L.R.Z., J.V., C.J.B.), NSF DEB-1557190 (to E.B.R., C.J.B., R.A.K.), NSF DEB-1551488 (to E.B.R., C.L.R.Z., J.V.), NSF DEB CAREER - 1846403 (to J.V.), NSF DEB-166311 (to C.L.R.Z.), and the National Park Service. Sequencing done at IBEST Genomics Resources Core at the University of Idaho was supported in part by NIH COBRE grant P30GM103324. All sample collections were authorized by research permits provided by NPS, USFWS and the Institutional Animal Care and Use Committees of UC Berkeley, UC Santa Barbara, University of Nevada—Reno and University of Pittsburgh. ==== Refs References 1. Daszak P, Cunningham AA, Hyatt AD. 2000 Emerging infectious diseases of wildlife—threats to biodiversity and human health. Science 287 , 443-449. (10.1126/science.287.5452.443)10642539 2. Tompkins DM, Carver S, Jones ME, Krkošek M, Skerratt LF. 2015 Emerging infectious diseases of wildlife: a critical perspective. Trends Parasitol. 31 , 149-159. (10.1016/j.pt.2015.01.007)25709109 3. Altizer S, Ostfeld RS, Johnson PT, Kutz S, Harvell CD. 2013 Climate change and infectious diseases: from evidence to a predictive framework. Science 341 , 514-519. (10.1126/science.1239401)23908230 4. Crowl TA, Crist TO, Parmenter RR, Belovsky G, Lugo AE. 2008 The spread of invasive species and infectious disease as drivers of ecosystem change. Front. Ecol. Environ. 6 , 238-246. (10.1890/070151) 5. McCallum H, Dobson A. 2002 Disease, habitat fragmentation and conservation. Proc. R. Soc. Lond. B 269 , 2041-2049. (10.1098/rspb.2002.2079) 6. O'Hanlon SJ et al . 2018 Recent Asian origin of chytrid fungi causing global amphibian declines. Science 360 , 621-627. (10.1126/science.aar1965)29748278 7. Grenfell BT, Pybus OG, Gog JR, Wood JLN, Daly JM, Mumford JA, Holmes EC. 2004 Unifying the epidemiological and evolutionary dynamics of pathogens. Science 303 , 327-332. (10.1126/science.1090727)14726583 8. Volz EM, Pond SLK, Ward MJ, Brown AJL, Frost SDW. 2009 Phylodynamics of infectious disease epidemics. Genetics 183 , 1421-1430. (10.1534/genetics.109.106021)19797047 9. Benton CH, Delahay RJ, Trewby H, Hodgson DJ. 2015 What has molecular epidemiology ever done for wildlife disease research? Past contributions and future directions. Eur. J. Wildl. Res. 61 , 1-16. (10.1007/s10344-014-0882-4) 10. Blanchong JA, Robinson SJ, Samuel MD, Foster JT. 2016 Application of genetics and genomics to wildlife epidemiology. J. Wildl. Manag. 80 , 593-608. (10.1002/jwmg.1064) 11. Stuart SN, Chanson JS, Cox NA, Young BE, Rodrigues ASL, Fischman DL, Waller RW. 2004 Status and trends of amphibian declines and extinctions worldwide. Science 306 , 1783-1786. (10.1126/science.1103538)15486254 12. Wake DB, Vredenburg VT. 2008 Are we in the midst of the sixth mass extinction? A view from the world of amphibians. Proc. Natl Acad. Sci. USA 105 , 11 466-11 473. (10.1073/pnas.0801921105)18172195 13. Longcore JE, Pessier AP, Nichols DK. 1999 Batrachochytrium dendrobatidis gen. et sp. nov., a chytrid pathogenic to amphibians. Mycologia 91 , 219-227. (10.2307/3761366) 14. Voyles J et al . 2009 Pathogenesis of chytridiomycosis, a cause of catastrophic amphibian declines. Science 326 , 582-585. (10.1126/science.1176765)19900897 15. Lips KR et al . 2006 Emerging infectious disease and the loss of biodiversity in a Neotropical amphibian community. Proc. Natl Acad. Sci. USA 103 , 3165-3170. (10.1073/pnas.0506889103)16481617 16. Vredenburg VT, Knapp RA, Tunstall TS, Briggs CJ. 2010 Dynamics of an emerging disease drive large-scale amphibian population extinctions. Proc. Natl Acad. Sci. USA 107 , 9689-9694. (10.1073/pnas.0914111107)20457913 17. Byrne AQ et al . 2019 Cryptic diversity of a widespread global pathogen reveals expanded threats to amphibian conservation. Proc. Natl Acad. Sci. USA 116 , 20 382-20 387. (10.1073/pnas.1908289116)30563855 18. Farrer RA et al . 2011 Multiple emergences of genetically diverse amphibian-infecting chytrids include a globalized hypervirulent recombinant lineage. Proc. Natl Acad. Sci. USA 108 , 18 732-18 736. (10.1073/pnas.1111915108)21173276 19. Rosenblum EB et al . 2013 Complex history of the amphibian-killing chytrid fungus revealed with genome resequencing data. Proc. Natl Acad. Sci. USA 110 , 9385-9390. (10.1073/pnas.1300130110)23650365 20. James TY et al . 2015 Disentangling host, pathogen, and environmental determinants of a recently emerged wildlife disease: lessons from the first 15 years of amphibian chytridiomycosis research. Ecol. Evol. 5 , 4079-4097. (10.1002/ece3.1672)26445660 21. Schloegel LM et al . 2012 Novel, panzootic and hybrid genotypes of amphibian chytridiomycosis associated with the bullfrog trade. Mol. Ecol. 21 , 5162-5177. (10.1111/j.1365-294X.2012.05710.x)22857789 22. Grinnell J, Storer TI. 1924 Animal life in the Yosemite. Berkeley, CA: University of California Press. 23. Vredenburg VT, Bingham R, Knapp R, Morgan JA, Moritz C, Wake D. 2007 Concordant molecular and phenotypic data delineate new taxonomy and conservation priorities for the endangered mountain yellow-legged frog. J. Zool. 271 , 361-374. (10.1111/j.1469-7998.2006.00258.x) 24. Vredenburg VT et al . 2019 Pathogen invasion history elucidates contemporary host pathogen dynamics. PLoS ONE 14 , e0219981. (10.1371/journal.pone.0219981)31536501 25. Zhou H, Hanson T, Knapp R. 2015 Marginal Bayesian nonparametric model for time to disease arrival of threatened amphibian populations. Biometrics 71 , 1101-1110. (10.1111/biom.12345)26148536 26. Bradford DF. 1991 Mass mortality and extinction in a high-elevation population of Rana muscosa. J. Herpetol. 25 , 174-177. (10.2307/1564645) 27. Rachowicz LJ, Knapp RA, Morgan JAT, Stice MJ, Vredenburg VT, Parker JM, Briggs CJ. 2006 Emerging infectious disease as a proximate cause of amphibian mass mortality. Ecology 87 , 1671-1683. (10.1890/0012-9658(2006)87[1671:EIDAAP]2.0.CO;2)16922318 28. Knapp RA, Fellers GM, Kleeman PM, Miller DAW, Vredenburg VT, Rosenblum EB, Briggs CJ. 2016 Large-scale recovery of an endangered amphibian despite ongoing exposure to multiple stressors. Proc. Natl Acad. Sci. USA 113 , 11 889-11 894. (10.1073/pnas.1600983113) 29. Lips KR. 1998 Decline of a tropical montane amphibian fauna. Conserv. Biol. 12 , 106-117. (10.1046/j.1523-1739.1998.96359.x) 30. Pounds JA, Fogden MP, Savage JM, Gorman GC. 1997 Tests of null models for amphibian declines on a tropical mountain. Conserv. Biol. 11 , 1307-1322. (10.1046/j.1523-1739.1997.95485.x) 31. Crawford AJ, Lips KR, Bermingham E. 2010 Epidemic disease decimates amphibian abundance, species diversity, and evolutionary history in the highlands of central Panama. Proc. Natl Acad. Sci. USA 107 , 13 777-13 782. (10.1073/pnas.0914115107) 32. Voyles J et al . 2018 Shifts in disease dynamics in a tropical amphibian assemblage are not due to pathogen attenuation. Science 359 , 1517-1519. (10.1126/science.aao4806)29599242 33. Whiles MR et al . 2013 Disease-driven amphibian declines alter ecosystem processes in a tropical stream. Ecosystems 16 , 146-157. (10.1007/s10021-012-9602-7) 34. Zipkin EF, DiRenzo GV, Ray JM, Rossman S, Lips KR. 2020 Tropical snake diversity collapses after widespread amphibian loss. Science 367 , 814-816. (10.1126/science.aay5733)32054766 35. Lips KR, Diffendorfer J, Mendelson III JR, Sears MW. 2008 Riding the wave: reconciling the roles of disease and climate change in amphibian declines. PLoS Biol. 6 , e72. (10.1371/journal.pbio.0060072)18366257 36. Collins JP, Storfer A. 2003 Global amphibian declines: sorting the hypotheses. Divers. Distrib. 9 , 89-98. (10.1046/j.1472-4642.2003.00012.x) 37. Knapp RA, Briggs CJ, Smith TC, Maurer JR. 2011 Nowhere to hide: impact of a temperature-sensitive amphibian pathogen along an elevation gradient in the temperate zone. Ecosphere 2 , 1-26. (10.1890/ES11-00028.1) 38. Puschendorf R, Carnaval AC, VanDerWal J, Zumbado-Ulate H, Chaves G, Bolaños F, Alford RA. 2009 Distribution models for the amphibian chytrid Batrachochytrium dendrobatidis in Costa Rica: proposing climatic refuges as a conservation tool. Divers. Distrib. 15 , 401-408. (10.1111/j.1472-4642.2008.00548.x) 39. Olson DH et al . 2013 Mapping the global emergence of Batrachochytrium dendrobatidis, the amphibian chytrid fungus. PLoS ONE 8 , e56802. (10.1371/journal.pone.0056802)23463502 40. Morgan JAT et al . 2007 Population genetics of the frog-killing fungus Batrachochytrium dendrobatidis. Proc. Natl Acad. Sci. USA 104 , 13 845-13 850. (10.1073/pnas.0701838104) 41. Farrer RA, Henk DA, Garner TWJ, Balloux F, Woodhams DC, Fisher MC. 2013 Chromosomal copy number variation, selection and uneven rates of recombination reveal cryptic genome diversity linked to pathogenicity. PLoS Genet. 9 , e1003703. (10.1371/journal.pgen.1003703)23966879 42. Byrne AQ, Rothstein AP, Poorten TJ, Erens J, Settles ML, Rosenblum EB. 2017 Unlocking the story in the swab: a new genotyping assay for the amphibian chytrid fungus Batrachochytrium dendrobatidis. Mol. Ecol. Resour. 17 , 1283-1292. (10.1111/1755-0998.12675)28371384 43. Jombart T. 2008 adegenet: a R package for the multivariate analysis of genetic markers. Bioinformatics 24 , 1403-1405. (10.1093/bioinformatics/btn129)18397895 44. Korneliussen TS, Albrechtsen A, Nielsen R. 2014 ANGSD: analysis of next generation sequencing data. BMC Bioinf. 15 , 356. (10.1186/s12859-014-0356-4) 45. Bodenhofer U, Bonatesta E, Horejš-Kainrath C, Hochreiter S. 2015 msa: an R package for multiple sequence alignment. Bioinformatics 31 , 3997-3999. (10.1093/bioinformatics/btv494)26315911 46. Nguyen L-T, Schmidt HA, Von Haeseler A, Minh BQ. 2015 IQ-TREE: a fast and effective stochastic algorithm for estimating maximum-likelihood phylogenies. Mol. Biol. Evol. 32 , 268-274. (10.1093/molbev/msu300)25371430 47. Rambaut A, Lam TT, Max Carvalho L, Pybus OG. 2016 Exploring the temporal structure of heterochronous sequences using TempEst (formerly Path-O-Gen). Virus Evol. 2 , vew007. (10.1093/ve/vew007)27774300 48. Duchêne S, Duchêne D, Holmes EC, Ho SY. 2015 The performance of the date-randomization test in phylogenetic analyses of time-structured virus data. Mol. Biol. Evol. 32 , 1895-1906. (10.1093/molbev/msv056)25771196 49. Bouckaert R, Heled J, Kühnert D, Vaughan T, Wu C-H, Xie D, Suchard MA, Rambaut A, Drummond AJ. 2014 BEAST 2: a software platform for Bayesian evolutionary analysis. PLoS Comput. Biol. 10 , e1003537. (10.1371/journal.pcbi.1003537)24722319 50. Hadfield J, Megill C, Bell SM, Huddleston J, Potter B, Callender C, Sagulenko P, Bedford T, Neher RA. 2018 Nextstrain: real-time tracking of pathogen evolution. Bioinformatics 34 , 4121-4123. (10.1093/bioinformatics/bty407)29790939 51. Rambaut A, Drummond AJ, Xie D, Baele G, Suchard MA. 2018 Posterior summarization in Bayesian phylogenetics using Tracer 1.7. Syst. Biol. 67 , 901-904. (10.1093/sysbio/syy032)29718447 52. Berger L et al . 1998 Chytridiomycosis causes amphibian mortality associated with population declines in the rain forests of Australia and Central America. Proc. Natl Acad. Sci. USA 95 , 9031-9036. (10.1073/pnas.95.15.9031)9671799 53. Walker SF et al . 2010 Factors driving pathogenicity vs. prevalence of amphibian panzootic chytridiomycosis in Iberia. Ecol. Lett. 13 , 372-382. (10.1111/j.1461-0248.2009.01434.x)20132274 54. Gillespie G, Hunter D, Berger L, Marantelli G. 2015 Rapid decline and extinction of a montane frog population in southern Australia follows detection of the amphibian pathogen Batrachochytrium dendrobatidis. Anim. Conserv. 18 , 295-302. (10.1111/acv.12174) 55. Velo-Antón G, Rodríguez D, Savage AE, Parra-Olea G, Lips KR, Zamudio KR. 2012 Amphibian-killing fungus loses genetic diversity as it spreads across the New World. Biol. Conserv. 146 , 213-218. (10.1016/j.biocon.2011.12.003) 56. De León ME, Zumbado-Ulate H, García-Rodríguez A, Alvarado G, Sulaeman H, Bolaños F, Vredenburg VT. 2019 Batrachochytrium dendrobatidis infection in amphibians predates first known epizootic in Costa Rica. PLoS ONE 14 , e0208969. (10.1371/journal.pone.0208969)31821326 57. Puschendorf R, Bolaños F, Chaves G. 2006 The amphibian chytrid fungus along an altitudinal transect before the first reported declines in Costa Rica. Biol. Conserv. 132 , 136-142. (10.1016/j.biocon.2006.03.010) 58. Jani AJ, Knapp RA, Briggs CJ. 2017 Epidemic and endemic pathogen dynamics correspond to distinct host population microbiomes at a landscape scale. Proc. R. Soc. B 284 , 20170944. (10.1098/rspb.2017.0944) 59. Kamoroff C, Goldberg C. 2017 Using environmental DNA for early detection of amphibian chytrid fungus Batrachochytrium dendrobatidis prior to a ranid die-off. Dis. Aquat. Org. 127 , 75-79. (10.3354/dao03183) 60. Paz A, Ibáñez R, Lips KR, Crawford AJ. 2015 Testing the role of ecology and life history in structuring genetic variation across a landscape: a trait-based phylogeographic approach. Mol. Ecol. 24 , 3723-3737. (10.1111/mec.13275)26080899 61. Richards-Zawacki CL. 2009 Effects of slope and riparian habitat connectivity on gene flow in an endangered Panamanian frog, Atelopus varius. Divers. Distrib. 15 , 796-806. (10.1111/j.1472-4642.2009.00582.x) 62. Zweifel RG. 1955 Ecology, distribution and systematics of frogs of the Rana boylei group. Univ. California Pub. Zool. 54 , 207-292. 63. Pope KL, Matthews KR. 2001 Movement ecology and seasonal distribution of mountain yellow-legged frogs, Rana muscosa, in a high-elevation Sierra Nevada basin. Copeia 2001 , 787-793. (10.1643/0045-8511(2001)001[0787:MEASDO]2.0.CO;2) 64. Rothstein AP, Knapp R, Bradburd G, Boiano D, Briggs CJ, Rosenblum EB. 2020 Stepping into the past to conserve the future: archived skin swabs from extant and extirpated populations inform genetic management of an endangered amphibian. Mol. Ecol. 29 , 2598-2611. (10.1111/mec.15515)32573039 65. Rothstein AP, Byrne AQ, Knapp RA, Briggs CJ, Voyles J, Richards-Zawacki CL, Rosenblum EB. 2021 Data from: Divergent regional evolutionary histories of a devastating global amphibian pathogen. Figshare.
PMC008xxxxxx/PMC8220267.txt
==== Front Proc Biol Sci Proc Biol Sci RSPB royprsb Proceedings of the Royal Society B: Biological Sciences 0962-8452 1471-2954 The Royal Society 34157876 10.1098/rspb.2021.0464 rspb20210464 100160144Palaeobiology Research Articles Vertically migrating Isoxys and the early Cambrian biological pump Vertically migrating Isoxys and the Early Cambrian biological pump http://orcid.org/0000-0001-8063-9469 Pates Stephen Conceptualization Data curation Formal analysis Funding acquisition Investigation Methodology Project administration Resources Software Validation Visualization Writing-original draft Writing-review & editing [email protected] 1 2 http://orcid.org/0000-0001-5369-5879 Daley Allison C. Writing-review & editing 3 Legg David A. Data curation Writing-review & editing 4 http://orcid.org/0000-0001-6598-6534 Rahman Imran A. Methodology Validation Writing-review & editing 5 1 Museum of Comparative Zoology and Department of Organismic and Evolutionary Biology, Harvard University, Cambridge, MA, USA 2 Department of Zoology, University of Cambridge, Cambridge, UK 3 ISTE, University of Lausanne, Lausanne, Vaud, Switzerland 4 Faculty of Science and Engineering, University of Manchester, Manchester, UK 5 Oxford University Museum of Natural History, University of Oxford, Oxford, UK 30 6 2021 June 30, 2021 23 6 2021 June 23, 2021 23 6 2021 June 23, 2021 288 1953 2021046424 2 2021 Feburary 24, 2021 3 6 2021 June 3, 2021 © 2021 The Authors. 2021 https://creativecommons.org/licenses/by/4.0/ Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited. The biological pump is crucial for transporting nutrients fixed by surface-dwelling primary producers to demersal animal communities. Indeed, the establishment of an efficient biological pump was likely a key factor enabling the diversification of animals over 500 Myr ago during the Cambrian explosion. The modern biological pump operates through two main vectors: the passive sinking of aggregates of organic matter, and the active vertical migration of animals. The coevolution of eukaryotes and sinking aggregates is well understood for the Proterozoic and Cambrian; however, little attention has been paid to the establishment of the vertical migration of animals. Here we investigate the morphological variation and hydrodynamic performance of the Cambrian euarthropod Isoxys. We combine elliptical Fourier analysis of carapace shape with computational fluid dynamics simulations to demonstrate that Isoxys species likely occupied a variety of niches in Cambrian oceans, including vertical migrants, providing the first quantitative evidence that some Cambrian animals were adapted for vertical movement in the water column. Vertical migration was one of several early Cambrian metazoan innovations that led to the biological pump taking on a modern-style architecture over 500 Myr ago. biological pump , Cambrian , computational fluid dynamics , Isoxys , pelagic University of Cambridge http://dx.doi.org/10.13039/501100000735 University of Manchester http://dx.doi.org/10.13039/501100000770 Oxford University Museum of Natural History Harvard University http://dx.doi.org/10.13039/100007229 cover-dateJune 30, 2021 ==== Body pmc1. Introduction The biological pump is the process by which organic nutrients are transported from shallow ocean to deep sea [1]. Today this consists of two major vectors: passive sinking of organic matter aggregates and vertical movement of animals [1] (figure 1). The biological pump, the main driver of the marine carbon cycle, is responsible for approximately two-thirds of the vertical gradient of carbon in the ocean [2]. While transport of carbon to the deep ocean through vertical mixing of dissolved organic carbon (DOC) carries significant amounts of carbon to depth, 95% of DOC cannot be used as food by marine organisms [1,3]. By contrast, aggregates and vertical migrants concentrate organic matter in a form that can be used by demersal animals [1] and are therefore crucial for the establishment and sustenance of deep-water communities. Figure 1. Simplified architecture of the modern biological pump. Arrows trace pathway of carbon (C) and nitrogen (N) and other nutrients. Numbers indicate two main biologically mediated vectors that transport nutrients fixed by phytoplankton from the surface ocean to demersal communities. (1) Aggregation vector of phytoplankton, faecal pellets and other organic matter which sinks passively through the water column. (2) Vertical migration vector driven by active two-way migration by metazoans. (Online version in colour.) The biological pump was very different before the Cryogenian Period (more than 720 Ma). Cyanobacterial picoplankton (0.2–2.0 µm) domination led to a stratified and turbid water column (e.g. [4–6]). Cells were too small to sink in significant numbers, so most nutrients were recycled within the surface waters, with little export to the deeper ocean [4–6]. The flux of the aggregation vector (figure 1) would have increased when primary productivity shifted to be eukaryote-dominated during the Cryogenian ∼ 650 Ma [7], and further when Cambrian suspension feeding sponges and pelagic phytoplanktivores applied a size-selective pressure for larger primary producers [8–12]. The carcasses, sloppy feeding and faecal pellets of macrozooplankton, including phytoplanktivorous crown-group crustaceans (e.g. [9]), would have further increased the formation and size of aggregates, and thus the flux along this vector. However, despite the importance of vertical migration in the transport of nutrients necessary to sustain mesopelagic and deep-water ecosystems in the modern ocean (e.g. [13]), this vector is poorly understood in the Cambrian. Comparisons of the morphology of Cambrian fossil organisms with modern vertically mobile pelagic animals provides the opportunity to infer whether vertical migration occurred in oceans over 500 Ma. Qualitative comparisons between the pelagic crustacean Gnathophausia and the nektonic stem group euarthropod [14] Isoxys (e.g. [15,16]) suggest the latter is a promising candidate. The 20 Isoxys species so far formally described are united by the presence of a bivalved carapace which bears both anterior and posterior spines [17] (figure 2), and the genus has a cosmopolitan distribution (e.g. [18]). The presence of eyes that can comprise approximately 10% of the body length, a digestive tract with paired serial midgut glands, and a pair of anteriorly positioned raptorial appendages (figure 2) support a predatory habit for Isoxys, which would have been well suited for capturing small soft-bodied invertebrates [19]. Isoxys is unusual for Cambrian animals as a pelagic lifestyle has been proposed (e.g. [18]), although some recent workers have suggested a potential hyperbenthic lifestyle (1–10 m above the bottom), with individuals capable of moving small distances vertically [19,20]. However, while Isoxys carapaces appear to show adaptations for hydrodynamic streamlining, interspecific differences in both carapace asymmetry and spine lengths (e.g. figure 2), as well as soft parts, suggest that different species may have occupied distinct niches, including some much closer to the seafloor [21]. Figure 2. Morphology of Isoxys. (a) line drawing of idealized Isoxys illustrating known soft parts. (b) YPM IP 005804, Isoxys acutangulus from the Burgess Shale, British Columbia, Canada (Cambrian: Wuliuan) (credit: W. K. Sacco). (c) USNM PAL 189170, Isoxys longissimus from the Burgess Shale, British Columbia, Canada (Cambrian: Wuliuan). Image courtesy of the Smithsonian Institution (EZID: http://n2t.net/ark:/65665/m372f2a644-97c3-441c-87e2-24b1dccb2e8c, credit: Xingliang Zhang). Abbreviations: as, anterior spine; en, endopod; ex, flap like exopod; ey, eye; g, gut with paired diverticulae; ps, posterior spine; ra, raptorial appendage. (Online version in colour.) Here, we provide the first quantitative assessment of carapace shape variation across the group and compare Isoxys to other Cambrian ‘bivalved’ euarthropods and Gnathophausia. Subsequently, through computational fluid dynamics (CFD) simulations, we test the importance of the spines and carapace outline for generating lift and reducing drag, and thus the ability of different taxa to move vertically through the water column. These analyses support the hypothesis that Isoxys taxa occupied a variety of distinct niches in Cambrian oceans, including vertical migrants. 2. Material and methods (a) Outline analyses Two-dimensional outlines of 20 Isoxys species, its sister taxon Surusicaris elegans [22], 6 Tuzoia species and 11 Gnathophausia species were constructed in Inkscape from the literature sources by S.P. (electronic supplementary material, table S1) and imported into R [23] for elliptical Fourier analysis (EFA). D.A.L. independently constructed 20 Isoxys and 1 Surusicaris outlines directly from photographs of fossil specimens to allow assessment of the error introduced in the creation of outlines (electronic supplementary material, table S2; and S1). Tuzoia was chosen for comparison as it is also common in Cambrian communities, and has been suggested to be closely related to Isoxys based on similarities in the structure of the eyes and carapace shape (e.g. [24]). Gnathophausia was selected because similarities in the carapace morphology of one species (G. zoea) and Isoxys have been repeatedly noted (e.g. [16,19]), the carapaces of these animals are similar in size (10–30 mm), and multiple species of Gnathophausia are known to be vertically mobile in the modern ocean, having been sampled from surface waters and at depths of over 3000 m [25]. Outlines were sampled at the same resolution (64 points provided sufficient detail to distinguish taxa), centred, scaled by centroid size and subjected to EFA using the Momocs package [26]. Harmonics describing 99.9% of the variation were retained. EFA results were visualized with a principal components analysis. A hierarchical clustering analysis (cluster package; [23]) quantitatively grouped similarly shaped carapaces together using all principal components. (b) Computational fluid dynamics Isoxys species reflecting the variation in carapace shape over both PC1 and PC2 were chosen for inclusion in CFD simulations, and Gnathophausia zoea was analysed for comparison. We chose to undertake a two-dimensional analysis to conserve computational resources and minimize errors in the modelled geometries (the exact three-dimensional shape is unknown for most taxa). This is justified because undeformed Isoxys specimens preserved in dorsal view show a narrow profile [16,21,27,28]. Propulsion during swimming derived from movement of the ventral appendages, and not from flapping of the bivalve carapace [15,27]. A lack of adductor muscles means that Isoxys was unable to alter the size of the gape during swimming [15], and the numerous specimens preserved in ‘butterfly’ orientation are considered exuviae [27]. In addition, the full variation of the carapaces considered in this study can be visualized in two dimensions. Two-dimensional analyses are commonly performed on analyses of wing outlines to assess aerodynamic performance by both biologists and engineers, (e.g. [29,30]). A two-dimensional analysis is suitable for this study as the wake behind the Isoxys carapaces is steady at the Reynolds numbers considered, so can be modelled with two-dimensional simulations [31,32]. The impact of soft parts such as eyes protruding from the carapace was analysed for two taxa: Isoxys acutangulus and I. longissimus. Isoxys was assumed to be negatively buoyant, like Gnathophausia [33]. All Isoxys species were assumed to have the same carapace composition and density. The cuticle ornamentation in some Isoxys species is not expected to impact the drag at the low Reynolds numbers considered in this study, as the roughened surface falls within the slowly moving fluid near the carapace surface [31]. Following validation and verification of the model and set-up for low Reynolds numbers, and mesh quality assessments using ANSYS Mesh and ANSYS Fluent (Ansys Academic, Release 2020 R2; electronic supplementary material, S2), outlines of the selected Isoxys species and G. zoea were exported from R as .txt files readable by ANSYS DesignModeler (Ansys Academic, Release 2020 R2; electronic supplementary material, S2), and standardized to a dorsal length (chord length) of 25 mm. This allowed size-independent comparisons of hydrodynamic performance of shapes. While some Isoxys taxa (e.g. I. communis, I. longissimus) can reach up to 50 mm, other adult forms only reach approximately 20 mm (e.g. I. glaessneri, I. volucris) [18,27]. A size of 25 mm represents a compromise size for comparison between these larger and smaller forms, with the influence of larger size able to be assessed by simulating higher Re (as Re depends on both size and swimming speed). CFD simulations were conducted using the steady-state laminar solver in ANSYS Fluent (Ansys Academic, Release 2020 R2). The laminar solver performed best of the three considered during validation and verification (laminar, SST, k-epsilon; electronic supplementary material, S2), as expected for the low Reynolds numbers in this study [31]. Coefficients of drag (Cd) and lift (Cl) (electronic supplementary material, S2) were calculated under three flow speeds equating to 0.75, 1.00 and 1.18 body lengths per second (chord Reynolds numbers, Re, 255, 340 and 400 respectively for saltwater conditions at 0°C; electronic supplementary material, S2). These Re were chosen as swimming speeds of between 75% and 100% of the body length per second have been observed in adult Gnathophausia ingens (carapace mean length approx. 25 mm [34]), and as the laminar model was validated against published drag and lift data for NACA aerofoils at exactly Re = 400 [29]. The chord length (25 mm) was taken as the reference area for both coefficients. Solutions were considered converged when residuals were less than 10–6. Simulations were run at numerous angles of attack, to evaluate the hydrodynamic performance of carapaces at multiple orientations. In tank experiments, Gnathophausia ingens has been observed to change angle of attack to generate more lift or less drag at different swimming speeds [34]. The angle of attack was increased from 0 to 8° at all Re, until the stall angle could be identified and/or unsteady flow was observed. If the stall angle was not reached, further experiments were run until the maximum lift coefficient was obtained. Negative angles of attack were also simulated to assess the negative lift generated by the different outlines. In all cases, the absolute value of the negative angle of attack was increased until the drag coefficient was equal to or greater than that of the stall angle. When unsteady flow was suspected to be the reason that steady-state simulations did not converge, the steadiness of the flow field was determined by carrying out a time-dependent analysis of 100 time steps, with each time step equal to the flow speed (so, for an inlet velocity of 0.01875 ms–1, the time step = 0.01875 s). 3. Results (a) Outline analysis In the outline analysis, a total of 18 harmonics were retained. Principal coordinates 1 and 2 described 79.3% of the total variation. Carapace asymmetry, narrowness and spine length increased as PC1 became more positive, while the length of the anterior spine relative to the posterior spine corresponded to an increase in PC2 (figure 3). Isoxys occupied the largest area in the morphospace. Visual overlap of the areas occupied by the genera demonstrated that some Isoxys taxa were more similar in shape to Gnathophausia than their Cambrian relatives Surusicaris and Tuzoia. Confirmation was provided by the cluster analysis (figure 3). All six Tuzoia species formed a single cluster, with Isoxys taxa spread over the three remaining groups (Surusicaris group, Gnathophausia groups 1 and 2). Species clustered with Surusicaris have symmetric and deep carapaces and relatively short spines. Species in the Gnathophausia groups displayed narrower outlines whose narrowness, asymmetry and spine length increased with PC1. The single species of Isoxys in Gnathophausia group 1, I. paradoxus, displayed an anterior spine much more elongate than its posterior one, that contrasted with the seven species in Gnathophausia group 2 with their spines of approximately equal length. Figure 3. Principal component analysis of results of EFA conducted on the outlines of 11 species of Gnathophausia, 20 Isoxys, 1 Surusicaris and 6 Tuzoia. Convex hulls indicate optimum four groupings as recovered by clustering analysis. Labelled Isoxys species chosen for subsequent hydrodynamic analysis. (Online version in colour.) (b) Computational fluid dynamics CFD simulations assessed the impact of increasing asymmetry, spine length and relative lengths of anterior and posterior spines on hydrodynamic performance. Inclusion of eyes did not significantly impact the hydrodynamic performance of carapaces (electronic supplementary material, S3). Isoxys zhurensis, the most symmetric species chosen for analysis, created an unsteady wake with Kármán vortex street at the lowest Re considered, and so no drag or lift coefficients were obtained (electronic supplementary material video). The flow around remaining Isoxys outlines was laminar with a steady wake, and there was no evidence for three dimensionality (electronic supplementary material, S4). Greater asymmetry and narrowness (more positive in PC1, figure 3) resulted in lower drag coefficients, as demonstrated by a comparison of the short-spined taxa I. chilhoweanus, I. acutangulus and I. mackenziensis. The most asymmetric of these forms, I. mackenziensis, produced lower drag coefficients than the other two, but ranges of lift coefficients were similar for all three (figure 4a). More elongate spines increased the range of lift coefficients (vertical bars, figure 4a) and, significantly, negative lift coefficients at negative angles of attack (e.g. compare I. mackenziensis, I. communis and I. longissimus). In I. paradoxus, where the anterior spine is much longer than the posterior one (more positive in PC2, figure 3), the range of lift coefficients further increased (figure 4a). Similar drag coefficients and ranges of lift coefficients were obtained in an analysis of the hydrodynamics of Gnathophausia zoea, when either the carapace alone or both the carapace and abdomen were considered (figure 4b). Figure 4. Drag polars (plot of Cd against Cl) of taxa analysed at Re = 255 (0.75 body lengths per second for an animal 25 mm long). Each point corresponds to a single simulation at a different angle of attack. Vertical bars show range of lift coefficients. Note that flow was unsteady for Isoxys zhurensis at Re = 255, and so no quantitative lift or drag coefficients were recorded. Drag polars at faster flow speeds and raw data presented in electronic supplementary material, S3. (Online version in colour.) 4. Discussion (a) Vertical migrations and niche partitioning in Isoxys Functional morphology of Isoxys fossil specimens supports an off-bottom (hyperbenthic or pelagic) life habit for this animal [15,16,18,19,32], based on the eye orientation (forwards, slightly ventral) and the elongate slender carapace shape of Isoxys. Our study combines outline morphometric and CFD analyses and suggests that Isoxys species occupied a variety of niches, including some as pelagic vertically mobile predators. Lift and drag coefficients of Isoxys carapaces indicate variation in the depth range and swimming speeds of these species. Isoxys taxa clustering with Surusicaris (figure 3) generate positive lift, but do not generate negative lift (figure 4). This supports suggestions of previous workers that these Isoxys species may have occupied a hyperbenthic (1–10 m above the seafloor) and/or nektobenthic [21] niche, perhaps moving vertically short distances in the water column [15,19,35]. Vertical movement would be achieved by altering the angle of attack to produce lift force greater than (ascent), equal to (horizontal swimming), or less than (descent) the impact of their negative buoyancy. Drag reduction associated with streamlining would have allowed some taxa (e.g. I. mackenziensis) to capture faster-moving prey animals, as faster swimming speeds could have been maintained over longer distances for the same metabolic cost. Isoxys species clustering with Gnathophausia show convergent adaptations to moving vertically in the water column, such as asymmetric carapaces with elongate anterior and posterior spines (figures 3 and 4). This does not preclude elongate spines from also acting as anti-predatory deterrents, as suggested by Vannier & Chen [15]. These adaptations provided hydrodynamic benefits that would have allowed the Isoxys species to operate over a wider bathymetric range. A streamlined carapace facilitates not only faster movement, but also more efficient swimming, beneficial for migrations over a long distance. The carapace shapes of I. longissimus and I. paradoxus generate lift coefficient ranges and minimum drag coefficients comparable to the modern crustacean Gnathophausia zoea (figure 4), which has been recovered at depths ranging from surface waters down to approximately 3000 m in the modern ocean [25]. These results also suggest that an elongate abdomen in G. zoea reduces the drag experienced by the animal slightly but does not greatly impact on the range of lift coefficients (figure 4b), though the abdomen may also play a physical role. Animals that move vertically in the water column do not have to cover the entire distance from the surface ocean to demersal communities, and instead sometimes migrate across only a shorter vertical distance. Thus, Isoxys taxa with the broadest ranges of lift coefficients (I. longissimus and I. paradoxus) probably covered a wider depth range than those with smaller ranges of lift coefficients (e.g. I. communis). Corroborating evidence for variation in bathymetric range for different Isoxys species comes from the fossil record itself. Members of different groups as resolved in the cluster analysis (convex hulls, figure 3) co-occur with different relative abundances in Cambrian deposits preserving soft-bodied fossils. In general, species with inferred vertically migrating lifestyles are much rarer than those that lived close to the seafloor. In the Chengjiang Biota, Isoxys auritus (Surusicaris group) greatly outnumbers both I. paradoxus and I. curvirostratus (Gnathophausia groups 1 and 2, respectively) (figure 3; [15,21,36–38]). A similar pattern of relative abundances can be observed in the two Burgess Shale taxa (figure 2): in the Walcott Quarry, I. acutangulus (Surusicaris group) comprises nearly 0.5% of the total community, vastly outnumbering the extremely rare inferred vertical migrant I. longissimus (Gnathophausia group 2 [16,39]). The relative abundances of these Isoxys species can be partly explained by the differences in lifestyle predicted by the carapace outline and soft anatomy. The less hydrodynamic taxa (those with higher drag coefficients and narrower ranges of lift coefficients) probably lived near the seafloor, with the more streamlined species living in the water column and occupying a broader bathymetric range. This broader bathymetric range would have included more open water settings, beyond the maximum depth of the shelf where Cambrian deposits preserving soft-bodied fossils occur—Gnathophausia zoea for example has been found at depths of up to 3000 m [25]. As modern euarthropod carapaces disarticulate quickly after death (e.g. [40]), the preservation potentials for pelagic euarthropods living high in the water column are lower than for those living closer to the seafloor. In addition, animals which occupy an ecological niche in the open water are less likely to find themselves over shelf environments such as those which preserve soft-bodied fossils or be trapped and transported by an obrution event responsible for the preservation of soft-bodied communities in these settings. The small numbers of vertically mobile Isoxys individuals observed may have been at the bottom of their vertical migrations and/or been transported horizontally by currents. Isoxys species are not globally distributed [41]. Many species (for example those clustering with Surusicaris) appear suited to hyperbenthic habits, and so would be expected to have provincial distributions. The limited geographical distribution of I. longissimus and I. paradoxus is most likely to be due to a combination of factors. Firstly, deposits where Isoxys is expected to be preserved are not evenly distributed in time and space—Stage 3 deposits are mostly in South China, while Wuliuan and younger are mostly in Laurentia [42], though the absence of the Chengjiang species I. paradoxus from Sirius Passet is notable. Secondly, the lower preservation potential of pelagic (compared to hyperbenthic) species means that they are rare even in Tier 1 Burgess Shale-type Lagerstätten (sensu [42]). However, despite its rarity, the Burgess Shale species I. longissimus has a wider known geographical range than the co-occurring I. acutangulus. The former has also been reported from the Wheeler Formation, House Range, UT, USA [43]. The situation appears more complex in the Emu Bay Shale, where the more hydrodynamic species I. communis greatly outnumbers the less streamlined I. glaessneri [27]. However, the Emu Bay Shale is not a traditional Burgess Shale-type deposit, instead it represents a localized deep-water micro basin on the inner shelf [44]. Here fluctuating oxygen levels may have periodically deoxygenated the water column, possibly killing pelagic taxa like I. communis in great numbers and creating a taphonomic bias that preferentially preserves pelagic taxa. Further support for the Chengjiang taxon Isoxys auritus occupying a niche closer to the seafloor than I. curvirostratus comes from a comparison of the soft anatomy (soft parts are unknown for I. paradoxus) [21]. The stout endopods of I. auritus appear well suited for interacting with the substrate, while exopods with broad fringing lamellae and a sophisticated vascular system in the more streamlined I. curvirostratus suggest it was a more powerful swimmer, providing additional support for a pelagic habit [21]. Lastly, a compendium of fossil, geochemical and phylogenetic data show that vertically mobile Isoxys species would have had access to a variety of pelagic prey items and an oxygenated water column to travel in. Cambrian oceans were not stratified, instead displaying wedge-shaped oxygen minimum zones broadly comparable to modern oceans [45]. Isoxys prey size range (approx. 5–20 mm [19]) includes Cambrian phytoplanktivores, such as crustaceans with setae and filter plates from Sirius Passet and Mount Cap (15–50 mm) [4,8,9,12], as well as crown-group branchiopods and copepods and total group ostracods [46], as well as bradoriids, some of which were also likely pelagic [47,48]. The presence of planktonic larvae, another possible prey item, can be inferred from tip-dated phylogenetic analyses that support the evolution of metamorphosis in euarthropods by the Cambrian [49]. Thus, a range of different data sources suggest multiple Isoxys taxa were vertically mobile, and that the Cambrian ocean could support such an ecology. (b) Metazoans and the Cambrian biological pump The presence of the likely vertically mobile Isoxys paradoxus in the Chengjiang Biota makes it the oldest confidently identified euarthropod vertical migrant, and probably among the first animals to employ this life habit. For this vector to be significant by the Cambrian Stage 3, a large biomass of Isoxys would need to move vertically. While pelagic animals have a lower preservation potential than benthic ones (for example, very few fossil copepods are known [50]), Isoxys species with inferred (hyper)benthic habits are extremely abundant in both the Chengjiang and Burgess Shale [15,16,21,36–38]), and their pelagic counterparts Isoxys longissimus and I. paradoxus may have been similarly numerous. The Chengjiang and coeval Qingjiang biotas also preserve the earliest evidence for gelatinous zooplankton, which move vertically small distances in the modern ocean [51,52]. However, an active swimming Isoxys would have covered greater distances more rapidly. Furthermore, the presence of a through gut would have increased processing time for food, vital for transporting nutrients consumed in surface waters to the deep ocean as faecal pellets. Vertical migration was one of many important eukaryotic and metazoan innovations key to establishing a modern-style biological pump. A series of metazoan innovations which appear in the fossil record in quick succession during the early Cambrian gave the biological pump a modern-looking structure (figure 5), which was strengthened during the Phanerozoic. Figure 5. First known appearance in the fossil record of metazoans that impacted the biological pump. Circle colour denotes preservation style. Fossils with more than one preservation style indicated with split circles, with the preservation style that provided the oldest evidence on the left of the circle. Abbreviations: Drum., Drumian; Guz., Guzhangian; Wul., Wuliuan. (Online version in colour.) The shift to primary production dominated by eukaryotes, and the innovation of active suspension feeding, would have ventilated the oceans, cleared organic matter in the water column, and increased transfer of oxygen and plankton of increasing sizes from the surface to the sediment–water interface (e.g. [6]). The first of these events occurred during the Cryogenian [7], but while benthic passive suspension feeders are known from the Ediacaran (e.g. [53]), early Cambrian stem and crown-group sponges represent the oldest benthic active suspension feeders [10,11]. As active suspension feeders, sponges were able to transport large volumes of water between benthic and pelagic realms (e.g. [54]), and the most abundant sponges from the Cambrian, the reef building archaeocyaths, display pore size differentiation within reef systems [55], presumed to be evidence of prey size-selectivity. This illustrates that predator–prey feedbacks were present in the Cambrian Stage 2 (figure 2), presenting a potential driver towards a larger size of plankton, increasing the sinking speed and efficiency of this part of the biological pump, and ventilating the water column. The invasion of the pelagic realm by eumetazoan zooplankton provides the next step towards a modern-style biological pump. These zooplankton would have further cleared surface waters and contributed faecal pellets to organic aggregates sinking to the deep ocean, and also increased oxygen levels at depth (e.g. [6,54]). The small shelly fossil (SSF) record provides a source of evidence for the invasion of the pelagic realm by eumetazoans. Terreneuvian SSFs include the possible chaetognath Protoherzina and the molluscs Watsonella, Aldanella and Oelandiella, whose widespread distributions are suggestive of a pelagic lifestyle, or at least a planktonic larval stage (figure 5) [56]. Euarthropods, likely early occupants of the plankton [35], are represented in the SSF record from the Cambrian Stage 3 by millimetre-scale bradoriids among others (figure 5; e.g. [47,57,58]). The first macroscopic nektonic suspension feeders, such as the radiodont Tamisiocaris, also appear at this time [59], while the first centimetre-scale phytoplanktivores are identified close to the Stage 3–4 boundary (figure 5) [12]. These data suggest that there was an increase in the diversity of millimetre-scale zooplankton at or close to the base of Stage 3, very close in time to the appearance of the first vertical migrants. Most bradoriids are considered benthic, however Anabarochilina increased its distribution in three phases, providing complementary evidence for a steady strengthening of the pump during the Cambrian. In Epoch 2 Anabarochilina was coupled with benthic assemblages, by the Wuliuan it spread to a wider spectrum of lithofacies, and by the Guzhangian two species became widely distributed [48]. (c) Significance of vertical migration for the early Cambrian radiation of animals The metazoan innovation of vertical migration would have impacted both demersal and pelagic communities. The strength of the impact depends on the amount of biomass undertaking vertical migration. Models based on Cambrian environmental parameters predict that vertical migration would have increased the efficiency of the carbon pump by around 7% [60], however, more significantly, vertical migrants transport organic nutrients to the deep sea more quickly than aggregates, with a different nutritional balance, and repackage decaying sinking organic matter (figure 1) [1,13,61–64]. In addition, vertical migrants are major contributors to ocean mixing and ventilation, spreading oxygen and nutrients throughout the water column (e.g. [54,65]). These effects probably played a role in contributing to the rapid rate of diversification during the Cambrian explosion, interwoven with numerous evolutionary and ecological feedbacks. For example, the higher metabolic needs and nutrient requirements of large biomineralizing animals and motile predators [66] may have been facilitated by increased quality and quantity of nutrient transport in the biological pump, and resulted in increased oxygenation of bottom waters. In turn, the increasing size and motility of predators would have provided a further ecological pressure for animals to ‘escape’ into the pelagic realm. The establishment of a biological pump with a modern-style architecture by the Cambrian Stage 3 does not mean that the fluxes along the aggregation and vertical migration vectors (figure 1) remained constant to the modern day. Indeed they likely strengthened through the Palaeozoic with an increase in biomass (from an increased number of taxa, individuals and size of individuals). Fossil evidence points to additional metazoan innovations during the Palaeozoic that would have affected the fluxes along these vectors and strengthened the pump. For example, the aggregation vector would have been strengthened following the evolution of centimetre-scale and decimetre-scale phytoplanktivores later in the Cambrian [12,67], and the major radiation of plankton and the evolution of metre-scale nektonic suspension feeders during the Great Ordovician Biodiversification Event [68–70]. The flux of nutrients along the vertical migration vector would have increased as pelagic and vertically migrating animals diversified and increased in size—an innovation that would also have increased the mixing of waters and ocean ventilation. For example, the evolution of large, fast moving fish as part of the Devonian nekton revolution is expected to have been especially significant [54,71]. In summary, the innovation of vertical migration in some Isoxys species was one of several interwoven and coevolutionary feedbacks during the early Cambrian that increased the efficiency and altered the architecture of the biological pump, likely contributing to the rapid expansion in metazoan diversity at this time. Supplementary Material Click here for additional data file. Acknowledgements We thank the associate editor, two anonymous referees, and Christian Klug, who provided helpful reviewer comments. We thank members of the Ortega-Hernández Lab for Invertebrate Paleobiology (Harvard University) for fruitful discussions. S. Butts (Yale Peabody Museum) and M. Florence (Smithsonian National Museum of Natural History) provided curatorial assistance. Data accessibility All data and supplementary materials are available through the Open Science Framework: https://doi.org/10.17605/OSF.IO/2JDRS. Authors' contributions S.P.: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, software, validation, visualization, writing–original draft, writing-review & editing; A.C.D.: writing–review & editing; D.A.L.: data curation, writing–review and editing; I.A.R.: methodology, validation, writing–review & editing. All authors gave final approval for publication and agreed to be held accountable for the work performed therein. Competing interests We declare we have no competing interests. Funding S.P. was supported by an Alexander Agassiz Postdoctoral Fellowship (Harvard University) and a Herchel Smith Postdoctoral Fellowship (University of Cambridge), D.A.L. by a Dame Kathleen Ollerenshaw Research Fellowship (University of Manchester), and I.A.R. by a Museum Research Fellowship (Oxford University Museum of Natural History). ==== Refs References 1. Turner JT. 2015 Zooplankton fecal pellets, marine snow, phytodetritus and the ocean's biological pump. Prog. Oceanogr. 130 , 205-248. (10.1016/j.pocean.2014.08.005) 2. Passow U, Carlson CA. 2012 The biological pump in a high CO2 world. Mar. Ecol. Prog. Ser. 470 , 249-271. (10.3354/meps09985) 3. Hansell DA. 2013 Recalcitrant dissolved organic carbon fractions. Ann. Rev. Mar. Sci. 5 , 421-445. (10.1146/annurev-marine-120710-100757) 4. Butterfield NJ. 1997 Plankton ecology and the Proterozoic–Phanerozoic transition. Paleobiology 23 , 247-262. (10.1017/S009483730001681X) 5. Butterfield NJ. 2009 Macroevolutionary turnover through the Ediacaran transition: ecological and biogeochemical implications. Geol. Soc. Spec. Publ. 326 , 55-66. (10.1144/SP326.3) 6. Lenton TM, Boyle RA, Poulton SW, Shields-Zhou GA, Butterfield NJ. 2014 Co-evolution of eukaryotes and ocean oxygenation in the Neoproterozoic era. Nat. Geosci. 7 , 257-265. (10.1038/ngeo2108) 7. Brocks JJ, Jarrett AJM, Sirantoine E, Hallmann C, Hoshino Y, Liyanage T. 2017 The rise of algae in Cryogenian oceans and the emergence of animals. Nature 548 , 578-581. (10.1038/nature23457)28813409 8. Butterfield NJ. 1994 Burgess Shale-type fossils from a Lower Cambrian shallow-shelf sequence in northwestern Canada. Nature 369 , 477-479. (10.1038/369477a0) 9. Harvey THP, Butterfield NJ. 2008 Sophisticated particle-feeding in a large Early Cambrian crustacean. Nature 452 , 868-871. (10.1038/nature06724)18337723 10. Antcliffe JB, Callow RHT, Brasier MD. 2014 Giving the early fossil record of sponges a squeeze. Biol. Rev. 89 , 972-1004. (10.1111/brv.12090)24779547 11. Botting JP, Muir LA. 2018 Early sponge evolution: a review and phylogenetic framework. Palaeoworld 27 , 1-29. (10.1016/j.palwor.2017.07.001) 12. Wallet E, Slater BJ, Willman S, Peel JS. In press. Small carbonaceous fossils (SCFs) from North Greenland: new light on metazoan diversity in early Cambrian shelf environments. Pap. Palaeontol. (10.1002/spp2.1347) 13. Steinberg DK et al . 2008 Bacterial vs. zooplankton control of sinking particle flux in the ocean's twilight zone. Limnol. Oceanogr. 53 , 1327-1338. (10.4319/lo.2008.53.4.1327) 14. Ortega-Hernández J. 2016 Making sense of ‘lower’ and ‘upper’ stem-group Euarthropoda, with comments on the strict use of the name Arthropoda von Siebold, 1848. Biol. Rev. 91 , 255-273. (10.1111/brv.12168)25528950 15. Vannier J, Chen J. 2000 The Early Cambrian colonization of pelagic niches exemplified by Isoxys (Arthropoda). Lethaia 33 , 295-311. (10.1080/002411600750053862) 16. García-Bellido DC, Vannier J, Collins D. 2009 Soft-part preservation in two species of the arthropod Isoxys from the middle Cambrian Burgess Shale of British Columbia, Canada. Acta Palaeontol. Pol. 54 , 699-712. (10.4202/app.2009.0024) 17. Legg DA, Vannier J. 2013 The affinities of the cosmopolitan arthropod Isoxys and its implications for the origin of arthropods. Lethaia 46 , 540-550. (10.1111/let.12032) 18. Williams M, Siveter DJ, Peel JS. 1996 Isoxys (Arthropoda) from the Early Cambrian Sirius Passet Lagerstätte, North Greenland. J. Paleontol. 70 , 947-954. (10.1017/S0022336000038646) 19. Vannier J, García-Bellido DC, Hu SX, Chen AL. 2009 Arthropod visual predators in the early pelagic ecosystem: evidence from the Burgess Shale and Chengjiang biotas. Proc. R. Soc. B 276 , 2567-2574. (10.1098/rspb.2009.0361) 20. Perrier V, Williams M, Siveter DJ. 2015 The fossil record and palaeoenvironmental significance of marine arthropod zooplankton. Earth-Science Rev. 146 , 146-162. (10.1016/j.earscirev.2015.02.003) 21. Fu DJ, Zhang XL, Shu DG. 2011 Soft anatomy of the early Cambrian arthropod Isoxys curvirostratus from the Chengjiang biota of South China with a discussion on the origination of great appendages. Acta Palaeontol. Pol. 56 , 843-852. (10.4202/app.2010.0090) 22. Aria C, Caron J-B. 2015 Cephalic and limb anatomy of a new isoxyid from the burgess shale and the role of ‘stem bivalved arthropods' in the disparity of the frontalmost appendage. PLoS ONE 10 , e0124979. (10.1371/journal.pone.0124979)26038846 23. R Core Team. 2020 R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. See http://www.r-project.org/. 24. Vannier J, Caron J-B, Yuan J, Briggs DEG, Collins D, Zhao Y, Zhu M. 2007 Tuzoia: Morphology and lifestyle of a large bivalved Arthropod of the Cambrian seas. J. Paleontol. 81 , 445-471. (10.1666/pleo05070.1) 25. Meland K, Aas PØ. 2013 A taxonomical review of the Gnathophausia (Crustacea, Lophogastrida), with new records from the northern mid-Atlantic ridge. Zootaxa 3664 , 199-225. (10.11646/zootaxa.3664.2.5)26266297 26. Bonhomme V, Picq S, Gaucherel C, Claude J. 2014 Momocs: outline analysis using R. J. Stat. Softw. 56 , 1-24. (10.18637/jss.v056.i13) 27. García-Bellido DC, Paterson JR, Edgecombe GD, Jago JB, Gehling JG, Lee MSY. 2009 The bivalved arthropods Isoxys and Tuzoia with soft-part preservation from the lower Cambrian Emu Bay Shale Lagerstätte (Kangaroo Island, Australia). Palaeontology 52 , 1221-1241. (10.1111/j.1475-4983.2009.00914.x) 28. Wang Y, Huang D, Liu Q, Hu S. 2012 Isoxys from the Cambrian Guanshan Fauna, Yunnan Province. Earth Sci. J. China Univ. Geosci. 37 , 156-164. 29. Mateescu D, Abdo M. 2010 Analysis of flows past airfoils at very low Reynolds numbers. Proc. Inst. Mech. Eng. Part G J. Aerosp. Eng. 224 , 757-775. (10.1243/09544100JAERO715) 30. Wang ZJ. 2000 Two dimensional mechanism for insect hovering. Phys. Rev. Lett. 85 , 2216-2219. (10.1103/PhysRevLett.85.2216)10970501 31. Vogel S. 1996 Life in moving fluids: the physical biology of flow, 2nd edn. Princeton, NJ: Princeton University Press. 32. Schoenemann B, Clarkson ENK. 2011 Eyes and vision in the Chengjiang arthropod Isoxys indicating adaptation to habitat. Lethaia 44 , 223-230. (10.1111/j.1502-3931.2010.00239.x) 33. Cowles DL, Childress JJ, Gluckj DL. 1986 New method reveals unexpected relationship between velocity and drag in the bathypelagic mysid Gnathophausia ingens. Deep. Res. 33 , 865-880. (10.1016/0198-0149(86)90002-6) 34. Cowles DL, Childress JJ. 1988 Swimming speed and oxygen consumption in the bathypelagic mysid Gnathophausia ingens. Biol. Bull. 175 , 111-121. (10.2307/1541898) 35. Vannier J. 2007 Early Cambrian origin of complex marine ecosystems. In Deep-time perspectives on climate change: marrying the signal from computer models and biological proxies (eds M Williams, AM Haywood, FJ Gregory, DN Schmidt), pp. 81-100. London, UK: The Geological Society. 36. Hou X. 1987 Early Cambrian large bivalved arthropods from Chengjiang. Acta Palaeontol. Sin. 26 , 286-297. 37. Fu D, Zhang X, Budd GE, Liu W, Pan X. 2014 Ontogeny and dimorphism of Isoxys auritus (Arthropoda) from the Early Cambrian Chengjiang biota, South China. Gondwana Res. 25 , 975-982. (10.1016/j.gr.2013.06.007) 38. Zhao FC, Zhu MY, Hu SX. 2010 Community structure and composition of the Cambrian Chengjiang biota. Sci. China Earth Sci. 53 , 1784-1799. (10.1007/s11430-010-4087-8) 39. Caron JB, Jackson DA. 2008 Paleoecology of the Greater Phyllopod Bed community, Burgess Shale. Palaeogeogr. Palaeoclimatol. Palaeoecol. 258 , 222-256. (10.1016/j.palaeo.2007.05.023) 40. Allison PA. 1986 Soft-bodied animals in the fossil record: The role of decay in fragmentation during transport. Geology 14 , 979-981. (10.1130/0091-7613(1986)14<979:SAITFR>2.0.CO;2) 41. Stein M, Peel JS, Siveter DJ, Williams M. 2010 Isoxys (Arthropoda) with preserved soft anatomy from the Sirius Passet Lagerstätte, lower Cambrian of North Greenland. Lethaia 43 , 258-265. (10.1111/j.1502-3931.2009.00189.x) 42. Gaines RR. 2014 Burgess Shale-type preservation and its distribution in space and time. Paleontol. Soc. Pap. 20 , 1-24. (10.1017/S1089332600002837) 43. Lerosey-Aubril R, Kimmig J, Pates S, Skabelund J, Weug A, Ortega-Hernández J. 2020 New exceptionally preserved panarthropods from the Drumian Wheeler Konservat-Lagerstätte of the House Range of Utah. Pap. Palaeontol. 6 , 501-531. (10.1002/spp2.1307) 44. Paterson JR, García-Bellido DC, Jago JB, Gehling JG, Lee MSY, Edgecombe GD. 2016 The Emu Bay Shale Konservat-Lagerstätte: a view of Cambrian life from East Gondwana. J. Geol. Soc. London. 173 , 1-11. (10.1144/jgs2015-083) 45. Guilbaud R, Slater BJ, Poulton SW, Harvey THP, Brocks JJ, Nettersheim BJ, Butterfield NJ. 2017 Oxygen minimum zones in the early Cambrian ocean. Geochemical Perspect. Lett. 6 , 33-38. (10.7185/geochemlet.1806) 46. Harvey THP, Vélez MI, Butterfield NJ, Stanley SM. 2012 Exceptionally preserved crustaceans from western Canada reveal a cryptic Cambrian radiation. Proc. Natl Acad. Sci. USA 109 , 1589-1594. (10.1073/pnas.1115244109)22307616 47. Williams M, Siveter DJ, Popov LE, Vannier JMC. 2007 Biogeography and affinities of the bradoriid arthropods: cosmopolitan microbenthos of the Cambrian seas. Palaeogeogr. Palaeoclimatol. Palaeoecol. 248 , 202-232. (10.1016/j.palaeo.2006.12.004) 48. Williams M, Vandenbroucke TRA, Perrier V, Siveter DJ, Servais T. 2015 A link in the chain of the Cambrian zooplankton: Bradoriid arthropods invade the water column. Geol. Mag. 152 , 923-934. (10.1017/S0016756815000059) 49. Wolfe JM. 2017 Metamorphosis is ancestral for crown euarthropods, and evolved in the Cambrian or Earlier. Integr. Comp. Biol. 57 , 499-509. (10.1093/icb/icx039)28957514 50. Selden PA, Huys R, Stephenson MH, Heward AP, Taylor PN. 2010 Crustaceans from bitumen clast in Carboniferous glacial diamictite extend fossil record of copepods. Nat. Commun. 1 , 50. (10.1038/ncomms1049)20975721 51. Hu S, Steiner M, Zhu M, Erdtmann BD, Luo H, Chen L, Weber B. 2007 Diverse pelagic predators from the Chengjiang Lagerstätte and the establishment of modern-style pelagic ecosystems in the early Cambrian. Palaeogeogr. Palaeoclimatol. Palaeoecol. 254 , 307-316. (10.1016/j.palaeo.2007.03.044) 52. Fu D et al. 2019 The Qingjiang biota-A Burgess Shale-type fossil Lagerstätte from the early Cambrian of South China. Science 363 , 1338-1342. (10.1126/science.aau8800)30898931 53. Rahman IA, Darroch SAF, Racicot RA, Laflamme M. 2015 Suspension feeding in the enigmatic Ediacaran organism Tribrachidium demonstrates complexity of Neoproterozoic ecosystems. Sci. Adv. 1 , e1500800. (10.1126/sciadv.1500800)26702439 54. Butterfield NJ. 2018 Oxygen, animals and aquatic bioturbation: an updated account. Geobiology 16 , 3-16. (10.1111/gbi.12267)29130581 55. Antcliffe JB, Jessop W, Daley AC. 2019 Prey fractionation in the Archaeocyatha and its implication for the ecology of the first animal reef systems. Paleobiology 45 , 652-675. (10.1017/pab.2019.32) 56. Steiner M, Li G, Qian Y, Zhu M, Erdtmann BD. 2007 Neoproterozoic to Early Cambrian small shelly fossil assemblages and a revised biostratigraphic correlation of the Yangtze Platform (China). Palaeogeogr. Palaeoclimatol. Palaeoecol. 254 , 67-99. (10.1016/j.palaeo.2007.03.046) 57. Kouchinsky A, Bengtson S, Runnegar B, Skovsted C, Steiner M, Vendrasco M. 2012 Chronology of early Cambrian biomineralization. Geol. Mag. 149 , 221-251. (10.1017/S0016756811000720) 58. Betts MJ, Paterson JR, Jago JB, Jacquet SM, Skovsted CB, Topper TP, Brock GA. 2017 Global correlation of the early Cambrian of South Australia: Shelly fauna of the Dailyatia odyssei Zone. Gondwana Res. 46 , 240-279. (10.1016/j.gr.2017.02.007) 59. Vinther J, Stein M, Longrich NR, Harper DAT. 2014 A suspension-feeding anomalocarid from the Early Cambrian. Nature 507 , 496-499. (10.1038/nature13010)24670770 60. Fakhraee M, Planavsky NJ, Reinhard CT. 2020 The role of environmental factors in the long-term evolution of the marine biological pump. Nat. Geosci. 13 , 812-816. (10.1038/s41561-020-00660-6) 61. Schnetzer A, Steinberg DK. 2002 Active transport of particulate organic carbon and nitrogen by vertically migrating zooplankton in the Sargasso Sea. Mar. Ecol. Prog. Ser. 234 , 71-84. (10.3354/meps234071) 62. Steinberg DK, Goldthwait SA, Hansell DA. 2002 Zooplankton vertical migration and the active transport of dissolved organic and inorganic nitrogen in the Sargasso Sea. Deep. Res. Part I Oceanogr. Res. Pap. 49 , 1445-1461. (10.1016/S0967-0637(02)00037-7) 63. Wilson SE, Steinberg DK, Buesseler KO. 2008 Changes in fecal pellet characteristics with depth as indicators of zooplankton repackaging of particles in the mesopelagic zone of the subtropical and subarctic North Pacific Ocean. Deep. Res. Part II Top. Stud. Oceanogr. 55 , 1636-1647. (10.1016/j.dsr2.2008.04.019) 64. Hannides CCS, Landry MR, Benitez-Nelson CR, Styles RM, Montoya JP, Karl DM. 2009 Export stoichiometry and migrant-mediated flux of phosphorus in the North Pacific Subtropical Gyre. Deep. Res. Part I Oceanogr. Res. Pap. 56 , 73-88. (10.1016/j.dsr.2008.08.003) 65. Bollens SM, Rollwagen-Bollens G, Quenette JA, Bochdansky AB. 2011 Cascading migrations and implications for vertical fluxes in pelagic ecosystems. J. Plankton Res. 33 , 349-355. (10.1093/plankt/fbq152) 66. Sperling EA, Frieder CA, Raman AV, Girguis PR, Levin LA, Knoll AH. 2013 Oxygen, ecology, and the Cambrian radiation of animals. Proc. Natl Acad. Sci. USA 110 , 13 446-13 451. (10.1073/pnas.1312778110) 67. Lerosey-Aubril R, Pates S. 2018 New suspension-feeding radiodont suggests evolution of microplanktivory in Cambrian macronekton. Nat. Commun. 9 , 1-9. (10.1038/s41467-018-06229-7)29317637 68. Servais T, Owen AW, Harper DAT, Kröger B, Munnecke A. 2010 The Great Ordovician Biodiversification Event (GOBE): the palaeoecological dimension. Palaeogeogr. Palaeoclimatol. Palaeoecol. 294 , 99-119. (10.1016/j.palaeo.2010.05.031) 69. Servais T et al. 2016 The onset of the ‘Ordovician Plankton Revolution’ in the late Cambrian. Palaeogeogr. Palaeoclimatol. Palaeoecol. 458 , 12-28. (10.1016/j.palaeo.2015.11.003) 70. Van Roy P, Daley AC, Briggs DEG. 2015 Anomalocaridid trunk limb homology revealed by a giant filter-feeder with paired flaps. Nature 522 , 77-80. (10.1038/nature14256)25762145 71. Klug C, Kröger B, Kiessling W, Mullins G, Servais T, Frýda J, Korn D, Turner S. 2010 The Devonian nekton revolution. Lethaia 43 , 465-477. (10.1111/j.1502-3931.2009.00206.x)
PMC008xxxxxx/PMC8220268.txt
==== Front Proc Biol Sci Proc Biol Sci RSPB royprsb Proceedings of the Royal Society B: Biological Sciences 0962-8452 1471-2954 The Royal Society 34157868 10.1098/rspb.2021.0692 rspb20210692 10016014470Palaeobiology Research Articles Deep-time biodiversity patterns and the dinosaurian fossil record of the Late Cretaceous Western Interior, North America Deep-time biodiversity patterns and the dinosaurian fossil record of the late cretaceous Western Interior, North America http://orcid.org/0000-0002-7741-2500 Maidment Susannah C. R. Conceptualization Data curation Formal analysis Investigation Methodology Project administration Writing-original draft Writing-review & editing [email protected] 1 3 Dean Christopher D. Formal analysis Investigation Methodology Writing-original draft Writing-review & editing 1 Mansergh Robert I. Data curation Writing-original draft 1 2 http://orcid.org/0000-0003-2136-7541 Butler Richard J. Conceptualization Data curation Formal analysis Investigation Methodology Writing-original draft Writing-review & editing 3 1 Department of Earth Sciences, Natural History Museum, Cromwell Road, London, SW7 5BD, UK 2 Department of Earth Sciences, 5 Gower Place, London WC1E 6BS, UK 3 School of Geography, Earth and Environmental Sciences, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK Electronic supplementary material is available online at https://doi.org/10.6084/m9.figshare.c.5459154. 30 6 2021 June 30, 2021 23 6 2021 June 23, 2021 23 6 2021 June 23, 2021 288 1953 2021069222 3 2021 March 22, 2021 27 5 2021 May 27, 2021 © 2021 The Authors. 2021 https://creativecommons.org/licenses/by/4.0/ Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited. In order for palaeontological data to be informative to ecologists seeking to understand the causes of today's diversity patterns, palaeontologists must demonstrate that actual biodiversity patterns are preserved in our reconstructions of past ecosystems. During the Late Cretaceous, North America was divided into two landmasses, Laramidia and Appalachia. Previous work has suggested strong faunal provinciality on Laramidia at this time, but these arguments are almost entirely qualitative. We quantitatively investigated faunal provinciality in ceratopsid and hadrosaurid dinosaurs using a biogeographic network approach and investigated sampling biases by examining correlations between dinosaur occurrences and collections. We carried out a model-fitting approach using generalized least-squares regression to investigate the sources of sampling bias we identified. We find that while the raw data strongly support faunal provinciality, this result is driven by sampling bias. The data quality of ceratopsids and hadrosaurids is currently too poor to enable fair tests of provincialism, even in this intensively sampled region, which probably represents the best-known Late Cretaceous terrestrial ecosystem on Earth. To accurately reconstruct biodiversity patterns in deep time, future work should focus on smaller scale, higher resolution case studies in which the effects of sampling bias can be better controlled. faunal provincialism , endemicity , Laramidia , sampling bias , Ceratopsidae , Hadrosauridae European Union 637483 cover-dateJune 30, 2021 ==== Body pmc1. Introduction In order to predict how biodiversity patterns on today's Earth will respond to climate change, the factors that cause biodiversity distributions must be understood [1,2]. Deep-time perspectives can provide novel insights into the controls on biodiversity distribution. By examining biodiversity distributions at times in Earth's history when climate, continental arrangement, and oceanic currents were different than today, ecological hypotheses about the causative mechanisms behind biodiversity distribution and the establishment of modern patterns can be tested [3–7]. However, if palaeontologists wish their data to be informative to those working on the causative mechanisms of modern-day biodiversity patterns, we must first demonstrate that actual biodiversity patterns are preserved in our reconstructions of past ecosystems, and that we are able to overcome the many sampling biases that affect the fossil record (e.g. [8–10]). Dinosaurs are an exceptional model system for studying biodiversity and macroevolution in terrestrial vertebrates. For greater than 150 million years, from the Late Triassic to the end of the Cretaceous, they dominated terrestrial ecosystems, occupied every continent, and radiated into a wide variety of ecological niches. Because of public interest, they are the best-sampled Mesozoic terrestrial vertebrate group and their fossils have been collected for well over 150 years [11–14]. Arguably, the best-sampled part of the dinosaurian fossil record is the Late Cretaceous of the Western Interior region of North America [15–17]. During the Late Cretaceous, North America was divided into two landmasses, Laramidia to the west and Appalachia to the east, by the epicontinental Western Interior Seaway. In the latter stages of the Late Cretaceous (Campanian and Maastrichtian), large-bodied herbivorous niches in Laramidia were dominated by two groups of ornithischian dinosaurs, the hadrosaurs and ceratopsids. The body fossil record of the latter is entirely restricted to North America at this time, with the exception of a single taxon [18,19]. Study of these herbivorous dinosaurs has provided major insight into dinosaur behaviour, palaeoecology, and biogeographic patterns (e.g. [20–32]). Numerous workers have argued for strong faunal provinciality in Laramidia throughout the Campanian and have divided the landmass into northern and southern faunal provinces (e.g. [33–35]). This signal is particularly clear in chasmosaurine ceratopsids, where virtually all species are recognized from either northern or southern Laramidia, but not both [34–36]. This endemicity in ceratopsids is thought to have driven high levels of diversity, underpinning their radiation [34]. Since no geological or geographical barrier has thus far been identified between northern and southern Laramidia [37], the boundary between the northern and southern provinces has been suggested to be related to latitudinal climate, temperature, or rainfall patterns [34,35,38] or was maintained due to competition between local populations [39]. The patterns of apparent provincialism decrease in the Maastrichtian, coincident with overall regression of the Western Interior Seaway [15,17,33,36]. These hypotheses of biogeographic provincialism, however, remain controversial. With very few exceptions [35,36], studies that advocate for provincialism are based on qualitative observations (e.g. [33,34]) and arise from comparisons of the fauna of specific geological formations (e.g. [33–35]). Recent research has, however, suggested that some of the formations used in such studies are not contemporaneous [17,40] and that the length of time intervals used results in the amalgamation of multiple successive faunas [17,37]. Many studies advocating faunal endemism are based on taxonomic decisions that have proven controversial and the conclusions have been called into question as a result (e.g. [37,39]). Additionally, it remains a possibility that faunal provinces within the Campanian are an artefact of sampling: most Campanian dinosaur occurrences are known from Alberta, Montana, southern Utah, and northern New Mexico, with far less sampling having occurred in northern Utah and Wyoming [16]. Three quantitative studies have investigated the provincialism hypothesis in dinosaurs of the Late Cretaceous Western Interior. Gates et al. [35] used a variety of statistical techniques to assess the similarity between Campanian northern and southern faunas, and found evidence for either two distinct provinces with a broad area of overlap between them, or a latitudinal diversity gradient. The statistical techniques employed were unable to distinguish between these two hypotheses, and their results regarding dinosaurs were inconclusive. They suggested a further investigation into the causes of dinosaur distribution in the Western Interior. Berry [36] used a phylogenetic approach to assess biogeography within the Campanian and found no evidence for endemic sub-clades of chasmosaurine ceratopsids, arguing that this would be expected if there was a major barrier to dispersal, or niche conservatism related to climate. Vavrek & Larsson [15] investigated faunal endemism in the Maastrichtian of Laramidia using measures of beta diversity. After correcting for sampling biases, they found little evidence of provincialism, instead suggesting a homogeneous dinosaurian fauna across the Western Interior region at the very end of the Cretaceous; however, they did not test to see whether apparent biogeographic patterns within the Campanian were also caused by sampling. Herein, we quantitatively test hypotheses of faunal endemism in both the Campanian and Maastrichtian using biogeographic and multivariate statistical approaches. We focus our study on ceratopsid and hadrosaurian dinosaurs, as these megaherbivores have well-understood phylogenies and taxonomies and have been at the centre of previous discussions of faunal provinciality in this region. The distinctiveness of northern and southern Laramidian provinces are tested using phylogenetically corrected Biogeographic Connectedness (pBC). This quantitative method uses a network approach to assess phylogenetic distances between taxa in different geographic areas, resulting in a metric that quantifies the degree of faunal provinciality versus cosmopolitanism. It has been used successfully to understand changes in faunal compositions through the Carboniferous–Permian transition [41], and the Permian–Triassic and Triassic–Jurassic extinction events [42]. We also introduce additions to the methodology that address concerns regarding variation in sampling through time. To investigate the impact of sampling bias on our results, we examine correlations between occurrences (records of specimens) and collections (sites where specimens have been collected) with latitude, and use multivariate regression to examine which sources of sampling bias best explain sampling patterns. We use our results to determine whether it is possible to identify true geographic patterns of biodiversity on a continental scale in this very well-sampled area. 2. Methods (a) Taxon sampling and phylogeny Since no complete phylogenetic analysis of all ceratopsids is available, we built an informal supertree of all ceratopsid taxa considered valid in recent phylogenetic analyses by combining the phylogenetic results of [18] for chasmosaurines and [19] for centrosaurines. We resolved polytomies by removing Nedoceratops, a taxon some workers consider to be invalid ([43], but see [44]), from the data matrix in [18] and re-analysing the dataset. This resolved polytomies in the clade containing the common ancestor of Eotriceratops, Triceratops, and all of its descendants. The resulting supertree includes 67 taxa and represents a consensus of current views on ceratopsian phylogeny (electronic supplementary material, figure S1a). The structure of the hadrosaurid tree is based on several key recent analyses [45–47]. We resolved polytomies and added taxa considered valid but not included in those references using other recent phylogenetic analyses [48–51]. The resulting supertree includes 55 taxa and represents a current reasonable estimate of hadrosaur phylogeny (electronic supplementary material, figure S1b). (b) Stratigraphic age and geographic data Age for North American hadrosaur and ceratopsid species was obtained from the primary literature. The formations in which taxa occurred were found from the Paleobiology Database (PBDB; www.paleobiodb.org), and the most recent absolute age estimate of those formations was obtained from the primary literature (see electronic supplementary material, OSM, for sources). The age and geographic data for taxa outside of North America were obtained from the PBDB. pBC requires a priori assignment of geographic regions to test hypotheses of biogeographic connectedness, so we assigned dinosaurs to either northern Laramidia or southern Laramidia. Northern Laramidia includes taxa found in Wyoming and further north; southern Laramidia includes taxa found in Utah and further south, following previous studies. Age data were used to time-calibrate the phylogenetic trees using the ‘timePaleoPhy’ function of the Strap package [52] in R v. 3.5.2 [53] with the minimum branch length option specified (type=’mbl’). While it would be ideal to use high-resolution bins to test patterns of biogeography through the Late Cretaceous [17], too few taxa would be present in each bin to permit the use of pBC. Consequently, we divided taxa into Campanian and Maastrichtian time bins, which also has the benefit of allowing for comparison between previous studies of faunal provincialism in this area. Where a taxon's stratigraphic range/uncertainty crossed the Campanian–Maastrichtian boundary, it was included in both time intervals. (c) pBC We calculated pBC for Campanian ceratopsids, Maastrichtian ceratopsids, and Campanian hadrosaurs. The sampling of Maastrichtian hadrosaurs was too sparse, particularly in southern Laramidia, to calculate meaningful pBC values. Trees were pruned to exclude taxa from timeslices other than the one being analysed, and were made ultrametric prior to analysis. pBC was calculated using the function BC of the package ‘dispeRse’ (available at github.com/laurasoul/disperse). We initially varied the constant µ (see [41]) from 1 to 15 million years; subsequent analyses used a constant µ of 10. Data were jack-knifed 1000 times to produce a distribution of possible pBC values. To address concerns about the potential for a relationship between pBC and taxon sample size [54], we calculated rarefaction curves for pBC for the ceratopsian data (to facilitate comparisons between the Campanian and Maastrichtian). Sample sizes were rarefied down to a minimum number of five taxa. Ninety-five per cent confidence intervals for the rarefaction curves were generated using 1000 replicates at each sampling level. (d) Randomization of data (null model) In order to determine whether pBC for each clade and time interval was significantly different from random, we randomly permuted the geographic areas in which taxa are found. We generated 1000 permutations of the data for each clade and time interval and calculated pBC for each permutation. The pBC for the unpermuted data was compared to the distribution of permuted pBC values to establish statistical significance (p < 0.05). (e) Sampling bias To investigate whether biogeographic patterns we observed in the pBC results were influenced by sampling bias, we downloaded raw occurrence data for ceratopsids and hadrosaurs for the Campanian and Maastrichtian from the PBDB. We then downloaded North American dinosaur-bearing collections and North American tetrapod-bearing collections for each timeslice, and plotted occurrences and collections with latitude. We compared the curves using Spearman's rho and Kendall's tau. To investigate the possible causes of sampling bias we identified, we statistically examined correlations between occurrences and outcrop area, depositional environment, and proxies for exposure. First, we imported publicly available United States Geological Survey (USGS) state-level and Canadian Province digital geological maps (www.ngmbd.usgs.gov; https://ags.aer.ca/publication/map-600; https://geohub.saskatchewan.ca/datasets/bedrock-geology) into ArcMap 10 (www.esri.com), identified Campanian and Maastrichtian strata, and assigned an environmental attribute determining whether strata were deposited in a terrestrial, marine, or mixed setting (OSM). These data, along with maximum green vegetation fraction (MGVF) and slope, both proxies for exposure, were imported into R (version 3.5.0). Methods for generating MGVF and slope are provided in OSM. Level plots of total outcrop area, terrestrial, mixed and marine outcrop area, slope, and MGVF were produced using the levelplot() function of the rasterVis() package [55] (electronic supplementary material, figure S2). To investigate the power of each or a combination of these variables to explain the dinosaur occurrence data, we carried out a model-fitting approach using generalized least-squares regression (GLS). Ceratopsian and hadrosaur occurrences from the PBDB were counted in each 1-degree latitudinal bin (latitude is modern latitude, rather than palaeolatitude). Models compared latitudinal changes in ceratopsian and hadrosaur occurrences to changes in four different measures of outcrop area (see OSM), MGVF, and slope. GLS autoregressive models were fitted to combinations of potential explanatory variables. We used a first-order autoregressive model (corARMA) fitted to the data to account for spatial autocorrelation using the function gls() in the R package nlme v. 3.1–150 [56]. GLS reduces the chance of overestimating the statistical significance of regression lines due to serial correlation in the latitudinal series. Data series were ln-transformed prior to analysis to ensure normality and homoskedasticity of residuals. We calculated likelihood-ratio-based pseudo-R2 values using the function r.squaredLR() of the R package MuMIn [57]. Results were compared using Akaike's information criterion for small sample sizes (AICc) and Akaike weights were calculated to identify the best combination of explanatory variables from those tested. AICc was calculated using the function AICc() of the R package qpcR [58], and Akaike weights calculated using the aic.w() function of the R package phytools [59]. (i) Sampling bias and pBCs To test the impact of the Campanian bimodal sampling distribution on pBC results, we ran a second pBC test where we randomly removed 95% of ceratopsian taxa from the Maastrichtian that occurred between 35 and 50 degrees of latitude. We chose these latitudinal boundaries to enforce a similar bimodal latitudinal diversity gradient on the Maastrichtian data as seen in the Campanian (see Results). The remaining distribution of occurrences was used to re-run pBC analyses (with a µ of 10), and this process was repeated 1000 times for increased accuracy of results. pBC scores were recorded for each run, and the resulting distribution was used to calculate the mean pBC to compare against the original Maastrichtian ceratopsian pBC score and produce a probability density curve to estimate the probability of different values of pBC scores. 3. Results The observed value of pBC for Campanian ceratopsids was 0.05, while that for Campanian hadrosaurs was 0.11, and for Maastrichtian ceratopsids the observed value was 0.16. pBC was therefore lower for ceratopsids in the Campanian than in the Maastrichtian, and endemism was correspondingly higher, in agreement with previous studies [33,36]. Jack-knifed distributions of ceratopsid pBC for the Campanian and Maastrichtian overlap (figure 1a–c), but their median values are strongly significantly different from each other (Wilcox Test, W = 60235, p = 0.00). Rarefaction curves for ceratopsids for the Campanian and Maastrichtian indicate a much higher pBC in the Maastrichtian than in the Campanian at equivalent levels of sampling, although the confidence intervals do overlap, particularly at lower sampling levels (figure 1d). This demonstrates that the higher pBC of the Maastrichtian is not a consequence of sampling lower numbers of species in that interval in comparison to the Campanian. Higher pBC equates to more cosmopolitan faunas, and thus this result supports lower endemism in Laramidia during the Maastrichtian when compared to the Campanian. Figure 1. (a–c) Jack-knifed distributions of pBC values for Campanian (green) and Maastrichtian (purple) ceratopsids. (d) Rarefaction curves for Campanian (green circles) and Maastrichtian (purple triangles) ceratopsids. Error bars show the 95% confidence intervals of values obtained during rarefaction. (Online version in colour.) Values for pBC for both Campanian and Maastrichtian ceratopsid data are statistically significantly lower than for datasets in which the geographic areas are randomized (Campanian, p = 0.00; Maastrichtian, p = 0.015; electronic supplementary material, figure S3), and the same is true for the Campanian hadrosaur data (p = 0.00; electronic supplementary material, figure S3). This indicates that endemism was statistically significantly higher than in all randomized datasets across both time intervals and supports previous qualitative hypotheses of distinct northern and southern provinces in Laramidia (e.g. [33–35]) Curves of raw occurrence data with latitude for hadrosaurs and ceratopsids in both the Campanian and the Maastrichtian correlate strongly and statistically significantly with both dinosaur-bearing and tetrapod-bearing collections (figure 2; electronic supplementary material, figure S4; OSM). During the Campanian, sampling and occurrences are focused at two latitudes: 51–49 degrees north, which corresponds with the Dinosaur Park, Oldman and, to a lesser extent, the Foremost formations, and 36–37 degrees north, which corresponds primarily with the Kirtland/Fruitland, Aguja, and Kaiparowits Formations (figure 2a,b; electronic supplementary material, figure S4a,b). These two areas have been sampled orders of magnitude better than the surrounding latitudinal bins [16], although there are tetrapod- and dinosaur-bearing formations across the majority of the Western Interior at this time (figure 2a,b; electronic supplementary material, figure S4a,b). In the Maastrichtian, sampling is more evenly spread across the range of latitudes for which we have hadrosaur and ceratopsid body fossils (figure 2c,d; electronic supplementary material, figure S4c,d; [16]). These data are strongly indicative that the provinciality observed based on raw data in the Campanian could be due to intensive sampling in the Dinosaur Park Formation and Kirtland/Fruitland Formations with a lack of sufficient sampling between, and our observed increase in pBC (= reduced endemism) in the Maastrichtian is due to increased latitudinal coverage of sampling. Figure 2. (a,c) Ceratopsid occurrences and dinosaur-bearing collections with latitude in the (a), Campanian and (c), Maastrichtian. (b,d) Hadrosaurid occurrences and dinosaur-bearing collections with latitude in the (b), Campanian and (d), Maastrichtian. τ = Kendall's tau; ρ = Spearman's rho; DBCs, dinosaur-bearing collections. (Online version in colour.) The mean pBC score of Maastrichtian ceratopsians subjected to a Campanian-style sampling distribution was 0.0351 with a standard deviation of 0.0476, significantly lower than the original pBC score of 0.16. The probability of a pBC score less than or equal to 0.8 was 0.78 (OSM and electronic supplementary material, figure S5). These results provide a further indication that sampling bias is driving pBC scores of Campanian fauna. A lack of sampling in the area between 49 degrees north and 37 degrees north (the ‘sampling peaks’) in Campanian strata could be caused by a variety of factors. It has long been known that rock outcrop area is strongly correlated with raw diversity (e.g. [60,61]); if there is less outcrop, there are fewer opportunities for palaeontologists to sample the rocks, and fewer fossils found as a consequence. As terrestrial organisms, the vast majority of dinosaur fossils are found in formations that were deposited on land. If Campanian rocks between the sampling peaks are primarily marine, there will be fewer opportunities for dinosaur fossils to be preserved, and thus fewer opportunities for them to be sampled by palaeontologists. Fossils are primarily found where bare rock is exposed at the surface. If less rock is exposed between the sampling peaks than in the areas of the peaks themselves, there will be fewer opportunities for fossils to come to light. GLS analyses recovered the following best models (highest AICc weights) for outcrop and tetrapod occurrence masks (see OSM for additional results): Campanian hadrosaurs, summed outcrop area + MGVF + slope; Campanian ceratopsians, non-marine total outcrop area; Maastrichtian hadrosaurs, null model; Maastrichtian ceratopsians, null model. However, in nearly all cases the correlations are non-significant (OSM) and only the Campanian hadrosaur model results had a strong overall explanatory power (OSM). This indicates that the potential sampling bias with latitude in the Campanian that we have identified cannot be fully explained by any of these variables and other sources of sampling bias that are hard to quantify may additionally be responsible. 4. Discussion Several authors have suggested that the apparent faunal provincialism in Laramidia during the Late Cretaceous is an artefact, either because the formations in which dinosaurian taxa have been found are not contemporaneous [17,37,40] or due to uneven sampling of the fossil record [15,16]. Our results show that while the raw data clearly supports faunal endemicity, particularly in the Campanian, this pattern is driven by a lack of sampling outside of two specific latitudinal belts on Laramidia (51–49 degrees north, which corresponds with the Dinosaur Park, Oldman and Foremost formations, and 36–37 degrees north, which corresponds primarily with the Kirtland/Fruitland, Aguja, and Kaiparowits formations). This sampling bias cannot be fully explained by differences in outcrop area across the region, or by differences in slope or vegetation, which are factors that affect rock exposure. There are numerous other factors that can bias sampling, but these are very difficult to quantify. Low sampling between the northern and southern sampling peaks could occur if palaeontologists have yet to prospect the area to the same degree that they have in the north and south. The Late Cretaceous of the Western Interior has been intensively sampled for dinosaur fossils for over 100 years, and it is now probably the best-known Late Cretaceous ecosystem anywhere on Earth [15,16]. It is therefore highly unlikely that large parts of it remain unexplored for dinosaurs, and the fact that dinosaur fossils are known from the area between the sampling peaks during the Maastrichtian suggest the area has been explored. The lack of exploration for fossils is therefore unlikely to be the primary driver of the uneven sampling patterns we have observed. The ‘common cause hypothesis’ (e.g. [62]) suggests that correlations between raw diversity and sampling proxies (e.g. numbers of formations) are driven by a third factor, usually sea level. Although initially formulated for marine environments, the possibility of a sea-level driven common cause on land has also been discussed (e.g. [63]). During sea-level high stands sediment flux to inner shelves and marginal marine areas is high; this results in both high potential for the preservation of fossils due to rapid burial and high diversity due to habitat fragmentation leading to endemism and increased beta diversity. Conversely, sediment bypasses inner shelf environments during low stands, reducing sediment flux and leading to poorer preservation of fossils due to a lower chance of rapid burial, while diversity is lower due to cosmopolitanism. Although the effect of eustatic sea-level changes on the global terrestrial fossil record of vertebrates has been questioned [63], Chiarenza et al. [16] demonstrated that the areas of our northern and southern sampling peaks correlated with high sediment fluxes and low runoff rates during the Campanian. It is therefore possible that reduced sampling between our sampling peaks is because this area was less suitable for fossil preservation in the Campanian. Indeed, Chiarenza et al. [16] suggested that faunal provinicialism in the Campanian was a sampling bias at least partially due to variation in climatic-induced taphonomic suitability between northern and southern regions. Historical collecting practises and/or land ownership might also play a role in the sampling patterns we have observed. If the proportion of outcrop on public land was reduced in the areas outside of our sampling peaks, this might mean palaeontologists have less access to explore there for fossils. Furthermore, if there is a particularly field-active palaeontological institution close to an area of Campanian outcrop, or long-term agreements in place with landowners, this may have allowed prospecting to occur more regularly over a longer period of time in specific areas. A bias may also be introduced by uneven regional entry of data into online databases such as the PBDB. Such a bias could stem from monographs on specific formations or museums whose collections focus on specific areas that also have online databases. Data from these sources are comparatively easy to enter into the PBDB and thus could be contributing to the sampling patterns we observed. It seems highly likely that a combination of available outcrop area, rocks suitable for the fossilization of vertebrate remains, and an interplay between climate, topography, and historical collection and data entry practises is responsible for variations in sampling across the Western Interior, which have resulted in apparent northern and southern faunal provinces on Laramidia. (a) Taxonomic differences in northern and southern Laramidia Despite the fact that we find faunal provincialism in the Late Cretaceous Western Interior to mostly likely be due to sampling bias based on currently available data, it is clear that different taxa are found in the northern and southern areas of Laramidia [33–35]. This is especially clear in chasmosaurine ceratopsids, where there is almost no overlap at all between taxa found in the north and those found in the south [34], but see [36,39]. It has been demonstrated that many of these taxa were not contemporaneous [40], which would at least partially explain taxonomic differences. But, in addition, the study area covers 12 degrees of latitude and climate would have varied significantly over that area, even in a greenhouse world where latitudinal temperature gradients were reduced relative to today [39,64]. General circulation models for the Campanian show significant variation in mean annual temperature and rainfall patterns with latitude across Laramidia [16] and recent research has suggested elevated temperature gradients in a climatic transition zone between the northern and southern faunal provinces [38]. Given that there is evidence for both spatial [26] and functional [29] niche partitioning in Laramidia's large herbivores, taxonomic differences between the north and south could be related to climatic preference, and there may well have been a latitudinal biodiversity gradient across the area. Unfortunately, we have demonstrated here that that raw data is currently too influenced by sampling biases for such biodiversity patterns to be reconstructed. 5. Conclusion We show that data quality of Campanian and Maastrichtian ceratopsids and hadrosaurs, two of the most abundant clades of dinosaurs in the Late Cretaceous of North America, is currently too poor to enable fair tests of endemicity and provincialism. In order to effectively test hypotheses regarding the causative mechanisms of biodiversity distribution, palaeontologists must demonstrate either that the fossil record preserves true biodiversity patterns at high levels of temporal resolution, or that methods exist that can adequately overcome sampling biases. The Western Interior region represents probably the most densely sampled Late Cretaceous terrestrial region worldwide [15,16], but even in this intensively sampled area, it is not currently possible to reconstruct diversity patterns at the regional scale. In order for palaeontologists to make a meaningful contribution to ecological hypotheses about future biodiversity change, we must focus our efforts on smaller scale case studies, where temporal resolution is high, stratigraphic correlation is well established, and where sampling biases are likely to be more homogeneous and can be more easily quantified. A good example of a recent such study is [65]. The results of multiple high-resolution case studies can then be compared globally to establish the rules that governed past biodiversity distributions. Supplementary Material Click here for additional data file. Acknowledgements Graeme Lloyd (University of Leeds) and David Button (Natural History Museum) wrote the original code to implement pBC. Caleb Brown (Royal Tyrrell Museum of Palaeontology) and an anonymous reviewer provided detailed and thoughtful comments that improved this manuscript. This is Paleobiology Database official publication 404. Data accessibility Raw data, additional methods, results, and figures can be found in the electronic supplementary material. Further raw data, all code and a copy of the electronic supplementary material are available from the Dryad Digital Repository: https://doi.org/10.5061/dryad.bcc2fqzbz [66]. Author contributions S.C.R.M. and R.J.B. designed the study. S.C.R.M., R.J.B., and C.D.D. wrote the manuscript. R.I.M. and S.C.R.M. collected data and generated supertrees. S.C.R.M., R.J.B., and C.D.D. ran analyses. All authors provided critical comments on the manuscript. Competing interests We declare we have no competing interests. Funding R.J.B. was funded by European Union's Horizon 2020 research and innovation programme under grant agreement no. 637483 during the course of this work. ==== Refs References 1. Gaston KJ. 2000 Global patterns in biodiversity. Nature 405 , 220-227. (10.1038/35012228)10821282 2. Willig MR, Kaufman DM, Stevens RD. 2003 Latitudinal gradients of biodiversity: pattern, process, scale, synthesis. Ann. Rev. Ecol. Evol. Syst. 34 , 273-309. (10.1146/annurev.ecolsys.34.012103.144032) 3. Jablonski D, Roy K, Valentine JW. 2006 Out of the tropics: evolutionary dynamics of the latitudinal diversity gradient. Science 314 , 102-106. (10.1126/science.1130880)17023653 4. Jetz W, Thomas GH, Joy JB, Hartmann K, Mooers AO. 2012 The global diversity of birds in space and time. Nature 491 , 444-448. (10.1038/nature11631)23123857 5. Dornelas M et al. 2013 Quantifying temporal change in biodiversity: challenges and opportunities. Proc. R. Soc. B 280 , 20121931. (10.1098/rspb.2012.1931) 6. Mannion PD, Upchurch P, Benson RJB, Goswami A. 2014 The latitudinal biodiversity gradient through deep time. TREE 29 , 42-50. (10.1016/j.tree.2013.09.012)24139126 7. Price SA, Schmitz L. 2016 A promising future for integrative biodiversity research: an increased role of scale-dependency and functional biology. Phil. Trans. R. Soc. B 371 , 30150228. (10.1098/rstb.2015.0228) 8. Fraser D. 2017 Can latitudinal richness gradients be measured in the terrestrial fossil record? Paleobiology 43 , 479-494. (10.1017/pab.2017.2) 9. Jones LA, Dean CD, Mannion PD, Farnsworth A, Allison PA. 2021 Spatial sampling heterogeneity limits the detectability of deep time latitudinal biodiversity gradients. Proc. R. Soc. B 288 , 20202762. (10.1098/rspb.2020.2762) 10. Darroch SAF, Fraser D, Casey MM. 2021 The preservation potential of terrestrial biogeographic patterns. Proc. R. Soc. B 288 , 20202927. (10.1098/rspb.2020.2927) 11. Brusatte SL, Benton MJ, Ruta M, Lloyd GT. 2008 Superiority, competition, and opportunism in the evolutionary radiation of dinosaurs. Science 321 , 1485-1488. (10.1126/science.1161833)18787166 12. Lloyd GT, Davis KE, Pisani D, Tarver JE, Ruta M, Sakamoto M, Hone DWE, Jennings R, Benton MJ. 2008 Dinosaurs and the Cretaceous terrestrial revolution. Proc. R. Soc. B 275 , 2483-2490. (10.1098/rspb.2008.0715) 13. Benson RBJ, Hunt G, Carrano MT, Campione N. 2017 Cope's rule and the adaptive landscape of dinosaur body size evolution. Palaeontology 61 , 13-48. (10.1111/pala.12329) 14. Benson RBJ. 2018 Dinosaur macroevolution and macroecology. Ann. Rev. Ecol. Evol. Syst. 49 , 379-408. (10.1146/annurev-ecolsys-110617-062231) 15. Vavrek MJ, Larsson HCE. 2010 Low beta diversity of Maastrichtian dinosaurs of North America. PNAS 107 , 8265-8268. (10.1073/pnas.0913645107)20404176 16. Chiarenza AA, Mannion PD, Lunt DJ, Farnsworth A, Jones LA, Kelland S-J, Allison PA. 2019 Ecological niche modelling does not support climatically-driven dinosaur diversity decline before the Cretaceous/Paleogene mass extinction. Nat. Comm. 10 , 1091. (10.1038/s41467-019-08997-2) 17. Dean CD, Chiarenza AA, Maidment SCR. 2020 Formation binning: a new method for increased temporal resolution in regional studies, applied to the Late Cretaceous dinosaur fossil record of North America. Palaeontology 63 , 881-901. (10.1111/pala.12492) 18. Mallon JC, Ott C, Larson PL, Iuliano EM, Evans DC. 2016 Spiclypeus shipporum gen. et sp. nov., a boldly audacious new chasmosaurine ceratopsid (Dinosauria: Ornithischia) from the Judith River Formation (Upper Cretaceous: Campanian) of Montana, USA. PLoS ONE 11 , e0154218. (10.1371/journal.pone.0154218)27191389 19. Chiba K, Ryan MJ, Fanti F, Loewen MA, Evans DC. 2018 New material and systematic re-evaluation of Medusaceratops lokii (Dinosauria, Ceratopsidae) from the Judith River Formation (Campanian, Montana). J. Paleontol. 92 , 272-288. (10.1017/jpa.2017.62) 20. Carrano MT, Janis CM, Sepkoski JJ. 1999 Hadrosaurs as ungulate parallels: lost lifestyles and deficient data. Acta Palaeontol. Pol. 44 , 237-261. 21. Horner JR, Goodwin MB. 2006 Major cranial changes during Triceratops ontogeny. Proc. R. Soc. B 273 , 2757-2761. (10.1098/rspb.2006.3643) 22. Witmer LM, Ridgely RC. 2008 Structure of the brain cavity and inner ear of the centrosaurine ceratopsid dinosaur Pachyrhinosaurus based on CT scanning and 3D visualization. In A New Horned Dinosaur from an Upper Cretaceous Bone Bed in Alberta (eds PJ Currie, W Langston, DH Tanke), pp. 117-144. Ottawa, Canada: NRC Research Press. 23. Bell PR, Snively E, Schychoski L. 2009 A comparison of jaw mechanics in hadrosaurid and ceratopsid dinosaurs using finite element analysis. Anat. Rec. 292 , 1338-1351. (10.1002/ar.20978) 24. Evans DC, Ridgely R, Witmer LM. 2009 Endocranial anatomy of Lambeosaurine hadrosaurids (Dinosauria: Ornithischia): a sensorineural perspective on cranial crest function. Anat. Rec. 292 , 1315-1337. (10.1002/ar.20984) 25. Schweitzer MH et al. 2009 Biomolecular characterization and protein sequences of the Campanian hadrosaur B. canadensis. Science 324 , 626-631. (10.1126/science.1165069)19407199 26. Lyson TR, Longrich NR. 2011 Spatial niche partitioning in dinosaurs from the latest Cretaceous (Maastrichtian) of North America. Proc. R. Soc. B 278 , 1158-1164. (10.1098/rspb.2010.1444) 27. Prieto-Márquez A. 2010 Global historical biogeography of hadrosaurid dinosaurs. Zoo. J. Linn. Soc. 159 , 503-525. (10.1111/j.1096-3642.2010.00642.x) 28. Mallon JC, Evans DC, Ryan MJ, Anderson JS. 2012 Megaherbivorous dinosaur turnover in the Dinosaur Park Formation (upper Campanian) of Alberta, Canada. Palaeogeog. Palaeoclimat. Palaeoecol. 350–352 , 124-138. (10.1016/j.palaeo.2012.06.024) 29. Mallon JC, Anderson JS. 2013 Skull ecomorphology of megaherbivorous dinosaurs from the Dinosaur Park Formation (upper Campanian) of Alberta, Canada. PLoS ONE 8 , e67182. (10.1371/journal.pone.0067182)23874409 30. Knapp A, Knell RJ, Farke AA, Loewen MA, Hone DWE. 2018 Patterns of divergence in the morphology of ceratopsian dinosaurs: sympatry is not a driver of ornament evolution. Proc. R. Soc. B 285 , 20180312. (10.1098/rspb.2018.0312) 31. Terrill DF, Henderson CM, Anderson JS. 2020 New application of strontium isotopes reveals evidence of limited migratory behaviour in Late Cretaceous hadrosaurs. Biol. Lett. 16 , 20190930. (10.1098/rsbl.2019.0930)32126185 32. Lehman TM. 1987 Late Maastrichtian paleoenvironments and dinosaur biogeography in the Western Interior of North America. Palaeogeog. Palaeoclimat. Palaeoecol. 60 , 187-217. (10.1016/0031-0182(87)90032-0) 33. Lehman TM. 2001 Late Cretaceous dinosaur provinciality. In Mesozoic vertebrate life (eds DH Tanke, K Carpenter), pp. 310-328. Bloomington, IN: Indiana University Press. 34. Sampson SD, Loewen MA, Farke AA, Roberts EM, Forster CA, Smith JA, Titus AL. 2010 New horned dinosaurs from Utah provide evidence for intracontinental dinosaur endemism. PLoS ONE 5 , e12292. (10.1371/journal.pone.0012292)20877459 35. Gates TA et al. 2010 Biogeography of terrestrial and freshwater vertebrates from the Late Cretaceous (Campanian) Western Interior of North America. Palaeogeog. Palaeoclimat. Palaeoecol. 291 , 371-387. (10.1016/j.palaeo.2010.03.008) 36. Berry K. 2018 Paleontological evidence against a major geographic barrier at about the paleolatitude of Colorado, USA, during the Late Campanian (Late Cretaceous): the conspicuous absence of endemic subclades of chamosaurine ceratopsid (horned) dinosaurs and its significance. Mountain Geol. 55 , 5-18. (10.31582/rmag.mg.55.1.5) 37. Lucas SG, Sullivan RM, Lichtig AJ, Dalman SG, Jasinski SE. 2016 Late Cretaceous dinosaur biogeography and endemism in the Western Interior Basin, North America: a critical re-evaluation. In Cretaceous Period: biotic diversity and biogeography, vol. 71 (eds A Khosla, SG Lucas), pp. 195-215. Albuquerque, NM: New Mexico Museum of Natural History Science Bulletin. 38. Burgener L, Hyland E, Griffith E, Mitášová H, Zanno LE, Gates TA. 2021 An extreme climate gradient-induced ecological regionalization in the Upper Cretaceous Western Interior Basin of North America. GSA Bull. (10.1130/B35904.1) 39. Longrich NR. 2014 The horned dinosaurs Pentaceratops and Kosmoceratops from the upper Campanian of Alberta and implications for dinosaur biogeography. Cret. Res. 51 , 292-308. (10.1016/j.cretres.2014.06.011) 40. Fowler DW. 2017 Revised geochronology, correlation and dinosaur stratigraphic ranges of the Santonian-Maastrichtian (Late Cretaceous) formations of the Western Interior of North America. PLoS ONE 12 , e0188426. (10.1371/journal.pone.0188426)29166406 41. Dunne EM, Close RA, Button DJ, Brocklehurst N, Cashmore DD, Lloyd GT, Butler RJ. 2018 Diversity change during the rise of tetrapods and the impact of the ‘Carboniferous rainforest collapse’. Proc. R. Soc. B 285 , 20172730. (10.1098/rspb.2017.2730) 42. Button DJ, Lloyd GT, Ezcurra MD, Butler RJ. 2017 Mass extinctions drove increased faunal cosmopolitanism on the supercontinent Pangaea. Nat. Comm. 8 , 733. (10.1038/s41467-017-00827-7) 43. Scannella JB, Horner JR. 2011 ‘Nedoceratops’: an example of a transitional morphology. PLoS ONE 6 , e28705. (10.1371/journal.pone.0028705)22194891 44. Farke AA. 2011 Anatomy and taxonomic status of the chasmosaurine ceratopsid Nedoceratops hatcheri from the Upper Cretaceous Lance Formation of Wyoming, USA. PLoS ONE 6 , e16196. (10.1371/journal.pone.0016196)21283763 45. Zhang J, Wang X, Wang Q, Jiang S, Cheng X, Li N, Qiu R. 2019 A new saurolophine hadrosaurid (Dinosauria: Ornithopoda) from the Upper Cretaceous of Shandong, China. Ann. Brazil. Acad. Sci. 91 , e20160920. (10.1590/0001-3765201720160920) 46. Cruzado-Caballero P, Powell J. 2017 Bonapartesaurus rionegrensis, a new hadrosaurine dinosaur from South America: implications for phylogenetic and biogeographic relations with North America. J. Vert. Palaeontol. 37 , e1289381. (10.1080/02724634.2017.1289381) 47. Mori H, Druckenmiller PS, Erickson GM. 2016 A new Arctic hadrosaurid from the Prince Creek Formation (lower Maastrichtian) of northern Alaska. Acta Palaeontol. Pol. 61 , 15-32. (10.4202/app.00152.2015) 48. Lehman TM, Wick SL, Wagner JR. 2016 Hadrosaurian dinosaurs from the Maastrichtian Javelina Formation, Big Bend National Park, Texas. J. Paleontol. 90 , 333-356. (10.1017/jpa.2016.48) 49. Gates TA, Jinnah Z, Levitt C, Getty MA. 2014 New hadrosaurid (Dinosauria, Ornithopoda) specimens from the lower-middle Campanian Wahweap Formation of southern Utah. In Hadrosaurs (eds DA Eberth, DC Evans), pp. 156-173. Bloomington, IN: Indiana University Press. 50. Prieto-Márquez A, Dalla Vecchia FM, Gaete R, Galobart À. 2013 Diversity, relationships and biogeography of the Lambeosaurine dinosaurs from the European Archipelago, with description of the new aralosaurin Canardia garonnensis. PLoS ONE 8 , e69835. (10.1371/journal.pone.0069835)23922815 51. Godefroit P, Bolotsky YL, Bolotsky IY. 2012 Osteology and relationships of Olorotitan arharensis, a hollow-crested hadrosaurid dinosaur from the latest Cretaceous of far eastern Russia. Acta Palaeontol. Pol. 57 , 526-560. (10.4202/app.2011.0051) 52. Bell MA, Lloyd GT. 2015 Strap: stratigraphic tree analysis for Palaeontology. Available from: cran.r-project.org/web/packages/strap. 53. R Core Team. 2019 R: a language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. See www.R-project.org. 54. Kubo T. 2019 Biogeographic network analysis of Cretaceous terrestrial tetrapods: a phylogeny-based approach. Syst. Biol. 68 , 1034-1051. (10.1093/sysbio/syz024)31135923 55. Lamigueiro OP, Hijmans R. 2021. rasterVis: visualization methods for Raster Data. Available from: oscarperpinan.github.io/rastervis. 56. Pinheiro J, Bates D, DebRoy S, Sarkar D, ESIPACK authors, Hesiterkamp S, Van Willigen B, Ranke J, R Core Team. 2018. nlme: Linear and nonlinear mixed effects models. Available from: cran.r-project.org/web/packages/nlme. 57. Barton K. 2018. MuMIn: multi-model inference. Available from: cran.r-project.org/web/packages/MuMIn. 58. Spiess AN. 2018. qpcR: modelling and analysis of real-time PCR data. Available from: cran.r-project.org/web/packages/qpcR. 59. Revell LJ. 2017. phytools: phylogenetic tools for comparative biology (and other things). Available from: cran.r-project.org/web/packages/phytools. 60. Raup DM. 1976 Species diversity in the Phanerozoic: an interpretation. Paleobiology 2 , 289-297. (10.1017/S0094837300004929) 61. Smith AB, McGowan AJ. 2005 Cyclicity in the fossil record mirrors rock outcrop area. Biol. Lett. 1 , 443-445. (10.1098/rsbl.2005.0345)17148228 62. Smith AB, Gale AS, Monks NEA. 2001 Sea-level change and rock-record bias in the Cretaceous: a problem for extinction and biodiversity studies. Paleobiology 27 , 241-253. (10.1666/0094-8373(2001)027<0241:SLCARR>2.0.CO;2) 63. Butler RJ, Benson RJB, Carrano MT, Mannion PD, Upchurch P. 2011 Sea level, dinosaur diversity and sampling biases: investigating the ‘common cause’ hypothesis in the terrestrial realm. Proc. R. Soc. B 278 , 1165-1170. (10.1098/rspb.2010.1754) 64. Huber BT, Norris RD, MacLeod KG. 2002 Deep-sea paleotemperature record of extreme warmth during the Cretaceous. Geology 30 , 123-126. (10.1130/0091-7613(2002)030<0123:DSPROE>2.0.CO;2) 65. Cantalpiedra JL, Soledad Domingo M, Domingo L. 2018 Multi-scale interplays of biotic and abiotic drivers shape mammalian sub-continental diversity over millions of years. Sci. Rep. 8 , 13413. (10.1038/s41598-018-31699-6)30194335 66. Maidment SCR, Dean CD, Mansergh RI, Butler RJ. 2021 Data from: Deep-time biodiversity patterns and the dinosaurian fossil record of the Late Cretaceous Western Interior, North America. Dryad Digital Repository. (10.5061/dryad.bcc2fqzbz)
PMC008xxxxxx/PMC8220270.txt
==== Front Proc Biol Sci Proc Biol Sci RSPB royprsb Proceedings of the Royal Society B: Biological Sciences 0962-8452 1471-2954 The Royal Society 34157873 10.1098/rspb.2021.0226 rspb20210226 10017058Evolution Research Articles A highly conserved ontogenetic limb allometry and its evolutionary significance in the adaptive radiation of Anolis lizards A highly conserved ontogenetic limb allometry and its evolutionary significance in the adaptive radiation of Anolis lizards http://orcid.org/0000-0003-4648-6950 Feiner Nathalie Conceptualization Data curation Formal analysis Funding acquisition Investigation Methodology Project administration Resources Supervision Visualization Writing-original draft Writing-review & editing [email protected] http://orcid.org/0000-0002-7948-2860 Jackson Illiam S. C. Data curation Investigation Methodology Resources Writing-review & editing http://orcid.org/0000-0003-4931-8880 Van der Cruyssen Eliane Data curation Formal analysis Methodology Writing-review & editing http://orcid.org/0000-0003-1293-5842 Uller Tobias Conceptualization Formal analysis Funding acquisition Investigation Methodology Project administration Resources Writing-original draft Writing-review & editing Department of Biology, Lund University, Sweden Electronic supplementary material is available online at https://doi.org/10.6084/m9.figshare.c.5464696. 30 6 2021 June 30, 2021 23 6 2021 June 23, 2021 23 6 2021 June 23, 2021 288 1953 2021022628 1 2021 January 28, 2021 1 6 2021 June 1, 2021 © 2021 The Authors. 2021 https://creativecommons.org/licenses/by/4.0/ Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited. Diversifications often proceed along highly conserved, evolutionary trajectories. These patterns of covariation arise in ontogeny, which raises the possibility that adaptive morphologies are biased towards trait covariations that resemble growth trajectories. Here, we test this prediction in the diverse clade of Anolis lizards by investigating the covariation of embryonic growth of 13 fore- and hindlimb bones in 15 species, and compare these to the evolutionary covariation of these limb bones across 267 Anolis species. Our results demonstrate that species differences in relative limb length are established already at hatching, and are resulting from both differential growth and differential sizes of cartilaginous anlagen. Multivariate analysis revealed that Antillean Anolis share a common ontogenetic allometry that is characterized by positive allometric growth of the long bones relative to metapodial and phalangeal bones. This major axis of ontogenetic allometry in limb bones deviated from the major axis of evolutionary allometry of the Antillean Anolis and the two clades of mainland Anolis lizards. These results demonstrate that the remarkable diversification of locomotor specialists in Anolis lizards are accessible through changes that are largely independent from ontogenetic growth trajectories, and therefore likely to be the result of modifications that manifest at the earliest stages of limb development. Anolis lizards , adaptive radiation , allometries , diversification , limb morphology , macroevolution John Templeton Foundation http://dx.doi.org/10.13039/100000925 60501 Kungliga Fysiografiska Sällskapet i Lund http://dx.doi.org/10.13039/501100005753 cover-dateJune 30, 2021 ==== Body pmc1. Introduction Morphological diversification of ecologically specialized forms is a hallmark of adaptive radiations. The morphological differences that accrue are often substantial, but comparison of adult phenotypes reveals that even extreme cases of diversification tend to preserve covariation between characters [1–3]. One explanation for this is that the phenotypic variation that selection can act upon is highly structured by development [4]. Developmental integration can be observed and quantified in terms of the covariation between morphological characters, such as the bones of the limb, as they develop and grow [5–9]. If the regulation of development and growth were to exercise a substantial effect on adaptive diversification, distantly related species should share similar patterns of morphological covariation in ontogeny, and this may force adult morphologies to diverge along the same dimensions (i.e. developmental lines of least resistance [4,10]). While an increasing number of studies compares trait covariation across species (e.g. [1,11–14]), and sometimes relate these to ontogenetic trajectories [6,15,16], little has been done in the context of adaptive radiations. Thus, the extent to which ontogenetic covariation of morphological characters evolves, as well as the extent to which it contributes to the pattern of morphological diversification across the phylogeny, remains poorly understood. One outstanding opportunity to investigate these issues is the adaptive radiation of Anolis lizards, a highly diverse clade of almost 400 species that demonstrate extensive morphological diversification. While roughly two-thirds of all Anolis lizards are native to the mainland of Central and Southern America, many species are found on the islands of the Greater and Lesser Antilles. The match between morphology and microhabitat, best documented on the four Greater Antillean islands (Cuba, Hispaniola, Puerto Rico and Jamaica), is often spectacular and strongly pronounced in limb morphology, in particular hindlimb length [17,18]. For example, species with long limbs tend to be agile runners on tree trunks, whereas species with short limbs tend to be specialized for perching on twigs and branches [19,20]. This match is captured by the ecomorph classification, which recognizes up to six habitat specialist types (ecomorphs) that have repeatedly and convergently evolved on the four Greater Antillean islands [21–23]. However, not all Greater Antillean species are classified as ecomorphs and roughly 10% exhibit more unique morphologies. While limb morphology is arguably the most defining ecological feature of Anolis, the developmental basis of adaptive variation has received little attention. One study comparing limb ontogenetic growth between pairs of Anolis species belonging to different ecomorphs found that relative limb length differences are established very early in development and that differential growth contributes little, if anything, to ecomorph-specific differences [24]. Another study on two Anolis species confirmed the early appearance of species-specific limb length differences, but found that these increase over developmental time, even after hatching [25]. Besides this question of which period of limb development that is evolutionarily labile, the extent to which growth trajectories have shaped evolutionary diversification of the Anolis adaptive radiation is yet to be addressed. Here, we fill these gaps in knowledge by investigating the covariation of 13 limb bones during embryonic development of 15 species, and subsequently comparing the ontogenetic covariation in these species against patterns of diversification of the same limb bones across adults of 267 island and mainland Anolis species. The 15 species of the ontogenetic series include eight species of four ecomorphs and seven non-ecomorph species, all native to the Greater or Lesser Antilles (figure 1). Note that the ontogenetic series included only developmental stages up until hatching, but no juvenile stages. Using these datasets, we addressed the following questions. Firstly, we tested if species-specific relative limb length is established already at hatching, or even earlier, and if these differences emerge through differential growth rate of individual bones or differential size of limb anlagen. Secondly, we used multivariate ontogenetic allometry to assess if the covariation of limb bones during embryonic development is conserved between species. Thirdly, we applied the same approach to an equivalent morphological dataset comprising adult limb data of 267 species and assessed if evolutionary covariation is conserved between the three major biogeographic groups. Finally, we tested if the major axis of embryonic bone growth is aligned with evolutionary allometry, as would be expected if growth trajectories represented a developmental line of least resistance during adaptive radiation. Figure 1. Phylogenetic relationship of species included in this study and selected adult phenotypes. Phylogenetic relationships (a) are adapted from Poe et al. [26]. The 15 species that are comprised in the ontogenetic dataset are highlighted with their scientific names. Circles on tree tips indicate ecomorph classifications, and biogeographic groups are marked by colour-coded arcs. Sample sizes of each species in the ontogenetic dataset are given in brackets after species names. Adult phenotypes are shown for six Anolis species (b). Note that all individuals are perching on the same stick to convey relative differences in body size. (Online version in colour.) 2. Material and methods (a) Ontogenetic dataset We established breeding groups for 15 Anolis species at Lund University. A. sagrei were collected at Palm Coast, Florida and A. equestris in Miami, Florida in April 2016 and brought to the animal facility at Lund University. The remaining species were purchased through the commercial pet trade or private breeders. Each breeding group consisted of one male and two females (except for A. porcus that consisted of a single female) and a minimum of one breeding group per species was established. Since our aim was to compare embryonic development across species, individual variation within species was deemed sufficiently small to be neglected in this study. Breeding groups were housed in 80 l plastic cages (Wham Crystal box with mesh on top, 590 × 390 × 415 mm), except for A. equestris and A. porcus that were kept in Zoo Med Reptibreeze open air screen cages (610 × 610 × 1200 mm). During breeding, lizards were kept at a light cycle of 12 L : 12 D and given access to basking lights (60 W) for 10 h per day and a UV light (Exo-Terra 10.0 UVB fluorescent tube) for 6 h per day. Live food (mealworms, crickets, wax worms, cockroaches (only for A. equestris) and snails (only for A. porcus)) were provided ad libitum. Eggs were collected every day and incubated at 26°C in individual small plastic containers filled two-thirds with moist vermiculite (5 : 1 vermiculite:water volume ratio) and sealed with clingfilm. While different Anolis species probably have different optimal incubation temperatures [27], 26°C is a good compromise for the species included in the present work. To first establish the incubation time for each species, at least one embryo was allowed to hatch (electronic supplementary material, table S1). On the basis of this species-specific incubation time, embryos were dissected at 12 evenly spaced time points from oviposition to hatching to obtain an ontogenetic series of each species that provides a broad overview of the variation of morphogenesis (electronic supplementary material, figure S1). Eggs were dissected in phosphate-buffered saline (PBS) and embryos were sacrificed, fixed in 4% paraformaldehyde for 1–4 days and stored in 100% methanol. Embryos were staged according to the Anolis staging table [28]. To accurately measure individual limb bones, embryos were subjected to a clearing and staining procedure [29]. We found that this procedure produced satisfactory results for embryos from stage 11 onwards and therefore excluded younger stages. Stained limbs (dissociated from the rest of the body) were mounted horizontally on a petri-dish and photographed under a dissection microscope (Olympus SZX10). ImageJ [30] was used to record measurements of the limb bones (in millimetre to the closest 0.001 mm) that span the maximal proximo-distal axis of both fore- and hindlimbs. These include the long bones (humerus, ulna, femur and tibia), metapodials (metatarsus and metacarpal), and the phalanges of the longest digit of both fore- and hindlimb (electronic supplementary material, figure S2). In total, we recorded six individual bone elements for the forelimb, and seven elements for the hindlimb. The body posture of embryos makes the common measure of body size in lizards, snout-vent length, prone to measurement error. Therefore, we recorded a dorsal view of the flattened embryo to obtain a measure of the length from the first cervical vertebra to the sacral vertebra, an estimate of body size that we term spine length. These measurements were highly repeatable (Pearson's product-moment correlation rP = 0.980, n = 47). A small number (n = 253; 2.95%) of individual measurements were missing due to fractured bones, and we imputed these missing values using the ‘pcaMethods’ package [31] based on all linear measurement per species. (b) Evolutionary dataset Measurements of adult limb elements were obtained from micro-CT scanning of museum specimens that were published elsewhere [12,13]. In brief, specimens were scanned using a GE phoenix v|tome|x m system (source voltage 100 kV; source current 200 µA; isometric voxel size 55–75 µm) at the Nanoscale Facility of the University of Florida, US. Reconstructed image stacks (software GE phoenix datos|x CT) were further processed using VGStudio MAX software (v. 3.2) by applying manual thresholding to extract surface models of skeletal structures. Linear measurements were directly obtained using the VGStudio MAX software. For each lizard, we measured the same limb elements as detailed above (to the closest 0.01 mm). This was achieved by placing one landmark each on the proximal and on the distal end of a bone and extracting the distance between these two points in 3D space. These measurements were highly repeatable (Pearson's product-moment correlation rP = 0.992, n = 40). A small number (n = 51; 0.47%) of individual measurements were missing due to fractured bones, and we imputed these missing values using the ‘pcaMethods’ package [31] based on all linear measurements of all individuals. Estimates of body size were inferred from the centroid size of the hip girdles and were shown to be highly correlated with snout–vent length (see Methods in [12]). (c) Statistical analyses All limb elements of the ontogenetic and the phylogenetic dataset were log-transformed prior to analyses. Linear models were used to assess the relationship between total limb length and spine length, and to test if species differ in their relative limb length at early embryonic stages, at hatching and as adults. The relative limb length for fore- and hindlimbs was captured by the sum of their respective limb bones divided by spine length for embryos and hatchlings, and divided by centroid size for adults. We used ANCOVA and MANCOVA to test for species differences in the bivariate allometry between limb length (total or individual bones) and spine length. Multivariate allometry, the relative proportions of individual limb bones in relation to overall size, was quantified for fore- and hindlimbs separately, using a principal component analysis (PCA) approach [32]. Determining the dimension in multivariate space that maximizes variation, the patterns of covariation in a given dataset can be quantified and this approach is commonly used for studying ontogenetic, static and evolutionary allometries [33,34]. To test if species differed in their multivariate ontogenetic allometric slopes, we performed principal component analyses on log-transformed bone measurements for each species. The coefficients of the first principal components (i.e. the allometric vectors) serve as an indicator of allometry for each limb element and is effectively a measure of ontogenetic covariation between limb elements [33]. We used these coefficients in three ways. First, as observations in a second PCA that resulted in the construction of an ‘allometric space’ [15,35], which visually presents patterns of multivariate allometry. Second, coefficients were used to test if individual limb elements scale isometrically with size. If all individual limb elements would contribute equally to the increase in size over ontogeny, the coefficients of the allometric vector would be equal to p−0.5 where p is the number of elements [34]. We applied non-parametric bootstrapping (1000 iterations) of the dataset to test if the range of resampled coefficients includes the value expected under multivariate isometry (coefficient = p−0.5). Third, coefficients were used as vectors in multivariate space that can be compared between pairs of species to extract an angle (arc cosine of the inner product of the two vectors) that describes how well two ontogenetic allometries are aligned [33]. To assess whether or not the observed angles between pairs of species were larger than expected if all species share a common allometric pattern, we created a dataset that conformed to a shared allometric pattern and applied non-parametric bootstrapping to generate a distribution of test statistics [32] (see electronic supplementary material for extended methods). To test if the ontogenetic and phylogenetic allometries were aligned, we compared the common ontogenetic allometric vector with the evolutionary allometric vector of the same major group. The Anolis clade originated on the mainland, but has colonized Caribbean islands in two waves—one to the Southern Lesser Antilles, and one to the Greater and Northern Lesser Antilles—and the latter gave rise to a re-colonization of the mainland. Since the Southern Lesser Antillean clades contain only few species, we divide Anolis lizards into three major groups. The first two include the group that gave rise to an independent diversification on the mainland (the Primary Mainland group, MLpri; 29 species in our dataset) and the group that diversified on the mainland following a re-colonization from the Greater Antilles (the Secondary Mainland group, MLsec; 107 species in our dataset). The remaining group includes all the Greater Antillean (110 species in our dataset) and the Northern Lesser Antillean Anolis (11 species in our dataset). We refer to this group as the Antillean group (Ant). One species in our embryonic dataset, A. roquet, is native to the Southern Lesser Antilles and does not belong to any of the three major groups, and we therefore excluded this species from the ontogenetic dataset for this analysis. We assessed the alignment of the common ontogenetic allometric vector (derived for the remaining 14 species) to the evolutionary allometric vector of their native Antillean group (Ant), and to the two mainland groups (MLpri and MLsec). Evolutionary allometric vectors (pPC1) were derived from phylogenetic PCAs in the R package ‘phytools’ [36] for each of the three major groups based on species means of each measurement and the phylogenetic tree published by Poe et al. [26] with modifications described in [13]. We first assessed how well the phylogenetic allometric vectors of the three major groups are aligned with each other. Secondly, we calculated the angle between the common ontogenetic allometric vector (cPC1) and the evolutionary allometric vectors (pPC1) of each major Anolis group (observed angles Θont-Ant, Θont-MLpri, Θont-MLsec) and assessed their statistical significance by bootstrapping from rotated datasets (see electronic supplementary material for extended methods). Since we compared ontogenetic and phylogenetic major axis that were derived by slightly different methodologies (common PCA versus phylogenetic PCA), we repeated all analyses that involve phylogenetic PCAs and replaced these with standard PCAs and accordingly performed non-parametric bootstrapping to generate the sampling distributions. This alternative approach produced qualitatively identical results that are presented in the electronic supplementary material. All statistical analyses were conducted in R (v. 3.6.1) [37]. 3. Results The final dataset comprised linear morphometric measurement of 374 individual embryos and hatchlings belonging to 15 Anolis species (mean, 24.93 individuals per species ± 12.99 standard deviations, s.d.) and 693 individual adults belonging to 267 Anolis species (mean, 2.60 individuals per species ± 1.58 s.d.). Eleven species of the embryonic dataset belong to the Greater Antillean group, three to the Northern Lesser Antillean group that is nested within the Greater Antillean group and one to the Southern Lesser Antillean group that diverged from the Primary Mainland group early in the evolutionary history of Anolis (figure 1). (a) Bivariate allometry shows that species-specific differences in relative limb length are fully expressed at hatching We first focused on the total fore-and hindlimb length (sum of all individual bones) in relation to body size to address when in ontogeny species-specific differences emerge. Species differences in relative limb length were substantial in both adults and hatchlings (all p-values < 0.001), and were detectable, albeit to a lesser degree, in embryos at stage 11, the earliest stage at which repeatable measurements were possible (forelimb: F1,13 = 2.28, p = 0.025; hindlimb: F1,13 = 1.89, p = 0.064; electronic supplementary material, table S2). Using mean values for each of the 15 species at hatching and as adults, these two life stages show a significant correlation of their relative hindlimb length (Pearson's product-moment correlation r = 0.616, p-value = 0.014) and a weak correlation of their relative forelimb length (r = 0.361, p-value = 0.187). These results suggest that species differ both in the size of initial limb anlagen and in growth trajectories. Accordingly, an ANCOVA on total limb length revealed species differences in both intercept and bivariate allometric slopes (electronic supplementary material, table S3). For total fore- and hindlimb length, nearly all slopes were significantly larger than 1, indicating that limbs grow proportionally faster than the trunk (figure 2; electronic supplementary material, table S4). The largest species contrasts in growth rates (i.e. slopes) tended to involve species with exceptionally low allometric slopes (e.g. A. equestris, A. rejectus; electronic supplementary material, table S5). In agreement with these results on total limb length, MANCOVA on individual limb elements showed significant species and species-by-size effects (electronic supplementary material, table S3) which were equally distributed across individual limb elements (electronic supplementary material, table S6). Figure 2. Patterns of bivariate allometry across 15 Anolis species. Regressions of (a) fore- and (b) hindlimbs against spine length are plotted for each of the 15 Anolis species. The two species with exceptionally large body size and relatively short limbs are A. equestris and A. porcus. (Online version in colour.) (b) Multivariate ontogenetic allometries are conserved across Anolis species While these results demonstrate that limbs of different Anolis species grow apart in terms of overall length relative to body size, they do not address if the relative proportions of individual limb elements are changing over ontogeny, and if this change is consistent across species. To address these questions, we first tested if the multivariate ontogenetic allometries differ between Anolis species. Calculating the angles between allometric vectors for pairs of species revealed that the alignment between allometric trajectories generally lies within the 95% CIs of the bootstrapped angles, and are therefore consistent with one ontogenetic allometry shared by all 15 Anolis species (electronic supplementary material, tables S7 and S8; mean angle for forelimb was 6.03° and for hindlimb 4.10°). Out of 105 species pairs, only 14 (forelimb; 13.3%) and 11 (hindlimb; 10.5%) had angles between allometric trajectories that were significantly larger than those expected from a shared ontogenetic allometric (electronic supplementary material, tables S7 and S8). The difference in multivariate allometry between species can be visualized using a second PCA on the allometric vectors of all 15 species. The result describes an allometric space that is commonly used to reveal clustering of species according to shared patterns of allometric growth [15,35]. Constructing such allometric spaces revealed that A. rejectus (forelimb and hindlimb) and A. porcus (hindlimb) are differentiated from all other species on PC1 of this allometric space (figure 2). The other species formed a continuous distribution along PC1 and PC2. It is noticeable that species belonging to the same ecomorph did not cluster tightly with each other and that ecomorphs did not occupy a private location in allometric space (figure 3). This again illustrates how similar ontogenetic multivariate allometries are for these 15 species, despite the large differences in limb length relative to body size. Figure 3. Allometric space of limb growth of six forelimb and seven hindlimb bones for 15 Anolis species. Allometric spaces of (a) fore- and (b) hindlimbs visualize the broad patterns of allometric growth of the 15 Anolis species. Allometric spaces were constructed from allometric vectors (PC1 coefficients) of each species. Species belonging to the same ecomorph as classified by Losos [23] are marked by a square (grass-bush) or a circle (trunk-ground). (Online version in colour.) (c) Major axis of evolutionary allometries are not well aligned with ontogenetic allometries We first approached phylogenetic patterns of covariation in fore- and hindlimb elements through a phylogenetic PCA (pPCA) of all 267 Anolis species. This revealed no clear differentiation between the three major groups (Primary Mainland, Secondary Mainland and Antilles; electronic supplementary material, figure S3). The first principal component (pPC1) explained 93.64% (forelimb; using a standard PCA: 96.21%) and 95.01% (hindlimb; standard PCA: 96.13%) of the total variation and was tightly correlated with the centroid size of the pelvic girdle, a proxy for body size (forelimb: rP = 0.971; hindlimb: rP = 0.951; both p-value < 0.001; standard PCA: forelimb: rP = 0.971; hindlimb: rP = 0.950; both p-value < 0.001). We therefore interpret pPC1 as an allometric vector that describes how individual limb elements scale with overall limb length. Before we addressed to what extent ontogenetic and evolutionary allometries are aligned with each other, we first asked whether patterns of evolutionary allometries are already established at the hatching stage. To this end, we compared the angles between evolutionary allometric vectors derived from hatchlings and adults of our 15 focal species. We found that the observed angles (Θforelimb: 5.42°; Θhindlimb: 4.97°; for results from an alternative analysis using standard PCA instead of phylogenetic PCA, see electronic supplementary material, table S5) lie within the 95% confidence interval of the bootstrap distribution derived from the sampling distribution of perfectly aligned allometric vectors (95% CIforelimb: 8.19°; 95% CIhindlimb: 9.28°; electronic supplementary material, table S9). This indicates that the evolutionary allometry of the 15 species is indistinguishable between hatchlings and adults. To test if the major axis of ontogenetic growth of the limb bones is similar to the pattern of evolutionary diversification of limb morphology in the major groups, we assessed whether or not the coefficients of the allometric vectors showed similar patterns. Using bootstrapping of the ontogenetic major axis (common PC1 of 14 species since A. roquet was excluded from this analysis; see Material and methods), we found that the ontogenetic allometric vectors (coefficients of cPC1) of both fore- and hindlimbs deviated significantly from isometry (electronic supplementary material, tables S10 and S11). The ontogenetic coefficients for the four long bones (humerus, ulna, femur and tibia) were positive (i.e. a proportional increase relative to the other limb elements). By contrast, evolutionary coefficients for the long bones were nearly all significantly more negative. Similarly, the metatarsus showed a positive allometric growth pattern in ontogeny, but its evolutionary allometry revealed isometry in all three major groups. In general, the three major groups exhibited largely similar patterns of evolutionary allometry, evidenced by, for example, consistently positive coefficients of the last phalangeal element of the hindlimb. This was supported by quantifying the alignment between pairs of evolutionary allometric vectors that demonstrated that evolutionary allometries were highly similar between the three groups (all angles < 5.35°; electronic supplementary material, tables S12–15). This qualitative pattern was quantitatively confirmed by assessing the alignment in terms of angles between the cPC1 vector of the ontogenetic dataset to the pPC1 vector of its native group (i.e. the Antillean group) and to the two more distantly related mainland groups. We found that the ontogenetic allometric vector (cPC1) of our 14 focal species was most closely aligned with the pPC1 of its native group for the forelimb (Θont-Ant: 7.27°; Θont-MLpri: 8.60°; Θont-MLsec: 8.10°), but not for the hindlimb (Θont-Ant: 7.29°; Θont-MLpri: 7.45°; Θont-MLsec: 5.24°). Comparing these angles against a bootstrapped distribution derived from a shared onto-phylogenetic allometric axis demonstrated that ontogenetic and phylogenetic axis are not well aligned with each other (figure 4; electronic supplementary material, figure S4). Figure 4. Alignment between ontogenetic allometry and phylogenetic allometry. Each plot represents a comparison between the ontogenetic allometric vector and the phylogenetic allometric vectors of the Antilles group (a,d), the Primary Mainland group (b,e) and the Secondary Mainland group (c,f) for fore- (a–c) and hindlimbs (d–f). The ontogenetic allometry is derived from species that belong to the Antillean group. Each plot shows the observed angles (Θont-Ant, Θont-MLpri, Θont-MLsec) between ontogenetic and phylogenetic allometric vectors as a red arrow. Grey bars indicate the frequency distribution of angles Θ drawn from modified datasets that simulated perfect alignment (derived from 1000 bootstrap replicates), and blue dashed lines mark their 95% confidence intervals. For results of an alternative analysis that used standard PCA instead of phylogenetic PCA, see electronic supplementary material, figure S4. (Online version in colour.) 4. Discussion Unravelling the factors that shape adaptive evolution requires studies of both generative and selective processes. Decades of ecological research on Anolis lizards have come a long way in establishing the adaptive function of species differences in limb morphology, and many studies have demonstrated that traits such as relative limb length are under selection in contemporary populations [38–42]. By contrast, the developmental basis of variation in limb morphology and its consequences for the adaptive diversification of Anolis are little studied to date (but see [24,25]). Studies that investigate how morphological characters co-develop and coevolve are useful in this respect, and here we emphasize three aspects in particular. Firstly, comparisons of growth trajectories between diverse morphologies can identify the developmental periods that are most evolutionarily labile, thereby pointing towards developmental mechanisms that are responsible for morphological evolvability. Our results demonstrate that, in Anolis, species differences in limb morphology are well established at hatching (both in terms of total relative length as well as the proportion of individual limb elements). This suggests that post-hatching regulation of limb growth might be a less important mechanism for the generation of adaptive variation than pre-hatching mechanisms. However, since our study did not include post-hatching stages, we cannot infer if growth trajectories in juveniles differ between Anolis species. Nevertheless, this conclusion is consistent with a previous study [12], which showed that modification of the plastic responses to mechanical stress imposed by running versus climbing contributed little to adaptive evolution of the locomotor skeleton in Anolis. However, plasticity might still play a role in explaining differential bone growth in embryonic development since mechanical stress induced by motility of limbs in ovo is known to modulate growth rates of long bones (see e.g. [43]). The molecular mechanisms that promote limb bone growth include the paracrine signalling of the insulin-like growth factor (IGF) pathway [44,45] and of the bone morphogenetic protein (BMP) pathway [46] that stimulate chondrocyte proliferation and differentiation. Irrespective of the mechanism underlying differential growth of limb element, our results demonstrate that differential growth alone is not responsible for the species differences in relative limb length in Anolis lizards, since these differences are to some extent manifested already in the early limb anlagen. Furthermore, we show that differential growth cannot account for species differences in the relative proportion of limb bones. Thus, the results from both bivariate and multivariate analysis of limb allometry suggest that the adaptive features that define Anolis habitat specialists are to a large extent established at the earliest phase of limb development, prior to or during the emergence of cartilaginous condensations. Further pinpointing the exact emergence of species-specific differences will require a modified methodology that does not rely on quantifying cartilaginous structures (e.g. [47]). These results broadly agree with the findings of Wakasa et al. [25], who found that the difference in relative limb length between the short-limbed A. angusticeps and the long-limbed A. sagrei are evident before the formation of cartilaginous anlagen, and accumulate throughout embryonic and even post-hatching development. By contrast, another study by Sanger and colleagues [24] found that the long bones of two pairs of trunk-ground and trunk-crown specialists (ecomorphs that differ in relative limb length [23]) exhibit equal growth rates in their long bones relative to body size (i.e. bivariate allometric slopes [24]). While these observations agree with our conclusions in that species differences indeed are to a large extent established early in limb development (see also [47]), the lack of differential growth rates reported in this previous study [24] may be attributed to a more narrow taxon sampling that did not include more unique morphologies (e.g. our study included both ecomorph and non-ecomorph species). However, since in our study differential growth did not account for differences in the relative proportions of different limb bones, we suggest that much of the adaptive diversity in limb morphology in Anolis involves regulatory changes to genes involved in the early patterning of the limb, such as HoxA10, HoxD11 and HoxD12 [48]. Interestingly, the mechanisms of bone element morphogenesis differ between long bones, metapodials and phalangeal elements [49], which suggests that changes in limb length and proportions can result from a variety of different mechanisms, providing promising targets for future studies. Unfortunately, comparative developmental genetic data on the regulation of limb morphology are limited (but see [46]), but successful artificial selection for long-limbed mice largely occurred through an increase in the initial number of proliferative chondrocytes [50]. This is consistent with the early-acting mechanisms we predict to be responsible for the adaptive diversification of limb morphology in Anolis. Unravelling the developmental genetic basis of these mechanisms, together with genomic association studies, would be a powerful approach for identifying the developmental variabilities upon which selection can act, as exemplified by the research program on the beaks of Darwin's finches [51–53]. Since adaptive phenotypic change has to arise through developmental change, a second reason to study ontogenetic variability is that it can identify conserved developmental lines of least resistance that help to explain why adaptive diversification occurred in some directions and not others [4,54,55]. For example, jaws of cichlid fish show variational properties in development that appear to have directed evolutionary diversification [56,57]. While the ontogenetic allometry of the Anolis limb is highly conserved, we found that it was poorly aligned with the major axis of evolutionary allometry of both Antillean and mainland Anolis. In particular, the proportional increase in the length of long bones (proximal elements) was positive in embryonic development, but tended to be negative across species. Similarly, Anolis phalangeal (distal) elements exhibit patterns of coevolution with other limb bones that are distinct from their co-development. As explained above, these results suggest that the evolution of the Anolis limb involved modifications of developmental mechanisms early in ontogeny. These mechanisms may themselves exhibit developmental bias, which has been suggested to explain why the distal-most phalanges show the greatest variability and evolvability in birds [58]. Distal elements do appear to be particularly evolvable in Anolis as well. Across the phylogeny, differences in the length of distal bones account for much of the variation in relative limb length of different Anolis ecomorphs [25]. For example, Toro et al. [59] found that differences in the relative limb length between A. sagrei and A. carolinensis were most pronounced in the metatarsus and the longest toe. Similarly, Mahler et al. [60] reported that, among 22 morphological traits measured for 81 primarily Greater Antillean Anolis species, the length of the longest digit loaded highest on the first principal component. In addition to a possible developmental bias, the results for Anolis emphasize the functional importance of the phalangeal elements, which might be driven by their importance for the lizard's clinging ability, enabled through lamellae that cover the ventral side of phalangeal elements [61]. Finally, exploring the consistency of morphological covariation across different clades or environments can reveal insights into the relationship between adaptive diversification and developmental innovation [62]. For example, the butterfly genus Heteropsis has evolved correlations between eyespot colour patterns and elevated evolutionary rates that deviate from other Mycalesina butterflies, and this macroevolutionary shift coincides with a developmental innovation in eye spot development [63]. In Anolis, the major axis of evolutionary limb allometry has been conserved throughout the Antillean and two mainland radiations. Thus, despite Greater Antillean Anolis having generated ecomorph specialists that are (nearly [64]) absent from the mainland Anolis fauna (e.g. grass-bush, twig and crown-giant [65,66]), this adaptive radiation does not appear to involve profound changes to the underlying generative processes. Instead, the morphological variability of the Anolis limb seems perfectly capable of exploiting ecological opportunities, and the absence of certain morphologies from the mainland is better explained by ecological factors than by developmental innovations on islands [13,67]. This is consistent with a macroevolutionary perspective that suggests that drastic modifications of limb proportions are rare. In mammals, such innovations are restricted to bats [46,62], jerboas [68] and arguably humans [69] (see also the conflicting conclusions regarding kangaroos and their allies; e.g. [70,71]). Squamate reptiles are understudied in this respect, although there are several recent studies on the evolutionary modularity of the skull that demonstrate the utility of this macroevolutionary approach [72,73]. Given that limb reduction is a rather frequent evolutionary event in squamates [74], it would be interesting to address the extent to which this is associated with changes in modularity and integration of different limb bones. 5. Conclusion Our results show that both ontogenetic and phylogenetic allometries in limb bones are surprisingly conserved across Anolis species. However, evolutionary diversification of limb morphology has not followed the ontogenetic growth trajectory of limb bones, but is the result of modifications of developmental processes that act during early developmental stages. Supplementary Material Click here for additional data file. Acknowledgements We are grateful to private breeders for supplying us with stock animals. We thank David Blackburn and Edward L Stanley for providing us with access to the Nanoscale Research Facility (University of Florida, US) where all micro-CT scanning was conducted. We thank Dan Warner and Tim Mitchell for assistance with collecting animals in the field. Ethics All work was conducted according to the Lund University Local Ethical Review Process under the permit number Dnr M 31-16. Data accessibility Raw CT scans of adult Anolis specimens are deposited at MorphoSource under the project ID P1059, title ‘Anolis sp.’. Code and datasets, including a list of DOIs of individual CT scans deposited at MorphoSource, are available from the Dryad digital Repository: https://doi.org/10.5061/dryad.jsxksn08d [75]. Authors' contributions N.F.: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, supervision, visualization, writing-original draft, writing-review and editing; I.S.C.J.: data curation, investigation, methodology, resources, writing-review and editing; E.V.d.C.: data curation, formal analysis, methodology, writing-review and editing; T.U.: conceptualization, formal analysis, funding acquisition, investigation, methodology, project administration, resources, writing-original draft, writing-review and editing All authors gave final approval for publication and agreed to be held accountable for the work performed therein. Competing interests We declare we have no competing interests. Funding This research was supported by a grant from the John Templeton Foundation (60501) to T.U., a grant from the Royal Physiographic Society of Lund to N.F., a Wenner-Gren postdoctoral fellowship to N.F., and a Wallenberg Academy Fellowship from the Knut and Alice Wallenberg Foundation to T.U. ==== Refs References 1. Navalon G, Marugan-Lobon J, Bright JA, Cooney CR, Rayfield EJ. 2020 The consequences of craniofacial integration for the adaptive radiations of Darwin's finches and Hawaiian honeycreepers. Nat. Ecol. Evol. 4 , 270-278. (10.1038/s41559-019-1092-y)32015429 2. Kavanagh KD, Evans AR, Jernvall J. 2007 Predicting evolutionary patterns of mammalian teeth from development. Nature 449 , 427-432. (10.1038/nature06153)17898761 3. Felice RN, Randau M, Goswami A. 2018 A fly in a tube: macroevolutionary expectations for integrated phenotypes. Evolution 72 , 2580-2594. (10.1111/evo.13608)30246245 4. Uller T, Moczek AP, Watson RA, Brakefield PM, Laland KN. 2018 Developmental bias and evolution: a regulatory network perspective. Genetics 209 , 949-966. (10.1534/genetics.118.300995)30049818 5. Cheverud JM. 1996 Developmental Integration and the evolution of Pleiotropy. Am. Zool. 36 , 44-50. (10.1093/icb/36.1.44) 6. Klingenberg CP. 1996 Individual variation of ontogenies: a longitudinal study of growth and timing. Evolution 50 , 2412-2428. (10.1111/j.1558-5646.1996.tb03628.x)28565691 7. Cheverud JM. 1982 Relationships among ontogenetic, static, and evolutionary allometry. Am. J. Phys. Anthropol. 59 , 139-149. (10.1002/ajpa.1330590204)7149015 8. Klingenberg CP. 2010 Evolution and development of shape: integrating quantitative approaches. Nat. Rev. Genet. 11 , 623-635. (10.1038/nrg2829)20697423 9. Klingenberg CP. 2014 Studying morphological integration and modularity at multiple levels: concepts and analysis. Phil. Trans. R. Soc. B 369 , 20130249. (10.1098/rstb.2013.0249)25002695 10. Schluter D. 1996 Adaptive radiation along genetic lines of least resistance. Evolution 50 , 1766-1774. (10.2307/2410734)28565589 11. Haber A. 2016 Phenotypic covariation and morphological diversification in the ruminant skull. Am. Nat. 187 , 576-591. (10.1086/685811)27104991 12. Feiner N, Jackson ISC, Munch KL, Radersma R, Uller T. 2020 Plasticity and evolutionary convergence in the locomotor skeleton of Greater Antillean Anolis lizards. eLife 9 , e57468. (10.7554/eLife.57468)32788040 13. Feiner N, Jackson ISC, Stanley EL, Uller T. 2021 Evolution of the locomotor skeleton in Anolis lizards reflects the interplay between ecological opportunity and phylogenetic inertia. Nat. Commun. 12 , 1525. (10.1038/s41467-021-21757-5)33750763 14. Machado FA. 2020 Selection and constraints in the ecomorphological adaptive evolution of the skull of living Canidae (Carnivora, Mammalia) . Am. Nat. 196 , 197-215. (10.1086/709610) 15. Wilson LA, Sanchez-Villagra MR. 2010 Diversity trends and their ontogenetic basis: an exploration of allometric disparity in rodents. Proc. R. Soc. B 277 , 1227-1234. (10.1098/rspb.2009.1958) 16. Watanabe J. 2018 Clade-specific evolutionary diversification along ontogenetic major axes in avian limb skeleton. Evolution 72 , 2632-2652. (10.1111/evo.13627)30328113 17. Losos JB, Sinervo B. 1989 The effects of morphology and perch diameter on sprint performance of Anolis lizards. J. Exp. Biol. 145 , 23. (10.1242/jeb.145.1.23) 18. Losos JB. 1990 The evolution of form and function: morphology and locomotor performance in West Indian Anolis lizards. Evolution 44 , 1189-1203. (10.1111/j.1558-5646.1990.tb05225.x)28563896 19. Irschick DJ, Losos JB. 1998 A comparative analysis of the ecological significance of maximal locomotor performance in Caribbean Anolis lizards. Evolution 52 , 219-226. (10.2307/2410937)28568148 20. Mattingly WB, Jayne BC. 2004 Resource use in arboreal habitats: structure affects locomotion of four ecomorphs of Anolis lizards. Ecology 85 , 1111-1124. (10.1890/03-0293) 21. Rand AS, Williams EE. 1969 The anoles of La Palma: aspects of their ecological relationships. Breviora. 327 , 1-19. 22. Williams EE. 1972 The origin of faunas. Evolution of lizard congeners in a complex island fauna: A trial analysis. In Evol. Biol. (eds T Dobzhansky, M Hecht, W Steere), pp. 47-89. Berlin, Germany: Springer US. 23. Losos JB. 2009 Lizards in an evolutionary tree: ecology and adaptive radiation of anoles. Berkeley, CA: University of California Press. 24. Sanger TJ, Revell LJ, Gibson-Brown JJ, Losos JB. 2012 Repeated modification of early limb morphogenesis programmes underlies the convergence of relative limb length in Anolis lizards. Proc. R. Soc. B 279 , 739-748. (10.1098/rspb.2011.0840) 25. Wakasa H, Cadiz A, Echenique-Diaz LM, Iwasaki WM, Kamiyama N, Nishimura Y, Yokoyama H, Tamura K, Kawata M. 2015 Developmental stages for the divergence of relative limb length between a twig and a trunk-ground Anolis lizard species. J. Exp. Zool. Part B Mol. Dev. Evol. 324 , 410-423. (10.1002/jez.b.22627) 26. Poe S et al . 2017 A phylogenetic, biogeographic, and taxonomic study of all extant species of Anolis (Squamata; Iguanidae). Syst. Biol. 66 , 663-697. (10.1093/sysbio/syx029)28334227 27. Fläschendräger A, Wijffels L. 2009 Anolis, 2nd edn. Münster, Germany: NTV Natur und Tier-Verlag. 28. Sanger TJ, Losos JB, Gibson-Brown JJ. 2008 A developmental staging series for the lizard genus Anolis: a new system for the integration of evolution, development, and ecology. J. Morphol. 269 , 129-137. (10.1002/jmor.10563)17724661 29. Taylor WR, Dyke GV. 1985 Revised procedures for staining and clearing small fishes and other vertebrates for bone and cartilage study. Cybium 9 , 107-119. 30. Schneider CA, Rasband WS, Eliceiri KW. 2012 NIH Image to ImageJ: 25 years of image analysis. Nat. Methods 9 , 671-675. (10.1038/nmeth.2089)22930834 31. Stacklies W, Redestig H, Scholz M, Walther D, Selbig J. 2007 pcaMethods–a bioconductor package providing PCA methods for incomplete data. Bioinformatics 23 , 1164-1167. (10.1093/bioinformatics/btm069)17344241 32. Klingenberg CP. 1996 Multivariate allometry. In Advances in morphometrics (eds LF Marcus, M Corti, A Loy, GJP Naylor), pp. 23-49. New York, NY: Plenum. 33. Klingenberg CP, Zimmermann M. 1992 Static, ontogenetic, and evolutionary allometry: a multivariate comparison in nine species of water striders. Am. Nat. 140 , 601-620. (10.1086/285430) 34. Jolicoeur P. 1963 The multivariate generalization of the allometry equation. Biometrics 19 , 497-499. (10.2307/2527939) 35. Gerber S, Eble GJ, Neige P. 2008 Allometric space and allometric disparity: a developmental perspective in the macroevolutionary analysis of morphological disparity. Evolution 62 , 1450-1457. (10.1111/j.1558-5646.2008.00370.x)18346223 36. Revell LJ. 2012 phytools: an R package for phylogenetic comparative biology (and other things). Methods Ecol. Evol. 3 , 217-223. (10.1111/j.2041-210X.2011.00169.x) 37. R Core Team. 2020 R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. 38. Donihue CM, Herrel A, Fabre AC, Kamath A, Geneva AJ, Schoener TW, Kolbe JJ, Losos JB. 2018 Hurricane-induced selection on the morphology of an island lizard. Nature 560 , 88-91. (10.1038/s41586-018-0352-3)30046104 39. Kolbe JJ, Leal M, Schoener TW, Spiller DA, Losos JB. 2012 Founder effects persist despite adaptive differentiation: a field experiment with lizards. Science 335 , 1086. (10.1126/science.1209566)22300849 40. Losos JB, Schoener TW, Warheit KI, Creer D. 2001 Experimental studies of adaptive differentiation in Bahamian Anolis lizards. Genetica 112–113 , 399-415. (10.1023/A:1013387705408) 41. Losos JB, Schoener TW, Spiller DA. 2004 Predator-induced behaviour shifts and natural selection in field-experimental lizard populations. Nature 432 , 505-508. (10.1038/nature03039)15565155 42. Losos JB, Irschick DJ. 1996 The effect of perch diameter on escape behaviour of Anolis lizards: laboratory predictions and field tests. Anim. Behav. 51 , 593-602. (10.1006/anbe.1996.0063) 43. Pollard AS, Charlton BG, Hutchinson JR, Gustafsson T, McGonnell IM, Timmons JA, Pitsillides AA. 2017 Limb proportions show developmental plasticity in response to embryo movement. Sci. Rep. 7 , 41926. (10.1038/srep41926)28165010 44. Cooper KL, Oh S, Sung Y, Dasari RR, Kirschner MW, Tabin CJ. 2013 Multiple phases of chondrocyte enlargement underlie differences in skeletal proportions. Nature 495 , 375-378. (10.1038/nature11940)23485973 45. Wang J, Zhou J, Bondy CA. 1999 Igf1 promotes longitudinal bone growth by insulin-like actions augmenting chondrocyte hypertrophy. FASEB J. 13 , 1985-1990. (10.1096/fasebj.13.14.1985)10544181 46. Sears KE, Behringer RR, Rasweiler JJ, Niswander LA. 2006 Development of bat flight: morphologic and molecular evolution of bat wing digits. Proc. Natl Acad. Sci. USA 103 , 6581. (10.1073/pnas.0509716103)16618938 47. Andrews RM, Skewes SA. 2017 Developmental origin of limb size variation in lizards. Evol. Dev. 19 , 136-146. (10.1111/ede.12221)28463470 48. Tollis M et al . 2018 Comparative genomics reveals accelerated evolution in conserved pathways during the diversification of anole lizards. Genome Biol. Evol. 10 , 489-506. (10.1093/gbe/evy013)29360978 49. Hall BK. 2005 Bones and cartilage: developmental and evolutionary skeletal biology. Amsterdam; The Netherlands: Elsevier Academic Press. 50. Marchini M, Rolian C. 2018 Artificial selection sheds light on developmental mechanisms of limb elongation. Evolution 72 , 825-837. (10.1111/evo.13447)29436719 51. Abzhanov A, Protas M, Grant BR, Grant PR, Tabin CJ. 2004 Bmp4 and morphological variation of beaks in Darwin's finches. Science 305 , 1462-1465. (10.1126/science.1098095)15353802 52. Lamichhaney S et al . 2015 Evolution of Darwin's finches and their beaks revealed by genome sequencing. Nature 518 , 371-375. (10.1038/nature14181)25686609 53. Campas O, Mallarino R, Herrel A, Abzhanov A, Brenner MP. 2010 Scaling and shear transformations capture beak shape variation in Darwin's finches. Proc. Natl Acad. Sci. USA 107 , 3356-3360. (10.1073/pnas.0911575107)20160106 54. Brakefield PM. 2006 Evo-devo and constraints on selection. Trends Ecol. Evol. 21 , 362-368. (10.1016/j.tree.2006.05.001)16713653 55. Salazar-Ciudad I. 2006 Developmental constraints vs. variational properties: how pattern formation can help to understand evolution and development. J. Exp. Zool. Part B Mol. Dev. Evol. 306 , 107-125. (10.1002/jez.b.21078) 56. Hu Y, Albertson RC. 2017 Baby fish working out: an epigenetic source of adaptive variation in the cichlid jaw. Proc. R. Soc. B 284 , 20171018. (10.1098/rspb.2017.1018) 57. Meyer A. 1987 Phenotypic plasticity and heterochrony in Cichlasoma managuense (Pisces, Cichlidae) and their implications for speciation in cichlid fishes. Evolution 41 , 1357-1369. (10.2307/2409100)28563603 58. Kavanagh KD, Shoval O, Winslow BB, Alon U, Leary BP, Kan A, Tabin CJ. 2013 Developmental bias in the evolution of phalanges. Proc Natl Acad. Sci. USA 110 , 18 190-18 195. (10.1073/pnas.1315213110) 59. Toro E, Herrel A, Vanhooydonck B, Irschick DJ. 2003 A biomechanical analysis of intra- and interspecific scaling of jumping and morphology in Caribbean Anolis lizards. J. Exp. Biol. 206 , 2641-2652. (10.1242/jeb.00473)12819270 60. Mahler DL, Revell LJ, Glor RE, Losos JB. 2010 Ecological opportunity and the rate of morphological evolution in the diversification of Greater Antillean anoles. Evolution 64 , 2731-2745. (10.1111/j.1558-5646.2010.01026.x)20455931 61. Donihue CM et al. 2020 Hurricane effects on Neotropical lizards span geographic and phylogenetic scales. Proc. Natl Acad. Sci. USA 117 , 10 429-10 434. (10.1073/pnas.2000801117) 62. Young NM, Hallgrimsson B. 2005 Serial homology and the evolution of mammalian limb covariation structure. Evolution 59 , 2691-2704. (10.1111/j.0014-3820.2005.tb00980.x)16526515 63. Brattstrom O, Aduse-Poku K, van Bergen E, French V, Brakefield PM. 2020 A release from developmental bias accelerates morphological diversification in butterfly eyespots. Proc. Natl Acad. Sci. USA 117 , 27 474-27 480. (10.1073/pnas.2008253117) 64. Irschick DJ, Vitt LJ, Zani PA, Losos JB. 1997 A comparison of evolutionary radiations in mainland and Caribbean Anolis lizards. Ecology 78 , 2191-2203. (10.1890/0012-9658(1997)078[2191:ACOERI]2.0.CO;2) 65. Velasco JA, Herrel A. 2007 Ecomorphology of Anolis lizards of the Choco′ region in Colombia and comparisons with Greater Antillean ecomorphs. Biol. J. Linn. Soc. 92 , 29-39. (10.1111/j.1095-8312.2007.00885.x) 66. Vanhooydonck B, Irschick D. 2002 Is evolution predictable? Evolutionary relationships of divergence in ecology, performance and morphology in old and new world lizard radiations. In Topics in functional and ecological vertebrate morphology (eds P Aerts, K D'Août, A Herrel, R Van Damme). Maastricht, The Netherlands: Shaker Publishing. 67. Poe S, Anderson CG. 2019 The existence and evolution of morphotypes in Anolis lizards: coexistence patterns, not adaptive radiations, distinguish mainland and island faunas. PeerJ 6 , e6040. (10.7717/peerj.6040)30627481 68. Moore TY, Organ CL, Edwards SV, Biewener AA, Tabin CJ, Jenkins Jr FA, Cooper KL. 2015 Multiple phylogenetically distinct events shaped the evolution of limb skeletal morphologies associated with bipedalism in the jerboas. Curr. Biol. 25 , 2785-2794. (10.1016/j.cub.2015.09.037)26455300 69. Young NM, Wagner GP, Hallgrimsson B. 2010 Development and the evolvability of human limbs. Proc. Natl Acad. Sci. USA 107 , 3400-3405. (10.1073/pnas.0911856107)20133636 70. Martin-Serra A, Benson RBJ. 2020 Developmental constraints do not influence long-term phenotypic evolution of marsupial forelimbs as revealed by interspecific disparity and integration patterns. Am. Nat. 195 , 547-560. (10.1086/707194)32097034 71. Kelly EM, Sears KE. 2011 Reduced phenotypic covariation in marsupial limbs and the implications for mammalian evolution. Biol. J. Linn. Soc. 102 , 22-36. (10.1111/j.1095-8312.2010.01561.x) 72. Watanabe A, Fabre A-C, Felice RN, Maisano JA, Müller J, Herrel A, Goswami A. 2019 Ecomorphological diversification in squamates from conserved pattern of cranial integration. Proc. Natl Acad. Sci. USA 116 , 14688. (10.1073/pnas.1820967116)31262818 73. Ebel R, Müller J, Ramm T, Hipsley C, Amson E. 2020 First evidence of convergent lifestyle signal in reptile skull roof microanatomy. BMC Biol. 18 , 185. (10.1186/s12915-020-00908-y)33250048 74. Greer AE. 1991 Limb reduction in squamates: identification of the lineages and discussion of the trends. J. Herpetol. 25 , 166-173. (10.2307/1564644) 75. Feiner N, Jackson ISC, Van der Cruyssen E, Uller T. 2021 Data from: A highly conserved ontogenetic limb allometry and its evolutionary significance in the adaptive radiation of Anolis lizards. Dryad Digital Repository. (10.5061/dryad.jsxksn08d)
PMC008xxxxxx/PMC8220278.txt
==== Front Proc Biol Sci Proc Biol Sci RSPB royprsb Proceedings of the Royal Society B: Biological Sciences 0962-8452 1471-2954 The Royal Society 34157872 10.1098/rspb.2021.0328 rspb20210328 10011292369Global Change and Conservation Research Articles Combined responses of primary coral polyps and their algal endosymbionts to decreasing seawater pH Combined responses of primary coral polyps and their algal endosymbionts to decreasing seawater pH http://orcid.org/0000-0001-7785-4758 Scucchia Federica Conceptualization Data curation Formal analysis Investigation Methodology Project administration Writing-original draft Writing-review & editing 1 2 Malik Assaf Data curation Methodology Software Visualization 1 Zaslansky Paul Data curation Methodology Resources Software Visualization 3 http://orcid.org/0000-0003-2322-3269 Putnam Hollie M. Conceptualization Data curation Funding acquisition Investigation Methodology Project administration Supervision Validation Writing-original draft Writing-review & editing 4 http://orcid.org/0000-0002-7298-290X Mass Tali Conceptualization Data curation Funding acquisition Investigation Project administration Resources Supervision Validation Writing-original draft Writing-review & editing [email protected] 1 5 1 Department of Marine Biology, The Leon H. Charney School of Marine Sciences, University of Haifa, Haifa 3498838, Israel 2 The Interuniversity Institute of Marine Sciences, Eilat 88103, Israel 3 Department for Operative and Preventive Dentistry, Charité—Center for Dental and Craniofacial Sciences, Universitätsmedizin Berlin, Berlin 14197, Germany 4 Department of Biological Sciences, University of Rhode Island, Kingston, RI 02881, USA 5 Morris Kahn Marine Research Station, The Leon H. Charney School of Marine Sciences, University of Haifa, Sdot Yam, Israel Electronic supplementary material is available online at https://doi.org/10.6084/m9.figshare.c.5463865. 30 6 2021 June 30, 2021 23 6 2021 June 23, 2021 23 6 2021 June 23, 2021 288 1953 202103288 2 2021 Feburary 8, 2021 2 6 2021 June 2, 2021 © 2021 The Authors. 2021 https://creativecommons.org/licenses/by/4.0/ Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited. With coral reefs declining globally, resilience of these ecosystems hinges on successful coral recruitment. However, knowledge of the acclimatory and/or adaptive potential in response to environmental challenges such as ocean acidification (OA) in earliest life stages is limited. Our combination of physiological measurements, microscopy, computed tomography techniques and gene expression analysis allowed us to thoroughly elucidate the mechanisms underlying the response of early-life stages of corals, together with their algal partners, to the projected decline in oceanic pH. We observed extensive physiological, morphological and transcriptional changes in surviving recruits, and the transition to a less-skeleton/more-tissue phenotype. We found that decreased pH conditions stimulate photosynthesis and endosymbiont growth, and gene expression potentially linked to photosynthates translocation. Our unique holistic study discloses the previously unseen intricate net of interacting mechanisms that regulate the performance of these organisms in response to OA. ocean acidification , scleractinian corals , Stylophora pistillata , primary polyps , Symbiodinium Israel Science Foundation http://dx.doi.org/10.13039/501100003977 2155/15 2288/16 312/15 Binational Science Foundation BSF 2016321 H2020 European Research Council http://dx.doi.org/10.13039/100010663 755876 cover-dateJune 30, 2021 ==== Body pmc1. Introduction Stony corals secrete calcium carbonate skeletons which create three-dimensional reef frameworks that support the most productive and biologically diverse marine ecosystems on Earth. Corals have a biphasic life cycle with swimming planktonic larvae and sessile adults [1]. Successful settlement of larvae and their subsequent survival (known as ‘recruitment’) is a key element for coral reef resilience and recovery following major disturbances [2]. Young coral settlers are highly vulnerable to mortality resulting from environmental changes, such as rising seawater temperature and ocean acidification (OA) associated with global climate change [3]. OA is expected to negatively affect mineral formation of marine calcifiers, such as corals, by making it more energetically difficult for these organisms to deposit their skeletons [4]. This circumstance has led to numerous studies on the effects of OA on coral calcification, that is predicted to substantially decline over the course of this century [5]. However, great uncertainty still remains about the extent of the threat of OA to coral persistence. This is partly due to the high variability of the coral response to OA and to the persistence of coral communities that thrive in naturally low pH environments [6,7], which has shed light on the potential for coral acclimation and adaptation to the projected acidification scenarios. Most of the studies conducted on primary polyps focused primarily on changes occurring in the skeletal development, which has been shown to be significantly delayed, with the potential for substantial deformities with decreases in seawater pH [8]. Studies on the functioning of the coral–algal symbiosis under OA are scarce, especially related to coral early-life stages. Coral symbiotic algae, which belong to the family Symbiodiniaceae, have a major role in the coral holobiont physiology and nutrition [9]. These dinoflagellates supply the coral host with photosynthetic products, thereby supporting the host metabolism, growth and reproduction [9]. A comprehensive understanding of how the earliest stages of coral, together with their algal partners, will respond to OA is critically needed to assess the capacity of these organisms to adapt and/or acclimatize to a future acidified ocean. To this end, we performed a controlled laboratory study by culturing larvae and primary polyps (metamorphosed and settled larvae) of the coral Stylophora pistillata under reduced pH values (high CO2 concentrations) predicted by the end of the century, specifically at pH 8.2, pH 7.8 and pH 7.6 [10]. Here, we investigate larval ability to survive and successfully recruit under OA conditions, and we examine key physiological processes and skeletal characteristics influenced by changes in seawater pH, at the level of both phenotype and gene expression. Our findings reveal that the response of young corals to OA involves an intricate network of interrelated changes in the coral holobiont. 2. Material and methods Coral larvae were collected from 20 randomly selected adult colonies of the stony coral S. pistillata on the reef adjacent to the Interuniversity Institute of Marine Sciences (IUI, 29°30′06.0″ N 34°54′58.3″ E) in the Gulf of Eilat (Israel), under a special permit from the Israeli Natural Parks Authority. All larvae were pooled together (approx. 50 larvae from each parental colony) and were transported to a controlled environment aquarium system at the Leon H. Charney School of Marine Science at the University of Haifa. The carbonate chemistry of seawater was manipulated in the experimental aquariums by injecting CO2 to reduce the ambient pH (pH 8.2) and obtain the target values of pH 7.8 and pH 7.6. After 9 days of experimental treatment exposure, the numbers of settled primary polyps, still swimming larvae (not settled) and dead larvae (larvae dissolved during the experiment) were recorded. The primary polyps were gently removed from the chambers and they were collected for subsequent physiological, skeletal and molecular analysis (for full details, see electronic supplementary material, Methods). 3. Results (a) Larval survival and settlement Following the 9 days experimental exposure to decreased pH conditions, we observed a reduction in the percentage of settled larvae from 82% at the control pH 8.2, to 64% at pH 7.8, and to 46% at pH 7.6, which resulted to be significantly lower than the control (Fisher's exact test, p < 0.0001) (electronic supplementary material, figure S2). At the intermediate pH, we detected a significantly higher percentage of larvae that were still at the planktonic or swimming stage relative to pH 8.2 (Fisher's exact test, p < 0.0001), while at the lowest pH, we found significantly higher larval mortality compared to the control (Fisher's exact test, p < 0.0001). (b) Physiological parameters of primary polyps and algal endosymbionts We then explored the physiological changes that allowed the surviving individuals to successfully settle and develop into a primary polyp. Measurements of the primary polyps dark-respiration rates, indicative of the coral metabolic rates, were significantly reduced at the lowest pH condition, with an average decrease of 64 and 69% at pH 7.6 relative to pH 8.2 and pH 7.8, respectively (one-way ANOVA, N = 6, F2,15 = 5.927, p = 0.01) (electronic supplementary material, figure S3A). In contrast with the reduction in respiration rate, coral host protein concentration, used as a proxy for tissue biomass, significantly increased in both pH 7.8 and pH 7.6 compared to the control pH 8.2 (electronic supplementary material, figure S3B) (one-way ANOVA, N = 3, F2,24 = 24.99, p < 0.0001). We then explored changes in the algal physiological attributes, and we found a significantly higher number of endosymbiotic algae per polyp surface area at both pH 7.8 and pH 7.6 relative to pH 8.2 (electronic supplementary material, figure S3C) (one-way ANOVA, N = 3, F2,24 = 99.25, p < 0.05), pointing to an enhancement of algal growth with decreasing pH. The number of symbiont cells per host protein did not show instead any significant change (electronic supplementary material, figure S3D). Furthermore, chlorophyll a concentration was higher in the polyps at the lowest pH (one-way ANOVA, N = 3, F2,24 = 73.41, p < 0.0001) (electronic supplementary material, figure S3E), even though its concentration within a single algal cell did not significantly vary relative to the control pH (electronic supplementary material, figure S3F), providing further evidence of the augmented algal density at pH 7.6. In addition, higher chlorophyll auto-fluorescence was detected in the live primary polyps at the lowest pH (one-way ANOVA, F2,6 = 73.38, p > 0.0001) (electronic supplementary material, figure S4A,B), matching the chlorophyll a concentration results (electronic supplementary material, figure S3E). (c) Endosymbiont photosynthetic traits The observed changes in the algal physiological properties are indicative of variations in photosynthetic traits. Indeed, we observed an increase of photochemical efficiency and activity in the endosymbiotic algae at pH 7.6. The maximal quantum yield (Fv/Fm) of photosystem II (PSII) was significantly higher in the primary polyps at pH 7.6 relative to the other two pH (electronic supplementary material, figure S4C) (Kruskal–Wallis test, H = 7.199, p = 0.027). In addition, although the measurements of the initial slope (α) and of the minimal photoinhibition point (Ek) showed no significant change (electronic supplementary material, table S5), the maximum values of the relative electron transport rate (rETR), were significantly elevated in the polyps at pH 7.6 relative to the control (electronic supplementary material, figure S4D) (Kruskal–Wallis test, H = 8.025, p = 0.018). Taken together, these results show that the observed increase of algal density under acidification conditions is paralleled by the enhancement of photosynthetic efficiency and activity. We also measured the maximum values of the non-photochemical quenching (NPQ) and found no significant change among pH treatments (electronic supplementary material, table S5). (d) Macro-scale skeletal growth patterns of primary polyps The existence of a linkage between tissue biomass and growth [11], and our observation of increased host tissue at reduced pH, prompted us to explore changes in the primary polyps skeletal characteristics. By employing the fluorescent dye calcein [12], we visualized patterns of incorporation of divalent ions (such as Ca2+) into the live primary polyps and subsequently into the skeleton. The in vivo imaging shows that in all three pH conditions, the highest calcein fluorescence intensity is found in the calyx area (see electronic supplementary material, figure S6 for additional details) and in the skeletal septa (electronic supplementary material, figure S5A). In corals, ions and other particles are ingested from seawater into the mouth cavity and are then transported to the site of calcification [13]. Our observations indicate that in all three pH conditions, corals are actively up-taking Ca2+ from the mouth, and that these ions are eventually incorporated into the skeleton. Measurements of the calcein fluorescence intensity of the primary polyps skeleton (N = 3 polyps per pH condition, only 1 polyp per pH is shown in electronic supplementary material, figure S5A) show a significantly lower calcein fluorescence at pH 7.8 and 7.6 compared to pH 8.2 (electronic supplementary material, figure S5B) (one-way ANOVA, F2,6 = 213.5, p < 0.0001). This indicates that a lower amount of Ca2+ was incorporated into the skeleton, thus pointing to a lower skeletal development at reduced pH conditions. Patterns of calcein incorporation correspond to the skeletal morphology imaged using micro-CT. Slices of the 3D reconstructions of the polyps skeleton (figure 1d–f) show a progressive reduction in the thickness of the skeletal septa as the pH decreases (figure 1g–l). Such reduction, that was significant at both pH 7.8 and 7.6 compared to the control (one-way ANOVA, F2,15 = 24.32, p > 0.001) (electronic supplementary material, figure S5C), indicates that at reduced pH, a lower amount of CaCO3 was deposited by the coral. Figure 1. Changes in skeletal growth patterns in coral primary polyps with decreasing pH. (a–c) Microscopy images showing the Calcein Blue fluorescence (expressed as percentage intensity) in the skeleton of the primary polyps at (a) pH 8.2, (b) pH 7.8 and (c) pH 7.6. Magnification: 4×. Scale bar: 400 µm. (d–f) Micro-CT images showing top views of three-dimensional reconstructions of the primary polyps' skeleton at (d) pH 8.2, (e) pH 7.8 and (f) pH 7.6. Scale bar: 400 µm. (g–i) Slices of the three-dimensional reconstructions showing thresholded two-dimensional slices of the base of the septa at (g) pH 8.2, (h) pH 7.8 and (i) pH 7.6. Scale bar: 200 µm. (j–l) Skeletal thickness distribution (μm) along the vertical axis of the septa of the primary polyps at (j) pH 8.2, (k) pH 7.8 and (l) pH 7.6. Scale bar: 200 µm. (Online version in colour.) (e) Skeletal micro-morphological features Further changes in skeletal growth patterns were additionally detected at a micro-scale. Measurements of the calyx area (figure 2a–c, cyan overlays) of the polyps imaged by scanning electron microscopy indicate that there was no significant difference between calyxes among the three pH conditions (figure 2m). Differently from the calyxes, the crown areas shown in figure 2a–c (yellow overlays), corresponding to the forming coenosteum (see electronic supplementary material, figure S6), were significantly smaller in the primary polyps at pH 7.8 and 7.6 relative to pH 8.2 (one-way ANOVA, N = 6, F2,15 = 8.17, p < 0.01). At higher magnification, the thickness of the septa (electronic supplementary material, figure S6) was significantly smaller in polyps grown at lower pH compared to the control (figure 2d–f,n) (one-way ANOVA, N = 27, F2,78 = 51.92, p < 0.0001). Overall, these observations show that the reduction of skeletal development detected at the macro-scale is also manifested at the microstructural level. Figure 2. Modifications of skeletal features of coral primary polyps at different pH conditions. (a–c) SEM images showing the micro-morphology of the calyx (cyan areas) and crown (orange areas) of the primary polyps at (a) pH 8.2, (b) pH 7.8 and (c) pH 7.6. Enlargements of the skeletal (d–f) septa and (g–i) spines at (d,g) pH 8.2, (e,h) pH 7.8 and (f,i) pH 7.6, respectively. (g) White arrows indicate the RADs on the spine. (j–l) Insets showing the RADs surface texture at (j) pH 8.2, (k) pH 7.8 and (l) pH 7.6. (m) Size of the calyx area (cyan bar) and the crown area (orange bar) calculated in the primary polyps (N = 6 per pH condition). (n) Thickness of the septa calculated in the primary polyps at (d) pH 8.2, (e) pH 7.8 and (f) pH 7.6 (N = 27 per pH condition). (o) RADs number measured per units of basal area of the spines (mm) at (g) pH 8.2, (h) pH 7.8 and (i) pH 7.6 (N = 9 per pH condition). (m–o) Asterisks (*) indicate statistical differences (p < 0.05, one-way ANOVA) relative to the control pH 8.2. Scale bars: (a–c) 200 µm, (d–f) 100 µm, (g–i) 10 µm and (j–l) 1 µm. Moreover, the terminal portion of the skeletal spines located on the septa showed a significantly lower number of rapid accretion deposits (RADs; globular elements in figure 2g–i; see electronic supplementary material, figure S6) in the polyps at reduced pH, when compared with the control (figure 2o) (one-way ANOVA, N = 9, F2,24 = 19.72, p < 0.0001). RADs correspond to areas of the skeleton with rapid CaCO3 deposition [14]. Hence, the lower abundance of RADs in the polyps at pH 7.8 and 7.6 indicates a reduced development of these regions in acidified seawater. In addition, we found prominent differences in the RADs texture. At the control pH, the texture of the RADs was smoother and more compact in comparison to the acidified conditions (figure 2j–l). At pH 7.6, the aragonite bundles of fibres, corresponding to the microcrystalline features on the RADs surface (figure 2j–l), were less compact and were characterized by granular aggregates forming needle-like structures with differing shapes and orientations (figure 2l) compared to the texture at both pH 8.2 and 7.8. This microstructural pattern conferred a more porous appearance to the RADs surface. The exposure to acidified seawater conditions, therefore, generates changes at multiple levels of the skeletal hierarchy, from macro- to micro-scale. (f) Coral host transcriptomic response to decreased pH We examined the molecular controls underlying the observed modifications in the primary polyps physiological and skeletal characteristics. Analysis of the coral host's differentially expressed genes (DEGs) show that, compared to pH 8.2, a higher number of DEGs was found at pH 7.6 as opposed to pH 7.8 (figure 3a). This pattern, which reflects the logarithmic nature of the pH scale, indicates that the slope of the progression in DEGs with decreasing pH sharply steepens between pH 7.8 and 7.6, where larger transcriptional modifications are needed to adjust to acidified seawater. Figure 3. Coral DEGs, functional enrichment and biomineralization-related genes expression between experimental pH conditions. (a) Number of DEGs in coral primary polyps detected, respectively, in the pH 7.8 treatment relative to the control pH 8.2 (left bar), in the pH 7.6 treatment relative to pH 8.2 (middle bar) and in the pH 7.6 treatment relative to pH 7.8 (right bar). Red indicates the upregulated genes and blue indicates the downregulated genes within each comparison. (b) Dendrogram on the left showing enriched GO terms and KEGG pathways clustered according to the portion of identical genes shared. The heat map shows the positive (upregulated genes) or negative (downregulated genes) percentages of DEGs per GO term or KEGG pathway, detected, respectively, at pH 7.6 relative to pH 8.2 (right column) and to pH 7.8 (left column). Red numbers indicate the total number of genes within each per GO term or KEGG pathway. DE (%), percentage of DEGs. (c) Heat map showing the log2 fold change of significantly (adjusted p-value > 0.05) upregulated (red) and downregulated (blue) biomineralization-related genes at pH 7.8 and 7.6 when compared with pH 8.2. (Online version in colour.) Gene Ontology (GO) enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses were performed to identify enriched genes groups and molecular pathways involved in coral response to pH reduction. Figure 3b shows that among all pH treatments comparisons, the largest number of enriched genes groups (both up- and downregulated) was found between pH 7.6 and pH 8.2. The upregulated gene groups at the lowest pH show a significant enrichment for genes associated with the activity of voltage-gated calcium channels, with cell differentiation and developmental processes, and with genes linked to sensory perception and detection of environmental stimuli (figure 3b). The downregulated genes groups at pH 7.6 include a large number of genes involved in lipid metabolism, mitochondria biosynthesis and activity, in energy production and respiration. The downregulation of these genes points to a reduced energy metabolism at acidified conditions, disclosing the molecular control behind the observed respiration rate decrease (electronic supplementary material, figure S3A). (g) Differential expression patterns of putative biomineralization-related genes Looking into the molecular mechanisms underlying the observed changes in skeletal features, we evaluated the differential expression of putative coral biomineralization-related genes (electronic supplementary material, table S4). Among these 91 total genes, 5 were differentially expressed in the polyps at pH 7.8, and 23 were differentially expressed in the polyps at pH 7.6 (figure 3c), indicating a greater response of the biomineralization molecular machinery at the extreme pH. In particular at pH 7.6 relative to pH 8.2, a significantly higher number of biomineralization-related genes was upregulated, compared to the number of downregulated genes (Fisher's exact test, p < 0.05). Most of these upregulated genes code for proteins of the skeletal organic matrix with structural and adhesion functions. These proteins are embedded in the skeletal framework and control the growth of the aragonite crystal-like fibres within the coral skeleton [15]. Notably, we observed the downregulation of the coral acid-rich protein 4 (CARP4) gene at the lowest pH treatment (figure 3c). As member of the CARPs group, this biomineralization toolkit protein is expected to transport Ca2+ to the calcification site that are then concentrated in centres of calcification, also known as RADs [15]. Moreover, among the upregulated genes, we detected the S. pistillata carbonic anhydrase 2 (STPCA-2) and the L-type calcium channel (figure 3c), which is located in the calicoblastic epithelium (cell layer overlying the coral skeleton) [16]. These two genes are involved in regulating the pH and carbonate chemistry of the calcifying medium, the confined space where skeletal deposition occurs [17,18]. In particular, the increased expression of the L-type calcium channel, together with the enrichment of genes associated with calcium channels activity (figure 3b), indicates that the exposure to acidic seawater stimulates in the corals the transport of Ca2+. (h) Symbiodinium microadriaticum transcriptomic response to decreased pH In S. pistillata algal endosymbiont, enriched genes groups (both up- and downregulated) were only found in pH 7.6 relative to the other two pH conditions, while no enrichment was detected in pH 7.8 relative to the control (figure 4). This indicates that, similar to the coral host, wider transcriptional changes take place at the lowest pH. Figure 4. Changes in expression levels of different functional groups of genes in the coral endosymbiotic algae. Dendrogram showing enriched functional groups clustered according to the portion of identical genes shared. The heat map shows the positive (upregulated genes) or negative (downregulated genes) normalized enrichment score (obtained using the functional enrichment analysis GSEA) per each functional group, detected, respectively, at pH 7.6 compared to pH 8.2 (right bar) and to pH 7.8 (left bar). No enrichment was found in pH 7.8 compared to pH 8.2. Genes groups are shown only for cases with significant (adjusted p-value < 0.05) enrichment in both comparisons. Red numbers indicate the total number of genes within each functional group. NES, normalized enrichment score. (Online version in colour.) The downregulated groups at pH 7.6 include genes related to photosynthesis (e.g. ‘thylakoid’), gene expression (e.g. ‘ribosome’) and proton transport (e.g. ‘proton transmembrane transport’). The upregulated groups comprise a large number of genes involved in photosynthetic carbon fixation (‘ribulose-bisphosphate carboxylate activity’, also known as RuBisCO, ‘carboxy-lyase activity’, etc.), carbohydrates binding and transport (e.g. ‘polysaccharide binding’, ‘substrate-specific channel activity’, ‘active transmembrane transporter activity’). As shown by the clustering (figure 4), terms related to carbohydrates binding and transport share a portion of common genes, suggesting that symbiont genes associated with transmembrane transport are also involved in the transport of sugars. Taken together, our observations indicate that the symbiotic algae response to reduced pH is governed by the modulation of multiple photosynthesis-related genes. 4. Discussion The current study on S. pistillata early-life stages reveals that the exposure to predicted future OA conditions leads to a reduction of larval survival and settlement (electronic supplementary material, figure S2). Nonetheless, the more acidic seawater did not entirely inhibit larval recruitment, suggesting potential acclimatory mechanisms that allowed the corals to successfully settle and start growing. Larval recognition of the settlement substrates occurs through environmental cues, that are exogenous factors capable of inducing larval settlement [19]. Seawater acidification was shown to reduce the induction of larval settlement in response to environmental cues [20]. Based on our transcriptomic analysis, genes shown to be related to larval selection of and attachment to the substrate (i.e. sensory perception and detection of environmental stimuli-related genes) [21,22] are significantly enriched at the lowest pH (figure 3b). This indicates that the modulation of genes with a role in sensory perception may account for the adjustment capacity and survival of the corals under OA conditions, as observed in other organisms exposed to a changing environment [23]. Diminished availability of settling juveniles could inhibit the replenishment of reefs after sporadic disturbances such as storms and bleaching events [2], with the potential to compromise coral reef resilience. However, coral populations can naturally persist in acidic environments, despite having a lower recruitment efficiency in comparison to corals living at ambient pH [24]. Moreover, improved adult performance could compensate for low recruitment rates, as observed for corals living around CO2 seeps [25]. In this study, larvae reared under acidified conditions that successfully metamorphosed into a primary polyp were characterized by reduced metabolic rates, as shown by the lower respiration rate (electronic supplementary material, figure S3A) and the downregulation of metabolism-related genes (figure 3b). A downward trend of energy metabolism seems a common response in young corals exposed to acidic seawater [26]. In an OA scenario, the formation of the skeleton is thought to be more energetically expensive for corals, requiring more energy to remove from the calcifying space the excess H+ produced during calcification [4]. The reduction of metabolic rates observed in the primary polyps may energetically limit the rates of H+ removal. This leads to a diminished ability to deposit CaCO3, and indeed our multifaceted examination of the skeleton reveals that under OA primary polyps were characterized by reduced skeletal development (figures 1a–c,j–l and 2a–i,m–o; electronic supplementary material, figure S5B,C). Furthermore, our observations of the reduction in primary polyp crown areas (figure 2a–c,m) and in the number of RADs (figure 2g–i,o) show that exposure to OA conditions involves changes within the skeleton that have never been investigated so far. In this context, the downregulation of the CARP4 gene detected at acidified conditions (figure 3c) reveals a potentially lower contribution of this protein to the formation of RADs, thereby contributing to the overall reduction in skeletal growth. In fact, this biomineralization toolkit protein was localized in S. pistillata skeleton in correspondence of RADs, and was suggested to form crystal binding substrates that lead to CaCO3 nucleation [15]. Given the reduction of skeletal development in acidified seawater, corals must put in place alternative strategies to sustain overall growth. Indeed, the greater amount of tissue biomass covering the skeleton, detected in this study (electronic supplementary material, figure S3B), could allow coral polyps to reach larger sizes, representing a critical strategy to counterbalance the lower skeletal development. Through a lower investment of energy into skeletal growth, polyps can instead allocate more of the available energy pool to tissue biomass [27,28]. For example, in the sea urchin Strongylocentrotus purpuratus, another marine calcifier, protein synthesis accounted for approximately 84% of the available energy pool consumption under exposure to seawater acidification [29]. With this energy-reallocation mechanism, corals could increase their tissue biomass despite having decreased metabolic rates. Similar to our study, a higher-biomass phenotype was observed under a long-term exposure to OA in the temperate coral Oculina patagonica [30]. When transferred back to ambient pH conditions, these corals returned to calcify normally despite being 12 months as soft-bodied polyps in low pH conditions [30], showing that corals possess a remarkable ability to adjust and survive to future acidification conditions. The potential ecological consequences of maintaining overall growth while reducing skeletal development are associated with reaching critical sizes for sexual maturity and assuring a high production of offspring, thereby perpetuating the species and sustaining the reef persistence [7]. The shift to a less-skeleton/more-tissue phenotype would not negatively impact the achievement of sexual maturity and gametogenesis, which appeared to develop similarly in both reduced-skeleton corals at acidified conditions and corals at ambient pH [30]. More porous skeletons with preserved rates of linear extension were observed in S. pistillata adults exposed to acidic conditions [31]. In addition in naturally acidic pH environments, adult corals build more porous skeletons to keep linear extension rate constant [7]. These observations indicate that, for corals, the achievement of larger sizes is indeed a critical need, and that the shift to a more porous skeleton is a common trade-off and an acclimatory strategy to cope with OA. In acidic seawater, our results show a more porous appearance of the primary polyps' skeleton at micro-scale (figure 2l). For corals, the accretion of a more porous skeleton under OA conditions is sustained by embedding more organic matrix proteins within the skeletal fibres [31]. Indeed, our molecular examination shows an enhanced expression of organic matrix proteins at acidic conditions (figure 3c), which points to an increased incorporation of these biomineralization toolkit proteins within the skeletal pores. Previous observations report changes in the expression of organic matrix genes [26] and in the skeletal fibres arrangement [32] of early coral stages reared under OA conditions. Nevertheless, these studies focused either on the molecular or on the morphological aspects, without integrating the coral response across multiple variables. It is noteworthy that energy investment into organic matrix synthesis is suggested to be three orders of magnitude less than the energy cost of active pumps regulating CaCO3 deposition [27]. This implies that increasing the organic matrix production represents a more cost-effective way to support overall polyp growth, when less favourable conditions for skeleton building occur. Despite the change in seawater carbonate chemistry, primary polyps still managed to favour CaCO3 accretion. With our work, we show that coral recruits finely tuned the expression of calcification-related genes. In particular, our molecular analysis implies a tight control of the calcification site chemistry exerted by the coral, through the modulation of Ca2+ transport (figure 3b,c) and of the activity of STPCA-2 (figure 3c). This latter enzyme catalyses the hydration of CO2 to HCO3¯ that is subsequently delivered to the calcification site [17]. The concentrations of both Ca2+ and HCO3¯ have a major role in the skeletal development under OA, as indicated in adults of S. pistillata and other coral species [33,34]. Furthermore, a recent study shows that genes related to Ca2+ management and to inorganic carbon regulation (i.e. carbonic anhydrases) have a major role in the persistence of coral populations adapted to naturally low pH environments [35]. In the light of the predicted decrease of oceanic pH, the ability of corals to modulate the Ca2+ transport and the STPCA-2 activity, especially during the delicate early-life stages, constitutes a crucial area of further research, that could disclose the susceptibility of different coral species to OA. In symbiotic corals, metabolic requirements for growth are critically supported by the photosynthetic activity of the symbiotic algae [9]. Through the carbon fixation process, these dinoflagellates incorporate CO2 into organic compounds, that are used as energy sources [9]. Growing evidence shows that enhanced CO2 levels in seawater could improve the symbiont photosynthetic performance, by alleviating the carbon limitation that is experienced by symbiotic algae inside the host tissues [36,37]. The predicted increase of [CO2] in seawater [10], simulated in our study, appears to boost the activity RuBisCO (the key photosynthetic enzyme that catalyses the first major step of carbon fixation), as suggested by the increased expression of carbon fixation genes in acidic seawater (figure 4). As a result, the algal photosynthetic activity and efficiency increase (electronic supplementary material, figure S4), ultimately enhancing algal growth (electronic supplementary material, figure S3C,E). It is noteworthy that endosymbiotic algae transfer a portion of the photosynthetically derived products, that are not ultimately used for algal growth, to the host [9]. The increased expression of genes linked to carbohydrate binding, that cluster with genes related to the activity of transmembrane transporters and channels (figure 4), suggests an enhanced transfer of photosynthates to the coral host. In corals, heterotrophic feeding requires additional energy to break down food particles and absorb nutrients, compared to sugars coming from the endosymbionts that are quickly metabolized [38]. Given the reduction of metabolic rates under OA, an increased transfer of sugars from the algal endosymbionts might constitute for corals a more cost-effective energy source, compared to heterotrophic sources. Notably, the energy derived from photosynthesis is hypothesized to have a 10–20-fold greater effect on tissue energetics compared to skeletal energetics [27]. Therefore, the augmented algal photosynthetic activity could substantially contribute to the tissue thickening response observed in the primary polyps. In addition, photosynthetic products have been suggested to be used as precursors for skeletal organic matrix biosynthesis [39]. An increased translocation of photosynthates from S. microadriaticum would thus considerably support the enhanced synthesis of organic matrix in the corals, and it would overall increase the host ability to cope with the effects of OA [40]. Earlier works on adult corals investigating the response to OA of the algal endosymbiont focused either on transcriptomic [41] or physiological changes [42,43], and reported negative or no physiological responses in the algal symbionts with the exposure to acidic seawater. However, it must be noted that broad physiological differences exist among different species across Symbiodiniaceae [44], and that gene expression changes of the endosymbiotic algae greatly vary among coral populations in response to acidification [45]. The possibility of symbiont-mediated changes in the coral response to seawater acidification, especially at the highly critical early-life stages, urges, therefore, more in-depth future examinations. In conclusion, we thoroughly describe the mechanisms underlying the response of S. pistillata early-life stages under OA. Our findings suggest that acclimatory mechanisms played a role in the observed coral response. However, we cannot definitively discern relative contributions of acclimation versus natural selection acting on corals genotypes, due to the differential survivorship among treatments and to genetic heterogeneity among coral larvae in the field. In any case, it is important to consider that coral recruits spawned from adults that are already experiencing the effects of decreasing pH may inherit from the parent colonies a greater tolerance to acidifying oceans [46]. These surviving individuals may be better able to cope with OA than prior generations, ultimately building coral reefs resilience through adaptive evolution (reviewed in [47]). As summarized in electronic supplementary material, figure S9, we show that in coral primary polyps, the increased energetic demands of the calcification process under OA resulted in reduced skeletal development and higher skeletal porosity. To counterbalance these changes and support overall growth, corals enhanced the incorporation of organic matric proteins into the skeletal framework and increased the production of tissue biomass. With more available CO2 dissolved in seawater, the growth and photosynthetic activity of the algal endosymbiont was greatly stimulated, leading to a potential enhancement of sugar translocation to the host. While extrapolating laboratory-based findings to projections of the effects of OA on corals in the field is possible, we caution that controlled aquarium systems often do not simulate the dynamic environment of the reef. Moreover, other environmental factors, such as elevated seawater temperature, might act synergistically with acidification, affecting coral fitness [48]. Although extensive, our study has analysed a single coral species and at only one time point, thus limiting the power of our predictions on the acclimatory nature of the observed coral response, and on the phenotypic changes that could occur at later coral life stages under long-term acidification exposure. Thus, future efforts should focus on assessing the degree of modifications in subsequent developmental stages of various coral species under acidification conditions, allowing us to make more accurate predictions on the vulnerability of corals to future OA scenarios. Our multidisciplinary approach reveals strong merit in investigating the response of both the host and the algal partner, as we show that the response to OA involves a wide and intricate net of interrelated parameters in both organisms. The overall correspondence of processes across different assays and biological scales showcases the robust nature of our work, and the importance of including interdisciplinary and complementary analyses towards understanding corals’ vulnerability to environmental change. Supplementary Material Click here for additional data file. Acknowledgements We would like to thank Maayan Neder for the help with collecting coral larvae from the wild. We also thank Jeana Drake for the valuable and critical comments on the manuscript. Lastly, we thank the Interuniversity Institute of Marine Sciences in Eilat for access to its infrastructure and services. Data accessibility All data needed to evaluate the conclusions in the paper are present in the paper and/or the electronic supplementary materials and/or are available from the Dryad Digital Repository, including the scripts employed to analyse the RNA-seq data and all the raw data of the study (i.e. raw data of physiological, morphological and photosynthetic measurements, and of the aquariums seawater chemistry: https://doi.org/10.5061/dryad.66t1g1k27 [49]. The RNA-Seq raw data were deposited in the National Center for Biotechnology Information (www.ncbi.nlm.nih.gov/bioproject/PRJNA640748). Authors' contributions F.S.: conceptualization, data curation, formal analysis, investigation, methodology, project administration, writing—original draft, writing—review and editing; A.M.: data curation, methodology, software, visualization; P.Z.: data curation, methodology, resources, software, visualization; H.M.P.: conceptualization, data curation, funding acquisition, investigation, methodology, project administration, supervision, validation, writing—original draft, writing—review and editing; T.M.: conceptualization, data curation, funding acquisition, investigation, project administration, resources, supervision, validation, writing—original draft, writing—review and editing. All authors gave final approval for publication and agreed to be held accountable for the work performed therein. Competing interests The authors declare no competing interests. Funding This project has received funding from the Israeli Binational Science Foundation (BSF 2016321 to H.M.P. and T.M.), the Israel Science Foundation (312/15 to T.M.) and from the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation programme (grant agreement no. 755876 to T.M.). The experiment was performed in a controlled aquarium system which was funded by Institutional ISF grants 2288/16. Computations presented in this work were performed on the Hive computer cluster at the University of Haifa, which is partly funded by ISF grant 2155/15 to T.M. ==== Refs References 1. Akiva A et al. 2018 Minerals in the pre-settled coral Stylophora pistillata crystallize via protein and ion changes. Nat. Commun. 9 , 1-9. (10.1038/s41467-017-02088-w)29317637 2. Adjeroud M, Kayal M, Penin L. 2017 Importance of recruitment processes in the dynamics and resilience of coral reef assemblages. Mar. Anim. For. 549-569. (10.1007/978-3-319-21012-4_12) 3. Albright R, Mason B, Miller M, Langdon C. 2010 Ocean acidification compromises recruitment success of the threatened Caribbean coral Acropora palmata. Proc. Natl Acad. Sci. USA 107 , 20 400-20 404. (10.1073/pnas.1007273107) 4. Ries JB. 2011 A physicochemical framework for interpreting the biological calcification response to CO2-induced ocean acidification. Geochim. Cosmochim. Acta. 75 , 4053-4064. (10.1016/j.gca.2011.04.025) 5. Chan NCS, Connolly SR. 2013 Sensitivity of coral calcification to ocean acidification: a meta-analysis. Glob. Change Biol. 19 , 282-290. (10.1111/gcb.12011) 6. Comeau S, Edmunds PJ, Spindel NB, Carpenter RC. 2013 The responses of eight coral reef calcifiers to increasing partial pressure of CO2 do not exhibit a tipping point. Limnol. Oceanogr. 58 , 388-398. (10.4319/lo.2013.58.1.0388) 7. Fantazzini P et al . 2015 Gains and losses of coral skeletal porosity changes with ocean acidification acclimation. Nat. Commun. 6 , 7785. (10.1038/ncomms8785)26183259 8. Foster T, Falter JL, McCulloch MT, Clode PL. 2016 Ocean acidification causes structural deformities in juvenile coral skeletons. Sci. Adv. 2 , e1501130. (10.1126/sciadv.1501130)26989776 9. Davy SK, Allemand D, Weis VM. 2012 Cell biology of cnidarian-dinoflagellate symbiosis. Microbiol. Mol. Biol. Rev. 76 , 229-261. (10.1128/MMBR.05014-11)22688813 10. Pörtner H-O et al. 2019 IPCC special report on the ocean in a changing climate. Geneva, Switzerland: Intergovernmental Panel on Climate Change. 11. Edmunds PJ, Putnam HM. 2020 Science-based approach to using growth rate to assess coral performance and restoration outcomes. Biol. Lett. 16 , 20200227. (10.1098/rsbl.2020.0227)32673540 12. Holcomb M, Cohen AL, McCorkle DC. 2013 An evaluation of staining techniques for marking daily growth in scleractinian corals. J. Exp. Mar. Biol. Ecol. 440 , 126-131. (10.1016/j.jembe.2012.12.003) 13. Ganot P, Tambutté E, Caminiti-Segonds N, Toullec G, Allemand D, Tambutté S. 2020 Ubiquitous macropinocytosis in anthozoans. eLife 9 , e50022. (10.7554/eLife.50022)32039759 14. Drake JL, Mass T, Stolarski J, Von Euw S, Schootbrugge B, Falkowski PG. 2020 How corals made rocks through the ages. Glob. Change Biol. 26 , 31-53. (10.1111/gcb.14912) 15. Mass T, Drake JL, Peters EC, Jiang W, Falkowski PG. 2014 Immunolocalization of skeletal matrix proteins in tissue and mineral of the coral Stylophora pistillata. Proc. Natl Acad. Sci. USA 111 , 12 728-12 733. (10.1073/pnas.1408621111) 16. Zoccola D, Tambutté E, Sénégas-Balas F, Michiels JF, Failla JP, Jaubert J, Allemand D. 1999 Cloning of a calcium channel alpha1 subunit from the reef-building coral, Stylophora pistillata. Gene. 227 , 157-167. (10.1016/S0378-1119(98)00602-7)10023047 17. Bertucci A, Tambutté S, Supuran CT, Allemand D, Zoccola D. 2011 A new coral carbonic anhydrase in Stylophora pistillata. Mar. Biotechnol. 13 , 992-1002. (10.1007/s10126-011-9363-x) 18. Vidal-Dupiol J et al. 2013 Genes related to ion-transport and energy production are upregulated in response to CO2-driven pH decrease in corals: new insights from transcriptome analysis. PLoS ONE 8 , e58652. (10.1371/journal.pone.0058652)23544045 19. Tebben J et al. 2015 Chemical mediation of coral larval settlement by crustose coralline algae. Sci. Rep. 5 , 10803. (10.1038/srep10803)26042834 20. Webster NS, Uthicke S, Botté ES, Flores F, Negri AP. 2013 Ocean acidification reduces induction of coral settlement by crustose coralline algae. Glob. Change Biol. 19 , 303-315. (10.1111/gcb.12008) 21. Meyer E, Aglyamova GV, Matz MV. 2011 Profiling gene expression responses of coral larvae (Acropora millepora) to elevated temperature and settlement inducers using a novel RNA-Seq procedure. Mol. Ecol. 20 , 3599-3616.21801258 22. Strader ME, Aglyamova GV, Matz MV. 2018 Molecular characterization of larval development from fertilization to metamorphosis in a reef-building coral. BMC Genomics 19 , 17. (10.1186/s12864-017-4392-0)29301490 23. Riveron J, Boto T, Alcorta E. 2013 Transcriptional basis of the acclimation to high environmental temperature at the olfactory receptor organs of Drosophila melanogaster. BMC Genomics 14 , 259. (10.1186/1471-2164-14-259)23590196 24. Caroselli E, Gizzi F, Prada F, Marchini C, Airi V, Kaandorp J, Falini G, Dubinsky Z, Goffredo S. 2019 Low and variable pH decreases recruitment efficiency in populations of a temperate coral naturally present at a CO2 vent. Limnol. Oceanogr. 64 , 1059-1069. (10.1002/lno.11097) 25. Fabricius KE, Noonan SHC, Abrego D, Harrington L, De'ath G. 2017 Low recruitment due to altered settlement substrata as primary constraint for coral communities under ocean acidification. Proc. R. Soc. B 284 , 20171536. (10.1098/rspb.2017.1536) 26. Moya A, Huisman L, Forêt S, Gattuso J-P, Hayward DC, Ball EE, Miller DJ. 2015 Rapid acclimation of juvenile corals to CO2-mediated acidification by upregulation of heat shock protein and Bcl-2 genes. Mol. Ecol. 24 , 438-452. (10.1111/mec.13021)25444080 27. Anthony KRN, Connolly SR, Willis BL. 2002 Comparative analysis of energy allocation to tissue and skeletal growth in corals. Limnol. Oceanogr. 47 , 1417-1429. (10.4319/lo.2002.47.5.1417) 28. Houlbreque F. 2004 Interactions between zooplankton feeding, photosynthesis and skeletal growth in the scleractinian coral Stylophora pistillata. J. Exp. Biol. 207 , 1461-1469. (10.1242/jeb.00911)15037640 29. Pan T-CF, Applebaum SL, Manahan DT. 2015 Experimental ocean acidification alters the allocation of metabolic energy. Proc. Natl Acad. Sci. USA 112 , 4696-4701. (10.1073/pnas.1416967112)25825763 30. Fine M, Tchernov D. 2007 Scleractinian coral species survive and recover from decalcification. Science 315 , 1811-1811. (10.1126/science.1137094)17395821 31. Tambutté E, Venn AA, Holcomb M, Segonds N, Techer N, Zoccola D, Allemand D. 2015 Morphological plasticity of the coral skeleton under CO2-driven seawater acidification. Nat. Commun. 6 , 7368. (10.1038/ncomms8368)26067341 32. Cohen AL, McCorkle DC, de Putron S, Gaetani GA, Rose KA. 2009 Morphological and compositional changes in the skeletons of new coral recruits reared in acidified seawater: insights into the biomineralization response to ocean acidification. Geochem. Geophys. Geosyst. 10 , Q07005. (10.1029/2009GC002411) 33. Zoccola D, Innocenti A, Bertucci A, Tambutté E, Supuran CT, Tambutté S. 2016 Coral carbonic anhydrases: regulation by ocean acidification. Mar. Drugs 14 , 109. (10.3390/md14060109)27271641 34. DeCarlo TM, Comeau S, Cornwall CE, McCulloch MT. 2018 Coral resistance to ocean acidification linked to increased calcium at the site of calcification. Proc. R. Soc. B 285 , 20180564. (10.1098/rspb.2018.0564) 35. Teixidó N et al . 2020 Ocean acidification causes variable trait shifts in a coral species. Glob. Change Biol. 26 , 6813-6830. 36. Barott KL, Venn AA, Perez SO, Tambutté S, Tresguerres M. 2015 Coral host cells acidify symbiotic algal microenvironment to promote photosynthesis. Proc. Natl Acad. Sci. USA 112 , 607-612. (10.1073/pnas.1413483112)25548188 37. Wooldridge SA. 2013 Breakdown of the coral-algae symbiosis: towards formalising a linkage between warm-water bleaching thresholds and the growth rate of the intracellular zooxanthellae. Biogeosciences 10 , 1647-1658. (10.5194/bg-10-1647-2013) 38. Leal MC, Hoadley K, Pettay DT, Grajales A, Calado R, Warner ME. 2015 Symbiont type influences trophic plasticity of a model cnidarian–dinoflagellate symbiosis. J. Exp. Biol. 218 , 858-863. (10.1242/jeb.115519)25617454 39. Muscatine L, Cernichiari E. 1969 Assimilation of photosynthetic products of zooxanthellae by a reef coral. Biol. Bull. 137 , 506-523. (10.2307/1540172)28368714 40. Guillermic M et al. 2021 Thermal stress reduces pocilloporid coral resilience to ocean acidification by impairing control over calcifying fluid chemistry. Sci. Adv. 7 , eaba9958. (10.1126/sciadv.aba9958)33523983 41. Lin Z, Wang L, Chen M, Chen J. 2018 The acute transcriptomic response of coral-algae interactions to pH fluctuation. Mar. Genomics 42 , 32-40. (10.1016/j.margen.2018.08.006)30197044 42. Kaniewska P, Campbell PR, Kline DI, Rodriguez-Lanetty M, Miller DJ, Dove S, Hoegh-Guldberg O. 2012 Major cellular and physiological impacts of ocean acidification on a reef building coral. PLoS ONE 11 , e34659. (10.1371/journal.pone.0034659) 43. Baghdasarian G, Osberg A, Mihora D, Putnam H, Gates RD, Edmunds PJ. 2017 Effects of temperature and pCO2 on population regulation of Symbiodinium spp. in a tropical reef coral. Biol. Bull. 232 , 123-139. (10.1086/692718)28654331 44. Díaz-Almeyda EM, Prada C, Ohdera AH, Moran H, Civitello DJ, Iglesias-Prieto R, Carlo TA, Lajeunesse TC, Medina M. 2017 Intraspecific and interspecific variation in thermotolerance and photoacclimation in Symbiodinium dinoflagellates. Proc. R. Soc. B 284 , 20171767. (10.1098/rspb.2017.1767) 45. Kenkel CD, Moya A, Strahl J, Humphrey C, Bay LK. 2018 Functional genomic analysis of corals from natural CO2 seeps reveals core molecular responses involved in acclimatization to ocean acidification. Glob. Change Biol. 24 , 158-171. (10.1111/gcb.13833) 46. Putnam HM, Gates RD. 2015 Preconditioning in the reef-building coral Pocillopora damicornis and the potential for trans-generational acclimatization in coral larvae under future climate change conditions. J. Exp. Biol. 218 , 2365-2372. (10.1242/jeb.123018)26246609 47. Ward BA, Collins S, Dutkiewicz S, Gibbs S, Bown P, Ridgwell A, Sauterey B, Wilson JD, Oschlies A. 2019 Considering the role of adaptive evolution in models of the ocean and climate system. J. Adv. Model Earth Syst. 11 , 3343-3361. (10.1029/2018MS001452)32025278 48. Prada F et al. 2017 Ocean warming and acidification synergistically increase coral mortality. Sci. Rep. 7 , 40842. (10.1038/srep40842)28102293 49. Scucchia F, Malik A, Zaslansky P, Putnam HM, Mass T. 2021 Data from: Combined responses of primary coral polyps and their algal endosymbionts to decreasing seawater pH. Dryad Digital Repository. (10.5061/dryad.66t1g1k27)
PMC008xxxxxx/PMC8220279.txt
==== Front Proc Biol Sci Proc Biol Sci RSPB royprsb Proceedings of the Royal Society B: Biological Sciences 0962-8452 1471-2954 The Royal Society 34157874 10.1098/rspb.2021.0326 rspb20210326 10017014Evolution Research Articles Hosts elevate either within-clutch consistency or between-clutch distinctiveness of egg phenotypes in defence against brood parasites Hosts elevate either within-clutch consistency or between-clutch distinctiveness of egg phenotypes in defence against brood parasites http://orcid.org/0000-0003-3497-5925 Caves Eleanor M. Conceptualization Data curation Formal analysis Investigation Methodology Software Validation Visualization Writing-original draft Writing-review & editing [email protected] 1 2 http://orcid.org/0000-0001-5604-7965 Dixit Tanmay Conceptualization Writing-review & editing 1 Colebrook-Robjent John F. R. 3 † Hamusikili Lazaro Resources 3 http://orcid.org/0000-0001-7768-3426 Stevens Martin Conceptualization Methodology Software Supervision Writing-review & editing 1 2 http://orcid.org/0000-0001-5010-2177 Thorogood Rose Formal analysis Software Writing-review & editing 1 4 5 http://orcid.org/0000-0003-3232-9559 Spottiswoode Claire N. Conceptualization Funding acquisition Project administration Resources Supervision Writing-review & editing [email protected] 1 6 1 Department of Zoology, University of Cambridge, Downing Street, Cambridge CB2 3EJ, UK 2 Centre for Ecology and Conservation, University of Exeter, Penryn Campus, Penryn TR10 9FE, UK 3 Musumanene Farm, Choma, Zambia 4 HiLIFE Helsinki Institute of Life Sciences, University of Helsinki, Helsinki FI-00011, Finland 5 Research Programme in Organismal and Evolutionary Biology, Faculty of Biological and Environmental Sciences, University of Helsinki, Helsinki FI-00011, Finland 6 FitzPatrick Institute of African Ornithology, DST-NRF Centre of Excellence, University of Cape Town, Rondebosch 7701, South Africa † Deceased 17/11/2008 Electronic supplementary material is available online at https://doi.org/10.6084/m9.figshare.c.5459514. 30 6 2021 June 30, 2021 23 6 2021 June 23, 2021 23 6 2021 June 23, 2021 288 1953 202103268 2 2021 Feburary 8, 2021 1 6 2021 June 1, 2021 © 2021 The Authors. 2021 https://creativecommons.org/licenses/by/4.0/ Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited. In host–parasite arms races, hosts can evolve signatures of identity to enhance the detection of parasite mimics. In theory, signatures are most effective when within-individual variation is low (‘consistency’), and between-individual variation is high (‘distinctiveness’). However, empirical support for positive covariation in signature consistency and distinctiveness across species is mixed. Here, we attempt to resolve this puzzle by partitioning distinctiveness according to how it is achieved: (i) greater variation within each trait, contributing to elevated ‘absolute distinctiveness’ or (ii) combining phenotypic traits in unpredictable combinations (‘combinatorial distinctiveness’). We tested how consistency covaries with each type of distinctiveness by measuring variation in egg colour and pattern in two African bird families (Cisticolidae and Ploceidae) that experience mimetic brood parasitism. Contrary to predictions, parasitized species, but not unparasitized species, exhibited a negative relationship between consistency and combinatorial distinctiveness. Moreover, regardless of parasitism status, consistency was negatively correlated with absolute distinctiveness across species. Together, these results suggest that (i) selection from parasites acts on how traits combine rather than absolute variation in traits, (ii) consistency and distinctiveness are alternative rather than complementary elements of signatures and (iii) mechanistic constraints may explain the negative relationship between consistency and absolute distinctiveness across species. coevolution , avian brood parasitism , egg colour , egg pattern , host defence , egg signatures Balfour Studentship, Department of Zoology, University of Cambridge Biotechnology and Biological Sciences Research Council http://dx.doi.org/10.13039/501100000268 BB/G022887/1 BB/J014109/1 Natural Environment Research Council http://dx.doi.org/10.13039/501100000270 NE/K00929X/1 Pomona College and Downing College Exchange Fellowship University of Helsinki Institute of Life Science (HiLIFE) Helsinki Institute of Life Science Royal Society Dorothy Hodgkin Fellowship cover-dateJune 30, 2021 ==== Body pmc1. Introduction Whenever antagonistic coevolution involves mimicry as an offence, individuals are under selection to improve their detection of enemies [1,2]. In host–parasite arms races, this can result in hosts evolving ‘signatures’ of identity, which are individually distinctive phenotypes that facilitate detection of a parasite imposter while minimizing error [1,3]. Signatures are expected to be most effective if (i) host individuals vary distinctively with respect to other host signatures in a population, yet (ii) their individual signatures are consistent and vary as little as possible. Together, these adaptations are hypothesized to make it difficult for parasites to fool hosts because consistent signatures require mimics to be an even better match, but distinctive signatures mean parasites can only mimic a small number of host phenotypes [1,3–6]. However, the questions of whether both adaptations co-occur within single systems, and how they manifest, have largely remained untested (but see [6–8]). Avian brood parasitism, in which some brood-parasitic birds lay eggs that visually mimic those of their hosts to trick hosts into accepting the foreign egg as one of their own [9], provides a model system for exploring these questions. Over a century ago, Charles Swynnerton conducted a pioneering study of avian egg rejection in south-eastern Africa [3]. He hypothesized that if individual host females in a population were more phenotypically distinct from one another (adaptation (i) above), it would be harder for a parasite to lay eggs sufficiently mimetic to be accepted by a large proportion of host females, thus reducing Type II errors (false negatives; accepting a parasite egg). Subsequently, adaptation (ii) has also been applied to brood parasitism: if, within a clutch, females lay eggs that appear very similar to one another (i.e. more phenotypically consistent within individuals), it should be easier for females to identify parasitic eggs (lowering Type II errors [10]). Consistency within clutches should also have an added benefit by helping hosts to avoid erroneously rejecting their own eggs (thus also reducing Type I errors [11]). These hypotheses have generated the predictions that across species or populations, selection from brood parasites should be associated with both (i) greater interclutch variation (increased ‘distinctiveness’) between clutches laid by different females and (ii) reduced intraclutch variation (increased ‘consistency’) between eggs in a clutch laid by the same female, with respect to egg colour and pattern [12]. Here onwards, we use the terms ‘distinctiveness’ and ‘consistency’ (rather than ‘interclutch variation’ and ‘intraclutch variation’) because they are more generalizable to signature systems outside of avian brood parasitism, and because predictions are easier to visualize when selection for defence predicts elevation of both adaptations (rather than predicting elevation of one adaptation but reduction of the other, as for interclutch and intraclutch variation). Despite the long-standing predictions that selection from parasitism should be associated with both elevated distinctiveness and consistency in host egg clutches, evidence supporting them is so far mixed. Studies relating to egg signature evolution have typically inferred the strength of selection from brood parasites from either the incidence of parasitism, or the intensity of host egg rejection. According to the predictions above, greater estimated selection from parasites should be correlated with increased distinctiveness and increased consistency. When testing these predictions, some studies have found that estimates of parasitism were positively correlated with distinctiveness across species [13–16], but others have not [4,17]. The comparative evidence for an association between parasitism and consistency is similarly mixed, with some support for a positive association [13,17–20] (but see for example [14,21]). Many studies within single species have also failed to find that consistency is associated with improved egg rejection ability [22–26] (but see [17,27,28]). However, compelling evidence for an association between parasitism and both distinctiveness and consistency has come from two studies of introduced populations, where hosts released from parasitism exhibited reduced distinctiveness and reduced consistency over 100–200 years of subsequent evolution [6–8]. How, then, can we reconcile these contrasting results? One reason why studies have yielded conflicting results may be that distinctiveness and consistency need not occur hand-in-hand as defences against parasites. Rather, they may function as alternatives [29–31], if selection on one defence reduces the frequency of successful parasitism and thus weakens selection on the other [29]. Under such a ‘strategy-blocking’ scenario [32], hosts with high interclutch distinctiveness will rarely encounter a good match by the parasite, so hosts will not experience strong selection to further refine their detection and rejection via intraclutch consistency [12]. Reciprocally, hosts with high intraclutch consistency will often be able to detect a parasite even if it is a good match, and so not experience strong selection for interclutch distinctiveness. When comparing across host species, this strategy-blocking hypothesis generates different predictions to the traditional hypothesis. The latter predicts that consistency and distinctiveness in egg appearance should both be elevated, and positively correlated, in parasitized species, but only weakly correlated or uncorrelated across unparasitized species. By contrast, the strategy-blocking hypothesis predicts that either consistency or distinctiveness should be elevated in parasitized species. Therefore, there should be a negative correlation between consistency and distinctiveness in parasitized, but not unparasitized, species. A second reason for the conflicting results may relate to how distinctiveness is achieved. Two key traits of bird eggs include colour and pattern, which can be quantified using a variety of metrics (e.g. hue, saturation, marking size, etc.). Typically, distinctiveness is quantified by summing the absolute level of variation within each metric (‘absolute distinctiveness’). Higher absolute distinctiveness is conferred by greater interclutch variation in individual colour and pattern metrics (figure 1, left column). However, higher distinctiveness can arise in a second way if colour and pattern trait values co-occur in unpredictable combinations (‘combinatorial distinctiveness’; figure 1, top row). In a species with high combinatorial distinctiveness, absolute levels of variation within each metric may be low, but because each individual's eggs comprise a different combination of values for each metric, they are all distinct from those of other individuals (figure 1c). However, combinatorial distinctiveness is rarely investigated, or separated from absolute distinctiveness. Figure 1. An illustration (a–d) and examples (e–h) of eggs from species with high and low levels of absolute and combinatorial distinctiveness. In each panel, the four eggs are from four different females. In the schematic (a–d), eggs vary in two traits (colour and pattern); for simplicity, each is described here by one metric (background hue and spot size). (a) The best scenario (for a host) is high variation in the metrics of spot size and background hue compared to other clutches (high absolute distinctiveness), and for the two to be uncorrelated (high combinatorial distinctiveness), making these eggs hardest for a parasite to mimic. (b) When spot size and background hue are correlated, one can be predicted from the other, so combinatorial distinctiveness is low, a situation which is not ideal for a host as any given parasite phenotype might be a good enough match to a larger subset of host phenotypes. (c) A species constrained for whatever reason to have low absolute distinctiveness is able to maximize variation among females by having high combinatorial distinctiveness. (d) The worst scenario (for a host) is low variation in both spot size and background hue and for the two to also be correlated, such that one can be predicted from the other. (e–h) Eggs from representative species in our dataset that exhibit high or low levels of absolute and combinatorial distinctiveness, relative to other species in our dataset. (Online version in colour.) We previously found that the eggs of host species had consistently elevated combinatorial distinctiveness (reflected by the correlation component of entropy), but not absolute distinctiveness (reflected by the variance component of entropy), compared with the eggs of non-hosts [33]. This suggested that selection from brood parasites may act on how hosts deploy phenotypic variation into combinations of trait values, rather than on absolute levels of variation within each metric. If selection acts more strongly on combinatorial than absolute distinctiveness, then this may account for the inconsistent findings of previous studies that have focussed only on absolute distinctiveness. Additionally, examining how consistency relates to each type of distinctiveness in both parasitized and unparasitized species may help clarify how selection from brood parasites acts on each defence. Here we tested the predictions of the two hypotheses outlined above for the evolution of host egg signatures as a defence against brood parasites, by examining whether, and how, distinctiveness and consistency covary. The traditional hypothesis predicts that both consistency and distinctiveness are elevated, and positively correlated, across host species. The strategy-blocking hypothesis predicts that either consistency or distinctiveness should be elevated, and negatively correlated with one another, across parasitized species. We studied the same two families of African birds that inspired Swynnerton a century ago: the African warblers (Cisticolidae), many of which are parasitized by the cuckoo finch Anomalospiza imberbis [34]; and the weavers (Ploceidae), many of which are parasitized by the diederik cuckoo Chrysococcyx caprius [35]. These two systems have evolved independently from each other and have an ancient history of parasitism [36,37]. They are also notable for their diversity of egg phenotypes, with some species showing very high levels of interclutch distinctiveness resulting in elaborate ‘signatures’ of colour and pattern, and others not [33,38] (figure 2). Although each parasitic species has distinct host-races that mimic the variation in egg colour and pattern of their specialist host [35,38], parasitic females lay eggs haphazardly among nests of their host species rather than targeting individual females with phenotypes that match their own [39,40]. Therefore, distinctiveness (whether achieved by absolute and/or combinatorial mechanisms) could be important for anti-parasitic defence in both systems. Figure 2. (a) Intraclutch consistency and (b) absolute (downward triangles) and combinatorial (upward triangles) distinctiveness for species of weaver (i) and warbler (ii). Black symbols indicate species currently parasitized at our study site. Dashed symbols indicate species not currently parasitized at our study site: light-grey dashed symbols indicate species with parasitism records from elsewhere in Africa; white dashed symbols indicate species with no parasitism records. Violin plots show distributions and ranges, diamonds show means, and bars show standard error. Consistency is a clutch-level measure, whereas distinctiveness is a species-level measure, so ranges are provided for consistency but not distinctiveness. Egg photos are representative examples of egg phenotypes, each from a different clutch. It is not possible to calculate entropy, and thus absolute and combinatorial distinctiveness, for the three species that lay only immaculate eggs (see Material and Methods); therefore, no values are displayed for those three species. (Online version in colour.) 2. Material and methods (a) Study system We analysed 806 clutches (comprising 1942 eggs) from 11 warbler species (five parasitized, six unparasitized at our study site) and 14 weaver species (10 parasitized, four unparasitized at our study site) (electronic supplementary material, table S1). All eggs came from the Choma region of southern Zambia (near 16°47′S, 26°50′E) where they were collected by JFRCR and LH during the 1970s–1990s. We only analysed eggs collected in the Choma District and the districts of Monze and Mazabuka (centred 80 km and 130 km north-east of Choma, respectively). Data from this same set of eggs have previously been analysed as reported in [33,38]. To classify species as parasitized or unparasitized, we used data on 1490 collected clutches (range: 10–227 clutches, mean: 59.6 clutches) for our 25 study species, collected in the Choma region (including the Monze and Mazabuka districts) over 38 years by J.F.R.C.-R. and L.H. (electronic supplementary material, table S1). Each collected clutch was labelled as either parasitized or unparasitized, from which we calculated a parasitism rate in the Choma region (range: 2.17–42.7%, electronic supplementary material, table S1). These rates are an imperfect index of parasitism pressure (e.g. some parasitic eggs could have been rejected before nests were found), but give a reliable indication of whether a species was regularly parasitized during the study period. We therefore categorized all species with non-zero parasitism rates as ‘parasitized’. However, we cannot know with certainty whether any currently unparasitized species have previously acted as hosts. As a precaution, we therefore repeated all analyses treating as parasitized five locally unparasitized species with published parasitism records from elsewhere in their range [41] (electronic supplementary material, table S1); this left only one unparasitized warbler in our dataset, reducing statistical power. (b) Quantifying egg colour and pattern We focused on two traits which vary intraspecifically in the species in this study: colour and pattern. We quantified ten metrics following previously published methods [33] that included four measures of egg colour, one of egg luminance and five of egg pattern. In brief, to describe variation in colour, we used reflectance spectra to calculate avian photon catches (using data from the blue tit Cyanistes caeruleus [42]) for the UV, SW, MW, LW and double cones as measures of colour and luminance, respectively (as in [40]). To describe variation in pattern, we applied a granularity analysis [43] to digital photographs of eggs to quantify five pattern metrics, that have been previously applied to eggs [29,33,38,40,44]. These were (i) size of the predominant marking, (ii) contribution of the main marking to the overall pattern, (iii) contrast between pattern markings and the background, (iv) proportion of the egg covered by markings and (v) how dispersed markings were between the poles of the egg. Each pattern attribute and luminance was standardized for analyses by expressing it as a proportion of its maximum value within each family (warblers or weavers), so that the scale was comparable between all of the phenotypic attributes (as in [29,33]). Cone catch values were standardized to remove variation in absolute brightness, such that the standardized cone catch values sum to one (as in [33]). As most of these metrics predict egg discrimination in cuckoo finch host species [29], we assumed that they represent biologically relevant metrics on which selection may act. We then used these metrics to calculate consistency, absolute distinctiveness and combinatorial distinctiveness in colour and pattern for each species, which we analysed in relation to current selection from brood parasites. (c) Quantifying consistency using multi-dimensional phenotypic space To quantify phenotypic consistency within clutches, we calculated an index of overall diversity for our 10 focal colour and pattern metrics (four standardized single cone catches, luminance and five pattern metrics) combined. Following [29], we used a multi-dimensional phenotypic space (MDPS) analysis, in which each egg was mapped as a point in 10-dimensional space. The Euclidean distance between two points in this 10-dimensional space then provided a single measure of overall phenotypic distance between any two eggs in a group. Each group was the clutch of a given female, which ranged from one to five eggs. To make clutches of different sizes directly comparable, we first eliminated all one-egg clutches from our dataset (as consistency cannot be calculated for one-egg clutches; this resulted in 704 clutches comprising 1840 eggs) and then carried out analyses on an ‘effective clutch’ (following [7]), defined as two randomly selected eggs from each clutch. A species index of intraclutch variability was taken as the mean of all pairwise comparisons within the effective clutches of a given species; we then subtracted this from one to yield an index of intraclutch consistency. Thus, high consistency values indicate that the eggs within a clutch are highly similar to one another in colour and pattern. We also quantified distinctiveness using an MDPS method, which we show in the electronic supplementary material is conceptually analogous to absolute distinctiveness, and highly correlated with it. (d) Quantifying absolute and combinatorial distinctiveness As measures of absolute and combinatorial distinctiveness, we used published values from [33], which analysed the same set of host and non-host species as examined here (see [33] for detailed methods). Briefly, species-specific values of ‘differential’ entropy (an extension of Shannon entropy for continuous variables [45,46]) were calculated using the same 10 phenotypic metrics described above (see the electronic supplementary material of [33] for entropy formulae). Values of entropy depend on the sum of two components, the variance component (the sum of the absolute variation of each metric, ‘absolute distinctiveness’) and the correlation component (how values for each metric are assembled within individuals; ‘combinatorial distinctiveness’). Entropy was then decomposed into contributions from the variance and correlation components, which were uncorrelated (F1,20 = 1.83, R2 = 0.04, p = 0.19), and which here represent the ‘absolute’ and ‘combinatorial’ components of distinctiveness, respectively. One egg per clutch was randomly selected for calculations of absolute and combinatorial distinctiveness to avoid pseudoreplication. Three weaver species in our study lay only immaculate eggs (Anaplectes melanotis, Euplectes orix and Ploceus intermedius; figure 2), meaning that the values for the five pattern attributes in these species were zero. Although MDPS analyses are not confounded by zero values, entropy cannot be measured on a comparable scale when any of the ten phenotypic attributes has a value of zero. Therefore, while we were able to calculate consistency, we were not able to calculate absolute and combinatorial distinctiveness for these three species. (e) Statistical analyses We used R [47] to calculate distances in multi-dimensional phenotypic space and to implement linear models (using the lm function). First, we examined how consistency and both absolute and combinatorial distinctiveness were influenced by various factors. In these models, either consistency, absolute distinctiveness or combinatorial distinctiveness was the response variable, and parasitism status (parasitized or unparasitized), family membership (warbler or weaver) and sample size were fixed effects; where these contributed significantly to the model, we report their effects. Note that the models in which absolute or combinatorial distinctiveness were predicted by parasitism status, family membership and sample size are similar to those that are described in [33]. They differ only in that here we make use of an updated phylogeny for phylogenetic least-squares (PGLS) analyses and account for differences in sample size by including sample size as a covariate, rather than weighting the linear models by sample size, because the assumption of heteroscedasticity was met in our models, and because model coefficients are easier to interpret when they themselves are not weighted. For comparison with the previous study, however, we repeated all linear models weighting by sample size, and results were largely consistent with those using sample size as a covariate (electronic supplementary material, tables S2, S3 and S4). In the main text, we report the results of analyses that include sample size as a covariate, for consistency among the formulation of linear models in this study. Second, we asked whether there was a positive or a negative relationship between consistency and either absolute or combinatorial distinctiveness across species. We modelled consistency as the response variable, absolute or combinatorial distinctiveness as the predictor variable, and parasitism status, family membership, and sample size as covariates. Finally, we asked whether the relationships between consistency and either absolute or combinatorial distinctiveness differed for parasitized and unparasitized species, by including an interaction between the predictor variable and parasitism status in these models. Sample sizes varied among species (electronic supplementary material, table S1), so we checked for any bias caused by unequal sample sizes bias in two ways. First, we included sample size as a covariate in analyses using the full dataset. Second, we resampled each species to the minimum sample size (n = 5) and repeated analyses using consistency recalculated using this smaller dataset, to check whether results were affected (none were; see electronic supplementary material, table S2). In models analysing resampled datasets, sample size was not included as a fixed effect. We used Cook's distance [48] to identify outliers in any analyses; where outliers were identified, we repeated analyses with those species excluded (electronic supplementary material, table S3). In some models, there were minor deviations from normality of residuals, so as a precaution we repeated all of the above analyses with the dependent variable expressed in ranks, but all conclusions were unchanged (electronic supplementary material, table S2). (f) Accounting for phylogenetic relatedness We used the R package phytools [49] to calculate Pagel's λ [50], a measure of phylogenetic signal that indicates the level of match between the model's residuals and the structure of the phylogeny. λ normally varies between zero, indicating phylogenetic independence, and one, indicating direct covariance between the species' phenotypic values of interest and phylogenetic structure. Some species in our study have either not been formally placed on a phylogenetic tree, or placed but with low confidence; we therefore used birdtree.org [51] to compile 100 trees with branch lengths for our focal species. We then calculated both λ and a p-value for a log-likelihood test of significant phylogenetic signal and examined the mean and standard deviation of each across the 100 trees. We found no evidence of phylogenetic signal in intraclutch consistency (λ < 0.001 and p = 1 in all trees), absolute distinctiveness (λ < 0.001 and p = 1 in all trees) or combinatorial distinctiveness (λ mean ± s.d. = 0.29 ± 0.03; range of p across all trees: 0.07–0.12). However, all analyses were repeated using PGLS models implemented by the R package caper [52], to account for the fact that related species are not statistically independent owing to shared phylogenetic history [53]. PGLS models did not include family membership as a fixed effect. 3. Results (a) Consistency, absolute distinctiveness and combinatorial distinctiveness in parasitized versus unparasitized species First, we asked whether parasitized species have consistently higher levels of each defence than unparasitized species (figure 2). We found no significant differences in consistency between currently parasitized and unparasitized species (slope ± s.e. = 0.004 ± 0.033, t21 = 0.12, p = 0.91). Results were similar when we took a more restrictive definition of parasitism status, scoring the five species parasitized outside of Choma (our main study area) as parasitized (slope ± s.e. = −0.045 ± 0.041, t21 = −1.08, p = 0.29). Repeating the analyses above while taking phylogenetic structure into account (and therefore not modelling family membership as a factor) did not change any of the conclusions above: consistency did not differ in relation to parasitism status at our study site (slope ± s.e. = 0.005 ± 0.037, t16 = 0.15, p = 0.88), or when species parasitized outside of Choma were treated as parasitized (slope ± s.e. = −0.048 ± 0.052, t16 = −0.92, p = 0.37). We also repeated all analyses (i) with species-specific values for consistency calculated from data resampled to the sample size of the least sampled species; (ii) omitting the three species that lay only immaculate eggs and (iii) with the dependent variable expressed in ranks, and in all cases conclusions were unchanged (electronic supplementary material, table S2). According to a very similar previous analysis [33], differing only with respect to the exact formulation of the linear models used and in applying a more updated phylogenetic tree for PGLS analyses (see Material and Methods), we found that only combinatorial distinctiveness, and not absolute distinctiveness, was consistently elevated in parasitized over unparasitized species (combinatorial distinctiveness: slope ± s.e. = −0.31 ± 0.12, t18 = −2.67, p = 0.02; absolute distinctiveness: slope ± s.e. = −0.28 ± 0.30, t18 = −0.94, p = 0.36). These results were unchanged when performing the analyses using ranked data, with a phylogenetic correction, or when weighting the results by sample size (electronic supplementary material, table S4). In summary, we found no relationship between parasitism status and either consistency or absolute distinctiveness in egg phenotype, contrary to the traditional pair of predictions that selection from brood parasites should result in higher levels of both defences. We did, however, find that combinatorial distinctiveness is elevated in parasitized over unparasitized species, replicating our previous finding [33]. (b) The relationship between intraclutch consistency and both absolute and combinatorial distinctiveness We then examined how consistency related to both absolute and combinatorial distinctiveness. In an additive model, the relationship between consistency and absolute distinctiveness was significantly negative (slope ± s.e. = −0.06 ± 0.02, t17 = −3.04, p = 0.007). Allowing an interaction between parasitism status and absolute distinctiveness, the interaction term was not significant (slope ± s.e. = −0.039 ± 0.04, t16 = −0.97, p = 0.35). The interaction term remained non-significant when locally unparasitized species were treated as parasitized; when excluding two species (Euplectes axillaris and Plocepasser mahali) that were identified as outliers using Cook's distance; when applying a phylogenetic correction; or when using values of consistency generated from five randomly sampled clutches per species (electronic supplementary material, table S3). Therefore, the relationship between consistency and absolute distinctiveness did not detectably differ between parasitized and unparasitized species. In contrast with the negative relationship between consistency and absolute distinctiveness, we found no significant relationship between consistency and combinatorial distinctiveness across all species (slope ± s.e. = −0.11 ± 0.06, t17 = −1.79, p = 0.10). However, in a model allowing an interaction between combinatorial distinctiveness and consistency, the interaction term was significant (slope ± s.e. = 0.15 ± 0.07, t16 = 2.05, p = 0.05), showing that the slope of the relationship between consistency and combinatorial distinctiveness is different in parasitized versus unparasitized species (figure 3b). In particular, the relationship between consistency and combinatorial distinctiveness was significantly negative in a model including only parasitized species (slope ± s.e. = −0.18 ± 0.07, t8 = −2.72, p = 0.03), but not significant in a model including only unparasitized species (slope ± s.e. = −0.03 ± 0.12, t6 = −0.21, p = 0.84). The interaction between combinatorial distinctiveness and parasitism status became marginally non-significant when we incorporated a phylogenetic correction (p = 0.07) and used resampled data (p = 0.07), and was not significant when species parasitized elsewhere in Africa were treated as parasitized or when the outliers P. mahali and E. axillaris were removed (electronic supplementary material, table S3). However, as noted in the Material and Methods, these supplementary analyses had low power to distinguish between parasitized and unparasitized species due to low sample sizes. Figure 3. A visualization of the linear models describing the relationship between intraclutch consistency and (a) absolute distinctiveness and (b) combinatorial distinctiveness, while accounting for family membership and sample size, in warbler (circles) and weaver (squares) species. Black symbols and lines indicate species currently parasitized at our study site. Dashed symbols and lines indicate species not currently parasitized at our study site; light-grey dashed symbols indicate species with parasitism records from elsewhere in Africa, and white symbols indicate species with no parasitism records. The slopes for parasitized and unparasitized species significantly differ in (b) but not in (a) (see main text) and the line of best fit for all species combined (solid grey) is provided for comparison. The species identified as statistical outliers, Plocepasser mahali and Euplectes axillaris, have been marked with an asterisk. A visualization of the relationship between intraclutch consistency and both absolute and combinatorial distinctiveness, with 95% confidence intervals, is available in electronic supplementary material, figure S1. 4. Discussion The expectation that hosts under selection from parasites should evolve signatures that are distinctive among females, yet consistent within a female, to help them recognize their own eggs, has received extensive testing in avian brood parasite-host systems. Here, in two families of African birds in which several species are respectively parasitized by a cuckoo and the cuckoo finch, we found support for two potential reasons why previous results have been inconsistent in their support for this elegant hypothesis. First, in parasitized species, we found a negative relationship between consistency and combinatorial distinctiveness, which supports the hypothesis that these are alternative rather than simultaneous defences. This finding goes against the traditional expectation that selection for effective signatures of identity should elevate both consistency and distinctiveness [5]. Second, we found a negative relationship between consistency and absolute distinctiveness that was the same for parasitized and unparasitized species, suggesting that selection from brood parasites favours combinatorial rather than absolute distinctiveness (i.e. distinctiveness achieved by assembling unpredictable combinations of egg colour and pattern trait values, rather than higher levels of variation in each). Therefore, previous studies might have missed a relationship between consistency and distinctiveness depending upon how they defined and quantified distinctiveness. These results lead us to ask two questions: (i) why should hosts elevate only consistency or combinatorial distinctiveness rather than both? (ii) Why is the relationship between consistency and absolute distinctiveness negative in both hosts and non-hosts? (a) Why should hosts elevate only a single defense rather than both? Our results support the hypothesis that elevating either consistency or combinatorial distinctiveness allows hosts to successfully detect parasitic eggs. Experimental data currently exist for two of our focal species to examine whether this is also supported when parasitic egg detection is measured directly. The tawny-flanked prinia Prinia subflava lays eggs with high absolute distinctiveness (figure 2), while the red-faced cisticola Cisticola erythrops lays only blotched or stippled eggs with a turquoise background that have low absolute distinctiveness (the lowest of any warbler in our dataset). At first sight, the two species appear to have very different defences: P. subflava eggs appear quite variable while C. erythrops eggs are not. Both species, however, have high combinatorial distinctiveness and low consistency, and both reject parasitic eggs equally well [29]. Thus, this behavioural evidence supports the idea that parasitism pressure acts on combinatorial distinctiveness, and that species with high combinatorial distinctiveness will not experience strong selection for consistency. Experimental data on more species, especially those that exhibited high consistency and low combinatorial distinctiveness here (e.g. desert cisticola C. aridulus), will help to reveal whether hosts prioritizing either consistency or distinctiveness can have similar outcomes for the detection of parasitic eggs. Prioritizing one defence over another also provides support for strategy-blocking, whereby selection on one defence lowers selection on another defence [32]. For example, species with high distinctiveness will rarely encounter a parasitic egg that is a good match for their own, so hosts will not experience strong selection to further refine either consistency or egg rejection behaviour. Interestingly, support for a strategy-blocking hypothesis was only found when distinctiveness was measured as combinatorial distinctiveness and not as absolute distinctiveness. Thus, precisely what type of individually identifying information is encoded in an egg phenotype may determine how consistency and distinctiveness are expressed by a given female or species. This means that the way in which distinctiveness is measured can affect the conclusions of studies on interclutch variation. Why are some host species consistent in their egg phenotypes, and others (combinatorially) distinct? Which defence is elevated by coevolution with parasites may be influenced by selection from ecological factors aside from brood parasitism: certain egg phenotypes may be costly with respect to thermoregulation, protection from UV radiation, or camouflage [54–57], and increased susceptibility to host colonizations by other species or host races of the parasite [29,38], potentially limiting distinctiveness. Moreover, non-adaptive factors may also mean that the null hypothesis is not necessarily that both consistency and distinctiveness should be low in the absence of parasitism. For example, non-adaptive mechanisms might account for the high levels of consistency observed in some unparasitized species; we could speculate that certain egg phenotypes may be relatively invariant within females for mechanistic reasons during pigment deposition. Variation in such potential constraints may influence whether high consistency or distinctiveness is favoured, and thus why the closely related, sympatric species studied here have taken divergent trajectories from their similar phylogenetic and ecological starting points. Finally, one inevitable problem shared by all comparative studies relating defensive traits to current selection from brood parasites is that coevolution is dynamic. Species that are not currently parasitized may have experienced selection from brood parasites in the past. For example, the rattling cisticola Cisticola chiniana is currently unparasitized at our study site yet shows intermediate levels of both interclutch distinctiveness and discrimination behaviour. As it is parasitized elsewhere, this suggests that it may have locally defeated its parasite [29]. Similarly, the grey-backed camaroptera Camaroptera brevicaudata is parasitized by emerald cuckoos Chrysococcyx cupreus elsewhere in its range [39] although no parasitism records exist from the Choma region despite intensive searching (by J.F.R.C.-R. and L.H.). Gene flow from emerald cuckoo-parasitized populations might help to account for its variously blue, white and immaculate to heavily blotched eggs. We attempted to reduce any such confounding effects by reanalysing our data with such species re-assigned and excluded. An undetected history of parasitism might also account for the surprisingly high levels of distinctiveness shown by some species with no records of parasitism anywhere in their range, such as the red-billed buffalo weaver Bubalornis niger and fan-tailed widowbird Euplectes axillaris. (b) Why is the relationship between consistency and absolute distinctiveness negative in both hosts and non-hosts? We found a negative relationship between consistency and absolute distinctiveness for both parasitized and unparasitized species, suggesting that selection from brood parasites cannot entirely explain these results. One possibility is that this negative relationship arises from a mechanistic constraint that is not directly related to parasitism. Instead, it may be indicative of a fundamental trade-off between mechanisms which produce repeatability in signatures (which can increase consistency within clutches) and mechanisms which produce randomness in signatures (which can promote distinctiveness between the clutches of different females). Repeatability and randomness are somewhat antithetical, as repeatability requires predictability, whereas randomness entails unpredictability [58]. Very little is known about the mechanisms of colour and pattern generation in the shell gland [30,59]. However, we can speculate that consistency within clutches must be produced by mechanisms which ensure repeatability in pattern, whereas distinctiveness could be produced through some level of randomness in the generation of signatures. If this is the case, the mechanisms involved in producing consistency may reduce distinctiveness and vice versa. This could potentially explain why we find a negative relationship in both parasitized species (where we would expect selection on consistency and distinctiveness) and unparasitized species (where we would expect no or weaker selection). 5. Conclusion Overall, our data suggest that in these two bird families, the egg signatures of different species lie on a spectrum of low distinctiveness and high consistency to high distinctiveness and low consistency. When distinctiveness is generated specifically by combinatorial information, this spectrum is seen only in parasitized species, consistent with a strategy-blocking scenario whereby selection for one defence reduces the strength of selection favouring the other defence. Previous studies found varying levels of support for the hypotheses that selection from parasitism should elevate consistency and/or distinctiveness in host eggs. Our results suggest that this could be explained if those host families also rely primarily on combinatorial rather than absolute distinctiveness in defence against brood parasites. We suggest that using entropy as a way of conceptualizing and measuring both combinatorial and absolute distinctiveness may help to clarify the relationship between distinctiveness and either incidence of parasitism or rejection ability. This study is correlative, but comparative analyses together with behavioural and physiological experiments on a wider range of species will help to test the generality of the mechanisms proposed here. In particular, we should investigate the mechanisms involved in producing consistent or distinctive eggs, and explore the costs of these phenotypes from other sources of selection besides brood parasitism. Future research should also test whether negative relationships between consistency and measures of distinctiveness are a default property of birds' eggs, or indeed biological patterns in general, rather than an adaptation to parasitism. More broadly, by considering the relationship between consistency and distinctiveness across a range of systems involving discrimination of self from non-self, we may enhance our understanding of how and why these systems follow a variety of trajectories when hosts coevolve with mimetic antagonists. Supplementary Material Click here for additional data file. Acknowledgements In Zambia, we thank Emma and Ian Bruce-Miller for their hospitality, Jeroen Koorevaar for help with egg photography, and all of J.F.R.C.-R.'s many field assistants who found the nests involved in this study. We thank Jess Lund and two anonymous reviewers for helpful comments on earlier drafts. We thank Dr Edwin Iversen for helping to devise the application of entropy to egg signatures. Data accessibility Data and R codes associated with this manuscript are available from the Dryad Digital Repository https://doi.org/10.5061/dryad.02v6wwq34 [60]. Authors' contributions C.N.S., M.S. and E.M.C. conceived the study; J.F.R.C.-R. and L.H. collected eggs and breeding data and helped inspire the study; E.M.C. collected the data from J.F.R.C.-R.'s egg collection; M.S. and E.M.C. performed visual modelling; E.M.C. analysed the data; E.M.C., T.D., C.N.S., M.S. and R.T. interpreted the data and E.M.C., T.D., C.N.S. and R.T. wrote the manuscript, with contributions from M.S. Competing interests We declare we have no competing interests. Funding E.M.C. was supported by the Pomona College-Downing College Student Exchange Scholarship; T.D. by a Balfour Studentship from the Department of Zoology, University of Cambridge; R.T. by an Independent Research Fellowship from the Natural Environment Research Council UK (grant no. NE/K00929X/1) and a start-up grant from the Helsinki Institute of Life Science (HiLIFE), University of Helsinki; MS by a BBSRC David Phillips Fellowship (grant no. BB/G022887/1), and C.N.S. by a Royal Society Dorothy Hodgkin Fellowship and BBSRC David Phillips Fellowship (grant no. BB/J014109/1). ==== Refs References 1. Spottiswoode CN, Busch R. 2019 Vive la difference! Self/non-self recognition and the evolution of signatures of identity in arms races with parasites. Phil. Trans. R. Soc. B 374 , 20180206. (10.1098/rstb.2018.0206)30967089 2. Tibbetts EA, Dale J. 2007 Individual recognition: it is good to be different. Trends Ecol. Evol. 22 , 529-537. (10.1016/j.tree.2007.09.001)17904686 3. Swynnerton CFM. 1918 Rejections by birds of eggs unlike their own: with remarks on some of the Cuckoo problems. Ibis (Lond. 1859) 6 , 127-154. (10.1111/j.1474-919X.1918.tb00774.x) 4. Davies NB, Brooke MdL. 1989 An experimental study of co-evolution between the cuckoo, Cuculus canorus, and its hosts. II. Host egg markings, chick discrimination and general discussion. J. Anim. Ecol. 58 , 225-236. (10.2307/4996) 5. Moskát C, Avilés JM, Bán M, Hargitai R, Zölei A. 2008 Experimental support for the use of egg uniformity in parasite egg discrimination by cuckoo hosts. Behav. Ecol. Sociobiol. 62 , 1885-1890. (10.1007/s00265-008-0618-0) 6. Lahti DC. 2006 Persistence of egg recognition in the absence of cuckoo brood parasitism: pattern and mechanism. Evolution 60 , 157-168. (10.1111/j.0014-3820.2006.tb01090.x)16568640 7. Lahti DC. 2005 Evolution of bird eggs in the absence of cuckoo parasitism. Proc. Natl Acad. Sci. USA 102 , 18 057-18 062. (10.1073/pnas.0508930102) 8. Yang C, Liu Y, Zeng L, Liang W. 2014 Egg color variation, but not egg rejection behavior, changes in a cuckoo host breeding in the absence of brood parasitism. Ecol. Evol. 4 , 2239-2246. (10.1002/ece3.1096)25360264 9. Davies NB. 2000 Cuckoos, cowbirds, and other cheats. London, UK: T and AD Poyser. 10. Davies NB, Brooke MdL. 1989 An experimental study of co-evolution between the cuckoo, Cuculus canorus, and its hosts. I. Host egg discrimination. J. Anim. Ecol. 58 , 207-224. (10.2307/4995) 11. Marchetti K. 1992 Costs to host defence and the persistence of parasitic cuckoos. Proc. R. Soc. Lond. B 248 , 41-45. (10.1098/rspb.1992.0040) 12. Davies NB, Brooke MdL, Kacelnik A. 1996 Recognition errors and probability of parasitism determine whether reed warblers should accept or reject mimetic cuckoo eggs. Proc. R. Soc. Lond. B 263 , 925-931. (10.1098/rspb.1996.0137) 13. Wang L, Liang W, Yang C, Cheng SJ, Hsu YC, Lu X. 2016 Egg rejection and clutch phenotype variation in the plain prinia Prinia inornata. J. Avian Biol. 47 , 788-794. (10.1111/jav.00786) 14. Landstrom M, Heinsohn R, Langmore NE. 2010 Clutch variation and egg rejection in three hosts of the pallid cuckoo, Cuculus pallidus. Behaviour 147 , 19-36. (10.1163/000579509X12483520922043) 15. Øien IJ, Moksnes A, Røskaft E. 1995 Evolution of variation in egg color and marking pattern in European passerines: adaptations in a revolutionary arms race with the cuckoo, Cuculus canorus. Behav. Ecol. 6 , 166-174. (10.1093/beheco/6.2.166) 16. Kilner RM. 2006 The evolution of egg colour and patterning in birds. Biol. Rev. 81 , 383-406. (10.1017/S1464793106007044)16740199 17. Moskát C, Szentpéteri J, Barta Z. 2002 Adaptations by great reed warblers to brood parasitism: a comparison of populations in sympatry and allopatry with the common cuckoo. Behaviour 139 , 1313-1329. (10.1163/156853902321104181) 18. Avilés JM, Møller AP. 2003 Meadow pipit (Anthus pratensis) egg appearance in cuckoo (Cuculus canorus) sympatric and allopatric populations. Biol. J. Linn. Soc. 79 , 543-549. (10.1046/j.1095-8312.2003.00208.x) 19. Soler JJ, Møller AP. 1996 A comparative analysis of the evolution of variation in appearance of eggs of European passerines in relation to brood parasitism. Behav. Ecol. 7 , 89-94. (10.1093/beheco/7.1.89) 20. Stokke BG, Moksnes A, Røskaft E. 2002 Obligate brood parasites as selective agents for evolution of egg appearance in passerine birds. Evolution 56 , 199-205. (10.1111/j.0014-3820.2002.tb00861.x)11913664 21. Abernathy VE, Peer BD. 2014 Intraclutch variation in egg appearance of brown-headed cowbird hosts. Auk 131 , 467-475. (10.1642/AUK-12-186-R1.1) 22. Avilés JM, Soler JJ, Soler M, Møller AP. 2004 Rejection of parasitic eggs in relation to egg appearance in magpies. Anim. Behav. 67 , 951-958. (10.1016/j.anbehav.2003.08.022) 23. Lovászi P, Moskát C. 2004 Break-down of arms race between the red-backed shrike (Lanius collurio) and common cuckoo (Cuculus canorus). Behaviour 141 , 245-262. (10.1163/156853904322890843) 24. Procházka P, Honza M. 2004 Egg discrimination in the yellowhammer. Condor 106 , 405-410. (10.1093/condor/106.2.405) 25. Stokke G, Rudolfsen G, Moksnes A, Røskaft E. 2004 Rejection of conspecific eggs in chaffinches: the effect of age and clutch characteristics. Ethology 470 , 459-470. (10.1111/j.1439-0310.2004.00978.x) 26. Cherry MI, Bennett ATD, Moskát C. 2007 Host intra-clutch variation, cuckoo egg matching and egg rejection by great reed warblers. Naturwissenschaften 94 , 441-447. (10.1007/s00114-007-0216-4)17252240 27. Polačiková L, Honza M, Procházka P, Topercer J, Stokke BG. 2007 Colour characteristics of the blunt egg pole: cues for recognition of parasitic eggs as revealed by reflectance spectrophotometry. Anim. Behav. 74 , 419-427. (10.1016/j.anbehav.2006.10.023) 28. Yang C, Liu Y, Liang W. 2015 Egg trait variation in a large hawk-cuckoo (Hierococcyx sparverioides) host population of Chinese babax (Babax lanceolatus). Integr. Zool. 10 , 295-301. (10.1111/1749-4877.12128)25664780 29. Spottiswoode CN, Stevens M. 2011 How to evade a coevolving brood parasite: egg discrimination versus egg variability as host defences. Proc. R. Soc. B 278 , 3566-3573. (10.1098/rspb.2011.0401) 30. Stoddard MC, Kilner RM, Town C. 2014 Pattern recognition algorithm reveals how birds evolve individual egg pattern signatures. Nat. Commun. 5 , 4117. (10.1038/ncomms5117)24939367 31. Medina I, Troscianko J, Stevens M, Langmore NE. 2016 Brood parasitism is linked to egg pattern diversity within and among species of Australian passerines. Am. Nat. 187 , 351-362. (10.1086/684627)26913947 32. Britton NF, Planqué R, Franks NR. 2007 Evolution of defence portfolios in exploiter-victim systems. Bull. Math. Biol. 69 , 957-988. (10.1007/s11538-006-9178-5)17265120 33. Caves EM, Stevens M, Iversen ES, Spottiswoode CN. 2015 Hosts of avian brood parasites have evolved egg signatures with elevated information content. Proc. R. Soc. B 282 , 20150598. (10.1098/rspb.2015.0598) 34. Vernon CJ. 1964 The breeding of the cuckoo-weaver (Anomalospiza imberbis (Cabanis)) in Southern Rhodesia. Ostrich 35 , 2-5. (10.1080/00306525.1964.9639425) 35. Colebrook-Robjent JFR. 1984 The breeding of the Didric cuckoo Chrysococcyx caprius in Zambia. In Proc. 5th Pan-African Ornithological Congress (ed. J Ledger), pp. 763-777. Pretoria, South Africa: Southern African Ornithological Society. 36. Sorenson MD, Payne RB. 2001 A single ancient origin of brood parasitism in African finches: implications for host-parasite coevolution. Evolution 55 , 2550-2567. (10.1111/j.0014-3820.2001.tb00768.x)11831669 37. Sorenson MD, Payne RB. 2005 A molecular genetic analysis of the cuckoo phylogeny. In The Cuckoos (ed. RB Payne), pp. 68-94. Oxford, UK: OUP. 38. Caves EM, Stevens M, Spottiswoode CN. 2017 Does coevolution with a shared parasite drive hosts to partition their defences among species? Proc. R. Soc. B 284 , 20170272. (10.1098/rspb.2017.0272) 39. Jensen RAC, Vernon CJ. 1970 On the biology of the Didric cuckoo in Southern Africa. Ostrich 41 , 237-246. (10.1080/00306525.1970.9634377) 40. Spottiswoode CN, Stevens M. 2010 Visual modeling shows that avian host parents use multiple visual cues in rejecting parasitic eggs. Proc. Natl Acad. Sci. USA 107 , 8672-8676. (10.1073/pnas.0910486107)20421497 41. Hockey PAR, Dean WRJ, Ryan PG. 2005 Birds of Southern Africa, 7th edn. Cape Town, South Africa: The Trustees of the John Voelcker Bird Book Fund. 42. Hart NS, Partridge JC, Cuthill IC, Bennett ATD. 2000 Visual pigments, oil droplets, ocular media and cone photoreceptor distribution in two species of passerine bird: the blue tit (Parus caeruleus L.) and the blackbird (Turdus merula L.). J. Comp. Physiol. A 186 , 375-387. (10.1007/s003590050437)10798725 43. Barbosa A, Mäthger LM, Buresch KC, Kelly J, Chubb C, Chiao C-C, Hanlon RT. 2008 Cuttlefish camouflage: the effects of substrate contrast and size in evoking uniform, mottle or disruptive body patterns. Vision Res. 48 , 1242-1253. (10.1016/j.visres.2008.02.011)18395241 44. Stoddard MC, Stevens M. 2010 Pattern mimicry of host eggs by the common cuckoo, as seen through a bird's eye. Proc. R. Soc. B 277 , 1387-1393. (10.1098/rspb.2009.2018) 45. Beecher MD. 1989 Signalling systems for individual recognition: an information theory approach. Anim. Behav. 38 , 248-261. (10.1016/S0003-3472(89)80087-9) 46. Shannon CE, Weaver W. 1949 A mathematical theory of communication, p. 117. Champaign, IL: University of Illinois Press. 47. R Development Core Team. 2019 R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. See https://www.R-project.org/. 48. Cook RD. 1977 Detection of influential observation in linear regression. Technometrics 19 , 15-18. 49. Revell LJ. 2012 phytools: an R package for phylogenetic comparative biology (and other things). Methods Ecol. Evol. 3 , 217-223. (10.1111/j.2041-210X.2011.00169.x) 50. Pagel M. 1999 Inferring the historical patterns of biological evolution. Nature 401 , 877-884. (10.1038/44766)10553904 51. Jetz W, Thomas GH, Joy JB, Hartmann K, Mooers AO. 2012 The global diversity of birds in space and time. Nature 491 , 444-448. (10.1038/nature11631)23123857 52. Orme D, Freckleton R, Thomas G, Petzoldt T, Fritz S, Isaac N, Pearse W. 2013 caper: comparative analyses of phylogenetics and evolution in R. See caper.r-forge.r-project.org. 53. Harvey PH, Pagel MD. 1991 The comparative method in evolutionary biology. Oxford, UK: Oxford University Press. (10.1016/0169-5347(92)90117-T) 54. Montevecchi WA. 1976 Field experiments on the adaptive significance of avian eggshell pigmentation. Behaviour 58 , 26-39. (10.1163/156853976X00226) 55. Stoddard MC, Marshall KLA, Kilner RM. 2011 Imperfectly camouflaged avian eggs: artefact or adaptation? Avian Biol. Res. 4 , 196-213. (10.3184/175815511X13207484398647) 56. Lahti DC, Ardia DR. 2016 Shedding light on bird egg color: pigment as parasol and the dark car effect. Am. Nat. 187 , 547-563. (10.1086/685780)27104989 57. Wisocki PA, Kennelly P, Rivera IR, Cassey P, Burkey ML, Hanley D. 2020 The global distribution of avian eggshell colours suggest a thermoregulatory benefit of darker pigmentation. Nat. Ecol. Evol. 4 , 148-155. (10.1038/s41559-019-1003-2)31659308 58. Eagle A. 2005 Randomness is unpredictability. Br. J. Phil. Sci. 56 , 749-790. (10.1093/bjps/axi138) 59. Gosler AG, Barnett PR, Reynolds SJ. 2000 Inheritance and variation in eggshell patterning in the great tit Parus major. Proc. R. Soc. B Biol. Sci. 267 , 2469-2473. (10.1098/rspb.2000.1307) 60. Caves E, Dixit T, Colebrook-Robjent JFR, Hamusikili L, Stevens M, Thorogood R, Spottiswoode CN. 2021 Data and code for ‘Hosts elevate either within-clutch consistency or between-clutch distinctiveness of egg phenotypes in defence against brood parasites'. Dryad Digital Repository. (10.5061/dryad.02v6wwq34).
PMC008xxxxxx/PMC8242826.txt
==== Front Proc Biol Sci Proc Biol Sci RSPB royprsb Proceedings of the Royal Society B: Biological Sciences 0962-8452 1471-2954 The Royal Society 34187195 10.1098/rspb.2021.0785 rspb20210785 10016960Ecology Research Articles Supplemental forage ameliorates the negative impact of insecticides on bumblebees in a pollinator-dependent crop Supplemental forage ameliorates the negative impact of insecticides on bumblebees in a pollinator-dependent crop Ingwell Laura L. Conceptualization Data curation Investigation Methodology Project administration Writing-original draft Writing-review & editing 1 † Ternest John J. Conceptualization Data curation Formal analysis Investigation Methodology Project administration Resources Software Visualization Writing-original draft Writing-review & editing 1 2 † Pecenka Jacob R. Investigation Methodology 1 http://orcid.org/0000-0003-4469-2750 Kaplan Ian Conceptualization Data curation Formal analysis Funding acquisition Investigation Project administration Supervision Visualization Writing-original draft Writing-review & editing [email protected] 1 1 Department of Entomology, Purdue University, 901 West State Street, West Lafayette, IN, USA 2 Department of Entomology and Nematology, University of Florida, 1881 Natural Area Drive, Gainesville, FL, USA † These authors contributed equally to this study. Electronic supplementary material is available online https://doi.org/10.6084/m9.figshare.c.5469506. 30 6 2021 June 30, 2021 30 6 2021 June 30, 2021 30 6 2021 June 30, 2021 288 1953 2021078525 4 2021 April 25, 2021 4 6 2021 June 4, 2021 © 2021 The Authors. 2021 https://creativecommons.org/licenses/by/4.0/ Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited. Insecticide use and insufficient forage are two of the leading stressors to pollinators in agroecosystems. While these factors have been well studied individually, the experimental designs do not reflect real-world conditions where insecticide exposure and lack of forage occur simultaneously and could interactively suppress pollinator health. Using outdoor enclosures, we tested the effects of insecticides (imidacloprid + lambda-cyhalothrin) and non-crop forage (clover) in a factorial design, measuring the survival, behaviour and performance of bumblebees (Bombus impatiens), as well as pollination of the focal crop, watermelon. Colony survival was synergistically reduced to 17% in watermelon alone + insecticides (survival was 100% in all other treatments). However, behavioural shifts in foraging were mainly owing to insecticides (e.g. 95% reduced visitation rate to watermelon flowers), while impacts on hive performance were primarily driven by clover presence (e.g. 374% increase in the number of live eggs). Insecticide-mediated reductions in foraging decreased crop pollination (fruit set) by 32%. Altogether, these data indicate that both insecticides and non-crop forage play integral roles in shaping pollinator health in agricultural landscapes, but the relative importance and interaction of these two factors depend on which aspect of ‘health’ is being considered. agroecosystem , Bombus impatiens , crop pollination , herbicide , insecticide , pollinator health National Institute of Food and Agriculture http://dx.doi.org/10.13039/100005825 2016-51181-25410 cover-dateJune 30, 2021 ==== Body pmc1. Introduction Uncovering the factors causing declines in pollinator health is complicated by the fact that environmental stressors probably act in concert rather than individually. Thus, efforts should be made to experimentally test the leading stressors both alone and in combination to tease apart the main and interactive effects. For bees, most experts agree that two of the primary causes underlying population declines or reductions in performance are insecticide exposure and lack of forage [1]. Insecticidal toxicity is largely driven by the widespread integration of neonicotinoid seed treatments, which has dramatically elevated the toxic load for bees across agricultural landscapes, despite lower overall amounts of product applied [2,3]. The lack of forage is a result of multiple factors surrounding the general phenomenon of agricultural intensification, e.g. conversion of once-diverse prairies into monocultures of one or a few flowering crops; increased herbicide inputs, such as glyphosate and dicamba, owing to the adoption of herbicide-tolerant transgenic crops. Collectively, this process has sterilized landscapes by removing non-crop flowering plants, often with negative outcomes for bees inhabiting these regions [4–7]. The two mechanisms, however, are not independent. Insecticide-mediated toxicity is likely to interact with changes to the flowering plant community in agroecosystems for several reasons. First, non-crop pollen is expected to dilute the higher insecticide load derived from a monofloral crop pollen diet, thereby acting as a toxicity buffer. This assumes that crop pollen contains more toxic insecticides than non-crop pollen, which is not always the case [8,9]. Second, a lack of non-crop flowers can funnel pollinators into crop fields in search of food, increasing spatial overlap with areas targeted for insecticide application. Last, a lack of non-crop flowers reduces pollen quantity and diversity, both of which can make bees more stressed and physiologically vulnerable to insecticide toxicity [10–12]. Regardless of mechanism, the general prediction is that reduced access to non-crop flowers will exacerbate the detrimental effect of insecticides on pollinators. A corollary of this prediction is that managing herbicide regimes, land use and/or wildflower habitat can attenuate the negative impact of insecticides. The presence of natural habitat in the landscape is known to buffer the negative effects of insecticide use [13,14]. However, few studies have experimentally manipulated supplemental forage and insecticide use in tandem (but see [15,16]), and these mostly employ controlled laboratory experiments that do not resemble a natural foraging arena [17,18]. Virtually, all of the existing work in this area also employs artificial feeders that titrate insecticides or nutritional resources via sugar water rather than simulating actual exposure routes and concentrations that a bee encounters while foraging in the field [19]. Furthermore, the interactive effects of non-crop flowers and insecticides, at present, are only studied from the perspective of pollinator health; we know far less about consequences for crop pollination, even though the two, in theory, should go hand in hand (i.e. healthier pollinators are expected to be more effective pollinators). Insecticide studies in general tend to only measure pollinator performance, with few simultaneously quantifying crop pollination in realistic, outdoor environments [20]. Similarly, the impacts of supplemental forage for crop pollination are complicated. Increased availability of non-crop flowers is assumed to benefit pollinator nutrition and thus performance, but this benefit may not extend to crop pollination [21,22]. A widespread concern among growers is that heterospecific pollen will result in stigma clogging, whereby bees deposit wild plant pollen onto crop flowers, reducing yield. Another worry is that supplemental forage is a distraction that will ‘pull’ bees away from the focal crop, reducing visitation rates during periods of peak bloom when they are most needed. The likelihood for these outcomes depends on factors such as phenological synchrony between non-crop and crop plants [23], and the degree to which pollinators overlap in foraging across adjacent habitats, which is high for polylectic managed bees but could be low for oligolectic wild taxa [24]. Here, we use large, semi-field enclosures to evaluate the singular and combined impacts of insecticides and clover presence as a supplemental forage on: (i) the survival, performance and foraging behaviours of managed bumblebees (Bombus impatiens); and (ii) pollination of the focal crop, watermelon (Citrullus lanatus) (electronic supplementary material, figure S1). Bumblebees, while less popular than honeybees, are commonly used as a secondary pollinator on commercial watermelon farms [25]. This species is also highly sensitive to insecticides applied to crops with much of the recent emphasis on systemic neonicotinoids [26–30] and responds favourably to flowering resource availability [31]. We tested the hypothesis that insecticide use has a comparatively stronger detrimental effect on bees and, correspondingly, crop pollination in the absence of non-crop forage. 2. Methods (a) Site description and plant propagation Experiments were conducted at the Meigs Horticultural Farm, part of the Throckmorton Purdue Agricultural Center (TPAC), located in Lafayette, IN, USA. Six high tunnels were used as field cages to control the foraging radius of experimental bumblebee hives. Each tunnel measured 14.6 × 7.9 × 3.7 m; length, width, height (LWH) and was covered in a single layer of plastic allowing 80% light transmission (12-mil ClearSpan™ PolyMax®). Openings were covered with insect exclusion screening (Anti-Insect Netting, 25 Mesh, Greenhouse Megastore, Danville, IL, USA) to keep bumblebees inside the tunnel. In addition, each tunnel was divided in half, lengthwise, using exclusion screens to separate the tunnel into two arenas. There were a total of 12 arenas used in this experiment, each measuring 7.3 × 7.9 × 3.7 m LWH. The focal crop used in this system was watermelon, which is highly reliant on managed bees for pollination, requiring at least 12 bumblebee visits per flower for optimal yield [25,32,33]. We used a diploid pollenizer variety, AcePlus, to optimize flower availability within the limited foraging space. See the electronic supplementary material, Methods for additional detail on crop and experiment management. (b) Experimental design The experiment employed a 2 × 2 factorial design with two levels of insecticide use (±) and two levels of non-crop forage (±), resulting in the following four treatments: (i) untreated watermelon [insecticide (−)/non-crop forage (−)]; (ii) insecticide-treated watermelon [insecticide (+)/non-crop forage (−)]; (iii) untreated watermelon with clover [insecticide (−)/non-crop forage (+)]; and (iv) insecticide-treated watermelon with clover [insecticide (+)/non-crop forage (+)]. Each of these treatments was assigned to one of the 12 caged arenas. The experiment was organized in a split-plot design where insecticide was the main plot factor applied to the whole tunnel and forage was the subplot factor nested within the main plot using bisected tunnels on either side of the exclusion netting (i.e. one half was seeded with forage; the other was not). For a visual depiction, see the electronic supplementary material, figure S2. A pair of adjacent tunnels (east–west orientation) consisted of a single block, replicated three times along a north–south gradient. The experiment was conducted twice (once in May–July 2018 and again in July–September 2018), resulting in a total of six replicates of each treatment combination. In arenas assigned to non-crop forage, we directly seeded (15 kg ha−1; 1–2 cm depth) in October 2017 and April 2018 with a clover mix consisting of 44% yellow blossom sweet clover (Melilotus officinalis), 33% red clover (Trifolium pratense) and 23% Ladino white clover (Trifolium repens) (3-Way Clover Mix; King's AgriSeeds, Ronks, PA, USA). This seed mix is marketed as a pollinator-friendly blend and clovers, in general, are among the most nutritionally important forages for bumblebees [34,35]. In the insecticide (+) treatment, a systemic neonicotinoid was applied as a root drench to the soil at transplant. To do so, watermelon seedlings received imidacloprid (Admire Pro®) at a rate of 730.5 ml ha−1 by watering each seedling with 500 ml of the insecticide solution. This concentration was based on the recommended label rate for Admire Pro used by growers of 511.6–768.9 ml ha−1. In addition, the pyrethroid lambda-cyhalothrin (Warrior II with Zeon Technology® at 140.2 ml ha−1) was applied to the watermelon plants using an electrostatic sprayer four weeks post-transplant. This combination of neonicotinoid at planting, followed by a subsequent foliar pyrethroid spray, is a commonly used insecticide regime targeting striped cucumber beetles (Acalymma vittatum) and other insect pests by watermelon growers in the midwestern USA [36]. (c) Bumblebee performance and crop pollination When watermelon vines began flowering (ca two weeks after transplant), each arena received one bumblebee hive (B. impatiens; Koppert Biological Systems, Howell, MI, USA). Excel research hives were used, which contain a minimum of 70 workers, one queen and an observation lid. Although nectar sacks were installed in the hives, we did not remove the lid and thus bees were unable to access this resource. Similarly, pollen patties were not added to hives. As a result, observed changes in hive survival and performance are a direct reflection of the experimentally imposed treatments and not buffered by supplemental nutrients. We also assumed this would encourage bees to forage on flowers provided in the arenas. Nest covering material was excluded so we could observe hives throughout the experiment. Hives were placed on 25 May and 15 August where they remained for six weeks. Within arenas, hives were positioned in a corner, elevated from the soil on a plastic crate and shaded beneath an umbrella. The entire hive box (nest-box + nectar + outer cardboard box) was weighed immediately prior to placement in arenas and again at the end of the six-week foraging period to estimate weight gain (final − initial weight) as a measure of performance. Hives were also checked daily to track survival over time. They were considered dead when no active bees were observed tending the colony or out foraging on flowers for two consecutive surveys. In the first trial, one hive in the insecticide (−)/non-crop forage (−) treatment was knocked over owing to high wind and died soon after placement (29 May). This replicate was excluded from statistical analyses of all bumblebee and crop variables, resulting in five replicates rather than six. Forager surveys were conducted twice per week for the six-week duration of bumblebee placement. During surveys, observers walked each row in a transect fashion, recording the number of total foragers active and identity of the flower that each was visiting (i.e. watermelon versus clover). At the end of observations, the hive lid was opened and the number of bees working inside the hive was counted. The queen, if visible, was recorded as dead or alive and the presence of honeypots or pollen stores was also noted. Pollination efficiency was evaluated by flagging 15 female watermelon flowers in each arena (11 June in trial 1; 27 August in trial 2) and following those flowers for one week to verify if pollination had occurred. This was determined by measuring the presence and size of the developing fruit (electronic supplementary material, figure S3). Flagged fruits were then tracked through maturity and eventually weighed to calculate individual melon weights; however, fruit yield data were only measured in trial 2. We monitored within-hive behaviour during each trial on 5 July and 12 September. GoPro cameras (Hero 5) were placed in each arena on a tripod focused directly over the observation lid, draped with a white cloth for shade. Videos were analysed for hive activity using EthoVision software. See the electronic supplementary material, Methods for protocol details. At the conclusion of each experiment or upon colony death, nest-boxes were placed in a freezer. In the laboratory, we dissected each hive, recording the number and weight of workers and queens. Weights were calculated by combining all bees from each group (worker or queen), measuring total weight and dividing by the number of individuals, resulting in a single average value per hive. We also recorded: the number of open worker cells and open queen cells (distinguished by size; see the electronic supplementary material, figure S4); number of worker cells filled with nectar; number of live egg cells and number of dead larval cells. These variables are commonly used to estimate bumblebee hive health in response to insecticides and other stressors [30,37]. Eggs were considered dead when they were desiccated and/or black upon visual inspection. Viable eggs are milky in colour, oval-shaped and free of secondary pathogens. (d) Pesticide residue sampling and quantification Soil, clover and watermelon flowers, and bee nest material were collected to measure the residual amount of insecticides present in each arena. Sampling and analytical details are provided in the electronic supplementary material, Methods. (e) Data analysis Survival analysis was performed using the duration of colony survival in days since placement in the foraging arenas. All hives that survived the entire six-week trial period were right-censored. The Kaplan–Meier estimates were used to create survival curves and a log-rank test was performed to compare among the four treatments. For bee foraging data, we only included observations when hives survived across all experimental arenas (i.e. we censored data from dates after hive death began). To avoid pseudoreplication and provide a single value per arena, we summed foraging observations collected over multiple dates. The effects of the two focal treatments on bee foraging behaviour were analysed using negative binomial regression because count data were overdispersed. For watermelon observations, we tested the main and interactive effects of insecticide and non-crop forage; however, for clover, we only included the effect of insecticide in the non-crop forage (+) treatment because there was no clover to observe in the (−) treatment. For both flower types, we also included trial (1 versus 2) and spatial block (high tunnels 1–2, 3–4, 5–6) as predictor variables. For hive activity and colony performance, main and interactive effects of insecticide and non-crop forage were tested for each response. Continuous data were analysed using a normal distribution for final hive weight, while a log(x + 1) transformation was used for queen and worker weights to improve normality. Discrete count data (worker no., worker larvae, worker pupae, worker honeypots, dead larval cells, live eggs) were analysed with a negative binomial or zero-inflated negative binomial model. As with foraging data, we included trial (1 versus 2) and spatial block (high tunnels 1–2, 3–4, 5–6) as predictor variables. We also included initial colony weight as a covariate. To assess fruit set, we performed an arcsine square root transformation on the proportion of flagged flowers that developed into a fruit per experimental arena. Insecticide, non-crop forage, trial and block were used as predictor variables using a general linear model. The mean fruit weight (kg) per melon per tunnel was analysed with a regression model using a lognormal distribution; we did not include trial as a factor as fruit weights were only measured in trial 2. All statistical analyses were conducted using JMP Pro 15 (SAS Institute Inc., Cary, NC, USA). 3. Results (a) Bumblebee survival In both treatments containing clover (insecticide ±), bumblebee hive survival was 100% over the full six-week experiment period, regardless of pesticide application. In the absence of clover, however, survival depended on whether the crop was treated with insecticides; namely, in the insecticide (−) treatment all hives survived, while in the insecticide (+) treatment, survival was reduced to 17% (figure 1; log-rank test, χ32 = 19.8, p = 0.0002). Figure 1. Survival curves with 95% confidence interval showing the persistence of bumblebee hives in the presence/absence of insecticides and alternative non-crop forage in the four treatment combinations. Hives remained in arenas for six weeks and colony survival was assessed daily. Lines at 1.0 were jittered for visibility. Ins., insecticide; NCF, non-crop forage (clover). (Online version in colour.) (b) Bumblebee foraging and in-hive behaviours Insecticides reduced bee visitation to watermelon flowers by 95% compared with untreated control arenas (insecticide (−) versus (+), mean = 28.0 and 1.3 observations arena−1, respectively), but clover presence did not affect watermelon visitation rate (figure 2a and table 1a). Similarly, insecticides caused a 68% reduction in foraging on clover (insecticide (−) versus (+), mean = 52.5 and 16.8 observations arena−1, respectively), despite only being applied to the crop (figure 2b). Importantly, we found no difference in bee foraging rates (p = 0.9040) comparing early observation periods before pyrethroids were applied as a foliar spray in the insecticide (+) treatments (i.e. isolating only the effect of the systemic neonicotinoid at-planting) with later dates after the watermelon was sprayed. Figure 2. Visitation rate of bumblebees to flowers from observational transect walks, including watermelon (a) and clover (b) flowers. Box plots show the median, quartiles, range and individual data points across the four treatments. NCF, non-crop forage (clover). ***p < 0.001. (Online version in colour.) Table 1. Statistical table describing the main effects of insecticide and non-crop forage (clover) on all bumblebee and crop response variables (d.f. = 1 throughout). (Insecticide × NCF denotes statistical interaction between the two main effects. Significant (p < 0.05) and marginally significant (p < 0.07) insecticide and non-crop forage effects are italicized for emphasis. Yellow and red shading denote increases and decreases, respectively, of the response variable to treatments that were significant for main effects (note a few interactions were also significant; left unshaded). Effects of non-treatment-related variables (e.g. block, trial, initial hive weight) are reported in the electronic supplementary material, table S1. EthoVision analysis of video recordings from within-hive activity also revealed a strong negative influence of insecticides with no corresponding impact of clover (figure 3 and table 1b). Bees from insecticide-exposed hives exhibited a 57% reduction in activity (insecticide (−) versus (+), mean = 1.9 and 0.8% activity within observation area, respectively). Figure 3. Within-hive bee activity from EthoVision analysis of video recordings. Insecticide use significantly reduced activity, whereas non-crop forage had no significant effect. GoPro cameras focused directly over the observation lid recorded for at least 20 min with activity (%) quantified by measuring the amount of pixel change in the videos over time. The arena size for each analysis was standardized to a 9.5 × 7 cm area focused over the largest portion of brood in the hive. Box plots show the median, quartiles, range and individual data points across the four treatments. NCF, non-crop forage (clover). *p < 0.05. (Online version in colour.) (c) Bumblebee colony performance Unlike foraging behaviour, colony performance showed the opposite pattern with stronger overall effects of clover than insecticides (table 1c). Insecticide use increased the number of worker honeypots (insecticide (−) versus (+), mean = 0.09 and 19.3 hive−1, respectively), while decreasing the number of live eggs (insecticide (−) versus (+), mean = 22.18 and 10.5, respectively). Clover presence increased worker weight (clover (−) versus (+), mean = 0.92 and 3.23 g, respectively) and number of live eggs (clover (−) versus (+), mean = 5.45 and 25.83, respectively), but decreased the number of worker pupae (clover (−) versus (+), mean = 55.91 and 20.17, respectively) and dead larval cells (clover (−) versus (+), mean = 47.64 and 22.25, respectively) (means for all treatments reported in the electronic supplementary material, table S2). Two variables—queen weight and worker larvae—were affected by the interaction between insecticide use and clover presence. Hive weight and worker count were the only variables unaffected by either factor. Data on number of reproductives (queens and males) were excluded as they were always either zero or one per hive. No new queens were produced by the colonies in this experiment (i.e. in cases where one was found, it was probably the old mother queen originating with the hive). (d) Crop pollination The mean fruit set (proportion of female flowers producing fruit) was 0.31 across all treatments combined (n = 360 flowers); this rate is normal for watermelon, which ranges from 0.2 to 0.4, even under optimal conditions [38]. Insecticides reduced fruit set by 32% (insecticide (−) versus (+), mean = 0.37 and 0.25, respectively) with no corresponding effect owing to clover presence (figure 4 and table 1d). By contrast, individual fruit weights (kg) were unaffected by insecticide use but were reduced by 21% owing to clover presence (clover (−) versus (+), mean ± s.e. = 1.51 ± 0.09 and 1.20 ± 0.04, respectively). Because fruit set was not significantly different, and trended towards being slightly lower, in the presence of clover, changes in this variable could not compensate for the decline in individual fruit weight to affect total fruit weight per arena. Figure 4. Fruit set of tagged watermelon flowers in the four treatments. Insecticide use significantly reduced fruit set, whereas non-crop forage had no significant effect. Fruit set was calculated as the proportion of flowers developing into a fruit per experimental arena (n = 15 flowers arena−1) and used as a measure of pollination. Box plots show the median, quartiles, range and individual data points across the four treatments. NCF, non-crop forage (clover). *p < 0.05. (Online version in colour.) (e) Pesticide residues The neonicotinoid imidacloprid was detected in all matrices in which it was tested, including soil, watermelon pollen, clover pollen and bumblebee nest material (summary data reported in the electronic supplementary material, table S3). In virtually all cases, detection rates increased dramatically with insecticide use; for example, detection in watermelon pollen ranged from 0 to 20% in the insecticide (–) treatment and increased to 92–100% in the insecticide (+) treatment. It is notable that imidacloprid was detected at high rates in clover pollen (100%) and nest material (80–83%) since it was not applied to these areas. However, the mean concentrations were two to three times higher in crop than non-crop pollen. Overall, the highest concentrations were found in the soil (max. 5574 ng g−1), followed by pollen (max. 377 ng g−1), and lowest levels were in nest material (max. 8 ng g−1). The related neonicotinoids, thiamethoxam and clothianidin were also measured in samples, but were mostly below the limit of detection (thiamethoxam: soil—0% samples detected, pollen—3% samples detected (median = 0.99 ng g−1, max. = 1.36 ng g−1); clothianidin: soil—50% samples detected (median = 0.77 ng g−1, max. = 3.59 ng g−1), pollen—1% samples detected (1.46 ng g−1 in a single sample)). 4. Discussion Overall, we found partial support for our hypothesis that insecticides have stronger negative effects on bees and crop pollination in the absence of non-crop forage. The strongest evidence came from hive survival, where insecticide use and lack of forage synergistically increased mortality. It was particularly striking that survival was 100% in all three treatments with zero or one stressor alone, but the combination of two stressors caused survival to plummet to 17%. These data suggest that cultivating or encouraging supplemental flowers in and around crop fields can buffer the negative non-target effects from insecticides [14,16]. In the field, growers could accomplish this by implementing wildflower strips along field borders or adopting a less stringent herbicide regime to encourage flowering weeds. Alternatively, investing in integrated pest management (IPM) programmes or technologies that reduce insecticide use (e.g. pest scouting and action thresholds; host plant resistance) could generate the same outcome. We were unable to pinpoint the mechanism underlying this synergy. However, it is unlikely a behavioural result of ‘forcing’ bees to forage on the insecticide-treated crop owing to a lack of alternatives, because clover presence did not affect the visitation rate on watermelon flowers. Watermelon is not a preferred resource for bumblebees in the field [32,33,39], but in our experiment, they readily visited these flowers even when surrounded by clover, which is considered a high-quality forage whose flowers are open and accessible during the same time of day as watermelon (i.e. temporal coincidence in bloom times). Clover flowers also contained non-trivial amounts of imidacloprid, albeit in lower concentrations than watermelon. This means that dietary insecticide exposure would have occurred, even if bees exclusively foraged from clover. The detection of imidacloprid in neighbouring clover is not surprising, given their close proximity (1–2 m) to treated watermelon in arenas and the propensity for water-soluble neonicotinoids to move laterally in the soil profile. A more likely explanation for the synergistic reduction in bumblebee survival is that co-exposure to both stressors acted at a physiological level to increase vulnerability. This stressor combination may have been exacerbated by the fact that screened tunnels are considerably warmer than ambient conditions owing to reduced ventilation, resulting in heat stress as a likely contributing factor. Moreover, we removed nest covering material to observe hives, which could have further impeded thermoregulation. The temporal pattern of hive death also supports this explanation. Several hives died near-simultaneously within a few days in early September. The 5 days immediately preceding their death experienced the highest daily maximum temperatures of the entire experiment (avg. 41.2°C; for comparison, the two weeks prior to this (16–31 August) had maximum temperatures that were greater than 5° cooler (avg. 35.1°C)). Interestingly, a recent study demonstrated that poor diet quality increases susceptibility of the bumblebee Bombus terrestris to heat waves [40], and bumblebees in general seem to respond poorly to warm environments [41]. Experimentally testing forage/insecticide interactions under a range of temperatures is necessary to tease apart these relationships. In our study, the specific route of insecticide exposure for bees is unknown. Imidacloprid was detected in all matrices tested (soil, pollen, nest material) with highest levels in soil. This is not particularly surprising since the insecticide was applied as a soil drench, rather than a foliar spray, and the product moved systemically through the plant to reach flowers. As a result, soil exposure is likely to pose a threat to ground-nesting wild bees that come in close contact with these residues [42–44]; however, given that bumblebees were housed in aboveground structures in the experiment, we assume that insecticide detected in nest materials derived from oral exposure via collection of contaminated floral resources. The specific concentrations reported for imidacloprid in crop-treated pollen were high (median in watermelon: 75.4 ng g−1) but within the range detected in other studies of neonicotinoids in cucurbits [45–47]. For example, Bloom et al. [45] reported the median thiamethoxam concentration in seed-treated cucumber pollen from commercial fields at 73.7 ng g−1. Similarly, Dively & Kamel [46] found the mean imidacloprid residues in pumpkin pollen at 60.9 ng g−1, using the same soil drench technique and product used here. Because insecticides were applied in covered high tunnels in our study, the residues measured could differ from open-field values. Protected structures prevent rain downpour from leaching pesticides from the soil and the plastic covering filters sunlight, potentially leading to altered pesticide degradation rates. We also detected imidacloprid in untreated watermelon and clover flowers, which could underestimate the effect of insecticides if the control is contaminated. This off-site detection is not surprising because: (i) imidacloprid is among the commonly used insecticides in agricultural areas where the study was conducted; and (ii) our analytical approach using liquid chromatography-mass spectrometry (LC-MS) is highly sensitive to trace amounts that, while detectable, may not be biologically relevant. For context, the median value of imidacloprid in treated flowers was ca 250 and 50 times higher than untreated flowers for watermelon and clover, respectively. Thus, we are confident that bees in the untreated arenas were not inadvertently exposed to high levels of imidacloprid. An additional factor strongly affecting our interpretation of forage/insecticide relationships is the response variable measured. While survival clearly showed a synergistic reduction, the other variables were dominated by main effects of either forage availability or insecticides. These main effects were remarkably consistent with short-term behaviours, such as in-hive activity and flower visitation, negatively affected by insecticides. Reductions to in-hive activity were expected based on the fact that imidacloprid and other neonicotinoids are neurotoxins that impede motor function in bees [48]. Despite the dramatically lower in-hive movements and overall poor health, these colonies were still alive after insecticide exposure when behaviours were monitored. Unexpectedly, insecticide use also led to an increased number of honeypots. Bumblebee production of honeypots is a complex response that integrates resource quality with colony-level regulations and feedbacks [49,50]. Given that bumblebees cannot taste neonicotinoids and, in some conditions, prefer imidacloprid-laced nectar [51], the high concentrations in our flowers could have led to preferential nectar (versus pollen) foraging, even though these are complementary rather than substitutable resources. Alternatively, bees may shift to relatively more nectar-foraging as an avoidance response because neonicotinoids tend to occur at much lower concentrations in nectar than pollen [52]. Because we only measured floral visitation, potential shifts in nectar versus pollen collection across treatments are unknown. Insecticides suppressed watermelon flower visitation by greater than 90%, which ultimately led to lower fruit set. This occurred in spite of using a relatively high stocking rate in our arenas, which contained one hive for a 58 m2 area. For comparison, recommendations for commercial vegetable production are one to three bumblebee hives per 1000 m2 of enclosed space for indoor pollination. This means that bumblebees in our arenas were at least five times higher than the recommended stocking rate for optimal pollination. Furthermore, watermelon requires relatively few visits from bumblebees for successful pollination [25,32,33]. As a result, strong insecticide-mediated reductions in crop foraging appeared to offset the fact that efficient pollinators were present at such high densities. It is also important to note that clover benefited bumblebee performance compared to a monofloral diet of the crop alone (e.g. increasing live eggs by nearly 400%) without reducing watermelon visitation or fruit set. The average fruit weight was lower when bordered by clover; however, we strongly suspect this is not a pollination-related mechanism but rather a consequence of the unusually close proximity and large size of the clover, resulting in resource competition with the crop (see the picture in the electronic supplementary material, figure S2 for context). This means that non-crop forage as a habitat manipulation can achieve the goal of enhancing pollinator health without detracting from crop pollination by ‘pulling’ bees away from the crop or clogging stigmas with non-crop pollen. This conclusion is noteworthy and probably extends to more realistic field conditions, given the nature of our experimental set-up. Inside arenas, watermelon and clover were at roughly equal proportions and directly neighbouring one another (i.e. clover was cultivated as a uniform ground cover between each crop row). This is an extreme case compared to a typical agricultural field where the crop/non-crop flower ratio would be far more skewed towards crop dominance, even in diverse systems, and most of the non-crop forage would probably be relegated to field edges rather than growing in such close proximity to the crop itself. Because clover did not interfere with crop pollination under such extreme conditions, it seems highly unlikely in open-field settings that wildflowers would be anything aside from beneficial. More work is needed to understand why clover or other non-crop flowers benefit bees. In this case, it remains unclear whether the beneficial properties derive from higher resource quantity (i.e. twice the number of flowers), quality (i.e. clover pollen is considered high quality with protein levels ranging from 22 to 25%; [53]) or diversity (i.e. one versus several species of pollen). These three features are confounded in our design and thus we cannot disentangle them from one another. However, a recent study using a novel stoichiometric approach to evaluate bee diets based on elemental ratios in pollen, highlighted clover—out of ca 100 taxa considered—as a nutritionally well-balanced species [54]. Cucurbit pollen, on the other hand, is detrimental for B. impatiens performance [39], and, therefore, clover may simply dilute the effects of a suboptimal diet consisting of watermelon alone. 5. Conclusion As a whole, our data indicate that both insecticide use and non-crop flower presence contribute strongly to pollinator health in agroecosystems. An important caveat to this conclusion is that our experimental set-up represents an extreme test of what bees might face under real world conditions. For instance, non-crop flowers (e.g. unmanaged habitats, border vegetation) are typically within the flight range, particularly for large taxa like bumblebees or honeybees, and thus, it would be rare for foragers to be unable to access any floral resource aside from the crop itself. Similarly, insecticide use is usually not an all-or-nothing endeavour; growers vary along a continuum of application frequency and product toxicity. Thus, our experiment should be viewed as a proof-of-concept for the relative importance and interactions among these common stressors, rather than a true estimate of the magnitude of effects experienced in the field. It is difficult to make broad statements about the relative importance of these factors and the potential for ecological interactions between the two. These conclusions are shaped by the specific aspects of ‘health’ that are measured or targeted for enhancement. Given that hive-scale survival is critical to bumblebees fulfilling their functional role as crop pollinators, we view the synergistic increase in mortality to be an overriding outcome from this work. The challenge will be to find opportunities for reducing pesticide inputs while maintaining farm productivity. This seems possible both in watermelon [36] and farming systems more generally [55] using IPM and related approaches aimed at increasing sustainability. Achieving this balance is particularly critical in pollinator-dependent crops, like watermelon, where pest management and pollinator conservation are central, but sometimes conflicting, goals. Finally, it should be noted that, while bumblebee hives were used in this study as a tool for measuring pollinator responses, the broader implications of these findings extend beyond managed species to inform the conservation of wild bees on or near agricultural lands. Managed bumblebee hives are commonly used as convenient proxies for wild taxa. Compared to bumblebees, however, many wild bees tend to have a smaller foraging radius and more limited diet, making the experimental design perhaps even more relevant to these functionally and taxonomically related species. Extending stressor interaction studies to wild bees will be a critical step in understanding the factor(s) contributing to their persistence and diversity in agroecosystems. Supplementary Material Click here for additional data file. Acknowledgements We thank Amber Jannasch and Yu Han-Hallet from the Purdue Bindley Bioscience Center for their assistance in the LC–MS pesticide residue quantification; Danielle Madison and Natalie Eason for helping with data collection and plot maintenance; and Allison Bistline-East, Wadih Ghanem, Ashley Leach, Christie Shee and Emily Tronson for helpful edits on later drafts of the manuscript. Data accessibility Raw data are available to download via the Purdue University Research Repository at: https://purr.purdue.edu/publications/3754/1 [56]. Authors' contributions L.L.I.: conceptualization, data curation, investigation, methodology, project administration, writing—original draft, writing—review and editing; J.J.T.: conceptualization, data curation, formal analysis, investigation, methodology, project administration, resources, software, visualization, writing—original draft, writing—review and editing; J.R.P.: investigation, methodology; I.K.: conceptualization, data curation, formal analysis, funding acquisition, investigation, project administration, supervision, visualization, writing—original draft, writing—review and editing. All authors gave final approval for publication and agreed to be held accountable for the work performed therein. Competing interests The authors declare we have no competing interests. Funding This research was funded by a USDA NIFA grant (no. 2016-51181-25410) to I.K. ==== Refs References 1. Goulson D, Nicholls E, Botías C, Rotheray EL. 2015 Bee declines driven by combined stress from parasites, pesticides, and lack of flowers. Science 347 , 1255957. (10.1126/science.1255957)25721506 2. Douglas MR, Sponsler DB, Lonsdorf EV, Grozinger CM. 2020 County-level analysis reveals a rapidly shifting landscape of insecticide hazard to honey bees (Apis mellifera) on US farmland. Sci. Rep. 10 , 797. (10.1038/s41598-019-57225-w)31964921 3. DiBartolomeis M, Kegley S, Mineau P, Radford R, Klein K. 2019 An assessment of acute insecticide toxicity loading (AITL) of chemical pesticides used on agricultural land in the United States. PLoS ONE 14 , e0220029. (10.1371/journal.pone.0220029)31386666 4. Naug D. 2009 Nutritional stress due to habitat loss may explain recent honeybee colony collapses. Biol. Conserv. 142 , 2369-2372. (10.1016/j.biocon.2009.04.007) 5. Dolezal AG, St. Clair AL, Zhang G, Toth AL, O'Neal ME. 2019 Native habitat mitigates feast–famine conditions faced by honey bees in an agricultural landscape. Proc. Natl Acad. Sci. USA 116 , 25 147-25 155. (10.1073/pnas.1912801116) 6. Alaux C et al. 2017 A ‘landscape physiology’ approach for assessing bee health highlights the benefits of floral landscape enrichment and semi-natural habitats. Sci. Rep. 7 , 40568. (10.1038/srep40568)28084452 7. Smart MD, Pettis JS, Euliss N, Spivak MS. 2016 Land use in the Northern Great Plains region of the U.S. influences the survival and productivity of honey bee colonies. Agric. Ecosyst. Environ. 230 , 139-149. (10.1016/j.agee.2016.05.030) 8. Long EY, Krupke CH. 2016 Non-cultivated plants present a season-long route of pesticide exposure for honey bees. Nat. Commun. 7 , 11629. (10.1038/ncomms11629)27240870 9. Wood TJ, Kaplan I, Zhang Y, Szendrei Z. 2019 Honeybee dietary neonicotinoid exposure is associated with pollen collection from agricultural weeds. Proc. R. Soc. B 286 , 20190989. (10.1098/rspb.2019.0989) 10. Wahl O, Ulm K. 1983 Influence of pollen feeding and physiological condition on pesticide sensitivity of the honey bee Apis mellifera carnica. Oecologia 59 , 106-128. (10.1007/bf00388082)25024157 11. Huang Z. 2012 Pollen nutrition affects honey bee stress resistance. Terrestr. Arthrop. Rev. 5 , 175-189. (10.1163/187498312X639568) 12. Schmehl DR, Teal PEA, Frazier JL, Grozinger CM. 2014 Genomic analysis of the interaction between pesticide exposure and nutrition in honey bees (Apis mellifera). J. Insect Physiol. 71 , 177-190. (10.1016/j.jinsphys.2014.10.002)25450567 13. Kremen C, Williams NM, Thorp RW. 2002 Crop pollination from native bees at risk from agricultural intensification. Proc. Natl Acad. Sci. USA 99 , 16 812-16 816. (10.1073/pnas.262413599) 14. Park MG, Blitzer EJ, Gibbs J, Losey JE, Danforth BN. 2015 Negative effects of pesticides on wild bee communities can be buffered by landscape context. Proc. R. Soc. B 282 , 20150299. (10.1098/rspb.2015.0299) 15. Stuligross C, Williams NM. 2020 Pesticide and resource stressors additively impair wild bee reproduction. Proc. R. Soc. B 287 , 20201390. (10.1098/rspb.2020.1390) 16. Klaus F, Tscharntke T, Bischoff G, Grass I. 2021 Floral resource diversification promotes solitary bee reproduction and may offset insecticide effects—evidence from a semi-field experiment. Ecol. Lett. 24 , 668-675. (10.1111/ele.13683)33524201 17. Barraud A, Vanderplanck M, Nadarajah S, Michez D. 2020 The impact of pollen quality on the sensitivity of bumblebees to pesticides. Acta Oecolog. 105 , 103552. (10.1016/j.actao.2020.103552) 18. Leza M, Watrous KM, Bratu J, Woodard SH. 2018 Effects of neonicotinoid insecticide exposure and monofloral diet on nest-founding bumblebee queens. Proc. R. Soc. B 285 , 20180761. (10.1098/rspb.2018.0761) 19. Tosi S, Nieh JC, Sgolastra F, Cabbri R, Medrzycki P. 2017 Neonicotinoid pesticides and nutritional stress synergistically reduce survival in honey bees. Proc. R. Soc. B 284 , 20171711. (10.1098/rspb.2017.1711) 20. Stanley DA, Garratt MPD, Wickens JB, Wickens VJ, Potts SG, Raine NE. 2015 Neonicotinoid pesticide exposure impairs crop pollination services provided by bumblebees. Nature 528 , 548-550. (10.1038/nature16167)26580009 21. Feltham H, Park K, Minderman J, Goulson D. 2015 Experimental evidence that wildflower strips increase pollinator visits to crops. Ecol. Evol. 5 , 3523-3530. (10.1002/ece3.1444)26380683 22. Zamorano J, Bartomeus I, Grez AA, Garibaldi LA. 2020 Field margin floral enhancements increase pollinator diversity at the field edge but show no consistent spillover into the crop field: a meta-analysis. Insect Conserv. Divers. 13 , 519-531. (10.1111/icad.12454) 23. Wood TJ, Gibbs J, Rothwell N, Wilson JK, Gut L, Brokaw J, Isaacs R. 2018 Limited phenological and dietary overlap between bee communities in spring flowering crops and herbaceous enhancements. Ecol. Appl. 28 , 1924-1934. (10.1002/eap.1789)30184292 24. Nicholson CC et al. 2020 Mismatched outcomes for biodiversity and ecosystem services: testing the responses of crop pollinators and wild bee biodiversity to habitat enhancement. Ecol. Lett. 23 , 326-335. (10.1111/ele.13435)31797535 25. Stanghellini MS, Ambrose JT, Schultheis JR. 1998 Using commercial bumble bee colonies as backup pollinators for honey bees to produce cucumbers and watermelons. HortTech. 8 , 590-594. (10.21273/HORTTECH.8.4.590) 26. Arce AN, David TI, Randall EL, Rodrigues AR, Colgan TJ, Wurm Y, Gill RJ. 2017 Impact of controlled neonicotinoid exposure on bumblebees in a realistic field setting. J. Appl. Ecol. 54 , 1199-1208. (10.1111/1365-2664.12792) 27. Baron GL, Jansen VAA, Brown MJF, Raine NE. 2017 Pesticide reduces bumblebee colony initiation and increases probability of population extinction. Nat. Ecol. Evol. 1 , 1308-1316. (10.1038/s41559-017-0260-1)29046553 28. Laycock I, Cotterell KC, O'Shea-Wheller TA, Cresswell JE. 2014 Effects of the neonicotinoid pesticide thiamethoxam at field-realistic levels on microcolonies of Bombus terrestris worker bumble bees. Ecotox. Environ. Saf. 100 , 153-158. (10.1016/j.ecoenv.2013.10.027) 29. Stanley DA, Raine NE. 2016 Chronic exposure to a neonicotinoid pesticide alters the interactions between bumblebees and wild plants. Funct. Ecol. 30 , 1132-1139. (10.1111/1365-2435.12644)27512241 30. Whitehorn PR, O'Connor S, Wackers FL, Goulson D. 2012 Neonicotinoid pesticide reduces bumble bee colony growth and queen production. Science 336 , 351-352. (10.1126/science.1215025)22461500 31. Spiesman BJ, Bennett A, Isaacs R, Gratton C. 2017 Bumble bee colony growth and reproduction depend on local flower dominance and natural habitat area in the surrounding landscape. Biol. Conserv. 206 , 217-223. (10.1016/j.biocon.2016.12.008) 32. Campbell JW, Stanley-Stahr C, Bammer M, Daniels JC, Ellis JD. 2019 Contribution of bees and other pollinators to watermelon (Citrullus lanatus Thunb.) pollination. J. Apicult. Res. 58 , 597-603. (10.1080/00218839.2019.1614271) 33. Campbell JW, Daniels JC, Ellis JD. 2018 Fruit set and single visit stigma pollen deposition by managed bumble bees and wild bees in Citrullus lanatus (Cucurbitales: Cucurbitaceae). J. Econ. Entomol. 111 , 989-992. (10.1093/jee/toy008)29385500 34. Carvell C, Roy DB, Smart SM, Pywell RF, Preston CD, Goulson D. 2006 Declines in forage availability for bumblebees at a national scale. Biol. Conserv. 132 , 481-489. (10.1016/j.biocon.2006.05.008) 35. Goulson D, Hanley ME, Darvill B, Ellis JS, Knight ME. 2005 Causes of rarity in bumblebees. Biol. Conserv. 122 , 1-8. (10.1016/j.biocon.2004.06.017) 36. Ternest JJ, Ingwell LL, Foster RE, Kaplan I. 2020 Comparing prophylactic versus threshold-based insecticide programs for striped cucumber beetle (Coleoptera: Chrysomelidae) management in watermelon. J. Econ. Entomol. 113 , 872-881. (10.1093/jee/toz346)31901943 37. Ellis C, Park KJ, Whitehorn P, David A, Goulson D. 2017 The neonicotinoid insecticide thiacloprid impacts upon bumblebee colony development under field conditions. Environ. Sci. Technol. 51 , 1727-1732. (10.1021/acs.est.6b04791)28079366 38. Adlerz WC. 1966 Honey bee visit numbers and watermelon pollination. J. Econ. Entomol. 59 , 28-30. (10.1093/jee/59.1.28) 39. Brochu KK, van Dyke MT, Milano NJ, Petersen JD, McArt SH, Nault BA, Kessler A, Danforth BN. 2020 Pollen defenses negatively impact foraging and fitness in a generalist bee (Bombus impatiens: Apidae). Sci. Rep. 10 , 3112. (10.1038/s41598-020-58274-2)32080216 40. Vanderplanck M, Martinet B, Carvalheiro LG, Rasmont P, Barraud A, Renaudeau C, Michez D. 2019 Ensuring access to high-quality resources reduces the impacts of heat stress on bees. Sci. Rep. 9 , 12596. (10.1038/s41598-019-49025-z)31467366 41. Rasmont P, Iserbyt S. 2012 The bumblebees scarcity syndrome: are heat waves leading to local extinctions of bumblebees (Hymenoptera: Apidae: Bombus)? Ann. Soc. Entomol. Fr. 48 , 275-280. (10.1080/00379271.2012.10697776) 42. Willis CD, Raine NE. 2021 Population decline in a ground-nesting solitary squash bee (Eucera pruinosa) following exposure to a neonicotinoid insecticide treated crop (Cucurbita pepo). Sci. Rep. 11 , 4241. (10.1038/s41598-021-83341-7)33608633 43. Anderson NL, Harmon-Threatt AN. 2019 Chronic contact with realistic soil concentrations of imidacloprid affects the mass, immature development speed, and adult longevity of solitary bees. Sci. Rep. 9 , 3724. (10.1038/s41598-019-40031-9)30842465 44. Willis CDS et al . 2019 Assessment of risk to hoary squash bees (Peponapis pruinosa) and other ground-nesting bees from systemic insecticides in agricultural soil. Sci. Rep. 9 , 11870. (10.1038/s41598-019-47805-1)31413274 45. Bloom EH, Wood TJ, Hung KLJ, Ternest JJ, Ingwell LL, Goodell K, Kaplan I, Szendrei Z. 2021 Synergism between local- and landscape-level pesticides reduces wild bee floral visitation in pollinator-dependent crops. J. Appl. Ecol. 58 , 1187-1198. 46. Dively GP, Kamel A. 2012 Insecticide residues in pollen and nectar of a cucurbit crop and their potential exposure to pollinators. J. Agric. Food Chem. 60 , 4449-4456. (10.1021/jf205393x)22452667 47. Stoner KA, Eitzer BD. 2012 Movement of soil-applied imidacloprid and thiamethoxam into nectar and pollen of squash (Cucurbita pepo). PLoS ONE 7 , e39114. (10.1371/journal.pone.0039114)22761727 48. Williamson SM, Willis SJ, Wright GA. 2014 Exposure to neonicotinoids influences the motor function of adult worker honeybees. Ecotoxicology 23 , 1409-1418. (10.1007/s10646-014-1283-x)25011924 49. McAulay MK, Otis GW, Gradish AE. 2015 Honeypot visitation enables scent learning and heightens forager response in bumblebees (Bombus impatiens). Learn. Motivat. 50 , 22-31. (10.1016/j.lmot.2014.11.001) 50. Dornhaus A, Chittka L. 2005 Bumble bees (Bombus terrestris) store both food and information in honeypots. Behav. Ecol. 16 , 661-666. (10.1093/beheco/ari040) 51. Kessler S et al. 2015 Bees prefer foods containing neonicotinoid pesticides. Nature 521 , 74-76. (10.1038/nature14414)25901684 52. Goulson D. 2013 An overview of the environmental risks posed by neonicotinoid insecticides. J. Appl. Ecol. 50 , 977-987. (10.1111/1365-2664.12111) 53. Pernal SF, Currie RW. 2001 The influence of pollen quality on foraging behavior in honeybees (Apis mellifera L.). Behav. Ecol. Sociobiol. 51 , 53-68. (10.1007/s002650100412) 54. Filipiak M et al . 2017 Ecological stoichiometry of the honeybee: pollen diversity and adequate species composition are needed to mitigate limitations imposed on the growth and development of bees by pollen quality. PLoS ONE 12 , e0183236. (10.1371/journal.pone.0183236)28829793 55. Lechenet M, Dessaint F, Py G, Makowski D, Munier-Jolain N. 2017 Reducing pesticide use while preserving crop productivity and profitability on arable farms. Nat. Plants 3 , 17008. (10.1038/nplants.2017.8)28248316 56. Kaplan I. 2021 Ingwell et al. data for high tunnel bee study 2018. Purdue University Research Repository. (10.4231/86ZT-Y828)
PMC008xxxxxx/PMC8242831.txt
==== Front Proc Biol Sci Proc Biol Sci RSPB royprsb Proceedings of the Royal Society B: Biological Sciences 0962-8452 1471-2954 The Royal Society 34187194 10.1098/rspb.2021.0787 rspb20210787 10016014Ecology Research Articles Self and non-self recognition affects clonal reproduction and competition in the pea aphid Self and non-self recognition affects clonal reproduction and competition in the pea aphid http://orcid.org/0000-0001-7955-6009 Li Yang Formal analysis Investigation Writing-original draft 1 2 http://orcid.org/0000-0002-7038-9678 Akimoto Shin-ichi Conceptualization Data curation Formal analysis Funding acquisition Investigation Project administration Visualization Writing-original draft Writing-review & editing [email protected] 1 1 Department of Ecology and Systematics, Graduate School of Agriculture, Hokkaido University, Sapporo 060-8589, Japan 2 College of Biology and Agriculture, Zunyi Normal University, Zunyi Guizhou 563006, People's Republic of China Electronic supplementary material is available online at https://doi.org/10.6084/m9.figshare.c.5469509. 30 6 2021 June 30, 2021 30 6 2021 June 30, 2021 30 6 2021 June 30, 2021 288 1953 202107873 4 2021 April 3, 2021 4 6 2021 June 4, 2021 © 2021 The Authors. 2021 https://creativecommons.org/licenses/by/4.0/ Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited. The spatial interaction of clonal organisms is an unsolved but crucial topic in evolutionary biology. We evaluated the interactions between aphid clones using a colour mutant (yellow) and an original (green) clone. Colonies founded by two aphids of the same clone and mixed colonies, founded by a green aphid and a yellow aphid, were set up to observe population growth for 15 days. We confirmed positive competition effects, with mixed colonies increasing in size more rapidly than clonal colonies. In mixed colonies where reproduction started simultaneously, green aphids overwhelmed yellow aphids in number, and yellow aphids restrained reproduction. However, when yellow aphids started to reproduce earlier, they outnumbered the green aphids. To test whether aphids have the ability to control reproduction according to the densities of self and non-self clones, one yellow aphid or one antennae-excised yellow aphid was transferred into a highly dense green clone colony. Intact yellow aphids produced fewer nymphs in crowded green colonies, whereas the fecundity of antennae-excised aphids did not change. Thus, we conclude that aphid clones can discriminate between self and non-self clones, and can regulate their reproduction, depending on whether they are superior or inferior in number to their competitors. Acyrthosiphon pisum , colony , colour mutant , density-dependent effect , reproductive restraint The Japan Society for the Promotion of Science 19K06848 cover-dateJune 30, 2021 ==== Body pmc1. Introduction Clonal reproduction is the major mode of propagation for some plants, parasites, marine benthic animals and aphids. In clonal organisms, competition among clones shapes evolutionary outcomes [1–5]. Several empirical studies have demonstrated that competition among unrelated clones is so strong that it leads to a great reduction in clone diversity within a local population [6–10]. In addition to intense competition among unrelated clones, clonal organisms reportedly often have the ability to discriminate between kin and non-kin clones, and to control the strength of competition depending on their relatedness to the neighbouring clone [1,11–14]. An increasing number of studies have documented that, as with other immobile clonal organisms, plants are capable of self–non-self discrimination and can adjust root and shoot growth, depending on the identity of the neighbouring plants [15–18]. Aphids (Aphididae; Insecta) are peculiar among clonal organisms in that clone members are independent of each other and mobile [19]. In the wild, different aphid clones often coexist, forming mixed colonies on a host plant [20–24]. Clonal mixing can lead to competition among aphid clones over nutrition and space on the same host plant [25–27]. Furthermore, clonal mixing can hinder the evolution of altruistic behaviour such as soldier production [23,28]. Considering the role of aphids as important crop pests, it is an evolutionarily and practically crucial issue to understand whether aphid clones can discriminate between closely related and unrelated clones, and how they regulate their reproduction in mixed colonies. Nevertheless, very few studies have addressed this issue [20,26,27,29]. Regarding aphid species that produce a solder caste or induce communal galls, it has been reported that they have no ability to discriminate between kin and non-kin members [20,30,31]. However, it is not known whether this feature applies to non-social aphid species. Some authors have compared reproductive rates, distribution patterns on plants, and reactions to parasitoids between clonal colonies and mixed colonies of aphids [26,27,29], confirming the existence of significant interactions between clones in some cases. Mixed colonies are reported to reproduce faster [26,27] and exhibit a different response to parasitoids than clonal colonies [29]. However, these studies used DNA markers to distinguish clones, which has the disadvantage of precluding the observation of detailed interactions between live aphid clones. Furthermore, the rearing of aphid colonies on growing plants makes it difficult to precisely count aphid numbers throughout the rearing period. To overcome these disadvantages, we used a colour mutant and its original clone of the pea aphid, Acyrthosiphon pisum, and employed the agar-leaf rearing method [32] to separately evaluate the daily reproduction of the different clones. Acyrthosiphon pisum is a serious pest aphid that feeds on leguminous plants and is distributed worldwide [33]. The two clones were so distinct in colour so as to discriminate clones in all stages. Because we reared the aphids on cut leaves attached to agar medium, we were able to evaluate daily reproduction of the respective clones for 15 days, as well as the distribution patterns of the clones on the leaves. In the present study, we tested the hypothesis that aphids can discriminate self and non-self clones, and can alter their reproduction accordingly. For this purpose, we evaluated (1) the population growth of each clone in colonies founded by a single aphid, two clonal aphids, and two non-clonal aphids (i.e. a mixed colony), (2) the population growth in mixed colonies where two aphid clones start to reproduce simultaneously or with different timing, (3) the reproductive rate of an antennae-equipped or antennae-excised aphid when it was transferred to highly dense self or non-self colonies and (4) the distribution pattern of the members of each clone in a mixed colony to understand whether the clones avoid each other. Synthesizing the results from these experiments, we will, for the first time, indicate that aphids can discriminate self and non-self clones and regulate their reproduction depending on whether they are surrounded by their own clonal members or by members of a distinct clone. 2. Material and method (a) Study organisms We used two A. pisum clones for our experiments, one of which (CR13028c) originated from laboratory crosses, and the other being a colour mutant (CR13028c-Y) derived from the former. CR13028c was a backcross strain, in which sexual females from a hybrid clone (paternal origin: a Vicia cracca-associated clone from Hachinohe, Japan; maternal origin: a V. sativa subsp. nigra-associated clone from Hachinohe) were backcrossed to males from the V. cracca-associated clone. The colour mutant appeared in the stock culture of the original clone in spring, 2017, and had been reared separately. Thereafter, all aphids that were produced parthenogenetically exhibited yellow body colour (hereafter referred to as the yellow clone), which contrasted well with the green body colour of the original clone (hereafter referred to as the green clone) (electronic supplementary material, figure S1). This colour difference enabled us to count the two clones separately within mixed colonies and, thus, the number of newborns daily. In the experiments, apterous adults that were reared at a low density were used for both clones. (b) Competition between the clones To evaluate population growth patterns in the yellow and green clones, we transferred fourth instar aphids onto broad bean leaves, allowed them to reproduce, and counted the total number of aphids, daily, for 15 days (from the first day of larviposition (day 1) to day 15). For rearing, we used the agar-leaf method [32], in which aphids were transferred onto cut leaves on agar medium containing nutrient solution. In this system, aphids grow and reproduce on cut leaves as successfully as they do on broad bean seedlings. Three treatments were prepared, the first of which was a single aphid treatment, which started with one aphid of each clone (1G or 1Y) being transferred onto a leaf. To evaluate density-dependent effects among clonal members, the two-aphid treatment was created by transferring two aphids of each pure clone (2G or 2Y) onto different leaves. In the mixed clone treatment, two aphids of different clones (1G + 1Y) were simultaneously transferred onto different leaves to test whether the coexistence of different clones leads to competitive interactions. For each treatment, 10 replicates were prepared. In the mixed clone treatment, it is likely that a clone with high reproductive rate is dominant in the competition. However, the outcome of the competition may be affected by the timing of reproduction [34]. In the mixed clone treatment, we additionally prepared experiments where a yellow aphid reproduced 2 or 3 days earlier than a green aphid to test whether the timing of reproduction affects the outcome of competition. For these experiments, the final colony size at 16 days after the transfer of yellow adults (and 12 or 13 days after the transfer of green adults) was determined. We examined how the initial difference in the numbers of yellow and green nymphs affect the final sizes of these clones using all experiments of the simultaneous and time lag installation. Round plastic containers (10 cm diameter and 5 cm height) containing agar medium (3 cm height) were used for all of the experiments. A new leaf was placed on the agar surface every 4 days. During observation, if any foundress died, the replicate was removed and a new replicate was prepared. All the containers were placed in a climatic chamber (MIR-254; Sanyo Corporation, Mito, Japan) that was set to 20°C, 50–60% relative humidity, under a 16 L : 8 D photoperiod at 5.8–7.3 W m−2. (c) Reproduction in highly dense colonies We tested whether an aphid's reproductive activity changes when it is placed in highly dense colonies of self or non-self clones. A yellow adult prior to larviposition was transferred to a colony of 130 to 150 aphids consisting of either self or green clone, and the aphid's fecundity was evaluated for 3 days after larviposition. The highly dense colonies were prepared by transferring five aphids of the same clone onto one leaf as foundresses and allowing them to reproduce freely for 5 days. Another leaf was added 3 days later. When a yellow adult was transferred to a highly dense colony of its own clone, it was impossible to distinguish nymphs produced by that aphid and by other aphids. Therefore, all adults were removed from the colony before transfer and a young adult was subsequently transferred into the colony so that we were able to guarantee that newborns were produced by the transferred aphid. To clarify how an aphid detects self and non-self clones, for some yellow adults, we excised both antennae at the third antennal segment using fine forceps. Antennae-excised yellow aphids were then individually transferred to a highly dense colony of the green clone or of self clone to observe reproduction for 3 days. (d) Spatial distribution of the two clones We expected that, if the different clones avoided each other, non-clonal members would be situated more distantly from each other than from clonal members. To examine this, we used images of the mixed colonies from day 12, when the difference in population size between the clones became obvious (see results). The positions of all aphids were located within an X–Y coordinate configuration using ImageJ [35], with the clones being discriminated. We measured the distances between all possible pairs of aphids and between all members of the same clone (yellow or green) and compared the mean distance among green–green pairs (G–G), yellow–yellow pairs (Y–Y) and green–yellow pairs (G–Y). The number of distances less than 5 mm (short distance) and distances ≥ 5 mm (long distance) were then counted for clonal and non-clonal members. For the measurement and categorization of the distances, the R functions ‘dist’ and ‘hist' were used. (e) Statistics Differences in population growth pattern were tested with repeated-measures MANOVA [36] because its objective variable, colony size, involved multiple measures of the same colony over 15 days. The statistical significance of within-subject interactions between days and clones (or colonies under different conditions) was checked to determine whether or not the population growth patterns were different. When the assumption of sphericity was satisfied for all possible pairs of days (i.e. the equivalence of the variance of difference in colony size between all possible pairs of days), the interaction would be tested with F statistics. However, if this assumption was violated, then the degrees of freedom were modified such that a valid F ratio could be obtained. Our analyses showed that the assumption of sphericity was violated in all cases, and so we used the Greenhouse–Geisser correction [35] to adjust the degrees of freedom. When population growth was compared between colonies founded by a single foundress and colonies founded by two clonal foundresses, the population size of the latter on each day was divided by two to obtain the per capita growth pattern. The relationship between the timing of reproduction by the two clones and the final proportions of the clones was examined by the generalized linear model (GLM) with a binomial error structure. In the transfer experiments, differences among treatments in the number of nymphs produced over 3 days were tested using the Tukey–Kramer method. These statistical analyses were performed using JMP v. 13 (SAS Institute, Cary, NC, USA). Regarding the distribution patterns of aphids on leaves in mixed colonies, the mean distances between aphids were compared among G–G pairs, Y–Y pairs and G–Y pairs using ANOVA. Differences in the number of short distances relative to that of long distances were also tested among the three groups using the generalized linear mixed model (GLMM). In the model, aphid density in each container was included as a covariate and variation among containers was treated as a random effect. The analysis was conducted using the glmmML function in the package ‘glmmML' in R v. 3.2.1 [37], with a binomial error structure. 3. Results In all of the treatments, the second generation that was produced on day 1 started reproducing on day 10, resulting in a steeper colony growth curve after day 10 (figure 1). Colonies founded by a single green aphid (1G) increased more rapidly in number than those founded by a single yellow aphid (1Y) (figure 1a; within-subject interaction between time and clone, d.f. = 1.68,30.23, F = 6.49, p = 0.0066). Similarly, colonies founded by two green aphids (2G) increased more quickly than those founded by two yellow aphids (2Y) (figure 1b; within-subject interaction between time and clone, d.f. = 2.84,51.04, F = 7.56, p = 0.0004), but no significant difference was detected in final colony size between the two clones (ANOVA, d.f. = 1,18, F = 0.21, p = 0.65). The final colony sizes for 2G and 2Y colonies were on average 1.17 and 1.35 times that of 1G and 1Y colonies, respectively. Figure 1. Population growth of colonies (cumulative number of aphids) founded by a single aphid, two clonal aphids and two non-clonal aphids over 15 days (mean ± s.e.). For each curve, n = 10. (a) Colonies founded by a single aphid (upper: 1G, lower: 1Y); (b) colonies founded by two clonal or two non-clonal aphids (upper: 1G + 1Y, middle: 2G, lower: 2Y); (c) colonies founded by a single green aphid or a green aphid coexisting with a yellow or a green aphid (upper: 1G, middle: 1G coexisting with 1Y, lower: 1G coexisting with 1G); (d) colonies founded by a single yellow aphid or a yellow aphid coexisting with a yellow or a green aphid (upper: 1Y, middle on the final day: 1Y coexisting with 1Y, lower on the final day: 1Y coexisting with 1G). 1G = one green aphid; 2G = two green aphids; 1Y = one yellow aphid; 2Y = two yellow aphids. Mixed colonies (1G + 1Y) increased more rapidly and attained a larger final size than colonies founded by two clonal aphids (within-subject interaction between time and clone, 1G+1Y versus 2G, d.f. = 2.71,48.86, F = 9.35, p < 0.0001; 1G+1Y versus 2Y, d.f. = 2.58,46.47, F = 9.22, p < 0.0001; ANOVA on day 15, 1G+1Y versus 2G, d.f. = 1,18, F = 80.7, p < 0.0001; 1G+1Y versus 2Y, d.f. = 1,18, F = 49.2, p < 0.0001). The final sizes of mixed colonies were 1.14 and 1.13 times that of 2G and 2Y colonies, respectively (for mixed colonies versus 2G, d.f. = 1,18, F = 80.7, p < 0.0001; for mixed colonies versus 2Y, d.f. = 1,18, F = 49.2, p < 0.0001). In mixed colonies, green aphids overwhelmed yellow aphids in number (figure 1c,d; within-subject interaction, d.f. = 2.13,38.27, F = 96.89, p < 0.0001), and yellow aphids largely restrained their reproduction after day 9 (figure 2). On day 11, the mean numbers of green and yellow aphids in mixed colonies were 88.4 (±5.52 s.e.) and 61.8 (± 2.76), respectively. The difference negatively affected the production of newborns by yellow aphids the next day. On day 12, yellow aphids produced five times fewer nymphs (8.0 ± 1.24) than did green aphids (40.9 ± 4.72) (ANOVA, d.f. = 1,18, F = 45.46, p < 0.0001). Figure 2. Daily production of newborns by the green (upper) and the yellow (lower) clone in mixed colonies (mean ± s.e.). For each curve, n = 10. (Online version in colour.) We compared per capita reproductive rates between the two-clonal aphid treatment and the mixed clone treatment (figure 1c,d). The green aphid reproduced more rapidly when it coexisted with a yellow aphid (1G in 1G + 1Y) than when it coexisted with a fellow clone (1G in 2G) (figure 1c; within-subject interaction between time and clone, d.f. = 2.06,37.01, F = 49.61, p < 0.0001). Colonies founded by 1G alone increased more rapidly than 1G in 2G colonies (figure 1c; within-subject interaction, d.f. = 1.79,32.28, F = 87.65, p < 0.0001) and 1G in 1G+1Y colonies (d.f. = 1.77,31.82, F = 3.64, p = 0.0425). Regarding yellow aphids, since they were outnumbered by green aphids in mixed colonies, their per capita reproduction in mixed colonies (1Y in 1G + 1Y) was much lower than that in the 2Y treatment (1Y in 2Y) (figure 1d; within-subject interaction, d.f. = 2.64,47.51, F = 33.07, p < 0.0001). Nevertheless, the yellow aphid produced a larger number of nymphs by day 10 when it coexisted with a green aphid than when it coexisted with a fellow clone (d.f. = 3.07,55.24, F = 15.57, p < 0.0001). Colonies founded by 1Y alone increased more rapidly than 1Y in 2Y colonies (figure 1d; within-subject interaction, d.f. = 1.96,35.35, F = 41.93, p < 0.0001) and 1Y in 1G+1Y colonies (d.f. = 1.97,35.38, F = 91.40, p < 0.0001). When yellow adults were manipulated to start reproduction 2 or 3 days earlier than green adults, yellow aphids outnumbered green aphids (figure 3a: minus numbers in the horizontal axis mean the numbers of pre-existing green nymphs). If a yellow adult had produced a larger number of nymphs prior to the competitor's reproduction, yellow aphids accounted for higher proportions in the final mixed colonies; logistic regression predicted that if on average 5.5 and 18.9 yellow nymphs pre-exist in the mixed colony, then the final proportion of yellow aphids account for 50.0% (48.8–51.0%; 95% CI) and 75.0% (73.5–76.4%), respectively. Yellow aphids produced on average 10.9 (±3.33 s.d.) and 15.2 (±4.15) nymphs for the first 2 and 3 days, respectively. In addition, there was a link between the proportion of yellow aphids in a mixed colony and the final colony size (figure 3b). The final size of mixed colonies was larger when the proportion of yellow aphids was intermediate (45–71%) than for smaller or larger values (AIC for quadratic regression 268.4, AIC for linear regression 275.1). This result shows that aphids continue to reproduce when they are not inferior in number to their competitors. Figure 3. Relationship (a) between the timing of reproduction of the two clones and the final percentage of yellow aphids in mixed colonies and (b) between the final percentage of yellow aphids and the total number of aphids in mixed colonies. The horizontal axis in (a) indicates the number of yellow nymphs pre-existing when green aphids start reproduction. Negative numbers represent the numbers of green nymphs pre-existing. (Online version in colour.) When a single yellow aphid was transferred to a highly dense colony of the green clone, it produced a significantly smaller number of nymphs than did the control (1Y) (figure 4). When a single yellow aphid was transferred to a highly dense colony of its own clone, its fecundity did not change compared to that of 1Y. Interestingly, when an antennae-excised yellow aphid was transferred to a high-density colony of the green clone, its fecundity was not significantly different from that of 1Y, suggesting that the aphid was not able to perceive the presence or crowding of the different clones. The same was true for an antennae-excised yellow aphid that was transferred to a high-density colony of the fellow clone. Figure 4. Comparison of the number of nymphs produced by a single yellow aphid for the first 3 days after larviposition (mean ± s.e.). Comparison was made among a control aphid and a single aphid transferred to a highly dense colony of the green clone (G) or the yellow clone (Y). The comparison included an antennae-excised yellow aphid transferred to a highly dense colony of the green or yellow clone. (Online version in colour.) The mean distance and the percentage of distances less than 5 mm between aphids for each of G–G pairs, Y–Y pairs and G–Y pairs in each replicate are shown in the electronic supplementary material, table S1 and figure S2. G–Y pairs were located more distantly than G–G pairs or Y–Y pairs in every cage (Tukey–Kramer test: p < 0.05). The percentage of G–Y pairs situated within short distances was significantly smaller than that of G–G pairs (GLMM, z = 7.54, p < 0.0001), but not significantly different from that of Y–Y pairs (z = 1.87, p = 0.0618). Aphid density had no significant effect on the percentage (z = –1.42, p = 0.154). 4. Discussion Using a colour mutant and a novel rearing method, we were able to examine the competition dynamics between and within aphid clones. Remarkably, we found that an aphid can regulate its reproductive rate depending on whether it coexists with a clonal or non-clonal member. When a colony starts from two clonal members, their reproduction is self-regulated from the initial founding stage, and this birth control becomes more prominent in the subsequent generation. By contrast, in the case of an aphid coexisting with a non-clonal aphid, no density-dependent control of reproduction exists at the initial stage, leading to moderate but significant competition effects. The positive effects of competition in mixed colonies, that is, higher reproductive rates, have also been reported in other aphids [26,27], as well as in Chlamydomonas and malaria parasites [38,39]. Competing clones exhibited different reproductive patterns after the onset of reproduction by the offspring generation. When a given clone outnumbered its competitor, it maintained high reproductive rates. However, when the clone was outnumbered by the competitor, it restrained its reproduction by a considerable degree (figures 2 and 4), suggesting that the clone can evade competition with little chance of winning. As a result, competition between the different clones resulted in a clear-cut winner and loser; the green clone far exceeded the yellow clone in the final colony size (2.1 : 1) compared to the ratio of the final colony size of 1G and 1Y colonies (1.1 : 1). These results indicate that aphid clones can regulate their reproductive rate in competitive situations. The results of the experiments manipulating birth timing corroborated this observation. The outcome of clone–clone competition was reversed when yellow aphids started reproduction a little earlier than green aphids, and this finding indicates that the outcome of competition is affected by the relative timing at which the two clones start reproduction as well as by their reproductive rates. Manipulation of birth timing also indicated that where the two clones did not recognize their numerical superiority, the regulation mechanism for reproduction did not work, leading to high densities in both clones (figure 3b). These results suggest that aphids have a flexible ability to regulate their reproduction according to the relative density of self and non-self clones. Given the ability to regulate reproduction, natural selection would favour not only aphids with higher reproductive rates but also those that can access food resources earlier. For example, the timing of egg hatching could be subject to intense selection because foundresses hatching earlier in spring can access preferred resources and start reproduction earlier. Several studies have reported instances of intense competition among nymphal foundresses [40,41]. Similarly, aphid clones that tend to produce a high proportion of winged females can access unused host plants earlier, albeit with a low reproductive rate. The ability of aphids to access resources earlier could compensate for low reproductive rates, possibly leading to a fecundity–dispersal trade-off [42]. The transfer experiment of single aphids corroborated the hypothesis of reproductive restraint only for the yellow clones. We observed that reproductive activities changed significantly between yellow aphids that were transferred to highly dense green colonies and those that were transferred to highly dense fellow colonies. A high density of the green clone has a stronger negative effect on the reproduction of the yellow clone than that of fellow clonal members. The reason for the difference in reproductive activities may be due to the relatedness of colony members. Where an adult is placed in a fellow colony, all aphids are clonal, so that they may behave altruistically, sharing food resources. However, where an adult is surrounded by members of non-self clone, different clones probably behave selfishly such that sharing of food resources may be limited, resulting in lower viability or performance of its offspring. In this situation, reproductive restraint might be advantageous. Although the present study was not able to determine the fate of defeated clones, in the wild, defeated clones probably moved to different parts of the host plant or to different plants to resume reproduction [25,43]. There was evidence to suggest that the yellow clone was a dominant mutant (electronic supplementary material, table S2) and had different phenotypes from the green clone with regard to body colour and reproductive rates. The hypothesis that competing clones could detect the presence of each other through the antennation was tested in the present study by the antennae-excision experiment. The fact that the 3-day fecundity of an antennae-excised yellow aphid that was transferred to a highly dense green colony was not significantly different from that of a control yellow aphid suggests that some substances that can be detected by antennation are involved in the recognition of self and non-self clones or crowding. Comparisons of 3-day fecundity between intact and antennae-excised aphids in highly dense fellow colonies suggest that antennae excision had negligible effects on their reproduction. The body surface of aphids is covered with cuticular hydrocarbons [44], which are used by the attending ants and predators as cues [44,45]. Similarly, it is likely that aphid clones partially use cuticular hydrocarbons as cues for recognizing self and non-self clones. However, the result can be explained if antennae-excised aphids lose the ability to perceive crowding [46]. If aphids can recognize non-clonal members, one might ask whether the members of two different clones would be distributed randomly or distantly from each other on a leaf. Comparisons of the distances between aphid pairs suggested that the distribution of yellow aphids did not completely overlap with that of green aphids in every cage. However, the fact that the percentage of short distances between G–Y pairs was not significantly different from that of Y–Y pairs implies that members of a given clone were not isolated from non-clonal members, such that chemical communication was possible between them. The present study highlighted, for the first time, the ability of aphids to differentiate between self and non-self clones, and to employ reproductive restraint. Throughout the reproductive season from spring to autumn, aphids of a clone always coexist with different clonal members on host plants [19–25], with pre-empting suitable plant parts and overwhelming the competitors on some plants, while, being forestalled and outnumbered on other plants. Our study indicates that aphids have been selected to evolve the capability of context-dependent decision making such that they accelerate or restrain reproduction to flexibly cope with unexpected competitive situations. In another aphid species, Myzus persicae, an increase in colony growth rate was detected under clone–clone competition, and the cause was ascribed to rapid evolutionary changes in clonal composition [26,27]. Although different species may respond differently to competitive situations, it may be important to test whether M. persicae has the ability to accelerate or restrain reproduction depending on the relative density of self and non-self clones. Since our observation was limited to 15 days of aphid reproduction, the outcome of longer-term competition was not elucidated. It has been reported that, in some pairs of aphid clones, the outcome of competition is reversed between low-density and high-density treatments [47]. Thus, more studies are needed to understand the outcome of clone–clone competition when the population density becomes higher or after the population is drastically decreased due to extremely high density and attacks by natural enemies [48]. Nevertheless, long-term studies on clonal competition in aphids, Daphnia, and malaria parasites have reported that clones with higher reproductive rates or clones that infested their hosts first had absolute advantages over competitors [25–27,38,49,50]. Therefore, if aphids have the potential to regulate their reproduction, thereby evading hopeless competition, then it could be predicted that clone–clone competition leads to a distinct winner/loser outcome during the initial stage of colony growth. The next step for future studies would be to explore how defeated clones manage to resume reproduction, to what extent the relatedness of coexisting clones is associated with the strength of competition, and to what extent the components of cuticular hydrocarbons are related to self–non-self recognition in aphids. Supplementary Material Click here for additional data file. Acknowledgements We thank Takashi Kanbe for having found the yellow mutant from a number of A. pisum clone stocks and providing it for us, Toshiko Taniguchi for helping with cross experiments, and Mayako Kutsukake for helpful suggestions for the experimental design. Data accessibility The data used in this manuscript are available from the Dryad Digital Repository: https://doi.org/10.5061/dryad.8gtht76mh [51]. The data are provided in electronic supplementary material [52]. Authors' Contributions Y.L.: formal analysis, investigation, writing-original draft; S.A.: conceptualization, data curation, formal analysis, funding acquisition, investigation, project administration, visualization, writing-original draft, writing-review & editing. All authors gave final approval for publication and agreed to be held accountable for the work performed therein. Competing interests We declare we have no competing interests. Funding This work was supported by Grants-in-Aid (19K06848) for Scientific Research from the Japan Society for the Promotion of Science to S.A. ==== Refs References 1. Buss LW. 1990 Competition within and between encrusting clonal invertebrate. Tren. Ecol. Evol. 5 , 352-356. (10.1016/0169-5347(90)90093-S) 2. Semlitsch RD, Hotz H, Guex G. 1997 Competition among tadpoles of coexisting hemiclones of hybridogenetic Rana esculenta: support for the frozen niche variation model. Evolution 51 , 1249-1261. (10.1111/j.1558-5646.1997.tb03972.x)28565500 3. Oborny B, Kun Á, Czárán T, Bokros S. 2000 The effect of clonal integration on plant competition for mosaic habitat space. Ecology 81 , 3291-3304. (10.1890/0012-9658(2000)081[3291:TEOCIO]2.0.CO;2) 4. Agarwal AA, Underwood N, Stinchcombe JR. 2004 Intraspecific variation in the strength of density dependence in aphid populations. Ecol. Entomol. 29 , 521-526. (10.1111/j.0307-6946.2004.00635.x) 5. Schön I, Martens K, Dijk P. 2009 Lost sex: the evolutionary biology of parthenogenesis. Dordrecht, The Netherlands: Springer. 6. Ayre DJ, Willis BL. 1988 Population structure in the coral Pavona cactus: clonal genotypes show little phenotypic plasticity. Mar. Biol. 99 , 495-505. (10.1007/BF00392557) 7. Coffroth MA, Lasker HR. 1998 Population structure of a clonal gorgonian coral: the interplay between clonal reproduction and disturbance. Evolution 52 , 379-393. (10.1111/j.1558-5646.1998.tb01639.x)28568334 8. Wilson ACC, Sunnucks P, Hales DF. 1999 Microevolution, low clonal diversity and genetic affinities of parthenogenetic Sitobion aphids in New Zealand. Mol. Ecol. 8 , 1655-1666. (10.1046/j.1365-294x.1999.00751.x)10583829 9. Vorburger C. 2006 Temporal dynamics of genotypic diversity reveal strong clonal selection in the aphid Myzus persicae. J. Evol. Biol. 19 , 97-107. (10.1111/j.1420-9101.2005.00985.x)16405581 10. Kanbe T, Akimoto S. 2009 Allelic and genotypic diversity in long-term asexual populations of the pea aphid, Acyrthosiphon pisum in comparison with sexual populations. Mol. Ecol. 18 , 801-816. (10.1111/j.1365-294X.2008.04077.x)19207245 11. Grosberg RK. 1988 The evolution of allorecognition specificity in clonal invertebrates. Q. Rev. Biol. 63 , 377-412. (10.1086/416026) 12. Strassmann JE, Zhu Y, Queller DC. 2000 Altruism and social cheating in the social amoeba Dictyostelium discoideum. Nature 408 , 965-967. (10.1038/35050087)11140681 13. Strassmann JE, Queller DC. 2001 Selfish response by clone invaders. Proc. Natl Acad. Sci. USA 98 , 11 839-11 841. (10.1073/pnas.221450998)11136242 14. Hennige SJ, Morrison CL, Form AU, Büscher J, Kamenos NA, Roberts JM. 2014 Self-recognition in corals facilitates deep-sea habitat engineering. Sci. Rep. 4 , 6782. (10.1038/srep06782)25345760 15. Karban R, Shiojiri K. 2009 Self-recognition affects plant communication and defense. Ecol. Lett. 12 , 502-506. (10.1111/j.1461-0248.2009.01313.x)19392712 16. Yamawo A, Mukai H. 2017 Seeds integrate biological information about conspecific and allospecific neighbours. Proc. R. Soc. B 284 , 20170800. (10.1098/rspb.2017.0800) 17. Fukano Y, Yamawo A. 2015 Self-discrimination in the tendrils of the vine Cayratia japonica is mediated by physiological connection. Proc. R. Soc. B 282 , 20151379. (10.1098/rspb.2015.1379) 18. Fukano Y, Guo W, Noshita K, Hashida S, Kamikawa S. 2019 Genotype-aggregated planting improves yield in Jerusalem artichoke (Helianthus tuberosus) due to self/non-self-discrimination. Evol. Appl. 12 , 508-518. (10.1111/eva.12735)30828371 19. Foster WA. 2002 Soldier aphids go cuckoo. Tren. Ecol. Evol. 17 , 199-200. (10.1016/S0169-5347(02)02486-2) 20. Miller DG. 1998 Consequences of communal gall occupation and a test for kin discrimination in the aphid Tamalia coweni (Cockerell) (Homoptera: Aphididae). Behav. Ecol. Sociobiol. 43 , 95-103. (10.1007/s002650050471) 21. Abbot P, Withgott JH, Moran NA. 2001 Genetic conflict and conditional altruism in social aphid colonies. Proc. Natl Acad. Sci. USA 98 , 12 068-12 071. (10.1073/pnas.201212698) 22. Johnson PCD, Whitfield JA, Foster WA. 2002 Clonal mixing in the soldier-producing aphid Pemphigus spyrothecae (Hemiptera: Aphididae). Mol. Ecol. 11 , 1525-1531. (10.1046/j.1365-294X.2002.01530.x)12144671 23. Abbot P. 2009 On the evolution of dispersal and altruism in aphids. Evolution 63 , 2687-2696. (10.1111/j.1558-5646.2009.00744.x)19500147 24. Vantaux A, Billen J, Wenseleers T. 2011 Levels of clonal mixing in the black bean aphid Aphis fabae, a facultative ant mutualist. Mol. Ecol. 20 , 4772-4785. (10.1111/j.1365-294X.2011.05204.x)21777319 25. Rochat J, Vanlerberghe-Masutti F, Chavigny P, Boll R, Lapchin L. 1999 Inter-strain competition and dispersal in aphids: evidence from a greenhouse study. Ecol. Entomol. 24 , 450-464. (10.1046/j.1365-2311.1999.00223.x) 26. Turcotte MM, Reznick DN, Hare JD. 2011 Experimental assessment of the impact of rapid evolution on population dynamics. Evol. Ecol. Res. 13 , 113-131. 27. Turcotte MM, Reznick DN, Hare JD. 2011 The impact of rapid evolution on population dynamics in the wild: experimental test of eco-evolutionary dynamics. Ecol. Lett. 14 , 1084-1092. (10.1111/j.1461-0248.2011.01676.x)21827586 28. Stern DL, Foster WA. 1996 The evolution of soldiers in aphids. Bio. Rev. 71 , 27-79. (10.1111/j.1469-185X.1996.tb00741.x)8603120 29. Muratori FB, Rouyar A, Hance T. 2014 Clonal variation in aggregation and defensive behavior in pea aphids. Behav. Ecol. 25 , 901-908. (10.1093/beheco/aru064) 30. Carlin NF, Gladstein DS, Berry AJ, Pierce NE. 1994 Absence of kin discrimination behavior in a soldier-producing aphid, Ceratovacuna japonica (Hemiptera: Pemphigidae; Cerataphidini). J. New York Entomol. Soc. 102 , 287-298. 31. Shibao H. 1999 Lack of kin discrimination in the eusocial aphid Pseudorcgma bambucicola (Homoptera: Aphididae). J. Ethol. 17 , 17-24. (10.1007/BF02769293) 32. Li Y, Akimoto S. 2018 Evaluation of an aphid-rearing method using excised leaves and agar medium. Entomol. Sci. 21 , 210-215. (10.1111/ens.12296) 33. Blackman RL, Eastop VF 2006 Aphids on the world's herbaceous plants and shrubs. Chichester, UK: John Wiley & Sons. 34. Mooney KA, Jones P, Agrawal AA. 2008 Coexisting congeners: demography, competition, and interactions with cardenolides for two milkweed-feeding aphids. Oikos 117 , 450-458. (10.1111/j.2007.0030-1299.16284.x) 35. Abramoff MD, Magalhaes PJ, Ram SJ. 2004 Image processing with ImageJ. Biophoton. Inter. 11 , 36-42. 36. Ende CN. 2001 Repeated-measures analysis. In Design and analysis of ecological experiments (eds SM Scheiner, J Gurevitch), 2nd edn. New York, NY: Oxford University Press. 37. R Core Team. 2018 R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. See https://www.R-project.org/. 38. Bell G. 1991 The ecology and genetics of fitness in Chlamydomonas. IV. The properties of mixtures of genotypes of the same species. Evolution 45 , 1036-1046. (10.1111/j.1558-5646.1991.tb04368.x)28564047 39. De Roode JC, Read AF, Chan BHK, Mackinnon MJ. 2003 Rodent malaria parasites suffer from the presence of conspecific clones in three-clone Plasmodium chabaudi infections. Parasitology 127 , 411-418. (10.1017/S0031182003004001)14653530 40. Whitham TG. 1979 Territorial behaviour of Pemphigus gall aphids. Nature 279 , 324-325. (10.1038/279324a0) 41. Muramatsu K, Akimoto S. 2016 Spatiotemporal fluctuations in natural selection acting on the gall-parasitic aphid Tetraneura sorini. J. Evol. Biol. 29 , 1423-1436. (10.1111/jeb.12881)27087064 42. Friedenberg NA. 2003 Determinism in a transient assemblage: the roles of dispersal and local competition. Am. Nat. 162 , 586-596. (10.1086/378782)14618537 43. Hazell SP, Gwynn DM, Ceccarelli S, Fellowes MDE. 2005 Competition and dispersal in the pea aphid: clonal variation and correlations across traits. Ecol. Entomol. 30 , 293-298. (10.1111/j.0307-6946.2005.00703.x) 44. Lang C, Menzel F. 2011 Lasius niger ants discriminate aphids based on their cuticular hydrocarbons. Anim. Behav. 82 , 1245-1254. (10.1016/j.anbehav.2011.08.020) 45. Lohman DJ, Liao Q, Pierce N. 2006 Convergence of chemical mimicry in a guild of aphid predators. Ecol. Entomol. 31 , 41-51. (10.1111/j.0307-6946.2006.00758.x) 46. Kunert G, Weisser WW. 2005 The importance of antennae for pea aphid wing induction in the presence of natural enemies. Bull. Entomol. Res. 95 , 125-131. (10.1079/ber2004342)15877861 47. Turcotte MM, Reznick DN, Hare JD. 2013 Experimental test of an eco-evolutionary dynamic feedback loop between evolution and population density in the green peach aphid. Am. Nat. 181 , S46-S57. (10.1086/668078)23598359 48. Karley AJ, Pitchford JW, Douglas AE, Parker WE, Howardh JJ. 2003 The causes and processes of the mid-summer population crash of the potato aphids Macrosiphum euphorbiae and Myzus persicae (Hemiptera: Aphididae). Bull. Entomol. Res. 93 , 425-438. (10.1079/ber2003252)14641981 49. Loaring JM, Hebert PDN. 1981 Ecological differences among clones of Daphnia pulex Leydig. Oecologia 51 , 162-168. (10.1007/BF00540595)28310076 50. Bell AS, De Roode JC, Sim D, Read AF. 2006 Within-host competition in genetically diverse malaria infections: parasite virulence and competitive success. Evolution 60 , 1358-1371. (10.1111/j.0014-3820.2006.tb01215.x)16929653 51. Li Y, Akimoto S. 2021 Data from: Self and non-self recognition affects clonal reproduction and competition in the pea aphid. Dryad Digital Repository. (10.5061/dryad.8gtht76mh) 52. Li Y, Akimoto S. 2021 Self and non-self recognition affects clonal reproduction and competition in the pea aphid. Figshare.
PMC008xxxxxx/PMC8242833.txt
==== Front Proc Biol Sci Proc Biol Sci RSPB royprsb Proceedings of the Royal Society B: Biological Sciences 0962-8452 1471-2954 The Royal Society 34187197 10.1098/rspb.2021.0874 rspb20210874 1001606970Ecology Research Articles Introgression dynamics from invasive pigs into wild boar following the March 2011 natural and anthropogenic disasters at Fukushima Introgression dynamics from invasive pigs into wild boar following the March 2011 natural and anthropogenic disasters at Fukushima http://orcid.org/0000-0001-6716-4059 Anderson Donovan Conceptualization Data curation Formal analysis Investigation Methodology Resources Software Validation Visualization Writing-original draft Writing-review & editing [email protected] 1 † Negishi Yuki Data curation Formal analysis Investigation Visualization 1 http://orcid.org/0000-0001-7712-8051 Ishiniwa Hiroko Investigation Validation Writing-review & editing 2 http://orcid.org/0000-0001-7997-0701 Okuda Kei Funding acquisition Investigation Resources Validation 3 Hinton Thomas G. Funding acquisition Investigation Resources Validation Writing-review & editing 4 Toma Rio Conceptualization Investigation Software 1 http://orcid.org/0000-0001-9182-319X Nagata Junco Resources Validation Writing-review & editing 5 http://orcid.org/0000-0001-5255-3350 Tamate Hidetoshi B. Investigation Methodology Resources Validation 6 http://orcid.org/0000-0002-9021-8155 Kaneko Shingo Conceptualization Data curation Formal analysis Funding acquisition Investigation Methodology Project administration Resources Software Supervision Validation Visualization Writing-review & editing [email protected] 1 2 1 Symbiotic Systems Science and Technology, Fukushima University, Fukushima City, Fukushima, Japan 2 Institute of Environmental Radioactivity, Fukushima University, Fukushima City, Fukushima, Japan 3 Faculty of Human Environmental Studies, Hiroshima Shudo University, Hiroshima, Hiroshima, Japan 4 Centre for Environmental Radioactivity, Faculty of Environmental Sciences and Natural Resource Management, Norwegian University of Life Sciences, Ås, Norway 5 Forestry and Forest Products Research Institute, Tsukuba, Ibaraki, Japan 6 Yamagata University, Yamagata City, Yamagata, Japan † Present address: Center for Research in Isotopes and Environmental Dynamics, University of Tsukuba, Tsukuba, Ibaraki, Japan. Electronic supplementary material is available online at https://doi.org/10.6084/m9.figshare.c.5470434. 30 6 2021 June 30, 2021 30 6 2021 June 30, 2021 30 6 2021 June 30, 2021 288 1953 2021087414 4 2021 April 14, 2021 7 6 2021 June 7, 2021 © 2021 The Authors. 2021 https://creativecommons.org/licenses/by/4.0/ Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited. Natural and anthropogenic disasters have the capability to cause sudden extrinsic environmental changes and long-lasting perturbations including invasive species, species expansion and influence evolution as selective pressures force adaption. Such disasters occurred on 11 March 2011, in Fukushima, Japan, when an earthquake, tsunami and meltdown of a nuclear power plant all drastically reformed anthropogenic land use. Using genetic data, we demonstrate how wild boar (Sus scrofa leucomystax) have persevered against these environmental changes, including an invasion of escaped domestic pigs (Sus scrofa domesticus). Concurrently, we show evidence of successful hybridization between pigs and native wild boar in this area; however in future offspring, the pig legacy has been diluted through time. We speculate that the range expansion dynamics inhibit long-term introgression and introgressed alleles will continue to decrease at each generation while only maternally inherited organelles will persist. Using the gene flow data among wild boar, we assume that offspring from hybrid lineages will continue dispersal north at low frequencies as climates warm. We conclude that future risks for wild boar in this area include intraspecies competition, revitalization of human-related disruptions and disease outbreaks. invasive species , introgression , naturalization , microsatellite , pig , boar Norges Forskningsråd http://dx.doi.org/10.13039/501100005416 223268/F50 Nippon Life Insurance Foundation http://dx.doi.org/10.13039/501100010233 cover-dateJune 30, 2021 ==== Body pmc1. Introduction Anthropogenic land use can cause ecological impacts to animals inhabiting the landscape, such as limiting range expansion, driving some species to extinction and overall loss of biodiversity [1,2]. Drastic shifts in land use occur as a result of anthropogenic or natural disasters; and these changes are often coupled with new anthropogenic pressures, such as escaped invasive species [3], habitat disruptions [4] and, in rare incidents, an influx of radioactivity contamination [5]. However, while sudden environmental changes or new anthropogenic pressures have caused severe reduction in biodiversity and extinction [6,7], ecosystems have demonstrated ecological resilience and sometimes animal populations experience greater dispersal and increased abundance [8,9]. Such extrinsic ecological changes could cause long-lasting perturbations and challenges, including invasive species, fitness declines and sudden fluctuations in population densities, which may considerably influence evolution. Radioactive materials were dispersed as a result from both the Chernobyl and Fukushima Dai-ichi Nuclear Power Plant (CNPP and FDNPP) accidents in 1986 and 2011, respectively [5], causing widespread radiological contamination of the environments and chronic exposure to ionizing radiation. Consequently, both national governments mandated an evacuation of humans from large areas surrounding the nuclear accident sites (i.e. 4300 km2 for the CNPP and 1150 km2 for the FDNPP). In the evacuated areas, wildlife census studies have demonstrated a rewilding occurred over time, with an abundance of large mammals in both these ecological communities [10–12]. Increased animal abundance of various species in the evacuated area might cause new pressures, such as overpopulation and resource depletion [9]. In the Chernobyl evacuated area, wild boar (Sus scrofa) saw a drastic population increase, but it was stabilized and even declined as predatory populations and disease increased [11,12]. Wild boar in the Fukushima evacuation area could be experiencing changes in population size and behaviour, as suggested by an increased abundance and shifts to more diurnal behaviour [10]; however, these populations lack predators to counter the population increase, other than government-sponsored culling programmes. Evaluation of genetic data and gene flow among populations with indication of expansion trends and high population densities, presumably as a result of sudden ecological changes or a rewilding, may shed light on the adaptive process. Human abandonment of such a large area at Fukushima may have provided favourable conditions for a rapid increase in those wildlife species that were able to benefit from landscapes that were formally anthropocentric [10]. Concurrently, the natural disasters and forced abandonment of farming communities at Fukushima resulted in the release of domesticated livestock into the same landscapes, and it is known that escaped porcine livestock can naturalize in the wild and breed with their wild relatives [13,14]. The favourable environments established new invasive species interactions with the pre-existing species, resulting in intraspecies hybridization, which may alter genetic traits and genomes of the native species [15,16]. The interactions and genetic mixing from intraspecies hybridization can be used to provide key insight on the naturalization, range expansion dynamics and natural selection phenomena of influenced species in these landscapes. Additionally, any introgressed alleles or organelles in the native population that increase, protract or decrease over a period of time can be used to determine if such invasions from a suddenly introduced population (i.e. release of livestock following a natural disaster or forced evacuations) have ecological benefits. We focus on wild boar in Fukushima's evacuated area because wild boar here are experiencing: (i) sudden population expansion as evidenced by greater population abundance [10] and population growth in the area (estimated 49 000 to 62 000 wild boar from 2014 to 2018 [17]); (ii) reduced anthropogenic disturbances because of human evacuations [4], which included approximately 300 km2 of urban and agricultural lands [5]; and (iii) challenges from a recent invasion of about 30 000 pigs (Sus scrofa domesticus), released from abandoned farmlands [18] causing population intraspecies hybridization (see phenotypical colour alteration in electronic supplementary material, figure S1). Presumably, the native wild boar populations may benefit from an introgression of pigs, known for high genetic diversity, novel genotypes and heterosis [13,14,19], which may enhance the chances of adaptation and expansion [20]. However, intraspecies hybridization may also lead to the replacement of the native species by the invading species [21], may cause maladaptation to the natural environment [22] and specifically, intraspecies hybridization between pigs and wild boar may cause alterations of genetic traits including litter sizes, immunology and population expansion dynamics [23]. Thus, investigating intraspecies hybridization and introgression in a natural population influenced by a recent large invasion could infer important information on the complex histories of hybrids including selection, fitness in hybrids or the invaders, and admixture dynamics from range expansion. Here, we evaluate the hypotheses that range expansion acts as resistance against hybridization, and that invasion success is dependent on the invaders' response to naturalization; both hypotheses have been often embedded in evolution [22,24,25]. Specifically, we determined the extent of the genetic introgression from invasive pigs into wild boar in Fukushima Prefecture after the FDNPP accident, and observed the longevity of introgressed alleles or organelles to evaluate the above hypotheses. In addition, we were able to predict the initial contact zone for pigs and wild boar and evaluate the dispersal of offspring from hybrid lineages using genetic data. We show that if future hybridization events occur, they will occur at low frequencies and that the introgressed pig genes should continue to dilute along the expansion of hybrids. Such results would provide evidence of demic diffusion in the wild and that intraspecies competition or lack of survivability of the invasive species has prevented the invaders’ naturalization. 2. Material and methods (a) Study area Due to high-dose rates from radioactive material dispersed in March 2011, the Japanese Government issued an evacuation order for 164 845 people living within a 20 km radius of the FDNPP (https://www.pref.fukushima.lg.jp/site/portal/). The evacuation order was subsequently modified to include areas as far as 40 km northwest of the FDNPP that were also contaminated by the radioactive plume, comprising a total area of approximately 1150 km2 [5,10]. Beginning in 2016, some of the evacuated areas have been remediated and a small percentage of the former residents have returned; however, human access remains severely restricted in the most contaminated areas as a result of the radiation exposure exceeding the safety threshold for humans set by the Japanese Government. The reduced human activities and altered anthropogenic pressures created favourable habitat conditions for mid- to large mammal species, such as wild boar, as evidenced by increased densities in the evacuated area [10]. (b) Sampling and DNA extraction Muscle samples were collected from 191 wild boar, all of which were morphologically identified as typical wild boar in Japan [26], captured in or nearby the Fukushima evacuated zone (described above) from 2015 to 2018 (figure 1). Additional wild boar samples from the period prior to the 2011 evacuations were provided by prefectural hunters, which included 25 muscle samples from a wild boar population in Ibaraki Prefecture, south of Fukushima Prefecture; 10 muscle samples from Yamagata Prefecture and seven muscle samples from Miyagi Prefecture, both north of Fukushima Prefecture. GPS coordinates, sex and estimated age based on tooth erosion patterns were recorded at the trap sites prior to sampling. See electronic supplementary material for approximating the birth period using sampling year and age data. Ten pig muscle samples were obtained from a Fukushima pig slaughterhouse or purchased from a Fukushima prefectural meat market in 2016. In total, 243 samples were assayed for this study. Among the samples, the samples from Fukushima, Miyagi and Yamagata Prefectures, and 10 pig samples were previously used in Anderson et al. [27]; and the 25 samples from Ibaraki Prefecture were used in Nagata et al. [28], which both analysed the mitochondrial DNA (mtDNA) control region. All samples were stored individually at −20°C in 99.5% ethanol until DNA extraction. Total genomic DNA was extracted using the Gentra PureGene Blood & Tissue kit (QIAGEN, CA, USA), according to the manufacturer's instructions. Figure 1. Distribution map of wild boar samples collected in the evacuated zone and nearby areas impacted by radiation dispersal from the Fukushima Dai-ichi Nuclear Power Plant accident in 2011. Other samples were collected from Yamagata, Miyagi and Ibaraki Prefectures, which neighbour Fukushima and are indicated in the inset map. Ambient dose rate (μSv h−1) measurements (1 m above ground surface) are shown for November 2016, provided by the Ministry of Education, Culture, Sports, Science and Technology (MEXT) and Nuclear Regulation Authority (NSR) airborne monitoring project. Map is sourced from Extension Site of Distribution Map of Radiation Dose (MEXT/NSR) site (https://ramap.jmc.or.jp/map). (Online version in colour.) (c) Genetic analysis (i) Mitochondrial DNA The mtDNA sequences provide material lineage information and can be used to identify wild boar with pig maternal ancestry for many distant generations. The mtDNA sequences were obtained from previous literature for all samples in this study [27,28] . In brief, the mtDNA control region was successfully amplified and partial sequences (713 bp) were determined from all samples. DNA sequencing data were viewed from FinchTV chromatogram viewer v. 1.5.0 (Geospiza Inc., WA, USA). (ii) Nuclear microsatellites Nuclear microsatellite (STR) data were analysed to determine the genetic structure of wild boar and the extent of domestic pig introgression into the local wild boar population. A total of 24 STR loci were selected and genotyped for our study populations based on the allele frequencies and the amplification for each of these markers in pure species individuals (see Anderson et al. [29]). All 24 markers were developed by previous studies [30–32] and recommended by the Food and Agriculture Organisation of the United Nations database [33]. For 31 samples already analysed in Anderson et al. [29], we used the genotypes reported therein for the 24 STR loci. The remaining 212 samples were analysed as described in Anderson et al. [29]. In brief, PCR amplification was performed in 5 µl reactions using the QIAGEN Multiplex PCR Kit (QIAGEN, CA, USA) and a protocol for fluorescent labelling [34]. Each sample reaction contained 10–20 ng of genomic template DNA, 2.5 µl of Multiplex PCR Master Mix, 0.1 µM of forward primer, 0.2 µM of reverse primer and 0.1 µM of fluorescently labelled primer. Product sizes were determined using an ABI PRISM 3130 Genetic Analyzer and GeneMapper software (Applied Biosystems, MA, USA). (d) Data analyses Our genetic data, both mtDNA and STR, were divided into two groups for analyses: (i) individual-based data analyses to assess introgression from pigs into wild boar, i.e. intraspecies hybridization in the evacuation zone or nearby areas. Samples from the period prior to the 2011 evacuations and outside Fukushima Prefecture were excluded from this group. (ii) Population-based data analyses to predict possible dispersal patterns of hybrids using gene flow and genetic structure of the native wild boar populations within the evacuated area, to the north (Miyagi and Yamagata Prefectures) and to the south (Ibaraki Prefecture) of the evacuation zone. (i) Hybrid analyses of crossed wild boar and pigs The STR genetic structures and genomic mixtures between wild boar and pigs were investigated using a Bayesian clustering algorithm available in STRUCTURE v. 2.3.4 [35,36]. STRUCTURE was run using the settings of the admixture model with correlated allele frequencies; all parameters were set to default with the LOCPRIOR model implemented. To estimate the number of clusters (K), 10 independent runs with K = 1–10 were performed using 106 Markov chain Monte Carlo (MCMC) iterations following a burn-in of 105 iterations. STRUCTURE HARVESTER [37] was used to calculate the probability of the data for each K (LnP(D)) and ΔK [38]. The LnP(D) remained at 0 with each K except a high peak at K = 2 (electronic supplementary material, figure S2), and the highest ΔK was detected when K = 2. Taken together, K = 2 was the optimal number and we retained K = 2 to identify the proportions of admixture in wild boar and pig hybrids. For the selected K = 2, we assessed the average proportion of the membership coefficient (Q1) to the inferred clusters. We assigned each individual to the inferred clusters, using a conservative threshold Q1 ≥ 0.99 for the assignment of individuals genomes to the pig cluster, or, Q1 < 0.01 to the wild boar cluster. Admixed individuals were jointly assigned to the two clusters and were considered to have possible hybrid ancestry or were the offspring from ancestors that were a cross between wild boar and pig. Admixed individuals were considered hybrids only if neither of the thresholds were true in all independent runs. If an individual admixed in more than 50% of the runs, but not all, then the individual was considered as a ‘suggested hybrid’. For clarity, terms used in this manuscript are defined as follows. First, ‘hybridization’ refers to mixing between evolutionarily distinct lineages, whereas ‘introgression’ refers to gene flow between species as a consequence of hybridization. In this study, introgressed or hybrid wild boar were determined based on the inherited pig lineage haplotype from mtDNA, or an admixed individual based on STRUCTURE (see above), or both. Frequencies of wild boar identified as hybrids were partitioned by the corresponding evacuation zones based on GPS coordinates: initial evacuation zone, extended evacuated zone and outside the evacuation zone. To geographically evaluate wild boar considered as hybrids based on either method, two maps were created using corresponding GPS coordinates and (i) colour based on haplotype or (ii) a colour gradient scale based on Q1 values. Distances from wild boar trap sites and the FDNPP (37°25′23″ N, 141°01′59″ E) were calculated using the geographical distance-based analysis in GenALEx v. 6.41 [39]. We calculated pairwise codominant genotypic distances [40], and performing principal coordinates analysis (PCA) using GenALEx v. 6.41 [39]. (ii) Population-based genetic analyses of wild boar For the subsequent analyses, only individuals that had typical mtDNA haplotypes of wild boar, i.e. J10 and J5, and assigned to the inferred wild boar cluster were used to determine gene flow among the native wild boar populations in this region, as such information could help provide predictions on future dispersal. Samples used for these genetic analyses were considered as pure strain wild boar, or wild boar with only wild boar ancestry and lineages, and are henceforth referred to as ‘pure-wild boar’. The STR data from 202 pure-wild boar, including 160 pure-wild boar from the evacuated zone and 42 pure-wild boar from the neighbouring prefectures to the north and south, were used. Genetic structure of pure-wild boar was investigated by performing a STRUCTURE analysis and a test of isolation by distance. We ran STRUCTURE with the same setting parameters as discussed above and the same criteria were used for K and ΔK. We examined the association between the matrix of the geographical distances and Queller & Goodnights genetics relatedness (r) [41] with three groups: (i) pure-wild boar inside the evacuated zone, (ii) the northern cluster as identified by STRUCTURE, and (iii) among all pure-wild boar, using a Mantel test with 9999 random permutations in GenALEx v. 6.41 [39]. 3. Results (a) Genetic analyses of hybridization in wild boar MtDNA. Two typical wild boar haplotypes (J10 and J5; D42172 and AB015085, respectively) and one typical pig lineage haplotype (H1; MK801664) were observed in wild boar surveyed for mtDNA. For the 191 wild boar samples from within the evacuated zone, the number (and proportion) of each haplotype was 170 (89%), three (1.6%) and 18 (9.4%) for J10, J5 and H1, respectively. In other words, 173 (90.6%) of the mtDNA haplotypes were typical for wild boar lineages and 18 (9.4%) showed evidence of pig ancestry. Nuclear STR. The STRUCTURE and PCA analysis used to investigate the introgression of pig alleles into native wild boar based on 24 loci gave similar genetic structuring compared to that obtained from mtDNA and distinguished wild boar, pig and hybridized individuals. The most probable number of clusters that captured greatest proportion of the data in the STRUCTURE analysis was K = 2 (figure 2) when the LnP(D) and the ΔK were evaluated (electronic supplementary material, figure S2). All individuals were assigned to two clusters at K = 2, a first group comprised the 10 pigs, and a second group comprised all the wild boar within the Fukushima evacuation area with 10 admixed individuals in 100% of the independent runs and an additional five admixed individuals in more than 50% of the runs. The estimated Q1 values for all members of the first group were Q1 ≥ 0.99, while the second group consisted of 160 individuals with a Q1 < 0.01. The 10 admixed individuals in 100% of the runs and additional five admixed individuals in more than 50% of the runs ranged from 0.03 ≤ Q1 ≤ 0.51 and 0.01 ≤ Q1 ≤ 0.03, respectively (figure 2; see more detail in electronic supplementary material, table S1). PCA analysis of the STR data showed similar genetic clustering and separated the dataset into the two genetic clusters, with the identified hybrid wild boar positioned separately from the wild boar cluster, trailing off in the direction of the pig cluster (electronic supplementary material, figure S3). Figure 2. Results from genetic clustering conducted using STRUCTURE analysis (K = 2) based on 24 loci for hybrid analysis. The membership bar-plot indicates per cent membership to pig, where values less than 1% indicated pure-wild boar ancestry and greater than 99% indicated pig ancestry. ‘Introgressed’ indicates admixed individuals. The morphology information was determined in the field from Anderson et al. [27]. Mitochondria results of ‘pig’ indicated the typical pig haplotype, and ‘boar’ indicated typical Japanese wild boar haplotypes. Combined mtDNA and STR results indicated that 31 wild boar, or 16% of the wild boar from the evacuated zone, were hybrids (table 1). The distribution map of wild boar in the evacuated area coupled with either the mtDNA information or Q1 values (figure 3) revealed that the distribution of hybrids was analogous for each dataset and showed similar tendencies with distance from the FDNPP. The greatest number of hybrids detected, or relatively 75% of all hybrids, were within the initial 20 km radius evacuation zone (table 1). The shared ancestry to pig was higher for hybrids in the initial 20 km radius evacuation zone or, in other words, there was evidence of higher introgression inside the initial evacuated zone (figure 4). Outside the evacuation zone, only one female hybrid wild boar was identified using mtDNA (see electronic supplementary material for sex information and table S1). Using the estimated birth year of hybrid wild boars identified using STR analysis, five hybrids were born in 2016 that ranged from 0.03 to 0.51 (average = 0.27) and five hybrids were born in 2018 that ranged from 0.04 to 0.33 (average = 0.21) (electronic supplementary material, table S1). Table 1. Number of hybrids detected using mtDNA and STR data from wild boar samples within the initial evacuation zone, extended evacuated area and outside the evacuation zone, which had 20, 20–40 and greater than 40 km radii, respectively, from the FDNPP. Q1, estimated proportion of the membership coefficient by STRUCTURE. distance (km) n no. of hybrids detected Q1 value of hybrids mtDNA STRa mtDNA and STR total (per cent of all) range average <20 149 11 11 (4) 2 24 (16%) 0.01–0.51 0.1 20–40 24 4 2 (1) 0 6 (25%) 0.01–0.11 0.02 >40 18 1 0 0 1 (6%) n.a. n.a. all 191 16 13 (5) 2 31 (16%) 0.01–0.51 0.08 aNo. of ‘suggested hybrid’ in parentheses (see Material and methods). Figure 3. Distribution map of all wild boar samples in the evacuated area and either their (a) mtDNA haplotype data or (b) the estimated proportion of the membership coefficient (Q1) obtained by STRUCTURE. (Online version in colour.) Figure 4. The estimated proportion of the membership coefficient (Q1) plotted against the distance from the Fukushima Dai-ichi Nuclear Power Plant. Only identified hybrid wild boar were plotted. (Online version in colour.) (b) Genetic structure and admixture among pure-wild boar Closely related populations and a separation of a southern and northern cluster among wild boar populations across our study's wide region, spanning four prefectures of Japan, were observed by the pure-wild boar STRUCTURE analysis. The LnP(D) increased progressively from K = 2 to 3 and then plateaued after K = 3 (electronic supplementary material, figure S4). Additionally, the highest ΔK was detected when K = 2 and the second highest when K = 3. Thus, the most likely number of clusters was K = 2 followed by K = 3. At K = 2, the range of the wild boar was divided into two clusters (figure 5). A first group was detected with a shared ancestry in all populations and dominated the wild boar ancestry in Yamagata Prefecture, Miyagi Prefecture and the Fukushima evacuated zones. A second group was detected that dominated in Ibaraki Prefecture, with a decreasing share of ancestry moving north. At K = 3, a third cluster showed similar share of ancestry between northern prefectures, and the Fukushima evacuated zones, but drastically decreased moving south. The clustering results agreed with PCA results (electronic supplementary material, figure S5), which also indicated wild boar scattered across two genetic clusters. In other words, the Fukushima evacuated zone and northern prefectures' populations tended to overlap positions, while the southern Ibaraki population was positioned separately. In addition, the northern group wild boar collected from the period prior to 2011, suggested little genetic drift from the samples collected after 2011 (F = 0.197, figure 5), while the southern group samples diverged from the northern group (figure 5; see electronic supplementary material, figure S5). Figure 5. Distribution map of pure-wild boar populations and corresponding membership bar plots using STRUCTURE (K = 2 and K = 3) based on 24 loci. (Online version in colour.) Limited gene flow or mixing between the populations from the two main regions identified by STRUCTURE, the southern and northern separation, was also evident from the association between the matrix of the geographical distances and genetics relatedness analysis (electronic supplementary material, figure S6). A clear significant correlation (r = 0.55; p < 0.001) was observed among all wild boar subpopulations in the southern and northern regions. However, no significant correlation was observed (r = 0.03; p = 0.131) for wild boar within the Fukushima evacuated zones. This result indicates that there is genetic mixture and similarity between the Fukushima evacuated area and northern prefecture's wild boar populations, while there is limited gene flow moving south. 4. Discussion (a) Genetic introgression from pigs, an invasive species, into wild boar The data presented here reveal recent hybridization and genetic introgression from invasive pigs into wild boar. Hybrids and the extent of introgression from pigs were identified by maternally inherited markers and analysing multi-loci genotype data, which showed that 31 individuals, morphologically identified as wild boar, in Fukushima Prefecture had pig ancestry (figure 2). In a PCA analysis of wild boar and pigs (electronic supplementary material, figure S2), the majority of the inferred hybrids were positioned between the wild boar cluster and pig cluster. The observed maternally inherited pig haplotype in two of the admixed individuals further support a scenario of hybridization. A series of results showed that while a few individuals with pig ancestry were born into the Fukushima wild boar population, the proportion of pig genes in those individuals' genomes was low (average 8%, table 1). The proportion of hybrid wild boar in this dataset is 16%, which is higher than previously reported figures between 2 and 10% for introgression in European wild boar populations [15,42,43] and 10% for introgression in Ryukyu wild boar [44] using genetic markers. A higher frequency of hybrid wild boar was expected in this study because of the recent release of about 30 000 pigs into the wild following the Japanese Government ordered evacuations in 2011 [18]. However, despite a higher frequency of hybrids in our dataset, the estimated admixture or the average proportion of hybrids’ genomes that originated from a pig population remained low (table 1) with all inferred hybrids containing an admixture of less than 50 and 30%, 5 and 7 years after the FDNPP accident, respectively (electronic supplementary material, table S1). Here, our results may be underestimating the full extent of introgression or hybrid occurrences because mtDNA data can only infer about female lineages (i.e. a cross between a female pig and a male wild boar) and our limiting sample size of pigs may not necessarily be representative of the complete gene pool of escaped pigs for STR. It may also be the case that our low pig ancestry results suggest a scenario that hybrids in this study were probably the offspring of one or more backcross generation of pure-wild boar, which would halve the number of introgressed pig alleles at every generation. While the selected STR markers consistently detected third backcross (BC2) generation hybrids (about 88%; see electronic supplementary material, table S2), it is feasible to assume a fourth or more backcross generation hybrid would likely be beyond the marker's detection ability. The decrease of introgressed genes in the wild boars' genomes with the increase in the distance from the invasion source (figure 4) and with time is likely to be caused from the wild boar being most abundant in the evacuated zone [10] and the random mating tendencies of wild boar in this area (electronic supplementary material, figure S6), which in turn, increases the opportunity of backcross generations with wild boar and less chance of direct contact between remaining invasive pigs or other hybrids. Particularly, our result may be an indication of demic diffusion [25] in the wild, in which the invader gene pool is progressively diluted along the expansion due to recurrent admixture with the native species. This contrasts with reported hybridization in European wild boar populations where, despite new developments of intensive indoor pig farming, older hybridization events are being followed-up by interbreeding among hybrids, or recent escaped pigs, resulting in ongoing introgression on multiple occasions [42]. The difference in introgression observed in Europe versus our study may also be due to the rapid decline and total cessation of pig farming in Fukushima's evacuated zone because radiation dose rates exceed Japanese Government regulations. Therefore, the probability of follow-up hybridization events will remain low; direct contact between domestic pigs and wild boar is highly unlikely until pig farming practices are fully restored to pre-accident levels. Despite such a sudden and large invasive force from pigs, our data likely suggest that most pigs failed to naturalize in the wild and survival of male pigs may have been higher. Here, we observed minimal selection of the pigs' invasion from the remaining invasive pig mtDNA haplotype in wild boar and the reduction of pig alleles, which is more pronounced in the STR loci because they are a mixture from both sexes. However, our results also detected only two individuals as hybrids by both the mtDNA and the STR markers. Previous studies have also pointed out inconsistencies in hybrid determination by mitochondrial and nuclear markers [15]. Given that our selected markers consistently detected third-generation backcrosses (see electronic supplementary material, table S2), we can speculate that female pigs and maternal offspring had difficulty surviving. If pig farms tended to raise more female pigs than male pigs, then this survival bias may explain the lack of concurrence in hybrid determination by mitochondrial and nuclear markers, and the seemingly unnatural distribution of Q1 values in mitochondrial hybrids. Additionally, in the USA, male wild pigs tended to have higher survivability contributed to larger mass at birth compared to females [45]. Thus, it may be possible that male wild pigs or their paternal lineages in this area were more successful than female pigs. (b) Introgression source, dispersal prediction, and what this means for native wild boar Our results suggest that abandoned livestock farms, inside the initial Japanese Government ordered evacuation zone (20 km radius from the FDNPP), were the source of the invasive pigs and the initial contact zone of introgression, with an indication of outward dispersal (figure 3). The highest number of hybrid wild boar was within this area and, on average, the hybrids in this area revealed the highest pig ancestry (table 1). Thus, we can assume from this study that the hybrids have shown gradual range expansion from the contact zone while backcrossing on multiple occasions with wild boar. When we compare our findings from the hybrid and pure-boar analyses to wild boar expansion trends reported in the literature, a probable prediction can be made of how hybrid lineages from pigs will disperse in the future. From the 1700s to 1970s, wild boar were thought to be regionally extinct in the northern mainland of Japan [1], which included our vast study area, presumably due to harsh winter climates and predatory populations. As climates warmed [45] and some predator species went extinct in Japan [46], wild boar expanded northward [1] filling new niches with little competition from conspecific species. Both our pure-wild boar genetic clustering (figure 5) and gene flow data among populations (electronic supplementary material, figure S6) inferred northern population expansion with limited gene flow moving south. We can predict, as climates warm, that wild boar and the offspring from hybrid lineages will follow a similar dispersal pattern north. Thus, if either the hybrid wild boar organelle inherited from invasive pigs or the introgressed pig alleles continue to persist with time, then we are more likely to observe these in northern wild boars’ genomes at low frequencies. The introgression of pig genes is suggested to have contributed to the rapid population growth rates of wild boar in Europe [23,47,48] and in the USA [49], and our evidence of introgression from pigs into the Japanese wild boars' gene pool should be a concern for conservation biologists, as increased abundances of wild boar may affect plant cover and influence food web dynamics in their environment [50,51]. However, our data also suggest that the introgression is most likely not ongoing in this area. Thus, an evolutionary shift due to the mixing of multiple lineages and predominant expansion of intraspecies hybridization is not likely to occur in Japan as seen in Europe and the USA. If our highly supported prediction that the observed hybridization was sourced from the contact zone with reduced ancestry to pigs at greater distances, and if this is coupled with reduced pig farming in the area, then we can expect further dilution of remaining pig genes in the wild boar gene pool. The greater abundance and higher densities of wild boar in the Fukushima evacuated area [10] coupled with our findings hybridization from invasive pigs establish a high-risk population to potential diseases [52,53]. As Japanese Government evacuation orders are lifted and human activities return to these landscapes, we can expect anthropogenic pressures to cause environmental stressors to wild boar and push expansion to less disruptive environments [4]. As this expansion occurs, the opportunities of disease transmission or contacting infected individuals may increase, which could be catastrophic for the high-density wild boar in this region. While there is limited information regarding the mechanisms of disease transmission [52], wild boar at similar population densities in Europe were mathematically modelled and estimated a 60–70% reduction in population density following an infectious outbreak [54]. Extensive introgression from pigs could alter wild boar immunology characteristics and resistance to these diseases [23] and although our investigation of introgression from pigs into the wild boar genome suggests the frequency of pig alleles has decreased at greater distances from the contact zone (table 1), the wild boar within the initial contact zone will likely continue to pass on introgressed pig alleles to the next generations. Therefore, the wild boar in this area should be periodically monitored using genetic markers, especially as pig farming communities are re-established and the chance of new hybridization events increase. 5. Conclusion Here, using genetic markers, we demonstrated evidence of hybridization and introgression of invasive genes as a result of extrinsic environmental changes that reformed anthropogenic land use in Fukushima Prefecture in 2011. We found that there were likely successful hybridization events in the evacuation zone following a sudden and large biological invasion, that thereafter spread and diluted through time. However, a massive introgression was not observed in this area and we speculate two hypotheses for these patterns: (i) an abundant wild boar population caused increasing introgression of wild boar genes into the invasive pigs and a decreasing introgression of invasive genes into the wild boar with the increase in the distance from the invasion source. (ii) The pig legacy passed on to the next-generation hybrids was dependent on the ability of pigs to naturalize in the wild area of Japan. In either case, if the invasion occurred in an environment that was not occupied already by high densities of wild boar, or if invaders had more advantageous survival traits for the wild, then maybe the invasive pigs would have had the same successful adaption as they had in other countries. We recommend that future studies assess the fitness of these hybrids and better characterize their ecological niche using range expansion models and their ecological interactions. Such studies could determine if demic diffusion occurred after range expansion, or if natural selection played a role. Both scenarios suggest that the introgressed genes will eventually disappear in this area. Supplementary Material Click here for additional data file. Acknowledgements We are thankful to all prefectural hunters, and members of the Japanese Government-sponsored culling programme for their support in obtaining samples. We thank the Institute of Environmental Radioactivity at Fukushima University for their support in conducting this research. This work was supported by the support programme of FFPRI for researchers having family obligations. Ethics No animal was killed specifically for this research and all animals were legally culled by licensed hunters. The study was approved by Fukushima University's Institutional Animal Care and Use Committee. All experiments were performed in accordance with relevant guidelines and regulations. Data accessibility All data used in this study, supplemental tables and figures are provided in the electronic supplementary material available from the Dryad Digital Repository: https://doi.org/10.5061/dryad.qnk98sfgz [55]. Authors' contributions D.A.: conceptualization, data curation, formal analysis, investigation, methodology, resources, software, validation, visualization, writing—original draft, writing—review and editing; Y.N.: data curation, formal analysis, investigation, visualization; H.I.: investigation, validation, writing—review and editing; K.O.: funding acquisition, investigation, resources, validation; T.G.H.: funding acquisition, investigation, resources, validation, writing—review and editing; R.T.: conceptualization, investigation, software; J.N.: resources, validation, writing—review and editing; H.B.T.: investigation, methodology, resources, validation; S.K.: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, software, supervision, validation, visualization, writing—review and editing. All authors gave final approval for publication and agreed to be held accountable for the work performed therein. Competing interests We declare we have no competing interests. Funding Funding for this study was partially provided by the Nippon Life Insurance Foundation, and the Research Council of Norway through its Centres of Excellence funding scheme, project no. 223268/F50. ==== Refs References 1. Tsujino R, Ishimaru E, Yumoto T. 2010 Distribution patterns of five mammals in the Jomon period, middle Edo period, and the present, in the Japanese Archipelago. Mammal Study 35 , 179-189. (10.3106/041.035.0304) 2. Pimm SL, Raven P. 2000 Extinction by numbers. Nature 403 , 843-845. (10.1038/35002708) 3. Randi E. 2008 Detecting hybridization between wild species and their domesticated relatives. Mol. Ecol. 17 , 285-293. (10.1111/j.1365-294X.2007.03417.x)18173502 4. Doherty TS, Hays GC, Driscoll DA. 2021 Human disturbance causes widespread disruption of animal movement. Nat. Ecol. Evol. 5 , 513-519. (10.1038/s41559-020-01380-1)33526889 5. Steinhauser G, Brandl A, Johnson TE. 2014 Comparison of the Chernobyl and Fukushima nuclear accidents: a review of the environmental impacts. Sci. Total Environ. 470-471 , 800-817. (10.1016/J.SCITOTENV.2013.10.029)24189103 6. Todesco M et al. 2016 Hybridization and extinction. Evol. Appl. 9 , 892-908. (10.1111/eva.12367)27468307 7. Foley JA et al. 2005 Global consequences of land use. Science 309 , 570-574. (10.1126/science.1111772)16040698 8. Gunderson LH. 2000 Ecological resilience—in theory and application. Annu. Rev. Ecol. Syst. 31 , 425-439. (10.1146/annurev.ecolsys.31.1.425) 9. Perino A et al. 2019 Rewilding complex ecosystems. Science 364 , eaav5570. (10.1126/science.aav5570)31023897 10. Lyons PC, Okuda K, Hamilton MT, Hinton TG, Beasley JC. 2020 Rewilding of Fukushima's human evacuation zone. Front. Ecol. Environ. 18 , 127-134. (10.1002/fee.2149) 11. Webster SC, Byrne ME, Lance SL, Love CN, Hinton TG, Shamovich D, Beasley JC. 2016 Where the wild things are: influence of radiation on the distribution of four mammalian species within the Chernobyl exclusion zone. Front. Ecol. Environ. 14 , 185-190. (10.1002/fee.1227) 12. Deryabina TG, Kuchmel SV, Nagorskaya LL, Hinton TG, Beasley JC, Lerebours A, Smith JT. 2015 Long-term census data reveal abundant wildlife populations at Chernobyl. Curr. Biol. 25 , R824-R826. (10.1016/j.cub.2015.08.017)26439334 13. Iacolina L et al. 2018 Hotspots of recent hybridization between pigs and wild boars in Europe. Sci. Rep. 8 , 1-10. (10.1038/s41598-018-35865-8)29311619 14. Larson G et al. 2005 Worldwide phylogeography of wild boar reveals multiple centers of pig domestication. Science 307 , 1618-1621. (10.1126/science.1106927)15761152 15. Goedbloed DJ et al. 2013 Genome-wide single nucleotide polymorphism analysis reveals recent genetic introgression from domestic pigs into Northwest European wild boar populations. Mol. Ecol. 22 , 856-866. (10.1111/j.1365-294X.2012.05670.x)22731769 16. Harrison RG, Larson EL. 2014 Hybridization, introgression, and the nature of species boundaries. J. Hered. 105 , 795-809. (10.1093/jhered/esu033)25149255 17. JFPG. 2019 Wild boar management plan of Fukushima prefecture (3rd term). In Japanese. See https://www.pref.fukushima.lg.jp/sec/16035b/yaseityouyuuhogokannrijigyoukeikaku-fukushima.html (accessed 15 May 2021). 18. Takahashi R. 2018 Detection of inobuta from wild boar population in Japan by genetic analysis. Rev. Agric. Sci. 6 , 61-71. (10.7831/RAS.6.61-71) 19. Hu C, Pan T, Wu Y, Zhang C, Chen W, Chang Q. 2020 Spatial genetic structure and historical demography of East Asian wild boar. Anim. Genet. 51 , 557-567. (10.1111/age.12955)32510675 20. Pfennig KS, Kelly AL, Pierce AA. 2016 Hybridization as a facilitator of species range expansion. Proc. R. Soc. B 283 , 20161329. (10.1098/rspb.2016.1329) 21. Huxel GR. 1999 Rapid displacement of native species by invasive species: effects of hybridization. Biol. Conserv. 89 , 143-152. (10.1016/S0006-3207(98)00153-0) 22. Verhoeven KJF, MacEl M, Wolfe LM, Biere A. 2011 Population admixture, biological invasions and the balance between local adaptation and inbreeding depression. Proc. R. Soc. B 278 , 2-8. (10.1098/rspb.2010.1272) 23. Goedbloed DJ, Van Hooft P, Megens H-J, Langenbeck K, Lutz W, Crooijmans RPMA, van Wieren SE, Ydenberg RC, Prins HHT. 2013 Reintroductions and genetic introgression from domestic pigs have shaped the genetic population structure of Northwest European wild boar. BMC Genet. 14 , 43. (10.1186/1471-2156-14-43)23688182 24. Lee CE. 2002 Evolutionary genetics of invasive species. Trends Ecol. Evol. 17 , 386-391. (10.1016/S0169-5347(02)02554-5) 25. Quilodrán CS, Tsoupas A, Currat M. 2020 The spatial signature of introgression after a biological invasion with hybridization. Front. Ecol. Evol. 8 , 569620. (10.3389/fevo.2020.569620) 26. Ohdachi S, Ishibashi Y, Iwasa M, Saitoh T. 2009 The wild mammals of Japan, 2nd edn. Kyoto, Japan: Shoukadoh Book Sellers and the Mammal Society of Japan. 27. Anderson D, Toma R, Negishi Y, Okuda K, Ishiniwa H, Hinton TG, Nanba K, Tamate HB, Kaneko S. 2019 Mating of escaped domestic pigs with wild boar and possibility of their offspring migration after the Fukushima Daiichi Nuclear Power Plant accident. Sci. Rep. 9 , 1-6. (10.1038/s41598-019-47982-z)30626917 28. Nagata J, Maruyama T, Asada M, Ochiai K, Yamazaki K, Yamada F, Kawaji N, Yasuda M. 2006 Genetic characteristics of the wild boars in Tochigi Prefecture and neighboring prefectures. Wildl. Tochigi Pref. 32 , 58-62. 29. Anderson D, Negishi Y, Toma R, Nagata J, Tamate H. 2020 Robust microsatellite markers for hybrid analysis between domesticated pigs and wild boar. Genet. Resour. 1 , 29-41. (10.46265/genresj.BNHB8715) 30. Rohrer GA, Alexander LJ, Keele JW, Smith TP, Beattie CW. 1994 A microsatellite linkage map of the porcine genome. Genetics 136 , 231-245. (10.1093/genetics/136.1.231)8138161 31. Krause E, Morrison L, Reed KM, Alexander LJ. 2002 Radiation hybrid mapping of 273 previously unreported porcine microsatellites. Anim. Genet. 33 , 477-485. (10.1046/j.1365-2052.2002.00938_9.x)12464033 32. Karlskov-Mortensen P, Hu ZL, Gorodkin J, Reecy JM, Fredholm M. 2007 Identification of 10 882 porcine microsatellite sequences and virtual mapping of 4528 of these sequences. Anim. Genet. 38 , 401-405. (10.1111/j.1365-2052.2007.01609.x)17559553 33. FAO. 2011 Molecular genetic characterization of animal genetic resources. Rome, Italy: FAO Animal Production and Health Guidelines. No. 9. See https://papers2://publication/uuid/1815F6E0-03DF-4283-ADEE-8EA900C802C2. 34. Blacket MJ, Robin C, Good RT, Lee SF, Miller AD. 2012 Universal primers for fluorescent labelling of PCR fragments—an efficient and cost-effective approach to genotyping by fluorescence. Mol. Ecol. Resour. 12 , 456-463. (10.1111/j.1755-0998.2011.03104.x)22268566 35. Pritchard JK, Stephens M, Donnelly P. 2000 Inference of population structure using multilocus genotype data. Genetics 155 , 945-959. (10.1007/s10681-008-9788-0)10835412 36. Falush D, Stephens M, Pritchard JK. 2007 Inference of population structure using multilocus genotype data: dominant markers and null alleles. Mol. Ecol. Notes 7 , 574-578. (10.1111/j.1471-8286.2007.01758.x)18784791 37. Earl DA, vonHoldt BM. 2012 STRUCTURE HARVESTER: a website and program for visualizing STRUCTURE output and implementing the Evanno method. Conserv. Genet. Resour. 4 , 359-361. (10.1007/s12686-011-9548-7) 38. Evanno G, Regnaut S, Goudet J. 2005 Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Mol. Ecol. 14 , 2611-2620. (10.1111/j.1365-294X.2005.02553.x)15969739 39. Peakall R, Smouse PE. 2006 GenALEx6: genetic analysis in Excel. Population genetic software for teaching and research. Mol. Ecol. Notes 6 , 288-295. (10.1111/j.1471-8286.2005.01155.x) 40. Smouse PE, Peakall R. 1999 Spatial autocorrelation analysis of individual multiallele and multilocus genetic structure. Heredity 82 , 561-573. (10.1038/sj.hdy.6885180)10383677 41. Queller DC, Goodnight KF. 1989 Estimating relatedness using genetic markers. Evolution 43 , 258. (10.2307/2409206)28568555 42. Scandura M, Iacolina L, Crestanello B, Pecchioli E, Di Benedetto MF, Russo V, Davoli R, Apollonio M, Bertorelle G. 2008 Ancient vs. recent processes as factors shaping the genetic variation of the European wild boar: are the effects of the last glaciation still detectable? Mol. Ecol. 17 , 1745-1762. (10.1111/j.1365-294X.2008.03703.x)18371016 43. Dzialuk A, Zastempowska E, Skórzewski R, Twarużek M, Grajewski J. 2018 High domestic pig contribution to the local gene pool of free-living European wild boar: a case study in Poland. Mammal Res. 63 , 65-71. (10.1007/s13364-017-0331-3) 44. Murakami K, Yoshikawa S, Konishi S, Ueno Y, Watanabe S, Mizoguchi Y. 2014 Evaluation of genetic introgression from domesticated pigs into the Ryukyu wild boar population on Iriomote Island in Japan. Anim. Genet. 45 , 517-523. (10.1111/age.12157)24754898 45. Root TL, Price JT, Hall KR, Schneider SH, Rosenzweig C, Pounds JA. 2003 Fingerprints of global warming on wild animals and plants. Nature 421 , 57-60. (10.1038/nature01333)12511952 46. Knight J. 1997 On the extinction of the Japanese wolf. Asian Folkl. Stud. 56 , 129. (10.2307/1178791) 47. Goulding MJ. 2001 Possible genetic sources of free-living wild boar (Sus scrofa) in southern England. Mamm. Rev. 31 , 245-248. (10.1046/j.1365-2907.2001.00086.x) 48. Frantz LAF et al. 2015 Evidence of long-term gene flow and selection during domestication from analyses of Eurasian wild and domestic pig genomes. Nat. Genet. 47 , 1141-1148. (10.1038/ng.3394)26323058 49. Smyser TJ et al. 2020 Mixed ancestry from wild and domestic lineages contributes to the rapid expansion of invasive feral swine. Mol. Ecol. 29 , 1103-1119. (10.1111/mec.15392)32080922 50. Bueno CG, Alados CL, Gómez-García D, Barrio IC, García-González R. 2009 Understanding the main factors in the extent and distribution of wild boar rooting on alpine grasslands. J. Zool. 279 , 195-202. (10.1111/j.1469-7998.2009.00607.x) 51. Schlichting PE, Beasley JC, VerCauteren KC. 2020 The naturalized niche of wild pigs in North America. In Invasive wild pigs in North America: ecology, impacts, and management (eds KC VerCauteren, JC Beasley, SS Ditchkoff, JJ Mayer, GJ Roloff, BK Strickland), 1st edn. Boca Raton, FL: CRC Press. 52. Kodera Y. 2019 C.S.F. prevention of epidemics from a point of view of the ecology of wild boar. J. Vet. Epidemiol. 23 , 4-8. (10.2743/jve.23.4) 53. Ito S, Jurado C, Bosch J, Ito M, Sánchez-Vizcaíno JM, Isoda N, Sakoda Y. 2019 Role of wild boar in the spread of classical swine fever in Japan. Pathogens 8 , 1-12. (10.3390/pathogens8040206) 54. O'Neill X, White A, Ruiz-Fons F, Gortázar C. 2020 Modelling the transmission and persistence of African swine fever in wild boar in contrasting European scenarios. Sci. Rep. 10 , 5895. (10.1038/s41598-020-62736-y)32246098 55. Anderson D, Negishi Y, Ishiniwa H, Okuda K, Hinton TG, Toma R, Nagata J, Tamate HB, Kaneko S. 2021 Data from: Introgression dynamics from invasive pigs into wild boar following the march 2011 natural and anthropogenic disasters at Fukushima. Dryad Digital Repository. (10.5061/dryad.qnk98sfgz)
PMC008xxxxxx/PMC8242924.txt
==== Front Proc Biol Sci Proc Biol Sci RSPB royprsb Proceedings of the Royal Society B: Biological Sciences 0962-8452 1471-2954 The Royal Society 34187193 10.1098/rspb.2021.1082 rspb20211082 100125202Development and Physiology Research Articles Insect flight metabolic rate revealed by bolus injection of the stable isotope 13C Insect flight metabolic rate revealed by bolus injection13C http://orcid.org/0000-0001-9044-2977 Urca Tomer Conceptualization Data curation Formal analysis Investigation Methodology Software Visualization Writing-original draft Writing-review & editing 1 http://orcid.org/0000-0002-9972-7028 Levin Eran Conceptualization Data curation Methodology Project administration Resources Supervision Validation Writing-original draft Writing-review & editing [email protected] 1 2 http://orcid.org/0000-0002-6267-5471 Ribak Gal Conceptualization Data curation Methodology Project administration Resources Software Supervision Validation Visualization Writing-original draft Writing-review & editing [email protected] 1 2 1 School of Zoology, Faculty of Life Sciences, Tel Aviv University, Tel Aviv 6997801, Israel 2 Steinhardt Museum of Natural History, Israel National Center for Biodiversity Studies, Tel Aviv 6997801, Israel Electronic supplementary material is available online at https://doi.org/10.6084/m9.figshare.c.5470435. 30 6 2021 June 30, 2021 30 6 2021 June 30, 2021 30 6 2021 June 30, 2021 288 1953 2021108210 5 2021 May 10, 2021 8 6 2021 June 8, 2021 © 2021 The Authors. 2021 https://creativecommons.org/licenses/by/4.0/ Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited. Measuring metabolic rate (MR) poses a formidable challenge in free-flying insects who cannot breathe into masks or be trained to fly in controlled settings. Consequently, flight MR has been predominantly measured on hovering or tethered insects flying in closed systems. Stable isotopes such as labelled water allow measurement of MR in free-flying animals but integrates the measurement over long periods exceeding the average flight duration of insects. Here, we applied the ‘bolus injection of isotopic 13C Na-bicarbonate’ method to insects to measure their flight MR and report a 90% accuracy compared to respirometry. We applied the method on two beetle species, measuring MR during free flight and tethered flight in a wind tunnel. We also demonstrate the ability to repeatedly use the technique on the same individual. Therefore, the method provides a simple, reliable and accurate tool that solves a long-lasting limitation on insect flight research by enabling the measurement of MR during free flight. stable isotopes , free flight , flight energetics , Coleoptera cover-dateJune 30, 2021 ==== Body pmc1. Background Insect flight is a highly demanding activity, with flight metabolic rate (MR) exceeding resting MR 100-fold in some cases [1]. As the smallest actively flying animals, insects use unique unsteady aerodynamics related to their small size and high flapping frequency [2]. They are also unique in being the only flying animals with an invertebrate body plan entailing an open circulatory system and a unique gas exchange system where oxygen is delivered directly to the flight muscles through trachea [3]. With such fundamental differences, it would be plausible to assume that insect flight physiology differs substantially from flying homeothermic vertebrate (birds and bats). However, comparing flight physiology between flying insects and vertebrates is met with the technical challenge of performing respirometry on insects during flight. Unlike flying vertebrates, where the respired air can be sampled directly via a mask placed over the mouth and nose, insects' tracheal systems limit the study of their respiration to measuring changes in O2 and/or CO2 levels in the surrounding medium. Such measurements are feasible in a small, closed metabolic chamber, as the insect hovers within the confined space [4,5]. However, the measurement of MR during forward flight requires the detection of small changes in gas levels from an enormous volume of air. This was performed successfully by Ellington et al. [6] on bumblebees in a closed wind tunnel showing that their MR hardly changed between flight speeds 0 and 4 m s−1. Their finding suggested that theories on the relationship between power output and flight speed, developed for birds [7] inadequately explain insect flight energetics. The use of radioactively labelled isotopes (e.g. doubly labelled water, [8]) circumvents the challenges of continuous measurement of respiration in flying animals. By injecting the flyer with the labelled metabolic substrates and measuring their body levels after activity, one can estimate their turnover and, therefore, MR. However, the measurement integrates MR over a relatively long period (hours–days), during which the animal is likely to be engaged in more than one activity such as rest, perching, flying and walking. By contrast, the ‘bolus injection of isotopic 13C Na-bicarbonate’ method [9] can measure the MR during activities lasting several minutes. The injected 13C enters the animal's bicarbonate pool and is gradually removed from the body with exhaled CO2. The appearance and decline of 13C in the CO2 from the metabolic chamber (measured by a carbon analyser relative to universal standard as δ13CVDPB) is, therefore, indicative of the rate of CO2 production. To date, the technique has been used to measure MR in free-flying vertebrates (birds and bats), but not on insects. Here, for the first time, we apply the method to measure the flight MR of large beetles. We used the mango stem borer (Batocera rufomaculata, De Geer 1775) as our model for insects relying on flight for dispersal. These large beetles are capable of prolonged flights and exhibit a high degree of intraspecific variance in adult body mass (figure 1), resulting from differences in nutrient availability during larval growth [10]. We injected beetles varying in body mass with 13C isotope and measured their MR in a metabolic chamber before and after flight. From the depletion of δ13C at the two measurements, we estimated the amount of CO2 expired during flight. We used the technique to measure MR of the beetles during free flight and during tethered flight in a wind tunnel, simulating fast (3.5 m s−1) forward flight conditions. We then demonstrated the applicability of the method to other species of beetles by measuring the MR of free-flying rose chafers (Protaetia cuprea) who are smaller and fly more frequently while foraging for pollen and nectar. Figure 1. Variation in body mass of the insects used in the study. (a) Batocera rufomaculata has a high variance in adult body mass associated with differing conditions during the larval growth period. The body mass of field-collected beetles (n = 34, 5.23 ± 1 g) is significantly higher than that of laboratory-reared beetles (n = 27, 2.2 ± 0.6) (t-test, p < 0.001). The body mass of P. cuprea (n = 27, 0.97 ± 0.22 g) is shown in green. Boxes denote the 1st–3rd interquartile range, whiskers denote the range of observed values, the ‘times’ symbol denotes the mean and horizontal lines denote the median body mass. (b) An extreme example of intraspecific variation in adult body mass. Shown are two B. rufomaculata adult males. Differences in adult body mass can reach sevenfold. (Online version in colour.) 2. Methods (a) Animals Mango stem borer beetles (B. rufomaculata) used for the establishment, calibration and validation of the technique were either taken from a laboratory population maintained at Tel Aviv University or collected in the field from infected fig trees. Laboratory-reared beetles emerged from eggs placed in individual test tubes and kept in the dark at 28°C and 70% relative humidity. The hatched larvae were fed fresh fig branches twice a week up to pupation. Experiments started at least a week after adult emergence, once the beetles began feeding. The field-collected beetles were larger than the laboratory-reared ones (mean body mass = 5.1 ± 1.56 and 2 ± 0.6 g, respectively, t-test, p < 0.001, figure 1), enabling measurement of MR as a function of intraspecific variation in body mass. Rose chafer beetles (P. cuprea) were collected around Tel Aviv University during March 2021 and tested within 2 days of collection. In between, the beetles were housed in large 2.5 l transparent plastic containers with moist paper and food (apple) ad libitum. (b) Isotope injection 13C-labelled Na-bicarbonate (Cambridge Isotope Laboratories Inc., Andover, MA, USA) was mixed with insect saline (mM: 150 NaCl, 5 KCl, 5 CaCl2, 2 MgCI2, 10 HEPES, 25 sucrose at pH 7.4) and injected through the intertergal membrane to the dorsolateral abdomen using a microsyringe (HAMILTON, Reno, NV). The molar concentration of 13C and the injected volume were determined by trial and error on 16 specimens of B. rufomaculata (body mass range 1.1–3.3 g) to give a peak of δ13C enrichment (2500–4500‰) in the metabolic chamber (chamber volume = 40 ml), within less than 10 min after the injection, and then a gradual depletion of δ13C in the chamber lasting for at least 30 min. The long δ13C depletion period was necessary to allow ample time to perform the flight experiments and return the insect to the metabolic chamber before total depletion of 13C from the beetle (figure 2). We found that injecting the beetles with a solution of 0.145 M 13C is optimal, with the injected volume (Vol) adjusted according to beetle body mass (m) using the equation:2.1 Vol (μl) =5⋅m(g) +0.3. Figure 2. Flight MR measurement using bolus injections of 13C Na-bicarbonate. (a) Typical isotope enrichment (δ13C) readings of five injected beetles (B. rufomaculta), varying in body mass. δ13C first rises in the metabolic chamber, reaching a maximum within up to 10 min and then starts to decline with time. The rate of increase and decline in δ13C in the metabolic chamber depends on the MR of the insect and not its body mass. (b) The linear relationship between the measured logarithmic 13C-isotope elimination rate (Kc) and the corresponding carbon dioxide production (V˙CO2 ml min−1) as the beetles are resting or harassed in the metabolic chamber (r = 0.93, p < 0.001) provide a calibration for CO2 exhaled by the beetles: V˙CO2(ml min−1)=3.9⋅Kc−0.022. The calibration was then tested on beetles harassed inside the metabolic chamber and yielded a measurement of MR with an accuracy of 89.4 ± 5.4% (c). (d) Outline of the method and calculations: δ13C measurements in rest before and after flight (i, iii) are carried out in a metabolic chamber (MC) connected to a carbon analyser (CA) and used to measure 13C elimination during tethered flight in a wind tunnel (ii). The decline in log-transformed δ13C, over time, measured before and after flight are fitted with a linear regression and is extrapolated forward and backwards to the beginning and ending of the flight bout (broken lines). The slope between the extrapolated points gives the elimination rate (kc) during flight (dotted lines). ‘WT’, wind tunnel (flight) trials. (Online version in colour.) (c) Metabolic chamber and 13C measurement Each injected beetle was placed in the metabolic chamber. Room air (25°C) was pulled by the analyser pump at a rate of 30 ml min−1 (STP) through an Ascarite® column to scrub any atmospheric CO2. The air continued throughout the chamber to a G212-i isotope analyser (PICARRO, Santa Clara, CA, USA). Isotope enrichment (δ13C) and CO2 levels were measured simultaneously and automatically corrected by the analyser for H2O present from the beetle's respiratory water loss. The data were recorded on a desktop computer at 1 Hz frequency. (d) Calibration We calibrated δ13C depletion rate against CO2 emission rate using 24 injected beetles (13 females varying in body mass 1.9–6.5 g; 11 males varying in body mass 1.8–6.8 g, data from all beetles were pooled together). CO2 concentration and δ13C levels were recorded simultaneously in the metabolic chamber for approximately 2 h (the time needed for total 13C depletion from the beetles). The slope (kc) of the linear decline of log (δ13C) with time was then correlated with the mean CO2 emission rate during the same time period. Data from the first 5 min of the experiment were removed to allow the CO2 emission measurement to stabilize. To increase metabolic activity levels beyond resting MR, a small steel plate was glued to the prothorax of eight of the beetles, and the beetles were evoked to move by disturbing them (for 10 min) from outside the chamber using a strong magnet. We began the harassment procedure 5 min after δ13C began to decline in the chamber. (e) Validation and error estimate Once a calibration between kc and V˙CO2 was established (figure 2b), we moved to validate the accuracy of the 13C technique in determining MR during activity compared to standard CO2 measurements. We injected six beetles (body mass range 1.7–3.3 g) with the isotope and glued a piece of metal plate to their prothorax as described above. We allowed the beetle 10 min of rest in the metabolic chamber, during which a clear decline pattern of δ13C was established. The metabolic chamber was then held vertically, evoking the beetles to climb upwards. Once reaching the top of the chamber, the beetles were dragged by a magnet to the lower end of the chamber, triggering the agitated beetles to climb up again rapidly. This ‘climbing activity’ was continuously repeated for 5–8 min. Subsequently, the beetles were left to rest, and the decline in δ13C while resting post-activity was recorded. We then estimated the MR during the activity period as in [11]: linear functions were fitted to the log-transformed measurements of δ13C, as a function of time during rest before and after the climbing activity. The fitted functions were then extrapolated forward and backwards to the start and end of the climbing activity bout. A third linear function was fitted between the start and end of the activity bout (e.g. figure 2d). The slope of this new function (kc) was then translated into CO2 emission rate using the formula obtained from the ‘Calibration’ section. Since the entire experiment was carried out in the metabolic chamber, we were able to compare our estimate of climbing MR using δ13C with the actual V˙CO2 measurements carried out during the climbing activity. (f) Tethered-flight experiments Laboratory-reared beetles (body mass 1.1–3.3 g, n = 16, eight males and eight females) and field-collected beetles (body mass 3–6.9 g, n = 17, nine males and eight females) were injected with the isotope solution and placed in the metabolic chamber until log(δ13C) in the chamber showed a distinct trend of depletion over time (approx. 5 min after reaching maximum δ13C values, figure 2d). The beetles were then taken out of the chamber and a 7 cm long (diameter = 2 mm) metal rod was glued to their dorsal prothorax using hot glue. The rod with the attached beetle was fastened vertically to a tethering system inside the working section of the wind tunnel as described by Urca et al. [12]. The tethering technique allowed the beetles to control their body angle (pitch) and the stroke plane angle relative to the oncoming wind inside the wind tunnel as their meso- and meta-thorax were free to move relative to the fixed prothorax. Preliminary free-flight experiments suggested that body angle during flight (relative to the horizontal plane) is correlated (N = 7, r = 0.7, p < 0.01) with flight speed, V(m s−1) giving:2.2 body angle∘=−14V+83, which, for the 3.5 m s−1 wind (flight) speed used in the wind tunnel, gives an expected body angle of 34°. The tethered beetles (n = 33) in this experiment assumed a mean flight body angle of 20.4° ± 10.4°, which was 16.2° ± 2.6° lower than their body angle at rest prior to flight. The tethered beetle started flying once the wind tunnel was turned on, and we allowed 2 min of continuous flight before the beetles were returned to the metabolic chamber and their δ13C depletion and CO2 emission recorded up to 13C depletion. Flight MR was extracted as in Hambly & Voigt [11]. The decline in δ13C before and after flight was used to linearly extrapolate the decline in δ13C over the period of flight as described in the ‘Validation and error estimate’ section, but replacing the climbing trials in the metabolic chamber with tethered flight in the wind tunnel (figure 2d). (g) Calculation of power output during tethered flight To estimate the power output of the beetles flying in the wind tunnel, we used two synchronized high-speed cameras (FASTCAM SA3_120 K, Photron, Japan) filming the beetles at 2000 fps from two different viewpoints. From the two movies, we reconstructed flapping kinematics in 3D and estimated power from the flapping kinematics using a quasi-steady blade-element model [12] (see also electronic supplementary material, S1). The estimate of mechanical power needed to move the wings through air (Pmech) enabled us to calculate the aerobic efficiency of converting metabolic work to mechanical energy. The MR measured with the 13C technique as ml CO2 s−1 was first converted to Watts using the equation2.3 Pmetabolic(W)=V˙CO2(ml s−1)RQ×(16+5.164×RQ), where the second parenthesis on the right side of the equation is the oxyjoule equivalent (joules O2 ml−1) and RQ is the respiration quotient (ratio between CO2 emission rate and O2 intake rate, RQ=V˙CO2/V˙O2) [13]. RQ can range from 0.7 to 1 depending on the metabolic fuel used at the time of MR measurement. Batocera rufomaculata has a mean RQ of 1.05 ± 0.17 during rest (n = 30, T.U., unpublished data 2018) implying that carbohydrates are the main source of fuel prior to flight. For the sake of simplicity, we assumed here that carbohydrates remain the main source of fuel during short flights [14] and, therefore, used RQ = 1 in equation (2.3). The aerobic efficiency (ηaerobic) can then be calculated as2.4 ηaerobic=PmechPmetabolic. We also assumed RQ = 1 to compare our measurements with available data from the literature (electronic supplementary material, S2). Simultaneously with high-speed filming, the forces generated by the flying beetle were directly measured at 100 Hz using two force transducers attached to the tether arm, as described by Urca et al. [12]. The force transducers measured forces in two axes: parallel to the horizontal flight direction and vertical. The vertical force was measured directly as a decrease in the weight of the beetle during flight. The horizontal force was measured as the difference between horizontal force reading when the beetle was (i) flying and (ii) resting, with the wind tunnel operating at 3.5 m s−1 wind speed in both cases. We used the measured forces to evaluate the tethering effect on the beetle flight. The resultant of horizontal and vertical forces (Ftotal, in Newton) divided by the beetle's body weight (mg) gave a non-dimensional number2.5 Inv= Ftotalmg, where g is the gravitational acceleration (9.8 m s−2) and m is beetle mass in kg. Inv scores the beetle's investment in flight: a value of Inv = 1 implies sufficient force generated to support body mass in air, whereas lower values imply insufficient force production for free-flight conditions (i.e. tethering effect). (h) Free-flight experiments Three injected B. rufomaculata were treated as described above, but rather than fixed to a tether in the wind tunnel, they were placed on a wooden perch at the lobby of the Life Sciences building (12 × 6 × 3 m L × W × H). They were encouraged to take off by directing warm air from a hairdryer towards them. A stopwatch was activated each time the beetles flew, and we stimulated the beetles to fly repeatedly. The total duration of flight attempts did not exceed 8 min before returning the beetles to the metabolic chamber. The cumulative duration of each beetle's flight was used in the calculation of flight MR as described above for the wind tunnel study. Following the post-flight measurement, we allowed the three injected beetles at least an hour of rest (a duration sufficient for complete depletion of the remaining 13C levels from the animals, as confirmed in the metabolic chamber). The beetles were then injected again and the entire flight MR measurement procedure was repeated, giving measurements of two flight bouts from each of the three beetles. To test the technique's applicability to other species, we performed additional MR measurements on seven free-flying P. cuprea beetles. The beetles were injected with the same mass-specific dose of 13C (equation (2.1)) and stimulated to fly freely in a large well-lit room (7.2 × 3.6 × 2.5, L × W × H m) as described above. The calibration between V˙CO2 and Kc for P. cuprea was established on 20 injected beetles (body mass range 0.68–1.4 g), as described in the ‘Calibration’ section. However, because the size of the prothorax in these beetles was much smaller compared to that of B. rufomaculata, we did not glue a metal plate on them. Rather, we increased their activity level by raising the ambient temperature (25–40°C) to give a broader range of MR for the calibration [15]. (i) Statistics We used the general linear model (GLM) with body mass as a covariate to compare mass-specific flight MR. Data were linearized by log transformation, and tethered flight power was corrected for tethering effects (Inv) either by using Inv as a covariate or by dividing the measured power by Inv (see specific tests below). All results are reported as mean ± s.d. 3. Results (a) Technique calibration and validation V˙CO2 measured directly from injected B. rufomaculata in a metabolic chamber was tightly correlated with the logarithmic depletion (kc) of δ13C in the chamber over time (r = 0.93, p < 0.001, figure 2b). When beetles were harassed, forcing them to climb vertically inside the metabolic chamber, the 13C Na-bicarbonate method was able to predict their climbing V˙CO2 with an error of ± 10.2 ± 5.4% compared to V˙CO2 measured directly (n = 6, figure 2c). (b) Tethered flight The tethered flight MR, averaged from all the beetles, was 45.6 ± 24.8 ml CO2 h−1 (figure 3a). This value was, on average, 16 ± 17.0 fold higher than the measured resting MR of the same beetle. The tethered beetles exerted lower forces than required for supporting their body weight in the air (Inv = 0.61 ± 0.22). When divided by Inv to correct for tethering effects, the flight MR Inv−1 increased to 33 ± 47.5 fold the resting MR. Figure 3. Flight energetics of B. rufomaculata. Males are denoted by full circles and females by empty circles. Blue and red denote laboratory-reared and field-collected beetles, respectively. (a) The log converted flight MR corrected for investment show a significant correlation with log (body mass) (r = 0.46, p = 0.012) with an allometric equation of 36⋅M0.56. (b) Mass-specific mechanical power increased significantly with investment (r = 0.6, p < 0.001) and showed a significant interaction between beetle source and sex (GLM, p = 0.018) resulting from field-collected males (full red circles) having a higher mass-specific power output compared to field-collected females (empty red circles). (c) The aerobic efficiency (mechanical power output/metabolic power) increased with body mass (r = 0.52; p = 0.006) with no significant beetle source (laboratory-reared/field-collected) or sex effects (GLM with body mass as a covariate, source: p = 0.1; sex: p = 0.25). (Online version in colour.) Flight MR Inv−1 increased with body mass (figure 3a) but did not differ between males and females (p = 0.75) or between laboratory-reared and field-collected beetles (p = 0.71) once corrected for body mass (GLM with body mass as a covariate). The data of all beetles suggested marginal hypoallometry of FMR Inv−1 (figure 3a). The exponent of the allometric equation was 0.57 but it was not significantly different than 1.0 (confidence interval 0.13–1.0). The mass-specific power needed to move the wings through the air (Pmech∗), estimated from the flapping kinematics (electronic supplementary material, S1), was positively correlated with Inv (r = 0.64, p < 0.001, figure 3b) but not with the mass-specific flight MR (r = 0.12, p = 0.54). Pmech∗ did not vary between laboratory-reared and field-collected beetle or between males and females (GLM with Inv as a covariate, beetle source: p = 0.074; Sex: p = 0.37, figure 3b) but did have a significant interaction between beetle source and sex (p = 0.018). The significance was due to field-collected males having a higher mass-specific mechanical power output than field-collected females (Tukey's post hoc, p = 0.008). The mechanical power divided by metabolic power (Pmech/MR) gave an estimate of the aerobic efficiency (ηaerobic) of the tethered beetles ranging between 9 and 55%. ηaerobic increased with beetle body mass (r = 0.54, p = 0.003, figure 3c). (c) Free-flight experiments The three B. rufomaculata beetles that were injected twice and allowed to fly freely within a large room had a cumulative flight duration of 26–43 s in each of two flight bouts (electronic supplementary material, S2). The beetles survived both injections and remained vital for at least a week after the experiments. Free-flight MR exceeded the MR during tethered flight approximately threefold even after correcting for tethering effects (i.e. correcting for Inv, figure 4) and reached a mean mass-specific MR of 93.6 ± 19.8 ml CO2 g−1 h−1. The smaller species of beetles (P. cuprea, n = 7) flew for 43–68 s in a flight bout and had a mean mass-specific flight MR of 69.5 ± 30.9 ml CO2 g−1 h−1 (figure 4; electronic supplementary material, figure S2). Figure 4. Mass-specific MR in tethered and free flight. The mass-specific flight MR of tethered B. rufomaculata corrected for investment (grey, N = 29) was on average 3.6-fold lower than flight MR during free flight (brown, N = 6). The MR of free-flying P. cuprea (green, N = 7) was on average 1.34-fold lower than that of the free-flying B. rufomaculata. Boxes denote the 1st–3rd interquartile range, whiskers denote the range of observed values, the ‘times’ symbol denotes the mean and horizontal line denotes the median body mass. (Online version in colour.) 4. Discussion To the best of our knowledge, this is the first study showing the applicability of the 13C-bicarbonate bolus injection method for the measurement of insect flight MR. Using the method, we were able to measure the MR of B. rufomaculata during tethered flight in a wind tunnel under conditions simulating forward flight and during free flight within a large room. Such flight conditions are not amenable to MR measurement using respirometry due to the limitation of detecting small changes in O2 or CO2 concentrations in the large air volumes flowing past the beetle. The relatively large body size of B. rufomaculata makes it easier to inject it with the isotope solution, and we found no apparent adverse short- or long-term side effects for the animals. The injected beetles flew readily in the experiments and remained viable for four to six weeks after the injection. Moreover, we were able to inject the same beetles more than once, thus allowing us to measure their flight MR repeatedly. Hence, the technique enables repeated measurements of flight MR, enabling a comparative paired design of ‘before’ and ‘after’ treatment. Our ability to apply the method to the smaller P. cuprea (body mass 0.59–1.45 g) as well, suggests that the method can be applied to even smaller insects. The δ13C enrichment and depletion in the metabolic chamber as a function of time and the linear relationship between kc and V˙CO2 (figure 2) resembled those reported previously for vertebrates [11]. As opposed to the study of homeothermic vertebrates, where the ambient temperature was used to elevate the resting MR during calibration, we either used magnets to mechanically agitate the beetles and boost their activity levels (B. rufomaculata) or changing temperature (P. cuprea). In both cases, increasing the MR during calibration significantly improved the correlation between kc and V˙CO2. The inclusion of elevated activity in the calibration may be particularly important in insects where resting MR can be extremely low and highly variable. Using such calibration, the yielded accuracy of the technique for measuring MR was 89.8 ± 5.4%, a relatively high accuracy considering that MR is estimated through indirect measurements of CO2 production based on 13C depletion before and after activity. Certain specific experiments require insects to be tethered in a wind tunnel for force or electrophysiological measurements. We were able to measure MR in such a set-up despite the high wind speed (3.5 m s−1) needed to simulate forward flight conditions. While tethered-flight experiments have obvious drawbacks, as revealed by the Inv < 1 found in our study, tethering the beetles allowed us to ensure constant flight conditions during continuous flight for 2 min while measuring the forces applied by the insect on the tether. Moreover, it allowed us to measure the flapping kinematics at high precision to estimate the beetle's mechanical power output during flight. Although the conditions are not representative of free flight, conducting the experiments only on free-flying beetles would have resulted in MR measurements of insects flying at different speeds, accelerating, manoeuvring and landing, thus, limiting our ability to relate the MR measurement to specific flight conditions and, therefore, making it difficult to compare the MR measurement between beetles varying in body mass. Our tethered flight set-up revealed how: (i) flight MR Inv−1 increased with body size with no difference in mass-specific cost of flight between small and larger individuals or males and females; (ii) it also showed that larger (field-collected) males required higher mass-specific mechanical power than field-collected females (after correcting for investment) to fly in the wind tunnel at the same flight (wind) speed. Their increase in Pmech∗ was not mirrored by an increase in mass-specific flight MR contributing to a higher estimate of aerobic efficiency in larger beetles. The aerobic efficiency found for B. rufomaculata may be an overestimate because the metabolic power was converted to watts assuming carbohydrates were used as fuel (RQ = 1). If beetles use other substrates during flight, RQ can become as small as 0.7 [16], implying that our estimate of metabolic power may be underestimated by up to 24.5% (equation (2.3)). Higher metabolic power for the same mechanical power implies lower aerobic efficiency (equation (2.4)). More importantly, for RQ = 0.7, our flight MR data converted to VO2 in electronic supplementary material, table S3 would have been approximately 25% higher. Since shifts in metabolic substrates may occur during long flight, future studies on insect flight MR should focus on determining these substrates and their RQ to allow more accurate, species-specific, assessments of MR and power. Our free and tethered flight measurements of MR were in general agreement with previously reported values in tethered and free-flying insects (electronic supplementary material, S3). Previous studies have estimated that the lack of need to support body weight in the air during tethered flight underestimates flight MR by 20–50% [17], which agrees with our finding of a 3.6-fold higher mass-specific MR during free flight even after correcting for Inv in the tethered-flight experiment. The mean mass-specific MR measurements during free flight of B. rufomaculata were 35% higher than those of the rose chafer, possibly reflecting taxonomic and anatomical differences (rose chafers are more streamlined, fly with their elytra closed and have approximately threefold higher flapping frequency). Both species have mass-specific flight MR that falls within the range of reported values from other insects (electronic supplementary material, S3). 5. Conclusion The development of real-time portable stable carbon isotopes analysers and the variety of labelled nutrients and molecules has many promising directions in the study of metabolism, substrate turnover and ecophysiology of animals [18–20]. Our study demonstrates that the bolus 13C-bicarbonate injection technique can be used efficiently and effectively on flying insects under free-flight and controlled flight conditions. It thereby enables flight MR to be measured outside the confines of a closed system. The method provides a novel transformative tool for studying flight MR in insects under natural flight modes such as foraging, commuting and aerial hawking. Supplementary Material Click here for additional data file. Acknowledgements We thank the staff of the School of Zoology at Tel Aviv University for logistical support as well as the reviewers of this manuscript for their valuable input. Data accessibility The data are provided in the electronic supplementary material associated with this paper [21]. Authors' contributions T.U.: conceptualization, data curation, formal analysis, investigation, methodology, software, visualization, writing—original draft, writing—review and editing; E.L.: conceptualization, data curation, methodology, project administration, resources, supervision, validation, writing—original draft, writing—review and editing; G.R.: conceptualization, data curation, methodology, project administration, resources, software, supervision, validation, visualization, writing—original draft, writing—review and editing. All authors gave final approval for publication and agreed to be held accountable for the work performed therein. Competing interests We declare we have no competing interests. Funding We received no funding for this study. ==== Refs References 1. Ellington CP. 1984 The aerodynamics of hovering insect flight. VI. Lift and power requirements. Phil. Trans. R. Soc. Lond. B 305 , 145-181. (10.1098/rstb.1984.0054) 2. Sane SP. 2003 The aerodynamics of insect flight. J. Exp. Biol. 206 , 4191-4208. (10.1242/jeb.00663)14581590 3. Weis-fogh T. 1964 Diffusion in insect wing muscle, the most active tissue known. J. Exp. Biol. 41 , 229-256. (10.1242/jeb.41.2.229)14187297 4. Chappell MA, Morgan KR. 1987 Temperature regulation, endothermy, resting metabolism, and flight energetics of tachinid flies (Nowickia sp.). Physiol. Zool. 60 , 550-559. (10.1086/physzool.60.5.30156129) 5. Bartholomew GA, Casey TM. 1978 Oxygen consumption of moths during rest, pre-flight warm-up, and flight in relation to body size and wing morphology. J. Exp. Biol. 76 , 11-25. (10.1242/jeb.76.1.11) 6. Ellington C, Machin K, Casey T. 1990 Oxygen consumption of bumblebees in forward flight. Nature 347 , 472-473. (10.1038/347472a0) 7. Pennycuick CJ. 1969 The mechanics of bird migration. Ibis (Lond. 1859) 111 , 525-556. (10.1111/j.1474-919X.1969.tb02566.x) 8. Lifson N, Gordon GB, McClintock R. 1955 Measurement of total carbon dioxide production by means of D2O18. J. Appl. Physiol. 7 , 704-710. (10.1152/jappl.1955.7.6.704)14381351 9. Speakman JR, Thomson SC. 1997 Validation of the labeled bicarbonate technique for measurement of short-term energy expenditure in the mouse. Z. Ernahrungswiss. 36 , 273-277. (10.1007/BF01617797)9467215 10. Brown S, Soroker V, Ribak G. 2017 Effect of larval growth conditions on adult body mass and long-distance flight endurance in a wood-boring beetle: do smaller beetles fly better? J. Insect Physiol. 98 , 327-335. (10.1016/j.jinsphys.2017.02.008)28237580 11. Hambly C, Voigt CC. 2011 Measuring energy expenditure in birds using bolus injections of 13C-labelled Na-bicarbonate. Comp. Biochem. Physiol. A Mol. Integr. Physiol. 158 , 323-328. (10.1016/j.cbpa.2010.05.012)20510385 12. Urca T, Debnath AK, Stefanini J, Gurka R, Ribak G. 2020 The aerodynamics and power requirements of forward flapping flight in the mango stem borer beetle (Batocera rufomaculata). Integr. Org. Biol. 2 , obaa026. (10.1093/iob/obaa026)33796817 13. Lighton JRB, Bartholomew GA, Feener DH. 1987 Energetics of locomotion and load carriage and a model of the energy cost of foraging in the leaf-cutting ant Atta colombica Guer. Physiol. Zool. 60 , 524-537. (10.1086/physzool.60.5.30156127) 14. Weis-Fogh T. 1952 Fat combustion and metabolic rate of flying locusts (Schistocerca gregaria Forskål). Phil. Trans. R. Soc. Lond. B 237 , 1-36. (10.1098/rstb.1952.0011) 15. Clarke A, Fraser KPP. 2004 Why does metabolism scale with temperature? Funct. Ecol. 18 , 243-251. (10.1111/j.0269-8463.2004.00841.x) 16. Thompson SN, Bennett RB. 1971 Oxidation of fat during flight of male Douglas-fir beetles, Dendroctonus pseudotsugae. J. Insect Physiol. 17 , 1555-1563. (10.1016/0022-1910(71)90162-4) 17. Niven JE, Scharlemann JPW. 2005 Do insect metabolic rates at rest and during flight scale with body mass? Biol. Lett. 1 , 346-349. (10.1098/rsbl.2005.0311)17148203 18. Levin E, Lopez-Martinez G, Fane B, Davidowitz G. 2017 Hawkmoths use nectar sugar to reduce oxidative damage from flight. Science 355 , 733-735. (10.1126/science.aah4634)28209896 19. McCue MD, Welch KC. 2016 13C-Breath testing in animals: theory, applications, and future directions. J. Comp. Physiol. B 186 , 265-285. (10.1007/s00360-015-0950-4)26660654 20. Welch KC, Péronnet F, Hatch KA, Voigt CC, McCue MD. 2016 Carbon stable-isotope tracking in breath for comparative studies of fuel use. Ann. N. Y. Acad. Sci. 1365 , 15-32. (10.1111/nyas.12737)25817456 21. Urca T, Levin E, Ribak G. 2021 Insect flight metabolic rate revealed by bolus injection of the stable isotope 13C. FigShare.
PMC008xxxxxx/PMC8242925.txt
==== Front Proc Biol Sci Proc Biol Sci RSPB royprsb Proceedings of the Royal Society B: Biological Sciences 0962-8452 1471-2954 The Royal Society 34187196 10.1098/rspb.2021.0774 rspb20210774 10011460Behaviour Research Articles Noise matters: elephants show risk-avoidance behaviour in response to human-generated seismic cues Elephants response to human-generated seismic cues http://orcid.org/0000-0002-7230-3647 Mortimer Beth Conceptualization Data curation Formal analysis Funding acquisition Investigation Methodology Project administration Resources Validation Visualization Writing-original draft Writing-review & editing [email protected] 1 Walker James A. Formal analysis Investigation Methodology Validation Writing-review & editing 1 Lolchuragi David S. Investigation Methodology Validation Writing-review & editing 2 http://orcid.org/0000-0001-8481-8392 Reinwald Michael Visualization Writing-review & editing 1 Daballen David Methodology Supervision Validation Writing-review & editing 2 1 Department of Zoology, University of Oxford, Oxford OX1 3PS, UK 2 Save the Elephants, Marula Manor, Karen, Nairobi 00200, Kenya Electronic supplementary material is available online at https://doi.org/10.6084/m9.figshare.c.5470431. 30 6 2021 June 30, 2021 30 6 2021 June 30, 2021 30 6 2021 June 30, 2021 288 1953 202107741 4 2021 April 1, 2021 8 6 2021 June 8, 2021 © 2021 The Authors. 2021 https://creativecommons.org/licenses/by/4.0/ Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited. African elephants (Loxodonta africana) use many sensory modes to gather information about their environment, including the detection of seismic, or ground-based, vibrations. Seismic information is known to include elephant-generated signals, but also potentially encompasses biotic cues that are commonly referred to as ‘noise’. To investigate seismic information transfer in elephants beyond communication, here we tested the hypothesis that wild elephants detect and discriminate between seismic vibrations that differ in their noise types, whether elephant- or human-generated. We played three types of seismic vibrations to elephants: seismic recordings of elephants (elephant-generated), white noise (human-generated) and a combined track (elephant- and human-generated). We found evidence of both detection of seismic noise and discrimination between the two treatments containing human-generated noise. In particular, we found evidence of retreat behaviour, where seismic tracks with human-generated noise caused elephants to move further away from the trial location. We conclude that seismic noise are cues that contain biologically relevant information for elephants that they can associate with risk. This expands our understanding of how elephants use seismic information, with implications for elephant sensory ecology and conservation management. Loxodonta africana , noise , risk-avoidance , vibrational communication , seismic vibration British Ecological Society http://dx.doi.org/10.13039/501100000409 LRB18/1010 Royal Society http://dx.doi.org/10.13039/501100000288 URF/R1/191033 Royal Commission for the Exhibition of 1851 http://dx.doi.org/10.13039/501100000700 St Anne's College, Oxford cover-dateJune 30, 2021 ==== Body pmc1. Introduction African elephants use a variety of sensory modes, whether for communication through signals [1,2] or for information gathering by detecting cues generated by other animals (e.g. predator odour [3]), humans (e.g. voices [4], roads/railways [5]) or natural earth processes (e.g. rainfall [6]). Information transfer through ground-based, or seismic, vibrations is the least well understood [7]. Elephants likely detect seismic information using the Pacinian corpuscles on the feet and/or the inner ear, picking up ground vibrations via bone conduction [8]. Elephants are thought to use seismic signals to communicate with each other, as seismic vibrations are generated by elephants during certain infrasonic vocalizations, known as rumbles [9–11]. The rumbles contain both acoustic and seismic components in the frequency range 20–40 Hz and under [12–14], which are modelled to propagate to a maximum of 6 km under differing favourable conditions [15–17]. Sensory modes can be investigated using playback experiments, including seismic information transfer in elephants. Behavioural changes during these experiments indicate what signals and cues animals can detect and what they can discriminate between, which involves classifying the potential information according to the source identity and status [18]. A series of acoustic (no seismic component) playback experiments with elephants has shown that elephants can detect, for example, elephant rumbles, the sound of bees, human voices and big cat growls [4,12,19–21]. Elephants can also discriminate between subtle differences in acoustic sources, changing their behavioural response, for example human voices versus bees [11], or even gender, age and ethnicity of human voices [4]. Two studies to date have used seismic playback experiments to explicitly test elephant responses to ground-based vibrations [22,23]. These showed that seismic playback of (the acoustic component of) an elephant alarm rumble is sufficient to elicit a behavioural response [23] and the elephants responded significantly to seismic playback of alarm rumbles from familiar elephants, but not from unfamiliar elephants or artificial warble tones [22]. Yet potential seismic information for elephants is wider than elephant-generated signals, encompassing what may traditionally be considered as noise, but can be used to inform biological decision making, making them cues [24]. What is generally referred to as seismic noise is generated by a mixture of biotic, abiotic and human sources, so can be natural or artificial (human-generated). These seismic vibrations could potentially provide direct reliable information about the seismic source for decision making in elephants, so are potential seismic cues. Relevant seismic sources include elephants or other animals that will generate seismic vibrations as they move around [13,15,25], the processes of the natural environment (e.g. thunder [14]), and humans that purposefully or incidentally generate seismic vibrations (e.g. wind turbines [26]). Elephants could use these seismic vibrations as cues to determine the presence of potential threats, for example from the movement patterns of other animals or from human activity [4,11]. Aside from potentially acting as cues, the detection of seismic noise is also useful for decision making as higher noise decreases communication efficacy [18]. A lower signal-to-noise ratio (SNR; i.e. higher noise level) reduces the ability to detect and discriminate between seismic signals and cues (when the signal magnitude is constant) [15]. Therefore, detecting and responding to seismic noise can allow elephants to mitigate these negative effects (e.g. by using repeated or louder calls [18]) or choose environments where they are more likely to be able to communicate effectively through seismic vibrations (the seismic domain). However, seismic noise is inevitable as it is superimposed during wave propagation [15]. SNR naturally decreases as propagation distance increases, so elephants must have methods to cope with increasing noise levels since they can respond to acoustic rumbles over a variety of distances, up to the kilometre range [12,27]. Despite the potential sources of seismic information, it remains untested whether and how elephants respond to seismic vibrations beyond elephant-generated signals (their infrasonic rumbles) [22,23]. Here, we focus on whether seismic ‘noise’ acts as a cue for elephants, testing how wild elephants respond to seismic vibrations of different noise types (elephant- versus human-generated noise). Since there is increasing scope for anthropogenic seismic noise within the elephants' natural and captive landscapes [28,29], determining whether human-generated seismic noise can act as cues for elephants has important implications for their conservation management. This includes understanding the potential impacts on elephants of land-use change that is expected to increase noise, for example due to infrastructure development (roads, railways, wind farms, etc.) [30], which includes seismic noise [26,31]. 2. Methods (a) Seismic playback tracks and calibration Three 6 min seismic playback tracks of different noise types were generated as stimuli to play to wild elephants in the field (figures 1 and 2a; see also electronic supplementary material, Methods). The first treatment was a source function [15] of recordings of elephant-generated seismic vibrations (Ele). The Ele source function was 6 s long and was repeated thrice (18 s long). The second seismic playback treatment was 18 s of human-generated white noise (WN; random magnitude over time, with many frequencies at equal intensities) generated in MATLAB. This track was used to test whether the source type of the seismic vibration influenced elephant behaviour, where the frequency range was equivalent (electronic supplementary material, figures S1 and S2). For the final treatment, Ele and WN were added together, by adding amplitudes over every time step, to overlay the WN onto the elephant-generated track (Ele&WN), generating a track of longer exposure to simulate elephant-generated vibrations in a noisy environment. Figure 1. Design of playback tracks. (a) Two-minute treatments (light grey) were between 2 min of silence (dark grey) for each playback track. Treatments were three types: (b) elephant-generated seismic vibrations (Ele); (c) white/human-generated noise (WN); (d) combined track of WN superimposed on elephant track (Ele&WN). Colour gives spectral amplitude on a log scale from dark blue (1E−5 V) to yellow (1 V). Note that propagation alters the tracks, as seen when a geophone records the seismic vibrations 10 m from the source (electronic supplementary material, figures S1 and S2). (Online version in colour.) Figure 2. Elephant behavioural responses suggest detection and discrimination of seismic treatments that differ in noise type. (a) Seismic playback set-up, where car 1 plays the seismic tracks to resting elephants (grey) under trees (green) and car 2 records their behaviour within a camera's field of view (indicated with solid black lines), as well as audio and seismic data. Three seismic playback treatments were used: elephant-generated seismic vibrations (Ele; orange), white/human-generated noise (WN; pink) and a combined track of WN superimposed on the elephant track (Ele&WN; blue), where time-voltage for 2 min is shown (+1 V to −1 V). (b) Mean distance moved over time (t) since t = 0 s (n = 7, families where data were collected across all three seismic treatments). Dashed line gives the maximum distance that could be moved since t = 120 s, as capped by the distance-coding method. This method was used to code the data from [32], with acoustic WN (grey squares, n = 15) and acoustic bee noise (grey triangles, n = 17) played from t = 120–360 s. Grey shaded area gives experimental period b, t = 120–240 s, when the seismic treatment is applied. (c) The same seismic treatment data as B, but plotting the mean across each experimental period. Error bars give standard error of the mean between families. Lines, brackets and asterisks denote significant difference from Friedman's test, where p < 0.05. (Online version in colour.) Twelve seconds of silence (for Ele and WN) or WN (for Ele&WN) were added to the end of the 18 s segments in Audacity (freeware) software. These 30 s segments were repeated four times to generate the 2 min seismic treatment period (b). Two minutes of silence were added either side to create the silent control periods (a1 and a2) to finish the 6 min playback tracks (figure 1). For the playback, a custom-built and portable system was used to generate the seismic cues using a partially buried modified speaker (see electronic supplementary material, Methods). The playback system was portable to allow it to be deployed in the field at the elephants’ chosen resting sites under trees. The tracks were played into the speaker from an iPad at full volume (Apple, USA) via an amplifier (Pyle PLMRA400). Tracks were trialled and recorded using a microphone (Earthworks M30 microphone) and vertical geophone (Raspberry Shake 4D, USA; 100 fps) at 3, 10 and 30 m from the playback location (encompassing playback distances to elephants). This was to ensure that there was a high signal in the seismic domain compared to (i) background seismic levels when tracks were silent and (ii) the acoustic domain (electronic supplementary material, figures S1 and S2). The amplitudes of tracks and geophone recordings differed—for the input tracks, amplitudes were similar across all three treatments (figures 1 and 2a). From our recordings, the mean maximum amplitude of seismic vibrations was lower for 10 s of Ele track compared to WN and Ele&WN 10 m from the speaker (5E−6 ± 1E−6, 1.2E−5 ± 4E−6 and 1.2E−5 ± 4E−6 m s−1, respectively, standard deviation given n = 4; electronic supplementary material, figures S1 and S2), which was also seen at 3 m (7E−6 ± 4E−6, 1.3E−5 ± 2E−6, 1.4E−5 ± 2E−6 m s−1) and 30 m (4E−6 ± less than 1E−6, 8E−6 ± 1E−6, 7E−6 ± 2E-6 m s−1) from the speaker. These were on average above background seismic noise levels (4E−6 ± 2E−6, 3E-6 ± 3E−6, 2E−6 ± less than 1E−6 at 3, 10 and 30 m, respectively). (b) Field experiments All institutional and national guidelines for the use of protected animals for scientific research were followed during this project. The project was approved through an ethical review process within the University of Oxford by the Zoology Animal Welfare and Ethical Review Board (ref.: APA/1/5/ZOO/NAPSA/Mortimer/ElephantVibrations) and the project was approved locally via a research permit granted from National Commission for Science, Technology and Innovation, Kenya for research in Samburu county (ref.: NACOSTI/P/16/69501/9147) and via an approved research affiliation with the Kenya Wildlife Service. Park permits were gathered for Samburu and Buffalo Springs National Reserves. Animal welfare standards were ensured during fieldwork via ongoing evaluation of elephant behaviours during experiments. In the field, the seismic tracks were played to wild elephants within the Samburu and Buffalo Springs National Reserves in Kenya (electronic supplementary material, Methods). For each trial, one randomly chosen 6 min track was played to each family/sub-family (table 1). The focal elephant was selected based on their size and visibility. The family or sub-family of the focal elephant was taken to be an independent group of elephants for analysis and were identified both in the field and using photographs. Table 1. Independent families/sub-families sampled during field trials, indicating the order of three seismic playback treatments (1st, 2nd and 3rd indicated with yellow, red and blue respectively) and group size (1–6). The speaker was an average of 17.3 m (standard deviation: 3.9 m, n = 32) from the position of the focal individual at the start of the trial. From reference to our trial experiments (at 3, 10 and 30 m), this propagation distance gives maximum amplitude levels at ca less than 1E−5 m s−1. For approximate comparison, this is under the range of maximum seismic amplitude recorded from elephant steps and car noise (2.9E−5 and 4.4E−5 m s−1; electronic supplementary material, figure S3). To avoid habituation, each family/sub-family was left a minimum of 5 days before another experimental trial with a different treatment [4,20,22]. Therefore, treatment was only 2 min every 6 days (or more) for each family/sub-family group. Sixteen independent families/sub-families were sampled in total, with seven families receiving all three seismic treatments. Group size during playback (defined as individuals within 10 m of the focal individual that was greater than three quarters the size of the largest elephant) varied from 1 to 6 (table 1). During each experiment, video, audio and seismic data were recorded and synchronized. Recording equipment was deployed an average of 8.4 ± 3.6 m (n = 32) from the speaker. The video camera was used to record the behavioural responses of the focal individual (Sony RX II; 30 fps). The geophone and microphone recordings were used as reference data at each field location. (c) Analysis of behaviour To code most behaviours (for distance travelled, see below), each video was cropped into the three 2 min experimental periods (a1, b, a2; figure 1a) and given a random code to eliminate bias. The videos were then analysed blind using BORIS software to extract the durations of behaviours of interest [33], which included vigilant, social, eating and other types of behaviours (electronic supplementary material, table S1). In MATLAB, behaviour durations were converted to time budgets, which was defined as the total duration divided by the time the relevant body part was visible within an experimental period. Distance moved from trial location (i.e. initial focal elephant position) was calculated from frames 30 s apart, which were compared to code whether the focal elephant had moved more than a body length during that period (adding 2 m for yes, or 0 m for no). The distance moved since the start of the trial (t = 0 s) was a cumulative sum over the length of the trial. This allowed a conservative measure of the distance moved from the trial location to be calculated, without estimating the absolute distance moved from the video frames, which would add error. Using this method, there was a maximum (+2 m every 30 s) and a minimum possible distance moved (0 m throughout). The direction of travel was not coded, but for all elephants, distance moved was always away from the speaker, with none coming back towards their initial position (see video data on Dryad data repository). The videos were coded blind by two researchers independently. This method was also applied to the data of King et al. [32], which gives the number of families remaining stationary in response to acoustic WN and the acoustic sound of African bees. The remaining stationary over a 30 s period was coded as +0 m, and not remaining stationary over a 30 s period was coded as +2 m. Again, the direction of travel was not explicitly coded, but elephant responses were away from the trial location [32]. Non-parametric statistics were used to analyse the data, including Friedman's tests, Mann–Whitney U tests and Wilcoxon signed-rank tests, depending on the number of variables being compared and whether there were repeated measures or not (electronic supplementary material, Methods). Data are shown and p-values are given where statistical tests indicated a significant difference, defined as p < 0.05. 3. Results and discussion Our study supported the hypothesis that elephants are able to detect and discriminate between seismic vibrations of different noise types (figure 2b,c). Most notably, we found significant differences in the distance moved by elephants in response to the three seismic treatments (figure 2b,c). We found that treatments containing human-generated noise led to elephants significantly increasing the distance that they moved away from the trial location: in both the WN and combined noise treatment (Ele&WN), the elephants significantly increased the distance moved away from the trial location over the experimental periods (p < 0.01 for both treatments; Friedman's test χ2(2) = 9.29 and 10). Furthermore, the elephants did not respond in the same way to the three treatments: the three treatments differed significantly in the distance moved during the post-treatment period (p = 0.048; Friedman's test χ2(2) = 6.08). Over all the trials, 1/10 (Ele), 2/11 (WN) and 5/10 (Ele&WN) focal elephants left the site during or following treatments, with instances of running away from the trial location for the combined treatment (Ele&WN; electronic supplementary material, movie S1). For the families that were exposed to all three treatments (n = 7), treatment order did not have a significant effect on distance moved: the distance moved did not significantly differ due to whether it was first, second or third track for any of the treatment periods (p = 0.607, 0.466, 0.619; Friedman's test χ2(2) = 1, 1.53, 0.96 for a1, b and a2 periods respectively). This finding differs from previous seismic playback experiments, where warble tone ‘noise’ (with frequency content and duration similar to an alarm rumble) did not elicit a behavioural response [22]. This indicates that the amplitude and/or time of exposure of seismic noise is important to elephants when assessing its risk, which requires further study in the field. Elephants have been shown to increase their distance moved in response to the playback of the acoustic sound of bees, human voices, tiger growls and elephant alarm rumbles [4,20–22,32], as well as an increased number of tourist vehicles [34]. To allow a quantitative comparison, we applied our analytical method (see Methods) to the distance moved in response to the sound of bees and acoustic WN, taken from King et al. [32] (figure 2b). This analysis does not reveal anything about the distance moved by elephants in response to playback treatment beyond the cap of 2 m per 30 s. Interestingly, the distance moved in response to acoustic and seismic WN was similar at 21.0 and 26.0% of the capped maximum and they were not significantly different from each other during any experimental period (p = 0.172, p = 0.639 and p = 0.241; Mann–Whitney U = 165, 166.5, 156). Therefore, elephant responses to acoustic and seismic WN for these two studies were comparable. However, elephants were more likely to move further due to the sound of bees than the combined noise seismic track: whereas the response to bees averaged 86.9% of the maximum capped distance moved since the start of treatment (using our coding method; figure 2b dashed line; where 100% = +2 m every 30 s, 0% = +0 m), the response to Ele&WN we saw here was 48.1%. Therefore, for the sound of bees, almost every elephant moved at least 2 m every 30 s (13/17 from 150 s onwards), whereas fewer elephants showed this response rate to the seismic cue, with a latency of response (4/7 elephant maximum rate starting 180 s or later). We also found significant differences in head scanning and freezing behaviour for the treatments containing human-generated noise (WN and Ele&WN; figure 3). Head scanning and freezing sometimes increased during the periods where the seismic tracks were playing: we found increased head scanning and freezing in the treatment versus pretreatment periods (b versus a1, p = 0.031 and p < 0.01 for WN and Ele&WN head scanning; p < 0.01 for Ele&WN freezing; Wilcoxon signed-rank = 27, 36, 36) and increased freezing in the treatment versus post-treatment periods for WN (b versus a2, p < 0.01; Wilcoxon signed-rank = 36). Both head scanning and freezing are thought to indicate vigilance or listening in elephants [12,19,22], with freezing also thought to promote seismic wave detection in elephants [7]. Elephants have previously been shown to increase vigilant behaviours in response to seismic and acoustic playback of conspecific calls [12,22]. Figure 3. Elephant time budget responses show significant differences in behaviour for control and treatment periods. Three seismic playback treatments were used: elephant-generated seismic vibrations (Ele; orange), white/human-generated noise (WN; pink) and a combined track of WN superimposed on elephant track (Ele&WN; blue). (a) Mean time budget of elephants showing head scanning behaviour during each treatment period (n = 10 families for Ele and Ele&WN, n = 11 for WN): a1 was pretreatment sampling, t = 0–120 s, and a2 was post-treatment sampling, t = 240–360 s. (b) Mean time budget of elephants showing freezing behaviour during each treatment period (sample size as in A). Error bars give standard error of the mean between families. Brackets and asterisks denote significant paired difference from Wilcoxon signed-rank test between treatment periods, where p < 0.05. (Online version in colour.) No other significant differences in behaviours were recorded when the elephants were present, whether trunk swing/manipulation (can be a sign of apprehension [19]) or social and foraging behaviours, although behaviour could not be recorded once the elephant left the site. Compared to acoustic playback of conspecific calls [12,19], there were no differences in ‘listening’ behaviour (ears extended stiff), which supports that elephants respond differently to seismic and acoustic playback of conspecific calls [12,22]. Retreat response in animals is regarded as defensive, indicating an association between the stimulus and increased risk [10,35]. The retreat response is an output of a trade-off between the advantages of the response (risk-avoidance) versus the costs (less time for beneficial activities such as feeding) [4]. In the case of the sound of bees, the elephants consistently retreat to this acoustic stimulus as they associate it with the risk of stinging [20,32,36,37]. In the case of human voices, elephants are more likely to retreat to Maasai men voices as they associate it with a higher threat [4]. This shows that elephants employ high-level discrimination of acoustic vibrational cues to assess relative risk [4,22], which also appears to be the case with seismic vibrational cues (figures 2 and 3) and signals [22]. The trade-off underlying a retreat response is hypothesized to be similar between predation and non-lethal human disturbance [35], of which seismic noise is one example. Our findings suggest that human-generated noise (e.g. Ele&WN) was associated with a higher risk in elephants. The specific risks encoded within these seismic cues that necessitate a quick retreat response are hard to untangle and warrant further research. It could be that elephants associate these seismic cues with unusual or unknown human activity, which they could associate with risk [4,11]; i.e. human activities outside of the elephants' experiences could be associated with higher risk (even if the risk is actually low). For example, known/normal human activity associated with low risk could be seismic cues from tourist vehicles during the day for habituated elephants, whereas unusual/unknown human activity might be larger vehicles at night that may be associated with higher risk. This may also explain why there was not a significant response to the elephant-generated treatment—because they were less ‘unusual’ to the elephants, although the lower seismic amplitude of this treatment could also explain this. Less likely (due to the fast response), it could be that these seismic cues indicate a risk as they are able to mask seismic signals of greater biological importance, or at least reduce the efficacy of necessary communication [18]. Finally, miscategorization of the seismic cues by the elephants cannot be ruled out, where the seismic cues might be misinterpreted as something else elephants associate with risk (e.g. landslide, earthquake). Compared to the sound of bees, the quantitative comparison indicates that the association with risk is not as universal or strongly linked as the risk of bee stings, as fewer individuals showed the retreat response to the combined seismic noise treatment compared to the sound of bees treatment, and the response took longer. Our results indicate that seismic noise is a cue that matters to elephants in the wild: it is a stimulus type that contains information that elephants responded to, in some cases with risk-avoidance responses. Avoidance behaviour in elephants has also been correlated with an increased presence of stress hormone [38]. Furthermore, seismic noise is often an overlooked form of anthropogenic noise [28,29], so elephant behavioural responses to the seismic noise generated by infrastructure development and use (including roads, railways and human settlements) will be an important avenue for future research. Tracking data suggest that African elephants avoid or change the timing and speed of their movements around newly built roads/railways [5]. Compared to our human-generated treatments, an example recording of car noise was the same order of magnitude for the maximum amplitude levels (ca 3E−5 car versus ca 1E−5 m/s Ele&WN), was more sustained (ca 15% points greater than 1E−5 m/s car versus ca 4% WN) and contained similar frequency content (electronic supplementary material, figures S1–S3). This suggests that there will be significant challenges to elephants resulting from their seismic sensitivity as human-generated seismic noise increases in their natural habitats. Overall, we show that the seismic sensory domain contains a wider variety of information for elephants than previously considered, revealing deeper connections between elephants and their dynamic and challenging physical environment. Supplementary Material Click here for additional data file. Acknowledgements All authors thank Iain Douglas-Hamilton, George Wittemeyer and all the staff at Save the Elephants for kindly agreeing to support the fieldwork in Kenya. We thank Lucas Wilkins and John Hogg for their help with the modified speaker set-up. B.M. thanks Lucy Taylor and Tom Mulder for their comments on the manuscript. B.M. thanks Fritz Vollrath for his valuable discussions throughout the study. We thank Tarje Nissen-Meyer for the loan of the Raspberry Shake 4D. Ethics All institutional and national guidelines for the use of protected animals for scientific research were followed during this project. The project was approved through an ethical review process within the University of Oxford by the Zoology Animal Welfare and Ethical Review Board (ref.: APA/1/5/ZOO/NAPSA/Mortimer/ElephantVibrations). Research permit and project approval was granted locally from National Commission for Science, Technology and Innovation, Kenya for research in Samburu county (ref.: NACOSTI/P/16/69501/9147). B.M. had project approval and research affiliation from the Kenya Wildlife Service. Park permits were also gathered for Samburu and Buffalo Springs National Reserves. Animal welfare standards were ensured during fieldwork via ongoing evaluation of elephant risk and behaviours during and after experiments. Specifically, the risk of injury to the elephants was sufficiently low to continue planned experiments. All treatments were non-contact in their nature, so caused no physical harm. Treatment periods were kept to 2 min, repeated up to every 6 days for some family/sub-family groups, representing 2/8640 min over this period. The maximum amplitude seismic levels of the chosen treatments were within the natural range that elephants would encounter: they were within the order of magnitude (and slightly lower) than the seismic levels that elephants can generate themselves and those made by cars. Data accessibility Electronic supplementary material is provided to support this article [39]. This includes electronic supplementary material, figures S1 and S2 showing geophone and microphone recordings of treatments on different deployments (time–velocity plots and spectrograms respectively); electronic supplementary material, figure S3 showing example geophone recordings of elephant steps and car noise; electronic supplementary material, table showing definitions of behaviours used for video analysis; and electronic supplementary material, movie showing an example of elephant responses to human- and elephant-generated seismic vibrations combined treatment (Ele&WN). The cropped video data (videos separated as a1 silent, b treatment and a2 silent segments: 96 videos) collected and used as part of this manuscript and an explanatory spreadsheet (with video name to treatment key, list of focal individuals and where they are on the video) are available from the Dryad Digital Repository: https://doi.org/10.5061/dryad.3tx95X6gb [40]. Authors' contributions B.M.: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, resources, validation, visualization, writing-original draft, writing-review and editing; J.A.W.: formal analysis, investigation, methodology, validation, writing-review and editing; D.S.L.: investigation, methodology, validation, writing-review and editing; M.R.: visualization, writing-review and editing; D.D.: methodology, supervision, validation, writing-review and editing. All authors gave final approval for publication and agreed to be held accountable for the work performed therein. Competing interests We declare we have no competing interests. Funding B.M. thanks the British Ecological Society (grant no. LRB18/1010), the Royal Commission for the Exhibition of 1851, St Anne's College, Oxford and the Royal Society (grant no. URF/R1/191033) for funding. ==== Refs References 1. Soltis J. 2010 Vocal communication in African elephants (Loxodonta africana). Zoo Biol. 29 , 192-209. (10.1002/zoo.20251)19434672 2. Hollister-Smith JA, Alberts SC, Rasmussen LEL. 2008 Do male African elephants, Loxodonta africana, signal musth via urine dribbling? Anim. Behav. 76 , 1829-1841. (10.1016/j.anbehav.2008.05.033) 3. Valenta K, Schmitt MH, Ayasse M, Nevo O. 2021 The sensory ecology of fear: African elephants show aversion to olfactory predator signals. Conserv. Sci. Pract. 3 , e333. (10.1111/csp2.333). 4. McComb K, Shannon G, Sayialel KN, Moss C. 2014 Elephants can determine ethnicity, gender, and age from acoustic cues in human voices. Proc. Natl Acad. Sci. USA 111 , 5433-5438. (10.1073/pnas.1321543111)24616492 5. Okita-Ouma B, Koskei M, Tiller L, Lala F, King L, Moller R, Amin R, Douglas-Hamilton I. 2021 Effectiveness of wildlife underpasses and culverts in connecting elephant habitats: a case study of new railway through Kenya's Tsavo National Parks. Afr. J. Ecol. 00 , 1-17. (10.1111/aje.12873) 6. Garstang M, Davis RE, Leggett K, Frauenfeld OW, Greco S, Zipser E, Peterson M. 2014 Response of African elephants (Loxodonta africana) to seasonal changes in rainfall. PLoS ONE 9 , e108736. (10.1371/journal.pone.0108736)25299514 7. O'Connell-Rodwell CE. 2007 Keeping an ‘Ear’ to the ground: seismic communication in elephants. Physiology 22 , 287-294. (10.1152/physiol.00008.2007)17699882 8. Bouley DM, Alarcon CN, Hildebrandt T, O'Connell-Rodwell CE. 2007 The distribution, density and three-dimensional histomorphology of Pacinian corpuscles in the foot of the Asian elephant (Elephas maximus) and their potential role in seismic communication. J. Anat. 211 , 428-435. (10.1111/j.1469-7580.2007.00792.x)17711421 9. Poole JH, Payne K, Langbauer WR, Moss CJ. 1988 The social contexts of some very low-frequency calls of African elephants. Behav. Ecol. Sociobiol. 22 , 385-392. (10.1007/bf00294975) 10. Sharma N, Prakash V, Kohshima S, Sukumar R. 2020 Asian elephants modulate their vocalizations when disturbed. Anim. Behav. 160 , 99-111. (10.1016/j.anbehav.2019.12.004) 11. Soltis J, King LE, Douglas-Hamilton I, Vollrath F, Savage A. 2014 African elephant alarm calls distinguish between threats from humans and bees. PLoS ONE 9 , e89403. (10.1371/journal.pone.0089403)24586753 12. Langbauer WR Payne KB Charif RA Rapaport, L, Osborn, F. 1991 African elephants respond to distant playbacks of low-frequency conspecific calls. J. Exp. Biol. 157 , 35-46. (10.1242/jeb.157.1.35) 13. O'Connell-Rodwell CE, Arnason BT, Hart LA. 2000 Seismic properties of Asian elephant (Elephas maximus) vocalizations and locomotion. J. Acoust. Soc. Am. 108 , 3066-3072. (10.1121/1.1323460)11144599 14. O'Connell-Rodwell CE, Hart LA, Arnason BT. 2001 Exploring the potential use of seismic waves as a communication channel by elephants and other large mammals. Am. Zool. 41 , 1157-1170. (10.1668/0003-1569(2001)041[1157:etpuos]2.0.co;2) 15. Mortimer B, Rees WL, Koelemeijer P, Nissen-Meyer T. 2018 Classifying elephant behaviour through seismic vibrations. Curr. Biol. 28 , R547-R548. (10.1016/j.cub.2018.03.062)29738725 16. Larom D, Garstang M, Payne K, Raspet R, Lindeque M. 1997 The influence of surface atmospheric conditions on the range and area reached by animal vocalizations. J. Exp. Biol. 200 , 421-431. (10.1242/jeb.200.3.421)9057305 17. Larom D, Garstang M, Lindeque M, Raspet R, Zunckel M, Hong Y, Brassel K, Obeirne S, Sokolic F. 1997 Meterology and elephant infrasound at Etosha National Park, Namibia. J. Acoust. Soc. Am. 101 , 1710-1717. (10.1121/1.418180) 18. Bradbury JW, Vehrencamp SL. 1998 Principles of animal communication. Sunderland, MA: Sinauer Associates. 19. Poole JH. 1999 Signals and assessment in African elephants: evidence from playback experiments. Anim. Behav. 58 , 185-193. (10.1006/anbe.1999.1117)10413556 20. King LE, Soltis J, Douglas-Hamilton I, Savage A, Vollrath F. 2010 Bee threat elicits alarm call in African elephants. PLoS ONE 5 , e10346. (10.1371/journal.pone.0010346)20436682 21. Thuppil V, Coss RG. 2013 Wild Asian elephants distinguish aggressive tiger and leopard growls according to perceived danger. Biol. Lett. 9 , 20130518. (10.1098/rsbl.2013.0518)24026347 22. O'Connell-Rodwell CE, Wood JD, Kinzley C, Rodwell TC, Poole JH, Puria S. 2007 Wild African elephants (Loxodonta africana) discriminate between familiar and unfamiliar conspecific seismic alarm calls. J. Acoust. Soc. Am. 122 , 823-830. (10.1121/1.2747161)17672633 23. O'Connell-Rodwell CE, Wood JD, Rodwell TC, Puria S, Partan SR, Keefe R, Shriver D, Arnason BT, Hart LA. 2006 Wild elephant (Loxodonta africana) breeding herds respond to artificially transmitted seismic stimuli. Behav. Ecol. Sociobiol. 59 , 842-850. (10.1007/s00265-005-0136-2) 24. Geipel I, Smeekes MJ, Halfwerk W, Page RA. 2019 Noise as an informational cue for decision-making: the sound of rain delays bat emergence. J. Exp. Biol. 222 , jeb192005. (10.1242/jeb.192005)30665972 25. Wood JD, O'Connell-Rodwell CE, Klemperer SL. 2005 Using seismic sensors to detect elephants and other large mammals: a potential census technique. J. Appl. Ecol. 42 , 587-594. (10.1111/j.1365-2664.2005.01044) 26. Hu WF, Barthelmie RJ, Letson F, Pryor SC. 2020 Seismic noise induced by wind turbine operation and wind gusts. Seismol. Res. Lett. 91 , 427-437. (10.1785/0220190095) 27. McComb K, Reby D, Baker L, Moss C, Sayialel S. 2003 Long-distance communication of acoustic cues to social identity in African elephants. Anim. Behav. 65 , 317-329. (10.1006/anbe.2003.2047) 28. Roberts L, Laidre ME. 2019 Finding a home in the noise: cross-modal impact of anthropogenic vibration on animal search behaviour. Biol. Open 8 , bio041988. (10.1242/bio.041988)31292133 29. Caorsi V et al. 2019 Anthropogenic substrate-borne vibrations impact anuran calling. Sci. Rep. 9 , 19456. (10.1038/s41598-019-55639-0)31857629 30. Blickley JL, Patricelli GL. 2010. Impacts of Anthropogenic noise on wildlife: research priorities for the development of standards and mitigation. J. Int. Wildl. Law Policy 13 , 274-292. (10.1080/13880292.2010.524564) 31. Green DN, Bastow ID, Dashwood B, Nippress SEJ. 2017 Characterizing broadband seismic noise in central London. Seismol. Res. Lett. 88 , 113-124. (10.1785/0220160128) 32. King LE, Douglas-Hamilton I, Vollrath F. 2007 African elephants run from the sound of disturbed bees. Curr. Biol. 17 , R832-R833. (10.1016/j.cub.2007.07.038)17925207 33. Friard O, Gamba M. 2016 BORIS: a free, versatile open-source event-logging software for video/audio coding and live observations. Methods Ecol. Evol. 7 , 1325-1330. (10.1111/2041-210x.12584) 34. Szott ID, Pretorius Y, Koyama NF. 2019 Behavioural changes in African elephants in response to wildlife tourism. J. Zool. 308 , 164-174. (10.1111/jzo.12661) 35. Frid A, Dill L. 2002 Human-caused disturbance stimuli as a form of predation risk. Conserv. Ecol. 6 , 11. (10.5751/es-00404-060111) 36. King L, Pardo M, Weerathunga S, Kumara TV, Jayasena N, Soltis J, de Silva S. 2018 Wild Sri Lankan elephants retreat from the sound of disturbed Asian honey bees. Curr. Biol. 28 , R64-R65. (10.1016/j.cub.2017.12.018)29374444 37. Ndlovu M, Devereux E, Chieffe M, Asklof K. 2016 Responses of African elephants towards a bee threat: its application in mitigating human–elephant conflict. S. Afr. J. Sci. 112 , 52-56. (10.17159/sajs.2016/20150058) 38. Oduor S, Brown J, Macharia GM, Boisseau N, Murray S, Obade P. 2020 Differing physiological and behavioral responses to anthropogenic factors between resident and non-resident African elephants at Mpala Ranch, Laikipia County, Kenya. PeerJ 8 , e10010. (10.7717/peerj.10010)33062433 39. Mortimer B, Walker JA, Lolchuragi DS, Reinwald M, Daballen D. 2021 Noise matters: elephants show risk-avoidance behaviour in response to human-generated seismic cues. Figshare. 40. Mortimer B, Walker JA, Lolchuragi DS, Reinwald M, Daballen D. 2021 Data from: Noise matters: elephants show risk-avoidance behaviour in response to human-generated seismic cues. Dryad Digital Repository. (10.5061/dryad.3tx95X6gb)
PMC008xxxxxx/PMC8242928.txt
==== Front Proc Biol Sci Proc Biol Sci RSPB royprsb Proceedings of the Royal Society B: Biological Sciences 0962-8452 1471-2954 The Royal Society 34187189 10.1098/rspb.2021.1021 rspb20211021 10011442203Behaviour Research Articles Win-Stay-Lose-Shift as a self-confirming equilibrium in the iterated Prisoner’s Dilemma Win-Stay-Lose-Shift as a self-confirming equilibrium in the iterated Prisoner’s Dilemma Kim Minjae Formal analysis Investigation Visualization 1 Choi Jung-Kyoo Conceptualization Supervision Validation 2 http://orcid.org/0000-0002-4143-1187 Baek Seung Ki [email protected] Methodology Project administration Writing-original draft 1 1 Department of Physics, Pukyong National University, Busan 48513, Korea 2 Department of Economics, Kyungpook National University, Daegu 41566, Korea 30 6 2021 June 30, 2021 30 6 2021 June 30, 2021 30 6 2021 June 30, 2021 288 1953 2021102130 4 2021 April 30, 2021 4 6 2021 June 4, 2021 © 2021 The Authors. 2021 https://creativecommons.org/licenses/by/4.0/ Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited. Evolutionary game theory assumes that players replicate a highly scored player’s strategy through genetic inheritance. However, when learning occurs culturally, it is often difficult to recognize someone’s strategy just by observing the behaviour. In this work, we consider players with memory-one stochastic strategies in the iterated Prisoner’s Dilemma, with an assumption that they cannot directly access each other’s strategy but only observe the actual moves for a certain number of rounds. Based on the observation, the observer has to infer the resident strategy in a Bayesian way and chooses his or her own strategy accordingly. By examining the best-response relations, we argue that players can escape from full defection into a cooperative equilibrium supported by Win-Stay-Lose-Shift in a self-confirming manner, provided that the cost of cooperation is low and the observational learning supplies sufficiently large uncertainty. evolution of cooperation , reciprocity , Bayesian inference , Win-Stay-Lose-Shift , observational learning Ministry of Education http://dx.doi.org/10.13039/501100002701 2020R1A6A3A13075972 2020R1I1A2071670 cover-dateJune 30, 2021 ==== Body pmc1. Introduction Evolutionary game theorists often assume that behavioural traits can be genetically transmitted across generations [1]. Along this line, researchers have investigated the genetic basis of cooperative behaviour [2,3]. However, humans learn many culture-specific behavioural rules through observational learning [4], and this mechanism mediates ‘cultural’ transmission that has been proved to exist among a number of non-human animals as well [5,6]. The mirror neuron research suggests that the primate brain may even have a specialized circuit for imitating each other’s behaviour, which facilitates social learning [7–9]. In comparison with the direct genetic transmission, the non-genetic inheritance through social learning can provide better adaptability by responding faster to environmental changes [10]. In contrast with genetic inheritance, however, observational learning may lead to imperfect mimicry if observation is not sufficiently informative or involved with a systematic bias. The notion of self-confirming equilibrium (SCE) has been proposed by incorporating such imperfectness of observation in learning [11]: when an SCE strategy is played, some of the possible information sets may not be reached, so the players do not have exact knowledge but only certain untested belief about what their co-players would do at those unreached sets. It is nevertheless sustained as an equilibrium in the sense that no player can expect a better payoff by unilaterally deviating from it once given such belief, and that the beliefs do not conflict with observed moves. Dynamics of learning based on a limited set of information has been investigated in the context of the coordination game [12,13], in which the opponent’s observed decision is assumed to be his or her strategy. However, the subtlety of cultural transmission manifests itself clearly when a strategy is regarded as a decision rule, hidden from the observer, rather than the decision itself. In this work, we investigate the iterated Prisoner’s Dilemma (PD) game among players with memory-one strategies, who infer the resident strategy from observation and optimizes their own strategies against it. By memory-one, we mean that a player refers to the previous round to choose a move between cooperation and defection [14]. If we restrict ourselves to memory-one strategies, it is already well known in evolutionary game theory that ‘Win-Stay-Lose-Shift (WSLS)’ [15–17] can appear through mutation and take over the population from defectors if the cost of cooperation is low [14]. Compared with such an evolutionary approach, we will impose ‘less bounded’ rationality in that our players are assumed to be capable of computing the best response to a given strategy within the memory-one pure-strategy space. We will identify the best-response dynamics in this space and examine how the dynamics should be modified when observational learning introduces uncertainty in Bayesian inference about strategies. If every player exactly replicated each other’s strategy, full defection would be a Nash equilibrium (NE) for any cost of cooperation. Under uncertainty in observation, however, our finding is that defection is not always an SCE so that the population can move to a cooperative equilibrium supported by WSLS, which is both an SCE and an NE and can thus be called a SCENE. 2. Method and result (a) Best-response relations without observational uncertainty Let us define the one-shot PD game in the following form: 2.1 where we abbreviate cooperation and defection as C and D, respectively, and c is the cost of cooperation assumed to be 0 < c < 1. In this work, the game of equation (2.1) will be repeated indefinitely. Furthermore, the environment is noisy: Even if a player intends to cooperate, it can be misimplemented as defection, or vice versa, with probability ε. In the analysis below, we will take ε as an arbitrarily small positive number. We will restrict ourselves to the space of memory-one (M1) pure strategies. By a M1 pure strategy, we mean that it chooses a move between C and D as a function of the two players’ moves in the previous round. We thus describe such a strategy as [pCC, pCD, pDC, pDD], where pXY = 1 means that C is prescribed when the players did X and Y, respectively, in the previous round, and pXY = 0 if D is prescribed in the same situation. Note that the initial move in the first round is irrelevant to the long-term average payoff in the presence of error so that it has been discarded in the description of a strategy. The set of M1 pure strategies, denoted by Δ, contains 16 elements from d0 ≡ [0, 0, 0, 0] to d15 ≡ [1, 1, 1, 1]. Let us assume that a player, say, Alice, takes a M1 pure strategy dα as her strategy. The noisy environment effectively modifies her behaviour to 2.2 sAϵ≡(1−ϵ) dα+ϵ(1−dα) as if she were playing a mixed strategy, where 1 ≡ [1, 1, 1, 1]. Likewise, Alice’s co-player Bob chooses dβ, and his effective behaviour is described by 2.3 sBϵ≡(1−ϵ)dβ+ϵ(1−dβ). The repeated interaction between Alice and Bob is Markovian, and it is straightforward to obtain the stationary probability distribution 2.4 v(dα,dβ,ϵ)=(vCC,vCD,vDC,vDD), where vXY means the long-term average probability to observe Alice and Bob choosing X and Y, respectively [18–20] (see appendix A for more details). The presence of ε > 0 guarantees the uniqueness of v. Alice’s long-term average payoff against Bob is then calculated as 2.5 Π(dα,dβ,ϵ)=v⋅P, where P ≡ (1 − c, − c, 1, 0) is a payoff vector corresponding to equation (2.1). As long as Alice can exactly identify Bob’s strategy dβ with no observational uncertainty, she can find the best response to Bob within the set of M1 pure strategies by applying every dα∈Δ to equation (2.5). In table 1, we list the best response to each strategy in Δ in the limit of small ε (see also figure 1 for its graphical representation). In most cases, the best-response dynamics ends up with d0 = [0, 0, 0, 0], which is the best response to itself and often called Always-Defect (AllD). For example, if we start with Tit-for-Tat (TFT), represented as d10 = [1, 0, 1, 0], table 1 shows that the best response to TFT within Δ is Always-Cooperate (AllC), represented as d15 = [1, 1, 1, 1], to which AllD is the best response for obvious reasons. Table 1. Best response among M1 pure strategies. Against each strategy in the first column, we obtain the best response (the second column), and the resulting average payoff (equation (2.5)) earned by the best response is given as a power series of ε in the third column. In the second column, we have placed a dagger next to a strategy when it is the best response to itself. opponent strategy best response payoff of the best response to the opponent strategy Misc. d0 d0† (1 − c)ε AllD d1 d0 1/2 − (1/4 + c)ε + O(ε2) d2 d11 (1 − c)/2 − (1 + c)ε/2 + O(ε2) d3 d0 1/2 − ce + O(ε3) d4 d0 1/3 + (2/9 − c)ε + O(ε2) d5 d0 1 − (2 + c)ε + O(ε2) d6 d9 1 − 3(1 + c)ε + O(ε2) d7 d0 1 − (2 + c)ε + 4ε2 + O(ε3) d8 {d8†,c>1/3d15,c<1/3} {3(1−c)ϵ/2+O(ϵ2)1/3−c+O(ϵ)} GT1 d9 {d0,c>1/2d9†,c<1/2} {1/2+O(ϵ)1−c+O(ϵ)} WSLS d10 d15 (1 − c) − (2 − c)ε + O(ε2) TFT d11 {d0,c>1/2d13,c<1/2} {1/2+(1/4−c)ϵ+O(ϵ2)(1−c)−(2−c)ϵ+O(ϵ2)} d12 d0 1/2 + O(ε) d13 d0 1 − (1 + c)ε + O(ε2) d14 d1 1 − 2(1 + c)ε + O(ε2) d15 d0 1 − (1 + c)ε + O(ε3) AllC Figure 1. Graphical representation of best-response relations in table 1. If dμ is the best response to dν, we represent it as an arrow from dν to dμ. The blue node (Win-Stay-Lose-Shift) means an efficient NE with 1 − vCC ∼ O(ε), whereas the red nodes (Always-Defect and M1 Grim Trigger) mean inefficient ones with vCC≲O(ϵ) as shown in table 2. (Online version in colour.) However, two exceptions exist: The first one is d8 = [1, 0, 0, 0], which we may call M1 Grim Trigger (GT1). If c > 1/3, this strategy is the best response to itself, and it is an inefficient equilibrium giving each player an average payoff of O(ε). The other exception is WSLS, represented by d9 = [1, 0, 0, 1], which is the best response to itself when c ≤ 1/2. It is an efficient NE, at which each player earns 1 − c + O(ε) per round on average. (b) Observational learning Now, let us imagine a monomorphic population of players who have adopted a strategy dγ in common. The population is in equilibrium in the sense that a large ensemble of their states XY ∈ {CC, CD, DC, DD} can represent the stationary probability distribution v(dγ,dγ,ϵ). We have an observer, say, Alice, with a potential strategy dα. As we learn social norms in childhood, it is assumed that Alice does not yet participate in the game but has a learning period to observe M ( ≫1) pairs of players, all of whom have used the resident strategy dγ. How their mind works is a black box to her: Just by observing their states XY and subsequent moves, Alice has to form belief about dγ, based on which she chooses her own strategy dα to maximize the expected payoff. If Alice’s optimal strategy turns out to be identical to the resident strategy dγ, it constitutes an SCE. To see how Alice can specify dγ∈Δ from observation, let us consider an example that the observed probability distribution over states XY is best described as v ≈ (0, 1/4, 1/4, 1/2). If Alice has computed v for every strategy in Δ as listed in table 2, the observation suggests that the resident strategy is unlikely to be TFT (d10 = [1, 0, 1, 0]) because the corresponding stationary distribution would be v = (1/4, 1/4, 1/4, 1/4). She finds that dγ can be either d2 = [0, 0, 1, 0] or d4 = [0, 1, 0, 0]. To distinguish between them, she has to check how people react to CD or DC. According to table 2, these states will be observed frequently because vCD = vDC = 1/4. Thus, in this example, Alice succeeds in identifying dγ as long as M ≫ 1. Eight strategies have this property, constituting Category I in Δ (table 2). As another example, if v ≈ (1/2, 0, 0, 1/2), Alice sees that dγ must be either d1 = [0, 0, 0, 1] or d7 = [0, 1, 1, 1]. To resolve the uncertainty, she has to further check how people react to CD or DC, but she may actually save this effort because the best response turns out to be d0 in either case (table 1). This is the case of Category II in Δ (table 2). Table 2. Stationary probability distribution v(dγ,dγ,ϵ), where we have retained only the leading-order term in the ε-expansion for each vXY. When we describe a strategy in binary, the boldface digits are the ones that are frequently observed with vXY ∼ O(1) and thus readily identifiable as long as M ≫ 1. In this table, the eight strategies in Category I have three or four such digits, so if the population is using one of these strategies, Alice can tell which one is being played after M ( ≫ 1) observations. As for Category II, the member strategies d1 and d7 would be indistinguishable if M ≪ ε−1 because they differ at their non-boldface digits. Still, Alice can find the best response d0 which is common to both of them (table 1). In Category III, each member strategy has just one boldface digit, so the strategies as well as the best responses can be identified only if M ≫ ε−1. category strategy vCC vCD vDC vDD I d3=[0,0,1,1] 14 14 14 14 d5=[0,1,0,1] d10=[1,0,1,0] d12=[1,1,0,0] d2=[0,0,1,0] 12ϵ 14 14 12 d4=[0,1,0,0] d11=[1,0,1,1] 12 14 14 12ϵ d13=[1,1,0,1] II d1=[0,0,0,1] 12 ε ε 12 d7=[0,1,1,1] III d0=[0,0,0,0] ε2 ε ε 1 d6=[0,1,1,0] 2ε ε ε 1 d8=[1,0,0,0] 12ϵ ε ε 1 d9=[1,0,0,1] 1 ε ε 2ε d14=[1,1,1,0] 1 ε ε 12ϵ d15=[1,1,1,1] 1 ε ε ε2 In general, the first important piece of information to infer dγ is the stationary distribution v because it heavily depends on dγ (table 2). However, the information of v may be insufficient to single out the answer: Suppose that v gives multiple candidate strategies which prescribe different moves at a certain state XY and thus have different best responses. Alice then needs to observe what players actually choose at XY, and such observations should be performed sufficiently many times, i.e. M vXY ≫ 1, for the sake of statistical power. If we check every dγ∈Δ one by one in this way, we see that the best response to the resident strategy can readily be identified as long as M ≫ ε−1, in which case the result of observational learning would be the same as that of exact identification of strategies. If M ≪ ε−1, on the other hand, Alice cannot fully resolve such uncertainty through observation. Still, note that M should be taken as far greater than O(1) for statistical inference to be meaningful. Furthermore, ε has been introduced as a regularization parameter whose exact magnitude is irrelevant, so we look at the behaviour in the limit of small ε. When 1 ≪ M ≪ ε−1, uncertainty in the best response remains only when v ≈ (0, 0, 0, 1) or (1, 0, 0, 0), both of which are characteristic of Category III in table 2. In the former case, d0, d6 and d8 are the candidate strategies for dγ, whereas in the latter case, the candidates are d9, d14 and d15. From the Bayesian perspective, it is reasonable to assign equal probability to each of the candidate strategies. However, if Mε ≪ 1, the number of observations cannot be enough to update this prior probability (see appendix B for a detailed discussion). Therefore, when v ≈ (0, 0, 0, 1), yielding dγ=d0 or d6 or d8, Alice tries to maximize the expected payoff 2.6 Π¯α=Π(dα,d0,ϵ)+Π(dα,d6,ϵ)+Π(dα,d8,ϵ)3, and the calculation shows that it can be achieved by playing 2.7 {d8,if c>1633d9,if c<1633 in the limit of ε → 0. Likewise, when v ≈ (1, 0, 0, 0), yielding dγ=d9 or d14 or d15, Alice tries to maximize her expected payoff from the three possibilities, which is achieved when she plays 2.8 {d1,if c>29d9,if c<29 as ε → 0. Now, AllD ceases to be the best-looking response to itself (figure 2): The expected payoff against AllD will be higher when WSLS is played, if c < 16/33. On the other hand, if we consider a WSLS population with c < 2/9, its cooperative equilibrium is protected from invasion of defectors because Alice under observational uncertainty will keep choosing WSLS, which is truly the best response to itself. Figure 2. Best-looking responses to maximize the expected payoff under uncertainty in observation, when 1 ≪ M ≪ ε−1. Compared with figure 1, the first difference is that Alice uses equation (2.7) against d0, d6 and d8. In addition, she will use equation (2.8) against d9, d14 and d15. (Online version in colour.) The above analysis concerns the uniform prior among three candidate strategies in each case. Let fi denote the fraction of di. For an observer who almost always sees defection from the population, the prior in equation (2.6) can be written as (f0, f6, f8) = (1/3, 1/3, 1/3). For a general prior (f0, f6, f8) with 0 < fi < 1 and f8 = 1 − f0 − f6, the condition for WSLS to give the highest expected payoff is summarized as the intersection of the following two inequalities (figure 3a): 2.9 f6>13f8−(5c4+3c) and 2.10 f6>(3c2+3c)−35(2−c2+3c)f8. The above inequalities are written for f6 because it is d6 that has WSLS as the best response (table 1). If c > 1/3, the former inequality becomes trivial because of the positivity of f6. Note that WSLS still gives the highest expected payoff for a significant part of the simplex even when the cost of cooperation is as high as c = 0.9 (figure 3b). Figure 3. Effect of the prior on the observer’s choice. A point in the triangle represents three fractions, which sum up to one, and its distance to an edge is proportional to the fraction of the strategy at the opposite vertex [21]. (a) When the observer sees nearly defection only, the prior takes the form of (f0, f6, f8), for which we can find the strategy that gives the best expected payoff as written in each region. When c is low, d9 (WSLS) gives the highest expected payoff for most of the prior. (b) Even when the cost increases to c = 0.9, the observer should choose WSLS if the prior contains a sufficiently high fraction of d6. (c) If the observer sees cooperation almost all the time, the prior can be expressed as (f9, f14, f15). If c is low, WSLS can be the observer’s choice when f9 is high enough. (d) The region of WSLS disappears as c exceeds 1/2, and the only possible choice is between d1 and d0 (AllD). Similarly, we can check what an observer would conclude after observing nearly cooperation only, although it is of less importance compared with the above case of a defecting population (figure 2). For a general prior represented by (f9, f14, f15), where f14 = 1 − f9 − f15, WSLS gives the highest expected payoff when 2.11 f9>(c1−c)(1+ f152), as can be seen in figure 3c. This inequality can be satisfied only if c ≤ 1/2: Otherwise, it is better to be a defector by playing d0 or d1 (figure 3d). 3. Summary and discussion In summary, we have investigated the iterated PD game in terms of best-response relations and checked how it is modified by observational learning. Thereby we have addressed a question about how cooperation is affected by cultural transmission, which may be systematically involved with observational uncertainty. The notion of SCE takes this systematic uncertainty into account, and its intersection with NE can be an equilibrium refinement. It is worth pointing out the following: If everyone plays a certain strategy di with belief that everyone else does the same, the whole situation is self-consistent in the sense that observation will always confirm the belief, which in turn agrees with the actual behaviour. The importance of SCENE becomes clear when someone happens to play a different strategy or begins to doubt the belief: If di is not an NE, the player will benefit from the deviant behaviour and reinforce it. If di is not an SCE, the player may fail to dispel the doubt, which will undermine the prevailing culture. Therefore, the strategy has to be a SCENE for being transmitted in a stable manner through observational learning. As a reference point, we have started with the conventional assumption that one can identify a strategy without uncertainty, and checked the best-response relations within the set of M1 pure strategies. Our finding is that a symmetric NE is possible if one uses one of the following three strategies: AllD, GT1 and WSLS (figure 1). Only the last one is efficient. Although we have restricted ourselves to pure strategies, we can discuss the idea behind it as follows: Let us consider a monomorphic population playing a mixed strategy q = [qCC, qCD, qDC, qDD], where each element means the probability to cooperate in a given circumstance. Such a mixed strategy can be represented as a point inside a four-dimensional unit hypercube. The observer seeks the best response to it, say, p = [pCC, pCD, pDC, pDD]. Suppose that p also turns out to be a mixed strategy, say, containing dk and dl with k ≠ l. According to the Bishop–Cannings theorem [22], it implies that 3.1 Π(dk,q,ϵ)=Π(dl,q,ϵ), and this equality imposes a set of constraints on q, rendering the dimensionality of the solution manifold lower than four. Therefore, to almost all q in the four-dimensional hypercube, only one pure strategy will be found as the best response. In appendix C, we provide an explicit proof for this argument in case of reactive strategies. Even if our theoretical framework of Bayesian best-response dynamics is an idealization, we believe that it captures certain aspects of animal behaviour. For example, although the best-response dynamics per se shows poor performance in explaining learning behaviour because of its deterministic character [23], its modified versions can provide reasonable description for experimental results [24,25]. In addition, some studies show that Bayesian updating yields consistent results with observed behaviour of animals, including mammals, birds, a fish and an insect, in the foraging and reproduction activities [26]. These studies support the Bayesian brain hypothesis, which argues that the brain has to successfully simulate the external world in which Bayes’ theorem holds [27]. We also point out that the posterior can be calculated correctly even if the observer has short-term memory as implied by the M1 assumption: As long as input observations are exchangeable with each other, Bayesian updating can be done in a sequential manner (i.e. by modifying the prior little by little every time a new observation arrives), and it is mathematically equivalent to a batch update that uses all the observations at once. To conclude, if we take observational learning into consideration, our result suggests that WSLS can be a SCENE to a Bayesian observer, whereas AllD cannot under observational uncertainty. That is, if the number of observations is too small to see how to behave after error, the uncertainty provides a way to escape from full defection, whereas WSLS can still maintain cooperation: The point is that AllD is not easy to learn by observing defectors because it is difficult to tell what they would choose if someone actually cooperated. WSLS is also difficult to learn, but the uncertainty works in an asymmetric way because one can expect more from mutual cooperation than from full defection by the very definition of the PD game. Supplementary Material Click here for additional data file. Data accessibility Source codes used in this work are available from the Dryad Digital Repository: https://doi.org/10.5061/dryad.n02v6wwwz [28]. Authors' contributions J.-K.C. conceived of the study. S.K.B. designed the study and wrote the paper. M.K. analysed the model. All authors gave final approval for publication and agree to be held accountable for the work performed therein. Competing interests We declare we have no competing interests. Funding M.K. was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (NRF-2020R1A6A3A13075972). S.K.B. was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (NRF-2020R1I1A2071670). Appendix A. Stationary distribution Let us consider two players, Alice and Bob, playing the PD game repeatedly. As written in equation (2.2), Alice’s effective behaviour in the noisy environment is described by a mixed strategy sAϵ=(qCC,qCD,qDC,qDD), where qXY ∈ {ε, 1 − ε} denotes Alice’s probability of cooperation when she and Bob did X and Y, respectively, in the previous round. In the same manner, another mixed strategy sBϵ=(rCC,rCD,rDC,rDD) applies to Bob’s effective behaviour (equation (2.3)), where rXY ∈ {ε, 1 − ε} denotes Bob’s probability of cooperation when he and Alice did X and Y, respectively, in the previous round. Let vXY(t) be the probability to see Alice and Bob choosing X and Y, respectively, in round t. The condition vCC(t)+vCD(t)+vDC(t)+vDD(t)=1 is satisfied all the time. The probability distribution v(t)≡(vCC(t),vCD(t),vDC(t),vDD(t)) evolves as v(t+1) = Wv(t) with A 1 W=[qCCrCCqCDrDCqDCrCDqDDrDDqCCr¯CCqCDr¯DCqDCr¯CDqDDr¯DDq¯CCrCCq¯CDrDCq¯DCrCDq¯DDrDDq¯CCr¯CCq¯CDr¯DCq¯DCr¯CDq¯DDr¯DD], where q¯XY≡1−qXY and r¯XY≡1−rXY. Note that it is a positive stochastic matrix for ε > 0. According to the Perron–Frobenius theorem, it has a unique largest eigenvalue 1, and the corresponding eigenvector can be chosen to have positive entries. Thus, by solving Wv = v, we can obtain the stationary distribution v = (vCC, vCD, vDC, vDD). Each element vXY can be interpreted as the long-time average frequency of XY, and it can readily be expanded as a Taylor series in terms of ε. To determine the best response to dβ as shown in table 1, we calculate the long-term average payoff of dα against it for every α ∈ {0, …, 15} (equation (2.5)) and compare the Taylor-expanded expressions order by order. As for table 2, we set α = β and retain only the leading order terms in the Taylor series for v. Appendix B. Bayesian inference To illustrate the inference procedure, let us assume that v ≈ (0, 0, 0, 1) is given to Alice. She has a set of candidate strategies Λ≡{d0,d6,d8} for the resident strategy q. Alice assigns equal prior probability to each of these candidate strategies. In a certain round t, she observes interaction between Eve and Frank both of whom use q. Let Et and Ft denote Eve’s and Frank’s moves, respectively, in round t. If Alice sees Eve cooperate (i.e. Et = C) after St−1 ≡ (Et−1, Ft−1) = (C, C), she may use this additional information in a Bayesian way to calculate the posterior probability of q = d0 as follows: B 1 P(q=d0|Et,St−1)=P(Et|St−1,d0)P(St−1|d0)P(d0)∑di∈ΛP(Et|St−1,di)P(St−1|di)P(di) B 2 =ϵ⋅ϵ2⋅(1/3)ϵ⋅ϵ2⋅(1/3)+ϵ⋅(2ϵ−5ϵ2+4ϵ3)⋅(1/3)+ϵ⋅ϵ/2⋅(1/3), where P(Et|St−1, di) is directly obtained from di, and P(St−1|di) is taken from the stationary probability distribution v. This posterior probability is used as prior probability for the next observation. If q is actually d6, the average number of times to observe Et = C after St−1 = (C, C) will be B 3 MP(Et,St−1|q=d6)=MP(Et|St−1,d6)P(St−1|d6). In this way, Alice obtains the final posterior probability of q = d0 after observing interaction between M pairs of players, when their actual strategy is d6. If ε is fixed as a small positive value, this inference procedure approaches the correct answer as M → ∞. The effect of observational uncertainty manifests itself when Mε ≪ 1. For example, we may choose M ≈ ε−1/2 as a representative value for 1 ≪ M ≪ ε−1 and check various values of ε from 10−2 to 10−6. Then, the above calculation confirms that the posterior probabilities should remain identical to the prior ones due to the lack of observation. Appendix C. Best-response relations among reactive strategies Let us consider two reactive strategies p=[p C,p D,p C,p D] and q=[q C,q D,q C,q D]. The long-term average payoff of p against q is C 1 Π=(p Dq C−p Dq D+q D)−c(p D+p Cq D−p Dq D)1−(p C−p D)(q C−q D) in the limit of ε → 0. After some algebra, we find the following: First, if qC − qD > c, both ∂Π/∂pC and ∂Π/∂pD are positive, so the best response is given by p C=p D=1. Or, if qC − qD < c, both ∂Π/∂pC and ∂Π/∂pD are negative, so the best response is given by p C=p D=0. Note that we have neglected the measure-zero line defined by qC − qD = c, on which the best response is not uniquely determined. ==== Refs References 1. Maynard Smith J. 1982 Evolution and the theory of games. Cambridge, UK: Cambridge University Press. 2. Kasper C et al. 2017 Genetics and developmental biology of cooperation. Mol. Ecol. 26 , 4364-4377. (10.1111/mec.14208)28626971 3. Manfredini F, Brown MJ, Toth AL. 2018 Candidate genes for cooperation and aggression in the social wasp Polistes dominula. J. Comp. Physiol. A 204 , 449-463. (10.1007/s00359-018-1252-6) 4. Bandura A. 1977 Social learning theory. Englewood Cliffs, NJ: Prentice Hall. 5. Krützen M, Mann J, Heithaus MR, Connor RC, Bejder L, Sherwin WB. 2005 Cultural transmission of tool use in bottlenose dolphins. Proc. Natl Acad. Sci. USA 102 , 8939-8943. (10.1073/pnas.0500232102)15947077 6. Frith CD, Frith U. 2012 Mechanisms of social cognition. Annu. Rev. Psychol. 63 , 287-313. (10.1146/annurev-psych-120710-100449)21838544 7. Di Pellegrino G, Fadiga L, Fogassi L, Gallese V, Rizzolatti G. 1992 Understanding motor events: a neurophysiological study. Exp. Brain Res. 91 , 176-180. (10.1007/BF00230027)1301372 8. Gallese V, Fadiga L, Fogassi L, Rizzolatti G. 1996 Action recognition in the premotor cortex. Brain 119 , 593-609. (10.1093/brain/119.2.593)8800951 9. Ferrari PF, Rizzolatti G. 2014 Mirror neuron research: the past and the future. Phil. Trans. R. Soc. B 369 , 20130169. (10.1098/rstb.2013.0169)24778369 10. Leimar O, McNamara JM. 2015 The evolution of transgenerational integration of information in heterogeneous environments. Am. Nat. 185 , E55-E69. (10.1086/679575)25674697 11. Fudenberg D, Levine DK. 1998 The theory of learning in games. Cambridge, MA: MIT Press. 12. Sandholm WH. 2001 Almost global convergence to p-dominant equilibrium. Int. J. Game Theory 30 , 107-116. (10.1007/s001820100067) 13. Kreindler GE, Young HP. 2013 Fast convergence in evolutionary equilibrium selection. Games Econ. Behav. 80 , 39-67. (10.1016/j.geb.2013.02.004) 14. Baek SK, Jeong HC, Hilbe C, Nowak MA. 2016 Comparing reactive and memory-one strategies of direct reciprocity. Sci. Rep. 6 , 25676. (10.1038/srep25676)27161141 15. Kraines D, Kraines V. 1989 Pavlov and the prisoner’s dilemma. Theory Decis. 26 , 47. (10.1007/BF00134056) 16. Nowak M, Sigmund K. 1993 A strategy of win-stay, lose-shift that outperforms tit-for-tat in the Prisoner’s Dilemma game. Nature 364 , 56-58. (10.1038/364056a0)8316296 17. Imhof LA, Fudenberg D, Nowak MA. 2007 Tit-for-tat or win-stay, lose-shift? J. Theor. Biol. 247 , 574-580. (10.1016/j.jtbi.2007.03.027)17481667 18. Nowak M. 1990 Stochastic strategies in the prisoner’s dilemma. Theor. Popul. Biol. 38 , 93-112. (10.1016/0040-5809(90)90005-G) 19. Nowak MA, Sigmund K, El-Sedy E. 1995 Automata, repeated games and noise. J. Math. Biol. 33 , 703-722. (10.1007/BF00184645) 20. Press WH, Dyson FJ. 2012 Iterated Prisoner’s Dilemma contains strategies that dominate any evolutionary opponent. Proc. Natl Acad. Sci. USA 109 , 10 409-10 413. (10.1073/pnas.1206569109) 21. Harper M et al . 2015 python-ternary: ternary plots in Python. Zenodo. (10.5281/zenodo.594435) 22. Bishop D, Cannings C. 1978 A generalized war of attrition. J. Theor. Biol. 70 , 85-124. (10.1016/0022-5193(78)90304-1)564432 23. Nagel R, Tang FF. 1998 Experimental results on the centipede game in normal form: an investigation on learning. J. Math. Psychol. 42 , 356-384. (10.1006/jmps.1998.1225)9710555 24. Van Huyck JB, Battalio RC, Rankin FW. 1997 On the origin of convention: evidence from coordination games. Econ. J. 107 , 576-596. (10.1111/j.1468-0297.1997.tb00028.x) 25. Cheung YW, Friedman D. 1997 Individual learning in normal form games: some laboratory results. Games Econ. Behav. 19 , 46-76. (10.1006/game.1997.0544) 26. Valone TJ. 2006 Are animals capable of Bayesian updating? An empirical review. Oikos 112 , 252-259. (10.1111/j.0030-1299.2006.13465.x) 27. Friston K. 2012 The history of the future of the Bayesian brain. NeuroImage 62 , 1230-1233. (10.1016/j.neuroimage.2011.10.004)22023743 28. Baek SK, Kim M, Choi JK. 2021 Data from ‘Win-Stay-Lose-Shift as a self-confirming equilibrium in the iterated Prisoner’s Dilemma’. Dryad Digital Repository. (10.5061/dryad.n02v6wwwz)
PMC008xxxxxx/PMC8356502.txt
==== Front Campbell Syst Rev Campbell Syst Rev 10.1002/(ISSN)1891-1803 CL2 Campbell Systematic Reviews 1891-1803 John Wiley and Sons Inc. Hoboken 37131510 10.1002/cl2.1055 CL21055 Campbell Title Registrations Campbell Title Registrations Campbell Title Registrations to Date ‐ September 2019 03 10 2019 9 2019 15 3 10.1002/cl2.v15.3 e1055© 2019 The Authors. Campbell Systematic Reviews published by John Wiley & Sons Ltd on behalf of The Campbell Collaboration. https://creativecommons.org/licenses/by/4.0/ This is an open access article under the terms of the http://creativecommons.org/licenses/by/4.0/ License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. source-schema-version-number2.0 cover-dateSeptember 2019 details-of-publishers-convertorConverter:WILEY_ML3GV2_TO_JATSPMC version:6.3.2 mode:remove_FC converted:14.07.2023 Campbell Title Registrations to Date ‐ September 2019. Campbell Systematic Reviews. 2019;15 :e1055. 10.1002/cl2.1055 37131510 ==== Body pmc1 Details of new titles for systematic reviews or evidence and gap maps that have been accepted by the Editor of a Campbell Coordinating Group are published in each issue of the journal. If you would like to receive a copy of the approved title registration form, please send an email to the Managing Editor of the relevant Coordinating Group. 1 BUSINESS AND MANAGEMENT Work autonomy and management control as predictors of well‐being and performance Jonny Gifford, Rossella Barilli, Niklas Frewel 26 June 2019 Performance pay and employee health Karsten Albæk, Tine Jeppesen, Trine Filges, Bjørn Christian Arleth Viinholt 31 May 2019 Work autonomy and management control as predictors of well‐being and performance Jonny Gifford, Rossella Barilli, Niklas Frewel 19 June 2019 2 CRIME AND JUSTICE What are the effects of different elements of media on radicalisation outcomes? Michael Wolfowicz, Badi Hasisi, David Weisburd 21 June 2019 3 EDUCATION A systematic review of mobile device use in the primary school classroom and impact on pupil literacy and numeracy attainment Claire Dorris, Liam O'Hare, Karen Winter 19 July 2019 4 INTERNATIONAL DEVELOPMENT Impact of conservation agricultural interventions on farm outcomes among smallholder farmers in sub‐Saharan Africa: A systematic review and meta‐analysis Edward N. Mwavu, Onan Mulumba, Paul Mukwaya, Enock Ssekuubwa, Obuku Ekwaro, Moses Ocan 24 July 2019 A map of development evaluations conducted in Uganda 2000‐2018 Howard White, Timothy Lubanga, Roland Bless Taremwa, Benjamin Kachero, Caroline Otike, Robert Apunyo, Alison Kinengyere, Francis Rathinam, Zeba Siddiqui, Ekwaro Obuku, Ashrita Saran 12 August 2019 An evidence and gap map of development evaluations conducted in Kenya 2000‐2018 Ashrita Saran, Rodney Malesi, Timothy Okech, Virginia Thuku, Evans Mokeni, Margaret Githinji, Bosco Okumu, David Ameyaw, Francis Rathinam, Zeba Siddiqui, Howard White 9 September 2019 Impact assessment of electronic waste recycling practices Salsabil Shaikh, Keith Thomas, Segu Zuhair 30 August 2019 Mitigating climate change in food consumption and food waste: A systematic map of behavioural interventions Lucia A. Reisch, Mark A. Andor, Friederike C. Doebbe, Neal R. Haddaway, Johanna Meier 12 September 2019 Effects of performance‐based financing interventions on the health system and across health sectors in low‐ and middle‐income countries: An evidence gap map Miriam N. Nkangu, Lawrence Mbuagbaw, Loveline Lum, Patrick M. Okwen, Orvill Adams, Christine Mathew, Janet Hatcher Roberts, Sanni Yaya 15 March 2019 Early detection and intervention for reducing the impact of hearing loss in infancy to late adolescence: an evidence and gap map Dhanshree R Gunjawate, Rohit Ravi, Carlie Driscoll 16 July 2019 Gender equality in the distribution of capabilities and resources: An evidence gap map Sujatha Srinivasan, Shreya Jha, R Balagopal, Lakshmi Kumar, Santha Sheela Nair 18 March 2019 Effects of prenatal flavour exposure on subsequent food preferences and birth outcomes of offspring Beyza Nur Ustun, Dilara Ozsoy, Judith Covey, Nadja Reissland 16 May 2019 Examination and reporting of intersectionality in systematic reviews: capturing multidimensional experiences in public health and nutrition interventions Manasee Mishra, Andreea Brabete, Jennifer Petkovic, Kate Ghezzi‐Kopel 13 May 2019 5 SOCIAL WELFARE Effectiveness of intervention for health promotion of refugees and migrant population: an evidence and gap map Baby S Nayak, Sheela Shetty, Tenzin Phagdol, Ashrita Saran, Preethy D'souza 15 September 2019 Technology‐based, digital and digitally‐delivered interventions addressing all forms of intimate partner and domestic violence (IPDV) Chuka Emezue, Tina L. Bloom 23 May 2019 Adult/child ratio and group size in early childhood education and care to promote the development of children aged 0‐5 years Nina Thorup Dalgaard, Anja Bondebjerg, Bjørn Christian Arleth Viinholt, Rasmus Henriksen Klokker, Jens Dietrichson 14 August 2019 Home care package use for preventing patient decline, death, and entrance into hospital and residential care facilities Zorah Hilvert‐Bruce, Dimity Crisp, Zita Hilvert‐Bruce 29 May 2019 Parent/child attachment interventions to improve parent/child attachment and psychosocial adjustment in foster and adoptive parents and children Nina Thorup Dalgaard, Maiken Pontoppidan, Morten Kjær Thomsen, Bjørn Christian Arleth Viinholt, Trine Filges 27 June 2019 Relationship of emotional regulation on video games dependence Jorge R. Saraiva, Cátia M. de Carvalho, Pedro Hubert, Luís Azevedo, Jorge Negreiros, Marta Pinto 28 June 2019
PMC008xxxxxx/PMC8356522.txt
==== Front Campbell Syst Rev Campbell Syst Rev 10.1002/(ISSN)1891-1803 CL2 Campbell Systematic Reviews 1891-1803 John Wiley and Sons Inc. Hoboken 37133288 10.1002/cl2.1043 CL21043 Protocol PROTOCOL Crime and Justice PROTOCOL: Body‐worn cameras’ effects on police officers and citizen behavior: A systematic review LUM et al. Lum Cynthia 1 [email protected] Koper Christopher S. 1 Wilson David B. 1 Stoltz Megan 1 Goodier Michael 1 Eggins Elizabeth 2 Higginson Angela 3 Mazerolle Lorraine 2 1 Department of Criminology, Law and Society George Mason University Fairfax Virginia 2 School of Social Science University of Queensland Brisbane Queensland Australia 3 School of Justice Queensland University of Technology Brisbane Queensland Australia * Correspondence Cynthia Lum, Department of Criminology, Law and Society, Center for Evidence‐Based Crime Policy, George Mason University, 4400 University Drive, MS 6D12, Fairfax, Virginia, 22030. Email: [email protected] 19 9 2019 9 2019 15 3 10.1002/cl2.v15.3 e1043© 2019 The Authors. Campbell Systematic Reviews published by John Wiley & Sons Ltd on behalf of The Campbell Collaboration https://creativecommons.org/licenses/by/4.0/ This is an open access article under the terms of the http://creativecommons.org/licenses/by/4.0/ License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. source-schema-version-number2.0 cover-dateSeptember 2019 details-of-publishers-convertorConverter:WILEY_ML3GV2_TO_JATSPMC version:6.3.2 mode:remove_FC converted:14.07.2023 Lum C , Koper CS , Wilson DB , et al. PROTOCOL: Body‐worn cameras’ effects on police officers and citizen behavior: A systematic review. Campbell Systematic Reviews. 2019;15 :e1043. 10.1002/cl2.1043 37133288 Linked Article Systematic review ==== Body pmc1 BACKGROUND 1.1 The problem, condition or issue Body‐worn cameras (BWCs) are one of the most rapidly diffusing technologies in policing today, costing agencies and their municipalities millions of dollars. This adoption has been propelled by highly publicized events involving police use of force or misconduct, often linked to concerns of racial and ethnic discrimination (see general discussions by Braga, Sousa, Coldren, & Rodriguez, 2018; Lum, Stoltz, Koper, & Scherer, 2019; Maskaly, Donner, Jennings, Ariel, & Sutherland, 2017; Nowacki & Willits, 2018; White, 2014). In culmination, these contexts fostered enough public and political will to generate an urgent call for BWCs. This demand was matched with a prepared supplier; technology companies had already been developing both BWCs and other similar surveillance devices (e.g., in‐car cameras, license plate readers, and closed‐circuit televisions). In the United States, an estimated 60% of local police departments have fully deployed BWCs (Hyland, 2018). Similar widespread testing, piloting, and adoption of BWCs have also occurred in the United Kingdom, Australia, and Europe. Given the rapid and widespread adoption of BWCs, their significant costs, and their potential impacts on law enforcement agencies and the communities they serve, an important question for practitioners, government officials, and researchers is whether the cameras effectively achieve the expectations of them. In their narrative review of empirical BWC research, Lum et al. (2019) suggest there may be equivocal answers to the question of BWC effects. This systematic review of BWCs reviews and synthesizes existing research to examine these concerns. 1.2 The intervention and how it might work BWCs are small surveillance and information technologies that law enforcement officers wear on their clothing or glasses. These cameras can be turned on manually or automatically based on a variety of procedures, policies, rules, or prompts that are usually determined by an agency, government, or other municipal oversight groups. When operating, BWCs record interactions, activities, and events from an officer's vantage point. Some cameras can also record a small time period before and after the cameras are activated to capture a wider time frame around events that officers choose to record. Given their recording capabilities, BWCs are believed to serve a number of functions. Because of the context in which they were adopted, BWCs are intended to document interactions between police and citizens to increase transparency and accountability in the investigation of police misconduct. It is also believed that BWCs can reduce the use of force by officers, deter civilian assaults on officers, and have a mitigating effect on complaints against officers. The primary mechanisms behind how BWCs might work to achieve some of these outcomes is deterrence and self‐awareness. BWCs are theorized to have a deterrent effect on excessive use of force or unlawful actions by officers because officers become self‐aware that they are being recorded by their camera or other cameras worn by fellow officers. As noted by Ariel et al. (2017) in their discussion of the application of deterrence and self‐awareness theories to BWCs: “It is hypothesized that the self‐awareness that arises when we are aware of being watched/filmed drives us to comply with rules/norms, primarily because of the perceived certainty of punishment.” (p. 297) BWCs may also similarly serve to deter individuals that officers encounter. For example, civilians may see the BWCs (or be alerted to them verbally by officers) and then moderate their behavior accordingly. Existing research both supports and challenges the deterrence and self‐awareness hypothesis of BWCs. As Lum et al.’s (2019) narrative review has found, early research seemed to show that BWCs reduced the use of force by officers. More recent findings, however, have been mixed. The hypothesized self‐awareness imposed on officers by the cameras may also affect their use of arrests, citations, and proactive activities, which may have further implications (both positive and negative) for crime prevention and police‐citizen relations. However, research in this area is also ambiguous. Some studies indicate arrests or proactive activities increase, while other studies indicate these activities do not change or even decrease. Further, Lum et al. (2019) also point out that while BWCs seem to reduce complaints against officers, it is unclear whether this reduction is due to the changes in officer behavior (which would indicate a deterrent effect on officers); changes in citizen behavior (which may also indicate a deterrent effect on civilians); or changes in reporting practices by officers and citizens (which does not necessarily point to a deterrent effect). A further complication: McClure et al. (2017) found that many citizens who interact with police cannot even remember whether officers were wearing BWCs (see also White, Todak, & Gaub, 2017). An additional challenge to understanding BWC effectiveness is that survey research has shown there is an incongruence in the expectations that police and community members have for BWCs (Lum et al., 2019). The meaning of “effectiveness” for the police may not be the same as what “effectiveness” means for civilians. The police may view cameras as effective when they protect officers from frivolous complaints and assaults and when they strengthen officers’ ability to arrest and prosecute offenders. Civilians, in contrast, may judge BWC effectiveness by whether cameras provide greater accountability and transparency for officer actions and protect the public against excessive use of force and officer misconduct. And, while some of these effects can be explained from a deterrence perspective, other effects are purely technical or organizational. For example, BWCs are also believed to improve investigations and case clearances. Here, the theoretical mechanism is straightforward: if a crime or an important piece of evidence is captured on an officer's BWC, it can be used to more effectively prosecute offenders. Organizational effects may include the effects of BWCs on training, which delves into the realms of educational theory (i.e., visual aids for learning may create better retention of experiential knowledge for application in the future). 1.3 Why it is important to do this review Because the rapid adoption of BWCs was driven by public protest, law enforcement concerns, government funding, and the development and marketing of portable video technology, it should not be any surprise that BWCs were quickly adopted in a low‐research environment (Lum, Koper, Merola, Scherer, & Reioux, 2015). The importance of scientific inquiry about police technologies like BWCs, however, cannot be overstated. If law enforcement—and ultimately, citizens—intend to invest heavily in BWCs, then BWCs should produce the outcomes we expect of them. Unfortunately, however, researchers have consistently found that police technologies may not lead to the outcomes sought and often have unintended consequences for police officers, their organizations, and citizens (Chan, Brereton, Legosz, & Doran, 2001; Colton, 1980; Koper, Lum, Willis, Woods, & Hibdon, 2015; Lum, Hibdon, Cave, Koper, & Merola, 2011; Lum, Koper, & Willis, 2017; Manning, 2008; Orlikowski & Gash, 1994). Without the results of rigorous research and evaluation, law enforcement leaders are left to rely on best guesses, hunches, notions about “craft,” and “group think” about the impact of technologies like BWCs (see discussion by Lum & Koper, 2017). Research knowledge about technologies, if mind, can help law enforcement agencies better anticipate unintended consequences from technologies, optimize their use of already acquired technologies, or decide whether to invest in specific technologies. The first review of BWCs was conducted by White (2014), who discovered that only five evaluation studies had been completed as of September 2013, even though almost a third of U.S. agencies had already adopted BWCs and widespread adoption was being planned in the UK and Australia. In other words, agencies had already begun rapidly adopting BWCs without clear knowledge about whether the technology could deliver on the high expectations that many had for it (i.e., to increase police accountability, reduce the use of force, reduce disparity, and improve community relationships). Fortunately, researchers have taken a major interest in studying BWCs in the last 5 years and have tried to keep up with its rapid adoption. For example, by November 2015, Lum et al. (2015) found that completed studies about BWCs had grown to more than a dozen, with 30+ additional studies underway. Most of the studies included in both White's and Lum et al.’s reviews were focused on BWCs’ impacts on officer behavior, as measured by complaints and use of force, and on officer perceptions about BWCs. Maskaly et al. (2017), in a review of police and citizen outcomes more specifically, found 21 empirical studies as of January 2017, which led them to conclude that police are generally receptive to BWCs, and that the cameras can exert positive effects on police behavior. In their most current and comprehensive narrative review of BWCs that included all empirical studies found or accepted for publication through June 2018, Lum et al. (2019) discovered approximately 70 published or publicly available studies of BWCs that contained over 110 sub‐studies examining various outcomes and aspects of BWCs. They grouped these studies into six topical categories: (a) the impact of BWCs on officer behavior; (b) officer attitudes about BWCs; (c) the impact of BWCs on citizen behavior; (d) citizen and community attitudes about BWCs; (e) the impact of BWCs on criminal investigations; and (f) the impact of BWCs on law enforcement organizations. Many of these studies were outcome evaluations, an unusual development in technology research where outcome evaluations are often lacking. In their narrative review, Lum et al. (2019) concluded that although it appears that many agencies and officers support BWCs, BWCs have not consistently had the effects intended by either police officers or community members. They argue that anticipated effects may have been “overestimated” and that behavioral changes in the field may be “modest and mixed.” They also discuss that while study findings have indicated that complaints have declined in many evaluations of BWCs, it is unclear why the decline occurs and whether the actual interactions or relationships between the police and the public have improved. There are some outcomes that have not been investigated—in particular, the impact of BWCs on racial and ethnic disparities in policing outcomes, the alleviation of which was a major reason for some communities to push for BWC adoption. At the same time, Lum et al. state that BWCs will continue to be adopted by police agencies, which makes the production and synthesis of rigorous research even more essential to this policy area. Lum et al. (2019) did not conduct a systematic review and meta‐analysis of BWC studies, and therefore many questions remain about what we can conclude about the impacts of BWCs from the existing research. Perhaps findings might be conditioned by the quality of research studies and designs, the location and timing of evaluations conducted, or even by the groups involved in the research (much of the BWC research has been clustered amongst groups of researchers at specific universities). A major concern with BWC outcome evaluations has been the extent of contamination between the treatment and control groups, as well as how outcomes are measured. Lum et al. (2019) acknowledge the importance of a systematic review to parse out these important aspects of studies. 2 OBJECTIVES The primary objective of this review is to synthesize and explore the evidence on the impacts of BWCs on several outcomes of interest to police, policymakers, and the wider community. Specifically, given the existing research found by Lum et al. (2019), this review will focus on examining two categories of effects of BWCs: The impact of BWCs on officer behaviors, as measured by officer use of force, complaints, arrest and citation behavior, and proactive activities. We note that changes in citizen complaints might also be a measure of civilian behavior, as discussed above. However, for this review, it will be used as a measure of officer behavior. The impact of BWCs on civilian behaviors, as measured by community members’ compliance with police commands (to include resisting arrest or assaults against officers). Studies examining this type of impact may also include evaluations of whether BWCs deter criminal or disorderly conduct of community members. Additionally, some studies have examined citizen willingness to call the police (either as a victim or witness) or cooperate in criminal investigations. The second objective of this review is to explore explanations for variations in effect sizes and directions of effects that are likely to be found across studies. Explanations could be due to variations in the location, context, quality, or characteristics of research studies. Toward this end, a number of post‐hoc moderator analyses will be run. For example, there may be differences in findings between experimental and quasi‐experimental studies, although what the difference might be is unclear. While evaluations in criminal justice have indicated that experimental studies may find smaller effects, this is not the case in other related areas. Another possible moderator might be the study's country. Findings may differ, for example, between BWCs studies conducted on U.S. jurisdictions compared to other jurisdictions because of the context of BWC adoption in the U.S. Another possible post‐hoc moderator analysis could be done on the author group involved in the study (see example in Petrosino et al., 2014). A sizeable portion of BWC research has been conducted by research groups at Arizona State University and also Cambridge University. Yet another possible post‐hoc moderator analysis that might be examined is the year of BWC adoption by agencies. Although BWCs impact the specific agency examined, the use and public discourse around BWCs have been widespread. Evaluations of early adopters may yield different results than evaluations of later adopters. Again, these moderator analyses and others will be post hoc, given that there is little theory or empirical research to predict outcomes. Overall, the goal of the review will be to provide practical information to police agencies, municipalities, governments, and citizens as to whether to adopt BWCs or to more carefully consider what BWCs might do for them. As with Lum et al.’s narrative review, this review should also serve as a starting point for debate and conversation about incongruent expectations of BWCs among officers and civilians. Because technologies often lead to unintended consequences for both agencies and the communities they serve, research syntheses can also help to highlight the possibility of such consequences and help agencies and communities plan for potential future impacts of BWCs. 3 METHODOLOGY Given the two objectives discussed above as well as variations in outcomes measured, this systematic review will be organized into two sections (impacts on officer behavior and impacts on civilian behavior). The review may be further broken down into different subareas depending on the outcomes measured (see Gill, Weisburd, Telep, Vitter, & Bennett, 2014, who took a similar approach in their Campbell systematic review of community‐oriented policing). 3.1 Criteria for including and excluding studies 3.1.1 Types of study designs Both experimental and quasi‐experimental designs will be included in this review. Experimental designs will be eligible if treatment is randomly assigned to the units of analysis. BWC studies have used various units of analysis, from individual officers to officer‐shift combinations, to investigative cases. Thus, eligible experimental designs will include those in which BWCs are randomly assigned to officers, officer‐shifts, or other units. Quasi‐experimental studies will also be eligible for this review if a similar comparison group is evident in the study. Study authors may develop a comparable comparison group using propensity score or other matching techniques achieved through the use of statistical controls. Matching may be at the individual level of at the group level. Statistical control methods can include regression, analysis‐of‐covariance, and propensity score matching, among others. Use of a statistical control method is sufficient for inclusion; we will not exclude studies based on a subjective assessment of the quality of the statistical controls. Rather, any quasi‐experimental design that controls for possible explanations for BWC outcomes, such as officer characteristics (race, gender, age, time in service, rank, etc.) or civilian or event characteristics (race, gender, age, situation, the reason for the stop, etc.) will be eligible. Quasi‐experimental designs that do not have a comparison group or do not use the above methods to achieve comparability are not eligible for inclusion in this review. 3.1.2 Types of participants Given the objectives outlined above, the population of interest is law enforcement officers and civilians. However, it should be noted that the units of analysis could vary, to include officers, groups of officers, officer‐shift combinations, and non‐law enforcement personnel (community members, citizens, etc.). In the original title for this protocol, we also listed the “police organization” as being impacted by BWCs. However, given that scarcity of experimental or quasi‐experimental research on BWC impacts on organizations, we will not be pursuing this area of BWC research. Discussions of the scarcity of research in this area have been explored by Lum et al. (2019). 3.1.3 Types of interventions Studies that examine the use of BWCs by law enforcement officers will be eligible for this review. Excluded are studies that focus solely on the use of BWCs for interrogations in an interrogation room within the police agency. 3.1.4 Types of outcome measures Given the possible impacts of BWCs on a wide variety of outcomes as discussed by Lum et al. (2019), multiple outcomes measures are considered in this review. Thus, subanalyses will be conducted and organized similarly to Gill et al.’s (2014) systematic review of community‐oriented policing. Specifically, the following outcomes will be examined within each of the three general categories of studies outlined above: Officer behavior officer use of force complaints against officers arrest and citation behavior proactive activities Civilian behavior compliance with police commands as measured by resisting arrest assaults against officers (which may overlap with 2a, above) criminal or disorderly conduct willingness to call the police or cooperate in an investigation We note that in the Title submitted for this review, we suggested the possibility of examining officer and citizen attitudes towards BWCs for this review. As Lum et al. (2019) discovered, survey research reflects a significant portion of the empirical research on BWCs. However, we do not include officer or citizen attitudes toward BWCs in this systematic review for the following reasons: First, most of this research is neither experimental nor quasi‐experimental, but descriptive. A systematic review of such research would require a separate protocol and review. Second, that literature is not focused on the impact of BWCs on behavior, but rather perceptions of BWCs generally, or perceived beliefs about behaviors. The focus of this systematic review is on measured outcomes, rather than perceived ones. Finally, Lum et al. (2019) have already conducted a preliminary review of this survey research and are currently undertaking a more in‐depth exploratory review of this research outside of the Campbell framework. We also note that in the Title submitted for this review, we suggested the possibility of examining case or investigative outcomes. For example, Lum et al. (2019) found two experimental studies that examined the impact of BWCs for domestic violence case outcomes (Morrow, Katz, & Choate, 2016; Owens, Mann, and Mckenna, 2014). However, we decided to exclude this category of studies from this review. There is a large body of research that examines the impact of videotaping more generally on interrogations and interviewing of suspects, witnesses, and victims, and the use of videos and court outcomes. Given that these outcomes do not specifically focus on the impact of cameras on officer and citizen behavior and given the overlap of this area with other unrelated areas (investigatory effectiveness using video technologies), we excluded this area of research from this review. 3.1.5 Duration of follow‐up The expected effects of BWCs are immediate. That is, they are presumed to have an effect while they are being used. As such, the outcomes are measured concurrently with the intervention and no follow‐up period is needed in assessing their effects. Some studies do measure the longer‐term effects of BWCs, but in these studies, the BWCs are still in use. While we do not expect to find studies that measure effects at a follow‐up period after the BWCs are no longer in use, if such a study is located during the search and screening process, it will be included in the review. 3.2 Search strategy and screening process The search for BWC research will be led by the Global Policing Database research team at the University of Queensland (Elizabeth Eggins and Lorraine Mazerolle) and Queensland University of Technology (Angela Higginson). The University of Queensland is home to the Global Police Database (see http://www.gpd.uq.edu.au), which will serve as the main search location for this review. As detailed by Higginson, Eggins, Mazerolle, and Stanko (2015), the GPD “is a web‐based and searchable database designed to capture all published and unpublished experimental and quasi‐experimental evaluations of policing interventions conducted since 1950. There are no restrictions on the type of policing technique, type of outcome measure or the language of the research” (p. 1). The GPD is compiled using systematic search and screening techniques, which are reported in the Higginson et al. (2015) and summarised in Appendices A and B. Broadly, the GPD search protocol includes an extensive range of search locations to ensure that both published and unpublished research is captured across criminology and allied disciplines. To capture studies for this review, we will use BWC specific terms to search the GPD corpus of full‐text documents that have been screened as reporting a quantitative impact evaluation of a policing intervention. Specifically, we use the following terms to search the title and abstract fields of the corpus of documents published between January 20041 and December 2018: camera⁎video⁎ORBWC⁎ORBWV⁎ The results of this search will then be processed using a two‐coder system. The abstract for each study found from the search above will be examined by two coders, who will separately determine whether a study is “potentially eligible” for further full‐text review given this protocol's criteria; “not eligible” for further full‐text review; “unclear” (the coder could not make a determination given the information given); or a “relevant review” (the article is not a study, but should be flagged as a relevant review of studies. The codes from both coders will be reviewed for differences by a principal investigator (Lum or Koper). Each difference will then be discussed and if needed, mitigated by a third coder. Studies with differences that persist and cannot be mitigated (specifically if one coder continues to believe a study is “potentially eligible”) will be retained and the full text of the study will be examined in the next screening process. After reviewing the initial abstracts from the GPD, the research team will also examine whether any relevant studies from Lum et al. (2019) and from the Bureau of Justice Assistance's BWC Toolkit Resources2 were missed in the GPD search. Both sources of information contain comprehensive research collections on the impacts of BWCs on officer and citizen behavior and may contain studies relevant for this review. Once studies are determined by at least one coder to be “potentially eligible”, the full‐text document of each study will be obtained and examined separately by two coders for eligibility according to the “Criteria for Including and Excluding Studies” as described above. Studies must satisfy these criteria in order to be included in the systematic review. If the coders differ in their assessment, a third coder will be used to examine the study for eligibility. If a study continues to draw debate, other coders and expert may be consulted to determine its eligibility for the systematic review. 3.3 Criteria for determination of independent findings The primary unit‐of‐analysis for this review will be a research study defined as a distinct sample of study participants involved in a common research project. Multiple reports (e.g., publications, technical reports, etc) from a common research study will be coded as a single study. Stated differently, a research study will only be treated as unique if the study sample does not include study participants included in any other coded study. Multiple effect sizes will be coded, if possible, from studies when multiple outcomes are analyzed. Statistical independence will be maintained or modeled in all statistical analyses. The choice of outcomes in this review will be prioritized. For example, while there are many different types of use of force (e.g., hands only, nonlethal instruments, firearm use) and complaints (i.e., complaints of rudeness, service delivery), we will select the most general measure of use of force or complaints measured (i.e., counts of reports of use of force or complaints generated). Additionally, there are many different types of crimes and infractions that may receive arrest and citations, but only the most general measure of arrest and citation will be measured (i.e., “all arrests” or “all citations”). Similarly, for non‐police civilian behaviors, the more general behavioral categories will be measured (i.e., “resisting arrest,” “assault on officers,” etc). With regard to officer proactivity, a decision may need to be made as to whether to examine the overall levels of proactivity or specific types of proactivity (i.e., stop‐question‐and‐frisks, traffic stops, pedestrian stops, problem‐solving, community policing, etc). As Lum et al. (2019) discuss, not all proactive activities are viewed similarly by either the police or community members, and may need to be parsed out during the analysis to examine BWCs impacts on different types of proactive police behaviors. Finally, for the impacts of BWCs on investigative case files, general categories will be used, including “arrest” or “conviction.” 3.4 Details of study coding categories Per Campbell policy, all studies will be double‐coded. Detailed coding categories and instructions are presented in the Appendix C. Coding will include information on the nature of the BWC use, comparator condition, contextual features of the agency, method and design features, dependent measures and effect sizes for the above outcomes, and risk‐of‐bias indicators. Data will be maintained in a relational database (MySQL) with coding forms developed in LibreOffice Base (similar to MS Access). 3.5 Statistical procedures and conventions Based on prior work by Lum et al. (2019), we expect to find a sufficient number of studies to conduct a meta‐analysis for the three broad outcomes described above. However, given the various outcomes and study designs that are likely to be found, a variety of approaches to calculating effect sizes will have to be used, as described by Lipsey and Wilson (2001). Various effect sizes will be then converted to Cohen's d except for outcomes that are more naturally measured dichotomously, in which case the odds ratio will be used. Calculation techniques as described by Lipsey and Wilson and the online effect size calculator developed by David Wilson will be employed. A meta‐analysis will be conducted using random‐effects models estimated via full‐information maximum likelihood. Primary analyses will be performed using Stata packages developed by David B. Wilson and available at http://mason.gmu.edu/~dwilsonb/ma.html. The robust standard error method of modeling statistical dependences will be implemented with the Stata package robumeta (see http://www.northwestern.edu/ipr/qcenter/RVE‐meta‐analysis.html for details). Moderator analyses of a single categorical variable will be fit using the analog‐to‐the‐ANOVA method, also under a random‐effects model. Moderator analyses of continuous moderators or multiple moderators will be conducted with meta‐analytic regression methods, also under a random‐effects model. Results will be presented separately for experimental (randomized) and quasi‐experimental designs, although these may be combined in moderator analyses. Publication‐selection bias will be assessed in three ways. First, analyses will compare the results from published and unpublished reports. Published documents will include peer‐reviewed journal articles, books, and book chapters. All other report forms, such as theses, technical reports, government and agency reports, will be considered unpublished. Second, we will perform a trim‐and‐fill analysis on the major outcome categories. Third, we will visually inspect a funnel plot on the major outcome categories. Sensitivity analysis will be conducted if needed, based on initial findings. As already discussed, moderator analysis will be conducted for this review, and can include (but is not limited to) the following: Whether the study was done inside or outside of the United States (where most BWC research has been conducted) Whether research was conducted by specific dominant research teams that conduct BWC research Type of research design (i.e., experimental versus quasi‐experimental) Publication type (published versus unpublished) Evaluation on earlier or later adopters We do not plan to include qualitative research in this systematic review, except as to provide context for interpreting results. ROLES AND RESPONSIBILITIES Content: Lum, Koper, Wilson, Stoltz, and Goodier Systematic review methods: Wilson, Lum, and Koper Statistical analysis: Wilson, Lum, and Koper Information retrieval: Eggins, Higginson, Mazerolle, Stoltz, and Goodier SOURCES OF SUPPORT This systematic review is supported by funding from the Laura and John Arnold Foundation DECLARATION OF INTERESTS None of the authors have been involved in an evaluation of BWCs in a police agency. We also have not received any payment or have had any contractual agreements with any technology company who manufactures or sells BWCs, or who profits from sales of BWCs to law enforcement agencies. PRELIMINARY TIMEFRAME 1/31/2019 Submission of Protocol 3/31/2019 Completion of the search process using the GPD 12/31/2019 Approximate date of submission of the systematic review John Wiley & Sons, Ltd. PLANS FOR UPDATING THE REVIEW This review will be updated every 4 years, under the primary responsibility of Cynthia Lum, unless all authors agree that another author takes primary responsibility. The updated review is contingent on availability and funding. AUTHOR DECLARATION Authors’ responsibilities By completing this form, you accept responsibility for preparing, maintaining, and updating the review in accordance with the Campbell Collaboration policy. Campbell will provide as much support as possible to assist with the preparation of the review. A draft review must be submitted to the relevant Coordinating Group within 2 years of protocol publication. If drafts are not submitted before the agreed deadlines, or if we are unable to contact you for an extended period, the relevant Coordinating Group has the right to deregister the title or transfer the title to alternative authors. The Coordinating Group also has the right to deregister or transfer the title if it does not meet the standards of the Coordinating Group and/or Campbell. You accept responsibility for maintaining the review in light of new evidence, comments and criticisms, and other developments, and updating the review at least once every 5 years, or, if requested, transferring responsibility for maintaining the review to others as agreed with the Coordinating Group. Publication in the Campbell Library The support of the Coordinating Group in preparing your review is conditional upon your agreement to publish the protocol, finished review, and subsequent updates in the Campbell Library. Campbell places no restrictions on publication of the findings of a Campbell systematic review in a more abbreviated form as a journal article either before or after the publication of the monograph version in Campbell Systematic Reviews. Some journals, however, have restrictions that preclude publication of findings that have been, or will be, reported elsewhere and authors considering publication in such a journal should be aware of possible conflict with the publication of the monograph version in Campbell Systematic Reviews. Publication in a journal after publication or in press status in Campbell Systematic Reviews should acknowledge the Campbell version and include a citation to it. Note that systematic reviews published in Campbell Systematic Reviews and coregistered with Cochrane may have additional requirements or restrictions for copublication. Review authors accept responsibility for meeting any co‐publication requirements. Appendix A GPD systematic search strategy Search terms To ensure optimum sensitivity and specificity, the GPD search strategy utilizes a combination of free‐text and controlled vocabulary search terms. Because controlled vocabularies and search capabilities vary across databases, the exact combination of search terms and field codes are adapted to each database. Final search syntax for each location will be reported in the final review. The free‐text search terms for the GPD are provided in Table A1 and are grouped by substantive (i.e., some form of policing) and evaluation terminology. Although the search strategy across search locations may vary slightly, the search follows a number of general rules: Search terms will be combined into search strings using Boolean operators “AND” and “OR”. Specifically, terms within each category will be combined with “OR” and categories will be combined with “AND”. For example: (police OR policing OR “law#enforcement”) AND (analy* OR ANCOVA OR ANOVA OR …). Compound terms (e.g., law enforcement) will be considered single terms in search strings by using quotation marks (i.e., “law*enforcement”) to ensure that the database searches for the entire term rather than separate words. Wild cards and truncation codes will be used for search terms with multiple iterations from a stem word (e.g., evaluation, evaluate) or spelling variations (e.g., evaluat* or randomi#e). If a database has a controlled vocabulary term that is equivalent to “POLICE”, we will combine the term in a search string that includes both the policing and evaluation free‐text search terms. This approach will ensure that we retrieve documents that do not use policing terms in the title/abstract but have been indexed as being related to policing in the database. An example of this approach is the following search string: (((SU: “POLICE”) OR (TI,AB,KW: police OR policing OR “law*enforcement”)) AND (TI,AB,KW: intervention* OR evaluat* OR compar* OR …)). For search locations with limited search functionality, we will implement a broad search that uses only the policing free‐text terms. Multidisciplinary database searches will be limited to relevant disciplines (e.g., include social sciences but exclude physical sciences). Search results will be refined to exclude specific types of documents that are not suitable for systematic reviews (e.g., newspapers, front/back matter, book reviews). Table A1 Free‐text search terms for the GPD systematic search Policing search terms Evaluation search terms police policing “law*enforcement” constab* detective* sheriff* analy* ANCOVA ANOVA “ABAB design” “AB design” baseline causa* “chi#square” coefficient* “comparison condition*” “comparison group*” “control condition*” “control group*” correlat* covariat* “cross#section*” data effect* efficacy eval* experiment* hypothes* impact* intervent* interview* longitudinal MANCOVA MANOVA “matched group” measure* “meta‐analy*” “odds#ratio* outcome* paramet* “post‐test” posttest “post test” predict* “pre‐test” pretest program* “propensity score*” quantitative “quasi#experiment*” questionnaire* random* RCT regress* result* “risk#ratio*” sampl* “standard deviation*” statistic* studies study survey* “systematic review*” “t#test*” “time#series” treatment* variable* variance John Wiley & Sons, Ltd. Search locations To reduce publication and discipline bias, the GPD search strategy adopts an international scope and involves searching for literature across a number of disciplines (e.g., criminology, law, political science, public health, sociology, social science, and social work). The search captures a comprehensive range of published (i.e., journal articles, book chapters, books) and unpublished literature (e.g., working papers, governmental reports, technical reports, conference proceedings, dissertations) by implementing a search strategy across bibliographic/academic, grey literature, and dissertation databases or repositories. We note that there is a substantial overlap of the content coverage between many of the databases. Therefore, we have used the Optimal Searching of Indexing Databases (OSID) computer program (Neville & Higginson, 2014) to analyze the content crossover for all databases that have accessible content coverage lists. OSID analyses the content coverage and creates a search location solution that provides the most comprehensive coverage via the least number of databases. For example, if the content for the set of databases seen in Figure A1 were imported, OSID would provide a solution that entails searching only databases 3 and 4 because the content covered by databases 1 and 2 is covered by database 4. Another advantage of using OSID when designing a search strategy is the reduction in the number of duplicates that would need to be removed prior to the screening phase. Databases with >10 unique titles were being searched in full, whereas for databases with ≤10 unique titles, we searched only the unique titles and any nonserial content (e.g., reports and conference proceedings). Where a modified search of a database would be more labor intensive than a full search and export results, conducted a full search of the database. The final search locations and solution are reported in Table A2. Figure A1  The GPD Flowchart [Color figure can be viewed at wileyonlinelibrary.com] Table A2 GPD search locations and protocol (January 1st 1950 to December 2018) Indexed and academic databases Content coverage fed into osid? Full or modified search? Search modifications ProQuest Criminal Justice Yes Full None. Dissertation and Theses Database Global Not Available Modified Social Sciences subset. Political Science Yes Full None. Periodical Archive Online Yes Full None. Research Library Yes Modified Social Sciences subset. Social Science Journals Yes Full None. Sociology Yes Modified Search 2 unique journal titles and nonserial content only. Applied Social Sciences Index and Abstracts Yes Full None. International Bibliography of the Social Sciences Yes Full None. Public Affairs Information Service Yes Full None. Social Services Abstracts Yes Modified Search 5 unique journal titles and nonserial content only. Sociological Abstracts Yes Full None. Worldwide Political Sciences Abstracts Yes Modified Search 9 unique journal titles and nonserial content only. EBSCO Academic Search Premier Yes Full None. Criminal Justice Abstracts Yes Full None. EconLit Yes Full None. MEDLINE with Full‐Text Yes Full None. Social Sciences Full‐Text Yes Full None. OVID International Political Science Abstracts Not Available Full None. PsycARTICLES Yes Modified Search 4 unique journal titles only. PsycEXTRA Not Available Full None. PsycINFO Yes Full None. Social Work Abstracts Not Available Full None. Web of Science Current Contents Connect – Social and Behavioural Sciences Edition Yes Modified Search 1 unique journal title and nonserial content only. Book Citation Index (Social Sciences and Humanities) Not Available Full None. Conference Proceedings Citation Index (Social Sciences and Humanities) Not Available Full None. Social Science Citation Index Yes Full None. Informit Australian Attorney‐General Information Service Yes Full None. Australian Criminology Database (CINCH) Yes Full None. Australian Federal Police Database Yes Full None. Australian Public Affairs Full‐Text Yes Full None. DRUG Yes Full None. Health & Society Database Yes Modified Search unique journal titles and nonserial content only. Humanities and Social Sciences Collection Yes Full None. Gale‐Cengage Expanded Academic ASAP Yes Full None. Standalone and open access databases Cambridge Journals Online Yes Modified Search 4 unique journal titles in Law and Political Science collections and a full search of Social Studies collection. Directory of Open Access Journals Yes Full None. HeinOnline Yes Modified Law Journals Online collection only. JSTOR Yes Modified Search unique titles across the Law, Political Science, Public Health, Public Policy, Social Work and Sociology collections only. The Criminal Justice collection had no unique content and so will be excluded from the search. Only 10% of the content in this database have abstracts and a full‐text search returns > 250,000 results because of inability to construct complex search strings. Therefore, a modified search of the unique titles across these collections will be more pragmatic than a full search of the database. Oxford Scholarship Online Yes Full None. Sage Journals Online and Archive (Sage Premier) Yes Modified Search 5 unique journal titles and nonserial content only. ScienceDirect Yes Full None. SCOPUS Yes Full None. SpringerLink Yes Full Although this database has low uniqueness when combined with the full set of databases, a full search using only the policing search terms will be more pragmatic than a modified search on unique titles because of the restricted search functionality of this database. Taylor & Francis Online Yes Modified Although this database has low uniqueness when combined with the full set of databases, a full search using only the policing search terms will be more pragmatic than a modified search on unique titles because of the restricted search functionality of this database. Wiley Online Library Yes Full None. California Commission on Peace Officer Standards & Training Library No Full None. Cochrane Library No Full None. CrimeSolutions.gov No Full None. Database of Abstracts of Reviews of Effectiveness (DARE) No Full None. FBI – The Fault (Reports and Publications) No Full None. Evidence‐Based Policing Matrix No Full None. International Initiative for Impact Evaluation Database (3ie) No Full None. National Criminal Justice Reference Service No Full None. Safety Lit Database No Full None. Australian Institute of Criminology No Full None. Bureau of Police Research and Development (India) No Full None. Canadian Police Research Catalogue No Full None. Centre for Problem‐Oriented Policing No Full None. College of Policing (including POLKA and Crime Reduction Toolkit) No Full None. European Police College (CEPOL) No Full None. Evidence for Policy and Practice Information and Coordinating Centre No Full None. National Research Institute of Police Science (Japanese) No Full None. Office of Community Oriented Policing Services No Full None. Police Executive Research Forum (US) No Full None. Police Foundation (US) No Full None. Tasmania Institute of Law Enforcement Studies (Australia) No Full None. Policing Online Information System (POLIS, Europe) No Full None. Scottish Institute for Policing Research No Full None. Centre of Excellence in Policing and Security (Australian, now archived) No Full None. John Wiley & Sons, Ltd. Appendix B GPD systematic compilation strategy Inclusion criteria Each record captured by the GPD systematic search must satisfy all inclusion criteria to be included in the GPD: timeframe, intervention, and research design. There are no restrictions applied to the types of outcomes, participants, settings or languages considered eligible for inclusion in the GPD. Types of interventions Each document must contain an impact evaluation of a policing intervention. We define a policing intervention is some kind of a strategy, program, technique, approach, activity, campaign, training, directive, or funding /organizational change that involves the police in some way (other agencies or organizations can be involved). Police involvement is broadly defined as Police initiation, development or leadership Police are recipients of the intervention or the intervention is related, focused or targeted to police practices Delivery or implementation of the intervention by police Types of study designs The GPD includes quantitative impact evaluations of policing interventions that utilize randomized experimental (e.g., RCTs) or quasi‐experimental evaluation designs with a valid comparison group that does not receive the intervention. The GPD includes designs where the comparison group receives “business‐as‐usual” policing, no intervention or an alternative intervention (treatment‐treatment designs). The specific list of research designs included in the GPD are as follows: Systematic reviews with or without meta‐analyses Crossover designs Cost‐benefit analyses Regression discontinuity designs Designs using multivariate controls (e.g., multiple regression) Matched control group designs with or without pre‐intervention baseline measures (propensity or statistically matched) Unmatched control group designs with pre‐post intervention measures which allow for difference‐in‐difference analysis Unmatched control group designs without pre‐intervention measures where the control group has face validity Short interrupted time‐series designs with a control group (less than 25 preintervention and 25 postintervention observations (Glass, 1997)) Long interrupted time‐series designs with or without a control group (≥25 pre‐ and post‐intervention observations (Glass, 1997)) Raw unadjusted correlational designs where the variation in the level of the intervention is compared to the variation in the level of the outcome The GPD excludes single group designs with pre‐ and post‐intervention measures as these designs are highly subject to bias and threats to internal validity. Systematic screening To establish eligibility, records captured by the GPD search progress through a series of systematic stages which are summarised in Figure A1, with additional detail provided in the following subsections. All research staff working on the GPD undergo standardized training before beginning work within any of the stages detailed below. Staff then complete short training simulations to enable an assessment of their understanding of the GPD protocols and highlight any areas for additional training. In addition, random samples of each staff's work are regularly cross‐checked to ensure adherence to protocols. Disagreements between staff are mediated by either the project manager or GPD chief investigators. Title and abstract screening After removing duplicates, the title and abstract of record captured by the GPD systematic search is screened by trained research staff to identify potentially eligible research that satisfy the following criteria: Document is dated between 1950 to present Document is unique (i.e., not a duplicate) Document is about police or policing Document is an eligible document type (e.g., not a book review) Records are excluded if the answer to any one of the criteria is unambiguously “No,” and will be classified as potentially eligible otherwise. Records classified as potentially eligible progress to full‐text document retrieval and screening stages. Full‐text eligibility screening Wherever possible, a full‐text electronic version of eligible records will be imported into SysReview. For records without an electronic version, a hardcopy of the record will be located to enable full‐text eligibility screening. The full text of each document will be screened to identify studies that satisfy the following criteria: Document is dated between 1950 to present; Document is unique; Document reports a quantitative statistical comparison; Document reports on policing evaluation; Document reports in quantitative impact evaluation of a policing intervention; and Evaluation uses an eligible research design. Appendix C Coding forms and instructions Version: July 8, 2019 Note: This is a living document that will be updated and modified during the coding process as decisions are made regarding how to handle edge cases or other refinements that are made to the coding protocol. Coding will be done directly into a MySQL relational database using Libreoffice Base as the front‐end with detailed coding forms that reflect the coding protocol. Initial eligibility screening Initial eligibility screening will be performed on all titles and abstracts identified by the bibliographic search. Two coders will independently assess whether the title and abstract suggest that the study may meet the full eligibility criteria (see protocol). Each coder will determine if the reference is “potentially eligible”, “not eligible”, a “relevant review”, or that it is “unclear”. A reference marked as unclear will be assessed by another coder. Any reference marked as “potentially eligible” by either coder will move forward to a full‐text eligibility assessment. Final (full‐text) eligibility screening For full‐text screening, each document (reference) will be assessed against the four criteria of the eligibility criteria (see protocol or database coding form for Final Screening). For each criterion, answer “yes”, “no” or “uncertain”. If the coder answers “yes” to all four criteria, code the reference as “Eligible”. If the coder answers “no” to any item, code the reference as “Not Eligible”. If there are mix of “yes” and “uncertain” then another coder must make the final assessment. Two coders will assess each reference for eligibility. Any discrepancies will be resolved through a consensus process. Study level coding Study ID & Substudy ID: These two field uniquely identify the study and sub‐study and should correspond to a record coded at the study level. Label substudies sequentially by letter (i.e., A, B, C, …). This field cannot be left blank. Thus, if there is only one study, the substudy is “A”. Author/year label. Create a unique author label for each study/substudy combination that suitable for use on a forest plot (e.g., “Ariel 2017 Denver” and “Braga et al 2018 Las Vegas”). Coder's Initials: Insert your initials. This is case‐insensitive. Date Modified: This is auto‐generated and is a timestamp for the last time any changes were made to this record. Publication Types (check all that apply [multiple documents may be coded as part of one study]): Journal Book Book Chapter Technical report (government agency) Technical report (university/research institute) Dissertation/thesis Presentation Other Context of the study 6. Specific location or jurisdiction in which the study was conducted [text box, this could be a city, county, state, country such as “Tucson” or “Colombia” or “Maryland” (if conducted across the entire state)] 7. State [text box, if in the U.S., use full state name, including “District of Columbia”] 8. Country [text box, use full name of the country – “United States”, not “U.S.”] 9. Population density (population per square mile from the 2010 U.S. Census of the specific geographic area that study location is a part of) 10. Diversity Index (Calculated and added from 2010 Census information based on established diversity index formula) 11. Law enforcement agency name if known [text box]. 12. Number of full‐time sworn officers in the agency in the year in which the intervention was initiated [number box, found from annual UCR LEOKA files] 13. Agency type law enforcement only law enforcement agency with correctional duties [i.e., Sheriff's agencies; requirement of the review is that officers with BWCs have law enforcement duties] regional agency that oversees multiple jurisdictions or a national agency 14. Primary university or research organization in which the study was conducted [text box, full name of university (“Florida State University”) or the research organization (“Police Executive Research Forum”)]. 15. Were BWCs used in the agency prior to the study? BWCs were already in use by agency before the study began across the agency (not selective use) BWCs were already in use by agency before the study began but only by specific and very limited units/people or in a pilot testing phase Use of BWCs began very close to the time of the study or for the purposes of the study 16. If known, the year of BWC adoption by the agency [text box year] 17. In the 2 years prior to camera adoption, had this agency or jurisdiction undergone a collaborative reform, consent decree, or USDOJ review? YES/NO Intervention information 18. Nature of BWC use during the intervention Uniformed patrol only Specialized units only (traffic, investigative, community‐oriented, etc.) Combination uniformed patrol and specialized units Training environments (academy, in‐service) Other [text box] 19. Activation requirements for BWC use specific to the intervention condition (agency may have had different activation policies prior to the intervention, but this is specific to the intervention and study) Activated at all times with exceptions as described in official policy (bathroom breaks, hospitals, child victims) Activation policies are specified for categories of incidents and activities as described in policy (turned off at all other times) Discretionary activation (officers can choose whether to turn on/off) No activation policy Unknown/unclear activation policy 20. Month/year study intervention started 21. Month/year study intervention finished 22. Data collection start date (month/year) (this includes the date when baseline data/measures started to be collected) 23. Data collection end date (month/year) 24. Challenges to implementation (fidelity issues) [text box] Research methods and design 25. Unit of Analysis Officer Shift Officer‐shift combinations Enforcement groups (squads, specialized units) Police‐defined geographic areas (beats, districts, sectors) Other geographic areas (hot spots, census blocks, etc.) Other [text box] 26. Type of experimental or quasi‐experimental design RCT (randomized experiment) Cluster RCT (higher‐order units randomized) Quasi‐experiment: simple matching Quasi‐experiment: propensity score matching Quasi‐experiment: statistical adjustments for baseline features (including time trend) Quasi‐experiment: historical/cohort design Quasi‐experiment: other Eligible time series 27. How the sample was selected? All available units of analysis were used Convenience sampling: only those who volunteered were used Specific sample: Agency only wanted specific units to use BWCs and/or be involved in the study Other [text box] 28. Description of sample selection bias [text box] 29. Treatment description [text box] 30. Treatment group sample size at the start of the study [number box] 31. Control group description [text box] 32. Control group sample size at the start of the study [number box] 33. Attrition concerns described [text box] 34. A priori power calculations? Yes (explicitly stated) No (not explicitly stated; no information) Cochrane risk‐of‐bias (Y = Yes; PY = Probably Yes; PN = Probably No; N = No; NI = No information) Domain 1: Risk of bias arising from the randomization process 1.1Was the allocation sequence random? Y/PY/PN/N/NI 1.2Was the allocation sequence concealed until participants were enrolled and assigned to interventions? Y/PY/PN/N/NI 1.3Did baseline differences between intervention groups suggest a problem with the randomization process? [For quasi‐experiments, were there meaningful differences on baseline variables?] Y/PY/PN/N/NI Optional: Were there violations to the randomization? That is, were members of the control group exposed to the treatment and/or members of the treatment group who were not exposed to the treatment? Y/PY/PN/N/NI Optional: What is the predicted direction of bias arising from the randomization process? [For quasi‐experimental designs, what is the predicted direction of bias of the selection process?] Favors Experimental/Favors Comparator/ Towards Null/Away From Null/Unpredictable Domain 2: Risk of bias due to deviations from the intended interventions (effect of assignment to intervention) 2.6Was an appropriate analysis used to estimate the effect of assignment to intervention? Y/PY/PN/N/NI 2.7Was there potential for a substantial impact (on the result) of the failure to analyze participants in the group to which they were randomized? Y/PY/PN/N/NI Domain 5: Risk of bias in selection of the reported result Is [Are] the numerical result [results] being assessed likely to have been selected, on the basis of the results, from… 5.2… multiple outcome measurements (e.g. scales, definitions, time points) within the outcome domain? Y/PY/PN/N/NI 5.3… multiple analyses of the data? Y/PY/PN/N/NI Optional: What is the predicted direction of bias due to selection of the reported result? Favors Experimental/Favors Comparator/ Towards Null/Away From Null/Unpredictable Outcome level coding Each eligible dependent variable or measure will be coded as a separate record. The Primary Key that is the unique identifier for each record in the outcome table (in the database, this table is named "3_Outcome") and is the combination of the following fields: StudyID, SubStudyID, OutcomeID, and CoderID. The StudyID and SubStudyID. Study ID & Substudy ID: These two fields uniquely identify the study and sub‐study and should correspond to a record coded at the study level (see study level coding instructions). Outcome ID: This field uniquely identifies the outcome associated with this effect size. This should correspond to a record coded at the outcome level (see outcome level coding instructions). Outcome label. Type a label that describes this outcome. Ideally, this should match what is used in the text unless additional clarification is needed. Coder's Initials. Date Modified: This is auto‐generated and is a timestamp for the last time any changes were made to this record. Construct measured by this outcome Officer behavior officer use of force complaints against officers arrest and citation behavior proactive activities Civilian behavior compliance with police commands as measured by resisting arrest assaults against officers (which may overlap with 2a, above) criminal or disorderly conduct willingness to call the police or cooperate in an investigation Nature of outcome data Dichotomous Discrete ordinal scaled measure (<10 categories) Discrete ordinal scaled measure (10+ categories) Count Ratios Continuous measures Unclear Unit‐of‐measurement Officer Citizen Incident Shift Police unit Time period Geographic area Unclear Cochrane Risk‐of‐Bias Domain 3: Missing outcome database 3.1Were data for this outcome available for all, or nearly all, participants randomized? Y/PY/PN/N/NI 3.2If N/PN/NI to 3.1: Is there evidence that the result was not biased by missing outcome data? Y/PY/PN/N/NI 3.3If N/PN to 3.2: Could missingness in the outcome depend on its true value? Y/PY/PN/N/NI 3.4If N/PN to 3.2: Could missingness in the outcome depend on its true value? Y/PY/PN/N/NI Optional: What is the predicted direction of bias due to missing outcome data? NA Favors experimental Favors comparator Towards the null Away from the null Unpredictable Cochrance Risk‐of‐Bias Domain 4: Risk of bias in measurement of the outcome 4.2Could measurement or ascertainment of the outcome have differed between intervention groups? Y/PY/PN/N/NI 4.3If N/PN/NI 4.2: Were outcome assessors aware of the intervention received by study participants? Y/PY/PN/N/NI 4.4If Y/PY/NI to 4.3: Could assessment of the outcome have been influenced by knowledge of intervention received? Y/PY/PN/N/NI 4.5If Y/PY/NI to 4.4: Is it likely that assessment of the outcome was influenced by knowledge of intervention received? Y/PY/PN/N/NI Optional: What is the predicted direction of bias in the measurement of the outcome? Favors Experimental/Favors Comparator/ Towards Null/Away From Null/Unpredictable Effect size level coding Each unique codeable effect for an eligible outcome should be coded as a separate record. The Primary Key that is the unique identifier for each record in the effect size table (in the database, this table is named “4_EffectSize”) is the combination of the following fields: StudyID, SubStudyID, OutcomeID, ESID, and Coder. The StudyID and SubStudyID should uniquely identify which study/substudy is being coded. The OutcomeID indicates which outcome coded at the outcome level is associated with this effect. The coding form has input fields for nine different ways to compute an effect size. There are times when a unique effect can be computed in more than one way. Select the method that is most accurate. For example, a study may report both means, standard deviations, and sample size information along with an independent t‐test associated with these means. Both the former and latter can be used to compute the effect size and should produce the same value. However, the t‐test, unless reported to 3 or more digits, is likely to be less precise due to rounding error than the raw means and standard deviations. In contrast to the above, the same outcome may be analyzed in different ways that would produce different effect sizes. For example, a study may report the raw means and t‐test but also report a regression model with a treatment dummy code that adjusts for baseline covariates. In such a situation, code two effect sizes, one based on the means and one on the regression model (i.e., method 7 on the coding form). These should be coded as separate records in the database. 1. Study ID and substudy ID: These two fields uniquely identify the study and sub‐study and should correspond to a record coded at the study level (see study level coding instructions). 2. Outcome ID: This field uniquely identifies the outcome associated with this effect size. This should correspond to a record coded at the outcome level (see outcome level coding instructions). 3. Effect Size ID: Assign each effect size for a given StudyID+SubStudyID combination a unique number, starting at 1. For example, if a Study/SubStudy has four effect sizes, these should be numbered 1, 2, 3, 4 in this field. 4. Coder's Initials: Insert your initials. This is case‐insensitive. 5. Date Modified: This is auto‐generated and is a timestamp for the last time any changes were made to this record. 6. Page number. Indicate the page number where the effect size data can be found. 7. Notes: Use this text box to indicate any notes you have recording this effect size. Use liberally. If you are uncertain about how anything is coded, please specify this in the note. Also, if you computed the effect size by hand, include relevant information so that another coder could replicate your computations in this field (e.g., insert R code if applicable). 8. Description of the timing for the effect size. Use this text box to describe the timing for the effect size. For example, the data for the effect size may reflect a 6‐month period following the start of the use of BWCs. 9. Timing for the effect size. Indicate if this effect size is measured at baseline (or pretest). For all effect sizes after the start of the use of BWC, select the post‐test 1, 2, 3, or 4 sequentially (e.g., the first post‐test is 1, next is 2, etc.). 1. Baseline 2. Post‐test 1 3. Post‐test 2 4. Post‐test 3 5. Post‐test 4 10. Direction of effect size. Indicate if the direct for the effect favors the treatment or control. If you select “neither” then the effect size must equal 0.00. If you select “unclear” this effect size will not be used unless the direct of effect can be resolved. 1. Treatment 2. Control 3. Neither 4. Unclear 11. Effect size adjusted for baseline variables (e.g., effect size based on the regression model that includes baseline variables) 1. Yes 2. No 3. Unclear 12. Is the effect reported as significant at P <= .05? 1. Yes 2. No 3. Not tested 4. Unclear 13. Unit‐of‐analysis for this effect (i.e., what does the sample size reflect) 1. Officer 2. Incident 3. Shift 4. Other (specify: __________) 14. Clustered or nested data for this effect? 1. Yes 2. No 3. Unclear 15. If this is a clustered effect, does the study report the ICC (intra‐class correlation, the variance of the random intercept in a mixed‐effects model)? 16. Method used to compute the effect size 1. Means and standard deviations 2. Means and standard errors 3. the t‐test (between two means) 4. Frequencies (success/failure)—could be represented as a 2 by 2 (treatment/control by success/failure or some other binary outcome) 5. Proportions (success/failure)—proportion of successes or failures in each group (could be converted to a 2 by 2 frequency table using the sample sizes) 6. Count or rate (check with Dave Wilson before using this one) 7. OLS/HLM regression. Results from a regression model with a continuous type‐dependent variable (OLS type regression, etc.). Treatment/control must be reflected by a dummy variable (e.g., 0/1). 8. Logistic regression. Results from a logistic regression model. Treatment/control must be reflected by a dummy variable (e.g., 0/1). 9. Hand calculated. This is for more complex situations where you, for example, use the online effect size calculator or computed the effect size using R. 17. Treatment group sample size (needed for all methods of computing effect sizes) 18. Control group sample size (needed for all methods of computing effect sizes) Method 1 19. Treatment group mean 20. Control group mean 21. Treatment group standard deviation 22. Control group standard deviation Method 2 23. Treatment group mean 24. Control group mean 25. Treatment group standard error 26. Control group standard error Method 3 27. t‐value (t‐test from an independent t comparing two means) 28. P‐value from above (only needed if t‐value is not available Method 4 29. Frequency of failures (or successes) in treatment group (must be a fraction of sample size) 30. Frequency of failures (or successes) in control group (must be a fraction of sample size) Method 5 31. Proportion of failures (or successes) in treatment group (must be a proportion of sample size) 32. Proportion of failures (or successes) in control group (must be a proportion of sample size) Method 6 33. Counts of events (or event rate) in treatment group 34. Counts of events (or event rate) in control group Method 7 35. Unstandardized regression coefficient for treatment dummy variable (B) (OLS or similar regression) 36. Standard error for the unstandardized regression coefficient for treatment dummy 37. Standardized regression coefficient 38. Standard deviation for the dependent variable Method 8 39. Unstandardized regression coefficient for treatment dummy variable (B) (logistic regression) 40. Standard error for the unstandardized regression coefficient for treatment dummy 41. Odds‐ratio Method 9 42. Hand calculated Cohen's d effect size 43. Hand calculated variance (standard error squared) for Cohen's d effect size 1 While the GPD data extends back to 1950, to date full‐text documents have only been screened back to 2003. For this systematic review, the authors believe the use of the GPD is justified, as the earliest recorded evaluation for BWCs according to Lum et al. was Goodall (2007). Additionally, per the GPD search protocol, grey literature will also be searched (see Appendix B). 2 See https://bwctta.com/resources/bwc‐resources/impacts‐bwcs‐use‐force‐directory‐outcomes and https://bwctta.com/resources/bwc‐resources/impact‐bwcs‐citizen‐complaints‐directory‐outcomes. ==== Refs REFERENCES Ariel, B. , Sutherland, A. , Henstock, D. , Young, J. , Drover, P. , Sykes, J. , … Henderson, R. (2017). “Contagious accountability”: A global multisite randomized controlled trial on the effect of police body‐worn cameras on citizens’ complaints against the police. Criminal Justice and Behavior, 44 (2 ), 293–316. Braga, A. A. , Sousa, W. H. , Coldren, J. R., Jr. , & Rodriguez, D. (2018). The effects of body‐worn cameras on police activity and police‐citizen encounters: A randomized controlled trial. Journal of Criminal Law and Criminology, 108 (3 ), 511–538. Chan, J. , Brereton, D. , Legosz, M. , & Doran, S. (2001). E‐policing: The impact of information technology on police practices. Brisbane, Australia: Queensland Criminal Justice Commission. Colton, K. W. (1980). Police and computer technology: The case of the San Diego computer‐aided dispatch system. Public Productivity Review, 4 (1 ), 21–42. Gill, C. , Weisburd, D. , Telep, C. W. , Vitter, Z. , & Bennett, T. (2014). Community‐oriented policing to reduce crime, disorder and fear and increase satisfaction and legitimacy among citizens: A systematic review. Journal of Experimental Criminology, 10 (4 ), 399–428. Glass, G.V. (1997). Interrupted time series quasi‐experiments. In Jaeger, R.M. (Ed.), Complementary methods for research in education (2nd ed., pp. 589–608). Washington, DC: American Educational Research Association. Goodall, M. (2007). Guidance for the police use of body‐worn video devices: Police and crime standards directorate, London, UK: Home Office. Higginson, A. , Eggins, E. , Mazerolle, L. , & Stanko, E. (2015). The Global Policing Database [Database and Protocol]. Retrieved from http://www.gpd.uq.edu.au Hyland, S. S. (2018). Body‐worn camera in law enforcement agencies, 2016. U.S. Department of Justice, Bureau of Justice Statistics. Retrieved from https://www.bjs.gov/content/pub/pdf/bwclea16.pdf Koper, C. S. , Lum, C. , Willis, J. , Woods, D. J. , & Hibdon, J. (2015). Realizing the potential of technology in policing: a multisite study of the social, organizational, and behavioral aspects of implementing police technologies. Washington, DC: National Institute of Justice. Lipsey, M. W. , Wilson, D. B. (2001). Practical meta‐analysis. Thousand Oaks, CA: Sage. Lum, C. , Hibdon, J. , Cave, B. , Koper, C. S. , & Merola, L. (2011). License plate reader (LPR) police patrols in crime hot spots: An experimental evaluation in two adjacent jurisdictions. Journal of Experimental Criminology, 7 (4 ), 321–345. Lum, C. , & Koper, C. S. (2017). Evidence‐based policing: Translating research into practice. Oxford, U.K.: Oxford University Press. Lum, C. , Koper, C. S. , Merola, L. , Scherer, A. , & Reioux, A. (2015). Existing and Ongoing Body Worn Camera Research: Knowledge Gaps and Opportunities (Report for the Laura and John Arnold Foundation). Fairfax, VA: Center for Evidence‐Based Crime Policy, George Mason University. Lum, C. , Koper, C. S. , & Willis, J. (2017). Understanding the limits of technology's impact on police effectiveness. Police Quarterly, 20 (2 ), 135–163. Lum, C. , Stoltz, M. , Koper, C. S. , & Scherer, J. A. (2019). The research on body‐worn cameras: What we know, what we need to know. Criminology and Public Policy, 18 (1 ), 93–118. Manning, P. K. (2008). The Technology of Policing. New York: New York University Press. Maskaly, J. , Donner, C. , Jennings, W. G. , Ariel, B. , & Sutherland, A. (2017). The effects of body‐worn cameras (BWCs) on police and citizen outcomes: A state‐of‐the‐art review. Policing: An International Journal of Police Strategies & Management, 40 (4 ), 672–688. McClure, D. , La Vigne, N. , Lynch, M. , Golian, L. , Lawrence, D. , & Malm, A. (2017). How Body Cameras Affect Community Members' Perceptions of Police: Results From a Randomized Controlled Trial of One Agency's Pilot. Washington, DC: Justice Policy Center, Urban Institute. Retrieved from urban.org/sites/default/files/publication/91331/2001307‐how‐body‐cameras‐affect‐community‐members‐perceptions‐of‐police_1.pdf Morrow, W. J. , Katz, C. M. , & Choate, D. E. (2016). Assessing the impact of police body‐worn cameras on arresting, prosecuting, and convicting suspects of intimate partner violence. Police Quarterly, 19 , 303–325. Neville, R. , & Higginson, A. (2014). Optimal Search of Indexing Databases (OSID) (Computer Software). Brisbane, Australia: University of Queensland. Nowacki, J. S. , & Willits, D. (2018). Adoption of body cameras by United States police agencies: An organisational analysis. Policing & Society: An International Journal of Research and Policy, 28 (7 ), 841–853. Orlikowski, W. J. , & Gash, D. C. (1994). Technological frames: Making sense of information technology in organizations. ACM Transactions on Information Systems, 12 (2 ), 174–207. Owens, C. , Mann, D. , & Mckenna, R. (2014). The Essex body worn video trial: The impact of body worn video on criminal justice outcomes of domestic abuse incidents, Ryton‐on‐Dunsmore, Coventry, England: U.K. College of Policing. Petrosino, A. , Turpin‐Petrosino, C. , & Guckenburg, S. (2014). The impact of juvenile system processing on delinquency. In D. Farrington and J. Murray (eds.) Labeling Theory—Empirical Tests. Advances in Criminological Theory, 18 : 113‐147. White, M. D. (2014). Police officer body‐worn cameras: Assessing the evidence. Washington, DC: Office of Community Oriented Policing Services. White, M. D. , Todak, N. , & Gaub, J. E. (2017). Assessing citizen perceptions of body‐worn cameras after encounters with police. Policing: An International Journal of Police Strategies and Management, 40 (4 ), 689–703.
PMC008xxxxxx/PMC8484611.txt
==== Front Nat Commun Nat Commun Nature Communications 2041-1723 Nature Publishing Group UK London 34593804 26053 10.1038/s41467-021-26053-w Article Efficient CO2 electroreduction on facet-selective copper films with high conversion rate http://orcid.org/0000-0002-9100-4810 Zhang Gong 123 http://orcid.org/0000-0002-8856-5078 Zhao Zhi-Jian 123 http://orcid.org/0000-0003-3509-699X Cheng Dongfang 123 Li Huimin 123 Yu Jia 123 Wang Qingzhen 123 Gao Hui 123 Guo Jinyu 4 Wang Huaiyuan 123 http://orcid.org/0000-0002-6315-0925 Ozin Geoffrey A. 5 http://orcid.org/0000-0002-9862-5038 Wang Tuo [email protected] 123 http://orcid.org/0000-0001-7263-318X Gong Jinlong [email protected] 1236 1 grid.33763.32 0000 0004 1761 2484 School of Chemical Engineering and Technology, Tianjin University, Tianjin, 300072 China 2 grid.33763.32 0000 0004 1761 2484 Key Laboratory for Green Chemical Technology of Ministry of Education, Tianjin University, Tianjin, 300072 China 3 grid.509499.8 Collaborative Innovation Center of Chemical Science and Engineering (Tianjin), Tianjin, 300072 China 4 grid.168010.e 0000000419368956 Department of Chemical Engineering, Stanford University, Stanford, CA USA 5 grid.17063.33 0000 0001 2157 2938 Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, ON Canada 6 grid.33763.32 0000 0004 1761 2484 Joint School of National University of Singapore and Tianjin University, International Campus of Tianjin University, Binhai New City, Fuzhou 350207 China 30 9 2021 30 9 2021 2021 12 57453 6 2021 6 9 2021 © The Author(s) 2021, corrected publication 2023 https://creativecommons.org/licenses/by/4.0/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Tuning the facet exposure of Cu could promote the multi-carbon (C2+) products formation in electrocatalytic CO2 reduction. Here we report the design and realization of a dynamic deposition-etch-bombardment method for Cu(100) facets control without using capping agents and polymer binders. The synthesized Cu(100)-rich films lead to a high Faradaic efficiency of 86.5% and a full-cell electricity conversion efficiency of 36.5% towards C2+ products in a flow cell. By further scaling up the electrode into a 25 cm2 membrane electrode assembly system, the overall current can ramp up to 12 A while achieving a single-pass yield of 13.2% for C2+ products. An insight into the influence of Cu facets exposure on intermediates is provided by in situ spectroscopic methods supported by theoretical calculations. The collected information will enable the precise design of CO2 reduction reactions to obtain desired products, a step towards future industrial CO2 refineries. Regulation of Cu facets to promote electrocatalytic CO2 reduction is interesting and challenging. Here the authors describe a deposition-etch-bombardment synthetic approach to prepare Cu(100)-rich thin film electrodes for CO2 electroreduction with over 50% ethylene Faradaic efficiency at a total current of 12 A. Subject terms Electrocatalysis Solar fuels Electrochemistry https://doi.org/10.13039/501100001809 National Natural Science Foundation of China (National Science Foundation of China) 22038009 Gong Jinlong issue-copyright-statement© Springer Nature Limited 2021 ==== Body pmcIntroduction The renewable electricity-powered CO2 reduction has been considered as one promising route to the product of chemical feedstocks, which might close the carbon loop1,2. Multi-carbon (C2+) products such as ethylene, ethanol, n-propanol, etc. are important raw materials in the chemical industry or can be used directly as fuels, thus the efficient CO2 reduction to C2+ products system is essential to the production of high-value commodity chemicals with a net negative carbon emissions footprint. Among various CO2 reduction catalyst materials, Cu has been studied extensively as it can produce C2+ products with appreciable selectivity3. Among numerous ways to promote C2+ formation4–8, adjusting the facet of Cu-based catalysts is an effective method since the dimerization reaction is facet sensitive9. According to the theoretical calculations, the Cu(100) facet can significantly lower the dimerization energy barrier10. Thus, it is an effective approach to improve the selectivity toward C2+ by designing Cu catalysts with Cu(100) as the dominant exposed facet10. Colloidal chemistry is an inexpensive, simple, and widely used method for facet exposure control. It produces catalysts with preferential exposure of specific facets by using capping agents to manipulate surface energies, which changes the growth rates of different facets to alter the shape of nanocatalysts11,12. Recently, colloidally synthesized Cu nanocubes rich in Cu(100) facets were shown to achieve a Faradaic efficiency (F.E.) of ~57% towards ethylene13. However, the corresponding electricity conversion efficiency (E.C.E.) and single-pass yield have been rarely reported. Some studies also propose that the intermediates along the CO2 reduction pathways can control the formation of specific facets, where the adsorption of the intermediates plays a role analogous to that of capping agents14,15. Nonetheless, these approaches require specific chemicals, such as capping agents, to be selectively adsorbed on particular facets to reduce the surface energy11,12. However, the effectiveness of the colloidal method might compromise to some extent since low-index facets of fcc transition metals often possess similar surface energies (1.25 J cm−2 for Cu(111) and 1.43 J cm−2 for Cu(100))15,16. Besides, residual capping agents left on the catalyst surface could block catalytically active sites17,18. Moreover, the colloidally synthesized catalyst needs to be dispersed in a solution containing polymer binders (such as Nafion) to form a well-mixed ink before being drop-casted on the conductive support to form an electrode, which is not always compatible with the catalytic electrode system due to agglomeration and the peeling-off of catalysts especially when scaling up the electrodes19,20. At the same time, due to the existence of the cross-linked network formed by polymer binders21, the contact between the colloidally synthesized catalyst and the conductive substrate is weak22,23, and plenty of active sites are further encapsulated20, which will lead to slow electron transfer and low E.C.E. Thus, it is highly desirable to develop a novel approach to replace colloidal synthesis to prepare Cu electrodes with dominant (100) facets in one-step without using capping agents and polymer binders to achieve a high E.C.E. and a single-pass yield. In this work, we describe the design and realization of a dynamic deposition-etch-bombardment process to produce Cu(100)-rich films as the CO2 reduction electrode in one-step, which could break the limitation of using capping agents while avoiding the issue of catalyst loading faced by conventional methods. This Cu(100)-rich film yields a full-cell electricity conversion efficiency of 40.5% towards C2+ products in the 4 cm2-membrane electrode assemble (MEA) system. This facile one-step method also makes it easier for scaling up electrodes, with a 25 cm2 electrode exhibiting a 13.2% single-pass yield of C2+ product at a total current of 12 A in the MEA system. Thanks to the controllable structure, the influence of facets on the adsorption and activation of intermediates is further revealed by in situ spectroscopy and density functional theory (DFT) calculation. Results To amplify, when Cu atoms with high kinetic energy bombard the deposited Cu film, the high local temperature causes dynamic recrystallization24,25. Facets with the relatively loose atomic arrangement, like Cu(100) (Supplementary Fig. 1c), would receive less damage and remain at lower temperatures thereby acting as recrystallization centers26,27. As a result, the Cu(100) will preferentially grow, replacing the more densely packed Cu(111) facet (Supplementary Fig. 1b). Therefore, by controlling the kinetic energy of bombarding copper atoms, the exposure ratio of Cu(100) facets could be adjusted. To realize the simultaneous deposition, etching, and bombardment, high-energy radio frequency (RF) sputtering was adopted to prepare Cu(100)-rich films (details in the “Methods”), which enables the direct deposition of catalysts on carbon-based gas diffusion layers (GDLs) as electrodes in one-step (Supplementary Fig. 1a). Three typical types of sputtered Cu films with different Cu(100) proportions were obtained by adjusting the RF power to control the kinetic energy (Ek) of the bombarding Cu atoms (Fig. 1a, details in “Methods”), resulting in low-power, medium-power, and high-power sputtered Cu films (denoted as LS-Cu, MS-Cu, HS-Cu, respectively). The obtained HS-Cu film is prone to expose the Cu(100) facet, as evidenced by transmission electron microscopy (TEM, Supplementary Fig. 2a) and X-ray diffraction (XRD, Supplementary Fig. 3a), while LS-Cu tends to form Cu(111) facet (Supplementary Figs. 2c and 3a). The energy of Cu atoms generated by thermal evaporation2 or low-rate sputtering process4,28 is usually not high enough, rendering Cu films that favor the Cu(111) facet exposure similar to LS-Cu. As different Cu facets feature distinctive OH− electrochemical adsorption behaviors29,30, surface structures of these samples were probed by using the OH− electroadsorption technique (the features were labeled by comparing the CVs to those of single crystals shown in Supplementary Fig. 4)31. The cyclic voltammograms (CVs) of OHads peaks (Supplementary Fig. 3b) reveal that the presence of Cu(111) is suppressed in the HS-Cu film. Therefore, this deposition-etch-bombardment process successfully realizes the replacement of Cu(100) to Cu(111) facets, adjusting the facet exposure of Cu.Fig. 1 characterization of various Cu films. a A schematic illustration of the synthesis process for LS/MS/HS-Cu and LRS/MRS/HRS-Cu. b Voltammograms of resulting electrodes collected immediately after pre-reduction. Typical TEM images of c HRS-Cu, d MRS-Cu, and e LRS-Cu. f Typical XRD patterns of resulting electrodes. As visual aids, the red dash line indicates the surfaces of the resulting electrodes. The precatalysts of HRS/MRS/LRS-Cu were pre-reduced under −0.5 V versus the RHE for 1 h, and the other conditions of prereduction are the same as that of the CO2 reduction. The labeling of Cu(200) is used for easier comparison with XRD, in which only (200), the second-order diffraction of (100) could be detected. The peaks marked with an asterisk in XRD patterns originate from carbon-based GDL substrate. However, these obtained Cu films (i.e. LS-Cu, MS-Cu, HS-Cu) lack a large electrochemically active surface area (ECSA, Supplementary Fig. 26 and Supplementary Table 1) to realize high-rate CO2 electrolysis32. To increase the ECSA, an oxidation−reduction procedure was included, during which the preferential exposure of Cu(100) could be maintained. Specifically, molecular O2 (with a constant partial pressure of 0.67 Pa) was introduced during this deposition-etch-bombardment process to obtain Cu2O as precatalysts, which could be further reduced during the pre-reduction process (−0.5 V vs. the RHE) under the same CO2 reduction conditions to metallic Cu with the Cu(100) facets retained (Fig. 1a, details in “Methods”). The same crystallographic relationship exists between the oxidation and reduction processes for Cu(100)/(111) and Cu2O(100)/(111)10,33, since oxygen atoms could pack into the interlayer sites of different planes of Cu, causing a lattice expansion while retaining the underlying fcc crystal structure with the inclusion of oxygen33. Therefore, the introduction of an oxidation-reduction step will be able to increase the ECSA while retaining the same Cu(100) exposure. Correspondingly, samples obtained by reactive sputtering followed by reduction are referred to as low-power, medium-power, and high-power reactively sputtered Cu films followed by reduction (denoted as LRS-Cu, MRS-Cu, and HRS-Cu). Typical XRD patterns (Supplementary Fig. 5a) show that all the obtained precatalysts are mainly Cu2O, which is confirmed by X-ray photoelectron spectroscopy (XPS), Auger spectroscopy (Supplementary Fig. 5b, c), and Raman spectroscopy (Supplementary Fig. 4d). The same type of Cu facets distribution could be maintained for these electrodes regardless of the oxidation-reduction process according to TEM (Fig. 1c–e, more images of different samples are also provided in Supplementary Figs. 7–9) and XRD (Fig. 1f), consistent with previous reports10,34. The different proportions of Cu(100) facet on these samples (i.e. HRS-Cu, MRS-Cu, and LRS-Cu) are also evidenced by the electrochemical OH− adsorption peaks on Cu(100) and Cu(111) at potentials of ~0.33 and 0.43 V versus the RHE (Fig. 1b). Meanwhile, large ECSA is obtained for these samples (Supplementary Table 1). Thus, this one-step surfactant-free route indeed leads to nanostructured Cu films with preferred Cu(100) exposure and high surface area that could be further used as CO2 reduction electrodes. To further explore the importance of the high-energy atom bombardment proposed above, another control electrode was prepared. Cu2O precatalyst (Supplementary Fig. 10) was obtained through wet chemistry without using capping agents (denoted as W-Cu). Then this control Cu2O precatalyst was airbrushed onto GDEs and pre-reduced to Cu under CO2 reduction conditions to form a control electrode (details in “Methods”). The typical TEM image shows that the lattice fringes corresponding to Cu(111) are widely distributed in this control electrode (Supplementary Fig. 11a)32. These observations are in good agreement with the XRD pattern (Supplementary Fig. 11b) and CV of OHads peaks (Supplementary Fig. 11c). It is also worth noting that LRS-Cu, MRS-Cu, HRS-Cu, and W-Cu all consist of metallic Cu only without residual oxides after the pre-reduction, as confirmed by in-situ Raman spectroscopy (Supplementary Fig. 12, details in “Methods”). Thus, this control experiment reveals that the lack of high-energy atom bombardment during the synthesis process would result in the dominant exposure of Cu(111) facets, similar to the scenario of lacking capping agents that are widely used in the synthesis of Cu nanocubes34,35. Before performance testing, it was confirmed that different samples (LRS-Cu, MRS-Cu, HRS-Cu, and W-Cu) possess similar mass loadings (Supplementary Table 2), catalyst layer thickness (Supplementary Fig. 13) and morphology (Supplementary Fig. 14) of Cu, which ensures a fair comparison to explore the activity difference among various samples36–39. The activities were evaluated at different potentials using 2 M KOH (aq.) as the electrolyte (Fig. 2a and Supplementary Figs. 16–19, details in “Methods”) in a flow cell electrolyzer (Supplementary Fig. 15) for CO2 reduction with an effective electrode geometric area of ~0.64 cm2 for both cathode and anode. As for HRS-Cu, the products detected in significant quantities were ethylene, ethanol, n-propanol, and CO. At more negative potentials, a small amount of methane was produced, and the remaining charge was attributed to the competing hydrogen evolution reaction (Fig. 2b). Compared with the control samples, the HRS-Cu exhibits a maximum F.E. of 58.6% for ethylene, 86.6% for C2+ products (containing ethylene). Consistently, the HRS-Cu sample also exhibits the largest ethylene and C2+ products partial current densities (Fig. 2c, d and Supplementary Fig. 17c, d) among the four samples at all applied potentials. Moreover, at the applied potential of −0.85 V versus the RHE, the C2+-to-C1 ratio of the HRS-Cu electrode reaches about 15.2, which largely outperforms that of its counterparts (Supplementary Fig. 18). Due to the close contact between the catalytic Cu film and the GDL substrate, the reaction system could exhibit higher E.C.E. The corresponding full-cell E.C.E. towards ethylene and C2+ products of this HRS-Cu-based reaction system reaches 24.8 and 36.5%, respectively, exceeding the efficiency of many other reaction systems reported so far (see below). The stability of the HRS-Cu was also examined under a constant applied potential of −0.75 V versus the RHE for 4.5 h (270 min), where ethylene and C2+ F.E.s remained stable over the test duration (Fig. 2e and Supplementary Fig. 19). Although the surface of the Cu-based catalysts may undergo reconstruction during the reaction to obtain Cu(100)40, the HRS-Cu, MRS-Cu, and LRS-Cu samples after use show similar surface exposures as before the reaction (Supplementary Figs. 20–22), indicating that the reconstruction process could not significantly change the main facet exposure, and thus surface reconstruction alone cannot guarantee electrodes similar to the HRS-Cu (showing a predominant Cu(100) exposure without undergoing reconstruction). The mass of the HRS-Cu sample after the reaction is also almost the same as that before the reaction (Supplementary Table 3), which further indicates that the firm conjunction between the substrate and the active composition. If the hydrophobicity of GDLs can be further improved2 and the reaction rate of carbonate formation can be reduced41, the stability of the sample in the flow cell will be longer.Fig. 2 Catalytic performance of various Cu films. a Current−voltage (I–V) curves on various electrodes measured in CO2-flowed 2 M KOH electrolytes. b Faradaic efficiencies of CO2 reduction products on the HRS-Cu sample as a function of different applied potentials. c Ethylene and d C2+ products partial current densities obtained on various samples at different applied potentials. e Stability test over a span of 4.5 h (270 min) of CO2− electrolysis in 2 M KOH (aq.) at −0.75 V versus the RHE. The oscillation of the current density is due to the repeated release of O2 bubbles in the anode side of the flow cell. C2+ products include ethylene, ethanol, and n-propanol. Error bars represent the standard deviation from at least three independent measurements. To examine the influence of Cu(100) exposure over C2+ products formation in CO2 reduction, Tafel analyses (Supplementary Fig. 23) were conducted on various samples. It is clear that C2+ products formation depends strongly on the surface structure. With the increasing exposure of Cu(100) facets, a lower change of the Tafel slope can be observed, which indicates that the Cu(100) surface is beneficial to the formation of C2+ products. However, the Tafel slope of all samples locates at approximately 120−140 mV dec−1, implying that they might share the same rate-determining step (details in “Methods”), which is further validated by the results of CO partial pressure dependence experiments (Supplementary Figs. 24, 25). To further compare the reaction rates, the performances are normalized to the ECSA (details in “Methods”). Obviously, the current densities of specific ethylene and C2+ products of HRS-Cu are larger than other electrodes (Supplementary Fig. 27), which indicates the highest intrinsic activities of HRS-Cu. Understanding the behaviors of adsorbates is essential to the inquiry into the nature of the catalytic activity. Thus, in-situ attenuated total reflectance (ATR)-surface-enhanced infrared absorption spectroscopy (ATR-SEIRAS, details in “Methods”) was applied to investigate the effect of facets exposure on the adsorption of intermediates. When the potential is swept from −0.1 to −1.5 V versus the RHE, a positive band centered at ~2050 cm−1 which corresponds to the linear-bond CO (COL, a reactive adsorbed species) is observed on all electrodes (Fig. 3a–c and Supplementary Fig. 28)42–44. Meanwhile, a small positive band centered at ~1800 cm−1, which corresponds to the bridge-bond CO (COB, an unreactive adsorbed species)44 is also observed on the surface of other samples except the HRS-Cu. These results of the ATR-SEIRAS are also corroborated by the corresponding Raman spectra, where the HRS-Cu surface was covered by abundant COL45,46, while the LRS-Cu surface exhibited a pronounced appearance of COB (Supplementary Fig. 30). The appearance of COB might result from the interconversion of COL. According to the literature, when the *CO (*denotes the adsorbed species) coverage of the electrode decreases, a fraction of COL would convert to COB44. Therefore, the increase of COB ratio reveals the reduction of surface *CO coverage. By calculating the COB to COL ratio of different samples, the *CO coverage might decrease as the exposure of Cu(100) facets decreases (Supplementary Fig. 29). Normally, *H and *CO occupy most of the surface sites, so they are in direct competition with each other for surface sites. Meanwhile, the surface coverage of *CO eventually influences the distribution of products derived from CO2 reduction. It has been previously speculated that when the surface cannot maintain a high *CO coverage, the corresponding *H coverage will increase, inhibiting the dimerization of *CO to produce C2+ products, thereby shifting the selectivity to C1 products and H247–49, which is also demonstrated by the results of our CO partial pressure dependence experiments (Supplementary Figs. 24, 25). Meanwhile, the results of CO2 reduction performance support the above speculation. The LRS-Cu and the W-Cu with a lower exposure of Cu(100) facets exhibit the highest F.E. of methane, meanwhile, the methane F.E. increases rapidly as the potential becomes negative (Supplementary Fig. 31). Thus, these ATR-SEIRAS results prove that Cu(100) facet might be a type of strong *CO adsorption site, and a higher exposure of Cu(100) facets is beneficial to the increase of strong adsorption sites on the surface that maintains a higher *CO coverage on the surface, thereby leading to the catalyst with a higher C−C coupling performance, which is also supported by other experimental and computational studies50–52.Fig. 3 Spectroscopic investigations of various Cu films and DFT calculations on different Cu facets. In situ ATR-SEIRAS spectra of a HRS-Cu; b MRS-Cu; c LRS-Cu; d adsorption geometry for different C−C coupling precursors (i.e., COL−COL, COB−COL, COB−COB), where solvent molecules are not presented to show the adsorbate configurations; e reaction barriers for C−C coupling from different precursors on Cu(100) and Cu(111) facets. Since the ATR-SEIRAS cannot directly provide the adsorption energy and activation barrier of the intermediate, DFT calculations on Cu(111) and Cu(100) (3×3) models were performed to further understand the facet exposure effect (Supplementary Fig. 32, details in “Methods”). Based on the Tafel analysis and CO partial pressure dependence study (Supplementary Figs. 23, 24), the energetics of *OCCO formation is chosen as the main consideration. Previous work has found that the solvent and cation effects can stabilize *COCO, hence an appropriate electrochemical interface was built up to explore the mechanism of this dimerization procedure (Fig. 3d and Supplementary Figs. 34–36)53. From the DFT calculation results, the binding of *CO is the weakest on the Cu(111) facet (Supplementary Fig. 33), which is not suitable to build up a sufficient coverage of *CO on the surface to promote the kinetics of C−C coupling. Therefore, LRS-Cu and W-Cu would shorten the *CO stay and reduce the *CO coverage as compared to HRS-Cu, leading to a higher CO F.E. (Supplementary Fig. 16). In addition, the barriers of various C−C coupling processes are always higher on the Cu(111) facet than on Cu(100) facet, implying that *CO dimerization is the most sluggish step on the Cu(111) facet (Fig. 3e). Thus, LRS-Cu and W-Cu are less active in catalyzing the reduction of CO2 to C2+ products. These theoretical calculations are in good agreement with our ATR-SEIRAS and performance test results. Due to the impressive E.C.E of this HRS-Cu-based reaction system, a PV-EC system (Fig. 4a) was further constructed to demonstrate the photosynthesis of C2+ products. Using the same test conditions (electrolyzer, cathode, anode, and electrolyte, details in “Methods”) for measuring the CO2 reduction performance on the cathode side (as in Fig. 2), it is found that an overall cell voltage of about 2.5 V is required (Fig. 4b) to obtain an operating current density varying from 60 to 70 mA cm−2 (Fig. 2a), which could be translated to a cathodic potential of ~−0.55 V versus the RHE that yields ~45 and ~72% F.E. for ethylene and C2+ products (containing ethylene), respectively (Fig. 2b). The deviation of these F.E.s from the maximized F.E.s for ethylene and C2+ products is because of the matching between electrolyzer and the solar panel towards the maximum power point (MPP) of the solar panel. Considering that the effective area of the HRS-Cu sample is ~0.64 cm2, the operating current varies from 38.4 to 44.8 mA. The widely available p−n+ solar cells, with an open circuit potential of ~0.6 V and a short-circuit current density of ~36 mA cm−2, were selected and cut into ~1.14 cm2 pieces. By connecting five of them in series, the solar panel would provide the suitable voltage and current for the CO2 reduction electrolyzer near the MPP of the solar panel (i.e., the PV system). I−V curve of the obtained solar panel was measured under simulated AM 1.5G 1-sun illumination (Fig. 4b). This curve crosses the I−V curve of the electrolyzer at the point (i.e., the operating point) where its cathodic current and voltage are ~41.3 mA and ~2.41 V, respectively, which matches well with the MPP output metrics of the solar panel (39.8 mA at 2.53 V), indicating the optimum solar-to-electricity conversion process in our integrated device.Fig. 4 Solar-powered CO2 reduction. a Schematic of the PV-EC system. b Photovoltaic and electrocatalytic I-V behaviors. The photovoltaic performance is shown under light (green) with the MPP marked by a red dot. The measured operating current of the CO2 electrolysis system (cathode, anode, and anion exchange membrane) at different voltages has been marked by the purple curve. The observed long-term operating point is marked by a blue dot, with the black dashed lines showing the corresponding current and voltage. c Faradaic efficiency towards ethylene and C2+ products, solar current, and solar conversion efficiency as a function of reaction time. Carbon feedstocks include CO, methane, ethylene, ethanol, and n-propanol. C2+ products include ethylene, ethanol, and n-propanol. During ~3.7 h (220 min) of electrolysis powered by 1-sun solar illumination, the cathodic current of the electrolyzer was stable at ~41.3 mA and the F.E. of ethylene and C2+ products stabilized at ~45 and ~72%, respectively (Fig. 4c and Supplementary Figs. 37–39). Although this selectivity is already deviated from the maximum F.E.s owing to MPP matching, such a PV-EC system still yields a solar-to-ethylene efficiency of ~4.0% and a solar-to-C2+ products efficiency of ~6.0% (Fig. 4c) under simulated 1 Sun illumination, which exceeds the efficiency of general natural photosynthesis for producing carbohydrates (3−6%)54. Such efficiency is also able to match a recent state-of-the-art perovskite solar cells-powered Cu−Ag bimetallic reaction system55, while a longer stability and larger CO2 reduction current are obtained in this system. The performance of this work also exceeds that of most previously reported copper-based PV-EC systems (Supplementary Table 4), which provides a benchmark for solar conversion efficiency while using feasibly available Si-based solar cells and earth-abundant low-cost electrode materials. The membrane electrode assemble system (MEA) was further adopted to scale up the CO2 reduction system (Supplementary Fig. 40, details in “Methods”). On the cathode side, humidified CO2 gas was supplied (Fig. 5a), which reduces the direct contact between the catalyst and the aqueous electrolyte, while reducing the ohmic resistance of the electrolyte56,57. At the same time, the one-step deposition-etch-bombardment process proposed in this work is based on a widely used vacuum deposition process similar to the photovoltaic industries, thus its marginal cost for producing large-area electrodes can be largely reduced with mass production. In addition, this method does not require additional catalyst loading steps and does not use additional chemicals, accelerating the continuous preparation of large-area electrodes and avoiding potential electrode contamination. Therefore, the combination of a MEA reaction system and this preparation method is particularly suitable for the scale-up of CO2 reduction.Fig. 5 Scaling up the CO2 reduction system. a Schematic of the MEA system. The gaskets are not shown. b Stability test over a span of ~4.5 h of CO2-electrolysis in a 4 cm2-MEA system at the total current density of 120 mA cm−2. c Stability test over a span of ~3.5 h of CO2-electrolysis in a 25 cm2-MEA system at the total current of 12 A. The decrease in ethylene F.E. and the increase in cell voltage may be attributed to the formation of carbonate on the cathode side and the cathode water flooding. The electrode area was enlarged to 4 cm2, and the CO2 reduction performance test was performed in the MEA system (the area of cathodic flow fields is 4 cm2, referred as to the 4 cm2-MEA) at a current density of 120 mA cm−2 corresponding to the maximum ethylene F.E. (58.6%) in the flow cell system (Fig. 5b). The optimal ethylene F.E. reaches 55.8%, while the corresponding full-cell E.C.E of ethylene and C2+ products increase to 26.4 and 40.2%, respectively. After ~4.5 h of operation (Supplementary Fig. 41), the ethylene selectivity (50.4%) still remains above 90% of the initial value (55.8%). Although many works have obtained impressive CO2 reduction selectivity, the single-pass yield for C2+ products is generally low (mostly below 3%, Supplementary Table 5). In order to improve the single-pass yield of C2+ products, it is vital to further increase the current density and electrode area without affecting the transport of CO2 (details in the “Methods”). Using the one-step method proposed in this work, a HRS-Cu electrode with 25 cm2 activity area could be easily fabricated with the deposition-etch-bombardment process (Supplementary Fig. 42). In a 25 cm2 MEA system (i.e., the area of cathodic flow fields is 25 cm2), the optimal ethylene F.E. of HRS-Cu reached 50.9% and maintained above 45% (Fig. 5c and Supplementary Fig. 43) after ~3.5 h of operation at a total current of 12 A (i.e., 480 mA cm−2). The corresponding single-pass yield of CO2 towards C2+ products increased to and 13.2%, with ethylene increased to 12.0%. However, a corresponding decrease in E.C.E. was observed (Supplementary Table 5). Therefore, using more efficient anode materials, while controlling the operating conditions of the device such as temperature and pressure, or designing a new flow field structure to enhance CO2 mass transfer may be effective ways to further improve the E.C.E. Discussion In conclusion, this work demonstrates a deposition-etch-bombardment process that feasibly increases the exposure of Cu(100) facets in low-cost nanostructured Cu films. This strategy obviates the use of capping agents, achieving precise control of surface structures in a one-step approach. The obtained Cu(100)-rich film functions as a high-performance electrode for CO2 reduction towards ethylene and C2+ products. In flow cell, it realized ethylene and C2+ (containing ethylene) F.E.s of 58.6 and 86.6%, respectively. In addition, this deposition-etch-bombardment method bypasses the electrode assembly process, avoids the use of polymer binders, enhances the contact between the catalytic film and the substrate. Due to the above advantages, in flow cell, a corresponding full-cell E.C.E. of 24.8 and 36.5% for ethylene and C2+ products were obtained, respectively, which is a notable advance over existing single metallic Cu-based systems. Moreover, this preparation method is flexible and easy to achieve electrode scale-up. When a 4 cm2 electrode was applied to the MEA system, the corresponding full-cell E.C.E. could be increased to 40.5 and 26.8%, respectively, without compromising product selectivity. When the electrode area of the MEA system is increased to 25 cm2, the single-pass yield of C2+ products can be further increased to 13.2%, with an ethylene yield of 12.0%. In situ ATR-SEIRAS studies and theoretical calculations provide insights into the role of Cu(100) facets for increasing *CO coverage and reducing the energy barrier of C−C coupling, vital for ethylene and C2+ products formation. The potential of using this Cu(100)-rich film for photosynthesis was also demonstrated using renewable electricity generated by Si solar panels, achieving a solar-to-C2+ products efficiency of ~6.0% under simulated 1 Sun illumination. Future research endeavors may focus on the generation of energetic particles (atoms, molecules, ions, etc.) in other environments such as solutions in ambient conditions, which may further reduce the cost of one-step facet control by this vacuum-based deposition-etch-bombardment method. Methods Synthesis of LS-Cu, MS-Cu, HS-Cu The deposition-etch-bombardment process was conducted in a custom-designed radio frequency (RF, 13.56 MHz) magnetron sputtering system (Supplementary Fig. 1a). For LS-Cu, Ar was delivered into the system. The deposition chamber was pumped down by a mechanical pump and a molecular pump that resulted in a base pressure of 2.0 × 10−4 Pa. The flow rate of Ar was set as 20 standard cubic centimeters per minute (sccm). During the deposition, the RF power was 40 W, and the working pressure was 4 Pa. The deposition time was 15 min. The target to substrate distance was set to 7 cm. Commercial GDLs were cut into squares (5 × 5 cm) for use as substrates. For MS-Cu, the RF power was set as 100 W and the deposition time was 7.5 min, other conditions were the same as LS-Cu. For HS-Cu, the RF power was set as 200 W and the deposition time was 3.5 min, other conditions were the same as LS-Cu. Synthesis of LRS-Cu, MRS-Cu, and HRS-Cu During the synthesis process of precatalysts of LRS-Cu, MRS-Cu, and HRS-Cu, O2 was also introduced to the Ar atmosphere. The flow rates of Ar and O2 were 20 and 4 sccm, respectively, to achieve a partial pressure of O2 of 0.67 Pa. Other conditions were the same as those of LS-Cu, MS-Cu, and HS-Cu. The obtained precatalysts were pre-reduced at −0.5 V versus the RHE for 1 h under the same conditions as CO2 electroreduction. Pre-reduction was carried out with a potentiostat (CompactStat.e20250, IVIUM). After pre-reduction, the final samples (i.e., LRS-Cu, MRS-Cu, and HRS-Cu) were obtained. Synthesis of W-Cu The W-Cu was fabricated through the electrochemical pre-reduction of the Cu2O nanorods at −0.5 V versus the RHE for 1 hour under the same conditions as CO2 electroreduction. After pre-reduction, the electrode was directly used for CO2 reduction. The Cu2O nanorods were prepared by annealing the Cu(OH)2 nanorods in the Ar atmosphere at 500 °C for 2 h with a heating rate of 10 °C min−1. The Cu(OH)2 nanorods were prepared by a previous method with some modifications32. The obtained Cu2O nanorods were airbrushed onto the commercial GDLs at an approximate loading of 0.9 mg/cm2, measured through weighing GDLs before and after airbrushing. The catalyst ink was prepared by dispersing 200 mg of Cu2O nanorods and 50 μL of Nafion Solution (Sigma-Aldrich) in 750 μL isopropyl alcohol and 250 μL of ultra-purity water (18.2 MΩ cm) and sonicated for 1 h before airbrushing (H-SET, Paasche). Pre-reduction was carried out with a potentiostat (CompactStat.e20250, IVIUM). Characterizations Field-emission scanning electron microscopy (FESEM) (Hitachi S-4800, 3 kV) was used to characterize the morphology and microstructure of the samples. Transmission electron microscopy (TEM), High-resolution TEM (HRTEM) images were obtained at 200 kV (JEOL JEM-2100F). The crystal structure was determined by X-ray Diffractometer (XRD, Bruker D8 Focus) with Cu Kα radiation (λ = 1.54056 Å) at 40 kV and 40 mA. XRD spectra were collected over a 2θ range of 30−60° at a scanning speed of 8°/min. XPS analyses of precatalysts were carried out on a Physical Electronics PHI 1600 ESCA system with an Al Kα X-ray source (1486.6 eV). The binding energy was calibrated using the C 1s photoelectron peak at 284.6 eV as the reference. In-situ Raman spectroscopy measurements In-situ Raman spectroscopy was carried out in a custom-designed flow cell (Supplementary Fig. 12a), which was manufactured by Gaossunion Co., Ltd., Tianjin. The electrode was encased in a PEEK fitting, with an exposed circular geometric surface area of ~1 cm2. A platinum wire and an Ag/AgCl electrode (saturated KCl, Gaossunion Co., Ltd., Tianjin) were used as the counter and the reference electrode, respectively. The counter electrode is separated from the working electrode by an anion exchange membrane (FAA-3-PK-75, Fumatech) to avoid cross-contamination. In situ Raman spectroscopy was performed with a Raman microscopy system (LabRAM HR Evolution, Horiba Jobin Yvon). A He−Ne laser (λ = 532 nm) served as the excitation source. All spectra were collected at a constant potential (−0.5 V versus the RHE). Electrochemical measurements were carried out with a potentiostat (CompactStat.e20250, IVIUM). In-situ ATR-SEIRAS measurements In-situ ATR-SEIRAS was performed with an ATR configuration. Au nanofilms were deposited directly on the reflecting plane of a Si prism using a modified electroless chemical deposition method outlined by Xu et al.58. The spectroelectrochemical cell was based on the design of Xu et al.59 and manufactured by Gaossunion Co., Ltd., Tianjin. In order to reduce the corrosion of Si crystal, 0.1 M KOH was used as the electrolyte. The counter electrode (a graphite rod) was separated from the working and reference electrodes, i.e., the catalyst film and a saturated Ag/AgCl electrode (saturated KCl, Gaossunion Co., Ltd., Tianjin), respectively, with a piece of anion exchange membrane (AEM, FAA-3-PK-75, Fumatech). This cell is integrated into the FTIR (is50, Nicolet) spectrometer with a modified accessory at a 60° incident angle (VeeMax III, PIKE Technology). All spectra were collected with a 4 cm−1 resolution. Spectra are presented in absorbance, with positive and negative peaks showing an increase and decrease in signal, respectively. As for LRS-Cu, MRS-Cu and HRS-Cu, they were deposited on the Au nanofilm coated-Si prisms like the process described above, while W-Cu was drop-casted onto the Au nanofilm coated-Si prisms. The background was taken at +0.1 V versus the RHE in Ar saturated electrolyte for each electrode. Electrochemical measurements are carried out with a potentiostat (CompactStat.e20250, IVIUM). OH− electroadsorption measurements In-situ OHads studies were conducted by flowing Ar in the flow cell (Supplementary Fig. 15). First, CO2 electrolysis was conducted at a constant potential of −0.5 V versus the RHE for 1 h by switching the gas feed to CO2 and flowing the electrolyte. Immediately after electrolysis, the electrolyte (1 M NaOH (a.q.)) flow rate was stopped to minimize the fluctuation in the voltammogram, and the gas feed was switched to Ar, the electrolyte flow rate was stopped, and then cyclic voltammetry (20 mV/s) was performed. Electrochemical measurements are carried out with a potentiostat (Autolab PGSTAT204, Metrohm). ECSA measurements The ECSA was determined by measuring the double-layer capacitance (CDL) of various electrodes in Ar-purged 2 M KOH (aq.) in the flow cell (Supplementary Fig. 15) and the ECSA was measured after CO2 electrolysis at a constant potential of −0.5 V versus the RHE for 1 h. Immediately after electrolysis, the gas feed was switched to Ar, and then the electrolyte flow rate was stopped to minimize the fluctuation in the voltammogram. The scan rate was varied from 25 to 125 mV s−1 in the non-faradaic potential region and the observed current was plotted as a function of scan rate to obtain the CDL. ECSA was determined by normalizing the CDL to that of a Cu foil. Electrochemical measurements are carried out with a potentiostat (CHI 660E, CH Instruments Inc.). Electrochemical reduction of CO2 in a flow cell CO2 reduction was conducted in a custom-designed three-chamber flow cell manufactured by Gaossunion Co., Ltd. (Supplementary Fig. 15), where the CO2 gas was supplied directly to the catalyst layer (cathode, working electrode). The CO2 gas flow rate was controlled using a mass flow controller (MC-Series, Alicat Scientific) and set to 10 sccm. However, it is well known that OH− can react with CO2 to form HCO3−or CO32−. Therefore, the calculation based on the inlet CO2 flow rate will result in overestimated F.E. results60,61. For this reason, we used another flowmeter (M-Series, Alicat Scientific) to detect the CO2 flow rate at the outlet of the reactor and used this number as the basis for calculating F.E. Aqueous KOH solution (2 M) was used as both the catholyte and the anolyte. Activated Ni foam was used as the anode (counter electrode). Peristaltic pumps (EC200-01, Gaossunion Co., Ltd.) were used to control the flow rate of the electrolytes at ~10 ml min−1. An AEM (FAA-3-PK-75, Fumatech) was used to separate the cathode and anode chambers. Electrolysis experiments were conducted using chronoamperometry with a potentiostat (CompactStat.e20250, IVIUM). The cathode potentials were measured against a Hg/HgO reference electrode (1 M KOH, Gaossunion Co., Ltd., Tianjin). For each measurement, products were quantified after the amount of electron flowing through the cathode achieved 50 C and at least three replicates were conducted to obtain an average value with the standard deviation. It should be noted that iR correction was not performed. Electrochemical reduction of CO2 in the MEA system The MEA cell (manufactured by Gaossunion Co., Ltd.) consists of a titanium anode (cathode) bipolar plate with serpentine flow field, associated nuts, bolts, and insulating kit. The geometric area of each flow field is 4 or 25 cm2 (Supplementary Fig. 40). An AEM membrane (FAA-3-PK-75, Fumatech) was activated in 0.1 M KOH for 24 h, washed with ultra-purity water prior to use. The anode consisted of iridium oxide supported on titanium mesh (IrOx/Ti mesh) was prepared by a dip-coating and thermal decomposition method62. The MEA was assembled in a way as illustrated in Fig. 5a. A direct current power supply (UTP1300, UNI-T Group Co., Ltd) was used to apply current to the MEA. A Corrtest CS350M in a galvanostatic mode was used to measure the cell voltage. No iR compensation was applied. Aqueous KHCO3 electrolyte (0.1 M) was used as the anolyte and was circulated using a peristaltic pump (EC200-01, Gaossunion Co., Ltd.). The electrolyte flow rate was kept at 10 mL min–1. As the current density and electrode area increase, the CO2 flow rate should be adjusted upwards to avoid mass transfer limitations of CO2 while maintaining optimum selectivity. The flow rate of the CO2 gas flowing into the cathode flow field was kept at 20 or 60 sccm by a mass flow controller (MC-Series, Alicat Scientific) for different geometries of the flow field. CO2 was flowed through a homemade humidifier (7/8 full of Milli-Q water, room temperature) prior to the MEA. The flow rate of the CO2 gas flowing out the cathode flow field was also measured by a flowmeter (M-Series, Alicat Scientific). The liquid products carried by CO2 gas are absorbed by low-temperature ultra-purity water obtained from an ice salt bath Analysis of CO2 reduction products During electrolysis, gas products were quantified using an on-line gas chromatography system (GC7890B, Agilent Technologies, Inc.). The thermal conductivity detector (TCD) connected to a MolSieve 5A packed column (Agilent Technologies, Inc.) to detect H2, O2, and N2 and a back flame ionization detector (FID) connected to a Porapak Q packed column (Agilent Technologies, Inc.) to detect CO. A methanizer was installed to enable the back FID to detect CO with 1000 times higher sensitivity. A front FID connected to an HP-PLOT Al2O3 capillary column (Agilent Technologies, Inc.) to detect hydrocarbons (C1~C3). Ar was used as the carrier gas. After passing through the reactor, the gas was allowed to flow directly into the gas sampling loop of the gas chromatography for online gaseous product analysis. In the performance test using flow cell and the MEA system, the liquid products were collected from the cathode and anode chambers60. The liquid products were analyzed by headspace gas chromatography (HS-GC) and 1H-NMR. HS-GC measurements were carried out using a BCHP HS-2 Headspace Sampler with GC2060 gas chromatography (Shanghai Ruimin Instrument Co., Ltd.). Typically, 10 mL vials were filled with 3 mL of the liquid sample and sealed. They were heated to 70 °C over 20 min in the headspace sampler and 1 mL of the headspace gas composition was automatically injected into the GC. The sample loop (110 °C) and transfer line (110 °C) were both heated to avoid condensation. Ar was used as the carrier gas. An HP-INNOWax capillary column (Length: 60 m; ID: 0.32 mm; Film: 0.5 μm, Agilent) was used to separate the compounds in the sample. 1H-NMR was performed using AVANCE IIITM HD 400 MHz NanoBAY. The water suppression method was used. DMSO (10 mM) and phenol (50 mM) were added as internal standards. For CO2 reduction products showing multiple sets of NMR peaks, the set of peaks with the highest intensity were chosen for calibration and quantification. Construction of the photovoltaic-electrolyzer (PV-EC) system The simulated solar illumination was obtained from a 300 W Xenon arc lamp (Microsolar 300 UV, Beijing Perfectlight Technology Co. Ltd.) equipped with an air mass 1.5 global (AM 1.5G) filter, and the power intensity of the incident light was calibrated to 100 mW/cm2 using a Si photodiode (FDS100, Thorlabs). The solar panel was based on five p−n+ Si solar cells connected in series (effective illuminated area of ~5.7 cm2). A Source Measure Unit (2450, Keithely) was wired in series with 0 V applied to monitor the current. The electrolyzer is the flow cell. Computational methods Vienna ab initio simulation package (VASP) was used to carry out calculations with the PBE exchange-correlation functional63,64. Van der Waals interactions were accounted for by using the DFT-D3 method65. The cut-off energy is 400 eV. The interactions between the atomic cores and electrons were described by the projector augmented wave (PAW) method66. All structures were optimized until the force on each atom has been less than 0.02 eV/Å. The transition state search was conducted with the climbing image nudged elastic band (CI-NEB) method, followed by the dimer method to converge the saddle point within 0.05 eV/Å. We access CO dimerization on the three models. A four-layer Cu(111)-(3×3) slab with a (3×3×1) k-point grid and a four-layer Cu(100)-(3×3) slab with a (3×3×1) k-point grid were used as models for DFT calculations. The bottom two layers are fixed while the upper two layers were relaxed during optimization. One layer of water with a simple hydronium ion was chosen to simulate the electrochemical interface67–69. Supplementary information Supplementary Information Supplementary information The online version contains supplementary material available at 10.1038/s41467-021-26053-w. Acknowledgements We acknowledge the National Key R&D Program of China (2016YFB0600901), the National Natural Science Foundation of China (22038009, 21722608, 51861125104, 2212100031), the Natural Science Foundation of Tianjin City (18JCJQJC47500), and the Program of Introducing Talents of Discipline to Universities (B06006) for financial support. Author contributions J.L.G. supervised the project. J.L.G., T.W., G.Z., and Z.J.Z. conceptualized the project. G.Z. H.M.L., and H.G synthesized Cu-based catalysts. G.Z., H.Y.W., J. Y., and J.Y.G. conducted the catalytic tests and the related data processing. Q.Z.W. fabricated the solar panel. Z.J.Z. and D.F.C. carried out the theoretical calculations. All the authors participate in the writing of the paper. Data availability The authors declare that all data supporting the results of this study are available within the paper and its supplementary information files or from the corresponding authors upon reasonable request. Competing interests The authors declare no competing interests. Peer review information Nature Communications thanks Wei An, Antonio José Martín, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. These authors contributed equally: Gong Zhang, Zhi-Jian Zhao. Change history 7/13/2023 In this article several affiliation details were incomplete: the affiliation ‘School of Chemical Engineering and Technology, Tianjin University, Tianjin, China’ should have been ‘School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, China’; ‘Key Laboratory for Green Chemical Technology of Ministry of Education, Tianjin University, Tianjin, China’ should have been ‘Key Laboratory for Green Chemical Technology of Ministry of Education, Tianjin University, Tianjin 300072, China’; ‘Collaborative Innovation Center of Chemical Science and Engineering (Tianjin), Tianjin, China’ should have been ‘Collaborative Innovation Center of Chemical Science and Engineering (Tianjin), Tianjin 300072, China’; ‘Joint School of National University of Singapore and Tianjin University, International Campus of Tianjin University, Binhai New City, Fuzhou, China’ should have been ‘Joint School of National University of Singapore and Tianjin University, International Campus of Tianjin University, Binhai New City, Fuzhou 350207, China’. The original article has been corrected. ==== Refs References 1. Kattel S Ramirez PJ Chen JG Rodriguez JA Liu P Active sites for CO2 hydrogenation to methanol on Cu/ZnO catalysts Science 2017 355 1296 1299 10.1126/science.aal3573 28336665 2. Dinh CT CO2 electroreduction to ethylene via hydroxide-mediated copper catalysis at an abrupt interface Science 2018 360 783 787 10.1126/science.aas9100 29773749 3. Nitopi S Progress and perspectives of electrochemical CO2 reduction on copper in aqueous electrolyte Chem. Rev. 2019 119 7610 7672 10.1021/acs.chemrev.8b00705 31117420 4. Li F Molecular tuning of CO2-to-ethylene conversion Nature 2020 577 509 513 10.1038/s41586-019-1782-2 31747679 5. Yang PP Protecting copper oxidation state via intermediate confinement for selective CO2 electroreduction to C2+ fuels J. Am. Chem. Soc. 2020 142 6400 6408 10.1021/jacs.0c01699 32176485 6. Zhong M Accelerated discovery of CO2 electrocatalysts using active machine learning Nature 2020 581 178 183 10.1038/s41586-020-2242-8 32405017 7. Ma WC Electrocatalytic reduction of CO2 to ethylene and ethanol through hydrogen-assisted C−C coupling over fluorine-modified copper Nat. Catal. 2020 3 478 487 10.1038/s41929-020-0450-0 8. Jung H Electrochemical fragmentation of Cu2O nanoparticles enhancing selective C−C coupling from CO2 reduction reaction J. Am. Chem. Soc. 2019 141 4624 4633 10.1021/jacs.8b11237 30702874 9. Durand WJ Peterson AA Studt F Abild-Pedersen F Norskov JK Structure effects on the energetics of the electrochemical reduction of CO2 by copper surfaces Surf. Sci. 2011 605 1354 1359 10.1016/j.susc.2011.04.028 10. Jiang K Metal ion cycling of Cu foil for selective C−C coupling in electrochemical CO2 reduction Nat. Catal. 2018 1 111 119 10.1038/s41929-017-0009-x 11. Xia Y Xiong Y Lim B Skrabalak SE Shape-controlled synthesis of metal nanocrystals: simple chemistry meets complex physics? Angew. Chem. Int. Ed. 2009 48 60 103 10.1002/anie.200802248 12. Wang Z Yang G Zhang Z Jin M Yin Y Selectivity on etching: creation of high-energy facets on copper nanocrystals for CO2 electrochemical reduction ACS Nano 2016 10 4559 4564 10.1021/acsnano.6b00602 26974506 13. De Gregorio GL Facet-dependent selectivity of Cu catalysts in electrochemical CO2 reduction at commercially viable current densities ACS Catal. 2020 10 4854 4862 10.1021/acscatal.0c00297 32391186 14. Wang HX Self-selective catalyst synthesis for CO2 reduction Joule 2019 3 1927 1936 10.1016/j.joule.2019.05.023 15. Wang Y Catalyst synthesis under CO2 electroreduction favours faceting and promotes renewable fuels electrosynthesis Nat. Catal. 2019 3 98 106 10.1038/s41929-019-0397-1 16. Tao AR Habas S Yang PD Shape control of colloidal metal nanocrystals Small 2008 4 310 325 10.1002/smll.200701295 17. Niu Z Li Y Removal and utilization of capping agents in nanocatalysis Chem. Mater. 2013 26 72 83 10.1021/cm4022479 18. Hou YH Photonic curing: activation and stabilization of metal membrane catalysts (MMCs) for the electrochemical reduction of CO2 ACS Catal. 2019 9 9518 9529 10.1021/acscatal.9b03664 19. Nwabara UO Cofell ER Verma S Negro E Kenis PJA Durable cathodes and electrolyzers for the efficient aqueous electrochemical reduction of CO2 ChemSusChem 2020 13 855 875 10.1002/cssc.201902933 31863564 20. Yang H., et al. Recent progress in self-supported catalysts for CO2 electrochemical reduction. Small Methods 4, 1900826 (2020). 21. Ott S Ionomer distribution control in porous carbon-supported catalyst layers for high-power and low Pt-loaded proton exchange membrane fuel cells Nat. Mater. 2020 19 77 85 10.1038/s41563-019-0487-0 31570820 22. Zhou Y-G Kang Y Huang J Fluidized electrocatalysis CCS Chem. 2020 2 31 41 10.31635/ccschem.020.201900065 23. Wei X Li Y Chen L Shi J Formic acid electro-synthesis by concurrent cathodic CO2 reduction and anodic CH3OH oxidation Angew. Chem. Int. Ed. 2021 60 3148 3155 10.1002/anie.202012066 24. Dobrev D Ion-beam-induced texture formation in vacuum-condensed thin metal films Thin Solid Films 1982 92 41 53 10.1016/0040-6090(82)90186-9 25. Abadias G Stress and preferred orientation in nNitride-based PVD coatings Surf. Coat. Technol. 2008 202 2223 2235 10.1016/j.surfcoat.2007.08.029 26. Lv YH Influence of substrate bias voltage on structure and properties of the CrAlN films deposited by unbalanced magnetron sputtering Appl. Surf. Sci. 2012 258 3864 3870 10.1016/j.apsusc.2011.12.048 27. Zheng JY Bao SH Guo Y Jin P Anatase TiO2 films with dominant {001} facets fabricated by direct-current reactive magnetron sputtering at room temperature: oxygen defects and enhanced visible-light photocatalytic behaviors ACS Appl. Mater. Interfaces 2014 6 5940 5946 10.1021/am500979j 24720367 28. Li F Cooperative CO2-to-ethanol conversion via enriched intermediates at molecule–metal catalyst interfaces Nat. Catal. 2019 3 75 82 10.1038/s41929-019-0383-7 29. Cao L Mechanistic insights for low-overpotential electroreduction of CO2 to CO on copper nanowires ACS Catal. 2017 7 8578 8587 10.1021/acscatal.7b03107 30. Droog JMM Schlenter B Oxygen electrosorption on copper single crystal electrodes in sodium hydroxide solution J. Electroanal. Chem. Interfacial Electrochem. 1980 112 387 390 10.1016/S0022-0728(80)80421-9 31. Schouten KJ Qin Z Perez Gallent E Koper MT Two pathways for the formation of ethylene in CO reduction on single-crystal copper electrodes J. Am. Chem. Soc. 2012 134 9864 9867 10.1021/ja302668n 22670713 32. Lv JJ A highly porous copper electrocatalyst for carbon dioxide reduction Adv. Mater. 2018 30 e1803111 10.1002/adma.201803111 30368917 33. LaGrow AP Ward MR Lloyd DC Gai PL Boyes ED Visualizing the Cu/Cu2O interface transition in nanoparticles with environmental scanning transmission electron microscopy J. Am. Chem. Soc. 2017 139 179 185 10.1021/jacs.6b08842 27936677 34. Feng X Jiang K Fan S Kanan MW Grain-boundary-dependent CO2 electroreduction activity J. Am. Chem. Soc. 2015 137 4606 4609 10.1021/ja5130513 25835085 35. Jin M Shape-controlled synthesis of copper nanocrystals in an aqueous solution with glucose as a reducing agent and hexadecylamine as a capping agent Angew. Chem. Int. Ed. 2011 50 10560 10564 10.1002/anie.201105539 36. Wang X Varela AS Bergmann A Kuhl S Strasser P Catalyst particle density controls hydrocarbon product selectivity in CO2 electroreduction on CuOx ChemSusChem 2017 10 4642 4649 10.1002/cssc.201701179 28776946 37. Kas RK Electrochemical CO2 reduction on nanostructured metal electrodes: fact or defect? Chem. Sci. 2020 11 1738 1749 10.1039/C9SC05375A 34123269 38. Reske R Controlling catalytic selectivities during CO2 electroreduction on thin Cu metal overlayers J. Phy. Chem. Lett. 2013 4 2410 2413 10.1021/jz401087q 39. Tan YC Lee KB Song H Oh J Modulating local CO2 concentration as a general strategy for enhancing C−C coupling in CO2 electroreduction Joule 2020 4 1104 1120 10.1016/j.joule.2020.03.013 40. Lee SH Oxidation state and surface reconstruction of Cu under CO2 reduction conditions from in situ x-ray characterization J. Am. Chem. Soc. 2021 143 588 592 10.1021/jacs.0c10017 33382947 41. Jouny M Luc W Jiao F High-rate electroreduction of carbon monoxide to multi-carbon products Nat. Catal. 2018 1 748 755 10.1038/s41929-018-0133-2 42. Heyes J Dunwell M Xu B CO2 reduction on Cu at low overpotentials with surface-enhanced in situ spectroscopy J. Phy. Chem. C. 2016 120 17334 17341 10.1021/acs.jpcc.6b03065 43. Sartin MM Effect of particle shape and electrolyte cation on CO adsorption to copper oxide nanoparticle electrocatalysts J. Phys. Chem. C. 2018 122 26489 26498 10.1021/acs.jpcc.8b08541 44. Gunathunge CM Ovalle VJ Li YW Janik MJ Waegele MM Existence of an electrochemically inert CO population on Cu electrodes in alkaline pH ACS Catal. 2018 8 7507 7516 10.1021/acscatal.8b01552 45. Ren D Fong J Yeo BS The effects of currents and potentials on the selectivities of copper toward carbon dioxide electroreduction Nat. Commun. 2018 9 925 10.1038/s41467-018-03286-w 29500358 46. Ren D Gao J Zakeeruddin SM Grätzel M New insights into the interface of electrochemical flow cells for carbon dioxide reduction to ethylene J. Phys. Chem. Lett. 2021 12 7583 7589 10.1021/acs.jpclett.1c02043 34347495 47. Goodpaster JD Bell AT Head-Gordon M Identification of possible pathways for C−C bond formation during electrochemical reduction of CO2: new theoretical insights from an improved electrochemical model J. Phys. Chem. Lett. 2016 7 1471 1477 10.1021/acs.jpclett.6b00358 27045040 48. Schreier M Yoon Y Jackson MN Surendranath Y Competition between H and CO for active sites governs copper-mediated electrosynthesis of hydrocarbon fuels Angew. Chem. Int. Ed. 2018 57 10221 10225 10.1002/anie.201806051 49. Clark EL Hahn C Jaramillo TF Bell AT Electrochemical CO2 reduction over compressively strained CuAg surface alloys with enhanced multi-carbon oxygenate selectivity J. Am. Chem. Soc. 2017 139 15848 15857 10.1021/jacs.7b08607 28988474 50. Li J Constraining CO coverage on copper promotes high-efficiency ethylene electroproduction Nat. Catal. 2019 2 1124 1131 10.1038/s41929-019-0380-x 51. Malkani AS Dunwell M Xu B Operando spectroscopic investigations of copper and oxide-derived copper catalysts for electrochemical CO reduction ACS Catal. 2018 9 474 478 10.1021/acscatal.8b04269 52. Sandberg RB Montoya JH Chan K Norskov JK CO−CO coupling on Cu facets: coverage, strain, and field effects Surf. Sci. 2016 654 56 62 10.1016/j.susc.2016.08.006 53. Resasco J Promoter effects of alkali metal cations on the electrochemical reduction of carbon dioxide J. Am. Chem. Soc. 2017 139 11277 11287 10.1021/jacs.7b06765 28738673 54. Ren D Loo NWX Gong L Yeo BS Continuous production of ethylene from carbon dioxide and water using intermittent sunlight. ACS Sustain. Chem. Eng. 2017 5 9191 9199 10.1021/acssuschemeng.7b02110 55. Gao J Selective C−C coupling in carbon dioxide electroreduction via efficient spillover of intermediates as supported by operando Raman pectroscopy J. Am. Chem. Soc. 2019 141 18704 18714 10.1021/jacs.9b07415 31656072 56. Weng LC Bell AT Weber AZ Modeling gas-diffusion electrodes for CO2 reduction Phys. Chem. Chem. Phys. 2018 20 16973 16984 10.1039/C8CP01319E 29900441 57. Burdyny T Smith WA CO2 reduction on gas-diffusion electrodes and why catalytic performance must be assessed at commercially-relevant conditions Energy Environ. Sci. 2019 12 1442 1453 10.1039/C8EE03134G 58. Dunwell M The central role of bicarbonate in the electrochemical reduction of carbon dioxide on gold J. Am. Chem. Soc. 2017 139 3774 3783 10.1021/jacs.6b13287 28211683 59. Li J Electrokinetic and in situ spectroscopic investigations of CO electrochemical reduction on copper Nat. Commun. 2021 12 3264 10.1038/s41467-021-23582-2 34075039 60. Zhang J Luo W Zuttel A Crossover of liquid products from electrochemical CO2 reduction through gas diffusion electrode and anion exchange membrane J. Catal. 2020 385 140 145 10.1016/j.jcat.2020.03.013 61. Ma M Insights into the carbon balance for CO2 electroreduction on Cu using gas diffusion electrode reactor designs Energy Environ. Sci. 2020 13 977 985 10.1039/D0EE00047G 62. Luc W Rosen J Jiao F An Ir-based anode for a practical CO2 electrolyzer Catal. Today 2017 288 79 84 10.1016/j.cattod.2016.06.011 63. Kresse G Hafner J Ab initio molecular-dynamics simulation of the liquid-metal-amorphous-semiconductor transition in germanium Phys. Rev. B Condens. Matter 1994 49 14251 14269 10.1103/PhysRevB.49.14251 10010505 64. Grimme S Antony J Ehrlich S Krieg H A consistent and accurate ab initio parametrization of density functional dispersion correction (DFT-D) for the 94 elements H-Pu J. Chem. Phys. 2010 132 154104 10.1063/1.3382344 20423165 65. Liu S Zong J Zhao Z-J Gong J Exploring the initial oxidation of Pt, Pt3Ni, Pt3Au (111) surfaces: a genetic algorithm based global optimization with density functional theory Green. Chem. Eng. 2020 1 56 62 10.1016/j.gce.2020.09.006 66. Kresse G Joubert D From ultrasoft pseudopotentials to the projector augmented-wave method Phys. Rev. B 1999 59 1758 1775 10.1103/PhysRevB.59.1758 67. Gauthier JA Dickens CF Chen LD Doyle AD Norskov JK Solvation effects for oxygen evolution reaction catalysis on IrO2(110) J. Phys. Chem. C. 2017 121 11455 11463 10.1021/acs.jpcc.7b02383 68. Montoya JH Shi C Chan K Norskov JK Theoretical insights into a CO dimerization mechanism in CO2 electroreduction J. Phys. Chem. Lett. 2015 6 2032 2037 10.1021/acs.jpclett.5b00722 26266498 69. Cheng D The nature of active sites for carbon dioxide electroreduction over oxide-derived copper catalysts Nat. Commun. 2021 12 395 10.1038/s41467-020-20615-0 33452258
PMC008xxxxxx/PMC8729023.txt
==== Front F1000Res F1000Res F1000Research 2046-1402 F1000 Research Limited London, UK 35035896 10.12688/f1000research.73657.4 Research Article Articles Prevalence of anisometropia in children and adolescents [version 4; peer review: 2 approved] Nunes Amélia F Conceptualization Data Curation Formal Analysis Investigation Methodology Project Administration Resources Software Supervision Validation Visualization Writing – Original Draft Preparation Writing – Review & Editing https://orcid.org/0000-0002-5001-3534 a1234 Batista Maria Data Curation Formal Analysis Investigation Resources Writing – Original Draft Preparation 134 Monteiro Pedro Conceptualization Methodology Project Administration Supervision Validation Visualization Writing – Review & Editing 1234 1 Universidade da Beira Interior, Portugal, Covilhã, Portugal 2 Health Sciences Research Centre (CICS-UBI), Universidade da Beira Interior, Covilhã, Portugal 3 UBIMedical, Universidade da Beira Interior, Covilhã, Portugal 4 Clinical and Experimental Center in Vision Sciences (CCECV), Universidade da Beira Interior, Covilhã, Portugal a [email protected] No competing interests were disclosed. 8 6 2022 2021 10 11016 6 2022 Copyright: © 2022 Nunes AF et al. 2022 https://creativecommons.org/licenses/by/4.0/ This is an open access article distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Background: This research was developed to study the epidemiology of anisometropia. It aims to estimate the prevalence of anisometropia in Portuguese children and adolescents at various educational stages, studying its association with sociodemographic variables. Methods: Observational cross sectional study envolving 749 children and adolescents (from 3 to 16 years old) from the central region of Portugal. The refraction was performed with a paediatric, open field auto refractometer (PlusOptix), without cycloplegia and under binocular conditions, to determine the rate of anisometropia and its association with gender, study cycle and area of residence. Results: The prevalence of anisometropia in the studied sample was 6.1%, varying from 2.9% in pre-school education to 9.4% in the 3rd study cycle. Myopic anisometropia was the most prevalent and hyperopic and astigmatic anisometropia showed identical proportions of occurrence. No statistical differences were found between genders or between areas of residence regarding the rate of anisometropia. Regarding spherical equivalent anisometropia, there was a pattern of variation that increased with the cycle of studies (p = 0.012), with myopic anisometropia being the main contributor to this variation. Conclusions: This study found an increase in anisometropia with the educational stage. The high rate of anisometropia found in adolescents (9.4%) as well as the progressive increase in this rate throughout school progress (from 2.9% to 9.4%) suggests the need to extend the detection strategies of this condition beyond childhood. Pediatrics Child Teenager Refraction School vision The author(s) declared that no grants were involved in supporting this work.Revised Amendments from Version 3 The new version includes minor changes: - Correction of a spelling error and removal of a sentence in the methods section. - Removal of a statistical data in the abstract and in the results section ==== Body pmcIntroduction Anisometropia is an ocular disorder characterized by an interocular difference (IOD) in refractive error. It represents a specific refractive condition where the two eyes of an individual, can have asymmetric eye growth. 1 This condition can occur in situations of myopic, hyperopic or astigmatic asymmetry and is strongly associated with the development of other eye changes such as aniseikonia, amblyopia, diplopia and strabismus. 2 , 3 Although there is no uniformly defined dioptric value for its clinical classification, an IOD in the spherical equivalent (SE) of 1 diopter or more is accepted as the threshold, for most authors. 1 , 4 – 8 However, even using this limit, the scientific literature presents a significant variation in the prevalence values of anisometropia in terms of age, gender and ethnicity. 4 , 5 , 8 , 9 Factors associated with lifestyle and educational level have also been referred to as risk factors for anisometropia. 7 , 8 , 10 Early detection and early treatment is crucial to prevent permanent visual loss. Although it is not clear what is the ideal age to perform the correction, in order to guarantee an ideal visual development and maturation, the early correction of anisometropia is important. 11 This prevents the development of other changes such as aniseikonia, amblyopia and strabismus, 3 , 12 , 13 and even in small degrees (<1D) facilitates emmetropization. 11 It also improves quality of life, reducing or eliminating symptoms of visual discomfort. In this way, visual screening at a young age is useful in identifying who is most likely to benefit from early optical correction or preventive treatment. 11 , 14 , 15 The clinical methods used to characterize anisometropia are refractive techniques, with autorefraction, using Plusoptix, one of the most recommended techniques for screening activities. 16 , 17 This instrument allows quantifying the refractive error in which it is possible to obtain a very similar value to that obtained by cycloplegic refraction 17 – 19 and with excellent precision in anisometropia signalling. 19 Studies on the prevalence of anisometropia focus on children or adults, with less research being found in adolescence. The aim is to estimate the prevalence of anisometropia (spherical and astigmatic) and to analyse its pattern of variation in a sample of children and adolescents, from preschool education (from three to six years old) to the various cycles of basic education (from the 1 st to the 9 th school year in Portugal, from six to fifteen years old). Methods This is an observational cross sectional study to evaluate the prevalence of anisometropia in children and adolescents. Data from the Portuguese Census 2011, showed a population of 319284 from 0 to 14 years old in the central region. 20 For a confidence level of 95% and a 5% margin of error, taking into account that the prevalence of the studied condition is unknown, it was fixed at 50% to obtain a large enough sample size. The result was a minimum of 384 subjects. The data collection took place in five schools of the central region of Portugal, including all students that were authorized to participate by their legal guardians, between October 2018 and February 2019. Participants were 749 children and adolescents aged between three and sixteen years. Students data for which it was not possible to obtain refraction were excluded, due to technical issues associated with the performance of the instrument (presence of strabismus, opacities, retinal anomalies or when the refractive error exceeded the instrument's measurement limit - spherical measurement range or cylindrical from −7.00 to +5.00D) or due to lack of cooperation from the participant. Only two participants were excluded. The refractive error was measured with the paediatric auto refractometer model A09 by PlusOptix, Nuremberg, Germany, by the average of three consecutive measurements, binocular and without the use of a cycloplegic. The PlusOptix allows measuring the refractive error in open field simultaneously in both eyes and under the same conditions, in a fast, easy, safe, non-invasive way, at a distance of one meter from the subject's eyes. Sample characterization In Portugal the compulsory education includes basic education, which is divided into 3 cycles, followed by the secondary education. The 1 st cycle of basic education includes the 1 st to 4 th school year (ages 6 to 10), the 2 nd cycle includes the 5 th and 6 th year (ages 10 to 12) and the 3 rd cycle includes the 7 th to the 9 th year (ages 12 to 15). The sample under study had students from preschool to the 3 rd cycle of basic education and was characterized according to gender, area of residence and cycle of studies. Rural and urban areas classification was based on the information provided by the municipalities, considering the residence area of each participant. Regarding the use of glasses, 177 participants wore glasses or contact lenses and 572 did not use any type of optical correction. Table 1 summarizes the characteristics of the sample. Table 1. Sample characteristics. Factor Sample dimension n (%) Wearing glasses Yes (n) No (n) missing Gender Male 399 (53.3%) 80 318 1 Female 350 (46.7%) 95 254 1 Residence area Rural 320 (42.7%) 69 249 2 Urban 423 (56.5%) 104 319 0 Missing 6 (0.8%) - - - Study stage Preschool 103 (13.8%) 3 100 0 1 st cycle 231 (30.8%) 36 195 0 2 nd cycle 181 (24.2%) 44 136 1 3 rd cycle 234 (31.2%) 92 141 1 Data analysis In order to calculate the average of the three refractive measurements, the power was converted from its spherical-cylindrical form to its power vector representation, described by Thibos, 21 using the following expressions: SE=S+C2 (1) J0=−C2cos2α (2) J45=−C2sin2α (3) Where SE represents the spherical equivalent; J 0 represents Jackson's crossed cylinders on the 90° or 180° axis, which stands for the amount of direct or indirect astigmatism; J 45 represents Jackson's crossed cylinders on the 45° or 135° axis, which stands for the amount of oblique astigmatism; S, C and α represent the spherical, cylindrical (in negative form) and cylinder axis component, respectively, of the auto refractometer measurement. Refractive state classification The participants were classified in emmetropes, myopes, hyperopes, astigmats or anisometropes, according to the average value of the auto refractometer. In order to carry out this classification, the criteria recommended for the auto refractometer used were applied ( Table 2). Table 2. Classification criteria for refractive state. 16 Refractive state < 6 years old ≥6 years old Emmetrope −1.00D < SE < +1.25D −1.00D < SE < +1.00D Myope SE ≤ −1.00D Hyperope SE ≥ +1.25D SE ≥ +1.00D Astigmat |C|≥1.00D |C|≥1.25D Anisometrope |IOD|≥1.25 (SE or C) Anisometropia classification The absolute value of IOD of the refractive error in terms of SE was designated as spherical anisometropia (SA), the absolute IOD in astigmatism was designated as meridional anisometropia (MA), SA=SERE−SELE (4) MA=CRE−CLE (5) where RE refers to the right eye and LE to the left eye. When there was no IOD in the spherical equivalent and the absolute IOD was only in the astigmatic component, it was designated by simple meridional anisometropia (sMA). The anisometropia classification was done according to the cutoff points referred to in Table 2. The presence of at least one of the previous conditions was designated as total anisometropia (TA). Low anisometropia was considered for IOD values below 2.00D, high anisometropia for values between 2.00D and 6.00D and very high anisometropia for IOD values above 6.00D. According to the type of refractive error, anisometropia was classified as myopic, when both eyes were myopic or when one eye was myopic and the other was emmetropic; hyperopic, when both eyes were hyperopic or when one eye was hyperopic and the other emmetropic; antimetropic, when one eye was myopic and the other hyperopic; simple meridional anisometropia when there was no SA, but there was MA. Statistical analysis A descriptive statistical analysis was carried out, using SPSS version 26 package, characterizing the sample in the variables of interest, sociodemographic and refractive, presenting means and standard deviations, frequencies and percentages both in the whole of the sample and also according to several stratifications to which it was subjected. In all the sociodemographic factors in which the sample was categorized, groups with a large size (n > 30) were obtained. The proportion of subjects with anisometropia was analysed according to gender, area of residence and school cycle, and through the Chi-square test, it was evaluated whether these variables were associated with the occurrence of anisometropia in the studied population. All the results of the statistical inference tests were interpreted to a 95% confidence level, that is, the significance level of 0.05 was used. Ethics approval This study was conducted in accordance with the principles of the Declaration of HELSINKI and written informed consent as obtained from parents of each participant in the study. It was approved by the Ethics Committee from Universidade da Beira Interior (CE-UBI-Pj-2019-043). Consent to participate Written informed consent was obtained from parents or legal guardians of all participants. Results Refractive state According to the classification criteria previously defined for the classification of refractive state, it was concluded that in the study sample 71.16% (n = 533) were emmetropic. Among subjects with significant refractive error (n = 216), it was found that hyperopia was the most prevalent refractive error (n = 109, corresponding to 14.6% in the studied population) followed by myopia (n = 49, corresponding to 6.5% in the population), anisometropia (n = 46, corresponding to 6.1% in the population) and astigmatism (n = 25 where 12 were cases of simple astigmatism and 13 compound astigmatism). Figure 1 shows graphically the distribution of the different refractive states in the study sample. The representation of astigmatism, refers only to the occurrence of simple astigmatism, without a significant SE. The mean values of refractive errors, according to the previous classification, are shown in Table 3. Figure 1. Refractive error distribution. Table 3. Mean values and standard deviation of refractive errors of subjects with significant ametropia (right eye data). Refractive group Wearing glasses (n) Spherical equivalent (D) Cylindrical component (D) Yes No Mean ± SD Range [min; max] Mean ± SD Range [min; max] Myopia 44 5 −2.67 ± 1.32 −5.79; −0.92 −0.39 ± 0.33 −1.29; 0.00 Hyperopia § 22 85 +1.46 ± 0.59 +1.00; +5.25 −0.41 ± 0.47 −2.50; 0.00 Simple astigmatism 6 6 +0.29 ± 0.55 −0.83; +1.21 −1.50 ± 0.33 −2.07; −1.14 Anisometropia ϯ 37 8 Interocular difference 1.53 ± 0.82 0.00; 3.79 0.78 ± 0.85 0.00; 3.9 § Two cases without data regarding the use of glasses. ϯ One case without data regarding the use of glasses. Anisometropia Anisometropia was identified in 46 participants (6.1% of the studied population). Regarding the use of glasses, it was observed that eight cases did not use any type of optical correction. According to the magnitude of the refractive error, no child was found with very high anisometropia, 15 were registered with high anisometropia and 31 with low anisometropia, that is, of the anisometrope subjects, most (67.4%) had low anisometropia. The dispersion of interocular diference in SA and MA is shown in Table 3. In the classification of anisometropia according to the type of refractive error, 37 subjects (corresponding to 4.9% in the studied population) were found with SA, integrating 21 with myopia, 15 with hyperopia and 1 with antimetropia; 15 subjects (corresponding to 2% in the population) with MA, and only 9 of these did not have SA, and were classified as simple meridional anisometropia (sMA). This distribution is represented graphically in Figure 1. It is possible to observe that myopic anisometropia was the most prevalent (46%), followed by hyperopic anisometropia (33%) and sMA (19%). Antimetropic anisometropia was the least prevalent (2%). Influence of sociodemographic variables The study of the influence of sociodemographic variables in the occurrence of anisometropia is presented in Table 4. The Chi-square test indicates that there was no association between MA and any of the factors under analysis. Table 4. Relationship of gender, area of residence and study cycle, in anisometropia. Gender Chi-square test Female (350) Male (399) N % N % P TA 25 7.1 21 5.3 0.29 SA 21 6 16 4 0.21 MA 8 2.3 7 1.8 0.60 Area of residence Chi-square test Rural (320) Urban (423) N % N % p TA 22 6.8 23 5.4 0.42 MA 7 2.2 8 1.9 0.78 SA 18 5.6 18 4.3 0.39 School stage Chi-square test Preschool (103) 1 st cycle (231) 2 nd cycle (181) 3 rd cycle (234) N % N % N % N % p TA 3 2.9 10 4.3 11 6.1 22 9.4 0.06 MA 2 1.9 3 1.3 6 3.3 4 1.7 0.52 SA 1 1.0 9 3.9 7 3.9 20 8.6 0.012 * p (adjust) 0.36 <0.001 * <0.001 * 0.015 * TA – Total anisometropes (spherical and meridional); SA - spherical equivalent anisometropia; MA – meridional anisometropia. The highest rates are shown in bold. Multiple comparisons with Chi-squared test for study cycles (Z: adjusted standardized residual to Chi-Square and p adjusted to Bonferroni correction). * Significant at the 0.05 level. No significantly different occurrence of anisometropia was found according to gender or area of residence (p > 0.05). In relation to the school, there was a pattern of variation that increased with the cycle of studies, ranging from 2.9% in preschool education to 9.4% in the 3 rd cycle of studies, however this association was only statistically significant for SA (p = 0.012), where there was a rate of 1% in preschool education and 8.6% in the 3 rd cycle. It should be noted that the study cycle is dependent on age. Figure 2 illustrates the distribution of anisometropia according to the type of refractive error, for each school cycle. It is observed that myopic anisometropia was present in all study cycles, registering a considerable increase from the beginning of school, that is, from the 1 st cycle to the 3 rd cycle. On the other hand, hyperopic anisometropia was manifested on a larger scale in children of the 1 st cycle, with a lower occurrence in the following cycles. As for the simple MA, it was present in all study cycles and there was no specific pattern in its variation with the study cycle progress. Figure 2. Relative frequency of the variation of the various types of anisometropia, with the school cycle. Discussion An anisometropia rate of 6.1% was found, considering the spherical and meridional anisometropia. It was found that this rate varies with the cycle of studies, showing an increase as the level of education advances, ranging from 2.9% in preschool education to 9.4% in the 3 rd cycle of basic education. No statistically significant differences were found in the distribution of anisometropia either between genders or between areas of residence. It was also found that myopic anisometropia was the most prevalent (46%), with a considerable increase in the 3 rd cycle of studies. Table 5 summarises worldwide epidemiological data regarding studies in children and adolescents, which used the same criterion for anisometropia definition (IOD ≥ 1D). Table 5. Anisometropia prevalence studies. TA - Total Anisometropia; SA - Spherical Equivalent Anisometropia; MA - Meridional Anisometropia. Author/year Location Age (years) Sample Refractive test Definition Prevalence (%) Ohlsson/2003 22 America Monterrey, Mexico 12-13 1035 Retinoscopy AT ≥ 1.00 15 Dobson/2008 13 Toronto, Canada 4-13 1041 Cycloplegic Autorefraction SA ≥ 1.00 MA ≥ 1.00 TA ≥ 1.00 6.7 15 18.1 Deng/2012 5 Boston, USA 5 12-15 395 312 Retinoscopy AS ≥ 1.00 1.3 5.8 Quek/2004 23 Asia Singapore 15-19 946 N/Cycloplegic Autorefraction SA ≥ 1.00 11.2 Yekta/2010 24 Shiraz, Iran 7-15 1872 N/Cycloplegic Autorefraction SA ≥ 1.00 2.6 Hu/2016 7 Shandong, China 4-18 6025 Cycloplegic Autorefraction SA ≥ 1.00 MA ≥ 1.00 7 Lee/2017 8 Taiwan, China 8 23114 Cycloplegic Autorefraction SA ≥ 1.00 3.7 Alrahili/2017 25 Medina, Saudi Arabia 3-10 1893 N/Cycloplegic Autorefraction TA ≥ 1.00 7.4 Hendriks/2009 4 Europe Maastricht, Netherlands 11-13 520 N/Cycloplegic Autorefraction SA ≥ 1.00 4.6 Flitcroft/2020 11 Ireland 6-7 362 Cycloplegic Autorefraction SA ≥ 1.00 6.9 Huynh/2006 26 Oceania Sydney, Australia 6 1765 Cycloplegic Autorefraction AS ≥ 1.00 AM ≥ 1.00 1.6 1.0 Comparing with others studies, some authors point to a frequency similar to the one found on this research, 4 , 7 , 8 , 11 , 25 others point to a lower frequency 24 and others still refer to a higher frequency. 22 , 23 Studies that included only children aged five and six years report rates of 1.3% and 1.6% (SA). 5 , 26 The frequency of SA, found in the present study, in preschool education (children from three to six years old) was 1%, this value being closer to those studies. Other studies included participants from 15 to 19 years old and report rates of 11,2% for SA. 23 For the 3 rd cycle of studies (average between 12 and 15 years of age), the present study indicated a frequency of 8.6% (SA) and according to the observed variation pattern, at a more advanced age and school stage, a higher rate is expected. Geographical and methodological issues make it difficult to compare prevalence studies. There is a pattern of greater variability in studies on the Asian continent, where for an identical age group, there are records ranging from 2.5% 24 to more than 10%; 7 however this is the continent where more studies on the subject are found. In Europe, more similar results are found, 4,6% and 6,9%. 4 , 11 In Portugal, a study in children from 6 to 11 years old found a 0.7% rate (5 in 672) of uncorrected anisometropia, 27 similar to the 1.1% (8 in 749) found in the present study. In the literature, the pattern of variation of anisometropia as a function of age and during the school period is not clear, presenting discordant results. Although many studies focus on children, it is common to have relatively wide age ranges. Most authors conclude that anisometropia varies with age, 4 , 5 , 7 , 28 although there are also studies where this relationship has not been found. 10 , 13 Studies on the subject, at school age, which showed an increasing prevalence with age were carried out in populations with a high frequency of myopia, a condition whose prevalence increases in adolescence. 28 On the other hand, longitudinal studies show that the prevalence of anisometropia increases after children start attending school. 5 , 7 , 25 Given these two lines that justify the variation of anisometropia during school age, it appears that they are related to each other, since the progress in the school path is accompanied by increasing age. In the present study, the anisometropia prevalence was also found to increase with advancement in the level of education. Consequently the age factor is also contributing to this situation, with myopic anisometropia being the one that most contributes to this variation pattern. Regarding the influence of gender on anisometropia, contradictory data are found in the scientific literature. While in one study it is reported that the prevalence found was higher in males 28 in others it is reported that the prevalence rates are higher in females. 10 , 23 The results of the present study revealed a higher frequency in females (7.1%) than in males (5.3%), however these differences were not statistically significant. This finding is in line with the results of other authors. 9 , 13 , 25 , 26 The influence of living in rural or urban areas has also been the object of study by several researchers, considering the development of myopia and, consequently, myopic anisometropia, which is more pronounced in urban areas. 7 , 8 , 10 The present study did not prove whether the area of residence and the frequency of anisometropia were related. This parameter is highly dependent on the way in which each author classifies the area of residence, as rural or urban, and the limits of these regions are sometimes difficult to define. Also, living in a rural area and working in an urban area means that the daily experience of some populations turns out to be more urban, in both environments. Some studies show that lifestyle parameters, such as reading and writing habits and time spent indoors, contribute to the prevalence of anisometropia variation. 7 , 8 , 10 , 11 This study has several strengths. The study was carried out in children and adolescents in a school environment. This allows minimizing the potential bias that occurs in the sampling process and avoids overestimation of the problem, when the investigation is carried out in a clinical environment. For this study, spherical anisometropia and astigmatic anisometropia were considered. The latter being disregarded in several studies and there is a record that this anisometropia is the one that most varies in terms of ethnicity. 9 Certain limitations can be pointed out in this study. Firstly, the use of autorefraction without cycloplegia is highlighted, which is not the gold standard method of refraction. This fact limits the analysis of the distribution of the different types of refractive errors found in the population studied. Despite the use of an open field auto refractometer, recognized as an instrument with good agreement with cycloplegic retinoscopy and which eliminates the need for cycloplegia in children 17 , 18 tends to underestimate hyperopia. 17 , 19 However, for the present study, this situation does not weaken the conclusions, as the refraction was evaluated in an open field and binocularly, both eyes are exposed to the same conditions. In addition to that studies by other authors with the same methodology, point out a sensitivity of 100% for the diagnosis of anisometropia. 19 Secondly, the choice of the cut-off point for the classification of anisometropia is pointed out. The literature recommends considering an IOD of at least 1.00D, however the sensitivity studies of the auto refractometer used, recommend considering 1.25D. 16 Thirdly, the signalling of the area of residence of the participants as rural or urban is reported, since the limits of these regions were at times difficult to define. Noting also that being a study carried out in an area of the inner country, whose territorial classification is of low density, it is expectable that habits and behaviours are more uniform between rural and urban areas. This wouldn’t be the case if the same study had been carried out in an area of greater population density. For future studies, data regarding reading habits and time spent outdoors will be collected, since these can be risk factors for the development of refractive errors. Conclusions The present study estimated the prevalence of anisometropia in Portuguese children from preschool to the 3 rd cycle of basic education finding an occurrence rate of 6.1%. The results of this work also showed that the level of the study cycle and the spherical anisometropia are related, verifying that it is low in preschool education and higher in the 3 rd cycle of basic education. It was also found that, hyperopic anisometropia was more prevalent in younger children and that with the progress of the school path, myopic anisometropia predominated. Regarding anisometropia, 17% of the subjects were uncorrected, which can be associated to an absence of visual related symptoms due to a possible amblyopia. Taking into account the cases of uncorrected anisometropia, in our opinion the implementation of visual screening programs is essential for the timely detection and correction of possible eye problems. This course of action will lead to better development, learning and school outcomes. Data availability Underlying data Dryad: Portuguese Children Refractive data - VER+ Project, https://doi.org/10.5061/dryad.h44j0zpm5. 29 This project contains the following underlying data: • PortugueseChildrenRefractiveDataVERmaisProjectV2.xlsx Data are available under the terms of the Creative Commons Zero “No rights reserved” data waiver (CC0 1.0 Public domain dedication). Acknowledgments To Rafaela Venâncio for the support and help in data collection. To the Vision Sciences Clinical and Experimental Centre and UBImedical for the availability of the equipment used and to the Beira Interior University Mission Group in Optometry and Vision Sciences for encouraging this study. 10.5256/f1000research.134743.r140056 Reviewer response for version 4 Lança Carla 12Referee https://orcid.org/0000-0001-9918-787X 1 Escola Superior de Tecnologia da Saúde de Lisboa (ESTeSL), Instituto Politecnico de Lisboa, Lisbon, Lisbon, Portugal 2 Comprehensive Health Research Center (CHRC), Escola Nacional de Saúde Pública, Universidade Nova de Lisboa, Lisboa, Portugal 12 7 2023 Copyright: © 2023 Lança C 2023 https://creativecommons.org/licenses/by/4.0/ This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Version 4recommendationapprove NA Is the work clearly and accurately presented and does it cite the current literature? Partly If applicable, is the statistical analysis and its interpretation appropriate? Partly Are all the source data underlying the results available to ensure full reproducibility? Yes Is the study design appropriate and is the work technically sound? Partly Are the conclusions drawn adequately supported by the results? Partly Are sufficient details of methods and analysis provided to allow replication by others? Partly Reviewer Expertise: Refractive errors and strabismus I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. 10.5256/f1000research.119923.r119238 Reviewer response for version 3 Lança Carla 12Referee https://orcid.org/0000-0001-9918-787X 1 Escola Superior de Tecnologia da Saúde de Lisboa (ESTeSL), Instituto Politecnico de Lisboa, Lisbon, Lisbon, Portugal 2 Comprehensive Health Research Center (CHRC), Escola Nacional de Saúde Pública, Universidade Nova de Lisboa, Lisboa, Portugal 13 1 2022 Copyright: © 2022 Lança C 2022 https://creativecommons.org/licenses/by/4.0/ This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Version 3recommendationapprove The authors have revised the paper and addressed my previous comments. I still have 3 minor suggestions as described below.  Page 1 of 28: and page 7 of 28 Please, delete “χ2(3) = 10.918;” Page 5 of 28: Please, delete the final “e” from “componente”. Please, delete “and by applying the central limit theorem, it can be considered that the violation of the assumptions does not have serious consequences." Is the work clearly and accurately presented and does it cite the current literature? Partly If applicable, is the statistical analysis and its interpretation appropriate? Partly Are all the source data underlying the results available to ensure full reproducibility? Yes Is the study design appropriate and is the work technically sound? Partly Are the conclusions drawn adequately supported by the results? Partly Are sufficient details of methods and analysis provided to allow replication by others? Partly Reviewer Expertise: Refractive errors and strabismus I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. Nunes Amelia Universidade da Beira Interior, Portugal 2 6 2022 Dear reviewer, All corrections have been made. Thank you so much for your careful and valuable review. 10.5256/f1000research.79990.r102249 Reviewer response for version 2 Serra Pedro M. 1Referee https://orcid.org/0000-0003-0471-0213 1 Research and Development Department, Ophthalmology Clinic Vista Sánchez Trancón, Badajoz, Spain 24 12 2021 Copyright: © 2021 Serra PM 2021 https://creativecommons.org/licenses/by/4.0/ This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Version 2recommendationapprove I would like to thank the authors for having implemented my recommendations in the current version of the manuscript. Generally, the manuscript improved considerably but it requires linguist editing. Specific comments Abstract: I recommend using the word prevalent instead of “frequent”, it highlights the aim of the study. When possible avoid passive voice “ An increase in anisometropia with the educational stage, was found in this study”. May be better: This study found an increase in… Introduction: “ Anisometropia is an ocular disorder characterized by an interocular difference (IOD) in refractive error. It represents a specific refractive condition where the two eyes of an individual, with similar sociodemographic, environmental, and genetic influences, can have asymmetric eye growth” Recommended - Anisometropia is an ocular disorder characterized by an interocular difference (IOD) in refractive error. It represents a specific refractive condition where the two eyes of an individual, can have asymmetric eye growth. “Scientific studies on the prevalence of anisometropia focus on children or adults, with less research being found in adolescence, table 1. Recommended: remove the word Scientific and place table 1 in the discussion. Methods: I think it would be useful to indicate how rural and urban schools were segmented. For instance, say rural Schools were defined when settled in a living area with less than X thousand inhabitants. “Only two volunteers were excluded.” Be consistent with the terminology used, participants or volunteers? I believe Table 2 would be better placed in the results part. Discussion The discussion needs to have the sentences shortened and the sentences need to be more objective. Is the work clearly and accurately presented and does it cite the current literature? Partly If applicable, is the statistical analysis and its interpretation appropriate? Yes Are all the source data underlying the results available to ensure full reproducibility? Yes Is the study design appropriate and is the work technically sound? Yes Are the conclusions drawn adequately supported by the results? Partly Are sufficient details of methods and analysis provided to allow replication by others? Partly Reviewer Expertise: Vision Sciences, Ophthalmic Optics. I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. Nunes Amelia Universidade da Beira Interior, Portugal 30 12 2021 Thank you so much for the review. We have made the changes accordingly based on the comments. Abstract: I recommend using the word prevalent instead of “frequent”, it highlights the aim of the study. Response: The wording was replaced in the abstract and in the remaining text. When possible avoid passive voice “ An increase in anisometropia with the educational stage, was found in this study”. May be better: This study found an increase in… Response: The sentence was modified. Introduction: “ Anisometropia is an ocular disorder characterized by an interocular difference (IOD) in refractive error. It represents a specific refractive condition where the two eyes of an individual, with similar sociodemographic, environmental, and genetic influences, can have asymmetric eye growth” Recommended - Anisometropia is an ocular disorder characterized by an interocular difference (IOD) in refractive error. It represents a specific refractive condition where the two eyes of an individual, can have asymmetric eye growth. Response: The sentence was modified. “Scientific studies on the prevalence of anisometropia focus on children or adults, with less research being found in adolescence, table 1. Recommended: remove the word Scientific and place table 1 in the discussion. Response: The word was removed and table 1 moved to the discussion section. Methods: I think it would be useful to indicate how rural and urban schools were segmented. For instance, say rural Schools were defined when settled in a living area with less than X thousand inhabitants. Response: That segmentation wasn’t done by this research team, it was provided by the municipalities representative. The following sentence was added to the sample characterization methods section “Rural and urban areas classification was based on the information provided by the municipalities, considering the residence area of each participant.”  “Only two volunteers were excluded.” Be consistent with the terminology used, participants or volunteers? Response: The terminology was all changed to participants. I believe Table 2 would be better placed in the results part. Response: In the first version of the manuscript, table 2 was in the results section. However since it presented data regarding the sample characterization, the other reviewer requested that it should be moved to the methods section. Discussion The discussion needs to have the sentences shortened and the sentences need to be more objective. Response: Several sentences have been changed in the discussion section. 10.5256/f1000research.79990.r102250 Reviewer response for version 2 Lança Carla 12Referee https://orcid.org/0000-0001-9918-787X 1 Escola Superior de Tecnologia da Saúde de Lisboa (ESTeSL), Instituto Politecnico de Lisboa, Lisbon, Lisbon, Portugal 2 Comprehensive Health Research Center (CHRC), Escola Nacional de Saúde Pública, Universidade Nova de Lisboa, Lisboa, Portugal 10 12 2021 Copyright: © 2021 Lança C 2021 https://creativecommons.org/licenses/by/4.0/ This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Version 2recommendationapprove The paper is much improved, and the authors have addressed most of my comments. However, I still have a few minor comments:   Introduction: Table 1 should be moved to the discussion section. It is not common to see tables in the introduction section of scientific papers. The table needs a legend for the abbreviations mentioned.  Results: Table 5: The values of Z(adjust) do not add much to the table and the p-value is best. I suggest deleting the Z(adjust) values. Conclusions: The authors refer that screening programmes are necessary. However, only 8 of the 46 participants did not have a correction, meaning that the majority had been already diagnosed. The conclusion needs further refinement.  Is the work clearly and accurately presented and does it cite the current literature? Partly If applicable, is the statistical analysis and its interpretation appropriate? Partly Are all the source data underlying the results available to ensure full reproducibility? Yes Is the study design appropriate and is the work technically sound? Partly Are the conclusions drawn adequately supported by the results? Partly Are sufficient details of methods and analysis provided to allow replication by others? Partly Reviewer Expertise: Refractive errors and strabismus I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. Nunes Amelia Universidade da Beira Interior, Portugal 30 12 2021 Thank you so much for the review. We have made the changes accordingly based on the comments. Introduction: Table 1 should be moved to the discussion section. It is not common to see tables in the introduction section of scientific papers. The table needs a legend for the abbreviations mentioned. Response:   Table 1 was moved to the discussion section (table and reference numbers will be updated) Results: Table 5: The values of Z(adjust) do not add much to the table and the p-value is best. I suggest deleting the Z(adjust) values. Response:   Table 5, Z(adjust) deleted Conclusions: The authors refer that screening programs are necessary. However, only 8 of the 46 participants did not have a correction, meaning that the majority had been already diagnosed. The conclusion needs further refinement.  Response: The last paragraph of the conclusion was changed to “Regarding anisometropia, 17% of the subjects were uncorrected, which can be associated to an absence of visual related symptoms due to a possible amblyopia. Taking into account the cases of uncorrected anisometropia, in our opinion the implementation of visual screening programs is essential for the timely detection and correction of possible eye problems. This course of action will lead to better development, learning and school outcomes.” 10.5256/f1000research.77322.r98419 Reviewer response for version 1 Serra Pedro M. 1Referee https://orcid.org/0000-0003-0471-0213 1 Research and Development Department, Ophthalmology Clinic Vista Sánchez Trancón, Badajoz, Spain 19 11 2021 Copyright: © 2021 Serra PM 2021 https://creativecommons.org/licenses/by/4.0/ This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Version 1recommendationapprove-with-reservations General comment Thank you for considering my name as a reviewer of the manuscript titled “Frequency of anisometropia in children and adolescents”. I would like to congratulate the authors for their work in this field of vision sciences which has a special interest in countries where is a shortage of epidemiologic data in the field of vision. Regarding the manuscript, my feeling is that it requires additional work before it is considered for indexing. In this report I will cover the technical issues that in my opinion would improve the quality of the manuscript, however, I would recommend the author's linguistic editing of the final manuscript. Examples of linguistic improvement are very long sentences with multiple ideas and some lack of concise writing. Specific topics Abstract: Provide a statement defining the topic of research and a second sentence where you state the aim of the study (eg. This study aims to estimate the prevalence of anisometropia in the Portuguese population of children and adolescents.) I suggest changing the word frequency to prevalence throughout the text. Methods: Indicate the study design? (e.g. Observational Cross-sectional study) How did you build the sample? Describe the dependent and independent variables Remove any data from the methods Results Use direct messages (e.g. No association was found between the presence of anisometropia with gender or area of residence) Conclusion The first sentence should be deleted this is not a conclusion from the present study. Introduction: I think the introduction requires some improvement 1 st paragraph – Provides a definition of the research topic “Anisometropia is an ocular disorder characterized by an interocular difference (IOD) in refractive error, representing a specific refractive condition insofar as the two eyes of an individual, with presumably similar sociodemographic, environmental and genetic influences, can have asymmetric eye growth.” Please reformulate this sentence, particularly remove the word presumably. 2 nd paragraph and 3 rd paragraph – mention the risk factors for anisometropia and the importance of assessing it 4 th paragraph- describes the instrument used. I would suggest moving this information to the methods 5 th paragraph- states the aim of the study. Missing information- the authors should mention the prevalence studies published, especially in Portugal, and compare with other European countries.   Methods: Please include: Type of study Nature of the sample, how did you select the different groups? There are differences in the socioeconomic nature of the study sites? Please indicate how did you define Rural and Urban populations? (e.g. number of inhabitants covered by each school)? Please indicate specifically the region where this study was conducted, this is important for instance in future meta-analysis studies. Duration of the study (e.g September 2021 – November 2021) Could you add information about the use of spectacles, you could try to see if the anisometropia had already been identified. Instrument description Include PlusOptix manufacturer city and country Provide the repeatability of the instrument (published) This is just a comment - “ Students' data for which it was not possible to obtain refraction were excluded, due to technical issues associated with the performance of the instrument (presence of strabismus, opacities, retinal anomalies or when the refractive error exceeded the instrument's measurement limit - spherical measurement range or cylindrical from −7.00 to +5.00D) or due to lack of cooperation from the participant.” This is probably one of the most valuable pieces of information from the data collection and it is a shame not being presented. Data Analysis Please indicate how the reader can interpret Thibos’ notation in regards to the J0 and J45. Say that the Cylinder component needs to be in negative form Anisometropia Classification I suggest writing the formulas used for calculating the SA, MA, and sMA. Statistical Analysis Results Sample characterization – 1 st paragraph move it to the methods section. Refractive State Please provide a table or figure (Box plot) with the refractive error information, where is possible to see the mean, standard deviation and range Figure 1 has limited information, I suggest including the percentages in the table mentioned above Anisometropia Please, use written numbers when placing them at the beginning of a sentence. (e.g Forty-six not 46) Provide information about the level of anisometropia per type. Describe the mean, standard deviation, and range Suggest creating a figure (bar plot with the percentages of anisometropia, total, young age group and older age group) Discussion The authors found a statistically significant association between the frequency of myopic anisometropia and age/ scholarly. Could the authors extend the discussion on the etiology of myopic anisometropia, which probably implies an asymmetric eye growth? Discussing this point could help the authors to suggest actions/mechanisms to prevent anisometropia. Discuss the consequences of hyperopic anisometropia, use the values (mean and standard deviation) for arguing about the number of subjects in risk of developing amblyopia. Is the work clearly and accurately presented and does it cite the current literature? Partly If applicable, is the statistical analysis and its interpretation appropriate? Yes Are all the source data underlying the results available to ensure full reproducibility? Yes Is the study design appropriate and is the work technically sound? Yes Are the conclusions drawn adequately supported by the results? Partly Are sufficient details of methods and analysis provided to allow replication by others? Partly Reviewer Expertise: Vision Sciences, Ophthalmic Optics. I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above. Nunes Amelia Universidade da Beira Interior, Portugal 30 11 2021 We thank the reviewer for his time and very constructive comments. We appreciate the reviewers' voluntary contributions in the form of helpful comments that allowed us to strengthen our article.  We updated the article according to your comments. Specific topics Abstract: Provide a statement defining the topic of research and a second sentence where you state the aim of the study (eg. This study aims to estimate the prevalence of anisometropia in the Portuguese population of children and adolescents.) I suggest changing the word frequency to prevalence throughout the text. R: The abstract was modified to include the topic and the aims “This research was developed to study the epidemiology of anisometropia. It aims to estimate the prevalence of anisometropia in Portuguese children and adolescents at various educational stages, studying its association with sociodemographic variables” Methods: Indicate the study design? (e.g. Observational Cross-sectional study) How did you build the sample? Describe the dependent and independent variables Remove any data from the methods R: All information was included and data was removed. The sample was collected from several kindergartens and schools in the central region of Portugal. The central region was included in the abstract and more details were added to the methods section. Results Use direct messages (e.g. No association was found between the presence of anisometropia with gender or area of residence) R: Section changed Conclusion The first sentence should be deleted this is not a conclusion from the present study. R: Section changed Introduction: I think the introduction requires some improvement 1 st paragraph – Provides a definition of the research topic “Anisometropia is an ocular disorder characterized by an interocular difference (IOD) in refractive error, representing a specific refractive condition insofar as the two eyes of an individual, with presumably similar sociodemographic, environmental and genetic influences, can have asymmetric eye growth.” Please reformulate this sentence, particularly remove the word presumably. 2 nd paragraph and 3 rd paragraph – mention the risk factors for anisometropia and the importance of assessing it 4 th paragraph- describes the instrument used. I would suggest moving this information to the methods 5 th paragraph- states the aim of the study. Missing information- the authors should mention the prevalence studies published, especially in Portugal, and compare with other European countries.   R: All information was included. A new table (new table 1) was introduced containing prevalence studies. We did not find any studies regarding anisometropia prevalence in Portugal, in indexed research papers. We found an indexed research paper by Lança, Serra, and Prista in Portuguese children, which indirectly refers to the number of uncorrected anisometropic children, but the methodology is very different from ours, including the anisometropia classification criterium. However, the percentage of uncorrected anisometropia is similar. We considered this fact relevant and it was included in the discussion. Methods: Please include: Type of study Nature of the sample, how did you select the different groups? There are differences in the socioeconomic nature of the study sites? Please indicate how did you define Rural and Urban populations? (e.g. number of inhabitants covered by each school)? Please indicate specifically the region where this study was conducted, this is important for instance in future meta-analysis studies. Duration of the study (e.g September 2021 – November 2021) Could you add information about the use of spectacles, you could try to see if the anisometropia had already been identified. R: This information is important and was included to some extent in the methods section. Rural and Urban definitions were based on the information provided by the municipalities representative. Instrument description Include PlusOptix manufacturer city and country Provide the repeatability of the instrument (published) This is just a comment - “ Students' data for which it was not possible to obtain refraction were excluded, due to technical issues associated with the performance of the instrument (presence of strabismus, opacities, retinal anomalies or when the refractive error exceeded the instrument's measurement limit - spherical measurement range or cylindrical from −7.00 to +5.00D) or due to lack of cooperation from the participant.” This is probably one of the most valuable pieces of information from the data collection and it is a shame not being presented. R: Information included. Although a study reports repeatability of around 0,6D for MSE with the A09 model (Plusoptix Vision Screener: the accuracy and repeatability of refractive measurements using a new autorefractor) this and other studies are not focused on anisometropia, for which IOD MSE errors could be minimized. Another study with a different plus optix model, reports a sensitivity of 100% for anisometropia (A comparison of plusoptiX A12 measurements with cycloplegic refraction). We included this study in the introduction. Data Analysis Please indicate how the reader can interpret Thibos’ notation in regards to the J0 and J45. Say that the Cylinder component needs to be in negative form R: This information has been added Anisometropia Classification I suggest writing the formulas used for calculating the SA, MA, and sMA. R: The formulas were added Statistical Analysis Results Sample characterization – 1 st paragraph move it to the methods section. R: Paragraph moved   Refractive State Please provide a table or figure (Box plot) with the refractive error information, where is possible to see the mean, standard deviation, and range Figure 1 has limited information, I suggest including the percentages in the table mentioned above R: A new table (table 4) containing this data was added to this section Anisometropia Please, use written numbers when placing them at the beginning of a sentence. (e.g Forty-six not 46) Provide information about the level of anisometropia per type. Describe the mean, standard deviation, and range Suggest creating a figure (bar plot with the percentages of anisometropia, total, young age group and older age group) R: Sentence rewritten Mean, standard deviation, and range of anisometropia per type was added in table 4. An extra figure would duplicate the information. Information about the age group is already included in table 5, distributed by the study cycle which indirectly represents age groups. An extra figure would duplicate the information. Discussion The authors found a statistically significant association between the frequency of myopic anisometropia and age/ scholarly. Could the authors extend the discussion on the etiology of myopic anisometropia, which probably implies an asymmetric eye growth? Discussing this point could help the authors to suggest actions/mechanisms to prevent anisometropia. R: This is an excellent suggestion for future research. We can hypothesise that an excess of near-work activities, associated with incorrect posture, can lead to asymmetric eye growth. However, those factors were not controlled in the present study.   Discuss the consequences of hyperopic anisometropia, use the values (mean and standard deviation) for arguing about the number of subjects at risk of developing amblyopia. R: Only 2 uncorrected anisometropes were hyperopic, and since no BCVA data was collected, we feel that there are no sufficient data to support this kind of analysis. However, we agree that it is an excellent suggestion for future work. 10.5256/f1000research.77322.r98416 Reviewer response for version 1 Lança Carla 12Referee https://orcid.org/0000-0001-9918-787X 1 Escola Superior de Tecnologia da Saúde de Lisboa (ESTeSL), Instituto Politecnico de Lisboa, Lisbon, Lisbon, Portugal 2 Comprehensive Health Research Center (CHRC), Escola Nacional de Saúde Pública, Universidade Nova de Lisboa, Lisboa, Portugal 12 11 2021 Copyright: © 2021 Lança C 2021 https://creativecommons.org/licenses/by/4.0/ This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Version 1recommendationapprove-with-reservations The study reports on the prevalence of anisometropia in a sample of Portuguese children and adolescents. The wording needs major revision. I have several comments as shown below. Title I suggest replacing the word “frequency” by “prevalence”. This suggestion also applies throughout all sections of the manuscript. Abstract Methods - Please replace “cycloplegic” by "cycloplegia”. Results - Please revise the wording of the following sentence “to state that the occurrence…”; The last sentence of the results it’s not clear and should be revised. Introduction The first sentence is too long, and the wording needs to be revised. I suggest splitting it into two sentences. I have a similar comment for the third paragraph of the introduction section. There is no mention to previous studies on anisometropia prevalence. Also, there is the need to show more detailed results on the studies reporting risk factors and its association with anisometropia (e.g., sample size, age, p values, etc.). Methods The methods section is incomplete. The methods section of a research paper provides the information by which a study's validity is judged. Therefore, it requires a clear and precise description of how an experiment was done, and the rationale for why specific experimental procedures were chosen. For example, more information is necessary on the sampling method and authors should describe their study design. Results The first paragraph should be moved to the methods section. Table 2 – Replace “statistical test” by p value only (follow the same recommendation in the text results). Also, include a note referring to which comparisons are being made (for example, is it the distribution of children by cycles or the age by cycles?). At the time you are writing your article, you have already completed your study, so you should use past tense in your methodology and results section to record what you did and found (for example, “this difference was statistically significant”). Refractive state and anisometropia sections need to be re-written, as it is confusing to present two percentages at the same time. Perhaps the authors can consider showing separate sentences for those results. Influence of sociodemographic variables: the first sentence of this section should be moved to the statistical analysis; Delete “therefore this parameter is not in the table” and replace by the p value.   Table 3 – delete the X2 value and only present p value. Non-significant p values may be presented with only 2 decimal places. For risk factor analysis the multivariate adjusted analysis should be additionally shown. Figure 2. The colour of the legend is not clear, and it is difficult to distinguish the categories presented in the graph. The y axis wording needs revision. Discussion Please, delete the 1 st sentence. The differences between prevalence of anisometropia in Asia and Europe need a more careful explanation. Environmental factors play a major role in the development of refractive errors, especially myopia. Another limitation of the study that was not referred is that factors such as reading habits or outdoor time were not collected and those may be risk factors for the development of refractive errors. Is this the first study in Portugal? Are your results comparable with other studies in Portugal? Conclusions Please revise the wording on the sentence that includes the following: “we are of the opinion”. Is the work clearly and accurately presented and does it cite the current literature? Partly If applicable, is the statistical analysis and its interpretation appropriate? Partly Are all the source data underlying the results available to ensure full reproducibility? Yes Is the study design appropriate and is the work technically sound? Partly Are the conclusions drawn adequately supported by the results? Partly Are sufficient details of methods and analysis provided to allow replication by others? Partly Reviewer Expertise: Refractive errors and strabismus I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above. Nunes Amelia Universidade da Beira Interior, Portugal 30 11 2021 We would like to thank the reviewer for her valuable time and comments on our article; this will be very useful for us in future research and has allowed us to improve this document. We have made the required corrections. Title I suggest replacing the word “frequency” by “prevalence”. This suggestion also applies throughout all sections of the manuscript. R: The title was replaced Abstract Methods - Please replace “cycloplegic” by "cycloplegia”. R: The word was replaced   Results - Please revise the wording of the following sentence “to state that the occurrence…”; The last sentence of the results it’s not clear and should be revised. R: Sentence revised to “No statistical differences were found between genders or between areas of residence regarding the rate of anisometropia”; Last sentence revised to “Regarding spherical equivalent anisometropia, there is a pattern of variation that increases with the cycle of studies (�� 2 (3)= 10.918; p = 0.012), with myopic anisometropia being the main contributor to this variation.” Introduction The first sentence is too long, and the wording needs to be revised. I suggest splitting it into two sentences. I have a similar comment for the third paragraph of the introduction section. R: The first sentence was split and wording revised to:” Anisometropia is an ocular disorder characterized by an interocular difference (IOD) in refractive error. It represents a specific refractive condition where the two eyes of an individual, with presumably similar sociodemographic, environmental and genetic influences, can have asymmetric eye growth.” The third paragraph split and wording revised to:” Although it is not clear what is the ideal age to perform the correction, in order to guarantee an ideal visual development and maturation, the early correction of anisometropia is important. 11 This prevents the development of other changes such as aniseikonia, amblyopia and strabismus 3,12,13 and even in small degrees (<1D) facilitates emmetropization. 11 It also improves quality of life, reducing or eliminating symptoms of visual discomfort.”   There is no mention to previous studies on anisometropia prevalence. Also, there is the need to show more detailed results on the studies reporting risk factors and its association with anisometropia (e.g., sample size, age, p values, etc.). R: These data were introduced in a new table, and the reference numbering was changed accordingly. Methods The methods section is incomplete. The methods section of a research paper provides the information by which a study's validity is judged. Therefore, it requires a clear and precise description of how an experiment was done, and the rationale for why specific experimental procedures were chosen. For example, more information is necessary on the sampling method and authors should describe their study design. R: The method section was completed including the following text “Data from the Portuguese Census 2011, shows a population of 319284 from 0 to 14 years old in the central region. For a confidence level of 95% and a 5% margin of error, taking into account that the prevalence of the studied condition is unknown, it was fixed at 50% to obtain a large enough sample size. The result was a minimum of 384 subjects. The data collection took place in 5 schools of the central region of Portugal, including all students that were authorized to participate by their legal guardians.” Results The first paragraph should be moved to the methods section. R: The paragraph was moved   Table 2 – Replace “statistical test” by p value only (follow the same recommendation in the text results). Also, include a note referring to which comparisons are being made (for example, is it the distribution of children by cycles or the age by cycles?). R: Done. This table was moved to the methods section, and the statistical analysis was deleted since it did not provide significant information. At the time you are writing your article, you have already completed your study, so you should use past tense in your methodology and results section to record what you did and found (for example, “this difference was statistically significant”). R: Text rewritten to past tense.   Refractive state and anisometropia sections need to be re-written, as it is confusing to present two percentages at the same time. Perhaps the authors can consider showing separate sentences for those results. R: Both sections were rewritten to avoid double percentages.  Influence of sociodemographic variables: the first sentence of this section should be moved to the statistical analysis; Delete “therefore this parameter is not in the table” and replace by the p-value.   R: The sentence was moved to the methods section. “Therefore..” was removed and the p-value information was added to the table.   Table 3 – delete the X2 value and only present p-value. Non-significant p values may be presented with only 2 decimal places. R: X2 values deleted and p values with 2 decimal places For risk factor analysis the multivariate adjusted analysis should be additionally shown. R: The multivariate analysis was added to the table.   Figure 2. The colour of the legend is not clear, and it is difficult to distinguish the categories presented in the graph. The y axis wording needs revision. R: Legend colour changed. Y-axis caption revised. Discussion Please, delete the 1 st sentence. R: Sentence deleted The differences between prevalence of anisometropia in Asia and Europe need a more careful explanation. Environmental factors play a major role in the development of refractive errors, especially myopia. R: An extra sentence was added to explain to acknowledge the contribution of environmental factors “Some studies show that lifestyle parameters, such as reading and writing habits and time spent indoors, contribute to the prevalence of anisometropia variation. 7,8,10,11” Another limitation of the study that was not referred to is that factors such as reading habits or outdoor time were not collected and those may be risk factors for the development of refractive errors. R: We recognise the relevance of these factors, and they will be included in future studies. Accordingly to your suggestions, the following sentence was added to the end of the text “For future studies, data regarding reading habits and time spent outdoors will be collected, since these can be risk factors for the development of refractive errors.” Is this the first study in Portugal? Are your results comparable with other studies in Portugal? R: This is not the first study in Portugal, however in indexed research papers we couldn’t find any comparable studies for Portugal. We found an indexed research paper by Lança, Serra, and Prista in Portuguese children, which indirectly refers to the number of uncorrected anisometropic children, but the methodology is very different from ours, including the anisometropia classification criterium. However, the percentage of uncorrected anisometropia is similar. We considered this fact relevant and it was included in the discussion. Conclusions Please revise the wording on the sentence that includes the following: “we are of the opinion”. R: The sentence was revised and divided to: “Taking into account the high frequency of anisometropia found in 3 rd cycle students, in our opinion the implementation of visual screening programs at this age is essential for the timely detection and correction of possible eye problems.  This course of action will lead to better development, learning, and school outcomes at these ages.” Competing interests: No competing interests were disclosed. Competing interests: No competing interests were disclosed. Competing interests: No competing interests were disclosed. Competing interests: No competing interests were disclosed. Competing interests: No competing interests were disclosed. Competing interests: No competing interests were disclosed. Competing interests: No competing interests were disclosed. Competing interests: No competing interests were disclosed. Competing interests: No competing interests. Competing interests: No competing interests were disclosed. Competing interests: No competing interests. ==== Refs References 1 Vincent SJ Collins MJ Read SA : Myopic anisometropia: Ocular characteristics and aetiological considerations. Clin. Exp. Optom. 2014;97 (4 ):291–307. 10.1111/cxo.12171 24939167 2 Smith EL III Hung LF Arumugam B : Observations on the relationship between anisometropia, amblyopia and strabismus. Vis. Res. 2017;134 :26–42. 10.1016/j.visres.2017.03.004 28404522 3 South J Gao T Collins A : Aniseikonia and anisometropia: implications for suppression and amblyopia. Clin. Exp. Optom. 2019;102 (6 ):556–565. 10.1111/cxo.12881 30791133 4 Hendricks TJ De Brabander J Vankan-Hendricks MH : Prevalence of habitual refractive errors and anisometropia among Dutch schoolchildren and hospital employees. Acta Ophthalmol. 2009;87 (5 ):538–543. 10.1111/j.1755-3768.2008.01251.x 19416112 5 Deng L Gwiazda JE : Anisometropia in children from infancy to 15 years. Investig. Ophthalmol. Vis. Sci. 2012;53 (7 ):3782–7. 10.1167/iovs.11-8727 22589429 6 Yoo YC Kim JM Park KH : Refractive errors in a rural Korean adult population: the Namil Study. Eye. 2013;27 (12 ):1368–1375. 10.1038/eye.2013.195 24037232 7 Hu YY Wu JF Lu TL : Prevalence and associations of anisometropia in Children. Investig. Ophthalmol. Vis. Sci. 2016;57 (3 ):979–988. 10.1167/iovs.15-18647 26962694 8 Lee CW Fang SY Tsai DC : Prevalence and association of refractive anisometropia with near work habits among young schoolchildren: The evidence from a population-based study. PLoS One. 2017;12 (3 ):1–15. 10.1371/journal.pone.0173519 9 Afsari S Rose KA Gole GA : Prevalence of anisometropia and its association with refractive error and amblyopia in preschool children. Br. J. Ophthalmol. 2013;97 (9 ):1095–1099. 10.1136/bjophthalmol-2012-302637 23613508 10 Wei S Sun Y Li S : Refractive errors in university students in central China: the Anyang University Students Eye Study. Investig. Ophthalmol. Vis. Sci. 2018;59 (11 ):4691–4700. 10.1167/iovs.18-24363 30267091 11 Flitcroft I Mccullough S Saunders K : What can anisometropia tell us about eye growth?. Br. J. Ophthalmol. 2021;105 :bjophthalmol-2020-316406–bjophthalmol-2020-311215. 10.1136/bjophthalmol-2020-316406 12 Jeon HS Choi DG : Stereopsis and fusion in anisometropia according to the presence of amblyopia. Graefes Arch. Clin. Exp. Ophthalmol. 2017;255 (12 ):2487–2492. 10.1007/s00417-017-3798-3 28889226 13 Dobson V Miller JM Clifford-Donaldson CE : Associations between anisometropia, amblyopia, and reduced stereoacuity in a school-aged population with a high prevalence of astigmatism. Investig. Ophthalmol. Vis. Sci. 2008;49 (10 ):4427–4436. 10.1167/iovs.08-1985 18539935 14 Nunes AF Monteiro PM Ferreira FB : Convergence insufficiency and accommodative insufficiency in children. BMC Ophthalmol. 2019;19 (1 ):1–8. 10.1186/s12886-019-1061-x 30606142 15 Nunes AF Sena F Calado R : Reduced visual acuity in children from 5 to 6 years old, with LEA chart. Graefes Arch. Clin. Exp. Ophthalmol. 2021;259 :759–768. 10.1007/s00417-020-04927-x 32945936 16 Silbert D Matta N Tian J : Comparing the SureSight autorefractor and the plusoptiX photoscreener for pediatric vision screening. Strabismus. 2014;22 (2 ):64–67. 10.3109/09273972.2014.904896 24738949 17 Dikkaya F Erdur SK : Comparison of the PlusOptix S09 and Spot Vision photorefractor to cycloretinoscopy. Int. Ophthalmol. 2018;2018 :3–10. 18 Yilmaz I Ozkaya A Alkin Z : Comparison of the Plusoptix A09 and Retinomax K-Plus 3 With Retinoscopy in Children. J. Ped. Ophthalmol. Strab. 2015;52 (1 ):37–42. 2015. 19 Fogel-Levin M Doron R Wygnanski-Jaffe T : A comparison of plusoptiX A12 measurements with cycloplegic refraction. J. AAPOS. 2016;20 (4 ):310–4. 10.1016/j.jaapos.2016.04.006 27422572 20 Census 2011 Resultados por Freguesia. Portugal: Instituto Nacional de Estatística;2011 [cited 2021 Nov 23]. Reference Source 21 Thibos LN Wheeler W Horner D : Power vectors: an application of Fourier analysis to the description and statistical analysis of refractive error. Optom. Vis. Sci. 1997;74 (6 ):367–375. 10.1097/00006324-199706000-00019 9255814 22 Ohlsson J Villarreal G Sjöström A : Visual acuity, amblyopia, and ocular pathology in 12- to 13-year-old children in Northern Mexico. J. AAPOS. 2003;7 (1 ):47–53. 10.1067/mpa.2003.S1091853102420113 12690370 23 Quek TP Chua CG Chong CS : Prevalence of refractive errors in teenage high school students in Singapore. Ophthalmic Physiol. Opt. 2004;24 (1 ):47–55. 10.1046/j.1475-1313.2003.00166.x 14687201 24 Yekta A Fotouhi A Hashemi H : Prevalence of refractive errors among schoolchildren in Shiraz, Iran. Clin. Exp. Ophthalmol. 2010;38 (3 ):242–248. 10.1111/j.1442-9071.2010.02247.x 20447119 25 Alrahili NHR Jadidy ES Alahmadi BSH : Prevalence of uncorrected refractive errors among children aged 3-10 years in western Saudi Arabia. Saudi Med. J. 2017;38 (8 ):804–810. 10.15537/smj.2017.8.20412 28762432 26 Huynh SC Wang XY Ip J : Prevalence and associations of anisometropia and aniso-astigmatism in a population based sample of 6 year old children. Br. J. Ophthalmol. 2006;90 (5 ):597–601. 10.1136/bjo.2005.083154 16622090 27 Lança C Serra H Prista J : Strabismus, visual acuity, and uncorrected refractive error in Portuguese children aged 6 to 11 years. Strabismus. 2014;22 (3 ):115–119.25019452 28 Wajuihian SO Mashige KP : Gender and age distribution of refractive errors in an optometric clinical population. J. Optom. 2021. 10.1016/j.optom.2020.09.002 29 Nunes AF Batista MJ Monteiro PL : Portuguese Children Refractive data - VER+ Project. 2021. 10.5061/dryad.h44j0zpm5
PMC008xxxxxx/PMC8913730.txt
==== Front Nat Commun Nat Commun Nature Communications 2041-1723 Nature Publishing Group UK London 35273166 28869 10.1038/s41467-022-28869-6 Article Engineered assembly of water-dispersible nanocatalysts enables low-cost and green CO2 capture Alivand Masood S. 1 Mazaheri Omid 12 Wu Yue 1 http://orcid.org/0000-0002-5860-8938 Zavabeti Ali 13 http://orcid.org/0000-0003-0904-6630 Christofferson Andrew J. 34 Meftahi Nastaran 4 http://orcid.org/0000-0003-3589-3040 Russo Salvy P. 4 http://orcid.org/0000-0002-5788-4682 Stevens Geoffrey W. 1 Scholes Colin A. 1 http://orcid.org/0000-0002-7056-5600 Mumford Kathryn A. [email protected] 1 1 grid.1008.9 0000 0001 2179 088X Department of Chemical Engineering, The University of Melbourne, Melbourne, Vic 3010 Australia 2 grid.1008.9 0000 0001 2179 088X School of Agriculture and Food, Faculty of Veterinary and Agricultural Sciences, The University of Melbourne, Melbourne, Vic 3010 Australia 3 grid.1017.7 0000 0001 2163 3550 School of Science, RMIT University, Melbourne, Vic 3001 Australia 4 grid.1017.7 0000 0001 2163 3550 ARC Centre of Excellence in Exciton Science, School of Science, RMIT University, Melbourne, Vic 3000 Australia 10 3 2022 10 3 2022 2022 13 12491 7 2021 31 1 2022 © The Author(s) 2022 https://creativecommons.org/licenses/by/4.0/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Catalytic solvent regeneration has attracted broad interest owing to its potential to reduce energy consumption in CO2 separation, enabling industry to achieve emission reduction targets of the Paris Climate Accord. Despite recent advances, the development of engineered acidic nanocatalysts with unique characteristics remains a challenge. Herein, we establish a strategy to tailor the physicochemical properties of metal-organic frameworks (MOFs) for the synthesis of water-dispersible core-shell nanocatalysts with ease of use. We demonstrate that functionalized nanoclusters (Fe3O4-COOH) effectively induce missing-linker deficiencies and fabricate mesoporosity during the self-assembly of MOFs. Superacid sites are created by introducing chelating sulfates on the uncoordinated metal clusters, providing high proton donation capability. The obtained nanomaterials drastically reduce the energy consumption of CO2 capture by 44.7% using only 0.1 wt.% nanocatalyst, which is a ∽10-fold improvement in efficiency compared to heterogeneous catalysts. This research represents a new avenue for the next generation of advanced nanomaterials in catalytic solvent regeneration. Catalytic solvent regeneration is of interest to reduce energy consumption in CO2 separation, however, the development of engineered nanocatalysts remains a challenge. Here, a new avenue is presented for the next generation of advanced metal-organic frameworks (MOFs) in energy-efficient CO2 capture. Subject terms Carbon capture and storage Chemical engineering Metal-organic frameworks issue-copyright-statement© Springer Nature Limited 2022 ==== Body pmcIntroduction The growing trend of fossil fuel consumption and anthropogenic CO2 emission has gained global attention as an urgent environmental issue1,2. To overcome this challenge, the Paris Agreement was signed in 2015 by consensus to keep the Earth’s temperature rise well below 2 °C by mid-century3,4. Despite this, CO2 emission has been relentlessly increasing and approaching 40 GtCO2/year due to persistent release from key industrial sectors5. A strategy to reduce CO2 emissions while keeping existing industrial assets is carbon capture and storage (CCS). However, given the limited number of large-scale CO2 capture plants and the noticeable reduction in fossil fuel prices, hundreds of CCS facilities will need to be constructed by 2030 for the successful achievement of the Paris Agreement targets6. Chemical solvent CO2 absorption–desorption, as the most viable and widely accepted technique for carbon separation, is not inherently green and its high energy demand for solvent regeneration indirectly contributes to global CO2 emissions. Additionally, the energy-intensive nature of separation strongly impacts the economics of the process, leading to a reluctance to invest in CCS projects. Hence, it is of high importance to deploy a diverse portfolio of energy-efficient and green CO2 capture technologies with net-zero emissions. The remarkable energy consumption of CO2 separation is mainly attributed to the high solvent regeneration temperature (above 100 °C) required to accelerate CO2 desorption kinetics7,8. Recently, catalytic solvent regeneration has emerged as an interesting approach to effectively promote the CO2 desorption rate and reduce the required regeneration energy9,10. One of the distinguishing features of catalytic regeneration is its potential to critically reduce the CO2 desorption temperature below 100 °C and accordingly pave the way for utilizing lower grade heat resources, such as solar hot water, as a green approach for solvent regeneration7,11. In this regard, a wide range of commercial and synthetic heterogeneous catalysts have been tested for catalyst-aided solvent regeneration, though their low efficiency and operational difficulties are still a significant barrier toward large-scale implementation11. Metal–organic frameworks (MOFs) are an emerging class of porous materials with highly favorable catalytic properties owing to their large pore volume, abundant metal sites, and tailorable structure12,13. These features have made MOFs potential candidates for developing advanced nanocatalysts. Yaghi et al. recently reported the synthesis of an acidic MOF-SO4 catalyst by treating a zirconium-based MOF (MOF-808) with aqueous sulfuric acid14,15. They showed the strong Brønsted acidity of MOF-808-SO4 originates from the chelating sulfates attached to the unsaturated metal clusters which release protons into the reaction medium. However, the microporosity of MOFs is still a major concern for the fabrication of super-efficient nanomaterials, particularly in catalysis16–18. Additionally, there are other acid-stable MOFs desired for the fabrication of MOF-SO4 nanocatalysts which have not yet been explored. To this end, it is of great interest to exploit advanced MOF-SO4 materials with targeted physicochemical features for low-temperature CO2 desorption, thereby taking a step further toward establishing green CO2 capture technologies and achieving the objectives of the Paris Climate Accord. Herein, we present a new strategy for the fabrication of engineered water-dispersible nanocatalysts with hierarchical micro-mesoporous structures (Fig. 1 and Supplementary Fig. 4). In this technique, we used acidic Fe3O4 nanoclusters as a versatile substrate for the modulated self-assembly of different MOFs with broad structural diversity. The presence of carboxylates surrounding the core surface leverages the creation of active mesoporosity throughout the network of assembled MOFs. Specifically, the grown mesoporous shell facilitates the subsequent coordination of chelating sulfate moieties onto the metal clusters, thus allowing for a higher density of Brønsted acid sites. The accommodation of approachable acidic sites through the tailored hierarchical structure, as well as the nanofluidic aspect readily enables water-dispersible Fe3O4@MOF-SO4 nanocatalyst to actively participate in CO2 desorption reactions and drastically reduce energy consumption.Fig. 1 Synthetic strategy for engineering water-dispersible nanocatalysts. Schematic illustration of sulfated core–shell nanomaterials through modulated self-assembly of MOFs on the carboxylate-rich surface of Fe3O4–COOH nanoclusters. The acidic cover of magnetic nanocluster enables the creation of missing-ligand defects in the assembled structure and forms a hierarchical micro-mesoporous network. The coordination of chelating sulfate on metal clusters results in the formation of water-dispersible nanocatalysts with superacidity. Results Water-dispersible acidic Fe3O4 substrate We first examined the potential of acidic Fe3O4 nanoclusters as favorable water-dispersible supports for the formation of advanced nanocatalysts. Ferric ammonium citrate, containing both iron (Fe3+) and citric acid, was selected as a cheap and environmentally benign precursor for the fabrication of carboxylated Fe3O4 nanoclusters (Fe3O4–COOH). One of the benefits of the citrate assembly approach is that the carboxylate groups on the surface of the nanocluster can serve as proton donor sites. Additionally, carboxylates can also provide stable hydrogen bonds to water molecules, allowing great dispersibility and facilitating the nanofluidic behavior of Fe3O4–COOH in aqueous solvents (Supplementary Fig. 5). The combination of these two features enables the Fe3O4–COOH to be simply utilized as a water-dispersible acidic catalyst during the continuous CO2 absorption–desorption operation without any modification to the process configuration. The surface modification is achieved via two steps by the assembly and adhesion of citrate ions to the Fe3O4 surface (Fig. 2a). Upon partial reduction of Fe3+ to Fe2+ in the non-aqueous solution and formation of single-crystal Fe3O4 nanoparticles, citrate ions easily attach to the surface owing to the strong coordination affinity between carboxylate groups and Fe3+/Fe2+ ions, resulting in aggregation of single nanoparticles to form large Fe3O4 nanoclusters19,20. Helium ion microscopy and scanning electron microscopy (SEM) showed the spherical clusters ranging in size from ⁓100 to 300 nm (Fig. 2b, c and Supplementary Fig. 6). Specifically, transmission electron microscopy (TEM) indicated the aggregated nanoclusters are composed of small adhesive Fe3O4 nanoparticles with about ⁓2-5 nm size (Fig. 2d). High-angle annular dark-field (HAADF), energy-dispersive X-ray spectroscopy (EDX) mapping, and elemental line scanning confirmed the homogeneous morphology of Fe3O4 nanoclusters with uniform Fe and O distribution throughout the structure (Fig. 2e and Supplementary Fig. 7). In addition, T2g and Eg peaks were identified as the prominent vibrational modes in the Raman spectrum which confirm the successful formation of the Fe3O4 structure (Fig. 2f).Fig. 2 Synthesis of the water-dispersible magnetic substrate with acidic characteristics. a Schematic of the formation of nanoclusters in a one-pot synthesis methodology. b HIM and c SEM images of the acidic Fe3O4–COOH substrate. Scale bars are 200 nm. d TEM image of a single Fe3O4–COOH nanocluster and its corresponding e HAADF and EDX mapping. Scale bars are 50 nm. f Raman and g FTIR spectra of Fe3O4–COOH structure. Insets in g are the images of Fe3O4–COOH dispersion in pure water (a), followed by adding silicon oil (b), toluene (c), and diethyl ether (d). Before taking the picture, the mixtures were completely shaken and incubated for 15 min and nanocatalyst concentration was kept constant at 0.1 wt%; the transparency of the nonaqueous phase indicates the hydrophilicity of Fe3O4–COOH nanoclusters and their accumulation in the aqueous phase. h Full survey and i high-resolution XPS spectra showing the chemical moieties on the exterior surface of Fe3O4–COOH nanoclusters. Different characterization methods were utilized to confirm the successful assembly of carboxylic acid groups (COOH) on the surface of Fe3O4–COOH nanoclusters. Fourier transform infrared spectroscopy (FTIR) displayed two characteristic peaks at ⁓1340 and ⁓1610 cm−1 corresponding to the carbonyl group (C=O) and C–OH stretching of the carboxylic acid group (Fig. 2g)21. X-ray photoelectron spectroscopy (XPS) revealed that the majority of Fe3O4–COOH surface is covered by O (⁓58.9%) and C (⁓29.4%), while the share of Fe was ⁓9.1% (Fig. 2h, i and Supplementary Fig. 8). In contrast, the elemental analysis showed the bulk is mainly composed of Fe (⁓87.0%), which is ⁓28-fold greater than that of C (⁓2.9%) in the bulk sample (Supplementary Fig. 9), suggesting that COOH groups have mainly accumulated on the external surface of the Fe3O4–COOH nanocluster. The anchored citrate groups can subsequently favor the hydrophilic nature of Fe3O4–COOH, its electrostatic stabilization, and dispersibility in the aqueous media, due to the intense negative charge density of the surface. To further demonstrate the hydrophilicity of Fe3O4–COOH, its aqueous solution was mixed with silicon oil, toluene, and dichloromethane, and no Fe3O4–COOH was observed in the non-aqueous solvents, highlighting the catalytic potential of Fe3O4–COOH in phase induced areas, such as CO2 capture using liquid–liquid phase change solvents (Fig. 2g)22. These results indicate that Fe3O4–COOH can be viably used as a water-dispersible substrate for further surface modification, having tunable physicochemical motifs that are of interest for catalytic solvent regeneration in CO2 capture. Tailoring the properties of Fe3O4@MOF To appraise the modulated self-assembly of different MOFs on acidic Fe3O4–COOH substrate, seven MOFs, including ZIF-8, ZIF-67, MIL-100(Fe), MOF-Fe(II), HKUST-1, UiO-66, and UiO-66-NH2 with various metal (zinc (Zn), zirconium (Zr), cobalt (Co), copper (Cu), ferrous (Fe2+), and ferric (Fe3+)) and ligand (benzene-1,4-dicarboxylic acid (H2BDC), 2-aminoterephthalic acid (H2BDC-NH2), 2,5-pyridinedicarboxylic acid (H2BDC-N), benzene-1,3,5-tricarboxylic acid (H3BTC) and 2-methylimidazole (2-Melm)) combinations were utilized (Fig. 3a). After growing MOFs, the zeta potential of Fe3O4–COOH substrate in all cases changed from −43.1 mV to greater values (ranging from −1.7 to −19.5 mV), indicating that the surface of the substrate is primarily covered by metal–ligand assembly (Supplementary Fig. 10). This variation was due to the coordination of citrate ions and thereby diminution in the negative surface charge onto the nanocluster. In detail, the citrate groups of Fe3O4–COOH nanocluster can be easily dissociated and create a negatively charged surface, while hydroxyl groups of MOFs have a lower dissociation constant than those of carboxylates resulting in a decrease in the negative surface charge of Fe3O4@MOFs. The successful preparation of Fe3O4@MOF core–shell structure was further confirmed by TEM images, nitrogen adsorption–desorption analysis, FTIR, XRD, Raman spectroscopy, and XPS characterization (Fig. 3b–d and Supplementary Figs. 11–16).Fig. 3 Structural tailoribility of MOFs via modulated self-assembly. a Schematic diagram of different metal ions and organic ligands used to grow various MOF coatings with unique physicochemical characteristics. b Raman spectra of the Fe3O4–COOH, UiO-66, and Fe3O4@UiO-66. c Full survey XPS spectra of Fe3O4@HKUST-1 (I), Fe3O4@UiO-66 (II), and Fe3O4@UiO-66-NH2 (III). d Nitrogen adsorption-desorption isotherms of UiO-66 and Fe3O4@UiO-66 at 77 K. e TGA profiles of the Fe3O4-COOH, HKUST-1, and Fe3O4@HKUST-1 under nitrogen atmosphere. SEM images of f Fe3O4@ZIF-8, g Fe3O4@ZIF-67, h Fe3O4@MOF-Fe(II), i Fe3O4@MIL-100(Fe), j Fe3O4@HKUST-1, and k Fe3O4@UiO-66. Scale bars are 200 nm. l Pore volume of the hierarchical Fe3O4@MOF materials with missing-linker defects and mesopore-induced network. m The coating weight and comparative weight reduction of various Fe3O4@MOF core–shells. The comparative weight reduction was calculated by the weight loss of core–shell material (ΔWC-S) divided by that of its corresponding pristine MOF (ΔWP) in the 100–350 °C range. SEM images were also employed to observe the exterior of Fe3O4@MOF materials. As shown in Fig. 3f–k, Fe3O4@MOFs exhibited almost different surface morphologies from highly uniformed spherical shapes, similar to that of pristine Fe3O4–COOH without MOF coating, to semispherical conjunct core-shell structure, like Fe3O4@MIL-100(Fe). This structural diversity is mainly imputed to the unique characteristic features of MOFs. As MOFs are created from different metal–ligand coordination, a series of frameworks with various physicochemical properties, crystal size, and surface roughness can be obtained (Supplementary Figs. 17–19). Furthermore, the high density of carboxylic acid sites on the rough surface of the substrate, as a potential nucleation site, can effectively manipulate the size and shape of the crystal, justifying the super small MOF crystals spotted on the exterior surface of core–shell structures compared to their pristine MOFs with large crystals. In addition to the surface morphology, we investigated the porosity of Fe3O4@MOFs, particularly relative to those of pristine MOFs. As observed in Fig. 3l, the total pore volume of core-shell materials increased to ⁓0.02–0.20 cm3/g, whereas a negligible porosity was detected for the acidic core itself. The difference in the porosity of Fe3O4@MOFs can likely be attributed to the different pore architectures of MOFs on the exterior shell side. Notably, a comparison of pore volume exhibited that Fe3O4@UiO-66 has the maximum pore volume of ⁓0.20 cm3/g among core–shell materials, highlighting the determining role of MOF type in the created porosity. Interestingly, we found that the modulated core–shell nanomaterials benefit from a hierarchical micro-mesoporous structure, while pristine MOFs are generally microporous materials (Supplementary Figs. 20–22). For example, the mesoporosity, i.e., mesopore divided by total pore volume, of the modulated UiO-66 shell (⁓85.4%) was much greater than that of pristine UiO-66 (⁓11.1%) synthesized by the conventional procedure. The created mesoporosity in the coating could be ascribed to the induced missing-linker defects during the modulated self-assembly of MOFs on the acidic core. Similar to the carboxylic acid linkers used for the synthesis of MOFs, Fe3O4–COOH can partially play the role of ligand, change the bonding energy of metal–ligand coordination and manipulate the MOF formation mechanism. Accordingly, the MOF layers cannot smoothly grow on the carboxylic acid-rich substrate which can create mesoporosity and engineer the pore structure by hindering the bridging linkers and changing the coordination environment during their modulated self-assembly. To confirm the role of missing-linker defects in the formation of engineered core–shell materials, thermogravimetric analysis was used (Fig. 3e and Supplementary Fig. 23). Because of the thermal decomposition of organic ligands from 100 to 350 °C, the weight reduction ratio of core–shell structure (ΔWC-S) to pristine MOF (ΔWP) in this temperature range was considered as an indicator of missing-linker deficiency23. Although Zn-, Co-, and Fe-based MOFs showed a high linker deficiency, the predominance of metal coordination through the network could have a negative impact on the self-assembly process, resulting in the poor formation of MOF structure (Fig. 3m). This is compatible with their low coating layer weights. In contrast, both Cu- and Zr-based MOFs displayed a good linker deficiency (⁓56.2–91.5%), as well as high coating weight (⁓30.1–34.7 wt%) and homogenous coating (Supplementary Figs. 24–27), likely due to the optimum surface energy of Fe3O4–COOH@Cu and Fe3O4–COOH@Zr compared to their corresponding organic ligands24–26. These results demonstrate that the self-assembly of MOFs on Fe3O4–COOH can be used as a simple platform to prepare advanced mesopore-induced core–shell materials with tailored properties, specifically more defects, and unsaturated metal sites. The formation of Fe3O4@MOF-SO4 Defect-engineered MOFs are promising platforms for developing advanced functional nanomaterials for various catalytic applications, including photo- and electrocatalysis, as they can provide active metal sites with a strong affinity towards a broad range of functional moieties12. We, therefore, with respect to our interest in acidic nanocatalysts, treated the engineered Fe3O4@MOFs with diluted sulfuric acid to introduce sulfate moieties through their defected structures. The general procedure for Fe3O4@MOF sulfation in the present study is schematically demonstrated in Fig. 4a and Supplementary Fig. 28. The primary advantage of the reported sulfation method is that Fe3O4@MOF-SO4 can be simply prepared by dispersing Fe3O4@MOFs into the aqueous solution of sulfuric acid (0.05 M, pH ⁓1.3) at room conditions which makes it a viable approach for large-scale implementation. FTIR analysis of Fe3O4@MOF-SO4 disclosed the presence of sulfur compounds in 800–1300 cm−1 region, including both S–O (⁓800–950 cm−1) and S=O (⁓1000–1300 cm−1) bonds, thus confirming that SO4−2 species were successfully coordinated with active metal sites (Fig. 4b, c and Supplementary Fig. 29)27. In addition, elemental line scanning profiles also revealed that there is a good distribution of sulfur across the treated core–shells, while no sulfur was detected before acid treatment (Fig. 4d, e). Since the sulfates can take different coordination positions on the surface of uncoordinated metal clusters, missing-linker deficiency in defect-engineered core–shells can positively manipulate the chelating mode of sulfate by providing additional space, resulting in the improved sulfation yield.Fig. 4 Coordination of chelating sulfate species in defected core–shell structure. a Schematic illustration of the preparation of water-dispersible Fe3O4@MOF–SO4 nanocatalyst via aqueous sulfation process. b, c FTIR spectra, and d, e sulfur line scanning profile of Fe3O4@UiO-66 and Fe3O4@UiO-66-SO4. TEM, HAADF, and EDX images of f–h Fe3O4@HKUST-SO4, i–k Fe3O4@UiO-66-SO4, and l–n Fe3O4@UiO-66-NH2–SO4. Scale bars are 50 (f, i, l), 10 (g, j, m), and 100 nm (h, k, n). High-resolution XPS spectra of sulfur species in o Fe3O4@HKUST-SO4, p Fe3O4@UiO-66-SO4, and q Fe3O4@UiO-66-NH2-SO4. To explore the versatility of the aqueous sulfation method for Fe3O4@MOF, Fe3O4@HKUST-1, Fe3O4@UiO-66, and Fe3O4@UiO-66-NH2 with the highest pore volumes were selected for post-treatment. TEM analysis illustrated the Fe3O4@MOF structures well preserved their core–shell structures at low pH values (i.e., ⁓1.3), with a shell thickness changing from ⁓10 nm in Fe3O4@HKUST-SO4 (Fig. 4f, g) to ⁓20 nm in Fe3O4@UiO-66–SO4 (Fig. 4i, j) and ⁓40 nm in Fe3O4@UiO-66-NH2–SO4 (Fig. 4l, m). HAADF and EDX maps reconfirmed the homogeneous distribution of metallic and organic elements corresponding to their MOF structures (Fig. 4h, k, n). Nevertheless, it was found that the amount of sulfur elements varies among Fe3O4@MOF–SO4 materials, and Fe3O4@HKUST–SO4 possesses the least sulfur content when compared with those of Zr-based core–shell structures. This observation was further examined by high-resolution XPS analysis, showing the sulfur species on the exterior surface of Fe3O4@MOF–SO4 materials (Fig. 4o–q and Supplementary Fig. 30). Notably, XPS peaks did not appear at 164–174 eV (the typical range of binding energy for sulfur, S 2p), validating the low sulfur signals in its EDX. These results suggest that the chemical stability of Fe3O4@MOF materials is a key contributor to the aqueous sulfation process. For instance, HKUST-1 framework is composed of Cu2+ ion pairs chelated by carboxylate bridges with paddle-wheel units; however, the high concentration of protons at low pH values can potentially accelerate the hydrolysis of Cu–O bonds, resulting in the partial disassembly of MOF crystals28. Similarly, Fe3O4@HKUST-1 could not entirely preserve its structural stability at harsh acidic conditions, as the shell side partially dissociated (Supplementary Fig. 31) and the average shell thickness remarkably diminished after the sulfation process. Furthermore, the unstable Cu bond could prohibit both bridging and chelating mode of sulfate coordination which verifies the poor efficiency of Fe3O4@HKUST-1 sulfation. In contrast, both Fe3O4@UiO-66–SO4 and Fe3O4@UiO-66-NH2–SO4 exhibited that sulfate species were successfully coordinated to the Zr metals, owing to the tolerance of Zr–O bond in a broad pH range from 1 to 1029. It is worth noting that good stability was also observed for Fe3O4–COOH nanoclusters at pH 1–3, whereas conventional Fe3O4 nanoparticles immediately digested into the acidic solution at the same conditions. This difference may be due to the citrate groups covering the surface of nanoparticles with lower pKa values (⁓3.0–5.5) than those of unfunctionalized Fe3O4 nanoparticles (⁓5.3–8.8), which could be deprotonated at low pH values and prevent the subsequent disassembly of Fe–O structure. Catalytic CO2 desorption performance Catalytic solvent regeneration, a recently emerged technique, has garnered wide attention, because of its low-temperature operation and high energy efficiency. The acidic nanocatalysts can act as a good proton donor, supplying the excess amount of protons required for the carbamate breakdown reaction, and promote CO2 desorption at temperatures less than 100 °C (the boiling temperature of water at atmospheric pressure)30. To illustrate that the prepared water-dispersible nanomaterials can be used to accelerate CO2 desorption reactions, we first examined the catalytic performance of acidic Fe3O4–COOH nanoclusters during the regeneration of CO2-rich monoethanolamine (MEA, 5 M) at 88 °C (Fig. 5a). As shown in Fig. 5b, adding a small amount of Fe3O4–COOH (0.1 wt%) significantly increased the kinetics of CO2 desorption, resulting in ⁓27.3% less energy consumption, when compared with that of blank MEA solution without using any catalyst. By increasing the amount of Fe3O4–COOH from 0.1 to 1 wt%, CO2 was released more quickly from the solvent; however, the energy efficiency parameter (i.e., the absolute of relative heat duty reduction versus the amount of catalyst used) significantly dropped from 2.73 to 0.66, respectively. Thus, a concentration of 0.1 wt% was established to be the optimal value of water-dispersible nanocatalysts, almost 10 times less than previously reported values, owing to the abundant active acidic sites on the surface of Fe3O4–COOH, as well as the Brownian motion and nanofluidic behavior of magnetic nanoclusters in the solvent31,32. These results highlight the potential of Fe3O4–COOH as a versatile substrate for the synthesis of acidic water-dispersible nanocatalysts which can be easily added at low concentrations during the continuous operation of CO2 absorption–desorption processes.Fig. 5 Application of water-dispersible nanocatalysts for energy-efficient CO2 capture. a Schematic of amine solvent regeneration with and without using a catalyst (Cat.). b CO2 desorption profile of amine solution with different concentrations of Fe3O4–COOH. c Relative heat duty (RH) and d cyclic CO2 uptake capacity of amine solution in the presence of various water-dispersible core–shell nanomaterials. e Schematic of catalytic CO2 desorption during thermal solvent regeneration. f The effect of time on the accumulative desorbed CO2 with and without using Fe3O4@UiO-66–SO4. g RH of amine solution without (a) and with heterogeneous catalysts (Al2O3 (b), V2O5 (c), H-Beta (d), HZSM-5 (e), SO42−/ZrO2/Al2O3 (f), SO42−/ZrO2/SBA-15 (g)), water-dispersible Fe3O4–COOH (h), and Fe3O4@UiO-66-SO4 (I). h Schematic of cyclic CO2 absorption–desorption of catalyst-aided solvent regeneration. i The effect of catalyst concentration on the cyclic CO2 capacity; blank (a), with H-Bata (b), Fe3O4–COOH (c), Fe3O4@UiO-66 (d), and Fe3O4@UiO-66-SO4 (e). j Recyclability of the Fe3O4@UiO-66–SO4 during consecutive CO2 absorption–desorption cycles. The amine solution in all cases is 5 M MEA in water and no catalyst was used for the blank solvent. The concentration of catalyst was fixed at 0.1 wt% in (c, d, f, g, j). Besides the CO2 desorption performance of Fe3O4–COOH, the catalytic behavior of defect-engineered Fe3O4@MOFs was explored. Figure 5c, d shows that all prepared core–shell materials improved the kinetics of CO2 desorption, ranging from Fe3O4@ZIF-8 to Fe3O4@HKUST-1 with the least and the best performance, respectively. From these observations, it appears that a MOF coating had a negative impact on the catalytic efficiency of Fe3O4–COOH substrates. Indeed, Fe3O4@HKUST-1 (as the best of core–shell materials) depicted a relative heat duty of 74.1%, whereas 72.7% was recorded when solely using Fe3O4–COOH. This could be attributed to the replacement of Brønsted acid sites (i.e., carboxylates) with Lewis acid sites (uncoordinated metal clusters) with less proton donation capability33. Contrarily, enriching Brønsted acid sites through the hierarchical structure of core–shell structures (i.e., Fe3O4@MOF–SO4) resulted in a distinct catalytic performance. The Fe3O4@UiO-66–SO4 succeeded to desorb 80.9% more CO2 compared with that of the blank solution at similar operating conditions. In addition, we found a decreasing trend in the differential desorbed CO2 with the regeneration time (from 24 to 36 min), highlighting the substantial influence of catalyst on the kinetics of CO2 desorption (Fig. 5e, f). To further investigate the catalytic performance of the Fe3O4@UiO-66–SO4, its corresponding relative heat duty was compared with those of Fe3O4–COOH and commercialized solid acid catalysts, including metal oxides (Al2O3, V2O5) and zeolites (H-Beta and HZSM-5) (Fig. 5g and Supplementary Fig. 32). The Fe3O4@UiO-66–SO4 exhibited the lowest required energy for the regeneration of CO2-rich MEA solution with a relative heat duty of 55.3%, mainly due to its large pore volume, predominant mesoporous structure, abundant Brønsted acid sites, and nanofluidic behavior (Supplementary Figs. 33 and 34). These findings also suggested that the performance of other prevalent heterogeneous nanocatalysts is not comparable with the water-dispersible materials at low catalyst concentrations (less than 0.1 wt%). To specifically explore the performance of water-dispersible nanomaterials at low concentrations (0.01, 0.05, and 0.1 wt%), their cyclic CO2 absorption–desorption capacity was measured (Fig. 5h, i). As discussed, H-Beta zeolite failed to display any sensible promotion throughout the course of the CO2 desorption operation. Unlike H-Beta, all Fe3O4–COOH, Fe3O4@UiO-66, and Fe3O4@UiO-66–SO4 demonstrated comparable cyclic CO2 capacity, even at an extremely low concentration of 0.01 wt%. For instance, adding Fe3O4@UiO-66–SO4 nanocatalysts increased the cyclic capacity of CO2 from 0.21 mol CO2/mol MEA in the blank solution to 0.30, 0.33, and 0.38 mol CO2/mol MEA with 0.01, 0.05, and 0.1 wt% concentrations of nanocatalyst, respectively, which are comparable with those of commercialized catalysts with ⁓10- to ⁓100-fold higher concentrations (⁓1.0–1.1 wt%)34. Since acidic catalyst allows for enhanced cumulative CO2 desorption during the solvent regeneration process, the solvent can absorb more CO2 in the next absorption cycle, leading to the better performance of the solvent (in terms of equilibrium and kinetics) in the absorption column. The stability of Fe3O4@UiO-66–SO4 was assessed via ten cycles of consecutive CO2 absorption–desorption operation (Fig. 5j). As seen, the relative heat duty of MEA solution increased by only ⁓9% over the first four cycles (region 1) and remained stable throughout the last six cycles (region 2). In addition, no significant changes were observed in the XPS spectra, XRD patterns, and SEM images of Fe3O4@UiO-66–SO4, before and after the cyclic operation (Supplementary Fig. 35), confirming the excellent recyclability of these water-dispersible nanocatalysts. Besides all targeted features of Fe3O4@UiO-66–SO4 for catalyst-aided solvent regeneration, its distinguished catalytic performance, particularly when compared with Fe3O4–COOH, could be assigned to its special proton donation mechanism. Throughout the solvent regeneration, CO2 molecules are generally released by the carbamate breakdown reaction. However, the yield of this reaction is highly dependent on the number of active protons in the reaction medium supplied by the amine deprotonation reaction (Fig. 6)35. Owing to the endothermic nature of all reactions, a high operating temperature (⁓120–140 °C) is required for spontaneous proton transfer and bond cleavage, resulting in high-quality steam use and subsequently high energy consumption36. In Fe3O4@UiO-66–SO4, the adsorbed water molecules on the surface of uncoordinated Zr clusters can participate in a hydrogen bond with a sulfate moiety chelated to another neighboring Zr center14. This specific arrangement of sulfate and water moieties results in the formation of superacid sites (H0 ≤ −14.5; see Supplementary Table 1) with distinct proton donation ability, accelerating the carbamate breakdown reaction and allowing for enhanced CO2 desorption at low regeneration temperatures (less than 100 °C). We note that all water-dispersible nanocatalysts could successfully recover their released protons during the CO2 absorption process, as compared to heterogeneous solid acid catalysts which need to be protonated via an acid washing processes. The results indicate that the unique privileges of water-dispersible nanomaterials (i.e., ease of use, low operating concentration, and high efficiency) can potentially make the implementation of catalytic solvent regeneration industrially affordable.Fig. 6 Catalytic CO2 desorption mechanism. a The acidic nanocatalyst enables an accelerated proton transfer from protonated amine (MEAH+) to carbamate (MEA-COO−), and facilitated the decomposition of the zwitterion (MEA-COOH). It resulted in enhanced CO2 desorption at low operating temperatures with the lower required energy. b The possible proton transfer routes from MEAH+ to MEA-COO− using the Brønsted acid sites of water-dispersible nanocatalyst. c The proposed carbamate breakdown mechanism in the presence of acidic nanocatalyst resulted in reusable MEA and CO2. In order to gain a greater understanding of the carbamate breakdown catalyzed by Fe3O4@UiO-66–SO4, density functional theory (DFT) calculations were performed. Starting from a defect active site with HSO4− and H2O coordinating the two Zr atoms, the addition of carbamate resulted in the displacement of the H2O by the carbamate carboxylate group (Fig. 7a). This configuration places the HSO4− hydroxyl oxygen within ~3 Å of the carbamate nitrogen. While the exact order of proton transfer is unclear, the protonation of the carbamate nitrogen by either the HSO4− or the nearby H2O resulted in the breakage of the N–C bond, coupled with the formation of a C–O bond from the H2O, leading to stable products of MEA and HCO3− (Fig. 7b). Attempts to protonate the carbamate nitrogen and break the C–N bond in the absence of the water molecule were unsuccessful, suggesting that the presence of H2O and the formation of HCO3− as a product are essential in the catalytic reaction.Fig. 7 Reactants and products for the decomposition of carbamate on a Fe3O4@UiO-66–SO4 defect active site. a Reactants. b Products. The full fragment is shown in the upper panel, and a rotated view of the defect active site alone is shown in the lower panel. Zirconium, sulfur, oxygen, nitrogen, carbon, and hydrogen are colored cyan, yellow, red, blue, gray, and white, respectively. Discussion In summary, we have developed a new and effective methodology to synthesize a series of water-dispersible nanocatalysts with engineered properties and nanofluidic behavior for low-temperature CO2 desorption. A promising feature of the current method is that the carboxylate-rich surface of Fe3O4–COOH nanoclusters allows for the modulated growth of MOFs on its structure. Importantly, we observed a transition from the microporous structure in pristine MOFs to a hierarchical micro-mesoporous network in Fe3O4@MOFs. Different metal ion and organic ligand combinations can be selected for the tailored self-assembly of MOFs on the magnetic core, resulting in a broad diversity of physicochemical properties. The obtained materials with structural defects and active metal centers can serve as a potential platform for the coordination of chelating sulfates with superacidity. Enhanced kinetics of CO2 desorption was obtained for all Fe3O4–COOH, Fe3O4@MOF, and Fe3O4@MOF–SO4 materials. As a proof of concept, Fe3O4@UiO-66-SO4 exhibited the best catalytic performance (44.7% reduction in energy consumption) with excellent recyclability (⁓9% loss over five cycles) using a low catalyst concentration of only 0.1 wt%. The comprehensive characterizations revealed that the missing-linker defects and superacidic sites through the hierarchical network of Fe3O4@UiO-66–SO4 are key contributors to its advanced proton donation capability and thus accelerated CO2 desorption performance. We anticipate that this approach will open up a new avenue to the synthesis of water-dispersible nanocatalysts in the area of catalytic CO2 absorption–desorption and take one further step toward the utilization of green energy (e.g., solar hot water) in the portfolio of energy-efficient CO2 capture. Methods Synthesis of acidic Fe3O4–COOH nanoclusters Typically, ferric ammonium citrate (3.25 g) and NaOAc (6 g) were dissolved in 100 mL EG under vigorous stirring at room temperature. The formed hazy solution was transferred to a stainless-steel autoclave (150 mL capacity) and heated at 200 °C for 10 h. Then, the autoclave was gradually cooled to room temperature and the black solid precipitants were separated via the application of an external magnet. The obtained nanoclusters were repeatedly washed with acetone, ethanol, and water to remove unreacted or physically attached molecules. During each washing step, the nanoclusters were suspended in the solvent, ultrasonicated for 30 min, and separated by an external magnet. Finally, the resultant materials were dried at 80 °C overnight under a severe vacuum and labeled as Fe3O4–COOH. Synthesis of conventional Fe3O4 nanoparticles To compare the structural properties of Fe3O4–COOH nanoclusters with a benchmark, conventional Fe3O4 nanoparticles were synthesized according to the previously reported co-precipitation method with minor modifications37. Firstly, 298.5 mg of FeCl2·4H2O and 810 mg of FeCl3·6H2O were dissolved in 100 mL of water and sonicated for 15 min. The temperature was increased to 60 °C, and while keeping the mixture under vigorous agitation, NH4OH solution was added as an oxidation agent to trigger the iron oxide precipitation. When the pH of the solvent reached 11, the temperature was increased to 80 °C and held for 2 h. The black precipitates were separated by an external magnet and washed with a copious amount of water to remove extra NH4OH reactants and OH− ions. Eventually, the obtained products were dried in a vacuum oven at 80 °C overnight, labeled as Fe3O4 nanoparticles and stored for future use. Synthesis of Fe3O4@ZIF-8 For the synthesis of Fe3O4@ZIF-8 core–shell particles with a magnetic core, Fe3O4–COOH nanoclusters were first dispersed in 100 mL methanol (5 mg/mL) under vigorous stirring for 30 min. Then, 0.325 g of Zn(NO3)2·6H2O was added and the suspension was sonicated for 1 h to facilitate zinc metal ion coordination on the carboxylate groups of the Fe3O4–COOH cluster surface. For the self-assembly of the ZIF-8 shell, 100 mL of 2-MeIm solution (2.82 mg/mL in methanol) was added dropwise to the suspension and the mixture was stirred at room temperature for 12 h. The fabricated Fe3O4@ZIF-8 core–shell particles were separated by an external magnet. For the washing process, Fe3O4@ZIF-8 particles were sonicated in methanol for 15 min, while the solvent was replaced three times. The particles produced were vacuum dried at 120 °C overnight and labeled as Fe3O4@ZIF-8. Synthesis of Fe3O4@ZIF-67 For the synthesis of Fe3O4@ZIF-67 core–shell particles with a magnetic core, Fe3O4–COOH nanoclusters were first dispersed in 100 mL methanol (50 mg/mL) under vigorous stirring for 30 min. Then, 0.5 g of Co(NO3)2·6H2O was added, and the suspension was sonicated for 1 h to facilitate cobalt metal ion coordination on the carboxylate groups of the Fe3O4–COOH cluster surface. For the self-assembly of the ZIF-67 shell, 100 mL of 2-MeIm solution (5.6 mg/mL in methanol) was added dropwise to the suspension and the mixture was stirred at room temperature for 24 h. The fabricated Fe3O4@ZIF-67 core–shell particles were separated by an external magnet. For the washing process, Fe3O4@ZIF-67 particles were sonicated in methanol for 15 min, and the solvent was replaced three times. The particles were vacuum dried at 120 °C overnight and labeled as Fe3O4@ZIF-67. Synthesis of Fe3O4@MIL-100(Fe) For the synthesis of Fe3O4@MIL-100(Fe) core–shell particles with a magnetic core, Fe3O4-COOH nanoclusters (5 mg/mL) were first dispersed in 100 mL of ethanol under vigorous stirring for 30 min. Then, 0.464 g of FeCl3·6H2O was added and the suspension was sonicated for 1 h to facilitate iron metal ion coordination on the carboxylate groups of the Fe3O4-COOH cluster surface. The obtained suspension was transferred to a 250 mL round bottom flask and heated at 70 °C using an external oil bath with a magnetic stirrer. After 1 h, 100 mL of H3BTC solution (3.61 mg/mL in water) was added dropwise to the suspension and the mixture was stirred at 70 °C for a further 6 h to homogeneously grow the MIL-100(Fe) shell. The fabricated Fe3O4@MIL-100(Fe) core–shell particles were naturally cooled to room temperature and separated by an external magnet. For the washing process, Fe3O4@MIL-100(Fe) particles were sonicated three times in ethanol for 15 min. The particles were vacuum dried at 110 °C overnight and labeled as Fe3O4@MIL-100(Fe). Synthesis of Fe3O4@MOF–Fe(II) For the synthesis of Fe3O4@MOF–Fe(II) core–shell particles with a magnetic core, Fe3O4–COOH nanoclusters were first dispersed in 50 mL water (10 mg/mL) under vigorous stirring for 30 min. Then, 0.278 g of FeSO4 was added and the suspension was sonicated for 1 h to facilitate iron metal ion coordination on the carboxylate groups of the Fe3O4–COOH cluster surface. The obtained suspension was transferred to a 250 mL round bottom flask and heated at 130 °C using an external oil bath with a magnetic stirrer. After 1 h, 150 mL of H2BDC-N solution (1.67 mg/mL in DMF) was added dropwise to the suspension and the mixture was stirred at 130 °C for a further 4 h to homogeneously grow the MOF-Fe(II) shell. The fabricated Fe3O4@MOF–Fe(II) core–shell particles were naturally cooled to room temperature and separated by an external magnet. For the washing process, Fe3O4@MOF–Fe(II) particles were sonicated three times in DMF, water, and ethanol for 15 min. The particles were vacuum dried at 110 °C overnight and labeled as Fe3O4@MOF–Fe(II). Synthesis of Fe3O4@HKUST-1 For the synthesis of Fe3O4@HKUST-1 core-shell particles with a magnetic core, Fe3O4–COOH nanoclusters were first dispersed in 100 mL of ethanol (5 mg/mL) under vigorous stirring for 30 min. Then, 0.343 g of Cu(NO3)2·3H2O was added and the suspension was sonicated for 1 h to facilitate copper metal ion coordination on the carboxylate groups of the Fe3O4–COOH cluster surface. The obtained solution was transferred to a 250 mL round bottom flask and heated at 85 °C using an external oil bath with a magnetic stirrer. After 1 h, 100 mL of H3BTC solution (3.61 mg/mL in ethanol) was added dropwise to the suspension and the mixture was stirred at 85 °C for a further 24 h to homogeneously grow the HKUST-1 shell. The fabricated Fe3O4@HKUST-1 core–shell particles were naturally cooled to room temperature and separated by an external magnet. For the washing process, Fe3O4@HKUST-1 particles were sonicated three times in ethanol and dichloromethane for 15 min. The particles were vacuum dried at 120 °C overnight and labeled as Fe3O4@HKUST-1. Synthesis of Fe3O4@UiO-66 For the synthesis of Fe3O4@UiO-66 core–shell particles with a magnetic core, Fe3O4–COOH nanoclusters were first dispersed in 100 mL of DMF (5 mg/mL) under vigorous stirring for 30 min. Then, 0.64 g of ZrCl4 was added and the suspension was sonicated for 1 h to facilitate zirconium metal ion coordination on the carboxylate groups of the Fe3O4–COOH cluster surface. The obtained solution was transferred to a 250 mL round bottom flask, mixed with 2 mL of AcOH, and heated at 120 °C using an external oil bath with a magnetic stirrer. After 1 h, 100 mL of H2BDC solution (4.56 mg/mL in DMF) was added dropwise to the suspension and the mixture was stirred at 120 °C for a further 24 h to homogeneously grow the UiO-66 shell. The fabricated Fe3O4@UiO-66 core–shell particles were naturally cooled to room temperature and separated by an external magnet. For the washing process, Fe3O4@UiO-66 particles were sonicated three times in hot DMF, water, and ethanol for 15 min. The particles were vacuum dried at 110 °C overnight and labeled as Fe3O4@UiO-66. Synthesis of Fe3O4@UiO-66–NH2 For the synthesis of Fe3O4@UiO-66–NH2 core-shell particles with a magnetic core, Fe3O4–COOH nanoclusters were first dispersed in 100 mL of DMF (5 mg/mL) under vigorous stirring for 30 min. Then, 0.64 g of ZrCl4 was added and the suspension was sonicated for 1 h to facilitate zirconium metal ion coordination on the carboxylate groups of the Fe3O4–COOH cluster surface. The obtained solution was transferred to a 250 mL round bottom flask, mixed with 2 mL of AcOH, and heated at 120 °C using an external oil bath with a magnetic stirrer. After 1 h, 100 mL of H2BDC solution (4.97 mg/mL in DMF) was added dropwise to the suspension and the mixture was stirred at 120 °C for a further 24 h to homogeneously grow the UiO-66–NH2 shell. The fabricated Fe3O4@UiO-66–NH2 core–shell particles were naturally cooled to room temperature and separated by an external magnet. For the washing process, Fe3O4@UiO-66–NH2 particles were sonicated three times in hot DMF, water, and ethanol for 15 min. The particles were vacuum dried at 110 °C overnight and labeled as Fe3O4@UiO-66–NH2. Synthesis of Fe3O4@MOF–SO4 nanomaterials The pre-prepared core-shell materials were used to synthesize Fe3O4@MOF–SO4 nanocatalysts. Typically, 1 g of Fe3O4@MOF was dispersed in 500 mL of 0.05 M aqueous H2SO4 solution. After sonicating for 15 min, the solution was gently stirred for 24 h at room temperature. The nanocatalysts were washed three times to remove excess H2SO4 molecules. In each washing step, Fe3O4@MOF–SO4 were magnetically separated, dispersed in 250 mL ultra-pure hot water (ca. 60 °C), and sonicated for 15 min, followed by magnetic separation and supernatant removal. Finally, the products were dried using a vacuum oven at 150 °C for 48 h and stored for future use. Theoretical calculations The initial structure of UiO-66 was taken from Cavka et al.38, and truncated and protonated based on the work of Sittiwong et al.39. This structure comprises four 1,4-benzene-dicarboxylate (BDC) linkers surrounding the defect active site on the Zr6O4(OH)4 node created by removing one linker, and seven linkers truncated to formate, for computational efficiency. To balance the charge of the system, hydrogens were added to the terminal carboxylates of BDC, and hydrogen was removed from the bridging OH of the defect active site to compensate for the removal of the BDC linker (Supplementary Fig. 3a). Following initial geometry optimization, a water molecule was added to the active site. The protonation of bridging oxygen and coordination of a hydroxyl group to a Zr atom was found to be more energetically favorable than an intact H2O by ~0.26 eV (Supplementary Fig. 3b). To model the effects of sulfation, an H2SO4 molecule was added to the hydroxylated defect active site, which spontaneously reacted with the adjacent hydroxyl group coordinated to the Zr to form HSO4- coordinated to one Zr, and H2O coordinated to the other Zr, during geometry optimization (Supplementary Fig. 3c). All DFT calculations were performed using Gaussian 16 Revision C.0140. The 6-31G(d,p) basis set was employed for all atoms except Zr, which was treated using the double-ζ of the Stuttgart–Dresden pseudopotential41,42 from the Basis Set Exchange43. The M06-L DFT functional7 was used for all calculations. For all geometry optimizations, the formate hydrogens and terminal carboxyl carbons were frozen to maintain the overall structure of UiO-66, while all other atoms were allowed to relax. This methodology has been validated previously for catalytic reactions on UiO-6639. Supplementary information Supplementary Information Supplementary information The online version contains supplementary material available at 10.1038/s41467-022-28869-6. Acknowledgements This research was conducted in part at the Advanced Separation Technologies, Bio21 Advanced Microscopy Facility, and Materials Characterization and Fabrication Platform (MCFP) the Bio21 Advanced Microscopy Facility at the University of Melbourne. We especially thank Dr. Alex Duan for his technical support and data analysis on XPS measurements performed at Melbourne TrACEES Platform (Trace Analysis for Chemical, Earth, and Environmental Sciences). The authors also acknowledge Dr. Anders Barlow for his expert advice on HIM microscopy conducted in MCFP. This research was undertaken with the assistance of supercomputing resources from the National Computational Infrastructure (NCI), which is supported by the Australian Government, under the National Computational Merit Allocation Scheme. The assistance of the Pawsey Supercomputer Centre is also acknowledged. Author contributions M.S.A. and K.A.M. conceptualized the work and designed the synthesis methodology. M.S.A. conducted all the experiments and characterizations with help of A.Z. and O.M; M.S.A. and Y.W. performed the CO2 absorption-desorption experiments. A.J.C., N.M., and S.P.R. conducted the DFT simulations. M.S.A., K.A.M, G.W.S., and C.A.S. contributed to the data analysis and scientific discussions. All authors revised the paper and provided helpful comments. Peer review Peer review information Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. Data availability The data that support the findings of this study are available from the corresponding authors upon reasonable request. Competing interests The authors declare no competing interests. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Samset BH Fuglestvedt JS Lund MT Delayed emergence of a global temperature response after emission mitigation Nat. Commun. 2020 11 3261 10.1038/s41467-020-17001-1 32636367 2. Peters GP Carbon dioxide emissions continue to grow amidst slowly emerging climate policies Nat. Clim. Change 2020 10 3 6 10.1038/s41558-019-0659-6 3. Pauw P Mbeva K van Asselt H Subtle differentiation of countries’ responsibilities under the Paris Agreement Palgrave Commun. 2019 5 86 10.1057/s41599-019-0298-6 4. Wei Y-M Self-preservation strategy for approaching global warming targets in the post-Paris Agreement era Nat. Commun. 2020 11 1624 10.1038/s41467-020-15453-z 32286257 5. Realmonte G An inter-model assessment of the role of direct air capture in deep mitigation pathways Nat. Commun. 2019 10 3277 10.1038/s41467-019-10842-5 31332176 6. Wei Y-M A proposed global layout of carbon capture and storage in line with a 2 °C climate target Nat. Clim. Change 2021 11 112 118 10.1038/s41558-020-00960-0 7. Lai Q Catalyst-TiO(OH)2 could drastically reduce the energy consumption of CO2 capture Nat. Commun. 2018 9 2672 10.1038/s41467-018-05145-0 29991689 8. Xing L One-step synthesized SO42–/ZrO2-HZSM-5 solid acid catalyst for carbamate decomposition in CO2 capture Environ. Sci. Technol. 2020 54 13944 13952 10.1021/acs.est.0c04946 33054187 9. Zhang X Reducing energy penalty of CO2 capture using Fe promoted SO42–/ZrO2/MCM-41 catalyst Environ. Sci. Technol. 2019 53 6094 6102 10.1021/acs.est.9b01901 31008586 10. Bhatti UH Nam S Park S Baek IH Performance and mechanism of metal oxide catalyst-aided amine solvent regeneration ACS Sustain. Chem. Eng. 2018 6 12079 12087 10.1021/acssuschemeng.8b02422 11. Alivand MS Catalytic solvent regeneration for energy-efficient CO2 capture ACS Sustain. Chem. Eng. 2020 8 18755 18788 10.1021/acssuschemeng.0c07066 12. Wang Q Astruc D State of the art and prospects in metal–organic framework (MOF)-based and MOF-derived nanocatalysis Chem. Rev. 2020 120 1438 1511 10.1021/acs.chemrev.9b00223 31246430 13. Wei Y-S Zhang M Zou R Xu Q Metal–organic framework-based catalysts with single metal sites Chem. Rev. 2020 120 12089 12174 10.1021/acs.chemrev.9b00757 32356657 14. Trickett CA Identification of the strong Brønsted acid site in a metal–organic framework solid acid catalyst Nat. Chem. 2019 11 170 176 10.1038/s41557-018-0171-z 30455431 15. Jiang J Superacidity in sulfated metal–organic framework-808 J. Am. Chem. Soc. 2014 136 12844 12847 10.1021/ja507119n 25157587 16. Wang H Topologically guided tuning of Zr-MOF pore structures for highly selective separation of C6 alkane isomers Nat. Commun. 2018 9 1745 10.1038/s41467-018-04152-5 29717138 17. Feng D A highly stable zeotype mesoporous zirconium metal–organic framework with ultralarge pores Angew. Chem. Int. Ed. 2015 54 149 154 10.1002/anie.201409334 18. Kim Y Hydrolytic transformation of microporous metal–organic frameworks to hierarchical micro- and mesoporous MOFs Angew. Chem. Int. Ed. 2015 54 13273 13278 10.1002/anie.201506391 19. Wang L Deliberate modification of Fe3O4 anode surface chemistry: impact on electrochemistry ACS Appl. Mater. Interfaces 2019 11 19920 19932 10.1021/acsami.8b21273 31042346 20. Liu J Highly water-dispersible biocompatible magnetite particles with low cytotoxicity stabilized by citrate groups Angew. Chem. Int. Ed. 2009 48 5875 5879 10.1002/anie.200901566 21. Jang J Janus graphene oxide sheets with Fe3O4 nanoparticles and polydopamine as anodes for lithium-ion batteries ACS Appl. Mater. Interfaces 2021 13 14786 14795 10.1021/acsami.1c02892 33739082 22. Alivand MS Development of aqueous-based phase change amino acid solvents for energy-efficient CO2 capture: the role of antisolvent Appl. Energy 2019 256 113911 10.1016/j.apenergy.2019.113911 23. He T Fast and scalable synthesis of uniform zirconium-, hafnium-based metal–organic framework nanocrystals Nanoscale 2017 9 19209 19215 10.1039/C7NR06274E 29188246 24. De Luna P Metal–organic framework thin films on high-curvature nanostructures toward tandem electrocatalysis ACS Appl. Mater. Interfaces 2018 10 31225 31232 10.1021/acsami.8b04848 30129364 25. Gao F-Y High-curvature transition-metal chalcogenide nanostructures with a pronounced proximity effect enable fast and selective CO2 electroreduction Angew. Chem. Int. Ed. 2020 59 8706 8712 10.1002/anie.201912348 26. Liu X Photoconductivity in metal–organic framework (MOF) thin films Angew. Chem. Int. Ed. 2019 58 9590 9595 10.1002/anie.201904475 27. Peng W-L Accelerating biodiesel catalytic production by confined activation of methanol over high-concentration ionic liquid-grafted UiO-66 solid superacids ACS Catal. 2020 10 11848 11856 10.1021/acscatal.0c03261 28. Álvarez JR Structure stability of HKUST-1 towards water and ethanol and their effect on its CO2 capture properties Dalton Trans. 2017 46 9192 9200 10.1039/C7DT01845B 28678240 29. Wang C Liu X Chen JP Li K Superior removal of arsenic from water with zirconium metal-organic framework UiO-66 Sci. Rep. 2015 5 16613 10.1038/srep16613 26559001 30. Bhatti UH Practical and inexpensive acid-activated montmorillonite catalysts for energy-efficient CO2 capture Green. Chem. 2020 22 6328 6333 10.1039/D0GC01887B 31. Yu W Wang T Park A-HA Fang M Review of liquid nano-absorbents for enhanced CO2 capture Nanoscale 2019 11 17137 17156 10.1039/C9NR05089B 31517369 32. Elhambakhsh A Keshavarz P Investigation of carbon dioxide absorption using different functionalized Fe3O4 magnetic nanoparticles Energy Fuels 2020 34 7198 7208 10.1021/acs.energyfuels.0c00234 33. Bhatti UH Effects of transition metal oxide catalysts on MEA solvent regeneration for the post-combustion carbon capture process ACS Sustain. Chem. Eng. 2017 5 5862 5868 10.1021/acssuschemeng.7b00604 34. Bairq ZAS Gao H Murshed FAM Tontiwachwuthikul P Liang Z Modified heterogeneous catalyst-aided regeneration of CO2 capture amines: a promising perspective for a drastic reduction in energy consumption ACS Sustain. Chem. Eng. 2020 8 9526 9536 10.1021/acssuschemeng.0c02582 35. Lv B Guo B Zhou Z Jing G Mechanisms of CO2 capture into monoethanolamine solution with different CO2 loading during the absorption/desorption processes Environ. Sci. Technol. 2015 49 10728 10735 10.1021/acs.est.5b02356 26236921 36. Bhatti UH Sivanesan D Nam S Park SY Baek IH Efficient Ag2O–Ag2CO3 catalytic cycle and its role in minimizing the energy requirement of amine solvent regeneration for CO2 capture ACS Sustain. Chem. Eng. 2019 7 10234 10240 10.1021/acssuschemeng.9b01709 37. Dheyab MA Simple rapid stabilization method through citric acid modification for magnetite nanoparticles Sci. Rep. 2020 10 10793 10.1038/s41598-020-67869-8 32612098 38. Cavka JH A new zirconium inorganic building brick forming metal organic frameworks with exceptional stability J. Am. Chem. Soc. 2008 130 13850 13851 10.1021/ja8057953 18817383 39. Sittiwong J Density functional investigation of the conversion of furfural to furfuryl alcohol by reaction with i-propanol over UiO-66 metal–organic framework Inorg. Chem. 2021 60 4860 4868 10.1021/acs.inorgchem.0c03764 33764784 40. Gaussian 16 Rev. C.01 (Wallingford, CT, 2016). 41. Andrae D Häußermann U Dolg M Stoll H Preuß H Energy-adjustedab initio pseudopotentials for the second and third row transition elements Theor. Chim. Acta 1990 77 123 141 10.1007/BF01114537 42. Martin JML Sundermann A Correlation consistent valence basis sets for use with the Stuttgart–Dresden–Bonn relativistic effective core potentials: the atoms Ga–Kr and In–Xe J. Chem. Phys. 2001 114 3408 3420 10.1063/1.1337864 43. Pritchard BP Altarawy D Didier B Gibson TD Windus TL New basis set exchange: an open, up-to-date resource for the molecular sciences community J. Chem. Inf. Model. 2019 59 4814 4820 10.1021/acs.jcim.9b00725 31600445
PMC009xxxxxx/PMC9054240.txt
==== Front RSC Adv RSC Adv RA RSCACL RSC Advances 2046-2069 The Royal Society of Chemistry 35517722 d0ra03220d 10.1039/d0ra03220d Chemistry Ionic liquids of superior thermal stability. Validation of PPh4+ as an organic cation of impressive thermodynamic durability† † Electronic supplementary information (ESI) available: Representative synthetic details; multi-nuclear NMR spectra; ESI-MS. See DOI: 10.1039/d0ra03220d Soltani Mohammad a McGeehee Jimmie L. a Stenson Alexandra C. a O'Brien Richard A. a https://orcid.org/0000-0002-2839-3227 Duranty Edward R. a Salter E. Alan a Wierzbicki Andrzej a Glover T. Grant a https://orcid.org/0000-0001-9573-3597 Davis James H. Jr a a Department of Chemistry, University of South Alabama Mobile Alabama 36688 USA [email protected] 29 5 2020 27 5 2020 29 5 2020 10 35 2052120528 13 4 2020 21 5 2020 This journal is © The Royal Society of Chemistry 2020 The Royal Society of Chemistry https://creativecommons.org/licenses/by/3.0/ Recent work by Wasserscheid, et al. suggests that PPh4+ is an organic molecular ion of truly exceptional thermal stability. Herein we provide data that cements that conclusion: specifically, we show that aliphatic moieties of modified PPh4+-based cations incorporating methyl, methylene, or methine C–H bonds burn away at high temperatures in the presence of oxygen, forming CO, CO2, and water, leaving behind the parent ion PPh4+. The latter then undergoes no further reaction, at least below 425 °C. When appended to the tetraphenylphosphonium cation, organic groups containing aliphatic C–H bonds are burned away at high temperatures in the presence of O2. However, the parent cation remains unscathed, demonstrating its remarkable thermal stability. National Science Foundation 10.13039/100000001 CHE-1800122 pubstatusPaginated Article ==== Body pmcSince 2013 we have sought to design, prepare, and study ionic liquids that have degrees of thermal stability superior to those exhibited by commonplace imidazolium, quaternary ammonium, and tetraalkylphosphonium-type ILs.1–8 And, while high thermal stability is important to ILs slated to be used as solvents in synthetic applications,9 their development is especially important to their practical utility as additives to high-performance polymers, advanced lubricants, and as liquid- or ‘liquifiable’ materials for energy-related uses such as energy storage via phase change and heat transfer fluids in solar thermal arrays.10 We note that the latter applications are those in which we especially envision future practical roles for the present class of ILs. This new family of ionic liquids of improved thermal stability is formulated around perarylated cations (cations commonly being the thermal weak link of organic salts),11 and the “success stories” among them (typified by compounds 1 and 2, Fig. 1) are routinely stable for months (perhaps indefinitely), in air, at temperatures of 300 °C–400 °C; as a consequence they rank among the most thermally stable organic materials heretofore described.7 Furthermore, they constitute a class of salts which bridge a melting point/thermal stability (“mesothermal”) gap that arguably exists between classical, all-inorganic molten salts and compounds more widely referred to as ionic liquids. The basis for the superior thermal stability of these perarylated cations is thought to be rooted in their relative invulnerability to retro-SN2 or Hoffman elimination reactions, these being the principal mechanistic pathways by which more widely studied IL-cations are known to thermally decompose.7,11 Fig. 1 Ionic liquids previously established to be thermally stable in air for periods of 90 days and greater at T ≥ 350 °C. However, even from our first study it was clear that certain modifications to these otherwise thermally stable peraryl onium cations led to changes in this attribute compared to the parent ion. Notably, those cations which are modified with aliphatic moieties are the most impacted, in-line with what is known about the relative thermal reactivity of aliphatic [C(sp3)–H] versus aromatic [C(sp2)–H] C–H bonds.12–17 Once melted, as the temperature to which the water-white liquids are subjected continues to be increased, their colour changes to brown, clear evidence of heat-induced changes to the structures and/or composition of the salts. At the same time, however, these changes are usually accompanied by only modest mass losses. In this light, it seems clear that the observed thermal ‘decomposition’ process likely produces two classes of product: volatile substances of low-molecular weight that account for the observed mass losses, and non- or less-volatile materials, likely those remaining ionic, that account for the residual mass. Their curiosity piqued by a progress report incorporating the former observations, the reviewers of the renewal application for the grant which has provided funding to conduct work on thermally stable ILs expressed a strong interest in having us seek insights into how/why this introduction of aliphatic components into the cations diminishes their thermal stability; interest was also expressed in our gaining insight into the nature of the decomposition products. Concurring that acquiring this information would be potentially useful, we undertook the present study. Here we have synthesized and investigated the pyrolysis of a number of tetraarylphosphonium Tf2N salts, the cations of 3, 6, 12–14, 16, 18, and 20 each having an arene-ring appendage containing an aliphatic C–H bond that is part of a methyl, methylene, or methine moiety. The structures of these salts, all but one of which (3) is new, are illustrated in Fig. 2. In every case, the Tf2N− anion was employed. Our decision to employ it was driven by the established nature of its high intrinsic thermal stability, and its demonstrated lack of reactivity (versus several other ions) towards tetraarylphosphonium cations.2,4 Fig. 2 New mesothermal ionic liquids prepared for the present study. The Tf2N− anion is common to all. As in all our prior work, these were prepared using a modification of a synthetic methodology used by Charette and co-workers.18 In brief, this consists of the dissolution of equimolar amounts of triphenylphosphine and an appropriately substituted aryl bromide, chloride, or iodide in ethylene glycol, to which has been added ca. 5 mol% NiCl2·6H2O. The mixture is sealed in a thick-walled, screw-capped tube, in which is a magnetic stir bar. The tube is immersed in an oil bath maintained at 180 °C, then heated with stirring for 2 hours. Upon cooling, the solution is added to a two-phase system consisting of water and dichloromethane, the catalyst and ethylene glycol partitioning into the water, and the desired tetraarylphosphonium salt into the dichloromethane. The organic phase is separated, dried over anhydrous MgSO4, and the solvent removed to leave a white solid. Replacement of the bromide anion of this intermediate salt is accomplished by dissolving it in hot water, then adding to the stirred solution a slight (ca. 5%) molar excess of solid KTf2N. The rapid dissolution of the latter in the aqueous phase is quickly followed by the precipitation of the final phosphonium Tf2N salt, usually as a white solid. On occasion, the product separates as an oil which solidifies over the course of a few hours. Yields of both the intermediate halide salts and the final products are commonly quantitative. Thermochemical evaluations were carried out in the following fashions. First, gradient and isothermal thermogravimetric analyses (TGA) were conducted in air using a TGA 550 with quartz furnace, air being used to maximize chances for compound decomposition/oxidation. During the gradient measurements, T was increased at a rate of 2 °C per minute from ambient to 500 °C. Isothermal experiments were carried out by ramping to the target hold temperature – 200 °C, 250 °C, or 300 °C, at a rate of 5 °C per minute. Once reached, the hold temperature was maintained for 10 hours, after which time the system T was ramped to 500 °C at a rate of 5 °C per minute. TGA traces are supplied in the ESI.† In separate experiments, samples of each IL were heated in air to 200 °C, 250 °C, and 300 °C in a Thermolyne FB-1315M furnace for a 96 hour period, with each sample then being analysed post-heating by 1H-, 13C-, 31P-, and 19F-NMR (JEOL ECA 500), as well as by electrospray ionization mass spectra (ThermoFisher LTQ-Velos). This particular thermoanalytic approach – developed specifically to subject ILs to harsh, ‘real-world’ conditions – was conceived and its utility validated by Fox et al. at Savannah River National Laboratory, before being employed by our group.19 Finally, to complement both of the former experimental protocols, samples were also aerobically pyrolyzed under air in a sealed IR gas cell in order to capture and analyse the volatile components of the thermo-oxidative breakdown (ThermoFisher Nicolet is50 with MCTa detector). The TGA data proved to be rather unrevealing in the sense that they betrayed no clear-cut relationship between mass changes at/in any given temperature or temperature range and assignable chemical changes in the salts, unlike the relationships that proved to be understandable by NMR, ESI-MS, and FTIR in conjunction with the furnace experiments (vide infra). In retrospect, this not particularly surprising, since concurrent volatile-fragment elimination (mass loss) and incorporation of oxygen (mass gain) into the still-ionic (hence non-volatile) perarylphosphonium cation core is likely to have occurred (vide infra). Such behaviour, while infrequently observed with ordinary, wholly covalent organic materials nevertheless has precedents, especially in association with intrinsically non-volatile and/or non-combustible materials.20 Such a course of events, depicted conjecturally for compound 3 (Scheme 1), does not appear to be commonly considered in TGA studies of ILs, probably because the experiments are routinely conducted in inert nitrogen or argon atmospheres rather than in a reactive atmosphere of air or O2. Although the data generated in the former instance is unarguably valuable within its context, it should be noted that those TGA outcomes may be of little utility in anticipating the thermal stability of an IL in an eventual application in which it is deliberately or incidentally exposed to air while operating at high temperatures. Making the latter such determinations is a central theme of our work. Scheme 1 Overlapping mass gain and loss from different chemical reactions occurring during aerobic thermolysis can frustrate the use of TGA to connect specific chemical changes to particular temperatures.18 As described previously, each material was also heated in air (porcelain crucible) to 200 °C, 250 °C, and 300 °C in a furnace for a 96 hour period, followed by residue analysis using 1H-, 13C-, 31P-, and 19F-NMR, as well as ESI-MS. From these experiments, it was plainly apparent that the condensed-phase pyrolysis products contained no fluorine other than that contained in the Tf2N− anion; the 19F spectrum of all compounds was invariant over time, consistent with there being no heat-induced changes to this anion at 200 °C, 250 °C or 300 °C. This fully comports with experiments by Wasserscheid, et al. establishing that the decomposition temperature of Tf2N− is ≥420 °C.2 In contrast, there were routinely specific, clearly discernible changes observed in the 1H-, 13C- and ESI mass spectra of the post-heating condensed phases, and by using FTIR (vide infra) gaseous off-products were readily identified. By-in-large, the data emergent from these analyses was such that the identities of the major decomposition product(s) could be assigned. The aerobic, heat-induced transformations of the p-tolyl salt 3 is typical of those observed to occur with most of the present alkyl-bearing tetraarylphosphonium salts (Scheme 2). Consequently, a detailed discussion of these will be a useful primer for grasping the changes observed with the other salts as well. Scheme 2 Aerobic, pyrolytic conversion of 3 into 1, and a proposed mechanism for the process. Over the lifetime of the pyrolysis, the 3 : 4 relative intensity of the tolyl CH3 group in the 1H-NMR spectrum versus that of the tolyl aromatic hydrogens progressively diminished. These changes were easily identified as being associated with the substituted phenyl ring since samples of 3 were also prepared using d15-PPh3 to suppress overlapping signals from aromatic protons distal to the expected locus of change, e.g. the aliphatic-substituted ring. Parenthetically, the 2H spectra of 3 and a number of like-deuterated counterparts among the other salts were also acquired, and none showed any sign of thermolytic change on the part of the three non-alkylated phenyls. Simultaneously, in the aromatic region, the original P-coupled AA′BB′ signals of parent 3 centred at 7.44 and 7.51 ppm grew much smaller, and were replaced with two emerging sets of AA′BB′ patterns at lower field. One pattern set, centred at 7.74 and 8.36 ppm, was considerably larger than the other, which appeared at 7.90 and 8.09 ppm. Coincident with these changes, a small new peak appeared at 10.12 ppm, while a broader, much larger resonance grew in at 9.03 ppm. The location of the former was consistent with an aldehyde –CHO; the latter, while at atypically high field for a carboxylic acid COOH, was nevertheless concluded to be such. This assignment was buttressed by (a) the intensity of the signal, which integrates as one relative to the integrated intensity of four arising specifically from the 7.74/8.36 ppm AA′BB′ signal set noted above, and (b) the complete disappearance of the signal upon addition of a trace of D2O. Note that D2O addition had no effect on the peak at 10.12 ppm, consistent with it arising from the non-exchangeable H of an aldehyde group. Progressive changes in the 13C-NMR of 3 likewise support the emergence over time of aldehyde (4) and acid-bearing (5) derivatives originating in the reaction of the tolyl CH3, per Scheme 1. The clearly-identifiable resonance of the latter feature at 21.52 ppm grows steadily smaller upon heating, while a new peak emerges (168.08 ppm) that is assigned as a –COOH carbon. A second, smaller carbonyl peak (–CHO) was observed at 165.65 ppm. In addition, the very small AA′BB′ set of 1H peaks at 7.90 and 8.09 ppm comport well with an aldehyde-bearing arene ring, while the large AA′BB′ set of 1H peaks at 7.74 and 8.36 ppm are consistent with a –COOH functionalized ring. Further validating these assignments, the ESI-MS of the pyrolysis product is clearly consistent with conversion of the tolyl-bearing cation (m/z = 368) into one bearing a carboxylic acid group in its place (m/z = 398), while an additional small peak at m/z = 382 is likewise consistent with the conversion of the methyl group into an aldehyde. The NMR and MS data mutually support the proposition that aerobic thermolysis of 3 converts the tolyl methyl group into oxygenated species, specifically an aldehyde and a carboxylic acid; further, we posit that this transformation is sequential, oxidation to the aldehyde preceding oxidation to the acid. Such a pattern of reactivity closely tracks that observed for the high-temperature oxidative transformation of structures in coal and coal model compounds.21 In those instances, aliphatic moieties attached to aromatic rings form benzyl radicals which then react with O2. Further support for the proposition that the same pattern is plausible with alkylated phosphonium salts is provided by calculations‡ which show that the spin densities of a putative benzyl radical formed by 3 and that formed by toluene are virtually indistinguishable. Also, the idealized exchange reaction that connects these two radicals and their parents is energetically flat, with formation of toluene's benzyl radical favoured by just 0.53 kcal mol−1. Even so, in order to provide direct experimental support for thermolysis involving sequential changes to 3, we independently prepared salt 4, the proposed aldehyde intermediate. This material was then subjected to pyrolysis under the conditions previously stated. Doing so, as anticipated, resulted in the conversion of the aldehyde group to the corresponding acid 5 (1H-, 13C-NMR; ESI-MS). However, a careful examination of the NMR spectra suggested an additional, unexpected change; while the peaks for the aldehyde group were now gone, and those corresponding to the acid were in evidence, the aromatic region of both the 1H- and 13C-spectra were more complex than should have been the case were only 5 now present. Proceeding on conjecture, we returned the sample to the 300 °C oven for an additional 96 hour period, at the end of which the sample was again analysed. Our suspicions were confirmed; the sample had undergone complete decarboxylation, forming 1, the parent salt PPh4+Tf2N−, a species previously having been established to be stable to ca. 425 °C (at which point the anion, not the PPh4+ cation, is observed to decompose).2 Significantly, the parallel thermal decarboxylation of benzoic acid to benzene is known.22 Taking a cue from the previously discussed aldehyde-to-acid experiment, we independently prepared 5, the p-CO2H-functionalized PPh4+ Tf2N− salt, which we then thermolyzed at 300 °C in order to further validate our earlier-stated hypothesis that the acid, once formed, undergoes decarboxylation to form 1. Once again, the outcome matched our expectation. Accordingly, it is manifestly clear that if the acid-functionalized 5 underwent decarboxylation in those instances, any other alkyl-functionalized salts which produce it are destined to fall, upon extended thermolysis at high temperatures, into the rather deep thermodynamic well which PPh4+Tf2N− (1) appears to constitute. In addition to studying the nature of thermo-oxidative transformations of p-tolyl salt 3, we likewise prepared and investigated the behaviour of m-tolyl congener 6 (Scheme 3). As illustrated, the sequential conversion of 6 first into the corresponding aldehyde, followed by further oxidation to the derivative acid, and ending with decarboxylation of the latter to form 1, was apparent from changes over time to the 1H- and 13C-NMR of the heated material, as well as from changes in its ESI-MS. Clearly, the nature of the putative, initially-formed benzylic radical or the last-step carboxylic acid group as being in- or not-in-conjugation with the phosphonium centre had no manifest impact on the eventual outcome of the reaction. Scheme 3 Aerobic, pyrolytic conversion of 6 into 1, and a proposed mechanism for the process. Noting that each of the forgoing methylated tetraphenylphosphonium-cored cations bears its methyl group on a phenyl ring attached directly to the centre of cationic charge, the question arises as to whether this placement contributes to its relative thermal instability. In order to probe this possibility, we prepared a methylated analogue of 2 (compound 8, Scheme 4), that former parent cation having been previously demonstrated (like 1) to be thermally stable at 300 °C, in air, for months. Scheme 4 Aerobic, pyrolytic conversion of 8 into 2 and 11, and a proposed mechanism for the process. As with 3 and 6, samples of 8 were heated to 200 °C, 250 °C, and 300 °C for 96 hours, after which time they were analysed by multi-nuclear NMR and ESI-MS. Insofar as we could determine, there were no changes to the sample subjected to heating at 200 °C. However, a by-now familiar set of changes were observed with the samples of 8 heated to 250 °C and 300 °C. In the former case, a substantial decrease in the integrated intensity (1H-NMR) of the tolyl methyl group peak versus those from the collective aromatic protons was observed – the parent compound exhibiting a ratio of ≈1 : 7, while that from the sample heated to 250 °C was ≈1 : 11, reflecting a significant diminution of the methyl group content in the sample. Equally significant, the 1H-NMR of the sample showed an emerging singlet at 9.94, consistent with the oxidation of the tolyl methyl group into an aldehyde, as observed when 3 and 6 were pyrolyzed. Likewise, a new resonance was observed at 169.27 ppm in the 13C-NMR, also consistent with the incipient transformation of the –CH3 group into –CHO (9, Scheme 4). After 96 hours at 300 °C, the 1H- and 13C-spectra alike indicated the complete absence of the tolyl methyl group. Likewise, the peaks that had indicated its transformation into an aldehyde group in the 250 °C experiment were absent. Present instead was a peak (166.67 ppm) in the 13C-NMR consistent with the presence of a –COOH group (10), presumably formed by the oxidation of the foregoing intermediate aldehyde; its occurrence was also confirmed by ESI-MS. Perhaps most significant was a large MS peak indicating the presence of a material having a mass corresponding to that of a phenol-substituted tetraphenylphosphonium ion, i.e.11. So, in order to more definitively confirm the presence of 11 in the pyrolyzate, we prepared an authentic sample on which data could be acquired for purposes of comparison. The reference compound – pure 11 – proved to have a highly distinctive, P-coupled doublet in its 13C spectrum centred at 104.72 ppm from the P-bonded ring carbon para to the phenolic OH; likewise, 11 featured characteristic peaks in the 1H-NMR assignable to the ring protons immediately adjacent to the –OH group. The formation of 11 in the present case, when it is not produced by the pyrolysis of 2, indicates that at least under those conditions, the diphenylether linkage in 8 appears to be more labile than the comparable bond in the non-alkylated counterpart 2. Our simplest explanation is that the para methyl group of 8 gives rise to an easily-formed benzylic radical, which then leads to fragmentation at the ether bond. We note that radical exchange calculations‡ show that the diphenylether linkage of such a distal benzylic radical is slightly easier to break than that of parent 8, by about 2–3 kcal mol−1. Once satisfied that we had clarified (to a reasonable degree) the thermal fate of simple methyl groups appended to PPh4+-type cores, we next explored the nature of transformations (if any) that would occur when a methylene group was present in the ion. In this instance, two salts, 12 and 14, were prepared and subjected to the standard thermal treatment. Per Scheme 5, thermolysis of the ethyl-substituted salt 12 appears to proceed in a fashion closely related to that involved in the tolyl-methyl oxidation. At all temperatures, MS analysis shows the emergence of an unambiguous peak for the same carboxylic-acid functionalized salt (5) as that formed by the pyrolysis of 3, indicating that the –CH3 of the ethyl is altogether shed, with the –CH2 group being oxidized to –COOH. Even so, it is equally apparent, especially from MS and NMR of the 200 °C experiments, that the reaction proceeds by the formation of a ketone, from which the –CH3 group is subsequently cleaved and replaced by –OH, forming the carboxylic acid. We note that oxidation of the neutral counterpart molecule ethylbenzene has also been proposed to begin at the methylene position,23 a logical surmise since benzylic radicals are known to be considerably more stable than those on a primary carbon centre.24 In the case of the present IL, further confirmation of this pathway was obtained when we independently prepared the putative ketone-functionalized intermediate, 13, and aerobically pyrolyzed it, finding the product of this process to be carboxylic acid 5. This, in turn, upon prolonged heating underwent decarboxylation to form the PPh4+ cation (1). Scheme 5 Aerobic, pyrolytic conversion of 12 into 1, and a proposed mechanism for the process. The outcome of the thermolysis experiments focused on the second type of methylene-bearing cation, 14, were unexpected (Scheme 6). We anticipated the clean conversion of 14 to benzophenone-like 15, largely on the basis of 15 having been previously prepared and established by us to be as thermally stable as the parent cation 1.5 To our thinking, the degree of oxidation of 14 would be limited by the presence of two phenyl groups bracketing the aliphatic moiety; then, lacking a means to eliminate CO2, the system would be ‘stuck’ at that stage of transformation. That, however, did not prove to be the case. Whereas 15 as previously reported5 was prepared from a pre-assembled benzophenone tecton, its formation in the present case would be expected to proceed through an intermediate such as one of those shown in Scheme 6. In that event, it seems plausible to propose that the potential energy surface in the vicinity of the intermediate(s) might have multiple saddle points sufficiently low in energy to be crossed at the high temperatures of the present experiments, leading to more than one product. Significantly, the formation of more than one product is precisely what we observed. Scheme 6 Aerobic, pyrolytic conversion of 14 into 1 and 15, and a proposed mechanism for the process. Based upon the intensities of select signals in the pyrolysate 1H-NMR spectra, within 96 hours at 300 °C, the bulk of 14 was converted to the expected product 15, although acid 5 and parent cation 1 were observed in the product mixture as well. However, while the lower-temperature experiments (200 °C and 250 °C) also showed 15 being formed after 96 hours, an observable amount of 4 was also present, in keeping with it being an intermediate in the formation of 5, which in turn is eventually pyrolyzed to 1. It is notable, however, that based on prior observation, 15 is not further oxidized, nor is its side chain cleaved;515 is not converted into 4. Consequently, it seems apparent that 4 must emerge from one of the proposed intermediates, or another that is not anticipated in Scheme 5. The wholesale cleavage of an aromatic ring from 14 – made clear by the formation of 4, 5, and eventually 1 during the pyrolysis – was a surprise to us. Even so, on reflection it made clear that whatever intermediate(s) was first formed in the oxidation of 14, it had to be of a character that could lead to ring loss. In the interest of addressing this possibility, we separately prepared prospective intermediate 16, which was then pyrolyzed to ascertain whether this actually resulted in the formation of any of the arene ring-cleavage salts (note: regrettably, we were not able to prepare 17 in order to study its decomposition products). The heating of 16 in air resulted in the formation of observable quantities of 4 in fewer than 24 hours, whether at 200 °C, 250 °C, or 300 °C. After 72 hours at 300 °C, neither 4 nor 5 were still observed in the NMR, having been fully driven to 1. Insofar as could be ascertained, the latter, along with 15, constituted the final reaction products. In all, this constellation of derivatives of 14 is very much in keeping with that established to emerge from the decomposition of the molecular counterparts (diphenylcarbinol and diphenylmethyl hydroperoxide) of the intermediates proposed in Scheme 6.25–27 The methyl- and methylene-appendage experiments concluded, we proceeded to explore the behaviour of the next aliphatic C–H type in the planned evaluation sequence, methine moieties as found within isopropyl- (18) and triphenylmethyl groups (19) (Schemes 7 and 8). In these instances, the study of the behaviour of 18 is of particular note since a high-volume industrial process, the Hock synthesis of phenol and acetone, proceeds through the radical-based thermal reaction of cumene (isopropyl benzene) with O2.28–32 Scheme 7 Aerobic, pyrolytic conversion of 18 into 11 and 1, and a proposed mechanism for the process. Scheme 8 Aerobic, pyrolytic conversion of 19 into 1 and 15, and a proposed mechanism for the process. In the Hock (cumene hydroperoxide) process, complete scission of the isopropyl group from the aromatic ring is accomplished, resulting in its replacement by a phenolic –OH, and accompanied by the liberation of acetone. However, the second step in the process requires catalysis by acid, and in the present experiments involving the pyrolysis of 18, none is added (our current interest being in outcomes driven by aerobic pyrolysis alone). That being noted, we can report that no phosphonium counterpart to phenol, i.e., 11, is observed by NMR, although a small peak of the appropriate mass for this is observed by ESI-MS, a more sensitive technique. Over the course of 96 hours, the 1H- and 13C-NMR spectra of 18 did show the steady disappearance of signals associated with the appended isopropyl group. In the 1H-spectrum, these signals were a doublet of intensity six at 1.31 ppm and a heptet of intensity one at 3.08 ppm; the associated signals in the 13C-spectrum were observed at 23.04 and 34.10 ppm. As these diminished, a large new singlet gradually grew in and then faded out over time at 2.69 ppm in the 1H-spectrum, and a similar signal grew in/and faded out at 26.73 ppm in the 13C spectrum. These signals are identical with ones from an authentic reference sample of 13, and are taken as clearly indicative of its formation in significant amounts during pyrolysis. Interestingly, the formation of acetophenone, the non-ionic counterpart of 13, is known to be a side product in the oxidation step Hock process, as is dimethylbenzyl alcohol. In the case of salt 18, only a trace of the phosphonium dimethylbenzyl alcohol counterpart was observed, and it by ESI-MS only. In the final analysis, after 96 hours at 300 °C, 18 was converted to a mixture of 5 and 1, and based upon outcomes already discussed, the 5 would have fully transformed to 1 given additional time; here again, the P-bonded phenyl constituted a firewall that sharply limited the overall degree of thermal decomposition of salt 18. The final aliphatic-C–H-bearing salt to be studied was 19 (Scheme 8), a species in which a solitary methine-type C–H was flanked only by aromatic rings. We speculated that this compound might prove to be more robust, having its C–H moiety isolated from any other aliphatic components. However, even after only 12 hours at 250 °C, the compound showed substantial degradation; after 96 hours at 250 °C and 300 °C, it was totally decomposed (note: the sample at only 200 °C also exhibited a degree of decomposition after 96 hours). Interestingly, the pyrolytic residue consisted of only two phosphonium ions – 15 and 1 – suggesting that the postulated oxidative intermediate(s) 20 and 21 thermally rearranged by cleavage of an aromatic module from the phosphonium cation. In order to investigate this possibility, a sample of 19 was pyrolyzed at 300 °C overnight, the sample flask being provided with a cold trap to capture volatiles released during the experiment. Having successfully trapped a bolus of colourless liquid as a result, we subjected it to analysis by NMR, and were able to confirm that the liquid consisted of benzene and benzophenone, the exactly expected materials complimentary to the phosphonium ions remaining after pyrolysis. Having sussed out the nature of the non-volatile thermolysis products, and having proposed mechanisms for their formation, we turned to an evaluation of the volatile products to see if the aforementioned mechanisms remained reasonable in light of the character of these gases. In doing so, we observed three gaseous off-products the presence of which was universal in the IR spectra of samples examined (ESI†), and which could be unambiguously assigned: H2O, CO2, and CO. Significantly, the comparative off-gas spectra in the 2000–2400 cm−1 range of compounds 3–5 (Fig. 3) provide especially valuable insights supportive of the decomposition mechanisms already proposed (vide supra). Specifically, the spectrum of 5 – bearing a COOH group – is essentially devoid of CO, and dominated by the loss of CO2, as previously suggested. However, the pyrolysis of 4 – incorporating a “pre-made” COH group – generates a large quantity of CO, again consistent with our proposed decomposition mechanism (see Schemes 2 and 3). Here too, however, CO2 is also in evidence, consistent with the proposition that the conversion of 4 to 1 can proceed by direct CO elimination or further oxidation to 5, followed by CO2 loss. In turn, note that the direct, start-to-finish pyrolysis of 3 shows both CO and CO2 generation, again as previously predicted. As a side note, the absorptions between those of CO2 and CO in the spectrum of the gaseous products from 3 could not be unambiguously assigned, but may be from a C/N/O unsaturated moiety. Interestingly, these peaks only appeared when pyrolysis temperatures eventually reached 400 °C, beyond which point the Tf2N− anion is known to decompose. Given the inertness of the elemental nitrogen which would have been present in the air under which the pyrolyses were conducted, it seems reasonable to posit that these absorptions stem from a material generated by anion decomposition, since the only N-containing element of the salt was the Tf2N− anion. Fig. 3 Left to right: Key region of IR spectra of off-gases from 3–5. Note that 4 gives off a considerable amount of CO, consistent with decarbonylation. In turn, 5 gives off considerable CO2, but virtually no CO. Pyrolysis of 3 generates observable amounts of both gases. Each of these observations is consistent with the mechanisms proposed for the respective compound's decomposition. The third substance apparent from compound 3 was not conclusively assigned. However, it appears in a range consistent with a cyanate-type moiety. Notably, it only appears after the sample is heated to T ≥ 400 °C, at which point the Tf2N− anion of the salt – it's only N-containing component – is already known to decompose. To conclude our study, we thought it worth examining what happens to select representatives of the present salts when pyrolyzed in the absence of air. Specifically, we hoped to acquire insight into whether O2 is strictly necessary to “decompose” an initial alkyl group (methyl, for sake of simplicity), and whether – once oxygen is incorporated into an ion structure – ongoing access to O2 is needed to further the progression of thermal changes. Accordingly, samples of 3–6 were sealed into glass ampoules under vacuum and heated to 300 °C for 24 hours, this shorter period being utilized in order to increase our chances for observing the (prospective) multitude of intermediate materials we thought might be generated in the reaction. Significantly, pyrolysis of 3 and 6 in the absence of oxygen resulted in no significant changes to the materials as indicated by NMR, and only a few new, very small peaks in the ESI-MS. In contrast, samples of 4 and 5 – both of which began with oxygen as a compositional element – continued to show significant changes when heated anaerobically; recall that each of these had been prepared in a pure form to provide reference material for use in characterizing proposed products/intermediates being generated by the aerobic pyrolysis of 3. In the case of aldehyde-bearing 4, only a trace of that parent material remained after 24 hours. Most (based upon ESI-MS and NMR) had been converted into 1. We find this interesting, since (in the absence of an exogenous source of oxygen), the further oxidation of 4 into 5 that needs to occur to enable decarboxylation of the latter into 1 should not have been possible. However, we again note that another mechanism for the conversion of it into 1 is possible: decarbonylation. The latter has been reported previously as a reaction to which benzaldehyde is subject under high-temperature conditions.33,34 In the anaerobic thermolysis of pure 5, major product was clearly the expected decarboxylation product 1. However, an unexpected peak was also observed at m/z = 395.25, equal to the mass of 5 (m/z = 383.17) plus an additional carbon atom. We have been unable to further clarify its structure. Whatever it may be, we speculate that this material was possibly formed in the high-energy environment of the ESI-MS experiment. Collectively, the foregoing data provide a reasonable basis to conclude that the thermally-induced cleavage of aliphatic C–H bonds in tetraarylphosphonium cations, when done in the presence of oxygen, leads to the loss or wholesale transformation of the aliphatic component, but almost exclusively without damage to the central, tetraphenylphosphonium core; the Tf2N− salt of 1 remains, it clearly being highly rugged with respect to thermolysis. Indeed, we have been able to routinely recover pure 1 by simple recrystallization of the post-pyrolysis residues. In addition, these results may also be of utility with respect to new classes of high-performance polymers (based upon tetraarylphosphonium ions) such as those under development by Smith and co-workers, materials that are envisioned for use in the chemically and thermally demanding conditions found in fuel cells.35–37 We also anticipate that thermally stable, perarylphosphonium salts may prove to be interesting and thermally stable property modifiers if added to commercial high-performance polymers such as PEEK and PES.38,39 Not only might their addition modify properties such as plasticity, but they could prove to be effective flame retardants since many high-valent phosphorous compounds are used for that purpose. Conflicts of interest There are no conflicts to declare. Supplementary Material RA-010-D0RA03220D-s001 We thank the National Science Foundation for support under grant number CHE-1464740. This work was also made possible in part by a grant of high performance computing resources and technical support from the Alabama Supercomputer Authority. ‡ Calculations were UB3LYP/6-311+g(d,p) optimizations of the radical doublets performed using Gaussian'16; spin densities were rendered with Spartan'08. The exchange reaction energy was estimated using B3LYP/aug-cc-pvtz single-point energies and zero-point corrections from B3LYP/6-31g(d) frequency calculations. The atomic spin density assignments are based upon natural charges as determined by population analysis of the UB3LYP/aug-cc-pvtz densities. ==== Refs Notes and references Cassity C. G. Mirjafari A. Mobarrez N. Strickland K. J. O'Brien R. A. Davis Jr J. H. Chem. Commun. 2013 49 7590 7592 10.1039/C3CC44118K 23873358 Scheuermeyer M. Kusche M. Agel F. Schreiber P. Maier F. Steinrueck H.-P. Davis Jr J. H. Heym F. Jess A. Wasserscheid P. New J. Chem. 2016 40 7157 7161 10.1039/C6NJ00579A Siu B. Cassity C. G. Benchea A. Hamby T. Hendrich J. Strickland K. J. Wierzbicki A. Sykora R. E. Salter E. A. O'Brien R. A. West K. N. Davis Jr J. H. RSC Adv. 2017 7 7623 7630 10.1039/C6RA25774G Benchea A. Siu B. Soltani M. McCants J. H. Salter E. A. Wierzbicki A. West K. N. Davis Jr J. H. New J. Chem. 2017 41 7844 7848 10.1039/C7NJ01788J Cassity C. A. Siu B. Soltani M. McGeehee J. L. Strickland K. J. Vo M. Salter E. A. Stenson A. C. Wierzbicki A. West K. N. Rabideau B. D. Davis Jr J. H. Phys. Chem. Chem. Phys. 2017 19 31560 31571 10.1039/C7CP06278H 29165458 Soltani M. Siu B. Salter E. A. Wierzbicki A. West K. N. Davis Jr J. H. Tetrahedron Lett. 2017 58 4628 4631 10.1016/j.tetlet.2017.10.073 Rabideau B. D. West K. N. Davis Jr J. H. Chem. Commun. 2018 54 5019 5031 10.1039/C8CC01716F 29637207 Rabideaux B. D. Soltani M. Parker R. A. Siu B. Salter E. A. Wierzbicki A. West K. N. Davis Jr J. H. Phys. Chem. Chem. Phys. 2020 22 10.1039/D0CP01214A Fraser K. J. MacFarlane D. R. Aust. J. Chem. 2009 62 309 321 10.1071/CH08558 MacFarlane D. R. Tachikawa N. Forsyth M. Pringle J. M. Howlett P. C. Elliott G. D. Davis Jr J. H. Watanabe M. Simon P. Angell C. A. Energy Environ. Sci. 2014 7 232 250 10.1039/C3EE42099J Maton C. De Vos N. Stevens C. V. Chem. Soc. Rev. 2013 42 5963 5977 10.1039/C3CS60071H 23598738 Blake E. S. Hammann W. C. Edwards J. W. Reichard T. E. Ort M. R. J. Chem. Eng. Data 1961 6 87 98 10.1021/je60009a020 Cassidy P. E. , Thermally Stable Polymers: Synthesis and Properties, Marcel Dekker, Inc., 1980 Johns I. B. McElhill E. A. Smith J. O. Ind. Eng. Chem. Prod. Res. Dev. 1962 1 2 6 10.1021/i360001a001 Johns I. B. McElhill E. A. Smith J. O. J. Chem. Eng. Data 1962 7 277 281 Madison J. J. Roberts R. M. Ind. Eng. Chem. 1958 50 237 250 Baum G. and Short F. R. , Thermal stability of organic compounds determined by the isoteniscope method, Technical Report No. AFML-TR-65-347, Air Force Materials Laboratory, Wright-Patterson Air Force Base, 1966 Marcoux D. Charette A. B. Adv. Synth. Catal. 2008 350 2967 2974 10.1002/adsc.200800542 Fox E. B. Smith L. T. Williamson T. K. Kendrick S. E. Energy Fuels 2013 27 6355 6367 10.1021/ef401148j Personal communication from Charles Potter, PhD, applications specialist, TA Instruments – Waters LLC: “TGA can be used to look at the interconversion and concurrent weight change. Since the kinetics of the weight changes for the concurrent events are typically different, changing experimental variable (SiC) such as heating rate and purge gas type can sometimes suppress or enhance certain weight change events or move them with respect to time or temperature. In many case (SiC), the multiple weight change events are too many to permit there (SiC) separation using only gravimetric information” (emphasis added). See: Stonhill L. G. Can. J. Chem. 1959 37 454 459 10.1139/v59-059 Qi X. Chen L. Xin H. Ji Y. Bai C. Song R. Xue A. Liu F. Energy Fuels 2018 32 10469 10477 10.1021/acs.energyfuels.8b02165 Smith R. E. Hinshelwood C. N. Proc. R. Soc. London, Ser. A 1940 175 961 131 142 Gutmann B. Elsner P. Roberge D. Kappe C. O. ACS Catal. 2013 3 2669 2676 10.1021/cs400571y Davies D. I. and Parrot M. J. , Free Radicals in Organic Synthesis, Springer Verlag, Berlin, 1978 Du Z. Sun Z. Zhang W. Miao H. Ma H. Xu J. Tetrahedron Lett. 2009 50 1677 1680 10.1016/j.tetlet.2009.01.077 Cadogan J. I. G. Hey D. H. Sanderson W. A. J. Chem. Soc. 1958 4498 4503 Rossi M. J. McMillan D. F. Golden D. M. J. Phys. Chem. B 1984 88 5031 5039 Duh Y.-H. Kao C.-S. Hwang H.-H. Lee W. W.-L. Trans. Inst. Chem. Eng. 1998 76 271 276 Seubold Jr F. H. Vaughan W. E. J. Am. Chem. Soc. 1953 75 3790 3792 10.1021/ja01111a055 Yuan W. Li Y. Gagaut P. Yang J. Qi F. Combust. Flame 2015 162 3 21 Weber M. , Daldrup J.-B. G. and Weber M. , Non-catalyzed Radical Chain Oxidation: Cumene Hydroperoxide, in Liquid Phase Aerobic Oxidation Catalysis: Industrial Applications and Academic Perspectives, ed. S. S. Stahl and P. L. Alsters, Wiley-VCH, 1st edn, 2016, ch. 2 Weissermel K. and Arpe H.-J. , Industrial Organic Chemistry, Wiley-VCH, Weinheim, 4th edn, 2003 Nagai Y. Matubayasi N. Nobuyuki N. Nakahara M. Chem. Lett. 2004 33 622 623 10.1246/cl.2004.622 Allen J. K. , US Pat. US4266084A, 1981-05-05 Yang X. Wen Y. Chumanov G. Smith R. C. J. Polym. Sci., Part A: Polym. Chem. 2017 55 1620 1625 10.1002/pola.28535 Wan W. Yang X. Smith R. C. J. Polym. Sci., Part A: Polym. Chem. 2017 55 1984 1990 Moritz R. Wagner M. Schollmeyer D. Baumgarten M. Muellen K. Chem.–Eur. J. 2015 21 9119 9125 25965029 Mittal V. , in High-Performance Polymers and Engineering Plastics, ed. V. Mittal, Scrivener, Beverly, MA, USA, 2011, pp. 1–20 Levchik S. V. Ivashkevich O. A. Costa L. Gaponik P. N. Andreeva T. N. Polym. Degrad. Stab. 1994 46 225 234 10.1016/0141-3910(94)90054-X
PMC009xxxxxx/PMC9087897.txt
==== Front Front Vet Sci Front Vet Sci Front. Vet. Sci. Frontiers in Veterinary Science 2297-1769 Frontiers Media S.A. 35558891 10.3389/fvets.2022.871549 Veterinary Science Original Research Blood B Cell Depletion Reflects Immunosuppression Induced by Live-Attenuated Infectious Bursal Disease Vaccines Courtillon Céline 1 * Allée Chantal 1 Amelot Michel 2 Keita Alassane 2 Bougeard Stéphanie 3 Härtle Sonja 4 Rouby Jean-Claude 5 Eterradossi Nicolas 6 Soubies Sebastien Mathieu 1 † 1Ploufragan-Plouzané-Niort Laboratory, OIE Reference Laboratory for Infectious Bursal Disease, French Agency for Food, Environmental and Occupational Health & Safety (ANSES), VIPAC Unit, Ploufragan, France 2Ploufragan-Plouzané-Niort Laboratory, French Agency for Food, Environmental and Occupational Health & Safety (ANSES), SELEAC Service, Ploufragan, France 3Ploufragan-Plouzané-Niort Laboratory, French Agency for Food, Environmental and Occupational Health & Safety (ANSES), EPISABE Unit, Ploufragan, France 4Ludwig-Maximilians-Universität München, Veterinärwissenschaftliches Department, München, Germany 5French Agency for Veterinary Medicinal Products (ANMV), French Agency for Food, Environmental and Occupational Health & Safety (ANSES), Javené, France 6Ploufragan-Plouzané-Niort Laboratory, French Agency for Food, Environmental and Occupational Health & Safety (ANSES), Management Department, Ploufragan, France Edited by: Yulong Gao, Harbin Veterinary Research Institute (CAAS), China Reviewed by: Ping Wei, Guangxi University, China; Venugopal Nair, The Pirbright Institute, United Kingdom *Correspondence: Céline Courtillon [email protected] This article was submitted to Veterinary Infectious Diseases, a section of the journal Frontiers in Veterinary Science †Present Address: Sebastien Mathieu Soubies, UMR 1225 IHAP, Toulouse, France 25 4 2022 2022 9 87154908 2 2022 08 3 2022 Copyright © 2022 Courtillon, Allée, Amelot, Keita, Bougeard, Härtle, Rouby, Eterradossi and Soubies. 2022 Courtillon, Allée, Amelot, Keita, Bougeard, Härtle, Rouby, Eterradossi and Soubies https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Immunosuppression in poultry production is a recurrent problem worldwide, and one of the major viral immunosuppressive agents is Infectious Bursal Disease Virus (IBDV). IBDV infections are mostly controlled by using live-attenuated vaccines. Live-attenuated Infectious Bursal Disease (IBD) vaccine candidates are classified as “mild,” “intermediate,” “intermediate-plus” or “hot” based on their residual immunosuppressive properties. The immunosuppression protocol described by the European Pharmacopoeia (Ph. Eur.) uses a lethal Newcastle Disease Virus (NDV) infectious challenge to measure the interference of a given IBDV vaccine candidate on NDV vaccine immune response. A Ph. Eur.-derived protocol was thus implemented to quantify immunosuppression induced by one mild, two intermediate, and four intermediate-plus live-attenuated IBD vaccines as well as a pathogenic viral strain. This protocol confirmed the respective immunosuppressive properties of those vaccines and virus. In the search for a more ethical alternative to Ph. Eur.-based protocols, two strategies were explored. First, ex vivo viral replication of those vaccines and the pathogenic strain in stimulated chicken primary bursal cells was assessed. Replication levels were not strictly correlated to immunosuppression observed in vivo. Second, changes in blood leukocyte counts in chicks were monitored using a Ph. Eur. - type protocol prior to lethal NDV challenge. In case of intermediate-plus vaccines, the drop in B cells counts was more severe. Counting blood B cells may thus represent a highly quantitative, faster, and ethical strategy than NDV challenge to assess the immunosuppression induced in chickens by live-attenuated IBD vaccines. IBDV live-attenuated vaccine vaccine safety immunosuppression B cells replication ==== Body pmcIntroduction Immunosuppression is, as defined by Gimeno and Schat (1), “A state of temporary or permanent dysfunction of the immune response resulting from insults to the immune system with suboptimal antibody production and/or suboptimal innate and/or suboptimal cell-mediated responses leading to increased susceptibility to disease.” In poultry production, immunosuppression is a multifactorial and recurrent problem worldwide. Infection with the Birnaviridae family's infectious bursal disease virus (IBDV) is one of many possible causes of immunosuppression (2). This virus mainly replicates in immature B cells found in the bursa of Fabricius (BF), a bird-specific immune organ responsible for the development and selection of B cells during the first month of life (3). Infection by this agent may result in the destruction of the organ (3). Depending on the viral pathotype, the age of chickens, their genetics, and their IBDV immune status, infection can induce an acute and sometimes lethal disease called Infectious Bursal Disease (IBD), also known as Gumboro disease, or be asymptomatic. In both cases, surviving birds, due to BF atrophy, will undergo a transient or permanent immunosuppression, in turn resulting in secondary infections and vaccination failures (4). Infectious bursal disease control is achieved by a combination of biosecurity and medical prophylaxis (vaccination). Several vaccination strategies exist. These include the use of live-attenuated vaccines, inactivated vaccines, immune complexes, and vectored vaccines (4). Live-attenuated IBD vaccines are particularly adapted to mass vaccination due to low costs and to an easy distribution in the drinking water. Several types have been developed based on their residual virulence. Hence, live-attenuated vaccines are classified as mild, intermediate (further abbreviated as “I”), intermediate-plus (“I+”), and hot. The higher the residual virulence of an attenuated IBD vaccine, the earlier it can be administered to chicks to break through maternal antibodies and induce an immune response. I+ and hot vaccines can be used in epidemiological contexts (for instance circulation of very virulent strains of IBDV) in which chicks are particularly at risk. Those vaccines can nevertheless induce significant B cell depletion in the bursa and cause immunosuppression by themselves. Vaccine challenge strategies, which are used to check vaccine safety during licensing processes, can demonstrate this immunosuppression. In those strategies, specific pathogen-free chicks, after receiving IBDV vaccination, are immunized with a model antigen (e.g., killed Brucella abortus) or a live-attenuated vaccine [e.g., Hitchner B1 strain of Newcastle disease virus (NDV)] (5). The ability of the chicks to mount an immune response to those immunogens is then evaluated through their humoral response and/or an infectious challenge. In particular, the assessment of immunosuppressive properties of IBD vaccine candidates following the European Pharmacopeia (Ph. Eur.) protocol requires administering IBD vaccine candidate to specific pathogen-free (SPF) chicks. After 15 days, around the time when IBDV-induced bursal lesions peak, chicks are immunized with live-attenuated Newcastle Disease (ND) vaccine. A lethal NDV infectious challenge is performed 15 days after the NDV vaccination. For the vaccine candidate to be considered as safe both NDV serology and clinical protection results are expected to be similar in chicks vaccinated against IBD or not (6). Still according to the Ph. Eur., vaccine candidates that induce a significant degree of bursal microscopic damages in preliminary studies fall into the category of I+ vaccines candidates. Therefore, the standards retained by the Ph. Eur. monograph with regard to the immunosuppression testing are not applicable for licensing. However, a special warning is introduced in the Summary of Product Characteristics to describe how these vaccines are subjected to additional precautions for use. The aim of this study was to explore, in the interest of improving ethics and cost, alternatives to challenge-based protocols to quantify immunosuppression induced by IBD vaccine candidates. First, a Ph. Eur. —derived protocol was implemented to confirm the immunosuppressive properties of mild, I, and I+ live-attenuated IBD vaccines as well as a pathogenic viral strain. Next, two alternative strategies were then explored, namely (i) comparing the replicative capacity of vaccine viruses ex vivo using bursal B cells and (ii) measuring blood leucocyte counts after IBDV vaccination in SPF chicks. Materials and Methods Ethic Statements All animal trials were conducted in an animal facility approved for animal experiments (n°C-22–745–1). They were approved by Anses Ploufragan ethical committee (committee number C2EA-016/ ComEth ANSES/ENVA/UPEC) and authorized by French Ministry for higher education and research (permit number APAFiS#3925-2016020310538529, application number 16-016, statement number 08/03/16-1). Chickens were raised and humanely euthanized in agreement with EU directive number 2010/63/UE. Viruses Vaccine viruses used in this study were obtained from a local veterinary practice, reconstituted, and administered according to the manufacturer's instructions. To avoid any conflict of interest, currently distributed IBD vaccines used in this study were anonymized. All IBD I+ vaccines (coded as I+1, I+2, I+3, and I+4) marketed in France as of 2016 were included. As representative intermediate vaccines, two intermediate vaccines (I1 and I2) were also included. The NDV Poulvac Hitchner B1 strain (Zoetis) was used for NDV vaccination and the velogenic NDV strain “Herts33” (Anses Ploufragan laboratory) was used to challenge the birds. The classical IBDV virulent strain “F52/70” and the mild IBD vaccine strain “PBG98” (no longer marketed in France; passaged twice on chicken fibroblasts) were obtained from the collection of IBDV strains maintained by the OIE reference Laboratory (Anses, Ploufragan, France). Animal Experiments In total three experiments were performed. The layouts of the experiments are presented in Figure 1A (experiments 1 and 2) and Figure 3A (experiment 3), respectively. For each experiment, six groups of 10 1-day-old SPF chicks (Anses, Ploufragan, France) were housed in negative-pressure filtered-air isolators and were administered (or not, in case of controls) one dose of IBD vaccine by the eye-drop route. In the case of PBG98 or F52/70, chicks received 100 tissue culture median infectious doses (TCID50) or 100 embryo median infectious doses (EID50, corresponding for this virus to 740 TCID50), respectively. Chicks received (or not, in case of the non-vaccinated controls) one dose of ND vaccine by the eye-drop route after 14 days of IBDV vaccination. Controls included a group that received neither IBD nor ND vaccine (“no vaccine control”), and a group that received only the ND vaccine (“NDV vaccine control”). In the case of experiment 3, prior to ND vaccination, chicks were blood-sampled at the occipital sinus with EDTA-coated tubes (S-Monovette EDTA K2 2.7 mL, Sarstedt, reference 04.1915.100) for blood cell quantification. For all experiments, chickens were blood-sampled at the occipital sinus 2 weeks post-NDV vaccination, either in serology tubes (S-Monovette serum-gel, 2.6 mL, Sarstedt, reference 04.1904) for serological analyses only (experiments 1 and 2) or in EDTA-coated tubes for blood cell quantification as well as serological analyses (experiment 3). On the same day, chicks were inoculated intra-muscularly with 105 EID50 of NDV strain Herts33. A clinical follow-up was performed for 14 days following the challenge. Birds were rated as “clinically protected” when no clinical sign was observed or “non-protected” when they displayed neurological, respiratory signs, or prostration. Animals showing clinical signs clearly unrelated to the NDV challenge were excluded from the clinical follow-up. Animals exhibiting neurological signs and/or prostration reached the ethical endpoint of the experiment and were consequently humanely euthanized. Figure 1 Implementation of a European Pharmacopoeia-derived protocol for IBDV immunosuppression testing. (A) Layout of animal experiments 1 and 2. (B) Percentage of clinically protected chicks observed after velogenic Newcastle Disease Virus Challenge. Different letters indicate statistically significant differences (p < 0.05) between groups (within one given experiment) using Fisher's exact test with FDR adjustment method for multiple pairwise comparisons. (C) Serological response to NDV vaccination during experiments 1 and 2. Horizontal bars indicate the median of each group. Different letters indicate statistically significant differences (p < 0.05) between groups (within one given experiment) using Kruskal-Wallis test. NDV Serological Analyses In all experiments, serum (experiments 1 and 2) or plasma (experiment 3) obtained after blood sampling were analyzed by hemagglutination inhibition (HI) to quantify levels of anti-NDV antibodies according to Agence Française de Normalisation Norm U 47-011. Four hemagglutination units of NDV LaSota strain were used as antigen. Results were expressed as the reciprocal of the highest dilution of serum able to completely inhibit the hemagglutination of 4 units of NDV antigen in log2. In this assay, titers equal to or below 2 log2 are considered negative. To exclude any impact of EDTA from the blood-collection tube on HI titers, plasmas from experiment 3 were treated with CaCl2 (to a final concentration of 2.5 mM) to induce fibrin precipitation (7). After 2 h of incubation at room temperature, the fibrin clot was mechanically disrupted by stirring with micropipette tips, the tubes were centrifuged at room temperature for 10 min at 14 000 × g and supernatants were collected. Blood Leukocytes Quantification In the case of experiment 3, collected blood was analyzed by flow cytometry to quantify leukocytes following the approach developed by Seliger et al. (8) with the few modifications already described (9). Accordingly, blood thrombocytes (defined as K1+ cells), blood monocytes (defined as K1+, Kul01+ and CD45+ cells), blood T cells (defined as CD45+ cells that are additionally either CD4+ or CD8+ or TCRγδ+), and blood B cells (defined as CD45+ and Bu1+ cells) were quantified. Ex vivo Bursal Cell Infection Bursal cells were isolated from 9 to 15 week-old SPF chickens as previously described (10). Cells were maintained at 40°C, 5% CO2 in a lymphocyte culture medium consisting of Iscove's Modified Dulbecco's Medium (IMDM) with L-glutamine and HEPES (reference 21980–032, Gibco, Thermo Fisher), supplemented with 8% FBS, 2% SPF chicken serum (ANSES, Ploufragan, France), 1X Insulin Transferrin Selenium (reference 41400–045, Gibco, Thermo Fisher), 50 μM beta-mercaptoethanol, 1 μg/mL Phorbol 12-myristate 13-acetate (reference tlrl-pma, Invivogen), penicillin (200 IU/mL), streptomycin (0.2 mg/mL), and fungizone (2 μg/mL). For infection, one million bursal cells per well seeded in 96-well plates were inoculated at a target multiplicity of infection of 5 × 10−4. Inocula were titrated and titers are presented as the 0 h post-infection (h.p.i.) time-point in Figure 2A. Supernatants were collected at 16, 23, and 48 h.p.i. and were frozen at −70°C until titration. Additionally, at 16 and 23 h.p.i., cells were collected for immunolabeling and flow cytometry analyses. A total of 5 biological replicates (consisting of cells from distinct birds) were analyzed during two distinct experiments. Figure 2 Replication curves of IBDV vaccine and pathogenic strains in ex vivo chicken bursal cells. (A) Layout of the experiment. (B) Viral growth kinetics of IBDV vaccine and pathogenic strains. Intermediate vaccine strains appear in blue grey while intermediate-plus and pathogenic strains are in black. Supernatants collected at the indicated time-points were titrated on freshly isolated bursal cells. Viral titers, expressed as median tissue culture infective doses (TCID50) were calculated using the Read and Muench formula. Icons at each time-point indicate the median value obtained from a total of five biological replicates. Different letters indicate statistically significant differences (p < 0.05) between groups using Kruskall-Wallis test. (C) Example of flow cytometry dot-plot. P1: IBDV-negative dead cells; P2: IBDV-positive dead cells; P3: IBDV-negative live cells; P4: IBDV-positive live cells. (D) Percentages of live infected cells at 16 and 23 h post-infection. Dots indicate values for cells from individual bursae (color-coded). Box-plots indicate extreme values, quartiles and median. Different letters indicate statistically significant differences (p < 0.05) between groups using Kruskal-Wallis test. Immunolabeling and Flow Cytometry Bursal cells collected at 16 and 23 h.p.i. were labeled as previously described (10). Briefly, viability was evaluated using the Fixable Viability Dye eFluor 750 (Thermofisher, reference 65-0865-14) and cells were immunostained using an anti-VP3 monoclonal antibody [mAb 18, (11)] followed by a goat anti-mouse IgG2a Alexa Fluor 546 antibody (Thermofisher, reference A21133). Cells were analyzed on a FC500 MPL flow cytometer (Beckman Coulter). Virus Titration Virus titration of inocula and supernatants at 16, 23, and 48 h.p.i. was carried out as previously described (10). Briefly, serial dilutions of viral suspensions were incubated into 96-well plates with chicken bursal cells in lymphocyte culture medium at 40°C for 48 h in a humidified 5% CO2 incubator. Cells were then subjected to immunocytochemistry using the IBDV VP3 specific mAb 18, a horse-radish peroxidase coupled anti-mouse IgG serum and KPL Trueblue peroxidase substrate (Seracare, reference 5510-0030). Viral titers, expressed as TCID50/mL, were determined using Reed and Muench formula (12). Statistical Analyses Statistical analyses were performed using R version 4.0.3 (13). Differences in percentages of protection were analyzed using the Fisher's exact test followed by pair-wise comparisons using the “fisher.multcomp” function from RVAideMemoire package version 0.9-79 (14). All other quantitative parameters were analyzed using Kruskal-Wallis test followed by Fisher's least significant difference test with Holm adjustment method for multiple comparisons using the “kruskal” function from the package “Agricolae” version 1.3-2 (15). Results The Ph. Eur. IBDV Immunosuppression Protocol Confirms the Expected Immunosuppressive Properties of the Selected IBDV Strains In both experiments 1 and 2, following challenge with the virulent NDV strain Herts 33, all the NDV non-vaccinated control birds rapidly exhibited intense prostration and died or were euthanized within 72 h post-inoculation. All the NDV vaccine control birds showed complete clinical protection, without any sign of disease. Chicks that received both IBD and ND vaccines showed varying degrees of protection from the NDV challenge. Chicks from the PBG98-treated group showed 100% clinical protection. Those that received I vaccines had 80–90% protection, which did not differ statistically from those who received NDV vaccine controls. In contrast, protection in chicks that received I+ vaccines was significantly lower than that received NDV vaccine controls, with only 10–40% protection (Figure 1B). Finally, no animals from the F52/70 group were protected. HI testing for anti-NDV antibodies revealed expected background levels (1–2 log2) for non-vaccinated controls and high titers for NDV-vaccinated controls, with median values of 6 and 4.5 log2 in experiments 1 and 2, respectively (Figure 1C). Chicks vaccinated with PBG98 or I vaccines did not show any statistically significant difference in HI titers compared with NDV vaccine controls. In contrast, birds that received I+2, I+3, I+4, and F52/70 showed severe reduction in HI titers, almost reaching the background levels of non-vaccinated controls. Interestingly, chicks that received I+1 vaccine showed a numerically mild and statistically non-significant reduction in HI titers in spite of a statistically reduced clinical protection (only 40% birds protected). Ex vivo Replicative Fitness in Chicken Primary Bursal Cells Does Not Accurately Reflect Immunosuppression in vivo Primary bursal cells obtained from SPF chickens were infected with various IBDV strains to determine whether replication parameters (e.g., replicative fitness or cytopathogenicity) could mirror immunosuppressive properties. Strain F52/70 as well as the six vaccines tested in vivo were included in this approach. Since the PBG98 vaccine does not replicate in bursal B cells (data not shown), it was not included in the panel. For each of the tested strains, a dramatic increase in viral titers measured from bursal cells supernatants was visible as early as 16 h.p.i. and viral titers increased until 48 h.p.i. (Figure 2B). Interestingly, two trends in replication curves could be observed. On the one hand, I2, I+2, I+3, I+4, and strain F52/70 reached a viral titer higher than 107 TCID50/mL at 23 h.p.i. and then reached almost 109 TCID50/mL at 48 h.p.i. On the other hand, I1 and I+1 presented viral titers of about 106 TCID50/mL at 23 h.p.i. and then reached a plateau phase of about 107 TCID50/mL at 48 h.p.i., which was statistically lower compared with the first trend for this time point. In parallel of these infectious titrations, both the percentages of viable and IBDV-infected cells were analyzed by flow cytometry (example of dot-plot in Figure 2C). Significant differences in the percentage of live infected cells were detected at 16 h.p.i. with higher values for I2, I+2, I+4, and F52/70 (Figure 2D). At 23 h.p.i., a high variability in percentages between cells originating from distinct birds was observed, and differences between viruses were no longer significant. It is noteworthy that a bursa-specific trend was observed: for instance, at least 3 fold-higher percentages of infected cells were observed for bursa number 3 compared with bursa 4 for almost all viruses at 23 h.p.i. Intermediate-Plus Vaccines Induce a Dramatic Depletion of Blood B Cells in SPF Chickens Based on the hypothesis that I+ vaccines, inducing higher bursal damages, may reduce blood B cell concentrations, a third in vivo experiment was implemented in which four vaccines were tested (Figure 3A): I1 and I+3, for which ex vivo replication levels were correlated to the intensity of in vivo immunosuppression, and I2 and I+1, for which this correlation was not observed. Clinical protection from NDV challenge was consistent with results from experiments 1 and 2, except for the I+1 treated group, which no longer differed statistically from the NDV vaccine control (Figure 3B). Anti-NDV antibody levels were significantly reduced for all vaccines except I1 compared with NDV vaccine controls (Figure 3C), unlike in experiment 1 (Figure 1C). To check the lack of interference of the anti-coagulant, HI testing was performed again after adding CaCl2 to the plasmas to induce clotting. No difference in HI titers was observed after this treatment (data not shown). At both time-points, blood leukocyte counts revealed significant differences in B cell concentrations between virus strains. At 15 days of age, prior to NDV vaccination, both control groups exhibited concentrations close to 1,000 B cells/μL (Figure 3D, left panel). For I1 vaccinated birds, the median concentration significantly dropped to 143 B cells/μL. For I2 vaccinated birds, the drop was even more pronounced (median of 42 cells/μL, p < 0.05). For I+1 and I+3 vaccinated birds, B cell counts were statistically lower than in I vaccine groups, with medians below the detection limit of 8 cells/μL. In addition, no bird in these two groups presented more than 60 B cells/μL. Blood cell counts performed at 29 days of age (prior to NDV challenge) revealed a global increase of B cell counts in all groups compared with 15 days of age (Figure 3D, right panel). Both controls groups showed median concentrations above 2,000 B cells/μL. Median B cell concentrations significantly dropped to 956 for I1 vaccinated group. Those median concentrations were lower in I2, I+1, and I+3 vaccinated groups, with 130, 91, and 10 B cells/μL, respectively. Some birds from I+1 and I+3 treated groups still did not show any blood B cell at that time-point. Some milder variations were observed in other blood leukocyte concentrations: for instance, at 15 days of age, a statistically significant reduction in both monocytes and T cell concentrations were seen for I2 and I+3 groups compared with control groups (Supplementary Figure 1). Figure 3 Implementation of a modified European Pharmacopoeia-derived protocol for IBDV immunosuppression testing. (A) Layout of animal experiment 3. Extra analyses compared to animal experiments 1 and 2 appear in red text. (B) Percentage of clinically protected chicks observed after velogenic Newcastle Disease Virus challenge. Different letters indicate statistically significant differences (p < 0.05) between groups using Fisher's exact test with FDR adjustment method for multiple pairwise comparisons. (C) Serological response to NDV vaccination during experiment 3. Horizontal bars indicate the median of each group. Different letters indicate statistically significant differences (p < 0.05) between groups using Kruskal-Wallis test. (D) Blood B cell concentrations prior to NDV vaccination (“15 days,” left panel), and prior to NDV challenge (“29 days,” right panel). Different letters indicate statistically significant differences (p < 0.05) between groups using Kruskal-Wallis test. Discussion Evaluating the immunosuppressive properties of live-attenuated IBD vaccines is crucial for a precise assessment of their risk-benefit ratio. In this study, alternative strategies to the Ph. Eur. protocol, which is based on lethal viral challenges, were explored to quantify IBDV immunosuppressive properties in the interest of improving ethics and cost. The Ph. Eur. derived in vivo immunosuppression protocol that was implemented provided results in agreement with the official status (mild, intermediate, or intermediate-plus) of the tested vaccines. Accordingly, only mild and intermediate vaccines gave both protection and HI results not differing from NDV vaccine controls. These results are also in agreement with previous data obtained in the Ploufragan laboratory for some of the tested vaccines (16). NDV HI titer patterns were higher in experiment 1 than in experiments 2 and 3. As the presence of anti-coagulants has been linked to differences in values observed during serological testing (7), CaCl2 was added to plasma samples from experiment 3 to artificially induce fibrinogen precipitation allowing coagulation. The lack of impact of CaCl2 addition on HI titers, combined with the comparable titers of NDV vaccine controls from experiments 2 and 3 argue against an effect of EDTA on HI titers. A more likely hypothesis may thus be that a higher efficacy of NDV vaccination in experiment 1 resulted in higher HI titers. As a first alternative strategy to the current in vivo testing scheme, ex vivo infections of primary bursal cells were performed. This strategy was based on the assumption that the varying degrees of immunosuppression were caused by differences in IBDV replicative fitness in the BF (with higher replication levels producing higher immunosuppression) (17, 18) and that those differences may persist during ex vivo infection. Among the seven viruses tested, two of them (I2 and I+1) went against that assumption: while I+1 mimicked I1 replication curve, I2 replicated as well as I+ vaccines. Flow cytometry analyzes of live-infected cells, although revealing virus-specific differences at 16 h.p.i., did not allow differentiating I from I+ vaccines. Those differences were no longer visible at 23 h.p.i., partly due to high inter-individual differences. It was a surprise to see that cells from some individuals presented consistently high percentages of infection irrespective of the infecting viral strain. Higher numbers of biological replicates would be necessary to ascertain those tendencies in the flow cytometry analyses. Similarly, analysis of virus-induced cell death by flow cytometry did not reveal any difference between viral strains (data not shown). Thus, flow cytometry analyses, alone or in combination with ex vivo viral replication studies, did not allow discrimination of IBDV immunosuppressive properties. B cells continuously exit the BF to reach the bloodstream during the first weeks post-hatching, resulting in a steady increase in blood B cells during this period (8). Early studies have shown that IBDV infection at 1 day of age with a pathogenic strain results in a dramatic and durable reduction of blood B cell concentrations (19). Based on these studies, the second alternative strategy consisted in checking whether the reduction in blood B cell concentration would be more severe in the case of I+ vaccines, and if other leukocyte populations could be modified. This strategy revealed striking differences for blood B cells, in particular at 15 days of age. All IBDV-infected groups showed a significant reduction relative to controls, and I+ infected groups showed a significantly more severe reduction than I infected groups, with some birds without any detectable blood B cells. Noteworthy, I1 and I2 infections, which had no effect on NDV HI titers or protection results of experiment 1, induced significant 8- and 28-fold reductions in B cell counts, respectively, when compared to controls. Furthermore, the approximate 3.4-fold reduction between I2 and I1 also appeared significant, making it possible to differentiate the impact of those two vaccines. At 29 days of age, a global increase in blood B cell concentration was observed. Apart from the increase normally observed in control birds, these higher values may reflect, in IBDV-infected groups, the recovery of the BF from IBDV-induced lesions (20, 21). Although all IBDV-infected groups still showed a significant blood B cell depletion compared with control birds, it was no longer possible to completely differentiate I+ from I infected groups. It is worth mentioning that some birds from I+ infected groups still lacked any detectable blood B cell at that time-point. In summary, following circulating B cell counts may help to accurately quantify the impact of IBDV vaccination on the B cell compartment and to distinguish intermediate from intermediate-plus vaccines. It therefore represents an attractive potential strategy, alone or in combination with NDV HI testing, to assess the immunosuppressive properties of IBDV vaccine candidates or field strains in a faster and more ethical way to the current Ph.Eur. procedure, which relies on a lethal NDV infectious challenge. Further investigations, using other IBDV vaccines and IBDV strains with varying degree of pathogenicity are required to establish the threshold of this technique in differentiating intermediate from intermediate-plus vaccines. Data Availability Statement Inquiries concerning raw data availability can be directed to the corresponding author. Ethics Statement The animal study was reviewed and approved by Anses Ploufragan Ethical Committee (Committee Number C2EA-016/ComEth ANSES/ENVA/UPEC) and French Ministry for higher education and research. Author Contributions NE and SS proposed the original idea. SH, J-CR, NE, and SS participated in the design of the study. SS supervised the experiments. CC, CA, MA, AK, SB, and SS participated to experiments and/or to their analysis. CC and SS wrote the manuscript. All authors read, improved, and approved the final manuscript. Funding This work was supported by the French Agency for Veterinary Medicinal Products (CRD number 2015-112). Anses Ploufragan research was supported by Département des Côtes d'Armor and Saint Brieuc Armor Agglomération. Conflict of Interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Publisher's Note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. The authors thank Mélissa Sizorn for her contribution on some ex vivo experiments, and Paul Alun Brown for his advices and the reading of the manuscript. Supplementary Material The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fvets.2022.871549/full#supplementary-material Click here for additional data file. Abbreviations BF Bursa of Fabricius EDTA Ethylene Diamine Tetra Acetic acid EID50 50% Embryo Infective Dose EU European Union HI Hemagglutination Inhibition h.p.i. hours post-infection I Intermediate I+ Intermediate-plus IBD Infectious Bursal Disease IBDV Infectious Bursal Disease Virus ND Newcastle Disease NDV Newcastle Disease Virus Ph. Eur. European Pharmacopeia SPF Specific Pathogen-Free TCID50 50% Tissue Culture Infective Dose. ==== Refs References 1. Gimeno IM Schat KA . Virus-induced immunosuppression in chickens. Avian Dis. (2018) 62 :272–85. 10.1637/11841-041318-Review.1 30339511 2. Delmas B Attoui H Ghosh S Malik YS Mundt E Vakharia VN . ICTV virus taxonomy profile: Birnaviridae. J Gen Virol. (2019) 100 :5–6. 10.1099/jgv.0.001185 30484762 3. Ratcliffe MJH Härtle S . Chapter 4 - B cells, the bursa of fabricius and the generation of antibody repertoires. In: Schat KA Kaspers B Kaiser P editors. Avian Immunology. 2nd ed. Boston, MA: Academic Press (2014). p. 65–89. 4. Eterradossi N Saif YM . Infectious bursal disease. In: Swayne DE Boulianne M Logue CM McDougald LR Nair V Suarez DL editors. Diseases of Poultry. Hoboken: Wiley-Blackwell (2020). p. 257–83. 5. Infectious Bursal disease (Gumboro disease). In: Manual of Diagnostic Tests and Vaccines for Terrestrial Animals. Paris: Organisation mondiale de la santé animale (2021). p. 931–51. 6. Avian infectious bursal disease vaccine (live). In: European Pharmacopoeia (Ph Eur). 10th ed. p. 1217–9. 7. Nilson E Ekholm B Rees Smith B Törn C Hillman M . Calcium addition to EDTA plasma eliminates falsely positive results in the RSR GADAb ELISA. Clin Chim Acta. (2008) 388 :130–4. 10.1016/j.cca.2007.10.021 18023278 8. Seliger C Schaerer B Kohn M Pendl H Weigend S Kaspers B . A rapid high-precision flow cytometry based technique for total white blood cell counting in chickens. Vet Immunol Immunopathol. (2012) 145 :86–99. 10.1016/j.vetimm.2011.10.010 22088676 9. Cubas-Gaona LL Flageul A Courtillon C Briand FX Contrant M Bougeard S . Genome evolution of two genetically homogeneous infectious bursal disease virus strains during passages in vitro and ex vivo in the presence of a mutagenic nucleoside analog. Front Microbiol. (2021) 12 :678563. 10.3389/fmicb.2021.678563 34177862 10. Soubies SM Courtillon C Abed M Amelot M Keita A Broadbent A . Propagation and titration of infectious bursal disease virus, including non-cell-culture-adapted strains, using ex vivo-stimulated chicken bursal cells. Avian Pathol. (2018) 47 :179–88. 10.1080/03079457.2017.1393044 29039212 11. Eterradossi N Toquin D Rivallan G Guittet M . Modified activity of a VP2-located neutralizing epitope on various vaccine, pathogenic and hypervirulent strains of infectious bursal disease virus. Arch Virol. (1997) 142 :255–70. 10.1007/s007050050075 9125042 12. Reed LJ Muench H . A simple method of estimating fifty per cent endpoints. Am J Epidemiol. (1938) 27 :493–7. 10.1093/oxfordjournals.aje.a118408 13. R Core Team. R: A Language and Environment for Statistical Computing. 4.0.3 ed. Vienna: R Foundation for Statistical Computing (2020). 14. Hervé M . RVAideMemoire: Testing and Plotting Procedures for Biostatistics. R package version 0.9-79. ed (2021). 15. Mendiburu Fd . agricolae: Statistical Procedures for Agricultural Research. R package version 1.3-3. ed (2020). 16. Guittet M Le Coq H Picault JP Eterradossi N Bennejean G . Safety of infectious bursal disease vaccines: assessment of an acceptability threshold. Dev Biol Standard. (1992) 79 :147–52.1337527 17. Wang W Huang Y Ji Z Chen G Zhang Y Qiao Y . The full region of n-terminal in polymerase of ibdv plays an important role in viral replication and pathogenicity: either partial region or single amino acid V4I substitution does not completely lead to the virus attenuation to three-yellow chickens. Viruses. (2021) 13 :107. 10.3390/v13010107 33466596 18. Wang W Huang Y Zhang Y Qiao Y Deng Q Chen R . The emerging naturally reassortant strain of IBDV (genotype A2dB3) having segment A from Chinese novel variant strain and segment B from HLJ 0504-like very virulent strain showed enhanced pathogenicity to three-yellow chickens. Transbound Emerg Dis. (2021). 10.1111/tbed.14336 .[Epub ahead of print].34581009 19. Sivanandan V Maheswaran SK . Immune profile of infectious bursal disease: I. Effect of infectious bursal disease virus on peripheral blood T and B lymphocytes of chickens. Avian Dis. (1980) 24 :715–25. 10.2307/1589807 6255926 20. Withers DR Davison TF Young JR . Diversified bursal medullary B cells survive and expand independently after depletion following neonatal infectious bursal disease virus infection. Immunology. (2006) 117 :558–65. 10.1111/j.1365-2567.2006.02332.x 16556270 21. Kim IJ Gagic M Sharma JM . Recovery of antibody-producing ability and lymphocyte repopulation of bursal follicles in chickens exposed to infectious bursal disease virus. Avian Dis. (1999) 43 :401–13. 10.2307/1592637 10494408
PMC009xxxxxx/PMC9179939.txt
==== Front Cureus Cureus 2168-8184 Cureus 2168-8184 Cureus Palo Alto (CA) 35702476 10.7759/cureus.24877 Physical Medicine & Rehabilitation Preventive Medicine Nutrition The Effects of a Pre-workout Supplement on Measures of Alertness, Mood, and Lower-Extremity Power Muacevic Alexander Adler John R Curtis Jason 1 Evans Cassandra 2 Mekhail Veronica 3 Czartoryski Paulina 3 Santana Juan Carlos 4 Antonio Jose 5 1 Exercise and Sports Science, Keiser University, West Palm Beach, USA 2 Nutrition/Exercise and Sports Sciences, Rocky Mountain University of Health Professions, Ft. Lauderdale, USA 3 Nutrition/Exercise and Sports Sciences, Nova Southeastern University, Ft. Lauderdale, USA 4 Fitness and Performance, Institute of Human Performance, Boca Raton, USA 5 College of Health Care Sciences, Nova Southeastern University, Davie, USA Jason Curtis [email protected] 10 5 2022 5 2022 14 5 e2487710 5 2022 Copyright © 2022, Curtis et al. 2022 Curtis et al. https://creativecommons.org/licenses/by/3.0/ This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. This article is available from https://www.cureus.com/articles/95900-the-effects-of-a-pre-workout-supplement-on-measures-of-alertness-mood-and-lower-extremity-power Objective The purpose of this study was to investigate the effects of a multi-ingredient pre-workout supplement (MIPS) on mental and physical performance. Materials and methods Fourteen exercise-trained men (n=7) and women (n=7) completed this randomized, placebo-controlled, double-blind, counterbalanced, crossover trial. Participants consumed either a MIPS or a placebo in a randomized, counterbalanced order. Forty-five minutes after consumption, the following assessments were conducted: psychomotor vigilance test (PVT), Profile of Mood States (POMS), vertical jump test, and heart rate and blood pressure. There was a one-week washout period between assessments. Results Statistically significant differences were observed between the treatment and placebo groups for the PVT (reaction time: treatment: 286 ± 28 ms, placebo: 306 ± 46 ms, p=0.0371) and POMS (i.e., vigor: treatment: 15.2 ± 14.9, placebo: 9.7 ± 9.6, p=0.0403; fatigue: treatment: 1.0 ± 1.1, placebo: 3.3 ± 3.4, p=0.0100). There were no significant differences between groups for the other indices of mood, false starts from the PVT, and vertical jump. Conclusion Based on our findings, the acute consumption of a MIPS produced a significant improvement in a sustained-attention, reaction-timed task as well as measures of vigor and fatigue. performance exercise focus supplement caffeine ==== Body pmcIntroduction Multi-ingredient pre-workout supplements (MIPS) are commonly marketed as “pre-workouts” and have been shown to positively affect mental and physical performance [1-6]. Even though the primary ingredient in most MIPS is caffeine (CAF) [1-6], a multitude of other ingredients is also found in these supplements. In active females, the acute ingestion of MIPS has been shown to improve factors such as focus, anaerobic capacity, and upper body muscular endurance following high-intensity exercise [7]. Energy drinks, which are often consumed prior to training, have demonstrated an improved time to exhaustion as well as improvement in terms of subjective measures of energy, fatigue, and focus [6,8]. A study that examined the effects of a CAF-containing energy drink versus a placebo found that those who consumed the energy drinks had fewer false starts during a psychomotor vigilance test (PVT), suggesting it elicited some improved performance in PVT testing [9]. Other ingredients found in MIPS may also favorably affect mental and physical performance [e.g., tyrosine, theanine, L-citrulline, alpha-glycerophosphocholine (alpha-GPC)]. Cutrufello et al. have explored the effects of a single dose of L-citrulline prior to exercise [6]. Subjects consumed one of three beverages (L-citrulline, watermelon juice, or placebo) one or two hours prior to exercise, and it was found that the single dose of L-citrulline did not have any effect on exercise performance [10]. Creatine supplementation has been shown to enhance cognitive functioning and work as a neuroprotective supplement [higher levels of creatine reduced injuries to the brain and spinal cord due to planned traumatic brain injuries (TBIs) in rat studies] [11]; however, this is due to chronic (daily use) not acute (single occurrence) consumption. Every MIPS has different blends of primary ingredients; thus, it is difficult to make a direct comparison between different MIPS. The MIPS studied in the current investigation consisted of several potentially ergogenic ingredients including CAF, L-citrulline, betaine, beta-alanine, creatine, and alpha-GPC. Different combinations of ingredients in MIPS may produce different results vis-à-vis the mental and physical performance. In light of this, the purpose of this study was to investigate the effects of MIPS on mental and physical performance. Materials and methods Participants Fourteen exercise-trained men (n=7; mean age: 19.9 ± 1.4 years) and women (n=7; mean age 19.9 ± 1.4 years) volunteered for this randomized, placebo-controlled, double-blind, counterbalanced, crossover trial. Subjects reported to the lab on two separate occasions a week apart to participate in testing. In accordance with the Declaration of Helsinki, the Institutional Review Board for Keiser University approved all procedures involving human subjects prior to the beginning of the data collection (IRB# IRB000SP21JC101, Keiser University). This study was approved through an expedited review in February 2021. All participants signed a written informed consent that was approved by the Institutional Review Board of the university (IRB# IRB000SP21JC101) prior to conducting any study-related activities. Daily CAF consumption and exercise history were assessed via self-reporting. Body composition A multi-frequency bioelectrical impedance device (InBody 270) was used on the participants to assess their body composition on their first visit to the laboratory (body mass, fat mass, lean body mass, body fat percentage, and total body water in liters). Each subject was instructed to arrive after having fasted for at least three hours. Participants stood on the device platform with bare feet on the electrodes. They were then instructed by the device to grasp the handles (which contain additional electrodes on the thumb and fingers) while they kept their arms straight and horizontally adducted at approximately 30 degrees. This assessment was completed within approximately one minute. Multi-ingredient pre-workout supplement (MIPS) The MIPS and placebo were donated by Shifted, LLC (Eugene, OR). The placebo comprised maltodextrin. Both the placebo and the MIPS were placed in identical bags with different codes. Both powders were identical in appearance and taste and were mixed with ~8-12 fluid ounces of water in opaque water bottles for consumption. Table 1 lays out the nutritional facts panel of the Shifted MIPS. Table 1 Nutritional facts panel of the Shifted MIPS *Percentage DV is based on a 2,000-calorie diet. **DV not established DV: daily value; MIPS: multi-ingredient pre-workout supplement Supplement facts     Serving size: 1 scoop (30 g)     Servings per container: 20       Amount per serving % DV* Calories 5   Total carbohydrate 1 g <1% Niacin (as nicotic acid) 15 mg 94% Vitamin B6 (as pyridoxine HCL) 1 mg 59% Vitamin B12 (as methylcobalamin) 100 mg 4167% Iron 1 mg 6% Magnesium (red spinach leaf extract and dimagnesium malate) 9 mg 2% Sodium (as pink Himalayan sea salt) 40 mg 2% Potassium (from red spinach leaf extract and potassium chloride) 248 mg 5% L-citrulline 8 g ** Creatine monohydrate 5 g ** Taurine 3 g ** Beta-alanine (as Carnosyn®) 2.5 g ** Betaine anhydrous 2.5 g ** L-tyrosine 2 g ** Red spinach leaf extract (as Oxystorm®) 1 g ** Beetroot extract 1 g ** Alpha-GPC (50%) 300 mg ** Caffeine blend 300 mg **             Caffeine anhydrous (250 mg)   **             zümXR® Delayed-Release Caffeine (50 mg)   ** L-theanine 150 mg ** ElevATP® (ancient peat and apple fruit extract) 150 mg ** Pink Himalayan sea salt 100 mg ** AstraGin® [astragalus membranaceus (root) extract and Panax notoginseng (root) extract] 25 mg ** BioPerine® (black pepper fruit extract) 5 mg ** Other ingredients: citric acid, natural flavor, malic acid, silicon dioxide, calcium silicate, sucralose, spirulina powder (color)     Testing procedures Participants arrived at the lab on two separate occasions separated by one week between the hours of 1200 and 1400. The subjects read and signed the consent form instructing them to not change dietary or exercise habits during the treatment period. The subjects’ physical characteristics were only assessed on the first visit; during both visits, they filled out the Profile of Mood States (POMS) questionnaire and performed the five-minute PVT and a three-trial max vertical jump test using a vertec. The subjects either consumed the MIPS drink or the placebo (every odd-number subject received the powder from bags with the first code, and every even-number subject received powder from the bags with the second code). On the second visit, they received a coded bag of powder that they had not received on the first visit. The research team was blinded to the treatment product and the placebo, and the company that had sent the bags of MIPS or placebo did not release the information of the codes to the research team until all trials were complete. Forty-five minutes post-consumption (during which all participants were instructed not to leave the lab or consume any food or drink), the subjects filled out the POMS questionnaire and performed the five-minute PVT and the three-trial vertical jump, and then the post-test heart rate (HR) and blood pressure (BP) were measured after all other tests. Profile of Mood States (POMS) The POMS is a 65-word standard validated psychological test to assess transient and distinct mood states [12]. The test listed words such as “angry”, “tense”, “lively”, and 62 other words commonly used to describe different mood states. Next to each word is a drop-down menu with the words “how I have felt” above them. The options in the drop-down menu were “not at all”, “a little”, “moderately”, “quite a bit”, and “extremely”. This test scores total mood disturbance, anger, depression, fatigue, tension, and vigor. Those participants who have lower scores would be categorized as having more stable mood profiles [12]. Psychomotor vigilance test (PVT) A PVT was administered using standard electronic tablets (Apple iPads; Apple Inc., Cupertino, CA) and automated testing software (Vigilance Buddy by Research Buddies). The PVT test displays stimuli (i.e., a number on the screen). As soon as the participants saw the number appear in the middle of the screen, they then touched the screen as fast as they could to get the lowest possible time. Each participant was given instructions on how the test works, and how to properly do the test. The subjects were instructed to respond as fast as possible to the stimulus while not responding prematurely [13]. The PVT assesses vigilant attention and activates the visual cortex, motor cortex, and prefrontal cortex [14]. Statistical analysis The GraphPad Prism 6 statistical software (GraphPad Software Inc., San Diego, CA) was used to perform all statistical analyses. All data for this study is presented as mean ± standard deviation (SD). A paired t-test was performed to determine whether statistically significant differences (p<0.05) existed between the treatment and the placebo (i.e., for PVT and vertical jump); a Wilcoxon matched-pairs signed-rank test was performed (due to the types of data we analyzed) to determine if differences existed for the various mood assessments. Results The descriptive characteristics (Table 2) of the subjects were as follows: 14 total exercise-trained individuals, men (n=7) and women (n=7) (mean ± SD: age: 19.9 ± 1.4 years; height: 168.2 ± 11.3 cm; body mass: 68.9 ± 10.7 kg). CAF consumption for the treatment was approximately 4 mg per kg based on the mean body mass of the subjects. Table 2 Physical characteristics N=14, seven males, seven females SD: standard deviation (SD) Variables Values (mean ± SD) Age (years) 19.9 ± 1.4 Height (cm) 168.2 ± 11.3 Body mass (kg) 68.9 ± 10.7 Lean body mass (kg) 54.9 ± 12.5 Fat mass (kg) 14.0 ± 5.5 % body fat 21.0 ± 9.0 Total body water (liters) 40.2 ± 9.2 Total number of years of training 6.3 ± 3.5 Average hours of aerobic exercise per week 3.6 ± 3.5 Average hours of resistance exercise per week 7.4 ± 4.9 Other exercises per week (exercises other than resistance training or aerobic training) 1.8 ± 2.8 Average caffeine consumed per day 207 ± 112 HR and BP were measured after testing at 45-minute post-consumption (Table 3), and the results showed no differences between the treatment and the placebo. Table 3 Heart rate and blood pressure Heart rate and blood pressure were assessed 45-min post-consumption of the treatment and placebo. There were no significant differences between the groups except for diastolic BP, which was higher in the placebo group BP: blood pressure; SD: standard deviation   Treatment, mean ± SD Placebo, mean ± SD P-value Heart rate (bpm) 80 ± 15 79 ± 14 0.8947 Systolic BP (mmHg) 138 ± 15 136 ± 19 0.6696 Diastolic BP (mmHg) 80 ± 11 87 ± 16 0.0077 The results of the POMS (Table 4) are as follows: total mood disturbance score (TMDS): treatment: 1.1 ± 1.5, placebo: 6.2 ± 7.1, p=0.1912; anger: treatment: 2.0 ± 2.0, placebo: 1.4 ± 1.5, p=0.4257; confusion: treatment: 3.4 ± 3.6, placebo: 4.3 ± 4.5, p=0.1386; depression: treatment: 1.2 ± 1.3, placebo: 1.1 ± 1.1, p=0.7701; fatigue: treatment: 1.0 ± 1.1, placebo: 3.3 ± 3.4, p=0.0100; tension: treatment: 8.6 ± 8.4, placebo: 5.9 ± 6.1, p=0.1345; vigor: treatment: 15.2 ± 14.9, placebo: 9.7 ± 9.6, p=0.0403. Table 4 Profile of Mood States (POMS) Self-reported fatigue and vigor were significantly lower and higher, respectively, in the treatment group vs. the placebo TMDS: total mood disturbance score; SD: standard deviation   Treatment (mean ± SD) Placebo (mean ± SD) P-value TMDS 1.1 ± 1.5 6.2 ± 7.1 0.1912 Anger 2.0 ± 2.0 1.4 ± 1.5 0.4257 Confusion 3.4 ± 3.6 4.3 ± 4.5 0.1386 Depression 1.2 ± 1.3 1.1 ± 1.1 0.7701 Fatigue 1.0 ± 1.1 3.3 ± 3.4 0.0100 Tension 8.6 ± 8.4 5.9 ± 6.1 0.1345 Vigor 15.2 ± 14.9 9.7 ± 9.6 0.0403 The results of the PVT (Table 5) and vertical jump test (Table 5) were as follows: reaction time: treatment: 286 ± 28 ms, placebo: 306 ± 46 ms, p=0.0371; number of false starts: treatment: 3.1 ± 2.4, placebo: 2.5 ± 2.8, p=0.4473; vertical jump (cm): treatment: 54 ± 8, placebo: 54 ± 10, p=0.4927. Table 5 Psychomotor vigilance and vertical jump The treatment group had a faster reaction time vs. the placebo SD: standard deviation   Treatment (mean ± SD) Placebo (mean ± SD) P-value Reaction time (ms) 286 ± 28 306 ± 46 0.0371 Number of false starts 3.1 ± 2.4 2.5 ± 2.8 0.4473 Vertical jump (cm) 54 ± 8 54 ± 10 0.4927 Discussion The current study found an improvement in sustained attention (i.e., mean reaction time on the PVT) as well as levels of perceived fatigue (perception of feeling tired) and vigor (perception of feeling energetic) with the consumption of MIPS; however, there was no effect on vertical jump performance. It should be noted that CAF is a common ingredient in both energy drinks and MIPS; however, there are other ingredients that may also play a role vis-à-vis an ergogenic effect (e.g., L-citrulline, tyrosine, alpha-GPC) [15-30]. The strengths of this study are that it was a double-blind, crossover study. Since all participants belonged to both groups, there were no differences associated with subjects belonging to different groups. This study provides additional insights into ergogenic ingredients found in MIPS. Adenosine receptors in the central nervous receptors (CNS) are the primary targets for antagonism when CAF is consumed [15]. In a study investigating the effects of CAF and vigilance on the Special Forces personnel by McLellan et al., they found that during 27 hours of wakefulness, vigilance was maintained, and running performance improved [17]. Our current investigation found that 45 minutes after the consumption of a CAF-containing MIPS, the reaction time of the treatment group was significantly better than the placebo group. Our investigation used a dose within the normal range of 3-6 mg per kg of body weight; 300 mg of CAF equates to approximately 4 mg of caffeine per kg of body weight in the current study. This shows that in sports, jobs, and other activities that require elevated levels of vigilance, the consumption of MIPS would be a beneficial ergogenic aid. Our study also found significant improvements in vigor and a reduction in fatigue when comparing the POMS scores between the treatment group and the placebo group. This would also endorse the ability to be more vigilant gained by consuming a MIPS product. The MIPS contained other potentially ergogenic ingredients. L-citrulline is a non-essential amino acid and a precursor for L-arginine. Several studies have observed improvements in exercise performance following supplementation with L-citrulline. For instance, supplementing with 6-8 grams of citrulline malate 40-60 minutes prior to exercise improves muscular endurance (i.e., increased repetitions) [17]. Several clinical trials have found alpha-GPC supplementation to improve cognitive impairments in patients with cerebral disorders such as dementia and Alzheimer’s disease [18]. Other studies have evaluated the effects of alpha-GPC on physical and psychomotor performance. Marcus et al. have reported no significant differences in psychomotor vigilance task results between alpha-GPC (two capsules of either 250 or 500 mg/day for seven days) and placebo groups [16]. Improvements in cognitive function are more likely in individuals with cognitive impairments compared to healthy, exercise-trained individuals [18]. In a recent study, supplementation of 600 mg alpha-GPC for six days showed improvements in lower body force production compared to placebo [20]. Additionally, 600 mg of alpha-GPC administration 90 minutes prior to resistance exercise showed improvements in peak bench press force, but not peak power or rate of force development [21]. These studies suggest that alpha-GPC administration could improve physical performance, but not cognitive performance, in healthy, exercise-trained individuals. Tyrosine and theanine also show promise as potential ergogenic aids that may enhance cognitive functions. Older adults were found to have higher plasma tyrosine concentrations in a dose-dependent fashion [22]. Interestingly, working memory in older adults was best at a lower dose (i.e., 100 mg per kg body weight) in comparison to higher doses (i.e., 150 and 200 mg per kg body weight) [22]. An investigation by Kahathuduwa et al. found that a 200-mg dose of theanine improved cognitive measures of selective attention [12] with CAF also demonstrating similar effects. Moreover, the combination of caffeine and theanine ostensibly had additive effects on attention [23]. One of the more intriguing studies involved the acute consumption of caffeine, theanine, and tyrosine [24]. Twenty-four male former collegiate athletes completed mental and physical performance assessments using a Makoto Arena testing device that assessed physical as well as mental performance. Compared to the placebo, the supplement combination of CAF, theanine, and tyrosine improved movement accuracy [24]. Rhodiola rosea is another supplement that is purported to enhance cognitive function [25]. Rhodiola is not one compound but a myriad of different compounds that includes flavonoids and phenylethanol (i.e., Spanish sage) [25]. Scholey et al. discovered that a 333-mg dose of sage improved memory in a group of older adults [22]. Tildesley et al. showed that sage improved immediate word recall [23]. Notwithstanding the intriguing data from these trials, it is evident that more research is needed regarding Rhodiola and/or its constituent compounds. Other ingredients found in MIPS that purported to show promise for enhancing physical and cognitive function include creatine monohydrate, beta-alanine, and betaine. However, it would appear that chronic supplementation is required to elicit such an effect with these supplements. Current research suggests that creatine and beta-alanine require a chronic loading phase of approximately four weeks to meaningfully impact body stores of creatine and carnosine, respectively [28]. Although MIPS contained 2500 mg of beta-alanine and 5000 mg of creatine monohydrate, it is unlikely that these ingredients contributed to any of the effects seen in this study due to the necessity of chronic use [29]. An investigation by Borsook et al. reported increased feelings of well-being and decreased feelings of fatigue due to betaine supplementation in a clinical population [4]. These studies suggest that betaine may improve mood. It is plausible to think betaine could contribute to reported improvements in feelings of vigor, and fatigue; however, more research is needed to validate this. There are several limitations to this study. We observed elevated BP and HR in both placebo and treatment groups with no significant difference between the groups. This could be due to the “White Coat Syndrome” or the fact that we measured these diagnostics immediately after the vertical jump protocols. Additionally, we only had a total of 14 subjects for this study. A study involving a larger sample size would allow us to discern any differences between sexes or find other significant differences in other measures that may result from a greater sampling number. Also, we did not look at the effects of CAF by genotypes. However, the results of this study do support the findings of previous research regarding the effects CAF products have on PVT. Conclusions The acute consumption of MIPS significantly improved sustained attention as well as perceived vigor and fatigue. It is not clear if there is a specific single ingredient (e.g., CAF) that contributed to this improvement or if it was the result of a combination of ingredients (i.e., CAF, theanine, alpha-GPC, L-tyrosine, etc.). Since there are multiple ingredients within the MIPS, it is not possible to determine which ingredient(s) contributed the most to the ergogenic effect. CAF is perhaps the only ingredient that is present in all MIPS; thus, it behooves future researchers to include a CAF-only positive control. Nevertheless, one can sensibly conclude that the MIPS, as shown in the current investigation, may serve to enhance performance in tasks (e.g., driving) or sports (e.g., baseball) that require sustained attention. We thank Shifted LLC (Eugene, OR) for providing the multi-ingredient pre-workout supplement as well as the placebo. The Institute of Human Performance (Boca Raton, FL) served as the clinical site for the investigation. Human Ethics Animal Ethics Consent was obtained or waived by all participants in this study. Keiser University IRB issued approval IRB000SP21JC101 Animal subjects: All authors have confirmed that this study did not involve animal subjects or tissue. The authors have declared that no competing interests exist. ==== Refs References 1 Dose response effects of a caffeine-containing energy drink on muscle performance: a repeated measures design J Int Soc Sports Nutr Del Coso J Salinero JJ González-Millán C Abián-Vicén J Pérez-González B 21 9 2012 22569090 2 The effect of a multi-ingredient pre-workout supplement on time to fatigue in NCAA Division I cross-country athletes (Online) Nutrients Fye H Pass C Dickman K Bredahl E Eckerson J Siedlik J 2021 3 Effects of energy drink major bioactive compounds on the performance of young adults in fitness and cognitive tests: a randomized controlled trial J Int Soc Sports Nutr Kammerer M Jaramillo JA García A Calderón JC Valbuena LH 44 11 2014 25389379 4 Effects of a pre-workout energy drink supplement on upper body muscular endurance performance Int J Exerc Sci Magrini MA Colquhoun RJ Dawes JJ Smith DB 667 676 9 2016 https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5154715/ 27990227 5 Preexercise energy drink consumption does not improve endurance cycling performance but increases lactate, monocyte, and interleukin-6 response J Strength Cond Res Phillips MD Rola KS Christensen KV Ross JW Mitchell JB 1443 1453 28 2014 24172719 6 Improved time to exhaustion following ingestion of the energy drink Amino Impact™ J Int Soc Sports Nutr Walsh AL Gonzalez AM Ratamess NA Kang J Hoffman JR 14 7 2010 20398312 7 The acute effects of a multi-ingredient pre-workout supplement on resting energy expenditure and exercise performance in recreationally active females J Int Soc Sports Nutr Cameron M Camic CL Doberstein S Erickson JL Jagim AR 1 15 2018 29311763 8 Thermogenic effect of meltdown RTD energy drink in young healthy women: a double blind, cross-over design study Lipids Health Dis Rashti SL Ratamess NA Kang J Faigenbaum AD Chilakos A Hoffman JR 57 8 2009 20017916 9 The effects of an energy drink on measures of cognition and physical performance J Exerc Physiol Online Evans C Mekhail V Kaminski J Peacock C Tartar J Santana JC 75 82 24 2021 https://go.gale.com/ps/i.do?id=GALE%7CA667938350&sid=googleScholar&v=2.1&it=r&linkaccess=abs&issn=10979751&p=HRCA&sw=w&userGroupName=anon%7Ec0bb80d2 10 The effect of l-citrulline and watermelon juice supplementation on anaerobic and aerobic exercise performance J Sports Sci Cutrufello PT Gadomski SJ Zavorsky GS 1459 1466 33 2015 https://doi.org/10.1080/02640414.2014.990495 25517106 11 International Society of Sports Nutrition position stand: safety and efficacy of creatine supplementation in exercise, sport, and medicine J Int Soc Sports Nutr Kreider RB Kalman DS Antonio J 18 14 2017 28615996 12 Manual Profile of Mood States McNair DM Lorr M Droppleman LF 1971 https://www.statisticssolutions.com/free-resources/directory-of-survey-instruments/profile-of-mood-states-poms/ 13 Validity and sensitivity of a brief psychomotor vigilance test (PVT-B) to total and partial sleep deprivation Acta Astronaut Basner M Mollicone D Dinges DF 949 959 69 2011 22025811 14 The neural basis of the psychomotor vigilance task Sleep Drummond SP Bischoff-Grethe A Dinges DF Ayalon L Mednick SC Meloy MJ 1059 1068 28 2005 https://watermark.silverchair.com/sleep-28-9-1059.pdf?token=AQECAHi208BE49Ooan9kkhW_Ercy7Dm3ZL_9Cf3qfKAc485ysgAAAs8wggLLBgkqhkiG9w0BBwagggK8MIICuAIBADCCArEGCSqGSIb3DQEHATAeBglghkgBZQMEAS4wEQQMbcU5Tg_mXXZ6t0isAgEQgIICgjVna1iEcD-QIQCGXZ_ZlqKNl0TW11n4qlnsiNxItMgg7qSWUctzlAEf-OTevMzPMwV4M34K43mJMLEry5-LESFusnoSN1j0vFXgVFHVuohZC3kOSTs-kR4pymEaZgoOSsSkNpbFi5rgiM_74xJjbRER9ZxmXCwaBXmwNgNVt7ipg1gfTXnK5gXo7iJT7u6CjqTmuUxsYBvzAT9DL2BmwESdItvDVlU6DwHfVWrDdmFGjA8V2AFxcmb0182v9lqjtbYP4moM-U-gxtk1czcvzyB2yVuAUxrQpvAliuSClmn6JkcAn4hN5x2FilBOB83q7V_s0kgAQqXh1Lqadlldq-fNXWrFzQNEfY7lGkaOPV6tgddbLwV1sxWSyCyrCYy8pJ3g4CVWDN-DnY0KagXU-0cLY4-criZgpT1-KQXaORqqLn37iz02YhQAl5uGuPAuDPZaE0h22xWtol7RRRtDVzOQhHtougxxkqTrfV0uMTv46_QrMPo4IDEOObeNx0SoFuSxSUqG9vwl50ZuwwOfrAx7Ug7n04zR3dNEylLATMFn5LMl41NCn032QQVtavt29j3SUrNx7DjR3ymdPVYZzJjK2on9RZQH6H_SmUX1Mz5rKE606J4fers-DKn6letLIoPBS1mFcHa2FPmByyIyLeHGzQiuTIFIf35EXSCth8nN1dF0PWdfWlnRYqL3YO6Z0Kw2sUfXZfNWdgRyrqqRKO8SH72aWU1ZTjNZcEw8zD_vZbhmP2kX1DYZ4XZX_-maD4m5LYSJQ5OrVDhfV-vnTnFsXOf7VNvyOrEG-DquPqWQUgEGc7QmjeCjqjMLiTQ0QGQSoUyz0kXmCgg-wrdTBfi6vg 16268374 15 Caffeine eliminates psychomotor vigilance deficits from sleep inertia Sleep Van Dongen HP Price NJ Mullington JM Szuba MP Kapoor SC Dinges DF 813 819 24 2001 11683484 16 Caffeine maintains vigilance and improves run times during night operations for Special Forces Aviat Space Environ Med McLellan TM Kamimori GH Voss DM Bell DG Cole KG Johnson D 647 654 76 2005 https://www.ingentaconnect.com/content/asma/asem/2005/00000076/00000007/art00005 16018347 17 Acute effect of citrulline malate on repetition performance during strength training: a systematic review and meta-analysis Int J Sport Nutr Exerc Metab Vårvik FT Bjørnsen T Gonzalez AM 350 358 31 2021 34010809 18 Choline alphoscerate in cognitive decline and in acute cerebrovascular disease: an analysis of published clinical data Mech Ageing Dev Parnetti L Amenta F Gallai V 2041 2055 122 2001 11589921 19 Evaluation of the effects of two doses of alpha glycerylphosphorylcholine on physical and psychomotor performance J Int Soc Sports Nutr Marcus L Soileau J Judge LW Bellar D 39 14 2017 29042830 20 The effect of 6 days of alpha glycerylphosphorylcholine on isometric strength J Int Soc Sports Nutr Bellar D LeBlanc NR Campbell B 42 12 2015 26582972 21 Acute supplementation with alpha-glycerylphosphorylcholine augments growth hormone response to, and peak force production during, resistance exercise J Int Soc Sports Nutr Ziegenfuss T Landis J Hofheins J 0 5 2008 22 Dose-dependent effects of oral tyrosine administration on plasma tyrosine levels and cognition in aging Nutrients van de Rest O Bloemendaal M de Heus R Aarts E 1 5 9 2017 23 Acute effects of theanine, caffeine and theanine-caffeine combination on attention Nutr Neurosci Kahathuduwa CN Dassanayake TL Amarakoon AM Weerasinghe VS 369 377 20 2017 26869148 24 Effects of acute caffeine, theanine and tyrosine supplementation on mental and physical performance in athletes J Int Soc Sports Nutr Zaragoza J Tinsley G Urbina S 56 16 2019 31771598 25 Acute effects of dietary constituents on motor skill and cognitive performance in athletes Nutr Rev Baker LB Nuccio RP Jeukendrup AE 790 802 72 2014 25400063 26 An extract of Salvia (sage) with anticholinesterase properties improves memory and attention in healthy older volunteers Psychopharmacology (Berl) Scholey AB Tildesley NT Ballard CG Wesnes KA Tasker A Perry EK Kennedy DO 127 139 198 2008 18350281 27 Salvia lavandulaefolia (Spanish sage) enhances memory in healthy young volunteers Pharmacol Biochem Behav Tildesley NT Kennedy DO Perry EK Ballard CG Savelev S Wesnes KA Scholey AB 669 674 75 2003 https://www.sciencedirect.com/science/article/abs/pii/S0091305703001229 12895685 28 The absorption of orally supplied beta-alanine and its effect on muscle carnosine synthesis in human vastus lateralis Amino Acids Harris RC Tallon MJ Dunnett M 279 289 30 2006 16554972 29 International Society of Sports Nutrition position stand: beta-alanine J Int Soc Sports Nutr Trexler ET Smith-Ryan AE Stout JR 30 12 2015 26175657 30 The biochemical basis of betaine-glycocyamine therapy Ann West Med Surg Borsook H Borsook ME 825 829 5 1951 https://www.ncbi.nlm.nih.gov/pubmed/14878413 14878413
PMC009xxxxxx/PMC9326105.txt
==== Front J Med Imaging (Bellingham) J Med Imaging (Bellingham) JMIOBU JMI Journal of Medical Imaging 2329-4302 2329-4310 Society of Photo-Optical Instrumentation Engineers 35911208 10.1117/1.JMI.9.4.047501 JMI-22101GR 22101GR Digital Pathology Paper Pilot study to evaluate tools to collect pathologist annotations for validating machine learning algorithms https://orcid.org/0000-0002-7089-0534 Elfer Katherine [email protected] Dudgeon Sarah [email protected] Garcia Victor [email protected] Blenman Kim [email protected] Hytopoulos Evangelos [email protected] Wen Si [email protected] Li Xiaoxian [email protected] Ly Amy [email protected] Werness Bruce [email protected] https://orcid.org/0000-0002-5183-4131 Sheth Manasi S. [email protected] https://orcid.org/0000-0001-7599-6162 Amgad Mohamed [email protected] Gupta Rajarsi [email protected] Saltz Joel [email protected] https://orcid.org/0000-0002-7536-1746 Hanna Matthew G. [email protected] Ehinger Anna [email protected] Peeters Dieter [email protected] Salgado Roberto [email protected] https://orcid.org/0000-0001-7332-1620 Gallas Brandon D. a*[email protected] a United States Food and Drug Administration, Center for Devices and Radiological Health, Office of Science and Engineering Laboratories, Division of Imaging Diagnostics & Software Reliability, Silver Spring, Maryland, United States b National Institutes of Health, National Cancer Institute, Division of Cancer Prevention, Cancer Prevention Fellowship Program, Bethesda, Maryland, United States c Yale University Computational Biology and Bioinformatics, New Haven, Connecticut, United States d Yale New Haven Hospital, Center for Outcomes Research and Evaluation, New Haven, Connecticut, United States e School of Medicine, Yale Cancer Center, Department of Internal Medicine, Section of Medical Oncology, New Haven, Connecticut, United States f Yale University, School of Engineering and Applied Science, Department of Computer Science, New Haven, Connecticut, United States g iRhythm Technologies Inc., San Francisco, California, United States h Emory University School of Medicine, Department of Pathology and Laboratory Medicine, Atlanta, Georgia, United States i Massachusetts General Hospital, Boston, Massachusetts, United States j Inova Health System Department of Pathology, Falls Church, Virginia, United States k Arrive Bio LLC, San Francisco, California, United States l United States Food and Drug Administration (FDA), Center for Devices and Radiologic Health, Office of Product Evaluation and Quality, Office of Clinical Evidence and Analysis, Division of Biostatistics, White Oak, Maryland, United States m Northwestern University Feinberg School of Medicine, Department of Pathology, Chicago, Illinois, United States n SUNY Stony Brook Medicine, Department of Biomedical Informatics, Stony Brook, New York, United States o SUNY Stony Brook Medicine, Department of Pathology, Stony Brook, New York, United States p Memorial Sloan Kettering Cancer Center, New York, New York, United States q Lund University, Laboratory Medicine, Region Skåne, Department of Genetics and Pathology, Lund, Sweden r Sint-Maarten Hospital, Department of Pathology, Mechelen, Belgium s University of Antwerp, Department of Biomedical Sciences, Antwerp, Belgium t Peter Mac Callum Cancer Centre, Division of Research, Melbourne, Australia u GZA-ZNA Hospitals, Department of Pathology, Antwerp, Belgium * Address all correspondence to Brandon D. Gallas, [email protected] 27 7 2022 7 2022 27 7 2023 9 4 04750114 4 2022 28 6 2022 © 2022 The Authors 2022 The Authors https://creativecommons.org/licenses/by/4.0/ Published by SPIE under a Creative Commons Attribution 4.0 International License. Distribution or reproduction of this work in whole or in part requires full attribution of the original publication, including its DOI. Abstract. Purpose Validation of artificial intelligence (AI) algorithms in digital pathology with a reference standard is necessary before widespread clinical use, but few examples focus on creating a reference standard based on pathologist annotations. This work assesses the results of a pilot study that collects density estimates of stromal tumor-infiltrating lymphocytes (sTILs) in breast cancer biopsy specimens. This work will inform the creation of a validation dataset for the evaluation of AI algorithms fit for a regulatory purpose. Approach Collaborators and crowdsourced pathologists contributed glass slides, digital images, and annotations. Here, “annotations” refer to any marks, segmentations, measurements, or labels a pathologist adds to a report, image, region of interest (ROI), or biological feature. Pathologists estimated sTILs density in 640 ROIs from hematoxylin and eosin stained slides of 64 patients via two modalities: an optical light microscope and two digital image viewing platforms. Results The pilot study generated 7373 sTILs density estimates from 29 pathologists. Analysis of annotations found the variability of density estimates per ROI increases with the mean; the root mean square differences were 4.46, 14.25, and 26.25 as the mean density ranged from 0% to 10%, 11% to 40%, and 41% to 100%, respectively. The pilot study informs three areas of improvement for future work: technical workflows, annotation platforms, and agreement analysis methods. Upgrades to the workflows and platforms will improve operability and increase annotation speed and consistency. Conclusions Exploratory data analysis demonstrates the need to develop new statistical approaches for agreement. The pilot study dataset and analysis methods are publicly available to allow community feedback. The development and results of the validation dataset will be publicly available to serve as an instructive tool that can be replicated by developers and researchers. Keywords: digital pathology reader studies reader variability trothing artificial intelligence machine learning FDA Office of Women’s HealthFDA-OWH-2021-Gallas DOE | Oak Ridge Institute for Science and Education (ORISE)https://doi.org/10.13039/100006229 running-headElfer et al.: Pilot study to evaluate tools to collect pathologist annotations for validating machine learning… ==== Body pmc1 Introduction Recent advancements in the fields of digital pathology and artificial intelligence and machine learning (AI/ML) have the potential to transform the field of pathology by increasing the speed and accuracy of diagnosis.1–5 However, before implementing novel AI/ML algorithms for clinical use, their performance must be validated against a reference standard (ground truth) to assess their accuracy.6 In cases where an appropriate independent reference standard is not available, the reference standard may be established by human experts. Using human experts as a reference standard occurs in fields that rely on the interpretation of multi-factored visual information, like pathology.5,7 Using pathologist annotations as a reference standard is susceptible to interobserver variability.8–10 Here, as is often done in pathology, we use “annotations” to refer to any marks, segmentations, measurements, or labels a pathologist adds to a report, image, region of interest (ROI), or biological feature/target. Potential sources of variability in pathologist annotations include preanalytic and analytic variables. Preanalytic sources may include variability from tissue acquisition, tissue processing, tissue section thickness, tissue staining, slide digitization, and viewing digital displays.11–13 Analytic variability may result from differences in knowledge, residency training, subspecialty expertise, and experience as a practicing pathologist.7,14 This variability is compounded in validation studies when these differences are not standardized or captured in the data-collection process, which is further complicated by variable numbers of annotating pathologists between studies.15 Moreover, there is a lack of consensus agreement on the statistical methods for calculating interpathologist variability, which makes determining whether an AI/ML algorithm meets the acceptance criteria of performance difficult to ascertain. Therefore, there is a need to establish standardized workflows for efficiently collecting and using pathologist annotations and to construct the statistical methods needed to establish agreement between pathologists. One example of where AI/ML algorithms may improve pathologist evaluation is in the estimation of the density of stromal tumor-infiltrating lymphocytes (sTILs) in triple-negative breast cancer (TNBC). sTILs density is estimated by pathologist assessment of hematoxylin and eosin (H&E) stained tumor-associated stromal tissue.16–18 In TNBC, the density of sTILs is a quantitative, prognostic, and predictive biomarker.19–22 High levels of sTILs density correlate with improved survival without and with therapy and predict positive response to therapy.19,23–25 Additional evidence indicates that the density of sTILs is a viable biomarker for other tumor types, such as nonsmall cell lung cancer, colorectal cancer, and prostate cancer.17,26–29 Yet, density evaluations of sTILs by pathologists are difficult and have notable variability, which makes developing a validation dataset for AI/ML algorithms for sTILs density estimation difficult.26,30–32 While flow cytometry, immunohistochemistry evaluation, and quantitative multiplex immunofluorescence also characterize sTILs in tissue, these approaches can be complex, time-consuming, expensive, and consume large quantities of tissue compared with estimating sTILs density in H&E slides during routine diagnostic examination.29,33,34 Therefore, a validation dataset using pathologist annotations on H&E slides will be accessible to the largest number of users. One method to develop such a validation dataset is to collect annotations from multiple pathologists on multiple cases to understand and improve the reference standard. To maximize the efficiency of the annotation process, we proposed a standardized workflow and methodology to create a validation dataset of pathologist-annotated sTILs density estimates in H&E stained breast cancer biopsy slides.35 We launched a pilot study in February 2020 and collected pathologist annotations on an optical microscope platform and two digital whole slide image (WSI) web-viewer platforms through May 2021. The microscope platform (eeDAP: an evaluation environment for digital and analog pathology) is the reference technology.36 Pathologists are familiar with this technology, and it does not suffer from the limitations of slide digitization. In addition to the microscope platform, we used two digital image viewing and manipulation platforms, PathPresenter37,38 and caMicroscope.39 These platforms enable data collection worldwide regardless of physical proximity to the slides. An overview of this work can be found in an earlier publication.35 This manuscript presents initial results and analyses from a pilot study of the high throughput truthing (HTT) project. We discuss modifications to the original protocol and data-collection platforms made during and since the pilot study. These modifications were necessary as gaps were identified from pathologist feedback and data analysis. Such changes are the purpose of a pilot study and would not be permitted in a pivotal study. We also explore the validity of our quantitative assessment methods using the annotations collected on the PathPresenter digital platform, which will inform the development of statistical methods for interpathologist agreement using annotations. This assessment will inform the creation of a pivotal validation dataset (slides, images, annotations, and statistical methods) for the evaluation of AI/ML algorithms in digital pathology fit for a regulatory purpose that can be replicated by other researchers and developers. 2 Methods 2.1 Purpose of the Pilot Study The purpose of our pilot study (or any pilot study) is to have an opportunity to evaluate the workflows and tools that will be used in the final pivotal study. Modifications to protocols and data-collection platforms can be made during and after the pilot study without spoiling the goals and hypotheses of a pivotal study. Here, we discuss issues, gaps, and inconsistencies identified from pathologist feedback and data analysis. We also discuss the specific changes made to our original protocol and data-collection tools to address these issues. 2.2 Clinical Use Case: sTILs Density Estimates The clinical use case of this work is the annotation of sTILs in breast cancer H&E slides. We digitized 64 slides of invasive ductal carcinoma core biopsies prepared at a single site by recutting sections from formalin-fixed paraffin-embedded blocks. The slides were scanned on a Hamamatsu Nanozoomer-RS 2.0 (0.23  μm/pixel=40× magnification). Each slide represents a single case of breast cancer. A collaborating pathologist preselected ten 500  μm×500  μm regions of interest (ROIs) per slide for 640 total unique ROIs. The ROIs were selected to sample tumor and nontumoral tissue.35 For this pilot study, the number of ROIs was selected to provide both a representation of breast cancer tissue and sTILs variability and serve as the basis for sizing the pivotal study. The 64 cases in the pilot study were subdivided into eight batches of eight cases each with ten ROIs per case. Therefore, each batch consisted of 80 ROIs, and the full study had 640 ROIs. This subdivision into batches allowed for easier workload management and slide handling. Pathologists annotated each ROI by providing comments on the sTILs content through a series of input tasks. The original description of the pilot study included two annotation tasks: label the ROI and, if the ROI is of a type that is evaluable for sTILs, estimate the density of sTILs (0% to 100%). The primary characteristic for an ROI to be evaluable is whether it contains tumor-associated stroma. With assistance from our collaborating pathologists, we created four options for labeling the ROI. Two of the label categories (“intratumoral stroma” and “invasive margin”) are considered evaluable for sTILs density estimation.16,40 If a pathologist labels an ROI with the other two labels (“tumor with no intervening stroma” or “other region”), the ROI is not evaluable for sTILs. We added the task of estimating the percent of tumor-associated stroma after the launch of the pilot study. Percent tumor-associated stroma enables evaluation of an algorithm’s ability to identify tumor-associated stroma, the tissue where sTILs density is to be evaluated. This is a modification from the protocol previously described.35 The sTILs density is the percent of area of tumor-infiltrating lymphocytes within the area of tumor-associated stroma.16 It is calculated as sTILs Density=(Area of sTILsArea of tumor-associated stroma)×100%. Following emerging evidence and recommendations from clinical experts, the sTILs density estimates can be grouped into three bins corresponding to anticipated patient management decisions and outcomes: “low” or 0% to 10%; “moderate” or 11% to 40%; and “high” or 41% to 100%.18,30,31 We provided pathologists with a visual reference created by Salgado et al.16 to help pathologists calibrate their estimates of the sTILs density. The reference sheet pairs representative ROIs of a range of sTILs densities in H&E images with a black-and-white image of that ROI with the sTILs digitally sectioned from the background content. 2.3 Recruitment of Readers and Description of Data Collection Technology Following an IRB exempt determination protocol, we crowdsourced volunteer pathologists by recruiting at in-person conferences, online research forums, and emailing pathology communities for distribution to their networks. Crowdsourcing provides open access for data collection, enlisting the services of a large number of volunteers for the data collection process. These crowdsourced pathologists completed data-collection tasks via two modalities: an optical light microscope (eeDAP36) and two digital WSI viewing platforms (caMicroscope39 and PathPresenter37). Pathologists recruited in-person were directed to the eeDAP platform, and pathologists recruited online were directed to choose their preferred digital platform, PathPresenter, or caMicroscope. Pathologists were allowed to use multiple platforms during the pilot study. The workflow for the study is shown in Fig. 1. Fig. 1 Workflow for the pilot study of the HTT Project. Study materials: the project includes a study of 640 ROIs from 64 breast cancer H&E slides. Pathologist recruitment: pathologists were recruited to either collect data in-person (eeDAP) and/or using a digital web-application (PathPresenter or caMicroscope). Study analysis: the results of the pilot study were then analyzed across all platforms and interobserver agreement analysis was evaluated using the results from the PathPresenter platform. Pathologists were recruited to contribute at an on-site data-collection event held in conjunction with the annual meeting of the United States and Canadian Academy of Pathology (USCAP) 2020. Further in-person data collection on eeDAP was paused due to the COVID-19 pandemic. The digital platforms replicated the eeDAP data-collection workflow and allowed continuous and world-wide data collection during the COVID-19 pandemic. Using webinars, training videos, and scientific publications, we provided training on sTILs identification and density estimation. In total, ∼90  min of training was available for this study. We did not enforce completion of training prior to data collection. Pathologists were asked to annotate complete batches but were not required to annotate all 640 ROIs in the study. We also asked participants to complete batches of 80 ROIs in a random order, and late in the pilot study, we introduced a random-number-generator to assist pathologists with randomizing the sequence of batches. Use of the random-number-generator was not enforced. 2.4 Time to Collect Data Across All Three Platforms We examined the time-associated burden of this study on pathologists. The time it takes to complete each ROI annotation is automatically recorded by the annotation software for all platforms. The time recorded by eeDAP does not include the time required for the microscope stage to move between ROIs. Additionally, the eeDAP system requires additional time to register each slide in a batch to the corresponding digital image. eeDAP consists of two levels of registration (low and high resolution) and a camera-eyepiece offset correction. The high-resolution registration and camera-eyepiece offset correction are optional but add additional time to the study. When these optional steps are performed, registration accuracy is ∼42  μm.41 We calculated the time for annotation on eeDAP by combining (1) the time it takes to set up and run through slides without deliberative annotation (an experiment completed by the study investigators, not the study pathologists) and (2) the average time it took study pathologists to annotate ROIs on eeDAP. 2.5 Quantitative Analysis of the PathPresenter Pilot Study Data For this work, we analyzed the distribution of sTILs annotations from the PathPresenter digital platform collected from February 2020 to May 2021. Variability from platform differences were eliminated by analyzing only PathPresenter annotations. In this analysis, we calculated the sample mean and variance of the sTILs density estimates over the readers for each ROI, and we examined these variance estimates and corresponding coefficients of variation (CV) across the entire range and within the three density bins defined above (low, moderate, and high). We also characterized the interpathologist agreement with the root-mean-squared difference, which is similar to the root-mean-squared error but does not require an independent reference standard.42 Specifically, the pathologist-to-pathologist expected mean-squared difference (MSD) is given as MSD=E[(Xj′kl−Xjkl)2], where the X’s denote the sTILs density estimates from two different pathologists (j and j′) for the same ROI (k and l denote the case and ROI, respectively). This metric is an average over pathologists, cases, and ROIs. In a full multireader and multicase study, we should account for the variability and correlations from readers and cases,43 as well as the correlation of ROIs that are nested within a case. In this work, we analyze two readers who each completed the entire study and ignore the correlation of ROIs. 2.6 Pilot Study Feedback from Expert Collaborators After collecting and exploring the pilot study data, we convened an expert panel of seven practicing pathologists and one clinical scientist with expertise in microscopy and pathologist data collection.44 All expert panelists were pilot study collaborators with expertise in breast cancer pathology, training in sTILs assessment, and involvement in the design of our pilot study. One of these pathologists completed the pilot study. Two other expert pathologists fully annotated batch 1 (80 ROIs) of the pilot study. These annotations were completed at least 6 months before the expert panel was convened; crowdsourced pathologists were not invited. Through a series of eight recorded one-hour virtual sessions, the experts discussed each ROI regarding their sTILs assessment and experience with the data-collection process. The virtual sessions included a minimum of three expert panelists with study investigators guiding the discussion and note-taking. From these discussions, we identified limitations in our platforms and workflows. We also discussed potential resolutions to these limitations that will improve the pivotal study.44 3 Results From the start of the pilot study in February 2020 through May 2021, 29 pathologists generated 7373 annotations across the three annotation platforms, with eight pathologists annotating on multiple platforms. Information about the readers in the study varies based on completion of the data-collection registration form. The registration form was provided to readers but not mandated. From the registration survey, 19 readers were board-certified pathologists with a range of 3 to 44 years since certification (median = 11 years and IQR = 17.5). Four readers reported current enrollment in a pathology residency program. Six readers did not provide information regarding their level of experience. Annotations varied across platforms by number of ROIs scored and whether a pathologist completed a full study. By the conclusion of the pilot study, seven pathologists annotated all 640 ROIs on at least one platform. On PathPresenter, 50.2% (321/640) of the ROIs were annotated more than two times and two pathologists completed the full study (640 ROIs). Annotation within the study was uneven. The collection of the first 80 ROIs known as batch 1 was annotated 16 times, but all other batches were only annotated seven to nine times. Table 1 shows the results of the data-collection efforts of the pilot study across all platforms. Information on total time to complete a study was estimated from available data. A study of the eeDAP platform examined the additional time needed to complete the study in-person due to slide registration and hardware operation. On average, the additional time needed for in-person data-collection was 30:54 min per four slides. Table 1 Tabulated annotation counts and average annotation time of the pilot study by data collection platform. Included are the number of readers and individual observations per platform, the average time to annotate a single ROI per platform with standard deviation in brackets, additional time needed for eeDAP hardware operation, and the estimated total average time for a pathologist to complete a full study (n=640 ROIs). Values for batch time and study time are calculated by multiplying average time for annotation per ROI (ROI/sec) and average time for hardware accounting for the number of ROIs. As no full study set was completed on the eeDAP platform, these numbers are estimated from a small subset of readers (nReaders = 7 and nROI 80 to 240). Platform Readers Observations Avg. time for annotation [st. dev] (ROIs = 1, s) Avg. time for hardware (ROIs = 1, s) Batch time (ROIs = 80, h) Study time (ROIs = 640, h) eeDAP 7 440 27.41 [24.94] 46.34 1.64 13.11 PathPresenter 10 1833 29.98 [29.37] N/A 0.67 5.33 caMicroscope 20 5100 39.06 [34.73] N/A 0.87 6.94 All 37 7373 32.15 [35.52]       We investigated the distribution of annotations across all platforms and summarized total ROIs stratified by tumor classification and sTILs density bin in Table 2. We tallied the instances for each of the four ROI labels and the mean sTILs density per ROI within each of the pre-selected density bins. The distribution was counted within individual platforms and summarized across all platforms. As shown in Table 2, over 75.8% (5592/7373) of all labels were listed as “intratumoral stroma.” Additionally, 60.9% (4492/7373) of all sTILs density estimates fell within the 0% to 10% and only 4.4% (322/7373) were estimates between 41% and 100%. Table 2 Distribution of pathologist annotations (n=7373) in the pilot study by ROI label and the density bins. All ROIs are assigned an ROI label. If the label is considered evaluable, the pathologist then proceeds to assign the ROI a density estimate. Density estimates fell into one of three bins: low, moderate, or high.   ROI label Density Bins Evaluable (n=5958) Not evaluable Low: 0% to 10% Moderate: 11% to 40% High: 41% to 100% Platform Intratumoral stroma Invasive margin Tumor with no intervening stroma Other eeDAP 349 5 9 77 323 21 10 PathPresenter 1326 91 288 128 1127 211 79 caMicroscope 3917 270 210 703 3042 912 233 All 5592 366 507 908 4492 1144 322 We analyzed the sTILs density estimates for every ROI with at least two annotations on the PathPresenter platform (n=495). As four of the ten PathPresenter readers did not complete registration, we did not factor our analysis by experience level. In Fig. 2(a), we show the sample variance of the estimates as a function of the mean sTILs density estimates on the PathPresenter platform. From this figure, we observe that the pathologist variance depends on the ROI and increases with the mean. Figure 2(b) shows the CV of the sTILs density estimates as a function of the mean sTILs density estimates. We observe high CV at low sTILs density that decreases as the sTILs density increases. The CV plot also shows that the standard deviation of the data is not proportional to the mean. Fig. 2 (a) Variance as a function of the mean calculated as averages over all readers on PathPresenter. Each blue circle represents one ROI (n=495) with at least two sTILs density estimates. The horizontal lines show the average variance in 10% bins of the data (49 ROIs). Vertical dashed lines split the data into low (0% to 10%), moderate (11% to 40%), and high (41% to 100%) sTILs density bins. (b) Coefficient of variation calculated as averages over all readers on PathPresenter. We characterized interpathologist agreement by analyzing the annotations of the two pathologists who completed a full study on the PathPresenter platform. Table 3 shows the number of instances each pathologist gave a density estimate of an ROI or indicated an ROI was not evaluable. We analyzed the observations for each ROI as paired observations and averaged observations. For example, only instances where ROIs were marked as not evaluable by both pathologists were counted as a paired observation. Thus, 89 ROIs were labeled as not evaluable by both pathologists. In total, 448 ROIs were given density estimates by both pathologists (Table 3). This metric also helps us understand the number of ROIs assigned similar scores, but observations are lost when scores cross between density bins. Averaged observations are paired according to the mean density estimate of the ROI. Therefore, an ROI with density estimates across two bins (e.g., 9% and 11%) will be binned according to the mean of the estimates (10% = low-density bin). Finally, we calculated the root mean square difference (RMSD) within each density bin for the averaged observations. Table 3 Individual and paired agreement analysis of the two pathologists who annotated all 640 ROIs in the study (anonymized ReaderIDs shown). The number of times each pathologist annotated an ROI as not evaluable or assigned a density estimate within the low, moderate, or high-density bin is shown in the first two rows. The third row shows the paired ROI observations for both pathologists. The fourth row shows the averaged observations for both pathologists. The last row is the calculated root mean square difference (RMSD) between the two pathologists across all densities and within the low, moderate, and high-density bins.   Not evaluable All densities: 0% to 100% Low densities: 0% to 10% Mod. densities: 11% to 40% High densities: 41% to 100% ReaderID 4776 92 548 426 94 28 ReaderID 0455 189 451 344 68 39 Paired observations 89 448 313 44 24 Averaged observations N/A 448 319 93 36 RMSD   10.62 4.62 14.25 26.25 Figure 3 shows Bland–Altman plots for each density bin to better visualize the difference between the two pathologists. In these plots, the differences between the paired observations are plotted against the mean of the estimates. The mean difference (m¯) for each bin is represented by the solid horizontal black line in each plot. The limits of agreement (horizontal red dashed lines) were constructed using twice the standard deviation of the differences in sTILs densities (RMSD2^−m¯2). As there were very few paired observations within the moderate and high-density bins, any further trend-identification analysis or confidence interval calculations were deemed inappropriate. Fig. 3 (a)–(d) Mean-difference (Bland–Altman) plots of the two pathologists with complete sets of annotations on the PathPresenter platform. Each blue circle represents a group of ROIs that share the same mean and different values. The size of the light blue circle is proportional to the number of ROIs; the largest circle corresponds to 50 ROIs. The horizontal black line shows the bias from 0 of each density bin. Red horizontal dashed lines indicate the upper and lower limits of agreement derived from the RMSD in Table 3. 4 Discussion The pilot study of the HTT project used three data collection platforms to generate 7373 annotations from 29 pathologists over a 15-month period. This manuscript summarizes the context, methods, and experience of the data collection with a focus on understanding processes to speed data collection and reduce pathologist variability. We focus on data collection and assessment time across platforms but only present quantitative the PathPresenter annotation results to understand pathologist variability on a single platform. Future work will look at the data from the other platforms, building on the lessons learned and presented here. We estimated the time-burden on pathologists by a data-collection platform (Table 1). While the two digital platforms have similar estimated times (estimated average is 36 s/ROI), a pathologist using the eeDAP system can expect to spend double the amount of time on the study due to stage movements and registering the images to the slides. Thus, we are investigating upgrades to the eeDAP hardware and annotation workflow for the pivotal study. These time estimates are useful for planning future data collection activities and determining potential reimbursement rates for annotators. We did not expect crowdsourced volunteer pathologists to annotate all 640 ROIs of the pilot study. The complete study takes more than five hours (Table 1), which is beyond reasonable expectations for most volunteers. Our plan for the pivotal study is to manage the workload of each pathologist by balancing the distribution of batches among the pathologists in the spirit of a split-plot study design.45–47 We did not effectively manage the workload of the pathologists in the pilot study as the data-collection tools did not yet have methods to create pathologist-specific work lists. We found that batch 1 was annotated double the number of times as other batches. This finding suggests a need to either enforce the use of the random-number-generator or integrate another strategy to direct readers within the workflow. We analyzed the pilot study data as continuous data (0% to 100%) and within three predefined clinical bins (low 0% to 10%, moderate 11% to 40%, and high 41% to 100%) as prior studies of sTILs found associations between these density bins and clinical outcomes.18,24,30,31 We also observed 60.9% of the annotations were in the low sTILs density bin (Table 2). The two pathologists who annotated all 640 ROIs also identified over 70% of evaluable ROIs as low density (319/448), demonstrating that low-density ROIs persist throughout the study and are not confined to the most annotated batches (Table 4). These findings indicate that the ROIs selected for the pivotal study will need to avoid oversampling of the low-density bin within individual batches as well as within the study. Moreover, the entire pathologist workflow must be managed to ensure equitable distribution of annotations between batches. Consistent with prior studies, there is high variability in pathologist sTILs density estimates (Fig. 2).26,30–32 Plotting variance as a function of the mean shows that the variance of the density estimates increases with the mean. We quantify this finding with the RMSD results in Table 3. The CV plot shows a complicated relationship with the mean with large relative variability for small sTILs densities (Fig. 2). Together, Figs. 2 and 3 indicate that the data are not independent and identically distributed or well-behaved as a function of the mean sTILs density. This precludes the use of many standard analysis methods for quantitative data. We have explored log and other transformations, but they did not stabilize the mean-variance relationship. As such, we will explore methods that treat the data as ordinal (ranks) and nominal (bins). Furthermore, the analyses in this work ignored the correlation between ROIs from the same slide. We will address this limitation in future work and properly account for those correlations. This work identified several limitations in the technical workflows of the pilot study, including completion of the prestudy participant registration survey. The low completion rate for the pre-study registration translated to an inability to identify the experience level of 6 of the 29 pilot study readers. The original prestudy registration did not request information on a reader’s clinical specialty or familiarity with the clinical task, eliminating our ability to stratify expert readers from those new to the task or who had not completed training. These issues informed the design and integration of a mandatory registration survey that will prevent readers from starting the study until all fields of the registration survey are complete. An additional component contributing to variability within the HTT project was the moderately structured and unenforced training provided to readers. As completion of the training modules was unmonitored and honor-based, we have no ability to determine the actual completion rates for study participants. We cannot evaluate the impact of existing training modules on pathologists’ annotations or extrapolate the potential reduction in variability resulting from enforcement and standardization of training. To resolve these issues, we are in the process of developing a standardized training module consisting of short videos describing the clinical task for each data collection platform, a training set of ROIs that will provide feedback on participant annotations, and a mandatory proficiency test. A passing grade on this proficiency test will be required to proceed with the study. Standardizing registration and training workflows will assist with evaluating pathologist performance in the pivotal study. Finally, our data analysis results should be taken with caution. We made changes to the protocol and data-collection platforms during the pilot study. Another limitation of our analysis is the methods used to determine the average time needed for an annotation (Table 1). The eeDAP platform did not record the time related to microscope stage movements and both digital and microscope platforms can only determine the time spent on the annotation input screen (and will not reflect any interruptions). Nonetheless, we believe these measurements can help us plan the time it will take to conduct future pivotal studies, which is their true purpose here. 5 Conclusions This work contributes to the field of digital pathology and AI/ML validation by establishing the methods and tools useful for the standardized collection of pathologist annotations. It also demonstrates that new analysis methods are needed as prior assumptions of the appropriate statistical methods for establishing interpathologist agreement were not supported, like treating the data as independent and identically distributed and finding variance differs as a function of the mean. This pilot study and the associated interpathologist agreement methods we are currently exploring will be used to size and create a validation dataset for AI/ML algorithms that yield sTILs density estimates in breast cancer. Our pilot dataset and analysis methods are available on a public HTT Github repository (https://github.com/didsr/htt) to allow open access to our methods and feedback from the digital pathology and statistics communities. The data and methods serve as an instructive tool for AI/ML algorithm developers and researchers and can be replicated for other tumor sites and quantitative biomarkers. Acknowledgments This work was supported by the FDA Office of Women’s Health. This project was supported in part by an appointment (V.G.) to the ORISE Research Participation Program at the CDRH, U.S. Food and Drug Administration, administered by the Oak Ridge Institute for Science and Education through an interagency agreement between the U.S. Department of Energy and FDA/Center. The FDA IRB determined that the research study was exempt from the requirements of 45 CFR part 46; 45 CFR 46.104(d)(2)(ii) (No. 2019-CDRH-109). The mention of commercial products, their sources, or their use in connection with material reported herein is not to be construed as either an actual or implied endorsement of such products by the Department of Health and Human Services. Katherine Elfer is an NCI-FDA Interagency Oncology Task Force Cancer Prevention Fellow. She received her BS degree in nanosystems engineering from Louisiana Tech University, her MPH degree from Johns Hopkins University Bloomberg School of Public Health, and her PhD in biomedical engineering from Tulane University in 2018. Her research interests include the development and validation of tools for digital pathology, human-technology interactions, and the regulatory science of diagnostic imaging devices. Sarah Dudgeon is a doctoral student at Yale University in the computational biology and bioinformatics program. She received her Master of Public Health degree from Johns Hopkins University and her Bachelor of Science degree from the University of Michigan. Her research interests include graph analytics, healthcare data formatting, and innovative methods for verification and validation. Victor Garcia, MD, is an ORISE fellow in the FDA’s Division of Imaging, Diagnostics, and Software Reliability. After completing an internal medicine residency and clinical informatics fellowship, he became a member of the high-throughput truthing project and is leading efforts to pursue qualification of this dataset as a medical device development tool. His prior research includes natural language processing of radiology reports and analysis of electronic health record data. Kim Blenman, PhD, MS, is an immunologist and clinical chemist with expertise in drug development and regulatory science. She uses and develops innovative methods, hardware, and software tools to understand the mechanisms responsible for disparities in disease pathogenesis and therapeutic response. Evangelos Hytopoulos is currently a senior director of data science at iRhythm Technologies, leading the development of deep learning algorithms that are at the heart of the detection of arrhythmias through iRhythm’s wearable biosensing technology. He has more than 20 years of experience in AI/ML, algorithm development, high-performance computing, 3D visualization, and technical management. He received his master’s and PhD degrees in aerospace engineering and an undergraduate degree in naval architecture and marine engineering. Si Wen works at the FDA as a statistician to support the evaluation of medical imaging devices and related algorithms. She received her PhD in statistics from Stony Brook University in 2018. Her research is focused on statistical analysis for multireader multicase (MRMC) studies. Manasi S. Sheth offers 10+ years of combined modeling and analytical experience within various applications for statistics. Throughout her career, she has demonstrated flexible positions in academia, industry, and government sectors. Her responsibilities have included quality management, health authority interface, regulatory compliance management, pilot testing and development, and technical deliverables. Her successes include multiple initiatives resulting in significant quality improvement and increased operational efficiency. Mohamed Amgad is a physician-scientist in training, working in the interdisciplinary field of computational pathology. He received his MD degree from Cairo University in 2016 and his PhD from Emory University in 2021. His research interests include the development of reliable and explainable machine learning models and systems to facilitate diagnostic pathology, and to discover prognostic visual biomarkers in scanned histopathology slides. Biographies of the other authors are not available. Disclosure KNE holds equity in Instapath Bioptics LLC, New Orleans, LA. DP serves as a part-time consultant at CellCarta NV, Antwerp, Belgium. BW is AL serves as a part-time consultant at Ultivue, Inc. KB is on the Scientific Advisory Board of CDI Labs (Mayaguez, Puerto Rico). RS serves an Advisory Board role for BMS, Roche. RS has received research funding by Roche, Puma, Merck. RS has received travel and congress-registration support by Roche, Merck, and Astra Zeneca. The authors report no conflicts related to the current work. Data Availability Statement The expert annotations on the select ROIs analyzed in this study will be publicly released before or in parallel with the publication of this work. The pilot study annotations are immediately available on this public repository: https://github.com/DIDSR/HTT. ==== Refs References 1. Marble H. D. et al. , “A regulatory science initiative to harmonize and standardize digital pathology and machine learning processes to speed up clinical innovation to patients,” J. Pathol. Inf. 11 (1 ), 22 (2020).10.4103/jpi.jpi_27_20 2. Cucoranu I. C. et al. , “Digital pathology: a systematic evaluation of the patent landscape,” J. Pathol. Inform. 5 , 16 (2014).10.4103/2153-3539.133112 25057430 3. Cui M. Zhang D. Y. , “Artificial intelligence and computational pathology,” Lab Invest. 101 , 412–422 (2021).10.1038/s41374-020-00514-0 33454724 4. Bera K. et al. , “Artificial intelligence in digital pathology - new tools for diagnosis and precision oncology,” Nat. Rev. Clin. Oncol. 16 (11 ), 703–715 (2019).10.1038/s41571-019-0252-y 31399699 5. Rashidi H. H. et al. , “Artificial intelligence and machine learning in pathology: the present landscape of supervised methods,” Acad. Pathol. 6 , 2374289519873088 (2019).10.1177/2374289519873088 31523704 6. FDA/CDRH, Software as a Medical Device (SAMD): Clinical Evaluation, U.S. Food and Drug Administration (2017). 7. Steiner D. F. Chen P.-H. C. Mermel C. H. , “Closing the translation gap: AI applications in Digital Pathology,” Biochim. Biophys. Acta Rev. Cancer 1875 , 188452 (2020).10.1016/j.bbcan.2020.188452 33065195 8. Janowczyk A. Madabhushi A. , “Deep learning for digital pathology image analysis: a comprehensive tutorial with selected use cases,” J. Pathol. Inform. 7 , 29 (2016).10.4103/2153-3539.186902 27563488 9. Dano H. et al. , “Interobserver variability in upfront dichotomous histopathological assessment of ductal carcinoma in situ of the breast: the DCISion study,” Mod. Pathol. 33 (3 ), 354–366 (2020).MODPEO 0893-3952 10.1038/s41379-019-0367-9 31534203 10. Tramm T. et al. , “Standardized assessment of tumor-infiltrating lymphocytes in breast cancer: an evaluation of inter-observer agreement between pathologists,” Acta Oncol. 57 (1 ), 90–94 (2018).10.1080/0284186X.2017.1403040 29168428 11. Engel K. B. Moore H. M. , “Effects of preanalytical variables on the detection of proteins by immunohistochemistry in formalin-fixed, paraffin-embedded tissue,” Arch. Pathol. Lab. Med. 135 (5 ), 537–543 (2011).APLMAS 0003-9985 10.5858/2010-0702-RAIR.1 21526952 12. Yildiz-Aktas I. Z. Dabbs D. J. Bhargava R. , “The effect of cold ischemic time on the immunohistochemical evaluation of estrogen receptor, progesterone receptor, and HER2 expression in invasive breast carcinoma,” Mod. Pathol. 25 (8 ), 1098–1105 (2012).10.1038/modpathol.2012.59 22460807 13. Khoury M. J. et al. , “Health-centers for disease control and prevention multidisiplinary workshop,” Genet. Med. 11 (8 ), 559–567 (2009).10.1097/GIM.0b013e3181b13a6c 19617843 14. Allison K. H. et al. , “Understanding diagnostic variability in breast pathology: lessons learned from an expert consensus review panel,” Histopathology 65 (2 ), 240–251 (2014).HISTDD 1365-2559 10.1111/his.12387 24511905 15. Nagendran M. et al. , “Artificial intelligence versus clinicians: systematic review of design, reporting standards, and claims of deep learning studies,” BMJ 368 , m689 (2020).10.1136/bmj.m689 32213531 16. Salgado R. et al. , “The evaluation of tumor-infiltrating lymphocytes (TILs) in breast cancer: recommendations by an International TILs Working Group 2014,” Ann. Oncol. 26 (2 ), 259–271 (2015).ANONE2 0923-7534 10.1093/annonc/mdu450 25214542 17. Hendry S. et al. , “Assessing tumor-infiltrating lymphocytes in solid tumors: A practical review for pathologists and proposal for a standardized method from the international immunooncology biomarkers working group: part 1: assessing the host immune response, TILs in invasive breast carcinoma and ductal carcinoma in situ, metastatic tumor deposits and areas for further research.,” Adv. Anat. Pathol. 24 (5 ), 235–251 (2017).10.1097/PAP.0000000000000162 28777142 18. Kos Z. et al. , “Pitfalls in assessing stromal tumor infiltrating lymphocytes (sTILs) in breast cancer,” NPJ Breast Cancer 6 (1 ), 17 (2020).10.1038/s41523-020-0156-0 32411819 19. Mao Y. et al. , “The prognostic value of tumor-infiltrating lymphocytes in breast cancer: a systematic review and meta-analysis,” Plos One 11 (4 ), e0152500 (2016).POLNCL 1932-6203 10.1371/journal.pone.0152500 27073890 20. Loi S. et al. , “Tumor-infiltrating lymphocytes and prognosis: a pooled individual patient analysis of early-stage triple-negative breast cancers,” J. Clin. Oncol. 37 (7 ), 559–569 (2019).10.1200/JCO.18.01010 30650045 21. Savas P. et al. , “Clinical relevance of host immunity in breast cancer: from TILs to the clinic,” Nat. Rev. Clin. Oncol. 13 (4 ), 228–241 (2016).10.1038/nrclinonc.2015.215 26667975 22. Denkert C. et al. , “Tumour-infiltrating lymphocytes and prognosis in different subtypes of breast cancer: a pooled analysis of 3771 patients treated with neoadjuvant therapy,” Lancet Oncol. 19 (1 ), 40–50 (2018).LOANBN 1470-2045 10.1016/S1470-2045(17)30904-X 29233559 23. Park J. H. et al. , “Prognostic value of tumor-infiltrating lymphocytes in patients with early-stage triple-negative breast cancers (TNBC) who did not receive adjuvant chemotherapy,” Ann. Oncol. Off. J. Eur. Soc. Med. Oncol. 30 (12 ), 1941–1949 (2019).ANONE2 0923-7534 10.1093/annonc/mdz395 24. Luen S. J. et al. , “Prognostic implications of residual disease tumor-infiltrating lymphocytes and residual cancer burden in triple-negative breast cancer patients after neoadjuvant chemotherapy,” Ann. Oncol. 30 (2 ), 236–242 (2019).ANONE2 0923-7534 10.1093/annonc/mdy547 30590484 25. Dieci M. V. et al. , “Update on tumor-infiltrating lymphocytes (TILs) in breast cancer, including recommendations to assess TILs in residual disease after neoadjuvant therapy and in carcinoma in situ: a report of the International Immuno-Oncology Biomarker Working Group on Breast Cancer,” Semin. Cancer Biol. 52 (Part 2 ), 16–25 (2017).SECBE7 1044-579X 10.1016/j.semcancer.2017.10.003 29024776 26. Amgad M. et al. , “Report on computational assessment of Tumor Infiltrating Lymphocytes from the International Immuno-Oncology Biomarker Working Group,” NPJ Breast Cancer 6 , 16 (2020).10.1038/s41523-020-0154-2 32411818 27. Garutti M. et al. , “Find the flame: predictive biomarkers for immunotherapy in melanoma,” Cancers 13 (8 ), 1819 (2021).10.3390/cancers13081819 33920288 28. Uryvaev A. et al. , “The role of tumor-infiltrating lymphocytes (TILs) as a predictive biomarker of response to anti-PD1 therapy in patients with metastatic non-small cell lung cancer or metastatic melanoma,” Med. Oncol. Northwood Lond. Engl. 35 (3 ), 25 (2018).10.1007/s12032-018-1080-0 29. Gartrell R. D. et al. , “Quantitative analysis of immune infiltrates in primary melanoma,” Cancer Immunol. Res. 6 (4 ), 481 (2018).10.1158/2326-6066.CIR-17-0360 29467127 30. Van Bockstal M. R. et al. , “Interobserver variability in the assessment of stromal tumor-infiltrating lymphocytes (sTILs) in triple-negative invasive breast carcinoma influences the association with pathological complete response: the IVITA study,” Mod. Pathol. 34 (12 ), 2130–2140 (2021).10.1038/s41379-021-00865-z 34218258 31. Swisher S. K. et al. , “Interobserver agreement between pathologists assessing tumor-infiltrating lymphocytes (TILs) in breast cancer using methodology proposed by the International TILs Working Group,” Ann. Surg. Oncol. 23 (7 ), 2242–2248 (2016).10.1245/s10434-016-5173-8 26965699 32. O’Loughlin M. et al. , “Reproducibility and predictive value of scoring stromal tumour infiltrating lymphocytes in triple-negative breast cancer: a multi-institutional study,” Breast Cancer Res. Treat. 171 (1 ), 1–9 (2018).BCTRD6 10.1007/s10549-018-4825-8 29774470 33. Ju X. et al. , “Predictive relevance of PD-L1 expression with pre-existing TILs in gastric cancer,” Oncotarget 8 (59 ), 99372–99381 (2017).10.18632/oncotarget.22079 29245908 34. Tan W. C. C. et al. , “Overview of multiplex immunohistochemistry/immunofluorescence techniques in the era of cancer immunotherapy,” Cancer Commun. Lond. Engl. 40 (4 ), 135–153 (2020).10.1002/cac2.12023 35. Dudgeon S. et al. , “A pathologist-annotated dataset for validating artificial intelligence: A project description and pilot study,” J. Pathol. Inf. 12 (1 ), 45 (2021).10.4103/jpi.jpi_83_20 36. Gallas B. D. et al. , “Evaluation environment for digital and analog pathology (eeDAP): a platform for validation studies,” J Med Img 1 (3 ), 037501 (2014).10.1117/1.JMI.1.3.037501 37. AI PathPresenter, https://pathpresenter.net (accessed 16 Sept. 2020). 38. Singh R. , “Introducing PathPresenter, an innovative platform of pathology by pathologists for pathology practices,” Int. Acad. Pathol. Bull. 61 (1 ), 4 (2021). 39. Saltz J. et al. , “A containerized software system for generation, management, and exploration of features from whole slide tissue images,” Cancer Res. 77 (21 ), e79–e82 (2017).CNREA8 0008-5472 10.1158/0008-5472.CAN-17-0316 29092946 40. “Home,” Int. TILS Work. Group, https://www.tilsinbreastcancer.org/ (accessed 5 Nov. 2021). 41. Gong Q. et al. , “Registration accuracy between whole slide images and glass slides in eeDAP workflow,” Proc. SPIE 10581 , 1058118 (2018).PSISDG 0277-786X 10.1117/12.2293189 42. Obuchowski N. A. , “Can electronic medical images replace hard-copy film? Defining and Testing the equivalence of dagnostic tests,” Stat. Med. 20 (19 ), 2845–2863 (2001).SMEDDA 1097-0258 10.1002/sim.929 11568944 43. Wen S. Gallas B. D. , “Three-way mixed effect ANOVA to estimate MRMC limits of agreement,” Stat. Biopharm. Res. 1–10 (2022).10.1080/19466315.2022.2063169 44. Garcia V. et al. , “Development of training materials for pathologists to provide machine learning validation data of tumor-infiltrating lymphocytes in breast cancer,” Cancers 14 (10 ), 2467 (2022).10.3390/cancers14102467 35626070 45. Obuchowski N. A. Gallas B. D. Hillis S. L. , “Multi-reader ROC studies with split-plot designs: a comparison of statistical methods,” Acad. Radiol. 19 (12 ), 1508–1517 (2012).10.1016/j.acra.2012.09.012 23122570 46. Gallas B. D. et al. , “A collaborative project to produce regulatory-grade pathologist annotations to validate viewers and algorithms, in: abstracts,” J. Pathol. Inf. 10 (1 ), 28 (2019).10.4103/2153-3539.266902 47. Chen W. Gong Q. Gallas B. D. , “Paired split-plot designs in MRMC studies,” J. Med. Imaging 5 (3 ), 031410 (2018).10.1117/1.JMI.5.3.031410
PMC009xxxxxx/PMC9532384.txt
==== Front Cell Death Dis Cell Death Dis Cell Death & Disease 2041-4889 Nature Publishing Group UK London 36195585 5225 10.1038/s41419-022-05225-6 Article Single-cell sequencing resolves the landscape of immune cells and regulatory mechanisms in HIV-infected immune non-responders Li Haiyu 12 Tang Yongyao 3 Wang Yujing 3 Li Yue 4 Yang Yi 5 Liao Kui [email protected] 6 Song Fangzhou [email protected] 4 Deng Shixiong [email protected] 4 http://orcid.org/0000-0003-3403-9782 Chen Yaokai [email protected] 1 1 grid.507893.0 0000 0004 8495 7810 Department of Infectious Disease, Chongqing Public Health Medical Center, 400036 Chongqing, China 2 grid.9227.e 0000000119573309 Chongqing Institute of Green and Intelligent Technology, Chinese Academy of Sciences, 400714 Chongqing, China 3 grid.203458.8 0000 0000 8653 0555 School of Medical Information, Chongqing Medical University, 400016 Chongqing, China 4 grid.203458.8 0000 0000 8653 0555 Basic Medical College, Chongqing Medical University, 400016 Chongqing, China 5 grid.203458.8 0000 0000 8653 0555 The First Clinical College of Chongqing Medical University, Chongqing Medical University, 400016 Chongqing, China 6 grid.452206.7 0000 0004 1758 417X Department of radiotherapy, The First Affiliated Hospital of Chongqing Medical University, 400016 Chongqing, China 4 10 2022 4 10 2022 10 2022 13 10 8495 5 2022 25 8 2022 31 8 2022 © The Author(s) 2022, corrected publication 2023 https://creativecommons.org/licenses/by/4.0/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Immune non-responder after highly active antiretroviral therapy (HAART) is the main cause of opportunistic infections and high mortality in AIDS patients, but the mechanism underlying immune reconstitution failure is poorly understood. Here, we performed scRNA-seq, and scATAC-seq analysis of peripheral blood mononuclear cells (PBMCs) derived from immune non-responder (INR) and responder (IR) HIV-1-infected subjects. We found low expression of mucosal-associated invariant T (MAIT) cells in INRs, which exhibited transcriptional profiles associated with impaired mitochondrial function and apoptosis signaling. Single-cell assays for transposase-accessible chromatin (scATAC-seq) and flow cytometry revealed diminished mitochondrial fitness in MAIT cells from INRs, and MAIT had low expression of transcription factor A for mitochondria (TFAM) and peroxisomal proliferator-activated receptor alpha (PPARA). These findings demonstrate that restoring mitochondrial function could modulate the immune dysfunction characteristic of MAIT against bacterial co-infections in INRs subjects. Subject terms HIPPO signalling HIV infections https://doi.org/10.13039/501100001809 National Natural Science Foundation of China (National Science Foundation of China) 82002154 Li Haiyu issue-copyright-statement© Associazione Differenziamento e Morte Cellulare ADMC 2022 ==== Body pmcIntroduction Acquired immune deficiency syndrome (AIDS) is a fatal infectious disease caused by the human immunodeficiency virus (HIV). The main pathological feature is a significant reduction in CD4+ lymphocyte count, resulting in a series of immunodeficiency-related diseases. An increasing body of evidence suggests that combined antiretroviral therapy (cART) can significantly reduce viral replication in HIV/AIDS patients, increase the CD4+T lymphocyte count, reconstitute the immune function and reduce mortality [1–5]. However, there are still 15–20% of AIDS patients with complete inhibition of virus replication after cART, and the CD4+ T-cell count cannot be restored to the normal level of uninfected people, a phenomenon termed immune non-response (INR) [6–10]. INR increases the susceptibility to opportunistic infections or non-AIDS-related diseases, such as cardiovascular, liver and kidney disease, leading to higher mortality rates than immune responders (IRs) with restored CD4+T-cell count [11–15]. Opportunistic infections are common in INR patients, and the incidence rate and mortality are also high. In addition to AIDS-related diseases and deaths, non-AIDS-related diseases and deaths are also higher in nonimmune responders [16–19]. Current evidence suggests that low CD4+T-cell counts increase the incidence rate and mortality of cardiovascular diseases and are closely related to the occurrence of AIDS-related or non-related tumors and HIV-related neurocognitive diseases. To better understand the mechanism underlying immune reconstitution failure, we used single-cell RNA sequencing (scRNA-seq) and single-cell assay for transposase-accessible chromatin with high-throughput sequencing (scATAC-seq) to examine the single-cell transcriptional profile of nonimmune and immune responders. We found lower mucosal-associated invariant T cells (MAIT) expression in INRs, which displayed transcriptional profiles associated with impaired mitochondrial function and apoptosis signaling. MAITs are part of the immune system; their main task is to control bacteria on the body barrier (such as skin and mucosa). MAIT cells represent unconventional T cells with the dual characteristics of innate and adaptive immunity. They can be activated in a TCR-dependent or independent manner [20, 21]. They respond quickly to stimuli and can affect the response of T cells and B cells in the early stage of immune response [22–26]. This study explores the molecular mechanism of MAIT mitochondrial dysfunction in HIV-infected immune non-responders. Results Single-cell transcriptome atlas of PBMCs in IRs and INRs To identify the immunological features of HIV-1-infected patients, we performed droplet-based single-cell RNA sequencing technology (10×Genomics) of fresh PBMCs derived from 2 immune non-responders (INR1 and INR2) and 2 immune responder controls (IR1 and IR2). After quality control and a unified single-cell analysis pipeline, the total number of recovered cells was 12086, comprising 5172 cells for INR and 6915 cells for controls IRs detecting a mean of 1,445 and 1,384 genes per cell, respectively (Fig. 1A, B and Fig. S1A). We calculated subsets that showed high differences between cells in the dataset (i.e., high and low expression in distinct cells). The screened differential genes with hypervariable characteristics were used for downstream principal component analysis. We visualized the cells in two-dimensional spaces according to their expression profiles using graph-based clustering of uniform manifold approximation and projection (UMAP), a nonlinear dimensionality reduction algorithm (Fig. 1A). Based on the expression of canonical gene markers [27], we identified cells to 11 major cell types or subtypes: CD8 effector T, NK cell, T helper, B cell, classical monocytes, nonclassical monocytes, pDCs, mucosal-associated invariant T cell, platelets, CD4 naïve T cells, and CD8 naïve T cells (Fig. 1B). A detailed list of marker genes defining cell subsets is provided in Supplementary Table 1 and Fig. S1B. To visualize single-cell RNA-Seq data, we constructed a website at http://150.158.212.80:8081. This website is open to the public and allows visualization of the expression of genes of interest in specific cell types without registration. These clusters and their specific markers included CD3D, CD8A, CD8B, and NKG7 (Fig. 1C, D).Fig. 1 Single-cell transcriptional profiling of PBMCs derived from IR and INR patients with HIV-1-infected. A Two-dimensional UMAP visualization of PBMCs for IRs and INRs. Different colors represent different sample sources. B Different colors represent 10 clusters (cell types) defined by the k-means clustering algorithm. Different colors represent 11 cell types. The UMAP projection of 12086 single cells from IR and INR samples shows 11 clusters with the respective labels. Each dot corresponds to a single cell, colored according to cell type. C, D Stacked violin plots showing the expression distribution of selected canonical cell markers in the CD8 MAIT T cells. Significant reduction and dysfunction of MAIT cells in INRs Single-cell transcriptome sequencing showed that the proportion of MAIT cells in INR decreased significantly. To assess the frequency of MAIT cells, FACS analyses were performed in peripheral blood mononuclear cells (PBMCs) from HIV-1–infected immune non-responders (n = 17), HIV-1–infected immune responders (n = 10), and uninfected healthy blood donors (n = 10). MAIT cells were identified using cell-surface markers CD3+ TCRV7.2+CD161high in PBMCs (Fig. 2A). We found that the frequency of MAIT cells was significantly lower in INRs than in IRs or HCs (Fig. 2B). INRs were defined as having CD4+ T-cell counts below 350 cells/μl and IRs as having CD4 + T-cell counts above 350/μl after at least 2 years of cART with virologic control. The level of IFN gamma was detected in MAIT cells derived from IR, INR, and HC stimulated by paraformaldehyde (PFA)-fixed E. coli. MAIT cells from INRs produced lower levels of IFN gamma in response to E. coli stimulation than the HC group (Fig. 2C, D). These results suggest that MAIT derived from INR exhibited dysfunction in resisting infection by other pathogens.Fig. 2 MAIT derived from INR is dysfunctional. A, B Proportions of MAIT cells detected in PBMCs from healthy controls (HC), immune non-responder (INR) HIV-1-infected subjects and immune responder (IR) HIV-1-infected subjects. Representative FC (flow cytometry) plots (A) and statistical analysis (B) are shown. C, D Detection of interferon-gamma produced by MAIT after PBMC isolated from HCs, INRs, and IRs was stimulated by paraformaldehyde (PFA)-fixed E. coli. MAIT cells from INRs have distinct gene expression profiles reflecting apoptotic signaling and inflammatory responses To define the molecular mechanism of MAIT dysfunction in INRs, we next explored differences in gene expression profiles of MAIT cells between INRs and IRs. We found that the expression of proapoptotic genes in MAIT cells was significantly (P < 0.05) upregulated in INRs than in IRs (Fig. 3A). Furthermore, we found that the expression of proapoptotic genes IFI27, IFIT2, and SAMD9 was significantly increased in MAIT cells from INRs (Fig. 3B). As shown in Fig. 3C, the expression of CD54, KLRB1, and JUN, regulators of T-cell function and MT-ND1, MT-ATP6 and RPS26, master regulators of mitochondrial biogenesis, were significantly decreased in MAIT cells from INRs. It is well-established that mitochondrial function is closely related to apoptosis [28–30]. We analyzed the transcriptional profile of mitochondrial-related genes (MT-ND1, MT-ATP6 and RPS26) in INR and found decreased expression of genes involved in different phases of mitochondrial function of MAIT cells from INRs. To confirm whether the mitochondrial function of MAIT is impaired, Mito-Tracker Green (MG) and Mito-Tracker Red CMXRos (MRC) were used to detect mitochondrial mass and mitochondrial membrane potential, respectively (Fig. 3D). The oxidative phosphorylation activity measured by the mean fluorescence intensity of MCR was significantly decreased in MAIT cells derived from INRs. Similarly, when the mean fluorescence intensity of MG was used to measure the quality of mitochondria, the MFI significantly decreased in MAIT cells from INR (Fig. 3D). The Gene Ontology (GO) and KEGG enrichment analysis further showed that the upregulated genes in MAIT cells from INRs were significantly enriched in Type I IFN and apoptosis pathways (Fig. 3E, F).Fig. 3 The apoptotic gene expression profile of MAIT cells derived from INR. A Heatmaps of differentially expressed genes in MAIT cells between IR and INR groups. B Violin plots of the proapoptotic gene expression in MAIT cells from INRs and IRs. C Violin plots of mitochondrial function-related gene expression in MAIT cells from INRs and IRs. D Mitochondrial mass and mitochondrial membrane potential are estimated by MG fluorescence and MRC fluorescence in MAIT cells from INRs and IRs, respectively. E, F GO and KEGG analyses conducted using DAVID for upregulated genes in MAIT cells from INRs compared with IRs. TFAM and PPARA low expression was associated with MAIT cell dysfunction of INRs To identify the mechanism that drives MAIT cell dysfunction, we performed scATAC-Seq on PBMC derived from INRs. After normalization and dimensional reduction, we identified nine major clusters of cells, together with open chromatin peaks specific to each cluster (Fig. 4A). We integrated the scRNAseq and scATAC-seq datasets to correlate and cross-validate gene expression profiles and chromatin accessibility landscape in INR (Fig. 4B). To identify transcription factors (TFs) essential for cell-type-specific gene expression, we profiled TF binding motifs overrepresented in the peaks that showed differential accessibility among cell types (Fig. 4C and Fig. S2A). We also found that the chromatin domain of BCL11B was active and open in MAIT cells derived from INR, indicating that BCL11B may be highly expressed in INR-derived MAIT cells (Fig. 4D). Having established the high concordance between the scRNAseq and scATAC-seq data, we found that transcription factor A for mitochondria (TFAM) binding activity was significantly decreased in INR (Fig. 4E and Fig. S2B). Furthermore, the expression of TFAM and peroxisomal proliferator-activated receptor alpha (PPARA), the master regulator of mitochondrial biogenesis, were detected by flow cytometry. As shown in Fig. 4F, MAIT cells from INRs showed significantly decreased expression of PPARA and TFAM.Fig. 4 scATAC-Seq reveals open chromatin landscapes of single cells in INR. A tSNE plots showing single-cell open chromatin profiles analyzed in the study, color-coded for cell clusters. B Display after integration of scATAC-Seq and scRNA-seq. C scATAC-seq tracks showing open chromatin peaks associated with cell-type-specific genes across different cell clusters. D Chromatin open state of BCL11B gene in MAIT cells derived from INR. E Transcription factor motif enrichment (upper) and gene expression profiles (bottom row) for TFAM. F Mean fluorescence intensity of TFAM and PPARA expression in MAIT cells from HCs, IRs and INRs. Cytotoxic CD4 T cells are significantly reduced in INRs Our single-cell analysis established that the proportion of Th cells in INR was significantly decreased. It is widely acknowledged that Th cells play an important role in the immune system, and their function is to produce a variety of cytokines, transmit antigen information, promote the differentiation and proliferation of T and B cells, and assist B cells in producing antibodies. Th cells were extracted from all cells to better understand the function of this Th cell-specific population in INR and further analyzed using the Seurat R software package. The Th cluster was divided into four small cell subsets using a clustering algorithm based on shared nearest neighbor modular optimization (Fig. 5A). Based on the expression of established marker genes, we identified a marked increase of cytotoxic CD4 T cells (CD4 cytotoxic T lymphocytes [CTLs]) in IR and INR groups (Fig. 5A). We mapped CD4 CTL to the two-dimensional UMAP of single-cell transcriptome and found that the proportion of CD4 CTL in INR was significantly reduced compared with IR (Fig. 5B). We then compared these 2 groups and identified 440 differentially expressed genes of CD4 CTL. GO analysis of highly expressed genes in CD4 CTL derived from INR showed significant enrichment in cell over-activation (Fig. 5B).Fig. 5 Profile of Th cells in INRs. A Two-dimensional UMAP visualization of Th cells using the Seurat R package. Different colors represent two groups (up). Different colors represent different cell subsets (down). B Expression and proportion of cytotoxic genes in CD4 CTL. C GO and KEGG analyses conducted using DAVID for upregulated genes in CD4 CTL cells from INRs compared with IRs. The profile of CD8 effector T cells in INRs CD8 effector T cells are specific T cells that secrete various cytokines to participate in immune function and kill antigenic substances such as viruses and tumor cells. Our previous results showed no significant difference in the CD8 effector T-cell proportion between INR and IR. To better understand the function of CD8 effector T cells, we extracted all cells and subdivided them into 11 small cell clusters (C1–11) using the Seurat R package (Fig. 6A, B). The proportion of CD8 effector T cells C1 and C2 was significantly different in INR and IR, possibly due to different biological functions (Fig. 6C). We then compared these two groups and identified differentially expressed genes of CD8 effector T, CD8 effector T C1 and C2 (Fig. 6D–F). GO analysis of highly expressed genes in CD8 effector T C1 derived from INR showed significant enrichment in the type 1 interferon signaling pathway (Fig. 6G).Fig. 6 Profile of CD8 effector T cells in INRs. A UMAP visualization of CD8 effector T cells using the Seurat R package. Different colors represent two groups (left). Different colors represent different cell subsets (right) (B). C Proportion of CD8 effector T cells. D Differentially expressed genes in CD8 effector T cells between IRs and INRs ©. E Heatmap of top 40 genes significantly expressed in CD8 effector T C1. F Heatmap of top 40 genes expressed in CD8 effector T C2. G GO and KEGG were analyzed using DAVID for upregulated genes in CD4 CTL cells from INRs compared with IRs. Cell state transition of CD8 effector T cells, MAIT cells, and CD4 CTL during T-cell differentiation To understand the difference in T-cell differentiation between INRs and IRs, we constructed single-cell trajectories using the Monocle2 R package. According to the gene expression profile changes, all T cells, including CD8 effector T cells, MAIT cells and CD4 CTL, were placed on these trajectories (Fig. 7A). Consistent with the clustering analyses, there was continuity in the differentiation of MAIT and Th cells in IRs, while there was a gap in the differentiation trajectory in INRs (Fig. 7B). There was no significant difference in the differentiation trajectories of CD8 effector T cells in IRs and INRs (Fig. 7C).Fig. 7 The differentiation state of T cells in INRs and IRs. A Pseudotime trajectory of T cells estimated using Monocle 2. A continuous value from 0 to 12 was assigned to each cell as a pseudotime. (Right) Expression transition of differentiation-associated genes along the pseudotime. B Pseudotime trajectory of CD4 CTL, MAIT, and Th cells estimated using Monocle 2. (Right) Expression transition of differentiation-associated genes along the pseudotime. C Pseudotime trajectory of CD8 effector T cells estimated using Monocle 2. Identification of the cytokine profile for INR and IR To screen differential cytokines between INR and IR, L-serious 507 antibody-based protein microarrays were performed to measure the inflammatory cytokine expression profile. Expression levels of 62 cytokines were significantly different between the INR and IR patients (P < 0.05). The top 20 cytokines with the most significant differences are shown in Fig. 8A. To verify the cytokine microarray results, a liquid-suspension cytokine microarray was performed to detect concentrations of IL-4, MCP-1, IL-7, and IL-15 in plasma samples from IRs (n = 51) and INRs (n = 24) (Fig. 8B–E). GO and KEGG enrichment analysis further showed that the upregulated cytokines in plasma from INR subjects were enriched in cellular activation pathways, including cytokine–cytokine receptor interaction, inflammatory bowel disease etc, indicating that INRs sustained abnormal immune activation (Fig. 8F).Fig. 8 Cytokine profile between INR and IR. A Heatmap of significantly expressed cytokine between INRs and IRs using cytokine antibody microarray. B, C, D, and E Concentrations (pg/ml) of IL-4, IL-7, IL-15, and MCP-1 in the plasma samples obtained from INR and IR patients. F GO and KEGG analyses were conducted using DAVID for upregulated cytokine in plasma from INRs compared with IRs. Discussion Current evidence suggests that highly active antiretroviral therapy (HAART) can significantly reduce viral replication in HIV/AIDS patients, increase the CD4+ T lymphocyte count and reconstruct the immune function of patients [31]. However, some patients still experience immune reconstitution failure during viral inhibition [32]. Many factors are related to the failure of immune reconstitution in HIV/AIDS patients. Over the years, several retrospective studies have been conducted on patients with poor immune reconstitution. Overwhelming evidence substantiates that age, baseline CD4+ T lymphocyte level, HCV co-infection and other factors can affect the immune reconstitution of patients [33–36]. However, the mechanism underlying immune reconstitution failure is poorly understood. Single-cell sequencing technology is a new technology for high-throughput sequencing analysis of genome, transcriptome and epigenome at the single-cell level. It can reveal the gene structure and gene expression state of a single cell, reflect the heterogeneity between cells, and reveal the biological mechanism of disease. In this study, we sequenced the single-cell transcriptome of PBMC from INR and IR patients. Cluster analysis identified 11 cell subsets, including CD8 effector T, NK cell, T helper cell, B cell, classical monocytes, nonclassical monocytes, pDCs, mucosal-associated invariant T cell, platelets, CD4 naïve T, and CD8 naïve T. Single-cell transcriptome sequencing and flow cytometry results confirmed that the proportion of MAIT cells in INR decreased significantly compared with IR subjects. We found that MAIT cells derived from INR were dysregulated under the stimulation of E. coli and could not produce an interferon response. MAIT cells derived from INR are in a state of depletion, resulting in dysfunction. About 5% of T cells in healthy subjects are composed of MAIT cells, mainly used to control bacteria, but the immune system can also recruit them to fight viral infections. The results of scATAC-seq and flow cytometry showed that the low expression of TFAM and PPARA predominantly accounted for mitochondrial dysfunction in MAIT cells derived from INR. Furthermore, we identified a marked increase of cytotoxic CD4 T cells (CD4 cytotoxic T lymphocytes [CTLs]) in IR and INR subjects. In 1977, it was first reported that CD4 T cells cultured in vitro exhibited cytotoxicity. There are few studies on CD4 T cytotoxic cells, and much controversy surrounds the presence of CD4 T CTL in the human body. In 2001, sunI et al. reported that CD4+ CD8-T lymphocytes had strong cytokine expression and proliferation ability and showed cytotoxic activity when infected with HCV and HIV-1 virus. Since then, research on CD4 cytotoxic T cells has gradually attracted attention. In this study, we also found a cluster of CD4+ T lymphocytes with high specific expression of cytotoxicity-related genes, which further confirmed that there might be CD4+ CTL in the human body with potential cytotoxicity-related functions. Therefore, this cluster was defined as “CD4+ CTL”. In summary, we revealed the single-cell expression profiles in INR and IR by scRNA-seq and scATAC and established that low expression of TFAM accounted for mitochondrial dysfunction in MAIT derived from INR and the poor clinical outcomes. Materials and methods Human blood samples Human blood samples involved in this study were approved by the Ethics Committee of Chongqing Public Health Medical Center. Informed consent was obtained from all subjects. Immune non-responder and responder HIV-1 infected patients were recruited at the AIDS Outpatient Department of Geleshan Hospital of Chongqing Public Health Medical Center. The blood samples of healthy individuals came from Chongqing Public Health Medical Center (Pingdingshan hospital area). PBMCs were isolated from fresh whole blood in a 15 ml centrifuge tube using Ficoll-Paque PREMIUM (Cytiva, USA, 17544203) according to the manufacturer’s instructions. After mixing the blood sample with sterile PBS in equal volume, the mixture was slowly added to the Ficoll cell separation media and centrifuged 500 × g horizontally for 15 min at room temperature. Cell numbers and viability were measured using a hemocytometer with Trypan blue staining. scRNA-seq library construction The single-cell library was constructed using the ChromiumTM Controller and Chromium Single Cell 5ʹ Library Kit. Briefly, single cells, reagents and Gel Beads containing barcoded oligonucleotides were encapsulated into nanoliter-sized GEMs using the GemCode Technology. Lysis and barcoded reverse transcription of polyadenylated mRNA from single cells was performed inside each GEM. Post-RT-GEMs were cleaned up, and cDNA was amplified. cDNA was then fragmented, fragment ends were repaired, and A-tailing was added to the 3’ end. The adaptors were ligated to fragments that were double-sided SPRI selected. Another double-sided SPRI selection was carried out after sample index PCR. The quality and quantity of the final library were validated using two methods: checking the distribution of different fragment sizes using the Agilent 2100 bioanalyzer and quantifying the library using real-time quantitative PCR. The final products were sequenced using the MGISEQ (BGI-Shenzhen, China). Single-cell data analysis Cell Ranger Single Cell Software Suite (v3.1.0) was used to align complementary DNA reads to the reference genome. Single-cell FASTQ sequencing reads from each sample were processed and converted to digital gene expression matrices. The dataset was trimmed of cells with fewer than 200 genes. The number of genes, UMI counts and percentage of mitochondrial genes were examined to identify outliers. Principal component analysis was used for dimensionality reduction, followed by clustering in principal component analysis space using a graph-based clustering approach. UMAP was then used for two-dimensional visualization of the resulting clusters. For each cluster, the marker genes were identified using the FindConservedMarkers function implemented in the Seurat package (logFC. threshold > 0.25 and minPct > 0.25). Then, clusters were remarked to a known cell type according to Cell Marker database [27, 37, 38]. Differently expressed genes across different samples were identified using the FindConservedMarkers function in Seurat using the screening criteria: logFC. threshold > 0.25, minPct > 0.25, and adjusted P-value ≤0.05’. Pseudotime trajectory analysis was conducted with the R package Monocle2 [38]. GO and KEGG analysis GO and KEGG pathway analyses were performed using the phyper R function. Only GO terms or KEGG pathways with FDR ≤ 0.05 were significantly enriched. R package WGCNA was used to identify highly correlated genes using parameters “biweight midcorrelation > 0.7”. The highly correlated genes were used for the final correlation network construction. Antibodies and flow cytometry analysis Fluorochrome-conjugated monoclonal antibodies specific for human antigens were used anti-CD4-FITC (BD Bioscience, catalog 561005), anti-TCR Vα7.2-APC (BioLegend, catalog 351708), anti-IFN-γ-APC (BD Bioscience, catalog 562017), Anti-CD3-FITC (Bioscience, catalog 555339), Anti-CD161-PE (BD Bioscience, catalog 556081). Intracellular staining was performed using eBioscience™ Intracellular Fixation & Permeabilization Buffer Set (ThermoFisher) according to the manufacturer’s protocols. Cytokine antibody microarray assay The AAH-BLG-507 cytokine antibody microarray (Raybiotech, Norcross, GA, USA) was used to detect cytokines in plasma samples derived from IR and INR subjects. The detection of cytokines was performed according to the manufacturer’s instructions. Briefly, plasma was subjected to sample dialysis, sample labeling, antibody array blocking and incubation, and antibody array scanning detection. When the raw data were analyzed, the background removed fluorescent signal FI (F532 medium-B532 medium) was generally used for analysis, the mean value of two repetitions of all antibodies was calculated, and the mean value was used as the signal value of each antibody for subsequent analysis. All antibodies of each sample were normalized with all positive controls to obtain the normalized signal value. Generally, in the chip experiment results, for proteins with low signal value, a signal value FI < 25 was regarded as “not detected”. Luminex liquid-suspension microarray The concentrations of selected cytokines in the plasma samples derived from 51 IR subjects and 24 INR subjects were detected using a human high sensitivity T-cell magnetic bead panel (Millipore) according to the manufacturer’s instructions. The samples underwent incubation with magnetic beads, detection antibodies and Luminex200 detection. After the samples and standards tested in this experiment were detected by Luminex 200 detector, the fluorescence obtained was automatically detected and optimized by software to form the raw data. According to the fluorescence detection value obtained from the standard, the standard curve and its equation are obtained by fitting the standard curve with multi-parameter mode, and the concentration unit is pg/ml. The original fluorescence detected by each sample was substituted into the standard curve formula to calculate the sample concentration, used for comparison between samples. Statistical analysis All data are expressed as means ± SD. Statistical analyses were carried out using GraphPad Prism 8.0 software. P-value was calculated by t-test or one-way analysis. P-value < 0.05 was considered significant. Supplementary information Reproducibility checklist supplementary supplementary Table 1 Supplementary information The online version contains supplementary material available at 10.1038/s41419-022-05225-6. Acknowledgements This work was supported by the National Natural Science Foundation of China [No. 82002154], National Science and Technology Major Project of China during the 13th Five-year plan period [No. 2018ZX10302104] and the General Project of Chongqing Basic Research and Frontier Exploration Project (Natural Science Foundation), China [No. cstc2019jcyj-msxmX0029]. Author contributions HL, KL, FS, SD, and YC performed the experiment and wrote and drafted the manuscript. YT, YW, and YL retrieved the data and performed the statistical analysis. YY, YX, and KL performed the bioinformatics analysis. All authors read and approved the final manuscript. Data availability All datasets generated/analyzed for this study are included in this published article. Competing interests The authors declare no competing interests. Edited by Yufang Shi Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Change history 7/12/2023 A Correction to this paper has been published: 10.1038/s41419-023-05940-8 ==== Refs References 1. Grinsztejn B Hosseinipour M Ribaudo H Swindells S Eron J Chen Y Effects of early versus delayed initiation of antiretroviral treatment on clinical outcomes of HIV-1 infection: results from the phase 3 HPTN 052 randomised controlled trial Lancet Infect Dis 2014 14 281 90 10.1016/S1473-3099(13)70692-3 24602844 2. Lundgren J Babiker A Gordin F Emery S Grund B Sharma S Initiation of antiretroviral therapy in early asymptomatic HIV infection N. Engl J Med 2015 373 795 807 10.1056/NEJMoa1506816 26192873 3. Pierre S Jannat-Khah D Fitzgerald D Pape J McNairy M 10-year survival of patients with AIDS receiving antiretroviral therapy in Haiti N. Engl J Med 2016 374 397 8 10.1056/NEJMc1508934 26816026 4. Melgar M Nichols C Cavanaugh J Kirking H Surie D Date A Tuberculosis preventive treatment scale-up among antiretroviral therapy patients—16 countries supported by the U.S. President’s emergency plan for AIDS Relief, 2017-2019 MMWR Morbidity Mortal Wkly Rep 2020 69 329 34 10.15585/mmwr.mm6912a3 5. Ganatra S Bucşan A Alvarez X Kumar S Chatterjee A Quezada M Antiretroviral therapy does not reduce tuberculosis reactivation in a tuberculosis-HIV coinfection model J Clin Investig 2020 130 5171 9 10.1172/JCI136502 32544085 6. Guihot A Bourgarit A Carcelain G Autran B Immune reconstitution after a decade of combined antiretroviral therapies for human immunodeficiency virus Trends Immunol 2011 32 131 7 10.1016/j.it.2010.12.002 21317040 7. Autran B Carcelaint G Li TS Gorochov G Blanc C Renaud M Restoration of the immune system with anti-retroviral therapy Immunol Lett 1999 66 207 11 10.1016/S0165-2478(98)00159-X 10203056 8. Grabar S Le Moing V Goujard C Leport C Kazatchkine M Costagliola D Clinical outcome of patients with HIV-1 infection according to immunologic and virologic response after 6 months of highly active antiretroviral therapy Ann Intern Med 2000 133 401 10 10.7326/0003-4819-133-6-200009190-00007 10975957 9. Yang X Su B Zhang X Liu Y Wu H Zhang T Incomplete immune reconstitution in HIV/AIDS patients on antiretroviral therapy: Challenges of immunological non-responders J Leukoc Biol 2020 107 597 612 10.1002/JLB.4MR1019-189R 31965635 10. Powderly W Landay A Lederman M Recovery of the immune system with antiretroviral therapy: the end of opportunism? JAMA 1998 280 72 7 10.1001/jama.280.1.72 9660367 11. Engsig F Gerstoft J Kronborg G Larsen C Pedersen G Røge B Long-term mortality in HIV patients virally suppressed for more than three years with incomplete CD4 recovery: a cohort study BMC Infect Dis 2010 10 318 10.1186/1471-2334-10-318 21044307 12. Monforte A Abrams D Pradier C Weber R Reiss P Bonnet F HIV-induced immunodeficiency and mortality from AIDS-defining and non-AIDS-defining malignancies AIDS (Lond, Engl) 2008 22 2143 53 10.1097/QAD.0b013e3283112b77 13. Bowen L Smith B Reich D Quezado M Nath A HIV-associated opportunistic CNS infections: pathophysiology, diagnosis and treatment Nat Rev Neurol 2016 12 662 74 10.1038/nrneurol.2016.149 27786246 14. Tan I Smith B von Geldern G Mateen F McArthur J HIV-associated opportunistic infections of the CNS Lancet Neurol 2012 11 605 17 10.1016/S1474-4422(12)70098-4 22710754 15. Picat M Lewis J Musiime V Prendergast A Nathoo K Kekitiinwa A Predicting patterns of long-term CD4 reconstitution in HIV-infected children starting antiretroviral therapy in sub-Saharan Africa: a cohort-based modelling study PLoS Med 2013 10 e1001542 10.1371/journal.pmed.1001542 24204216 16. Cribbs S Crothers K Morris A Pathogenesis of HIV-related lung disease: immunity, infection, and inflammation Physiological Rev 2020 100 603 32 10.1152/physrev.00039.2018 17. Baker J Peng G Rapkin J Abrams D Silverberg M MacArthur R CD4+ count and risk of non-AIDS diseases following initial treatment for HIV infection AIDS (Lond, Engl) 2008 22 841 8 10.1097/QAD.0b013e3282f7cb76 18. Kelley C Kitchen C Hunt P Rodriguez B Hecht F Kitahata M Incomplete peripheral CD4+ cell count restoration in HIV-infected patients receiving long-term antiretroviral treatment Clin Infect Dis: Off Publ Infect Dis Soc Am 2009 48 787 94 10.1086/597093 19. Shive C Mudd J Funderburg N Sieg S Kyi B Bazdar D Inflammatory cytokines drive CD4+ T-cell cycling and impaired responsiveness to interleukin 7: implications for immune failure in HIV disease J Infect Dis 2014 210 619 29 10.1093/infdis/jiu125 24585897 20. Awad W Ler G Xu W Keller A Mak J Lim X The molecular basis underpinning the potency and specificity of MAIT cell antigens Nat Immunol 2020 21 400 11 10.1038/s41590-020-0616-6 32123373 21. Le Bourhis L Martin E Péguillet I Guihot A Froux N Coré M Antimicrobial activity of mucosal-associated invariant T cells Nat Immunol 2010 11 701 8 10.1038/ni.1890 20581831 22. Toubal A Nel I Lotersztajn S Lehuen A Mucosal-associated invariant T cells and disease Nat Rev Immunol 2019 19 643 57 10.1038/s41577-019-0191-y 31308521 23. Legoux F Bellet D Daviaud C El Morr Y Darbois A Niort K Microbial metabolites control the thymic development of mucosal-associated invariant T cells Science (NY) 2019 366 494 9 10.1126/science.aaw2719 24. Koay H Gherardin N Enders A Loh L Mackay L Almeida C A three-stage intrathymic development pathway for the mucosal-associated invariant T cell lineage Nat Immunol 2016 17 1300 11 10.1038/ni.3565 27668799 25. Jeffery H van Wilgenburg B Kurioka A Parekh K Stirling K Roberts S Biliary epithelium and liver B cells exposed to bacteria activate intrahepatic MAIT cells through MR1 J Hepatol 2016 64 1118 27 10.1016/j.jhep.2015.12.017 26743076 26. Legoux F Salou M Lantz O MAIT cell development and functions: the microbial connection Immunity 2020 53 710 23 10.1016/j.immuni.2020.09.009 33053329 27. Zhang X Lan Y Xu J Quan F Zhao E Deng C CellMarker: a manually curated resource of cell markers in human and mouse Nucleic acids Res 2019 47 D721 D728 10.1093/nar/gky900 30289549 28. Milasta S Dillon C Sturm O Verbist K Brewer T Quarato G Apoptosis-inducing-factor-dependent Mitochondrial Function Is Required for T Cell But Not B cell function Immunity 2016 44 88 102 10.1016/j.immuni.2015.12.002 26795252 29. Bock F Tait S Mitochondria as multifaceted regulators of cell death Nat Rev Mol cell Biol 2020 21 85 100 10.1038/s41580-019-0173-8 31636403 30. Liang J Cao R Wang X Zhang Y Wang P Gao H Mitochondrial PKM2 regulates oxidative stress-induced apoptosis by stabilizing Bcl2 Cell Res 2017 27 329 51 10.1038/cr.2016.159 28035139 31. Katz I Leister E Kacanek D Hughes M Bardeguez A Livingston E Factors associated with lack of viral suppression at delivery among highly active antiretroviral therapy-naive women with HIV: a cohort study Ann Intern Med 2015 162 90 9 10.7326/M13-2005 25599347 32. Lewden C, Chene G, Morlat P, Raffi F, Dupon M, Dellamonica P, et al. HIV-infected adults with a CD4 cell count greater than 500 cells/mm3 on long-term combination antiretroviral therapy reach same mortality rates as the general population. J Acquir Immune Defic Syndr. 2007;46:72–7. 33. Ahuja S Kulkarni H Catano G Agan B Camargo J He W CCL3L1-CCR5 genotype influences durability of immune recovery during antiretroviral therapy of HIV-1-infected individuals Nat Med 2008 14 413 20 10.1038/nm1741 18376407 34. Younes S Talla A Pereira Ribeiro S Saidakova E Korolevskaya L Shmagel K Cycling CD4+ T cells in HIV-infected immune nonresponders have mitochondrial dysfunction J Clin Investig 2018 128 5083 94 10.1172/JCI120245 30320604 35. Scherpenisse M, Kootstra N, Bakker M, Berkhout B, Pasternak A. Cell-associated HIV-1 unspliced-to-multiply-spliced RNA ratio at 12 weeks of ART predicts immune reconstitution on therapy. mBio. 2021;12:e00099-21. 36. Lu CM, Dezube BJ, Pantanowitz L. HIV infection masquerading as monoclonal gammopathy of unknown significance. N Engl J Med. 2003;349: 1192–3. 37. Butler A Hoffman P Smibert P Papalexi E Satija R Integrating single-cell transcriptomic data across different conditions, technologies, and species Nat Biotechnol 2018 36 411 20 10.1038/nbt.4096 29608179 38. Qiu X Hill A Packer J Lin D Ma Y Trapnell C Single-cell mRNA quantification and differential analysis with Census Nat Methods 2017 14 309 15 10.1038/nmeth.4150 28114287
PMC009xxxxxx/PMC9678860.txt
==== Front Sci Rep Sci Rep Scientific Reports 2045-2322 Nature Publishing Group UK London 36411304 24478 10.1038/s41598-022-24478-x Article The association between expressive language skills and adaptive behavior in individuals with Down syndrome del Hoyo Soriano Laura [email protected] 12 Villarreal Jennifer Catalina 12 Sterling Audra 3 Edgin Jamie 4 Berry-Kravis Elizabeth 5 Hamilton Debra R. 6 Thurman Angela John 12 Abbeduto Leonard 12 1 grid.27860.3b 0000 0004 1936 9684 MIND Institute, University of California Davis, 2825 50Th Street, Sacramento, CA USA 2 grid.27860.3b 0000 0004 1936 9684 Department of Psychiatry and Behavioral Sciences, Davis Health, University of California, Sacramento, CA USA 3 grid.14003.36 0000 0001 2167 3675 Waisman Center and Department of Communication Sciences and Disorders, University of Wisconsin-Madison, Madison, WI USA 4 grid.134563.6 0000 0001 2168 186X Department of Psychology, University of Arizona, Tucson, AZ USA 5 grid.240684.c 0000 0001 0705 3621 Departments of Pediatrics, Neurological Sciences and Biochemistry, Rush University Medical Center, Chicago, IL USA 6 grid.189967.8 0000 0001 0941 6502 Department of Human Genetics, Emory University School of Medicine, Atlanta, GA USA 21 11 2022 21 11 2022 2022 12 2001422 7 2022 16 11 2022 © The Author(s) 2022 https://creativecommons.org/licenses/by/4.0/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The primary goal of this study was to determine whether expressive language skills contribute to adaptive behavior (e.g., socialization and daily living skills) in children, adolescents, and young adults with Down syndrome (DS) whilst controlling for age and nonverbal cognitive ability. Expressive language was assessed using the psychometrically validated Expressive Language Sampling (ELS) conversation and narration procedures. The language produced was transcribed and analyzed to yield measures of expressive vocabulary, syntax, and intelligibility. Socialization and daily living skills of participants with DS were measured with the Vineland Adaptive Behavior Scales, 2nd edition (VABS-2) parent/caregiver rating form. Our results show that the three ELS measures were significantly correlated with multiple measures from the VABS-2 when controlling for age. Several correlations remained significant even when nonverbal cognitive ability was included as a control variable. Our results suggest that expressive language skills contribute to adaptive behavior in children, adolescents, and young adults with DS regardless of age and some of these associations are not explained solely by overall cognitive delays. Further studies including longitudinal data are needed to extend our results. Subject terms Genetics Psychology Human behaviour Quality of life National Institutes of HealthR01HD074346, P50HD103526, and UL1TR001860 R01HD074346, P50HD103526, and UL1TR001860 R01HD074346, P50HD103526, and UL1TR001860 R01HD074346, P50HD103526, and UL1TR001860 R01HD074346, P50HD103526, and UL1TR001860 R01HD074346, P50HD103526, and UL1TR001860 R01HD074346, P50HD103526, and UL1TR001860 del Hoyo Soriano Laura Sterling Audra Edgin Jamie Berry-Kravis Elizabeth Hamilton Debra R. Thurman Angela John Abbeduto Leonard issue-copyright-statement© Springer Nature Limited 2022 ==== Body pmcIntroduction Down syndrome (DS) is the leading genetic cause of intellectual disability (ID), with approximately 1 in 691 live births affected1. The syndrome is typically caused by an extra copy of all or part of chromosome 212, which results in a complex condition that affects both physical (e.g., facial dysmorphology, a disproportionately large tongue) and cognitive development3. Indeed, DS is associated with prototypical pattern of cognitive and behavioral strengths and challenges. More specifically, when compared either to individuals with typical development (TD) or with other neurodevelopmental disabilities (NDDs) of similar mental ages (MAs), individuals with DS demonstrate, as a group, relative strengths in the areas of nonverbal communication, gross motor skills, visual motor integration, and visual imitation4,5. In contrast, auditory short-term memory, episodic memory, aspects of visuospatial construction and executive function, and language are areas that are particularly challenging for individuals with DS6–12. Language skills constitute one of the most impaired aspects of the cognitive profile of individuals with DS. Research has also shown, however, that within the area of language some aspects are generally more affected than others among those with DS9. For example, receptive language is usually an area of relative strength compared to expressive language skills (i.e., syntax, vocabulary, phonology). Indeed, there is considerable research documenting that multiple areas of expressive language lag behind MA expectations13–15. For example, individuals with DS score less well on measures of vocabulary and syntax when compared to younger TD peers10,16–19 and similarly-aged peers with other developmental disabilities, both matched on MA17,20,21. There is also evidence suggesting that expressive syntax skills are more impaired than are expressive vocabulary skills in individuals with DS22 as expressive syntax appears to lag relative to both nonverbal cognition and vocabulary level-expectations13–15,23,24. Therefore, severity of impairment also varies across different areas of expressive language in DS. In addition, hearing loss and differences in oral-facial anatomy also negatively impact speech perception and production in individuals with DS25, contributing to high rates of unintelligible speech. Even areas of relative strength are delayed relative to expectations based on chronological age (CA). Nonetheless, there is considerable heterogeneity in terms of level and profile of strengths and weaknesses among individuals with DS26,27. Adaptive behavior is the collection of conceptual, social, and practical skills that are learned through experiences and are essential to meeting the demands of everyday life28. Consistent with the definition of ID, adaptive behavior is also an area of challenge for individuals with DS29. That is, ID is, by definition, characterized by deficits in cognitive function that would impact how the individual functions in their everyday life and, therefore, cognitive deficits are highly associated with difficulties in adaptive behavior. Interestingly, previous research in DS using the Vineland Adaptive Behavior Scales (VABS)30 has shown that children and adolescents may demonstrate better adaptive behavior performance, on average, in the domain of socialization than in the domain of daily living skills29,31,32. Language skills are an important factor associated with the level of adaptive behavior achieved by individuals with DS. Indeed, language challenges appear to be a substantial barrier to independent functioning and meaningful inclusion in the community for individuals with DS16. A study of adolescents and young adults with DS33 documented associations between expressive and receptive language and adaptive behavior in functional skill areas such as communication, community use, and functional academics, as well as in the overall score for adaptive behavior as measured with the Adaptive Behavior Assessment System, Second Edition (ABAS-2)34. In fact, among other cognitive constructs evaluated (e.g., attention, memory, executive function), language was the skill most closely related to higher competence in overall adaptive skills for participants with DS33. It also has been found that the speech scale (which assesses articulation skills and phonology) of the parent-report Children's Communication Checklist, Second Edition (CCC-2)35, which assesses receptive and expressive language (speech, syntax, semantics, and coherence) and pragmatics, accounted for variance in adaptive behavior assessed with the ABAS-2 in adolescents with DS36. The close association between language, particularly expressive language, and daily functioning has also been documented for other developmental disabilities such as fragile X syndrome (FXS)37. Because expressive language is one of the most impaired aspects of the cognitive profile of individuals with DS, and consists of diverse but related constructs (e.g., syntax, vocabulary, intelligibility) affected at variable levels in DS and possibly differentially allied to different dimensions of adaptive behavior, exploring these concurrent relationships is of special interest in this population. Examining the relationships between specific aspects of adaptive behavior (i.e., socialization and daily living skills) and specific aspects of expressive language (e.g., syntax, vocabulary, intelligibility) can help identify the language skills to be targeted in interventions designed to support greater independent functioning in those with DS. The aim of the present study was to examine these concurrent relationships for children, adolescents, and young adults with DS. In doing so, we assessed dimensions of spoken language (i.e., syntax, vocabulary, intelligibility) likely to be related to socialization (i.e., interpersonal relationships, play and leisure time, coping skills) and daily living skills (i.e., personal care, domestic skills, and community use). We measured expressive language using a set of expressive language sampling (ELS) procedures, which have been psychometrically validated for use in DS38 and in other NDDs such as FXS37. ELS procedures entail collecting and analyzing relatively brief samples of spoken language in a naturalistic contexts that are representative of the individual’s “everyday” language activities37. The particular ELS procedures we used have been standardized to ensure consistency of measurements across individuals and occasions of assessment37. Importantly, there are multiple advantages to using ELS, compared to typical standardized assessments of language skills, when considering the link between expressive language and adaptive behavior in DS. For example, ELS procedures (1) use a format more closely aligned with real-world contexts and, therefore, are more likely to generalize to activities that are functional and meaningful for the participant39; (2) can yield multiple dependent variables, reflecting different domains of skill, that can be examined separately; (3) are less prone to noncompliance and floor effects as compared to standardized tests; and (4) can be collected quickly and often with minimal training of examiners, making ELS especially attractive for multi-site studies38,40. The research questions and hypotheses of the current study were:Are expressive language skills related to concurrent levels of adaptive behavior in the domains of daily living skills and socialization in children, adolescents and young adults with DS whilst controlling for CA? It was hypothesized that stronger expressive language skills, as assessed through the ELS procedures, would be associated with greater levels of adaptive behavior regardless of age. Such statistical control was needed given the anticipated correlation between age and expressive language skill, with the latter improving with age even among individuals with DS. Are expressive language skills associated with specific aspects of daily living and socialization skills whilst controlling for age and nonverbal cognitive ability in children, adolescents, and young adults with DS? It was hypothesized that expressive language skills would be associated with specific socialization and daily living skills even after controlling for both CA and nonverbal cognitive ability. Such statistical control was needed given the anticipated correlation of nonverbal cognition and expressive language skill, with the latter improving with increases in cognitive ability in individuals with DS, as in individuals with TD. Methods The data for the current study were collected as part of a larger multi-site study evaluating the psychometric properties of variables derived from ELS procedures37,38. Study procedures were reviewed and approved by Institutional Review Boards at all participating university sites located in Arizona, Georgia, California, and Wisconsin. Written informed consent was obtained from caregivers, and verbal assent was obtained from the youth with DS prior to beginning study procedures. The authors affirm that all procedures contributing to this work comply with the ethical standards of the relevant national and international committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008. All data for the present study were collected at each participant’s initial annual visit. Data source and study sample A total of 95 participants with DS between the ages of 7 and 23 years were included in the current study. The same inclusion criteria were followed in the current study as in the larger project. All participants provided medical documentation of DS (i.e., trisomy 21 or translocation) without mosaicism and met IQ criteria for ID (IQ ≤ 70). Based on parental report, the participant with DS also had to use speech as the primary mode of communication, with English as their primary language given that the ELS procedures were at that time available only in English. In addition, the participant could not have serious (uncorrected) visual and/or hearing impairments and could not be enrolled in a randomized controlled clinical trial or experience changes in medication, treatment, or educational interventions/programs within the 8 weeks prior to the initial visit. Participants who were non-compliant with the two ELS procedures used were excluded from the current analyses (see below for operationalization of compliance). Thus, a total of 10 participants (3 non-compliant in conversation, 6 in narration and 1 in both procedures) were excluded from the present analyses. Characteristics of participating youth with DS are described in Table 1.Table 1 Characteristics of participating youth with DS (n = 95). Measure M SD Range Chronological age (years) 15.89 4.92 7.05–23.72 SB-5 NV fluid reasoning raw score 10.63 4.55 1–23 SB-5 nonverbal deviation IQ 40.62 12.99 7.43–73.03 Distribution of participants Race Ethnicity Sex Family Income African American -7 Hispanic/Latinx -20 Female -50 USD < 25,000 -5 Asian Pacific islander- 2 Non-Hisp/Latinx- 75 Male -45 USD 25,000–50,000 -20 White -61 Multiracial -6 USD 50,001–75,000 -14 Other- 3 USD 75,001–100,000 -16 Not reported- 16 USD 100,001–150,000 -20 USD 150,001–550,000 -10 USD > 250,001 -6 Not reported- 14 Measures The measures reported here are a subset of a larger battery of direct assessments, questionnaires, and interviews from the project. The measures for the present study were the ELS procedures, the Stanford-Binet Intelligence Scales, Fifth Edition (SB-5) to assess nonverbal cognitive ability of participants, and the Vineland Adaptive Behavior Scales, Second Edition (VABS-2) to address socialization and daily living skills through parent/caregiver report. Expressive language sampling Expressive language samples (ELS) were collected in two contexts—conversation and narration—from each participant. Conversation was always administered before narration, and each participant completed other measures between the two ELS procedures. Manuals describing ELS administration, training, and the assessment of fidelity are available at https://ctscassist.ucdmc.ucdavis.edu/ctscassist/surveys/?s=W9W99JLMNX. The conversation consists of a 12-min interview-style interaction with a trained examiner. The examiner relies primarily on open-ended prompts to topics (e.g., “Tell me everything you did at school yesterday”) and broad follow-up questions and prompts (e.g., “What do you like about school?”) to encourage participant talk while minimizing their own talk. The examiner introduces predetermined topics in a standard order. The goal is to introduce at least three topics in addition to an initial idiosyncratic topic reflecting an interest of the participant (according to parent/caregiver report). Two alternate versions of the conversation task were administered, with different topics in each. Approximately half the participants received Version A and half Version B. Additional details of the conversation procedures can be found in Abbeduto et al.37 The narration consists of the participant telling the story depicted in a wordless picture book. The participant first looks at each page spread of the book without talking to gain a sense of the story. The participant then tells the story page by page, with the examiner controlling the page turning. The examiner’s prompts and responses are standardized and limited largely to the first page of the book. There is no set time limit for the narration task. Two books, each including 16 pages of story content from the Mercer Mayer’s “Frog” series, were used: Frog Goes to Dinner (Version A) and Frog on His Own (Version B). Although these books were created for children, they are actually rather sophisticated in the sense that they allow for a range of different types of language, form very concrete description of actions to explanation of character motivation, mental states, and emotions and have been used successfully even with TD adolescents and adults41. Approximately half the participants received Version A and half Version B. Non-compliance in the ELS procedures was defined as refusal to complete the procedures, no response, or repeated off-task behavior (e.g., saying, “I’m done” or refusing to talk) or, in conversation, failing to engage in talk for at least 9.5 min or, in narration, failing to produce relevant talk on at least 12 of the 16 pages in the book. All ELS sessions were digitally audio recorded and analyzed using SALT: Systematic Analysis of Language Transcripts42. All transcripts were prepared by a primary transcriber and reviewed by a secondary transcriber before being finalized. Transcribers were blind to individual participant results for other measures. Talk was segmented into C-units, the upper bound of which is an independent clause and any modifiers. Inter-transcriber agreement was randomly assessed for 10 transcripts (4 Narration, 6 Conversations), with at least two from each site. In the larger study, inter-transcriber was 87% for utterance segmentation, 87% for identification of partly or fully unintelligible C-units, and 84% for identification of the exact lexical and morphemic content of each C-unit. In addition, inter-transcriber agreement was 76% for identification of the exact number of morphemes in each C-unit and 80% for the exact number of words in each C-unit38. We focused on the three ELS outcome measures shown to have the strongest psychometric properties37,38. The measures were: (1) Lexical Diversity, which indexes the size of the participant’s expressive vocabulary and is operationalized as the number of different word roots in 50 complete and fully intelligible C-units (or the full sample of complete and fully intelligible C-units if the participant produces fewer than 50 C-units). Higher scores indicate more advanced expressive vocabulary. (2) Syntax, which indexed expressive syntactic complexity and is computed as the mean length of C-unit measured in morphemes (MLU) for complete and fully intelligible C-units. Higher scores indicate more advanced expressive syntax. (3) Unintelligibility, which is an index of speech articulation problems and is computed as the proportion of the total C-units that are either partly or fully unintelligible to the transcriber. Higher scores indicate more problems with articulation. A composite score for each measure was derived by first computing scores separately for conversation and narration tasks and then averaging performance across the two tasks. In the larger study, conversation and narration were readministered after 4 weeks to assess practice effects and test–retest reliability. Minimal practice effects were observed, and test–retest reliability was very strong for each of the 10 ELS measures for participants with DS, with the intraclass correlations ranging from 0.79 to 0.95 with all but one value above 0.83 and all p values > 0.00538. See Thurman et al. 2020 for detailed values. Note that each participant received the alternate versions of the conversation and narration materials in the two administrations, suggesting comparability of the versions. Importantly, comparability of version with a different sample of participants had been established previously43. Nonverbal cognitive ability The Stanford-Binet Intelligence Scales, Fifth Edition (SB-5)44 was used to assess nonverbal cognitive ability. In our analyses, we used the Nonverbal Fluid Reasoning (NVFR) raw score derived from the Object Series/Matrices subtest as the control variable in assessing whether expressive language skills were associated with concurrent independent functioning. Note that we present the raw NVFR in Table 1. The deviation NVIQ is also derived from Object Series/Matrices subtest, indexing Nonverbal Fluid Reasoning (NVFR), and it is a z-score transformation based on the general population norms. The score is calculated following procedures outlined by Sansone and colleagues45 to avoid floor effects and obtain a normal distribution of the measurement for our sample of participants. This score is also presented in Table 1 for descriptive purposes. Note that we used NV raw scores rather than standard scores because interest was in the association between the absolute levels of ability achieved in expressive language and nonverbal cognition. Socialization and daily living skills The Vineland Adaptive Behavior Scales-2nd edition (VABS-2)46 is a caregiver report measure of adaptive behavior. The measure was normed on individuals aged birth to 90 years, including individuals with ID. In the present study, we used raw scores for the three Socialization subdomains—Interpersonal Relationships (i.e., how the individual interacts with others), Play and Leisure (i.e., how the individual plays and uses leisure time), and Coping Skills (i.e., how the individual demonstrates responsibility and sensitivity to others)—as well as for the three Daily Living subdomains—Personal (i.e., how the individual manages tasks such as dressing and personal hygiene), Domestic (i.e., how the individual manages household tasks such as cleaning and laundry) and Community (i.e., how the individual manages time, money, technology, and employment opportunities). Raw scores range from 0 to near 130 depending on the subdomain, with higher scores reflecting greater ability. We focused on the subdomains rather than the superordinate domains of socialization and daily living because Shaffer et al.47 did not find any significant correlations of the superordinate domains with the ELS measures for their sample of participants with FXS. In addition, we were interested in evaluating the possibility that there are differential patterns of associations with the ELS measures and the specific adaptive skills represented in the various subdomains of the VABS-2. The VABS-2 has strong psychometric properties, including significant correlations with earlier versions of the VABS48 and the ABAS-2 and strong test–retest reliabilities for the six subdomains of interest in the norming sample (e.g., average correlations range between 0.76 and 0.92 across domains). Statistical analysis plan First, descriptive statistics were computed for all variables of interest. We also examined the variables and their residuals for the assumptions of normality required in the parametric tests. Most VABS-2 variables violated assumptions of normality and transformations did not improve the distribution; thus, nonparametric correlations were implemented. To address the first research question, we computed partial Spearman correlations between each of the three ELS measures (derived from averaging the scores for conversation and narration) and each of the socialization and daily living skills raw scores while controlling for CA. To address the second research question, we computed nonparametric partial correlations between the ELS variables and the socialization and daily living skills raw scores while controlling for both CA and NV cognitive ability. In each set of analyses, one-tailed inferential tests were used in all cases because we had clear hypotheses about the directionality of the relationships of interest. Familywise alpha levels for the correlations and for the partial correlations were maintained at p < 0.050 levels through application of Benjamini and Hochberg’s False Discovery Rate (FDR) procedures49. Ethical approval This study was approved by the Institutional Review Boards (IRB) of each participating site: University of California Davis IRB Clinical Committee B (IRB403210), Education and Social/Behavioral Science IRB at the University of Wisconsin-Madison (IRB2013-0512), Rush University Medical Center IRB (IRB11112301), Emory University IRB (IRB00065271 UCDMC Language Study) and University of Arizona IRB (IRB1300000331). The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008. Informed consent was obtained from the parent or guardian of each participant before testing. Results Descriptive statistics Means and standard deviations are presented for the ELS measures and the measures of adaptive behavior in Tables 2 and 3, respectively.Table 2 Means and standard deviations for ELS measures. Measure Conversation Narration Combineda N M SD (range) M SD (range) M SD (range) Syntax 3.37 1.55 (1.03–6.78) 4.43 2.09 (1–8.76) 3.91 1.74 (1.16–7.22) 95 Lexical diversity 73.76 33.53 (10–154) 62.21 35.13 (1–141) 68.27 32.45 (7.5–135) 95 Unintelligibility 0.25 0.19 (0.02–0.77) 0.23 0.21 (0–0.78) 0.24 0.18 (0.01–0.7) 95 n = 95; aAverage of conversation plus narration. Table 3 Means and standard deviations for VABS-2 measures. Measure Raw Score Standardized Score N M SD M SD Interpersonal relationships subdomain 60.24 8.604 10.75 4.706 89 Play and leisure subdomain 46.25 10.274 9.95 4.321 91 Coping skills subdomain 40.22 11.161 11.48 2.259 82 Personal subdomain 65.52 11.837 9.73 3.662 92 Domestic subdomain 26.87 9.896 10.63 4.938 93 Community use subdomain 45.14 15.943 8.95 6.334 86 Socialization domain – – 72.25 14.27 80 Daily living skills domain – – 67.61 11.14 84 Adaptive behavior composite – – 71.89 27.32 72 *Note that missing values were due to parents not completing certain questions on each specific VABS-2 subdomain. Primary analyses As seen in Table 4, nearly all the ELS measures were significantly correlated with the functional measures of adaptive behavior while controlling for CA, even after applying the FDR, with better performance on the ELS measures associated with greater levels of socialization and daily living skills. As seen in Table 4, correlations with the Personal and Domestic subdomains were weak. In contrast, ELS measures showed strong correlations with Interpersonal Relationships, Coping, and Community Use subdomains. None of the ELS measures were correlated significantly with the Play and Leisure subdomain raw score on the VABS-2. In addition, the correlation between ELS Unintelligibility and the VABS-2 Interpersonal Relationships subdomain just failed to reach significance (p = 0.051).Table 4 Nonparametric partial correlations between ELS and VABS-2 raw scores controlling for chronological age. Measures VABS-2 VABS-2 VABS-2 VABS-2 VABS-2 VABS-2 Interpersonal relationships Play and leisure Coping Personal Domestic Community use Syntax 0.33*** 0.17 0.42**** 0.26* 0.23* 0.51**** df 85 88 79 89 90 83 Lexical Diversity 0.39**** 0.05 0.46**** 0.21* 0.23* 0.54**** df 85 88 79 89 90 83 Unintelligibility  − 0.21  − 0.13  − 0.30**  − 0.23*  − 0.25**  − 0.37**** df 85 88 79 89 90 83 *p ≤ 0.050, **p ≤ 0.010, ***p ≤ 0.005, ****p ≤ 0.001, *****p ≤ 0.0005. Missing values were due to parents not completing certain questions on each specific VABS-2 subdomain. Values in bold are significant at p ≤ 0.05 after FDR correction. Partial correlations between the ELS and selected measures of adaptive behavior controlling for both CA and nonverbal ability (SB-5 NVFR raw score) are presented in Table 5. When including the SB-5 NVFR as a control variable along with CA, fewer correlations reached statistical significance in comparison to those represented in Table 4. In particular, after application of FDR, Syntax and Lexical Diversity were both still significantly correlated with raw scores on the Interpersonal Relationships (moderate to strong), the Coping Skills (strong), and Community Use (moderate to strong) subdomains, whereas Unintelligibility was significantly correlated with Coping skills and Community Use only (mild to moderate). Finally, raw scores on the Play and Leisure, Personal, and Domestic VABS-2 subscales were not correlated with any of the ELS measures.Table 5 Nonparametric partial correlations between ELS and VABS-2 raw scores controlling for chronological age and the SB-5 NVFR raw score. Measures VABS-2 VABS-2 VABS-2 VABS-2 VABS-2 VABS-2 Interpersonal relationships Play and leisure Coping Personal Domestic Community use Syntax 0.29** 0.11 0.35**** 0.20 0.17 0.37**** df 83 86 77 87 88 81 Lexical Diversity 0.35**** 0.11 0.40**** 0.16 0.15 0.39**** df 83 86 77 87 88 81 Unintelligibility  − 0.18  − 0.08  − 0.26*  − 0.20  − 0.22*  − 0.31*** df 83 86 77 87 88 81 *p ≤ 0.050, **p ≤ 0.010, ***p ≤ 0.005, ****p ≤ 0.001, *****p ≤ 0.0005. Missing values were due to parents not completing certain questions on each specific VABS-2 subdomain. Values in bold are significant at p ≤ 0.05 after FDR correction. Discussion The goal of this study was to further understand the relationship between components of expressive language skill and specific aspects of adaptive behavior in children, adolescents, and young adults with DS whilst controlling for CA and NV cognitive ability. We found that all but 4 of the 18 bivariate correlations between the ELS measures and the parent/caregiver-reported measures of adaptive behavior were significant when controlling for CA. This pattern of findings is consistent with previous research in DS showing that more advanced language skills are related with more functional adaptive behaviors33,36,50, such as advanced skills for social and daily living functioning. The relationships among some of the constructs of interest, however, could be bidirectional. For example, more frequent and closer interpersonal relationships could create opportunities for practicing and acquiring new expressive language skills. The concurrent nature of the present data does not allow us to address these bidirectional possibilities. However, our data do establish the important link between expressive language abilities and the capacity for meaningful engagement in socialization and management of the tasks of daily living. This suggests a need for further research including longitudinal data. Findings for our first research question indicate that limited expressive syntax and lexical diversity, as well as greater problems in intelligibility, are each associated with a less well-developed capacity for interpersonal relations, coping skills and daily living skills in children, adolescents, and young adults with DS. Interestingly, each of the expressive language skills measured in our study (e.g., syntax, lexical diversity, and unintelligibility) is associated with each of the subdomains of daily living skills (e.g., personal, domestic and community use skills). Importantly, research shows that people with DS with higher levels of daily living skills are more likely to participate in open employment-related activities and to eventually acquire a job, which is in turn related to a better quality of life for the family51. Therefore, interventions in DS with a special emphasis on expressive language should be considered as this could have a long-term impact on the level of independence achieved due to the highly verbal nature of the normative tasks of adulthood. With regards to the relationship between expressive language and socialization skills, there is a correlation with interpersonal relations and coping skills (or how the individual demonstrates responsibility and sensitivity to others). This pattern of findings may indicate that better expressive language skills promote social interactions and meaningful relationships or that greater socialization opportunities may facilitate development of expressive language skills. Note, however, that none of the ELS variables was related to play and leisure, which is surprising as previous research has shown a link between expressive language and play and leisure abilities in children with DS52, as well as in older adolescents and young adults with FXS53. In this regard, the ability to play and use leisure time in DS could be more closely aligned with other aspects of communication (i.e., pragmatics) than the structural language skills assessed in this study, which could be one of the reasons why we do not see any relationship between domains. It could also be that the association between expressive language and play and leisure skills is age-dependent in DS, and we have not observed such association in our sample as our age range was wide enough to control for CA given that we used raw scores in our analyses. In terms of our second research question, we found that when controlling for CA and NV cognitive ability, far fewer correlations (8 out of the 18) remained significant. Specifically, only the links between ELS measures, Interpersonal Relationships, Coping and Community Use remained significant after controlling for NV cognition and age. These results suggest that specific expressive language skills make unique contributions to the prediction of specific aspects of adaptive functioning in individuals with DS over and above the contribution of nonverbal cognitive ability. It appears that expressive language skills are closely related to skills in the areas of interpersonal relationships, coping, and the use of community resources. Thus, how the individual participates in the community (e.g., how they use time, money, the telephone, public transportation, gain and maintain employment, or complete school tasks) might require expressive language skills to a greater degree than do domestic skills (e.g., cooking, laundry) and personal care skills (e.g., eating, dressing personal hygiene). These latter adaptive skills could be less verbal-dependent and more related to overall nonverbal cognition and/or other cognitive skills, which could be the reason why we do not see an association between these functional domains and our targeted expressive language constructs when controlling for NV cognition. For example, executive function skills may be more relevant than expressive language to follow cooking and laundry instructions. Memory skills are also necessary for personal hygiene (e.g., when was the last day I washed my hair, or whether I replaced the used towels for clean ones). The observed specificity in the relationships between expressive language skills and adaptive behavior skills could be related to the nature of the adaptive behaviors targeted and the extent to which these adaptive behaviors require the specific expressive language skills assessed in individuals with DS. Interestingly, the specific pattern of relationships observed for individuals with DS differs from the patters observed for individuals with ID due to other causes. For example, Shaffer et al., did not find any link between the ELS measures of syntax and lexical diversity and the Socialization or Daily Living Skills domains but did find a link between unintelligible speech and the Daily Living Skills in individuals with FXS54. It is important to note, however, that Shaffer did not examine the subdomains but instead focused on the superordinate domains, which raises the possibility that there were also differential patterns of associations with the ELS measures were obscured by the use of the superordinate domain scores. Similarly, Abbeduto and colleagues55 found that lexical diversity, syntax, and unintelligibility were each related to all three Socialization subdomains, with the only exception being that unintelligibility was not linked to Play and Leisure in individuals with FXS. Abbeduto et al., however, did not examine the Daily Living subdomains. Taken together, these results suggest that the role that specific expressive language skills play in specific areas of adaptive behavior may be syndrome specific. Limitations The current study has several limitations. First, it is important to be cautious about the generalizability of our results as we included only participants who were compliant on the administration of both conversation and narration, and for whom primary language was English. Second, although the inclusion of CA and NVFR as covariates controls for potential differences in participant characteristics, variability in samples across correlations (e.g., variable missing data in concrete VABS-2 subdomains) should be considered when interpreting our results. In addition, although the VABS-2 is available in multiple languages, additional work in terms of translation and cultural appropriateness is needed for the ELS procedures to be administered to participants for whom primary language is other than English. Importantly, this effort has begun for Spanish speakers56,57 but there is still a lot of work to do. Third, establishing a potential causal link between language and adaptive behavior will require simultaneous examination of both concurrent and longitudinal associations, as it is impossible to determine the direction of causality in the relationships observed through current cross-sectional analyses. In addition, it is important to note that we explored only a few possible dimensions of adaptive functioning as well as concrete expressive language constructs relative to structural language (e.g., we did not explore pragmatics). The relative contributions of language as well as other specific cognitive abilities to multiple dimensions of adaptive behaviors should be explored in future longitudinal research with an expanded range of measures of language (including receptive language, pragmatics, etc.), cognition, adaptive behavior, and several adaptive and functional skills relevant to independent functioning and inclusion in the community. We also recommend future studies stratifying analyses by age groups to better understand whether some of the targeted associations are age-dependent in DS. Conclusion The present study has demonstrated that expressive language skills are linked to socialization and daily living skills in individuals with DS, which raises the possibility that interventions focused on improvements in expressive language may lead to, or at least provide a foundation for, improved adaptive behavior for those with DS. Unfortunately, there are few such evidence-based language interventions for this population focused on the targeted expressive language domains58,59. Therefore, there is a pressing need to develop such interventions. This study also addresses the call for psychometrically sound measures for evaluating treatment efficacy in studies of individuals with ID. In particular, we have shown previously that ELS-derived measures are feasible, are subject to only minimal practice effects, have strong test–retest reliability, and have construct validity for individuals with DS and other developmental disabilities37,38,57. The data from the present study show that differences on ELS-derived measures are associated with real-world functional competence (i.e., socialization and daily living skills) in children, adolescents, and young adults with DS, which is an association valued by treatment-regulating bodies, such as the U.S. Food and Drug Administration (FDA), when deciding on the utility of an outcome measure for establishing treatment efficacy. Based on our results, we encourage additional study of the link between other commonly used outcome measures in NDDs (e.g., clinician or parent ratings of perceived improvement, eye-tracking, computerized tests, etc.) and meaningful functional outcomes to support selection of functionally relevant efficacy endpoints for clinical trials in these populations. Acknowledgements This research was supported by the following grant from the National Institutes of Health, R01HD074346, P50HD103526, and UL1TR001860. The authors thank the many staff across all the participating data collection sites who worked on the project. We are indebted to the families who participated for their time, support, and partnership. Author contributions L.d.H.S. conceived and designed the analytical plan for this study, analyzed the data, interpreted the results, and wrote the manuscript.; J.C.V. interpreted the results and co-wrote the manuscript.; A.T. was responsible for constructing and validating all data sets and reviewing the entire manuscript.; L.A. conceived the overall design of the larger study and protocol from which the data for this study were drawn, oversaw data collection, and contributed to the analytical plan for this study and revision of the manuscript.; A.S., J.E., E.B.K. and D.H. reviewed the manuscript. All authors have read and agreed to the published version of the manuscript. Data availability The datasets used and/or analyzed for the present paper can be made available upon a reasonable request to the corresponding author. Competing interests The authors declare no competing interests. Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. Parker SE Updated national birth prevalence estimates for selected birth defects in the United States, 2004–2006 Birth Defects Res. Part A Clin. Mol. Teratol. 2010 88 1008 1016 10.1002/bdra.20735 2. Patterson D Genetic mechanisms involved in the phenotype of Down syndrome Ment. Retard. Dev. Disabil. Res. Rev. 2007 13 199 206 10.1002/mrdd.20162 17910086 3. Sherman SL Allen EG Bean LH Freeman SB Epidemiology of Down syndrome Ment. Retard. Dev. Disabil. Res. Rev. 2007 13 221 227 10.1002/mrdd.20157 17910090 4. Brock J Jarrold C Serial order reconstruction in Down syndrome: Evidence for a selective deficit in verbal short-term memory J. Child Psychol. Psychiatry 2005 46 304 316 10.1111/j.1469-7610.2004.00352.x 15755306 5. Fidler DJ Philofsky A Hepburn SL Rogers SJ Nonverbal requesting and problem-solving by Toddlers with Down syndrome Am. J. Ment. Retard. 2005 110 312 10.1352/0895-8017(2005)110[312:NRAPBT]2.0.CO;2 15941367 6. Edgin JO Cognition in Down syndrome: A developmental cognitive neuroscience perspective Wiley Interdiscip. Rev. Cogn. Sci. 2013 4 307 317 10.1002/wcs.1221 26304208 7. Edgin JO Development and validation of the Arizona cognitive test battery for Down syndrome J. Neurodev. Disord. 2010 2 149 164 10.1007/s11689-010-9054-3 21274406 8. Fidler DJ Barrett KC Most DE Age-related differences in smiling and personality in down syndrome J. Dev. Phys. Disabil. 2005 17 263 280 10.1007/s10882-005-4384-x 9. del Hoyo Soriano L Thurman AJ Abbeduto L Specificity: A phenotypic comparison of communication-relevant domains between youth with Down syndrome and fragile X syndrome Front. Genet. 2018 9 424 10.3389/fgene.2018.00424 30327664 10. Del Hoyo Soriano L Expressive language development in adolescents with down syndrome and fragile X syndrome: Change over time and the role of family-related factors J. Neurodev. Disord. 2020 12 1 1 18 31906846 11. McDuffie A Early language intervention using distance video-teleconferencing: A pilot study of young boys with fragile X syndrome and their mothers Am. J. Speech-Lang. Pathol. 2016 25 46 10.1044/2015_AJSLP-14-0137 26502382 12. Daunhauer LA Fidler DJ The Down syndrome behavioral phenotype: Implications for practice and research in occupational therapy Occup. Ther. Heal. Care 2011 25 7 25 10.3109/07380577.2010.535601 13. Finestack LH Abbeduto L Expressive language profiles of verbally expressive adolescents and young adults with Down syndrome or fragile X syndrome J. Speech. Lang. Hear. Res. 2010 53 1334 1348 10.1044/1092-4388(2010/09-0125) 20643789 14. Channell MM McDuffie AS Bullard LM Abbeduto L Narrative language competence in children and adolescents with Down syndrome Front. Behav. Neurosci. 2015 9 283 10.3389/fnbeh.2015.00283 26578913 15. Lee M A multi-method investigation of pragmatic development in individuals with Down syndrome Am. J. Intellect. Dev. Disabil. 2017 122 289 309 10.1352/1944-7558-122.4.289 28654411 16. Chapman RS Hesketh LJ Kistler DJ Predicting longitudinal change in language production and comprehension in individuals with Down syndrome: Hierarchical linear modeling J. Speech. Lang. Hear. Res. 2002 45 902 915 10.1044/1092-4388(2002/073) 12381048 17. Martin GE Losh M Estigarribia B Sideris J Roberts J Longitudinal profiles of expressive vocabulary, syntax and pragmatic language in boys with fragile X syndrome or Down syndrome Int. J. Lang. Commun. Disord. 2013 48 432 443 10.1111/1460-6984.12019 23889838 18. Næss K-AB Lyster S-AH Hulme C Melby-Lervåg M Language and verbal short-term memory skills in children with Down syndrome: A meta-analytic review Res Dev Disabil 2011 32 6 2225 2234 10.1016/j.ridd.2011.05.014 21628091 19. Price JR Syntactic complexity during conversation of boys with fragile X syndrome and Down syndrome J. Speech Lang. Hear. Res. 2008 51 3 10.1044/1092-4388(2008/001) 18230852 20. Abbeduto L The linguistic and cognitive profile of Down syndrome: Evidence from a comparison with fragile X syndrome Down Syndr. Res. Pract. 2001 7 9 15 10.3104/reports.109 21. Eadie PA Fey ME Douglas JM Parsons CL Profiles of grammatical morphology and sentence imitation in children with specific language impairment and Down syndrome J. Speech. Lang. Hear. Res. 2002 45 720 732 10.1044/1092-4388(2002/058) 12199402 22. Laws G Bishop DVM A Comparison of Language Abilities in Adolescents With Down Syndrome and Children With Specific Language Impairment J. Speech. Lang. Hear. Res. 2003 46 1324 1339 10.1044/1092-4388(2003/103) 14700358 23. Finestack LH Palmer M Abbeduto L Macrostructural narrative language of adolescents and young adults with Down syndrome or fragile X syndrome Am. J. speech-language Pathol. 2012 21 29 46 10.1044/1058-0360(2011/10-0095) 24. del Hoyo Soriano L Thurman AJ Harvey D Kover ST Abbeduto L Expressive language development in adolescents with Down syndrome and fragile X syndrome: Change over time and the role of family-related factors J. Neurodev. Disord. 2020 12 18 10.1186/s11689-020-09320-7 32593286 25. Stoel-Gammon C Down syndrome phonology: Developmental patterns and intervention strategies Down Syndr. Res. Pract. 2001 7 93 100 10.3104/reviews.118 26. Onnivello S Cognitive profiles in children and adolescents with Down syndrome Sci. Rep. 2022 12 1 1 14 10.1038/s41598-022-05825-4 34992227 27. Karmiloff-Smith, A. et al. The importance of understanding individual differences in Down syndrome. F1000Research 5, (2016). 28. Tassé MJ Schalock RL Balboni G Spreat S Navas P Validity and reliability of the diagnostic adaptive behaviour scale J. Intellect. Disabil. Res. 2016 60 80 88 10.1111/jir.12239 26553772 29. Del Cole CG Caetano SC Ribeiro W Kümmer AMEE Jackowski AP Adolescent adaptive behavior profiles in Williams-Beuren syndrome, Down syndrome, and autism spectrum disorder Child Adolesc. Psychiatry Ment. Health 2017 11 1 9 28077965 30. Sparrow SS Cicchetti DV Saulnier CA Vineland Adaptive Behavior Scales 2016 Pearson 31. Dykens EM Hodapp RM Evans DW Profiles and development of adaptive behavior in children with Down syndrome Down Syndr. Res. Pract. 2006 9 45 50 10.3104/reprints.293 32. Schworer EK Patterns and predictors of adaptive skills in 2- to 7-year-old children with Down syndrome J. Neurodev. Disord. 2022 14 1 11 10.1186/s11689-022-09430-4 34983360 33. de Sola S A new cognitive evaluation battery for Down syndrome and its relevance for clinical trials Front. Psychol. 2015 6 708 26089807 34. Burns, M. Test review of the adaptive behavior assessment system. 2nd edn. (eds. Spies, R.A., & Plake, B.S.) Sixt. Ment. Meas. Yearb. (Springer, New York, 2005). 35. Bishop, D. V. M. The children’s communication checklist, version 2 (CCC-2). (2003). 36. Stephan C Clasen L Adeyemi E Lee NR Speech impairments explain unique variance in adaptive behavior skills in young people with Down syndrome Am. J. Speech-Lang. Pathol. 2020 30 253 259 10.1044/2020_AJSLP-20-00054 33197320 37. Abbeduto L Expressive language sampling as a source of outcome measures for treatment studies in fragile X syndrome: Feasibility, practice effects, test-retest reliability, and construct validity J. Neurodev. Disord. 2020 12 10 10.1186/s11689-020-09313-6 32204695 38. Thurman AJ Spoken language outcome measures for treatment studies in Down syndrome: Feasibility, practice effects, test-retest reliability, and construct validity of variables generated from expressive language sampling J. Neurodev. Disord. 2021 13 13 10.1186/s11689-021-09361-6 33827417 39. Westerveld MF Gillon GT Miller JF Spoken language samples of New Zealand children in conversation and narration Adv. Speech Lang. Pathol. 2004 6 4 195 208 10.1080/14417040400010140 40. Berry-Kravis E Outcome measures for clinical trials in fragile X syndrome J. Dev. Behav. Pediatr. 2013 34 508 522 10.1097/DBP.0b013e31829d1f20 24042082 41. Channell MM Loveall SJ Conners FA Harvey DJ Abbeduto L Narrative language sampling in typical development: implications for clinical trials Am. J. Speech-Lang. Pathol. 2018 27 123 10.1044/2017_AJSLP-17-0046 29222570 42. Miller, J. & Iglesias, A. Systematic analysis of language transcripts (SALT version 9) [computer software]. (2006). 43. Kover ST McDuffie A Abbeduto L Brown WT Effects of sampling context on spontaneous expressive language in males with fragile X syndrome or Down syndrome J. Speech. Lang. Hear. Res. 2012 55 1022 1038 10.1044/1092-4388(2011/11-0075) 22232386 44. Roid G Stanford Binet Intelligence Scales 2003 5 Riverside Publishing 45. Sansone SM Improving IQ measurement in intellectual disabilities using true deviation from population norms J. Neurodev. Disord. 2014 6 16 10.1186/1866-1955-6-16 26491488 46. Sparrow, S. S., Cicchetti, D. & Balla, D. A. Vineland Adaptive Behavior Scales, 2nd edn (2012). 47. Shaffer RC The relationship between expressive language sampling and clinical measures in fragile X syndrome and typical development Brain Sci. 2020 10 66 10.3390/brainsci10020066 31991905 48. Sparrow, S. S. & Cicchetti, D. V. The Vineland Adaptive Behavior Scales. In Major Psychol. Assess. instruments Allyn Bacon Allyn Bacon, vol. 2 (ed. Newmark, C. S.) 199–231 (1984). 49. Benjamini Y Yekutieli D The control of the false discovery rate in multiple testing under dependency Ann. Stat. 2001 29 1165 1188 10.1214/aos/1013699998 50. Hoyo LD Semantic verbal fluency pattern, dementia rating scores and adaptive behavior correlate with plasma Aβ42 concentrations in down syndrome young adults Front. Behav. Neurosci. 2015 9 301 10.3389/fnbeh.2015.00301 26635555 51. Foley KR Relationship between family quality of life and day occupations of young people with Down syndrome Soc. Psychiatry Psychiatr. Epidemiol. 2014 49 1455 1465 10.1007/s00127-013-0812-x 24414088 52. Joginder Singh S Gan B Chu SY Association between symbolic play and language: A comparison between typically developing children and children with Down syndrome J. Dev. Behav. Pediatr. 2021 42 146 151 10.1097/DBP.0000000000000873 33027107 53. Abbeduto L Concurrent associations between expressive language ability and independence in adolescents and adults with fragile X syndrome Brain Sci. 2021 11 1179 10.3390/brainsci11091179 34573200 54. Shaffer RC The relationship between expressive language sampling and clinical measures in fragile X syndrome and typical development Brain Sci. 2020 10 2 66 10.3390/brainsci10020066 31991905 55. Abbeduto L Concurrent associations between expressive language ability and independence in adolescents and adults with fragile X syndrome Brain Sci. 2021 11 9 1179 10.3390/brainsci11091179 34573200 56. del Hoyo Soriano L Bullard L Thurman AJ Alvarez CH Abbeduto L Providing a parent-administered outcome measure in a bilingual family of a father and a mother of two adolescents with ASD: Brief report Dev. Neurorehabilit. 2021 25 2 140 144 10.1080/17518423.2021.1942281 57. del Hoyo Soriano L Bullard L Alvarez CH Thurman AJ Abbeduto L Using telehealth-delivered procedures to collect a parent-implemented expressive language sampling narrative task in monolingual and bilingual families with Autism Spectrum Disorder: A pilot study Front. Rehabil. Sci. 2021 2 77 58. Seager E Sampson S Sin J Pagnamenta E Stojanovik V A systematic review of speech, language and communication interventions for children with Down syndrome from 0 to 6 years Int. J. Lang. Commun. Disord. 2022 57 441 463 10.1111/1460-6984.12699 35191587 59. Smith E Hokstad S Næss KAB Children with Down syndrome can benefit from language interventions; Results from a systematic review and meta-analysis J. Commun. Disord. 2020 85 105992 10.1016/j.jcomdis.2020.105992 32445828
PMC009xxxxxx/PMC9684568.txt
==== Front Nat Commun Nat Commun Nature Communications 2041-1723 Nature Publishing Group UK London 36418855 34798 10.1038/s41467-022-34798-1 Article Engineered disorder in CO2 photocatalysis Li Zhao 123 Mao Chengliang 2 Pei Qijun 4 http://orcid.org/0000-0002-9046-1198 Duchesne Paul N. 5 http://orcid.org/0000-0003-2900-7612 He Teng 4 Xia Meikun 2 Wang Jintao 4 Wang Lu 6 Song Rui 123 Ali Feysal M. 2 http://orcid.org/0000-0002-7529-2736 Meira Débora Motta 78 http://orcid.org/0000-0003-3345-014X Ge Qingjie 4 http://orcid.org/0000-0002-3668-5043 Ghuman Kulbir Kaur [email protected] 9 http://orcid.org/0000-0002-4520-0482 He Le 13 http://orcid.org/0000-0002-6732-2499 Zhang Xiaohong [email protected] 13 http://orcid.org/0000-0002-6315-0925 Ozin Geoffrey A. [email protected] 2 1 grid.263761.7 0000 0001 0198 0694 Institute of Functional Nano & Soft Materials (FUNSOM), Soochow University, 199 Ren’ai Road, Suzhou, 215123 Jiangsu PR China 2 grid.17063.33 0000 0001 2157 2938 Solar Fuels Group, Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, ON M5S 3H6 Canada 3 grid.263761.7 0000 0001 0198 0694 Jiangsu Key Laboratory of Advanced Negative Carbon Technologies, Soochow University, Suzhou, 215123 Jiangsu PR China 4 grid.9227.e 0000000119573309 Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian, 116023 Liaoning PR China 5 grid.410356.5 0000 0004 1936 8331 Department of Chemistry, Queen’s University, 90 Bader Lane, Kingston, ON K7L 3N6 Canada 6 grid.511521.3 The Chinese University of Hong Kong, Shenzhen, 518172 Shenzhen, Guangdong People’s Republic of China 7 grid.187073.a 0000 0001 1939 4845 CLS@APS, Advanced Photon Source, Argonne National Laboratory, Lemont, IL 60439 USA 8 grid.423571.6 0000 0004 0443 7584 Canadian Light Source Inc., 44 Innovation Boulevard, Saskatoon, SK S7N 2V3 Canada 9 grid.418084.1 0000 0000 9582 2314 Institut National de la Recherche Scientifique, Centre Énergie, Matériaux et Télécommunications, 1650 Boul. Lionel Boulet, Varennes, QC J3X 1S2 Canada 23 11 2022 23 11 2022 2022 13 720530 9 2021 7 11 2022 © The Author(s) 2022 https://creativecommons.org/licenses/by/4.0/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Light harvesting, separation of charge carriers, and surface reactions are three fundamental steps that are essential for an efficient photocatalyst. Here we show that these steps in the TiO2 can be boosted simultaneously by disorder engineering. A solid-state reduction reaction between sodium and TiO2 forms a core-shell c-TiO2@a-TiO2-x(OH)y heterostructure, comprised of HO-Ti-[O]-Ti surface frustrated Lewis pairs (SFLPs) embedded in an amorphous shell surrounding a crystalline core, which enables a new genre of chemical reactivity. Specifically, these SFLPs heterolytically dissociate dihydrogen at room temperature to form charge-balancing protonated hydroxyl groups and hydrides at unsaturated titanium surface sites, which display high reactivity towards CO2 reduction. This crystalline-amorphous heterostructure also boosts light absorption, charge carrier separation and transfer to SFLPs, while prolonged carrier lifetimes and photothermal heat generation further enhance reactivity. The collective results of this study motivate a general approach for catalytically generating sustainable chemicals and fuels through engineered disorder in heterogeneous CO2 photocatalysts. Developing an efficient photocatalyst for CO2 reduction is appealing. Here, the authors report a core-shell c-TiO2@a-TiO2-x(OH)y heterostructure with surface frustrated Lewis pairs on amorphous shells which enables a new genre of chemical reactivity. Subject terms Photocatalysis Energy Heterogeneous catalysis https://doi.org/10.13039/501100002790 Canadian Network for Research and Innovation in Machining Technology, Natural Sciences and Engineering Research Council of Canada (NSERC Canadian Network for Research and Innovation in Machining Technology) https://doi.org/10.13039/100013216 Imperial Oil Limited (L'Impériale) GAO acknowledges the financial support of the Ontario Ministry of Research and Innovation (MRI), the Ministry of Economic Development, Employment, and Infrastructure (MEDI), the Ministry of the Environment and Climate Change’s (MOECC) Best in Science (BIS) Award, Ontario Centre of Excellence Solutions 2030 Challenge Fund, Ministry of Research Innovation and Science (MRIS) Low Carbon Innovation Fund (LCIF), the University of Toronto’s Connaught Innovation Fund (CIF), Connaught Global Challenge (CGC) Fund.issue-copyright-statement© Springer Nature Limited 2022 ==== Body pmcIntroduction Amorphous solids may behave as metals1,2, semiconductors3,4, insulators5,6, or superconductors7,8 and can exhibit electrical, optical, thermal, and mechanical properties distinct from those of their crystalline analogues. Disorder, for example, can introduce both deep and shallow defect states and extended tail states into the band structure of solids, leading to profound changes in the optical and electrical properties of solid-state devices, including those of solar cells9,10. The broadly tunable band structure of amorphous metal oxides, together with other merits such as transparency and good stability, has even led them to supersede their crystalline counterparts as promising electron injection layer materials for organic light-emitting diodes11,12. In the context of electron mobility, extensive disorder can also impede electron and phonon transport, lowering electrical and thermal conductivity relative to the corresponding ordered phase, thereby making them promising materials for use in thermoelectric power generation and solid-state refrigeration devices13–15. These examples demonstrate how engineering the physical properties of solids using disorder enables advances in materials science and technology16,17. Little is known, however, about the influence of disorder on surface chemistry and catalysis18,19, which is the focus of the research described herein. As a key example, disordered (black) titania is renowned for its enhanced photocatalytic properties relative to its crystalline (white) counterpart20–23. Despite great efforts to understand the origin of its reactivity (focusing, for example, on its electronic bandgap24,25, photon absorption26,27, and charge-carrier separation kinetics28,29, pivotal aspects of its surface chemistry remain essentially unknown to this day, due largely to the challenges of defining the surface structure and understanding the physicochemical properties of the amorphous state. In this article, by means of in situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS), solid-state magic angle spinning proton nuclear magnetic resonance (1H-MAS-NMR), electron paramagnetic resonance (EPR), and density functional theory (DFT) modeling of the amorphous state, we present a molecular-level description of surface frustrated Lewis pairs (SFLPs), active for the photocatalytic reduction of CO2 to CO, in a titania heterostructure material comprised of a crystalline core surrounded by an amorphous shell denoted c-TiO2@a-TiO2-x(OH)y. At the crux of this new kind of SFLP30, embedded in a disordered a-TiO2-x(OH)y surface, is the quasi-dative bonding that exists between the Lewis base surface hydroxyl group (OH) and the Lewis acid mixed valence Ti(III,IV) induced by the presence of an oxygen vacancy ([O]) denoted HO-Ti(III)-[O]-Ti(IV), (Fig. 1). Such SFLPs engender high reactivity towards heterolytic H2 dissociation to produce a protonated/hydridic SFLP represented as HOH-Ti-[O]-Ti-H. The capture of CO2, via interaction of the electrophilic carbon of CO2 with the hydride and nucleophilic oxygen of the hydroxide of the SFLP HOH-Ti-[O]-Ti-H site, creates a formic acid (HCOOH) intermediate, which fragments to CO and H2O.Fig. 1 SFLPs, confined to the surface of the amorphous shell of the c-TiO2@a-TiO2-x(OH)y heterostructure, facilitate record photocatalysis CO2 performance. a Depiction of the central role played by the physicochemical properties in the catalytic hydrogenation of CO2-to-CO by SFLPs. b–d Reactive chemistry of an SFLP in c-TiO2@a-TiO2-x(OH)y, highlighting the heterolysis of H2 to form a protonated hydroxyl and a hydridic titanium SFLP. Structures are shown of (b) an unmodified crystalline surface, (c) formation of an SFLP site, (d) a ground-state SFLP site following activation by hydrogen, and (e) the enhanced activity excited-state SFLP generated by photoexcitation of the ground-state SFLP. The SFLPs, confined to the surface of the amorphous shell covering the c-TiO2@a-TiO2-x(OH)y heterostructure, serve multiple functions: they improve the harvesting of visible and near-infrared photons, facilitate the generation and separation of photoexcited charge carriers, prolong charge-carrier lifetimes, and promote local heat generation. Collectively, these effects boost the chemical reactivity of the SFLPs, resulting in a record CO2-to-CO conversion rate of 5.3 mmol gcat−1 h−1 (350 times superior to that of crystalline c-TiO2) and a turnover frequency of 592 h−1, under solar light irradiation of 4.0 W cm−2. Thus, the results of this study motivate a general approach to sustainable fuels through engineered disorder in metal oxide-based CO2 photocatalysts. Results and discussion Na/c-TiO2-H2O synthesis of oxygen-vacancy- and hydroxyl-laden c-TiO2@a-TiO2-x(OH)y The core@shell c-TiO2@a-TiO2-x(OH)y was prepared by ball-milling c-TiO2 (crystalline, 2.2 g) with Na/NaCl mixture (8.8 g) at solid state in Ar atmosphere under ambient pressure. A matrix of excess NaCl was added to facilitate the thorough mixing of the two reactants. Through the mechanochemical-driven redox reaction, mixed oxides with low-valent titanium (Ti(III)) atoms and oxygen vacancies ([O]) were obtained. During this process, we managed to confine the mechanochemistry to a nanometer-thick c-TiO2 surface layer by controlling the ball-milling time. To understand the charge and stoichiometric balance of the sample preparation, electron paramagnetic resonance (EPR) analysis was employed (Fig. 2a–c). The room-temperature first-derivative spectrum of the reacted mixture showed two major peaks at g = 2.003 and g = 1.977, which are typical for an unpaired free electron (such as in [O] or Na(0)) and Ti(III), respectively31–33 (Fig. 2a)). Spectra acquired at 110 K provided better resolution and intensity for these signals, allowing anatase and rutile Ti(III) to be distinguished by g values of (1.990, 1.990, 1.962) and (1.975, 1.975, 1.951), respectively (Fig. 2b). Although the [O] signal in TiO2 typically demonstrates a symmetric Lorentzian line shape, while the Na(0) shows a Dysonian line34, it is difficult to distinguish the two in the spectra of the reacted mixture due to the complexity of overlapping peaks (Supplementary Fig. 1). Removing the residuals in the synthesis mixture by thorough washing (by deionized water) helped to resolve this dilemma, resulting in the desired product c-TiO2@a-TiO2-x(OH)y, the EPR spectra of which showed prominently increased Ti(III) and [O] signals (Fig. 2c). These results evidence that the low-valent Ti(III) and oxygen vacancies were generated mainly during sample washing. Based on this information, the following reaction sequence with balanced equations is proposed to explain the synthesis:1 TiO2+(2x−y)Na→Na(2x−y)TiO2 2 Na(2x−y)TiO2/H2O→{(x−0.5y)Na2O+TiO2−x+0.5y}/H2O 3 Na2O+H2O→2NaOH 4 TiO2−x+0.5y+0.5yH2O→TiO2−x(OH)y Fig. 2 The preparation of c-TiO2@a-TiO2-x(OH)y from c-TiO2. a–c EPR spectra of the pristine c-TiO2, the mixture of Na + c-TiO2 + NaCl after ball-milling and the product c-TiO2@a-TiO2-x(OH)y. d Schematic of SFLP formation during c-TiO2@a-TiO2-x(OH)y synthesis. e PXRD patterns of starting c-TiO2, c-TiO2@a-TiO2-x(OH)y product, and the intermediate phase. Upon the contact of Na and c-TiO2, Na(0) clusters inserted into the TiO2 lattice inducing new coordination of Ti(IV)-ONa-Ti(IV). Subsequently, an electron was transferred from Na(0) to Ti(IV) forming a Ti(III) ion, the Ti(III)-ONa-Ti(IV), followed by reacting with water to generate a Ti(III)-[O]-Ti(IV)-OH. The last step is also believed to be the origin of surface disorder in c-TiO2@a-TiO2-x(OH)y (Fig. 2d). Powder X-ray diffraction (PXRD) confirmed the reaction model, showing evidence of a crystalline intermediate Na2TiO3 being formed by ball-milling c-TiO2 with metallic Na(0), with a decrease in crystallinity for the TiO2, Fig. 2e. Then, this intermediate was transformed by Na de-intercalation during the washing process and did not remain in the resulting c-TiO2@a-TiO2-x(OH)y according to the PXRD and extended X-Ray absorption fine structure (EXAFS) results (Supplementary Fig. 2). Ti(III)-[O]-Ti(IV)-OH-related atomic/electronic structure The thickness of the surface a-TiO2-x(OH)y in the c-TiO2@a-TiO2-x(OH)y was approximated to be 2 to 6 nm, as indicated by the formation of an amorphous shell/crystalline core heterostructure in the high-resolution transmission electron microscopy (HRTEM) image (Fig. 3a and Supplementary Figs. 3 and 4). HRTEM video and Raman analysis proved that both the c-TiO2 and the c-TiO2@a-TiO2-x(OH)y surfaces are stable under measurement conditions, and thus excluded the possibility of the electron beam induced surface amorphization for the c-TiO2@a-TiO2-x(OH)y (Supplementary Movie 1 and Supplementary Fig. 5). Ti L-edge electron energy loss spectroscopy (EELS) scans taken at several points on a c-TiO2@a-TiO2-x(OH)y nanoparticle supported this model. Conventional Ti(IV) L2,L3-edge features of tg-e2g splitting emerged in the core region, which were similar to those of c-TiO2, but no such splitting features were observed on the nanoparticle surface, and the L-edge energies shifted to low values by ~2.2 eV (Fig. 3b, c and Supplementary Fig. 6). Given the fact that the tg-e2g splitting stems from the energy difference between Ti 3d-orbitals with different geometric orientations, it is sensitive to the valence of Ti and the presence of point defects in titania35. Therefore, the above result suggests a low-valent Ti(III)- and [O]-rich amorphous a-TiO2-x(OH)y shell with Ti-O6 octahedra distorted and reconstructed to lower symmetry (e.g., distorted Ti-O5 pyramids).Fig. 3 Characterization of crystalline-amorphous heterostructured c-TiO2@a-TiO2-x(OH)y. a HR-TEM micrograph of c-TiO2@a-TiO2-x(OH)y (anatase phase). The amorphous/crystalline interfaces are marked with dotted lines. b, c High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) image of c-TiO2@a-TiO2-x(OH)y (b) and corresponding Ti-L edge EELS spectra at different positions (c). d Comparison of geometrical and electronic structures between bridging and terminal hydroxyl. e, f 1H MAS-NMR (e) and ATR-FTIR (f) spectra of c-TiO2@a-TiO2-x(OH)y and c-TiO2. Furthermore, the X-ray photoelectron spectroscopy (XPS) spectra of c-TiO2@a-TiO2-x(OH)y showed consistent results. Compared with the c-TiO2, the Ti 2p spectra of c-TiO2@a-TiO2-x(OH)y exhibited a slight peak shift from 459.5 eV to 459.3 eV with an emerging shoulder peak at 456.9 eV, signifying the presence of Ti(III) (Supplementary Fig. 7a). The temperature-dependent EPR spectra of c-TiO2@a-TiO2-x(OH)y showed that the intensity ratio of Ti(III) to [O] decreased from 15.28 at 207 K to 1.716 at room temperature (Fig. 2c), indicating their vicinity enabled an electron shuttle. These results supported our model of mixed-valence [O]-laden Ti(III)-[O]-Ti(IV)-OH with the presence of surface hydroxyl groups, confirmed by a peak at 533.4 eV in the O1s XPS spectra of c-TiO2@a-TiO2-x(OH)y (Supplementary Fig. 7b). Theoretically, to obtain the above Ti(III)-[O]-Ti(IV)-OH via the TiO2-x+0.5y + 0.5yH2O → TiO2-x(OH)y reaction, the H2O should dissociate on the surface O(-II) and Ti(IV) sites to make charge-balancing Lewis base hydroxide OH(-I) on the Ti(III,IV) sites and protons on the O(-II) as Bronsted acid OH, namely the SFLP. Alternatively, the H2O can bind at the oxygen vacancy through its oxygen and dissociate to hydroxide and a proton where the latter binds to a lattice oxygen—this reaction pathway generates only Bronsted acid sites on the oxide lattice contrasting with the SFLP model. To confirm the function of the c-TiO2@a-TiO2-x(OH)y surface as a SFLP [Ti(III)-[O]-Ti(IV)-OH] instead of a conventional Bronsted acid [Ti(IV)-OH-Ti(IV)], we now focus on distinguishing the two models. Theoretically, the terminal hydroxyl Ti(IV)-OH in the SFLP possesses a higher degree of freedom compared to that of the bridging hydroxyl (Fig. 3d). Thus, their charge density and molecular vibrations should be distinct, which can be monitored through solid-state 1H magic-angle spinning nuclear magnetic resonance (1H MAS NMR) and attenuated total reflection Fourier transform infrared (ATR-FTIR) spectroscopy, respectively. The 1H MAS NMR spectrum of c-TiO2 showed peaks at 5.8, 1.6, and −1.65 ppm. The peak around 5.8 ppm is typically ascribed to absorbed water, and the peak at 2.2 ppm on neat titanium dioxide is assigned to hydroxyl group36. Both water and hydroxyl group are naturally existing species when a metal oxide sample is exposed to air. The peak at −1.65 ppm appeared to be a background signal as similar signal emerged on the 1H MAS NMR spectra of c-TiO2@a-TiO2-x(OH)y as well. Interestingly, the water and hydroxyl peaks of c-TiO2@a-TiO2-x(OH)y shifted to 5.2 ppm and 1.7 ppm, respectively, and two additional peaks (2.2 and 0.6 ppm) emerged around the hydroxyl peak. Since the chemical shift of 1H-NMR reflects the shielding effect, compared to c-TiO2, the lower chemical shift of water on c-TiO2@a-TiO2-x(OH)y indicated the possible electron donation from oxygen vacancies and/or Ti(III) in a-TiO2-x(OH)y, while the higher chemical shift of hydroxyl (1.7 and 2.2 ppm)37 suggested the proton of less electron density, which agreed well the Lewis base -OH in the SFLP model. The peak at 0.6 ppm can be assigned to several species38, including terminal hydroxyl group and proton bonded to oxygen vacancy, which are closely related to the amorphous a-TiO2-x(OH)y shell (Fig. 3e). Furthermore, ATR-FTIR spectra of c-TiO2@a-TiO2-x(OH)y demonstrated a 1340 cm−1 peak, which is likely the bending mode of terminal OH39 induced by the symmetry-breaking of a bridging OH (Fig. 2d and Supplementary Fig. 8). Additionally, new peaks in the 3610~3680 cm−1 region, belonging to the stretching mode of terminal OH39,40, appeared on the ATR-FTIR spectra of c-TiO2@a-TiO2-x(OH)y, while that of c-TiO2 only showed the bridging OH feature located in the 3690~3750 cm−1 region (Supplementary Fig. 8). Despite the OH stretching peaks being low in intensity compared to the OH bending peaks, it was quite clear that the terminal OH species increased with the introduction of disorder TiO2-x(OH)y on the c-TiO2 surface. Both theoretic and experimental results confirm the SFLP feature of c-TiO2@a-TiO2-x(OH)y, where mixed-valence Ti(III, IV), [O], and terminal OH are in close proximity, forming the Ti(III)-[O]-Ti(IV)-OH. SFLPs-related electronic properties, light absorption, and charge carrier separation The reduction extent of titania during sample preparation determined the degree of non-stoichiometry (x) and surface disorder as well as the population of Ti(III), oxygen vacancies, and OH sites, which further corresponds to their optical spectra and band gaps41. The optical reflectance spectrum (UV-Vis-NIR) of c-TiO2 demonstrated UV light absorption with a band edge at ~380 nm, while that for c-TiO2@a-TiO2-x(OH)y (0.0015 < x < 0.0031) demonstrated strong broadband light absorption (Fig. 4a). Increasing the x value from 0.0002 < x < 0.0004 to 0.0015 < x < 0.0031 in c-TiO2@a-TiO2-x(OH)y samples (1→4) showed the evolution of Vis-NIR absorption, and red shifting of the absorption edges (Fig. 4a), corresponding to bandgaps of 3.49, 3.46, 3.43 and 3.38 eV (Supplementary Fig. 9), respectively; all of these bandgap energies were smaller than the value of 3.57 eV for pristine c-TiO2, indicating the increased sunlight harvest of c-TiO2@a-TiO2-x(OH)y compared with c-TiO2. The detailed band-edge positions of c-TiO2@a-TiO2-x(OH)y was provided by ultraviolet photoelectron spectroscopy (UPS). Valence band (EVB) positions of c-TiO2 and a-TiO2-x(OH)y (0.0015 < x < 0.0031) were found at −7.7 and −7.59 eV (vs. vacuum), and the corresponding conduction band (ECB) positions were found at −4.13 and −4.21 eV (vs. vacuum), respectively (Supplementary Fig. 10).Fig. 4 Opto-electronic properties of c-TiO2@a-TiO2-x(OH)y. a UV–Vis-NIR diffuse reflectance spectra (DRS) of c-TiO2@a-TiO2-x(OH)y as a function of reduction degree. The x was quantified in each sample. b, c The total density of states (black line, right axis) and the corresponding values of the inverse participation ratio (IPR) (blue dots, left axis) for a-TiO2-x(OH)y (b) and crystalline TiO2-x (c) surfaces. d Schematic of charge carrier separation and transfer pathways in c-TiO2@a-TiO2-x(OH)y. The colors of arrows indicate the wavelength of incident light. e Time-resolved photoluminescence spectroscopy decay curves of c-TiO2 and c-TiO2@a-TiO2-x(OH)y. Density of states (DOS) and the delocalization degree of electrons were then simulated computationally in both the amorphous and crystalline components to provide mechanistic insights. Specifically, inverse participation ratio (IPR) analysis was conducted to investigate charge mobility and the localization of energy states on amorphous TiO2-x(OH)y and crystalline TiO2-x surfaces. In Fig. 4b, c, the IPR and DOS of amorphous TiO2-x(OH)y and crystalline TiO2-x are depicted, where a large IPR value represents highly localized states and a small IPR value represents delocalized states. The IPR results showed that the Ti 3d in the bottom of the CB (conduction band) and O 2p states at the top of the VB (valence band) for amorphous TiO2-x(OH)y are strongly localized as compared to crystalline TiO2-x. Specifically, the OH and [O] vacancy resulted in shallow band states near the VB edge and the [O] vacancy resulted in shallow band states near the CB edge, respectively (Supplementary Fig. 11). These results indicate the high-concentration self-trapped polarons and excitons of a-TiO2-x(OH)y compared to crystalline TiO2-x, which facilitate charge carrier separation by trapping photo-generated holes and electrons by SLFPs OH and Ti(III), respectively. Based on CB and VB positions obtained from the UPS experiment, DOS and charge localization analyses via DFT, a schematic illustration for electron-hole transfer, separation, and recombination, is proposed in Fig. 4d. The UV light can excite both the core c-TiO2 and the shell a-TiO2-x(OH)y because the photon energy of the UV light (λ < 347 nm or E > 3.57 eV) is larger than the bandgap energy of c-TiO2 (Eg = 3.57 eV) or a-TiO2-x(OH)y (Eg = 3.38 eV). The calculated penetration depth (Dp) of UV light for TiO2 further supported the above results (Supplementary Figs. 12 and 13). c-TiO2 and a-TiO2-x(OH)y demonstrate similar UV-absorbing capacities, thus showing a similar penetration depth in UV region. Given the Dp is larger than the 3─6 nm thickness of the shell, the UV light will be partially absorbed by surface a-TiO2-x(OH)y and then reach the core c-TiO2 in the c-TiO2@a-TiO2-x(OH)y. Therefore, both the core c-TiO2 and shell a-TiO2-x(OH)y will be excited by the UV light to yield photo electrons and holes. Simultaneously, the visible light penetration depth of black titania is within several micrometers, so the visible light cannot be fully absorbed by a 3–6 nm surface a-TiO2-x(OH)y layer but will be absorbed by the sub-surface catalyst particles with a stacking thickness of several micrometers. Accordingly, the visible and IR light can excite electrons on the defect- and OH-related energy levels of the a-TiO2-x(OH)y. Because the CB of a-TiO2-x(OH)y is lower than that of c-TiO2, the photogenerated electrons on the c-TiO2 preferentially transfer across the crystalline-amorphous interface to reach the a-TiO2-x(OH)y and get trapped by an unsaturated Ti(III) and/or oxygen-vacancy sites. Therefore, the crystalline core-amorphous shell structure facilitates the charge carrier spatial separation, which enhances the excited-state Lewis basicity and acidity of the Ti(III)-[O]-Ti(IV)-OH, namely the ES SFLP, by the trapped electrons and holes. This ES SFLP not only enhances chemical reactivity compared to its GS counterpart but also possibly prolongs the lifetime of the photogenerated electrons and holes during the H2 and CO2 reaction. Inspired by above calculation, time-resolved photoluminescence (PL) spectroscopy was performed at room temperature to testify the diffusion and recombination processes of photogenerated electron-hole pairs in c-TiO2@a-TiO2-x(OH)y. All PL decay curves exhibited extremely fast relaxation (Fig. 4e) and were fitted by a bi-exponential model. The pristine TiO2 gave an average exciton-decay time of 10.4 ns, while a longer exciton-decay time of 23.6 ns was observed for the c-TiO2@a-TiO2-x(OH)y (Supplementary Table 1). This supports the proposed model (depicted in Fig. 4d) of charge transfer between the c-TiO2 core and the a-TiO2-x(OH)y shell, prolonging the lifetime of charge carriers for subsequent photoreaction. Activation of CO2 and H2 by SFLPs on c-TiO2@a-TiO2-x(OH)y Previous studies evidenced a key feature of the frustrated Lewis pair as triggering H2 activation42 under ambient temperatures. The obtained products are a proton and a hydride, of which the latter shows high chemical reactivity for CO2 reduction. Therefore, to check the chemical function of the SFLP in the c-TiO2@a-TiO2-x(OH)y, its reaction with H2 was monitored through the NMR. Solid state 1H MAS NMR of c-TiO2@a-TiO2-x(OH)y showed a major peak at ca. 5.2 ppm which was typically ascribed to Ti-OH2 species43, and peaks between −1.6─2.2 ppm with negligible intensities compared to that at 5.2 ppm. After exposure to H2 at room temperature, a protonated Ti-OH2 peak at 3.8 ppm, and two intense hydridic Ti-H peaks44,45 at 0.14 ppm & −0.2 ppm emerged, signifying the heterolytic dissociation of H2 on SFLPs (Fig. 5a). In contrast, there was no significant change in the NMR spectra of c-TiO2 before and after introducing H2 (Fig. 5b).Fig. 5 Activation of CO2 by SFLPs on c-TiO2@a-TiO2-x(OH)y. a, b Ex situ solid-state 1H MAS-NMR spectra of c-TiO2@a-TiO2-x(OH)y (a) and c-TiO2 (b) before (purple) and after (orange) exposure to H2 at atmospheric pressure and room temperature. c–e H2 adsorption resulting in the formation of two surface OH (c) or a Ti-H and an OH on amorphous titania surfaces (d), or absorbed H atoms on crystalline TiO2-x (e). The Ti, O and H atoms are represented in gray, red, and blue, respectively. f, g In situ DRIFTS spectra of CO2 reduction over c-TiO2@ a-TiO2-x(OH)y (f) and c-TiO2 (g) at various temperatures while flowing 2 sccm H2, 2 sccm CO2, and 16 sccm He. Theoretical calculations on the amorphous structure were then conducted to provide mechanistic understanding of the origin of the facile H2 heterolysis over the SFLP site of c-TiO2@a-TiO2-x(OH)y. The interaction of H2 with amorphous TiO2 (a-TiO2) was screened on three Ti sites (Ti-4c, Ti-5c, and Ti-6c) and ten O sites (a to i) via geometry optimization (Supplementary Fig. 14). In all cases (except sites ‘i’ and ‘Ti-4c’), H2 adsorption on surface was exothermic but only resulted in the formation of surface OH, as shown in Fig. 5c. Surprisingly, the introduction of O vacancies on amorphous surfaces resulted in facile exothermic H2 heterolysis, with δEads = −0.06 Ry and the generation of a surface OH and a Ti-H, Fig. 5d. Bader charge analysis showed that the hydrogen in the OH bore a charge of +0.99e while −0.6e in the hydride (Supplementary Table 2), bonding to weakly non-coordinative Lewis basic O (−1.7e) and Lewis acidic Ti (+2.1e), respectively. In contrast, the H2 could not dissociate to form OH when it interacted with the crystalline c-TiO2-x surface having vacancies at O-2c (Fig. 5e) and O-3c sites (Supplementary Fig. 15). Subsequently, in situ DRIFTS was applied to monitor the CO2 reduction over the SFLP-laden c-TiO2@a-TiO2-x(OH)y at molecular level. In the presence of both H2 and CO2, three major C-related peaks emerged at 1307 cm−1, 1370 cm−1, and 1677 cm−1 (Fig. 5f), which could be assigned as carboxylate species (1307 cm−1)46,47 and formate (1370 cm−1 and 1677 cm−1)48,49. Such a formate species agreed well with the reaction model of CO2 addition to the protonated and hydridic SFLPs by its nucleophilic O(-II) and electrophilic C(IV), respectively (Fig. 1). Consistently, the peak intensities of formate species decreased at high temperature (200 → 300 °C) as they were converted to CO, with the increased intensity at 1643 cm−1 corresponding to the H2O product induced scissor modes. By contrast, carbonates (1345 cm−1, 1625 cm−1, and 1574 cm−1) and bicarbonates (1248 cm−1, 1407 cm−1, and 1672 cm−1)48,50,51 were generated on the surface of c-TiO2 (Fig. 5g). The distinct surface chemistry between c-TiO2 and amorphous TiO2-x(OH)y in RWGS reaction was also confirmed by the apparent activation energy (Ea). Arrhenius plots of temperature-dependent CO rates suggest an Ea of 75.40 kJ mol−1 for c-TiO2@a-TiO2-x(OH)y, which decreased by ~17% compared with 90.27 kJ mol−1 of c-TiO2 (Supplementary Fig. 16). Photocatalytic performance With the favorable band structure and surface reactivity, the SFLPs-laden c-TiO2@a-TiO2-x(OH)y (0.0015 < x < 0.0031) demonstrated CO2 photocatalytic activity of 5.3 mmol gcat−1 h−1 under solar light irradiation (4.0 W cm−2), which was 350 times that of the pristine c-TiO2 (Fig. 6a). The turnover frequency (TOF) of c-TiO2@a-TiO2-x(OH)y was as high as 592 h−1, superb among known catalysts tested under comparable photocatalytic conditions52 (Supplementary Table 3). The apparent quantum yield (AQY) of the full Xe lamp spectrum can achieve 0.09% (Supplementary Note). This is likely a lower limit as neither the photocatalyst nor photoreactor architectures have been engineered and optimized for high photon capture efficiency, the key to high energy efficiency of the integrated system.Fig. 6 Photocatalytic performance of oxides in the batch reactor without external heating. Initial reactants were 15 psi of CO2 and 15 psi of H2 (total pressure of ca. 2 bar). a Non-stoichiometry dependent CO production rates of the c-TiO2@a-TiO2-x(OH)y under full-spectrum Xe light, 4.0 W cm−2. b Stability of c-TiO2@a-TiO2-x(OH)y under full-spectrum Xe light (4.0 W cm−2) in 8 h. c CO production rates of c-TiO2@a-TiO2-x(OH)y, pristine c-TiO2 and 2% Na/c-TiO2 under visible light (2.8 W cm−2). Notably, the c-TiO2@a-TiO2-x(OH)y (0.0015 < x < 0.0031) demonstrated no apparent activity decrease in eight consecutive runs in the batch reactor (Fig. 6b) or a 48-h on-stream reaction under 300 °C in a flow reactor (Supplementary Fig. 17). The structural stability of spent c-TiO2@a-TiO2-x(OH)y was supported by PXRD and XPS, which showed no significant changes compared to that before reaction (Supplementary Figs. 18 and 19). Furthermore, using an isotopically labeled 13CO2-H2 feedstock gave 13CO as the only product for c-TiO2@a-TiO2-x(OH)y, confirming the carbon source from the CO2 feedstock (Supplementary Fig. 20). ICP-OES analysis of c-TiO2@a-TiO2-x(OH)y suggested a 2 wt% residual Na, while our control sample prepared by loading 2 wt% Na (NaOH) on c-TiO2 demonstrated 0.19 mmol gcat−1 h−1 activity, a negligible increase compared with the pristine sample, excluding the possibility that Na-dominated activity enhancement of the c-TiO2@a-TiO2-x(OH)y sample (Supplementary Fig. 21). These results support the SFLP-laden c-TiO2@a-TiO2-x(OH)y (0.0015 < x < 0.0031) as a promising CO2 photocatalyst. Interestingly, the aforementioned synergy effect between the c-TiO2 core and the a-TiO2-x(OH)y shell, as shown in Fig. 4d, could also be reflected in the activity by varying the non-stoichiometry of the c-TiO2@a-TiO2-x(OH)y catalyst. To amplify, the CO production rate of c-TiO2@a-TiO2-x(OH)y samples first increased from 0.17 mmol gcat−1 h−1 (0.0002 < x < 0.0004) to 5.3 mmol gcat−1 h−1 (0.0015 < x < 0.0031), then decreased to 4.7 mmol gcat−1 h−1 by further increasing the non-stoichiometry to 0.0021 < x < 0.0039, demonstrating a volcano activity-non-stoichiometry relationship (Fig. 6a, Supplementary Fig. 22). HRTEM images evidenced the optimal sample (0.0015 < x < 0.0031) with an amorphous shell of 3–6 nm in thickness, while the samples of 0.0002 < x < 0.0004 and 0.0021 < x < 0.0039 showed negligible and ~10 nm shell thickness, respectively (Supplementary Fig. 23). Given the efficiency of a photocatalytic reaction is determined by the efficiencies of light absorption, charge separation and finally the surface reaction, above results suggested the synergy effect resulting from a tradeoff between 1) the high charge-carrier separation efficiency and high photo-available surface area related to the core (c-TiO2) and 2) the high surface reactivity induced by the SFLP shell (a-TiO2-x(OH)y) (Supplementary Fig. 13). In addition, wavelength-dependent activity tests further shed light on the contribution of the core-shell synergy effect in the c-TiO2@a-TiO2-x(OH)y (0.0015 < x < 0.0031). After deactivating the core c-TiO2 by filtering the UV part (λ < 420 nm) of the Xe lamp, the c-TiO2@a-TiO2-x(OH)y sample exhibited visible light-driven activity of 3.0 mmol CO gcat−1 h−1 (Fig. 6c). By comparing the UV-corresponded activity [(5.3–3.0) mmol gcat−1 h−1] with that of the pure c-TiO2 (0 mmol gcat−1 h−1), the contribution of the synergy effect is 43% of the full-spectrum activity (2.3 vs. 5.3 mmol gcat−1 h−1), which is the upper limit. While the lower limit is estimated by comparing the UV activity (2.3 mmol gcat−1 h−1) with that of the TiO2@a-TiO2-x(OH)y sample bearing the highest x value (0.0021 < x < 0.0039; 1.9 mmol gcat−1 h−1), which is 8% [(2.3–1.9) vs. 5.3 mmol gcat−1 h−1]. In the Vis-IR region devoid of the synergy effect, wavelength-dependent CO production rate agreed well with the light-absorbing capability as reflected by DRS spectra (Supplementary Fig. 24). The photothermal contribution to the activity was identified by changing the CO2/H2 gas ratio in the reactor from 1:1 to 5:1 without varying the light intensity (Supplementary Fig. 25). The CO production rate doubled to 11.2 mmol gcat−1 h−1 due to the better insulating effect of CO2 relative to H2. ASPEN Plus was also employed to calculate the local temperature of c-TiO2@a-TiO2-x(OH)y using a CO2/H2 ratio of 5:1. The predicted temperature reached as high as 200 °C (Supplementary Table 4). Intentionally engineered SFLPs in the amorphous shell of c-TiO2@a-TiO2-x(OH)y boost the harvesting potential of solar photons and the effectiveness of the photothermal effect, while the heterostructure benefits the generation, separation, and lifetime of electron-hole pairs on the photo reactivity of HOTi-[O]-Ti SFLPs towards H2 heterolysis and CO2 reduction. The paradigm of integrating disorder engineering with surface frustrated Lewis pairs in a core-shell crystalline-amorphous c-TiO2@a-TiO2-x(OH)y heterostructure provides a new paradigm for designing and implementing photocatalysts for synthesizing sustainable fuels from carbon dioxide feedstock. Methods Material synthesis All chemicals used were of analytical grade and used without any further purification. Commercial P25(c-TiO2) (99%) was purchased from Alfa Aesar. The Na/NaCl mixture, composed of small Na particles dispersed within NaCl, is expected to be an effective agent for the reduction of metal oxides. The weight ratio between Na and NaCl is about 1:10 and the milling speed is 150–200 rpm. Following a typical experiment for synthesizing c-TiO2@a-TiO2-x(OH)y, P25 was milled with Na/NaCl powder in a weight ratio of 1:4 under argon atmosphere at room temperature using a Retsch PM400 planetary ball mill, at a milling rate of 180 rpm. All above procedures were conducted in a glovebox filled with Ar. The samples (4-c-TiO2@a-TiO2-x(OH)y) were milled for 4 h before being collected and washed several times with deionized water to remove the residual Na/NaCl. Finally, the samples were vacuum-dried at room temperature to isolate the final, dark-blue a-TiO2-x(OH)y@c-TiO2 powder. All characterized and tested c-TiO2@a-TiO2-x(OH)y samples were washed 3 times with Na content of about 2 wt% Na if not mentioned or emphasized the wash times. Other samples with different oxygen-vacancy concentrations (1- c-TiO2@a-TiO2-x(OH)y, 2- c-TiO2@a-TiO2-x(OH)y and 3-c-TiO2@a-TiO2-x(OH)y) were prepared by milling P25 with Na/NaCl powder in a weight ratio of 1:1 at a milling rate of 150 rpm for 4 h, milling with Na/NaCl powder in a weight ratio of 1:1 at a milling rate of 180 rpm for 4 h, and milling with Na/NaCl powder in a weight ratio of 1:2 at a milling rate of 180 rpm for 4 h, respectively. Materials characterization EPR measurements were conducted on a Bruker EMXmicro-6/1 EPR spectrometer. The powdered samples were packed into a Wilmad EPR quartz tube and closed with a cap. At least two co-added scans were performed for each sample (receiver gain: 30 dB) using a 1000.0 G sweep width of 60.0 s (or a 1500.0 G sweep width of 90.0 s), and a microwave power of ~5.0 mW (attenuation: 16 dB). The DPPH was used as a standard to calibrate the central field every time before the sample measurements. Low-temperature experiments were enabled by liquid nitrogen. PXRD was performed on a Bruker D2-Phaser X-ray diffractometer, using Cu Kα radiation at 30 kV. XPS was performed in an ultrahigh vacuum chamber with base pressure of 10−9 torr. This system used a Thermo Scientific K-Alpha XPS spectrometer with an Al Kα X-ray source (1,486.7 eV) operating at 12 kV and 6 A. The spectra were obtained using an analyzer pass energy of 50 eV with a resolution of 0.1 eV. HR-TEM images were acquired on an aberration-corrected FEI Titan 380–300 microscope operated at a 300 kV accelerating voltage. HR-TEM images were also evaluated by calculating their 2D Fourier transform, yielding information regarding their crystal lattices. ICP-OES was carried out on a Thermo Scientific iCAP 7000 Series ICP Spectrometer. The surface areas of samples were measured on an Autosorb-1 system (Quantachrome, USA) from N2 adsorption isotherms obtained via the BET method. UV-visible diffuse reflectance spectra were obtained for dry-pressed disk samples using a Cary 500 Scan Spectrophotometer (Varian, USA) over a range of 200 to 800 nm. BaSO4 was used as a reflectance standard in the UV-visible diffuse reflectance experiment. The photoluminescence (PL) spectra were measured on an Andor Shamrock SR-750 fluorescence spectrometer with a Xe-lamp as an excitation source (Andor Technology Ltd, Belfast, UK). A CCD detector combined with a monochromator was used for signal collection. The 1H solid state MAS NMR spectra were obtained at a spinning rate of 20 kHz. The NMR Spectra were calibrated to reference adamantane with optimized parameters: pulse width (pwX90) = 3.45 microseconds, fine power (aX90) = 2700, course power (tpwr) = 59, and synthesizer offset (tof) = 1192.9. Number of scans = 64, delay time = 6 s. Samples for 1H solid state MAS NMR were treated by H2 and then transferred to a glovebox with an Ar atmosphere for sample loading. The overall local temperature of the sample in a high-intensity batch reactor (CO2/H2 ratio = 5:1, 40 suns illumination) was estimated by allowing the system to approach equilibrium and comparing the final composition with ASPEN Plus V9 output, using the Peng-Robinson property package with the Gibbs reactor. Measurements of the gas-phase photocatalytic reduction CO2 For gas-phase photocatalytic testing in a batch reactor, the samples were prepared by drop-casting samples, dispersed in DI water, onto binder-free borosilicate glass microfiber filters with an area of 1 cm2 and then drying them under ambient conditions prior to reactor testing. Gas-phase photocatalytic experiments were conducted in a custom-fabricated 11.8 mL stainless steel batch reactor with a fused silica view port sealed with a Viton O-ring. The reactor was evacuated using an Alcatel dry pump before being purged with H2 (99.9995%) at a flow rate of 15 sccm. The pressure inside the reactor was monitored during the reaction using an Omega PX309 pressure transducer. After H2 purging, the reactor was infiltrated with H2 and CO2 gas, to achieve the desired pressure ratio, before being sealed. The reactor was irradiated using a 300 W Xe lamp at 4.0 W cm−2 light intensity for a duration of 1 h per run. For the tests using a CO2/H2 gas ratio of 1:1, the reactor was infiltrated with 15 psi CO2 and 15 psi H2. For the tests using a CO2/H2 gas ratio of 5:1, CO2 was infiltrated until pressures of 25 psi CO2 and 5 psi H2 were reached. Visible light with an intensity of 2.8 W cm−2 was obtained using a 420 nm optical cut-off filter to remove UV light from the Xe source. Product gases were analyzed using flame ionization and thermal conductivity detectors installed in an SRI-8610 gas chromatograph equipped with 30 Mole Sieve 13a and 60 Haysep D column. Isotope product gases were measured using an Agilent 7890 A gas chromatographic mass spectrometer with a 60 m GS-Carbon PLOT column fed to the mass spectrometer. The hydrogenation experiments in a flow system were conducted in a tubular quartz reactor with an inner diameter of 2 mm, in which ∼20 mg of catalyst sample was packed and irradiated using an unfiltered 300 W Xe lamp. The diameter of the light spot was ~2 cm, with an area of about 3.14 cm2, which could fully cover the sample. An OMEGA temperature controller was attached to a heating cartridge inserted into the copper block, along with a thermocouple inserted into the quartz tube (in contact with the catalyst bed) for control of the catalyst temperature. In situ DRIFTS measurement The in situ DRIFTS experiments were performed on a Thermo Scientific Nicolet iS50 FTIR Spectrometer with a mercury cadmium telluride detector cooled with liquid nitrogen. The spectrometer was equipped with a Harrik Praying Mantis diffuse reflection accessory and a Harrick high-temperature reaction chamber. The temperature was controlled with a Harrick ATC-024-3 Temperature Controller. Fresh reduced oxide sample was loaded onto a sampling accessory with a diamond-coated zinc selenide window and then sealed in the reactor. The reactor was heated to 300 °C and kept for 2 h in He (20 sccm) to remove the moisture and surface contaminants on the sample. After cooling the reactor down to 25 °C in He, the IR spectrum was collected to provide the baseline for subsequent differential spectra. The spectra were collected at 25 °C in H2 (5 sccm) and He (15 sccm) within 0 to 20 min. The H2 was then turned off while He (15 sccm) was kept purging at 25 °C for 30 min to remove residual H2 within the reactor. Afterward, the spectra were collected under flowing H2 (2 sccm), CO2 (2 sccm), and He (16 sccm) at various temperatures (25 °C, 100 °C, 150 °C, 200 °C, 250 °C, and 300 °C). Each temperature was held for 20 min. Computational details The amorphous TiO2 (a-TiO2) model was prepared via the melting and quenching method as demonstrated elsewhere53. The properties of the prepared a-TiO2 model were in agreement with experimental and theoretical data available. The structural analysis of these models suggests that local structural features of bulk crystalline TiO2 (c-TiO2) are retained in the a-TiO2 model. In order to obtain the a-TiO2 surface in the present study, a vacuum of about 20 Å was added in the z-direction of 2 × 2 × 4 bulk supercell (96-atom model)54. The a-TiO2-x model was created by removing the surface O atoms coordinated to Ti having coordination numbers of four, five, and six, denoted by Ti-4c, Ti-5c, and Ti-6c, respectively (Fig. 4a), creating, in total, 10 a-TiO2-x model surfaces. For c-TiO2, we investigated the most thermodynamically stable (101) surface of anatase with two models of crystalline TiO2-x (c-TiO2-x) surface created by removing a two-fold coordinated O atom (denoted as O2c) and a three-fold coordinated O atom (denoted as O3c) (Supplementary Fig. 26). The anatase (101) surface is represented by a periodic supercell slab of 144 atoms and a vacuum region greater than 12 Å. The plane wave pseudopotential approach, together with the Perdew-Burke-Ernzerhof55 exchange-correlation functional and Vanderbilt ultrasoft pseudopotentials56, was utilized throughout. The kinetic energy cutoffs of 544 and 5440 eV were used for the smooth part of the electronic wavefunctions and augmented electron density, respectively. The Quantum-ESPRESSO software package57, was used to perform the calculations. All calculations were spin polarized. The structures were relaxed by using a conjugate gradient minimization algorithm until the magnitude of residual Hellman-Feynman force on each atom was less than 10−2 Ry/Bohr. In all electronic density of states (DOS) plots, a conventional Gaussian smearing of 0.007 Ry was utilized. Appreciable underestimation of band gap and delocalization of d and f electrons are well-known limitations of DFT. Therefore, DFT with Hubbard energy correction (DFT + U) formalism was used in this study with U = 4.2 eV applied to Ti 3d electrons for analyzing electronic properties of crystalline and amorphous TiO2. The value of U for Ti was chosen not solely on the basis of band gap but also depending on the property of interest58, which in the current study is the catalytic behavior of TiO2 that in-turn depends upon the position of band gap states and their effect on the electronic structure. This value of U for Ti is consistent with theoretical investigations by ref. 59, who calculated it by fitting the peak positions for surface oxygen vacancies to experimental X-ray photoelectron spectroscopy data. In order to understand the charge mobility (i.e., localization of energy states in a-TiO2-x and c-TiO2-x), we further conducted Inverse Participation Ratio (IPR) analysis. IPR is capable of determining the charge localization of the tail states and the mobility band gap of materials60. The IPR of an orbital ψn(ri→), I(ψn), is accordingly defined by5 I(ψn)=N∑i=1N∣ψn(r→i)∣4[∑i=1N∣ψn(r→i)∣2]2 where N is the number of volume elements in the cell and i is the index of the volume element. Ideally, a localized orbital means I(ψ) = N, whereas a delocalized orbital means I(ψ) = 1. The IPR can identify a level as belonging to the delocalized band, to the partially localized band tail, or to the highly localized band gap. Supplementary information Supplementary Information Peer Review File Description of Additional Supplementary Files Supplementary Movie 1 Supplementary information The online version contains supplementary material available at 10.1038/s41467-022-34798-1. Acknowledgements G.A.O. acknowledges the financial support of the Ontario Ministry of Research and Innovation (MRI), the Ministry of Economic Development, Employment and Infrastructure (MEDI), the Ministry of the Environment and Climate Change’s (MOECC) Best in Science (BIS) Award, Ontario Center of Excellence Solutions 2030 Challenge Fund, Ministry of Research Innovation and Science (MRIS) Low Carbon Innovation Fund (LCIF), Imperial Oil, the University of Toronto’s Connaught Innovation Fund (CIF), Connaught Global Challenge (CGC) Fund, and the Natural Sciences and Engineering Research Council of Canada (NSERC). L.H. and X.H.Z. acknowledges the support from the National Natural Science Foundation of China (51920105005, 51821002, 21902113), 111 Project, and the Collaborative Innovation Center of Suzhou Nano Science & Technology. K.K.G. acknowledges Calcul Québec (www.calculquebec.ca) and Compute Canada (www.computecanada.ca) supercomputing facility; infrastructural support from Canada Foundation for Innovation, and the financial support from the Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery grant program, (RGPIN-2020-05924) and the Canada Research Chairs Program. C.M. acknowledges financial support from the UofT Faculty of Arts & Science Postdoctoral Fellowship. P.N.D. acknowledges financial support from the NSERC PDF program. This research used resources of the Advanced Photon Source, a U.S. Department of Energy (DOE) Office of Science User Facility operated for the DOE Office of Science by Argonne National Laboratory under Contract no. DE-AC02-06CH11357. A. Tountas’ help on ASPEN estimation is acknowledged by all authors. Author contributions Z.L. and C.M. contributed equally to this work. G.A.O., X.Z., and Z.L. conceived and designed the experiments. Z.L. and Q.P. carried out the synthesis of the materials and catalytic testing. K.K.G. conducted and analyzed the DFT calculations. Z.L., C.M., M.X., J.W., F.M.A., and D.M.M. performed materials characterizations. Z.L., C.M., L.W., T.H., R.S., Q.G., and P.N.D. contributed to data analysis. G.A.O., X.Z., and K.K.G. supervised the project. G.A.O., X.Z., K.K.G., L.H., C.M., and Z.L. wrote the paper. P.N.D. edited the paper. All authors discussed the results and commented on the manuscript. Peer review Peer review information Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. Peer reviewer reports are available. Data availability The data that support the plots within this paper and other finding of this study are available from the corresponding author upon reasonable request. Competing interests The authors declare no competing interests. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. These authors contributed equally: Zhao Li, Chengliang Mao. ==== Refs References 1. Finney JL Modelling the structures of amorphous metals and alloys Nature 1977 266 309 314 10.1038/266309a0 2. Pei Y Synthesis and catalysis of chemically reduced metal–metalloid amorphous alloys Chem. Soc. Rev. 2012 41 8140 8162 10.1039/c2cs35182j 22907172 3. Kamiya T Hosono H Material characteristics and applications of transparent amorphous oxide semiconductors NPG Asia Mater. 2010 2 15 22 10.1038/asiamat.2010.5 4. Jeon S Gated three-terminal device architecture to eliminate persistent photoconductivity in oxide semiconductor photosensor arrays Nat. Mater. 2012 11 301 305 10.1038/nmat3256 22367002 5. Mitchell NP Nash LM Hexner D Turner AM Irvine WTM Amorphous topological insulators constructed from random point sets Nat. Phys. 2018 14 380 385 10.1038/s41567-017-0024-5 6. Wesenberg D Liu T Balzar D Wu M Zink BL Long-distance spin transport in a disordered magnetic insulator Nat. Phys. 2017 13 987 993 10.1038/nphys4175 7. Sacépé B Localization of preformed Cooper pairs in disordered superconductors Nat. Phys. 2011 7 239 244 10.1038/nphys1892 8. Dubouchet T Collective energy gap of preformed Cooper pairs in disordered superconductors Nat. Phys. 2019 15 233 236 10.1038/s41567-018-0365-8 9. Ball JM Petrozza A Defects in perovskite-halides and their effects in solar cells Nat. energy 2016 1 16149 10.1038/nenergy.2016.149 10. Taylor FH Buckeridge J Catlow CRA Defects and oxide ion migration in the solid oxide fuel cell cathode material LaFeO3 Chem. Mater. 2016 28 8210 8220 10.1021/acs.chemmater.6b03048 11. Yu X Marks TJ Facchetti A Metal oxides for optoelectronic applications Nat. Mater. 2016 15 383 396 10.1038/nmat4599 27005918 12. Hosono H Kim J Toda Y Kamiya T Watanabe S Transparent amorphous oxide semiconductors for organic electronics: application to inverted OLEDs Proc. Natl Acad. Sci. USA 2017 114 233 238 10.1073/pnas.1617186114 28028243 13. Cahill DG Pohl RO Thermal conductivity of amorphous solids above the plateau Phys. Rev. B 1987 35 4067 4073 10.1103/PhysRevB.35.4067 14. Beekman M Morelli DT Nolas GS Better thermoelectrics through glass-like crystals Nat. Mater. 2015 14 1182 1185 10.1038/nmat4461 26585077 15. Moya X Kar-Narayan S Mathur ND Caloric materials near ferroic phase transitions Nat. Mater. 2014 13 439 450 10.1038/nmat3951 24751772 16. Simonov A Goodwin AL Designing disorder into crystalline materials Nat. Rev. Chem. 2020 4 657 673 10.1038/s41570-020-00228-3 37127977 17. Catlow, R. Defects and disorder in crystalline and amorphous solids, vol. 418. Springer Science & Business Media, 2012. 18. Jia J Heterogeneous catalytic hydrogenation of CO2 by metal oxides: defect engineering—perfecting imperfection Chem. Soc. Rev. 2017 46 4631 4644 10.1039/C7CS00026J 28635998 19. Goldsmith BR Peters B Johnson JK Gates BC Scott SL Beyond ordered materials: understanding catalytic sites on amorphous solids ACS Catal. 2017 7 7543 7557 10.1021/acscatal.7b01767 20. Chen X Liu L Huang F Black titanium dioxide (TiO2) nanomaterials Chem. Soc. Rev. 2015 44 1861 1885 10.1039/C4CS00330F 25590565 21. Liu X Progress in black titania: a new material for advanced photocatalysis Adv. Energy Mater. 2016 6 1600452 10.1002/aenm.201600452 22. Dhakshinamoorthy A Navalon S Corma A Garcia H Photocatalytic CO2 reduction by TiO2 and related titanium containing solids Energy Environ. Sci. 2012 5 9217 9233 10.1039/c2ee21948d 23. Xu H Recent advances in TiO2-based photocatalysis J. Mater. Chem. A 2014 2 12642 12661 10.1039/C4TA00941J 24. Tao J Luttrell T Batzill M A two-dimensional phase of TiO2 with a reduced bandgap Nat. Chem. 2011 3 296 300 10.1038/nchem.1006 21430688 25. Scanlon DO Band alignment of rutile and anatase TiO2 Nat. Mater. 2013 12 798 801 10.1038/nmat3697 23832124 26. Chen X Liu L Yu PY Mao SS Increasing solar absorption for photocatalysis with black hydrogenated titanium dioxide nanocrystals Science 2011 331 746 750 10.1126/science.1200448 21252313 27. Li L Sub-10 nm rutile titanium dioxide nanoparticles for efficient visible-light-driven photocatalytic hydrogen production Nat. Commun. 2015 6 5881 10.1038/ncomms6881 25562287 28. Sachs M Effect of oxygen deficiency on the excited state kinetics of WO3 and implications for photocatalysis Chem. Sci. 2019 10 5667 5677 10.1039/C9SC00693A 31293751 29. Zhao Y Ultrafine NiO nanosheets stabilized by TiO2 from monolayer NiTi-LDH precursors: An active water oxidation electrocatalyst J. Am. Chem. Soc. 2016 138 6517 6524 10.1021/jacs.6b01606 27159825 30. Stephan DW The broadening reach of frustrated Lewis pair chemistry Science 2016 354 aaf7229 10.1126/science.aaf7229 27940818 31. Chiesa M Paganini MC Livraghi S Giamello E Charge trapping in TiO2 polymorphs as seen by Electron Paramagnetic Resonance spectroscopy Phys. Chem. Chem. Phys. 2013 15 9435 9447 10.1039/c3cp50658d 23695705 32. Hurum DC Agrios AG Gray KA Rajh T Thurnauer MC Explaining the enhanced photocatalytic activity of degussa P25 mixed-phase TiO2 using EPR J. Phys. Chem. B 2003 107 4545 4549 10.1021/jp0273934 33. Zhao Y Tuning oxygen vacancies in ultrathin TiO2 nanosheets to boost photocatalytic nitrogen fixation up to 700 nm Adv. Mater. 2019 31 1806482 10.1002/adma.201806482 34. Damay P David T Sienko MJ Conduction electron spin resonance in metallic calcium J. Chem. Phys. 1974 61 4369 4371 10.1063/1.1681751 35. Guo Y Hydrogen-location-sensitive modulation of the redox reactivity for oxygen-deficient TiO2 J. Am. Chem. Soc. 2019 141 8407 8411 10.1021/jacs.9b01836 31083914 36. Nosaka AY Fujiwara T Yagi H Akutsu H Nosaka Y Characteristics of water adsorbed on TiO2 photocatalytic systems with increasing temperature as studied by solid-state 1H NMR spectroscopy J. Phys. Chem. B 2004 108 9121 9125 10.1021/jp037297i 37. Li G Ionothermal synthesis of black Ti3+-doped single-crystal TiO2 as an active photocatalyst for pollutant degradation and H2 generation J. Mater. Chem. A 2015 3 3748 3756 10.1039/C4TA02873B 38. Crocker M 1H NMR spectroscopy of titania. Chemical shift assignments for hydroxy groups in crystalline and amorphous forms of TiO2 J. Chem. Soc., Faraday Trans. 1996 92 2791 2798 10.1039/ft9969202791 39. Jackson P Parfitt GD Infra-red study of the surface properties of rutile. Water and surface hydroxyl species Trans. Faraday Soc. 1971 67 2469 2483 10.1039/tf9716702469 40. Liu C Structure–activity relationship of surface hydroxyl groups during NO2 adsorption and transformation on TiO2 nanoparticles Environ. Sci. Nano 2017 4 2388 2394 10.1039/C7EN00920H 41. Davies DW Descriptors for electron and hole charge carriers in metal oxides J. Phys. Chem. Lett. 2020 11 438 444 10.1021/acs.jpclett.9b03398 31875393 42. Stephan DW Frustrated lewis pairs: from concept to catalysis Acc. Chem. Res. 2015 48 306 316 10.1021/ar500375j 25535796 43. Nosaka AY Fujiwara T Yagi H Akutsu H Nosaka Y Photocatalytic reaction sites at the TiO2 surface as studied by solid-state 1H NMR spectroscopy Langmuir 2003 19 1935 1937 10.1021/la0268146 44. Yang Y An unusual strong visible-light absorption band in red anatase TiO2 photocatalyst induced by atomic hydrogen-occupied oxygen vacancies Adv. Mater. 2018 30 1704479 10.1002/adma.201704479 45. Chen X Properties of disorder-engineered black titanium dioxide nanoparticles through hydrogenation Sci. Rep. 2013 3 1510 10.1038/srep01510 23528851 46. Ramis G Busca G Lorenzelli V Low-temperature CO2 adsorption on metal oxides: spectroscopic characterization of some weakly adsorbed species Mater. Chem. Phys. 1991 29 425 435 10.1016/0254-0584(91)90037-U 47. Nelson NC Nguyen M-T Glezakou V-A Rousseau R Szanyi J Carboxyl intermediate formation via an in situ-generated metastable active site during water-gas shift catalysis Nat. Catal. 2019 2 916 924 10.1038/s41929-019-0343-2 48. Liao LF Lien CF Shieh DL Chen MT Lin JL FTIR study of adsorption and photoassisted oxygen isotopic exchange of carbon monoxide, carbon dioxide, carbonate, and formate on TiO2 J. Phys. Chem. B 2002 106 11240 11245 10.1021/jp0211988 49. Chuang CC Wu WC Huang MC Huang IC Lin JL FTIR study of adsorption and reactions of methyl formate on powdered TiO2 J. Catal. 1999 185 423 434 10.1006/jcat.1999.2516 50. Mino L Spoto G Ferrari AM CO2 capture by TiO2 anatase surfaces: A combined DFT and FTIR study J. Phys. Chem. C. 2014 118 25016 25026 10.1021/jp507443k 51. Baltrusaitis J Schuttlefield J Zeitler E Grassian VH Carbon dioxide adsorption on oxide nanoparticle surfaces Chem. Eng. J. 2011 170 471 481 10.1016/j.cej.2010.12.041 52. Wang Z-J Song H Liu H Ye J Coupling of solar energy and thermal energy for carbon dioxide reduction: status and prospects Angew. Chem. Int. Ed. 2020 59 8016 8035 10.1002/anie.201907443 53. Kaur K Singh CV Amorphous TiO2 as a photocatalyst for hydrogen production: A DFT study of structural and electronic properties Energy Procedia 2012 29 291 299 10.1016/j.egypro.2012.09.035 54. Staykov A Kamachi T Ishihara T Yoshizawa K Theoretical study of the direct synthesis of H2O2 on Pd and Pd/Au surfaces J. Phys. Chem. C. 2008 112 19501 19505 10.1021/jp803021n 55. Perdew JP Burke K Ernzerhof M Generalized gradient approximation made simple Phys. Rev. Lett. 1996 77 3865 3868 10.1103/PhysRevLett.77.3865 10062328 56. Vanderbilt D Soft self-consistent pseudopotentials in a generalized eigenvalue formalism Phys. Rev. B 1990 41 7892 7895 10.1103/PhysRevB.41.7892 57. Giannozzi P QUANTUM ESPRESSO: a modular and open-source software project for quantum simulations of materials J. Phys.: Condens. Matter 2009 21 395502 21832390 58. Deskins NA Dupuis M Electron transport via polaron hopping in bulk TiO2: A density functional theory characterization Phys. Rev. B 2007 75 195212 10.1103/PhysRevB.75.195212 59. Morgan BJ Watson GW A DFT+U description of oxygen vacancies at the TiO2 rutile (110) surface Surf. Sci. 2007 601 5034 5041 10.1016/j.susc.2007.08.025 60. Justo JF de Brito Mota F Fazzio A First-principles investigation of a-SiNx:H Phys. Rev. B 2002 65 073202 10.1103/PhysRevB.65.073202
PMC009xxxxxx/PMC9704871.txt
==== Front Medicine (Baltimore) Medicine (Baltimore) MD Medicine 0025-7974 1536-5964 Lippincott Williams & Wilkins Hagerstown, MD 36451406 00030 10.1097/MD.0000000000031699 3 3400 Research Article Study Protocol Systematic Review Opioids for treating refractory dyspnea in patients with heart failure: A protocol for systematic review and meta-analysis Lyu Jianguo Bachelor [email protected] a Sun Xianghong Master [email protected] a Chen Mingming Bachelor [email protected] b Li Sijia Bachelor b* a Intensive Care Unit, Shengjing Hospital of China Medical University, Liaoning, China b Interventional Ward, Shengjing Hospital of China Medical University, Liaoning, China. * Correspondence: Sijia Li, Interventional Ward, Shengjing Hospital of China Medical University, Liaoning 110000, China (e-mail: [email protected]). 25 11 2022 25 11 2022 101 47 e3169915 10 2022 17 10 2022 Copyright © 2022 the Author(s). Published by Wolters Kluwer Health, Inc. 2022 https://creativecommons.org/licenses/by/4.0/ This is an open access article distributed under the Creative Commons Attribution License 4.0 (CCBY), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Background: Dyspnea is a hallmark symptom of heart failure. The existing clinical studies have indicated that opioid can effectively improve the clinical symptoms of heart failure patients with dyspnea. However, there has not been any relevant systematic review and meta-analysis. We performed a protocol for systematic review and meta-analysis to evaluate the safety and efficacy of opioid therapy for heart failure patients with refractory dyspnea. Methods: We searched 3 foreign electronic databases (Cochrane Library, Embase, Pubmed) and 4 Chinese electronic databases (China National Knowledge Infrastructure, Wang Fang Database, Chinese Biomedical Literature Database and Chinese Scientific Journal Database) to collect potential studies from their inceptions to October 2022. The risk of bias in the included articles was assessed according to the Risk of Bias Assessment Tool in Cochrane Handbook of Systematic Reviews (5th edition). The quality of evidence was assessed by the Grading of Recommendations Assessment, Development and Evaluation approach. Data were analyzed using the RevMan Version 5.4.1. Results: This study will evaluate whether opioid is effective and safe for treating refractory dyspnea in patients with heart failure. Conclusion: This meta-analysis will provide comprehensive evidence of opioid therapy for heart failure patients with dyspnea. dyspnea heart failure meta-analysis opioid OPEN-ACCESSTRUE ==== Body pmc1. Introduction Heart failure (HF) is a syndrome characterized by symptoms (such as breathlessness, ankle swelling, and fatigue) and signs (e.g., raised jugular venous pressure, pulmonary crackles, and peripheral edema) caused by structural or functional cardiac abnormalities that lead to elevated intracardiac pressures or a reduced cardiac output at rest or during stress.[1–3] HF is a leading and increasing cause of morbidity and mortality worldwide. The prevalence of HF is age-dependent, ranging from < 2% of people younger than 60 years to more than 10% of those older than 75 years.[4] As a result of aging of the general population and improved treatment of acute cardiovascular events, the prevalence of heart failure is projected to increase by 25% in the next 20 years.[5] For HF patients, healthcare professionals should focus on not only prolonging survival but also improving their quality of life. Refractory dyspnea (defined as dyspnea that occurs despite optimization of traditional medical management, such as inhaled medications) is a frequently encountered and challenging problem in HF.[6,7] Refractory dyspnea has profound implications on patient quality of life and contributes to significant psychological distress.[8] Although an off-label usage, there is a theoretical role for opioids in the management of refractory dyspnea.[9,10] By binding to opioid receptors that are heavily concentrated in the medulla, where respiratory rhythm is generated, opioids may theoretically lessen refractory dyspnea by decreasing minute ventilation and/or dulling respiratory response to chemoreceptor stimulation from hypoxemia and hypercapnea.[11] Opioid receptors are also located in the deeper cerebral cortex (e.g., insula, thalamus, anterior cingulate cortex, etc.) and opioid action at these sites may help modulate the perception of dyspnea.[12] Currently, the opioid therapy for HF patients with refractory dyspnea have not been widely used. Therefore, we performed a protocol for systematic review and meta-analysis to assess the safety and efficacy of opioid therapy for HF patients with refractory dyspnea. 2. Methods The protocol of this review was registered in PROSPERO (CRD42021252201). Meanwhile, it was reported follow the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses.[13] Ethical approval is not required because this review will retrieve publicly available scientific literature. 2.1. Inclusion and exclusion criteria PICOS will be applied, including population, intervention, comparison, outcome, and study.[14] 2.1.1. Type of participants. Adults patients (age > 18 years) from HF with dyspnea were included. Regardless of gender, race, occupation, education, nationality, etiology, and severity. 2.1.2. Type of intervention. Patients in intervention group received opioid therapy, regardless of the route of administration used for the treatment of dyspnea. 2.1.3. Type of comparator (S)/control. The control group’s treatment is not limited, including no treatment and placebo. 2.1.4. Type of outcome measurements. Primary outcomes include dyspnea relief at 8, 12, 24 hours/day 1, 48, and 72 hours, and the incidence of worsening renal function (defined as an increase in serum creatinine of ≥ 0.3mg/dl). Secondary outcomes include heart rate, left ventricular ejection fraction, blood urea nitrogen, systolic blood pressure, serum creatinine and adverse events. 2.1.5. Type of study design. Only randomized controlled trials are included. The language will be limited to Chinese and English. 2.2. Search methods for identification of studies We searched 3 foreign electronic databases (Cochrane Library, Embase, Pubmed) and 4 Chinese electronic databases [China National Knowledge Infrastructure (CNKI), Wang Fang Database, Chinese Biomedical Literature Database (CBM) and Chinese Scientific Journal Database (VIP)] to collect potential studies from their inceptions to October 2022. The following search terms will be used: heart failure, opioid and dyspnea. A draft search strategy using Pubmed, 1 of the planned electronic databases to be searched, is presented in Table 1. Table 1 Search strategy (PubMed). Number search terms #1 opioid[Ti/Ab] #2 morphine[Ti/Ab] #3 fentanyl[Ti/Ab] #4 demerol[Ti/Ab] #5 codeine[Ti/Ab] #6 oxycodone[Ti/Ab] #7 pethidine[Ti/Ab] #8 remifentanil[Ti/Ab] #9 sufentanil[Ti/Ab] #10 buprenorphine[Ti/Ab] #11 #1 OR #2 OR #3 OR #4 OR #5 OR #6 OR #7 OR #8 OR #9 OR #10 #12 heart failure[Ti/Ab] #13 cardiac failure[Ti/Ab] #14 cardiac insufficiency[Ti/Ab] #15 #12 OR #13 OR #14 #16 dyspnea[Ti/Ab] #17 breathless[Ti/Ab] #18 #16 OR #17 #19 #11 AND #15 AND #18 Ab = abstract, Ti = title. 2.3. Studies selection Studies will be identified using Note Express 3.2.0. After the initial removal of duplicate studies, 2 reviewers will independently screen titles and abstracts based on the eligibility criteria. If studies contain insufficient information to make a decision about eligibility, we will try to contact authors of the original reports to obtain further details. During the procedure, disagreements will be resolved by discussion or consensus with the third reviewer. Study selection will be performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses flowchart (Fig. 1). Figure 1. Flow diagram. 2.4. Data extraction Two researchers extracted literature information based on inclusion and exclusion criteria, including the following: Study characteristics: author, year, study design, sample size and follow-up time; Patient characteristics: age, sex, BMI, race, and nationality; Intervention: intervention measures in the experimental group, intervention measures in the control group; Outcome of the study: 2 researchers cross-checked the extraction results of the documents. Disagreements are resolved through discussion among all authors. 2.5. Evaluation of risk of bias The risk of bias in the included articles was assessed according to the Risk of Bias Assessment Tool in Cochrane Handbook of Systematic Reviews (5th edition).[15] Two investigators evaluated the risk of bias for each item as unclear, low risk, and high risk of bias. In case of disagreement, the third investigator was responsible for resolving it. 2.6. Evaluation of the evidence quality of the included studies The quality of evidence was assessed by the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach.[16] This tool aims to assess the quality of evidence for each outcome indicator in the study. The 2 authors will independently evaluate the evidence of the results and should describe in detail the degradation or upgrade factors that affect the quality of the evidence to ensure the reliability and transparency of the results. Any disagreements will be resolved through discussion by 2 authors. The overall quality of evidence was judged as “high,” “moderate” or “very low.” 2.7. Data synthesis and statistical analysis Meta-analyses were conducted where applicable; otherwise, outcomes were presented in narrative form. Data were analyzed using the RevMan Version 5.4.1. Next, relative risk (RR) for dichotomous outcomes with corresponding 95% confidence intervals (CIs) were computed for individual trials. χ2 and Higgins I2 tests were used to assess heterogeneity among the included studies. If significant heterogeneity (P ≤ .10 for χ2 test results or I2 ≥ 50%) was obtained, we used a random-effects model, otherwise a fixed-effects model was used. And a P < .05 was taken to indicate statistical significance. To assess the robustness of the results, meta-regression analyses (STATA version 12.0) were carried out for sensitivity analysis to test the influence of potential effect modifiers. The P value of Egger’s linear regression test was used to assess the presence of publication bias in included articles for each outcome. 3. Discussion The incidence of HF increases year by year with a large aging population in the world. Dyspnea is a hallmark symptom of HF. Therefore, the search for a safe and effective drug has become a much-talked-about and received widespread attention from the global medical community. The existing clinical studies have indicated that opioid can effectively improve the clinical symptoms of patients with HF, such as dyspnea, and has fewer side effects.[17–19] However, the exact mechanism remains to be further explored. There has not been any meta-analysis of the clinical efficacy and safety of opioid for treating refractory dyspnea in patients with HF. We hope that this study can provide more rigorous medical evidence. However, there may be some potential deficiencies in this study. For instance, different doses and courses of treatment in the intervention group in the included trials may lead to relatively significant heterogeneity in the meta-analysis results. Further high quality randomized controlled trials are still required. Author contributions Data curation: Xianghong Sun. Formal analysis: Mingming Chen. Investigation: Mingming Chen. Writing – original draft: Jianguo Lyu. Writing – review & editing: Sijia Li. Abbreviation: HF = heart failure Data sharing not applicable to this article as no datasets were generated or analyzed during the current study. The authors have no conflicts of interest to disclose. This work is supported by the Natural Science Foundation of Liaoning Province (5737461124). How to cite this article: Lyu J, Sun X, Chen M, Li S. Opioids for treating refractory dyspnea in patients with heart failure: A protocol for systematic review and meta-analysis. Medicine 2022;101:47(e31699). ==== Refs References [1] Dharmarajan K Rich MW . Epidemiology, pathophysiology, and prognosis of heart failure in older adults. Heart Fail Clin. 2017;13 :417–26.28602363 [2] Rogers C Bush N . Heart failure: pathophysiology, diagnosis, medical treatment guidelines, and nursing management. Nurs Clin North Am. 2015;50 :787–99.26596665 [3] Skrzypek A Mostowik M Szeliga M . Chronic heart failure in the elderly: still a current medical problem. Folia Med Cracov. 2018;58 :47–56.30745601 [4] Orso F Fabbri G Maggioni AP . Epidemiology of heart failure. Handb Exp Pharmacol. 2017;243 :15–33.27718059 [5] Roger VL . Epidemiology of heart failure: a contemporary perspective. Circ Res. 2021;128 :1421–34.33983838 [6] van Riet EE Hoes AW Limburg A . Prevalence of unrecognized heart failure in older persons with shortness of breath on exertion. Eur J Heart Fail. 2014;16 :772–7.24863953 [7] Shariat M Mohan R Iwanochko MR . Unusual intense fluorodeoxyglucose uptake in the intercostal muscles due to severe shortness of breath in a patient with heart failure. World J Nucl Med. 2015;14 :66–7.25709551 [8] Breaden K Phillips J Agar M . The clinical and social dimensions of prescribing palliative home oxygen for refractory dyspnea. J Palliat Med. 2013;16 :268–73.23289922 [9] Verberkt CA van den Beuken-van EM Franssen FM . A randomized controlled trial on the benefits and respiratory adverse effects of morphine for refractory dyspnea in patients with COPD: protocol of the MORDYC study. Contemp Clin Trials. 2016;47 :228–34.26825021 [10] van Dijk M Mooren K van den Berg JK . Opioids in patients with COPD and refractory dyspnea: literature review and design of a multicenter double blind study of low dosed morphine and fentanyl (MoreFoRCOPD). BMC Pulm Med. 2021;21 :289.34507574 [11] Emery MJ Groves CC Kruse TN . Ventilation and the response to hypercapnia after morphine in Opioid-naive and Opioid-tolerant rats. Anesthesiology. 2016;124 :945–57.26734964 [12] Yamamoto Y Watanabe H Sakurai A . Effect of continuous intravenous oxycodone infusion in opioid-naive cancer patients with dyspnea. Jpn J Clin Oncol. 2018;48 :748–52.29893883 [13] Moher D Shamseer L Clarke M .; PRISMA-P Group. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev. 2015;4 :1.25554246 [14] Amir-Behghadami M Janati A . Population, intervention, comparison, outcomes and study (PICOS) design as a framework to formulate eligibility criteria in systematic reviews. Emerg Med J. 2020;37 :387–387.32253195 [15] Higgins JP Altman DG Gotzsche PC . The Cochrane collaboration’s tool for assessing risk of bias in randomised trials. BMJ. 2011;343 :d5928.22008217 [16] Atkins D Best D Briss PA . Grading quality of evidence and strength of recommendations. BMJ. 2004;328 :1490.15205295 [17] Williams SG Wright DJ Marshall P . Safety and potential benefits of low dose diamorphine during exercise in patients with chronic heart failure. Heart. 2003;89 :1085–6.12923038 [18] Johnson MJ McDonagh TA Harkness A . Morphine for the relief of breathlessness in patients with chronic heart failure--a pilot study. Eur J Heart Fail. 2002;4 :753–6.12453546 [19] Chua TP Harrington D Ponikowski P . Effects of dihydrocodeine on chemosensitivity and exercise tolerance in patients with chronic heart failure. J Am Coll Cardiol. 1997;29 :147–52.8996307
PMC009xxxxxx/PMC9708394.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36446560 10.1055/s-0042-1751074 220044 Original Article Lower Genital Tract Disease Association of Swede Score and 2011 IFCPC Nomenclature in Women with Abnormal Cytology A Associação entre o escore Swede e a nomenclatura IFCPC 2011 em mulheres com citologia anormalhttp://orcid.org/0000-0002-9011-4428 Campos Priscila Loyola 1 http://orcid.org/0000-0002-9885-2724 Guimarães Isabel Cristina Chulvis do Val 1 http://orcid.org/0000-0002-4212-0022 Fialho Susana Cristina Aidé Viviani 1 http://orcid.org/0000-0002-7207-7969 Martins Caroline Alves Oliveira 1 http://orcid.org/0000-0001-5062-1065 Rodrigues Fabiana Resende 1 http://orcid.org/0000-0003-3110-5270 Velarde Luiz Guilhermo Coca 1 http://orcid.org/0000-0001-5511-285X Monteiro Daniela da Silva Alves 1 1 Universidade Federal Fluminense, Maternal – Infant Department Niterói, RJ, Brasil Address for correspondence Priscila Loyola Campos, MSc Rua Marques do Paraná303, Niterói 24033-900, Rio de [email protected] 29 11 2022 10 2022 1 11 2022 44 10 938944 06 2 2022 02 5 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  To assess the association between two colposcopic indices, the Swede score and the 2011 International Federation of Cervical Pathology and Colposcopy (IFCPC) Nomenclature as well as to determine the efficacy of the Swede score with cutoffs of 7 and 8. Methods  In the present cross-sectional pilot study, 34 women who had at least 1 colposcopy-directed biopsy due to abnormal cytology were enrolled. The colposcopic findings were scored by both the Swede score and the 2011 IFCPC Nomenclature and were compared with each other. The Kappa coefficient and the McNemar test were used. Accuracy, sensitivity, specificity, and positive and negative predictive values (NPV and PPV, respectively) were calculated, as well as the effectiveness with cutoffs of 7 and 8 in identifying cervical intraepithelial neoplasm (CIN) 2+ when using the Swede score. Results  The correlation between the 2 colposcopic indices was 79.41%. The Kappa coefficient and the McNemar p-value were 0.55 and 0.37, respectively. The IFCPC Nomenclature had sensitivity, specificity, accuracy, PPV, and NPV of 85.71, 55.00, 67.64, 57.14, and 84.61%, respectively. The Swede score had sensitivity, specificity, accuracy, PPV, and NPV of 100, 63.15, 79.41, 68.18, and 100%, respectively. A Swede score cutoff of 7 for CIN 2+ detection had a specificity of 94.73%, while with a cutoff of 8 it increased to 100%. The sensitivity for both values was 60%. The PPV and NPV for cutoffs of 7 and 8 were 90 and 75 and 100 and 76%, respectively. Conclusion  Although both colposcopic indices have good reproducibility, the Swede score showed greater accuracy, sensitivity, and specificity in identifying CIN 2 + , especially when using a cutoff of 8. Resumo Objetivo  Avaliar a associação entre dois índices colposcópicos, o escore Swede e a Nomenclatura International Federation of Cervical Pathology and Colposcopy (IFCPC, na sigla em inglês) 2011, assim como determinar a eficácia do escore Swede com os pontos de corte 7 e 8. Métodos  Trata-se de um estudo transversal, com 34 mulheres incluídas, que realizaram colposcopia com biópsia dirigida devido a uma citologia anormal. Os achados colposcópicos foram categorizados pelo escore Swede e pela Nomenclatura IFCPC 2011 e comparados um com o outro. Foram avaliados o coeficiente Kappa e o teste de McNemar e foram calculados a acurácia, a sensibilidade, a especificidade e valores preditivos negativos e positivos (VPN e VPP, respectivamente) de cada índice, assim como a eficácia com os pontos de corte 7 e 8 do escore Swede para determinar as lesões de neoplasia intraepitelial cervical (NIC) 2 + . Resultados  A concordância entre os 2 índices foi de 79,41% e o coeficiente Kappa e o valor-p do teste de McNemar foram 0.55 e 0.37, respectivamente. Pela Nomenclatura IFCPC 2011, obtivemos como sensibilidade, especificidade, acurácia, VPP e VPN, respectivamente: 85,71, 55,00, 67,64, 57,14 e 84,61%. Pelo escore Swede obtivemos como sensibilidade, especificidade, acurácia, VPP e VPN, respectivamente: 100, 63,15, 79,41, 68,18 e 100%. O uso do escore Swede para detecção das lesões NIC 2+ obteve como especificidade 94,73% com o valor de corte de 7, enquanto o valor de corte 8 obteve 100%. A sensibilidade para ambos os cortes foi de 60%. O VPP e o VPN com os cortes 7 e 8 foram, respectivamente: 90,00 e 75,00 e 100,00 e 76,00%. Conclusão  Ambos os índices colposcópicos tiveram boa reprodutibilidade; no entanto, o escore Swede mostrou melhor acurácia, sensibilidade e especificidade em identificar as lesões NIC 2+ e o melhor ponto de corte para identificar as lesões NIC2+ foi com o valor 8. Keywords HPV colposcopy squamous intraepithelial lesions cervical neoplasm Palavras-chave HPV colposcopia lesões escamosas intraepiteliais neoplasia cervical ==== Body pmcIntroduction Cervical cancer is the fourth most common cancer worldwide, with a standardized incidence of 6.0 per 100,000 and a mortality of 3.41 per 100,000. 1 Although human papillomavirus (HPV) infection is a causative agent of cervical cancer, it is transient in most cases (70–90%) and lasts 1–2 years on average. 2 HPV16 is the most prevalent and carcinogenic high-risk HPV genotype, followed by HPV18, accounting for 70% of all cases of cervical cancer. 3 The evolution time of the precursor lesion for cervical cancer is, on average, between 10 and 15 years, which allows for early identification and treatment. In Brazil, cervical cancer screening is performed by cytology alone, which shows sensitivity and specificity for CIN 2+ (cervical intraepithelial neoplasm) ranging from 30 to 87% and from 86 to 100%, respectively. 4 5 Colposcopic classifications were created to categorize the findings observed in colposcopy. However, the colposcopic examination depends on the experience of the examiner and is considered to have low accuracy and reproducibility among specialists. 6 Due to this subjectivity, colposcopic indices that categorize these findings were created. The International Federation of Cervical Pathology and Colposcopy (IFCPC) has developed several colposcopic nomenclatures, including the 2011 IFCPC Nomenclature, which is the latest in effect. 7 This classification is descriptive and categorizes colposcopic findings according to their severity. 8 Another classification model is the Swede score, which was developed by Strander et al. 9 This system is comprised of 5 variables, each scored as 0, 1, or 2 points, consisting of acetowhiteness, type of margin, vessel patterns, iodine staining, and lesion size, which is evaluated as an independent variable. 10 The final value determines the score that will categorize the clinical suspicion of the lesion. 9 A cutoff point ≥ 7 is suggested for predicting premalignant lesions. 11 In the present study, we proposed a comparison of these two colposcopic indices in order to assess the association between them and the effectiveness of each individually. We compared the effectiveness between the Swede score cutoffs 7 and 8 in identifying high-grade cervical intraepithelial neoplasia lesions. Methods We carried out a cross-sectional pilot study approved by the Research Ethics Committee (number: 89861218.0.0000.5243) of the Antonio Pedro University Hospital of the Federal Fluminense University, Niterói, state of Rio de Janeiro, Brazil, between December 2019 and November 2020. Thirty-four women who attended the colposcopic clinic and had had at least one colposcopy-directed biopsy due to abnormal cytology with the recommendation to perform colposcopy were enrolled. Patients with normal colposcopic impressions in both colposcopic indices were excluded, as were pregnant women. The colposcopic findings were categorized according to the Swede score and the 2011 IFCPC Nomenclature by a single colposcopist. The 2011 IFCPC Nomenclature classifies colposcopic findings into normal, abnormal (minor, major, and nonspecific), suspected invasion, and miscellaneous ( Chart 1 ). Chart 1 Colposcopic findings description by the 2011 IFCPC Nomenclature 2011 IFCPC NOMENCLATURE Normal findings Original squamous epithelium mature or atrophic; columnar epithelium ectopy; metaplastic squamous epithelium. Abnormal findings Minor grade: Thin acetowhite epithelium, irregular or geographic border, and fine mosaic. Major grade: Dense acetowhite epithelium, rapid appearance of acetowhitening, cuffed crypt, and coarse mosaic with sharp or inner border. Nonspecific: leukoplakia, erosion, Lugol staining. Suspicious for invasion Atypical vessels, fragile vessels, irregular surface, necrosis, ulceration, tumor Miscellaneous Condyloma, polyp, inflammation, stenosis, endometriosis Abbreviation: IFCPC, International Federation for Cervical Pathology and Colposcopy. The Swede colposcopic scoring system is comprised of five variables: acetowhiteness, margins plus surface, vessel patterns, lesion size, and iodine staining, each of which is scored with 0, 1, or 2 points, and depends of the grade of these findings ( Chart 2 ). Chart 2 Colposcopic findings description by the Swede score and correlation by grade Swede score 0 1 2 Acepto uptake Zero or transparent Milky Opaque white Margins/Surface Diffuse Sharp and geographical satellites Sharp and surface level Vessels Fine and regular Absent Coarse or atypical Lesion size < 5 mm 5–15mm or 2 quadrants > 15 mm or 3–4 quadrants Iodine staining Brown Yellow Distinct yellow For interpretation of the study, according to the 2011 IFCPC Nomenclature, colposcopic reports with major alterations, nonspecific, or suggestive of invasion were categorized as high grade of suspicion, and those that presented minor alterations were classified as low grade of suspicion. According to the Swede score, those with results ≥ 5 were categorized as high-grade, and scores < 5 were categorized as low-grade ( Chart 3 ). Chart 3 Interpretation of the colposcopic indices by grade of suspicion 2011 IFCPC Swede score High grade Low grade High grade Low grade Major alterations, nonspecific and invasion Minor alterations ≥ 5 < 5 Abbreviation: IFCPC, International Federation for Cervical Pathology and Colposcopy. When evaluating the histopathology, we categorized the results into two groups. One group included results with minor alterations, which could be nonspecific atypias, CIN1, and normal findings. The second group included the cases with CIN 2 + , which could be CIN2, CIN3, and invasion or cancer ( Chart 4 ). Chart 4 Histopathological groups by lesion grade Histopathological evaluation MINOR CIN 2 + Nonspecific atypias, normal, and CIN 1 CIN 2, CIN 3, invasion/cancer Thus, we proceeded with the analysis of the effectiveness of each test. Accuracy, sensitivity, specificity, and positive and negative predictive values (PPV and NPV, respectively) were calculated. We also evaluated the correlation between the two indices and their reproducibility. The Kappa coefficient was also calculated, and the positive paired samples were analyzed using the McNemar test, which was calculated using the free software R version 3.6.1 (R Foundation, Vienna, Austria). Regarding the Swede score, we evaluated the effectiveness with cutoffs of 7 and 8 for identifying CIN 2 + . Results Characteristics of the Patients and Histopathological Results A total of 34 women were recruited into the study. The mean and median ages were 44 years old. The cytology included were: Atypical squamous cells of indeterminate significance (ASC-US); atypical squamous cell of undeterminated significance which can not be excluded from high-grade intraepithelial lesions (ASC-H); low-grade squamous intraepithelial lesion (LSIL); high-grade squamous intraepithelial lesion (HSIL); squamous carcinoma; atypical glandular cells of undetermined significance (AGC-NOS); atypical glandular cells favor neoplasic (AGC-H); and atypical cells with undertemined origin (ACs) ( Table 1 ). Table 1 Abnormal cytologies included CYTOLOGY Quantify ASC-US 05 LSIL 04 ASC-H 06 HSIL 05 AGC-NOS 08 AGC-H 03 INVASION/CANCER 02 AC 01 T otal 34 Abbreviations: AC, atypical cells with undetermined origin; AGC-NOS, atypical glandular cells of undetermined significance; ASC-H, atypical squamous cells of undetermined significance; ASC-US, atypical squamous cells of indeterminate significance; HSIL; LSIL, low-grade squamous intraepithelial lesion. Of the 34 cases, 9 (26.47%) had a histopathological report of CIN 3, 5 (14.7%) of CIN 2, 1 (2.9%) of squamous carcinoma, and 5 (17.64%) of CIN 1. The other 13 cases (38.23%) had minor alterations with normal results or nonspecific atypias. A total of 15 outcomes were identified with CIN 2+ (44.1%), and there were 19 minor alterations (55.88%) ( Table 2 ). Table 2 Histopathological classification by lesion grade Minor Quantify X CIN 2+ Quantity Normal 10 CIN 2 05 Nonspecific 03 CIN 3 09 CIN 1 06 Cancer 01 Total 19 Total 15 2011 IFCPC Nomenclature Results Based on the findings of the 2011 IFCPC Nomenclature, we had 21 cases (61.76%) categorized as having high-grade suspicion. The remaining 13 (38.23%) were categorized as low-grade suspicion ( Table 3 ). Table 3 Colposcopy results according to the 2011 IFCPC Nomenclature 2011 IFCPC Quantity n (%) Normal 1 (2.94) Minor changes 12 (35.29) Major changes 17 (50) Invasion 4 (11.76) Miscellaneous 0 (00) Total 34 (100) When evaluating the diagnostic performance of the 2011 IFCPC Nomenclature, we obtained an accuracy of 67.64%, a PPV of 57.14% (12/21), and a NPV of 84.61% (11/13). We calculated the sensitivity and specificity at 85.71 and 55%, respectively ( Table 4 ). Table 4 Association of histopathological results with the 2011 IFCPC Nomenclature 2011 IFCPC/Histopathological Normal/ Nonspecific CIN 1 CIN 2/3 Invasive cancer Total Normal − 01 − − 01 Minor changes 07 03 02 − 12 Major changes/nonspecific 07 02 08 − 17 Invasive cancer − − 03 01 04 Total 14 06 13 01 34 Abbreviation: IFCPC, International Federation for Cervical Pathology and Colposcopy. Swede Score Results Using the Swede score, we identified 22 cases (64.7%) with a high-grade suspicion and 12 (35.29%) with a low-grade suspicion ( Table 5 ). Table 5 Colposcopy results according to the Swede score Swede score Quantify 01 01 02 02 03 02 04 07 05 06 06 06 07 03 08 03 09 00 10 04 When evaluating the diagnostic performance of the Swede score, we obtained an accuracy of 79.41%. We identified a VPP of 68.18% (15/22) and a NPV of 100% (12/12). Sensitivity and specificity were 100 and 63.15%, respectively. When we used the cutoff ≥ 7 for CIN 2 + , we obtained 94.73% specificity, and with the cutoff ≥ 8, this percentage rose to 100%, with the sensitivity for both scores being 60%. The PPV and NPV for cutoffs of 7 and 8 were 90 and 75% and 100 and 76%, respectively ( Table 6 ). Table 6 Association of histopathological results with the Swede score Swede/Histopathological Normal/Nonspecific CIN 1 CIN 2/3 Invasive cancer Total < 5 08 04 − − 12 5–6 04 02 06 − 12 7 01 − − − 01 ≥ 8 − − 08 01 09 Total 13 06 14 01 34 Correlation Between the Colposcopic Indices When we evaluated the correlation between the two colposcopic classifications, we found 18 results that were concordant for high-grade suspicion and 9 that were concordant for low-grade suspicion, totaling 27 cases and a correlation value of 79.41%. The calculated Kappa coefficient was 0.55 ( Table 7 ). Table 7 Correlation between the 2011 IFCPC Nomenclature and the Swede score 2011 IFCPC Nomenclature X Swede score High-grade Low-grade Total High-grade 18 03 21 Low-grade 04 09 13 Total 22 12 34 Abbreviation: IFCPC, International Federation for Cervical Pathology and Colposcopy. The positive results of the 2011 IFCPC Nomenclature and Swede score colposcopic classifications were used for the McNemar test. The p-value found was 0.37 ( Table 8 ). Table 8 Matching of concordant results according to the 2011 IFCPC Nomenclature and to the Swede score 2011 IFCPC Nomenclature X Swede score Positive Negative Positive 16 01 Negative 04 09 Abbreviation: IFCPC, International Federation for Cervical Pathology and Colposcopy. Discussion Colposcopy is a subjective exam that is difficult to reproduce because it depends on the skill of the specialist who performs it. To reduce variations in the observations of specialists that could result in making the test less effective, nomenclatures and scores were created that serve to broadly guide the colposcopic reports. Among them are two used by the IFCPC, which is the 2011 IFCPC Nomenclature, and the Swede score. In our study, we determined a correlation value of 79.41% between the 2011 IFCPC Nomenclature and the Swede score, and we obtained a Kappa value of 0.55%, which we interpreted as moderate association. 10 This means that the results were not mostly random, and we can trust the reproducibility of the rankings among experts. With the McNemar test, we found a p-value equal of 0.37; that is, we can conclude that there is no significant difference between the proportions of positive results between the groups and that the samples are comparable and similar. In the literature, we found few studies that assess the 2011 IFCPC Nomenclature. One of them was performed at the Fundan University Hospital in Shanghai by Li et al., 12 who reviewed 525 colposcopies to evaluate them according to the 2011 IFCPC Nomenclature and its histopathological examination for high-grade lesions. 12 13 14 The results obtained were: 64.95% accuracy, 63.64% sensitivity, and 96.01% specificity. Another important study was carried out by Fan et al., 15 which included 2,262 patients whose colposcopic evaluations were reviewed according to the 2011 IFCPC Nomenclature. The accuracy, sensitivity, and specificity values found were 65.5, 71.6, and 98%, respectively. Analyzing the applicability of the 2011 IFCPC Nomenclature alone in predicting high-grade lesions, our study found accuracy, sensitivity, and specificity values of 67.64, 85.71, and 55%. Comparing these figures with the results of Li et al. 12 and Fan et al., 15 our study found very similar totals for accuracy and sensitivity, with more varied results regarding specificity. For the Swede score, we used the article by Strander et al., 9 in which 297 patients were evaluated at the Care Hospital in Eastern Sweden. This study obtained 100% sensitivity with a cutoff of 5, and 90% specificity with a cutoff of 8 to identify CIN 2+ lesions. Another study was carried out by Kushwah et al., 16 in which the Swede score was correlated with that of Reid. This was carried out at the Ghandi Memorial Hospital and recruited 80 patients who were included in the study, in which the performance of the new score was individually evaluated, finding a sensitivity of 100% and a specificity of 91.3% for a score of 5 for high-grade lesions. When using a score of 8, the specificity increased to 100% and the sensitivity dropped to 36.8%. Bowring et al. 11 also evaluated the efficacy of the Swede score in 200 patients who underwent colposcopies at the Royal Hospital in London, Great Britain. In this study, a sensitivity of 38% and a specificity of 95% for high-grade lesions were obtained with a score of 8. When evaluating the use of the Swede score alone to predict high-grade lesions by using a score ≥ 5, our study showed an accuracy of 79.41%. In addition, it demonstrated a sensitivity of 100% and a specificity of 63.15%. When a score of 7 was used, the specificity increased to 94.73%, and with a score of 8, it increased to 100%, while the sensitivity remained the same, at 60%. Comparing our results with the study by Strander et al., 9 we found an association with their results, showing the same 100% sensitivity for the cutoff with a score of 5 and a greater specificity with a score of 8 for invasive disease. The same value of 100% specificity was found in the study by Kushwah et al. 16 Therefore, there is an agreement between the results found in our study and those of other authors. Thus, due to the results found in our study, we can suggest that the Swede score appears to be more effective in identifying patients with CIN 2 + , as it presents results demonstrating greater accuracy, sensitivity, and specificity. Since it demonstrated 100% sensitivity, that is, it identified all patients with a NPV value of 100%, safely excluding tests for low suspicion that could be wrongly excluded from the investigation. Furthermore, it is evident that the reliability of the test increases with a score of 8, since all included cases have a CIN 2+ lesion, which indicates that there was no overtreatment. This increases confidence in utilizing the “see and treat” method, which consists of carrying out the excisional treatment immediately during the diagnostic examination. 17 This measure makes it possible to do away with biopsies and for the patient to have therapeutic treatment performed in a single visit to the clinic. In turn, it would benefit the health system by presenting lower costs, 18 and it would benefit the patient by allowing for a reduction in the period of absence from work and in the anxiety about waiting for the procedure. It would also prevent the delay of treatment, thus avoiding loss of follow-up. Thus, despite the small sample size, the present study is unprecedented due to being the only one to compare these two colposcopic classifications with each other. We still have as limitations the inclusion of low-grade cytology as an indication for colposcopy, which may have reduced the prevalence of the disease, leading to a selection bias; the inclusion of the 2011 IFCPC Nomenclature of the nonspecific classification as a high degree that may have led to overestimating the diagnosis; and the fact that the accuracy of colposcopy may be related to the amount of experience of the colposcopist, which was not taken into account. These are initial data for a larger and more comprehensive study that revealed that the Swede score appears to be superior in identifying high-grade cervical lesions. It is important that studies be carried out to assess the reproducibility of the concepts advocated by the IFCPC in order to improve care for women and to allow for the early diagnosis of precursor lesions of cervical cancer. Conclusion The nonrandom agreement between the colposcopic Swede score and the IFCPC 2011 Nomenclature demonstrated moderate correlation in our study; that is, most of it did not occur by chance. This indicates that there is reproducibility of the two tests among experts. This assessment is important, as it is a subjective examination and depends on the expertise of the person performing it. Using the McNemar test, we concluded that there are no significant differences between the samples and that they can be compared. Assessing each classification individually, we obtained better results in terms of accuracy, sensitivity, and specificity with the Swede score. The greater the sensitivity and the NPV, the lower the possibility of misdiagnosis, making it more interesting as a screening test. Thus, we can suggest in the present study, with the limitations already described, the hypothesis that the Swede score is more effective than the 2011 IFCPC Nomenclature in identifying precursor lesions of cervical cancer. According to the Swede score, the cutoff of 8 for CIN 2+ disease proved to be slightly higher than the score of 7 that is used by the IFCPC, which configures the absence of overtreatment. Contributions Conflict of Interests The authors have no conflict of interests to declare. All authors contributed to the design of the study and were involved in the data collection and data analysis and/or interpretation. Also, all authors contributed to the writing/substantive editing and review of the manuscript and approved the final draft of the manuscript. ==== Refs References 1 Sung H Ferlay J Siegel R L Lauversanne M Soerjomataram I Jemal A Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries CA Cancer J Clin 2021 71 03 209 249 10.3322/caac.21660 33538338 2 Bruni L Diaz M Castellsagué X Ferrer E Bosch F X de Sanjosé S Cervical human papillomavirus prevalence in 5 continents: meta-analysis of 1 million women with normal cytological findings J Infect Dis 2010 202 12 1789 1799 10.1086/657321 21067372 3 Saslow D Solomon D Lawson H W Killackey M Kulasigam S L Cain J M American Cancer Society, American Society for Colposcopy and Cervical Pathology, and American Society for Clinical Pathology screening guidelines for the prevention and early detection of cervical cancer J Low Genit Tract Dis 2012 16 03 175 204 10.1097/LGT.0b013e31824ca9d5 22418039 4 Ministério da Saúde Instituto Nacional de Câncer José Alencar Gomes da Silva Coordenação de Prevenção e Vigilância Divisão de Detecção Precoce e Apoio à Organização de Rede Diretrizes brasileiras para o rastreamento do câncer do colo do útero 2a ed. Rio de Janeiro INCA 2016 5 Schiffman M Wentzensen N Wacholder S Kinney W Gage J C Castle P E Human papillomavirus testing in the prevention of cervical cancer J Natl Cancer Inst 2011 103 05 368 383 10.1093/jnci/djq562 21282563 6 ALTS Group Ferris D G Litaker M Interobserver agreement for colposcopy quality control using digitized colposcopic images during the ALTS trial J Low Genit Tract Dis 2005 9 01 29 35 10.1097/00128360-200501000-00007 15870519 7 Bornstein J Bentley J Bösze P Girardi F Haefner H Menton M 2011 colposcopic terminology of the International Federation for Cervical Pathology and Colposcopy Obstet Gynecol 2012 120 01 166 172 10.1097/AOG.0b013e318254f90c 22914406 8 Quaas J Reich O Frey Tirri B Küppers V Explanation and Use of the Colposcopy Terminology of the IFCPC (International Federation for Cervical Pathology and Colposcopy) Rio 2011 Geburtshilfe Frauenheilkd 2013 73 09 904 907 10.1055/s-0033-1350824 24771940 9 Strander B Ellström-Andersson A Franzén S Milsom I Rådberg T The performance of a new scoring system for colposcopy in detecting high-grade dysplasia in the uterine cervix Acta Obstet Gynecol Scand 2005 84 10 1013 1017 10.1111/j.0001-6349.2005.00895.x 16167921 10 International Federation of Cervical Pathology and Colposcopy Swede Score [Internet] 2020[cited 2020 Nov 13]. Available from:http://www.ifcpcdev.net/newWP/wp-content/uploads/2017/10/Swede-Score-En.docx 11 Bowring J Strander B Young M Evans H Walker P The Swede score: evaluation of a scoring system designed to improve the predictive value of colposcopy J Low Genit Tract Dis 2010 14 04 301 305 10.1097/LGT.0b013e3181d77756 20885156 12 Li Y Duan X Sui L Xu F Xu S Zhang H Closer to a uniform language in colposcopy: study on the potential application of 2011 International Federation for Cervical Pathology and Colposcopy terminology in clinical practice BioMed Res Int 2017 2017 8.984516E6 10.1155/2017/8984516 13 Cooper D B Goyal M Colposcopy Treasure Island StatPearls Publishing 2020[cited 2020 Nov 13]. Available from:https://www.ncbi.nlm.nih.gov/books/NBK564514/ 14 Landis J R Koch G G The measurement of observer agreement for categorical data Biometrics 1977 33 01 159 174 843571 15 Fan A Wang C Zhang L Yan Y Han C Xue F Diagnostic value of the 2011 International federation for cervical pathology and colposcopy terminology in predicting cervical lesions Oncotarget 2018 9 10 9166 9176 10.18632/oncotarget.24074 29507681 16 Kushwah S Kushwah B Correlation of two colposcopic indices for predicting premalignant lesions of cervix J Midlife Health 2017 8 03 118 123 10.4103/jmh.JMH_22_17 28983158 17 Ebisch R M Rovers M M Bosgraaf R P van der Pluijm-Schouten H W Melchers W JG Evidence supporting see-and-treat management of cervical intraepithelial neoplasia: a systematic review and meta-analysis BJOG 2016 123 01 59 66 10.1111/1471-0528.13530 26177672 18 Bedell S L Goldstein L S Goldstein A R Goldstein A T Cervical cancer screening: past, present and future Sex Med Rev 2020 8 01 28 37 10.1016/j.sxmr.2019.09.005 31791846
PMC009xxxxxx/PMC9708395.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36067798 10.1055/s-0042-1751060 220018 Original Article High Risk Pregnacy Is There a Place for Family-centered Cesarean Delivery during Placenta Accreta Spectrum Treatment? Existe um lugar para cesariana centrada na família durante o tratamento do espectro da placenta acreta?http://orcid.org/0000-0001-5639-9127 Nieto-Calvache Albaro José 1 http://orcid.org/0000-0002-7049-2982 Hidalgo Alejandra 2 http://orcid.org/0000-0002-6467-160X Maya Juliana 2 http://orcid.org/0000-0001-7603-8740 Sánchez Beatriz 13 http://orcid.org/0000-0003-1189-1040 Blanco Luisa Fernanda 13 http://orcid.org/0000-0001-6971-2098 Sinisterra-Díaz Stiven Ernesto 4 http://orcid.org/0000-0002-0141-4809 Benavides-Calvache Juan Pablo 1 http://orcid.org/0000-0002-0024-2662 Padilla Iván 5 http://orcid.org/0000-0003-4906-3064 Aldana Ivonne 5 http://orcid.org/0000-0003-3790-7389 Jaramillo Martha 5 http://orcid.org/0000-0002-8703-3360 Gómez Ana Maria 5 http://orcid.org/0000-0003-1555-0948 Castillo Angela María Olarte 6 http://orcid.org/0000-0001-6025-2590 Bryon Adriana Messa 1 1 Clínica de Espectro de Acretismo Placentario, Fundación Valle del Lili, Cali, Colombia 2 Universidad Icesi, Programa de Medicina, Cali, Colombia 3 Departamento de Anestesiología, Fundación Valle del Lili, Cali, Colombia 4 Centro de Investigaciones Clínicas, Fundación Valle del Lili, Cali, Colombia 5 Unidad de Cuidad Intensivo Neonatal, Fundación Valle del Lili, Cali, Colombia 6 Unidad de Cuidado Intensivo Pediátrico, Fundación Valle del Lili, Cali, Colombia Address for correspondence Albaro José Nieto Calvache, MD Fundación Valle Del LiliCali, Carrera 98# 18-49 Cali [email protected] 06 9 2022 10 2022 1 9 2022 44 10 925929 17 1 2022 28 4 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  Placenta accreta spectrum (PAS) is a cause of massive obstetric hemorrhage and maternal mortality. The application of family-centered delivery techniques (FCDTs) during surgery to treat this disease is infrequent. We evaluate the implementation of FCDTs during PAS surgeries. Methods  This was a prospective, descriptive study that included PAS patients undergoing surgical management over a 12-month period. The patients were divided according to whether FCDTs were applied (group 1) or not (group 2), and the clinical outcomes were measured. In addition, hospital anesthesiologists were surveyed to evaluate their opinions regarding the implementation of FCDTs during the surgical management of PAS. Results  Thirteen patients with PAS were included. The implementation of FCDTs during birth was possible in 53.8% of the patients. The presence of a companion during surgery and skin-to-skin contact did not hinder interdisciplinary management in any case. Conclusion  Implementation of FCDTs during PAS care is possible in selected patients at centers with experience in managing this disease. Resumo Objetivo  O espectro da placenta acreta (do inglês placenta accreta spectrum - PAS) é causa de hemorragia obstétrica maciça e mortalidade materna. A aplicação de técnicas de parto centrado na família (do inglês family-centered delivery techniques - FCDTs) durante a cirurgia para tratar esta doença é pouco frequente. Avaliamos a implementação das FCDTs durante as cirurgias do PAS. Métodos  Estudo prospectivo e descritivo que incluiu pacientes com PAS submetidos a tratamento cirúrgico durante um período de 12 meses. Os pacientes foram divididos de acordo com a aplicação de FCDTs (grupo 1) ou não (grupo 2), e os resultados clínicos foram medidos. Além disso, anestesiologistas hospitalares foram entrevistados para avaliar suas opiniões sobre a implementação das FCDTs durante o manejo cirúrgico do PAS. Resultados  Foram incluídos 13 pacientes com PAS. A implementação de FCDTs durante o parto foi possível em 53,8% das pacientes. A presença do acompanhante durante a cirurgia e o contato pele a pele não prejudicou o manejo interdisciplinar em nenhum caso. Conclusão  A implementação de FCDTs durante o atendimento do PAS é possível em pacientes selecionados em centros com experiência no manejo dessa doença. Keywords placenta accreta humanization birthing centers patient-centered care Palavras-chave placenta acreta humanização centros de parto cuidado centrado no paciente ==== Body pmcIntroduction Placenta accreta spectrum (PAS) is a cause of maternal morbidity. 1 Its management may require highly complex interventions, including prolonged surgeries with extensive dissection and massive transfusions. 2 Those conditions make general anesthesia the most widely used anesthetic method for PAS surgery, especially in low- and middle-income countries. 3 4 Although the benefits of family-centered delivery techniques (FCDTs), such as having a companion during the surgical procedure or allowing skin-to-skin contact between neonates and the mother, have been demonstrated when part of a routine cesarean section, 5 6 7 8 their use in PAS surgery is infrequent. In patients affected by PAS, being aware of the possibility of suffering from serious complications (even death) and being subjected to highly complex interventions have been found to be related to posttraumatic stress disorder 9 and to the request for psychological support. 10 We evaluate the implementation of FCDTs during PAS surgeries, as well as the opinions of a group of anesthesiologists about this activity. Methods A prospective descriptive study was conducted between August 2018 and March 2020. We included all patients undergoing PAS surgical management on a scheduled basis, after a prenatal diagnosis was established and interdisciplinary surgery planning at the Fundación Valle de Lili (FVL), Cali, Colombia. Cases with an intraoperative diagnosis of PAS were excluded. Details of the anesthetic and surgical protocols are described elsewhere. 11 12 Details about the FCDTs applied (having a companion during the surgery and skin-to-skin contact between the newborn and the mother) are described in Chart 1 . Chart 1 Details of the presence of a companion in the operating room and skin-to-skin contact In August 2018, the decision was made to favor FCDT among patients with PAS at FVL. Two interventions were proposed to the anesthesiologists and neonatologists: the presence of a companion during the surgery and skin-to-skin contact between the newborn and her mother. The final decision to allow these two strategies in a specific case rested directly with the treating anesthesiologist. The presence of a companion in the operating room was allowed when the surgery was performed under scheduled conditions, in the absence of active vaginal bleeding, and with normal vital signs at the beginning of the surgery. The patient and her companion were educated on the interventions to be performed during the surgery and informed that if the patient's clinical condition deteriorated, it would be necessary for the companion to leave the operating room. In all cases managed with neuraxial anesthesia, at the time of fetal extraction, the newborn was presented to her mother, and skin-to-skin contact was favored when the newborn's condition was stable. Hence, the baby was delivered to the neonatologist who performed the neonatal adaptation on the mother's breast, involving the companion when possible. The companion was removed from the operating room when the newborn was transferred to the recovery room or the neonatal unit. The cases attended in the study period in which these strategies were applied were recorded. The time between birth and breastfeeding, the time between birth and visual or tactile contact between the newborn and mother, and whether the anesthesiologist perceived that the surgical or anesthetic procedure was hampered by the presence of the companion were evaluated. Abbreviations: FCDT, family-centered delivery techniques; FVL, Fundación Valle de Lili; PAS, placenta accreta spectrum. In July 2019, a survey was conducted among all FVL anesthesiologists seeking their opinion about the application of FCDTs during the surgical management of PAS. We divided the patients into those who had skin-to-skin contact with their neonates and/or the presence of a relative during the surgery (group 1) and those who did not (group 2). The clinical results were compared between the two groups, as were the mother-neonate interaction and the anesthesiologist's perception of the impact of FCDTs on the anesthetic procedure. Continuous variables were expressed as central tendency measurements (median) and dispersion measurements (interquartile range [IQR]) based on normal distribution criteria. Categorical variables were expressed as absolute and relative frequencies. The ethics committee of FVL approved the study protocol (protocol no. 1023), and informed consents were obtained from all the study participants. Results During the study period, 16 patients with suspected PAS were treated. Three of them had an intraoperative diagnosis and were excluded, and the remaining 13 patients were included in the analysis. In 7 patients (53.8%), FCDTs were applied (group 1), while the remaining 6 patients underwent the surgery unaccompanied, and skin-to-skin contact was not allowed (group 2). Table 1 presents the opinions of the 28 FVL anesthesiologists about the application of FCDTs during the care of PAS patients Table 2 shows the clinical outcomes of groups 1 and 2. Table 1 Opinions of anesthesiologists on FCDT during the care of patients with PAS ( n  = 28) Variable n (%) Are you familiar with FCDT employed during c-section birth? ● Yes 26 (92.8) What importance do you give to FCDT during birth? * 5 (4–5) ** Do you apply FCDT during elective c-section? ● Yes 26 (92. 8) Do you apply FCDT during emerging c-section? ● Yes 14 (50) What is your preferred type of anesthesia during the management of PAS? ● Neuraxial 27 (96.4) ● General 1 (3.6) Would you allow the presence of a companion during PAS surgery? ● Yes 7 (25) Would you allow skin-to-skin contact during PAS surgery? ● Yes 10 (35.7) What are your reasons for not allowing a companion during the PAS surgery? • High risk of complications 19 (67.8) • The presence of a companion hinders the work of health professionals during the management of complications • Caring for the companion can become one more task for an already busy team • During this surgery, information is handled that can be misinterpreted by the companion 2 (7.1) 3 (10.7) 1 (3.6) Would skin-to-skin contact hinder the proper management of the patient affected by PAS? • Yes 5 (17.8) If you currently do not allow FCDT, in what situation would you allow them in cases of PAS? • Under no condition 17 (60.7) • Only after the severity of the lesion has been determined, and control of the situation has been established 4 (14.3) • Only if clear criteria are established for FCDT implementation, including time limits for each of the interventions 2 (7.1) • Personnel in charge of supervising these interventions (skin-to-skin contact and companion in the room) are present 5 (17.9) Abbreviations: FCDT, family-centered delivery techniques; PAS, placenta accreta spectrum. * score from 1 to 5, with 1 representing “not important” and 5 representing “very important.” ** medium (interquartile range). Table 2 Clinical characteristics of the patients, according to the strategy used for emotional support to the patient during birth Variables Group 1 ( n  = 7) Group 2 ( n  = 6) Gestational age (weeks) * 36 (34–38) 34.5 (33–37) Histological confirmation of PAS. N (%) ** 6 (85.7) 5 (83.3) Intraoperative blood loss (milliliters) * 1,048 (459–1,480) 1,578 (1,053–2,055) RBCU transfusion. N (%) 3 (42.8) 2 (33.3) Hysterectomy. n (%) 4 (57.1) 3 (50) Uterine tamponade. N (%) 2 (28.5) 3 (50) Uterine tourniquet. N (%) 5 (71.4) 5 (83.3) Uterus compression sutures. N (%) 2 (28.5) 3 (50) Aortic occlusion. n (%) 2 (28.5) 3 (50) Uterus S2 involvement. N (%) *** 4 (57.1) 4 (66.6) Neuraxial anesthesia. N (%) 7 (100) 3 (50) Duration of surgery (minutes) * 195 (172–220) 178 (135–233) Neonate weight (grams) * 2,713 (1,846–2,885) 2,564 (2,109–2,968) APGAR at 10 minutes * 8 (8–8) 9 (8–9) Hospitalization of the newborn. N (%) 3 (42.8) 3 (50) Abbreviations: PAS, placenta accreta spectrum; RBCU, red blood cell units. * median (Interquartile Range). ** accreta, increta, or percreta. *** involvement of the cervix, parametria, or lower uterine segment. Table 3 shows the differences in the interaction between mother and neonate between groups 1 and 2. Table 3 Results observed with the strategy used for emotional support of the patient during birth Variable Group 1 ( n  = 7) Group 2 ( n  = 6) Mother-baby skin contact in the operating room. N (%) 7 0 Companion-baby skin contact in the operating room. N (%) 3 0 Presence of a companion during surgery. N (%) 5 0 Time between birth and mother-newborn contact (hours) * 0 6.5 (4–36) Breastfeeding within the first 12 hours of birth. N (%) 3 3 Time between birth and breastfeeding (hours) * 8 (2–44) 28 (4–144) Anesthesiologist report of additional difficulty of anesthetic management due to the implementation of FCDT by the management team. N (%) 0 Does not apply Abbreviation: FCDT: family-centered delivery techniques. * median (interquartile range); Discussion The presence of a companion during surgery and skin-to-skin contact did not hinder anesthetic management in any case ( Table 3 ). Similar to observations in other high-risk obstetric populations, 13 the implementation of FCDTs has difficult challenges to overcome, such as the burden of responsibility for the health of the mother and fetus, the added challenges posed by the presence of a companion, and the fear of legal conflicts about complications during surgery ( Table 1 ). Although almost all of the anesthesiologists surveyed ( n  = 26, 92.8%) stated that they were familiar with FCDTs and applied them during elective cesarean sections ( Table 1 ), only 25% and 35.7% of them reported that they would allow the presence of a companion and skin-to-skin contact during PAS surgery, respectively ( Table 1 ). Furthermore, during the period studied, only 7 of the 13 patients with PAS undergoing surgery benefited from skin-to-skin contact with their neonates. Among those patients, only 5 had a companion present during the surgical procedure. In some PAS reference centers, the frequency of neuraxial anesthesia can be as high as 95%, 14 which allows the mother to be alert, as she anxiously awaits the resolution of a serious disease, and the birth of her child. Although greater maternal satisfaction with the experience of childbirth has been reported when a companion is allowed during the cesarean section, 15 and although benefits are reported in terms of breastfeeding, neonatal temperature regulation, and neonatal cutaneous bacterial colonization when maternal-neonatal skin contact is made within the first hour of life, 5 8 these observations are often ignored during “complicated” cesarean sections. The most common reason mentioned by the anesthesiologists surveyed for not allowing a companion in surgery was the high risk of complications in this type of surgery ( n  = 19, 67.8%). However, none of the 13 patients who were treated on a scheduled basis due to prenatal suspicion of PAS had sudden or uncontrollable massive bleeding. The management of PAS in developing countries faces difficulties much higher in a priority list than the use of FCDT. 16 However, even in pregnancies with serious conditions, it is worth asking whether there is a place for FCDT, especially given that posttraumatic stress disorder is reported in 40% of patients with PAS, 9 which highlights the importance to improve these women's birthing experience , and considering that the presence of a companion during surgery or skin-to-skin contact was not perceived by the anesthesiologist as hindering the normal performance of the activities necessary for PAS treatment among any patients in group 1. Our study has limitations; we described a small number of cases from a single center, so that the external validity of our observations depends on the strategies used in other centers. Additionally, selection bias during candidates' selection for FCDTs cannot be ruled out. However, this study provides an initial observation of the viability of FCDTs during surgeries for selected cases of PAS. Finally, although we assume a benefit of applying FCDTs in terms of psychological impact during the surgical management of PAS, our study does not have the capacity to evaluate these types of implications. We only intend to draw attention to the issue, as multicenter studies are necessary to evaluate the receptivity of PAS teams to this type of intervention as well as the possible benefits for patients, their neonates, and their families. Conclusion Applying FCDTs during PAS care is possible in selected patients at centers with experience in managing this disease. Contributions Conflict of Interests The authors have no conflict of interests to declare. All authors contributed to the design of the study, were involved in the data collection, data analysis and/or interpretation. Also, all authors contributed to the writing/substantive editing and review of the manuscript and approved the final draft of the manuscript. ==== Refs References 1 Maternal Death Exploratory Committee in Japan and the Japan Association of Obstetricians and Gynecologists Hasegawa J Tanaka H Katsuragi S Sekizawa A Ishiwata I Ikeda T Maternal deaths in Japan due to abnormally invasive placenta Int J Gynaecol Obstet 2018 140 03 375 376 10.1002/ijgo.12368 29080305 2 Stafford I Belfort M Placenta accreta, increta, and percreta: lifesaving strategies to stop the bleeding Contemp Ob Gyn 2008 53 05 48 53 3 Muñoz L A Mendoza G J Gomez M Reyes L E Arevalo J J Anesthetic management of placenta accreta in a low-resource setting: a case series Int J Obstet Anesth 2015 24 04 329 334 10.1016/j.ijoa.2015.05.005 26343175 4 Urfalıoglu A Öksüz G Bilal B Retrospective evaluation of anesthetic management in cesarean sections of pregnant women with placental anomaly Anesthesiol Res Pract 2020 2020 1.358258E6 10.1155/2020/1358258 5 Dominguez-Bello M G Costello E K Contreras M Delivery mode shapes the acquisition and structure of the initial microbiota across multiple body habitats in newborns Proc Natl Acad Sci U S A 2010 107 26 11971 11975 10.1073/pnas.1002601107 20566857 6 Hung K J Berg O Early skin-to-skin after cesarean to improve breastfeeding MCN Am J Matern Child Nurs 2011 36 05 318 324, quiz 325–32610.1097/NMC.0b013e3182266314 21743355 7 Guala A Boscardini L Visentin R Skin-to-skin contact in cesarean birth and duration of breastfeeding: a cohort study Scientific World Journal 2017 2017 1.940756E6 10.1155/2017/1940756 8 Moore E R Bergman N Anderson G C Medley N Early skin-to-skin contact for mothers and their healthy newborn infants Cochrane Database Syst Rev 2016 11 11 CD003519 10.1002/14651858.CD003519.pub4 27885658 9 Tol I D Yousif M Collins S L Post traumatic stress disorder (PTSD): The psychological sequelae of abnormally invasive placenta (AIP) Placenta 2019 81 42 45 10.1016/j.placenta.2019.04.004 31138430 10 Welz J Keyver-Paik M D Gembruch U Merz W M Self-reported physical, mental, and reproductive sequelae after treatment of abnormally invasive placenta: a single-center observational study Arch Gynecol Obstet 2019 300 01 95 101 10.1007/s00404-019-05175-z 31069488 11 López-Erazo L J Sánchez B Blanco L F Nieto-Calvache A J Placenta accreta spectrum anaesthetic management with neuraxial technique can be facilitated by multidisciplinary groups Indian J Anaesth 2021 65 02 153 156 10.4103/ija.IJA_1216_20 33776091 12 Nieto-Calvache A J Vergara-Galliadi L M Rodríguez F A multidisciplinary approach and implementation of a specialized hemorrhage control team improves outcomes for placenta accreta spectrum J Trauma Acute Care Surg 2021 90 05 807 816 10.1097/TA.0000000000003090 33496549 13 Behruzi R Hatem M Goulet L Fraser W Leduc N Misago C Humanized birth in high risk pregnancy: barriers and facilitating factors Med Health Care Philos 2010 13 01 49 58 10.1007/s11019-009-9220-0 19669934 14 Markley J C Farber M K Perlman N C Carusi D A Neuraxial anesthesia during cesarean delivery for placenta previa with suspected morbidly adherent placenta: a retrospective analysis Anesth Analg 2018 127 04 930 938 10.1213/ANE.0000000000003314 29481427 15 Tessier España E Camaño Gutiérrez I García Burguillo A Cesárea humanizada Prog Obstet Ginecol. 2013 56 02 73 78 16 Nieto-Calvache A J López-Girón M C Nieto-Calvache A Messa-Bryon A Benavides-Calvache J P Burgos-Luna J M A nationwide survey of centers with multidisciplinary teams for placenta accreta patient care in Colombia, observational study J Matern Fetal Neonatal Med 2022 35 12 2331 2337 10.1080/14767058.2020.1786052 32627610
PMC009xxxxxx/PMC9708396.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36446565 10.1055/s-0042-1755340 220203 Letter to the Editor Adequacy of Antenatal Care during the COVID-19 Pandemic: Observational Study with Postpartum Women Adequação da assistência pré-natal durante a pandemia de COVID-19: Estudo observacional com puérperashttp://orcid.org/0000-0003-0000-0740 Rosa Vanessa Martins 1 http://orcid.org/0000-0001-6061-2541 Iser Betine Pinto Moehlecke 1 1 Universidade do Sul de Santa Catarina, Programa de Pós-Graduação em Ciências da Saúde, Palhoça, SC, Brazil Address for correspondence Betine Pinto Moehlecke Iser, PhD Avenida José Acácio Moreira787, 88704-900, Dehon, Tubarão, Santa [email protected] 29 11 2022 10 2022 1 11 2022 44 10 995998 24 6 2022 29 6 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. ==== Body pmcDear Editor, We enthusiastically read the analysis conducted in the study Adequacy of Prenatal Care During The COVID-19 Pandemic: An Observational Study with Postpartum Women, 1 published in volume 44 issue #4 of the Brazilian Journal of Gynecology and Obstetrics in 2022, and would like to bring some considerations about the topic. According to the World Health Organization (WHO), investing in prenatal care is essential to improve one of the most important indicators of quality of life for the world population, which is maternal mortality, in which Brazil is below the rate stipulated as ideal. 2 The state of Santa Catarina and the South region, according to national data, stand out for the lowest rates of maternal mortality, better prenatal care coverage, and quality of prenatal care. 3 4 However, the few current studies analyzing the quality of prenatal care in this region differ from the data made available by the Ministry of Health (MH). 1 5 Thus, it is pertinent to emphasize that this study has great value by making a careful analysis of this assistance. As the authors point out in the limitations of the study, the diversity of criteria used to evaluate the quality of prenatal care makes a more complete and comparative analysis of the data difficult. 1 The Brazilian government uses criteria with few variables for prenatal care adequacy, using the Kotelchuck index, 6 which takes into account only the date of prenatal care initiation and the number of appointments. Thus, the data found in national studies differ from the MH database. Between the years 2012 and 2013, Tomasi et al. 7 found in a national survey with 50,791 participants only 21% adequacy to prenatal care, while for the MH, in the year 2014, the national adequacy was 63%. Therefore, the development of robust studies that confront the findings and justify such divergence becomes pertinent. Most studies analyze how well the current MH guidelines 8 are being met, while others analyze national and even international indexes. However, there are no uniform and well-accepted criteria for the classification of the adequacy of prenatal care; besides, there is no uniformity in the way data is collected, resulting in values with large variations and making it difficult to compare the populations studied. 9 As an example, we can cite the Northeast region of Brazil, which, according to the WHO, has a prenatal care quality index lower than the national average; 10 however, in a study conducted in Paraiba in 2018, a good quality was verified with > 80% of prenatal visits starting at the ideal time and with an adequate number of appointments, and having > 90% of the necessary exams performed. 11 According to the study by Martin et al., 1 in the year 2020, amid the COVID-19 pandemic, 35.8% of prenatal care were considered adequate, 46.8% intermediate, and 17.4% inadequate. In the previous year, according to the MH, 79.3% of prenatal care for Southern Brazil were considered adequate. This important difference is probably not only related to the pandemic, since other studies out of the pandemic period have even lower values of adequacy. 5 12 Weaknesses in assistance were found in a study from the most populous city of Santa Catarina: only 17% of the patients received proper guidance during pregnancy and less than half of the patients underwent 3 rd trimester exams (42.5%). 5 Regarding the study in question, the authors claim to use not very strict parameters to assess the quality of prenatal care, such as the week of the beginning of prenatal care being appropriate under 16 weeks; however, since 2012, the MH recommends starting before 12 weeks. 8 The procedures recommended as adequate were the performance of three tests: syphilis, HIV, and urinary tract infection in the 3 rd trimester. Additionally, since 2012, the following are recommended as essential for a quality prenatal care: immunization for hepatitis B and Diphtheria and tetanus vaccine (DT), and performance of. obstetric clinical procedures such as weight, blood pressure, edema, body mass index, uterine height, fetal heartbeat, fetal movements, fetal presentation in the appointments conducted, and laboratory tests for each trimester. In addition to the prescription of ferrous sulfate and folic acid supplements and guidance related to the prenatal period, delivery, puerperium, breastfeeding, newborn care, healthy habits, and emotional and body modifications. 8 It is assumed, therefore, that the percentage of adequacy in this study was overestimated by considering only the minimum necessary, and the impact of the pandemic on prenatal care must have been even greater than estimated. As long as studies are conducted in a nonuniform manner, using different assessment approaches and not meeting the recommendations of the MH, 8 we will not know exactly how the healthcare of pregnant women is at the regional and national levels. Thus, a complete analysis would require the use of the different criteria available in order to allow the applicability of the results in the development of assertive improvement policies. Author's Reply Resposta do autor Margot Marie Martin 1 Roxana Knobel 2 Vitor Nandi 1 Jessica Goedert Pereira 2 Alberto Trapani Junior 2 Carla Betina Andreucci 3,4 1 Hospital Universitário Polydoro Ernani de São Thiago, Florianópolis, SC, Brazil 2 Universidade Federal de Santa Catarina, Florianópolis, SC, Brazil 3 Hospital Regional Homero de Miranda Gomes, São José, SC, Brazil 4 Universidade Federal de São Carlos, São Carlos, SP, Brazil Address for correspondence Roxana Knobel, Departamento de Ginecologia e Obstetrícia – Bloco pedagógico – 2° andar Hospital Universitário- Rua Professora Maria Flora, Pausewang, s/n°, 88036-800, Campus Universitário, Trindade, Florianópolis, Santa Catarina, Brazil (e-mail: [email protected]). Dear Authors, We appreciate the feedback on our article “Adequacy of Antenatal Care during the COVID-19 Pandemic: Observational Study with Postpartum Women”. 1 As you pointed out, the method applied may underestimate the percentage of people with access to adequate antenatal care, mainly because quantitative criteria were applied, and as mentioned, in a quite permissive way. However, the adopted index has a high positive predictive value on adequate antenatal care access and good levels of adjusted agreement with other antenatal assessment indices. 2 The same method was also used in previous publications, 3 4 allowing comparisons. We agree that qualitative and more in-depth assessments are needed, but quantitative assessment may also reveal inequities and differences in outcomes. 5 In our study, for example, we found barriers to black or brown skin color women in receiving adequate antenatal care, as already described in the literature, 3 6 a situation that was aggravated by the pandemic. 7 Modifying the beginning of antenatal care to ≤ 12 weeks (without modifying further criteria) in our sample, the number of postpartum people with adequate antenatal care drops from 91 to 67. Despite the highlighted limitations, our study demonstrates a decrease in access and quality of antenatal care in 2020. Instead of the decrease, an increase in adequate coverage would be expected using less rigid criteria for antenatal adequacy. Due to scheduling difficulties, many users stopped seeking medical care or going to routine appointments in 2020, 8 as shown in our study. Virtual consultations were also a risk factor for inadequate antenatal care, probably because of their abrupt start during the pandemic, without former training or preparation. 9 Additionally, our data showed how social isolation impacted pregnant women during the 1 st year of the pandemic, when no vaccines were available, and how none of the investigated social isolation variables were related to inadequate antenatal care. 1 We agree with the need to standardize antenatal adequacy criteria or a series of criteria suitable for each research objective, allowing better comparisons among studies, different locations, and institutions. We also agree that qualitative approach studies are needed to assess antenatal care quality. However, we believe that this should not prevent researchers from carrying out quantitative studies. We consider that even the official public data available on the Information System on Live Births (SINASC, in the Portuguese acronym) of the Information Technology Department of the Brazilian Unified Public Healthcare System (DATASUS, in the Portuguese acronym) 10 may help to identify barriers to quality antenatal care access in the Brazilian population, leading to public policies to optimize obstetric care. Conflict of Interests The authors have no conflict of interests to declare. References 1 Martin MM, Knobel R, Nandi V, Pereira JG, Trapani Junior A, Andreucci CB. Adequacy of antenatal care during the COVID-19 pandemic: observational study with postpartum women. Rev Bras Ginecol Obstet. 2022;44(04):398–408. Doi: 10.1055/s-0041-1741450 2 Tomasi E, Fernandes PA, Fischer T, Siqueira FCV, Silveira DS, Thumé E, et al. Qualidade da atenção pré-natal na rede básica de saúde do Brasil: indicadores e desigualdades sociais. Cad Saude Publica. 2017;33(03):e00195815. Doi: 10.1590/0102-311 × 00195815 3 Melo EC, Oliveira RR, Mathias TA. [Factors associated with the quality of prenatal care: an approach to premature birth]. Rev Esc Enferm USP. 2015;49(04):540–549. Doi: 10.1590/S0080-623420150000400002 4 Nunes JT, Gomes KR, Rodrigues MT, Mascarenhas MD. Qualidade da assistência pré-natal no Brasil: revisão de artigos publicados de 2005 a 2015. Cad Saude Colet. 2016;24(02):252–261. Doi: 10.1590/1414-462 × 201600020171 5 Rodrigues KMD, Zoldan C, Silva CBO, Santana EFM, Araujo Júnior E, Peixoto AB. Relationship between the number of prenatal care visits and the occurrence of adverse perinatal outcomes. Rev Assoc Med Bras. 2022;68(02):256–260. Doi: 10.1590/1806-9282.20211239 6 Leal MDC, Gama SGND, Pereira APE, Pacheco VE, Carmo CND, Santos RV. The color of pain: racial iniquities in prenatal care and childbirth in Brazil. Cad Saude Publica. 2017;33(33, Suppl 1)e00078816. Doi: 10.1590/0102-311 × 00078816 7 Gur RE, White LK, Waller R, Barzilay R, Moore TM, Kornfield S, et al. The disproportionate burden of the covid-19 pandemic among pregnant black women. Psychiatry Res. 2020;293:113475. Doi: 10.1016/j.psychres.2020.113475 8 Horta BL, Silveira MF, Barros AJD, Hartwig FP, Dias MS, Menezes AMB, et al. COVID-19 and outpatient care: a nationwide household survey. Cad Saude Publica. 2022;38(04):e00194121. Doi: 10.1590/0102-311 × 00194121 9 Caetano R, Silva AB, Guedes ACCM, Paiva CCN, Ribeiro GR, et al. Challenges and opportunities for telehealth during the COVID-19 pandemic: ideas on spaces and initiatives in the Brazilian context. Cad Saude Publica. 2020;36(05):e00088920. Doi: 10.1590/0102-311 × 00088920 10 Ministério da Saúde. Secretaria de Vigilância em Saúde. Departamento de Análise de Situação de Saúde (MS/SVS/DASIS) Nascidos vivos por ano e adequação quantitativa do pré-natal, por região do Brasil. Sistema de informações sobre Nascidos Vivos (SINASC) [Internet]. 2022 [cited 2022 Jul 4]. Available from: http://tabnet.datasus.gov.br/cgi/deftohtm.exe?sinasc/cnv/nvuf.def Conflict of Interests The authors have no conflict of interests to declare. ==== Refs References 1 Martin M M Knobel R Nandi V Pereira J G Trapani Junior A Andreucci C B Rev. Bras. Ginecol. Obstet. 2022 44 398, 10.1055/s-0041-174145035176779 2 Morón-Duarte L S Varela A R Bertoldi A D Domingues M R Wehrmeister F C Silveira M F BMC Health Serv. Res. 2021 21 1070, 10.1186/s12913-021-07053-434627235 3 Ministério da Saúde, Secretaria de Vigilância em SaúdeCoordenação Geral de Informações e Análises Epidemiológicas - Sistema de Informações sobre Mortalidade – SIM. Óbitos maternos por ano, por região do Brasil [Internet]. 2022 [cited 2022 Jun 16]. Available from:http://tabnet.datasus.gov.br/cgi/tabcgi.exe?sim/cnv/mat10uf.def 4 Ministério da Saúde, Secretaria de Vigilância em SaúdeDepartamento de Análise de Situação de Saúde (MS/SVS/DASIS). Nascidos vivos por ano e adequação quantitativa do pré-natal, por região do Brasil. Sistema de informações sobre Nascidos Vivos (SINASC) [Internet]. 2022 [cited 2022 Jun 16]. Available from:http://tabnet.datasus.gov.br/cgi/deftohtm.exe?sinasc/cnv/nvuf.def 5 Vaichulonis C G Silva R R Pinto A I Cruz I R Mazzetti A C Haritsch L Avaliação da assistência pré-natal segundo indicadores do Programa de Humanização no Pré-natal e Nascimento Rev Bras Mater Infant 2021 21 02 451 50. 10.1590/1806-93042021000200006 6 Kotelchuck M Am. J. Public Health. 1994 84 1414; 10.2105/ajph.84.9.14148092364 7 Tomasi E Fernandes P A Fischer T Siqueira F C Silveira D S Thumé E Duro S M Saes M O Nunes B P Fassa A G Facchini L A Cad. Saude Publica. 2017 33 e00195815; doi: 10.1590/0102-311 × 0019581528380149 8 Ministério da Saúde Secretaria de Atenção à SaúdeDepartamento de Atenção Básica. Atenção ao pré-natal de baixo risco.Brasília (DF) Ministério da Saúde 2012 9 Cruz G C Ruiz P C Ribeiro Junior O C Sousa A D Pereira R M Barroso C O Métodos de avaliação da qualidade de assistência ao pré-natal no Brasil: revisão integrativa da literatura Rev Eletrônica Acervo Saúde 2019 27 e521. 10.25248/reas.e521.2019 10 Leal M C Esteves-Pereira A P Viellas E F Domingues R M Gama S G Rev. Saude Publica. 2020 54 8; 10.11606/s1518-8787.20200540014531967277 11 Pedraza D F Gomes A A Rev. Cienc. Salud. 2021 19 55; 10.12804/revistas.urosario.edu.co/revsalud/a.10600 12 Pilau G M Volpato L K Nunes R D Arq Catarin Med. 2014 43 9
PMC009xxxxxx/PMC9708397.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36446557 10.1055/s-0042-1756147 rbgo-20-0027 Original Article Obstetrics Nutritional Status of Iodine in a Group of Pregnant Women from the State of Minas Gerais Correlated with Neonatal Thyroid Function Status nutricional de iodo em um grupo de gestantes mineiras correlacionado com função tireoidiana neonatalhttp://orcid.org/0000-0003-1099-7177 Scherr Natália Campos Gonçalves 1 http://orcid.org/0000-0002-4278-7783 Nogueira Anelise Impelizieri 1 http://orcid.org/0000-0003-3463-0951 Rajão Kamilla Maria Araújo Brandão 1 http://orcid.org/0000-0003-2806-8560 Leite Henrique Vitor 1 1 Departamento de Ginecologia e Obstetrícia, Universidade Federal de Minas Gerais, Belo Horizonte, MG, Brazil Address for correspondence Natália Campos Gonçalves Scherr, MSc Av. Prof. Alfredo Balena, Universidade Federal de Minas Gerais190, 30130-100, Belo Horizonte, [email protected] 29 11 2022 10 2022 1 11 2022 44 10 909914 17 8 2020 17 12 2020 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  To evaluate the iodine sufficiency of pregnant women assisted in a University Hospital of Minas Gerais, and to correlate the urinary concentrations of maternal iodine with the fetal thyroid hormone levels at birth. Methods  Urinary iodine concentrations from 30 pregnant women with a singleton pregnancy and gestational age lower than 20 weeks were analyzed. Occasional samples of the mothers' urine were collected for the urinary iodine concentration dosage, and these were correlated with the newborns' thyroid-stimulating hormone (TSH) levels. Results  The median iodine urinary concentration of this study's pregnant women population was 216.73 mcg/l, which is proper for the group, following the World Health Organization (WHO). No cases of neonatal hypothyroidism were reported in the study, which corroborates the iodine sufficiency in this population sample. Conclusion  This study shows that despite the increased demand for iodine from pregnant women and the Brazilian Health Regulatory Agency (ANVISA) recommendation of 2013 for reduction of salt iodization levels, the population of pregnant women attended in the prenatal ambulatory of normal risk from the Federal University of Minas Gerais is considered sufficient in iodine. As a higher sample is necessary for the confirmation of these findings, it is too early to recommend the universal supplementation of iodine for Brazilian pregnant women, and more studies must be carried out, considering that iodine supplementation for pregnant women in an area of iodine sufficiency is associated to the risks of the fetus's excessive exposure to iodine. Resumo Objetivo  Avaliar a suficiência iódica de gestantes atendidas em um Hospital Universitário de Minas Gerais e correlacionar as concentrações urinárias de iodo materno com os níveis de hormônios tireoidianos fetais ao nascimento. Métodos  Foi analisada a concentração urinária de iodo de 30 gestantes com gestação única e idade gestacional menor que 20 semanas. Foram coletadas amostras ocasionais de urina materna para dosagem da concentração urinária de iodo, e estas foram correlacionadas com os níveis de hormônio estimulante da tiroide (TSH) dos recém-nascidos. Resultados  A mediana da concentração urinária de iodo das gestantes estudadas foi de 216,73 mcg/L, sendo adequada para o grupo segundo a Organização Mundial de Saúde (OMS). Não houve nenhum caso de hipotireoidismo neonatal no estudo, o que corrobora a suficiência de iodo nesta amostra populacional. Conclusão  Esse estudo demonstra que apesar do aumento da demanda de iodo pelas gestantes e da recomendação da Agência Nacional de Vigilância Sanitária (ANVISA) de 2013 da redução dos níveis de iodação do sal, a população de gestantes atendidas no ambulatório de pré-natal de risco habitual da Universidade Federal de Minas Gerais é considerada suficiente em iodo. Apesar de uma maior amostragem ser necessária para a confirmação destes achados, é cedo para recomendar a suplementação universal de iodo para as gestantes brasileiras e mais estudos precisam ser realizados, levando-se em conta que a suplementação de iodo para gestantes em áreas suficientes em iodo está associada aos riscos da exposição excessiva de iodo ao feto. Keywords urinary iodine iodine sufficiency iodine pregnancy Palavras-chave iodúria suficiência iódica iodo gestação ==== Body pmcIntroduction Thyroid dysfunctions during pregnancy result in higher maternal and fetal morbidity. One of the causes of thyroid dysfunction is the deficiency or excess of iodine. During pregnancy, there is an increase in the production of thyroid hormones, a significant increment of kidney clearance, and an increase of the (T4) thyroxine turnover through placental type 3 deiodinase in the second half of the pregnancy, thus occasioning an increased need for iodine. 1 2 Over the last years, international consensus on thyroid diseases during pregnancy have recommending the iodine supplementation for pregnant women. These recommendations appeared based on studies that showed iodine insufficiency in pregnant women originated from populations with proper iodine intake. 3 4 5 However, there is a lack of Brazilian studies on this subject. Iodine deficiency has multiple effects on the growth and development of animals and humans. Studies in animal models show that iodine deficiency in mothers and, consequently, the lack of thyroid hormones causes goiter, cortical, and cerebellar changes in newborns. 6 However, the excess of iodine may also lead to an increase in the incidence of hypothyroidism and hyperthyroidism in susceptible subjects. 3 7 Despite iodine being an essential nutrient to produce thyroid hormones, the excess of this substance inhibits three processes in the production of thyroid hormones: iodine capture, iodization of thyroglobulin (Wolff-Chaikoff effect), and the release of thyroid hormones from the gland. This way, the excess of iodine causes a transitory inhibition of the thyroid hormones' synthesis. However, from 2 to 4 weeks of continued exposure to iodine excess, organification and the synthesis of thyroid hormone return to normal, probably due to a decrease in the NIS activity, the sodium-iodide symporter, a carrier protein of iodine that captures iodine to the interior of the thyroid cell. 8 Some subjects cannot escape from this effect, and the iodine excess triggers the hypothyroidism. 9 Fetuses are particularly susceptible, as this escape ability is not totally mature until 36 weeks of pregnancy. 4 The best parameter to evaluate the nutritional status of the population is the urinary iodine concentration. 10 The absorption of iodide in healthy adults is higher than 90%, and more than 90% of the ingested iodine is released in the urine. 11 The thyroid-stimulating hormone (TSH) in newborns is also a great indicator of iodine deficiency. The thyroid in newborns has a low input of iodine compared to adults, so the iodine turnover is much larger. This high turnover is exaggerated in iodine deficiency and requires a high stimulation from the TSH. 10 11 12 The increase in the number of newborns with moderate elevation of TSH concentrations (above 5 mIU/l) is proportional to the level of iodine deficiency during pregnancy. A frequency of < 3% of neonatal TSH values > 5 mIU/l indicates iodine sufficiency in a population. 13 We carried out this study to evaluate the nutritional status of iodine in a sample of pregnant women from the state of Minas Gerais, Brazil, correlating these values with the thyroid function of newborns. Methods This was an observational study in a population of healthy pregnant women, all of whom attended for a first elective appointment in the prenatal ambulatory of normal risk from the Federal University of Minas Gerais in the period from May to December 2017. These pregnant women were invited to participate in the study, informed on its purpose, and received an informed consent term, in compliance with the recommendations of the Declaration of Helsinki. Pregnant women with a singleton pregnancy and gestational age of up to 20 weeks were included. Patients with a history of thyroid disease and/or using multivitamins containing iodine or food supplements were excluded. The patients answered questionnaires regarding age, gestational age, parity, use of multivitamins and food supplements, use of drugs, smoking habits, alcoholism, and family history of thyroid diseases. Pregnant women were also subjected to height and weight measurements. Initially, 39 patients were recruited; however, 9 of them did not present the inclusion criteria of the study. Therefore, 30 patients were enrolled. Generally, according to World Health Organization (WHO), thirty urinary iodine determinations in a defined sampling group are sufficient for the interpretation about iodine sufficiency. 10 During the recruitment, a single sample of urine was collected in the afternoon for urinary iodine dosage. The samples were stored in a freezer at −20° C until the dosage time. It is known that freezing does not interfere with sample stability. 14 The urinary iodine concentration measurement (in mcg/L) of occasional aliquots is considered as the real laboratory biochemical parameter for iodine use by the community. 15 Urinary iodine concentration was determined using inductively coupled plasma mass spectrometry. The manufacturer's reference interval was from 70 to 530 μg/g of creatinine. The iodine status in pregnant women at population level was classified according to the WHO, UNICEF, and ICCIDD guidelines from 2007; with the median urinary iodine measures being < 150 mcg/l considered as insufficient iodine intake; from 150 to 240 mcg/l as adequate iodine intake; from 250 to 249 mcg/k as more than adequate; and ≥ 500 mcg/l as excessive intake. 10 The results of neonatal TSH measurements were provided by the Núcleo de Ação e Pesquisa em Apoio Diagnóstico (NUPAD). The quantitative determinations of TSH were made in blood samples dried in filter paper, using the automatic system of fluoroimmunoassay 1235 AutoDELFIA (PerkinElmer Inc., Waltham, MA, USA). The cut-off values recommended by the manufacturer are: < 9 µIU/ml for normal values; from 9 to 18 µIU/ml for borderline; and  > 18 µIU/ml for hypothyroidism. The statistical analyses were made using the R (R Foundation for Statistical Computing, Vienna, Austria) software, version 3.6.1. In the descriptive analysis of categorical variables (qualitative variables), the absolute and relative frequencies were used. In the description of numerical variables, measurements of position, central trend, and dispersion were used. Initially, to relate the urinary iodine concentration of pregnant women with the variables, a univariate analysis was made. The Mann Whitney test was used for the categorical variables with two levels, ; while the Kruskal-Wallis test was used for the categorical variables with three or more levels, and the Nemenyi test for multiple comparisons. To relate the urinary iodine concentration with the numerical variables, the Spearman correlation was used. 16 The Stepwise method was implemented for the selection of variables for multivariate analysis. The Stepwise method is a mix of the Backward and Forward methods. Thus, first, using the Forward method, the variables with p -value < 0.25 were selected for multivariate analysis. Then, the Backward method was implemented, by removing the variable with the highest p -value, and repeating the procedure until only significant variables remain in the model. For the Backward method, a 5% significance level was adopted. 17 In the multivariate analysis, linear regression was used. Results A total of 30 pregnant women with a singleton pregnancy and gestational age ranging from 4 weeks and 5 days to 19 weeks and 3 days were evaluated. The average age among them was 27.03 years old, with the minimum age being 18 and the maximum 42 years. Among them, 13 pregnant women (43.33%) were overweight, 8 (26.67%) had an adequate weight, 6 (20%) were obese, and 3 (10%) were underweight. The weight classification of pregnant women was established through the Technical Standard from the Brazilian Food and Nutrition Surveillance System (SISVAN). 18 The group counted with 15 nulliparous women, 11 primiparas, and 4 multiparas. The median of urinary iodine concentration samples was 216.73 mcg/l, which is proper for the group, according to WHO. 10 The urinary iodine average (mcg/l) was 247.28 with 95% confidence interval (95% CI = 192.26; 302.29) and standard deviation (SD: 153.35); the minimum value was 75.06, and the maximum was 714.05. In the context of the entire population of pregnant women attended in the prenatal ambulatory of normal risk from the Federal University of Minas Gerais from May to December 2017 (82 patients), for a sample of 30 pregnant women, an error margin of 43 mcg/l would be expected. The neonatal TSH average (µIU/ml) was 1.4, with the minimum value being 0.29 and the maximum 2.84. The collected data are described in Table 1 . Table 1 Descriptive analysis of numeric variables Variable n Average Minimum 1 st Q 2 nd Q 3 rd Q Maximum Urinary iodine (mcg/L) 30 247.28 75.06 127.65 216.73 309.79 714.02 Neonatal TSH 29 1.4 0.29 0.96 1.22 1.83 2.84 BMI (kg/m 2 ) 30 26.75 18.4 23.01 26.82 30.22 36.71 Gestational age (days) 30 86.63 33.0 67.0 86.5 108.0 136.0 Age 30 27.03 18.0 22.0 24.5 32.0 42.0 Abbreviations: BMI, body mass index; TSH, thyroid-stimulating hormone. Table 2 shows the relationship between the urinary iodine concentration of pregnant women and the categorical variables. Thus, it is observed that there was a significant difference ( p  = 0.021) between the urinary iodine concentration of pregnant women and family history of thyroid disease. The average urinary iodine concentration of pregnant women with a family history of thyroid disease was higher when compared with those without ( Table 3 ). Table 2 Relation between the urinary iodine concentration of pregnant women and the categorical variables Variable n Average S.E. 1 st Q 2 nd Q 3 rd Q p -value Smoking 29 no 1 yes 250.3 159.8 28.81 - 127.7 159.8 230.3 159.8 309.8 159.8 0.729 Family history for thyroid disease 23 without 7 with 202.2 395.3 19.94 81.76 126.1 280.9 164.7 292.8 278.0 524.8 0.021 Ex-smoker 23 no 6 yes 236.6 302.9 29.83 82.2 127.2 230.3 166.6 268.1 312.7 296.6 0.572 Ex-drinker 13 no 17 yes 221.2 267.2 28.94 44.38 126.8 148.6 164.7 230.3 315.7 292.8 0.786 Parity 15 nulliparous 15 not nulliparous 253.2 241.4 40.57 39.95 137.0 143.7 230.3 203.1 332.9 292.3 0.868 BMI classification 8 adequate 3 underweight 19 overweight 151.8 207.5 293.8 16.66 54.17 39.37 124.3 153.5 156.6 140.2 159.8 291.7 184.9 237.7 352.8 0.071 Abbreviations: BMI, body mass index. Notes:1 Mann Whitney test; 2 Kruskal-Wallis test. Table 3 Relation between the urinary iodine concentration of pregnant women and the numerical variables Variable/ioduria (mcg/L) r 1 p -value 2 Neonatal TSH (µIU/ml) -0.37 0.051 BMI (Kg/m 2 ) 0.08 0.676 Gestational age 0.2 0.287 Age -0.11 0.547 Abbreviations: BMI, body mass index; TSH, thyroid-stimulating hormone. Notes:1 Correlation coefficient; 2 Spearman correlation. The existence of a marginally significant ( p  = 0.051) and negative (r = -0.37) relationship between the urinary iodine concentration of pregnant women and neonatal TSH was observed ( Table 3 ). In the multivariate analysis, it may be concluded that there was a significant influence ( p  = 0.023) of the BMI classification over the urinary iodine concentration, and when compared with pregnant women with adequate BMI, those with overweight or obesity BMI showed an increase of 82.74 units in the urinary iodine concentration ( Table 4 ). Table 4 Factors that influence the urinary iodine concentration of pregnant women Variables Initial Model Final Model β S.E. (β) p -value 1 β S.E. (β) p -value 1 Family history for thyroid disease = No 0.00 − − 0.00 − − Family history for thyroid disease = Yes 215.58 81.02 0.014 160.73 81.659 0.060 BMI classification = proper 0.00 − − 0.00 − − BMI classification = underweight 37.82 52.87 0.481 55.75 46.897 0.245 BMI classification = overweight 82.74 28.89 0.009 82.74 34.365 0.023 Neonatal TSH -43.48 35.11 0.228 − − − VIF 1.223 1.214 R 2 39.04% 26.56% Abbreviations: BMI, body mass index; TSH, thyroid-stimulating hormone. Notes:1 Solid linear regression. Discussion In several countries that have already adopted salt iodination programs, despite normal levels of urinary concentration in schoolchildren, an iodine insufficiency in pregnant women has been shown. 19 20 Studies with American pregnant women also found iodine insufficiency in this subpopulation. 14 In London, Manchester, and Leeds, the median urinary iodine concentration in pregnant women was also below 150 mcg/l, showing the insufficiency of iodine intake for this particular population, even though the United Kingdom's population is generally considered iodine sufficient. 21 This way, the last guidelines from the American Thyroid Association, Endocrine Society, and European Thyroid Association, recommend the administration of iodine supplements to pregnant women and those planning to get pregnant. 3 4 5 Brazilian data on iodine sufficiency in pregnant women are controversial. In 2001, Barca et al. 15 showed iodine sufficiency in 20 pregnant women attended in a public hospital of São Paulo, with iodine excretion represented by the median of urinary concentration of 167.8 mcg/l. In 2008, Soares et al. 22 evaluated 147 pregnant women in Porto Alegre and obtained urinary iodine levels between 22 and 534 mcg/l, with 19.6% of the pregnant women showing urinary iodine concentration lower than 150 mcg/l, and median of 224 mcg/l, that is, adequate levels of urinary iodine concentration. 22 In 2014, in the state of São Paulo, a high rate of iodine insufficiency was observed in pregnant women (median: 137.7 mcg/l). Among the 191 pregnant women, 57% showed urinary iodine concentration lower than recommended by WHO. 23 It is important to emphasize that these three Brazilian studies were carried out when the level of salt iodination was still from 20 to 60 mg iodine/ Kg salt. 24 There are more recent Brazilian studies were carried out in the states of São Paulo and Rio de Janeiro, after ANVISA's resolution of 2013. Saraiva et al. 25 analyzed 629 urine samples from 244 pregnant women in the first 12 weeks of pregnancy in the state of Rio de Janeiro. The median of urinary iodine concentration was adequate (221 mcg/l), with 48.7% of women with insufficiency (<150 mcg/l) and 4.5% with excess (> 500 mcg/l) in at least one of the samples. 25 The median of urinary iodine concentration is used as a populational marker in iodine status. In our study, the concerned population showed iodine sufficiency. In a study conducted in the city of São Paulo, pregnant women in all the trimesters of pregnancy were evaluated, and a median urinary iodine concentration of 144 mcg/l was found, with rates from 27.3 to 403 mcg/l, which indicates iodine insufficiency. A urinary iodine concentration below 150 mcg/l was found in 52% of pregnant women, and lower than 50 mcg/l in only 2%. The urinary iodine concentration was proper for 44% of the women, and 4% presented with levels above the requirement. 26 Our study corroborates the results found by Saraiva et al. in the state of Rio de Janeiro, 25 which also showed iodine sufficiency in their population. The fact that these two studies have different results from the investigation carried out in São Paulo by Mioto et al. 26 may be justified by the time of the urinary iodine samples collection (in the afternoon in our study, three random samples in the study from Rio de Janeiro, and a morning sample in São Paulo). The urinary iodine release levels reflect the recent iodine intake, and the morning period is when iodine concentration is at its lowest. This way, studies carried out using urine collected in morning periods cannot be directly compared to studies in which urine samples were collected throughout the day. 27 The main limitations of this study are its small sample size and the lack of information on dietary intake. However, until this moment, our research is one of the few Brazilian studies evaluating the iodine nutritional status in pregnant women while correlating it with the thyroid function of newborns, an outcome that is one of the primary and most concerning consequences of iodine deficiency in pregnant women. Therefore, despite the limitation of the small sample of pregnant women, no neonatal TSH was higher than 5 mIU/l in the studied sample, showing iodine sufficiency in this population sample. In our results, overweight was associated with iodine dosages above the normal. Similarly, the study carried out by Saraiva et al. 25 also found that correlation. This result may be explained due to higher food intake with higher levels of iodine by this subpopulation. This study shows that the population of pregnant women attended in the prenatal ambulatory of normal risk from the Federal University of Minas Gerais is considered iodine sufficient, even after the decrease of salt iodine levels in Brazil. It should be emphasized that there were no cases of neonatal hypothyroidism reported in the study. In this context, we may state that there is insufficient data to recommend the iodine supplementation for all Brazilian pregnant women, in the face of the heterogeneity of studies' results. Brazil is a continental country, which makes uniform standardization hard to be implanted. Several factors may interfere in the iodine sufficiency of pregnant women. Differences in the diet, social and demographic characteristics, easiness of people and food transportation, business negotiations, and processed food consumption may explain the coexistence of areas with iodine insufficiency and excessive iodine intake in the country. 28 Conclusion The results of this study reinforce the importance of regional studies, such as this one. Studies in different regions and the collection of a higher number of urinary samples per pregnant woman might improve the accuracy of the results, considering that urinary iodine concentration may vary day-to-day, including within-day variation. 29 30 This way, it is too early to recommend a universal iodine supplementation for Brazilian pregnant women, and more studies need to be carried out. It is worthy to emphasize that iodine supplementation for pregnant women in iodine sufficient areas is associated with the risks of the fetus's excessive exposure to iodine. Acknowledgments Núcleo de Ações e Pesquisa em apoio diagnóstico (NUPAD) and Centro de Pós-graduação (Saúde da Mulher). Contributors Conflict of Interests The authors have no conflict of interests to declare. All of the authors contributed with the project and data interpretation, the writing of the article, the critical review of the intellectual content, and with the final approval of the version to be published. ==== Refs References 1 Glinoer D The regulation of thyroid function during normal pregnancy: importance of the iodine nutrition status Best Pract Res Clin Endocrinol Metab 2004 18 02 133 152 10.1016/j.beem.2004.03.001 15157832 2 Glinoer D The regulation of thyroid function in pregnancy: pathways of endocrine adaptation from physiology to pathology Endocr Rev 1997 18 03 404 433 10.1210/edrv.18.3.0300 9183570 3 De Groot L Abalovich M Alexander E K Amino N Barbour L Cobin R H Management of thyroid dysfunction during pregnancy and postpartum: an Endocrine Society clinical practice guideline J Clin Endocrinol Metab 2012 97 08 2543 2565 10.1210/jc.2011-2803 22869843 4 Alexander E K Pearce E N Brent G A Brown R S Chen H Dosiou C Guidelines of the American Thyroid Association for the diagnosis and management of thyroid disease during pregnancy and the postpartum Thyroid 2017 27 03 315 389 10.1089/thy.2016.0457 28056690 5 Lazarus J Brown R S Daumerie C Hubalewska-Dydejczyk A Negro R Vaidya B 2014 European thyroid association guidelines for the management of subclinical hypothyroidism in pregnancy and in children Eur Thyroid J 2014 3 02 76 94 10.1159/000362597 25114871 6 Eastman C J Zimmermann M B The iodine deficiency disorders [Internet]South Dartmouth: MDText.com; 2000 [cited 2019 Oct 1]. Available from:https://www.ncbi.nlm.nih.gov/books/NBK285556/ 7 Todd C H Allain T Gomo Z A Hasler J A Ndiweni M Oken E Increase in thyrotoxicosis associated with iodine supplements in Zimbabwe Lancet 1995 346 (8989):1563 1564 10.1016/s0140-6736(95)92095-1 7491075 8 Farebrother J Zimmermann M B Andersson M Excess iodine intake: sources, assessment, and effects on thyroid function Ann N Y Acad Sci 2019 1446 01 44 65 10.1111/nyas.14041 30891786 9 Bürgi H Iodine excess Best Pract Res Clin Endocrinol Metab 2010 24 01 107 115 10.1016/j.beem.2009.08.010 20172475 10 World Health Organization Assessment of iodine deficiency disorders and monitoring their elimination: a guide for programme managers 3rd ed. Geneva WHO 2007 11 Zimmermann M B Iodine deficiency Endocr Rev 2009 30 04 376 408 10.1210/er.2009-0011 19460960 12 Sullivan K M May W Nordenberg D Houston R Maberly G F Use of thyroid stimulating hormone testing in newborns to identify iodine deficiency J Nutr 1997 127 01 55 58 10.1093/jn/127.1.55 9040544 13 Zimmermann M B Aeberli I Torresani T Bürgi H Increasing the iodine concentration in the Swiss iodized salt program markedly improved iodine status in pregnant women and children: a 5-y prospective national study Am J Clin Nutr 2005 82 02 388 392 10.1093/ajcn.82.2.388 16087983 14 Caldwell K L Pan Y Mortensen M E Makhmudov A Merrill L Moye J Iodine status in pregnant women in the National Children's Study and in U.S. women (15-44 years), National Health and Nutrition Examination Survey 2005-2010 Thyroid 2013 23 08 927 937 10.1089/thy.2013.0012 23488982 15 Barca M F Knobel M Tomimori E Cárdia M S Zugaib M Medeiros-Neto G Aspectos ultra-sonográficos e prevalência da tireoidite pós-parto em gestantes sem disfunção tireóidea atendidas em Hospital Público de São Paulo. Arq Bras Endocrinol Metab 2001 45 02 180 9 10.1590/S0004-2730200100020001 16 Hollander M Wolfe D A Nonparametric statistical methods New York John Wiley & Sons 1999 17 Efroymson M A Multiple regression analysis New York John Wiley 1960. p. 191-203 18 Ministério da Saúde Secretaria de Atenção à Saúde. Departamento de Atenção Básica. Orientações para a coleta e análise de dados antropométricos em serviços de saúde: Norma Técnica do Sistema de Vigilância Alimentar e Nutricional - SISVAN Brasília, DF Ministério da Saúde 2011 19 Anaforoğlu İ Algün E İnceçayır Ö Topbaş M Erdoğan M F Iodine status among pregnant women after mandatory salt iodisation Br J Nutr 2016 115 03 405 410 10.1017/S0007114515004559 26596695 20 Delshad H Touhidi M Abdollahi Z Hedayati M Salehi F Azizi F Inadequate iodine nutrition of pregnant women in an area of iodine sufficiency J Endocrinol Invest 2016 39 07 755 762 10.1007/s40618-016-0438-4 26951055 21 Snart C JP Keeble C Taylor E Cade J E Stewart P M Zimmermann M Maternal iodine status and associations with birth outcomes in three major cities in the United Kingdom Nutrients 2019 11 02 441 10.3390/nu11020441 30791590 22 Soares R Vanacor R Manica D Dorneles L B Resende V L Bertoluci M C Thyroid volume is associated with family history of thyroid disease in pregnant women with adequate iodine intake: a cross-sectional study in southern Brazil J Endocrinol Invest 2008 31 07 614 617 10.1007/BF03345612 18787379 23 Ferreira S M Navarro A M Magalhães P K Maciel L M Iodine insufficiency in pregnant women from the State of São Paulo. Arq Bras Endocrinol Metab 2014 58 03 282 7 10.1590/0004-2730000002979 24 Medeiros-Neto G Iodine nutrition in Brazil: where do we stand? Arq Bras Endocrinol Metab 2009 53 04 470 4 10.1590/S0004-27302009000400014 25 Saraiva D A Morais N AOES Martins Corcino C Berbara T MBL Schtscherbyna A Santos M Iodine status of pregnant women from a coastal Brazilian state after the reduction in recommended iodine concentration in table salt according to governmental requirements Nutrition 2018 53 109 114 10.1016/j.nut.2018.02.001 29677691 26 Mioto V CB Monteiro A CCNG de Camargo R YA Borel A R Catarino R M Kobayashi S High prevalence of iodine deficiency in pregnant women living in adequate iodine area Endocr Connect 2018 7 05 762 767 10.1530/EC-18-0131 29700098 27 Als C Helbling A Peter K Haldimann M Zimmerli B Gerber H Urinary iodine concentration follows a circadian rhythm: a study with 3023 spot urine samples in adults and children J Clin Endocrinol Metab 2000 85 04 1367 1369 10.1210/jcem.85.4.6496 10770167 28 Campos RdeO Barreto IdosS Maia L R Rebouças S C Cerqueira T L Oliveira C A Iodine nutritional status in Brazil: a meta-analysis of all studies performed in the country pinpoints to an insufficient evaluation and heterogeneity Arch Endocrinol Metab 2015 59 01 13 22 10.1590/2359-3997000000004 25926109 29 Rasmussen L B Ovesen L Christiansen E Day-to-day and within-day variation in urinary iodine excretion Eur J Clin Nutr 1999 53 05 401 407 10.1038/sj.ejcn.1600762 10369497 30 König F Andersson M Hotz K Aeberli I Zimmermann M B Ten repeat collections for urinary iodine from spot samples or 24-hour samples are needed to reliably estimate individual iodine status in women J Nutr 2011 141 11 2049 2054 10.3945/jn.111.144071 21918061
PMC009xxxxxx/PMC9708398.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36446556 10.1055/s-0042-1758467 RBGO-22-0510-Editorial Editorial Fear of Childbirth: It is Time to Talk About It! http://orcid.org/0000-0003-3244-2930 Imakawa Cibele Santini de Oliveira 1 http://orcid.org/0000-0002-6706-7117 Nadai Mariane Nunes de 2 http://orcid.org/0000-0001-6938-6288 Reis Monica Iassana dos 3 http://orcid.org/0000-0002-9311-786X Quintana Silvana Maria 1 http://orcid.org/0000-0001-9099-4544 Moises Elaine Christine Dantas 1 1 Ribeirão Preto Medical School, Universidade de São Paulo, Ribeirão Preto, SP, Brazil 2 Bauru School of Dentistry, Universidade de São Paulo, Bauru, SP, Brazil 3 National Institute for Women's, Children's and Adolescent Health Fernandes Figueira, Oswaldo Cruz Foundation, Rio de Janeiro, RJ, Brazil Address for correspondence Elaine Christine Dantas Moises, PhD Av. Bandeirantes, 3900, 14049-900, Campus da Universidade de São Paulo, Ribeirão Preto, [email protected] 29 11 2022 10 2022 1 11 2022 44 10 907908 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. ==== Body pmc Fear encompasses concerns on a spectrum ranging from mild fear to phobia. 1 When it comes to fear related to childbirth, there is no consensus on its exact definition. However, there is no doubt about its importance in obstetric care. 2 3 4 When fear of childbirth is intense, it can harm the woman's health, 5 6 becoming a disabling factor that interferes with the daily routine, such as occupational and domestic activities and social life. 7 There are disagreements regarding the prevalence of fear of childbirth (FOC) due to the lack of consensus on its definition, as well as the differences in the diagnostic methods adopted by the studies. 7 8 9 In general, it is known that some degree of concern and fear can be presented by up to 80% of pregnant women. 10 11 In Brazil, this topic is rarely discussed among health professionals, and the assessment of its real prevalence remains unknown. Some protective factors can lead to lower incidences of FOC, such as relationship stability and duration. In turn, there are risk factors that can increase its incidence, such as lack of social support; unplanned pregnancy; infertility; previous negative experiences; anxiety, depression, and other psychological disorders. 12 Younger women tend to be more afraid of childbirth than older women. The woman's nationality also seems to influence the occurrence of FOC. 13 Among the obstetric aspects, nulliparity, advanced gestational age and a history of previous cesarean section stand out as risk factors for FOC. 10 14 The preference for cesarean section is expressed by 27% of women who use the public health system and by 44% of those who access private health services in Brazil. In addition to social, economic, and cultural aspects, fear of labor pain and childbirth itself seem to be factors that could influence the mother's preference for cesarean section. 15 16 17 Besides the maternal request for cesarean section, the increase in other adverse obstetric outcomes such as preterm birth, prolonged labor, postpartum depression, and post-traumatic stress have already been related to the fear of childbirth. It is extremely important to establish treatment plans for identified cases of FOC to minimize its negative impacts on obstetric outcomes. Accessible and timely interventions, which can contribute to the reduction of FOC, include educational and reassuring measures such as the construction of a birth plan; discussion groups with pregnant women and health professionals; access to educational materials on delivery mechanisms and obstetric care; establishment of a support network; empathic follow-up; psychological support. In more severe cases, it is essential to refer the pregnant woman to professionals specialized in mental health for better monitoring and control of symptoms. There are still little data in the literature on the prevalence, predictive and protective factors concerning fear of childbirth according to each population, considering the obstetric history and social, economic, and cultural aspects. A research group linked to the Ribeirão Preto Medical School of the University of São Paulo is evaluating this issue, to characterize the context of FOC in the Brazilian population and analyze the effectiveness of possible interventions to reduce the impact of this condition. Furthermore, it is intended to promote the discussion on the need for public policies aimed at identifying the fear of childbirth during prenatal care, implementing strategies to improve the education of the population regarding the physiology of childbirth and obstetric care (including information on the mode of delivery and its risks/benefits, delivery plan, labor pain management, evidence-based interventions that can occur during labor and delivery) and, consequently, expand access to these resources. FOC exists, it is a real problem, but it remains masked by the knowledge gaps on the subject. Thus, it is essential to develop studies, establish guidelines, and train health professionals on this important issue. There is an urgent need to implement diagnostic and management measures for these cases in clinical practice. It is time to talk about it! Conflicts of Interest The authors have no conflict of interest to declare. ==== Refs References 1 Rondung E Thomtén J Sundin Ö Psychological perspectives on fear of childbirth J Anxiety Disord 2016 44 80 91 10.1016/j.janxdis.2016.10.007 27788373 2 Räisänen S Lehto S M Nielsen H S Gissler M Kramer M R Heinonen S Fear of childbirth in nulliparous and multiparous women: a population-based analysis of all singleton births in Finland in 1997-2010 BJOG 2014 121 08 965 970 10.1111/1471-0528.12599 24494605 3 Aksoy A N Ozkan H Gundogdu G Fear of childbirth in women with normal pregnancy evolution Clin Exp Obstet Gynecol 2015 42 02 179 183 26054114 4 Richens Y Smith D M Lavender D T Fear of birth in clinical practice: A structured review of current measurement tools Sex Reprod Healthc 2018 16 98 112 10.1016/j.srhc.2018.02.010 29804785 5 O'Connell M A Leahy-Warren P Khashan A S Kenny L C O'Neill S M Worldwide prevalence of tocophobia in pregnant women: systematic review and meta-analysis Acta Obstet Gynecol Scand 2017 96 08 907 920 10.1111/aogs.13138 28369672 6 Saisto T Halmesmäki E Fear of childbirth: a neglected dilemma Acta Obstet Gynecol Scand 2003 82 03 201 208 12694113 7 Rouhe H Salmela-Aro K Halmesmäki E Saisto T Fear of childbirth according to parity, gestational age, and obstetric history BJOG 2009 116 01 67 73 10.1111/j.1471-0528.2008.02002.x 19055652 8 Haines H M Pallant J F Fenwick J Identifying women who are afraid of giving birth: A comparison of the fear of birth scale with the WDEQ-A in a large Australian cohort Sex Reprod Healthc 2015 6 04 204 210 10.1016/j.srhc.2015.05.002 26614602 9 Hildingsson I Rubertsson C Karlström A Haines H Exploring the Fear of Birth Scale in a mixed population of women of childbearing age-A Swedish pilot study Women Birth 2018 31 05 407 413 10.1016/j.wombi.2017.12.005 29249331 10 Szeverényi P Póka R Hetey M Török Z Contents of childbirth-related fear among couples wishing the partner's presence at delivery J Psychosom Obstet Gynaecol 1998 19 01 38 43 10.3109/01674829809044219 9575467 11 Melender H L Experiences of fears associated with pregnancy and childbirth: a study of 329 pregnant women Birth 2002 29 02 101 111 10.1046/j.1523-536X.2002.00170.x 12051188 12 Dencker A Nilsson C Begley C Causes and outcomes in studies of fear of childbirth: A systematic review Women Birth 2019 32 02 99 111 10.1016/j.wombi.2018.07.004 30115515 13 Ternström E Hildingsson I Haines H Rubertsson C Higher prevalence of childbirth related fear in foreign born pregnant women–findings from a community sample in Sweden Midwifery 2015 31 04 445 450 10.1016/j.midw.2014.11.011 25529841 14 Adams S S Eberhard-Gran M Eskild A Fear of childbirth and duration of labour: a study of 2206 women with intended vaginal delivery BJOG 2012 119 10 1238 1246 10.1111/j.1471-0528.2012.03433.x 22734617 15 Mazzoni A Althabe F Gutierrez L Women's preferences and mode of delivery in public and private hospitals: a prospective cohort study BMC Pregnancy Childbirth 2016 16 34 10.1186/s12884-016-0824-0 26857448 16 Rattner D Moura E C Nascimentos no Brasil: associação do tipo de parto com variáveis temporais e sociodemográficas Rev Bras Saúde Mater Infant 2016 16 01 39 47 10.1590/1806-93042016000100005 17 Reiter M Betrán A P Marques F K Torloni M R Systematic review and meta-analysis of studies on delivery preferences in Brazil Int J Gynaecol Obstet 2018 143 01 24 31 10.1002/ijgo.12570 29920679
PMC009xxxxxx/PMC9708399.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36446564 10.1055/s-0042-1756148 RBGO-20-0297 Review Article Efficacy of Hormonal and Nonhormonal Approaches to Vaginal Atrophy and Sexual Dysfunctions in Postmenopausal Women: A Systematic Review Eficácia das abordagens hormonais e não hormonais para atrofia vaginal e disfunções sexuais em mulheres na pós-menopausa: Uma revisão sistemáticahttp://orcid.org/0000-0001-9131-1952 Sarmento Ayane Cristine Alves 1 http://orcid.org/0000-0002-4511-4373 Costa Ana Paula Ferreira 1 http://orcid.org/0000-0002-9309-0513 Lírio Juliana 2 http://orcid.org/0000-0003-4617-7269 Eleutério Jr José 3 http://orcid.org/0000-0001-5335-6770 Baptista Pedro Vieira 4 http://orcid.org/0000-0002-8351-5119 Gonçalves Ana Katherine 1 1 Postgraduate Program student in Health Science, Universidade Federal do Rio Grande do Norte, Natal, RN, Brazil 2 Department of Obstetrics and Gynecology, Universidade Estadual de Campinas, Campinas, SP, Brazil 3 Department of Obstetrics and Gynecology, Universidade Federal do Ceará, Fortaleza, CE, Brazil 4 Lower Genital Tract Unit, Centro Hospital de São João, Porto, Portugal Address for correspondence Ana Katherine Gonçalves, Professor rua Major Laurentino de Morais, 1.218/1.301, Natal, RN, [email protected] 29 11 2022 10 2022 1 11 2022 44 10 986994 17 8 2020 06 10 2020 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  To evaluate the efficacy of the hormonal and nonhormonal approaches to symptoms of sexual dysfunction and vaginal atrophy in postmenopausal women. Data Sources  We conducted a search on the PubMed, Embase, Scopus, Web of Science, SciELO, the Cochrane Central Register of Controlled Trials (CENTRAL), and Cumulative Index to Nursing and Allied Health Literature (CINAHL) databases, as well as on clinical trial databases. We analyzed studies published between 1996 and May 30, 2020. No language restrictions were applied. Selection of Studies  We selected randomized clinical trials that evaluated the treatment of sexual dysfunction in postmenopausal women. Data Collection  Three authors (ACAS, APFC, and JL) reviewed each article based on its title and abstract. Relevant data were subsequently taken from the full-text article. Any discrepancies during the review were resolved by consensus between all the listed authors. Data Synthesis  A total of 55 studies were included in the systematic review. The approaches tested to treat sexual dysfunction were as follows: lubricants and moisturizers (18 studies); phytoestrogens (14 studies); dehydroepiandrosterone (DHEA; 8 studies); ospemifene (5 studies); vaginal testosterone (4 studies); pelvic floor muscle exercises (2 studies); oxytocin (2 studies); vaginal CO 2 laser (2 studies); lidocaine (1 study); and vitamin E vaginal suppository (1 study). Conclusion  We identified literature that lacks coherence in terms of the proposed treatments and selected outcome measures. Despite the great diversity in treatment modalities and outcome measures, the present systematic review can shed light on potential targets for the treatment, which is deemed necessary for sexual dysfunction, assuming that most randomized trials were evaluated with a low risk of bias according to the Cochrane Collaboration risk of bias tool. The present review is registered with the International Prospective Register of Systematic Reviews (PROSPERO; CRD42018100488). Resumo Objetivo  Avaliar a eficácia das abordagens hormonais e não hormonais para os sintomas de disfunção sexual e atrofia vaginal em mulheres na pós-menopausa. Fontes de Dados  Pesquisamos as bases de dados PubMed, Embase, Scopus, Web of Science, SciELO, Cochrane Central Register of Controlled Trials (CENTRAL), e Cumulative Index to Nursing and Allied Health Literature (CINAHL), assim como bancos de dados de ensaios clínicos. Foram analisados estudos publicados entre 1996 e 30 de maio de 2020. Nenhuma restrição de idioma foi aplicada. Seleção dos Estudos  Foram selecionados ensaios clínicos randomizados que avaliavam o tratamento das disfunções sexuais em mulheres na pós-menopausa. Coleta de Dados  Três autores (ACAS, APFC e JL), revisaram cada artigo com base em seu título e resumo. Os dados relevantes foram posteriormente retirados do texto completo do artigo. Quaisquer discrepâncias durante a revisão foram resolvidas por consenso entre todos os autores listados. Síntese dos Dados  Ao todo, 55 estudos foram incluídos na revisão sistemática. As abordagens testadas para tratar a disfunção sexual foram: lubrificantes e hidratantes (18 estudos); fitoestrogênios (14 estudos); deidroepiandrosterona (DHEA; 8 estudos); ospemifeno (5 estudos); testosterona vaginal (4 estudos); exercícios para os músculos do assoalho pélvico (2 estudos); oxitocina (2 estudos); laser de CO 2 vaginal (2 estudos); lidocaína (1 estudo), e vitamina E vaginal (1 estudo). Conclusão  Identificou-se falta de coerência na literatura quanto aos tratamentos propostos e medidas de resultados selecionadas. Apesar da grande diversidade de modalidades de tratamento e medidas de resultados, esta revisão sistemática pode lançar luz sobre alvos potenciais para o tratamento, que é considerado necessário para a disfunção sexual, assumindo que a maioria dos estudos randomizados foi avaliada com baixo risco de viés de acordo com a ferramenta de avaliação de risco de viés de Cochrane Collaboration. Esta revisão tem cadastro no International Prospective Register of Systematic Reviews (PROSPERO; CRD42018100488). Keywords menopause dyspareunia orgasm therapeutics estrogens Palavras-chave menopausa dispareunia orgasmo terapêutica estrogênios ==== Body pmcIntroduction According to the North American Menopause Society (NAMS), natural menopause is defined as the final menstrual period, diagnosed after 12 consecutive months of spontaneous amenorrhea without an apparent pathological cause. The most common symptoms associated with menopause are hot flushes, night sweats, sleep disturbance, vaginal atrophy, and dyspareunia. 1 2 Studies 3 4 5 6 7 8 indicate that postmenopausal women with vulvovaginal atrophy (VVA) have a higher probability of developing sexual dysfunction, including difficulties regarding sexual desire, arousal, lubrication, and orgasm. Vulvovaginal atrophy leads to thinning of the mucus and tissues of the vulva and vagina caused by the estrogen deprivation that occurs in women during this period. Patients with VVA complain of vaginal irritation and discharge, itching, dryness, dysuria, and dyspareunia. Despite the decrease in sexual function indices during the menopausal transition, 3 it is not clear whether this is caused by lower levels of ovarian hormones, aging, or both. 4 According to the American Psychiatric Association (APA), 5 6 7 8 sexual dysfunction is defined as disorders of sexual desire and psychophysiological changes that characterize the sexual-response cycle, causing marked suffering and interpersonal difficulty. Female sexual dysfunction (FSD) can be assessed in different domains, including sexual interest, arousal, orgasm, and pain,9 and is not easy to define or investigate, because it depends on several factors, such as health and well-being, cultural habits, socioeconomic status, relationship problems, and sexual partner's existence or not. 5 The measurement of sexual function in women has never been an easy task. Furthermore, the literature 6 describes a diverse assortment of outcome measures across various symptom levels, and, as a result, pooling of data is imprecise. Because of the complexity of FSD, it is often difficult to define the factors primarily responsible for the disorder and to establish meaningful steps in the treatment. The treatments usually aim to deal with individual symptoms, but no single treatment modality addresses the entire spectrum of the disorder. Therefore, multimodal therapy is required. The great variety of distinct classes of medication described in the literature 7 confirms the complexity of the disorder and indicates the need for a systematic review to define the best practices. Several options are available to treat this condition, and hormonal therapies (HTs; estrogens and androgens) and nonhormonal therapies (non-HTs; lubricants and long-acting vaginal moisturizers) are the most commonly used. It is recognized that vaginal estrogen may improve the symptoms; conversely, non-HTs can be useful in specific cases in which HTs are harmful or not recommended (such as breast cancer). 7 10 The real effect of the treatments on sexual function in menopausal women is particularly difficult to understand based on the literature published so far, partly because of different drugs, routes, and dosages, as well as the diversity of tools used to assess sexual function and the studied population in each trial. Similarly, it is essential to have guidance on effective alternative treatments for women who cannot use HTs. Thus, there is a real need for a systematic review of the best available evidence to facilitate evidence-based decisions. Thus, the present systematic review aimed to evaluate the efficacy of hormonal and nonhormonal approaches for symptoms of sexual dysfunction and vaginal atrophy in postmenopausal women. Methods Study Design, Data Search, and Inclusion and Exclusion Criteria The present systematic review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement 11 and has been registered with PROSPERO (CRD42018100488). Ethical approval was not required because the review uses published patient data. A full systematic search was carried out on the following databases: PubMed, Embase, Scopus, Web of Science, SciELO, the Cochrane Central Register of Controlled Trials (CENTRAL), and Index to Nursing and Allied Health Literature CINAHL, as well as clinical trial databases ( www.trialscentral.org ; www.controlled-trials.com ). We included studies published from 1996 until May 30, 2020 without language restrictions. Grey literature was not searched. Combinations of the following keywords were used to identify the studies: menopause ; postmenopause ; dyspareunia ; orgasm ; therapeutics ; dehydroepiandrosterone ; testosterone ; oxytocin ; phytoestrogens ; lidocaine ; hyaluronic acid ; lubricants ; and laser . The PubMed search strategy is shown in Chart 1 . Chart 1 Search strategy on PubMed Number Search items 1 Menopause 2 Postmenopause 3 Genitourinary syndrome of menopause 4 Dyspareunia 5 Orgasm 6 Vaginal atrophy 7 Female sexual dysfunction 8 Sexual function 9 Or/1-8 10 Therapeutics 11 Vaginal estrogen 12 Estrogens 13 Vaginal therapy 14 Non-hormonal treatments 15 Dehydroepiandrosterone 16 Testosterone 17 Oxytocin 18 Phytoestrogens 19 Local lidocaine 20 Hyaluronic acid 21 Lubricants 22 Moisturizers 23 Laser therapy 24 Radio waves 25 Microablative fractional radiofrequency 26 Physical therapy modalities 27 Or/10-26 28 Randomized clinical trial 29 Clinical trial 30 Or/28-29 31 9 and 27 and 30 Three authors (ACAS, APFC, and JL) reviewed each article based on its title and abstract. The relevant data were subsequently taken from the full-text article. We evaluated the full text based on the following inclusion criteria: 1) studies with postmenopausal women; 2) randomized clinical trials evaluating the efficacy of hormonal or nonhormonal approaches to sexual dysfunction; 3) intervention groups undergoing treatment for sexual dysfunction; 4) any HT (considering HTs, according to the NAMS, as all the prescription drugs used most often to treat hot flashes and symptoms of genitourinary syndrome of menopause [GSM], which include vaginal dryness) or non-HT being compared to a placebo or no intervention or vaginal estrogen (all types of estrogen, progesterone, and androgens); 1 and 5) improvement in sexual dysfunction as the primary outcome. Cross-sectional and observational studies were excluded, and published studies were excluded if only their abstract was available. The full texts of potentially-eligible studies were extracted and examined for the following data: year the study was conducted; number of subjects; location of subject recruitment; mean age of the subjects; randomization and blinding processes; inclusion and exclusion criteria; description of the therapies used, and definition of the outcomes measured; length of follow-up; and side effects. Any discrepancies during the review were resolved by consensus between all the listed authors. Primary and Secondary Outcomes The primary outcomes were vaginal dryness, arousal, desire, orgasm, lubrication, satisfaction, pain, and dyspareunia. The secondary outcomes included side effects, treatment duration, and adhesion. The outcomes of the randomized controlled trials included had to be measured using gynecological and visual examinations, pH, cell maturation index, and validated scales. For the secondary outcomes, the side effects were evaluated according to the records of their onset and throughout treatment by self-assessment questionnaires. Assessment of the Risk of Bias and Qualitative Analysis To assess the risk of bias, the Cochrane Collaboration risk of bias tool was applied to evaluate the following criteria: adequate sequence generation; allocation concealment; blinding; incomplete outcome data; selective reporting; and other risks of bias. 12 Three authors assessed each original study, and the quality of the data is shown in Chart 2 . Relevant data were subsequently extracted from the full-text article, according to the data extraction protocol. Each of the aforementioned criteria received one of the following classifications: “low risk of bias”; “high risk of bias”; or “unclear risk of bias”. Disagreements were resolved by consulting a third author. Moreover, regarding the primary outcomes, we assessed the certainty of the evidence according to the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) classification. 13 Chart 2 Quality assessment of the included studies using the Cochrane risk of bias tool Study Random sequence generation Allocation concealment Blinding of participants and personnel Blinding of outcome assessment Incomplete outcome data Selective reporting Other bias Bygdeman and Swahn (1996) 14 Loprinzi et al. (1997) 15 Balk et al. (2002) 32 Labrie et al. (2009) 45 Bachmann and Komi (2010) 52 Oh et al. (2010) 33 Raghunandan et al. (2010) 16 Ekin et al. (2011) 67 Genazzani et al. (2011) 46 Jonasson et al. (2011) 61 Labrie et al. (2011) 27 Le Donne et al. (2011) 28 Lee et al. (2011) 17 Loprinzi et al. (2011) 18 Amato et al. (2013) 34 Grimaldi et al. (2012) 29 Tedeschi and Benvenuti (2012) 35 Chen et al. (2013) 68 Lima et al. (2013) 36 Portman et al. (2013) 53 Zheng et al. (2013) 30 Constantine et al. (2015) 54 Fernandes et al. (2014) 19 Labrie et al. (2014) 47 Lima et al. (2014) 37 Portman et al. (2014) 55 Archer et al. (2015) 48 Bouchard et al. (2015) 49 Goetsch et al. (2015) 65 Labrie et al. (2015) 50 Tungmunsakulchai et al. (2015) 57 Hickey et al. (2016) 20 Jokar et al. (2016) 31 Juliato et al. (2017) 21 Labrie et al. (2016) 51 Melisko et al. (2017) 58 Postigo et al. (2016) 38 Seyyedi et al. (2016) 22 De Souza et al. (2016) 39 Yaralizadeh et al. (2016) 40 Cruz et al. (2018) 63 Malakouti et al. (2017) 41 Nappi et al. (2017) 23 Nazarpour et al. (2017) 59 Suwanvesh et al. (2017) 42 Diem et al. (2018) 24 Golmakani et al. (2019) 66 Mitchell et al. (2018) 25 Nazarpour et al. (2018) 60 Torky et al. (2018) 62 Archer et al. (2019) 56 Ghorbani et al. (2019) 43 Mitchell et al. (2019) 26 Palma et al. (2019) 44 Politano et al. (2019) 64 Key: High risk of bias; Unclear risk of bias; Low risk of bias As for the quantitative studies, we could not perform a formal statistical analysis because of the heterogeneity of the measurements. However, the findings relevant to the aims of the present review were extracted. Results Study Selection and Characteristics After searching the databases, 25,631 articles were identified ( Figure 1 ). After a review of their titles and abstracts, 25,488 records were excluded, and 143 records remained for full-text review. In total, 44 articles were excluded because they were duplicates. After reviewing the full-text articles, an additional 44 records were excluded. This process resulted in 55 articles, which were further reviewed using a manual search approach. Fig. 1 Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) diagram of the search strategy used in the present systematic review. A total of 55 studies were deemed eligible for inclusion ( Table 1, Supplementary material ). The approaches tested to treat vaginal atrophy and sexual dysfunction were as follows: lubricants and moisturizers (18 studies, 14 15 16 17 18 19 20 21 22 23 24 25 26 5 27 28 29 30 31 of which evaluated hyaluronic acid); phytoestrogens (14 studies); 30 32 33 34 35 36 37 38 39 40 41 42 43 44 dehydroepiandrosterone (DHEA; 8 studies); 27 45 46 47 48 49 50 51 ospemifene (5 studies); 52 53 54 55 56 vaginal testosterone (4 studies); 16 19 57 58 pelvic floor muscle exercises (2 studies); 59 60 oxytocin (2 studies); 61 62 vaginal CO 2 laser (2 studies); 63 64 lidocaine (1 study); 65 and vitamin E vaginal suppository (1 study). 66 Overall, 18 studies evaluated the parameters of dyspareunia and vaginal dryness, 2 evaluated symptoms of VVA, and the other studies used validated questionnaires for the evaluation of sexual dysfunction. Side Effects Regarding side effects, the results for lubricants and moisturizers indicated that all adverse events (AEs) were considered mild in severity and self-limited. 14 15 16 17 18 19 20 21 22 23 24 25 26 The most common AEs were burn on the application site, discomfort on the application site, vulvovaginal pruritus, and general pruritus. The compound hyaluronic acid has been shown to be well tolerated without side effects among most patients tested. 28 29 31 67 68 For soy flour and fennel cream, no adverse effects were reported in any of the included studies. 32 40 Cimicifuga foetida extract was found to be effective and safe during three months of treatment. 30 For ginseng , 12 events were reported during a clinical trial; 33 of these, 2 were of vaginal bleeding, which were assumed to be possibly related to the trial. Among the studies selected, there were only two reports of participants experiencing serious AEs. 34 35 36 37 For Tribulus terrestris , the most frequent adverse effects were diarrhea, nervousness, dizziness, and nausea. 38 39 In relation to the other phytoestrogens, the studies did not provide data on side effects. Approximately half of the subjects treated with DHEA experienced the most frequently reported treatment-emergent AEs (TEAEs). The three most commonly reported preferred terms were discharge at the application site , urinary tract infection , and headache . No adverse effects were observed regarding hepatic tests, hematocrit, or any hematological or biochemical parameters. 27 45 46 47 48 49 50 51 Ospemifene was shown to be safe and well tolerated. Most TEAEs were classified as not related to the study drug or were unlikely to be related. The most frequent were urinary tract infection, hot flush, and nasopharyngitis. Only two patients experienced serious AEs, one of which was probably treatment-related (deep vein thrombosis). 52 53 54 55 56 All related AE treatments with vaginal testosterone were classified grade 1 (mild) or 2 (moderate). The most frequent were vaginal discharge, facial hair growth, vaginal or vulvar itching and/or irritation, vaginal odor, and urinary tract or yeast infection. 16 19 57 58 None of the study participants reported any side effects related to the use of oxytocin. 61 62 There were no data available on the side effects of the treatment with pelvic floor muscle exercises, 59 60 vaginal CO 2 laser, 63 lidocaine, 65 and vitamin E vaginal suppository. 66 Discussion The female population is aging and facing new health issues, including menopause symptoms and FSD. Despite this, the best approach to sexual dysfunction in menopausal women is still unclear. Until now, estrogens and androgens for FSD have been shown to improve not only dyspareunia and vaginal dryness but also the desire and orgasm domains. However, not all women can use them; breast cancer, endometrial cancer, and deep venous thrombosis, among other conditions, can be contraindications for HTs. Moreover, there is insufficient evidence to attest to the effectiveness of the non-HTs studied to improve orgasm, lubrication, and sexual satisfaction in menopausal women. 1 2 The present review suggests that the use of lubricants and moisturizers has a lower impact than that of vaginal estrogens. The domains in which improvement was more evident were dyspareunia and vaginal dryness. The results of the present review point out the efficacy of hyaluronic acid for the symptoms of vaginal atrophy when compared with vaginal estrogen. Previous studies by Ekin et al. 67 and Chen et al. 68 have also demonstrated its effectiveness. Laser therapy, along with radiofrequency, is among the new non-HTs proposed. 69 70 Only two clinical trials 63 64 on laser therapy were found, which showed that there was a significant increase in the sexual score after two laser applications. Although pelvic floor muscle exercises have been widely discussed, only a single clinical trial has been found. The results concerning pelvic floor muscle showed that exercises have the potential to improve sexual function and are thus suggested to accompany other healthcare treatments designed for postmenopausal women. 59 60 A study 65 on the use of lidocaine in women with dyspareunia showed that it could enable comfortable intercourse; thus, it is a treatment that can be considered. Moreover, vitamin E vaginal suppository may be an alternative to vaginal estrogen in relieving the symptoms of vaginal atrophy in postmenopausal women, especially those not able to use HTs or those who have low compliance. 66 Ospemifene is a selective estrogen receptor modulator (SERM) that has been approved for the treatment of dyspareunia associated with VVA due to menopause. According to the studies 52 53 54 55 56 included in the present review, it was shown to be effective and well tolerated for the treatment of vaginal dryness and dyspareunia symptoms associated with VVA. A wide variety of phytoestrogens are being used to improve sexual function in symptomatic postmenopausal women. The phytoestrogens found in the present review were as follows: soy flour , 32 ginseng , 33 62 isoflavones , 34 35 44 C. foetida , 30 Glycine max ( L .), 36 37 T. terrestris , 38 39 fennel cream , 40 ginkgo biloba , 41 and Pueraria mirifica . 42 Studies 38 39 have shown that T. terrestris could be a safe alternative for the treatment of sexual desire disorder, since its probable mechanism of action involves an increase in the serum levels of free and bioavailable testosterone. Vaginal HTs, DHEA, and testosterone may have a positive impact on sexual desire and sexual function. The present review indicates that the vaginal HT with DHEA may have a positive impact on sexual desire and sexual function. This beneficial effect was apparent in all domains of sexual function, not only in those related to dyspareunia and vaginal dryness. 27 45 46 47 48 49 50 51 A meta-analysis conducted by Scheffers et al. 71 showed that oral and intravaginal DHEA used in peri- and postmenopausal women resulted in slight improvement in sexual function over placebo. Peixoto et al. 72 concluded that intravaginal DHEA is effective in improving several aspects of sexual function; however, there was no evaluation of vaginal dryness. Regarding testosterone, what is known so far is that the efficacy of the therapy may rely on synergistic effects with estrogen. 16 19 57 58 Previous reviews have revealed controversial conclusions. A study conducted by Khera 73 demonstrated that systemic testosterone might improve sexual desire, arousal, pleasure, and overall satisfaction. The data obtained from studies by Pitsouni et al. 7 and Reed et al. 74 pointed out that, in cases of low sexual desire, testosterone is not indicated. The results obtained in the present review are in agreement with those of the literature, as there was no difference in sexuality scores and sexual satisfaction when comparing topical testosterone and estrogen levels. Most treatments are considered safe, and no side effects were considered severe, only moderate to mild. It is important to highlight the lack of data about the side effects in some studies, which ends up generating uncertainties the true safety of treatment. According to the GRADE classification, the present study provided moderate or very low certainty of evidence that HTs or non-HTs are useful to improve sexual function in postmenopausal women. Consequently, the findings of the present study cannot be generalized until new randomized clinical trials are performed to confirm the strength of the evidence. Conclusion The present systematic review identified that the literature lacks coherence in terms of the proposed treatments and selected outcome measures. There is great diversity and heterogeneity in the scales used for the assessment of sexual dysfunction, despite the fact that they have been validated, hence the need for standardization in the scores. One of the most commonly used, the Female Sexual Function Index, is problematic regarding non-sexually active women. The present review should be considered in the context of its limitations. Despite the great diversity in treatment modalities and outcome measures, the review can shed light on potential targets for treatment, which is deemed necessary for sexual dysfunction, assuming that most randomized trials were evaluated with a low risk of bias, according yours methodological characteristics. Supplementary Material Supplementary Material Supplementary Material Conflict of Interests The authors have no conflict of interests to declare. ==== Refs References 1 North American Menopause Society The role of local vaginal estrogen for treatment of vaginal atrophy in postmenopausal women: 2007 position statement of The North American Menopause SocietyMenopause 2007;14(3 Pt 1):355–369, quiz 370–371 2 Endocrine Society Santen R J Allred D C Ardoin S P Archer D F Boyd N Braunstein G D Postmenopausal hormone therapy: an Endocrine Society scientific statement J Clin Endocrinol Metab 2010 95 (7, Suppl 1)s1 s66 10.1210/jc.2009-2509 20566620 3 Rosen R C Maserejian N N Connor M K Krychman M L Brown C S Goldstein I Characteristics of premenopausal and postmenopausal women with acquired, generalized hypoactive sexual desire disorder: the Hypoactive Sexual Desire Disorder Registry for women Menopause 2012 19 04 396 405 10.1097/gme.0b013e318230e286 22076307 4 da Silva Lara L A Useche B Rosa E Silva J C Ferriani R A Reis R M Sá M F Sexuality during the climacteric period Maturitas 2009 62 02 127 133 10.1016/j.maturitas.2008.12.014 19186014 5 Davis S R Understanding female sexual function Menopause 2009 16 03 425 426 10.1097/gme.0b013e31819c67a7 19225425 6 Weinberger J M Houman J Caron A T Anger J Female sexual dysfunction: a systematic review of outcomes across various treatment modalities Sex Med Rev 2019 7 02 223 250 10.1016/j.sxmr.2017.12.004 29402732 7 Pitsouni E Grigoriadis T Douskos A Kyriakidou M Falagas M E Athanasiou S Efficacy of vaginal therapies alternative to vaginal estrogens on sexual function and orgasm of menopausal women: A systematic review and meta-analysis of randomized controlled trials Eur J Obstet Gynecol Reprod Biol 2018 229 45 56 10.1016/j.ejogrb.2018.08.008 30103082 8 American Psychiatric Association Diagnostic and statistical manual of mental disorders: text revision (DSM-IV-TR) Washington, DC APA 2000 9 Binik Y M Brotto L A Graham C A Segraves R T Response of the DSM-V Sexual Dysfunctions subworkgroup to commentaries published in JSM J Sex Med 2010 7 07 2382 2387 10.1111/j.1743-6109.2010.01899 20653837 10 Kingsberg S A Wysocki S Magnus L Krychman M L Vulvar and vaginal atrophy in postmenopausal women: findings from the REVIVE (REal Women's VIews of Treatment Options for Menopausal Vaginal Changes) survey J Sex Med 2013 10 07 1790 1799 10.1111/jsm.12190 23679050 11 PRISMA Group Moher D Liberati A Tetzlaff J Altman D G Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement PLoS Med 2009 6 07 e1000097 10.1371/journal.pmed.1000097 19621072 12 Higgins J P Thomas J Chandler J Cumpston M Li T Page M J Cochrane handbook for systematic reviews of interventions. 2nd ed Chichester John Wiley & Sons 2019 13 Guyatt G H Oxman A D Vist G Kunz R Brozek J Alonso-Coello P GRADE guidelines: 4. Rating the quality of evidence–study limitations (risk of bias) J Clin Epidemiol 2011 64 04 407 415 10.1016/j.jclinepi.2010.07.017 21247734 14 Bygdeman M Swahn M L Replens versus dienoestrol cream in the symptomatic treatment of vaginal atrophy in postmenopausal women Maturitas 1996 23 03 259 263 10.1016/0378-5122(95)00955-8 8794418 15 Loprinzi C L Abu-Ghazaleh S Sloan J A vanHaelst-Pisani C Hammer A M Rowland K M Phase III randomized double-blind study to evaluate the efficacy of a polycarbophil-based vaginal moisturizer in women with breast cancer J Clin Oncol 1997 15 03 969 973 10.1200/jco.1997.15.3.969 9060535 16 Raghunandan C Agrawal S Dubey P Choudhury M Jain A A comparative study of the effects of local estrogen with or without local testosterone on vulvovaginal and sexual dysfunction in postmenopausal women J Sex Med 2010 7 03 1284 1290 10.1111/j.1743-6109.2009.01667.x 20102444 17 Lee Y K Chung H H Kim J W Park N H Song Y S Kang S B Vaginal pH-balanced gel for the control of atrophic vaginitis among breast cancer survivors: a randomized controlled trial Obstet Gynecol 2011 117 04 922 927 10.1097/aog.0b013e3182118790 21422866 18 Loprinzi C L Balcueva E P Liu H Sloan J A Kottschade L A Stella P J A phase III randomized, double-blind, placebo-controlled study of pilocarpine for vaginal dryness: North Central Cancer Treatment group study N04CA J Support Oncol 2011 9 03 105 112 10.1016/j.suponc.2011.02.005 21702402 19 Fernandes T Costa-Paiva L H Pinto-Neto A M Efficacy of vaginally applied estrogen, testosterone, or polyacrylic acid on sexual function in postmenopausal women: a randomized controlled trial J Sex Med 2014 11 05 1262 1270 10.1111/jsm.12473 24612478 20 Hickey M Marino J L Braat S Wong S A randomized, double-blind, crossover trial comparing a silicone- versus water-based lubricant for sexual discomfort after breast cancer Breast Cancer Res Treat 2016 158 01 79 90 10.1007/s10549-016-3865-1 27306420 21 Juliato P T Rodrigues A T Stahlschmidt R Juliato C R Mazzola P G Can polyacrylic acid treat sexual dysfunction in women with breast cancer receiving tamoxifen? Climacteric 2017 20 01 62 66 10.1080/13697137.2016.1258396 27876429 22 Seyyedi F Kopaei M R Miraj S Comparison between vaginal royal jelly and vaginal estrogen effects on quality of life and vaginal atrophy in postmenopausal women: a clinical trial study Electron Physician 2016 8 11 3184 3192 10.19082/3184 28070251 23 Nappi R E Cagnacci A Becorpi A M Nappi C Paoletti A M Busacca M Monurelle Biogel ® vaginal gel in the treatment of vaginal dryness in postmenopausal women Climacteric 2017 20 05 467 475 10.1080/13697137.2017.1335703 28657769 24 Diem S J Guthrie K A Mitchell C M Reed S D Larson J C Ensrud K E Effects of vaginal estradiol tablets and moisturizer on menopause-specific quality of life and mood in healthy postmenopausal women with vaginal symptoms: a randomized clinical trial Menopause 2018 25 10 1086 1093 10.1097/gme.0000000000001131 29738424 25 Mitchell C M Reed S D Diem S Larson J C Newton K M Ensrud K E Efficacy of vaginal estradiol or vaginal moisturizer vs placebo for treating postmenopausal vulvovaginal symptoms: a randomized clinical trial JAMA Intern Med 2018 178 05 681 690 10.1001/jamainternmed.2018.0116 29554173 26 Mitchell C M Guthrie K A Larson J Diem S LaCroix A Z Caan B Sexual frequency and pain in a randomized clinical trial of vaginal estradiol tablets, moisturizer, and placebo in postmenopausal women Menopause 2019 26 08 816 822 10.1097/GME.0000000000001341 30994576 27 Labrie F Archer D F Bouchard C Fortier M Cusan L Gomez J L Intravaginal dehydroepiandrosterone (prasterone), a highly efficient treatment of dyspareunia Climacteric 2011 14 02 282 288 10.3109/13697137.2010.535226 21244215 28 Le Donne M Caruso C Mancuso A Costa G Iemmo R Pizzimenti G The effect of vaginally administered genistein in comparison with hyaluronic acid on atrophic epithelium in postmenopause Arch Gynecol Obstet 2011 283 06 1319 1323 10.1007/s00404-010-1545-7 20577750 29 Grimaldi E F Restaino S Inglese S Foltran L Sorz A Di Lorenzo G Role of high molecular weight hyaluronic acid in postmenopausal vaginal discomfort Minerva Ginecol 2012 64 04 321 329 22728576 30 Zheng T P Sun A J Xue W Wang Y P Jiang Y Zhang Y Efficacy and safety of Cimicifuga foetida extract on menopausal syndrome in Chinese women Chin Med J (Engl) 2013 126 11 2034 2038 23769553 31 Jokar A Davari T Asadi N Ahmadi F Foruhari S Comparison of the hyaluronic acid vaginal cream and conjugated estrogen used in treatment of vaginal atrophy of menopause women: a randomized controlled clinical trial Int J Community Based Nurs Midwifery 2016 4 01 69 78 26793732 32 Balk J L Whiteside D A Naus G DeFerrari E Roberts J M A pilot study of the effects of phytoestrogen supplementation on postmenopausal endometrium J Soc Gynecol Investig 2002 9 04 238 242 10.1177/107155760200900410 33 Oh K J Chae M J Lee H S Hong H D Park K Effects of Korean red ginseng on sexual arousal in menopausal women: placebo-controlled, double-blind crossover clinical study J Sex Med 2010 7 (4 Pt 1):1469 1477 10.1111/j.1743-6109.2009.01700.x 20141583 34 Amato P Young R L Steinberg F M Murray M J Lewis R D Cramer M A Effect of soy isoflavone supplementation on menopausal quality of life Menopause 2013 20 04 443 447 10.1097/gme.0b013e318275025e 23211877 35 Research Group EG Tedeschi C Benvenuti C Comparison of vaginal gel isoflavones versus no topical treatment in vaginal dystrophy: results of a preliminary prospective study Gynecol Endocrinol 2012 28 08 652 654 10.3109/09513590.2011.650764 22316315 36 Lima S M Yamada S S Reis B F Postigo S Galvão da Silva M A Aoki T Effective treatment of vaginal atrophy with isoflavone vaginal gel Maturitas 2013 74 03 252 258 10.1016/j.maturitas.2012.11.012 23312487 37 Lima S M Bernardo B F Yamada S S Reis B F da Silva G M Galvão M A Effects of Glycine max (L.) Merr. soy isoflavone vaginal gel on epithelium morphology and estrogen receptor expression in postmenopausal women: a 12-week, randomized, double-blind, placebo-controlled trial Maturitas 2014 78 03 205 211 10.1016/j.maturitas.2014.04.007 24856055 38 Postigo S Lima S M Yamada S S dos Reis B F da Silva G M Aoki T Assessment of the effects of Tribulus terrestris on sexual function of menopausal women Rev Bras Ginecol Obstet 2016 38 03 140 146 10.1055/s-0036-1571472 26902700 39 de Souza K Z Vale F B Geber S Efficacy of Tribulus terrestris for the treatment of hypoactive sexual desire disorder in postmenopausal women: a randomized, double-blinded, placebo-controlled trial Menopause 2016 23 11 1252 1256 10.1097/gme.0000000000000766 27760089 40 Yaralizadeh M Abedi P Najar S Namjoyan F Saki A Effect of Foeniculum vulgare (fennel) vaginal cream on vaginal atrophy in postmenopausal women: A double-blind randomized placebo-controlled trial Maturitas 2016 84 75 80 10.1016/j.maturitas.2015.11.005 26617271 41 Malakouti J Jabbari F Asghari-Jafarabadi M Javadzadeh Y Farshbaf-Khalili A The impact of ginkgo biloba tablet and aromatherapy inhaler combination on sexual function in females during postmenopausal period: a double-blind randomized controlled trial Int J Women's Health Reprod Sci 2017 5 02 129 136 10.15296/ijwhr.2017.24 42 Suwanvesh N Manonai J Sophonsritsuk A Cherdshewasart W Comparison of Pueraria mirifica gel and conjugated equine estrogen cream effects on vaginal health in postmenopausal women Menopause 2017 24 02 210 215 10.1097/gme.0000000000000742 27749740 43 Ghorbani Z Mirghafourvand M Charandabi S M Javadzadeh Y The effect of ginseng on sexual dysfunction in menopausal women: A double-blind, randomized, controlled trial Complement Ther Med 2019 45 57 64 10.1016/j.ctim.2019.05.015 31331583 44 Palma F Fontanesi F Facchinetti F Cagnacci A Acupuncture or phy(F)itoestrogens vs. (E)strogen plus progestin on menopausal symptoms. A randomized study Gynecol Endocrinol 2019 35 11 995 998 10.1080/09513590.2019.1621835 31142156 45 Labrie F Archer D Bouchard C Fortier M Cusan L Gomez J L Effect of intravaginal dehydroepiandrosterone (Prasterone) on libido and sexual dysfunction in postmenopausal women Menopause 2009 16 05 923 931 10.1097/gme.0b013e31819e85c6 19424093 46 Genazzani A R Stomati M Valentino V Pluchino N Potì E Casarosa E Effect of 1-year, low-dose DHEA therapy on climacteric symptoms and female sexuality Climacteric 2011 14 06 661 668 10.3109/13697137.2011.579649 21942655 47 Labrie F Archer D Bouchard C Fortier M Cusan L Gomez J Lack of influence of dyspareunia on the beneficial effect of intravaginal prasterone (dehydroepiandrosterone, DHEA) on sexual dysfunction in postmenopausal women J Sex Med 2014 11 07 1766 1785 10.1111/jsm.12517 24774442 48 VVA Prasterone Group Archer D F Labrie F Bouchard C Portman D J Koltun W Cusan L Treatment of pain at sexual activity (dyspareunia) with intravaginal dehydroepiandrosterone (prasterone) Menopause 2015 22 09 950 963 10.1097/gme.0000000000000428 25734980 49 VVA Prasterone Group Bouchard C Labrie F Archer D F Portman D J Koltun W Elfassi E Decreased efficacy of twice-weekly intravaginal dehydroepiandrosterone on vulvovaginal atrophy Climacteric 2015 18 04 590 607 10.3109/13697137.2014.992012 25511551 50 Members of the VVA Prasterone Research Group Labrie F Derogatis L Archer D F Koltun W Vachon A Young D Effect of intravaginal prasterone on sexual dysfunction in postmenopausal women with vulvovaginal atrophy J Sex Med 2015 12 12 2401 2412 10.1111/jsm.13045 26597311 51 VVA Prasterone Research Group Labrie F Archer D F Koltun W Vachon A Young D Frenette L Efficacy of intravaginal dehydroepiandrosterone (DHEA) on moderate to severe dyspareunia and vaginal dryness, symptoms of vulvovaginal atrophy, and of the genitourinary syndrome of menopause Menopause 2016 23 03 243 256 10.1097/gme.0000000000000571 26731686 52 Ospemifene Study Group Bachmann G A Komi J O Ospemifene effectively treats vulvovaginal atrophy in postmenopausal women: results from a pivotal phase 3 study Menopause 2010 17 03 480 486 10.1097/gme.0b013e3181c1ac01 20032798 53 Ospemifene Study Group Portman D J Bachmann G A Simon J A Ospemifene, a novel selective estrogen receptor modulator for treating dyspareunia associated with postmenopausal vulvar and vaginal atrophy Menopause 2013 20 06 623 630 10.1097/gme.0b013e318279ba64 23361170 54 Constantine G Graham S Portman D J Rosen R C Kingsberg S A Female sexual function improved with ospemifene in postmenopausal women with vulvar and vaginal atrophy: results of a randomized, placebo-controlled trial Climacteric 2015 18 02 226 232 10.3109/13697137.2014.954996 25252699 55 Portman D Palacios S Nappi R E Mueck A O Ospemifene, a non-oestrogen selective oestrogen receptor modulator for the treatment of vaginal dryness associated with postmenopausal vulvar and vaginal atrophy: a randomised, placebo-controlled, phase III trial Maturitas 2014 78 02 91 98 10.1016/j.maturitas.2014.02.015 24679891 56 Archer D F Goldstein S R Simon J A Waldbaum A S Sussman S A Altomare C Efficacy and safety of ospemifene in postmenopausal women with moderate-to-severe vaginal dryness: a phase 3, randomized, double-blind, placebo-controlled, multicenter trial Menopause 2019 26 06 611 621 10.1097/GME.0000000000001292 30694917 57 Tungmunsakulchai R Chaikittisilpa S Snabboon T Panyakhamlerd K Jaisamrarn U Taechakraichana N Effectiveness of a low dose testosterone undecanoate to improve sexual function in postmenopausal women BMC Womens Health 2015 15 113 10.1186/s12905-015-0270-6 26631095 58 Melisko M E Goldman M E Hwang J De Luca A Fang S Esserman L J Vaginal testosterone cream vs estradiol vaginal ring for vaginal dryness or decreased libido in women receiving aromatase inhibitors for early-stage breast cancer: a randomized clinical trial JAMA Oncol 2017 3 03 313 319 10.1001/jamaoncol.2016.3904 27832260 59 Nazarpour S Simbar M Ramezani Tehrani F Alavi Majd H Effects of sex education and kegel exercises on the sexual function of postmenopausal women: a randomized clinical trial J Sex Med 2017 14 07 959 967 10.1016/j.jsxm.2017.05.006 28601506 60 Nazarpour S Simbar M Majd H A Tehrani F R Beneficial effects of pelvic floor muscle exercises on sexual function among postmenopausal women: a randomised clinical trial Sex Health 2018 15 05 396 402 10.1071/SH17203 30048604 61 Jonasson A F Edwall L Uvnäs-Moberg K Topical oxytocin reverses vaginal atrophy in postmenopausal women: a double-blind randomized pilot study Menopause Int 2011 17 04 120 125 10.1258/mi.2011.011030 22120944 62 Torky H A Taha A Marie H El-Desouky E Raslan O Moussa A A Role of topical oxytocin in improving vaginal atrophy in postmenopausal women: a randomized, controlled trial Climacteric 2018 21 02 174 178 10.1080/13697137.2017.1421924 29347848 63 Cruz V L Steiner M L Pompei L M Strufaldi R Fonseca F L Santiago L H Randomized, double-blind, placebo-controlled clinical trial for evaluating the efficacy of fractional CO2 laser compared with topical estriol in the treatment of vaginal atrophy in postmenopausal women Menopause 2018 25 01 21 28 10.1097/GME.0000000000000955 28763401 64 Politano C A Costa-Paiva L Aguiar L B Machado H C Baccaro L F Fractional CO2 laser versus promestriene and lubricant in genitourinary syndrome of menopause: a randomized clinical trial Menopause 2019 26 08 833 840 10.1097/GME.0000000000001333 31246661 65 Goetsch M F Lim J Y Caughey A B A practical solution for dyspareunia in breast cancer survivors: a randomized controlled trial J Clin Oncol 2015 33 30 3394 3400 10.1200/jco.2014.60.7366 26215946 66 Golmakani N Parnan Emamverdikhan A Zarifian A Sajadi Tabassi S A Hassanzadeh M Vitamin E as alternative local treatment in genitourinary syndrome of menopause: a randomized controlled trial Int Urogynecol J Pelvic Floor Dysfunct 2019 30 05 831 837 10.1007/s00192-018-3698-z 67 Ekin M Yaşar L Savan K Temur M Uhri M Gencer I The comparison of hyaluronic acid vaginal tablets with estradiol vaginal tablets in the treatment of atrophic vaginitis: a randomized controlled trial Arch Gynecol Obstet 2011 283 03 539 543 10.1007/s00404-010-1382-8 20135132 68 Chen J Geng L Song X Li H Giordan N Liao Q Evaluation of the efficacy and safety of hyaluronic acid vaginal gel to ease vaginal dryness: a multicenter, randomized, controlled, open-label, parallel-group, clinical trial J Sex Med 2013 10 06 1575 1584 10.1111/jsm.12125 23574713 69 Vicariotto F DE Seta F Faoro V Raichi M Dynamic quadripolar radiofrequency treatment of vaginal laxity/menopausal vulvo-vaginal atrophy: 12-month efficacy and safety Minerva Ginecol 2017 69 04 342 349 10.23736/S0026-4784.17.04072-2 28608667 70 Leibaschoff G Izasa P G Cardona J L Miklos J R Moore R D Transcutaneous Temperature Controlled Radiofrequency (TTCRF) for the treatment of menopausal vaginal/genitourinary symptoms Surg Technol Int 2016 29 149 159 27608749 71 Scheffers C S Armstrong S Cantineau A E Farquhar C Jordan V Dehydroepiandrosterone for women in the peri- or postmenopausal phase Cochrane Database Syst Rev 2015 1 CD011066 10.1002/14651858.cd011066.pub2 25879093 72 Peixoto C Carrilho C G Barros J A Ribeiro T T Silva L M Nardi A E The effects of dehydroepiandrosterone on sexual function: a systematic review Climacteric 2017 20 02 129 137 10.1080/13697137.2017.1279141 28118059 73 Khera M Testosterone therapy for female sexual dysfunction Sex Med Rev 2015 3 03 137 144 10.1002/smrj.53 27784606 74 Reed B G Bou Nemer L Carr B R Has testosterone passed the test in premenopausal women with low libido? A systematic review Int J Womens Health 2016 8 599 607 10.2147/ijwh.s116212 27785108
PMC009xxxxxx/PMC9708400.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36446563 10.1055/s-0042-1755456 210409 Review Article The Effectiveness of Herbal Medicines on Cyclic Mastalgia: A Systematic Review on Meta-analysis A eficácia dos medicamentos fitoterápicos na mastalgia cíclica: Uma revisão sistemática em metanálisehttp://orcid.org/0000-0002-1114-0813 Mirzaee Firoozeh 1 http://orcid.org/0000-0001-7497-4180 Fakari Farzaneh Rashidi 2 http://orcid.org/0000-0002-6128-8023 Babakhanian Masoudeh 3 http://orcid.org/0000-0002-1749-4743 Roozbeh Nasibeh 4 http://orcid.org/0000-0003-4639-3711 Ghazanfarpour Masumeh 1 1 Department of Midwifery, Nursing Research Center, Kerman University of Medical Sciences, Kerman, Iran 2 Department of Midwifery, School of Medicine, North Khorasan University of Medical Sciences, Bojnurd, Iran 3 Social Determinants of Health Research Center, Semnan University of Medical Sciences, Semnan, Iran 4 Department of Midwifery, University of Medical Sciences, Bandar Abbas, Iran Address for correspondence Masumeh Ghazanfarpour, PhD Department of Midwifery, Nursing Research Center, Kerman University of Medical [email protected] 29 11 2022 10 2022 1 11 2022 44 10 972985 27 10 2021 23 2 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  Different drugs are used to treat mastalgia, such as danazol and bromocriptine, and both are associated with side effects, due to which most of women and healthcare providers are interested in herbal medicines. Therefore we aim to study the effectiveness of phytoestrogens on the severity of cyclic mastalgia. Methods  To carry out the present study, English electronic resources such as the Cochrane Library, ISI Web of Science, Scopus, and PubMed were used systematically and with no time limitation up to February 10, 2020. Results  In total, 20 studies were included in the present meta-analysis. The results of the meta-analysis showed that herbal medicines versus the control group (standard mean difference [SMD] = - 0.585; 95% confidence interval [CI]: - 0.728–- 0.44; heterogeneity; p  = 0.02; I2 = 42%), herbal medicines versus the B group (SMD = - 0.59; 95%CI: - 0.75–- 0.44; heterogeneity; p  = 0.03; I2 = 42%), and its subgroups, such as phytoestrogen (SMD = - 0.691; 95%CI: - 0.82–- 0.55; heterogeneity; p  = 0.669; I2 = 0%), Vitex-agnus-castus (SMD = - 0.642; 95%CI: - 0.84–- 0.44; p  < 0.001; p  = 203; I2 = 32%), flaxseed (SMD = - 0.63; 95%CI: - 0.901–- 0.367; p  = 0.871; I2 = 0%), and evening primrose (SMD= - 0.485; 95%CI:- 0.84–- 0.12; p  = 0.008; heterogeneity; p  = 0.06; I2 = 56%] may have effective and helpful effects on improving cyclic breast mastalgia. Also, chamomile, isoflavone, cinnamon, and nigella sativa significantly reduced mastalgia symptoms. Conclusion  Herbal medicines and their subgroups may have effective and helpful effects on improving cyclic breast mastalgia. The findings of our meta-analysis must be done cautiously because low methodological quality in some evaluated studies of this systematic review. Resumo Objetivo  Diferentes drogas são utilizadas para tratar a mastalgia, como danazol e bromocriptina, e ambas estão associadas a efeitos colaterais, devido aos quais a maioria das mulheres e dos profissionais de saúde está interessada em medicamentos fitoterápicos. Portanto, nosso objetivo no presente estudo é estudar a eficácia dos fitoestrogênios na gravidade da mastalgia cíclica. Métodos  Para a realização do presente estudo, foram utilizados recursos eletrônicos em inglês como a Cochrane Library, ISI Web of Science, Scopus e PubMed, de forma sistemática e sem limitação de tempo até 10 de fevereiro de 2020. Resultados  No total, 20 estudos foram incluídos na presente metanálise. Os resultados da metanálise mostraram que fitoterápicos versus grupo controle (SMD = - 0,585; intervalo de confiança (IC) 95%: - 0,728–- 0,44; heterogeneidade; p  = 0,02; I2 = 42%), fitoterápicos versus grupo B (SMD = - 0,59; IC95%: - 0,75–- 0,44; heterogeneidade; p  = 0,03; I2 = 42%) e seus subgrupos, como fitoestrogênios (SMD = - 0,691; IC95%: - 0,82–- 0,55; heterogeneidade; p  = 0,669; I2 = 0%), Vitex-agnus-castus (SMD = - 0,642; IC95%: - 0,84–- 0,44; p  < 0,001; p  = 203; I2 = 32%), linhaça (SMD = - 0,63; IC95%: - 0,901–- 0,367; p  = 0,871; I2 = 0%) e prímula (SMD = - 0,485; IC95%: - 0,84–- 0,12; p  = 0,008; heterogeneidade; p  = 0,06; I2 = 56%) podem ter efeitos eficazes e úteis na melhora da mastalgia cíclica da mama. Além disso, camomila, isoflavona, canela e Nigella sativa reduziram significativamente a mastalgia. Conclusão  Os medicamentos fitoterápicos e seus subgrupos podem ter efeitos eficazes e úteis na melhora da mastalgia mamária cíclica. Os achados do presente estudo devem ser explantados com atenção devido ao pequeno número de estudos existentes sobre o tema, a maioria dos quais com um tamanho de amostra pequeno. Keywords mastodynia phytoestrogens systematic review herbal medicine Palavras-chave mastodinia fitoestrogênios revisão sistemática fitoterapia ==== Body pmcIntroduction Breast pain may be divided in two major categories: cyclic pain and noncyclic pain. 1 2 Cyclic breast pain exacerbates with the onset of the second half of the menstrual period and alleviates with the onset of menstrual bleeding; it is also distributed bilaterally toward the upper arms and armpits. 3 It may last > 5 days and ∼ 30% of women with mastalgia 4 and 11% of women may suffer from pain for 7 days. There are different studies on the prevalence of cyclic breast pain, which has been reported to range from 30 to 70%. 3 Breast pain may provoke anxiety and concern regarding breast cancer among patients; in turn, this concern may impose a high financial burden on the healthcare system due to unnecessary medical referrals and the performance of various diagnostic procedures, such as mammography and biopsy. 5 It also interferes with daily activities, sexual, physical, and social activities. 6 The etiology of breast pain is still unknown. However, the most accepted etiology is related to disturbance in concentration of estrogen, progesterone, and prolactin and the responsiveness of target organs to these hormones Nutritional and psychological causes, water retention in the body, and body and breast weight gain are considered other causes of cyclic mastalgia. 7 Different methods have been used for decreasing mastalgia. Pharmaceutical treatments include danazol, bromocriptine, and tamoxifen, and nonpharmaceutical treatments include supplements, oils, and herbal medicines. 6 Different studies proved that vitamin E is not effective for mastalgia. Although using drugs is associated with excessive expenses, there are also common side effects that renders them inneficient. 8 Meanwhile, tamoxifen has better therapeutic effects and fewer side effects than danazol, so it is mostly used. 9 The use of medicinal plants and herbal medicines has increased recently. Many studies were carried out on the use of herbal medicines for complications of menopause, dysmenorrhea, premenstrual syndrome, mastalgia, etc. 10 Most women, researchers, and healthcare providers have been interested in herbal medicines and phytoestrogens. Phytoestrogens are some compounds that are similar to 17-β-Sterol in terms of structure and function, or may have some effects similar to estrogens. 11 Phytoestrogens include several groups of compounds such as lignans, isoflavones, and coumestans. 12 There is much research on the effects of phytoestrogens on the severity of cyclic mastalgia. 13 14 15 16 Currently, danazol is used as the only effective treatment licensed for mastalgia associated with side effects. Tamoxifen as a third-line therapy is not currently licensed for breast pain treatment. 17 We have identified new studies that met the inclusion criteria that were not included in the previous systematic reviews. The purpose of the present study was to investigate the effectiveness of phytoestrogens on the severity of cyclic mastalgia. Methods English electronic resources such as ISI Web of Science, Scopus, PubMed, and Cochrane Library were used systematically and with no limitations up to February 10, 2020, in order to carry out the present study. The following keywords were used to find out research articles related to the effects of herbal medicines on cyclic mastalgia: ( Mastalgia ) and ( Complementary treatments OR alternative treatments OR phytomedicine OR herbal treatments OR alternative medicine OR complementary medicine OR Vitex agnus-castus OR chaste OR flaxseed OR isoflavones OR soy OR Matricaria chamomilla OR chamomile OR Nigella Sativa OR Cinnamon . The references of the included articles and review articles on the subject of the present study were also carefully reviewed to complete the search. The search results from these five databases were merged and duplicates were deleted (based on the same title, year of publication, and name of the author). Two authors independently investigated the title and abstract of articles, and the complete articles were extracted and investigated when they found that the subject is related to the purpose of the current research. All clinical trials investigated the effect of oral or topical herbal therapies in the treatment of cyclic or noncyclic mastalgia. The intervention included women receiving herbal medicines as monotherapy or in combination with other chemical or herbal medications. Placebo, herbal medicine, chemical medication, usual care, and no interven tion considered as control group. We also excluded conference papers, review papers, Editor's Notes, letters, case reports, and animal studies. In cases in which several reports from a study appeared to have been published, only one with more complete information was included, and the others were deleted. These cases were identified by controlling the similarity of the team of authors, the center and the period of the study, and the reported statistical results. The selection of related articles was carried out by two independent reviewers within two steps. In the screening phase, the titles were read first, and a decision was made to enter the analysis. In case of any ambiguity in the inclusion of the article, the abstracts were reviewed to match their title and abstract with the inclusion and exclusion criteria. Cases that were suspicious and required to be fully read entered the second stage. In the second phase, the full text of the reviewed articles and the articles that fully complied with the inclusion and exclusion criteria were entered into a systematic review. All included articles, review articles, and references of articles on the study subject were also carefully reviewed to complete the search. The data extraction table was designed by the research team and each article in the present study was reviewed by two independent researchers. The following data were extracted and reported in the table: Authors, country, age of the patients, duration of treatment, number of subjects submitted to the intervention, type of control of the intervention, and assessment tool results ( Chart 1 ). Chart 1 Specifications of the studies included in the present systematic review article Author (year) Type of clinical trial Age (years old) Outcome Intervention (dose and duration of treatment) Comparison (dose and duration of treatment) Duration of follow-up Intensity of mastalgia Assessment tools Inclusion criteria Results Vaziri et al. (2014) 16 Single-blind 20–45 Treatment of mastalgia 180 g of flaxseed for 2 cycles Omega 3 fatty acids (180 mg of eicosapentaenoic acid and 120 mg of docosahexaenoic acid) 3 months Cyclicmastalgia Visual analog scale 181 Flaxseed was more effective in reducing mastalgia Sekhavat et al. (2009) 18 Double-blind 18–40 Treatment of mastalgia 60 drops of Vitagnus daily Placebo 3 months Cyclic or noncyclic mastalgia Visual analog scale 117 Vitagnus reduced mastalgia more than placebo. Saghafi et al. (2018) 15 Double-blind > 18 Treatment of mastalgia 5 drops of chamomile 3 times a day for 2 consecutive months Placebo(distilled water) 2 months Cyclicmastalgia Visual analog scale 55 Chamomile reduced mild to moderate mastalgia. Rajaby Gharaiy et al. (2017) 13 Double-blind 18–40 Treatment of mastalgia 400 mg of cinnamon 3 times a day Placebo 2 months Cyclic mastalgia Cardiff checklist 74 Cinnamon can be effective in reducing the severity of mastalgia in women. Mirmolaei et al. (2017) 14 Triple-blind 15–49 Treatment of mastalgia 10 ml (2 tablespoons) of Nigella sativa syrup Placebo(paraffin oil syrup) 2 months Cyclic mastalgia McGill questionnaire and visual analog scale 72 Nigella sativa syrup reduced pain intensity compared with placebo. Jahdi et al. (2019) 19 Triple-blind 18–50 Treatment of mastalgia 1000 mg evening primrose every 12 hours, 50 mg vitamin B6 every 12 hours, Placebo 1, 2, and 3 months Cyclic mastalgia Visual analog scale 94 B6 and evening primrose have the same therapeutic effects in the treatment of cyclical mastalgia Alvandipour et al. (2011) 9 Double-blind – Treatment of mastalgia Evening primrose 2 g/day and vitamin E 400 mg/day Placebo After 1 month and 6 months Cyclic mastalgia McGill questionnaire 100 women with cyclic mastalgia Evening primrose and vitamin E had a similar effect in the treatment of mastalgia Gateley et al. (1992) 20 Clinical trial > 17 Treatment of mastalgia Danazol 200 mg daily/bromocriptine 1.25 mg daily Evening primrose oil, 3 g/day 2 months Cyclicmastalgia Cardiff checklist 478 women with cyclic mastalgia Danazol was more effective in reducing the severity of mastalgia in women than bromocriptine and evening primrose oil. Blommers et al. (2002) 21 Double-blind clinical trial − Treatment of mastalgia 3 g of evening primrose oil and control oil 3 g of fish oil and control oil 3 and 6 months Cyclic or noncyclic mastalgia Clinical examinations 120 women with cyclicmastalgia Both groups showed a similar reduction in pain. Aydin et al. (2012) 22 Prospective clinical trial 19–54 Treatment of mastalgia Group 1 = vitex agnus castus and group 2 = meloxicam Placebo 3months Cyclic mastalgia Visual analog scale 108 women with cyclicmastalgia for at least 5 days in 1 cycle with normal and high prolactin Vitex-agnus-castus was more effective in reducing mastalgia than meloxicam and placebo. Jaafarnejad et al. (2017) 23 Quasiexperimental clinical trial 18–45 Treatment of mastalgia Group 1 = flaxseed, group 2 = 1000-mg capsules of evening primrose oil daily Vitamin E group, 400-IU capsules 1–2 months Cyclic mastalgia Researcher-made checklist Women with cyclicmastalgia Flaxseed, evening primrose oil, and vitamin E reduced the duration of mastalgia, but this decrease was significant only inin the flaxseed group. Ingram et al. (2002) 24 Double-blind >18 Treatment of mastalgia Isoflavones, 80 and 40 mg daily Placebo 2 months Cyclic mastalgia Breast pain checklist 12 women with cyclicmastalgia Isoflavones could be effective as complementary therapy in the treatment of mastalgia. Mirghafourvand et al. (2016) 2 Double-blind 18–45 Treatment of mastalgia Group 1 = 25 g flaxseed powder and group 2 = 3.2–8.8 mg of Vitagnus daily Placebo 2 months Cyclic mastalgia Cardiff checklist 159 women with cyclical mastalgia Flaxseed and Vitagnus were effective in reducing mastalgia in the short term. Kiliç et al. (2016) 25 Prospective clinical study > 18 Treatment of mastalgia Group 1 = evening primrose oil and group 2= fructus agni casti/reassurance Placebo 3 months Cyclic or noncyclic mastalgia Cardiff checklist 128 women with cyclicmastalgia Fructus agni casti was more effective in reducing mastalgia than evening primrose and placebo. Ataollahi et al. (2015) 26 Triple-blind – Treatment of the symptoms of premenstrual syndrome 400 g wheat germ 3 times a day from the 16th day of the cycle until the next 5 periods Placebo 2 months Cyclic mastalgia Daily Symptom Record 84 women with premenstrual syndrome Wheat germ was effective in treating mastalgia Ghazanfarpour et al. (2011) 27 Double-blind 31 Treatment of the symptoms of premenstrual syndrome Hypericum perforatum (1360-μg hypericin tablets per day) Placebo 2 months Undermine Premenstrual syndrome questionnaire 170 Hypericum perforatum was more effectiveness than placebo Pruthi et al. (2010) 28 Double-blind > 18 Treatment of mastalgia 3 g of evening primrose Placebo 6 months Mastalgia McGill questionnaire 85 women > 18 years old who develop mastalgia for at least 2 cycles 2 weeks before menstruation Evening primrose and vitamin E, either alone or in combination, had a similar effect in the treatment of mastalgia Masumi et al. (2017) 29 Double-blind > 18 years – 1000 mg of evening primrose daily 400 mg of vitamin E daily 60 days Undermine Premenstrual Symptoms Screening Tool 70 women with moderate to severe menstrual syndrome Evening primrose caused a greater decrease in the treatment of premenstrual syndrome symptoms than vitamin E. Pakgohar et al. (2005) 30 Double-blind – Treatment of premenstrual syndrome 60 drops of Hypiran daily 7 days before menstruation for 2 cycles Placebo (60 drops daily 7 days before menstruation in 2 cycles) 2 months Undermine Daily Symptom Record 70 students with at least 5 symptoms of premenstrual syndrome Hypiran was more effective than placebo in treating the symptoms of premenstrual syndrome, including mastalgia. Mirmolaei et al. (2017) 14 Triple-blind 15–49 Treatment of mastalgia Daily Vitagnus (8 cc) Placebo (Oral paraffin) 3 months Cyclic mastalgia McGill questionnaire and visual analog scale 67 women aged 15 to 49 years old with a visual analog scale score > 4 Vitagnus was more effective in reducing mastalgia than placebo. Evaluating the Quality of Articles The Final Jadad scale including three items was used to evaluate the quality of articles. 31 These items was considered in terms of randomization (whether randomization was done and whether it was done appropriately), blinding (whether the trial was blinded and whether it was done appropriately), reporting account of all patients ( Chart 2 ). Chart 2 Assessment of the quality of studies by the Jadad Scale Authors Blinding Randomization Account of all patients Mentions randomization Method: appropriate Method: inappropriate Mention srandomization Method: appropriate Method: inappropriate Vaziri et al. (2014) 16 + + − − − − + Sekhavat et al. (2009) 18 + + − + + − + Saghafi et al. (2018) 15 + + − + + − + Rajaby Gharaiy et al. (2017) 13 + + − + + − + Mirmolaei et al. (2017) 14 + + − + + − + Jahdi et al. (2019) 19 + + − + + − + Alvandipour et al. (2011) 9 + + − − − − + Gateley et al. (1992) 20 + + − + + − − Aydin et al. (2012) 22 − − − − − − − Blommers et al. (2002) 21 + + − + − + + Jaafarnejad et al. (2017) 23 + − − − − − + Ingram et al. (2002) 24 − + + + + − + Mirghafourvand et al. (2016) 2 + + − + + − + Kiliç et al. (2016) 25 + − − − − − − Ataollahi et al. (2015) 26 + + − + − − − Ghazanfarpour et al. (2011) 27 + + − + + − + Pruthi et al. (2010) 28 + + − + + − + Masumi et al. (2017) 29 + + − + + − + Pakgohar et al. (2005) 30 + + − + + − + Mirmolaei et al. (2016) 10 + + − + + − + Statistical Analysis The software Comprehensive Meta-analysis (CMA) version 2 (Biostat Inc. Englewood, NJ, USA) was used to perform the data analysis. The heterogeneity index of studies was determined by the I2 test and the Q Cochran test. According to the results of Higgins et al. 32 , it is considered that values < 25% show low heterogeneity; values between 25 and 75% show moderate heterogeneity; and values > 75% shows high levels of heterogeneity. According to the results of heterogeneity, random was used to report the effect of phytoestrogens if heterogeneity was 25 percent or higher instead of fixed effect. Forest plot was used to demonstrate the results of the meta-analysis in which the size of the squares shows the number of samples of each, and lines drawn on both sides show the 95% confidence interval (CI) for the effects of each study. Results Herbal Medicines versus Control Group The results of the Q Cochran test demonstrate the heterogeneity between the results of the different studies and a random model of meta-analysis was used instead of a fixed model ( p  = 0.02; I2 = 42%). The standardized mean difference (SMD) value between the intervention group and the control group was SMD = - 0.585; 95%CI: - 0.728–- 0.44; heterogeneity; p  = 0.02; I2 = 42%) ( Fig. 1 ), with statistical significance ( p  < 0.001). The findings showed that the severity of the pain was lower in the herbal medicine group in comparison with the control group ( p  < 0.001). 2 9 10 13 14 15 16 18 19 21 22 24 25 26 28 30 Fig. 1 Effects of herbal medicines versus control on mastalgia. The horizontal lines denote the 95% confidence interval; □ point estimate (size of the square corresponds to its weight); ♦, combined overall effect of treatment Herbal Medicines versus Placebo The SMD value between the herbal medicines group and the placebo group was SMD = - 0.59; 95%CI: - 0.75–- 0.44; heterogeneity; p  = 0.03; I2 = 42% ( Fig. 2 ). The heterogeneity between the studies was moderate. Sensitivity analysis was conducted based on the type and severity of mastalgia. 2 9 10 13 14 15 18 19 22 24 25 26 28 30 Fig. 2 Effects of herbal medicines versus placebo on mastalgia. The horizontal lines denote the 95% confidence interval; □ point estimate (size of the square corresponds to its weight); ♦, combined overall effect of treatment The intensity of mastalgia was reported mild, therefore Sensitivity analysis was performed to exclude Saghafi et al. 15 The SMD and heterogeneity did not change after the removal of Saghafi study (SMD = - 0.58; 95%CI: - 0.75–- 0.42; heterogeneity; p  = 0.03; I2 = 44%; random effect model). The second sensitivity analysis was performed to exclude studies that reported both cyclical and noncyclical mastalgia. The SMD values increased from - 0.59 to 0.65, and heterogeneity was slightly reduced to 40% (SMD = - 0.65; 95%CI: - 0.81–- 0.48; heterogeneity; p  < 0.001; I2 = 40%; p  = 0.059; random effect model) ( Fig. 3 ). 2 9 10 13 14 15 19 22 24 26 28 30 Fig. 3 Effects of herbal medicines versus placebo on cyclical mastalgia. The horizontal lines denote the 95% confidence interval; □ point estimate (size of the square corresponds to its weight); ♦, combined overall effect of treatment. Phytoestrogen versus Control The standardized mean difference value between the intervention and control groups was SMD = - 0.691; 95%CI: - 0.82–- 0.55; heterogeneity; p  = 0.669; I2 = 0%) ( Fig. 4 ). 2 10 13 14 15 16 18 22 24 25 26 27 33 The severity of pain was lower in the phytoestrogen group compared with in the control group ( p  < 0.001). Fig. 4 Effects of phytoestrogens on mastalgia. The horizontal lines denote the 95% confidence interval; □ point estimate (size of the square corresponds to its weight); ♦, combined overall effect of treatment. Vitex-agnus-castus versus Control The results of the analysis of Vitex-agnus-castus with five studies 2 10 18 22 25 showed that the severity of pain was lower in the Vitex-agnus-castus group compared with in the control group (SMD = - 0.642; 95%CI: - 0.84–-0.44; p  < 0.001) ( Fig. 5 ). According to the values of the heterogeneity index ( p  = 203; I2 = 32%), it has been found that there is moderate heterogeneity between studies. Sensitivity analysis was done due to detect potential resource in our meta-analysis. Sekhavat et al. study 18 considered as potential resource heterogeneity and removal of this study decreased heterogeneity to 0%. SMD = 0.793) 95 CI: -1.03 to -0.55; P < 0.001; heterogeneity; p = 0.663; I2 = 0%). Fig. 5 Effects of Vitex-agnus-castus on mastalgia. The horizontal lines denote the 95% confidence interval; □ point estimate (size of the square corresponds to its weight); ♦, combined overall effect of treatment. Flaxseed versus Placebo The results of analyzing flaxseed with two studies 10 16 showed that women in the flaxseed group reported significantly less pain than those in the control group (SMD = - 0.63; 95%CI: - 0.901–- 0.367; p  = 0.871; I2 = 0%) ( Fig. 6 ). Fig. 6 Effects of flaxseed on mastalgia. The horizontal lines denote the 95% confidence interval; □ point estimate (size of the square corresponds to its weight); ♦, combined overall effect of treatment. Hypericum Perforatum L. versus Placebo The analysis results showed that the effects of Hypericum perforatum L. and placebo were the same in relieving breast pain (SMD = - 0.16; 95%CI: - 0.41–- 0.08; p  = 0.2; heterogeneity; p  = 0.55; I2 = 0%; fixed effect model; 2 trials) ( Fig. 7 ). 28 30 Fig. 7 Effects of Hypericum perforatum L. on mastalgia. The horizontal lines denote the 95% confidence interval; □ point estimate (size of the square corresponds to its weight); ♦, combined overall effect of treatment. Evening Primrose versus Placebo The analysis results showed that women in the evening primrose group reported significantly less pain than those in the control group (SMD = - 0.485; 95%CI: - 0.84–- 0.12 ; p  = 0.008; heterogeneity; p  = 0.06; I2 = 56%; random effect model) ( Fig. 8 ). 9 19 21 25 28 Sensitivity analysis was conducted due to moderate heterogeneity between studies, and the effect of each study on the final result was evaluated separately. None of the studies had a significant effect on the final result and heterogeneity of the present study. Fig. 8 Effects of evening primrose on mastalgia. The horizontal lines denote the 95% confidence interval; □ point estimate (size of the square corresponds to its weight); ♦, combined overall effect of treatment. Chamomile A significant reduction in pain was observed in both groups (chamomile and placebo) after 2 months ( p  < 0.0001 and p  = 0.048, respectively) compared with baseline and between the two groups ( p  = 0.007). 15 Isoflavone The reduction in pain was 13% for placebo, 44% for 40 mg of isoflavone per day, and 31% for 80 mg per day. There was a significant difference between groups. 24 Cinnamon There was a statistically significant difference between the two groups in the mean pain scores at the end of the 1 st and 2 nd months ( p  < 0.001 and p  = 0.02), meaning that the intensity of the pain at the end of the 1 st and 2 nd months were significantly lower in the intervention group than in the control group. 13 Nigella Sativa A significant reduction in pain was seen in the Nigella Sativa group compared with the placebo group based on the visual analogue scale (VAS) ( p  = 0.002). 10 Evening primrose versus Vitagnus The mean pain decreased significantly in both the evening primrose ( p  = 0.004) and in the vitagnus ( p  < 0.001) groups. Vitagnus was more effective than primrose. The authors did not report a p-value. 33 Vitamin E versus Flaxseed Oil Breast pain decreased significantly in both the vitamin E and flaxseed oil groups during the 1 st and 2 nd months ( p -value among groups < 0.001). However, the mean breast pain was not significantly different between the two groups, which were not different from each other at the end of the 1 st ( p  = 0.54) and 2 nd months ( p  = 0.73). 34 Danazol versus Evening Primrose The overall response with danazol was 76%, in contrast with a 68% response in patients treated with evening primrose. 35 A clinically useful response was observed in 76% cyclical mastalgia and in 36% of those with noncyclical mastalgia treated with danazol, and in 55% of the patients with cyclical mastalgia and in 33% for those with noncyclical mastalgia treated with evening primrose oil. 20 Evening Primrose versus Bromocriptine A clinically useful response was observed in 50% of the patients with cyclical mastalgia and in 24% of those with noncyclical mastalgia treated with bromocriptine, and in 55% of the patients with cyclical mastalgia and in 33% of those with noncyclical mastalgia treated with evening primrose oil. 20 Vitex Agnus Castus with Meloxicam No significant difference was observed between Vitex-agnus-castus and meloxicam. 22 The present study showed that GLA (Efamast) efficacy did not differ from that of placebo fatty acids, regardless of whether or not antioxidant vitamins were present. Flaxseed and Vitex-agnus-castus Patients with mastalgia in both the flaxseed and the Vitex-agnus-castus groups reported a significant decrease in breast pain intensity and breast pain length in comparison with placebo ( p  < 0.01). However, no significant difference was observed between flaxseed and Vitex-agnus-castus in the 1 st and 2 nd months. 2 Discussion Mastalgia is one of the common problems experienced by women worldwide during reproductive period effects. Drugs like tamoxifen, danazol, and bromocriptine were associated with side effects. As a result, it caused both patients and health providers are interested in herbal medicines. 35 36 According to our investigation, the present research is considered the first meta-analysis on clinical trials that studied the effectiveness of herbal medicines and their subgroups on cyclic mastalgia. Three studies were carried out on vitagnus, 2 10 22 25 one study on nigella sativa, 14 one study on cinnamon, 13 one on isoflavones, 24 two studies on Hypericum perforatum L, 28 30 one study on chamomile, 15 five studies on evening primrose, 9 19 21 25 28 one study on isoflavone, 24 one study compared evening primrose with bromocriptine, 20 vitex agnus castus with meloxicam, 22 and flaxseed with Vitex-agnus-castus. 2 The results of the present research demonstrate that phytoestrogen leads to improvement of cyclic mastalgia compared with placebo. 2 10 13 14 15 18 22 24 25 26 Similarly, nigella sativa, chamomile, cinnamon, and red clover may have helpful effects in improving cyclic mastalgia. According to the result of a study, it can be said that chamomile can significantly reduce the severity of cyclic mastalgia compared with placebo. 15 In vitro, chamomile can inhibit both the function of cyclooxygenase and lipoxygenase; consequently, the production of prostaglandins and leukotrienes is inhibited. 37 This plant is also used as antioxidant, analgesic, antiviral, anti-inflammatory, and antiseptic. 38 According to Gharaiy et al. study, cinnamon is more effective than placebo to reduce the severity of breast pain. 13 Cinnamon contains eugenol, a compound that can prevent prostaglandin biosynthesis and also has anti-inflammatory effects. Research on cinnamon pharmacology and toxicology demonstrate no risk in consuming it. 39 Nigella sativa can relieve breast pain from cyclical mastalgia. 14 This finding is consistent with animal models, as the aqueous extract of N. sativa had anti-inflammatory and analgesic antipyretic effects in albino Wistar rats and albino Swiss mice. 40 Thymoquinone is one of the major compounds of N. sativa, 41 with analgesic, 42 anti-inflammatory, 43 antioxidative, 44 and antioxidative stress effects (Bhandari, 2014). Nigella sativa inhibits inflammatory mediators such as prostaglandins and leukotrienes, macrophage function, NK antitumor activity, amends splenocyte proliferation and Th1/Th2 cytokine profile. 45 The physiological effect of nigella sativa is related to its volatile oil, and thymoquinone is considered one of its active elements with anti-inflammatory effect. 46 The process of inflammation would be regulated by lipoxygenase and cyclooxygenase. Inhibiting lipoxygenase and cyclooxygenase processes prevents the metabolism of arachidonic acid and controls the production of prostaglandins and leukotrienes. 47 Our meta-analysis with studies study shows that flaxseed can reduce the severity of cyclic mastalgia. 2 16 Flaxseed inhibited prostaglandin (PGE2), leukotriene-, histamine-, and bradykinin-induced inflammation, and arachidonic acid-induced inflammation. Flaxseed oil also inhibited both the cyclooxygenase and lipoxygenase pathways of the arachidonate metabolism. 48 Vitex-agnus-castus is known as chaste tree. 49 It is a tree with fingered leaves and cylindrical flowers that grows in the Mediterranean region and Eastern Asia; it has brown fruits which smell of pepper. Its extract is used to treat postpartum hemorrhage. 50 This plant contains progestins, essential oils, diterpenoids, iridoids, flavonoids and ketosteroids. 51 Vitex agnus castus contains iridoids, flavonoids, diterpenoids, and progestins, as well as essential and fatty oils in the fruits, flowers, and leaves. 52 The antiprolactin effect of this plant was evident in previous studies and it is more effective in the treatment premenstrual syndrome. 53 The results of the present meta-analysis showed that, compared with placebo, Vitex-agnus-castus can significantly reduce the severity of cyclic mastalgia. 2 10 18 22 25 The exact mechanism of its effects has not been proved yet but it seems that its effect on the hypothalamus-pituitary axis reduces FSH and prolactin levels and also increases the level of LH. 54 The effect of Vitex Agnus- castus extract on the treatment of luteal phase defects due to latent hyperprolactinemia is investigated. The findings of the study showed a decrease in prolactin levels, normalization of shortened luteal phases duration, and elimination of luteal progesterone. 55 Estimated high to moderate heterogeneity was observed between studies in the evening primrose and Vitex-agnus-castus subgroups, which may be related to different amounts of effective ingredients, different ages of the participants, and different mastalgia severity and pattern of (cyclical and noncyclical). Cyclic mastalgia is more frequent in the 2 nd and 3 rd decades of life. 56 Thus, age may have an effect on mastalgia; therefore, to clarify the relationship between age and mastalgia severity, confounders such as age need to be adjusted in future studies. 56 The limited number of studies is a reason for the second limitation of the present study, because it was not possible to evaluate the heterogeneity of studies by metaregression. The third limitation is the level of methodological quality in some evaluated studies of this systematic review was evident. The problems we have observed in these studies were inappropriate methods of randomization and double-blinding and lack of reportingintention-to-treat analysis. It is suggested that future studies future studies should adhere to CONSORT guidelines. The limited number of studies and their small sample sizes are the fourth imitation of the present study, and more studies with larger sample sizes are warranted. The fifth limitation is that some of the studies have a small sample size, and any enhancement in sample size may have changed the results of these studies. Some studies had no placebo group, so they were not included in the present study. Some studies were designed with test and post-test with no control group and were excluded from the present study, so we suggested designing future studies with placebo and control groups. The sixth limitation is that phytoestrogens are divided into four groups (isoflavones, gensitin, dydizin, and glycine). We suggest that future studies investigate the effect on cyclic mastalgia of other phytoestrogen compounds in the aforementioned groups so that their results can help us to reach a better understanding of their function. The seventh limitation is that generalizing the research findings must be done cautiously because all of the studies were conducted in Iran. The length of treatment was short in most studies and therapeutic effects have not been evaluated after discontinuing the drugs. Finally, all phytoestrogens evaluated in the present meta-analysis had positive effects on the severity of mastalgia, but the function of these phytoestrogens was not investigated in any of the included studies. We recommend the investigation of this important issue in future studies. Since phytoestrogens have a positive effect on cyclic mastalgia, we suggest investigating the effect of phytoestrogens on noncyclic mastalgia; further studies with longer duration and with a follow-up phase should be performed in order to investigate the maintenance of their effectiveness. Some studies have reported that such an assurance the women obtain when clinical breast examination can effect on the intensity of cyclical mastalgia 16 the confounding variables should be controlled in future study. Conclusion Due to the important effect of the health of women in their function in the family and in society and to the fact that mastalgia may cause disruption on their activities and also the positive effect of the effectiveness of herbal medicines, this study was peformed to investigate the effectiveness of herbal medicines on the severity of cyclic mastalgia. The findings of this study showed that herbal medicines such as nigella sativa, chamomile, flaxseed, vitex-agnus-castus and red clover can be considered as an effective and helpful method in improving cyclic mastalgia. The findings of the included studies must be interpreted cautiously due to the high level of heterogeneity between studies, the limited number of studies, and their small sample sizes. Conflict of Interests The authors have no conflict of interests to declare. ==== Refs References 1 Hafiz S P Barnes N LP Kirwan C C Clinical management of idiopathic mastalgia: a systematic review J Prim Health Care 2018 10 04 312 323 10.1071/HC18026 31039960 2 Mirghafourvand M Mohammad-Alizadeh-Charandabi S Ahmadpour P Javadzadeh Y Effects of Vitex agnus and Flaxseed on cyclic mastalgia: A randomized controlled trial Complement Ther Med 2016 24 90 95 10.1016/j.ctim.2015.12.009 26860808 3 Shobeiri F Oshvandi K Nazari M Cyclical mastalgia: prevalence and associated determinants in Hamadan City, Iran Asian Pac J Trop Biomed 2016 6 03 275 278 10.1016/j.apjtb.2015.12.008s 4 Ader D N South-Paul J Adera T Deuster P A Cyclical mastalgia: prevalence and associated health and behavioral factors J Psychosom Obstet Gynaecol 2001 22 02 71 76 10.3109/01674820109049956 11446156 5 Sharami S Sobhani A Asgharnia M Shabaani M Prevalence of cyclic mastalgia and it's relation with age, marriage and employment outside the house J Guil Uni Med Sci 2000 9 (33, 34):111 116 6 Hajizadeh K Alizadeh Charandabi S M Hasanzade R Mirghafourvand M Effect of vitamin E on severity and duration of cyclic mastalgia: A systematic review and meta-analysis Complement Ther Med 2019 44 1 8 10.1016/j.ctim.2019.03.014 31126540 7 Olfati F Parsay S Kazemnejad A Farhad M Comparison of two-month and four-month effect of vitamin E on cyclic mastalgia J Inflamm Dis 2006 10 02 60 64 8 Ahmadinezhad M Delfan B Tarahi M J Pouria A Hashemi S M Taheri T M Comparison of efficacy of local piroxicam and diclofenac in benign mastalgiaIran J Surg 2007;15(03): 9 Alvandipour M Tayebi P Alizadeh Navaie R Khodabakhshi H Comparison between effect of evening primrose oil and vitamin E in treatment of cyclic mastalgia JBUMS 2011 13 02 7 11 10 Mirmolaei S Olfatbakhsh A Fallahhosseini H Kazemnejad E Sotodeh A The effect of vitagnus on cyclic breast pain in women of reproductive age JBUMS 2016 18 09 7 13 11 de Oliveira J M Santos D B Cardoso F S Silva M S Boakari Y L Lira S RS Study of reproductive toxicity of Combretum leprosum Mart and Eicher in female Wistar rats Afr J Biotechnol 2013 12 16 2105 2109 10.5897/AJB11.2798 12 Kurzer M S Xu X Dietary phytoestrogens Annu Rev Nutr 1997 17 353 381 10.1146/annurev.nutr.17.1.353 9240932 13 Gharaiy N R Shahnazi M Kia P Y Javadzadeh Y The effect of cinnamon on cyclical breast pain Iran Red Crescent Med J 2017 19 06 e26442 14 Mirmolaei S T Olfatbakhsh A Falah Huseini H Kazemnezhad Leyli E Sotoodeh Moridiani A The effect of Nigella Sativa syrup on the relief of cyclic mastalgia: a triple-blind randomized clinical trial J Hayat 2017 23 01 33 43 15 Saghafi N Rhkhshandeh H Pourmoghadam N Pourali L Ghazanfarpour M Behrooznia A Effectiveness of Matricaria chamomilla (chamomile) extract on pain control of cyclic mastalgia: a double-blind randomised controlled trial J Obstet Gynaecol 2018 38 01 81 84 10.1080/01443615.2017.1322045 29072514 16 Vaziri F Zamani Lari M Samsami Dehaghani A Salehi M Sadeghpour H Akbarzadeh M Comparing the effects of dietary flaxseed and omega-3 Fatty acids supplement on cyclical mastalgia in Iranian women: a randomized clinical trial Int J Family Med 2014 2014 174532 10.1155/2014/174532 25197571 17 Efamast Study Group Goyal A Mansel R E A randomized multicenter study of gamolenic acid (Efamast) with and without antioxidant vitamins and minerals in the management of mastalgia Breast J 2005 11 01 41 47 10.1111/j.1075-122X.2005.21492.x 15647077 18 Sekhavat L Zare Tarzejani T Kholase Zadeh P The effect of vitex agnus-castus on mastalgia in women Iran South Med J 2009 11 02 147 152 19 Jahdi F Tolouei R Neisani Samani L Hashemian M Haghani H Mojab F Effect of evening primrose oil and vitamin b6 on pain control of cyclic mastalgia associated with fibrocystic breast changes: a triple-blind randomized controlled trial Shiraz E Med J 2019 20 05 e81243 10.5812/semj.81243 20 Gateley C A Miers M Mansel R E Hughes L E Drug treatments for mastalgia: 17 years experience in the Cardiff Mastalgia Clinic J R Soc Med 1992 85 01 12 15 1548647 21 Blommers J de Lange-De Klerk E S Kuik D J Bezemer P D Meijer S Evening primrose oil and fish oil for severe chronic mastalgia: a randomized, double-blind, controlled trial Am J Obstet Gynecol 2002 187 05 1389 1394 10.1067/mob.2002.127377a 12439536 22 Aydin I Baltaci D Turkyilmaz S Oncu M Comparison of vitex agnus castus with meloxicam and placebo in treatment of patients with cyclical mastalgia Duzce Med J 2012 14 01 1 5 23 Jaafarnejad F Adibmoghaddam E Emami S A Saki A Compare the effect of flaxseed, evening primrose oil and Vitamin E on duration of periodic breast pain J Educ Health Promot 2017 6 85 10.4103/jehp.jehp_83_16 29114553 24 Ingram D M Hickling C West L Mahe L J Dunbar P M A double-blind randomized controlled trial of isoflavones in the treatment of cyclical mastalgia Breast 2002 11 02 170 174 10.1054/brst.2001.0353 14965665 25 Kiliç MÖ Şen M Içen D The comparison of evening primrose oil, fructus agni casti and reassurance in the treatment of mastalgia Int J Surg Med. 2016 2 02 83 88 26 Ataollahi M Akbari S A Mojab F Alavi Majd H The effect of wheat germ extract on premenstrual syndrome symptoms Iran J Pharm Res 2015 14 01 159 166 25561922 27 Ghazanfarpour M Kaviani M Asadi N Ghaffarpasand F Ziyadlou S Tabatabaee H R Hypericum perforatum for the treatment of premenstrual syndrome Int J Gynaecol Obstet 2011 113 01 84 85 10.1016/j.ijgo.2010.11.007 21315349 28 Pruthi S Wahner-Roedler D L Torkelson C J Cha S S Thicke L S Hazelton J H Vitamin E and evening primrose oil for management of cyclical mastalgia: a randomized pilot study Altern Med Rev 2010 15 01 59 67 20359269 29 Masoumi Z Khalili A Delforooz A Faradmal J Shayan A [Comparison the effect of evening primrose oil and vitamin E on premenstrual syndrome] CMJA 2017 7 02 1931 1943Persian. 30 Pakgohar M Ahmadi M Salehi Surmaghi M Mehran A Akhondzadeh S Effect of Hypericum perforatum L. for treatment of premenstrual syndrome Faslnamah-i Giyahan-i Daruyi 2005 4 15 33 42 31 Jadad A R Moore R A Carroll D Jenkinson C Reynolds D J Gavaghan D J Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996 17 01 1 12 10.1016/0197-2456(95)00134-4 8721797 32 Higgins J P Thompson S G Quantifying heterogeneity in a meta-analysis Stat Med 2002 21 11 1539 1558 10.1002/sim.1186 12111919 33 Seraji A Salehi A Momeni H Kerami A Naeimi N The effects of evening primrose and vitex agnus on pain scale of the women with cyclic mastalgia a clinical trial CMJA 2014 3 04 639 653 34 Godazandeh G Ala S Motlaq T M Sahebnasagh A Bazi A The comparison of the effect of flaxseed oil and vitamin E on mastalgia and nodularity of breast fibrocystic: a randomized double-blind clinical trial J Pharm Health Care Sci 2021 7 01 4 10.1186/s40780-020-00186-4 33407941 35 Parveen S Sarwar G Ali M Channa G A Danazol versus oil of evening primrose in the treatment of mastalgia Pak J Surg 2007 23 01 10 13 36 Ghazanfor R Qureshi U Adil R G Malik S Tariq M Khan J S Comparative study of effectiveness of vitamin E and evening primrose oil for pain relief in moderate cyclical mastalgia The Professional Medical Journal 2019 26 08 1328 1332 10.29309/TPMJ/2019.26.08.3878 37 Srivastava J K Pandey M Gupta S Chamomile, a novel and selective COX-2 inhibitor with anti-inflammatory activity Life Sci 2009 85 (19-20):663 669 10.1016/j.lfs.2009.09.007 19788894 38 Chauhan E S Jaya A Chamomile an ancient aromatic plant-a review J Ayurveda Med Sci. 2017 2 04 251 255 10.5530/jams.2017.2.26 39 Mirabi P Alamolhoda S H Esmaeilzadeh S Mojab F Effect of medicinal herbs on primary dysmenorrhoea- a systematic review Iran J Pharm Res 2014 13 03 757 767 25276177 40 Al-Ghamdi M S The anti-inflammatory, analgesic and antipyretic activity of Nigella sativa J Ethnopharmacol 2001 76 01 45 48 10.1016/s0378-8741(01)00216-1 11378280 41 Qadri S M Mahmud H Föller M Lang F Thymoquinone-induced suicidal erythrocyte death Food Chem Toxicol 2009 47 07 1545 1549 10.1016/j.fct.2009.03.037 19358869 42 Falahieh F M Jafarnejad F Rakhshandeh H Shakeri M T Motavasselian M Comparison of the effects of Nigella sativa and Mefenamic acid on the severity, duration, and systemic symptoms of primary DysmenorrheaMed Sci 2019;23(96): 43 Hadi V Kheirouri S Alizadeh M Khabbazi A Hosseini H Effects of Nigella sativa oil extract on inflammatory cytokine response and oxidative stress status in patients with rheumatoid arthritis: a randomized, double-blind, placebo-controlled clinical trial Avicenna J Phytomed 2016 6 01 34 43 27247920 44 Al-Okaily B N Mohammed R S Al-Mzain K A Khudair K K Effect of flavonoids extracted from black cumin (Nigella sativa) and vitamin E in ameliorating hepatic damage induced by sodium nitrate in adult male rats Iraqi J Vet Med 2012 36 (0E):172 181 10.30539/iraqijvm.v36i0E.413 45 Abd El-Hack M E Alagawany M Farag M R Tiwari R Karthik K Dhama K Nutritional, healthical and therapeutic efficacy of black cumin (Nigella sativa) in animals, poultry and humans Int J Pharmacol 2016 12 03 232 248 10.3923/ijp.2016.232.248 46 Dwita L P Yati K Gantini S N The anti-inflammatory activity of Nigella sativa balm sticks Sci Pharm 2019 87 01 3 10.3390/scipharm87010003 47 Kalus U Pruss A Bystron J Jurecka M Smelakova A Lichius J J Effect of Nigella sativa (black seed) on subjective feeling in patients with allergic diseases Phytother Res 2003 17 10 1209 1214 10.1002/ptr.1356 14669258 48 Kaithwas G Mukherjee A Chaurasia A K Majumdar D K Anti-inflammatory, analgesic and antipyretic activities of Linum usitatissimum L. (flaxseed/linseed) fixed oil Indian J Exp Biol 2011 49 12 932 938 22403867 49 Niroumand M C Heydarpour F Farzaei M H Pharmacological and therapeutic effects of Vitex agnus-castus L.: a review Pharmacogn Rev 2018 12 23 103 114 10.4103/phrev.phrev_22_17 50 Ulukanli Z Çenet M Öztürk B Bozok F Karabörklü S Demirci S C Chemical characterization, phytotoxic, antimicrobial and insecticidal activities of Vitex agnus-castus' essential oil from East Mediterranean Region J Essent Oil-Bear Plants 2015 18 06 1500 1507 10.1080/0972060X.2015.1004125 51 Chhabra G S Kulkarni K S Evaluation of anti-inflammatory activity of Vitex agnus castus leaves. Quantitative analysis of flavonoids as possible active constituents J Pharmacogn Phytochem 2014 3 01 183 189 52 Maltaş Cagil E Uysal A Yildiz S Durak Y Evaluation of antioxidant and antimicrobial activity of Vitex agnus-castus L Fresenius Environ Bull 2010 19 (12b):3094 3099 53 Rafieian-Kopaei M Movahedi M Systematic review of premenstrual, postmenstrual and infertility disorders of vitex agnus castus Electron Physician 2017 9 01 3685 3689 10.19082/3685 28243425 54 Nasri S Oryan S Haeri Rohani A Amin G Yahyavi H The effects of vitex agnus castus L. extract on gonadotrophines and testosterone in male mice Iran Int J Sci 2004 5 01 25 31 55 Milewicz A Gejdel E Sworen H Sienkiewicz K Jedrzejak J Teucher T [Vitex agnus castus extract in the treatment of luteal phase defects due to latent hyperprolactinemia. Results of a randomized placebo-controlled double-blind study] Arzneimittelforschung 1993 43 07 752 756 8369008 56 Giri R Bhandari R Mahato I Poudel M Kumari S Yadav A Descriptive study of breast problems in women who presents to general outpatient of a tertiary hospital in eastern region of Nepal Health Renaiss 2013 11 01 33 37 10.3126/hren.v11i1.7599
PMC009xxxxxx/PMC9708401.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36446559 10.1055/s-0042-1751058 210454 Original Article Human Reproduction and Assisted Fertilization The Addition of Dydrogesterone after Frozen Embryo Transfer in Hormonal Substituted Cycles with Low Progesterone Levels Uso adicional de didrogesterona após transferência de blastocisto em ciclos substituídos com valores baixos de progesteronahttp://orcid.org/0000-0003-3480-8406 Metello Jose 12 http://orcid.org/0000-0002-4869-8554 Tomas Claudia 2 http://orcid.org/0000-0001-8061-187X Ferreira Pedro 12 http://orcid.org/0000-0002-1336-9662 Santos-Ribeiro Samuel 3 1 Hospital Garcia de Orta, Cirma, Almada, Portugal. 2 Ginemed, Ginemed-Lisboa, Lisboa, Portugal. 3 Instituto Valenciano de Infertilidade (IVI-RMA) Lisboa, Lisboa, Portugal. Address for correspondence Jose Metello Av. Torrado da Silva2805-267, [email protected] 29 11 2022 10 2022 1 11 2022 44 10 930937 06 12 2021 18 4 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  To determine whether a rescue strategy using dydrogesterone (DYD) could improve the outcomes of frozen embryo transfer cycles (FET) with low progesterone (P4) levels on the day of a blastocyst transfer. Methods  Retrospective cohort study including FET cycles performed between July 2019 and October 2020 following an artificial endometrial preparation cycle using estradiol valerate and micronized vaginal P4 (400 mg twice daily). Whenever the serum P4 value was below 10 ng/mL on the morning of the planned transfer, DYD 10 mg three times a day was added as supplementation. The primary endpoint was ongoing pregnancy beyond 10 weeks. The sample was subdivided into two groups according to serum P4 on the day of FET: low (< 10 ng/mL, with DYD supplementation) or normal (above 10 ng/mL). We performed linear or logistic generalized estimating equations (GEE), as appropriate. Results  We analyzed 304 FET cycles from 241 couples, 11.8% (n = 36) of which had serum P4 below 10 ng/mL on the FET day. Baseline clinical data of patients was comparable between the study groups. Overall, 191 cycles (62.8%) had a biochemical pregnancy, of which 131 (44,1%) were ongoing pregnancies, with a 29,8% miscarriage rate. We found no statistically significant differences in the hCG positive (63 vs 64%) or ongoing pregnancy rates (50 vs 43,3%) between those FETs with low or normal serum P4 values, even after multivariable logistic regression modelling. Conclusion  Our results indicate that DYD 10 mg three times a day administered in women who perform FET with P4 serum levels < 10 ng/mL, allows this group to have pregnancy rates beyond 12 weeks at least as good as those with serum levels above 10 ng/mL. Resumo Objetivo  Determinar se uma estratégia de resgate usando didrogesterona (DYD) pode melhorar os resultados dos ciclos de transferência de embriões congelados (TEC) com baixos níveis de progesterona (P4) no dia de uma transferência de blastocisto. Métodos  Estudo de coorte retrospectivo que incluiu ciclos TEC realizados entre julho de 2019 e outubro de 2020 após um ciclo de preparação endometrial artificial usando valerato de estradiol e P4 vaginal micronizado (400 mg duas vezes ao dia). Sempre que o valor de P4 sérico estava abaixo de 10 ng/mL na manhã da transferência planejada, adicionou-se 10 mg de DYD tri-diário como suplementação. O desfecho primário foi gravidez evolutiva após 10 semanas. A amostra foi subdividida em dois grupos de acordo com o P4 sérico no dia da TEC: baixo (< 10 ng/mL, com suplementação de DYD) ou normal (acima de 10 ng/mL). Realizamos equações de estimativa generalizada linear ou logística (GEE), conforme apropriado. Resultados  Analisaram-se 304 ciclos de FET de 241 casais, dos quais 11,8% (n = 36) tinham valores de P4 sérico abaixo de 10 ng/mL no dia da TEC. Os dados clínicos e demográficos dos pacientes eram comparáveis entre os grupos. Globalmente, 191 ciclos (62,8%) tiveram uma gravidez bioquímica, dos quais 131 (44,1%) foram gestações em curso, com uma taxa de aborto espontâneo de 29,8%. Não encontramos diferenças estatisticamente significativas na taxa de gravidez bioquímica (63 vs. 64%) ou nas taxas de gravidez evolutiva (50 vs. 43,3%) entre TEC com valores séricos de P4 baixos ou normais, mesmo após modelação com regressão logística multivariável. Conclusão  Nossos resultados indicam que a suplementação com DYD 10 mg três vezes ao dia em mulheres com níveis séricos de P4 abaixo de 10 ng/mL em ciclos de TEC substituídos parecem conseguir resultados pelo menos tão bons como nos ciclos com valores superiores para taxas de gravidez em curso além de 12 semanas. Keywords dydrogesterone frozen embryo transfer artificial cycle assisted reproductive technology outcomes luteal phase Descritores didrogesterona transferência de embriões congelados ciclo artificial sucesso em procriação medicamente assistida fase lútea ==== Body pmcIntroduction Frozen embryo transfers (FETs) have increased considerably over the past decade. 1 Endometrial preparation in FETs can be achieved in natural, modified, or artificial cycles, but up to now there is no recommended evidence to support one protocol over another. 2 3 4 Nonetheless, some evidence is building on the benefits of the corpus luteum, especially its influence on neonates' size, and its ability to reduce hypertensive disorders in pregnancy. 5 6 7 In hormone-substituted cycles for FET, the administration of exogenous progesterone (P4) after estrogen priming plays a fundamental role in the secretory transformation of the endometrium, given the absence of any endogenous production of P4 by the corpus luteum, 2 or placenta, until at least the 7 th to 9 th week of gestation. 8 9 In FET cycles, estradiol is often given in patches, oral, or vaginal pills, while exogenous P4 may be administered orally, vaginally, subcutaneously, or intramuscularly. The vaginal and intramuscularly administration have both been extensively studied, with neither showing superiority over the other in terms of pregnancy outcomes. 10 Although the vaginal route is often preferred, 11 12 many have advocated the need for serum P4 monitoring during these FET cycles given that the absorption and metabolization in each patient is very variable. 13 14 15 More recently, the oral route has also been evaluated as a potentially noninferior alternative, at least when used in fresh embryo transfers. 12 16 17 During the last years, several retrospective and prospective studies have demonstrated a relationship between low serum P4 levels on the day of embryo transfer and a reduction in ongoing and delivery rates. 13 14 18 19 20 21 22 23 In such cases, researchers have advocated the use of enhanced P4 “rescue” strategies, although evidence of the benefit of such an approach is currently lacking. Dydrogesterone (DYD) is an orally administered synthetic molecule. Despite having a bioavailability lower than 10% following first hepatic passage, 15 it is still significantly more active than orally administered micronized P4, given its high selectivity for the P4 receptor, 24 25 allowing it to be used in much lower doses with lesser side effects. 12 26 The active metabolite of DYD is 20-alpha-dihydrodydrogesterone, which has a half-life of up to 17 hours, with no additional agonist or antagonist activity in other receptors. That said, the main drawback of this drug is the difficulty of monitoring serum 20-alpha-dihydrodydrogesterone levels. 8 We sought to determine whether a rescue strategy with DYD starting immediately after a blastocyst transfer could improve the outcome of FET cycles with low P4 values on the day of a blastocyst transfer. Methods This was a retrospective study conducted at the Centro de Infertilidade e Reprodução Medicamente Assistida (CIRMA) between July 2019 and October 2020. We included only FET cycles in women aged between 18 and 40 years who transferred 1 or 2 frozen blastocysts with an expansion degree equal to or greater than 2 and with a grade 1 or 2 internal cell mass and trophectoderm (Istanbul Consensus, 2011) 27 in a hormone-substituted cycle. Patients with an endometrial thickness below 6 mm prior to P4 administration, untreated endocavitary disease, uncorrected Müllerian anomaly, or those who were demonstrated to have serum P4 values below 3 ng/mL on the day of embryo transfer (n = 2) were excluded from analysis. The study was approved by the local Institutional Review Board. On the second day of a spontaneous or post-pill menstrual cycle, patients started estradiol (Zumenon, Bayer Portugal, SA., Carnaxide, Portugal) at a dose of 2 mg every 12 hours vaginally, with the first transvaginal ultrasound control being performed 12 to 20 days later. If the endometrial thickness was over 6 mm, with resting ovaries, patients started vaginal administration of P4 (Progeffik, Laboratórios Effik, Algés, Portugal) at a dose of 400 mg every 12 hours starting on the following morning. Whenever necessary, luteinizing hormone (LH) and P4 serum assessments were performed in order to exclude the occurrence of spontaneous ovulation prior to starting exogenous P4 administration. Another serum P4 assay was performed on the morning of the transfer after the 11 th vaginal P4 administration. Whenever the P4 value was below 10 ng/mL, 10 mg of DYD (Duphaston, BGP products LTD., Lisboa, Portugal) three times a day was started and maintained until at least 8 weeks of pregnancy. The cut-off used for DYD administration was set according to a previous evaluation of our own population, which corresponds to 11.5% of the pool of the cycles. 28 In this cohort women with lower progesterone levels had a trend towards lower delivery rates (26 vs. 39%; p  > 0.05). A maximum of two good quality blastocysts (ALPHA group) were warmed according to the following protocol: the Cryotop straw (Kitazato Corp., Shizuoka, Japan) was removed from liquid nitrogen and immediately submerged in 300µl thawing solution (Kitazato Corp., Shizuoka, Japan), previously heated to 37˚C. After 1 minute, the embryos were placed in a 60µl drop of diluent solution (Kitazato Corp., Shizuoka, Japan) for 3 minutes at room temperature. Finally, the embryos were placed in a 60µl drop of washing solution (Kitazato Corp., Shizuoka, Japan) for 5 minutes at room temperature, and then washed for 1 minute in another drop of 60µl washing solution at room temperature. They were then placed into 30µl drops of Sequential Blast medium (CooperSurgical Fertility Solutions, Måløv, Denmark) and covered with Liquid Paraffin (CooperSurgical Fertility Solutions, Måløv, Denmark), where they remained for at least 2 hours prior to transfer. Embryo transfers were routinely performed under ultrasound guidance using either Cook or Wallace embryo catheters introduced until passing the middle of the endometrial cavity, where the embryos were later deposited. The β-HCG test was performed 9 to 12 days after the transfer and, in those who conceived, estradiol and DYD was maintained until at least until the 8 th week of pregnancy and vaginal progesterone until the 10 th week. The P4 hormone assays were performed using the electrochemiluminescence (ECLIA) and the Cobas 8000 Roche Diagnostics (Hoffmann–La Roche AG., Switzerland) equipment in the morning (between 9–11 am ). A hCG positive pregnancy was diagnosed in all cycles with a serum β-HCG value > 10 IU/L. Meanwhile, an ongoing pregnancy was defined as the presence of at least one embryo with heart beating detected after the 12 th week of pregnancy. Miscarriage was defined as hCG positive pregnancy that did not go beyond the 12 th week. Cycles with progesterone values below 10 ng/mL were compared with those above. The following variables were subject to comparative statistical analysis between the study groups: female and male age on the day of the oocyte retrieval, female age on the FET day, infertility duration, female weight, female body mass index (BMI), female and male smoking habits, female and male ethnicity, serum anti-Müllerian hormone (AMH) values, antral follicle count (AFC), total dose of gonadotropins used in the IVF/ICSI cycle, number of oocytes collected and used, number of embryos obtained, blastocyst development day (D5 or D6), number of transferred embryos, FET rank, endometrial thickness prior to FET, and serum P4 value on the day of FET. Using the Statistical Package Social Sciences (SPSS, IBM Corp., Armonk, NY, USA) software version 22.0, we performed linear or logistic generalized estimating equations (GEE), as appropriate, in order to account for the inclusion of more than one FET cycle performed by the same patient. 29 A p -value below < 0.05 was considered statistically significant. Finally, for the outcome ongoing pregnancy, we also performed multivariable GEE in order to account for potential confounding factors, with all variables in which the univariable model presented a p  < 0.15 being included in the final multivariable model. Results Overall, a total of 304 cycles from 241 couples were included. There was a total of 191 (62.8%) pregnancies, of which 131 (44.1%) were ongoing pregnancies, with a 29.8% miscarriage rate. Among these, 36 (11.5%) had P4 levels below 10 ng/mL on the day of FET and, hence, were administered DYD 10 mg three times a day in addition to vaginal micronized P4. These were compared with cycles with progesterone ≥ 10 ng/mL. Comparison of the Cycles with and without DYD Supplementation Table 1 shows the baseline clinical data of patients according to serum progesterone levels on the day of embryo transfer. Overall, women with serum P4 levels < 10 ng/mL weighed more. There were no other statistically significant differences between the study groups amongst the other variables evaluated. Table 1. Baseline characteristics of both groups according to the progesterone value on the day of embryo transfer ≥ or < than 10 ng/mL Progesterone ≥ 10 ng/mL (n = 268) Progesterone < 10 ng/mL (n = 36) p -value Female age at pickup (years), mean ± SD 34.1 ± 3.6 33.8 ± 3.5 0.64 Female age at FET (years), mean ± SD 35.0 ± 3.6 34.4 ± 3.6 0.40 Infertility duration (months), mean ± SD 60.1 ± 30.2 (n = 265)* 65.5 ± 34.0 0.40 Endometrium thickness (mm), mean ± SD 9.7 ± 1.8 (n = 253)* 9.8 ± 1.7 (n = 30)* 0.79 Female weight (kg), mean ± SD 63.2 ± 12.7 69.5 ± 14.4 0.03 Female BMI (kg/m 2 ), mean ± SD 24.1 ± 4.7 25.8 ± 5.5 0.10 Male age at pickup (years), mean ± SD 36.0 ± 4.5 (n = 265)* 37.5 ± 7.4 0.30 AMH (ng/mL), mean ± SD 3.7 ± 3.7 (n = 264)* 4.3 ± 4.2 (n = 35)* 0.42 AFC, mean ± SD 18.7 ± 9.8 (n = 263)* 20.2 ± 11.8 0.49 Gonadotrophins dose (UI), mean ± SD 2606.3 ± 814.5 2626.9 ± 793.5 0.88 Oocytes collected, mean ± SD 15.5 ± 8.2 15.8 ± 7.0 0.82 Oocytes used, mean ± SD 14.3 ± 7.4 14.8 ± 6.6 0.66 2PN, mean ± SD 9.8 ± 5.7 8.9 ± 4.7 0.30 Female smoking 0.06  Never 144 (53.7%) 26 (72.2%)  Current/past 124 (46.3%) 10 (27.8%) Male smoking 0.69  Never 108 (40.8%) 16 (44.4%)  Current/past 157 (59.2%) 20 (55.6%) Female race 0.54  Non-Caucasian 32 (11.9%) 3 (8.3%)  Caucasian 236 (88.1%) 33 (91.7%) Male race  Non-Caucasian 29 (10.9%) 3 (8.3%) 0.65  Caucasian 236 (89.1%) 33 (91.7%) FET rank 0.71  1 180 (67.2%) 25 (69.4%)  2 67 (25.0%) 9 (25.0%)  3 15 (5.6%) 2 (5.6%)  4 5 (1.9%) 0 (0.0%)  5 1 (0.4%) 0 (0.0%) Transfer day  Day 5 216 (80.6%) 30 (83.3%) 0.70  Day 6 52 (19.4%) 6 (16.7%) Number of embryos transferred 0.81  1 213 (79.5%) 28 (77.8%)  2 55 (20.5%) 8 (22.2%) Abbreviations: 2PN, two-pronuclear zygote; AFC, antral follicle count; AMH, anti-Müllerian hormone; BMI, body mass index; FET, frozen embryo transfer; SD, standard deviation. Notes : * Excluded cases with missing values. Clinical Outcomes According to the Progesterone Values and DYD administration Figure 1 shows the outcomes according to the groups: progesterone < 10 ng/mL with DYD addition versus progesterone > 10 ng/mL. No statistically significant differences were found concerning biochemical pregnancy (64 vs 63%; p  = 0.889), ongoing pregnancy (50 vs 43%; p  = 0.446), and miscarriage (22 vs 31%; p  = 0.365) rates. Fig. 1. Outcomes according to the groups: progesterone < 10 ng/mL with addition of DYD three times a day versus progesterone > 10 ng/mL. Clinical Outcomes According to Progesterone Distribution The P4 serum concentration on the day of the embryo transfer varied between 4.4 and 32.3 ng/mL, with an average of 14.7 ng/mL. Figure 2 shows a sensitivity analysis with the distribution and outcomes according to different P4 intervals (presented in regular intervals of 2 ng/mL). Fig. 2. Outcomes according to the percentile of progesterone distribution. Taking into account the sensitivity analysis, we decided to subdivide the group with P4 ≥ 10 ng/mL into an upper and lower group. For this, we used the cut-off on 20% (12.2 ng/mL) of the sample ( Table 2 ). Cycles with a lower level of serum progesterone had significantly lower hCG positive pregnancy rates. Finally, we also performed stepwise logistic GEE regression using all the variables with a p -value below 0.15 in the univariable analysis for the outcomes ongoing pregnancy rate: AMH, transfer day, number of embryos transferred, plus female age at pickup, and progesterone value below or above 10 ng/mL ( Table 3 ). In this final model, only the number of transferred embryos was significantly associated with ongoing pregnancy rate. More importantly, we also found that the use of DYD and a serum progesterone level below 10 ng/mL on the day of embryos transfer was not associated with the outcome in the adjusted model. Table 2. Sub-group evaluation P4 10–12.2 ng/mL (n = 53) P4 12.3–32.3 ng/mL (n = 215) p -value βHCG + 49.1% 66.0% 0.026 Ongoing pregnancy 34.0% 45.6% 0.130 Miscarriage 30.8% 31.0% 0.181 Notes: The 20% cycles with lower serum P4 values that did not have the addition of DYD. Table 3. Adjusted odds ratio for clinical pregnancy rate Univariable Multivariable OR 95% CI p -value OR 95% CI p- value Female age at pickup 0.952 (0.895–1.014) 0.125 0.948 (0.886–1.014) 0.121 AMH 1.054 (0.991–1.120) 0.092 1.051 (0.986–1.121) 0.125 Transfer day (D5 vs. D6) 1.776 (0.978–3.22) 0.059 1.635 (0.879–3.041) 0.120 Number of embryos transferred (2 vs. 1) 2.578 (1.457–4.456) 0.001 2.708 (1.510–4.856) 0.001 P4 < 10 ng/ml + DYD 1.304 (0.652–2.607) 0.453 1.163 (0.560–2.418) 0.685 Abbreviations: AMH, anti-Müllerian hormone; CI, confidence interval; DYD, dydrogesterone; OR, odds ratio. Discussion There seems to be a relationship between circulating P4 in the peri-implantation period and the transfer outcome. Accumulating evidence suggest a negative relation between lower levels of P4 and treatment outcome. 13 14 18 19 23 24 25 26 27 28 29 We sought to determine whether a rescue strategy using DYD starting immediately after a blastocyst transfer could improve the outcome of cycles with low P4 values on the day of a blastocyst transfer. Our results seem to demonstrate that women with lower levels of plasmatic P4 on the day of transfer of a blastocyst can indeed be rescued with the DYD. A similar rescue strategy concept had already been presented by Cédrin-Durnerin et al. 23 using vaginal P4 (cut-off of 10 ng/mL), and by Brady et al. 20 with intramuscular P4 (cut-off of 20 ng/mL). In both cases, the measurement of serum P4 was performed on the day of the transfer, and the patient was supplemented with an increase in the dose of the medication already on course, if the values were below the cut-offs established in the respective cohort. Both studies failed to show any improvement in terms of pregnancy outcomes. Recently, a retrospective study assessing 599 cycles concluded that hCG positive and clinical pregnancy rates were comparable between DYD three times a day and micronized vaginal P4, 30 but no sub-analysis regarding the serum progesterone values was performed. Conversely, we opted to use DYD after taking into consideration several clinical trial that evaluated the efficacy of its oral form in luteal phase support, both in fresh and frozen transfers. 11 16 17 26 31 32 33 34 35 36 37 The cut-off used for DYD administration was set according to a previous evaluation of our own population which corresponds to 11.5% of the pool of the cycles. 28 However, when compared to other series, 18 19 the P4 distribution in our population seems to be much higher. For this reason, we decided to assess what occurred in the 20% lower cycles with P4 above 10 ng/mL and we concluded that this group presented a statistically significant lower rate of hCG positive pregnancies. For this reason, we believe that center-specific rescue strategy, based on a percentile distribution of serum P4, might be better than an arbitrary threshold derived from other publications. According to our data, we believe that FET cycles in the 30% lower level of P4 might benefit from this strategy. Two recent publications support our conclusions. Specifically, Álvarez et al. 38 reported that the addition of progesterone injections whenever serum P4 levels (measured the day prior to an euploid blastocyst transfer) were < 10.6 ng/mL was an effective rescue strategy in terms of clinical, ongoing pregnancy rate and live birth rates. Furthermore, Vuong et al. 39 also recently reported on a prospective open-label nonrandomized study with two groups of patients. One group performed luteal support with vaginal micronized progesterone 400 mg twice daily plus oral DYD 10 mg twice daily (732 patients), while the other had only vaginal micronized progesterone 400 mg twice daily (632 patients). The researchers concluded that live birth rates were statistically significantly better in the vaginal progesterone plus DYD group (46.3 vs. 41.3%). Concerning safety issues on the use of DYD in early pregnancy, the literature is sparse. A review of 28 cases of potential links between maternal DYD use during pregnancy and congenital birth defects was reported, 40 but no conclusions could be drawn. Although Mahmoud et al. 41 found a positive association between its use in embryonic phases and congenital heart disease in the fetus, in the LOTUS I trial 16 the rates of adverse effects associated with treatment, including congenital, familial, and genetic effects were similar between the DYD and micronized vaginal P4 group (1.0% DYD vs. 1.2% MVP). A recent systematic review on the assessment of the evidence on the efficacy and safety of oral DYD versus micronized vaginal P4 for luteal phase support reported an overall similar incidence of congenital, familial, and genetic disorders between both groups. 42 The strengths of our study include the fact that that data come from a single center, during a short period of time (16 months) and that all the patients went through the same protocol. Furthermore, all serum P4 assessments were performed in the same laboratory during the morning. Conversely, the interpretation of our results may be limited by the fact that this is still a retrospective study presenting some heterogeneity between groups in terms of patient and cycle characteristics. According to our results transferring 1 or 2 embryos and transferring a day 5 or day 6 embryo might be related to the outcome. However, we tried to control for these variables, with both groups having a quite similar distribution of both these variables ( Table 1 ). Conclusion In conclusion, DYD 10 mg three times a day, administered in women who perform a FET with P4 serum levels below 10 ng/mL, allows this group to have pregnancy rates beyond 12 weeks at least as good as those with serum levels above 10 ng/mL. Contributions Conflict of interests The authors have no conflict of interests to declare. Jose Metello: conceived the presented idea; retrieved the data; performed the computations. Claudia Tomas: developed the theoretical formalism. Pedro Ferreira: retrieved the data. Samuel Santos-Ribeiro: conceived the presented idea; performed the computations; supervised the findings of this work. All authors discussed the results and contributed to the final manuscript. ==== Refs References 1 European IVF-monitoring Consortium (EIM) for the European Society of Human Reproduction and Embryology (ESHRE) De Geyter C Calhaz-Jorge C Kupka M S Wyns C Mocanu E Motrenko T ART in Europe, 2014: results generated from European registries by ESHRE: The European IVF-monitoring Consortium (EIM) for the European Society of Human Reproduction and Embryology (ESHRE) Hum Reprod 2018 33 09 1586 1601 10.1093/humrep/dey242 30032255 2 Ghobara T Gelbaya T A Ayeleke R O Cycle regimens for frozen-thawed embryo transfer Cochrane Database Syst Rev 2017 7 07 CD003414 10.1002/14651858.CD003414.pub3 28675921 3 Groenewoud E R Cohlen B J Macklon N S Programming the endometrium for deferred transfer of cryopreserved embryos: hormone replacement versus modified natural cycles Fertil Steril 2018 109 05 768 774 10.1016/j.fertnstert.2018.02.135 29778369 4 Yarali H Polat M Mumusoglu S Yarali I Bozdag G Preparation of endometrium for frozen embryo replacement cycles: a systematic review and meta-analysis J Assist Reprod Genet 2016 33 10 1287 1304 10.1007/s10815-016-0787-0 27549760 5 Jing S Li X F Zhang S Gong F Lu G Lin G Increased pregnancy complications following frozen-thawed embryo transfer during an artificial cycle J Assist Reprod Genet 2019 36 05 925 933 10.1007/s10815-019-01420-1 30924053 6 Ginström Ernstad E Wennerholm U B Khatibi A Petzold M Bergh C Neonatal and maternal outcome after frozen embryo transfer: Increased risks in programmed cycles Am J Obstet Gynecol 2019 221 02 1260 1.26E20 10.1016/j.ajog.2019.03.010 7 Roque M Bedoschi G Cecchino G N Esteves S C Fresh versus frozen blastocyst transfer Lancet 2019 394 (10205):1227 1228 10.1016/S0140-6736(19)31393-5 8 Neumann K Depenbusch M Schultze-Mosgau A Griesinger G Characterization of early pregnancy placental progesterone production by use of dydrogesterone in programmed frozen-thawed embryo transfer cycles Reprod Biomed Online 2020 40 05 743 751 10.1016/j.rbmo.2020.01.019 32336650 9 Csapo A I Pulkkinen M O Wiest W G Effects of luteectomy and progesterone replacement therapy in early pregnant patients Am J Obstet Gynecol 1973 115 06 759 765 10.1016/0002-9378(73)90517-6 4688578 10 van der Linden M Buckingham K Farquhar C Kremer J A Metwally M Luteal phase support for assisted reproduction cycles Cochrane Database Syst Rev 2015 07 CD009154 10.1002/14651858.CD009154.pub3 26148507 11 Casper R F Yanushpolsky E H Optimal endometrial preparation for frozen embryo transfer cycles: window of implantation and progesterone support Fertil Steril 2016 105 04 867 872 10.1016/j.fertnstert.2016.01.006 26820769 12 Griesinger G Tournaye H Macklon N Petraglia F Arck P Blockeel C Dydrogesterone: pharmacological profile and mechanism of action as luteal phase support in assisted reproduction Reprod Biomed Online 2019 38 02 249 259 10.1016/j.rbmo.2018.11.017 30595525 13 Labarta E Mariani G Holtmann N Celada P Remohí J Bosch E Low serum progesterone on the day of embryo transfer is associated with a diminished ongoing pregnancy rate in oocyte donation cycles after artificial endometrial preparation: a prospective study Hum Reprod 2017 32 12 2437 2442 10.1093/humrep/dex316 29040638 14 Yovich J L Conceicao J L Stanger J D Hinchliffe P M Keane K N Mid-luteal serum progesterone concentrations govern implantation rates for cryopreserved embryo transfers conducted under hormone replacement Reprod Biomed Online 2015 31 02 180 191 10.1016/j.rbmo.2015.05.005 26099447 15 Nahoul K Dehennin L Jondet M Roger M Profiles of plasma estrogens, progesterone and their metabolites after oral or vaginal administration of estradiol or progesterone Maturitas 1993 16 03 185 202 10.1016/0378-5122(93)90064-o 8515718 16 Tournaye H Sukhikh G T Kahler E Griesinger G A Phase III randomized controlled trial comparing the efficacy, safety and tolerability of oral dydrogesterone versus micronized vaginal progesterone for luteal support in in vitro fertilization Hum Reprod 2017 32 05 1019 1027 10.1093/humrep/dex023 28333318 17 Rashidi B H Ghazizadeh M Nejad E ST Bagheri M Gorginzadeh M Oral dydrogesterone for luteal support in frozen-thawed embryo transfer artificial cycles: a pilot randomized controlled trial Asian Pac J Reprod 2016 5 06 490 494 10.1016/j.apjr.2016.10.002 18 Gaggiotti-Marre S Martinez F Coll L Garcia S Álvarez M Parriego M Low serum progesterone the day prior to frozen embryo transfer of euploid embryos is associated with significant reduction in live birth rates Gynecol Endocrinol 2019 35 05 439 442 10.1080/09513590.2018.1534952 30585507 19 Labarta E Mariani G Paolelli S Rodriguez-Varela C Vidal C Giles J Impact of low serum progesterone levels on the day of embryo transfer on pregnancy outcome: a prospective cohort study in artificial cycles with vaginal progesterone Hum Reprod 2021 36 03 683 692 10.1093/humrep/deaa322 33340402 20 Brady P C Kaser D J Ginsburg E S Ashby R K Missmer S A Correia K F Serum progesterone concentration on day of embryo transfer in donor oocyte cycles J Assist Reprod Genet 2014 31 05 569 575 10.1007/s10815-014-0199-y 24619510 21 Basnayake S K Volovsky M Rombauts L Osianlis T Vollenhoven B Healey M Progesterone concentrations and dosage with frozen embryo transfers - What's best? Aust N Z J Obstet Gynaecol 2018 58 05 533 538 10.1111/ajo.12757 29271471 22 Alsbjerg B Thomsen L Elbaek H O Laursen R Povlsen B B Haahr T Progesterone levels on pregnancy test day after hormone replacement therapy-cryopreserved embryo transfer cycles and related reproductive outcomes Reprod Biomed Online 2018 37 05 641 647 10.1016/j.rbmo.2018.08.022 30385142 23 Cédrin-Durnerin I Isnard T Mahdjoub S Sonigo C Seroka A Comtet M Serum progesterone concentration and live birth rate in frozen-thawed embryo transfers with hormonally prepared endometrium Reprod Biomed Online 2019 38 03 472 480 10.1016/j.rbmo.2018.11.026 30642638 24 Schindler A E Campagnoli C Druckmann R Huber J Pasqualini J R Schweppe K W Classification and pharmacology of progestins Maturitas 2003 46 01 S7 S16 10.1016/j.maturitas.2003.09.014 14670641 25 Stanczyk F Z Hapgood J P Winer S Mishell D R Jr Progestogens used in postmenopausal hormone therapy: differences in their pharmacological properties, intracellular actions, and clinical effects Endocr Rev 2013 34 02 171 208 10.1210/er.2012-1008 23238854 26 Griesinger G Blockeel C Tournaye H Oral dydrogesterone for luteal phase support in fresh in vitro fertilization cycles: a new standard? Fertil Steril 2018 109 05 756 762 10.1016/j.fertnstert.2018.03.034 29778368 27 Alpha Scientists in Reproductive Medicine and ESHRE Special Interest Group of Embryology The Istanbul consensus workshop on embryo assessment: proceedings of an expert meeting Hum Reprod 2011 26 06 1270 1283 10.1093/humrep/der037 21502182 28 Metello J Tomás C Ferreira P Bravo I Branquinho M Santos-Ribeiro S Impact of plasmatic progesterone on the day of frozen embryo transfer in hormone-induced cycles Rev Bras Ginecol Obstet 2021 43 08 608 615 10.1055/s-0041-1735229 34547795 29 EARTH Study Team Yland J Messerlian C Mínguez-Alarcón L Ford J B Hauser R Williams P L Methodological approaches to analyzing IVF data with multiple cycles Hum Reprod 2019 34 03 549 557 10.1093/humrep/dey374 30576499 30 Kofinas J D Blakemore J McCulloh D H Grifo J Serum progesterone levels greater than 20 ng/dl on day of embryo transfer are associated with lower live birth and higher pregnancy loss rates J Assist Reprod Genet 2015 32 09 1395 1399 10.1007/s10815-015-0546-7 26238390 31 Atzmon Y Aslih N Estrada D Bilgory A Ellenbogen A Shalom-Paz E Comparable outcomes using oral dydrogesterone vs. micronized vaginal progesterone in frozen embryo transfer: a retrospective cohort study Reprod Sci 2021 28 07 1874 1881 10.1007/s43032-020-00376-3 33140324 32 Chakravarty B N Shirazee H H Dam P Goswami S K Chatterjee R Ghosh S Oral dydrogesterone versus intravaginal micronised progesterone as luteal phase support in assisted reproductive technology (ART) cycles: results of a randomised study J Steroid Biochem Mol Biol 2005 97 05 416 420 10.1016/j.jsbmb.2005.08.012 16213136 33 Patki A Pawar V C Modulating fertility outcome in assisted reproductive technologies by the use of dydrogesterone Gynecol Endocrinol 2007 23 01 68 72 10.1080/09513590701584857 17943542 34 Ganesh A Chakravorty N Mukherjee R Goswami S Chaudhury K Chakravarty B Comparison of oral dydrogestrone with progesterone gel and micronized progesterone for luteal support in 1,373 women undergoing in vitro fertilization: a randomized clinical study Fertil Steril 2011 95 06 1961 1965 10.1016/j.fertnstert.2011.01.148 21333984 35 Tomic V Tomic J Klaic D Z Kasum M Kuna K Oral dydrogesterone versus vaginal progesterone gel in the luteal phase support: randomized controlled trial Eur J Obstet Gynecol Reprod Biol 2015 186 49 53 10.1016/j.ejogrb.2014.11.002 25622239 36 Saharkhiz N Zamaniyan M Salehpour S Zadehmodarres S Hoseini S Cheraghi L A comparative study of dydrogesterone and micronized progesterone for luteal phase support during in vitro fertilization (IVF) cycles Gynecol Endocrinol 2016 32 03 213 217 10.3109/09513590.2015.1110136 26486011 37 Zarei A Sohail P Parsanezhad M E Alborzi S Samsami A Azizi M Comparison of four protocols for luteal phase support in frozen-thawed Embryo transfer cycles: a randomized clinical trial Arch Gynecol Obstet 2017 295 01 239 246 10.1007/s00404-016-4217-4 27761732 38 Álvarez M Gaggiotti-Marre S Martínez F Coll L García S González-Foruria I Individualised luteal phase support in artificially prepared frozen embryo transfer cycles based on serum progesterone levels: a prospective cohort study Hum Reprod 2021 36 06 1552 1560 10.1093/humrep/deab031 33686413 39 Vuong L N Pham T D Le K TQ Ly T T Le H L Nguyen D T Micronized progesterone plus dydrogesterone versus micronized progesterone alone for luteal phase support in frozen-thawed cycles (MIDRONE): a prospective cohort study Hum Reprod 2021 36 07 1821 1831 10.1093/humrep/deab093 33930124 40 Queisser-Luft A Dydrogesterone use during pregnancy: overview of birth defects reported since 1977 Early Hum Dev 2009 85 06 375 377 10.1016/j.earlhumdev.2008.12.016 19193503 41 Zaqout M Aslem E Abuqamar M Abughazza O Panzer J De Wolf D The impact of oral intake of dydrogesterone on fetal heart development during early pregnancy Pediatr Cardiol 2015 36 07 1483 1488 10.1007/s00246-015-1190-9 25972284 42 Griesinger G Blockeel C Kahler E Pexman-Fieth C Olofsson J I Driessen S Dydrogesterone as an oral alternative to vaginal progesterone for IVF luteal phase support: A systematic review and individual participant data meta-analysis PLoS One 2020 15 11 e0241044 10.1371/journal.pone.0241044 33147288
PMC009xxxxxx/PMC9708402.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36446566 10.1055/s-0042-1758490 febrasgostatement Febrasgo Position Statement Screening, diagnosis and management of hypothyroidism in pregnancy Number 10 – October 2022 http://orcid.org/0000-0002-9086-2303 Solha Sara Toassa Gomes 1 http://orcid.org/0000-0003-1405-5371 Mattar Rosiane 2 http://orcid.org/0000-0002-1215-7190 Teixeira Patrícia de Fátima dos Santos 3 http://orcid.org/0000-0002-2890-9094 Chiamolera Maria Izabel 4 http://orcid.org/0000-0001-5773-1590 Maganha Carlos Alberto 5 http://orcid.org/0000-0001-6440-4395 Zaconeta Alberto Carlos Moreno 6 http://orcid.org/0000-0003-0932-6835 Souza Renato Teixeira 7 1 Policlínicas Municipal, Sorocaba, SP, Brazil 2 Departamento de Obstetrícia, Escola Paulista de Medicina, São Paulo, SP, Brazil 3 Universidade Federal do Rio de Janeiro, Rio de Janeiro, RJ, Brazil 4 Escola Paulista de Medicina, Universidade Federal de São Paulo, São Paulo, SP, Brazil 5 Faculdade de Ciências Médicas de São José dos Campos, São José dos Campos, SP, Brazil 6 Universidade de Brasília, Brasília, DF, Brasil 7 Universidade Estadual de Campinas, Campinas, SP, Brazil 29 11 2022 10 2022 1 11 2022 44 10 9991010 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. ==== Body pmcKey points Pregnancy places a metabolic overload on the maternal thyroid, especially in the first trimester, mainly because of the demand imposed by the conceptus. The fetal thyroid becomes functionally mature only around pregnancy week 20. Until then, the fetus depends on the transfer of maternal thyroid hormones (THs). Thyroid hormones are essential for the adequate fetal neurofunctional and cognitive development. Hypothyroidism brings higher risks of obstetric and fetal complications, namely, first-trimester miscarriage, preeclampsia and gestational hypertension, placental abruption, prematurity, low birth weight, and higher perinatal morbidity and mortality. Primary hypothyroidism (involvement of the gland with difficulty in producing and/or releasing TH) is the most common form of disease presentation, with the main etiology of Hashimoto’s thyroiditis of autoimmune origin. In about 85%-90% of cases of Hashimoto’s thyroiditis, antithyroid antibodies are present; the antithyroperoxidase (ATPO) is the most frequent. Positivity for ATPO is determined when circulating values exceed the upper limit of the laboratory reference. It implies greater risks of adverse maternal-fetal outcomes. Such a correlation occurs even in ranges of maternal euthyroidism. The critical point for the diagnosis of hypothyroidism during pregnancy is an elevation of thyroid-stimulating hormone (TSH). The measurement of free thyroxine (FT4) differentiates between subclinical and overt hypothyroidism. In subclinical hypothyroidism, FT4 is within the normal range, whereas in overt hypothyroidism, FT4 values are below the lower limit of the laboratory reference. Treatment of hypothyroidism is performed with levothyroxine (LT4) replacement with the aim of achieving adequate TSH levels for pregnancy. Some women have a previous diagnosis of hypothyroidism, and may or may not be compensated at the beginning of pregnancy. Even in compensated cases, the increase in LT4 dose is necessary as soon as possible. In the postpartum period, adjustment of the LT4 dose depends on the condition of previous disease, on the positivity for ATPO, and also on the value of LT4 in use at the end of pregnancy. Recommendations In places with full technical and financial conditions, TSH testing should be performed for all pregnant women (universal screening) as early as possible, ideally at the beginning of the first trimester or even in preconception planning. In places with less access to laboratory tests, screening is reserved for cases with greater risk factors for decompensation, namely: previous thyroidectomy or radioiodine therapy, type 1 diabetes mellitus or other autoimmune diseases, presence of goiter, previous history of hypo or hyperthyroidism or previous ATPO positivity. The TSH dosage should be repeated throughout pregnancy only in these cases. The diagnosis of hypothyroidism is made from the TSH value > 4.0 mIU/L. Pregnant women with previous hypothyroidism, overt hypothyroidism diagnosed during pregnancy or those with the above-mentioned higher risk factors for decompensation should be referred for risk antenatal care, preferably in conjunction with the endocrinologist. Overt hypothyroidism in pregnancy is identified when TSH > 10 mIU/L, and treatment with LT4 is readily recommended at an initial dose of 2 mcg/kg/day. TSH values > 4.0 mUI/L and ≤ 10.0 mUI/L require FT4 measurement with two diagnostic possibilities: overt hypothyroidism when FT4 levels are below the lower limit of the laboratory reference, or subclinical hypothyroidism when FT4 levels are normal. The treatment for subclinical hypothyroidism is LT4 at an initial dose of 1 mcg/kg/day, and the dose should be doubled upon diagnosis of overt hypothyroidism. In cases of TSH > 2.5 and ≤ 4.0 mIU/L, if there are complete conditions, ATPO should be measured. If positive (above the upper limit of normal), treatment with LT4 at a dose of 50 mcg/day is indicated. If conditions are not complete, the repetition of the TSH dosage should be done only for cases at higher risk. In these cases, treatment with LT4 will be established when TSH > 4.0 mIU/L at a dose of 1 mcg/kg/day; if needed, the dose can be adjusted after FT4 evaluation. Women with previous hypothyroidism should have their LT4 dose adjusted to achieve TSH < 2.5 mIU/L at preconception. As soon as they become pregnant, they need a 30% increase in LT4 as early as possible. In practice, they should double the usual dose on two days a week. Levothyroxine should be given 30-60 minutes before breakfast or three hours or more after the last meal. Concomitant intake with ferrous sulfate, calcium carbonate, aluminum hydroxide and sucralfate should be avoided. The target of LT4 therapy during pregnancy is to achieve a TSH value < 2.5 mIU/L. Once the therapy is started, monthly control must be performed until the mentioned goal is reached. In the postpartum period, women with previous disease should resume the preconception dose. Cases diagnosed during pregnancy in use of LT4 ≤ 50 mcg/day may have the medication suspended. The others should reduce the current dose by 25% to 50% and repeat the TSH measurement in six weeks. Cases of ATPO positivity are at higher risk of developing postpartum thyroiditis and de-escalation of LT4 should be performed as explained. Background There is an increase in the stimulation of the hypothalamic-pituitary-thyroid axis during pregnancy through different mechanisms: 1 Increase in serum concentration of estrogens accompanied by an increase in thyroid hormone-binding globulin (TBG) and consequent reduction in free fractions of THs; Greater iodine clearance; Greater degradation of THs by placental deiodinases; Increase in the serum concentration of human chorionic gonadotropin (hCG), which stimulates the thyroid tissue by cross-reacting with the TSH receptor that can generate goiter and gestational transient hyperthyroidism (GTT). This stimulus to the axis explains why pregnant women have lower TSH concentrations than non-pregnant women, especially in the first trimester. 1 In the first trimester of pregnancy, there are greater metabolic demands and given the changes in the stimulation of the gland, this a critical period for the occurrence of thyroid dysfunctions. All described changes in the physiology of the hypothalamic-pituitary-thyroid axis ensure the supply of THs to the fetus, especially in the period when the fetal thyroid is not yet functionally mature. Although the development of the gland begins at week 8, it functions fully only between weeks 18-20 of pregnancy. Therefore, until that moment, the fetus is totally dependent on the placental transfer of maternal THs. Note that this is a critical period for the formation of the fetal nervous system. 1 There is a compensatory mechanism, via feedback, between the thyroid, pituitary and hypothalamus that regulates glandular functioning. Knowing this mechanism helps in signaling where the cause of a possible dysfunction may be ( Figure 1 ). Figure 1. Schematic representation of the hypothalamic-pituitary-thyroid axis indicating the types of hypothyroidism. TRH: thyrotropin releasing hormone; TSH: thyroid stimulating hormone; T4: thyroxine; T3: triiodothyronine. Source: Prepared by the Working Group for Thyroid Dysfunctions in Pregnancy (Brazilian Society of Endocrinology and Metabology – SBEM). Primary hypothyroidism is the most common form of the disease and occurs due to failure in the production and/or release of thyroxine (T4) and triiodothyronine (T3) hormones by the thyroid gland. 2 3 4 When central hypothyroidism is secondary to a failure in the production or release of TSH by the pituitary gland, manifestations from other pituitary sectors often occur. As gonadotropic or corticotrophic impairment is common, patients usually need to replace other hormones before they need TH replacement. 5 More rarely, central hypothyroidism has a tertiary origin resulting from hypothalamic disorders impairing the production or release of thyrotropin releasing hormone (TRH). In this document, we will address primary hypothyroidism during pregnancy. As in non-pregnant women, Hashimoto’s thyroiditis is its main cause, as well as the most prevalent among organ-specific autoimmune diseases, affecting about 12-13% of the population. 6 7 In approximately 85-90% of cases, Hashimoto’s thyroiditis presents with circulating antithyroid antibodies; the most common is the ATPO. 2 6 Primary hypothyroidism caused by Hashimoto’s thyroiditis may present with progressive thyroid failure, in which, initially, small elevations in serum TSH levels may compensate for slight reductions in the production of T4 and T3 by the thyroid. 3 4 Other causes of hypothyroidism include previous thyroidectomy and/or radioiodine therapy for definitive treatment of hyperthyroidism. Less commonly, primary hypothyroidism is caused by glandular infiltrative diseases (amyloidosis, hemochromatosis, sarcoidosis), drug use, and severe deficiency or excess of iodine. 2 3 4 A critical point in the investigation of hypothyroidism in pregnant women refers to what TSH elevation should be considered, defining the presence of primary hypothyroidism. When local reference values specific for pregnant women are unavailable, reference values from other centers that have been determined in a population of similar characteristics using the same TSH measurement methodology can be adopted. 8 According to the latest North American guideline, the upper limits of normality for TSH in the first trimester of pregnancy can be calculated by reducing 0.5 mIU/L from the upper limit of TSH reference values for non-pregnant women. 8 Such a strategy was confirmed in a study conducted with a population of pregnant women in the city of Rio de Janeiro. 9 Therefore, for most centers, the upper limit of TSH for pregnant women will be around 4.0 mIU/L, since most laboratory kits present a value of 4.5 mIU/L as the upper limit of normal for non-pregnant women. When treatment of primary hypothyroidism is indicated, LT4 replacement is required with the aim to reduce TSH levels to appropriate values for the patient’s age group and health condition. 2 3 4 These values are different for pregnant women compared to non-pregnant women in the same age group. 2 3 4 10 Common scenarios in clinical practice When addressing hypothyroidism in pregnancy, the main scenarios present patients previously diagnosed with hypothyroidism and those who received this diagnosis during pregnancy. The following conditions are possible for patients with a previous diagnosis: Under treatment (appropriate or not) with LT4; Untreated and decompensated; No treatment, but compensated, especially in cases of subclinical hypothyroidism. In general, patients with hypothyroidism who get pregnant, even if compensated, need an increase in LT4 dose to maintain adequate TSH concentrations. Regarding diagnosis during pregnancy, we may have: Overt hypothyroidism; Subclinical hypothyroidism; Euthyroidism with positive ATPO. The laboratory finding that combines TSH levels above reference values considered normal for a given population with normal circulating levels of THs is called subclinical hypothyroidism (mild or initial). 10 Overt hypothyroidism is considered when TSH levels during pregnancy are above 10.0 mIU/L 11 12 or when TSH levels are > 4 mIU/L and ≤ 10.0 mIU/L and the patient has TH levels below the lower limits of the respective reference values. 8 By knowing that most of the hormone produced by the thyroid is T4, which circulates preferentially linked to TBG and the latter is under direct interference from different factors, it is appropriate to measure FT4 as a way of estimating thyroid hormone production. Measurement of FT4 concentrations is performed to differentiate subclinical from clinical thyroid dysfunction. Obstetricians and endocrinologists must be familiar with thyroid function and dysfunction, as the correct diagnosis and treatment of thyroid diseases surely leads to better maternal and fetal outcomes. What are the possible maternal-fetal repercussions of hypothyroidism? Overt hypothyroidism has been consistently associated with a higher risk of pregnancy complications, and impaired neurocognitive development of the conceptus. 8 12 13 Normal TH levels are essential for neuronal migration, myelination, and proper brain formation of the fetal brain. 8 14 15 Complications more commonly associated with overt hypothyroidism are: 8 11 13 16 First trimester miscarriage; Preeclampsia and gestational hypertension; Placental abruption; Change in fetal vitality; Prematurity; Low birth weight; Cesarean delivery; Postpartum hemorrhage; Perinatal morbidity and mortality; Neuropsychological and cognitive impairment in children. Although there are controversies regarding the causal relationship of subclinical hypothyroidism with all situations mentioned above, many studies indicate that women with subclinical hypothyroidism are also at a higher risk for: Preeclampsia; Premature birth; Placental abruption; Neonatal respiratory distress syndrome and/or pregnancy loss. 8 11 13 With regard to fetal neurocognitive development, overt hypothyroidism should be considered a risk factor for its alteration, and logic suggests that subclinical hypothyroidism, in the same way, can cause intellectual impairment in a variable spectrum. 8 However, some studies argue there is no neurocognitive benefit associated with treatment of subclinical hypothyroidism in pregnancy. 17 Some limitations of these studies, such as fixed or excessive dose of LT4, as well as late initiation of treatment should be considered. Further research is needed to determine if early initiation of treatment (before week 13) is beneficial. 11 17 Note that women with circulating ATPO, even with TSH within normal limits, are at a higher risk of pregnancy complications compared to women with negative antibodies. 8 12 13 Regarding the timing of delivery, there is no evidence to support the resolution of pregnancy before 40 weeks for women with hypothyroidism. In places with complete technical and financial conditions, for which pregnant women and when to request an evaluation of thyroid function? Considering the possible complications of hypothyroidism during pregnancy, all women with pregnancy plans who known they have thyroid disease, should have their TSH levels evaluated, preferably before conception. On the other hand, there is no consensus on the best form of screening women without known thyroid disease. Despite the scarcity of robust evidence supporting universal screening, which proposes the evaluation of thyroid function for all pregnant women in the first trimester, several authors recommend this strategy under justification that different studies have shown the ineffectiveness of selective screening in diagnosing cases of hypothyroidism during pregnancy. 18 19 In a systematic review with meta-analysis, the two strategies were compared and almost half of cases of thyroid dysfunction went undiagnosed when only high-risk women were investigated. 20 In a recent study conducted in the city of Rio de Janeiro, 41.2% of pregnant women would remain undiagnosed because they did not meet any criteria for selective screening. 21 Thus, thyroid dysfunction during pregnancy presents high prevalence and is usually asymptomatic or oligosymptomatic. On the other hand, the serum TSH measurement, considered the test of choice for the diagnosis of thyroid dysfunction, is a sensitive, simple, robust, low-cost and widely available test, in addition to not offering risks to the patient when applied. Furthermore, the treatment of hypothyroidism, LT4, is an effective, inexpensive, available, and safe drug. 8 22 The benefits of treating overt hypothyroidism during pregnancy are crystal clear. According to recommendations of the World Health Organization (WHO), the only point that still lacks evidence for the complete inclusion of TSH in criteria establishing a screening test regards the lack of reported benefits of treating subclinical hypothyroidism during pregnancy. 22 Considering these premises, the present group recommends that under complete financial and technical conditions, the evaluation of thyroid function should be performed universally during pregnancy, as early as possible, at the beginning of antenatal care. In places where there are no complete financial and technical conditions, for which pregnant women and when to request the evaluation of thyroid function? If technical and/or financial limitations exist, criteria should be used to select those at greater risk for developing hypothyroidism during pregnancy. Evaluation of thyroid function should be performed as early as possible, preferably at the beginning of the first trimester of pregnancy. According to the latest recommendation from the American Thyroid Association (ATA), the physician should identify patients at greater risk of developing thyroid dysfunction according to criteria presented in chart 1 . In these selected cases, TSH dosage will be recommended before conception or at the first antenatal visit. 8 Considering that some factors, especially age > 30 years and ≥ 2 pregnancies, significantly increase the number of pregnant women selected for screening, in situations of even more limited economic and technical conditions, this group suggests the priority evaluation of thyroid function for women with the following risk factors: History of head/neck irradiation; Prior thyroid surgery or radioiodine therapy; Type 1 diabetes mellitus or other autoimmune disease; Previously known thyroid autoimmunity or presence of goiter; History of previous hypo or hyperthyroidism. Chart 1. Risk criteria for thyroid dysfunction during pregnancy according to the American Thyroid Association *History of head/neck irradiation *Previous thyroid surgery or radioiodine therapy *Type 1 diabetes mellitus or other autoimmune disease *Previously known thyroid autoimmunity or presence of goiter *History of hypo/hyperthyroidism - Signs/symptoms of thyroid dysfunction - Morbid obesity (BMI ≥ 40 kg m²) - History of fetal loss, premature birth or infertility - Use of amiodarone or lithium, or recent administration of iodinated contrast - Family history of autoimmune thyroid disease or thyroid dysfunction - Resident in an area with moderate to severe iodine insufficiency - Age > 30 years - ≥2 pregnancies BMI: body mass index Source: Adapted from Alexander et al. 8 In the presence of symptoms such as fatigue, constipation, anemia and weight gain in addition to the expected during pregnancy, the possibility of evaluating thyroid function should be considered ( Figure 2 ). Figure 2. Hypothyroidism: evaluation of thyroid function. TSH: thyroid stimulating hormone; BMI: body mass index. Source: Prepared by the Working Group for Thyroid Dysfunctions in Pregnancy (CNEGAR and Brazilian Society of Endocrinology and Metabology – SBEM). How to laboratory evaluate hypothyroidism? Measurement of TSH is considered the test of choice for screening for thyroid dysfunction ( Chart 2 ). Given the known physiological reduction in TSH levels during pregnancy, elevations in these levels are highly sensitive and suggestive of dysfunction. Preferably, the trimester-specific reference values provided by the laboratory for the local pregnant population should be taken into account. However, this group suggests that in the absence of validated references in local populations, a TSH level ≤ 4.0 mIU/L should be considered as a normal value. When TSH levels are at concentrations above 10.0 mIU/L, the diagnosis of overt hypothyroidism is confirmed and the patient should be treated with LT4 immediately, regardless of FT4 levels. Thus, there is no need to measure FT4 or ATPO to guide management. If TSH values are > 4.0 mUI/L and ≤ 10.0 mUI/L, FT4 must be requested, with two possibilities: Subclinical hypothyroidism: when FT4 levels are normal; Overt hypothyroidism: when FT4 levels are below the lower limit of the laboratory reference. Chart 2. Laboratory evaluation in the investigation of hypothyroidism in pregnancy TSH T4L Diagnosis ATPO HYPOTHYROIDISM DIAGNOSED PRIOR TO PREGNANCY ≤2.5 mUI/L Do not request (Result will not change management) Properly treated hypothyroidism during pregnancy Do not request (Result will not change management) >2.5 mUI/L Hypothyroidism not properly treated during pregnancy HYPOTHYROIDISM DIAGNOSED IN PREGNANCY >10.0 mUI/L Do not request (Result will not change management) Overt hypothyroidism Do not request (Result will not change management) >4.0 and ≤10.0 mUI/L Within reference values informed by the laboratory Subclinical hypothyroidism Do not request (Proposed management in subclinical hypothyroidism is independent of ATPO) >4.0 and ≤10.0 mUI/L Below the lower limit of reference informed by the laboratory Overt hypothyroidism Do not request (Result will not change management) >2.5 until ≤4.0 mUI/L Within reference values informed by the laboratory Euthyroidism ATPO positive case (above the upper limit of reference informed by the laboratory), consider treatment ≤2.5 Do not request (Result will not change management) Euthyroidism Do not request (Result will not change the proposed management for the first trimester in most patients) TSH: thyroid stimulating hormone; FT4: free thyroxine; ATPO: antithyroperoxidase antibody; SCH: subclinical hypothyroidism. Source: Prepared by the Working Group for Thyroid Dysfunctions in Pregnancy (CNEGAR and Brazilian Society of Endocrinology and Metabology – SBEM). Therefore, the ideal is to have immediate access to FT4 levels to determine the management, as its value defines the subclinical or clinical condition of hypothyroidism. 8 12 13 However, if this access is unavailable, our group suggests immediate initiation of LT4 treatment. With TSH values > 4.0 mUI/L and ≤ 10.0 mUI/L and normal FT4, our group understands there is no need to request ATPO. Evidence regarding the benefits obtained with treatment of subclinical hypothyroidism when there is no circulating ATPO is weak, but these studies have methodological limitations and adequate LT4 replacement has low cost and few risks. Thus, we consider that treatment may be indicated in these patients. If TSH values are > 2.5 mIU/L and ≤ 4 mIU/L, ATPO should be tested. Justification is that when combined with positive ATPO (values above the upper limit of normality of the reference), this group indicates the beginning of treatment as early as possible, preferably at the beginning of the first trimester. These patients are at higher risk of TSH elevation. Studies point to the relevant role of autoimmunity in the occurrence of maternal-fetal adverse outcomes. However, if ATPO testing is unavailable, the recommendation for the range of TSH > 2.5 mIU/L and ≤ 4 mIU/L, is as follows: It should not be treated; Patients who present the most relevant risk factors should have the TSH test repeated as soon as possible; If a TSH value > 4 mIU/L is identified, treatment with LT4 1 mcg/kg/day should be started and adjusted, if necessary, after FT4 evaluation. If the TSH level is normal in screening, should the test be repeated as the pregnancy progresses? There is no need to repeat the evaluation of thyroid function throughout pregnancy. It should be performed at the beginning of antenatal care, preferably in the first trimester. If TSH concentrations are ≤ 2.5 mIU/L or > 2.5 and ≤ 4 mIU/L with negative ATPO, the patient is considered to be euthyroid. The dosage will only be repeated when there is any clinical suspicion of thyroid dysfunction or, exceptionally, in the presence of more relevant risk factors ( Figure 1 ). How to calculate the LT4 dose to start hypothyroidism treatment? Recent studies have shown that the timing of introduction of TH can play an important role in the effectiveness of the intervention, and treatment in the first trimester has shown effectiveness in reducing the overall rate of pregnancy complications and the rate of preterm births. 23 24 25 Pregnant women with hypothyroidism should be treated with LT4 starting with doses of 1-2 mcg/kg daily. Since the objective is to reach euthyroidism as quickly as possible, the ideal is not to use the strategy of dose staggering, but to start with a full dose immediately. In a practical way, we recommend the following scheme ( Figure 3 ): 13 TSH > 10 mUI/L: 2 mcg/kg/day; TSH > 4 mUI/L and ≤ 10 mUI/L with FT4 below the lower limit of the laboratory reference: 2 mcg/kg/day; TSH > 4 mUI/L and ≤ 10 mUI/L with T4L within the normal range: 1 mcg/kg/day; TSH > 2.5 and ≤ 4 mUI/L with positive ATPO: 50 mcg/day. 8 13 Figure 3. Hypothyroidism: management during pregnancy. Source: Prepared by the Working Group for Thyroid Dysfunctions in Pregnancy (CNEGAR and Brazilian Society of Endocrinology and Metabology – SBEM). What should be the type of antenatal care for pregnant woman with hypothyroidism? Patients with hypothyroidism prior to pregnancy and those with a diagnosis of overt hypothyroidism in pregnancy screening should be followed up in high-risk antenatal care (preferably together with an endocrinologist). High surveillance is recommended for these patients. 8 Furthermore, in patients with more relevant risk factors for decompensation of thyroid metabolism during pregnancy, the same procedure should be followed. Such factors are: Prior head and neck irradiation; History of thyroid surgery and/or radioiodine therapy; Type 1 diabetes mellitus and other autoimmune diseases; Autoimmunity (positive ATPO) identified prior to pregnancy or presence of goiter; Previous history of hypo or hyperthyroidism. In this differentiated follow-up, continuous interaction between obstetrician and endocrinologist is recommended, with serial laboratory measurements to obtain adequate and early metabolic control. Cases of subclinical hypothyroidism without major risk factors should be followed up in Primary Care in usual risk antenatal care. How to perform dose adjustment for women with hypothyroidism who are planning to get pregnant or in early stages of pregnancy? Different cutoff points for TSH have been advocated as a target for preconception, ranging from < 1.2 mIU/L to < 2.5 mIU/L. It has already been found that only 17% of women with TSH < 1.2 mIU/L had to increase their LT4 dose during pregnancy. However, from a practical point of view, the recommendation so far is that women with hypothyroidism undergoing treatment optimize the dose in preconception, aiming to obtain TSH < 2.5 mIU/L ( Figure 4 ). 8 26 Pregnancy is associated with a higher need for TH in approximately one third of women already undergoing treatment. 26 27 This increased demand is believed to be related to greater estrogen production. 28 Clinical studies have confirmed that the higher need for exogenous LT4 occurs as early as weeks 4 to 6 of pregnancy. 26 27 This need gradually increases throughout weeks 16-20 of pregnancy and stabilizes thereafter until the time of delivery. For this reason, most pregnant women need to receive a higher dose of exogenous LT4 during pregnancy. 29 The importance of correcting doses early in pregnancy is noteworthy. The increase in LT4 dose varies depending on a number of factors: Underlying etiology of hypothyroidism: patients without functional thyroid tissue (following radiofrequency ablation or total thyroidectomy) have a greater need for dose increments compared to patients with autoimmune thyroid disease; 27 The preconception TSH level: lower preconception TSH values can lead to lower TSH elevations during the first trimester; 26 30 Variation in maternal estrogen levels during pregnancy correlates with variations in LT4 requirements during pregnancy; 27 After pregnancy confirmation, LT4 adjustment should be done as soon as possible, without the need for a gradual increase. The recommendation is to increase the dose by about 30%. In practice, this is achieved by doubling the usual dose two days a week. The routine dose is maintained on the other days. This can effectively mimic gestational physiology and therefore, prevent maternal hypothyroidism during the first trimester. Figure 4. Hypothyroidism: dose adjustment and therapeutic goal. TSH: thyroid stimulating hormone. Source: Prepared by the Working Group for Thyroid Dysfunctions in Pregnancy (CNEGAR and Brazilian Society of Endocrinology and Metabology – SBEM). After starting treatment, what is the goal to be achieved? How often should exams be requested? The treatment target is to achieve TSH in the lower half of the specific reference range of the pregnancy trimester. When this reference value is unavailable, it is reasonable to target maternal TSH concentrations below 2.5 mIU/L ( Figure 3 ). The TSH level should be evaluated every four weeks during dose adjustment. 8 13 31 What should be the guidelines for prescribing LT4? Preparations containing T3 should be avoided during pregnancy, as they increase T3 levels in comparison with T4, generating supraphysiological levels of maternal T3 and low levels of T4; maternal T4 is essential for the development of the fetal central nervous system. 8 As concomitant administration of LT4 and food can impair hormone absorption, LT4 administration is recommended 30-60 minutes before breakfast or at bedtime (three or more hours after the evening meal) for optimal and consistent absorption. In addition, LT4 should be taken separately from other potentially interfering drugs and supplements (eg ferrous sulfate, calcium carbonate, aluminum hydroxide, and sucralfate). 32 Switching LT4 brands can result in variations in the administered dose and should be avoided. 32 33 How should postpartum dose adjustment be done? After delivery, the maternal dose of LT4 should be adjusted to the dose used before pregnancy, and serum TSH should be measured after six weeks ( Figure 5 ). Some patients with Hashimoto’s thyroiditis who required an increase in the dose of TH during pregnancy may continue to need an additional dose in the postpartum period, in case of postpartum exacerbation of autoimmune thyroid dysfunction. On the other hand, some women with autoimmune thyroid disease who started taking LT4 during pregnancy may not need LT4 after delivery, especially when the dose is ≤ 50 mcg/d. If LT4 is discontinued, serum TSH should be re-evaluated in approximately six weeks. 34 When the dose used is greater than 50 mcg/d, a 25-50% reduction with further laboratory evaluation is recommended. 35 Figure 5. Hypothyroidism: postpartum. Source: Prepared by the Working Group for Thyroid Dysfunctions in Pregnancy (CNEGAR and Brazilian Society of Endocrinology and Metabology – SBEM). Final considerations Proper screening, diagnosis and management of hypothyroidism during pregnancy are essential, especially given the risks to fetal neurocognitive development and obstetric complications caused by improper treatment. Early identification is one of the major challenges, ideally at the beginning of the first trimester, allowing the initiation of therapy at the stage when the conceptus’ neuronal structures are developing. This way, the window of opportunity for real neurocognitive benefits for the offspring is reached. Thyroid hormones are essential for proper myelination and neuronal migration. The fetal thyroid is functionally mature only after week 20 of pregnancy; until then, the conceptus depends on the transfer of maternal THs. Furthermore, obstetric complications, especially those linked to first trimester abortion, preeclampsia, placental abruption and prematurity are at greater risk, especially when overt hypothyroidism is untreated. The laboratory test that guides the diagnosis of hypothyroidism is TSH, but there are several methodological limitations for the assertive establishment of TSH reference values representative of the local population. Current recommendations state that in this scenario, the normality value for TSH adopted during pregnancy should be ≤ 4.0 mIU/L. Universal screening is recommended provided that complete technical and financial conditions are available and should be carried out as early as possible. In adverse conditions, screening is reserved for cases with higher risk criteria for thyroid function decompensation during pregnancy. An important risk group is of women previously thyroidectomized or who underwent radioiodine therapy for the treatment of hyperthyroidism and who are already on LT4 therapy. Other critical factors include type 1 diabetes mellitus or other autoimmune diseases, the presence of goiter, a history of previous hypo or hyperthyroidism, or even the presence of autoimmunity. The T4L dosage helps to elucidate the overt or subclinical context of hypothyroidism. The role of autoimmunity –mainly ATPO positivity – is currently much studied. Women in this condition, even in euthyroidism, are at higher risk for adverse maternal-fetal outcomes. Although the literature is still controversial about the real benefits of treating subclinical hypothyroidism during pregnancy, evidence points to a relevant role of autoimmunity in the occurrence of potential complications in pregnancy and neurocognitive impairment of the offspring. Furthermore, the risk of occurrence of postpartum thyroiditis is greater in the presence of this autoimmune condition. In the preconception period, women who are known to be hypothyroid, especially those with positive ATPO, should be advised to adjust their LT4 therapy in advance, in order to enter the pregnancy-puerperal cycle under compensated conditions. De-escalation of LT4 in the postpartum period should be judicious and adequately monitored. It is a drug for use during breastfeeding. The challenges for coping with this pathology involve, above all, excessive diagnosis based on erroneous laboratory values that guide erroneous treatments and determine a non-existent gestational risk. At the other extreme, important conditions of thyroid dysfunction deserve adequate attention and guidance in high-risk antenatal care, receiving early and assertive treatments that avoid serious maternal-fetal complications. National Commission Specialized in High Risk Pregnancy of the Brazilian Federation of Gynecology and Obstetrics Associations (Febrasgo) President: Rosiane Mattar Vice-president: Alberto Carlos Moreno Zaconeta Secretary: Mylene Martins Lavado Members: Arlley Cleverson Belo da Silva Carlos Alberto Maganha Elton Carlos Ferreira Felipe Favorette Campanharo Fernanda Santos Grossi Inessa Beraldo de Andrade Bonomi Janete Vettorazzi Maria Rita de Figueiredo Lemos Bortolotto Renato Teixeira Souza Sara Toassa Gomes Solha Vera Therezinha Medeiros Borges Thyroid Department of the Brazilian Society of Endocrinology and Metabology President: Patrícia de Fátima dos Santos Teixeira (RJ) Vice-president: Danilo Glauco Pereira Villagelin Neto (SP) Directors: Rafael Selbach Scheffel (RS) Cléo Otaviano Mesa Júnior (PR) Gláucia Maria Ferreira da Silva Mazeto (SP) Maria Izabel Chiamolera (SP) Helton Estrela Ramos (BA) Conflicts of interest None to declare. The National Commission Specialized in High Risk Pregnancy of the Brazilian Federation of Gynecology and Obstetrics Associations (Febrasgo) and the Thyroid Department of the Brazilian Society of Endocrinology and Metabology (SBEM) endorse this document. The production of content is based on scientific evidence on the proposed theme and the results presented contribute to clinical practice. ==== Refs References 1 Korevaar T I Medici M Visser T J Peeters R P Thyroid disease in pregnancy: new insights in diagnosis and clinical management Nat Rev Endocrinol 2017 13 10 610 22 10.1038/nrendo.2017.93 28776582 2 Garber J R Cobin R H Gharib H Hennessey J V Klein I Mechanick J I Clinical practice guidelines for hypothyroidism in adults: cosponsored by the American Association of Clinical Endocrinologists and the American Thyroid Association Thyroid 2012 22 12 1200 35 10.1089/thy.2012.0205 22954017 3 Almandoz J P Gharib H Hypothyroidism: etiology, diagnosis, and management Med Clin North Am 2012 96 02 203 21 10.1016/j.mcna.2012.01.005 22443971 4 McDermott M T Hypothyroidism Ann Intern Med 2020 173 01 ITC1 16 10.7326/AITC202007070 32628881 5 Fleseriu M Hashim I A Karavitaki N Melmed S Murad M H Salvatori R Hormonal replacement in hypopituitarism in adults: an Endocrine Society Clinical Practice Guideline J Clin Endocrinol Metab 2016 101 11 3888 921 10.1210/jc.2016-2118 27736313 6 Caturegli P De Remigis A Rose N R Hashimoto thyroiditis: clinical and diagnostic criteria Autoimmun Rev 2014 13 (4-5):391 7 10.1016/j.autrev.2014.01.007 24434360 7 Sichieri R Baima J Marante T Vasconcellos M T Moura A S Vaisman M Low prevalence of hypothyroidism among black and Mulatto people in a population based study of Brazilian women Clin Endocrinol (Oxf) 2007 66 06 803 7 10.1111/j.1365-2265.2007.02816.x 17381480 8 Alexander E K Pearce E N Brent G A Brown R S Chen H Dosiou C 2017 Guidelines of the American Thyroid Association for the diagnosis and management of thyroid disease during pregnancy and the postpartum Thyroid 2017 27 03 315 89 10.1089/thy.2016.0457 28056690 9 Morais N S Assis A S Corcino C M Saraiva D A Berbara T Ventura C D Recent recommendations from ATA guidelines to define the upper reference range for serum TSH in the first trimester match reference ranges for pregnant women in Rio de Janeiro Arch Endocrinol Metab 2018 62 04 386 91 10.20945/2359-3997000000064 30304102 10 Pearce S H Brabant G Duntas L H Monzani F Peeters R P Razvi S 2013 ETA Guideline: management of subclinical hypothyroidism Eur Thyroid J 2013 2 04 215 28 10.1159/000356507 24783053 11 Lazarus J Brown R S Daumerie C Hubalewska-Dydejczyk A Negro R Vaidya B 2014 European Thyroid Association Guidelines for the Management of Subclinical Hypothyroidism in Pregnancy and in Children Eur Thyroid J 2014 3 02 76 94 10.1159/000362597 25114871 12 Thyroid disease in pregnancy: ACOG Practice Bulletin, Number 223 Obstet Gynecol 2020 135 06 e261 74 10.1097/AOG.0000000000003893 32443080 13 Ross D S Hypothyroidism during pregnancy: clinical manifestations, diagnosis, and treatment [Internet] 2022[cited 2022 Jun 10]. Available from:https://www.uptodate.com/contents/hypothyroidism-during-pregnancy-clinical-manifestations-diagnosis-and-treatment 14 Spencer L Bubner T Bain E Middleton P Screening and subsequent management for thyroid dysfunction pre-pregnancy and during pregnancy for improving maternal and infant health Cochrane Database Syst Rev 2015 2015 09 CD011263 10.1002/14651858.CD011263.pub2 26387772 15 Ausó E Lavado-Autric R Cuevas E Del Rey F E Morreale De Escobar G Berbel P A moderate and transient deficiency of maternal thyroid function at the beginning of fetal neocorticogenesis alters neuronal migration Endocrinology 2004 145 09 4037 47 10.1210/en.2004-0274 15087434 16 Abalovich M Gutierrez S Alcaraz G Maccallini G Garcia A Levalle O Overt and subclinical hypothyroidism complicating pregnancy Thyroid 2002 12 01 63 8 10.1089/105072502753451986 11838732 17 Lazarus J H Bestwick J P Channon S Paradice R Maina A Rees R Antenatal thyroid screening and childhood cognitive function N Engl J Med 2012 366 06 493 501 10.1056/NEJMoa1106104 22316443 18 Vaidya B Anthony S Bilous M Shields B Drury J Hutchison S Detection of thyroid dysfunction in early pregnancy: Universal screening or targeted high-risk case finding? J Clin Endocrinol Metab 2007 92 01 203 7 10.1210/jc.2006-1748 17032713 19 Chang D L Pearce E N Screening for maternal thyroid dysfunction in pregnancy: a review of the clinical evidence and current guidelines J Thyroid Res 2013 2013 851326 10.1155/2013/851326 23762776 20 Jouyandeh Z Hasani-Ranjbar S Qorbani M Larijani B Universal screening versus selective case-based screening for thyroid disorders in pregnancy Endocrine 2015 48 01 116 23 10.1007/s12020-014-0385-9 25173191 21 Berbara T M Morais N S Saraiva D A Corcino C M Schtscherbyna A Moreira K L Selective case finding versus universal screening for detecting hypothyroidism in the first trimester of pregnancy: a comparative evaluation of a group of pregnant women from Rio de Janeiro Arch Endocrinol Metab 2020 64 02 159 64 10.20945/2359-3997000000209 32236307 22 Wilson J M Jungner Y G [Principles and practice of mass screening for disease] Bol Oficina Sanit Panam 1968 65 04 281 393Spanish4234760 23 Dong A C Stephenson M D Stagnaro-Green A S The need for dynamic clinical guidelines: a systematic review of new research published after release of the 2017 ATA Guidelines on thyroid disease during pregnancy and the postpartum Front Endocrinol (Lausanne) 2020 11 193 10.3389/fendo.2020.00193 32318026 24 Nazarpour S Ramezani Tehrani F Simbar M Tohidi M Minooee S Rahmati M Effects of levothyroxine on pregnant women with subclinical hypothyroidism, negative for thyroid peroxidase antibodies J Clin Endocrinol Metab 2018 103 03 926 35 10.1210/jc.2017-01850 29126290 25 Casey B M Thom E A Peaceman A M Varner M W Sorokin Y Hirtz D G Treatment of subclinical hypothyroidism or hypothyroxinemia in pregnancy N Engl J Med 2017 376 09 815 25 10.1056/NEJMoa1606205 28249134 26 Abalovich M Alcaraz G Kleiman-Rubinsztein J Pavlove M M Cornelio C Levalle O The relationship of preconception thyrotropin levels to requirements for increasing the levothyroxine dose during pregnancy in women with primary hypothyroidism Thyroid 2010 20 10 1175 8 10.1089/thy.2009.0457 20860419 27 Alexander E K Marqusee E Lawrence J Jarolim P Fischer G A Larsen P R Timing and magnitude of increases in levothyroxine requirements during pregnancy in women with hypothyroidism N Engl J Med 2004 351 03 241 9 10.1056/NEJMoa040079 15254282 28 Arafah B M Increased need for thyroxine in women with hypothyroidism during estrogen therapy N Engl J Med 2001 344 23 1743 9 10.1056/NEJM200106073442302 11396440 29 Mandel S J Larsen P R Seely E W Brent G A Increased need for thyroxine during pregnancy in women with primary hypothyroidism N Engl J Med 1990 323 02 91 6 10.1056/NEJM199007123230204 2359428 30 Loh J A Wartofsky L Jonklaas J Burman K D The magnitude of increased levothyroxine requirements in hypothyroid pregnant women depends upon the etiology of the hypothyroidism Thyroid 2009 19 03 269 75 10.1089/thy.2008.0413 19265498 31 Yassa L Marqusee E Fawcett R Alexander E K Thyroid hormone early adjustment in pregnancy (the THERAPY) trial J Clin Endocrinol Metab 2010 95 07 3234 41 10.1210/jc.2010-0013 20463094 32 Jonklaas J Bianco A C Bauer A J Burman K D Cappola A R Celi F S Guidelines for the treatment of hypothyroidism: prepared by The American Thyroid Association task force on thyroid hormone replacement Thyroid 2014 24 12 1670 751 10.1089/thy.2014.0028 25266247 33 Mayor G H Orlando T Kurtz N M Limitations of levothyroxine bioequivalence evaluation: analysis of an attempted study Am J Ther 1995 2 06 417 32 11850687 34 Galofré J C Haber R S Mitchell A A Pessah R Davies T F Increased postpartum thyroxine replacement in Hashimoto’s thyroiditis Thyroid 2010 20 08 901 8 10.1089/thy.2009.0391 20615129 35 Melmed S Auchus R J Goldfine A B Koenig R J Rosen C J Williams texbook of endocrinology 14th ed. Philadelphia Elsevier 2020
PMC009xxxxxx/PMC9708403.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36446562 10.1055/s-0042-1755458 210480 Review Article Pregestational Diabetes and Congenital Heart Defects Diabetes pré-gestacional e defeitos cardíacos congênitoshttp://orcid.org/0000-0002-9929-9668 Maduro Catarina 1 http://orcid.org/0000-0001-6740-1990 Castro Luís Ferreira de 2 http://orcid.org/0000-0002-1807-5014 Moleiro Maria Lúcia 2 http://orcid.org/0000-0002-0491-0752 Guedes-Martins Luís 1234 1 Instituto de Ciências Biomédicas Abel Salazar, Universidade do Porto, Porto, Portugal 2 Departamento da Mulher e da Medicina Reprodutiva, Centro Hospitalar do Porto EPE, Centro Materno Infantil do Norte, Largo Prof. Abel Salazar, Porto, Portugal 3 Unidade de Investigação e Formação, Centro Materno Infantil do Norte, Porto, Portugal 4 Instituto de Investigação e Inovação em Saúde, Universidade do Porto, Porto, Portugal Address for correspondence Catarina Maduro, MSc Rua Jorge de Viterbo Ferreira 228, 4050-313, [email protected] 29 11 2022 10 2022 1 11 2022 44 10 953961 22 12 2021 03 6 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Studies have consistently shown a significant increase in the risk of congenital heart defects in the offspring of diabetic mothers compared with those of nondiabetic pregnancies. Evidence points that all types of pregestational diabetes have the capacity of generating cardiac malformations in a more accentuated manner than in gestational diabetes, and there seems to be an increased risk for all congenital heart defects phenotypes in the presence of maternal diabetes. Currently, the application of some therapies is under study in an attempt to reduce the risks inherent to diabetic pregnancies; however, it has not yet been possible to fully prove their effectiveness. The present review aims to better understand the mechanisms that govern the association between pregestational diabetes and congenital heart defects and how maternal diabetes interferes with fetal cardiac development, as there is still a long way to go in the investigation of this complex process. Resumo Estudos têm demonstrado consistentemente um aumento significativo no risco de defeitos cardíacos congênitos em filhos de mães diabéticas em comparação com os de gestações não diabéticas. Evidências apontam que todos os tipos de diabetes pré-gestacional têm capacidade de gerar malformações cardíacas mais acentuadas do que no diabetes gestacional e parece haver um risco aumentado para todos os fenótipos de cardiopatias congênitas na presença de diabetes materno. Atualmente, está em estudo a aplicação de algumas terapias na tentativa de diminuir os riscos inerentes à gravidez diabética; no entanto, ainda não foi possível comprovar totalmente a sua eficácia. A presente revisão visa compreender melhor os mecanismos que regem a associação entre diabetes pré-gestacional e cardiopatias congênitas e como o diabetes materno interfere no desenvolvimento cardíaco fetal, pois ainda há um longo caminho a percorrer na investigação deste processo complexo. Keywords pregestational diabetes diabetes mellitus hyperglycemia congenital heart defects Palavras-chave diabetes pré-gestacional diabetes melito hiperglicemia defeitos cardíacos congênitos ==== Body pmcIntroduction Congenital heart defects (CHDs), which affect 40,000 births per year in the United States, 1 represent the most prevalent congenital defects. 1 2 3 4 5 6 7 8 In addition, they are a major cause of noninfectious death in infants 7 and convey an increase in healthcare costs, 9 so their prenatal diagnosis through fetal echocardiography is essential. 4 6 It is described that, worldwide, ∼ 130 million women aged between 20 and 49 years old are diagnosed with diabetes mellitus (DM) and ∼ 21 million births are complicated by maternal diabetes (matDM). 7 Statistical data from Europe and the United States report that pregestational diabetes (PGD) affects 0.3% of pregnant women. 5 10 In addition, there has been an increasing rise in its prevalence over time, especially for type II diabetes. 2 4 8 Pregestational diabetes is associated with an increased risk of congenital defects and maternal and perinatal morbidity and mortality. 2 10 It threatens normal fetal cardiac development at several levels, which explains the wide spectrum of associated CHDs, from small structural and/or functional defects to major heart disease, with potential long-term sequelae. 7 9 10 11 12 Some studies point to a three times higher risk of CHD in the offspring of women with PGD compared with the offspring of nondiabetic women. 4 7 8 Similarly, there is a higher prevalence for each CHD phenotype in this population. 1 3 It is known that in pregnancies associated with prior matDM, hyperglycemia acts as a primary teratogen. 2 4 7 9 11 13 Its presence in early stages of the embryonic development of the cardiovascular system promotes the occurrence of embryopathies, culminating in cardiac defects. 1 3 7 10 12 14 Despite the clear role of hyperglycemia, other factors inherent to matDM, namely placental dysfunction, metabolic disorders such as obesity, and increased oxidative stress appear to be players that also modulate the disturbance of cardiogenesis. 4 11 15 Despite the apparent association between PGD as an environmental risk factor for CHDs, 1 3 5 7 8 9 13 there is still a long way to go in the investigation of this complex process and the mechanisms by which matDM interferes with fetal cardiac development. 3 7 9 11 13 Methods The present bibliographic review was based on a literature search of articles published between 2016 and 2021on the PubMed and Medline databases, restricted to articles written in English. Experimental and observational studies involving humans or animals were included. The keywords used were pregestational diabetes , diabetes mellitus , hyperglycemia , and congenital heart defects . From the analysis of the abstracts of the articles obtained, those that corresponded to the objective of the review were selected and, additionally, a search of the references of all the analyzed studies was performed to obtain additional information whenever necessary ( Fig. 1 ). Fig. 1 . Flow diagram of the literature review. Maternal Hyperglycemia and Fetal Cardiac Development The fetal environment in utero influences the development of the fetus during gestation, impacting on the likelihood of developing lifelong disease. Fetal effects resulting from deleterious conditions in utero appear to be proportional to the aggressiveness of these conditions. 16 During pregnancy, changes in glucose metabolism take place, in particular the increase in maternal insulin needs and its resistance in the last trimester, as well as hormones that inhibit its action, stimulating an increase in the amount of insulin supplied by the pancreas. 17 In normal situations, there is the maintenance of a balance of the fetal blood glucose level. At a late stage of gestation, there is a marked drop in glucose uptake in fetal cardiac cells in order to promote proper embryonic cardiogenesis. 18 Several studies highlight a strong correlation between matDM and a significantly increased risk of CHDs in the offspring of affected pregnant woman. 4 7 8 19 Hyperglycemia is the main teratogenic factor in diabetic pregnancies, and its presence prior to conception and in the 1 st trimester is associated with an increased risk of disturbed embryonic cardiac development. 7 17 20 21 22 23 24 25 Fetal hyperinsulinemia, inherent in pregnancies of diabetic mothers, is also thought to underlie diabetic embryopathy. 26 The spectrum of congenital cardiopathies associated with PGD involves looping, situs anomalies, conotruncal, septal 7 22 and valvular malformations, transposition of great vessels, double-outlet right ventricle, tetralogy of Fallot, 17 24 27 and aortic arch discontinuation. 20 Malformations of the cardiac outflow tract 21 and of the auriculoventricular septum 19 are particularly frequent. Claudio Gutierrez et al. 23 27 found an association between the hyperglycemic environment during pregnancy and the expansion of the ventricular compartment, decreased area of the ventricular myocardium, and dilation of the ascending aorta in the late stages of pregnancy. Other studies point to a correlation between PGD and hypertrophic fetal cardiomyopathy, 17 24 28 29 possibly due to the hyperinsulinemia that the fetus acquires in the context of maternal hyperglycaemia. 17 28 An evident increase in the thickness of several cardiac structures is also described, especially the interventricular septum, in pregnancies of diabetic mothers compared with fetuses of normal pregnancies, 17 28 resulting in disturbed cardiac function. 17 24 28 Several mechanisms have been implicated in this association between PGD and CHDs. 7 Hyperglycemia plays a role as a promoter of oxidative stress by increasing reactive oxygen and nitrogen species, 7 22 23 27 which will promote the occurrence of genetic changes and abnormalities of the usual pattern of apoptosis in cardiac cells, 23 27 particularly in the neural crest, 17 which is a key part in cardiac development. 7 On the other hand, in an hyperglycemic environment, the alteration of multiple signaling pathways with repercussions in cardiac development is observed: (1) exacerbation of the expression of transforming growth factor beta 1 (TGF-β1), originating an excessive accumulation of extracellular matrix proteins in cardiac tissues; 27 (2) decreased levels of nitric oxide, which is essential for the proper functioning of cardiac endothelial cells and whose reduction leads to inhibition of other signaling pathways dependent on its effect; 22 and (3) excessive stimulation of nucleotide biosynthesis via pentose phosphate, which is responsible for glucose metabolization, preventing proper maturation of cardiac cells. 18 An association between high glucose levels and consequent placental vascular dysfunction due to dysregulation of vascular endothelial growth factor, with a consequent impact on cardiogenesis, has also been described. 17 19 20 It is also known that the association between CHD and PGD does not change according to the type of PGD 3 4 9 or to the type of treatment implemented in the context of diabetes. 3 9 In fact, even with optimal glycemic control, there is an increased risk of developing CHD; 7 17 22 24 27 28 in clinical trials, it has been found that a negligible increase in glucose levels in the mother is associated with defects such as tetralogy of Fallot in the offspring, 7 22 which presupposes that hyperglycemia is potentiated or interacts concomitantly with other conditions in its teratogenic process. 22 Some studies point to a gene-environment interaction in which external factors inherent to the fetal environment may act together with genetic predisposition in modulating cardiac embryogenesis. 7 30 In short, how the fetus reacts to maternal hyperglycemia is subject to several factors, such as the developmental stage in which there was contact with maternal hyperglycemia ( Chart 1 ), its severity, the presence of pathologies or concomitant risk factors, and genetic background, inducing epigenetic changes and a complex interaction with repercussions on fetal cardiogenesis. 7 22 Chart 1 Spectrum of congenital cardiopathies associated with pregestational diabetes Congenital cardiopathies associated with pregestational diabetes Looping, situs, conotruncal, septal, and valvular anomalies Transposition of great vessels Double-outlet right ventricle Tetralogy of Fallot Aortic arch discontinuation Hypertrophic fetal cardiomyopathy Disturbed cardiac function Placental vascular dysfunction Effect of Pregestational Diabetes on Placental Function and Fetal Cardiac Function in the 1 st Trimester Placental development, which takes place in the 1 st trimester, corresponds to a stage of marked susceptibility, so PGD may be a disruptive factor. 31 32 The interface between the placental vascular system and fetal vessels exposes the placenta to maternofetal endocrine imbalances, with possible harmful repercussions on fetal development. 31 Fetuses exposed to the effects of hyperglycemia have a five-fold increased risk of death in utero. 33 For this reason, diabetic women should be the target of a careful preconceptional assessment and close monitoring from the 1 st trimester in order to maintain a regular and balanced metabolic control, minimizing the associated risks. 34 35 36 37 Oxygen levels, and consequently reactive oxygen species, are known to increase significantly in the placenta throughout the 1 st trimester of pregnancy, especially in the presence of PGD, with potential consequences on placental development. It is suggested that this increased oxygen tension amplifies the effects of hyperglycemia at the trophoblast level, culminating in decreased trophoblast proliferation during this period of gestation. As a result, the fetus will receive a deficient nutritional intake, compromising its development. Thus, there seems to be an association between PGD, deficient trophoblast proliferation, and disorders such as fetal growth restriction, pre-eclampsia, and miscarriage. 38 The higher propensity for congenital anomalies in pregnancies of diabetic mothers associated with elevated maternal glucose levels early in gestation is notorious. 39 Maternal hyperglycemia is thought to convey changes in the blood flow established between the mother, the placenta, and the fetus, which may have molecular effects promoting CHD. Placental abnormalities seem to propitiate inflammation and oxidative stress, with disruption of signaling pathways involved in fetal cardiac development. 7 Hyperglycemia is also known to impact proliferation and migration of neural crest cell tissues. 40 These are important for an adequate evolution of fetal cardiac function throughout pregnancy; 40 therefore, this interference in the 1 st trimester interferes with organogenesis, promoting the appearance of CHDs. 33 Russel et al. demonstrated a higher incidence of fetal cardiac function irregularities in the 1 st trimester in PGD compared with nondiabetic pregnancies. A deterioration of diastolic function and global cardiac function is noted in this context, highlighting a decrease in the ratio between passive and active ventricular filling and an increase in the isovolumetric relaxation period and in the myocardial performance index. 41 Turan et al. 33 identified a shortening of the isovolumetric contraction period, failure of cardiac contraction capacity, and deterioration of the ejection fraction. It was found that the worse the maternal glycemic control, the greater the deterioration of fetal diastolic function. 33 Sirico et al. 40 also described an increase in the mean 1 st -trimester fetal heart rate in matDM compared with nondiabetic pregnancies. Some studies seem to indicate that the structural cardiac abnormalities that occur in PGD are noticed after the deterioration of cardiac function shown on ultrasound, raising the suspicion that the latter may occur first. 41 In summary, an adequate functional and structural cardiovascular development of the fetus is determined by the interactions between the maternal, placental, and fetal environments ( Chart 2 ), which are closely dependent on maternal glycemic control in PGD, since glucose levels in the mother influence multiple aspects of fetal cardiogenesis. 33 Chart 2 Fetal cardiac dysfunctions in pregestational diabetes and methods for fetal heart function assessment Fetal cardiac dysfunctions in PGD Fetal heart function assessment Deterioration of diastolic function and global cardiac function Fetal echocardiography Decrease in the ratio between passive and active ventricular filling Increase in the isovolumetric relaxation period and myocardial performance index Shortening of the isovolumetric contraction period, failure of cardiac contraction capacity, and deterioration of the ejection fraction Increase in the mean 1 st -trimester fetal heart rate Abbreviation: PGD, pregestational diabetes. Risk of Congenital Heart Defects in Offspring Exposed to Maternal Diabetes Clinical trials have demonstrated an increased risk of CHDs in the offspring of diabetic mothers compared with those of nondiabetic mothers. 42 43 44 45 46 47 48 49 What remains to be clarified is the extent of this association, something that differs from study to study, as well as the relationship between matDM and particular subtypes of CHDs, 1 10 42 since the spectrum of associated CHDs seems to encompass > 20 phenotypes. 42 The literature shows that all types of PGD appear to be more likely to cause cardiac malformations than gestational diabetes. 9 42 Similarly, there appears to be an increased risk for all phenotypes of CHDs in the presence of matDM. 1 However, conotruncal defects, auriculoventricular septal defects, heterotaxy, ventricular outflow tract obstruction, and double-outflow right ventricle 10 42 43 have been particularly identified. It is estimated that the risk of CHD is about three times higher in pregnancies of diabetic mothers compared with those of nondiabetic mothers. 2 8 It is also noteworthy that, among congenital anomalies associated with matDM, CHDs correspond to the most frequent class. 2 9 36 Pregestational diabetes is, therefore, a modifiable risk factor for the incidence of adverse pregnancy outcomes. 9 It is known that the decisive period of fetal cardio genesis is between the 3 rd and 7 th weeks of gestation. 1 42 Thus, matDM, by promoting a hyperglycemic environment, generates imbalances in molecular pathways crucial to cardiac embryogenesis, with consequent damage to it. 1 8 36 42 The inherent alterations in insulin resistance favour glucose transfer through the placental interface, promoting a greater secretion of insulin by the pancreas, with increased levels of fetal insulin. 12 Hyperglycemia and subsequent fetal hyperinsulinemia may have teratogenic effects at this early stage of pregnancy. One of its apparent repercussions is myocardial hyperplasia and hypertrophy through insulin receptors on the cardiac surface, which mediate the increase in nutrient synthesis, with subsequent increase in cardiac muscle mass. 10 It has been found that there is an intensification of the expression of these receptors in the presence of poor glycemic control. 12 They are especially numerous in the interventricular septum, which is consistent with the hypertrophy often found in this septum in the offspring of diabetic mothers. 10 Similarly, studies identify an association between interventricular septal thickness and glycated hemoglobin (HgA1c) values. 12 Based on these findings, we conclude that the measurement of HgA1c at preconception and in the 1 st trimester is crucial for the surveillance of these pregnant women and for the assessment of the risk of congenital malformations. 36 Simultaneously, some authors argue that, following the oxidative stress intrinsic to PGD, there is a decrease in cell proliferation and an increase in apoptosis, as well as suppression of the expression of certain genes, blocking cardiomyocyte maturation and differentiation, inhibiting embryonic cardiac development. Thus, the regeneration potential of cardiac progenitor cells to restore injured cells is affected, which ultimately may also lead to cardiac abnormalities. 50 Although the extent of cardiac involvement is dependent on maternal glycemic control, 10 it has not yet been possible to quantify how current prenatal measures modulate the risk of CHDs. Measures to reduce the risk of cardiac abnormalities in PGD include strict control of blood glucose and body mass index 8 at preconception and in the early stages of pregnancy. In addition, early fetal ultrasound monitoring allows the diagnosis of a part of the cardiac anomalies in the 1 st half of pregnancy, making it possible to establish a timely course of action in the course of pregnancy. 51 The role of insulin analogues in 1 st -trimester pregnancies of diabetic mothers is currently under evaluation. In fact, there seems to be a decreased risk of CHDs in the offspring exposed to insulin analogues as opposed to human insulin. 36 The feasibility of stem cell therapies in CHDs is also under discussion, since PGD impairs the biological performance of progenitor cells and cardiac stem cells. 50 In conclusion, the pathogenesis of CHDs remains unclear, but seems to involve multiple players, with a crucial interaction between genetic and environmental factors. 8 42 52 These factors seem to lead to cardiac developmental disorders, both at morphological and functional levels, conditioning a wide spectrum of CHDs. 11 12 Therefore, the study of glycemic control interventions in pregnant women is essential to reduce the risk of these malformations. 36 Pregestational Diabetes Mellitus and Obstetric Outcomes Pregnancies complicated by PGD present a greater association with unfavorable maternofetal outcomes compared with pregnancies of nondiabetic mothers, 53 54 55 56 culminating in increased morbidity, 35 57 58 mortality, and hospitalizations. 56 The complications resulting from matDM with greater emphasis in the literature encompass fetal macrosomia, congenital anomalies (previously discussed), and miscarriage. 54 59 60 61 62 Also of note is the increased likelihood that the pregnant woman will suffer from hypertensive disorders, such as pre-eclampsia, or that the fetus will develop complications such as growth restriction, 35 54 jaundice, respiratory disorders, and neonatal hypoglycaemia. 63 It seems that the damage inherent to each of these complications is greater the greater the severity and duration of diabetes, pre-existing comorbidities, and glycemic control in early pregnancy. 35 59 60 Interestingly, even in pregnancies of diabetic mothers with better blood glucose levels, adverse outcomes continue to be recorded, and it remains unclear how much glycemic control effectively mitigates the risks inherent to matDM. On the other hand, maternal hypoglycemia also has the potential to generate adverse effects in pregnancy. Its presence in a fetus usually with high glucose levels seems to be associated with a greater threat of miscarriage. 60 Therefore, the assessment of fetal well-being during pregnancy involves several factors, and amniotic fluid volume is a key tool when we talk about diabetes in pregnancy. In pregnancies of diabetic mothers, there is a correlation between poor glycemic control and excessive accumulation of amniotic fluid (polyhydramnios). 61 The detection of this and other complications involves a multidisciplinary surveillance, with analytical and echographic controls, whose frequency and most effective management is still to be clarified, since all of them have limitations. 55 60 64 65 Since the obstetric prognosis is largely influenced by the follow-up implemented in diabetic mothers, 54 it would be ideal to initiate a line of preconception care. This would include closer monitoring of diabetic women who are planning to become pregnant in the near future, making efforts to control blood glucose values prior to pregnancy and implementing a multidisciplinary approach to optimize care, 58 66 which should include the regular screening for nephropathy and retinopathy and the verification of potentially teratogenic prescribed drugs, among other measures, in order to reduce as much as possible the risk of complications during pregnancy. 58 Despite advances in glycemic control and prenatal surveillance, improving obstetric care in this population remains a challenge: not all patients have access to healthcare and a large proportion do not use preconception care, missing a key window of opportunity to institute effective disease control before pregnancy to prevent or mitigate adverse outcomes ( Chart 3 ). 66 Chart 3 Some methods for the assessment of fetal well-being in pregestational diabetes Methods for the assessment of fetal well-being in PDG Analytical controls (glycaemia, serum levels of Pregnancy-Associated Plasma Protein A [PAPP-A]…) Fetal ultrasound monitoring Amniotic fluid volume Placental vascularization indices (uterine artery pulsatility levels…) Abbreviation: PGD, pregestational diabetes. Discussion The incidence of CHDs is clearly higher in the offspring of mothers with PGD compared with in the offspring of nondiabetic women, 1 3 4 7 8 with malformations of the cardiac outflow tract 21 and of the auriculoventricular septum 19 being particularly frequent. There is an association between PGD and fetal hypertrophic cardiomyopathy, 17 24 28 29 with an evident increase in the thickness of cardiac structures such as the interventricular septum, 17 28 leading to negative effects on long-term cardiac function. 17 24 28 Hyperglycemia is identified as the primary teratogen in this relationship, 2 4 7 9 11 13 and its presence in the early stages of cardiac embryogenesis seems to favour the occurrence of CHDs. 1 3 7 10 12 14 In addition, other factors inherent to matDM, such as placental dysfunction, increased oxidative stress, and alteration of multiple molecular signaling pathways appear to be players that also negatively modulate cardiogenesis. 4 11 15 Thus, cardiac abnormalities in the context of matDM have a multifactorial basis, highlighting the gene-environment interaction; that is, environmental factors, such as PGD, act together with genetic predisposition in modulating cardiovascular development 7 30 ( Fig. 2 ). Fig. 2 Evidence points to an association between pregestational diabetes and a higher propensity of the offspring to develop congenital heart disease (CHD). This correlation seems to be justified, on the one hand, by the presence of fetal hyperglycemia and hyperinsulinemia and, on the other hand, by a deficient placental development. Thus, the teratogenesis of maternal diabetes will reside in the generation of reactive oxygen and nitrogen species (oxidative stress), culminating in epigenetic and cell cycle changes, which condition a defective cardiogenesis. Simultaneously, studies highlight the role of genetic predisposition for abnormal fetal cardiac development, so that this interrelation between fetal environment and genetic background will be at the basis of fetal heart defects. How the fetus reacts to maternal hyperglycemia depends on several factors, such as the developmental stage in which it came into contact, its severity, the presence of concomitant diseases, and genetic background. This interaction results in epigenetic changes with considerable repercussions on fetal cardiogenesis. 7 22 Research in this area shows an important correlation between CHDs and maternal blood glucose levels at an early stage of pregnancy; therefore, the risk of CHD increases in pregnancies based on poor glycemic control or with repeated episodes of acute complications of diabetes at an earlystage. 4 8 9 Furthermore, pregnancies complicated by PGD are more associated with unfavorable maternal and fetal outcomes 53 54 55 56 and higher fetal and maternal morbidity and mortality. 35 57 58 At the fetal level, complications involve macrosomia, congenital anomalies, miscarriage, 54 59 60 61 62 shoulder dystocia or contusions at delivery, 53 64 jaundice, respiratory disorders, and neonatal hypoglycaemia. 63 For the mother, there is a higher risk of hypertensive disorders 35 54 and higher rates of caesarean sections or perineal injuries. 53 64 There are also several characteristics that, when present in pregnancies of diabetic mothers, are imminently promoters of perinatal mortality, namely a low socioeconomic status, smoking, advanced maternal age, obesity, or twin pregnancies. 31 61 67 68 Early fetal ultrasound monitoring with a set of diagnostic and prognostic markers, such as amniotic fluid volume assessment and fetal echocardiography, allows the identification of some complications and some cardiac anomalies. This surveillance is essential to define an appropriate course of action and to plan the eventual intervention required after birth. 51 The uterine environment experienced by the fetus clearly influences its development during pregnancy and, possibly, will also have repercussions in adulthood. Thus, diabetic women should receive individualized care, ideally from preconception, in order to maintain regular metabolic control and minimize the associated risks. 34 35 36 37 It is necessary to implement a continuous improvement of preconceptional and prenatal care, since there are still women who do not benefit from it, losing the possibility to prevent or mitigate deleterious outcomes. 66 In addition, it is necessary to continue to implement and improve surveillance and intervention programs to address the complications that arise in the context of maternal mortality, since the prevalence of PGD is increasing. 9 Although various resources exist for the early diagnosis of some of the complications of pregnancy in diabetic women, constant research into new markers is crucial, as the current methods have limitations. 55 60 64 65 There is still a significant list of answers to be found: why the teratogenesis associated with hyperglycemia has a more profound impact on certain organs; why the risk of CHDs in pregnancies of diabetic women does not equal the same risk in nondiabetic women, despite optimal glycemic control; or what mechanisms explain the existence of pregnancies in the context of matDM, which record much higher HgA1c values than what is considered acceptable for a pregnancy without birth defects and that, despite this, follow a normal course. 4 In fact, the extent of cardiac impairment is found to be partly dependent on maternal glycemic control, 10 but it is not yet possible to quantify how current prenatal measures modulate the risk of CHDs in this setting. 8 In this scenario, the question that arises is what should be the HgA1c threshold considered adequate for a woman with PGD to become pregnant without increasing risks, which remains unanswered. 4 Further studies will be needed to understand how this gene-environment interface occurs and why infants who have been exposed to teratogenic agents such as hyperglycemia are vulnerable to fetal cardiac development disorders. 7 Genetic mechanisms that potentiate susceptibility to certain environmental factors may be involved, something that will need to be clarified in future investigations. 7 Some treatments for diabetic pregnant women are currently under investigation, such as insulin analogues, which, compared to the use of human insulin, appear to have a superior ability to maintain more adequate blood glucose levels. Future investigations should test whether they effectively minimize the risk of CHDs in offspring exposed to them. 36 New therapies under study include the use of stem cells, given the role of maternal diabetes in cardiomyocyte development and repair; however, their efficacy has not yet been proven. 50 Conclusion Pregestational diabetes has an irrefutable negative influence on pregnancy and fetal cardiac development, even in women with adequate glycemic control. Given the increase of women with this condition in recent years, a proactive attitude is imperative in the information, prevention, and metabolic control of these patients in order to minimize the associated disorders and complications. It is necessary to continue research in this area in order to understand the various aspects of the association between maternal diabetes and fetal cardiac anomalies so that we can have an early and effective intervention in its development and prenatal detection. Conflict of Interest The authors have no conflict of interests to declare. ==== Refs References 1 Hoang T T Marengo L K Mitchell L E Canfield M A Agopian A J Original findings and updated meta analysis for the association between maternal diabetes and risk for congenital heart disease phenotypes Am J Epidemiol 2017 186 01 118 128 10.1093/aje/kwx033 28505225 2 Agha M M Glazier R H Moineddin R Booth G Congenital abnormalities in newborns of women with pregestational diabetes: A time-trend analysis, 1994 to 2009 Birth Defects Res A Clin Mol Teratol 2016 106 10 831 839 10.1002/bdra.23548 27511615 3 Øyen N Diaz L J Leirgul E Boyd H A Priest J Mathiesen E R Prepregnancy diabetes and offspring risk of congenital heart disease: a nationwide cohort study Circulation 2016 133 23 2243 2253 10.1161/CIRCULATIONAHA.115.017465 27166384 4 Gabbay-Benziv R Reece E A Wang F Yang P Birth defects in pregestational diabetes: Defect range, glycemic threshold and pathogenesis World J Diabetes 2015 6 03 481 488 10.4239/wjd.v6.i3.481 25897357 5 Garne E Loane M Dolk H Barisic I Addor M-C Arriola L Spectrum of congenital anomalies in pregnancies with pregestational diabetes Birth Defects Res A Clin Mol Teratol 2012 94 03 134 140 10.1002/bdra.22886 22371321 6 Gomez K J Dowdy K Allen G Tyson-Thomas M Cruz A C Evaluation of ultrasound diagnosis of fetal anomalies in women with pregestational diabetes: University of Florida experience Am J Obstet Gynecol 1988 159 03 584 586 10.1016/s0002-9378(88)80013-9 3048097 7 Basu M Garg V Maternal hyperglycemia and fetal cardiac development: Clinical impact and underlying mechanisms Birth Defects Res 2018 110 20 1504 1516 10.1002/bdr2.1435 30576094 8 Leirgul E Brodwall K Greve G Vollset S E Holsmtrøm H Tell G S Maternal diabetes, birth weight, and neonatal risk of congenital heart defects in Norway, 1994-2009 Obstet Gynecol 2016 128 05 1116 1125 10.1097/AOG.0000000000001694 27741197 9 Correa A Pre-gestational diabetes and congenital heart defects Circulation 2016 133 23 2219 2221 10.1161/CIRCULATIONAHA.116.022960 27166385 10 Dervisoglu P Kosecik M Kumbasar S Effects of gestational and pregestational diabetes mellitus on the foetal heart: a cross-sectional study J Obstet Gynaecol 2018 38 03 408 412 10.1080/01443615.2017.1410536 29355062 11 Pauliks L B The effect of pregestational diabetes on fetal heart function Expert Rev Cardiovasc Ther 2015 13 01 67 74 10.1586/14779072.2015.988141 25431859 12 Babović I Arandjelović M Plešinac S Kontić-Vučinić O Radunović N Maternal glycoregulation in pregnancies complicated by diabetes mellitus in the prediction of fetal echography findings and perinatal outcomes J Obstet Gynaecol Res 2018 44 03 432 439 10.1111/jog.13537 29266610 13 Correa A Gilboa S M Besser L M Botto L D Moore C A Diabetes mellitus and birth defects Am J Obstet Gynecol 2008 199 03 2370 2.37E11 10.1016/j.ajog.2008.06.028 14 Parnell A S Correa A Reece E A Pre-pregnancy obesity as a modifier of gestational diabetes and birth defects associations: a systematic review Matern Child Health J 2017 21 05 1105 1120 10.1007/s10995-016-2209-4 28120287 15 Moazzen H Lu X Liu M Feng Q Pregestational diabetes induces fetal coronary artery malformation via reactive oxygen species signaling Diabetes 2015 64 04 1431 1443 10.2337/db14-0190 25422104 16 De Blasio M J Dodic M Jefferies A J Moritz K M Wintour E M Owens J A Maternal exposure to dexamethasone or cortisol in early pregnancy differentially alters insulin secretion and glucose homeostasis in adult male sheep offspring Am J Physiol Endocrinol Metab 2007 293 01 E75 E82 10.1152/ajpendo.00689.2006 17356009 17 Corrigan N Brazil D P McAuliffe F Fetal cardiac effects of maternal hyperglycemia during pregnancy Birth Defects Res A Clin Mol Teratol 2009 85 06 523 530 10.1002/bdra.20567 19180650 18 Nakano H Minami I Braas D Pappoe H Wu X Sagadevan A Glucose inhibits cardiac muscle maturation through nucleotide biosynthesis eLife 2017 6 e29330 10.7554/eLife.29330 29231167 19 Madri J A Enciso J Pinter E Maternal diabetes: effects on embryonic vascular development–a vascular endothelial growth factor-A-mediated process Pediatr Dev Pathol 2003 6 04 334 341 10.1007/s10024-003-5051-9 14692647 20 Molin D G Roest P A Nordstrand H Wisse L J Poelmann R E Eriksson U J Disturbed morphogenesis of cardiac outflow tract and increased rate of aortic arch anomalies in the offspring of diabetic rats Birth Defects Res A Clin Mol Teratol 2004 70 12 927 938 10.1002/bdra.20101 15578651 21 Zhao Z TGFβ and Wnt in cardiac outflow tract defects in offspring of diabetic pregnancies Birth Defects Res B Dev Reprod Toxicol 2014 101 05 364 370 10.1002/bdrb.21120 25231192 22 Basu M Zhu J Y LaHaye S Majumdar U Jiao K Han Z Epigenetic mechanisms underlying maternal diabetes-associated risk of congenital heart disease JCI Insight 2017 2 20 95085 10.1172/jci.insight.95085 29046480 23 Claudio Gutierrez J Prater M R Hrubec T C Smith B J Freeman L E Holladay S D Heart changes in 17-day-old fetuses of diabetic ICR (Institute of Cancer Research) mothers: improvement with maternal immune stimulation Congenit Anom (Kyoto) 2009 49 01 1 7 10.1111/j.1741-4520.2008.00213.x 19243410 24 Han S S Wang G Jin Y Ma Z-I Jia W-j Wu X Investigating the mechanism of hyperglycemia-induced fetal cardiac hypertrophy PLoS One 2015 10 09 e0139141 10.1371/journal.pone.0139141 26418041 25 Negrato C A Mattar R Gomes M B Adverse pregnancy outcomes in women with diabetes Diabetol Metab Syndr 2012 4 01 41 10.1186/1758-5996-4-41 22964143 26 Tshiyombo T D Oulton M R Comparative effects of chronic exposure to glucose or sodium butyrate on surfactant development in fetal rabbits J Perinat Med 2001 29 06 476 485 10.1515/JPM.2001.067 11776678 27 Gutierrez J C Hrubec T C Prater M R Smith B J Freeman L E Holladay S D Aortic and ventricular dilation and myocardial reduction in gestation day 17 ICR mouse fetuses of diabetic mothers Birth Defects Res A Clin Mol Teratol 2007 79 06 459 464 10.1002/bdra.20357 17335049 28 Wong M L Wong W H Cheung Y F Fetal myocardial performance in pregnancies complicated by gestational impaired glucose tolerance Ultrasound Obstet Gynecol 2007 29 04 395 400 10.1002/uog.3957 17330321 29 Loffredo C A Wilson P D Ferencz C Maternal diabetes: an independent risk factor for major cardiovascular malformations with increased mortality of affected infants Teratology 2001 64 02 98 106 10.1002/tera.1051 11460261 30 Fung A Manlhiot C Naik S Rosenberg H Smythe J Lougheed J Impact of prenatal risk factors on congenital heart disease in the current era J Am Heart Assoc 2013 2 03 e000064 10.1161/JAHA.113.000064 23727699 31 Gonzalez Gonzalez N L Gonzalez Davila E Castro A Padron E Plasencia W Effect of pregestational diabetes mellitus on first trimester placental characteristics: three-dimensional placental volume and power Doppler indices Placenta 2014 35 03 147 151 10.1016/j.placenta.2014.01.002 24456657 32 Leach L Taylor A Sciota F Vascular dysfunction in the diabetic placenta: causes and consequences J Anat 2009 215 01 69 76 10.1111/j.1469-7580.2009.01098.x 19563553 33 Turan S Turan O M Miller J Harman C Reece E A Baschat A A Decreased fetal cardiac performance in the first trimester correlates with hyperglycemia in pregestational maternal diabetes Ultrasound Obstet Gynecol 2011 38 03 325 331 10.1002/uog.9035 21538641 34 Allen A J Snowden J M Lau B Cheng Y Caughey A B Type-2 diabetes mellitus: does prenatal care affect outcomes? J Matern Fetal Neonatal Med 2018 31 01 93 97 10.1080/14767058.2016.1276558 28076991 35 Tetileanu A V Berceanu C Paitici Ş Ciurea L E Berceanu S Prenatal, obstetric and perinatal aspects in pregnancy associated with pregestational diabetes Curr Health Sci J 2017 43 04 376 380 10.12865/CHSJ.43.04.15 30595906 36 Wang H Wender-Ozegowska E Garne E Morgan M Loane M Morris J K Insulin analogues use in pregnancy among women with pregestational diabetes mellitus and risk of congenital anomaly: a retrospective population-based cohort study BMJ Open 2018 8 02 e014972 10.1136/bmjopen-2016-014972 37 Melamed N Hod M Perinatal mortality in pregestational diabetes Int J Gynaecol Obstet 2009 104 01 S20 S24 10.1016/j.ijgo.2008.11.024 19155005 38 Fröhlich J D Huppertz B Abuja P M König J Desoye G Oxygen modulates the response of first-trimester trophoblasts to hyperglycemia Am J Pathol 2012 180 01 153 164 10.1016/j.ajpath.2011.09.012 22056361 39 Benhalima K Van Crombrugge P Verhaeghe J Vandeginste S Verlaenen H Vercammen C The Belgian Diabetes in Pregnancy Study (BEDIP-N), a multi-centric prospective cohort study on screening for diabetes in pregnancy and gestational diabetes: methodology and design BMC Pregnancy Childbirth 2014 14 226 10.1186/1471-2393-14-226 25015413 40 Sirico A Sarno L Zullo F Martinelli P Maruotti G M Pregestational diabetes and fetal heart rate in the first trimester of pregnancy Eur J Obstet Gynecol Reprod Biol 2019 232 30 32 10.1016/j.ejogrb.2018.11.003 30465928 41 Russell N E Foley M Kinsley B T Firth R G Coffey M McAuliffe F M Effect of pregestational diabetes mellitus on fetal cardiac function and structure Am J Obstet Gynecol 2008 199 03 3120 3.12E9 10.1016/j.ajog.2008.07.016 42 Chen L Yang T Chen L Wang L Wang T Zhao L Risk of congenital heart defects in offspring exposed to maternal diabetes mellitus: an updated systematic review and meta-analysis Arch Gynecol Obstet 2019 300 06 1491 1506 10.1007/s00404-019-05376-6 31713644 43 Nasri H Z Houde Ng K Westgate M N Hunt A T Holmes L B Malformations among infants of mothers with insulin-dependent diabetes: Is there a recognizable pattern of abnormalities? Birth Defects Res 2018 110 02 108 113 10.1002/bdr2.1155 29377640 44 Sharpe P B Chan A Haan E A Hiller J E Maternal diabetes and congenital anomalies in South Australia 1986-2000: a population-based cohort study Birth Defects Res A Clin Mol Teratol 2005 73 09 605 611 10.1002/bdra.20172 16007590 45 Eidem I Stene L C Henriksen T Hanssen K F Vangen S Vollset S E Congenital anomalies in newborns of women with type 1 diabetes: nationwide population-based study in Norway, 1999-2004 Acta Obstet Gynecol Scand 2010 89 11 1403 1411 10.3109/00016349.2010.518594 20929418 46 Canadian Perinatal Surveillance System (Public Health Agency of Canada) Liu S Joseph K S Lisonkova S Rouleau J Van den Hof M Sauve R Association between maternal chronic conditions and congenital heart defects: a population-based cohort study Circulation 2013 128 06 583 589 10.1161/CIRCULATIONAHA.112.001054 23812182 47 Mills J L Baker L Goldman A S Malformations in infants of diabetic mothers occur before the seventh gestational week. Implications for treatment Diabetes 1979 28 04 292 293 10.2337/diab.28.4.292 437367 48 Arjmandnia M Besharati M Rezvan S Studying the determinant factors leading to congenital heart disease in newborns J Educ Health Promot 2018 7 53 10.4103/jehp.jehp_146_17 29693034 49 Janssen P A Rothman I Schwartz S M Congenital malformations in newborns of women with established and gestational diabetes in Washington State, 1984-91 Paediatr Perinat Epidemiol 1996 10 01 52 63 10.1111/j.1365-3016.1996.tb00026.x 8746431 50 Zhong J Wang S Shen W B Kaushal S Yang P The current status and future of cardiac stem/progenitor cell therapy for congenital heart defects from diabetic pregnancy Pediatr Res 2018 83 (1-2):275 282 10.1038/pr.2017.259 29016556 51 Asoglu M R Yao R Seger L Turan O M Turan S Applicability of standardized early fetal heart examination in the obese population J Ultrasound Med 2019 38 05 1269 1277 10.1002/jum.14807 30251391 52 Patel S S Burns T L Nongenetic risk factors and congenital heart defects Pediatr Cardiol 2013 34 07 1535 1555 10.1007/s00246-013-0775-4 23963188 53 Egan A M Dennedy M C Al-Ramli W Heerey A Avalos G Dunne F ATLANTIC-DIP: excessive gestational weight gain and pregnancy outcomes in women with gestational or pregestational diabetes mellitus J Clin Endocrinol Metab 2014 99 01 212 219 10.1210/jc.2013-2684 24187402 54 Fernandes R S Simões A F Figueiredo A C Ribeiro A RS Aleixo F MF Aragüés S MDTG [Pregnancy outcomes in women with pre-existing diabetes] Rev Bras Ginecol Obstet 2012 34 11 494 498 10.1590/s0100-72032012001100003 23288259 55 Wong S F Petersen S G Idris N Thomae M McIntyre H D Ductus venosus velocimetry in monitoring pregnancy in women with pregestational diabetes mellitus Ultrasound Obstet Gynecol 2010 36 03 350 354 10.1002/uog.7744 20617505 56 Günter H H Scharf A Tzialidou I Hillemanns P Wenzlaff P Maul H [Hospitalization, type of obstetrical clinic and anamnestic risk profile of pregnant women with pregestational diabetes mellitus–results of the perinatal registry of Lower Saxony, Germany] Zentralbl Gynäkol 2006 128 06 341 346 10.1055/s-2006-955205German.17213973 57 Northern Diabetic Pregnancy Survey Steering Group Bell R Bailey K Cresswell T Hawthorne G Critchley J Lewis-Barned N Trends in prevalence and outcomes of pregnancy in women with pre-existing type I and type II diabetes BJOG 2008 115 04 445 452 10.1111/j.1471-0528.2007.01644.x 18271881 58 Van Zyl H Levitt N S Pregnancy outcome in patients with pregestational and gestational diabetes attending Groote Schuur Hospital, Cape Town, South Africa S Afr Med J 2018 108 09 772 776 10.7196/SAMJ.2018.v108i9.12992 30182903 59 Klemetti M Nuutila M Tikkanen M Kari M A Hiilesmaa V Teramo K Trends in maternal BMI, glycaemic control and perinatal outcome among type 1 diabetic pregnant women in 1989-2008 Diabetologia 2012 55 09 2327 2334 10.1007/s00125-012-2627-9 22752076 60 Lucas M J Diabetes complicating pregnancy Obstet Gynecol Clin North Am 2001 28 03 513 536 10.1016/s0889-8545(05)70215-1 11512498 61 Yehuda I Nagtalon-Ramos J Trout K Fetal growth scans and amniotic fluid assessments in pregestational and gestational diabetes J Obstet Gynecol Neonatal Nurs 2011 40 05 603 614, quiz 614–616.10.1111/j.1552-6909.2011.01283.x 62 García-Domínguez M Herranz L Hillman N Martín-Vaquero P Jáñez M Moya-Chimenti E Use of insulin lispro during pregnancy in women with pregestational diabetes mellitus Med Clin (Barc) 2011 137 13 581 586 10.1016/j.medcli.2010.11.021 21376350 63 Lloreda-García J M Sevilla-Denia S Rodríguez-Sánchez A Muñoz-Martínez P Díaz-Ruiz M Perinatal outcome of macrosomic infants born to diabetic versus non-diabetic mothers Endocrinol Nutr 2016 63 08 409 413 10.1016/j.endonu.2016.04.010 27267696 64 Tuuli M G Kapalka K Macones G A Cahill A G Three-versus two-dimensional sonographic biometry for predicting birth weight and macrosomia in diabetic pregnancies J Ultrasound Med 2016 35 09 1925 1930 10.7863/ultra.15.08032 27466257 65 Moodley S Arunamata A Stauffer K J Nourse S E Chen A Quirin A Maternal arterial stiffness and fetal cardiovascular physiology in diabetic pregnancy Ultrasound Obstet Gynecol 2018 52 05 654 661 10.1002/uog.17528 28508434 66 Easter S R Rosenthal E W Morton-Eggleston E Nour N Tuomala R Zera C A Disparities in care for publicly insured women with pregestational diabetes Obstet Gynecol 2017 130 05 946 952 10.1097/AOG.0000000000002252 29016514 67 Lauenborg J Mathiesen E Ovesen P Wetergaard J G Ekbom P Mølsted-Pedersen L Audit on stillbirths in women with pregestational type 1 diabetes Diabetes Care 2003 26 05 1385 1389 10.2337/diacare.26.5.1385 12716793 68 Ganer Herman H Dekalo A Jubran L Schreiber L Bar J Kovo M Obstetric outcomes and placental findings in gestational diabetes patients according to maternal prepregnancy weight and weight gain J Matern Fetal Neonatal Med 2019 32 10 1682 1687 10.1080/14767058.2017.1416078 29402158
PMC009xxxxxx/PMC9708404.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36446558 10.1055/s-0042-1756149 RBGO-21-0387 Original Article High Risk Pregnacy Inflammatory Bowel Disease and Pregnancy: Is It a Marker for Adverse Outcomes? Doença inflamatória intestinal e gravidez: Será um marcador de desfechos adversos?http://orcid.org/0000-0002-1943-1316 Costa Rita Vicente 1 http://orcid.org/0000-0002-9694-1563 Simões Carolina 2 http://orcid.org/0000-0002-3443-5366 Correia Luís 2 http://orcid.org/0000-0001-9003-0868 Pinto Luísa 3 1 Gynecology and Obstetris Service, Hospital Distrital de Santarém, Santarém, Portugal 2 Gastroenterology Service, Hospital de Santa Maria, Lisboa, Portugal 3 Gynecology and Obstetrics Service, Hospital de Santa Maria, Lisboa, Portugal Address for correspondence Rita Vicente Costa, MSc Av. Bernardo Santareno, Santarém [email protected] 29 11 2022 10 2022 1 11 2022 44 10 915924 05 10 2021 21 6 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  To assess obstetric/puerperal/neonatal outcomes in an inflammatory bowel disease (IBD) population and to analyze disease characteristics that may be associated to adverse outcomes. Methods  Retrospective descriptive analysis including 47 pregnant wom e n with IBD (28 with Crohn's disease – CD and 19 with ulcerative colitis – UC) who delivered between March 2012 and July 2018 in a tertiary hospital. We reviewed clinical records to extract demographic information, previous medical history, disease subtype, activity, severity, treatment, and obstetric, puerperal, and neonatal outcome measures. Results  Obstetric and neonatal complications (composite outcomes) occurred in 55.3% and 14.6% of the IBD population, respectively, and were more frequent in UC patients. Preterm birth (PTB), preeclampsia, anemia, low birth weight (LBW), and neonatal death were also more frequent in UC patients. The rate of postpartum hemorrhage (PPH) was 14.9%, and it was higher in CD patients. Women with active IBD had more obstetric/neonatal adverse outcomes (fetal growth restriction and LBW in particular) and cesarean sections. Patients with medicated IBD had less obstetric/neonatal complications (PTB and LBW in specific) and cesarean sections but more PPH. Conclusion  Women with IBD may have an increased risk of obstetric/puerperal/neonatal adverse outcomes. Ulcerative colitis patients had more obstetric and neonatal complications, whereas PPH was more frequent if CD patients. Other disease characteristics were considered, which allowed a better understanding of their possible influence. Although more research is needed, this work reinforces the importance of adequate surveillance to allow prompt recognition and treatment of complications. Resumo Objetivo  Avaliar os desfechos obstétricos/puerperais/neonatais em uma população com doença inflamatória intestinal (DII) e analisar as características da doença, que podem estar associadas a desfechos adversos. Métodos  Análise descritiva retrospectiva incluindo 47 gestantes com DII (28 com doença de Crohn – DC e 19 com retocolite ulcerativa – RCU) que deram à luz entre março de 2012 e julho de 2018 em um hospital terciário. Revisamos os registros clínicos para extrair informações demográficas, histórico médico prévio, subtipo da doença, atividade, gravidade, tratamento e medidas de resultados obstétricos, puerperais e neonatais. Resultados  As complicações obstétricas e neonatais (desfechos compostos) ocorreram em 55,3% e 14,6% da população com DII, respectivamente; e foram mais frequentes em pacientes com RCU. Nascimento prematuro (PTB), pré-eclâmpsia, anemia, baixo peso ao nascer (BPN) e óbito neonatal também foram mais frequentes em pacientes com RCU. A taxa de hemorragia pós-parto (HPP) foi de 14,9% e foi maior em pacientes com DC. Mulheres com DII ativa tiveram mais desfechos obstétricos/neonatais adversos (restrição de crescimento fetal e BPN em particular) e cesarianas. Pacientes com DII medicada tiveram menos complicações obstétricas/neonatais (PTB e BPN em específico) e cesarianas, mas mais HPP. Conclusão  Mulheres com DII podem ter um risco aumentado de desfechos adversos obstétricos/puerperais/neonatais. As pacientes com RCU apresentaram mais complicações obstétricas e neonatais, enquanto a HPP foi mais frequente em pacientes com DC. Outras características da doença foram consideradas, o que permitiu uma melhor compreensão de sua possível influência. Embora mais pesquisas sejam necessárias, este trabalho reforça a importância de uma vigilância adequada para permitir o reconhecimento e o tratamento imediatos das complicações. Keywords inflammatory bowel diseases pregnancy retrospective studies Palavras-chave doenças inflamatórias intestinais gravidez estudos retrospectivos ==== Body pmcIntroduction Inflammatory bowel disease (IBD) is characterized by a chronic non-infectious inflammation of the gastrointestinal tract, and it includes Crohn's disease (CD), ulcerative colitis (UC), and indeterminate colitis. 1 The prevalence in Portugal is around 40 to 60 per 100,000. 1 The peak incidence coincides with reproductive age. 2 Inflammatory bowel disease can reduce fertility, especially if the disease is active and is associated with an increased risk of obstetric complications due to malnutrition, inflammation, and medication. 2 3 Some studies indicate that IBD may contribute to the increased incidence of preterm birth (PTB), low birth weight (LBW), and antepartum hemorrhage, but the conclusions of the articles are not unanimous. 4 5 6 7 8 9 Although there is also some controversy, most publications report that IBD does not confer a greater risk of congenital anomalies. The incidence of preeclampsia does not seem to be affected by IBD. 10 Adverse obstetric outcomes are more frequent if the disease is active in the periconceptional period or if there are exacerbations during pregnancy, so the risk of discontinuing the medication may be greater than the possible adverse effects. 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 However, some of the drugs used in the treatment of IBD can affect fertility or induce pregnancy complications and should be replaced by other alternatives when planning a future pregnancy. Cesarean delivery tends to be more frequent in women with IBD and is indicated in cases of active perianal disease. 4 26 In other cases of IBD, the type of delivery should be dictated by obstetric criteria. This study analyses the potential effect of IBD in obstetric, puerperal, and perinatal outcomes, reinforcing the importance of adequate gastroenterological and obstetric surveillance. The type of IBD (CD or UC), disease activity, extent, treatment, and other IBD characteristics were also described, which provides additional information that most previous studies did not include. Methods We carried out a retrospective descriptive analysis study in which we included 47 pregnant women with IBD (28 with CD and 19 with UC) followed at the Maternal-Fetal Medicine Department at Santa Maria Hospital, who delivered between March 2012 and July 2018. We described obstetric, puerperal, and neonatal adverse outcomes in our IBD sample in its totality but also in addressing several disease particularities (subtype, activity, severity, duration, and treatment). We reviewed medical records from gastrenterology and obstetric routine appointments and urgency room visits to extract the demographic information, previous medical history, disease type/activity/severity/therapy, and outcome measures related to pregnancy, delivery, puerperium, and newborn. The software used for statistical analysis was the IBM SPSS Statistics for Windows version 23.0 (IBM Corp, Armonk, NY, USA). This study was approved by the institutional review board. Obstetric complications included PTB (birth before 37 weeks of gestation), vaginal bleeding in the 3 rd trimester, fetal growth restriction (FGR), spontaneous and therapeutic abortion, stillbirth, placental abruption (placental detachment from the uterus before delivery), premature rupture of membranes (which, by definition, occurs before labour), (PROM)—at term and preterm—prolonged PROM (if >18 hours prior to delivery), chorioamnionitis, preeclampsia, eclampsia, congenital anomalies, intrahepatic cholestasis of pregnancy, anemia, oligohydramnios, short cervix, and gestational diabetes. Abortion was defined as a nonviable intrauterine pregnancy up to 20 weeks of gestation (spontaneous if without intervention, or therapeutic if it was a purposed termination for medical reasons). Stillbirth refers to delivery of a fetus ≥ 20 weeks of gestation with no signs of life. Fetal growth restriction was defined as estimated fetal weight below 3 rd centile or below 10 th centile when associated to fluxometric changes. Congenital anomalies comprise a wide range of abnormalities of body structure or function that are present at birth and are of prenatal origin. Preeclampsia refers to the new onset of hypertension (systolic/diastolic blood pressure ≥ 140 mmHg/90 mmHg in 2 occasions 4 hours apart and proteinuria or the new onset of hypertension and significant end-organ dysfunction with or without proteinuria after 20 weeks of gestation or postpartum in a previously normotensive woman). Eclampsia refers to the occurrence of new-onset, generalized, tonic-clonic seizures or coma in a patient with preeclampsia or gestational hypertension. Intrahepatic cholestasis of pregnancy was diagnosed when there was presence of pruritus associated with elevated total serum bile acid levels, elevated aminotransferases or both, and the absence of diseases that may produce similar laboratory findings and symptoms. Anemia during pregnancy was defined by hemoglobin < 11 g/dL in the first or third trimester or < 10.5 g/dl in the second trimester. The diagnosis of oligohydramnios and short cervix was made either by routine ultrasound or ultrasound in the emergency room. Oligohydramnios was diagnosed by obstetric ultrasound when single deepest pocket < 2 cm or amniotic fluid index ≤ 5 cm. Short cervix was detected when transvaginal ultrasound cervical length was ≤ 25 mm before 24 weeks of gestation. Gestational diabetes was diagnosed by the results of routine blood tests (either if fast glucose levels ≥ 92 mg/dl in the first trimester or between 24 th and 28 th weeks of gestation) or if glucose levels ≥ 180 mg/dl or ≥ 153 mg/dl 1 or 2 hours after ingesting 75 g of glucose, respectively (between 24 th and 28 th weeks of gestation). Chorioamnionitis was defined as maternal fever (axilary temperature ≥ 38°C or tympanic temperature ≥ 38.5°C) associated with at least 2 of the following: maternal tachycardia (> 100 bpm)/fetal tachycardia (> 160 bpm), painful uterine palpation, purulent vaginal discharge/amniotic fluid with fetid odor or leucocitosys (> 15 × 10 ^9 /L)/reactive protein-C elevation (> 10 mg/L). Fetal growth restriction, oligohydramnios, and short cervix were consistently investigated in routine ultrasounds as well as anemia and gestational diabetes, in routine blood tests. We revised clinical files from routine appointments and emergency room visits to verify the presence or absence of the remaining obstetric adverse outcome. Postpartum hemorrhage (PPH) was evaluated as the perception of ≥ 500 ml of blood after a vaginal birth or ≥ 1,000 ml of blood after a cesarean delivery. Neonatal adverse outcomes, as a composite outcome, included: LBW (birth weight < 2,500 g), Apgar index < 7 (in the 1 st , 5 th or 10 th minute after birth), neonatal death (death in the first 28 days of life), and neonatal intensive care unit admission. Results Forty-seven pregnant women with IBD were included (28 with CD and 19 with UC). Their mean age was 33.7 ± 3.7 years. The sample included 25 nulliparous and 22 multiparous patients as well as 45 single-born and 2 twin pregnancies. Some of the women had other relevant diseases or history (4 smokers, 4 with asthma, 1 with previous conization, 1 with septate uterus, 1 with previous uterine perforation, 5 with obesity, 3 with chronic anemia and 1 with chronic gastritis and osteoporosis). Extraintestinal manifestations were reported in 30.8% (n = 12) of the women (7.7%, n = 3 musculoskeletal, 7.7%, n = 3 dermatological/oral and 7.7%, n = 3 hematological, 5.1%, n = 2 hepatobiliary and 2.6%, n = 1 with > 1 extraintestinal manifestation). We collected the necessary data to characterize IBD with the Montreal classification in the 18 cases for which this information was available. Regarding the location of CD, the terminal ileon and the ileon and colon were the most affected areas (37.5%, n = 6 each), followed by the colon alone (25%, n = 4). We also evaluated the behavior of CD. In 43.8% (n = 7) of the cases, the disease was non-stenosing and non-penetrating; in 25% (n = 4), it was strict; in 12.5% (n = 2), it was penetrating; and in 18.8%, (n = 3) a combination of the above was described. As for the extension of the UC, it was characterized only in 2 cases: 1 with left colitis and 1 with pancolitis. Active perianal disease was documented in 29.4% of cases (n = 10), which constituted an indication for elective cesarean section. The age at diagnosis was ≤ 16 years in 5.6% (n = 1) and 17 to 40 years in 94.4% (n = 17) of cases. The average time since the diagnosis of IBD was 7.61 years. The diagnosis of IBD had been made ≤ 5 years ago in 41.7% (n = 15) of pregnant women, > 5 years ago in 58.3% (n = 21), and > 10 years ago in 36.1% (n = 13). During the periconceptional period, according to the endoscopic Mayo classification, 60% (n = 3) of pregnant women with UC had a score 3, and 40% (n = 2) had a score 2. Among those with CD, the majority (87%; n = 20) was classified as being in remission, while 13% (n = 3) had mild disease (Harvey-Bradshaw clinical classification). Most pregnant women (85.7%; n = 36) with IBD were medicated for the pathology, and 26.2% (n = 11) were medicated with more than 1 drug. Salicylates was the most used class of drugs (47.6%; n = 20), followed by thiopurines (38.1%; n = 16), biologicals (23.8%; n = 10 - discontinued on average at 23 weeks of pregnancy), corticosteroids (4.76%; n = 2), and antibiotics (2.4%; n = 1). Almost half of those with IBD (45.2%; n = 19) were medicated with immunosuppressants. Among pregnant women with IBD, 25.5% (n = 12) had previously undergone intestinal surgery due to CD. Almost a third (31.4%; n = 11) of pregnant women with IBD had active disease in the periconceptional period, of which 81.8% (n = 9) remained active during pregnancy. Overall, 41.7% (n = 15) of pregnant women manifested IBD activity during pregnancy. Among pregnant women with inactive disease in the periconceptional period, 25% (n = 6) had exacerbation of the disease during pregnancy. Only 9.1% (n = 3) of pregnant women had active disease in the puerperium and, in all of these cases, IBD activity had already been documented during pregnancy. Obstetric complications, evaluated as a composite outcome (PTB, vaginal bleeding in the 3 rd trimester, FGR, placental abruption, PROM at term and preterm, prolonged PROM, spontaneous and therapeutic abortions, stillbirths, congenital anomalies, chorioamnionitis, preeclampsia, eclampsia, intrahepatic cholestasis of pregnancy, anemia, oligohydramnios, short cervix and gestational diabetes) occurred in 55.3% (n = 26) of pregnant women with IBD, and women with UC had a higher rate of obstetric adverse outcomes than those with CD (68.4%, n = 13 vs 46.4%, n = 13) ( Table 1 ). Table 1 Obstetric complications IBD (n = 47) CD (n = 28) UC (n = 19) PROM* at term 19.1% (9) 21.4% (6) 15.8% (3) Preterm birth 8.5% (4) 7.1% (2) 10.5% (2) Fetal growth restriction 10.6% (5) 10.7% (3) 10.5% (2) Preterm PROM* 2.1% (1) 3.6% (1) 0% (0) Placental abruption 2.1% (1) 0% (0) 5.3% (1) Preeclampsia 4.3% (2) 3.6% (1) 5.3% (1) Intrahepatic cholestasis of pregnancy 8.5% (4) 7.1% (2) 7.1% (2) Anemia 4.3% (2) 0% (0) 7.1% (2) Vaginal bleeding in the third trimester 4.3% (2) 3.6% (1) 5.3% (1) Oligohydramnios 2.1% (1) 3.6% (1) 0% (0) Short cervix 6.4% (3) 7.1% (2) 5.3% (1) Gestational diabetes 8.5% (4) 14.3% (4) 0% (0) Therapeutic abortion (selective feticide) 2.1% (1) 3.6% (1) 0% (0) Abbreviations: CD, Crohn's disease; IBD, inflammatory bowel disease; PROM, premature rupture of membranes; UC, ulcerative colitis. After a joint decision with the parents, 1 selective feticide was performed at 18 weeks (by ultrasound-guided umbilical cord occlusion) due to severe FGR in a monochorionic twin pregnancy, so there were only 48 live births. There were no spontaneous abortions, stilllbirth, cases of eclampsia, chorioamnionitis, prolonged rupture of membranes, or congenital anomalies. Preterm and term PROM, FGR, oligohydramnios, short cervix, gestational diabetes, and feticide were more frequent in CD than in UC ( Table 1 ). Preterm birth, placental abruption, preeclampsia, anemia, and vaginal bleeding in the third trimester were more prevalent in pregnant women with UC than in those with DC. The same prevalence was observed in the DC and CU groups regarding intrahepatic cholestasis of pregnancy (7.1%) ( Table 1 ). Operative delivery, which included vaccum and forceps deliveries as well as all cesarean sections, was more frequent in pregnant women with CD (78.6% in CD vs 65% in UC) ( Table 2 ). Table 2 Route of delivery Route of delivery IBD (n = 48) CD (n = 28) UC (n = 20) Operative delivery 72.9% (35) 78.6% (22) 65% (13) Cesarean section 39.6% (19) 50% (14) 25% (5) Elective 68.4% (13) 85.7% (12) 20% (1) Urgent/emergent 31.6% (6) 14.3% (2) 80% (4) Forceps delivery 6.2% (3) 10.7% (3) 0% (0) Vaccum delivery 27.1% (13) 17.9% (5) 40% (8) Eutocic delivery 27.1% (13) 21.4% (6) 35% (7) Abbreviations: CD, Crohn's disease; IBD, inflammatory bowel disease; UC, ulcerative colitis. *Operative delivery included vaccum delivery, forceps delivery and all cesarean sections There were 13 elective cesarean sections: 10 due to active perianal disease, 1 due to a previous uterine perforation, 1 due to pelvic fetal presentation and 1 due to 2 previous cesarean sections. Within the remaining cesarean sections, 1 was emergent due to placental abruption and 5 were urgent (2 due to labour dystocia and 3 to failed induction of labour). Prolonged second stage of labour was the indication for 3 forceps and 11 vaccum deliveries. The remaining 2 vaccum deliveries were motivated by fetal bradicardia in the 2 nd stage of labour. Considering IBD characteristics, obstetric complications (composite outcome) were more frequent in women with active IBD in the periconceptional period (63.6% vs 45.8%), IBD duration > 5 years (52.4% vs 46.7%), or active perianal disease (60% vs 45.8%), but not with more severe IBD or with activity during pregnancy ( Table 3 ). Specifically regarding PTB, it was more frequent if IBD duration > 5 years (14.3% vs 6.7% if ≤ 5 years), more severe UC (33.3% if score Mayo 3 vs 0% if score 2), or active perianal disease (10% vs 8.3%) ( Table 3 ). Fetal growth restriction had a higher occurrence in women with active IBD in the periconceptional period (but not during pregnancy), more recent IBD diagnosis, or active perianal disease ( Table 3 ). We found a higher frequency of operative delivery in cases of active IBD in the periconceptional period (100%, n = 11 vs 66.7%, n = 16 if IBD in remission) and pregnancy (80% vs 76.2%), fewer time since IBD diagnosis (80% if ≤ 5 years vs 76.2% if > 5 years), more severe CD (100% if mild disease vs 80% if disease in remission), or active perianal disease (100% vs 75%) ( Table 3 ). We also verified more cesarean sections in pregnant women with active IBD in the periconceptional period (63.6%, n = 7 vs 29.2%, n = 7 if IBD in remission), diagnosis of IBD > 5 years ago (52.4%, n = 11 vs 20%, n = 3 in those with diagnosis ≤ 5 years ago), more severe IBD (0% if UC with Mayo score 2 vs 66.7% if Mayo score 3 and 45% if CD in remission vs 100% if mild disease), and also those with active perianal disease (100%, n = 10 vs 25%, n = 6) ( Table 3 ). Neonatal adverse outcomes (composite outcome) were present in 14.6% (n = 7) of the newborns of mothers with IBD, higher if UC patients (20%, n = 4 vs 10.7%, n = 3 if CD). These adverse outcomes were more frequent in cases of active IBD in the periconceptional period and during regnancy (27.3% vs 12.5% and 20% vs 14.3%, respectively), more severe CD (33.3% if mild disease vs 5% CD in remission), and also in cases of IBD diagnosis > 5 years ago (19% vs 13.3% if diagnosis ≤ 5 years), but not in cases of active perianal disease ( Table 3 ). Low birth weight was also more frequent if the disease was active (27.3%, n = 3 vs 4.2%, n = 1 in the periconceptional period and 20%, n = 3 vs 4.8%, n = 1 during pregnancy), diagnosis made > 5 years ago (14.3%, n = 3 vs 6.7%, n = 1 if diagnosis ≤ 5 years ago), and in active perianal disease cases (10%, n = 1 vs 8.3%, n = 2) ( Table 3 ). The rate of PPH in the IBD sample was 14.9% (n = 7), higher in CD patients (17.9%, n = 5) than in women with UC (10.5%, n = 2). The occurrence of PPH was higher in women with more years since the IBD diagnosis (13.3%, n = 2 if ≤ 5 years vs 19%, n = 4 if > 5 years) and more severe CD (33.3%, n = 1 if mild disease vs 15%, n = 3 if CD in remission), but not with active IBD (during the periconceptional period or pregnancy) or perianal disease ( Table 3 ). Table 3 Inflammatory bowel disease characteristics and obstetric/neonatal/puerperal adverse outcomes Obstetric adverse outcomes (composite outcome) a) Preterm birth Fetal growth restriction Operative delivery b) Cesarean section Neonatal adverse outcomes (composite outcome) c) Low birth weight Postpartum hemorrhage IBD activity (periconceptional period) Remission (n = 24) 45.8% (11) 12.5% (3) 4.2% (1) 66.7% (16) 29.2% (7) 12.5% (3) 4.2% (1) 20.8% (5) Active (n = 11) 63.6% (7) 9.1% (1) 18.2% (2) 100% (11) 63.6% (7) 27.3% (3) 27.3% (3) 9.1% (1) IBD activity (pregnancy) Remission (n = 21) 52.4% (11) 14.3% (3) 9.5% (2) 76.2% (16) 45.2% (9) 14.3% (3) 4.8% (1) 23.8% (5) Active (n = 15) 40% (6) 6.7% (1) 6.1% (1) 80% (12) 33.3% (5) 20% (3) 20% (3) 6.7% (1) IBD duration ≤ 5 years (n = 15) 46.7% (7) 6.7% (1) 13.3% (2) 80% (12) 20% (3) 13.3% (2) 6.7% (1) 13.3% (2) > 5 years (n = 21) 52.4% (11) 14.3% (3) 9.5% (2) 76.2% (16) 52.4% (11) 19% (4) 14.3% (3) 19% (4) CD severity (Harvey-Bradshaw classification) d) Remission (n = 20) 45% (9) 10% (2) 10% (2) 80% (16) 45% (9) 5% (1) 0% (0) 15% (3) Mild disease (n = 3) 33.3% (1) 0% (0) 0% (0) 100% (3) 100% (3) 33.3% (1) 33.3% (1) 33.3% (1) UC severity (Mayo classification) e) Score 2 (n = 2) 100% (2) 0% (0) 50% (1) 100% (2) 0% (0) 50% (1) 50% (1) 0% (0) Score 3 (n = 3) 66.7% (2) 33.3% (1) 0% (0) 100% (3) 66.7% (2) 33.3% (1) 33.3% (1) 0% (0) Perianal disease Absent/inactive (n = 24) 45.8% (11) 8.3% (2) 8.3% (2) 75% (18) 25% (6) 16.7% (4) 8.3% (2) 20.8% (5) Active (n = 10) 60% (6) 10% (1) 20% (2) 100% (10) 100% (10) 10% (1) 10% (1) 20% (2) Abbreviations: CD, Crohn's disease; IBD, inflammatory bowel disease; UC, ulcerative colitis. a) Obstetric adverse outcomes (composite outcome) included preterm birth, vaginal bleeding in the 3 rd trimester, fetal growth restriction, spontaneous and therapeutic abortion, stillbirth, placental abruption, premature rupture of membranes at term and preterm, prolonged premature rupture of membranes, preeclampsia, eclampsia, chorioamnionitis, congenital anomalies, intrahepatic cholestasis of pregnancy, anemia, oligohydramnios, short cervix, gestational diabetes. b) Operative delivery included vaccum delivery, forceps delivery and all cesarean sections. c) Neonatal adverse outcomes (composite outcome) included low birth weight, Apgar index < 7, neonatal death and neonatal intensive care unit admission. d) Harvey-Bradshaw classification includes 5 variables (general well-being, abdominal pain, number of liquid dejections per day, abdominal mass and complications), each with an individual score. Total score: < 5 (remission), 5–7 (mild disease), 8–16 (moderate disease) or > 16 (severe disease). e) Mayo endoscopic classification – score 0 (absence of disease or inactive disease – absence of endoscopic alterations), 1 (mild disease – erythema, decreased vascular pattern, mild friability), 2 (moderate disease – marked erythema, absent vascular pattern, friability, erosions) or 3 (severe disease – spontaneous bleeding, ulceration) Women with medicated IBD had less obstetric and neonatal complications (composite outcomes; including PTB and LBW in specific), operative delivery, and cesarean sections but more PPH ( Table 4 ). In women treated specifically with immunossuppressants, the occurrence of obstetric and neonatal complications was also lower (as well as PTB, FGR, and LBW in specific), but not PPH ( Table 4 ). Regarding obstetric adverse outcomes (composite outcome), women treated with salicylates or costicosteroids had a higher rate of complications, as well as women with previous bowel surgery ( Table 4 ). Women treated with salicylates had a higher frequency of operative delivery (90%, n = 18 vs 68.2%, n = 15), as well as those medicated with corticosteroids or biologics (100%, n = 2 vs 80%, n = 32 and 90%, n = 9 vs 78.1%, n = 25, respectively). Neonatal adverse outcomes (composite outcome) were less frequent in women treated with salicylates (10%, n = 2 vs 13.6%, n = 3), thiopurines (6.3%, n = 1 vs 15.4%, n = 4), corticosteroids (0%, n = 0 vs 12.5%, n = 5), or biologics (10%, n = 1 vs 12.5%, n = 4), and also women with previous bowel surgery (0%, n = 0 vs 14.3%, n = 5) ( Table 4 ). Regarding PPH, women treated with salicylates, thiopurines, or biologics had a higher rate of complications, as well as women with previous bowel surgery ( Table 4 ). Table 4 Inflammatory bowel disease treatment and obstetric/neonatal/puerperal adverse outcomes IBD treatment Obstetric adverse outcomes (composite outcome) a) Preterm birth Fetal growth restriction Operative delivery b) Cesarean section Neonatal adverse outcomes (composite outcome) c) Low birth weight Postpartum hemorrhage Medicated IBD Yes (n = 36) 52.8% (19) 8.3% (3) 13.9% (5) 80.6% (30) 41.7% (15) 8.3% (3) 8.3% (3) 19.4% (7) No (n = 6) 66.7% (4) 16.7% (1) 0% (0) 83.3% (5) 50% (3) 33.3% (2) 16.7% (1) 0% (0) Immunossuppressants Yes (n = 19) 42.1% (8) 5.3% (1) 10.5% (2) 78.9% (15) 47.4% (9) 5.3% (1) 0% (0) 26.3% (5) No (n = 23) 65.2% (15) 13% (3) 13% (3) 78.3% (18) 39.1% (9) 17.4% (4) 17.4% (4) 8.7% (2) Salicylates Yes (n = 20) 75% (15) 10% (2) 25% (5) 90% (18) 35% (7) 10% (2) 10% (2) 25% (5) No (n = 22) 40.9% (9) 9.1% (2) 0% (0) 68.2% (15) 50% (11) 13.6% (3) 9.1% (2) 9.1% (2) Thiopurines Yes (n = 16) 50% (8) 6.3% (1) 12.5% (2) 75% (12) 37.5% (6) 6.3% (1) 0% (0) 25% (4) No (n = 26) 50% (13) 11.5% (3) 11.5% (3) 84.6% (22) 46.2% (12) 15.4% (4) 15.4% (4) 11.5% (3) Corticosteroids Yes (n = 2) 100% (2) 0% (0) 0% (0) 100% (2) 50% (1) 0% (0) 0% (0) 0% (0) No (n = 40) 52.5% (21) 10% (4) 12.5% (5) 80% (32) 42.5% (17) 12.5% (5) 10% (4) 17.5% (7) Biologics Yes (n = 10) 50% (5) 10% (1) 0% (0) 90% (9) 60% (6) 10% (1) 0% (0) 30% (3) No (n = 32) 59.4% (19) 9.4% (3) 15.5% (5) 78.1% (25) 37.5% (12) 12.5% (4) 12.5% (4) 12.5% (4) Previous bowel surgery Yes (n = 12) 58.3% (7) 8.3% (1) 16.7% (2) 75% (9) 50% (6) 0% (0) 0% (0) 16.7% (2) No (n = 35) 54.3% (19) 8.6% (3) 8.6% (3) 74.3% (26) 37.1% (13) 14.3% (5) 11.4% (4) 14.3% (5) Abbreviation: IBD, inflammatory bowel disease. a) Obstetric adverse outcomes (composite outcome) included preterm birth, vaginal bleeding in the 3 rd trimester, fetal growth restriction, spontaneous and therapeutic abortion, stillbirth, placental abruption, premature rupture of membranes at term and preterm, prolonged premature rupture of membranes, preeclampsia, eclampsia, chorioamnionitis, congenital anomalies, intrahepatic cholestasis of pregnancy, anemia, oligohydramnios, short cervix, and gestational diabetes. b) Operative delivery included vaccum delivery, forceps delivery, and all cesarean sections. c) Neonatal adverse outcomes (composite outcome) included low birth weight, Apgar index < 7, neonatal death, and neonatal intensive care unit admission Newborns of women with UC had a higher rate of LBW, neonatal death, and neonatal intensive care unit admission than women with CD but not Apgar index < 7 ( Table 5 ). Table 5 Neonatal adverse outcomes IBD (n = 48) CD (n = 28) UC (n = 20) LBW 10.4% (5) 3.6% (1) 20% (4) Apgar index < 7 2.1% (1) 3.6% (1) 0% (0) Neonatal intensive care unit admission 2.1% (1) 0% (0) 5% (1) Neonatal death 2.1% (1) 0% (0) 5% (1) Abbreviations: CD, Crohn's disease; IBD, inflammatory bowel disease; LBW, low birth weight; UC, ulcerative colitis. Discussion In the cohort study carried out by Uma Mahadevan's team, adverse pregnancy outcomes/complications (including PTB, small for gestational age, stillbirth, placental abruption, preeclampsia and PROM, among other outcomes) were more frequent in pregnant women with IBD than in healthy pregnant women. 2 Most studies suggest IBD can contribute to increased frequency of PTB, although they do not specify the degree of disease activity or the therapy used. 4 5 6 7 8 9 12 20 21 26 27 28 29 In our study, PTB rate was 10.5% in women with UC, which is higher than the prevalence found in healthy populations according to most works (4.5–9.5%). 2 7 12 22 26 28 29 However, the frequency of PTB found in women with IBD in general (8.5%) or women with CD (7.1%) was similar to the refered baseline prevalence in the population. 2 7 12 22 26 28 29 Stephansson et al. 28 showed higher occurrence of PTB and small size for gestational age in women with IBD in a population-based prevalence study. However, Bortoli et al. 7 conducted a prospective case-control study that did not show significant differences regarding pregnancy outcomes such as spontaneous and therapeutic abortions, infant death in utero , and PTB between pregnant women with or without IBD. It is important to mention that in Bortoli's study and in the present article most pregnant women with IBD were in remission in the periconceptional period and remained with inactive disease during pregnancy. Spontaneous and/or therapeutic abortions are also another of the outcomes generally studied in pregnant women with IBD, as was the case of the work carried out by Boyd et al., 27 without differences between the groups. 27 In our IBD population, there was 1 selective feticide in a woman with CD, corresponding to 2.1% of IBD women and 3.6% of CD, which is a higher rate than the rate of therapeutic abortion reported by Bortoli et al. 7 in a population without IBD (0.5–1.4%). The possible association of IBD with congenital anomalies is controversial, despite having been addressed in several publications; in the IBD population included in our work there where no cases of congenital anomalies, whereas the prevalence in the general population is higher and disparate in the literature (0.7–7%). 2 4 7 9 12 22 27 28 29 30 Regarding other pregnancy complications included in our study, preeclampsia rates (4.3% in all IBD types, 3.6% if CD and 5.3% if UC) were similar to those reported in healthy populations in the literature (1.8–5.6%). 2 9 12 27 28 31 The placental abruption rates of our IBD population were higher than the ones relative to the general population (2.1% vs 0.4-1.2%), just like those of gestational diabetes (8.5% vs 1.9-5.8%) and PROM (21.2% vs 2.8%), specialy in women with CD (14.3% vs 1.9% regarding gestational diabetes and 25% vs 2.8% respecting PROM). 2 9 26 31 The literature suggests a higher frequency of cesarean sections in pregnant women with IBD compared to healthy ones. 4 5 10 22 26 28 This is also supported by our results: 39.6% of women diagnosed with IBD (50% of those with CD and 25% of women with UC) had cesarean sections, whereas the frequency of c-section in the general population is 9.5–28.2%, according to the literature. 2 7 12 22 26 31 According to our work, operative delivery was more frequent in pregnant women with IBD than in the population without IBD of the Sultan's team study (72.9% vs 35.6%), specially if IBD was active in the periconceptional period (100% vs 35.6%). 31 As expected, all pregnant women with perianal disease had cesarean section, as this is an indication for this according to the European Crohn's and Colitis organization's (ECCO) guidelines. Cesarean delivery was also more frequent if IBD had been diagnosed > 5 years ago, which may be explained by the cumulative complications of the disease. In the literature, there are few studies that evaluated puerperal outcomes, and, in those that did, PPH was the most studied, and the results were mixed. 26 31 In the IBD population of our study, the rate of PPH was 14.9%, higher than the prevalence described by the work of Sultan et al. 31 —8.7%. The literature is not consensual regarding adverse neonatal outcomes. Some studies observed a higher frequency of LBW in children of mothers with IBD. 4 20 29 These results were supported by our work: newborns of women with IBD, and speciffically UC, had higher rates of LBW (10.4% and 20%, respectively) than the general population (4.5–10%), unlike newborns of women with CD (3.6%). 2 29 31 In one publication, UC was associated with neonatal death, and in our IBD population, the prevalence of neonatal death in IBD women, specially if UC (2,1% and 5%, respectively), was higher than the rate reported in the literature for women without IBD (0.2%). 2 22 In our work, the rate of Apgar index < 7 in newborns of women with IBD was higher than that reported for healthy population in the literature (2.1% vs 0.7–1.4%). 22 27 28 However, other studies did not find differences between the groups regarding adverse neonatal outcomes, either as a composite or isolated outcome. 2 5 6 7 8 9 26 Most publications suggest that disease activity is associated with adverse obstetric outcomes, like in our study, regarding IBD activity in the periconceptional period (but not during pregnancy). 32 33 34 However, in the prospective cohort study by Lima-Karagiannis et al., 35 the activity of IBD did not confer a greater risk of obstetric or neonatal complications. In our study, women with active IBD (in the periconceptional period and during pregnancy) had more operative delivery and neonatal adverse outcomes (as a composite outcome and LBW in especific) but not higher rate of PPH. Inflammatory bowel disease severity and its potential association with obstetric/neonatal/puerperal complications is generally not addressed in publications in this area. In our work, women with more severe CD had higher rates of neonatal adverse outcomes and PPH but not obstetric complications; UC severity did not seem to relate to poor outcomes. However, it is important to state that IBD severity was the parameter with more missing information, and this may lead to more unreliable results. Women with diagnosis of IBD > 5 years prior had more obstetric and neonatal complications, assessed as composite outcomes (including PTB and LBW individually), and also more PPH. To justify these results, we hypothesize that women with a longer time since diagnosis are more likely to accumulate complications of the disease than women with a more recent diagnosis. Women with medicated IBD had less obstetric/neonatal complications (composite outcomes), including PTB and LBW individually, probably because the treatment leads to better control of the disease symptoms, activity, and/or severity. Women treated with salicylates had a higher frequency of operative delivery (90% vs 68.2%), and this may be explained by the fact that salicylates are a non-steroidal antiinflammatory substance, and it inhibits the synthesis of prostaglandins, which may contribute to the inhibition of labor and its prolonged duration, leading to a higher incidence of operative delivery. In this study's population, some pregnant women with IBD had medical comorbidities theoretically associated with a higher risk of PTB, LBW, and cesarean section, which could be a confounding factor in the interpretation of the results. One of the biases of this work, which comes from its retrospective nature, was the impossibility of collecting information on the degree of activity, severity, duration and therapy of IBD, as well as the presence/absence of perianal disease for all pregnant women with IBD. This aspect allied to the heterogenity of the IBD group may have contributed to possibly inaccurate and sometimes disparate results from the available literature. However, we consider that the analysis of these particularities of the IBD constituted a strong point of this work, as they are generally not addressed in publications in this area. Also, the fact that the sample refers to a tertiary center offers a different perspective from that usually portrayed in the literature—even more heterogeneous samples often coming from national databases, pregnant women less medicated and rarely undergoing biological therapy (either because of the length of the studies or because the severity of the disease does not justify it). 26 This disparity between the studies, not only in terms of methodology but also in sampling, may explain the differences in results. All in all, it is important to mention that all the results should be considered in a very specific context and their interpretation is limited, not only because of all the factors mentioned before but also because the sample was relatively small and heterogenous. Conclusion The ocurrence of some obstetric complications (placental abruption, gestational diabetes, and PROM) was higher in the IBD population (and specialy if CD) in this work than the prevalence described in the literature for healthy individuals, and the same tendency was verified with PTB but only in UC cases. 2 7 9 12 22 26 28 29 31 Other adverse obstetric outcomes in women with IBD had similar prevalences to the ones described in the literature for healthy pregnant women (PTB and preeclampsia), or even lower (congenital anomalies). 2 4 7 9 12 22 26 27 28 29 31 Regarding cesarean sections, the rate in our IBD population was higher than the reported in the literature for the general population, specialy if CD. 2 7 12 22 26 31 The same tendency occurred with PPH and some neonatal adverse outcomes (LBW, Apgar index < 7 and neonatal death). 2 22 27 28 29 31 Obstetric and neonatal complications (assessed as composite outcomes) were more frequent in women with UC than with CD but the opposite tendency was verified regarding PPH. Cesarean sections in the IBD population were more frequent if CD, active IBD in the periconceptional period, diagnosis of IBD > 5 years ago or active perianal disease. Women with active IBD (in the periconceptional period or during pregnancy) had more neonatal adverse outcomes (as a composite outcome and LBW individually) but not higher rate of PPH. Diagnosis of IBD >5 years prior may be associated to more obstetric and neonatal complications, assessed as composite outcomes, including PTB and LBW in particular. Women with more severe CD had higher rates of neonatal adverse outcomes (composite outcome) and PPH but not obstetric complications (composite outcome). Ulcerative colitis severity did not seem to relate to poor outcomes. Women with medicated IBD had less obstetric and neonatal complications (composite outcomes), including PTB and LBW individually, probably because the treatment lead to better control of the disease symptoms, activity, and/or severity. This work reinforces the importance of therapeutic optimization and control of IBD prior to pregnancy and adequate gastroenterological surveillance during pregnancy. We also highlight the importance of obstetric and postpartum vigilance in this population, allowing early diagnosis and treatment of potential complications. However, specially due to the retrospective nature of this study and to a small and heterogenous sample, these results should be interpreted carefully. Contributors Conflict of Interests The authors have no conflict to declare. All authors were involved in the design and interpretation of the analyses, contributed to the writing of the manuscript, and read and approved the final manuscript. ==== Refs References 1 GBD 2017 Inflammatory Bowel Disease Collaborators The global, regional, and national burden of inflammatory bowel disease in 195 countries and territories, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017 Lancet Gastroenterol Hepatol 2020 5 01 17 30 10.1016/S2468-1253(19)30333-4 31648971 2 Mahadevan U Sandborn W J Li D K Hakimian S Kane S Corley D A Pregnancy outcomes in women with inflammatory bowel disease: a large community-based study from Northern California Gastroenterology 2007 133 04 1106 1112 10.1053/j.gastro.2007.07.019 17764676 3 European Crohn's and Colitis Organisation (ECCO) Van Assche G Dignass A Reinisch W van der Woude C J Sturm A De Vos M The second European evidence-based Consensus on the diagnosis and management of Crohn's disease: Special situations J Crohn's Colitis 2010 4 01 63 101 10.1016/j.crohns.2009.09.009 21122490 4 Cornish J Tan E Teare J Teoh T G Rai R Clark S K A meta-analysis on the influence of inflammatory bowel disease on pregnancy Gut 2007 56 06 830 837 10.1136/gut.2006.108324 17185356 5 Nguyen G C Boudreau H Harris M L Maxwell C V Outcomes of obstetric hospitalizations among women with inflammatory bowel disease in the United States Clin Gastroenterol Hepatol 2009 7 03 329 334 10.1016/j.cgh.2008.10.022 19027089 6 Bröms G Granath F Linder M Stephansson O Elmberg M Kieler H Complications from inflammatory bowel disease during pregnancy and delivery Clin Gastroenterol Hepatol 2012 10 11 1246 1252 10.1016/j.cgh.2012.08.018 22922307 7 European Crohn-Colitis Organisation (ECCO) Study Group of Epidemiologic Committee (EpiCom) Bortoli A Pedersen N Duricova D D'Inca R Gionchetti P Panelli M R Pregnancy outcome in inflammatory bowel disease: prospective European case-control ECCO-EpiCom study, 2003-2006 Aliment Pharmacol Ther 2011 34 07 724 734 10.1111/j.1365-2036.2011.04794.x 21815900 8 Bröms G Granath F Linder M Stephansson O Elmberg M Kieler H Birth outcomes in women with inflammatory bowel disease: effects of disease activity and drug exposure Inflamm Bowel Dis 2014 20 06 1091 1098 10.1097/MIB.0000000000000060 24810137 9 Getahun D Fassett M J Longstreth G F Koebnick C Langer-Gould A M Strickland D Association between maternal inflammatory bowel disease and adverse perinatal outcomes J Perinatol 2014 34 06 435 440 10.1038/jp.2014.41 24651735 10 Tandon P Govardhanam V Leung K Maxwell C Huang V Systematic review with meta-analysis: risk of adverse pregnancy-related outcomes in inflammatory bowel disease Aliment Pharmacol Ther 2020 51 03 320 333 10.1111/apt.15587 31912546 11 Woolfson K Cohen Z McLeod R S Crohn's disease and pregnancy Dis Colon Rectum 1990 33 10 869 873 10.1007/BF02051924 2209277 12 Nørgård B Fonager K Sørensen H T Olsen J Birth outcomes of women with ulcerative colitis: a nationwide Danish cohort study Am J Gastroenterol 2000 95 11 3165 3170 10.1111/j.1572-0241.2000.03290.x 11095336 13 Reddy D Murphy S J Kane S V Present D H Kornbluth A A Relapses of inflammatory bowel disease during pregnancy: in-hospital management and birth outcomes Am J Gastroenterol 2008 103 05 1203 1209 10.1111/j.1572-0241.2007.01756.x 18422816 14 Baiocco P J Korelitz B I The influence of inflammatory bowel disease and its treatment on pregnancy and fetal outcome J Clin Gastroenterol 1984 6 03 211 216 6144706 15 Khosla R Willoughby C P Jewell D P Crohn's disease and pregnancy Gut 1984 25 01 52 56 10.1136/gut.25.1.52 6140209 16 Mayberry J F Weterman I T European survey of fertility and pregnancy in women with Crohn's disease: a case control study by European collaborative group Gut 1986 27 07 821 825 10.1136/gut.27.7.821 3732892 17 Fedorkow D M Persaud D Nimrod C A Inflammatory bowel disease: a controlled study of late pregnancy outcome Am J Obstet Gynecol 1989 160 04 998 1001 10.1016/0002-9378(89)90324-4 2712128 18 Nørgård B Hundborg H H Jacobsen B A Nielsen G L Fonager K Disease activity in pregnant women with Crohn's disease and birth outcomes: a regional Danish cohort study Am J Gastroenterol 2007 102 09 1947 1954 10.1111/j.1572-0241.2007.01355.x 17573787 19 Bush M C Patel S Lapinski R H Stone J L Perinatal outcomes in inflammatory bowel disease J Matern Fetal Neonatal Med 2004 15 04 237 241 10.1080/14767050410001668662 15280131 20 Fonager K Sørensen H T Olsen J Dahlerup J F Rasmussen S N Pregnancy outcome for women with Crohn's disease: a follow-up study based on linkage between national registries Am J Gastroenterol 1998 93 12 2426 2430 10.1111/j.1572-0241.1998.00698.x 9860403 21 Kornfeld D Cnattingius S Ekbom A Pregnancy outcomes in women with inflammatory bowel disease–a population-based cohort study Am J Obstet Gynecol 1997 177 04 942 946 10.1016/s0002-9378(97)70298-9 9369849 22 Stephansson O Larsson H Pedersen L Kieler H Granath F Ludvigsson J F Congenital abnormalities and other birth outcomes in children born to women with ulcerative colitis in Denmark and Sweden Inflamm Bowel Dis 2011 17 03 795 801 10.1002/ibd.21369 20564537 23 Elbaz G Fich A Levy A Holcberg G Sheiner E Inflammatory bowel disease and preterm delivery Int J Gynaecol Obstet 2005 90 03 193 197 10.1016/j.ijgo.2005.06.003 16043179 24 Baird D D Narendranathan M Sandler R S Increased risk of preterm birth for women with inflammatory bowel disease Gastroenterology 1990 99 04 987 994 10.1016/0016-5085(90)90617-a 2394353 25 Lin H C Chiu C C Chen S F Lou H Y Chiu W T Chen Y H Ulcerative colitis and pregnancy outcomes in an Asian population Am J Gastroenterol 2010 105 02 387 394 10.1038/ajg.2009.562 19809415 26 Shand A W Chen J S Selby W Solomon M Roberts C L Inflammatory bowel disease in pregnancy: a population-based study of prevalence and pregnancy outcomes BJOG 2016 123 11 1862 1870 10.1111/1471-0528.13946 26924786 27 Boyd H A Basit S Harpsøe M C Wohlfahrt J Jess T Inflammatory bowel disease and risk of adverse pregnancy outcomes PLoS One 2015 10 06 e0129567 26083614 28 Stephansson O Larsson H Pedersen L Kieler H Granath F Ludvigsson J F Crohn's disease is a risk factor for preterm birth Clin Gastroenterol Hepatol 2010 8 06 509 515 10.1016/j.cgh.2010.02.014 20202483 29 Dominitz J A Young J C Boyko E J Outcomes of infants born to mothers with inflammatory bowel disease: a population-based cohort study Am J Gastroenterol 2002 97 03 641 648 10.1111/j.1572-0241.2002.05543.x 11926208 30 Nørgård B Puho E Pedersen L Czeizel A E Sørensen H T Risk of congenital abnormalities in children born to women with ulcerative colitis: a population-based, case-control study Am J Gastroenterol 2003 98 09 2006 2010 10.1111/j.1572-0241.2003.07578.x 14499779 31 Abdul Sultan A West J Ban L Humes D Tata L J Fleming K M Adverse pregnancy outcomes among women with inflammatory bowel disease: a population-based study from England Inflamm Bowel Dis 2016 22 07 1621 1630 10.1097/MIB.0000000000000802 27306070 32 Morales M Berney T Jenny A Morel P Extermann P Crohn's disease as a risk factor for the outcome of pregnancy Hepatogastroenterology 2000 47 36 1595 1598 11149010 33 Nielsen O H Andreasson B Bondesen S Jacobsen O Jarnum S Pregnancy in Crohn's disease Scand J Gastroenterol 1984 19 06 724 732 6515312 34 Nielsen O H Andreasson B Bondesen S Jarnum S Pregnancy in ulcerative colitis Scand J Gastroenterol 1983 18 06 735 742 10.3109/00365528309182088 6669937 35 de Lima-Karagiannis A Zelinkova-Detkova Z van der Woude C J The effects of active IBD during pregnancy in the Era of Novel IBD Therapies Am J Gastroenterol 2016 111 09 1305 1312 10.1038/ajg.2016.254 27349339
PMC009xxxxxx/PMC9708405.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36174653 10.1055/s-0042-1756312 RBGO-21-0393 Review Article Management Strategies for Sexuality Complaints after Gynecologic Cancer: A Systematic Review Estratégias para enfrentamento de queixas da sexualidade após câncer ginecológico: uma revisão sistemáticahttp://orcid.org/0000-0002-6085-4695 Pizetta Luciane Machado 1 http://orcid.org/0000-0002-3200-8096 Reis Augusto da Cunha 1 http://orcid.org/0000-0003-4506-7009 Méxas Mirian Picinini 2 http://orcid.org/0000-0001-7662-3499 Guimarães Vanessa de Almeida 1 http://orcid.org/0000-0001-7457-3803 de Paula Carmen Lucia 1 1 Centro Federal de Educação Tecnológica Celso Suckow da Fonseca, Rio de Janeiro, RJ, Brazil 2 Universidade Federal Fluminense, Niterói, RJ, Brazil Address for correspondence Luciane Machado Pizetta Avenida Maracanã229, 20271-110, Maracanã, Rio de Janeiro, [email protected] 29 9 2022 10 2022 1 9 2022 44 10 962971 20 10 2021 07 7 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  To explore the main sexuality complaints of gynecologic cancer survivors after treatment and to identify the care strategies provided. Data Source  Searches were conducted in six electronic databases: Scopus, Web of Science, LILACS, MEDLINE, PsychINFO, and EMBASE. Study Selection  Articles published between 2010 and 2020 were selected and the following descriptors were used in the English language: female genital neoplasms and gynaecological cancer . The methodological quality of the studies used the Mixed Methods Appraisal Tool (MMAT). Data Collection  The primary data extracted were: names of the authors, year of publication, country of origin, objective and type of study, data collection instrument, sample size and age range, types of cancer, and symptoms affected with the strategies adopted. Data Summary  A total of 34 out of 2,536 screened articles were included. The main strategies found for patient care were patient-clinician communication, practices for sexuality care, individualized care plan, multiprofessional team support, and development of rehabilitation programs. For sexuality care, the most common practices are pelvic physiotherapy sessions and the use of vaginal gels and moisturizers. Conclusion  The main complaints identified in the scientific literature were low libido and lack of interest in sexual activity, vaginal dryness, pain during sexual intercourse, and stenosis. Different care strategies may be adopted, such as follow-up with a multidisciplinary health team and sexual health rehabilitation programs, which could minimize these symptoms and ensure the quality of life of patients. Resumo Objetivo  Explorar as principais queixas da sexualidade com sobreviventes de câncer ginecológico após o tratamento e identificar as estratégias de cuidados prestados. Fonte dos Dados  As buscas foram realizadas em seis bases eletrônicas: Scopus, Web of Science, LILACS, MEDLINE, PsychINFO e EMBASE. Seleção dos Estudos  Foram selecionados artigos publicados entre 2010 e 2020 e os descritores utilizados (em inglês) foram female genital neoplasms e gynaecological cancer . A qualidade metodológica dos estudos utilizou a ferramenta Mixed Methods Appraisal Tool (MMAT). Coleta de Dados  Os principais dados extraídos foram: nomes dos autores, ano de publicação, país de origem, objetivo e tipo de estudo, instrumento para coleta de dados, tamanho da amostra e faixa etária, tipos de câncer, os sintomas acometidos e as estratégias adotadas. Síntese dos Dados  Dos 2,536 artigos identificados, 34 foram incluídos. As principais estratégias encontradas para os cuidados aos pacientes foram a comunicação paciente-médico, práticas para os cuidados sexuais, plano de cuidados individualizado, apoio a equipes multiprofissionais e desenvolvimento de programas de reabilitação. Para os cuidados de sexualidade, as práticas mais comuns são sessões de fisioterapia pélvica e o uso de géis vaginais e hidratantes. Conclusão  As principais queixas identificadas na literatura científica foram baixa libido e falta de interesse na atividade sexual, secura vaginal, dor durante a relação sexual e estenose. Diferentes estratégias de cuidados podem ser adotadas, como o acompanhamento com uma equipe de saúde multidisciplinar e programas de reabilitação da saúde sexual, as quais poderiam minimizar estes sintomas e garantir a qualidade de vida dos pacientes. Keywords survivors gynecological cancer sexuality supportive care needs Palavras-chave sobreviventes câncer ginecológico sexualidade suporte de necessidades de cuidados ==== Body pmcIntroduction Early detection and treatment of gynecologic cancer at an initial stage allow high chances of cure and improved survival and, consequently, an increase in the number of survivors. 1 2 3 4 5 6 7 8 All forms of treatment for gynecologic neoplasia (pelvic surgery, radiation, and drug therapy) have the potential to impair physiological and psychological functions, and also generally affect the self-esteem, body image, femininity, and intimate relationships of women; sexual dysfunction is the most frequently mentioned in the literature. 9 10 11 12 Sexual health is an essential aspect of quality of life. As a result of the aggressive treatment received, > 40% of gynecologic cancer patients reported chronic and distressing sexual difficulties, even after 12 months of treatment. 13 14 15 Sexual dysfunction, in addition to being extremely distressing, negatively affects the relationship with their spouses, which can result in emotional estrangement. 16 17 18 19 Studies indicate that 74% of the surviving participants felt that communication with oncology professionals about sexual issues was important, but few had received this information. 20 21 Many women reported having no discussion about sexuality with their physicians during and after treatment, which represents a troubling gap in the provision of supportive care for health needs. 22 23 Recognizing that post-treatment gynecologic cancer causes various negative impacts on the health of the female sexual organs, it becomes important to seek to understand the subjective phenomena about the experience of sexuality after treatment. 24 25 Recently, topics such as “cancer survivors” and “supporting care needs” have been increasingly published, and many researchers are studying these concepts, addressing dimensions such as “sexuality”, “gynecologic cancer”, and “health dysfunctions after cancer.” 26 27 28 29 30 Various care strategies that can help survivor patients are mentioned in the scientific literature, such as the development of support groups with the purpose of providing an environment of cooperation, solidarity, and readaptation to the disease, multiprofessional care team support to minimize the difficulties associated with the disease, doctor-patient communication, and development of rehabilitation programs. 31 32 33 34 However, there are strategies that may present disadvantages such as the deficiency of clinical training by health professionals to perform an appropriate approach, because many patients have personal difficulties and taboos about sexuality. 35 36 About 65% of gynecologic cancer patients do not have their needs met; little is explored about knowledge on the experience of supportive care for gynecologic cancer survivors or the supportive resources available. 37 In this regard, there is a need for a literature review to investigate new resources that can be developed and incorporated into treatment and healthcare settings for the benefit of patients. The current review aimed to explore the main collateral effects on sexuality with gynecologic cancer survivors after treatment and to identify the care strategies provided. Methods Search Selection Strategy For the development of the present study, a systematic literature review was performed based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocol (PRISMA-P). 38 The search for articles was performed in the following electronic databases: Scopus, Web of Science, LILACS, MEDLINE, PsychINFO, and EMBASE. Searches in the electronic databases was performed from January to February 2021. The search strategies were changed according to the requirements of each database. Initially, two authors chose independently the articles selected for eligibility, following a two-phase process. The initial screening phase resulted in a restricted list of articles, including titles and abstracts. In the second stage, the screening process involved reading the articles in full text. The two reviewers independently assessed all articles for eligibility concerning the selection criteria until consensus was reached. When there was disagreement and no consensus was reached, a third author reviewed the articles and selected the one with the greatest adherence to the research. The qualitative synthesis of the selected papers was based on an integral lecture focusing on post-treatment gynecological neoplasms ( Chart 1 ). Chart 1 Search terms and strategy Search ID# Search terms Search mode S1 female genital neoplasm OR uterine neoplasm OR vulvar neoplasm OR vaginal neoplasm Boolean/Phrase S2 ovarian neoplasm OR female genital cancer OR uterine cancer OR vulvar cancer OR vaginal cancer Boolean/Phrase S3 ovarian cancer OR female genital tumors OR uterine tumors OR vulvar tumors OR vaginal tumors Boolean/Phrase S4 ovarian tumors OR gynaecological cancer OR gynecological cancer OR gynaecological tumors OR gynecological tumors Boolean/Phrase S5 female genital carcinoma OR uterine carcinoma OR vulvar carcinoma OR vaginal carcinoma OR ovarian carcinoma Boolean/Phrase S6 sexual health OR sexuality OR sexual function Boolean/Phrase S7 patient-centered care OR patient care OR supportive care OR supportive care need* OR helping behavior OR humanization OR survivorship OR survival o OR r cancer survivorship Boolean/Phrase Inclusion and Exclusion Criteria To be considered eligible, the following inclusion and exclusion criteria was applied to all identified scientific articles. Inclusion Criteria Studies that investigate care support needs of women survivors of gynecological cancer (post-treatment of the disease) with a focus on sexuality. Studies employing quantitative and/or qualitative methods, regardless of research design. Studies published in English with readily available abstracts. Studies performed with adult individuals ≥ 18 years old diagnosed with gynecological cancer (endometrial, cervical, ovarian, and vulvar) including the effects and aftertreatment mainly affecting sexuality. Articles focusing on the medium and long-term effects of gynecological cancer. Studies published as original articles in specialized journals between 2010 and 2020. Exclusion Criteria Studies in which care support needs of survivors of gynecological cancer focused on sexuality were not explicitly discussed. Studies conducted with patients with other types of cancer, whether primary or secondary. Systematic reviews and individual publications within systematic reviews were excluded so as not to duplicate the content. Data Extraction and Analysis Data from the included studies were extracted into a predefined data extraction table. Data extraction was performed by a reviewer and then checked by a second reviewer jointly and according to the agreement and consistency between them. A third reviewer analyzed discrepancies. Three time periods were established: Immediately after treatment (0 to 3 months post-treatment). Long-term post-treatment effects (> 3 months to 5 years). Mixed (Studies with patients from both time frames, that is, from 0 to 3 months to 5 years post-treatment. The included studies were classified according to the type of studies conducted and to the location of the tumor according to discussions with the research team. The quality of the studies was independently assessed by two reviewers using the Mixed Methods Appraisal Tool (MMAT). Risk of Bias The articles were independently assessed to maintain methodological quality. The negative and positive points were analyzed and, at the end, a score from 0 to 4 was attributed to each article (0 = no criteria fulfilled, 1 = fulfilled one criterion, 2 = fulfilled two criteria, 3 = fulfilled three criteria, and 4 = fulfilled all four criteria). 39 When any disagreement occurred, the authors discussed them to reach a final agreement. Qualitative Synthesis After the search for studies in the databases, a screening was performed independently by the review team by reading the titles and abstracts. This search in scientific databases generated the return of a large number of articles, and the strategy adopted was to ensure sensitivity on the specificity and compliance to the established criteria. Thus, for the screening of studies, the sum of the total number of articles in all databases was recorded; the titles were read quickly, allowing the selection of references, and discarding those that did not fit the eligibility criteria established by the review team. A free reference management software, Mendeley, was used to sort the articles, count duplicate articles, and organize the references, thus providing greater practicality and time optimization. The studies that passed the screening had their full text retrieved, and the eligibility of studies was confirmed after reading the full text and selection of studies that investigate support care needs of survivors of gynecological cancer (post-treatment of the disease) with a focus on sexuality. Results Study Selection Initially, 2,536 potential articles were identified: 102 through Scopus; 153 through PubMed; 1,062 through EMBASE with a “not PubMed” filter; 760 through Medline; 25 through Lilacs; 51 through a concurrent search of CINAHL and PsycINFO on the EBSCO platform; and 383 through Web of Science. Duplicates were removed and 2,329 remaining articles were evaluated based on title and abstract. Most were discarded because they did not meet the inclusion criteria. Only 60 potentially relevant articles were retained and assessed based on full text, leaving 34 included articles ( Fig. 1 ). Fig. 1 Flowchart of the process of search and selection of studies. Study Characteristics The studies included most frequently women who had been diagnosed with gynecological cancer ( n  = 19; 56%). Only 1 study on cervical cancer was conducted with young adult women (18 to 39 years old). The remaining studies ( n  = 5; 12%) reported experiences of gynecological cancer survivors regarding its adverse effects and the care that needs to be received in the context of sexuality. The most common countries where the survey was performed were: USA ( n  = 7; 21%), followed by Canada ( n  = 5; 15%) and the Netherlands ( n  = 4; 12%). Few studies ( n  = 3; 8.8%) included cancer survivors residing in rural or remote locations. Population studies occurred most frequently at an average of 6 months after the end of treatment ( n  = 18; 54%). The average sample size was 180 participants (range: 11 to 1,029). Most studies had a cross-sectional design ( n  = 19; 56%). More than 50% of the included studies ( n  = 19; 56%) were published in the past 5 years. The included studies were mapped using a descriptive synthesis of the domains of the needs, the study population (for example, type of cancer, age), and the post-treatment of the disease with after primary treatment/early survivorship ranging from 1 month to 3 years. Long-term survivorship after 3 years and mixed phase care when samples are composed of survivors with different periods after treatment ( Appendix 1, Supplementary material ). Evaluation of the Methodological Quality of the Studies All 34 studies selected and included in the systematic review were submitted to a methodological assessment by the MMAT instrument; 13 received a score of 4 because they met all criteria in the quality assessment, 19 received a score of 3 because some topics, such as the interpretation of results, were not sufficiently supported by valid data, 2 articles received a score of 2, and no article received the lowest score of 1. Main Side Effects of Sexuality Dominance in Women Survivors According to the selected articles, low libido was identified in 18 studies and was diagnosed as a major problem for gynecological cancer survivors. Therapeutic modalities such as chemotherapy, radiation therapy, and surgeries performed for the treatment of gynecological malignancies can contribute to this problem in various ways. 40 41 42 43 Sexual dysfunction is the main cause of suffering for the patient after gynecological cancer treatment, and ∼ 70% of survivors experience these problems, such as vaginal dryness, pain during intercourse, and vaginal stenosis. 14 15 16 17 18 44 45 These adverse effects can pose great challenges for survivors in the process of readaptation after the treatment phase. 46 47 48 49 50 51 Pain during intercourse was a predominant symptom in 10 of the selected studies and causes a great impact on the lives of women, as many avoid sex and abandon their sex life due to fear of the pain they may experience during intercourse. 9 15 18 46 Vaginal dryness was prevalent in several studies, and it has been identified through research that 70% of sexually active women experienced this side effect. 47 50 51 Currently, with the help of a variety of symptom management offerings, such as vaginal lubricators and targeted therapies, this discomfort can be alleviated for patients. 30 37 49 Studies have reported difficulties for women in adjusting to their altered sex lives and difficulties with intimacy with their partners. 52 53 As a consequence of these problems, many patients experience abandonment by their husband or partner, who does not understand the new context of female sexuality after treatment. 18 23 26 47 49 51 52 53 54 Menopausal symptoms are another side effect associated with female tract cancer, including vulvar, cervical and ovarian cancer. 41 42 Especially in younger women, this causes a devastating effect that interferes with their body image and psychology. 26 36 Sexual distress, although not a physical problem like vaginal symptoms, interferes directly with the psychological sphere (for example, anxiety or depression). 18 36 50 Many women have a poorer quality of life and unfortunately do not look for psychological intervention, which occasionally increases distress rates. 55 56 57 58 A summary of side effects associated with gynecologic cancer patients who had their sexuality affected after treatment is presented in Table 1 . Table 1 Main sexuality complaints of women survivors after treatment Sexuality Survival phase After treatment Long Mixed* Low libido (13, 23, 26, 37, 45, 51, 58) (27, 56) (10, 15, 17, 18, 30, 47, 48, 54, 55) Vaginal dryness (17, 23, 26, 37, 43, 50) (46, 49, 51) (18, 48, 47, 54, 52) Disinterest in sexual activity (15, 23, 26) (49) (14, 18, 41, 48, 50, 55, 59) Pain during intercourse (9, 15, 58) (51, 56) (14, 18, 41, 55, 59) Lower sexual satisfaction (17) (46, 56) (48, 47, 54, 55, 52) Vaginal stenosis / shortening (58) − (47, 54, 55, 52) Menopause symptoms (10, 26, 36) (51) (48) Sexual distress (18, 26, 36, 57, 58) (49, 51) (48, 54) Note: Mixed-phase comprehends studies that included participants from different treatment periods after cancer, i.e., it covers both studies just after the end of treatment as well as those that ended a few years ago. Care and Medical Support Strategies for Collateral Effects on Sexuality Six strategies were selected to manage adverse effects on sexuality after treatment: Patient-clinician communication Healthcare providers should understand the sexual, physical, and psychosocial needs of survivors because after treatment the psychological distress of cancer patients is already high, and many professionals overestimate this psychological suffering or fail to detect it, which can lead to detrimental results. 13 47 57 Open sexual communication is an opportunity for patients to seek information and emotional support to adapt positively to sexual changes, as they often have to face fear and stigma related to the disease, complications of treatment. It is an opportunity for patients to help adjust to a “new normal” even though they have ongoing and emerging health problems. 26 Developing strategies and practices for sexuality care The need to design and test known implementation strategies to integrate and disseminate scientific evidence-based interventions into practice for the benefit of cancer survivorship research. And the development of interventions such as implanting program in a gynecological oncology outpatient clinic formed by health professionals to receive information about sexual health care, the changes that may occur in their bodies and mainly an individual approach to realize the needs of each of the patients. 41 44 Individualized care plan A care plan should be made based on an assessment of the needs of the patients and reach those who need rehabilitation to obtain long-term quality of life of their sexual function. 10 27 37 58 59 This individualized approach would also allow for greater flexibility and effective sexual health assessment with the possibility of discussing symptom management and strategies to improve sexual function and satisfaction. 41 Multiprofessional team support Multiprofessional team support could help the sexual functioning of gynecological cancer survivors in the long term, as it is composed of physicians, psychologists, social workers, and physiotherapists. 43 Many women suffer with sequelae from treatment, such as vaginal discomfort, and physiotherapy treatment would help with pelvic floor exercises that could decrease this problem. The patients would also benefit from the help of psychologists accompanying them in this difficult period of changes, especially bodily and mental ones. 14 23 50 51 Development of rehabilitation programs The development of rehabilitation programs for survivors of gynecological cancer is very important because they can help analyzing the discomforts of sexual complaints and improve health as a whole. 9 Many female cancer survivors receive part of the medical treatment; however, frequently, psychosocial issues are not properly addressed. This is problematic because psychosocial needs impact the lives of patients in various ways, both personal, social, psychological, and financial. However, existing survival care is often targeted with a focus on the disease, not covering other aspects of the life of the patient, such as the social stigma on cancer, which is often seen by society as a death sentence and with little chance of the patient having quality of life. However, there are experiences in studies that confirm that, with the implantation of intervention groups formed by professionals such as physiotherapists, sexual counselors, and psychologists helped patients in their sexual difficulties and resulted in improvements in their psychological and physical aspects. 23 Support networks The creation of support networks aims to understand the emotional concerns and physical changes that interfere with the sexual identity of female cancer survivors and to offer the most appropriate support according to their specific needs. 42 Survivors have been living with or beyond cancer for a longer time, and there is a growing need to focus on health strategies that improve care and support to ensure a healthy and active life for as long as possible. 50 Because many women do not have their needs met, such as psychosocial and psychosexual needs, support networks corroborate so that they are welcomed, listened to, and supported to maintain their recovery and better manage the consequences of treatment. 9 40 44 Chart 2 demonstrates the supportive care strategies that can be offered to female cancer survivors post gynecological cancer treatment for sexuality complaints. Chart 2 Matrix of healthcare strategies for post-treatment sexuality complaints Study Care Strategies Patient-clinician communication Development of rehabilitation programs Multiprofessional team support Individualized care plan Development of strategies and practices for sexuality care Creating a support network for survivors Carter et al. (2010) 46 ✓ Levin et al. (2010) 47 ✓ Walton et al. (2010) 36 ✓ ✓ Beesley et al.(2013) 54 ✓ ✓ Grover et al. (2012) 37 ✓ Stavraka et al. (2012) 52 ✓ Le Borgne et al. (2013) 55 ✓ Afiyanti et al. (2013) 48 ✓ Sekse et al. (2013) 56 ✓ Loyd et al. (2014) 40 ✓ ✓ McCallum et al. (2014) 41 ✓ ✓ ✓ Vermeer et al. (2015) 18 ✓ ✓ Hopkins et al. (2015) 49 ✓ Lee et al. (2015) 13 ✓ Rowlands et al. (2015) 42 ✓ ✓ Westin et al. (2015) 58 ✓ Bakker et al. (2016) 44 ✓ ✓ Corrêa et al. (2016) 45 ✓ ✓ Mikkelsen et al. (2016) 10 ✓ Bakker et al. (2017) 9 ✓ McCallum et al. (2017) 16 Lutgendorf et al. (2017) 43 ✓ Teng et al. (2014) 57 ✓ ✓ Chow et al. (2018) 26 ✓ Mattsson et al. (2018) 50 ✓ ✓ Plotti et al. (2018) 27 ✓ Abbott-Anderson et al. (2020) 15 ✓ Fischer et al. (2019) 17 ✓ ✓ Hubbs et al. (2019) 30 ✓ Bacalhau et al. (2020) 53 Haryani et al. (2020) 51 ✓ Roberts et al. (2020) 23 ✓ Discussion The main objective of the present systematic review was to explore the main complaints in the field of sexuality with women survivors of cancer of the gynecological tract after treatment and to identify the care strategies that can be offered for these complaints. When doing oncological treatment of the female tract, many negative symptoms occur in women, and each type of tumor causes an effect, such as vulvar cancer. Although surgery and treatment have become less radical over the decades, it can still cause scarring and mutilation of the external genitalia, such as local excision to radical vulvectomy and removal of the clitoris, which can affect various nerves and blood vessels involved in important sexual, anal, and/or urinary functions. 26 31 Thus, the main complaints of affected women resulting from this treatment is the high risk of psychological distress, low libido, and dissatisfaction in the relationship with their partner. 60 61 And because vulvar cancer is a neoplasm with a rare condition, there is scarcity of studies on the impact of the disease and little is known about the real emotional, social, and psychological impacts on surviving patients. 2 4 8 Studies suggest that, for these cases, the care strategy that can be offered to surviving patients is psychological and therapeutic support, because many of them have significant levels of suffering, with an altered perception of their body image and feelings of isolation and embarrassment resulting from external genital mutilation. 11 12 13 Patients who underwent treatment for cervical cancer may also have a worsening quality of life and complaints in the field of sexuality, because certain treatments, such as radiotherapy, induce long-term toxicity with decreased urinary and gastrointestinal function; as a consequence, the main complaints in sexuality is vaginal shortening and atrophy, leading to dyspareunia and loss of sexual desire. 13 As a care strategy to prevent the atrophy process that affects sexual function, early initiation of local estrogen therapy is recommended; however, unfortunately, there is no definitive therapeutic option that fully preserves the quality of life of cervical cancer patients, because each treatment induces more or less low self-esteem and variations in reactions to the treatment. 26 31 In the biopsychosocial aspect, a serious problem to be faced arising from the treatment performed for cervical cancer is stigma and discrimination; as found in a study conducted in Brazil, the respondents reported that they suffered prejudice and were considered by the society in which they lived as dirty, lazy, and promiscuous, and due to this condition, many of them had their social life impaired. 62 The studies reviewed highlighted that ovarian cancer is usually diagnosed late and that the disease is detected in advanced stages; however, thanks to recent advances in treatment options, the number of surviving women is increasing. 29 42 However, patients who underwent chemotherapy, radiation, or surgery to remove the ovaries experience as a complaint in sexuality the psychosexual morbidity (physical and psychological sexual problems), vaginal atrophy due to the drop in estrogen levels, and reduction of libido as a consequence of the reduction in testosterone and androstenedione levels. 35 43 Throughout the trajectory of the patient, cancer-related symptoms and side effects of the treatment cause significant psychosocial morbidity that directly reflects on their lives. 23 The psychosocial suffering related to sexual dysfunction and visual effects such as scars and hair loss may directly affect the perception of the own self-image of the woman, therefore, referral to specialists is necessary to treat this multiplicity of chronic problems that occur after treatment. 9 21 58 In the literature, it is widely reported that regular and prolonged use of vaginal dilators or vibrators can prevent and delay the development of vaginal adhesions and stenosis; the use of water-based lubricants and estrogen application are also recommended to promote vaginal regeneration and decrease sexual complaints; however, there are cases of women whose sexual health problems have become chronic, needing additional support. 35 In this sense, the strategy of doctor-patient communication regarding sexual health is essential and can help patients face these challenges. 23 However, this care strategy can be challenging in the oncologic scenario, since despite its importance, there are obstacles in this bond, and on several occasions, health professionals find it difficult to approach the sexuality and intimacy of the patient due to lack of time and training and to the fear that the patient will be embarrassed to talk about her personal life. 12 45 An intimate relationship is a psychosocial process strongly experienced in the female sexual function and many difficulties faced by women survivors are a result of this. 19 34 The partner perceives the changes that occur after treatment, whether physical or emotional, and in some cases the roles may change, with the partner being the main caregiver of the patient. 15 19 In the biopsychosocial aspect, this has a very relevant factor, because many of them feel unprepared and without the correct information of care, which brings a negative psychological, social, and even financial burden, because in many situations it is necessary to give up the job to be able to give the support that the patient needs. 11 13 60 Intimate intercourse is a psychosocial process strongly experienced in the female sexual function and many difficulties faced by surviving women are a result of this. 19 34 In addition, sexual dysfunctions are closely linked to the stages of the cycle of sexual response, that is, the inability of the sexual act to be satisfactory for the couple. And with women who have undergone cancer treatment, this is more evident because many lose their will and libido, not only because of their physical appearance, but also because they are very psychologically shaken, which further distances them from their partners. The limitation of the present research lies in the design of the strategies used and are not considered to be other adverse symptoms that survivors face in the context of sexuality after gynecological treatment. Conclusion The main complaints identified in the scientific literature regarding the post-treatment of cancer survivors were low libido and lack of interest in sexual activity. The care strategy that could be adopted would involve initial consultations with medical personnel and a thorough review of complaints. and symptoms reported patient's hair to find cause, which may be physical or psychological. Another complaint often referred to in the selected studies is vaginal dryness and, as a treatment strategy, the use of vagina moisturizers could improve humidification and thus reduce this discomfort in women. Pain during intercourse and stenosis or shortening of the vagina were also identified as common symptoms in women after cancer treatment; therefore, adopting healthcare strategies and actions can reduce pain and better deal with these complaints, such as mobilizing rehabilitation focused on sexual health with the participation of specialist professionals such as physiotherapists to analyze the case of each patient to help her in the best possible way. And, finally, sexual anxiety is another common complaint cited in the academic literature that causes severe mental and psychological problems for patients. An approach to care that could be used would be open communication and sincere dialogue between the physician and the patient with an integrative therapeutic approach to identify the major problems affecting them to provide an individualized care plan with the support of a multidisciplinary health team. It is suggested that future research should conduct qualitative studies to explore strategies that are acceptable in the perceptions of patients of suggesting improvements. Supplementary Material Supplementary Material Supplementary Material Conflict of Interests The authors have no conflict of interests to declare. ==== Refs References 1 Ashmore L A Stewart H Hutton D Evans K Digital support for living with and beyond gynaecological cancer Radiography 2020 26 04 e270 e276 10.1016/j.radi.2020.03.014 32439292 2 Götze H Friedrich M Taubenheim S Dietz A Lordick F Mehnert A Depression and anxiety in long-term survivors 5 and 10 years after cancer diagnosis Support Care Cancer 2020 28 01 211 220 10.1007/s00520-019-04805-1 31001695 3 Sung Uk L Young Ae K Young-Ho Y Y-J K Cheol L M S-Y P General health status of long-term cervical cancer survivors after radiotherapy Strahlenther Onkol 2017 193 07 543 551 10.1007/s00066-017-1143-8 28492995 4 Kim J Y Chun M Kim S W Song J Identification of Korean cancer survivors' unmet needs and desired psychosocial assistance: A focus group study PLoS One 2020 15 01 e0228054 10.1371/journal.pone.0228054 31945143 5 Shin J Shin D W Lee J Hwang J H Yoo J E Ko H Association between perception of care coordination and health outcomes in Korean cancer survivors Health Qual Life Outcomes 2020 18 01 21 10.1186/s12955-020-1279-6 32019567 6 Mamguem Kamga A Dumas A Joly F Billa O Simon J Poillot M-L Long-term gynecological cancer survivors in Côte d'Or: health-related quality of life and living conditions Oncologist 2019 24 07 e490 e500 10.1634/theoncologist.2018-0347 30578310 7 Harris M G Sexuality and menopause: unique issues in gynecologic cancer Semin Oncol Nurs 2019 35 02 211 216 10.1016/j.soncn.2019.02.008 30871843 8 Maguire R Kotronoulas G Simpson M Paterson C A systematic review of the supportive care needs of women living with and beyond cervical cancer Gynecol Oncol 2015 136 03 478 490 10.1016/j.ygyno.2014.10.030 25462200 9 Bakker R M Kenter G G Creutzberg C L Stiggelbout A M Derks M Mingelen W Sexual distress and associated factors among cervical cancer survivors: A cross-sectional multicenter observational study Psychooncology 2017 26 10 1470 1477 10.1002/pon.4317 27862635 10 Mikkelsen T B Sørensen B Dieperink K B Prediction of rehabilitation needs after treatment of cervical cancer: what do late adverse effects tell us? Support Care Cancer 2017 25 03 823 831 10.1007/s00520-016-3466-x 27834003 11 Duimering A Walker L M Turner J Andrews-Lepine E Driga A Ayume A Quality improvement in sexual health care for oncology patients: a Canadian multidisciplinary clinic experience Support Care Cancer 2020 28 05 2195 2203 10.1007/s00520-019-05040-4 31428930 12 Frimer M Turker L B Shankar V Cardaci R Van Arsdale A R The association of sexual dysfunction with race in women with gynecologic malignancies Gynecol Oncol Rep 2019 30 100495 10.1016/j.gore.2019.100495 31656849 13 Lee J T Lin H H Tsai J L Chen C P Huang K G Lien A S Transformation of sexual expression in taiwanese women after treatment for gynecological cancer Cancer Nurs 2015 38 06 475 483 10.1097/NCC.0000000000000224 25629893 14 Harter P Schrof I Karl L M Kullmann V Traut A Scheller H Sexual function, sexual activity and quality of life in women with ovarian and endometrial cancer Geburtshilfe Frauenheilkd 2013 73 05 428 432 10.1055/s-0032-1328602 24771922 15 Abbott-Anderson K Young P K Eggenberger S K Adjusting to sex and intimacy: Gynecological cancer survivors share about their partner relationships J Women Aging 2020 32 03 329 348 10.1080/08952841.2019.1591888 30905277 16 McCallum M Jolicoeur L Lefebvre M Babchishin L K Le T Lebel S Filling in the gaps: Sociodemographic and medical predictors of sexual health and other supportive care needs, and desire for help in gynecological cancer survivors Can Oncol Nurs J 2017 27 03 251 258 10.5737/23688076273251258 31148785 17 Fischer O J Marguerie M Brotto L A Sexual function, quality of life, and experiences of women with ovarian cancer: a mixed-methods study Sex Med 2019 7 04 530 539 10.1016/j.esxm.2019.07.005 31501030 18 Vermeer W M Bakker R M Kenter G G de Kroon C D Stiggelbout A M ter Kuile M M Sexual issues among cervical cancer survivors: how can we help women seek help? Psychooncology 2015 24 04 458 464 10.1002/pon.3663 25858440 19 Afiyanti Y Setyowati S Milanti A Young A 'Finally, I get to a climax': the experiences of sexual relationships after a psychosexual intervention for Indonesian cervical cancer survivors and the husbands J Psychosoc Oncol 2020 38 03 293 309 10.1080/07347332.2020.1720052 32093583 20 Stabile C Goldfarb S Baser R E Goldfrank D J Abu-Rustum N R Barakat R R Sexual health needs and educational intervention preferences for women with cancer Breast Cancer Res Treat 2017 165 01 77 84 10.1007/s10549-017-4305-6 28547655 21 Chapman C H Heath G Fairchild P Berger M B Wittman D Uppal S Gynecologic radiation oncology patients report unmet needs regarding sexual health communication with providers J Cancer Res Clin Oncol 2019 145 02 495 502 10.1007/s00432-018-2813-3 30539283 22 Schofield P Juraskova I Bergin R Gough K Mikeshkin L Krishnasamy M A nurse- and peer-led support program to assist women in gynaecological oncology receiving curative radiotherapy, the PeNTAGOn study (peer and nurse support trial to assist women in gynaecological oncology): study protocol for a randomised controlled trial Trials 2013 14 39 10.1186/1745-6215-14-39 23399476 23 Roberts K Chong T Hollands E Tan J Mohan G RKA Cohen P A Screening for sexual health concerns in survivors of gynecological cancer Support Care Cancer 2020 28 02 599 605 10.1007/s00520-019-04872-4 31104133 24 Buzaglo J S Zaleta A K McManus S Golant M Miller M F CancerSupportSource®: validation of a revised multi-dimensional distress screening program for cancer patients and survivors Support Care Cancer 2020 28 01 55 64 10.1007/s00520-019-04753-w 30980258 25 Johnston S Hogel M Burchell A N Rebick G Antoninou T McLaren M Developing a performance framework for measuring comprehensive, community-based primary healthcare for people with HIV Prim Health Care Res Dev 2016 17 04 361 384 10.1017/s1463423615000572 26627002 26 Chow K M So W KW Choi K C Chan C WH Sexual function, psychosocial adjustment to illness, and quality of life among Chinese gynaecological cancer survivors Psychooncology 2018 27 04 1257 1263 10.1002/pon.4663 29430767 27 Plotti F Terranova C Capriglione S Crispino S Pomi AL Nardone C C Assessment of quality of life and urinary and sexual function after radical hysterectomy in long-term cervical cancer survivors Int J Gynecol Cancer 2018 28 04 818 823 10.1097/igc.0000000000001239 29538249 28 Millman R Jacox N Sears C Robinson J W Turner J Walker L M Patient interest in the Lowdown on Down There: attendance at a vulvovaginal and sexual health workshop post-cancer treatment Support Care Cancer 2020 28 08 3889 3896 10.1007/s00520-019-05162-9 31845008 29 Lokich E Gynecologic cancer survivorship Obstet Gynecol Clin North Am 2019 46 01 165 178 10.1016/j.ogc.2018.10.002 30683262 30 Hubbs J L Dickson Michelson E L Vogel R I Rivard C L Teoh D GK Geller M A Sexual quality of life after the treatment of gynecologic cancer: what women want Support Care Cancer 2019 27 12 4649 4654 10.1007/s00520-019-04756-7 30941579 31 Tsai L Y Tsai J M Tsay S L Life experiences and disease trajectories in women coexisting with ovarian cancer Taiwan J Obstet Gynecol 2020 59 01 115 119 10.1016/j.tjog.2019.11.032 32039777 32 Sekse R JT Dunberger G Olesen M L Østerbye M Seibaek L Lived experiences and quality of life after gynaecological cancer-An integrative review J Clin Nurs 2019 28 (9-10):1393 1421 10.1111/jocn.14721 30461101 33 Beesley V L Alemayehu C Webb P M A systematic literature review of trials of survivorship interventions for women with gynaecological cancer and their caregivers Eur J Cancer Care (Engl) 2019 28 03 e13057 10.1111/ecc.13057 31020737 34 Micaux Obol C Lampic C Wettergren L Ljungman L Eriksson L E Experiences of a web-based psycho-educational intervention targeting sexual dysfunction and fertility distress in young adults with cancer-A self-determination theory perspective PLoS One 2020 15 07 e0236180 10.1371/journal.pone.0236180 32697801 35 Liberacka-Dwojak M Izdebski P Sexual function and the role of sexual communication in women diagnosed with cervical cancer: a systematic review Int J Sex Health 2021 33 03 385 395 10.1080/19317611.2021.1919951 36 Walton L M Reeve J Brown P M Farquhar C M Gynaecologic cancer patients' needs and experiences of supportive health services in New Zealand Psychooncology 2010 19 02 201 208 10.1002/pon.1553 19253918 37 Grover S Hill-Kayser C E Vachani C Hampshire M K DiLullo G A Metz J M Patient reported late effects of gynecological cancer treatment Gynecol Oncol 2012 124 03 399 403 10.1016/j.ygyno.2011.11.034 22119992 38 Page M J McKenzie J E Bossuyt P M Boutron I Hoffmann T C Mulrow C D The PRISMA 2020 statement: an updated guideline for reporting systematic reviews Syst Rev 2021 10 01 89 10.1186/s13643-021-01626-4 33781348 39 Pace R Pluye P Bartlett G Macaulay A C Salsberg J Jagosh J Testing the reliability and efficiency of the pilot Mixed Methods Appraisal Tool (MMAT) for systematic mixed studies review Int J Nurs Stud 2012 49 01 47 53 10.1016/j.ijnurstu.2011.07.002 21835406 40 Lloyd P A Briggs E V Kane N Jeyarajah A R Shepherd J H Women's experiences after a radical vaginal trachelectomy for early stage cervical cancer. A descriptive phenomenological study Eur J Oncol Nurs 2014 18 04 362 371 10.1016/j.ejon.2014.03.014 24794078 41 McCallum M Jolicoeur L Lefebvre M Babschishin L K Supportive care needs after gynecologic cancer: where does sexual health fit in? Oncol Nurs Forum 2014 41 03 297 306 10.1188/14.ONF.297-306 24769594 42 Rowlands I J Lee C Beesley V L Janda M Nagle C M Webb P M Women's perceptions of their lifestyle and quality of life several years after a diagnosis of endometrial cancer Cancer Nurs 2015 38 01 E21 E28 10.1097/NCC.0000000000000129 43 Lutgendorf S K Shinn E Carter J Leighton S Baggerly K Guindani M Quality of life among long-term survivors of advanced stage ovarian cancer: A cross-sectional approach Gynecol Oncol 2017 146 01 101 108 10.1016/j.ygyno.2017.05.008 28527672 44 Bakker R M Mens J W de Groot H E Tuijnman-Raasveld C C Braat C Hompus W CP A nurse-led sexual rehabilitation intervention after radiotherapy for gynecological cancer Support Care Cancer 2017 25 03 729 737 10.1007/s00520-016-3453-2 27787681 45 Corrêa C S Leite I C Andrade A P de Souza Sérgio Ferreira A Carvalho S M Guerra M R Sexual function of women surviving cervical cancer Arch Gynecol Obstet 2016 293 05 1053 1063 10.1007/s00404-015-3857-0 26335186 46 Carter J Chi D S Brown C L Abu-Rustum N R Sonoda Y Aghajanian C Levine D A Baser R E Cancer-related infertility in survivorship Int J Gynecol Cancer 2010 20 01 2 8 10.1111/IGC.0b013e3181bf7d3f 20130497 47 Levin A O Carpenter K M Fowler J M Brothers B M Andersen B L Maxwell G L Sexual morbidity associated with poorer psychological adjustment among gynecological cancer survivors Int J Gynecol Cancer 2010 20 03 461 470 10.1111/IGC.0b013e3181d24ce0 20375814 48 Afiyanti Y Milanti A Physical sexual and intimate relationship concerns among Indonesian cervical cancer survivors: A phenomenological study Nurs Health Sci 2013 15 02 151 156 10.1111/nhs.12006 23171291 49 Hopkins T G Stavraka C Gabra H Fallowfield L Hood C Blagden S Sexual activity and functioning in ovarian cancer survivors: an internet-based evaluation Climacteric 2015 18 01 94 98 10.3109/13697137.2014.929104 25029389 50 Mattsson E Einhorn K Ljungman L Sundström-Poromaa I Stålberg K Wikman A Women treated for gynaecological cancer during young adulthood - A mixed-methods study of perceived psychological distress and experiences of support from health care following end-of-treatment Gynecol Oncol 2018 149 03 464 469 10.1016/j.ygyno.2018.03.055 29588102 51 Haryani H Afiyanti Y Besral B Gayatri D Wahidi K R Pramiasti H Cross-cultural adaptation and validation of the cancer survivors' unmet needs measure among gynecological cancer patients in Indonesia Arch Oncol 2020 26 02 23 29 10.2298/AOO190531011H 52 Stavraka C Ford A Ghaem-Maghami S Crook T Agarwal R Gabra H A study of symptoms described by ovarian cancer survivors Gynecol Oncol 2012 125 01 59 64 10.1016/j.ygyno.2011.12.421 22155797 53 do Rosário Ramos Nunes Bacalhau M Pedras S da Graça Pereira Alves M Attachment style and body image as mediators between marital adjustment and sexual satisfaction in women with cervical cancer Support Care Cancer 2020 28 12 5813 5819 10.1007/s00520-020-05423-y 32236767 54 Australian Ovarian Cancer Study Group Australian Ovarian Cancer Study-Quality of Life Study Investigators Beesley V L Price M A Webb P M O'Rourke P Marquart L Butow P N Changes in supportive care needs after first-line treatment for ovarian cancer: identifying care priorities and risk factors for future unmet needs Psychooncology 2013 22 07 1565 1571 10.1002/pon.3169 22936668 55 Le Borgne G Mercier M Woronoff A S Guizard A-V Abeilard E Caravati-Jouvenceau A Klein D Quality of life in long-term cervical cancer survivors: a population-based study Gynecol Oncol 2013 129 01 222 228 10.1016/j.ygyno.2012.12.033 23280088 56 Sekse R J Gjengedal E Råheim M Living in a changed female body after gynecological cancer Health Care Women Int 2013 34 01 14 33 10.1080/07399332.2011.645965 23216094 57 Teng F F Kalloger S E Brotto L McAlpine J N Determinants of quality of life in ovarian cancer survivors: a pilot study J Obstet Gynaecol Can 2014 36 08 708 715 10.1016/S1701-2163(15)30513-2 25222166 58 Westin S N Sun C C Tung C S Lacour R A Meyer L A Urbauer D L Survivors of gynecologic malignancies: impact of treatment on health and well-being J Cancer Surviv 2016 10 02 261 270 10.1007/s11764-015-0472-9 26245979 59 Vermeer W M Bakker R M Kenter G G Stiggelbout A M Ter Kuile M M Cervical cancer survivors' and partners' experiences with sexual dysfunction and psychosexual support Support Care Cancer 2016 24 04 1679 1687 10.1007/s00520-015-2925-0 26412245 60 Malandrone F Bevilacqua F Merola M Gallio N Ostacoli L Carletto S The impact of vulvar cancer on psychosocial and sexual functioning: a literature review Cancers (Basel) 2021 14 01 63 10.3390/cancers14010063 35008225 61 Aerts L Enzlin P Verhaeghe J Vergote I Amant F Psychologic, relational, and sexual functioning in women after surgical treatment of vulvar malignancy: a prospective controlled study Int J Gynecol Cancer 2014 24 02 372 380 10.1097/IGC.0000000000000035 24407574 62 Gregg J L An unanticipated source of hope: stigma and cervical cancer in Brazil Med Anthropol Q 2011 25 01 70 84 10.1111/j.1548-1387.2010.01137.x 21495495
PMC009xxxxxx/PMC9709142.txt
==== Front Front Mol Biosci Front Mol Biosci Front. Mol. Biosci. Frontiers in Molecular Biosciences 2296-889X Frontiers Media S.A. 36465563 1006917 10.3389/fmolb.2022.1006917 Molecular Biosciences Original Research No safe renal warm ischemia time—The molecular network characteristics and pathological features of mild to severe ischemia reperfusion kidney injury Chen et al. 10.3389/fmolb.2022.1006917 Chen Ya-Lei 1 † Li Huai-Kang 2 † Wang Lei 2 † Chen Jian-Wen 2 3 * Ma Xin 2 * 1 Department of Critical Care Medicine, Capital Medical University Electric Power Teaching Hospital/State Grid Beijing Electric Power Hospital, Beijing, China 2 Senior Department of Urology, The Third Medical Centre of PLA General Hospital, Beijing, China 3 Department of Nephrology, State Key Laboratory of Kidney Diseases, Beijing Key Laboratory of Kidney Disease Research, First Medical Center of Chinese PLA General Hospital, Nephrology Institute of the Chinese People’s Liberation Army, National Clinical Research Center for Kidney Diseases, Beijing, China Edited by: Xin Zhang, Jiangmen Central Hospital, China Reviewed by: Daniele Canale, University of São Paulo, Brazil Susann Patschan, Brandenburg Medical School Theodor Fontane, Germany Long Zheng, The second Affiliated Hospital of Zhejiang University School of Medicine, China *Correspondence: Xin Ma, [email protected]; Jian-Wen Chen, [email protected] † Those authors contributed equally to this work This article was submitted to Molecular Diagnostics and Therapeutics, a section of the journal Frontiers in Molecular Biosciences 16 11 2022 2022 9 100691729 7 2022 03 11 2022 Copyright © 2022 Chen, Li, Wang, Chen and Ma. 2022 Chen, Li, Wang, Chen and Ma https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Ischemic acute kidney injury (AKI) has always been a hot and difficult research topic in the field of renal diseases. This study aims to illustrate the safe warm ischemia time of kidney and the molecular network characteristics and pathological features of mild to severe ischemia reperfusion kidney injury. We established varying degrees of renal injury due to different ischemia time (0 min, 16 min, 18 min, 20 min, 22 min, 24 min, 26 min, 28 min, and 30 min) on unilateral (left kidney) ischemia-reperfusion injury and contralateral (right kidney) resection (uIRIx) mouse model. Mice were sacrificed 24 h after uIRIx, blood samples were harvested to detect serum creatinine (Scr), and kidney tissue samples were harvested to perform Periodic Acid-Schiff (PAS) staining and RNA-Seq. Differentially expressed genes (DEGs) were identificated, time-dependent gene expression patterns and functional enrichment analysis were further performed. Finally, qPCR was performed to validated RNA-Seq results. Our results indicated that there was no absolute safe renal warm ischemia time, and every minute of ischemia increases kidney damage. Warm ischemia 26min or above in mice makes severe kidney injury, renal pathology and SCr were both significantly changed. Warm ischemia between 18 and 26 min makes mild kidney injury, with changes in pathology and renal molecular expression, while SCr did not change. No obvious pathological changes but significant differences in molecular expression were found less than 16min warm ischemia. There are two key time intervals in the process of renal ischemia injury, 0 min–16 min (short-term) and 26 min–28 min (long-term). Gene expression of immune-related pathways were most significantly down-regulated in short-term ischemia, while metabolism-related pathways were the mainly enriched pathway in long-term ischemia. Taken together, this study provides novel insights into safe renal artery occlusion time in partial nephrectomy, and is of great value for elucidating molecular network characteristics and pathological features of mild to severe ischemia reperfusion kidney injury, and key genes related to metabolism and immune found in this study also provide potential diagnostic and therapeutic biomarkers for AKI. acute kidney injury pathological features ischemia reperfusion kidney injury molecular network characteristics mild to severe ischemia reperfusion injury This work was supported by grants from the National Natural Science Foundation of China (Grant No. 82100713, 82000631, 81970665, 81970594); China Postdoctoral Science Foundation (Grant No. 2021T140791); Fund of Chinese PLA 13th Five Year Plan for Medical Sciences (No. BLB19J009); National Sciences Foundation of Beijing (Grant No. 7222169); Military Medical Youth Special Project of PLA General Hospital (Grant No. QNF19035); and Young Elite Scientist Sponsorship Program by CAST (No. YESS20200400). ==== Body pmc1 Introduction Acute kidney injury (AKI) is a multidisciplinary disease, at present, the basic research work of AKI is mainly focused on the fields of nephrology, intensive care unit and kidney transplantation (Bellomo et al., 2012; Zhang et al., 2020). In these fields, bilateral or isolated renal ischemia is commonly faced, and the incidence of AKI is very high. For example, the incidence of hospital-acquired AKI is approximately up to 20% in the intensive care unit (Gonsalez et al., 2019). Due to the limited treatment of AKI, there are still a high proportion of patients with AKI who progress to chronic kidney disease (CKD) and end-stage renal disease, which need kidney dialysis or waiting for kidney transplantation (Chen et al., 2020; Uber and Sutherland, 2020). Despite extensive experimental and clinical investigations, the underlying pathogenesis and progression mechanisms of AKI have not been fully clarified. Ischemia-reperfusion is one of the main cause of AKI (Melo Ferreira et al., 2021). In urology, the renal artery needs to be clipped during nephron sparing surgery or partial nephrectomy (Weight et al., 2010; Soranno et al., 2019). Transient occlusion of the renal artery can control surgical blood loss, facilitate tumor resection, and wound suture. However, the occlusion of the renal artery also causes transient renal ischemia and potential renal injury (Thompson et al., 2010). At present, 25 min of renal warm ischemia is generally recognized as the threshold for AKI (Funahashi et al., 2009). Other studies have shown that warm ischemia time of laparoscopic partial nephrectomy over 20 min is correlated with poor preoperative and postoperative glomerular filtration rate, and the incidence of postoperative complications increases 2.54 times for each extension of warm ischemia time of laparoscopic partial nephrectomy for 9 min (Lane et al., 2008). Based on the analysis of 362 kidney patients undergoing partial nephrectomy, Thompson et al. of the Department of Urology at Mayo Clinic found that a longer warm ischemia time was associated with both short-term and long-term renal complications, suggested that every minute of renal artery occlusion during surgery was important (Thompson et al., 2010). These academic debates indicate that the safe time of warm ischemia is still controversial, and the specific relationship between renal ischemia time and AKI requires a larger clinical sample size and clear basic research evidence. The experimental mouse models of renal ischemia/reperfusion injury (IRI) have been widely applied to study the pathogenesis and injury outcome of ischemic AKI (Chen et al., 2020). IRI mouse model can well simulate renal injury caused by renal artery occlusion in human partial nephrectomy. In this study, we established a relatively stable IRI model with unilateral ischemia (left kidney) and contralateral (right kidney) resection (uIRIx) to explore the changes of injury, molecular network characteristics and pathological features of mouse kidney with different degrees of ischemia. We systematically examined the renal pathology at different ischemia times, and identified the molecular characteristics of gene expression associated with different ischemia times by performing RNA sequencing (RNAseq). Our study provides basic research evidence for the effect of ischemic time on renal damage in partial nephrectomy of urology, and offers resources for clinical and research communities, and brings basis and new ideas for AKI research. 2 Materials and methods 2.1 Animal model construction and detection Wild-type C57BL/6 mice (18–22 g) were purchased from the Animal Center of Chinese PLA General Hospital. These mice were housed in a specific pathogen-free facility under a 12 h light/12 h dark cycle with free access to food and water. Mice were randomly assigned to several groups based on different ischemia time: 0 min, 16 min, 18 min, 20 min, 22 min, 24 min, 26 min, 28 min, and 30 min. To induce unilateral ischemia reperfusion injury and the contralateral kidney resection AKI (uIRIx), mice were placed on 37°C warm plate (Yuyan instruments, Beijing, China) to keep the temperature of kidney all through the surgery; the right kidney was excised, and left kidney pedicle was clipped for indicated time using microaneurysm clamp. Mice were then sacrificed 24 h after uIRIx, blood and kidney tissue samples were harvested for further processing. The animal protocol and all animal procedures were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) of the Chinese PLA General Hospital, and the permit number is 2019-X15-97. 2.2 Blood and histopathological assessment of kidney injury Blood samples were collected from the vena cava at the indicated time points, and the serum was separated by centrifugation at 3,000 rpm for 15 min at 4°C and then sent to the PLA General Hospital Biochemistry Department to detect serum creatinine (SCr). A quarter of the left kidney was fixed in 4% formaldehyde, dehydrated, and embedded in paraffin. Histopathological examination was performed by Periodic Acid-Schiff (PAS) staining of kidney tissue section (2 μm) to evaluate tubular injury. 2.3 High-throughput RNA sequencing (RNA-Seq) and preprocessing In order to explore the molecular characteristics of kidney injury at different ischemic times, the left ischemia kidney in each mouse was collected and the RNA was extracted to perform RNA-Seq (Beijing Genomics Institute, Beijing, China). All raw RNA-Seq data were uploaded to the GEO database GSE192883 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE192883). The preprocessing of the gene expression profile data, which includes the background correction, the quantile normalization, the median polish summarization, and the log2 transformation, was performed by R software (https://www.r-project.org/) and RStudio software (https://www.rstudio.com/) as we described before (Chen et al., 2020). Principal component analysis (PCA) was performed by scatterplot3d package and princomp function in R software. 2.4 Identification of differentially expressed genes The linear model for high-throughput data analysis (limma) in Bioconductor (http://www.bioconductor.org/) was applied to find DEGs by comparing expression value between different ischemia groups. Differential expression was calculated using an empirical Bayes model. The criteria for the statistically significant difference of DEGs was | log2 fold change (FC)| > 1 in expression and adjusted p-value (false discovery rate, FDR) < 0.05. The overlapping DEGs were further analyzed by UpSetR package in R software. We obtained a union DEGs set, which included the combination of DEGs obtained in each ischemia group compared with the ischemia 0min and 16min groups. 2.5 Time-dependent gene expression patterns profiling We applied the fuzzy c-means algorithm in the Mfuzz R package (Kumar and Futschik, 2007) to perform noise-robust soft clustering analysis and visualize the time-dependent expression patterns of the genes in union DEGs set. Soft clustering is more accurate than hard clustering with more robustness to noise and less information loss, and Mfuzz R package is one of the software packages for soft clustering (Kumar and Futschik, 2007; Yang et al., 2021). The average RPKM value for each gene at each time point was used as the input. After standardization, genes in union DEGs set were assigned to 8 clusters, such as continuous increase, continuous decrease, increase first then decrease, and decrease first then increase etc., according to the changes in the gene expression with ischemia time. 2.6 Functional enrichment analysis Gene ontology (GO, http://www.geneontology.org) analysis has been widely used to annotate the gene function. The categories of GO include the cellular component (CC), the biological process (BP), and the molecular function (MF) terms (Chen et al., 2020). Kyoto Encyclopedia of Genes and Genomes (KEGG, http://www.genome.jp/kegg) is a database containing the information of genes related to metabolic and regulatory pathways (Yang et al., 2021). In the present study, the clusterProfiler R package (Yu et al., 2012) was used to conduct the GO and KEGG functional analysis for genes enriched in each cluster, respectively. The p-value of enriched pathways was calculated by a hyper geometric distribution test. Subsequently, the p-value was revised using BH (Benjamini and Hochberg) method, and the adjusted p-value < 0.05 was served as the cut-off criterion for selecting significant enrichment pathways. We also used clusterProfiler R package with the same threshold value, as mentioned above, to conduct the gene set enrichment analysis (GSEA) analyses for genes compared between 16 min and 0 min, between 28 min and 26 min, respectively. The criteria for the statistically significant difference of GSEA was |normalized enrichment score (NES)| >0.4 and adjusted p-value < 0.05. 2.7 PPI network construction and analysis of modules Considering that proteins rarely work alone, it is necessary to study the interactions among proteins. The Search Tool for the Retrieval of Interacting Genes/Proteins (STRING, http://string-db.org/) is an online biological resource database that is commonly used to identify the interactions between known and predicted proteins (Szklarczyk et al., 2015). By searching the STRING database, the PPI network of the genes in each cluster acquired by soft clustering above were selected with a score>0.4, and the PPI network was visualized by Cytoscape software (http://www.cytoscape.org/). In the PPI network, each node stands for a gene or a protein, and edges represent the interactions between the nodes. We then used the plug-in named Molecular Complex Detection (MCODE) by Cytoscape to filter the modules of the PPI network. The first module of each cluster was selected. 2.8 Gene expressions by quantitative real-time PCR The frozen tissue samples were lysed in TRIzol reagent (Invitrogen, Carlsbad, CA) and total RNA was extracted according to the manufacturer’s instructions. The levels of transcripts were determined by qPCR using TransStart® Top Green qPCR SuperMix (AQ131, Transgen, Beijing, China) on an Applied Biosystems 7,500 system PCR cycler (Applied Biosystems, Foster City, CA, United States). The data were normalized to 18s expression. Primers were obtained from Beijing Genomics Institute (Beijing, China). The list of primers was presented in Supplementary Table S1. 2.9 Statistical analysis The statistical analyses used are specified in the figure legends. 3 Results 3.1 Pathological features of kidney injury at different ischemia time pionts To explore the pathological characteristics of kidney injury at different ischemia times, we adopted the uIRIx model, which were more stable than bilateral ischemia reperfusion kidney injury AKI model. Mice were randomly assigned to 9 ischemia groups based on different ischemia times of left kidney: 0 min, 16 min, 18 min, 20 min, 22 min, 24 min, 26 min, 28 min, and 30 min (Figure 1A), ischemia 0min was the control group. As shown in Figure 1B, after reperfusion for 24 h, the serum creatinine of 28 min and 30 min groups were significantly higher than that of 0 min group (p < 0.05), while there was no statistical difference among 16 min, 18 min, 20 min, 22 min, 24 min, or 26 min with 0 min group. The serum creatinine dramatically changed between 26 min and 28 min of ischemia at the condition of 37°C. FIGURE 1 Pathological features of kidney injury at different ischemia times. (A) The experimental design of the unilateral ischemia reperfusion injury and the contralateral kidney resection AKI (uIRIx) model. Mice were randomly assigned to 9 ischemia groups based on different ischemia time of left kidney: 0 min, 16 min, 18 min, 20 min, 22 min, 24 min, 26 min, 28 min, and 30 min, and the mice were sacrificed 24 h after reperfusion. (B) The serum creatinine levels in different ischemia groups 24 h after reperfusion. Statistical analysis: firstly, we performed one-way ANOVA analysis on the 9 groups, and found statistically significant differences in mean values among the samples; And then, Dunnett-t test was used to make multiple comparisons of mean differences between each ischemia group and sham group. n = 6–9 mice per group, ns indicates no significant, ***p < 0.001, compared to 0min group (C–K) Representative micrographs of PAS staining showed the pathological features of kidney injury at different ischemia times: 0min group (C), 16 min group (D), 18 min group (E), 20 min group (F), 22 min group (G), 24 min group (H), 26 min group (I), 28 min group (J), and 30 min group (K). Scale bars: 100 μm. The arrow showed the brush border of normal tubule, the asterisk (*) showed the blocked tubule, the triangle showed the infiltrated inflammatory cell, and the well sign (#) showed the protein tubule. PAS staining of kidney (Figures 1C–K) showed that the renal cortex and medulla structures of mice in the 0min and 16min groups were normal, with no ablation of the brush edge of renal tubules, no dilation of renal tubules and no tubular formation. In ischemia 18min group, there was scattered cellular cast in the cortical and medulla area. In 20min, 22min, and 24min groups, the cellular casts in the cortex and medulla area increased gradually, and in 26min group, there was small patches of renal tubular obstruction. In ischemia 28min group, there was necrosis, tubule blockage formed by exfoliation of large patchy renal tubular epithelial cells, and inflammatory cell infiltration in the interstitium, as well as patchy protein tube type. In ischemia 30min group, almost all areas were filled with tubule cast, protein cast, and inflammatory cell infiltration. 3.2 Expression characteristics of injury-related molecules of kidney injury at different ischemia time points To further elucidate the expression of injury-related molecules at different renal ischemia time points, we performed immunofluorescence staining on KIM1 (Figure 2A). The results found that the expression of KIM1 began at 16 min of ischemia, and then increased with the increase of ischemia time. At the same time, we performed qPCR to detect the expression of Kim1 (Figure 2B), Lcn2 (Figure 2C), and Il6 (Figure 2D), and found that Kim1 was highly expressed at 16 min and continued to increase at 30 min after ischemia. The expression of Lcn2 (Ngal) began to increase in the 22 min ischemia group, and gradually increased with the prolongation of ischemia time. However, the expression of Il6 was significantly increased at 30 min of ischemia, with no significant in less than ischemia 28 min. FIGURE 2 Expression characteristics of injury-related molecules of kidney injury at different ischemia time points. (A) Representative immunofluorescence staining of KIM1 expression in renal tissues at different ischemia time points. Red stands for KIM1. Scale bars: 100 μm (B–D) The mRNA expression of Kim1 (B), Lcn2 (C), and Il6 (D) in renal tissues at different ischemia time points. One-way ANOVA and Dunnett-t test, n = 3 per group, ns indicates no significant, *p < 0.05, **p < 0.01, ***p < 0.001, compared to 0min group. 3.3 High-throughput RNA sequencing and identification of differentially expressed genes between groups In order to explore the molecular characteristics of kidney injury at above different ischemic times, triplicate left ischemia kidneys in every group were collected to perform RNA-Seq. Before the next analysis, we performed the quality control for every raw expression data to guarantee the quality. The normalized box diagram showed that the data was well preprocessed after normalization. Next, we applied PCA analysis to those 27 samples (different colors) to visualize sample clusters. The results showed that samples at the same time points were clustered (Figure 3), indicating good correlation between the biological repeats. Notably, samples of 16 min, 18 min, 20 min, 22 min, 24 min, and 26 min localized close to each another, suggesting that those time points have similarities in transcriptome. FIGURE 3 Principle component analysis (PCA) plot of samples from different ischemia times. Each point represents a sample. PCA plot showed the tight clustering of biological replicates and distinct clustering between groups. DEG analysis was performed in all samples using cutoff values of a fold change≥2 and an adjusted p-value < 0.05. With the increase of ischemia time from 16 min to 26 min, the number of DEGs was gradually but slowly increased, from 863 to 1,448. While in the group of 28 min vs 0 min, the number of DEGs suddenly increased at 2,782. The most pronounced DEGs were in the group of 30 min vs 0 min, with 1,698 upregulated and 1784 downregulated genes (Figure 4A). Overall, the kidney transcriptome was similar at different time points of the 16 min, 18 min, 20 min, 22 min, 24 min, and 26 min. These findings are consistent with PCA results. FIGURE 4 The variation trend of the number of differentially expressed genes (DEGs) between 0 min group and different ischemia groups. (A) The variation trend of the number of DEGs among different compared groups (B) The number of intersected DEGs between different compared groups. The number above stands for the number of DEGs that intersect between different groups. The overlapping DEGs were further analyzed by UpSetR package in R software. The number of unique DEGs that only found in 16 min–0 min group, 18 min–0 min group, 20 min–0 min group, 22 min–0 min group, 24 min–0 min group, 26 min–0 min group, 28 min–0 min group, and 30 min–0 min group were 17,43,29,50,49,66,1034,1,095, respectively (Figure 4B). Moreover, 498 genes were differentially expressed in each group. We also obtained a union DEGs set, which included the combination of DEGs obtained in each ischemia group compared with the ischemia 0 min and 16 min groups. This union DEGs set contains 4,519 genes. 3.4 The time-dependent gene expression patterns of genes in union DEGs set In order to visualize the expression patterns and characteristics of all 4,519 union DEGs, we applied the Mfuzz R package on them. After standardization, genes in union DEGs set were assigned to 8 clusters (Figure 5). The expressions of genes in Cluster 1 were highest in 0 min group, declined to a plateau in 16 min group to 26 min group, and reached the lowest in 28 min group and 30 min group. The expressions of genes in cluster 2 were opposite with cluster 1. The expressions of genes in Cluster 3 were lowest in 0 min group, increased to a plateau in 16 min group to 26 min group, increased then in 28 min group, and reached the highest in 30 min group. The difference between cluster 2 and cluster 3 is that the expressions of genes in Cluster 3 were much higher in 30 min group than 28 min group. The expressions of genes in Cluster 4 were low in 0min group to 26 min group, and continuous decrease in 28 min group and 30 min group, the expressions of genes in cluster 5 were opposite with cluster 4. The expressions of genes in Cluster 6 were highest in 0 min group, and decreased to a plateau in 16 min group to 30 min group. The expressions of genes in cluster 7 and cluster 8 were firstly increased and then decreased after reaching a plateau. The difference between these two clusters was that the expressions of genes of in cluster 7 start to decreased in the 30 min group, while those of cluster 8 began to decline in the 28 min group. FIGURE 5 The time-dependent gene expression patterns of genes in union DEGs set. Eight clusters identified based on gene expression of genes in union DEGs set by Mfuzz R package. Curves in yellow or green denote genes with small membership values, while curves in red or purple correspond to genes with big membership values. By searching the STRING database, the PPI network of the genes in each cluster acquired by soft clustering above were selected with a score>0.4. We then used the plug-in named Molecular Complex Detection (MCODE) by Cytoscape to filter the modules of the PPI networks. The first module of each cluster was selected (Supplementary Figure S1). 3.5 GO and KEGG pathway enrichment analyses of genes in each cluster To gain insight into the biological roles of each cluster, KEGG pathway and GO term analyses were performed. The results showed that genes in each cluster were associated with distinct biological processes. Genes in Cluster 1, cluster 4 and cluster 6 were down-regulated with the prolongation of ischemia time. Cluster 1 was mainly associated with metabolic-related processes, such as “fatty acid metabolic process”, “purine-containing compound metabolic process”, “valine, leucine and isoleucine degradation”, “pyruvate metabolism”, and “tryptophan metabolism”. In addition, immune-related processes such as “complement and coagulation cascades”, “antigen processing and presentation”, and “Th1 and Th2 cell differentiation” were also enriched in Cluster 1. Similar to Cluster 1, Cluster 4 genes were also mainly involved in metabolic-related processes. Genes in Cluster 6 were mainly involved in immune-related processes, such as “antigen processing and presentation”, “cytokine-cytokine receptor interaction”, “response to interferon-gamma”, “leukocyte mediated cytotoxicity”, and “T cell mediated cytotoxicity” (Supplementary Figure S2). These results indicated that, the main event of decline expression genes from ischemia 0 min to ischemia 16 min were immune-related processes, while genes with sharply decreased expression from ischemia 26 min to ischemia 28 min were mainly enriched in metabolism-related pathways. Genes in Cluster 2, cluster 3 and cluster 5 were up-regulated with the prolongation of ischemia time. Cluster 2 were mainly involved in cell junction-related processes, such as “tight junction”, “focal adhesion”, “cell junction assembly”, and “cell-matrix adhesion”. In addition, “apoptosis”, “PI3K-Akt signaling pathway”, “MAPK signaling pathway”, and “complement and coagulation cascades” were also enriched in Cluster 2. Genes in Cluster 3 were mainly associated with “wound healing”, “MAPK signaling pathway”, “TNF signaling pathway”, “PI3K-Akt signaling pathway” and “epithelial cell proliferation”. Cluster 5 were mainly involved in “renin-angiotensin system”, “cytokine-cytokine receptor interaction”, “HIF-1 signaling pathway”, “calcium signaling pathway”, “hydrolase activity” (Supplementary Figure S3). Genes in cluster 7 and cluster 8 showed trapezoidal expression. They mainly enriched in DNA-repair related pathway, such as “DNA replication”, “cell cycle”, “mismatch repair”, “nucleotide excision repair”, and “recombinational repair”. In addition, genes in cluster 8 also involved in “cellular senescence”, “p53 signaling pathway”, “apoptosis - multiple species”, “gap junction”, “tyrosine metabolism” (Supplementary Figure S4). 3.6 GSEA enrichment analysis of key genes and pathways in two key time intervals According to the results of time-dependent gene expression cluster analysis, gene expression changes were most significant in the interval of 0 min–16 min and 26 min to 28 min. Therefore, we further analyzed the GSEA enrichment changes of these two key time intervals in order to find common and characteristic enrichment pathways. GSEA analysis showed that there were many common enrichment pathways in the 16 min–0 min group and 28 min–26 min group, among which DNA-repair related pathways, such as “DNA replication”, “mismatch repair”, and “nucleotide excision repair”, were all up-regulated in both two key time intervals. While metabolism related pathways, such as “propanoate metabolism”, “steroid hormone biosynthesis”, “Valine, leucine and isoleucine degradation”, and “glyoxylate and dicarboxylate metabolism” were all down-regulated in both two key time intervals (Table 1). TABLE 1 The common enrichment pathways in GSEA analysis of genes in 16min–0min group and 28min–26min group. 16-0 min group 28-26 min group GSEA enrichment pathways NES p.adjust Express trend NES p.adjust Express trend DNA replication 0.766 0.002 up-regulation 0.62 0.004 up-regulation IL-17 signaling pathway 0.532 0.002 up-regulation 0.417 0.007 up-regulation p53 signaling pathway 0.555 0.002 up-regulation 0.44 0.007 up-regulation Mismatch repair 0.718 0.004 up-regulation 0.571 0.022 up-regulation Nucleotide excision repair 0.607 0.004 up-regulation 0.514 0.009 up-regulation Valine, leucine and isoleucine degradation −0.517 0.006 down-regulation −0.699 0.002 down-regulation beta-Alanine metabolism −0.625 0.008 down-regulation −0.656 0.002 down-regulation Propanoate metabolism −0.638 0.008 down-regulation −0.629 0.004 down-regulation Complement and coagulation cascades −0.432 0.011 down-regulation −0.463 0.005 down-regulation Peroxisome −0.448 0.011 down-regulation −0.614 0.002 down-regulation Tryptophan metabolism −0.544 0.013 down-regulation −0.672 0.002 down-regulation Pantothenate and CoA biosynthesis −0.625 0.022 down-regulation −0.673 0.002 down-regulation Butanoate metabolism −0.608 0.023 down-regulation −0.685 0.002 down-regulation Vitamin digestion and absorption −0.614 0.027 down-regulation −0.61 0.022 down-regulation Glyoxylate and dicarboxylate metabolism −0.567 0.034 down-regulation −0.654 0.002 down-regulation Phenylalanine metabolism −0.62 0.04 down-regulation −0.719 0.002 down-regulation Proteasome −0.594 0.004 down-regulation 0.595 0.002 up-regulation Protein digestion and absorption 0.467 0.006 up-regulation −0.592 0.002 down-regulation ECM-receptor interaction 0.456 0.009 up-regulation −0.435 0.01 down-regulation GSEA, gene set enrichment analysis; NES, normalized enrichment score; ECM, extra-cellular matrix. In addition, immune-related pathways, such as “antigen processing and presentation”, “Th1 and Th2 cell differentiation” were downregulated only in 16 min–0 min group (Table 2); “TNF signaling pathway”, “glutathione metabolism”, “gap junction”, “fatty acid biosynthesis”, and “cell cycle” were upregulated only in 16 min–0 min group. In contrary, metabolism-related pathways, such as “glycolysis/Gluconeogenesis”, “tyrosine metabolism”, “glycine, serine and threonine metabolism”, “PPAR signaling pathway”, “citrate cycle”, and “nitrogen metabolism”, and oxidative stress related pathways, such as “oxidative phosphorylation” were downregulated only in 28 min–26 min group (Table 3). TABLE 2 The unique enrichment pathways in GSEA analysis of genes in 16min–0min group. GSEA enrichment pathways NES p.adjust Express trend Cell cycle 0.707 0.002 up-regulation Homologous recombination 0.682 0.002 up-regulation Fatty acid biosynthesis 0.605 0.046 up-regulation Gap junction 0.505 0.002 up-regulation Glutathione metabolism 0.467 0.021 up-regulation TNF signaling pathway 0.427 0.009 up-regulation Antigen processing and presentation −0.792 0.002 down-regulation ABC transporters −0.583 0.004 down-regulation Th1 and Th2 cell differentiation −0.574 0.002 down-regulation Cell adhesion molecules −0.505 0.004 down-regulation Th17 cell differentiation −0.495 0.004 down-regulation Phagosome −0.474 0.004 down-regulation Natural killer cell mediated cytotoxicity −0.458 0.011 down-regulation GSEA, gene set enrichment analysis; NES, normalized enrichment score. TABLE 3 The unique enrichment pathways in GSEA analysis of genes in 28min–26min group. GSEA enrichment pathway NES p.adjust Express trend Nucleocytoplasmic transport 0.443 0.002 up-regulation C-type lectin receptor signaling pathway 0.442 0.002 up-regulation Oxidative phosphorylation −0.502 0.002 down-regulation PPAR signaling pathway −0.557 0.002 down-regulation Retinol metabolism −0.561 0.002 down-regulation Glycolysis/Gluconeogenesis −0.567 0.002 down-regulation Starch and sucrose metabolism −0.619 0.002 down-regulation Tyrosine metabolism −0.645 0.002 down-regulation Histidine metabolism −0.711 0.002 down-regulation Carbon metabolism −0.431 0.003 down-regulation Glycine, serine and threonine metabolism −0.567 0.004 down-regulation Citrate cycle (TCA cycle) −0.64 0.004 down-regulation Arginine and proline metabolism −0.526 0.005 down-regulation Renin-angiotensin system −0.604 0.005 down-regulation Nitrogen metabolism −0.739 0.005 down-regulation 2-Oxocarboxylic acid metabolism −0.7 0.006 down-regulation Biosynthesis of amino acids −0.439 0.025 down-regulation Arginine biosynthesis −0.647 0.026 down-regulation Pyruvate metabolism −0.491 0.027 down-regulation Ascorbate and aldarate metabolism −0.564 0.027 down-regulation Sulfur metabolism −0.709 0.027 down-regulation Fatty acid metabolism −0.449 0.032 down-regulation Collecting duct acid secretion −0.56 0.036 down-regulation Endocrine and other factor-regulated calcium reabsorption −0.437 0.038 down-regulation Fatty acid degradation −0.46 0.039 down-regulation Other glycan degradation −0.602 0.039 down-regulation Taurine and hypotaurine metabolism −0.654 0.039 down-regulation Cysteine and methionine metabolism −0.456 0.041 down-regulation GSEA, gene set enrichment analysis; NES, normalized enrichment score; PPAR, peroxisome proliferators-activated receptors; TCA, tricarboxylic acid cycle. These results were consistent with previous results of GO and KEGG pathway analysis, namely, gene expression of immune-related pathways was down-regulated in short-term ischemia, while in long-term ischemia, metabolism-related pathways were the mainly enriched pathway. Thus, we performed a more detailed analysis of genes in immune-related pathways and metabolism-related pathways (Figure 6). Genes, such as Pgam2, Idh2, Gnmt, Prdx6, Gsta1, Stat1, and Tnfsf10 were critical genes in maintaining immune and metabolism process. FIGURE 6 The relationship of unique enrichment pathways in GSEA analysis of genes in 16 min–0 min group or 28 min–26 min group. Two dotted boxes represent functional modules of immune related pathways and metabolism related pathways. The circles represent genes and the triangles represent GSEA enrichment pathways. Different colors represent genes enriched in different clusters. Genes are connected based on the pathways they belong to. GSEA, gene set enrichment analysis. Furthermore, as indicated above, inflammation related pathways and cell repair related pathways were critical upregulated processes that reflect the kidney injury and repair. Thus, we also performed a more detailed analysis of genes in the pertinent enriched pathways (Figure 7). Pathways in inflammation related pathways and cell repair related pathways were connected by mutually overlapping genes and form an integrated functional module. Mutual genes, such as Fos, Jun, Lcn2, Cxcl1, Cxcl2, Cxcl5, Ccl2, Ptgs2, Mapk11, Mapk12, Bcl3 are critical genes in maintaining inflammation interactions; Trp53, Cdk4, Pold4, Pole2, and Lig1 are critical genes in maintaining cell repair and replications. FIGURE 7 The relationship of upregulated enrichment pathways in GSEA analysis of genes in 16 min–0 min group and 28 min–26 min group. Two dotted boxes represent functional modules of inflamation related pathways and cell repair related pathways. The rectangles represent genes and the octagons represent GSEA enrichment pathways. Different colors represent genes enriched in different clusters. Genes are connected based on the pathways they belong to. GSEA, gene set enrichment analysis. 3.7 Verification of key genes expression In addition, to verify the expression pattern of the key genes, we performed qPCR assays for the four representative genes in the kidney sample (ischemia 0 min, 16 min, 26 min, and 28 min) and compared them with the RNA-seq results. As shown in Figure 8, the expression patterns of Stat1, Lcn2, Pgam2, and Ptgs2 were highly consistent between the RNA-seq and qPCR results. In conclusion, this finding suggests that our RNA-seq data were reliable, and the function and specific mechanism of these key genes in renal injury repair deserve further study. FIGURE 8 Validation of expression patterns of four representative key genes. (A–D) The expression level of Stat1(A), Lcn2(B), Pgam2(C), and Ptgs2(D). To verify the dynamic change of expression level, we selected four representative genes to perform the qPCR assay. The expression patterns of Stat1, Lcn2, Pgam2, and Ptgs2 are highly consistent between RNA-seq data and qPCR results. The x-axis represents the time points. The right black y-axis represents the RPKM value from RNA-seq data, and the left blue y-axis indicates the relative expression level of qPCR results. 4 Discussion How many minutes of ischemia time is a safe warm ischemia time of kidney? Different studies have drawn mixed conclusions (Lane et al., 2008; Funahashi et al., 2009; Thompson et al., 2010). What are the pathological changes and molecular characteristics of different renal ischemia time? There are also lacking systematic research. The experimental mouse models of renal ischemia/reperfusion injury (IRI) can well simulate renal injury caused by renal artery occlusion in human partial nephrectomy. However, there are various surgical methods for IRI model reported in different studies: bilateral IRI (bIRI) (Chen et al., 2018; Chen et al., 2022), contralateral kidney reservation plus unilateral IRI (uIRI) (Godwin et al., 2010; Huang et al., 2022), contralateral nephrectomy plus unilateral IRI (uIRIx) (Ferhat et al., 2018; Wang et al., 2022), unilateral IRI plus contralateral resection 14d before surgery (Chou et al., 2020), and so on (Wei and Dong, 2012; Fu et al., 2018; Wei et al., 2019a; Shiva et al., 2020). The setting of IRI ischemia time varies from 18min to 35min. It has been reported that uIRIx model mice with ischemia greater than 18 min, or bIRI model mice with ischemia greater than 21 min will die within 72 h (Wei et al., 2019a). There were also reports of bilateral ischemia of 21min (Liu et al., 2017), 26 min (Inoue et al., 2016), 28 min (Liu et al., 2014), 30min (Chen et al., 2019), 35 min (Chen et al., 2018), and mice can survive for at least 7 days. In this study, combined with the experience of previous studies, we established a relatively stable uIRIx model to explore the pathological features, and molecular network characteristics of mouse kidney with different degrees of ischemia. We found that there was no significant difference among 0 min–26 min of ischemia in SCr, while the SCr was dramatically changed between 26 min and 28 min ischemia groups at the condition of 37°C. However, the pathological changes were not exactly the same as SCr. Although there was no change in SCr, renal damage such as cellular casts and renal tubular obstruction could be observed from the group with more than 20 min of ischemia. Soranno et al. reported that SCr may be poor predictor of long-term histological and functional outcomes after uIRI (Soranno et al., 2019). Slocum et al. also reported that SCr is insensitive, nonspecific, and a late marker of disease (Slocum et al., 2012). Even though 24 h is the peak period of SCr after IRI (Hesketh et al., 2014; Liu et al., 2017), SCr remains unelevated in mild injury, and focusing solely on SCr may mask the severity of renal injury. The SCr at reperfusion 24 h reached the maximum in ischemia 30 min group, and there was no significant difference in SCr between ischemia 60 min and ischemia 30 min (data were not shown) at reperfusion 24 h, which means that SCr also did not totally reflect the extent of acute severe kidney injury. Some experts reported that both pathological and urea nitrogen (BUN) changes occurred after 18 min of renal ischemia (Kirita et al., 2020), while other reports showed that pathological changes were not significant after 20 min of ischemia (Ide et al., 2021). Such disunity indicates that there are complex influencing factors in the animal model of renal ischemia-reperfusion injury. So far, many factors affect the degree of renal injury after ischemia-reperfusion, including ischemia time, intraoperative kidney temperature, surgical method (unilateral or bilateral renal ischemia-reperfusion), surgical incision site (midline or lateral abdomen), age, gender, and even the type of the nontraumatic microaneurysm clamp, etc. Temperature has a great influence on ischemic tolerance. It has been reported that mice with 2 h of cold ischemia exhibited no significant changes in renal function or histopathology, and 3 h or 4 h of cold ischemia led to mild to moderate acute kidney injury with characteristic features (Wei et al., 2019b). Males are more susceptible to AKI than females, which means female sex protects against development of acute kidney injury (AKI) (Viñas et al., 2020). AKI onset age defines whether the kidney undergoes repair or maladaptive remodeling, the older was less tolerant to injury, and more susceptible to undergo maladaptive repair (Rudman-Melnick et al., 2020). Other than SCr and BUN, many other injury-related molecules, such as Kim1 (havcr1), Lcn2 (Ngal), and Il6, were also determined to predict kidney damage. Our results showed that the increase of Kim1 and Lcn2 expression emerged at 16 min and 22 min of ischemia respectively, and gradually enhanced according to the ischemia time. However, the expression of Il6 was significantly increased at 30 min of ischemia, with no significant changes in less than 28 min. As can be seen, the molecular expression patterns of each injured molecule are not the same under different ischemic time. Therefore, we further performed high-throughput RNA-Seq. High-throughput RNA-Seq results also confirmed that the molecular levels of kidney genes had started to change at 16 min–26 min of ischemia whereas SCr showed no difference. The pathology of the 863 DEGs in the 16 min ischemia group presented no alterations. Those results indicate that molecular alterations were faster and earlier than pathology changes and SCr. It is also worth noting whether there will be changes under electron microscopy at the organelle level rather than cellular level. Among the 863 DEGs in the 16 min–0 min group, Kim1 (Havcr1) presented the highest expression DEGs, and the log(fold change) was 4.99. Since Kim1 can be detected in blood and urine (Vinken et al., 2012), and its peak occurred 24 h after IRI (Chen et al., 2022), blood or urine tests for Kim1 may be considered a possible method to detect signs of mild kidney damage”. For a long time, AKI was considered to be a self-recovery disease, while recent studies have highlighted that AKI can lead to fibrosis and CKD (Venkatachalam et al., 2015; Kormann et al., 2020). When kidney injury is mild and baseline function is normal, the repair process can be adaptive with few long-term consequences. When the injury is more severe, repeated, or to a kidney with underlying disease, the repair can be maladaptive (Canaud and Bonventre, 2015). In order to find the molecular network characteristics of mild to severe ischemia reperfusion kidney injury, and explain key changes in the occurrence of mild and severe injuries, we classified all 4,519 union DEGs identified in high-throughput RNA-Seq into 8 clusters according to the expression patterns. The expression of genes in Cluster 2 and Cluster 3 were both enhanced with the increase of ischemia degree. Genes in Cluster1, Cluster 4, and Cluster 6 presented reduced expression after ischemia. Inhibiting the function of continuously increased expression genes in Cluster 2 and Cluster 3 or promoting the function of the continuously decreased expression genes in Cluster1 and Cluster 4 may have a potential renal protective effect on AKI. For example, Cdk4 and Cdk6 were members of Cluster 2, Pabla et al. found that G1/S-regulating cyclin dependent kinase 4/6 (CDK4/6) pathway is activated in nephrotoxic AKI (Pabla et al., 2015), and targeted inhibition of CDK4/6 pathway by small-molecule inhibitors resulted in inhibition of cell-cycle progression, amelioration of kidney injury, and improved overall survival. Targeted inhibition of CDK4/6 pathway by small-molecule inhibitors palbociclib (PD-0332991) and ribociclib (LEE011) resulted in inhibition of cell-cycle progression, amelioration of kidney injury, and improved overall survival (Pabla et al., 2015; Brown et al., 2020; Kim et al., 2020). Kim et al. also reported that mice treated with PD 0332991 before IRI exhibited dramatically reduced epithelial progression through S phase 24 h after IRI and ameliorated kidney injury, as reflected by improved SCr and blood urea nitrogen levels 24 h after injury. The inflammatory markers and macrophage infiltration were also significantly decreased in injured kidneys 3 days following IRI. The expression of genes in cluster 7 and cluster 8, which mainly enriched in DNA-repair related pathway, increased in mild injury and decreased in severe injury. It was reported that epithelial cell cycle arrest plays an important role in the development of fibrosis (Bonventre, 2014; Lovisa et al., 2015; Moonen et al., 2018). Thus, we speculate that these genes in cluster 7 and cluster 8 may participate in the process of injury repair and play an important role in physiological repair function. According to the results of time-dependent gene expression cluster analysis, gene expression changes were most significant in two key time intervals, 0 min–16 min and 26 min to 28 min. GSEA enrichment analysis of these two key time intervals were consistence with above results. These results revealed that gene expression of immune-related pathways was down-regulated in short-term ischemia, while in long-term ischemia metabolism-related pathways were the mainly enriched pathway. This finding suggests the important role of metabolic pathways in ischemia-reperfusion. Recently, single-cell transcriptomic data showed that metabolic processes are critical for maintaining the cellular identity of fully differentiated PT cells, and dysregulation of these pathways underlies the failure of damage-associated PT cells to redifferentiate into normal PT cell state (Ide et al., 2021). Matrix-assisted laser desorption/ionization−mass-spectrometry imaging (MALDI-MSI) results also showed that differential expression of characteristic lipid-degradation products between severe and mild ischemia was detected within 2 h after IRI, even though the histopathological examination revealed no significant difference between kidneys (van Smaalen et al., 2019). As validated by qPCR assays, the expression patterns of four representative genes, Stat1, Lcn2, Pgam2, and Ptgs2 were highly consistent between the RNA-seq and qPCR results. These key genes were very important in the process of AKI, which could be used to explore some new diagnostic and therapeutic strategies. Signal transducer and activator of transcription 1(STAT1) is a member of STATs, which are important transcription factors that regulate the expression of inflammatory genes (Leonard and O'Shea, 1998). JAK-STAT activation is involved in the pathogenesis of AKI, and is also a key route in the signaling cascade of cytokine mediated AKI (Yun et al., 2021). JAK-STAT inhibitors may prevent and/or treat AKI via blockade of the feedback loop of proinflammatory cytokines, and the inhibition of STAT phosphorylation could also significantly downregulate the production of proinflammatory cytokines during the process of renal inflammation and injury (Li et al., 2019). Lipocalin 2 (LCN2), also known as neutrophil gelatinase associated lipocalin (NGAL), is a secreted protein that belongs to the Lipocalins, and it is also known as an important and common marker of renal injury (Asimakopoulou et al., 2016). Lcn2 is expressed in kidney cells and its production markedly increases in response to stimulation such as ischemia, and plays a critical role in renal ischemia/reperfusion induced AKI by regulating autophagy activation. Glycolytic enzyme phosphoglycerate mutase 2 (Pgam2), is an isoform of Pgam, which converts 3-phosphoglycerate into 2-phosphoglycerate as an isomerase (Mikawa et al., 2021). It was reported that type-M of PGAM2 was specifically expressed in muscles and could be a potential biomarker for early myocardial ischemia (Li et al., 2012), while there is still no related research of Pgam2 in kidney injury. Prostaglandin-endoperoxide synthase 2 (Ptgs2), also called cyclooxygenase-2(Cox-2), is positively correlated with the severity of the inflammatory response (Fan et al., 2021). Ptgs2 was reported upregulated after AKI in our previously research (Chen et al., 2020). Inhibition of Ptgs2 plays as a renal protective role against ischemia-induced damage in AKI (Fan et al., 2021; Liu et al., 2021). The present study has some limitations and drawbacks. First, because we mainly focused on observing molecular network characteristics and pathological features of mild to severe ischemia reperfusion kidney injury, the main experimental observation time was 1d after AKI. Therefore, we did not monitor the long-term prognosis in mild to severe AKI. Second, this study lacks human kidney tissue sample data, which may result in not full applicability in human. Third, the above results, such as the gene expression level and gene function, should be validated by further mechanism experiments. These issues will be illustrated in our future research. In summary, our results indicated that there was no absolute safe renal warm ischemia time, and “every minute counts”, which means every minute of ischemia increases kidney damage. Warm ischemia 26 min or above in mice makes severe kidney injury, renal pathology and SCr were significantly changed. Warm ischemia between 18 and 26-min leads to mild kidney injury, with changes in pathology and renal molecular expression, while SCr did not changed. No obvious pathological changes under 16 min warm ischemia were observed, while significant differences were found in molecular expression. The results also showed that SCr could not completely reflect the severity of kidney injury. There are two key time intervals in the process of renal ischemia injury, 0 min–16 min and 26 min to 28 min. Gene expression of immune-related pathways were most significantly down-regulated in short-term ischemia, while metabolism-related pathways were the mainly enriched pathway in long-term ischemia. Taken together, this study provides novel insights into safe renal artery occlusion time in partial nephrectomy, and is of great value for elucidating molecular network characteristics and pathological features of mild to severe ischemia reperfusion kidney injury, and key genes related to metabolism and immune found in this study also provide potential diagnostic and therapeutic biomarkers for AKI. Data availability statement The datasets presented in this study are available in the article and online Supplementary Material. All raw RNA-Seq data discussed in this publication have been deposited in NCBI Gene Expression Omnibus with the following ID: GSE192883 and was available at: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE192883. Ethics statement The animal study was reviewed and approved by the Institutional Animal Care and Use Committee of the Chinese PLA General Hospital. Author contributions Y-LC and J-WC conceived the study, analyzed the RNAseq data, performed tissue analyses, and prepared the manuscript. H-KL and LW analyzed and interpreted the data, and performed the qPCR assays. H-KL and LW participated in animal experiment, data acquisition, and data analysis. XM and J-WC designed and directed the project. All authors drafted the work, revised it critically for important intellectual content, and approved the final version. Conflict of interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Publisher’s note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. Supplementary material The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fmolb.2022.1006917/full#supplementary-material Click here for additional data file. ==== Refs References Asimakopoulou A. Weiskirchen S. Weiskirchen R. (2016). Lipocalin 2 (LCN2) expression in hepatic malfunction and therapy. Front. Physiol. 7 , 430. 10.3389/fphys.2016.00430 27729871 Bellomo R. Kellum J. A. Ronco C. (2012). Acute kidney injury. Lancet 380 (9843 ), 756–766. 10.1016/S0140-6736(11)61454-2 22617274 Bonventre J. V. (2014). Maladaptive proximal tubule repair: Cell cycle arrest. Nephron. Clin. Pract. 127 (1-4 ), 61–64. 10.1159/000363673 25343823 Brown C. N. Atwood D. J. Pokhrel D. Ravichandran K. Holditch S. J. Saxena S. (2020). The effect of MEK1/2 inhibitors on cisplatin-induced acute kidney injury (AKI) and cancer growth in mice. Cell. Signal. 71 , 109605. 10.1016/j.cellsig.2020.109605 32194168 Canaud G. Bonventre J. V. (2015). Cell cycle arrest and the evolution of chronic kidney disease from acute kidney injury. Nephrol. Dial. Transpl. 30 (4 ), 575–583. 10.1093/ndt/gfu230 Chen J. Chen Y. Olivero A. Chen X. (2020). Identification and validation of potential biomarkers and their functions in acute kidney injury. Front. Genet. 11 , 411. 10.3389/fgene.2020.00411 32528518 Chen J. You H. Li Y. Xu Y. He Q. Harris R. C. (2018). EGF receptor-dependent YAP activation is important for renal recovery from AKI. J. Am. Soc. Nephrol. 29 (9 ), 2372–2385. 10.1681/ASN.2017121272 30072422 Chen J. W. Huang M. J. Chen X. N. Wu L. L. Li Q. G. Hong Q. (2022). Transient upregulation of EGR1 signaling enhances kidney repair by activating SOX9(+) renal tubular cells. Theranostics 12 (12 ), 5434–5450. 10.7150/thno.73426 35910788 Chen S. Fu H. Wu S. Zhu W. Liao J. Hong X. (2019). Tenascin-C protects against acute kidney injury by recruiting Wnt ligands. Kidney Int. 95 (1 ), 62–74. 10.1016/j.kint.2018.08.029 30409456 Chou Y. H. Pan S. Y. Shao Y. H. Shih H. M. Wei S. Y. Lai C. F. (2020). Methylation in pericytes after acute injury promotes chronic kidney disease. J. Clin. Invest. 130 (9 ), 4845–4857. 10.1172/JCI135773 32749240 Fan H. Le J. W. Sun M. Zhu J. H. (2021). Pretreatment with S-nitrosoglutathione attenuates septic acute kidney injury in rats by inhibiting inflammation, oxidation, and apoptosis. Biomed. Res. Int. 2021 , 6678165. 10.1155/2021/6678165 33604382 Ferhat M. Robin A. Giraud S. Sena S. Goujon J. M. Touchard G. (2018). Endogenous IL-33 contributes to kidney ischemia-reperfusion injury as an alarmin. J. Am. Soc. Nephrol. 29 , 1272–1288. 10.1681/ASN.2017060650 29436517 Fu Y. Tang C. Cai J. Chen G. Zhang D. Dong Z. (2018). Rodent models of AKI-CKD transition. Am. J. Physiol. Ren. Physiol. 315 (4 ), F1098–f1106. 10.1152/ajprenal.00199.2018 Funahashi Y. Hattori R. Yamamoto T. Kamihira O. Kato K. Gotoh M. (2009). Ischemic renal damage after nephron-sparing surgery in patients with normal contralateral kidney. Eur. Urol. 55 (1 ), 209–215. 10.1016/j.eururo.2008.07.048 18706758 Godwin J. G. Ge X. Stephan K. Jurisch A. Tullius S. G. Iacomini J. (2010). Identification of a microRNA signature of renal ischemia reperfusion injury. Proc. Natl. Acad. Sci. U. S. A. 107 (32 ), 14339–14344. 10.1073/pnas.0912701107 20651252 Gonsalez S. R. Cortês A. L. Silva R. C. D. Lowe J. Prieto M. C. Silva Lara L. d. (2019). Acute kidney injury overview: From basic findings to new prevention and therapy strategies. Pharmacol. Ther. 200 , 1–12. 10.1016/j.pharmthera.2019.04.001 30959059 Hesketh E. E. Czopek A. Clay M. Borthwick G. Ferenbach D. Kluth D. (2014). Renal ischaemia reperfusion injury: A mouse model of injury and regeneration. J. Vis. Exp. 88 , 51816. 10.3791/51816 Huang M. Li D. Chen J. Ji Y. Su T. Chen Y. (2022). Comparison of the treatment efficacy of umbilical mesenchymal stem cell transplantation via renal subcapsular and parenchymal routes in AKI-CKD mice. Stem Cell Res. Ther. 13 (1 ), 128. 10.1186/s13287-022-02805-3 35337372 Ide S. Kobayashi Y. Ide K. Strausser A. S. Abe A. Herbek S. (2021). Ferroptotic stress promotes the accumulation of pro-inflammatory proximal tubular cells in maladaptive renal repair. Elife 10 , e68603. 10.7554/eLife.68603 34279220 Inoue T. Abe C. Sung S. S. Moscalu S. Jankowski J. Huang L. (2016). Vagus nerve stimulation mediates protection from kidney ischemia-reperfusion injury through α7nAChR+ splenocytes. J. Clin. Invest. 126 (5 ), 1939–1952. 10.1172/JCI83658 27088805 Kim J. Y. Jayne L. A. Bai Y. Feng M. J. H. H. Clark M. A. Chung S. (2020). Ribociclib mitigates cisplatin-associated kidney injury through retinoblastoma-1 dependent mechanisms. Biochem. Pharmacol. 177 , 113939. 10.1016/j.bcp.2020.113939 32229099 Kirita Y. Wu H. Uchimura K. Wilson P. C. Humphreys B. D. (2020). Cell profiling of mouse acute kidney injury reveals conserved cellular responses to injury. Proc. Natl. Acad. Sci. U. S. A. 117 (27 ), 15874–15883. 10.1073/pnas.2005477117 32571916 Kormann R. Kavvadas P. Placier S. Vandermeersch S. Dorison A. Dussaule J. C. (2020). Periostin promotes cell proliferation and macrophage polarization to drive repair after AKI. J. Am. Soc. Nephrol. 31 (1 ), 85–100. 10.1681/ASN.2019020113 31690575 Kumar L. Futschik M. E. (2007). Mfuzz: A software package for soft clustering of microarray data. Bioinformation 2 (1 ), 5–7. 10.6026/97320630002005 18084642 Lane B. R. Novick A. C. Babineau D. Fergany A. F. Kaouk J. H. Gill I. S. (2008). Comparison of laparoscopic and open partial nephrectomy for tumor in a solitary kidney. J. Urol. 179 (3 ), 847–851. discussion 852. 10.1016/j.juro.2007.10.050 18221958 Leonard W. J. O'Shea J. J. (1998). Jaks and STATs: Biological implications. Annu. Rev. Immunol. 16 , 293–322. 10.1146/annurev.immunol.16.1.293 9597132 Li H. Li J. Wang Y. Yang T. (2012). Proteomic analysis of effluents from perfused human heart for transplantation: Identification of potential biomarkers for ischemic heart damage. Proteome Sci. 10 (1 ), 21. 10.1186/1477-5956-10-21 22443514 Li Y. Zhou H. Li Y. Han L. Song M. Chen F. (2019). PTPN2 improved renal injury and fibrosis by suppressing STAT-induced inflammation in early diabetic nephropathy. J. Cell. Mol. Med. 23 (6 ), 4179–4195. 10.1111/jcmm.14304 30955247 Liu D. Zhang C. Hu M. Su K. (2021). Scutellarein relieves the death and inflammation of tubular epithelial cells in ischemic kidney injury by degradation of COX-2 protein. Int. Immunopharmacol. 101 , 108193. 10.1016/j.intimp.2021.108193 34619498 Liu J. Krautzberger A. M. Sui S. H. Hofmann O. M. Chen Y. Baetscher M. (2014). Cell-specific translational profiling in acute kidney injury. J. Clin. Invest. 124 (3 ), 1242–1254. 10.1172/JCI72126 24569379 Liu J. Kumar S. Dolzhenko E. Alvarado G. F. Guo J. Lu C. (2017). Molecular characterization of the transition from acute to chronic kidney injury following ischemia/reperfusion. JCI Insight 2 (18 ), 94716. 10.1172/jci.insight.94716 28931758 Lovisa S. LeBleu V. S. Tampe B. Sugimoto H. Vadnagara K. Carstens J. L. (2015). Epithelial-to-mesenchymal transition induces cell cycle arrest and parenchymal damage in renal fibrosis. Nat. Med. 21 (9 ), 998–1009. 10.1038/nm.3902 26236991 Melo Ferreira R. Sabo A. R. Winfree S. Collins K. S. Janosevic D. Gulbronson C. J. (2021). Integration of spatial and single-cell transcriptomics localizes epithelial cell-immune cross-talk in kidney injury. JCI Insight 6 (12 ), 147703. 10.1172/jci.insight.147703 34003797 Mikawa T. Shibata E. Shimada M. Ito K. Ito T. Kanda H. (2021). Characterization of genetically modified mice for phosphoglycerate mutase, a vitally-essential enzyme in glycolysis. PLoS One 16 (4 ), e0250856. 10.1371/journal.pone.0250856 33914812 Moonen L. D'Haese P. C. Vervaet B. A. (2018). Epithelial cell cycle behaviour in the injured kidney. Int. J. Mol. Sci. 19 (7 ), E2038. 10.3390/ijms19072038 Pabla N. Gibson A. A. Buege M. Ong S. S. Li L. Hu S. (2015). Mitigation of acute kidney injury by cell-cycle inhibitors that suppress both CDK4/6 and OCT2 functions. Proc. Natl. Acad. Sci. U. S. A. 112 (16 ), 5231–5236. 10.1073/pnas.1424313112 25848011 Rudman-Melnick V. Adam M. Potter A. Chokshi S. M. Ma Q. Drake K. A. (2020). Single-cell profiling of AKI in a murine model reveals novel transcriptional signatures, profibrotic phenotype, and epithelial-to-stromal crosstalk. J. Am. Soc. Nephrol. 31 (12 ), 2793–2814. 10.1681/ASN.2020010052 33115917 Shiva N. Sharma N. Kulkarni Y. A. Mulay S. R. Gaikwad A. B. (2020). Renal ischemia/reperfusion injury: An insight on in vitro and in vivo models. Life Sci. 256 , 117860. 10.1016/j.lfs.2020.117860 32534037 Slocum J. L. Heung M. Pennathur S. (2012). Marking renal injury: Can we move beyond serum creatinine? Transl. Res. 159 (4 ), 277–289. 10.1016/j.trsl.2012.01.014 22424431 Soranno D. E. Gil H. W. Kirkbride-Romeo L. Altmann C. Montford J. R. Yang H. (2019). Matching human unilateral AKI, a reverse translational approach to investigate kidney recovery after ischemia. J. Am. Soc. Nephrol. 30 (6 ), 990–1005. 10.1681/ASN.2018080808 31072827 Szklarczyk D. Franceschini A. Wyder S. Forslund K. Heller D. Huerta-Cepas J. (2015). STRING v10: Protein-protein interaction networks, integrated over the tree of life. Nucleic Acids Res. 43 , D447–D452. 10.1093/nar/gku1003 25352553 Thompson R. H. Lane B. R. Lohse C. M. Leibovich B. C. Fergany A. Frank I. (2010). Every minute counts when the renal hilum is clamped during partial nephrectomy. Eur. Urol. 58 (3 ), 340–345. 10.1016/j.eururo.2010.05.047 20825756 Uber A. M. Sutherland S. M. (2020). Nephrotoxins and nephrotoxic acute kidney injury. Pediatr. Nephrol. 35 (10 ), 1825–1833. 10.1007/s00467-019-04397-2 31646405 van Smaalen T. C. Ellis S. R. Mascini N. E. Siegel T. P. Cillero-Pastor B. Hillen L. M. (2019). Rapid identification of ischemic injury in renal tissue by mass-spectrometry imaging. Anal. Chem. 91 (5 ), 3575–3581. 10.1021/acs.analchem.8b05521 30702282 Venkatachalam M. A. Weinberg J. M. Kriz W. Bidani A. K. (2015). Failed tubule recovery, AKI-CKD transition, and kidney disease progression. J. Am. Soc. Nephrol. 26 (8 ), 1765–1776. 10.1681/ASN.2015010006 25810494 Viñas J. L. Porter C. J. Douvris A. Spence M. Gutsol A. Zimpelmann J. A. (2020). Sex diversity in proximal tubule and endothelial gene expression in mice with ischemic acute kidney injury. Clin. Sci. 134 (14 ), 1887–1909. 10.1042/CS20200168 Vinken P. Starckx S. Barale-Thomas E. Looszova A. Sonee M. Goeminne N. (2012). Tissue Kim-1 and urinary clusterin as early indicators of cisplatin-induced acute kidney injury in rats. Toxicol. Pathol. 40 (7 ), 1049–1062. 10.1177/0192623312444765 22581811 Wang L. Peng C. Chen J. Li H. Jiao Q. Zhang Z. (2022). Intermittent hilar occlusion attenuates or prevents renal ischaemia-reperfusion in mice. Biomed. Pharmacother. 153 , 113457. 10.1016/j.biopha.2022.113457 36076492 Wei J. Wang Y. Zhang J. Wang L. Fu L. Cha B. J. (2019). A mouse model of renal ischemia-reperfusion injury solely induced by cold ischemia. Am. J. Physiol. Ren. Physiol. 317 (3 ), F616–f622. 10.1152/ajprenal.00533.2018 Wei J. Zhang J. Wang L. Jiang S. Fu L. Buggs J. (2019). New mouse model of chronic kidney disease transitioned from ischemic acute kidney injury. Am. J. Physiol. Ren. Physiol. 317 (2 ), F286–f295. 10.1152/ajprenal.00021.2019 Wei Q. Dong Z. (2012). Mouse model of ischemic acute kidney injury: Technical notes and tricks. Am. J. Physiol. Ren. Physiol. 303 (11 ), F1487–F1494. 10.1152/ajprenal.00352.2012 Weight C. J. Lieser G. Larson B. T. Gao T. Lane B. R. Campbell S. C. (2010). Partial nephrectomy is associated with improved overall survival compared to radical nephrectomy in patients with unanticipated benign renal tumours. Eur. Urol. 58 (2 ), 293–298. 10.1016/j.eururo.2010.04.033 20546991 Yang C. Ding M. Shao G. Jia S. Yin X. Cui Y. (2021). Kcnk3, Ggta1, and Gpr84 are involved in hyperbaric oxygenation preconditioning protection on cerebral ischemia-reperfusion injury. Exp. Brain Res. 239 (12 ), 3601–3613. 10.1007/s00221-021-06220-7 34591125 Yu G. Wang L. G. Han Y. He Q. Y. (2012). clusterProfiler: an R package for comparing biological themes among gene clusters. OMICS 16 (5 ), 284–287. 10.1089/omi.2011.0118 22455463 Yun Y. Chen J. Wang X. Li Y. Hu Z. Yang P. (2021). Tofacitinib ameliorates lipopolysaccharide-induced acute kidney injury by blocking the JAK-STAT1/STAT3 signaling pathway. Biomed. Res. Int. 2021 , 8877056. 10.1155/2021/8877056 33511217 Zhang J. Wang X. Wei J. Wang L. Jiang S. Xu L. (2020). A two-stage bilateral ischemia-reperfusion injury-induced AKI to CKD transition model in mice. Am. J. Physiol. Ren. Physiol. 319 (2 ), F304–f311. 10.1152/ajprenal.00017.2020
PMC009xxxxxx/PMC9738043.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36446561 10.1055/s-0042-1755457 210479 Original Article Sexual Violence Legal Pregnancy Interruption due to Sexual Violence in a Public Hospital in the South of Brazil Interrupção legal da gestação decorrente de violência sexual em um hospital público do Sul do Brasilhttp://orcid.org/0000-0002-5648-0598 Trapani Vitoria Finger 1 http://orcid.org/0000-0002-7561-4790 Feuerschuette Otto Henrique May 1 http://orcid.org/0000-0003-0196-4488 Júnior Alberto Trapani 2 1 Universidade do Sul de Santa Catarina, Tubarão, SC, Brazil 2 Universidade Federal de Santa Catarina, Florianópolis, SC, Brazil Address for correspondence Alberto Trapani Júnior, PhD, MD Rua Esteves Júnior,458/802, CEP 88015-130. Florianópolis, [email protected] 29 11 2022 10 2022 1 11 2022 44 10 945952 22 12 2021 03 6 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  To analyze the cases of all women who attend to a service of legal termination of pregnancy in cases of sexual violence in a public reference hospital and to identify the factors related to its execution. Methods  Cross-sectional observational study with information from medical records from January 2014 to December 2020. A total of 178 cases were included, with an evaluation of the data referring to the women who attended due to sexual violence, characteristics of sexual violence, hospital care, techniques used, and complications. The analysis was presented in relative and absolute frequencies, medians, means, and standard deviation. Factors related to the completion of the procedure were assessed using binary logistic regression. Results  Termination of pregnancy was performed in 83.2% of the cases; in 75.7% of the cases, the technique used was the association of transvaginal misoprostol and intrauterine manual aspiration. There were no deaths, and the rate of complications was 1.4%. Gestational age at the time the patient's sought assistance was the determining factor for the protocol not being completed. Pregnancies up to 12 weeks were associated with a lower chance of the interruption not occurring (odds ratio [OR]: 0.41; 95% confidence interval [CI]: 0.12–0.88), while cases with gestational age > 20 weeks were associated with a greater chance of the interruption not happening (OR: 29.93; 95%CI: 3.91–271.50). Conclusion  The service studied was effective, with gestational age being the significant factor for resolution. Resumo Objetivo  Analisar os casos de todas as mulheres que buscaram o serviço de interrupção legal da gestação em casos de violência sexual em um hospital público de referência, assim como identificar os fatores relacionados à realização do procedimento. Métodos  Estudo observacional transversal com informações de prontuários de janeiro de 2014 a dezembro de 2020. Foram incluídos 178 casos, com avaliação dos dados referentes às mulheres atendidas, características da violência sexual, atendimento hospitalar, técnicas utilizadas e complicações. A análise foi apresentada em frequência relativa e absoluta, medianas, médias e desvio padrão. Fatores relacionados à realização do procedimento foram avaliados por meio de regressão logística binária. Resultados  O procedimento de interrupção da gravidez foi realizado em 83,2% dos casos atendidos. Em 75,7% dos casos, a técnica utilizada foi a associação de misoprostol por via vaginal e aspiração manual intrauterina. Não houve óbitos e a taxa de complicações foi de 1,4%. A idade gestacional no momento da busca pela interrupção legal da gravidez foi o fator significativo pelo qual o protocolo não foi concluído. Gestações de até 12 semanas foram associadas a uma menor chance de não ocorrer a interrupção (odds ratio [OR]: 0,41; intervalo de confiança [IC] 95%: 0,12–0,88), enquanto casos com idade gestacional > 20 semanas foram associados a uma maior chance de não ocorrer a interrupção (OR: 29,93; IC95%: 3,91–271,50). Conclusão  O serviço estudado foi efetivo, sendo a idade gestacional o fator significativo para a resolutividade. Keywords abortion sexual violence legal abortion women's health health service Palavras-chave abortamento violência sexual aborto legal saúde da mulher serviço de saúde ==== Body pmcIntroduction Sexual violence is a humanitarian and public health issue, as the countless psychological and physical damage resulting from the sexual abuse remains with the victim after the crime. 1 The high incidence of cases of sexual violence brings to the health service patients who need assistance for possible physical and psychological damage and prophylaxis against sexually transmitted infections and pregnancy. 2 Furthermore, regarding psychological damage, it is common that the victims present anxiety and post-traumatic stress disorder, which interfere in both the individual and the collective spheres, affecting the interpersonal relationships and daily lives of the victims, also having an economic impact. 3 In 2009, a modification in the Brazilian Civil Code declared rape as a crime against sexual liberty and dignity. 4 In its article 213 (in the wording given by Law No. 12,015), rape is defined as: to constrain someone, through violence or serious threat, to have a carnal conjunction or to practice or allow another libidinous act to be performed with them. 5 Studies indicate that almost 90% of the cases of sexual violence go unreported. Deficiency, absence, or inadequate disclosure of the care network contributes to underreporting and health problems. 4 5 6 7 Among the consequences of sexual violence, pregnancy stands out due to the complexity of the psychological, social, and biological reactions it determines. Unwanted or forced pregnancy is seen as a second type of violence, intolerable for many women. 2 Women in a situation of pregnancy resulting from sexual violence, as well as adolescents and their guardians, must be informed about the legal alternatives regarding the possibilities of care in health services. It is the right of these women and adolescents to be informed of the possibility of a legal termination of pregnancy (LTP). 8 The Brazilian penal code of 1940, in its article 128, says that abortion caused by a doctor is not punishable if the pregnancy is due to rape and is preceded by the consent of the pregnant woman or of her guardian, in the case of a minor or incapable person; however, the concept of abortion was not defined in terms of gestational age. 5 In the English language and in some countries the procedure is called LTP and not abortion, and the gestational age is not limited. Although Brazilian law has made it possible to voluntarily terminate pregnancy due to rape since the 1940s, access to health services has not been regulated for nearly 50 years. 9 In 1989, the Municipal Government of the city of São Paulo implemented the first service to assist women victims of sexual violence. 10 To have access to LTP, the woman should present a copy of the Police Report (PR) and the expert report of the Legal Medical Institute (LMI). 8 The national regulation for LTP took place in 1999, with the launching of the technical norm Prevention and Treatment of Diseases Resulting from Sexual Violence against Women and Adolescents, which stimulated and regulated the structuring of services. Updated in 2005 and in 2011, the norm exempted women from submitting a PR or LMI report; the written consent of the woman and the evaluation of an interdisciplinary team would be enough. 8 However, this access is not yet widespread and not well-known across the country. There are few services that offer the LTP resulting from sexual violence program. In the state of Santa Catarina, for example, only four public institutions are referenced by the State Health Department to have an active program. 11 12 Linked to this, the problem of abortions considered to be unsafe, when they are not performed in a medical appropriate environment or by a suitable professional, must be recognized. 13 According to the World Health Organization (WHO), ∼ 45% of abortions in the world are performed unsafely. 14 In Brazil, abortion complications are the fourth cause of maternal mortality. Between 2006 and 2019, according to the Mortality Information System (MIS), 1,059 women had abortion registered as the underlying cause of death. 15 We must also consider that many of the deaths due to abortion complications are underreported, mainly because they are performed in illegal clinics. 16 17 The care of a woman who wants to terminate a pregnancy that resulted from sexual violence, according to Brazilian law, does not depend on a judicial or police process. In the health institution, five terms are required: the free and informed consent term, the liability term signed by the victim and/or by her legal representative, the detailed report term of the woman and/or her legal representative, a technical opinion, signed by a physician, attesting the compatibility of the gestational age with the date of the sexual assault, and the term of approval of the procedure for termination of the pregnancy, signed by the multidisciplinary team. 8 18 The present study aimed to analyze the profile of the women who sought a public service for abortion in case of sexual violence, as well as their results, to help in the construction of policies and practices for health assistance to women. Methods This is a cross-sectional observational study; the data has been extracted from medical records of women who sought assistance for LTP resulting from sexual violence at the Hospital Universitário Polydoro Ernani de São Thiago of the Universidade Federal de Santa Catarina (HU-UFSC, in the Portuguese acronym), Florianópolis, state of Santa Catarina, Brazil, as a source of information, during the period from January 2014 to December 2020. The study was submitted and approved by the Research Ethics Committee (REC) of the Universidade do Sul de Santa Catarina (UNISUL, in the Portuguese acronym) under the number 4.193.711. All women who sought the LTP service due to sexual violence were included in the survey. The adopted protocol follows the definitions of the technical norms of the Brazilian Ministry of Health and defines as abortions pregnancies terminated up to 20 weeks. 18 Regarding the sexual violence suffered, the following data were collected: city and date of occurrence, use of emergency contraception, previous police report about the violence, medical care after the rape, type of aggressor, number of aggressors, type of intimidation, and place of violence. Regarding the LTP, the following data were collected: date of the visit to the hospital, previous attempt at unsafe abortion, search for a previous hospital service, complaint to the police authority at the time of the search for the LTP, time between the visit to the hospital and hospitalization, days of hospitalization, performance or not of the procedure, method used, and observed complications. Regarding information about the right to LTP in case of rape, it was asked how the patient knew about the right to interrupt the pregnancy and the source regarding the referral service. Data were entered into Microsoft Excel (Microsoft Corporation, Redmond, WA, USA) and exported to SPSS PASW Statistics for Windows 18.0 (SPSS Inc., Chicago, IL, USA). They were analyzed and described as relative and absolute frequencies or by measures of central tendency and data dispersion (medians, means, and standard deviations [SDs]). In a second step, the difference between the dependent variable of performing the interruption procedure or not and the independent variables was tested through the association measure of the odds ratio (OR) with the respective 95% confidence intervals (95%CIs), through binary logistic regression. The level of significance established was p  < 0.05. Results From January 2014 to December 2020, 178 women who sought the HU-UFSC requesting the procedure for LTP due to sexual violence were found. After stability in the initial 3 years of the study, between 2016 and 2019 there was an average increase of 62.3% of cases per year. In 2020, there was a 26.5% reduction in visits by the LTP group ( Fig. 1 ). Fig. 1 Study flowchart. Among the 178 cases in which the LTP protocol for sexual violence was initiated, the procedure was performed in 148 (83.2%). Among the 30 cases (16.8%) in which the termination of pregnancy did not occur, most were due to noncompliance with the protocol: in 20 cases, the gestational age was incompatible with the date of aggression, the gestational age was > 20 weeks in 6 cases, and in 4 cases the interruption was discontinued at the request of the patient ( Fig. 2 ). Fig. 2 Distribution of patients who sought the service for legal termination of pregnancy due to sexual violence, according to the year and the procedure being performed or not. In the history collected by the team, only one patient reported using an effective contraceptive method at the time of violence; in this case, a combined hormonal contraceptive ( Table 1 ). Table 1 Characteristics of women who sought the legal termination of pregnancy service Characteristics n % Age (years old)  ≤14 13 7.3  15 to 19 18 10.1  20 to 29 76 42.7  30 to 39 58 32.6  ≥ 40 13 7.3 Skin color  White 143 80.3  Nonwhite 35 19.7 Scholarity  Middle school 50 28.1  High school 81 45.5  University education 47 26.4 Occupation  Employed 113 63.5  Unemployed 20 11.2  Student 45 25.3 Partner  Married or stable union 16 9.0  No partner 162 91.0 Religion ( n  = 88)  Catholic 36 40.9  Spiritism 9 10.2  Evangelicalism 13 14.8  No religion 27 30.7  Other 3 3.4 Primiparous  Yes 101 56.7  No 77 43.3 Housing  Florianópolis county 123 69.1  Other cities 55 30.9 Gestational age  ≤12 weeks 146 82.0  12w1d to 16w0d 17 9.6  16w1d to 20w0d 9 5.0  > 20 weeks 6 3.4 Most women had never been pregnant (56.7%), and in 12 cases (6.7%) they were virgins on the occasion of the sexual assault ( Table 2 ). Table 2 Characteristics of the aggression reported by women who sought legal termination of pregnancy ( n  = 178) Characteristics n % Police report at the time of the assault  Yes 17 9.6  No 161 90.4 Medical assistance at the time of the assault  Yes 10 5.4  No 168 94.6 Aggressor  Unknown 90 50.7  Family 13 7.4  Partner 16 8.8  Ex-partner 25 14.2  Friends 4 2.0  Other acquainted 30 16.9 Number of aggressors  Unknown 43 24.1  1 134 75.3  > 1 1 0.6 Type of intimidation* (combined or not)  Physical 105 59.0  Pharmacological 62 34.8  Verbal 28 15.7  Cutting weapon 10 5.6  Firearm 10 5.6  Other 6 3.4 Location of the assault  Housing 98 55.1  Public street 44 24.7  Party 15 8.4  Unknown 17 9.5  Other 4 2.3 *More than one type of intimidation = 46 The average time between the visit to the hospital and admission for the interruption procedure ranged from the same day of the first visit to up to 28 days, with a median of 3 days and an average of 4.7 days (SD ± 3.2). The total length of hospitalization ranged from 1 to 14 days, with a median of 2 days and a mean of 2.0 days (SD ± 0.7) ( Table 3 ). Table 3 Characteristics of the medical care and procedure for legal termination of pregnancy ( n  = 178) Characteristics n % Who informed about the right to LTP ( n  = 54)  Other health institution 20 37.0  Friend/family 8 14.8  Internet 16 29.6  Police or judiciary authority 4 7.4  Social services 6 11.2 Who referenced the HU-UFSC ( n  = 98)  Other health institution 46 47.0  Friend/family 19 19.4  Internet 15 15.3  Police or judiciary authority 10 10.2  Social services 8 8.1 History of attempted unsafe abortion?  Yes 4 2.2  No 174 97.8 Previously searched another hospital institution?  Yes 34 19.1  No 144 80.9 Police report at the time of the LTP?  Yes 24 13.5  No 154 86.5 Outcome  Pregnancy terminated 148 83.2  Denied 26 14.6  Quitclaim 4 2.2 Method for the LTP  Misoprostol 4 2.7  Misoprostol + curettage 31 20.9  Misoprostol + MVA 112 75.7  Other 1 0.7 Complications  Yes (allergy and hemorrhage) 2 1.4  No 146 98.6 Abbreviations: LTP, legal termination of pregnancy; MVA, manual vacuum aspiration. Table 4 shows the logistic regression analysis between the main characteristics of the cases and the performance or not of the LTP procedure. A gestational age ≤ 12 weeks was the factor associated with a lower chance of the LTP procedure not being performed ( p  = 0.021). There was an important association between gestational age > 20 weeks and not performing the interruption ( p  < 0.001). The other study factors were not associated with the performance or not of the LTP. Table 4 Logistic regression analysis between the characteristics of the cases and the performance or not of the procedure among women who sought the service for legal termination of pregnancy due to sexual violence Characteristics Performed* Not performed** Total Adjusted OR (95%CI) p -value n (%) n (%) n (%) Age (years old)  ≤ 19 26 (83.9) 5 (16.1) 31 (17.4) 0.97 (0.39–2.72) 0.934  20 to 29 62 (81.6) 14 (18.4) 76 (42.7) 1.28 (0.58–2.91) 0.681  ≥ 30 60 (84.5) 11 (15.5) 71 (39.9) 0.88 (0.40–1.97) 0.699 Scholarity  Middle school 43 (86.0) 7 (14.0) 50 (28.1) 0.81 (0.41–1.89) 0.581  High school 65 (80.2) 16 (19.8) 81 (45.5) 1.51 (0.71–3.55) 0.376  University education 40 (85.1) 7 (14.9) 47 (26.4) 0.87 (0.39–2.35) 0.691 Occupation  Employed 98 (86.7) 15 (13.3) 113 (63.5) 0.62 (0.31–1.28) 0.133  Unemployed 18 (90.0) 2 (10.0) 20 (11.2) 0.55 (0.29–2.71) 0.451  Student 32 (71.1) 13 (28.9) 45 (25.3) 2.10 (0.81–6.91) 0.068 Gestational age  ≤ 12 weeks 126 (85.1) 20 (66.6) 146 (82.0) 0.41 (0.12–0.88) 0.021  > 12 to ≤ 20 weeks 21 (14.2) 5 (16.7) 26 (14.6) 1.21 (0.42–3.51) 0.771  > 20 weeks 1 (0.7) 5 (16.7) 6 (3.4) 29.93 (3.91–271.5) <0.001 Skin color, nonwhite 2.01 (0.93–5.89) 0.075  Yes 25 (71.4) 10 (28.6) 35(19.7)  No 123 (86.0) 20 (14.0) 143 (80.3) Married or stable union 0.71 (0.23–3.21) 0.701  Yes 14 (87.5) 2 (12.5) 16 (9.0)  No 134 (82.7) 28 (17.3) 162 (91.0) Reports having a religion ( n  = 88) 3.41 (0.80–17.21) 0.190  Yes 48 (78.7) 13 (21.3) 61 (69.3)  No 25 (92.6) 2 (7.4) 27 (30.7) Primiparous 0.89 (0.42–1.92) 0.714  Yes 85 (84.2) 16 (15.8) 101 (56.7)  No 63 (81.8) 14 (18.2) 77 (43.3) Metropolitan area 1.60 (0.69–4.21) 0.390  Yes 100 (81.3) 23 (18.7) 123 (69.1)  No 48 (87.3) 7 (12.7) 55 (30.9) Known aggressor 0.91 (0.47–1.99) 0.781  Yes 74 (84.1) 14 (15.9) 88 (49.3)  No 74 (82.2) 16 (17.8) 90 (50.7) Abbreviations: CI, confidence interval; OR, odds ratio. *Performed: sought the service of legal termination of pregnancy and the procedure was performed **Not performed: sought the service of legal termination of pregnancy and the procedure was not performed Discussion In the hospital evaluated by the present research, there was a progressive increase in demand for LTP due to sexual violence, notably between 2016 and 2019. This upward trend was also observed in a study performed in the city of Rio de Janeiro, state of Rio de Janeiro, Brazil, during the same period. 19 Data available from the Brazilian Unified Health System Informatics Department (DATASUS) show that, in the period from 2014 to 2019, 7,699 cases of sexual violence were reported in the state of Santa Catarina, with a 70% increase in notifications over these years, which is in line with the growth curve in the present research. 20 The drop in demand for the service in 2020 may be a reflection of the SARS-Cov-2 pandemic, which induced an important change in behavior in the population. Considering that cases of sexual violence are underreported, a survey conducted in Santa Catarina showed that 6.4% of reported cases of sexual violence resulted in a pregnancy 21 and that there are few services in the state that carry out LTD resulting from sexual violence, 12 the number of cases studied may not match the probable existing demand. The disagreement between these data may be related to the lack of knowledge about women's rights to a LTP. Many women learn about the right to undergo an LTD and the institutions that offer it after discovering the pregnancy and looking for a health service. 11 In the present study, most women only learned about their rights and about the referral service through an institution, which was contacted after discovering the pregnancy. The present study showed that most women declared themselves white, without a partner, primiparous, and in the 1 st trimester of pregnancy, a profile similar to those observed in other studies. 22 23 24 The prevalent age group was 20 to 29 years old, which is in agreement with a study performed in the city of Porto Alegre, state of Rio Grande do Sul, Brazil. 22 Among the cases studied, the highest prevalence was of women who worked and had completed high school. This result was also found in a study in the city of Campinas, state of São Paulo, Brazil. 11 Unlike the present study and the state profile, 21 at the national level, incomplete middle school prevails for victims of sexual violence. 3 4 This difference may be related to the population characteristics of the state of Santa Catarina and to the possibility that women with a higher level of education have more knowledge about their rights. As for the bond between the victim and the aggressor (50.7% unknown) and the number of aggressors (75.3% with 1), the results obtained corroborate other similar studies. 6 22 23 In 46 cases, > 1 type of intimidation was reported, with a predominance of physical force, in 105 cases (59%). This characteristic corresponds to a national analysis of sexual violence; 4 however, it contrasts with a study performed in Campinas, in which verbal intimidation predominated (35.8%) over the use of physical force (29.1%); however, the present study did not specify the coexistence between types of intimidation. 23 In the present research, in 55.1% of the cases, sexual violence occurred in a residence, which is in line with data on cases of sexual violence in the state of Santa Catarina. 21 At the national level, the prevalence is of sexual violence by an unknown aggressor in public streets. 4 This disagreement may be a consequence of several regional differences in Brazil. Of the 178 cases analyzed, only 17 (9.6%) filed a police complaint after sexual violence and 10 (5.4%) sought medical care within the first 72 hours. Studies show that in only 10% of rape cases police complaints are filed. 25 This highlights the underreporting and hesitation of women to expose themselves after sexual violence and strengthens the need for an adequate reception in health services by a team capable of listening without judgment. The report of women who sought other hospitals for LTP before being referred to a reference service (19.1% in the present study) shows the need to expand the services that offer this type of care, both at the state and national level. 26 Of the cases analyzed, 83.2% had termination of pregnancy as an outcome, in agreement with a similar study performed in Porto Alegre, where this number was 90.5%. 22 In a nationwide study of legal abortion, < 50% of women completed the LTP. 26 However, the present study covers services with very different capacities, protocols, and local realities. It also does not clarify the reasons for not performing the procedure. In Brazilian legislation, there is lack of definition regarding a gestational age limit for LTP resulting from sexual violence. The most recent document that regulates the procedure fails to define a gestational age limit for abortion. 18 Based on Resolution No. 1779/2005 of the Federal Council of Medicine and on norms of the Brazilian Ministry of Health, the HU-UFSC defined as criteria for authorizing the procedure a gestational age < 20 weeks and the compatibility between the date of the occurrence of violence and the gestational age. 8 18 The service analyzed in the present study has a specific multidisciplinary team for LTP, following a well-defined protocol, which ensures performance in accordance with technical standards and reduces the possibility of conscientious objection. This structure is not found in 95% of the services registered to perform the LTP resulting from sexual violence, where care is provided by on-call professionals. 26 Training of the team in welcoming victims is also essential. The time between looking for the service and hospitalization for the procedure (median of 3 days) and the length of hospitalization (median of 2 days) demonstrate the effectiveness of the service and of the techniques used to terminate the pregnancy. Regarding the method used for the procedure, WHO recommendation is the association of misoprostol with manual vacuum aspiration (MVA), 27 which was prevalent in the current research with 75.7% of cases. This method also had the highest incidence in a similar study in Porto Alegre/RS with 49.1% of cases. 22 However, MVA is a method of uterine emptying that is not present in part of the services in Brazil that perform LTP. 26 Regarding attempted unsafe abortion, only 2.2% said they had tried it. However, it is noteworthy that this data does not portray the real number of women who tried unsafe abortion, as it only demonstrates those who spontaneously expressed it, as this question was not specifically included in the questionnaires of the service. In Porto Alegre, it was observed that 20.6% of the women stated that they had attempted an unsafe abortion. 22 The practice of unsafe abortion, which is not performed in a medical appropriate environment or by a suitable professional, in Brazil, has an estimated occurrence of > 1 million of these procedures per year. 8 The disagreement between the reported attempts and the actual number can be considered when analyzing that even with the low incidence in studies, postabortion curettage is the third most performed obstetric procedure in public health services in Brazil. 8 There were no maternal deaths in the present study, and in only 2 cases (1.4%) there was a reference to the occurrence of complications, 1 being allergic reaction, and the other, hemorrhaging. These results are compatible with those of a study performed in Campinas, where the rate of complications was 2.3%. 23 Unsafe abortions have a high rate of complications, reaffirming the importance of hospital services that adequately offer the procedure, as well as adequate disclosure of women's rights when faced with a pregnancy with the possibility of provided interruption foreseen by law. 28 29 Gestational age at the time of seeking LTP was the significant factor for not completing the protocol. Pregnancies ≤ 12 weeks were associated with a lower chance of interruption not occurring, whereas cases with a gestational age > 20 weeks were associated with a greater chance of not having interruption. Early search for an LTP service proved to be a protective factor against refusal of the procedure, possibly because the longer the time between the sexual violence and the search for a health service, the greater the probability of an error in the information. Gestational age > 20 weeks places the case outside the protocol of the service and contributes to the denial of the procedure. The reason and the refusal rate for the procedure are not demonstrated and discussed in the available literature. National policies involving both women victims of sexual violence and LTP resulting from it have evolved and have been regularized in recent decades. However, despite the legislation and technical norms, access to this right still requires certain improvements. The definition of a protocol and a specialized team for LTP care can offer an effective and resolving service with a minimum risk of complications. The lack of hospitals that offer this service in Brazil and of those that have an adequate staff and protocol is a limitation both for accessing this service and for conducting research on the procedure. The present research has as limitations its retrospective design and the fact that it is limited to a single institution. Multicenter and prospective studies are needed for a better view of LTP in Brazil. Conclusion The present study has demonstrated the profile of women who seek LTP resulting from sexual violence, which was defined by the regional characteristics of the study. The data demonstrated the effectiveness of the service, with gestational age being the significant factor for resolution. Contributions Conflict of Interests The authors have no conflict of interests to declare. All authors contributed to the design of the study, were involved in the data collection and analysis and/or interpretation. Also, all authors contributed to the writing/substantive editing and review of the manuscript and approved the final draft of the manuscript. ==== Refs References 1 Lima M C Larocca L M Nascimento D J Abortamento legal após estupro: histórias reais, diálogos necessários Saúde Debate. 2019 43 121 417 28 10.1590/0103-1104201912110 2 Ministério da Saúde, Ministério da Justiça, Secretaria de Políticas para as Mulheres. Norma técnica: atenção humanizada às pessoas em situação de violência sexual com registro de informações e coleta de vestígios Brasília (DF) MS/MJ/SPM 2015 3 Gaspar R S Pereira M UL [Trends in reporting of sexual violence in Brazil from 2009 to 2013] Cad Saude Publica 2018 34 11 e00172617 30427416 4 Cerqueira D Coelho D S . [Rape in Brazil: a radiograph according to health data (preliminary version)] [Internet]. Rio de Janeiro: IPEA; 2014 [cited 2021 Nov 10] (Technical Note; 11). Portuguese. Available from:http://www.ipea.gov.br/portal/index.php?option=com_content&view=article&id=21842 5 Dip R H A vida dos direitos humanos: bioética médica e jurídicaIn: Penteado JC. A vida dos direitos humanos: bioética médica e jurídica. Porto Alegre: SAFE; 1999. p. 355–402. 6 Oshikata C T Bedone A J Faúndes A Atendimento de emergência a mulheres que sofreram violência sexual: características das mulheres e resultados até seis meses pós-agressão Cad Saude Publica 2005 21 01 192 199 10.1590/S0102-311X2005000100021 15692652 7 Cividanes G C Mello A F Mello M F Revictimization as a high-risk factor for development of posttraumatic stress disorder: a systematic review of the literature Br J Psychiatry 2019 41 01 82 89 10.1590/1516-4446-2017-0013 8 Ministério da Saúde Secretaria de Atenção à Saúde. Departamento de Ações Programáticas EstratégicasAtenção humanizada ao abortamento: norma técnica. 2a ed.Brasília (DF) Ministério da Saúde 2011 9 Ministério da Saúde Secretaria de Atenção à Saúde. Departamento de Ações Programáticas Estratégicas[Prevention and treatment of the injuries resulted from sexual violence against women and teenagers] [Internet]. 3rd ed. Brasília (DF): Ministério da Saúde; 2012 [cited 2021 Sep 12]. Portuguese. Available from:https://bvsms.saude.gov.br/bvs/publicacoes/prevencao_agravo_violencia_sexual_mulheres_3ed.pdf 10 Cabral C DS Barbosa R M Arilha M The right to abortion: pathways in Brazil - an interview with Margareth Arilha Cad Saude Publica 2020 36 (36, Suppl 1):e00118319 10.1590/0102-311X00118319 11 Machado C L Fernandes A M Osis M J Makuch M Y Gravidez após violência sexual: vivências de mulheres em busca da interrupção legal Cad Saude Publica 2015 31 02 345 353 10.1590/0102-311X00051714 25760168 12 Delgado J S Serviço social e interrupção legal da gestação decorrente de situações de violência sexual [tcc] Florianópolis Universidade Fedederal de Santa Catarina 2017 13 Cisne M Castro V V Oliveira G M Unsafe abortion: a patriarchal and racialized portrayal of women's poverty Rev Katálysis. 2018 21 03 452 70 10.1590/1982-02592018v21n3p452 14 Cardoso B B Vieira F MDSB Saraceni V Abortion in Brazil: what do the official data say? Cad Saude Publica 2020 36 (36, Suppl 1)e00188718 10.1590/01002-311X00188718 15 Ministério da Saúde. Secretaria de Vigilância em SaúdeDepartamento de Análise em Saúde e Vigilância das Doenças Não Transmissíveis. [Maternal mortality monitoring panel] [Internet]. 2020 [cited 2020 May 13]. Portuguese. Available from:http://svs.aids.gov.br/dantps/centrais-de-conteudos/paineis-de-monitoramento/mortalidade/materna/ 16 Ministério da Saúde. Secretaria de Vigilância em Saúde. Departamento de Análise de Situação em Saúde. [Maternal death epidemiological surveillance guide]. Brasília (DF): Ministério da Saúde; 2009. Portuguese. 17 Galli B Challenges and opportunities for access to legal and safe abortion in Latin America based on the scenarios in Brazil, Argentina, and Uruguay Cad Saude Publica 2020 36 (36, Suppl 1)e00168419 10.1590/0102-311X00168419 18 Ministério da Saúde. [Decree No. 2.561, from September 23, 2020. Provides for the Procedure for Justification and Authorization of Interruption of Pregnancy in the cases provided for by law, within the scope of the Unified Health System-SUS]. Diário Oficial da União. 2020 Sep 24;Sec 1:89. Portuguese. 19 Mudjalieb A A Report of an experience to expand access to legal abortion for rape victims in the city of Rio de Janeiro, Brazil Cad Saude Publica 2020 36 (36, Suppl 1):e00181219 10.1590/0102-311X00181219 20 Ministério da Saúde DATASUS. SINAN[Domestic, sexual and other violences - 2014 a 2019. Sexual violence] [Internet]. 2020 [cited 2021 May 4]. Available from: http://tabnet.datasus.gov.br/cgi/tabcgi.exe?sinannet/violencia/bases/violebrnet.def. Portuguese. 21 Delziovo C R Bolsoni C C Nazário N O Coelho E BS [Characteristics of sexual violence against adolescent and adult women reported by the public health services in Santa Catarina State, Brazil] Cad Saude Publica 2017 33 06 e00002716 10.1590/0102-311X00002716 28724022 22 Patuzzi G C Aborto legal em gestações decorrentes de estupro: como as mulheres descobrem o seu direito?[monografia].Porto Alegre Grupo Hospitalar Conceição 2019 23 Mutta D S Yela D A Sociodemographic characteristics of women who underwent legal abortion due to sexual violence in a public hospital in Campinas: a cross-sectional study Sao Paulo Med J 2017 135 04 363 368 10.1590/1516-3180.2017.0048150317 28767994 24 Mattar R Abrahão A R Andalaft Neto J Colas O R Schroeder I Machado S JR Assistência multiprofissional à vítima de violência sexual: a experiência da Universidade Federal de São Paulo Cad Saude Publica 2007 23 02 459 464 10.1590/S0102-311X2007000200023 17221096 25 Sudário S Almeida P C Jorge M S Mulheres vítimas de estrupo: contexto e enfrentamento dessa realidade Psicol Soc. 2005 17 03 80 6 10.1590/S0102-71822005000300012 26 Madeiro A P Diniz D Legal abortion services in Brazil–a national study Cien Saude Colet 2016 21 02 563 572 10.1590/1413-81232015212.10352015 26910163 27 World Health Organization Medical management of abortion Geneva World Health Organization 2018 28 Singh S Maddow-Zimet I Facility-based treatment for medical complications resulting from unsafe pregnancy termination in the developing world, 2012: a review of evidence from 26 countries BJOG 2016 123 09 1489 1498 10.1111/1471-0528.13552 26287503 29 Domingues R MSM Fonseca S C Leal M DC Aquino E ML Menezes G MS Unsafe abortion in Brazil: a systematic review of the scientific production, 2008-2018 Cad Saude Publica 2020 36 (36, Suppl 1)e00190418 10.1590/0102-311X00190418
PMC009xxxxxx/PMC9800061.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36580952 10.1055/s-0042-1757957 RBGO-22-0273 Letter to the Editor Adverse Events Related to COVID-19 Vaccines in Pregnant Women: Correspondence http://orcid.org/0000-0002-8859-5322 Sookaromdee Pathum 1 http://orcid.org/0000-0003-1039-3728 Wiwanitkit Viroj 2 1 Private Academic Consultant, Bangkok, Thailand 2 Honorary Professor, Dr. DY Patil University, Pune, Maharashtra, India Address for correspondence Pathum Sookaromdee, PhD, Private Academic Consultant [email protected] 29 12 2022 11 2022 1 12 2022 44 11 10781078 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. ==== Body pmc Authors' reply Dear Editor, We would like to share ideas on the publication “Adverse Events Related to COVID-19 Vaccines Reported in Pregnant Women in Brazil.” 1 According to Kobayashi et al., 1 ten deaths were identified, one of which was thought to be causally related to the vaccine, and the other nine maternal deaths had causality C, that is, no causal relationship with the vaccine, and the majority were due to complications inherent in pregnancy, such as pregnancy-specific hypertensive disorder (PSHD) in four cases, and COVID-19 in three. 1 The problem of the efficacy and safety of the COVID-19 vaccine for pregnant women is intriguing, and evidence from real-world situations are needed. 2 Aside from death, there are further concerns about the safety of the COVID-19 vaccine for pregnant women, including the induction of abortion. 3 We are all concerned that, despite its benefits, the COVID-19 vaccine may be harmful. It is difficult to determine the exact source of the clinical issue in this case due to a lack of information on the health and immunological status of vaccination recipients prior to inoculation. People may reject vaccines and lose faith in them if they are given contradicting information. A clinical comorbidity may be at the basis of the problem. 4 5 The assessment of potential detrimental outcomes is now too early due to incomplete documentation of a person's health or immune condition prior to immunization. Conflict of Interests The authors have no conflict of interests to declare. ==== Refs References 1 Kobayashi C D Porto V BG da Nóbrega M EB Cabral C M Barros T D Martins C MR Adverse events related to COVID-19 vaccines reported in pregnant women in Brazil Rev Bras Ginecol Obstet 2022 ••• : [ ahead of print ]10.1055/s-0042-1755461 2 Joob B Wiwanitkit V Vaccine for Covid-19 and pregnant women Rev Bras Ginecol Obstet 2021 43 06 490 491 10.1055/s-0041-1731382 34318475 3 Sookaromdee P Wiwanitkit V Magnitude of abortion after COVID-19 vaccination: how about rate? Erciyes Med J 2022 44 02 244 10.14744/etd.2021.78379 4 Kebayoon A Wiwanitkit V Dengue after COVID-19 vaccination: possible and might be missed Clin Appl Thromb Hemost 2021 27 10760296211047229. Doi: 10.1177/10760296211047229 5 Sookaromdee P Wiwanitkit V Severe immune thrombocytopenia following COVID-19 vaccination Am J Emerg Med 2022 58 337 10.1016/j.ajem.2022.03.055 35437199
PMC009xxxxxx/PMC9800062.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36580946 10.1055/s-0042-1757954 RBGO-21-0359 Original Article Obstetrics Experience of a Tertiary Service in the Treatment of Women with Cervical Pregnancy Experiência de um serviço terciário no tratamento de mulheres com gravidez ectópica cervicalhttp://orcid.org/0000-0002-4950-6331 Mori Karen Hiromi 1 http://orcid.org/0000-0003-4060-8515 Tavares Bárbara Virgínia 1 http://orcid.org/0000-0003-3889-4778 Yela Daniela Angerame 1 http://orcid.org/0000-0002-8837-8061 Baccaro Luis Francisco Cintra 1 http://orcid.org/0000-0003-3197-1195 Juliato Cassia Raquel Teatin 1 1 Universidade Estadual de Campinas, Campinas, SP, Brazil Address for correspondence Cássia Raquel Teatin Juliato Rua Alexander Fleming, 101, Cidade Universitária Zeferino Vaz, Campinas, SP, [email protected] 29 12 2022 11 2022 1 12 2022 44 11 10141020 17 9 2021 10 8 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  Cervical pregnancy is challenging for the medical community, as it is potentially fatal. The treatment can be medical or surgical; however, there are no protocols that establish the best option for each case. The objective of the present study was to describe the cases of cervical pregnancy admitted to a tertiary university hospital over a period of 18 years. Methods  A retrospective study based on a review of the medical records of all cervical pregnancies admitted to the Women's Hospital at Universidade Estadual de Campinas, Southeastern Brazil, from 2000 to 2018. Results  We identified 13 cases of cervical pregnancy out of a total of 673 ectopic pregnancies; only 1 case was initially treated with surgery because of hemodynamic instability. Of the 12 cases treated conservatively, 7 were treated with single-dose intramuscular methotrexate, 1, with intravenous and intramuscular methotrexate, 1, with intravenous methotrexate, 1, with 2 doses of intramuscular methotrexate, and 2, with intra-amniotic methotrexate. Of these cases, one had a therapeutic failure that required a hysterectomy. Two women received blood transfusions. Four women required cervical tamponade with a Foley catheter balloon for hemostasis. There was no fatal outcome. Conclusion  Cervical pregnancy is a rare and challenging condition from diagnosis to treatment. Conservative treatment was the primary method of therapy used, with satisfactory results. In cases of increased bleeding, cervical curettage was the initial treatment, and it was associated with the use of a cervical balloon for hemostasis. Resumo Objetivo  A gravidez ectópica cervical é um desafio para a comunidade médica, pois pode ser fatal. O tratamento pode ser clínico ou cirúrgico, mas não existem protocolos que estabeleçam a melhor opção para cada caso. O objetivo deste estudo foi descrever os casos de gravidez ectópica cervical internados em um hospital universitário terciário durante 18 anos. Métodos  Estudo retrospectivo com revisão de prontuários de todas as gestações ectópicas cervicais internadas no Hospital da Mulher da Universidade Estadual de Campinas de 2000 a 2018. Resultados  Foram identificados treze casos de gestação ectópica cervical em um total de 673 gestações ectópicas; apenas 1 caso foi inicialmente tratado com cirurgia por causa de instabilidade hemodinâmica. Dos 12 casos tratados conservadoramente, 7 foram tratados com metotrexato por via intramuscular em dose única, 1, com metotrexato pelas vias intravenosa e intramuscular, 1, com metotrexato por via intravenosa, 1, com 2 doses de metotrexato por via intramuscular, e 2, com metotrexato por via intra-amniótica. Desses casos, um apresentou falha terapêutica, e realizou-se uma histerectomia. Duas mulheres receberam transfusões de sangue. Quatro mulheres necessitaram de tamponamento cervical com cateter balão de Foley para hemostasia. Não houve casos fatais. Conclusão  A gravidez cervical é uma condição rara e desafiadora desde o diagnóstico até o tratamento. O tratamento conservador foi o principal método terapêutico utilizado, com resultados satisfatórios. Nos casos de sangramento aumentado, a curetagem cervical foi o tratamento inicial, e foi associada ao uso de balão cervical para hemostasia. Keywords cervical pregnancy ectopic pregnancy methotrexate surgery treatment Palavras-chave gravidez cervical gravidez ectópica metotrexato cirurgia tratamento Funding/Acknowledgments The authors would like to thank Coordenação de Aperfeiçoamento de Pessoal de Nível Superior – Brasil (CAPES) for the master's scholarship to one of the co-authors of the present study (B.V.T). The study was financed in part by CAPES (finance code 001). ==== Body pmcIntroduction Cervical ectopic pregnancy is defined as the implantation of the embryo into the endocervical canal. The pathophysiology of the condition is not completely understood. Several studies 1 have suggested that surgical procedures such as dilatation and curettage, in addition to cesarean sections, are associated with a higher occurrence of this condition. Assisted reproduction procedures have also been associated with a higher frequency of cervical ectopic pregnancies. 2 3 4 5 6 The condition is rare, with an estimated incidence of 0.0001%, that is, 1 in every 10 thousand women will have a cervical pregnancy. 3 The most common symptoms of cervical pregnancy are vaginal bleeding and pain. However, the trophoblastic invasion that occurs in the epithelium and endocervical fibromuscular stroma can lead to rupture of the local vessels, resulting in profuse hemorrhage. If a cervical pregnancy is not promptly treated, there is a high risk of severe hemorrhage, which can lead to maternal morbidity and even death. 7 From 2000 to 2019, it is estimated that 1.9% of maternal deaths in the state of São Paulo (the most developed region in Brazil) occurred due to complications of cervical ectopic pregnancy. 8 Thus, the appropriate treatment for ectopic pregnancy is a challenge for the medical community. As it is a rare disease, there is no well-established protocol for its management. 9 Previously, cervical pregnancy was treated surgically, primarily by hysterectomy due to the high risk of severe hemorrhage. In recent years, most reports 10 11 have demonstrated the high efficacy and safety of the conservative medical treatment for cervical pregnancy with systemic or local methotrexate (MTX), uterine artery embolization, or dilation and curettage (D&C) with Foley catheter balloon tamponade. However, to date, no consensus has been reached among clinicians regarding the standard care for cervical ectopic pregnancy. The aim of the present study was to describe our experience with 13 cases of cervical pregnancy at a tertiary teaching hospital in Southeastern Brazil throughout 18 years. Methods The present is a retrospective study that involved the evaluation of medical records of women with cervical pregnancies admitted to a tertiary hospital (Centro de Atenção Integral à Saúde da Mulher [CAISM], the Women's Hospital at Universidade Estadual de Campinas) from 2000 to 2018. We identified the cases in the hospital records based on the International Classification of Diseases. The study was approved by the Research Ethics Committee under a Certificate of Presentation for Ethical Appreciation (number 53019116.6.0000.5404). The research followed the principles of the Declaration of Helsinki. A diagnosis of cervical ectopic pregnancy was made through transvaginal ultrasound and involved the identification of a gestational sac in the uterine cervix and the absence of intrauterine pregnancy. In cervical ectopic pregnancies with an embryo, the presence of cardiac activity in the embryo was evaluated, in addition to other ultrasound descriptors such as the shape of the uterus (if it was described as having an hourglass shape with a ballooned cervical canal). 12 We obtained the clinical and obstetric history by reviewing the medical records. The gestational age was calculated based on the date of the last menstrual period or was estimated by ultrasound (US) when necessary. The following factors associated with cervical ectopic pregnancy were evaluated: gestational age, ultrasound diagnosis, the presence of fetal cardiac activity, treatment methods, and success rate. The treatment (either medical or surgical) was considered successful if a negative beta-human chorionic gonadotropin (β-hCG) result was obtained. The quantitative variables were presented as means and standard deviations. The categorical variables were presented as absolute numbers and relative frequencies. Results We identified 673 women with ectopic pregnancies, and 13 (1.9%) presented ultrasonography criteria for a cervical ectopic pregnancy ( Fig. 1 ). Fig. 1 Ultrasonography aspects of cervical ectopic pregnancy. The mean age of the women with a cervical ectopic pregnancy was of 31.8 (±5.9) years, and the number of previous pregnancies was of 2.1 (±1.4). Approximately half of the women (7/13; 53.8%) had a history of cesarean section, with an average of 0.8 (±0.9). A history of uterine curettage was present in 2 of them, with an average of 0.3 (±0.8). None of the women used an intrauterine device as a contraceptive method or reported that they smoked ( Chart 1 ). Chart 1 Characteristics and treatment of the cases of cervical pregnancy N O Age (years) Parity index GA Initial diagnosis Ultrasonography findings β-hCG (IU/l) Initial treatment Additional treatment 1 39 G2C1 6 + 2 Anembryonic pregnancy 4-cm cervical gestational sac 129,57 Methotrexate IM None 2 26 G6C2A3 6 + 5 Cervical ectopic pregnancy Hematometra 79 × 47 × 62 mm, cervical gestation, and invasion of decidua into the myometrium, FHB present 88,878 Cervical curettage, vacuum aspiration, Foley catheter balloon None 3 30 G3C1A1 10 + 5 Cervical ectopic pregnancy Isthmic pregnancy, CRL 38 mm, FHB present 75,810 AF aspiration, intracardiac injection of xylocaine, and intra-amniotic injection of methotrexate 50 mg None 4 32 G6P5C3 9 + 5 Cervical ectopic pregnancy in abortion Cervical ectopic pregnancy in abortion: heterogeneous oval image, 5.7 × 8.4 mm 4,585 Methotrexate EV and methotrexate IM None 5 28 G4P2A1 8 + 0 Embryonic death in the cervical region Gestational sac 25 × 19 × 14 mm, with 12-mm embryo in the cervical region, without FHB 3,678 Methotrexate IM None 6 35 G3P1A1 9 + 3 Cervical ectopic pregnancy 28-mm gestational sac with an embryo, 11 mm distant from the IO, FHB present ≥ 10,00 AF aspiration, intracardiac injection of xylocaine and intra-amniotic injection of methotrexate 87.5 mg None 7 31 G3C2 6 + 1 Cervical ectopic pregnancy Cervical ectopic pregnancy of GA 6 + 2, with live embryo ≥ 10,000 Methotrexate IM Cervical curettage 8 25 G3C1A1 10 + 2 Ectopic pregnancy in Cesarean section scar Cervical region just below cesarean section scar – an image of 23 × 27 × 33 mm may correspond to cervical ectopic pregnancy 642 Methotrexate IM None 9 21 G3P1A1 9 + 0 Incomplete abortion in the cervical canal Heterogeneous nodular formation in the cervical canal with extension to the serosa measuring 42 × 38 × 35mm, with gestational sac without embryo 6,598 Methotrexate IM Curettage and Foley catheter balloon 10 37 G3P2C1 9 + 0 Cervical ectopic pregnancy Cervical ectopic pregnancy with the presence of the gestational sac only 10,185 Methotrexate IM None 11 44 G1P0 7 + 1 Cervical ectopic pregnancy Pregnancy in the cervical-uterine body transition, with embryo and FHB present 78,839 Methotrexate EV Curettage, Foley catheter balloon, hysterectomy 12 33 G2A1 5 + 6 Cervical ectopic pregnancy Anechoic image in cervical region 21 × 13 × 11mm, diameter: 15mm, 2.9-mm CRL embryo, FHB present 7,683 Methotrexate IM Methotrexate IM 13 33 G1 6 + 3 Cervical embryonic death Gestational sac 14 × 6 × 11 mm, with implantation on the cervix, with no sign of rupture, with Doppler flux, presence of an embryo 6,548 Methotrexate IM Cervical Foley catheter Abbreviations: A, abortion; AF, amniotic fluid; C, cesarean delivery; CRL, crown-rump length; EV, endovenous; FHB, fetal heartbeat; G, gravity; GA, gestational age; IM, intramuscular; IO, internal os; P, deliveries; β-hCG, beta-human chorionic gonadotropin. Among the 13 identified cases, the average gestational age was of 8.1 (±1.7) weeks. Of these women, 69.2% had vaginal bleeding on admission, and 2 (15.3%) had profuse bleeding. On admission, 4 women (30.7%) complained of pain in the lower abdomen. The mean β-hCG was of 2,3351.96 (±SD) IU/L. In the ultrasound evaluation, the embryos of 8 out of the 13 women (61.5%) could be visualized and the embryos of 6 of these had visible heartbeats. The medical management involved systemic or intra-amniotic MTX injection after an evaluation of the blood tests regarding liver and renal function and a complete blood count. The baseline serum β-hCG was recorded. The MTX injection was administered as follows: an intramuscular dose of 50 mg/m 2 , an intravenous dose of 1mg/kg, or an intra-amniotic dose of 1 mg/kg. The serum levels of β-hCG were assessed on days 4 and 7 after the injection, and a ≥ 15% drop between these days was considered a response to the medical treatment. Subsequently, the levels of β-hCG were monitored weekly until they dropped below 5 mIU/mL, when they were considered negative. Regarding the treatment, a flowchart of the identified cases is presented in Fig. 1 . Of the 13 cases, 1 required immediate surgical treatment after admission due to heavy bleeding and hemodynamic instability. In this case, the presence of an embryo with a fetal heartbeat was identified during the ultrasound. The patient was treated with cervical canal curettage, vacuum aspiration, and insertion of a Foley catheter balloon with 30 mL of water for hemostasis. The patient also received a blood transfusion. The 12 remaining hemodynamically stable cases were divided into 2 groups based on the ultrasound results: 7 women with an embryo and 5 women without an embryo ( Fig. 2 ). Fig. 2 Flowchart of the treatment for cervical ectopic pregnancy. Abbreviations: AF, amniotic fluid; EV, endovenous; IM, intramuscular; MTX, methotrexate; n, number. Cervical Pregnancy without an Embryo during an Ultrasound Evaluation Among the 5 women without embryos, 3 were successfully treated with a single administration of intramuscular MTX (50 mg/m 2 ). One woman received a double dose of MTX, one received an intramuscular dose, and another received an intravenous dose. The remaining patient was treated with a single intramuscular dose of MTX and also required complementary treatment with uterine curettage and homeostasis with a Foley catheter balloon. The patient returned eleven days after the MTX injection with severe abdominal pain and excessive vaginal bleeding. The bleeding was maintained despite the curettage, so the Foley balloon catheter was positioned with 60 mL of water. The balloon was removed after 24 hours, with no further complications. The patient required a blood transfusion after the curettage. Cervical Pregnancy with an Embryo but without a Heartbeat during an Ultrasound Evaluation The two women who had embryos without a heartbeat during the ultrasound evaluation were treated with a single dose of intramuscular MTX. In 1 of the women, the initial β-hCG level was of 3,678 UI/L. After four days of MTX administration, the patient developed vaginal bleeding, and the cervical pregnancy was aborted. The control ultrasound showed no evidence of ectopic pregnancy on the cervix. The β-hCG level on day 9 after the administration of MTX decreased to 10 UI/L. In the other patient, the initial β-hCG level was of 6,548 UI/L. On day 6 after the MTX injection, the woman presented with vaginal bleeding and the gestational sac was adhered to the cervix. An attempt to remove the sac was unsuccessful due to its strong adherence to the cervix. A Foley balloon catheter with 30 mL of water was the option of additional treatment. It was positioned, guided by an ultrasound, and was removed after 48 hours with no complications. The β-hCG level decreased from 30,188 UI/L on day 4 to 4,819 UI/L on day 7. Cervical Pregnancy with an Embryo with a Heartbeat during an Ultrasound Evaluation There were 5 cases of embryos with a heartbeat, and three different treatment protocols were used. Two of the cases were treated with an intramuscular MTX injection (50 mg/m 2 ). In these two cases, due to elevated serum levels of β-hCG, one woman underwent uterine curettage, and the other received a second dose of intramuscular MTX (50 mg/m 2 ). In this latter case, an ultrasound performed on day 4 after the MTX injection showed evidence of an embryo with a heartbeat. One of the 5 cases received a dose of intravenous MTX (1 mg/kg); however, due to therapeutic failure and an elevated β-hCG level, the patient required additional treatment and underwent curettage and hemostasis with a Foley catheter balloon. The condition progressed, and the patient suffered excessive bleeding; thus, a hysterectomy was performed. This was the only case that required a hysterectomy after failed hemostasis with a Foley catheter balloon. The other two cases required aspiration of 30 mL of amniotic fluid and intracardiac injection of xylocaine and intra-amniotic injection of MTX. One of the patients received 50 mg of intra-amniotic MTX after the heartbeat of the embryo stopped. The other case received 87.5 mg of intra-amniotic MTX after the heartbeat of the embryo stopped. Discussion The present study described 18 years of experience concerning the treatment of cervical ectopic pregnancy, and positive results were observed regarding the conservative treatment. However, after the medical treatment, there was a considerable requirement for complementary treatment. The results of the present study were positive because only one woman required a hysterectomy for a cervical pregnancy. A separate study 13 demonstrated that 70% of cervical pregnancies required a hysterectomy due to massive blood loss. Most of the women included in this study were hemodynamically stable. In these cases, medical treatment with MTX could be considered as a first-line treatment. Isolated curettage is associated with a 40% risk of requiring a hysterectomy. 14 Treatment with MTX was initially conducted in 1983 15 and, at present, the literature shows that the risk of bleeding is of 11% for drug treatment and of 3% in the case of a hysterectomy. 11 Recent studies 16 17 show that drug treatment can prevent a hysterectomy in up to 91% of cases. The conservative treatments include drug treatment (local or systemic treatments, or treatments with intracervical MTX, potassium chloride, vasopressin), and surgical procedures (local aspiration, curettage, cerclage with cervical tamponade), and the radical treatments include cervical amputation and hysterectomy. 11 In the present study, the treatment with MTX was initially only successful in 4 out of 8 women with a cervical pregnancy. However, an intramuscular single dose of this medication was used, and, in certain cases, it was used heterogeneously with the application of intravenous doses. Evidence of the use of MTX for the treatment of cervical pregnancy is limited to case reports. 18 19 20 21 Previous studies reported cases of successful treatment of cervical ectopic pregnancies with multiple doses of MTX, suggesting that it is an appropriate therapy for hemodynamically stable women. 22 23 A separate study 24 compared the effectiveness of the treatment with a single and with multiple doses of MTX, and the authors concluded that a single dose was as effective as multiple doses. Since the condition is rare and the number of cases was limited, in the present study, we could not conclude whether the failure of the initial treatment in our case series was solely due to the MTX regimen adopted. Although experience with cervical ectopic pregnancy is limited to conclusions concerning the criteria for medical therapy, it has been postulated that serum levels of β-hCG > 10,000 mUI/ mL, gestational age > 9 weeks, the presence of cardiac fetal activity, or a crown-rump length > 10 mm will result in unsuccessful treatment of the cervical ectopic pregnancy. 11 Similar to observations made regarding tubal pregnancy, the presence of fetal cardiac activity is a poor prognostic factor for successful MTX treatment. 25 However, in addition to this poor prognostic factor, the surgical risk of hemorrhage is higher in cervical pregnancies than in tubal pregnancies. In the present study, two women with fetal heartbeat on ultrasound were successfully treated with intra-amniotic MTX. This result corroborates those of a recent study 26 that demonstrated greater efficacy when intra-amniotic MTX was used after the systemic use failed. Previous studies 27 described the injection of potassium chloride or intracardiac ethanol in these cases, and the authors demonstrated an improvement in the effectiveness of the conservative treatment. However, the data are too limited to enable a comparison of the two regimes. 11 In the present study, there was a successful report of xylocaine injection associated with the withdrawal of amniotic fluid. The use of curettage and hemostasis with a Foley catheter tamponade was described in other studies 28 as an alternative and safe treatment to the use of MTX. One study 29 described a series of thirteen cases successfully treated using this method, and the author emphasized that it would eliminate the maternal MTX risks. However, in the cases examined in the present study, all the women who were stabilized received either an intramuscular, intravenous, or intra-amniotic injection of MTX. Moreover, similar to other studies, 28 curettage and a Foley catheter balloon were used as additional treatments, and occasionally it was necessary to reduce the massive hemorrhage caused by the MTX at the implantation site. The present study was conducted in a tertiary hospital, which is a reference regarding the treatment of ectopic pregnancies, as a considerable number of women with this condition have been treated there. Because cervical ectopic pregnancy is a rare condition, most of the works published in the literature are case reports. Therefore, the present was a retrospective study that included descriptions of various treatment modalities, and the results highlight the need for further studies and the establishment of protocols to assist women with cervical ectopic pregnancies. In conclusion, from diagnosis to treatment, cervical pregnancy is a rare and challenging condition. A single intramuscular injection of MTX was the main treatment performed on women with this condition, and satisfactory results were obtained. However, there was a frequent requirement for complementary treatment due to vaginal bleeding or failure of β-hCG levels to decrease. In cases of increased bleeding, surgical treatment with curettage was the initial treatment and, in three cases, additional treatment with a cervical Foley catheter balloon for hemostasis was administered. Therefore, conservative and fertility-sparing treatment was possible in most cases. Moreover, even though cervical pregnancy has been associated with considerable morbidity, no fatal outcomes were observed during the present study. Contribuitions Karen Hiromi Mori: writing – original draft; formal analysis; conceptualization; methodology; investigation; resources; writing – review and editing, project administration; Bárbara Virgínia Tavares: methodology; writing – review and editing; Daniela Angerami Yela: writing – review and editing; Luis Francisco Cintra Baccaro: writing – review and editing; Cassia Raquel Teatin Juliato: conceptualization; methodology; investigation; resources; writing – review and editing. Conflict of Interests The authors have no conflict of interests to declare. ==== Refs References 1 Ushakov F B Elchalal U Aceman P J Schenker J G Cervical pregnancy: past and future Obstet Gynecol Surv 1997 52 01 45 59 10.1097/00006254-199701000-00023 8994238 2 Modayil V Ash A Raio C Cervical ectopic pregnancy diagnosed by point-of-care emergency department ultrasound J Emerg Med 2011 41 06 655 657 10.1016/j.jemermed.2011.05.048 21820259 3 Kochi K Hidaka T Yasoshima K Yoneda K Arai K Arai T Cervical pregnancy: a report of four cases J Obstet Gynaecol Res 2014 40 02 603 606 10.1111/jog.12185 24118279 4 Rothe D J Birnbaum S J Cervical pregnancy: diagnosis and management Obstet Gynecol 1973 42 05 675 680 4749570 5 Shinagawa S Nagayama M Cervical pregnancy as a possible sequela of induced abortion. Report of 19 cases Am J Obstet Gynecol 1969 105 02 282 284 10.1016/0002-9378(69)90075-1 5810457 6 Hoyos L R Tamakuwala S Rambhatla A Brar H Vilchez J Allsworth J Risk factors for cervical ectopic pregnancy J Gynecol Obstet Hum Reprod 2019 ••• 101665; [ ahead of print ]10.1016/j.jogoh.2019.101665 7 Tinelli A Malvasi A Vergara D Casciaro S Emergency surgical procedure for failed methotrexate treatment of cervical pregnancy: a case report Eur J Contracept Reprod Health Care 2007 12 04 391 395 10.1080/13625180701502351 17853169 8 Ministério da Saúde . DATASUS. Óbitos de mulheres em idade fértil e óbitos maternos - São Paulo [Internet]. 2021; [cited 2021 Sep 10]. Available from:http://tabnet.datasus.gov.br/cgi/tabcgi.exe?sim/cnv/mat10sp.def 9 Shavell V I Abdallah M E Zakaria M A Berman J M Diamond M P Puscheck E E Misdiagnosis of cervical ectopic pregnancy Arch Gynecol Obstet 2012 285 02 423 426 10.1007/s00404-011-1980-0 21748311 10 Vela G Tulandi T Cervical pregnancy: the importance of early diagnosis and treatment J Minim Invasive Gynecol 2007 14 04 481 484 10.1016/j.jmig.2006.11.012 17630167 11 Hosni M M Herath R P Mumtaz R Diagnostic and therapeutic dilemmas of cervical ectopic pregnancy Obstet Gynecol Surv 2014 69 05 261 276 10.1097/OGX.0000000000000062 25101692 12 Kung F T Lin H Hsu T Y Chang C Y Huang H W Huang L Y Differential diagnosis of suspected cervical pregnancy and conservative treatment with the combination of laparoscopy-assisted uterine artery ligation and hysteroscopic endocervical resection Fertil Steril 2004 81 06 1642 1649 10.1016/j.fertnstert.2003.11.034 15193489 13 Hung T H Jeng C J Yang Y C Wang K G Lan C C Treatment of cervical pregnancy with methotrexate Int J Gynaecol Obstet 1996 53 03 243 247 10.1016/0020-7292(96)02653-7 8793626 14 Jurkovic D Hacket E Campbell S Diagnosis and treatment of early cervical pregnancy: a review and a report of two cases treated conservatively Ultrasound Obstet Gynecol 1996 8 06 373 380 10.1046/j.1469-0705.1997.08060373.x 9014275 15 Farabow W S Fulton J W Fletcher V Jr Velat C A White J T Cervical pregnancy treated with methotrexate N C Med J 1983 44 03 91 93 6302525 16 Kung F T Chang S Y Efficacy of methotrexate treatment in viable and nonviable cervical pregnancies Am J Obstet Gynecol 1999 181 06 1438 1444 10.1016/s0002-9378(99)70389-3 10601926 17 Kim T J Seong S J Lee K J Lee J H Shin J S Lim K T Clinical outcomes of patients treated for cervical pregnancy with or without methotrexate J Korean Med Sci 2004 19 06 848 852 10.3346/jkms.2004.19.6.848 15608396 18 Uludag S Z Kutuk M S Aygen E M Sahin Y Conservative management of cervical ectopic pregnancy: Single-center experience J Obstet Gynaecol Res 2017 43 08 1299 1304 10.1111/jog.13362 28586112 19 Jeng C J Ko M L Shen J Transvaginal ultrasound-guided treatment of cervical pregnancy Obstet Gynecol 2007 109 05 1076 1082 10.1097/01.AOG.0000262052.09350.52 17470585 20 Kirk E Bourne T The nonsurgical management of ectopic pregnancy Curr Opin Obstet Gynecol 2006 18 06 587 593 10.1097/GCO.0b013e3280105886 17099327 21 Mesogitis S Pilalis A Daskalakis G Papantoniou N Antsaklis A Management of early viable cervical pregnancy BJOG 2005 112 04 409 411 10.1111/j.1471-0528.2004.00447.x 15777436 22 Dziedzic J M Patel P V Cervical ectopic pregnancy: a rare site of implantation J Emerg Med 2019 56 06 e123 e125 10.1016/j.jemermed.2019.03.024 31003816 23 Weibel H S Alserri A Reinhold C Tulandi T Multidose methotrexate treatment of cervical pregnancy J Obstet Gynaecol Can 2012 34 04 359 362 10.1016/S1701-2163(16)35217-3 22472336 24 Murji A Garbedian K Thomas J Cruickshank B Conservative management of cervical ectopic pregnancy J Obstet Gynaecol Can 2015 37 11 1016 1020 10.1016/s1701-2163(16)30051-2 26629722 25 Guzowski G Sieroszewski P Invasive ultrasound in the management of cervical ectopic pregnancy Eur J Obstet Gynecol Reprod Biol 2014 172 7 9 10.1016/j.ejogrb.2013.10.016 24287286 26 Jachymski T Moczulska H Guzowski G Pomorski M Piątek S Zimmer M Conservative treatment of abnormally located intrauterine pregnancies (cervical and cesarean scar pregnancies): a multicenter analysis (Polish series) J Matern Fetal Neonatal Med 2020 33 06 993 998 10.1080/14767058.2018.1514009 30122076 27 Osada H Teramoto S Kaijima H Segawa T Miyauchi O Nagaishi M A novel treatment for cervical and cesarean section scar pregnancies by transvaginal injection of absolute ethanol to trophoblasts: efficacy in 19 cases J Minim Invasive Gynecol 2019 26 01 129 134 10.1016/j.jmig.2018.04.021 29723645 28 Adabi K Nekuie S Rezaeei Z Rahimi-Sharbaf F Banifatemi S Salimi S Conservative management of cervical ectopic pregnancy: systemic methotrexate followed by curettage Arch Gynecol Obstet 2013 288 03 687 689 10.1007/s00404-013-2807-y 23525594 29 Fylstra D L Cervical pregnancy: 13 cases treated with suction curettage and balloon tamponade Am J Obstet Gynecol 2014 210 06 5810 5.81E7 10.1016/j.ajog.2014.03.057
PMC009xxxxxx/PMC9800063.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36138537 10.1055/s-0042-1756212 RBGO-22-0025 Original Article Endometriosis Postoperative Complications and Stoma Rates After Laparoscopic Resection of Deep Infiltrating Endometriosis with Bowel Involvement Complicações pós-operatórias e taxas de estomia após ressecção laparoscópica de endometriose infiltrativa profunda com envolvimento intestinalhttp://orcid.org/0000-0002-5566-9284 Parra Rogério Serafim 1 http://orcid.org/0000-0002-4049-0185 Valério Fernando Passador 1 http://orcid.org/0000-0002-5733-0794 Zanardi José Vitor Cabral 1 http://orcid.org/0000-0002-4440-2023 Feitosa Marley Ribeiro 1 http://orcid.org/0000-0001-7267-1880 Camargo Hugo Parra 1 http://orcid.org/0000-0003-3593-0526 Féres Omar 1 School of Medicine of Ribeirão Preto, Universidade de São Paulo, Ribeirão Preto, SP, Brazil Address for correspondence Rogério Serafim Parra, MD, PhD Av. Bandeirantes, 3900, 14048-900, Ribeirão Preto, [email protected] 22 9 2022 11 2022 1 9 2022 44 11 10401046 27 1 2022 28 6 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  The purpose was to assess the rates of postoperative complications and the need of temporary stoma of laparoscopic surgical treatment for bowel endometriosis in a referral center. Methods  The surgical indication, type of operation, operative time, length of hospital stay, need for a temporary stoma, rate of conversion to open surgery, postoperative complications were evaluated. Results  One-hundred and fifty patients were included. The average duration of surgery was significantly longer for segmental resection (151 minutes) than for disc excision (111.5 minutes, p < 0.001) and shaving (96.8 minutes, p < 0.001). Patients with segmental resection had longer postoperative lengths of hospital stay (1.87 days) compared with patients with disc excision (1.43 days, p < 0.001) and shaving (1.03 days, p < 0.001). A temporary stoma was performed in 2.7% of patients. Grade II and III postoperative complications occurred in 6.7% and 4.7% patients, respectively. Conclusion  Laparoscopic intestinal resection has an acceptable postoperative complication rate and a low need for a temporary stoma. Resumo Objetivo  O objetivo foi avaliar as taxas de complicações pós-operatórias e a necessidade de estomia temporária do tratamento cirúrgico laparoscópico para endometriose intestinal em um centro de referência. Métodos  Foram avaliados a indicação cirúrgica, tipo de operação, tempo operatório, tempo de internação, necessidade de estomia temporária, taxa de conversão para cirurgia aberta, complicações pós-operatórias. Resultados  Cento e cinquenta pacientes foram incluídos. A duração média da cirurgia foi significativamente maior para a ressecção segmentar (151 minutos) do que para a excisão do disco (111,5 minutos, p  < 0,001) e shaving (96,8 minutos, p  < 0,001). Pacientes com ressecção segmentar tiveram maior tempo de internação pós-operatória (1,87 dias) em comparação com pacientes com excisão de disco (1,43 dias, p  < 0,001) e shaving (1,03 dias, p  < 0,001). Um estoma temporário foi realizado em 2,7% dos pacientes. Complicações pós-operatórias de grau II e III ocorreram em 6,7% e 4,7% dos pacientes, respectivamente. Conclusão  A ressecção intestinal laparoscópica apresenta taxa aceitável de complicações pós-operatórias e baixa necessidade de estomia temporária. Keywords colorectal surgery endometriosis laparoscopy postoperative complications surgical stomas Palavras-chave cirurgia colorretal endometriose laparoscopia complicações pós-operatórias estomas cirúrgicos ==== Body pmcIntroduction Endometriosis is a chronic inflammatory disease with a high prevalence among women of reproductive age. 1 Deep infiltrating endometriosis (DIE) is one of the most severe types of endometriosis, and in up to 25% of cases, it affects the bowel. 2 Lesions can be single or multifocal, and depending on the anatomical site affected, they can present symptoms that vary from dysmenorrhea to dyspareunia, and rectal bleeding. 3 Preoperative assessment is crucial to determine the best therapeutic approach for intestinal DIE. 4 This planning consists of clinical examinations and imaging, to identify the location, such as the rectum, vagina, uterosacral ligaments, bladder, and ureter. Imaging can predict the presence of DIE and define the size and extent of lesions compromising the bowel. 5 Lesions affecting the ileum, cecum, or the appendix are unusual and can be difficult to identify with preoperative imaging. Therefore, careful and methodical inspection of these structures must be performed during laparoscopy. 6 Patient age, pain intensity, risk of intestinal obstruction and desire for pregnancy are factors that should be considered for the management of DIE with bowel involvement. Surgery is mainly indicated in patients with pelvic pain who do not respond to medical therapy and in those with the desire for pregnancy. 2 6 Laparoscopic segmental resection, rectal disc excision, and rectal shaving are the main surgical techniques described to treat DIE infiltrating the intestine. The coexistence of two or more lesions affecting the bowel requires segmental resection of the bowel. 5 Rectal disc excision should be reserved for patients with single lesions smaller than 3 cm. Ileal disease also requires surgical excision because of the high risk of acute obstruction. 7 When the appendix is affected, surgery is always indicated to rule out carcinoid tumors. 8 Surgical removal of the lesion is also required when lesions are symptomatic, impairing bowel, urinary, sexual, and reproductive functions. Complete resection should be attempted to reduce the risk of residual lesions and recurrence. 6 9 Postoperative complications represent one of the greatest concerns for bowel surgeries, including leakage, fistulas, and severe peritonitis. 10 Another fear is whether the patients will need a temporary stoma. The aim of the present study was to assess the rates of postoperative complications and the need for temporary stoma in laparoscopic surgical treatment for bowel endometriosis at a single referral center and to compare them with the outcomes in the available literature. Methods We conducted a retrospective study using electronic data prospectively recorded. We included all patients with DIE with bowel involvement submitted to laparoscopy with intestinal resection from October 2014 to October 2019. Patients submitted to open surgery as the main surgery or those without bowel involvement were excluded. We evaluated the following patient's characteristics: age, body mass index (BMI), previous surgeries, surgical indication, type of operation, operative time, length of hospital stay, need for temporary stoma, conversion rate to open surgery, and need for a second bowel resection. We also evaluated grades II and III postoperative complications according to the Clavien-Dindo classification. 11 All surgical procedures were performed by the same surgical team, composed by a colorectal surgeon and three gynecologists. Surgery was mainly indicated in patients with pelvic pain who did not respond to medical therapy as well as those with desire for pregnancy. We also indicated surgery in ileal disease because of the high risk of bowel obstruction, and when the appendix was affected, to rule out carcinoid tumors. The criteria to perform either rectal shaving, disc excision or segmental resection depended on location, number, size, and circumferential involvement of the lesions as described elsewhere. 2 In summary, lesions larger than 3 cm in diameter, multifocal bowel lesions or extent of bowel circumference involvement higher than 40% required a segmental resection. Disc excision was the preferred technique in the presence of unique nodules, lower than 3 cm in diameter and with lower than 40% of circumference involvement. Rectal shaving was planned to be performed in patients without involvement of inner layer muscularis or deeper. Some patients required more than one bowel resection for DIE. When two or more surgical procedures for DIE were performed on the same patient, we considered the major surgical procedure in the statistical analysis, to assess the outcomes. Our imaging protocol includes evaluation with transvaginal ultrasound with a high-resolution linear transducer and bowel preparation in all patients to map the endometriosis lesions of the right iliac fossa and to detect lesions of the ileum, cecum, and appendix. Previous studies showed that ultrasonography has diagnostic accuracy not inferior to magnetic resonance imaging. 12 13 The diagnostic performance of transvaginal ultrasound and magnetic resonance imaging are similar for detecting deep endometriosis involving the rectosigmoid colon, uterosacral ligaments, and rectovaginal septum. Therefore, it must be considered the primary approach for deep endometriosis diagnosis. Categorical variables were expressed as frequencies/percentages and continuous variables as means ± standard deviations. The one-sample Kolmogorov-Smirnov test was used to assess the normality of continuous variables. Analysis of variance was used to compare continuous variables. Fisher exact test or the χ2 test was used to compare categorical variables. All p -values were 2-sided, and a significance level of 5% was established. Statistical analysis was performed with IBM SPSS Statistics for Windows, Version 20.0 (IBM Corp., Armonk, NY, USA). The study was approved by the ethics committee of the Hospital das Clínicas of Faculdade de Medicina de Ribeirão Preto, Universidade de São Paulo (HCFMRP-USP) (IRB protocol, CAAE: 31679420.3.0000.5440; Ethics Committee Approval: 4.029.839/2020). All procedures were conducted in accordance with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. Results Demographic Characteristics A total of 150 women with DIE and bowel involvement who underwent laparoscopic surgical management were included. The mean age was 35.1 ± 5.3 years, and the mean body mass index (BMI) was 25.1 ± 3.8 kg/m 2 . Most patients were nullipara (n = 111; 74%), and in 38.7% (n = 58), it was the first attempt at resection of DIE. Ninety-two patients (61.3%) had already been previously subjected to surgical management for endometriosis, and, of these, 7 patients had undergone recent surgical treatment (less than 1 year) in another hospital and were considered to have “frozen pelvis”. The demographic data are shown in Table 1 . Table 1 Demographic data (n = 150) Patients' characteristics Observed Age (mean ± SD) (min–max) (years) 35.1 ± 5.3 (20–53) BMI (mean ± SD) (kg/m 2 ) 25.1 ± 3.8 Previous surgeries for endometriosis n (%) 92(61.3) Previous pregnancies n (%)  Nulliparous  Previous birth 111 (74) 39 (26) Abbreviations: BMI, body mass index; SD, standard deviation Surgical Indication and Surgical Procedures The main surgical indication was severe chronic pelvic pain refractory to medical management (66%), followed by infertility (56%), dyspareunia (46%), and dysmenorrhea (34.7%). The rectosigmoid was the most frequently affected organ of the gastrointestinal tract (98.7%). The appendix was involved in 10.7% (n = 16) of cases. The analysis of the appendix of one patient revealed a well-differentiated neuroendocrine carcinoma at the appendix tip. 14 The ileum was involved in 6% of cases. Only 2 patients (1.3%) did not have rectosigmoid involvement and underwent laparoscopy: one due to isolated small bowel (ileal) involvement and the other due to ileal and appendix involvement. These results are shown in Table 2 . Table 2 Surgical indication and bowel involved in the patients with DIE Surgical indication N (%)  Chronic pelvic pain  Infertility  Dyspareunia  Dysmenorrhea 99 (66) 84 (56) 69 (46) 52 (34.7) Bowel involved N (%)  Rectosigmoid  Appendix  Small bowel  Cecum 148 (98.7) 16 (10.7) 5 (3.3) 4 (2.7) Abbreviations: DIE, deep infiltrating endometriosis Eighty-three percent of patients (n = 124) had uterosacral ligament or retrocervical space involved. An endometrioma was present in 70.6% (n = 106) of the patients, and adenomyosis was present in 35.3% of the cases (n = 53). The median size of the rectosigmoid nodules was 2.1 cm (range 0.8–6.5 cm). In 69.6% of the patients (n = 103), there was a single rectal nodule, and in 47 patients (31.7%), there were 2 of more rectosigmoid nodules. In 41 patients, the rectosigmoid nodule was located up to 8 cm from the anal verge (27.7%), and in 103 patients (69.6%), the lesion was located 9 to 15 cm from the anal verge. In 6 cases (4.1%), the lesions were located more than 15 cm from the anal verge. Segmental resection was performed in 73 patients (48.7%). Rectal disc excision and shaving for bowel endometriosis were performed in 45 (30.0%) and 35 (23.3%) patients, respectively. The associated surgical procedures performed were hysterectomy (n = 19; 12.7%), appendectomy (n = 16; 10.7%), ureter nodule excision (n = 8; 5.3%), and ileocolic and/or small bowel resection (n = 9; 6%). Hysterectomies were performed in 19 patients due to concomitant endometriosis and either adenomyosis or leiomyomatosis; these patients had previously failed clinical treatment and had no future reproductive desire. Bowel endometriosis was histologically confirmed in all patients. These results are shown in Table 3 . Table 3 Type of operations Type of operations N (%) Shaving Shaving (only) Shaving + Hysterectomy Shaving + rectal disc excision Shaving + appendectomy Shaving + ureter nodule excision Shaving + segmental resection + appendectomy Rectal disc excision Rectal disc excision (only) Rectal disc excision + hysterectomy Rectal disc excision + appendectomy Rectal disc excision + shaving Rectal disc excision + segmental resection Rectal disc excision + ureter nodule excision Rectal disc excision + ileal resection + appendectomy Segmental resection (rectosigmoid) Segmental resection (only) Segmental resection + Hysterectomy Segmental resection + ureter/bladder nodule excision Segmental resection + appendectomy Segmental resection + ileal resection + appendectomy Segmental resection + ileal resection Segmental resection + ileocecal resection Segmental resection + appendectomy + ureter nodule excision Segmental resection + shaving + appendectomy Segmental resection + rectal disc excision Ileal resection (without rectal involvement) 35 (24.0) 21 (14.0) 5 (3.3) 3 (2.0) 3 (2.0) 2 (1.3) 1 (0.67) 45 (30) 29 (19.3) 6 (4.0) 4 (2.7) 3 (2.0) 1 (0.67) 1 (0.67) 1 (0.67) 73 (48.7) 49 (32.7) 8 (5.3) 4 (2.7) 3 (2.0) 2 (1.3) 2 (1.3) 2 (1.3) 1 (0.67) 1 (0.67) 1 (0.67) 2 (1.3) The average operative time was 128 ± 55 minutes. The mean duration of the procedure was longer for segmental resection (151 ± 56.3 minutes) than for disc excision (111.5 ± 38.2 minutes, p  < 0.001) and shaving (96.8 ± 48.7 minutes, p  < 0.001). Patients with segmental resection had longer postoperative lengths of hospital stay (1.87 ± 0.69 days) than those with disc excision (1.43 ± 1.17 days, p  < 0.001) and shaving (1.03 ± 0.18 days, p  < 0.001). These results are shown in Figure 1 . Fig. 1 Comparative analysis of operative time (A), mean length of hospital stay (B). Statistically significant differences between groups were demonstrated by arrows with the corresponding p -value. Postoperative Complications Grade-II complications occurred in 6.7% of patients (n = 10). One patient who underwent disc excision had excessive rectal bleeding within 24 hours of the postoperative period and required transfusion without the need for endoscopic hemostasis. There were five patients with surgical site infections, all of whom were successfully treated clinically, and there was one patient with deep vein thrombosis. Two patients had diarrhea within 30 days postoperatively and required hospitalization and intravenous antibiotics. One patient who underwent disc excision required bladder self-catheterization and urinary physiotherapy for 6 months due to bladder atony. Upon clinical follow-up, there was complete resolution of symptoms. Grade-III complications requiring surgical intervention occurred in 4.7% of patients (n = 7). Two patients had thermal injuries of the rectum during laparoscopy, and both cases were detected within 36 hours of the 1 st surgery. In the patient who underwent disc excision, the abdominal cavity was cleaned, and sigmoid ostomy was performed. In the patient undergoing shaving, there was a need for segmental resection of the rectum, with anastomosis and protective ostomy. One patient developed a pelvic abscess, with the need for another laparoscopy 47 days after the 1 st surgery. In three patients, there was an accidental injury to the ureter detected during laparoscopy. Last, 1 patient presented with a urinoma, and a late thermal bladder injury was detected on the 17 th postoperative day requiring a laparoscopic surgical re-approach and ureter reimplantation. Intensive care was not required in any of the patients. The major postoperative surgical complications (grade III) are presented in Table 4 . Table 4 Major complications, classified as grade III of the Clavien-Dindo classification (n = 7) Complication N (%) Urinary complications 4 (2.7) Shaving Disc excision Segmental resection Infected hematoma of the Douglas cul-de-sac Shaving Disc excision Segmental resection 0 (0) 0 (0) 4 (2.7) 1 (0.67) 0 (0) 0 (0) 1 (0.67) Thermal bowel injury Shaving Disc excision Segmental resection 2 (1.3) 1 (0.67) 1 (0.67) 0 (0) Conversion Rates, and the Need for Temporary Stoma and fora Second Bowel Resection The rate of conversion to open surgery was 2.0% (n = 3). A temporary stoma was performed in 4 patients (2.7%), and this outcome was not different among the 3 surgeries. In 2 patients, the indication for the stoma was due to thermal injury to the intestine; one protective stoma was performed in a patient who underwent rectal disc excision with leakage detected intraoperatively, and another protective stoma was performed in a patient with very low (< 3 cm from the anal verge) colorectal anastomoses. The need for a 2 nd laparoscopy for the resection of intestinal nodules occurred in 5 patients during an average follow-up time of 28.9 ± 16.7 months. All of these patients had postoperative image evaluation showing no residual lesions in the intestine after the first surgery. There were no cases of anastomosis dehiscence or death. These results are summarized in Table 5 . Table 5 Short-term postoperative outcomes Outcome N (%) Conversion rate to open surgery Shaving Disc excision Segmental resection 3 (2.0) 0 (0) 1 (0.67) 2 (1.3) Need of temporary stoma Shaving Disc excision Segmental resection 4 (2.7) 1 (0.67) 2 (1.3) 1 (0.67) 30-day emergency room visit Shaving Disc excision Segmental resection 5 (3.3) 1 (0.67) 0 (0) 4 (2.7) 30-day reoperation rate Shaving Disc excision Segmental resection 3 (2.0) 1 (0.67) 1 (0.67) 1 (0.67) Surgical recurrence Shaving Disc excision Segmental resection Blood transfusion Shaving Disc excision Segmental resection Bladder atony Shaving Disc excision Segmental resection Dehiscence of intestinal anastomosis Overall mortality 5 (3.3) 2 (1.3) 2 (1.3) 1 (0.67) 1 (0.67) 0 (0) 1 (0.67) 0 (0) 1 (0.67) 0 (0) 1 (0.67) 0 (0) 0 (0) 0 (0) Discussion Our study showed that laparoscopic surgery for the treatment of DIE with bowel involvement was safe, with very low conversion rates to open surgery and without the routine use of stoma during the initial surgery. This did not negatively impact the results of our cohort of patients, who had low rates of serious postoperative complications and low need for second laparoscopy for the resection of intestinal nodules due to recurrence. Among all the patients in the study, four needed a temporary stoma. All stomas were subsequently reversed within a period of up to 3 months after the first surgery. The stoma rate was lower in our study than the average rate reported in the literature. Ruffo et al. 15 reported that 21.3% of patients required an ileostomy, and stoma was still present at the time of follow-up in 3.2% of patients (5 of 156 stomas). Another study compared conservative surgery (rectal shaving or disc excision) versus segmental resection and showed that 59.3% of patients who underwent conservative surgery needed a colostomy and 71.4% of patients who underwent segmental resection needed a colostomy (42.9%) or ileostomy (28.6%). 16 A protective stoma was performed in 27% of patients in a prospective cohort study, 17 and in a case-series study using mainly conventional laparoscopy, the stoma rate was 19.1%. 18 Our study showed no differences in the stoma rate among the three surgeries, even if we compared conservative surgery with segmental resection. The indication for a protective stoma should be individualized, depending on risk factors for leakage, and should be recommended only for a low coloanal anastomosis 3 or when the anastomosis is performed lower than 5 cm from the anal verge due to the risk of anastomotic leakage. 19 20 Laparoscopic surgery for DIE with bowel involvement is almost always elective and scheduled, mostly in young patients: those without comorbidities (ASA 1 or 2), those with no prolonged operative time (< 3 hours), and those with no anemia, malnutrition, or any other factors that could increase the risk of dehiscence, such as anemia or sepsis. 21 In our series, there were no cases of anastomotic leakage, which is very similar to other series that showed 0.8% of cases of anastomotic leakage. 18 A stoma may be associated with postoperative complications, such as abnormal healing of the stoma scar, wound or urinary infections following stoma closure, leakage, hernias, subcutaneous abscesses, and bowel obstruction. 22 Additionally, those patients will require another surgical intervention. Serious complications occurred in 7 patients and included thermal damage to the intestine and urinary system, damage to the ureter, and postoperative infection requiring hospitalization. Abo et al. 23 showed 11.8% of patients with Clavien 3b postoperative complications, two thirds of whom were managed by segmental colorectal resection. They also reported 14 cases (3.8%) of rectovaginal fistula and 24 cases (6.6%) of pelvic abscess. Malzoni et al. 24 described a rate of 8.06% of major postoperative complications. Tarjanne et al. 25 found that the rate of major complications was 12%, and a rectal nodule higher than 4 cm was associated with the development of a major complication. There are some frequently observed risk factors of major complications. One of them is the excessive use of electrocoagulation, which increases the risk of rectovaginal fistulae and abscesses due to the risk of necrosis of the posterior vaginal cuff. In all surgeries in our cohort, we used an ultrasonic energy device, and laparoscopic bipolar electrosurgical devices were used, which may be associated with lower postoperative complications. In one patient, a ureteral fistula was diagnosed, attributed to thermal injury. This is in accordance with a large series that reported an incidence of ureteral fistula of 0.7%. 18 Iatrogenic injury to the ureter occurred in 2% of patients. To avoid this complication, which can occur in patients with previous surgeries and distorted anatomy, preoperative planning is important, including imaging and an experienced surgical team familiar with advanced laparoscopic surgery in DIE. 26 One patient (0.67%) developed bladder atony. During the follow-up, there was complete resolution of symptoms. Marty et al. 27 described an incidence of 1.8% of bladder atony necessitating ≥ 3 weeks of daily self-catheterization. Some authors have explained that segmental rectal resection, which requires circular dissection of the rectum and wide dissection below the nodule, exposes patients to more functional morbidity than the discoid resection technique, which leads to less risk for damage to the plexus. 28 In our cohort, the conversion rate to laparotomy was very low, which is similar to other series. 29 Deep infiltrating endometriosis invading the intestine is a challenging condition, and intraoperative technical difficulty, level of laparoscopic complexity, and surgeon inexperience are some of the main risk factors for conversion to open surgery. 30 In colorectal surgery, patients who underwent conversion to open surgery had higher mortality, higher overall morbidity, longer lengths of hospitalization, and increased hospital charges. 31 Surgical recurrence, meaning a need for another bowel resection, was similar to other studies. 16 25 32 All of our patients were successfully treated with a second laparoscopy. Our study has several limitations. First, long-term postoperative functional outcomes, such as evacuation and/or incontinence outcomes and urinary disorders, were not evaluated. However, an assessment of subjective wellbeing and intestinal and urinary symptoms was carried out and published elsewhere. 33 Secondly, although a prospective database was used, this study was limited by its retrospective nature. Third, some important data was not available in many patients, such as opening of the vagina, which is known to be a cause of postoperative complications. However, in our cohort, we did not observe any cases of rectal fistula or dehiscence. Conclusion In summary, the present study confirms that laparoscopic intestinal resection for DIE has a low postoperative complication rate and a low need for temporary stoma. Contributions Conflict of interests The authors have no conflict of interests to declare. R. S. P.: Conception and design of the study, data collection, data analysis and interpretation, drafting of manuscript and figures, and critical revision of manuscript. F. P. V.: data collection, data analysis and interpretation; critical revision of the manuscript. H. P. C.: data collection, data analysis and interpretation, tables preparation; critical revision of the manuscript. M. R. F.: data analysis and interpretation, statistical analysis, and figure preparation; critical revision of the manuscript. J. V. C. Z.: data collection, data analysis and interpretation; critical revision of the manuscript. O. F.: supervision, critical revision of the manuscript. All authors contributed to the analysis and interpretation of data, revision of the manuscript for important intellectual content, granted final approval of the version to be published and agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. ==== Refs References 1 Giudice L C N. Engl. J. Med. 2010 362 2389 2398 10.1056/NEJMcp1000274 20573927 2 Abrão M S Petraglia F Falcone T Keckstein J Osuga Y Chapron C Hum. Reprod. Update 2015 21 329 339 10.1093/humupd/dmv003 25618908 3 Kent A Shakir F Rockall T Haines P Pearson C Rae-Mitchell W Jan H J. Minim. Invasive Gynecol. 2016 23 526 534 10.1016/j.jmig.2015.12.006 26724718 4 Laganà A S La Rosa V L Minerva Med. 2020 111 18 20 10.23736/S0026-4806.19.06370-5 31755676 5 Abrão M S J. Minim. Invasive Gynecol. 2016 23 461 462 10.1016/j.jmig.2016.02.007 26898896 6 Kho R M Andres M P Borrelli G M Neto J S Zanluchi A Abrão M S Best Pract. Res. Clin. Obstet. Gynaecol. 2018 51 102 110 10.1016/j.bpobgyn.2018.01.020 29545114 7 Barbosa R N Andres M P Kho R M Abrão M S J. Minim. Invasive Gynecol. 2018 25 759 760 10.1016/j.jmig.2017.10.001 29031678 8 Padovesi Mota I L Klajner S da Costa Goncalves M O Passman L J Podgaec S Appendiceal nodules in the setting of endometriosis can be carcinoid tumors. JSLS 2015 19 03 e2015.00028 10.4293/JSLS.2015.00028 9 Parra R S Féres O Am. J. Obstet. Gynecol. 2022 226 154 155 10.1016/j.ajog.2021.08.022 34453935 10 Parra R S Feitosa M R Féres O Arch. Gynecol. Obstet. 2021 304 1381 1382 10.1007/s00404-020-05890-y 33226487 11 Dindo D Demartines N Clavien P A Ann. Surg. 2004 240 205 213 10.1097/01.sla.0000133083.54934.ae 15273542 12 Noventa M Scioscia M Schincariol M Cavallin F Pontrelli G Virgilio B Vitale S G Laganà A S Dessole F Cosmi E D'Antona D Andrisani A Saccardi C Vitagliano A Ambrosini G Diagnostics (Basel) 2019 9 225 10.3390/diagnostics9040225 31861142 13 Guerriero S Saba L Pascual M A Ajossa S Rodriguez I Mais V Alcazar J L Ultrasound Obstet. Gynecol. 2018 51 586 595 10.1002/uog.18961 29154402 14 Parra R S Feitosa M R Biagi G B. B Brandão D F Moraes M M. F. D. S Silvestre L Zanardi J V. C Sato Junior N H Féres O da Rocha J J. R J. Med. Case Reports 2020 14 152 10.1186/s13256-020-02490-x 15 Ruffo G Scopelliti F Manzoni A Sartori A Rossini R Ceccaroni M Minelli L Crippa S Partelli S Falconi M BioMed Res. Int. 2014 2014 463058 10.1155/2014/463058 24877097 16 Roman H Tuech J J Huet E Bridoux V Khalil H Hennetier C Bubenheim M Branduse L A Hum. Reprod. 2019 34 2362 2371 10.1093/humrep/dez217 31820806 17 Belghiti J Ballester M Zilberman S Thomin A Zacharopoulou C Bazot M Thomassin-Naggara I Daraï E J. Minim. Invasive Gynecol. 2014 21 472 479 10.1016/j.jmig.2013.12.094 24378832 18 FRIENDS group (French coloRectal Infiltrating ENDometriosis Study group) Roman H J. Gynecol. Obstet. Hum. Reprod. 2017 46 159 165 10.1016/j.jogoh.2016.09.004 28403973 19 Bouaziz J Soriano D Minerva Ginecol. 2017 69 477 487 10.23736/S0026-4784.17.04052-7 28381079 20 Akladios C Messori P Faller E Puga M Afors K Leroy J Wattiez A J. Minim. Invasive Gynecol. 2015 22 103 109 10.1016/j.jmig.2014.08.001 25109779 21 Buchs N C Gervaz P Secic M Bucher P Mugnier-Konrad B Morel P Int. J. Colorectal Dis. 2008 23 265 270 10.1007/s00384-007-0399-3 18034250 22 Bonin E Bridoux V Chati R Kermiche S Coget J Tuech J J Roman H Eur. J. Obstet. Gynecol. Reprod. Biol. 2019 232 46 53 10.1016/j.ejogrb.2018.11.008 30468983 23 Abo C Moatassim S Marty N Saint Ghislain M Huet E Bridoux V Tuech J J Roman H Fertil. Steril. 2018 109 172 178 10.1016/j.fertnstert.2017.10.001 29307394 24 Malzoni M Di Giovanni A Exacoustos C Lannino G Capece R Perone C Rasile M Iuzzolino D J. Minim. Invasive Gynecol. 2016 23 512 525 10.1016/j.jmig.2015.09.024 26455526 25 Tarjanne S Heikinheimo O Mentula M Härkki P Acta Obstet. Gynecol. Scand. 2015 94 72-79 10.1111/aogs.12515 25256374 26 Paté M Hauss A S Faller E Colin J Lecointre L Akladios C Bladder and urinary deep pelvic endometriosis: a step-by-step standard approach. J Minim Invasive Gynecol. 2019 2020 27 06 1249 50 10.1016/j.jmig.2019.10.024 27 Marty N Touleimat S Moatassim-Drissa S Millochau J C Vallee A Stochino Loi E Desnyder E Roman H J. Minim. Invasive Gynecol. 2017 24 1121-1127 10.1016/j.jmig.2017.06.019 28673871 28 Possover M Fertil. Steril. 2014 101 754-758 10.1016/j.fertnstert.2013.12.019 24424366 29 Erdem S Imboden S Papadia A Lanz S Mueller M D Gloor B Worni M Dis. Colon Rectum 2018 61 733-742 10.1097/DCR.0000000000001047 29664797 30 Sokol A I Chuang K Milad M P J. Am. Assoc. Gynecol. Laparosc. 2003 10 469-473 10.1016/s1074-3804(05)60146-6 14738630 31 Moghadamyeghaneh Z Masoomi H Mills S D Carmichael J C Pigazzi A Nguyen N T Outcomes of conversion of laparoscopic colorectal surgery to open surgery. JSLS 2014 18 04 e2014.00230 10.4293/JSLS.2014.00230 32 Donnez O Roman H Fertil. Steril. 2017 108 931 942 10.1016/j.fertnstert.2017.09.006 29202966 33 Parra R S Feitosa M R Camargo H P Valério F P Zanardi J V. C Rocha J J. R. D Féres O J. Psychosom. Obstet. Gynaecol. 2021 42 75 80 10.1080/0167482X.2020.1773785 32538257
PMC009xxxxxx/PMC9800064.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36580948 10.1055/s-0042-1757955 RBGO-22-0064 Original Article High Risk Pregnancy Prevalence of Screening for Diabetes Mellitus in Patients Previously Diagnosed with Gestational Diabetes: Factors Related to its Performance Prevalência do rastreamento de diabetes mellitus em pacientes diagnosticadas previamente com diabetes gestacional: fatores relacionados à sua realizaçãohttp://orcid.org/0000-0002-9789-501X Kindermann Lucas 1 http://orcid.org/0000-0002-6739-9891 Costa Leandro de Liz 2 http://orcid.org/0000-0003-0196-4488 Trapani Júnior Alberto 12 1 Departament of Gynecology and Obstetrics, Hospital Regional de São José Dr. Homero de Miranda Gomes, São José, SC, Brazil 2 Faculdade de Medicina, Universidade do Sul de Santa Catarina, Palhoça, SC, Brazil Address for correspondence Lucas Kindermann Departament of Gynecology and Obstetrics, Hospital Regional de São José Dr. Homero de Miranda GomesSão José, [email protected] 29 12 2022 11 2022 1 12 2022 44 11 10321039 21 2 2022 10 8 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  To determine how many patients underwent screening for diabetes mellitus (DM) in the puerperium after a diagnosis of gestational DM (GDM) and which factors were related to its performance. Methods  The present is a prospective cohort study with 175 women with a diagnosis of GDM. Sociodemographic and clinico-obstetric data were collected through a questionnaire and a screening test for DM was requested six weeks postpartum. After ten weeks, the researchers contacted the patients by telephone with questions about the performance of the screening. The categorical variables were expressed as absolute and relative frequencies. The measure of association was the relative risk with a 95% confidence interval (95%CI), and values of p ≤ 0.05 were considered statistically significant and tested through logistic regression. Results  The survey was completed by 159 patients, 32 (20.1%) of whom underwent puerperal screening. The mean age of the sample was of 30.7 years, and most patients were white (57.9%), married (56.6%), and had had 8 or more years of schooling (72.3%). About 22.6% of the patients used medications to treat GDM, 30.8% had other comorbidities, and 76.7% attended the postnatal appointment. Attendance at the postpartum appointment, the use of medication, and the presence of comorbidities showed an association with the performance of the oral glucose tolerance test in the puerperium. Conclusion  The prevalence of screening for DM six weeks postpartum is low in women previously diagnosed with GDM. Patients who attended the postpartum consultation, used medications to treat GDM, and had comorbidities were the most adherent to the puerperal screening. We need strategies to increase the rate of performance of this exam. Resumo Objetivo  Determinar quantas pacientes realizaram o rastreamento puerperal para diabetes mellitus (DM) após diagnóstico de DM gestacional (DMG) e quais fatores estão relacionados com a sua realização. Métodos  Trata-se de um estudo com uma coorte prospectiva de 175 puérperas com diagnóstico de DMG. Informações sociodemográficas e clinico-obstétricas foram coletadas por meio de questionário, e solicitou-se a realização do rastreamento para DM às seis semanas de pós-parto. Após dez semanas, os pesquisadores contataram as pacientes por telefone com questões sobre a realização do rastreamento. As variáveis categóricas foram expressas em termos de frequências absoluta e relativa. A medida de associação foi o risco relativo com intervalo de confiança de 95% (IC95%), e valores de p ≤ 0,05 foram considerados estatisticamente significativos e testados por regressão logística. Resultados  A pesquisa foi concluída por 159 pacientes, 32 (20,1%) das quais fizeram o rastreamento puerperal. A idade média da amostra foi de 30,7 anos, e a maioria das pacientes eram brancas (57,9%), casadas (56,6%) e tinham 8 ou mais anos de escolaridade (72,3%). Cerca de 22,6% utilizaram medicações para tratamento da DMG, 30,8% eram portadoras de outras comorbidades, e 76,7% compareceram na consulta pós-natal. O comparecimento na consulta pós-parto, a utilização de medicamentos e a presença de comorbidades demonstraram associação com a realização do teste oral de tolerância à glicose no puerpério. Conclusão  O rastreamento de DM após seis semanas de puerpério é baixo em mulheres com diagnóstico prévio de DMG. Pacientes que compareceram na consulta pós-parto, utilizaram medicações para tratamento da DMG, e eram portadoras de comorbidades foram as mais aderentes à realização do rastreamento puerperal. Necessitamos de estratégias para aumentar a abrangência da realização desse exame. Keywords gestational diabetes screening programs postpartum period diabetes mellitus Palavras-chave diabetes gestacional programas de rastreamento período pós-parto diabetes mellitus ==== Body pmcIntroduction Diabetes mellitus (DM) is a chronic disease characterized by high levels of blood glucose due to insufficient insulin production or some degree of insulin resistance that can contribute to physical, psychological and clinical complications. 1 It is estimated that ∼ 9.3% of the world population of adults aged between 20 and 79 years live with this condition. 1 Brazil is the South-American country with the highest prevalence of this disease, with 16.8 million adults affected, and the prevalence is higher among women (10.4% of the population) than among men (8.4% of the population). 1 Women with hyperglycemia detected for the first time during pregnancy and with blood glucose levels that are not sufficient to achieve the diagnostic criteria for DM are characterized as having gestational DM (GDM). 2 It is estimated that 15.8% of the patients who delivered live newborns worldwide were affected by hyperglycemia in 2019, and 83.6% of the cases were due to GDM. 1 In Brazil, the diagnosis of GDM follows the consensus of the International Association of the Diabetes and Pregnancy Study Group (IADPSG): patients with fasting glucose levels between 92 mg/dL and 125 mg/dL or at least one altered value on the Oral Glucose Tolerance Test (OGTT) with 75 g of dextrose, performed between 24 and 28 weeks, with cutoff points of ≥ 92 mg/dL fasting, ≥ 180 mg/dL in the first hour, and ≥ 153 mg/dL in the second hour. 2 3 When compared with pregnant women with normoglycemia, GDM carriers are ∼ 7 to 10 times more likely to develop type-2 DM (DM2) after delivery. 4 5 The prevalence of hyperglycemia in patients with GDM varies from 25.9% in the first year postpartum to up to 53.7% when evaluated 5 years later. 6 Therefore, national and international protocols recommend performing the OGTT with 75 g of dextrose 6 weeks after delivery to screen for DM in patients diagnosed with GDM. 2 7 Early diagnosis of diabetes is important to avoid the impacts of prolonged hyperglycemia, such as increased incidence and mortality from cardiovascular diseases, since the diagnosis of GDM is also associated with higher rates of dyslipidemia, hypertension, vascular dysfunctions, atherosclerosis, and occurrence of cardiovascular events throughout life. 8 9 10 However, despite the recommendations, the rate of patients diagnosed with GDM who undergo postpartum screening is generally low, ranging from 19.5% to 54%. 11 12 It has been observed that many patients fail to undergo DM screening during the puerperium. There is a need to create a program to follow up these patients, perform exams to enable the establishment of early diagnoses, and act in the prevention and treatment of this comorbidity that has such a great long-term impact for the patient and the public health system. Thus, the present research aimed to determine how many GDM patients among the population under study undergo the screening test in the puerperal period and what are the factors related to its performance. Methods The present is a prospective cohort study. The population consisted of patients diagnosed with GDM, hospitalized for delivery, in the public health system, between May and September 2021, at the Obstetrics service of Hospital Regional de São José Dr. Homero de Miranda Gomes, in the city of São José, state of Santa Catarina, Southern Brazil. The sampling was carried out by convenience, with an estimated sample size of 175 cases. The calculation was performed with the OpenEpi software, version 3, considering the significant variables from previous studies, 13 14 with bilateral significance (1-α) of 0.05 and power (1-β) of 0.80, and an estimated loss of up to 30%. The inclusion criteria were patients diagnosed with GDM based on the aforementioned IADSPG consensus, 2 3 who knew how to read and write in Portuguese. Patients with blood glucose levels in the laboratory screening that enabled the establishment of a diagnosis of previous DM or who had already had a diagnosis prior to pregnancy were excluded. At first, during hospitalization, the patients answered a questionnaire through an interview with the researcher pertaining to: sociodemographic data (age, marital status, self-declared skin color, area of residence, years of schooling, family income in terms of the Brazilian minimum monthly wage of $213,09 according to federal law no. 14.158/2021 and commercial US dollar quotation at the beginning of the 2021, occupational status); obstetric history (parity, number of prenatal appointments, pregestational BMI, weight gain during pregnancy, mode of delivery, use of oral antidiabetics or insulin during pregnancy, birth weight, shoulder dystocia, amniotic fluid disorders diagnosed in any moment of pregnancy, gestational age at the time of the GDM diagnosis, history of GDM in previous pregnancies, breastfeeding during hospitalization, and associated comorbidities like hypertensive syndromes, thyroid disorders, and neurological or atopic diseases); family history of DM; and personal habits (sedentary lifestyle, smoking status). The data collected was complemented by the analysis of the hospital records and prenatal care documents. Birth weight greater than 4,000 g, presence of polyhydramnios, occurrence of shoulder dystocia and/or admission to the Neonatal Intensive Care Unit (NICU) were also grouped in a variable named “Composite outcome perinatal morbidity” for the statistical analysis. The patients were informed of the importance of screening for DM six weeks after delivery. At the time of hospital discharge, they received a request for an OGTT test with 75 g of glucose (before and 2 hours after ingestion), which should be scheduled and performed at the Basic Health Unit of their hometowns, through the Brazilian Unified Health System. 2 After four weeks of the period stipulated for the performance of the screening test, the researchers contacted the patients by telephone, asking if they had performed the requested test and if they kept breastfeeding in this period. The patients who failed to undergo the exam were asked why by the researchers. Patients who presented positive screening were instructed to seek their Basic Health Unit to continue the management of the disease. After the collection the data was inserted in a Windows Excel (Microsoft Corp., Redmond, WA, United States) spreadsheet and later exported to the Statistical Package for the Social Sciences (PASW Statistics for Windows, SPSS Inc., Chicago, IL, United States) software, version 18.0 for the descriptive and analytical analyses. To describe the quantitative variables, means and standard deviations were calculated. The categorical variables were expressed as absolute (n) and relative (%) frequencies. The bivariate analysis was performed using the Chi-squared test to present the distribution of the variables. The measure of association used was the relative risk (RR) and 95% confidence interval (95%CI). Differences were considered statistically significant when p ≤ 0.05, and they were tested through logistic regression. The present research project was submitted to the Research Ethics Committee and accepted under CAAE number 45699821.1.0000.0110, and data collection only started after the approval. Results The sample was composed of 175 puerperal women diagnosed with GDM, and all of them answered the questionnaire. In the second stage, 159 patients completed the survey, with a total of 16 losses: patients who did not answer the telephone contact, or whose number was wrong in the registration, or who requested exclusion from the study during the telephone contact ( Fig. 1 ). Fig. 1 Flowchart of the selection of the 175 patients who compose the study sample. The age of the patients ranged from 17 to 44 years, with a mean of 30.7 years, and most (66.7%) were aged between 20 and 34 years, self-reported as white (57.9%), were married (56.6%), and reported 8 or more years of schooling (72.3%). The other sociodemographic characteristics are shown in Table 1 . Table 1 Sociodemographic characteristics of the 159 patients who completed the survey of the present study Variables Underwent the OGTT (%) Did not undergo the OGTT (%) Total (%) p -value Age Mean 32.5 ± 5.8 30.2 ± 6.1 30.7 ± 5.8 < 19 years old 1 (11.1) 8 (88.9) 9 (5.7) 0.487 20–34 years old 20 (18.9) 86 (81.1) 106 (66.7) > 35 years old 11 (25.0) 33 (75.0) 44 (27.6) 0.343 Self-declared skin color White 19 (20.7) 73 (79.3) 92 (57.9) Non-white (others) 13 (19.4) 54 (80.6) 67 (42.1) 0.846 Marital status: living with partner Yes 31 (31) 69 (69) 152 (95.6) 0.693 No 1 (14.3) 6 (85.7) 7 (4.4) Years of schooling < 8 years 7 (15.9) 37 (84.1) 44 (27.7) 0.412 ≥ 8 years 25 (21.7) 90 (78.3) 115 (72.3) Family income < 1 minimum wage 1 (11.1) 8 (88.9) 9 (5.7) 0.124 Between 1 and 2 minimum wages 6 (13.6) 38 (86.4) 44 (27.7) Between 2 and 3 minimum wages 5 (11.6) 38 (88.4) 43 (27.0) > 3 minimum wages 20 (31.7) 43 (68.3) 63 (39.6) Pre gestational BMI < 18.5 Kg/m 2 0 (00.0) 1 (100.0) 1 (0.6) 18.5–24.9 Kg/m 2 5 (16.1) 26 (83.9) 31 (19.5) 25–29.9 Kg/m 2 15 (26.3) 42 (73.7) 57 (35.9) > 30 Kg/m 2 12 (17.1) 58 (82.9) 70 (44.0) 0.405 Occupational status: paid work Yes 19 (24.4) 59 (75.6) 78 (49.0) 0.191 No 13 (16.0) 68 (84.0) 81 (51.0) Family history of DM Yes 18 (22.5) 62 (77.5) 80 (50.3) 0.452 No 14 (17.7) 65 (82.3) 79 (49.7) History of GDM in previous pregnancies Yes 3 (14.3) 18 (85.7) 21 (13.2) 0.474 No 29 (21.0) 109 (79.0) 138 (86.8) Smoking Yes 7 (25.9) 20 (74.1) 27 (17.0) 0.409 No 25 (18.9) 107 (81.1) 132 (83.0) Breastfeeding Yes 27 (20.0) 108 (80.0) 135 (84.9) 0.633 No 5 (20.8) 19 (79.2) 24 (15.1) Abbreviations: BMI, body mass index; DM, diabetes mellitus; GDM, gestational diabetes mellitus; OGTT, oral glucose tolerance test. As for the obstetric data, most patients were not primiparous (73.6%) and had completed prenatal care with more than 6 appointments (83.6%). Of the 159 patients who completed the survey, 76.7% attended the postnatal appointment, 22.6% used medications (oral antidiabetics or insulin therapy) to treat GDM, and 30.8% had other comorbidities. There was 1 case of shoulder dystocia during delivery. The rest of the obstetric data is presented in Table 2 . Table 2 Obstetrics and perinatal characteristics of the 159 patients who completed the survey of the present study Obstetrics and perinatal characteristics Underwent the OGTT (%) Did not undergo the OGTT (%) Total (%) p -value Primiparity Yes 10 (23.8) 32 (76.2) 42 (26.4) 0.488 No 22 (18.8) 95 (81.2) 117 (73.6) Prenatal appointments ≤ 6 2 (7.7) 24 (92.3) 26 (16.4) 0.084 > 6 30 (22.6) 103 (77.4) 133 (83.6) Postnatal appointments Yes 30 (24.6) 92 (75.4) 122 (76.7) 0.011 No 2 (5.4) 35 (94.6) 37 (23.3) Use of medications Yes 15 (41.7) 21 (58.3) 36 (22.6) < 0.001 No 17 (13.8) 106 (86.2) 123 (77.4) Associated comorbidities * Yes 17 (34.7) 32 (65.3) 49 (30.8) 0.002 No 15 (13.6) 95 (86.4) 110 (69.2) Birth weight  > 4,000 g Yes 2 (14.3) 12 (85.7) 14 (8.8) 0.568 No 30 (20.7) 115 (79.3) 145 (91.2) Polyhydramnios Yes 1 (20.0) 4 (80.0) 5 (3.1) 0.994 No 31 (20.1) 123 (79.9) 154 (96.9) Mode of delivery Vaginal 16 (22.2) 56 (77.8) 72 (45.3) Cesarean section 16 (18.4) 71 (81.6) 87 (54.7) 0.549 Admission to the Neonatal Intensive Care Unit Yes 9 (27.3) 24 (72.7) 33 (20.8) 0.250 No 23 (18.3) 103 (81.7) 126 (79.2) Composite outcome perinatal morbidity # Yes 8 (17.8) 37 (82.2) 45 (15.7) 0.643 No 24 (21.1) 90 (78.9) 114 (71.7) Abbreviations: OGTT, oral glucose tolerance test. Notes: *Associated comorbidities, such as like hypertensive syndromes, thyroid disorders, and neurological or atopic diseases. # Birth weight > 4,000 g, polyhydramnios, shoulder dystocia, and admission to the Neonatal Intensive Care Unit. Of the total of 159 cases, 122 women (76.7%) attended the puerperal appointment. Among those, 43 (35.2%) reported they had been instructed about the OGTT and its performance. In total 32 patients (20.1%) underwent the exam 6 weeks postpartum. Among the 127 women who did not undergo the test, 51 (40.1%) reported having submitted the request for the exam at their Basic Health Unit and that the exam was not scheduled; 32 (25.2%) said they forgot to place the request; 7 (5.5%) reported they were instructed by the Basic Health Unit not to undergo it, or to do so at intervals not stipulated by the present study, and in 37 (29.2%) cases the justifications were different. Among the 32 OGTTs performed in the postpartum period, no case of DM was diagnosed, and glucose intolerance was evidenced in 2 exams. The univariate analysis ( Table 3 ) showed that attending the postnatal consultation, having used medication for GDM, and having comorbidities were factors related to performing the screening test in the puerperium. After logistic regression, these variables remained significant, as shown in Table 4 . Table 3 Analysis of factors associated with the performance of postpartum screening of the 159 patients who completed the survey of the present study Factors Univariete analysis Relative risk 95% confidence interval p -value Postnatal appointment 1.294 1.13–1.49 0.011 Medication use 2.835 1.66–4.85 < 0.001 Associated comorbidities* 2.108 1.35–3.28 0.002 Age < 19 years old 0.487 0.08–3.50 0.487 Age > 35 years old 1.323 0.75–2.32 0.343 Self-declared skin color non-white 0.955 0.60–1.52 0.846 Marital status: living with partner 1.017 0.94–1.09 0.693 < 8 years of schooling 0.751 0.37–1.53 0.412 Family income < 2 minimum wages 0.604 0.30–1.21 0.124 Pregestational BMI > 30 Kg/m 2 0.821 0.51–1.33 0.405 Work 1.278 0.91–1.80 0.191 History family of DM 1.152 0.81–1.64 0.452 GDM in previous pregnancies 0.661 0.21–2.11 0.474 Smoking 1.389 0.64–3.00 0.409 Breastfeeding 0.964 0.82–1.13 0.633 Primiparity 1.240 0.68–2.25 0.488 < 6 prenatal appointments 0.331 0.08–1.33 0.084 Birth weight > 4,000 g 0.661 0.16–2.81 0.568 Polyhydramnios 0.992 0.11–8.58 0.994 Cesarean section 0.894 0.61–1.31 0.549 Admission to the NICU 1.488 0.77–2.88 0.250 Composite outcome perinatal morbidity # 0.858 0.44–1.66 0.643 Abbreviations: BMI, body mass index; DM, diabetes mellitus; GDM, gestational diabetes mellitus; NICU, Neonatal Intensive Care Unit. Notes: *Associated comorbidities such as hypertensive syndromes, thyroid disorders, and neurological or atopic diseases. # Birth weight > 4,000 g, polyhydramnios, shoulder dystocia, and admission to the Neonatal Intensive Care Unit. Table 4 Logistic regression of the significant factors associated with the performance of postpartum screening of the 159 patients who completed the survey of the present study Factors Logistic regression Relative risk 95% confidence interval p -value Postnatal appointment 1.213 1.11–1.37 0.023 Medication use 2.455 1.54–4.21 < 0.001 Associated comorbidities* 2.017 1.23–3.12 0.004 Note: *Associated comorbidities such as hypertensive syndromes, thyroid disorders, and neurological or atopic diseases. Discussion The focus of the present study is to establish the prevalence of DM screening in the puerperium and what are the factors associated with its performance. In the present study, about 1 in 5 women (20.1%) completed the OGTT screening 6 weeks postpartum. This index is similar to that found by Ortiz et al., 15 who found a rate of 19%. The prevalence of DM screening in patients with a history of GDM is quite variable in the literature, ranging from 13.8% to 41%, but what all the studies 13 14 16 17 18 19 show, including the current study, is that even though screening is indicated for all postpartum women with a history of GDM, its performance is still low. Screening is important, as it enables the establishment of the risks of developing DM, so physicians can prescribe behavioral or even pharmacological changes to prevent or delay the its diagnosis in this population and avoid the impacts of this comorbidity in the long term. 8 9 10 20 The main reason given for not performing the screening was failure to schedule the examination by the Basic Health Unit (40.1%), despite clear guidelines provided by the Brazilian Ministry of Health. 2 This shows the fragility or lack or organizational alignment within the Brazilian public health care network. The other justifications, for the most part, were related to forgetfulness and other personal issues such as early return to work and difficulties with baby care. These cases could be avoided with more comprehensive assistance on the part of health agents. Women who attended the postpartum appointments were more likely to undergo the screening test (RR: 1.213; 95%CI: 1.11–1.37). Corroborating the present study, Lawrence et al. 21 observed that patients who attended follow-up postnatal appointments were three times more likely to complete the DM screening. This is probably due to the relationship between seeking medical care at the primary care unit with the delivery of the exam request at the Basic Health Unit and the admission into the process of its performance and also the positive feedback on screening guidance during the appointment, although in the present study only 35.2% of the patients who attended the appointment reported having received guidance about the examination by the professionals who assisted them. In another study, 22 whose focus was to observe the relationship between the level of training and the rate of DM screening, the authors found that puerperal women examined by resident physicians or midwives in their postnatal appointment, rather than by fully-trained physicians, tended to adhere more to the performance the test. Another positive factor for the performance of the OGTT was the association with the need to use medication to treat GDM (RR: 2.455; 95%CI 1.54–4.21). Cabizuca et al. 16 showed that the use of insulin during pregnancy increases screening rates by ∼ 6 times. This may be due to the perception that the use of medication implies a greater severity of the condition, which leads patients to be more concerned about the persistence of the disease in the puerperium, with a consequent greater demand for screening. In the present study, we were able to show the association between the presence of comorbidities and adherence to screening (RR: 2.017; 95%CI: 1.23- 3.12). To the best of our knowledge, the present is the first study to show the correlation between these factors. This association may be due to the fact that women who have other morbidities, in addition to GDM, are already more concerned about their health; therefore, a higher rate of them chose to undergo screening. Although these are the only significant associations in the present study, a few other articles 14 17 23 24 25 26 also relate primiparity, advanced maternal age, history of GDM in previous pregnancies and family history of DM as factors that increase patient adherence to screening. When interpreting the results of the present study, one should take into account that the patients received previous guidance on the performance of the screening and left the hospital with the request for the exam. The results might have been different if the guidance about the importance of the OGTT and the request for it depended on the primary care unit. Conclusion The rate of DM screening through the OGTT six weeks postpartum in women diagnosed with GDM is low. Women who attended the postpartum appointment, who needed medication for the treatment of GDM, and who had comorbidities were the most adherent to screening. It is essential to develop measures to cover a greater number of patients, such as follow-up programs, monitoring to prevent these women from leaving the hospital until the exam is performed, and improvements in the access to information regarding the importance of puerperal screening for DM, mainly in the primary health care setting. Contributions All researchers participated in the following steps: substantial contributions to conception and design, data collection or data analysis and interpretation; article writing or relevant critical review of the intellectual content; and final approval of the version to be published. Conflict of Interests The authors have no conflict of interests to declare. ==== Refs References 1 International Diabetes Federation IDF Diabetes Atlas. 9th ed Brussels IDF 2019 2 Organização Pan-Americana da Saúde, Ministério da Saúde, Federação Brasileira das Associações de Ginecologia e Obstetrícia, Sociedade Brasileira de Diabetes Rastreamento e diagnóstico de diabetes mellitus gestacional no Brasil Brasília (DF) OPAS 2016 3 International Association of Diabetes and Pregnancy Study Groups Consensus Panel Metzger B E Gabbe S G Persson B Buchanan T A Catalano P A Damm P International association of diabetes and pregnancy study groups recommendations on the diagnosis and classification of hyperglycemia in pregnancy Diabetes Care 2010 33 03 676 682 10.2337/dc09-1848 20190296 4 Bellamy L Casas J P Hingorani A D Williams D Type 2 diabetes mellitus after gestational diabetes: a systematic review and meta-analysis Lancet 2009 373 (9677):1773 1779 10.1016/S0140-6736(09)60731-5 19465232 5 Vounzoulaki E Khunti K Abner S C Tan B K Davies M J Gillies C L Progression to type 2 diabetes in women with a known history of gestational diabetes: systematic review and meta-analysis BMJ 2020 369 m1361 10.1136/bmj.m1361 32404325 6 Miao Z Wu H Ren L Ren L Bu N Jiang L Yang H Long-Term Postpartum Outcomes of Insulin Resistance and β -cell Function in Women with Previous Gestational Diabetes Mellitus Int J Endocrinol 2020 2020 7.417356E6 10.1155/2020/7417356 7 American College of Obstetricians and Gynecologists Practice Bulletin No. 180: gestational diabetes mellitus Obstet Gynecol 2017 130 01 e17 e37 10.1097/AOG.0000000000002159 28644336 8 Guo J Chen J L Whittemore R Whitaker E Postpartum lifestyle interventions to prevent type 2 diabetes among women with history of gestational diabetes: a systematic review of randomized clinical trials J Womens Health (Larchmt) 2016 25 01 38 49 10.1089/jwh.2015.5262 26700931 9 Miao Z R Wu H H Zhang Y Z Sun W J Lu D F Yang H X Evaluation of the gestational diabetes mellitus diagnostic criteria recommended by the international association of diabetes and pregnancy study group for long-term maternal postpartum outcomes in mainland China Medicine (Baltimore) 2020 99 08 e19242 10.1097/MD.0000000000019242 32080127 10 Tobias D K Stuart J J Li S Chavarro J Rimm E B Rich-Edwards J Association of history of gestational diabetes with long-term cardiovascular disease risk in a large prospective cohort of US Women JAMA Intern Med 2017 177 12 1735 1742 10.1001/jamainternmed.2017.2790 29049820 11 Rotem R Salem-Yaniv S Sandler-Rahat H Yohay D Shanny S Yahav L Adherence to postpartum diabetes mellitus screening, do associated pregnancy complications make a difference? Diabetes Res Clin Pract 2020 159 107972 10.1016/j.diabres.2019.107972 31805350 12 Stasenko M Liddell J Cheng Y W Sparks T N Killion M Caughey A B Patient counseling increases postpartum follow-up in women with gestational diabetes mellitus Am J Obstet Gynecol 2011 204 06 5220 5.22E8 10.1016/j.ajog.2011.01.057 13 Tang Y Guo J Long Q Yang J Luo J Yang S Factors influencing postpartum blood glucose screening among women with prior gestational diabetes mellitus in a rural community J Adv Nurs 2020 76 08 2151 2160 10.1111/jan.14440 14 de Gennaro G Bianchi C Aragona M Battini L Baronti W Brocchi A Postpartum screening for type 2 diabetes mellitus in women with gestational diabetes: Is it really performed? Diabetes Res Clin Pract 2020 166 108309 10.1016/j.diabres.2020.108309 32650034 15 Ortiz F M Jimenez E Y Boursaw B Huttlinger K Postpartum care for women with gestational diabetes MCN Am J Matern Child Nurs 2016 41 02 116 122 10.1097/NMC.0000000000000215 26909726 16 Cabizuca C A Rocha P S Marques J V Costa T F Santos A S Schröder A L Postpartum follow up of gestational diabetes in a Tertiary Care Center Diabetol Metab Syndr 2018 10 2 10.1186/s13098-017-0303-4 29308091 17 STRONG Study Collaborators and the AMD-SID Diabetes and Pregnancy Study Group Napoli A Sciacca L Pintaudi B Tumminia A Dalfra M G Festa C Screening of postpartum diabetes in women with gestational diabetes: high-risk subgroups and areas for improvements-the STRONG observational study Acta Diabetol 2021 58 09 1187 1197 10.1007/s00592-021-01707-9 33842997 18 Farid H Blake R Hacker M R Erlinger A L Modest A M Strategies to improve postpartum glucose screening rates are needed J Obstet Gynaecol Res 2021 47 08 2641 2645 10.1111/jog.14868 34041808 19 Paul J C Fitzpatrick J J Postpartum glucose screening among women with gestational diabetes Appl Nurs Res 2020 56 151341 10.1016/j.apnr.2020.151341 33280784 20 Diabetes Prevention Program Research Group Aroda V R Christophi C A Edelstein S L Zhang P Herman W H Barrett-Connor E The effect of lifestyle intervention and metformin on preventing or delaying diabetes among women with and without gestational diabetes: the Diabetes Prevention Program outcomes study 10-year follow-up J Clin Endocrinol Metab 2015 100 04 1646 1653 10.1210/jc.2014-3761 25706240 21 Lawrence J M Black M H Hsu J W Chen W Sacks D A Prevalence and timing of postpartum glucose testing and sustained glucose dysregulation after gestational diabetes mellitus Diabetes Care 2010 33 03 569 576 10.2337/dc09-2095 20040657 22 Rosenbloom J I Blanchard M H Compliance with postpartum diabetes screening recommendations for patients with gestational diabetes J Womens Health (Larchmt) 2018 27 04 498 502 10.1089/jwh.2017.6477 29211594 23 Jones K E Yan Y Colditz G A Herrick C J Prenatal counseling on type 2 diabetes risk, exercise, and nutrition affects the likelihood of postpartum diabetes screening after gestational diabetes J Perinatol 2018 38 04 315 323 10.1038/s41372-017-0035-1 29298984 24 Butalia S Donovan L Savu A Johnson J Edwards A Kaul P Postpartum diabetes testing rates after gestational diabetes mellitus in canadian women: a population-based study Can J Diabetes 2017 41 06 613 620 10.1016/j.jcjd.2016.12.013 28506815 25 Liu Z Y Zhao J J Gao L L Wang A Y Glucose screening within six months postpartum among Chinese mothers with a history of gestational diabetes mellitus: a prospective cohort study BMC Pregnancy Childbirth 2019 19 01 134 10.1186/s12884-019-2276-9 30999888 26 Quaresima P Visconti F Chiefari E Puccio L Foti D P Venturella R Barriers to postpartum glucose intolerance screening in an italian population Int J Environ Res Public Health 2018 15 12 2853 10.3390/ijerph15122853 30558120
PMC009xxxxxx/PMC9800065.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Medical Publishers 333 Seventh Avenue, New York, NY 10001, USA. 36580953 10.1055/s-0042-1757970 RBGO-22-0273 Reply to the Letter to the Editor Authors' reply http://orcid.org/0000-0002-8691-7260 Kobayashi Carla Dinamerica 1 http://orcid.org/0000-0002-3112-6154 Porto Victor Bertollo Gomes 1 http://orcid.org/0000-0003-4615-2280 Nóbrega Martha Elizabeth Brasil da 1 http://orcid.org/0000-0003-1934-3081 Cabral Cibelle Mendes 1 http://orcid.org/0000-0001-6960-3801 Barros Tiago Dahrug 1 http://orcid.org/0000-0001-5646-3977 Martins Cecília Maria Roteli 2 1 Coordenação Geral do Programa Nacional de Imunizações, Ministério da Saúde, Brasília, DF, Brazil 2 Centro Universitário Faculdade de Medicina do ABC, Santo André, SP, Brazil Address for correspondence Carla Dinamerica Kobayashi SRTVNQuadra 701, Bloco D, Ed. PO 700, 6° andar-CGPNI, 70719-040, Brasília, [email protected] 29 12 2022 11 2022 1 12 2022 44 11 10791080 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. ==== Body pmc Adverse Events Related to COVID-19 Vaccines in Pregnant Women: Correspondence Spontaneous abortion is the most frequent adverse outcome in pregnancy, with a rate of 10% to 20% of all pregnancies. In Brazil, the rate of abortion before the COVID-19 pandemic was of 3.5%. In our study, we observed 24 miscarriages after 678,025 doses of COVID-19 vaccines had been administered in pregnant women in Brazil, which corresponds to an incidence of 3.5 cases per 100,000 doses. 1 Such incidence is lower than the postvaccination rate of abortion related to COVID-19 vaccines described by Trostle et al. 2 and much lower than the historical 3.5% expected rate in Brazil. Our study was observational and used passive surveillance data. Such studies cannot describe the exact rates of abortion in vaccinated pregnant women. Nonetheless, given the historical rates in Brazil, the expected number of abortions in 678,025 normal pregnancies would be 23,731, which averages out to 616 abortions per week of observation. Even if we consider reporting rates lower than 5%, our incidence of abortions would still be far below the expected. 3 Currently, the Brazilian Ministry of Health and the Pan American Health Organization (PAHO) are conducting a prospective active surveillance study with women who were vaccinated against COVID-19 during pregnancy in the 5 regions of the country. Such study will enable a more accurate description of the true rates of abortion and other adverse pregnancy outcomes in vaccinated women. It is already known that pregnant women have a higher risk of developing severe COVID-19 and an increased risk of having adverse maternal and neonatal outcomes after COVID-19. We believe that our data, in association with the previously-published literature, suggests that the benefits of vaccinating pregnant women against COVID-19 far outweighs the risk of the disease. 4 Conflict of Interests The authors have no conflict of interests to declare. ==== Refs References 1 Kim C Barnard S Neilson J P Hickey M Vazquez J C Dou L Medical treatments for incomplete miscarriage Cochrane Database Syst Rev 2017 1 01 CD007223Doi: 10.1002/14651858.CD007223.pub428138973 2 Trostle M E Limaye M A Avtushka V Lighter J L Penfield C A Roman A S COVID-19 vaccination in pregnancy: early experience from a single institution Am J Obstet Gynecol MFM 2021 3 06 100464Doi: 10.1016/j.ajogmf.2021.10046434411758 3 Sookaromdee P Wiwanitkit V Magnitude of abortion after COVID-19 vaccination: how about rate? Erciyes Med J 2022 44 02 244Doi: 10.14744/etd.2021.78379 4 CDC v-safe COVID-19 Pregnancy Registry Team Shimabukuro T T Kim S Y Myers T R Moro P L Oduyebo T Panagiotakopoulos L Preliminary findings of mRNA Covid-19 vaccine safety in pregnant persons N Engl J Med 2021 384 24 2273 2282Doi: 10.1056/NEJMoa210498333882218
PMC009xxxxxx/PMC9800066.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36580945 10.1055/s-0042-1760116 RBGO-44.11-Editorial Editorial Facing Osteoporosis: Is Hormonal Therapy Losing an Opportunity to be Used? The Role of Gynecologists http://orcid.org/0000-0002-4813-6404 Sá Marcos Felipe Silva de 1 1 Editor –in – Chief RBGO Address for correspondence Marcos Felipe Silva de Sá [email protected] 29 12 2022 11 2022 1 12 2022 44 11 10111013 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. ==== Body pmc In the month of October, the World Day against Osteoporosis was celebrated and the date is dedicated to the global awareness of its prevention, diagnosis and treatment. This disease affects about 200 million people worldwide, causing more than 8.9 million fractures annually. 1 2 Globally, it affects 21.2% of women over 50 years of age. 3 Fractures caused by osteoporosis have great importance not only because of their high prevalence, but also because of their serious physical, psychosocial and financial consequences that affect both individuals and their families, the community and health systems. Considering these demographic aspects of the disease, the Brazilian Federation of Gynecology and Obstetrics Associations – FEBRASO - joined other organizations and medical societies in the campaign, since climacteric women are among the population most at risk for osteoporosis, who nowadays correspond to a large percentage of patients in gynecology offices. According to surveys by DataFolha, one of the main institutes of public opinion in Brazil, eight in every ten women consider the specialty of Gynecology and Obstetrics as the most important for women's health in Brazil. 4 For brazilian women , Gynecologist-Obstetricians are considered as reference physicians, whether for treating problems of the specialty, as well as for counseling and guidance when they need healthcare in another medical specialty. 4 Thus, the gynecologist's office becomes one of the main entry points into the health system for women, and this professional should become a true sentinel in the screening for osteoporosis, seeking to identify women at risk for fractures. Therefore, gynecologists must be prepared for this task and also Also to refer the patient when the osteoporosis etiology requires joint care with another medical specialty since this disease has multiple origins and patients often require multidisciplinary follow-up. Thus the Competence Matrix for Medical Residency Programs in Gynecology and Obstetrics stablished by FEBRASGO emphasize the attention to climacteric women's health. 5 In addition, Febrasgo created the National Specialized Commission on Osteoporosis, which has been developing an excellent work, aiming to reinforce the fundamentals to approach patients at risk for osteoporosis. . With the active participation of members of that Commission, the Brazilian Osteoporosis Manual was recently launched, conceived as a practical guidelines for health professionalss. 6 Once the osteoporosis have been identified, it is essential that the physician offer to the patient all the information about the disease in order that she must be awared about the risks of suffering fracture. Thousands of women with this silent disease are unaware of these risks. Currently there are international consensus regards the necessity of the patients to change her lifestyle, adequate diet, use of calcium and vitamin D, physical exercise, avoiding alcoholic beverages, among others, in addition to supplementing with medication when necessary. Several options of pharmacological treatment are available on the market, such as Hormone Replacement Therapy (HT), Bisphosphonates (BP), Selective Estrogen Receptor Modulator (SERMs), Denosumab, Teriparatide and others still being studied. All options have advantages and disadvantages and the choice will depend on the professional's experience in dealing with each medication, considering its possible side effects and/or complications of its use. Since the disease has multiple triggering factors and patients are treated by different medical specialties, drug therapy has varied according to treatment protocols established by different specialty societies. It is known that one of the main trigger for the development of osteoporosis in women is estrogen deficiency consequent to physiological or induced ovarian failure, which determines an increase in bone resorption that is not compensated by an equivalent increase in formation. 7 The medical literature has consistently and significantly shown that HT (encompassing both estrogenic therapy and estrogen-progestin therapy ) is indicated for climacteric women when they presented with vasomotor symptoms and genitourinary syndrome of menopause. Besides that, TH may be considered to be used to prevent bone loss and fragility fractures. 8 9 10 Estrogens have a positive effect on reducing the risk of fractures of the hip, vertebrae and other related fractures in postmenopausal women. It is the only therapy available with proven randomized clinical trials presenting effectiveness in reducing fractures, even in groups of women who do not have an identified risk for fractures or who have a T-score in the normal or osteopenic range in bone mineral densitometry (BMD). 11 Considering its well-known contraindications, 10 12 the HT in climacteric can be started in women at risk of fractures or osteoporosis before the age of 60 or within the first ten years after menopause (window of opportunity). There is an international consensus supported by influential Specialty Societies that indicate its use, evidently establishing individualized safety criteria for each patient. 9 HT would be the best choice for climacteric women in that period, because in addition to the undeniable benefits on bone mass, , patients have the opportunity to the additional benefits offered by HT, such as the prevention or abolition of hot flashes, , protection against genitourinary syndrome of menopause and its consequences on sexual health, positive effects on collagen and skin and significantly improving in sleep and quality of life. 10 Since HT is a routine prescription for gynecologists, it should obviously be the first choice for these patients, considering the extra benefits mentioned above. However, contrary to what is expected, an increasing use of BP has been observed as the first choice among gynecologists rather than HT, even for those climacteric patients considered within the window of opportunity who do not have any contraindication for its use. Why gyneologists are missing this opportunity to prescribe HT? Several factors may be contributing to this behavior. The first to be cited would still be the impact of the Women's Health Initiative (WHI) study published two decades ago, which raised fears among physicians and patients about the risk of the association between HT and breast cancer and cardiovascular diseases. 13 It is known that many patients refuse to use HT for fear of breast cancer, often as a result of misinformation or because they receive distorted information from their own doctors , a common fact in current times, where social networks have negatively contributed to the dissemination of information from unqualified origin that reaches both, patients and healthcare professionals. It is important that the physician has up-to-date information on these topics , through reliable sources to better guide patients about the real risks of its use. The negative impact of the WHI study has been revised in recent publications as its original data have been reviewed in more detail. When subdividing patients by age groups and analyzing separately the effects of therapy with estrogen alone versus the estrogen-progestin combination the results show that the risks within the window of opportunity period are minimized, with the benefits of HT being greater than the risks of its use 14 . It should be noted that in the WHI study the mean age of evaluated patients was 63 years (including patients up to 79 years old), 13 therefore, well above the age currently suggested in the international literature for the introduction of HT. HT may be safe for a period of five years, which can be extended to ten years, depending on the patient's response to treatment, always under careful supervision of the attending physician. 10 The fact that hormone doses recommended nowadays are much lower compared to those used two decades ago should also be taken into account. There are different therapeutic schemes with proven cost-effectiveness and available in most countries including Brazil. 10 12 There is a worldwide trend towards its use by the non-oral route given the lower risks and side effects. In their Medical Residency Program , gynecologists received guidance and practices for the use of HT, including training to deal with its side effects, especially the management of abnormal uterine bleeding, which are frequent and constitute additional difficulties for prescribers from other medical specialties who may assist these patients. Perhaps this is the main reason for non-gynecological specialists to choose therapies other than HT therapy and . in this sense, the gynecologist may have an advantage to safer prescribe the HT. It is worth adding that the possible risks of HT disappear quickly when its use is discontinued, unlike BP, which prolonged use, particularly beyond five years, compromises the bone structure and decreases its resistance, with the risk of significant adverse effects, since its residual effects may persist for several years after discontinuation. 15 Doctors specialists in Gynecology should reflect on this issue and, in order to avoid abuses in its prescription, it is good to remember that HT should not be recommended without a clear indication for its use and must be in accordance with the acceptance of the patient. and her priorities in terms of health as opposed to personal risks aiming her quality of life Conflicts to Interest None to declare. ==== Refs References 1 Lane N E Epidemiology, etiology, and diagnosis of osteoporosis Am J Obstet Gynecol 2006 194 (2, Suppl)S3 S11 10.1016/j.ajog.2005.08.047 16448873 2 Johnell O Kanis J A An estimate of the worldwide prevalence and disability associated with osteoporotic fractures Osteoporos Int 2006 17 12 1726 1733 10.1007/s00198-006-0172-4 16983459 3 Kanis J A McCloskey E V Johansson H Oden A Melton L J III Khaltaev N A reference standard for the description of osteoporosis Bone 2008 42 03 467 475 10.1016/j.bone.2007.11.001 18180210 4 Martins L Reconhecimento nacional Femina 2019 47 03 131 135 5 Romão G S Silva de Sá M F Competency-based training and the competency framework in gynecology and obstetrics in Brazil Rev Bras Ginecol Obstet 2020 42 05 272 288 10.1055/s-0040-1708887 32483808 6 Pedro A O Plapler P G Szejnfeld V L Manual brasileiro de osteoporose: orientações práticas para os profissionais de saúde [Internet] São Paulo Clannad 2021[cited 2022 Nov 12]. Available from:https://www.editoraclannad.com.br/wp-content/uploads/2022/11/Manual-Brasileiro-de-Osteoporose_Versao_04NOV22.pdf 7 Fernandes C E Steiner M L Pompei L M Strufaldi R Terapia hormonal e SERMs São Paulo Clannad 2021[cited 2022 Nov 12]. p. 91-5. Available from:https://www.editoraclannad.com.br/wp-content/uploads/2022/11/Manual-Brasileiro-de-Osteoporose_Versao_04NOV22.pdf 8 IMS Writing Group Baber R J Panay N Fenton A 2016 IMS Recommendations on women's midlife health and menopause hormone therapy Climacteric 2016 19 02 109 150 10.3109/13697137.2015.1129166 26872610 9 de Villiers T J Hall J E Pinkerton J V Revised global consensus statement on menopausal hormone therapy Maturitas 2016 91 153 155 10.1016/j.maturitas.2016.06.001 27389038 10 NAMS Position Statement. The 2022 hormone therapy position statement of The North American Menopause Society Menopause 2022 29 07 767 794 35797481 11 Zhu L Jiang X Sun Y Shu W Effect of hormone therapy on the risk of bone fractures: a systematic review and meta-analysis of randomized controlled trials Menopause 2016 23 04 461 470 10.1097/GME.0000000000000519 26529613 12 Pompei L M Quais são as reais contraindicações para a terapêutica hormonal? São Paulo Leitura Médica 2018p. 141-5. 13 Writing Group for the Women's Health Initiative Investigators Rossouw J E Anderson G L Prentice R L Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results From the Women's Health Initiative randomized controlled trial JAMA 2002 288 03 321 333 10.1001/jama.288.3.321 12117397 14 Manson J E Chlebowski R T Stefanick M L Menopausal hormone therapy and health outcomes during the intervention and extended poststopping phases of the Women's Health Initiative randomized trials JAMA 2013 310 13 1353 1368 10.1001/jama.2013.278040 24084921 15 Ensrud K E Crandall C J Bisphosphonates for postmenopausal osteoporosis JAMA 2019 322 20 2017 2018 10.1001/jama.2019.15781 31621799
PMC009xxxxxx/PMC9800067.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36580949 10.1055/s-0042-1759630 RBGO-22-0002 Original Article Assisted Fertilization Comparison of the Effects of GMCSF-Containing and Traditional Culture Media on Embryo Development and Pregnancy Success Rates http://orcid.org/0000-0002-5535-6296 Adanacıoglu Fatih 1 http://orcid.org/0000-0001-6733-592X Çetin Çağlar 2 http://orcid.org/0000-0001-6084-4043 Tokat Gozde 3 http://orcid.org/0000-0003-4578-0939 Adanacıoglu Durdane 4 http://orcid.org/0000-0001-7480-4691 Karasu Ayse Filiz Gokmen 2 http://orcid.org/0000-0003-4048-4882 Çetin Mehmet Turan 1 1 Turan Cetin Private IVF Center, Adana, Turkey 2 Department of Obstetrics and Gynecology, Bezmialem Vakif University School of Medicine, Istanbul, Turkey 3 Department of Embryology, Turan Cetin Private IVF Center, Adana, Turkey 4 Adana State Hospital, Adana, Turkey Address for correspondence Ayse Filiz Gokmen Karasu İskender Paşa Mh, Adnan Menderes Blv.34093 Fatih/İ[email protected] 29 12 2022 11 2022 1 12 2022 44 11 10471051 02 1 2022 09 9 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  The use of granulocyte macrophage colony-stimulating factor (GM-CSF)-containing medium, which is a commercial medium that is used for cultivation of embryos in in vitro fertilization (IVF) treatments, has been suggested to increase the efficiency of this procedure in patients with previous multiple unsuccessful attempts. In this retrospective study, we analyzed GM-CSF-containing embryo culture media compared with traditional culture media in terms of development of embryos, pregnancy, and ongoing pregnancy success and live birth rates. Methods  This is a prospective case control study conducted in a single center. A total of 131 unexplained infertility patients were included in the study. A cohort of 69 patients whose embryos were cultured in GM-CSF-containing medium and a control group of 62 age-matched patients whose embryos were cultured in conventional Sage One Step medium were included in the study. The major study outcomes were achievement of pregnancy and ongoing pregnancy rate at 12 weeks of gestation. Results  The pregnancy and ongoing pregnancy rates of the patients whose embryos were cultured in GM-CSF-containing medium were 39.13% and 36.23%, respectively. These were higher than the rates of the control group, which were 30.65% and 29.03%, respectively, although this difference was not statistically significant. In addition, the 5 th -day embryo transfer percentage in the GM-CSF group was higher than in the control group (34.78% versus 27.4%). Conclusion  The main findings of our study were that there was no difference between the GM-CSF-enhanced medium and the control group in terms of our major study outcomes. However, blastomere inequality rate and embryo fragmentation rates were lower in the GM-CSF group. Keywords in vitro fertilization embryo culture embryo transfer granulocyte macrophage colony-stimulating factor ==== Body pmcIntroduction Ongoing research in assisted reproductive technology, especially in vitro fertilization (IVF) methods, has increased the efficiency of treatment. The important problems associated with IVF treatment include implantation failures and miscarriages. The arrest of embryonic development results in miscarriage in some IVF-assisted pregnancies. Repeated implantation failure, which occurs when embryos fail to implant following several IVF treatment cycles, can be caused by reduced sensitivity of the endometrium, genetic factors, or stress. The success of implantation in the IVF process depends on three main parameters: a healthy embryo, a receptive endometrium, and a strong communication between these two to facilitate the subsequent events necessary for pregnancy. 1 The communication between the embryo and the endometrium is made possible by the interaction of cytokines, which are necessary for the regulation of normal embryonic development, enhancing implantation efficiency and normal development of the fetus and the placenta. 2 3 4 5 There are some differences between the in vitro culture conditions utilized during the IVF process and the in vivo conditions. 6 7 8 9 10 The lack of necessary ingredients and factors during in vitro culture may cause recurrent implantation failure and other clinical pathologies. 7 11 12 Since the culture conditions and the ingredients of the in vitro culture medium are important to ensure the success of the IVF treatment, there is ongoing research to find the optimal conditions for the in vitro culture and its effect on the embryo development and the health of the baby. Addition of the cytokines that are responsible for the embryo-endometrium communication to the culture medium during IVF has been shown to be important in normal blastocyst development, embryo implantation, and development of the fetus and placenta. 13 14 15 16 Numerous cytokines and growth factors which are secreted by the endometrium have exhibited in stimulating normal embryo development. Among the above-mentioned factors, there has been a focus on granulocyte-macrophage colony-stimulating factor (GM-CSF), also known as CSF (colony stimulating factor). Granulocyte-macrophage colony-stimulating factor is expressed in the epithelial cells lining the oviduct and uterus 17 and enhances embryo implantation and development of a normal embryo. 18 19 20 21 22 The embryos, which invade placental trophoblast cells and the abundant populations of leukocytes controlling maternal immune tolerance, are affected by GM-CSF adjustment. Granulocyte-macrophage colony-stimulating factor deficiency in pregnancy negatively alters fetal and placental development. Additionally, GM-CSF regulation is also important in growth after birth; thus, this cytokine is a central maternal determinant of pregnancy outcome with clinical relevance in human fertility. Preclinical studies show that GM-CSF has positive effects on development and implantation efficiency in mouse and human embryos. 20 21 22 During early pregnancy, GM-CSF in blood serum increases dramatically, and the levels decrease if pregnancy is terminated. 23 Reduced GM-CSF levels have been associated with miscarriage. 4 Research on the use of GM-CSF in the IVF process has shown promising results. 16 The addition of recombinant GM-CSF to the culture medium has been shown to increase the success rate of IVF. 24 25 26 Methods The institutional review board (71/10/2016) approved the present study. Informed verbal and written consent were obtained from all participating couples. A total of 131 unexplained infertility patients who received IVF treatment were chosen to participate in the study or the control group. Embryos of 69 of these patients were treated with GM-CSF-containing medium, whereas 62 were treated with conventional medium. All embryos in the control group were incubated with single-step culture medium (without GM-CSF). For the control group selection, matching was done considering female age, body mass index, number of M-II oocyte retrieved, and number of embryos transferred. All patients were treated between 2016 and 2018. The sperm parameters of all patients who participated in this study were evaluated according to the World Health Organization standards, and morphology was assessed by the Kruger criteria. The total sperm counts of all patients were over 15 million. The mean forward mobility was over 32%. The sperm morphologies were between 3 and 7%, according to the Kruger criteria. All patients had normal hysterosalpingography (HSG). The exclusion criteria were as follows: patients carrying genetic anomaly risk; those who had mature but unfertilized oocytes; those who had immature or low-quality oocytes; and patients with uterine cervical insufficiency, thrombophilia, and sperm morphology lower than 3%, according to the Kruger strict criteria. After oocyte pick up, the oocytes were either transferred to a GM-CSF-containing medium or Sage One Step medium (CooperSurgical, Denmark). In the study group, the oocytes that were collected at day 0 were transferred to GM-CSF-containing medium drops with 35 µL volume. Four hours after oocyte pick-up, denudation and intracytoplasmic sperm injection was performed. On day 1 after fertilization control, the zygotes were transferred to fresh GM-CSF-containing medium drops (Embryogen, Origio, Denmark). After this transfer, their development was monitored daily. In the control group, the oocytes that were collected on day 0 were placed in 35 µL Sage One Step droplets that were coated with paraffin oil. The oocytes were denuded after 4 hours, and an intracytoplasmic sperm injection was applied. After checking for fertilization on day 1, the zygotes were transferred to fresh Sage One Step drops, and their development was monitored daily. The day of embryo transfer was determined according to the development of embryos. The maximum number of transferred embryos was two per patient. Embryos were assessed morphologically at cleavage stage (day ⅔) on the basis of number and symmetry of blastomeres, degree of fragmentation, and presence of multinucleated cells according to the British Fertility Society and Association of Clinical Embryologists guidelines, published by Cutting et al. (2008). 27 After oocyte pick-up, 90 mg progesterone gel was applied vaginally the next day. On the second day posttransfer, 2 mg of estradiol was applied transdermally by using a 6.5 cm 2 patch. The day of embryo transfer was determined according to the development of the embryos. Intramuscular daily injections of 50 mg of progesterone and peroral application of 4 mg of methyl prednisolone were performed for a 12-day period after embryo transfer. On day 12 of the transfer, β-HCG levels were tested in serum blood samples obtained from the patients. The GraphPad Prism software was used for the statistical analyses. The difference between groups was analyzed using the Chi-squared test. The test results were presented with 95% confidence interval. A p -value < 0.05 was set as statistically significant. All data are presented as average ± standard deviation (SD). Results The effect of culturing embryos in GM-CSF-containing medium was studied in a total of 131 patients who underwent IVF treatment in the center. Sixty-nine of the patients were in the GM-CSF-containing medium group, whereas 62 patients constituted the control group whose embryos were cultured in Sage One Step. The ages of the patients ranged between 23 and 38. The average ages of the control group and GM-CSF-containing medium group were 30.61 ± 3.46 years and 30.73 ± 3.89 years, respectively. The endometrial thickness values of both groups were also found to be comparable ( Table 1 ). Also, the pregnancy outcomes and live birth ratios of the study group are demonstrated in Table 2 . Table 1 The pregnancy outcomes and clinical information of the study group Protocol # of Patients Mean age ± SD # of transferred embryos Endometrial thickness GM-CSF containing medium 69 30.73 ± 3.89 1.91 ± 0.28 9.36 ± 1.03  • Pregnant 27 (39.13%) 30.18 ± 4.12 1.96 ± 0.19 9.48 ± 0.75  • Not pregnant 42 (60.87%) 31.72 ± 5.62 1.91 ± 0.30 9.13 ± 0.99 No GM-CSF containing medium 62 30.61 ± 3.46 1.89 ± 0.32 9.37 ± 1.11  • Pregnant 19 (30.64%) 30.89 ± 3.90 1.79 ± 0.42 9.74 ± 1.10  • Not pregnant 43 (69.36%) 30.49 ± 3.29 1.86 ± 0.37 9.22 ± 1.52 Table 2 The pregnancy outcomes and live birth ratios of the study group Protocol Pregnancy Ongoing pregnancy Live birth Twin birth GM-CSF-containing medium 27 (39.13) 25 (36.23) 25 (36.23) 4 (5.79) Non-GM-CSF-containing medium 19 (30.64) 18 (29) 18 (29) 2 (1.24) The culture medium containing GM-CSF is usually recommended for IVF patients with a history of multiple unsuccessful treatment attempts. We analyzed the number of previous unsuccessful attempts for the GM-CSF-containing medium group and found that the average number of previous unsuccessful attempts was 3.03 ± 1.51. The average number of previous unsuccessful IVF attempts for the control group, on the other hand, was 3.00 ± 0.54. The inequality rates of blastomeres in the GMCSF were lower than in the control group. In the GMCSF group it was 10%, and in the control group the inequality rate was 40%. In addition, in the GM-CSF-containing medium group, 80% of the embryos had less than 5% fragmentation, while in the control group 50% of the embryos had less than 5% fragmentation. Among the control group, the ongoing pregnancy rate at 12 weeks was 30.65%, whereas this rate was 39.13% for the GM-CSF-containing medium group. However, the difference between groups was not statistically significant according to the Chi-squared test with a p -value of 0.3106 and an odds ratio of 0.6873 and 95% confidence interval of 0.3330 to 1.4188. Discussion The main findings of our study were that there was no difference between the GM-CSF-enhanced medium and the control group in terms of our major study outcomes. However, blastomere inequality rate and embryo fragmentation rates were less in the GM-CSF group. The success of the IVF treatment depends on, among many other factors, the environmental conditions of the embryo before implantation. The presence of growth factors results in increased efficiency of preimplantation embryo development. Among these growth factors, addition of GM-CSF into the culture medium has been shown to have favorable outcomes in preimplantation development and clinical pregnancy rates. Ziebe et al. 26 showed an increase in ongoing implantation rate in women with high incidences of miscarriage, from 17 to 24.5% when GM-CSF-containing medium was used. Among women who were not classified based on previous IVF failures, the ongoing implantation rates at 12 weeks were 23.0% (GM-CSF) and 18.7% (control). The live birth rates were 28.9% (GM-CSF) and 24.1%(control) in the same cohort, respectively. We also observed that the pregnancy rate increased from 29.032 to 39.13% in the GM-CSF-containing medium group compared with control group, but this difference was not statistically significant. This result is in line with the other studies mentioned above ( Fig. 1 ). Also similarly, in our study, the implantation rate was higher in the GM-CSF group too (12 weeks 39.13% vs 30.65%), and the live birth rates were 36.23% (GM-CSF) and 29%(control) in the same cohort, respectively. Fig. 1 The distribution of single births, miscarriages, and twin births among pregnant women who underwent IVF using either GM-CSF-containing medium or Sage One Step. Mignini Renzini et al. 24 showed that miscarriage rates were reduced, and live birth rates were increased in women with a history of miscarriage; however, there was no significant difference between the clinical pregnancy rates. The patient population in our study did not experience pregnancy before, there was no history of miscarriage. However, we observed that embryos developed better in GM-CSF medium, which may decrease the miscarriage rates. Sfontouris et al. 21 observed that clinical pregnancy rates (35.3% vs 22.9%), and implantation rates (12 weeks, 17.4% vs 11.4%) were higher in the GM-CSF containing medium group compared with controls, but these differences were not statistically significant. Similarly, in our study, the implantation rate was higher in the GM-CSF group (12 weeks 39.13% vs 30.65%). In another study, Tevkin et al. 28 showed that the frequency of clinical pregnancy increased, and early pregnancy losses decreased in the GM-CSF-containing medium treated group compared with the control group. 27 28 These results suggest that the use of GM-CSF increases the efficiency of IVF treatment in patients with multiple unsuccessful IVF attempts. It was suggested that the transfer of blastocysts to the uterus may lead to higher implantation rates and, thus, reduce the levels of multiple births after IVF treatment. 16 However, Mignini Renzini et al. 24 observed a higher live birth rate (19.5% versus 9.5%) and an increase in twin births in the GM-CSF-treated group, which is similar to our results (5.79%vs 1.24%). The reason for this appears to be a decrease in miscarriage rates; however, there may also be another mechanism involved. More studies will be important to define the mechanism of action through which GM-CSF may affect pregnancy and live birth rates. Singleton and twin births rates in our study are shown in Fig. 2 . Fig. 2 The distribution of pregnant versus non-pregnant patients who underwent IVF using either GM-CSF-containing medium or Sage One Step and their live birth ratios. The major limitation of our study is that the study population was small. However, the strengths of our study include prospective design, appropriate follow-up, and reporting of ongoing pregnancy results. Overall, our results are in line with previous observations showing that the addition of GM-CSF into the culture media increases clinical pregnancy by better mimicking the natural in vivo conditions and helping the embryo to better adapt to the in vitro conditions. Our study strengthens the view that clinical pregnancy and ongoing pregnancy rates after IVF treatment in patients with multiple failures may be enhanced by the use of GM-CSF-containing medium in vitro. Conclusion The main findings of our study were that there was no difference between the GM-CSF-enhanced medium and the control group in terms of our major study outcomes. However, blastomere inequality rate and embryo fragmentation rates were lower in the GM-CSF group. Contributions Conflict of Interests The authors have no conflict of interests to declare. All authors participated in the concept and design of the present study; analysis and interpretation of data; draft or revision of the manuscript, and they have approved the manuscript as submitted. All authors are responsible for the reported research. ==== Refs References 1 Simon A Laufer N Assessment and treatment of repeated implantation failure (RIF) J Assist Reprod Genet 2012 29 11 1227 1239 10.1007/s10815-012-9861-4 22976427 2 Donaghay M Lessey B A Uterine receptivity: alterations associated with benign gynecological disease Semin Reprod Med 2007 25 06 461 475 10.1055/s-2007-991044 17960531 3 Laird S M Tuckerman E M Li T C Cytokine expression in the endometrium of women with implantation failure and recurrent miscarriage Reprod Biomed Online 2006 13 01 13 23 10.1016/s1472-6483(10)62011-1 16820103 4 Robertson S A GM-CSF regulation of embryo development and pregnancy Cytokine Growth Factor Rev 2007 18 (3-4):287 298 10.1016/j.cytogfr.2007.04.008 17512774 5 Segerer S Kammerer U Kapp M Dietl J Rieger L Upregulation of chemokine and cytokine production during pregnancy Gynecol Obstet Invest 2009 67 03 145 150 10.1159/000174291 19023215 6 Chronopoulou E Harper J C IVF culture media: past, present and future Hum Reprod Update 2015 21 01 39 55 10.1093/humupd/dmu040 25035437 7 Dumoulin J C Land J A Van Montfoort A P Nelissen E C Coonen E Derhaag J G Effect of in vitro culture of human embryos on birthweight of newborns Hum Reprod 2010 25 03 605 612 10.1093/humrep/dep456 20085915 8 Hansen P J Dobbs K B Denicol A C Programming of the preimplantation embryo by the embryokine colony stimulating factor 2 Anim Reprod Sci 2014 149 (1-2):59 66 10.1016/j.anireprosci.2014.05.017 24954585 9 Kawamura K Chen Y Shu Y Cheng T Qiao J Behr B Promotion of human early embryonic development and blastocyst outgrowth in vitro using autocrine/paracrine growth factors PLoS One 2012 7 11 e49328 10.1371/journal.pone.0049328 23152897 10 Kwak S S Jeung S H Biswas D Jeon Y B Hyun S H Effects of porcine granulocyte-macrophage colony-stimulating factor on porcine in vitro-fertilized embryos Theriogenology 2012 77 06 1186 1197 10.1016/j.theriogenology.2011.10.025 22153263 11 Das M Holzer H E Recurrent implantation failure: gamete and embryo factors Fertil Steril 2012 97 05 1021 1027 10.1016/j.fertnstert.2012.02.029 22425200 12 Hansen M Kurinczuk J J Bower C Webb S The risk of major birth defects after intracytoplasmic sperm injection and in vitro fertilization N Engl J Med 2002 346 10 725 730 10.1056/NEJMoa010035 11882727 13 Agarwal A Durairajanayagam D du Plessis S S Utility of antioxidants during assisted reproductive techniques: an evidence based review Reprod Biol Endocrinol 2014 12 112 10.1186/1477-7827-12-112 25421286 14 Binder N K Evans J Gardner D K Salamonsen L A Hannan N J Endometrial signals improve embryo outcome: functional role of vascular endothelial growth factor isoforms on embryo development and implantation in mice Hum Reprod 2014 29 10 2278 2286 10.1093/humrep/deu211 25124669 15 Colazingari S Fiorenza M T Carlomagno G Najjar R Bevilacqua A Improvement of mouse embryo quality by myo-inositol supplementation of IVF media J Assist Reprod Genet 2014 31 04 463 469 10.1007/s10815-014-0188-1 24526355 16 Sjöblom C Wikland M Robertson S A Granulocyte-macrophage colony-stimulating factor promotes human blastocyst development in vitro Hum Reprod 1999 14 12 3069 3076 10.1093/humrep/14.12.3069 10601098 17 Giacomini G Tabibzadeh S S Satyaswaroop P G Bonsi L Vitale L Bagnara G P Epithelial cells are the major source of biologically active granulocyte macrophage colony-stimulating factor in human endometrium Hum Reprod 1995 10 12 3259 3263 10.1093/oxfordjournals.humrep.a135899 8822455 18 Agerholm I Loft A Hald F Lemmen J G Munding B Sørensen P D Ziebe S Culture of human oocytes with granulocyte-macrophage colony-stimulating factor has no effect on embryonic chromosomal constitution Reprod Biomed Online 2010 20 04 477 484 10.1016/j.rbmo.2009.12.026 20129825 19 Imakawa K Carlson K D McGuire W J Christenson R K Taylor A Enhancement of ovine trophoblast interferon by granulocyte macrophage-colony stimulating factor: possible involvement of protein kinase C J Mol Endocrinol 1997 19 02 121 130 10.1677/jme.0.0190121 9343304 20 Robertson S A Sjöblom C Jasper M J Norman R J Seamark R F Granulocyte-macrophage colony-stimulating factor promotes glucose transport and blastomere viability in murine preimplantation embryos Biol Reprod 2001 64 04 1206 1215 10.1095/biolreprod64.4.1206 11259269 21 Sfontouris I A Lainas G T Anagnostara K Kolibianakis E M Lainas T G Effect of granulocyte-macrophage colony-stimulating factor (GM-CSF) on pregnancy rates in patients with multiple unsuccessful IVF attempts Hum Reprod 2013 28 (Suppl 1):i62 10.1093/humrep/det173 22 Sjöblom C Roberts C T Wikland M Robertson S A Granulocyte-macrophage colony-stimulating factor alleviates adverse consequences of embryo culture on fetal growth trajectory and placental morphogenesis Endocrinology 2005 146 05 2142 2153 10.1210/en.2004-1260 15705781 23 Perricone R De Carolis C Giacomelli R Guarino M D De Sanctis G Fontana L GM-CSF and pregnancy: evidence of significantly reduced blood concentrations in unexplained recurrent abortion efficiently reverted by intravenous immunoglobulin treatment Am J Reprod Immunol 2003 50 03 232 237 10.1034/j.1600-0897.2003.00083.x 14629028 24 Mignini Renzini M Dal Canto M Coticchio G Clinical efficiency and perinatal outcome of ART cycles following embryo culture in the presence of GM-CSF in patients with miscarriage or early pregnancy loss history Hum Reprod 2013 28 (Suppl 1):i160 i202 25 Siristatidis C Vogiatzi P Salamalekis G Creatsa M Vrachnis N Glujovsky D Granulocyte macrophage colony stimulating factor supplementation in culture media for subfertile women undergoing assisted reproduction technologies: a systematic review Int J Endocrinol 2013 2013 704967 10.1155/2013/704967 23509457 26 Ziebe S Loft A Povlsen B B Erb K Agerholm I Aasted M A randomized clinical trial to evaluate the effect of granulocyte-macrophage colony-stimulating factor (GM-CSF) in embryo culture medium for in vitro fertilization Fertil Steril 2013 99 06 1600 1609 10.1016/j.fertnstert.2012.12.043 23380186 27 BFS and ACE Cutting R Morroll D Roberts S A Pickering S Rutherford A Elective single embryo transfer: guidelines for practice British Fertility Society and Association of Clinical Embryologists Hum Fertil (Camb) 2008 11 03 131 146 10.1080/14647270802302629 18766953 28 Tevkin S Lokshin V Shishimorova M Polumiskov V The frequency of clinical pregnancy and implantation rate after cultivation of embryos in a medium with granulocyte macrophage colony-stimulating factor (GM-CSF) in patients with preceding failed attempts of ART Gynecol Endocrinol 2014 30 (Suppl 1):9 12 10.3109/09513590.2014.945767 25200819
PMC009xxxxxx/PMC9800068.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36580950 10.1055/s-0042-1751286 220060 Original Article Mastology Oncological Outcomes of Nipple-Sparing Mastectomy in an Unselected Population Evaluated in a Single Center Desfechos oncológicos da mastectomia poupadora do complexo areolo-mamilar em uma população não selecionada avaliada em um único centrohttp://orcid.org/0000-0002-7156-2890 Cavalcante Francisco Pimentel 1 http://orcid.org/0000-0002-8123-5682 Araújo Mirella Macedo Parente 2 http://orcid.org/0000-0002-7885-225X Veras Igor Moreira 3 http://orcid.org/0000-0003-4145-8598 Freitas-Junior Ruffo 4 1 Hospital Geral de Fortaleza, Fortaleza, CE, Brazil 2 Grupo de Educação e Estudos Oncológicos, Fortaleza, CE, Brazil 3 Centro Regional Integrado de Oncologia, Fortaleza, CE, Brazil 4 Department of Obstetrics and Gynecology, Universidade Federal de Goiás, Goiânia, GO, Brazil Address for correspondence Francisco Pimentel Cavalcante, MD Rua Manuel Jacaré136/1401, Fortaleza 60175-110, Ceará[email protected] 29 12 2022 11 2022 1 12 2022 44 11 10521058 16 2 2022 02 5 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  Nipple-sparing mastectomy (NSM) has been traditionally used in selected cases with tumor-to-nipple distance > 2 cm and negative frozen section of the base of the nipple. Recommending NSM in unselected populations remains controversial. The present study evaluated the oncological outcomes of patients submitted to NSM in an unselected population seen at a single center. Methods  This retrospective cohort study included unselected patients with invasive carcinoma or ductal carcinoma in situ (DCIS) who underwent NSM in 2010 to 2020. The endpoints were locoregional recurrence, disease-free survival (DFS), and overall survival (OS), irrespective of tumor size or tumor-to-nipple distance. Results  Seventy-six patients (mean age 46.1 years) (58 invasive carcinomas/18 DCIS) were included. The most invasive carcinomas were hormone-positive (60%) (HER2 overexpression: 24%; triple-negative: 16%), while 39% of DCIS were high-grade. Invasive carcinomas were T2 in 66% of cases, with axillary metastases in 38%. Surgical margins were all negative. All patients with invasive carcinoma received systemic treatment and 38% underwent radiotherapy. After a mean of 34.8 months, 3 patients with invasive carcinoma (5.1%) and 1 with DCIS (5.5%) had local recurrence. Two patients had distant metastasis and died during follow-up. The 5-year OS and DFS rates for invasive carcinoma were 98% and 83%, respectively. Conclusion  In unselected cases, the 5-year oncological outcomes following NSM were found to be acceptable and comparable to previous reports. Further studies are required. Resumo Objetivo  A mastectomia poupadora do complexo areolo-mamilar (MPM) tem sido tradicionalmente utilizada em casos selecionados com distância tumor-mamilo > 2 cm e biópsia de congelação da base do mamilo negativa. Recomendar MPM em populações não selecionadas continua controverso. Este estudo avaliou os resultados oncológicos de pacientes submetidas à MPM em uma população não selecionada atendida em um único centro. Métodos  Coorte retrospectivo incluindo pacientes não selecionadas com carcinoma invasivo ou carcinoma ductal in situ (CDIS) submetidas à MPM entre 2010 e 2020. Os desfechos incluíram: recorrência locorregional, sobrevida livre de doença (SLD) e sobrevida global (SG), independentemente do tamanho do tumor ou da distância tumor-mamilo. Resultados  Setenta e seis pacientes (média: 46,1 anos de idade) (58 carcinomas invasivos/18 CDIS) foram incluídas. A maioria dos carcinomas invasivos era hormônio-positivo (60%) (superexpressão de HER2: 24%; triplo-negativo: 16%), enquanto 39% dos CDIS eram de alto grau histológico. Os carcinomas invasivos foram T2 em 66% dos casos, com metástases axilares em 38%. As margens cirúrgicas foram todas negativas. Todas as pacientes com carcinoma invasivo receberam tratamento sistêmico e 38% receberam radioterapia. Após um período médio de 34,8 meses, 3 pacientes com carcinoma invasivo (5,1%) e 1 com CDIS (5,5%) apresentaram recidiva local. Durante o acompanhamento, duas pacientes tiveram metástase à distância e vieram a óbito. As taxas de SG e SLD aos 5 anos para carcinoma invasivo foram de 98% e 83%, respectivamente. Conclusão  Em casos não selecionados, os resultados oncológicos de 5 anos após MPM foram considerados aceitáveis e comparáveis a resultados anteriores. Estudos adicionais são necessários. Keywords mastectomy nipple-sparing mastectomy breast neoplasms subcutaneous mastectomy segmental mastectomy Palavras-chave mastectomia mastectomia poupadora do complexo areolo-mamilar neoplasias mamárias mastectomia subcutânea mastectomia segmentar ==== Body pmcIntroduction The surgical treatment of early breast cancer has progressed significantly over recent decades. The concept of radical mastectomy has gradually given way to the development of breast-conserving surgery. Evidence from several randomized clinical trials shows that the survival rates after breast-conserving surgery are equivalent to those found with mastectomy. 1 2 3 4 5 6 Skin-sparing mastectomy (SSM) and nipple-sparing mastectomy (NSM), traditionally recommended for selected cases in which tumor-to-nipple distance is > 2 cm, also represented a major progress in the management of breast cancer, facilitating immediate breast reconstruction and becoming increasingly popular due to the more satisfactory cosmetic outcome. 7 8 Nevertheless, since no prospective randomized studies have been conducted to evaluate the oncological safety of these techniques, their use in clinical practice is based predominantly on retrospective studies, many of which involve small sample sizes. 9 10 11 12 13 14 Therefore, numerous questions remain regarding expanding the criteria for recommending NSM. Controversial issues include its use in locally advanced cases, the criterion for tumor-to-nipple distance, the routine use of frozen section of the base of the nipple, the need for postmastectomy radiotherapy, and the thickness of the skin flap. 15 16 17 18 19 The primary objective of the present study was to evaluate the oncological outcomes of patients submitted to NSM in an unselected population seen at a single Brazilian institute. Methods This cross-sectional retrospective study evaluated a cohort of unselected patients with invasive carcinoma or ductal carcinoma in situ (DCIS) in whom NSM had been performed between 2010 and 2020. All patients were operated by the same surgical team in an institute working exclusively within Brazil's public healthcare system (SUS, or Sistema Único de Saúde ). The institute's internal review board approved the study protocol under reference CAAE: 42697221.8.0000.5040. Since the data were collected retrospectively and anonymously, no specific written informed consent was required. The recommendation for NSM was made on an outpatient basis following physical examination and imaging evaluation, normally mammography and ultrasonography, with no breast magnetic resonance imaging (MRI) being performed. Neither tumor size nor tumor-to-nipple distance was taken into consideration in the recommendation for NSM irrespective of whether the patient had an invasive carcinoma or DCIS. On imaging evaluation, tumor-to-nipple distance was ≤ 2 cm in all cases. Patients who had previously had breast cancer or had Paget disease of the nipple, and those with metastatic disease at diagnosis were excluded, as were patients who had undergone a free nipple graft and those with incomplete data on their hospital records. The surgical technique used in NSM was individualized for each patient based on the subcutaneous plane of the breast, and the minimum thickness of the flap was not stipulated. In general, the type of incision in NSM depended on breast size/volume, on ensuring the best possible cosmetic outcome, and on oncological criteria. The skin flap was meticulously dissected using an electric scalpel after the superficial fascia had been identified, according to anatomical and oncological criteria, up to the limits of the breast silhouette, previously identified with the patient in a seated position. Frozen section analysis of the base of the nipple was not routinely performed; however, the nipple margin was always evaluated at histopathology of the surgical specimen, together with all the margins in the specimen. Clear margins were defined as “no ink on tumor” for invasive tumors, and as free margins greater than 2 mm for DCIS. Surgical evaluation of the axilla was carried out through a small separate incision in the axilla for identification of the sentinel lymph node and for performing axillary dissection whenever necessary. The decision regarding the type of immediate breast reconstruction to be performed was made at the discretion of the surgical team. The indication for systemic treatment and for radiotherapy was based on histopathological and immunohistochemical criteria, following the recommendations laid out in the international guidelines. Imaging tests were not used to measure flap thickness postoperatively. Data such as age, the date of surgery, the duration of follow-up after surgery, and the type of incision (inframammary fold, periareolar, radial or other) were collected. In addition, tumor staging was evaluated according to the eighth edition of the American Joint Committee on Cancer (AJCC) Cancer Staging Handbook: tumor size (T), final lymph node status (N), histological type, and whether the disease was invasive or in situ (low or high-grade). In the case of upfront surgery, T and N status were assessed through histological analysis of the surgical specimen, while in cases in which systemic neoadjuvant therapy was used, the initial stage prior to treatment was the factor taken into consideration. The subtype of tumor in cases of invasive disease was classified according to the results of immunohistochemistry performed on the biopsy specimen: patients expressing hormone receptors and who did not have HER2 overexpression (HER2) were considered hormone-positive or luminal, while patients with HER2 expression were classified as such, irrespective of hormone receptor status, and cases with no hormone receptor expression and no overexpression of HER2 were considered triple-negative. Any systemic treatment received was also evaluated: adjuvant or neoadjuvant chemotherapy, hormone therapy, and radiotherapy. The oncological outcomes evaluated were local recurrence (recurrence in the reconstructed breast), contralateral recurrence, regional recurrence (including axillary and internal mammary chains), distant recurrence and death. The 5-year disease-free survival (DFS) rate was defined as the proportion of live patients with no signs of locoregional disease, or contralateral or distant metastasis up to the time of the last follow-up. The overall survival (OS) rate was based on the proportion of patients alive at the last follow-up. Kaplan-Meier curves were developed for patients with invasive carcinoma beginning on the date of surgery to evaluate OS and DFS. Results The study population consisted of 76 consecutive patients with invasive or in-situ carcinomas operated on between November 2010 and December 2020. Thirteen of these (17%) were operated on between 2010 and 2015, while 63 (83%) underwent surgery between 2016 and 2020. The mean time of follow-up after NSM was 34.8 months (range 5–114 months). Clinical, surgical, and pathological characteristics are described in Table 1 . The mean age of patients was 46.1 years (27–73 years). The most common incisions used were inframammary fold in 35 cases (46%), periareolar in 34 (40%), and radial in 7 (9%). All patients underwent implant-based breast reconstruction. Twenty-seven patients (36%) were submitted to direct-to-implant reconstruction and 49 (64%) to 2-stage breast reconstruction. Fifty-eight patients had an invasive carcinoma, while 18 had a DCIS. The most common subtype of invasive carcinoma consisted of hormone-positive tumors (60%) followed by HER2 (24%) and triple-negative (16%). Of the cases of DCIS, histological grade was high in 39%. Table 1 Clinical, surgical, and pathological characteristics of the 76 consecutive patients submitted to nipple-sparing mastectomy for breast carcinomas Invasive carcinoma Ductal carcinoma in situ n % n % Number of cases diagnosed 58 18 Grading of ductal carcinoma in situ High grade 7 39 Low grade 11 61 Subtype of invasive carcinoma Luminal 35 60 − − Triple-negative 9 16 − − HER2 14 24 − − Type of incision Periareolar incision 23 40 11 61 Radial incision 7 12 0 0 Inframammary fold incision 28 48 7 39 Tumor size Tis 1 2 18 100 T1 16 28 0 0 T2 38 66 0 0 T3 3 5 0 0 Axillary status N0 36 62 18 100 N1 13 22 0 0 N2 8 14 0 0 N3 1 2 0 0 Treatment Neoadjuvant chemotherapy 22 38 0 0 Adjuvant chemotherapy 29 50 0 0 Endocrine therapy 41 71 0 0 Radiotherapy 22 38 1 6 In the majority of cases of invasive carcinoma, tumor size was T2 (66%), followed by T1 (28%), and T3 (5%). Axillary status was positive in 38% of cases and classified as follows: N1 (22%), N2 (14%), and N3 (2%). The resection margins of the surgical specimens were free of disease in all cases. All the patients with invasive carcinomas underwent some form of systemic treatment: chemotherapy (n = 51, 88%), with this consisting of neoadjuvant chemotherapy in 22 women (38%), and hormone therapy in 41 patients (71%). Radiotherapy was performed in 38% of cases of invasive carcinoma. Patients with DCIS underwent no systemic therapy. Four patients had a local recurrence: three cases of invasive carcinoma (5.1%), one of which was in the nipple-areola complex (NAC), and one in a patient with DCIS, also in the NAC (5.5%). In relation to cases of recurrence in the NAC, one patient had invasive HER2 disease (irregular mass; 1.8 cm from the NAC) and the other, a DCIS (microcalcifications; 2 cm from the nipple). Recurrence in the contralateral breast was recorded in two cases. Two patients had distant metastasis and two deaths occurred during the follow-up period ( Table 2 ), all in patients with an invasive carcinoma, resulting in a 5-year OS rate of 98% and a DFS rate of 83% ( Figure 1 ). Fig. 1 Five-year oncological outcome of nipple-sparing mastectomy in invasive carcinoma. Table 2 Recurrences following nipple-sparing mastectomy in invasive or in situ carcinomas Invasive carcinoma Ductal carcinoma in situ n % n % Local recurrence 3 5.1 1 5.5 Regional recurrence 0 0 0 0 Recurrence in the contralateral breast 1 1.7 0 0 Distant metastasis 2 3.4 0 0 Death 2 3.4 0 0 Discussion In the present analysis, the OS rate and the local recurrence rate were acceptable and comparable with data from previously published studies, both for invasive and for in-situ disease. Indeed, the rate of local recurrence detected over the study period was around 5%, with a 5-year OS rate of 98%, despite the fact that 71% of the sample of invasive tumors consisted of T2/T3 tumors, and the axilla was positive in 38%, the same proportion of patients who received neoadjuvant chemotherapy. An Italian study involving 1,989 patients reported similar results, with a local recurrence rate of 5.3% in cases of invasive breast cancer and 4% in cases of in-situ disease after a follow-up time of 94 months. Around 50% of that sample had T2/T3 disease or positive axilla. 10 Retrospective studies have shown that NSM is a safe alternative to conventional mastectomy. Although this is a technically more complex surgery, delays in providing adjuvant therapies are uncommon and locoregional control of the disease is generally excellent. 8 10 11 12 20 Most of those studies, however, evaluated selected populations, predominantly at clinical stages I or II. 8 13 21 Expansion of the use of NSM has been a subject of debate; however, the oncological safety of NSM in locally more advanced tumors (stage IIB or higher) or in patients receiving neoadjuvant chemotherapy has not been sufficiently demonstrated. 22 23 Tumor-to-nipple distance is also a topic that has generated much debate. Traditionally, a tumor-to-nipple distance of 2 cm has been suggested as the limit criterion for performing NSM; however, this is not uniformly accepted in the literature. 24 25 26 27 In the present study, the recommendation for performing NSM was not based on any specific tumor-to-nipple distance. The findings of the present study reject the concept of a tumor-to-nipple distance > 2 cm. In fact, our group has expanded the indication of NSM over the years to include patients with locally advanced disease. We believe that tumor-to-nipple distance < 2 cm should not be a contraindication for the technique, particularly in the era of multimodal treatment. This rationale is applied in breast-conserving surgery, in which, paradoxically, this type of surgery has decreased over the years without affecting control of the disease. 1 2 3 4 5 6 We considered that the absence of any signs that the NAC was affected by the disease at clinical evaluation and imaging tests, associated with a final histopathology result showing disease-free margins, would be sufficient. An evaluation of 193 patients with invasive disease and a tumor-to-nipple distance > or < 2 cm, measured using MRI, found no statistically significant difference in terms of DFS. 27 Another study reporting similar results following the analysis of 245 patients submitted to NSM and previous MRI measurements found no statistically significant differences in terms of DFS and survival without local recurrence after 60 months of follow-up irrespective of tumor-to-nipple distance (< or ≥ 2 cm). 26 Furthermore, a recent evaluation comparing the oncological outcome of patients in whom tumor-to-nipple distance was ≤ 1 cm versus a group in which tumor-to-nipple distance was > 1 cm, as measured by imaging tests, detected no statistically significant difference in local recurrence after more than 100 months of follow-up. 25 Other controversial issues concern intraoperative margin assessment and the need for NAC resection if disease-positive. 24 28 29 30 31 In the present study, frozen section examination was not routinely performed, with decisions regarding management being based on the final histological analysis of the surgical specimen. Initially, the possibility of a false-positive result is a motive of concern, particularly in low-grade lesions. 29 Furthermore, identification of disease at frozen section examination does not necessarily mean recurrence in the NAC. Analysis of 948 NSM and 88 false-negatives at frozen section examination showed that the 5-year accumulated rate of local recurrence in the NAC was 2.4%, suggesting that it is possible to preserve the NAC in selected cases following an interdisciplinary debate and discussion with the patient. 31 Another possible strategy in such cases is to remove only the nipple, preserving the rest of the areola. In a study involving 1,326 patients submitted to NSM in which 46 nipple margins were positive, the nipple alone was removed in 51% of cases, and no recurrences were found in the NAC after 36 months of follow-up, suggesting that partial resection of the NAC can represent a safe option. 30 Recently, the debate regarding the appropriate thickness of the NSM flap has intensified. Since this is a technically more complicated procedure and surgical access is more limited, the possibility of leaving residual breast tissue has been controversial. 19 32 33 34 A survey conducted with radiation oncologists and surgeons concluded that the ideal flap within the context of conservative mastectomy (SSM and NSM) would be between 1 and 5 mm. 33 In the present study, no specific cut-off point was stipulated for the thickness of the flap, and the flap was not evaluated with routine postsurgical imaging tests. In our opinion, the thickness of the flap in NSM depends on the anatomical structure of each individual patient and leaving breast tissue behind does not necessarily imply a greater likelihood of recurrence. A review of mastectomies (SSM) found residual breast tissue even in flaps with thickness < 5 mm, showing that “more radical” flaps may not eliminate the risk of residual breast tissue. 15 Furthermore, thinner flaps are known to be associated with a greater likelihood of ischemic complications, suggesting a need to individualize the technique among women. 32 Using presurgical MRI measurements, a study evaluated the flap thickness in NSM as a function of body mass index (BMI) categories. 33 The findings showed that BMI < 25, 25–30 and > 30, as well the weight categories of the mastectomy (< 400 g, 400–799 g and > 800 g), were associated with flap thicknesses, with each increase being statistically significant. Another related topic is the need for radiotherapy. Many argue that if NSM constitutes conservative surgery, then the patient is a natural candidate for radiotherapy. However, various authors have questioned the need for radiotherapy in NSM due to the low rate of local recurrence found: in a study conducted in Italy, radiotherapy of the whole thoracic wall was given in 6.7% of cases following NSM, with local recurrence reported in around 5% of the women, although the axillary lymph nodes were affected in around 50% of the patients. 10 Paradoxically, patients submitted to NSM still appear more likely to receive radiotherapy compared to those submitted to less conservative mastectomies. A review conducted using the Surveillance, Epidemiology, and End Results (SEER) program database identified 470 patients who had undergone NSM between 2006 and 2010 and 112,347 who had not, with results showing that the women in the NSM group were more likely to have undergone radiotherapy. 35 There are some limitations associated with the present study. Since this was a retrospective cohort design, and the study was conducted in one single institute, there is a possibility that biases could have affected the findings. The small number of events could also have affected the analysis, as well as the short follow-up time. On the other hand, the strongpoint of this study lies in the fact that the population analyzed was unselected, and healthcare was provided within the public sector. Conclusion The present findings corroborate those of other retrospective series on NSM, with a low rate of local recurrence, even in an unselected population, and an excellent OS rate. Nevertheless, further studies are required to improve understanding of this approach. Contributions Conflict of Interests The authors have no conflict of interests to declare. F. P. C. and M. M. P. A. contributed to the conception and design of the study and participated in the acquisition, analysis, and interpretation of the data. All authors contributed to the drafting of the article, revised the manuscript critically for important intellectual content, approved the final version to be published and agree to be accountable for all aspects of the work. ==== Refs References 1 Fisher B Anderson S Bryant J Margolese R G Deutsch M Fisher E R Twenty-year follow-up of a randomized trial comparing total mastectomy, lumpectomy, and lumpectomy plus irradiation for the treatment of invasive breast cancer N Engl J Med 2002 347 16 1233 1241 10.1056/NEJMoa022152 12393820 2 Veronesi U Cascinelli N Mariani L Greco M Saccozzi R Luini A Twenty-year follow-up of a randomized study comparing breast-conserving surgery with radical mastectomy for early breast cancer N Engl J Med 2002 347 16 1227 1232 10.1056/NEJMoa020989 12393819 3 van Dongen J A Voogd A C Fentiman I S Legrand C Sylvester R J Tong D Long-term results of a randomized trial comparing breast-conserving therapy with mastectomy: European Organization for Research and Treatment of Cancer 10801 trial J Natl Cancer Inst 2000 92 14 1143 1150 10.1093/jnci/92.14.1143 10904087 4 Poggi M M Danforth D N Sciuto L C Smith S L Steinberg S M Liewehr D J Eighteen-year results in the treatment of early breast carcinoma with mastectomy versus breast conservation therapy: the National Cancer Institute Randomized Trial Cancer 2003 98 04 697 702 10.1002/cncr.11580 12910512 5 Danish Breast Cancer Cooperative Group Blichert-Toft M Rose C Andersen J A Overgaard M Axelsson C K Andersen K W Danish randomized trial comparing breast conservation therapy with mastectomy: six years of life-table analysis J Natl Cancer Inst Monogr 1992 11 19 25 1627427 6 Institut Gustave-Roussy Breast Cancer Group Arriagada R Lê M G Rochard F Contesso G Conservative treatment versus mastectomy in early breast cancer: patterns of failure with 15 years of follow-up data J Clin Oncol 1996 14 05 1558 1564 10.1200/JCO.1996.14.5.1558 8622072 7 Cavalcante F P Lima M VA Nipple-sparing mastectomy with periareolar incision and two-stage reconstruction: Initial analysis of 31 cases Breast J 2018 24 06 940 943 10.1111/tbj.13114 30216598 8 Margenthaler J A Gan C Yan Y Cyr A E Tenenbaum M Hook D Oncologic safety and outcomes in patients undergoing nipple-sparing mastectomy J Am Coll Surg 2020 230 04 535 541 10.1016/j.jamcollsurg.2019.12.028 32032724 9 Lanitis S Tekkis P P Sgourakis G Dimopoulos N Al Mufti R Hadjiminas D J Comparison of skin-sparing mastectomy versus non-skin-sparing mastectomy for breast cancer: a meta-analysis of observational studies Ann Surg 2010 251 04 632 639 10.1097/SLA.0b013e3181d35bf8 20224371 10 Galimberti V Morigi C Bagnardi V Corso G Vicini E Fontana S KR Oncological outcomes of nipple-sparing mastectomy: a single-center experience of 1989 patients Ann Surg Oncol 2018 25 13 3849 3857 10.1245/s10434-018-6759-0 30225833 11 Wu Z Y Kim H J Lee J W Chung I Y Kim J S Lee S B Breast cancer recurrence in the nipple-areola complex after nipple-sparing mastectomy with immediate breast reconstruction for invasive breast cancer JAMA Surg 2019 154 11 1030 1037 10.1001/jamasurg.2019.2959 31461141 12 Adam H Bygdeson M de Boniface J The oncological safety of nipple-sparing mastectomy - a Swedish matched cohort study Eur J Surg Oncol 2014 40 10 1209 1215 10.1016/j.ejso.2014.07.037 25186915 13 Valero M G Muhsen S Moo T A Zabor E C Stempel M Pusic A Increase in utilization of nipple-sparing mastectomy for breast cancer: indications, complications, and oncologic outcomes Ann Surg Oncol 2020 27 02 344 351 10.1245/s10434-019-07948-x 31823173 14 Wu Z Y Kim H J Lee J Chung I Y Kim J S Lee S B Recurrence outcomes after nipple-sparing mastectomy and immediate breast reconstruction in patients with pure ductal carcinoma in situ Ann Surg Oncol 2020 27 05 1627 1635 10.1245/s10434-019-08184-z 31912259 15 Torresan R Z dos Santos C C Okamura H Alvarenga M Evaluation of residual glandular tissue after skin-sparing mastectomies Ann Surg Oncol 2005 12 12 1037 1044 10.1245/ASO.2005.11.027 16244800 16 Brachtel E F Rusby J E Michaelson J S Chen L L Muzikansky A Smith B L Occult nipple involvement in breast cancer: clinicopathologic findings in 316 consecutive mastectomy specimens J Clin Oncol 2009 27 30 4948 4954 10.1200/JCO.2008.20.8785 19720921 17 Kim S Lee S Bae Y Lee S Nipple-sparing mastectomy for breast cancer close to the nipple: a single institution's 11-year experience Breast Cancer 2020 27 05 999 1006 10.1007/s12282-020-01104-0 32372321 18 Eisenberg R E Chan J S Swistel A J Hoda S A Pathological evaluation of nipple-sparing mastectomies with emphasis on occult nipple involvement: the Weill-Cornell experience with 325 cases Breast J 2014 20 01 15 21 10.1111/tbj.12199 24438063 19 Giannotti D G Hanna S A Cerri G G Barbosa Bevilacqua J L Analysis of skin flap thickness and residual breast tissue after mastectomy Int J Radiat Oncol Biol Phys 2018 102 01 82 91 10.1016/j.ijrobp.2018.05.023 30102208 20 Albright E L Schroeder M C Foster K Sugg S L Erdahl L M Weigel R J Nipple-sparing mastectomy is not associated with a delay of adjuvant treatment Ann Surg Oncol 2018 25 07 1928 1935 10.1245/s10434-018-6446-1 29671138 21 Frey J D Alperovich M Kim J C Axelrod D M Shapiro R L Choi M Oncologic outcomes after nipple-sparing mastectomy: A single-institution experience J Surg Oncol 2016 113 01 8 11 10.1002/jso.24097 26628318 22 Kopkash K Sisco M Poli E Seth A Pesce C The modern approach to the nipple-sparing mastectomy J Surg Oncol 2020 122 01 29 35 10.1002/jso.25909 32219847 23 Wu Z Y Han H H Kim H J Chung I Y Kim J Lee S B A propensity score-matched analysis of long-term oncologic outcomes after nipple-sparing versus conventional mastectomy for locally advanced breast cancer Ann Surg 2020 •••: 10.1097/SLA.0000000000004416[ahead of print] 24 Fregatti P Gipponi M Zoppoli G Lambertini M Blondeaux E Belgioia L Tumor-to-nipple distance should not preclude nipple-sparing mastectomy in breast cancer patients. Personal experience and literature review Anticancer Res 2020 40 06 3543 3550 10.21873/anticanres.14343 32487656 25 Wu Z Y Kim H J Lee J Chung I Y Kim J Lee S B Oncologic safety of nipple-sparing mastectomy in patients with breast cancer and tumor-to-nipple distance ≤ 1 cm: a matched cohort study Ann Surg Oncol 2021 28 08 4284 4291 10.1245/s10434-020-09427-0 33423119 26 Alsharif E Ryu J M Choi H J Nam S J Kim S W Yu J Oncologic outcomes of nipple-sparing mastectomy with immediate breast reconstruction in patients with tumor-nipple distance less than 2.0 cm J Breast Cancer 2019 22 04 613 623 10.4048/jbc.2019.22.e48 31897334 27 Balci F L Kara H Dulgeroglu O Uras C Oncologic safety of nipple-sparing mastectomy in patients with short tumor-nipple distance Breast J 2019 25 04 612 618 10.1111/tbj.13289 31087467 28 Alperovich M Choi M Karp N S Singh B Ayo D Frey J D Nipple-sparing mastectomy and sub-areolar biopsy: to freeze or not to freeze? Evaluating the role of sub-areolar intraoperative frozen section Breast J 2016 22 01 18 23 10.1111/tbj.12517 26510917 29 D'Alonzo M Pecchio S Campisi P De Rosa G Bounous V E Villasco A Nipple-Sparing Mastectomy: Reliability of sub-areolar sampling and frozen section in predicting occult nipple involvement in breast cancer patients Eur J Surg Oncol 2018 44 11 1736 1742 10.1016/j.ejso.2018.07.059 30100362 30 Tang R Coopey S B Merrill A L Rai U Specht M C Gadd M A Positive nipple margins in nipple-sparing mastectomies: rates, management, and oncologic safety J Am Coll Surg 2016 222 06 1149 1155 10.1016/j.jamcollsurg.2016.02.016 27118712 31 Kneubil M C Lohsiriwat V Curigliano G Brollo J Botteri E Rotmensz N Risk of locoregional recurrence in patients with false-negative frozen section or close margins of retroareolar specimen in nipple-sparing mastectomy Ann Surg Oncol 2012 19 13 4117 4123 10.1245/s10434-012-2514-0 22820937 32 Frey J D Salibian A A Choi M Karp N S Mastectomy flap thickness and complications in nipple-sparing mastectomy: objective evaluation using magnetic resonance imaging Plast Reconstr Surg Glob Open 2017 5 08 e1439 10.1097/GOX.0000000000001439 28894660 33 Frey J D Salibian A A Choi M Karp N S Optimizing outcomes in nipple-sparing mastectomy: mastectomy flap thickness is not one size fits all Plast Reconstr Surg Glob Open 2019 7 01 e2103 10.1097/GOX.0000000000002103 30859052 34 Marta G N Poortmans P de Barros A C Filassi J R Freitas Junior R Audisio R A Multidisciplinary international survey of post-operative radiation therapy practices after nipple-sparing or skin-sparing mastectomy Eur J Surg Oncol 2017 43 11 2036 2043 10.1016/j.ejso.2017.09.014 28967564 35 Agarwal S Agarwal J Radiation delivery in patients undergoing therapeutic nipple-sparing mastectomy Ann Surg Oncol 2015 22 01 46 51 10.1245/s10434-014-3932-y 25037972
PMC009xxxxxx/PMC9800069.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36067799 10.1055/s-0042-1755460 220043 Review Article Mastology Iron Salts, High Levels of Hemoglobin and Ferritin in Pregnancy, and Development of Gestational Diabetes: A Systematic Review Sais de ferro, níveis elevados de hemoglobina e ferritina na gravidez e o desenvolvimento de diabetes gestacional: Uma revisão sistemáticahttp://orcid.org/0000-0002-9156-5036 Miranda Vanessa Iribarrem Avena 1 http://orcid.org/0000-0002-7566-7745 Pizzol Tatiane da Silva Dal 2 http://orcid.org/0000-0002-4987-146X Jesus Patricia Romualdo de 2 http://orcid.org/0000-0002-6453-8534 Silveira Marysabel Pinto Telis 3 http://orcid.org/0000-0002-4680-3197 Bertoldi Andréa Dâmaso 1 1 Post-Graduate Program in Epidemiology, Federal University of Pelotas, Pelotas, Brazil 2 Post-Graduate Program in Epidemiology, Federal University of Rio Grande do Sul, Porto Alegre, RS, Brazil 3 Institute of Biology, Department of Physiology and Pharmacology and Post-Graduate Program in Epidemiology, Federal University of Pelotas, Pelotas, Brazil Address for correspondence Patricia Romualdo de Jesus, Master's degree Ramiro Barcelos Street2400, Porto Alegre, [email protected] 06 9 2022 11 2022 1 9 2022 44 11 10591069 11 2 2022 03 6 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  The aim of this study was to systematically review literature on the use of iron supplements (not including iron derived from diet), increased levels of hemoglobin and/or ferritin, and the risk of developing gestational diabetes mellitus (GDM). Data source  The following databases were searched, from the study's inception to April 2021: PUBMED, Cochrane, Web of Science, Scopus, Embase, Cinahl and Lilacs. Selection of studies  A total of 6,956 titles and abstracts were reviewed, 9 of which met the final inclusion criteria, with 7,560 women in total. Data collection  Data extraction was performed by two independent reviewers and disagreements were resolved by a third researcher. Data synthesis  Methodological quality in controlled trials were assessed according to the Cochrane Collaboration tools (ROB-2 and ROBINS-1) and for the observational studies, the National Institutes of Health's (NIH) quality assessment tool was used. Among the 5 observational studies, women with a higher hemoglobin or ferritin level were more likely to develop GDM when compared with those with lower levels of these parameters. Among the 3 randomized clinical trials, none found a significant difference in the incidence of GDM among women in the intervention and control groups. However, we identified many risks of bias and great methodological differences among them. Conclusion  Based on the studies included in this review, and due to the important methodological problems pointed out, more studies of good methodological quality are needed to better establish the association between iron supplementation and GDM. Resumo Objetivo  O objetivo deste estudo foi revisar sistematicamente a literatura sobre o uso de suplementos de ferro (não incluindo o ferro derivado da dieta), aumento dos níveis de hemoglobina e/ou ferritina e o risco de desenvolver diabetes mellitus gestacional (DMG). Fontes dos dados  as bases de dados PUBMED, Cochrane, Web of Science, Scopus, Embase, Cinahl e Lilacs foram pesquisadas até abril de 2021. Seleção dos estudos  Foram revisados 6.956 títulos e resumos, dos quais 9 preencheram os critérios finais de inclusão, totalizando 7.560 mulheres. Coleta de dados  A extração de dados foi realizada por dois revisores independentes e as divergências foram resolvidas por um terceiro revisor. Síntese dos dados  A qualidade metodológica dos ensaios controlados foi avaliada de acordo com as ferramentas da Colaboração Cochrane (ROB-2 e ROBINS-1) e para os estudos observacionais, foi utilizada a ferramenta de avaliação de qualidade do National Institutes of Health (NIH). Entre os 5 estudos observacionais, as mulheres com maiores níveis de hemoglobina ou ferritina apresentaram maior probabilidade de desenvolver DMG quando comparadas àquelas com níveis mais baixos nesses parâmetros. Entre os 3 ensaios clínicos randomizados, nenhum deles encontrou diferença significativa na incidência de DMG entre as mulheres dos grupos de intervenção e controle. No entanto, identificamos muitos riscos de viés e enormes diferenças metodológicas entre eles. Conclusão  Com base nos estudos incluídos nesta revisão e devido aos importantes problemas metodológicos apontados, são necessários mais estudos de boa qualidade metodológica para melhor estabelecer a associação entre suplementação de ferro e DMG. Keywords ferritin gestational diabetes hemoglobin iron supplement pregnancy Palavras-chave ferritina diabetes gestacional hemoglobina suplemento de ferro gravidez ==== Body pmcIntroduction Gestational diabetes mellitus (GDM) is a temporary condition characterized by hyperglycemia, which occurs due to glucose intolerance, with onset during pregnancy and usually disappearing shortly after delivery. 1 This condition is responsible for several maternal-fetal health consequences, such as the increased risk of malformations, fetal loss, and neonatal, perinatal, and maternal mortality. 2 In the long term, the maternal effect of GDM is the increased risk of metabolic syndrome and type 2 diabetes. In children, it almost doubles the risk of developing childhood obesity and metabolic syndrome compared with children born from non-diabetic mothers. 3 The prevalence of GDM is increasing globally, in parallel with the increase in type 2 diabetes mellitus and female obesity. 1 However, there is no accurate estimate of the overall incidence of gestational diabetes because screening and diagnostic patterns are not uniform throughout the world and over the years. 3 Estimates from the American Diabetes Association (ADA) indicate a prevalence of 1 to 14%. 4 However, in 2013, the World Health Organization (WHO) generated new estimates from the update of its new criterion, where the prevalence ranged from 4 to 25%, with more than 90% occurring in low- and middle-income countries. 1 5 The pathogenesis of GDM is unclear, although some authors suggest it is a complex disease with a combination of genetic and environmental factors. 6 According to the Diabetes Canada Clinical Practice Guidelines Expert Committee, risk factors increasing the risk of GDM include increased maternal age, a high-risk group (African, Arab, Asian, Hispanic, Indigenous, or South Asian), corticosteroid medication, obese (body mass index, BMI > 30 kg/m2), prediabetic, prior history of GDM, family history of diabetes (parent, brother or sister), polycystic ovary syndrome or acanthosis nigricans. 7 The American Diabetes Association includes other risk factors, such as overweight in current pregnancy, central deposition of body fat, short stature (< 151 cm), excessive fetal growth, polyhydramnios, hypertension, or preeclampsia in the current pregnancy, as well as a history of fetal or neonatal death, and macrosomia. 4 From the 2000s onwards, several authors have hypothesized that iron supplementation in women with normal levels of hemoglobin and serum ferritin, a practice recommended by several entities to prevent gestational anemia, may contribute to an increased risk of GDM. 8 9 10 The glucose metabolism can be affected by iron accumulation in several ways. One of the suggested mechanisms is regarding the high oxidation capacity of iron. Iron promotes the formation of hydroxyl radicals that can attack β-cell membranes and affect insulin synthesis and secretion in the pancreas. 11 However, some studies have not confirmed this hypothesis. 12 13 14 15 16 The role of iron excess in the pathogenesis of GDM needs to be examined, specifically to regard the iron provided by supplements. The present systematic review was conducted to answer the following question: “Does the use of iron supplements (not including dietary iron) and high levels of hemoglobin and/or ferritin increase the risk of developing GDM in non-anemic women?” Methods A systematic review was performed according to the Preferred Reporting Items for Systematic Review and Meta-Analyzes (PRISMA). 17 The review protocol was registered in the international database PROSPERO, under the number CRD42018086269. Experimental or observational studies in humans were included, with GDM as primary or secondary outcome, and the exposure being use of iron supplement during pregnancy (not including iron derived from diet), and the presence of at least one biochemical parameter of iron (hemoglobin or serum ferritin). Animal studies, uncontrolled studies, reviews, meta-analysis, protocols, abstracts, charts, and those that measured exposure through dietary iron intake or supplementary iron intake in all groups of pregnant women regardless of hemoglobin level were excluded. The systematic review of the literature was performed in PUBMED, Cochrane, Web of Science, Scopus, Embase, Cinahl, and Lilacs databases, from inception to April 2021. No limits for year of publication and language were used. The gray literature was accessed through Capes Thesis database, the gray literature report ( http://www.greylit.org ) and Open Gray ( http://www.opengrey.eu ). The references of the included studies were reviewed to check for other possible studies to be included. For further details regarding search strategies for all databases, please refer to the supplementary file. Three authors independently screened the titles and abstracts of all retrieved studies. Then full-text screening of each potentially eligible study was done independently by the authors, and discrepancies were resolved by a third reviewer. The online tools Covidence (Covidence Melbourne, Australia) and Rayyan(Qatar Foundation, Qatar) were used for article selection. 18 The following data were extracted from the eligible articles: country of origin, year of publication, design, health service used, sample size, iron dose (duration and frequency), age, number of women with GDM, gestational week of GDM diagnosis, biochemical parameters of iron, and gestational trimester of collection of these parameters. Methods to control for confounders and covariables such as parity, age, and BMI were recorded. The authors were contacted when additional information was needed. Of the 16 authors we contacted by e-mail, 4 responded. Two of our authors independently extracted data using a standardized data extraction form, and discrepancies were resolved by a third reviewer. The outcome of interest, GDM, was measured differently in the studies, and the authors' data was extracted as reported. The risk of bias was independently assessed by two reviewers and the disagreements were resolved by a third reviewer. Risk of bias in randomized controlled trials (RCTs) was assessed according to the revised Cochrane risk of bias tool for randomized trials (RoB-2), 19 which recommended the evaluation of the following domains: risk of bias arising from the randomization process, bias due to deviations from the intended interventions, bias due to missing outcome data, bias in the measurement of the outcome, and bias in the selection of the reported result. The risk of bias for each domain assessed was classified as low, some concerns, and high. To evaluate the risk of bias in non-randomized controlled trials (NRSIs), we used the Risk Of Bias In Non-Randomized Studies of Interventions (ROBINS-I) tool, 20 with seven domains: bias due to confounding, bias in the selection of participants into the study, bias in classification of interventions, bias due to deviations from intended interventions, bias due to missing data, bias in the measurement of outcomes, and bias in the selection of the reported result. The risk of bias of domains was classified as low, moderate, serious, critical, and no information. For the cohort and case-control studies, the NHS's 21 Quality Assessment Tool was used, with 14 questions for cohort and cross-sectional studies, and another one with 12 questions for case-control studies, with each item being rated as either yes (criterion met), no (criterion not met), not applicable, cannot determine, or not reported. The results of the analysis of the risk of bias were presented in figures using the Risk of bias VISualization (robvis) tool. 22 The main measures to summarize our outcome of interest were obtained as reported by the authors (e.g., relative risk, odds ratio, mean difference). Due to the variation in the study design, participants, interventions, and measures of results reported, it was decided to describe the studies, their results, applicability, and limitations in a qualitative synthesis, instead of performing a meta-analysis. Results Of a total of 6,956 studies identified, 4,209 were screened after exclusion of duplicates. After reading the titles and abstracts, 75 manuscripts were selected for a full reading, and of these, 7 reports could not be retrieved. Most studies were ruled out for other undesired exposures, such as supplementation in all groups of pregnant women, or no data on iron supplementation. After reading the articles and reviewing references, 9 studies were included in the review ( Fig. 1 ). Fig. 1 The PRISMA flowchart of search strategy and selection process. The characteristics of the included articles were described in Charts 1 and 2 . The selected works included cohort studies ( n  = 3), 23 24 25 case controls ( n  = 2), 26 27 and controlled trials ( n  = 4). 13 28 29 30 The studies were conducted in the United Arab Emirates ( n  = 1), China ( n  = 4), Finland ( n  = 1), Iran ( n  = 1), Turkey ( n  = 1), and the United States ( n  = 1). Overall, 7,560 pregnant women participated in the studies, ranging from 58 to 3,289. Participants came from hospitals, 13 23 24 28 30 clinical centers, 27 polyclinics, 26 and primary health care settings. 29 One study did not report this data ( Chart 1 and 2 ). 13 Chart 1 Characteristics of the controlled trials included in the systematic review First author (year) Country Study design Health service GDM diagnosis: method/gestational week Control group Experimental group N Age Iron dose (frequency, duration, trimester) n Age Iron dose (frequency, duration, trimester) Asadi (2019) 28 * Iran NRCT Hospital NR 30 30 29.6 ± 12.3 27.5 ± 8.3 NR 30 29.6 ± 16.1 NR Chan (2009) 13 China RCT Hospital OGTT/ 28–30 and 36 weeks 599 31.3 ± 0.18 Placebo at minimum of 12 weeks (from 16–28), 1 st trimester 565 31.3 ± 0.19 60 mg EI daily, minimum of 12 weeks (16–28 weeks), 1 st trimester Kinnunen (2016) 29 Finland RCT Primary health care NR (diagnosis obtained through medical records) a 1,358 27.8 ± 5.6 50 mg EI twice a day, only if Hb fell below 100 g/L, 2 months or until their Hb increased to 110 g/L, 2 nd trimester 1,336 27.7 ± 4.9 100 mg EI daily b until delivery, all trimesters Ouladsahebmadarek (2011) 30 Arab Emirates RCT Hospital NR 372 25.48 ± 4.96 Placebo, 13 weeks, 1 st trimester 410 26.28 ± 5.25 30 mg EI daily, 13 weeks, 1 st trimester Abbreviations: CI, confidence interval; EI, elemental iron; GDM, gestational diabetes mellitus; Hb, Hemoglobin; NR, not related; NRCT, non-randomized clinical trial; OR, odds ratio; RCT, randomized controlled trial. Notes: *This study had three groups of comparison: women with normal serum ferritin levels (= 30 µg/dL) who received standard prophylactic iron supplementation during the pregnancy (designed as the experimental group); those who had minor thalassemia and El-evated serum ferritin levels (= 30 µg/dL) who did not receive prophylactic iron supplementation, or those with normal ferritin levels (≥ 30 μg/dL) who refused to take iron supplementation due to gastrointestinal upset (designed as control group 1); and those with iron deficiency anemia with low serum ferritin levels (< 30 μg/dL) who received standard iron supplementation during pregnancy (designed as control group 2). a The authors stated that GDM “was not assessed systematically among all participants, but was abstracted from patient records as recognized in the usual care.” b The participants received 100 mg elemental iron throughout the pregnancy regardless of Hb level. Chart 2 Characteristics of the observational studies included in the systematic review. Cohort study ( n  = 3) and case-control ( n  = 2) First author (year) Country Study design Health service GDM diagnosis: method/ gestational week Sample size Age Iron dose (frequency, duration) N Control group Exposed group Case group Control group Özyiğit (2008) 26 Turkey Case-control Polyclinic OGTT/ 24–28 weeks 58 23 35 − − 26 40 mg IS daily for at least 2 months Rawal (2016) 27 United States Case-control Clinical centers NR (diagnosis obtained through medical records)/ 10–14 and 15–26 weeks 321 − − 107 214 Control group: 30.4 ± 5.4 Case group: 30.5 ± 5.7 Dose: NR 83% at weeks 10–14; 87% in the 2 nd trimester during 15–26 gestational weeks reported using IS Liu and Pang (2018) 23 China Cohort Hospital NR/at delivery 259 124 135 − − Control group 31.1 ± 1.5 Exposed group 30.9 ± 1.3 300 mg of IS daily, 7 days weekly until the baby delivery Zhu (2019) 25 China Cohort NR OGTT 24–28 3,289 a 1 st 2,508 2 nd 2,805 1 st 777 2 nd 433 − − 26.4 ± 3.7 NR Si (2020) 24 China Cohort Hospital OGTT/ 24–28 1,128 b 1 st 831 2 nd 639 1 st 297 2 nd 489 − − GDM group 28.60 ± 3.33 No GDM group 27.91 ± 3.25 NR Abbreviation: GDM, gestational diabetes mellitus; IS, iron supplementation; NR, Not related; OGTT, oral glucose tolerance test. Notes: a The 3,289 pregnant women were evaluated in the 1 st and 2 nd trimesters of pregnancy, with 4 missing in the 1 st and 51 missing in the 2 nd trimester; b The 1,128 pregnant women were evaluated in the 1 st and 2 nd trimesters of pregnancy. In the controlled trials, the daily dose of iron supplementation ranged from 30 to 100 mg of elemental iron in the experimental groups. 13 29 30 In the observational studies, one study reported 40 mg/day of elemental iron 26 and another 300 mg daily. 23 The other studies did not report this data. 24 25 27 28 The duration of supplementation throughout pregnancy was inconsistently recorded throughout the studies. There is no standard in the duration of iron supplementation, occurring over different periods of pregnancy. 23 26 27 29 30 Three studies did not report the duration of iron supplementation. 24 25 28 Blood collection for hemoglobin and/or ferritin tests was performed at different times, during the first, 24 25 28 30 second, 24 25 27 28 and third trimesters, 13 28 as well as at delivery. 23 It was also performed at first visits, and at 12, 20, 28, and 36 weeks of gestation. 29 The diagnosis of GDM was measured with an oral glucose tolerance test (OGTT) in 3 studies: in 3, the test was done at 24 to 28 weeks, 24 25 26 and in 1 at 28 to 30 and 36 weeks. 13 Furthermore, 3 studies did not present any information on GDM diagnosis 23 28 30 and 2 mentioned that the diagnosis was extracted from patients' medical record. 27 29 The diagnosis of GDM was performed between 24 and 28 weeks of gestation in 3 of the studies; 24 25 26 Rawal et al. 27 reported it at 10 to 14 and 15 to 26 weeks; Chan et al. 13 reported it at 28 to 30 and 36 weeks; and Liu and Pang 23 reported it at delivery. Finally, 3 studies did not report the time of GDM diagnosis. 28 29 30 Among the RCTs, none of them found a significant difference in the incidence of GDM between women in the control and experimental groups ( Chart 3 ). 13 29 30 On the other hand, the non-randomized clinical trial (NRCT) found an association between iron supplementation in pregnant women with normal ferritin levels and the increased risk of GDM. 28 Chart 3 Quantitative results from controlled trials First author (year) Biomarker of iron Trimester collected Outcomes (incidence of GDM) Statistically significant difference Hemoglobin (g/dL) Ferritin (pmol/L) Control group Experimental group Asadi (2019) 28 Control group 1: 1 st 12.24 ± 1.5 2 nd 11.88 ± 1.4 3 rd 12.39 ± 2.29 Control group 2: 1 st 12.54 ± 1.23 2 nd 12.02 ± 1.06 3 rd 12.46 ± 1.12 Experimental group: 1 st 14.75 ± 1.42; p  = 0.452 2 nd 12.32 ± 1.27; p  = 0.387 3 rd 12.79 ± 1.18; p  = 0.59 NR 1 st , 2 nd , and 3 rd 0 (0%) 0 (0%) 5 (16.7%) Yes p  = 0.038 Chan (2009) 13 Control group ( n  = 531) 11.4 ± 0.04 Experimental group ( n  = 511) 11.4 ± 0.04 p < 0.001 Control group ( n  = 490) 49.14 ± 2.17 Experimental group ( n  = 469) 60.4 ± 2.22 p < 0.001 3 rd − OR = 1.04 CI = 0.7–1.53 No Kinnunen (2016) 29 Control group: 20 weeks 126.4 ± 8.9 28 weeks 123.9 ± 9.5 36 weeks 124.8 ± 9.6 Experimental group: 20 weeks 127.0 ± 8.6; p  = 0.11 28 weeks 127.0 ± 8.6; p  < 0.001 36 weeks 131.2 ± 9.3; p  < 0.001 NR 2 nd and 3 rd 11.0% 13.0% No p  = 0.12 Ouladsahebmadarek(2011) 30 Control group: 12.48 ± 0.91 Experimental group: 13.46 ± 0.75; p  = 0.03 Control group (μg/dl): 9.26 ± 0.62 Experimental group (μg/dl): 26.91 ± 2.11; p  = 0.048 At delivery 0.8% 0.5% No p  = 0.67 Abbreviations: CI, confidence interval; GDM, gestational diabetes mellitus; NR, not related; OR, odds ratio. Among the cohort studies ( Chart 2 ), two found no significant associations between iron supplement use and risk of GDM. 23 25 Si et al. 24 verified that the iron supplementation in pregnant women with high hemoglobin concentration (Hb > 11) increased the risks for GDM. Among the case-control studies, one used average hemoglobin and ferritin, 26 and another used only ferritin. 27 The average hemoglobin and ferritin levels were higher in women with GDM in both studies ( Chart 4 ). Chart 4 Quantitative results from observational studies First author, year Biomarker of iron Trimester collected OR crude (95% CI or p-value) OR adjusted (95% CI) Control for confounding Hemoglobin (g/dL) Ferritin (pmol/L) Özyiğit (2008) 26 Iron group 12.3 ± 1.3 Control group 12.1 ± 1.04 Iron group 16.5 ± 6.09 Control group 12.5 ± 9.7 (ng/mL) ( p  = 0.018) 1 st − − BMI, age, parity. Rawal (2016) 27 NR GDM group 94.5 (67.9–133.5) No GDM group 78.1 (49.7–119.4) 2 nd 3.06 (CI = 1.27–7.34) 15–26 weeks; highest vs. lowest quartiles 3.95 (CI= 1.38–11.3) 15–26 weeks; highest vs. lowest quartiles BMI, parity, education, family history of diabetes. Liu and Pang (2018) 23 Iron group 12.1 ± 0.06 Control group 11.7 ± 0.08 p <0.01 Iron group 65.8 ± 3.1 Control group 53.7 ± 2.9 p <0.01 At delivery 1.02 ( p  = 0.96) − NR Zhu (2019) 25 GDM group 1 st 128.8 ± 9.2 2 nd 115.4 ± 9.6 No GDM group 1 st 127.2 ± 9.3 2 nd 114.6 ± 9.9 (g/L) NR 1 st and 2 nd 1 st  = 1.1 (1.04–1.16) 2 nd  = 1.04 (0.99–1.1) 1 st ( n  = 113) Crude model = 1.19 (0.94–1.49) Model 1 = 1.12 (0.88–1.43) 2 nd ( n  = 63) Crude model = 1.13 (0.85–1.52) Model 1 = 1.06 (0.78–1.44) Age, prepregnancy BMI, family history of diabetes, parity, smoking, drinking, diastolic blood pressure, systolic blood pressure, gestational week at visit, income, and education. Si (2020) 24 GDM group 1 st 12.68 ± 0.91 2 nd 11.57 ± 0.83 No GDM group 1 st 12.54 ± 0.88 2 nd 11.43 ± 0.83 NR 1 st and 2 nd − IS in 1 st / Hb in 2 nd trimester Hb <11 ( n  = 80) 1.32 (0.67–2.51) Hb >11 ( n  = 217) 1.53 (1.05–2.24) IS in 2 nd / Hb in 2 nd trimester Hb <11 ( n  = 184) 1.23 (0.68–2.28) Hb >11 ( n  = 305) 1.92 (1.13–3.35 ) IS in 2 nd / Hb in 1 st trimester Hb < 11 ( n  = 54) 1.04 (0.13–2.21) Hb > 11 ( n  = 435) 2.15 (1.07–4.34) Maternal age, education, pre-pregnancy BMI, parity, weight gain during pregnancy, dietary iron intake, family history of diseases, isolated hypothyroxinemia in 1 st trimester, participation in physical exercise during pregnancy, gestational week of Hb measurement and OGTT measurement. Abbreviations: BMI, body mass index; CI, confidence interval; IS, iron supplementation; NR, not reported; OGTT, oral glucose tolerance test; OR, odds ratio; RR, relative risk. Based on ROB-2, no domain was classified as low risk of bias for all randomized clinical trial evaluated; 2 of the studies presented high risk of bias in relation to measurement of the outcome, 29 30 and 1 study presented high risk of bias due to missing outcome data. 30 ( Fig. 2 ) Regarding the NRCT, the evaluation based on ROBINS-I classified the study as serious risk of bias due confounding and classification of interventions, and low risk of bias in selection of participants into the study and deviations from intended interventions. No information was reported on missing data and measurement of outcomes to allow the evaluation of bias on these domains. 28 Fig. 2 Graphical representation of risk of bias assessment of the Randomized Clinical Trials included in the review ( n  = 3). Source of bias risk assessment: Cochrane Collaboration tool. Based on the NIH tool, cohort studies had a low risk of bias for most of the questions, mainly the study of Zhu et al. 25 Liu and Pang 23 did not clearly define their exposure measures, did not realize adjustments considered important in the analyses, such as parity and maternal age, and presented a high risk for evaluation of repeated exposure, that is, were not performed multiple measurements of the hemoglobin and/or ferritin, which would provide greater confidence that the exposure status was correctly classified. 23 Si et al. 24 did not provide information on loss of follow-up after the baseline and, presented high risk in different levels of the exposure of interest and the participation rate at least 50% of eligible persons either ( Fig. 3 ). In the case-control studies, Özyiğit et al. 26 selected the study sample from exposure to the use of iron, and Rawal et al. 27 did not have clearly defined and differentiated cases of controls; furthermore, both studies did not report a sample size justification. 26 27 ( Fig. 4 ) Fig. 3 Graphical representation of risk of bias assessment of the Cohort Studies included in the review ( n  = 3). Source of quality assessment: National Institutes of Health (NIH). *Abbreviations: CD, cannot determine; NR, not reported; NA, not applicable. Fig. 4 Graphical representation of risk of bias assessment of the Case-Control Studies included in the review ( n  = 2). Source of quality assessment: National Institutes of Health (NIH). *Abbreviations: CD, cannot determine; NR, not reported; NA, not applicable. Discussion A systematic review with a robust search strategy was conducted to evaluate the association between iron supplementation, hemoglobin levels and/or serum ferritin, and the risk of developing GDM. There were few studies of which only three were RCTs with exposure to the use of iron supplementation and information on iron biomarkers (ferritin and/or hemoglobin). The association investigated in this review has biological plausibility. Iron has a high oxidation capacity, and its free form can catalyze the formation of free radicals, which may lead to cellular damage, also known as oxidative stress. It is known that pregnancy by itself is a condition that favors the occurrence of oxidative stress because the placenta is rich in mitochondria, as well as for the fact that the transition metals, especially iron, are particularly abundant in the placenta, which would already boost the production of free radicals. 11 Thus, one of the suggested mechanisms involves the formation of the hydroxyl radical, which can damage β-cell membranes, affecting the synthesis and secretion of insulin by the pancreas. 11 Another proposed mechanism would be more peripheral. The deposition of iron in the muscle would damage the muscle tissue, which would consequently lead to a decrease in glucose uptake. Additionally, insulin would be responsible for stimulating iron uptake by the cell, which could lead to an even greater accumulation of cellular iron, forming vicious cycle that could induce insulin resistance and diabetes mellitus. 31 32 Most studies included in our review did not find a positive association between iron supplementation during pregnancy and GDM. We did find a positive association between hemoglobin or ferritin levels and increased risk of developing GDM, based on cohort and case-control studies; however, the same association was not found in RCTs. The positive association found in the case-control studies 26 27 may reflect the condition in which the increased ferritin in GDM was, as it is an acute phase reactant and the increased levels seen in women with GDM were a result of the inflammation associated with the disease. 13 Among the reasons for the differences found in the types of studies, the methodological aspects stand out; RCT is a type of study very similar to prospective cohort studies, with the difference being that its design makes it possible to remove several biases, such as confounding and selection bias, since the treatment and control groups are allocated using random techniques and the characteristics are similarly distributed in both groups. However, the RCTs included in this review presented methodological problems that may have compromised their ability to assess the causality between iron supplementation, elevated hemoglobin and ferritin, and the development of GDM. Among the problems found, stands out the bias in the measure of GDM. This flaw was detected in nonrandomized studies too. Some studies did not report what was considered for diagnosis, 23 28 30 and among those that did, some used the values of the OGTT test performed between the 24 th and 28 th gestational week for the diagnosis of GDM, 24 25 26 others assessed OGTT at different gestational weeks, 13 27 or considered other criteria, such as investigating GDM only in pregnant women with risk factors. 29 These inconsistencies regarding the type of method and screening period for the diagnosis of GDM may cause misclassification bias. In the NRCT, there was not enough information regarding missing data and measurement of outcomes to classify as serious or critical risk of bias. The measure of the diagnostic of GDM and the initial sample as well as the missing data were not related. 28 Regarding the case-control studies, the selection bias was the most important one, mainly regarding the representativeness. Information about sample size justification 26 27 and a specific description of case and control groups was not provided. 27 Furthermore, one of the included studies selected the cases and controls regarding the exposure. 26 Among cohort studies, although Si et al. 24 showed a positive association between iron supplementation and GDM in women with Hb > 11, some methodological flaws cause concerns regarding the rate of eligible persons and the follow-up baseline being less than 20%. Other limitations related to the revised data and the review process must be considered. Regarding the reviewed studies, we identified that there was no standard for the hemoglobin and/or ferritin collection in the trimester of pregnancy, which did not allow us to evaluate the dose–response effect of these parameters on the use of iron supplementation. The time of use, period of onset, and dosage of iron supplement were also different among the studies evaluated. Despite this review only including studies that assess ferritin and/or hemoglobin, two studies reported information on dietary iron intake. 13 24 There is a demand for studies that assess multiple factors that could influence iron status in pregnant women. Distinguishing the origin of iron, whether diet or supplementation, is a limiting factor. Thus, studies including the use of other parameters to assess the iron status and evaluated stores, supplementation, and dietary iron intake, as well as its role in the development of GDM are required. Although we performed a comprehensive review, including various databases, gray literature, and with no language restriction, a possible publication bias cannot be ruled out. On the other hand, this review, unlike other existing ones, 33 34 used as in inclusion criterion the presence of at least one hematological parameter of iron (hemoglobin and/or ferritin), commonly required during antenatal care and, as main exposure, the use of iron through supplements. Additionally, this is the only review on the topic that has included observational and interventional studies. Conclusion Based on the studies included in this review and the critical methodological problems, the association between iron supplementation and GDM remains undetermined. Therefore, this review highlights the necessity for more studies evaluating the relationship between iron supplementation and GDM. Conflicts to Interest None to declare ==== Refs References 1 World Health Organization Diagnostic criteria and classification of hyperglycaemia first detected in pregnancy: a World Health Organization Guideline Diabetes Res Clin Pract 2014 103 03 341 363 10.1016/j.diabres.2013.10.012 24847517 2 Hod M Kapur A Sacks D A The International Federation of Gynecology and Obstetrics (FIGO) Initiative on gestational diabetes mellitus: a pragmatic guide for diagnosis, management, and care Int J Gynaecol Obstet 2015 131 (Suppl 3):S173 S211 10.1016/S0020-7292(15)30033-3 3 Reece E A Leguizamón G Wiznitzer A Gestational diabetes: the need for a common ground Lancet 2009 373 (9677):1789 1797 10.1016/S0140-6736(09)60515-8 19465234 4 American Diabetes Association Diagnosis and classification of diabetes mellitus Diabetes Care 2014 37 (Suppl 1):S81 S90 10.2337/dc14-S081 24357215 5 Guariguata L Linnenkamp U Beagley J Whiting D R Cho N H Global estimates of the prevalence of hyperglycaemia in pregnancy Diabetes Res Clin Pract 2014 103 02 176 185 10.1016/j.diabres.2013.11.003 24300020 6 Quan W Zeng M Jiao Y Western dietary patterns, foods, and risk of gestational diabetes mellitus: a systematic review and meta-analysis of prospective cohort studies Adv Nutr 2021 12 04 1353 1364 10.1093/advances/nmaa184 33578428 7 Diabetes Canada Clinical Practice Guidelines Expert Committee Feig D S Berger H Donovan L Diabetes and Pregnancy Can J Diabetes 2018 42 (Suppl 1):S255 S282 10.1016/j.jcjd.2017.10.038 29650105 8 Afkhami-Ardekani M Rashidi M Iron status in women with and without gestational diabetes mellitus J Diabetes Complications 2009 23 03 194 198 10.1016/j.jdiacomp.2007.11.006 18413178 9 Amiri F N Basirat Z Omidvar S Sharbatdaran M Tilaki K H Pouramir M Comparison of the serum iron, ferritin levels and total iron-binding capacity between pregnant women with and without gestational diabetes J Nat Sci Biol Med 2013 4 02 302 305 10.4103/0976-9668.116977 24082721 10 Lao T T Chan P L Tam K F Gestational diabetes mellitus in the last trimester - a feature of maternal iron excess? Diabet Med 2001 18 03 218 223 10.1046/j.1464-5491.2001.00453.x 11318843 11 Casanueva E Viteri F E Iron and oxidative stress in pregnancy J Nutr 2003 133 (05, Suppl 2):1700S 1708S 10.1093/jn/133.5.1700S 12730487 12 Behboudi-Gandevani S Safary K Moghaddam-Banaem L Lamyian M Goshtasebi A Alian-Moghaddam N The relationship between maternal serum iron and zinc levels and their nutritional intakes in early pregnancy with gestational diabetes Biol Trace Elem Res 2013 154 01 7 13 10.1007/s12011-013-9703-y 23743666 13 Chan K K Chan B C Lam K F Tam S Lao T T Iron supplement in pregnancy and development of gestational diabetes–a randomised placebo-controlled trial BJOG 2009 116 06 789 797, discussion 797–79810.1111/j.1471-0528.2008.02014.x 19432567 14 Chen X Scholl T O Stein T P Association of elevated serum ferritin levels and the risk of gestational diabetes mellitus in pregnant women: The Camden study Diabetes Care 2006 29 05 1077 1082 10.2337/diacare.2951077 16644640 15 Helin A Kinnunen T I Raitanen J Ahonen S Virtanen S M Luoto R Iron intake, haemoglobin and risk of gestational diabetes: a prospective cohort study BMJ Open 2012 2 05 e001730 10.1136/bmjopen-2012-001730 16 Milman N Iron prophylaxis in pregnancy–general or individual and in which dose? Ann Hematol 2006 85 12 821 828 10.1007/s00277-006-0145-x 16763841 17 Page M J Moher D Bossuyt P M PRISMA 2020 explanation and elaboration: updated guidance and exemplars for reporting systematic reviews BMJ 2021 372 160 n160 10.1136/bmj.n160 33781993 18 Ouzzani M Hammady H Fedorowicz Z Elmagarmid A Rayyan-a web and mobile app for systematic reviews Syst Rev 2016 5 01 210 10.1186/s13643-016-0384-4 27919275 19 Sterne J AC Savović J Page M J RoB 2: a revised tool for assessing risk of bias in randomised trials BMJ 2019 366 l4898 10.1136/bmj.l4898 31462531 20 Sterne J A Hernán M A Reeves B C ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions BMJ 2016 355 i4919 10.1136/bmj.i4919 27733354 21 National Heart, Lung, and Blood Institute Study quality assessment tools [Internet] 2021[2022 Jan 26]. Available from:https://www.nhlbi.nih.gov/health-topics/study-quality-assessment-tools 22 McGuinness L A Higgins J PT Risk-of-bias VISualization (robvis): An R package and Shiny web app for visualizing risk-of-bias assessments Res Synth Methods 2021 12 01 55 61 10.1002/jrsm.1411 32336025 23 Liu X N Pang J A retrospective study of supplemental iron intake in singleton pregnancy women with risk of developing gestational diabetes mellitus Medicine (Baltimore) 2018 97 26 e10819 10.1097/MD.0000000000010819 29952938 24 Si S Shen Y Xin X Hemoglobin concentration and iron supplement during pregnancy were associated with an increased risk of gestational diabetes mellitus J Diabetes 2021 13 03 211 221 10.1111/1753-0407.13101 32755052 25 Zhu B Liang C Xia X Iron-Related Factors in Early Pregnancy and Subsequent Risk of Gestational Diabetes Mellitus: the Ma'anshan Birth Cohort (MABC) Study Biol Trace Elem Res 2019 191 01 45 53 10.1007/s12011-018-1595-4 30515713 26 Özyiğit E A Uğur M Ünlü S Özakşiï T G Avşar F The effect of oral iron supplementation on the glucose metabolism in non-anemic pregnant women: a prospective case-control study UHOD 2008 18 03 155 162 27 Rawal S Hinkle S N Bao W A longitudinal study of iron status during pregnancy and the risk of gestational diabetes: findings from a prospective, multiracial cohort Diabetologia 2017 60 02 249 257 10.1007/s00125-016-4149-3 27830277 28 Asadi N Vafaei H Kasraeian M Yoosefi S Faraji A Abbasi L Effects of prophylactic iron supplementation on outcome of nonanemic pregnant women: A non-randomized clinical trial J Chin Med Assoc 2019 82 11 840 844 10.1097/JCMA.0000000000000184 31517773 29 Kinnunen T I Luoto R Helin A Hemminki E Supplemental iron intake and the risk of glucose intolerance in pregnancy: re-analysis of a randomised controlled trial in Finland Matern Child Nutr 2016 12 01 74 84 10.1111/mcn.12139 24995700 30 Ouladsahebmadarek E Sayyah-Melli M Abbasalizadeh S Seyedhejazie M The effect of supplemental iron elimination on pregnancy outcome Pak J Med Sci 2011 27 03 641 645 31 Fernández-Real J M López-Bermejo A Ricart W Cross-talk between iron metabolism and diabetes Diabetes 2002 51 08 2348 2354 10.2337/diabetes.51.8.2348 12145144 32 Swaminathan S Fonseca V A Alam M G Shah S V The role of iron in diabetes and its complications Diabetes Care 2007 30 07 1926 1933 10.2337/dc06-2625 17429063 33 Fernández-Cao J C Aranda N Ribot B Tous M Arija V Elevated iron status and risk of gestational diabetes mellitus: A systematic review and meta-analysis Matern Child Nutr 2017 13 04 e12400 10.1111/mcn.12400 27966840 34 Fu S Li F Zhou J Liu Z The relationship between body iron status, iron intake and gestational diabetes: a systematic review and meta-analysis Medicine (Baltimore) 2016 95 02 e2383 10.1097/MD.0000000000002383 26765415
PMC009xxxxxx/PMC9800070.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36580951 10.1055/s-0042-1759773 RBGO-22-0105 Review Article Mastology How to Reach the Best Ultrasound Performance in the Delivery Room Como obter o melhor desempenho do ultrassom na sala de partohttp://orcid.org/0000-0002-3309-2613 Santana Eduardo Félix Martins 12 http://orcid.org/0000-0002-3705-8615 Castello Renata Gomes 2 http://orcid.org/0000-0002-4785-5386 Passos Maria Eduarda Tenório 2 http://orcid.org/0000-0002-0651-821X Ribeiro Gabriela Carneiro Freitas 1 http://orcid.org/0000-0002-6145-2532 Araujo Júnior Edward 34 1 Medical Course, Albert Einstein Medical School, São Paulo, SP, Brazil. 2 Fetal Medicine Unit, Albert Einstein Hospital, São Paulo, SP, Brazil. 3 Department of Obstetrics, Paulista School of Medicine, Federal University of Sao Paulo, São Paulo, SP, Brazil. 4 Medical Course, Municipal University of Sao Caetano do Sul, São Paulo, SP, Brazil. Address for correspondence Edward Araujo Junior, PhD Rua Borges Lagoa, 1341, 04038-034, Vila Clementino, São Paulo, [email protected] 29 12 2022 11 2022 1 12 2022 44 11 10701077 19 3 2022 21 9 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Ultrasonography is an instrument that is present in the maternal-fetal assessment throughout pregnancy and with widely documented benefits, but its use in intrapartum is becoming increasingly relevant. From the assessment of labor progression to the assessment of placental disorders, ultrasound can be used to correlate with physiological findings and physical examination, as its benefit in the delivery room cannot yet be proven. There are still few professionals with adequate training for its use in the delivery room and for the correct interpretation of data. Thus, this article aims to present a review of the entire applicability of ultrasound in the delivery room, considering the main stages of labor. There is still limited research in evidence-based medicine of its various possible uses in intrapartum, but it is expected that further studies can bring improvements in the quality of maternal and neonatal health during labor. Resumo A ultrassonografia é um instrumento que está presente na avaliação materno-fetal durante toda a gestação e com benefícios largamente documentados, porém sua utilização no intraparto vem sendo cada vez mais pertinente. Desde a avaliação de progressão de trabalho de parto a avaliação das desordens placentárias, a ultrassonografia pode ser empregada correlacionando com os achados fisiológicos e do exame físico, pois o seu benefício na sala de parto ainda não pode ser comprovado. Há ainda poucos profissionais com treinamento adequado para seu uso na sala de parto e para interpretação correta dos dados. Dessa forma, este artigo tem como finalidade apresentar uma revisão de toda a aplicabilidade do ultrassom na sala de parto, considerando as principais etapas do trabalho de parto. Ainda são limitadas as pesquisas em medicina baseada em evidências sobre os diversos usos possíveis no intraparto, mas espera-se que novos estudos possam trazer melhorias na qualidade da saúde materno-neonatal durante o trabalho de parto. Keywords ultrasound delivery room labor placental disorders Palavras-chave ultrassom sala de parto parto desordens placentárias ==== Body pmcIntroduction The use of intrapartum ultrasound has been widely reported as an additional tool for predicting the evolution of successful labor. 1 The sonographic evaluation was not shown to be superior to the vaginal examination (VE), but complementary, as the first is better for the evaluation of head station, position, and caput succedaneum, while cervix dilatation in the active stage of labor (> 4 cm) is better assessed by VE. 2 Sonographic assistance during the first and second stages of labor has the potential to improve labor outcomes, although its real benefits have not yet been proven in large randomized trials. 3 4 5 6 In contrast, the intrapartum Doppler assessment has shown no benefit in perinatal outcomes. 7 Differentiated normal and abnormal sonographic postpartum findings can also be an extra implement for the patients' well-being when the clinical evaluations are doubtful. 8 Despite great acceptability by patients, 9 specially during stressful situations such as prolonged labor (more than 12 hours from the beginning of active phase of the first stage) and unplanned operative delivery, 10 the use of intrapartum ultrasound requires a steep learning curve for good reproducibility; thus, younger obstetricians prefer to rely on clinical and digital examinations, 11 12 even though ultrasound has been proven to be more reliable than VE. The aim of this article is to present a revision of all ultrasound applicability in the delivery room, considering the main stages of labor. Placenta and Cord Anomalies Placenta and cord anomalies are associated with 30% of intrauterine death risk factors and a high risk of cerebral palsy. 13 Therefore, they are a great cause of concern during prenatal and intrapartum period. The best time to diagnose placental implantations abnormalities is during the second trimester of pregnancy, ideally with a gestational age between 18 and 26 weeks, 14 when is still possible to program the optimum time to perform cesarean section (c-section)—usually around 36 gestational weeks—modifying the neonatal and obstetric outcomes. The umbilical cord is protected from trauma and compression through the presence of the Warthon jelly and spiraling of blood vessels. 15 Literature has shown that both hypocoiled cords (spiral index below the 10 th percentile) and hypercoiled cords (spiral index above the 90 th percentile) are associated with unfavorable neonatal outcomes, 16 17 18 such as higher rates of fetal growth restriction, fetal death, intrapartum fetal heart decelerations, karyotype abnormalities, 19 low birth weight (< 2,500 g), and Appearance, Pulse, Grimace, Activity, and Respiration (APGAR) score < 7 on the 1 st and 5 th minutes of life. 15 A prenatal ultrasound assessment of cord coiling is possible; however, no benefit was found in this diagnostic screening since there are no revised means to prevent intrauterine death or a nonreassuring pattern of fetal heart rate in these cases. 20 While 97% of vasa previa cases are diagnosed during prenatal scanning, 12 the benefit of performing the intrapartum diagnosis to foresee possible complications such as maternal bleeding, fetal bleeding, and neonatal death is questioned. Due to the low prevalence of this pathology (0.02–0.27% of all pregnancies), 12 prenatal screening through transvaginal ultrasound becomes unfeasible and is recommended only for women at high risk: in vitro fertilization pregnancies, placenta previa, placenta with accessory lobe, velamentous cord insertion, and multiple gestations. 21 22 Data in the literature are very vague about intrapartum diagnosis of vasa previa using the Doppler ultrasound, with only two case reports. 23 24 In both cases, the correct diagnostic enables the performance of c-section before the rupture of the vasa previa, with a favorable outcome for the maternal-fetal binominal. Another condition that can lead to risk of maternal and fetal life due to bleeding is placental abruption, present in 0.4 to 1% of all pregnancies. 25 26 The sonographic visualization of retroplacental clots is a finding present in only 15 to 25% of cases and does not interfere with the conduct regarding the interruption of pregnancy, both in term and preterm pregnancies, since maternal and fetal conditions are more important for clinical management. 27 28 The intrapartum ultrasound represents a sensitivity of less than 30% for the diagnosis of placental abruption, and the clinical diagnosis remains the gold standard of this obstetric emergency. 29 On the other hand, the benefit of intrapartum ultrasound use has been proven in relation to the diagnosis of nuchal cord, with a sensibility of 90.2 to 96.8% when using the Doppler mode. 16 30 This finding is present in 22 to 45% of all pregnancies, and it is known that single nuchal cord is not associated with unfavorable perinatal outcomes. 16 31 32 33 However, multiple nuchal cord is associated with worse outcomes, such as perinatal mortality, APGAR score < 7 on the 1 st and 5 th minutes of life, fetal distress, and meconium ( Figure 1 ). 16 34 35 36 37 Fig. 1 Ultrasound imaging showing multiple nuchal cord loops. The intrapartum diagnostic of nuchal cord is a good tool in situations of variable deceleration in cardiotocography during labor, as it helps to recognize the cases in which the cardiotocographic pattern is not reassuring due to fetal distress and the cases when the deceleration is due to the presence of nuchal cord. 34 Lastly, the umbilical cord prolapse is a rare situation that affects 0.12 to 0.62% of all pregnancies, with a mortality rate of up to 10% due to compression of the umbilical cord. 20 Some risk factors for this comorbidity are polyhydramnios, prematurity, multiparity, multiple pregnancies, breech presentation, and low birth weight (< 2,500 g). 19 20 38 39 The evident umbilical cord prolapse occurs when the umbilical cord passes between the fetal parts after the premature rupture of membranes and the diagnosis is possible through the VE, while the occult umbilical cord prolapse occurs when the membranes are intact but the cord is ahead of fetal presentation, and the diagnosis is made by ultrasound. 20 The literature has shown low accuracy for the diagnosis of cord prolapse in routine ultrasound, 40 but has shown benefit in the use of transvaginal ultrasound to predict occult umbilical in breech presentation, 41 and the results were better when the occult cord prolapse was previously diagnosed when compared with the evident cord prolapse, suggesting that in high-risk situations, ultrasound evaluation could improve the neonatal outcomes. 42 Fetal Wellbeing During the Labor The use of Doppler ultrasound during the labor is still limited for research purposes. However, new studies are emerging, and the application of Doppler is increasingly being studied at this time. Sütterlin et al. 43 evaluated 70 pregnant women in early labor between 38 and 41 weeks of gestation, obtaining Doppler waveforms before and during abnormal fetal heart rate patterns. When an oxygen saturation level of < 30% was maintained for more than 2 minutes, the middle cerebral artery Doppler indices were reversed, indicating morbid fetal hypoxia. These results were considered consistent with the concept that the fetus maintains the oxygen supply to the brain by redistributing blood flow during active labor. Chainarong and Petpichetchian 44 evaluated the cerebroplacental ratio (CPR) during the labor, and no association was found between CPR and adverse perinatal outcomes with any CPR cut-off values. This study found that fetuses that ended up in a non-reassuring state, necessitating operative delivery, had significantly lower CPR compared with fetuses that did not. Dall'Asta et al. 45 studied the relationship between CPR measured at the beginning of labor and perinatal and delivery outcomes in a cohort of uncomplicated term pregnancies with a single child. The study's conclusion suggests that reduced CPR by itself, although associated with an increased risk of intrapartum distress, represents a poor predictor of adverse perinatal outcomes. Cochrane review assessed the effectiveness of fetal movement monitoring and Doppler ultrasound for the detection and surveillance of high-risk pregnancies and their effect in preventing stillbirths. The combined results of 16 studies showed that the umbilical arterial Doppler assessment in high-risk pregnancies leads to a 29% reduction in perinatal mortality compared with no Doppler assessment. 46 Intrapartum ultrasound (including Doppler) allowed for a greater understanding of the complex physiology of childbirth. Although promising, neither maternal nor fetal intrapartum Doppler has played a role in the true management of intrapartum ultrasound to date. 43 Labor Progression Through Ultrasound While digital VE are uncomfortable and subjective exams, 47 with an error rate ranging from 26.6 48 to 33.5% 49 due to interexaminer reproducibility, sonographic measurements are more reliable and could be an additional tool for the evaluation and estimation of a successful labor. 50 Besides, multiple digital VE are associated with ascending infection to the fetus and the uterus 51 52 and are contraindicated in some situations, such as preterm prelabor rupture of membranes and placenta previa. 53 It is possible to get valuable information that could not be obtained in a VE, such as angle of progression (AoP) which is the angle between a line in the midline of the pubic symphysis and a line running tangentially from the anterior edge of the symphysis to the fetal skull evaluated through transperineal ultrasound ( Figure 2 ); the head progression distance (HPD) which is the shortest distance between the infrapubic line and the leading edge of the fetal skull, also evaluated through transperineal ultrasound; and the head direction (HD) which is the angle between the infrapubic line, perpendicular to the most caudal part of the pubic symphysis, and a line drawn perpendicular to the widest diameter of fetal head, evaluated through abdominal ultrasound. 47 54 The AoP is the most useful measure to predict the success of vaginal delivery, with the manual parasagittal technique being the most reliable, 55 in which the angle is formed between a line drawn along the superior-inferior axis of the pubic bone and a line drawn along the inferior end of the hyperechogenic pelvic bone forming the vertex of the angle with the fetal head. Fig. 2 Ultrasound imaging demonstrating the angle of progression (AoP) access. A systematic review has shown that ultrasound is superior to digital VE for evaluation of fetal head position in the first stage of labor, in addition to the great agreement between the two methods in the assessment of cervical dilatation and a moderate correlation for fetal head station. 48 Although the success rate of digital VE increases with the progression of cervical dilatation, approximately a quarter of digital assessments differ by more than 45° when compared with the sonographic evaluation of fetal head position, 48 49 53 which may lead to unfavorable neonatal outcomes, especially when instrumented deliveries are necessary. 53 Nevertheless, ultrasound evaluation is associated with higher rates of instrumental vaginal delivery, 56 with no difference in maternal and neonatal morbidities when compared with exclusive VE evaluation before operative delivery, 56 57 58 nor in relation to c-section rate, even though ultrasound evaluation corresponds to a greater success in the diagnoses of fetal head position and station. 58 Kameyama et al. 54 described an optimal cut-off from prediction of spontaneous vaginal delivery of 83° for HPD (positive predictive value, PPV = 92.9%), 56 mm for PD (PPV = 94.4%) and AoP of 146° (PPV = 94.3%) right after full cervical dilatation. Ghi et al. 59 have shown that women with spontaneous vaginal delivery had a wider AoP in the begging of second stage of labor (140° ± 20.2°) than the women who had been submitted to operative delivery (122.9° ± 16.7°). Sainz et al. 60 have found that an AoP of 122° ( ± 17.8°) is associated with a complicated operative delivery in nulliparous woman, while an AoP of 149.2° ( ± 15.6°) and a HPD of 50.5 mm are good predictors of uncomplicated deliveries. These facts are consistent with the findings Bultez et al., 61 in which the median of the AoP of 145° is associated with a successful delivery with vacuum extraction, whereas the median of AoP of 136° corresponds to vacuum extraction failure. On the other hand, Kalache et al. 62 described an AoP of 120° as leading to the probability of an easy and successful vacuum or spontaneous vaginal delivery in 90% of the cases. Chan et al. 63 have shown that parasagittal AoP is an independent predictor for c-section and for non-progression before induction of labor: women with manual parasagittal AoP of 102° (93–111°) and automated parasagittal AoP of 108° (99–115°) were more likely to give birth through vaginal delivery, while women with manual parasagittal AoP of 93° (90–102°) and automated parasagittal AoP of 99° (93–104°) were submitted to c-sections, with no difference between nulliparous and multiparous women. Tse et al. 64 have also shown an additional decrease of 5.28° in the parasagittal AoP and an additional increase of 0.27 cm in HPD for a unit increase in fetal head station and cervical dilation in women requiring c-section, while the additional decrease was 1.35° in the parasagittal AoP and the additional increase was 0.12 cm in HPD in women who achieved vaginal delivery. Birth weight is an important predictor of neonatal morbidity and mortality and has a strong influence on obstetric and neonatal management. 65 Stubert et al. 66 confirmed that the ultrasound-derived estimated fetal weight during labor at term is an appropriate diagnostic tool, with an average accuracy of 70% within a relative difference of ± 10% to the real birth weight. Furthermore, term-estimated fetal weight has been shown to be unreliable for predicting macrosomia and is therefore not recommended. 66 Considering international guidelines, the cesarean delivery rate should not be higher if fetal weight is estimated immediately before delivery. However, overestimation of fetal weight was associated with an increased risk of c-section. 65 In this study, the increase in the rate of c-section was not accompanied by a decrease in fetal or maternal morbidity. No differences were observed in shoulder dystocia and in third- and fourth-degree perineal lacerations. 66 Yang et al. 67 found that biparietal diameter, abdominal circumference, and estimated fetal weight at 38 weeks of gestation were associated with c-sections for failure to progress in labor after adjusting for confounders. Routine biometry may help identify patients whose intrapartum c-section risk could be reduced by elective induction at 39 weeks. Faschingbauer et al. 65 found that the best results regarding intrapartum estimated fetal weight can be obtained with formulas that use biparietal diameter as the only head measurement. Little et al. 68 suggest that provider knowledge may be associated with a higher rate of c-section; therefore, limiting ultrasound check of fetal weight in the short term may help reduce c-section rate. A different use of ultrasound during the labor is by creating a sonopartogram, which is a conformation of the conventional partogram, with the use of ultrasound parameters of recording assessments during the labor. 69 70 It is possible to evaluate cervical dilatation, fetal head rotation, and fetal head descent, as it is in the conventional partogram, as well as to evaluate caput and molding 69 70 ( Figure 3 ). Although a good agreement was shown between VE and ultrasound evaluation regarding cervical dilatation and head rotation during the first period of labor, the evaluation of head descent was better estimated by VE. 60 Another possibility for the use of ultrasound in the delivery room would be the prediction of success for vaginal delivery on patients with leiomyomas located in regions close to cervix; however, we did not find any data about this topic ( Figure 4 ). Fig. 3 Ultrasound image in the delivery room showing the measurement of caput succedaneum. Fig. 4 Extensive leiomyoma in the anterior uterine wall. Ultrasonography at the Immediate Postpartum Considering postpartum hemorrhage is the leading cause of maternal death worldwide, 71 it is logical to think that ultrasound in the immediate postpartum could be an extra resource to identify possible cases of hemorrhagic complications, including uterine atony, retained products of conception, uterine arteriovenous malformations, and hematomas due birth canal trauma, as well as a good tool for specific treatments, such as curettage, embolization of uterine arteries, and the use of the Bakri balloon. In the immediate postpartum period, the transabdominal approach is preferable for uterus evaluation rather than transvaginal approach. 72 The main cause of postpartum hemorrhage is uterine atony, which can be identified by the loss of at least 500 ml of blood after vaginal delivery or 1000 ml after c-section, associated with the lack of the Pinnard security globe, which represents the uterus involution and can be diagnosed by physical exam, through palpation of the uterine height. 73 74 Through sonographic exam, the mean uterine length is 16.1 ± 1.7 cm, 75 while the maximum anterior-posterior uterine dimeter reported was 9.2 cm 76 ( Figure 5 ). Concerning endometrial evaluation, some studies have shown no correlation between the duration or amount of bleeding and the presence of echogenic material diagnosed by postpartum ultrasound. 75 77 78 79 A recent systematic review has found that the upper limit for endometrial thickness (95 th centile) measured by abdominal ultrasound within 24 hours postpartum is 22 mm, 80 with no statistically significant difference between vaginal delivery or c-section, or between nulliparous and multiparous women. 77 80 Fig. 5 Immediate postpartum ultrasound: endometrial echogenic and uterus size evaluation. Another possible cause of postpartum bleeding is the presence of retained placental tissue, which can happen in approximately 1% of term deliveries. 81 The literature has shown a variable sensitivity (42–94%) and specificity (62–92%) for the use of ultrasound in the uterus evaluation after placental removal. 82 The gray scale itself is not the best option as a diagnostic method, as the appearances of retained placental tissue in the immediate postpartum are highly variable and can be represented as echogenic mass, heterogeneous mixed density mass, and normal endometrial cavity. Therefore, it could not be correlated with a need for intervention and might not change patient outcomes. 81 82 The identification of thickened endometrial echo complex > 10 mm, associated with vascular flow detection on color Doppler, is highly suggestive for retained placental tissue. 83 However, a hypervascular area can be physiologic in the postpartum period and disappear spontaneously or after removal of placental remnants 84 85 and, therefore, there is no gold standard protocol for diagnosing retained placental tissue through imaging exams. 86 As for the placenta accreta spectrum, the diagnostic should preferably happen prenatally, so the best time and place for delivery can be arranged, as an intraoperative hysterectomy might be necessary. 87 However, when there is no prenatal diagnosis and the patient goes through labor, the diagnostic is made during the third stage of labor, which may lead to major bleeding. 87 88 The normal placental separation can be sonographically characterized by decreased blood flow while the placenta is detaching from the myometrium, whereas the presence of placenta accreta spectrum can be sonographically characterized by the persistent blood flow between the placenta and the myometrium. 88 A rare, but serious situation of postpartum hemorrhage, is the uterine rupture. Its prevalence is less than 1% after a vaginal delivery after one c-section, increasing to up to 2% when the vaginal delivery occurs after more than two previous c-sections. 89 It should be suspected in patients with vaginal delivery after c-section or any uterine surgery, presenting postpartum hemorrhage and hypovolemic shock. 90 The diagnostic must be done as soon as possible, and a transabdominal ultrasound would show an echo-free space or mass lesions, possibly corresponding to intraperitoneal bleeding or retroperitoneal hematoma. 90 Lastly, an unusual but possible cause of hemodynamic instability in the delivery room is the presence of a vulvar or paravaginal hematoma. 91 92 This complication might happen specially after direct injury of the perineum, from instrumental deliveries, vaginal laceration, or episiotomy. 92 The main symptoms are pelvic and perianal pain, swelling of the vulva, paravaginal mass, and urinary retention due mechanical urethral obstruction. 91 92 The use of transperineal or transabdominal ultrasound can provide precise information about the presence, location, and size of the vaginal hematoma, with similar results findings with computed tomography, but with the advantage that it can be performed in the labor ward, immediately after delivery. 91 Conclusion The performance of ultrasound in the delivery room is still a poorly explored resource in maternity hospitals. However, with the potential to improve the diagnosis and interpretations of situations and allow for more timely interventions, since it is a tool with the potential to complement (and not replace) clinical practice. There is still little evidence-based medical research on the several possibilities of its intrapartum use, but we expect that further studies could provide improvements in the quality of maternal-neonatal health during the labor. Conflict of Interests There is no conflict of interests to declare. ==== Refs References 1 Ghi T Eggebø T Lees C ISUOG Practice Guidelines: intrapartum ultrasound Ultrasound Obstet Gynecol 2018 52 01 128 139 10.1002/uog.19072 29974596 2 Usman S Wilkinson M Barton H Lees C C The feasibility and accuracy of ultrasound assessment in the labor room J Matern Fetal Neonatal Med 2019 32 20 3442 3451 10.1080/14767058.2018.1465553 29712501 3 Kahrs B H Eggebø T M Intrapartum ultrasound in women with prolonged first stage of labor Am J Obstet Gynecol MFM 2021 3 (6S):100427 10.1016/j.ajogmf.2021.100427 34273584 4 Gimovsky A C Intrapartum ultrasound for the diagnosis of cephalic malpositions and malpresentations Am J Obstet Gynecol MFM 2021 3 (6S):100438 10.1016/j.ajogmf.2021.100438 34302995 5 Chan V YT Lau W L Intrapartum ultrasound and the choice between assisted vaginal and cesarean delivery Am J Obstet Gynecol MFM 2021 3 (6S):100439 10.1016/j.ajogmf.2021.100439 34216834 6 Ghi T Intrapartum ultrasound and evidence-based medicine: a necessary but challenging marriage Am J Obstet Gynecol MFM 2021 3 (6S):100428 10.1016/j.ajogmf.2021.100428 34303655 7 Dall'Asta A Kumar S Prelabor and intrapartum Doppler ultrasound to predict fetal compromise Am J Obstet Gynecol MFM 2021 3 (6S):100479 10.1016/j.ajogmf.2021.100479 34496306 8 Steinkeler J Coldwell B J Warner M A Ultrasound of the postpartum uterus Ultrasound Q 2012 28 02 97 103 10.1097/RUQ.0b013e31824e6b7d 22572864 9 Solaiman S A Atwa K A Gad A A Al- Shatouri M. Transperineal ultrasound of fetal head progression in prolonged labor: women's acceptance and ability to predict the mode of delivery Egypt J Radiol Nucl Med 2020 51 94 10.1186/s43055-020-00215-0 10 Rizzo G Aloisio F Bacigalupi A Women's compliance with ultrasound in labor: a prospective observational study J Matern Fetal Neonatal Med 2021 34 09 1454 1458 10.1080/14767058.2019.1638903 31257977 11 Gilboa Y Perlman S Karp H Rabinovitch R Achiron R What do obstetricians really think about ultrasound in the delivery room? Isr Med Assoc J 2017 19 04 234 236 28480677 12 Plurien A Berveiller P Guerby P Ultrasound in delivery room: Does it have a place for the younger generation? J Gynecol Obstet Hum Reprod 2020 49 10 101915 10.1016/j.jogoh.2020.101915 13 Hasegawa J Ultrasound screening of umbilical cord abnormalities and delivery management Placenta 2018 62 66 78 10.1016/j.placenta.2017.12.003 29405969 14 Melcer Y Maymon R Jauniaux E Vasa previa: prenatal diagnosis and management Curr Opin Obstet Gynecol 2018 30 06 385 391 10.1097/GCO.0000000000000478 30102606 15 Degani S Lewinsky R M Berger H Spiegel D Sonographic estimation of umbilical coiling index and correlation with Doppler flow characteristics Obstet Gynecol 1995 86 06 990 993 10.1016/0029-7844(95)00307-d 7501354 16 Hayes D JL Warland J Parast M M Umbilical cord characteristics and their association with adverse pregnancy outcomes: A systematic review and meta-analysis PLoS One 2020 15 09 e0239630 10.1371/journal.pone.0239630 32970750 17 Pergialiotis V Kotrogianni P Koutaki D Christopoulos-Timogiannakis E Papantoniou N Daskalakis G Umbilical cord coiling index for the prediction of adverse pregnancy outcomes: a meta-analysis and sequential analysis J Matern Fetal Neonatal Med 2019 33 23 1 8 10.1080/14767058.2019.1594187 18 Mittal A Nanda S Sen J Antenatal umbilical coiling index as a predictor of perinatal outcome Arch Gynecol Obstet 2015 291 04 763 768 10.1007/s00404-014-3456-5 25209352 19 Moshiri M Zaidi S F Robinson T J Comprehensive imaging review of abnormalities of the umbilical cord Radiographics 2014 34 01 179 196 10.1148/rg.341125127 24428290 20 Hasegawa J Ultrasound assessment of the umbilical cord Donald School J Ultrasound Obstet Gynecol 2014 8 04 382 390 10.5005/jp-journals-10009-1378 21 Harding J A Lewis D F Major C A Crade M Patel J Nageotte M P Color flow Doppler–a useful instrument in the diagnosis of vasa previa Am J Obstet Gynecol 1990 163 (5 Pt 1):1566 1568 10.1016/0002-9378(90)90628-k 2240107 22 Ruiter L Kok N Limpens J Incidence of and risk indicators for vasa praevia: a systematic review BJOG 2016 123 08 1278 1287 10.1111/1471-0528.13829 26694639 23 Arts H van Eyck J Antenatal diagnosis of vasa previa by transvaginal color Doppler sonography Ultrasound Obstet Gynecol 1993 3 04 276 278 10.1046/j.1469-0705.1993.03040276.x 12797276 24 Baschat A A Gembruch U Ante- and intrapartum diagnosis of vasa praevia in singleton pregnancies by colour coded Doppler sonography Eur J Obstet Gynecol Reprod Biol 1998 79 01 19 25 10.1016/s0301-2115(98)00026-8 9643398 25 Tikkanen M Placental abruption: epidemiology, risk factors and consequences Acta Obstet Gynecol Scand 2011 90 02 140 149 10.1111/j.1600-0412.2010.01030.x 21241259 26 Downes K L Grantz K L Shenassa E D Maternal, labor, delivery, and perinatal outcomes associated with placental abruption: a systematic review Am J Perinatol 2017 34 10 935 957 10.1055/s-0037-1599149 28329897 27 Sholl J S Abruptio placentae: clinical management in nonacute cases Am J Obstet Gynecol 1987 156 01 40 51 10.1016/0002-9378(87)90200-6 3541621 28 Tikkanen M Nuutila M Hiilesmaa V Paavonen J Ylikorkala O Clinical presentation and risk factors of placental abruption Acta Obstet Gynecol Scand 2006 85 06 700 705 10.1080/00016340500449915 16752262 29 Qiu Y Wu L Xiao Y Zhang X Clinical analysis and classification of placental abruption J Matern Fetal Neonatal Med 2021 34 18 2952 2956 10.1080/14767058.2019.1675625 31608779 30 Jauniaux E Mawissa C Peellaerts C Rodesch F Nuchal cord in normal third-trimester pregnancy: a color Doppler imaging study Ultrasound Obstet Gynecol 1992 2 06 417 419 10.1046/j.1469-0705.1992.02060417.x 12796916 31 Clapp J F III Stepanchak W Hashimoto K Ehrenberg H Lopez B The natural history of antenatal nuchal cords Am J Obstet Gynecol 2003 189 02 488 493 10.1067/s0002-9378(03)00371-5 14520223 32 Mastrobattista J M Hollier L M Yeomans E R Effects of nuchal cord on birthweight and immediate neonatal outcomes Am J Perinatol 2005 22 02 83 85 10.1055/s-2005-837737 15731986 33 Lal N Deka D Mittal S Does the nuchal cord persist? An ultrasound and color-Doppler-based prospective study J Obstet Gynaecol Res 2008 34 03 314 317 10.1111/j.1447-0756.2007.00695.x 18686343 34 Tagliaferri S Esposito F G Esposito G Impact of nuchal cord on antenatal and intrapartum foetal heart rate surveillance and perinatal outcome J Obstet Gynaecol 2020 40 03 316 323 10.1080/01443615.2019.1621816 31976797 35 Peesay M Nuchal cord and its implications Matern Health Neonatol Perinatol 2017 3 28 10.1186/s40748-017-0068-7 29234502 36 Spellacy W N Gravem H Fisch R O The umbilical cord complications of true knots, nuchal coils, and cords around the body. Report from the collaborative study of cerebral palsy Am J Obstet Gynecol 1966 94 08 1136 1142 10.1016/0002-9378(66)90777-0 5293995 37 Pergialiotis V Fanaki M Bellos I Tzortzis A Loutradis D Daskalakis G Evaluation of umbilical cord entanglement as a predictive factor of adverse pregnancy outcomes: A meta-analysis Eur J Obstet Gynecol Reprod Biol 2019 243 150 157 10.1016/j.ejogrb.2019.10.038 31704532 38 Behbehani S Patenaude V Abenhaim H A Maternal risk factors and outcomes of umbilical cord prolapse: a population-based study J Obstet Gynaecol Can 2016 38 01 23 28 10.1016/j.jogc.2015.10.008 26872752 39 Dilbaz B Ozturkoglu E Dilbaz S Ozturk N Sivaslioglu A A Haberal A Risk factors and perinatal outcomes associated with umbilical cord prolapse Arch Gynecol Obstet 2006 274 02 104 107 10.1007/s00404-006-0142-2 16538441 40 Ezra Y Strasberg S R Farine D Does cord presentation on ultrasound predict cord prolapse? Gynecol Obstet Invest 2003 56 01 6 9 10.1159/000072323 12867760 41 Kinugasa M Sato T Tamura M Suzuki H Miyazaki Y Imanaka M Antepartum detection of cord presentation by transvaginal ultrasonography for term breech presentation: potential prediction and prevention of cord prolapse J Obstet Gynaecol Res 2007 33 05 612 618 10.1111/j.1447-0756.2007.00620.x 17845317 42 Japan Association of Obstetricians and Gynecologists, Tokyo, Japan Hasegawa J Ikeda T Sekizawa A Ishiwata I Kinoshita K Obstetric risk factors for umbilical cord prolapse: a nationwide population-based study in Japan Arch Gynecol Obstet 2016 294 03 467 472 10.1007/s00404-015-3996-3 26714678 43 Sütterlin M W Seelbach-Göbel B Oehler M K Heupel M Dietl J Doppler ultrasonographic evidence of intrapartum brain-sparing effect in fetuses with low oxygen saturation according to pulse oximetry Am J Obstet Gynecol 1999 181 01 216 220 10.1016/s0002-9378(99)70462-x 10411822 44 Chainarong N Petpichetchian C The relationship between intrapartum cerebroplacental ratio and adverse perinatal outcomes in term fetuses Eur J Obstet Gynecol Reprod Biol 2018 228 82 86 10.1016/j.ejogrb.2018.06.016 29909268 45 Dall'Asta A Ghi T Rizzo G Cerebroplacental ratio assessment in early labor in uncomplicated term pregnancy and prediction of adverse perinatal outcome: prospective multicenter study Ultrasound Obstet Gynecol 2019 53 04 481 487 10.1002/uog.19113 29900608 46 Alfirevic Z Stampalija T Dowswell T Fetal and umbilical Doppler ultrasound in high-risk pregnancies Cochrane Database Syst Rev 2017 6 06 CD007529 10.1002/14651858.CD007529.pub4 28613398 47 Wiafe Y A Whitehead B Venables H Nakua E K The effectiveness of intrapartum ultrasonography in assessing cervical dilatation, head station and position: A systematic review and meta-analysis Ultrasound 2016 24 04 222 232 10.1177/1742271X16673124 27847537 48 Akmal S Kametas N Tsoi E Hargreaves C Nicolaides K H Comparison of transvaginal digital examination with intrapartum sonography to determine fetal head position before instrumental delivery Ultrasound Obstet Gynecol 2003 21 05 437 440 10.1002/uog.103 12768552 49 Akmal S Tsoi E Kametas N Howard R Nicolaides K H Intrapartum sonography to determine fetal head position J Matern Fetal Neonatal Med 2002 12 03 172 177 10.1080/jmf.12.3.172.177 12530614 50 Yeo L Romero R Sonographic evaluation in the second stage of labor to improve the assessment of labor progress and its outcome Ultrasound Obstet Gynecol 2009 33 03 253 258 10.1002/uog.6336 19247999 51 Ahn K H Oh M J Intrapartum ultrasound: A useful method for evaluating labor progress and predicting operative vaginal delivery Obstet Gynecol Sci 2014 57 06 427 435 10.5468/ogs.2014.57.6.427 25469329 52 Benediktsdottir S Eggebø T M Salvesen KÅ Agreement between transperineal ultrasound measurements and digital examinations of cervical dilatation during labor BMC Pregnancy Childbirth 2015 15 273 10.1186/s12884-015-0704-z 26496894 53 Usman S Lees C Benefits and pitfalls of the use of intrapartum ultrasound Australas J Ultrasound Med 2015 18 02 53 59 10.1002/j.2205-0140.2015.tb00042.x 28191241 54 Kameyama S Sato A Miura H Prediction of spontaneous vaginal delivery by transperineal ultrasound performed just after full cervical dilatation is determined J Med Ultrason 2016 43 02 243 248 55 Frick A Kostiv V Vojtassakova D Akolekar R Nicolaides K H Comparison of different methods of measuring angle of progression in prediction of labor outcome Ultrasound Obstet Gynecol 2020 55 03 391 400 10.1002/uog.21913 31692170 56 Popowski T Porcher R Fort J Javoise S Rozenberg P Influence of ultrasound determination of fetal head position on mode of delivery: a pragmatic randomized trial Ultrasound Obstet Gynecol 2015 46 05 520 525 10.1002/uog.14785 25583399 57 Ramphul M Ooi P V Burke G Instrumental delivery and ultrasound : a multicentre randomised controlled trial of ultrasound assessment of the fetal head position versus standard care as an approach to prevent morbidity at instrumental delivery BJOG 2014 121 08 1029 1038 10.1111/1471-0528.12810 24720273 58 Mappa I Tartaglia S Maqina P Ultrasound vs routine care before instrumental vaginal delivery: A systematic review and meta-analysis Acta Obstet Gynecol Scand 2021 100 11 1941 1948 10.1111/aogs.14236 34314520 59 Ghi T Youssef A Maroni E Intrapartum transperineal ultrasound assessment of fetal head progression in active second stage of labor and mode of delivery Ultrasound Obstet Gynecol 2013 41 04 430 435 10.1002/uog.12379 23288706 60 Sainz J A García-Mejido J A Aquise A Intrapartum transperineal ultrasound used to predict cases of complicated operative (vacuum and forceps) deliveries in nulliparous women Acta Obstet Gynecol Scand 2017 96 12 1490 1497 10.1111/aogs.13230 28889406 61 Bultez T Quibel T Bouhanna P Popowski T Resche-Rigon M Rozenberg P Angle of fetal head progression measured using transperineal ultrasound as a predictive factor of vacuum extraction failure Ultrasound Obstet Gynecol 2016 48 01 86 91 10.1002/uog.14951 26183426 62 Kalache K D Dückelmann A M Michaelis S A Lange J Cichon G Dudenhausen J W Transperineal ultrasound imaging in prolonged second stage of labor with occipitoanterior presenting fetuses: how well does the ‘angle of progression’ predict the mode of delivery? Ultrasound Obstet Gynecol 2009 33 03 326 330 10.1002/uog.6294 19224527 63 Chan W WY Chaemsaithong P Lim W T Pre-induction transperineal ultrasound assessment of the prediction of labor outcome Fetal Diagn Ther 2019 45 04 256 267 10.1159/000489122 30121653 64 Tse W T Chaemsaithong P Chan W WY Labor progress determined by ultrasound is different in women requiring cesarean delivery from those who experience a vaginal delivery following induction of labor Am J Obstet Gynecol 2019 221 04 3350 3.35E20 10.1016/j.ajog.2019.05.040 65 Faschingbauer F Dammer U Raabe E Intrapartum sonographic weight estimation Arch Gynecol Obstet 2015 292 04 805 811 10.1007/s00404-015-3720-3 25870017 66 Stubert J Peschel A Bolz M Glass Ä Gerber B Accuracy of immediate antepartum ultrasound estimated fetal weight and its impact on mode of delivery and outcome - a cohort analysis BMC Pregnancy Childbirth 2018 18 01 118 10.1186/s12884-018-1772-7 29716537 67 Yang J M Hyett J A Mcgeechan K Phipps H de Vries B S Is ultrasound measured fetal biometry predictive of intrapartum caesarean section for failure to progress? Aust N Z J Obstet Gynaecol 2018 58 06 620 628 10.1111/ajo.12776 29355895 68 Little S E Edlow A G Thomas A M Smith N A Estimated fetal weight by ultrasound: a modifiable risk factor for cesarean delivery? Am J Obstet Gynecol 2012 207 04 3090 3.09E8 10.1016/j.ajog.2012.06.065 69 Hassan W A Eggebø T Ferguson M The sonopartogram: a novel method for recording progress of labor by ultrasound Ultrasound Obstet Gynecol 2014 43 02 189 194 10.1002/uog.13212 24105734 70 Dira L M Tudorache S Antsaklis P Sonographic Evaluation of the Mechanism of Active Labor (SonoLabor Study): observational study protocol regarding the implementation of the sonopartogram BMJ Open 2021 11 09 e047188 10.1136/bmjopen-2020-047188 71 Urner F Zimmermann R Krafft A Manual removal of the placenta after vaginal delivery: an unsolved problem in obstetrics J Pregnancy 2014 2014 274651 10.1155/2014/274651 24812585 72 Mulic-Lutvica A Postpartum ultrasound Donald School J Ultrasound Obstet Gynecol 2012 6 01 76 92 73 Üçyiğit A Johns J The postpartum ultrasound scan Ultrasound 2016 24 03 163 169 10.1177/1742271X16653779 27867409 74 Plunk M Lee J H Kani K Dighe M Imaging of postpartum complications: a multimodality review AJR Am J Roentgenol 2013 200 02 W143-54 10.2214/AJR.12.9637 23345378 75 Sokol E R Casele H Haney E I Ultrasound examination of the postpartum uterus: what is normal? J Matern Fetal Neonatal Med 2004 15 02 95 99 10.1080/14767050310001650798 15209115 76 Mulic-Lutvica A Bekuretsion M Bakos O Axelsson O Ultrasonic evaluation of the uterus and uterine cavity after normal, vaginal delivery Ultrasound Obstet Gynecol 2001 18 05 491 498 10.1046/j.0960-7692.2001.00561.x 11844171 77 Deans R Dietz H P Ultrasound of the post-partum uterus Aust N Z J Obstet Gynaecol 2006 46 04 345 349 10.1111/j.1479-828X.2006.00604.x 16866798 78 Fuller K P Feldman D M Ultrasound evaluation of the postpartum endometrial cavity J Reprod Med 2015 60 (1-2):3 5 25745744 79 Edwards A Ellwood D A Ultrasonographic evaluation of the postpartum uterus Ultrasound Obstet Gynecol 2000 16 07 640 643 10.1046/j.1469-0705.2000.00234.x 11169371 80 Ucci M A Di Mascio D Bellussi F Berghella V Ultrasound evaluation of the uterus in the uncomplicated postpartum period: a systematic review Am J Obstet Gynecol MFM 2021 3 03 100318 10.1016/j.ajogmf.2021.100318 33493704 81 Weissbach T Haikin-Herzberger E Bacci-Hugger K Shechter-Maor G Fejgin M Biron-Shental T Immediate postpartum ultrasound evaluation for suspected retained placental tissue in patients undergoing manual removal of placenta Eur J Obstet Gynecol Reprod Biol 2015 192 37 40 10.1016/j.ejogrb.2015.06.004 26149477 82 Carlan S J Scott W T Pollack R Harris K Appearance of the uterus by ultrasound immediately after placental delivery with pathologic correlation J Clin Ultrasound 1997 25 06 301 308 10.1002/(sici)1097-0096(199707)25:6<301:aid-jcu3>3.0.co;2-g 9142625 83 Kamaya A Ro K Benedetti N J Chang P L Desser T S Imaging and diagnosis of postpartum complications: sonography and other imaging modalities Ultrasound Q 2009 25 03 151 162 10.1097/RUQ.0b013e3181b5451e 19730078 84 Van den Bosch T Van Schoubroeck D Lu C De Brabanter J Van Huffel S Timmerman D Color Doppler and gray-scale ultrasound evaluation of the postpartum uterus Ultrasound Obstet Gynecol 2002 20 06 586 591 10.1046/j.1469-0705.2002.00851.x 12493048 85 Durfee S M Frates M C Luong A Benson C B The sonographic and color Doppler features of retained products of conception J Ultrasound Med 2005 24 09 1181 1186, quiz 1188–1189. Doi: 10.7863/jum.2005.24.9.118116123177 86 De Winter J De Raedemaecker H Muys J Jacquemyn Y The value of postpartum ultrasound for the diagnosis of retained products of conception: A systematic review Facts Views Vis ObGyn 2017 9 04 207 216 30250654 87 Society of Gynecologic Oncology American College of Obstetricians and Gynecologists and the Society for Maternal–Fetal Medicine Cahill A G Beigi R Heine R P Silver R M Wax J R Placenta accreta spectrum Am J Obstet Gynecol 2018 219 06 B2 B16 10.1016/j.ajog.2018.09.042 88 Krapp M Baschat A A Hankeln M Gembruch U Gray scale and color Doppler sonography in the third stage of labor for early detection of failed placental separation Ultrasound Obstet Gynecol 2000 15 02 138 142 10.1046/j.1469-0705.2000.00063.x 10775997 89 Habak P J Kole M Vaginal birth after cesarean delivery Treasure Island StatPearls Publishing 2021[cited 2022 Jan 20]. Available from:https://www.ncbi.nlm.nih.gov/books/NBK507844/ 90 Oba T Hasegawa J Sekizawa A Postpartum ultrasound: postpartum assessment using ultrasonography J Matern Fetal Neonatal Med 2017 30 14 1726 1729 10.1080/14767058.2016.1223034 27578054 91 Bellussi F Cataneo I Dodaro M G Youssef A Salsi G Pilu G The use of ultrasound in the evaluation of postpartum paravaginal hematomas Am J Obstet Gynecol MFM 2019 1 01 82 88 10.1016/j.ajogmf.2019.03.002 33319759 92 Oong G C Eruo F U Vulvar hematoma Treasure Island StatPearls Publishing 2021[cited 2022 Jan 20]. Available from:https://www.ncbi.nlm.nih.gov/books/NBK560753/
PMC009xxxxxx/PMC9800141.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36037813 10.1055/s-0042-1755462 220098 Original Article Endoscopy Misoprostol Administration Before Hysteroscopy Procedures – A Retrospective Analysis O uso do misoprostol prévio aos procedimentos histeroscópicos – Um estudo retrospectivohttp://orcid.org/0000-0003-1327-3377 Inácio Quênya Antunes Silveira 1 http://orcid.org/0000-0002-1819-622X Troncon Júlia Kefalás 1 http://orcid.org/0000-0002-4049-0185 Valério Fernando Passador 1 http://orcid.org/0000-0002-7746-3834 Herren Helmer 1 http://orcid.org/0000-0002-1376-008X Nogueira Antônio Alberto 1 http://orcid.org/0000-0003-0270-5496 Neto Omero Benedicto Poli 1 http://orcid.org/0000-0001-6930-1328 Silva Júlio César Rosa e 1 1 Departamento de Ginecologia e Obstetrícia, Faculdade de Medicina de Ribeirão Preto, Universidade de São Paulo, Ribeirão Preto, SP, Brazil Address for correspondence Júlio César Rosa e Silva, MD, PhD Av. Bandeirantes3900, 14049-900, Ribeirão Preto, [email protected] 29 8 2022 12 2022 1 8 2022 44 12 11021109 19 3 2022 26 3 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  To evaluate the use of misoprostol prior to hysteroscopy procedures regarding technical ease, the presence of side effects, and the occurrence of complications. Methods  This is a retrospective, observational, analytical, case-control study, with the review of medical records of 266 patients followed-up at the Gynecological Videoendoscopy Sector of the Hospital das Clínicas da Faculdade de Medicina de Ribeirão Preto of the Universidade de São Paulo (HCFMRP – USP, in the Portuguese acronym) from 2014 to 2019, comparing 133 patients who used the drug before the procedure with 133 patients who did not. Results  The occurrence of postmenopausal uterine bleeding was the main indication for hysteroscopy and revealed a statistical difference between groups ( p  < 0.001), being present in 93.23% of the patients in the study group and in 69.7% of the patients in the control group . Only 2 patients (1.5%) in the study group reported adverse effects. Although no statistical differences were observed regarding the occurrence of complications during the procedure ( p  = 0.0662), a higher total number of complications was noted in the group that used misoprostol ( n  = 7; 5.26%) compared with the group that did not use the drug ( n  = 1; 0.75%), a fact that is clinically relevant. When evaluating the ease of the technique (measured by the complete performance of all steps of the hysteroscopy procedure), it was verified that although there was no difference between groups ( p  = 0.0586), the control group had more than twice as many incompletely performed procedures ( n  = 17) when compared with the group that used misoprostol previously ( n  = 8), which is also clinically relevant. Conclusion  The use of misoprostol prior to hysteroscopy in our service indicated that the drug can facilitate the performance of the procedure, but not without side effects and presenting higher complication rates. Resumo Objetivo  Avaliação do misoprostol prévio à histeroscopia quanto à facilidade técnica, efeitos colaterais e a ocorrência de complicações durante o procedimento. Métodos  Estudo analítico observacional retrospectivo tipo caso controle com revisão de prontuários de 266 pacientes do Setor de Videoendoscopia Ginecológica do Hospital das Clínicas da Faculdade de Medicina de Ribeirão Preto - Universidade de São Paulo (HCFMRP – USP), de 2014 a 2019, sendo comparadas 133 pacientes que utilizaram o medicamento prévio ao procedimento com 133 pacientes que não o utilizaram. Resultados  Sangramento uterino após a menopausa foi a principal indicação de histeroscopia, apresentando diferença estatística ( p  < 0,001), estando presente em 93,23% das pacientes do grupo de estudo e em 69,17% das pacientes do grupo controle. Apenas 2 pacientes (1,5%) do grupo de estudo relataram efeitos adversos. Não foram observadas diferenças quanto à presença de complicações durante o procedimento ( p  =  0,0662), mas observamos um número total de complicações maior no grupo de estudo ( n  = 7; 5,26%) do que no grupo controle ( n  = 1; 0,75%), o que é clinicamente relevante. Não houve diferença entre os grupos quanto à facilidade técnica ( p  = 0,0586), mas o grupo controle apresentou mais do que o dobro de procedimentos não completamente realizados ( n  = 17) quando comparado com o grupo de estudo ( n  = 8), o que é clinicamente relevante. Conclusão  O uso de misoprostol prévio à histeroscopia no nosso serviço demonstrou que ele pode facilitar a realização do procedimento, mas não é isento de efeitos colaterais e apresenta maiores taxas de complicações. Keywords hysteroscopy misoprostol complications side effects Palavras-chave histeroscopia misoprostol complicações efeitos colaterais ==== Body pmcIntroduction The hysteroscopic procedure emerged ∼ 200 years ago, providing the direct visualization of diffuse or focal uterine abnormalities, the anatomical configuration of the cervical canal and the uterine cavity, path permeability, access for biopsies, and direct removal of lesions. 1 2 3 It is considered a minimally invasive procedure, and its use is quite common both in the diagnosis and in the treatment of several conditions, such as abnormal uterine bleeding, the evaluation of infertile patients, surgeries including myomectomy and polypectomy, the diagnosis of endometrial and endocervix hyperplasia and carcinoma, among others, often being performed in clinics or as outpatient follow-up procedures. 1 2 3 4 5 This method has an important advantage over other diagnostic techniques: the anatomopathological confirmation of lesions visually identified through guided biopsy. 4 In order to carry out the procedure, it is often necessary to dilate the cervix, especially in surgical hysteroscopies in which the equipment for performing the procedure is larger than the endocervical canal. 6 The most frequent causes of unsatisfactory exams are cervical stenosis, pain or patient intolerance, bleeding that hinders hysteroscopic view, and technical difficulties. 7 In an attempt to reduce these technical problems and the number of unsatisfactory exams, several methods of cervical dilation have been developed over the years, such as the use of hydrophilic laminators, bladder catheter balloons, and Hegar dilators. These techniques, however, cause great discomfort to patients and increase the risk of complications during the dilation process. 3 4 5 8 The complication rate varies between 0.3 and 5% according to the definition used, the most common being pain, vagal reaction, uterine perforation, false passage formation, and cervical lacerations. Serious complications, such as organ perforation and pelvic infection, are seldom reported. 7 9 Therefore, there was a need to develop new methods of cervical dilation, which should include cervix preparation for a limited time, acceptable for the patient, with ease of administration, quick action, and providing adequate cervical ripening to facilitate the procedure. Misoprostol is one of the most studied agents in this context. 10 It is a synthetic analog of prostaglandin E1 that has been used for cervical preparation prior to performing hysteroscopy because it promotes effective cervical ripening, as well as being an inexpensive, easy to store and administer, and widely available method. 4 5 The most common adverse effects of misoprostol occur mainly before the procedure and include cramping, abdominal and/or pelvic pain, nausea, changes in intestinal transit, vaginal bleeding of varying intensity, and fever and/or chills. However, these effects are generally described as tolerable and rarely motivate the cancellation or alteration of the procedure. 11 12 According to the scientific literature, misoprostol is effective in cervical ripening in the preoperative period of hysteroscopy, reducing the time needed for cervical dilation and increasing the mean cervical diameter. Nevertheless, the optimal dose, the route of administration, and the ideal time of administration prior to hysteroscopy, in addition to whether the drug reduces the rates of pre- and postmenopausal complications, remain unclear, a fact that justifies the importance of carrying out the present study in daily gynecological practice, whose objectives were to assess the ease of the operative hysteroscopic technique with the use of misoprostol, evaluated by the complete performance of the steps of the procedure, to assess the presence of side effects with the use of the drug, and to analyze the occurrence of hysteroscopic complications with its use. 5 8 Methods The present analytical, observational, case-control study was conducted by reviewing the medical records of patients followed up at the Gynecological Videoendoscopy Sector of the HCFMRP – USP in the period from 2014 to 2019. Using a list provided by the Medical Support Service (SAME, in the Portuguese acronym), a total of 508 patients with prescriptions for the drug misoprostol for intrahospital use at the HCFMRP– USP were identified in the analyzed period. By cross-referencing the data of patients followed-up at the Gynecological Videoendoscopy Sector of the same hospital and who had used misoprostol prior to hysteroscopy, 207 patients were identified. The medical records of these 207 patients were reviewed, evaluating: age, parity, type of delivery, time since menopause, associated diseases (systemic arterial hypertension [SAH], type 2 diabetes mellitus [DM2], obesity, and other comorbidities), use of continuous medication, symptoms, postmenopausal bleeding, presence of endometrial thickening, side effects of the drug, procedure complications, and complete performance of the procedure. Among the total patients, 74 lacked data in their medical records or had their hysteroscopy procedures suspended for various reasons unrelated to the use of misoprostol or to the procedure itself and were therefore excluded. The 133 patients included in the study were compared with another 133 age-matched patients also followed-up at the Gynecological Videoendoscopy Section of the HCFMRP – USP who underwent hysteroscopy procedures but who did not use misoprostol previously (control group), regarding the ease of the technique, considered easy when all steps of the hysteroscopy procedure were carried out completely, the presence of side effects, and the occurrence of complications. Patients who used misoprostol before the hysteroscopy did so by vaginally introducing 2 tablets of 200 micrograms (µg) each the night before the procedure, totaling a single dose of 400 µg. After filling out the study database, which evaluated age, parity, type of delivery, time since menopause, associated diseases (SAH, DM2, obesity, and other comorbidities), the use of continuous medication, symptoms, postmenopausal bleeding, presence of endometrial thickening, drug side effects, procedure complications, and its complete performance, a statistical analysis was carried out using SPSS Statistics for Windows, Version 17.0 (SPSS Inc., Chicago, IL, USA). The distribution of the variables was assessed using the Kolmogorov-Smirnov test. Normally-distributed data were analyzed with the t -test, while abnormally distributed data were evaluated using the Kruskal-Wallis test. Mean values were presented with statistical significance (SS), which was accepted for p  < 0.05. Numerical data were presented as mean ± standard deviation (SD) or median and range, depending on their distribution. The chi-squared test was used for variables expressed as percentages, considering a significance level of p  < 0.05. Since this was a retrospective study involving medical record analysis, the Research Ethics Committee of the HCRP – USP was asked to waive the application of informed consent through a Letter of Exemption from the collection of the consent form, given most of the patients were no longer being followed-up at the hospital. The present study, as well as the waiver of written informed consent, were approved by the Research Ethics Committee (REC) of the HCRP – USP on February 17, 2020, under CAAE Protocol No. 28983920.0.0000.5440. All ethical precepts were followed as recommended by the Helsinki Convention. Results Regarding the clinical characteristics of the studied patients, no significant difference was observed between those who used misoprostol or not prior to the hysteroscopy procedure in relation to age ( p  = 0.9005), the number of pregnancies ( p  = 0.4586), the number of vaginal deliveries ( p  = 0.5531), the time since menopause ( p  = 0.9193), history of previous cesarean delivery ( p  = 0.8723), or regarding the number of prior abortions ( p  = 0.8528) ( Tables 1 and 2 ). When analyzing the comorbidities presented by the patients at the time of hysteroscopy, we observed a significant difference in relation to SAH ( p  = 0.0041), but not regarding DM2 ( p  = 0.0622), obesity ( p  = 0.5082), or other comorbidities ( p  = 0.3510). The fact that the patients used continuous medications for these comorbidities also did not differ significantly between the 2 groups ( p  = 0.3023) ( Table 2 ). Table 1 Clinical characteristics of the studied population With misoprostol Without misoprostol p -value Mean ± SD Mean ± SD Age (years old) 60.08 ± 8.5 59.95 ± 8.15 0.9005 Pregnancy 3.44 ± 1.96 3.24 ± 2.47 0.4586 Vaginal delivery 2.13 ± 2.17 1.96 ± 2.37 0.5531 Time since menopause 10.98 ± 7.48 11.08 ± 8.22 0.9193 Abbreviation: SD, standard deviation. Table 2 Clinical characteristics of the studied population With misoprostol Without misoprostol p -value n (133) % n (133) % Cesarean delivery 0.8723 0 61 45.86 69 51.88 1 33 24.81 27 20.30 2 23 17.29 23 17.29 3 15 11.28 13 9.77 4 1 0.75 1 0.75 Miscarriage 0.8528 0 99 74.44 99 74.44 1 24 18.05 23 17.29 2 8 6.02 6 4.51 3 1 0.75 2 1.50 4 1 0.75 2 1.50 5 0 0.00 1 0.75 SAH 0.0041 Yes 111 83.46 91 68.42 No 22 16.54 42 31.58 DM2 0.0622 Yes 63 47.37 48 36.09 No 70 52.63 85 63.91 Obesity 0.5082 Yes 44 33.08 39 29.32 No 89 66.92 94 70.68 Other comorbidities 0.3510 Yes 89 66.92 96 72.18 No 44 33.08 37 27.82 Continuous use of medications 0.3023 Yes 127 95.49 123 92.48 No 6 4.51 10 7.52 Menopause 1.0000 Yes 123 92.48 123 92.48 No 10 7.52 10 7.52 Symptoms <0.001 Yes 124 93.23 96 72.18 No 9 6.77 37 27.82 Postmenopausal bleeding <0.001 Yes 124 93.23 92 69.17 No 9 6.77 41 30.83 Endometrial thickening 0.6419 Yes 124 93.23 122 91.73 No 9 6.77 11 8.27 Abbreviations: DM2, diabetes mellitus type 2; n , number of samples; SAH, systemic arterial hypertension. Both groups had the same number of patients before and after menopause, with no difference in hormonal status in relation to the use of misoprostol ( p  = 1.000). The presence of symptoms reported by the patients for the indication of the hysteroscopic procedure showed a difference between the 2 groups ( p  < 0.001); 93.23% of the patients in the group that used misoprostol had at least 1 symptom versus 72.18% of the patients in the control group. Among the most reported symptoms was postmenopausal uterine bleeding, which also showed a significant difference between the 2 groups ( p  < 0.001); 93.23% of the patients in the group that used misoprostol had this symptom versus 69.17% of the patients in the control group. Meanwhile, asymptomatic endometrial thickening showed no statistically significant difference between the 2 groups ( p  = 0.6419). Regarding the adverse effects observed in the patients who used the drug prior to hysteroscopy, only 2 patients (1.5%) reported symptoms: both presented tremors, and 1 presented with symptoms of anxiety. Although no significant difference was observed in relation to the occurrence of complications during the procedure ( p  = 0.0662), a higher total number of complications was observed in the group that used misoprostol ( n  = 7; 5.26%) compared with the group that did not ( n  = 1; 0.75%), which is clinically relevant. In the control group, the only reported complication was false passage formation. Meanwhile, in the group that used misoprostol, the most frequent complication was the absence of uterine cavity distension, which was observed in three of the patients. The other observed complications included cervical laceration ( n  = 1), uterine perforation ( n  = 2), and increased fluid absorption ( n  = 1), the latter 2 of which compelled the termination of the procedure. In the group of patients who used misoprostol, false passage formation was not reported ( Table 3 ). Table 3 Complications reported after the hysteroscopy procedure With misoprostol Without misoprostol p -value n (133) % n (133) % Complications 0.0662 Yes 7 5.26 1 0.75 No 126 94.74 132 99.25 Uterine Cervical Laceration 0.9999 Yes 1 0.75 0 0 No 132 99.25 133 100 Absence of uterine cavity distension 0.2619 Yes 3 2.26 0 0 No 130 97.74 133 100 Uterine perforation 0.5079 Yes 2 1.50 0 0 No 131 98.50 133 100 Increased fluid absorption 0.9999 Yes 1 0.75 0 0 No 132 99.25 133 100 False passage 0.9999 Yes 0 0 1 0.75 No 133 100 132 99.25 Abbreviation: n , number of samples. Finally, upon evaluating the technical ease (performing all steps of the procedure), we noted that although there was no difference between groups ( p  = 0.0586), the control group had more than twice as many incompletely performed procedures ( n  = 17) when compared with the group that used misoprostol previously ( n  = 8), which is clinically relevant. Discussion The present study compared 133 women who used misoprostol prior to hysteroscopy with 133 who did not use the medication. This drug is a prostaglandin E1 analog oxytocin that causes changes in the physicochemical structure of cervical collagen. After its administration, misoprostol undergoes de-esterification in the liver into misoprostolic acid. This active metabolite exerts direct action on prostaglandin receptors, leading to the softening and ripening of the cervix, favoring its dilation, in addition to promoting an increase in intracellular calcium, which is responsible for the contraction of uterine muscles. 13 All of these mechanisms enable progressive cervical effacement and dilation. 4 5 The systematic review of misoprostol suggests variations in the plasma concentration of this drug depending on the route of administration. Orally, the drug is rapidly and completely absorbed in the gastrointestinal tract; however, it is also quickly and extensively metabolized into its acidic form in the first hepatic pass (de-esterification). A single 400-µg dose of oral misoprostol reaches its peak concentration in 30 minutes and declines in ∼ 120 minutes, remaining at a low level. After vaginal administration, on the other hand, there is a gradual increase in the plasma concentration of misoprostol, reaching a maximum level after 70 to 80 minutes, followed by a slow decline, with the drug level still detectable after 6 hours. It has also been reported that the mean concentration peak via the sublingual route is higher than that achieved via the oral and vaginal routes, which is due to the rapid absorption by the sublingual mucosa, avoiding first-pass metabolism in the liver. When administered rectally, the absorption curve of the drug is similar to that seen when using the vaginal route. 14 Corroborating another study carried out with 77 women between January 2005 and March 2006, we did not observe a significant difference in age and the number or type of previous births between the study group and the control group. 15 Although no significant difference was found in the present study regarding pre- and postmenopausal patients, some studies suggest that the use of misoprostol is more effective in dilating the cervix of premenopausal patients, mainly due to the hormonal difference between these women. 4 Regarding the comorbidities presented by the patients at the time of the procedure, only SAH showed significance. In our study, postmenopausal uterine bleeding was the most prevalent symptom, observed in 93.23% of the patients in the study group and in 69.17% of those in the control group. A case review coordinated by Gimpelson et al. 16 revealed a high incidence of abnormal bleeding, which was observed in 76% of the cases as the chief complaint of patients to undergo hysteroscopy. In addition, a descriptive, cross-sectional study with 26 women showed that the primary complaint of 65.3% of the patients was uterine bleeding. 17 The occurrence of adverse effects to the use of misoprostol was reported by only 2 patients (1.5%), who presented tremors and anxiety. A meta-analysis evaluating 14 studies showed that significantly more adverse effects were reported when misoprostol was administered compared with procedures without previous use of the drug (odds ratio [OR] = 3.56; 95% confidence interval [CI]: 1.60–7.93). 18 Two other studies described the incidence of adverse effects among women randomized to 200 or 400 μg of misoprostol, but their data were conflicting: the first did not demonstrate a dose-related increase in adverse effects (nausea and abdominal pain) in women randomized to 200 or 400 μg ( p  = 1.0 and p  = 0.055, respectively); however, the second study showed a significant increase in the number of adverse events related only to 400 μg ( p  = 0.015), such as fever, abdominal pain, diarrhea, nausea, vomiting, and vaginal bleeding. 18 Regarding the administration time, these studies showed that there was no significant difference in the incidence of abdominal cramps ( p  = 0.64), nausea ( p  = 0.79), diarrhea ( p  > 0.99), genital tract bleeding ( p  = 0.62), and fever ( p  > 0.99) among women who received misoprostol 12 hours or 3 hours before hysteroscopy. 18 One study with 160 women comparing oral, sublingual, and vaginal administration of the drug concluded that the 3 groups were comparable, and all the adverse effects were similar in all groups and were tolerable. This result is in line with a recent meta-analysis that analyzed 7 randomized, controlled studies involving 568 individuals, evaluating the use of misoprostol in surgical hysteroscopy. Compared with the placebo group, there was an increase in side effects (cramps, vaginal bleeding, nausea, and diarrhea) in the misoprostol group (relative risk [95%CI]: 4.28 [1.43­–12.85]). 19 20 The incidence of complications in our study was low compared with the mean described in the literature. 9 A total of 7 complications occurred in patients who had previously used misoprostol (5.26%) and in 1 patient who did not use the medication (0.75%). The primary complication reported among patients who used misoprostol was the absence of uterine cavity distension, which occurred in 3 patients. This was also found by Batukan et al. 15 in a randomized, double-blind, placebo-controlled study carried out with the objective of evaluating the efficacy of 400 µg of misoprostol 10 to 12 hours before surgical hysteroscopy in premenopausal women, via the vaginal route. One of the reported disadvantages of vaginal administration was excessive cervical dilation, resulting in difficulty distending the uterine cavity due to fluid leakage through the cervical canal. 8 Another study, prospectively conducted between January 2005 and March 2006 at the Department of Obstetrics and Gynecology of the Faculty of Medicine of Erciyes University, with 77 women, showed that vaginal administration of misoprostol (400 μg) prior to operative hysteroscopy in premenopausal women was superior to the same dose of orally administered misoprostol. The complication rates during cervical dilation, as well as drug side effects, were comparable between the two regimens. Fluid leakage caused by excessive cervical dilation and effacement appeared to be the most important potential disadvantage of vaginal misoprostol. Since the intrauterine pressure did not reach the ideal desired level in these cases, the uterine distension was suboptimal and, therefore, the procedure was more difficult. 15 In the present study, the other complications observed in patients who used misoprostol were cervical laceration ( n  = 1; 0.75%), uterine perforation ( n  = 2; 1.5%), and increased absorption of distension media ( n  = 1; 0.75%). The patient who did not use misoprostol presented false passage formation as a complication. In a study carried out at Bringham and Women's Hospital, located in Boston, MA, USA, Propst et al. 21 verified a total number of 925 surgical hysteroscopies between 1995 and 1996, with the occurrence of complications in 2.7% of the patients. Among these complications was cervical laceration, which was present in all cases. In another study carried out by Jansen et al., 22 in which 13,600 hysteroscopies in 82 hospitals in the United States were evaluated in 1997, as well as in the study by Agostini et al., 23 performed in Marseille, France, in which 2,116 surgical hysteroscopies between 1990 and 1999 were analyzed, the most frequent complication among the surgical procedures was uterine perforation, with 0.76% and 1.61% of cases, respectively. 22 23 The study by Propst et al. 21 also reported complications related to uterine distension media in 1.4% of the surgical hysteroscopy cases. In a randomized, controlled, double-blind study conducted by Oppengard et al., 24 in which each participant received 1,000 µg of misoprostol or placebo, which was self-inserted vaginally at least 12 hours before operative hysteroscopy, there were a total of 9 (11%) complications reported. Regarding technical ease, which was evaluated by the complete performance of all steps of the procedure, we observed that although there was no statistically significant difference, the number of patients who did not use misoprostol and did not undergo the complete procedure was more than double in relation to the patients who used the medication (8 patients who used the drug versus 17 patients who did not), which is of substantial clinical relevance. Fernandez et al. 25 reported in a larger series that the administration of 400 µg of oral misoprostol 12 or 24 hours before surgery or 200 µg of vaginal misoprostol 9 to 10 hours before surgery, respectively, demonstrated greater ease of cervical dilation. However, to date, no placebo-controlled trials have shown a significant decrease in the rate of serious complications such as cervical laceration or perforation. 25 The prospective study conducted between January 2005 and March 2006 at the Department of Obstetrics and Gynecology of the Faculty of Medicine of Erciyes University showed that vaginally administered misoprostol (400 µg) prior to operative hysteroscopy in premenopausal women was superior to the same dose of orally administered misoprostol in terms of shorter cervical dilation and surgery duration, as well as the need for cervical dilation for No. 9 Hegar. 15 Other previous studies comparing patients who used misoprostol or placebo previously to prepare the cervix, as evidenced by Uckuyu et al. 26 in Ankara, Turkey, also show a relevant rate of failure in dilation with Hegar dilators, especially in patients who had had previous cesarean sections. In the aforementioned study, the reported failure rate of cervical dilation using Hegar dilators was 25%. 26 Our study had some limitations, such as the absence of a patient-reported pain assessment, given most of the analyzed patients underwent the procedure in the operating room under anesthesia; lack of evaluation of different doses and routes of administration of misoprostol, since all of the patients used the drug at a dose of 400 µg via the vaginal route; the absence of evaluation of the time of misoprostol administration, seeing that, in all patients, the drug was introduced the night before the procedure. Costa et al. 27 conducted a randomized study with 120 postmenopausal women who received 200 µg of vaginal misoprostol or placebo 8 hours before outpatient hysteroscopy. There was a significant reduction in the pain scale during the procedure, a fact that would facilitate the performance of outpatient surgical procedures. In 2018, Fouda et al. 28 evaluated the effect of timing of vaginal misoprostol administration (3 hours versus 12 hours) before diagnostic hysteroscopy in nulliparous women, who are at increased risk for cervical canal stenosis. The group of women given the drug just 3 hours before hysteroscopy reported more pain during the procedure than those given the medication 12 hours earlier. However, the pain intensity 30 minutes after the procedure, its mean duration, and the occurrence of side effects to misoprostol were similar between the 2 groups. The passing of the hysteroscope through the cervical canal was also assessed by the examiners and was found to be easier in the 12-hour group. Conclusion The use of misoprostol prior to hysteroscopy in our service showed that the drug can facilitate the performance of the procedure; however, this drug is not free from side-effects and higher complication rates. Also, misoprostol is a well-tolerated drug. We agree with most authors that there is a need for further studies to better identify the ideal dose, route of administration, and time to indicate misoprostol before the procedure. Acknowledgments The present study was carried out with the support of de Coordination for the Improvement of Higher Education Personnel - Brazil (CAPES, in the Portuguese acronym). Contributions Conflict of Interests The authors have no conflict of interests to declare. All authors contributed to the design of the study, were involved in the data collection, data analysis and/or interpretation. Also, all authors contributed to the writing/substantive editing and review of the manuscript and approved the final draft of the manuscript. ==== Refs References 1 Martins F N Martins N N Histeroscopia diagnóstica [Internet] Coimbra Federação das Sociedades Portuguesas de Obstetrícia e Ginecologia 2010[cited 2021 Dec 12]. p. 359-71. Available from:http://www.fspog.com/fotos/editor2/cap_42.pdf 2 Ganer Herman H Kerner R Gluck O Feit H Keidar R Bar J Different routes of misoprostol for same-day cervical priming prior to operative hysteroscopy: a randomized blind trial J Minim Invasive Gynecol 2017 24 03 455 460 10.1016/j.jmig.2016.12.024 28069480 3 Valente E P Uso de misoprostol vaginal para redução da dor em histeroscopia diagnóstica na menacme: ensaio clínico randomizado, triplamente mascarado, controlado com placebo [dissertação] Recife Instituto Materno Infantil Professor Fernando Figueira 2017 4 Bastu E Celik C Nehir A Dogan M Yuksel B Ergun B Cervical priming before diagnostic operative hysteroscopy in infertile women: a randomized, double-blind, controlled comparison of 2 vaginal misoprostol doses Int Surg 2013 98 02 140 144 10.9738/INTSURG-D-12-00024.1 23701149 5 Zhuo Z Yu H Jiang X A systematic review and meta-analysis of randomized controlled trials on the effectiveness of cervical ripening with misoprostol administration before hysteroscopy Int J Gynaecol Obstet 2016 132 03 272 277 10.1016/j.ijgo.2015.07.039 26797202 6 Fradique A Histeroscopia cirúrgica [Internet] Coimbra Federação das Sociedades Portuguesas de Obstetrícia e Ginecologia 2010[cited 2021 Dec 12]. p. 511-33. Available from:http://www.fspog.com/fotos/editor2/cap_49.pdf 7 Clark T J Voit D Gupta J K Hyde C Song F Khan K S Accuracy of hysteroscopy in the diagnosis of endometrial cancer and hyperplasia: a systematic quantitative review JAMA 2002 288 13 1610 1621 10.1001/jama.288.13.1610 12350192 8 Allen R O'Brien B M Uses of misoprostol in obstetrics and gynecology Rev Obstet Gynecol 2009 2 03 159 168 19826573 9 Costa A A Estudo randomizado, duplamente mascarado, placebo controlado do uso do misoprostol versus placebo para histeroscopia diagnóstica em mulheres na pós-menopausa [tese] Campinas Universidade Estadual de Campinas 2006 10 Moreira A R O uso de misoprostol em histeroscopia: revisão do seu papel em contexto diagnóstico e cirúrgico [monografia] Porto Universidade do Porto 2018 11 de Albuquerque L G Hardy E Bahamondes L Histerossonografia: avaliação da cavidade uterina com sangramento anormal Rev Assoc Med Bras (1992) 2006 52 04 247 250 10.1590/S0104-42302006000400025 16967143 12 Angioni S Loddo A Milano F Piras B Minerba L Melis G B Detection of benign intracavitary lesions in postmenopausal women with abnormal uterine bleeding: a prospective comparative study on outpatient hysteroscopy and blind biopsy J Minim Invasive Gynecol 2008 15 01 87 91 10.1016/j.jmig.2007.10.014 18262151 13 Koch D M Rattmann Y D Uso do misoprostol no tratamento da hemorragia pós-parto: uma abordagem farmacoepidemiológica Einstein (Sao Paulo) 2020 18 eAO5029 10.31744/einstein_journal/2020AO5029 31721897 14 Souza A S Amorim M M Costa A A Noronha Neto C Farmacocinética e farmacodinâmica do misoprostol em obstetrícia Femina 2019 37 12 679 684 15 Batukan C Ozgun M T Ozcelik B Aygen E Sahin Y Turkyilmaz C Cervical ripening before operative hysteroscopy in premenopausal women: a randomized, double-blind, placebo-controlled comparison of vaginal and oral misoprostol Fertil Steril 2008 89 04 966 973 10.1016/j.fertnstert.2007.03.099 17681307 16 Gimpelson R J Rappold H O A comparative study between panoramic hysteroscopy with directed biopsies and dilatation and curettage. A review of 276 cases Am J Obstet Gynecol 1988 158 (3 Pt 1):489 492 10.1016/0002-9378(88)90011-7 3348309 17 Bezerra Júnior A L Costa A A Fernandes G C Santos J V Pereira T M Resultados intra e pós-operatórios e a frequência de complicações em pacientes pré e pós-menopausadas submetidas à histeroscopia cirúrgica no IMIP [tcc] Recife Instituto Materno Infantil de Pernambuco 2019 18 De Silva P M Wilson L Carnegy A Smith P P Clark T J Cervical dilatation and preparation prior to outpatient hysteroscopy: a systematic review and meta-analysis BJOG 2021 128 07 1112 1123 10.1111/1471-0528.16604 33219606 19 Selk A Kroft J Misoprostol in operative hysteroscopy: a systematic review and meta-analysis Obstet Gynecol 2011 118 04 941 949 10.1097/AOG.0b013e31822f3c7b 21934459 20 Song T Kim M K Kim M L Jung Y W Yoon B S Seong S J Effectiveness of different routes of misoprostol administration before operative hysteroscopy: a randomized, controlled trial Fertil Steril 2014 102 02 519 524 10.1016/j.fertnstert.2014.04.040 24856464 21 Propst A M Liberman R F Harlow B L Ginsburg E S Complications of hysteroscopic surgery: predicting patients at risk Obstet Gynecol 2000 96 04 517 520 10.1016/s0029-7844(00)00958-3 11004351 22 Jansen F W Vredevoogd C B van Ulzen K Hermans J Trimbos J B Trimbos-Kemper T C Complications of hysteroscopy: a prospective, multicenter study Obstet Gynecol 2000 96 02 266 270 10.1016/s0029-7844(00)00865-6 10908775 23 Agostini A Bretelle F Cravello L Ronda I Roger V Blanc B [Complications of operative hysteroscopy] Presse Med 2003 32 18 826 829 12870384 24 Oppengard K S Nesheim B I Istre O Qvigstad E Comparison of self-administered vaginal misoprostol versus placebo for cervical ripening prior to operative hysteroscopy using a sequential trial design BJOG 2008 115 05 663,e 1–910.1111/j.1471-0528.2007.01628.x 18201279 25 Fernandez H Alby J D Tournoux C Chauveaud-Lambling A DeTayrac R Frydman R Vaginal misoprostol for cervical ripening before operative hysteroscopy in pre-menopausal women: a double-blind, placebo-controlled trial with three dose regimens Hum Reprod 2004 19 07 1618 1621 10.1093/humrep/deh302 15155607 26 Uckuyu A Ozcimen E E Sevinc F C Zeyneloglu H B Efficacy of vaginal misoprostol before hysteroscopy for cervical priming in patients who have undergone cesarean section and no vaginal deliveries J Minim Invasive Gynecol 2008 15 04 472 475 10.1016/j.jmig.2008.03.001 18588848 27 Costa HdeL Costa L O [Hysteroscopy in menopause: analysis of the techniques and accuracy of the method] Rev Bras Ginecol Obstet 2008 30 10 524 530 10.1590/s0100-72032008001000008 19082390 28 Fouda U M Elsetohy K A Elshaer H S Hammad B EM Shaban M Youssef M A Misoprostol prior to diagnostic office hysteroscopy in the subgroup of patients with no risk factors for cervical stenosis: a randomized double-blind placebo-controlled trial Gynecol Obstet Invest 2018 83 05 455 460 10.1159/000480234 28982101
PMC009xxxxxx/PMC9800142.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36580943 10.1055/s-0042-1760216 RBGO-22-0283 Letter to the Editor | Carta ao Editor Simulation, A Fundamental Component of Training to Treat Placenta Accreta Spectrum http://orcid.org/0000-0001-5639-9127 Nieto-Calvache Albaro José 1 http://orcid.org/0000-0002-6327-6149 Aguilera Lorgio Rudy 2 1 Clínica de Acretismo Placentario, Hospital Universitario Fundación Valle del Lili, Cali, Colombia 2 Departamento de Ginocología y Obstetricia, Hospital De La Mujer Dr. Percy Boland, Santa Cruz de la Sierra, Bolivia Address for correspondence Albaro José Nieto-Calvache, MD Hospital Universitario Fundación Valle del LiliCra 98 # 18-49, Cali, Valle del [email protected] 29 12 2022 12 2022 1 12 2022 44 12 11591160 15 9 2022 26 9 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. ==== Body pmcDear Editor, We thank Professor Chikazawa et al. 1 for their interest in our paper 2 and for highlighting the importance of simulation during training for the management of placenta accreta spectrum (PAS). There are multiple options to manage PAS and although the disease exhibits a wide variety of clinical presentations (spectrum), most groups choose a single therapeutic alternative and apply it to all their patients, making it difficult to respond when deviations from the original plan arise. Few publications propose a clear sequence of interventions applicable to all types of PAS. Our group uses the protocolized approach described by Palacios-Jaraquemada et al. 3 applicable to patients with suspected prenatal PAS, but also to those diagnosed intraoperatively, considering the nature (predominantly hypervascularization or presence of vesicouterine fibrosis) and the topography of the lesion (which uterine wall is affected, and which is the relationship of the lesion with the vesicouterine peritoneal fold). 3 4 This protocol includes four steps ( Fig. 1 ). Fig. 1 Protocolized approach to PAS. Abbreviations: MSTH, Modified subtotal hysterectomy; OSCS, One-step conservative surgery; PAS, Placenta accreta spectrum. *If the clinical condition of the patient or her fetus does not allow to defer the procedure, avoid manipulating the placenta. First, the evaluation of the available resources and the clinical situation of the patient (to define whether or not to go ahead with the surgery). Doctor Chikazawa et al. 1 rightly point out that the process of training to manage PAS is a long one, and that obstetricians without such training are likely to be faced with the intraoperative finding of PAS. As useful as training in what to do, it is necessary to be very clear about what to avoid in the event of a PAS intraoperative finding, without the appropriate resources (human or technological), the greatest success of the obstetrician would be to avoid a high number of interventions when the clinical situation of the patient allows it. Second, intraoperative staging through 4 actions: opening of the parametrium (to evaluate the lateral uterine wall), digital evaluation of the retrovesical space (Pelosi maneuver), dissection of the retrovesical space by ligating the vesicouterine pedicles (to evaluate the anterior uterine wall), and exteriorization of the uterus to evaluate the posterior uterine wall. Third, the recommended treatment will be chosen (one step conservative surgery, total hysterectomy or modified subtotal hysterectomy) based on the topographic classification, 3 4 and after answering the three following questions: Is it possible to separate the bladder from the uterus? Is there > 2 cm of healthy myometrium cephalic to the cervix and caudal to the PAS area? Does > 50% of the circumference of the uterus (in an axial section at the level of the PAS area) has healthy myometrium? ( Fig. 1 ). Fourth and last, it is essential to have photographic and video recording elements of the surgical procedures to later on debrief, self-assess, and provide research activities that facilitate learning and continuous improvement of the performance of the group. A standardized approach facilitates the construction of a mental map that obstetricians can internalize or consult immediately, facilitating decision-making in the face of a planned or unexpected PAS case. Conflict of Interests The authors have no conflict of interests to declare. ==== Refs References 1 Chikazawa K Matsubara S Kuwata T Difficulties in the management of placenta accreta spectrum disorders are not confined to low-/middle-income countries: a possible usefulness of simulation training Rev Bras Ginecol Obstet 2022 44 08 804 805 10.1055/s-0042-1751073 36075226 2 Aguilera L R Mojica-Palacios L M Urquizu F Gorena M Guzmán F T Galliadi L MV Difficulties in the management of placenta accreta spectrum in hospitals with limited resources Rev Bras Ginecol Obstet 2022 44 05 467 474 10.1055/s-0042-1742408 35472821 3 Palacios-Jaraquemada J M Fiorillo A Hamer J Martínez M Bruno C Placenta accreta spectrum: a hysterectomy can be prevented in almost 80% of cases using a resective-reconstructive technique J Matern Fetal Neonatal Med 2022 35 02 275 282 10.1080/14767058.2020.1716715 31984808 4 Nieto-Calvache A J Palacios-Jaraquemada J M Aryananda R A Rodriguez F Ordoñez C A Bryon A M How to identify patients who require aortic vascular control in placenta accreta spectrum disorders? Am J Obstet Gynecol MFM 2022 4 01 100498 10.1016/j.ajogmf.2021.100498 34610485
PMC009xxxxxx/PMC9800143.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36580939 10.1055/s-0042-1759635 220305 Expert Recommendations Expert Recommendation /Monkeypox Expert Recommendations on Monkeypox (MPX) in Pregnancy, Postpartum and Lactating Women http://orcid.org/0000-0003-1405-5371 Mattar Rosiane 1 http://orcid.org/0000-0002-2942-6182 Neto Antonio Rodrigues Braga 23 http://orcid.org/0000-0001-9863-9993 Luz Adriana Gomes 4 http://orcid.org/0000-0001-7306-533X Hatanaka Alan 1 http://orcid.org/0000-0001-6440-4395 Zaconeta Alberto 5 http://orcid.org/0000-0002-6985-0818 Guazzelli Cristina Aparecida Falbo 1 http://orcid.org/0000-0002-6856-8327 Traina Evelyn 1 http://orcid.org/0000-0002-5251-2556 Baptista Fernanda Spadotto 6 http://orcid.org/0000-0001-9130-8745 Osanan Gabriel 7 http://orcid.org/0000-0002-1689-6142 Duarte Geraldo 8 http://orcid.org/0000-0002-3789-885X Ramos Jose Geraldo Lopes 9 http://orcid.org/0000-0002-6578-995X Oppermann Maria Lucia 9 http://orcid.org/0000-0002-9981-8069 Francisco Rossana Pulcineli Vieira 6 http://orcid.org/0000-0003-1098-3269 Cardoso Sigrid Maria Loureiro de Queiroz 10 http://orcid.org/0000-0002-9311-786X Quintana Silvana Maria 8 http://orcid.org/0000-0001-6573-7386 Sun Sue Yazaki 1 http://orcid.org/0000-0002-9347-9641 Borges Vera Therezinha Medeiros 11 1 Escola Paulista de Medicina, Universidade Federal de São Paulo, São Paulo, SP, Brazil 2 Universidade Federal do Rio de Janeiro, Rio de Janeiro, RJ, Brazil 3 Universidade Federal Fluminense, Niterói, RJ, Brazil 4 Universidade Estadual de Campinas, Campinas, SP, Brazil 5 Universidade de Brasília, Brasília, DF, Brazil 6 Faculdade de Medicina, Universidade de São Paulo, São Paulo, SP, Brazil 7 Universidade Federal de Minas Gerais, Belo Horizonte, MG, Brazil 8 Faculdade de Medicina, Universidade de São Paulo, Ribeirão Preto, SP, Brazil 9 Universidade Federal do Rio Grande do Sul, Porto Alegre, RS, Brazil 10 Universidade Federal do Amazonas, Manaus, AM, Brazil 11 Faculdade de Medicina, Universidade Estadual Paulista, Botucatu, SP, Brazil Address for correspondence Rosiane Mattar Escola Paulista de Medicina, Universidade Federal de São PauloRua Botucatu, 740, 04023-062, Vila Clementino, São Paulo, [email protected] 29 12 2022 12 2022 1 12 2022 44 12 11221125 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. ==== Body pmcThe Monkeypox Disease In 2020, Brazil and the whole world faced the COVID-19 pandemic, which caused a high number of deaths. This disease was particularly severe for pregnant and postpartum women and determined a significant increase in the Maternal Death Ratio (MMR). To face the disease and assist health professionals in the qualification of the best care to the maternal-fetal binomial, the Ministry of Health and Febrasgo developed a working group formed by professors and researchers from several universities who worked to establish recommendations for the care of pregnant women and puerperal women by the time of the COVID-19 pandemic. In 2022, while we are still experiencing the COVID-19 pandemic, we are surprised by another disease caused by a virus that has been alarming the population and worrying public health authorities and gynecology and obstetrics societies in Brazil and worldwide. It is the infection that is caused by monkeypox virus (MPXV), which is still a not well-known disease, with many of its characteristics not well determined. The knowledge of this disease is fundamental for health professionals working in Obstetrics to plan forms of prevention, as well as the establishment of the diagnosis and treatment of the monkeypox (MPX) disease, preserving the health of the maternal-perinatal binomial. For this reason, the Brazilian Ministry of Health requested the same working group that acted diligently by the time of COVID-19 to establish recommendations for facing MPX, to provide adequate care for pregnant women and puerperal women. These recommendations, based on the knowledge that exists so far, are what guide these orientations and may change depending on new findings that may be presented over time. The MPXV was named after being identified in laboratory monkeys in 1958. The first case of this virus in humans was recorded in 1970 in a child in Congo and since then has become an endemic disease in West and Central Africa. 1 In 2003, the first cases were registered outside the African continent, in the United States, 2 3 4 but that was contained through hygienic measures and stock vaccines. In 2017, there was a major outbreak started in Nigeria and spread to some African countries. In early May 2022, another outbreak of MPXV was identified, this time in several countries outside the African continent, with fast dissemination of cases. As a result, on May 21, 2022, the World Health Organization (WHO) declared the existence of an emerging global outbreak of MPXV infection, and on July 23 has determined that this outbreak constituted a Public Health emergency of international concern. Pregnant women present clinically with similar characteristics to nonpregnant women, but may evolve with greater severity, being therefore considered a risk group. In addition to maternal clinical repercussions, there are also concerns specific to the pregnancy period, such as fetal vitality, the possibility of vertical transmission and perinatal outcome. It has been verified that MPXV infection can lead to adverse results in pregnancy, such as fetal death and spontaneous abortion. 5 6 A recent publication on the evolution of pregnancy in 4 MPXV-infected women showed spontaneous 1 st trimester abortion in 2 pregnant women, without testing of the conception products; an intrauterine death in the 2 nd trimester, with clinical, histological and laboratory evidence of intrauterine fetal infection evidencing the very probable vertical transmission of the disease, and a pregnant woman with MPXV infection that evolved with full-term delivery of healthy conceptus. 7 Close and prolonged skin-to-skin contact, including during sexual activity, seems to be the main means of transmission of MPXV. There are suspicions of transmission of this virus by droplets and aerosols. There is also transmission through biting of rodent animals or even the ingestion of those animals. In addition, contagion by phositis, especially used clothing, can transmit the disease. The quick identification and isolation of affected individuals is fundamental to prevent the spreading of the disease. 8 Transmission of MPXV occurs in the phase of active skin lesions and only ends when they heal completely, which usually requires isolation of 21 to 28 days. 7 9 10 There are doubts as to whether the contagion could be prior to the phase of skin lesions, since viral DNA has already been identified in the blood and respiratory system of patients prior to the lesions. 5 7 Sexual transmission has been discussed not only by contact, but also because the virus has been identified in seminal material. 5 11 Patients with MPXV should be isolated in a separate area of their home or in hospital services, especially if they present extensive lesions and/or respiratory symptoms. Skin lesions should be covered (for example, with the use of long sleeves and trousers) to minimize the risk of contact. Everyone should wear a face mask in the presence of an infected person. Sexual abstinence is also recommended in the phase of unhealed lesions and condom use for any form of sexual act (anal, oral, or vaginal) in the 12 weeks following the healing of the lesions. 7 Most patients with MPXV will have mild disease and can be cared for at home, where they should remain isolated. Standard cleaning and disinfection procedures should be performed, taking care of clothes and used objects. The diagnosis of infection can be made by anamnesis and clinical findings, with epidemiological suspicion. The incubation period is, on average, 6 to 13 days, and can be from 5 to 21 days. Next to this, a prodromal period occurs, when fever, sweating, headache, myalgia, fatigue and lymphadenomegaly, which is quite characteristic of the disease, are manifested. About 1 to 3 days later, the rash, which usually affects the face, genitals, and extremities, and has a centrifugal character, appears. The lesion evolves from macules to papules, vesicles, pustules and, later, crusts. 3 In general, they are well circumscribed and deep, and develop umbilication. They can also be uniform or at different stages of evolution. Lesions are often painful until healing (which happens normally when they are itching). The severity criteria consider the number of lesions: Mild (< 25 skin lesions), Moderate (25–99), Severe (100–250) and very severe (> 250). Another severity criterium is when one of these signs is present: fever for > 7 days, cervical lymph node, persistent vomiting, dehydration, retrocular pain, respiratory failure, mental confusion, hepatomegaly, and sepsis. Confirmatory laboratory diagnosis of MPXV can be performed by real-time polymerase chain reaction (qPCR). 12 Sample collection should preferably be performed with swab of ulcerated lesions. If there are only vesicles, the sample can be obtained by means of fine needle puncture, with extreme caution to avoid accidents that allow contagion. If there are only crusts, the material can be obtained by swab or by collection of small fragments. These samples should be stored in dry sterile vials without any preserving liquid. In individuals in whom MPXV is suspected, without any clinical suspicion, oral swab should be tested and vaginal or anal swab may be considered. 13 Blood samples are oriented for differential diagnosis and/or concomitance of other diseases that could cause lesions, such as syphilis, acquired immunodeficiency syndrome, and herpetic infection. 12 13 Although the disease is most of the time self-limited and with spontaneous cure, in some cases there may be a need for specific drug treatment. Most of the time, there is only indication of symptomatic treatment for fever and pain with dipyrone, acetaminophen, or even opiate derivatives in the most severe pain conditions. In cases with more important lesions, the use of antibiotics may be indicated to prevent secondary bacterial infection: systemic amoxicillin and ocular chloramphenicol. There are some patients who have worsening of the condition, and in these circumstances, antivirals are indicated. Antiviral drugs: Tecovirimat (TPOXX), cidofovir (Vistide) and brincidofovir (Tembexa) have been considered. About 5% of patients with MPXV require antivirals. 11 14 There are still no well-established protocols for their use in pregnancy. Immunoglobulin (VIG), a mixture of purified blood antibodies of individuals immunized with the smallpox vaccine, has already been used for the prevention/treatment of MPXV. There is no evidence on its effectiveness; however, it has already been considered as prophylaxis in exposed individuals with severe immunodeficiency as a prevention/treatment of MPXV. There are two vaccines developed to fight human smallpox, which are capable of inducing protective antibodies against MPXV. However, they are not available in Brazil and there are not enough doses for mass vaccination. The first is ACAM2000, with live vaccinia attenuated but replicating virus, applied in a single dose and with immune response 4 weeks after application. Because it is a live virus vaccine, it is contraindicated for individuals with immunodeficiency and pregnant women. The second vaccine is Modified Ankara (MVA-BN), produced by Bavarian Nordic, with live, attenuated, nonreplicating virus 15 is sold in Europe as Imvamune or Imvanex and in the United States as Jynneos. It has efficacy of 85%. 16 It can be used in immunosuppressed patients. It is applied in two doses with an interval of 4 weeks, and its protection begins 2 weeks after the second dose. Monkeypox in Pregnant, Postpartum, and Lactating Women There are few reports on MPXV during pregnancy. 6 It is known that the virus can cross the placenta and reach the fetus. Thus, as in other viral infections, it may increase the risk of abortion, fetal death, prematurity, and other fetal complications. There is still no way to quantify these risks. Therefore, the care with the pregnant woman and the fetus should be intense in the face of suspicion or confirmation of the infection. The WHO recognizes maternal-fetal transmission through the transplacental passage, originating the congenital disease, and/or transmission through intimate contact, during and after delivery. In postexposure asymptomatic pregnant women, if MPXV is undetectable, monitoring can be suspended. If MPXV is detectable, home isolation should be maintained for a minimum of 21 days. Self-monitoring of body temperature and skin lesions and teleservice monitoring by the health professionals should be orientated. In pregnant women with signs or symptoms of MPXV but with negative qPCR, isolation and self-monitoring of temperature and skin lesions should be indicated. Other potential causes should be ruled out, and retesting the patient is indicated if symptoms persist. If the MPXV test is positive, hospitalization is indicated in moderate and severe cases. There are still insufficient data on the use of vaccines in pregnant or lactating women and none of the vaccines are approved in pregnancy. Animal studies did not find adverse fetal effects, and a study with 300 pregnant women did not show an increase in adverse outcomes. 17 It is still unknown whether vaccines are excreted in breast milk. Vaccines with replicant viruses are contraindicated in pregnancy and infants. 18 19 20 The MVA-BN has been considered safe during breastfeeding. 1 Thus, any woman who is breastfeeding, with substantial exposure to the virus, should be vaccinated after considering the risks of MPXV infection for her and her child. It will usually only be necessary to use symptomatic for the treatment of MPVX during pregnancy. The use of antivirals is not approved. There are no studies of the antiviral drugs in humans. In animals, Tecovirimat did not induce teratogenicity, and Cidofovir and Brincidofovir were classified as FDA class C because they caused changes in the morphology of the animals undergoing the study. Although little is known about VIG during pregnancy, other immunoglobulins have already been used in pregnancy and have been shown to be safe. Until now, this type of treatment has not been indicated during pregnancy. 18 In the presence of acute infection, obstetric ultrasound (US) is recommended during the 1 st trimester to evaluate the viability of pregnancy. In moderate and severe cases during the 2 nd trimester, obstetric US is suggested to evaluate biometrics and fetal morphology as well as to quantify the amniotic fluid index. During the 3 rd trimester, when available, fetal biophysical profile and fetal Doppler flowmetry should be done to assess the well-being of the conceptus. After the 26 th week, cardiotocography is recommended in cases of infection considered moderate and severe. 19 After the maternal cure, fetal risks are low, but obstetric US is recommended every 4 weeks for evaluation of fetal growth and well-being. 10 Normally, there is no indication to anticipate delivery. 7 In severe cases or fetal impairment, we should consider delivery, evaluating gestational age and fetal weight. In cases in which preterm delivery is indicated, magnesium sulfate and corticosteroids should be used, according to obstetric indication. During the delivery, the presence of a healthy companion with the use of personal protective equipment should be ensured, which should be maintained throughout the hospital stay. The delivery should follow obstetrics indications, and there is no reason to indicate cesarean section because of the infection. If the patient presents genital lesions, because of a higher risk of neonatal infection during delivery, c-section will be indicated. 9 Timely clamping of the umbilical cord is recommended, although skin-to-skin contact between mother and newborn (NB) should be avoided. Immediate macroscopic examination of the NB should be taken and, when available, a swab of throat and any skin lesions. 9 It is recommended that the newborn be sanitized by bathing immediately after delivery. It is up to the doctor to inform the risks of infection and need to keep mother and child in separate rooms during the isolation phase. If this is not possible, strict precautions should be followed during contact: the NB should be fully clothed or wrapped in sheets, just as the mother should wear gloves and surgical mask well-adjusted to the face. Direct breastfeeding should be postponed, but support should be offered for the woman to maintain milk production and allow relactation later. Milked breast milk should be discarded. Precautions should be maintained until isolation criteria are met. If the NB is tested positive, the isolation can be cleared. Action of antiviral drugs and vaccine immunoglobulin are little known in milk production. 20 The discharge should be adjusted considering the isolation time and the ability to follow to the recommendations to avoid the contagion of the NB. Recommendations for Pregnant, Postpartum, and Lactating Women Use of masks, especially in environments with individuals potentially infected with the virus. Stay away from people who have suspected symptoms such as fever and mucosal skin lesions. Use condoms in all types of sexual intercourse (oral, vaginal, anal) since transmission through intimate contact has been the most frequent. Be alert if your sexual partnership presents any lesion in the genital area. Seek for medical attention if you have a suspicious symptom, so that a clinical and laboratory diagnosis can be established. Recommendations for healthcare professionals Pregnant women should be at home isolated with constant follow-up by the care team in case of mild illness. Cases of greater severity should be followed-up in hospital. There is still no specific treatment protocol with antivirals in the pregnancy-puerperal cycle. Monitoring fetal vitality should be carefully observed in patients with moderate or severe disease, because of the higher fetal morbidity and mortality in these cases. The delivery has obstetrics indications and cesarean section as a routine is not indicated. Breastfeeding should be postponed during the isolation period, offering specific support that allows for further relactation. Conflict of Interests The authors have no conflict of interests to declare. ==== Refs References 1 Centers for Disease Control and Prevention Monkeypox: Travel [Internet] 2022[cited 2022 Jul 24]. Available from:https://wwwnc.cdc.gov/travel/notices/alert/monkeypox 2 Beer E M Rao V B A systematic review of the epidemiology of human monkeypox outbreaks and implications for outbreak strategy PLoS Negl Trop Dis 2019 13 10 e0007791 10.1371/journal.pntd.0007791 31618206 3 World Health Organization Multi-country monkeypox outbreak: situation update [Internet] 2022[cited 2022 Jul 25]. Available from:https://www.who.int/emergencies/disease-outbreak-news/item/2022-DON396 4 Bunge E M Hoet B Chen L Lienert F Weidenthaler H Baer L R The changing epidemiology of human monkeypox-A potential threat? A systematic review PLoS Negl Trop Dis 2022 16 02 e0010141 10.1371/journal.pntd.0010141 35148313 5 Pittman P R Martin J W Kingebeni P M Clinical characterization of human monkeypox infections in the Democratic Republic of the CongomedRxiv [Preprint]. 2022 [cited 2022 Jul 24];2022.05.26.22273379. Available from:https://www.medrxiv.org/content/10.1101/2022.05.26.22273379v1 6 Mbala P K Huggins J W Riu-Rovira T Ahuka S M Mulembakani P Rimoin A W Maternal and fetal outcomes among pregnant women with human monkeypox infection in the Democratic Republic of Congo J Infect Dis 2017 216 07 824 828 10.1093/infdis/jix260 29029147 7 World Health Organization Clinical management and infection prevention and control for monkeypox: interim rapid response guidance [Internet] Geneva WHO 2022[cited 2022 Jul 24]. Available from:https://www.who.int/publications/i/item/WHO-MPX-Clinical-and-IPC-2022.1 8 Meaney-Delman D M Galang R R Petersen B W Jamieson D J A primer on monkeypox virus for obstetrician-gynecologists diagnosis, prevention, and treatment Obstet Gynecol 2022 140 03 391 397 10.1097/AOG.0000000000004909 36356237 9 Khalil A Samara A O'Brien P Morris E Draycott T Lees C Monkeypox and pregnancy: what do obstetricians need to know? Ultrasound Obstet Gynecol 2022 60 01 22 27 10.1002/uog.24968 35652380 10 Dashraath P Nielsen-Saines K Mattar C Musso D Tambyah P Baud D Guidelines for pregnant individuals with monkeypox virus exposure Lancet 2022 400 (10345):21 22 10.1016/S0140-6736(22)01063-7 35750071 11 SHARE-net Clinical Group Thornhill J P Barkati S Walmsley S Rokstroh J Antinori A Harrison L B Monkeypox virus infection in humans across 16 countries—April–June 2022 N Engl J Med 2022 387 08 679 691 10.1056/NEJMoa2207323 35866746 12 Ministério da Saúde. Secretaria de Atenção Primária à Saúde. Departamento de Saúde da Família. Coordenação-Geral de Garantia dos Atributos da Atenção Primária. Nota Informativa No. 6/2022-CGGAP/DESF/SAPS/MS, de 6 de julho de 2022. Trata-se de orientações às equipes que atuam na Atenção Primária à Saúde acerca da doença Monkeypox (MPX)[Internet]. 2022 [cited 2022 Jul 27]. Available from: NOTA INFORMATIVA N° 6_2022_CGGAP_DESF_SAPS_MS.pdf 13 UK Health Security Agency Monkeypox: background information [Internet] 2022[cited 2022 Jul 25]. Available from:https://www.gov.uk/guidance/monkeypox 14 Grosenbach D W Honeychurch K Rose E A Chinsangaram J Frimm A Maiti B Oral tecovirimat for the treatment of smallpox N Engl J Med 2018 379 01 44 53 10.1056/NEJMoa1705688 29972742 15 Verheust C Goossens M Pauwels K Breyer D Biosafety aspects of modified vaccinia virus Ankara (MVA)-based vectors used for gene therapy or vaccination Vaccine 2012 30 16 2623 2632 10.1016/j.vaccine.2012.02.016 22342706 16 UK Health Security Agency Monkeypox vaccination recommendations: recommendations for the use of pre and post exposure vaccination during a monkeypox incident[Internet]. 2022 [cited 2022 Jul 25]. Available from:https://www.gov.uk/government/publications/monkeypox-vaccination 17 European Medicines Agency Imvanex: Smallpox and monkeypox vaccine (Live Modified Vaccinia Virus Ankara)[Internet]. 2022 [cited 2022 Jul 27]. Available from:https://www.ema.europa.eu/en/medicines/human/EPAR/imvanex 18 Centers for Disease Control and Prevention Patient's guide to monkeypox treatment with TPOXX[Internet]. 2022 [cited 2022 Jul 25]. Available from:https://www.cdc.gov/poxvirus/monkeypox/treatment.html 19 Khalil A Samara A O'Brien P Morris E Draycott T Lees C Monkeypox vaccines in pregnancy: lessons must be learned from COVID-19 Lancet Glob Health 2022 10 09 e1230 e1231 10.1016/S2214-109X(22)00284-4 35772413 20 Centers for Disease Control and Prevention Clinical considerations for monkeypox in people who are pregnant or breastfeeding[Internet]. 2022 [cited 2022 Jul 25]. Available from:https://www.cdc.gov/poxvirus/monkeypox/clinicians/pregnancy.html
PMC009xxxxxx/PMC9800144.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36580934 10.1055/s-0042-1760208 RBGO-44.12-Editorial Editorial The Use of Three-dimensional Printer Molds for Treatment of Vaginal Agenesis http://orcid.org/0000-0001-5698-0245 Takano Claudia Cristina 1 http://orcid.org/0000-0002-3001-6076 Sartori Marair Gracio Ferreira 1 1 Department of Gynecology, Escola Paulista de Medicina, Universidade Federal de São Paulo, São Paulo, SP, Brazil Address for correspondence Marair Gracio Ferreira Sartori R. Botucatu, 740, 04023-062, Vila Clementino, São Paulo, [email protected] 29 12 2022 12 2022 1 12 2022 44 12 10811082 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. ==== Body pmcAlthough rare, vaginal agenesis is a relevant condition for gynecologists, who must be familiar with its current treatment. It results from agenesis of the Mullerian ducts, known as Mayer-Rokitansky-Kuster-Hauser Syndrome (MRKHS), and the incidence is 1:5000 women. In this congenital malformation, genetic alterations affect the development of Mullerian ducts during the embryonic period and there is complete absence or significant hypoplasia of the uterus and vagina, with normal development of the external genitalia and breasts. More rarely, the absence of the uterus and vagina is identified in patients with 46, XY Disorders of Sex Development (DSD) in which the presence of anti-Mullerian hormone inhibits the formation of the Mullerian ducts. In complete androgen insensitivity (Morris syndrome), the absence of testosterone action on its receptors leads to female differentiation of the external genital organs, and the conversion of testosterone to estrogen in peripheral tissues leads to the development of breasts at puberty. The clinical picture is similar to that of Rokitansky Syndrome, and in most cases, this is the initial diagnosis. The gynecologist will differentiate one from the other; in some cases, the suspicion is based on the lack of pubic and axillary hair and/or the presence of palpable gonads in the inguinal canal, but is confirmed by elevated levels of testosterone and the karyotype. Treatment will be the same as that of Rokitansky's syndrome, except for the recommendation to evaluate the gonads, given the higher risk of developing gonadoblastoma. The current recommendation is to wait for the end of puberty to consider gonadectomy, so that secondary characteristics can develop without the need for hormone replacement therapy. 1 As soon as the diagnosis is confirmed, the treatment of vaginal agenesis involves the steps established by the American College of Obstetricians and Gynecologists (ACOG). It begins by informing and advising the patient and her family about the condition, options and timing of treatment, and explaining about sexual relationships and reproductive future. It also involves referrals to psychological support and encouraging participation in support groups. 1 The approach regarding the formation of the neovagina is well established. The time to perform it is decided by the woman, when she manifests the desire to start a sexual relationship and demonstrates maturity and motivation to understand and participate in treatment, which generally occurs at the end of adolescence. Individual aspects inherent to this decision must be considered, such as the family context, religion and sexual orientation. Since 2006, the ACOG recommends that “Nonsurgical creation of the vagina is the appropriate first-line approach in most patients”. 2 This approach is based on a success rate greater than 90%, which is similar to surgery, although with unquestionably smaller morbidity and costs. 1 3 Dilation is performed by the patient at home, on a daily basis, after detailed guidance from the gynecologist and supervision and follow-up throughout the process. Monitoring with a specialized physiotherapist is always beneficial, and essential when hypertonicity of the pelvic floor muscles is identified. In Brazil, it is difficult to acquire rigid dilators, which are made of resistant material such as polylactic acid, since they are not commercially available in the country. Adapted devices such as acrylic candles and silicone dilators are commonly used. The use of Additive Manufacturing (AM) technology and the three-dimensional printing device (3D Printer) have shown great potential for contribution and innovation in the health area. The 3D printer can create an object through its digital design. The model is evaluated and recognized by the three-dimensional printing device (3D Printer) through Computer Aided Manufacturing (CAM), the software that performs the processes of reading, analysis and digital slicing. Additive manufacturing technology is based on the deposition of layers to build the physical object. 4 The high customization capacity and the possibility of creating prototypes quickly, as well as the production of objects with complex geometries, have enabled the use of this technology in the development of products in the medical field. Additive manufacturing is also an economically viable technology in the production of small batches of customized products compared to conventional methods, making it an interesting alternative in the production and research of customized products. 4 According to a study by Fernandes et al. 5 published in the current issue, the application of this technology in the production of dilators for vaginal agenesis proved to be effective, economically viable, accessible and reproducible. Therefore, dilators can be produced in a gynecological care service equipped with a 3D printer and a qualified professional, allowing women with vaginal agenesis to have access to the recommended treatment for their condition. In addition to women with agenesis, these molds can also be used in other conditions in which dilation may be necessary, such as strictures and shortening of the vagina after radiotherapy or surgery. There is also the possibility of using it in the manufacture of other devices in Urogynecology, such as customized pessaries for the treatment of genital prolapse and urinary incontinence. The use of 3D printing technology reveals the importance of combining knowledge in the field of technology and health, as it enables the development of products with direct impact on medical treatment, in addition to opening up promising perspectives in other areas of Gynecology. Conflicts to Interest None to declare. ==== Refs References 1 Committee on Adolescent Health Care ACOG Committee Opinion No. 728: Müllerian agenesis: diagnosis, management, and treatment Obstet Gynecol 2018 131 01 e35 e42 10.1097/AOG.0000000000002458 29266078 2 ACOG Committee on Adolescent Health Care ACOG Committee Opinion No. 355: Vaginal agenesis: diagnosis, management, and routine care Obstet Gynecol 2006 108 06 1605 1609 10.1097/00006250-200612000-00059 17138802 3 Apfel V R Takano C C Marquini G V de Jarmy di Bella Z IK Girão M JBC Sartori M GF Treatment for vaginal agenesis: A prospective and comparative study between vaginal dilation and surgical neovaginoplasty Int J Gynaecol Obstet 2022 157 03 574 581 10.1002/ijgo.13931 34534375 4 Chrispin T T Fernandez M S Novoa C C Sartori M G Development of personalized molds for neovagina creation by 3D printer Rev Assoc Med Bras (1992) 2020 66 11 1498 502 10.1590/1806-9282.66.11.1498 33295399 5 Fernandes M S Takano C C Chrispin T TB Marquini G V Girão M JBC Sartori M GF Three-dimensional printer molds for vaginal agenesis: An individualized approach as conservative treatment Rev Bras Ginecol Obstet 2022 44 12 1110 1116 36138536
PMC009xxxxxx/PMC9800145.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36580944 10.1055/s-0042-1760235 febrasgostatement Febrasgo Position Statement Abnormal uterine bleeding and chronic iron deficiency Number 11 – December 2022 http://orcid.org/0000-0002-5157-1597 Barros Venina Viana de 1 http://orcid.org/0000-0001-8059-5873 Hase Eliane Azeka 2 http://orcid.org/0000-0002-2814-1396 Salazar Cristiano Caetano 3 http://orcid.org/0000-0003-3569-3559 Igai Ana Maria Kondo 4 http://orcid.org/0000-0002-7908-9073 Orsi Fernanda Andrade 5 http://orcid.org/0000-0003-1898-520X Margarido Paulo Francisco Ramos 6 1 Departamento de Obstetrícia e Ginecologia, Hospital das Clínicas, Faculdade de Medicina, Universidade de São Paulo, São Paulo, SP, Brazil 2 Departamento de Obstetrícia e Ginecologia, Hospital das Clínicas, Faculdade de Medicina, Universidade de São Paulo, São Paulo, SP, Brazil 3 Hospital Moinhos de Vento, Porto Alegre, RS, Brazil 4 Departamento de Obstetrícia e Ginecologia, Hospital das Clínicas, Faculdade de Medicina, Universidade de São Paulo, São Paulo, SP, Brazil 5 Departamento de Hematologia, Hospital das Clínicas, Faculdade de Medicina, Universidade de São Paulo, São Paulo, SP, Brazil; Departamento de Hematologia da Universidade Estadual de Campinas, Campinas, SP, Brazil 6 Departamento de Obstetrícia e Ginecologia, Hospital das Clínicas, Faculdade de Medicina, Universidade de São Paulo, São Paulo, SP, Brazil 29 12 2022 12 2022 1 12 2022 44 12 11611168 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. ==== Body pmcKey points Abnormal uterine bleeding (AUB) in menacme is the leading cause of iron deficiency anemia (IDA) and iron deficiency (ID). All patients with AUB should be investigated and treated. Anemia is one of the most common problems in clinical practice that affects millions of people worldwide. Non-pregnant women account for 30% of all anemia cases in the world, and approximately 60% of them have ID. Oral iron replacement is the most widespread, especially in cases of milder IDA and ID. Intravenous (IV) formulations have gained more space in prescriptions as their safety and efficacy have become more evident. Recommendations Abnormal uterine bleeding is a very frequent complaint that negatively affects the quality of life since menacme. Investigation for IDA and ID is mandatory in these patients. The approach to patients with AUB prioritizes stabilization in acute cases, using mainly hormones and antifibrinolytics to stop bleeding. Etiological investigation will guide the therapy in non-acute cases. Treatment selection for ID is driven by several factors, including the presence of inflammation, the time available for iron replacement and the anticipated risk of side effects or intolerance. The treatment of choice for ID is preferably via oral (VO). The increase in hepcidin by oral iron supplements limits oral absorption when large amounts of iron need to be administered or in the presence of inflammatory conditions. Intravenous iron preparations are indicated for the treatment of ID when oral medications are ineffective or cannot be used. They have applicability in a wide range of clinical settings, including chronic inflammatory conditions, perioperative situations, and disorders associated with chronic blood loss. Serious adverse events that occur with IV iron are very rare and well-studied, which provides a basis for educating and preparing staff and patients on how iron infusions can be safely and effectively administered. Background One of the most common gynecological complaints worldwide is the occurrence of abnormal uterine bleeding (AUB), a term that refers to abnormalities in the amount, duration or frequency of bleeding from the uterus. With a prevalence of 10-30% among women of reproductive age, it can negatively affect the quality of life and is associated with financial losses, reduced productivity, inadequate health status and greater use of health services. 1 2 What are the main causes of AUB and how to classify them? Abnormal uterine bleeding is a symptom, not a diagnosis, and describes bleeding that deviates from the general menstrual pattern of the population. The terms and parameters currently used are described in chart 1 . Abnormal uterine bleeding can also be characterized as acute (severe enough episode that requires immediate intervention), chronic (occurring in most cycles in the previous six months) and intermenstrual bleeding (occurring between defined cycles and predictable menstruation). 1 3 Chart 1. Definitions of normal and abnormal menstrual bleeding Parameter Descriptive term Definition Frequency (interval between the start of each cycle) Amenorrhea No bleeding for 90 days Infrequent >38 days Normal 24 to 38 days Frequent <24 days Regularity (variation in duration between the longest and shortest cycle in 12 months) Regular ≤7 to 9 days Irregular ≥10 days Duration (duration of bleeding) Normal ≤8 days Prolonged >8 days Volume (total blood loss) Mild Patient perceives as mild Normal Patient considers normal Heavy Patient considers heavy Intermenstrual bleeding (bleeding between regular menstrual cycles) Absent (normal) No bleeding Random Present, not predictable Cyclic Present, predictable (at the beginning, middle or end of the cycle) Unscheduled bleeding in gonadal steroid users (estrogen ± progestin) Normal Absent Abnormal Present (Not applicable) No steroid use Causes The International Federation of Gynecology and Obstetrics (FIGO) classifies the causes of non-pregnancy-related AUB under the PALM-COEIN acronym, referring to Polyps, Adenomyosis, Leiomyoma, Malignancy and hyperplasia, Coagulopathy, Ovulatory dysfunction, Endometrial disorders, Iatrogenic and Not otherwise classified. In general terms, the first group (“PALM”) refers to structural causes (mostly identifiable by imaging exams or histopathology), and the other group (“COEIN”) refers to non-structural causes. The term “dysfunctional uterine bleeding” (DUB), in turn, refers to causes related to hemostasis (“C”), ovulatory dysfunction (“O”) and endometrial primary disorders (“E”), according to the current FIGO classification system. 3 Should all patients with AUB be investigated and treated? Management of patients with AUB includes assessment of hemodynamic instability and anemia, identification of the source of bleeding, and exclusion of pregnancy. Initially, it is important to determine whether it is acute or non-acute bleeding. The etiological diagnosis will guide therapy and treatment success. 4 5 However, in situations of acute and severe bleeding, treatment can be instituted to stop acute bleeding, followed by investigation. Although the uterus is often the source of abnormal bleeding, any part of the female genital tract (vulva, vagina) may have externalized vaginal bleeding, and this differential diagnosis is necessary. Initial physical examination may reveal vulvar or cervical lesions, guiding specific therapy. Anamnesis focused on the bleeding pattern, the use of medications and the association of other characteristics, signs and symptoms can guide the investigation, leading to the most likely etiologies ( Chart 2 ). Chart 2. Differential diagnosis of AUB in non-pregnant women of reproductive age Bleeding pattern Associated clinical features Etiologies to consider Regular, heavy, or prolonged periods Enlarged uterus on physical examination, with or without palpable masses Leiomyomas Dysmenorrhea Enlarged and softened uterus on physical examination Adenomyosis Family history of blood dyscrasia Symptoms of hemorrhagic diathesis Anticoagulant therapy Coagulopathies Risk factors for uterine cancer Endometrial carcinoma, uterine sarcoma Regular periods with intermenstrual bleeding   Endometrial polyp Risk factors for uterine cancer Endometrial carcinoma, uterine sarcoma Recent history of cervical or uterine procedure, or delivery/cesarean section, especially if infected Chronic endometritis Irregular bleeding, more or less frequent than normal periods, with variable volume and duration   Ovulatory dysfunction Hirsutism, acne, obesity Polycystic ovary syndrome Galactorrhea Hyperprolactinemia Recent weight loss or gain Cold or heat intolerance Family history of thyroid disease Thyroidopathy Risk factors for uterine cancer Endometrial carcinoma, uterine sarcoma How to perform the management and follow-up of AUB? Acute uterine bleeding When there is acute and severe blood loss and the patient is anemic and hypovolemic, hypotensive, tachycardic or with orthostatic hypotension, before determining the etiology, measures to stop bleeding are adopted. The first step is reestablishing hemodynamic stability with the use of crystalloids and eventually, the use of vasopressors and blood components. Pharmacological measures The use of high doses of intravenous estrogen causes rapid endometrial growth, stimulates contraction of the uterine arteries, and promotes platelet aggregation and clotting. Intravenous conjugated estrogen 25 mg every four to six hours for the first 24 hours is suggested, followed by a combination of estrogen and progestin for the following days. 1 4 6 Combined oral contraceptives (COCs), more widely available in our country, can also be used to treat acute AUB. A COC with 35 mcg of ethinyl estradiol (or other combination of pills to achieve this dose) three times a day for seven days is indicated. However, both estrogen treatments (oral or intravenous) should be avoided in patients at high risk of thromboembolism. An alternative is the use of multiple doses of progestins, especially in cases where estrogens are contraindicated. Medroxyprogesterone acetate 20 mg three times a day, norethisterone 5 mg three times a day, or another high-dose progestogen can be used for seven days, followed by one dose a day for three weeks. 1 4 Another suggested option in the literature is the use of a gonadotropin-releasing hormone (GnRH) agonist associated with an aromatase inhibitor or GnRh antagonist. 4 All these hormonal options, after a higher loading dose and a lower maintenance dose for a week or period of a menstrual cycle, in general, can be maintained while etiological investigation is performed. In addition to hormonal alternatives, tranexamic acid can be used to manage acute bleeding; 10 mg/kg of body weight are given intravenously every eight hours (most effective) or 20-25 mg/kg orally every eight hours. The use of antifibrinolytics can reduce bleeding by up to 50%. Caution should also be exercised in patients at high risk of thromboembolism. 1 Nonpharmacological measures In some emergencies in which hemodynamic instability persists despite the drug treatment instituted, it is necessary to resort to mechanical or surgical procedures. An alternative is to try to tamponade the uterus by inserting a Foley catheter and fill the balloon with 10-30 mL of saline or distilled water. Sometimes it is necessary to perform a uterine curettage to stop bleeding. If severe bleeding persists, uterine artery embolization or even hysterectomy should be considered, depending on reproductive desire and bleeding severity. Non-acute uterine bleeding For these patients, the objective is to continue the diagnostic investigation and institute management directed at the cause. Pelvic ultrasound is the complementary test that provides more data for the management of AUB cases and has sensitivity of 96% and specificity of 14% for uterine abnormalities. Saline-infused sonography may better reveal intracavitary pathologies (such as polyps and fibroids). Histopathological evaluation (endometrial biopsy) is always indicated in postmenopausal patients, in those aged 45 years or older and those at high risk for endometrial carcinoma. According to risk factors in patients with suspected coagulopathy, the platelet count, plasma fibrinogen, prothrombin time and activated partial thromboplastin time should be evaluated. Sometimes it is necessary to follow the investigation with a von Willebrand factor and platelet aggregation tests, as well as with a hemophilia test – especially in patients with reports of ecchymosis or easy bleeding and in those with a suggestive family history. Patients using anticoagulants (coumarins, heparins, direct oral anticoagulants) should have their therapeutic regimen optimized. When an infectious cause is suspected, tests for gonococcus, chlamydia, and trichomoniasis are performed. If hormonal causes are suspected, it is critical to evaluate prolactin, thyroid tests, gonadotropins, and androgens, as well as other ways to diagnose chronic anovulation or polycystic ovary syndrome. Generally speaking, management is different for structural and non-structural etiologies. Structural causes In causes grouped in the first part (“PALM”) of the PALM-COEIN acronym, the aim of treatment is mostly the structural pathology. Non-structural causes The causes grouped in the second part (“COEIN”) of the PALM-COEIN acronym, the so-called “non-structural”, and some structural causes of AUB have the treatment focus on satisfactory control of bleeding, regardless of etiology. Levonorgestrel intrauterine system (LNG-IUS) Continuously released levonorgestrel (20 mcg/day) from the LNG-IUS is the most effective measure to prevent heavy menstrual bleeding, leading to a 71-95% reduction in blood loss by promoting endometrial atrophy. Other isolated systemic progestins Progestins promote endometrial atrophy and have anti-inflammatory action, although it is not fully understood how they reduce uterine bleeding. They can be indicated for most women, especially those with contraindications to the use of estrogens. Continuous oral progestins are effective in the treatment of AUB, reducing bleeding by up to 87% and promoting amenorrhea in a large percentage of women (10%-15%). 7 Progestogens-only contraceptive pills can be used: norethisterone 0.35 mg, desogestrel 75 mcg, drospirenone 5 mg, or micronized progesterone (200-400 mg/day) in continuous use, or from day 5 to day 26 of the menstrual cycle. Injectable depot medroxyprogesterone acetate (150 mg intramuscularly every three months) can promote amenorrhea in up to 24% of women and is an option for women with increased bleeding. However, there is no conclusive evidence regarding the use of injectable progestin in AUB. Likewise, there are not enough studies to indicate the etonorgestrel implant to manage AUB, even though it promotes amenorrhea in 20% of users. 4 Estrogen and progesterone combinations Combined oral contraceptives are able to reduce menstrual bleeding by 35-69%. They are a widely available and effective therapeutic option for most cases of AUB without structural change. Monophasic formulations containing 30 to 35 mcg of ethinylestradiol are the most studied, but the most diverse presentations of COCs are effective. 1 4 A quadriphasic formulation containing dienogest with estradiol valerate (10-30 mcg/day) showed a reduction of menstrual volume thus, is an alternative. 8 Other routes (transdermal patch, vaginal ring) are probably as effective as the oral options, and may be superior when there is an indication to avoid first-pass effect. Non-hormonal treatments Tranexamic acid is the most frequently prescribed antifibrinolytic, associated with a 26-54% reduction in the amount of bleeding. When the patient is bleeding, it is used in doses ranging from 1 to 1.5 g, three to four times a day for about three to five days. It can be used by women with the intention of becoming pregnant, but not by women at higher risk of thromboembolism. Non-steroidal anti-inflammatory drugs (NSAIDs) can be used alone or as an adjuvant therapy to some hormonal treatment, reducing the amount of bleeding by 10% to 52%. The most studied medications are mefenamic acid (500 mg orally three times daily) and naproxen (500 mg orally twice daily) while patient is experiencing bleeding. They can be used by women who are trying to get pregnant, but should be avoided by patients with coagulopathies. Surgical treatments Endometrial ablation is a less invasive alternative to hysterectomy for patients with AUB without structural damage. The aim is to destroy the basal layer of the endometrium through a series of methods (laser, thermal balloon, vaporization, cryoablation, bipolar radiofrequency, microwave), preventing its regeneration. This is an option only for those who no longer wish to get pregnant. The amenorrhea rate is 40-50% in one year, with good results in uteri with hysterometry less than 10 cm. 4 Hysterectomy is an exception treatment for structural AUB reserved for patients with no reproductive desire and unsuccessful drug management. However, it is the most effective definitive treatment and achieves high levels of satisfaction. Iron deficiency anemia and iron deficiency: consequence of AUB conditions In the presence of profuse acute uterine bleeding, acute anemia, hypotension, shock and even death can occur if prompt intervention is not performed. Chronic AUB, in turn, is an important cause of ID, as are parasitic infections, gastrointestinal bleeding, and nutritional deficiencies, which can lead to anemia. 2 9 10 Anemia is defined as a condition in which the concentration of hemoglobin (Hb) in the blood is below normal and can be determined by several factors, with 50% of cases comprising IDA or ID. 11 12 The usual symptoms of IDA include weakness, headache, irritability, restless legs syndrome and varying degrees of fatigue and exercise intolerance or pica (perverted appetite for clay or soil, paper, starch, etc.). 12 13 Patients with low ferritin and without anemia may have the same symptoms. Ice pica may also occur, which is considered quite specific for low ferritin. Some patients with low ferritin, with or without anemia, may complain of tongue pain, decreased salivary flow with dry mouth, and atrophy of the lingual papillae. 13 Other patients may present with alopecia, dry skin, devitalized hair, and koilonychia. 14 However, many patients are asymptomatic, with no typical symptoms, and only recognize symptoms retrospectively after treatment. The differential diagnosis of IDA includes parasitic diseases such as malaria, hookworm and schistosomiasis, nutritional causes such as lack of folic acid, vitamin A and vitamin B12, and genetic causes such as hereditary thalassemia-type hemoglobinopathies. 15 How to make the laboratory diagnosis of IDA and ID? When IDA or ID is suspected, 16 17 a complete blood count (with RBC indices and peripheral smear evaluation) and ferritin levels should be requested ( Chart 3 ). 18 Chart 3. Laboratory parameters to define iron deficiency anemia (IDA) and iron deficiency (ID)   Adult normal values IDA Latent ID IDA refractory to iron treatment Anemia of chronic disease IDA + anemia of inflammation Serum iron (μmol/L) 10-30 ↓ N/↓ ↓ ↓ N/↓ Transferrin saturation (%) 20-45 <20 N/↓ <10 N/↓ <20 Serum ferritin (μg/L) 20-200 (F) 40-300 (M) <30 <30 Variable >100 30-300 Reticulocyte hemoglobin (pg) >29 <29 <29 <29 <29 <29 Hemoglobin (g/dL) >12 (F) >13 (M) ↓ N ↓ ↓ ↓ MCV (fL) 80-100 ↓ N/↓ ↓ N/↓ N/↓ Source: Adapted from Elstrott et al. (2020) 18 Other measurements, such as serum iron, transferrin, and transferrin saturation, are not mandatory. Patients with IDA have low serum iron, high transferrin, and low transferrin saturation. 12 13 Reticulocyte hemoglobin (Ret-Hb) is a good indicator of the amount of available iron, and its dosage does not interfere with inflammatory processes. 19 According to the diagnostic standards of the World Health Organization, IDA is mild to moderate if Hb is between 7 and 12 g/dL, and severe, if Hb is less than 7 g/dL, with small variations according to age, sex or presence of pregnancy. 20 For the adult female population, Hb values below 12 g/dL are considered as anemia and for men, Hb values below 13 g/dL. 20 21 Although Hb is widely used for the evaluation of IDA, it has low specificity and sensitivity, and a biomarker of iron status, such as serum ferritin, should be requested together. 21 Serum ferritin concentration is the most reliable marker of iron storage in the body. Normal values range from 30 to 200 ng/mL (mcg/L), and there is no clinical situation in which low rates do not mean ID. As long as patients with IDA do not have infection or associated inflammatory disease, the cutoff value of 30 ng/mL gives better diagnostic efficiency, with sensitivity of 92% and specificity of 98%. As ferritin is an acute-phase reactor with increased levels in inflammatory, infectious, malignant, or liver diseases, falsely elevated ferritin may be found in the presence of these diseases and IDA. The effect of inflammation on ferritin is to increase it threefold. Therefore, in these patients, the golden rule is to divide the ferritin value by 3, and values less than or equal to 20 ng/mL suggest concomitant IDA. 22 What are the main treatments for iron deficiency? The main treatments for IDA and ID are iron replacement, correction of nutritional aspects and treatment of AUB. The goal of iron replacement is to provide enough iron to normalize Hb concentrations and replenish iron storage, thereby improving quality of life and symptoms. 13 Regardless of the presence of symptoms, all patients with IDA and most of those with ID without anemia should be treated. 23 24 There are two distinct approaches: prevention strategies targeting populations at risk, such as patients with AUB, and active iron supplementation approaches in confirmed IDA. 13 Nutritional guidance It is recommended to increase the intake of meat, the main source of heme iron; it is estimated that 100 g of meat corresponds to 1 kg of beans (non-heme iron). Concomitant consumption of fruit juice with vitamin C enhances the absorption of iron from the diet, and the use of an iron pan to prepare meals is also part of the guidelines. It is recommended not to mix milk and tea at the same meal and avoid whole grain cereals and chocolate as a dessert during the period of treatment with ferrous salt. These recommendations are not necessary when ferric salts are used in the treatment, because in these compounds, iron is chelated with sugar or amino acids, and there is no interaction of its absorption with food in general. Foods rich in ascorbic acid (cashew, legumes, guava) and meats in general, favor the absorption of non-heme iron, while phytates, phosphates and carbonates (pineapple, vegetables, milk), tannin (tea, coffee), phosphoprotein (yolk eggs) and drugs that raise gastric pH (antacids, proton pump inhibitors, histamine H2 blockers) make absorption of non-heme iron difficult. Although intestinal iron absorption can increase significantly when iron is deficient (from less than 1% to more than 50% of the iron present in the diet), dietary correction alone is not usually sufficient to treat patients with IDA. 13 16 When and how to prescribe oral iron? What are the main indications and contraindications? Oral iron replacement is undoubtedly the most widespread, especially for lighter IDA and ID cases. However, doubts about the dosage are common and it is not uncommon to find situations when the drug is apparently ineffective. The recommended therapeutic dose is 2 to 5 mg/kg/day for a period sufficient to normalize Hb values – one to two months – and restore normal body iron stores – two to six months or until serum ferritin is greater than 30-50 ng/mL. 16 Therefore, the duration of treatment varies widely depending on the severity of ID and its cause. In practice, the recommended dose for adult individuals is 150 to 200 mg of elemental iron per day, and the administration of daily doses greater than 200 mg is not recommended, as, in this case, the intestinal mucosa acts as a barrier, preventing the absorption of the metal, and the proportion absorbed decreases significantly. 12 16 Chart 4 shows the products and doses available for oral treatment. Chart 4. Main compounds with iron salts available for the oral treatment of iron deficiency anemia (IDA) and iron deficiency (ID) COMPOUND Total amount of iron Amount of elemental iron Registration on the Anvisa website* Ferrous sulfat (mg) 190 60 40 Henfer/Anemifer Furp/Farmanguinhos/Nesh Ferric glycinate (mg) 150, 300, 500 30, 60, 100 Neutrofer Ferripolymaltose (mg) 333, 33/357 100 Endofer, Noripurum Research has shown that high doses of iron VO induce greater production of hepcidin, a protein produced by the liver that controls serum iron by blocking intestinal absorption and the release of iron from stores. 12 Therefore, the use in sequence or in overdose may paradoxically take away the effect of the drug. In some studies, lower doses of elemental iron per day, 15 to 20 mg, have shown equal effectiveness compared to higher doses, likely because of this mechanism. In addition to ensuring adequate absorption, dosage adjustments allow for better control of side effects (diarrhea, constipation, epigastric pain, nausea, dark-colored stools). 18 Numerous oral iron formulations are available and mostly, they are all equally effective, as long as they are taken. 23 Iron absorption from intestinal mucosal cells occurs through divalent metal transporter 1 (DMT1), a protein located in the duodenum and upper jejunum. Once in the cell, ferroportin transports iron through the cell into the blood, where it is bound by transferrin. The paradigm for iron replacement evolved as evidence began to emerge suggesting that excessive dosage is potentially counterproductive, as it decreases iron absorption and increases side effects without improving iron levels or anemia. 23 More research is needed to define the best strategy for oral iron administration. When and how to prescribe IV iron? What are the main indications and contraindications? Intravenous formulations have gained more space in prescriptions as the safety of their use has become more evident. Infusion reactions are rare, usually mild, and if they occur, drug administration can continue at a slower infusion rate. The various injectable formulations have the same efficacy and are especially useful in cases of more vigorous replacement, patients intolerant to oral administration or with malabsorptive processes (for example: patients with inflammatory bowel diseases) and chronic renal patients. Contraindications to use of IV iron are: anemia unrelated to ID, transferrin saturation > 45%, ferritin > 500 ng/mL, active infection/septicemia, severe dysfunction (hepatic or cardiac), pregnant women in the first trimester of pregnancy. Chart 5 shows the IV iron formulations available in Brazil and the main information for the use of these drugs. Chart 5. Intravenous iron formulations Compound Ferric hydroxide saccharate Ferric derisomaltose Ferric carboxymaltose Commercial name Noripurum; Sucrofer Monofer Ferinject Concentration 20 mg/mL 100 mg/mL 50 mg/mL Total Dose Determined individually, according to iron deficiency.* Determined individually, according to iron deficiency.* Hb < 10 g/dL: 1.500 mg if weight 35-70 kg; 2.000 mg if weight > 70 kg.   Simplified mean dosage: 100 to 200 mg one to three times a week Simplified mean dosage: Hb < 10 g/dL: 1.500 mg if weight 35-70 kg; 2.000 mg if weight > 70 kg Hb > 10 g/dL: 1.000 mg if weight 35-70 kg; 1.500 mg if weight > 70 kg. Hb > 10 g/dL: 1.000 mg if weight 35-70 kg; 1.500 mg if weight > 70 kg. Recommended maximum single dose 200 mg per day 500 mg daily, up to three times a week 1.000 mg daily, up to once a week Infusion time At least 1 hour 15 minutes 15 minutes Risk category in pregnancy B B B * calculation of iron need: total iron deficiency (mg) = [weight (kg) x DHb (g/dl) x 2.4] + iron reserves (mg) Adverse effects of IV iron Many physicians are reluctant to use IV iron because of concerns about anaphylaxis. True allergic reactions are extremely rare and overrated. In individuals with asthma, inflammatory rheumatic diseases, or multiple drug allergies, premedication with a glucocorticoid alone is generally recommended. Final considerations Although AUB is a very common condition, it should be valued and properly investigated, as it can significantly worsen a woman’s quality of life. According to the etiology, AUB can be effectively treated by quite effective pharmacological and surgical measures depending on age, reproductive desire and other associated conditions. Abnormal uterine bleeding in menacme is the main cause of IDA and ID. The tests requested for diagnosis must include, at least, the blood count, ferritin and iron profile. Iron replacement should be prescribed for these patients, and treatment monitoring is usually performed between 30 and 60 days, depending on the clinical picture. National Commission Specialized in Venous Thromboembolism and Hemorrhage in Women of the Brazilian Federation of Gynecology and Obstetrics Associations (Febrasgo) President: Venina Isabel Poco Viana Leme de Barros Vice-president: André Luiz Malavasi Longo de Oliveira Secretary: Paulo Francisco Ramos Margarido Members: Ana Maria Kondo Igai Cristiano Caetano Salazar Dênis José Nascimento Eduardo Zlotnik Egle Cristina Couto Eliane Azeka Hase Fernanda Andrade Orsi Joaquim Luiz de Castro Moreira Marcelo Melzer Teruchkin Marcos Arêas Marques Mônica Cristina da Costa Drago Souza Valeria Doria Mendes da Costa Conflicts of interest None to declare. The National Commission Specialized in Venous Thromboembolism and Hemorrhage in Women of the Brazilian Federation of Gynecology and Obstetrics Associations (Febrasgo) endorses this document. The production of content is based on scientific evidence on the proposed theme and the results presented contribute to clinical practice. ==== Refs References 1 Wouk N Helton M Abnormal uterine bleeding in premenopausal women Am Fam Physician 2019 99 07 435 43 30932448 2 Davis E Sparzak P B Abnormal uterine bleeding [Internet]. Treasure Island: StatPearls; 2022 [cited 2022 Jan 10]. Available from:https://www.ncbi.nlm.nih.gov/books/NBK532913/ 3 Munro M G Critchley H O Fraser I S Abnormal uterine bleeding in reproductive-age patients: FIGO System 1 terminology and symptoms and System 2 PALM-COEIN etiology classificationUpToDate2021 4 Benetti-Pinto A L Rosa e Silva A C Yela D A Soares Júnior J M Abnormal uterine bleeding Rev Bras Ginecol Obstet 2017 39 07 358 68 10.1055/s-0037-1603807 28605821 5 Kaunitz A M Abnormal uterine bleeding in nonpregnant reproductive-age patients: evaluation and approach to diagnosis [Internet]2021 [cited 2022 Jan 12]. Available from:https://www.uptodate.com/contents/abnormal-uterine-bleeding-in-nonpregnant-reproductive-age-patients-terminology-evaluation-and-approach-to-diagnosis 6 Machado L V Sangramento uterino disfuncional Arq Bras Endocrinol Metab 2001 45 04 375 82 10.1590/S0004-27302001000400010 7 Irvine G A Campbell-Brown M B Lumsden M A Heikkilä A Walker J J Cameron I T Randomised comparative trial of the levonorgestrel intrauterine system and norethisterone for treatment of idiopathic menorrhagia Br J Obstet Gynaecol 1998 105 06 592 8 10.1111/j.1471-0528.1998.tb10172.x 9647148 8 Rafie S Borgelt L Koepf E R Temple-Cooper M E Lehman K J Novel oral contraceptive for heavy menstrual bleeding: estradiol valerate and dienogest Int J Womens Health 2013 5 313 21 10.2147/IJWH.S31922 23788843 9 Munro M G Critchley H O Fraser IS; FIGO Menstrual Disorders Committee. The two FIGO systems for normal and abnormal uterine bleeding symptoms and classification of causes of abnormal uterine bleeding in the reproductive years: 2018 revisions Int J Gynaecol Obstet 2018 143 03 393 408 10.1002/ijgo.12666 30198563 10 Ministério da Saúde. Secretaria de Atenção à Saúde. Departamento de Atenção Básica Programa nacional de suplementação de ferro: manual de condutas gerais Brasília (DF) Ministério da Saúde 2013 11 World Health Organization Nutritional anaemias: tools for effective prevention and control [Internet] Geneva WHO 2017[cited 2021 Oct 20]. Available from:https://apps.who.int/iris/bitstream/handle/10665/259425/?sequence=1 12 Camaschella C Iron-deficiency anemia N Engl J Med 2015 372 19 1832 43 10.1056/NEJMra1401038 25946282 13 Lopez A Cacoub P Macdougall I C Peyrin-Biroulet L Iron deficiency anaemia Lancet 2016 387 (10021):907 16 10.1016/S0140-6736(15)60865-0 26314490 14 Pasricha S R Tye-Din J Muckenthaler M U Swinkels D W Iron deficiency Lancet 2021 397 (10270):233 48 10.1016/S0140-6736(20)32594-0 33285139 15 Berber I Diri H Erkurt M A Aydogdu I Kaya E Kuku I Evaluation of ferric and ferrous iron therapies in women with iron deficiency anaemia Adv Hematol 2014 2014 297057 10.1155/2014/297057 25006339 16 Cançado R D Chiattone C S Anemia ferropênica no adulto: causas, diagnóstico e tratamento Rev Bras Hematol Hemoter 2010 32 03 240 6 10.1590/S1516-84842010005000075 17 Grotto H Z Diagnóstico laboratorial da deficiência de ferro Rev Bras Hematol Hemoter 201032 Supl 2:22-810.1590/S1516-84842010005000046 18 Elstrott B Khan L Olson S Raghunathan V DeLoughery T Shatzel J J The role of iron repletion in adult iron deficiency anemia and other diseases Eur J Haematol 2020 104 03 153 61 10.1111/ejh.13345 31715055 19 Chinudomwong P Binyasing A Trongsakul R Paisooksantivatana K Diagnostic performance of reticulocyte hemoglobin equivalent in assessing the iron status J Clin Lab Anal 2020 34 06 e23225 10.1002/jcla.23225 32043622 20 World Health Organization Haemoglobin concentrations for the diagnosis of anaemia and assessment of severity [Internet] Geneva WHO 2011[cited 2021 Dec 21]. Available from:https://apps.who.int/iris/bitstream/handle/10665/85839/WHO_NMH_NHD_MNM_11.1_eng.pdf 21 Greig A J Patterson A J Collins C E Chalmers K A Iron deficiency, cognition, mental health and fatigue in women of childbearing age: a systematic review J Nutr Sci 2013 02 e14 10.1017/jns.2013.7 22 Means R T Jr Brodsky R A Diagnostic approach to anemia in adults [Internet] 2022[cited 2022 Jan 12]. Available from:https://www.uptodate.com/contents/diagnostic-approach-to-anemia-in-adults 23 Auerbach M Treatment of iron deficiency anemia in adults[Internet]. 2022 [cited 2022 Jan 12]. Available from:https://www.uptodate.com/contents/treatment-of-iron-deficiency-anemia-in-adults 24 Auerbach M Deloughery T Single-dose intravenous iron for iron deficiency: a new paradigm Hematology Am Soc Hematol Educ Program 2016 2016 01 57 66 10.1182/asheducation-2016.1.57 27913463
PMC009xxxxxx/PMC9800146.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36580938 10.1055/s-0042-1759632 RBGO-22-0164 Original Article Oncology Association of Obesity and Surgery Outcomes in Patients with Endometrial Cancer: A Single-Center Analysis http://orcid.org/0000-0003-3028-5804 Ozdemir Savas 1 http://orcid.org/0000-0001-9555-1128 Dogan Gul Ozel 1 1 Department of Gynecology and Obstetrics, University of Health Sciences, Prof. Dr. Cemil Tascioglu City Hospital, Istanbul, Turkey Address for correspondence Gul Ozel Dogan Sisli Hamidiye Etfal Egitim ve Arastirma Hastanesi Halaskargazi Cad.Etfal Sk., Post code: 34371 Sisli/[email protected] 29 12 2022 12 2022 1 12 2022 44 12 11171121 22 5 2022 02 9 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  Although obesity can result in high morbidity and mortality in surgical outcomes because of multiple comorbidities, determinants of outcome in obese patients who underwent endometrial cancer surgery remain unclear. The aim of this study is to assess the relationship between body mass index (BMI) and surgical outcomes in obese patients with endometrial cancer. Methods  An institutional retrospective review of the demographic details, clinical characteristics, and follow-up data of 142 patients with endometrial cancer who underwent surgery during a 72-month period was performed. The patients were divided into three groups based on their BMI; patients with BMI < 25 were identified as normal weight, patients with BMI between 25 and 30 were accepted as overweight, and those with BMI ≥ 30 kg/m 2 were identified as obese. The groups' demographic and clinical variables were compared. Results  Of the 142 patients, 42 were in the normal weight group, 55 in the overweight group, and 45 in the obese group. Age, surgical procedures, blood loss, preoperative health status, and metastatic lymph nodes did not show a significant difference between groups. However, surgery time and total lymph nodes were higher in the obese group. ( p  = 0.02, p  = 0.00, and p  = 0.00, respectively). Common complications were anemia, fever, intestinal injury, deep vein thrombosis, fascial dehiscence and urinary infection. There was no significant difference according to the complications. Conclusion  Our results indicated that higher BMI was significantly associated with a longer duration of endometrial cancer surgery. Minimally invasive surgeries and conventional laparotomy could be performed safely in obese patients. Keywords BMI endometrial cancer obesity surgery ==== Body pmcIntroduction Obesity is a well-established risk factor for developing endometrial cancer, more than any other cancer type. 1 Insulin resistance is responsible for releasing growth factors for cellular proliferation, higher levels of interleukins, tumor necrosis factors, and adipokines causing an obesity-related proinflammatory state, and high estrogen levels through increased aromatase activity in adipose tissue are proposed as contributors to the increased risk of developing endometrial cancer. 2 Endometrial cancer is known as a hormone-dependent type of cancer. Obesity affects hormone metabolism by increasing the aromatization of androstenedione to estrone in adipose tissues and causes an increase in the circulating levels of estrogen, creating a favorable environment for tumor formation. 3 The incidence of endometrial cancer is projected to rise as women's obesity rates continue to rise. In a study by Ward et al. 4 with 33,232 endometrial cancer patients, it was reported that the 10-year mortality due to endometrial cancer was associated with death due to cardiovascular disease. It was the most common reason related to morbid obesity. Obesity and endometrial cancer have been linked in numerous research studies. 5 6 7 8 Although excess body fat is a significant risk factor for endometrial cancer, its impact on survival is unclear. Surgical procedures for treating endometrial cancer are hysterectomy, bilateral salpingo-oophorectomy, and pelvic and para-aortic lymphadenectomy. Many studies have established that obese patients are at a higher risk of perioperative and postoperative complications, such as longer hospital stay and increased morbidity, even when minimally invasive surgeries or laparotomy are performed. 9 10 11 Obesity is defined by the World Health Organization (WHO) by using the body mass index (BMI) cutoff point of > 30 kg/m 2 , which is calculated as weight in kg divided by height in meters squared. Body mass index has significant public health importance because it correlates well with morbidity and/or mortality and endometrial cancer risk. Although surgery is the standard procedure in the staging and treatment of endometrial cancer, obesity may affect surgical outcomes due to its accompanying comorbid disorders. The impact of morbid obesity on endometrial cancer patients' survival is crucial as postoperative complications among obese women seem to be higher than among their normal-weight counterparts. 12 The evaluation of the effects of obesity on surgical outcomes may help decrease morbidity and improve prognosis in patients with endometrial cancer. However, there are insufficient data in the current literature that explain the impact of morbid obesity on the surgical outcomes of endometrial cancer and compare it with the endometrial cancer patients with normal weight. This study aims to determine the link between BMI and surgical outcomes in obese endometrial cancer patients. Methods Participants This study was performed in accordance with the ethical standards of the Helsinki declaration. Ethical approval was obtained for this retrospective, cross-sectional study by the ethics committee of the University of Health Sciences, Sisli Hamidiye Etfal Education and Research Hospital. Informed consent was obtained routinely. We reviewed the records of patients older than 18 years with endometrial cancer admitted to our hospital's gynecologic oncology department within a 72-month period from 2014 to 2020. Inclusion and Exclusion Criteria The inclusion criteria are listed below: Patients with pathologically proven endometrial cancer. Patients older than 18 years. A detailed medical record including patient's history, clinical findings, laboratory and pathology test results, treatment outcomes, etc. The exclusion criteria are as follows: Patients without a definite pathologic diagnosis Patients with secondary cancer. Patients with endometrial cancer who were treated conservatively. Data Collection The patients' demographic characteristics (age, sex), weight, height and body mass index (BMI), surgical procedure (total hysterectomy-bilateral salpingo-oophorectomy [via laparotomy or laparoscopy] with or without pelvic and para-aortic lymphadenectomy), duration of hospital stay, lymph node involvement, the average number of lymph nodes removed, routine biochemical examination, preoperative evaluation and preparation for anesthesia, perioperative and postoperative complications, and follow-up data were recorded. The patients were divided into three groups based on their BMI. Body mass index [kg/height (m)]2 was calculated and classified according to the World Health Organization (WHO) guidelines. Thus, patients with BMI < 25 were identified as normal weight, patients with BMI from 25 to 30 were accepted as overweight, and those with BMI ≥ 30 kg/m 2 were identified as obese. Later, we compared the variables mentioned above and surgical outcomes according to patients' BMI. Statistical Analysis Data were analyzed using the IBM SPSS Statistics for Windows, version 23.0 software (IBM Corp., Armonk, NY, USA). Descriptive statistics (mean, standard deviation, frequency, and percentage) were used for the demographic and clinical characteristics. The analysis of variance (ANOVA) test followed by Tukey multiple comparison methods among these three BMI groups was performed. The categorical variables were compared using the Chi-squared test. A p -value of < 0.05 was considered to be statistically significant. Results During the 6-year period, 142 patients were operated on for endometrial cancer. These surgical procedures were performed via laparotomy or conventional laparoscopy, which were performed at our hospital's gynecologic oncology unit. The mean age of the subjects was 60.52 ± 9.89 years (range, 18–82 years). The in-hospital mortality rate was 0%. The number of patients with BMI < 25 (normal weight) was 42 (29.6%), those with BMI from 25 to 30 (overweight) were 55 (38.7%), and patients with BMI > 30 (obese) were 45 (31.7%). The demographic and clinical characteristics of patients with a comparison between the three groups are demonstrated in Table 1 . There was no significant difference between the groups regarding age, surgical procedures, presence of menopause, intraoperative bleeding, preoperative health status (ASA), the mean duration of hospital stay, CA125 level, and the number of metastatic nodes. However, duration of surgery, the number of total nodes and non-metastatic nodes differed significantly and were higher in the group of patients with BMI > 30 ( p  = 0.02, p  = 0.00, and p  = 0.00, respectively). Table 2 shows the posthoc Tukey test results of these variables. Table 1 Demographic and clinical characteristics of patient groups Variables BMI < 25 (normal weight) BMI from 25 to 30 (overweight)  BMI > 30  (obese) P -value Number of patients 42 (29.6%) 55 (38.7%) 45 (31.7%) − Age 61.00 ± 11.15 60.00 ± 10.56 60.44 ± 10.42 0.901 Surgical procedure  Laparotomy 25 (17.6%) 31 (21.8%) 26 (18.3%) 0.952  Laparoscopy 17 (12.0%) 24 (16.9%) 19 (13.4%) Menopause  Premenopause 7 (4.9%) 10 (7.0%) 9 (6.3%) 0.922  Postmenopause 35 (24.6%) 45 (31.7%) 36 (25.4%) Estimated blood loss (ml) 253.81 ± 89.25 263.64 ± 108.77 273.33 ± 111.88 0.685 Surgery time (minutes) 142.14 ± 25.50 158.27 ± 36.69 169.89 ± 44.18 0.002 ASA 1.62 ± 0.66 1.49 ± 0.57 1.47 ± 0.59 0.451 Duration of hospital stay (days) 5.64 ± 1.91 5.16 ± 1.75 5.56 ± 2.03 0.404 Pre-CA125 23.98 ± 34.58 48.58 ± 139.33 54.16 ± 132.77 0.436 Number of total dissected pelvic lymph nodes 19.10 ± 10.62 23.47 ± 11.74 34.84 ± 14.31 0.000 Metastatic nodes 1.21 ± 2.08 1.05 ± 2.38 0.93 ± 2.06 0.837 Non-metastatic nodes 18.00 ± 10.76 22.42 ± 12.03 33.91 ± 13.90 0.000 Table 2 The posthoc comparisons using Tukey's HSD Factor Pairwise Comparison P -value Surgery time (minutes) Normal weight vs Overweight 0.083 Overweight vs Obese 0.256 Normal weight vs Obese 0.002 Total nodes Normal weight vs Overweight 0.196 Overweight vs Obese 0.000 Normal weight vs Obese 0.000 Non-metastatic Normal vs Overweighted 0.190 Overweight vs Obese 0.000 Normal weight vs Obese 0.000 Patient complaints at the time of admission and complications to our outpatient clinic according to patient groups were summarized in Table 3 . The reasons for patients' admissions were vaginal bleeding, abdominal pain, itching, and routine examination. The symptoms for hospital admission did not differ according to patients' BMI. Anemia (5.6%), fever (1.4%), intestinal injury (0.7%), deep vein thrombosis (0.7%), fascial dehiscence (0.7%), rupture of veins (0.7%), surgical site infections (0.7%), and urinary infection (2.8%) were common perioperative (intraoperative and postoperative) complications among all patients. There was no significant difference between the groups according to the perioperative complications. Table 3 Perioperative data and complications according to patient groups Clinical features BMI < 25 (normal weight) BMI from 25–30 (overweight)  BMI > 30  (obese) Total (n) Symptoms for hospital admission Bleeding 37 (26.1%) 38 (26.8%)  31 (21.8%) 106 (74.6%) Abdominal pain 0 (0.0%) 1 (0.7%)  1 (0.7%) 2 (1.4%) Itching 0 (0.0%) 1 (0.7%)  1 (0.7%) 2 (1.4%) Routine examination 5 (3.5%) 15 (10.6%) 12 (8.5%) 32 (22.5%) Perioperative complications Atrial fibrillation 1 (0.7%) 0 (0.0%) 0 (0.0%) 1 (0.7%) Anemia 3 (2.1%) 2 (1.4%) 3 (2.1%) 8 (5.6%) Fever 0 (0.0%) 1 (0.7%) 1 (0.7%) 2 (1.4%) Intestinal injury 0 (0.0%) 0 (0.0%) 1 (0.7%) 1 (0.7%) Deep vein thrombosis 0 (0.0%) 0 (0.0%) 1 (0.7%) 1 (0.7%) Fascial dehiscence 0 (0.0%) 0 (0.0%) 1 (0.7%) 1 (0.7%) Rupture of the iliac vein 0 (0.0%) 1 (0.7%) 0 (0.0%) 1 (0.7%) Bleeding of obturator vein 0 (0.0%) 1 (0.7%) 0 (0.0%) 1 (0.7%) Urinary infection 1 (0.7%) 2 (1.4%) 1 (0.7%) 4 (2.8%) Surgical site infections 0 (0.0%) 1 (0.7%) 0 (0.0%) 1 (0.7%) None 37 (26.1%) 47 (33.1%) 37 (26.1%) 121 (85.2%) Discussion The current study aimed to demonstrate the surgical outcomes pertaining to the management of endometrial cancer patients according to their BMI and to explore which variables were significant in patient morbidity. Our statistical analysis demonstrated that duration of surgery, number of total dissected pelvic lymph nodes, and non-metastatic nodes were higher in endometrial cancer patients with obesity. Demographic, clinical, and laboratory findings, such as patient's age, presence of menopause, blood loss, preoperative health status, hospital stay, and CA125 levels, were not related to patients' BMI. Moreover, our results indicate that surgical procedures which were performed via laparotomy or laparoscopy had no significant relationship with the subjects' weight. The symptoms for hospital admission did not differ according to patients' BMI. In addition, perioperative complications did not significantly differ among patient groups. Kokts-Porietis et al. 13 reviewed the studies that included estimated body fat with BMI to evaluate the relationship between obesity and mortality among endometrial cancer survivors. They reported that endometrial cancer survivors who were obese at the time of diagnosis had a higher risk of cancer recurrence and all-cause death but not endometrial cancer-specific mortality. In our study sample, the in-hospital mortality rate was also 0%, and we suggest that elevated BMI might not be related to mortality in endometrial cancer. In a retrospective study by Ward et al., 4 the authors evaluated the causes of death among women with endometrial cancer. They found that cardiovascular diseases were the leading reason of death from endometrial cancer. Although the causes of death in patients with endometrial cancer were not the main concern of the current study, we found that the patients with BMI > 30 had a lower preoperative health status, which shows the preoperative chronic medical conditions. In a review by Onstad et al., 14 the authors suggested that operating on obese patients was more difficult than on normal-weight patients with endometrial cancer due to technical aspects of the surgery that could affect visualization. In the current study, we detected that the duration of surgery in the obese patient group was longer than in the normal-weighted and overweight patients, which supports the researchers' conclusion. However, obesity did not significantly impact the surgical procedures in the three groups of this study. Similar to our results, Gabala et al. 10 reported that obesity did not affect the surgical techniques in endometrial cancer. In a study by Erkanli et al., 15 they investigated the effect of BMI on clinical and pathologic features and surgical morbidity in 42 patients with endometrial cancer. The number of participants was higher in the current study, and our findings support their results. They also did not find any difference in length of hospital stay and intraoperative or postoperative complications. It can be concluded that the surgical approach might be performed safely in morbidly obese endometrial cancer patients. There are mixed results in the retrospective studies regarding the impact of obesity on operative complications in the current literature. Similar to previous studies, we detected that postoperative complication rates did not differ significantly between the obese and non-obese patient groups. 10 11 15 16 On the contrary, Bouwman et al. 17 reported that elevated BMI was associated with an increased risk of postoperative surgical complications in morbidly obese patients who underwent laparotomy. Patients' characteristics may explain these different results, and they also depend on surgeons' experience and type of equipment. Only 14.8% of our patient sample had suffered from perioperative complications. Arrhythmia, blood loss, high fever, intestinal injury, deep vein thrombosis, fascial dehiscence, bleeding of veins, urinary infection and surgical site infections were detected. Even though obesity did not affect the course of surgery in this study, we believe that when considering endometrial cancer surgery, it is critical to recognize these complications to avoid them. Endometrial cancer continues to increase in incidence and mortality. Obesity is now recognized as an independent risk factor for endometrial cancer, accounting for more than half of all cases. Women diagnosed with endometrial cancer with a high BMI have a higher probability of morbidity. For this reason, obesity may adversely affect surgery outcomes. Therefore, we think that more studies determining this relationship between endometrial cancer and obesity are needed in the medical literature. This study has some limitations. First, only eligible data in the record were assessed because of its retrospective nature. Second, the study was performed at a single center. In addition, all the patients were operated on by the same experienced surgeons. However, a long-term period and a relatively high number of participants are the strengths of this study. Conclusion In conclusion, our findings indicate that increased BMI is related to higher surgery time in patients with endometrial cancer. However, obesity did not impact surgical outcomes such as blood loss, duration of hospital stay, and complication rates. Moreover, mortality rates in endometrial cancer were not affected by BMI. Endometrial cancer surgery can be safely and adequately managed among patients with obesity. Further prospective studies evaluating the impact of obesity on long-term follow-up of endometrial cancer surgery are warranted. Contributions Conflict of Interests The authors have no conflict of interests to declare. S. O. and G. O. D. designed the study. S. O. and G. O. D. collected the data. S. O. analyzed and interpreted the data. G. O. D. drafted the manuscript. All authors were comprehensively involved in all aspects of the study and in the preparation of the manuscript. All authors have read and approved the final version of the it. ==== Refs References 1 Million Women Study Collaboration Reeves G K Pirie K Beral V Green J Spencer E Bull D Cancer incidence and mortality in relation to body mass index in the Million Women Study: cohort study BMJ 2007 335 (7630):1134 10.1136/bmj.39367.495995 17986716 2 Balescu I Bacalbasa N Endometrial cancer prevention-does bariatric surgery play a role? Gıneco Eu 2017 13 47 37 39 3 Papadia A Ragni N Salom E M The impact of obesity on surgery in gynecological oncology: a review Int J Gynecol Cancer 2006 16 02 944 952 10.1111/j.1525-1438.2006.00577.x 16681794 4 Ward K K Shah N R Saenz C C McHale M T Alvarez E A Plaxe S C Cardiovascular disease is the leading cause of death among endometrial cancer patients Gynecol Oncol 2012 126 02 176 179 10.1016/j.ygyno.2012.04.013 22507532 5 Orekoya O Samson M E Trivedi T Vyas S Steck S E The mpact of obesity on surgical outcome in endometrial cancer patients: a systematic review J Gynecol Surg 2016 32 03 149 157 10.1089/gyn.2015.0114 27274182 6 Bergström A Pisani P Tenet V Wolk A Adami H O Overweight as an avoidable cause of cancer in Europe Int J Cancer 2001 91 03 421 430 10.1002/1097-0215(200002)9999:9999<:aid-ijc1053>3.0.co;2-t 11169969 7 Akbayır O Corbacıoglu Esmer A Numanoglu C Goksedef B PC Akca A Bakır L V Kuru O Influence of body mass index on clinicopathologic features, surgical morbidity and outcome in patients with endometrial cancer Arch Gynecol Obstet 2012 286 05 1269 1276 10.1007/s00404-012-2431-2 22729137 8 Linkov F Edwards R Balk J Yurkovetsky Z Stadterman B Lokshin A Taioli E Endometrial hyperplasia, endometrial cancer and prevention: gaps in existing research of modifiable risk factors Eur J Cancer 2008 44 12 1632 1644 10.1016/j.ejca.2008.05.001 18514507 9 Fambrini M Pieralli A Bitossi U Andersson K L Scarselli G Livi L Mini-laparotomy versus vaginal surgery for class II-III obese patients with early-stage endometrial cancer Anticancer Res 2012 32 02 707 712 22287767 10 Gaballa K Abdelkhalek M Refky B Gadelhak B Aboelnaga E M El-Beshbishi W The impact of obesity on surgical complications and disease recurrence in endometrial cancer: a retrospective study of 267 patients Resum Oncol 2020 6 01 11 14 10.21608/resoncol.2020.23506.1089 11 Rabischong B Larraín D Canis M Le Bouëdec G Pomel C Jardon K Long-term follow-up after laparoscopic management of endometrial cancer in the obese: a fifteen-year cohort study J Minim Invasive Gynecol 2011 18 05 589 596 10.1016/j.jmig.2011.05.015 21719359 12 Güzel A B Khatib G Küçükgöz Güleç Ü Gümürdülü D Vardar M A The impact of morbid obesity on survival of endometrial cancer Turk J Obstet Gynecol 2020 17 03 209 214 10.4274/tjod.galenos.2020.83773 33072426 13 Kokts-Porietis R L Elmrayed S Brenner D R Friedenreich C M Obesity and mortality among endometrial cancer survivors: A systematic review and meta-analysis Obes Rev 2021 22 12 e13337 10.1111/obr.13337 34476900 14 Onstad M A Schmandt R E Lu K H Addressing the role of obesity in endometrial cancer risk, prevention, and treatment J Clin Oncol 2016 34 35 4225 4230 10.1200/JCO.2016.69.4638 27903150 15 Erkanli S Kayaselçuk F Bagis T Kuşçu E Impact of morbid obesity in surgical management of endometrial cancer: surgical morbidity, clinical and pathological aspects Eur J Gynaecol Oncol 2006 27 04 401 404 17009636 16 Santoso J T Barton G Riedley-Malone S Wan J Y Obesity and perioperative outcomes in endometrial cancer surgery Arch Gynecol Obstet 2012 285 04 1139 1144 10.1007/s00404-011-2116-2 22020677 17 Bouwman F Smits A Lopes A Das N Pollard A Massuger L The impact of BMI on surgical complications and outcomes in endometrial cancer surgery—an institutional study and systematic review of the literature Gynecol Oncol 2015 139 02 369 376 10.1016/j.ygyno.2015.09.020 26407479
PMC009xxxxxx/PMC9800147.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36138536 10.1055/s-0042-1756214 rbgo-22-0136 Original Article Gynecological Endocrinology/Malformations Three-dimensional Printer Molds for Vaginal Agenesis: An Individualized Approach as Conservative Treatment Moldes de impressão tridimensionais para agenesia vaginal: Uma abordagem individualizada como tratamento conservadorhttp://orcid.org/0000-0003-0943-9573 Fernandes Marina Silva 1 http://orcid.org/0000-0001-5698-0245 Takano Claudia Cristina 1 http://orcid.org/0000-0003-2908-5215 Chrispin Thyeres Teixeira Bueno 1 http://orcid.org/0000-0003-4573-5361 Marquini Gisele Vissoci 1 http://orcid.org/0000-0002-1206-9377 Girão Manoel João Batista Castello 1 http://orcid.org/0000-0002-3001-6076 Sartori Marair Gracio Ferreira 1 1 Escola Paulista de Medicina, Universidade Federal de São Paulo, São Paulo, SP, Brazil Address for correspondence Gisele Vissoci Marquini, MD Escola Paulista de Medicina, Universidade Federal de São PauloRua Napoleão de Barros, 608, 04024-002, São Paulo, [email protected] 22 9 2022 12 2022 1 9 2022 44 12 11101116 22 4 2022 14 6 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  The aim of this study was to evaluate the use of vaginal molds, made with three-dimensional (3D) printing, for conservative treatment through vaginal dilation in patients with vaginal agenesis (VA). Methods  A total of 16 patients with a diagnosis of VA (Mayer-Rokitansky-Küster-Hauser syndrome, total androgen insensitivity syndrome, and cervicovaginal agenesis) from the Federal University of São Paulo were selected. Device production was performed in a 3D printer, and the polymeric filament of the lactic polyacid (PLA) was used as raw material. A personalized treatment was proposed and developed for each patient. Results  There were 14 patients who reached a final vaginal length of 6 cm or more. The initial total vaginal length (TVL) mean (SD) was 1.81(1.05) and the final TVL mean (SD) was 6.37 (0.94); the difference, analyzed as 95% confidence interval (95% CI) was 4.56 (5.27–3.84) and the effect size (95% CI) was 4.58 (2.88–6.28). Conclusion  The 3D printing molds for vaginal dilation were successful in 87.5% of the patients. They did not present any major adverse effects and offered an economical, accessible, and reproducible strategy for the treatment of VA. Resumo Objetivo  O objetivo deste estudo foi avaliar o uso de moldes dilatadores vaginais, confeccionados com impressão tridimensional (3D), para tratamento conservador através da dilatação vaginal em pacientes com agenesia vaginal (AV). Métodos  Foram selecionadas 16 pacientes com diagnóstico de AV (síndrome de Mayer-Rokitansky-Küster-Hauser, síndrome de insensibilidade androgênica total e agenesia cervicovaginal), da Universidade Federal de São Paulo. A produção dos dispositivos foi realizada em uma impressora 3D e, como matéria-prima, foi utilizado o filamento polimérico do poliácido lático (PLA). Um tratamento personalizado foi proposto e desenvolvido para cada paciente. Resultados  Quatorze pacientes atingiram um comprimento vaginal final (CVF) de 6 cm ou mais. A média inicial do CVF (DP) foi de 1,81 (1,05) e a média final do CVF (DP) 6,37 (0,94); a diferença (IC 95%) foi de 4,56 (5,27–3,84) e o tamanho do efeito (IC 95%) foi de 4,58 (2,88–6,28). Conclusão  Os moldes de impressão 3D para dilatação vaginal obtiveram sucesso em 87,5% das pacientes. Como impacto secundário, não apresentaram efeitos adversos importantes e ofereceram uma estratégia econômica, acessível e reprodutível para o tratamento da AV. Keywords 3D printing Mayer-Rokitansky-Küster-Hauser syndrome vaginal agenesis vaginal dilation Palavras-chave impressão 3D síndrome de Mayer-Rokitansky-Küster-Hauser agenesia vaginal dilatação vaginal ==== Body pmcIntroduction Vaginal agenesis (VA) is a congenital malformation and 90% of cases are associated with Mayer-Rokitansky-Küster-Hauser syndrome (MRKHS). The MRKHS has an incidence ranging from one case for every 4,000 to 5,000 female births, and is characterized by the congenital absence of the uterus and the upper ⅔ of the vagina. 1 Such patients have a female karyotype (46XX), function ovarian and normal sexual characters. Differential diagnosis should be made for patients with a transverse vaginal septum and an imperforate hymen. 1 2 According to the American College of Obstetricians and Gynecologists (ACOG), the treatment for VA can be conservative or surgical, and has the objective of restoring the vagina's anatomy and function. Vaginal dilation is the method used in the conservative treatment of VA. 3 4 5 This method was developed by Frank in 1938 and modified by Ingram in 1981, reaching an effectiveness of approximately 90%. 3 4 5 It is performed using progressively larger sized rigid vaginal molds until the proper vaginal length is achieved. While there is no consensus on the best therapeutic strategy, the ACOG and most of the scientific community recommend that the least invasive and effective treatment should be adopted, with vaginal dilation being the first line of treatment for VA. 3 4 5 Additionally, conservative treatment through vaginal dilation is the first choice due to the good results and low rate of complications. 5 Three-dimensional (3D) printing is gaining wide use in the health care field. Especially in gynecology, it is possible to manufacture various devices, such as pessaries and vaginal molds, using a wide variety of materials. 6 Using this technology, devices that are fit for purpose and cost-effective are created. The objective of this study was to evaluate the use of personalized vaginal molds made with 3D printing for conservative treatment through vaginal dilation in patients with VA. 1 Methods The study was performed at the Federal University of São Paulo (UNIFESP), between June 2017 and October 2019, after approval by the Human Research Ethics Committee of the same institution, under the Certificate of Presentation of Ethical Appreciation number (CAAE): 91233917–9 and opinion number 2970405. The interventions were only made after approval, with the aim of offering an individualized and conservative treatment for each patient who voluntarily proposed to participate in the study. The patients who were willing to participate in the study signed an informed consent form. The present study protocol was purely observational, which obviated the need for registration on clinical trial platforms. Additionally, it was not a randomized study or clinical trial, due to the rarity of the pathology, requiring patients to make a clear choice for conservative treatment before the beginning of the study. The patients were selected from the Female Genital Malformations Sector at UNIFESP according to the following inclusion criteria: confirmed diagnosis of VA due to SMRKH, androgen insensitivity syndrome (AIS), or cervicovaginal agenesis, and desire to undergo conservative treatment. The exclusion criterion was not wanting conservative treatment or not having free will or availability to participate in the research. All patients underwent evaluation by a multidisciplinary team: physician, physiotherapist, and psychologist. All patients were properly advised about the anatomy of the external genitalia before treatment. The study's protocol had three phases: prototype development, patient selection, and mold application. Characteristics, dimensions, and initial parameters were defined through research of devices already available on the market and adjusted for each patient. The initial geometric parameters were defined (cylindrical mold with the tip tapering progressively). For modeling the prototypes, the AutoCAD (Autodesk Inc., Mill Valley, California, USA) and FreeCAD 3D parametric modeling software were used based on the requirements defined above; computer aided design (CAD) system is the generic name for software used by engineering, geology, geography, health systems, and architecture and design to facilitate technical design and drawing. 6 The production of the devices was performed using the 3D cube printer, developed by the company 3D Systems (Rock Hill, SC, USA), using the polymeric filament of lactic polyacid (PLA) as raw material. The devices were evaluated by the medical staff for adjustments before application in the study. Three standard molds were created ( Fig. 1 ) with the following sizes (from right to left): Dilator A (1.5 × 8 cm); Dilator B (2 × 9 cm); and Dilator C (2.5 × 12 cm). 6 Fig. 1 Vaginal dilators created with 3D printing. Source : Marta Maria Kemp. The individual molds were reused only by the patient herself. The molds were not sterile, but the patients were instructed to clean them with soap and water and use them with condom protection. During the test period, the authors realized that the sizes described above were the most used. These 3 sizes were used because there was no need to use larger or smaller sizes at any time. Despite the biological plausibility of having no contraindication for use in the supine position, as long as the patient is well oriented, the present study prioritized the systematic methodology of the same position of introduction of the casts in the gynecological position. The applicability of the molds and the success of the vaginal dilation treatment were evaluated considering variables such as final total vaginal length (TVL), patient satisfaction, complication rate, and cost of mold production. The patients were instructed to perform light pressure exercises from the vaginal introitus, positioning as shown in Fig. 2 . The first return was within 15 days, and after that there was a monthly follow-up during the first year of the study. Patients who reached 6 cm or more in TVL were considered treated. After 6 cm of TVL, the patients were allowed to attempt sexual intercourse. Fig. 2 Positioning the dilator into vaginal introitus (purple). Source : Marta Maria Kemp. The orientation of the direction and strength of the perineal massage vectors was standardized as introduction movements into the vagina and circular movements, toward the sides of the vagina and posterior wall or toward the perineal region, to preserve possible urethral trauma in the anterior region, according to Fig. 3 , for 20 minutes during the study. The strength guidance was individualized, being gradual and progressive according to the sensitivity to pain or discomfort of each patient. Fig. 3 Orientation of perineal massage with the dilator. Source : Marta Maria Kemp. All patients were followed up every month from the beginning of the treatment, during the 1 st year, and every 2 months in the 2 nd year of the study, with free return as needed. During follow-up, the newly created vaginal canal was analyzed using a speculum and by digital vaginal exam. The aspect, amplitude, and length were observed. All patients were allowed to have sexual intercourse when vaginal length was greater than 6 cm. The clinical aspects of the vagina were subjectively analyzed by the first two independent authors, with no difference regarding the appearance of the neovagina, such as presence or absence of active bleeding, the color of the vaginal mucosa, and granulation tissue in all patients, in the period of 3, 6, 12 and 24 months, and after treatment. The patients remain under follow-up at the same service with possible long-term results in future studies. To compare the means of the initial and final TVL variable, the Student t- test for related samples was used. The D'Agostino normality test was performed for the assumptions of the analysis. To assess the magnitude of the difference, the effect size (d) with a 95% confidence interval (95% CI) was used. According to Cohen, 7 it was agreed that the values of d are considered small if (20 ≤ d < 50); medium if (50 ≤ d < 80); and large if (d ≥ 80). Results A total of 16 patients were treated with the 3D printing dilators between October 2017 and October 2019. The patients in the present study had no exposure to previous treatments. The mean age was 19 years (standard deviation, SD: 2.84), mean initial vaginal length of 1.8 cm, and 60% of the patients had an associated malformation. Furthermore, 14 patients (87,5%) achieved a TVL greater than 6 cm at the end of the evaluation ( Table 1 ). The median time taken to reach treatment TVL was 5.6 months. The 2 patients who did not achieve TVL (patient number 15 and 16–Table 1) had used the molds for only 2 and 3 months, respectively. Table 1 Initial and final TVL Patient Initial TVL (cm) Final TVL (cm) 1 2 7 2 4 6.5 3 1.5 7 4 1 7 5 2 7 6 4 8 7 1.5 6 8 2 6 9 1 6 10 0.5 6.5 11 1 6 12 1 6 13 1.5 7 14 1 7 15 2 4 16 3 5 Abbreviation: TLV, total vaginal length. As shown in Table 2 , there was a significant difference ( p  < 0.05) between the initial and final TVL measurements. The effect size was 4.58, reinforcing the great magnitude of this difference. Table 2 Statistical Analysis: Initial and Final TVL Initial TVL mean (SD) Final TVL mean (SD) Difference (95% CI) Effect size (95% CI) p -value* 1.81 (1.05) 6.37 (0.94) 4.56 (5.27–3.84) 4.58 (2.88–6.28) 0.0001 Abbreviations: CI, confidence interval; SD, standard deviation; TLV, total vaginal length. Note: * p -values < 0.05 were considered statistically significant. The patients did not report any adverse effects, such as pain, discomfort, or bleeding, as they were instructed to perform dilation according to their ability, respecting their limitations of pain or discomfort. The authors believe that the absence of complications may be related to the previous educational guidelines and follow-up. Only 2 patients had inadvertent dilation of the urethra at the beginning of the treatment. Both had a smaller vaginal introitus (shorter distance between the urethral meatus and the vaginal furcula). They were reoriented in relation to anatomy and perineal massage exercises. After that, both were able to reach the treatment TVL. The patients were not operated on later because they were satisfied with the conservative treatment's results. They remain under follow-up at the specialized outpatient clinic of the same service for future follow-ups. Fig. 4 shows the evolution analysis of the time of prothesis use by each patient. According to these results, there was no statistically significant difference between the time of use of the 14 patients that patients who adhered and were successful during the evolution of the treatment ( p  = 0.189). Fig. 4 Time of use of the prosthesis and the evolution of total vaginal length of each patient. Discussion In 1938, Frank 4 described the first conservative treatment for vaginal dilation using Pyrex (Corning Inc., Corning, NY, USA) tubes of gradually increasing sizes (0.8, 1.5, and 2.0 cm in diameter). This was used to force the mucous membrane into the vaginal introitus region. No incisions were required for this procedure. 4 The main criticism of this therapeutic modality is that it requires a special dedication from the patients, as the exercises with the dilators make it possible to create a vaginal canal that enables sexual intercourse. Maintaining a vaginal prosthesis is sometimes necessary to keep the vaginal canal patent, as well as performing exercises in the absence of regular sexual practice. Decades after Frank's first description, several studies reported favorable results using his method. 8 9 10 11 In 1981, Ingram 12 suggested that the failures in the technique used by Frank were due to tiredness of the hands and fingers during the procedure, the need to use the embarrassing position, and the inability to perform other productive activities during the procedure. 10 In an attempt to overcome these limitations, Ingram 12 proposed a modification of the original Frank method. In the Ingram method, the patient's weight is used to replace manual and digital effort. The specially designed bicycle seat bench was used to facilitate perineal mold pressure. 12 Additionally, corroborating the results of the successful experiment by Ingram, 12 Roberts et al. 13 reported a 91% success rate using the Ingram method in their study of 51 patients with MRKHS. When well advised and emotionally prepared, almost all patients (90–96%) will achieve a satisfactory anatomical and functional result with vaginal dilation. 5 A recent study performed at the same reference center of the present project compared, in terms of anatomical, functional, and sexual aspects, two types of treatment for women with VA: progressive dilation (using the Frank method) or surgical neovaginoplasty (using the modified Abbé-McIndoe technique with oxidized cellulose). According to that study, both treatments had satisfactory efficacy and positive outcomes regarding the analyzed aspects. These data reinforce the reliability of the results from the present study, which indicate that dilation treatment can remain the first-line therapy for VA. 14 The literature lacks more consistent and robust studies comparing the different surgical techniques with each other and with vaginal dilation. However, so far, no surgical technique has surpassed the success rate of nonsurgical treatment; a fact that, together with the benefit of being a safer technique, places vaginal dilation as the first line in the treatment of VA. 5 15 16 17 18 19 20 21 22 In this context, the present study suggests the development of a personalized conservative treatment for each patient through vaginal dilation with 3D molds. In 1984, Charles Hull 23 founded the world's first 3D printing company, with the use of production technologies such as Additive Manufacturing (AM) and Rapid Prototyping. The AM is used in the synthesis of a given physical object by adding layers to form a part based on data generated by CAD. These technologies are widely used to quickly prototype products and tools for commercial purposes. Over time, it has also been integrated into other areas, such as the health area, since these tools have enabled the assistance of health professionals in diagnosis, surgical planning, and synthesis of orthotics and prostheses for the rehabilitation of patients. 23 The use of 3D printing is gaining considerable acceptance in many medical fields, including surgery. The resulting tactile feedbacks significantly help the comprehension of anatomical details, especially the spatial relations between structures. Currently, an increasing number of applications have been successfully tested in many surgical disciplines, extending the range of possible uses to preoperative planning, counselling with patients, education of students and residents, surgical training, intraoperative navigation, and others. 23 24 In a recent systematic review, Barbosa et al. assess previous publications within 3D printing in human reproduction and gynecology. Based on the included studies, it was possible to design 3D models (uterus, ovaries, uterine cervix, and uterus with fibroids) that provided enriched information to improve presurgical planning, medical training, fertility-sparing surgery, patient comprehension of surgical procedures, and assisted reproduction applications. 25 Future expectations for 3D printing concern the reduction of manufacturing costs and time to further increase accessibility, as well as the development of novel techniques and suitable materials for biological structures, making it possible to recreate the architecture and functionality of real human organs and tissues. 23 24 The choice of devices made using a 3D printer was based on the possibility of offering an individualized treatment for each patient at a low cost and with a low rate of complications, in line with the plausibility of mold development already demonstrated in other areas, such as in Gynecology. The devices can be made with the most diverse formats and materials, which allows them to be adapted to the needs of each patient. Of the 16 patients treated, 14 reached the vaginal length considered for treatment, representing an 86% success rate. The only 2 patients who did not achieve a TVL of 6 cm or greater were still starting treatment (only 2 and 3 months ago). According to the results of this present study, as well as in the literature, 5 15 16 17 18 19 20 21 22 conservative treatment for VA remains an excellent choice, with good efficacy and few complications, through personalized vaginal molds made with 3D printing. These results highlight the good applicability of the devices, bringing a cost-effective and easily reproducible option for the treatment of VA, making this a promising and accessible tool. Therefore, it would be a fruitful option to facilitate its use over the country, train professionals to apply the treatment, and shorten the distances so that more patients could benefit from it, thus eliminating the bias of distance and regularity in performing the exercises. The main relevance of this research is the possibility of offering an individualized treatment option that is recommended in the scientific community with ethics, efficacy, and safety for a patient in an international reference center for the treatment of VA. Additionally, this is the first study to analyze reproductible 3D molds with conservative treatment and improve sexual function in women with VA. The homogeneous patient sample, standardized procedures, and prospective model are also strong points. Another positive impact of this project was the effective response to the guidance of perineal massage exercises with the cast in patients with smaller vaginal introitus. This strategy can be used before the beginning of the dilation itself, aiming to reduce the chance of inadvertent dilation of the urethra. The present study was not a randomized trial because of ethical issues. The main limitation of this trial was the small sample size. However, VA is a rare disease, with an incidence of 1:4.000 female births. The seriousness and scientific effort of this study are not diminished because of the difficulty to include more patients. The main difficulties encountered in this study were the lack of motivation, lack of privacy in the patients' home, and distance from the city of origin to the hospital. Another obstacle was the attendance to outpatient follow-ups, since many patients lived in different cities and some even in other districts, as well as the regularity in the performance of the exercises, which depended on the personal motivation of each patient, home privacy, and emotional situation during treatment. Considering the statements above, the authors believe that the study's strengths overcome its limitations. As VA is a rare disease that affects young women and involves the sensitive issues of sexuality and self-esteem, disclosure of well-structured trials can contribute to gaining knowledge so that an increasing number of women can benefit from the results of the studies. Conclusion Based on the present findings, a 3D model device can be offered in a personalized and individualized way as the first-line conservative treatment for VA in nonspecialized health centers in developed and developing countries. Furthermore, the use of 3D printing for making the molds proved to be a promising, effective and reproductible strategy, especially to be applied in health care centers with limited financial resources or a shortage of professionals specialized in the surgical treatment of VA, with low rate of complications. Considering the encouraging outcomes of this project and the rarity of the evaluated clinical condition affecting young women, the authors suggest more well-structured trials should be performed to better treat and benefit this population. Acknowledgments The authors would like to thank all patients who participated in this study; Marta Maria Kemp (gynecologist and urogynecologist) who designed the didactic figures; and the Board of the Urogynecology and Vaginal Surgery Sector of the Department of Gynecology from the Federal University of São Paulo for the inspiration of the values of medicine. This study was funded by the Brazilian National Research Council (CNPq) (code 001). No other funding was provided. Contributions Conflict of Interests The authors have no conflict of interests to declare. MSF contributed to project development, data collection, and writing the manuscript; CCTN contributed to project development and proofreading/editing the manuscript; TTBC and GVM proofread/ edited the manuscript; MJBCG contributed to project development and proofreading/editing the manuscript; MGFS contributed to project development and proofreading/editing the manuscript. ==== Refs References 1 Fontana L Gentilin B Fedele L Gervasini C Miozzo M Genetics of Mayer-Rokitansky-Küster-Hauser (MRKH) syndrome Clin Genet 2017 91 02 233 246 10.1111/cge.12883 27716927 2 Sarpel U Palmer S K Dolgin S E The incidence of complete androgen insensitivity in girls with inguinal hernias and assessment of screening by vaginal length measurement J Pediatr Surg 2005 40 01 133 136, discussion 136–13710.1016/j.jpedsurg.2004.09.012 15868573 3 McQuillan S K Grover S R Dilation and surgical management in vaginal agenesis: a systematic review Int Urogynecol J Pelvic Floor Dysfunct 2014 25 03 299 311 10.1007/s00192-013-2221-9 4 Frank R T The formation of an artificial vagina without operation Am J Obstet Gynecol 1938 35 06 1053 1055 10.1016/S0002-9378(38)90400-4 5 ACOG Committee Opinion No. 728 Summary: Müllerian Agenesis: Diagnosis, Management, And Treatment Obstet Gynecol 2018 131 01 196 197 10.1097/AOG.0000000000002452 29266072 6 Chrispin T T Fernandez M S Novoa C C Sartori M G Development of personalized molds for neovagina creation by 3D printer Rev Assoc Med Bras (1992). 2020 66 11 1498 502 10.1590/1806-9282.66.11.1498 33295399 7 Cohen J Statistical power analysis for the behavioral sciences 2nd ed. New York Routledge 1988 8 Edmonds D K Rose G L Lipton M G Quek J Mayer-Rokitansky-Küster-Hauser syndrome: a review of 245 consecutive cases managed by a multidisciplinary approach with vaginal dilators Fertil Steril 2012 97 03 686 690 10.1016/j.fertnstert.2011.12.038 22265001 9 Costa E M Mendonca B B Inácio M Arnhold I J Silva F A Lodovici O Management of ambiguous genitalia in pseudohermaphrodites: new perspectives on vaginal dilation Fertil Steril 1997 67 02 229 232 10.1016/s0015-0282(97)81902-4 9022594 10 Robson S Oliver G D Management of vaginal agenesis: review of 10 years practice at a tertiary referral centre Aust N Z J Obstet Gynaecol 2000 40 04 430 433 10.1111/j.1479-828x.2000.tb01175.x 11194430 11 Lappöhn R E Congenital absence of the vagina–results of conservative treatment Eur J Obstet Gynecol Reprod Biol 1995 59 02 183 186 10.1016/0028-2243(94)02037-f 7657013 12 Ingram J M The bicycle seat stool in the treatment of vaginal agenesis and stenosis: a preliminary report Am J Obstet Gynecol 1981 140 08 867 873 10.1016/0002-9378(81)90076-4 7270598 13 Roberts C P Haber M J Rock J A Vaginal creation for müllerian agenesis Am J Obstet Gynecol 2001 185 06 1349 1352, discussion 1352–135310.1067/mob.2001.119075 11744908 14 Apfel V R Takano C C Marquini G V de Jarmy di Bella Z IK Girão M JBC Sartori M GF Treatment for vaginal agenesis: A prospective and comparative study between vaginal dilation and surgical neovaginoplasty Int J Gynaecol Obstet 2022 157 03 574 581 10.1002/ijgo.13931 34534375 15 Baldwin J F XIV. The formation of an artificial vagina by intestinal transplantation Ann Surg 1904 40 03 398 403 17861520 16 Vecchietti G [Creation of an artificial vagina in Rokitansky-Küster-Hauser syndrome] Attual Ostet Ginecol 1965 11 02 131 147Italian.5319813 17 Davydov S N Zhvitiashvili O D Formation of vagina (colpopoiesis) from peritoneum of Douglas pouch Acta Chir Plast 1974 16 01 35 41 4141200 18 McIndoe A H Banister J B An operation for the cure of congenital absence of the vagina J Obstet Gynaecol Br Emp 1938 45 490 494 19 McIndoe A H The treatment of vaginal atresia by free grafts Baltimore Williams & Wilkins 1957 414 7 20 Crema L C Neovaginoplastia com celulose oxidada aspectos clínicos e morfológicos [dissertação] São Paulo Universidade Federal de São Paulo 2008 21 Dornelas J Jármy-Di Bella Z I Heinke T Kajikawa M M Takano C C Zucchi E V Vaginoplasty with oxidized cellulose: anatomical, functional and histological evaluation Eur J Obstet Gynecol Reprod Biol 2012 163 02 204 209 10.1016/j.ejogrb.2012.04.018 22739655 22 Kajikawa M M Jármy-Di Bella Z I Focchi G R Dornelas J Girão M J Sartori M G Role of estrogen receptor alpha on vaginal epithelialization of patients with Mayer-Rokitansky-Kuster-Hauser syndrome submitted to neovaginoplasty using oxidized regenerated cellulose Int Urogynecol J Pelvic Floor Dysfunct 2012 23 04 467 472 10.1007/s00192-011-1604-z 23 National Inventors Hall of Fame Charles Hull: Stereolithography (3D Printing): Patent No. 4,575,330 [Internet] 2014[cited 2021 Oct 18]. Available from:https://web.archive.org/web/20140312010450/http://invent.org/inductees/hull-charles/ 24 Pugliese L Marconi S Negrello E Mauri V Peri A Gallo V The clinical use of 3D printing in surgery Updates Surg 2018 70 03 381 388 10.1007/s13304-018-0586-5 30167991 25 Barbosa M Z Zylbersztejn D S de Mattos L A Carvalho L F Three-dimensionally-printed models in reproductive surgery: systematic review and clinical applications Minerva Ginecol 2019 71 03 235 244 10.23736/S0026-4784.19.04319-3 30756546
PMC009xxxxxx/PMC9800148.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36580936 10.1055/s-0042-1758712 RBGO-22-0191 Original Article | Artigo Original High Risk Prenancy Relationship between the Prenatal Diagnosis of Placenta Acreta Spectrum and Lower Use of Blood Components Relação entre o diagnóstico prenatal de espectro da placenta acreta e menor uso de hemoderivadoshttp://orcid.org/0000-0003-4590-0945 Pavón-Gomez Néstor 12 http://orcid.org/0000-0003-1500-4920 López Rita 12 http://orcid.org/0000-0003-3437-2446 Altamirano Luis 2 http://orcid.org/0000-0002-5838-4612 Cabrera Sugey Bravo 1 http://orcid.org/0000-0002-1485-2277 Rosales Gusmara Porras 1 http://orcid.org/0000-0001-5874-4262 Chamorro Sergio 12 http://orcid.org/0000-0002-9777-0153 González Karen 1 http://orcid.org/0000-0002-2247-9994 Morales Amparo 1 http://orcid.org/0000-0002-6467-160X Maya Juliana 3 http://orcid.org/0000-0001-6971-2098 Sinisterra Stiven 4 http://orcid.org/0000-0001-5639-9127 Nieto-Calvache Albaro José 25 1 Hospital Bertha Calderón Roque, Managua, Nicaragua 2 Latin American Group for the Study of Placenta Accreta Spectrum, Cali, Colombia 3 Programa de Medicina, Facultad de Ciencias de la Salud, Universidad Icesi, Cali, Colombia 4 Centro de Investigaciones Clínicas, Fundación Valle del Lili, Cali, Colombia 5 Clínica de Espectro de Acretismo Placentario, Fundación Valle del Lili, Cali, Colombia Address for correspondence Néstor Pavón-Gómez Hospital Bertha Calderón [email protected] 29 12 2022 12 2022 1 12 2022 44 12 10901093 13 6 2022 05 8 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  To describe the clinical results of patients admitted and managed as cases of placenta accreta spectrum (PAS) at a Central American public hospital and the influence of the prenatal diagnosis on the condition. Materials and Methods  A retrospective analysis of PAS patients treated at Hospital Bertha Calderón Roque, in Managua, Nicaragua, between June 2017 and September 2021. The diagnostic criteria used were those of the International Federation of Gynecology and Obstetrics (Fédération Internationale de Gynécologie et d'Obstétrique, FIGO, in French). The population was divided into patients with a prenatal ultrasonographic diagnosis of PAS (group 1) and those whose the diagnosis of PAS was established at the time of the caesarean section (group 2). Results:  During the search, we found 103 cases with a histological and/or clinical diagnosis of PAS; groups 1 and 2 were composed of 51 and 52 patients respectively. Regarding the clinical results of both groups, the patients in group 1 presented a lower frequency of transfusions (56.9% versus 96.1% in group 2), use of a lower number of red blood cell units (RBCUs) among those undergoing transfusions (median: 1; interquartile range: [IQR]: 0–4 versus median: 3; [IQR]: 2–4] in group 2), and lower frequency of 4 or more RBCU transfusions (29.4% versus 46.1% in group 2). Group 1 also exhibited a non-significant trend toward a lower volume of blood loss (1,000 mL [IQR]: 750–2,000 mL versus 1,500 mL [IQR]: 1,200–1,800 mL in group 2), and lower requirement of pelvic packing (1.9% versus 7.7% in group 2). Conclusion  Establishing a prenatal diagnosis of PAS is related to a lower frequency of transfusions. We observed a high frequency of prenatal diagnostic failures of PAS. It is a priority to improve prenatal detection of this disease. Keywords placenta accreta prenatal ultrasonographic diagnosis surgical procedure blood transfusion ==== Body pmcIntroduction Placenta accreta spectrum (PAS) is a condition associated to massive hemorrhage and polytransfusion, 1 and patients should be cared for by interdisciplinary groups in experienced centers. 2 3 4 However, the participation of these expert groups relies on a prenatal diagnosis that enables the patient to be guided towards this type of care. The frequency of cases of PAS not diagnosed before laparotomy is variable, but it can be as high as 50%. 5 6 In Nicaragua, among the factors that contribute to the low rate of prenatal diagnoses are the difficulties in training to identify PAS, the absence of centers with a high influx of patients, and the lack of feedback between the centers that carry out the diagnosis and those who deliver treatment. The present work describes the clinical results of the PAS patients managed at a Central American public hospital and the importance of establishing a prenatal diagnosis. Materials and Methods A retrospective analysis of medical records was carried out in search for patients with PAS treated at Hospital Bertha Calderón Roque, in Managua, Nicaragua, between June 2017 and December 2021. The diagnostic criteria used was those of the International Federation of Gynecology and Obstetrics (Fédération Internationale de Gynécologie et d'Obstétrique, FIGO, in French). 7 The population was divided into patients with a prenatal PAS diagnosis by ultrasound submitted to surgery for that reason (group 1) and patients in whom PAS was only detected at the time of the caesarean section (group 2). During this period, the management protocol was standard, with no variations. All patients with a diagnosis of PAS underwent cesarean section at 35 weeks, with a plan for total hysterectomy after extraction of the fetus through the uterine fundus. Specific vascular control strategies nor ureteral catheters were used. All patients included had placenta previa and underwent cesarean hysterectomy, same surgical technique was applied. The present retrospective study was approved by the Institutional Review Board/Ethics Committee for Biomedical Research (under no. 1494). A descriptive statistical analysis was carried out; The continuous variables were expressed as median and interquartile range (IQR) values, and they were analyzed using the Mann-Whitney U test. The qualitative variables were expressed as absolute and relative frequencies, and the comparison between them was made using the Chi-squared test or the Fisher exact test according to the case. Statistical significance was defined as p  < 0.05. The analyses were performed using the STATA (StataCorp LLC, College Station, TX, United States) software, version 14. Results During the study period, 114 women with a histological and/or clinical diagnosis of PAS were found: 51 patients had a prenatal PAS diagnosis by ultrasound (group 1), and 63 were only diagnosed when they underwent laparotomy (group 2). Chart 1 summarizes the clinical results of both groups, showing a lower frequency of transfusions in group 1 (56.9% versus 87.3% in group 2), as well as the use of a lower number of red blood cell units (RBCUs) in said transfused patients (median: 1; IQR: 0–4 versus median: 3; IQR: 2–4 in group 2). The frequency of 4 or more RBCU transfusions was also lower in group 1 (29.4% versus 44.4% in group 2). Group 2 underwent surgery at a higher gestational age (mean: 38 weeks; IQR: 35–39 weeks versus median: 34 weeks; IQR: 32–36 weeks in group 1), with a lower rate of participation of interdisciplinary groups (62.4% versus 90.2% in group 1) and of elective surgeries than group 1 (22.2% in group 2 versus 78.4% in group 1). Group 1 also exhibited a non-significant trend toward a lower volume of blood loss (median: 1,000 mL; IQR: 750–2,000 mL versus median: 1,500 mL; IQR: 1,300–2,200 mL in group 2), lower requirement of pelvic packing with compresses to control bleeding (1.9% versus 7.9% in group 2), surgical reinterventions (11.8% versus 17.5% in group 2), and surgical site infection (1.9% versus 4.8% in group 2) than group 2. In a high percentage of patients (35; 30.9%), the histological diagnosis was not available because the tissue had not been not processed by the pathology department. Chart 1 Comparison of the clinical results of PAS patients with and without a prenatal diagnosis Group 1 (n = 51): WITH prenatal diagnosis Group 2 (n = 63): WITHOUT prenatal diagnosis p -value Gestational age at surgery (in weeks)* 34 (32-36) 38 (35-39) 0.003 Surgical time (in minutes)* 103 (82-145) 94 (74-129) 0.235 Interdisciplinary group participation: n (%) 46 (90.2) 33 (62.4) 0.003 Elective surgery: n (%) 40 (78.4) 14 (22.2) 0.002 Bleeding volume (mL)* 1000 (750-2000) 1500 (1300-2200) 0.347 Transfusions: n (%) 29 (56.9) 55 (87.3) 0.005 Number of RBCUs transfused 1 (0-4) 3 (2-4) 0.027 4 or more RBCUs: n (%) 15 (29.4) 28 (44.4) 0.039 Bladder injury: n (%) 7 (13.7) 9 (14.2) 0.923 Ureteral injury: n (%) 2 (3.9) 0 − Urinary fistula: n (%) 0 1 (1.6) − Pelvic packing with compresses: n (%) 1 (1.9) 5 (7.9) 0.380 Reintervention: n (%) 6 (11.8) 11 (17.5) 0.996 Wound infection: n (%) 1 (1.9) 3 (4.8) 0.666 Death: n (%) 1 (1.9) 1 (1.6) 0.104 Histological analysis: n (%) Placenta acreta 24 (47.1) 34 (53.9) 0.890 Placenta increta 11 (21.6) 6 (9.5) Placenta percreta 2 (3.9) 2 (3.2) No histological study 14 (27.4) 21 (33.3) Abbreviations: PAS, placenta accreta spectrum; RBCU, red blood cells unit. Note: *Median (interquartile range). Discussion Less than half of our cases (44.7%) had a prenatal PAS diagnosis (group 1). Patients in group 1 had a lower frequency of transfusions and those among them who received blood components required a lower number of RBCUs. Although some expert groups have reported excellent performance of the PAS ultrasonographic diagnosis, 8 9 even in some high-income countries the frequency of false positives is close to that observed in the Nicaraguan population, with a rate of intraoperative diagnosis close to 50%. 5 6 There are multiple factors that explain a poor performance in establishing a PAS prenatal diagnosis in our population. Although all the patients included in the present study underwent prenatal follow-up visits and periodic ultrasonographic scans, Nicaragua has not established protocols to diagnose PAS. Additionally, there are few maternal-fetal medicine specialists or prenatal ultrasonography experts. Finally, there is no chair in the diagnosis and treatment of PAS in the obstetrics training programs in our country. It is important to point out these difficulties as the first step towards improving the prenatal identification of PAS. It is likely that the knowledge of a prenatal PAS diagnosis in group 1 facilitated the scheduling of the surgical procedure, which was elective in 78.4% of these patients, unlike group 2, in which it was elective in 22.2% of the cases, and at a higher gestational age (median: 34 weeks; IQR: 32–36 weeks in group 1 versus median: 38 weeks; IQR: 35–39 weeks in group 2). One of the advantages of knowing the PAS diagnosis is the possibility of “scheduling” the participation of the interdisciplinary groups during surgery. 10 11 12 Our hospital does not have an interdisciplinary group dedicated to the treatment of PAS (a “PAS team”); however, patients from group 1 were treated by the more experienced surgeons available, which included the urologist and the teneral surgeon on duty that day. This was possible in 90.2% of the cases in group 1, and only in 62.4% of the cases in group 2. In the event that the diagnosis of PAS was a “surprise” during the laparotomy, calling the surgeon and the urologist on duty was left at the discretion of the obstetricians in charge of the surgery. Other authors 5 have pointed out the importance of prenatal diagnosis and its relationship with lower levels of blood loss and lower use of transfusions, and they also coincide in documenting important differences in the management of patients with and without a prenatal diagnosis. Our hospital is a reference center for the most critical obstetric conditions in the country, but like many other hospitals with similar characteristics, it does not have a PAS team. 13 Our flaws in the prenatal diagnosis (intraoperative finding of PAS in 55.2% of our cases) and histological analysis (absence of analysis by a pathologist in 35 cases) result in an opportunity to improve the quality of care in our center. The present study has limitations. Its retrospective design makes it more susceptible to bias. The absence of histological confirmation in 30.7% of the cases enabled the inclusion of non-PAS cases; however, patients whose medical record described macroscopic findings compatible with the FIGO definition were included. The present is the first evaluation of the clinical results of PAS management in Nicaragua, and one of the few that have been carried out in Central America. The results shed light on the need to design improvement plans at the national level, with the need for multicenter prospective studies to confirm our observations and evaluate the effect of interventions already implemented, such as the creation of a PAS team that is provided with specific training in the management of this disease, the proposal of including PAS screening in routine prenatal care appointments for women with risk factors, and the project of including a section about PAS in the national guidelines for the treatment of postpartum hemorrhage. Conclusion The presence of a prenatal diagnosis of PAS is related to a lower frequency of RBCU transfusions. We observed a high frequency of failures within the prenatal PAS diagnostic steps. It is a priority to improve the prenatal detection of this disease. Contributions Conflict of Interests The authors have no conflict of interests to declare. All authors made substantial contributions to the conception and design, data collection or analysis, and interpretation of data, writing of the article or critical review of the intellectual content, and final approval of the version to be published. ==== Refs References 1 DeSimone R A Leung W K Schwartz J Transfusion medicine in a multidisciplinary approach to morbidly adherent placenta: preparing for and preventing the worst Transfus Med Rev 2018 32 04 244 248 10.1016/j.tmrv.2018.05.007 30025661 2 Silver R M Fox K A Barton J R Center of excellence for placenta accreta Am J Obstet Gynecol 2015 212 05 561 568 10.1016/j.ajog.2014.11.018 25460838 3 International Society for Abnormally Invasive Placenta (IS-AIP) Collins S L Alemdar B van Beekhuizen H J Evidence-based guidelines for the management of abnormally invasive placenta: recommendations from the International Society for Abnormally Invasive Placenta Am J Obstet Gynecol 2019 220 06 511 526 10.1016/j.ajog.2019.02.054 30849356 4 FIGO Placenta Accreta Diagnosis and Management Expert Consensus Panel Allen L Jauniaux E Hobson S Papillon-Smith J Belfort M A FIGO consensus guidelines on placenta accreta spectrum disorders: Nonconservative surgical management Int J Gynaecol Obstet 2018 140 03 281 290 10.1002/ijgo.12409 29405317 5 Fitzpatrick K E Sellers S Spark P Kurinczuk J J Brocklehurst P Knight M The management and outcomes of placenta accreta, increta, and percreta in the UK: a population-based descriptive study BJOG 2014 121 01 62 70, discussion 70–71. Doi: 10.1111/1471-0528.12405 6 Silveira C Kirby A Melov S J Nayyar R Placenta accreta spectrum: We can do better Aust N Z J Obstet Gynaecol 2022 62 03 376 382 10.1111/ajo.13471 34984671 7 FIGO Placenta Accreta Diagnosis and Management Expert Consensus Panel Jauniaux E Ayres-de-Campos D Langhoff-Roos J Fox K A Collins S FIGO classification for the clinical diagnosis of placenta accreta spectrum disorders Int J Gynaecol Obstet 2019 146 01 20 24 10.1002/ijgo.12761 8 Cali G Forlani F Timor-Trisch I Diagnostic accuracy of ultrasound in detecting the depth of invasion in women at risk of abnormally invasive placenta: A prospective longitudinal study Acta Obstet Gynecol Scand 2018 97 10 1219 1227 10.1111/aogs.13389 29797715 9 Jauniaux E Bhide A Prenatal ultrasound diagnosis and outcome of placenta previa accreta after cesarean delivery: a systematic review and meta-analysis Am J Obstet Gynecol 2017 217 01 27 36 10.1016/j.ajog.2017.02.050 28268196 10 Chantraine F Braun T Gonser M Henrich W Tutschek B Prenatal diagnosis of abnormally invasive placenta reduces maternal peripartum hemorrhage and morbidity Acta Obstet Gynecol Scand 2013 92 04 439 444 10.1111/aogs.12081 23331024 11 Weiniger C F Einav S Deutsch L Ginosar Y Ezra Y Eid L Outcomes of prospectively-collected consecutive cases of antenatal-suspected placenta accreta Int J Obstet Anesth 2013 22 04 273 279 10.1016/j.ijoa.2013.04.014 23932233 12 Hall T Wax J R Lucas F L Cartin A Jones M Pinette M G Prenatal sonographic diagnosis of placenta accreta–impact on maternal and neonatal outcomes J Clin Ultrasound 2014 42 08 449 455 10.1002/jcu.22186 24975386 13 Nieto-Calvache A J Palacios-Jaraquemada J M Hidalgo A Management practices for placenta accreta spectrum patients: a Latin American hospital survey J Matern Fetal Neonatal Med 2021 ••• 1 8; [ahead of print].10.1080/14767058.2021.1906858
PMC009xxxxxx/PMC9800149.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36580940 10.1055/s-0042-1760215 RBGO-22-0201 Review Article | Artigo de Revisão Maternal Mercury Exposure and Hypertensive Disorders of Pregnancy: A Systematic Review Exposição materna a mercúrio e distúrbios hipertensivos na gestação: Uma revisão sistemáticahttp://orcid.org/0000-0002-8007-0142 Dantas Aline de Oliveira 1 http://orcid.org/0000-0002-8528-4254 Castro Thiania dos Santos da Silva de 1 http://orcid.org/0000-0002-6596-6653 Câmara Volney de Magalhães 1 http://orcid.org/0000-0002-5498-3992 Santos Aline de Souza Espindola 1 http://orcid.org/0000-0002-9864-6656 Asmus Carmen Ildes Rodrigues Froes 1 http://orcid.org/0000-0003-0657-2141 Vianna Angelica dos Santos 1 1 Faculdade de Medicina, Universidade Federal do Rio de Janeiro, Rio de Janeiro, RJ, Brazil Address for correspondence Aline de Oliveira Dantas, medical student Universidade Federal do Rio de JaneiroAv. Carlos Chagas Filho, 373, 2° andar, Sala 49, 21044-020, Cidade Universitária, Rio de Janeiro, [email protected] 29 12 2022 12 2022 1 12 2022 44 12 11261133 27 6 2022 26 9 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  The present review aimed to synthesize the evidence regarding mercury (Hg) exposure and hypertensive disorders of pregnancy (HDP). Data Sources  The PubMed, BVS/LILACS, SciELO and UFRJ's Pantheon Digital Library databases were systematically searched through June 2021. Study Selection  Observational analytical articles, written in English, Spanish, or Portuguese, without time restriction. Data Collection  We followed the PICOS strategy, and the methodological quality was assessed using the Downs and Black checklist. Data Synthesis  We retrieved 77 articles, of which 6 met the review criteria. They comprised 4,848 participants, of which 809 (16.7%) had HDP and 4,724 (97.4%) were environmentally exposed to Hg (fish consumption and dental amalgam). Mercury biomarkers evaluated were blood (four studies) and urine (two studies). Two studies found a positive association between Hg and HDP in the group with more exposure, and the other four did not present it. The quality assessment revealed three satisfactory and three good-rated studies (mean: 19.3 ± 1.6 out 28 points). The absence or no proper adjustment for negative confounding factor, such as fish consumption, was observed in five studies. Conclusion  We retrieved only six studies, although Hg is a widespread toxic metal and pregnancy is a period of heightened susceptibility to environmental threats and cardiovascular risk. Overall, our review showed mixed results, with two studies reporting a positive association in the group with more exposure. However, due to the importance of the subject, additional studies are needed to elucidate the effects of Hg on HDP, with particular attention to adjusting negative confounding. Resumo Objetivo  A presente revisão busca sintetizar as evidências em relação à exposição ao mercúrio (Hg) e os distúrbios hipertensivos da gestação (DHG). Fontes Dos Dados Os bancos de dados PubMed, BVS/LILACS, SciELO e a Biblioteca Digital da UFRJ Pantheon foram sistematicamente pesquisadas durante junho de 2021. Seleção de estudos  Artigos observacionais analíticos, escritos em inglês, espanhol ou português, sem restrição temporal. Coleta de Dados  A estratégia PICOS foi seguida e a qualidade metodológica foi avaliada usando o checklist Downs and Black. Síntese de dados  Foram encontrados 77 artigos, dos quais 6 atenderam aos critérios da revisão. Foram 4.848 participantes, dos quais 80 (16,7%) tinham DHG e 4.724 (97,4%) estavam expostos ambientalmente ao Hg (consumo de peixe e amálgama dental). Os biomarcadores de mercúrio avaliados foram sangue (quatro estudos) e urina (dois estudos). Dois estudos encontraram associação positiva entre Hg e DHG no grupo com maior exposição e os outros quatro não a apresentaram. A avaliação de qualidade metodológica revelou 3 estudos satisfatórios e 3 bons (média: 19,3 ± 1,6 em 28 pontos). A ausência ou não de ajuste adequado para fator de confusão negativo, como consumo de pescado, foi observada em cinco estudos. Conclusão  Recuperamos apenas seis estudos, embora o Hg seja um metal tóxico generalizado e a gravidez seja um período de maior suscetibilidade a ameaças ambientais e risco cardiovascular. No geral, nossa revisão mostrou resultados mistos, com dois estudos relatando associação positiva no grupo com maior exposição. No entanto, devido à importância do assunto, estudos adicionais são necessários para elucidar os efeitos do Hg sobre DHG, com atenção especial ao ajuste de confundimento negativo. Keywords mercury pregnancy-induced hypertension preeclampsia eclampsia and gestational hypertension Palavras-chave mercúrio hipertensão induzida pela gestação pré-eclâmpsia eclâmpsia e hipertensão gestacional ==== Body pmcIntroduction Systemic arterial hypertension (SAH) is a highly prevalent health issue worldwide, leading to significant morbidity and costs for health systems. 1 It is equally an important public health issue during pregnancy and deserves special attention since it is one of the leading causes of maternal and perinatal mortality worldwide. 2 Besides, the traditional risk factors for SAH, including overweight/obesity, age > 60 years old, daily ingestion of sodium > 2 g, and sedentarism, multifetal pregnancy, primigravid women, and multiparas > 35 years old are additional factors for hypertensive disorders of pregnancy (HDP). 1 3 4 Environmental exposure to heavy metals, such as mercury (Hg), have been associated with adverse cardiovascular effects, including changes in blood pressure levels. 5 6 7 8 9 10 Although the mechanisms by which Hg may induce hypertension are not yet fully elucidated, some evidence points to an increase in angiotensin-converting enzyme activity, stimulation of the proliferation of vascular smooth muscle cells, induction of renal dysfunction, and an imbalance of the redox system, with an increase in oxidative stress and consequent reduction in nitric oxide bioavailability, endothelial dysfunction, and decreased smooth muscle relaxation. 6 8 11 Also, Hg can accumulate in the placenta tissue and leads to its dysfunction. 9 Mercury is a ubiquitous environmental toxic substance with adverse results for health. 10 There are three distinct forms of Hg: elemental mercury (Hg 0 ), inorganic mercury (IHg), and organic mercury (ethylmercury [ethylHg], methylmercury [MeHg]). Its main sources of exposure include gold mining, Chlor-alkali industry, biomass burning, and deforestation, dentist activities (Hg 0 ), presence of dental amalgams, skin cosmetics use (IHg), vaccines conservative (ethylHg), and fish and shellfish intake (MeHg). 12 13 14 15 The association between Hg exposure and hypertension has produced inconsistent findings. 16 Differences in study populations, and exposure levels, different Hg species, Hg biomarkers used to assess the exposure and absence of proper adjustment for confounding factors may contribute to the discrepancies observed in studies. 8 Considering the widespread distribution of Hg, the great impact of HDP on public health, and the controversial evidence about their association, the present systematic review aimed to address this topic. Methods We followed the guidelines of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) to conduct and report the present review. 17 In addition, the study protocol was submitted to the International Prospective Register of Systematic Reviews (PROSPERO), approved under number CRD42022297367. A search strategy was developed in three electronic databases (BVS/LILACS, PubMed/Medline, and SciELO) and one Digital Library of Theses and Dissertations (Pantheon – Universidade Federal do Rio de Janeiro) in June 2021. We used various combinations of MeSH descriptors associated with the text words: mercury AND hypertension, pregnancy-induced hypertensive disorders of pregnancy OR preeclampsia OR eclampsia OR gestational hypertension . Articles were considered for inclusion based on the PICOS strategy, as follows: P articipants comprised pregnant or puerperal women; I ntervention included assessment of Hg exposure through its measurement in any biological matrix; C omparison with normotensive pregnant or puerperal women and documented Hg measurement in any biological matrix; O utcome comprised gestational hypertension syndromes with their criteria reported by the authors. S tudy: original observational analytical article, written in Spanish, English, or Portuguese, without time restriction. We excluded any article without Hg exposure assessment in a biological matrix, without the criteria used to classify HDP, editorial articles, author's opinions, books, case reports, experimental studies (animal and in vitro); and reviews. The PubMed database was the reference database for cases of duplicate articles. Two reviewers (Dantas A. O. and Castro T. S. S.) independently assessed the entire study selection process. Any disagreements about study selection were resolved by discussion and, if necessary, a third reviewer (Vianna A. S.) was consulted. The flowchart started by analyzing the titles, followed by the abstract, and later by the full text. Finally, we checked the reference lists of eligible papers to identify additional relevant studies. One reviewer (Dantas A. O.) extracted the data from the eligible studies using a form that included: 1. Study characteristics: name of the first author, year of publication, country of study; 2. Methods: design, sample size, and exposure site; 3. gestational hypertension (GH) cases: number of cases, age, ethnicity; 4. Hg exposure: source, biological matrix, laboratory technique; 5. Statistical analysis including parametric (Student t -test and analysis of variance [ANOVA]) and nonparametric tests (Mann-Whitney and Kruskal-Wallis) for comparison (mean difference), regression tests for measure of association (risk ratio, odds ratio [OR] and/or hazard ratio), and prevalence ratio; 6. Methodological quality score. Another reviewer (Vianna A. S.) checked this step. Two reviewers (Dantas A. O. and Castro T. S. S.) independently assessed the quality of each eligible study according to the Downs and Black (DB) checklist. It contains 27 items, subdivided into 5 sub-scales, which assess reporting (9 items), external validity (3 items), internal validity (bias and confounding – 13 items), and power (1 item). The 25-item score is: yes = 1, no = 0 and unable to determine = 0. Item 5 (distribution of main confounding variables) presents the score: yes = 2; partially = 1 and not = 2. 18 Item 27 (power) was modified, scoring yes or no for the power calculation. We adopted the categorization of quality proposed by Hooper et al.: excellent (26–28), good (20–25), satisfactory (15–19), or poor (≤14). 19 Although we had originally planned to perform a quantitative meta-analysis, we considered it inappropriate due to methodological limitations of the selected articles and to the high heterogeneity in exposure assessment with different cutoffs. Therefore, we reported the findings as a systematic qualitative review. Results The present systematic review retrieved 77 potentially eligible studies. Of these, 6 met our inclusion criteria, 4 from the electronic databases and 2 from the manual reference consultation, published between 2006 and 2020. The main reason for exclusion was out of scope, comprising 42 articles (30 without Hg and 12 without pregnant women). A flowchart of the search and screening process is displayed in Fig. 1 . Fig. 1 PRISMA flowchart of the study selection process. Abbreviations: DLTD, Digital Library of Theses and Dissertations; BVS/LILACS, Biblioteca Virtual em Saúde/Literatura Latino-Americana e do Caribe em Ciências da Saúde; SciELO, Scientific Library Online; UFRJ, Universidade Federal do Rio de Janeiro The six studies had the following design: three were cohorts and three were case control. They covered 4,848 participants from 5 countries, 3 conducted in North America (2 in the USA and 1 in Canada), 2 in Asia, and 1 in North Africa. Four studies comprised 4,724 participants (97.4%) involved primarily in environmental exposure to Hg, and 1 study with 124 participants (2.6%) had both environmental and occupational exposure. 20 21 22 23 24 25 Out of 4,848 participants, 4,039 were controls (2,514 pregnant women and 1,525 postpartum), and 809 (16.7%) had a HDP diagnosis, comprising 187 (23.1%) GH, and 622 (76.9%) preeclampsia (PE). The participants had the following characteristics: age ranging between 15 and 49 years old, 406 (8.4%) were smokers, and 2.685 (55.4%) reported their ethnicity, with 1,794 (66.8%) white individuals. Regarding the source of exposure, half of the studies reported it as follows: amalgam use (64 dentists), presence of dental amalgam (905 participants) and fish intake (1,817 individuals). 23 25 Concerning the latter, one study (1,817 participants) reported the frequency, but not the type of fish consumed. 23 Four studies measured Hg concentrations in blood (whole maternal blood [three], umbilical cord blood [one], and/or red blood cell [one]), and two in urine. The laboratory method more frequently used was inductively coupled plasma mass spectrometry (ICP-MS). 20 21 22 23 24 The detection limit was described in three studies, ranging from 0.12 to 0.33 μg/l (total Hg). No study investigated the association with hypertension according to the type of Hg. In addition, four studies measured other toxicants (metals) during the research. 20 21 22 24 Two studies investigated the association between the metal mixture and HDP (PE). 21 24 Statistical analysis of studies included mean difference (two studies no and the other two yes), and measures of association. 20 22 23 25 Concerning the latter, two studies reported positive association (more exposed group [Hg urine  = 41.8μg/g]: RR = 3.67; 95%CI = 1.25–10.76 and more exposed group [Hg blood ≥1.89μg/L]: aOR multi-metal = 1.60; 95%CI = 1.08–2.38; p  = 0.039), and the other four studies found no association (One unit increase p > 0.05; HR single model = 0.90; 95%CI = 0.63–1.28; HR (As, Hg and Sn) = 0.75; 95%CI = 0.39–1.46; Prevalence ratio = 1.03; 95%CI = 0.88–1.20; p  = 0.71; 1 to 4 dental amalgams group: aOR = 1.31; 95%CI = 0.92–1.85 or ≥ 5 dental amalgams group - aOR = 1.32; 95%CI = 0.86, 2.04). 20 21 22 23 24 25 Regarding the latter, although the authors did not observe any association with HDP, they reported an inverse association with systolic blood pressure (dental amalgam replacement group: β = - 1.58; 95%CI = - 2.95–- 0.02; p  = 0.02). 23 The characteristics of all six studies are summarized in Chart 1 . Chart 1 Characteristics of the selected studies Author year country Study characteristics design; number Characteristics of the participants Hg assessment Outcome Mean difference/ measure of association Age (Mean: years old) Ethnicity Site of exposure (source) Biological matrix (laboratory technique) Mercury level Vigeh et al. (2006), 20 Iran Case-control; n  = 396 27 396 NI Environ Maternal blood and UBC (ICP-MS) Maternal blood: PE: 1.35 (0.74); Control: 1.34 (1.19) UBC: PE: 1.69 (1.19) Control: 1.70 (1.33) PE Mean difference: no El-Badry et al. (2018), 25 Egypt Cohort; n  = 124 Ex:25.6 NEx: 25.9 124 NI Environ (fish) Occupat (dentist) Urine (CVAAS) Urine 3rd quarter: Ex: 42.8 (13.7); NEx: 7.1 (3.9) PE Mean difference: yes; RR Ex  = 3.67 (1.25–10.76) Bommarito et al. (2019), 21 USA Case-control; n  = 383 32.7 231 white 57 black 95 NI Environ Urine (ICP-MS) Urine 3rd quarter: PE: 0.50 (0.24, 0.76); Control: 0.51 (0.27, 0.97) PE HR unimetal = 0.90 (0.63 - 1.28; p  = 0.55); HR multimetal As, Hg, Sn = 0.75 (0.39–1.46; p  = 0.40) Liu et al. (2019), 22 USA Cohort; n  = 1.274 27.99 739 black 535 NI Environ Maternal red blood cells (ICP-MS) Maternal blood: PE: 2.1 (1.0–4.7); Control: 2.0 (1.0–3.6) PE Mean difference: no; Prevalence ratio = 1,03 (0.88–1.20; p  = 0.71) Louopou et al. (2020), 23 Canada Cohort; n  = 1.817 31.86 1.563 white 254 NI Environ (fish, amalgam) Maternal blood (ICP-MS) Maternal blood 1st trimester: zero amalgams: 0.58 1 to 4 amalgams: 0.74 ≥ 5 amalgams: 0.90 GH Mean difference: yes; aOR = 1.31 (0.92, 1.85) < 5 dental amalgams group; aOR = 1.32 (0.86, 2.04) ≥ 5 dental amalgams group Wang et al. (2020), 24 China Case-control; n  = 854 20–30 854 NI Environ Maternal blood (ICP-MS) Maternal blood: PE: 1.52 (0.97–2.36); Control: 1.49 (0.96–2.08) PE aOR = 1,60 (1,08–2,38; p  = 0,039) in high Hg ≥ 1,89 Abbreviations: aOR, adjusted odds ratio; As, arsenic; CVAAS, cold vapor atomic absorption spectroscopy; Environ, environmental; Ex, exposed; GH, gestational hypertension; Hg, mercury; ICP-MS, inductively coupled plasma mass spectrometry; NEx, not exposed; NI, not informed; Occupat, occupational; PE, preeclampsia; Sn, tin; UBC, umbilical cord blood. The assessment of the methodological quality of the articles by the Downs and Black checklist showed that 3 were considered satisfactory and three were rated as good (mean = 19.3 ± 1.6 out of 28 points). The representativeness of the samples and the adjustment for confounding factors were the most often not clearly described items. For example, two studies did not adjust for any confounding factors, four adjusted for them, but only one made an adjustment for fish intake among these three studies. 23 The quality assessments for the selected studies are provided in Table 1 . Table 1 Methodological assessment of the selected studies Downs and black checklist – subscales Vigeh et al. 20 El-Badry et al. 25 Bommarito et al. 21 Liu et al. 22 Louopou et al. 23 Wang et al. 24 Reporting (10 items) 7 7 9 9 9 8 External Validity (3 items) 2 3 3 3 3 3 Internal validity – bias (7 items) 5 5 5 5 5 5 Power (1 item) 1 1 1 1 1 1 Total score 17 18 20 21 21 19 Discussion The present systematic review identified six studies that focused on Hg exposure and HDP, with mixed results. Previously, two systematic reviews had addressed the association of Hg exposure with blood pressure/hypertension in general population. 8 26 Together, they gathered 30 studies, but only 2 comprised pregnant women. Very few studies have investigated the association between Hg exposure and hypertension during pregnancy and, in general, the ones that did it reported inconsistent findings. These discrepancies may partially be explained by the study methodology differences, such as sample size, exposure levels, chemical forms of Hg and its toxicokinetics, Hg biomarkers used to assess the exposure, role of metal mixture, as well as the absence or proper adjustment for confounding factors, including fish intake, a probable cause of negative confounding. 8 16 24 27 Our review also observed mixed results, with four studies reporting no association, despite the level of exposure. 20 21 22 23 The other two studies reported a positive association in groups with more exposure, although the authors used different cutoff levels for classification. 24 25 A recent systematic review with meta-analysis reported an association among those exposed to high Hg levels (hair Hg ≥ 2 µg/g) and hypertension and blood pressure. The authors suggested these levels might be considered the threshold of the toxic effect of Hg on hypertension. 8 We highlight two studies that addressed the association in both exposure scenarios, single metal, and multiple metals. 21 24 One study evaluated 28 preeclamptic women and reported no association in neither model. 21 The other investigated 854 pregnant women and found an association only in the multi-metal model (aOR multi-metal = 1.60; 95%CI = 1.08–2.38 versus aOR single metal = 1.23. 95%CI = 0.87–1.73). 24 As metals are usually dispersed in the environment, it is essential to examine their possible interactions. 28 In addition, four studies investigated the mean difference and two found greater levels in pregnant women with HDP. 23 25 However, it is pretty challenging to compare mean Hg levels between biomarkers as there is uncertainty about how mercury accumulates and is distributed across tissues. 29 Although Hg is largely distributed worldwide and hypertension is the most common medical problem encountered during pregnancy, we could retrieve only five studies for the analysis. Only one was from North Africa and none were from Latin America and the Caribbean, despite their high birth rate and low- and middle-income countries. According to 2019 data from the World Bank, 30 the fertility global tax (FGT) was 2.4 children per woman, while in the Sub-Saharan African countries, it reached 4.6. When comparing incomes, high-income countries had a FGT of 1.6, while low- and middle-income countries had 2.5 and low-income countries had 4.6. 30 All humans are exposed to some level of Hg during their lifetime. In the general population, it mainly occurs through consuming fish and shellfish contaminated with MeHg. Also, they are exposed to relatively low levels of Hg 0 /IHg, primarily through dental amalgam, and through inhalation from anthropogenic sources. 8 15 On the other hand, elevated exposure to Hg 0 /IHg is found at workplaces, such as gold mines and dentist offices. 8 In our review, most (97.4%) participants were environmentally exposed, probably through diet, although only 1 study did report its frequency, but not the type of fish. 23 The direct measurement of the level of exposure, one of the major types of biomarkers, lessens the possibility of misclassification. 31 In our review, instead of relying on the history of exposure, we chose to select studies that measured Hg levels in any biological matrix. However, we should point out the different toxicological characteristics of the three types of Hg. MeHg has a higher absorption in the gastrointestinal tract and is usually measured in blood or hair. The first indicates a recent exposure, while it points to long-term average exposure in hair. The target organ for MeHg is the brain. On the other hand, Hg 0 and IHg have high absorption through the respiratory system and usually are detected in urine, suggesting a recent exposure. The target organs for Hg 0 are the brain and kidney, and for IHg, it is the kidney. Of note, only MeHg and Hg 0 readily pass placental barriers, and Hg levels measured in umbilical cord blood suggest an exposure in the 3 rd trimester. 32 In our review, four studies assessed Hg exposure through blood samples (maternal blood, maternal red blood cell, and umbilical cord blood), 20 22 23 and two did it in urine samples. 21 25 Thus, we had access to information on recent exposures, not on past ones, due to the biological matrices used. 26 Overall, the selected studies were considered satisfactory according to the quality assessment tool used. As all studies were observational, confounding is potentially present. The adjustment for confounding factors was one of the items with significant gaps in our review. Two studies ignored it and four adjusted for confounding factors. Among those, only one adjusted for fish intake. 23 Fish is a food source of MeHg and essential nutrients, such as selenium and n-3 polyunsaturated fatty acids, which may have important cardiovascular benefits, such as a small but significant decline in blood pressure. 27 33 When exposure to a toxicant occurs from a food source, such as fish, negative confounding occurs, resulting in underestimating Hg toxicity and fish benefits. 27 Therefore, the four studies that did not adjust for this variable could have hampered the results. To our knowledge, the present review was the first one to focus on the association between Hg exposure and HDP. As Hg is one of the most toxic substances widely dispersed in nature and pregnancy is a period of heightened susceptibility to environmental threats and cardiovascular risk, addressing their association is of utmost importance for public health. 16 To that end, we followed prespecified methods to review the evidence systematically. However, as a systematic review of observational studies, there are also some inherent limitations. First, the absence or no proper adjustment for confounding factors, especially fish intake, may be a significant reason the evidence is still inconclusive. Second, although we chose to accept studies that assess Hg exposure through measuring it in biological matrices (biomarkers), interindividual variations in the Hg kinetics cannot be disregarded as they are not well known. 32 Besides, using four different biomarkers (maternal blood, maternal red blood cells, umbilical cord blood, and urine) may introduce uncertainty to assess Hg exposure. Third, we observed substantial heterogeneity between the classification of groups according to Hg exposure level (low, middle, or high), even though there is a recommendation regarding human blood levels of Hg for pregnant women of up to 3.5 μg/L. 34 Fourth, we should acknowledge the lack of studies from developing countries, representing a significant gap in the literature, as populations with high fertility rates and living in low- and middle-income countries were also not investigated. Finally, we evaluated the relationship between Hg and HDP (categorical variable) but not with blood pressure levels (numerical variable). Not including the latter may lose studies addressing the Hg effect on blood pressure without necessarily leading to hypertension. Conclusion Although Hg is a toxicant widely dispersed worldwide and pregnancy is a life stage of heightened susceptibility, our review retrieved only six studies addressing the association between Hg and HDP. We found mixed results, and two of these studies found a positive association in the groups with more Hg exposure. Besides, absence or no proper adjustment for confounding factors, especially the negative one (fish intake), could hamper the results. Due to the public health impact of this topic, future studies must focus on the potential effect of Hg exposure on HDP, with particular attention to adjusting for negative confounding. Conflict of Interests The authors have no conflict of interests to declare. ==== Refs References 1 Barroso W KS Rodrigues C IS Bortolotto L A Mota-Gomes M A Brandão A A Feitosa A DM Brazilian guidelines of hypertension – 2020 Arq Bras Cardiol 2021 116 03 516 658 10.36660/abc.20201238 33909761 2 Gestational Hypertension and Preeclampsia: ACOG Practice Bulletin, Number 222 Obstet Gynecol 2020 135 06 e237 e260 10.1097/AOG.0000000000003891 32443079 3 Sousa M G Lopes R G Rocha M L Lippi U G Costa E S Santos C M Epidemiologia da hipertensão arterial em gestantes Einstein (Sao Paulo) 2020 18 1 7 10.31744/einstein_journal/2020AO4682 4 Subki A H Algethami M R Baabdullah W M Alnefaie M J Alzanbagi M A Alsolami R M Prevalence, risk factors, and fetal and maternal outcomes of hypertensive disorders of pregnancy: a retrospective study in Western Saudi Arabia Oman Med J 2018 33 05 409 415 10.5001/omj.2018.75 30210720 5 Farzan S F Howe C G Chen Y Gilbert-Diamond D Korrick S Jackson B P Prenatal and postnatal mercury exposure and blood pressure in childhood Environ Int 2021 146 106201 10.1016/j.envint.2020.106201 33129000 6 Houston M C Role of mercury toxicity in hypertension, cardiovascular disease, and stroke J Clin Hypertens (Greenwich) 2011 13 08 621 627 10.1111/j.1751-7176.2011.00489.x 21806773 7 Valera B Dewailly E Poirier P Environmental mercury exposure and blood pressure among Nunavik Inuit adults Hypertension 2009 54 05 981 986 10.1161/HYPERTENSIONAHA.109.135046 19805642 8 Hu X F Singh K Chan H M Mercury exposure, blood pressure, and hypertension: a systematic review and dose-response meta-analysis Environ Health Perspect 2018 126 07 76002 10.1289/EHP2863 9 Kahn L G Trasande L Environmental toxicant exposure and hypertensive disorders of pregnancy: recent findings Curr Hypertens Rep 2018 20 10 87 10.1007/s11906-018-0888-5 30090982 10 Ha E Basu N Bose-O'Reilly S Dórea J G McSorley E Sakamoto M Current progress on understanding the impact of mercury on human health Environ Res 2017 152 419 433 10.1016/j.envres.2016.06.042 27444821 11 Fillion M Mergler D Sousa Passos C J Larribe F Lemire M Guimarães J R A preliminary study of mercury exposure and blood pressure in the Brazilian Amazon Environ Health 2006 5 29 10.1186/1476-069X-5-29 17032453 12 Vassallo D V Simões M R Giuberti K Azevedo B F Ribeiro Junior R F Salaices M Effects of chronic exposure to mercury on angiotensin-converting enzyme activity and oxidative stress in normotensive and hypertensive rats Arq Bras Cardiol 2019 112 04 374 380 10.5935/abc.20180271 30624528 13 Lorscheider F L Vimy M J Summers A O Mercury exposure from “silver” tooth fillings: emerging evidence questions a traditional dental paradigm FASEB J 1995 9 07 504 508 7737458 14 Malm O Gold mining as a source of mercury exposure in the Brazilian Amazon Environ Res 1998 77 02 73 78 10.1006/enrs.1998.3828 9600798 15 Driscoll C T Mason R P Chan H M Jacob D J Pirrone N Mercury as a global pollutant: sources, pathways, and effects Environ Sci Technol 2013 47 10 4967 4983 10.1021/es305071v 23590191 16 Wells E M Herbstman J B Lin Y H Hibbeln J R Halden R U Witter F R Methyl mercury, but not inorganic mercury, associated with higher blood pressure during pregnancy Environ Res 2017 154 247 252 10.1016/j.envres.2017.01.013 28110211 17 PRISMA-P Group Moher D Shamseer L Clarke M Ghersi D Liberati A Petticrew M Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement Syst Rev 2015 4 01 1 10.1186/2046-4053-4-1 25554246 18 Downs S H Black N The feasibility of creating a checklist for the assessment of the methodological quality both of andomized and non-randomised studies of health care interventions J Epidemiol Community Health 1998 52 06 377 384 10.1136/jech.52.6.377 9764259 19 Hooper P Jutai J W Strong G Russell-Minda E Age-related macular degeneration and low-vision rehabilitation: a systematic review Can J Ophthalmol 2008 43 02 180 187 10.3129/i08-001 18347620 20 Vigeh M Yokoyama K Ramezanzadeh F Dahaghin M Sakai T Morita Y Lead and other trace metals in preeclampsia: a case-control study in Tehran, Iran Environ Res 2006 100 02 268 275 10.1016/j.envres.2005.05.005 16029873 21 Bommarito P A Kim S S Meeker J D Fry R C Cantonwine D E McElrath T F Urinary trace metals, maternal circulating angiogenic biomarkers, and preeclampsia: a single-contaminant and mixture-based approach Environ Health 2019 18 01 63 10.1186/s12940-019-0503-5 31300062 22 Liu T Zhang M Guallar E Wang G Hong X Wang X Trace minerals, heavy metals, and preeclampsia: findings from the Boston Birth Cohort J Am Heart Assoc 2019 8 16 e012436 10.1161/JAHA.119.012436 31426704 23 Louopou R C Trottier H Arbuckle T E Fraser W D Dental amalgams and risk of gestational hypertension in the MIREC study Pregnancy Hypertens 2020 21 84 89 10.1016/j.preghy.2020.04.015 32447273 24 Wang Y Wang K Han T Zhang P Chen X Wu W Exposure to multiple metals and prevalence for preeclampsia in Taiyuan, China Environ Int 2020 145 106098 10.1016/j.envint.2020.106098 32916414 25 El-Badry A Rezk M El-Sayed H Mercury-induced oxidative stress may adversely affect pregnancy outcome among dental staff: a cohort study Int J Occup Environ Med 2018 9 03 113 119 10.15171/ijoem.2018.1181 29995016 26 Rosen E M Muñoz M I McElrath T Cantonwine D E Ferguson K K Environmental contaminants and preeclampsia: a systematic literature review J Toxicol Environ Health B Crit Rev 2018 21 05 291 319 10.1080/10937404.2018.1554515 30582407 27 Choi A L Cordier S Weihe P Grandjean P Negative confounding in the evaluation of toxicity: the case of methylmercury in fish and seafood Crit Rev Toxicol 2008 38 10 877 893 10.1080/10408440802273164 19012089 28 Singh N Gupta V K Kumar A Sharma B Synergistic effects of heavy metals and pesticides in living systems Front Chem 2017 5 70 10.3389/fchem.2017.00070 29075624 29 Dack K Fell M Taylor C M Havdahl A Lewis S J Mercury and prenatal growth: a systematic review Int J Environ Res Public Health 2021 18 13 7140 10.3390/ijerph18137140 34281082 30 The World Bank Fertility rate, total (births per woman) [Internet] 2019[cited 2022 Mar 14]. Available from:https://data.worldbank.org/indicator/SP.DYN.TFRT.IN?locations%3DZG 31 Mayeux R Biomarkers: potential uses and limitations NeuroRx 2004 1 02 182 188 10.1602/neurorx.1.2.182 15717018 32 Berglund M Lind B Björnberg K A Palm B Einarsson O Vahter M Inter-individual variations of human mercury exposure biomarkers: a cross-sectional assessment Environ Health 2005 4 20 10.1186/1476-069X-4-20 16202128 33 Morris M C Sacks F Rosner B Does fish oil lower blood pressure? A meta-analysis of controlled trials Circulation 1993 88 02 523 533 10.1161/01.cir.88.2.523 8339414 34 Stone J Sutrave P Gascoigne E Givens M B Fry R C Manuck T A Exposure to toxic metals and per- and polyfluoroalkyl substances and the risk of preeclampsia and preterm birth in the United States: a review Am J Obstet Gynecol MFM 2021 3 03 100308 10.1016/j.ajogmf.2021.100308 33444805
PMC009xxxxxx/PMC9800150.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36580937 10.1055/s-0042-1753546 220089 Original Article High Risk Prenancy Perinatal Outcomes in Women with Chronic Kidney Diseases Resultados perinatais em mulheres com doenças renais crônicashttp://orcid.org/0000-0003-1202-2592 Zilli Marcus Vinicius Pinheiro 1 http://orcid.org/0000-0002-2659-6012 Borovac-Pinheiro Anderson 1 http://orcid.org/0000-0001-8280-3234 Costa Maria Laura 1 http://orcid.org/0000-0003-4335-0337 Surita Fernanda Garanhani 1 1 Department of Obstetrics and Gynecology, Universidade Estadual de Campinas, Campinas, SP, Brazil Address for correspondence Fernanda Garanhani Surita, Full Professor Department of Obstetrics and Gynecology, Universidade Estadual de CampinasRua Alexander Fleming, 101, Campinas, SP [email protected] 29 12 2022 12 2022 1 12 2022 44 12 10941101 08 3 2022 19 5 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  To assess maternal and neonatal outcomes in women with chronic kidney disease (CKD) at a referral center for high-risk pregnancy. Methods  A retrospective cohort of pregnant women with CKD was followed at the Women's Hospital of Universidade Estadual de Campinas, Brazil, between 2012 and 2020. Variables related to disease etiology, treatment duration, sociodemographic variables, lifestyle, other associated diseases, obstetric history, and perinatal outcomes were assessed. The causes of CKD were grouped into 10 subgroups. Subsequently, we divided the sample according to gestational age at childbirth, as preterm and term births, comparing maternal and neonatal outcomes, and baseline characteristics as well as outcomes among such groups. Results  A total of 84 pregnancies were included, in 67 women with CKD. Among them, six pregnancies evolved to fetal death, five to miscarriage, and one was a twin pregnancy. We further analyzed 72 single pregnancies with live births; the mean gestational age at birth was 35 weeks and 3 days, with a mean birth weight of 2,444 g. Around half of the sample (51.39%) presented previous hypertension, and 27.7% developed preeclampsia. Among the preterm births, we observed a higher frequency of hypertensive syndromes, longer maternal intensive care unit (ICU) stay in the postpartum period, higher incidence of admission to the neonatal ICU, higher neonatal death, lower 5-minute Apgar score, and lower birth weight. Conclusion  This study demonstrates increased adverse outcomes among pregnancies complicated by CKD and expands the knowledge on obstetric care among such women in an attempt to reduce maternal risks and identify factors related to prematurity in this population. Resumo Objetivo  Avaliar os desfechos maternos e neonatais em mulheres com doença renal crônica (DRC) em um centro de referência para gestação de alto risco. Métodos  Coorte retrospectiva de gestantes com DRC acompanhadas no Hospital da Mulher da Universidade Estadual de Campinas, Brasil, entre 2012 e 2020. Variáveis relacionadas à etiologia da doença, duração do tratamento, variáveis sociodemográficas, estilo de vida, outras doenças associadas, história obstétrica, número de consultas de pré-natal e os resultados perinatais foram avaliados. As causas da DRC foram agrupadas em 10 subgrupos. Posteriormente, dividimos a amostra de acordo com a idade gestacional no parto, pois os nascimentos pré-termo e a termo comparam os desfechos maternos e neonatais bem como as características basais e desfechos entre esses grupos. Resultados  Um total de 84 gestações foram incluídas em 67 mulheres com DRC. Dentre elas, seis gestações evoluíram para óbito fetal, cinco para aborto espontâneo, e uma era gestação gemelar. Foram analisadas ainda 72 gestações únicas, com nascidos vivos; a idade gestacional média ao nascer foi de 35 semanas e 3 dias, e o peso médio ao nascer foi 2.444 g. Cerca de metade da amostra (51,39%) apresentava hipertensão prévia e 27,7% desenvolveram pré-eclâmpsia. Entre os casos de prematuridade (34 casos), observamos maior frequência de síndromes hipertensivas, mais dias de internação materna na UTI no pós-parto, maior incidência de internação na UTI neonatal, óbito neonatal, menor índice de Apgar de 5 minutos e menor peso ao nascimento. Conclusão  Este estudo demonstra o aumento de desfechos adversos em gestações complicadas por DRC e amplia o conhecimento sobre cuidados obstétricos entre essas mulheres na tentativa de reduzir os riscos maternos e identificar fatores relacionados à prematuridade nessa população. Keywords kidney disease high-risk pregnancy antenatal care perinatal outcomes Palavras-chave doença renal gravidez de alto risco cuidado pré-natal cuidados perinatais ==== Body pmcIntroduction Chronic kidney disease (CKD) is a global health problem that affects ∼ 10% of the population. The prevalence has increased in recent decades and is higher among low- and middle-income countries. 1 In Brazil, more than ten million people have CKD. Chronic kidney disease occurs in women and men equally, and reproductive function can be affected in women, in addition to influencing maternal and neonatal outcomes. 2 Approximately 3 to 4% of women of reproductive age and ∼ 1 to 3% of pregnant women have CKD, regardless of the underlying cause. 3 4 In these patients, there is a greater risk of maternal hypertensive complications, fetal growth restriction, and premature birth; therefore, there is a greater chance of hospitalization of the newborn in a neonatal intensive care unit (ICU), stillbirth and neonatal death, in addition to morbidities related to prematurity. 5 Women with CKD are 10 times more likely to develop preeclampsia than women at usual risk, with a reported prevalence of preeclampsia of up to 40% among pregnant women with CKD. 6 The reported overall prevalence of preterm birth (before 37 weeks) in Brazil is around 10 to 12%. In other countries, rates vary according to other health indicators. One of the factors that can influence this is pregestational creatinine levels. A Canadian study of 56,000 pregnancies showed an increase in therapeutic preterm birth in women with pregestational creatinine above the 95 th percentile (0.87 mg/dL), which was not observed in patients with spontaneous preterm births. In this same study, a graph was constructed that illustrates a J-curve supporting the association of serum creatinine and the probability of preterm delivery, with a 1.23-fold increase in the chance of preterm delivery in patients who had some renal dysfunction compared with pregnant women with normal renal function. 7 Despite the possible unfavorable perinatal outcomes in pregnant women with CKD reported in the literature, there is still a lack of Brazilian studies on the subject. The aim of this study is to evaluate the maternal and perinatal outcomes of women with CKD who underwent prenatal care and delivery at a single Brazilian reference center for high-risk pregnancies, and further compare cases with preterm and term childbirth. Methods We performed a retrospective cohort study at the Women's Hospital of Universidade Estadual de Campinas, Brazil, a referral university hospital in southeast Brazil, accounting for a surrounding population of 3,100,000 inhabitants. This study was approved by the research ethics committee of the institution (CAAE report 15429419.5.0000.5404). We included all pregnancies of women with a previous diagnosis of CKD who underwent prenatal follow-up at the specialized antenatal care (ANC) outpatient clinic and who gave birth at the Women's Hospital between 2012 and 2020. All patients with high risk of preeclampsia were given prophylaxis with low dose aspirin and calcium supplementation, as recommended by institutional protocol. We collected data from the medical records on an electronic system by completing a data collection form specifically created for the study. We evaluated variables related to CKD etiology, duration of kidney disease treatment, sociodemographic variables, lifestyle variables, other associated diseases, and obstetric history and perinatal outcomes. In the case of patients with more than one pregnancy during the study period, each index pregnancy was considered, that is, the unit of study was the pregnancy. The data obtained were entered into a database created for this study, in Excel format, which was reviewed to identify inconsistencies. The underlying causes of kidney disease were later grouped into 10 subgroups according to similar characteristics and frequency of diagnoses. To describe the profile of the sample according to the variables under study, frequency tables of categorical variables were made with absolute (n) and percentage (%) frequency values, and descriptive statistics of numerical variables, with mean values and standard deviation. Subsequently, women who had a viable pregnancy (excluding abortions and fetal deaths) were divided into 2 groups according to the occurrence or not of prematurity (gestational age [GA] < 37 weeks). To compare the categorical variables between pregnancies that ended in preterm birth, the chi-squared test or Fisher exact test (for expected values lower than 5) were used. The significance level adopted for the statistical tests was 5%. The software used was the SAS System for Windows version 9.2. (SAS Institute Inc, 2002–2008, Cary, NC, USA). 8 9 10 11 All Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) requirements for an observational study were followed and verified in this article. Results A total of 84 pregnancies were included, in 67 women with CKD who underwent absolute neutrophil count (ANC) between 2012 and 2020. Of these, 6 pregnancies evolved with fetal deaths and 5 with abortion, totaling 11 gestational losses, which corresponds to 13.1% of this sample. A diamniotic dichorionic twin pregnancy occurred in a 32-year-old primigravid patient with a history of systemic lupus erythematosus (SLE), CKD on dialysis, kidney transplant in 2010 and viral infection with loss of the transplanted kidney. The patient progressed to preterm labor at 32 weeks and underwent a cesarean section; the newborns were born weighing 1,460 g and 1,197 g, with favorable neonatal outcomes. Considering the single pregnancies that progressed to childbirth ( n  = 72), the mean age of the pregnant women was 28.58 years (standard deviation [SD] = 6.34), with a mean time since diagnosis of CKD of 10.61 years (SD = 8.82). Most of the women were white, were in a stable relationship, + and had high school education; none of the participants reported using alcohol or illicit drugs. Table 1 shows the sociodemographic data of the patients included in the study. Table 1 Characteristics of women with chronic kidney disease and singleton pregnancy that progressed to childbirth N = 72 % Age (years)  < 20 3 4.17  20–29 35 48.61  30–39 31 43.05  ≥ 40 3 4.17 Age (mean)  28.58 years Marital status  Single 26 36.11  Stable relationship 46 63.89 Occupation*  Paid work 11 15.28  Unpaid work 28 38.89 Schooling**  Elementary 13 13.06  High school 33 45.83  University 5 6.94 Skin color***  White 46 63.89  Non-White 24 33.33 Smoking 2 4.17 Number of Pregnancies  1 31 43.05  2 16 22.22  ≥ 3 25 34.73 Previous miscarriage  0 59 81.94  ≥ 1 13 18.06 Frequency missing *33 **21 ***2. We grouped the main causes of CKD into 10 categories, presented from the most frequent to the least frequent: SLE ( n  = 21), glomerulopathy ( n  = 12), nephrotic syndrome ( n  = 11), transplant ( n  = 10), infection ( n  = 8), dialysis ( n  = 3), hypertension ( n  = 3), diabetes ( n  = 2), and other diseases with lower frequency that were included in the “others” group ( n  = 2: one patient with Wegener granulomatosis and the other with rheumatoid arthritis and nephrolithiasis). Fig. 1 shows the cause of CKD of the patients included in the study. Fig. 1 Main causes of chronic kidney disease among pregnancies followed at high-risk antenatal care ( n  = 84). Among the 72 pregnancies that resulted in live births, the mean gestational age at birth was 35 weeks and 3 days (SD = 7.21), with a mean birth weight of 2,443.7 g (SD = 722.48). The majority had a birth at term (52.78%) or late preterm (34.72%). Eight newborns were classified as small for gestational age (11.1%), and 5 of them were children of mothers with SLE (⅝ = 62.5%). Five children were classified as low birth weight newborns (0.69%). Table 2 shows obstetric and neonatal data for patients with CKD included in the study. These women attended an average of 9.29 (SD = 3.88) prenatal consults. Approximately a quarter (20/72; 27.78%) of the study population developed preeclampsia during pregnancy. Table 2 Obstetric history, data on pregnancy, childbirth, and neonatal outcomes of women with chronic kidney disease N = 72 % Hypertensive syndrome Chronic hypertension 17 23.61 Chronic hypertension with superimposed preeclampsia 12 16.67 Preeclampsia 8 11.11 Without hypertension 35 48.61 Mode of birth Vaginal 23 31.94 C-Section 49 68.06 Maternal morbidities after birth* Hemorrhage 6 8.33 Infection 4 5.56 Others**** 6 8.33 None 53 76.61 Maternal ICU admission** Yes 9 12.50 No 62 86.11 Apgar score at the first minute < 7 14 19.44 ≥ 7 57 79.17 Apgar score at the fifth minute** < 7 3 4.17 ≥ 7 68 94.44 Gestational age at birth GA < 28 4 5.55 ≥ 28 GA < 32 5 6.95 ≥ 32 GA < 37 25 34.72 GA ≥ 37 38 52.78 Neonatal ICU admission** Yes 26 36.11 No 45 62.50 Neonatal death*** Yes 6 8.83 No 62 86.11 Abbreviation: GA, gestational age; ICU, intensive care unit. Missing *3 **1 ***4 (no information due to medical transfer of the newborn to another hospital) ****Others: one hypoglycemia, four with hypertensive spikes, one with hypervolemia and dialysis. Considering the high prevalence of preterm birth among the considered cases, and the burden of this condition for mothers and their children in the short and long term, we aimed to investigate conditions associated with this event. We compared the two groups according to the occurrence or not of preterm birth. There was a significant association between the occurrence of premature birth (gestational age < 37 weeks) and the need for a woman to be hospitalized in the ICU after childbirth, a higher occurrence of complications after childbirth, and a greater number of days of hospitalization ( Table 3 ). The majority cause of preterm birth among our population was preeclampsia (12/34 = 35.3%) and premature rupture of membranes (4/34 = 11.7%). This population had more prevalence of ICU admission and more days staying in the hospital. Table 3 Comparison of maternal characteristics and outcomes according to the occurrence of preterm birth ( n  = 72) Variables Preterm birth ( n  = 34) Term birth ( n  = 38) p -value Maternal age (mean / SD) 29.03 (6.44) 28.18 (6.31) 0.560 * Years since diagnosis (mean / SD) 10.00 (7.68) 11.13 (9.77) 0.894* Maternal hospitalization in days (mean / SD) 4.85 (2.93) 3.13 (1.44) 0.001* Group of kidney disease 0.143** SLE 8 (23.5%) 13 (34.2%) Glomerular disease 4 (11.7%) 8 (21.1%) Nephrotic syndrome 5 (14.7%) 6 (15.8%) Transplant 4 (11.7%) 6 (15.8%) Infectious 7 (20.6%) 1 (2.6%) Hemodialysis 2 (5.9%) 1 (2.6%) Hypertensive nephropathy 2 (5.9%) 1 (2.6%) Diabetes nephropathy 2 (5.9%) 0 Others 0 2 (5.3%) Skin color 0.892** White 23 (67.3%) 23 (63.9%) Non-white 11(32.7%) 13 (36.1%) Hypertensive syndrome 0.009** yes 23 (67.7%) 14 (36.8%) no 11 (32.3%) 24 (63.2%) Preeclampsia 0.178** yes 12 (35.3%) 8 (21.0%) no 22 (64.7%) 30 (78.9%) Mode of birth 0.663** vaginal 10 (29.4%) 13 (34.2%) cesarean 24 (70.6%) 25 (65.8%) Postpartum maternal ICU 0.006** yes 11 (34.4%) 3 (11.9%) no 21 (65.6%) 35 (92.1%) 0.029** Adverse maternal outcome # yes 11 (35.5%) 5 (13.2%) no 20 (64.5%) 33 (86.8%) Abbreviations: ICU, intensive care unit; SD, standard deviation; SLE. * Kruskal-Wallis test, ** Qui-square test, *** Fisher test # adverse maternal outcome: bleeding, infection or other; #2 missing data. Discussion The present study reports increased adverse maternal and neonatal outcomes among cases of CKD followed at a referral maternity hospital. Overall, around 13.1% of pregnancies progressed to abortion or stillbirth, and, among the cases of livebirths, almost half were preterm deliveries, with around one quarter complicated by preeclampsia. The cases of preterm delivery were associated with increased adverse outcomes, with 6 neonatal deaths. According to international epidemiological data, the average rate of preterm births in the general population is 7 to 12% of births and, of these, ∼ 12% occur due to preeclampsia. 12 In our study, as expected, the incidence of premature births was much higher than in the general population (almost 50%), which is most likely due to CKD itself or secondary to the development of hypertensive syndromes and their consequences. The incidence of prematurity among pregnant women with CKD found in our study is similar to the data reported in the literature. A meta-analysis of 23 studies and 506,340 pregnant women concluded that CKD increased the risk of preeclampsia 10-fold, the risk of premature delivery and small-for-gestational-age newborns and led to a 3-fold increased risk of cesarean section. 6 Another study showed an association between CKD stage and its implications, with an increased incidence of prematurity as the CKD stage increased (CKD stage 1: 23.5% preterm; stage 2: 50.6%; stage 3: 78.4%; stage 4–5: 88.9%), using a serum creatinine threshold of 1.9 mg/dL, observed 93% newborn survival, 59% preterm birth, and fetal growth restriction of 37%. 13 14 In our study, we did not distinguish the CKD stage of the patients; however, our data are similar to the data from this study, since the incidence is within this range. Other studies have already shown that there is an association between CKD and prematurity, and one showed a 1.23-fold increase in relative risk of preterm birth in patients with prepregnancy kidney dysfunction, compared with those with normal renal function. 2 7 13 15 Our data show that prematurity was most likely associated with more severe cases, with a greater number of women hospitalized in adult ICUs among those who had premature births (around one-third of cases), while in patients with term birth, the number of ICU admissions was much lower (less than 10%). Maternal factors associated with CKD can increase the chance of patients being hospitalized in the ICU, due to the complexity of their cases, especially in those undergoing dialysis. Epidemiologically, the most prevalent risk factors for CKD in the general population are arterial hypertension and diabetes mellitus; however, among pregnant women (mostly young women), other comorbidities are associated with the loss of renal function. 16 In our study, the mean age of the patients was 28.5 years, with multiple other causes for CKD. Additionally, the study hospital is the referral hospital for some diseases, including SLE. It was also possible to verify that CKD was a risk factor for the development of hypertensive syndromes during pregnancy, more frequently observed in the group of patients whose outcome was preterm birth. It is known that CKD is a factor for hypertensive syndromes, as observed in several studies. One study of 778 women with CKD reported that 25.3% presented chronic arterial hypertension, and the incidence of preeclampsia was 9.3%. 17 Another study reported an incidence of chronic arterial hypertension of 30.5% and of preeclampsia of 24.6%, with a higher rate of preeclampsia, explained in the study by almost a quarter of patients having CKD at more advanced stages (3–5). 18 In addition, meta-analyses and cohorts show that women with CKD are 10 times more likely to develop preeclampsia, and up to 40% of patients with preeclampsia have had CKD previously. 2 6 19 20 21 22 Acetylsalicylic acid has been recommended as an effective intervention to reduce the incidence of preeclampsia, especially in women with known risk factors, including those with CKD, preferably introduced between 12 and 16 weeks. 19 23 24 A study showed that its use may reduce the chance of developing preeclampsia and intrauterine growth restriction 24 , while another showed a reduction in the incidence of severe preeclampsia among patients with CKD stages 3 to 5, with no evidence in this study among patients with CKD stages 1 to 2. 18 24 However, another controlled trial showed a reduction in preterm preeclampsia in patients that used aspirin. 25 Our patients, guided by the institution protocol, used aspirin prophylaxis throughout the pregnancy, as well as calcium supplementation, as this is recommended in some groups of patients such as those with SLE. 26 The overall incidence of preeclampsia in the general population is ∼ 4.6 to 8.1%, depending on the region. 5 19 27 In this study group, we saw a rate of 27.78%, high compared with the general population, but close to the values found in other studies that evaluated populations with CKD. We currently know that CKD, even in its early stages, is associated with the production of proinflammatory cytokines, which triggers endothelial inflammation and consequently increases the chance of developing hypertension. In a normal pregnancy, there is a balance between angiogenic factors (among them PlGF and VEGF) and anti-angiogenic factors (sFlt-1), favoring good placental implantation. However, when there is an imbalance between these factors, poor implantation of the placenta or worsening of placental perfusion can occur, leading to a reduction in factors such as PlGF and VEGF and an increase in sFlt-1, causing endothelial inflammation and increasing the chances of developing preeclampsia. 5 Biomarker assessment may be an interesting way to adequately distinguish preeclampsia from other complications that can present with worsening proteinuria and hypertension. 28 In addition, there is evidence that these proinflammatory factors can cause glomerular damage, leading to proteinuria, which can worsen renal function or favor the development of preeclampsia. 15 29 Some of these studies report that the decline in kidney function is even worse among patients with CKD stages 3 to 5, while other studies did not see any worsening of renal function in stages 1 to 3. 15 18 These data in the literature still lacks consensus and require further investigation. Our study presented some limitations. Given that our data was from a single center, it was not possible to further investigate the association between the reported adverse outcomes and the diverse CKD reported. In addition, pregnant women were referred to our service after the diagnosis of kidney disease, implying that we had no data related to the kidney biopsy, details about the infections that caused kidney failure or previous treatments. Our patients had a very heterogeneous treatment during pregnancy based on the causes and evaluation of their kidney disease. It was possible only to the group of women who underwent dialysis. The others received basic care to treat the underlying disease, such as hypertension, lupus, and diabetes. However, our sample is relevant to referral centers in a middle-income setting. In our study, 68.06% of the patients underwent cesarean section, a number higher than that recommended by the World Health Organization (WHO) and higher than the Brazilian average. 30 31 However, in this case, these are patients with a greater possibility of complications, or acute or chronic fetal distress, which may be factors that increase the chances of opting for cesarean delivery. Even so, it is a high rate of cesarean sections, similar to that seen in other studies, which found cesarean section rates of 37 to 59%. 13 14 This corroborates the results of the study by Zhang et al., which showed a 3-fold increase in the number of cesarean sections in patients with CKD. 6 Conclusion These data reinforce that pregnancy complicated by CKD can present increased adverse maternal and perinatal outcomes, in addition to worsening the underlying disease or renal function. It is necessary to counsel these women on adequate family planning, to help plan their pregnancies when their kidney disease is stable and controlled. These are patients who require multiprofessional evaluation at a referral center, with special attention and care during high-risk prenatal care. With a planned pregnancy, it is possible to better evaluate risk factors and prognosis, and evaluate the indication of prophylaxis for preeclampsia, in addition to undertaking maternal-fetal surveillance and monitoring. Acknowledgments The authors would like to thank Helymar Costa Machado, a statistician who made it possible to analyze the data of the study. Contributors Conflict of Interests The authors have no conflict of interests to declare. Marcus Vinicius Pinheiro Zilli: investigation, data curation and writing - original draft. Anderson Borovac-Pinheiro: methodology, data curation and writing - original draft. Maria Laura Costa: investigation, review & editing. Fernanda Garanhani Surita: conceptualization, methodology, data curation, supervision, review & editing. All authors approved the final version to be published. ==== Refs References 1 GBD Chronic Kidney Disease Collaboration Global, regional, and national burden of chronic kidney disease, 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017 Lancet 2020 395 (10225):709 733 10.1016/S0140-6736(20)30045-3 32061315 2 Hui D Hladunewich M A Chronic kidney disease and pregnancy Obstet Gynecol 2019 133 06 1182 1194 10.1097/AOG.0000000000003256 31135733 3 Williams D Davison J Chronic kidney disease in pregnancy BMJ 2008 336 (7637):211 215 10.1136/bmj.39406.652986.BE 18219043 4 Kumakura S Okamoto K Takeuchi S Kidney function, blood pressure and proteinuria were associated with pregnancy outcomes of pregnant women with chronic kidney disease: a single-center, retrospective study in the Asian population Clin Exp Nephrol 2020 24 06 547 556 10.1007/s10157-020-01865-0 32162117 5 Kattah A Preeclampsia and kidney disease: deciphering cause and effect Curr Hypertens Rep 2020 22 11 91 10.1007/s11906-020-01099-1 32926258 6 Zhang J J Ma X X Hao L Liu L J Lv J C Zhang H A systematic review and meta-analysis of outcomes of pregnancy in CKD and CKD outcomes in pregnancy Clin J Am Soc Nephrol 2015 10 11 1964 1978 10.2215/CJN.09250914 26487769 7 Harel Z Park A L McArthur E Prepregnancy renal function and risk of preterm birth and related outcomes CMAJ 2020 192 30 E851 E857 10.1503/cmaj.200089 32719020 8 Conover W J Practical nonparametric statistics. 3rd ed New York John Wiley & Sons 1999 9 Siegel S Castellan N J Jr Estatística não-paramétrica para ciências do comportamento. 2a ed. Porto AlegreArt Med2006 10 Tabachnick B G Fidell L S Using multivariate statistics. 4th ed Boston Allyn and Bacon 2001 11 Fleiss J L Levin B Paik M C Statistical methods for rates and proportions. 3rd ed Hoboken John Wiley & Sons 2003 12 Goldenberg R L Culhane J F Iams J D Romero R Epidemiology and causes of preterm birth Lancet 2008 371 (9606):75 84 10.1016/S0140-6736(08)60074-4 18177778 13 Piccoli G B Cabiddu G Attini R Risk of adverse pregnancy outcomes in women with CKD J Am Soc Nephrol 2015 26 08 2011 2022 10.1681/ASN.2014050459 25766536 14 Jones D C Hayslett J P Outcome of pregnancy in women with moderate or severe renal insufficiency N Engl J Med 1996 335 04 226 232 10.1056/NEJM199607253350402 8657238 15 Cabiddu G Castellino S Gernone G A best practice position statement on pregnancy in chronic kidney disease: the Italian Study Group on Kidney and Pregnancy J Nephrol 2016 29 03 277 303 10.1007/s40620-016-0285-6 26988973 16 Marinho A W Penha A P Silva M T Galvão T F Prevalência de doença renal crônica em adultos no Brasil: revisão sistemática da literatura. Cad Saúde Coletiva 2017 25 03 379 88 10.1590/1414-462×201700030134 17 Kendrick J Sharma S Holmen J Palit S Nuccio E Chonchol M Kidney disease and maternal and fetal outcomes in pregnancy Am J Kidney Dis 2015 66 01 55 59 10.1053/j.ajkd.2014.11.019 25600490 18 Wang M Chen S He Y Zhao M Yang H Chen Q Low-dose aspirin for the prevention of severe preeclampsia in patients with chronic kidney disease: a retrospective study : This is the study for kidney and pregnancy J Nephrol 2021 34 05 1631 1639 10.1007/s40620-021-01049-3 33914255 19 Wiles K Chappell L C Lightstone L Bramham K Updates in diagnosis and management of preeclampsia in women with CKD Clin J Am Soc Nephrol 2020 15 09 1371 1380 10.2215/CJN.15121219 32241779 20 International Society for the Study of Hypertension in Pregnancy (ISSHP) Brown M A Magee L A Kenny L C Karumanchi S A McCarthy F P Saito S The hypertensive disorders of pregnancy: ISSHP classification, diagnosis & management recommendations for international practice Pregnancy Hypertens 2018 13 291 310 10.1016/j.preghy.2018.05.004 29803330 21 High Risk of Pre-eclampsia Identification Group Bartsch E Medcalf K E Park A L Ray J G Clinical risk factors for pre-eclampsia determined in early pregnancy: systematic review and meta-analysis of large cohort studies BMJ 2016 353 i1753 10.1136/bmj.i1753 27094586 22 Saudan P Brown M A Buddle M L Jones M Does gestational hypertension become pre-eclampsia? Br J Obstet Gynaecol 1998 105 11 1177 1184 9853766 23 The National Institute for Health and Care Excellence Hypertension in pregnancy: diagnosis and management [Internet]2019 [cited 2022 Feb 1]. (NICE guideline; 133). Available from:www.nice.org.uk/guidance/ng133 24 Ray J G Bartsch E Park A L Shah P S Dzakpasu S Estimated reductions in provider-initiated preterm births and hospital length of stay under a universal acetylsalicylic acid prophylaxis strategy: a retrospective cohort study CMAJ Open 2017 5 02 E508 E516 10.9778/cmajo.20160092 25 Rolnik D L Wright D Poon L C Aspirin versus placebo in pregnancies at high risk for preterm preeclampsia N Engl J Med 2017 377 07 613 622 10.1056/NEJMoa1704559 28657417 26 Pastore D EA Costa M L Parpinelli M A Surita F G A critical review on obstetric follow-up of women affected by systemic lupus erythematosus Rev Bras Ginecol Obstet 2018 40 04 209 224 10.1055/s-0038-1625951 29702718 27 Giordano J C Parpinelli M A Cecatti J G The burden of eclampsia: results from a multicenter study on surveillance of severe maternal morbidity in Brazil PLoS One 2014 9 05 e97401 10.1371/journal.pone.0097401 24825164 28 de Jesús G R Lacerda M I Rodrigues B C Soluble Flt-1, placental growth factor, and vascular endothelial growth factor serum levels to differentiate between active lupus nephritis during pregnancy and preeclampsia Arthritis Care Res (Hoboken) 2021 73 05 717 721 32583963 29 Di Leo V Capaccio F Gesualdo L Preeclampsia and glomerulonephritis: a bidirectional association Curr Hypertens Rep 2020 22 05 36 10.1007/s11906-020-1033-9 32200430 30 Rudey E L Leal M DC Rego G Cesarean section rates in Brazil: Trend analysis using the Robson classification system Medicine (Baltimore) 2020 99 17 e19880 10.1097/MD.0000000000019880 32332659 31 Boerma T Ronsmans C Melesse D Y Barros A J Barros F C Juan L Global epidemiology of use of and disparities in caesarean sections Lancet 2018 392 (10155):1341 1348 10.1016/S0140-6736(18)31928-7 30322584
PMC009xxxxxx/PMC9800151.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36580935 10.1055/s-0042-1759728 220113 Original Article | Artigo Original Obstetrics Randomized Clinical Trial Comparing Quadratus Lumborum Block and Intrathecal Morphine for Postcesarean Analgesia Ensaio Clínico Randomizado Comparando Bloqueio do Quadrado Lombar e Morfina Intratecal Para Analgesia Pós-cesarianahttp://orcid.org/0000-0002-7655-7433 Araújo Karoline Moura de 1 http://orcid.org/0000-0002-5389-8796 Ferraro Leonardo Henrique Cunha 1 http://orcid.org/0000-0001-6573-7386 Sun Sue Yasaki 1 http://orcid.org/0000-0003-1405-5371 Mattar Rosiane 1 1 Escola Paulista de Medicina, Universidade Federal de São Paulo, São Paulo, SP, Brazil Address for correspondence Karoline Moura de Araújo Escola Paulista de Medicina, Universidade Federal de São PauloRua Napoleão de Barros, 715, 04024002, São Paulo, [email protected] 29 12 2022 12 2022 1 12 2022 44 12 10831089 06 4 2022 08 9 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  To compare the efficacy of quadratus lumborum (QL) block and intrathecal morphine (M) for postcesarean delivery analgesia. Methods  Thirty-one pregnant women with ≥ 37 weeks of gestation submitted to elective cesarean section were included in the study. They were randomly allocated to either the QL group (12.5 mg 0.5% bupivacaine for spinal anesthesia and 0.3 ml/kg 0.2% bupivacaine for QL block) or the M group (12.5 mg bupivacaine 0.5% and 100 mcg of morphine in spinal anesthesia). The visual analog scale of pain, consumption of morphine and tramadol for pain relief in 48 hours, and side effects were recorded. Results  Median pain score and/or pain variation were higher in the morphine group than in the QL group ( p  = 0.02). There was no significant difference in the consumption of morphine or tramadol between groups over time. Side effects such as pruritus, nausea, and vomiting were observed only in the morphine group. Conclusion  Quadratus lumborum block and intrathecal morphine are effective for analgesia after cesarean section. Patients undergoing QL block had lower postoperative pain scores without the undesirable side effects of opioids such as nausea, vomiting, and pruritus. Resumo Objetivo  Comparar a eficácia do bloqueio do quadrado lombar (QL) e da morfina intratecal (M) na analgesia pós-cesariana. Métodos  Trinta e uma gestantes com ≥ 37 semanas de gestação submetidas a cesariana eletiva foram incluídas no estudo. Eles foram alocados aleatoriamente no grupo QL (12,5 mg de bupivacaína a 0,5% para raquianestesia e 0,3 ml/kg de bupivacaína a 0,2% para bloqueio de QL) ou no grupo M (12,5 mg de bupivacaína a 0,5% e 100 mcg de morfina na raquianestesia). A escala visual analógica de dor, consumo de morfina e tramadol para alívio da dor em 48 horas e efeitos colaterais foram registrados. Resultados  A mediana do escore de dor e/ou variação da dor foi maior no grupo morfina do que no grupo QL ( p  = 0,02). Não houve diferença significativa no consumo de morfina ou tramadol entre os grupos ao longo do tempo. Efeitos colaterais como prurido, náuseas e vômitos foram observados apenas no grupo morfina. Conclusão  O bloqueio QL e a morfina intratecal são eficazes para analgesia após cesariana. Os pacientes submetidos ao bloqueio do QL apresentaram menores escores de dor pós-operatória sem os efeitos colaterais indesejáveis dos opioides, como náuseas, vômitos e prurido. Keywords analgesia cesarean section morphine spinal anesthesia anestesia obstetrical Palavras-chave analgesia cesariana morfina anestesia espinhal anestesia obstétrica ==== Body pmcIntroduction Cesarean section is the most frequently performed surgical procedure in obstetrics. 1 The number of cesarean deliveries has increased in recent years, accounting for ∼ 21% of births worldwide. 2 During the postoperative period, ∼ 1 of 5 women experience acute pain. 3 The effective management of postoperative pain is highly important as it facilitates early recovery, ambulation, and breastfeeding, allowing mothers to provide better care to their newborns. 4 5 Moreover, it helps preventing venous thromboembolism 6 and respiratory complications, decreasing hospital stay. 7 In addition to these benefits, an adequate management of acute postcesarean pain is associated with a 3-fold decrease in the risk of postpartum depression 8 and development of chronic pelvic pain. 9 10 Morphine (M) has been widely used for postoperative pain relief, due to its favorable pharmacokinetic profile, ease of administration during spinal block and low cost. 11 12 However, the use of opioids is associated with undesirable side effects such as nausea, vomiting, pruritus, and urinary retention, which can reduce patient satisfaction. Furthermore, the fact that M may produce severe maternal respiratory depression underscores the importance of investigating alternative opioid-free analgesia approaches. 13 14 Another opioid used for postoperative pain relief is Tramadol. This is a synthetic 4-phenyl-piperidine analogue of codeine with a dual mechanism of action. It stimulates µ receptors and, to a lesser extent, δ and Ϗ receptors. Like tricyclic antidepressants, tramadol also activates spinal pain inhibition by decreasing the reuptake of norepinephrine and serotonin. 15 The blockade of peripheral nerves for analgesia of the abdominal wall after surgery has become more frequent, especially with the development of ultrasound technology. 16 17 Quadratus lumborum (QL) block is a technique used to inject local anesthetic in the posterior abdominal wall around the quadratus lumborum muscle to anesthetize the thoracolumbar nerves. 18 It can provide somatic as well as visceral analgesia due to its paravertebral spread. 19 20 According to a systematic review by Jin et al., 21 QL block significantly reduces opioid requirement in cesarean delivery and in renal surgery. Similarly, Graça et al. 22 have reported a reduction in postoperative opioid consumption with the use of QL block in laparoscopic nephrectomy, while Zhu et al. 23 have found that anterior QL block significantly alleviated pain in patients undergoing open liver resection. The objective of the present study was to compare the efficacy of QL block and intrathecal M for postcesarean delivery analgesia by measuring M/tramadol consumption during the first 48 hours after surgery. Methods The present randomized clinical trial was registered at the Brazilian Clinical Trials Registry (RBR-5RHP9J) and was approved by the Research Ethics Committee of the Universidade Federal de São Paulo (UNIFESP, in the Portuguese acronym) (CEP/UNIFESP.CAAE:83549817.3.0000.5505), was conducted at the Hospital São Paulo, UNIFESP. The inclusion criteria were as follows: pregnant women > 18 years old with gestational age of at least 37 weeks, normal singleton pregnancy, physical status classified according to the American Society of Anesthesiologists as ASAII (mild systemic disease without functional limitations), and ASA III (mild systemic disease with functional limitations), elective cesarean section performed under spinal anesthesia at the Obstetric Center of the Faculdade Paulista de Medicina between June 2019 and December 2019. Exclusion criteria: inability to understand or provide a verbal self-report of pain on a scale, congenital or acquired coagulation disorders, allergy to local anesthetics, anatomical disorders of the spine leading to neuraxial block failure, BMI > 35 kg/m 2 , and local infection. Initially, we defined that the sample size would be for an established period of time. However, during our study, Salama 24 published a study similar to ours. Considering total postoperative morphine consumption as significantly lower in the QL block group than in the M group, as reported by Salama, 24 and assuming a statistical power of 80% at a significance level of 0.5%, a sample size of 4 participants in each group was estimated to be enough to compare efficacy between QL block and intrathecal morphine for postcesarean delivery analgesia on the basis of total postoperative M consumption. Thus, we used a secondary outcome obtained in the study by Salama, 24 total postoperative morphine consumption, to calculate our sample size. However, at that time, we had already collected more patients: 15 to the QL block group and 16 to the M group. So, we decided to present data from all patients studied. Eligible parturients were invited to participate in the study during the preanesthetic visit. Patients who agreed to participate in the research signed an informed consent form. Study participants were randomly allocated into two groups: Group M (spinal anesthesia with bupivacaine and M), and Group QL (spinal anesthesia with bupivacaine + QL block). The randomization was performed using software available at http://github.com/Gear61/Random-Number-Generator (v. 2015). This software generated a numerical sequence of 1 or 2. The patient who was randomized with the number 1 would belong to the M group and the one randomized with the number 2 would belong to the QL group. Before the procedure, the patients did not know which group they belonged to. The study procedures were performed by two anesthesiologists. “Anesthesiologist One” (AO) conducted randomization, filled the syringes with the study medication, and performed the QL block. Anesthesiologist One is a specialist in regional anesthesia, with 5 years of experience in regional blocks. “Anesthesiologist Two” (AT) administered spinal anesthesia but did not know the volume of drugs that each group would receive during spinal anesthesia. Anesthesiologist Two was blind to patient allocation and performed the postoperative assessment. Both groups received 12.5 mg of 0.5% hyperbaric bupivacaine for spinal anesthesia. To group M, 100 mcg of M, whose onset of action occurs in 60 minutes, was added to the syringe containing bupivacaine. To group QL, bilateral QL block was performed by injecting 0.3 mL/kg of 0.2% bupivacaine on each side. Spinal anesthesia was performed in the sitting position at vertebral levels L3-L4 or L4-L5, using a 27-gauge pencil point needle according to the standard hospital protocol. After surgery, AT left the operating room to return 1 hour later to reevaluate the patient. In the QL group, bilateral US-guided QL block was performed by AO using a Sonosite M-Turbo R System with HFL 38x, 5–8 MHz convex transducer (Sonosite, Bothell, WA, USA). With the patient in the lateral decubitus position, the transducer was placed at the anterosuperior iliac spine level and advanced cranially to visualize the three abdominal muscle layers. The external oblique muscle was followed laterally until its posterior border was visualized with the internal oblique muscle underneath, like a roof over the QL muscle. The transducer was directed downwards to identify the middle layer of the thoracolumbar fascia as a bright hyperechoic line. After antisepsis of the anterior abdominal wall with alcoholic chlorhexidine, a 22G x 100 mm needle (AEq. 2250, BMD Group, Venice, Italy) was inserted in-plane from the anteromedial to the poster direction, at an angle of 45 degrees to the skin for the injection of bupivacaine. Thus, all patients received type 2 QL block. The performance of QL block took ∼ 5 minutes. Postoperative analgesia for all patients consisted of dipyrone 1 g every 6 hours, and ketoprofen 100 mg every 12 hours, both intravenous. We followed the postoperative rescue analgesia protocol adopted at the hospital: at any time during the postoperative period, patients could request rescue medication for pain relief. If the patient reported a moderate pain score (4–6), tramadol 100 mg was administered intravenously every 6 hours. If there was no improvement in pain after tramadol or if the patient reported severe pain score (7–10), the rescue medication used was M 4 mg, intravenously, every 6 hours. The AD evaluated the parameters, described below, in all patients at predetermined intervals after surgery (1, 2, 4, 6, 12, 24, and 48 hours); with the main outcome being the consumption of M/tramadol. Morphine/tramadol consumption was measured in milligrams (mg), heart rate in beats per minute (bpm), oxygen saturation in percentage of oxygen carried by the blood (%), noninvasive blood pressure in millimeters of mercury (mmHg). Pain scores were evaluated using the visual analogue pain scale, whose values range from 0 to 10; zero means total absence of pain and 10 the maximum level of pain bearable by the patient, with a score of 1 to 3 considered as mild, moderate from 4 to 6, and severe from 7 to 10. The sedation level was measured as follows: grade 1 = anxious and agitated patient; grade 2 = cooperative, oriented and calm patient; grade 3 = sleepy patient and attentive to commands; grade 4 =patient sleeping, responds quickly to vigorous sound stimulus, grade 5 = patient sleeping, responds slowly to vigorous sound stimulus, and grade 6 = patient sleeping, no response. Pruritus was evaluated as follows: grade 0 = absent, grade 1 = mild, grade 2 = moderate, grade 3 = severe. Nausea was measured as follows: grade 0 = absent, grade 1 = mild, grade 2 = moderate, grade 3 = severe or vomiting. In addition to these parameters, the presence or absence of residual block and other complications was evaluated. Quantitative variables were compared using the parametric Student t -test or the nonparametric Mann-Whitney test. For the comparison of qualitative variables between groups, the chi-squared test, the Fisher exact test or the likelihood ratio test were performed. To compare quantitative variables between groups over time, analysis of variance (ANOVA) with repeated measures or repeated measures ANOVA with rank transform were used. To compare qualitative variables between groups over time, the generalized estimating equation (GEE) model was used. Significance was set at 5% ( p  < 0.05). Results As shown in the diagram below, 44 patients were invited to participate in the study; 13 were not included in the study for the following reasons: 1 refused to participate, 12 met the exclusion criteria; 5 had BMI > 36 kg/m2, 4 were < 37 weeks pregnant, and 3 were with multiple pregnancies. Thus, the study population consisted of 31 parturients. Of these, 15 were assigned to the QL group, and 16 to the M group ( Fig. 1 ). Fig. 1 Study flow diagram. As shown in Table 1 , there was no significant difference between the groups regarding age, BMI, and number of previous cesarean deliveries. Only one patient in the QL group was classified as ASA III, all others in both groups were ASA II, showing the homogeneity of the samples. Table 1 Clinical and demographic characteristics of study patients Variables by group QL M p-value Age (Years old) Mean (SD) 33.5 (6.7) 31.1 (6.4) 0.293 (**) Median (Minimum–Maximum) 37 (19–41) 31 (18–41) Total 15 16 ASA − n (%) II 14 (93.3) 16 (100) not calculated III 1 (6.7) 0 (0) Total 15 (100) 16 (100) BMI Mean (SD) 30.8 (3.9) 30.2 (4.8) 0.669 (*) Median (Minimum–Maximum) 30 (24.5–36) 30.3 (21.6–36) Total 15 16 Previous cesarean delivery − n (%) No 5 (33.3) 10 (62.5) Yes 10 (66.7) 6 (37.5) 0.104 (#) Total 15 (100) 16 (100) Previous cesarean delivery Mean (SD) 1.2 (1.15) 0.81 (1.17) 0.279 (**) Median (Minimum–Maximum) 1 (0–4) 0 (0–3) Total 15 16 Abbreviation: SD. standard deviation. (*) Parametric t-Student test; (**) Mann-Whitney nonparametric test; (#) Chi-squared test. Table 2 shows that most patients in both groups received neither morphine nor tramadol over 48 hours after surgery. Morphine consumption ( Table1 ) was lower in the QL group (33.4%) compared with the M group (43.9%), but no statistical difference was reached. Moreover, there were twice as many patients who used M in group M compared with group QL at 12 hours and 24 hours after cesarean section. In the QL group, tramadol was used by 26.7% of the patients at 2, 6, and 12 hours and by 13.3% at 24 hours. In contrast, the use of tramadol in group M was higher at 24 hours (18.8%) than at 6 and 12 hours (6.3%). Nonetheless, no statistical difference was observed between groups. During the period in which they were evaluated, six patients used both medications: M and tramadol, all of which were in the M group. Table 2 Opioid consumption and urinary retention between groups Time Consumption of tramadol Consumption of morphine Urinary retention (Post- operative)  M group n (%) QL group n (%) p a  M  group n (%)  QL  group n (%) p b M group n (%) QL group n (%) 1 hour 1(6.25) 0 0.059 1(6.25) 1(6.66) 0.631 0 0 2 hours 2(12.5) 4(26.66) 1(6.25) 1(6.66) 0 0 4 hours 3(18.75) 0 1(6.25) 2(13.33) 1(6.25) 0 6 hours 1(6.25) 4(26.66) 1(6.25) 0 4(25) 0 12 hours 1(6.25) 4(26.66) 2(12.5) 1(6.66) 4(25) 0 24 hours 3(18.75) 2(13.33) 2(12.5) 1(6.66) 4(25) 1(6.66) 48 hours 0 1(6.66) 0 0 1(6.25) 0 Abbreviations: M group, morphine group. QL group: quadratus lumborum group. a Generalized estimating equations model (GEE). b Mann-Whitney nonparametric test. As shown in Fig. 2 , the pain scores significantly differed between groups ( p  = 0.002) independently of time ( p  = 0.162). Median pain scores and/or pain variation, despite being higher in group M than in group QL. Fig. 2 Distribution of pain score over time among patients in the quadratus lumborum and morphine groups. M: Morphine Group, QL: Quadratus Lumborum Group. p-value = 0.002. Model of Analysis of Variance (ANOVA). Pruritus was not observed in any of the QL patients. On the other hand, in group M, pruritus was present in half of the women after 4 hours, in nearly 70% after 12 hours, and was still present in 12.6% of them 48 hours after surgery. Nausea was not seen in any of the women in the QL group but was present from the first hour to 24 hours postoperatively in 12 to 20% of the participants in group M. Urinary retention occurred in only 1 woman in the QL group, whereas in the M group it occurred in 25% of the patients from 6 hours to 24 hours after cesarean section. No patient had respiratory depression in any of the groups. Residual block was present in only one patient in the QL group. Heart rate, respiratory rate, systolic blood pressure, diastolic blood pressure, oxygen saturation, and sedation score did not significantly differ between groups. All patients in the study had a grade 1 sedation score, that is, they remained cooperative, oriented, and calm. Discussion The present study demonstrated that QL block and intrathecal M injection are effective in providing postoperative analgesia in patients undergoing cesarean section as there was no significant difference in opioid consumption in 48 hours. In the QL group, pain scores were significantly lower and side effects such as pruritus, nausea, and vomiting were not observed. Our results show that QL block and intrathecal morphine can effectively relief postoperative pain. Indeed, 86% of the patients in both groups did not require the administration of opioids. Quadratus lumborum block type 2 was the technique of choice in the present study because it was demonstrated by Blanco et al. 25 to be superior to transversus abdominis plane blocks (TAP blocks) in providing postoperative analgesia. However, a study by Kang et al. 26 comparing the effects of epidural M and major QL block approaches showed that the combination of QL block type 2 and 3 can provide superior postcesarean analgesic effect. Morphine and tramadol consumption did not differ between the study groups ( Table 2 ). Morphine consumption 6 hours after cesarean section was lower in the QL group, suggesting that QL block has a longer lasting effect. However, this finding did not reach statistical significance, in opposition to Salama, 24 who observed a significant lower morphine consumption in the QL group. Our results also differ from those of Kang et al., 26 who reported significantly higher morphine consumption with QL block compared with peridural morphine. These diverging results may be explained by differences in the local anesthetic dose and volume used. While higher concentrations (0.375% ropivacaine) were used by Salama, 24 the participants of the present study received the same concentration used by Kang et al. 26 (0.2% ropivacaine) and Blanco et al. 19 (0.2% bupivacaine). Furthermore, anesthetic volumes also differed; Salama 24 used 24 ml on each side of the block, whereas the volume adopted in the present study, as well as by Blanco et al. 19 was 0.3 ml/kg, and that used by Kang et al. 26 was 30 ml on each side. Pain scores over 48 hours among our patients were significantly lower in the QL group ( Fig. 2 ), indicating that QL block was effective as an anesthetic technique. The spreading of QL block into the paravertebral space 24 and into thoracic and lumbar sympathetic nerves 27 seems to be the major mechanism of action of this anesthetic approach and might explain the lower pain scores found in our study. Pain intensity and elective cesarean section have been associated with a negative birth experience, 28 and are related to post-traumatic stress symptoms, and postpartum depressive symptoms. 29 Within this framework, QL block stands as an effective alternative, given that it not only provides analgesia but is also free of undesirable side effects that could render the experience of childbearing more negative. The incidence of pruritus, nausea, and vomiting in the postoperative period was higher in the M group. As a matter of fact, these symptoms were not seen in the QL group. Urinary retention was more frequent in the M group than in the QL group, which had only one patient with this symptom. However, this difference was not statistically significant. No case of respiratory depression was observed in any of the study participants. Residual block was seen in only one patient of the QL group up to 12 hours after cesarean section. Kang et al. 26 also described this event in two patients undergoing QL block. It is possible that a posterior dispersion of the local anesthetic occurred, and therefore the QL block behaved as type 3, a complication that has been previously described. 30 Hemodynamic parameters were similar in both study groups, which did not differ regarding heart rate, respiratory rate, systolic and diastolic blood pressure, and oxygen saturation. It is noteworthy that the groups of patients herein investigated were homogeneous, and that the same investigator performed all blocks. However, the study had some limitations. Obese patients with BMI ≥ 35 kg/m 2 were not included. It was not possible to install a patient-controlled analgesia pump, as is done in large centers, due to the infrastructure of the institution, so we chose to use the rescue analgesia protocol adopted in our hospital, which includes the use of tramadol for moderate pain scores. This setback did not affect the progress of the research or the results we arrived at, allowing us to proceed with the research. Conclusion The QL block can be seen as a valuable option for those patients with a previous history of nausea, vomiting, and itching. Perhaps performing the quadratus lumborum block with a greater volume and concentration of local anesthetic can provide analgesia for a period longer than 48 hours. In brief, both QL block and intrathecal M were demonstrated to be effective for postcesarean section analgesia. Nonetheless, QL block seemed to be more advantageous, given that it was associated with lower postoperative pain scores and absence of pruritus, nausea, and vomiting. Contributors Conflict of Interests The authors have no conflict of interests to declare. All authors participated in the concept and design of the present study; analysis and interpretation of data; draft or revision of the manuscript; and they have approved the manuscript as submitted. All authors are responsible for the reported research. ==== Refs References 1 Betrán A P Merialdi M Lauer J A Bing-Shun W Thomas J Look P V Rates of caesarean section: analysis of global, regional and national estimates Paediatr Perinat Epidemiol 2007 21 02 98 113 10.1111/j.1365-3016.2007.00786.x 17302638 2 Betran A P Ye J Moller A B Souza J P Zhang J Trends and projections of caesarean section rates: global and regional estimates BMJ Glob Health 2021 6 06 e005671 10.1136/bmjgh-2021-005671 3 Eisenach J C Pan P H Smiley R Lavand'homme P Landau R Houle T T Severity of acute pain after childbirth, but not type of delivery, predicts persistent pain and postpartum depression Pain 2008 140 01 87 94 10.1016/j.pain.2008.07.011 18818022 4 Wong C A Girard T Undertreated or overtreated? Opioids for postdelivery analgesia Br J Anaesth 2018 121 02 339 342 10.1016/j.bja.2018.05.061 30032870 5 Verma K Malawat A Jethava D Jethava D D Comparison of transversus abdominis plane block and quadratus lumborum block for post-caesarean section analgesia: A randomised clinical trial Indian J Anaesth 2019 63 10 820 826 10.4103/ija.IJA_61_19 31649394 6 Kehlet H Dahl J B Anaesthesia, surgery, and challenges in postoperative recovery Lancet 2003 362 (9399):1921 1928 10.1016/S0140-6736(03)14966-5 14667752 7 Kerai S Saxena K N Taneja B Post-caesarean analgesia: What is new? Indian J Anaesth 2017 61 03 200 214 10.4103/ija.IJA_313_16 28405033 8 Yurashevich M Carvalho B Butwick A J Ando K Flood P D Determinants of women's dissatisfaction with anaesthesia care in labour and delivery Anaesthesia 2019 74 09 1112 1120 10.1111/anae.14756 31264207 9 Allegri M Clark M R De Andrés J Jensen T S Acute and chronic pain: where we are and where we have to go Minerva Anestesiol 2012 78 02 222 235 22095106 10 Niklasson B Georgsson Öhman S Segerdahl M Blanck A Risk factors for persistent pain and its influence on maternal wellbeing after cesarean section Acta Obstet Gynecol Scand 2015 94 06 622 628 10.1111/aogs.12613 25714852 11 Palmer C M Emerson S Volgoropolous D Alves D Dose-response relationship of intrathecal morphine for postcesarean analgesia Anesthesiology 1999 90 02 437 444 10.1097/00000542-199902000-00018 9952150 12 Rathmell J P Pino C A Taylor R Patrin T Viani B A Intrathecal morphine for postoperative analgesia: a randomized, controlled, dose-ranging study after hip and knee arthroplasty Anesth Analg 2003 97 05 1452 1457 10.1213/01.ANE.0000083374.44039.9E 14570664 13 McMorrow R C Ni Mhuircheartaigh R J Ahmed K A Aslani A Ng S-C Conrick Martin I Comparison of transversus abdominis plane block vs spinal morphine for pain relief after Caesarean section Br J Anaesth 2011 106 05 706 712 10.1093/bja/aer061 21498494 14 Kanazi G E Aouad M T Abdallah F W Khatib M I Adham A MB Harfoush D W The analgesic efficacy of subarachnoid morphine in comparison with ultrasound-guided transversus abdominis plane block after cesarean delivery: a randomized controlled trial Anesth Analg 2010 111 02 475 481 10.1213/ANE.0b013e3181e30b9f 20488929 15 Kazuhiko F Analgésicos opioides Rio de Janeiro Elsevier 2019 903 16 Ripollés J Marmaña Mezquita S Abad A Calvo J Eficácia analgésica do bloqueio ecoguiado do plano transverso do abdome - revisão sistemática Rev Bras Anestesiol 2015 65 04 255 280 10.1016/j.bjan.2013.10.014 26145531 17 Abrahams M S Horn J L Noles L M Aziz M F Evidence-based medicine: ultrasound guidance for truncal blocks Reg Anesth Pain Med 2010 35 (2, Suppl)S36 S42 10.1097/AAP.0b013e3181d32841 20216023 18 Blanco R Tap block under ultrasound guidance: the description of a “no pops technique” Reg Anesth Pain Med 2007 32 01 130 17350524 19 Blanco R Ansari T Girgis E Quadratus lumborum block for postoperative pain after caesarean section: A randomised controlled trial Eur J Anaesthesiol 2015 32 11 812 818 10.1097/EJA.0000000000000299 26225500 20 Mieszkowski M M Mayzner-Zawadzka E Tuyakov B Mieszkowska M Żukoski M Waśniewski T Evaluation of the effectiveness of the Quadratus Lumborum Block type I using ropivacaine in postoperative analgesia after a cesarean section - a controlled clinical study Ginekol Pol 2018 89 02 89 96 10.5603/GP.a2018.0015 29512813 21 Jin Z Liu J Li R Gan T J He Y Lin J Single injection Quadratus Lumborum block for postoperative analgesia in adult surgical population: A systematic review and meta-analysis J Clin Anesth 2020 62 109715 10.1016/j.jclinane.2020.109715 31968297 22 Graça R Miguelez P Cardoso J M Sá M Brandão J Pinheiro C [Continuous quadratus lumborum type II block in partial nephrectomy] Braz J Anesthesiol 2018 68 06 653 656 10.1016/j.bjan.2018.03.001 29784431 23 Zhu Q Li L Yang Z Shen J Zhu R Wen Y Ultrasound guided continuous Quadratus Lumborum block hastened recovery in patients undergoing open liver resection: a randomized controlled, open-label trial BMC Anesthesiol 2019 19 01 23 10.1186/s12871-019-0692-z 30777027 24 Salama E R Ultrasound-guided bilateral quadratus lumborum block vs. intrathecal morphine for postoperative analgesia after cesarean section: a randomized controlled trial Korean J Anesthesiol 2020 73 02 121 128 10.4097/kja.d.18.00269 30852882 25 Blanco R Ansari T Riad W Shetty N Quadratus lumborum block versus transversus abdominis plane block for postoperative pain after cesarean delivery: a randomized controlled trial Reg Anesth Pain Med 2016 41 06 757 762 10.1097/AAP.0000000000000495 27755488 26 Kang W Lu D Yang X Zhou Z Chen X Chen K Postoperative analgesic effects of various quadratus lumborum block approaches following cesarean section: a randomized controlled trial J Pain Res 2019 12 2305 2312 31413627 27 Elsharkawy H El-Boghdadly K Kolli S Esa W AS DeGrande S Soliman L M Injectate spread following anterior sub-costal and posterior approaches to the quadratus lumborum block: A comparative cadaveric study Eur J Anaesthesiol 2017 34 09 587 595 10.1097/EJA.0000000000000680 28731927 28 Nystedt A Hildingsson I Women's and men's negative experience of child birth-A cross-sectional survey Women Birth 2018 31 02 103 109 10.1016/j.wombi.2017.07.002 28789827 29 Gosselin P Chabot K Béland M Goulet-Gervais L Morin A J [Fear of childbirth among nulliparous women: Relations with pain during delivery, post-traumatic stress symptoms, and postpartum depressive symptoms] Encephale 2016 42 02 191 196 10.1016/j.encep.2016.01.007 26924001 30 Carline L McLeod G A Lamb C A cadaver study comparing spread of dye and nerve involvement after three different quadratus lumborum blocks Br J Anaesth 2016 117 03 387 394 10.1093/bja/aew224 27543534
PMC009xxxxxx/PMC9800152.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36580942 10.1055/s-0042-1759742 RBGO-22-0229 Review Article Assessment of Pelvic Floor Disorders due to the Gestational Diabetes Mellitus Using Three-Dimensional Ultrasonography: A Narrative Review Avaliação das desordens do assoalho pélvico decorrentes do diabetes gestacional usando a ultrassonografia tridimensional: Uma revisão narrativahttp://orcid.org/0000-0003-4710-3392 Sartorão Filho Carlos Izaias 12 http://orcid.org/0000-0003-0817-9511 Barbosa Angélica Mércia Pascon 1 http://orcid.org/0000-0003-4761-4336 Calderon Iracema de Mattos Paranhos 1 http://orcid.org/0000-0002-9227-832X Rudge Marilza Vieira Cunha 1 1 Department of Gynecology and Obstetrics, Faculdade de Medicina de Botucatu, Universidade Estadual de São Paulo (UNESP), Botucatu, SP, Brazil 2 Department of Medical School, Fundação Educacional do Município de Assis (FEMA), Assis, SP, Brazil Address for correspondence Marilza Vieira Cunha Rudge Distrito de Rubião Jr s/n 18.618-000, Botucatu, [email protected] 29 12 2022 12 2022 1 12 2022 44 12 11341140 26 7 2022 21 9 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Gestational diabetes mellitus (GDM)is an entity with evolving conceptual nuances that deserve full consideration. Gestational diabetes leads to complications and adverse effects on the mother's and infants' health during and after pregnancy. Women also have a higher prevalence of urinary incontinence (UI) related to the hyperglycemic status during pregnancy. However, the exact pathophysiological mechanism is still uncertain. We conducted a narrative review discussing the impact of GDM on the women's pelvic floor and performed image assessment using three-dimensional ultrasonography to evaluate and predict future UI. Resumo O diabetes gestacional (DG)é uma entidade com nuances conceituais em evolução que merecem total consideração. O DG leva a complicações e efeitos adversos na saúde da mãe e do bebê durante e após a gravidez. As mulheres também apresentam maior prevalência de incontinência urinária (IU) relacionada ao estado hiperglicêmico durante a gravidez. No entanto, o mecanismo fisiopatológico exato ainda é incerto. Realizamos uma revisão narrativa discutindo o impacto do DG no assoalho pélvico das mulheres e utilizamos o exame de ultrassonografia tridimensional para avaliar e predizer a ocorrência de IU. Keywords diabetes gestational pelvic floor pelvic floor disorders urinary incontinence ultrasonography Palavras-chave diabetes gestacional assoalho pélvico desordens do assoalho pélvico incontinência urinária ultrassonografia ==== Body pmcIntroduction We performed a narrative review of the literature with the intention of summarizing a qualitative interpretation of the prior knowledge towards gestational diabetes mellitus (GDM), the implication of the hyperglycemia status to the pelvic floor, and the outcome of postpartum urinary incontinence (UI). Moreover, we also evaluated the importance of the pelvic floor assessment using three-dimensional ultrasonography. In the last few decades, the Diamater Research Group, located at Faculdade de Medicina de Botucatu at Universidade Estadual de São Paulo, has been studying these important physiopathological mechanisms and assessment tools related to pelvic floor disorders. Their primary goal is to synthesize the extent of the body of knowledge regarding these particular research topics. We selected studies that support the critical findings in these areas. Hence, using the method of narrative review, we did not intend to formally assess the quality or the risk of bias in the literature provided. Gestational Diabetes Mellitus: An Entity with Evolving Conceptual Nuances Gestational diabetes mellitus is defined as hyperglycemia first detected during pregnancy, with glycemic blood levels that do not meet the diagnostic criteria for diabetes mellitus (DM). 1 It differs from diabetes mellitus (DM) diagnosed during pregnancy, also called overt diabetes, which is when women, without a prior diagnosis, have hyperglycemia detected during pregnancy and present blood glycemic levels that meet the World Health Organization (WHO) criteria for DM in the absence of pregnancy. 1 Brazil has high rates of DM in the adult population, with an estimated total of 14.3 million people aged 20 to 79 years. The population estimated prevalence of hyperglycemia during pregnancy in Brazil is approximately 18%, using the diagnostic criteria currently proposed in the literature. 2 The International Association of Diabetes and Pregnancy Study Groups (IADPSG) defines that if the pregnant woman presents, in the first prenatal consultation, diagnostic criteria equal to those predetermined for the diagnosis of diabetes outside pregnancy (glycated hemoglobin ≥ 6.5%; fasting glycemia ≥ 126 mg/dL, or glycemia at any time ≥ 200 mg/dL), she will be considered as a carrier of previous DM or overt diabetes, diagnosed in pregnancy. 3 It also defines that the GDM diagnosis should be established when fasting glucose is ≥ 92 mg/dL and < 126 mg/dL. Alternatively, at least one of the values of the oral glucose tolerance test with 75 g (75g-OGTT), performed between 24 and 28 weeks of gestational age, is ≥ 92 mg/dL at fasting; ≥ 180 mg/dL in the 1st hour; and ≥ 153 mg/dL in the 2nd hour. The 75g-OGTT is universally recommended for all pregnant women who did not present previous DM or hyperglycemia at the beginning of pregnancy. 4 The HAPO study determined the cutoff points of the 75g-OGTT because they corresponded to an increase in the odds ratio of 1.75 for one of the following neonatal outcomes studied: birth weight above the 90th percentile, percentage of neonatal body fat above the 90th percentile, or C-peptide value in the umbilical cord above the 90th percentile. Thus, pregnant women with one or more points in the 75g-OGTT would have a 75% higher risk of having a newborn with one of these three neonatal outcomes when compared to pregnant women without any of these altered values. 5 6 Given the need to move towards a single criterion for the diagnosis of GDM, the WHO adopted the IADPSG. Two warnings were inserted: 1) that these criteria were valid for any gestational age, and 2) that the blood glucose value of 2 hours of 75g-OGTT should be between 153 and 199 mg/dL for the diagnosis of GDM since values ≥ 200 mg/dL correspond to the diagnosis of DM. 7 8 In prenatal routine, fasting glucose is recommended up to 20 weeks of gestational age to diagnose GDM and overt diabetes. All pregnant women with fasting glucose below 92 mg/dL should perform 75g-OGTT from 24 to 28 weeks. If the onset of prenatal care is delayed after 20 weeks of gestational age, 75g-OGTT should be completed as soon as possible. 1 Gestational diabetes mellitus leads to complications and adverse effects on the mother's and infant's health during pregnancy. In addition, in the immediate postpartum period, it can delay the onset of breastfeeding and affect the health of the woman and the infant. 9 Women diagnosed with GDM in the first half of pregnancy represent a high-risk subgroup for increased obstetric and clinical complications. 10 11 Women with GDM have a higher chance of recurrence of GDM in future pregnancies and also a higher risk of developing type 2 DM (T2DM) throughout life. Those with obesity or who require insulin for glycemic control during pregnancy have a higher risk of T2DM. Insulin resistance is the pathophysiological basis of both GDM and T2DM and can be addressed with measures that lead to increased insulin sensitivity, such as nutritional adequacy, exercise, and medications. These interventions reduce the risk of T2DM in high-risk women, such as those with a previous history of GDM. 12 Gestational Diabetes and Postpartum Urinary Incontinence: A Neglected but Common Association Urinary incontinence (UI) is defined by the International Continence Society as any involuntary loss of urine. 13 It is associated with patients' physical, psychological, and social discomforts. In addition, there are well-established risk factors for UI, including advanced age, obesity, and vaginal delivery. 14 A systematic review and meta-analysis conducted by Tähtinen et al., 14 in 2016, reported that vaginal delivery is associated with almost twice as long-term UI, an increase of about 8% when compared to cesarean delivery. Gyhagen et al. 15 conducted a national cohort study in Sweden to investigate UI's prevalence and risk factors 20 years after a vaginal delivery or cesarean section. The study population consisted of 5,236 women who returned the questionnaire by mail, primiparous with a single pregnancy, had vaginal or cesarean delivery between 1985 and 1988, and had no later births. The prevalence of UI was higher after vaginal delivery (40.3%) than after cesarean section (28.8%); odds ratio (OR) 1.67; 95% confidence interval (CI) 1.45–1.92. In addition, there was an 8% increase in UI risk for each unit of body mass index (BMI) plus, and maternal age at delivery increased the risk of UI by 3% each year. The weakening of the pelvic floor muscle (PFM) causes hypermobility of the bladder neck, and urethra, leading to the woman's incompetence of the urethral sphincter. Pregnancy itself is a significant risk factor for UI. The exact causes associated with pregnancy remain not fully understood. 16 During pregnancy, UI is more frequent as pregnancy progresses, compromising women's quality of life. There are few publications on the prevalence of the pregnancy-specific UI (PS-UI). In addition, little is known about the clinical implications regarding the time of onset of UI during pregnancy, and the factors involved in its pathophysiology remain unexplored. 17 A study conducted in Norway by Wesnes et al. 18 described a cumulative incidence of 46% of pregnant women with UI, and multiparity was the main and most relevant risk factor. In a large cohort study with 81,845 women to assess the association between type 2 DM and UI, development risk was higher in diabetic women. 19 Women with GDM also have a higher prevalence of UI. However, the exact pathophysiological mechanism is still uncertain. Nevertheless, weight gain, obesity, fetal macrosomia, and any conditions that increase bladder pressure and urethral mobility may be implicated. In addition, hyperglycemia can cause polyuria and detrusor instability. Hence, the risk of UI is higher during pregnancy and persists after childbirth. 20 Kim et al. 20 examined the prevalence of UI among women with GDM. They found that 49% of women reported urine loss during pregnancy, and 50% reported UI in the first 5 years after delivery. 20 Chuang et al., 21 in a survey of 6,653 women with GDM, described that incontinent women who had GDM had a higher severity of UI 2 years after delivery. Thus, they conclude that GDM is an independent risk factor for postpartum UI, with an essential impact on the severity of symptoms. 21 A pioneer cross-sectional study in Brazil conducted by Barbosa et al. 22 with 832 selected women evaluated the prevalence of 2-years postpartum UI and found it was 18.9% after cesarean section and 17% after vaginal delivery, with no statistical differences between delivery routes. Women who had increased weight gain during pregnancy were at increased risk for PF dysfunction during pregnancy. Women with GDM had a significantly higher UI prevalence 2-years postpartum (OR: 8.6, 95% CI: 3.0–24.3). 22 The Pelvic Floor The PF deep muscles consist of the levator ani muscle (LAM), formed by the puborectalis, pubococcygeus, and iliococcygeus muscles. The superficial muscles of the PF form the urogenital diaphragm and include the cavernous ischium, spongy bulb, and the superficial transverse muscle of the perineum. Fascia interposes these muscles continued with the pelvic endofascia, which involves the pelvic viscera, and contributes to the PF support. 23 The LAM has a tapered shape, with a central slit through the urethra, vagina, and anus. The puborectalis part is the lowest and is placed in the lower branches of the pubis and later borders the anal canal. The function of the PF muscles is to make voluntary and involuntary contractions, responsible for urinary and fecal continence. The puborectalis portion of LAM is essential in supporting and conserving continence. 24 Pelvic floor disorders (PFDs), including UI, genital prolapse, and anal incontinence, are highly prevalent in women of all ages. Imaging evaluation methods are essential for diagnosing and treating diseases and studying the integrity of pelvic structures. Many imaging modalities are used to evaluate the PF, such as computed tomography (CT), magnetic resonance imaging (MRI), and contrast-enhanced defecography. 25 As limitations, CT and defecography use contrasts and employ ionizing radiation. Defecography can replicate and evaluate patients' symptoms in real-time during defecation. However, it employs X-ray, it is unpleasant for the patient, and it is challenging to reproduce. Magnetic resonance imaging provides a good evaluation of the soft tissues of the PF without the use of ionizing radiation. However, it requires the use of contrasts, and it is not operator-dependent. In addition, the high cost of the examination, the prolonged time of image acquisition, and the difficulty of availability of the device limit its use in practice and are disadvantages. More importantly, MRI does not allow the proper evaluation of the functional maneuvers of the PF. 26 Ultrasound (US) imaging is widely used for morphological and functional evaluation of the PF. Studies have demonstrated the superiority of the US in conjunction with clinical evaluation compared with manometry, electromyography, and defecography. It is also helpful as a biofeedback tool for functional PF training. Anal ultrasound of the PF is beneficial for evaluating the anal sphincter and diagnosing fecal incontinence. It uses specific high-frequency transducers that ideally reproduce a 360-degree panoramic image to visualize the anal sphincter complex properly. Transvaginal US for PF evaluation employs identical transducers to study internal pelvic structures. However, it has limited use due to the very close proximity to the PF structures, the possible interference of the transducer in the functional evaluation, and especially the limited angle of insonation provided by the transducer for good acquisition of the images. 27 The transperineal, or translabial, US was one of the first ultrasound modalities used to study the PF. It is a handy and widely available tool, of low cost, little invasive, and easy to reproduce. It also allows the evaluation of the structures of the PF during functional maneuvers. The three-dimensional US (US3D) of the PF is a technique described more than 20 years ago which has gained more notoriety in recent years. It reproduces three-dimensional images of the PF similarly to those obtained by MRI, with the advantages of being more widely available and more affordable, being faster and mainly allowing clinical and functional evaluation of the patient in real-time. In addition, it does not cause more significant discomfort for the patient and does not require contrasts for its execution. It is less user-dependent than the two-dimensional US, which contributes to greater accuracy of the evaluations and measurements of the PF. The 3DUS provides an adequate and reliable assessment of the anatomy and function of muscles and structures, essential in the clinical and complementary diagnosis, treatment, and follow-up of PF disorders. Therefore, many authors argue that the PF 3DUS should be routinely used to evaluate and manage diseases and dysfunctions. In addition to assessing the structures and functions of LAM, imaging methods are essential to exclude coexisting diseases, propose individualized treatments according to the findings, monitor therapy, better understand therapeutic failures, and support the study of conditions that can cause damage to the LAM. 28 The comparison between MRI and 3DUS is frequently found in the literature, and the benefits and practicality of using the PF 3DUS are well established and validated. 29 The etiology of PF disorders is multifactorial. Traumatic damage to support structures during labor and vaginal delivery may be important factors contributing to UI and genital prolapse development. 30 In 2009, Shek and Dietz, 31 using 3DUS from the PF before and after delivery, concluded that vaginal delivery results in enlargement of the hiatal area (HA), especially after LAM avulsion. However, even without macroscopic alteration of the muscle, there may be greater distensibility of the HA, which may be related to other mechanisms. 31 The same authors, in 2010, conducted a prospective longitudinal study in 468 nulliparas in the 3rd trimester and after delivery to determine whether the prediction of trauma to the LAM is feasible with the 3DUS, without success. They concluded that prediction is very difficult or even impossible. 32 In a prospective longitudinal study in Germany, in 2013, Falkert et al. 33 used 3DUS after immediate delivery and 18 to 24 months postpartum. The objective was to evaluate whether the changes observed in the PF after prompt delivery persisted after 18 to 24 months. A total of 59% of women completed the follow-up, and a significant increase in HA was observed in vaginal postpartum compared to cesarean section. However, there were no significant UI changes between the vaginal and cesarean groups. Independently of the mode of delivery, the UI incidence was higher in the larger HA group. 33 A study conducted in Brazil, in 2013, by Araujo Júnior et al. 34 at Universidade Federal de São Paulo evaluated the changes in the 3DUS of the PF of primiparous women with different delivery modes. They demonstrated higher HA in the postpartum vaginal group and forceps about cesarean delivery. 34 Chan et al., in 2013, 35 investigated PF biometrics during pregnancy and its correlation with symptoms of PF disorders in each trimester of pregnancy. The HA significantly increased by 15.1 +/- 24.8% at rest and 24.7 +/- 28.5% at Valsalva from the first to the third trimester. Symptoms of UI, bladder neck descent, and prolapse were associated with increased HA. 35 Siafarikas et al. 36 investigated the association between the PF dimensions at the end of the pregnancy with the second stage of labor duration and the type of delivery. In conclusion, they found a significant association between HA and the shorter duration of the active phase of the second delivery stage and expected vaginal delivery. However, the process of parturition is highly complex, and the pelvic anatomy is only an influencing factor. The clinical findings are inconclusive in determining the risk predictors for dystocic or instrumentalized deliveries. 36 In contrast, Van Veelen et al. 37 showed that smaller HA dimensions during the contraction of the LAM in the first pregnancy were associated with instrumentalized or cesarean delivery. In another study, Van Veelen et al. 38 demonstrated that the HA values and the contractility and distensibility of the LAM increase during the first pregnancy. Thus, regardless of the type of delivery, this more significant distension of the HA persists after birth and may be related to future PF dysfunctions in the woman's life. 38 Staer-Jensen et al. 39 studied the morphological changes of the PF in a cohort of primiparous women. In conclusion, the LAM can recover after pregnancy and delivery, although not all women recover from the levels demonstrated during pregnancy. 39 Siafarikas et al. (2013) 40 published a study on the learning process to perform and analyze the images of the PF 3DUS. They concluded that the exam can be learned quickly and that the technique is reliable. 40 The publication addresses the length of the learning process for multiple measures of hiatal functional anatomy, showing that the measurement of all assessed hiatal dimensions could be taught to an acceptable standard within 23 hours of total training, confirming several other studies demonstrating good repeatability of levator hiatal dimensions. We published a study by Sartorão Filho et al. 41 that evaluated the PF biometry using 3DUS at 2 time points of gestation in pregnant women with GDM. We performed a prospective cohort study at the Perinatal Diabetes Research Center, including 44 pregnant women with GDM and 66 pregnant women without GDM at 24 to 28 weeks of gestation. The minimal hiatal dimensions plane was used to determine the HA biometry at 24 to 28 and 34 to 38 weeks of pregnancy by 3DUS. Of a 110 pregnant women, 100 (90.9%) completed the follow-up. The 3DUS measurements showed a negative biometric change between the 2 time points in pregnancy in women with GDM; in the HA (β coefficient: estimative of effect in biometric progression according to GDM diagnosis, using the non-GDM group as reference = − 6.76; P  = .020), anteroposterior diameter (β = − 5.07; P  = .019), and levator ani thickness (β = − 12.34; P  = 0.005). Pregnant women with GDM had a significantly lower than expected percentage of changes in biometry of levator ani thickness and HA from 24 to 28 to 34 to 38 weeks of gestation when compared with the group of pregnant women with non-GDM. Thus, GDM altered the biometric morphology of PF structures assessed by the 3DUS. This reported complication may be implicated in adverse birth outcomes and may play a role in developing PF dysfunction. 41 The Pelvic Floor 3DUS Exam Technique The US3D biometry data of the PF used by the Diamater study group were anteroposterior diameter, transversal diameter, and HA, collected at rest, during maximum contraction, and maximum Valsalva maneuver. Women were positioned in the lithotomy position after voiding. The equipment used was the GE P8 or the GE Voluson i system with a 2-to-6 MHz curved array three-dimensional transducer (GE Healthcare, Zipf, Austria). We acquired the volume angle setting maximum in the sagittal plane and 85° in the coronal plane. Offline analysis of the rendered volume datasets was blinded using the 4D View (GE Healthcare) software program. Finally, we used the method proposed by Dietz, 27 obtaining the image of the three orthogonal planes as seen in Figure 1 . Fig. 1 Image of the three orthogonal planes: A - Mid-sagittal plane, B - Coronal plane, C - Axial plane, D - Axial plane with the rendered image. Figure 2 demonstrates the levator hiatal dimensions, measured in the axial plane of minimal levator hiatal distances, identified in the mid-sagittal image as the minimal distance between the inferior margin of the symphysis pubis and the anorectal junction. The anteroposterior diameter of the levator hiatus was defined as the minimum distance in a mid-sagittal direction and was measured from the symphysis pubis' inferior border to the levator's posterior margin ani. The levator hiatal transverse diameter was measured at its widest part from the internal border of the levator ani muscle, perpendicular to the anteroposterior diameter. The levator hiatus area was measured as the internal area bordered by the LAM, pubic symphysis, and the inferior pubic ramus. 27 The LAM's thickness is another possible measurement, as shown in Figure 1 , although we did not perform or consider it for our research. The transperineal 3DUS learning process is reliable, repeatable, and practical. Thus, it should be incorporated into the modern arsenal of PF evaluation. Fig. 2 Axial plane of pelvic floor: 1 - anteroposterior diameter, 2 - transverse diameter, and 3 - Levator ani muscle thickness 4 - hiatal area. Acknowledgments We are very thankful for the support from the Diamater Research Group, led by Professor Emeritus Marilza Vieira Cunha Rudge. Diamater Study Group - Rudge M. V. C., Barbosa, A. M. P., Caldeiron I. M. P., Souza F. P., Berghmans B., de Bie R., Thabane L., Junginger B., Graeff C. F. O., Magalhães C. G., Costa R. A., Lima S. A. M., Kron-Rodrigues M. R., Felisbino S., Barbosa W., Campos, F. J., Bossolan G., Corrente J. E., Nunes H. R. C. Abbade J., Rossignoli P. S., Pedroni C. R., Atallah A. N., Di Bella Z. I. K. J., Uchoa S. M. M., Duarte M. A. H., Mareco E. A., Sakalem M. E., Martinho N., Hallur L. S. R., Reyes D. R. A., Alves F. C. B., Marcondes J. P. C., Prudencio C. B., Pinheiro F. A., Sartorao Filho C. I., Quiroz S. B. C. V., Pascon T., Nunes S. K., Catinelli B. B., Reis F. V. D. S., Oliveira R. G., Barneze S., Enriquez E. M. A., Takano L., Carr A. M., Magyori A. B. M., Iamundo L. F., Carvalho C. N. F., Jacomin M., Avramidis R. E., Silva A. J. B., Orlandi M. I. G., Dangió T. D., Bassin H. C. M., Melo J. V. F., Takemoto M. L. S., Menezes M. D., Calderon T. D., Santos N. J., Lourenço I. O., Marostica de Sá J., Caruso I. P., Rasmussen L. T., Garcia G. A., Nava G. T. A., Pascon C., Bussaneli D. G., Nogueira V. K. C., Rudge C. V. C., Piculo F., Prata G. M. Financial Support Conflict of Interests The authors have no conflict of interests to declare. The São Paulo Research Foundation (FAPESP) #2016/09710-9 awarded the corresponding author financial support for this research. ==== Refs References 1 Organização Pan-Americana da Saúde Ministério da Saúde. Federação Brasileira das Associações de Ginecologia e ObstetríciaSociedade Brasileira de Diabetes. Rastreamento e diagnóstico de diabetes mellitus gestacional no Brasil [Internet]. Brasília (DF): OPAS; 2017 [cited 2021 Jul 22]. Available from:https://iris.paho.org/bitstream/handle/10665.2/34278/9788579671180-por.pdf?sequence=1&isAllowed=y 2 Trujillo J Vigo A Duncan B B Falavigna M Wendland E M Campos M A Schmidt M I Impact of the International Association of Diabetes and Pregnancy Study Groups criteria for gestational diabetes Diabetes Res Clin Pract 2015 108 02 288 295 10.1016/j.diabres.2015.02.007 25765668 3 McIntyre H D Colagiuri S Roglic G Hod M Diagnosis of GDM: a suggested consensus Best Pract Res Clin Obstet Gynaecol 2015 29 02 194 205 10.1016/j.bpobgyn.2014.04.022 25242583 4 Laafira A White S W Griffin C J Graham D Impact of the new IADPSG gestational diabetes diagnostic criteria on pregnancy outcomes in Western Australia Aust N Z J Obstet Gynaecol 2016 56 01 36 41 10.1111/ajo.12394 26293845 5 International Association of Diabetes & Pregnancy Study Groups (IADPSG) Consensus Panel Writing Group and the Hyperglycemia & Adverse Pregnancy Outcome (HAPO) Study Steering Committee. The diagnosis of gestational diabetes mellitus: new paradigms or status quo? J Matern Neonatal Med. 2012 25 12 2564 2569 10.3109/14767058.2012.7180022 6 HAPO Study Cooperative Research Group Lowe L P Metzger B E Dyer A R Lowe J McCance D R Lappin T RJ Hyperglycemia and Adverse Pregnancy Outcome (HAPO) Study: associations of maternal A1C and glucose with pregnancy outcomes Diabetes Care 2012 35 03 574 580 10.2337/dc11-1687 22301123 7 Hod M Kapur A Sacks D A Hadar E Agarwal M Di Renzo G C FIGO initiative on gestational diabetes mellitus: a pragmatic guide for diagnosis, management, and care Int J Gynaecol Obstet 2015 131 03 S173 S211 10.1016/S0020-7292(15)30033-3 8 Kapur A Mahmood T Hod M FIGO's response to the global challenge of hyperglycemia in pregnancy - toward a global consensus Gynecol Endocrinol 2018 34 01 1 3 10.1080/09513590.2017.1381682 28980832 9 Goedegebure E AR Koning S H Hoogenberg K Korteweg F J Lutgers H L Diekman M JM Pregnancy outcomes in women with gestational diabetes mellitus diagnosed according to the WHO-2013 and WHO-1999 diagnostic criteria: a multicentre retrospective cohort study BMC Pregnancy Childbirth 2018 18 01 152 10.1186/s12884-018-1810-5 29747601 10 Riskin-Mashiah S Younes G Damti A Auslender R First-trimester fasting hyperglycemia and adverse pregnancy outcomes Diabetes Care 2009 32 09 1639 1643 10.2337/dc09-0688 19549728 11 Gupta S Dolin C Jadhav A Chervenak J Timor-Tritsch I Monteagudo A Obstetrical outcomes in patients with early onset gestational diabetes J Matern Fetal Neonatal Med 2016 29 01 27 31 10.3109/14767058.2014.991711 25424373 12 Dickens L T Thomas C C Updates in gestational diabetes prevalence, treatment, and health policy Curr Diab Rep 2019 19 06 33 10.1007/s11892-019-1147-0 31073850 13 International Urogynecological Association International Continence Society Haylen B T de Ridder D Freeman R M Swift S E Berghmans B Lee J An International Urogynecological Association (IUGA)/International Continence Society (ICS) joint report on the terminology for female pelvic floor dysfunction Neurourol Urodyn 2010 29 01 4 20 10.1007/s00192-009-0976-9 19941278 14 Tähtinen R M Cartwright R Tsui J F Aaltonen R L Aoki Y Cárdenas J L Long-term impact of mode of delivery on stress urinary incontinence and urgency urinary incontinence: a systematic review and meta-analysis Eur Urol 2016 70 01 148 158 10.1016/j.eururo.2016.01.037 26874810 15 Gyhagen M Bullarbo M Nielsen T F Milsom I The prevalence of urinary incontinence 20 years after childbirth: a national cohort study in singleton primiparae after vaginal or caesarean delivery BJOG 2013 120 02 144 151 10.1111/j.1471-0528.2012.03301.x 22413831 16 Sangsawang B Sangsawang N Stress urinary incontinence in pregnant women: a review of prevalence, pathophysiology, and treatment Int Urogynecol J Pelvic Floor Dysfunct 2013 24 06 901 912 10.1007/s00192-013-2061-7 17 Martínez Franco E Parés D Lorente Colomé N Méndez Paredes J R Amat Tardiu L Urinary incontinence during pregnancy. Is there a difference between first and third trimester? Eur J Obstet Gynecol Reprod Biol 2014 182 86 90 10.1016/j.ejogrb.2014.08.035 25262291 18 Wesnes S L Rortveit G Bø K Hunskaar S Urinary incontinence during pregnancy Obstet Gynecol 2007 109 04 922 928 10.1097/01.AOG.0000257120.23260.00 17400855 19 Lifford K L Curhan G C Hu F B Barbieri R L Grodstein F Type 2 diabetes mellitus and risk of developing urinary incontinence J Am Geriatr Soc 2005 53 11 1851 1857 10.1111/j.1532-5415.2005.53565.x 16274364 20 Kim C McEwen L N Sarma A V Piette J D Herman W H Stress urinary incontinence in women with a history of gestational diabetes mellitus J Womens Health (Larchmt) 2008 17 05 783 792 10.1089/jwh.2007.0616 18537481 21 Chuang C M Lin I F Horng H C Hsiao Y H Shyu I L Chou P The impact of gestational diabetes mellitus on postpartum urinary incontinence: a longitudinal cohort study on singleton pregnancies BJOG 2012 119 11 1334 1343 10.1111/j.1471-0528.2012.03468.x 22901044 22 Barbosa A M Dias A Marini G Calderon I M Witkin S Rudge M V Urinary incontinence and vaginal squeeze pressure two years post-cesarean delivery in primiparous women with previous gestational diabetes mellitus Clinics (São Paulo) 2011 66 08 1341 1346 10.1590/s1807-59322011000800006 21915481 23 Dietz H P Shek C Clarke B Biometry of the pubovisceral muscle and levator hiatus by three-dimensional pelvic floor ultrasound Ultrasound Obstet Gynecol 2005 25 06 580 585 10.1002/uog.1899 15883982 24 Durnea C M O'Reilly B A Khashan A S Kenny L C Durnea U A Smyth M M Dietz H P Status of the pelvic floor in young primiparous women Ultrasound Obstet Gynecol 2015 46 03 356 362 10.1002/uog.14711 25359670 25 Chamié L P Ribeiro D MFR Caiado A HM Warmbrand G Serafini P C Translabial US and dynamic MR imaging of the pelvic floor: normal anatomy and dysfunction Radiographics 2018 38 01 287 308 10.1148/rg.2018170055 29320316 26 Dietz H P Pelvic floor assessment Fetal Matern Med Rev 2009 20 01 49 66 10.1017/S096553950900237X 27 Dietz H P Pelvic floor ultrasound: a review Clin Obstet Gynecol 2017 60 01 58 81 10.1097/GRF.0000000000000264 28005595 28 Dietz H P Pelvic floor ultrasound: a review Am J Obstet Gynecol 2010 202 04 321 334 10.1016/j.ajog.2009.08.018 20350640 29 Majida M Braekken I H Bø K Benth J S Engh M E Validation of three-dimensional perineal ultrasound and magnetic resonance imaging measurements of the pubovisceral muscle at rest Ultrasound Obstet Gynecol 2010 35 06 715 722 10.1002/uog.7587 20178105 30 Memon H U Handa V L Vaginal childbirth and pelvic floor disorders Womens Health (Lond Engl) 2013 9 03 265 277, quiz 276–277. Doi: 10.2217/whe.13.1723638782 31 Shek K L Dietz H P The effect of childbirth on hiatal dimensions Obstet Gynecol 2009 113 06 1272 1278 10.1097/AOG.0b013e3181a5ef23 19461422 32 Shek K L Dietz H P Can levator avulsion be predicted antenatally? Am J Obstet Gynecol 2010 202 06 5860 5.86E8 10.1016/j.ajog.2009.11.038 33 Falkert A Willmann A Endress E Meint P Seelbach-Göbel B Three-dimensional ultrasound of pelvic floor: is there a correlation with delivery mode and persisting pelvic floor disorders 18-24 months after first delivery? Ultrasound Obstet Gynecol 2013 41 02 204 209 10.1002/uog.11214 22745047 34 Araujo Júnior E de Freitas R C Di Bella Z I Alexandre S M Nakamura M U Nardozz L MM Moron A F Assessment of pelvic floor by three-dimensional-ultrasound in primiparous women according to delivery mode: initial experience from a single reference service in Brazil Rev Bras Ginecol Obstet 2013 35 03 117 122 10.1590/s0100-72032013000300005 23538470 35 Chan S S Cheung R Y Yiu K W Lee L L Leung T Y Chung T K Pelvic floor biometry during a first singleton pregnancy and the relationship with symptoms of pelvic floor disorders: a prospective observational study BJOG 2014 121 01 121 129 10.1111/1471-0528.12400 24148651 36 Siafarikas F Staer-Jensen J Hilde G Bø K Ellström Engh M The levator ani muscle during pregnancy and major levator ani muscle defects diagnosed postpartum: a three- and four-dimensional transperineal ultrasound study BJOG 2015 122 08 1083 1091 10.1111/1471-0528.13332 25716540 37 van Veelen G A Schweitzer K J van Hoogenhuijze N E van der Vaart C H Association between levator hiatal dimensions on ultrasound during first pregnancy and mode of delivery Ultrasound Obstet Gynecol 2015 45 03 333 338 10.1002/uog.14649 25158301 38 van Veelen G A Schweitzer K J van der Vaart C H Ultrasound imaging of the pelvic floor: changes in anatomy during and after first pregnancy Ultrasound Obstet Gynecol 2014 44 04 476 480 10.1002/uog.13301 24436146 39 Stær-Jensen J Siafarikas F Hilde G Benth JŠ Bø K Engh M E Postpartum recovery of levator hiatus and bladder neck mobility in relation to pregnancy Obstet Gynecol 2015 125 03 531 539 10.1097/AOG.0000000000000645 25730212 40 Siafarikas F Staer-Jensen J Braekken I H Bø K Engh M E Learning process for performing and analyzing 3D/4D transperineal ultrasound imaging and interobserver reliability study Ultrasound Obstet Gynecol 2013 41 03 312 317 10.1002/uog.11192 22605574 41 Sartorão Filho C I Pinheiro F A Prudencio C B Nunes S K Takano L Enriquez E MA Impact of gestational diabetes on pelvic floor: A prospective cohort study with three-dimensional ultrasound during two-time points in pregnancy Neurourol Urodyn 2020 39 08 2329 2337 10.1002/nau.24491 32857893
PMC009xxxxxx/PMC9800153.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36580941 10.1055/s-0042-1759774 RBGO-22-0156 Review Article | Artigo de Revisão Commercial Surrogacy: An Overview Gestação de substituição comercial: Uma visão globalhttp://orcid.org/0000-0002-1470-184X Brandão Pedro 1 http://orcid.org/0000-0001-8271-5218 Garrido Nicolás 23 1 Department of Reproductive Medicine, Instituto Valenciano de Infertilidad, Valencia, Spain 2 University of Valencia, Valencia, Spain 3 IVI Foundation, Valencia, Spain Address for correspondence Pedro Brandão, MD, MSc, EFOG, EFRM IVIRMA Valencia, Plaza de la Policia Local 3, 46015, [email protected] 29 12 2022 12 2022 1 12 2022 44 12 11411158 16 5 2022 25 8 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  Surrogacy is the process in which a woman carries and delivers a baby to other person or couple, known as intended parents. When carriers are paid for surrogacy, this is known as commercial surrogacy. The objective of the present work is to review the legal, ethical, social, and cultural aspects of commercial surrogacy, as well as the current panorama worldwide. Methods  This is a review of the literature published in the 21 st century on commercial surrogacy. Results  A total of 248 articles were included as the core of the present review. The demand for surrogate treatments by women without uterus or with important uterine disorders, single men and same-sex male couples is constantly increasing worldwide. This reproductive treatment has important ethical dilemmas. In addition, legislation defers widely worldwide and is in constant change. Therefore, patients look more and more for treatments abroad, which can lead to important legal problems between countries with different laws. Commercial surrogacy is practiced in several countries, in most of which there is no specific legislation. Some countries have taken restrictive measures against this technique because of reports of exploitation of carriers. Conclusion  Commercial surrogacy is a common practice, despite important ethical and legal dilemmas. As a consequence of diverse national legislations, patients frequently resort to international commercial surrogacy programs. As of today, there is no standard international legal context, and this practice remains largely unregulated. Resumo Objetivo  A gestação de substituição é o processo no qual uma mulher engravida e entrega um bebê a outra pessoa ou casal, conhecidos como pais pretendidos. Quando as gestantes são pagas, isto é conhecido como gestação de substituição comercial. O objetivo do presente trabalho é rever os aspectos legais, éticos, sociais e culturais da gestação de substituição comercial, bem como o panorama atual em todo o mundo. Métodos  Trata-se de uma revisão da literatura publicada no século XXI sobre a gestação de substituição comercial. Resultados  Um total de 248 artigos foi incluído nesta revisão. A demanda por tratamentos com gestação de substituição por mulheres sem útero ou com distúrbios uterinos importantes, homens solteiros e casais masculinos está aumentando constantemente em todo o mundo. Este tratamento reprodutivo tem dilemas éticos importantes. Além disso, a legislação é amplamente adiada em todo o mundo e está em constante mudança. Portanto, os pacientes procuram cada vez mais por tratamentos no exterior, o que pode levar a importantes problemas legais entre países com leis diferentes. A gestação de substituição comercial é praticada em vários países, na maioria dos quais não há legislação específica. Alguns países tomaram medidas restritivas contra esta técnica por causa de relatos de exploração destas mulheres. Conclusão  A gestação de substituição comercial é uma prática comum, apesar de importantes dilemas éticos e legais. Como consequência de diversas legislações nacionais, os pacientes frequentemente recorrem a programas de gestação de substituição comercial internacionais. Atualmente, não existe um contexto jurídico internacional padrão e esta prática permanece em grande parte não regulamentada. Keywords bioethics fertilization in vitro legislation medical medical tourism surrogacy Palavras-chave bioética fertilização em vitro legislação médica turismo médico gestação de substituição ==== Body pmcIntroduction Surrogacy is the process in which a woman carries and delivers a baby to another person or couple, known as the commissioning or intended parents (IPs). 1 2 The first historical report of surrogacy is in the book of Genesis – the case of Sarah and Abraham's child, Ishmael, carried by their servant. 3 However, the first officially recognized surrogacy was performed in 1991. Traditional or genetic surrogacy occurs when the carrier also provides the oocyte, hence she is the genetic female progenitor of the child. This modality has progressively been abandoned and the American Society of Reproductive Medicine (ASRM) clearly recommends against it. 4 On the other hand, in gestational surrogacy, also called treatment with a gestational carrier (GC), both gametes are provided by other people, so the surrogate woman has no genetic links with the child. 5 Nomenclature regarding this matter may be somewhat misleading – some also call this last modality “partial” (and traditional would be total surrogacy) since the carrier is only providing the womb, while others call it “total” because the embryo is completely genetically unrelated to the carrier (and traditional would be partial surrogacy). 6 Other terms have been used less frequently, such as classical or straight. 7 Gestational carriers may be a relative, a friend, or a person chosen by the surrogacy agency or in vitro fertilization (IVF) clinic. In the last case, the surrogacy process may be completely anonymous or there may be a direct contact between the IPs and the surrogate. 8 Intended parents may be heterosexual couples, female couples or single women, usually with a uterine condition that limits pregnancy, as well as male couples or single men. Surrogacy may be altruistic or gestational carriers may be paid for the process, which is called “commercial” or “compensated” surrogacy. Commercial surrogacy has been practiced for the last decades and is usually associated with important costs. In general lines, there is a written agreement which outlines both the intentions of both parties, defines duties, delineates reimbursements and payments, and allows decision-making when contingencies occur. 9 There is neither an absolute number of babies born through his technique, nor an exact estimation of its value. In the beginning of the decade of 2010, an estimated 2% of all the IVF cycles in the United States of America (USA) and Canada ended in an embryo transfer to a gestational carrier, 18% of them by foreign patients. In California, in 2015, the percentage of foreign patients was estimated to be 44%. 10 11 12 In the early 2000s (it was legal by then), the commercial surrogacy business was evaluated by Thai and Indian national departments of health as 125 and 449 million US dollars (USD), respectively. 13 In 2012, the industry of surrogacy in the USA was estimated to be worth around 6 billion USD. 5 There was a 4-fold increase in the number of GC cycles in the USA between 1999 and 2013. Between 2006 and 2010, there was a 1,000% increase in the market of international surrogacy. There are reports of international surrogacy agencies stating a growth of 6,000% in 12 years. 14 In the USA, a complete process of surrogacy may cost as much as $200,000. Programs usually include $20,000–80,000 for medical expenses, $3,000–15,000 for legal support, $6,000–54,000 for surrogacy recruiting programs and $20,000–55,000 for carrier compensation. 15 In low-income countries, a full surrogacy process usually costs less than half of the USA. 16 The aim of the present work is to review and summarize information published in scientific journals about commercial surrogacy, with particular attention to the legal, ethical, and sociocultural aspects of this reproductive treatment as well as current practices worldwide. Methods This is a review of all articles listed in PubMed concerning commercial surrogacy published in the 21 st century. The search was conducted in November 2021 using the query “surrogacy” and limiting to articles published from 2000 (including) on. Articles not related to commercial surrogacy were excluded, as well as editorials, letters to the editor, comments, corrigenda, replies, book chapters and study protocols. Articles written in English, French, Portuguese, or Spanish were included. References of the selected articles were thoroughly reviewed in order to include other potentially related articles. Study Appraisal From the search using the query, a total of 1,278 results were retrieved. All titles and/or abstracts of the articles were analyzed. Duplicates were removed ( n =  3). Studies not related to the study question ( n  = 965), editorials/letters to the editor ( n =  7) and answers or comments ( n =  17) were excluded. From the remaining articles, 5 mentioned only altruistic surrogacy, 28 were not available (mainly because of the year or journal of publication) and 5 were excluded due to language. At end, a total of 248 articles were included as core of this review. Ethical Issues Principle of Autonomy If a surrogate treatment is performed by free will of both intended parents and carrier, one may assume that autonomy is guaranteed. 17 If surrogacy was to be prohibited, both IPs and surrogate would be restrained from having the option to participate in it, which violates their autonomy and free will. 18 Nevertheless, the offspring is not autonomous to take any decision on the matter, and will be restricted the right to completely know his/her biological origins. 19 Principle of Beneficence In the same way, if intended parents wish to have a child, their benefit is obvious. The gestational carrier may have her benefit from her personal satisfaction of aiding others and considerable economic compensation. The offspring will be granted the right to live. 20 Principle of Nonmaleficence On the other hand, IPs will face important costs, they won't be able to experience pregnancy, they will not have any control on the surrogate or pregnancy, and they will be imposed a potential additional stress caused by the distance to their children-to-be. Couples seldom report fears related to the surrogate process, particularly related to social judgment and legal troubles arising from this procedure. 21 However, studies on IPs show they usually have a good experience throughout the surrogacy process. 22 Carriers will be exposed to the risks of pregnancy, may eventually suffer from social stigma and ostracism (including by their family, which often leads them to move away from their communities to hide pregnancy) and will not have any motherhood rights on that child. 23 24 25 26 27 While studies with surrogates in high-income countries show that GCs perceive surrogacy as a positive experience, studies in low-income countries show high rates of depression and negative feelings in GCs. 28 29 30 A curious comparison has been made between the carrier and a nanny – if the child is already adopted prenatally by the IPs, the surrogate would play the role of a "prenatal nanny”. These authors question the potential maleficence of being a carrier and ask if it would be more outrageous to be a nanny before birth than after. 31 Children may eventually suffer from social stigma and may have difficulties when being told their true origins. They will not be able to know or have any contact with their birth mother. 32 33 Nevertheless, studies about the opinion of children (born after surrogacy) are reassuring, even though evidence to date is limited. 24 34 35 In addition, some doubts have been raised about raising a child in a non-traditional (non “mother and father”) family. 36 However, much research has been done on child rearing by same-sex couples and the results are reassuring. 37 Principle of Justice If all people were to have the same access to surrogacy, the principle of justice would be granted for everyone, but this implies that surrogacy is not to be limited based on marital status, ethnicity, religion, sexual orientation, or any another kind of discrimination. 38 39 In fact, surrogacy is a way to provide fertility to some infertile couples, singles and same-sex couples. 40 However, in case of commercial surrogacy, the costs are high and not affordable by everyone. As opposed to patients, carriers are not given any legal motherhood rights. In addition, the process of selecting only healthy young women as carriers may lead to discrimination of other candidates willing to participate. 41 As opposed to conventionally-born children, children born through surrogacy are not granted the right to grow with their gestational mother. 17 Ethical Aspects of Commercial Surrogacy An important ethical aspect of commercial surrogacy is that women may be regarded as a “way to conception” and children as mere products of conception. 17 42 43 44 45 Concerning children, some regard surrogacy as “selling babies” or human trafficking. 46 47 48 49 Others consider it does not violate any of the children's rights, it cannot be regarded as a market of babies and that if the conditions of the surrogate arrangement are fulfilled at the end of the process, the best interest of the child is implicitly protected, since this was the manifested desired of both parties, a carrier who was always aware she was not going to be mother, and the IPs who are receiving their most desired child. 50 Some consider these treatments to be exploitative to women. 51 52 53 In fact, commercial surrogacy opens a door to illegal exploitation if not adequately ruled and monitored, especially in low-income countries. 54 55 In many cases, third party organizations or people receive their compensation and little, if any, is given to the surrogate. 56 57 58 59 It is also not uncommon that women are not aware of the risks of this procedure and do not have an opinion on the decision to become a carrier. 60 61 62 Others regard surrogacy as a different way of prostitution. 63 64 Surrogacy is also seldom compared to donating or selling a kidney in vivo . 25 65 66 Interestingly, some argue that paid surrogacy is in no worse position than many other exploitative commercial transactions which take place against a backdrop of global inequality and constrained options, such as poorly paid and dangerous construction work. Hence, there would be no reason for special condemnation. 67 The criminalization of commercial surrogacy may result in undesirable consequences, removing opportunity for evidence-based law reforms which would regulate the process. 68 69 70 71 In the end, some authors argue that the theoretical “do no harm” reasons to refuse surrogacy are far from being proven. Thus, there would be no reasons for banning commercial surrogacy. 72 73 74 Motivations There are several reasons that may lead women to become a GC, such as economic compensation, pure altruism, the wish of going through a different kind of motherhood or cultural beliefs. 63 75 Nevertheless, the majority of carriers undergo this process for the compensation, mostly people with dependent families. 76 77 78 In fact, some women tend to accept it as a work and regard the surrogacy agency as their “boss”. 23 76 79 Other gestational carriers perceive this process as an “exchange of gifts”. 79 On the other side, patients who resort to surrogacy are usually women (single or part of a couple) without uterus or with important uterine disorders impairing pregnancy, single men, and same-sex male couples. 80 81 82 It is not uncommon that clinics, agencies or international intermediates advertise commercial surrogacy treatments specifically to single men and male couples. 83 Compensation Depending on national legislation, surrogacy may be commercial or altruistic. Nevertheless, in some countries where only altruistic surrogacy is allowed, carriers may be given a compensation for specific matters related to the process (such as health care expenses, sick leaves, etc.). However, given the considerable costs of healthcare in some countries, a large reward could eventually be acceptable as a mere compensation for these expenses. Thus, it may not be easy to clearly define the border between altruistic and commercial. Depending on the maximum amount allowed (if any), these compensations are seldom used off the record to mask monetary payment for surrogacy. 84 As a consequence, many authors consider the distinction between ‘altruistic’ and ‘commercial’ surrogacy increasingly unsustainable both in law and policy. 85 The amounts given to carriers in a commercial surrogacy process vary widely between countries. In the USA a gestational carrier usually receives an average amount of USD 20,000 to 55,000 per pregnancy. 15 86 Multiple pregnancies are usually paid a supplement. A monetary compensation may be regarded as a win-win situation for both parties, as the surrogate gets the money and the IPs get the child, while some believe surrogacy must not be reduced to a business transaction. 87 88 The monetary compensation for surrogacy may lead to a contradiction. On the one hand, paying a low amount may be regarded as compensation for expenses and damage, but also as an exploitation. 89 90 On the other hand, paying higher amounts leads carriers to be better compensated for their efforts, but may also lead to a competitive reproductive market, “machinerizing” of women and treating children and reproductive treatments as commodities. 51 In the Netherlands, some attempts have been made to define lower and upper limits for compensations, mainly based on the oocyte donation models. However, it is not easy to define what would be the true labor associated with surrogacy and if it should be considered as a full or partial-time job, since women will be pregnant 24 hours a day, but they are able to combine surrogacy with their daily activities, including other jobs. 51 Especially in low-income countries, it is not uncommon for surrogates to expect an extra-contractual compensation for the process. 87 In 2005, the European Society of Human Reproduction and Embryology (ESHRE) published its position regarding commercial surrogacy – “Payment for services is unacceptable; only reimbursement of reasonable expenses and compensation for loss of actual income should be considered”. 91 The International Federation of Gynecology and Obstetrics (FIGO) also stated that surrogate arrangements should not be commercial. 92 93 On the other hand, the American Society of Reproductive Medicine (ASRM) compares gestational surrogacy to medical research, in which individuals are paid for activities demanding time, stress, physical effort and risk, so they consider financial compensation for surrogacy ethically justifiable. 4 Likewise, the American College of Obstetricians and Gynecologists (ACOG) affirms that compensation is ethical and appropriate for the time, effort and risks taken by a gestational carrier. 94 Anonimity Regimen In many states, anonymity of surrogacy is to be guaranteed, which means the choice of the surrogate mother and all the communication between her and the IPs is indirect and mediated by the clinic or agency. This is the perfect regimen for some couples who prefer anonymity and not to know the carrier in any point of the process. 95 On the other hand, there are some states where carriers and IPs are not only obliged to know and approve each other, after checking they match their expectations, but they are also encourage to actively communicate during the process and participate during all its steps. 87 96 Some authors believe this involvement between IPs and carriers and an eventual further relationship of the latter with the child may be beneficial to all parties and may ease some of its ethical issues. 97 98 Medical Risks Being a gestational carrier is associated with important adverse medical or psychological outcomes. Obstetric complications are not higher (if not lower) in surrogate singleton gestations, since surrogate mothers are usually young and healthy. 99 100 101 102 103 Nevertheless, no gestation is exempt from risk. 104 105 Also, double embryo transfers are quite common in surrogate processes, because it is usually cheaper than having two separate pregnancies, resulting in more multiple pregnancies. 106 107 108 109 In addition, it is believed that cesarean section rates are high among surrogates, not only because IPs may ask for it, but also because low-income surrogates receive medical care in private clinics while in other situation they would be treated in public health systems. 110 Candidates to |GC are often quite misinformed about the procedure and lack of psychological and legal support. 111 Particularly in developing countries, women are seduced into being GCs. Many of these women live in precarious conditions and use this resource for a better future for themselves and their families. In some cases, women are even forced to be GCs. 112 In some countries, such as India, gestational carriers often live in a hotel hired for the purpose, in order to have more dignified conditions, a healthier lifestyle for their pregnancy and more easily be able to maintain obstetric surveillance, particularly women who live in remote areas. 113 Likewise, children born from surrogacy are not risk free. Nevertheless, current scientific data suggest this option is safe as long as all parties have adequate screening and medical, psychological, and social supports. 99 114 In order to optimize the outcomes of a surrogate gestation, both the United States Food and Drug Administration (FDA) and the ASRM have developed guidelines to help choosing the most adequate gestational carrier. 115 Ideally, candidates to gestational carriers must be between 21 and 45 years old, with an optimal BMI, have at least one previous term uncomplicated pregnancy, but no more than 5 deliveries or 3 cesarean sections and with a 12 to 18 months pregnancy interval. The optimal selection of GC candidates also includes assessment of their mental health, since this may be a very demanding process. 116 Adequate medical counseling to the surrogate candidates must be done in order to promote healthy habits both before and during pregnancy. Women must be encouraged to receive preconception immunizations, if applicable, to avoid potential teratogenic medications, to take folic acid supplements, to restrain from smoking, drinking alcohol, and excessive caffeine intake. 92 Surrogacy Agencies and Marketing There are several international agencies exclusively dedicated to intermediate surrogate treatments. 117 The websites of these agencies seldom advertise surrogacy treatments abroad focusing on the needs of IPs, referring to surrogacy as a solution to their problem, privileging genetic parenthood. Many online advertisements of global medical tourism offer "special deals" on commercial surrogacy. 118 119 They seldom include basic and guarantee plans. The difference is that the latter includes all necessary embryo transfer to have a live newborn. The potential for exploitation of the carriers is obviously not exposed and the surrogacy arrangements are advertised as a mutual benefit. In fact, this subject is often a taboo and avoided as much as possible during all the surrogacy process. 120 Surrogacy agencies usually include staff trained in international legislation and marketing. Interestingly, most of the staff of these agencies have also undergone a similar process or is quite familiar with other transnational reproductive treatments by personal experience. 121 They usually provide legal assistance, included in all their plans. Regardless of the countries and their legal context, it is not uncommon for these agencies to advertise that there are no legal risks and there will be no litigation. They take it for granted that the surrogates have no legal rights over the child-to-be, that both the country of treatment and the country of origin will only recognize the motherhood of IPs. These agencies also state that in case of litigation, the law always protects the IPs, when actually in most cases there is no legal framework. 122 However, these agencies are an important means for IPs to easily reach a surrogacy contract, including recruitment of donors, carriers, reproductive treatments, obstetric follow-up, and legal assistance. 123 124 Legal Issues Legal conflicts may appear in the country where surrogacy is performed, but also in the country of origin of the IPs (“receiving country”), when returning home with the child. 125 126 127 Country Where Surrogacy is Performed National legislation varies substantially worldwide. 128 Some countries explicitly prohibit any type of surrogacy, others allow surrogacy of any type, while others have some restrictions concerning marital status, sexual orientation, nationality, country of residence, medical reason to undergo a surrogate treatment and the altruistic/commercial nature of the process. In most countries, surrogacy is not regulated at all. 14 All surrogacy arrangements beginning by signing a contract between the IPs and the GC. There are innumerous important points that should be clearly settled in the contract in order to avoid future potential litigation. 129 These include setting out both parties legal parentage and nonparentage rights, agreements on prenatal and delivery issues, compensations and fees, insurances, and assumptions of risks. 130 The central and most important party in any reproductive treatment is the offspring because he/she is the only party that cannot have a word in any preconception contract or agreement. As a consequence, most countries worldwide recognize that the child, regardless of the way in which he/she was conceived, has the same rights guaranteed by the national and international framework of human rights. 131 132 Regarding the mother, defining biological motherhood may be quite challenging in the modern era, especially in assisted reproductive treatment (ART) involving third parties, such as donated gametes or surrogacy. An interesting example is the reception of oocytes from partner (ROPA) method, or lesbian shared IVF, in which both women share biological motherhood, one will be the gestational mother (the one giving birth) and the other will be the genetic mother (the one providing the oocytes). 133 In surrogacy, 3 people may be involved in motherhood: the carrier (which will be the birth mother), the oocyte provider and the intended mother (depending if it is with own or donated oocytes, these last two will be the same or different women, respectively). 134 In the majority of countries, legal motherhood is based upon the fact of birth. The “anonymous” or “secret birth,” where a woman may choose to give birth without revealing her identity, is not legal in most countries. Thus, as a rule, the woman giving birth is automatically recognized as mother, until proven otherwise. The requirement for a man to be registered as a father of a child depends upon the circumstances of the case, especially the couple's marital status. In most countries, in a married heterosexual couple, the man is automatically assumed as the father. However, in most cases, a man may voluntarily acknowledge his legal paternity. Once a child is registered and receives a birth certificate, parents are legally recognized as so for all purposes. However, in most states, it is possible to reverse this process upon genetic proof. 135 One of the main obstacles for couples who resort to surrogacy is the registration of the newborn in their name and the cession of motherhood rights by the carrier. 136 Countries where surrogacy is contemplated by law, as is the case of some states of the USA, a prepregnancy contract is signed between the two parties in which the surrogate waives any rights to motherhood after birth. Therefore, in these cases, the birth certificate is automatically conceived with the name of the intended parents. On the other hand, in countries where surrogacy is not regulated and it is performed not because it is legal, but because it is not illegal, the original birth certificate is usually issued with the surrogate as mother, and the IPs have to ask national authorities to amend the certificate with their names. 137 However, litigation may arise in various points throughout the surrogacy journey, in view of obstetric complications, decisions regarding pregnancy interruption, lack of agreement between the IPs and the surrogate, divorce or separation of the IPs, or changes of mind of one of the parties during the process. 138 Even in the presence of a prior contract, if this practice is not regulated and there is no specific legislation, its legal value is doubtful. The most troublesome scenario is if the surrogate decides not to abdicate her motherhood rights. 139 140 141 In these cases, a DNA test is seldom required. Thus, parenthood is determined on a genetic basis and the court is asked to declare the motherhood rights of the carrier null. In some cases, parents have to wait months after birth to have the birth certificate amended. 142 This may be even more problematic if pregnancy is a result of double donation (donated oocytes and sperm), in which none of the IPs shares a genetic link with the baby. In any case, in the absence of a deferment by a court, the carrier has full motherhood rights over the child, which prevents the child from leaving the country with the IPs without her consent. There are reports of large amounts of bribes paid to the carrier to finally cooperate by ceding her rights. 143 Receiving Country Due to the absence of uniform international legislation, cross-border surrogacy treatments may pose legal issues when returning to the home country of IPs with children who, according to the legislation of the receiving country, have been conceived illegally. 144 145 146 147 The main steps where IPs face difficulties most frequently in their home countries are when requiring a passport or any travel documentation at their consular authorities overseas to return home with the child, and when the IPs, back home, wish to register their children as a national citizen. 148 If their native countries do not recognize surrogacy, patients may struggle to register the child as theirs. 149 150 151 152 Further problems may arise in cases of singles or same-sex couples from countries where they are not allowed to have children. In these cases, surrogacy itself may not be the sole problem, but the lack of legal framework to recognize both same-sex IPs as legal parents. 143 There are reports of people who were criminally accused of having filed an illegal process abroad. Nevertheless, national courts ended up acquitting them for lack of legal support regarding international affairs, because these procedures were officially recognized in the country they were performed, and because this decision was ultimately considered to be in the best interest of all parties involved, especially the child. 143 153 154 As a consequence of these disparities between legislations and issues of countries regarding international private laws, many judicial authorities of several states have attempted to create solutions to enable children born from an international surrogacy arrangement to return home. The Hague Conference on Private International Law (HCCW) is an intergovernmental organization in the area of private international law that administers international conventions, protocols, and legal instruments. It is an important organization that deals with conflicting international affairs. In 2012, the Permanent Bureau of the HCCW released “A Preliminary Report on The Issues Arising from International Surrogacy Arrangements”. Since then, this institution has been trying to create guidelines to standardize the international recognition of surrogacy performed abroad. As of 2021, the HCCW had 90 country as members. 14 Transnational Surrogacy The denial of surrogacy in most countries, for all or for some (such as single people or same-sex couples), its cost or the lack of available carriers led to an important transnational search for these (and other) reproductive treatments. 155 156 This phenomena has been called reproductive, procreative or fertility tourism, transnational reproduction or cross border reproductive care. 157 158 159 160 161 162 In European countries alone and concerning any kind of ART, in 2010, a total of 24,000 to 30,000 cycles of cross border fertility treatment within the continent were estimated each year, involving 11,000 to 14,000 patients. 163 Transnational surrogacy is one of the fastest-growing cross-border reproductive treatments. 164 Choosing where to perform the surrogacy treatment usually entails finding the right equilibrium between legal guarantees and costs. 165 Due to the variety of legislations, costs and availability of donors and carriers between countries, patients may search for other countries to do the entire process of surrogacy, or different phases of the surrogate treatment in more than one country. 158 As an example, a male couple may get their donated oocytes from South Africa, where there are many donors available, do the IVF, recruit the surrogate and embryo transfer in Georgia ( Sakartvelo ), due to attractive prices, and fly the gestational carrier to the USA to deliver the baby, where children may be registered by both parents. 166 167 Countries for gamete donation (when needed) are usually chosen based on the availability of donors, anonymity regimen of donation, costs of the process, compensation to the donors, and ethnic issues. In vitro fertilization, in turn, may pose some legal obstacles in some countries. Legal requirements, as well as costs, the possibility of freezing embryos, performing preimplantation genetic test (PGT) and sex selection are important aspects. Surrogacy itself is the most complex part of the process. The legal status of surrogacy is by far the most important aspect when it comes to choose the country, not only the presence or not of specific legislation concerning the matter, but also the legal value of surrogate contracts in more delicate situations, such as pregnancy interruption and in case the carrier decides to keep the baby. In addition, same-sex couples may choose the country of delivery in order to be able to share parenthood since birth. The exclusion of motherhood rights from the gestational carrier and the attribution of these rights to IPs may be done immediately after birth, or it may be a court decision after DNA tests to the child, genetic IPs and the carrier. 168 Furthermore, in the case of gay couples, the process of sharing legal parenthood may be much easier if their country of origin accepts joint adoption of a child by same-sex couples. It is very common to cross borders between neighboring countries to undergo surrogacy. Both parents or carrier may be required to cross the border, as well as gametes or embryos, depending on the case. Examples of frequent neighboring border crossings are between the USA and Mexico, and Thailand and Vietnam or Laos. 169 Diverse measures have been taken by many governments to avoid the so called “reproductive tourism”. Some countries where treatments used to be performed banned these treatments, at least for foreign patients. Other countries, such as Portugal, decided to approve surrogacy only to national or resident citizens since its very beginning, to avoid reproductive tourism and legal litigation with other countries. 170 On the other hand, receiving countries face important dilemmas when it comes to attribute nationality to the offspring, but they are also in an ungrateful position to limit reproductive treatments abroad. 171 The vast majority of countries have no specific legislation concerning children conceived abroad via surrogacy. 157 Some countries, such as Australia, the Netherlands or the UK, are trying to draw preconception agreements for surrogate treatments abroad. 84 Several scandals have been reported during the last decades, such as the Baby Gammy, a child with Down Syndrome, whose intended parents left him in Thailand while taking home his twin sister, who was not affected by the condition. 172 Another famous case was a Japanese man who tried to conceive seventeen children via surrogacy. 169 In India, a Japanese couple refused to receive the baby because they divorced 1 month before delivery. 173 Following these occurrences, some popular destinations, especially in Asia, have taken legal measures to limit commercial surrogacy or access to foreign patients. Commercial surrogacy was banned in Thailand and Nepal in 2015 and in Cambodia in 2016. 76 In India, same-sex couples were excluded in 2013 and in 2018 this country limited surrogacy to national patients. 166 174 Consequently, the offer of surrogacy destinations has decreased. Over time, the “one-stop” surrogacy destinations have become increasingly rare, especially due to the partial limitation of some of the steps of the process, requiring intended parents to do a “puzzle” with various countries to complete their journey in surrogacy. On the other hand, demand for surrogacy from high-income countries such as European countries and Australia is continuously rising, due to increasing maternal age, single men, and male same-sex couples. Cultural Aspects Some social and cultural aspects influence the way society is more or less receptive to gestational surrogacy, especially the country of origin, religion, activism, and the whole social context. 175 176 Religion is one of the most important points, since different religions have various points of view regarding motherhood, marriage, life, and the status of the embryo. 177 Studies show that the vast majority of Muslims are against surrogate treatments, since procreation and childbearing must be carried out only under the framework of marriage. 178 179 However, some acknowledge this may be ethically justified and medically necessary. 180 Polls in Iran, Jordan, and Lebanon revealed a predominant negative attitude among healthcare workers and students toward surrogacy, mainly driven by religious beliefs. 181 182 183 In Jewish society, cases of donor eggs or surrogacy are also hard to deal with. If one of the women involved is not Jewish, rabbinic authorities disagree about the Jewish status of the child, which may imply that the child undergoes religious conversion. 184 The Catholic church is against any form of ART, especially if there is a third party involved, since reproduction is to be practiced in a marital context. Other branches of Christianism do accept IVF treatments. The opinions concerning surrogacy within Christians are diverse, even though they are, in general, in disfavor of this technique. 185 Hindus regard infertility as a curse, which means they accept ART and surrogacy as a cure for infertility. 185 Regarding Buddhism, since there are few theories written about ART, as long as pain and harm are avoided, all practices are acceptable. However, the very desire for a child through extraordinary means can also be seen as an unhealthy material attachment. As so, the matter of surrogacy is conflicting. 185 Studies report a duality of criteria in high income countries regarding public opinion about surrogacy. Poll-based studies in Australia, France, Germany, Japan, Philippines, Spain and the USA revealed that more than half of general population would be in favor of surrogate treatments for heterosexual and same-sex couples. 186 187 188 189 190 191 The same goes for reproductive care professionals and students. 192 193 194 On the other hand, feminists are against any kind of surrogacy. 63 195 196 Curiously, a recent meta-analysis showed that the majority of infertile women were not in favor of surrogacy. 197 198 A study in Romania revealed that women (general population) would rather adopt than resort to surrogacy. 199 On the other hand, studies with Iranian infertile couples reported that the majority has a positive view on surrogacy. 200 201 Legal Context Worldwide America By 2021, 22 USA states have no legal statutes for commercial surrogacy, 16 states explicitly and 7 states implicitly allow it, and in 5 states it is forbidden. 114 202 In Canada, commercial surrogacy is banned, even though altruistic surrogacy is permitted in all states except in Quebec. 5 203 204 In Mexico, legal status of surrogacy is not regulated at a federal level, thus, only a few states, like Tabasco, used to offer commercial arrangements. Consequently, Tabasco used to be a major destination for transnational surrogacy. In 2016, Tabasco changed the state law to limit surrogacy to heterosexual infertile couples. In June 2021, a Supreme Court decision upheld surrogacy in Mexico – the court endorsed both free and paid surrogacy and even invalidated the provisions of one state that prohibited access to same-sex and foreign couples. Since then, a door opened to any Mexican state to perform commercial surrogacy agreements. 205 In most countries of South America, surrogacy is not regulated, apart from Brazil and Uruguay. In Brazil, surrogacy is allowed only in the altruistic regime. There are 2 circumstances in which a person can resort to surrogacy, a woman who has ovarian reproductive potential but a uterine condition that prevents pregnancy, or a same-sex couple. In either case, the surrogate must be a 1 st to 4 th degree relative of one of the PIs, such as mother, sister, aunt, or cousin. As a consequence of the lack of legislation banning commercial agreements, some countries, such as Colombia, have become popular surrogacy destinations in the last years ( Fig. 1 ). 206 Fig. 1 Map showing the current legal status of surrogacy in America. Europe Legislation in Europe varies between different countries. 207 Surrogacy in Europe is allowed or not banned in Albania, Armenia, Belarus, Belgium, Cyprus, Czech Republic, Georgia, Greece, Ireland, Macedonia, Portugal, Romania, Russia, the Netherlands, the UK, and Ukraine. 208 On the other hand, it is completely banned in Austria, Estonia, France, Germany, Italy, Lithuania, Norway, Spain, Sweden, and Switzerland. 157 In most countries where it is regulated, only altruistic surrogacy is permitted, such as Belgium, Greece, Ireland, the Netherlands, Portugal, and the UK ( Fig. 2 ). 170 208 209 In Georgia, Russia, and Ukraine commercial surrogacy is possible, but in general limited to heterosexual couples. 5 Georgia and Ukraine became a major destination for commercial surrogacy, due to its attractive prices and easiness of the process. 5 Since there is no uniform legislation, the European Court of Human Rights (ECtHR) has gained importance regarding transcontinental surrogacy for European citizens, especially for receiving countries with no specific legislation or where it is forbidden. 17 153 210 211 212 213 This entity has mediated some complicated processes, in particular in France, ultimately ruling in favor of the legal recognition of the nationality and affiliation of children conceived through international surrogacy, bearing in mind that this would be in their best interest. 85 143 214 215 216 In addition, in some countries such as the UK, courts have accepted foreign commercial surrogacy, as national legislation supports the concept of surrogacy, provided that the foreign surrogacy is lawful, there are adequate safeguards for the child, the interests of the child being paramount, the arrangements are ethical and not exploitative, and the costs are reasonable. 217 218 219 220 221 In 2009, Spain made an ad hoc regulation of the national registry to facilitate the often unpredictable process of recognition of the filiations resulting from cross-border surrogacy. 222 223 Norway does not allow surrogacy of any kind but recognizes the citizenship of children of Norwegian parents born by surrogacy abroad. 224 Fig. 2 Map showing the current legal status of surrogacy in Europe. Asia Asian countries used to be a major destination for commercial surrogacy, until more restricted legislation on the subject has progressively been imposed. 225 226 Since 2018, when commercial and international surrogacy were both banned in India – the Surrogacy (Regulation) Bill – most countries in south or southeast Asia do not recognize commercial surrogacy. 227 228 229 230 231 India, Nepal, Thailand, and Vietnam recognize altruistic surrogacy (if not for all, in some specific situations or for national citizens only) but all these countries have explicitly banned commercial surrogacy. 173 232 233 Japan and South Korea do not have specific regulation regarding surrogacy. 8 234 Mongolia, Pakistan, People's Republic of China and Taiwan explicitly prohibit any kind of surrogacy. 235 Even though prohibited, in People's Republic of China there is an important practice of clandestine commercial surrogacy ( Fig. 3 ). 236 Nevertheless, an important part of these countries have no regulation at all regarding surrogacy, so it is not considered illegal and in some of them it keeps on being performed. 237 As a consequence of the ban to commercial surrogacy imposed by most south or southeast Asian countries, in particular India and Thailand, Laos became a popular choice, sometimes in an hybrid regimen with Thailand. 173 In the Middle East, Israel allows altruistic surrogacy only. 155 238 Iranian legislation is not clear regarding surrogacy and it is not an uncommon practice in the country. 179 In Saudi Arabia and in the United Arab Emirates, surrogacy is completely forbidden. 178 179 Fig. 3 Map showing the current legal status of surrogacy in Asia. Oceania In Oceania, altruistic surrogacy may be performed in Australia and New Zealand, but commercial surrogacy is illegal ( Fig. 4 ). 232 239 240 241 242 243 Fig. 4 Map showing the current legal status of surrogacy in Oceania. Africa Most African countries do not have any regulation concerning surrogacy. In Kenya, surrogacy is not regulated, hence it became a popular destination for this practice. 244 In South Africa, altruistic surrogacy is allowed ( Fig. 5 ). 157 Fig. 5 Map showing the current legal status of surrogacy in Africa. International Affairs In order to avoid transnational surrogacy, some models of “ideal” commercial surrogacy laws and arrangements have been proposed. Some Australian groups proposed a model targeting a fair and just compensation, enforceability of surrogacy agreements, amended parentage presumptions and the ability to obtain prebirth parenting orders, regulation of surrogacy agencies and brokers, and recognition of approved international surrogacy arrangements. 245 246 Given the legal diversity and the frequent difficulty of fitting foreign activity into national law, there are many calls of action at the international, national, and professional levels to establish a human rights based system of international governance based on three regulatory models: public health monitoring, intercountry adoption, and trafficking in human beings, organs and tissues. 247 248 249 250 251 As stated before, many international intermediates make the connection between the IPs, the gametes donors and the carriers, most of the times via specific gamete banks, IVF clinics and surrogacy agencies. Most of these agencies are based in a unique country but they operate with IPs of any nationality, offering surrogacy programs in different countries, adjustable to any case. Conclusion Surrogacy is an important means for some people to achieve biological parenthood, in particular women with uterine disorders, single men and male couples. However, this procedure entails important ethical dilemmas and legal issues. As a consequence of the diverse legal contexts worldwide, transnational surrogacy programs are frequently used, despite the possible legal complications. Commercial surrogacy is a common practice, although not regulated in most countries. This technique raises even more ethical and legal dilemmas. Various countries and international organizations made important attempts to regulate this practice in order to standardize its legal context worldwide and avoid litigation. Nevertheless, the situation remains largely unregulated and, as such, there is still a long way to go. Conflicts to Interests The authors have no conflict of interests to declare. ==== Refs References 1 Joseph R A Rafanello A M Morris C J Fleming K F Surrogacy: Pathway to Parenthood Neonatal Netw 2018 37 01 19 23 10.1891/0730-0832.37.1.19 29436354 2 Hobzová H Surrogate motherhood: the contradicitons in terminology Ceska Gynekol 2018 83 06 464 467 30848153 3 Klock S C Lindheim S R Gestational surrogacy: medical, psychosocial, and legal considerations Fertil Steril 2020 113 05 889 891 10.1016/j.fertnstert.2020.03.016 32312561 4 Ethics Committee of the American Society for Reproductive Medicine. Electronic address: [email protected] Ethics Committee of the American Society for Reproductive Medicine Consideration of the gestational carrier: an Ethics Committee opinion Fertil Steril 2018 110 06 1017 1021 10.1016/j.fertnstert.2018.08.029 30396538 5 Deharo G Madanamoothoo A Is international surrogacy the lark's glimmer?: When Covid-19 reveals the legal insecurity of surrogacy use Eur J Health Law 2020 27 04 345 367 10.1163/15718093-BJA10025 33652391 6 Cui L Li L Adashi E Y Chen Z J Surrogacy: a family-building option in search of legitimacy BJOG 2016 123 03 65 68 10.1111/1471-0528.14195 27627601 7 Beeson D Darnovsky M Lippman A What's in a name? Variations in terminology of third-party reproduction Reprod Biomed Online 2015 31 06 805 814 10.1016/J.RBMO.2015.09.004 26526414 8 Ha J O Current issues on a standard for surrogate pregnancy procedures Clin Exp Reprod Med 2012 39 04 138 143 10.5653/CERM.2012.39.4.138 23346523 9 Swain M E Rogerson C J Addressing legal issues in cross-border gestational surrogacy: current topics and trends Fertil Steril 2021 115 02 268 273 10.1016/j.fertnstert.2020.11.031 33579520 10 Swanson K Letourneau J M Kuppermann M Einerson B D Reproductive travel of intended parents for delivery of gestational carrier pregnancies Obstet Gynecol 2020 136 03 591 596 10.1097/AOG.0000000000003898 32769637 11 National Assisted Reproductive Technology Surveillance System (NASS) Group Perkins K M Boulet S L Jamieson D J Kissin D M Trends and outcomes of gestational surrogacy in the United States Fertil Steril 2016 106 02 435 44200 10.1016/j.fertnstert.2016.03.050 27087401 12 White P M Canada's surrogacy landscape is changing: should Canadians care? J Obstet Gynaecol Can 2017 39 11 1046 1048 10.1016/J.JOGC.2017.06.017 28760567 13 Howard S Taming the international commercial surrogacy industry BMJ 2014 349 g6334 10.1136/BMJ.G6334 25342063 14 Hague Conference on Private International Law. Permanent BureauA preliminary report on the issues arising from international surrogacy arrangements. The Hague: Permanente Bureau;2012 15 Tsai S Shaia K Woodward J T Sun M Y Muasher S J Surrogacy laws in the United States Obstet Gynecol 2020 135 03 717 722 10.1097/AOG.0000000000003698 32028508 16 Knoche J W Health concerns and ethical considerations regarding international surrogacy Int J Gynaecol Obstet 2014 126 02 183 186 10.1016/J.IJGO.2014.03.020 24834850 17 Gracias C S Reproductive surrogacy in Ireland - an ethical and legal context Ir J Med Sci 2021 190 03 1063 1070 10.1007/s11845-020-02402-4 33131030 18 Sifris R Commercial surrogacy and the human right to autonomy J Law Med 2015 23 02 365 377 26939504 19 Wade K Reconceptualising the interest in knowing one's origins: a case for mandatory disclosure Med Law Rev 2020 28 04 731 752 10.1093/medlaw/fwaa032 33118029 20 Brinsden P R Gestational surrogacy Hum Reprod Update 2003 9 05 483 491 10.1093/humupd/dmg033 14640380 21 Golboni F Jalali A Dinmohammadi M Taghizadeh Z Nouri P Salahsoor M R Iranian model of decision making to use surrogacy: A grounded theory Study Health Care Women Int 2020 41 07 853 865 10.1080/07399332.2020.1758107 32357079 22 MacCallum F Lycett E Murray C Jadva V Golombok S Surrogacy: the experience of commissioning couples Hum Reprod 2003 18 06 1334 1342 10.1093/humrep/deg253 12773469 23 Rozée V Unisa S de La Rochebrochard E The social paradoxes of commercial surrogacy in developing countries: India before the new law of 2018 BMC Womens Health 2020 20 01 234 10.1186/s12905-020-01087-2 33059640 24 Riddle M P The psychological impact of surrogacy on the families of gestational surrogates: implications for clinical practice J Psychosom Obstet Gynaecol 2022 43 02 122 127 10.1080/0167482X.2020.1814729 32909488 25 Hodson N Townley L Earp B D Removing harmful options: The law and ethics of international commercial surrogacy Med Law Rev 2019 27 04 597 622 10.1093/medlaw/fwz025 31867634 26 Bagcchi S Mothers who turn to surrogacy to support their families face ostracism, study shows BMJ 2014 348 g3257 10.1136/bmj.g3257 24821314 27 Agnafors M The harm argument against surrogacy revisited: two versions not to forget Med Health Care Philos 2014 17 03 357 363 10.1007/S11019-014-9557-x 24664239 28 Yee S Hemalal S Librach C L “Not my child to give away”: A qualitative analysis of gestational surrogates' experiences Women Birth 2020 33 03 e256 e265 10.1016/j.wombi.2019.02.003 30824376 29 Lamba N Jadva V Kadam K Golombok S The psychological well-being and prenatal bonding of gestational surrogates Hum Reprod 2018 33 04 646 653 10.1093/humrep/dey048 29566176 30 Jadva V Murray C Lycett E MacCallum F Golombok S Surrogacy: the experiences of surrogate mothers Hum Reprod 2003 18 10 2196 2204 10.1093/humrep/deg397 14507844 31 Nisand I Grossesse pour autrui : pour le cas par cas Gynécol Obstét Fertil 2010 38 03 226 229 10.1016/j.gyobfe.2010.01.007 20189435 32 Mitra P Invisible women in reproductive technologies: Critical reflections Indian J Med Ethics 2018 3 02 113 119 10.20529/IJME.2018.031 29724696 33 Montanari Vergallo G Marinelli E di Luca N M Zaami S Gamete donation: are children entitled to know their genetic origins? a comparison of opposing views. The Italian state of affairs Eur J Health Law 2018 25 03 322 337 10.1163/15718093-12530378 34 Jadva V Postdelivery adjustment of gestational carriers, intended parents, and their children Fertil Steril 2020 113 05 903 907 10.1016/j.fertnstert.2020.03.010 32312560 35 Zadeh S Ilioi E C Jadva V Golombok S The perspectives of adolescents conceived using surrogacy, egg or sperm donation Hum Reprod 2018 33 06 1099 1106 10.1093/humrep/dey088 29701833 36 Montanari Vergallo G A child of two mothers: what about the father? Italian overview Acta Biomed 2019 90 03 319 325 10.23750/abm.v90i3.7970 31580321 37 Pennings G Evaluating the welfare of the child in same-sex families Hum Reprod 2011 26 07 1609 1615 10.1093/humrep/der109 21493632 38 Mackenzie S C Wickins-Drazilova D Wickins J The ethics of fertility treatment for same-sex male couples: Considerations for a modern fertility clinic Eur J Obstet Gynecol Reprod Biol 2020 244 71 75 10.1016/j.ejogrb.2019.11.011 31760265 39 Smith M K Willmott L Trowse P White B Back to the future: prohibiting surrogacy for singles, same-sex and shorter-term heterosexual couples in Queensland J Law Med 2013 20 03 638 654 23600195 40 Ekberg M E Ethical, legal and social issues to consider when designing a surrogacy law J Law Med 2014 21 03 728 738 24804538 41 Markens S The global reproductive health market: U.S. media framings and public discourses about transnational surrogacy Soc Sci Med 2012 74 11 1745 1753 10.1016/J.SOCSCIMED.2011.09.013 22014871 42 Aznar J Martínez Peris M Gestational surrogacy: current view Linacre Q 2019 86 01 56 67 10.1177/0024363919830840 32431389 43 Wade K The regulation of surrogacy: a children's rights perspective Child Fam Law Q 2017 29 02 113 131 28781570 44 Sharma R S Social, ethical, medical & legal aspects of surrogacy: an Indian scenario Indian J Med Res 2014 140 (Suppl):S13 S16 25673533 45 Frydman R Contre la grossesse pour autrui (GPA) Gynécol Obstét Fertil 2010 38 03 224 225 10.1016/j.gyobfe.2010.01.005 20206573 46 Makinde O A Olaleye O Makinde O O Huntley S S Brown B Baby factories in Nigeria: starting the discussion toward a national prevention policy Trauma Violence Abuse 2017 18 01 98 105 10.1177/1524838015591588 26209095 47 Overall C Reproductive ‘surrogacy’ and parental licensing Bioethics 2015 29 05 353 361 10.1111/BIOE.12107 25082172 48 Qadeer I The ART of marketing babies Indian J Med Ethics 2010 7 04 209 215 10.20529/IJME.2010.079 22106569 49 Parks J A Care ethics and the global practice of commercial surrogacy Bioethics 2010 24 07 333 340 10.1111/j.1467-8519.2010.01831.x 20690917 50 Hanna J K Revisiting child-based objections to commercial surrogacy Bioethics 2010 24 07 341 347 10.1111/j.1467-8519.2010.01829.x 20690918 51 Blazier J Janssens R Regulating the international surrogacy market:the ethics of commercial surrogacy in the Netherlands and India Med Health Care Philos 2020 23 04 621 630 10.1007/s11019-020-09976-x 32929622 52 Anu Kumar P Inder D Sharma N Surrogacy and women's right to health in India: issues and perspective Indian J Public Health 2013 57 02 65 70 10.4103/0019-557X.114984 23873191 53 Shetty P India's unregulated surrogacy industry Lancet 2012 380 (9854):1633 1634 10.1016/S0140-6736(12)61933-3 23162822 54 Deonandan R Green S van Beinum A Ethical concerns for maternal surrogacy and reproductive tourism J Med Ethics 2012 38 12 742 745 10.1136/MEDETHICS-2012-100551 23047836 55 Yeshua-Katz D Khvorostianov N “Only my husband and my doctor know. And you, girls”: online discussions of stigma coping strategies for Russian surrogate mothers Int J Environ Res Public Health 2021 18 21 11325 10.3390/ijerph182111325 34769841 56 Trowse P Cooper D The agony and the ecstasy: sacrifice and pain for financial gain: have Indian surrogate mothers been exploited by their intended parents in commercial surrogacy arrangements? J Law Med 2018 25 02 388 407 29978644 57 Albert M M La explotación reproductiva de mujeres y el mito de la subrogación altruista: una mirada global al fenómeno de la gestación por sustitución Cuad Bioet 2017 28 93 177 197 28621960 58 Saravanan S Global justice, capabilities approach and commercial surrogacy in India Med Health Care Philos 2015 18 03 295 307 10.1007/S11019-015-9640-y 25894236 59 Damelio J Sorensen K Enhancing autonomy in paid surrogacy Bioethics 2008 22 05 269 277 10.1111/J.1467-8519.2008.00629.x 18447862 60 Naik Africawala A Kapadia S Women's control over decision to participate in surrogacy J Bioeth Inq 2019 16 04 501 514 10.1007/s11673-019-09931-3 31399942 61 Tanderup M Reddy S Patel T Nielsen B B Reproductive ethics in commercial surrogacy: decision-making in IVF clinics in New Delhi, India J Bioeth Inq 2015 12 03 491 501 10.1007/S11673-015-9642-8 26133889 62 Tanderup M Reddy S Patel T Nielsen B B Informed consent in medical decision-making in commercial gestational surrogacy: a mixed methods study in New Delhi, India Acta Obstet Gynecol Scand 2015 94 05 465 472 10.1111/aogs.12576 25581093 63 Yanagihara Y Reconstructing feminist perspectives of women's bodies using a globalized view: The changing surrogacy market in Japan Bioethics 2020 34 06 570 577 10.1111/bioe.12758 32488873 64 Patrone T Is paid surrogacy a form of reproductive prostitution? A kantian perspective Camb Q Healthc Ethics 2018 27 01 109 122 10.1017/S0963180117000445 29214965 65 Beier K Wöhlke S An ethical comparison of living kidney donation and surrogacy: understanding the relational dimension Philos Ethics Humanit Med 2019 14 01 13 10.1186/s13010-019-0080-9 31533778 66 McCartney C Uterine surrogacy is morally equivalent to selling a kidney South Med J 2019 112 01 32 32 10.14423/SMJ.0000000000000910 30608628 67 Wilkinson S Exploitation in international paid surrogacy arrangements J Appl Philos 2016 33 02 125 145 10.1111/JAPP.12138 27471338 68 Stuhmcke A The regulation of commercial surrogacy: The wrong answers to the wrong questions J Law Med 2015 23 02 333 345 26939502 69 Makinde O A Makinde O O Olaleye O Brown B Odimegwu C O Baby factories taint surrogacy in Nigeria Reprod Biomed Online 2016 32 01 6 8 10.1016/J.RBMO.2015.10.001 26602942 70 Mahajan T (Mis)regulation–the case of commercial surrogacy Indian J Med Ethics 2015 12 01 38 41 10.20529/ijme.2015.008 25716437 71 Kirby J Transnational gestational surrogacy: does it have to be exploitative? Am J Bioeth 2014 14 05 24 32 10.1080/15265161.2014.892169 72 Gaffney P Why the “widespread agreement” is wrong: contesting the non-harm arguments for the prohibition of full commercial surrogacy J Law Med 2009 17 02 280 296 19998597 73 Osberg B For your first born child: an ethical defense of the exploitation argument against commercial surrogacy Penn Bioeth J 2006 2 02 42 45 17146914 74 Wilkinson S The exploitation argument against commercial surrogacy Bioethics 2003 17 02 169 187 10.1111/1467-8519.00331 12812183 75 Ferolino A P Camposo M AD Estaño K CL Tacbobo J MR Mothers for others: an interpretative phenomenological analysis of gestational surrogates' child relinquishment experiences J Patient Exp 2020 7 06 1336 1340 10.1177/2374373519892415 33457584 76 Attawet J Wang A Sullivan E ‘Womb for work’ experiences of Thai women and gestational surrogacy practice in Thailand Hum Fertil (Camb) 2021 ••• 1 12 10.1080/14647273.2021.1937716 77 Rozée V Unisa S de La Rochebrochard E Sociodemographic characteristics of 96 Indian surrogates: Are they disadvantaged compared with the general population? PLoS One 2019 14 03 e0214097 10.1371/journal.pone.0214097 30908500 78 Berend Z The social context for surrogates' motivations and satisfaction Reprod Biomed Online 2014 29 04 399 401 10.1016/J.RBMO.2014.07.001 25129693 79 Smietana M Rudrappa S Weis C Moral frameworks of commercial surrogacy within the US, India and Russia Sex Reprod Health Matters 2021 29 01 1 17 10.1080/26410397.2021.1878674 80 Golboni F Jalali A Dinmohammadi M Taghizadeh Z Nouri P Salahsoor M R Factors affecting on couple's decisions to use surrogacy: a qualitative study J Family Reprod Health 2019 13 04 201 208 32518570 81 Hammarberg K Stafford-Bell M Everingham S Intended parents' motivations and information and support needs when seeking extraterritorial compensated surrogacy Reprod Biomed Online 2015 31 05 689 696 10.1016/j.rbmo.2015.08.008 26371710 82 Henrion R La gestation pour autrui pour les personnes de même sexe Rev Prat 2014 64 10 1340 1341 25665307 83 Jacobson H A limited market: the recruitment of gay men as surrogacy clients by the infertility industry in the USA Reprod Biomed Soc Online 2018 7 14 23 10.1016/j.rbms.2018.10.019 30555951 84 van Beers B Bosch L A revolution by stealth: a legal-ethical analysis of the rise of pre-conception authorization of surrogacy agreements New Bioeth 2020 26 04 351 371 10.1080/20502877.2020.1836464 33112204 85 Jackson E Millbank J Karpin I Stuhmcke A Learning from cross-border reproduction Med Law Rev 2017 25 01 23 46 10.1093/MEDLAW/FWW045 28073821 86 Luk J Petrozza J C Evaluation of compliance and range of fees among American Society for Reproductive Medicine-listed egg donor and surrogacy agencies J Reprod Med 2008 53 11 847 852 19097517 87 Gunnarsson Payne J Korolczuk E Mezinska S Surrogacy relationships: a critical interpretative review Ups J Med Sci 2020 125 02 183 191 10.1080/03009734.2020.1725935 32070166 88 Saxena P Mishra A Malik S Surrogacy: ethical and legal issues Indian J Community Med 2012 37 04 211 213 10.4103/0970-0218.103466 23293432 89 Ramskold L A Posner M P Commercial surrogacy: how provisions of monetary remuneration and powers of international law can prevent exploitation of gestational surrogates J Med Ethics 2013 39 06 397 402 10.1136/MEDETHICS-2012-100527 23443211 90 Humbyrd C Fair trade international surrogacy Developing World Bioeth 2009 9 03 111 118 10.1111/J.1471-8847.2009.00257.x 91 ESHRE Task Force on Ethics and Law Shenfield F Pennings G Cohen J Devroey P de Wert G Tarlatzis B ESHRE Task Force on Ethics and Law 10: surrogacy Hum Reprod 2005 20 10 2705 2707 10.1093/humrep/dei147 15979996 92 Kim H H Selecting the optimal gestational carrier: medical, reproductive, and ethical considerations Fertil Steril 2020 113 05 892 896 10.1016/j.fertnstert.2020.03.024 32386618 93 FIGO Committee for Ethical Aspects of Human Reproduction and Women's Health Milliez J Surrogacy Int J Gynaecol Obstet 2008 102 03 312 313 10.1016/J.IJGO.2008.04.016 18603243 94 ACOG Committee Opinion No 660 Summary: family building through gestational surrogacy Obstet Gynecol 2016 127 03 620 621 10.1097/AOG.0000000000001350 26901333 95 Saito Y Matsuo H Survey of Japanese infertile couples' attitudes toward surrogacy J Psychosom Obstet Gynaecol 2009 30 03 156 161 10.1080/01674820802429435 19591054 96 Jadva V Gamble N Prosser H Imrie S Parents' relationship with their surrogate in cross-border and domestic surrogacy arrangements: comparisons by sexual orientation and location Fertil Steril 2019 111 03 562 570 10.1016/j.fertnstert.2018.11.029 30827525 97 Stuvøy I Troublesome reproduction: surrogacy under scrutiny Reprod Biomed Soc Online 2018 7 33 43 10.1016/j.rbms.2018.10.015 30560206 98 Van Zyl L Walker R Beyond altruistic and commercial contract motherhood: the professional model Bioethics 2013 27 07 373 381 10.1111/J.1467-8519.2012.01962.x 22500585 99 Swanson K Letourneau J M Kuppermann M Einerson B D Obstetric morbidity in gestational carrier pregnancies: a population-based study J Assist Reprod Genet 2021 38 01 177 183 10.1007/s10815-020-02000-4 33145724 100 Swanson K Ayala N K Barnes R B Desai N Miller M Yee L M Understanding gestational surrogacy in the United States: a primer for obstetricians and gynecologists Am J Obstet Gynecol 2020 222 04 330 337 10.1016/j.ajog.2020.01.037 31982386 101 Peters H E Schats R Verhoeven M O Mijatovic V de Groot C JM Sandberg J L Gestational surrogacy: results of 10 years of experience in the Netherlands Reprod Biomed Online 2018 37 06 725 731 10.1016/j.rbmo.2018.09.017 30420169 102 Rumpik D Rumpikova T Pohanka M Ventruba P Belaskova S Gestational surrogacy in the Czech Republic Biomed Pap Med Fac Univ Palacky Olomouc Czech Repub 2019 163 02 155 160 10.5507/bp.2018.040 30238935 103 Murugappan G Farland L V Missmer S A Correia K F Anchan R M Ginsburg E S Gestational carrier in assisted reproductive technology Fertil Steril 2018 109 03 420 428 10.1016/J.FERTNSTERT.2017.11.011 29428314 104 Phillips A M Magann E F Whittington J R Whitcombe D D Sandlin A T Surrogacy and Pregnancy Obstet Gynecol Surv 2019 74 09 539 545 10.1097/OGX.0000000000000703 31830299 105 Söderström-Anttila V Blomqvist T Foudila T Experience of in vitro fertilization surrogacy in Finland Acta Obstet Gynecol Scand 2002 81 08 747 752 12174160 106 Birenbaum-Carmeli D Montebruno P Incidence of surrogacy in the USA and Israel and implications on women's health: a quantitative comparison J Assist Reprod Genet 2019 36 12 2459 2469 10.1007/s10815-019-01612-9 31673853 107 Simopoulou M Sfakianoudis K Tsioulou P Rapani A Anifandis G Pantou A Risks in surrogacy considering the embryo: from the preimplantation to the gestational and neonatal period BioMed Res Int 2018 2018 6.287507E6 10.1155/2018/6287507 108 White P M Hidden from view: Canadian gestational surrogacy practices and outcomes, 2001-2012 Reprod Health Matters 2016 24 47 205 217 10.1016/J.RHM.2016.06.006 27578354 109 Wang A Y Dill S K Bowman M Sullivan E A Gestational surrogacy in Australia 2004-2011: treatment, pregnancy and birth outcomes Aust N Z J Obstet Gynaecol 2016 56 03 255 259 10.1111/AJO.12451 26916591 110 Hovav A Cutting out the surrogate: Caesarean sections in the Mexican surrogacy industry Soc Sci Med 2020 256 113063 10.1016/j.socscimed.2020.113063 32464414 111 Ethics Committee of the American Society for Reproductive Medicine Consideration of the gestational carrier: a committee opinion Fertil Steril 2013 99 07 1838 1841 10.1016/j.fertnstert.2013.02.042 23541404 112 Atreya A Kanchan T The ethically challenging trade of forced surrogacy in Nepal Int J Gynaecol Obstet 2018 140 02 254 255 10.1002/IJGO.12362 29072782 113 Goswami L Larmar S A Boddy J The impacts of the Covid-19 pandemic on surrogacy in India: The role of social work Qual Soc Work: Res Pract 2021 20 (1-2):472 478 10.1177/1473325020981082 114 Yau A Friedlander R L Petrini A Holt M C White II D E Shin J Medical and mental health implications of gestational surrogacy Am J Obstet Gynecol 2021 225 03 264 269 10.1016/j.ajog.2021.04.213 33839094 115 Kim H H Family building by same-sex male couples via gestational surrogacy Semin Reprod Med 2017 35 05 408 414 10.1055/S-0037-1607333 29073678 116 Riddle M P Psychological assessment of gestational carrier candidates: current approaches, challenges, and future considerations Fertil Steril 2020 113 05 897 902 10.1016/j.fertnstert.2020.02.104 32312559 117 Galbraith M McLachlan H V Swales J K Commercial agencies and surrogate motherhood: a transaction cost approach Health Care Anal 2005 13 01 11 31 10.1007/s10728-005-2567-3 15889679 118 Frankford D M Bennington L K Ryan J G Womb out sourcing: commercial surrogacy in India MCN Am J Matern Child Nurs 2015 40 05 284 290, quiz E19–E20. Doi: 10.1097/NMC.000000000000016326110574 119 Deonandan R Loncar M Rahman P Omar S Measuring reproductive tourism through an analysis of Indian ART clinic Websites Int J Gen Med 2012 5 763 773 10.2147/IJGM.S36426 23049270 120 Arvidsson A Johnsdotter S Essén B Views of Swedish commissioning parents relating to the exploitation discourse in using transnational surrogacy PLoS One 2015 10 05 e0126518 10.1371/JOURNAL.PONE.0126518 25955178 121 Millbank J The Role of professional facilitators in cross-border assisted reproduction Reprod Biomed Soc Online 2018 6 60 71 10.1016/j.rbms.2018.10.013 30533538 122 Walker N Houlahan L Johnston C Kelleher L International surrogacy: be prudent and well informed Aust Nurs Midwifery J 2013 21 05 17 123 Gezinski L B Karandikar S Levitt A Ghaffarian R “We want to offer you peace of mind”: Marketing of transnational commercial surrogacy services to intended parents Health Mark Q 2017 34 04 302 314 10.1080/07359683.2017.1375244 29173112 124 Hunt J Cross border treatment for infertility: the counselling perspective in the UK Hum Fertil (Camb) 2013 16 01 64 67 10.3109/14647273.2013.770565 23527752 125 Davies T Cross-border reproductive care: quality and safety challenges for the regulator Fertil Steril 2010 94 01 e20 e22 10.1016/j.fertnstert.2009.11.005 20056208 126 Henrion R Bergoignan-Esper C [Gestational surrogacy] Bull Acad Natl Med 2009 193 03 583 618 19883013 127 Gamble N Crossing the line: the legal and ethical problems of foreign surrogacy Reprod Biomed Online 2009 19 02 151 152 10.1016/S1472-6483(10)60064-8 19712546 128 Armour K L An overview of surrogacy around the world: trends, questions and ethical issues Nurs Womens Health 2012 16 03 231 236 10.1111/J.1751-486X.2012.01734.x 22697226 129 James S Chilvers R Havemann D Phelps J Y Avoiding legal pitfalls in surrogacy arrangements Reprod Biomed Online 2010 21 07 862 867 10.1016/j.rbmo.2010.06.037 21050815 130 Crockin S L Edmonds M A Altman A Legal principles and essential surrogacy cases every practitioner should know Fertil Steril 2020 113 05 908 915 10.1016/j.fertnstert.2020.03.015 32327239 131 Lamm E Argumentos para la necesaria regulación de la gestación por sustitución Gac Sanit 2017 31 06 539 540 10.1016/J.GACETA.2017.04.009 28709691 132 Sifris A The Family Courts and parentage of children conceived through overseas commercial surrogacy arrangements: A child-centred approach J Law Med 2015 23 02 396 412 26939506 133 Brandão P Ceschin N Gómez V H The pathway of female couples in a fertility clinic Rev Bras Ginecol Obstet 2022 44 07 660 666 10.1055/s-0042-1744444 35668678 134 Chang C L Surrogate motherhood Formos J Med Humanit 2004 5 12 48 62 135 Hague Conference on Private International LawA study of legal parentage and the issues arising from international surrogacy arrangements. The Hague;2014 136 Johnson K M Contingent maternities? Maternal claims-making in third party reproduction Sociol Health Illn 2017 39 08 1349 1364 10.1111/1467-9566.12589 28594091 137 Thompson M Plater D An Issue that is not Going Away: Recent Developments in Surrogacy in South Australia J Bioeth Inq 2019 16 04 477 481 10.1007/s11673-019-09952-y 31823187 138 Urquiza M F Carretero I Quaini F M Inciarte F Pasqualini R A Pasqualini R S Subrogación uterina. Aspectos médicos y jurídicos del primer caso con sustento legal en la Argentina Medicina (B Aires) 2014 74 03 233 238 24918676 139 Kalantry S Regulating markets for gestational care: comparative perspectives on surrogacy in the United States and India Cornell J Law Public Policy 2018 27 03 685 715 30382700 140 Purshouse C Bracegirdle K The problem of unenforceable surrogacy contracts: can unjust enrichment provide a solution? Med Law Rev 2018 26 04 557 584 10.1093/medlaw/fwy001 29425316 141 Oultram S One mum too few: maternal status in host surrogate motherhood arrangements J Med Ethics 2015 41 06 470 473 10.1136/MEDETHICS-2012-100949 25139932 142 Latham S R The United Kingdom revisits its surrogacy law Hastings Cent Rep 2020 50 01 6 7 10.1002/hast.1076 143 Igareda González N Legal and ethical issues in cross-border gestational surrogacy Fertil Steril 2020 113 05 916 919 10.1016/j.fertnstert.2020.03.003 32327242 144 Bassan S Different but same: a call for a joint pro-active regulation of cross-border egg and surrogacy markets Health Matrix 2018 28 01 323 374 30636839 145 Couture V Drouin R Moutquin J M Monnier P Bouffard C Reproductive outsourcing: an empirical ethics account of cross-border reproductive care in Canada J Med Ethics 2019 45 01 41 47 10.1136/medethics-2017-104515 30301813 146 Lozanski K Transnational surrogacy: Canada's contradictions Soc Sci Med 2015 124 383 390 10.1016/J.SOCSCIMED.2014.10.003 25308235 147 Nelson E Global trade and assisted reproductive technologies: regulatory challenges in international surrogacy J Law Med Ethics 2013 41 01 240 253 10.1111/JLME.12016 23581668 148 Deomampo D Defining parents, making citizens: nationality and citizenship in transnational surrogacy Med Anthropol 2015 34 03 210 225 10.1080/01459740.2014.890195 24611832 149 Sifris R Sifris A Parentage, surrogacy and the perplexing state of australian law: a missed opportunity J Law Med 2019 27 02 369 386 32129042 150 Merchant J Les nouvelles technologies de reproduction - État des lieux aux États-Unis Med Sci (Paris) 2019 35 03 253 257 10.1051/medsci/2019032 30931910 151 Arvidsson A Johnsdotter S Emmelin M Essén B Being questioned as parents: An interview study with Swedish commissioning parents using transnational surrogacy Reprod Biomed Soc Online 2018 8 23 31 10.1016/j.rbms.2018.08.001 30911689 152 Courduriès J At the nation's doorstep: the fate of children in France born via surrogacy Reprod Biomed Soc Online 2018 7 47 54 10.1016/j.rbms.2018.11.003 30723815 153 Margaria A Parenthood and cross-border surrogacy: what is ‘new’? The ECtHR's first advisory opinion Med Law Rev 2020 28 02 412 425 10.1093/medlaw/fwz042 32057071 154 Marinelli S Medically-assisted procreation and the rise of off-center, new types of “parenthood”: it is incumbent upon lawmakers to intervene Clin Ter 2019 170 04 e241 e244 10.7417/CT.2019.2140 31304508 155 Birenbaum-Carmeli D Thirty-five years of assisted reproductive technologies in Israel Reprod Biomed Soc Online 2016 2 16 23 10.1016/J.RBMS.2016.05.004 29892712 156 Whittaker A Speier A “Cycling overseas”: care, commodification, and stratification in cross-border reproductive travel Med Anthropol 2010 29 04 363 383 10.1080/01459740.2010.501313 21082483 157 Piersanti V Consalvo F Signore F Del Rio A Zaami S Surrogacy and “procreative tourism”. What does the future hold from the ethical and legal perspectives? Medicina (Kaunas) 2021 57 01 47 10.3390/medicina57010047 33429930 158 Jacobson H Cross-border reproductive care in the USA: Who comes, why do they come, what do they purchase? Reprod Biomed Soc Online 2020 11 42 47 10.1016/j.rbms.2020.09.003 33204864 159 Zannettino L Lines L Grant J de Lacey S L Untangling the threads: stakeholder perspectives of the legal and ethical issues involved in preparing Australian consumers for commercial surrogacy overseas J Law Med 2019 27 01 94 107 31682344 160 Whittaker A Inhorn M C Shenfield F Globalised quests for assisted conception: Reproductive travel for infertility and involuntary childlessness Glob Public Health 2019 14 12 1669 1688 10.1080/17441692.2019.1627479 31204900 161 Salama M Isachenko V Isachenko E Rahimi G Mallmann P Westphal L M Cross border reproductive care (CBRC): a growing global phenomenon with multidimensional implications (a systematic and critical review) J Assist Reprod Genet 2018 35 07 1277 1288 10.1007/s10815-018-1181-x 29808382 162 Gerrits T Reproductive travel to Ghana: testimonies, transnational relationships, and stratified reproduction Med Anthropol 2018 37 02 131 144 10.1080/01459740.2017.1419223 29333881 163 ESHRE Taskforce on Cross Border Reproductive Care Shenfield F de Mouzon J Pennings G Ferraretti A P Nyboe Andersen A de Wert G Cross border reproductive care in six European countries Hum Reprod 2010 25 06 1361 1368 10.1093/humrep/deq057 20348165 164 Deonandan R Recent trends in reproductive tourism and international surrogacy: ethical considerations and challenges for policy Risk Manag Healthc Policy 2015 8 111 119 10.2147/RMHP.S63862 26316832 165 Pérez Navarro P Surrogacy wars: notes for a radical theory of the politics of reproduction J Homosex 2020 67 05 577 599 10.1080/00918369.2018.1553351 30562154 166 König A Jacobson H Reprowebs: a conceptual approach to elasticity and change in the global assisted reproduction industry Biosocieties 2021 ••• 1 23 10.1057/s41292-021-00260-6 167 Gürtin Z B Banning reproductive travel: Turkey's ART legislation and third-party assisted reproduction Reprod Biomed Online 2011 23 05 555 564 10.1016/J.RBMO.2011.08.004 21962527 168 O'Callaghan E Surrogacy reform and its impact on the child's right to birth registration Reprod Biomed Soc Online 2021 13 46 50 10.1016/J.RBMS.2021.06.001 34258448 169 Hibino Y Non-commercial surrogacy in Thailand: ethical, legal, and social implications in local and global contexts Asian Bioeth Rev 2020 12 02 135 147 10.1007/s41649-020-00126-2 33717334 170 Raposo V L The new Portuguese law on surrogacy - The story of how a promising law does not really regulate surrogacy arrangements JBRA Assist Reprod 2017 21 03 230 239 10.5935/1518-0557.20170044 28837033 171 Nakash A Herdiman J Surrogacy J Obstet Gynaecol 2007 27 03 246 251 10.1080/01443610701194788 17464803 172 Whittaker A From ‘Mung Ming’ to ‘Baby Gammy’: a local history of assisted reproduction in Thailand Reprod Biomed Soc Online 2016 2 71 78 10.1016/J.RBMS.2016.05.005 29892719 173 Attawet J Mapping transnational commercial surrogacy arrangements in South and Southeast Asia Med Leg J 2021 89 02 128 132 10.1177/0025817220985099 33715522 174 Chaudhuri M New Indian visa rules exclude single people and gay couples from child surrogacy BMJ 2013 346 f475 10.1136/BMJ.F475 23349412 175 Alabi O J Perceptions of surrogacy within the Yoruba socio-cultural context of Ado-Ekiti, Nigeria F1000 Res 2020 9 103 10.12688/f1000research.20999.3 176 Ahmad N An international view of surgically assisted conception and surrogacy tourism Med Leg J 2011 79 (Pt 4):135 145 10.1258/MLJ.2011.011029 22156300 177 Nazari Tavakkoli S Personhood and moral status of the embryo: it's effect on validity of surrogacy contract revocation according to Shia jurisprudence perspective Int J Fertil Steril 2017 11 03 226 233 10.22074/ijfs.2017.4970 28868846 178 Bokek-Cohen Y Tarabeih M What do Sunni Muslims think about religiously forbidden reproductive options? Hum Fertil (Camb) 2021 ••• 1 12; [ ahead of print ]10.1080/14647273.2021.1921289 179 Aramesh K Iran's experience with surrogate motherhood: an Islamic view and ethical concerns J Med Ethics 2009 35 05 320 322 10.1136/jme.2008.027763 19407039 180 Muaygil R A Reexamining the prohibition of gestational surrogacy in Sunni Islam Developing World Bioeth 2017 17 02 112 120 10.1111/DEWB.12133 181 Saadeh R Abdulrahim N Alfaqih M Khader Y Attitude of Jordanian health care workers toward surrogacy J Family Reprod Health 2020 14 01 5 13 32863833 182 Mustafa A G Alzoubi K H Khabour O F Alfaqih M A Perspectives and attitudes of Jordanian medical and paramedical students toward surrogate pregnancy Int J Womens Health 2018 10 617 622 10.2147/IJWH.S184434 30410408 183 Salehi K Shakour M Pashaei Sabet F Alizadeh S The opinion of Iranian students about the society's perception on using surrogacy as an infertility treatment in the future community Sex Reprod Healthc 2015 6 01 19 22 10.1016/J.SRHC.2014.06.005 25637420 184 Wolowelsky J B Grazi R V Maternal identity for Orthodox Jewish couples using donor oocytes or surrogacy is not well established J Assist Reprod Genet 2020 37 05 1261 10.1007/s10815-020-01760-3 32239334 185 Deonandan R Thoughts on the ethics of gestational surrogacy: perspectives from religions, Western liberalism, and comparisons with adoption J Assist Reprod Genet 2020 37 02 269 279 10.1007/s10815-019-01647-y 31897847 186 Roth A E Wang S W Popular repugnance contrasts with legal bans on controversial markets Proc Natl Acad Sci U S A 2020 117 33 19792 19798 10.1073/PNAS.2005828117 32727903 187 Yamamoto N Hirata T Izumi G Nakazawa A Fukuda S Neriishi K A survey of public attitudes towards third-party reproduction in Japan in 2014 PLoS One 2018 13 10 e0198499 10.1371/journal.pone.0198499 30379816 188 Petitfils C Munoz Sastre M T Sorum P C Mullet E Mapping people's views regarding the acceptability of surrogate motherhood J Reprod Infant Psychol 2017 35 01 65 76 10.1080/02646838.2016.1222358 29517291 189 de Costa C Compensated surrogacy - What do Australians think? Aust N Z J Obstet Gynaecol 2016 56 06 549 551 10.1111/AJO.12572 27914112 190 Tremellen K Everingham S For love or money? Australian attitudes to financially compensated (commercial) surrogacy Aust N Z J Obstet Gynaecol 2016 56 06 558 563 10.1111/AJO.12559 27757962 191 Suzuki K Hoshi K Minai J Yanaihara T Takeda Y Yamagata Z Analysis of national representative opinion surveys concerning gestational surrogacy in Japan Eur J Obstet Gynecol Reprod Biol 2006 126 01 39 47 10.1016/j.ejogrb.2005.07.030 16171926 192 Stenfelt C Armuand G Wånggren K Skoog Svanberg A Sydsjö G Attitudes toward surrogacy among doctors working in reproductive medicine and obstetric care in Sweden Acta Obstet Gynecol Scand 2018 97 09 1114 1121 10.1111/AOGS.13342 29512820 193 Essén B Johnsdotter S Transnational surrogacy - reproductive rights for whom? Acta Obstet Gynecol Scand 2015 94 05 449 450 10.1111/AOGS.12636 25869766 194 Bruce-Hickman K Kirkland L Ba-Obeid T The attitudes and knowledge of medical students towards surrogacy J Obstet Gynaecol 2009 29 03 229 232 10.1080/01443610802712926 19358032 195 Merchant J Dead-end in sight: France struggles with surrogacy and cross-border practices New Bioeth 2020 26 04 314 327 10.1080/20502877.2020.1835207 33164712 196 Creux H Diaz M Grynberg M Papaxanthos-Roche A Chansel-Debordeaux L Jimenez C National survey on the opinions of French specialists in assisted reproductive technologies about social issues impacting the future revision of the French Bioethics laws J Gynecol Obstet Hum Reprod 2020 49 09 101902 10.1016/j.jogoh.2020.101902 32889113 197 Musavi S Mashhadi Abdolahi H Ghojazadeh M Abbasalizad Farhangi M Nikniaz Z Nikniaz L Infertile women's opinion concerning gestational surrogacy: a systematic review and meta-analysis Iran J Public Health 2020 49 08 1432 1438 10.18502/ijph.v49i8.3866 33083319 198 Baykal B Korkmaz C Ceyhan S T Goktolga U Baser I Opinions of infertile Turkish women on gamete donation and gestational surrogacy Fertil Steril 2008 89 04 817 822 10.1016/J.FERTNSTERT.2007.04.022 18406837 199 Maftei A Holman A C Moral women, immoral technologies? Romanian women's perceptions of assisted reproductive technologies versus adoption New Bioeth 2020 26 03 253 272 10.1080/20502877.2020.1796256 32716279 200 Kian E M Riazi H Bashirian S Attitudes of Iranian infertile couples toward surrogacy J Hum Reprod Sci 2014 7 01 47 51 10.4103/0974-1208.130847 24829531 201 Rahmani A Howard F Sattarzadeh N Ferguson C Asgari A Ebrahimi H Viewpoints of fertile women on gestational surrogacy in East Azerbaijan Province, Iran Indian J Med Ethics 2014 11 01 29 33 10.20529/IJME.2014.008 24509106 202 Perkins K M Boulet S L Levine A D Jamieson D J Kissin D M Differences in the utilization of gestational surrogacy between states in the U.S Reprod Biomed Soc Online 2017 5 1 4 10.1016/j.rbms.2017.08.002 29774269 203 Baylis F O Canada's prohibition on payment for surrogacy, eggs, and sperm J Obstet Gynaecol Can 2018 40 12 1569 1570 10.1016/j.jogc.2018.08.005 30361159 204 Chen M Wombs for rent: an examination of prohibitory and regulatory approaches to governing preconception arrangements Health Law Can 2003 23 03 33 50 12674583 205 Cabra R Alduncin A Cabra J R Ek L H Briceño M Mendoza P B Gestational surrogacy. Medical, psychological and legal aspects: 9 years of experience in Mexico Hum Reprod Open 2018 2018 01 hox029 10.1093/HROPEN/HOX029 30895241 206 Torres G Shapiro A Mackey T K A review of surrogate motherhood regulation in south American countries: pointing to a need for an international legal framework BMC Pregnancy Childbirth 2019 19 01 46 10.1186/s12884-019-2182-1 30691390 207 Gianaroli L Ferraretti A P Magli M C Sgargi S Current regulatory arrangements for assisted conception treatment in European countries Eur J Obstet Gynecol Reprod Biol 2016 207 211 213 10.1016/J.EJOGRB.2016.10.002 27832922 208 Calhaz-Jorge C De Geyter C H Kupka M S Wyns C Mocanu E Montrenko T Survey on ART and IUI: legislation, regulation, funding and registries in European countries: The European IVF-monitoring Consortium (EIM) for the European Society of Human Reproduction and Embryology (ESHRE) Hum Reprod Open 2020 2020 01 hoz044 10.1093/hropen/hoz044 32042927 209 Guerra-Palmero M J Contra la llamada gestación subrogada. Derechos humanos y justicia global versus bioética neoliberal Gac Sanit 2017 31 06 535 538 10.1016/J.GACETA.2017.05.009 28711230 210 März J W Challenges posed by transnational commercial surrogacy: the jurisprudence of the European Court of Human Rights Eur J Health Law 2021 28 03 263 280 10.1163/15718093-BJA10045 33975281 211 Mulligan A Identity rights and sensitive ethical questions: the European convention on human rights and the regulation of surrogacy arrangements Med Law Rev 2018 26 03 449 475 10.1093/MEDLAW/FWX066 29415260 212 Griffith R Surrogacy: why the Law Commission is reviewing current arrangements Br J Nurs 2018 27 03 164 165 10.12968/BJON.2018.27.3.164 29412031 213 Ramsey J Regulating surrogacy–a contravention of human rights? Med Law Int 2000 5 01 45 64 10.1177/096853320000500103 15040377 214 Depadt V La GPA: vers la légalisation? J Int Bioethique Ethique Sci 2015 26 03 139 151, 267–268 10.3917/jib.262.0139 27356352 215 Foret F Bolzonar F How the European Union deals with surrogacy. Birth without borders as a driver of value conflicts? Gend Technol Dev 2021 25 02 131 145 10.1080/09718524.2021.1933347 216 de Montgolfier S Mirkovic A Maternité pour autrui : du désir d'enfant à l'enfant à tout prix Med Sci (Paris) 2009 25 04 419 422 10.1051/MEDSCI/2009254419 19409196 217 Samuels A Surrogacy and the law: Possible reforms Med Leg J 2020 88 03 144 147 10.1177/0025817220923688 32437632 218 Dickens B M Paid surrogacy abroad does not violate public policy: UK Supreme Court Int J Gynaecol Obstet 2020 150 01 129 133 10.1002/ijgo.13197 32402094 219 Jadva V Prosser H Gamble N Cross-border and domestic surrogacy in the UK context: an exploration of practical and legal decision-making Hum Fertil (Camb) 2021 24 02 93 104 10.1080/14647273.2018.1540801 30537445 220 Fenton-Glynn C Outsourcing ethical dilemmas: regulating international surrogacy arrangements Med Law Rev 2016 24 01 59 75 10.1093/MEDLAW/FWV044 26785890 221 Dyer C UK should change law on surrogacy to help commissioning parents, report says BMJ 2015 351 h6302 10.1136/BMJ.H6302 26596286 222 Calvo Caravaca A L Carrascosa González J Gestación por sustitución y derecho internacional privado. Más allá del Tribunal Supremo y del Tribunal Europeo de Derechos Humanos Cuad Derecho Transnac. 2015 7 02 45 113 223 Marre D San Román B Guerra D On reproductive work in Spain: transnational adoption, egg donation, surrogacy Med Anthropol 2018 37 02 158 173 10.1080/01459740.2017.1361947 28768116 224 Stuvøy I Accounting for the money-made parenthood of transnational surrogacy Anthropol Med 2018 25 03 280 295 10.1080/13648470.2017.1392100 29927615 225 Pande A Transnational commercial surrogacy in India: gifts for global sisters? Reprod Biomed Online 2011 23 05 618 625 10.1016/J.RBMO.2011.07.007 21958916 226 Menezes R G Bhagavath P Adiga P K Surrogacy in India J Forensic Leg Med 2007 14 05 307 10.1016/j.jcfm.2006.05.006 16914355 227 Saran J Padubidri J R New laws ban commercial surrogacy in India Med Leg J 2020 88 03 148 150 10.1177/0025817219891881 32216696 228 Patel N H Jadeja Y D Bhadarka H K Patel M N Patel N H Sodagar N R Insight into different aspects of surrogacy practices J Hum Reprod Sci 2018 11 03 212 218 10.4103/jhrs.JHRS_138_17 30568349 229 Timms O Ending commercial surrogacy in India: significance of the Surrogacy (Regulation) Bill, 2016 Indian J Med Ethics 2018 3 02 99 102 10.20529/IJME.2018.019 29550749 230 Howard S Proposed ban on foreigners using Indian surrogacy services sparks protests BMJ 2015 351 h5854 10.1136/BMJ.H5854 26527734 231 Heng B C Proposed ethical guidelines and legislative framework for permitting gestational surrogacy in Singapore Reprod Biomed Online 2007 15 01 7 11 10.1016/S1472-6483(10)60352-5 232 Asia and Oceania Federation of Obstetrics and Gynaecology Li H WR Tank J Haththotuwa R Updated status of assisted reproductive technology activities in the Asia-Oceania region J Obstet Gynaecol Res 2018 44 09 1667 1672 10.1111/jog.13742 30058253 233 Stasi A Protection for children born through assisted reproductive technologies act, B.E. 2558: the changing profile of surrogacy in Thailand Clin Med Insights Reprod Health 2017 11 1.179558117749603E15 10.1177/1179558117749603 234 Kisu I Banno K Mihara M Iida T Yoshimura Y Current status of surrogacy in Japan and uterine transplantation research Eur J Obstet Gynecol Reprod Biol 2011 158 02 135 140 10.1016/J.EJOGRB.2011.04.037 21632170 235 Chiang W T Chou T Y A survey of judicial decisions concerning surrogacy disputes in Taiwan Taiwan J Obstet Gynecol 2018 57 04 517 521 10.1016/j.tjog.2018.06.006 30122570 236 Tang Q Surrogacy in China: public opinion, litigations, and court rulings Asian Soc Sci 2019 15 10 84 10.5539/ass.v15n10p84 237 Ding C Surrogacy litigation in China and beyond J Law Biosci 2015 2 01 33 55 10.1093/JLB/LSU036 27774179 238 Setabouha F Surrogacy in Australia and Middle Eastern Countries [Internet] 2016[cited 2022 Apr 21]. Available from:https://www.academia.edu/40552161/Surrogacy_in_Australia_and_Middle_Eastern_Countries 239 Newson A J Compensated transnational surrogacy in Australia: time for a comprehensive review Med J Aust 2016 204 01 33 35 10.5694/MJA15.00166 26763818 240 van Zyl L Walker R Surrogacy, compensation, and legal parentage: against the adoption model J Bioeth Inq 2015 12 03 383 387 10.1007/S11673-015-9646-4 26133892 241 Stuhmcke A The criminal act of commercial surrogacy in Australia: a call for review J Law Med 2011 18 03 601 613 21528744 242 Kirkman M Bourne K Fisher J Johnson L Hammarberg K Gamete donors' expectations and experiences of contact with their donor offspring Hum Reprod 2014 29 04 731 738 10.1093/humrep/deu027 24549216 243 Stuhmcke A Looking backwards, looking forwards: judicial and legislative trends in the regulation of surrogate motherhood in the UK and Australia Aust J Fam Law 2004 18 01 13 40 17058335 244 Scherer L Curran M Alvarez M Expanding Kenya's protected areas under the Convention on Biological Diversity to maximize coverage of plant diversity Conserv Biol 2017 31 02 302 310 10.1111/cobi.12792 27346759 245 Feiglin J Savulescu J A new ethical model of commercial surrogacy arrangements for Australia J Law Med 2018 25 04 919 928 29978675 246 Millbank J Rethinking “commercial” surrogacy in Australia J Bioeth Inq 2015 12 03 477 490 10.1007/S11673-014-9557-9 25015592 247 Shalev C Moreno A Eyal H Leibel M Schuz R Eldar-Geva T Ethics and regulation of inter-country medically assisted reproduction: a call for action Isr J Health Policy Res 2016 5 59 10.1186/s13584-016-0117-0 27980721 248 Kapfhamer J Van Voorhis B Gestational surrogacy: a call for safer practice Fertil Steril 2016 106 02 270 271 10.1016/J.FERTNSTERT.2016.04.028 27181925 249 Sifris R Ludlow K Sifris A Commercial surrogacy: what role for law in Australia? J Law Med 2015 23 02 275 296 26939494 250 Schover L R Cross-border surrogacy: the case of Baby Gammy highlights the need for global agreement on protections for all parties Fertil Steril 2014 102 05 1258 1259 10.1016/J.FERTNSTERT.2014.08.017 25241370 251 Drabiak K Wegner C Fredland V Helft P R Ethics, law, and commercial surrogacy: a call for uniformity J Law Med Ethics 2007 35 02 300 309 10.1111/j.1748-720X.2007.00139.x 17518856
PMC009xxxxxx/PMC9846656.txt
==== Front J Behav Med J Behav Med Journal of Behavioral Medicine 0160-7715 1573-3521 Springer US New York 36652087 390 10.1007/s10865-022-00390-7 Correction Correction to: A remotely delivered, peer-led intervention to improve physical activity and quality of life in younger breast cancer survivors https://orcid.org/0000-0003-1380-1774 Weiner Lauren S. [email protected] 12 Nagel Stori 3 Su H. Irene 24 Hurst Samantha 1 Levy Susan S. 5 Arredondo Elva M. 67 Hekler Eric 18 Hartman Sheri J. 12 1 grid.266100.3 0000 0001 2107 4242 Herbert Wertheim School of Public Health and Human Longevity Science, UC San Diego, La Jolla, CA USA 2 grid.516081.b 0000 0000 9217 9714 UC San Diego Moores Cancer Center, 3855 Health Sciences Drive, La Jolla, CA USA 3 Haus of Volta, Murrieta, CA USA 4 grid.266100.3 0000 0001 2107 4242 Division of Reproductive Endocrinology and Infertility, Department of Obstetrics, Gynecology, and Reproductive Sciences, UC San Diego, La Jolla, CA USA 5 grid.263081.e 0000 0001 0790 1491 School of Exercise and Nutritional Sciences, San Diego State University, San Diego, CA USA 6 grid.263081.e 0000 0001 0790 1491 Institute for Behavioral and Community Health, San Diego State University Research Foundation, San Diego, CA USA 7 grid.263081.e 0000 0001 0790 1491 School of Public Health, San Diego State University, San Diego, CA USA 8 Center for Wireless & Population Health Systems, Qualcomm Institute, San Diego, CA USA 18 1 2023 18 1 2023 2023 46 4 707708 26 12 2022 27 12 2022 © The Author(s) 2023 https://creativecommons.org/licenses/by/4.0/ Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. issue-copyright-statement© Springer Science+Business Media, LLC, part of Springer Nature 2023 ==== Body pmcCorrection to: Journal of Behavioral Medicine 10.1007/s10865-022-00381-8 Due to publisher processing error, the following manuscript sessions were omitted from the original version of this article: Supplemental Material 1, Acknowledgements. Additionally, the original version of this article unfortunately contained errors in Table 4. The omitted sections and corrected Table 4 are given below.Table 4 Physical activity and quality of life outcomes at T0, T1, and T2 and changes between time points among Pink Body Spirit participants Outcome N Time point Change Repeated measures ANOVA T0 T1 T2 T0 to T1 T0 to T2 T1 to T2 F value p value n2 M (SD) M (SD) M (SD) M (SD) M (SD) M (SD) MVPAa 30 90.2 (49.9) 129.9 (84.2) 127.4 (86.8)  + 39.7 (63.4)  + 37.2 (76.4)  − 2.5 (71.6) 5.94  < 0.005 0.06 Strength trainingb 32 0.1 (0.3) 0.9* (1.3) 1.1* (1.4)  + 0.8 (1.2)  + 1.0 (1.4)  + 0.2 (1.2) 11.2  < 0.001 0.14 Stretching & flexibilityc 31 0.23 (0.5) 1.06* (1.6) 1.32* (1.7)  + 0.8 (1.2)  + 1.1 (1.5)  + 0.3 (1.7) 8.46  < 0.001 0.11 Body imaged 34 17.2 (8.4) 13.2 (8.2) 11.3* (7.1)  − 4.0 (5.7)  − 5.9 (6.1)  − 1.9 (4.1) 21.24  < 0.001 0.09 Sexual functione,f 34 14.8 (10.7) 15.2 (11.0) 15.2 (11.9)  + 0.4 (6.2)  + 0.1 (5.9)  − 0.1 (7.0) 0.117  < 0.889 0.004 Fatigueg,h 33 60.2 (8.2) 54.3* (8.1) 54.0* (9.0)  − 5.8 (8.5)  − 6.1 (8.4)  − 0.3 (7.5) 11.94  < 0.001 0.1 Anxietyg,i 33 60.7 (7.6) 57.1 (9.5) 57.0 (9.3)  − 3.6 (9.7)  − 3.7 (7.5)  − 0.1 (9.8) 3.58  < 0.034 0.04 Depressiong 34 54.0 (9.0) 52.2 (9.2) 51.7 (8.0)  − 1.8 (7.1)  − 2.0 (6.2)  − 0.5 (6.6) 2.18  < 0.121 0.01 Emotional supportg 34 46.5 (8.2) 49.2 (8.6) 50.8 (9.6)  + 2.7 (6.4)  + 4.3 (6.3)  + 1.6 (6.4) 7.86  < 0.001 0.04 M Mean, SD Standard Deviation aMVPA Moderate to Vigorous Physical Activity measured with ActiGraph accelerometer, min/week, removed 4 outliers due to high MVPA at T0; bdays/week, removed 2 outliers with high strength training at T0; cdays/week, removed 3 outliers with high stretching & flexibility at T0 or T1; dScore range = 10–40; escore range = 2–36; fRemoved 12 participants who were not sexually active at T0; gPROMIS measures reported as standardized t-score with mean = 50 and SD = 10; hRemoved 1 outliers with high low fatigue at T0; iRemoved 1 outlier with low anxiety at T0; *significantly (p < 0.05) different from T0; η2 = eta-squared effect size Supplementary Information Below is the link to the electronic supplementary material.Supplementary file1 (DOCX 16 kb) Acknowledgements The authors would like to thank the peer mentors and participants for their commitment to the study, particularly through the COVID-19 pandemic. The authors also wish to acknowledge the UC San Diego interns and research staff who contributed to this project, including Sierra Ottilie-Kovelman, Valerie Zhang, Danya Greenberg, May Le, Umme Sheik, Shoshi Barkai, Lindsay Dillon, Brittany Lewars, Carissa Jantz, Hillary Nguyen, and Erin Deiotte. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
PMC009xxxxxx/PMC9870081.txt
==== Front JCI Insight JCI Insight JCI Insight JCI Insight 2379-3708 American Society for Clinical Investigation 36625347 162987 10.1172/jci.insight.162987 Research Article CCN2 deficiency in smooth muscle cells triggers cell reprogramming and aggravates aneurysm development Wang Yu [email protected] 1 Liu Xuesong [email protected] 12 Xu Qian [email protected] 1 Xu Wei [email protected] 1 Zhou Xianming [email protected] 13 Leask Andrew [email protected] 4 http://orcid.org/0000-0002-9663-3886 Lin Zhiyong [email protected] 1 1 Cardiology Division, Emory University School of Medicine, Atlanta, Georgia, USA. 2 Department of Cardiology, The First Affiliated Hospital of Kunming Medical University, Kunming, China. 3 Department of Cardiovascular Surgery, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China. 4 University of Saskatchewan, College of Dentistry, Saskatoon, Saskatchewan, Canada. Address correspondence to: Zhiyong Lin, 101 Woodruff Circle, Room 3004, Atlanta, Georgia 30322, USA. Phone: 404.712.0974; Email: [email protected]. Authorship note: YW and XL contributed equally to this work. 10 1 2023 10 1 2023 10 1 2023 8 1 e16298724 6 2022 17 11 2022 © 2023 Wang et al. 2023 Wang et al. https://creativecommons.org/licenses/by/4.0/ This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. This article is available online at https://insight.jci.org/articles/view/162987 Vascular smooth muscle cell (SMC) phenotypic switching is widely recognized as a key mechanism responsible for the pathogenesis of several aortic diseases, such as aortic aneurysm. Cellular communication network factor 2 (CCN2), often upregulated in human pathologies and animal disease models, exerts myriad context-dependent biological functions. However, current understanding of the role of SMC-CCN2 in SMC phenotypic switching and its function in the pathology of abdominal aortic aneurysm (AAA) is lacking. Here, we show that SMC-restricted CCN2 deficiency causes AAA in the infrarenal aorta of angiotensin II–infused (Ang II–infused) hypercholesterolemic mice at a similar anatomic location to human AAA. Notably, the resistance of naive C57BL/6 WT mice to Ang II–induced AAA formation is lost upon silencing of CCN2 in SMC. Furthermore, the pro-AAA phenotype of SMC-CCN2-KO mice is recapitulated in a different model that involves the application of elastase–β-aminopropionitrile. Mechanistically, our findings reveal that CCN2 intersects with TGF-β signaling and regulates SMC marker expression. Deficiency of CCN2 triggers SMC reprograming associated with alterations in Krüppel-like factor 4 and contractile marker expression, and this reprograming likely contributes to the development of AAA in mice. These results identify SMC-CCN2 as potentially a novel regulator of SMC phenotypic switching and AA biology. Vascular Biology Extracellular matrix NHLBI5R01HL144741,5R01HL152074 American Heart Association https://doi.org/10.13039/100000968 20TPA35490431 ==== Body pmcIntroduction Aortic aneurysm is a life-threatening vascular pathology that poses significant challenges to health care systems worldwide. Despite the advances in medical science and technology, surgical or endovascular interventions remain the only available therapies for high-risk patients. Findings from human and animal studies have suggested common aneurysm pathophysiology involving vascular endothelial and smooth muscle cell (SMC) dysfunction, extracellular matrix (ECM) deterioration, and increased vascular inflammation (1, 2). However, a comprehensive understanding of the aneurysm biology has yet to be acquired. Sticking points include lack of early-stage tissue samples and the presence of confounding factors, such as atherosclerosis and hypertension, that can be major obstacles to the next therapeutic leap. In addition, even with the help of animal models, very few studies were carefully designed to dissect the early molecular triggers in a cell-specific manner, which the multifactorial, multistage nature of aneurysm development is likely to rely upon. Cellular communication network factor 2 (CCN2), also known as connective tissue growth factor (CTGF), is a member of the CCN matricellular protein family. Similar to the majority of the other CCN proteins, CCN2 plays critical roles in the regulation of normal cell function and signaling. CCN2 conveys its biological effects through interaction with specific receptors (e.g., integrins, LRP1, HSPG) (3), the ECM protein fibronectin (4), and cytokines such as TGF-β and VEGF (5, 6). Although initially best known for promoting tissue fibrosis in multiple fibrotic diseases, CCN2 has also been implicated in cancer (7), inflammatory diseases (8), neural (9) and ocular disorders (10), and autoimmune diseases (11, 12). In vascular pathologies, increase of CCN2 has been reported in atherosclerosis (13, 14), aortic aneurysms (15), aorta dissection (16), and restenosis (17). Published studies hint at a causal role of elevated CCN2 in these diseases, largely owing to CCN2’s established function in promoting the formation and remodeling of the ECM (18, 19). However, in part due to the lack of critical genetic evidence derived from transgenic animals, the precise physiological role of CCN2 from different cellular origins in vivo remains largely undefined for vascular diseases including aortic aneurysm. SMC integrity and function are central for maintaining vascular homeostasis. Alterations in SMC function, as well as transdifferentiation of SMC populations, have long been hypothesized to play important roles in aneurysm formation, with evidence documented in both abdominal aortic aneurysm (AAA) and thoracic aortic aneurysm (TAA). In patients with Marfan or Loeys-Dietz syndrome, aneurysm development has been associated with an activated TGF-β signaling in SMCs (20). However, a recent study using SMC-specific TGFBR2-deficient mice demonstrated that a loss of smooth muscle TGF-β signaling, when combined with hyperlipidemia, stimulates the transdifferentiation of media SMC to a mesenchymal stem cell–like (MSC-like) state and promotes aortic aneurysm (21). Given that (a) CCN2 is a potential downstream target of TGF-β signaling (22) and (b) CCN2 has been suggested to potentiate TGF-β signaling, including TGF-β’s effect on inducing SMC contractile marks (3), it is conceivable to speculate that SMC-specific CCN2 may be critical for maintaining vascular homeostasis. The present study uses tissue-specific inducible CCN2-deficient mice (CCN2fl/fl–Myh11–Cre-ERT2) and murine models of AAA (induced by systemic angiotensin II [Ang II] infusion or local adventitia elastase exposure) to characterize the role of SMC-specific CCN2 in the development of AAA. We show that ablation of CCN2 in SMCs results in an unanticipated AAA phenotype in the infrarenal aortic region, a location dissimilar to Ang II–induced murine AAA yet identical to human AAA pathology. Subsequent in vivo and cell culture–based studies point to SMC phenotypic switching as a key driver predisposing the AAA phenotype in SMC-CCN2–deficient mice. Our studies provide evidence supporting the protective role of SMC-CCN2 against AAA. Results CCN2 protein is upregulated in both human and mouse AAA tissues. Little information is available regarding the impact of CCN2 modulation in vascular diseases. Studies performed to date are inconclusive, as the levels of CCN2 expression reported in human aneurysms varies and is, at times, contradictory (16, 21, 23). To clarify this, we compared the CCN2 protein levels between normal human aortic tissue and AAA samples. CCN2 protein was expressed minimally in normal aortas yet was remarkably induced in aortas with aneurysms (Figure 1A). Immunofluorescence staining of the aortic tissue revealed that increased CCN2 expression was mostly observed in the media layer and was colocalized with SMC marker αSMA (Figure 1B), suggesting that SMC is an important source of vascular CCN2 in response to vascular injury. A similar increase of CCN2 protein was observed in Ang II–induced AAA samples compared with aortas from saline-treated mice (Figure 1, C and D). It is unclear, however, whether a boost in CCN2 levels in the vasculature is a reparative mechanism to combat aneurysm progression, an agitator, or simply a consequence with little relevance of disease pathology. Since vascular SMCs (VSMCs) contribute to a large percentage of total CCN2 in the vascular bed and considering that SMC dysfunction has long been implicated in the development of AAA (24, 25), we explored the impact of SMC-specific CCN2 deficiency in the development of AAA in mice. Deficiency of CCN2 in SMC exacerbates Ang II–induced AAA in hypercholesterolemia mice. To evaluate the impact of SMC-CCN2 deficiency on murine AAA, we first generated CCN2fl/fl–Myh11–Cre-ERT2 mice by crossing CCN2-floxed mice with Myh11–Cre-ERT2. At 8 weeks of age, maleCCN2fl/fl–Myh11–Cre-ERT2 mice were subjected to 5 consecutive days of tamoxifen injection to induce CCN2 deletion in SMCs. After an additional 2 weeks, the resultant mice and CCN2-floxed control animals were rendered hypercholesterolemic (achieved via AAV-proprotein convertase subtilisin/kexin type 9 [PCSK9] injection to deplete LDL receptor [LDLR]) followed by Ang II infusion. Systemic Ang II infusion at the dose of 500 ng/kg/min successfully induced AAA in hypercholesterolemic mice in both groups. In the CCN2fl/fl group, 46% of the mice developed aneurysms in the suprarenal region of the abdominal aortas, while 91% of the SMC-specific CCN2-KO mice, henceforth CCN2SMCΔ mice, developed aneurysms exclusively in the infrarenal segments, the identical anatomic location where human AAA occurs (Figure 2, A and C). Gross and ultrasound data show that the aneurysms in CCN2SMCΔ groups exhibited greater uniformity and size (Figure 2, B and D) and were particularly extensive (from renal branch to femoral bifurcation) and tortuous, reminiscent of advanced human aortic aneurysms. No differences in systolic blood pressure (SBP) were observed between groups at various time points before and after Ang II infusion (Figure 2E). Histological analysis of the aneurysm sections of the CCN2SMCΔ mice revealed expected changes, including severe media deterioration, elastin degradation, and immune cell infiltration (Figure 2F and Supplemental Figure 1, A and B; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.162987DS1). Immunofluorescence studies confirmed massive aortic wall dilation and media thickening; moreover, cells in the expanded media layer and adventitia showed minimum expression of SMC contractile markers, αSMA and smooth muscle myosin heavy chain (SMMHC) (Figure 2G). Notably, in CCN2SMCΔ mice, diminished expression of SMC markers was also evident in regions of the aorta with no aneurysm or dilations (Supplemental Figure 1C), suggesting intrinsic molecular changes associated with SMC CCN2 deficiency prior to the development of aneurysm. Ang II alone induces infrarenal aortic aneurysm formation in CCN2SMCΔ mice. The profound AAA pathology manifestations in hypercholesterolemic CCN2SMCΔ mice prompted us to speculate on what might occur to these animals when infusing Ang II without being hypercholesterolemic. With this in mind, and to capture early cellular changes and explore possible molecular mechanisms fueling aneurysm development in CCN2SMCΔ mice, we infused Ang II (500 ng/kg/min) or saline (as vehicle control) to CCN2fl/fl and CCN2SMCΔ mice. At day 7, mice were euthanized for tissue harvesting. To our surprise, Ang II alone was sufficient to induce remarkable AAA development in 20 of 21 mice examined in the CCN2SMCΔ group (again in the infrarenal aortic region), with most CCN2SMCΔ mice exhibiting strong “ballooning” of abdominal and elongated aorta. In contrast, only a few (3 of 18 mice) developed mild aneurysms in the thoracic or suprarenal region of the aorta in the control group (Figure 3, A–C). Consistent with higher AAA incidence (Figure 3B), in CCN2SMCΔ mice, the maximum abdominal aortic diameters of aneurysms in the CCN2SMCΔ group were significantly larger than those in the control group (Figure 3C; 0.86 mm versus 2.72 mm, P < 0.0001). Similar to what is observed in the long-term Ang II model (Figure 1), massive wall expansion, media thickening, and elastin breakdown were evident but with no sign of lipid accumulation or atherosclerosis in the aneurysm sites from the CCN2SMCΔ group (Figure 3D and Supplemental Figure 2A). Cells in the thickened media layer exhibited increased proliferation, as indicated by Ki67+ staining (Figure 3E) and apoptosis (Figure 3F), suggesting accelerated cell turnover. Moreover, most cells in the remodeled area were CD68– and showed little to no expression of αSMA (Figure 3G), but they exhibited strong vimentin and collagen I staining, markers for mesenchymal-like cells (Figure 3, H and I), suggesting a distinctive cell population from conventional SMCs. Correspondingly, tissue MMP activity was also elevated in the aneurysm sites (Supplemental Figure 2B). However, this is likely not attributed to the infiltration of immune cells, since only scarce and sporadic CD68+ cells were observed mostly along the edge of the adventitia (Figure 3G). Expectedly, all of the cellular and molecular changes stated above were not seen in the saline-treated groups (Supplemental Figure 3). Considering the important role of SMC apoptosis in aneurysm pathology, we also studied the effects of CCN2 deficiency on apoptosis in cultured human primary aortic SMCs (HASMCs). In line with augmented aortic cell apoptosis in Ang II–treated CCN2SMCΔ mice (Figure 3F), deficiency of CCN2 resulted in a significant increase in Staurosporine-induced SMC apoptosis (Supplemental Figure 4A) and elevated caspase 8 and caspase 3 activities (Supplemental Figure 4B), supporting CCN2 as an important factor for SMC survival against harmful insults. Exacerbation of elastase-BAPN induced infrarenal aortic aneurysm in CCN2SMCΔ mice. To corroborate that the aneurysm phenotype was not limited to the Ang II model, we tested the consequence of SMC-specific CCN2 deficiency in the elastase–β-aminopropionitrile (BAPN) model (26). A brief 10-minute soaking of abdominal mouse aortas with topical elastase in conjunction with 14 days of BAPN supplementation in the drinking water successfully induced AAA formation in both CCN2fl/fl and CCN2SMCΔ mice. However, the sizes of the dilations were markedly bigger in CCN2SMCΔ mice compared with controls, as can be seen through gross images (Supplemental Figure 5A) and the ultrasonic measurements of maximal external diameters of the abdominal aortas (Supplemental Figure 5, B and C). In addition, greater aortic wall expansion was observed in KO mice, with the remodeling and thickening of the media layers more severe in KO mice but to a lesser extent when compared with those in the Ang II model (Supplemental Figure 5D). Notably, the losses of media elastin were comparable between the 2 groups, with a complete loss at the anterior and residual elastin at the posterior of the aortic wall, suggesting that the differences in aneurysms were not the consequence of the surgical variation (Supplemental Figure 5D). Since the Ang II model is arguably more clinically relevant, and considering the possible interplay between CCN2 and Ang II (27), we decided to use the Ang II model for subsequent studies. RNA-Seq analysis reveals SMC-derived CCN2 participates in vascular smooth muscle contraction and transmembrane receptor signaling. To unbiasedly evaluate the functional role of SMC-derived CCN2 in AAA, we isolated total RNA from infrarenal abdominal aortic tissue samples in CCN2fl/fl and CCN2SMCΔ mice infused with saline or Ang II for RNA-Seq. The data revealed drastically different gene expression profiles associated with CCN2 deficiency in SMC even under saline treatment. Much greater divergences in gene expression were discovered between groups under Ang II treatment (Figure 4A), in which 3,723 genes were upregulated and 2,413 genes were downregulated in aortic tissues of CCN2SMCΔ mice compared with those of control mice (Figure 4B). We then performed KEGG and GO pathway analysis on differentially expressed genes to explore biological processes important for AAA development in mice with SMC-specific deficiency. Data suggest that the top affected processes include ECM-receptor interactions, vascular smooth muscle contraction (Figure 4C), inflammation, and transmembrane receptor protein tyrosine kinase signaling (Figure 4D). To explore why deficiency of CCN2, an ECM signaling protein, could lead to enormous differences in gene expression, transcription factor binding motif (TFBM) enrichment analysis was performed. Expectedly, transcription factors with well-documented roles in SMC biology and aneurysm, including Krüppel-like factors (KLFs; KLF4, KLF5) (28–30) and Smads (Smad2, Smad3, Smad4) (31–33), were significantly affected in response to SMC-specific CCN2 deficiency (Figure 4E). These findings provided additional guidance for the subsequent exploration of the molecular traits of AAA formation in SMC-specific CCN2-deficient mice. SMC-specific deficiency of CCN2 represses SMC contractile markers both in vivo and in vitro. To confirm the involvement of CCN2 deficiency–associated loss of SMC contractile markers, quantitative PCR (qPCR) analysis was performed to compare the mRNA levels of SMC markers in abdominal aortas of CCN2fl/fl and CCN2SMCΔ mice with 7-day Ang II infusion. Deficiency of CCN2 in SMC resulted in a greater than 90% reduction of total CCN2 mRNA in mouse aortas. Correspondingly, Myh11 (SMMHC), Acta2 (αSMA), CNN1 (Calponin1), and Tagln (SM22), which are markers of SMC, as well as Eln (Elastin) and Fn1 (fibronectin 1), were significantly downregulated in CCN2SMCΔ mice. Not surprisingly, however, Lgals3 (a marker of cell activation) and Tnf (a marker of tissue inflammation) were upregulated in CCN2SMCΔ mice (Figure 5A). In a separate cohort of animals, abdominal aortas from CCN2fl/fl and CCN2SMCΔ mice infused with Ang II or saline were harvested for protein expression analysis. CCN2 protein is expressed at low levels at baseline in saline-treated WT mice, and its expression increased substantially under the effects of Ang II. Nevertheless, knockdown of CCN2 in SMC significantly reduced the protein content of CCN2 in mouse aortas in both groups but with greater changes in magnitude in the Ang II group (Figure 5, B and C). Along with the reduction in CCN2, SMC contractile markers SMMHC, αSMA, and SM22 were moderately reduced in saline-treated CCN2-KO mice; Ang II infusion notably suppressed SMC markers in WT mice; and the protein levels of SMC markers were further reduced in CCN2-KO mice. Interestingly, CCN2 deficiency in SMC was also associated with a reduction in phosphorylated serum response factor (p-SRF), a positive regulator of SMC gene transcription (34–36), in mouse aortas of both treatment groups. In contrast, KLF4, a known transcriptional repressor of SMC marker genes (37, 38), was markedly induced by Ang II, and its levels were further upregulated in conjunction with CCN2 deficiency (Figure 5, B and C). To further decipher the biological effects of CCN2 in SMC specifically and to test the validity of our findings in vivo, HASMCs were cultured and infected with adenoviruses to either knock down or overexpress CCN2. In line with the in vivo findings, knockdown of CCN2 significantly reduced the protein contents of p-SRF and SMC markers — including αSMA, calponin, and SM22 — but increased the KLF4 transcription factor (Figure 6A). In contrast, CCN2 overexpression substantially augmented the levels of p-SRF and SMC markers but reduced KLF4 (Figure 6B). Our data suggest that CCN2 is critical in maintaining an SMC contractile phenotype both in vivo and in vitro, and they show that deficiency of CCN2 leads to the loss of contractile markers in SMC and predisposes those mice to aneurysm development following vascular injury. CCN2 interacts with and potentiates TGF-β1 receptor signaling in HASMCs. As an ECM protein, CCN2 conveys its signaling potential through interaction with an array of receptors, ECM components, and growth factors. Specifically, studies have demonstrated that CCN2 is an essential cofactor for TGF-β and could potentiate TGF-β activity (5, 39). Considering the well-reported role of TGF-β signaling in SMC phenotype regulation in conjunction with RNA-Seq results that point to the involvement of the Smads and the receptor tyrosine kinase pathway associated with SMC CCN2 deficiency, we decided to take a deep look at the interplay between CCN2 and TGF-β. In cultured HASMCs, an intermediate dose of TGF-β1 at 1 ng/mL markedly stimulated CCN2 protein expression as early as 1 hour after treatment. Correspondingly, significant increases in αSMA and SM22 protein levels were also observed (Figure 7A). We then performed a co-IP assay in CCN2-overexpressing HASMCs to determine whether CCN2 could physically interact with TGF-β1 and/or the TGF-β receptor. After targeted pull-down with anti-CCN2 antibody, along with various positive and negative assay controls, the blots clearly showed the presence of both TGF-β1 and TGF-β receptor 2 in a CCN2 protein complex (Figure 7B). To test whether CCN2 could mediate TGF-β1 activity, HASMCs were incubated with recombinant human CCN2 (rhCCN2) with or without a low dose of TGF-β1 (0.1 ng/mL) for 30 minutes. The data demonstrate that rhCCN2 alone is capable of stimulating TGF-β2 receptor phosphorylation to some degree, yet it failed to evoke downstream Smad3 phosphorylation. However, in the presence of TGF-β1, a supplement of rhCCN2 significantly increased the phosphorylation of Smad3 (Figure 7C). Conversely, deficiency of CCN2 led to significant increases in KLF4 in both vehicle and TGF-β1–treated HASMCs. Moreover, TGF-β1–mediated Smad3 phosphorylation was markedly reduced in CCN2-deficient cells (Figure 8A). In line with the previous observation (Figure 7A), extended TGF-β1 stimulation (6 hours) significantly increased CCN2, αSMA, and SM22 protein levels; however, the stimulative effects of TGF-β1 on SMC marker expression were significantly blunted with CCN2 deficiency in HASMCs (Figure 8B). Together, these data suggest that CCN2 is both a target and mediator of TGF-β signaling and at least partially accounts for TGF-β1–mediated SMC contractile protein expression. TGF-β receptors are required for CCN2-mediated SMC contractile marker expression. To explore the involvement of TGF-β receptors in CCN2-mediated regulation of SMC marker expression, we knocked down TGFBR1 and TGFBR2, either individually or together in HASMCs, followed by adenovirus-mediated overexpression of CCN2. Knockdown of TGF-β receptors using siRNA was verified by qPCR, which indicated over 90% efficiency (Figure 9A). While we consistently observed the induction of SMC markers, both mRNAs (Figure 9B) and proteins (Figure 9C), in CCN2 overexpressing cells, CCN2-mediated increases in αSMA (ACTA2) and SM22 (TAGLN) were eliminated in cells deficient of TGFBR1, TGFBR2, or both (Figure 9, B and C). Taken together, these data indicate that CCN2-mediated SMC contractile marker expression is likely dependent on TGF-β receptor involvement. Discussion The results of this study are the first to our knowledge to demonstrate that the loss of SMC-specific CCN2 exacerbates aneurysm development in mice, both in the presence and absence of hypercholesterolemia. Aneurysms formed in CCN2-deficient mice were extensive and tortuous, with a focal infrarenal preference (Figure 2A). Tissue characterization of the aneurysm sites reveals severe media deterioration, elastin breakdown, immune cell infiltration, and lipid accumulation — features closely reminiscent of advanced human aortic aneurysm (Figure 2F and Supplemental Figure 1B). Intriguingly, resistance of C57BL/6 to Ang II–induced AAA is completely lost upon silencing of VSMC–specific CCN2. In a majority of murine AAA studies reported, hypercholesterolemia is required for successfully modeling AAA using Ang II (40), either by genetic ablation of apolipoprotein E (Apoe–/–) (41) or LDL receptor (Ldlr–/–) (42). Our acute studies revealed that Ang II alone (as short as 3 days) is sufficient to induce AAA at the infrarenal site in CCN2SMCΔ mice but not in the controls (Figure 3). These findings clearly underscore SMC-derived CCN2 as a key ECM component of the vascular wall essential for the maintenance of vascular integrity and the prevention of structural damages, especially in response to injury. Formation of AAA in the infrarenal aortic region of CCN2SMCΔ mice in response to Ang II infusion is an intriguing finding. To the best of our knowledge, since the introduction of the Ang II model for AAA studies in mice, studies over the last 2 decades have reported AAA pathology at the suprarenal aortic region, a location different than the infrarenal aorta where human AAA occurs (40, 43). While our understanding of the pathology of AAA formation has grown substantially due to experiments derived from murine models of AAA that recapitulate several major features of human AAA, there is still a lack of progress in developing effective medical therapy for AAA, indicating that our current understanding of the pathophysiology of human AAA remains inadequate. Thus, new models that closely mimic human AAA are strongly desired. Aside from the obvious limitations of the use of murine species, which the research community relies upon to glean insights into pathophysiological mechanisms that may or may not translate into human AAA disease, the disparity in the anatomic locations of murine and human AAA is another major limitation. While it is conceivable that humans being biped while mice are quadruped could render regional hemodynamic and structural differences between human and mice that might account for the AAA location difference, the exact underlying mechanisms are unknown. Nevertheless, our findings of profound infrarenal AAA pathology in both Ang II and elastase models not only unequivocally demonstrate an unanticipated antianeurysmal role of SMC-CCN2, but also support that CCN2 mutant mouse lines may serve as a novel model for studying infrarenal AAA disease. An important point to note is that our studies were carried out in the context of inducible deficiency of CCN2 in adult mice. Whether germline loss of SMC-CCN2 has any impact on vessel integrity, on vascular homeostasis, or in murine aortic aneurysm is a subject of future investigation. The proaneurysmal phenotype of CCN2SMCΔ mice is corroborated by a recent study using whole-body CCN2-deficient mice (inducible) in which augmented aneurysm development in both thoracic and abdominal aorta with Ang II infusion was observed (44). Despite all the structural and molecular changes associated with CCN2 deficiency, change in VSMC function was not a particular focus in that study. Interestingly, data from previous a study implied that hemizygous KO of CCN2 is protective from elastase-induced AAA (45), findings that appear in disagreement with our results. However, those experiments were performed in constitutive whole-body CCN2+/–, a condition that is very different from our current approach, where SMC-CCN2 ablation is induced in the adult mice. It is likely that, in comparison with inducible loss of SMC-CCN2 in adult mice, the constitutive and systemic deficiency of CCN2 could trigger distinct mechanisms leading to developmental changes that might render an overall antianeurysm effect in the elastase model. Also, in light of the importance of all vascular wall intrinsic cell types (EC, SMC, and fibroblasts) and hemopoietic lineage cells in AAA pathology, it is possible that haploinsufficiency of CCN2 in different cellular compartments might have differential responses in the elastase-induced AAA model, an issue that can only be addressed with additional cell lineage–restricted KO animals. Studies herein focused on SMC-CCN2, and this is in part attributed to the central importance of this cell type in the pathobiology of AAA. In support of this, our data show that SMC-CCN2 limits MMP activity, inflammation, and cellular apoptosis and preserves the SMC contractility, culminating in an overall protection against AAA (Figure 3, E–I; Figure 5; and Supplemental Figure 2B). However, we are not excluding the possibility that CCN2 produced by other cell types such as endothelial cells, adventitial fibroblasts, or immune cells may affect AAA — a possibility that necessitates follow-up studies in animals with restricted deficiency in the cell types mentioned above. Contradictory to its profibrotic role in fibroblasts, CCN2 deficiency in SMC results in reduced elastin mRNA but increased breakage in aneurysm tissue that could compromise the overall strength of the vascular wall (Figure 3D, Figure 5A, and Supplemental Figure 2A). Interestingly, when elastin was eliminated via elastase and BAPN, the AAA phenotype in CCN2SMCΔ mice was still significantly more severe than that in control mice, suggesting that CCN2 deficiency in SMC mediates deleterious effects beyond elastin dysregulation (Supplemental Figure 5). One important mechanism that likely contributes to the observed AAA phenotype in CCN2SMCΔ mice is the reduction in SMC contractile marker expression. In the chronic Ang II setting, signals of SMMHC and αSMA are significantly diminished in the media layer of aortas at the aneurysm site in CCN2SMCΔ mice compared with those in both the aneurysm and nonaneurysm sites in CCN2-floxed control mice, henceforth CCN2fl/fl mice (Figure 2G and Supplemental Figure 1C). Similar reductions in SMC markers were also evident in the acute setting (Figure 5, A and B). Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) analyses with Metascape (https://metascape.org/gp/index.html#/main/step1) of the RNA-Seq data reveal that vascular smooth muscle contraction is one of the most affected pathways in SMC-CCN2–deficient aortas (Figure 4C). Moreover, cells within the aberrantly remodeled vessel wall show minimal SMC markers, and they were CD68– (marker of macrophage lineages) and positive for vimentin and collagen I (markers for mesenchymal-like cells). This represents a unique cell population distinct from conventional SMCs (Figure 3, G–I). In addition, increases in cell proliferation and apoptosis were observed in that population, suggesting a synthetic phenotype with greater cell turnover (Figure 3, E and F). This unique cell population likely contributes to the increased MMP activity and elastin disruption in the aortas of the CCN2SMCΔ mice (Supplemental Figure 2). These findings, while not against the prevailing view that VSMCs undergo a phenotypic switch from contractile to synthetic states in vascular diseases, coincide with a trending view that an array of SMC with diverse fates are present in vascular disease, as revealed by accumulating lineage tracing and single-cell RNA-Seq studies (46–50). Additional lineage-tracing studies are required in order for researchers to completely understand SMC fate due to CCN2 loss in murine AAA models. As a matricellular protein, CCN2 binds to multiple receptors and growth factors and mediates crosstalk between the ECM and a variety of cells. In tissue fibrosis, it has been suggested that CCN2 partially mediates the profibrotic effects of both TGF-β and Ang II (3, 27, 51). Studies have shown that, in fibroblasts, TGF-β1–mediated induction of CCN2 mRNA occurs within 30 minutes of TGF-β treatment (22), and this effect is dependent on Smad3 signaling (52). This transcriptional regulation of CCN2 by TGF-β is further supported by studies showing the presence of a unique sequence located in the CCN2 promoter upon which Smad3 and Smad4 are able to bind (52). In addition, studies using biomolecular techniques demonstrated that CCN2 may function as a chaperone to facilitate TGF-β1/TGF-β receptor interaction and potentiate TGF-β1–mediated Smad activation (5). Here, our in vitro results echo previous findings and advance the knowledge regarding the interplay between CCN2 and TGF-β signaling in SMCs by providing evidence that (a) CCN2 is a target of TGF-β signaling and that TGF-β1–mediated induction of CCN2 protein is fast and robust (Figure 7A); (b) CCN2 can physically interact with the TGF-β1–TGF-β receptor complex and potentiate TGF-β1–mediated Smad3 phosphorylation (Figure 7, B and C); and (c) CCN2 and TGF-β signaling are mutually dependent on maintaining the contractile function of SMC, as deficiency of either CCN2 or TGF-β receptors suppress SMC markers (Figures 8 and 9). Concordant with the intricate functional relationship between CCN2 and TGF-β, mice with SMC-specific ablation of TGF-β signaling on a proatherosclerotic background (either Apoe–/– or Ldlr–/–) developed aortic aneurysms in response to a hypercholesterolemic diet (21). Moreover, the extensive and tortuous presentation of the aneurysms is very similar to those observed in our chronic Ang II model in CCN2SMCΔ mice. Mechanistically, the data suggest that a combination of medial SMC loss with increases in transdifferentiation of contractile SMCs to MSC-like cells contribute to aneurysm development in those mice. In addition, an elevated expression of KLF4, a direct target suppressed by TGF-β, was observed in the mouse aortas with either CCN2 or TGF-βR2 deficiency in SMCs (21). Therefore, it is tempting to speculate that augmented aneurysm development in SMC-specific TGF-βR2 deficient mice might partially attribute to the disruption of signaling normally mediated by SMC-derived CCN2. To gain in-depth insight into the functional importance of CCN2 for TGF-β pathway or vice versa in AAA, it would be interesting to know whether CCN2 expression/function is altered in SMC–TGF-βR2–KO mice; additional in vivo studies to determine whether CCN2 sufficiency or deficiency affects the AAA phenotype in animals deficient in SMC–TGF-βR2 are required. Together, our studies point to the potential importance of the cooperation of CCN2 and the TGF-β pathway in vascular homeostasis, an exciting topic that merits further interrogation. In summary, our findings demonstrate that deficiency of SMC-specific CCN2 alters SMC phenotype and function to result in compromised vascular integrity and aneurysm development. Owing to the uniqueness of intrarenal AAA pathology in CCN2SMCΔ mice, our studies suggest that CCN2SMCΔ mice might serve a tool to gain knowledge in disease mechanisms for human AAA. Finally, it is exciting to envision the perceived requirement of CCN2 for TGF-β action and vice versa for normal vascular homeostasis in general and in AAA disease, in particular. Methods Animals CCN2-floxed mice (originally acquired from Andrew Leask, The University of Western Ontario, London, Ontario, Canada) were crossed with Myh11–Cre-ERT2 transgenic mice (The Jackson Laboratory, strain no. 019079) to generate male CCN2-floxed mice harboring SMC-specific Cre (CCN2fl/fl–Myh11–Cre-ERT2). SMC-specific deletion of CCN2 was achieved by i.p. injection of 5 consecutive doses of tamoxifen (0.1 mL at 20 mg/mL dissolved in sunflower seed oil; T5648, Sigma-Aldrich) to 8-week-old maleCCN2fl/fl–Myh11–Cre-ERT2 mice. Age-matched male CCN2-floxed mice without Cre expression subjected to the same tamoxifen regimen were used as controls. After a 2-week washout period, CCN2SMCΔ mice and CCN2fl/fl mice were used for the modeling of AAA. Ang II infusion model. For the long-term Ang II model, tamoxifen-injected CCN2fl/fl (n = 13) and CCN2SMCΔ mice (n = 11) were subjected to a single injection of adeno-associated virus overexpressing a mouse form of PCSK9 (1 × 1011 VG, AAV8-D337y-mPCSK9, Vector Biolabs) via tail vein and were fed a high-fat diet (D12108C, Research Diets) for the duration of the studies. Two weeks after PCSK9 injection, Ang II was systemically infused by ALZET miniosmotic pumps (Model 2006, ALZET) as previously described (53) at the dose of 500 ng/kg/min (A9525, Sigma-Aldrich) for 42 days. For short-term Ang II infusion, tamoxifen-injected CCN2fl/fl (n = 18) and CCN2SMCΔ mice (n = 21) only received Ang II infusion via miniosmotic pumps (Model 1007D, ALZET) at the dose of 500 ng/kg/min for 7 days. Additional CCN2fl/fl (n = 10) and CCN2SMCΔ mice (n = 10) receiving saline-filled osmotic pumps were used as saline controls. BAPN-elastase model. The 0.2% BAPN (A3134, Sigma-Aldrich) containing drinking water (in light-shield red bottles) was provided to mice 2 days before surgery until the end of the studies. BAPN containing drinking water was freshly prepared and changed every other day. Tamoxifen-injected CCN2fl/fl (n = 10) and CCN2SMCΔ mice (n = 10) were subjected to the surgery for elastase exposure as described elsewhere (26). Briefly, surgeries were performed on mice under anesthesia to expose the lower infrarenal abdominal aorta. In total, 5 μL of porcine pancreas elastase (Sigma-Aldrich) was carefully applied to the exposed aortic adventitia at the bottom of the boat-shaped area and incubated for 5 minutes. The exposed area was gently washed afterward with normal saline, and surgical sites were carefully sorted and restored before closing. As for the control groups, additional CCN2fl/fl (n = 5) and CCN2SMCΔ mice (n = 5) were subjected to the same surgical procedure, except for the substitution of the active elastase with the heat-inactivated elastase (100°C for 30 minutes). In vitro studies HASMCs were purchased from American Type Culture Collection (ATCC, PCS-100-012) and cultured in vascular cell basal medium (ATCC, PCS-100-030) supplemented with growth kit (ATCC, PCS-100-042). Cells at passage 5–7 were used for various studies. Adenoviruses used for CCN2 overexpression (Ad-GFP-hCTGF, ADV-206232) and knockdown (Ad-GFP-U6-h-CTGF-shRNA, shADV-206232) studies, henceforth AdCCN2 and shCCN2, and control viruses (Ad-GFP, Ad-GFP-U6-scrmb-shRNA), henceforth AdCtrl and shCtrl, were purchased from Vector Biolabs. siRNAs used for human TGF-β receptor knockdown (TGFBR1-7046, TGFFBR2-7048, and Nontargeting Pool) were purchased from Horizon. rhCCN2 (9190-cc) and TGF-β1 (rhTGF-β1, 240-B/CF) were purchased form R&D Systems and reconstituted in sterile PBS and 4 mM HCl-0.1% BSA, respectively, as recommended by the manufacturer. For CCN2 knockdown– or overexpression-related studies, HASMCs at 80% confluence were infected with shCtrl, shCCN2, AdCtrl, or AdCCN2 at the multiplicity of infection (MOI) of 10. After overnight incubation, virus containing medium was replaced with fresh medium and sat for 48 hours, followed by 6-hour starvation (0.5% FBS) before harvesting or subsequent treatment. In one study, naive cells at confluence were fasted and treated with rhTGF-β1 (1 ng/mL) for 1, 3 and 6 hours to assess the expression of CCN2 and SMC markers. In another study, lower doses of rhTGF-β1 (0.1 ng/mL) and rhCCN2 (10 ng/mL) were used in combination to assess TGF-β receptor and Smad3 phosphorylation. In the context of CCN2 deficiency, short-term rhTGF-β1 exposure (1 ng/mL, 0.5 hours) and prolonged exposure (1 ng/mL, 6 hours) were performed to study the quick (phosphorylation) and slow (protein synthesis) responses respectively. For Co-IP, HASMCs were infected with AdCCN2 and cultured for 66 hours, followed by 6-hour starvation (0.5% FBS). Cells were then incubated with rhTGF-β1 (1 ng/mL) and rhCCN2 (100 ng/mL) for 30 minutes. After that, culture medium was completely removed, and cells were rinsed with 5 mL HBSS (Corning) and fixed with 1% formaldehyde on ice for 10 minutes. After another rinse with HBSS, 1 mL of Co-IP buffer with Halt protease and phosphatase inhibitor cocktail (Thermo Fisher Scientific) was added to each 100 mm dish, incubated on ice for 10 minutes, and cells were scraped and collected to 1.5 mL microcentrifuge tubes. Cell lysates were centrifuged at 10,000g (15 minutes, 4°C), and 500 μL clear supernatant was used for each i.p. injection using preprepared antibody-bound protein G beads (Santa Cruz Biotechnology Inc., sc-2002). Normal rabbit IgG (Sigma-Aldrich, I5006) in substitution of anti-CCN2 antibody was included as a negative assay control, while additional precipitations using anti–TGF-β1 and anti-TGFBR2 antibodies were used as positive controls. Detailed antibody information is provided in Supplemental Table 2. For the siRNA-mediated TGF-β receptor knockdown study, HASMCs were cultured in a 6-well culture plate until subconfluent (70%–80% confluency) and transfected with target siRNAs (siTGFBR1, siTGFBR2, siTGFBR1 + siTGFBR2) or control siRNA with lipofectamine RNAiMAX reagent (Invitrogen, 56532) at the dose of 30 nM. Six hours After siRNA transfection, medium was changed, and cells were infected overnight with either AdCtrl or AdCCN2 and cultured for an additional 48 hours before harvesting. For cell apoptosis studies, control (shCtrl) or CCN2-deficient (shCCN2) HASMCs (shRNA-mediated gene knockdown) were treated with vehicle (DMSO) or 1 μM Staurosporine (AdipoGen, AG-CN2-0022-M001) for 6 hours and subjected to the Western blot analysis of Caspase 8 (cleaved) and Caspase 3 (cleaved). In vitro TUNEL assay was performed on control and CCN2-deficient HASMCs cultured on cover glasses and treated with vehicle (DMSO) or 1 μM Staurosporine for 3 hours using an In Situ Cell Death Detection Kit, TMR red (Roche) per manufacturer instructions. Aortic ultrasound imaging Abdominal aortas were visualized in isoflurane-anesthetized mice at baseline and various time point after treatment and at study endpoints using the portable Vevo 3100 Micro-Imaging System with a D550 probe (FUJIFILM). Longitudinal and cross-sectional images of the abdominal aortas were captured and normal and maximal diameters of the abdominal aortas at aneurysm sites were measured. All recordings were made by 2 researchers in a blinded manner. Histochemistry and immunofluorescence staining At study endpoints, mice were euthanized via isoflurane overdose and aorta segments were carefully isolated and harvested. For histochemistry, aorta segments were rinsed in ice-cold PBS, fixed in 10% buffered formalin for less than 24 hours, and transferred to 70% ethanol for paraffin embedding. Serial sections (5 μm sections) from each group were deparaffinized and used for morphological study by H&E staining and for visualization of elastin fibers assessed by Verhoeff-van Gieson staining using the Elastic Stain Kit (Abcam, ab150667). Images were acquired using NanoZoomer (Hamamatsu) at 20× resolution and processed by NDP.veiw 2. For immunofluorescence staining, aorta segments were rinsed in ice-cold PBS, directly embedded in OCT compound (Sakura, 4583), and slowly frozen down to solid. Cryosections (5 μm sections) from each group were air-dried, washed, and briefly fixed with 10% PBS-buffered formalin for 10 minutes. Aortic sections were then permeabilized with 0.1% Triton X-100 for 10 minutes and blocked with 10% BSA for 30 minutes at room temperature (RT). Primary antibodies (αSMA, SM-MHC, Ki67, and CD68; see Supplemental Table 2 for specific details) at 2 μg/mL were prepared in 1% BSA, added to each section, and incubated overnight at 4°C. Sections were then incubated with host-specific secondary antibodies with Alexa Fluor 488 or 594 (2 μg/mL) for 2 hours at RT protected from light. DAPI at 1 μg/mL was used to counterstain the nuclei. Fluorescence signals were visualized, and images were captured using a fluorescence microscope (Olympus IX71). In situ cell death (TUNEL) Cell apoptosis was detected on freshly cut OCT-embedded abdominal aorta sections (5 μm) from CCN2fl/fl and CCN2SMCΔ mice (n = 5 per group) infused with saline or Ang II for 7 days using a In Situ Cell Death Detection Kit, Fluorescein (Roche), per manufacturer’s instructions. Briefly, DNA breaks were labeled at the free 3’-OH termini, with modified nucleotides (fluorescein-dUTP), during enzymatic reaction in cells, underwent apoptosis; green fluorescence signals were visualized; and images were captured by fluorescence microscopy. Negative controls were performed on parallel sections in the absence of enzyme solution. In situ MMP activity In situ MMP activity was performed on freshly cut OCT-embedded abdominal aorta sections (5 μm) from CCN2fl/fl and CCN2SMCΔ mice (n = 5 per group) infused with saline or Ang II for 7 days using a Gelatinase/Collagenase Assay Kit (Invitrogen, E12055) per manufacturer instructions. Briefly, fluorescein-conjugated gelatin substrate was prepared and applied to sections and allowed to incubate at RT for 36 hours. Green fluorescence develops in the presence of enzymatic activities and was examined using a fluorescence microscope. Negative controls were performed on parallel sections in the presence of 5 mM EDTA. RNA extraction and qPCR Total RAN was isolated using the RNase Mini Kit (Qiagen). RNA concentrations were determined using a Nanodrop Onec spectrophotometer (Thermo Fisher Scientific). In total, 100 ng of total RNA was used for reverse transcription to make cDNA using iScript Reverse Transcription kit (Bio-Rad, 1708841). Gene expression was assessed by SYBR green on QuantStudio 7 (Applied Biosystems). Data were normalized to GAPDH or β-actin using the ΔΔCt method. As a standard practice for qPCR experiments, no template and no RT controls were always included for quality control purposes. Detailed primer information is provided in Supplemental Table 1. Tissue lysis and Western blot Snap-frozen aortic tissues were smashed in 1.5 mL microcentrifuge tubes using a pestle in liquid nitrogen. Tissues were homogenized in 100–200 μL ice-cold T-PER Tissue Protein Extraction Reagent (Thermo Fisher Scientific) supplemented with Halt protease and phosphatase inhibitor cocktail (Thermo Fisher Scientific) using a pellet pestle motor (Kimble-Kontes) and centrifuged at 10,000g (15 minutes, 4°C). For protein extraction of cultured cells, 200 μL per well (6-well plate) of ice-cold protein extraction reagent was directly added to cell culture dish after PBS wash and left on ice for 10 minutes. Cells with extraction buffer were scraped and collected into 1.5 mL Eppendorf tube and centrifuged at 10,000g (15 minutes, 4°C). In both cases, supernatants were collected, and protein concentrations were determined by BAC assay (Pierce). Equal amounts of resultant protein were run on 4%–20% SDS-PAGE gel under denaturing conditions, transferred to nitrocellulose membrane, and subjected to Western blot analyses using specific antibodies. Depending on the compatibility and availabilities of the antibodies, membranes were reprobed with antibodies against either β-actin, GAPDH, or Vinculin to confirm equal loading. Detailed antibody information is provided in Supplemental Table 2. Target proteins were detected by Immobilon Western HRP substrate (MilliporeSigma) and exposure (Kodak). Bands’ intensity was analyzed by ImageJ (NIH). RNA-Seq studies Total RNAs extracted from abdominal aortas in mice infused with saline or Ang II for 7 days (4 groups, n = 3 per group) passed RNA sample quality control were subjected to commercial RNA-Seq (Novogene Corporation Inc.) using the NovaSeq PE150 platform at a depth of 6 G raw data per sample. Differentially expressed genes were identified by DEG-Seq based on P < 0.05 and |log2 fold change| > 1.5. Heatmap package (version 1.0.8, https://cran.r-project.org/web/packages/pheatmap/index.html) in R was utilized to perform the hierarchical cluster analysis. A volcano plot was generated to visualize up- and downregulated genes in CSA versus CFA aortic tissues. KEGG and GO enrichment were performed using Metascape (http://metascape.org/gp/index.html#/main/step1) to show the biological pathways and processes affected in the abdominal aorta associated with SMC-specific CCN2 deletion with Ang II infusion. RNA-Seq data have been deposited in the NCBI’s Gene Expression Omnibus (accession number GSE221399). Statistics Data are presented as mean ± SEM. For data form 2 groups, an unpaired Student’s 2-tailed t test was performed. One-way ANOVA (followed by Tukey post hoc test) was performed to compare a single variable in multiple groups, and 2-way ANOVA (followed by Dunnett post hoc test) was used for 2-factor analysis. χ2 test was preformed to assess statistical significance for aneurysm incidence between 2 groups. Statistical analyses were performed using Prism 9.0 software. P < 0.05 was considered statistically significant. Study approval All studies using mice were approved by an Institutional Animal Review Committee at Emory University and were conducted in accordance with the Guide for the Care and Use of Laboratory Animals (National Academies Press, 2011). Human aortic tissue samples were gifted by Guo-Ping Shi (Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts, USA). Frozen human AAA lesions and adjacent normal abdominal aortas were obtained from the Brigham and Women’s Hospital as previously reported (54). Discarded and decoded human aortas were reused according to the protocol 2010P001930 preapproved by the Human Investigation Review Committee at the Brigham and Women’s Hospital. Author contributions The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. YW, XL, and ZL designed the experiments. YW, XL, and WX performed the experiments. YW, XL, and QX analyzed the data. WX and XZ assisted with mouse breeding and tail vein injections. YW and ZL wrote the manuscript. ZL supervised the project. AL provided floxed-Ccn2 mouse strain. First authorship order position is based on intellectual contribution to design of the study and interpretation of data. Supplementary Material Supplemental data This work was supported by NIH grants 5R01HL144741 and 5R01HL152074, American Heart Association Transformative Project Award 20TPA35490431, and Emory University Startup Funds (all to ZL). Version 1 01/10/2023 Electronic publication Version 2 07/13/2023 Addition of author Andrew Leask Figure 1 CCN2 protein is upregulated in both human and mouse abdominal aortic aneurysm (AAA) tissues. (A) Western blot analysis of CCN2 protein in healthy human aortas versus AAA samples (n = 3). (B) Representative immunofluorescence staining images of CCN2 in human aortic sections with or without AAA (n = 3). (C) Western blot analysis of mouse CCN2 protein in abdominal aortic tissue lysates from normal (saline-treated) versus AAA (Ang II–treated) samples (n = 4). (D) Representative immunofluorescence staining images of CCN2 in mouse aortic sections with or without AAA (n = 3). Scale bars: 100 μm. Data were quantified and represented as mean ± SEM. ***P < 0.001, ****P < 0.0001, 2-tailed Student’s t test. Figure 2 SMC-specific deficiency of CCN2 exacerbates Ang II–induced AAA in hypercholesterolemia mice. (A) Representative gross images of aortas from CCN2fl/fl (WT, n = 13) and CCN2SMCΔ (KO, n = 11) mice infused with Ang II. (B) Representative abdominal ultrasound images (long- and short-axis views) of mice in both groups. (C) AAA incidence in mice from both groups. *P < 0.05, χ2 test. (D) Maximal external aortic diameters at study endpoint in mice from both groups. Data were quantified and represented as mean ± SEM. ***P < 0.001, 2-tailed Student’s t test. (E) Systolic blood pressure (SBP) measured at various time points before and after Ang II infusion for mice in both groups (n = 8). (F) Representative images of H&E-stained (5× and 20×) and Elastin-stained (20×) abdominal aortic sections from both groups (n = 5). Red arrows indicate elastin break points. (G) Immunofluorescence double staining of abdominal aortic sections from both groups with specific antibodies against αSMA (red) and SMMHC (green). Nuclei were stained with DAPI (blue). n = 4. Scale bars: 100 μm. Figure 3 SMC-specific deficiency of CCN2 results in the development of infrarenal AAA in mice with 7-day Ang II infusion. (A) Representative gross images of abdominal aortas from CCN2fl/fl (n = 18) and CCN2SMCΔ (n = 21) mice with Ang II infusion. (B) AAA incidence in mice from both groups. ***P < 0.001, χ2 test. (C) Maximal external aortic diameters at study endpoint (n = 9). Data were quantified and represented as mean ± SEM. ****P < 0.0001, 2-tailed Student’s t test. (D) Representative images of H&E-stained (5× and 20×) and Elastin-stained (20×) infrarenal aortic sections from both groups. Red arrows indicate elastin break points. n = 5. (E and G–I) Immunofluorescence double staining of αSMA (red) and Ki67 (green) (E), αSMA (red) and CD68 (green) (G), αSMA (red) and Vimentin (green) (H), and αSMA (red) and collagen I (green) (I) in infrarenal aortic sections from Ang II–infused CCN2fl/fl and CCN2SMCΔ mice. n = 4. Nuclei were stained with DAPI (blue). (F) In situ cell death assay of infrarenal aortic sections from CCN2fl/fl and CCN2SMCΔ mice; positive green signals indicate apoptotic cells. n = 4. Scale bars: 100 μm. Figure 4 RNA-Seq analysis reveals SMC-derived CCN2 participates in vascular smooth muscle contraction and transmembrane receptor signaling. (A) Heatmap of differentially expressed genes (DEGs) in the CFS (CCN2fl/fl + Saline), CSS (CCN2SMCΔ + Saline), CFA (CCN2fl/fl + Ang II), and CSA (CCN2SMCΔ + Ang II) aortic tissues (n = 3), as obtained by RNA-Seq (FDR < 0.05). (B) A volcano plot showing DEG expression between CFA and CSA aortic tissues. Genes in colored dots have an adjusted P < 0.05 and |log2 fold change| > 1.5. (C) Top enriched KEGG pathways analysis of RNA-Seq data obtained as in B. (D) Top enriched GO pathways analysis of RNA-Seq data obtained as in B. (E) Transcription factor binding motif (TFBM) enrichment analysis by pScan of RNA-Seq data obtained as in B. Colored dots indicate significant hits. P < 0.05 in a z test. Figure 5 SMC-specific deficiency of CCN2 represses SMC contractile markers in vivo. (A) qPCR analysis of various mRNA expression in abdominal aortic lysates from CCN2fl/fl and CCN2SMCΔ mice with 7-day Ang II infusion (n = 5). *P < 0.05, ***P < 0.001, ****P < 0.0001, 2-tailed Student’s t test. (B) Western blot analysis of various protein expression in abdominal aortic lysates from mice in each group (n = 4). (C) Quantification of protein expression in B. Target protein levels were normalized to Vinculin. Data were quantified and represented as mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 (effect of CCN2 deficiency within same treatment group); #P < 0.05, ##P < 0.01 (effect of Ang II within genotype group); 2-way ANOVA. Figure 6 CCN2 regulates SMC contractile markers in vitro. Western blot analysis of various protein expression in cultured human primary aortic SMCs (HASMCs) with CCN2 deficiency (Ad-CCN2-shRNA versus Ad-scrmb-shRNA) (A) or overexpression (Ad-GFP-CCN2 versus Ad-GFP) (B). Target protein levels were normalized to GAPDH. Data were quantified and represented as mean ± SEM. n = 3. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, 2-tailed Student’s t test. Figure 7 CCN2 interacts with and potentiates TGF-β1 receptor signaling. (A) TGF-β1 (1 ng/mL) induces protein expression of CCN2, αSMA, and SM22 in HASMCs by Western blot (n = 3). Target protein levels were normalized to Vinculin. *P < 0.05, **P < 0.01, compared with time point 0, using 1-way ANOVA. (B) Validation of physical interaction between CCN2 and TGF-β1/TGF-β receptor II via Co-IP followed by Western blot. Column 3 indicate positive bands for TGFBR2 (75 kD) and TGF-β1 (25 kD) in sediment precipitated with anti-CCN2 antibody; column 1, 20% input; column 2, IgG (negative control); column 4, TGFBR2 (positive control); column 5, TGF-β1 (positive control). (C) Supplementation of recombinant human CCN2 (rhCCN2) increases TGF-β receptor II phosphorylation and potentiates TGF-β1–induced Smad-3 phosphorylation in HASMCs by Western blot (n = 3). Target protein levels were normalized to β-actin. **P < 0.01, ***P < 0.001; ****P < 0.0001 (compared with nontreatment [N] group); ##P < 0.01 (compared with TGF-β1-only [T] group); 1-way ANOVA. Figure 8 Deficiency of CCN2 attenuates TGF-β1–mediated signaling and SMC markers in vitro. (A) CCN2 deficiency attenuates TGF-β1–induced (0.5 hours) Smad3 phosphorylation and augments KLF4 expression in HASMCs. (B) CCN2 deficiency attenuates the expression of SMC contractile markers with and without TGF-β1 stimulation (6 hours) in HASMCs. Target protein levels were normalized to β-actin. Data were quantified and represented as mean ± SEM. n = 3. ***P < 0.001, ****P < 0.0001 (effect of CCN2 deficiency within same treatment group); ##P <0.01, ###P < 0.001, ####P < 0.0001 (effect of TGF-β1 within same genotype group); 2-way ANOVA. Figure 9 TGF-β receptors are required for CCN2-mediated SMC contractile markers expression. (A) Validation of siRNA-mediated knockdown of TGF-β receptors (TGFBR1, TGFBR2) by qPCR. n = 4. ****P < 0.001, 2-tailed Student’s t test. (B) Knockdown of TGF-β receptors blunted the stimulating effect of CCN2 overexpression on ACTA2 and TAGLN mRNA by qPCR. n = 4. *P < 0.05, **P < 0.01 (effect of CCN2 overexpression within same siRNA group); ##P < 0.01 (effect of siRNA treatment within same Ad group); 2-way ANOVA. (C) Knockdown of TGF-β receptors blunted the stimulating effect of CCN2 overexpression on αSMA and SM22 protein by Western blot. n = 3. Target protein levels were normalized to Vinculin. *P < 0.05, ***P < 0.001, 1-way ANOVA. Conflict of interest: The authors have declared that no conflict of interest exists. Copyright: © 2023, Wang et al. This is an open access article published under the terms of the Creative Commons Attribution 4.0 International License. Reference information: JCI Insight. 2023;8(1):e162987.https://doi.org/10.1172/jci.insight.162987. ==== Refs 1 Isselbacher EM Thoracic and abdominal aortic aneurysms Circulation 2005 111 6 816 828 10.1161/CIRCRESAHA.118.312436 15710776 2 Pinard A et al Genetics of thoracic and abdominal aortic diseases Circ Res 2019 124 4 588 606 10.1161/CIRCRESAHA.118.312436 30763214 3 Leask A Abraham DJ All in the CCN family: essential matricellular signaling modulators emerge from the bunker J Cell Sci 2006 119 pt 23 4803 4810 10.1242/jcs.03270 17130294 4 Chen Y et al CCN2 (connective tissue growth factor) promotes fibroblast adhesion to fibronectin Mol Biol Cell 2004 15 12 5635 5646 10.1091/mbc.e04-06-0490 15371538 5 Abreu JG et al Connective-tissue growth factor (CTGF) modulates cell signalling by BMP and TGF-beta Nat Cell Biol 2002 4 8 599 604 10.1038/ncb826 12134160 6 Inoki I et al Connective tissue growth factor binds vascular endothelial growth factor (VEGF) and inhibits VEGF-induced angiogenesis FASEB J 2002 16 2 219 221 10.1096/fj.01-0332fje 11744618 7 Aguiar DP et al New strategy to control cell migration and metastasis regulated by CCN2/CTGF Cancer Cell Int 2014 14 61 10.1186/1475-2867-14-61 25120383 8 Chen Z et al Connective tissue growth factor: from molecular understandings to drug discovery Front Cell Dev Biol 2020 8 593269 10.3389/fcell.2020.593269 33195264 9 Gonzalez D Brandan E CTGF/CCN2 from skeletal muscle to nervous system: impact on neurodegenerative diseases Mol Neurobiol 2019 56 8 5911 5916 10.1007/s12035-019-1490-9 30689195 10 Klaassen I et al The role of CTGF in diabetic retinopathy Exp Eye Res 2015 133 37 48 10.1016/j.exer.2014.10.016 25819453 11 Tsai CC et al Essential role of connective tissue growth factor (CTGF) in transforming growth factor-β1 (TGF-β1)-induced myofibroblast transdifferentiation from Graves’ orbital fibroblasts Sci Rep 2018 8 1 7276 10.1038/s41598-018-25370-3 29739987 12 Bassyouni IH et al Clinical significance of CCN2/connective tissue growth factor in Behçet’s disease patients Int J Rheum Dis 2019 22 8 1459 1465 10.1111/1756-185X.13597 31124308 13 Oemar BS et al Human connective tissue growth factor is expressed in advanced atherosclerotic lesions Circulation 1997 95 4 831 839 10.1161/01.CIR.95.4.831 9054739 14 Cicha I et al Connective tissue growth factor is overexpressed in complicated atherosclerotic plaques and induces mononuclear cell chemotaxis in vitro Arterioscler Thromb Vasc Biol 2005 25 5 1008 1013 10.1161/01.ATV.0000162173.27682.7b 15761189 15 Sachdeva J et al Smooth muscle cell-specific Notch1 haploinsufficiency restricts the progression of abdominal aortic aneurysm by modulating CTGF expression PLoS One 2017 12 5 e0178538 10.1371/journal.pone.0178538 28562688 16 Wang X et al Increased collagen deposition and elevated expression of connective tissue growth factor in human thoracic aortic dissection Circulation 2006 114 1 suppl I200 I205 10.1161/CIRCULATIONAHA.105.000240 16820572 17 Oemar BS Luscher TF Connective tissue growth factor. Friend or foe? Arterioscler Thromb Vasc Biol 1997 17 8 1483 1489 10.1161/01.ATV.17.8.1483 9301624 18 Duncan MR et al Connective tissue growth factor mediates transforming growth factor beta-induced collagen synthesis: down-regulation by cAMP FASEB J 1999 13 13 1774 1786 10.1096/fasebj.13.13.1774 10506580 19 Lau LF Lam SC The CCN family of angiogenic regulators: the integrin connection Exp Cell Res 1999 248 1 44 57 10.1006/excr.1999.4456 10094812 20 Lindsay ME Dietz HC The genetic basis of aortic aneurysm Cold Spring Harb Perspect Med 2014 4 9 a015909 10.1101/cshperspect.a015909 25183854 21 Chen PY et al Smooth muscle cell reprogramming in aortic aneurysms Cell Stem Cell 2020 26 4 542 557 10.1016/j.stem.2020.02.013 32243809 22 Grotendorst GR et al A novel transforming growth factor beta response element controls the expression of the connective tissue growth factor gene Cell Growth Differ 1996 7 4 469 480 9052988 23 Kanazawa S et al The expression of platelet-derived growth factor and connective tissue growth factor in different types of abdominal aortic aneurysms J Cardiovasc Surg (Torino) 2005 46 3 271 278 15956925 24 Petsophonsakul P et al Role of vascular smooth muscle cell phenotypic switching and calcification in aortic aneurysm formation Arterioscler Thromb Vasc Biol 2019 39 7 1351 1368 10.1161/ATVBAHA.119.312787 31144989 25 Lu H et al Vascular smooth muscle cells in aortic aneurysm: from genetics to mechanisms J Am Heart Assoc 2021 10 24 e023601 10.1161/JAHA.121.023601 34796717 26 Bhamidipati CM et al Development of a novel murine model of aortic aneurysms using peri-adventitial elastase Surgery 2012 152 2 238 246 10.1016/j.surg.2012.02.010 22828146 27 Ruperez M et al Connective tissue growth factor is a mediator of angiotensin II-induced fibrosis Circulation 2003 108 12 1499 1505 10.1161/01.CIR.0000089129.51288.BA 12952842 28 Salmon M et al KLF4 regulates abdominal aortic aneurysm morphology and deletion attenuates aneurysm formation Circulation 2013 128 11 suppl 1 S163 S174 10.1161/CIRCULATIONAHA.112.000238 24030402 29 Salmon M et al Klf4, Klf2, and Zfp148 activate autophagy-related genes in smooth muscle cells during aortic aneurysm formation Physiol Rep 2019 7 8 e14058 10.14814/phy2.14058 31025534 30 Ma S et al Leveraging cell-type-specific regulatory networks to interpret genetic variants in abdominal aortic aneurysm Proc Natl Acad Sci U S A 2022 119 1 e2115601119 10.1073/pnas.2115601119 34930827 31 Leung A Natarajan R Forgetting to switch off SMAD2 in aneurysmal disease Circ Res 2013 113 7 843 845 10.1161/CIRCRESAHA.113.302138 24030017 32 Dai X et al SMAD3 deficiency promotes vessel wall remodeling, collagen fiber reorganization and leukocyte infiltration in an inflammatory abdominal aortic aneurysm mouse model Sci Rep 2015 5 10180 10.1038/srep10180 25985281 33 Zhang P et al Smad4 deficiency in smooth muscle cells initiates the formation of aortic aneurysm Circ Res 2016 118 3 388 399 10.1161/CIRCRESAHA.115.308040 26699655 34 Garat C et al Induction of smooth muscle alpha-actin in vascular smooth muscle cells by arginine vasopressin is mediated by c-Jun amino-terminal kinases and p38 mitogen-activated protein kinase J Biol Chem 2000 275 29 22537 22543 10.1074/jbc.M003000200 10807920 35 Hirschi KK et al Transforming growth factor-beta induction of smooth muscle cell phenotpye requires transcriptional and post-transcriptional control of serum response factor J Biol Chem 2002 277 8 6287 6295 10.1074/jbc.M106649200 11741973 36 Owens GK et al Molecular regulation of vascular smooth muscle cell differentiation in development and disease Physiol Rev 2004 84 3 767 801 10.1152/physrev.00041.2003 15269336 37 Liu Y et al Kruppel-like factor 4 abrogates myocardin-induced activation of smooth muscle gene expression J Biol Chem 2005 280 10 9719 9727 10.1074/jbc.M412862200 15623517 38 Shankman LS et al KLF4-dependent phenotypic modulation of smooth muscle cells has a key role in atherosclerotic plaque pathogenesis Nat Med 2015 21 6 628 637 10.1038/nm.3866 25985364 39 Shi-wen X et al CCN2 is necessary for adhesive responses to transforming growth factor-beta1 in embryonic fibroblasts J Biol Chem 2006 281 16 10715 10726 10.1074/jbc.M511343200 16484225 40 Daugherty A Cassis LA Mouse models of abdominal aortic aneurysms Arterioscler Thromb Vasc Biol 2004 24 3 429 434 10.1161/01.ATV.0000118013.72016.ea 14739119 41 Daugherty A et al Angiotensin II promotes atherosclerotic lesions and aneurysms in apolipoprotein E-deficient mice J Clin Invest 2000 105 11 1605 1612 10.1172/JCI7818 10841519 42 Rateri DL et al Angiotensin II induces region-specific medial disruption during evolution of ascending aortic aneurysms Am J Pathol 2014 184 9 2586 2595 10.1016/j.ajpath.2014.05.014 25038458 43 Busch A et al Translating mouse models of abdominal aortic aneurysm to the translational needs of vascular surgery JVS Vasc Sci 2021 2 219 234 10.1016/j.jvssci.2021.01.002 34778850 44 Rodrigues-Diez RR et al CCN2 (cellular communication network factor 2) deletion alters vascular integrity and function predisposing to aneurysm formation Hypertension 2022 79 3 e42 e55 10.1161/HYPERTENSIONAHA.121.18201 35138869 45 Sharma N Hans CP Interleukin 12p40 deficiency promotes abdominal aortic aneurysm by activating CCN2/MMP2 pathways J Am Heart Assoc 2021 10 3 e017633 10.1161/JAHA.120.017633 33470127 46 Clement M et al Vascular smooth muscle cell plasticity and autophagy in dissecting aortic aneurysms Arterioscler Thromb Vasc Biol 2019 39 6 1149 1159 10.1161/ATVBAHA.118.311727 30943775 47 Li Y et al Single-cell transcriptome analysis reveals dynamic cell populations and differential gene expression patterns in control and aneurysmal human aortic tissue Circulation 2020 142 14 1374 1388 10.1161/CIRCULATIONAHA.120.046528 33017217 48 Pan H et al Single-cell genomics reveals a novel cell state during smooth muscle cell phenotypic switching and potential therapeutic targets for atherosclerosis in mouse and human Circulation 2020 142 21 2060 2075 10.1161/CIRCULATIONAHA.120.048378 32962412 49 Pedroza AJ et al Single-cell transcriptomic profiling of vascular smooth muscle cell phenotype modulation in Marfan syndrome aortic aneurysm Arterioscler Thromb Vasc Biol 2020 40 9 2195 2211 10.1161/ATVBAHA.120.314670 32698686 50 Zhao G et al Single-cell RNA sequencing reveals the cellular heterogeneity of aneurysmal infrarenal abdominal aorta Cardiovasc Res 2021 117 5 1402 1416 10.1093/cvr/cvaa214 32678909 51 Ruiz-Ortega M et al TGF-beta signaling in vascular fibrosis Cardiovasc Res 2007 74 2 196 206 10.1016/j.cardiores.2007.02.008 17376414 52 Holmes A et al CTGF and SMADs, maintenance of scleroderma phenotype is independent of SMAD signaling J Biol Chem 2001 276 14 10594 10601 10.1074/jbc.M010149200 11152469 53 Lu H et al Subcutaneous angiotensin II infusion using osmotic pumps induces aortic aneurysms in mice J Vis Exp 2015 103 53191 10.3791/53191 26436287 54 Liu CL et al Eosinophils protect mice from angiotensin-II perfusion-induced abdominal aortic aneurysm Circ Res 2021 128 2 188 202 10.1161/CIRCRESAHA.120.318182 33153394
PMC009xxxxxx/PMC9886497.txt
==== Front Rev Bras Ginecol Obstet Rev Bras Ginecol Obstet 10.1055/s-00030576 RBGO Gynecology & Obstetrics 0100-7203 1806-9339 Thieme Revinter Publicações Ltda. Rua do Matoso 170, Rio de Janeiro, RJ, CEP 20270-135, Brazil 36580947 10.1055/s-0042-1756213 RBGO-22-0123 Original Article Obstetrics Development and Validation of a Protocol for Pregnant Women Based on the Brazilian Dietary Guidelines Desenvolvimento e validação de um protocolo para orientação de gestantes baseado no guia alimentar para a população brasileirahttp://orcid.org/0000-0001-7264-9317 Tramontt Cláudia Raulino 1 http://orcid.org/0000-0003-0166-5568 Jesus Juliana Giaj Levra de 1 http://orcid.org/0000-0003-4087-1815 Santos Thanise Sabrina Souza 12 http://orcid.org/0000-0001-9693-7954 Rauber Fernanda 3 http://orcid.org/0000-0002-3756-2301 Louzada Maria Laura da Costa 1 http://orcid.org/0000-0001-7598-9229 Couto Vanessa Del Castillo 1 http://orcid.org/0000-0002-5871-5450 Hochberg Jacqueline Resende Berriel 1 http://orcid.org/0000-0003-2291-8536 Jaime Patrícia Constante 1 1 Núcleo de Pesquisas Epidemiológicas em Nutrição e Saúde da Universidade de São Paulo, São Paulo, SP, Brazil 2 Universidade Federal de Minas Gerais, Belo Horizonte, MG, Brazil 3 Faculdade de Medicina, Universidade de São Paulo, SP, Brazil Address for correspondence Cláudia Raulino Tramontt, PhD Dr. Arnaldo715, São Paulo, SP, [email protected] 29 12 2022 11 2022 1 12 2022 44 11 10211031 19 4 2022 21 6 2022 Federação Brasileira de Ginecologia e Obstetrícia. This is an open access article published by Thieme under the terms of the Creative Commons Attribution License, permitting unrestricted use, distribution, and reproduction so long as the original work is properly cited. ( https://creativecommons.org/licenses/by/4.0/ ) 2022 Federação Brasileira de Ginecologia e Obstetrícia. https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Objective  To develop and validate a protocol for the use of the Dietary Guidelines for the Brazilian Population (DGBP) in the individual dietary advice for pregnant women assisted in primary healthcare (PHC). Methods  Methodological study that involved the elaboration of a protocol in six steps: definition of the format, definition of the instrument to evaluate food consumption, systematization of evidence on food and nutrition needs of pregnant women, extraction of DGBP recommendations, development of messages of dietary guidelines and content, and face validity. The analyses of the validation steps were carried out by calculating the Content Validity Index (CVI) and thematic content analysis. Results  As products of the steps, the protocol structure was defined and the dietary advice for pregnant women were elaborated, considering physiological changes, food consumption, nutritional and health needs, and socioeconomic conditions of this population. The protocol was well evaluated by experts and health professionals in terms of clarity, relevance (CVI > 0.8), and applicability. In addition, the participants made some suggestions to improve the clarity of the messages and to expand the applicability of the instrument with Brazilian pregnant women. Conclusion  The instrument developed fills a gap in clinical protocols on dietary advice for pregnant women focused on promoting a healthy diet, contributing to a healthy pregnancy. In addition, it demonstrates potential to contribute to the qualification of PHC professionals and to the implementation of the DGBP recommendations. Resumo Objetivo  Desenvolver e validar um protocolo para uso do Guia Alimentar para a População Brasileira (GAPB) na orientação alimentar individual de gestantes atendidas na atenção primária à saúde (APS). Métodos  Estudo metodológico que envolveu a elaboração de um protocolo em seis etapas: definição do formato do documento, definição do instrumento de avaliação do consumo alimentar, sistematização de evidências sobre necessidades de alimentação e nutrição de gestantes, extração das recomendações do GAPB, desenvolvimento de mensagens de orientação alimentar e validação de conteúdo e validação aparente. As análises das etapas de validação foram realizadas através do cálculo do índice de validade de conteúdo e análise temática de conteúdo. Resultados  Como produtos das etapas, a estrutura do protocolo foi definida e as orientações alimentares de gestantes foram elaboradas, considerando as alterações fisiológicas, consumo alimentar, necessidades nutricionais e de saúde e condições socioeconômicas desta população. O protocolo foi bem avaliado por especialistas e profissionais de saúde nos critérios de clareza, pertinência (índice de validade de conteúdo > 0,8) e aplicabilidade. Além disso, os participantes deram sugestões para melhoria da clareza das mensagens e para ampliar a aplicabilidade do instrumento em gestantes brasileiras. Conclusão  O instrumento desenvolvido preenche uma lacuna sobre protocolos clínicos de orientação alimentar para gestantes focado na promoção da alimentação saudável, contribuindo para uma gestação saudável. Além disso, o instrumento demonstra potencial para contribuir na qualificação de profissionais da APS e implementação das recomendações do GAPB. Keywords dietary guidelines practice guidelines prenatal nutrition primary health care validation study Palavras-chave guias alimentares guias de prática clínica nutrição pré-natal atenção primária à saúde estudo de validação ==== Body pmcIntroduction Food and nutrition have been widely discussed in the global public policy agenda due to their importance in human health. 1 2 During pregnancy, a healthy diet favors good fetal development and women's health and well-being, in addition to preventing the onset of diseases such as gestational diabetes, hypertension, and excessive weight gain, and is associated with better health conditions in childhood and adult life. 3 4 For this reason, many countries have invested in the development of advices, recommendations, and guidelines to support adequate nutrition to ensure healthy pregnancy and delivery. 5 6 7 8 Pregnancy can be a window of opportunity for dietary interventions, since pregnant women tend to change their diet due to their new condition and in favor of the fetus health. However, this change does not always occur in the sense of improving the diet. Recent data on the diet of Brazilian pregnant women show a high consumption of ultraprocessed foods and a low-quality diet, evidenced by the low consumption of fruits and vegetables, especially in pregnant teens, mulatto and black women, and smokers. 9 10 11 12 13 The monitoring carried out by health professionals during prenatal care is a good opportunity to provide pregnant women with adequate advice on nutrition. However, dietary advice in prenatal care does not always occur, and when it does occur, it is carried out inappropriately or limited in content, not always consistent with current dietary guidelines, insufficient to meet the health needs of pregnant women, in addition to occurring from a vertical relationship between professional and user, perpetuating the care discontinuity, and nonadherence to dietary advice. 14 15 16 17 18 The Dietary Guidelines for the Brazilian Population (DGBP) contains healthy diet recommendations officially adopted by the Ministry of Health throughout the national territory which are applicable to the entire population > 2 years old. 19 However, which recommendations on healthy diets and how they should be directed to pregnant women have not yet been established. Clinical practice protocols are considered the “basis of accurate advising” and the foundation of evidence-based practice. 20 The absence of instruments such as these to guide the clinical practice of health professionals is a limiting factor for adequate dietary advice to occur. A systematic review on the quality of clinical practice guidelines for nutrition during pregnancy found most of the guidelines inadequate in terms of development rigor, content of nutritional recommendations, applicability, and the declaration of funding sources and conflict of interests. 21 These findings highlight the lack of tools for dietary advice during pregnancy that help clinical practice, are updated based on the best available scientific evidence, improve prognosis and qualify the health care. Considering the importance of adequate dietary advice in the gestational period and the absence of instruments to support clinical practice, the present study aimed to develop and validate a protocol for using the DGBP in dietary advice for pregnant women assisted at primary healthcare (PHC). Methods This is a methodological study on the development and validation of a protocol for the use of the DGBP for dietary advice for pregnant women assisted at PHC. The Dietary Guidelines for the Brazilian Population Published in 2014, the DGBP is an official document that presents recommendations on healthy eating based on evidence on the relationship between diet and health, systematized in 5 chapters. Its elaboration derived from a holistic perspective, comprising the biological, cultural, social, and environmental dimensions of food. This document is an instrument to support nutrition education actions at the Brazilian Unified Health System (SUS, in the Portuguese acronym). 19 The development of this protocol is part of a matrix project that assumed the DGBP as the technical reference for guidance and promotion of healthy eating. The project aimed to develop a series of individual dietary protocols for applying the DGBP in different life cycles/events, aimed at health promotion and care qualification in relation to the most prevalent dietary needs in the PHC context. The methodological basis that guided the development of the series was described in a previous publication. 22 Development of the Protocol to Use the Brazilian Dietary Guidelines on Dietary Advice for Pregnant Women To elaborate the series of protocols, a team of eight researchers free of conflict of interests and experts in dietary advice based on the DGBP and on PHC was formed. The elaboration of the protocol to use the DGBP for pregnant women was carried out following six steps. The first addressed the definition of the protocol format whereas the second dealt with the definition of the instrument for assessing food consumption; the third step consisted of extracting the recommendations from the DGBP applicable to individual dietary advice; the fourth was the evidence systematization on the dietary and nutrition needs of pregnant women; the fifth addressed the development of dietary advice messages for pregnant women; and, finally, in the last step, the content and face validity was carried out. Step 1. Format Definition At this step, documents on the elaboration of guidelines for clinical practice were analyzed, as well as PHC protocols already published in Brazil. The objective was to investigate the possible formats, analyze their characteristics and identify the most appropriate ones to guide dietary advice for promoting healthy eating in individual appointments in PHC. Step 2. Definition of the Instrument to Assess Food Consumption To support decision-making for dietary advice, a search was carried out on existing instruments for assessing food consumption. The objective was to identify an assessment instrument of the main recommendations from the DGBP that could be used by any PHC professional during individual appointments. Step 3. Evidence Systematization on Food and Nutrition Needs of Pregnant Women To identify the scientific evidence on the dietary needs and specificities and food consumption of pregnant women, a literature review was carried out in the Scielo, Lilacs, PubMed, and gray literature databases (academic Google). We searched for original articles published in English, Portuguese, or Spanish with evidence from national surveys. The search was conducted using the terms food consumption OR food intake AND pregnant woman OR pregnancy OR prenatal nutrition AND brazil . Additionally, analyses were carried out on the 2017–2018 Family Budget Survey 23 with a focus on pregnant women and searches in books and in technical materials that included dietary advice for pregnant women. Step 4. Extracting the Recommendations Two researchers carried out a systematic reading of chapters 2 (Food choice), 3 (From food to meals), 4 (Ways of eating), and 5 (Understanding and overcoming obstacles) of the DGBP in order to identify appropriate and relevant recommendations that were in line with the indicators of the food consumption assessment instrument identified in Step 2 (Definition of the instrument). Furthermore, additional recommendations considered relevant to be included in the nutritional advice for pregnant women were identified, even if they were not addressed by the food consumption assessment instrument. The other five researchers in the expert team reviewed the recommendations to elaborate the final list. Step 5. Development of dietary advice messages for pregnant women At this step, dietary advice messages were prepared based on the recommendations of the DGBP extracted in Step 3 (Evidence systematization) considering the specificity of the diet of pregnant women systematized in Step 4 (Extraction of recommendations), also including the obstacles to a healthy diet faced by this population. Step 6. Content and Face Validity As a product of the previous steps, version 1 of the protocol was elaborated. This first version was sequentially submitted to content and face evaluations. 6.1–Content validity: 19 experts in the areas of health and nutrition for pregnant women, dietary guideline, and nutritional care in PHC were invited to compose a panel of judges. The experts received by email the first version of the protocol, support materials explaining the validation process, the systematization of life cycle evidence (result of Step 4), and an online form to assess the clarity and relevance of each protocol component using a 4-point scale (1–the item is not clear/relevant; 2–major revisions are needed to make the item clear/relevant; 3–minor revisions are necessary to make the item clear/relevant; and, 4–the item is clear/relevant). For each item judged as not clear/relevant, a justification was requested. The form also had a field for additional suggestions. The protocol components are identified in Table 1 . After completing the form, the experts were invited to participate in an online focus group to collect general impressions about the protocol. Upon the collaboration from the experts, changes were made to version 1 of the protocol in order to contemplate the pertinent suggestions, resulting in version 2. Table 1 Description of the content validity index by the protocol component of the protocols for the Dietary Guidelines for the Brazilian Population use in nutrition advising for pregnant women Content Validity Index Component Clarity (min-max) Relevance (min-max) Average CVI Introductory text 1 (3–4) 0.86 (2–4) 0.93 How to use the protocol? 0.93 (2–4) 0.93 (2–4) 0.93 Food consumption assessment 0.93 (1–4) 0.86 (1–4) 0.895 Flowchart 0.93 (2–4) 0.86 (1–4) 0.895 Recommendation 1 (R1): Encourage the daily consumption of beans” R1 - Advising 1 (3–4) 0.93 (2–4) 0.965 R1 - Suggestions for variations 1 (3–4) 0.93 (2–4) 0.965 R1 - Reason 1 (3–4) 0.93 (2–4) 0.965 R1 - Obstacles 1 (3–4) 0.93 (2–4) 0.965 Recommendation 2 (R2): Advise to avoid the consumption of sweetened beverages R2 - Advising 0.86 (1–4) 0.93 (1–4) 0.895 R2 - Suggestions for variations 0.86 (2–4) 0.86 (2–4) 0.86 R2 - Reason 0.93 (2–4) 0.86 (2–4) 0.895 R2 - Obstacles 1 (3–4) 0.79 (2–4) 0.895 Recommendation 3 (R3): Advise to avoid ultraprocessed foods. R3 - Advising 0.93 (1–4) 0.93 (1–4) 0.93 R3 - Suggestions for variations 0.93 (2–4) 0.86 (2–4) 0.895 R3 - Reason 1 (3–4) 1 (3–4) 1 R3 - Obstacles 1 (3–4) 1 (3–4) 1 Recommendation 4 (R4) - Advise the daily consumption of vegetables R4 - Advising 0.93 (1–4) 0.93 (1–4) 0.93 R4 - Suggestions for variations 0.86 (1–4) 0.93 (1–4) 0.895 R4 - Reason 1 (3–4) 1 (3–4) 1 R4 - Obstacles 0.79 (2–4) 0.93 (2–4) 0.86 Recommendation 5 (R5) - Advise the daily consumption of fruits R5 - Advising 1 (3–4) 1 (3–4) 1 R5 - Suggestions for variations 1 (3–4) 1 (3–4) 1 R5 - Reason 1 (3–4) 1 (3–4) 1 R5 - Obstacles 0.93 (1–4) 0.93 (1–4) 0.93 Recommendation 6 (R6) - Advise the user to eat in appropriate environments and with attention. R6 - Advising 0.93 (1–4) 0.93 (1–4) 0.93 R6 - Suggestions for changing habits 1 (3–4) 1 (3–4) 1 R6 - Reason 1 (3–4) 1 (3–4) 1 Additional Messages 0.86 (1–4) 0.79 (1–4) 0.825 Valuing the existing eating practice 0.93 (2–4) 1 (3–4) 0.965 Abbreviation: CVI, content validity index. Note: The numbers in parentheses correspond to the minimum and maximum scores obtained in the clarity and relevance assessment using the 4-point scale: (1) “the item is not clear/relevant”; (2) “major revisions are required to make the item clear/relevant”; (3) “minor revisions are required to make the item clear/relevant”; and (4) “the item is clear/relevant”. 6.2–Face validity: Subsequently, the face validity of version 2 of the protocol was conducted, which aimed to identify the understanding and applicability of the protocol by its potential users – health professionals. Sixteen higher education-level health professionals who worked in PHC from the 5 regions of the country were invited to participate in online focus groups. The contributions from health professionals were considered to elaborate the final version of the protocol. All focus groups were conducted by the team researchers, playing the roles of moderator and observer. Participants were instructed to keep the microphone off, activating it upon speaking. The desire to express an opinion was indicated by chat or by using the raise your hand feature, available on the virtual platform, and followed by the team members. All focus groups were recorded and later transcribed. Data Analysis of the Validity Step For the analysis of the answers filled in the online form, the Content Validity Index (CVI), which measures the proportion of experts who expressed agreement regarding the clarity and relevance of each protocol component, was used. The CVI was calculated by the proportion of grades (3) and (4) divided by the total number of specialists, separated for clarity and relevance. The average CVI was also calculated considering a simple average of the CVI values for clarity and relevance. Components with CVI > 0.80 were considered adequate; that is, they did not need to be modified. 24 The justifications indicated in the online form were analyzed according to Bardin 25 in order to identify suggestions and comments related to the specificity of the pregnant woman. Based on the recognition of the most recurrent suggestions made by the specialists, two researchers defined themes for suggestions and selected demonstrative examples. Suggestions considered relevant were incorporated into version 2 of the protocol. The transcripts of the focus groups with professionals were also evaluated with thematic content analysis. 25 Two team members read the transcripts, predefined the analysis categories, and extracted the related statements. The predefinition of categories and the extraction of the statements were monitored and verified by a third researcher of the team. Ethical Aspects All participants were voluntarily invited to participate in the study and signed a Free and Informed Consent Form. The focus groups were recorded and transcribed upon permission of the participants. The present study was approved by the ethics committee of the Universidade de São Paulo (4,232,862) and was conducted in partnership with the Ministry of Health. Results The results of the steps described in the methodology will be presented in four topics: (1) Structuring the protocol, referring to steps 1 and 2 presented in the methods; (2) evidence on the dietary and nutrition needs of pregnant women, summarizing steps 3 and 4; (3) dietary advice messages for pregnant women, as a result of step 5; and (4) protocol validity, responding to step 6. Protocol Structuring The adopted methodology resulted in a “Protocol of Use,” in compliance with the definition of the National Commission for the Incorporation of Technologies (CONITEC, in the Portuguese acronym) of the SUS, for dietary advice for pregnant women. The consumption assessment instrument that best met the protocol needs was the Food Consumption Markers form the Food and Nutrition Surveillance System (SISVAN, in the Portuguese acronym), since it is an instrument already implemented in PHC that addresses the main recommendations of the DGBP concisely and objectively through healthy and unhealthy eating markers. The protocol was structured in the following subsections: Introduction: with brief information about the DGBP, characteristics of the pregnancy, and outline of the target audience of the nutritional advice, of the purpose of the protocol, and general instructions on the conduct when carrying out dietary advice. Step-by-step on how the protocol should be used. Instrument to assess food consumption: SISVAN food consumption markers. Decision flowchart: The flowchart directs the health professional to dietary advice according to the questions on the SISVAN form, to priority recommendations, and to the golden rule of the DGBP. Recommendations: Deals with dietary advice messages related to each consumption marker on the SISVAN form. Six recommendations were established. Valuing the existing eating practice: A message to encourage the continuity of healthy eating practices. Additional messages: Messages that were not covered by the SISVAN markers but are relevant in the dietary advice for the pregnant woman according to the literature review. Evidence on the Dietary and Nutrition Needs of Pregnant Women The search in the databases found 24, 37, and 189 results in Scielo, Lilacs, and PubMed, respectively. Out of these, reading by title identified 57 articles eligible for abstract reading, excluding duplicates. After reading the abstracts, 17 manuscripts were excluded because they were not related to the purpose of the present review. Eventually, 35 scientific articles that investigated food consumption by pregnant women, key nutrients and nutritional demand, and the rapport between food and health outcomes were included for full reading. Additionally, information from technical materials, such as the “Pregnant Woman Booklet” published by the Ministry of Health, and original analyses of the 2017–2018 POF report, also supported the preparation of the protocol. Evidence has shown that the consumption of a wide variety of in natura and minimally processed foods, as well as water, is particularly relevant to meet the need for essential nutrients for pregnant women, such as iron, folic acid, calcium, vitamin A and vitamin D, among others. According to data from the 2017-2018 POF, 23 in natura or minimally processed foods had a caloric share of 53% of the total calories consumed among pregnant women, with rice, beans, red meat, and poultry corresponding, jointly, to 29% of the calories in the diet of Brazilian pregnant women. Cooking ingredients, processed, and ultraprocessed foods, in turn, had an approximate caloric share of 16, 10, and 20%, respectively. 23 When comparing the food consumption patterns of pregnant and nonpregnant women between 18 and 38 years old, we could see a better diet in pregnant women. This is due to the greater caloric participation from in natura and minimally processed foods (1.82% higher in pregnant women compared with nonpregnant women), with emphasis on the consumption of fruit and offal items and the lower caloric participation of ultraprocessed foods in the diet of pregnant women (1.4% lower in pregnant women). Among ultraprocessed foods, there was a lower consumption of soft drinks and higher consumption of sausages and artificial fruit juice during pregnancy compared with nonpregnant women, leading to practices that require attention. 23 It is also noteworthy that, for a more adequate dietary advice, it is essential that the health professional is attentive to aspects related to social vulnerability and income, support network, the age of pregnant women, and their working conditions. Other factors that require attention in the gestational period are the physiological changes and symptoms that can influence food consumption in this period, such as nausea, vomiting and dizziness, heartburn, gastric fullness, constipation, weakness, fainting, physical pain, swelling, and oral changes. These symptoms may vary throughout pregnancy quarters and are manageable with adequate dietary advice for this period. Finally, the process of extracting content from the DGBP identified 99 recommendations related to food consumption indicators evaluated by the SISVAN instrument and 5 additional recommendations considered relevant for inclusion in the protocol. Dietary Advice Messages for Pregnant Women Based on the results presented in the previous topics, the dietary advice messages were developed contemplating six recommendations linked to the SISVAN markers: 1) encourage the daily consumption of beans; 2) advise avoiding the consumption of sweetened beverages; 3) advise avoiding the consumption of ultraprocessed foods; 4) encourage the daily consumption of vegetables; 5) encourage the daily consumption of fruits; and 6) encourage the user to eat in appropriate environments and with attention. Each recommendation addresses the dietary advice that should be offered, suggestions and healthy alternatives, justification that supports the professional about the advice, possible obstacles that can make it difficult to adhere to the advice, and strategies to overcome them. Furthermore, additional messages were elaborated for important issues regarding the diet pregnant women not covered by the SISVAN markers. Example of the elaborated messages can be found in Chart 1 . Chart 1 Example of messages elaborated for dietary advice for pregnant women SISVAN form question Individual's answer Dimension of recommendation Dietary advice messages Did the user consume sweetened beverages yesterday? No Value the existing eating practice “Value the healthy eating practice as it strengthens the practice, encourages its continuity, and can provide important information around healthy diets. You may use the justification to value the existing eating practice.” Yes General recommendation “Recommend pregnant women to avoid the consumption of sugar-sweetened beverages such as soft drinks and artificial juices, which are also known as ultraprocessed beverages.” Suggestion on healthy alternatives “Encourage the consumption throughout the day of pure water that is proper for human consumption, or 'enriched' with slices of lime, mint leaves, or pineapple peel. In case of nausea, encourage the consumption of cold water”. Reason “Sweetened beverages are not recommended for pregnant women, as they are usually added with a lot of sugar, flavorings, dyes, and other cosmetic additives, in addition to many of them having caffeine in the composition. Such compounds can worsen common symptoms of pregnancy, such as nausea and vomiting, and, if they contain caffeine, can increase the risk of miscarriages, premature births, low birthweight, and stillbirth. The consumption of these drinks can also interfere with the water intake, which, during this period, is especially important to boost blood circulation and irrigation of the uterus and the placenta, to maintain amniotic fluid at adequate levels, to stabilize blood pressure, in addition to eliminating toxins that increase the risk of urinary infections”. Understanding and overcoming obstacles “For pregnant women who do not frequently drink much water, advise them to carry bottles of water in their bags when leaving home, ensuring access to drinking water during the day, in addition to saving money. At home or at work, always keep a bottle of water at hand.” Additional messages "Sweeteners are not recommended during pregnancy, as they may increase the likelihood of premature birth and of development of asthma in children up to 7 years old, in addition to being associated with metabolic changes in women and with a higher likelihood of overweight in children in early childhood." Source: Ministério da Saúde and Universidade de São Paulo (2021). 34 Protocol Validity Version 1 of the protocol, submitted to content validity by a panel of experts, had 14 female participants, from the South ( n =  3), Southeast ( n =  8), Midwest ( n =  1), and Northeast ( n =  2) regions of the country, with higher education in nutrition ( n =  11), nursing ( n =  2), or medicine ( n =  1), and working in the academic and research, public management, assistance, or primary healthcare areas. The panel took place in November 2020. All protocol components had an average CVI above the cutoff point (0.80). Even so, all contributions from experts in each component were analyzed individually. Out of the 29 components evaluated, 14 obtained total agreement (CVI = 1.0) for clarity and 09 for relevance ( Table 1 ). According to the assessors, the components that most required changes were recommendations 2 and 4 as well as additional recommendations. Some suggestions made by the specialists in the components that required more attention according to the results of the CVI analysis and that were incorporated into the protocol are presented in Chart 2 . In addition to suggestions for changes and comments on relevance and clarity, the experts brought elements about signs/symptoms and dietary strategies to be considered for pregnant women, specific nutritional needs, health outcomes, need for comprehensive care by health professionals, food safety, and broadening the look at issues related to eating habits and the regional diversity of the diets of pregnant women. Chart 2 Suggestion theme and its definition, example of suggestion and modification in the Protocols for the Dietary Guidelines for the Brazilian Population use in nutrition advising for pregnant women Suggestion theme and definition Suggestion example Example of change made to the protocol Clarity Suggestions related to the adequacy of technical terms, to the understanding of meanings and information, and technical content qualification. “Specify the size of the coffee cup that can be consumed.” Regarding caffeine consumption, a safe intake should not exceed 100 mg of caffeine, which is equivalent to a cup of caffeinated coffee or tea (such as mate tea, black and green teas, etc.). Specific needs Suggestions related to nutritional needs and specific recommendations for pregnant women. “Regarding specifically liver, the professional needs to know what to recommend. It is common that its consumption is encouraged during pregnancy. Should this be done? what is a safe weekly consumption without risk of excess vitamin A?” Offal, such as liver, is food source of iron and vitamin A, an essential nutrient in this period. If the pregnant woman is used to eating that, remind her of the importance of including it in a meal along the week. Health outcomes Suggestions regarding the most common health outcomes (weight gain, gestational diabetes, and hypertension). "Warn about excessive weight gain especially in ultraprocessed and sweetened drinks advice". Greater consumption of ultraprocessed foods during pregnancy is related to excessive weekly weight gain and greater chances of postpartum weight retention, factors that are related to negative maternal and child health outcomes, such as the development of obesity in the mother, and higher birthweight in infants. Food safety Suggestions related to hygiene and food handling relevant to pregnant women. “I suggest incorporating advice on the relevance of the correct hygiene of vegetables that will be consumed raw”. Many people may not consume fresh foods for fear of contamination by toxoplasmosis. Advise them to wash the fruits or vegetables under running water and then place them in a solution of 1 spoon of sodium hypochlorite or 1 spoon of bleach suitable for using in food to a liter of water for 15 minutes. After this period, rinse in running water. Food will thus be ready and safe for consumption. The face validity with health professionals, carried out with version 2 of the protocol, had the participation of 6 professionals, mostly women ( n =  4), working in PHC in the south ( n =  1), southeast ( n =  4), and northeast (1) regions of Brazil, including a nutritionist ( n =  1), doctors ( n =  2), and nurses ( n =  3), and took place in December 2020. The thematic analysis of the focus group transcripts identified two major themes and their respective categories and subcategories of analysis. The theme "technology in health" covered presentations that demonstrate the ability of the protocol to qualify and update health professionals for dietary advice for pregnant women based on the best technical and scientific evidence, in addition to contributing to promote the confidence of professionals to address the healthy eating issue. The “applicability” theme, on the other hand, gathered reports that point out possible challenges related to the work process with potential influence on the applicability of the protocol in prenatal appointments, while indicating pregnancy as an opportune phase for diet changes and subsequent application of the protocol. The themes and their respective categories and subcategories are presented in Table 1 with examples ( Chart 3 ). Chart 3 Themes, categories and subcategories found in the face validity with health professionals in relation to the Protocol for the Dietary Guidelines for the Brazilian Population use in nutrition advising for pregnant woman Topic: Health technology Category: Qualification of dietary advice Subcategory 1: Qualification of dietary advice when handling common pregnancy signs and symptoms: Example: “... food is the biggest challenge. Because when she arrives to start prenatal care at the unit, what is her main complaint? Nausea, vomiting, difficulty in eating... these are the advice that I use to not focus on. We do use to focus on other things and forget about it a little... I myself have a hard time saying: oh, what food items can you avoid? You talk a little about coffee. Which I found amazing. And that was something I didn't even talk to them about, which I now intend to start talking about. And…but which food items can they avoid that will cause less nausea? Which food items cause flatulence the most, right? They have a lot of questions about it. And we used to approach it in a very generic way: 'Ah, avoid sweets, you know, avoid saturated fats that are deep-fried foods. Avoid excess salt. These foods that are processed, right…” Subcategory 2: Qualification of dietary advice by other categories of health professionals (in addition to nutritionists): Example: “… so, here in my job, for the pregnant woman to get to an individual appointment with the nutritionist, usually faces some changes already… [such as] excessive weight gain, some pathology, or a high-risk prenatal care. In addition, we make group approaches and who will provide this service are the doctors, nurses, and nursing technicians. In addition, we make group approaches and who will provide this service are the doctors, nurses, and nursing technicians. And many times, … in my practice, with the rush, diet goes a little unnoticed.” Subcategory 3: Qualification of dietary advice for pregnant women based on scientific evidence and the food culture of Brazilian people: Example: “... you talked in the leaflet a little, you know, about social classes, right, and about consumption, you know, that white women, white pregnant women usually consume a little more... not a little, right... more healthy foods. The mulatto and black women, yes, consume more industrialized foods, right, processed. And for me, this is essential. Very important. So yeah, I wish I had material on that, because I can't find it, I can't get it. It is not accessible. Most of the women that I start prenatal care today, in my region, right, I work in the south region, extreme south, are mostly black and mulatto women. So, this woman is the woman who consumes more ultraprocessed foods… but for me it is essential that I look at this data that exists about black and mulatto women, who consume more ultraprocessed foods, and that I can advise them in an accessible way, the foods that are also available in other ways and that are healthier for them during this time…” Subcategory 4: Qualification of dietary advice for pregnant women based on the recommendations of the Dietary Guidelines for the Brazilian Population: Example: “... going back to the case of pregnant women, I usually use the Guideline very little, because it has been very comprehensive. And maybe, it is, to think of a protocol, in specific guidelines, for pregnant women. For me, it's very interesting.” Category: Promoting the trust of professionals Example: “So, yeah, for me, yes, it gives me a lot more confidence. I realize that, over time, as we are in a very dynamic job, there is a lot going on, right, no one was expecting, for example, this pandemic, and there are guidelines all the time, every day we receive something new; then I feel that, during my practice, many things end up fading into oblivion, or we end up losing a little, right. So, I want material that makes me think always, yes, that doesn't let me fade into oblivion, right, that makes me confident so that I can always offer this protocol, right, and that is always relevant. Of course, things are changing all the time, but with the material that I can, that I can offer that and I think it's going to offer me a lot more confidence, for sure.” Topic: Applicability Category: Challenges and spaces for using the protocol in prenatal appointments Subcategory 1: Challenges linked to the health work process that influence the applicability of the protocol: Example: “... yes, what I think will be a great challenge, J., at least in my job, is that I still see professionals who are not nutritionists, with difficulty, … the routine of fitting in the filling of the food consumption marker, which is the starting point for the protocol. The protocol, it will be based on you filling in the marker and from the marker responses, you should guide that. Yeah, and then I think that, I don't know, it's up to us to think of strategies to encourage this filling, if not in all appointments or for everybody, yes, the entire population served, but focused, at least in this period, to be able to provide advice for pregnant women. So, kind of, first, it would be important for us to understand food as a health determinant and that it belongs to all of us and how important it is to be addressed. And the marker is a kickoff. So, I think this is a big initial challenge.” Subcategory 2: Pregnancy as an opportune phase for changes in diet and subsequent application of the protocol. Example: “... and especially the pregnant women, right, which are a group that I see like that, in my job, that is looking for care, self-care, is looking for more information. It's a group that's more interested, mainly, in nutritional information, right, in nutrient information, vitamins, and everything else, and then, I think the more knowledge we have, the simpler, both for us to learn and to pass on to these pregnant women, I think it will flow much better and it is very helpful. Because we have a lot of extensive material, a lot of long stuff and we can't stop to read it. And we end up losing super useful knowledge, but that we can't read. You can't stop to acquire that knowledge.” Discussion The present study describes the process of elaborating and validating a protocol for the use of the DGBP in dietary advice for pregnant women in individual appointments. The protocol elaboration steps included the formation of a technical team, definition of the protocol format, definition of the food consumption assessment instrument, extraction of recommendations from the DGBP, literature review related to the dietary needs of pregnant women, development of advice messages and, finally, content and face validity of the protocol. The product resulting from these steps was well evaluated by specialists in food, nutrition, care, and women's health, as well as by the target audience, proving to be clear and relevant for use in the clinical practice of professionals working in PHC. Although the importance of adequate nutrition during pregnancy is known, to the best of the authors knowledge, there are no studies to date that developed recommendations on healthy eating during pregnancy aimed at health professionals, guided by theoretical-scientific references, applicable throughout the country, and comprising different regional and socioeconomic contexts. Nutritional advice during pregnancy nationwide is only found in the Pregnant Woman Booklet, in the form of “10 steps to a healthy diet.” 26 However, this instrument is aimed at pregnant women, with general information about nutrition during pregnancy, and it is not possible to identify which diet issues need to be prioritized. In Brazil, other publications are found in the academic literature, such as the validation of an educational booklet for healthy eating during pregnancy with regional foods 27 and the elaboration and evaluation of educational material for pregnant women in the northeast region of the country on the importance of adequate diet during pregnancy, with emphasis on vitamin D. 28 However, the first publication adopts a technical reference on nutrition that is already outdated, and both publications use a language that addresses the pregnant woman as a target audience. In addition, although significant, they deal with local efforts and, in general, they were not developed with the same objectives as the instrument presented in the present publication, limiting their comparability. Studies that propose to develop protocols to guide clinical practice in pregnant women care usually have a nutrient-centered approach or focus on preventing excessive gestational weight gain. 14 29 In the second case, the purpose of these propositions considers, above all, the relevance of this prognosis in negative outcomes in the health of the mother and of the child. The protocol proposed in the present study advances by proposing an instrument that, by recognizing the DGBP as a theoretical-scientific reference, focuses both on the promotion of healthy eating and on the prevention of excessive weight gain. This occurs because the Brazilian Guidelines adopt a multidimensional paradigm of healthy eating, including cultural, biological, economic, and environmental aspects, in addition to considering the level of food processing. This approach includes robust evidence on the rapport between healthy eating and positive health outcomes and between consumption of ultraprocessed foods and excessive weight gain during pregnancy, as well as postpartum weight retention, and other negative maternal and child health outcomes, such as development of maternal obesity, and higher birth weight in infants. 13 30 31 In this sense, the protocol, in addition to guaranteeing recommendations on healthy eating for pregnant women who are not at nutritional risk, also favors those who present some risk of excessive weight gain, allowing its use in the range of demands of individual care for pregnant women. Another relevant feature of the protocol is its potential for use in in-service training initiatives to qualify the workforce in food and nutrition. The need to invest in the training of health professionals to promote healthy eating is recognized, in view of the global emergence of preventable chronic noncommunicable diseases through healthy eating. 32 The developed protocol was satisfactorily identified in the evaluation panels with professionals as a health technology that focuses on the qualification of professional practice. They recognized the potential of the instrument to assist in the clinical management of pregnant women, in addition to feeling more confident to advise on food, especially those who are not nutritionists, as they are using a tool based on a theoretical framework, giving value to the DGBP and implementing it in professional practice. Although the importance of following methodological steps in the development of clinical protocols is consolidated, few studies elucidate in detail the process of building protocols in the dietary guidelines for pregnant women. 21 The protocol presented in the present study was developed following rigorous methodological steps of development and evaluation proposed by guidelines, enabling its reproducibility. 8 20 21 Although the validation performed by a panel of experts is the most used to validate protocols in the health area, 33 the process carried out in the present study may have some limitations. The difficulties posed by the COVID-19 pandemic required adaptations in the methodology, such as carrying out the evaluation panels virtually. This format may have limited the interaction between participants and hampered the discussion. In addition, there was difficulty for health professionals to participate due to the work overload on the front line and the unavailability for prior reading of the material and participation in the panels. As a way of trying to overcome these obstacles, the meetings were held after working hours and with reduced time. Finally, there is a need for additional efforts in the dissemination and implementation of the material among health professionals working in PHC, promotion of nutrition training in the permanent health education agenda, and maintenance of the routine service in the use of the tool. In the focus groups carried out in the present study, professionals pointed out the potential for protocol inclusion in prenatal appointments. The protocol was published by the Ministry of Health in 2021 34 and is available for consultation on the website of the virtual health library (available at: https://bvsms.saude.gov.br/bvs/publicacoes/protocolos_guia_alimentar_fasciculo3.pdf ). Future works will evaluate the effectiveness of an educational intervention for the use of the protocol in the qualification of the practice of health professionals. Conclusion The developed and validated instrument presented in the present study sought to fill an important gap in the development of clinical protocols for individual dietary advice for pregnant women. In addition, the protocol seems promising to qualify the performance of health professionals working in PHC and enables the implementation of the DGBP in clinical practice in individual care. Acknowledgements Financial support: Organização Pan-Americana da Saúde (OPAS- escritório Brasil) – Carta-Acordo SCON2019-00489. Contributors Conflict of Interests The authors have no conflict of interests to declare. All authors were involved in the design and interpretation of the analyses, contributed to the writing of the manuscript, and read and approved the final manuscript. ==== Refs References 1 Food and Agriculture Organization of the United Nations (FAO) Conference outcome document: Framework for Action–from committments to action [Internet] Rome FAO 2014[cited 2022 Mar 7]. Available from:https://www.fao.org/3/mm323e/mm323e.pdf 2 Global nutrition report: action on equity to end malnutrition [Internet]. Bristol: Development Initiatives 2020[cited 2022 Mar 7]. Available from:https://reliefweb.int/report/world/2020-global-nutrition-report-action-equity-end-malnutrition 3 Danielewicz H Myszczyszyn G Dębińska A Myszkal A Boznański A Hirnle L Diet in pregnancy-more than food Eur J Pediatr 2017 176 12 1573 1579 10.1007/s00431-017-3026-5 29101450 4 Santos I DS Crivellenti L C Franco L J Sartorelli D S Relationship between the quality of the pregnant woman's diet and birth weight: a prospective cohort study Eur J Clin Nutr 2021 75 12 1819 1828 10.1038/s41430-021-00894-6 33742155 5 Mason J B Shrimpton R Saldanha L S Ramakrishnan U Victora C G Girard A W The first 500 days of life: policies to support maternal nutrition Glob Health Action 2014 7 23623 10.3402/gha.v7.23623 24909407 6 Shrimpton R Global policy and programme guidance on maternal nutrition: what exists, the mechanisms for providing it, and how to improve them? Paediatr Perinat Epidemiol 2012 26 01 315 325 10.1111/j.1365-3016.2012.01279.x 22742618 7 World Health Organization Good maternal nutrition: the best start in life [Internet]. Copenhagen: WHO Regional Office for Europe 2016[cited 2022 Mar 7]. Available from:https://www.euro.who.int/__data/assets/pdf_file/0008/313667/Good-maternal-nutrition-The-best-start-in-life.pdf 8 Tsakiridis I Kasapidou E Dagklis T Leonida I Leonida C Bakaloudi D R Nutrition in pregnancy: A comparative review of major guidelines Obstet Gynecol Surv 2020 75 11 692 702 10.1097/OGX.0000000000000836 33252699 9 Holand B L Fonseca S G Drehmer M Bosa V L Adequacy of prenatal care considering nutritional assistance in Southern Brazil: Maternar Cohort Study Cad Saude Publica 2021 37 06 e00130320 10.1590/0102-311X00130320 34231762 10 Lacerda E M Kac G Braga da Cunha C Leal MdoC Consumo alimentar na gestação e no pós-parto segundo cor da pele no município do Rio de Janeiro Rev Saude Publica 2007 41 06 985 994 10.1590/S0034-89102007000600014 18066468 11 Gomes C B Vasconcelos L G Cintra R M Dias L C Carvalhaes M A Hábitos alimentares das gestantes brasileiras: revisão integrativa da literatura. Ciênc Saúde Coletiva 2019 24 06 2293 306 10.1590/1413-81232018246.14702017 12 Pereira M T Cattafesta M Santos Neto E TD Salaroli L B Maternal and sociodemographic factors influence the consumption of ultraprocessed and minimally-processed foods in pregnant women Rev Bras Ginecol Obstet 2020 42 07 380 389 10.1055/s-0040-1712996 32559797 13 Naspolini N F Machado P P Fróes-Asmus C IR Câmara V M Moreira J C Meyer A Food consumption according to the degree of processing, dietary diversity and socio-demographic factors among pregnant women in Rio de Janeiro, Brazil: The Rio Birth Cohort Study of Environmental Exposure and Childhood Development (PIPA project) Nutr Health 2021 27 01 79 88 10.1177/0260106020960881 33076740 14 McDonald S D Pullenayegum E Taylor V H Lutsiv O Bracken K Good C Despite 2009 guidelines, few women report being counseled correctly about weight gain during pregnancy Am J Obstet Gynecol 2011 205 04 3330 3.33E8 10.1016/j.ajog.2011.05.039 15 de Jersey S J Nicholson J M Callaway L K Daniels L A An observational study of nutrition and physical activity behaviours, knowledge, and advice in pregnancy BMC Pregnancy Childbirth 2013 13 115 10.1186/1471-2393-13-115 23688111 16 Whitaker K M Wilcox S Liu J Blair S N Pate R R Patient and provider perceptions of weight gain, physical activity, and nutrition counseling during pregnancy: a qualitative study Womens Health Issues 2016 26 01 116 122 10.1016/j.whi.2015.10.007 26621605 17 Verma B A Nichols L P Plegue M A Moniz M H Rai M Chang T Advice given by community members to pregnant women: a mixed methods study BMC Pregnancy Childbirth 2016 16 01 349 10.1186/s12884-016-1146-y 27829393 18 Pires C C Rodrigues M L Capelli J C Santos M M Baião M R Atenção nutricional e práticas alimentares na perspectiva de gestantes com excesso de peso Demetra. 2020 15 e40566 10.12957/demetra.2020.40566 19 Ministério da Saúde Secretaria de Atenção à Saúde Departamento de Atenção Básica Guia alimentar para a população brasileira 2a ed. Brasília (DF) Ministério da Saúde 2014 20 Johnston A Kelly S E Hsieh S C Skidmore B Wells G A Systematic reviews of clinical practice guidelines: a methodological guide J Clin Epidemiol 2019 108 64 76 10.1016/j.jclinepi.2018.11.030 30529647 21 Grammatikopoulou M G Theodoridis X Gkiouras K Lampropoulou M Petalidou A Patelida M Methodological quality of clinical practice guidelines for nutrition and weight gain during pregnancy: a systematic review Nutr Rev 2020 78 07 546 562 10.1093/nutrit/nuz065 31755916 22 Louzada M LDC Tramontt C R de Jesus J GL Rauber F Hochberg J RB Santos T SS Developing a protocol based on the Brazilian Dietary Guidelines for individual dietary advice in the primary healthcare: theoretical and methodological bases Fam Med Community Health 2022 10 01 e001276 10.1136/fmch-2021-001276 35140130 23 Instituto Brasileiro de Geografia e Estatística Pesquisa de orçamentos familiares 2017-2018: análise do consumo alimentar pessoal no Brasil Rio de Janeiro IBGE 2020 24 Rubio D M Berg-Weger M Tebb S S Lee E S Rauch S Objectifying content validity: conducting a content validity study Soc Work Res 2003 27 02 94 104 10.1093/swr/27.2.94 25 Bardin L Análise de conteúdo 5a ed. Lisboa Edições 70 2008 26 Ministério da Saúde Caderneta da gestante [Internet] 3a ed. Brasília (DF) Ministério da Saúde 2016[cited 2022 Jan 10]. Available from:http://www.as.saude.ms.gov.br/wp-content/uploads/2016/07/Caderneta-da-Gestante-2016.pdf 27 Oliveira S C Lopes M V Fernandes A F Development and validation of an educational booklet for healthy eating during pregnancy. Rev Latino-Am Enfermagem 2014 22 04 611 20 10.1590/0104-1169.3313.2459 28 Oliveira F B Santos K E Barros N V Teixeira S A Cavalcante R M Elaboração e avaliação de material educativo sobre alimentação saudável para gestantes Rev Eletrônica Extensão. 2020 17 37 18 33 10.5007/1807-0221.2020v17n37p18 29 Kramer M S WITHDRAWN: Nutritional advice in pregnancy Cochrane Database Syst Rev 2007 04 CD000149 10.1002/14651858.CD000149 17636602 30 Martins A P Benicio M H Influência do consumo alimentar na gestação sobre a retenção de peso pós-parto Rev Saude Publica 2011 45 05 870 877 10.1590/S0034-89102011005000056 21829976 31 Crivellenti L C Zuccolotto D C Sartorelli D S Association between the Diet Quality Index Adapted for Pregnant Women (IQDAG) and excess maternal body weight Rev Bras Saúde Mater Infant 2019 19 02 285 294 10.1590/1806-93042019000200002 32 Lepre B Trigueiro H Johnsen J T Khalid A A Ball L Ray S Global architecture for the nutrition training of health professionals: a scoping review and blueprint for next steps. BMJ Nutr Prev Health 2022 e000354 10.1136/bmjnph-2021-000354 33 Vieira T W Sakamoto V TM Moraes L C Blatt C R Caregnato R CA Validation methods of nursing protocols: an integrative review Rev Bras Enferm 2020 73 05 e20200050 10.1590/0034-7167-2020-0050 33084808 34 Ministério da Saúde, Universidade de São Paulo Fascículo 3 protocolos de uso do guia alimentar para a população brasileira na orientação alimentar de gestantes Brasília (DF) Ministério da Saúde 2021
PMC009xxxxxx/PMC9951512.txt
==== Front BMC Neurosci BMC Neurosci BMC Neuroscience 1471-2202 BioMed Central London 36829110 783 10.1186/s12868-023-00783-7 Research Similar cortical morphometry trajectories from 5 to 9 years in children with perinatal HIV who started treatment before age 2 years and uninfected controls Nwosu Emmanuel C. [email protected] 1 Holmes Martha J. 12 Cotton Mark F. 3 Dobbels Els 3 Little Francesca 4 Laughton Barbara 3 van der Kouwe Andre 156 Robertson Frances 127 Meintjes Ernesta M. [email protected] 127 1 grid.7836.a 0000 0004 1937 1151 Biomedical Engineering Research Centre, Division of Biomedical Engineering, Department of Human Biology, Faculty of Health Sciences, University of Cape Town, Anzio Road, Observatory, Cape Town, 7925 South Africa 2 grid.7836.a 0000 0004 1937 1151 Neuroscience Institute, University of Cape Town, Cape Town, South Africa 3 grid.417371.7 0000 0004 0635 423X Department of Pediatrics & Child Health, Family Centre for Research With Ubuntu (FAMCRU), Tygerberg Hospital, Stellenbosch University, Cape Town, South Africa 4 grid.7836.a 0000 0004 1937 1151 Department of Statistical Sciences, University of Cape Town, Cape Town, South Africa 5 grid.32224.35 0000 0004 0386 9924 A.A. Martinos Centre for Biomedical Imaging, Massachusetts General Hospital, Boston, MA USA 6 grid.38142.3c 000000041936754X Department of Radiology, Harvard Medical School, Boston, MA USA 7 grid.7836.a 0000 0004 1937 1151 Cape Universities Body Imaging Centre, University of Cape Town, Cape Town, South Africa 24 2 2023 24 2 2023 2023 24 1525 10 2022 14 2 2023 © The Author(s) 2023, corrected publication 2023 https://creativecommons.org/licenses/by/4.0/ Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Background Life-long early ART (started before age 2 years), often with periods of treatment interruption, is now the standard of care in pediatric HIV infection. Although cross-sectional studies have investigated HIV-related differences in cortical morphology in the setting of early ART and ART interruption, the long-term impact on cortical developmental trajectories is unclear. This study compares the longitudinal trajectories of cortical thickness and folding (gyrification) from age 5 to 9 years in a subset of children perinatally infected with HIV (CPHIV) from the Children with HIV Early antiRetroviral therapy (CHER) trial to age-matched children without HIV infection. Methods 75 CHER participants in follow-up care at FAMCRU (Family Centre for Research with Ubuntu), as well as 66 age-matched controls, received magnetic resonance imaging (MRI) on a 3 T Siemens Allegra at ages 5, 7 and/or 9 years. MR images were processed, and cortical surfaces reconstructed using the FreeSurfer longitudinal processing stream. Vertex-wise linear mixed effects (LME) analyses were performed across the whole brain to compare the means and linear rates of change of cortical thickness and gyrification from 5 to 9 years between CPHIV and controls, as well as to examine effects of ART interruption. Results Children without HIV demonstrated generalized cortical thinning from 5 to 9 years, with the rate of thinning varying by region, as well as regional age-related gyrification increases. Overall, the means and developmental trajectories of cortical thickness and gyrification were similar in CPHIV. However, at an uncorrected p < 0.005, 6 regions were identified where the cortex of CPHIV was thicker than in uninfected children, namely bilateral insula, left supramarginal, lateral orbitofrontal and superior temporal, and right medial superior frontal regions. Planned ART interruption did not affect development of cortical morphometry. Conclusions Although our results suggest that normal development of cortical morphometry between the ages of 5 and 9 years is preserved in CPHIV who started ART early, these findings require further confirmation with longitudinal follow-up through the vulnerable adolescent period. Supplementary Information The online version contains supplementary material available at 10.1186/s12868-023-00783-7. Keywords Longitudinal Children with HIV CHER Cortical thickness Gyrification FreeSurfer Vertex-wise issue-copyright-statement© BioMed Central Ltd., part of Springer Nature 2023 ==== Body pmcBackground As of mid-2021, there were an estimated 1.3–2.1 million children with perinatal HIV infection (CPHIV) under 15 years of age living in sub-Saharan Africa (SSA) [1]. Although the number of new vertical infections is declining due to implementation of strategies for prevention of vertical transmission, there are still about 310,000 children (age ≤ 14 years) living with HIV in South Africa [1]. Early antiretroviral therapy (ART) initiation upon testing HIV seropositive is the current standard of care for CPHIV [2–4]. However, the consequences of long-term ART on the developing brain are not well understood. As such, there is a need for longitudinal studies of brain development in well-described cohorts of CPHIV in the context of early ART. A key study that led to the global adoption of early ART was the children with HIV early antiretroviral therapy (CHER) trial, which was conducted between 2005 and 2011 in Johannesburg and Cape Town, South Africa. The study aimed to investigate clinical outcomes of early (age ≤ 12 weeks) time-limited ART in asymptomatic infants with perinatal HIV (PHIV) compared to deferred, continuous ART. Clinical, neuropsychological and neuroimaging follow-up of the Cape Town CHER participants was performed to investigate the long-term effects of perinatal infection and early ART initiation on the central nervous system (CNS), which is a key viral reservoir, even in the combination ART (cART) era [5]. In this cohort, both neuropsychological testing [6–8] and cross-sectional neuroimaging investigations suggest small but persistent effects of HIV on the developing brain, despite viral suppression from a young age. In subcortical regions, our group has found HIV-related volume increases and decreases at different ages [9–11]. At 5- and 9 years HIV-related differences in metabolite levels in the basal ganglia (BG) were observed [12, 13], and at 11 years in the midfrontal gray matter (MFGM) and peritrigonal white matter (PWM) [14]. In terms of white matter development, at 5- and 7 years we found HIV-associated abnormalities in localized white matter integrity [15, 16]. Even though no HIV-related microstructural differences were evident in the corpus callosum at these ages, corpus callosum volumes were smaller in children with PHIV at age 5 [9]. Looking at cortical gray matter, at 5 years we found HIV-associated cortical thickness (CT) increases in bilateral frontal and left temporal regions, as well as reductions in local gyrification indices (LGI) in bilateral medial orbitofrontal regions extending into the rostral anterior cingulate and left superior frontal regions [11]. At 7 years, CPHIV demonstrated thicker cortex in a small left lateral occipital region and less gyrification in bilateral paracentral and right temporal regions [10]. ART interruption could exacerbate the ongoing impact of HIV on brain development in CPHIV. Several circumstances could lead to ART interruption including, intolerance of ART in children since pediatric formulation and pill sizes are not readily available, stock-outs and, more recently, the test for HIV remission which requires intensive monitored antiretroviral pause [17–20]. Lewis et al. [21] indicated that while ART interruption may have limited effect on children in the short term, its impact may be significant during critical periods of neural development. We previously identified region specific differences in cortical thickness and folding in relation to ART interruption at age 5 years in a sub-group of the CHER cohort using a cross-sectional approach but have not investigated further the long-term impact of interruption using a longitudinal approach [11]. While HIV and ART interruption-related alterations have been observed cross-sectionally at ages 5, 7 and/or 9 years, many of the alterations are not observed across multiple time points. Longitudinal analysis across these ages (5–9 years) may clarify the longer-term consequences of cross-sectional findings. Longitudinal data are informative because they allow the modeling of developmental changes as either physiologically normal or pathological in relation to various disease conditions [22], thereby enhancing understanding of developmental patterns in the presence of these conditions. Longitudinal neuroimaging studies are increasingly desirable in neuroimaging research due to their high sensitivity and potential to account for inter-subject variability [23]. For example, our recent longitudinal analysis of metabolite levels in children from the CHER trial revealed elevated choline levels indicative of neuroinflammation in CPHIV compared to uninfected controls from 5–11 years in all three regions studied, namely BG, MFGM and PWM [24]. Brain morphometric measures such as CT—the distance between the inner white/gray matter boundary and the outer gray/pial interface in the cerebral cortex [25–27] and LGI—a measure of cortical folding, which increases the number of neurons and cortical surface area within a limited skull space [22, 28, 29]—can provide valuable information on cortical anatomical development. Important micro-anatomical changes, including synaptic pruning, neuronal specialization, rewiring of fibre tracts, and structural folding and compacting occur during development from late childhood into adolescence, and determine the subsequent optimal function of the brain in adult life [22, 30–32]. Longitudinal investigation of these microstructural parameters in children living with HIV is key to understanding whether the developmental trajectory of cortical anatomy during a critical period of brain development [22, 30–32] differs from that of uninfected controls. The aim of this study was therefore to describe the longitudinal trajectory of morphometric development of the cerebral cortex between the ages of 5 and 9 years in the Cape Town arm of the CHER cohort, all of whom started ART before 2 years of age. We hypothesized that early-treated CPHIV would demonstrate altered morphometric development compared to uninfected controls in regions where we previously observed HIV-related differences on cross-sectional analyses. In addition, based on effects of ART interruption seen at age 5 years, we posited that changes in cortical thickness development would only be evident in CPHIV in whom ART had been interrupted, while LGI development would be altered similarly in all CPHIV (irrespective of whether treatment had been interrupted or not). Methods Study participants Study participants were 141 children (75 with PHIV; 72 male) who are being followed longitudinally at the Family Centre for Research with Ubuntu (FAMCRU) at Tygerberg Children’s Hospital in Cape Town, South Africa. The CPHIV were from the Cape Town cohort of the CHER clinical trial on which infants with CD4 percentage ≥ 25% were randomized at age 6 -12 weeks to receive either immediate limited ART—interrupted at 40 or 96 weeks—and restarted when clinical and/or immunological criteria were met, or to start ART only when they developed HIV symptoms or CD4 percentage dropped below 20% (25% in the first year) as per guidelines at the time [33]. Enrolled children received comprehensive immunological and clinical follow-up, including assessment for HIV-related encephalopathy. All CPHIV in the present sample started ART before 76 weeks of age. First line ART regimen consisted of Zidovudine (ZDV) + Lamivudine (3TC) + Lopinavir-Ritonavir (LPV/r, Kaletra®) [2, 34]; all CPHIV were on first line ART regimen until the end point of the clinical trial [2]. Socioeconomically- and sociodemographically matched children without HIV infection (HIV-) were recruited from an interlinking vaccine trial [35]. The uninfected control group comprised both children born to mothers with HIV infection (children who are HIV exposed uninfected, CHEU; n = 31) and to HIV- mothers (children who are HIV unexposed; CHU; n = 35). To increase power to detect HIV-related differences, we combined CHEU and CHU groups in our analyses. Posthoc analyses were performed in any regions showing HIV-related differences in morphometric development to confirm whether exposed and unexposed children differed in those regions. While CPHIV were motivated to remain in the longitudinal study as they received clinical care and neuropsychological assessments, there was limited incentive for children in the control group to remain in the study. This resulted in higher attrition rates among the control group. To combat this, additional controls were recruited during the study from similar neighbourhoods as the CHER children. As far as possible, newly recruited controls were matched with respect to exposure status to those lost to follow-up (LTFU). Eligibility criteria included documented HIV status of the mother and the child at birth, birthweight > 2000 g, being clinically healthy with normal medical history, no history of CNS insults or birth complications, nor a medical or psychiatric disorder that might affect cognitive performance [6, 36]. The current sample includes 24 additional controls (6 CHEU, 18 CHU) who were recruited at age 7 years. Image acquisition and analysis Study participants underwent brain MRI scanning on a Siemens 3 T Allegra scanner at 5, 7 and 9 years of age, using a protocol and procedure described previously [10]. Scans were performed without sedation according to protocols approved by the Faculty of Health Sciences Human Research Ethics Committees of the Universities of Cape Town and Stellenbosch. Parents/guardians of study participants provided written informed consent and the children gave oral assent at ages 5 and 7, and written assent at 9 years. To limit motion due to restlessness, children watched a movie via a mirror and rear projection screen during scanning. Furthermore, we used a 3D echo planar imaging (EPI) navigated [37] multiecho magnetization prepared rapid gradient echo (MEMPRAGE) [38] sequence (FOV 224 × 224 mm2, TR 2530 ms, TI 1160 ms, TEs = 1.53/3.19/4.86/6.53 ms, bandwidth 657 Hz/px, 144 slices, 1.3 × 1.0 × 1.0 mm3) that prospectively corrects for motion during the scan. MR images were visually inspected and those with poor visual quality were excluded from analyses. MR images were reviewed by a senior radiologist; if abnormalities were noted, children were referred for a clinical scan. T1-weighted high-resolution structural volumes that met visual quality control criteria were processed using the automated 3-stage longitudinal processing stream in FreeSurfer version 6.0 to extract reliable longitudinal surface CT and LGI measures (https://surfer.nmr.mgh.harvard.edu/fswiki/LongitudinalProcessing). The longitudinal processing stream first performs cross-sectional parcellation and cortical surface reconstruction for all subjects at all time points. The next stage of the pipeline re-samples the cross-sectional data of each subject to a base template. From the base template, longitudinal data are generated at all time points [23, 39, 40]. Outputs from the segmentation and parcellation steps were visually checked for errors and outlier values for CT/LGI were identified. Reconstructed data were then sampled to the FreeSurfer average subject template for vertex-wise analysis. For CT analyses, spatial smoothing was applied using a Gaussian kernel with 10 mm full-width, half-maximum (FWHM). No smoothing was used for LGI. Longitudinal vertex-wise analyses of cortical thickness and LGI Morphometric data were analysed using FreeSurfer’s spatiotemporal linear mixed effect model (LME) toolbox in MATLAB R2017a (https://www.mathworks.com/) [41]. LME models can handle longitudinal data such as ours that contains observations at different time points and varying numbers of time points for different subjects. To establish regional CT and LGI trajectories in typically developing children, we first created an LME model with HIV- children only. To investigate the influence of HIV in the context of early treatment on morphometric development across this age range, we performed an analysis on all participants using an LME model that additionally included both HIV status and an age by HIV status interaction term to identify possible HIV-related differences in developmental trajectories. Further, the rates of change in CT and LGI of CPHIV who had interrupted ART (Interrupted ART) and those who had continuous ART (Continuous ART) were separately compared to HIV- controls. To assess the possible effect of encephalopathy on the trajectory of cortical development in PHIV, we also compared children who had previously had a diagnosis of HIVE to HIV- controls. All analyses were done using a vertex-wise LME model, with group as a categorical fixed effect and random participant-specific intercepts. Rate of CT and LGI change with age in HIV- control children were modelled as follows:1 Yij=β1+β2∗tij+β3∗sexi+b1i+eij where Yij = cortical thickness or gyrification of subject i at time point j β1 = intercept tij = time variable—age in years of subject i at scan time point j sexi = 1 if subject i was female and 0 if male b1i = participant-specific intercept eij = error To examine potential non-linear effects of time, we repeated our analyses with an additional term for the square of time (tij 2) added in the model. For both CT and LGI, we present colour maps of the regression coefficients for age (i.e. β2 from Eq. 1) across the whole cortical surface. To allow for different means or slopes in PHIV (or Interrupted ART or Continuous ART or HIVE) and HIV- children, the following model was used:2 Yij=β1+β2∗tij+β3∗HIVi+β4∗HIVi∗tij+β5∗sexi+b1i+eij where HIVi = variable for HIV status of subject i (or HIVE status or Interrupted ART or Continuous ART). The null hypothesis of differences in either the means or rate of change of CT/LGI between PHIV (or Interrupted ART or Continuous ART or HIVE) and HIV- groups was tested. We also investigated potential non-linear effects of time by repeating the analyses with a term for the square of time (tij 2) added into the model. To correct for multiple comparisons we used a false discovery rate (FDR) corrected threshold of p < 0.05 for a two-tailed test. For both CT and LGI, we present colour maps of the relevant regression coefficients (i.e. β3 and β4 from Eq. 2 for group and age by group interaction effects, respectively) across the entire cortical surface. Results We present results for 141 participants who received MRI at least once from 5 to 9 years on the same scanner. Fifty-six children (23 PHIV) received a single scan, 63 children (44 PHIV) were scanned twice, and 22 children (8 PHIV) had 3 scans, for a total of 248 time points. Viral loads of 93.5% of the children in this study were suppressed (< 400 copies/mL) at the first scan around age 5 years. The 5-year data of one HIV- control child were excluded due to being incomplete, and the data at age 7 years of one CPHIV were excluded due to poor image quality. There were no failures in automated LGI computation, nor errors in cortical segmentation. There were no outliers in output values of CT and LGI for processed MR images. Table 1 shows biographical data of children without (HIV-) and with PHIV at ages 5, 7 and 9 years. Table 2 presents clinical data for the CPHIV.Table 1 Biographical data for all participants (N = 141; 75 PHIV, 66 HIV-) CPHIV HIV- t/χ2 p Age: 5 years (N = 82) Sample size (N) 52 30 Age at scan (years) 5.39 ± 0.31 5.60 ± 0.43 − 2.64 0.01 Number of males (% males) 21 (40%) 16 (53%) 1.37 0.24 Birth weight (g) 3125 ± 386 2973 ± 603 1.47 0.14 Estimated total intracranial volume (ETIV) (cm3) 1375 ± 96 1363 ± 110 0.53 0.60 Neuropsychological measures GMDSa All CPHIV IntART (n = 24) ConART (n = 28) All HIV- CHEU (n = 17) CHU (n = 13)  Performance (EQ) 74 ± 10 73 ± 11 76 ± 10 76 ± 18 73 ± 20 79 ± 17 − 0.55 0.58  Practical reasoning (FQ) 77 ± 8 76 ± 10 77 ± 7 76 ± 11 73 ± 12 79 ± 7 0.60 0.55  Sub-scales aggregate (GQ) 83 ± 6 83 ± 7 83 ± 6 83 ± 8 82 ± 10 84 ± 6 0.23 0.82 Age: 7 years (N = 125) Sample size (N) 71 54 Age at scan (years) 7.20 ± 0.13 7.24 ± 0.13 − 1.40 0.16 Number of males (% males) 36 (51%) 30 (56%) 0.13 0.72 Birth weight (g) 3081 ± 454 3081 ± 481 − 0.01 0.99 Estimated total intracranial volume (ETIV) (cm3) 1419 ± 101 1444 ± 137 − 1.17 0.24 Neuropsychological measures KABCb COMPOSITE SCORES All CPHIV IntART (n = 32) ConART (n = 39) All HIV- CHEU (n = 23) CHU (n = 31)  Mental processing index 72 ± 8 72 ± 8 73 ± 8 75 ± 10 74 ± 10 76 ± 11 − 1.90 0.06  Non-verbal index 74 ± 11 73 ± 11 75 ± 11 76 ± 12 73 ± 12 78 ± 11 − 0.96 0.34 Age: 9 years (N = 41) Sample size (N) 12 29 Age at scan (years) 9.13 ± 0.27 9.03 ± 0.56 0.58 0.57 Number of males (% males) 7 (58%) 16 (55%)  < 0.001 1.00 Birth weight (g) 3126 ± 479 3171 ± 466 − 0.28 0.78 Estimated total intracranial volume (ETIV) (cm3) 1422 ± 683 1437 ± 135 − 0.38 0.71 Neuropsychological measures KABCb COMPOSITE SCORES All CPHIV IntART (n = 8) ConART (n = 4) All HIV- CHEU (n = 17) CHU (n = 12)  Mental processing index 73 ± 10 73 ± 10 72 ± 10 79 ± 12 77 ± 12 81 ± 12 − 1.49 0.14  Non-verbal index 78 ± 10 78 ± 8 77 ± 15 80 ± 14 80 ± 14 80 ± 13 − 0.46 0.65 All values are mean ± standard deviation except where indicated otherwise; Bold indicates significance at p < 0.05 CPHIV children with perinatal HIV; HIV- children without HIV, IntART children with PHIV in whom ART was interrupted; ConART children with PHIV who received continuous ART; CHEU uninfected children who were exposed to HIV; CHU uninfected children unexposed to HIV aGriffiths’ Mental Development Scales—extended revised: 2–8 years bKaufman Assessment Battery for Children, second edition Table 2 Clinical data for all children with PHIV (N = 75) Interrupted ART Continuous ART Sample size (N) 39 36 Clinical data at baseline (age 6–8 weeks)  CD4 count (cells/mm3) 1827 ± 955 1739 ± 833  CD4% 33 ± 10 33 ± 10  CD4/CD8 ratio 1.4 ± 0.9 [0.2–4.4]a 1.2 ± 0.7 [0.2–3.4]a  CD8 count (cells/mm3) 1660 ± 1218 1686 ± 871  Viral load at enrolment   High (>750,000 copies/mL): n (%) 21 (53.8%) 23 (63.9%)   Low (400–750,000 copies/mL): n (%) 18 (46.2%) 13 (36.1%)   Suppressed (< 400 copies/mL): n (%) 0 (0%) 0 (0%) Other clinical / treatment measures  Age at ART initiation (weeks) 8.3 [7.4, 9.7] [6.6–12.0]a 18.9 [7.7, 27.2] [6.0–75.7]a  Age at ART interruption (weeks) 51.7 [47.3, 72.6] -  Age of first viral load suppression (weeks) 33.9 [31.5, 49.8] [30.6–213.3]a 47.6 [33.5, 79.2] [29.1–285.6]a  Duration of ART interruption (weeks) 34.4 [20.4, 52.3] [5.7–299.4]a 0.00 (not interrupted)  Nadir CD4% 19 ± 6 20 ± 6  Age at nadir CD4% (weeks) 91.0 [55.9, 125.7] 27.9 [18.8, 51.0]  CDC classification   A: n (%) 8 (20%) 1 (3%)  B: n (%) 9 (23%) 7 (19%)  Severe B: n (%) 2 (5%) 7 (19%)  C: n (%) 20 (51%) 19 (53%)  Unknown: n (%) 0 (0%) 2 (6%)  HIV encephalopathy diagnosis: n (%) 7 (18%) 5 (14%) Values are mean ± standard deviation or median [IQR] arange Groups were matched for age at each time point, except at age 5 when CPHIV were 2.5 months younger on average than HIV- children. Scores on the Griffiths’ Mental Development Scales (GMDS) Extended Revision for South Africa at age 5 and on the Kaufman Assessment Battery for Children (KABC) at age 9, did not differ between CPHIV and HIV- control groups. At age 7, scores for mental processing on the KABC tended to be lower, albeit below conventional levels of significance, in CPHIV than in controls. Nonetheless, KABC developmental trajectories from age 7 to 9 were similar between CPHIV and uninfected controls (Additional file 1: Table S1). To support combining PHIV and control sub-groups in our analyses, we also present means of GMDS and KABC scores for each sub-group in Table 1. Notably, scores on the GMDS and KABC did not differ between CPHIV on Interrupted and Continuous ART, nor between CHEU and CHU at ages 5, 7 or 9 years. The number of children diagnosed with HIVE was similar in the continuous and interrupted ART groups (Table 2). Rate of change in CT and LGI in HIV- control children Thickness Among uninfected control children we observe widespread cortical thinning from age 5 to 9 years. This decrease in thickness was significant at an FDR-corrected p-value of 0.05 (β2 <–0.09 mm/year) over much of the lateral surface of the cortex, as well as in medial superior parietal, occipital and frontal regions. The decrease was greatest in postcentral/supramarginal, lateral frontal and temporal regions (0.11 mm/year), followed by bilateral precentral and medial occipital regions (0.07 mm/year) (Fig. 1). Adding the square of time (tij 2) to the model did not change the results.Fig. 1 Colour map of regression coefficients showing rate of change from age 5 to 9 years of (left) cortical thickness and (right) local gyrification indices in HIV-uninfected (HIV-) children controlling for sex. Positive regression coefficients (red/yellow) indicate increases with age, while negative coefficients (cyan/blue) indicate decreases with age. The colour bar scale in each figure applies to both lateral (top) and medial (bottom) views. (N = 66 HIV- controls) Local gyrification indices Children without HIV demonstrate increasing gyral folding from 5 to 9 years at an FDR-corrected p of 0.05 (β2 > 0.03) in large bilateral lateral frontal and right temporal regions, as well as right medial inferior parieto-occipital regions. The greatest changes (β2 in Eq. (1) = 0.047 units/year, FDR p < 0.001) were observed in bilateral rostral middle frontal and right superior temporal regions, and in parts of the right postcentral gyrus and parieto-occipital sulcus areas (Fig. 1). Adding the square of time (tij 2) to the model did not change the findings. Group comparison of means and rate of change of CT After correction for multiple comparisons, vertex-wise analyses did not reveal any regions where the mean CT (β3 in Eq. (2)), nor the rate of change of CT from age 5 to 9 years [β4 in Eq. (2)], differed significantly between CPHIV and their uninfected peers (Fig. 2A and C). However, small regions in the left supramarginal gyrus, insula, lateral occipital gyri, and inferior temporal gyrus (indicated by arrows in Fig. 2A) demonstrated thicker cortex in CPHIV that fell just short of significance at a strict FDR corrected threshold. To reduce the risk of type 2 errors, we identified clusters showing differences (in means or rate of change of CT) between CPHIV and HIV- controls at an uncorrected p < 0.005 and extracted mean CT values in these clusters for each subject at each age for which the subject provided data. At this threshold, we identified 6 regions where the cortex was thicker from age 5–9 years in CPHIV than controls (numbered arrows in Fig. 2B), but none where the rates of change of CT differed between groups. In Table 3 we present the MNI coordinates and sizes of the 6 clusters where CPHIV demonstrate thicker cortex, as well as regression coefficients from a post-hoc LME analysis. Plots of cortical thickness as a function of age are presented for each of these clusters in Fig. 3.Fig. 2 Colour map of the vertex-wise regression coefficients of the effect on cortical thickness of A, B HIV status [β3 in Eq. (2)] and C age by HIV status interactions [β4 in Eq. (2)], respectively. In A and B, positive regression coefficients (red/yellow) indicate thicker cortex in CPHIV compared to uninfected controls and negative coefficients (cyan/blue) thinner cortex. In C, positive regression coefficients (red/yellow) indicate a greater rate of change of cortical thickness from age 5 to 9 years in CPHIV than uninfected controls, and negative coefficients (cyan/blue) indicate a slower rate of change. The colour bar scale in each figure indicates the effect size (Cohen’s d). While the scales in A and C extend to effect sizes required for significance at FDR-corrected p-values, the scale in B has been right censored to highlight clusters showing group differences that survive at uncorrected p < 0.005. For HIV status, effect sizes ≥ 0.06 and ≥ 0.29 are significant at uncorrected and FDR corrected p = 0.05, respectively. For age by HIV status interactions, effect sizes ≥ 0.09 and ≥ 0.31 are significant at uncorrected and FDR corrected p = 0.05, respectively. (N = 141: PHIV = 75, HIV − = 66) Table 3 Regions where children with PHIV have thicker cortex than HIV- controls from age 5 to 9 years at an uncorrected p < 0.005, and regression coefficients from a post-hoc LME analysis performed on regional means of cortical thickness extracted in these clusters for each child at each age for which the child provided data. Numbering of regression coefficients is as per Eq. (2) Cluster region MNI coordinates (x,y,z) Size (mm2) HIV status effects Age effects HIV status by age interactions Sex effects β3 (SE) p-value β2 (SE) p-value β4 (SE) p-value β5 (SE); p-value L Supramarginal − 59.0, − 46.4, 21.6 245 0.38 (0.11) < 0.001 − 0.05 (0.01) < 0.001 0.02 (0.01) 0.11 − 0.07 (0.06) 0.21 L Insula − 32.9, 15.0, − 1.3 124 0.38 (0.16) < 0.001 − 0.06 (0.02) < 0.001 0.02 (0.02) 0.31 − 0.13 (0.07) 0.05 L Lateral orbitofrontal − 18.0, 24.6, − 18.9 120 0.20 (0.10) < 0.001 − 0.05 (0.01) < 0.001 0.01 (0.01) 0.63 0.06 (0.04) 0.17 L Banks of Superior Temporal Sulcus − 53.0, − 41.8, 8.6 77 0.33 (0.16) < 0.001 − 0.06 (0.01) < 0.001 0.02 (0.02) 0.46 − 0.06 (0.07) 0.34 R Medial Superior frontal 12.9, 25.3, 29.5 504 0.19 (0.08)  < 0.001 − 0.02 (0.01)  < 0.001 − 0.0002 (0.01) 0.98 − 0.05 (0.04) 0.19 R Insula 37.2, − 3.8, − 0.4 314 0.20 (0.12)  < 0.001 − 0.03 (0.01)  < 0.001 − 0.03 (0.01) 0.04 0.01 (0.05) 0.79 L Left; R Right; β unstandardised regression coefficients from LME model (Eq. 2); SE standard error; p-values are uncorrected Fig. 3 Developmental trajectories of cortical thickness from age 5 to 9 years in regions where the cortex of children with PHIV was thicker than in uninfected controls at an uncorrected p < 0.005. The shadows represent the 95% confidence interval (CI). (N = 141: PHIV = 75, HIV − = 66) To ensure that our results are not biased by the absence of within-subject information for participants who provided data at only a single time point, we repeated the post-hoc LME analysis using regional CT data from only children who provided data at multiple time points (52 CPHIV, 33 HIV-). Although effect sizes were smaller than from analyses including all children, cortical thickening in CPHIV compared to controls remained evident in all the regions except the left banks of the superior temporal sulcus, and below conventional levels of significance in left lateral orbitofrontal cortex. Age, group by age interactions, and sex effects were similar to before (Additional file 1: Table S2). We also repeated our post-hoc LME analyses to examine whether CT developmental trajectories differed within the HIV- or PHIV sub-groups. Notably, cortical thickness development did not differ in any of the regions between exposed and unexposed uninfected control children (all p’s > 0.26, Additional file 1: Table S3), nor between CPHIV in whom treatment had been interrupted compared to those on continuous treatment (all p’s > 0.08, Additional file 1: Table S4). We did not find any regions showing differences in the means nor rate of change of CT between either the Interrupted ART, Continuous ART, or HIVE groups and the HIV- control group at either FDR corrected p < 0.05 or uncorrected p < 0.005 thresholds. Adding the square of time (tij 2) to the LME model did not change the findings. Group comparison of means and rate of change of LGI After correction for multiple comparisons, vertex-wise analyses did not reveal any regions where the means (β3 in Eq. (2)), nor the rate of change from age 5 to 9 years (β4 in Eq. (2)), of LGI (Fig. 4A and B) differed significantly between CPHIV and their uninfected peers, nor between either the Interrupted- or Continuous ART groups and the HIV- control group. We also did not find any differences at an uncorrected p < 0.005. Adding the square of time (tij 2) to the model did not change the findings.Fig. 4 Colour map of the vertex-wise regression coefficients of the effect on local gyrification indices of A HIV status (β3 in Eq. (2)) and B HIV status by age interactions (β4 in Eq. (2)), respectively. In A, positive regression coefficients (red/yellow) indicate greater gyrification in CPHIV compared to uninfected controls and negative coefficients (cyan/blue) less gyrification. In B, positive regression coefficients (red/yellow) indicate a greater rate of change from age 5 to 9 years in CPHIV than uninfected controls, and negative coefficients (cyan/blue) indicate a slower rate of change. The colour bar scale in each figure indicates the effect size (Cohen’s d). For HIV status and status by age interactions, respectively, effect sizes ≥ 0.30 and ≥ 3.77 are required for significance at an FDR corrected p = 0.05. (N = 141: PHIV = 75, HIV- = 66) We did, however, find greater mean (β3 (SE) = 0.34 (0.12), FDR p = 0.04) and slower rate of change (β4 (SE) = − 0.03 (0.05), FDR p = 0.01) of gyrification in children who had previously had a diagnosis of HIVE compared to HIV- controls in a left rostral middle frontal region (MNI co-ordinates at peak: − 38.6, 45.9, 1.5; cluster size: 1702.56 mm2; Additional file 1: Figure S1). Discussion To the best of our knowledge, this study is the first longitudinal neuroimaging investigation of cortical development in CPHIV from a resource-limited setting and who received ART before age 2 years. Similar to previous studies, typically-developing control children showed widespread decreases in cortical thickness from age 5 to 9 years [22, 42–44]. Gyrification showed less change during this period, but increases were observed in rostral middle frontal, temporal, postcentral and parieto-occipital regions. Although the developmental trajectories of cortical thickness and gyrification in CPHIV were similar to that of matched typically-developing HIV- children during this important period of brain development, we identified 6 regions, albeit at a lower significance level, where the cortex of CPHIV was thicker than in uninfected children, namely bilateral insula, left supramarginal, lateral orbitofrontal and superior temporal, and right medial superior frontal regions. Morphometric development in typically-developing HIV- children from age 5–9 years Cortical thinning from childhood into early adolescence is associated with physiological processes related to neurodevelopment and regional specialization of the developing brain, and faster thinning of cortex has been associated with higher intelligence quotients (IQ) around age 10 years [45]. These physiological processes include myelinogenesis [22, 42, 46], synaptic pruning and dendritic arborisation [26, 42, 46, 47], which lead to reduction in the size and number of neurons and their synaptic processes [26, 42, 46]. These neurodevelopmental processes work together to reinforce and strengthen fibres and connections that are used consistently for transmitting brain impulses, while redundant fibres are “pruned” or removed [48]. The process of myelination in the lower cortical layer closest to the cortical white/grey matter interface may be the main process that influences cortical thinning with brain growth [22, 42, 49]. We also observed increased gyrification with age in bilateral frontal regions, which is consistent with findings by Schnack et al. [45] of increasing cortical surface area with age. Although, the developmental trajectory of cortical folding is not yet fully understood, the frontal cortex controls the complex executive skills of planning, working memory and cognitive flexibility that develops in late childhood [22, 50, 51], corresponding to the period studied here. Maturation of cortical morphology follows a back-to-front—occipital-to-prefrontal—[48] developmental sequence, with prefrontal cortex maturing last in the continuous process of development from childhood into adolescence. The regional variation in rate of change and development in brain morphometry may be related to the complexity associated with the cytoarchitecture, information processing and co-ordination functionalities of different cortical regions [22]. In terms of functional complexity, primary sensory areas change and develop faster than the more complex polymodal, higher-order association areas [22, 48, 52]. The increases in cortical folding we observed in this study may be functionally related to the acquiring of critical developmental skills in late childhood into adolescence [53, 54]. Our finding of increasing gyrification in rostral middle frontal regions through age 9 years is similar to that of Cao et al. [28], who reported that although regional gyrification decreases from medial to lateral regions starting as early as age 4, frontal gyral complexity continues to increase, possibly into the teenage years [28, 55]. Effect of perinatal HIV infection and early ART initiation on longitudinal morphometry development The main focus of this study was to examine whether PHIV in the context of early ART initiation affects brain development across childhood. Previous cross-sectional neuroimaging studies comparing brain developmental indices of children and youth with PHIV to healthy controls found associations of HIV infection with both regionally thicker or thinner cortex [56–59], as well as regionally-specific lower gyrification [56, 59]. In our own cross-sectional analyses of the cohort studied here, we previously reported thicker cortex at age 5 years in CPHIV compared to uninfected controls in bilateral frontal and left temporo-insular regions, and lower LGIs in left superior medial frontal and bilateral medial orbitofrontal cortices extending into rostral anterior cingulate [11]. In contrast, at age 7 years, we found thicker cortex only in a small left lateral occipital region, and less gyrification in bilateral paracentral and right temporal regions [10]. These regions seen in our cohort at ages 5 and 7 years are similar to the regions implicated in previous studies of CPHIV [56–59]. Few longitudinal studies exist of effects of HIV infection on neurodevelopment, especially in children initiated to ART early (< 2 years) who have been virally suppressed from a young age. We identified 3 longitudinal studies [57, 60, 61] that investigated brain development in children and adolescents living with HIV. Although all 3 demonstrated regionally altered brain morphometry in CPHIV, only one showed an effect of HIV on the trajectory of morphometric development [57]. Van den Hof et al. [60] reported a slower rate of normal grey matter volume decline over a period of approximately 5 years in children living with HIV aged 8—18 years, but this was attributable to variation in participant height. Although white matter volume was lower in CPHIV the developmental trajectory was no different from controls and these changes were not associated with cognitive development. Wade et al. [61] identified a subcortical pallidal region in the basal ganglia that was thinner in CPHIV with a mean age of 11 years. However, they found no effect of HIV on longitudinal shape or volume changes of the pallidum over a period of one year. By contrast, in addition to finding areas of lower GM volume and altered CT in CPHIV, Yu et al. [57] showed HIV status by age interaction effects in 13-year-old children after 1 year on GM volume, especially in bilateral parietal lobes, and on CT in the bilateral central sulcus, insula, frontal lobe and cingulate sulcus. Overall, our findings of thicker cortex in CPHIV in bilateral insula, right medial frontal, and left supramarginal, lateral orbitofrontal and superior temporal regions (at a lower uncorrected level of significance), but similar developmental trajectories of cortical thickness and gyrification between CPHIV and HIV- children are consistent with previous studies suggesting that HIV effects on brain morphology are subtle. The slightly relaxed level of significance was to address concerns regarding insufficient power and to reduce the risk of type 2 error. Unexpectedly, only the insular clusters in this study overlap with regions where cortex was thicker at either ages 5 or 7 years in our previous studies. The fact that brain regions implicated at either ages 7 or 9 years in our cohort do not overlap, together with the absence of HIV-related differences (at conventional levels of significance) in cortical development from 5–9 years in our current analyses, suggest that early ART initiation combined with ongoing treatment may prevent persistent HIV-associated damage to the cortex in young children. It is possible, therefore, that previously reported HIV-related alterations in CT and LGI at 5 and 7 years represent transient localized maturation differences that may be a side effect of long-term HIV infection and ongoing treatment. However, to confirm that the absence of HIV-related differences is not merely due to a lack of power, cortical development should be studied through adolescence, which will also increase power to detect long-term HIV-related differences. Notably, evidence of persistent cognitive deficits remains. At age 5 years, Beery visual perception scores were lower in CPHIV [6], and at both ages 7 and 9 years CPHIV performed worse than uninfected controls on tasks assessing executive function and auditory working memory [36]. Previously, in the same cohort studied here, we reported elevated levels of choline (glycerophosphocholine + phosphocholine) from 5 to 11 years in CPHIV compared to uninfected controls in all three regions where MR spectroscopy data were acquired, namely MFGM, BG and PWM, as well as higher myo-inositol in the MFGM [24]. Notably, the right medial superior frontal region where we find thicker cortex from age 5–9 years in CPHIV overlaps the MFGM region-of-interest (ROI) from our MRS study, and the supramarginal gyrus is adjacent to the PWM—although we find thicker cortex in the left supramarginal gyrus while the PWM ROI where MRS data were acquired was on the right. The elevated choline and myo-inositol levels in these regions suggest that the thicker cortex observed here in CPHIV may be related to persistent neuroinflammation and astrocytosis, despite early ART. Effect of ART interruption on morphometry development Since children with PHIV start a lifetime of ART at birth, there is much interest in the possibility of safe treatment interruption(s). One of the goals of this follow up study was to assess the long-term implications of planned interruption compared to continuous ART. Cross-sectional comparison at age 5 years showed that children in whom ART had been interrupted had thicker cortex than HIV- controls in the left rostral and superior frontal, and right insula regions, and lower gyrification in the left precuneus and right rostral and caudal anterior cingulate regions, but higher gyrification in the right lateral occipital regions [11]. Children on continuous ART demonstrated lower gyrification compared to HIV- controls in right posterior superior frontal and superior parietal regions, but higher gyrification in left fusiform, and right rostral middle frontal and lateral occipital regions [11]. However, longitudinal analysis did not reveal any regions where CT nor LGI, nor their developmental trajectories from 5 to 9 years, differed between either CPHIV on interrupted or continuous treatment and HIV- controls. According to previous studies, immune compromise related to a planned short ART interruption has a similar recovery pattern as continuous ART [21, 62]. Research has shown infants’ immune systems to be dynamic and malleable, allowing them to recover from short periods of ART interruption [19, 21]. Several studies demonstrate that the most important factors in perinatal HIV treatment is ART initiation timing and duration on treatment before interruption [20, 34, 62]. Early ART initiation is beneficial for improved immune system and viral suppression, as well as improving neurodevelopmental outcomes [2, 8, 34]. These results suggest that when ART is initiated before critical developmental processes are affected by viral replication, a short interruption may not affect long-term immune health. In contrast, a longer interruption may lead to viral rebound, resistance and neuropsychological consequences [19, 20, 63]. The range of interruption durations for the children in the cohort under study is quite large (6–300 weeks). Although ART interruption may lead to age specific alterations in brain morphometry development [11], the absence of HIV-related differences in the means and developmental trajectories of CT and LGI suggest that differences do not persist to later ages. Effect of HIVE on morphometry development With the introduction of early ART, there has been a decline in the incidence of HIVE [64, 65]. In our cohort of early-treated CPHIV, only 16% had been diagnosed with HIV-related encephalopathy, compared to incidences of 20–60% in previous studies of perinatally infected but ART-naïve pediatric cohorts [64, 66, 67]. We did not find any differences in the developmental trajectories of cortical thickness or gyrification between the HIVE group and HIV- controls, except in the left rostral middle frontal region where children with HIVE demonstrated greater mean gyrification than HIV- controls and failed to show the age-related increase in gyrification seen in controls. Although the small number of HIVE subjects raises concerns regarding the reliability of this finding, the low number of HIVE children in our cohort and the limited number of regions found to show altered developmental trajectories, reinforce the benefits of early ART in reducing the occurrence and severity of neurologic conditions such as HIVE. Since brain development moves from posterior (occipital lobe) to anterior (frontal lobe), the frontal lobe of the brain is expected to develop last [49]. The reduced age-related changes in gyrification in the rostral middle frontal region in HIVE children may therefore point to developmental delay. However, it may simply be a consequence of the higher gyrification evident in HIVE children in this region at age 5 years. Limitations and future work The small sample size at age 9 years, with fewer CPHIV compared to HIV- controls, was due to the decommissioning of the 3 T Siemens Allegra brain scanner used for this study. The small number of children diagnosed with HIVE precludes strong conclusions about group differences. Although FreeSurfer’s longitudinal pipeline was not originally designed for pediatric cortical segmentation, manual quality checks were performed on the cortical model to ensure accurate extraction and segmentation of cortical surfaces. A key theory of cortical folding formation links gyral and sulcal formation in the brain to neural connectivity [52]. It postulates that regions with greater neural connectivity are associated with greater tension, ensuring that such brain regions are in close proximity during brain growth, forming gyri, while regions with lower connectivity follow a similar pattern to form sulci [47, 52]. Functional connections are associated with acquiring new skills and abilities during neurodevelopment, leading to changes in surface morphology of the cortex. Hence it may be worthwhile for future longitudinal studies to investigate gyrification development in relation to neural connections using DTI and functional connectivity measures. The relationship between cortical developmental trajectories and longitudinal neuropsychological outcomes should also be examined. Conclusion We present results from a follow-up neuroimaging study of a subset of children from the CHER trial during an important stage of neurodevelopment (5–9 years). In our control group, generalized cortical thinning was observed from age 5 to 9 years, with the rate of thinning varying by region. Age-related increases in gyrification were observed in large bilateral frontal, right temporal, and right medial inferior parieto-occipital regions. After correction for multiple comparisons, we did not find any regions where the means or developmental trajectories of cortical thickness and gyrification differed between CPHIV who initiated ART early and uninfected controls. However, at an uncorrected p < 0.005, we found 6 regions where CPHIV had thicker cortex than uninfected controls. Planned ART interruption did not affect development of cortical morphometry. HIV-related encephalopathy was associated with greater gyrification and the absence of age-related gyrification increases in bilateral rostral middle frontal regions. Although our results suggest that early ART initiation preserves normal development of cortical morphometry between the ages of 5 and 9 years in perinatal HIV infection, these findings need to be confirmed with longitudinal follow-up through the vulnerable adolescent period. Supplementary Information Additional file 1: Table S1. Regression coefficients from a linear mixed effect analysis examining effects of HIV status on KABC developmental trajectories from age 7 to 9 years. Table S2. Regression coefficients from a post-hoc LME analysis of CT trajectories in regions where CPHIV showed cortical thickening using only children who provided data at more than one time point (N=85; 52 PHIV). Numbering of regression coefficients is as per equation (2). Table S3. Regression coefficients from a post-hoc LME analysis comparing CT developmental trajectories between CHEU and CHU in the 6 regions where children with PHIV demonstrated thicker cortex from age 5 to 9 years than HIV- controls. Numbering of regression coefficients is as per equation (2). Table S4. Regression coefficients from a post-hoc LME analysis comparing CT developmental trajectories between CPHIV in whom treatment was interrupted and those on continuous ART in the 6 regions where children with PHIV demonstrated thicker cortex from age 5 to 9 years than HIV- controls. Numbering of regression coefficients is as per equation (2). Figure S1. Effects of HIV status (HIVE vs HIV-) by age interactions on local gyrification indices (LGIs). (A) Colour map of effects sizes. Positive regression coefficients (red/yellow) indicate greater rates of change of LGI in children who had previously been diagnosed with HIVE than HIV-, and negative coefficients (cyan/blue) indicate lower rates of change in HIVE than HIV- children. The colour bar scale applies to both lateral (top) and medial (bottom) views. Children with a previous diagnosis of HIVE did not demonstrate the increasing gyrification from ages 5 to 9 years evident in HIV- controls in bilateral rostral middle frontal regions. In the left hemisphere in the region outlined in red (white arrow), the difference is significant at an FDR-corrected p<0.05. (B) Plots showing the rate of change of gyrification in the left rostral middle frontal regions from ages 5 and 9 across HIVE and HIV- groups (left) and a spaghetti plot of individual trajectories (right). (HIV- = 66, HIVE = 12). Abbreviations ART Antiretrovial therapy BG Basal ganglia cART Combination ART CD4 Cluster of differentiation 4 (Glycoprotein in immune cells) CHER Children with HIV Early antiRetroviral therapy CHEU Children who are HIV exposed and uninfected CHU Children who are HIV unexposed CNS Central nervous system CPHIV Children perinatally infected with HIV CT Cortical thickness DTI Diffusion tensor imaging EPI Echo planar imaging FAMCRU FAMily Centre for Research with Ubuntu, Tygerberg Children’s Hospital Cape Town, South Africa FDR False discovery rate FOV Field of view FWHM Full-width, half-maximum HIV Human immunodeficiency virus HIVE HIV-related encephalopathy LGI Local gyrification index LME Linear mixed effect model LPR/r Lopinavir-Ritonavir MATLAB MATrix LABoratory (programming language) MEMPRAGE MultiEcho Magnetization Prepared RApid Gradient Echo MFGM MidFrontal gray matter MRI Magnetic resonance imaging MRS Magnetic resonance spectroscopy PHIV Perinatal HIV PWM Peritrigonal white matter ROI Region of interest SD Standard deviation SSA Sub-Saharan Africa TR Repetition time TI Inversion time TE Time-to-echo ZDV Zidovudine 3D 3-dimensional 3 T 3 Tesla MRI 3TC Lamivudine Acknowledgements We wish to acknowledge the contribution of the following to this study: Staff of the Family Centre for Research with Ubuntu (FAMCRU), Department of Pediatrics & Child Health, Tygerberg Children’s Hospital, Cape Town, South Africa; Staff of the Cape Universities Brain Imaging Centre (CUBIC), Cape Town, South Africa; the children who participated in this study and their caregivers. We acknowledge that this study is part of a doctoral thesis submitted at the University of Cape Town by Emmanuel Nwosu Author contributions EMM, BL, AVK conceived the study design. EN conducted the analyses and drafted the manuscript for the study under supervision and guidance of FR and EMM. BL, ED and MC gave access to and provided some of the primary data for this study and provided critical review and suggestions on the draft manuscript. AVK provided technical support in designing tool for neuroimaging analyses and FL provided support with statistical analyses. MJH provided critical review and feedback on the manuscript. All authors were involved in writing and reviewing the manuscript and approved the final manuscript. All authors read and approved the final manuscript Funding Funding for this study was provided by NIH Grants R01HD099846, R01DC015984, R01HD071664, R21MH108346 and R21MH096559; South African National Research Foundation (NRF) Grants 48337, 99069 and 78737; UCT VC Interim Funding; US National Institute of Allergy and Infectious Diseases (NIAID) through the CIPRA network, Grant number U19AI53217; and the South African Medical Research Council (MRC). The Departments of Health of the Western Cape and Gauteng, South Africa and ViiV Healthcare/GlaxoSmithKline plc provided additional support for the CHER clinical trial; University of Cape Town, South Africa provided seed fund in the form of University of Cape Town International student scholarship.Except for antiretroviral therapy supplied by GSK/Viiv for participants on the initial CHER trial, from which this study follows. All funders were not involved in the study design, collection, analysis, interpretation of data, the writing of this article or the decision to submit it for publication. Availability of data and materials The datasets used and/or analysed during the current study are available from the corresponding authors on reasonable request. Declarations Ethics approval and consent to participate This study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving research study participants were outlined in the study protocols approved by the Faculty of Health Sciences Human Research Ethics Committees of the Universities of Cape Town and Stellenbosch. Parents/guardians of study participants provided written informed consent and the children gave oral assent at ages 5 and 7, and written assent at 9 years. Consent for publication Not Applicable. Competing interests The authors declare that they have no competing interests. The original version of this article was revised: There was some errors in tables 1, 2 and 3. This has been corrected. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Change history 7/12/2023 A Correction to this paper has been published: 10.1186/s12868-023-00802-7 ==== Refs References 1. UNAIDS. Global HIV & AIDS statistics—Fact sheet. 2022. Source: 2022. https://www.unaids.org/en/resources/fact-sheet. Accessed Aug 24. 2. Violari A Cotton MF Gibb DM Babiker AG Steyn J Madhi SA Early antiretroviral therapy and mortality among HIV-infected infants N Engl J Med 2008 359 21 2233 2244 10.1056/NEJMoa0800971 19020325 3. Lindsey JC Malee KM Brouwers P Hughes MD Neurodevelopmental functioning in HIV-infected infants and young children before and after the introduction of protease inhibitor–based highly active antiretroviral therapy Pediatrics 2007 119 3 e681 e693 10.1542/peds.2006-1145 17296781 4. World Health Organization (WHO) Report of the WHO technical reference group. Paediatric HIV/ART care guideline group meeting 2008 Geneva WHO Headquarter 5. Van Rie A Harrington PR Dow A Robertson K Neurologic and neurodevelopmental manifestations of pediatric HIV/AIDS: a global perspective Eur J Paediatr Neurol 2007 11 1 1 9 10.1016/j.ejpn.2006.10.006 17137813 6. Laughton B Cornell M Kidd M Springer PE Dobbels EFMT Rensburg AJV Five year neurodevelopment outcomes of perinatally HIV-infected children on early limited or deferred continuous antiretroviral therapy J Int AIDS Soc 2018 21 5 e25106 10.1002/jia2.25106 29722482 7. Laughton B Cornell M Boivin M Van Rie A Neurodevelopment in perinatally HIV-infected children: a concern for adolescence J Int AIDS Soc 2013 16 1 18603 10.7448/IAS.16.1.18603 23782482 8. Laughton B Cornell M Grove D Kidd M Springer PE Dobbels E Early antiretroviral therapy improves neurodevelopmental outcomes in infants AIDS (London, England) 2012 26 13 1685 10.1097/QAD.0b013e328355d0ce 22614886 9. Randall SR Warton CM Holmes MJ Cotton MF Laughton B van der Kouwe AJ Larger subcortical gray matter structures and smaller corpora callosa at age 5 years in HIV infected children on early ART Front Neuroanat 2017 11 95 10.3389/fnana.2017.00095 29163068 10. Nwosu EC Robertson FC Holmes MJ Cotton MF Dobbels E Little F Altered brain morphometry in 7-year old HIV-infected children on early ART Metab Brain Dis 2018 33 2 523 535 10.1007/s11011-017-0162-6 29209922 11. Nwosu EC Holmes MJ Cotton MF Dobbels E Little F Laughton B Cortical structural changes related to early antiretroviral therapy (ART) interruption in perinatally HIV-infected children at 5 years of age IBRO Neuroscience Reports 2021 10 161 170 10.1016/j.ibneur.2021.02.001 34179869 12. Mbugua KK Holmes MJ Cotton MF Ratai EM Little F Hess AT HIV-associated CD4/8 depletion in infancy is associated with neurometabolic reductions in the basal ganglia at age 5 years despite early antiretroviral therapy AIDS 2016 30 9 1353 10.1097/QAD.0000000000001082 26959509 13. Robertson FC Holmes MJ Cotton MF Dobbels E Little F Laughton B Perinatal HIV infection or exposure is associated with low N-acetylaspartate and glutamate in Basal Ganglia at age 9 but not 7 Years Front Hum Neurosci 2018 12 145 10.3389/fnhum.2018.00145 29867401 14. Graham AS Holmes MJ Little F Dobbels E Cotton MF Laughton B MRS suggests multi-regional inflammation and white matter axonal damage at 11 years following perinatal HIV infection Neuro Image Clin 2020 28 102505 15. Ackermann C Andronikou S Saleh MG Laughton B Alhamud AA van der Kouwe A Early antiretroviral therapy in HIV-infected children is associated with diffuse white matter structural abnormality and corpus callosum sparing Am J Neuroradiol 2016 37 12 2363 2369 10.3174/ajnr.A4921 27538904 16. Jankiewicz M Holmes MJ Taylor PA Cotton MF Laughton B van der Kouwe AJ White matter abnormalities in children with HIV infection and exposure Front Neuroanat 2017 11 1662 5129 10.3389/fnana.2017.00088 17. Li JZ Smith DM Mellors JW The critical roles of treatment interruption studies and biomarker identification in the search for an HIV cure AIDS 2015 29 12 1429 10.1097/QAD.0000000000000658 25870989 18. Dubrocq G Rakhmanina N Antiretroviral therapy interruptions: impact on HIV treatment and transmission HIV/AIDS 2018 10 91 19. Wamalwa D Benki-Nugent S Langat A Tapia K Ngugi E Moraa H Treatment interruption after 2-year antiretroviral treatment initiated during acute/early HIV in infancy AIDS 2016 30 15 2303 2313 10.1097/QAD.0000000000001158 27177316 20. Ananworanich J Melvin D Amador JT Childs T Medin G Boscolo V Neurocognition and quality of life after reinitiating antiretroviral therapy in children randomized to planned treatment interruption AIDS 2016 30 7 1075 1081 10.1097/QAD.0000000000001011 26730569 21. Lewis J Payne H Walker AS Otwombe K Gibb DM Babiker AG Thymic output and CD4 T-cell reconstitution in HIV-infected children on early and interrupted antiretroviral treatment: evidence from the children with HIV early antiretroviral therapy trial Front Immunol 2017 8 1162 10.3389/fimmu.2017.01162 28979264 22. Shaw P Kabani NJ Lerch JP Eckstrand K Lenroot R Gogtay N Neurodevelopmental trajectories of the human cerebral cortex J Neurosci 2008 28 14 3586 3594 10.1523/JNEUROSCI.5309-07.2008 18385317 23. Reuter M Schmansky NJ Rosas HD Fischl B Within-subject template estimation for unbiased longitudinal image analysis Neuroimage 2012 61 4 1402 1418 10.1016/j.neuroimage.2012.02.084 22430496 24. Van Biljon N Robertson F Holmes M Cotton MF Laughton B van der Kouwe A Multivariate approach for longitudinal analysis of brain metabolite levels from ages 5–11 years in children with perinatal HIV infection Neuroimage 2021 237 118101 10.1016/j.neuroimage.2021.118101 33961998 25. Moeskops P Benders MJ Kersbergen KJ Groenendaal F de Vries LS Viergever MA Development of cortical morphology evaluated with longitudinal MR brain images of preterm infants PLoS ONE 2015 10 7 e0131552 10.1371/journal.pone.0131552 26161536 26. Mills KL Tamnes CK Methods and considerations for longitudinal structural brain imaging analysis across development Dev Cogn Neurosci 2014 9 172 190 10.1016/j.dcn.2014.04.004 24879112 27. Fischl B Dale AM Measuring the thickness of the human cerebral cortex from magnetic resonance images Proc Natl Acad Sci 2000 97 20 11050 11055 10.1073/pnas.200033797 10984517 28. Cao B Mwangi B Passos IC Wu MJ Keser Z Zunta-Soares GB Lifespan gyrification trajectories of human brain in healthy individuals and patients with major psychiatric disorders Sci Rep 2017 7 1 511 10.1038/s41598-017-00582-1 28360420 29. Li G Wang L Shi F Lyall AE Lin W Gilmore JH Mapping longitudinal development of local cortical gyrification in infants from birth to 2 years of age J Neurosci 2014 34 12 4228 4238 10.1523/JNEUROSCI.3976-13.2014 24647943 30. Lebel C Beaulieu C Longitudinal development of human brain wiring continues from childhood into adulthood J Neurosci 2011 31 30 10937 10947 10.1523/JNEUROSCI.5302-10.2011 21795544 31. Aubert-Broche B Fonov VS García-Lorenzo D Mouiha A Guizard N Coupé P A new method for structural volume analysis of longitudinal brain MRI data and its application in studying the growth trajectories of anatomical brain structures in childhood Neuroimage 2013 82 393 402 10.1016/j.neuroimage.2013.05.065 23719155 32. Tamnes CK Walhovd KB Dale AM Østby Y Grydeland H Richardson G Brain development and aging: overlapping and unique patterns of change Neuroimage 2013 68 63 74 10.1016/j.neuroimage.2012.11.039 23246860 33. World Health Organization (WHO). Antiretroviral therapy of HIV infection in infants and children in resource-limited settings: towards universal access. Recommendations for a public health approach. 2006. Source:http://www.who.int/hiv/pub/guidelines/WHOpaediatric.pdf. Accessed 12 October, 2017. 34. Cotton MF Violari A Otwombe K Panchia R Dobbels E Rabie H Early time-limited antiretroviral therapy versus deferred therapy in South African infants infected with HIV: results from the children with HIV early antiretroviral (CHER) randomised trial Lancet 2013 382 9904 1555 1563 10.1016/S0140-6736(13)61409-9 24209829 35. Madhi SA Adrian P Cotton MF McIntyre JA Jean-Philippe P Meadows S Effect of HIV infection status and anti-retroviral treatment on quantitative and qualitative antibody responses to pneumococcal conjugate vaccine in infants J Infect Dis 2010 202 3 355 361 10.1086/653704 20583920 36. van Wyhe KS Laughton B Cotton MF Meintjes EM van der Kouwe AJ Boivin MJ Cognitive outcomes at ages seven and nine years in South African children from the children with HIV early antiretroviral (CHER) trial: a longitudinal investigation J Int AIDS Soc 2021 24 7 e25734 10.1002/jia2.25734 34259393 37. Tisdall MD Hess AT Reuter M Meintjes EM Fischl B van der Kouwe AJ Volumetric navigators for prospective motion correction and selective reacquisition in neuroanatomical MRI Magn Reson Med 2012 68 2 389 399 10.1002/mrm.23228 22213578 38. Van der Kouwe AJ Benner T Salat DH Fischl B Brain morphometry with multiecho MPRAGE Neuroimage 2008 40 2 559 569 10.1016/j.neuroimage.2007.12.025 18242102 39. Reuter M Rosas HD Fischl B Highly accurate inverse consistent registration: a robust approach Neuroimage 2010 53 4 1181 1196 10.1016/j.neuroimage.2010.07.020 20637289 40. Reuter M Fischl B Avoiding asymmetry-induced bias in longitudinal image processing Neuroimage 2011 57 1 19 21 10.1016/j.neuroimage.2011.02.076 21376812 41. Bernal-Rusiel JL Greve DN Reuter M Fischl B Sabuncu MR Alzheimer’s disease neuroimaging initiative. Statistical analysis of longitudinal neuroimage data with linear mixed effects models Neuroimage 2013 66 249 260 10.1016/j.neuroimage.2012.10.065 23123680 42. Sowell ER Thompson PM Leonard CM Welcome SE Kan E Toga AW Longitudinal mapping of cortical thickness and brain growth in normal children J Neurosci 2004 24 38 8223 31 10.1523/JNEUROSCI.1798-04.2004 15385605 43. Raznahan A Shaw P Lalonde F Stockman M Wallace GL Greenstein D How does your cortex grow? J Neurosci 2011 31 19 7174 7177 10.1523/JNEUROSCI.0054-11.2011 21562281 44. Ducharme S Albaugh MD Nguyen TV Hudziak JJ Mateos-Pérez JM Labbe A Trajectories of cortical thickness maturation in normal brain development—the importance of quality control procedures Neuroimage 2016 125 267 279 10.1016/j.neuroimage.2015.10.010 26463175 45. Schnack HG Van Haren NE Brouwer RM Evans A Durston S Boomsma DI Changes in thickness and surface area of the human cortex and their relationship with intelligence Cereb Cortex 2015 25 6 1608 1617 10.1093/cercor/bht357 24408955 46. Sowell ER Peterson BS Thompson PM Welcome SE Henkenius AL Toga AW Mapping cortical change across the human life span Nat Neurosci 2003 6 3 309 10.1038/nn1008 12548289 47. White T Su S Schmidt M Kao CY Sapiro G The development of gyrification in childhood and adolescence Brain Cogn 2010 72 1 36 45 10.1016/j.bandc.2009.10.009 19942335 48. Santos E Noggle CA Goldstein S Naglieri JA Synaptic pruning Encyclopedia of child behavior and development 2011 Boston Springer 49. Gogtay N Giedd JN Lusk L Hayashi KM Greenstein D Vaituzis AC Dynamic mapping of human cortical development during childhood through early adulthood Proc Natl Acad Sci 2004 101 21 8174 8179 10.1073/pnas.0402680101 15148381 50. Huizinga M Dolan CV van der Molen MW Age-related change in executive function: developmental trends and a latent variable analysis Neuropsychologia 2006 44 11 2017 2036 10.1016/j.neuropsychologia.2006.01.010 16527316 51. Diamond A Stuss D Knight R Normal development of prefrontal cortex from birth to young adulthood: cognitive functions, anatomy, and biochemistry Principles of frontal lobe function 2002 New York Oxford University Press 466 503 52. Van Essen DC A tension-based theory of morphogenesis and compact wiring in the central nervous system Nature 1997 385 6614 313 10.1038/385313a0 9002514 53. Chung YS Hyatt CJ Stevens MC Adolescent maturation of the relationship between cortical gyrification and cognitive ability Neuroimage 2017 158 319 331 10.1016/j.neuroimage.2017.06.082 28676299 54. Kersbergen KJ Leroy F Išgum I Groenendaal F de Vries LS Claessens NH Relation between clinical risk factors, early cortical changes, and neurodevelopmental outcome in preterm infants Neuroimage 2016 142 301 310 10.1016/j.neuroimage.2016.07.010 27395393 55. Blanton RE Levitt JG Thompson PM Narr KL Capetillo-Cunliffe L Nobel A Mapping cortical asymmetry and complexity patterns in normal children Psychiatr Res Neuroimaging 2001 107 1 29 43 10.1016/S0925-4927(01)00091-9 56. Lewis-de Los Angeles CP Williams PL Jenkins LM Huo Y Malee K Alpert KI Brain morphometric differences in youth with and without perinatally-acquired HIV: a cross-sectional study NeuroImage Clinical. 2020 26 102246 10.1016/j.nicl.2020.102246 32251906 57. Yu X Gao L Neuroanatomical changes underlying vertical HIV infection in adolescents Front Immunol 2019 10 814 10.3389/fimmu.2019.00814 31110499 58. Yadav SK Gupta RK Garg RK Venkatesh V Gupta PK Singh AK Altered structural brain changes and neurocognitive performance in pediatric HIV Neuro Image Clin 2017 14 316 322 59. Hoare J Fouche JP Phillips N Joska JA Myer L Zar HJ Structural brain changes in perinatally HIV-infected young adolescents in South Africa AIDS 2018 32 18 2707 2718 10.1097/QAD.0000000000002024 30234601 60. Van den Hof M Jellema PE Ter Haar AM Scherpbier HJ Schrantee A Kaiser A Caan MW Majoie CB Reiss P Wit FW Mutsaerts HJ Normal structural brain development in adolescents treated for perinatally acquired HIV: a longitudinal imaging study AIDS 2021 35 8 1221 10.1097/QAD.0000000000002873 33710018 61. Wade BS Valcour VG Puthanakit T Saremi A Gutman BA Nir TM Watson C Aurpibul L Kosalaraksa P Ounchanum P Kerr S Mapping abnormal subcortical neurodevelopment in a cohort of Thai children with HIV NeuroImage Clin. 2019 23 101810 10.1016/j.nicl.2019.101810 31029050 62. Bunupuradah T Duong T Compagnucci A McMaster P Bernardi S Kanjanavanit S Outcomes after reinitiating antiretroviral therapy in children randomized to planned treatment interruptions AIDS 2013 27 4 579 589 10.1097/QAD.0b013e32835c1181 23135172 63. Montserrat M Plana M Guardo AC Andrés C Climent N Gallart T Impact of long-term antiretroviral therapy interruption and resumption on viral reservoir in HIV-1 infected patients AIDS 2017 31 13 1895 1897 10.1097/QAD.0000000000001560 28590333 64. Donald KA Hoare J Eley B Wilmshurst JM Neurologic complications of pediatric human immunodeficiency virus: implications for clinical practice and management challenges in the African setting Semin Pediatr Neurol 2014 21 1 3 11 10.1016/j.spen.2014.01.004 24655398 65. Chriboga CA Fleishman S Champion S Gaye-Robinson L Abrams EJ Incidence and prevalence of HIV encephalopathy in children with HIV infection receiving highly active anti-retroviral therapy (HAART) J Pediatr 2005 146 3 402 407 10.1016/j.jpeds.2004.10.021 15756229 66. Foster CJ Biggs RL Melvin D Walters MDS Tudor-Williams G Lyall EGH Neurodevelopmental outcomes in children with HIV infection under 3 years of age Dev Med Child Neurol 2006 48 8 677 682 10.1017/S0012162206001423 16836781 67. Lobato MN Caldwell MB Ng P Oxtoby MJ Pediatric spectrum of disease clinical consortium. Encephalopathy in children with perinatally acquired human immunodeficiency virus infection J Pediatr. 1995 126 5 710 715 10.1016/S0022-3476(95)70397-7 7751993
PMC010xxxxxx/PMC10108471.txt
==== Front Radiat Oncol Radiat Oncol Radiation Oncology (London, England) 1748-717X BioMed Central London 37069641 2259 10.1186/s13014-023-02259-8 Research Application of four-dimensional cone beam computed tomography in lung cancer radiotherapy Abulimiti Muyasha [email protected] 1 Yang Xu [email protected] 2 Li Minghui [email protected] 2 Huan Fukui [email protected] 2 Zhang Yanxin [email protected] 2 Jun Liang [email protected] 2 1 grid.506261.6 0000 0001 0706 7839 National Cancer Center, National Clinical Research Center for Cancer/Cancer Hospital & Shenzhen Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Shenzhen, 518116 China 2 grid.506261.6 0000 0001 0706 7839 Department of Radiation Oncology, National Clinical Research Center for Cancer/Cancer Hospital, National Cancer Center, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, 100021 China 17 4 2023 17 4 2023 2023 18 6929 10 2022 5 4 2023 © The Author(s) 2023, corrected publication 2023 https://creativecommons.org/licenses/by/4.0/ Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Objective This study explored the application of four-dimensional cone beam computed tomography (4D CBCT) in lung cancer patients, seeking to improve the accuracy of radiotherapy and to establish a uniform protocol for the application of 4D CBCT in radiotherapy for lung cancer. Methods 4D CBCT was applied to evaluate tumor volume response (TVR), motion, and center coordinates during radiotherapy in 67 eligible individuals with lung cancer diagnoses. The differences between 4D CBCT and 3D CBCT in different registration methods were compared. Results TVR was observed during treatment in 41% of patients (28/67), with a mean volume reduction of 41.7% and a median time to TVR of 19 days. Tumor motion was obvious in 16 patients, with a mean value of 0.52 cm (0.22 to 1.34 cm), and in 3 of 6 tumors close to the diaphragm (0.28 to 0.66 cm). Gray value registration based on mean density projection could still achieve close results to the 4D gray value registration. However, when the registration was based on bone alone, partial off-targeting occurred in the treatment in 41.8% of cases. The off-target rate was 19.0% when the tumor motion was ≤ 0.5 cm and 52.2% when the motion was > 0.5 cm. Conclusion Tumor volume and motion of intrapulmonary lesions in individuals diagnosed with lung cancer varied significantly in the third week of radiotherapy. 4D CBCT may be more advantageous for isolated lesions without reference to relative anatomical structures or those near the diaphragm. Grayscale registration based on mean density projection is feasible. Keywords 4D cone beam CT Lung cancer Radiotherapy http://dx.doi.org/10.13039/501100009592 Beijing Municipal Science and Technology Commission No.Z151100004015052 No.Z151100004015052 No.Z151100004015052 No.Z151100004015052 No.Z151100004015052 No.Z151100004015052 issue-copyright-statement© BioMed Central Ltd., part of Springer Nature 2023 ==== Body pmcIntroduction Lung cancer is the most frequently diagnosed cancer worldwide, and the number of lung cancer diagnoses continues to rise, showing a high incidence rate and a low survival rate, with the associated mortality accounting for 19.4% of deaths among all individuals diagnosed with cancer [1]. In the multidisciplinary treatment of lung cancer, radiotherapy is mainly used for stereotactic large-split radiotherapy for early-stage lung cancer, adjuvant or radical treatment is chosen for intermediate stage tumors, and palliative treatment is available for the advanced stage. Stereotactic body radiotherapy (SBRT) is the standard treatment for early inoperable cases [2]. However, uncertainty in the spatial location of the tumor can significantly impact the implementation of SBRT. Furthermore, studies have demonstrated that off-target radiation based solely on population data can lead to inaccurate treatment volumes, which in turn can cause decreased local control rates and increased dose-limiting toxicity in normal tissues [3]. Radiotherapy for lung cancer, therefore, requires strict control of tumor motion to ensure that the tumor is irradiated at an adequate dose and the surrounding normal tissues are maximally protected. Four-dimensional cone beam computed tomography (4D CBCT) is an on-board imaging system that acquires respiratory motion data during the frame rotation, enabling the generation of four-dimensional images before, during, and after treatment [4]. The 4D CBCT images can also be aligned with Four-dimensional computed tomography (4DCT) positioning images to correct setup errors and to monitor the range of tumor motion in real-time, ensuring that the moving tumor is within the irradiation range and facilitating the implementation of precision radiotherapy. The implementation of 4D CBCT overcomes the defect that three-dimensional (3D) CBCT is a quick scan, reflects the one-sided range of tumor motion, and can evaluate the amplitude and range of tumor motion online. However, because of the long scanning time of 4D CBCT and the demanding technical requirements, its use in the clinic has been limited, and no uniform application scheme is recommended. This paper explores the use of 4D CBCT in lung cancer radiotherapy for monitoring tumor volume response (TVR) and tumor motion and attempts to establish a protocol for the use of 4D CBCT in lung cancer radiotherapy. Materials and methods Patients A total of 67 lung cancer patients who received thoracic radiotherapy in the Cancer Hospital of the Chinese Academy of Medical Sciences from 2015 to 2018 were enrolled. The inclusion criteria were: (1) age ≥ 18 years; (2) Karnofsky Performance Scale (KPS) ≥ 90; (3) able to tolerate long scanning times; (4) visible intrapulmonary lesions. The general clinical characteristics of the enrolled patients are displayed in Table 1. Table 1 Clinical characteristics of study patients (N = 67) Group Number (%) Mean Median Range Gender Male 50 (74.6) - - - Female 17 (25.4) - - - Age 36–60 27 (40.3) - - - 61–70 24 (35.8) - - - ≥ 71 16 (23.9) 63 63 36–82 BMI (kg/m2) - - 24.79 24.49 15.85–32.53 Tumor volume 0–50 45 (67.2) 42.31 13.81 0.44–285.78 50–300 22 (32.8) Tumor site Upper lobe of the right lung 19 (28.4) - - - Middle lobe of the right lung 5 (7) Lower lobe of the right lung 17 (25.4) Upper lobe of the left lung 8 (13.4) Lower lobe of the left lung 18 (26.8) Adjacent to surrounding structures Mediastinum 19 (28.4) - - - Chest wall 15 (22.4) Diaphragm 8 (11.9) Pulmonary atelectasis or obstructive pneumonia Yes 5 (7.5) - - - No 62 (92.5) Image datasets 4D CBCT 260 - - - 3D CBCT 228 Methods Localization of all tumors was by 4DCT and image-guided radiotherapy was performed with 4D CBCT after the completion of the treatment plan. At least 3 to 4 CBCT scans were completed in every case. Along with the planning 4DCT data, all 4D CBCT images and setup error data were collected, and a database and image library was established. In Part1: Tumor Volume、Motion and Center coordinates of tumor were compared between 4DCBCT and planning 4DCT, and the clinical value of 4DCBCT was determined by the characteristics of the tumor changes. During comparison of the tumor motion, the position of the tumor in three axes in 4DCBCT in the complete respiratory phase (0–100%) was compared with the position of the tumor in 4DCT, and the point with the greatest difference of activity was selected and compared with the extreme difference of action in each scan to explore the activity pattern. In Part 2: Different images were registrated with 4DCT for comparison of setup errors and to explore the most feasible registration method. SPSS11.5 software was used to calculate the mean offset values and the means ± standard deviations of setup errors according to the data, and a t-test was performed. Setup error1 was compared between 4DCBCT bone-based registration and grayscale + manual registration, Grayscale + manual registration was defined according to gray-based automatic registration with 4DCT and adjusted by the radiotherapist based on soft tissue(mainly the primary tumor),setup error 2 was compared between 4DCBCT grayscale + manual registration and average intensity projection(AIP) of 4DCBCT grayscale + manual registration and setup error 3 was compared between 4DCBCT grayscale + manual registration and 3DCBCT grayscale + manual registration. Figure 1 shows the study flowchart. Fig. 1 Study flowchart The evaluation index included: Dynamic change of tumor volume. Dynamic analysis of tumor motion. Stability of tumor centers. 4D and 3D bone and gray value registration data. Results Sixty-seven individuals who were diagnosed with lung cancer were enrolled. All of them had multiple 4D and 3D CBCT scans during the treatment. In total, 260 4D CBCT image datasets and 228 3D CBCT image datasets were collected. Analysis of tumor volume during treatment The target area was outlined for 39 4D CBCT series, All tumors were delineated manually by two physicians with more than 5 years of clinical experience and reviewed by physicians on the same team with more than 20 years of clinical experience. The tumor volumes were calculated and compared. The mean tumor volume in this group was 38.63 mL (0.44 to 285.78 mL). Tumor volume response (TVR) was observed in 16 (41.0%) cases during the treatment. Tumor volume change curves for each 4D CBCT in the patients with TVR are shown in Fig. 2. The mean tumor volume was 57.06 mL (0.56 to 285.78 mL) in this group and the mean volume reduction was 41.70% (17–77%),The average volume reduction was 23.79mL(9.7mL-43.9mL). TVR was usually observed within the first 3 weeks, with a median time to response of 19 days (3 to 40 days). Fig. 2 Dynamic change of target volume in 16 patients with tumor volume reduction. The median time to observed tumor volume response was 19 days (3 to 40 days) Analysis of tumor motion during treatment In the initial 4D CBCT validation, tumor motion could be traced along three axes in 64 cases (Table 2), of which 13 (20%) were volume reduced and 3 (5%) were near the diaphragm. The X-axis is left-right,the Y-axis is superior-inferior, the Z-axis is anterior-posterior. The mean value of the maximum difference in motion (the difference between the maximum and minimum motion) along the Y-axis was 0.24 cm (0.00 to 1.34 cm). In 16 cases (25%), the maximum difference was > 0.20 cm and the mean value of the maximum difference was 0.52 cm (0.22 to 1.34 cm) in these cases. Tumor motion decreased in 10 cases and increased in 3 cases. The median time to maximum motion was 13 days after the first irradiation (5 to 40 days). Three of the 6 lesions near the diaphragm demonstrated fluctuating tumor motion (0.28, 0.33, and 0.66 cm) with stable tumor volumes. Table 2 Tumor motion along three axes (N = 64) N Mean (cm) Range (cm) X-axis 20 0.33 0.12–0.80 Y-axis 64 0.91 0.13–2.40 ≤ 0.20 cm 50 0.15 0.00–0.20 > 0.20 cm 14 0.52 0.22–1.34 Z-axis 18 0.31 0.11–0.55 Tumor motion was observed in all directions, mostly during the first 3 weeks of treatment; the most movement was measured along the Y-axis, in volume-reduced patients, and in the lesions near the diaphragm, which may show greater motion regardless of volume changes. Analysis of tumor center coordinates The geometric center of the tumor was identified by outlining the target area in 39 4D CBCT images and determining the coordinates of the geometric center. The tumor center was calibrated by using 4D CBCT grayscale + manual registration data. The average range of motion of geometric centers was 0.25 cm (0.03 to 1.44 cm) on the X-axis, 0.28 cm (0.03 to 2.50 cm) on the Y-axis, and 0.22 cm (0.01 to 0.73 cm) on the Z-axis. Among 39 lesions, those with a range of motion < 0.5 mm were defined as centrally stable (N = 22) and those with a range of motion ≥ 2.5 mm as unstable (N = 17). The result of a univariate analysis of the mean tumor volume, tumor motion, and patient BMI in these two groups is shown in Table 3. Table 3 Univariate analysis of factors influencing central stability (N = 39) Stable (N = 22) Unstable (N = 17) P Average tumor volume 19.35 mL (0.44–85.12 mL) 63.60 mL (0.56–285.78 mL) 0.008 Range of motion 0.74 cm (0.00–2.38 cm) 1.03 cm (0.00–2.40 cm) 0.140 Body mass index 24.04 (20.00–27.76) 24.69 (15.85–28.73) 0.350 Tumor volume was found to be an independent factor affecting central stability (P = 0.008), and central stability was worst in patients with large volumes and a range of motion ≥ 2.5 mm. There was no statistical difference in range of tumor motion and body mass index between the two groups (P > 0.05) Registration with planning 4DCT and data analysis Setup error1: Comparison of 4DCBCT bone-based registration and grayscale + manual registration Patients (N = 67) were divided into a small error group (N = 39) and a large error group (N = 28); large error was present when the deviation between bone registration and 4D grayscale + manual registration was > 0.5 cm. Grayscale + manual registration was defined according to gray-based automatic registration with 4DCT and adjusted by the radiotherapist based on soft tissue(mainly the primary tumor).Univariate analysis of mean tumor volume, motion, and BMI was performed for both groups (Table 4). Table 4 Factors influencing the error of four-dimensional cone beam computed tomography bone- based registration and grayscale + manual registration Small error group (N = 39) Large error group (N = 28) P Average volume (mL) 43.30 (0.44–285.34) 40.98 (1.28–164.47) > 0.05 Tumor Motion (cm) 0.74 (0.00–2.40) 1.03 (0.17–2.38) 0.044 Motion ≤ 0.5 cm (N = 21) 17 (43.5%) 4 (14.2%) - Motion > 0.5 cm (N = 46) 22 (56.5%) 24 (85.8%) 0.032 Body mass index 24.19 (15.85–20.05) 25.63 (22.49–32.53) > 0.05 Tumor motion was the most influential factor in the difference between the two registration methods (P < 0.05); tumor motion was > 1 cm in the large error group. Upon further analysis of the effect of tumor motion on registration errors, the partial volume off-target rate was 41.8% if based on bone registration only, 19.0% when the motion was ≤ 0.5 cm, and 52.2% when the motion was > 0.5 cm. Setup error2: Comparison of 4DCBCT grayscale + manual registration and average intensity projection(AIP) of 4DCBCT grayscale + manual registration Two sets of registration data were obtained: 4D grayscale + manual registration and AIP of 4DCBCT grayscale + manual registration. In 260 sets of 4D CBCT images, the margin of error was < 0.5 cm in three axes. There were only four occasions out of 260 (1.5%) where the registration verification process demonstrated and error of ≥ 0.30 cm (0.30 to 0.34 cm).The frequency of errors in each axis is shown in Table 5. There was no difference between the 4D CBCT grayscale + manual registration and AIP of 4DCBCT grayscale + manual registration. Table 5 Registration errors in 4DCBCT grayscale + manual registration and AIP of 4DCBCT grayscale + manual registration Registration error (cm) X-axis (frequency, proportion) Y-axis (frequency, proportion) Z-axis (frequency, proportion) 0.00–0.10 203 (78.1%) 198 (76.3%) 198 (76.2%) 0.10–0.20 52 (20.0%) 62 (23.8%) 47 (18.1%) >0.20 5 (1.9%) 0 15 (5.7) Total 260 (100%) 260 (100%) 260 (100%) Setup error3: Comparison of 4DCBCT grayscale + manual registration and 3DCBCT grayscale + manual registration Because of the constraints of facility resources, patient treatment time, and radiation exposure, our trial used alternating 4D and 3D CBCT validation, which can reduce machine depreciation, decrease the chance of malfunction, reduce patient treatment time and the potential for movement during prolonged treatment, and limit the risk of excessive X-ray irradiation. However, setup errors, baseline differences, tumor volume, location, and motion changes may influence differences between the two methods of registration. Therefore, the 4D and 3D CBCT validation data were analyzed by excluding cases with significant tumor volume, motion changes, and unstable tumor centers. In the end, 21 cases were eligible for analysis. The bone registration data were subtracted from 4D and 3D CBCT grayscale + manual registration data to eliminate the influence of setup errors. When the two groups of data were compared, no difference was found in the 4D and 3D CBCT registration data of 14 cases on each axis, with 10 cases (71.4%) having tumors adjacent to the chest wall, mediastinum, or vertebral body. In 7 cases, there was a difference between the two kinds of registration (X-axis, 5 cases; Y-axis, 2 cases). Among these, the tumor was adjacent to the chest wall, mediastinum, or vertebral body in 3 cases (42.9%) (P = 0.296). Because of the relatively rigid anatomical reference, there was no difference between 4D CBCT and 3D CBCT registration when the tumor was adjacent to the chest wall, mediastinum, or vertebral body. Discussion Accelerator-borne 3D CBCT is widely used to correct setup errors and guide precise radiotherapy because of its low radiation dose and fast scanning speed. 4D CBCT, which contains respiratory motion information, is the gold standard for positional correction in imaging-guided radiotherapy (IGRT) because it allows real-time monitoring of patient setup errors and target area motion. However, specific indications for 4D CBCT have not yet been clarified. This study explored the applicability of 4D CBCT by analyzing the changes in tumor volume, position, and respiratory motion during treatment. In addition, it explored the differences between bone and grayscale registration in 4D CBCT and 3D CBCT to provide additional clinical application schemes for the precise treatment of lung cancer. 4DCT has become the standard for outlining the intended treatment volume (ITV) throughout the respiratory cycle in precision radiotherapy for lung cancer. Mean intensity projection 4DCT images contain significantly more tumor motion information than 3D CBCT and 3DCT [5]. Wang et al. [6] compared ITV3DCBCT and ITV4DCT and found that the volume difference was 8% and Liu et al. [7] found that ITV3DCBCT was 11.8% smaller than ITV4DCT. Among the three methods, 4DCT was the most accurate in determining the tumor volume, followed by 3D CBCT, which had the largest error. However, the above methods did not consider respiratory motion and TVR during treatment. This study further explored the TVR by 4D CBCT during the treatment. TVR was observed in 41% of patients, with an average reduction of tumor volume of 41.7%; the response was observed within a median time of 19 days after starting radiotherapy. If 4D CBCT can be used to evaluate the TVR, especially at around the third week of treatment, it would allow for more precise second-course planning based on the extent of tumor reduction, reducing unnecessary irradiation of normal tissues and irradiation complications. In addition, 4D CBCT can indicate tumor regression during treatment and reflect the radiosensitivity, thus providing a reference for the individualized treatment mode. Tumor motion is a hot topic in 4D CBCT research. Gottlieb et al. [8] used 4D CBCT to analyze baseline variability and motion patterns in 23 cases of lung cancer invading the mediastinum and found that 4D CBCT differed from localized 4DCT in the anterior-posterior direction before fractions 3, 10, and 20 (P < 0.05), suggesting that the motion of near-mediastinal tumors is prone to fluctuations with physiological conditions. Purdie et al. [9] compared the tumor motion trajectories of 4DCT and 4D CBCT images from 12 individuals with lung cancer diagnoses; the 4D CBCT motion trajectories matched those of 4DCT in 10 cases but showed significant deviations in the superior-inferior and anterior-posterior directions in 2. Sonke et al. [10] reported inter fractionated tumor motion on 4D CBCT in 56 individuals with lung cancer diagnoses; the range of motion range was mostly < 1.5 mm, with the most significant variation in the superior-inferior direction and in lower lobe lesions, and linear analysis revealed that the random error in baseline variation in three directions correlated with the mean tumor motion amplitude (P < 0.05). In this study, we confirmed that tumors could move in all directions, and also found that the maximum motion was along the Y-axis and in volume-reduced and diaphragm-adjacent lesions; volume reduction was most commonly observed in the third week of treatment. The tumors with greater motion were mostly those with TVR and those located near the diaphragm (0.28 to 0.66 cm). Similarly, previous studies suggest that tumors near the diaphragm or mediastinum or those with reduced volume might show greater motion in the Y-axis direction. This study also found a statistical difference between tumor volume and central stability. Large-volume tumors (average tumor volume of 63 mL) have the greatest degree of central instability, i.e. the range of motion exceeded 2.5 mm. Small-volume tumors (average volume of 19 mL) had central motion < 2.5 mm. Tan et al. [11] included small lung cancers (maximum tumor diameter between 18 and 27 mm) in a comparison of 4D CBCT and 3D CBCT and found that setup errors were significant in all directions. Therefore, regardless of the tumor size, 4D CBCT can reflect the movement of the tumor center and reduce the uncertainty of tumor location caused by internal movement due to respiratory motion. A comparison of 4D and 3D CBCT bone registration and manual grayscale registration reveals that when the tumor is more active, there is a 50% or higher probability of partial volume off-target when the registration is based solely on bone landmarks. If grayscale + manual registration is based on mean density projection, the effect is close to that of 4D CBCT grayscale + manual registration. There is no statistical difference between the registration data of 4D CBCT and 3D CBCT when the tumor is adjacent to the chest wall, mediastinum, or vertebral bodies because of the relative rigidity of the anatomical structures. Conversely, for lesions with isolated motion within the lung, especially for those with large ranges of motion, 4D CBCT may be more advantageous. Wang et al. [12] studied the difference between manual registration and gray registration, and found that gray registration could achieve better results than manual registration in the early stage of treatment, while the advantages of manual registration gradually emerged as radiotherapy progressed. Li et al. [13] selected the spine, spine + ITV, and lung for registration, with similar registration errors in the three groups. Schreibman et al. [14] explored the registration of full-time 4DCT and 4D CBCT in a case of liver cancer and demonstrated significant registration errors after bone registration, which were corrected after grayscale registration. Sweeney et al. [15] used 4DCT and 4D CBCT end-expiratory image registration errors as the gold standard (IG-4DCBCT) to compare 3D CBCT registration errors with IG-4DCBCT in 21 lung cancer patients and found that the difference between IG-4DCBCT and IG-3DCBCT was the largest (4.3 mm) in the superior-inferior direction; linear regression analysis indicated that the difference in registration between 4D CBCT and 3D CBCT gradually increased with increasing tumor motion. As per the findings of this study, when the tumor is adjacent to the chest wall, mediastinum, or vertebral body, because of its relatively rigid anatomical structure, accurate registration can be achieved by combining grayscale and manual 3D CBCT registration with bone registration. However, 4D CBCT with grayscale registration based on mean density projection is recommended when the tumor is located in an isolated region within the lung or is more mobile. Conclusion Our results showed that patients treated with radiotherapy for lung cancer with intrapulmonary lesions have greater TVR and motion changes in the third week of treatment. As a result, 4D CBCT may be more advantageous for isolated lesions without reference to relative anatomical structures or proximity to the diaphragm, and grayscale registration based on mean density projection is feasible. Authors’ contributions MA, J Liang and X Yang contributed to the study conception and design. Data collection and analysis were performed by MH Li, FK Huan, YX Zhang. The first draft of the manuscript was written by MA. and all authors commented on previous versions of the manuscript. J Liang and X Yang helped perform the analysis with constructive discussions. All authors read and approved the final manuscript. Funding This study was supported by Beijing Municipal Science and Technology Commission (No.Z151100004015052). Data availability The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. Declarations Ethical approval and Consent to participate The study was performed in line with the principles of the Declaration of Helsinki and was approved by the Ethics Committee of Cancer Hospital, Chinese Academy of Medical Sciences (NCC2015ST-10). All of the patients were recruited after providing informed consent for analysis of their clinical data. Consent for publication Written informed consent for publication was obtained from all participants. Competing interests The authors have no relevant financial or non-financial interests to disclose. The original online version of this article was revised: the second author ‘Xu Yang’ should be co-first author with ‘Muyasha Abulimiti’. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Muyasha Abulimiti and Xu Yang are co-first authors. Change history 7/12/2023 A Correction to this paper has been published: 10.1186/s13014-023-02280-x ==== Refs References 1. Xia C Dong X Li H Cancer statistics in China and United States, 2022: profiles, trends, and determinants Chin Med J 2022 135 5 584 90 10.1097/cm9.0000000000002108 35143424 2. Ettinger D Wood D Aisner D NCCN Guidelines Insights: Non-Small Cell Lung Cancer, Version 2.2021 J Natl Compr Cancer Network: JNCCN 2021 19 3 254 66 10.6004/jnccn.2021.0013 33668021 3. Guckenberger M Krieger T Richter A Potential of image-guidance, gating and real-time tracking to improve accuracy in pulmonary stereotactic body radiotherapy Radiotherapy and oncology: journal of the European Society for Therapeutic Radiology and Oncology 2009 91 3 288 95 10.1016/j.radonc.2008.08.010 18835650 4. Kida S, Masutani Y, Yamashita H et al. In-treatment 4D cone-beam CT with image-based respiratory phase recognition. 2012;5(2):138–147. 10.1007/s12194-012-0146-5 5. Li F, Li J, Zhang Y, Min X, Shao Q. Geometrical differences in gross target volumes between 3DCT and 4DCT imaging in radiotherapy for non-small-cell lung cancer. J Radiat Res. 2013;54(5). 10.1093/jrr/rrt017. 6. Wang Z Wu QJ Marks LB Larrier N Yin FF Cone-beam CT localization of internal target volumes for stereotactic body radiotherapy of lung lesions Int J Radiat Oncol Biol Phys 2007 69 5 1618 24 10.1016/j.ijrobp.2007.08.030 18035215 7. Liu H Khan R D’Ambrosi R Krobutschek K Nugent Z Lau H The influence of target and patient characteristics on the volume obtained from cone beam CT in lung stereotactic body radiation therapy Radiotherapy and oncology: journal of the European Society for Therapeutic Radiology and Oncology 2013 106 3 312 6 10.1016/j.radonc.2013.01.001 23395064 8. Gottlieb KL, Hansen CR, Hansen O, Westberg J, Brink CJAO. Investigation of respiration induced intra- and inter-fractional tumour motion using a standard Cone Beam CT. 2010;49(7):1192–1198. 10.3109/0284186X.2010.498834 9. Purdie TG Moseley DJ Bissonnette JP Sharpe MB Oncologica DJJA Respiration correlated cone-beam computed tomography and 4DCT for evaluating target motion in Stereotact Lung Radiation Therapy 2006 45 7 915 22 10.1080/02841860600907345 10. Sonke JJ Rossi M Wolthaus J Herk MV Belderbos JJIjoro, biology, physics. Frameless Stereotactic Body Radiotherapy for Lung Cancer using Four-Dimensional Cone Beam CT Guidance 2008 74 2 567 74 10.1259/bjr.20140620 11. Tan Z, Liu C, Ying Z, Shen WJJoRR. Preliminary comparison of the registration effect of 4D-CBCT and 3D-CBCT in image-guided radiotherapy of Stage IA non-small-cell lung cancer. 2017;58(6):854–861. 10.1093/jrr/rrx040 12. Wang Jiahao L Xiadong X Bing The application research about registration methods of 4D-CBCT in non-small cell lung cancer with SBRT Chin J Radiat Oncol 2016 25 1 71 5 13. Li J, Harrison A et al. Evaluation of Elekta 4D cone beam CT-based automatic image registration for radiation treatment of lung cancer. 2015; 10.1259/bjr.20140620 14. Schreibmann E, Thorndyke B, Li T, Wang J, Xing L. Four-Dimensional Image Registration for Image-Guided Radiotherapy. Int J Radiat Oncol Biol Phys. 2008;71(2):578–86. 10.1016/j.ijrobp.2008.01.042. 15. Sweeney RA, Seubert B, Stark S, Oncology VHJR. Accuracy and inter-observer variability of 3D versus 4D cone-beam CT based image-guidance in SBRT for lung tumors. 2012;7(1):81. 10.1186/1748-717X-7-81
PMC010xxxxxx/PMC10119427.txt
==== Front Front Pharmacol Front Pharmacol Front. Pharmacol. Frontiers in Pharmacology 1663-9812 Frontiers Media S.A. 37089942 1170637 10.3389/fphar.2023.1170637 Pharmacology Original Research Curcumin alleviates experimental colitis in mice by suppressing necroptosis of intestinal epithelial cells Zhong et al. 10.3389/fphar.2023.1170637 Zhong Yuting 1 † Tu Ye 2 3 † Ma Qingshan 4 Chen Linlin 3 Zhang Wenzhao 3 Lu Xin 3 Yang Shuo 3 * Wang Zhibin 2 3 * Zhang Lichao 1 * 1 Department of Pharmacy, Shanghai Municipal Hospital of Traditional Chinese Medicine, Shanghai University of Traditional Chinese Medicine, Shanghai, China 2 Department of Pharmacy, Shanghai East Hospital, School of Medicine, Tongji University, Shanghai, China 3 Department of Critical Care Medicine, School of Anesthesiology, Naval Medical University, Shanghai, China 4 Longhua Hospital Affiliated to Shanghai University of Traditional Chinese Medicine, Shanghai, China Edited by: Jie Liu, Zunyi Medical University, China Reviewed by: Jiasi Wu, Chengdu University of Traditional Chinese Medicine, China Ayaz Shahid, Western University of Health Sciences, United States *Correspondence: Lichao Zhang, [email protected]; Zhibin Wang, [email protected]; Shuo Yang, [email protected] This article was submitted to Ethnopharmacology, a section of the journal Frontiers in Pharmacology † These authors have contributed equally to this work 07 4 2023 2023 14 117063721 2 2023 30 3 2023 Copyright © 2023 Zhong, Tu, Ma, Chen, Zhang, Lu, Yang, Wang and Zhang. 2023 Zhong, Tu, Ma, Chen, Zhang, Lu, Yang, Wang and Zhang https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Curcumin, the primary bioactive substance in turmeric, exhibits potential therapeutic effects on ulcerative colitis. However, its mechanism for regulating necroptosis in colitis has not been fully elucidated. In this study, the effect of curcumin on experimental colitis-induced necroptosis of intestinal epithelial cells was investigated, and its molecular mechanism was further explored. We found that curcumin blocked necroptosis in a dose-dependent manner by inhibiting the phosphorylation of RIP3 and MLKL instead of RIP1 in HT-29 cells. Co-Immunoprecipitation assay showed that curcumin weakened the interaction between RIP1 and RIP3, possibly due to the direct binding of curcumin to RIP3 as suggested by drug affinity responsive target stability analysis. In a classical in vivo model of TNF-α and pan-caspase inhibitor-induced necroptosis in C57BL/6 mice, curcumin potently inhibited systemic inflammatory responses initiated by the necroptosis signaling pathway. Then, using a dextran sodium sulfate-induced colitis model in C57BL/6 mice, we found that curcumin inhibited the expression of p-RIP3 in the intestinal epithelium, reduced intestinal epithelial cells loss, improved the function of the intestinal tight junction barrier, and reduced local intestinal inflammation. Collectively, our findings suggest that curcumin is a potent targeted RIP3 inhibitor with anti-necroptotic and anti-inflammatory effects, maintains intestinal barrier function, and effectively alleviates colitis injury. curcumin necroptosis colitis intestinal epithelial cells RIP3 This work was supported by grants from the Natural Science Foundation of China (81973551 and 81872880), the Science and Technology Commission of Shanghai Municipality (21140905300), and the Shanghai Municipal Health Commission (ZY (2021-2023)-0203-04).section-at-acceptanceEthnopharmacology ==== Body pmc1 Introduction The incidence of inflammatory bowel disease (IBD) has shown an upward trend year by year, but the etiology and pathogenesis of IBD are still not fully understood. Ulcerative colitis (UC), as a typical IBD, is characterized by chronic and recurrent inflammation of the gastrointestinal tract and has become a refractory gastrointestinal disease due to the lack of effective treatment (Katz, 2002). Intestinal epithelial cells (IECs) play vital roles in host defense, maintenance of mucosal homeostasis, and immune responses (Eissa et al., 2019), among which apoptosis of IECs is an important early event in the progress of UC (Iwamoto et al., 1996). Most previous studies have reported that excessive IECs apoptosis, regulated by multiple signaling pathways, disrupts the epithelial barrier function, contributing to chronic inflammatory bowel diseases (Schulzke et al., 2009; Seidelin and Nielsen, 2009). In addition to apoptosis, there may be many other forms of cell death involved in UC. Of note is necroptosis, a caspase-independent programmed cell death, in which receptor-interacting serine/threonine protein kinases 1 and 3 (RIP1/3) play an important role (Cho et al., 2009; Zhang et al., 2009). It has been confirmed that altered intestinal epithelial necroptosis contributes to uncontrolled microbial translocation and amplifying inflammation (Vereecke et al., 2011; Günther et al., 2013; Pasparakis and Vandenabeele, 2015). The occurrence of necroptosis is positively correlated with the degree of intestinal injury. Inhibition of necroptosis by either chemical or genetic intervention can reduce intestinal injury and is considered a potential therapeutic strategy for the treatment of IBD. One potential approach to treating IBD is oral natural products that precisely target inflamed areas of the colon to reduce unwanted side effects and improve therapeutic efficacy. Curcumin, a yellow natural product isolated from the roots of the turmeric plant, displays multiple pharmacological properties, including anti-cancer, anti-oxidant, anti-bacterial, anti-inflammatory, and neuroprotective (Bandgar et al., 2012; Dai et al., 2013). Several preclinical studies have demonstrated that curcumin effectively protects the intestinal mucosa and repairs intestinal tissue function (Holt et al., 2005; Yue et al., 2019). Clinical studies further reveal that curcumin combined with conventional drugs can effectively maintain UC remission and prevent recurrence (Simadibrata et al., 2017; Sadeghi et al., 2020). In the dextran sulfate sodium (DSS)-induced colitis model, curcumin is also found to repair intestinal mucosa and significantly ameliorate intestinal inflammation (Camacho-Barquero et al., 2007; Villegas et al., 2011). Recently the effects of curcumin on necroptosis have garnered significant interest. Li et al. reported that curcumin has a protective effect on aflatoxin B1-induced liver necroptosis and inflammation (Li et al., 2022). Dai et al. found that curcumin protects primary cortical neurons from iron-induced neurotoxicity by attenuating necroptosis (Dai et al., 2013). Nevertheless, the regulatory effect of curcumin on colitis-associated intestinal epithelial necroptosis has yet to be reported. Therefore, the aim of this study is to investigate whether the protection of curcumin in DSS-induced colitis correlates with its inhibition of intestinal epithelial necroptosis. 2 Materials and methods 2.1 Reagents and primers Curcumin, Demethoxycurcumin, Bisdemethoxycurcumin, SM-164 Hydrochloride, Necrostain-1, and Cycloheximide were purchased from MedChemExpress; z-VAD-fmk was purchased from TargetMol; mouse-TNF-alpha and human-TNF-alpha were purchased from Novoprotein. Anti-RIP1(3493S), anti-phospho-RIP1 (65746s, human-specific), anti-phospho-RIP1(38662S, mouse-specific) were purchased from Cell Signaling Technology. Occludin (33–1,500) and ZO-1 (61–7,300) antibodies were purchased from Invitrogen, anti-GAPDH (ab181602), anti-RIP3/p-RIP3 (ab209384, human-specific), anti-MLKL (ab184718, human-specific), anti-phospho-MLKL (ab187091, human-specific), anti-MLKL (66675-1-Iq, mouse-specific), anti-phospho-MLKL (ab196436, mouse-specific), anti-RIP3(17563-1-AP, mouse-specific), anti-phospho-RIP3 (ab195117, mouse-specific) were purchased from Abcam. IRDye 800CW goat anti-mouse secondary antibody (926–32210) and 680RD goat anti-rabbit secondary antibody (926–68071) were obtained from LI-COR Biosciences. 2.2 In vivo experiment 2.2.1 Animal Four weeks of C57BL/6 mice were purchased from SIPPR-BK biochemistry Co. (Shanghai, China, SCXK 2018-0006). In a plastic cage (6 mice/cage), standard laboratory food and water were freely provided during the acclimation period and throughout the experiment. Animal experiments were approved by the Animal Care Committee of Shanghai Municipal Hospital of Traditional Chinese Medicine (2019SHL-KYYS-07). 2.2.2 Ulcerative colitis model induced by DSS Male C57BL/6 mice (20–23 g) at 6 weeks were randomly divided into five groups, including control group (CON), DSS group (DSS), curcumin 50 mg/kg group (CUR50), curcumin 100 mg/kg group (CUR100) and Necrostain 1 group (Nec-1), with 8 mice in each group. All mice were given 2% DSS (MP Biomedicals, 36,000–50,000 kDa) to induce UC for 10 days except the CON, which was given a normal diet and water. On the 11th day, curcumin group mice were intragastric administrated with 50 mg/kg and 100 mg/kg curcumin, respectively. Mice in the Nec-1 group were intraperitoneally injected with 5 mg/kg Necrostain-1 once a day, the DSS group was only given 95%CMC-Na+5% DMSO. The body weight, feces, and activity of mice were documented every day to calculate the disease activity index (DAI). On the 21st day, all mice were sacrificed by the cervical dislocation method, blood and colon tissue were collected. 2.2.3 Systemic inflammatory response syndrome (SIRS) C57BL/6 mice (6 weeks, 21–24 g) were randomly divided into Vehicle group (Vehicle), 50 mg/kg curcumin group (CUR50), 100 mg/kg curcumin group (CUR100) and Necrostain 1 group (Nec-1), with 12 mice in each group. 15 min before caudal vein injection of m-TNF-α, mice in each group were intraperitoneally injected with corresponding concentrations of curcumin and Necrostain-1, respectively. The Vehicle group was given the same dose of control solvent, and initial body temperature was recorded. 15 min later, the first dose of z-VAD-fmk (180 μg) was intraperitoneally injected. One hour after m-TNF-α was injected, the second dose of z-VAD-fmk (70 μg) was given. 2.3 Cell culture Human HT-29 cells (NCI-DTP Cat#HT-29), L929 cells (ECACC Cat# 141 12101), or EOL-1 cells (DSMZ Cat# ACC-386) were cultured in a high-glucose DMEM (Procell) containing 10% FBS (Biological Industries,04-001-1ACS), 100 U/mL streptomycin/penicillin, and 1% glutamine, Throughout the experiment, cells with no more than 10 generations were grown in an incubator with 5% CO2 at 37°C. 2.4 Induction of necroptosis, apoptosis, and cell viability assay HT-29 cells cultured in 96-well plates, each well containing 2 × 104 cells. Curcumin and Necrostain-1 were added to the corresponding concentration. SM-164 hydrochloride (10 nM) and caspase inhibitor z-VAD-fmk (20 μM) were joined together half an hour before the h-TNF-α (2 ng/100 μL), a total of 12 h of stimulation in the incubator. EOL-1 cells cultured in 96-well plates, each well was seeded with 2 × 105 cells. SM-164 hydrochloride (10 nM) and caspase inhibitor z-VAD-fmk (20 μM) were mixed with the corresponding concentrations of curcumin, and half an hour later h-TNF-α (2 ng/100 μL) was added for 6 h of co-stimulation. L929 cells were seeded in 96-well plates at 2 × 105 cells per well, and after overnight growth, caspase inhibitor z-VAD-fmk (20 μM) was mixed with the corresponding concentration of curcumin using medium, and after half an hour h-TNF-α (2 ng/100 μL) was added for co-stimulation for 6 h. TCZ experiment is the use of cycloheximide (10 nM) and caspase inhibitor z-VAD-fmk (20 μM) and h-TNF-α (2 ng/100 μL) induced by 12 h. Apoptosis was induced by TS (h-TNF-α plus smac mimetic) or TC (h-TNF-α plus cycloheximide) stimulation for 24 h respectively. Cells were stimulated with a specific time, 100 μL CellTiter-LumiTM luminescent reagents (Beyotime, C0065L) were added to each well, 37°C incubation after 10 min, every well liquid was transferred to 96-well board, using SpectraMaxM5 microplate reader to measure luminescent value. 2.5 Western blot 2 × 104 human colon cancer HT-29 cells grew to 90% in 6-well plates. Curcumin was pre-administered with different concentrations and then stimulated by TSZ (h-TNF-α + SM-164 + Z-VAD-FMK) to observe the anti-necrosis activity of curcumin. In another experiment, TSZ was used to stimulate for 2 h, 4 h, and 6 h respectively to observe the anti-necrosis effect of curcumin at different time points. The cells were lysed with NP40 (Beyotime, P0013F) to obtain supernatant. After the concentration was determined by BCA, Western blot was performed to detect the phosphorylated and total protein expression of RIP1, RIP3 and MLKL. The 20 μg sample was isolated by 10% SDS-PAGE and transferred to NC membrane, sealed with 5% Not-fat milk for 1 h, and then added with 5% Not-fat milk diluted primary antibody (1:1,000) overnight at 4°C. On the second day, a fluorescent secondary antibody (1:8,000) was incubated in the dark for 1 h. Western blots were displayed and analyzed by Image Studio Ver 5. 2 system. 2.6 Immunoprecipitation Human colon cancer HT-29 cells stimulated by TSZ were lysed with NP40 to obtain supernatant. BCA Protein Assay Kit (Beyotime, P0010) was used to determine protein concentration in cell lysate. PBS was added to keep the concentration consistent, and RIP1 primary antibody was joined and incubated at 4°C in a shaker for 24 h. Magnetic beads add rec-ProteinG-Sepharose (Thermo Fisher Scientific, 101242) 4°C for 24 h incubation, add 2x loading acquired immune coprecipitation protein buffer, 100°C, and degeneration, using Western blot detection RIP1 with the combination of RIP3. 2.7 RNA extraction and quantitative real-time PCR Colonic tissue RNA was obtained using RNA isoPlus reagent (TaKaRa, 9109) PrimeScript RT Master mixing kit (TaKaRa, RR036A). Power qPCR SYBR Green master mixture (Thermo Fisher Scientific, A25742) was performed for real-time fluorescence quantitative PCR on an instrument (Roche, LightCyccler96). The 2−ΔΔCt method was used to calculate the relative mRNA expression. Primer information is as follows: IL-1β, 5′-GCA​ACT​GTT​CCT​GAA​CTC​AAC​T-3′ and 5′-ATC​TTT​TGG​GGT​CCG​TCA​ACT-3′; IL-6, 5′-TAGTC CTTCCT AC CC C AATTTCC′ and 5′-TAG​TCC​TTC​CTA​CCC​CAA​TTT​CC-3′; CCR2, 5′-AT C CAC​GGC​ATA​CTA​TCA​ACA​TC′ and 5′-CAAGG CTCACCATC ATC GT A G′; CCR6, 5′-GTT CAA​CTT​TAA​CTG​TGG​GAT​G-3′ and 5′-GGTGT CT C ACCATC ATC GT A G′; GAPDH, 5′AGGCGAGGACTTTCT-3′ and 5′-GG GGTCG TTGATGGCAACA-3′. 2.8 Drug affinity responsive target stability assay The supernatant protein solution was collected after the cells were lysed with NP40 solution. The protein solution was evenly divided into the drug administration group and the blank control group. Curcumin was added to the drug administration group, and the same amount of DMSO was joined to the blank control group. The samples were incubated for 1 h at room temperature, followed by the corresponding dose of pronase protease (Roche, 10165921001) incubated on ice for 20 min and heated with 5xloading buffer. Using Western blot methods to analyze the samples. 2.9 FITC-dextran assay The mice fasted for 4 h, and then orally gavage 600 mg/kg FITC-dextran (Sigma, FD4, mol wet:3,000–5,000) 4 h before death. After blood collection, the supernatant was obtained by centrifugation and at 1:5 diluted with PBS. Finally, the serum FITC-dextran levels were detected by SpectraMax M5 fluorescence assay (Ex: 488 nm, Em: 525 nm). 2.10 Immunofluorescence Animal tissue specimens and paraffin-embedded intestinal sections were divided into 1 μm sections. The epithelial monolayer paraffin was removed and immunostained. The tissue slides were incubated overnight with antibodies at 4°C, and after 4 rinses with PBS, the second antibody was incubated for 1 h away from light. Next, stain with DAPI (Beyotime, C1005) for 10 min at room temperature. After the cells were rinsed with PBS, an anti-fluorescence quencher was dropped (Beyotime, P0191-3). Representative results were photographed by confocal microscopy (C2+, Nikon). 2.11 HE and TUNEL staining 4% paraformaldehyde was performed to fix with colon tissue. After dehydration, the tissue was embedded with paraffin, then sliced into 4 mm sections and stained with hematoxylin and eosin (G1005) and TUNEL (GB1502). Before staining, xylene was used to remove paraffin from sections. Finally, a microscopic examination (NIKON ECLIPSE TI-SR, Japan) was performed, and the images were collected and analyzed. 2.12 Statistical analysis Statistical analysis was performed by GraphPad Prism 8.0. An ANOVA and Tukey’s post hoc tests were performed to evaluate experiments involving multiple groups. P < 0.05 was set as statistically significant in all figures. All data were represented by mean ± SD. Statistical tests and the number of repetitions (n) were given in the figure legends. ImageJ 8.0 was used to calculate the gray value of each Western blot. The mechanism diagram is drawn using Figdraw software. 3 Results 3.1 Curcumin inhibits necroptosis in HT-29 cell line The active ingredients of curcuminoid compounds extracted from turmeric mainly include curcumin (CUR), demethoxycurcumin (DMC), and bisdemethoxycurcumin (BDMC). Among them, curcumin (Dai et al., 2013; Lee et al., 2021; Chung et al., 2022; Li et al., 2022)and BDMC (Córdoba-David et al., 2020) have been reported to have anti-necroptosis activity. To further assess the anti-necroptosis activity of the three (Figure 1A), we first examined cell viability after pretreatment with these three curcuminoids compounds in TSZ (h-TNF-α + SM-164 + Z-VAD-FMK) induced necroptosis in human colon cancer HT-29 cells. The results showed that three curcuminoids compounds were both efficient in blocking TSZ-induced necroptosis in HT-29 cells (Figure 1B). Since curcumin exhibited the strongest anti-necroptotic activity, we further evaluated its biological effects in subsequent experiments. (Figure 1B). FIGURE 1 The anti-necroptotic activities of curcumin compounds were compared. (A) The chemical structural formula of curcumin, dimethoxycurcumin, and bisdemethoxycurcumin. (B)HT-29 cells were pretreated with curcumin, demethoxycurcumin, and bisdemethoxycurcumin, Necrostain-1 was used as a positive control, followed by a final concentration of 20 ng/ml h-TNF-α (T), 10 nM SM-164 Hydrochloride (S), and 20 mM z-VAD-fmk (Z) to induce necroptosis. (***p < 0.001, compared with TSZ group; ###p < 0.001, compared with CUR group). Curcumin could dose-dependently inhibit TSZ (Figure 2A) or TCZ (h-TNF-α + cycloheximide + Z-VAD-FMK) -induced necroptosis. The same anti-necroptosis results were demonstrated in EOL-1 (Supplementary Figure S1A, B) and L929 (Supplementary Figure S2A, B) cell lines, which have been identified as mature necroptotic cellular models (Figure 2B). Besides, we also found that curcumin showed weak protective effects at high concentrations in TS (h-TNF-α plus Smac mimetic) or TC (hT-NF-α plus cycloheximide) induced apoptosis model (Figures 2C,D). Moreover, we confirmed that curcumin alone had no cytotoxicity on HT-29 cells within 24 h at doses lower than 30 μM (Figure 2E). Collectively, these results indicated that curcumin effectively inhibited activation of necroptosis in vitro. FIGURE 2 In vitro anti-necroptosis pharmacological activity of curcumin. (A, B) In TSZ (h-TNF-α + SM-164 + Z-VAD-FMK) and TCZ (h-TNF-α + cycloheximide + Z-VAD-FMK) assay, curcumin protected HT-29 cells in a dose-dependent manner. (C, D) Effects of curcumin on HT-29 cells in TS (h-TNF-α + smac mimetic) and TC (h-TNF-α + cycloheximide) induced apoptosis experiments. (E) Curcumin had no obvious toxic effect on HT-29 cells within 24 h (###p < 0.001, compared with CON group; *p < 0.05, **p < 0.01, ***p < 0.001 compared with the model group as indicated in the figure). 3.2 Curcumin inhibits necroptosis in vitro by selectively targeting RIP3 The RIP1-RIP3-MLKL signaling pathway has been reported to be involved in the induction of necroptosis (Yang et al., 2016). Therefore, we examined the protein expression and phosphorylation of RIP1, RIP3, and MLKL after curcumin treatment in TSZ-induced HT-29 cells, aiming to elucidate the inhibitory effect of curcumin on the necroptosis pathway. Firstly, Western blotting showed that curcumin treatment dose-dependently downregulated the phosphorylation of RIP3 (p-RIP3) and phosphorylation of MLKL (p-MLKL), but had no influence on total protein expression of RIP3 (t-RIP3) and MLKL (t-MLKL) (Figure 3A). Interestingly, curcumin had no significant influence on the expression of p-RIP1 or t-RIP1 (Figure 3A). Furthermore, we found that curcumin could completely inhibit the p-RIP3 at 30 μM within 6 h. Subsequently, the downstream p-MLKL was also markedly inhibited (Figure 3B). FIGURE 3 Molecular mechanism of curcumin blocking the activation of the necroptosis signaling pathway. (A) Curcumin inhibited p-RIP3 and p-MLKL in a dose-dependent manner. Statistical analysis of the gray value showed three independent repeated experiments. (B) The inhibitory effect of curcumin on p-RIP1, p-RIP3, and p-MLKL was observed using TSZ stimulation at different time points. (C) HT-29 cells were treated with DMSO or curcumin (30 µM) for 6 h. Cell lysates were immunoprecipitated with an anti-RIP1 antibody (IP: RIP1) and subjected to immunoblot analysis with the indicated antibodies. (D) Cell lysates from HT-29 were incubated with curcumin (200 µM) and DMSO for 1 hour, respectively, followed by the addition of 0.05% and 0.03% protease. The expression of RIP1 and RIP3 was then examined by immunoblotting. (*p < 0.05, compared with TSZ group). The formation of RIP1/3 complex bodies is necessary for the development of necroptosis (Cho et al., 2009). Therefore, we further evaluated the interaction between RIP1 and RIP3 by co-immunoprecipitation in curcumin-treated HT-29 cells. As compared with the DMSO-treated cells, the level of co-immunoprecipitated RIP3 decreased significantly in the curcumin-treated cells when RIP1 was immunoprecipitated with its antibody (Figure 3C). This suggested that curcumin could block TSZ-induced necrosome formation for inhibiting the necroptosis process. To examine whether curcumin directly interacted with RIP3, we performed drug affinity responsive target stability assay (DARTS) (Chen et al., 2019) to detect the potential interaction between curcumin and RIP3. In the sample of curcumin-treated HT-29 cells. It was found that RIP3 but not RIP1 was protected from protease digestion (Figure 3D), indicating that curcumin may prefer to interact with RIP3. Overall, these results suggested that curcumin may be a selective inhibitor of RIP3. 3.3 Curcumin inhibits systemic inflammatory response syndrome (SIRS) in vivo To evaluate the anti-necroptosis activity of curcumin in vivo, we established a SIRS model characterized by hypothermia (Duprez et al., 2011). After intraperitoneal injection of z-VAD-fmk (180 μg) and curcumin (50 and 100 mg/kg) for 15 min, m-TNF-α (65 ug/kg) was injected via the tail vein into the mice. Half an hour later, a second dose of z-VAD-fmk (70 μg) was administered. The results showed that both Necrostatin-1 (Nec-1, RIP1-targeted inhibitor) and curcumin significantly protected mice from extremely low hypothermia within 20 h compared with the vehicle group (Figure 4A). Similarly, prophylactic curcumin greatly increased the survival of mice in a dose-dependent manner. Consistent with the Nec-1 group, the percent survival of mice in the CUR100 group was 100% and completely protected SIRS mice (Figure 4B). Moreover, curcumin displayed comparable levels of anti-necroptosis effects in female mice as it did in male mice (Supplementary Figure S3A, B). FIGURE 4 Curcumin potently inhibited systemic inflammatory responses initiated by the necroptosis signaling pathway in an in vivo model. (A) Body temperature of C57BL/6 mice (n = 6) injected with m-TNF-α (65 μg/kg) and z-VAD-fmk (250 µg) after treatment with indicated doses of curcumin (100 mg/kg) and the survival curves (B). (C) The expression of necroptosis biomarkers in colon tissue was analyzed by Western blot. And quantitative analysis of gray scales of protein bands. (D) Quantitative analysis of inflammatory cytokine RNA in mouse colon tissue. (**p < 0.01, ***p < 0.001, compared with Vehicle Group). We next sought to demonstrate whether curcumin prevents the occurrence of necroptosis by inhibiting RIP3 in vivo, we extracted mouse colon tissue proteins for immunoblotting, and the outcomes displayed that curcumin remarkably inhibited the expression of p-RIP3 and p-MLKL (Figure 4C). Moreover, the real-time quantitative PCR experiments showed that curcumin significantly reduced the mRNA levels of IL-1β and IL-6 in intestinal tissues (Figure 4D). These findings suggested that curcumin pretreatment alleviated symptoms, death rates, and the production of inflammatory cytokines in necroptosis-characteristic SIRS mice. 3.4 Curcumin alleviates DSS-induced ulcerative colitis in vivo Curcumin has been reported to protect against DSS-induced UC (Holt et al., 2005; Camacho-Barquero et al., 2007), we speculated that curcumin may alleviate UC by inhibiting intestinal epithelial cell necroptosis. Therefore, we performed 2% DSS in C57BL/6 mice for 10 days to induce UC, followed by intragastric administration of curcumin to mice daily, intraperitoneal injection of Nec-1 as a positive control and evaluated the therapeutic efficacy of curcumin (Figure 5A). We observed that the trend of weight loss was curbed on the fourth day of treatment, and the body weight of the mice treated with curcumin was significantly restored as compared to that in the DSS group from the 11th day to the 21st day (Figure 5B). We measured clinical signs of DSS-induced UC in each group of mice by disease activity index (DAI). The index consists of three parameters (fecal consistency, rectal bleeding, and body weight changes of mice) and provides the order of severity of the inflammatory process associated with intestinal mucosal damage. As expected, the DAI of curcumin-treated mice was decreased compared to the DSS control group (Figure 5C). FIGURE 5 Curcumin alleviates symptoms of DSS-induced colitis. (A) Experimental protocol for the treatment of ulcerative colitis with curcumin (n = 6-8). (B) Body weight change. (C) Disease activity index (DAI). (D) FITC content of blood from mice 4 h after gavage. (E) Analysis of colon length and colon condition of mice in each group on the last day. (F) Representative images of H&E staining of mouse colon as well as histologic scores. (G) Quantitative analysis of inflammatory cytokine RNA in mouse colon tissue. (##p < 0.01, ###p < 0.001, compared with CON group; *p < 0.05, **p < 0.01, ***p < 0.001, compared with DSS group). In addition, as compared to the serum of DSS mice, the level of FITC-dextran in that of curcumin-treated mice was also robustly lower (Figure 5D), indicating that curcumin protected intestinal barrier function in DSS-induced colitis mice, with an effect comparable to that of Nec-1. We examined colon length after 10 days of curcumin treatment as it is a well-recognized indicator of DSS-induced colitis. As expected, administration of curcumin significantly prevented DSS-induced colon shortening, robustly ameliorated the ulcer area, reduced loss of mucosal epithelium and goblet cells, improved inflammatory infiltration, and overall decreased histopathological score (Figures 5E,F). Moreover, the mRNA expression of inflammatory cytokines (IL-6 and IL-1β) and chemokine receptors (Ccr-2 and Ccr-6) were significantly reduced through curcumin treatment (Figure 5G). Together, our results suggest that curcumin is effective in relieving DSS-induced colonic injury. 3.5 Curcumin ameliorates the loss of intestinal barrier function following DSS-induced colitis It was demonstrated that administration of DSS resulted in intestinal epithelial barrier (IEB) dysfunction in mice, mainly manifested by increased paracellular permeability and decreased transepithelial electrical resistance (TER). In order to further examine the effect of curcumin on intestinal epithelial barrier function, we observed the death of intestinal epithelial cells in different groups of mice by TUNEL staining and found that curcumin can effectively reduce cell death in DSS-induced mice (Figure 6A). Subsequently, we performed occludin and Zonula occludens-1 (ZO-1) as tight junction (TJs) protein markers to test the IEB function of intestinal epithelial cells. The results of both immunofluorescence and Western blotting showed that curcumin treatment greatly raised the protein levels of occludin and ZO-1, but also decreased the expression of p-RIP3 and p-MLKL in the downstream pathway (Figures 6B,C). The above results indicated that curcumin could effectively inhibit the necroptosis pathway by reducing the p-RIP3 while enhancing the expression of occludin and ZO-1 in IECs to maintain intestinal epithelial barrier function. FIGURE 6 Curcumin inhibits intestinal epithelial necroptosis activation and tight junction barrier damage in mice with colitis. (A) Analysis of intestinal epithelial cell injuries by TUNEL. (B) Immunofluorescence staining showed intestinal epithelial TJ proteins including ZO-1 and occludin, and necroptosis-related protein p-RIP3. (C) The TJ proteins ZO-1 and occludin and necroptosis-related proteins p-RIP1, p-RIP3, and p-MLKL were detected by Western blot. And quantitative analysis of gray scales of protein bands. (##p < 0.01, ###p < 0.001, compared with CON group; *p < 0.05, **p < 0.01, ***p < 0.001, compared with DSS group). 4 Discussion In the present study, we found that the inhibitory effect of curcumin on necroptosis in intestinal epithelial cells may be associated with targeting RIP3. In an experimental colitis model, curcumin ameliorated necroptosis-associated intestinal epithelial cell loss, local intestinal inflammation, and intestinal barrier permeability damage associated with tight junction injury (Figure 7). These results suggest that curcumin may be a potential therapeutic agent for ulcerative colitis as an inhibitor of RIP3. FIGURE 7 Curcumin alleviates experimental colitis in mice by suppressing necroptosis of intestinal epithelial cells. Curcumin inhibits IECs necroptosis by targeting RIP3 to ameliorate intestinal epithelial cell loss, local intestinal inflammation, and intestinal epithelial tight junction barrier injury. Turmeric has been used as an herbal remedy for various ailments for thousands of years. Curcuminoids are active ingredients extracted from the rhizome of Zingiberaceae (Curcuma longa) (Mazzio et al., 2016; Jeong, 2017; Sun et al., 2019), mainly including curcumin (71.5%), demethoxycurcumin (19.4%), and bisdemethoxycurcumin (9.1%), which are all diarylheptanoids (Pfeiffer et al., 2003; Hadi, 2018). The curcumin molecule contains multiple double bonds, as well as active groups such as phenolic hydroxyl and carbonyl, so it has strong physiological activity. Among them, curcumin is the most studied active ingredient, and it is also the major compound of curcuminoids to exert their medicinal values. The chemical properties of curcumin have an important influence on its biological activity. Several studies proposed the key role of methoxy groups and thiol-reactive α, ß unsaturated carbonyl groups of curcuminoids (Yang et al., 2017). We speculate that the biological ability of curcuminoids may be positively correlated with the number of methoxy groups. In our results, compared to DMC or BDMC, there were stronger inhibitory effects in curcumin treatment for TSZ-induced cell death, which implies that curcumin may have a greater value for the prevention and treatment of UC. However, the major obstacle to the clinical efficacy of curcumin is its poor bioavailability. Studies have shown that curcumin nanoformulations such as Polylactic-co-glycolic acid (PLGA) or hydroxypropyl beta-cyclodextrin (HP-β-CD) encapsulated curcumin significantly enhanced bioavailability after oral administration in rats (Shaikh et al., 2009). In addition, synthetic analogues can also enhance the solubility and stability of curcumin while preserving its pharmacological properties. The potential medicinal values of curcumin have been widely recognized in terms of diseases with the discovery of its biological functions (Akbar et al., 2016). Moreover, it plays a vital role in protecting the intestinal mucosa and repairing intestinal epithelial barrier function (Holt et al., 2005; Shen and Durum, 2010). It has been reported that curcumin treatment effectively relieved experimental colitis by modulating the re-equilibration of Th17/Treg, closely related to the regulation of the IL-23/Th17 pathway (Wei et al., 2021). It was proved that curcumin could strongly alleviate DSS-induced experimental colitis by inhibiting NLRP3 inflammasome activation (Gong et al., 2018). Zhong et al. also stated that curcumin availably ameliorated DSS-induced UC through regulatory mechanism related to immune memory homeostasis of T-cells (Zhong et al., 2020). While the potential molecular modulatory effect of curcumin on necroptosis in DSS-induced UC is still unclear. An insight into the mechanisms of necroptosis in UC may provide new therapeutic strategies for the prevention and treatment of IBD. Necrotizing apoptosis is a regulated form of programmed cell death independent of caspase, induced by cytokines, toll-like receptors, oxidative stress, or other death receptors (Dunai et al., 2011; Vanden Berghe et al., 2014; Humphries et al., 2015). Recent insights into the molecular mechanism of TNF-induced necroptosis (Vanlangenakker et al., 2011) reveal that RIP3, a critical factor for necroptosis, together with RIP1 and MLKL proteins, constitutes the main necroptosis signaling pathway (Sun et al., 2012; Zhang et al., 2016). When apoptosis is restrained, activated RIP1 binds to RlP3, generating a necrosome complex, followed by recruiting and phosphorylating MLKL. Then, the p-MLKL induces its oligomerization and translocation to the plasma membrane, causing necroptotic cell death by activating ion channels or promoting the formation of pore structures (Murphy et al., 2013; Dondelinger et al., 2014). Thus, necroptosis is generally considered to be a pro-inflammatory death on account of the release of intracellular immunostimulatory components following with cell lysis (Chan et al., 2015; Silke et al., 2015; Shlomovitz et al., 2017). Recent evidence points to necroptosis involved in ischemic injury (Linkermann et al., 2012; Luedde et al., 2014), liver injury (Dara et al., 2016), systemic inflammatory syndrome (Meng et al., 2015), and inflammatory bowel disease (Günther et al., 2011). Notably, Pierdomenico, M. et al. reported that both RIP3 and MLKL are highly expressed in children with IBD, who are often accompanied by symptoms of intestinal mucosal inflammation (Pierdomenico et al., 2014). These results above imply that RlP3-dependent necroptosis may functionally involve in the pathogenesis of colitis. Here, we show that curcumin pretreatment remarkably decreased the expression of p-RIP3 and p-MLKL in TNF-α-induced SIRS mice. Interestingly, curcumin does not affect RIP1 protein or phosphorylation levels either in vivo or in vitro. Similar to ours, Moujalled, D. M. et al. stated that in the absence of RIP1, TNF-α can still activate RIP3 and cause necroptosis (Moujalled et al., 2013). Combined with the results reported by Newton, K. et al. that “tissue-specific RIP3 deletion identified intestinal epithelial cells as the major target organ” (Newton et al., 2016), it shows that curcumin selectively inhibits necroptosis by targeting RIP3 to achieve the purpose of treating UC. Additionally, the intestinal mechanical barrier is mainly composed of the mucus layer, intestinal epithelial cells and intercellular TJs, which is the basis for maintaining the structure and function of the intestinal mucosal barrier (Zeisel et al., 2019). IBD is often accompanied by abnormal expression of TJs (Tanaka et al., 2015), so we also investigate curcumin’s involvement in probable mechanisms of intestinal epithelial cell permeability. We speculate that a further mechanism may be to disrupt the epithelial barrier by inhibiting the expression of junctional molecules, promoting the penetration of bacteria and cytotoxic substrates into intestinal mucosa, thereby exacerbating inflammation in vivo. Therefore, we analyze the curative effect of curcumin on the expression of TJs in the DSS-induced UC model by immunofluorescence and Western blotting. Here, we find a robust increase of occludin and ZO-1 levels with curcumin treatment in undergoing necroptosis, which confirms the close relationship between RIP3-dependent necroptosis and intestinal epithelial barrier dysfunction (Ibrahim et al., 2020). In conclusion, our study shows that curcumin has a positive role in colitis-induced intestinal epithelial injury by inhibiting the expression of p-RIP3 and p-MLKL by targeting RIP3, decreasing inflammatory cytokines expression, and inducing intestinal epithelial barrier integrity. We demonstrated that curcumin treatment may be an effective therapeutic approach for the management or collaborative treatment of IBD, including UC in the future. Data availability statement The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors. Ethics statement The animal study was reviewed and approved by The Animal Care Committee of the Shanghai Municipal Hospital of Traditional Chinese Medicine. Author contributions ZW, YT and LZ contributed to the conception and design of the study; YZ, LC, WZ and XL performed the experiments; YZ and SY contributed to the acquisition, analysis, and interpretation of data; SY, YZ and ZW drafted the manuscript; LZ, ZW and YT supervised the project; YT and LZ revised the manuscript critically for important intellectual content. All authors have approved the final vision of this manuscript. Conflict of interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Publisher’s note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. Supplementary material The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fphar.2023.1170637/full#supplementary-material Click here for additional data file. ==== Refs References Akbar A. Kuanar A. Joshi R. K. Sandeep I. S. Mohanty S. Naik P. K. (2016). Development of prediction model and experimental validation in predicting the curcumin content of turmeric (curcuma longa L.). Front. Plant Sci. 7 , 1507. 10.3389/fpls.2016.01507 27766103 Bandgar B. P. Hote B. S. Jalde S. S. Gacche R. N. (2012). Synthesis and biological evaluation of novel curcumin analogues as anti-inflammatory, anti-cancer and anti-oxidant agents. Med. Chem. Res. 21 (10 ), 3006–3014. 10.1007/s00044-011-9834-7 Camacho-Barquero L. Villegas I. Sanchez-Calvo J. M. Talero E. Sanchez-Fidalgo S. Motilva V. (2007). Curcumin, a Curcuma longa constituent, acts on MAPK p38 pathway modulating COX-2 and iNOS expression in chronic experimental colitis. Int. Immunopharmacol. 7 (3 ), 333–342. 10.1016/j.intimp.2006.11.006 17276891 Chan F. K. M. Luz N. F. Moriwaki K. (2015). Programmed necrosis in the cross talk of cell death and inflammation. Annu. Rev. Immunol. 33 (1 ), 79–106. 10.1146/annurev-immunol-032414-112248 25493335 Chen X. Zhuang C. Ren Y. Zhang H. Qin X. Hu L. (2019). Identification of the Raf kinase inhibitor TAK-632 and its analogues as potent inhibitors of necroptosis by targeting RIPK1 and RIPK3. Br. J. Pharmacol. 176 (12 ), 2095–2108. 10.1111/bph.14653 30825190 Cho Y. S. Challa S. Moquin D. Genga R. Ray T. D. Guildford M. (2009). Phosphorylation-driven assembly of the RIP1-RIP3 complex regulates programmed necrosis and virus-induced inflammation. Cell. 137 (6 ), 1112–1123. 10.1016/j.cell.2009.05.037 19524513 Chung H. Lee S. W. Hyun M. Kim S. Y. Cho H. G. Lee E. S. (2022). Curcumin blocks high glucose-induced podocyte injury via RIPK3-dependent pathway. Front. Cell. Dev. Biol. 10 , 800574. 10.3389/fcell.2022.800574 35706905 Córdoba-David G. Duro-Castano A. Castelo-Branco R. C. Gonzalez-Guerrero C. Cannata P. Sanz A. B. (2020). Effective nephroprotection against acute kidney injury with a star-shaped polyglutamate-curcuminoid conjugate. Sci. Rep. 10 (1 ), 2056. 10.1038/s41598-020-58974-9 32029842 Dai M. C. Zhong Z. H. Sun Y. H. Sun Q. F. Wang Y. T. Yang G. Y. (2013). Curcumin protects against iron induced neurotoxicity in primary cortical neurons by attenuating necroptosis. Neurosci. Lett. 536 , 41–46. 10.1016/j.neulet.2013.01.007 23328441 Dara L. Liu Z. X. Kaplowitz N. (2016). Questions and controversies: The role of necroptosis in liver disease. Cell. Death Discov. 2 , 16089. 10.1038/cddiscovery.2016.89 27924226 Dondelinger Y. Declercq W. Montessuit S. Roelandt R. Goncalves A. Bruggeman I. (2014). MLKL compromises plasma membrane integrity by binding to phosphatidylinositol phosphates. Cell. Rep. 7 (4 ), 971–981. 10.1016/j.celrep.2014.04.026 24813885 Dunai Z. Bauer P. I. Mihalik R. (2011). Necroptosis: Biochemical, physiological and pathological aspects. Pathol. Oncol. Res. 17 (4 ), 791–800. 10.1007/s12253-011-9433-4 21773880 Duprez L. Takahashi N. Van Hauwermeiren F. Vandendriessche B. Goossens V. Vanden Berghe T. (2011). RIP kinase-dependent necrosis drives lethal systemic inflammatory response syndrome. Immunity 35 (6 ), 908–918. 10.1016/j.immuni.2011.09.020 22195746 Eissa N. Hussein H. Diarra A. Elgazzar O. Gounni A. S. Bernstein C. N. (2019). Semaphorin 3E regulates apoptosis in the intestinal epithelium during the development of colitis. Biochem. Pharmacol. 166 , 264–273. 10.1016/j.bcp.2019.05.029 31170375 Gong Z. Zhao S. Zhou J. Yan J. Wang L. Du X. (2018). Curcumin alleviates DSS-induced colitis via inhibiting NLRP3 inflammsome activation and IL-1β production. Mol. Immunol. 104 , 11–19. 10.1016/j.molimm.2018.09.004 30396035 Günther C. Martini E. Wittkopf N. Amann K. Weigmann B. Neumann H. (2011). Caspase-8 regulates TNF-α-induced epithelial necroptosis and terminal ileitis. Nature 477 (7364 ), 335–339. 10.1038/nature10400 21921917 Günther C. Neumann H. Neurath M. F. Becker C. (2013). Apoptosis, necrosis and necroptosis: Cell death regulation in the intestinal epithelium. Gut 62 (7 ), 1062–1071. 10.1136/gutjnl-2011-301364 22689519 Hadi S. (2018). “Curcuminoid content of Curcuma longa L. and Curcuma xanthorrhiza rhizome based on drying method with NMR and HPLC-UVD,” in IOP conference series: Materials science and engineering (IOP Publishing). Holt P. R. Katz S. Kirshoff R. (2005). Curcumin therapy in inflammatory bowel disease: A pilot study. Dig. Dis. Sci. 50 (11 ), 2191–2193. 10.1007/s10620-005-3032-8 16240238 Humphries F. Yang S. Wang B. Moynagh P. N. (2015). RIP kinases: Key decision makers in cell death and innate immunity. Cell. Death Differ. 22 (2 ), 225–236. 10.1038/cdd.2014.126 25146926 Ibrahim S. Zhu X. Luo X. Feng Y. Wang J. (2020). PIK3R3 regulates ZO-1 expression through the NF-kB pathway in inflammatory bowel disease. Int. Immunopharmacol. 85 , 106610. 10.1016/j.intimp.2020.106610 32473571 Iwamoto M. Koji T. Makiyama K. Kobayashi N. Nakane P. K. (1996). Apoptosis of crypt epithelial cells in ulcerative colitis. J. Pathol. 180 (2 ), 152–159. 10.1002/(SICI)1096-9896(199610)180:2<152:AID-PATH649>3.0.CO;2-Y 8976873 Jeong H. J. (2017). Antioxidant activities and protective effects of hot water extract from Curcuma longa L. on oxidative stress-induced C2C12 myoblasts. J. Korean Soc. Food Sci. Nutr. 46 (11 ), 1408–1413. Katz J. A. (2002). Advances in the medical therapy of inflammatory bowel disease. Curr. Opin. Gastroenterol. 18 (4 ), 435–440. 10.1097/00001574-200207000-00007 17033318 Lee Y. J. Park K. S. Lee S. H. (2021). Curcumin targets both apoptosis and necroptosis in acidity-tolerant prostate carcinoma cells. Biomed. Res. Int. 2021 , 8859181. 10.1155/2021/8859181 34095313 Li S. Liu R. Xia S. Wei G. Ishfaq M. Zhang Y. (2022). Protective role of curcumin on aflatoxin B1-induced TLR4/RIPK pathway mediated-necroptosis and inflammation in chicken liver. Ecotoxicol. Environ. Saf. 233 , 113319. 10.1016/j.ecoenv.2022.113319 35189522 Linkermann A. Brasen J. H. Himmerkus N. Liu S. Huber T. B. Kunzendorf U. (2012). Rip1 (Receptor-interacting protein kinase 1) mediates necroptosis and contributes to renal ischemia/reperfusion injury. Kidney Int. 81 (8 ), 751–761. 10.1038/ki.2011.450 22237751 Luedde M. Lutz M. Carter N. Sosna J. Jacoby C. Vucur M. (2014). RIP3, a kinase promoting necroptotic cell death, mediates adverse remodelling after myocardial infarction. Cardiovasc. Res. 103 (2 ), 206–216. 10.1093/cvr/cvu146 24920296 Mazzio E. A. Li N. Bauer D. Mendonca P. Taka E. Darb M. (2016). Natural product HTP screening for antibacterial (E.coli 0157:H7) and anti-inflammatory agents in (LPS from E. coli O111:B4) activated macrophages and microglial cells; focus on sepsis. BMC Complement. Altern. Med. 16 (1 ), 467. 10.1186/s12906-016-1429-x 27846826 Meng L. Jin W. Wang X. (2015). RIP3-mediated necrotic cell death accelerates systematic inflammation and mortality. Proc. Natl. Acad. Sci. U. S. A. 112 (35 ), 11007–11012. 10.1073/pnas.1514730112 26283397 Moujalled D. M. Cook W. D. Okamoto T. Murphy J. Lawlor K. E. Vince J. E. (2013). TNF can activate RIPK3 and cause programmed necrosis in the absence of RIPK1. Cell. Death Dis. 4 (1 ), e465. 10.1038/cddis.2012.201 23328672 Murphy J. M. Czabotar P. E. Hildebrand J. M. Lucet I. S. Zhang J. G. Alvarez-Diaz S. (2013). The pseudokinase MLKL mediates necroptosis via a molecular switch mechanism. Immunity 39 (3 ), 443–453. 10.1016/j.immuni.2013.06.018 24012422 Newton K. Dugger D. L. Maltzman A. Greve J. M. Hedehus M. Martin-McNulty B. (2016). RIPK3 deficiency or catalytically inactive RIPK1 provides greater benefit than MLKL deficiency in mouse models of inflammation and tissue injury. Cell. Death Differ. 23 (9 ), 1565–1576. 10.1038/cdd.2016.46 27177019 Pasparakis M. Vandenabeele P. (2015). Necroptosis and its role in inflammation. Nature 517 (7534 ), 311–320. 10.1038/nature14191 25592536 Pfeiffer E. Hohle S. Solyom A. M. Metzler M. (2003). Studies on the stability of turmeric constituents. J. food Eng. 56 (2-3 ), 257–259. 10.1016/s0260-8774(02)00264-9 Pierdomenico M. Negroni A. Stronati L. Vitali R. Prete E. Bertin J. (2014). Necroptosis is active in children with inflammatory bowel disease and contributes to heighten intestinal inflammation. Am. J. Gastroenterol. 109 (2 ), 279–287. 10.1038/ajg.2013.403 24322838 Sadeghi N. Mansoori A. Shayesteh A. Hashemi S. J. (2020). The effect of curcumin supplementation on clinical outcomes and inflammatory markers in patients with ulcerative colitis. Phytotherapy Res. 34 (5 ), 1123–1133. 10.1002/ptr.6581 Schulzke J. D. Ploeger S. Amasheh M. Fromm A. Zeissig S. Troeger H. (2009). Epithelial tight junctions in intestinal inflammation. Ann. N. Y. Acad. Sci. 1165 , 294–300. 10.1111/j.1749-6632.2009.04062.x 19538319 Seidelin J. B. Nielsen O. H. (2009). Epithelial apoptosis: Cause or consequence of ulcerative colitis? Scand. J. gastroenterology 44 (12 ), 1429–1434. 10.3109/00365520903301212 Shaikh J. Ankola D. D. Beniwal V. Singh D. Kumar M. N. V. R. (2009). Nanoparticle encapsulation improves oral bioavailability of curcumin by at least 9-fold when compared to curcumin administered with piperine as absorption enhancer. Eur. J. Pharm. Sci. 37 (3-4 ), 223–230. 10.1016/j.ejps.2009.02.019 19491009 Shen W. Durum S. K. (2010). Synergy of IL-23 and Th17 cytokines: New light on inflammatory bowel disease. Neurochem. Res. 35 (6 ), 940–946. 10.1007/s11064-009-0091-9 19915978 Shlomovitz I. Zargrian S. Gerlic M. (2017). Mechanisms of RIPK3-induced inflammation. Immunol. Cell. Biol. 95 (2 ), 166–172. 10.1038/icb.2016.124 27974745 Silke J. Rickard J. A. Gerlic M. (2015). The diverse role of RIP kinases in necroptosis and inflammation. Nat. Immunol. 16 (7 ), 689–697. 10.1038/ni.3206 26086143 Simadibrata M. Halimkesuma C. C. Suwita B. M. (2017). Efficacy of curcumin as adjuvant therapy to induce or maintain remission in ulcerative colitis patients: An evidence-based clinical review. Acta Med. Indones. 49 (4 ), 363–368.29348389 Sun J. Jiang T. Xu W. Feng Z. Quan X. Leng P. (2019). Quantification of 1D, a novel derivative of curcumin with potential antitumor activity, in rat plasma by liquid chromatography-tandem mass spectrometry: Application to a pharmacokinetic study in rats. Pharm. Biol. 57 (1 ), 287–294. 10.1080/13880209.2019.1603243 31017510 Sun L. Wang H. Wang Z. He S. Chen S. Liao D. (2012). Mixed lineage kinase domain-like protein mediates necrosis signaling downstream of RIP3 kinase. Cell. 148 (1 ), 213–227. 10.1016/j.cell.2011.11.031 22265413 Tanaka H. Takechi M. Kiyonari H. Shioi G. Tamura A. Tsukita S. (2015). Intestinal deletion of Claudin-7 enhances paracellular organic solute flux and initiates colonic inflammation in mice. Gut 64 (10 ), 1529–1538. 10.1136/gutjnl-2014-308419 25691495 Vanden Berghe T. Linkermann A. Jouan-Lanhouet S. Walczak H. Vandenabeele P. (2014). Regulated necrosis: The expanding network of non-apoptotic cell death pathways. Nat. Rev. Mol. Cell. Biol. 15 (2 ), 135–147. 10.1038/nrm3737 24452471 Vanlangenakker N. Bertrand M. J. M. Bogaert P. Vandenabeele P. Vanden Berghe T. (2011). TNF-induced necroptosis in L929 cells is tightly regulated by multiple TNFR1 complex I and II members. Cell. Death Dis. 2 (11 ), e230. 10.1038/cddis.2011.111 22089168 Vereecke L. Beyaert R. van Loo G. (2011). Enterocyte death and intestinal barrier maintenance in homeostasis and disease. Trends Mol. Med. 17 (10 ), 584–593. 10.1016/j.molmed.2011.05.011 21741311 Villegas I. Sánchez-Fidalgo S. de la Lastra C. A. (2011). Chemopreventive effect of dietary curcumin on inflammation-induced colorectal carcinogenesis in mice. Mol. Nutr. Food Res. 55 (2 ), 259–267. 10.1002/mnfr.201000225 20848615 Wei C. Wang J. Y. Xiong F. Wu B. H. Luo M. H. Yu Z. C. (2021). Curcumin ameliorates DSS-induced colitis in mice by regulating the Treg/Th17 signaling pathway. Mol. Med. Rep. 23 (1 ), 34. 10.3892/mmr.2020.11672 33179078 Yang H. Du Z. Wang W. Song M. Sanidad K. Sukamtoh E. (2017). Structure-activity relationship of curcumin: Role of the methoxy group in anti-inflammatory and anticolitis effects of curcumin. J. Agric. Food Chem. 65 (22 ), 4509–4515. 10.1021/acs.jafc.7b01792 28513174 Yang X. Chao X. Wang Z. T. Ding W. X. (2016). The end of RIPK1-RIPK3-MLKL-mediated necroptosis in acetaminophen-induced hepatotoxicity? Hepatology 64 (1 ), 311–312. 10.1002/hep.28263 26418225 Yue W. Liu Y. Li X. Lv L. Huang J. Liu J. (2019). Curcumin ameliorates dextran sulfate sodium-induced colitis in mice via regulation of autophagy and intestinal immunity. Turk J. Gastroenterol. 30 (3 ), 290–298. 10.5152/tjg.2019.18342 30923033 Zeisel M. B. Dhawan P. Baumert T. F. (2019). Tight junction proteins in gastrointestinal and liver disease. Gut 68 (3 ), 547–561. 10.1136/gutjnl-2018-316906 30297438 Zhang D. W. Shao J. Lin J. Zhang N. Lu B. J. Lin S. C. (2009). RIP3, an energy metabolism regulator that switches TNF-induced cell death from apoptosis to necrosis. Science 325 (5938 ), 332–336. 10.1126/science.1172308 19498109 Zhang J. Yang Y. He W. Sun L. (2016). Necrosome core machinery: Mlkl. Cell. Mol. Life Sci. 73 (11 ), 2153–2163. 10.1007/s00018-016-2190-5 27048809 Zhong Y. B. Kang Z. P. Zhou B. G. Wang H. Y. Long J. Zhou W. (2020). Curcumin regulated the homeostasis of memory T cell and ameliorated dextran sulfate sodium-induced experimental colitis. Front. Pharmacol. 11 , 630244. 10.3389/fphar.2020.630244 33597887
PMC010xxxxxx/PMC10130509.txt
==== Front Front Pediatr Front Pediatr Front. Pediatr. Frontiers in Pediatrics 2296-2360 Frontiers Media S.A. 37124187 10.3389/fped.2023.1167595 Pediatrics Original Research Epidemiology, risk factors and outcomes of prolonged mechanical ventilation with different cut-points in a PICU Chongcharoenyanon Tatchanapong 1 Samransamruajkit Rujipat 2 * Sophonphan Jiratchaya 3 1 Division of Pulmonology, Department of Pediatrics, King Chulalongkorn Memorial Hospital, Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand 2 Division of Pediatric Critical Care, Department of Pediatrics, King Chulalongkorn Memorial Hospital, Faculty of Medicine, Chulalongkorn University, Bangkok, Thailand 3 The HIV Netherlands Australia Thailand Research Collaboration (HIV-NAT), Thai Red Cross AIDS Research Centre, Bangkok, Thailand Edited by: Andreas Schwingshackl, University of California, Los Angeles, United States Reviewed by: Mary E. Hartman, Washington University in St. Louis, United States Oguz Dursun, Akdeniz University, Türkiye Hongxing Dang, Children's Hospital of Chongqing Medical University, China * Correspondence: Rujipat Samransamruajkit [email protected] Specialty Section: This article was submitted to Pediatric Critical Care, a section of the journal Frontiers in Pediatrics 12 4 2023 2023 11 116759516 2 2023 28 3 2023 © 2023 Chongcharoenyanon, Samransamruajkit and Sophonphan. 2023 Chongcharoenyanon, Samransamruajkit and Sophonphan https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Background A consensus on the definition of prolonged mechanical ventilation (PMV) for children does not exist. There is still lack of published work presenting the epidemiology, risk factors and outcomes at different cut-points for PMV patients. These are important for planning the goals of treatment and counseling of the prognosis for patient families. We aimed to determine the incidence, baseline characteristics, risk factors and outcomes of PMV in pediatric patients at various cut-points (>14, >21 or >30days). Methods A retrospective cohort study among children <18-years-old who were PMV > 14 days in the PICU of King Chulalongkorn Memorial Hospital was conducted. The primary outcomes were incidence of PMV with various cut-points. We stratified patients into three groups (Group 1; PMV > 14–21, Group 2; >21–30, Group 3; >30 days) for evaluating the baseline characteristics, risk factors, and outcomes of PMV (extubation success, tracheostomy status and death). Factors associated with PMV and deaths were analyzed using univariate and multivariate logistic regression. Results From January 2018 to August 2022, 1,050 patients were screened. Of these, 114 patients were enrolled. The incidence of PMV > 14, >21 and >30 days were 10.9%, 7.3% and 5.0% respectively. Extubation success was significantly lower in Group 3 than in Groups 1 & 2 (15.4% vs. 62.2% & 56.0%, P < 0.001). Consequently, the tracheostomy rate (63.5% vs. 16.2% & 12.0%, P < 0.001), VAP rate (98.1% vs. 59.5% & 80.0%, P < 0.001), mortality rate by disease (34.6% vs. 5.4% & 20.0%, P = 0.003), median PICU LOS (50.5 vs. 22.0 & 28.0 days, P < 0.001) and median hospital LOS (124.5 vs. 55.0 & 62.0 days, P < 0.001) were also significantly higher for Group 3 compared with Groups 1 & 2. The factor associated with PMV > 30 days was VAP (aOR: 19.53, 95% CI: 2.38–160.34, P = 0.01). Factors associated with non-surviving patients were 3rd degree PEM (aOR: 5.14, 95% CI: 1.57–16.88, P = 0.01), PIM3 score ≥14 (aOR: 6.75, 95% CI: 2.26–20.15, P < 0.001) and muscle relaxant usage (aOR: 5.58, 95% CI: 1.65–18.86, P = 0.01). Conclusion Extubation failure, tracheostomy rate, VAP rate, mortality rate by disease, PICU LOS and hospital LOS were significantly higher for PMV >30 days. Consequently, we suggest that a 30-day duration as a cut-point for PMV in PICUs might be more appropriate. prolong mechanical ventilation children pediatrics definition cut-point long-term ventilation incidence tracheostomy ==== Body pmcIntroduction Medical advances, lower mortality and increased morbidity have increased the number of patients requiring prolonged mechanical ventilation (PMV) (1). In 2005, the National Association for Medical Direction of Respiratory Care (NAMDRC) defined PMV for adults as the need for ≥21 consecutive days of mechanical ventilation for ≥6 h/day of invasive and/or noninvasive methods of delivery. This cut-point was for most adult patients who were transferred to a long-term acute care (LTAC) hospital and had received ventilation for at least 21 days. The rate of PMV incidence in USA was estimated to be 3%–7% of mechanical ventilation (MV) patients (2–4). However, the PMV rate in China was surprisingly high, 36.1% of adults in ICUs (5). Currently, a consensus of the definition of PMV for children does not exist. Therefore, published literature reports show a variability in the duration of PMV ranging from 2 to 30 days (6–11), a lack of standardization about the inclusion of noninvasive ventilation (NIV), and a lack of standardization about the inclusion of times when a child is ventilator-free during weaning (12). José Colleti Jr et al. reviewed the existing definitions of PMV in children and found the most frequent thresholds in duration cited were 7, 14, 21 and 30 days (11). However, existing evidence and our experience found that the majority of patients in PICUs can be extubated before 14 days and MV for more than 15 days is associated with extubation failure (6, 7, 13). Recently, Liu et al. reported the incidence of PMV > 14 days of a PICU in China was 33.2% (14). Three studies used a PMV cut-point >21 days showing an incidence of approximately 2.5%–9.0%. Chronic comorbidities were reported in 85% of the patients. Mortality ranged from 37.7% to 48.0% (1, 9–11). Ramírez et al. conducted a study of PMV > 30 days in thirty-three PICUs in Spain. The incidence where PMV > 30 days was 30% (15). The published risk factors associated with PMV were younger age, especially <1 year old, chronic comorbidities, prematurity, a higher PIM-3 score, inotropic or vasopressors usage, higher driving pressure and fluid overload (14, 16–21). To the best of our knowledge, there is still a lack of published work presenting the epidemiology, risk factors and outcomes after different cut-points of PMV in children. This study aimed to determine the incidence of PMV in pediatric patients after different cut-points (>14, >21 or >30 days) and discover baseline characteristics, risk factors and outcomes of PMV in three groups (Group 1; >14–21 days, Group 2; >21–30 days, and Group 3; >30 days). Methods Study design and settings This retrospective cohort study was conducted at King Chulalongkorn Memorial Hospital (KCMH), a tertiary university hospital in Bangkok, Thailand. It has a pediatric department of 300 beds and usually an average 6,500 outpatient visits, 1,000 inpatient and 40–50 PICU admissions per month. The study population included patients aged from 37 weeks postmenstrual age (PMA) to 18 years old admitted into the PICU during January 2018 to August 2022. Eligibility criteria for enrollment included (1) patients who were ventilated >14 consecutive days (short interruptions with <48 h of ventilation), (2) those on an invasive mechanical ventilator (via an endotracheal tube, nasotracheal tube or tracheostomy tube). Patients were excluded if they were already included in this study, were ventilated via NIV, were brain dead, premature children <37 weeks PMA at enrollment, or were home ventilator-dependent patients. The criteria of brain death were adopted from the Determination of Brain Death/Death by Neurologic Criteria: The World Brain Death Project (22). This study was approved by the Institutional Review Board of the Faculty of Medicine, Chulalongkorn University (IRB No. 273/65) and informed consent requirements were waived. All data were kept confidential. Data collection and definition of variables Data were extracted from medical record to case record forms, including age (months), gender, weight, underlying disease, main diagnosis related with the current PICU admission, underlying disease, preterm birth, nutritional status, cumulative fluid accumulation (FA) from Days 1–7 of MV, treatment information (maximum dynamic driving pressure, MV duration, respiratory complications during MV, minimum P/F ratio in the first 72 h, vasopressor or inotropes, maximum vasoactive-inotropic score (VIS) in the first 72 h, neuromuscular blocking agents and duration of neuromuscular blocking agents). Severity of illness scores such as pediatric index of mortality (PIM)-3 (23), pediatric logistic organ dysfunction (PELOD)-2 (24) scores were collected. Pre-term was defined as patient born at a gestational age of <37 weeks. Underlying diseases were the condition of a patient prior to PICU admission. Nutritional status was categorized as either normal, 1st degree protein energy malnutrition (PEM), 2nd degree PEM or 3rd degree PEM (25, 26). Cumulative FA was calculated as the fluid inputs minus fluid outputs as a proportion of the admission (or preoperative) weight multiplied by 100. Fluids were expressed in liters and weights in kilograms. Fluid inputs included enteral and intravenous fluid administration. Fluid outputs included urine plus all other sources of fluid loss except insensible fluid loss (16, 27). Dynamic driving pressure (DP) was PIP minus PEEP. The dynamic DP overestimated the actual DP, but it could be used when patients are ventilated using modes that do not have a zero-flow state (17, 21). MV duration was defined as the time elapsed from intubation to extubation or successful disconnection from MV for tracheostomized patients (28, 29). The diagnosis of ventilator-associated pneumonia (VAP) was based on the CDC/NNIS definition related to age-specific criteria or qualified intensivist diagnosis by clinical assessment of clinical, laboratory, radiographic, and culture results (30). Hospital length of stay (LOS) was defined as the length of time elapsed between a patient's hospital admittance and discharge. PICU LOS was calculated as number of calendar days from the day of PICU admission (counted as 1 day) to the day of PICU discharge (31). Outcomes The primary outcome was incidence of PMV by different cut-points (>14 days, >21 days or >30 days). The secondary outcomes were baseline characteristics, risk factors, PICU length of stay, hospital length of stay, respiratory complications (VAP, pneumomediastinum, pneumothorax, pulmonary hemorrhage and unplanned extubation). Outcomes of PMV include extubation success, tracheostomy with or without home ventilator-dependence, transfer back with MV to previous hospital and death. The patients were placed into one of three groups (Group 1; >14 to ≤21 days, Group 2; >21 and ≤30 days, Group 3; >30 days). Indications and timing of tracheostomy were considered by the treating pediatric intensive care physician. Indications for tracheostomy were classified into three categories: upper airway obstruction, PMV and neurological impairment (such as cerebral palsy, encephalopathies with a Glasgow Coma Scale <8, central and peripheral neuromuscular disorders) (32, 33). All outcomes were collected from the day ventilation was started until the patient was discharged. The patients who died after withdrawal therapy were also excluded because the MV duration might have been longer if treatment was continued. Statistical analyses Demographic and clinical data were determined for each patient. Continuous variables are expressed as median (interquartile range: IQR) and percentage for categorical variables. Differences in continuous variables between the three groups were assessed using a Kruskal-Wallis test. A Wilcoxon rank sum test was used for comparisons between two groups. The categorical variables between groups were evaluated using χ2 or Fisher exact tests. Logistic regression (LR) was used to determine the factors associated PMV and death. Multivariate models were developed by adjusting for covariates with P < 0.1 in univariate models and stepwise backward LR to select a final model. All P-values reported are two-sided. Statistical significance was defined as P < 0.05. Stata version 15.1 (Stata Corp., College Station, Texas) was used for analysis. Results Patient demographic data and PMV incidence at various cut-points From January 2018 to August 2022, 1,050 mechanically ventilated pediatric patients were screened. One hundred and forty patients fulfilled the eligibility criteria. Twenty-six patients were excluded. The most common reason for study exclusion is that patients were previous included in the study (n = 19/26). One hundred and fourteen patients were included in the statistical analysis. No patients were being mechanically ventilated through a tracheostomy tube. We stratified them into three groups, Group 1 (n = 37, 32.5%), Group 2 (n = 25, 21.9%) and Group 3 (n = 52, 45.6%). The incidence of PMV > 14 days was 10.9% (95% CI: 9.0–12.9), PMV > 21 days was 7.3% (95% CI: 5.8–9.1) and PMV > 30 days was 5.0% (95% CI: 3.7–6.4) (Figure 1). Figure 1 Flow diagram of study design. MV, mechanical ventilation; PMA, postmenstrual age; PMV, prolong mechanical ventilation. The median age of all patients was 13 (IQR: 3–84) months. Females were 50.9% of all patients. The median MV duration was 26.5 days (IQR: 20.0–50.0). Most common main diagnoses were pulmonary disease, cardiac disease and neurological disease (58.8%, 15.8% and 10.5% respectively). Seventy-nine percent of the patients had underlying diseases. The most common underlying disease was cardiovascular disease which required significantly longer MV. A quarter of the patients were born prematurely (26.3%). Sixty percent of patients were malnourished, which was profoundly higher in Group 3 than Group 1 (71.2% vs. 48.7%, P < 0.05). The duration of MV was significantly longer for those with higher PIM-3 and PELOD2 scores. Baseline demographic data are shown in Table 1. Table 1 Demographic and baseline data characteristics. Total (n = 114) Group 1 (n = 37) Group 2 (n = 25) Group 3 (n = 52) P-value Age (months), median (IQR) 13 (3.0–84.0) 13 (5.0–48.0) 11 (4.0–34.0) 13 (3.0–115.0) 0.78  0–11 months old 54 (47.4%) 18 (48.7%) 13 (52.0%) 23 (44.2%)  12–59 months old 27 (23.7%) 10 (27.0%) 6 (24.0%) 11 (21.2%)  60–119 months old 12 (10.5%) 6 (16.2%) 0 (0) 6 (11.5%)  120–215 months old 21 (18.4%) 3 (8.1%) 6 (24.0%) 12 (23.1%) Female sex, n (%) 58 (50.9) 21 (56.8) 10 (40.0) 27 (51.9) 0.42 Weight (kg), median (IQR) 8.2 (4.0–22.0) 7.3 (4.4–15.9) 8.5 (5.6–15.0) 9.1 (3.8–24.5) 0.92 MV duration (days), median (IQR) 26.5 (20.0–50.0) 17.0 (16.0–19.0) 25.0 (23.0–26.0) 54.5 (38.5–100.5) <0.001 Main diagnosis, n (%) 0.44  Pulmonary disease 67 (58.8) 21 (56.8) 16 (64) 30 (57.7)  Cardiac disease 18 (15.8) 4 (10.8) 3 (12) 11 (21.2)  Neurological disease 12 (10.5) 3 (8.1) 2 (8) 7 (13.5)  Post-surgery 7 (6.1) 5 (13.5) 1 (4) 1 (1.9)  Shock 10 (8.8) 4 (10.8) 3 (12) 3 (5.8) Underlying disease, n (%) 0.84  No underlying disease 24 (21.1) 9 (24.3) 5 (20) 10 (19.2)  Yes 90 (79) 28 (75.7) 20 (80) 42 (80.8)   Cardiovascular disease 20 (17.5) 2 (5.4)* 6 (24) 12 (23.1)* 0.04   Pulmonary disease 17 (14.9) 6 (16.2) 2 (8) 9 (17.3) 0.59   Gastrointestinal disease 15 (13.2) 7 (18.9) 4 (16) 4 (7.7) 0.28   Oncological disease 8 (7) 2 (5.4) 1 (4) 5 (9.6) 0.71   Neuromuscular disease 7 (6.1) 2 (5.4) 1 (4) 4 (7.7) 0.80   Primary immune deficiency 4 (3.5) 1 (2.7) 2 (8) 1 (1.9) 0.43   Organ transplant 3 (2.6) 3 (8.1) 0 (0) 0 (0) 0.07   Others 16 (14) 5 (13.5) 4 (16) 7 (13.5) 0.95 Prematurity, n (%) 30 (26.3) 8 (21.6) 6 (24) 16 (30.8) 0.60 Malnutrition, n (%) 0.09  No 45 (39.5) 19 (51.4) 11 (44.0) 15 (28.9)  Yes 69 (60.5) 18 (48.7)* 14 (56.0) 37 (71.2)*   1st degree PEM 22 (19.3) 7 (18.9) 4 (16.0) 11 (21.2) 0.86   2nd degree PEM 16 (14.0) 4 (10.8) 4 (16.0) 8 (15.4) 0.79   3rd degree PEM 31 (27.2) 7 (18.9) 6 (24.0) 18 (34.6) 0.24 PIM-3 score (%), median (IQR) 6.9 (4.4–17.2) 5.5 (3.1–9.3)* 7.5 (5.7–22.6)* 7.6 (4.1–17.3) 0.04 PELOD-2 score, median (IQR) 5.0 (3.0–7.0) 4.0 (3.0–5.0)* 5.0 (4.0–6.0) 5.0 (4.0–7.0)* 0.04 MV, mechanical ventilation; PEM, protein energy malnutrition; PIM-3, pediatric index of mortality 3; PELOD-2, pediatric logistic organ dysfunction 2; Group 1, MV > 14–21 days; Group 2, MV > 21–30 days; Group 3, MV > 30 days; Differences in continuous and categorical variables between three groups using a Kruskal-Wallis test; Differences in continuous variables between two groups using a Wilcoxon rank sum test; Differences in categorical variables used χ2 test or Fisher exact test. * Statistically significant for pairwise comparison (P < 0.05). Treatment information The median of the minimum P/F ratio in the first 72 h was 169 (IQR: 118–268). It was significantly lower in Group 3 than Group 1 (150, IQR: 104–206 vs. 222, IQR: 139–283, P = 0.04). More than half of the patients (58.8%) were using inotropic or vasopressor drugs with a median maximum VIS in the first 72 h of 10.0 (IQR: 0–20.0). Muscle relaxant was used for 49.1% of the patients with a median duration of 4.5 days (IQR: 2.5–11.5). The median maximum DP was 16 cm H2O (IQR: 13–20). The median maximum cumulative FA was 9.5% (IQR: 5.6–17.7) with a median of 6 days after MV (IQR: 3–7). The median cumulative FA from Day 1 to Day 7 is shown in Table 2. However, there was no statistical difference in these parameters between groups. Table 2 Treatment information. Total (n = 114) Group 1 (n = 37) Group 2 (n = 25) Group 3 (n = 52) P-value Inotropic or vasopressor usage, n (%) 67 (58.8) 19 (51.4) 14 (56.0) 34 (65.4) 0.39 Maximum VIS in the first 72 h, median (IQR) 10.0 (0–20.0) 5.0 (0–15.0) 5.0 (0–25.0) 10.0 (0–27.9) 0.31 Minimum P/F ratio in the first 72 h, median (IQR) 169 (118–268) 222 (139–283)* 182 (122–272) 150 (104–206)* 0.04 Maximum DP (cmH2O), median (IQR) 16 (13–20) 15 (12–20) 15 (14–18) 17 (14–20) 0.18 Muscle relaxant usage, n (%) 56 (49.1) 15 (40.5) 12 (48.0) 29 (55.8) 0.36 Duration of muscle relaxant (days), median (IQR) 4.5 (2.5–11.5) 3 (2.0–7.0) 3 (2.5–11.5) 5 (3.0–18.0) 0.13 Cumulative FA Day 1 (%), median (IQR) 2.7 (0.5–4.4) 2.8 (0.6–5.1) 2.9 (0.9–4.4) 2.4 (0–4.2) 0.55 Cumulative FA Day 2 (%), median (IQR) 4.1 (0.5–6.8) 2.5 (0.8–8.2) 4.7 (−0.1–6.4) 4.1 (0.8–6.3) 0.84 Cumulative FA Day 3 (%), median (IQR) 4.8 (0.6–8.7) 5.5 (1.5–10.3) 5.1 (0.4–7.4) 4 (−0.2–8.1) 0.50 Cumulative FA Day 4 (%), median (IQR) 5.2 (0.6–9.9) 5.7 (0.8–10.1) 5.3 (1.2–9.3) 4.1 (0.2–10.4) 0.80 Cumulative FA Day 5 (%), median (IQR) 5.3 (0.3–12.1) 6.6 (0.4–12.1) 5.2 (−0.1–10.8) 5.2 (0.3–12.3) 0.90 Cumulative FA Day 6 (%), median (IQR) 7.1 (0.2–13.8) 4.9 (−1.0–13.8) 7.5 (4.0–12.7) 7.1 (−1.0–18) 0.82 Cumulative FA Day 7 (%), median (IQR) 6.5 (1.0–16.3) 5.5 (−1.2–15.6) 7.9 (4.7–14.7) 6.6 (−0.5–22.6) 0.54 Maximum cumulative FA (%), median (IQR) 9.5 (5.6–17.7) 8.9 (5.3–16.4) 9.6 (6.1–16.2) 9.7 (4.5–23.1) 0.79 Days of maximum cumulative FA, median (IQR) 6 (3–7) 5 (3–7) 6.5 (4–7) 6 (3–7) 0.56 PMV, prolonged mechanical ventilation; VIS, vasoactive-inotropic score; DP, driving pressure; FA, fluid accumulation; Group 1, MV > 14–21 days; Group 2, MV > 21–30 days; Group 3, MV > 30 days; Differences in continuous and categorical variables between three groups using a Kruskal-Wallis test; Differences in continuous variables between two groups using a Wilcoxon rank sum test; Differences in categorical variables used χ2 test or Fisher exact test. * Statistically significant for pairwise comparisons (P < 0.05). Clinical outcomes after PMV The extubation success rate was at 45/114 (39.5%) and it was significantly higher for Group 1 and Group 2 than for Group 3 (62.2% and 56.0% vs. 15.4%, P < 0.001). Consequently, Group 3 statistically had the highest tracheostomy rate (n = 33/52, 63.5%), VAP rate (n = 51/52, 98.1%), PICU LOS (median 50.5 days, IQR: 39.5–76.0) and hospital LOS (median 124.5 days, IQR: 74.5–206). Overall, the median mortality rate was 33/114 (29.0%) and 25/114 (21.9%) after exclusion of patients withdrawn from therapy. It is noteworthy that the highest mortality was in Group 3 (n = 18/52, 34.6%). Almost all patients of Group 1 were tracheostomized with indications of upper airway obstruction (n = 5/6, 83.3%). In contrast, within Group 3, the most common indication was PMV (n = 18/33, 54.6%). The median days intubated before tracheostomy was significantly higher for Group 3 (39 days, IQR: 22–59, P < 0.02) (Table 3). Table 3 Outcomes after PMV. Parameter Total (n = 114) Group 1 (n = 37) Group 2 (n = 25) Group 3 (n = 52) P-value Extubation success, n (%) 45 (39.5) 23 (62.2)* 14 (56.0)# 8 (15.4)*,# <0.001 Tracheostomy, n (%) 42 (36.8) 6 (16.2)* 3 (12.0)# 33 (63.5)*,# <0.001 Tracheostomy without ventilator-dependent, n (%) 28 (24.6) 5 (13.5)* 2 (8.0)# 21 (40.4)*,# 0.001 Tracheostomy with ventilator-dependent, n (%) 14 (12.3) 1 (2.7)* 1 (4.0) 12 (23.1)* 0.006 Indication for tracheostomy, n (%)  Upper airway obstruction 16/42 (38.1) 5/6 (83.3)* 1/3 (33.3) 10/33 (30.3)* 0.04  PMV 19/42 (45.2) 1/6 (16.7) 0/3 (0) 18/33 (54.6) 0.08  Neuromuscular disease 7/42 (16.7) 0/6 (0) 2/3 (66.7) 5/33 (15.2) 0.06 Days intubated before tracheostomy, median (IQR) 30 (18–54) 18 (17–18)*,# 26 (19–42)* 39 (22–59)# 0.02 Death, n (%) 33 (29.0) 8 (21.6) 7 (28.0) 18 (34.6) 0.41  Death by disease, n (%) 25 (21.9) 2 (5.4)* 5 (20.0) 18 (34.6)* 0.03  Death by withdrawal therapy, n (%) 8 (7.0) 6 (16.2)* 2 (8.0) 0 (0)* 0.005 Transfer back with MV, n (%) 6 (5.3) 0 (0) 1 (4.0) 5 (9.6) 0.10 Respiratory complication during MV, n (%) <0.001  No complication 17 (14.9) 12 (32.4) 4 (16.0) 1 (1.9)  Yes 97 (85.1) 25 (67.6) 21 (84.0) 51 (98.1)   VAP 93 (81.6) 22 (59.5)* 20 (80.0)# 51 (98.1)*,# <0.001   Pneumothorax 14 (12.3) 4 (10.8) 3 (12.0) 7 (13.5) 0.94   Pulmonary hemorrhage 8 (7.0) 2 (5.4) 4 (16.0) 2 (3.9) 0.17   Unplanned extubation 2 (1.8) 0 1 (4.0) 1 (1.9) 0.50   Pneumomediastinum 0 0 0 0 PICU LOS (days), median (IQR) 35.0 22.0 28.0 50.5 <0.001¥ (23.0–54.0) (18.0–28.0) (25.0–35.0) (39.5–76.0) Hospital LOS (days), median (IQR) 79.0 55.0 62.0 124.5 <0.001 (52.0–141.0) (43.0–87.0)* (48.0–102.0)# (74.5–206.0)*,# PMV, prolonged mechanical ventilation; MV, mechanical ventilation; VAP, ventilator-associated pneumonia; LOS, length of stay; Group 1, MV > 14–21 days; Group 2, MV > 21–30 days; Group 3, MV > 30 days; Differences in continuous and categorical variables between three groups using a Kruskal-Wallis test; Differences in continuous variables between two groups using a Wilcoxon rank sum test; Differences in categorical variables used a χ2 test or Fisher exact test. * Statistical significant for pairwise comparison (P < 0.05). # Statistical significant for pairwise comparison (P < 0.05). ¥ All pairwise comparisons are statistical significantly. PMV > 14–30 days vs. PMV > 30 days According to data for outcomes after PMV, we decided to re-categorize study participants to underscore the obvious differences between two groups: PMV > 14–30 days (combining Group 1 and Group 2) and PMV > 30 days (Group 3) (Supplementary material Tables S1–S3). In the PMV > 30-days group, the extubation success rate (15.4% vs. 59.7%, P < 0.001) and median minimum P/F ratio in the first 72 h were profoundly lower than for PMV > 14–30 days [150 (104–206) vs. 195.5 (132–275), P = 0.02]. The tracheostomy rate (63.5% vs. 14.5%, P < 0.001), PMV as an indication of tracheostomy rate (54.6% vs. 11.1%, P < 0.001), mortality rate after excluded withdrawal therapy (34.6% vs. 11.3%, P = 0.003), VAP rate (98.1% vs. 67.7%, P < 0.001), median duration of muscle relaxant use [5 (3–18) vs. 3 (2–8) days, P = 0.05], median PICU LOS [50.5 (39.5–76.0) vs. 25.0 (20.0–35.0) days, P < 0.001] and median hospital LOS [124.5 (74.5–206.0) vs. 57.0 (43.0–89.0) days, P < 0.001] were higher in the PMV > 30-days group (Figure 2). Figure 2 Outcomes of patients with PMV > 14 to 30 days and PMV > 30 days (A) Percentage of extubation success, tracheotomy, death by disease and VAP (B) Median PICU and hospital LOS (days). PMV, Prolonged Mechanical Ventilation; VAP, Ventilator-Associated Pneumonia; LOS, Length Of Stay. Univariate and multivariate analyses through binary logistic regression were performed to compare the patients in the PMV > 14–30 days and PMV > 30 days groups (Supplementary material Table S4). Multivariable analysis showed that only VAP was associated with PMV > 30 days (aOR: 19.53, 95% CI: 2.38–160.34, P = 0.01) (Figure 3A). Figure 3 Factors associated with (A) PMV > 30 days (B) deaths; Data were analyzed using a multivariate logistic regression model. PEM, protein energy malnutrition; PIM-3, pediatric index of mortality 3. Survivors vs. deaths A comparison between survivors (n = 81/114, 71.1%) and deaths (n = 33/114, 28.9%) is shown in Supplementary material Tables S5–S7. Deaths had higher rates of cardiac disease as a main diagnosis (27.3% vs. 11.1%, P = 0.03), cardiovascular diseases (30.3% vs. 12.4%, P = 0.02) and gastrointestinal diseases (24.2% vs. 8.6%, P = 0.03) as underlying diseases, 3rd degree PEM (54.6% vs. 16.1%, P < 0.001), inotropic or vasopressor usage (78.8% vs. 50.6%, P = 0.006), muscle relaxant usage (75.8% vs. 38.3%, P < 0.001), median maximum VIS in the first 72 h [20 (7–47) vs. 2.8 (0–15), P < 0.001], median duration of muscle relaxant use [10 (5–15) vs. 3 (2–5) days, P < 0.001], median maximum DP [19 (16–22) vs. 15 (12–18) cm H2O, P < 0.001], median PIM-3 score [17.3 (8.5–22.4) vs. 5.7 (3.3–9.6), P < 0.001] and median of PELOD-2 score [5 (3–6) vs. 6 (3–7), P = 0.03]. Risk factors associated with death patients were 3rd degree PEM (aOR: 5.14, 95% CI: 1.57–16.88, P = 0.01), PIM3 score ≥14 (aOR: 6.75, 95% CI: 2.26–20.15, P < 0.001) and muscle relaxant use (aOR: 5.58, 95% CI: 1.65–18.86, P = 0.01) (Figure 3B). Discussion The current study reports the incidence of PMV > 14, >21 and >30 days was 10.9%, 7.3% and 5.0%, respectively. We decided to not include the NIV patients because they can be managed outside the PICU. This might be the reason that our incidence of PMV > 14 days was lower than a previous study reported from China PICU (33.2%) (14). The incidence of PMV > 21 days was close to that reported from an Argentinian PICU (9%), but higher than a study in Brazilian PICUs (2.5%) and an adult ICU (3–7%). However, our mortality rate for PMV > 21 days was 32.5% (n = 25/77), which was lower than for Brazilian PICUs (48.0%), Argentinian PICUs (43.0%), in a Taiwan database (37.7%) and an adult ICU (42.4%) (1, 4, 9, 10, 34). This indicates high mortality rates for PMV > 21 days. Furthermore, a study in 33 Spanish PICUs reported 30% of patients required PMV > 30 days, whereas our institution reported a much lower level. Differences in patient characteristics might explain this. In our study, we observed that 26.3% of patients were born prematurely and 60% of patients suffered from malnutrition. These factors could have been potential contributors to the requirement for PMV in some cases. Premature birth can lead to underdeveloped lungs and respiratory muscles, which can result in respiratory failure and the need for mechanical ventilation. Furthermore, complications like bronchopulmonary dysplasia may prolong the necessity for mechanical ventilation in premature infants. Meanwhile, malnutrition can weaken respiratory muscles, thereby increasing the likelihood of respiratory failure and the subsequent need for mechanical ventilation. Additionally, malnutrition may cause other complications that can contribute to PMV, such as infections or organ failure. Teka et al. concluded that malnutrition in PICU was a frequent cause of morbidity that was linked with a higher requirement for mechanical ventilation (35). Several studies have reported baseline characteristics, risk factors and outcomes using arbitrary cut-points for PMV (11). To identify the PMV effect on clinical outcomes, we illustrate this by using different cut-point durations for PMV. In 2005, Fontela et al. first showed that MV ≥ 15 days was associated with a low extubation success rate (4.5%), whereas our study revealed a much higher success rate (39.5%) (13). This finding suggests that knowledge advancements about weaning and support care can improve the extubation success rate over time. Furthermore, we also found that extubation success was not different between PMV > 14–21 days (62.2%) and PMV > 21–30 days (56.0%), but it was significantly lower for PMV > 30 days (15.4%). Damuth et al. revealed a 29.0% pooled mortality at hospital discharge in PMV > 14 days for adults, which is similar to our results (36). We underscore that the mortality rate after excluded withdrawal therapy was significantly higher for PMV > 30 days compared with PMV > 14–30 days (34.6% vs. 11.3%, P = 0.003). Factors associated with mortality were inotropic or vasopressor usage, PIM-3 score ≥14% and muscle relaxant usage. Additionally, our analysis revealed that individuals who died had a significantly higher incidence of cardiac disease as a main diagnosis (27.3% vs. 11.1%, P = 0.03), as well as higher rates of cardiovascular (30.3% vs. 12.4%, P = 0.02) and gastrointestinal diseases (24.2% vs. 8.6%, P = 0.03) as underlying diseases. They also had a higher median maximum DP [19 (16–22) vs. 15 (12–18) cm H2O, P < 0.001] compared to the survivors. Our findings suggest that patients with cardiac disease are more vulnerable to changes in fluid and electrolyte balance, and poor cardiac function can worsen respiratory complications such as atelectasis or pulmonary edema, which can ultimately lead to death. Although it was recommended for adult patients that tracheostomy should be performed within 1–2 weeks of ventilation, no consensus criteria of time to tracheostomy exists for children (32, 37). Some studies reported that early tracheostomy (within 14 days of ventilation) could reduce MV duration and hospital LOS (38, 39). From another point of view, caregivers of children with tracheostomy tubes have high caregiver workloads and financial burden (40, 41). Our institution is highly concern about these problems. Therefore, we tend to delay tracheostomy tube for indications of PMV. The tracheostomy rate due to PMV was significantly higher for PMV > 30 days compared with PMV for 14–30 days (n = 18/52, 34.6% vs. n = 1/62, 1.6%, P = 0.02). The results of our study could be applied for family counseling regarding the prognosis for PMV in children. We found that if we used a cut point of PMV > 30 days, we could identify the highest percentage of children who eventually required tracheostomy. Thus, we suggest that use of a 30-day duration as a cut-point for PMV in PICUs might be more appropriate. The mortality rate and other morbidity data that we found in our study such as the VAP rate, PICU LOS and hospital LOS were also the highest for PMV > 30 days. High DP, more than 15 cm H2O, in acute hypoxemic respiratory failure children was independently associated with longer MV duration and PICU LOS, but not in the mortality rate. There is currently no study of DP in PMV. The median maximum DP in our study was 16 cm H2O (IQR: 13–20) and DP tended to be greater when comparing PMV > 30 days with PMV > 14–30 days [17 (14–20) vs. 15 (13–18), P = 0.09] which supports previous studies (17, 21). Although deaths had a significantly higher median maximum DP than survivors [19 (16–22) vs. 15 (12–18) cm H2O, P < 0.001], it was not an independent predictor of mortality after multivariate analysis (even with a cutoff of 17 cm H2O). Several studies reported an association between fluid accumulation and MV duration, PICU LOS and mortality rate (18–20). Recently, Gelbart et al. reported moderate FA up to 10% in MV critically ill children was not associated with harm in contrast with extremely greater than 20% which was associated with higher morbidity and mortality (16). Our institution is also greatly concerned about this issue. Consequently, the median cumulative FA from Day 1 to 7 was less than 10% and a median maximal FA was 9.5%. Therefore, we did not find any significance of FA among the three groups of PMV patients. There were no significant differences between survivors and deaths. Our particular findings highlight that a FA of less than 10% may not be associated with morbidity and mortality in PMV patients. To the best of our knowledge, this is the first study that evaluates baseline characteristics, risk factors and outcomes using various definitions instead of arbitrary cut-points for duration of PMV in a PICU. However, our study has some limitations. First, it was conducted in a single tertiary care setting where the incidence of PMV is higher and mortality rate is lower than in other settings. This may limit the generalizability of the findings. Nevertheless, we would like to propose the idea of conducting further studies that use various cut-points instead of single cut-points for duration of PMV. Second, it was a retrospective study that mainly relied on medical chart review. However, we can extract more than 95% of the data due to the use of electronic records. Third, most of our patients had an underlying disease that might influence the results. Fourth, we only included invasively mechanically ventilated children. In the future, a multicenter study using different cut-points for PMV in children seems warranted. Data availability statement The original contributions presented in the study are included in the article/Supplementary Materials, further inquiries can be directed to the corresponding author. Ethics statement The studies involving human participants were reviewed and approved by the Institutional Review Board of the Faculty of Medicine, Chulalongkorn University (IRB No. 273/65). Written informed consent from the participants’ legal guardian/next of kin was not required to participate in this study in accordance with the national legislation and the institutional requirements. Author contributions TC and RS: designed the study. TC: collected the data and wrote the original manuscript. RS: was supervision. JS: performed the statistical analysis. All authors contributed to the article and approved the submitted version. Conflict of interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Publisher's note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. Supplementary material The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fped.2023.1167595/full#supplementary-material. ==== Refs References 1. Pai S-C Kung P-T Chou W-Y Kuo T Tsai W-C . Survival and medical utilization of children and adolescents with prolonged ventilator-dependent and associated factors. PLoS One. (2017) 12 (6 ):e0179274. 10.1371/journal.pone.0179274 28628663 2. Hassenpflug M Scheinhorn DJ Chao DC Palma C . Post-ICU mechanical ventilation: treatment of 2,369 patients over 16 years at a regional weaning center. Chest. (2004) 126 (4 ):750S. 10.1378/chest.126.4_MeetingAbstracts.750S-a 3. Scheinhorn DJ Chao DC Hassenpflug MS Gracey DR . Post-ICU weaning from mechanical ventilation: the role of long-term facilities. Chest. (2001) 120 (6, Supplement ):482S–4S. 10.1378/chest.120.6_suppl.482S 11742970 4. MacIntyre NR Epstein SK Carson S Scheinhorn D Christopher K Muldoon S . Management of patients requiring prolonged mechanical ventilation: report of a NAMDRC consensus conference. Chest. (2005) 128 (6 ):3937–54. 10.1378/chest.128.6.3937 16354866 5. Li J Zhan QY Wang C . Survey of prolonged mechanical ventilation in intensive care units in mainland China. Respir Care. (2016) 61 (9 ):1224–31. 10.4187/respcare.04295 27460102 6. Wakeham MK Kuhn EM Lee KJ McCrory MC Scanlon MC . Use of tracheostomy in the PICU among patients requiring prolonged mechanical ventilation. Intensive Care Med. (2014) 40 (6 ):863–70. 10.1007/s00134-014-3298-4 24789618 7. Payen V Jouvet P Lacroix J Ducruet T Gauvin F . Risk factors associated with increased length of mechanical ventilation in children. Pediatr Crit Care Med. (2012) 13 (2 ):152–7. 10.1097/PCC.0b013e3182257a24 21760567 8. Polito A Patorno E Costello JM Salvin JW Emani SM Rajagopal S Perioperative factors associated with prolonged mechanical ventilation after complex congenital heart surgery*. Pediatr Crit Care Med. (2011) 12 (3 ):e122–6. 10.1097/PCC.0b013e3181e912bd 20625334 9. Monteverde E Fernández A Poterala R Vidal N Siaba Serrate A Castelani P Characterization of pediatric patients receiving prolonged mechanical ventilation. Pediatr Crit Care Med. (2011) 12 (6 ):e287–91. 10.1097/PCC.0b013e3182191c0b 21499185 10. Traiber C Piva JP Fritsher CC Garcia PCR Lago PM Trotta EA Profile and consequences of children requiring prolonged mechanical ventilation in three Brazilian pediatric intensive care units. Pediatr Crit Care Med. (2009) 10 (3 ):375–80. 10.1097/PCC.0b013e3181a3225d 19325502 11. Colleti JJ Azevedo RT de Oliveira Caino FR de Araujo OR . Prolonged mechanical ventilation in children: review of the definition. Pediatr Crit Care Med. (2021) 22 (11 ):e588–93. 10.1097/PCC.0000000000002773 34028375 12. Sauthier M Rose L Jouvet P . Pediatric prolonged mechanical ventilation: considerations for definitional criteria. Respir Care. (2017) 62 (1 ):49. 10.4187/respcare.04881 27879381 13. Fontela PS Piva JP Garcia PC Bered PL Zilles K . Risk factors for extubation failure in mechanically ventilated pediatric patients. Pediatr Crit Care Med. (2005) 6 (2 ):166–70. 10.1097/01.PCC.0000154922.65189.48 15730603 14. Liu Y Wang Q Hu J Zhou F Liu C Li J Characteristics and risk factors of children requiring prolonged mechanical ventilation vs. non-prolonged mechanical ventilation in the PICU: a prospective single-center study. Front Pediatr. (2022) 10 :830075. 10.3389/fped.2022.830075 35211431 15. Ramírez JB Cid JL-H Alapont VM . Prevalencia de la ventilación mecánica en las unidades de cuidados intensivos pediátricos en españa. Anales de Pediatría. (2004) 61 (6 ):533–41. 10.1016/S1695-4033(04)78440-4 15574254 16. Gelbart B Serpa Neto A Stephens D Thompson J Bellomo R Butt W Fluid accumulation in mechanically ventilated, critically ill children: retrospective cohort study of prevalence and outcome. Pediatr Crit Care Med. (2022) 23 (12 ):990–8. 10.1097/PCC.0000000000003047 36454001 17. van Schelven P Koopman AA Burgerhof JGM Markhorst DG Blokpoel RGT Kneyber MCJ . Driving pressure is associated with outcome in pediatric acute respiratory failure. Pediatr Crit Care Med. (2022) 23 (3 ):e136–44. 10.1097/PCC.0000000000002848 34669679 18. Kong X Zhu Y Zhu X . Association between early fluid overload and mortality in critically-ill mechanically ventilated children: a single-center retrospective cohort study. BMC Pediatr. (2021) 21 (1 ):474. 10.1186/s12887-021-02949-w 34702226 19. Lopes CLS Piva JP . Fluid overload in children undergoing mechanical ventilation. Rev Bras Ter Intensiva. (2017) 29 (3 ):346–53. 10.5935/0103-507X.20170045 28977099 20. van Mourik N Metske HA Hofstra JJ Binnekade JM Geerts BF Schultz MJ Cumulative fluid balance predicts mortality and increases time on mechanical ventilation in ARDS patients: an observational cohort study. PLoS One. (2019) 14 (10 ):e0224563-e. 10.1371/journal.pone.0224563 31665179 21. Rauf A Sachdev A Venkataraman ST Dinand V . Dynamic airway driving pressure and outcomes in children with acute hypoxemic respiratory failure. Respir Care. (2021) 66 (3 ):403–9. 10.4187/respcare.08024 33024000 22. Greer DM Shemie SD Lewis A Torrance S Varelas P Goldenberg FD Determination of brain death/death by neurologic criteria: the world brain death project. JAMA. (2020) 324 (11 ):1078–97. 10.1001/jama.2020.11586 32761206 23. Straney L Clements A Parslow RC Pearson G Shann F Alexander J Paediatric index of mortality 3: an updated model for predicting mortality in pediatric intensive care*. Pediatr Crit Care Med. (2013) 14 (7 ):673–81. 10.1097/PCC.0b013e31829760cf 23863821 24. Leteurtre S Duhamel A Salleron J Grandbastien B Lacroix J Leclerc F PELOD-2: an update of the PEdiatric logistic organ dysfunction score. Crit Care Med. (2013) 41 (7 ):1761–73. 10.1097/CCM.0b013e31828a2bbd 23685639 25. Bhattacharyya AK . Protein-energy malnutrition (kwashiorkor-marasmus syndrome): terminology, classification and evolution. World Rev Nutr Diet. (1986) 47 :80–133. 10.1159/000412332 3088855 26. Waterlow JC . Classification and definition of protein-energy malnutrition. Monogr Ser World Health Organ. (1976) (62 ):530–55. PMID: 824854.824854 27. Abulebda K Cvijanovich NZ Thomas NJ Allen GL Anas N Bigham MT Post-ICU admission fluid balance and pediatric septic shock outcomes: a risk-stratified analysis. Crit Care Med. (2014) 42 (2 ):397–403. 10.1097/CCM.0b013e3182a64607 24145842 28. Gamberini L Tonetti T Spadaro S Zani G Mazzoli CA Capozzi C Factors influencing liberation from mechanical ventilation in coronavirus disease 2019: multicenter observational study in fifteen Italian ICUs. J Intensive Care. (2020) 8 (1 ):80. 10.1186/s40560-020-00499-4 33078076 29. Contentin L Ehrmann S Giraudeau B . Heterogeneity in the definition of mechanical ventilation duration and ventilator-free days. Am J Respir Crit Care Med. (2014) 189 (8 ):998–1002. 10.1164/rccm.201308-1499LE 24735035 30. Srinivasan R Asselin J Gildengorin G Wiener-Kronish J Flori HR . A prospective study of ventilator-associated pneumonia in children. Pediatrics. (2009) 123 (4 ):1108–15. 10.1542/peds.2008-1211 19336369 31. Williams TA Ho KM Dobb GJ Finn JC Knuiman M Webb SAR . Effect of length of stay in intensive care unit on hospital and long-term mortality of critically ill adult patients. Br J Anaesth. (2010) 104 (4 ):459–64. 10.1093/bja/aeq025 20185517 32. Watters KF . Tracheostomy in infants and children. Respir Care. (2017) 62 (6 ):799. 10.4187/respcare.05366 28546379 33. Can FK Anıl AB Anıl M Gümüşsoy M Çitlenbik H Kandoğan T The outcomes of children with tracheostomy in a tertiary care pediatric intensive care unit in Turkey. Turk Pediatri Ars. (2018) 53 (3 ):177–84. 10.5152/TurkPediatriArs.2018.6586 30459517 34. Hill AD Fowler RA Burns KE Rose L Pinto RL Scales DC . Long-term outcomes and health care utilization after prolonged mechanical ventilation. Ann Am Thorac Soc. (2017) 14 (3 ):355–62. 10.1513/AnnalsATS.201610-792OC 28033033 35. Teka SG Kebede RA Sherman C . The prevalence of malnutrition during admission to the pediatric intensive care unit, a retrospective cross-sectional study at tikur anbessa specialized hospital, Addis Ababa, Ethiopia. Pan Afr Med J. (2022) 41 :77. 10.11604/pamj.2022.41.77.31284 35382053 36. Damuth E Mitchell JA Bartock JL Roberts BW Trzeciak S . Long-term survival of critically ill patients treated with prolonged mechanical ventilation: a systematic review and meta-analysis. Lancet Respir Med. (2015) 3 (7 ):544–53. 10.1016/S2213-2600(15)00150-2 26003390 37. Andriolo BN Andriolo RB Saconato H Atallah ÁN Valente O . Early versus late tracheostomy for critically ill patients. Cochrane Database Syst Rev. (2015 1 . 10.1002/14651858.CD010959.pub2 38. de Araujo OR Azevedo RT de Oliveira FRC Junior JC . Tracheostomy practices in children on mechanical ventilation: a systematic review and meta-analysis. J Pediatr. (2022) 98 (2 ):126–35. 10.1016/j.jped.2021.07.004 39. Chorath K Hoang A Rajasekaran K Moreira A . Association of early vs late tracheostomy placement with pneumonia and ventilator days in critically ill patients: a meta-analysis. JAMA Otolaryngol Head Neck Surg. (2021) 147 (5 ):450–9. 10.1001/jamaoto.2021.0025 33704354 40. Baddour K Mady LJ Schwarzbach HL Sabik LM Thomas TH McCoy JL Exploring caregiver burden and financial toxicity in caregivers of tracheostomy-dependent children. Int J Pediatr Otorhinolaryngol. (2021) 145 :110713. 10.1016/j.ijporl.2021.110713 33882339 41. Hartnick CJ Bissell C Parsons SK . The impact of pediatric tracheotomy on parental caregiver burden and health Status. Arch Otolaryngol Head Neck Surg. (2003) 129 (10 ):1065–9. 10.1001/archotol.129.10.1065 14568788
PMC010xxxxxx/PMC10150118.txt
==== Front Front Microbiol Front Microbiol Front. Microbiol. Frontiers in Microbiology 1664-302X Frontiers Media S.A. 37138607 10.3389/fmicb.2023.1018242 Microbiology Original Research A westernized diet changed the colonic bacterial composition and metabolite concentration in a dextran sulfate sodium pig model for ulcerative colitis Panah Farhad M. 1 * † Nielsen Katrine D. 1 Simpson Gavin L. 1 Schönherz Anna 1 Schramm Andreas 2 Lauridsen Charlotte 1 Nielsen Tina S. 1 Højberg Ole 1 Fredborg Marlene 1 Purup Stig 1 Canibe Nuria 1 1Department of Veterinary and Animal Sciences, Aarhus University, Tjele, Denmark 2Department of Biology, Aarhus University, Aarhus, Denmark Edited by: Li Wang, Guangdong Academy of Agricultural Sciences (GDAAS), China Reviewed by: Sérgio D. C. Rocha, Norwegian University of Life Sciences, Norway; Tao Wu, Tianjin University of Science and Technology, China *Correspondence: Farhad M. Panah, [email protected] †Present address: Farhad M. Panah, Department of Food Science, University of Copenhagen, Copenhagen, Denmark This article was submitted to Microorganisms in Vertebrate Digestive Systems, a section of the journal Frontiers in Microbiology 17 4 2023 2023 14 101824212 8 2022 27 3 2023 Copyright © 2023 Panah, Nielsen, Simpson, Schönherz, Schramm, Lauridsen, Nielsen, Højberg, Fredborg, Purup and Canibe. 2023 Panah, Nielsen, Simpson, Schönherz, Schramm, Lauridsen, Nielsen, Højberg, Fredborg, Purup and Canibe https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Introduction Ulcerative colitis (UC) is characterized by chronic inflammation in the colonic epithelium and has a blurred etiology. A western diet and microbial dysbiosis in the colon were reported to play a role in UC development. In this study, we investigated the effect of a westernized diet, i.e., increasing fat and protein content by including ground beef, on the colonic bacterial composition in a dextran sulfate sodium (DexSS) challenged pig study. Methods The experiment was carried out in three complete blocks following a 2×2 factorial design including 24 six-week old pigs, fed either a standard diet (CT) or the standard diet substituted with 15% ground beef to simulate a typical westernized diet (WD). Colitis was induced in half of the pigs on each dietary treatment by oral administration of DexSS (DSS and WD+DSS, respectively). Samples from proximal and distal colon and feces were collected. Results and discussion Bacterial alpha diversity was unaffected by experimental block, and sample type. In proximal colon, WD group had similar alpha diversity to CT group and the WD+DSS group showed the lowest alpha diversity compared to the other treatment groups. There was a significant interaction between western diet and DexSS for beta diversity, based on Bray-Curtis dissimilarly. The westernized diet and DexSS resulted in three and seven differentially abundant phyla, 21 and 65 species, respectively, mainly associated with the Firmicutes and Bacteroidota phyla followed by Spirochaetota, Desulfobacterota, and Proteobacteria. The concentration of short-chain fatty acids (SCFA) was lowest in the distal colon. Treatment had a slight effect on the estimates for microbial metabolites that might have valuable biological relevance for future studies. The concentration of putrescine in the colon and feces and that of total biogenic amines was highest in the WD+DSS group. We conclude that a westernized diet could be a potential risk factor and an exacerbating agent for UC by reducing the abundance of SCFA-producing bacteria, increasing the abundance of pathogens such as Helicobacter trogontum, and by increasing the concentration of microbial proteolytic-derived metabolites in the colon. inflammatory bowel disease ulcerative colitis meat consumption colonic inflammation porcine model 16S rRNA gut metagenomics dextran sulfate sodium Independent Research Fund Denmark 10.13039/501100004836 1335-00116 ==== Body pmc1. Introduction Ulcerative colitis (UC) is a human inflammatory bowel disease (IBD) characterized by chronic, nonspecific inflammation in the colonic epithelium (Granlund et al., 2013) as a result of chronic dysregulation of the immune response in colonic mucosa (Eichele and Kharbanda, 2017). While the exact etiology of IBD is not yet known, host genetic predisposition (McGovern et al., 2015), aggravated immune response to host gut microbiota (Roselli and Finamore, 2020), and environmental factors such as diet and altered gut microbiota have been shown to play a role in the incidence of IBD (Guo et al., 2020; Galipeau et al., 2021). Consumption of a western diet, characterized by being rich in animal protein and fat and low in dietary fiber, is considered a main factor contributing to increased UC risk (Chiba et al., 2019). Red meat is one of the ingredients of a typical western diet (Thøgersen and Bertram, 2021). In human studies, red meat consumption has been associated with UC, by inducing or exacerbating this disease (Jantchou et al., 2010; Ge et al., 2015; Rashvand et al., 2015) due to increased proteolytic activity by the colonic bacteria resulting from higher substrate availability (Gilbert et al., 2018). Although the exact mechanism behind this effect is yet to be elucidated (Van Hecke et al., 2021), microbial fermentation of undigested dietary and endogenous protein sources results in various metabolites, several of which, e.g., biogenic amines, NH4+, indoles, phenols, and H2S, are considered detrimental for gut health as they are capable of stimulating the immune system and starting a cascade of inflammatory responses in the colon (Windey et al., 2012; Gilbert et al., 2018; Panah et al., 2021). These metabolites can result in the loss of microvilli and apoptosis of colonocytes, which are ultrastructural changes in the colon at the early stages of diarrhea due to colitis (Thomson, 2009; Gilbert et al., 2018). Red meat intake leads to the formation of reactive oxygen species and it may also damage the colonic epithelium through increased concentrations of heme (originated from red meat haemoglobin), and amines (Glei et al., 2006). Accordingly, Le Leu et al. (2013) reported that feeding red meat to mice with colitis (induced by dextran sulfate sodium; DexSS) exacerbated colitis as observed by increased histopathological scores in the colon and changes in the gut microbial composition. Dysbiosis of colonic microbiota was reported to be associated with IBD (Ni et al., 2017) as gut microbiota are involved in various physiological functions, critical in colonic inflammation such as energy and nutrient supply (Ng et al., 2013), maintenance of mucosal homeostasis (Zheng et al., 2020), and modulation and enhancement of host immunity (Yang et al., 2021). Dysbiosis has been defined as reduced biodiversity, abnormal and varying composition of large intestine microbiota, alongside changes in the interactions among different strains of microbiota and between microbiota and the host (Guo et al., 2020). However, since mucosal inflammation in the colon can lead to dysbiosis, it remains unclear whether microbial dysbiosis is the cause or the consequence of IBD (Ni et al., 2017). It has been speculated that increased fecal proteolytic activity, associated with changes in gut microbiota, preceding the clinical diagnosis of UC, is potentially indicative of a causal relationship (Galipeau et al., 2021). Studies have shown that patients with UC may have a less diverse bacterial community in the large intestine, with lower abundance of Firmicutes, Bacteroidetes and Verrucomicrobia phyla, and higher abundance of Proteobacteria (Manichanh et al., 2012; Zakerska-Banaszak et al., 2021). Microbial dysbiosis, in return, may alter metabolic conditions in the gut (Yang et al., 2021) by reducing carbohydrate-fermenting bacteria such as Lachnospiraceae (in particular Roseburia spp.), and Prevotella and Ruminococcus taxa, which are involved in production of short-chain fatty acids (SCFAs; Louis et al., 2007). Short-chain fatty acids such as acetate, butyrate and propionate, and lactate are produced from bacterial fermentation of indigestible carbohydrates, and to a lesser extent from protein, in the colon, and exert vital beneficial effects to the maintenance of mutualistic relationships between commensal bacteria and the host immune system (Rooks and Garrett, 2016). The intestinal immune system is affected by SCFAs, since they inhibit pathogenic growth (Vinolo et al., 2011), regulate inflammatory responses in the colon (Maslowski et al., 2009), participate in secretion of the epithelial repair cytokine interleukin-18, and they also enhance the integrity of epithelial cells by promoting goblet cell mucin production and modification of tight junctions (Guo et al., 2020). Among SCFAs, butyrate has a significant biological importance in the context of IBD. It is the preferred energy source for colonocytes (supplying 70–80% of their required energy; Galipeau et al., 2021; Gasaly et al., 2021), it regulates immune responses by activation of G protein-coupled receptors (GPRs; Sanchez et al., 2020), it renders the differentiation of monocytes to macrophages through inhibition of histone deacetylase (HDAC; Schulthess et al., 2019), and its oxidation by colonocytes generates hypoxia, helping to maintain luminal anaerobiosis, which is favorable for the hindgut microbiota (Gasaly et al., 2021). In a number of studies, IBD patients exhibit lower abundance of butyrate-producing bacteria, lower butyrate concentration (Marchesi et al., 2007) and depressed colonocyte butyrate oxidation, which lowers luminal anaerobiosis and facilitates the expansion of bacteria like Enterobacteriaceae that contribute to inflammation (Gasaly et al., 2021). Vieira et al. (2012) showed that oral administration of butyrate to mice with UC induced by dextran sulphate sodium (DexSS), ameliorated the inflammatory profile of the colon mucosa, possibly by activation of GPR109A, which contributes to the suppression of DexSS-induced acute colitis (Yang et al., 2021). Aimed at demystifying the complex pathology of IBD, many studies have been conducted using animal models (Knudsen et al., 2020; Mizoguchi et al., 2020; Nielsen et al., 2020; Mizoguchi et al., 2021). Colitis in pigs may share some symptoms with UC (Chase-Topping et al., 2007; Panah et al., 2021), and since pigs are closer to humans than rodents like mice or guinea pigs with regard to their colonic microbiota, anatomy and physiology (Liu et al., 2017), they are often used as surrogate models. Therefore, establishing inflammation in the colon of pigs could provide a better experimental model for human IBD (Pistol et al., 2020). One of the most common chemicals used to simulate active IBD in animal models is DexSS administered orally, which has previously been tested in pig (Nielsen et al., 2020; Pistol et al., 2020) and murine (Eichele and Kharbanda, 2017) colitis models. It has been combined with dietary interventions to investigate, e.g., the potential impact of food components on mitigation of IBD activity (Bassaganya-Riera and Hontecillas, 2006; Pistol et al., 2020). We have previously shown that offering a westernized diet to 6-week old pigs by replacing 15% of the pelleted standard diet with cooked minced beef 2 weeks prior to DexSS treatment, aggravated the severity of colitis (Nielsen et al., 2020). Using samples from these piglets in the current study, we aimed at investigating the impact of feeding a westernized diet on colonic and fecal microbial composition and microbial metabolites. We hypothesized that the exacerbating effect of a westernized diet on colitis in DexSS-challenged pigs is associated with changes in the gut microbiota induced by the diet composition. 2. Materials and methods 2.1. Ethics statement The care and housing of the animals used in the experiment complied with Danish laws and regulations for the humane care and use of animals in research (The Danish Ministry of Justice, Animal Testing Act no. 1306 of 23 November 2007) and was performed under the license obtained from the Danish Animal Experimentation Inspectorate, Ministry of Food, Agriculture and Fisheries. 2.2. Experimental design A total of 24 pigs in three complete blocks following a 2 × 2 factorial design was used, as described by Nielsen et al. (2020). Briefly, 24 6-week old pigs were fed either a standard weaner diet (Prime Midi Piller U; DLG, Randers, Denmark) or a diet in which 15% of the standard diet (by weight of the total pelleted diet) was substituted with minced, cooked and dried ground beef (Hørkram Foodservice, Hørning, Denmark) to simulate a westernized diet from experimental day 0 to 14. Throughout the three blocks of the experiment, two non-littermate pigs from different sows were housed in one pen to receive the same treatment, and piglets from two different sows were used in each block. The animals were fed one of the two experimental diets for 14 days. From day 14, an oral dose of DexSS was administered to half of the pigs fed the standard diet (DSS; n = 5) and half of the pigs fed the WD diet (WD + DSS; n = 6); the other half continued without DexSS administration to the end of the experiment, forming the control (CT; n = 6) and westernized diet (WD; n = 6) groups, respectively. One of the pigs treated with DexSS reached the humane end point prior to the end of experiment, hence it was euthanized and results are not included in the downstream analysis; therefore, the results are from total number of 23 pigs. The DexSS treatment was administered as 1.25 g DexSS (MW 36–50 kDa, MP Biomedicals, Santa Ana, CA, United States)/kg BW, dissolved in 20 ml of sterile saline solution (0.9% NaCl) + 5 ml apple juice. On days 18 and 19, fecal samples were collected (1.5 g) and, on day 19, all animals were sacrificed and colonic digesta were collected from the proximal (25% of colon length) and distal colon (75% of colon length) for subsequent metabolite (4.0 g) analysis and DNA extraction (1.5 g). Fecal and digesta samples for DNA extraction were snap frozen in liquid nitrogen and stored at −80°C until analysis. Samples for metabolite analysis were stored at −20°C. 2.3. DNA extraction and 16S rRNA gene sequencing Total DNA was extracted from 200 mg fecal (collected on day 18) and digesta (collected on day 19) samples using the E.Z.N.A. stool DNA Kit (Omega bio-tek) according to the manufacturer’s instructions. Amplicon libraries of the bacterial 16S rRNA gene were prepared according to Illumina’s 16S Metagenomic Sequencing Library Preparation protocol (Illumina, 2013), with few modifications as described in Tawakoli et al. (2017). Briefly, 20 PCR cycles were performed with the bacterial primers Bac341F and Bac805R (Herlemann et al., 2011) to applify V3-V4 regions of 16S rRNA gene, followed by 10 cycles to add the Illumina adapters and 8 cycles to add the sequencing barcodes. The pooled libraries were sequenced on a MiSeq desktop sequencer (Illumina) using 2 × 300 bp chemistry (Illumina) according to the manufacturer’s instructions. 2.4. Bioinformatics analysis of 16S rRNA gene data Raw sequencing reads were processed with Qiime 2 (Bolyen et al., 2019) using the DADA2 package (Callahan et al., 2016) for primer trimming, quality filtration, denoising, merging, chimera removal and inference of amplicon sequence variants (ASV). Forward and reverse primers were removed by truncating the forward (trim left forward: 17 nt) and reverse (trim left reverse: 21 nt) primer length from the raw sequence reads. Trimmed reads were quality filtered by truncating sequence reads at the points where 25% of reads dropped below a Phred quality score (Q) of 30 (truncation length of forward reads: 260; truncation length of reverse reads: 220). In this way, merging of the sequences took place with the minimum of 16 nucleotides overlap. For diversity analysis, a phylogenetic tree was inferred using the fragment insertion method based on SATé-Enabled Phylogenetic Placement (SEPP) implemented in Qiime 2, which is inserting query sequences into an existing phylogenetic tree and aligning them to full-length sequences for the same gene (Mirarab et al., 2011). For taxonomic classifications of detected ASVs, a region-specific classifier based on our primer set was created based on the SILVA v.138 (Quast et al., 2012) taxonomic reference database with 99% of similarity cutoff using the RESCRIPt plugin (Robeson et al., 2020) in Qiime 2. The classifier was trained using the Naïve-Bayes method and the output artifact was used to generate the classified taxonomy table. Our region-specific trained classifier is available on this directory: https://drive.google.com/file/d/1qwGbPvxhXJIC_bSmoB33Ala26LeN69BJ/view?usp=sharing. All files generated in Qiime 2, i.e., ASV tables, ASV sequences, taxonomy table, and rooted phylogenetic tree were transferred to R by qiime2R package (Bisanz, 2018) for subsequent use with the phyloseq package (Holmes and McMurdie, 2013). ASVs belonging to taxonomic domain other than bacteria (e.g., Archaea and Eukaryotes) were removed and subsequent results are based only on bacteria. Decontamination of the reads was performed based on prevalence of ASVs in negative samples (n = 5) using the Decontam package (Davis et al., 2018) in R from which Caldalkalibacillus uzonensis, belonging to the phylum Firmicutes was identified as contaminant and removed from the ASV table. After this step, the negative control samples were deleted from the entire dataset. Additionally, ASVs with the prevalence in less than three samples (out of 66 samples) were filtered out. This resulted in filtering ASVs appeared in less than 5% of all samples. Moreover, singletons were removed from the dataset based on their abundance as described in this repository: https://github.com/farhadm1990/Microbiota-analysis (Panah, 2022). Singletons were those ASV that occurred only in one sample across all the dataset, i.e., their mean abundance was equal to their sum of abundance. Relative abundance of taxa was determined through dividing the number of sequencing reads assigned to different taxa in each sample by the total number of sequencing reads. Finally, all samples were normalized to the same reading depth of 17,000 reads per sample by rarefaction (sampling without replacement) in phyloseq. A total of 66 samples and 2,925 ASVs passed preprocessing and were used for the downstream analysis. 2.4.1. Quantitative PCR and calculation of absolute abundance Bacterial 16S rRNA gene copy numbers per g of sample were quantified in triplicate by quantitative PCR (qPCR) according to Busck et al. (2022). In brief, a 167 bp long gene fragment was amplified using the primer set Bac908F/Bac1075R (Mateos-Rivera et al., 2018), with annealing at 59°C and extension at 78°C. In order to obtain absolute abundances, relative taxon abundances derived from the sequencing reads were multiplied to the total load of bacterial 16S rRNA gene copies (total load per gram of sample), detected by qPCR (Jian et al., 2020). All results presented in this study are based on absolute count data. 2.4.2. Alpha and beta diversity Alpha diversity, a measure of microbiota diversity within a sample, was estimated based on richness or total number of observed ASVs (Chao1) and evenness using the Shannon’s index and on phylogenic distances using the Faith Phylogenetic Diversity (FaithPD) index. Chao1 and Shannon indices were estimated from the ASV count table using the phyloseq package; for estimation of FaithPD, the ASV count table and the rooted phylogenetic tree were used as the inputs in the pd function of picante package (Kembel et al., 2010). For Chao1, the non-rarefied ASV count table was used and for Shannon and FaithPD the rarefied ASV count table was used as the input. Beta diversity, the difference in microbial composition between the treatment groups, was estimated by Bray–Curtis dissimilarity coefficients from the variance-stabilizing transformed (VST) ASV count table. Variance-stabilizing transformation was done using the varianceStabilizingTransformation function in DESeq2 package (Love et al., 2014). 2.4.3. Analysis of differentially abundant taxa Absolute ASV abundances were agglomerated to phylum and species level using the tax_glom function in phyloseq and all ASVs at family level classified as “uncultured” were removed. Normalization of the microbial data and differential abundance analysis was done using DESeq2 (Love et al., 2014) in R via a Negative Binomial Wald test statistics and associated p-value with the main effects of WD, DSS and their interaction as well as the effect of blocks and sample type as additional covariates. Before estimation of the dispersions, the geometric means of the counts in each sample were calculated and used to estimate size factors per sample using DESeq2 (Love et al., 2014). To correct for multiple hypothesis testing, p-values were adjusted for the False Discovery Rate (FDR) using the Benjamini-Hochberg (BH) method (Benjamini and Hochberg, 1995). ASVs were considered differentially abundant when the adjusted p-value ≤0.05 (FDR < 0.05) and a Log2 Fold Change (LFC) exceeding 0 (|LFC| > 0) at phylum level and |LFC| > 2 at species level. 2.5. Analysis of microbial metabolites For quantification of chemical concentrations, fecal and digesta samples were collected in stomacher bags and kept on ice and stored in −20°C until the day of analysis. Concentration of the SCFAs acetate, propionate, butyrate, iso-acids (iso-butyrate and iso-valerate), and valerate were analyzed by capillary gas chromatography as described by Jensen et al. (1995) with some modifications as described by Canibe et al. (2007). Biogenic amines (cadaverine, agmatine, putrescine and tyramine) were quantified by gradient elution on reverse phase HPLC chromatography, as described by Poulsen et al. (2018). A heatmap of associations between bacterial absolute abundance and the concentration of different metabolites produced by colonic bacteria was created using the Spearman’s rank correlation method with the pheatmap package in R (Kolde and Maintainer, 2018). A significance test was performed to identify the significant associations and the pairwise comparisons were corrected for FDR < 0.05. 2.6. Statistical analysis The relationships between predictor variables and the expected responses for alpha diversity metrics (i.e., Chao1, Shannon and FaithPD) and microbial metabolites were assessed using the R statistical software (Team R. C, 2021). A Generalized Linear Mixed-Effect Model with Gamma distribution and log link function was estimated using the glmer function of the lme4 package (Bates et al., 2015). Estimated marginal means (EMM) of treatment effects were computed using the emmeans package (Lenth et al., 2018). The model estimated had the following functional form: logEYijklm=α+Wi+Dj+Wi·Dj+Sk+Wi·Sk+Dj·Sk+Wi·Dj·Sk+Bl+Pm Where Y is the dependent variable and α the model constant term. The model includes the fixed effects of diet (Wi; i = noWD and yesWD), DexSS (Dj; j = noDexSS and yesDexSS), sample type (Sk; k = digesta from proximal and distal colon, and feces), all second- and third-order interaction between these fixed effects, experimental block (Bl; l = 1, 2, 3), and the random effect of pig (Pm; m = 1, …, 23). We report EMMs with 95% confidence interval and pairwise comparisons between treatment groups from the full model. Pairwise comparisons were adjusted for false discovery ratio by BH method and declared significantly different where p.adjust ≤ 0.05. Compositional differences between treatment groups were assessed using a distance-based redundancy analysis (dbRDA) based on Bray–Curtis dissimilarity matrix (litter effect used as a partial effect, whose effect was removed prior to estimation of treatment effects) in R using the vegan package (Oksanen et al., 2020) with 999 permutations (restricted by litter as a “block”-level restriction and by individual animal as plot strata, while permutation between segments was freely done; Simpson, 2022). Bray–Curtis dissimilarity was obtained from the distance function in phyloseq and the distance between samples in each treatment group was analyzed using non-metric multidimensional scaling (NMDS) using the metaMDS function in vegan (Oksanen et al., 2020) and phyloseq packages. The effect of segment was evaluated by a dbRDA model and tested with a permutation with restricted litter and samples from different sample type and free permutation between animals. The Bray–Curtis dissimilarity matrix was computed from log-transformed absolute counts. A test of homogeneity of group variances (dispersions around the centroids) of the four treatment groups was conducted prior to dbRDA with 999 restricted permutations (litter as a “block”-level restriction, animals were restricted via plot strata and observations from different segments were freely permuted) using the betadisper function in vegan. Unless otherwise stated, all multiple comparisons were corrected for FDR by the BH method and were declared significant at FDR ≤ 0.05. The scripts for bioinformatics and statistical analysis are available at https://github.com/farhadm1990/Microbiota-analysis (Panah, 2022). 3. Results 3.1. Feed composition Meat substitution of 15% by weight, resulted in increased protein and fat content of the diet by 40.8 and 66.7%, respectively, and reduced fiber content by 16% in westernized diet compared to the standard diet (Table 1). Table 1 Nutrient composition (% as-is basis) of the standard and westernized diet (WD). Item Standard weaner diet1 WD2 Crude protein 19.1 26.9 Crude fat 4.8 8.0 Crude fiber 2.5 2.1 Crude ash 4.7 5.2 % Protein from meat 0 39.7 % Fat from meat 0 48.9 % Ash from meat 0 23.8 1 Nutrient composition in the standard diet was based on the declared composition by the feed company (DLG, Fredericia, Denmark). 2 Crude protein in ground beef was determined as N × 6.25 and N was measured by Dumas (Hansen, 1989). Concentration of crude fat in ground beef was quantified using the Stoldt procedure (Stoldt, 1952). Total ash was analyzed according to the AOAC method (923.03; AOAC, 1990). 3.2. Alpha diversity in digesta and feces Average alpha diversity indices of feces and colonic samples from the experimental treatment are summarized in Table 2. There were no differences in alpha diversity between experimental blocks and sample type, i.e., proximal colon, distal colon, and fecal samples (Supplementary Table S1). In proximal and distal colon, WD group showed similar alpha diversity to CT group (Table 2). In proximal colon, the WD + DSS group had lowest Chao1, Shannon and FaithPD indices, while in distal colon, DSS group showed lowest alpha diversity (p < 0.05; Table 2). Chao1 and Shannon indices in distal colon were similar to that in DSS group. However, in fecal samples only FaithPD seemed to be affected by treatments and it was lowest in DSS group compared with WD group. Table 2 Alpha diversity metrics with their estimated marginal means and their 95% confidence interval for each treatment group. Groups1 CT WD DSS WD + DSS Proximal Chao1 533 (426–665)b 500 (399–625)ab 460 (360–588)ab 367 (296–462)a Shannon 5.30 (5.0–5.66)b 5.20 (4.92–5.57)b 5.0 (4.63–5.30)ab 4.80 (4.05–5.09)a Faith PD 35.4 (30.0–42.0)b 35.5 (30.0–42.0)b 29.4 (24.0–35.0)ab 27.4 (23.0–32.0)a Distal Chao1 535 (428–668)b 555 (443–695)b 397 (311–508)a 397 (315–501)a Shannon 5.30 (4.94–5.59)b 5.30 (4.96–5.63)b 4.80 (4.46–5.12)a 4.90 (4.59–5.21)ab Faith PD 36.4 (31.0–43.0)b 38.6 (32.0–46.0)b 26.9 (22.0–32.0)a 28.5 (24.0–34.0)a Feces Chao1 488 (387–615) 534 (426–668) 391 (301–506) 434 (347–542) Shannon 5.20 (4.90–5.57) 5.30 (4.95–5.62) 5.0 (4.64–5.35) 5.0 (4.71–5.33) Faith PD 34.3 (29–41)ab 37.9 (32–45)b 27.7 (23–34)a 30.6 (26–36)ab 1 Treatment groups: control (CT; n = 17), WD (CT + ground beef; n = 18), CT + dextran sulfate sodium (DSS; n = 14), WD + dextran sulfate sodium (WD + DSS; n = 17). Pairwise comparison for differences in EMMS between groups was adjusted with BH and EMMs are superscripted with different letters at p < 0.05. 3.3. Bacterial beta diversity of digesta and feces The test for differences of bacterial diversity between the samples was done by dbRDA based on Bray–Curtis dissimilarity metric. To verify the validity of the test, the homogeneity of variance around the centroids for each treatment was examined (Supplementary Figure S1). Principal Coordinate Analysis (PCoA) plot shows the within group dispersion based on Bray-Curtis dissimilarities (Supplementary Figure S1) and illustrates the homogeneity of variances test (betadisper). This test showed no significant differences between groups in terms of dispersion around the treatment centroids (F = 5.12; p = 0.18). Accordingly, PCo1 and PCo2 explained 24.0 and 12.0% of the total variance, respectively. The results of the dbRDA model (Table 3) showed that there was a significant interaction between diet and DexSS (W·D; Pseudo-F = 10.8; p < 0.01) based on Bray-Curtis dissimilarities. However, the main effects of diet (Pseudo-F = 4.90; p < 0.01) and DexSS (Pseudo-F = 21.1; p < 0.01) were also significant. The magnitude of these effects was observed in samples from DexSS-treated groups based on the dbRDA model, showing that DexSS treatment explained 22.0% of the total variance (p < 0.01), while diet accounted for 5.0% of the total variance (p < 0.01) in the beta diversity based on Bray-Curtis dissimilarities. There was a distinct separation in Bray-Curtis dissimilarities on NMDS plot (Figure 1A) between treatment groups and the resolution of this separation became sharper when the sample scores of different treatments, extracted from dbRDA model, were plotted on PCoA plot (Figure 1B). According to the dbRDA plot of the sample scores, dbRDA1 explained 70.1% of the fitted variance and 21.7% of the total variance and dbRDA2 explained 16.9 and 5.20% of the fitted and the total variance, respectively. Although statistically significant, the effect of sample type on Bray–Curtis dissimilarity was small as it only explained 1.0% of the total variance based on the dbRDA model (Pseudo-F = 0.55; p < 0.01). Graph-based analysis on the Bray–Curtis dissimilarity matrix and permutation test based on Minimum Spanning Tree (MST) showed pure edges for each of the treatment groups, i.e., pure edges indicate similar bacterial composition, while mixed edge connections could also be seen between CT and WD, WD + DSS and DSS groups, and between WD + DSS and WD groups, i.e., mixed edges indicate samples with tendency to similar composition of bacteria (Figure 1C). Table 3 Test statistics of the dbRDA model for the effect of treatment and sample type on Bray–Curtis dissimilarity. Degree of freedom % of variance explained Sum of squares Pseudo-F p-value WD 1.0 5.0 0.05 4.90 0.02 DSS 1.0 22.0 0.23 21.1 <0.01 WD · DSS 3.0 31.0 0.33 10.8 <0.01 Sample type 2.0 1.0 0.12 0.55 <0.01 The values are averaged over three blocks. Figure 1 Non-metric multidimensional scaling (NMDS) plot based on Bray–Curtis dissimilarity index (A), and dbRDA MDS plot (B) of sample scores from Bray–Curtis dissimilarity for different treatments (CT: control, WD: westernized diet, DSS: group treated with dextran sulfate sodium, and WDDSS: WD + DSS) extracted from dbRDA model (variance explained by our treatments = 31.0%, p < 0.01). On dbRDA plot, the treatments (constrained factors) explained 70.1 and 16.9% of the fitted variance and 21.7 and 5.20% of total variance for Bray–Curtis dissimilarity on dbRDA1 and dbRDA2 axis, respectively. The shape of the points represents the origin of the samples, i.e., digesta from proximal and distal colon, and fecal samples and the color of the points represent different treatments. Graph-based analysis of the distributions in bacterial composition for different treatments (C), based on Bray–Curtis dissimilarity matrix with maximum distance of 0.35. The histogram of permutation test based on MST for Bray-Curtis is presented (p < 0.01). 3.4. Bacterial composition in digesta and feces Our results showed that by targeting V3 and V4 of 16S rRNA gene, 175 out of 573 ASV (30.5%) was classified as known species and at genus level, 221 out of 235 unique genera (94%) were classified as known. Bacterial 16S rRNA gene load in the samples from feces, proximal and distal colon was analyzed by qPCR. Total bacterial load was 1.47 × 1010 16S rRNA gene copies per g sample and it was slightly higher in WD group and lowest in DSS and WD + DSS groups. Regardless of the treatment, Firmicutes and Bacteroidota were the most abundant phyla in all groups, followed by Spirochaetota and Proteobacteria. The absolute abundance of Firmicutes was numerically lowest in WD + DSS and DSS compared to the CT and WD groups, 1.98 × 109 and 2.04 × 109 vs. 2.52 × 109 and 2.97 × 109 gene copies per g, respectively (Figure 2A). Absolute abundance of Bacteroidota and Fibrobacterota was lowest in the DSS and WD + DSS groups, and that of Spirochaetota was highest in the DSS and WD + DSS groups. Fusobacteriota was absent only in samples from the CT group, and Cyanobacteria and WPS-2 were absent in the DSS group (Figure 2A). Figure 2 Absolute abundances of different phyla with total abundance across all sequence reads in different treatments (A). The number of different phyla present in the treatment groups was 16, 16, 11, and 17 for CT, WD, DSS, and WDDSS (WD + DSS), respectively. Venn diagram of shared taxa between different treatments at Phylum (B) and Species (C) levels. At Phylum level, 10 phyla were shared by all treatments: Actinobacteriota, Bacteroidota, Campilobacterota, Deferribacterota, Desulfobacterota, Firmicutes, Patescibacteria, Proteobacteria, Spirochaetota, and Verrucumicrobiota (Figure 2B). At species level, all groups shared 272 species and the most closely related groups, regarding uniquely shared species, were CT and WD, with 54 species in common followed by DSS and WD + DSS, with 28 shared species (Figure 2C). 3.5. Differential abundance analysis of taxa in feces and digesta Sample type, i.e., proximal and distal colon and feces, had no effect on differential abundance of bacteria at phylum level. At phylum level, the main effect of westernized diet and DexSS resulted in 3 and 7 differentially abundant phyla, respectively (Figures 3A,B). Figure 3A shows that the animals fed with the westernized diet had low abundance of Actinobacteriota (LFC = −0.76) and Firmicutes (LFC = −0.46) and increased abundance of Fusobacteriota (LFC = 16.8). Administering DexSS resulted in a significant decrease in the abundance of Verrucumicrobiota (LFC = −4.59), Actinobacteriota (LFC = −0.96), and Bacteroidota (LFC = −0.56); and it increased the LFC for Spirochaetota (0.81), Proteobacteria (1.34), Deferribacterota (7.50) and Fusobacteriota (25.0). There was a significant interaction between diet and DexSS administration for differential abundance of 7 phyla, and pairwise comparisons of the different treatment groups for these phyla are presented in Figures 3C–F. In the WD, DSS and WD + DSS treatment groups, the abundance of Planctomycetota was reduced compared to the CT group. The WD group had higher abundance of Cyanobacteria with LFC = 30.0 (Figure 3C), and DSS (Figure 3D) and WD + DSS (Figure 3E) showed lower abundance of WPS-2 compared to the CT group. Comparing WD + DSS with WD, WD + DSS had higher abundance of Fibrobacterota and Campilobacterota, while it showed lower abundance of Cyanobacteria and WPS-2 (Figure 3F). Figure 3 Differentially abundant Phyla for the main effect of diet (westernized diet groups vs. non-westernized group; A) and DexSS (DexSS groups vs. non-DexSS groups; B), and pairwise comparison for phyla with significant interaction for WD vs. CT (C), DSS vs. CT (D), WDDSS (WD + DSS) vs. CT (E), and WDDSS (WD + DSS) vs. WD (F) with FDR < 0.05. Figures 4A,B depict differentially abundant species for the main effects of diet and DexSS administration with |LFC| > 2 and FDR < 0.01. Diet and DexSS administration impacted the abundance of 21 and 65 ASVs, respectively. The groups fed the westernized diet showed reduced abundance of 15 species, among which 8 species belonged to the Firmicutes phylum and 5 to the Bacteroidota, compared to those fed the standard diet. In addition, the westernized diet resulted in increased abundance of 6 species belonging to the Firmicutes and Patescibacteria phyla (Figure 4A). Administration of DexSS decreased the abundance of 35 species (Figure 4B), while it was associated with increased abundance of 30 species. Differentially abundant species associated with administration of DexSS were chiefly from the phyla Firmicutes and Bacteroidota followed by Spirochaetota, Desulfobacterota, and Proteobacteria. Figure 4 Waterfall plot for differentially abundant species of the main effect of diet (westernized diet groups vs. non-westernized diet groups; A) and DexSS (DexSS groups vs. non-DexSS groups; B). Different colors represent different phyla. Only species with |LFC| > 2 and FDR < 0.01 are presented. Pairwise comparisons between treatment groups for species, where there was a significant interaction between the effect of diet and DexSS are presented in Supplementary Figure S2A for WD vs. CT, in Supplementary Figure S2B for DSS vs. CT, and in Supplementary Figure S2C for WD + DSS vs. CT. Compared to CT, the WD, DSS and WD + DSS groups had a different abundance of 25, 54, and 43 species, respectively. Animals in the WD group showed lower abundance of species, such as Anaerovibrio sp., Lachnospiraceae NK4A136_group uncultured organism, and Olsenella umbonata, and an increased abundance of Helicobacter trogontum, Clostridium perfringens, Mogibacterium sp., and Roseburia hominis compared with the CT group. 3.6. Concentration of short chain fatty acids and biogenic amines in colon and feces Overall, except for valerate, which was constant in different sample types, the concentrations of all SCFAs varied according to the origin of the sample (Table 4). In proximal colon, the concentration of total SCFAs, acetate, propionate, and butyrate was higher than in distal colon but there was no difference between samples from the proximal colon and feces. On the other hand, valerate concentration was higher in fecal samples compared with those from the proximal and distal colon and concentration of iso-acids was constant across all sample types. Table 4 Concentration (mmol/kg wet sample) of short-chain fatty acids (SCFA) in proximal and distal colon digesta and in fecal samples1. Sample type2 Proximal colon Distal colon Feces SCFA 99.9 (88.0–114)b 82.7 (72.0–95.0)a 98.6 (85.0–115)b Acetate 60.8 (53.0–70.0)b 48.9 (42.0–57.0)a 57.4 (49.0–68.0)ab Propionate 22.0 (19.0–25.0)b 17.4 (15.0–20.0)a 20.1 (17.0–23.0)b Butyrate 12.4 (10.0–15.0)ab 11.1 (9–14)a 13.2 (11.0–16.0)b Valerate 2.60 (2.0–4.0)a 2.50 (2.0–3.0)a 3.30 (2.0–5.0)b Iso-acids 1.70 (1.0–3.0) 2.10 (1.0–4.0) 3.30 (1.0–8.0) 1 Response values are reported with their corresponding 95% confidence intervals. Pairwise comparison of EMMs was adjusted using BH method and rows with different superscripts have different EMMs at p < 0.05. 2 Samples from different treatment groups are pooled in each segment, i.e., proximal colon (n = 23), distal colon (n = 22), and in feces (n = 21). Table 5 and Supplementary Figure S3 show the concentration of SCFAs in the three sample types for the four treatment groups. These results were derived from the full model due to non-zero and relatively large coefficient for the second- and third-order interactions between treatment groups and sample type. Therefore, we report EMMs (with lower and upper confidence intervals) and pairwise comparison between groups. Although we did not observe statistical differences in the adjusted pairwise comparisons, possibly due to relatively high uncertainty in the estimates, there are numerical patterns in the effect of treatments on microbial metabolites that might have valuable biological meanings for future studies. In all sample types, WD and DSS groups showed numerically higher concentrations of SCFA and butyrate compared to CT, and WD + DSS the lowest compared to all other groups. Table 5 Concentration (mmol/kg wet sample) of short-chain fatty acids (SCFA) measured in digesta from the proximal and distal colon and in feces. Groups1 CT WD DSS WD + DSS Proximal SCFA 96.4 (75.0–124) 110 (85.0–141) 114 (83.0–156) 82.8 (65.0–106) Acetate 61.3 (47.0–80) 69.1 (53.0–90.0) 66.3 (47.0–93.0) 48.8 (37.0–64.0) Propionate 20.9 (16.0–27.0) 22.8 (18.0–29.0) 25.2 (19.0–34.0) 19.4 (15.0–25.0) Butyrate 10.1 (7.0–15.0) 12.5 (9.0–18.0) 17.1 (11.0–27.0) 10.8 (8.0–15.0) Valerate 2.0 (1.0–4.0) 2.70 (1.0–5.0) 3.70 (2.0–8.0) 2.30 (1.0–4.0) Iso-acids 1.70 (1.0–4.0) 2.40 (1.0–6.0) 1.50 (1.0–4.0) 1.40 (1.0–3.0) Distal SCFA 74.3 (58.0–96.0) 91.2 (71.0–118) 95.4 (67.0–136) 72.5 (54.0–97.0) Acetate 44.0 (34.0–58.0) 53.4 (41.0–70.0) 58.0 (39.0–86.0) 42.0 (30.0–58.0) Propionate 15.4 (12.0–20.0) 17.4 (14.0–22.0) 20.2 (14.0–28.0) 17.1 (13.0–23.0) Butyrate 10.1 (7.0–15.0) 12.8 (9.0–19.0) 13.5 (8.0–22.0) 8.80 (6.0–13.0) Valerate 2.10 (1.0–4.0) 3.40 (2.0–6.0) 3.20 (1.0–7.0) 1.80 (1.0–3.0) Iso-acids 2.50 (1.0–6.0) 3.90 (2.0–9.0) 1.30 (0.98–5.0) 1.30 (0.97–4.0) Feces SCFA 88.3 (68.0–115) 120 (93.0–154) 94.2 (63.0–140) 94.9 (69.0–131) Acetate 55.1 (41.0–73.0) 72.2 (55.0–94.0) 49.8 (32.0–78.0) 54.6 (38.0–79.0) Propionate 17.7 (14.0–23.0) 22.6 (18.0–29.0) 20.5 (14.0–30.0) 20.0 (15.0–27.0) Butyrate 9.30 (6.0–14.0) 14.2 (10.0–21.0) 17.1 (10.0–28.0) 13.4 (9–20) Valerate 2.50 (1.0–5.0) 4.40 (2.0–8.0) 5.60 (3.0–12.0) 2.0 (1.0–4.0) Iso-acids 3.40 (1.0–9.0) 6.0 (3.0–14.0) 3.40 (1.0–15.0) 1.60 (0.96–5.0) Response EMM values are reported with their corresponding 95% confidence intervals. 1 Treatments group: control (CT; n = 17), westernized diet (CT + ground beef; n = 18), CT + dextran sodium sulphate (DSS; n = 9), WD + dextran sodium sulphate (WD + DSS; n = 13). Pairwise comparison for differences in EMMS between groups was adjusted with BH and EMMs are superscripted with different letters at p.adjust < 0.05. Concentrations of biogenic amines in proximal and distal colon and feces are summarized in Table 6. In distal colon, the concentration of total biogenic amines was highest in the WD + DSS group (p < 0.05). In proximal and distal colon, concentration of putrescine was increased in the DSS and WD + DSS groups (p < 0.05) compared with CT; and in feces, it was highest in the WD + DSS group (p < 0.05). In general, concentrations of biogenic amines were highest in proximal and distal colon samples compared to feces (Supplementary Table S2). Table 6 Concentration (mmol/kg wet sample) of biogenic amines in proximal colon, distal colon and fecal samples. Groups1 CT WD DSS WD + DSS Proximal Biogenic amines 162.5 (69.0–380) 356 (152–833) 280 (97.0–803) 535 (236–1,216) Agmatine 16.7 (5.0–55.0) 18.8 (6.0–62.0) 31.3 (7.0–140) 46.3 (14.0–154) Putrescine 38.0 (23.0–63.0)a 53.0 (32.0–87.0)a 143 (77.0–266)b 216 (132–352)b Cadaverine 105 (36.0–310) 271 (93.0–792) 115 (30.0–440) 248 (87.0–706) Distal Biogenic amines 135 (58.0–313)a 187 (80.0–438)ab 371 (120–1,142)ab 564 (226–1,403)b Agmatine 14.2 (4.0–47.0) 15.1 (5.0–50.0) 35.7 (6.0–199) 65.0 (15.0–281) Putrescine 34.1 (21.0–56.0)a 39.0 (24.0–64.0)a 213 (106–427)b 232 (130–415)b Cadaverine 81.1 (28.0–236) 116 (40–340) 121 (29.0–500) 253 (80.0–799) Feces Biogenic amines 177.2 (73.0–428) 144 (61–338) 215 (62.0–744) 389 (141–1,069) Agmatine 12.6 (3.0–47.0) 12.2 (4.0–40.0) 17.9 (2.0–151) 23.7 (4.0–141) Putrescine 30.4 (18.0–52.0)a 25.3 (15.0–42.0)a 43.7 (19.0–99)a 168 (84.0–336)b Cadaverine 107 (35.0–328) 91.6 (31.0–269) 132 (28.0–625) 202 (58.0–707) 1 Treatments group: control (CT; n = 17), westernized diet (CT + ground beef; n = 18), CT + dextran sodium sulphate (DSS; n = 9), WD + dextran sodium sulphate (WD + DSS; n = 13). Pairwise comparison for differences in EMMS between groups was adjusted with BH and EMMs are superscripted with different letters at p.adjust < 0.05. Variations between colonic segments was inconsiderable for beta diversity and differentially-abundant taxa specifically; therefore, the microbiota data were pooled for samples taken from the different segments to evaluate the association of top 100 abundant species (chosen based on the highest standard deviation in the association matrix) with fermentation products by a Spearman rank test. Figure 5 shows that concentration of SCFAs in the large intestine was positively associated with 9 species and negatively associated with 14 species (p < 0.05), mainly from Firmicutes. Five species (mainly from Bacteroidota, Actinobacteriota, and Firmicutes) were positively associated with butyrate production, while 9 species (mainly belonged to Firmicutes and Proteobacteria) had a negative association with butyrate concentration. Total concentration of biogenic amines showed a positive association with 25 species and negative association with 34 species (p < 0.05). Figure 5 Heatmap of Spearman’s rank correlation coefficient (rs) between the top 100 abundant species and SCFA and biogenic amines in all segments. Columns are clustered based on the standard deviations between species abundances, i.e., those with lower deviation are clustered together. Species are color-labeled with their correspondent phyla and significant associations are labeled with stars. 4. Discussion Due to technical variations in 16 s rRNA gene amplicons as the result of differences in the quality of sequencing machines and individual runs, the data generated from next generation sequencing (NGS) machines are often treated as relative abundance rather than their absolute counts (Fernandes et al., 2014). However, since reads in the compositional data have an inherit mutual dependence, i.e., with increase in one feature, the other should decrease to maintain the overall sum, relative abundant NGS data can lead to misinterpretation of microbial community structure (Jian et al., 2020). This mutual dependence, therefore, makes it difficult to build a reliable biological conclusion over the mere effect of the environmental factor on the microbial phylotypes (Galazzo et al., 2020); it also hampers the actual detection of the direction and the extent of changes in taxa abundance (Morton et al., 2017; Props et al., 2017). For quantitative microbiome profiling in their absolute counts, different techniques such as flow cytometry (for total cell count; Props et al., 2017), and DNA-based methods like qPCR (Liu et al., 2015) and spike-in bacteria (Stämmler et al., 2016) have been suggested to normalize the relative abundance data. In the present study, the results of absolute (from qPCR 16 s rRNA copy-number genes) abundance of NGS data are presented. Despite efforts to better understand the pathology of IBD, the underlying mechanisms associated with these diseases are not yet understood. Disruption in the compositional balance of gut microbiota, also known as dysbiosis, resulting from dietary factors, medication, oxygen availability, or immune reactions, has been associated with IBD (Wei et al., 2021). Among dietary factors, red meat, as an important constituent of a typical western diet, has repeatedly been identified as a risk factor for the incidence of IBD due to dysbiosis of gut microbiota (Le Leu et al., 2013; Li et al., 2021). Nielsen et al. (2020) showed that administering DexSS induced colitis in pigs and that performance (e.g., feed intake and average daily gain) was negatively affected by treatment with DexSS. Furthermore, the inflammation scores in the colon and colonic expression of pro-inflammatory cytokines such as interleukin 6 (IL-6) were increased when providing DexSS (Nielsen et al., 2020). This is in agreement with previous studies, reporting that DexSS administration successfully induced UC in porcine (Pistol et al., 2020) and murine (Eichele and Kharbanda, 2017) models. In our study, DexSS administration was associated with a significant reduction in the total number of ASVs and alpha diversity in colonic and fecal samples. Previous studies have also reported reduced diversity of gut microbiota in patients with IBD (Ott et al., 2004; Scanlan et al., 2006; Schirmer et al., 2019). According to Nielsen et al. (2020), although feeding a westernized diet by adding beef did not result in inflammation and expression of pro-inflammatory cytokines, it exacerbated the clinical and histological signs of colitis in pigs challenged with DexSS. In the current study, the same westernized diet did not affect alpha diversity in pigs, while it significantly resulted in differential abundance of phyla (Figure 3A) and species compared to a control group (Supplementary Figure S2A). Nevertheless, the impact of the westernized diet on beta diversity was relatively small and, in general, gut microbiota in the WD group had uniquely more species in common with the CT group (n = 54) than with DSS and WD + DSS animals, indicating that adding 15% beef to a control pig diet, with higher fiber content than a typical human western diet, as a model for a westernized diet may be insufficient to provoke the microbial shifts that lead to dysbiosis-related disease. On the other hand, the experimental period in the current study might have been too short in order to see a more significant impact on gut microbial composition or dysbiosis. 4.1. Bacterial composition Most species affected by diet and DexSS administration belonged to Firmicutes, Bacteroidota, Actinobacteriota and Proteobacteria, which represent core phyla in human (Alam et al., 2020) and pig (Holman et al., 2017) gut microbiota. Both, feeding the westernized diet and administering DexSS resulted in an increased abundance of Fusobacteriota, a result which has previously been reported in piglets with diarrhea (Hermann-Bank et al., 2015) and in patients with UC (Rajilić-Stojanović et al., 2013). Fusobacterium mortiferum, belonging to the Fusobacteriota, was increased as a result of feeding the westernized diet and of administrating DexSS (LFC > 10), and Fusobacterium sp. has been characterized as an opportunistic pathogen that thrives in patients with UC (Rajilić-Stojanović et al., 2013). Nevertheless, the impact of F. mortiferum on UC development is difficult to discern since it has been reported that a strain of F. mortiferum inhibited the growth of both Gram-negative and Gram-positive bacteria by producing a bacteriocin-like substance (Portrait et al., 2000). In the current study, the abundance of Actinobacteriota was slightly decreased in the groups receiving the westernized diet and in those receiving DexSS as compared to those fed the standard diet and those not receiving DexSS, respectively. This corresponded to the reduced abundance of an uncultured Olsenella species belonging to this phylum. Pairwise comparison between treatments also showed lower abundance of O. umbonata in the WD group than in the CT group (Supplementary Figure S2A). Compared to the CT group, the DSS group had lower abundance of Bifidobacterium pseudolongum and higher of O. umbonata and of an uncultured Colinsella species. Olsenella umbonata and Colinsella are Gram-positive lactic-acid-producing bacteria belonging to Actinobacteriota phylum (Kraatz et al., 2011), which O. umbonata might be involved in regulating the expression of SCFA receptors in the gut epithelium (Wang et al., 2020). Other studies have reported increased abundance of the Actinobacteriota phylum in patients with UC (Alam et al., 2020; Putignani et al., 2021), while Alam et al. (2020) reported a reduction of Actinobacteriota at class (Coriobacteriia) and family (Bifidobacteriaceae) levels compared to healthy groups. Therefore, to get a better picture of bacterial association with the disease status, studying changes at ASV/species taxonomic level would be more informative over changes at phylum level. At ASV level, in animals fed westernized diet, we observed significantly reduced abundance of species from Firmicutes, Bacteroidota and Actinobacteriota phyla, which are mainly involved in carbohydrate fermentation such as Lachnospiraceae NK4A136_group, an uncultured organism at species level within Firmicutes. Lachnospiraceae NK4A136_group was also observed at lower abundances in DexSS-treated pigs, which is in agreement with previous studies reporting reductions of these ASVs in mice challenged with DexSS and fed a diet high in red meat (Li et al., 2021). Ma et al. (2020) reported that the Lachnospiraceae NK4A136 group are SCFA-producing bacteria and its increased abundance was correlated with enhanced gut barrier function in mice. In contrast, our results could not confirm a positive correlation between Lachnospiraceae NK4A136_group and the concentration of SCFA. However, the DSS group had lower abundance of Lachnospiraceae NK4A136_group, when compared to CT group. Bacteroides vulgatus, a pathogenic Bacteroides in the development of IBD (Shiba et al., 2003), was increased in pigs challenged with DexSS. This was in agreement with the results of Galipeau et al. (2021), who observed increased abundance of B. vulgatus, expressing proteolytic activity. Moreover, the abundances of other pathogens like B. coprophilus (Bacteroidota phylum), Clostridium perfringens (Firmicutes phylum), H. trogontum and Campylobacter fetus (both Campilobacterota phylum), which have been reportedly associated with IBD in humans (Loss et al., 1980; Hansen et al., 2011; Banaszkiewicz et al., 2014; Wu et al., 2021), were increased in DexSS-treated animals. The WD group also resulted in increased abundance of C. perfringens and H. trogontum compared with CT group. Clostridium perfringens seems to be the most well-documented pathogen linked to IBD (Banaszkiewicz et al., 2014), which produces an enterotoxin that causes damages to the villus tip cells by binding to their receptors and forming pores (Smedley III et al., 2008) that may disrupt the integrity of colonic epithelium. Nomura et al. (2021) reported that Parabacteroides merdae (Bacteroidota phylum) was among other Bacteroidota species showing a strong correlation with the UC activity metrics and we observed a significant increase in the abundance of this species in DexSS-treated pigs. In agreement with our results, Galipeau et al. (2021) also reported a significant increase in the relative abundance of P. merdae in the feces of individuals prior to the onset of UC; and in those with active UC, its abundance was significantly higher than in healthy, control groups. This can suggest this species as a potential biomarker for monitoring UC. 4.2. Impact of the treatments on microbial metabolites Short-chain fatty acids, in particular butyrate, improve gut maturity and health, and help maintain gut integrity and barrier functionality (Panah et al., 2021). In different studies, the reduced diversity of carbohydrate-degrading anaerobic species in the colon such as Lachnospiraceae and Ruminococcaceae families (Schirmer et al., 2019) and Bacteroides, Eubacterium, and Lactobacillus species (Ott et al., 2004) have been associated with the development of IBD. In the present study, WD-fed animals exhibited an increase in the abundance of Pasteurella aerogenes and Anaerovorax uncultured bacterium, which were negatively correlated with the concentration of SCFA (Figure 5). On the other hand, WD-fed animals showed a decreased abundance of an uncultured Frisingicoccus species and the genus Sutterella, which also had a negative association with SCFA concentration. In DexSS-treated groups, the abundance of an uncultured Anaerovorax and Roseburia uncultured organism increased, and the abundances of both were negatively correlated with the production of SCFAs. However, the effects of WD and DSS on SCFAs did not completely follow the changes in the abundance of species correlated with SCFA production. Due to the presence of uncertainties in our estimated EMMs and perhaps also due to rather small sample sizes, we could not capture statistically significant differences in SCFA concentrations, while their numerical variations due to treatment effects might be of interest for future studies. In colon and in fecal samples, WD and DSS groups had numerically higher concentration of SCFAs regardless of the fact that in these animals, species negatively correlating with SCFA production, were in higher abundance. Butyrate was also numerically higher in the DSS and WD groups compared to CT animals in both colonic digesta and fecal samples, while many studies have reported an alleviating effect of butyrate on IBD severity (Chen et al., 2018; Silva et al., 2018; Couto et al., 2020). An explanation to this may in part be due to alterations in the capacity of the colonic epithelium to transport and oxidize butyrate, caused by feeding the westernized diet and DexSS administration. As shown in Table 4, overall, from proximal to distal colon, 22.9% of total produced SCFAs concentration was reduced, while that for WD, DSS, and WD + DSS groups was 16.3, 16.3, and 12.4%, respectively. Ferrer-Picón et al. (2020) also reported that reduced butyrate-producing bacteria in the feces of active IBD patients did not correlate with the concentration of butyrate recovered from fecal samples. They related this phenomenon to the fact that increased pro-inflammatory cytokine TNF-α in the inflamed colon of IBD patients rendered the epithelium less responsive to butyrate. Likewise, De Preter et al. (2011) reported that butyrate oxidation in the colonic epithelium of patients with UC was lower than in healthy controls and increasing the concentration of butyrate did not result in a higher oxidation of butyrate in the epithelium. Although this needs to be further investigated, it could be speculated that WD and DexSS rendered colonic epithelium less responsive to butyrate oxidation. On the other hand, Nielsen et al. (2020) did not observe any significant changes in the expression of mRNA related to TNF-α in the colonic mucosa of pigs receiving the westernized diet and/or DexSS. Furthermore, our results showed that except for valerate, the concentration of total SCFA and individual SCFAs in proximal colon and in feces were not different as it decreased from proximal through distal colon and increased again in feces. This might indicate that fecal samples could represent the concentration of SCFAs formed in proximal colon, but it may not be a good representative sample for explaining the fate of SCFAs across the colon, i.e., from proximal to distal colon. Biogenic amines are produced by gut bacteria through decarboxylation of amino acids and their recovery in feces was reported to be a relevant biomarker for IBD (Maráková et al., 2020). The concentration of putrescine increased by administering DexSS and the effect was exacerbated when combined with the westernized diet, with a similar trend observed for total biogenic amine concentration. This was corresponded with increased abundance of species positively correlated with biogenic amine concentration. Changes in microbiota composition and/or impaired absorption capacity of the inflamed colon epithelium could be contributing factors to the observed higher concentration of biogenic amines in the DSS group. An increased substrate availability, i.e., more protein, by providing red meat to the DSS-challenged animals (WD + DSS) and a more severe inflammation seen in this group (Nielsen et al., 2020) would then lead to the exacerbated values measured in this group. Gobert et al. (2018) and Liu et al. (2020) also observed increased concentration of putrescine in DexSS-treated mice with a positive correlation to the inflammation scale of challenged mice. The exact mechanisms behind our observations remain unclear and further investigations on the epithelial oxidative capacity in relation to SCFAs, especially butyrate, and functional analysis of gut microbiota in patients with UC are required. In agreement with the findings of Nielsen et al. (2020), who suggested potential worsening effects of westernized diet on UC status, westernized diet could be a risk factor and an exacerbating agent for IBD by reducing SCFA-producing bacteria, increasing the abundance of pathogens and microbial proteolytic activity, and possibly by reducing butyrate absorption and oxidation capacity of colonic epithelium. Nevertheless, these results were driven from a relatively small sample size of animals (n = 5 or 6) per treatment and require caution in interpretation as also stated by Nielsen et al. (2020). The current study was conducted with pigs; therefore, a typical pig diet was used as a basis to add the beef meat. Pig diets have a relatively higher fiber content than a typical human diet in the Western countries. This fact has probably affected the results obtained regarding the impact of adding beef to the diet, as indicated by Pituch-Zdanowska et al. (2015), supplementation of some types of dietary fiber can help alleviate lesions of the intestinal mucosa during the course of the UC disease. 5. Conclusion We observed that the westernized diet, especially in combination with DexSS, resulted in significant changes in abundance of bacterial species involved in SCFA production. Providing DexSS lead to increased biogenic amine concentration, mainly putrescine in colon and in feces, and the effect was exacerbated when the westernized diet was fed to DexSS-challenged pigs. However, further investigations are required for studying the mechanisms behind the changes observed. We also confirm a significant change in colonic and fecal microbial composition in DexSS-induced UC pigs. Using a pig model in which the control diet has a lower fiber content, simulating more a human diet, would most probably be a better model to be used as surrogate for humans when investigating UC in relation to dietary interventions. Data availability statement The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number (s) can be found at: https://www.ncbi.nlm.nih.gov/, PRJNA867563. Ethics statement The animal study was reviewed and approved by The Danish Ministry of Justice, Animal Testing Act no. 1306 of 23 November 2007. Author contributions FP was responsible for bioinformatics and statistical analysis and for writing the first draft of the manuscript. KN performed all DNA-based analyses under the guidance of AndS and the microbial metabolite analysis under the guidance of NC. GS contributed to checking statistical and bioinformatics analysis of the data and correcting the manuscript for language and grammar. AnnS contributed to checking bioinformatics analysis and contributed to scientific proof reading of the manuscript. AndS, CL, TN, OH, and SP contributed to scientific proof reading of the manuscript. MF conducted the animal experimentation. NC contributed to scientific proof reading of the manuscript and she is the study responsible. All authors contributed to the article and approved the submitted version. Funding This research was supported by the Independent Research Fund Denmark, grant no. 1335-00116. Conflict of interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Publisher’s note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. We thank Marie Rosenstand Hansen, Lykke Beinta Bjærge Bamdali, Britta Poulsen, and Susanne Nielsen for excellent assistance with the molecular lab work and Illumina sequencing. Kasper V. Poulsen, Trine Poulsen, Lene R. Dal, Winnie Ø. Thomsen, Lisbeth Märcher, Mette K. Christensen, Leif Tind, Nanna H. Jensen, and Lene Nørskov are thanked for excellent technical assistance. Supplementary material The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fmicb.2023.1018242/full#supplementary-material Click here for additional data file. ==== Refs References Alam M. T. Amos G. C. A. Murphy A. R. J. Murch S. Wellington E. M. H. Arasaradnam R. P. (2020). Microbial imbalance in inflammatory bowel disease patients at different taxonomic levels. Gut Pathog. 12 :1. doi: 10.1186/s13099-019-0341-6, PMID: 31911822 AOAC (1990). Official methods of analysis. AOAC, Washington, DC. Banaszkiewicz A. Kądzielska J. Gawrońska A. Pituch H. Obuch-Woszczatyński P. Albrecht P. . (2014). Enterotoxigenic Clostridium perfringens infection and pediatric patients with inflammatory bowel disease. J. Crohn's Colitis 8 , 276–281. doi: 10.1016/j.crohns.2013.08.018, PMID: 24060617 Bassaganya-Riera J. Hontecillas R. (2006). CLA and n-3 PUFA differentially modulate clinical activity and colonic PPAR-responsive gene expression in a pig model of experimental IBD. Clin. Nutr. 25 , 454–465. doi: 10.1016/j.clnu.2005.12.008, PMID: 16698153 Bates D. Mächler M. Bolker B. Walker S. (2015). Fitting linear mixed-effects models using lme4. J. Stat. Softw. 67 , 1–48. doi: 10.18637/jss.v067.i01 Benjamini Y. Hochberg Y. (1995). Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. R. Stat. Soc. Ser. B Methodol. 57 , 289–300. Bisanz J. E. (2018). qiime2R: Importing QIIME2 artifacts and associated data into R sessions. Bolyen E. Rideout J. R. Dillon M. R. Bokulich N. A. Abnet C. C. Al-Ghalith G. A. . (2019). Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat. Biotechnol. 37 , 852–857. doi: 10.1038/s41587-019-0209-9 31341288 Busck M. M. Lund M. B. Bird T. L. Bechsgaard J. S. Bilde T. Schramm A. (2022). Temporal and spatial microbiome dynamics across natural populations of the social spider Stegodyphus dumicola. FEMS Microbiol. Ecol. 98 :fiac015. doi: 10.1093/femsec/fiac015, PMID: 35147190 Callahan B. J. Mcmurdie P. J. Rosen M. J. Han A. W. Johnson A. J. A. Holmes S. P. (2016). DADA2: high-resolution sample inference from Illumina amplicon data. Nat. Methods 13 , 581–583. doi: 10.1038/nmeth.3869, PMID: 27214047 Canibe N. Højberg O. Badsberg J. H. Jensen B. B. (2007). Effect of feeding fermented liquid feed and fermented grain on gastrointestinal ecology and growth performance in piglets. J. Anim. Sci. 85 , 2959–2971. doi: 10.2527/jas.2006-744, PMID: 17591711 Chase-Topping M. E. Gunn G. Strachan W. D. Edwards S. A. Smith W. J. Hillman K. . (2007). Epidemiology of porcine non-specific colitis on Scottish farms. Vet. J. 173 , 353–360. doi: 10.1016/j.tvjl.2005.12.002, PMID: 16459113 Chen G. Ran X. Li B. Li Y. He D. Huang B. . (2018). Sodium butyrate inhibits inflammation and maintains epithelium barrier integrity in a TNBS-induced inflammatory bowel disease mice model. EBioMedicine 30 , 317–325. doi: 10.1016/j.ebiom.2018.03.030, PMID: 29627390 Chiba M. Nakane K. Komatsu M. (2019). Westernized diet is the Most ubiquitous environmental factor in inflammatory bowel disease. Perm. J. 23 , 18–107. doi: 10.7812/TPP/18-107, PMID: 30624192 Couto M. R. Gonçalves P. Magro F. Martel F. (2020). Microbiota-derived butyrate regulates intestinal inflammation: focus on inflammatory bowel disease. Pharmacol. Res. 159 :104947. doi: 10.1016/j.phrs.2020.104947, PMID: 32492488 Davis N. M. Proctor D. M. Holmes S. P. Relman D. A. Callahan B. J. (2018). Simple statistical identification and removal of contaminant sequences in marker-gene and metagenomics data. bioRxiv:221499. De Preter V. Geboes K. P. Bulteel V. Vandermeulen G. Suenaert P. Rutgeerts P. . (2011). Kinetics of butyrate metabolism in the normal colon and in ulcerative colitis: the effects of substrate concentration and carnitine on the β-oxidation pathway. Aliment. Pharmacol. Ther. 34 , 526–532. doi: 10.1111/j.1365-2036.2011.04757.x, PMID: 21707682 Eichele D. D. Kharbanda K. K. (2017). Dextran sodium sulfate colitis murine model: an indispensable tool for advancing our understanding of inflammatory bowel diseases pathogenesis. World J. Gastroenterol. 23 , 6016–6029. doi: 10.3748/wjg.v23.i33.6016, PMID: 28970718 Fernandes A. D. Reid J. N. S. Macklaim J. M. Mcmurrough T. A. Edgell D. R. Gloor G. B. (2014). Unifying the analysis of high-throughput sequencing datasets: characterizing RNA-seq, 16S rRNA gene sequencing and selective growth experiments by compositional data analysis. Microbiome 2 :15. doi: 10.1186/2049-2618-2-15, PMID: 24910773 Ferrer-Picón E. Dotti I. Corraliza A. M. Mayorgas A. Esteller M. Perales J. C. . (2020). Intestinal inflammation modulates the epithelial response to butyrate in patients with inflammatory bowel disease. Inflamm. Bowel Dis. 26 , 43–55. doi: 10.1093/ibd/izz119, PMID: 31211831 Galazzo G. Van Best N. Benedikter B. J. Janssen K. Bervoets L. Driessen C. . (2020). How to count our microbes? The effect of different quantitative microbiome profiling approaches. Front. Cell. Infect. Microbiol. 10 :403. doi: 10.3389/fcimb.2020.00403, PMID: 32850498 Galipeau H. J. Caminero A. Turpin W. Bermudez-Brito M. Santiago A. Libertucci J. . (2021). Novel fecal biomarkers that precede clinical diagnosis of ulcerative colitis. Gastroenterology 160 , 1532–1545. doi: 10.1053/j.gastro.2020.12.004, PMID: 33310084 Gasaly N. Hermoso M. A. Gotteland M. (2021). Butyrate and the fine-tuning of colonic homeostasis: implication for inflammatory bowel diseases. Int. J. Mol. Sci. 22 :3061. doi: 10.3390/ijms22063061, PMID: 33802759 Ge J. Han T. J. Liu J. Li J. S. Zhang X. H. Wang Y. . (2015). Meat intake and risk of inflammatory bowel disease: a meta-analysis. Turk J Gastroenterol 26 , 492–497. doi: 10.5152/tjg.2015.0106, PMID: 26575042 Gilbert M. S. Ijssennagger N. Kies A. K. Mil S. W. C. V. (2018). Protein fermentation in the gut; implications for intestinal dysfunction in humans, pigs, and poultry. Am. J. Physiol. Gastrointest. Liver Physiol. 315 , G159–G170. doi: 10.1152/ajpgi.00319.2017 29597354 Glei M. Klenow S. Sauer J. Wegewitz U. Richter K. Pool-Zobel B. L. (2006). Hemoglobin and hemin induce DNA damage in human colon tumor cells HT29 clone 19A and in primary human colonocytes. Mutat. Res. - Fundam. Mol. Mech. Mutagen. 594 , 162–171. doi: 10.1016/j.mrfmmm.2005.08.006, PMID: 16226281 Gobert A. P. Al-Greene N. T. Singh K. Coburn L. A. Sierra J. C. Verriere T. G. . (2018). Distinct immunomodulatory effects of spermine oxidase in colitis induced by epithelial injury or infection. Front. Immunol. 9 :1242. doi: 10.3389/fimmu.2018.01242, PMID: 29922289 Granlund A. V. B. Flatberg A. Østvik A. E. Drozdov I. Gustafsson B. I. Kidd M. . (2013). Whole genome gene expression meta-analysis of inflammatory bowel disease colon mucosa demonstrates lack of major differences between Crohn's disease and ulcerative colitis. PLoS One 8 :e56818. doi: 10.1371/journal.pone.0056818, PMID: 23468882 Guo X. Y. Liu X. J. Hao J. Y. (2020). Gut microbiota in ulcerative colitis: insights on pathogenesis and treatment. J. Dig. Dis. 21 , 147–159. doi: 10.1111/1751-2980.12849 32040250 Hansen B. (1989). Determination of nitrogen as elementary N, an alternative to Kjeldahl. Acta Agric. Scand. 39 , 113–118. doi: 10.1080/00015128909438504 Hansen R. Thomson J. M. Fox J. G. El-Omar E. M. Hold G. L. (2011). Could helicobacter organisms cause inflammatory bowel disease? FEMS Immunol. Med. Microbiol. 61 , 1–14. doi: 10.1111/j.1574-695X.2010.00744.x 20955468 Herlemann D. P. R. Labrenz M. Jürgens K. Bertilsson S. Waniek J. J. Andersson A. F. (2011). Transitions in bacterial communities along the 2000 km salinity gradient of the Baltic Sea. ISME J. 5 , 1571–1579. doi: 10.1038/ismej.2011.41, PMID: 21472016 Hermann-Bank M. L. Skovgaard K. Stockmarr A. Strube M. L. Larsen N. Kongsted H. . (2015). Characterization of the bacterial gut microbiota of piglets suffering from new neonatal porcine diarrhoea. BMC Vet. Res. 11 :139. doi: 10.1186/s12917-015-0419-4, PMID: 26099928 Holman D. B. Brunelle B. W. Trachsel J. Allen H. K. (2017). Meta-analysis to define a Core microbiota in the swine gut. mSystems 2 , e00004–e00017. doi: 10.1128/mSystems.00004-17 28567446 Holmes S. McMurdie P. J. (2013). Phyloseq: an R package for reproducible interactive analysis and graphics of microbiome census data. PLoS One 8 :e61217. doi: 10.1371/journal.pone.0061217 23630581 Illumina I. (2013). 16S metagenomic sequencing library preparation. Preparing 16S Ribosomal RNA Gene Amplicons for the Illumina MiSeq System, 1–28. Jantchou P. Morois S. Clavel-Chapelon F. Boutron-Ruault M.-C. Carbonnel F. (2010). Animal protein intake and risk of inflammatory bowel disease: the E3N prospective study. J. Am. College Gastroenterol. 105 , 2195–2201. doi: 10.1038/ajg.2010.192 Jensen M. T. Cox R. P. Jensen B. B. (1995). Microbial production of skatole in the hind gut of pigs given different diets and its relation to skatole deposition in backfat. Anim. Sci. 61 , 293–304. doi: 10.1017/S1357729800013837 Jian C. Luukkonen P. Yki-Järvinen H. Salonen A. Korpela K. (2020). Quantitative PCR provides a simple and accessible method for quantitative microbiota profiling. PLoS One 15 :e0227285. doi: 10.1371/journal.pone.0227285, PMID: 31940382 Kembel S. W. Cowan P. D. Helmus M. R. Cornwell W. K. Morlon H. Ackerly D. D. . (2010). Picante: R tools for integrating phylogenies and ecology. Bioinformatics 26 , 1463–1464. doi: 10.1093/bioinformatics/btq166, PMID: 20395285 Knudsen L. A. Desdorf R. Moller S. Sorensen S. B. Hansen A. K. Andersen V. (2020). Translational potential of metabolomics on animal models of inflammatory bowel disease-a systematic critical review. Int. J. Mol. Sci. 21 :3856. doi: 10.3390/ijms21113856, PMID: 32485793 Kolde R. Maintainer R. K. (2018). “Package ‘pheatmap’.” R package 1.10. Kraatz M. Wallace R. J. Svensson L. (2011). Olsenella umbonata sp. nov., a microaerotolerant anaerobic lactic acid bacterium from the sheep rumen and pig jejunum, and emended descriptions of Olsenella, Olsenella uli and Olsenella profusa. Int. J. Syst. Evol. Microbiol. 61 , 795–803. doi: 10.1099/ijs.0.022954-0 20435744 Le Leu R. K. Young G. P. Hu Y. Winter J. Conlon M. A. (2013). Dietary red meat aggravates dextran sulfate sodium-induced colitis in mice whereas resistant starch attenuates inflammation. Dig. Dis. Sci. 58 , 3475–3482. doi: 10.1007/s10620-013-2844-1, PMID: 23990000 Lenth R. Singmann H. Love J. Buerkner P. Herve M. (2018). Emmeans: estimated marginal means, aka least-squares means. R package version, 1, 3. Li D.-P. Cui M. Tan F. Liu X.-Y. Yao P. (2021). High red meat intake exacerbates dextran sulfate-induced colitis by altering gut microbiota in mice. Front. Nutr. 8 :646819. doi: 10.3389/fnut.2021.646819, PMID: 34355008 Liu C. M. Hungate B. A. Tobian A. A. R. Ravel J. Prodger J. L. Serwadda D. . (2015). Penile microbiota and female partner bacterial vaginosis in Rakai, Uganda. MBio 6 :e00589. doi: 10.1128/mBio.00589-15 26081632 Liu F. Liu J. Wang T. T. Y. Liu Z. Xue C. Mao X. . (2020). Molecular and microbial signatures predictive of prebiotic action of neoagarotetraose in a dextran sulfate sodium-induced murine colitis model. Microorganisms 8 :995. doi: 10.3390/microorganisms8070995, PMID: 32635315 Liu Y. Wang X. Hou Y. Yin Y. Qiu Y. Wu G. . (2017). Roles of amino acids in preventing and treating intestinal diseases: recent studies with pig models. Amino Acids 49 , 1277–1291. doi: 10.1007/s00726-017-2450-1, PMID: 28616751 Loss R. W. Mangla J. C. Pereira M. (1980). Campylobacter colitis presenting as inflammatory bowel disease with segmentai colonic ulcerations. Gastroenterology 79 , 138–140. doi: 10.1016/0016-5085(80)90088-8, PMID: 7380209 Louis P. Scott K. P. Duncan S. H. Flint H. J. (2007). Understanding the effects of diet on bacterial metabolism in the large intestine. J. Appl. Microbiol. 102 , 1197–1208. doi: 10.1111/j.1365-2672.2007.03322.x 17448155 Love M. I. Huber W. Anders S. (2014). Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15 :550. doi: 10.1186/s13059-014-0550-8, PMID: 25516281 Ma L. Ni Y. Wang Z. Tu W. Ni L. Zhuge F. . (2020). Spermidine improves gut barrier integrity and gut microbiota function in diet-induced obese mice. Gut Microbes 12 , 1832857–1832819. doi: 10.1080/19490976.2020.1832857, PMID: 33151120 Manichanh C. Borruel N. Casellas F. Guarner F. (2012). The gut microbiota in IBD. Nat. Rev. Gastroenterol. Hepatol. 9 , 599–608. doi: 10.1038/nrgastro.2012.152 22907164 Maráková K. Piešťanský J. Zelinková Z. Mikuš P. (2020). Simultaneous determination of twelve biogenic amines in human urine as potential biomarkers of inflammatory bowel diseases by capillary electrophoresis – tandem mass spectrometry. J. Pharm. Biomed. Anal. 186 :113294. doi: 10.1016/j.jpba.2020.113294, PMID: 32348953 Marchesi J. R. Holmes E. Khan F. Kochhar S. Scanlan P. Shanahan F. . (2007). Rapid and noninvasive metabonomic characterization of inflammatory bowel disease. J. Proteome Res. 6 , 546–551. doi: 10.1021/pr060470d, PMID: 17269711 Maslowski K. M. Vieira A. T. Ng A. Kranich J. Sierro F. Di Y. . (2009). Regulation of inflammatory responses by gut microbiota and chemoattractant receptor GPR43. Nature 461 , 1282–1286. doi: 10.1038/nature08530, PMID: 19865172 Mateos-Rivera A. Øvreås L. Wilson B. Yde J. C. Finster K. W. (2018). Activity and diversity of methane-oxidizing bacteria along a Norwegian sub-Arctic glacier forefield. FEMS Microbiol. Ecol. 94 :fiy059. doi: 10.1093/femsec/fiy059 Mcgovern D. P. B. Kugathasan S. Cho J. H. (2015). Genetics of inflammatory bowel diseases. Gastroenterology 149 , 1163–1176.e2. doi: 10.1053/j.gastro.2015.08.001, PMID: 26255561 Mirarab S. Nguyen N. Warnow T . (2011). SEPP: SATé-enabled phylogenetic placement. Biocomputing 2012. WORLD SCIENTIFIC. Mizoguchi E. Low D. Ezaki Y. Okada T. (2020). Recent updates on the basic mechanisms and pathogenesis of inflammatory bowel diseases in experimental animal models. Int. Res. 18 , 151–167. doi: 10.5217/ir.2019.09154, PMID: 32326669 Mizoguchi E. Subramaniam R. Okada T. Mizoguchi A. (2021). A review of selected IBD biomarkers: from animal models to bedside. Diagnostics 11 :207. doi: 10.3390/diagnostics11020207, PMID: 33573291 Morton J. T. Sanders J. Quinn R. A. Mcdonald D. Gonzalez A. Vázquez-Baeza Y. . (2017). Balance trees reveal microbial niche differentiation. mSystems 2 , e00162–e00116. doi: 10.1128/mSystems.00162-16 28144630 Ng K. M. Ferreyra J. A. Higginbottom S. K. Lynch J. B. Kashyap P. C. Gopinath S. . (2013). Microbiota-liberated host sugars facilitate post-antibiotic expansion of enteric pathogens. Nature 502 , 96–99. doi: 10.1038/nature12503, PMID: 23995682 Ni J. Wu G. D. Albenberg L. Tomov V. T. (2017). Gut microbiota and IBD: causation or correlation? Nat. Rev. Gastroenterol. Hepatol. 14 , 573–584. doi: 10.1038/nrgastro.2017.88, PMID: 28743984 Nielsen T. S. Fredborg M. Theil P. K. Yue Y. Bruhn L. V. Andersen V. . (2020). Dietary red meat adversely affects disease severity in a pig model of DSS-induced colitis despite reduction in colonic pro-inflammatory gene expression. Nutrients 12 :1728. doi: 10.3390/nu12061728, PMID: 32526985 Nomura K. Ishikawa D. Okahara K. Ito S. Haga K. Takahashi M. . (2021). Bacteroidetes species are correlated with disease activity in ulcerative colitis. J. Clin. Med. 10 :1749. doi: 10.3390/jcm10081749, PMID: 33920646 Oksanen J. Blanchet F. G. Friendly M. Kindt Roeland Legendre P. Mcglinn D. . (2020). Vegan: community ecology package. R package version 2.5-7. Ott S. J. Musfeldt M. Wenderoth D. F. Hampe J. Brant O. Fölsch U. R. . (2004). Reduction in diversity of the colonic mucosa associated bacterial microflora in patients with active inflammatory bowel disease. Gut 53 , 685–693. doi: 10.1136/gut.2003.025403, PMID: 15082587 Panah M. F. (2022). Gut microbiota 16S rRNA gene data analysis and bioinformatics. GitHub repository. Panah F. M. Lauridsen C. Højberg O. Nielsen T. S. (2021). Etiology of colitis-complex diarrhea in growing pigs: a review. Animals 11 :2151. doi: 10.3390/ani11072151, PMID: 34359279 Pistol G. C. Marin D. E. Rotar M. C. Ropota M. Taranu I. (2020). Bioactive compounds from dietary whole grape seed meal improved colonic inflammation via inhibition of MAPKs and NF-kB signaling in pigs with DSS induced colitis. J. Funct. Foods 66 :103708. doi: 10.1016/j.jff.2019.103708 Pituch-Zdanowska A. Banaszkiewicz A. Albrecht P. (2015). The role of dietary fibre in inflammatory bowel disease. Prz. Gastroenterol. 10 , 135–141. doi: 10.5114/pg.2015.52753, PMID: 26516378 Portrait V. Cottenceau G. Pons A. M. (2000). A Fusobacterium mortiferum strain produces a bacteriocin-like substance(s) inhibiting Salmonella enteritidis. Lett. Appl. Microbiol. 31 , 115–117. doi: 10.1046/j.1365-2672.2000.00774.x 10972711 Poulsen A.-S. R. Jonge N. D. Nielsen J. L. Højberg O. Lauridsen C. Cutting S. M. . (2018). Impact of Bacillus spp. spores and gentamicin on the gastrointestinal microbiota of suckling and newly weaned piglets. PLoS One 13 :e0207382. doi: 10.1371/journal.pone.0207382, PMID: 30481191 Props R. Kerckhof F.-M. Rubbens P. De Vrieze J. Hernandez Sanabria E. Waegeman W. . (2017). Absolute quantification of microbial taxon abundances. ISME J. 11 , 584–587. doi: 10.1038/ismej.2016.117, PMID: 27612291 Putignani L. Oliva S. Isoldi S. Del Chierico F. Carissimi C. Laudadio I. . (2021). Fecal and mucosal microbiota profiling in pediatric inflammatory bowel diseases. Eur. J. Gastroenterol. Hepatol. 33 , 1376–1386. doi: 10.1097/MEG.0000000000002050, PMID: 33470709 Quast C. Pruesse E. Yilmaz P. Gerken J. Schweer T. Yarza P. . (2012). The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res. 41 , D590–D596. doi: 10.1093/nar/gks1219, PMID: 23193283 Rajilić-Stojanović M. Shanahan F. Guarner F. De Vos W. M. (2013). Phylogenetic analysis of dysbiosis in ulcerative colitis during remission. Inflamm. Bowel Dis. 19 , 481–488. doi: 10.1097/MIB.0b013e31827fec6d, PMID: 23385241 Rashvand S. Somi M. H. Rashidkhani B. Hekmatdoost A. (2015). Dietary protein intakes and risk of ulcerative colitis. Med. J. Islam Repub. Iran 29 :253.26793644 Rechkemmer G. Rönnau K. Von Engelhardt W. (1988). Fermentation of polysaccharides and absorption of short chain fatty acids in the mammalian hindgut. Comp. Biochem. Physiol. A Comp. Physiol. 90 , 563–568. doi: 10.1016/0300-9629(88)90668-8, PMID: 2902962 Robeson M. S. O’rourke D. R. Kaehler B. D. Ziemski M. Dillon M. R. Foster J. T. . (2020). RESCRIPt: reproducible sequence taxonomy reference database management for the masses. bioRxiv 2020.10.05.326504 Rooks M. G. Garrett W. S. (2016). Gut microbiota, metabolites and host immunity. Nat. Rev. Immunol. 16 , 341–352. doi: 10.1038/nri.2016.42, PMID: 27231050 Roselli M. Finamore A. (2020). Use of synbiotics for ulcerative colitis treatment. Curr. Clin. Pharmacol. 15 , 174–182. doi: 10.2174/1574884715666191226120322, PMID: 31878863 Sanchez H. N. Moroney J. B. Gan H. Shen T. Im J. L. Li T. . (2020). B cell-intrinsic epigenetic modulation of antibody responses by dietary fiber-derived short-chain fatty acids. Nat. Commun. 11 :60. doi: 10.1038/s41467-019-13603-6, PMID: 31896754 Scanlan P. D. Shanahan F. Omahony C. Marchesi J. R. (2006). Culture-independent analyses of temporal variation of the dominant fecal microbiota and targeted bacterial subgroups in Crohn's disease. J. Clin. Microbiol. 44 , 3980–3988. doi: 10.1128/JCM.00312-06, PMID: 16988018 Schirmer M. Garner A. Vlamakis H. Xavier R. J. (2019). Microbial genes and pathways in inflammatory bowel disease. Nat. Rev. Microbiol. 17 , 497–511. doi: 10.1038/s41579-019-0213-6, PMID: 31249397 Schulthess J. Pandey S. Capitani M. Rue-Albrecht K. C. Arnold I. Franchini F. . (2019). The short chain fatty acid butyrate imprints an antimicrobial program in macrophages. Immunity 50 , 432–445.e7. doi: 10.1016/j.immuni.2018.12.018, PMID: 30683619 Shiba T. Aiba Y. Ishikawa H. Ushiyama A. Takagi A. Mine T. . (2003). The suppressive effect of Bifidobacteria on Bacteroides vulgatus, a putative pathogenic microbe in inflammatory bowel disease. Microbiol. Immunol. 47 , 371–378. doi: 10.1111/j.1348-0421.2003.tb03368.x, PMID: 12906096 Silva P. B. J. Navegantes-Lima C. K. Oliveira L. B. A. Rodrigues V. S. D. Gaspar L. F. S. Monteiro V. S. V. . (2018). Protective mechanisms of butyrate on inflammatory bowel disease. Curr. Pharm. Des. 24 , 4154–4166. doi: 10.2174/1381612824666181001153605, PMID: 30277149 Simpson G. L. (2022). Permute: Functions for generating restricted permutations of data. R package version 0.7-0. Available at: http://CRAN.R-project.org/package=permute (Accessed 27, 2022). Smedley J. G. III Saputo J. Parker J. C. Fernandez-Miyakawa M. E. Robertson S. L. Mcclane B. A. . (2008). Noncytotoxic Clostridium perfringens enterotoxin (CPE) variants localize CPE intestinal binding and demonstrate a relationship between CPE-induced cytotoxicity and enterotoxicity. Infect. Immun. 76 , 3793–3800. doi: 10.1128/IAI.00460-08, PMID: 18505809 Stämmler F. Gläsner J. Hiergeist A. Holler E. Weber D. Oefner P. J. . (2016). Adjusting microbiome profiles for differences in microbial load by spike-in bacteria. Microbiome 4 :28. doi: 10.1186/s40168-016-0175-0, PMID: 27329048 Stoldt W. (1952). Vorschlag zur vereinheitlichung der fettbestimmung in lebensmitteln. Fette und Seifen 54 , 206–207. doi: 10.1002/lipi.19520540406 Tawakoli P. N. Neu T. R. Busck M. M. Kuhlicke U. Schramm A. Attin T. . (2017). Visualizing the dental biofilm matrix by means of fluorescence lectin-binding analysis. J. Oral Microbiol. 9 :1345581. doi: 10.1080/20002297.2017.1345581, PMID: 28748044 Team R. C (2021). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Thøgersen R. Bertram H. C. (2021). Reformulation of processed meat to attenuate potential harmful effects in the gastrointestinal tract – a review of current knowledge and evidence of health prospects. Trends Food Sci. Technol. 108 , 111–118. doi: 10.1016/j.tifs.2020.12.015 Thomson J. (2009). Feed-associated colitis of growing pigs and its interaction with enteric infections. Acta Sci. Vet. 37 , s1–s9. Topping D. L. Clifton P. M. (2001). Short-chain fatty acids and human colonic function: roles of resistant starch and nonstarch polysaccharides. Physiol. Rev. 81 , 1031–1064. doi: 10.1152/physrev.2001.81.3.1031, PMID: 11427691 Van Hecke T. Vossen E. Goethals S. Boon N. De Vrieze J. De Smet S. (2021). In vitro and in vivo digestion of red cured cooked meat: oxidation, intestinal microbiota and fecal metabolites. Food Res. Int. 142 :110203. doi: 10.1016/j.foodres.2021.110203, PMID: 33773678 Vieira E. L. M. Leonel A. J. Sad A. P. Beltrão N. R. M. Costa T. F. Ferreira T. M. R. . (2012). Oral administration of sodium butyrate attenuates inflammation and mucosal lesion in experimental acute ulcerative colitis. J. Nutr. Biochem. 23 , 430–436. doi: 10.1016/j.jnutbio.2011.01.007, PMID: 21658926 Vinolo M. A. Rodrigues H. G. Nachbar R. T. Curi R. (2011). Regulation of inflammation by short chain fatty acids. Nutrients 3 , 858–876. doi: 10.3390/nu3100858, PMID: 22254083 Wang R. Yu H. Fang H. Jin Y. Zhao Y. Shen J. . (2020). Effects of dietary grape pomace on the intestinal microbiota and growth performance of weaned piglets. Arch. Anim. Nutr. 74 , 296–308. doi: 10.1080/1745039X.2020.1743607, PMID: 32308036 Wei S. Bahl M. I. Baunwall S. M. D. Hvas C. L. Licht T. R. (2021). Determining gut microbial dysbiosis: a review of applied indexes for assessment of intestinal microbiota imbalances. Appl. Environ. Microbiol. 87 , e00395–e00321. doi: 10.1128/AEM.00395-21 33741632 Windey K. De Preter V. Verbeke K. (2012). Relevance of protein fermentation to gut health. Mol. Nutr. Food Res. 56 , 184–196. doi: 10.1002/mnfr.201100542 22121108 Wu N. Mah C. Koentgen S. Zhang L. Grimm M. C. El-Omar E. . (2021). Inflammatory bowel disease and the gut microbiota. Proc. Nutr. Soc. 80 , 424–434. doi: 10.1017/S002966512100197X Yang Q. Wang Y. Jia A. Wang Y. Bi Y. Liu G. (2021). The crosstalk between gut bacteria and host immunity in intestinal inflammation. J. Cell. Physiol. 236 , 2239–2254. doi: 10.1002/jcp.30024 32853458 Zakerska-Banaszak O. Tomczak H. Gabryel M. Baturo A. Wolko L. Michalak M. . (2021). Dysbiosis of gut microbiota in polish patients with ulcerative colitis: a pilot study. Sci. Rep. 11 :2166. doi: 10.1038/s41598-021-81628-3, PMID: 33495479 Zheng D. Liwinski T. Elinav E. (2020). Interaction between microbiota and immunity in health and disease. Cell Res. 30 , 492–506. doi: 10.1038/s41422-020-0332-7, PMID: 32433595
PMC010xxxxxx/PMC10150333.txt
==== Front Front Immunol Front Immunol Front. Immunol. Frontiers in Immunology 1664-3224 Frontiers Media S.A. 37138862 10.3389/fimmu.2023.1134412 Immunology Original Research Identification of immune-associated genes in diagnosing osteoarthritis with metabolic syndrome by integrated bioinformatics analysis and machine learning Li Junchen 1 † Wang Genghong 1 † Xv Xilin 2 3 * Li Zhigang 4 Shen Yiwei 1 Zhang Cheng 1 Zhang Xiaofeng 3 5 * 1 The Graduate School, Heilongjiang University of Chinese Medicine, Harbin, China 2 The Third Affiliated Hospital of Heilongjiang University of Chinese Medicine, Harbin, China 3 Teaching and Research Section of Orthopedics and Traumatology, Heilongjiang University of Chinese Medicine, Harbin, China 4 The Second Department of Orthopedics and Traumatology, The Second Affiliated Hospital of Heilongjiang University of Chinese Medicine, Harbin, China 5 The Bone Injury Teaching Laboratory, Heilongjiang University of Chinese Medicine, Harbin, China Edited by: Michael V. Volin, Midwestern University, United States Reviewed by: Ming Lu, Anhui Medical University, China; Maria Andersson, Lund University, Sweden *Correspondence: Xilin Xv, [email protected]; Xiaofeng Zhang, [email protected] †These authors have contributed equally to this work and share first authorship This article was submitted to Inflammation, a section of the journal Frontiers in Immunology 17 4 2023 2023 14 113441230 12 2022 27 3 2023 Copyright © 2023 Li, Wang, Xv, Li, Shen, Zhang and Zhang 2023 Li, Wang, Xv, Li, Shen, Zhang and Zhang https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Background In the pathogenesis of osteoarthritis (OA) and metabolic syndrome (MetS), the immune system plays a particularly important role. The purpose of this study was to find key diagnostic candidate genes in OA patients who also had metabolic syndrome. Methods We searched the Gene Expression Omnibus (GEO) database for three OA and one MetS dataset. Limma, weighted gene co-expression network analysis (WGCNA), and machine learning algorithms were used to identify and analyze the immune genes associated with OA and MetS. They were evaluated using nomograms and receiver operating characteristic (ROC) curves, and finally, immune cells dysregulated in OA were investigated using immune infiltration analysis. Results After Limma analysis, the integrated OA dataset yielded 2263 DEGs, and the MetS dataset yielded the most relevant module containing 691 genes after WGCNA, with a total of 82 intersections between the two. The immune-related genes were mostly enriched in the enrichment analysis, and the immune infiltration analysis revealed an imbalance in multiple immune cells. Further machine learning screening yielded eight core genes that were evaluated by nomogram and diagnostic value and found to have a high diagnostic value (area under the curve from 0.82 to 0.96). Conclusion Eight immune-related core genes were identified (FZD7, IRAK3, KDELR3, PHC2, RHOB, RNF170, SOX13, and ZKSCAN4), and a nomogram for the diagnosis of OA and MetS was established. This research could lead to the identification of potential peripheral blood diagnostic candidate genes for MetS patients who also suffer from OA. differentially expressed genes osteoarthritis metabolic syndrome machine learning immune infiltration The State Administration of Traditional Chinese Medicine Construction Project of Inheritance Studio of National Famous Traditional Chinese Medicine Experts (No. 2021-2024). Youth Qihuang Scholar Support Project of the State Administration of Troditional Chinese Medicine (No. 2021-2023). Key Laboratory of Basic and Clinical Research of Osteonecrosis in Heilongjiang Province. ==== Body pmc1 Introduction Osteoarthritis (OA) is one of the common degenerative diseases of the musculoskeletal system in orthopedics, often involving one or more joints, and has a high prevalence worldwide (1), and according to WHO, there is a 10% chance of OA occurring in people over 60 years of age worldwide (2). So far, the causative factors and pathogenesis of OA are not clear, but most scholars believe that its pathogenesis is a combination of several factors, such as obesity, aging, trauma, excessive joint strain, metabolic disorders, inflammation, and genetics (3). Metabolic syndrome (MetS) is a metabolic disorder characterized by abdominal obesity, dyslipidemia, hyperglycemia, and hypertension (4). According to the National Institute for Health and Nutrition Examination Survey, around one in five adults in the United States suffer from MetS (5). Most patients with metabolic syndrome are associated with varying degrees of obesity, which increases stress on the entire musculoskeletal system and poses a higher risk for the development of osteoarthritis and various other musculoskeletal disorders. Osteoarthritis and metabolic syndrome share common mechanisms of inflammation, oxidative stress, and metabolic dysfunction in their etiology (6). Traditionally, osteoarthritis is a non-inflammatory disease affected by trauma or metabolic dysregulation, and age-related joint degeneration is thought to be a causal factor in the development of the disease (7, 8). However, a growing body of evidence suggests that low-grade inflammation may be a key factor driving the pathogenesis of OA (9). Inflammation and metabolic disorders play a very important role in the progression of osteoarthritis (10, 11). Risk factors, including diabetes, hypertension, and hyperlipidemia, are largely involved in osteoarthritis through the release of inflammatory and adipokines that accelerate the progression of osteoarthritis by driving articular cartilage degeneration and bone marrow lesions (12). Several theories describe how MetS risk factors affect the progression of OA, such as high blood pressure can cause subchondral ischemia, abnormal lipids can cause lipid deposition in chondrocytes, and high blood glucose can cause oxidative stress and low inflammation, eventually leading to cartilage destruction (13). In addition, central obese patients also have abnormal leptin and lipocalin levels, which can further aggravate the development of OA (6, 14). To date, metabolic syndrome-associated osteoarthritis (MetS-OA) has been well characterized as a distinct phenotype of osteoarthritis (15, 16), and the goal of this model is to investigate the relationship between inflammatory response and metabolic disorders in order to improve the concept of treating MetS and to aid in the reduction of inflammatory response in OA patients. There is some evidence to support this, but the mechanism of action in metabolic syndrome-related osteoarthritis is still being investigated. In recent years, comprehensive bioinformatics analysis has been used to identify novel genes associated with various diseases that can serve as diagnostic and prognostic biomarkers. However, the common diagnostic and interlinked genes of OA and MetS are not known. Therefore, this study used a bioinformatics approach to screen for biomarkers associated with immune infiltration in both, which could help identify immune-related potential diagnostic markers for OA in patients with MetS. 2 Materials and methods 2.1 Microarray data The datasets used were all from the NCBI Gene Expression Omnibus (GEO) database (17), and the OA datasets were GSE169077, GSE55457, and GSE55235 (18); while the MetS dataset was GSE98895 (19). Figure 1 depicts the study flowchart. Figure 1 Study flowchart, GSE, gene expression omnibus series; WGCNA, weighted gene co-expression network analysis; Limma, linear models for microarray data; DEGs, differentially expressed genes. 2.2 Data processing and differential gene screening The following steps were all carried out using the R software (https://www.bioconductor.org/), starting with “affy” to calibrate, log-transform, and normalize the three OA datasets, then merging them and using “SVA” to remove batch effects, with p-values set to 0.05 and |log2 fold change (FC)|>1.5, and “limma” for differential gene screening. Table 1 presents detailed dataset information, including the microarray platform, sample groups, and numbers. Table 1 Basic information of GEO datasets used in the study. GSE series Type Sample size Platform Control Osteoarthritis GSE169077 RNA 5 6 GPL96 GSE55457 RNA 10 10 GPL96 GSE55235 RNA 10 10 GPL96 Control Metabolic syndrome GSE98895 RNA 20 20 GPL6947 2.3 Weighted gene co-expression network analysis WGCNA (20) was used to investigate the gene modules most closely linked to MetS. First, the top 50% of genes with the highest median absolute deviation (MAD) were filtered. The expression matrix was then filtered to remove ineligible data. Third, use a “soft” threshold power (β) for co-expression of similarity to calculate adjacency. Then, using dynamic tree cuts and hierarchical clustering, a topological overlap matrix (TOM) was created to group genes into modules by random colors, and a gene dendrogram was constructed using a TOM-based measure of phase dissimilarity and a minimum gene cluster size (n=100). Fifth, for the next step of the study, the dissimilarity of the module genes was calculated, and the average linkage hierarchy clustering was performed. It is finally visualized. 2.4 Functional enrichment analysis Sangerbox (http://www.sangerbox.com/tool) was used for Gene Ontology (GO) (21) and Kyoto Encyclopedia of Genes and Genomes (KEGG) (22) enrichment analysis. 2.5 Machine learning Machine learning algorithms are used to screen the core genes for OA diagnosis. “glmnet” (23) was used for LASSO (24) regression, and “randomForest” (25) was used for RF (26) analysis. The two intersecting genes were used as the core genes for OA diagnosis. 2.6 Nomogram construction and evaluation of recipient operating characteristics Nomogram was created for the screened genes using the “rms” (27) R package, and its value in the clinic was determined by the area under the curve (AUC) and 95% CI using ROC. When the AUC exceeds 0.7, it is considered to have diagnostic value. 2.7 Immune infiltration analysis The CIBERSORT algorithm is used to determine the proportion of immune cells in cells or tissues (28). The bar graphs show the proportion of each type of immune cell in various samples, and the “corrplot” (29) R package is used to generate a heat map of the correlation between 22 immune cells. The vioplot was used to visualize the differences between the OA and normal immune cell groups. 2.8 Statistical analysis R software version 4.2.2 and Sangerbox were used for all statistical analyses. To compare normally distributed continuous variables between two groups, the Independent Student’s t-test was used, and the Wilcoxon rank sum test was used to analyze non-normally distributed variables. Every statistical test was two-sided. The statistical significance level was set at p-Value <0.05. 3 Results 3.1 Differentially expressed genes The integrated OA dataset yielded a total of 2263 DEGs after LIMMA analysis, with 1341 up-regulated genes and 922 down-regulated genes. Figures 2A, B show the heat map and volcano map generated from the above data. Furthermore, the MetS dataset yielded 1449 DEGs, including 605 up-regulated genes and 844 down-regulated genes ( Figures 3A, B ). Figure 2 Heatmap and valcano plot for the DEGs identified from the integrated OA dataset. (A) Each row shows the DEGs, and each column refers to one of the samples of OA cases or controls. The red and blue represent DEGs with upregulated and downregulated gene expression, respectively. (B) Red and green plot triangles represent DEGs with upregulated and downregulated gene expression, respectively. OA, Osteoarthritis: DEGs, differentially expressed genes. Figure 3 Identification of DEGs via Limma and module genes via WGCNA in MetS. (A) The heatmap displays the top 50 upregulated and downregulated DECs identified from MetS dataset. Each row represents the intersection of genes, and each column represents one of MetS cases or controls. Red and blue represent upregulated and downregulated ge expression. (B) The volcano plot shows all DEGs, of which red and green triangles refer to significant DEGs. (C, D) b= 14 is selected as the soft threshold with the combined analysis of scale independence and average connectivity. (E) Clustering dendrogram of the MetS and control samples. (F) Gene co-expression modules represented by different colors under the gene tree (G) Heatmap of eigengene adjacency. (H) Heatmap of the association between modules and Mers. The yellow module is shown to be correlated significantly with MetS. Numbers at the top and bottom brackets represent the correlation coefficient and p-value, respectively. (I) Correlation plot between module membership and gene significance of genes included in the yellow module WGCNA, weighted gene co-expression network analysis, Limma, linear models for microarray data; DEGs, differentially expressed genes. 3.2 Selection of key modules In MetS, use WGCNA to filter critical modules. When the soft threshold =14, the scale-free network performs best ( Figures 3C, D ). Figure 3E depicts the clustering tree graph for both groups and the ten randomly colored gene modules obtained ( Figures 3F, G ). The yellow module (691 genes) had the highest MetS correlation ( Figure 3H ) and can be used as a key module in the following analysis. Figure 3I shows the results of a correlation analysis for the genes in the yellow module, which showed a positive correlation (r=0.62). 3.3 Functional enrichment analysis of metabolic syndrome The intersection of 1449 DEGs and 691 modular genes yielded a total of 108 genes ( Figure 4A ). The “PI3K-Akt signaling pathway” and “glycerophospholipid metabolism” were primarily enriched in KEGG analysis ( Figure 4B ). According to GO analysis, the main enrichment in biological process (BP) terms was in “B cell activation involved in immune response” and “cellular response to DNA damage stimuli” ( Figure 4C ). It is primarily found in “cell membrane,” “nuclear lumen,” and “vesicles” in terms of cellular component (CC) ontology ( Figure 4D ). It was primarily enriched in “molecular function regulators,” “enzyme regulator activity,” and “signaling receptor binding” in molecular function (MF) analysis ( Figure 4E ). Figure 4 Enrichment analysis of the intersection of genes in MetS. (A) Venn diagram shows that 108 genes are identified from the intersection of DEGs via Limma and yellow module genes via WGCNA. (B) KEGG pathway analysis of the intersection of genes. Different colors represent various significant pathways and related enriched genes. (C–E) GO analysis of the Intersection of genes, including biological process, cellular component, and molecular function, respectively. The y-axis represents different GO terms, the x-axis represents gene ratio enriched in relative GO terms, the circle size refers to gene numbers, and the color represents p-value. MetS, Metabolic syndrome: KEGG, Kyoto Encyclopedia of Genes and Genomes GO, Gene Ontology: WGCNA, weighted gene co-expression network analysis: Limma, linear models for microarray datac DEGs, differentially expressed genes. 3.4 Enrichment analysis of osteoarthritis with metabolic syndrome To further explore whether key MetS-related genes might be associated with the pathogenesis of OA, 82 genes were identified by the Venn diagram from the intersection of OA DEGs and MetS key module genes ( Figure 5A ). According to KEGG analysis, 82 genes were primarily enriched in “inflammatory mediator regulation of TRP channels,” “cellular senescence,” and “MAPK signaling pathway,” all of which are closely related to the immune system ( Figure 5B ). GO analysis revealed that they were primarily enriched in “immune response,” “cell death,” and “immune system process” (BP); “integral component of endoplasmic reticulum membrane,” “stereocilium mem-brane” and “immunological synapse,” (CC); and “catalytic activity,” “ATP binding,” and “sequence-specific DNA binding” (MF) ( Figures 5C–E ). Figure 5 Enrichment analysis of common genes from OA with MetS. (A) Venn diagram shows that 82 common genes are identified from the intersection of genes in OA using Limma and MetS using WGCNA. (B) KEGG analysis of 82 common genes. (C–E) GO analysis (hiological process, cellular component, and molecular function) of 82 common genes. OA. Osteoarthritis: MetS, Metabolic syndrome, WGCNA, weighted gene co-expression network analysis. 3.5 Core genes screening using machine learning The Lasso regression and RF algorithms were used to identify core genes and create relevant nomograms for ROC analysis. Lasso regression screened 10 candidate genes ( Figures 6A, B ), and the RF algorithm identified 30 most important genes ( Figures 6C, D ), and the two were taken to intersect ( Figure 6E ), resulting in the identification of eight genes (FZD7, IRAK3, KDELR3, PHC2, RHOB, RNF170, SOX13, and ZKSCAN4). Figure 6 Machine learning in screening candidate diagnostic biomarkers for OA with MetS (A, B) Biomarkers screening in the Lasso model. The number of genes (n-10) corresponding to the lowest point of the curve is the most suitable for OA with MetS diamos. (C, D) The random forest algorithm shows the error in OA; control group and genes are ranked based on the importance score. (E) Venn diagram shows that eight candidate diagnostic genes are identified via the above two algorithms OA. Osteoarthritis: MetS Metabolic syndrome. 3.6 Determining diagnostic value We created a nomogram ( Figure 7A ) and plotted ROC curves based on the eight candidate genes to assess the diagnostic value of each gene. The calculated AUCs and 95% confidence intervals were as follows: FZD7 (AUC 0.86, CI 0.96–0.76), IRAK3 (AUC 0.92, CI 0.99–0.84), KDELR3 (AUC 0.94, CI 1.00–0.87), PHC2 (AUC 0.89, CI 0.99–0.80), RHOB (AUC 0.96, CI 1.00–0.90), RNF170 (AUC 0.82, CI 0.94–0.70), SOX13 (AUC 0.83, CI 0.94–0.72), ZKSCAN4 (AUC 0.90, CI 0.99–0.81) ( Figures 7B–I ). The findings indicated that the acquired genes had a high value for the diagnosis of OA in combination with MetS. Figure 7 Nogrim comtraction and the diagnostic value evaluation (A) The visible nomogram for diagnosing OA with MetS (B–I) The ROC curve of each candidate gene (RHOR, KDELR3, TRAKA, PHC2, ZKSCANA, SOX13, RNF 170 and FZD7) and nomogram show the significant OA with MetS diagnostic value. OA. Osteoarthritis, MetS, Metabolic syndrome; AUC, area under the curve. 3.7 Immune infiltration analysis We discovered that genes associated with MetS can also play a role in OA, primarily in immune regulation. An in-depth examination of the nomogram and ROC revealed that they could be used as potential biomarkers for the diagnosis of OA, which was confirmed by immune infiltration. The percentage of 22 immune cells in each sample is shown in the bar graph for both datasets ( Figure 8A ). Voltammograms showed higher levels of B cells memory, Macrophages M0, Dendritic cells activated, Mast cells resting and Eosinophils in OA patients, and lower levels of T cells CD4 memory resting, Macrophages M2, Mast cells activated and Neutrophils ( Figure 8B ). Correlation analysis of the 22 immune cell types showed that NK cells resting was positively correlated with Neutrophils (r=0.62), T cells CD4 naive was positively correlated with T cells CD4 memory activated (r=0.52), Mast cells activated was negatively correlated with Mast cells resting (r=-0.67), and Macrophages M2 was negatively correlated with T cells gamma delta (r=-0.52) ( Figure 8C ). In conclusion, OA patients have varying degrees of multiple immune cell infiltration that may be potential regulatory points for the treatment of OA. Figure 8 Immune cell infiltration analysis between OA and control. (A) The proportion of 22 kinds of immune cells in different samples visualized from the burplot. (B) Comparison regarding the proportion of 22 kinds of immune cells between OA and control groups visualized by the vioplot. (C) Correlation of 22 immune cell type compositions, *,p < 0.05, **,p < 0.01, ***,p < 0.001. Both horizontal and vertical axes demonstrate immune cell subtypes OA, Osteoarthritis. 4 Discussion OA is one of the most significant causes of disability in middle-aged and older adults and places a significant burden on public health (30). Several biomarkers for OA have been identified in recent studies, including osteoclast protein (OPN), cartilage oligomeric matrix protein (COMP), cartilage acidic protein 1, and CRTAC1 (31, 32). Studies that combine these two diseases are relatively rare. There were currently no markers for diagnosing OA using nomogram and machine learning methods, so we combined both bioinformatics analysis and machine learning and evaluated their diagnostic value using nomogram and ROC. Notably, we identified eight key immune-related candidate genes (FZD7, IRAK3, KDELR3, PHC2, RHOB, RNF170, SOX13, and ZKSCAN4) and developed a nomogram for the diagnosis of OA in MetS patients. We were able to approximate the odds of developing OA in MetS patients by simply testing the peripheral blood of MetS patients with known expression of core genes, given that the MetS samples used in this study were all from peripheral blood. Because of its simplicity and effectiveness, peripheral blood testing is used in the diagnosis of many diseases (33, 34). Following that, we will develop a more refined model that can accurately reflect gene expression and assign it a numerical value to make it more accurate for diagnosis (35). When significant changes in target indicators occur in MetS patients, early monitoring and intervention can be performed, which is more valuable for diagnosing MetS combined with OA. Frizzled 7 (FZD7) belongs to the G protein-coupled receptor family and is a 7-channel transmembrane cell surface receptor and a key factor in the Wnt signaling pathway (36). Frizzled proteins have a universal structure: a cysteine structural domain (CRD) that fills the extracellular space, followed by a structural domain containing seven putative transmembrane fragments (37). CRD can interact with Wnts and form a complex that reduces β-catenin phosphorylation (38). The FZD family proteins mediate the Wnt/β-catenin signaling pathway and contribute to the overall pathological progression of OA, especially in the remodeling of cartilage (39, 40). Exos derived from hADSC have been shown to contain MiR-376c-3p, which has the ability to target WNT3 or WNT9a and inhibit the Wnt/β-catenin signaling pathway, reducing synovial fibrosis and chondrocyte degradation in OA patients. In conclusion, FZD7 can regulate cell proliferation, differentiation, migration, and tumorigenesis development by activating downstream signaling pathways through binding to Wnt ligands (41, 42). IRAK3, an interleukin-1 receptor-associated kinase (IRAK) family member, is primarily expressed in immune cells such as macrophages, monocytes, dendritic cells, and epithelial cells (43). As an important regulatory protein of the TLRs/IL-1R pathway, it is involved in the inhibition of Toll-like receptor signaling. In addition, IRAK1 and IRAK2 can dissociate from the myosin complex, and IRAK3 inhibits it, rendering IRAK1 and IRAK2 unable to interact with TRAF6 and thus inhibiting NF-B pathway activation, and reduces inflammation by reducing the production of pro-inflammatory factors (44). Of course, another possibility is that the MyD88-IRAK puposome interacts with IRAK3, which stimulates NF-B activation via complex formation, where both MEKK3 and TRAF6 can form complexes with IRAK3, which is also dependent on the NF-B pathway, to promote anti-inflammatory-related expression (45). An experiment discovered that miR-33b-3p regulated IRAK3 and that it was effective in alleviating IL-1-induced apoptosis and inflammation. Overall, this appears to be a promising OA target (46). KDELR3, a member of the KDEL family, is in charge of encoding a protein related to the endoplasmic reticulum (ER) (47). The ER influences not only the synthesis and transport of lipids and steroids but also the activity of hormones and glucose metabolism. As a result, KDELR3 expression varies significantly between individuals, with higher levels of expression in non-atherosclerotic tissues. In addition, several studies have found that KDELR3 is significantly aberrantly expressed in a variety of malignancies, including prostate adenocarcinoma and hepatocellular carcinoma (48, 49). As a result, we hypothesize that KDELR3 plays a critical role in OA patients. The HD1 structural domain of PHC2 is close to its FCS zinc finger structural domain, which is capable of interacting with the catalytic ring structural domain of RING1B. RING1B, a core component of Polycomb Repressive Complex 1 (PRC1), can monoubiquitinate histone H2A (H2AK119ub1) and plays a catalytic role at lysine 119. It is a key regulator of the estrogen receptor alpha (ER) transcriptional process (50), through which estrogen regulates biological processes like reproductive maturation, energy homeostasis, and skeletal growth by binding to ERα (51). Most importantly, estrogen action on ERα can increase miR-140 expression levels and decrease the expression levels of matrix metalloproteinase-13 (MMP-13) in human articular chondrocytes (52). It has also been shown that PHC2 can bind to the Vcam1 locus and act to reduce systemic immunodeficiency (53). It is reasonable to believe that PHC2 can have an impact on the progression of OA. RHOB is a Rho GTPase with a C-terminus and an N-terminus, the latter of which contains a G domain, also known as the RhoA-like domain (54). Because chromosome 2 retains RHOB’s genetic information and lacks the alternative RHOB precursor mRNA, only one RHOB protein sequence can be translated (55). RHOB can be induced by transforming growth factor beta (TGF) signaling and is immune to p53 regulation, allowing for a timely response to non-genotoxic (polysaccharide, hypoxia, inflammatory factors) and genotoxic stimuli (radiation) (56). RHOB deficiency inhibited pathological angiogenesis in ischemic retina patients, implying that RHOB alleviates symptoms by promoting the formation of lymphatic vessels following injury or in the presence of inflammation (57). Interestingly, RHOB also activates IL-1β (interleukin 1β), LPS (lipopolysaccharide), and TNFα (tumor necrosis factor α) in an inflammatory setting (54), contrary to previous observations. However, its mechanism of action in OA is unclear and remains to be investigated. In general, RHOB can regulate a variety of cellular processes, including vesicle trafficking, apoptosis, DNA repair, angiogenesis, proliferation, migration, and invasion (58). RNF170 is a novel E3 ubiquitin ligase that mediates the ubiquitination of the endoplasmic reticulum calcium channel sarcoglycan 1,4,5 trisphosphate receptor leading to the degradation of the sarcoglycan 1,4,5 trisphosphate receptor via the proteasome pathway and subsequently affecting the calcium flow and content in the endoplasmic reticulum lumen and cytoplasm. By specifically targeting TLR receptor molecules, RNF170 mediates polyubiquitination of lysine at position 766 on the TIR domain of TLR3 (59), resulting in the degradation of TLR3 receptor molecules via the proteasome pathway and reducing the effect on TLR3 downstream signaling. activation of the TLR3 downstream signaling pathway, thereby inhibiting the production of inflammatory factors and type I interferons (60). A search of the GEO database revealed that RNF170 protein expression was significantly increased in human monocytes after infection with human immunodeficiency virus (HIV) (61), and when the body’s fibroblasts underwent an immune response to cytomegalovirus infection, the expression level of RNF170 protein in fibroblasts showed a trend of first increasing and then decreasing (62), and in general, RNF170 is extremely important in natural immunity. SOX13 is a member of the SOX protein family, and its coding gene is located at 1q31.3-32.1, which is found in a wide range of cells and tissues. SOX13 contains 604 amino acids, including three specific functional regions, namely the leucine zipper region, the glutamine-rich sequence region, and the HMG functional region (63). Its expression in three embryonic cell lineages suggests that it may be involved in various developmental processes, with expression detected in the neural tube, developing brain, kidney, liver, mesenchymal, and chondrogenic progenitor cells, and has a significant impact on sex determination, neurogenesis and endochondral bone formation (64). Innate lymphocytes, including type 1, type 2 (ILC2) and type 3 (ILC3) subpopulations, are increasingly recognized as key regulators of tissue homeostasis and inflammation through the release of cytokines (65), with NCR ILC3s producing mainly IL-17A, which is regulated by DNA binding inhibitor 2 (ID2), transcription factor 1, and SRY (sex-determining region Y)-box transcription factor 13 (Sox13) (66). Fida et al. simultaneously tested 297 cases of primary biliary sclerosis (PBC), 22 cases of autoimmune cholangitis, 29 cases of autoimmune hepatitis, and 90 patients with T1DM for SOX13-Ab and found a high rate of positivity (67). In addition, Sox13, as an autoimmune antigen, can also modulate T-cell specificity, and in summary, it is reasonable to suspect that SOX13 can modulate the inflammatory response in OA. ZKSCAN4, a zinc finger protein family member, localizes to chromosome 6p21-p22.1 and regulates genomic stability, stem cell generation, and telomere elongation (68, 69). It is found in a variety of tissues, including the cervix, testes, and trachea, as well as the kidneys, adrenal glands, mouth, skin, lungs, brain, spleen, uterus, liver, intestines, and muscles (70). Several authors examined European and Asian populations using genome-wide association studies (GWAS) of genes and discovered significant differences in some genes between healthy people and rheumatoid arthritis patients, especially ZKSCAN4, ABCF1, BTN3A3, BRD2, FLOT1, HLA-DMA, HLA-G, HLA-F, HSP90AB1, IER3, and TUBB (71). ZKSCAN4 improves the HDM2 promoter region’s association with YY1 by modifying the HDM2 chromatin structure to promote its expression (72). HDM2 is a ubiquitin protein ligase that not only promotes the degradation but also the transcription of the tumor suppressor p53. Under hypoxia, p53 promotes cell survival by promoting the production of cellular metabolic energy in OA, inhibiting reactive oxygen species (ROS) production and ROS detoxification (73). Although there are no functional studies to validate ZKSCAN4, it can be speculated that it is tightly linked to OA. Previous research has revealed that immune and inflammatory responses are present in all phases of OA. The position of inflammatory factors and cellular infiltration in the progression of OA, as well as the various manifestations of OA in different immune settings, provide a theoretical foundation for further research into the relationship between immunity and OA. Many diseases are now recognized as being influenced by the immune microenvironment, and in OA patients, both cartilage damage and repair processes involve immune cells (74). It has been established that B cells and macrophages, as the primary immune cells, are involved in cartilage damage and repair, in addition to NK cells, T cells, and DCs (75). In addition, mast cells (MC), T cells, and macrophages can be found in large numbers in the synovial tissue of OA patients. A unique pattern of immune infiltration has recently been identified, characterized by increased polarization of CD4+ T cells to activated Th1 cells and increased secretion of immunomodulatory cytokines (76). This has similarities to our findings, with OA patients having higher levels of CD8+ T cells, B-cell memory, activated dendritic cells, M0 macrophages, and eosinophils, and lower levels of Mast cells activated, T cells CD4 memory activated, M2 macrophages and Neutrophils. Above all, understanding inflammatory signaling mechanisms is critical for OA diagnosis and treatment. 5 Conclusion Based on bioinformatics analysis and machine learning, we systematically identified eight related candidate genes (FZD7, IRAK3, KDELR3, PHC2, RHOB, RNF170, SOX13, and ZKSCAN4) and provided a template for the diagnosis of OA combined with MetS. We also noticed that the immune system of MetS patients with OA is out of balance, that the percentage of immune cells can be affected by the immune microenvironment, and that the screened genes could be used for clinical diagnosis and treatment. 6 Limitation There are some restrictions on our research. First, despite pooling three OA datasets, the samples are still tiny, and because of the small sample size, the diagnostic value of the column line graphs is quite high. Additionally, we wanted to choose a different sample to verify the diagnostic results. The eligible ones weren’t accessible, and the Mets dataset was already small. As a result, we were unable to confirm the results. After that, the findings should be verified in a bigger study with a larger sample size. Second, even though the interactions between candidate hub genes and dysregulated immune cells warrant further study, the eight candidate hub genes are primarily concentrated in regulatory immune pathways. Data availability statement The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author. Author contributions All authors listed have made a substantial, direct, and intellectual contribution to the work, and approved it for publication. Conflict of interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Publisher’s note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. Supplementary material The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fimmu.2023.1134412/full#supplementary-material Click here for additional data file. Click here for additional data file. ==== Refs References 1 Prieto-Alhambra D Judge A Javaid MK Cooper C Diez-Perez A Arden NK . Incidence and risk factors for clinically diagnosed knee, hip and hand osteoarthritis: Influences of age, gender and osteoarthritis affecting other joints. Ann Rheum Dis (2014) 73 (9 ):1659–64. doi: 10.1136/annrheumdis-2013-203355 2 Disease GBD Injury I Prevalence C . Global, regional, and national incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, 1990-2017: A systematic analysis for the global burden of disease study 2017. Lancet (London England) (2018) 392 (10159 ):1789–858. doi: 10.1016/S0140-6736(18)32279-7 3 Blagojevic M Jinks C Jeffery A Jordan KP . Risk factors for onset of osteoarthritis of the knee in older adults: A systematic review and meta-analysis. Osteoarthritis Cartilage (2010) 18 (1 ):24–33. doi: 10.1016/j.joca.2009.08.010 19751691 4 Fahed G Aoun L Bou Zerdan M Allam S Bou Zerdan M Bouferraa Y . Metabolic syndrome: Updates on pathophysiology and management in 2021. Int J Mol Sci (2022) 23 (2 ):786. doi: 10.3390/ijms23020786 35054972 5 Beltran-Sanchez H Harhay MO Harhay MM McElligott S . Prevalence and trends of metabolic syndrome in the adult U.S. population, 1999-2010. J Am Coll Cardiol (2013) 62 (8 ):697–703. doi: 10.1016/j.jacc.2013.05.064 23810877 6 Zhuo Q Yang W Chen J Wang Y . Metabolic syndrome meets osteoarthritis. Nat Rev Rheumatol (2012) 8 (12 ):729–37. doi: 10.1038/nrrheum.2012.135 7 Chen D Shen J Zhao W Wang T Han L Hamilton JL . Osteoarthritis: Toward a comprehensive understanding of pathological mechanism. Bone Res (2017) 5 :16044. doi: 10.1038/boneres.2016.44 28149655 8 Courties A Gualillo O Berenbaum F Sellam J . Metabolic stress-induced joint inflammation and osteoarthritis. Osteoarthritis Cartilage (2015) 23 (11 ):1955–65. doi: 10.1016/j.joca.2015.05.016 9 Terkawi MA Ebata T Yokota S Takahashi D Endo T Matsumae G . Low-grade inflammation in the pathogenesis of osteoarthritis: Cellular and molecular mechanisms and strategies for future therapeutic intervention. Biomedicines (2022) 10 (5 ):1109. doi: 10.3390/biomedicines10051109 35625846 10 Zhang H Cai D Bai X . Macrophages regulate the progression of osteoarthritis. Osteoarthritis Cartilage (2020) 28 (5 ):555–61. doi: 10.1016/j.joca.2020.01.007 11 Zheng L Zhang Z Sheng P Mobasheri A . The role of metabolism in chondrocyte dysfunction and the progression of osteoarthritis. Ageing Res Rev (2021) 66 :101249. doi: 10.1016/j.arr.2020.101249 33383189 12 Ramirez-Perez S Reyes-Perez IV Martinez-Fernandez DE Hernandez-Palma LA Bhattaram P . Targeting inflammasome-dependent mechanisms as an emerging pharmacological approach for osteoarthritis therapy. iScience (2022) 25 (12 ):105548. doi: 10.1016/j.isci.2022.105548 36465135 13 Nemet M Blazin T Milutinovic S Cebovic T Stanojevic D Zvekic Svorcan J . Association between metabolic syndrome, its components, and knee osteoarthritis in premenopausal and menopausal women: A pilot study. Cureus (2022) 14 (7 ):e26726. doi: 10.7759/cureus.26726 35967163 14 Le Clanche S Bonnefont-Rousselot D Sari-Ali E Rannou F Borderie D . Inter-relations between osteoarthritis and metabolic syndrome: A common link? Biochimie (2016) 121 :238–52. doi: 10.1016/j.biochi.2015.12.008 15 Courties A Sellam J Berenbaum F . Metabolic syndrome-associated osteoarthritis. Curr Opin Rheumatol (2017) 29 (2 ):214–22. doi: 10.1097/BOR.0000000000000373 16 Jiang X Zhong R Dai W Huang H Yu Q Zhang JA . Exploring diagnostic biomarkers and comorbid pathogenesis for osteoarthritis and metabolic syndrome Via bioinformatics approach. Int J Gen Med (2021) 14 :6201–13. doi: 10.2147/IJGM.S325561 17 Barrett T Wilhite SE Ledoux P Evangelista C Kim IF Tomashevsky M . Ncbi geo: Archive for functional genomics data sets–update. Nucleic Acids Res (2013) 41 (Database issue ):D991–5. doi: 10.1093/nar/gks1193 18 Woetzel D Huber R Kupfer P Pohlers D Pfaff M Driesch D . Identification of rheumatoid arthritis and osteoarthritis patients by transcriptome-based rule set generation. Arthritis Res Ther (2014) 16 (2 ):R84. doi: 10.1186/ar4526 24690414 19 D'Amore S Hardfeldt J Cariello M Graziano G Copetti M Di Tullio G . Identification of mir-9-5p as direct regulator of Abca1 and hdl-driven reverse cholesterol transport in circulating Cd14+ cells of patients with metabolic syndrome. Cardiovasc Res (2018) 114 (8 ):1154–64. doi: 10.1093/cvr/cvy077 20 Langfelder P Horvath S . Wgcna: An r package for weighted correlation network analysis. BMC Bioinf (2008) 9 :559. doi: 10.1186/1471-2105-9-559 21 The Gene Ontology C . The gene ontology resource: 20 years and still going strong. Nucleic Acids Res (2019) 47 (D1 ):D330–D8. doi: 10.1093/nar/gky1055 22 Kanehisa M Goto S . Kegg: Kyoto encyclopedia of genes and genomes. Nucleic Acids Res (2000) 28 (1 ):27–30. doi: 10.1093/nar/28.1.27 10592173 23 Yang C Delcher C Shenkman E Ranka S . Machine learning approaches for predicting high cost high need patient expenditures in health care. BioMed Eng Online (2018) 17 (Suppl 1 ):131. doi: 10.1186/s12938-018-0568-3 30458798 24 Ellis K Kerr J Godbole S Lanckriet G Wing D Marshall S . A random forest classifier for the prediction of energy expenditure and type of physical activity from wrist and hip accelerometers. Physiol Meas (2014) 35 (11 ):2191–203. doi: 10.1088/0967-3334/35/11/2191 25 Zhang M Zhu K Pu H Wang Z Zhao H Zhang J . An immune-related signature predicts survival in patients with lung adenocarcinoma. Front Oncol (2019) 9 :1314. doi: 10.3389/fonc.2019.01314 31921619 26 Alderden J Pepper GA Wilson A Whitney JD Richardson S Butcher R . Predicting pressure injury in critical care patients: A machine-learning model. Am J Crit Care (2018) 27 (6 ):461–8. doi: 10.4037/ajcc2018525 27 Pan X Jin X Wang J Hu Q Dai B . Placenta inflammation is closely associated with gestational diabetes mellitus. Am J Transl Res (2021) 13 (5 ):4068–79. 28 Newman AM Liu CL Green MR Gentles AJ Feng W Xu Y . Robust enumeration of cell subsets from tissue expression profiles. Nat Methods (2015) 12 (5 ):453–7. doi: 10.1038/nmeth.3337 29 Hu K . Become competent within one day in generating boxplots and violin plots for a novice without prior r experience. Methods Protoc (2020) 3 (4 ):64. doi: 10.3390/mps3040064 32977580 30 Pereira D Peleteiro B Araujo J Branco J Santos RA Ramos E . The effect of osteoarthritis definition on prevalence and incidence estimates: A systematic review. Osteoarthritis Cartilage (2011) 19 (11 ):1270–85. doi: 10.1016/j.joca.2011.08.009 31 Silverwood V Blagojevic-Bucknall M Jinks C Jordan JL Protheroe J Jordan KP . Current evidence on risk factors for knee osteoarthritis in older adults: A systematic review and meta-analysis. Osteoarthritis Cartilage (2015) 23 (4 ):507–15. doi: 10.1016/j.joca.2014.11.019 32 Watt FE . Osteoarthritis biomarkers: Year in review. Osteoarthritis Cartilage (2018) 26 (3 ):312–8. doi: 10.1016/j.joca.2017.10.016 33 Ohe Y Fushida S Yamaguchi T Kinoshita J Saito H Okamoto K . Peripheral blood platelet-lymphocyte ratio is good predictor of chemosensitivity and prognosis in gastric cancer patients. Cancer Manag Res (2020) 12 :1303–11. doi: 10.2147/CMAR.S241069 34 Rao C Jin J Lu J Wang C Wu Z Zhu Z . A multielement prognostic nomogram based on a peripheral blood test, conventional mri and clinical factors for glioblastoma. Front Neurol (2022) 13 :822735. doi: 10.3389/fneur.2022.822735 35250826 35 OH B Gransar H Callister T Shaw LJ Schulman-Marcus J Stuijfzand WJ . Development and validation of a simple-to-Use nomogram for predicting 5-, 10-, and 15-year survival in asymptomatic adults undergoing coronary artery calcium scoring. JACC Cardiovasc Imaging (2018) 11 (3 ):450–8. doi: 10.1016/j.jcmg.2017.03.018 36 MacDonald BT Tamai K He X . Wnt/Beta-catenin signaling: Components, mechanisms, and diseases. Dev Cell (2009) 17 (1 ):9–26. doi: 10.1016/j.devcel.2009.06.016 19619488 37 Wang Y Chang H Rattner A Nathans J . Frizzled receptors in development and disease. Curr Top Dev Biol (2016) 117 :113–39. doi: 10.1016/bs.ctdb.2015.11.028 38 Schulte G . International union of basic and clinical pharmacology. lxxx. the class frizzled receptors. Pharmacol Rev (2010) 62 (4 ):632–67. doi: 10.1124/pr.110.002931 39 Lories RJ Corr M Lane NE . To wnt or not to wnt: The bone and joint health dilemma. Nat Rev Rheumatol (2013) 9 (6 ):328–39. doi: 10.1038/nrrheum.2013.25 40 Mariani E Pulsatelli L Facchini A . Signaling pathways in cartilage repair. Int J Mol Sci (2014) 15 (5 ):8667–98. doi: 10.3390/ijms15058667 41 Fernandez A Huggins IJ Perna L Brafman D Lu D Yao S . The wnt receptor Fzd7 is required for maintenance of the pluripotent state in human embryonic stem cells. Proc Natl Acad Sci U.S.A. (2014) 111 (4 ):1409–14. doi: 10.1073/pnas.1323697111 42 Simonetti M Agarwal N Stosser S Bali KK Karaulanov E Kamble R . Wnt-fzd signaling sensitizes peripheral sensory neurons Via distinct noncanonical pathways. Neuron (2014) 83 (1 ):104–21. doi: 10.1016/j.neuron.2014.05.037 43 Miyata M Lee JY Susuki-Miyata S Wang WY Xu H Kai H . Glucocorticoids suppress inflammation Via the upregulation of negative regulator irak-m. Nat Commun (2015) 6 :6062. doi: 10.1038/ncomms7062 25585690 44 Hubbard LL Moore BB . Irak-m regulation and function in host defense and immune homeostasis. Infect Dis Rep (2010) 2 (1 ):e9. doi: 10.4081/idr.2010.e9 21390243 45 Freihat LA Wheeler JI Wong A Turek I Manallack DT Irving HR . Irak3 modulates downstream innate immune signalling through its guanylate cyclase activity. Sci Rep (2019) 9 (1 ):15468. doi: 10.1038/s41598-019-51913-3 31664109 46 Tao T Zhang Y Wei H Heng K . Downregulation of Irak3 by mir-33b-3p relieves chondrocyte inflammation and apoptosis in an in vitro osteoarthritis model. Biosci Biotechnol Biochem (2021) 85 (3 ):545–52. doi: 10.1093/bbb/zbaa105 47 Trychta KA Back S Henderson MJ Harvey BK . Kdel receptors are differentially regulated to maintain the er proteome under calcium deficiency. Cell Rep (2018) 25 (7 ):1829–40 e6. doi: 10.1016/j.celrep.2018.10.055 30428351 48 Bai Y Qi W Liu L Zhang J Pang L Gan T . Identification of seven-gene hypoxia signature for predicting overall survival of hepatocellular carcinoma. Front Genet (2021) 12 :637418. doi: 10.3389/fgene.2021.637418 33912215 49 Xu Z Xu L Liu L Li H Jin J Peng M . A glycolysis-related five-gene signature predicts biochemical recurrence-free survival in patients with prostate adenocarcinoma. Front Oncol (2021) 11 :625452. doi: 10.3389/fonc.2021.625452 33954109 50 Zhang Y Chan HL Garcia-Martinez L Karl DL Weich N Slingerland JM . Estrogen induces dynamic eralpha and Ring1b recruitment to control gene and enhancer activities in luminal breast cancer. Sci Adv (2020) 6 (23 ):eaaz7249. doi: 10.1126/sciadv.aaz7249 32548262 51 Mauvais-Jarvis F Clegg DJ Hevener AL . The role of estrogens in control of energy balance and glucose homeostasis. Endocr Rev (2013) 34 (3 ):309–38. doi: 10.1210/er.2012-1055 52 Liang Y Duan L Xiong J Zhu W Liu Q Wang D . E2 regulates mmp-13 Via targeting mir-140 in il-1beta-Induced extracellular matrix degradation in human chondrocytes. Arthritis Res Ther (2016) 18 (1 ):105. doi: 10.1186/s13075-016-0997-y 27165343 53 Bae J Choi SP Isono K Lee JY Park SW Choi CY . Phc2 controls hematopoietic stem and progenitor cell mobilization from bone marrow by repressing Vcam1 expression. Nat Commun (2019) 10 (1 ):3496. doi: 10.1038/s41467-019-11386-4 31375680 54 Eckenstaler R Hauke M Benndorf RA . A current overview of rhoa, rhob, and rhoc functions in vascular biology and pathology. Biochem Pharmacol (2022) 206 :115321. doi: 10.1016/j.bcp.2022.115321 36306821 55 Schaefer A Reinhard NR Hordijk PL . Toward understanding rhogtpase specificity: Structure, function and local activation. Small GTPases (2014) 5 (2 ):6. doi: 10.4161/21541248.2014.968004 25483298 56 Wang XH Wang Y Diao F Lu J . Rhob is involved in lipopolysaccharide-induced inflammation in mouse in vivo and in vitro. J Physiol Biochem (2013) 69 (2 ):189–97. doi: 10.1007/s13105-012-0201-z 57 Gerald D Adini I Shechter S Perruzzi C Varnau J Hopkins B . Rhob controls coordination of adult angiogenesis and lymphangiogenesis following injury by regulating Vezf1-mediated transcription. Nat Commun (2013) 4 :2824. doi: 10.1038/ncomms3824 24280686 58 Gutierrez E Cahatol I Bailey CAR Lafargue A Zhang N Song Y . Regulation of rhob gene expression during tumorigenesis and aging process and its potential applications in these processes. Cancers (Basel) (2019) 11 (6 ):818. doi: 10.3390/cancers11060818 31200451 59 Song XQ . Negative regulation of Tlr3-triggered innate immune response by E3 ligase Rnf170 [D]. Peking Union Medical College (2018). 60 Oshiumi H Matsumoto M Funami K Akazawa T Seya T . Ticam-1, an adaptor molecule that participates in toll-like receptor 3-mediated interferon-beta induction. Nat Immunol (2003) 4 (2 ):161–7. doi: 10.1038/ni886 61 Tilton JC Johnson AJ Luskin MR Manion MM Yang J Adelsberger JW . Diminished production of monocyte proinflammatory cytokines during human immunodeficiency virus viremia is mediated by type I interferons. J Virol (2006) 80 (23 ):11486–97. doi: 10.1128/JVI.00324-06 62 Hertel L Mocarski ES . Global analysis of host cell gene expression late during cytomegalovirus infection reveals extensive dysregulation of cell cycle gene expression and induction of pseudomitosis independent of Us28 function. J Virol (2004) 78 (21 ):11988–2011. doi: 10.1128/JVI.78.21.11988-12011.2004 63 Mertin S McDowall SG Harley VR . The DNA-binding specificity of Sox9 and other sox proteins. Nucleic Acids Res (1999) 27 (5 ):1359–64. doi: 10.1093/nar/27.5.1359 64 Wang Y Ristevski S Harley VR . Sox13 exhibits a distinct spatial and temporal expression pattern during chondrogenesis, neurogenesis, and limb development. J Histochem Cytochem (2006) 54 (12 ):1327–33. doi: 10.1369/jhc.6A6923.2006 65 Vivier E Artis D Colonna M Diefenbach A Di Santo JP Eberl G . Innate lymphoid cells: 10 years on. Cell (2018) 174 (5 ):1054–66. doi: 10.1016/j.cell.2018.07.017 66 Malhotra N Narayan K Cho OH Sylvia KE Yin C Melichar H . A network of high-mobility group box transcription factors programs innate interleukin-17 production. Immunity (2013) 38 (4 ):681–93. doi: 10.1016/j.immuni.2013.01.010 67 Fida S Myers MA Whittingham S Rowley MJ Ozaki S Mackay IR . Autoantibodies to the transcriptional factor Sox13 in primary biliary cirrhosis compared with other diseases. J Autoimmun (2002) 19 (4 ):251–7. doi: 10.1006/jaut.2002.0622 68 Lee K Gollahon LS . Zscan4 interacts directly with human Rap1 in cancer cells regardless of telomerase status. Cancer Biol Ther (2014) 15 (8 ):1094–105. doi: 10.4161/cbt.29220 69 Zalzman M Falco G Sharova LV Nishiyama A Thomas M Lee SL . Zscan4 regulates telomere elongation and genomic stability in Es cells. Nature (2010) 464 (7290 ):858–63. doi: 10.1038/nature08882 70 Ecker K Lorenz A Wolf F Ploner C Bock G Duncan T . A ras recruitment screen identifies Zkscan4 as a glucocorticoid receptor-interacting protein. J Mol Endocrinol (2009) 42 (2 ):105–17. doi: 10.1677/JME-08-0087 71 Zhu H Xia W Mo XB Lin X Qiu YH Yi NJ . Gene-based genome-wide association analysis in European and Asian populations identified novel genes for rheumatoid arthritis. PloS One (2016) 11 (11 ):e0167212. doi: 10.1371/journal.pone.0167212 27898717 72 Wei SY Shih YT Wu HY Wang WL Lee PL Lee CI . Endothelial yin yang 1 phosphorylation at S118 induces atherosclerosis under flow. Circ Res (2021) 129 (12 ):1158–74. doi: 10.1161/CIRCRESAHA.121.319296 73 Humpton TJ Vousden KH . Regulation of cellular metabolism and hypoxia by P53. Cold Spring Harb Perspect Med (2016) 6 (7 ):a026146. doi: 10.1101/cshperspect.a026146 27371670 74 Li M Yin H Yan Z Li H Wu J Wang Y . The immune microenvironment in cartilage injury and repair. Acta Biomater (2022) 140 :23–42. doi: 10.1016/j.actbio.2021.12.006 34896634 75 de Lange-Brokaar BJ Ioan-Facsinay A van Osch GJ Zuurmond AM Schoones J Toes RE . Synovial inflammation, immune cells and their cytokines in osteoarthritis: A review. Osteoarthritis Cartilage (2012) 20 (12 ):1484–99. doi: 10.1016/j.joca.2012.08.027 76 Rosshirt N Hagmann S Tripel E Gotterbarm T Kirsch J Zeifang F . A predominant Th1 polarization is present in synovial fluid of end-stage osteoarthritic knee joints: Analysis of peripheral blood, synovial fluid and synovial membrane. Clin Exp Immunol (2019) 195 (3 ):395–406. doi: 10.1111/cei.13230 30368774
PMC010xxxxxx/PMC10157391.txt
==== Front Front Public Health Front Public Health Front. Public Health Frontiers in Public Health 2296-2565 Frontiers Media S.A. 37151582 10.3389/fpubh.2023.1177695 Public Health Perspective NGS implementation for monitoring SARS-CoV-2 variants in Chicagoland: An institutional perspective, successes and challenges Tartanian Aileen C. 1 Mulroney Nicole 1 Poselenzny Kelly 1 Akroush Michael 1 Unger Trevor 1 Helseth Donald L. Jr. 1 Sabatini Linda M. 1 2 Bouma Michael 1 Larkin Paige M.K. 1 2 * 1NorthShore University HealthSystem, Evanston, IL, United States 2Pritzker School of Medicine, The University of Chicago, Chicago, IL, United States Edited by: Peera Hemarajata, Los Angeles County Department of Public Health, United States Reviewed by: Emily Smith, Theiagen Genomics, United States; John Bell, California Department of Public Health, United States *Correspondence: Paige M.K. Larkin, [email protected] This article was submitted to Infectious Diseases: Epidemiology and Prevention, a section of the journal Frontiers in Public Health 20 4 2023 2023 20 4 2023 11 117769501 3 2023 30 3 2023 Copyright © 2023 Tartanian, Mulroney, Poselenzny, Akroush, Unger, Helseth, Sabatini, Bouma and Larkin. 2023 Tartanian, Mulroney, Poselenzny, Akroush, Unger, Helseth, Sabatini, Bouma and Larkin https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Identification of SARS-CoV-2 lineages has shown to provide invaluable information regarding treatment efficacy, viral transmissibility, disease severity, and immune evasion. These benefits provide institutions with an expectation of high informational upside with little insight in regards to practicality with implementation and execution of such high complexity testing in the midst of a pandemic. This article details our institution’s experience implementing and using Next Generation Sequencing (NGS) to monitor SARS-CoV-2 lineages in the northern Chicagoland area throughout the pandemic. To date, we have sequenced nearly 7,000 previously known SARS-CoV-2 positive samples from various patient populations (e.g., outpatient, inpatient, and outreach sites) to reduce bias in sampling. As a result, our hospital was guided while making crucial decisions about staffing, masking, and other infection control measures during the pandemic. While beneficial, establishing this NGS procedure was challenging, with countless considerations at every stage of assay development and validation. Reduced staffing prompted transition from a manual to automated high throughput workflow, requiring further validation, lab space, and instrumentation. Data management and IT security were additional considerations that delayed implementation and dictated our bioinformatic capabilities. Taken together, our experience highlights the obstacles and triumphs of SARS-CoV-2 sequencing. SARS-CoV-2 sequencing molecular microbiology molecular diagnostics next generating sequencing ==== Body pmcIntroduction Next generation sequencing (NGS) has been pivotal for understanding the impact of SARS-CoV-2 variants on transmission, pathogenicity, disease severity, vaccine and therapy efficacy, and diagnostic detection (1). For instance, the Omicron variant has been shown to evade the immune response in patients previously infected with SARS-CoV-2 or vaccinated against SARS-CoV-2 (2–5), render the majority of monoclonal antibody therapies ineffective (4), and cause more infections in younger patients compared to other variants (6). Conversely, the Omicron variant was associated with a lower 28-day mortality, ICU admission rate, and oxygen requirements compared to Delta (7). With the benefits of SARS-CoV-2 sequencing highlighted, implementing such testing is attractive to many healthcare systems, but there are numerous challenges and considerations that should be addressed. Here we describe our experience with implementing NGS for SARS-CoV-2 variant analysis at NorthShore University HealthSystem (NSUHS) molecular diagnostic laboratory (MDL). As a fully integrated healthcare system, NSUHS-Edward-Elmhurst Health (EEH) serves over 4.2 million residents across northeast Illinois, including the city of Chicago and six suburban counties. The system currently encompasses 8 hospitals and over 300 outpatient centers. The NSUHS MDL was the first clinical laboratory in Illinois to perform SARS-CoV-2 testing (8) and has performed over 800,000 SARS-CoV-2 diagnostic assays to date. Our initial goals of SARS-CoV-2 sequencing was to detect shifts and emergence of lineages in real time, but challenges with staffing, turn-around-time (TAT), and sample selection complicated this. Demand on the lab and lab staff The MDL has ample experience with NGS, given the breadth of oncology NGS assays performed, uniquely positioning the laboratory to bring in SARS-CoV-2 sequencing compared to laboratories without sequencing experience. We officially launched COVIDSeq on an Illumina NextSeq 550Dx (San Diego, CA, United States) in March of 2021 after delays due to installation, training, and reagent acquisition. Once launched, COVIDSeq productivity was constrained by the priority given to clinical diagnostic assays for staffing and freezer storage. Based on these factors, the percentage of SARS-CoV-2 samples tested by the MDL that progressed to sequencing ranged from 0 to 18.5% monthly for 2022, when sequencing was performed on a regular basis (Table 1). Samples were selected based on testing location and available media volume, with the exact number of samples tested fluctuating due to balancing the cost of a run and the availability of reagents and technologists to perform sequencing within a timeframe. Samples from 2020 and 2021 were run retrospectively, but due to delays discussed in detail below, most of 2021 was spent troubleshooting, validating, and optimizing. The use of a manual bioinformatics pipeline and analysis, as discussed below, complicated analysis. The addition of a new technologist to lead SARS-CoV-2 sequencing and move to more automation, both for the wet lab and dry lab components, facilitated more streamlined SARS-CoV-2 sequencing in 2022. Looking at 2022, there were factors that directly impacted the number of samples that could be sequenced per month. In September, we extracted and prepared libraries for 163 samples. However, our liquid handler malfunctioned by erroneously releasing all pipette tips and crashing the program. All samples had been depleted and the libraries were rendered unsavable. In November and December 2022, we exhausted our purchased sequencing reagents and did not have approval to order additional sequencing reagents due to the high cost that exceeded the allotted budget. All SARS-CoV-2 sequencing was self-funded by our institution, requiring careful planning and restricting the ability to expand sequencing capacity significantly. Table 1 Summary of SARS-CoV-2 samples tested clinically and sequenced at NSUHS for 2022. Month (2022) # SARS-CoV-2 clinically tested # Positive SARS-CoV-2 # Positives sequenced # Samples of sequenced with consensus sequence % Samples of positives sequenced % Samples of sequenced positives with a consensus sequence January 25,892 6,062 607 272 10.0 44.8 February 20,396 1,523 281 71 18.5 25.3 March 24,870 1,133 134 35 11.8 26.1 April 30,301 3,065 258 147 8.4 57.0 May 33,106 5,963 432 211 7.2 48.8 June 24,988 4,423 433 170 9.8 39.3 July 23,230 4,556 482 252 10.6 52.3 August 22,467 3,674 299 119 8.1 39.8 September 22,467 2,461 0 N/A 0.0 N/A October 24,972 2,394 297 75 12.4 25.3 November 28,592 2,908 0 N/A 0.0 N/A December 27,763 3,857 0 N/A 0.0 N/A Our health system utilizes several different RT-PCR platforms for SARS-CoV-2 testing, which supports large volume testing in a variety of settings, including point-of-care and at each of our hospitals. However, this also led to multiple different swabs, transport media, and sample volumes. These variations were due to different assay requirements, sporadic swab and transport media shortages, and testing locations stocking different swabs. Due to early implementation of SARS-CoV-2 RT-PCR testing, we performed testing for a number of outreach non-affiliated sites that used a variety of swabs. The utilization of multiple instruments, many without available cycle threshold (Ct) values prevented establishment and selection of samples with appropriate Ct values. Often, labs will set a minimum Ct value for sequencing to increase sequencing yield, but we did not have that ability given the lack of available Ct value data. With sequencing any positive sample in 2022, only 25.3–57.0% of positive samples resulted in a consensus sequence for a SARS-CoV-2 lineage (Table 1). Some explanations for this wide range include low viral load (most EUA platforms used for clinical testing at our healthcare system only provided positive/negative results without a Ct value), and user and instrument error, including issues described below when automation was implemented. Due to lack of staffing and available resources, we did not collect this data for 2021 as we had to retrospectively sequence samples. Complications with manual processing The lab performed SARS-CoV-2 library preparations and sequencing using Illumina’s COVIDSeq™ RUO Test. As with most NGS assays, the library preparation portion of COVIDSeq is costly in terms of time and labor. Initially, all library preparations were performed manually, requiring two full 8 h shifts for one technologist to complete (Table 2). While there were two technologists trained on the COVIDSeq assay, these technologists also were performing molecular diagnostic assays for clinical care, limiting their ability to prepare libraries for COVIDSeq to once per week. Table 2 Comparison of manual vs. automated library preparations. Manual (96 samples) Automated (48 samples) No. batches per sequencing run 4 8 No. of pipette tip boxes 33 5 Hands-on tech time 16 h 3 h→30 min Turn around time 2 days 7 h Final pool molarity >150 nM <8 nM The increased demand for clinical plastic consumables caused backorders and supply chain issues, restricting our ability to regularly perform COVIDSeq. During the first year, it was difficult to acquire an adequate amount of pipette tips to perform not only COVIDSeq but any of our routine molecular diagnostic clinical assays. Manually performing one COVIDSeq library preparation would consume 33 pipette tip boxes (Table 2). Therefore, to sequence a full run of 384 samples, over 130 tip boxes would be required. Regular and rapid sequencing during shifts and emergence of lineages such as Alpha, Delta, and Omicron would have been beneficial as these results would contribute to the local and global sequencing effort as well as guide hospital policies (e.g., allowed meeting size). However, the clinical assays consumed the necessary pipette tips and other plastic consumables so this was not feasible. Complications with automatic processing With technologist time and consumables preciously scarce, two automated liquid handlers were purchased to supplement the labor demand required for this initiative. The PerkinElmer Janus G3 liquid handler (Waltham, MA, United States) was chosen to facilitate RNA extraction because it had a high volume capacity and was relatively easy to use. Unfortunately, several calibration corrections were required after initial install due to persistent issues with probe pressure and pipette tip compatibility resulting in inconsistent reagent and sample volumes. Most calibrations would require the onsite visit of a field service technician, delaying implementation even further. Once resolved, Janus was compatible with our already implemented ThermoFisher KingFisher Flex (Waltham, MA, United States) instrument for viral RNA extraction/purification. The Beckman Coulter i7 liquid handler (Brea, CA, United States) was purchased to automate library preparation. Because COVIDSeq library preparations of 96 samples require two thermocyclers running in tandem, and the i7 had only one, the batches were halved from 96 to 48 to accommodate the missing thermocycler. The i7 reduced hands-on time from 2 days to 7 h. In addition, the i7 uses only 30% of the number of tip boxes (Table 2). While there was a greater supply of tips for the i7 compared to manual pipette tips, we went one step further to decrease our chances of competing with labs for tips by using the unusual pipette tip size of 190 μL. The i7 is convenient and improves workflow, but there were many challenges in establishing this assay automation. For instance, hard shell 96-well plates were on backorder when the i7 arrived, so we used non-hardshell 96-well plates. These plates melted and warped from the heat of the thermocycler, causing the i7 to drop, crush, and toss the plates as the grippers attempted to move them. Similarly, these plates proved to be incompatible with the reusable lid used by the i7 thermocycler. During a run, the i7 would sense every time the plate was improperly sealed and stop the program. These problems required constant attention by our lab staff, manually adjusting the fit of the thermocycler lid. This persisted until the correct 96-well plates could be obtained. Automated liquid handlers can be the source of numerous errors that are difficult to identify and troubleshoot. For example, we observed a consistent reduction in consensus sequence yield for samples positioned on the left side of the 96-well plates compared to the right side. After troubleshooting, the lab determined that the i7 instrument lacked steps within the run script to re-suspend magnetic beads prior to arraying them into samples. As the beads settled to the bottom of the source tube through the duration of the library preparation, the i7 would dispense bead storage buffer, absent of beads, to the samples on the left side of the plate while the samples on the right side received the majority of beads. Once the mixing step was supplemented into the run script, we noted an improved uniformity of sample performance coupled with vastly improved library concentration yield. Despite this fix, the library concentration from the i7 would remain inferior to the yield of manual library preparation. And, like the Janus, i7 calibrations, updates, and repairs would often be delayed because they required an onsite visit from a service technician. We continued to identify opportunities for improved efficiency. The COVIDSeq program on our i7 calls for all reagents to be placed in 1.5 ml tubes and kept chilled on a cold block on the deck, reducing hands-on work for technologists, but it could take up to 3 h for a technologist to prepare the 1.5 ml tubes of master mixes. Over time, we reduced this timeframe by over 80% because we found that many of these master mixes could be prepared and frozen in advance without sacrificing library preparation performance (Table 2). Manual library preparations would typically produce >150 nM pooled libraries, but switching to automation resulted in libraries of <8 nM, despite the corrections made to the run script. Pools with a molarity <0.5 nM would result in a total batch failure defined as a 0% consensus sequence. However, above 0.5 nM, we found no correlation (R2 = 0.0442) between a pool’s molarity and the percentage of samples resulting in a consensus sequence (Supplementary Figure S1). Ideally, we would quantify all individual specimens after RNA extraction, cDNA synthesis, and library preparation, removing low-concentration samples at each QC step. However, our lab does not have a high throughput way to quantify 96 samples at a time, so we only quantified each batch’s pooled library prior to combining the pools for sequencing. Bioinformatics and cybersecurity NGS generates millions of sequencing reads per sample, and analyzing these reads requires a robust bioinformatics pipeline in an effort to detect and track novel variants. When the bioinformatics infrastructure is insufficient to support this immense quantity of data, institutions typically opt for commercially available solutions; either cloud-based or local, for their bioinformatics pipeline needs due to ease of use and readily available customer support. Cloud-based applications have the benefit of ease-of-use and easily accessible vendor support; however, the ever-growing push for cloud application usage provides tremendous cybersecurity concern for institutions and often requires a lengthy and in-depth risk assessment, which can delay implementation. Using a local analysis bioinformatics application platform can reduce concern from a cybersecurity perspective, but it increases cost as these systems often require the purchase of licensed software and additional hardware. Our initiative to implement a SARS-CoV-2 NGS assay was driven by immediate need to contribute in variant tracking within our community. Due to urgent importance and to avoid further delay in implementation, we opted to purchase the local Illumina DRAGEN server (as opposed to Illumina’s cloud-based application BaseSpace) to be the primary source of our bioinformatics data analysis. At the time, a BaseSpace subscription would have forced an extensive, lengthy risk assessment by our cybersecurity team as these cloud-based applications do not always satisfy standard HIPAA requirements to protect personal health information (PHI). The DRAGEN COVIDSeq test local pipeline provided a summary report of positive or negative results along with output directories containing the desired FASTA and VCF files. FASTAs, BAMs and VCFs generated by the Illumina DRAGEN software on the NextSeq 550Dx sequencer were copied to a separate Linux server for analysis. Initially, we ran our own variant calling pipeline using open source software (using samtools), visualizing the results in IGV, and running a local copy of Ensembl VEP for COVID-19 to annotate the variant consequences. This labor-intensive effort was quickly abandoned when we began using more specialized open source software packages provided by Nextstrain, Pangolin, and Nextclade, reducing the necessity of manual analysis. After using Nextstrain (9) for a few months, we recognized that variant nomenclature was evolving away from Nextstrain clade names to Pango lineage (10) and WHO labels. To generate Pango names we analyzed merged FASTA files using the latest version of Pangolin (11). Nextclade (12) was also used to compare and summarize variant classifications by uploading our merged FASTA files (13). Launching the DRAGEN COVIDSeq local pipeline was initiated via the Linux command line terminal. This method is extremely foreign to users who are accustomed to GUI-based software with only little to moderate Linux command line experience. Customer support was a necessity, particularly support via vendor remote access, as we experienced frequent pipeline analysis failures along with connectivity issues between the DRAGEN server and the NextSeq 550Dx. Vendor bioinformatics support is generally equipped to support their customers remotely. NorthShore HIT did not permit vendor remote support access, limiting our only options to lengthy phone conversations or email correspondence. With restricted remote access to independently investigate and troubleshoot, vendors rely on these often mutually time consuming methods to investigate and eventually resolve the issue. Lack of proper vendor remote support to address these issues contributed to lengthy delays in data processing as resolution to these problems often extended across multiple days. With this workflow, we quickly realized that our goal to track lineage shifts in real time would be extremely difficult to accomplish. Available bandwidth for our highly talented yet small bioinformatics team was limited, as our established clinical oncology NGS assays were beginning to rebound to pre-pandemic volumes. Building and maintaining a local pipeline intended to track current lineages shifts required a considerable amount of bioinformatics support beyond the limits of our available institutional resources. Internal bioinformatics resources were not the only struggle experienced through this initial process. The laboratory workflow required for the DRAGEN COVIDSeq test pipeline included a requirement for a positive, negative, and no template control for each set of 96 indices to be included in each sequencing run. In the event of a control failure, the entire set of 96 samples became invalid. To avoid risk of control failure, each positive control required a fresh serial dilution prior to each library preparation. These dilutions were not recommended to be stored long term. Since our intent was to only sequence known SARS-CoV-2 positive samples, the inclusion of controls seemed to hold little value and only added complexity to the workflow. The local DRAGEN COVIDSeq pipeline did provide some upside. Each analysis completed rather quickly (usually within 1 to 2 h) and provided the necessary output data required for lineage identification. However, because the workflow to maintain this pipeline became unmanageable, we made the decision to purchase a BaseSpace subscription and shift our analysis to this cloud-based application. This transition required a lengthy approval process through our HIT cybersecurity team as cloud-based NGS data analysis increases potential risk to loss of PHI. To diminish this risk, we decided that all samples would remain de-identified throughout the wet bench, sequencing, and post sequencing analysis. All data would be presented as aggregated de-identified data with no link to clinical information. We did not have permission from HIT to submit any data to GISAID. Not only were our samples de-identified on the sequencer, but our institution considers date collected as PHI. This information is requested by GISAID for submission. Clinical microbiology laboratories at other institutions were able to submit completely anonymized samples to their academic colleagues for sequencing and in turn were able to successfully report de-identified metadata to GISAID and NCBI (14). These labs had IRBs that allowed patient-level data to be reported back to public health entities as the clinical labs retained access to the patient-level data while the academic sequencing partners did not have access (14). This approach, which requires institutional approval, infrastructure for de-identifying and re-identifying, and access to academic sequencing laboratories, would be ideal to allow dissemination of data to public health and biorepositories. In our case, sequencing was so delayed that our public health colleagues would have already sequenced those samples of interest, creating another hurdle for rapid collaboration. Using the DRAGEN COVIDSeq pipeline via BaseSpace Sequence Hub resolved many of the previously mentioned concerns, including ease of use. Although analysis times increased by four-fold due to the shared traffic of the cloud-based server, the data analysis process was exponentially easier as it required very little intervention from internal staff and remote support was easily available to resolve problems. However, when launching the DRAGEN COVID lineage application, the sample selection process seems to be the most taxing step. Samples can be selected in groups, but careful attention is required as it is easy to unintentionally include or exclude samples from analysis. Identifying samples to be analyzed through the application can be difficult as the sample list includes both completed and analysis-pending samples. These concerns are rather minor compared to our prior workflow and the DRAGEN COVID lineage application has provided a manageable data analysis workflow as the application provides mapping/alignment and variant calling features. Open source databases, like NextClade and Pangolin, are routinely updated and made available for analysis through the application, and the data is easily viewable and managed by multiple users. Clinical relevance/discussion Molecular diagnostic assays that directly impact patient care were prioritized over SARS-CoV-2 sequencing, posing a challenge to continue RUO sequencing at high capacity. This was particularly the case during SARS-CoV-2 waves, when staffing was reduced due to illness and supplies were in high demand (15). As a result, there were substantial delays (>1 month) in sequencing SARS-CoV-2 specimens, contrasting with our original plan of using COVIDSeq to capture shifts and emergence of SARS-CoV-2 lineages. In addition to the cost, sequencing all specimens would likely provide little additional information as most samples received during pandemic waves would have the same composition of lineages that would be better captured with a smaller representative sample selection. On the other hand, between waves, our sample volume was too low to form any statistically relevant conclusions. Furthermore, it would take substantial time to accumulate 384 specimens for a full sequencing run, delaying results or forcing a partial run, which was costly. Surges in cases led us to recruit additional resource staff and research lab team members to work additional shifts to propel sequencing efforts, manually sorting through the samples to confirm positives, creating specimen labels, aliquoting, and documenting. As previously discussed, our results were de-identified and mass aggregated to demonstrates shifts and trends within our patient population. While ideally we could share our results with our local health department to aid in their sequencing efforts, our results were not only delayed, but also did not have linked clinical data. This meant that sequencing efforts were unnecessarily duplicated due to inability to coordinate and share results, furthering the documented gap between public health labs and clinical labs (1, 16). We were, however, able to capture data categorized by symptomatic vs. asymptomatic cases and had these samples designated with their own test code for easy sorting and comparison. This comparison relied on trusting that physicians selected the correct test code indicating the presence or absence of symptoms. While we had planned to use these data to make comparisons between lineage and symptomatic state, upon review, we found that a small portion of physicians erroneously ordered the wrong test code and thus, accurate conclusions required substantial review. If the test codes had been appropriately ordered, the comparison in lineage between symptomatic and asymptomatic patients could have contributed to our knowledge in the field. In the end, we were able to share monthly trends with our healthcare system, modeling what other institutions have done (14). Our decisions to sequence various populations and ultimately switch to mostly inpatient and ED specimens likely resulted in selection bias toward patients whose SARS-CoV-2 infection was not only symptomatic, but severe enough to seek hospital treatment, as well as selecting toward patients from high risk ages (including infants and those over the age of 65 years old) and individuals with pre-existing health conditions. The challenges of inferring clinical impact of variants have been well documented (17) as it is impossible to get a truly representative sample. Severity of symptoms is subjective and testing restrictions fluctuated throughout the pandemic, with some hospital systems only allowing the sickest patients to get tested (17). Moreover, COVID-19 studies often focus on hospitalized patients, not representative of the general population (17). The shift to at-home antigen testing also biases against sequencing asymptomatic or mildly symptomatic patients (16). Despite the issues discussed previously, our data were useful in a broader capacity for our healthcare system. While there were detected cases of Omicron in our state, our sequencing confirmed the presence of Omicron in our patient population. This contributed to discussions on policies for masking and permitted meeting sizes. Furthermore, in conjunction with in silico analysis, we used COVIDSeq to test detection of sequence-confirmed variants in our lab-developed SARS-CoV-2 assay. We were able to confirm that the primers for this clinical assay could still detect even the most recently detected lineages of SARS-CoV-2. This was a concern for clinical laboratories across the world as the Alpha and Omicron variants exhibited spike (S) gene dropouts on assays that detect the S gene (18). While we do not currently utilize any assay that targets S gene, mutations can occur in any region of the genome and thus it is important to monitor whether these mutations impact the ability for our assays to detect SARS-CoV-2. The question of balancing cost, in terms of time and money, as well as staffing remains difficult and potentially unsustainable in the long-term for genomic surveillance. At times where multiple lineages are circulating, there was a push for more sequencing to better document lineage changes within our patient population, with the caveat that we do not have the capacity to provide rapid TAT for COVIDSeq. When there was an overwhelmingly predominant lineage, there was less institutional support for routine sequencing as, until a new variant of interest or concern is identified or mutations within current circulating variants would render treatments ineffective, the results would not impact hospital protocols. However, this approach would prevent detection of shifts in lineages as well as detection of novel lineages. The challenges described here were not unique to our health care system. Both the importance of localized surveillance efforts as well as the extensive challenges in terms of labor force, supply chain issues, and coordinated data acquisition, analysis and sharing became painfully evident. In recognition, Congress passed the “Tracking Pathogen Act” as part of pandemic preparedness measures within the FY2023 Omnibus legislation. This Act directs the Department of Health and Human Services to issue guidance and to support such efforts. Data availability statement The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation. Author contributions AT: writing—original draft, writing—review and editing, visualization, data curation, and investigation. NM, KP, MA, and TU: writing—review and editing, data curation, and investigation. DH: writing—review and editing, formal analysis, and software. LS: writing—review and editing, supervision, and conceptualization. MB: writing—conceptualization, original draft, writing—review and editing, supervision, data curation, and formal analysis. PL: writing—conceptualization, original draft, writing—review and editing, and project administration. All authors contributed to the article and approved the submitted version. Conflict of interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Publisher’s note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. We extend our gratitude to the NSUHS MDL for all of their work providing timely and accurate patient results as well as assisting with COVIDSeq endeavors through coverage and testing support. We would also like to thank our administrative team, Karen Kaul, Lakshmi Halasyamani, and Matt Charles, for their support. Brian Staes and Mark Delamar were instrumental in positive specimen collection outreach and the NSUHS Core Laboratory lead the specimen receiving challenge. The NSUHS Microbiology Laboratory was essential in collection and storage of patient specimens. Donna Schora manually printed all labels and tracked positive specimens while the Research Laboratory stored all samples. We would also like to thank all resource individuals who contributed to the COVIDSeq project. Supplementary material The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpubh.2023.1177695/full#supplementary-material Supplementary Figure 1 Lack of Correlation Between Library Pool Molarity and Percentage of Samples Generating a Consensus Sequence. Click here for additional data file. ==== Refs References 1. Martin MA VanInsberghe D Koelle K . Insights from SARS-CoV-2 sequences. Science. (2021) 371 :466–7. doi: 10.1126/science.abf3995 33510015 2. Wang M Zhou B Fan Q Zhou X Liao X Lin J . Omicron variants escape the persistent SARS-CoV-2-specific antibody response in 2-year COVID-19 convalescents regardless of vaccination. Emerg Microbes Infect. (2023) 12 :2151381. doi: 10.1080/22221751.2022.2151381 36444724 3. Dhama K Nainu F Frediansyah A Yatoo MI Mohapatra RK Chakraborty S . Global emerging omicron variant of SARS-CoV-2: impacts, challenges and strategies. J Infect Public Health. (2023) 16 :4–14. doi: 10.1016/j.jiph.2022.11.024, PMID: 36446204 4. Liu L Iketani S Guo Y Chan JF Wang M Liu L . Striking antibody evasion manifested by the omicron variant of SARS-CoV-2. Nature. (2022) 602 :676–81. doi: 10.1038/s41586-021-04388-0, PMID: 35016198 5. Ma C Chen X Mei F Xiong Q Liu Q Dong L . Drastic decline in sera neutralization against SARS-CoV-2 Omicron variant in Wuhan COVID-19 convalescents. Emerg Microbes Infect. (2022) 11 :567–72. doi: 10.1080/22221751.2022.2031311, PMID: 35060426 6. Lai A Bergna A Della Ventura C Menzo S Bruzzone B Sagradi F . Epidemiological and clinical features of SARS-CoV-2 variants circulating between April-December 2021 in Italy. Viruses. (2022) 14 :2508. doi: 10.3390/v14112508, PMID: 36423117 7. Beraud G Bouetard L Civljak R Michon J Tulek N Lejeune S . Impact of vaccination on the presence and severity of symptoms of hospitalised patients with an infection by the omicron variant (B.1.1.529) of the SARS-coV-2 (subvariant BA.1). Clin Microbiol Infect. (2022). doi: 10.1016/j.cmi.2022.12.020 8. Kaul K Singh K Sabatini L Konchak C McElvania E Larkin P . The value and institutional impact of an in-system laboratory testing during the COVID-19 pandemic. Acad Pathol. (2021) 8 :23742895211010253. doi: 10.1177/23742895211010253, PMID: 33997276 9. Hadfield J Megill C Bell SM Huddleston J Potter B Callender C . Nextstrain: real-time tracking of pathogen evolution. Bioinformatics. (2018) 34 :4121–3. doi: 10.1093/bioinformatics/bty407, PMID: 29790939 10. Rambaut A Holmes EC O'Toole A Hill V McCrone JT Ruis C . A dynamic nomenclature proposal for SARS-CoV-2 lineages to assist genomic epidemiology. Nat Microbiol. (2020) 5 :1403–7. doi: 10.1038/s41564-020-0770-5, PMID: 32669681 11. O'Toole A Scher E Underwood A Jackson B Hill V McCrone JT . Assignment of epidemiological lineages in an emerging pandemic using the pangolin tool. Virus Evol. (2021) 7 :veab064. doi: 10.1093/ve/veab064, PMID: 34527285 12. Aksamentov I Roemer C Hodcroft EB Neher RA . Nextclade: clade assignment, mutation calling, and quality control for viral genomes. J Open Source Softw. (2021) 6 :3773. doi: 10.21105/joss.03773 13. Nextstrain. Clade assignment, mutation calling, and sequence quality checks. Available at: https://clades.nextstrain.org (Accessed 02/01/2023) 14. Wang J Hawken SE Jones CD Hagan RS Bushman F Everett J . Collaboration between clinical and academic laboratories for sequencing SARS-CoV-2 genomes. J Clin Microbiol. (2022) 60 :e0128821. doi: 10.1128/jcm.01288-21, PMID: 34985985 15. Mahilkar S Agrawal S Chaudhary S Parikh S Sonkar SC Verma DK . SARS-CoV-2 variants: impact on biological and clinical outcome. Front Med. (2022) 9 :995960. doi: 10.3389/fmed.2022.995960, PMID: 36438034 16. Ling-Hu T Rios-Guzman E Lorenzo-Redondo R Ozer EA Hultquist JF . Challenges and opportunities for global genomic surveillance strategies in the COVID-19 era. Viruses. (2022) 14 :2532. doi: 10.3390/v14112532, PMID: 36423141 17. Griffith GJ Morris TT Tudball MJ Herbert A Mancano G Pike L . Collider bias undermines our understanding of COVID-19 disease risk and severity. Nat Commun. (2020) 11 :5749. doi: 10.1038/s41467-020-19478-2, PMID: 33184277 18. McMillen T Jani K Robilotti EV Kamboj M Babady NE . The spike gene target failure (SGTF) genomic signature is highly accurate for the identification of alpha and omicron SARS-CoV-2 variants. Sci Rep. (2022) 12 :18968. doi: 10.1038/s41598-022-21564-y, PMID: 36347878
PMC010xxxxxx/PMC10157571.txt
==== Front J Neurol J Neurol Journal of Neurology 0340-5354 1432-1459 Springer Berlin Heidelberg Berlin/Heidelberg 37140729 11748 10.1007/s00415-023-11748-5 Original Communication Self-reported life-space mobility in the first year after ischemic stroke: longitudinal findings from the MOBITEC-Stroke project http://orcid.org/0000-0001-6200-307X Hinrichs Timo [email protected] 1 Rössler Roland 12 Infanger Denis 1 Weibel Robert 34 Schär Janine 56 Peters Eva-Maria 5 Portegijs Erja 7 Rantanen Taina 8 Schmidt-Trucksäss Arno 1 Engelter Stefan T. 59 Peters Nils 569 1 grid.6612.3 0000 0004 1937 0642 Division of Sport and Exercise Medicine, Department of Sport, Exercise, and Health, University of Basel, Grosse Allee 6, 4052 Basel, Switzerland 2 grid.6612.3 0000 0004 1937 0642 Basel Mobility Center, University Department of Geriatric Medicine Felix Platter, University of Basel, Basel, Switzerland 3 grid.7400.3 0000 0004 1937 0650 Department of Geography, University of Zurich, Zurich, Switzerland 4 grid.7400.3 0000 0004 1937 0650 University Research Priority Program (URPP) Dynamics of Healthy Aging, University of Zurich, Zurich, Switzerland 5 grid.6612.3 0000 0004 1937 0642 Neurology and Neurorehabilitation, University Department of Geriatric Medicine Felix Platter, University of Basel, Basel, Switzerland 6 grid.417546.5 0000 0004 0510 2882 Neurology and Stroke Center, Klinik Hirslanden, Zurich, Switzerland 7 grid.4494.d 0000 0000 9558 4598 Center for Human Movement Sciences, University of Groningen, University Medical Center Groningen, Groningen, The Netherlands 8 grid.9681.6 0000 0001 1013 7965 Faculty of Sport and Health Sciences and Gerontology Research Center, University of Jyvaskyla, Jyväskylä, Finland 9 grid.410567.1 Department of Neurology and Stroke Center, University Hospital Basel and University of Basel, Basel, Switzerland 4 5 2023 4 5 2023 2023 270 8 39924003 14 3 2023 24 4 2023 26 4 2023 © The Author(s) 2023 https://creativecommons.org/licenses/by/4.0/ Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Background Life-space mobility is defined as the size of the area in which a person moves about within a specified period of time. Our study aimed to characterize life-space mobility, identify factors associated with its course, and detect typical trajectories in the first year after ischemic stroke. Methods MOBITEC-Stroke (ISRCTN85999967; 13/08/2020) was a cohort study with assessments performed 3, 6, 9 and 12 months after stroke onset. We applied linear mixed effects models (LMMs) with life-space mobility (Life-Space Assessment; LSA) as outcome and time point, sex, age, pre-stroke mobility limitation, stroke severity (National Institutes of Health Stroke Scale; NIHSS), modified Rankin Scale, comorbidities, neighborhood characteristics, availability of a car, Falls Efficacy Scale-International (FES-I), and lower extremity physical function (log-transformed timed up-and-go; TUG) as independent variables. We elucidated typical trajectories of LSA by latent class growth analysis (LCGA) and performed univariate tests for differences between classes. Results In 59 participants (mean age 71.6, SD 10.0 years; 33.9% women), mean LSA at 3 months was 69.3 (SD 27.3). LMMs revealed evidence (p ≤ 0.05) that pre-stroke mobility limitation, NIHSS, comorbidities, and FES-I were independently associated with the course of LSA; there was no evidence for a significant effect of time point. LCGA revealed three classes: “low stable”, “average stable”, and “high increasing”. Classes differed with regard to LSA starting value, pre-stroke mobility limitation, FES-I, and log-transformed TUG time. Conclusion Routinely assessing LSA starting value, pre-stroke mobility limitation, and FES-I may help clinicians identify patients at increased risk of failure to improve LSA. Keywords Cohort studies Spatial behavior Mobility limitation Physical functional performance Social participation http://dx.doi.org/10.13039/501100001711 Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung 182681 Hinrichs Timo University of BaselOpen access funding provided by University of Basel issue-copyright-statement© Springer-Verlag GmbH Germany, part of Springer Nature 2023 ==== Body pmcIntroduction Ischemic stroke is one of the main aging-related diseases [1, 2] and is a major risk factor for incident disability in activities of daily living (ADL) [3, 4]. It frequently results in permanent functional limitations [5] and—even in individuals with mild to moderate stroke—has a major impact on patients’ self-perceived mobility and participation in social life [6]. “Life-space mobility” refers to the spatial area in which people move about within their daily lives, potentially ranging from staying in the room in which one sleeps to traveling out of town. It includes the frequency of travel and assistance needed [7] and thus reflects the interplay between people’s goals, capabilities, opportunities, and demands of their environment [8, 9]. In the general older population, life-space mobility predicts disability [10], nursing home admission [11], health care utilization [12], and mortality [13]. Furthermore, life-space mobility is positively associated with quality of life [14], and its decline over time is associated with a decline of quality of life [15]. Based on its relevance for personal health and social interaction, life-space mobility can be considered an important patient-oriented outcome in geriatric rehabilitation [16]. So far, research on life-space mobility after stroke has been sparse. A longitudinal study that followed-up 89 patients post-stroke (median time after event at baseline 75 months) showed a significant decline in life-space mobility over a 2-year period [17]. After adjustment for a number of potential confounders, higher age and lower comfortable gait speed were significantly associated with a decrease in life-space mobility over time. There is also some evidence from cross-sectional studies that functional independence, lower extremity physical function, falls efficacy, and health-related quality of life are positively associated with life-space mobility after stroke [18–20]. Overall, there is a lack of prospective, longitudinal studies assessing people’s life-space mobility at several clearly defined points in time after stroke. In our study, we aimed to (a) describe life-space mobility, (b) identify factors associated with its course, and (c) detect typical trajectories in the first year after an ischemic stroke. Methods Study design MOBITEC-Stroke (“Recovery of mobility function and life-space mobility after ischemic stroke”; ISRCTN85999967) was a prospective observational study approved by the Ethics Committee of Northwestern and Central Switzerland (Reg.-No. 2019-00989). All participants provided written informed consent. Assessments were conducted at the research center 3 (T0), 6 (T1), 9 (T2) and 12 (T3) months after stroke. Clinical data from the time of event were retrieved from clinical records. The full study protocol is available elsewhere [21]. Target group, inclusion and exclusion MOBITEC-Stroke targeted community-dwelling, ambulatory patients after a first ischemic stroke. Inclusion criteria were: first ischemic stroke (confirmed by brain imaging) within the previous 3 months; age ≥ 18 years; ability to communicate verbally; ability to understand the study information and to provide written informed consent; ability to get up from a chair and sit down without help; ability to walk for a minimum of 20 m at their own pace, with or without pauses, with or without a walking aid, but without personal assistance; and presence of at least one of the following stroke-related symptoms potentially affecting gait and mobility: lower limb paralysis or ataxia, stance/gait ataxia (cerebellar or sensory), visual disturbance/field defect, central vestibular deficit or attentional deficit/neglect. Exclusion criteria included the following: not living in one’s own home; inability to walk without assistance (modified Rankin Scale, mRS, > 3); severe cognitive impairment (Montreal Cognitive Assessment score < 21 or, for persons with ≤ 12 years of education, < 20) [22]; acute psychiatric disorder; advanced terminal illness; orthopedic surgery of the lower extremities within the previous year or on-going rehabilitation measures following an inpatient surgical procedure at the time of stroke. The following questions were used to assess pre-stroke mobility limitation: “In the week before the stroke, were you able to walk 2 km?” and “In the week before the stroke, were you able to climb 1 flight of stairs?” [23]. Response options were “Yes, without difficulty”; “Yes, but with some difficulty”; “Yes, but with a great deal of difficulty”; “Yes, but not without help”; and “Not even with help”. Those reporting at least “a great deal of difficulty” in 1 of the 2 activities were excluded from participation. Recruitment and participants All patients presenting at the Stroke Center, University Hospital Basel, with an acute ischemic stroke between October 2019 and March 2021 were screened for eligibility. All eligible patients were offered the opportunity to participate in the study. Recruitment was stopped once the targeted sample size of N = 59 (see study protocol [21]) was reached. Measures We used the University of Alabama at Birmingham Study of Aging Life-Space Assessment (LSA) to measure life-space mobility at four time points (T0–T3; 3, 6, 9 and 12 months post stroke) [7]. Participants were asked to report the extent of their movement within the previous 4 weeks, categorized into five spatial levels (1 = rooms in the house outside of the room in which they sleep, 2 = immediate outdoor area, 3 = own neighborhood, 4 = outside their own neighborhood but within town, and 5 = out of town), the frequency of traveling to these levels (1 = less than once/week, 2 = 1–3 times/week, 3 = 4–6 times/ week, and 4 = daily), and whether they needed assistance (ie, 1 = personal assistance, 1.5 = assistive devices, 2 = no assistance). First, a subscore for every level was calculated by multiplying the values (as stated in parentheses above) for level, frequency, and assistance; subscores were then added to derive the composite score (used for all analyses) ranging from 0 (completely bedridden) to 120 (visiting out-of-town places every day unassisted); ie, higher scores indicate better life-space mobility. Previous research suggests considering an LSA composite score of lower than 60 as being “restricted” in life-space mobility [11, 24]—indicating that a person generally remains at home or in their neighborhood—and a change of ≥ 5 as being clinically important [25]. LSA has been reported to be highly reliable, valid and sensitive to change [7, 26, 27]. At T0 (3 months post stroke), pre-stroke mobility limitation (ie, difficulty walking 2 km and/or climbing 1 flight of stairs in the week before the event) was assessed [23]. Participants also underwent a clinical-neurological examination to determine stroke severity (National Institutes of Health Stroke Scale; NIHSS) [28] and the level of functional independence (mRS). The presence of comorbidities (heart disease, high blood pressure, lung disease, diabetes, ulcer or stomach disease, kidney disease, liver disease, anemia or other blood disease, cancer, depression, osteoarthritis/degenerative arthritis, back pain, rheumatoid arthritis) was assessed using the Self-Administered Comorbidity Questionnaire (SCQ) (“Do you have the problem?” yes vs no) [29]. At T0 (3 months post stroke) and at all follow-up visits (T1–T3; 6, 9 and 12 months post stroke), the current type of neighborhood (urban vs suburban vs rural) [30], availability of a car (yes vs no), and the level of concern about falling when performing various activities (Falls Efficacy Scale–International Version; FES-I) [31] were assessed by self-report. FES-I scores may range from 16 (no concern about falling) to 64 (severe concern about falling). Lower extremity physical function was assessed using the timed up-and-go (TUG) test, a timed test in which the individual stands up from a chair, walks around a cone 3 m away and returns to sitting in the chair—a variation suggested by Rikli and Jones [32] as an alternative to the test in which the individual walks to a mark on the floor, turns around and walks back to the starting position [33]. Statistical analyses Participant characteristics (at T0; 3 months post stroke) as well as LSA (at T0–T3; 3, 6, 9 and 12 months post stroke) were analyzed descriptively (numbers/percentages or mean/SD/median/IQR, respectively). For all further analyses, the following variables were dichotomized: age (≤ vs > median; ie, ≤ vs > 74 years); pre-stroke mobility limitation (yes = at least some difficulty walking 2 km or climbing stairs vs no = no difficulty with either activity); NIHSS (< vs ≥ median; ie, 0–1 vs ≥ 2); mRS (0–1 vs ≥ 2; ie, no symptoms or no significant disability despite symptoms vs at least slight disability), and SCQ (< 2 vs ≥ 2 comorbidities). Due to its skewed distribution, TUG time was log-transformed. We used linear mixed effects models (LMMs) with LSA score as outcome and time point, sex, age category, pre-stroke mobility limitation category, NIHSS category, mRS category, comorbidity category, type of neighborhood, availability of a car for personal use, FES-I score, and log-transformed TUG time as independent variables. For sex, age category, mRS category and NIHSS category, interactions with time point were also included as independent variables. For sex, NIHSS category, age category, pre-stroke mobility-limitation category, and mRS category, the baseline values were used; type of neighborhood, availability of a car for personal use, FES-I score and log-transformed TUG time were included as time-varying variables. As specified in the protocol, we included time as a discrete variable [21]. The model included a random intercept for subjects and was fitted using maximum likelihood for unbiased estimation of the fixed effects. We used multiple imputation to account for missing data, which would have led to a loss of 28 (of 236) incomplete observations in the model (11.9%)—with “observation” referring to a set of data of all assessed variables of a specific participant at a specific time point [34]. Specifically, we assumed an MAR process for the missing data and imputed 60 datasets using weighted predictive mean matching for continuous variables and logistic or multinomial regression for binary or polytomous categorical variables. The R package “mice” was used for imputations. Estimates were pooled using Rubin’s rules. We used pooled likelihood ratio tests as implemented in the R package “mitml” (method D4) to assess the significance of the model terms [35]. Terms were tested according to the principle of marginality [36]. We calculated model-based marginal means using the R package “ggeffects” to illustrate the effect of NIHSS category, pre-stroke mobility limitation, comorbidities category, and FES-I score (quartiles) on the course of the LSA score. Continuous predictors were set at their respective means of the observed data during calculation of marginal means. As an exploratory analysis, we elucidated typical trajectories of LSA by latent class growth analysis (LCGA) using the “lcmm” R package [37, 38]. Because of the small sample size, we opted to model time as linear continuous effect and did not allow heterogeneity within groups to keep the model parsimonious. We used the following criteria to select the optimal number of classes: BIC (smallest value), at least 5% of subjects in each class, mean posterior probability > 0.7 for all classes [39]. We performed univariate tests (ANOVA for continuous, Chi2 tests for categorical variables) for differences between the identified classes of trajectories. The p-values for the Chi2 tests were calculated by Monte Carlo simulations with 1e6 replicates. In addition to the independent variables used in the previous analyses, we tested for differences between the classes regarding LSA starting (T0; 3 months post stroke) value. We used the Benjamini–Hochberg approach to control the false-discovery rate and present the adjusted p values in addition to the unadjusted p values [40]. No imputation was performed for these exploratory analyses. The level of significance was set at p ≤ 0.05; all tests were two-tailed. We used R version 4.2.1 for all statistical calculations. Results Participants Participant characteristics (N = 59) are shown in Table 1; the flow of participants through the study is depicted in Fig. 1. On average, data collection took place (SD; range; time point) 92 (9; 74–110; T0) days; 177 (8; 167–205; T1) days; 268 (7; 259–295; T2) days; and 360 (8; 351–384; T3) days post stroke. Four participants dropped out between T0 and T1 for the following reasons: lack of interest (n = 2), health-related reasons (n = 1), and fear of COVID-19 infection (n = 1). Two participants dropped out between T2 and T3 for health-related reasons. In addition to the participants who had already dropped out at the respective point in time, three other participants did not take part in the T1 assessment for health-related reasons. Two participants did not take part in the T2 assessments due to vacation (n = 1) and for health-related reasons (n = 1).Table 1 Participants' characteristics at T0 Characteristic N Category n (%) Mean SD Median IQR Sex 59 Female 20 (33.9) Age [years] 59 71.6 10.0 74 65, 78 NIHSS score 58 0 15 (25.9) 1 11 (19.0) 2 15 (25.9) 3 8 (13.8)  ≥ 4 9 (15.5) Modified Rankin score 57 0 3 (5.3) 1 27 (47.4) 2 22 (38.6) 3 5 (8.8) Comorbiditiesa [number] 59  ≥ 2 44 (74.6) Pre-stroke mobility limitation 59 Yes (at least some difficulty walking 2 km or climbing stairs) 8 (13.6) Type of neighborhood 59 Rural 19 (32.2) Suburban 20 (33.9) Urban 20 (33.9) Availability of a car for personal use 59 Yes 32 (54.2) FES-I score 59 20.9 5.7 19 17, 22.5 Timed up-and-go [s] 59 10.1 4.2 9.0 7.4, 12.0 SD standard deviation, IQR interquartile range, NIHSS National Institutes of Health Stroke Scale, FES-I Falls Efficacy Scale-International Version aAs specified in the Self-Administered Comorbidity Questionnaire (SCQ) Fig. 1 Flow of participants through the study Description of life-space mobility The descriptive analyses of life-space mobility in the first year after stroke (Table 2) showed a large variation of values around the mean and a wide IQR at each point.Table 2 Descriptive analyses of life-space mobility 3, 6, 9 and 12 months after ischemic stroke Measure T0 (3 months) T2 (6 months) T3 (9 months) T4 (12 months) N = 59 n = 52 n = 53 n = 53 LSA composite score  Mean (SD) 69.3 (27.3) 72.3 (27.4) 79.3 (26.4) 77.8 (33.1)  Median [IQR] 72 [55.75, 92] 72 [53, 92] 86 [60, 100] 88 [49.5, 100] LSA Life-Space Assessment, SD standard deviation, IQR interquartile range Factors associated with the course of life-space mobility Based on LMMs with LSA score as outcome (Table 3), there was evidence that the following factors were associated with the course of LSA: pre-stroke mobility limitation, NIHSS category, FES-I score, and comorbidities category. The relationship between these factors and LSA is illustrated in Fig. 2A to D. The model revealed no evidence for an effect of time point on LSA (relationship depicted in Fig. 3).Table 3 Results of linear mixed effects models with Life-Space Assessment (LSA) composite score as outcome (N = 59; imputed dataset) Independent variables F-value Degrees of freedom Beta coefficient (95% CI)a p valueb Time point 2.12 3, 4456.3 0.095  T1 vs T0 2.5 (− 11.7, 16.7)  T2 vs T0 11.2 (− 3.2, 25.5)  T3 vs T0 10.3 (− 4.4, 19.7) Sex (male vs female) 0.61 1, 30985.6 7.7 (− 4.4, 19.7) 0.434 Age category (≤ vs > median)c 0.01 1, 2948.8 0.3 (-12.2, 12.8) 0.922 NIHSS category (0–1 vs ≥ 2) 3.99 1, 49590.6 − 15.5 (− 27.5, − 3.5) 0.046 Modified Rankin category (0–1 vs ≥ 2) 0.09 1, 2869.1 3.1 (− 10.2, 16.4) 0.762 Comorbidities category (< vs ≥ 2)d 8.79 1, 122958.1 − 13.7 (− 23.1, − 4.3) 0.003 Pre-stroke mobility limitation (yes vs no)e 10.08 1, 34009.2 19.9 (7.5, 32.4) 0.001 Type of neighborhood 2.25 2, 12528.5 0.105  Suburban vs rural − 4.7 (− 13.9, 4.6)  Urban vs rural − 9.5 (− 19.2, 0.2) Availability of a car for personal use (no vs yes) 1.60 1, 622.2 5.2 (− 2.7, 13.1) 0.206 FES-I score 7.95 1, 813.8 − 1.2 (− 2.0, − 0.4) 0.005 Log-transformed TUG time [s] 2.80 1, 211.4 − 13.4 (− 26.7, − 0.1) 0.096 Interaction time point * sex 1.50 3, 5137.8 0.212 Interaction time point * age category 0.15 3, 4906.1 0.930 Interaction time point * NIHSS category 0.77 3, 3436.4 0.512 Interaction time point * modified Rankin category 0.51 3, 3096.8 0.676 CI confidence interval, NIHSS National Institutes of Health Stroke Scale, FES-I Falls Efficacy Scale-International Version, TUG timed up-and-go a95% confidence intervals of beta-coefficients are based on t-distribution and pooled standard error using Rubin’s rules bp-values are derived from likelihood-ratio tests; p-values ≤ 0.05 are bolded cMedian age was 74 years dAs specified in the Self-Administered Comorbidity Questionnaire (SCQ) eAt least some difficulty walking 2 km or climbing stairs vs no difficulty Fig. 2 Marginal means (N = 59) illustrating the relationship between National Institutes of Health Stroke Scale (NIHSS) score category (0–1 vs ≥ 2) (A), pre-stroke mobility limitation (yes = at least some difficulty walking 2 km or climbing stairs vs no = no difficulty with either activity) (B); comorbidities category (< 2 vs ≥ 2 comorbidities) (C), as well as falls efficacy (quartiles of FES-I score) (D) and the course of the Life-Space Assessment (LSA) composite score 3 (T0), 6 (T1), 9 (T2) and 12 (T3) months after stroke Fig. 3 Marginal means illustrating the relationship between time point (T0 = 3, T1 = 6, T2 = 9, and T3 = 12 months after stroke) and Life-Space Assessment (LSA) composite score for the total sample (N = 59) Typical trajectories The exploratory analysis of typical trajectories of LSA revealed three classes (Fig. 4) with n = 9 (class 1; “low stable”), n = 18 (class 2; “average stable”), and n = 32 (class 3; “high increasing”) participants, respectively. While for classes 1 (p = 0.767) and 2 (p = 0.956) we found no evidence for a change of LSA over time, there was evidence of an increase in LSA over time in class 3 (p < 0.001).Fig. 4 Illustration of the 3 typical trajectories of Life-Space Assessment (LSA) composite score revealed by latent class growth analysis (LCGA) (N = 59). The graph above shows the mean (modelled linearly over time) and the 95% confidence interval for every class; the graph below shows the data of each participant (thin lines) in the respective class and the empirical mean (thick lines) at 3 (T0), 6 (T1), 9 (T2) and 12 (T3) months post-stroke The univariate tests for differences between the classes (ANOVA or Chi2 tests respectively) revealed evidence (ie, unadjusted as well as adjusted p ≤ 0.05) that classes differed with regard to LSA starting value, pre-stroke mobility limitation, FES-I score, and log-transformed TUG time (Table 4). For the comorbidities category, only the unadjusted p value was ≤ 0.05 (adjusted p = 0.093). The mean LSA starting value (SD; median) was 32.1 (19.4; 20) in class 1; 61.2 (19.2; 62) in class 2; and 84.3 (20.5; 86) in class 3. In class 1, 4 out of 9; in class 2, 4 out of 14; and in class 3, 0 out of 32 had a pre-stroke mobility limitation. The mean (SD; median) FES-I score was 28.1 (9.0; 25) in class 1; 21.1 (3.9; 20) in class 2; and 18.8 (3.5; 17.5) in class 3. The mean (SD; median) log-transformed TUG time was 2.6 (0.4; 2.5) s in class 1; 2.3 (0.3; 2.4) s in class 2; and 2.1 (0.2; 2.0) s in class 3.Table 4 Results of univariate tests (ANOVA for continuous and Chi2 tests for categorical variables) for differences between the three classes of trajectories Variable N n class I/II/III Test type Test statistic Degrees of freedom Mean (95% CI) Class I Mean (95% CI) Class II Mean (95% CI) Class III Un-adjusted p valueb Adjusteda p valueb Sex 59 9/18/32 Chi2 3.21 NA 0.197 0.217 Age category (≤ vs > median)c 59 9/18/32 Chi2 4.10 NA 0.126 0.174 NIHSS category (0–1 vs ≥ 2) 59 9/18/32 Chi2 1.07 NA 0.625 0.625 Modified Rankin category (0–1 vs ≥ 2) 57 9/16/32 Chi2 4.63 NA 0.108 0.169 Comorbidities category (< vs ≥ 2) 59 9/18/32 Chi2 6.26 NA 0.046 0.093 Pre-stroke mobility limitation (yes vs no)d 59 9/18/32 Chi2 13.50 NA 0.002 0.006 Type of neighborhood (rural vs suburban vs urban) 59 9/18/32 Chi2 6.89 NA 0.148 0.180 Availability of a car for personal use (yes vs no) 59 9/18/32 Chi2 6.01 NA 0.051 0.093 FES-I score 59 9/18/32 ANOVA 13.23 2, 56 28.1 (24.9, 31.3) 21.1 (18.8, 23.3) 18.8 (17.2, 20.5)  < 0.001  < 0.001 Log-transformed TUG time [s] 58 9/17/32 ANOVA 11.36 2, 55 2.6 (2.4, 2.8) 2.3 (2.2, 2.5) 2.1 (2.0, 2.2)  < 0.001  < 0.001 LSA starting (T0) score 59 9/18/32 ANOVA 26.19 2, 56 32.1 (18.7, 45.4) 61.2 (51.8, 70.6) 84.3 (77.2, 91.3)  < 0.001  < 0.001 NIHSS National Institutes of Health Stroke Scale, FES-I Falls Efficacy Scale-International Version, TUG timed up-and-go, ANOVA analysis of variance, NA not applicable aBenjamini–Hochberg approach to control the false-discovery rate bp-values ≤ 0.05 are bolded cMedian age was 74 years dAt least some difficulty walking 2 km or climbing stairs vs no difficulty Discussion In this prospective observational study of 59 patients, we found evidence that stroke severity, the presence of 2 or more comorbidities, pre-stroke mobility limitation, and falls efficacy affected the course of life-space mobility within the first year after stroke. Analyses of typical trajectories revealed three classes which can be described as “low stable”, “average stable”, and “high increasing”. Classes differed with regard to their starting LSA, pre-stroke mobility limitation, falls efficacy, and lower extremity physical function with a higher mean starting LSA, lower prevalence of pre-stroke mobility limitation, higher mean falls efficacy, and better mean lower extremity physical function in the “high increasing” class. Median LSA values of our sample ranged between 72 (IQR 55.75–92) at 3 months and 88 (IQR 49.5–100) at 12 months after stroke, respectively; these values were markedly higher than those reported by Tsunoda et al. in patients at a median time after stroke of 75 (IQR 19–120) months (median LSA 48.0; IQR 36.0–67.5) [17]. Despite some similarities between the two samples (our sample vs Tsunoda et al.), including a comparable age and sex distribution and a high median level of functional independence of both samples, comparability is limited by discrepancies in inclusion criteria (our study: ability to walk for 20 m vs Tsunoda et al.: ability to walk for 5 m) and in starting point and length of the time periods studied (our study: first year after event vs Tsunoda et al.: a period of 2 years starting at a median of 75 months after stroke). The much earlier starting point and the relatively high burden (four assessments) within the first year after stroke might have led to a selection of healthier and fitter participants in our study; indicated by a markedly better lower extremity function in our sample (Tsunoda et al.: median comfortable walking speed of 0.66 m/s vs. our sample: median TUG time of 9.0 s). Median LSA values of our sample are comparable to values found in population-based studies in community-dwelling older adults with median scores typically between 55 and 75 [10–12, 41]. While in our sample, there was no evidence of a change in LSA over time in the multivariate analyses, Tsunoda et al. reported a significant decline within the 2-years follow-up period [17]. We additionally conducted an exploratory LCGA, indicating that there was 1 group of participants (class 3; “high increasing”; about half of the sample) with a high LSA starting value who seemed to be able to increase their LSA even further in the first year. The other groups with lower LSA starting values remained stable. The only existing study [17]—to the best of our knowledge—on factors associated with longitudinal changes in LSA in patients after stroke identified comfortable gait speed and age as independent factors in multiple LMMs. The authors found no evidence for effects of sex, time after event, type of stroke (ischemic vs hemorrhagic), presence of diabetes, functional independence, or cognition. Positive relationships between lower extremity physical function and life-space mobility are well-documented in the general population [42, 43] and have also been demonstrated by a number of cross-sectional studies in post-stroke patients. As an example, a cross-sectional study by Tashiro et al. [18] in 46 community-dwelling individuals with a median time of 49.5 months (IQR 32–90.5) post-event showed a significant association between maximum walking speed (m/s) and LSA score in a multiple regression analysis (coefficient β = 12.85; 95% CI 2.46–23.23; p = 0.017). In another cross-sectional study in 112 people after stroke (average time post-event 73.6; SD 57.4 months), lower extremity physical function, assessed by the Five Times Sit-to-Stand Test (lower values indicate better function), correlated negatively with LSA in unadjusted analyses (Spearman correlation coefficient r = − 0.42; p < 0.001) [19]. A longitudinal study aiming to predict LSA scores 2 months after discharge from inpatient rehabilitation after stroke based on parameters assessed at discharge identified a TUG time of < 15 s as being predictive of higher LSA scores (p < 0.0001) [44]. In our study, the multivariable LMMs did not reveal evidence for an effect of lower extremity physical function (measured by TUG) on the course of LSA. However, those belonging to the class (identified by LCGA) with high starting LSA value and increase of LSA over time (class 3; “high increasing”) had better TUG values than their counterparts in the other classes (classes 1 and 2; “low stable” and “average stable”). Previous cross-sectional analyses of our sample at 3 months post-stroke showed that log-transformed TUG time was negatively associated with objective life-space measures assessed by Global Navigation Satellite System (GNSS) over a 1-week period, such as the maximum distance from home and the convex hull area (the smallest convex polygon on a map enclosing all GNSS fixes) [45]. Stroke severity (measured by NIHSS) has repeatedly been shown to be one of the main predictors of functional limitations and disability after stroke [5, 46]. Physical, cognitive as well as perceptual deficits associated with stroke may compromise the patients’ ability and confidence to navigate through their community environment and thereby limit their social participation [47]. Our data suggest that even within a sample of patients with predominantly mild stroke severity (95% with a NIHSS score of ≤ 5), the severity (NIHSS score of 0–1 vs ≥ 2) affected the course of life-space mobility in the first year after stroke—with better LSA scores in those belonging to the NIHSS 0–1 category. The observed gain in LSA in those with higher NIHSS score (≥ 2) (Fig. 2A) may—at least partly—reflect the recovery of their neurological deficits. Our findings also suggest that the presence of comorbidities was associated with the course of LSA after stroke—with higher LSA scores in those with fewer (0–1 vs ≥ 2) comorbidities. To the best of our knowledge, there are no previous reports on the association between comorbidities and life-space mobility in patients after stroke; however, comorbidities have repeatedly been shown to be prognostic of functional recovery, participation in life situations and survival post-stroke [4, 47–50]. The presence of comorbidities in patients after stroke may affect life-space mobility through various biopsychosocial pathways; besides potentially causing additional physical, cognitive or perceptual deficits, they may also contribute to an increased psychological distress [51] as well as physical and mental fatigue [52]. In our study, falls efficacy was positively associated with life-space mobility after stroke. This is in line with previous findings that the FES-I score at discharge from primary rehabilitation predicts the LSA score 2 months after discharge [44]. Our findings are also in line with the abovementioned cross-sectional study by Tashiro et al. [18] in individuals post-stroke, which showed a significant association between FES-I score and LSA in a multiple regression analysis (coefficient β = − 0.303; 95% CI − 0.590 to − 0.015; p = 0.039). In contrast to this study with a median FES-I score of the participants of 43.5 (IQR 34–59), the median FES-I score of our sample was much lower (median 19; IQR 17–22.5), indicating a lower median concern of falling. This illustrates that even slight deteriorations in perceived self-efficacy to perform daily activities without falling may lead to restrictions of life space and social participation in patients after stroke. Our data showed that self-reported mobility limitation in the week before the event was associated with the course of life-space mobility in the first year after the event. When clinicians treat patients after their first stroke, it can be difficult to differentiate between potentially pre-existing limitations and limitations caused by the event itself; especially considering that within the general population aged 75–84, 23% are unable to walk half a mile and 15% are unable to climb stairs [53]. In order to better predict the potential for recovery, it may be helpful to routinely apply a retrospective assessment of pre-existing mobility limitations. It should however be considered that such measures may be affected by recall bias and may therefore not be useful in patients with severe cognitive impairment (who were excluded from participation in our study). Limitations and strengths The limited sample size meant that the selection of covariables for the statistical analyses was not exhaustive. It is therefore possible that other relevant determinants or confounders were overlooked, leading to residual confounding. The inclusion of a relatively high number of independent variables in relation to the rather small sample size increased the chance of missing an existing association; ie, the fact that our study did not find evidence for associations between some of the independent variables and the outcome should be interpreted with great care [54]. Furthermore, the small sample size limits the generalizability of the results. Data collection took place between January 2020 and February 2022, ie, within a time period that was affected by social distancing recommendations due to the COVID-19 pandemic. Hence, the overall level of life-space mobility of our participants, particularly of those with higher age, may have been reduced within this time period [55]. Strengths of the study include the longitudinal design with repeated measurements at clearly defined time points after stroke. Conclusion Routinely assessing pre-stroke mobility limitation, LSA starting value and falls efficacy—in addition to traditional routine parameters such as the NIHSS and comorbidities—may help clinicians to identify patients at risk of a lack of progress in regaining life-space mobility. Falls efficacy can potentially be modified and improved through targeted rehabilitative measures. Funding Open access funding provided by University of Basel. MOBITEC-Stroke (“Recovery of mobility function and life-space mobility after ischemic stroke”; PI/Co-PI: Timo Hinrichs/Nils Peters) was funded by the Swiss National Science Foundation (SNSF) from August 2019 to July 2022 (Project No. 182681; https://p3.snf.ch/Project-182681). The SNSF had no role in in study design, in the collection, analysis or interpretation of data, in the writing of the report or in the decision to submit the article for publication. Data availability The data that support the findings of this study are available from the corresponding author, TH, upon reasonable request. Declarations Conflicts of interest The authors declare that there are no conflicts of interest. Ethical standard statement This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Ethics Committee of Northwestern and Central Switzerland (Reg.-No. 2019-00989). ==== Refs References 1. Feigin VL Stark BA Johnson CO Roth GA Bisignano C Abady GG Abbasifard M Abbasi-Kangevari M Abd-Allah F Abedi V Global, regional, and national burden of stroke and its risk factors, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019 Lancet Neurol 2021 20 795 820 10.1016/S1474-4422(21)00252-0 34487721 2. Le Couteur DG Thillainadesan J What is an aging-related disease? An epidemiological perspective J Gerontol A Biol Sci Med Sci 2022 77 2168 2174 10.1093/gerona/glac039 35167685 3. Spiers NA Matthews RJ Jagger C Matthews FE Boult C Robinson TG Brayne C Diseases and impairments as risk factors for onset of disability in the older population in England and Wales: findings from the medical research council cognitive function and ageing study J Gerontol A Biol Sci Med Sci 2005 60 248 254 10.1093/gerona/60.2.248 15814870 4. Hankey GJ Long-term outcome after ischaemic stroke/transient ischaemic attack Cerebrovasc Dis 2003 16 14 19 10.1159/000069936 12698014 5. Lee EY Sohn MK Lee JM Kim DY Shin YI Oh GJ Lee YS Lee SY Song MK Han JH Changes in long-term functional independence in patients with moderate and severe ischemic stroke: comparison of the responsiveness of the Modified Barthel Index and the Functional Independence Measure Int J Environ Res Public Health 2022 19 9612 10.3390/ijerph19159612 35954971 6. Skoglund E Westerlind E Persson HC Sunnerhagen KS Self-perceived impact of stroke: a longitudinal comparison between one and five years post-stroke J Rehabil Med 2019 51 660 664 10.2340/16501977-2595 31478056 7. Baker PS Bodner EV Allman RM Measuring life-space mobility in community-dwelling older adults J Am Geriatr Soc 2003 51 1610 1614 10.1046/j.1532-5415.2003.51512.x 14687391 8. Saajanaho M Rantakokko M Portegijs E Tormakangas T Eronen J Tsai LT Jylha M Rantanen T Personal goals and changes in life-space mobility among older people Prev Med 2015 81 163 167 10.1016/j.ypmed.2015.08.015 26348450 9. Tsuji T Rantakokko M Portegijs E Viljanen A Rantanen T The effect of body mass index, lower extremity performance, and use of a private car on incident life-space restriction: a two-year follow-up study BMC Geriatr 2018 18 271 10.1186/s12877-018-0956-3 30409120 10. Portegijs E Rantakokko M Viljanen A Sipilä S Rantanen T Identification of older people at risk of ADL disability using the Life-Space Assessment: a longitudinal cohort study J Am Med Dir Assoc 2016 17 410 414 10.1016/j.jamda.2015.12.010 26805752 11. Sheppard KD Sawyer P Ritchie CS Allman RM Brown CJ Life-space mobility predicts nursing home admission over 6 years J Aging Health 2013 25 907 920 10.1177/0898264313497507 23965310 12. Kennedy RE Williams CP Sawyer P Lo AX Connelly K Nassel A Brown CJ Life-space predicts health care utilization in community-dwelling older adults J Aging Health 2019 31 280 292 10.1177/0898264317730487 29254407 13. Boyle PA Buchman AS Barnes LL James BD Bennett DA Association between life space and risk of mortality in advanced age J Am Geriatr Soc 2010 58 1925 1930 10.1111/j.1532-5415.2010.03058.x 20831722 14. Rantakokko M Portegijs E Viljanen A Iwarsson S Rantanen T Life-space mobility and quality of life in community-dwelling older people J Am Geriatr Soc 2013 61 1830 1832 10.1111/jgs.12473 24117303 15. Rantakokko M Portegijs E Viljanen A Iwarsson S Kauppinen M Rantanen T Changes in life-space mobility and quality of life among community-dwelling older people: a 2-year follow-up study Qual Life Res 2016 25 1189 1197 10.1007/s11136-015-1137-x 26407605 16. Taylor JK Buchan IE van der Veer SN Assessing life-space mobility for a more holistic view on wellbeing in geriatric research and clinical practice Aging Clin Exp Res 2019 31 439 445 10.1007/s40520-018-0999-5 30078096 17. Tsunoda S Shimizu S Suzuki Y Tsunoda A Yamada R Shimose R Kawabata M Ogura M Matsunaga A Longitudinal changes in life-space mobility and the factors influencing it among chronic community-dwelling post-stroke patients Disabil Rehabil 2022 44 7872 7876 10.1080/09638288.2021.2001054 34894964 18. Tashiro H Isho T Takeda T Nakamura T Kozuka N Hoshi F Life-space mobility and relevant factors in community-dwelling individuals with stroke in Japan: a cross-sectional study Prog Rehabil Med 2019 4 20190014 10.2490/prm.20190014 32789261 19. Ho LYW Lai CKY Ng SSM Psychometric properties testing of a Cantonese version of the Life-Space Assessment in people with stroke Sci Rep 2021 11 20614 10.1038/s41598-021-00140-w 34663852 20. Yang YN Kim BR Uhm KE Kim SJ Lee S Oh-Park M Lee J Life-space assessment in stroke patients Ann Rehabil Med 2017 41 761 768 10.5535/arm.2017.41.5.761 29201814 21. Rössler R Bridenbaugh SA Engelter ST Weibel R Infanger D Giannouli E Sofios A Iendra L Portegijs E Rantanen T Recovery of mobility function and life-space mobility after ischemic stroke: the MOBITEC-Stroke study protocol BMC Neurol 2020 20 348 10.1186/s12883-020-01920-z 32938425 22. Nasreddine ZS Phillips NA Bedirian V Charbonneau S Whitehead V Collin I Cummings JL Chertkow H The Montreal Cognitive Assessment, MoCA: a brief screening tool for mild cognitive impairment J Am Geriatr Soc 2005 53 695 699 10.1111/j.1532-5415.2005.53221.x 15817019 23. Hinrichs T Bücker B Klaassen-Mielke R Brach M Wilm S Platen P Mai A Home-based exercise supported by general practitioner practices: ineffective in a sample of chronically ill and mobility-limited older adults (the HOMEfit randomized controlled trial) J Am Geriatr Soc 2016 64 2270 2279 10.1111/jgs.14392 27676362 24. Phillips J Dal Grande E Ritchie C Abernethy AP Currow DC A population-based cross-sectional study that defined normative population data for the Life-Space Mobility Assessment-Composite Score J Pain Symptom Manage 2015 49 885 893 10.1016/j.jpainsymman.2014.09.010 25546285 25. Kennedy RE Almutairi M Williams CP Sawyer P Allman RM Brown CJ Determination of the minimal important change in the Life-Space Assessment J Am Geriatr Soc 2019 67 565 569 10.1111/jgs.15707 30536982 26. Peel C Sawyer Baker P Roth DL Brown CJ Brodner EV Allman RM Assessing mobility in older adults: the UAB Study of Aging Life-Space Assessment Phys Ther 2005 85 1008 1119 10.1093/ptj/85.10.1008 16180950 27. Ullrich P Werner C Abel B Hummel M Bauer JM Hauer K Assessing life-space mobility: a systematic review of questionnaires and their psychometric properties Z Gerontol Geriatr 2022 55 660 666 10.1007/s00391-022-02035-5 35244765 28. Brott T Adams HP Jr Olinger CP Marler JR Barsan WG Biller J Spilker J Holleran R Eberle R Hertzberg V Measurements of acute cerebral infarction: a clinical examination scale Stroke 1989 20 864 870 10.1161/01.STR.20.7.864 2749846 29. Sangha O Stucki G Liang MH Fossel AH Katz JN The Self-Administered Comorbidity Questionnaire: a new method to assess comorbidity for clinical and health services research Arthritis Rheum 2003 49 156 163 10.1002/art.10993 12687505 30. Rantanen T Portegijs E Viljanen A Eronen J Saajanaho M Tsai LT Kauppinen M Palonen EM Sipila S Iwarsson S Individual and environmental factors underlying life space of older people—study protocol and design of a cohort study on life-space mobility in old age (LISPE) BMC Public Health 2012 12 1018 10.1186/1471-2458-12-1018 23170987 31. Dias N Kempen G Todd CJ Beyer N Freiberger E Piot-Ziegler C Yardley L Hauer K The German version of the Falls Efficacy Scale-International Version (FES-I) Z Gerontol Geriatr 2006 39 297 300 10.1007/s00391-006-0400-8 16900450 32. Rikli RE Jones CJ Development and validation of a functional fitness test for community-residing older adults J Aging Phys Act 1999 7 129 161 10.1123/japa.7.2.129 33. Podsiadlo D Richardson S The timed “Up & Go”: a test of basic functional mobility for frail elderly persons J Am Geriatr Soc 1991 39 142 148 10.1111/j.1532-5415.1991.tb01616.x 1991946 34. van Buuren S Flexible imputation of missing data 2018 New York Chapman and Hall/CRC 35. Chan KW, Meng XL (2021) Multiple improvements of multiple imputation likelihood ratio tests. arXiv:1711.08822 36. Nelder J A reformulation of linear models J R Stat Soc Ser A 1977 140 48 63 10.2307/2344517 37. Nguena Nguefack HL Page MG Katz J Choiniere M Vanasse A Dorais M Samb OM Lacasse A Trajectory modelling techniques useful to epidemiological research: a comparative narrative review of approaches Clin Epidemiol 2020 12 1205 1222 10.2147/CLEP.S265287 33154677 38. Proust-Lima C Philipps V Liquet B Estimation of extended mixed models using latent classes and latent processes: the R package lcmm J Stat Softw 2017 78 1 56 10.18637/jss.v078.i02 39. van der Nest G Passos VL Candel MJJM van Breukelen GJP An overview of mixture modelling for latent evolutions in longitudinal data: modelling approaches, fit statistics and software Adv Life Course Res 2020 43 100323 10.1016/j.alcr.2019.100323 36726256 40. Benjamini Y Hochberg Y Controlling the false discovery rate—a practical and powerful approach to multiple testing J R Stat Soc B 1995 57 289 300 41. Fristedt S Kammerlind AS Fransson EI Bravell ME Physical functioning associated with life-space mobility in later life among men and women BMC Geriatr 2022 22 364 10.1186/s12877-022-03065-9 35473475 42. Kuspinar A Verschoor CP Beauchamp MK Dushoff J Ma J Amster E Bassim C Dal Bello-Haas V Gregory MA Harris JE Modifiable factors related to life-space mobility in community-dwelling older adults: results from the Canadian Longitudinal Study on Aging BMC Geriatr 2020 20 35 10.1186/s12877-020-1431-5 32005107 43. Johnson J Rodriguez MA Al Snih S Life-space mobility in the elderly: current perspectives Clin Interv Aging 2020 15 1665 1674 10.2147/CIA.S196944 32982200 44. Nakao M Izumi S Yokoshima Y Matsuba Y Maeno Y Prediction of life-space mobility in patients with stroke 2 months after discharge from rehabilitation: a retrospective cohort study Disabil Rehabil 2020 42 2035 2042 10.1080/09638288.2018.1550533 30676134 45. Rössler R Rommers N Kim E-K Iendra L Sofios A Giannouli E Portegijs E Rantanen T Infanger D Bridenbaugh S Timed up-and-go performance is associated with objectively measured life-space in patients 3 months after ischemic stroke: a cross-sectional observational study J Neurol 2022 270 1999 2009 10.1007/s00415-022-11524-x 36547716 46. Meyer MJ Pereira S McClure A Teasell R Thind A Koval J Richardson M Speechley M A systematic review of studies reporting multivariable models to predict functional outcomes after post-stroke inpatient rehabilitation Disabil Rehabil 2015 37 1316 1323 10.3109/09638288.2014.963706 25250807 47. Kim M Cho K Lee W Community walking training program improves walking function and social participation in chronic stroke patients Tohoku J Exp Med 2014 234 281 286 10.1620/tjem.234.281 25483170 48. Barow E Probst AC Pinnschmidt H Heinze M Jensen M Rimmele DL Flottmann F Broocks G Fiehler J Gerloff C Effect of comorbidity burden and polypharmacy on poor functional outcome in acute ischemic stroke Clin Neuroradiol 2022 33 147 154 10.1007/s00062-022-01193-8 35831611 49. Goldstein LB Samsa GP Matchar DB Horner RD Charlson Index comorbidity adjustment for ischemic stroke outcome studies Stroke 2004 35 1941 1945 10.1161/01.STR.0000135225.80898.1c 15232123 50. Ezekiel L Collett J Mayo NE Pang L Field L Dawes H Factors associated with participation in life situations for adults with stroke: a systematic review Arch Phys Med Rehabil 2019 100 945 955 10.1016/j.apmr.2018.06.017 29981316 51. Fortin M Bravo G Hudon C Lapointe L Dubois MF Almirall J Psychological distress and multimorbidity in primary care Ann Fam Med 2006 4 417 422 10.1370/afm.528 17003141 52. Hardy SE Studenski SA Qualities of fatigue and associated chronic conditions among older adults J Pain Symptom Manage 2010 39 1033 1042 10.1016/j.jpainsymman.2009.09.026 20538185 53. Jette AM Branch LG The Framingham Disability Study. 2. Physical disability among the aging Am J Public Health 1981 71 1211 1216 10.2105/AJPH.71.11.1211 7294262 54. Altman DG Bland JM Absence of evidence is not evidence of absence BMJ 1995 311 485 10.1136/bmj.311.7003.485 7647644 55. Rantanen T Eronen J Kauppinen M Kokko K Sanaslahti S Kajan N Portegijs E Life-space mobility and active aging as factors underlying quality of life among older people before and during COVID-19 lockdown in Finland—a longitudinal study J Gerontol A Biol Sci Med Sci 2021 76 e60 e67 10.1093/gerona/glaa274 33125043
PMC010xxxxxx/PMC10159678.txt
==== Front BMC Health Serv Res BMC Health Serv Res BMC Health Services Research 1472-6963 BioMed Central London 37143071 9412 10.1186/s12913-023-09412-9 Research Evaluation of the implementation progress through key performance indicators in a new multimorbidity patient-centered care model in Chile Varela Teresita 1 Zamorano Paula [email protected] 123 Muñoz Paulina 1 Rain Carolina 4 Irazoqui Esteban 1 Sapag Jaime C. 456 Tellez Alvaro 14 1 grid.7870.8 0000 0001 2157 0406 Centro de Innovación en Salud ANCORA UC, Facultad de Medicina, Pontificia Universidad Católica de Chile, Santiago, Chile 2 grid.7870.8 0000 0001 2157 0406 Health Technology Assessment Unit, Center of Clinical Research, Pontificia Universidad Católica de Chile, Santiago, Chile 3 Current Address: Diagonal Paraguay, Santiago, 362 Chile 4 grid.7870.8 0000 0001 2157 0406 Department of Family Medicine, Pontificia Universidad Católica de Chile, Santiago, Chile 5 grid.7870.8 0000 0001 2157 0406 Department of Public health, Pontificia Universidad Católica de Chile, Santiago, Chile 6 grid.17063.33 0000 0001 2157 2938 Dalla Lana School of Public Health, University of Toronto, Toronto, Canada 4 5 2023 4 5 2023 2023 23 4393 11 2022 17 4 2023 © The Author(s) 2023, corrected publication 2023 https://creativecommons.org/licenses/by/4.0/ Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Background Complex health interventions involve deep organizational, structural, and cultural changes that challenge health teams and decision-makers. The explosion of chronic diseases has made the multimorbidity approach a global priority. The Centro de Innovación en Salud ANCORA UC implemented a Multimorbidity Patient-Centered Care Model in the Chilean public health system. Objective This study aims to evaluate the progress of the implementation of the Multimorbidity Patient-Centered Care Model in seven primary care centers through key performance indicators. Methods a set of indicators was designed to evaluate change management, operations, installation of new roles, and services and activities of the intervention strategy of the model. Key performance indicators were identified to monitor the implementation progress on minimal components for the model’s sustainability. Each item was assigned against an expected minimum score of 67% of progress from the overall score. They were monitored twice in seven primary health centers in 2019 and 2020, which intervened 22,642 patients with the intervention. Results The results showed that six of the seven primary care centers reached the minimum implementation threshold. The main advances were in operational conditions, and those with minor progress in implementation were the clinical services. Population size, organization, coordination of the health care teams, additional training, and decision-makers support were key factors that determined the degree of progress in a complex intervention. Conclusion It was possible to measure the progression of the implementation of a complex intervention through key performance indicators delivering relevant information for decision-makers that pursue a successful and faithful implementation. This study provides a valuable tool for the national scale-up of a similar model started in Chile by the Ministry of Health and other countries. Supplementary Information The online version contains supplementary material available at 10.1186/s12913-023-09412-9. Keywords Multimorbidity Key performance indicators Implementation Progress Patient-centered Model issue-copyright-statement© BioMed Central Ltd., part of Springer Nature 2023 ==== Body pmcIntroduction Complex changes in health represent a real challenge for health systems, clinical teams, and individuals not only because of their inherent complexity but also in terms of ensuring sustainability over time. For example, in recent years, important epidemiological changes have modified the burden of disease, health services use, and life expectancy, making the reorganization of health services a priority [1]. Therefore, complex changes require core elements that allow change to be executed and its sustainability, such as changes in associated resources, adequate competencies, clear leadership, and culture and behaviors that support change [2]. Methodologies and frameworks for implementing complex changes in health described usually include stages of theory exploration, development of clinical intervention, identifying core aspects, as well as feasibility and implementation studies [3]. The evaluation is often focused on intermediate or final outcomes, but they lack performance indicators that can deliver valuable information from the implementation process. Even more, challenges of the complex interventions are born within the implementation process, from operational and leading changes barriers that need to be appropriately addressed to pursue a sustainable incorporation [4]. Measurements during complex interventions are core in monitoring the degree of implementation progress of the proposed intervention. A healthcare Key Performance Indicator (KPI) is a clear-cut measure used to observe, monitor, optimize, manage, and transform the performance of a healthcare process to ensure effectiveness, quality, and efficiency and increase patient satisfaction and healthcare providers [5, 6]. Therefore, their use in complex interventions could provide a broader perspective with quantitative and qualitative information that can help decision-makers during the process. For example, when a complex intervention involves organizational, operational, and cultural changes, KPIs can monitor/track progress and make objective comparisons between different contexts enabling opportune response to those who are experiencing a harder process [7]. However, according to the quality and amount of health services data, it is key to choose the indicators that can be fulfilled simply and provide relevant information to the process and progress of the implementation of a complex intervention. There are experiences in Chile and internationally that show how complex the profound changes to the organization and delivery of health services are. In the field of multimorbidity defined as two or more chronic conditions in the same person [8], the Chilean public health system and its primary care centers are organized in the traditional single diagnostic approach. They are offering fragmented disintegrated, and inefficient care, which has shown negative results in 11 million people (70% of the national population) living with chronic disease outcomes during the last years [9, 10]. Therefore, the Centro de Innovacion en Salud ANCORA UC (CISAUC), together with the Servicio Metropolitano Sur Oriente (SSMSO) and National Found of Health (FONASA), implemented a complex change in health. The objective was to change health services organized in diagnosis towards a patient centered care organized according to each patient multimorbidity risk. The Multimorbidity Patient-Centered Care Model (MPCM) enhances the family and community health model implemented in the primary health care centers (PHC) of the country [11] and adds core elements such as case management, risk stratification, and multimorbidity as shown in Fig. 1. The intervention strategies were designed and offered in primary and tertiary care centers, according to each person’s risk. Fig. 1 Multimorbidity Patient-Centered Care Model (MPCM) [12] The implementation process had three stages: preparation, implementation, and evaluation activities. During the preparation, process activities were carried out to disseminate and communicate the model, together with training of health teams and operational preparation. In the implementation, clinical activities corresponding to the intervention strategy were executed (Fig. 2), and the CISAUC expert team monitored each center’s implementation’s particularities and execution times. In the evaluation stage, an impact analysis on the use of health services and an evaluation of patient and health team satisfaction were carried out showing positive results [12]. Still, measuring the impact on avoidable hospitalization would have complemented those results. Similar interventions have shown a decrease in unplanned hospitalizations [13]. The MPCM intervention decreased the total number of hospitalizations, and we could infer that those results are related to the decrease in avoidable hospitalization. But at the time of evaluation, there was a lack of consensus about the kind or list of avoidable hospitalizations, limiting the data extraction and evaluation. Given that there were multiple barriers and facilitators that influence the progress of implementation and its sustainability over time, this study aimed to evaluate the progress in implementing the Multimorbidity patient-center care model in seven Primary Care Centers in Chile. Fig. 2 Intervention strategy of the MPCM Adapted from Care Strategy for patients with chronic diseases in the community of Madrid, November 2013, Ministry of Health, Social Services and Equality [14]. The study’s objective was to evaluate the progress of implementing the Multimorbidity Patient-Centered Care Model in seven primary care centers through key performance indicators. Methodology The study used a quantitative approach to assess the progress of implementing the MPCM in seven primary health care centers in the southeast of Santiago, Chile, that intervened in 22,642 adult patients with multimorbidity. The PHCs are organized by the Family and community model [15]. Their size ranged from 3 to 4 multidisciplinary health teams to offer care from 22,000 to 35,000 patients (covered population) with vulnerable conditions. The intervention strategy shown above (Fig. 2) had several components from which developed a set of indicators in four main areas: change management, operational items, new roles, and services and activities. In addition, some KPIs were identified from the overall set of indicators to reflect the minimum conditions required for the intervention sustainability. Figure 3 shows the process of the setup and monitoring of KPI. Fig. 3 Key performance indicators setup, scoring and monitoring process Indicators Assignment Areas Four areas were considered for grouping the KPIs according to the complex interventions challenges [4] and the intervention strategy main characteristics, as shown in Fig. 2. In the change management area, the organization of local governance to plan, lead and coordinate the actions necessary to achieve change is required to activate a gradual, strategic, and responsible process. Therefore, the objective of this intervention strategy area items was to activate local teams, perform constant communication and dissemination activities, and deliver the necessary training for health teams. To achieve the minimum implementation of this section, the center must have managerial support, internal leadership for the installation of the model, and a local induction plan for the strategy for new employees. The measurement of these last three corresponds to the KPIs. In the operational area, it is necessary to perform modifications to the structure and health services delivery organization to allow the installation of the new care model. The objective was to assess the incorporation of multimorbidity stratification, changes in the protocols for electronic clinical records (ECR), and health services delivery according to each patient’s complexity. To achieve the minimum implementation of this section, the center must have the adult population stratification, unified drug prescriptions, alerts for consultations in the emergency service, and hospitalization activated and modify the acts on the agenda toward comprehensive care. Incorporating new roles is expected to provide new activities of the intervention strategy, such as Case Manager, Transition Nurse, Clinical Pharmaceutical Chemist, and High-Risk Family Physicians. The objective was to measure the degree of implementation of the new roles proposed to guarantee the execution of the new clinical services and improve continuity of care and patient follow-up efficiency. To achieve the minimum implementation of this section, the center must have the new roles installed. In the activities and services area, the differentiation of health care delivery by multimorbidity risk is a core aspect of the new care model and reflects the transition from a single diagnostic to a person-centered approach. The objective was to evaluate the core activities and services that would be the foundation for the sustainability of the change in healthcare delivery. To achieve the minimum implementation of this section, the center must have included the implementation of agreed plans, telephone counseling, continuity of care with a professional from the team, rescue after hospital discharge, implementation of an induction plan, and transition care. Key performance Identification The objective was to identify components of the intervention strategy that were core for the change towards a multimorbidity approach and the implementation success. They were chosen based on the minimum conditions required for the intervention sustainability, on the representation of the implementation progress, on the availability of measurement information (either because it was available or because it was simple to download by the health team), and on accessible and sustainable monitoring over time. From a total of 32 components, 17 were identified and assigned key performance indicators to track their implementation progress (Table 1). Table 1 Indicators area and components Area Components Change Management Decision makers support (PHC director and managers) Leader for the implementation of the MPCM at the PHC Local training plan of MPCM for new employees Operational Adult population stratified by risk, available and with patients ID Unified drug prescription Alert system informing PHC teams of patients consulting at emergency room and hospitalization Integrated multimorbidity scheduled appointments New Roles Clinical Pharmacist High-Risk primary physician Case Manager Transition Nurse Activities and services Individualized Care Plans Phone counseling Continuity of care with a professional from the team Rescue after hospital discharge Implementation of an induction plan Transition care PHC: Primary healthcare center MPCM: Multimorbidity Patient-Centered Care Model ID: identification number Complementary indicators identification In addition to the KPIs, we developed another 15 indicators where the intervention strategy components open evaluation in greater depth if necessary. In the present study, we only evaluated the KPIs. The performance indicators for the MPCM are available in the supplementary material. Setup, measurement, and score assignment of KPI The monitorization of the KPIs was self-reported, with dichotomous responses, and was completed by the implementation health care teams composed by clinicians such as nurses, physicians, nutritionist and physiotherapist. The setting-up, measurement, and scoring of KPI were provided by the study’s researchers and the expert team of CISAUC. The KPIs were designed according to each area and component. And a score was defined according to the level of complexity and relevance where the component was tracked. For scoring, the individual scores of the KPI of each area were summed (example in change management score of 3) and divided by the maximum expected (score 28) to obtain a percentage of progress for each area (example: (3/28) * 100 = 10%). Finally, an average between areas was calculated for an overall percentage score. Table 2 represents the four groups of KPI scoring. The full description of scoring and measurement for each KPI is in the supplementary material. Table 2 Key performance indicators areas Area Score (min-max) Percentage of the total score (= group score/total score) (min-max) Change Management 0–3 0–10% Operational 0–9 0–29% Incorporation of New Roles 0–4 0–13% Activities and services 0–15 0–48%​ Total Score 0–31 0-100% Threshold and minimum implementation period An overall threshold of 67% was defined with a group of experts and local teams to determine the minimal expected progress after 12 months of implementation in activities that are core to reflect the change. The implementation of MPCM represents a complex change, and the implementation of the complete intervention strategy is expected to be longer than the piloting period. Therefore, defining a minimal implementation period and a minimal percentage of implementation progress was relevant. Review with the primary care team The KPIs’ setup, measurement, scoring, and pertinence were reviewed and discussed with the healthcare teams of the seven PHCs. Then a new draft was produced and checked for a second time to proceed and consolidate a final draft. The objective was to evaluate (i) the assertiveness of the KPIs with the minimal required conditions, (ii) the monitoring feasibility, and (iii)the understanding of a variety of healthcare professionals. Finally, the CISAUC team collected the information and made the necessary adjustments to the components and the indicators. This process was done twice, first after the indicators’ preliminary draft (December 2019). The second consisted in adapting the indicators to the global and national context of the COVID-19 pandemic (November 2020). Monitoring KPI The seven PHCs had 30 days to monitor, collect the necessary information and fulfill the information of the indicators. This process was carried out in September 2020. During this period, the process was conducted by a local health care professional in charge of implementing the MPCM and supported by the CISAUC team. In addition, a document was prepared and delivered to the teams to facilitate the monitoring, collection, and completion of the information required and standardize the process. The data with the results was collected and analyzed by the CISAUC expert team from each PHC that implemented the MPCM. Results The intervened PHC were located at the southeast of the capital of Chile, Santiago and implemented the MPCM between 2017 and 2020. The population covered ranged from 17,487 to 35,240 patients. Three of the PHC were located at the municipality of La Pintana, two in La Florida and two in Puente Alto. The number of local care teams ranged from two to six for each PHC [12] (local team integrated by physician, midwives, nutritionists, physical therapists, psychologists, social workers, dentists, nurses and paramedic technician). The overall results on the seven PHCs on 2020 showed positive implementation progress of the MPCM. The average total score was of 22 out of a maximum of 31. The overall threshold was met with a score of 72% (min 45% - max 100%) (Table 3). Table 3 Results of the KPIs by area and component for each municipality and PHC. Area Component Score Municipality 1 Municipality 2 Municipality 3 PHC 1 PHC 2 PHC3 PHC4 PHC5 PHC 6 PHC 7 Change Management Decision makers support (PHC director and managers) 1 1 1 1 1 1 1 1 Leader for the implementation of the MPCM at the PHC 1 1 1 1 1 1 1 1 Local training plan of MPCM for new employees 1 1 0 1 1 0 1 1 Change Management Total Score 3 3 2 3 3 2 3 3 Operational Adult population stratified by risk, available and with patients ID 3 3 3 3 3 3 3 1 Unified drug prescription 3 3 3 3 3 3 1 3 Alert system informing PHC teams of patients consulting at emergency room and hospitalization 1 1 1 1 1 1 1 1 Integrated multimorbidity scheduled appointments 2 2 1 2 2 0 2 2 Operational Conditions Total Score 9 9 8 9 9 7 7 7 New Roles Clinical Pharmacist 1 1 1 1 1 1 0 0 High-complexity primary physician 1 1 1 1 1 1 1 1 Case Manager 1 1 1 1 1 1 1 1 Transition Nurse 1 1 1 1 1 1 1 1 New Roles total Score 4 4 4 4 4 4 3 3 Activities and services Individualized Care Plans 3 3 1 3 3 0 3 2 Phone counseling 3 3 0 2 3 0 0 2 Continuity of care with a professional from the team 3 0 0 0 3 0 0 2 Rescue after hospital discharge 1 0 0 0 1 0 0 1 Implementation of an induction plan 3 3 0 3 3 0 3 0 Transition care 2 2 2 1 2 1 2 2 Activities and services total Score 15 11 3 9 15 1 8 9 Total Score 31 27 17 25 31 14 21 22 Implementation progress percentage 100% 87% 55% 81% 100% 45% 68% 71% PHC: Primary healthcare center MPCM: Multimorbidity Patient-Centered Care Model ID: identification number *The indicators description is available in supplementary material. The municipalities that implemented the MPCM offered health services for similar populations showing differences in between. In Municipality 1, one of the PHCs obtained the highest level of implementation. On the contrary, the other PHC didn´t reach the minimum implementation threshold, scoring a 55% of implementation progress and lower results in activities and services. In Municipality 2 had similar results, where two of the three PHCs scored 81% and 100% on implementation progress, with high scores in components in the four areas. The third PHC didn’t reach the threshold and scored 45% in the implementation progress. Finally, in Municipality 3, both PHCs reached the threshold with scores of 68% and 71%. Regarding the areas of evaluation, the highest scores were in change management and new roles. The lowest score was in services and activities. This is where the indicators reflect substantial changes in the real practice and the execution of the components of the intervention strategy of the MPCM. The results by each component study showed that there were six that scored the highest: Decision makers support (PHC director and managers), Leaders for the Implementation of the MPCM at the PHC, Alert System Informing PHC Teams of Patients Consulting at the Emergency Room and Hospitalization, High-complexity Primary Physician, Case Manager, and Transition Nurse. In contrast, the components that obtained the lowest scores across all centers were: Continuity of Care and Rescue of High-risk Patients After Discharge. Finally, regarding the review process with primary care teams, adjustments were provided in components mainly deciding if they were a “minimum or not” for the sustainability of the model. For example, in change management area the induction plan was a complementary indicator and after the review it was assessed as a KPI. In operational items, the Integrated multimorbidity scheduled appointments indicator was modified from a percentage of change to a dichotomic answer in yes/no. New roles had no modifications. Services and activities the Implementation of an induction plan and transition care were identified as KPI instead of complementary. Discussion The results of the study showed that the MPCM intervention strategy can be monitored by the health care teams in terms of implementation progress through key performance indicators. Of the seven pilot centers, five (71,4%) reached the expected threshold reflecting the presence of the minimum intervention strategies required for the sustainability of the MPCM. Only two (28,6%) didn’t meet the threshold, demanding further attention to improve quality and performance. The results of this monitoring of the KPIs delivered relevant information for decision-makers and implementation teams to analyze and optimize the implementation progress. Regarding the territory and the PHC where the MPCM pilot was implemented, the centers that did not reach the implementation progress threshold are from different municipalities but have in common the absence of important intervention strategy components. For example, the absence of integrated multimorbidity scheduled appointments refers to health professionals’ schedules by multimorbidity risk instead of by pathology or the program. Also, individualized plans and continuity of care were absent. These three missing components require a deeper and structural change in the organization and operation of the daily routine. Therefore, the barriers within the diagnostic approach for chronic [16–18] diseases are captured by the KPIs monitoring. Thus, strong decision-makers support is needed to authorize and facilitate the transition and sustainable change over the structural organization. The areas of the KPIs also showed a relationship with the implementation process. The differences between the areas of progress may be related the stages of the pilot. In the pre-implementation phase, interventions were carried out first with a focus in a cultural and paradigm change, therefore, change management, operational, and new role changes began executed and obtained the highest scores. In contrast, the services and activities had less score reflecting that structural and operational chances diverse areas of the health services require a longer time [3]. Therefore, the importance to invest time and perform actions to properly install the basis of a further change [2]. Thus, a gradual process should be performed to ensure success in the implementation and sustainability overall as described in other studies. [2] Concerning the areas, Change Management and New Roles reached the highest score. These results could be a consequence of the time invested in the pre-implementation period, where the action of socializing with the health teams, managers, and local leaders was frequent and essential to the change and its urgency. In addition, these changes don’t necessarily involve a structural change in the real context. Therefore, its implementation doesn’t face those barriers that are more difficult to address. Thus, the human resources inserted by the pilot study for the performance of the new roles was a challenge that had a positive acceptance from the health care teams, which probably positively influenced this area’s results [19]. Even more, the national scale-up of a similar intervention by the Ministry of Health included the new roles piloted [20]. The strength of the indicators is that they provide a simple, quantitative, and practical tool to monitor progress in multicomponent and interdisciplinary complex interventions. Methodologies described in the literature for health intervention design and implementation don’t usually include performance indicators or measures from the implementation process itself [3, 21]. Rather, they look for health outcomes [21, 22]. Therefore, complementing both could give health professionals and decision-makers a wider perspective with concrete gaps that certainly facilitate planning opportune quality improvement and addressing gaps in core areas to favor sustainability over time. The limitations of the indicators are that they focus on the primary care components of the intervention strategy. Due to the piloting time, indicators for the performance in secondary and tertiary care were not provided but we included the most relevant network coordination activities performed, indicators such as transition care and rescue after hospital discharge measuring the continuity of care between care levels. Another limitation is that the second measurement was done in the first six months of the pandemic in Chile, which could have affected the results. Also, the validity of the indicators [23], therefore the construction of the KPI was reviewed, discussed twice with health care teams from the pilot centers. Finally, these are self-reported indicators, which could generate bias in their measurement as a proper limitation of the KPIs [5]. Hence, driving to an automatized monitoring could mitigate bias and maintain the strengths of a tool that delivers opportune, concrete, and relevant information for decision makers [24]. Finally, the set of key performance indicators has the potential to reflect the progress in a complex intervention in health like the MPCM, even though in a pandemic context. The automatization and extrapolation to other complex interventions in other groups of patients could provide early useful information to make opportune necessary changes and increase the expected outcomes of the intervention. The setup, monitoring and knowledge performed by the study it is potentially valuable for the similar intervention that it is scaling up the Ministry of Health [20]. Further studies could complement the indicators in the performance of the secondary and tertiary level providing a complete overview of the progress implementation of complex health interventions. Electronic supplementary material Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements The authors would like to thank the Family Health Center (CESFAM) La Florida, La Florida; CESFAM Villa O’Higgins, La Florida; CESFAM Santiago Nueva Extremadura, La Pintana; CESFAM El Roble, La Pintana; ANCORA UC Mother Teresa Calcutta; ANCORA UC San Alberto Hurtado; ANCORA UC Juan Pablo II. La Florida Hospital, Dr. Sotero del Rio Hospital and Padre Hurtado Hospital and also the Department of Public Health, School of Nursing, Pontificia Universidad Católica de Chile and SSMSO and FONASA. Author contributions T.V Researched data and wrote the manuscript. P.Z. Researched data and wrote the manuscript. P.M. Researched data and wrote the manuscript. C.R. reviewed/edited the manuscript. E.I. reviewed/edited the manuscript. J.S. reviewed/edited the manuscript. A.T. researched data and reviewed/edited the manuscript. All authors read and approved the final manuscript. Funding Centro de Innovacion en Salud ANCORA UC. Data Availability All data generated or analyzed during this study are included in this published article and its supplementary information files. Declarations Ethics approval and consent to participate The Ethical approval and informed consent were waived by Pontificia Universidad Católica de Chile ethical committee. ID 190402003: “Centro de Innovación en Salud Ancora UC: una contribución al necesario cambio del sistema de atención en salud”. The study was performed in accordance with guidelines and regulations from Helsinki. Consent for publication Not applicable. Competing interests The authors declare that they have no competing interests List of abbreviations CISAUC Centro de Innovacion en Salud ANCORA UC ECR Electronic clinical records CESFAM Family Health Center KPI Key Performance Indicator MPCM Multimorbidity Patient-Centered Care Model FONASA National Found of Health PHC Primary health care centers SSMSO Servicio Metropolitano Sur Oriente The original version of this article was revised: multiple typesetting errors (missing words) were corrected. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Change history 7/12/2023 A Correction to this paper has been published: 10.1186/s12913-023-09797-7 ==== Refs References 1. Hunter DJ Bengoa R Meeting the challenge of health system transformation in european countries Policy Soc 2022 00 0 1 14 2. Le-Dao H Chauhan A Walpola R Fischer S Schwarz G Minbashian A Managing complex healthcare change: a qualitative exploration of current practice in New South Wales, Australia J Healthc Leadersh 2020 12 143 51 10.2147/JHL.S274958 33328776 3. Skivington K, Matthews L, Simpson SA, Craig P, Baird J, Blazeby JM et al. A new framework for developing and evaluating complex interventions: Update of Medical Research Council guidance. BMJ [Internet]. 2021 Sep 30 [cited 2022 May 29];374. Available from: https://www.bmj.com/content/374/bmj.n2061. 4. Khalil H Kynoch K Implementation of sustainable complex interventions in health care services: the triple C model BMC Health Serv Res 2021 21 1 1 10 10.1186/s12913-021-06115-x 33388053 5. Khalifa M Khalid P Developing strategic health care key performance indicators: a case study on a tertiary care hospital Procedia Comput Sci 2015 63 Icth 459 66 10.1016/j.procs.2015.08.368 6. Health Information and Quality Authority. Heal Inf Qual Auth. 2013;1(1February):70. Guidance on Developing Key Performance Indicators and Minimum Data Sets to Monitor Healthcare Quality. 7. Amer F Hammoud S Khatatbeh H Lohner S Boncz I Endrei D A systematic review: the dimensions to evaluate health care performance and an implication during the pandemic BMC Health Serv Res 2022 22 1 1 22 10.1186/s12913-022-07863-0 34974828 8. Pearson-Stuttard J Ezzati M Gregg EW Multimorbidity—a defining challenge for health systems Lancet Public Heal 2019 4 12 e599 600 10.1016/S2468-2667(19)30222-1 9. Margozzini P Passi Á Encuesta Nacional de Salud ENS 2016–2017: un aporte a la planificación sanitaria y políticas públicas en Chile ARS Med Rev Ciencias Médicas 2018 43 1 30 10.11565/arsmed.v43i1.1354 10. Ministerio de Salud; Gobierno de Chile Encuesta Nac de Salud ENS Chile 2010 15 5 322 11. Martinez MayraSapag, Jaime Z, Paula M, Paulina V, Teresita TA. Contribution of a multimorbidity person-centered care strategy to the Comprehensive Family and Community Healthcare Model in Chile. 2022;782–7. 12. Paula Z Teresita V Alvaro T Manuel E Paulina M Francisco S Impact of a patient-centered care model implemented in public health facilities in Chile: a real world evidence evaluation J Public Heal Epidemiol 2022 14 1 1 9 10.5897/JPHE2021.1371 13. Finlayson K Chang AM Courtney MD Edwards HE Parker AW Hamilton K Transitional care interventions reduce unplanned hospital readmissions in high-risk older adults BMC Health Serv Res 2018 18 1 1 9 10.1186/s12913-018-3771-9 29291745 14. de Madrid M. S. Estrategia de Atención a Pacientes con Enfermedades Crónicas en la Comunidad de Madrid. 2013. 15. Ministerio de Salud; Gobierno de Chile. Orientaciones Para La Implementacion Del Modelo De Atencion Integral De Salud Familiar Y Comunitaria. 2012;1–143. Available from: http://web.minsal.cl/portal/url/item/e7b24eef3e5cb5d1e0400101650128e9.pdf. 16. Orientación Técnica. Programa de Salud Cardiovascular. 2017. 17. DIVAP MINSAL. Manual Operativo Programas de Salud Respiratoria (MINSAL). Man Oper Programas Salud Respir. 2015;51. 18. MINSAL. Plan Nacional De Salud Mental. Plan Nac Salud Ment. 2017;206. 19. Zamorano Pichard P, Tellez A, Mu P, Sapag JC. Effect of COVID-19 pandemic on the implementation of a multimorbidity person- centered care model: A qualitative study from health teams ’ perspective. 2022;1–10. 20. Ministerio de Salud (MINSAL). Estrategia de cuidado integral centrado en las personas para la promoción, prevención y manejo de la cronicidad en contexto de multimorbilidad. Minist Salud. 2021;80. 21. Craig P Dieppe P Macintyre S Mitchie S Nazareth I Petticrew M Developing and evaluating complex interventions: the new Medical Research Council guidance BMJ 2008 337 7676 979 83 22. Campbell NC Murray E Darbyshire J Emery J Farmer A Griffiths F Designing and evaluating complex interventions to improve health care Br Med J 2007 334 7591 455 9 10.1136/bmj.39108.379965.BE 17332585 23. Lima T de Aguiar M Storpirtis PM Development and validation of key performance indicators for medication management services provided for outpatients Res Soc Adm Pharm 2019 15 9 1080 7 10.1016/j.sapharm.2018.09.010 24. Basu S Andrews J Kishore S Panjabi R Stuckler D Comparative performance of private and public healthcare systems in low- and middle-income countries: a systematic review PLoS Med 2012 9 6 19 10.1371/journal.pmed.1001244
PMC010xxxxxx/PMC10160385.txt
==== Front Front Plant Sci Front Plant Sci Front. Plant Sci. Frontiers in Plant Science 1664-462X Frontiers Media S.A. 37151567 10.3389/fpls.2022.1000877 Plant Science Original Research Foliar application of putrescine alleviates terminal drought stress by modulating water status, membrane stability, and yield- related traits in wheat (Triticum aestivum L.) Wasaya Allah 1 2 * Rehman Iqra 1 2 Mohi Ud Din Atta 3 Hayder Bin Khalid Muhammad 3 Ahmad Yasir Tauqeer 2 Mansoor Javaid Muhammad 4 El-Hefnawy Mohamed 5 Brestic Marian 6 * Rahman Md Atikur 7 El Sabagh Ayman 8 9 * 1 Department of Agronomy, Bahauddin Zakariya University Multan, Multan, Pakistan 2 College of Agriculture, University of Layyah, Layyah, Pakistan 3 National Research Center of Intercropping, The Islamia University of Bahawalpur, Multan, Pakistan 4 Department of Agronomy, College of Agriculture, University of Sargodha, Sargodha, Pakistan 5 Department of Chemistry, Rabigh College of Sciences and Arts, King Abdulaziz University, Jeddah, Saudi Arabia 6 Department of Plant Physiology, Slovak University of Agriculture, Nitra, Slovakia 7 Grassland and Forage Division, National Institute of Animal Science, Rural Development Administration, Cheonan, Republic of Korea 8 Department of Agronomy, Faculty of Agriculture, Kafrelsheikh University, Kafr al-Sheik, Egypt 9 Department of Field Crops, Faculty of Agriculture, Siirt University, Siirt, Türkiye Edited by: Vicent Arbona, University of Jaume I, Spain Reviewed by: Sajad Hussain, Islamia University of Bahawalpur, Pakistan; Stefania Toscano, University of Catania, Italy; Muhammad Mehood Iqbal, Central Cotton Research Institute (CCRI), Pakistan *Correspondence: Allah Wasaya, [email protected]; Marian Brestic, [email protected]; Ayman El Sabagh, [email protected] This article was submitted to Plant Abiotic Stress, a section of the journal Frontiers in Plant Science 21 4 2023 2022 13 100087722 7 2022 29 12 2022 Copyright © 2023 Wasaya, Rehman, Mohi Ud Din, Hayder Bin Khalid, Ahmad Yasir, Mansoor Javaid, El-Hefnawy, Brestic, Rahman and El Sabagh 2023 Wasaya, Rehman, Mohi Ud Din, Hayder Bin Khalid, Ahmad Yasir, Mansoor Javaid, El-Hefnawy, Brestic, Rahman and El Sabagh https://creativecommons.org/licenses/by/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Drought stress is one of the major limitations to the growth and yield productivity of cereal crops. It severely impairs the early growing and grain -filling stages of wheat. Therefore, cost- effective and eco-friendly approaches for alleviating drought stress in cereal crops are in high demand. Polyamines, such as putrescine, have a significant effect on improving crop yield under drought- stress conditions. Therefore, the current study was executed with the aim of exploring the significance of putrescine in alleviating drought stress and improving yield- related traits in wheat. Two distinct wheat cultivars (Fakhar-e-Bhakkar and Anaj-2017) were treated with the foliar application of different concentrations (control, 0.5, 1.0, and 1.5 PPM) of putrescine (put) under two moisture conditions (well- watered and terminal drought stress). The results demonstrate that the imposition of terminal drought stress significantly reduces different physiological and yield- related traits of both wheat cultivars. The reduction of relative water content (RWC%), membrane stability index (MSI), leaf area, tillers per plant, biomass yield, number of spikelets per spike, 100-grain weight, grain yield per plant, and straw yield was greater in Anaj-2017 than in Fakhar-e-Bhakkar cultivar. The results further explain that the foliar application of increased concentrations of putrescine from 0.0 to 1.0 PPM gradually improved physiological and yield traits, whereas these traits declined with the application of putrescine at the highest dose (1.5 PPM). The exogenous application of 1.0 PPM putrescine improved the relative water content (19.76%), specific leaf area (41.47%), and leaf area ratio (35.84%) compared with the controlled treatment. A higher grain yield (28.0 g plant-1) and 100-grain weight (3.8 g) were obtained with the foliar application of 1.0 PPM putrescine compared with controlled treatments. The findings of this study confirm the protective role of putrescine against terminal drought stress. It is therefore recommended to use putrescine at a concentration of 1.0 PPM, which could help alleviate terminal drought stress and attain better wheat yield. bread wheat yield terminal drought putrescine leaf area ratio membrane stability index This research work was funded for publication purpose by Institutional Fund Projects under grant no (IFPIP:1166-662-1443) provided the Ministry of Education and King Abdulaziz University, DSR, Jeddah, Saudi Arabia. ==== Body pmcIntroduction Drought stress is a serious threat to cereal crop growth, development, and productivity (Toulotte et al., 2022). The early growth and grain- filling phases of cereal crops are significantly affected by drought (Vijayaraghavareddy et al., 2021). Low water levels reduce vegetative growth, post-anthesis photosynthetic capacity, and grain yield in wheat (Cui et al., 2015). Drought occurs frequently in fields that alter the physiological and molecular characteristics of plants (Rahman et al., 2022; Islam et al., 2023 ). Drought stress reduces cell division, cell proliferation, and stem elongation, which leads to severe impairments to plant growth and productivity (Azooz and Youssef, 2010; Farooq et al., 2013; Wasaya et al., 2018). Plant morphology and physiology are adversely affected by environmental stresses, including drought, which ultimately reduce crop growth and agricultural yield (Sabagh et al., 2020). Drought stress leads to regulating several oxidative stress indicators, including the generation of reactive oxygen species (ROS), cellular injury, peroxidation of lipids, and membrane stability in plants (Kar, 2011; Rahman et al., 2015; Sabagh et al., 2021). Therefore, affordable and eco-friendly approaches are highly desirable for the improvement of economically important cereal crops under drought- stress conditions. Several eco-friendly approaches have been documented for abiotic stress tolerance and plant improvements (Sabagh et al., 2021; Khan et al., 2022; Raza et al., 2022a; Raza et al., 2022b). A numerical group of signaling molecules and growth regulators has been reported to alleviate drought stress in economically important crop plants, including cereals (Abd Elbar et al., 2019; Rahman et al., 2021; Islam et  al.,2022). Wheat (Triticum aestivum) is one of the most widely cultivated cereal crops across the globe, from subtropical to temperate climates, predominantly in the Mediterranean and semi-arid regions (Ahmed et al., 2019). It directly contributes to food security as it is considered the second-most significant staple food crop in the world (Zaheer et al., 2021). However, global plant production is at risk due to environmental changes, a decline in water levels, and irregular rain patterns, which severely influence crop growth stages (Nielsen et al., 2010; Rahman et al., 2018; Schmidt et al., 2020; Ahmed et al., 2022; Chauhan et al., 2022). For instance, the flowering and grain- filling stages of wheat are highly critical for yield protection, whereas the onset of terminal drought stress during these phases may cause a reduction in the number of kernels/ears and kernel weight and may be responsible for huge yield loss (Liu et al., 2015; Liu et al., 2016; Ebeed et al., 2017; Dong et al., 2017; Habib-ur-Rahman et al., 2022). Similarly, drought stress at anthesis can inflict a more profound impact on grain filling, leading to a shortened period of grain filling and altered enzymatic activities (Ahmadi and Baker, 2001; Shah and Paulsen, 2003; Pireivatlou, 2010; Farooq et al., 2014). Drought causes tissue dehydration, leading to metabolic impairment at critical growth stages of wheat crops (Farooq et al., 2009b; Hasan et al., 2021). Moreover, terminal drought disturbs stomatal oscillations, reduces transpiration and water- use efficiency (WUE), and decreases chlorophyll content and nutrient uptake, collectively resulting in yield reduction (Farooq et al., 2009a; Dawood et al., 2019; Yasir et al., 2019). It also imposes adverse effects on the physio-morphological responses of plants, resulting in significant yield reduction (Iqbal et al., 2019; Ejaz et al., 2022). For instance, drought stress induces the excessive production of ROS and inhibits antioxidant enzyme activity, which impairs photosynthesis, increases lipid peroxidation, and destroys cell membrane structure (Farooq et al., 2009a; Miller et al., 2010; Ju et al., 2020). Since crop genotype is an important determinant of drought response, different cultivars within crop species may differ strongly in their response and adaptation to drought stress (Abid et al., 2018; Mohammadi et al., 2018). Therefore, the response of wheat genotypes against drought stress, combined with their stress mitigation potential, must be studied to improve the drought tolerance of wheat genotypes in Pakistan. Nowadays, different strategies are being employed to mitigate drought stress in crop plants. One such strategy is the exogenous application of polyamines (PAs) (Phornvillay et al., 2019; Seifikalhor et al., 2020; Zhang et al., 2020; Çığ et al., 2021; Irshad et al., 2021). PAs, such as putrescine, are responsible for accumulating osmolytes and endogenous PAs in stressful environments, which help crop plants to protect themselves against stressful conditions (Ebeed et al., 2017). Putrescine (Put) is a type of PA found in all living organisms (Abd Elbar et al., 2019). The exogenous application of putrescine alleviated drought stress by regulating physio-biochemical traits in sugar beet (Islam et  al.,2022), regulating protein and fatty acids in thylakoid membrane under several abiotic stresses, including drought in plants (Rahman et al., 2014; Shu et al., 2015), and improving shoot/root dry matter ratio, morph-anatomical changes in roots, stems, and leaves to avoid tissue dehydration in Thymus vulgaris (Abd Elbar et al., 2019). Additionally, putrescine helps enhance the stability of cell membranes by preventing excessive water loss, ultimately retaining photosynthetic efficiency under adverse environmental conditions (Grigorova et al., 2012). The foliar spray of putrescine can trigger different physiological processes and induce osmotic adjustment in plants (Chen et al., 2019). It was also reported that putrescine foliar spray might slow aging and protect cell membranes from oxidative damage by removing free radicals in plants under drought-stress conditions (Hussein et al., 2019). These studies justify the beneficial effect of Put in plant abiotic stress. Under the current climate change scenario, there is a high demand for eco-friendly strategies for abiotic stress alleviation in crop plants in arid and semi-arid regions. Therefore, this study was undertaken to explore the insights of Put-mediated drought stress alleviation in wheat, along with how Put regulates agricultural and physiological yield attributes under water- deficient conditions. These studies open a new avenue of drought- stress alleviation in field crop research that might be useful to the plant breeder and farmer for improving drought stress tolerance in plants through breeding programs. Materials and methods Experimental sites, plant material, and treatments The present experiment was conducted under Lath House at the College of Agriculture, Bahauddin Zakariya University, Bahadur Sub-Campus Layyah, Punjab, Pakistan. Viable seeds of two different wheat varieties, Fakhar-e-Bhakkar (V1) and Anaj-2017 (V2), arranged by the Arid Zone Research Institute (AZRI) Bhakkar, Punjab, Pakistan, were used as experimental material. The pots were filled with soil and farmyard manure at a ratio of 2:1. Earthen pots (14 cm in diameter and 70 cm in height) were filled with 11 kg of well- pulverized sieved soil. The pots were irrigated uniformly and then, upon reaching optimum moisture conditions, eight seeds per pot were sown at a depth of 2 cm on 20th November 2020. The pots were irrigated regularly, and thinning was performed at the three-leaf stage of the seedlings to keep three plants per pot for observations. Commercial fertilizers were applied according to the recommended doses as follows: 0.66 g pot-1 nitrogen (as urea), 0.55 g pot-1 phosphorus as DAP (di-ammonium phosphate), 0.44 g pot-1 potassium as SOP (sulfate of potash) respectively. The crop was grown until the heading stage without drought imposition by watering the pots as and when required. At the heading stage, the pots were divided into two groups: well-watered (normal or control) and terminal drought (40% water holding capacity (WHC)). To maintain WHC, the pots were weighed regularly after every alternative day of drought imposition, and a difference in weight was achieved by adding water to attain the required weight. After one week of drought imposition at the heading stage, the wheat plants were sprayed with four different concentrations (PPM) of putrescine viz. putrescine 0 (control, put1), 0.5 (put 2), 1.0 (put 3), and 1.5 PPM (put 4). In the control treatment, plants were sprayed with distilled water, while the rest of the treatment groups were sprayed with different concentrations of putrescine, as per the treatment doses. The experiment containing three replications of each treatment under the factorial arrangement was laid out as a Completely Randomized Design (CRD). The experiment consisted of 48 experimental pots in total. Wheat plants were harvested on 14th April 2021 at physiological maturity. The harvested plants were sundried for a week and yield-related traits were collected. Measurement of leaf area, specific leaf area, and leaf area ratio Leaf area (LA), specific leaf area (SLA), and leaf area ratio (LAR) were determined after 7 days of putrescine treatment. The LA, SLA, and LAR were calculated from destructive flag leaf samples by following a distinct formula. The following Quarrie and Jones LA equation was used for measuring leaf area (Aldesuquy et al., 2014). Leaf area (cm2)=Length ×Width×0.75 The SLA and LAR were calculated using following formula (Amanullah et al., 2007). Specific Leaf Area=Leaf area÷leaf dry weight plant −1(cm2g−1) Leaf Area Ratio=Leaf area plant−1÷total dry weight plant −1(cm2g−1) The plant height (cm), number of tillers per plant, and plant biomass were measured for non-treated and treated plants. Plant samples were kept at 71°C in a hot air oven for 48 hours, then total biomass yields (dry weight basis) were recorded for two distinct wheat varieties, respectively. Determination of membrane stability index After 7 days of foliar application of putrescine, fully matured leaves were detached from each treated plant. Two pieces of fresh flag leaves, each of 0.2 g, were taken and divided into small strips then put in glass test tubes containing 10 ml of distilled water. Two sets of test tubes were prepared. One set of test tubes was put in a water bath for approximately 30 minutes with the temperature set at 40°C, then the EC (electrical conductivity) of each test tube sample was recorded using an EC meter and was designated C1. Meanwhile, the second set of test tubes was put in a water bath for about 15 minutes at 100°C; its EC (electrical conductivity) was noted and represented as C2. The MSI was estimated according to the given formula presented by Sairam et al. (1997). MSI=1−C1C2×100 Determination of relative water content Fresh, fully emerged flag leaves were collected from each experimental pot after 7 days of putrescine application and kept in a polythene bag, which was quickly shifted in the laboratory to estimate the relative water content (RWC). After moving the plant samples to the laboratory, the fresh weight (FW) of the detached leaves was recorded. Afterward, these leaves were dipped in distilled water and kept in dark conditions at room temperature for 24 hrs to record the turgid weight (TW). After recording the TW, leaf samples were air -dried in a hot air oven at a temperature of 70°C. The leaf samples were completely dried until a constant weight was achieved, then the dry weight (DW) was recorded using an electronic digital balance. The RWC was estimated according to the formula presented by Barrs and Weatherley (1962). RWC=[FW−DW/TW−DW]×100 Determination of yield -related traits Yield and related traits, like spike length (cm) per plant, spike weight (g) per plant, number of spikelets per spike, 100-grain weight (g), grain yield per plant, harvest index (HI), and straw yield were recorded for the wheat varieties, respectively, after plant harvesting. Three plants from each pot were randomly selected and the productive tillers per plant were counted manually. Moreover, the height (from the stem base to the apex of the ear) of three plants at maturity was recorded via a meter scale and was later averaged. Similarly, ten spikes per pot were selected randomly, and the length of each spike was measured using a scale then averaged to determine the spike length per plant. T hese selected spikes were then used for the calculation of the number of spikelets per spike and averaged. Three plants were harvested manually after reaching physiological maturity and sundried for a week, and biological yield was recorded using an electrical balance. The spikes were then separated from these three plants and threshed to record grain yield and the number of grains per spike. After threshing, 100 grains were counted from the seed lot of each pot to record the 100-grain weight. The harvest index (HI) was calculated considering the ratio of grain and biological yield as a percentage (%). Statistical analysis Data regarding all the traits were analyzed statistically by performing an analysis of a variance (ANOVA) test using the software, Statistix 8.1. Tukey’s test was applied in order to draw data means and the significant difference at P<0.05 (Steel et al., 1997). Bar graphs for the mean and interactive study were generated on graphing and analysis software, OriginPro Version 9.8.0.200. Results Effect of putrescine on the leaf area of wheat genotypes under drought stress Drought stress significantly reduced the LA, LAR, and SLA of both wheat genotypes. Moreover, these indices were considerably reduced in the Anaj-2017 (V2) wheat variety compared to the Fakhar-e-Bhakkar (V1) variety ( Figure 1 ). However, the putrescine application mitigated the impact of drought stress in both the genotypes, and the foliar application of 1.0 PPM seemed most suitable against drought stress. Maximum LA and LAR were recorded in Fakhar-e-Bhakkar with the foliar application of 1.0 PPM putrescine, while the minimum LA and LAR were recorded under controlled conditions. The foliar application of 1.0 PPM putrescine produced 39% more LA, 30% SLA, and 26% more LAR compared with controlled treatments ( Figure 1 ). Figure 1 Effect of putrescine on the leaf area (A), leaf area ratio (B), and specific leaf area (C) of two wheat cultivars under normal and terminal drought conditions. Abbreviation: PPM, parts per million, Put1, putrescine 0 PPM (untreated control); Put2, putrescine 0.5 PPM; Put3, putrescine 1.0 PPM; Put4, putrescine 1.5 PPM. The different letters above the bar columns indicate significant differences in mean ± SE at p< 0.05. At least three independent biological replications were considered for each treatment. Effect of putrescine on plant height, tillers per plant, and biomass accumulation of wheat genotypes under drought stress The plant height of both genotypes was significantly affected under both normal and drought conditions ( Figure 2 ). Among wheat cultivars, V1 attained significantly greater plant height, tillers per plant, and biomass accumulation than cultivar V2, under both normal and terminal stress conditions. Although putrescine improved the performance of both genotypes, especially under drought conditions, genotype V1 was more responsive to Putrescine than V2. Regarding plant height, putrescine application under all levels produced greater plant height than controlled treatments. Regarding the biomass per plant under all putrescine treatments, the 1.0 PPM was statistically superior to other concentrations, especially under drought conditions. Figure 2 The effect of putrescine on the plant height (A), tiller per plant (B), and biomass (g plant-1) (C) of two wheat cultivars under normal and terminal drought conditions. Abbreviation: PPM, parts per million, Put1, putrescine 0 PPM (untreated control); Put2, putrescine 0.5 PPM; Put3, putrescine 1.0 PPM; Put4, putrescine 1.5 PPM. The different letters above the bar columns indicate significant differences in mean ± SE at p< 0.05. At least three independent biological replications were considered for each treatment. Effect of putrescine on spike length, spike weight, and spikelet per spike of wheat genotypes under drought stress Wheat genotype Fakhar-e-Bhakkar (V1) showed a larger and heavier spike than the Anaj-2017 (V2), particularly under drought conditions ( Figure 3 ), which suggests that V2 is more sensitive to terminal drought. The application of putrescine, particularly at 1.00 PPM, significantly reduced drought impact, which was more obvious in V1. The interaction between drought stress, genotypes, and putrescine application showed an insignificant effect on spike length and spikelets per spike ( Figure 3 ), while it was found to significantly affect spike weight per plant. Under terminal drought stress, significantly, the spike weight per plant was observed with the foliar application of 1.0 PPM putrescine compared with other treatments ( Figure 3 ). Figure 3 The effect of putrescine on the spike length (A), spike weight (B), and spikelets per spike (C) of two wheat cultivars under normal and terminal drought conditions. Abbreviation: PPM, parts per million, Put1, putrescine 0 PPM (untreated control); Put2, putrescine 0.5 PPM; Put3, putrescine 1.0 PPM; Put4, putrescine 1.5 PPM. The different letters above the bar columns indicate significant differences in mean ± SE at p< 0.05. At least three independent biological replications were considered for each treatment. Effect of putrescine on MSI and RWC of wheat genotypes under drought stress The figure shows that higher RWC and MSI were recorded in Fakhar-e-Bhakkar under well-watered conditions with the foliar application of 1.0 PPM putrescine compared with other treatments. However, considering the drought treatments, the foliar application of 1.0 PPM putrescine maintained higher RWC and MSI in Fakhar-e-Bhakkar compared with other combinations, while less RWC and MSI was observed in Anaj2017 with water spray under terminal drought stress ( Figure 4 ). Figure 4 The effect of putrescine on the relative water content (A) and membrane stability index (%) (B) of two wheat cultivars under normal and terminal drought conditions. Abbreviation: PPM, parts per million, Put1, putrescine 0 PPM (untreated control); Put2, putrescine 0.5 PPM; Put3, putrescine 1.0 PPM; Put4, putrescine 1.5 PPM. The different letters above the bar columns indicate significant differences in mean ± SE at p< 0.05. At least three independent biological replications were considered for each treatment. Effect of putrescine on grain yield of wheat genotypes under drought stress Drought stress significantly reduced the grain yield and 100-grain weight in both genotypes ( Figure 5 ). Statistically, a higher 100-grain weight was recorded in Fakhar-e-Bhakkar under both well-watered and terminal drought stress with the foliar application of putrescine. However, a higher grain yield was recorded under well-watered conditions with putrescine application. Regarding terminal drought stress, Fakhar-e-Bhakkar produced a higher grain yield per plant with the foliar application of 1.0 PPM putrescine compared with other treatments. Similarly, a higher straw yield was recorded in Fakhar-e-Bhakkar with the foliar application of 1.0 PPM putrescine compared with other combinations. A higher harvest index was observed in Fakhar-e-Bhakkar under well-watered conditions, which was statistically on par with all other combinations ( Figure 5 ). Figure 5 The effect of putrescine on the 100-grain weight (A), grain yield (B), harvest index (C), and straw yield (D) of two wheat cultivars under normal and terminal drought conditions. Abbreviation: PPM, parts per million, Put1, putrescine 0 PPM (untreated control); Put2, putrescine 0.5 PPM; Put3, putrescine 1.0 PPM; Put4, putrescine 1.5 PPM. The different letters above the bar columns indicate significant differences in mean ± SE at p< 0.05. At least three independent biological replications were considered for each treatment. Discussion Climate change has worsened the outbreak of drought stress, particularly in the later reproductive stages of wheat. Therefore, the evaluation of different strategies to mitigate drought stress is necessary to enable wheat genotypes to perform well under drought conditions. This study investigates the drought tolerance and response of two wheat cultivars against putrescine application. The overall results of our study show that the wheat cultivar Fakhar-e-Bhakkar performed better against drought stress than Anaj-2017. In addition, in both genotypes, among four treatments of putrescine; the application of 1.0 PPM was found most effective against drought stress ( Figure 1 ). Putrescine application improved the leaf area and leaf area ratio in wheat Under drought stress, plants experience numerous morphological changes, such as reduced leaf size, leaf rolling, leaf senescence, and limited leaf extension, which could be the reason for lower LAR (leaf area ratio), LA (leaf area), and SLA in drought conditions. As Basu et al. (2016) reported drought stress could be the silent cause of leaf area reduction due to leaf rolling that minimizes water loss and decreases the activity of photosynthesis and the accumulation of dry matter. Reduced leaf area under drought stress might be due to reduced cell turgidity leading to a reduction in cell elongation, an imbalanced antioxidant defense system, and a higher number of reactive oxygen species (ROS) (Alishah et al., 2006; DaCosta and Huang, 2007; Zahoor et al., 2020). Similarly, Put improved leaf area, LAR, and SLA in the present study due to the regulation of LAI, turgidity, and photosynthetic activity (Farooq et al., 2009b). Putrescine application helped the wheat to retain higher biomass under drought stress The current study shows that drought stress had no significant effect on plant height and the number of tillers per plant in wheat, as the timing and length of drought stress execution are important factors affecting plant height and tiller number. Drought stress at the heading stage had little influence on plant height due to the cessation of stem elongation and vegetative growth. Wheat tillers cease to grow between the commencement of stem elongation and anthesis. Drought stress during the grain- filling stage in wheat reduces plant height by up to 7% (Caverzan et al., 2016a). Many researchers have found that water stress causes a considerable reduction in plant growth, which can be reflected in leaf area, dry weight, and other important growth parameters and functions (Kilic and Yağbasanlar, 2010; Budak et al., 2013; El Sabagh et al., 2019). Similarly, drought stress significantly reduced the biomass and tillers in wheat genotypes in this study, but the 0.5-1.0 PPM application of putrescine mitigated the drought effects. In addition, the higher number of tillers in the Fakhr-e-Bhakkar genotype suggests that it has higher drought tolerance ( Figure 2 ). Putrescine application improved the spike length and number of spikelets under drought stress Data analysis unveiled that drought and putrescine both considerably affect grain yield. Drought execution in wheat genotypes at the heading stage reduced the number of spikelets per plant, causing a reduction in yield. Exogenous putrescine application of 1.0 PPM recorded a positive effect on spikelets per spike and the spike weight when compared with the control treatment ( Figure 3 ) because of improved and higher RWC, MSI, and spikelet numbers. Scientists declared that grain yield was boosted due to increased spikelet numbers per spike (Philipp et al., 2018; Sakuma and Schnurbusch, 2020). Similarly, increased grain yield under putrescine application might be due to an improvement in growth and yield components as putrescine is involved in the production of abscisic acid (ABA) as well as osmotic adjustment and free radicals under abiotic stress (Velikova et al., 2000; Al-Kandari et al., 2009; Alcázar et al., 2010). Putrescine application reduced the oxidative damage caused by drought stress in wheat genotypes Data analysis of RWC and MSI revealed that drought-stressed plants have a lower water content than well-watered plants. Drought stress enforces physiological changes in the integrity of plant cell membranes. Consistent with Awasthi et al. (2014), drought stress boosts electrolyte leakage and enhances the instability of the membrane, together with lessening the chlorophyll content, having a huge impact on RWC. According to Reddy et al. (2004), drought stress has an inverse relationship with the water content of plant tissue; an increase in the severity and duration of drought stress results in reduced water content. Our investigation results of RWC and MSI also show that prolonged drought conditions diminish the membrane stability index and RWC. Alternatively, the exogenous putrescine application at the heading stage under terminal drought stress helped to retain the RWC inside the plant and reduced electrolyte leakage and ROS in contrast with no putrescine application. Putrescine, being a scavenger of “free radicals” or ROS detoxifiers, protected the membrane from oxidative injury. There is strong evidence to suggest that polyamines (PAs), like putrescine, have the ability to eliminate ROS by binding the conjugates directly (Hussain et al., 2011). Putrescine application within the 0.5-1.0 PPM range under drought stress alleviated the cell membrane injury in both wheat genotypes ( Figure 4 ). Foliar putrescine application upgraded the status of RWC (relative water content) under drought stress in wheat. It enhanced the organic compatible solute accretion that helps to avoid water loss in plants by increasing solute potential (Alcázar et al., 2010). The improvement in RWC through the foliar application of putrescine might be due to ROS’s scavenging role and ability to regulate osmotic balance (Al-Kandari et al., 2009; Alcázar et al., 2010). Exogenous application of putrescine improved MSI under drought stress. This might be due to the regulation of membrane integrity and antioxidant activity mediated by the translation of specific genes, which helps plants to resist ROS produced under drought stress (Bajguz, 2000; Takahashi and Kakehi, 2009). Putrescine application helps to retain the yield and harvest index of wheat under drought stress Data analysis demonstrated that water stress imposes the most prominent adverse effects in the Anaj-2017 genotype. Drought stress at the heading stage influenced the yield of terminal stress-sensitive wheat genotype and attributed to the severe reduction in yield components, namely biological yield, spikelet number, and grain weight. It was observed that drought stress reduced the yield of drought-sensitive wheat genotypes as a result of a decrease in photosynthetic parameters and chlorophyll content (Wasaya et al., 2021). Thus, it is obvious that stomata behavior, water status, and photosynthesis response to drought stress distinguish between sensitive genotypes and drought-tolerant genotypes (Munns et al., 2010). In relative terms, the foliar application of putrescine at the heading stage not only improves the yield of wheat genotypes but also assists in mitigating drought stress ( Figure 5 ). Data analysis showed that Fakhar-e-Bhakkar produced higher biological and grain yields than Anaj-2017 under drought stress, which could be attributed to the higher number of spikelets per spike, spike weight per plant relative to water content, and 100-grain weight. It was observed that some genotypes are able to complete the life cycle under moisture stress and maintain their yield compared with drought-susceptible genotypes due to their genetic makeup (Ahmad et al., 2009). Grain weight and spikelet number are the most effective and essential variables that could influence the grain yield of the wheat genotype (Leilah and Alkhateeb, 2005). Drought stress execution in the reproductive phase affects the traits contributing to yield parameters (Bayoumi et al., 2008). Fakhar-e-Bhakkar is a drought- resistant wheat genotype and attained the highest plant height, biomass, number of tillers, spike weight, number of spikelets per spike, grain weight per spike, 100-grain weight, and grain yield under drought stress and well-watered conditions. Because of its prodigious genetic makeup and its enzymatic defense system and osmolyte accumulation, the adverse effects of drought stress in the Fakhar-e-Bhakkar genotype are mitigated. Meanwhile, the wheat genotype Anaj-2017 recorded the lowest grain yield, being drought-sensitive, because sink capacity for dry matter accumulation is susceptible to drought conditions; nonetheless, this may be responsible for its low yield (Bayoumi et al., 2008). Conclusion The current study recommends that the foliar application of polyamine, such as putrescine, is the best approach to tackle and mitigate undesirable drought effects in wheat. A higher grain yield was achieved under the putrescine application, probably due to improved spike weight, spikelet numbers, and 100-grain weight, particularly in the stress-tolerant genotype (Fakhar-e-Bhakkar). According to the research findings, the foliar application of putrescine improved the membrane stability index and relative water content, which results in enhanced 100-grain weight and spike weight. Wheat cultivar Fakhar-e-Bhakkar showed more tolerance against terminal drought than Anaj-2017. Similarly, Fakhar-e-Bhakkar was most responsive to the putrescine application, and the foliar application of putrescine (1.0 PPM) has the potential to stabilize wheat against terminal drought stress. Our study suggests that Fakhar-e-Bhakkar has the potential to grow and survive in low- water and well-watered conditions, so farmers can grow this wheat genotype for longer with better outcomes. Data availability statement The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/supplementary material. Author contributions AW and TY: conceptualization. IR: data collection. AW, IR and TY: writing the original draft. AD, MK, MJ, GR, IA, ME-H, MB, MR, and AS: review and editing. All authors contributed to the article and approved the submitted version. Acknowledgments The authors gratefully acknowledge the technical and financial support provided by the Ministry of Education and King Abdulaziz University, DSR, Jeddah, Saudi Arabia, to this research work through project number IFPIP: 1166-662-1443. Conflict of interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Publisher’s note All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher. ==== Refs References Abd Elbar O. H. Farag R. E. Shehata S. A. (2019). Effect of putrescine application on some growth, biochemical and anatomical characteristics of thymus vulgaris l. under drought stress. Ann. Agric. Sci. 64 (2 ), 129–137. doi: 10.1016/j.aoas.2019.10.001 Abid M. Ali S. Qi L. K. Zahoor R. Tian Z. Jiang D. . (2018). Physiological and biochemical changes during drought and recovery periods at tillering and jointing stages in wheat (Triticum aestivum l.). Sci. Rep. 4615 (8 ), 1–15. doi: 10.1038/s41598-018-21441-7 Ahmad S. Ahmad R. Ashraf M. Y. Ashraf M. Waraich E. A. (2009). Sunflower (Helianthus annuus l.) response to drought stress at germination and seedling growth stages. Pak. J. Bot. 41 (2 ), 647–654. Ahmadi A. Baker D. A. (2001). The effect of water stress on the activities of key regulatory enzymes of the sucrose to starch pathway in wheat. Plant Growth Regul. 35 , 81–91. doi: 10.1023/A:1013827600528 Ahmed H. G. M. D. Sajjad M. Li M. Azmat M. A. Rizwan M. Maqsood R. H. . (2019). Selection criteria for drought-tolerant bread wheat genotypes at seedling stage. Sustainability 11 (9 ), 2584. doi: 10.3390/su11092584 Ahmed H. Zeng Y. Raza H. Muhammad D. Iqbal M. Uzair M. . (2022). Characterization of wheat accessions under salinity stress. Front. Plant Sci. 2583 . doi: 10.3389/fpls.2022.953670 Alcázar R. Altabella T. Marco F. Bortolotti C. Reymond M. Koncz C. . (2010). Polyamines: molecules with regulatory functions in plant abiotic stress tolerance. Planta 231 (6 ), 1237–1249. doi: 10.1007/s00425-010-1130-0 20221631 Aldesuquy H. Baka Z. Mickky B. (2014). Kinetin and spermine mediated induction of salt tolerance in wheat plants: Leaf area, photosynthesis and chloroplast ultrastructure of flag leaf at ear emergence. Egyp. J. Basic Appl. Sci. 1 , 77–87. doi: 10.1016/j.ejbas.2014.03.002 Alishah H. M. Heidari R. Hassani A. Dizaji A. (2006). Efect of water stress on some morphological and biochemical characteristics of purple basil (Ocimum basilicum). Res. J. Biol. Sci. 6 (4 ), 763–767. doi: 10.3923/jbs.2006.763.767 Al-Kandari M. Redha A. Suleman P. (2009). Polyamine accumulation and osmotic adjustment as adaptive responses to water and salinity stress in Conocarpus lancifolius . Func. Plant Sci. Biotech. 3 (1 ), 42–48. doi: 10.3390/plants10071313 Amanullah M. J. H. Nawab K. Ali A. (2007). Response of specific leaf area (SLA), leaf area index (LAI) and leaf area ratio (LAR) of maize (Zea mays l.) to plant density, rate and timing of nitrogen application. World Appl. Sci. J. 2 (3 ), 235–243. Awasthi R. Kaushal N. Vadez V. Turner N. C. Berger J. Siddique K. H. . (2014). Individual and combined effects of transient drought and heat stress on carbon assimilation and seed filling in chickpea. Funct. Plant Biol. 41 , 1148–1167. doi: 10.1071/FP13340 32481065 Azooz M. Youssef M. (2010). Evaluation of heat shock and salicylic acid treatments as inducers of drought stress tolerance in hassawi wheat. Amer. J. Plant Physiol. 5 (2 ), 56–70. doi: 10.3923/ajpp.2010.56.70 Bajguz A. (2000). Efect of brassinosteroids on nucleic acid and protein content in cultured cell of Chlorella vulgaris . Plant Physiol. Biochem. 38 , 209–215. doi: 10.1016/S0981-9428(00)00733-6 Barrs H. D. Weatherley P. E. (1962). A re-examination of the relative turgidity technique for estimating water deficits in leaves. Aust. J. Biol. Sci. 15 , 413–428. doi: 10.1071/BI9620413 Basu S. Ramegowda V. Kumar A. Pereira A. (2016). Plant adaptation to drought stress. F1000Res. 5 F1000 Faculty, Rev–1554. doi: 10.12688/f1000research.7678.1 Bayoumi T. Y. Eid M. H. Metwali E. M. (2008). Application of physiological and biochemical indices as a screening technique for drought tolerance in wheat genotypes. Afr. J. Biotechnol. 7 , 2341–2352. Budak H. Kantar M. Kurtoglu K. Y. (2013). Drought tolerance in modern and wild wheat. Sci. C World J. Caverzan A. Casassola A. Brammer S. P. (2016a). Antioxidant responses of wheat plants under stress. Gene. Mol. Biol. 39 (1 ), 1–6. doi: 10.1590/1678-4685-GMB-2015-0109 Chauhan J. Srivastava J. P. Singhal R. K. Soufan W. Dadarwal B. K. Mishra U. N. . (2022). Alterations of oxidative stress indicators, antioxidant enzymes, soluble sugars, and amino acids in mustard [Brassica juncea (L.) czern and coss.] in response to varying sowing time, and field temperature 13 . doi: 10.3389/fpls.2022.875009 Chen D. Shao Q. Yin L. Younis A. Zheng B. (2019). Polyamine function in plants: metabolism, regulation on development, and roles in abiotic stress responses. Front. Plant Sci. 9 , 1945. doi: 10.3389/fpls.2018.01945 30687350 Çığ F. Sönmez F. Nadeem M. A. El Sabagh A. (2021). ). effect of biochar and pgpr on the growth and nutrients content of einkorn wheat (Triticum monococcum l.) and post-harvest soil properties. Agronomy 11 (12 ), 2418. doi: 10.3390/agronomy11122418 Cui Y. Tian Z. Zhang X. Muhammad A. Han H. Jiang D. . (2015). Effect of water deficit during vegetative growth periods on post-anthesis photosynthetic capacity and grain yield in winter wheat (Triticum aestivum l.). Acta Physiol. Plant 37 (10 ), 1–10. doi: 10.1007/s11738-015-1944-2 DaCosta M. Huang B. (2007). Changes in antioxidant enzyme activities and lipid peroxidation for bent grass species in response to drought stress. J. Am. Soc Hortic. Sci. 132 , 319–326. doi: 10.21273/JASHS.132.3.319 Dawood M. F. A. Abeed A. H. A. Aldaby E. E. S. (2019). Titanium dioxide nanoparticles model growth kinetic traits of some wheat cultivars under different water regimes. Indian J. Plant Physiol. 24 , 129–140. doi: 10.1007/s40502-019-0437-5 Dong B. Zheng X. Liu H. Able J. A. Yang H. Zhao H. (2017). Effects of drought stress on pollen sterility, grain yield, abscisic acid and protective enzymes in two winter wheat cultivars. Front. Plant Sci. 8 , 1008. doi: 10.3389/fpls.2017.01008 28676806 Ebeed H. T. Hassan N. M. Aljarani A. M. (2017). Exogenous applications of polyamines modulate drought responses in wheat through osmolytes accumulation, increasing free polyamine levels and regulation of polyamine biosynthetic genes. Plant Physiol. Biochem. 118 , 438–448. doi: 10.1016/j.plaphy.2017.07.014 28743037 Ejaz M. K. Aurangzaib M. Iqbal R. Shahzaman M. Habib-ur-Rahman M. El-Sharnouby M. . (2022). The use of soil conditioners to ensure a sustainable wheat yield under water deficit conditions by enhancing the physiological and antioxidant potentials. Land 11 (3 ), 368. doi: 10.3390/land11030368 El Sabagh A. Hossain A. Barutcular C. Islam M. S. Awan S. I. Galal A. . (2019). Wheat (Triticum aestivum l.) production under drought and heat stress–adverse effects, mechanisms and mitigation: A review. Appl. Ecol. Environ. Res. 17 (4 ), 8307–8332. doi: 10.15666/aeer/1704_83078332 Farooq M. Hussain M. Siddique K. H. M. (2014). Drought stress in wheat during flowering and grain-filling periods. CRC Crit. Rev. Plant Sci. 33 , 331–349. doi: 10.1080/07352689.2014.875291 Farooq M. Irfan M. Aziz T. Ahmad I. Alam S. (2013). Seed priming with ascorbic acid improves drought resistance of wheat. J. Agron. Crop Sci. 199 , 12–22. doi: 10.1111/j.1439-037X.2012.00521.x Farooq M. Wahid A. Kobayashi N. Fujita D. Basra S. M. A. (2009a). Plant drought stress: effects, mechanisms and management. Agron. Sustain. Dev. 29 , 185–212. doi: 10.1051/agro:2008021 Farooq M. Wahid A. Lee D. J. (2009b). Exogenously applied polyamines increase drought tolerance of rice by improving leaf water status, photosynthesis, and membrane properties. Acta Physiol. Plant 31 , 937–945. doi: 10.1007/s11738-009-0307-2 Grigorova B. Vassileva V. Klimchuk D. Vaseva I. Demirevska K. Feller U. (2012). Drought, high temperature, and their combination affect ultrastructure of chloroplasts and mitochondria in wheat (Triticum aestivum l.) leaves. J. Plant Interact. 7 , 204–213. doi: 10.1080/17429145.2011.654134 Habib-ur-Rahman M. Raza A. Ahrends H. E. Hüging H. Gaiser T. (2022). Impact of in-field soil heterogeneity on biomass and yield of winter triticale in an intensively cropped hummocky landscape under temperate climate conditions. Precis. Agric. 23 , 912–938. doi: 10.1007/s11119-021-09868-x Hasan M. M. Skalicky M. Jahan M. S. Hossain M. N. Anwar Z. Zhengfei N. . (2021). Spermine: Its emerging role in regulating drought stress responses in plants. Cells 10 , 261. doi: 10.3390/cells10020261 33525668 Hussain S. S. Ali M. Ahmad M. Siddique K. H. M. (2011). Polyamines: natural and engineered abiotic and biotic stress tolerance in plants. Biotechnol. Adv. 29 , 300–311. doi: 10.1016/j.biotechadv.2011.01.003 21241790 Hussein H. A. A. Mekki B. B. El-Sadek M. E. A. El-Lateef E. E. (2019). Effect of l-ornithine application on improving drought tolerance in sugar beet plants. Heliyon 5 , e02631. doi: 10.1016/j.heliyon 31667428 Iqbal N. Hussain S. Raza M. A. Yang C. Q. Safdar M. E. Brestic M. . (2019). Drought tolerance of soybean (Glycine max l. merr.) by improved photosynthetic characteristics and an efficient antioxidant enzyme activities under a split-root system. Front. Physiol. 10 . doi: 10.3389/fphys.2019.00786 Irshad M. Ullah F. Fahad S. Mehmood S. Khan A. U. Brtnicky M. . (2021). Evaluation of Jatropha curcas l. leaves mulching on wheat growth and biochemical attributes under water stress. BMC Plant Biol. 21 (1 ), 1–12. doi: 10.1186/s12870-021-03097-0 33386069 Islam M.-R. Kamal M.-M. Hossain M.-F. Hossain J. Azam M.-G. Akhter M.-M. . (2022). Drought tolerance in mung bean is associated with the genotypic divergence, regulation of proline, photosynthetic pigment and water relation. Phyton. doi: 10.32604/phyton.2023.025138 Islam M.-R. Kamal M. -M. Hossain M. -F. Hossain J. Azam M. G. Akhter M. -M. . (2023). Drought tolerance in mung bean is associated with the genotypic divergence, regulation of proline, photosynthetic pigment and water relation). Phyton 92 (3 ), 955–981. doi: 10.32604/phyton.2023.025138 Ju Y. L. Yue X. F. Min Z. Wang X. H. Fang Y. L. Zhang J. X. (2020). VvNAC17, a novel stress-responsive grapevine (Vitis vinifera l.) NAC transcription factor, increases sensitivity to abscisic acid and enhances salinity, freezing, and drought tolerance in transgenic arabidopsis. Plant Physiol. Biochem. 146 , 98–111. doi: 10.1016/j.plaphy.2019.11.002 31734522 Kar R. K. (2011). Plant responses to water stress: Role of reactive oxygen species . Plant Signal. Behav 6 , 1741–1745. Khan I. Muhammad A. Chattha M. U. Skalicky M. Chattha M. B. Ayub M. A. . (2022). Mitigation of salinity-induced oxidative damage, growth, and yield reduction in fine rice by sugarcane press mud application. Front. Plant Sci. 13 . doi: 10.3389/fpls.2022.840900 Kilic H. Yağbasanlar T. (2010). The effect of drought stress on grain yield, yield components and some quality traits of durum wheat (Triticum turgidum) cultivars. Not. Bot. Horti. Agrobot. 38 , 164. doi: 10.15835/nbha3814274 Leilah A. A. Alkhateeb S. A. (2005). Statistical analysis of wheat yield under drought conditions. J. Arid. Environ. 61 , 483–496. doi: 10.1016/j.jaridenv.2004.10.011 Liu Y. Liang H. Lv X. Liu D. Wen X. Liao Y. (2016). Effect of polyamines on the grain filling of wheat under drought stress. Plant Physiol. Biochem. 100 , 113–129. doi: 10.1016/j.plaphy.2016.01.003 26812255 Liu H. Searle I. R. Mather D. E. Able A. J. Able J. A. (2015). Morphological, physiological and yield responses of durum wheat to pre-anthesis water-deficit stress are genotype-dependent. Crop Pasture Sci. 66 , 1024–1038. doi: 10.1071/CP15013 Miller G. Suzuki N. Ciftci-yilmaz S. Mittler R. (2010). Reactive oxygen species homeostasis and signalling during drought and salinity stresses. Plant Cell Environ. 33 , 453–467. doi: 10.1111/j.1365-3040.2009.02041.x 19712065 Mohammadi H. Ghorbanpour M. Brestic M. (2018). Exogenous putrescine changes redox regulations and essential oil constituents in field-grown thymus vulgaris l. under well-watered and drought stress conditions Ind . Crops Prod. 122 , 119–132. doi: 10.1016/j.indcrop.2018.05.064 Munns R. James R. A. Sirault X. R. R. Furbank R. T. Jones H. G. (2010). New phenotyping methods for screening wheat and barley for beneficial responses to water deficit. J. Exp. Bot. 61 (13 ), 3499–3507. doi: 10.1093/jxb/erq199 20605897 Nielsen D. C. Halvorson A. D. Vigil M. F. (2010). Critical precipitation period for dryland maize production. Field Crops Res. 118 , 259–263. doi: 10.1016/j.fcr.2010.06.004 Nishizawa-Yokoi A. Yabuta Y. Shigeoka S. (2008). The contribution of carbohydrates including raffinose family oligosaccharides and sugar alcohols to protection of plant cells from oxidative damage. Plant Signal. Behav. 3 , 1016–1018. doi: 10.4161/psb.6738 19704439 Philipp N. Weichert H. Bohra U. Weschke W. Schulthess A. W. Weber H. (2018). Grain number and grain yield distribution along the spike remain stable despite breeding for high yield in winter wheat. PloS One 13 , e0205452. doi: 10.1371/journal.pone.0205452 30304020 Phornvillay S. Pongprasert N. Wongs-Aree C. Uthairatanakij A. Srilaong V. (2019). Exogenous putrescine treatment delays chilling injury in okra pod (Abelmoschus esculentus) stored at low storage temperature. Sci. Hortic. Amsterdam 256 , 1–7. doi: 10.1016/j.scienta.2019.108550 Pireivatlou A. S. (2010). Evaluation of wheat (Triticum aestivum l.) genotypes under pre-and post-anthesis drought stress conditions. J. Agric. 10 , 109–121. Rahman M. A. Alam I. Kim Y. G. Ahn N.-Y Ahn N.-Y Heo S.-Y Lee D.-Y (2015). Screening for salt-responsive proteins in two contrasting alfalfa cultivars using a comparative proteome approach . Plant Physiol. Biochem. 89 , 112–122. Rahman M. H. U. Ahmad A. Wang X. Wajid A. Nasim W. Hussain M. . (2018). Multi-model projections of future climate and climate change impacts uncertainty assessment for cotton production in Pakistan. Agric. For. Meteorol. 253 , 94–113. doi: 10.1016/j.agrformet.2018.02.008 Rahman M. A. Alam I. Sharmin S. A. Kabir A. H. Kim Y.-G. Liu G. . (2021). Physiological and proteomic analyses reveal the protective roles of exogenous hydrogen peroxide in alleviating drought stress in soybean plants. Plant Biotechnol. Rep. 15 (6 ), 805–818. doi: 10.1007/s11816-021-00719-9 Rahman M. A. Kim Y.-G. Lee B.-H. (2014). Proteomic response of alfalfa subjected to aluminum (Al) stress at low pH soil. J. Korean Soc. Grassl. Forage Sci. 34 (4 ), 262–268. doi: 10.5333/KGFS.2014.34.4.262 Rahman M. A. Woo J. H. Song Y. Lee S.-H. Hasan M. M. Azad M. A. K. . (2022). Heat shock proteins and antioxidant genes involved in heat combined with drought stress responses in perennial rye grass. Life 12 (9 ), 1426. doi: 10.3390/life12091426 36143461 Raza A. Charagh S. García-Caparrós P. Rahman M. A. Ogwugwa V. H. Saeed F. . (2022a). Melatonin-mediated temperature stress tolerance in plants. GM Crops Food . 13 (1 ), 196–217. doi: 10.1080/21645698.2022.2106111 35983948 Raza A. Salehi H. Rahman M. A. Zahid Z. Haghjou M. M. Najafi-Kakavand S. . (2022b). Plant hormones and neurotransmitter interactions mediate antioxidant defenses under induced oxidative stress in plants. Front. Plant Sci. 13 . doi: 10.3389/fpls.2022.961872 Reddy A. R. Chaitanya K. V. Vivekanandan M. (2004). Drought-induced responses of photosynthesis and antioxidant metabolism in higher plants. J. Plant Physiol. 161 , 1189–1202. doi: 10.1016/j.jplph.2004.01.013 15602811 Sabagh A. E. Hossain A. Islam M. S. Iqbal M. A. Raza A. Karademir Ç. . (2020). Elevated CO2 concentration improves heat-tolerant ability in crops. In Abiotic Stress Plants IntechOpen, 1–17. doi: 10.5772/intechopen.94128 Sabagh A. E. Mbarki S. Hossain A. Iqbal M. A. Islam M. S. Raza A. . (2021). Potential role of plant growth regulators in administering crucial processes against abiotic stresses. Front. Agron. 3 . doi: 10.3389/fagro.2021.648694 Sairam R. K. Deshmukh P. S. Shukla D. S. (1997). Tolerance to drought and temperature stress in relation to increased antioxidant enzyme activity in wheat. J. Agron. Crop Sci. 178 , 171–177. doi: 10.1111/j.1439-037X.1997.tb00486.x Sakuma S. Schnurbusch T. (2020). Of floral fortune: tinkering with the grain yield potential of cereal crops. New Phytol. 225 , 1873–1882. doi: 10.1111/nph.16189 31509613 Schmidt J. Claussen J. Worlein N. Eggert A. Fleury D. Garnett T. . (2020). Drought and heat stress tolerance screening in wheat using computed tomography. Plant Methods 16 , 1–12. doi: 10.1186/s13007-020-00565-w 31911810 Seifikalhor M. Aliniaeifard S. Bernard F. Seif M. Li T. (2020). γ-aminobutyric acid confers cadmium tolerance in maize plants by concerted regulation of polyamine metabolism and antioxidant defense systems. Sci. Rep. 10 , 3356. doi: 10.1038/s41598-020-59592-1 32098998 Shah N. H. Paulsen G. M. (2003). Interaction of drought and high temperature on photosynthesis and grain-filling of wheat. Plant Soil 257 , 219–226. doi: 10.1023/A:1026237816578 Shu S. Yuan Y. Chen J. Sun J. Zhang W. Tang Y. . (2015). The role of putrescine in the regulation of proteins and fatty acids of thylakoid membranes under salt stress. Sci. Rep. 5 (1 ), 14390. doi: 10.1038/srep14390 26435404 Steel R. G. D. Torrie J. H. Dicky D. A. (1997). “Principles and procedures of statistics,” in A biometrical approach, 3rd edition (New York: McGraw Hill Book Int. Co.), 172–177. Takahashi T. Kakehi J. I. (2009). Polyamines: ubiquitous polycations with unique roles in growth and stress responses. Ann. Bot. 105 , 1–6. doi: 10.1093/aob/mcp259 Toulotte J. M. Pantazopoulou C. K. Sanclemente M. A. Voesenek L. Sasidharan R. (2022). Water stress resilient cereal crops: lessons from wild relatives. J. Integr. Plant Biol. 64 (2 ), 412–430. doi: 10.1111/jipb.13222 35029029 Velikova V. Yordanov I. Edreva A. (2000). Oxidative stress and some antioxidant systems in acid rain-treated bean plants: protective role of exogenous polyamines. Plant Sci. 151 (1 ), 59–66. doi: 10.1016/S0168-9452(99)00197-1 Vijayaraghavareddy P. Akula N. N. Vemanna R. S. Math R. G. H. Shinde D. D. Yin X. . (2021). Metabolome profiling reveals impact of water limitation on grain filling in contrasting rice genotypes. Plant Physiol. Biochem. 162 , 690–698. doi: 10.1016/j.plaphy.2021.02.030 33780742 Wasaya A. Manzoor S. Yasir T. A. Sarwar N. Mubeen K. Ismail I. A. . (2021). Evaluation of fourteen bread wheat (Triticum aestivum l.) genotypes by observing gas exchange parameters, relative water and chlorophyll content, and yield attributes under drought stress. Sustainability 13 (9 ), 4799. doi: 10.3390/su13094799 Wasaya A. Zhang X. Fang Q. Yan Z. (2018). Root phenotyping for drought tolerance: a review. Agronomy 8 (11 ), 241. doi: 10.3390/agronomy8110241 Yasir T. A. Wasaya A. Hussain M. Ijaz M. Farooq M. Farooq O. . (2019). Evaluation of physiological markers for assessing drought tolerance and yield potential in bread wheat. Physiol. Molec. Bio. Plants. 25 , 1163–1174. doi: 10.1007/s12298-019-00694-0 31564779 Zaheer M. S. Ali H. H. Soufan W. Iqbal R. Habib-ur-Rahman M. Iqbal J. . (2021). Potential effects of biochar application for improving wheat (Triticum aestivum l.) growth and soil biochemical properties under drought stress conditions. Land 10 (11 ), 1125. doi: 10.3390/land10111125 Zahoor A. Waraich E. A. Barutcular C. Hossain A. Erman M. Çig F. . (2020). Enhancing drought tolerance in wheat through improving morphophysiological and antioxidants activities of plants by the supplementation of foliar silicon. Phyton 89 , 52. doi: 10.32604/phyton.2020.09143 Zhang F. Zou Y. N. Wu Q. S. Kuca K. (2020). Arbuscular mycorrhizas modulate root polyamine metabolism to enhance drought tolerance of trifoliate orange. Environ. Exp. Bot. 171 , 103926. doi: 10.1016/j.envexpbot.2019.103926
PMC010xxxxxx/PMC10163230.txt
==== Front Nat Commun Nat Commun Nature Communications 2041-1723 Nature Publishing Group UK London 37147282 37194 10.1038/s41467-023-37194-5 Article Clarifying the effect of biodiversity on productivity in natural ecosystems with longitudinal data and methods for causal inference http://orcid.org/0000-0003-0471-1371 Dee Laura E. [email protected] 1 http://orcid.org/0000-0002-4777-5108 Ferraro Paul J. [email protected] 23 Severen Christopher N. 4 Kimmel Kaitlin A. 1 http://orcid.org/0000-0003-2259-5853 Borer Elizabeth T. 5 http://orcid.org/0000-0002-9791-9472 Byrnes Jarrett E. K. 6 Clark Adam Thomas 7 http://orcid.org/0000-0003-4347-7741 Hautier Yann 8 http://orcid.org/0000-0002-1309-7716 Hector Andrew 9 Raynaud Xavier 10 http://orcid.org/0000-0003-4424-662X Reich Peter B. 111213 http://orcid.org/0000-0002-5205-8576 Wright Alexandra J. 14 Arnillas Carlos A. 15 Davies Kendi F. 1 MacDougall Andrew 16 http://orcid.org/0000-0002-8422-1198 Mori Akira S. 17 http://orcid.org/0000-0003-4920-6985 Smith Melinda D. 1819 http://orcid.org/0000-0002-4216-4009 Adler Peter B. 20 http://orcid.org/0000-0002-8973-2771 Bakker Jonathan D. 21 http://orcid.org/0000-0002-8099-285X Brauman Kate A. 22 http://orcid.org/0000-0002-6754-9669 Cowles Jane 5 http://orcid.org/0000-0001-7056-4547 Komatsu Kimberly 23 Knops Johannes M. H. 24 http://orcid.org/0000-0002-2393-0599 McCulley Rebecca L. 25 http://orcid.org/0000-0001-9809-5092 Moore Joslin L. 26 Morgan John W. 27 Ohlert Timothy 18 http://orcid.org/0000-0002-2723-8671 Power Sally A. 13 Sullivan Lauren L. 2829 http://orcid.org/0000-0002-2390-1763 Stevens Carly 30 Loreau Michel 31 1 grid.266190.a 0000000096214564 Department of Ecology and Evolutionary Biology, University of Colorado, Boulder, CO USA 2 grid.21107.35 0000 0001 2171 9311 Department of Environmental Health and Engineering, Bloomberg School of Public Health & Whiting School of Engineering, Johns Hopkins University, Baltimore, MD USA 3 grid.21107.35 0000 0001 2171 9311 Carey Business School, Johns Hopkins University, Baltimore, MD USA 4 grid.507405.3 0000 0001 0740 3062 Research Department, Federal Reserve Bank of Philadelphia, Philadelphia, PA USA 5 grid.17635.36 0000000419368657 Department of Ecology, Evolution, and Behavior, University of Minnesota, St. Paul, MN 55108 USA 6 grid.266685.9 0000 0004 0386 3207 Department of Biology, University of Massachusetts Boston, 100 Morissey Blvd, Boston, MA 02125 USA 7 grid.5110.5 0000000121539003 Institute of Biology, University of Graz, Holteigasse 6, 8010 Graz, Austria 8 grid.5477.1 0000000120346234 Ecology and Biodiversity Group, Department of Biology, Utrecht University, Padualaan 8, 3584 CH Utrecht, The Netherlands 9 grid.4991.5 0000 0004 1936 8948 Department of Plant Sciences, University of Oxford, Oxford, OX1 3RB UK 10 grid.462350.6 Sorbonne Université, Université Paris Cité, UPEC, IRD, CNRS, INRA, Institute of Ecology and Environmental Sciences, iEES Paris, Paris, France 11 grid.214458.e 0000000086837370 Institute for Global Change Biology, and School for Environment and Sustainability, University of Michigan, Ann Arbor, MI USA 12 grid.17635.36 0000000419368657 Department of Forest Resources, University of Minnesota, St. Paul, MN 55108 USA 13 grid.1029.a 0000 0000 9939 5719 Hawkesbury Institute for the Environment, Western Sydney University, Penrith, NSW 2751 Australia 14 grid.253561.6 0000 0001 0806 2909 Department of Biological Sciences, California State University Los Angeles, Los Angeles, CA USA 15 grid.17063.33 0000 0001 2157 2938 Department of Physical and Environmental Sciences, University of Toronto at Scarborough, Toronto, 1265 Military Trail, ON M1C 1A4 Canada 16 grid.34429.38 0000 0004 1936 8198 Department of Integrative Biology, University of Guelph, Guelph, Ontario, N1G 2W1 Canada 17 grid.26999.3d 0000 0001 2151 536X Research Center for Advanced Science and Technology, The University of Tokyo, 4-6-1 Komaba, Meguro, Tokyo, 153-8904 Japan 18 grid.47894.36 0000 0004 1936 8083 Department of Biology, Colorado State University, Fort Collins, CO 80523 USA 19 grid.47894.36 0000 0004 1936 8083 Graduate Degree Program in Ecology, Colorado State University, Fort Collins, CO 80523 USA 20 grid.53857.3c 0000 0001 2185 8768 Department of Wildland Resources and the Ecology Center, Utah State University, Logan, UT 84322 USA 21 grid.34477.33 0000000122986657 School of Environmental and Forest Sciences, University of Washington, Box 354115, Seattle, WA 98195-4115 USA 22 grid.411015.0 0000 0001 0727 7545 Global Water Security Center, The University of Alabama, Box 870206, Tuscaloosa, AL 35487 US 23 grid.419533.9 0000 0000 8612 0361 Smithsonian Environmental Research Center, Edgewater, MD 21037 USA 24 grid.440701.6 0000 0004 1765 4000 Department of Health and Environmental Sciences, Xián Jiaotong-Liverpool University, Suzhou, China 25 grid.266539.d 0000 0004 1936 8438 Department of Plant and Soil Sciences, University of Kentucky, Lexington, KY 40546-0312 USA 26 grid.1002.3 0000 0004 1936 7857 School of Biological Sciences, Monash University, Clayton, VIC 3800 Australia 27 grid.1018.8 0000 0001 2342 0938 Department of Ecology, Environment and Evolution, La Trobe University, Bundoora, VIC 3086 Australia 28 grid.17088.36 0000 0001 2150 1785 Department of Plant Biology, Michigan State University, East Lansing, MI 48824 USA 29 grid.17088.36 0000 0001 2150 1785 Kellogg Biological Station, Michigan State University, Hickory Corners, MI 49060 USA 30 grid.9835.7 0000 0000 8190 6402 Lancaster Environment Centre, Lancaster University, Lancaster, LA1 4YQ UK 31 grid.462549.8 Theoretical and Experimental Ecology Station, CNRS, 09200 Moulis, France 5 5 2023 5 5 2023 2023 14 260710 3 2022 3 3 2023 © The Author(s) 2023, corrected publication 2023 https://creativecommons.org/licenses/by/4.0/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Causal effects of biodiversity on ecosystem functions can be estimated using experimental or observational designs — designs that pose a tradeoff between drawing credible causal inferences from correlations and drawing generalizable inferences. Here, we develop a design that reduces this tradeoff and revisits the question of how plant species diversity affects productivity. Our design leverages longitudinal data from 43 grasslands in 11 countries and approaches borrowed from fields outside of ecology to draw causal inferences from observational data. Contrary to many prior studies, we estimate that increases in plot-level species richness caused productivity to decline: a 10% increase in richness decreased productivity by 2.4%, 95% CI [−4.1, −0.74]. This contradiction stems from two sources. First, prior observational studies incompletely control for confounding factors. Second, most experiments plant fewer rare and non-native species than exist in nature. Although increases in native, dominant species increased productivity, increases in rare and non-native species decreased productivity, making the average effect negative in our study. By reducing the tradeoff between experimental and observational designs, our study demonstrates how observational studies can complement prior ecological experiments and inform future ones. Isolating the relationships between biodiversity and ecosystem functioning in natural ecosystems is challenging. Here, the authors apply a causal inference approach to observational data from grasslands and find a negative effect of biodiversity on productivity driven by non-native and rare species. Subject terms Biodiversity Community ecology Grassland ecology Ecological modelling issue-copyright-statement© Springer Nature Limited 2023 ==== Body pmcIntroduction Motivated by global changes in biodiversity, ecologists have advanced our understanding of the consequences of biodiversity change for ecosystem functioning1–11. One particularly active area of this research has focused on how plant species diversity affects ecosystem productivity1–4,12. To shed light on this causal relationship, studies have used both experimental and non-experimental designs, each of which presents distinct advantages and disadvantages for elucidating causal relationships in natural ecosystems. Experimentalists that manipulate plant species richness often infer that increases in richness cause increases in biomass1,2,7,13. Although experimental manipulations facilitate causal inferences, most experiments that manipulate richness are designed to test theory14,15 rather than to simulate how species richness changes in natural ecosystems16–18. If the effect of richness on productivity depends on the specific species gained and lost, and how they are gained and lost, inferences from experiments may not generalize to natural ecosystems (Fig. 1A)16,17,19. For example, many biodiversity experiments simulate random gains and losses of species (but see20–22), which may not mimic changes in species richness in nature. Moreover, most experiments plant common, native species (but see16,19–21,23). However, in diverse natural ecosystems, most species are rare24 and non-native species are increasingly prevalent25.Fig. 1 Challenges in estimating the causal effect of species richness on productivity. A Experimental designs permit credible causal inferences with few modeling assumptions. Yet experiments often manipulate richness in random permutations, plant limited sets of species, and weed out colonizing species. Such designs can yield ecological processes that differ from processes in natural systems. In A, when common species are more likely to be planted in experiments than rare species, the proportion of common species is higher than the proportion of rare species regardless of the planted richness level. In contrast, in natural communities, higher species richness is associated with greater numbers of rare species than common species. B Observational designs include natural processes but causal inferences are challenged by confounding variables (U) associated with both richness (R) and productivity (P); e.g., precipitation can increase both R and P, thereby inducing a positive correlation between the two, even if the true causal relationship were zero or negative. Some of these confounding variables, like topography, may be time-invariant (or slow-changing) over the study period at the level of the plot (Up) or site (Us). Others may be time-varying at the site (Ust), such as weather, or the plot (Upt), such as micro-climate. To estimate the effect of R on P without bias does not require data on variables I that only affect P, or Z that only affect R, nor on mechanisms (M), such as selection or complementarity. However, data on Z and M can help address unobserved confounders and differentiate the effect of R on P from the effect of P on R. Observational studies can capture the consequences of changes in species richness that occur in nature. However, determining the causal effect of richness on productivity in observational studies requires strong assumptions26,27. Confounding variables associated with both richness and productivity15 can mask or mimic a causal relationship between them (Fig. 1B). For example, unobserved differences in soil nitrogen across locations can mask a positive relationship between richness and productivity if more nitrogen reduces richness and increases productivity28. To eliminate confounding effects, common study designs in ecology require identifying, measuring, and statistically controlling for all confounding variables29. This task is daunting in natural ecosystems given myriad confounding variables that could influence both richness and productivity (e.g., land-use history, herbivory, disturbance). Yet, failure to control for all confounding variables can lead to inferences of the wrong sign or magnitude (i.e., due to statistical bias)26,30. Consequently, the mixed evidence on the effect of species richness on productivity in observational studies3,4,12,13,31,32 may reflect differences in the degree of control for confounding factors across studies. To isolate and quantify causal relationships between biodiversity and ecosystem function, the ideal study design would combine the strength of experiments in enabling causal inferences from correlations with the strength of observational designs in facilitating generalizable inferences about natural ecological processes. Experimental designs with more realistic extinction processes are one step in that direction (e.g.,16,20,21). Here, we develop a complementary approach by leveraging a global grassland dataset33 and methods designed for inferring causality from observational data26,27,34,35. This suite of methods now comprises the dominant approach to causal inference in fields outside of ecology, such as economics, medicine, and public health. When combined with our global longitudinal dataset, they allow us to account for the ecological complexity of grasslands without making strong assumptions about our ability to measure all confounding variables36, and they allow us to isolate the effect of biodiversity on productivity separate from the reverse relationship35,37. Applying traditional methods to our data, we would conclude that, on average, an increase in biodiversity increases productivity in grasslands, a result found in many prior studies. However, applying the suite of methods that control for a broader set of confounding variables, we come to the opposite conclusion: an increase in biodiversity reduces productivity in grasslands, on average. Results and discussion Study context and design We use repeated observations between 2007-2017 from 151 unmanipulated plots in 43 grassland sites in 11 countries33 from the Nutrient Network (https://nutnet.org), including mesic grasslands and prairies, savanna, desert grasslands, montane meadows, old fields, and alpine tundra (Table S1 in Supplementary Information (SI); SI Section 3). We define “productivity” as aboveground live biomass per year per 1m2 (following refs. 3,9,31). Each 1 m2 plot has between 1 and 37 species in a year, with an average of 11.3 (SD = 5.7) and median of 10. We use plots with five or more years of data, in contrast to most observational studies of biodiversity effects on productivity, which use a single year3,9,12,31,38. Data from multiple years offer three advantages: (1) an opportunity to study natural changes in richness; (2) enhanced generalizability; and (3) ways to control for a broad set of confounding variables, including unobserved ones (see “Methods”). Our study differs from prior ecology publications by combining three features: (1) causal diagrams to inform the design and transparently communicate the assumptions required for inferring a causal relationship from a correlation3,39; (2) regression models that leverage repeated observations on the same plots and sites to control for confounding variables (see Fig. 1B), both observable and unobservable26,30,36,40; and (3) rigorous assessments of the robustness of our inferences to violations of the assumptions required for inferring causal relationships from the data (these assumptions are described in more detail in Figs. 2 and 3, “Results”, and “Methods”). The SI offers a primer on these ideas and compares them to approaches widely used in ecology (SI Sections S2, S4, S7 and S9).Fig. 2 The effect of biodiversity on productivity, estimated as the mean % change in productivity from a 1% increase in richness. The top panel summarizes the assumptions needed for drawing causal inferences in each design. A red X on an arrow implies that the design blocks the confounding pathways described in the box with dotted lines. The Common Design, a multivariate mixed model that is common in ecological analyses of observational data, requires much stronger assumptions to interpret estimates as a causal effect than our Main Design, assumptions that are unlikely to be met in these data. The bottom panel shows the estimates with 95% confidence intervals for two designs: A Our Main Design, a panel data design (see SI Section S4) with n = 1231 observations; and B The Common Design, a multivariate mixed model (see SI Section S7) with n = 675. The positive estimated effect from the Common Design is not driven by having to drop sites that did not measure all the covariates (the sites in France, Portugal, and South Africa did not collect the soil data). If we use only the 675 observations from the multilevel modeling in our Main Design, we still obtain a negative estimated effect of richness on productivity, albeit less precisely estimated because of the smaller sample size (see SI Section S7). In prior observational studies of how richness affects productivity, controlling for the wide range of potential confounding variables in multilevel or structural equation models has posed challenges (e.g.,3,8,9,31,41). In those studies, researchers who wanted to interpret an estimated effect as a causal effect had to assume that no confounding variables were left out of the models30,36. In complex ecological systems, however, it is unlikely that one can measure all possible confounding variables. Moreover, when these confounding variables are measured with error, statistically controlling for them can introduce other biases36. In other words, prior studies require a strong assumption for interpreting the correlation between richness and productivity as causal: any site or plot attributes not included in the statistical estimation model are assumed to be uncorrelated with species richness and therefore not a source of statistical bias. Our design relaxes this strong assumption. We improve upon prior observational studies by controlling for a broader suite of confounding variables without needing to directly measure them as covariates (see “Methods” and refs. 36,40). To understand the intuition for how this control is possible, recall that, in contrast to most prior observational studies on this topic (e.g.,3,8,9,31,36), our multi-site data is longitudinal (‘panel data’) and thus includes variation in species richness in both time and space. Confounding variables that affect both richness and productivity could arise from conditions at the plot or the site. The values of these variables may be essentially invariant during the study period (e.g., soil texture, topography, land-use history) or they may vary through time (e.g., surrounding land-use change, drought conditions that differ by both site and year). With our multi-site panel data, we can directly control for time-varying, site-level conditions, whether they are observable or not, via a regression estimator that includes a simple interaction of binary variables for each site and year (see “Methods”, Eq. 2). Further, we can eliminate the confounding effects of time-invariant plot and site conditions by taking deviations from mean conditions, after which variables that do not change over time no longer have any explanatory role and thus are eliminated as a source of bias (“Methods”). Using alternative designs, we can also quantify the potential threat of additional sources of bias from unobserved, time-varying plot-level confounders and from reverse causality (by bias from reverse causality, we mean bias that could arise when a causal effect also runs from productivity to richness; see “Methods”). In contrast to our approach, virtually all observational analyses reviewed in4 omit important confounding variables (e.g., from human activities and land management) and ignore the potential for reverse causality (reviewed in ref. 42). To demonstrate how our study design builds on and advances prior research, we first apply two study designs that have been used in prior studies and then contrast them to our design. Specifically, we estimate a simple bivariate correlation of richness and productivity (like ref. 31) and then we estimate the relationship between richness and productivity using a multivariate design that mirrors advanced statistical designs that aim to control for confounding variables by directly measuring and including them as covariates in regression models (a “conditioning on observables” analysis, like ref. 6). The multivariate design, which we label “Common Design in Ecology,” controls for over 60 variables (far more than prior studies), including attributes of the soil, habitat, historical management, and weather (Table S10). More details are provided in “Methods” section. Main results We first report the bivariate correlation between-plot richness and productivity. Consistent with prior studies31, we find a statistically weak, positive relationship between richness and productivity when we do not control for any confounding variables: a 10% increase in richness is associated with a 1.4% increase in productivity, 95% CI [−0.6, 3.4]. To give that estimated correlation a causal interpretation requires an implausible assumption that there are no confounding variables in the system (or that they perfectly cancel each other out). Consistent with prior multivariate studies3, the Common Design in Ecology yields a statistically significant, positive relationship: a 10% increase in plot richness increased productivity by 3.8% on average, 95% CI [0.01, 7.5] (Fig. 2B). To give that estimated effect a causal interpretation, however, requires a strong assumption: all possible confounding variables are measured accurately and are included in the model. In contrast to prior analyses, our Main Design controls for a much broader set of potential confounders and comes to the opposite conclusion (Fig. 2A; Tables S2–S3; SI sections S4, S7 and S9). We find a 10% increase in plot richness decreased plot productivity by 2.4% on average, 95% CI [−4.1, −0.74]. The estimate is similar if we measure biodiversity using Simpson’s Diversity (Table S2), control for concomitant changes in species evenness (Fig. 2A), or measure species richness and productivity as untransformed variables in linear or non- linear specifications (Table S3). In extended analyses (Tables S4–S6), we find no evidence that the effect of species richness on productivity is moderated by the site’s productivity or total number of species (as in38). Switching from the Common Design to our Main Design flips the estimated effect of richness on productivity from positive to negative (Fig. 2). This sign-switching likely occurs for two reasons. First, on average, many of the observed site and plot variables at these 43 sites are negatively associated with richness and positively associated with productivity (or vice-versa). Controlling for them in the Common Design moves the estimated effect in the positive direction from the bivariate correlation. Second, unobserved site and plot variables (Us, Up, Ust in Fig. 1B) are, on average, positively associated with both richness and productivity. We can infer the sign of these associations by observing how the estimated effect changes with and without the controls for unobserved time-varying, site-level conditions and time-invariant plot and site conditions (Fig. 2A versus Fig. 2B; see “Methods”). Failing to control for the time-varying confounders is a particular problem in the Common Design. In other words, the Nutrient Network sites experience site-specific “shocks” that vary each year (e.g., weather shocks, like a particularly dry April, or herbivory shocks, like higher herbivore pressure than the prior year) and failing to control for them creates statistical biases in the positive direction. We cannot observe the exact components of these shocks, but because we observe the same sites over many years, we can control for them. The Main Design, with its greater set of controls, is thus less biased36. More details are available in SI Sections S7 and S9. Future research could elucidate what shocks are most relevant, thereby providing a way for researchers without longitudinal data to potentially control for the confounding effects of these shocks. Results are robust to alternative assumptions for inferring causality A hallmark of modern approaches to causal inference is to probe the robustness of results to potential violations in the assumptions used to infer causality from correlation35. Using four additional approaches, we use assumptions different from those made in our Main Design (Fig. 2A) and assess how our conclusions change (see “Methods”). In all four approaches, the estimated effect of richness on productivity is negative (Fig. 3; Tables S7–S8).Fig. 3 The effect of species richness on productivity (robustness checks), estimated as the mean % change in productivity from a 1% increase in richness. The bottom left panel shows the estimates of mean effect with 95% confidence intervals, except for the sensitivity test estimate, from (1) our main study design (n = 1231), (2) a dynamic panel design (n = 1063), (3) a sensitivity test (n = 1231), (4) a mechanism-blocking design (n = 1063), and (5) an instrumental variable design (n = 1212). The diagrams summarize the assumptions needed for drawing causal inferences in each design, where a red X on an arrow implies that the confounding pathway is blocked by the design. Using four approaches that make assumptions that differ from the assumptions in our Main Design (Fig. 2A), we find no evidence for a positive effect of species richness on productivity. Based on the first two approaches, we conclude that we are not mistaking the effect of productivity on richness3,15 for the effect of richness on productivity (“reverse causality”). The first approach employs an instrumental variable design, which uses an observable source of variation in richness (Z in Fig. 1B) that is assumed to have no connection to productivity after conditioning on the site and plot variables addressed in the Main Design (see “Methods”). When this assumption is valid, the design addresses both reverse causality and all forms of confounding in Fig. 1B, at the cost of drawing inferences from only a subset of the data, which can dramatically decrease the precision of the estimate. Our second approach assumes, based on3, that a negative effect of productivity on richness would be mediated (M in Fig. 1B), at least in part, by shading or factors for which shading is a proxy (e.g., overcrowding). To block the effect of this mechanism, we add a shading variable to the Main Design (see “Methods”). If the estimated effect changes, reverse causality may be a source of bias. In both approaches, the estimated effect remains negative, suggesting that, if either of the approaches’ assumptions are valid, reverse causality is not driving our results. This conclusion does not mean productivity cannot affect richness, only that such a relationship is not a likely source of bias in our Main Design. Based on the final two approaches, we conclude that the estimated negative effect in the Main Design is robust to potential biases from unobserved confounding variables at the plot level that vary over time (Upt in Fig. 1B). First, in a dynamic panel design, we address bias that would arise if the prior year’s productivity affects richness and productivity in the current year (e.g., via soil fertility43). The estimated effect is similar to the estimate from the Main Design. Second, in a sensitivity analysis34, we assess how the Main Design estimate would change if there were a strong, unobserved confounding variable that is negatively associated with species richness and positively associated with productivity (e.g., measurement error or plot-level drivers of disturbance). If such a confounding variable were to exist, it could create a spurious negative correlation between richness and productivity. The analysis implies that, even in the presence of a such an unobserved confounder, we would still infer that there is a negative relationship between richness and productivity. The role of rare species and non-native species In contrast to our study, many experimental studies report positive effects of richness on productivity1,2,7,13. One difference between experimental and natural systems is that most species in natural ecosystems are rare, whereas most species planted in experiments are not rare (Figs. S11 and S12). Rare and dominant species can affect productivity differently44. Thus, the effect on productivity from an increase in richness (e.g., from 4 to 8 species) could differ when the additional species are rare versus not rare. In the jargon of the causal inference literature, richness is a compound treatment with multiple versions, or a “heterogenous treatment”45. (Fig. 4A). Another difference between experimental and natural systems is the number of non- native species, which are absent in many experiments but increasingly prevalent in real ecosystems25. If non-native are more competitive but less productive, as in ref. 46, this could also explain the divergence between our results and those of experimental studies.Fig. 4 Composition and its role in the effect of species richness on productivity. This figure illustrates why results from observational and experimental studies may differ. Composition, the identities of species that could potentially grow at a site at a given level of richness, makes species richness a heterogeneous treatment in both experimental and non- experimental systems (A). The way richness changes in nature can thus influence how changes in richness affect productivity (B). In B (left), Richness has changed from 2 to 3 species in case (i), a case where rare species take space formerly occupied by more productive dominant species. In case (ii), where Richness also has changed from 2 to 3 species, rare species have negative effects on dominant species productivity (e.g., via below-ground competition or allelopathy). B (right) A species rank abundance curve that illustrates that most species in high diversity grasslands are rare. Thus, as diversity changes from low values to higher values, the way in which rare species affect productivity will become more influential in affecting productivity levels. For simplicity, figure focuses on a contrast between rare and dominant species, but the ideas can also apply to differences between native and non- native species. Plant species images in Panel B are from Tracy Saxby, IAN Image Library (https://ian.umces.edu/imagelibrary). Rare and non-native species could reduce productivity through multiple channels (Fig. 4B, left). For example, these species may compete with more productive species (e.g., via allelopathy of rare invaders47). Further, they may produce less aboveground biomass than common, native species44 and so when they enter a plot, they may take space formerly occupied by more productive species. These productivity-reducing effects would be strengthened if, as richness increases in a plot, rare and non-native species are more likely to be the incrementally added species (Fig. 4B, right). In experimental systems, species enter plots with equal probability. In contrast, as richness increases in our 43 grassland sites, the probability that the incremental species is a rare or non-native species also increases (Figs. S11 and S12). Given differences in the species pools studied in our natural systems versus many experimental systems, we explore whether changes in rare species and non-native species richness affect productivity differently from changes in native, non-rare species richness. We classify species into four categories: (1) rare, native species; (2) non-rare, native species; (3) non-rare, non-native species; and (4) rare, non-native species (see “Methods”). We then estimate the effect of each category’s richness on productivity using our Main Design. Our results imply that the negative average effect of richness on productivity in Fig. 2 is driven by changes in the numbers of rare, native species and non-native, non-rare species (Fig. 5). Consistent with results from experimental studies, an increase in species richness that came from a non-rare, native species increased productivity. But increases in richness decreased productivity when these increases came from non-rare, non-native species or rare, native species (inferences are similar using different definitions for rarity; Tables S11–S15). We acknowledge that there could also be positive and negative interactions across these species’ types, but we do not have sufficient statistical power to explore these potential interactions.Fig. 5 Estimates of the mean effect of species richness on biomass production conditional on species type. All estimated mean effects are on a log-inverse-hyperbolic-sine scale and shown with 95% confidence intervals and with n = 1175 within 42 sites. Given the inverse hyperbolic-sine transformation of the richness variable, the estimated effects cannot be interpreted as elasticities without further manipulation, but their signs and relative magnitudes can be compared. We can reject the null hypothesis that the estimated effects of these four types of species are equal (ChiSq = 9.82, Pr(ChiSq = 0.02)). Dropping observations from one site without data to define the species types does not change our estimates in Figs. 2 and 3. We conjecture that the proposed mechanisms through which richness positively affects productivity in archetypical experiments—i.e., niche complementarity and positive selection2,14,15—may operate primarily among non-rare, native species. Testing this conjecture would require experiments that successfully grow representative proportions of rare species (in experiments that planted rare species, these species failed to consistently emerge, see SI Section S9 – Box 2). Implications for experimental and observational biodiversity research Leveraging methodological advances for causal inference in observational designs, our study uncovers ecological relationships in grasslands that deserve closer attention. In the 43 grassland sites in our study, an increase in species richness decreases productivity on average (Fig. 2). This effect appears to arise because an increase in rare species and non-rare, non-native species decreases productivity on average (Fig. 5) and these species comprise most species in an ecosystem (Fig. 4). These effects will be missed in observational designs that do not adequately control for a wide range of confounding factors and in experiments that do not plant a representative mix of species in diversity patterns that occur in natural systems. Our results also highlight the challenge of determining the representativeness of experimental systems. For example, a recent comparison of natural and experimental systems identified many similarities in attributes but did not assess whether the patterns of rare or non-native species in experimental systems match the patterns in natural systems48. Our results point to promising areas for future research, including studies that experimentally manipulate rare species and non-native species and seek to identify traits of these species that drive their effects on productivity. For example, in a recent study22, researchers experimentally removed non-dominant species from randomly assembled communities and reported an increase in biomass one year after the removal; a result consistent with our results. Extensions of their study can help elucidate the traits of these species that drive their effects on productivity. Our results also imply that any estimated effect of changes in species richness on productivity may not generalize to different spatial or temporal scales, or to other ecosystems, forms of biodiversity, or ecosystem functions. Multiple ecological mechanisms underlie a relationship between richness and productivity (e.g., below-ground competition, niche complementarity), and their strength may vary across places and time depending on which types of species are changing and how (Fig. 4). When the operative mechanisms depend on the version of richness that is changing, ecologists face what has been called “treatment-variation relevance”45 or “consequential variation of the treatment”49. In such cases, interpreting and generalizing causal effects is challenging, whether ecologists use experimental or observational designs. The challenge is best met by using large observational datasets that capture the types of species changes that occur in nature and then, using methods like the ones in our study, determine which changes – in terms of which species are changing and how (Fig. 5) – are consequential for the effect of richness on ecosystem function. The results of these observational studies could then guide experimentalists in selecting experimental designs that can help confirm the results from observational studies and elucidate the underlying mechanistic processes. Most importantly, our study extends prior research3 that highlights the importance of study design in credibly isolating causal relationships in natural ecosystems. Other fields have made important advances in observational analyses — advances that have not yet permeated into ecology and other natural sciences. By demonstrating how to apply these advances to an ecological question and data, our study aims to spur broader adoption of these advances in ecology. Given the challenges of randomizing all the important elements of ecosystems at larger spatial and temporal scales, observational designs like ours that leverage these advances offer important complements to experimental designs in research to elucidate how natural ecosystems function. Methods To ensure reproducible results, we implemented all analyses in two software programs (R using the ‘fixest’ package v 0.8.2 and Stata v.16) and multiple researchers confirmed the results. The code for reproducing all analyses, figures, and tables in this study are available through Zenodo (10.5281/zenodo.7675340). A RMarkdown tutorial on the main methods can also be found on our Zenodo release (10.5281/zenodo.7675340) and as Supplementary Data. Target causal effect To formalize the causal relationship we seek to estimate, we use the potential outcomes framework30,50,51. The causal effect of a change in richness from R′ to R″ on productivity P in plot i is defined as [Pi(R″) − Pi(R′)], where Pi(R″) is the potential productivity outcome when R = R″ and Pi(R′) is the potential productivity outcome when R = R′ (R′ ≠ R″). The difference in these two potential productivity outcomes (i.e., productivity under two potential richness conditions) is the causal effect of a change in richness in a plot. For a specific location and time, only one of these potential outcomes will be directly observable; the counterfactual values for the other potential outcomes must be estimated from data. The average causal effect of a change in biodiversity from R′ to R″ across all plots is E[Pi(R″) − Pi(R′)], where E[·] is the expectation operator. We seek to estimate the average causal response of an incremental change in R across all plots (i.e., the average effect across all possible one-unit changes). When used for causal inference, non-experimental studies aim to replicate, conceptually, the idealized experimental design in which the factor or factors that affect variation in R only affect P via their effects on R. In other words, to permit credible causal inferences, a non-experimental design seeks, via design and statistical methods, to eliminate the confounding effects of Up, Us, Ust, and Upt in Fig. 1, as well as the effects of reverse causality in Fig. 3. Data description Study sites include mesic grasslands and prairies, savanna, desert grasslands, montane meadows, old fields, and alpine tundra from 11 countries. To measure productivity, we use plant above- ground live mass (biomass) (see Fig. S1). Biomass production supports many ecosystem processes and services and this measure of productivity has been widely used in studying the relationship between diversity and productivity with both observational (e.g.,3,9,31) and experimental data (reviewed in ref. 4,52,53). For herbaceous vegetation, aboveground live biomass provides a reasonable estimate of primary productivity54. Biodiversity measures are determined from species cover data from the Nutrient Network (SI Section 3a). Common design estimator To show how our Main Design differs from more common designs in ecology, we constructed what we call a “Common Design in Ecology”: a multivariate design that controls for over 60 variables (far more than prior studies), including attributes of the soil, habitat, historical management, and weather (Table S10). This design captures the strong assumption that is inherent in prior observational ecological studies that aim to estimate the causal effect of richness on productivity: there are no variables omitted from the statistical model that are correlated with both richness and productivity. When this assumption is not met, the design suffers from bias. We compare this Common Design to our Main Design, which relaxes that strong assumption. For the Common Design in Ecology, we estimate the effect of richness on productivity using the following regression equation:1 lnLiveMasspst=βlnRichnesspst+Xp+Xs+Xst+εpst. We use a ln-ln model specification (for rationale, see SI Section S3b). Recall that observations in our data come from a plot p located within a site s in a year t. Thus, Xp is a vector of plot-specific attributes that do not vary over the study period (e.g., soil type), Xs is a vector of site attributes that do not vary over the study period (e.g., habitat, historical management, elevation), and Xst is a vector of site attributes that vary by year (e.g., temperature seasonality, maximum and mean temperatures of the warmest month). Together, these vectors include over 60 variables, which are directly controlled for in the regression (see Table S10). In this equation, we can see that the effects of any omitted variables on productivity (i.e., variables not controlled for in the X vectors in the model) reside in the error term εpst. We can rewrite this error term as a combination of a random error, σpst, which only affects productivity, and unobserved confounding variables, U, which affect both richness and productivity at either the plot or site level and either in all years or only some years. Thus the error term can be rewritten as εpst = Up + Us + Ust + Upt + σpst, where Up and Us are vectors of plot and site-level variables that do not change over the study period, and Ust are vectors of time-varying site-level variables, and Upt are vectors of time-varying plot-level variables. If a study design has any of these omitted U variables, the estimator (β) would be biased—known as omitted variables bias. In other words, to interpret the estimate of β as an estimate of our target causal effect would require one to assume that the observed covariates in the model capture all relevant Up, Us, and Ust and Upt does not exist (because no time-varying covariates are measured at the plot level). Main design estimator Each observation in our study comes from a plot p located within a site s in a year t. With this longitudinal data structure, i.e. with repeated observations of the same plots, one can control for all dimensions of confounding variables that do not vary over the study period and all dimensions of time-varying site-level confounding variables without having to observe all of these dimensions. To achieve this control, we estimate an equation of the following form:2 lnLiveMasspst=βlnRichnesspst+δp+μst+εpst As noted above, we use a ln-ln model specification (for rationale, see SI Section S3b). Given that we have a ln-ln specification, β can be interpreted as an elasticity: the expected percent change in productivity given a one percent change in richness. We also tested the robustness of results to this modeling decision (see SI Section S5). The time-invariant plot attributes (δp) are modeled in a fully flexible way that allows each plot to have its own effect on productivity (details on estimation procedure below). In the Economics literature, δp would be called “plot-level fixed effects.” Note that the phrase “fixed effects” has a different meaning in economics than in ecology (see S1 Glossary). In economics, including δp is said to control for “unobserved heterogeneity” across plots that can be a potential source of bias. Note that δp is not part of the error term, as it would be in mixed (multilevel) models55 or in a Common Design as in Eq. 1 (i.e., Up and Us). Rather, it is a parameter to be estimated, just like β. In other words, β and δp are assumed to be fixed and estimable, rather than assumed to follow a distribution. Time-invariant site attributes are not explicitly included in the equation because they are subsumed into the time-invariant plot attributes (i.e., plots are nested within sites and so fixed site attributes are controlled via fixed plot attributes). In other words, this variable captures all attributes of a plot at a given site that vary little over the study period. Thus, δp captures both Up and Us in the decomposition of the error term from the previous section (i.e., εpst = Up + Us + Ust + Upt + σpst, where σpst is a random error that only affects productivity and U are unobserved confounding variables that affect both richness and productivity). To show how the estimator in Equation 2 can efficiently control for time-invariant confounders, we subtract the productivity observation within a plot in one year (t − 1) from the productivity observation within the same plot in the next year (t), yielding an equation for the change (Δ) in productivity from one year to the next:3 ΔlnLiveMasspst=βΔlnRichnesspst+Δμst+Δεpst The variable δp, which captures the effects of time-invariant plot attributes, has been differenced away, allowing for efficient estimation of β; in other words, we control for time-invariant plot attributes without having to estimate them and use up many degrees of freedom. In this differenced version of Equation 2, one can see that we are estimating the effect of richness on biomass from changes in richness within plots, where the confounding effects of between-plot differences are absent. Rather than first-differencing the equations to eliminate δp, one can instead take deviations from plot-level means, which is the approach we take to estimating Equation 2 because it can be more efficient. Thus, we can also describe our estimation strategy as estimating a correlation between deviations of productivity around its mean and the corresponding deviations in richness around its mean. The time-varying site attributes (μst) are also modeled in a fully flexible way that allows a year- specific effect for each site (in the estimation, an indicator for each year is interacted with an indicator for each site). Explicitly estimating μst flexibly controls for confounding variation due to conditions at a site that change through time, such as weather (e.g., time-varying patterns of temperature and precipitation), herbivory, and surrounding land management conditions. In other words, this variable captures all year-specific conditions experienced by every plot at a given site. Thus μst captures Ust in the decomposition of the error term from the previous section (εpst = Up + Us + Ust + Upt + σpst). The estimator in Eq. (2) is often called a “two-way fixed effects estimator” because, by taking deviations from the means, one controls for time- invariant confounding and, by including the site-by-year effects, one controls for time-varying confounding. Thus, in contrast to the Common Design in Ecology, we control for a broad suite of plot-level and site-level confounders without having to measure them directly. The Main Design also controls for non-linear relationships between the confounding variables and productivity or richness, as well as linear and non-linear interactions among those variables. Not having to measure the confounding variables also yields another benefit: if the observable confounders were measured with error and that error were correlated with the measure of richness, the Common Design would have another source of bias. Moreover, because the model specification comprises only the richness variable and a set of binary indicator variables and their interactions, the risk of misspecification bias from how confounders are modeled is lower. To better understand how our design differs from more common designs in Ecology such as mixed-effect modeling approaches and convergent cross-mapping approaches56, see SI Sections S4c and S7. The term εpst in Eq. (2) is a time-varying random error term at the plot level, assumed to have mean zero and no correlation with ln Richness, i.e., it corresponds to Ipst in Fig. 1B. These plot-level errors may be serially correlated (i.e., temporally dependent even after conditioning on richness and site-by-year effects), and thus we cluster the standard errors at the plot level57. Our clustered estimation of the variance allows for arbitrary serial correlation within each plot, as well as heteroskedasticity across plots36,58. Errors at a given site may also be correlated (even after conditioning on site-by-year effects) and thus, as a robustness check, we also estimate standard errors clustered at the site level (Table S2). Our Main Design has weaker assumptions than the Common Design in Ecology, but both have one assumption in common: there are no unobserved time-varying plot-level confounders in the error term (no Upt). In other words, we assume that, after controlling for time-invariant plot and site attributes that are correlated with richness and productivity, and time-varying site attributes that are correlated with richness and productivity, the remaining temporal variation in richness in a plot is “as if randomly assigned,” independently across time. This assumption is equivalent to assuming that the remaining variation in richness is driven by variables that have no link to productivity other than through their effect on richness (i.e., Zpst in Fig. 1) and thus there is no correlation between Δln Richnesspst and Δεpst in Eq. (3). If our assumption is correct, we can give a causal interpretation to the estimate of β. Unlike prior ecology studies, however, we assess the sensitivity of our results to violations of this assumption (see Robustness Checks next). Robustness checks: modifying the main design Robustness checks: model specifications In the SI (Table S3), we present the results from variations in the specification of Eq. (2): (1) we include a control for species evenness; (2) we change the measure of diversity from species richness to Simpson’s Diversity index; (3) we include the lagged effect of species richness in the prior year (ln SpeciesRichnesst-1); and (4) we vary the functional form by (i) taking the natural logarithm of productivity but using the untransformed richness values, (ii) using both untransformed richness and untransformed productivity values, and (iii) using untransformed richness and untransformed productivity values in a non-linear, quadratic specification (i.e., we include ln SpeciesRichnesst and ln SpeciesRichness2t). Robustness checks: causal assumptions As noted above, the key, untestable assumption for drawing a causal inference from the estimator in our Main Design is the following: after controlling for time-invariant plot confounders and time-varying site confounders, the remaining factors that drive changes in richness only affect productivity via their effects on richness. We consider potential violations of this assumption and the implications for our inferences -- i.e., whether our conclusions could change -- by conducting a series of analyses that rely on alternative assumptions for causal inference (Fig. 3). Instrumental variable design for unobservable confounders and reverse causality First, we explore the potential violation of our assumption that the effect we are estimating goes from richness to productivity, and not the other way around. Richness and biomass measures in our data are taken simultaneously each year, as they are in most ecological datasets. Thus, we cannot rely on temporal sequencing of the data to rule out reverse causality. To assess the potential threat of reverse causality, we adopt a statistical approach that is common in economics and public health, but rare in ecology: an instrumental variable design59–63. When its underlying causal assumptions are valid, this design allows researchers to eliminate not only the influence of reverse causality but also the influence of unobservable confounders, both static and dynamic. To implement this design, we need to measure an attribute of the system that has a relationship with richness, but, after conditioning on the other plot and site attributes in Equation 2, has no relationship with productivity other than through its relationship with richness. Such an attribute is conceptually illustrated by the variable Z in Figs. 1, S4 and S5. In economics and biostatistics, Z is called an instrumental variable (IV) or a surrogate variable. An example of a potential IV is randomization of planted richness by an experimenter. In field experiments, randomization of richness helps isolate the causal effect of richness on productivity, but only when the randomization affects productivity in a plot solely through its effect on richness, an assumption called excludability or the exclusion restriction64. In other words, one must assume there is no arrow going from Z directly to P in Fig. 1. In the absence of randomization, one must use theory and experience to identify a naturally occurring IV (reviewed in ref. 62). Each of the plots in our sample are unmanipulated plots that are embedded in blocks of manipulated plots in the Nutrient Network. In other words, each unmanipulated plot in our sample is surrounded by a set of plots with experimental nutrient additions (see ref. 65). These manipulated experimental plots received randomized amounts of nutrient additions, which subsequently affected the experimental plots’ richness66. We assume that the experimentally manipulated richness in these plots can also affect the richness in unmanipulated plots in the same block through ecological dispersal channels but does not affect the productivity of these unmanipulated plots except through the effect on the plots’ richness (an assumption made more plausible by the randomization of nutrients in the neighboring plots). If that assumption is correct (called an “excludability assumption”), we can use the average richness of an unmanipulated plot’s neighboring manipulated plots in the same block as an IV for richness in the unmanipulated plot. The SI Section 6bii provides justification and further discussion of this IV. The cost of using the IV design is that we can only estimate the average effect on productivity for the subset of the changes in richness that are affected by the IV. This subset is comprised of what are called “compliers” – plot-year observations for which the richness value would have been different had the average richness in surrounding plots been different. Thus, the IV design has much lower statistical power than our Main Design62,67. To implement the IV design, we use a two-stage, least squares estimator26:4 FirstStage:lnRichnesspst=γIV+δp+μst+Ppst 5 SecondStage:lnLiveMasspst=βlnRic^hnesspst+δp+μst+εpst In the first stage (Eq. 4), we predict richness, and, in the second stage, we use the predicted values of richness to estimate the effect of richness on productivity (see Table S8 for the results from both stages). We can reject the null hypothesis of a weak instrument using the Montiel-Pflueger effective F-statistic, which is a test that is robust to heteroscedasticity, serial correlation, and clustering68;. For further discussions of the IV design and its assumptions, see SI Section 6b.ii. Blocking a mechanism for reverse causality As an alternative approach to address the potential threat of reverse causality in our design, we posit a mechanism through which productivity affects richness: shading (based on3,69). Although productivity could affect species richness through non-light pathways, such as soil resource use, the effect of productivity on richness is expected to be, at least in part, mediated by reductions in light from increased biomass that, in turn, reduces richness in a plot69. As an estimate of shading, we measure the fraction of photosynthetically active radiation (e.g., light used by plants) that reaches the soil. See SI Section S7b for details. If the estimated negative relationship between richness and productivity in Fig. 2 were an artifact of reverse causality, then putting our shading variable in Equation 2 as a covariate would block the effect of productivity on richness that arises via shading. The sign of the coefficient on richness (β) would then become positive (or small and statistically insignificant if the true relationship between richness and productivity were zero). Yet the estimated effect remains unchanged (Fig. 3). If shading were not an important mechanism through which productivity would affect richness in our sample, or if our measure of shading is a poor measure of the shading mechanism, our mechanism-blocking design would fail to quantify the potential threat of reverse causality. For this reason, we also implement an instrumental variables design, described above, that makes different assumptions to account for reverse causality. Bracketing the “True Causal Effect”: accounting for potential bias from dynamics The IV design not only addresses reverse causality but it also addresses all forms of dynamic, plot-level confounding variables (e.g., past productivity). However, it relies on untestable assumptions that may not be satisfied (e.g., excludability assumption). To supplement that analysis, we also explored a range of potential sources of bias from dynamic confounders. Here, we report on the methods used for one of these analyses, with results shown in Fig. 3. The other analyses and methods are reported in the SI Section S6. In this analysis, we consider the possibility that prior productivity affects both current richness and productivity. We re-estimate the effect of richness on productivity using a lagged-dependent variable (LDV) design70, which relies on different causal assumptions for identifying a causal effect. The Main Design assumes that the relevant confounders are time-invariant over the study period, or they vary over time at the site level rather than the plot level (e.g., site and plot-level differences in evolutionary history, age in community assembly, grazing intensity at a site, and history of disturbances and recovery stage in each plot). Instead, the LDV design considers: “What if current species richness and productivity were determined by last year’s productivity, in addition to, or instead of, site-level conditions varying through time (e.g., precipitation)?” The LDV design, in contrast to the Main Design, assumes that the relevant confounders vary over time at the site level and, at the plot level, their static and dynamic effects can be controlled by controlling for past productivity (in other words, the effects of confounders are mediated directly and indirectly through prior productivity at the plot level). To achieve this control, we estimate an equation of the following form:6 lnLiveMasspst=βlnRichnesspst+θlnLiveMassps(t−1)+μst+εpst Under certain conditions, the estimated effects of richness in our Main Design and in the LDV design “bracket” the true causal effect26,71. If the assumptions of the Main Design are valid, but the LDV design are invalid, the estimate from the LDV design provides an upper bound estimate. If the assumptions of the LDV are valid, but the Main Design are invalid, the estimate from the Main Design provides a lower bound estimate. As observed in Fig. 3B, this bracketing exercise implies the true effect is negative. Sensitivity test: would unobserved confounding variables change our conclusions? To further explore the potential effect of violations in our assumption that there are no time- varying plot attributes that are systematically correlated with richness and productivity, we explore how our estimated effect would change if there were an unobserved confounder that was negatively correlated with richness and positively correlated with productivity (i.e., a source of bias that would yield a spurious negative causal relationship between richness and productivity in our design). Said another way, this analysis answers the question, “How much correlation between the unobserved variable and the richness and productivity variables would be sufficient to change our conclusions?” We applied a sensitivity test following the method introduced by Altonji et al.72 and further developed by Oster34. More details on the method are in SI Section S7a and Table S7. We set π = −0.10 and Rmax = 1, which would mimic a powerful potential unobserved confounder in our design: a confounder that is so strongly correlated with productivity and richness that, were we able to observe it (along with the other variables in the equation), we could predict with near certainty which of two plots would have higher productivity and which would have higher richness. Estimating the effect of richness on productivity with those implausible parameter values yields an upper bound on the impact of this confounder on the estimated effect of richness on productivity in our Main Design. The estimated upper bound is still negative: a 10% increase in richness implies a 2% decrease in productivity, on average. In other words, in the presence of an unobserved confounder that is negatively associated with richness and positively associated with productivity relationship (thus creating some spurious negative correlation between richness and productivity), we would still infer that there is a negative relationship between richness and productivity. To infer a positive relationship between the two variables would require an infeasible value for π: it requires π > 1, which implies the confounder would have to be more influential in explaining variation of productivity than the plot-level, time-invariant attributes and the site-level, time-varying attributes that are in Equation 2. For completeness, we also considered an unobservable confounder that was positively associated with both richness and productivity and thus could be masking some of the negative effect of richness on productivity (i.e., positive selection bias). In other words, we also calculate a lower bound on the estimated effect by setting π = 0.10 (see SI Section S6b.ii). Testing hypotheses about moderators of the causal effect In the SI (Tables S4–S6), we present results from hypothesis tests about moderators of the plot-level richness effect on productivity. The potential moderators are: (1) the average level of productivity at a site (i.e., does the effect of richness on productivity differ between high versus low productivity sites, as reported by38?); and (2) the average level of richness at a site (i.e., does the effect of richness on productivity differ between high versus low richness sites, as hypothesized by73?). To conduct these tests, we expanded Equation 2 by adding an interaction term between ln Richnesspst and the moderator variables. We measured site-level productivity in four ways, which vary by the discreteness of the measure and by the way time is incorporated into the measure. In terms of discreteness of the measure, we measured productivity both as a continuous variable and, using classifications from Wang et al.38, as categorical variables for high, medium, and low productivity. Wang et al.38 used cross-sectional analyses to study this moderator. Because we have longitudinal data, we can measure the continuous and categorical measures of site-level productivity in two ways: average productivity per site over the entire time series and site-level productivity per year. More details on the motivations for selecting these moderating variables for analysis can be found in the SI (Section S5). Exploring effect of species richness on biomass conditional on species type We assign the labels “rare” and “non-rare” based on relative abundance at each site, and species’ origin was origin was determined by the site coordinators in the Nutrient Network (SI: Section S8). We define relative abundance based on relative aboveground cover. We use relative cover as our metric for abundance because we believe it better captures the range of mechanisms through which rare species may decrease productivity, including taking space formerly occupied by more productive species. However, we test the sensitivity of our results to this decision by also using a relative frequency metric to define rarity (see SI Section S8c.ii). In Section S8c.iii, we also test the sensitivity of our conclusions to different cutoff values for assigning a species to the “rare” and “non-rare” categories. We modify and use the Main Design for each way of defining the four species groups. Finally, using Chi Squared tests, we tested the null hypothesis that the species richness of the groups had the same effects on live biomass. Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Supplementary information Supplementary Information Peer Review File Supplementary Dataset 1 Reporting Summary Supplementary information The online version contains supplementary material available at 10.1038/s41467-023-37194-5. Acknowledgements We thank E. Seabloom, A. Asmus, H. Correa, J. Firn, and F. Isbell for their input. The views herein are solely those of the authors and do not necessarily reflect those of the Federal Reserve Bank of Philadelphia or the Federal Reserve System. We thank UMN Supercomputer Institute for hosting data and IoE for hosting Network meetings. We acknowledge support from NASA BioSCape to L.E.D. and P.A. We acknowledge support from SF Long- 354 Term Ecological Research (LTER) Network Communications Office and US NSF 823 DEB- 355 1545288. This study used data from the Nutrient Network (http://www.nutnet.org), funded at the site-scale by individual researchers. Coordination and data management was supported by funding to E. Borer and E. Seabloom from NSF-DEB-1042132 and Institute on the Environment (DG-358 0001-13) and to E.B., and P.B.R from LTER (DEB-1234162, DEB-1831944). We also acknowledge support from DEB-0620652, NSF grants DEB-1242531, DEB-1753859, and DBI-202189 to P.B.R.; TULIP Laboratory of Excellence (ANR-10-LABX-41) to M.L.; and CEREEP-Ecotron Ile De France (CNRS/ENS UMS 3194), Regional Council of Ile- de-France (DIM Program R2DS I-05-098/R), GoF/ANR’s Investissements d’Avenir program (ANR-11-INBS-0001 AnaEE France; ANR-10-IDEX- 0001-02 PSL) to X.R. This is Kellogg Biological Station Contribution no. 2338. Soil analyses were supported, in part, by Oregon State University, University of Minnesota, and USDA-ARS grant 58-3098-7-007 to E.T.B. Author contributions L.E.D. led the paper. L.E.D and C.N.S. initiated the study concept, developed the study questions, and conceived the outline of study. L.E.D., C.N.S., and P.J.F. designed the study and conducted the data analyses. L.E.D. and P.J.F. wrote the paper and Supplementary Information, performed the revisions, and designed and made the figures. L.E.D. and K.A.K. processed and cleaned the data with input from the Nutrient Network and code checking provided by the coauthors listed below. C.N.S., E.T.B., M.L., P.B.R., J.C., K.A.K., A.T.C., J.E.K.B., K.B., A.J.W., A.S.M, and A.H. contributed to the conception and framing of the paper, the interpretation, and writing. C.N.S., A.T.C., K.K., J.M, R.L.C., and S.A.P. contributed to the Supplementary Information. C.N.S. and A.T.C. contributed to revisions/response to reviewers. E.T.B., K.D., P.A., A.M., YH, J.M.K, T.O., M.D.S., L.S., C.S., K.K., J.E.K.B., A.H., X. R., S.A.P., J.W.M., C.A.A., and J.L.M. contributed data and edited and commented on the manuscript. E.B. planned and coordinated experimental network. A.T.C., K.A.K., C.A.A., T.O., X.R., C.N.S., and L.E.D. checked R and/or STATA code for reproducibility. L.E.D. and C.N.S. created the supplementary RMarkdown tutorial and GitHub repository. Peer review Peer review information Nature Communications thanks Jakob Runge, and the other, anonymous, reviewers for their contribution to the peer review of this work. Peer reviewer reports are available. Data availability The processed data used in this study have been deposited on Zenodo under 10.5281/zenodo.7675340. The raw data for unmanipulated plots that were not included in the analyses, because they did not meet the inclusion criteria, are available under restricted access for which permission can be obtained by contacting the Nutrient Network at https://nutnet.org. Code availability All analysis code and output are available through our GitHub project site https://github.com/LauraDee/NutNetCausalinf and are released on Zenodo (10.5281/zenodo.7675340). All code for data processing, including of the raw data, main analyses, and supplemental analyses is available. Competing interests The authors declare no competing interests. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Change history 7/12/2023 A Correction to this paper has been published: 10.1038/s41467-023-39743-4 ==== Refs References 1. Hector A Plant diversity and productivity experiments in European grasslands Science 1999 286 1123 1127 10.1126/science.286.5442.1123 10550043 2. Tilman D Diversity and productivity in a long-term grassland experiment Science 2001 294 843 845 10.1126/science.1060391 11679667 3. Grace JB Integrative modelling reveals mechanisms linking productivity and plant s pecies richness Nature 2016 529 390 393 10.1038/nature16524 26760203 4. van der Plas F Biodiversity and ecosystem functioning in naturally assembled communities Biol. Rev. 2019 94 1220 1245 30724447 5. Tilman D Isbell F Cowles JM Biodiversity and ecosystem functioning Annu. Rev. Ecol. Evol. Syst. 2014 45 471 493 10.1146/annurev-ecolsys-120213-091917 6. Loreau M Biodiversity and ecosystem functioning: current knowledge and future challenges Science 2001 294 804 808 10.1126/science.1064088 11679658 7. Hooper DU Effects of biodiversity on ecosystem functioning: a consensus of current knowledge Ecol. Monogr. 2005 75 3 35 10.1890/04-0922 8. Liang J Positive biodiversity-productivity relationship predominant in global forests Science 2016 354 196 10.1126/science.aaf8957 9. Duffy JE Godwin CM Cardinale BJ Biodiversity effects in the wild are common and as strong as key drivers of productivity Nature 2017 549 261 264 10.1038/nature23886 28869964 10. Cardinale BJ Biodiversity loss and its impact on humanity Nature 2012 486 59 67 10.1038/nature11148 22678280 11. Isbell F High plant diversity is needed to maintain ecosystem services Nature 2011 477 199 202 10.1038/nature10282 21832994 12. Grace JB Does species diversity limit productivity in natural grassland communities? Ecol. Lett. 2007 10 680 689 10.1111/j.1461-0248.2007.01058.x 17594423 13. Balvanera P Quantifying the evidence for biodiversity effects on ecosystem functioning and services Ecol. Lett. 2006 9 1146 1156 10.1111/j.1461-0248.2006.00963.x 16972878 14. Tilman D Lehman CL Thomson KT Plant diversity and ecosystem productivity: theoretical considerations Proc. Natl Acad. Sci. USA 1997 94 1857 1861 10.1073/pnas.94.5.1857 11038606 15. Loreau M Biodiversity and ecosystem functioning: a mechanistic model Proc. Natl Acad. Sci. USA 1998 95 5632 5636 10.1073/pnas.95.10.5632 9576935 16. Smith MD Knapp AK Dominant species maintain ecosystem function with non- random species loss Ecol. Lett. 2003 6 509 517 10.1046/j.1461-0248.2003.00454.x 17. Brose, U. & Hillebrand, H. Biodiversity and ecosystem functioning in dynamic landscapes. Philos. Trans. R. Soc. B Biol. Sci. 371 (2016). 18. Srivastava DS Vellend M Biodiversity-ecosystem function research: is it relevant to conservation? Annu. Rev. Ecol. Evol. Syst. 2005 36 267 294 10.1146/annurev.ecolsys.36.102003.152636 19. Wolf AA Trait-based filtering mediates the effects of realistic biodiversity losses on ecosystem functioning Proc. Natl Acad. Sci. USA 2021 118 e2022757118 10.1073/pnas.2022757118 34162704 20. Bracken MES Friberg SE Gonzalez-Dorantes CA Williams SL Functional consequences of realistic biodiversity changes in a marine ecosystem Proc. Natl Acad. Sci. USA 2008 105 924 928 10.1073/pnas.0704103105 18195375 21. Zavaleta ES Realistic species losses disproportionately reduce grassland resistance to biological invaders Science 2004 306 1175 1177 10.1126/science.1102643 15539600 22. Schmid B Removing subordinate species in a biodiversity experiment to mimic observational field studies Grassl. Res. 2022 1 53 62 10.1002/glr2.12009 23. Wilsey BJ Wayne Polley H Aboveground productivity and root-shoot allocation differ between native and introduced grass species Oecologia 2006 150 300 309 10.1007/s00442-006-0515-z 16927104 24. Enquist BJ The commonness of rarity: Global and future distribution of rarity across land plants Sci. Adv. 2019 5 1 14 10.1126/sciadv.aaz0414 25. Seebens H No saturation in the accumulation of alien species worldwide Nat. Commun. 2017 8 14435 10.1038/ncomms14435 28198420 26. Angrist, J. D. & Pischke, J. Mostly Harmless Econometrics: An Empiricist’s Companion (Princeton University Press, 2009). 27. Rubin DB Causal inference using potential outcomes J. Am. Stat. Assoc. 2005 100 322 331 10.1198/016214504000001880 28. Isbell F Nutrient enrichment, biodiversity loss, and consequent declines in ecosystem productivity Proc. Natl Acad. Sci. USA 2013 110 11911 11916 10.1073/pnas.1310880110 23818582 29. Ferraro, P. J., Sanchirico, J. N. & Smith, M. D. Causal inference in coupled human and natural systems. Proc. Natl. Acad. Sci. USA, 201805563. 10.1073/pnas.1805563115 (2018). 30. Larsen AE Meng K Kendall BE Causal analysis in control–impact ecological studies with observational data Methods Ecol. Evol. 2019 10 924 934 10.1111/2041-210X.13190 31. Adler PB Productivity is a poor predictor of plant species richness Science 2011 1750 1750 1754 10.1126/science.1204498 32. Lewandowska AM The influence of balanced and imbalanced resource supply on biodiversity-functioning relationship across ecosystems Philos. Trans. R. Soc. Lond. B. Biol. Sci. 2016 371 20150283- 10.1098/rstb.2015.0283 27114584 33. Borer ET Finding generality in ecology: a model for globally distributed experiments Methods Ecol. Evol. 2014 5 65 73 10.1111/2041-210X.12125 34. Oster E Unobservable selection and coefficient stability: theory and evidence J. Bus. Econ. Stat. 2017 0 1 18 35. Athey S Imbens GW The state of applied econometrics: causality and policy evaluation J. Econ. Perspect. 2017 31 3 32 10.1257/jep.31.2.3 29465214 36. Wooldridge, J. M. Econometric Analysis of Cross Section and Panel Data, 2nd edn (The MIT Press, 2010). 37. Angrist JD Krueger AB Instrumental variables and the search for identification: from supply and demand to natural experiments J. Econ. Perspect. 2001 15 69 85 10.1257/jep.15.4.69 38. Wang Y Global evidence of positive biodiversity effects on spatial ecosystem stability in natural grasslands Nat. Commun. 2019 10 1 9 30602773 39. Schoolmaster, D. R., Zirbel, C. R. & Cronin, J. P. A graphical causal model for resolving species identity effects and biodiversity–ecosystem function correlations. Ecology 101 (2020). 40. Ferraro PJ Miranda JJ Panel data designs and estimators as substitutes for randomized controlled trials in the evaluation of public programs J. Assoc. Environ. Resour. Econ. 2017 4 281 317 41. Naeem S Thompson LJ Lawler SP Lawton JH Woodfin RM Declining biodiversity can alter the performance of ecosystems Nature 1994 368 734 737 10.1038/368734a0 42. Dee, L. E., Kimmel, K. & Hayden, M. Biodiversity and ecosystem functioning in observational analyses. In The Ecological and Societal Consequences of Biodiversity Loss (eds. Loreau, M. et al.) 119–145 (Wiley, 2022). 43. Reich PB Impacts of biodiversity loss escalate through time as redundancy fades Science 2012 336 589 592 10.1126/science.1217909 22556253 44. Parker SS Harpole WS Seabloom EW Plant species natural abundances are determined by their growth and modification of soil resources in monoculture Plant Soil 2019 445 273 287 10.1007/s11104-019-04299-0 45. Vander Weele TJ Concerning the consistency assumption in causal inference Epidemiology 2009 20 880 883 10.1097/EDE.0b013e3181bd5638 19829187 46. Vilà M Weiner J Are invasive plant species better competitors than native plant species? - Evidence from pair-wise experiments Oikos 2004 105 229 238 10.1111/j.0030-1299.2004.12682.x 47. Zhang Z Liu Y Yuan L Weber E van Kleunen M Effect of allelopathy on plant performance: a meta-analysis Ecol. Lett. 2021 24 348 362 10.1111/ele.13627 33085152 48. Jochum M The results of biodiversity–ecosystem functioning experiments are realistic Nat. Ecol. Evol. 2020 4 1485 1494 10.1038/s41559-020-1280-9 32839545 49. Ferraro PJ Agrawal A Synthesizing evidence in sustainability science through harmonized experiments: community monitoring in common-pool resources Proc. Natl Acad. Sci. USA 2021 118 e2106489118 10.1073/pnas.2106489118 34257156 50. Rubin D Causal inference using potential outcomes: design, modeling, decisions J. Am. Stat. Assoc. 2005 100 322 331 10.1198/016214504000001880 51. Rubin DB Estimating causal effects of treatments in randomized and nonrandomized studies J. Educ. Psychol. 1974 66 688 701 10.1037/h0037350 52. Cardinale BJ The functional role of producer diversity in ecosystems Am. J. Bot. 2011 98 572 592 10.3732/ajb.1000364 21613148 53. Isbell FI Biodiversity increases the resistance of ecosystem productivity to climate extremes Nature 2015 526 574 577 10.1038/nature15374 26466564 54. Lauenroth WK Hunt HW Switft DM Singh J Estimating aboveground net primary production in grasslands: a simulation approach Ecol. Model 1986 33 297 314 10.1016/0304-3800(86)90045-1 55. Bolker BM Generalized linear mixed models: a practical guide for ecology and evolution Trends Ecol. Evol. 2009 24 127 135 10.1016/j.tree.2008.10.008 19185386 56. Sugihara G Detecting causality in complex ecosystems Science 2012 338 496 500 10.1126/science.1227079 22997134 57. Bertrand M Duflo E Mullainathan S How much should we trust differences-in- differences estimates? Q. J. Econ. 2004 119 249 275 10.1162/003355304772839588 58. Cameron AC Miller DL A practitioner’s guide to cluster- robust inference J. Hum. Resources 2015 50 317 372 10.3368/jhr.50.2.317 59. MacDonald AJ Larsen AE Plantinga AJ Missing the people for the trees: Identifying coupled natural–human system feedbacks driving the ecology of Lyme disease J. Appl. Ecol. 2019 56 354 364 10.1111/1365-2664.13289 60. Creel S Creel M Density dependence and climate effects in Rocky Mountain elk: an application of regression with instrumental variables for population time series with sampling error J. Anim. Ecol. 2009 78 1291 1297 10.1111/j.1365-2656.2009.01581.x 19549144 61. Angrist JD Identification of causal effects using instrumental variables J. Am. Stat. Assoc 1996 91 444 455 10.1080/01621459.1996.10476902 62. Imbens GW InstrumentaL Variables: an Econometrician’s Perspective Stat. Sci. 2014 29 323 358 10.1214/14-STS480 63. Kendall, B. E. A statistical symphony: Instrumental variables reveal causality and control measurement error. Ecol. Stat. Contemp. Theory Appl. 149–167 10.1093/acprof:oso/9780199672547.001.0001. (2015) 64. Kimmel K Dee LE Avolio ML Ferraro PJ Causal assumptions and causal inference in ecological experiments Trends Ecol. Evol. 2021 36 1141 1152 10.1016/j.tree.2021.08.008 34538502 65. Borer ET Herbivores and nutrients control grassland plant diversity via light limitation Nature 2014 508 517 520 10.1038/nature13144 24670649 66. Seabloom EW Increasing effects of chronic nutrient enrichment on plant diversity loss and ecosystem productivity over time Ecology 2021 102 e03218 10.1002/ecy.3218 33058176 67. Morgan, S. L. & Winship, C. Counterfactuals and Causal Inference (Cambridge University Press, 2015). 68. Olea, J. L. M. & Pflueger, C. A Robust Test for Weak Instruments. J. Bus. Econ. Stat. 10.1080/00401706.2013.806694 (2013). 69. Hautier Y Niklaus PA Hector A Competition for light causes plant biodiversity loss after eutrophication Science 2009 324 636 638 10.1126/science.1169640 19407202 70. Ashenfelter O Estimating the effect of training programs on earnings Rev. Econ. Stat. 1978 60 47 57 10.2307/1924332 71. Ding P Li F A bracketing relationship between difference-in-differences and lagged-dependent-variable adjustment Polit. Anal. 2019 27 605 615 10.1017/pan.2019.25 72. Altonji JG Elder TE Taber CR Selection on observed and unobserved variables: assessing the effectiveness of catholic schools J. Polit. Econ. 2005 113 151 184 10.1086/426036 73. Leibold MA Chase JM Ernest SKM Community assembly and the functioning of ecosystems: how metacommunity processes alter ecosystems attributes Ecology 2017 98 909 919 10.1002/ecy.1697 27984663
PMC010xxxxxx/PMC10164145.txt
==== Front Transl Psychiatry Transl Psychiatry Translational Psychiatry 2158-3188 Nature Publishing Group UK London 37149657 2448 10.1038/s41398-023-02448-9 Article NAPE-PLD deletion in stress-TRAPed neurons results in an anxiogenic phenotype http://orcid.org/0000-0002-2225-682X Tevosian Margaryta 12 Todorov Hristo 3 Lomazzo Ermelinda 1 http://orcid.org/0000-0001-5538-2008 Bindila Laura 1 Ueda Natsuo 4 Bassetti Davide 56 Warm Davide 6 Kirischuk Sergei 6 Luhmann Heiko J. 6 http://orcid.org/0000-0001-9513-0729 Gerber Susanne 3 Lutz Beat [email protected] 12 1 grid.410607.4 Institute of Physiological Chemistry, University Medical Center of the Johannes Gutenberg University, Mainz, Germany 2 grid.509458.5 0000 0004 8087 0005 Leibniz Institute for Resilience Research (LIR), Mainz, Germany 3 grid.410607.4 Institute of Human Genetics, University Medical Center of the Johannes Gutenberg University, Mainz, Germany 4 grid.258331.e 0000 0000 8662 309X Department of Biochemistry, Kagawa University School of Medicine, Kagawa, Japan 5 grid.7645.0 0000 0001 2155 0333 Department of Mathematics, Technical University of Kaiserslautern, Kaiserslautern, Germany 6 grid.410607.4 Institute of Physiology, University Medical Center of the Johannes Gutenberg University, Mainz, Germany 6 5 2023 6 5 2023 2023 13 15214 12 2022 21 4 2023 25 4 2023 © The Author(s) 2023 https://creativecommons.org/licenses/by/4.0/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Anandamide (AEA) is an endogenous ligand of the cannabinoid CB1 and CB2 receptors, being a component of the endocannabinoid signaling system, which supports the maintenance or regaining of neural homeostasis upon internal and external challenges. AEA is thought to play a protective role against the development of pathological states after prolonged stress exposure, including depression and generalized anxiety disorder. Here, we used the chronic social defeat (CSD) stress as an ethologically valid model of chronic stress in male mice. We characterized a genetically modified mouse line where AEA signaling was reduced by deletion of the gene encoding the AEA synthesizing enzyme N-acyl-phosphatidylethanolamine-hydrolyzing phospholipase D (NAPE-PLD) specifically in neurons activated at the time of CSD stress. One week after the stress, the phenotype was assessed in behavioral tests and by molecular analyses. We found that NAPE-PLD deficiency in neurons activated during the last three days of CSD stress led to an increased anxiety-like behavior. Investigating the molecular mechanisms underlying this phenotype may suggest three main altered pathways to be affected: (i) desensitization of the negative feedback loop of the hypothalamic-pituitary-adrenal axis, (ii) disinhibition of the amygdala by the prefrontal cortex, and (iii) altered neuroplasticity in the hippocampus and prefrontal cortex. Subject terms Molecular neuroscience Depression https://doi.org/10.13039/501100001659 Deutsche Forschungsgemeinschaft (German Research Foundation) CRC1193 B04 CRC1193 B04 Luhmann Heiko J. Lutz Beat issue-copyright-statement© Springer Nature Limited 2023 ==== Body pmcIntroduction The endocannabinoid (eCB) system consists of the cannabinoid type 1 and type 2 receptors (CB1, CB2), their main endogenous ligands 2-arachidonoyl glycerol (2-AG) and N-arachidonoyl ethanolamine (anandamide, AEA), and the synthesizing and degrading enzymes of eCBs. At synapses, the eCB system was originally proposed to be mainly involved in retrograde suppression of neurotransmitter release, mediated by postsynaptically synthesized and released 2-AG and binding to the Gi/o protein-coupled CB1 at the presynaptic site. Recent insights, though, have revealed a much higher complexity of how eCBs influence synaptic transmission and plasticity, including AEA, also due to the involvement of glial cells, such as astrocytes [1, 2]. AEA belongs to the family of N-acylethanolamines (NAEs), which are synthesized by the Ca2+ dependent enzyme N-acyl phosphatidylethanolamine phospholipase D (NAPE-PLD), which is dominantly expressed at the presynapse [3]. AEA can be produced on demand, whereby synthesized overshoots are rapidly eradicated by degrading enzymes, such as fatty acid amide hydrolase (FAAH) [4]. NAPE-PLD deficient mice were independently generated by three groups [5–7]. While AEA levels were not reduced in the NAPE-PLD mutant line generated by the Cravatt’s group [6], AEA was significantly reduced in the lines created by Ueda’s group [7] and Luquet’s group [5]. In our study, we used the NAPE-PLDfl/fl mouse line by Ueda’s group [7]. Some studies on NAPE-PLD deficient mice from Cravatt’s laboratory showed that AEA levels in the brain of these mice were not different from wild-type littermates [8], suggesting alternative synthesis pathways involving other enzymes: the glycerophosphodiesterase (GDE) enzyme family, GDE1, GDE4, and GDE7, as well as α/β-hydrolase domain-containing protein 4 (ABHD4) [9]. Furthermore, NAPE-PLD deficiency was shown to lead to lipid changes beyond AEA, including different NAEs [10]. Stress can be defined as an external or internal challenge that is perceived as threatful and evokes specific physiological and behavioral responses in an organism [11]. Prolonged exposure to an aversive environment causes chronic stress, which can exhaust the resources, leading to the development of pathological states, such as depression and generalized anxiety disorders. Several studies involving the CB1 antagonist rimonabant and CB1 deficient mice introduced the notion that the eCB system exerts an inhibitory action on the hypothalamic-pituitary-adrenal (HPA) axis [12–15]. eCBs reportedly act at the site of fast, non-genomic negative feedback [15]. CB1 is expressed in many brain regions involved in the processing of anxiety and stress, including the hippocampus (Hip), the prefrontal cortex (PFC), the bed nucleus of the stria terminalis (BNST), the basolateral (BLA) and central amygdala (CeA) and various hypothalamic nuclei (Hypo) [16]. BNST integrates neuronal inputs from these regions for the inhibitory control of emotional stress induced by HPA axis activity [17]. eCB signaling in the BLA modulates both excitatory and inhibitory neurotransmission, and the content of 2-AG and AEA in the BLA is modulated in response to stress [18]. Studies on stress-induced human psychopathologies further pinpoint an important role of AEA in the stress response. For example, the human C385A single-nucleotide polymorphism (SNP) in the Faah gene was studied with regards to pain insensitivity and elevated AEA levels [19], as well as in more moderate response to stress in posttraumatic stress disorder (PTSD) patients [20–23]. AEA content in the amygdala, medial PFC (mPFC), Hip, and Hypo was shown to decrease upon chronic stress exposure [24]. These changes are linked to the concomitant elevation of FAAH activity in these regions, resulting in enhanced hydrolysis of AEA [25]. It is hypothesized that reduction of AEA in BLA results in inadequate gating of excitatory inputs leading to increased glutamate release, causing imbalanced behavioral response to stress. Whole-brain AEA levels were shown to negatively correlate with anxiety-like behavior [26]. Interestingly, pharmacological inhibition or genetic ablation of FAAH leads to the absence of anxiety-like behavior in mice following stress [27], whereas the anxiolytic effect is only observed during aversive testing conditions or stress exposure and absent at baseline non-stressful conditions [28]. The mPFC, a region responsible for top-down control of the emotional response, is highly intertwined with AEA content and responses to stress [29]. As mentioned above, chronic stress results in a decrease of AEA content in the mPFC, leading to disinhibition and basal hyperactivity of the HPA axis. Moreover, low levels of AEA correlate with hypersecretion of corticosterone (CORT). This effect is reversed by elevation of AEA through FAAH inhibition [24]. It has also been observed that cannabinoid CB1 receptor (Cnr1) mRNA is upregulated in the mPFC after chronic stress, possibly to compensate for the deficiency of AEA [30]. The maintenance of the AEA levels in the mPFC could buffer exaggerated responses to stress via mPFC top-down control and thus could be an important factor for stress resilience [31, 32]. Although the involvement of AEA in fear conditioning and extinction paradigms [33–35], as well as its role in the response to chronic restraint stress and chronic unpredictable stress (CUS) [36–38] has been well studied, the involvement of AEA in the framework of chronic social defeat and resulting behavior alterations has not yet been investigated. The widely used CSD model combines elements of both physiological and psychological stress, follows construct, face, and predictive validity and has high translational and ethological value [39]. CSD has been well established as an animal stress model with a robust outcome: a depression-like phenotype with prominent anxiety, anhedonia, and social avoidance behaviors [40]. The present study aims at investigating the role of the eCB system in regulating the response to chronic stress. To this end, we employed a Targeted Recombination in Active Populations (TRAP) strategy to enable a permanent ensemble-specific KO of Napepld. The TRAP approach is commonly used for mapping physiological responses of defined neuronal populations using activity-dependent gene expression as a guide for Cre recombination [41–44]. Here, we analysed the effect of NAPE-PLD deletion in neurons active during the last three days of CSD stress on the phenotype of mice using a battery of behavioral tests. To increase the translational value of this research we applied a “therapeutic” approach and focused on the adaptation phase of chronic stress, instead of the reactive phase. Materials and methods Animals All experiments were performed according to the European Community’s Council Directive of 22 September 2010 (2010/63EU) and approved by the respective agency (Landesuntersuchungsamt) of the State Rhineland-Palatinate (registration number G-17-1-005). Male mice were group-housed in temperature- and humidity-controlled rooms with a 12 h light-dark cycle with water and food provided ad libitum. Seven days prior to behavioral experiments, animals were single-housed. Mice used in this study were 8 weeks old by the start of the experiments. We used transgenic mice derived from crossing B6.Cg-Tg(Arc-cre/ERT2)MRhn/CdnyJ (JAX Nr. 022357), called Arc-Cre here for simplicity [42], with NAPE-PLD-floxed mice [7], further referred to as Arc-NAPE-PLD line. Arc-Cre mice express Cre-ERT2 (a Cre recombinase, fused to a mutated human estrogen receptor ligand binding domain) under the gene-regulatory elements of the activity regulated cytoskeletal-associated protein (Arc). Arc-NAPE-PLD are lacking the NAPE-PLD enzyme specifically on neurons, expressing Arc, in the presence of tamoxifen (TAM, a synthetic estrogen receptor ligand). Arc-Cretg/wt x NAPE-PLDfl/fl mice will be referred to as KO and compared to control littermate Arc-Crewt/wt x NAPE-PLDfl/fl mice referred to as WT. Genotyping was performed before and after experiments. Chronic social defeat Mice used in this study (both KO and WT) were exposed to CSD, which was carried out as previously reported [45]. Briefly, Arc-NAPE-PLD and adult retired breeder CD1 mice (Charles River) were housed in the same cage but were physically separated by a grid for 14 days. During this period, the separating grid was removed every day for 2 min per day to allow mouse interaction and the occurrence of multiple defeat episodes of attack from CD1 mice. After 14 days of CSD, mice were allowed to rest for one week (no exposure to CD1, single housing in home cage) before starting the behavioral tests. Tamoxifen injection Arc-Cre-ERT2 is only active in the presence of its ligand TAM. To induce Cre-dependent recombination, Arc-NAPE-PLD KO, as well as control WT mice, were injected with TAM (Sigma Aldrich, Germany) intra peritoneally (i.p.) at a concentration of 20 mg/ml in a volume of 100 µl per mouse. TAM was dissolved in corn oil and ethanol (9:1 ratio, Sigma Aldrich), immediately frozen in aliquots, which were thawed at room temperature only once, mixed by inversion, and checked for precipitates before injection. Notably, TAM administration was performed on the last three days of CSD to trigger deletion of the AEA synthetizing NAPE-PLD in Arc-expressing neurons at a specific time-point (after/during stress exposure in the adaptation phase of CSD). Behavioral assays Behavioral assessments were carried out during the light phase and trials were video-recorded and analyzed with Ethovision XT software, version 13 (Noldus, the Netherlands). After each trial the set-ups were cleaned with water. The social interaction test (SI) SI followed the procedure reported by [45] and was used to assess social behavior. Briefly, experimental animals were placed into an open field box (40 × 27 x 40 cm) with a small circular enclosure at one wall of the box and allowed to explore the arena for 2.5 min and were then placed back to their home-cage. After introducing the social target – naïve CD1 retired breeder mouse (target) – into the enclosure, the experimental animal was re-introduced to the box and allowed to explore for another 2.5 min. Time in the interaction zone, defined as a circular zone 2.5 cm in diameter around the enclosure, was recorded automatically with tracking software. The time the experimental animal spent in the interaction zone when the target mouse was absent (enclosure empty) was compared to the time spent in the interaction zone when the target mouse was present. SI ratio was calculated as 100 x (time in the interaction zone with a target mouse present) / (time in the interaction zone with target absent). Light/dark test (LDT) LDT was carried out as previously reported [46] and used to test anxiety-like behavior. Briefly, experimental animals were placed into a custom box (39 × 39 cm), divided into a light zone (two thirds of the box, white walls) and a dark zone (one third of the box, separated and covered by a 26 cm high lid, black walls). The light and dark compartments were connected by a small entry zone (5 ×5 cm). Animals were allowed to explore freely and move between the dark and the light zone for 6 min. Time spent in the light zone was assessed using video recording and subsequent automated analysis. The elevated plus maze (EPM) EPM was performed to assess anxiety-like behavior as previously reported [46], using a custom-made cross-shaped set-up having two open and two closed arms, elevated 100 cm above the floor. The arms of the maze were 35 cm long and 6 cm wide. Experimental mice were placed into the center of the maze facing the closed arms and were allowed to freely explore for 10 min. Animals were video-recorded and tracked automatically. The tail suspension test (TST) TST to measure emotional despair or depressive-like behavior followed the original reported procedure with modifications [47]. Mice were fixated by the tail onto a bright illuminated screen with fixative tape and left hanging for 6 min. The experimental animal was visually isolated from nearest objects. The mouse was recorded using a video-camera. Immobility was assessed automatically using tracking software. Nesting behavior test (Nesting) Nesting was performed and scored according to a previously established procedure [48] and was used to evaluate global well-being in mice and monitor whether CSD affects routinely performed tasks and shelter-seeking behavior. Briefly, a pad of compressed cotton was placed overnight into the home cage of the mouse. The following morning the quality and complexity of the nest were evaluated. A combined score (1 – no nest, 5 – perfect nest) was assigned for each mouse, combining the complexity of nest building and the amount of cotton that was left not processed. RNA and lipid analysis Tissue extraction and processing Three days after the last behavioral test, mice were deeply anesthetized using isoflurane (Abbott, USA) and decapitated. Trunk blood was collected in EDTA-coated tubes and immediately centrifuged at 10,000 rpm for 12 min at 4 °C. Serum was transferred to new tubes and frozen on dry ice. Brains were dissected and snap-frozen on a metal plate over dry ice. Tissue was stored at −80 °C until further processing. RNA and lipid co-extraction RNA and lipids were simultaneously isolated from the same tissue samples according to a recently published method for dual extraction of RNA and lipids from brain tissue [49]. Briefly, RNA was extracted by using the RNeasy® mini kit (Qiagen, Germany) according to standard procedure after adding a spike solution containing 10 µl internal standard (ISTDs) mixture of phospholipids, endocannabinoids, and chloroform to the samples to allow subsequent lipid isolation. Reverse transcription and real-time quantitative PCR (RT-qPCR) RNA (~ 500 ng per sample) was converted in cDNA by reverse transcription using the high-capacity cDNA reverse transcription kit (Applied Biosystems/Life Technologies, Germany). The cDNA (100 ng per reaction) was amplified by qPCR using the TaqMan Gene Expression Mastermix (Applied Biosystems/Life Technologies, Germany) and FAM dye-labelled TaqMan probes targeting the following genes: N-acyl phosphatidylethanolamine phospholipase D (Napepld) Mm00724596_m1; activity-regulated cytoskeleton-associated protein (Arc) Mm01204954_g1; FBJ osteosarcoma oncogene (Fos) Mm00487425_m1, Mm01302932_g1 and Mm00487426_g1; early growth response 1 (Egr1) Mm00656724_m1; regulator of G-protein signaling (Rgs2), Mm00501385_m1; fatty acid amide hydrolase (Faah) Mm00515684_m1; cannabinoid type-1 receptor (Cnr1) Mm00432621_s1; brain-derived neurotrophic factor (Bdnf) Mm04230607_s1; neuropeptide Y (Npy) Mm03048253_m1; FK506 binding protein 5 (Fkbp5) Mm00487406_m1; RAS-related C3 botulinum substrate 1 (Rac1) Mm01201653_mH; phosphodiesterase 11 A (Pde11a) Mm01327347_m1; and, as reference genes, glyceraldehyde-3-phosphate dehydrogenase (Gapdh) Mm99999915_g1. Data were analysed with an ABI 7300 real-time PCR cycler (Applied Biosystems/Life Technologies, Germany). Liquid chromatography/mass spectrometry (LC/MS) Lipid profiling was carried by using a SCIEX 5500 QTrap triple-quadrupole linear ion trap mass spectrometer (Concord, ON, Canada). LC conditions for eCBs and phospholipids (PLs) measurements followed in detail the procedure described by Post et al. [49]. The calibration standards N-arachidonoyl ethanolamine (AEA), 2-arachidonoyl glycerol (2-AG), arachidonic acid (AA), palmitoyl ethanolamide (PEA), and corresponding internal standards (ISTDs) (AEA-d4, 2-AG-d5, AA-d8, OEA-d2, PEA-d4 and 1-AG-d5) were purchased from BIOMOL Research Laboratories Inc. The calibration standards and corresponding ISTDs for phosphatidylethanolamine (PE 17:0–14:1) and phosphatidylcholine (PC 17:0–14:1) were purchased from Avanti Polar Lipids, Inc. Calibration curves were used for quantification of all target lipids using the MultiQuant 3.0. Software (AB SCIEX). Electrophysiological experiments Modified social defeat stress Arc-NAPE-PLD KO and WT female and male mice were subjected to modified social defeat stress (mSDS). Animals were placed into the cage of a retired breeder CD1 mouse 3 times for 5 min each. Between the 5 min exposure episodes, animals were separated by a metal mesh for 15 min. After the last exposure episode, the Arc-NAPE-PLD animals were placed back into their home-cage. This procedure was repeated on 5 consecutive days. TAM was injected as previously described during the last 3 days of stress 1 h prior to first exposure episode. After the stress, animals were allowed to rest in their home cages for 7 days prior to sacrifice. Brain slice preparation Brain slices were prepared from postnatal days (P) 50 to P71 in WT and KO mice. The animals were deeply anesthetized with isoflurane (Abbott, USA) and decapitated. The brain was rapidly removed and placed into ice-cold protective artificial cerebrospinal fluid containing (in mM): 110 choline chloride, 2.5 KCl, 1.25 NaH2PO4, 25 NaHCO3, 20 glucose, 11.6 sodium L-ascorbate, 3.1 sodium pyruvate, 0.2 CaCl2, 5 MgCl2. The solution was continuously bubbled with carbogen (95% O2 and 5% CO2), pH 7.4. Coronal slices of 300 μm thickness containing the prelimbic region of the mPFC [50] were prepared using a vibratome (Campden Instruments Ltd., UK). Slices were transferred into a storage chamber filled with artificial cerebrospinal fluid (aCSF) with a composition of (in mM): 126 NaCl, 2.5 KCl, 10 glucose, 1.25 NaH2PO4, 25 NaHCO3, 2 CaCl2, 1 MgCl2, continuously bubbled with carbogen, pH 7.4. Slices were stored for at least 60 min prior to recording. MEA recordings and analysis Extracellular recordings were performed with microelectrode arrays (MEA) consisting of 120 planar titanium nitrite electrodes with 4 internal references (120MEA100/30iR-Ti-pr, Multi-Channel Systems, Germany) using a MEA2100-System (Multi Channel Systems, Germany). Under a stereomicroscope, slices were carefully arranged onto MEAs in order to place the mPFC in correspondence with the field of electrodes. During recordings, slices were perfused with aCSF (equilibrated with 95% O2 / 5% CO2) at a rate of 3 ml/min through a peristaltic perfusion system and kept at 32 °C with a TC02 temperature controller connected to a heating plate and to a PH01 heatable perfusion cannula (Multi Channel Systems, Germany). Electrophysiological recordings were carried out for 10 min. Raw data from 120 channels were acquired at 50 kHz using MC_Rack software (Multi Channel Systems, Germany) and down-sampled to 200 Hz for local field potential (LFP) analysis. For LFP events detection, a threshold of 7 times the standard deviation of the noise was set on the negative slope of the signal and waveforms of max. 1100 ms were stored. Extracted datasets from all channels were then imported into Matlab 7.7 (Mathworks, MA, USA) for analysis of LFP event rates using a custom written routine. Only channels showing at least 4 LFP events in 10 min recordings were taken in consideration and counted as active. For comparisons, LFP rates were pooled across all recordings. Whole-cell patch-clamp recordings and analysis Brain slices were placed into a recording chamber (~0.7 ml volume) on the microscope stage (Axioscope FS, Zeiss, Germany). Slices were submerged with a constant flow of carbogenated aCSF. Flow rate was set to 1–1.5 ml/min. All experiments were performed at 31–32 °C. Pyramidal neurons in layers 2/3 of mPFC were selected visually according to morphological criteria. A 40x objective (Zeiss, Germany) was used. Patch pipettes were prepared from borosilicate glass capillaries using P-87 puller (Sutter Instrument Co., USA) and filled with an intracellular fluid (ICF) composed of (in mM): 130 KCl, 5 NaCl, 5 EGTA, 25 HEPES, 0.5 CaCl2, 2 Mg-ATP, 0.3 Na-GTP. pH was adjusted to 7.3 with KOH. Pipette resistance was 3–6 MOhms when filled with intracellular solution. Electrophysiological signals were acquired using an EPC-10 amplifier and TIDA 5.24 software (both from HEKA Elektronik, Germany). The signals were filtered at 3 kHz and sampled at a rate of 10 kHz. Liquid junction potential (<5 mV) was not corrected. Cells were patched in whole-cell configuration and the holding potential was set to −70 mV. Hyperpolarizing pulses of 10 mV were used to access cell capacity, series and access resistance. Only recordings with series resistance below 30 MOhm were accepted. Series resistance compensation was not applied. Cells exhibiting more than 20% changes in access resistance during an experiment were discarded. Action potentials (APs) were recorded in current clamp mode. Suprathreshold depolarizing stimuli of varying increasing amplitudes (500 ms) were applied to elicit APs. Miniature post-synaptic currents (mPSCs), both excitatory (mEPSCs) and inhibitory (mIPSCs), were recorded in presence of 0.5 μM tetrodotoxin (TTX), a blocker of voltage-gated sodium channels. mEPSCs were recorded in aCSF supplemented with 10 μM gabazine (Sigma, Germany), a blocker of γ-aminobutyric acid (GABA)A receptors, and 50 μM DL-2-amino-5-phosphonopentanoic acid (DL-APV), a N-methyl-D-aspartic acid (NMDA) receptor blocker. mIPSCs were recorded in the presence of 10 μM 6,7-dinitroquinoxaline-2,3-dione (DNQX), an α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA)/kainate receptors antagonist, and 50 μM DL-APV. All drugs were purchased from BioTrend (Germany) unless otherwise specified. Evoked inhibitory postsynaptic currents (eIPSCs) were recorded in the presence of 10 µM DNQX and 50 µM DL-APV. Stimulation electrodes were prepared using the same method as for patch pipettes but filled with aCSF. Responses were evoked using focal electrical stimulation in the vicinity of the patched cell (100–150 µm laterally) using rectangular current pulses. The latter were delivered using a custom-built stimulation unit controlled by the amplifier. The intensity of the stimulation was regulated in order to elicit a unitary synaptic input (minimal stimulation). Electrophysiological data were evaluated using TIDA 5.24 (HEKA Elektronik, Germany). mPSCs were analyzed using PeakCount V3 software. The program employs a derivative threshold-crossing algorithm to detect individual PSCs. Each automatically detected event was displayed for visual inspection. Data analysis Analysis of behavior testing, RNA and lipid analysis and electrophysiological experiments No formal sample size estimation or randomization was performed, and all available KO mice and WT littermates were included in the experiments. Investigators were not blinded to allocation and outcome analysis. Data are represented as the mean±standard error of the mean (SEM). Statistical analysis was performed using GraphPad Prism 6 (GraphPad Software, CA, USA). D’Agostino-Pearson normality test was performed for all groups. Unpaired two-tailed Student’s t-test was used to analyze normally distributed data. Mann-Whitney test was performed in case of not normally distributed data. One or two-way analysis of variance (ANOVA) with Tukey’s or Bonferroni multiple comparison post-hoc test was performed where applicable. p < 0.05 was set as value to determine statistical significance. All reported p-values are two-tailed. Statistical significance was presented using the following rules: *p < 0.05, **p < 0.005, ***p < 0.001, ns- not significant. Multivariate analysis Multivariate analyses were performed using the R language and environment for statistical computing version 4.0.2 [51]. RNA expression and lipid data were first screened for outliers by transforming raw values to z-scores. In case that a z-score exceeded the critical threshold of |3.29 | (p < 0.001, two-sided z-test), the corresponding measurement was marked as an outlier and excluded from univariate statistical analysis (linear mixed effects models). Outliers were treated as missing values in the multivariate analyses. Missing values were imputed using the predictive mean matching method from the mice R package v3.10.0 with 10 multiple imputations. Molecular markers from different brain regions were represented in reduced multivariate space using principal component analysis (PCA) as implemented in the stats R package v4.0.2. Redundancy analysis (RDA) followed by permutational analysis of variance was performed to test the multivariate hypothesis if RNA expression and lipid data differ between brain regions and genotypes using the vegan R package v2.5–6. Sparse partial least squares discriminant analysis (sPLS-DA) models were calculated using the mixOmics package v6.12.1. Correlational analysis was performed by calculating pairwise Pearson correlation coefficients. Correlation matrices were visualized using the corrplot package v0.84. To identify important molecular features which were most predictive of key behavioral outcome measures in the EPM, LDT and SI tests, we fit multiple linear regression models with a lasso penalty for the maximum likelihood estimates using glmnet v4.1-3. The shrinkage penalty λ was estimated using 10-fold cross-validation implemented in the cv.glmnet function. The λ value associated with the smallest mean cross-validated error was chosen for the final penalized regression model. Univariate differences in RNA expression and lipid data between different brain regions and genotypes were investigated using linear mixed effects models as implemented in the lme4 package v1.1–23. Genotype and brain region were included as fixed effects in the model and animal as a random effect. Pairwise comparisons of model means were calculated using emmeans v1.4.8. Results Mice devoid of NAPE-PLD in neurons TRAPed at the last three days of CSD exhibit an anxiety-like phenotype Arc-NAPE-PLD WT and KO animals were injected with TAM i.p. every day during the last three days of the CSD protocol (day 12–14), one hour before the defeat episode (Fig. 1A). We observed a pronounced anxiety-like phenotype in stressed Arc-NAPE-PLD KO mice in LDT and EPM. Mice spent significantly less time in the light compartment of the LDT (Fig. 1B, p = 0.0443). In the EPM, KO mice spent less time in the open arms of the apparatus than WT (Fig. 1C, p = 0.0070). Furthermore, KO mice showed a tendency of keeping their body longer in a contracted position than WT (Fig. 1D, p = 0.0692) as well as spending more time not moving (Fig. 1E, p = 0.0596), indicating a decrease in exploration behavior, which is a manifestation of an anxiety-like state. The time spent in the open arms of the EPM and in the light compartment in the LDT test were positively correlated when genotypes were pooled together (Fig. 1F), pointing to the coherent anxiety-like phenotype and validity of the behavioral tests. This effect was even more pronounced in the KO group. There were no differences found in other behavioral tests such as SI (p = 0.8336), Nesting (p = 0.8082), and TST (0.5225) between Arc-NAPE-PLD WT and KO, as well as no differences in bodyweight during the CSD and at the beginning of behavioral testing (Supplementary Fig. 1). In a separate experiment, we compared control non-stressed Arc-NAPE-PLD WT and KO mice that were single-housed for 12 days and then injected with TAM i.p. for 3 consecutive days (equivalent to day 12–14 of CSD). 7 days after the last TAM injection, the battery of behavioral tests was performed. We did not observe any genotype effects in these non-stressed mice (Supplementary Fig. 2).Fig. 1 Anxiety-like phenotype of Arc-NAPE-PLD KO mice. A Scheme of experimental approach. CSD stress lasted for 14 days, TAM injections were performed on the last three days prior to CSD. The genetic manipulation in TRAPed neurons is illustrated: in Arc-expressing neurons, upon TAM application, the TAM-inducible CreERT2 recombinase excises NAPE-PLD gene sequences that are flanked by loxP sites. After seven days of housing in the home cage, behavioral tests were performed on day 21 to 25, followed by sacrifice, and mRNA and lipids extraction. B–E Group comparisons of Arc-NAPE-PLD WT and KO in standard behavioral tests. LDT: light-dark test (LDT), and elevated plus maze (EPM). Unpaired t test was used to identify significant differences, *p<0.05, **p<0.01. Data are represented as mean ± SEM. F Regression lines were drawn irrespective of genotype (black) as well as separately for WT (blue; n = 16) and KO (orange; n = 24) animals. The overall Pearson correlation coefficient as well as the correlation coefficients for each genotype and the respective p-value are shown on the plots. G, H AEA and Napepld mRNA levels in brain regions of stressed Arc-NAPE-PLD WT and KO mice. G AEA levels were measured using LC/MS. H Napepld mRNA levels were normalized to Gapdh. Genotype and brain region differences were assessed using linear mixed effects models. Genotype and brain region were included as fixed effects in the model and animal as random effect. Pairwise comparisons were based on estimated model means with Tukey p value adjustment in case of multiple comparisons. Asterisks indicate significant differences between genotypes in a specific region: *** p < 0.001. Letters indicate the significance of pairwise comparisons of different brain regions within the same genotype. Groups sharing the same letter are significantly different from each other at p < 0.05. Data are represented as mean ± SEM and individual values; WT/KO n = 16/24. PFC prefrontal cortex, dHip dorsal hippocampus, vHip ventral hippocampus, Hypo hypothalamus, Gapdh glyceraldehyde-3-phosphate-dehydrogenase, LC/MS liquid chromatography/mass spectrometry. TRAP recombination leads to decreased Napepld expression and reduced AEA content in selected brain regions of Arc-NAPE-PLD KO mice The observed phenotype of Arc-NAPE-PLD KO mice, exposed to CSD and TRAPed during the last three days of the stress paradigm, led us to investigate the molecular drivers of the anxiety-like behavior. WT and KO mice were sacrificed two days after the last behavioral test; brain tissue was collected and immediately frozen. The right hemisphere was manually dissected into the prefrontal cortex (PFC), dorsal and ventral hippocampus (dHip and vHip, respectively), and the hypothalamus (Hypo). The tissue was later processed in the dual extraction procedure [49] to obtain the lipid fraction and the RNA fraction. We validated the TRAP recombination by measuring the AEA content and Napepld mRNA in the brains of stressed Arc-NAPE-PLD KO and stressed littermate WT mice. We observed a significant decrease of AEA in the dHip of KO mice (Fig. 1G, p < 0.001) compared to WT. The same trend was present in the PFC, but the difference was not statistically significant (p = 0.09). Paradoxically, there was an increase of AEA in the Hypo of the KO mice compared to WT. We also detected a significant decrease of PEA in dHip and vHip of KO mice (Supplementary Fig. 3A), as well as an increase of 2-AG in dHip and a decrease in Hypo of KO mice (Supplementary Fig. 3B). PEA is a N-acylethanolamines and similarly to AEA it is synthesized by NAPE-PLD, among other enzymes. The mRNA levels of Napepld were significantly decreased in the dHip of KO compared to WT (Fig. 1H, p < 0.001). No changes of Napepld expression were observed in the PFC, vHip, and Hypo of KO animals compared to the WT group. Herewith we confirmed the TRAPing of a subset of neurons, mostly in the dHip and partially in PFC. Distinct molecular signatures of selected brain regions of Arc-NAPE-PLD KO mice Next, we asked whether the different brain regions of stressed Arc-NAPE-PLD WT and KO mice were associated with distinct molecular signatures based on the selected eCBs and eCB-related lipids: AEA, 2-AG, arachidonic acid (AA), palmitoyl ethanolamide (PEA), and mRNA markers: Napepld; activity-regulated cytoskeleton-associated protein (Arc); FBJ osteosarcoma oncogene (Fos); early growth response factor 1 (Egr1); regulator of G-protein signaling (Rgs2); fatty acid amide hydrolase (Faah); Cnr1; brain-derived neurotrophic factor (Bdnf); neuropeptide Y (Npy); FK506 binding protein 5 (Fkbp5); RAS-related C3 botulinum substrate 1 (Rac1). To this end, we performed a principal component analysis (PCA), revealing that each brain region formed a distinct cluster (Fig. 2A, B). dHip was separated from PFC/vHip and Hypo along the first multivariate dimension (PC1), which captured 43.87% of the total variance. While the pattern was similar for WT and KO mice, differences between brain regions were more pronounced for WT animals, as indicated by the brain clusters for KO mice appearing closer to each other in the reduced multivariate space (Fig. 2B). To evaluate which of the original variables included in the analysis can explain this multivariate pattern, we calculated the loadings on PC1, which correspond to correlations of the original variables with the multivariate dimension (Fig. 2C). Altogether, 12 out of the 15 investigated molecular markers significantly contributed to differences between the brain regions. Rgs2 (controls signaling through G-protein coupled receptors) and the eCB 2-AG were positively correlated with PC1, indicating that, on average, the levels of these markers were highest in the Hypo. All additional markers were negatively correlated with PC1. Furthermore, we calculated the loadings on PC2, which explained the differences between the PFC and vHip (Fig. 2D). Six out of the 15 molecular markers significantly contributed to this pattern. 2-AG and Rgs2 were again positively correlated with PC2, pointing to increased levels in vHip compared to PFC. Apart from the differences between brain regions, we were also interested in elucidating the impact of genotype on the multivariate molecular signature. We did not observe a distinct separation between genotypes in the PCA analysis (Fig. 2E). Therefore, we employed redundancy analysis (RDA), which is the constrained version of PCA. We included genotype and brain region as well as their interaction as explanatory variables in the ordination procedure. WT mice were not visibly separated from KO animals on the RDA ordination plot (Fig. 2F). However, permutational analysis of variance revealed that the effects of genotype (p = 0.017) and the interaction term genotype/brain region (p = 0.001) were statistically significant. To more specifically investigate genotype differences, we performed sparse partial least squares discriminant analyses (sPLS-DA) for each brain region separately. PLS-DA offers higher sensitivity for detecting group differences compared to PCA and RDA while simultaneously being a robust method against false positives [52]. Results from this analysis for the PFC are shown in Fig. 2G. KO animals were separated from the WT littermates along the first multivariate dimension and the important molecular features explaining this pattern were Napepld, AEA, AA, and Cnr1, which on average had higher levels in WT animals as well as Npy, Egr1, and Arc, which were associated with increased levels in KO mice (Fig. 2H). To confirm these patterns we detected with the multivariate techniques, we also performed univariate analyses investigating the impact of brain region and genotype on specific molecular targets (Fig. 3).Fig. 2 Principal component analysis (PCA) of endocannabinoids, arachidonic acid and analysed transcripts in different brain regions of stressed Arc-NAPE-PLD WT and KO mice. The PCA model was calculated using all data from WT and KO mice. After the ordination procedure, WT (A) and KO (B) animals were visualized separately to investigate differences between brain regions within each genotype. The percentage of variance explained by each dimension is indicated on the respective axis in A and B. Confidence ellipses around each brain region were drawn at the 95% level. Loadings of the original variables on the first (PC1) and second principal component (PC2) are shown in C and D, respectively. Variables with an absolute loading greater than 0.4 were considered to significantly contribute to the observed multivariate pattern. E–H Multivariate differences in endocannabinoids, arachidonic acid and mRNA levels between stressed Arc-NAPE-PLD WT and KO mice. The effect of genotype for all brain regions was investigated using PCA or RDA. G The impact of genotype on the RDA ordination was examined statistically using permutational analysis of variance. Multivariate differences between WT and KO mice in the PFC were evaluated additionally using sPLS-DA. The percentage of variance explained by each dimension is indicated on the respective axis in E–G. Confidence ellipses for each genotype were drawn at the 95% level. H Correlations of the original variables with the first multivariate dimension (sPLS-DA1). Fig. 3 mRNA levels of selected molecular targets in brain regions of stressed Arc-NAPE-PLD WT and KO mice. A Npy mRNA levels, B Arc mRNA levels, C Egr1 mRNA levels, D Cnr1 mRNA levels, E Faah mRNA levels, F Bdnf mRNA levels; mRNA levels were normalized to Gapdh. Genotype and brain region differences were assessed using linear mixed-effects models. Genotype and brain region were included as fixed effects in the model and animal as random effect. Pairwise comparisons were based on estimated model means with Tukey p-value adjustment in case of multiple comparisons. Asterisks indicate significant differences between genotypes in a specific region: *p < 0.05; **p < 0.01; ***p < 0.001. Letters indicate the significance of pairwise comparisons of different brain regions within the same genotype. Groups sharing the same letter are significantly different from each other at p ≤ 0.05. Data are represented as mean ± SEM and individual values; WT/KO n = 16/24. Npy neuropeptide Y, Erg1 early growth response factor 1, Cnr1 cannabinoid CB1 receptor, Faah fatty acid hydrolase, Bdnf brain-derived neurotrophic factor. There was a significant increase of Npy in the PFC and a decrease in dHip of KO mice compared to WT. Additionally, we observed high variability of expression patterns of Npy across the analyzed brain regions with the highest expression in PFC and the lowest in vHip (Fig. 3A). To our surprise, the expression of the immediate-early gene (IEG) Arc was increased in PFC (p = 0.0006) and dHip (p = 0.013) of KO animals as compared to WT (Fig. 3B). These findings might indicate increased neuronal activity in these regions after CSD stress. Increased expression of another plasticity and activity-associated gene Egr1 was detected in PFC (p = 0.0038) of KO mice, compared to WT (Fig. 3C). However, no changes between genotypes were found in other regions. Another protein associated with neuronal plasticity is BDNF. Bdnf mRNA was slightly downregulated in dHip of KO mice (Fig. 3F) but did not reach significance (p = 0.0865). The expression of Cnr1 was significantly increased in dHip of KO mice (Fig. 3D, p = 0.0061), which might indicate the recruitment of CB1 following the abolishment of NAPE-PLD-mediated synthesis of AEA. Moreover, the Faah expression was significantly reduced in dHip of KO mice (Fig. 3E, p < 0.001) to potentially compensate for the decreased level of AEA in dHip (see Fig. 1G). Individual behavioral performance correlates with the brain’s molecular signature Additionally, we were interested in elucidating the potential relationship between the molecular signature of individual mice and their performance in the behavioral tests. Therefore, we calculated the pairwise correlations between all molecular and behavioral outcome measures, which identified multiple potential associations between behavioral and molecular features (Supplementary Fig. 4). To limit the analysis to the most representative outputs of the behavioral tests, we focused on the time spent in the light compartment in the LDT and the time the nose point of the mouse was detected in the open arms of the EPM as a proxy of anxiety-like behavior. The time spent interacting with the CD1 mouse in the SI test (nose point detected in proximity to CD1 mouse) was used as a proxy for sociability and generalization after CSD. We performed lasso regression for each selected behavioral measure as response variable to identify important molecular markers, which had the highest predictive value for the performance in the respective behavioral test and could thereby serve as biomarkers of anxiety-like behavior and social avoidance. The lasso regularization method facilitates feature selection by setting the regression coefficients of covariates with low predictive value to 0. This analysis revealed that the AEA and AA levels in dHip were the most important molecular markers predictive of the time spent in the open arms. A higher concentration of both AEA and AA in dHip was significantly correlated with an increased duration of stay in the open arms when both genotypes were analyzed together (Fig. 4A, B). The same trend was also present for WT and KO animals when separately analyzed; however, the association was no longer significant in WT mice, likely due to small sample size.Fig. 4 Correlation between selected behavioral response measurements and molecular markers in selected brain regions of Arc-NAPE-PLD WT and KO mice. A, B Correlation between AEA and AA levels in dHip and the time mice spent in the open arms of the EPM. C, D Correlation between Cnr1 and Napepld mRNA levels in the PFC and the time mice spent in the light compartment of the LDT. E Correlation between Fkbp5 mRNA levels in vHip and the time spent interacting with the CD1 mouse in the SI test. Regression lines were drawn irrespective of genotype (black) as well as separately for WT (blue; n = 16) and KO (orange; n = 24) animals. The overall Pearson correlation coefficient as well as the correlation coefficients for each genotype and the respective p-value are shown on the plots. Fkbp5 FK506 binding protein 5. Additionally, expression levels of Cnr1 and Napepld in PFC were the most relevant variables for predicting the time spent in the light compartment of the LDT (Fig. 4C, D). Increased mRNA levels of Cnr1 in the PFC of both KO and WT were associated with reduced anxiety behavior, as indicated by more time spent in the light compartment of the LDT (r = 0.54, p < 0.0001). This effect was even more pronounced in the WT subgroup (r = 0.69, p = 0.003). Increased Napepld expression in PFC was also positively correlated with a reduced anxiety-like phenotype. However, the association was no longer statistically significant when analyzing WT animals only. Finally, we identified Fkbp5 expression levels in vHip as an important molecular marker that was significantly positively correlated with social behavior in the SI test (Fig. 4E). Increasing mRNA concentrations of Fkbp5 were associated with more pronounced social avoidance indicated by less time spent interacting with the CD1 aggressor mouse. Electrophysiological profiling of the PFC of stressed Arc-NAPE-PLD KO mice Large-scale network activity To address the observed increase of Arc expression in the PFC (see Fig. 3B) as well as to characterize neurophysiological properties of the PFC of stressed mice, Arc-NAPE-PLD WT and KO female and male mice were subjected to a modified social defeat stress (mSDS), injected with TAM in the last three days of mSDS and sacrificed 7 days later. The effectiveness of the mSDS procedure to trigger Arc-mediated recombination was controlled by a genomic PCR performed on brain slices used for electrophysiological measurements (Supplementary Fig. 5). Extracellular recordings of the local field potential (LFP) were performed in acute brain slices of the prelimbic region of the mPFC using a 120 channels microelectrode array (MEA) (Fig. 5A). Spontaneous LFP activity in slices from Arc-NAPE-PLD WT and KO animals was rather low and characterized by both isolated and synchronized events on multiple channels (Fig. 5C). Two types of LFP events were detected: the vast majority had a shorter (<500 ms) duration and lower amplitude, whereas the residual ones had a slower kinetic (~1 s) and larger amplitude (Fig. 5B). Slices from KO animals displayed a higher (22.4 ± 5.2, n = 17 slices), although not significant number of active channels than WT ones (16.9 ± 4.2, n = 11 slices) (Fig. 5C, D). The average frequency of LFP events was significantly (p < 0.05) higher in KO (1.69 ± 0.10 min−1, n = 380 channels) than in WT animals (1.17 ± 0.07 min−1, n = 186).Fig. 5 Large-scale network activity in prelimbic cortex of Arc-NAPE-PLD WT and KO mice. A Coronal slice of the prelimbic region of the mPFC on a MEA; scale bar 100 µm. B Representative traces of two typical LFP events, raw traces (top) and low-pass filtered (bottom). C Representative raster plots (top) and frequency bar plots (bottom) of spontaneous LFP events in WT and KO animal. D Comparison of number of active channels (top) and mean LFP event rates (bottom) in slices from WT and KO mice. Statistical significance was evaluated using Mann–Whitney´s U test, *p < 0.05. MEA microelectrode array, LFP local field potential. Intrinsic and synaptic properties Pyramidal neurons in layer (L) 2/3 of the prelimbic region of the mPFC were recorded using whole cell patch clamp. Average membrane resistance was significantly (p < 0.05) higher in Arc-NAPE-PLD KO neurons (133.4 ± 4.4 MΩ, n = 54) as compared to WT (119.1 ± 4.2 MΩ, n = 38 (Supplementary Fig. 6B). In contrast, resting membrane potential (WT: −61 ± 0.5 mV; KO: −60.6 ± 0.4 mV, p = 0.61), membrane capacitance (WT: 32.9 ± 1.5 pF; KO: 30.8 ± 0.9 pF, p = 0.2) and membrane time constant (WT: 0.7 ± 0.02 ms; KO: 0.69 ± 0.02 ms, p = 0.85) were not significantly different between WT and KO (Supplementary Fig. 6A, C, D). Action potentials were significantly (p < 0.05) larger in KO cells (92.4 ± 0.8 mV, n = 22) than in WT neurons (89.6 ± 0.7 mV, n = 12) (Supplementary Fig. 6H). Other action potential properties such as threshold (WT: −35.5 ± 0.9 mV vs KO: −36.7 ± 0.8 mV, p = 0.31) and duration (WT: 3.54 ± 0.14 ms vs KO: 3.6 ± 0.13 ms, p = 0.77) were not different between WT and KO (Supplementary Fig. 6G, I). Depolarizing current pulse injections of increasing amplitude did not reveal any significant differences in firing frequencies in WT and KO (Supplementary Fig. 6E, F). Miniature excitatory postsynaptic currents (mEPSCs) did not reveal any differences between WT and KO neurons in either frequency (WT: 2.06 ± 0.36 Hz vs KO: 2.19 ± 0.34 Hz, p = 0.79) or amplitude (WT: 12.36 ± 0.56 pA vs KO: 13.45 ± 1.31 pA, p = 0.48) (Fig. 6A). Rise time (WT: 0.96 ± 0.07 ms vs KO: 0.94 ± 0.08 ms, p = 0.86) and decay time of the events (WT: 5.77 ± 0.23 ms vs KO: 6.20 ± 0.42 ms, p = 0.4) were not significantly different in WT and KO neurons. While amplitude (WT: 35.86 ± 2.83 pA vs KO: 38.05 ± 2.49 pA, p = 0.29) and kinetic properties (rise time - WT: 0.94 ± 0.03 ms vs KO: 0.98 ± 0.05 ms, p = 0.48, decay time - WT: 12.70 ± 0.40 ms vs KO: 12.86 ± 0.57 ms) of miniature inhibitory postsynaptic currents (mIPSCs) were comparable in WT and KO neurons, their mean mIPSC frequency was significantly (p < 0.001) decreased in KO neurons (1.53 ± 0.21 Hz, n = 15) as compared to WT (3.066 ± 0.30 Hz, n = 14) (Fig. 6B). To elucidate the mechanisms underlying the lower frequency of mIPSCs in KO neurons, we recorded evoked inhibitory postsynaptic currents (eIPSCs).Fig. 6 Miniature postsynaptic currents and evoked inhibitory postsynaptic currents in prelimbic cortical neurons of Arc-NAPE-PLD WT and KO mice. A Representative traces of miniature excitatory postsynaptic currents (mEPSCs) in neurons from WT and KO mice and analysis of mean frequency, amplitude, rise time and decay time in neurons from WT (n = 8) and KO (n = 9) mice. B Representative traces of miniatures inhibitory postsynaptic currents (mIPSCs) in neurons from WT (n = 14) and KO (n = 15) mice. The mean frequency of mIPSCs was significantly lower in KO animals than in WT (Student´s t test, ***p < 0.001). C Representative traces of evoked inhibitory postsynaptic currents at interstimulus interval (ISI) of 50 ms (left) and 1000 ms (center). Scaled events superimposed (right). Each trace is the average of 10 repetitions. D Statistical analysis demonstrated reduced amplitude of the mean first event as well as longer decay time in neurons from KO mice compared to WT. The paired-pulse ratio (PPR) was significantly increased in KO mice for ISI up to 250 ms (ANOVA, Bonferroni´s multiple comparisons test). Statistical significance was evaluated using Mann-Whitney´s U test for amplitude of first event and Student´s t test for decay time. *p < 0.05, **p < 0.005. Paired pulse ratio (PPR) was measured using four different interstimulus intervals (ISI): 50, 100, 250 and 1000 ms. The average PPR was significantly increased in KO neurons (Fig. 6C, D, F7.116 = 12.75, p < 0.0001, ANOVA), suggesting a decrease in release probability. In particular, the difference in PPR was larger (p < 0.0001, Bonferroni´s multiple comparisons post hoc test) at the shortest time interval, 0.59 ± 0.03 (n = 16) for WT and 0.79 ± 0.03 (n = 15) for KO at 50 ms ISI. This increase in PPR in KO neurons was present as well at ISI of 100 ms (WT: 0.66 ± 0.03 vs KO: 0.79 ± 0.02, p < 0.05, Bonferroni multiple comparisons test) and 250 ms (WT: 0.72 ± 0.04 vs KO: 0.83 ± 0.03, p < 0.05, Bonferroni multiple comparisons test). At the longest ISI of 1000 ms, the mean PPR did not differ between the two genotypes, as both displayed a PPR of 0.89 ± 0.03 ms. Another indication of reduced release probability in KO animals was the reduction (p < 0.005, Mann–Whitney´s U test) in amplitude of the first response in KO neurons (116.2 ± 12.24 pA, n = 15), as compared to WT (260.7 ± 30.92 pA, n = 16). Interestingly, the decay time was significantly slower (p < 0.05) in events from KO cells (19.80 ± 0.83 ms, n = 15) relative to responses recorded in WT neurons (17.81 ± 0.49 ms, −n = 16). The measurements are consistent with Arc-NAPE-PLD KO animals having morphologically similar L2/3 pyramidal neurons (at least, the perisomatic size), with an increased membrane resistance. Moreover, the evoked APs show slightly increased amplitude. Those cells have a relatively similar composition of synaptic GABAA receptors as WT neurons but receive a lower level of GABAergic synaptic inhibition. Investigation using evoked inhibitory responses suggests a decrease in release probability of inhibitory terminals on L2/3 pyramidal neurons. Discussion In the current study, we leveraged a recently developed TRAP system to genetically manipulate only a subset of neurons that were active at a selected time point during social stress. This sophisticated experimental system allowed us to identify a pronounced anxiety-like phenotype in Arc-NAPE-PLD KO mice. By subsequently employing the dual lipid/mRNA extraction from target key brain regions of these mice, we examined crucial molecular components and targets of the eCB system and the stress response. Our results shed light into the potential mechanisms leading to an increased anxiogenic phenotype as a consequence of reduced AEA signaling in stress-activated neurons. Impairment of the negative feedback loop of the HPA axis Glucocorticoid receptors (GR), the major mediators of the negative feedback regulation of the HPA axis, are abundantly expressed in the PFC, Hip, and Hypo. The interaction of the eCB system, in particular of the CB1, and glucocorticoid signaling arguably plays an important role in controlling the emotional, physiological, and adaptive responses to stress [53]. We observed a strong decrease of Napepld mRNA in the dHip of stressed Arc-NAPE-PLD KO mice and a concomitant decrease of AEA in this brain region as compared to stressed WT. The expression of Faah in dHip was downregulated, possibly as a compensatory mechanism for the decrease of AEA. Similarly, the expression of Cnr1 was upregulated in dHip. These findings are in accordance with several stress studies on rodents, as discussed below. Exposure to repeated stress, such as restraint and social defeat, has previously been reported to be associated with the reduction of AEA levels in the hippocampus [24]. The authors hypothesized that decreased AEA levels contributed to an increase in HPA axis activity and observed a negative correlation between the levels of AEA and basal CORT, a marker for HPA axix activity. Moreover, a high basal expression of Cnr1 was found in Wistar Kyoto (WKY) rats, a model of depressive-like behavior [54]. The upregulation of CB1 is believed to be a compensatory mechanism in response to a reduced AEA signaling in the hippocampus of these rats, mediated by the elevation of basal FAAH activity. When treated with a FAAH inhibitor, URB597, the WKY rats displayed an increase in AEA and a decrease in manifestations of depressive-like behavior. Interestingly, in the context of the glucocorticoid feedback loop, we also observed a negative correlation between FK506 binding protein 51 (Fkbp5) mRNA levels and the performance in the SI test across Arc-NAPE-PLD WT and KO mice, whereby the performance in the SI test is considered to be a behavioral correlate of stress resilience. Prolonged expression of FKBP5 following CORT release in response to a stressor leads to inhibition of the negative feedback loop, resulting in prolonged elevated circulation of CORT and consequent maladaptive stress response [55]. FKBP5 KO mice exhibit mild GR hypersensitivity and improved coping behavior after acute stress exposure [56]. Moreover, FKBP5 KO mice were shown to be less responsive to the deleterious effects of CSD stress in terms of behavioral and neuroendocrine responses [57]. Additionally, mice treated with a recent FKBP5 inhibitor during CSD show reduced social withdrawal and anxiety-like behavior [58], which is in line with our data. Our findings suggest an impairment of the negative feedback loop after prolonged stress exposure in Arc-NAPE-PLD KO mice. To validate this hypothesis, it is necessary to include measurements of the basal activity (plasma levels of CORT) and the functional state of the HPA axis (HPA reactivity assay) in future experiments [59]. Prefrontal cortex exhibits reduced control over amygdala NPY is involved in stress processing; it is thought to be implicated in the termination of stress response and interaction with the HPA axis [60]. We observed an increase of the NPY in PFC of Arc-NAPE-PLD KO mice after CSD stress compared to WT. This finding is in agreement with a study where chronic variable stress induced an increase of NPY in the PFC, while reduced levels were observed in the amygdala [61]. An increase of NPY in cortical interneurons might dampen the inhibiting output of the PFC to the amygdala [62]. A reduction in NPY signaling in the amygdala leads to negative behavioral consequences, such as anxiety [63]. Therefore, reduced PFC control over amygdala may lead to a stress-susceptible phenotype. In future studies, amygdala should be integrated into the analysis of molecular targets to investigate this hypothesis in more detail. Neuroplasticity and stress response In contrast to an increase of NPY in the PFC, we discovered a decrease of NPY mRNA in the dHip of stressed Arc-NAPE-PLD KO mice. This finding is in accordance with a study where mice, exposed to CUS and exhibiting a highly disrupted phenotype, displayed a reduction of NPY in several brain regions, such as the periaqueductal grey, the amygdala, and the hippocampus, compared to non-stressed controls. Furthermore, a single dose of NPY microinfused into the dHip of stressed mice one hour after stress exposure reduced manifestations of anxiety and avoidance behavior [64]. The microinjection of NPY was accompanied by an increase of BDNF. There is an anticipated connection between NPY and BDNF, i.e., TrkB (BDNF receptor) activity was shown to influence the expression of NPY in hippocampal slice cultures [65]. We observed a tendency of decreased BDNF levels in dHip of stressed Arc-NAPE-PLD KO compared to stressed WT. A reduction of BDNF in hippocampus was shown in response to stress [66], whereas an overexpression of BDNF played a stress-protective role [67]. It is hypothesized that increased CORT secretion due to HPA axis activity suppresses BDNF expression. These effects can be restored by prolonged antidepressant treatment [68]. In a recent study, a connection between the NPY and eCBs was established. NPY was shown to mediate anxiolytic and antidepressant effects of elevated AEA due to the action of URB597 in shock and reminders models of PTSD in rats. Moreover, the authors hypothesized that AEA is upstream of the effects of NPY, possibly modulating the sensitivity of the NPY receptor [69]. We observed an increase in the early growth response factor 1 (Egr1) mRNA levels in the PFC of stressed Arc-NAPE-PLD KO mice. EGR1 is an IEG involved in synaptic plasticity, neuronal activity, learning and memory, response to emotional stress, and reward [70]. Interestingly, there is a reported crosstalk between EGR1, BDNF, and GR. GR represses BDNF expression, acting on the activity-regulated BDNF transcript 4 by trans-repression on a DNA site for EGR1 binding, located on the transcription start site of transcript 4, thus establishing a direct functional interaction between GR and EGR1 [71]. To conclude, NPY, BDNF and EGR1 are interconnected factors important for neuronal plasticity. Deficits in synaptic plasticity in key regions such as hippocampus and PFC, responsible for emotional memory and stress processing, can impair the flexibility and adaptation capacity of neural circuits and form a substrate for the development of behavioral abnormalities and pathologies after chronic stress [72]. It is therefore important to further investigate the mechanisms of action of these and other genes involved in neuroplasticity as well as their interactions with other systems, such as the eCB system. Hyperexcited state of the PFC It is important to mention that another IEG, Arc, was also upregulated in the PFC of KO mice, indicating a possible hyperexcited state. Using in vitro electrophysiological measurements of the prelimbic region of the mPFC of stressed Arc-NAPE-PLD WT and KO mice, we found differences in passive and active intrinsic membrane properties. The largest effect was a strong decrease in GABAergic synaptic transmission (by ~50%), which suggests how the putative hyperactivity of this area, as described above, might arise from a reduced inhibition rather than an intrinsic membrane origin or an increase in glutamatergic drive. Additional analysis using evoked inhibitory currents highlights how the reduced inhibitory drive may be explained by a reduced release probability in inhibitory synapses onto pyramidal neurons in KO neurons. This reduced inhibition could be explained by a reduced release probability in inhibitory inputs onto KO neurons [73]. Due to methodological restrictions, we were bound to using a mixed cohort of male and female mice, which could dilute the effect, since there are sex differences in neuronal excitability of mPFC [74]. Furthermore, younger mice were used, which required using a milder stress protocol. In future studies, same sex- and age-matched subjects should be used for assessing electrophysiological properties of the mPFC of Arc-NAPE-PLD KO mice. In summary, our study provides compelling evidence for the involvement of AEA signaling in modulating anxiety-like phenotypes and indicates that targeting of the eCB system provides a very promising and translationally relevant approach for developing successful therapies in the context of stress-induced psychiatric disorders. In particular for the prevention of the emergence of such disorders, when treatment starts immediately after the stressful event. Supplementary information Text Suppl Figure Suppl Figure 1 Suppl Figure 2 Suppl Figure 3 Suppl Figure 4 Suppl Figure 5 Suppl Figure Supplementary information The online version contains supplementary material available at 10.1038/s41398-023-02448-9. Acknowledgements The authors would like to acknowledge Andrea Conrad for her help with planning the mouse breeding, Claudia Schwitter for performing the LC/MS experiments, and Dr. Konstantin Radyushkin for maintaining the Mouse Behavior Unit, where the behavioral experiments were performed. We would also like to thank Michael Plenikowski for the graphical design of the figures. Author contributions EL conceptualized the study, MT and EL planned and performed behavior experiments, tissue collection, RNA and lipid extraction, RT-qPCR and basic statistical analysis. LB performed the analysis of the lipidomic data. DB, DW and MT performed the electrophysiological experiments. NU contributed the NAPE-PLDfl/fl mouse line. DB, DW, SK and HJL performed the analysis and interpretation of the electrophysiological experiment. HT performed all bioinformatic analyses with inputs from SG. Writing – Original Draft Preparation, MT and HT; Writing – Review & Editing, MT, HT, HJL and BL. Funding Acquisition, BL; Project Administration, BL; Supervision, BL and SG. Funding MT was supported by the DFG through subproject B04 of the Collaborative Research Center (CRC) 1193 (“Neurobiology of Resilience”). HT and SG acknowledge funding from the Landesinitiative Rheinland-Pfalz and the ReALity initiative of the Johannes Gutenberg University Mainz. Open Access funding enabled and organized by Projekt DEAL. Data availability The raw data presented in this study are available on request from the corresponding author. Code availability Analysis code is available on request from the corresponding author. Competing interests The authors declare no competing interests. Ethics approval All animal experiments were performed according to the European Community’s Council Directive of 22 September 2010 (2010/63EU) and approved by the respective agency of the State Rhineland-Palatinate (Landesuntersuchungsamt), registration number G-17-1-005. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Deceased: Ermelinda Lomazzo. ==== Refs References 1. Kano M Ohno-Shosaku T Hashimotodani Y Uchigashima M Watanabe M Endocannabinoid-mediated control of synaptic transmission Physiol Rev 2009 89 309 80 10.1152/physrev.00019.2008 19126760 2. Busquets-Garcia A Bains J Marsicano G CB1 Receptor Signaling in the Brain: Extracting Specificity from Ubiquity Neuropsychopharmacol 2018 43 4 20 10.1038/npp.2017.206 3. Hussain Z Uyama T Tsuboi K Ueda N Mammalian enzymes responsible for the biosynthesis of N-acylethanolamines Biochim Biophys Acta Mol Cell Biol Lipids 2017 1862 1546 61 10.1016/j.bbalip.2017.08.006 28843504 4. Ahn K McKinney MK Cravatt BF Enzymatic pathways that regulate endocannabinoid signaling in the nervous system Chem Rev 2008 108 1687 707 10.1021/cr0782067 18429637 5. Leishman E Mackie K Luquet S Bradshaw HB Lipidomics profile of a NAPE-PLD KO mouse provides evidence of a broader role of this enzyme in lipid metabolism in the brain Biochimica Et Biophysica Acta Bba - Mol Cell Biol Lipids 2016 1861 491 500 6. Leung D Saghatelian A Simon GM Cravatt BF Inactivation of N-acyl phosphatidylethanolamine phospholipase D reveals multiple mechanisms for the biosynthesis of endocannabinoids Biochemistry 2006 45 4720 6 10.1021/bi060163l 16605240 7. Tsuboi K Okamoto Y Ikematsu N Inoue M Shimizu Y Uyama T Enzymatic formation of N-acylethanolamines from N-acylethanolamine plasmalogen through N-acylphosphatidylethanolamine-hydrolyzing phospholipase D-dependent and -independent pathways Biochimica Et Biophysica Acta Bba - Mol Cell Biol Lipids 2011 1811 565 77 8. Egertova M Simon GM Cravatt BF Elphick MR Localization of N-acyl phosphatidylethanolamine phospholipase D (NAPE-PLD) expression in mouse brain: A new perspective on N-acylethanolamines as neural signaling molecules J Comp Neurol 2008 506 604 15 10.1002/cne.21568 18067139 9. Maccarrone M Metabolism of the Endocannabinoid Anandamide: Open Questions after 25 Years Front Mol Neurosci 2017 10 166 10.3389/fnmol.2017.00166 28611591 10. Palese F Pontis S Realini N Piomelli D A protective role for N-acylphosphatidylethanolamine phospholipase D in 6-OHDA-induced neurodegeneration Sci Rep 2019 9 15927 10.1038/s41598-019-51799-1 31685899 11. McEwen BS The neurobiology of stress: from serendipity to clinical relevance Brain Res 2000 886 172 89 10.1016/S0006-8993(00)02950-4 11119695 12. Manzanares J Corchero J Fuentes JA Opioid and cannabinoid receptor-mediated regulation of the increase in adrenocorticotropin hormone and corticosterone plasma concentrations induced by central administration of delta(9)-tetrahydrocannabinol in rats Brain Res 1999 839 173 9 10.1016/S0006-8993(99)01756-4 10482810 13. Patel S Roelke CT Rademacher DJ Cullinan WE Hillard CJ Endocannabinoid signaling negatively modulates stress-induced activation of the hypothalamic-pituitary-adrenal axis Endocrinology 2004 145 5431 8 10.1210/en.2004-0638 15331569 14. Wade MR Degroot A Nomikos GG Cannabinoid CB1 receptor antagonism modulates plasma corticosterone in rodents Eur J Pharm 2006 551 162 7 10.1016/j.ejphar.2006.08.083 15. Cota D The role of the endocannabinoid system in the regulation of hypothalamic-pituitary-adrenal axis activity J Neuroendocrinol 2008 20 35 8 10.1111/j.1365-2826.2008.01673.x 18426497 16. Lutz B Marsicano G Maldonado R Hillard CJ The endocannabinoid system in guarding against fear, anxiety and stress Nat Rev Neurosci 2015 16 705 18 10.1038/nrn4036 26585799 17. Radley JJ Sawchenko PE A common substrate for prefrontal and hippocampal inhibition of the neuroendocrine stress response J Neurosci 2011 31 9683 95 10.1523/JNEUROSCI.6040-10.2011 21715634 18. Hill MN McLaughlin RJ Morrish AC Viau V Floresco SB Hillard CJ Suppression of amygdalar endocannabinoid signaling by stress contributes to activation of the hypothalamic-pituitary-adrenal axis Neuropsychopharmacol 2009 34 2733 45 10.1038/npp.2009.114 19. Cajanus K Holmstrom EJ Wessman M Anttila V Kaunisto MA Kalso E Effect of endocannabinoid degradation on pain: role of FAAH polymorphisms in experimental and postoperative pain in women treated for breast cancer Pain 2016 157 361 9 10.1097/j.pain.0000000000000398 26808012 20. Spagnolo PA Ramchandani VA Schwandt ML Kwako LE George DT Mayo LM FAAH Gene Variation Moderates Stress Response and Symptom Severity in Patients with Posttraumatic Stress Disorder and Comorbid Alcohol Dependence Alcohol Clin Exp Res 2016 40 2426 34 10.1111/acer.13210 27716956 21. Pinna G Biomarkers for PTSD at the Interface of the Endocannabinoid and Neurosteroid Axis Front Neurosci 2018 12 482 10.3389/fnins.2018.00482 30131663 22. Navarrete F Garcia-Gutierrez MS Jurado-Barba R Rubio G Gasparyan A Austrich-Olivares A Endocannabinoid System Components as Potential Biomarkers in Psychiatry Front Psychiatry 2020 11 315 10.3389/fpsyt.2020.00315 32395111 23. Dincheva I Drysdale AT Hartley CA Johnson DC Jing D King EC FAAH genetic variation enhances fronto-amygdala function in mouse and human Nat Commun 2015 6 6395 10.1038/ncomms7395 25731744 24. Hill MN McLaughlin RJ Bingham B Shrestha L Lee TT Gray JM Endogenous cannabinoid signaling is essential for stress adaptation Proc Natl Acad Sci 2010 107 9406 11 10.1073/pnas.0914661107 20439721 25. Hill MN Kumar SA Filipski SB Iverson M Stuhr KL Keith JM Disruption of fatty acid amide hydrolase activity prevents the effects of chronic stress on anxiety and amygdalar microstructure Mol Psychiatr 2013 18 1125 35 10.1038/mp.2012.90 26. Bluett RJ Gamble-George JC Hermanson DJ Hartley ND Marnett LJ Patel S Central anandamide deficiency predicts stress-induced anxiety: behavioral reversal through endocannabinoid augmentation Transl Psychiatry 2014 4 e408 10.1038/tp.2014.53 25004388 27. Moreira FA Kaiser N Monory K Lutz B Reduced anxiety-like behaviour induced by genetic and pharmacological inhibition of the endocannabinoid-degrading enzyme fatty acid amide hydrolase (FAAH) is mediated by CB1 receptors Neuropharmacol 2008 54 141 50 10.1016/j.neuropharm.2007.07.005 28. Haller J Barna I Barsvari B Pelczer KG Yasar S Panlilio LV Interactions between environmental aversiveness and the anxiolytic effects of enhanced cannabinoid signaling by FAAH inhibition in rats Psychopharmacol 2009 204 607 16 10.1007/s00213-009-1494-7 29. McLaughlin RJ Hill MN Gorzalka BB A critical role for prefrontocortical endocannabinoid signaling in the regulation of stress and emotional behavior Neurosci Biobehav Rev 2014 42 116 31 10.1016/j.neubiorev.2014.02.006 24582908 30. Bortolato M Mangieri RA Fu J Kim JH Arguello O Duranti A Antidepressant-like activity of the fatty acid amide hydrolase inhibitor URB597 in a rat model of chronic mild stress Biol Psychiatry 2007 62 1103 10 10.1016/j.biopsych.2006.12.001 17511970 31. Hill MN Lee FS Endocannabinoids and Stress Resilience: Is Deficiency Sufficient to Promote Vulnerability? Biol Psychiatry 2016 79 792 3 10.1016/j.biopsych.2016.03.2099 27130852 32. Worley NB Hill MN Christianson JP Prefrontal endocannabinoids, stress controllability and resilience: A hypothesis Prog Neuropsychopharmacol Biol Psychiatry 2018 85 180 8 10.1016/j.pnpbp.2017.04.004 28392485 33. Morena M Aukema RJ Leitl KD Rashid AJ Vecchiarelli HA Josselyn SA Upregulation of Anandamide Hydrolysis in the Basolateral Complex of Amygdala Reduces Fear Memory Expression and Indices of Stress and Anxiety J Neurosci 2019 39 1275 92 10.1523/JNEUROSCI.2251-18.2018 30573646 34. Mayo LM Asratian A Linde J Morena M Haataja R Hammar V Elevated Anandamide, Enhanced Recall of Fear Extinction, and Attenuated Stress Responses Following Inhibition of Fatty Acid Amide Hydrolase: A Randomized, Controlled Experimental Medicine Trial Biol Psychiatry 2020 87 538 47 10.1016/j.biopsych.2019.07.034 31590924 35. Micale V Stepan J Jurik A Pamplona FA Marsch R Drago F Extinction of avoidance behavior by safety learning depends on endocannabinoid signaling in the hippocampus J Psychiatr Res 2017 90 46 59 10.1016/j.jpsychires.2017.02.002 28222356 36. Hartmann A Fassini A Scopinho A Correa FM Guimaraes FS Lisboa SF Role of the endocannabinoid system in the dorsal hippocampus in the cardiovascular changes and delayed anxiety-like effect induced by acute restraint stress in rats J Psychopharmacol 2019 33 606 14 10.1177/0269881119827799 30789299 37. Griebel G Stemmelin J Lopez-Grancha M Fauchey V Slowinski F Pichat P The selective reversible FAAH inhibitor, SSR411298, restores the development of maladaptive behaviors to acute and chronic stress in rodents Sci Rep. 2018 8 2416 10.1038/s41598-018-20895-z 29403000 38. Buran I Etem EO Tektemur A Elyas H Treatment with TREK1 and TRPC3/6 ion channel inhibitors upregulates microRNA expression in a mouse model of chronic mild stress Neurosci Lett 2017 656 51 57 10.1016/j.neulet.2017.07.017 28716528 39. Gray JM Chaouloff F Hill MN To Stress or Not to Stress: A Question of Models Curr Protoc Neurosci 2015 70 1 22 40. Golden SA Covington HE 3rd Berton O Russo SJ A standardized protocol for repeated social defeat stress in mice Nat Protoc 2011 6 1183 10.1038/nprot.2011.361 21799487 41. DeNardo L Luo L Genetic strategies to access activated neurons Curr Opin Neurobiol 2017 45 121 9 10.1016/j.conb.2017.05.014 28577429 42. Denny CA Kheirbek MA Alba EL Tanaka KF Brachman RA Laughman KB Hippocampal Memory Traces Are Differentially Modulated by Experience, Time, and Adult Neurogenesis Neuron 2014 83 189 201 10.1016/j.neuron.2014.05.018 24991962 43. Mo A Mukamel EA Davis FP Luo C Henry GL Picard S Epigenomic Signatures of Neuronal Diversity in the Mammalian Brain Neuron 2015 86 1369 84 10.1016/j.neuron.2015.05.018 26087164 44. Guenthner CJ Miyamichi K Yang HH Heller HC Luo L Permanent genetic access to transiently active neurons via TRAP: targeted recombination in active populations Neuron 2013 78 773 84 10.1016/j.neuron.2013.03.025 23764283 45. Krishnan V Han M-H Graham DL Berton O Renthal W Russo SJ Molecular Adaptations Underlying Susceptibility and Resistance to Social Defeat in Brain Reward Regions Cell 2007 131 391 404 10.1016/j.cell.2007.09.018 17956738 46. Ruehle S Remmers F Romo-Parra H Massa F Wickert M Wörtge S Cannabinoid CB1 receptor in dorsal telencephalic glutamatergic neurons: distinctive sufficiency for hippocampus-dependent and amygdala-dependent synaptic and behavioral functions J Neurosci 2013 33 10264 77 10.1523/JNEUROSCI.4171-12.2013 23785142 47. Steru L Chermat R Thierry B Simon P The tail suspension test: a new method for screening antidepressants in mice Psychopharmacol 1985 85 367 70 10.1007/BF00428203 48. Deacon RMJ Assessing nest building in mice Nat Protoc 2006 1 1117 10.1038/nprot.2006.170 17406392 49. Post JM, Lerner R, Schwitter C, Lutz B, Lomazzo E, Bindila L. Lipidomics and Transcriptomics in Neurological Diseases. J Vis Exp. 2022. 10.3791/59423. 50. Paxinos G Franklin KBJ The Mouse Brain in Stereotaxic Coordinates 2001 Academic Press 51. R: A language and environment for statistical computing. R Core Team, R Foundation for Statistical Computing: Vienna, Austria, 2019. https://www.R-project.org/. 52. Todorov H Searle-White E Gerber S Applying univariate vs. multivariate statistics to investigate therapeutic efficacy in (pre)clinical trials: A Monte Carlo simulation study on the example of a controlled preclinical neurotrauma trial Plos One 2020 15 e0230798 10.1371/journal.pone.0230798 32214370 53. Scarante FF Vila-Verde C Detoni VL Ferreira-Junior NC Guimaraes FS Campos AC Cannabinoid Modulation of the Stressed Hippocampus Front Mol Neurosci 2017 10 411 10.3389/fnmol.2017.00411 29311804 54. Vinod KY Xie S Psychoyos D Hungund BL Cooper TB Tejani-Butt SM Dysfunction in fatty acid amide hydrolase is associated with depressive-like behavior in Wistar Kyoto rats Plos One 2012 7 e36743 10.1371/journal.pone.0036743 22606285 55. Binder EB The role of FKBP5, a co-chaperone of the glucocorticoid receptor in the pathogenesis and therapy of affective and anxiety disorders Psychoneuroendocrinol 2009 34 S186 95 10.1016/j.psyneuen.2009.05.021 56. Touma C Gassen NC Herrmann L Cheung-Flynn J Bull DR Ionescu IA FK506 binding protein 5 shapes stress responsiveness: modulation of neuroendocrine reactivity and coping behavior Biol Psychiatry 2011 70 928 36 10.1016/j.biopsych.2011.07.023 21907973 57. Hartmann J Wagner KV Liebl C Scharf SH Wang XD Wolf M The involvement of FK506-binding protein 51 (FKBP5) in the behavioral and neuroendocrine effects of chronic social defeat stress Neuropharmacol 2012 62 332 9 10.1016/j.neuropharm.2011.07.041 58. Codagnone MG Kara N Ratsika A Levone BR Wouw M van de, Tan LA Inhibition of FKBP51 induces stress resilience and alters hippocampal neurogenesis Mol Psychiatry 2022 27 4928 38 10.1038/s41380-022-01755-9 36104438 59. Armario A The hypothalamic-pituitary-adrenal axis: what can it tell us about stressors? CNS Neurological Disord - Drug Targets 2006 5 485 501 10.2174/187152706778559336 60. Heilig M Koob GF Ekman R Britton KT Corticotropin-releasing factor and neuropeptide Y: role in emotional integration Trends Neurosci 1994 17 80 5 10.1016/0166-2236(94)90079-5 7512773 61. McGuire JL Larke LE Sallee FR Herman JP Sah R Differential Regulation of Neuropeptide Y in the Amygdala and Prefrontal Cortex during Recovery from Chronic Variable Stress Front Behav Neurosci 2011 5 54 10.3389/fnbeh.2011.00054 21954381 62. Reichmann F Holzer P Neuropeptide Y: A stressful review Neuropeptides 2016 55 99 109 10.1016/j.npep.2015.09.008 26441327 63. Sajdyk TJ Johnson PL Leitermann RJ Fitz SD Dietrich A Morin M Neuropeptide Y in the amygdala induces long-term resilience to stress-induced reductions in social responses but not hypothalamic-adrenal-pituitary axis activity or hyperthermia J Neurosci 2008 28 893 903 10.1523/JNEUROSCI.0659-07.2008 18216197 64. Cohen H Liu T Kozlovsky N Kaplan Z Zohar J Mathe AA The neuropeptide Y (NPY)-ergic system is associated with behavioral resilience to stress exposure in an animal model of post-traumatic stress disorder Neuropsychopharmacol 2012 37 350 63 10.1038/npp.2011.230 65. Xapelli S Bernardino L Ferreira R Grade S Silva AP Salgado JR Interaction between neuropeptide Y (NPY) and brain-derived neurotrophic factor in NPY-mediated neuroprotection against excitotoxicity: a role for microglia Eur J Neurosci 2008 27 2089 102 10.1111/j.1460-9568.2008.06172.x 18412629 66. Daskalakis NP Kloet ERD Yehuda R Malaspina D Kranz TM Early Life Stress Effects on Glucocorticoid-BDNF Interplay in the Hippocampus Front Mol Neurosci 2015 8 68 10.3389/fnmol.2015.00068 26635521 67. Leschik J Gentile A Cicek C Péron S Tevosian M Beer A Brain-derived neurotrophic factor expression in serotonergic neurons improves stress resilience and promotes adult hippocampal neurogenesis Prog Neurobiol 2022 217 102333 10.1016/j.pneurobio.2022.102333 35872219 68. Zaletel I Filipovic D Puskas N Hippocampal BDNF in physiological conditions and social isolation Rev Neurosci 2017 28 675 92 10.1515/revneuro-2016-0072 28593903 69. Maymon N Zer-Aviv TM Sabban EL Akirav I Neuropeptide Y and cannabinoids interaction in the amygdala after exposure to shock and reminders model of PTSD Neuropharmacol 2020 162 107804 10.1016/j.neuropharm.2019.107804 70. Duclot F Kabbaj M The Role of Early Growth Response 1 (EGR1) in Brain Plasticity and Neuropsychiatric Disorders Front Behav Neurosci 2017 11 35 10.3389/fnbeh.2017.00035 28321184 71. Chen H Amazit L Lombes M Menuet DL Crosstalk Between Glucocorticoid Receptor and Early-growth Response Protein 1 Accounts for Repression of Brain-derived Neurotrophic Factor Transcript 4 Expression Neurosci 2019 399 12 27 10.1016/j.neuroscience.2018.12.012 72. Morena M Patel S Bains JS Hill MN Neurobiological Interactions Between Stress and the Endocannabinoid System Neuropsychopharmacol 2016 41 80 102 10.1038/npp.2015.166 73. Kavalali ET The mechanisms and functions of spontaneous neurotransmitter release Nat Rev Neurosci 2015 16 5 16 10.1038/nrn3875 25524119 74. Velasco EMF de, Hearing M Xia Z Victoria NC Luján R Wickman K Sex differences in GABABR-GIRK signaling in layer 5/6 pyramidal neurons of the mouse prelimbic cortex Neuropharmacol 2015 95 353 60 10.1016/j.neuropharm.2015.03.029
PMC010xxxxxx/PMC10167354.txt
==== Front Nat Commun Nat Commun Nature Communications 2041-1723 Nature Publishing Group UK London 37156840 38171 10.1038/s41467-023-38171-8 Article Cell-selective proteomics segregates pancreatic cancer subtypes by extracellular proteins in tumors and circulation Swietlik Jonathan J. 1 http://orcid.org/0000-0001-5986-4864 Bärthel Stefanie 234 http://orcid.org/0000-0001-9269-2582 Falcomatà Chiara 234 Fink Diana 5 Sinha Ankit 6 Cheng Jingyuan 1 Ebner Stefan 5 Landgraf Peter 7 Dieterich Daniela C. 78 Daub Henrik 9 http://orcid.org/0000-0001-5874-0210 Saur Dieter [email protected] 234 http://orcid.org/0000-0003-1000-7989 Meissner Felix [email protected] 15 1 grid.418615.f 0000 0004 0491 845X Experimental Systems Immunology, Max Planck Institute of Biochemistry, Martinsried, Germany 2 grid.7497.d 0000 0004 0492 0584 Division of Translational Cancer Research, German Cancer Research Center and German Cancer Consortium, Heidelberg, Germany 3 grid.6936.a 0000000123222966 Chair of Translational Cancer Research and Institute of Experimental Cancer Therapy, University Hospital Rechts der Isar, School of Medicine, Technical University of Munich, Munich, Germany 4 grid.6936.a 0000000123222966 Center for Translational Cancer Research (TranslaTUM), School of Medicine, Technical University of Munich, Munich, Germany 5 grid.10388.32 0000 0001 2240 3300 Institute of Innate Immunity, Department of Systems Immunology and Proteomics, Medical Faculty, University of Bonn, Bonn, Germany 6 grid.418615.f 0000 0004 0491 845X Department of Proteomics and Signal Transduction, Max Planck Institute of Biochemistry, Martinsried, Germany 7 grid.5807.a 0000 0001 1018 4307 Institute for Pharmacology and Toxicology, Otto-von-Guericke-University Magdeburg, Magdeburg, Germany 8 grid.452320.2 0000 0004 0404 7236 Center for Behavioral Brain Sciences, Magdeburg, Germany 9 NEOsphere Biotechnologies GmbH, Martinsried, Germany 8 5 2023 8 5 2023 2023 14 26426 8 2022 14 4 2023 © The Author(s) 2023 https://creativecommons.org/licenses/by/4.0/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Cell-selective proteomics is a powerful emerging concept to study heterocellular processes in tissues. However, its high potential to identify non-cell-autonomous disease mechanisms and biomarkers has been hindered by low proteome coverage. Here, we address this limitation and devise a comprehensive azidonorleucine labeling, click chemistry enrichment, and mass spectrometry-based proteomics and secretomics strategy to dissect aberrant signals in pancreatic ductal adenocarcinoma (PDAC). Our in-depth co-culture and in vivo analyses cover more than 10,000 cancer cell-derived proteins and reveal systematic differences between molecular PDAC subtypes. Secreted proteins, such as chemokines and EMT-promoting matrisome proteins, associated with distinct macrophage polarization and tumor stromal composition, differentiate classical and mesenchymal PDAC. Intriguingly, more than 1,600 cancer cell-derived proteins including cytokines and pre-metastatic niche formation-associated factors in mouse serum reflect tumor activity in circulation. Our findings highlight how cell-selective proteomics can accelerate the discovery of diagnostic markers and therapeutic targets in cancer. “In-depth cell-selective proteomics and secretomics has remained challenging. Here, the authors devise an optimised azidonorleucine labelling, mass spectrometry method and detect over 10,000 proteins in a pancreatic ductal adenocarcinoma model. Subject terms Tumour immunology Proteomics Proteome informatics https://doi.org/10.13039/501100001659 Deutsche Forschungsgemeinschaft (German Research Foundation) Project-ID 360372040 (SFB 1335), Cluster of Excellence EXC 2151 ImmunoSensation2 Project-ID 360372040 (SFB 1335) Project-ID 450149205 (TRR333) Project-ID 165054336 (SFB 914) Swietlik Jonathan J. Meissner Felix https://doi.org/10.13039/501100004189 Max-Planck-Gesellschaft (Max Planck Society) n.a. Meissner Felix issue-copyright-statement© Springer Nature Limited 2023 ==== Body pmcIntroduction Cells in multicellular organisms adapt their phenotypes and function by crosstalk with other cell types. Short- and long-ranged intercellular signals are an integral part of organismal homeostasis and, when altered, drive the pathogenesis of diverse diseases. For example, in cancer, vivid interactions between transformed cells and non-transformed stromal cells promote or inhibit tumor development, metastasis, and the efficacy of drugs. A rising incidence and high lethality make pancreatic ductal adenocarcinoma (PDAC) one of the leading causes of cancer-related deaths1. Since PDAC is typically discovered in advanced stages and refractory to most treatment modalities, there is a pressing need for more effective therapy and biomarkers that allow early detection. However, hallmark features of PDAC, such as a dense and fibrotic stroma, an immunosuppressive tumor microenvironment (TME), and often low neoplastic cellularity, exacerbate its molecular characterization and therapy development2,3. Based on the transcriptional profile and pathological features, PDAC is stratified into two major molecular subtypes4. Classical PDAC is characterized by a well-differentiated histopathology and epithelial gene expression signature. In contrast, mesenchymal (basal-like) PDAC shows an undifferentiated, non-glandular histology, a mesenchymal gene expression profile, and is associated with a poor prognosis and high resistance to standard-of-care chemotherapy compared to the classical subtype5–9. Despite the substantial clinicopathological differences between the two PDAC subtypes, the underlying differences in the intercellular signaling of cancer cells with their TME have not been studied systematically so far. Important insights into tumor cell composition and phenotype have been gained by systems-wide transcriptional approaches. However, the correlation between mRNA and protein copy numbers can vary widely10,11, especially for proteins with roles in intercellular crosstalk12,13. Therefore, systems-wide and unbiased tools for comprehensive quantitative protein analyses can provide unique perspectives on the context-dependent crosstalk of cancer cells with their microenvironment14. Mass spectrometry (MS)-based proteomics is today’s gold standard for high throughput protein analysis and has significantly improved our understanding of cancer pathogenesis15–18. The combination of proteomics with cell-selective metabolic protein labeling strategies promises to resolve context-dependent cell behavior and interaction in complex heterocellular systems like tumors. One of the emerging methods uses the specially engineered methionyl-tRNA-synthetaseL274G (MetRS*), which enables the time-controlled and cell-specific introduction of the non-canonical amino acid azidonorleucine (Anl) into proteomes19–21. Azide-alkyne click chemistry allows the subsequent extraction of MetRS*-expressing cell-derived proteins from cell mixtures. Successful application in living animals provided evidence for broad tissue compatibility22 and revealed, for example, differential expression of 200 proteins in hippocampal excitatory neurons in mice exposed to an enriched environment23. In contrast to cell-sorting-based strategies such as FACS or MACS, intact tissues are snap-frozen directly after harvesting and subsequently lysed without cell dissociation. This effectively avoids cell-damage-related losses, selection bias for more robust cell populations, and potential protein expression or modification state artifacts by stresses and environmental changes during the enzymatic and mechanical treatment necessary to extract cells from tissues24–27. However, the achieved proteome coverage has generally been low, and even the deepest studies remained under 4000 specifically enriched proteins23,28, leaving open the feasibility of comprehensive Anl enrichment-based proteomics analysis. Here, we developed an improved workflow that enables an unprecedented proteomics depth for cell type-specific cellular proteome and secretome profiling in vitro and in vivo. This vastly increased the detection capacity of often low abundant intercellular signaling proteins such as secreted cytokines or receptors and therefore raised the potential for MetRS*/ Anl-based cellular communication analyses. We applied our approaches in the context of primary PDAC co-culture and orthotopic transplantation models and demonstrated unique advantages in capturing extracellular proteins compared to conventional cell sorting-based proteomics. We used the strength of our comprehensive cell-type specific proteomics workflow to reveal functional differences between classical and mesenchymal PDAC subtypes in tumors and circulation, such as context-specific secretion of cancer cell-derived EMT-promoting molecules and immunomodulators that correlated with differential immune cell recruitment in vivo, as well as distinct qualitative and quantitative contributions of cancer cell-derived proteins to the tumor extracellular matrix (ECM). Results An improved workflow enables highly efficient and specific cell-selective enrichment of proteins Conceptually, methionyl-tRNA synthetaseL274G (MetRS*)-based azidonorleucine (Anl) labeling offers unique possibilities for analyzing intercellular interactions in complex heterocellular systems. However, the achieved proteomic depth in our initial experiments and previously published MetRS*-based studies did not exceed 4000 proteins23,28 and was therefore significantly lower than state-of-the-art with modern mass spectrometers and software29, limiting the discovery potential. Hence, we set out to identify and overcome technical bottlenecks. We first evaluated the Anl-incorporation rates of MetRS*-expressing cells in vitro by conventional MS-based shotgun proteomics without specific enrichment. Quantifying Anl-containing peptides compared to their unmodified counterparts showed that Anl incorporation was indeed highly specific to MetRS*-expressing cells but much slower than the incorporation of methionine (l-methionine-methyl-13C,d3) or the MetRS*-independent Met-substitute azidohomoalanine (Aha) (Supplementary Fig. 1a, b). Furthermore, Anl labeling is strongly dosage-dependent and reduced with methionine competition, as shown in previous studies19,30,31. We reasoned that the Anl-protein abundance would be very low in most applications, especially in vivo, considering often pronounced cell type heterogeneity and limited Anl bioavailability in tissues. Consequently, the demands for both recovery and specificity of the enrichment workflow are very high when aiming for deep proteomics analyses. We chose a straightforward copper(I)-catalyzed azide-alkyne cycloaddition (CuAAC) and alkyne agarose-based strategy for scalability and high reaction rates32 as the basis for protocol optimization. We individually evaluated key experimental steps to improve protein extraction from tissue and click chemistry efficiency by systematic implementation of previous findings33,34 and empirical testing of reactant ratios, buffer components, and new reagents, including next-generation Cu(I)-stabilizing agents35. A direct comparison of our improved alkyne-agarose CuAAC protocol with frequently used dibenzocyclooctyne (DBCO) resin- and cleavable disulfide biotin alkyne-tag (DST)-based procedures (Fig. 1a) demonstrated substantial advantages: Using MetRS*-expressing and negative control wild-type primary PDAC cells that were both incubated in Anl-containing media, our protocol showed minimal unspecific background and a drastically increased yield of specifically enriched peptides (Fig. 1b). This advantage translated well into the MS analysis: While DST-based enrichment provided good specificity but reduced overall coverage, DBCO-based enrichment led to many identifications in both MetRS* and negative control samples, concordant with higher side reactivity of strained alkynes32,36. In contrast, our protocol yielded deep proteome coverage but with the fewest identifications in negative controls (Fig. 1c). The technical reproducibility was equal to or better than alternative protocols, with 84% of all MetRS* sample identifications quantified in all three replicates and a median precursor coefficient of variation (CV) of 11.5% (Supplementary Fig. 2a, b). Importantly, low overlap and high-intensity differences of proteins identified in both MetRS* samples and controls demonstrated very low background interference from unspecific enrichment with our workflow (Fig. 1d). We defined proteins as specifically enriched if they were exclusively identified in MetRS* samples or quantified with at least threefold higher intensity than in negative controls and excluded all other proteins from further analysis, as described by Alvarez-Castelao et al.23,37. Accordingly, our workflow identified a total of 6576 specifically enriched protein groups (compared to 4416 and 4736 with DST- or DBCO-based enrichment, respectively), including almost all proteins covered with both other methods together plus 1039 exclusive identifications (Supplementary Fig. 3).Fig. 1 A sensitive workflow for comprehensive cell type-selective proteomics and secretomics. a Scheme of cell-selective proteomics workflows: The methionyl-tRNA synthetaseL274G (MetRS*) activates azidonorleucine (Anl) by loading it onto methionyl-tRNAs. MetRS*-expressing cells incorporate Anl as a methionine substitute into newly synthesized proteins. Lentivirally transduced primary MetRS*-expressing or wild-type (Ctrl) PDAC cells isolated from mouse tumors with a conditional pancreatic expression of KrasG12D were grown for 8 h in Met-depleted medium supplemented with 4 mM Anl. 1 × 107 MetRS* and Ctrl cells were processed by DST enrichment, DBCO enrichment, and our improved alkyne-agarose CuAAC enrichment protocols (n = 3, workflow replicates). b Peptide yields (mean ± SD) determined by absorbance at 280 nm after enrichment, digestion, and solid phase peptide extraction. c Identified protein groups (mean ± SD) after MS-based analysis using 2 h chromatographic gradient length and data-dependent acquisition (DDA). d Intensity ratios of proteins identified in MetRS* and Ctrl samples. Counts of overlapping identifications with ratios are indicated. e Specifically enriched protein groups (exclusive or >3-fold higher intensity compared to Ctrl samples) identified after alkyne-agarose enrichment and single run DDA, DDA analysis of 16 fractions separated by offline high-pH reverse phase chromatography, or single run data-independent acquisition (DIA) (mean ± SD, fractionation n = 1, single shots n = 3, workflow replicates). The latter was used for all further experiments. f Scheme of cell-selective secretomics workflow: MetRS* and Ctrl 8661 PDAC cells were cultured for 8 h in 5% FBS containing Met-depleted medium with 4 mM Anl (n = 3, workflow replicates). MetRS*-expressing cell-derived Anl-proteins were enriched from cell supernatants after buffer exchange and concentration. g Specifically-enriched PDAC cell-released proteins ranked by label-free quantification (LFQ) intensity. Proteins with cytokine function are indicated. Source data are provided as a Source Data file. To optimize deep proteomics investigations, we combined our workflow with offline high pH reverse phase fractionation of peptides after enrichment and digestion, resulting in the identification of 10,146 specifically enriched protein groups, demonstrating exceptional proteome coverage (Fig. 1e). Furthermore, using a data-independent acquisition (DIA) method, we achieved an average of 8770 specifically enriched protein groups per sample in 2-h runs without fractionation. The use of DIA also improved data completeness between replicates and decreased precursor CVs compared to data-dependent acquisition (DDA) (Supplementary Fig. 2c, d). To further evaluate the technical reproducibility of our enrichment, we repeated the experiment with ten negative control replicates, confirming the previously observed high signal-to-noise ratio between specifically enriched proteins and unspecific background (Supplementary Fig. 4a, c). While the very low signal intensity in negative controls caused more stochastic identifications than in MetRS* samples (Supplementary Fig. 4b), results remained very consistent when control samples were divided into groups of three and used separately to evaluate background interference in MetRS* samples (Supplementary Fig. 4c). The vast majority of proteins with sparse identifications in controls had high ratios far above our chosen specificity cutoff (Supplementary Fig. 4d). Conversely, the majority of proteins with lower MetRS*/Ctrl ratios had very high data completeness. Thus, not only were there a very limited number of proteins with higher background interference overall, but the controls were also effective in capturing most of these proteins consistently. We applied the described filtering strategy to all subsequent MetRS* experiments in this study, using at least three experiments with wild-type cells as negative controls for corresponding MetRS* sample groups to define specifically enriched proteins and ensure high confidence in cell selectivity. To enable both very deep and MS time-efficient analysis with high throughput for larger-scale experiments, we used our enrichment workflow together with DIA single-shot analyses. Comprehensive cell-selective secretomics analysis in serum-containing media Encouraged by the strongly increased specific peptide recovery, we next aimed to adapt our methods for investigations of intercellular signaling, specifically for the comprehensive analysis of secreted proteins. Previously, non-cell-selective incorporation of azide amino acids has improved the detection of comparably low abundant secreted cellular proteins in the presence of highly abundant serum proteins in serum-containing conditioned media38. MetRS*-based Anl-labeling could expand this concept for cell-selective analyses in heterocellular systems such as co-culture experiments. To establish proof-of-concept for in-depth secretomics with our enrichment protocol, we analyzed supernatants of primary PDAC cells in the presence of 5% serum (Fig. 1f). This yielded deep coverage of PDAC cell-released proteins, with a total of 2229 specifically enriched protein groups and 788 protein groups annotated with UniProtKB Keywords “secreted” and/or “signal.” Of those, 103 protein groups are known ligands for intercellular communication according to CellPhoneDB39, including 46 with described cytokine function (Fig. 1g). Despite their often small size and low abundance, 83 (81%) and 41 (89%) of the detected intercellular signaling proteins and cytokines were identified with at least two peptides. Increased yields and extracellular protein coverage of MetRS*-based cell-selective proteomics compared to FACS in vivo A key feature of MetRS*-based Anl labeling is its applicability in living animals. As shown previously in Falcomatà and Bärthel et al.40, we modeled molecular PDAC subtypes in vivo by orthotopic transplantation of primary low-passaged cancer cells in the pancreas of fully immunocompetent syngeneic mice. We evaluated our enrichment workflow with tissue samples from this model by directly comparing Anl-based enrichments with conventional fluorescence-activated cell sorting (FACS) from MetRS* and eGFP co-expressing cells. After cell injection and an initial tumor growth period, we supplemented Anl by intraperitoneal injection and then used one-half of each tumor for Anl-enrichment or FACS (Fig. 2a). In total, 13–17% of the dissociated cells were cancer cells, as indicated by eGFP-fluorescence (Supplementary Fig. 5).Fig. 2 MetRS*- and FACS-based cancer cell-selective proteomics in vivo. a Scheme of PDAC transplantation for cell-selective proteomics: MetRS*-eGFP expressing 8661 PDAC (>90% eGFP-positive cells before transplantation, see Supplementary Fig. 5a) and wild-type (Ctrl) cells were orthotopically transplanted into fully immunocompetent syngeneic mice (n = 3, biological replicates). After a 16 days tumor growth period, mice were interperitoneally injected with Anl twice daily for 5 days. Afterward, tumors were harvested and cut in half. One half was snap-frozen for subsequent click chemistry enrichment, the other half was used fresh for cell dissociation and eGFP-FACS. b Peptide yields (mean ± SD) determined by absorbance at 280 nm. c Exclusively identified and overlap of (specifically enriched) cancer cell-derived protein groups with either method. d Distribution of precursor coefficients of variation (CVs) between biological replicates. e Analysis of enriched GO annotations (Fisher’s exact test) within exclusively identified proteins with either method compared to all other identified proteins (full list in Supplementary Data 4). Source data are provided as a Source Data file. Peptide yields revealed striking differences with an over 50-fold higher average recovery of cancer cell-derived proteins by click chemistry enrichment compared to FACS, indicating significant cell losses during the dissociation and sorting procedure (Fig. 2b). However, both methods yielded a sufficient peptide amount for single-shot proteomics analyses with modern MS instrumentation. Both methods resulted in more than 8100 protein groups, with a lower median coefficient of variation between replicates for Anl-enrichment samples, indicating better quantitative precision (Fig. 2c, d). While around 70% of the identified protein groups overlapped between both FACS and Anl-enrichment-based analysis (Fig. 2c), exclusive identifications with each method revealed distinct strengths. Flow cytometry-sorted samples showed, for example, enrichment of transmembrane proteins, likely facilitated by strong ionic detergent-based lysis, which enhances transmembrane protein extraction and digestion41,42 but can interfere with CuAAC reactions34. Conversely, cell-selective labeling captured proteins released by cells, such as ECM components and cytokines, specifically well (Fig. 2e). We primarily attribute this to the enrichment of proteins from the interstitial space in tumors, which are accessible for MetRS*-based cell-selective proteomics but lost in tissue-dissociation and sorting-based protocols. Co-culture promotes inflammatory responses of PDAC cells and polarization of primary macrophages After closing the gap to state-of-the-art proteomics performance and extending Anl labeling applications to in-depth cell-selective secretomics, we applied our toolkit to study pancreatic cancer biology. Both the coverage of lower abundant proteins and cell type-resolved information on released signaling proteins are invaluable for understanding intercellular communication. In cancer, complex interactions between transformed cells and tumor stromal cells shape each other’s phenotypes and the overall tumor biology. Macrophages, for example, are a major component of solid tumors and are among the earliest tumor-infiltrating immune cells in PDAC43,44. To evaluate the potential of MetRS*-based cell-type specific proteomics for the molecular dissection of such intercellular crosstalk, we explored the bidirectional interaction between PDAC cells and macrophages in a controlled in vitro setting. All primary PDAC cell cultures were derived from a genetically engineered KrasG12D -driven autochthonous mouse PDAC model45. They are representative of the classical subtype, displaying an epithelial morphology (“8661” and “8442”), or of the basal-like mesenchymal subtype (“8513” and “9091”), characterized by increased oncogenic Kras gene dosage (Kras-mut iGD) and a particularly unfavorable prognosis. By generating LysM-Cre-MetRS* mice, which specifically express MetRS* in the myeloid compartment, we were able to obtain primary MetRS*-expressing bone marrow-derived macrophages (BMMs). We then cultured the four PDAC lines, and the BMMs alone or in co-culture (Fig. 3a) and cell type-selectively analyzed proteins from cells and cell supernatants. Principal component analyses (PCAs) showed reciprocal adaptions of cancer cells and BMMs to co-culture with changes in both global proteome expression and protein secretion, although less clear for PDAC secretomes (Supplementary Fig. 6). PCAs further indicated distinct differences between PDAC subtypes and PDAC line-specific BMM responses.Fig. 3 Co-culture of cancer cells with primary macrophages induces bidirectional adaptions. a Scheme of PDAC and primary macrophage cells in co-culture. Wild-type (WT) or MetRS*-expressing mesenchymal (8513 and 9091) or classical (8442 and 8661) PDAC cells and BMMs were cultured in isolation or co-culture for 36 h with Anl labeling during the last 8 h (n = 3, workflow replicates). Asterisks indicate MetRS* expression. b–e Strongly enriched gene ontology (GO) terms and UniProtKB keywords in upregulated proteins after PDAC—BMM co-culture compared to each cell type in isolation (two-sided 1D annotation enrichment128 (full list in Supplementary Data 4)). f Heatmap of protein intensities associated with macrophage polarization states in BMM proteomes and secretomes. We first investigated broad trends and processes in the proteome and secretome dynamics between each cell type in isolation and co-culture. A gene ontology (GO)46 enrichment analysis showed increased expression of antigen-presentation and major histocompatibility complex (MHC) class I-associated proteins in classical, and to a lesser degree in mesenchymal PDAC cells (Fig. 3b), which was previously observed in breast cancer cells co-cultured with macrophages in transwell systems47. Interaction with BMMs also induced strong upregulation of chemokine production and interferon response signatures in both classical and mesenchymal PDAC cells, while, in particular, mesenchymal cells strongly increased structural matrix protein deposition (Fig. 3b, d). Secretomics analysis at the individual protein level revealed secretion of complex immunomodulatory signals with pronounced differences between PDAC subtypes and significant changes upon interaction with macrophages (68 signaling proteins with cytokine function and significant abundance differences (ANOVA, FDR = 5%, S0 = 0.1) between subtypes and culture conditions, see Supplementary Fig. 7). For example, co-culture induced increased Il6 release by both subtypes but with much higher levels in classical PDAC cells, whereas specifically mesenchymal PDAC cells strongly increased secretion of CCL8 and 9. Moreover, significant enrichment of surface-exposed plasma membrane proteins in secretomes, including MHCI proteins, suggested increased shedding activity in cancer cells (Fig. 3d). Upon interaction with classical PDAC cells, BMMs expressed higher levels of proteins associated with exogenous antigen presentation, T cell regulation, and regulation of key cytokines involved in the coordination of pro- and antitumoral response reactions48,49 (Fig. 3c). Although trends could also be observed upon co-culture with mesenchymal PDAC cells, effects were less pronounced and did not reach statistical significance. However, BMM secretomes showed strong enrichment of immunomodulatory proteins, hormones, and growth factors, and extracellular matrix (ECM)-modifying proteins after co-culture with both PDAC subtypes (Fig. 3e). In addition to many cytokines, interaction with cancer cells broadly induced ECM regulators such as matrix metalloproteinases (MMPs), a disintegrin and metalloproteinases (ADAMs), and ADAMs with thrombospondin motifs (ADAMTSs), with crucial functions in cancer50,51 (Supplementary Fig. 7). Co-cultured macrophages acquire TAM-like features driven by a complex mix of cancer cell-secreted and -displayed signaling proteins To further evaluate the cancer-cell-induced macrophage states in co-culture, we annotated proteins from BMM-selective secretome and cellular proteome datasets using a panel of markers commonly associated with macrophage polarization52. M1 and M2 states exemplify broadly clustered extremes on a spectrum of macrophage states—with M1 being associated with interferon and Toll-like receptor signals and efficient production of effector molecules and inflammatory cytokines, and M2 macrophages with the resolution of inflammation or TH2 response-driven physiological reponses52,53. As an experimental reference, we stimulated MetRS*-expressing BMMs with lipopolysaccharide (LPS), a Toll-like receptor 4 (TLR4) agonist, and acquired their proteome and secretome profiles. As expected, LPS-stimulated BMMs showed exclusive expression and strong upregulation of M1-associated marker proteins compared to unstimulated cells (Fig. 3f). PDAC co-cultured macrophage M1-associated marker expression was detected only sporadically and mostly at basal levels, except for increased secretion of the pro-inflammatory cytokines Il6 and Tnf upon interaction with 8661 classical PDAC cells. Instead, co-culture primarily induced upregulation of M2-associated markers such as Arg1 and Chil3, again often with stronger responses to classical PDAC cells. Together, cancer cell co-culture therefore induced fast and profound adaptions in BMMs reminiscent of tumor-associated macrophage (TAM) features, which often show M2-like differentiation, contribute to immune cell recruitment and regulation, and remodeling of the tumor ECM54–56. Cell-selective proteomes and secretomes also allowed us to investigate potentially active intercellular signaling circuits that shaped macrophage polarization. We mapped PDAC-released proteins with corresponding macrophage receptors using ligand-receptor interactions curated in CellPhoneDB39. While PDAC cells did not release hallmark M2 polarizing factors IL-4 and IL-13, we found a complex mix of other proteins that have been associated with macrophage polarization (Fig. 4a): Some proteins were secreted without significant differences between subtypes, such as Tgfb1, a known M2 promoter57, or Tnf (at much lower abundance, see also Supplementary Data 1), an important M2-suppressing factor in cancer58. Other Tgfb- and Tnf-family members, macrophage survival essential colony-stimulating factors (Csfs), and many other signaling proteins, however, showed strong and consistent differential expression between PDAC subtypes. Classical PDAC cells secreted, for example, more Il6, a pleiotropic cytokine that has been described to enhance both M1- or M2-like states59, and Tnfsf15 (Supplementary Fig. 7), which has recently been shown to promote macrophage differentiation toward an M1 phenotype and increased cancer cell phagocytosis60. Mesenchymal PDAC cells secreted higher levels of Tgfb2 and Tgfb3 as well as Mif, Ccl5, and the alarmin Hmgb1, which have been found to skew macrophage polarization toward M1- and M2-like states in a context-dependent manner61–64. Furthermore, both PDAC subtypes expressed similar levels of Cd47, a contact-dependent anti-phagocytotic signal often upregulated by cancer cells to escape elimination by phagocytes65,66. Also, there is increasing evidence for the contribution of semaphorins to macrophage recruitment and differentiation, a family of exclusively secreted (class 3) or membrane-bound proteins (that can act as contact-dependent signals) with important roles in cancer67. For example, increased Sema3a has been associated with poor outcomes in PDAC68 and attraction of tumor-associated macrophages69, while Sema7a was shown to recruit and polarize macrophages toward the M2 state in the context of sepsis70.Fig. 4 Intercellular signals and signal receptors involved in BMM adaptions to PDAC co-culture. a Macrophage polarization-associated intercellular signaling protein expressed by mesenchymal or classical PDAC cells in co-culture with corresponding receptors detected in BMM proteomes. Ligands with significantly different secretion among PDAC subtypes (two-sided Student’s t-test, permutation-based FDR = 0.05, S0 = 0.1) and BMM receptors with significantly different levels of abundance between culture conditions (ANOVA, permutation-based FDR = 0.05, S0 = 1) are indicated in bold. b Heatmaps of primarily macrophage and neutrophil chemoattractants secreted by PDAC cells and BMMs. c Macrophage and neutrophil count as a percentage of CD45-positive cells in tumors from orthotopically transplanted classical and mesenchymal cancer cells analyzed by flow cytometry. Numbers of biological replicates and two-sided Welch’s t-test p-values between cell ratios in classical or mesenchymal tumors are indicated. Source data are provided as a Source Data file. The majority of detected PDAC signal corresponding receptors on BMMs showed stable expression, but some were regulated upon co-culture with cancer cells (Fig. 4a). Notably, BMMs upregulated Pvr (Poliovirus receptor) expression after interaction with both PDAC subtypes. Pvr activation on macrophages has been linked to an anti-inflammatory phenotype71, and targeting the Pvr-Tigit axis is being explored as a potential cancer immunotherapy strategy72. PDAC cancer cell subtype-specific chemokine secretion patterns correlate with immune cell recruitment in vivo Overall, both macrophages and PDAC cells responded to co-culture with increased production and release of immunomodulatory signaling proteins. Looking specifically at expression differences of immune cell recruiting factors, we noticed clear trends between PDAC subtypes: Mesenchymal PDAC cells secreted high levels of key monocyte recruitment and macrophage survival signals such as Ccl5 and Csf1 (Fig. 4b). In isolation, all four PDAC cell lines secreted many primarily neutrophil attracting proteins at similar levels. Interestingly, interaction with BMMs, however, strongly induced neutrophil recruiting chemokines like Cxcl2, Cxcl3, Cxcl5, and Cxcl15 in classical PDAC cells, whereas release remained unchanged (9091) or increased much less (8513) in mesenchymal cancer cells. BMM chemokine secretion patterns followed similar trends. Intrigued, we investigated the TME composition of tumors formed by the four PDAC subtype lines after orthotopic transplantation into mice. Immunophenotyping by flow cytometry revealed subtype-specific differences in immune cell populations. Among the analyzed cell types, differences between macrophage and neutrophil recruitment were the most significant and reflected the recruitment factor expression patterns from our secretomics experiments (Fig. 4c and Supplementary Fig. 8). PDAC tumors show systematic differences in matrisome protein production between mesenchymal and classical cancer subtypes To further investigate subtype differences between classical and mesenchymal PDAC, we transplanted all four MetRS*-expressing cancer lines orthotopically into syngeneic mice and compared cancer cell protein expression in the complex TME in vivo. In total, we identified 9415 specifically enriched cancer cell-derived proteins, which makes this one of the deepest cell type-specific PDAC in vivo proteomics datasets to date. Gene ontology enrichment analysis indicated pronounced differences in hallmark processes of epithelial–mesenchymal transition (EMT), such as cytoskeleton organization, ECM modulation, and cell–cell junctions (Fig. 5a). Moreover, mesenchymal PDAC cells showed an enriched interferon response signature and elevated antigen presentation-related protein expression, reminiscent of the adaptions that we observed in co-culture with macrophages in vitro and coinciding with the higher macrophage infiltration in these tumors (Fig. 3b and Fig. 4c).Fig. 5 PDAC subtype-specific expression of cancer cell-derived matrisome proteins in primary tumors. a Significantly enriched gene ontology terms (1D annotation enrichment, Benjamini–Hochberg FDR = 0.05) in classical and mesenchymal cancer cell proteomes after orthotopic transplantation into fully immunocompetent syngeneic mice (8442, 8513, 8661: n = 3, 9091: n = 2, biological replicates) (full list in Supplementary Data 4). b Left: Cancer cell-derived matrisome proteins annotated accoding to Naba et al.75 Right: Summed total and relative LFQ intensities per matrisome category. c Volcano plot of cancer cell-derived proteins in mesenchymal and classical PDAC tumors. Significantly regulated core matrisome proteins are highlighted in blue (two-sided Student’s t-test, permutation-based FDR = 0.05, S0 = 0.1). Boxplots show quantitative distributions of core matrisome, matrisome-associated and all identified non-matrisome proteins between the PDAC subtypes. P-values were determined by two-sided Welch’s t-test: ***p ≤  0.001 (1: p = 1.7 × 10−5, 2: p = 5.9 × 10−6). d Heatmap of matrisome proteins with significant expression differences between classical and mesenchymal PDAC cells in vivo (two-sided Student’s t-test, permutation-based FDR = 0.05, S0 = 0.1). e Gene ontology terms and UniProtKB keywords found overrepresented by Fisher’s exact test within each cluster are indicated (Benjamini–Hochberg FDR = 0.05, full list in Supplementary Data 4). Source data are provided as a Source Data file. Notably, many of the most prominent differences between mesenchymal and classical PDAC cell protein expression in vivo were ECM-related. Among diverse functions in cancer progression, dysregulated ECM in tumors strongly contributes to drug resistance, immune suppression, and metastasis51. Recent research has shown that, in particular, pancreatic cancer cell- rather than stromal cell-derived matrix proteins correlate with poor patient survival, although contributing only a minor fraction of the total ECM mass73,74. This introduced cell type-resolved profiling of ECM in tumors as a promising resource for therapeutic target and biomarker discovery. In contrast to the previous studies that characterized the cancer cell-derived matrix using xenotransplants73,74, Anl-labeling allows cell type-resolved analysis in syngeneic immunocompetent mice. Therefore, our PDAC model integrates interactions with infiltrating immune cells, which directly modulate the tumor ECM and change the ECM-associated protein expression of other cell types such as cancer cells50,51 (see also Supplementary Fig. 7). Motivated by this and the previously demonstrated advantages of Anl-enrichment for extracellular protein characterization (Fig. 2e), we further investigated ECM-related proteins in our data. We annotated proteins that constitute the ECM using an in silico defined matrisome atlas by Naba et al.75, which specifies “core matrisome” proteins such as collagens and proteoglycans, or proteins that are “matrisome-associated” such as ECM remodeling enzymes or secreted growth factors and cytokines that are known to bind to the ECM. Cancer cells expressed a diverse representation of each category, covering 405 matrisome proteins with only minor differences in overall identification numbers and very similar class distribution between classical and mesenchymal subtypes (Fig. 5b). Mesenchymal PDAC cancer cells have been shown to suppress cancer-associated fibroblasts (CAFs), the most prominent producers of ECM proteins in PDAC tumor stroma, leading to tumors with lower overall stromal and collagen content than classical PDAC76. However, quantitative analysis of cancer cell-derived proteins showed a higher abundance of mesenchymal-derived matrisome proteins and over-proportional expression of core matrisome and ECM regulators (Fig. 5b), indicating an increased relative contribution to the tumor ECM. Rather than being driven by a few highly abundant outliers, increased abundance of core matrisome expression in mesenchymal cells was a broad and statistically significant motif (Fig. 5c). At the individual protein level, more than a hundred matrisome protein groups had significant expression differences between the two PDAC subtypes (Fig. 5d). This included proteins recently identified as promising therapeutic targets, such as the predominantly cancer cell- rather than stromal cell-expressed PDAC metastasis promoters Agrn, Serpinb5, and Cstb74. All three proteins were detected in our experiment, and classical PDAC cancer cells produced significantly more Agrn and SerpinB5, suggesting potential subtype-specific responses to inhibition. Mesenchymal cancer cells, however, consistently produced higher amounts of EMT-promoting matrisome components, for example, fibrillar collagen types I and V, fibronectin, Fgf2, Tgfb family proteins such as Tgfb3 and Bmp2 as well as proteins involved in Tgfb signaling modulation, indicating a feed-forward loop with sustained local EMT signals (Fig. 5d, e). Moreover, we detected a much higher expression of lysyl oxidases Lox, Loxl1, and Loxl3 in mesenchymal cancer cells. Lysyl oxidase-family members mediate crosslinking of collagens and elastin and regulate cellular processes like adhesion, motility, and invasion77. They correlate with unfavorable patient prognosis in many cancers, including PDAC, and have been shown to promote chemoresistance, EMT, and metastasis78. In vivo, secretomics reflects tumor subtype and pathogenesis based on more than 1000 cancer cell-derived proteins in circulation In addition to local effects, tumor cell-derived proteins can act in distant tissues after entering circulation through the lymph or leaky blood vessels. Contrary to inference from cell expression data, profiling of such proteins in the bloodstream would give spatially-specific insights into many crucial aspects of tumor progression that involve long-distance signals and effectors and would also be invaluable for biomarker discovery. However, the lack of cell type-selectivity and the high dynamic range, with extremely abundant functional blood proteins and comparatively low abundance of tissue leakage proteins79, make this challenging with conventional methods. To evaluate whether Anl-labeling could be used to enrich cancer cell-derived proteins directly from body fluids, we collected mouse serum after orthotopic PDAC transplantation and Anl labeling of MetRS*-expressing cancer cells (Fig. 6a). Particularly, serum samples from mice bearing the 8661 (classical) and 8513 (mesenchymal) PDAC subtype tumors showed a good signal-to-noise ratio, with all replicates distinctly clustering from negative controls and each other in a principal component analysis (Supplementary Fig. 9). In these samples, 1614 proteins passed filtering criteria for specific enrichment, including 64 CellPhoneDB-annotated intercellular signaling protein ligands such as 23 cytokines (Fig. 6b), with around 42% identification overlap between the subtypes (Fig. 6c).Fig. 6 PDAC subtype-specific detection of cancer cell-derived proteins in circulation. a Scheme of cancer cell-derived protein capture from serum: After orthotopic transplantation of MetRS*-expressing or WT (Ctrl) PDAC cells and ANL labeling, tumor-bearing mouse sera were collected, and tumor-derived proteins were enriched and analyzed. b Specifically enriched cancer cell-derived serum protein groups ranked by intensity. Proteins with cytokine function are indicated. c Exclusively identified and overlap of specifically enriched 8513 and 8661 cancer cell-derived proteins (8661: n = 4, 8513: n = 3, biological replicates). d Cancer cell-derived matrisome proteins counts, summed total and relative LFQ intensities per matrisome category. e Fold change distribution of non-matrisome, and matrisome class proteins between PDAC subtypes. P-values were determined by a two-sided Welch’s t-test. f Volcano plot of 8661 and 8513 cancer cell-derived proteins in serum. GOCC annotated Laminin complex proteins (blue), fibrillar collagens (red), and proteins with significant fold changes (dark gray) are highlighted (two-sided Student’s t-test, permutation-based FDR = 0.05, S0 = 0.1). g Fold changes of pre-metastatic niche formation-associated proteins. Proteins with significant fold changes (two-sided Student’s t-test, permutation-based FDR = 0.05, S0 = 0.1) are indicated in bold. Source data are provided as a Source Data file. After discovering significant differences in matrisome protein expression between both PDAC subtypes in primary tumors, we were interested if these would be reflected in circulation. In total, we detected 199 core matrix or matrix-associated cancer cell-derived proteins in serum. While identified proteins had a very similar qualitative matrisome class distribution as primary tumors, quantitative distribution in serum was distinct, with the top two highest summed intensity classes shifting from secreted factors and ECM-affiliated proteins to ECM glycoproteins and ECM regulators (Figs. 6d and 5b). However, differences between subtypes followed the trends observed in primary tumors: Mesenchymal-derived matrisome proteins were more abundant and had an over-proportional share of core matrix proteins, specifically collagens and proteoglycans (Fig. 6d, e). At the individual protein level, we again found a higher abundance of laminin complex proteins in classical and fibrillar collagens in mesenchymal PDAC samples (Fig. 6f). Also, key subtype differences in immunomodulatory and matrix-modifying protein release identified in our previous experiments could be captured in serum, such as high Loxl1 and Csf1 secretion by mesenchymal cancer cells, providing direct evidence for potential long-range effects (Fig. 6g). Both proteins have indeed been described to pre-condition future places of metastases and, strikingly, we identified many other previously described pre-metastatic niche conditioning factors80–83 (Fig. 6g). A supportive premetastatic niche is vital for metastatic colonialization, which is considered a rate-limiting step of the invasion-metastatic cascade84,85. The premetastatic-niche-promoting signature of cancer cell-derived proteins in circulation likely contributes to the previous observation that increased KrasG12D gene dosage (Kras mut-iGD) promotes EMT and metastasis in the mesenchymal PDAC subtype45. Discussion Increasing insight into intercellular communication in the tumor microenvironment has helped to identify cancer vulnerabilities, for example, crucial immunosuppressive intercellular signaling circuits40,86. The combination of MS-based proteomics and cell-selective labeling is emerging as a powerful strategy to further accelerate the knowledge gained about heterocellular processes driving the disease. Direct analyses of labeled peptides offer straightforward solutions for multiplexed cell type-resolved proteomics and the evaluation of enrichment specificity87. However, with thorough background interference controls, analyses of all peptides from labeled proteins yield increased sensitivity and protein quantification accuracy. Nonetheless, previous studies were limited by low proteome coverage and/ or used extensive offline fractionation and less stringent filtering criteria to increase identifications19,23,28,88,89. The latter comes at the cost of sample throughput and specificity of enriched proteins, which are both key for the discovery of specific pathophysiological mechanisms. Here, we strongly increased the achievable proteomics depth with Anl labeling-based cell-selective proteomics by improving the biochemical enrichment of azide-modified proteins from complex biomolecule backgrounds. Beyond that, the increased recovery and enrichment specificity enabled additional applications for this concept, where low signal-to-noise was previously prohibitive. Combined with high-end mass spectrometry, data-independent acquisition, and recently developed software90, our workflow provides comprehensive and MS time-efficient cell-selective proteomes and secretomes in vitro and in vivo. In this study, we applied our improved workflows for the in-depth exploration of TME features between classical and mesenchymal PDAC subtypes in model systems with different degrees of complexity. In vitro, co-culture experiments offer a very controlled environment for focused and detailed mechanistic investigation of heterocellular interaction. Frequently used indirect co-culture systems such as supernatant transfer experiments or transwell assays facilitate cell-selective analysis by keeping cell types physically separated. However, they cannot cover all communication channels and reciprocal signaling dynamics91,92. In contrast, cell-selective labeling enables analysis of cells in direct co-culture, which fully integrates reciprocal communication means, including cell contact formation. Cell-selective labeling using amino acid precursors (CTAP)93, for example, granted invaluable insight into reciprocal signaling between pancreatic cancer cells and fibroblasts18. Specific enrichment of cell-selectively Anl-labeled proteins allowed us to extend this concept to in-depth profiling of heterocellular secretomes. Our findings not only recapitulated the pro-inflammatory secretory programs of macrophages as determined previously by distinct proteomics methods independent of MetRS*12,38 but identified, for example, 68 cancer cell-derived proteins with cytokine function in serum-containing culture media. Co-culture of PDAC cells and macrophages underlined the advantage of experiments that allow bidirectional intercellular communication, revealing broad reciprocal adaptions and strong regulation of intercellular signals upon co-culture, with an overall trend toward increased secretion of cytokines and chemokines in both cell types. While macrophages will be exposed to many more stimuli in tumors in vivo, including signals from other stromal cell types, both PDAC subtypes secreted a complex mix of macrophage polarization-associated proteins. Direct interaction with PDAC cells for less than two days was sufficient for macrophages to acquire many TAM-associated features. Moreover, in vitro secretomics allowed in-depth analysis of cancer cell-released chemokines and showed systematic differences between PDAC subtypes that reflected significant differences in TME cell composition, with higher macrophage infiltration in mesenchymal and higher neutrophil infiltration in classical PDAC tumors. This suggests that pancreatic cancer cell-derived signals directly contribute to the recruitment of these cell types. Specifically, large parts of the classical PDAC cell neutrophil recruiting signature became only apparent in co-culture, exemplifying how heterocellular systems expand the intercellular signaling capacity of a single cell type. For cell type-specific proteomics analysis in vivo, cells are typically extracted from tissue and sorted by FACS or MACS. We have shown that MetRS*-based cell-selective protein labeling and enrichment can have a fundamental cell-type-of-interest protein recovery advantage in pancreatic tumors compared to cell sorting. The high specific yields promise more effective analysis of less abundant or less robust cell types and even provide enough peptides to reach the higher input material demands of extended proteomics techniques such as offline fractionation for the construction of large peptide libraries or post-translational modification-specific enrichment. Importantly, Anl-enrichment also allows freezing of the cell states in tissues directly after harvesting, which provides an additional major benefit for PTM analysis since PTM-states such as protein phosphorylation are often highly dynamic and can be enzymatically modified within minutes in response to environmental changes94,95. The combination of Anl-enrichment and PTM analysis, therefore, outlines highly promising avenues for future research. Here, we focused on another advantage of the technique—the accessibility of extracellular proteins for click chemistry enrichment. Anl labeling facilitates the cell-selective analysis of secreted proteins in tissue or body fluids, which is of great interest and difficult to achieve with conventional techniques. For example, recent pioneering work has demonstrated the high value of cancer cell-selective matrisome analyses in primary tumors and metastases73,96 but relied on xenotransplants and immunocompromised mice to achieve cell-selectivity. In comparison, MetRS*-based cell-selective proteomics can provide additional value by overcoming the need for species-distinguishing peptides and avoiding potential dynamic range issues caused by the co-analysis of abundant host proteins, which can both reduce the cell-selective matrisome coverage. Moreover, in principle, any cell type can be studied without the need for immunosuppression. Our MS-based proteomics approach revealed pancreatic cancer cell and subtype-specific matrisome proteins in fully immunocompetent mice and provided proof-of-concept for in-depth analysis of cancer cell-derived proteins in tumor-bearing mouse serum. Previous research has shown a higher cellularity, less activated CAFs, and a less pronounced desmoplastic reaction in mesenchymal PDAC tumors97. Our cell-selective tumor analysis revealed that, among the lines we tested, mesenchymal cancer cells themselves produce significantly higher levels of matrisome proteins, particularly core matrix proteins, compared to classical PDAC cells. Furthermore, mesenchymal cancer cells exhibited a distinct matrisome signature that promotes epithelial-to-mesenchymal transition (EMT). Studies of human PDAC patient cohorts have demonstrated a high ECM content for the mesenchymal subtype, and mesenchymal/ECM-high PDAC correlated with a poor prognosis in comparison to classical/immune-rich PDAC98. Matrisome protein release could clearly segregate PDAC subtypes in our experiments, even when analyzing cancer cell-derived proteins in circulation. Remodeled ECM in cancer has been appreciated as critical for tumor progression51,99. Early trials directed towards broad depletion of aberrant stroma, however, revealed a dichotomous nature of the ECM and underlined the need for a more precise understanding of stromal components and their role in pathogenesis100–102. Using cell type-specific MetRS* mouse models, stromal cell types and their contribution to the tumor ECM can be studied systematically in the future. This combined knowledge will help to evaluate the sources and functions of individual tumor stromal components and identify tumor-promoting candidates for targeted inhibition without simultaneous interference with protective functions. Specifically, we detected elevated expression levels of Lox, Loxl1, and Loxl3 in mesenchymal cancer cells. Loxl2 has been identified as an independent prognostic factor in pancreatic cancer patients associated with poor survival103,104. However, anti-Loxl2 mAb treatment in a PDAC transplantation mouse model has caused a significant reduction in matrix content and accelerated tumor growth105. Other studies have shown a significant reduction of metastasis, chemosensitization, and prolonged survival after lysyl oxidase inhibition in PDAC78 or other cancers106,107. Future studies should evaluate how cell-type specific lysyl oxidases shape the composition of the tumor microenvironment and contribute to cancer progression. Cell type-selective profiling of tumor-derived proteins in body fluids opens exciting opportunities for a more precise understanding of long-range intercellular processes such as tumor cell recruitment and the metastatic cascade. In this study, we detected more than 1600 cancer cell-derived proteins in serum, including more than 20 cytokines, strongly improving the coverage achieved in previously published in vivo secretomics approaches, which range from a few dozen to a few hundred cells selectively identified proteins108–111. Our data revealed a strong release of pre-metastatic niche formation-associated factors by a mesenchymal compared to a classical PDAC line. Because the abundance of tumor-derived proteins in serum was very low even after enrichment, we expect current developments towards high sensitivity proteomics112,113 to be highly complementary for even more comprehensive cell-selective in vivo secretomics. Our study has identified differentiating features among PDAC subtypes with high consistency in our selected models. However, the limited sample size, with only two cell lines per subtype and one line per subtype in the serum secretomics experiment, and the small number of replicates in in vivo mouse experiments do pose limitations to our findings. Despite this, our results demonstrate the unique strengths of cell-selective proteomics analyses in uncovering disease mechanisms and provide a foundation for further research with larger sample sizes to statistically validate and expand upon these findings. Although our cell type-specific metabolic labeling approach cannot be directly applied to human cancer patients, it offers several possibilities for clinical translation. MetRS* transduced human premalignant cells (e.g., from pancreatic intraepithelial neoplasia (PanIN) or intraductal papillary mucinous neoplasm (IPMNs)), as well as PDAC cells and organoids (e.g., from invasive tumors representing various stages of PDAC progression, differences in metastatic capacity or molecular subtypes), can be transplanted into immunodeficient mice or mice with a humanized immune system114. Subsequent MetRS*-based proteomic profiling of tumors and body fluids, such as the blood, enables not only a deeper understanding of PDAC development, progression, and subtype specification but has also the potential for biomarker identification. So far, biomarkers for PanIN/IPMN and early PDAC detection, subtype classification, prognostic and therapeutic stratification, and the monitoring of targeted interventions are widely lacking115–117. Together, MetRS* based proteomic profiling holds the promise of biomarker discovery in tumors and circulation, which can be subsequently tested and validated in prospective studies in cancer patients. Methods LysMCre-MetRS* mice LysMCre/Cre118 and LSL-R26CAG-GFP,Mars*L274G/CAG-GFP,Mars*L274G (MetRS*)23 mice have been described previously. Strains were on a C57Bl/6 J background and interbred to obtain homozygous LysMCre-MetRS* mice for bone marrow isolation. Cell culture 293 T cells were obtained from ATCC (CRL-3216) and maintained at 37 °C, 5% CO2 in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% (v/v) heat-inactivated FCS (FCS HI) (complete DMEM). Primary mouse PDAC cells were obtained from autochthonous PDAC tumors as described before119 and maintained in complete DMEM for less than 30 passages. The preparation of bone marrow-derived macrophages (BMMs) followed the procedure described in Weischenfeldt and Porse (2008)120. In brief, bone marrow was harvested from femurs and tibiae of C57BL/6J wild-type (WT) or LysM-Cre-MetRS* mice. Bone marrow cells were passed through 70 µm nylon mesh filters and then plated on sterile, non-tissue culture-treated Petri dishes (5 × 106 cells/dish). After culture for 7 days in macrophage differentiation medium (DMEM supplemented with 10% (v/v) FCS HI and 20% (v/v) Csf1-containing L929 cell-conditioned media (replenished on day 3), BMMs were harvested in cold PBS and used for experiments. Cell lines were authenticated by genotyping and regularly tested for mycoplasma contamination by PCR. Transfection, lentivirus production, and transduction For Met-substitute incorporation comparison experiments, 293 T cells were transfected with an eGFP-MetRS* expression vector based on the pEGFP-C1 (Clontech) plasmid. Transient transfections were done with Lipofectamine 3000 (Invitrogen) according to the manufacturer’s instructions. For stable MetRS* expression, PDAC cells were lentivirally transduced using a modified Precision LentiORF Collection (pLOC) library (GE Healthcare) plasmid (pLOC-CMV > MetRS*:IRES:TurboGFP:P2A:BlastR; enrichment method comparison experiment), generated as described previously121, or a pLV-EF1A > MetRS*:P2A:EGFP:T2A:Puro plasmid, constructed by VectorBuilder. For virus production, 293 T cells were transfected with helper plasmids pMD2.G (Addgene), and psPAX (Addgene), and a MetRS* plasmid at a ratio of 1:1.5:2 (3 µg of plasmid DNA in total) in six-well plates using Lipofectamine 3000 and following the manufacturer’s instructions for lentiviral production, but using a total of 2 ml complete DMEM for virus collection 48 h post-transfection. After harvesting, polybrene was added to virus-containing supernatants at 10 µg/ ml final concentration. 1 ml of virus and polybrene-containing supernatant was added to 2.5 × 105 PDAC cells seeded in 6-well plates and mixed with 1 ml complete DMEM. After 16 h, media were exchanged with 2 ml complete DMEM. Forty-eight hours post-transduction, successfully transduced cells were selected with antibiotics (Blasticidin for pLOC-MetRS* and Puromycin for pLV-MetRS*) at a final concentration of 10 µg/ml. Pulse labeling with azidonorleucine or other Met-substitutes in vitro Cells were washed twice with PBS and then pre-incubated in methionine (Met)-free DMEM supplemented with 10% FCS HI for 30 min. Afterward, the medium was replaced by Met-free DMEM, 10% FCS HI containing 4 mM azidonorleucine (Iris Biotech) or other Met-substitutes (azidohomoalanine, L-methionine-methyl-13C,d3 (Met+4)). Cells were incubated for 8 h, washed twice with PBS, and harvested by scraping. For Anl enrichment-based secretome experiments, serum concentration during labeling was reduced to 5% to avoid protein precipitation in the supernatant concentration steps of the enrichment workflow. BMM—PDAC co-culture and LPS stimulation of BMMs 2 × 107 primary WT or MetRS* BMMs and 5 × 106 primary WT or MetRS* PDAC cells were seeded in mono- or co-culture on 15 cm dishes, incubated for 28 h in complete DMEM, and subsequently labeled for 8 h in 4 mM Anl (see above for details). BMMs seeded and cultured in parallel to co-culture experiment samples were treated with 500 ng/ ml LPS or vehicle simultaneously with the Anl-labeling. Cells and supernatants were harvested for cell-selective in vitro global proteomics and secretomics analyses (see below for details). Orthotopic transplantation and Anl administration in vivo In vivo transplantation experiments were performed as described in Nature Cancer volume 3, pages 318–336 (2022)40. In brief, 1 × 104 MetRS*/WT mouse PDAC cells were orthotopically transplanted into the pancreas of syngeneic immunocompetent C57Bl/6J mice. PDAC cell lines isolated from female endogenous mice were transplanted in female recipients and vice versa for male mice. Two to three weeks after transplantation, mice were treated with Anl (200 μl 300 mM Anl, twice a day for 5 days, intraperitoneal injections). Animals were sacrificed when individual mice reached the human endpoint or after Anl treatment. All mice experiments were performed in compliance with the European and the ARRIVE guidelines for the care and use of laboratory animals and were approved by the Institutional Animal Care and Use Committees (IACUC) of the local authorities of Technische Universität München and the Regierung von Oberbayern. A tumor diameter of 1.5 cm and a specific burden score, defined by a cumulative burden score, allowed by the IACUC and Regierung von Oberbayern were not surpassed in this study. All mice were kept in dedicated facilities, with a light–dark cycle of 12:12 h, housing temperature between 20 and 24 °C, and relative air humidity of 55%. Serum collection Blood from orthotopically transplanted mice was collected from the submandibular vein in serum collection tubes and further processed for downstream analysis (see below). Flow cytometry analysis and FACS Acquisition of eGFP-positive PDAC cell cultures by flow cytometry Cultured MetRS* and WT 8661 PDAC cell lines were detached using trypsin, then washed three times with ice-cold PBS, filtered through a 30 µm mesh, and resuspended in an adjusted volume of ice-cold PBS. Cell acquisition was performed using the BD FACS Aria Fusion. Flow cytometry data were analyzed using FlowJo software (v10.6.2). Acquisition and sorting of eGFP-positive cells from in vivo tumors by flow cytometry Dissociation of fresh tumor samples was performed as described previously40. Next, the debris removal solution (Miltenyi #130-109-398) was used to discard cell debris from the cell suspension, and the removal of dead cells was performed using the dead cell removal kit (Miltenyi #130-090-101). The enriched fraction of live cells were collected in ice-cold 2% FCS/PBS buffer and filtered through a 30 µm mesh before acquisition. Cell sorting was performed using the BD FACS Aria Fusion. EGFP-positive cells were sorted in low-bind tubes in PBS, washed two times with PBS, and the resulting cell pellet was shortly dried and snap-frozen. Flow cytometry data were analyzed using FlowJo software (v10.6.2). Immunophenotyping by flow cytometry Dissociation of fresh tumor samples and antibody staining was performed as described previously40. Cells were blocked with anti-mouse CD16/CD32 FC block (Biolegend, 1:100) for 10 min on ice and stained with Zombie Aqua Fixable Viability Kit (Biolegend, 1:500) to discriminate live and dead cells. The following antibody cocktails were used: CD4 BUV805 (BD, 1:100), CD3εBUV395 (BD, 1:20), CD8a BV785 (Biolegend, 1:100), CD25 BV650 (Biolegend, 1:50), TCRγ/δ BV421 (Biolegend, 1:100), CD62L PE (Biolegend, 1:500), CD44 APC-Fire (Biolegend, 1:30), CD45 PerCP Cy5.5 (Biolegend, 1:100), CD19 FITC (Biolegend, 1:100), EpCAM APC/AF647 (Biolegend, 1:200) for acquisition of adaptive immune cells; CD11c BUV737 (BD, 1:30), NK1.1 BUV395 (BD, 1:25), Ly6C BV785 (Biolegend, 1:200), CD11b BV650 (Biolegend, 1:100), F4/80 BV421/PB (Biolegend, 1:30), CD45 PerCP Cy5.5 (Biolegend, 1:100), Ly6G PE (Biolegend, 1:200), CD68 APC-CY7 (Biolegend, 1:20), EpCAM APC/AF647 (Biolegend, 1:200) for acquisition of innate immune cells. 1 × 106  events were acquired per antibody panel on the BD LSRFortessa. Flow cytometry data were analyzed using FlowJo software (v10.6.2). Enrichment of Anl-containing proteins DST-based enrichment DST-based enrichment was done as described in Methods in Molecular Biology volume 1266 pages 199–215 (2015)122 with slight modifications. In brief, samples were lysed in 1% (w/v) SDS, 2% Triton X-100, PBS pH 7.8 supplemented with EDTA-free protease inhibitors (PI) (Roche), diluted with PBS PI 1:1 for DNA digestion by benzonase (added 1:1000 (v/v)), heated for 10 min at 95 °C, diluted further with PBS PI to a final concentration of 0.1% SDS and 0.2% Triton X-100, and cleared by centrifugation for 5 min at 3000×g, 4 °C. Lysates were reduced and alkylated with immobilized TCEP and iodoacetamide (IAA), and subsequently desalted with PD-10 columns (GE Healthcare) to remove excess of reduction and alkylation agents. Click reactions were started by sequential addition of 200 µM tris((1-benzyl-4-triazolyl)methyl)amine, 25 µM disulfide biotin alkyne-tag (DST) (Click Chemistry Tools), and 100 µg/ml Cu(I)Br suspension and samples were incubated in an end-over-end mixer overnight at 4 °C. Following a second desalting step with PD-10 columns, elution in 10.5 ml 0.05% SDS, PBS pH 7.5, and the addition of 1% (v/v) NP40, tagged proteins were bound to 300 µl washed NeutrAvidin agarose (Thermo Scientific, 29202) in an end-over-end mixer overnight at 4 °C. Afterwards, the resins were sequentially washed with a total of 36 ml 0.2% SDS, 1% Triton X-100, PBS pH 7.4, then 18 ml PBS pH 7.4, and finally 18 ml 50 mM ammonium bicarbonate. Tagged proteins were eluted in a two-step procedure with a 5% (v/v) 2-mercaptoethanol/ammonium bicarbonate solution and subsequently lyophilized. After drying, proteins were resolubilized in 8 M Urea, 50 mM Tris-HCl pH 8 with 1 µg trypsin and lysC, predigested for 4 h at room temperature, and then diluted with 50 mM Tris-HCl pH 8 to a final concentration of 2 M urea for overnight digestion. Digests were desalted with C18 SepPak cartridges and in-house-made styroldivinylbenzol reversed phase sulfonate (SDB-RPS) (3 M Empore, 2241) StageTips. DBCO-agarose SPAAC enrichment DBCO-agarose enrichment was done as described in Mahdavi et al.19 with slight modifications. In brief, samples were lysed in 1% SDS, 100 mM chloroacetamide, PBS PI pH 7.4, heated at 95 °C for 10 min, sonicated to shear DNA, and centrifuged at 14.000 × g for 30 min. Cleared lysates were incubated for 3 h at room temperature with 100 µl washed dibenzocyclooctyne (DBCO)-agarose in an end-over-end mixer, and unreacted DBCO groups were subsequently quenched for 30 min by addition of 4 mM Anl. Afterward, bound proteins were reduced with 10 mM DTT for 15 min at 70 °C and alkylated with 40 mM IAA for 30 min at room temperature. The resins were sequentially washed with a total of 40 ml 0.8% SDS in PBS, 40 ml 8 M urea in 100 mM Tris-HCl (pH 8.0), and 40 ml 20% acetonitrile. Washed resins were resuspended in 100 µl 10% acetonitrile, 50 mM ammonium bicarbonate, and bound proteins were on-bead digested at 37 °C overnight with 1 µg of trypsin and lysC. Digests were collected, resins were washed with 500 µl 50 mM ammonium bicarbonate, washes were combined with digests and desalted with C18 SepPak cartridges. Alkyne-agarose CuAAC enrichment Samples were lysed in gua-lysis buffer (6 M guanidinium chloride, 4% (w/v) CHAPS, 0.5 M NaCl, 200 mM Hepes (pH 8) PI), heated at 95 °C for 5 min, sonicated to shear DNA and centrifuged at 10,000×g for 30 min. Cleared lysates were mixed with 100 µl (enrichment method comparison) or 50 µl (all other experiments) washed alkyne-agarose and diluted with ddH2O and a premixed catalyst solution to a final concentration of 1.5 M guanidinium chloride, 1 mM CuSO4, 6.25 mM BTTAA (Click Chemistry Tools), and 10 mM sodium ascorbate. Samples were incubated at room temperature overnight in an end-over-end mixer. Afterward, resins were washed twice with ddH2O and once with SDS wash buffer (1% (w/v) SDS, 250 mM NaCl, 5 mM EDTA, 100 mM Tris pH 8). After protein reduction with 10 mM DTT for 15 min at 70 °C and alkylation with 40 mM IAA for 30 min at room temperature in SDS wash buffer, resins were sequentially washed with a total of 20 ml SDS wash buffer, 20 ml 20% isopropanol, 20 ml 6 M guanidinium chloride, 100 mM Tris-HCl (pH 8), and 20 ml 20% acetonitrile. Washed resins were resuspended in 100 µl 10% acetonitrile, 2 mM CaCl2, 50 mM Tris-HCl pH 8, and bound proteins were on-bead digested at 37 °C overnight with 1 µg of trypsin and lysC. Digests were collected, resins were washed with 500 µl ddH2O, washes were combined with digests and desalted with C18 SepPak cartridges (enrichment method comparison and in vitro experiments) or Pierce Peptide Desalting Spin Columns (Thermo Scientific) (in vivo experiments). For Anl-enrichment-based in vitro or in vivo secretomes experiments, 15 ml cell-conditioned media or 400 µl tumor-bearing mouse serum were collected after Anl labeling (if yields from individual animals were lower, serum from multiple mice was pooled to reach the total volume). Conditioned media were centrifuged for 5 min at 1000×g to remove cell debris and supplemented with protease inhibitors. Conditioned media or mouse sera were washed twice with 15 ml 50 mM Tris-HCl pH 8 and concentrated to a volume of 250 µl using Ultracel-3 regenerated cellulose centrifugation filter units with a 3 kDa molecular weight cutoff (Millipore). Samples were mixed 1:1 with gua-lysis buffer, heated for 5 min at 95 °C, 1200 rpm, and then further processed using the alkyne-agarose CuAAC enrichment workflow (see above). Before lysis and CuAAC Anl-protein enrichment, tissue samples were homogenized to a fine powder with a mortar and pestle in liquid nitrogen. Sample preparation for mass spectrometry For proteomics analysis without Anl-enrichment, cells were lysed in SDC buffer (1% sodium deoxycholate (SDC), 10 mM tris(2-carboxy(ethyl)phosphine) (TCEP), 40 mM 2-chloroacetamide (CAA), 100 mM Tris-HCl pH 8.5) heated at 95 °C for 10 min and sonicated to shear DNA. Proteins were digested with trypsin and lysC (1:100 enzyme/protein ratio, w/w) at 37 °C, 1000 rpm overnight. Digests were desalted using in-house-made SDB-RPS StageTips. Desalted peptides from workflows with or without Anl-enrichment were dried in a vacuum concentrator and resolubilized in 0.1% formic acid. Concentrations were determined using a NanoDrop spectrophotometer and normalized between samples for equal peptide injection. Negative control (WT) samples for evaluating Anl-enrichment specificity were adjusted with corresponding volumes to their corresponding MetRS* samples for injections of equal total yield proportions. For offline high pH reversed-phase fractionation of peptide samples into 16 fractions (Fig. 1e), a spider fractionator was used as described previously123. LC–MS/MS Peptide mixtures were analyzed with an EASY-nLC 1000 or 1200 ultrahigh-pressure system (Thermo Fisher Scientific) coupled to a Q Exactive HF (293 T Met-substitution), Q Exactive HF-X (enrichment and acquisition method comparisons) or Orbitrap Exploris 480 (all other experiments) instrument (Thermo Fisher Scientific). Peptides (500 ng injections for Q Exactives or 300 ng for Exploris machines) were separated on 50 cm in-house-made 75 µm inner diameter columns, packed with 1.9-µm ReproSil C18 beads (Dr. Maisch GmbH) at a flow rate of 300 nl min−1 and 60 °C maintained by an in-house-made column oven. Offline pre-fractionated samples used for acquisition method comparison (see Fig. 1) were eluted with a binary buffer system (buffer A: 0.1% formic acid; buffer B: 80% acetonitrile, 0.1% formic acid) and a nonlinear gradient starting at 3% buffer B followed by a stepwise increase to 23% in 82 min, 40% in 8 min and a wash-out step for 10 min with an increase to 98% buffer B. Spectra were acquired with a data-dependent Top15 MS/MS method: Full scans (300–1650 m/z, automatic gain control (AGC) target = 3e6, maximum injection time = 25 ms, resolution = 60,000 at 200 m/z) were followed by up to 15 MS/MS scans with higher-energy collisional dissociation (HCD) (AGC target = 1e5, maximum injection time = 25 ms, isolation window = 1.5 m/z, normalized collision energy (nce) = 27%, resolution = 15,000 at 200 m/z). All other samples were analyzed without prefractionation in single shot measurements with a nonlinear gradient starting at 5% buffer B followed by a stepwise increase to 30% in 95 min, 60% in 5 min and a wash-out step for 20 min with an increase to 95% buffer B and subsequent decrease to 5% buffer B. Spectra were acquired with a data-dependent Top15 MS/MS method (as described above, but full scans with maximum injection time = 20 ms and MS/MS scans with maximum injection time = 28 ms, isolation window = 1.4 m/z) or data-independent acquisition (used for acquisition method comparison (Fig. 1e) and all following experiments) using full scans with a range of 300–1650 m/z (AGC target = 3e6, maximum injection time = 60 ms, resolution = 120,000 at 200 m/z) followed by MS/MS scans with 32 windows (nce = 27%, AGC target = 1e6, maximum injection time = 54 ms, resolution = 30,000 at 200 m/z). Data acquisition was controlled by Xcalibur (version 4.4.16.14, Thermo Fisher Scientific). LC–MS/MS data analysis DDA MS raw files were processed by MaxQuant124 (version 2.0.1.0.) using default parameters for orbitrap instruments with 1% FDR at the peptide and protein level, enabling MaxLFQ for label-free quantification. For analysis of Met-substitute incorporation in 293 T cells, Met-Anl, Met-Aha, and Met-Met+4 substitutions were added as variable modifications. DIA MS raw files were processed by DIA-NN90 (version 1.8) with FASTA digest for library-free search and deep learning-based spectra, RTs, and IMs prediction enabled. Precursor FDR was set to 1%, and default parameters were used with the following changes: The precursor range was restricted to 300–1650 m/z, and the fragment ion range to 200 – 1650 m/z. The “--relaxed-prot-inf” option was enabled via the command line. Mass accuracies and scan windows were optimized for individual experiments as recommended by the developers. MBR was enabled, neural network classifier was set to “double-pass mode,” and the quantification strategy to “robust LC (high accuracy).” Spectra were matched against the human (June 2022, 79,276 entries) or mouse (January 2022, 55,105 entries) UniProt FASTA database. Raw files were processed in two separate analyses for optimal independence of FACS- and Anl enrichment-based cancer cell-selective proteomics results (Fig. 2). MetRS*/ WT 8661 tumor rawfiles were reprocessed together with samples from other PDAC subtypes for in vivo PDAC cell subtype comparison (Fig. 5). 8661 PDAC (solo) secretome results (Fig. 3) were also used for secretome benchmarking (Fig. 1). Evaluation of Anl-enrichment specificity Except for enrichment method benchmarking (Fig. 1c, d), data from all Anl-enrichment-based experiments were filtered for specifically enriched proteins before further analysis. To evaluate enrichment specificity, samples were compared to corresponding negative control samples (WT equivalents of MetRS*-expressing cells that were treated equally and processed in parallel) and only proteins that were not identified in controls or had an at least 3fold higher median intensity than in controls were retained. For technical experiments (Figs. 1 and 2), PDAC MetRS* tumor comparison experiments (Fig. 5), and serum secretomics experiments with the PDAC lines 8661 and 8513 (Fig. 6), corresponding WT controls were used in triplicates for each PDAC line. We used aggregated control sample groups for multiple experimental groups in the co-culture experiments (Figs. 3 and 4): Three BMM WT samples were used to control BMM MetRS* samples cultured in isolation. A group of four co-cultured BMM WT + PDAC WT samples (one with each of the four PDAC lines) was used as controls for all BMM MetRS* + PDAC WT co-culture samples. Both solo and co-cultured PDAC MetRS* samples were controlled with the more conservative corresponding co-culture control samples (PDAC WT + BMM WT in triplicates for each of the PDAC lines). Statistical analysis Bioinformatic analyses were performed with Perseus125 (version 1.6.10.43) and R (version 4.1.2). Before statistical analysis, quantified proteins were filtered for at least two valid values in at least one group of replicates. The remaining missing values were imputed by random draw from a normal distribution with a width of 0.3 and a downshift of 1.8 relatives to the standard deviation of measured values. Statistical tests and parameters used to evaluate annotation enrichment and significant abundance differences of quantified proteins are specified in the figure legends. For box-and-whisker plots, standard boxplot features (lower quartile, median, upper quartile) were used as defined by ggplot2 version 3.4.0. Intercellular communication analysis Interactions between PDAC cells and macrophages in co-culture were inferred based on annotated ligand–receptor interactions from CellPhoneDB39 (v.2.0) extended by proteins with secretomes-derived experimental evidence126. BMM receptor expression levels were sourced from global proteomes, PDAC cell ligand expression levels from secretomes for secreted ligands, and global proteomes for membrane-bound ligands after filtering and imputation of missing values (see above). Reporting summary Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article. Supplementary information Supplementary Information Peer Review File Reporting Summary Description of Additional Supplementary Files Supplementary Dataset 1 Supplementary Dataset 2 Supplementary Dataset 3 Supplementary Dataset 4 Source data Source Data Supplementary information The online version contains supplementary material available at 10.1038/s41467-023-38171-8. Acknowledgements We thank I. Paron, C. Deiml, A. Strasser, J. B. Mueller-Reif, L. Schweizer, S. Steigerwald and P. Skowronek for MS assistance; V. Boeck, S. Ettl, M. Zukowska, A. Kudaliyanage, Y. P. Alvarez Valdivia, and A. Ullrich for technical assistance; T. Viturawong for intercellular communication analysis framework; M. Oroshi for computer and database support; and M. Mann and Marc Schmidt-Supprian for helpful discussions. This work was funded by the Max Planck Society for the Advancement of Science, the German Cancer Consortium (DKTK), the Wilhelm Sander-Stiftung (2020.174.1 and 2017.091.2) and the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation), Cluster of Excellence EXC 2151 ImmunoSensation2, Project-ID 360372040 (SFB 1335), Project-ID 408885537 (TRR 274), Project-ID 432325352 (SFB 1454), Project-ID 450149205 (TRR 333), Project-ID 329628492 (SFB 1321), Project-ID 515991405 (ME 4766/2-1 and SA 1374/8-1), Project-ID 515571394 (SA 1374/7-1), Project-ID 360394750 (SCHO 1732/2-1), Project ID 458890590 (SA 1374/6-1), Project ID 219542602 (SA 1374/4-3), Project-ID 97850925 (SFB854), and Project-ID 362321501 (GRK2413) and the Bundesministerium für Bildung und Forschung (BMBF) Kmu-innovativ 031B0686B. Author contributions J.J.S. performed experiments. J.J.S. developed and implemented bioinformatics methods. J.J.S. and F.M. conceived the data analysis and interpreted the data. SB and CF bred MetRS*-LysMCre mice, performed PDAC cell culture, characterization, and quality control, in vivo mouse experiments, flow cytometry analyses, and FACS. DS analyzed and interpreted data with S.B. and C.F. A.S. contributed to initial mouse experiments, testing of in vivo labeling schemes, and consultation on data analysis methods. D.F. contributed to co-culture experiments and figure design. J.C. prepared communication plot visualizations. S.E. helped with bone marrow-derived macrophage generation and experiments. P.L. and D.D. supplied reagents, initial training, and consultation on Anl-enrichment procedures. H.D. provided advice and supervision of initial experiments. F.M. conceived the study. F.M. and D.S. provided funding and supervised the experiments. J.S.S. and F.M. wrote the paper. Peer review Peer review information Nature Communications thanks Arnaud Droit and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available. Funding Open Access funding enabled and organized by Projekt DEAL. Data availability The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository127 with the dataset identifier PXD040084, which is publicly available. Source data are provided in this paper. Competing interests The authors declare no competing interests. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. These authors contributed equally: Jonathan J. Swietlik, Stefanie Bärthel, Chiara Falcomatà. ==== Refs References 1. Siegel RL Miller KD Fuchs HE Jemal A Cancer statistics, 2022 CA Cancer J. Clin. 2022 72 7 33 10.3322/caac.21708 35020204 2. Neesse A Algul H Tuveson DA Gress TM Stromal biology and therapy in pancreatic cancer: a changing paradigm Gut 2015 64 1476 1484 10.1136/gutjnl-2015-309304 25994217 3. Hosein AN Brekken RA Maitra A Pancreatic cancer stroma: an update on therapeutic targeting strategies Nat. Rev. Gastroenterol. Hepatol. 2020 17 487 505 10.1038/s41575-020-0300-1 32393771 4. Collisson EA Bailey P Chang DK Biankin AV Molecular subtypes of pancreatic cancer Nat. Rev. Gastroenterol. Hepatol. 2019 16 207 220 10.1038/s41575-019-0109-y 30718832 5. Collisson EA Subtypes of pancreatic ductal adenocarcinoma and their differing responses to therapy Nat. Med. 2011 17 500 503 10.1038/nm.2344 21460848 6. Moffitt RA Virtual microdissection identifies distinct tumor- and stroma-specific subtypes of pancreatic ductal adenocarcinoma Nat. Genet. 2015 47 1168 1178 10.1038/ng.3398 26343385 7. Bailey P Genomic analyses identify molecular subtypes of pancreatic cancer Nature 2016 531 47 52 10.1038/nature16965 26909576 8. S NK Morphological classification of pancreatic ductal adenocarcinoma that predicts molecular subtypes and correlates with clinical outcome Gut 2020 69 317 328 10.1136/gutjnl-2019-318217 31201285 9. Aung KL Genomics-driven precision medicine for advanced pancreatic cancer: early results from the COMPASS trial Clin. Cancer Res. 2018 24 1344 1354 10.1158/1078-0432.CCR-17-2994 29288237 10. Schwanhausser B Global quantification of mammalian gene expression control Nature 2011 473 337 342 10.1038/nature10098 21593866 11. Jovanovic M Immunogenetics. Dynamic profiling of the protein life cycle in response to pathogens Science 2015 347 1259038 10.1126/science.1259038 25745177 12. Meissner F Scheltema RA Mollenkopf HJ Mann M Direct proteomic quantification of the secretome of activated immune cells Science 2013 340 475 478 10.1126/science.1232578 23620052 13. Rieckmann JC Social network architecture of human immune cells unveiled by quantitative proteomics Nat. Immunol. 2017 18 583 593 10.1038/ni.3693 28263321 14. Swietlik JJ Sinha A Meissner F Dissecting intercellular signaling with mass spectrometry-based proteomics Curr. Opin. Cell Biol. 2020 63 20 30 10.1016/j.ceb.2019.12.002 31927463 15. Mertins P Proteogenomics connects somatic mutations to signalling in breast cancer Nature 2016 534 55 62 10.1038/nature18003 27251275 16. Jayavelu AK Splicing factor YBX1 mediates persistence of JAK2-mutated neoplasms Nature 2020 588 157 163 10.1038/s41586-020-2968-3 33239784 17. Mund, A. et al. Deep Visual Proteomics defines single-cell identity and heterogeneity. Nat. Biotechnol. 10.1038/s41587-022-01302-5 (2022). 18. Tape CJ Oncogenic KRAS regulates tumor cell signaling via stromal reciprocation Cell 2016 165 910 920 10.1016/j.cell.2016.03.029 27087446 19. Mahdavi A Engineered aminoacyl-tRNA synthetase for cell-selective analysis of mammalian protein synthesis J. Am. Chem. Soc. 2016 138 4278 4281 10.1021/jacs.5b08980 26991063 20. Ngo JT Cell-selective metabolic labeling of proteins Nat. Chem. Biol. 2009 5 715 717 10.1038/nchembio.200 19668194 21. Tanrikulu IC Schmitt E Mechulam Y Goddard WA 3rd Tirrell DA Discovery of Escherichia coli methionyl-tRNA synthetase mutants for efficient labeling of proteins with azidonorleucine in vivo Proc. Natl Acad. Sci. USA 2009 106 15285 15290 10.1073/pnas.0905735106 19706454 22. Liu Y Application of bio-orthogonal proteome labeling to cell transplantation and heterochronic parabiosis Nat. Commun. 2017 8 643 10.1038/s41467-017-00698-y 28935952 23. Alvarez-Castelao B Cell-type-specific metabolic labeling of nascent proteomes in vivo Nat. Biotechnol. 2017 35 1196 1201 10.1038/nbt.4016 29106408 24. Hines WC Su Y Kuhn I Polyak K Bissell MJ Sorting out the FACS: a devil in the details Cell Rep. 2014 6 779 781 10.1016/j.celrep.2014.02.021 24630040 25. Elkord E Williams PE Kynaston H Rowbottom AW Human monocyte isolation methods influence cytokine production from in vitro generated dendritic cells Immunology 2005 114 204 212 10.1111/j.1365-2567.2004.02076.x 15667565 26. Andra I An evaluation of T-cell functionality after flow cytometry sorting revealed p38 MAPK activation Cytometry A 2020 97 171 183 10.1002/cyto.a.23964 31944553 27. Llufrio EM Wang L Naser FJ Patti GJ Sorting cells alters their redox state and cellular metabolome Redox Biol. 2018 16 381 387 10.1016/j.redox.2018.03.004 29627745 28. Azizian NG Selective labeling and identification of the tumor cell proteome of pancreatic cancer in vivo J. Proteome Res. 2021 20 858 866 10.1021/acs.jproteome.0c00666 33289385 29. Bekker-Jensen DB A compact quadrupole-orbitrap mass spectrometer with FAIMS interface improves proteome coverage in short LC gradients Mol. Cell Proteom. 2020 19 716 729 10.1074/mcp.TIR119.001906 30. Erdmann I Cell-selective labelling of proteomes in Drosophila melanogaster Nat. Commun. 2015 6 7521 10.1038/ncomms8521 26138272 31. Muller A Stellmacher A Freitag CE Landgraf P Dieterich DC Monitoring astrocytic proteome dynamics by cell type-specific protein labeling PLoS ONE 2015 10 e0145451 10.1371/journal.pone.0145451 26690742 32. McKay CS Finn MG Click chemistry in complex mixtures: bioorthogonal bioconjugation Chem. Biol. 2014 21 1075 1101 10.1016/j.chembiol.2014.09.002 25237856 33. Presolski SI Hong VP Finn MG Copper-catalyzed azide-alkyne click chemistry for bioconjugation Curr. Protoc. Chem. Biol. 2011 3 153 162 10.1002/9780470559277.ch110148 22844652 34. Yang Y Yang X Verhelst SH Comparative analysis of click chemistry mediated activity-based protein profiling in cell lysates Molecules 2013 18 12599 12608 10.3390/molecules181012599 24126377 35. Besanceney-Webler C Increasing the efficacy of bioorthogonal click reactions for bioconjugation: a comparative study Angew. Chem. Int. Ed. Engl. 2011 50 8051 8056 10.1002/anie.201101817 21761519 36. van Geel R Pruijn GJ van Delft FL Boelens WC Preventing thiol-yne addition improves the specificity of strain-promoted azide-alkyne cycloaddition Bioconjug. Chem. 2012 23 392 398 10.1021/bc200365k 22372991 37. Alvarez-Castelao B Schanzenbacher CT Langer JD Schuman EM Cell-type-specific metabolic labeling, detection and identification of nascent proteomes in vivo Nat. Protoc. 2019 14 556 575 10.1038/s41596-018-0106-6 30610240 38. Eichelbaum K Winter M Berriel Diaz M Herzig S Krijgsveld J Selective enrichment of newly synthesized proteins for quantitative secretome analysis Nat. Biotechnol. 2012 30 984 990 10.1038/nbt.2356 23000932 39. Efremova M Vento-Tormo M Teichmann SA Vento-Tormo R CellPhoneDB: inferring cell-cell communication from combined expression of multi-subunit ligand-receptor complexes Nat. Protoc. 2020 15 1484 1506 10.1038/s41596-020-0292-x 32103204 40. Falcomata C Selective multi-kinase inhibition sensitizes mesenchymal pancreatic cancer to immune checkpoint blockade by remodeling the tumor microenvironment Nat. Cancer 2022 3 318 336 10.1038/s43018-021-00326-1 35122074 41. Varnavides G In search of a universal method: a comparative survey of bottom-up proteomics sample preparation methods J. Proteome Res. 2022 21 2397 2411 10.1021/acs.jproteome.2c00265 36006919 42. Alfonso-Garrido J Garcia-Calvo E Luque-Garcia JL Sample preparation strategies for improving the identification of membrane proteins by mass spectrometry Anal. Bioanal. Chem. 2015 407 4893 4905 10.1007/s00216-015-8732-0 25967148 43. Clark CE Dynamics of the immune reaction to pancreatic cancer from inception to invasion Cancer Res. 2007 67 9518 9527 10.1158/0008-5472.CAN-07-0175 17909062 44. Ruffell B Coussens LM Macrophages and therapeutic resistance in cancer Cancer Cell 2015 27 462 472 10.1016/j.ccell.2015.02.015 25858805 45. Mueller S Evolutionary routes and KRAS dosage define pancreatic cancer phenotypes Nature 2018 554 62 68 10.1038/nature25459 29364867 46. Ashburner M Gene ontology: tool for the unification of biology. The Gene Ontology Consortium Nat. Genet. 2000 25 25 29 10.1038/75556 10802651 47. Hollmen M Roudnicky F Karaman S Detmar M Characterization of macrophage–cancer cell crosstalk in estrogen receptor positive and triple-negative breast cancer Sci. Rep. 2015 5 9188 10.1038/srep09188 25776849 48. Castro F Cardoso AP Goncalves RM Serre K Oliveira MJ Interferon-gamma at the crossroads of tumor immune surveillance or evasion Front. Immunol. 2018 9 847 10.3389/fimmu.2018.00847 29780381 49. Setrerrahmane S Xu H Tumor-related interleukins: old validated targets for new anti-cancer drug development Mol. Cancer 2017 16 153 10.1186/s12943-017-0721-9 28927416 50. Winkler J Abisoye-Ogunniyan A Metcalf KJ Werb Z Concepts of extracellular matrix remodelling in tumour progression and metastasis Nat. Commun. 2020 11 5120 10.1038/s41467-020-18794-x 33037194 51. Cox TR The matrix in cancer Nat. Rev. Cancer 2021 21 217 238 10.1038/s41568-020-00329-7 33589810 52. Murray PJ Macrophage polarization Annu Rev. Physiol. 2017 79 541 566 10.1146/annurev-physiol-022516-034339 27813830 53. Mantovani A Sica A Locati M Macrophage polarization comes of age Immunity 2005 23 344 346 10.1016/j.immuni.2005.10.001 16226499 54. Candido JB CSF1R(+) macrophages sustain pancreatic tumor growth through T cell suppression and maintenance of key gene programs that define the squamous subtype Cell Rep. 2018 23 1448 1460 10.1016/j.celrep.2018.03.131 29719257 55. Kemp, S. B. et al. Pancreatic cancer is marked by complement-high blood monocytes and tumor-associated macrophages. Life Sci. Alliance10.26508/lsa.202000935 (2021). 56. Zhu Y Tissue-resident macrophages in pancreatic ductal adenocarcinoma originate from embryonic hematopoiesis and promote tumor progression Immunity 2017 47 323 338.e326 10.1016/j.immuni.2017.07.014 28813661 57. Zhang F TGF-beta induces M2-like macrophage polarization via SNAIL-mediated suppression of a pro-inflammatory phenotype Oncotarget 2016 7 52294 52306 10.18632/oncotarget.10561 27418133 58. Kratochvill F TNF counterbalances the emergence of M2 tumor macrophages Cell Rep. 2015 12 1902 1914 10.1016/j.celrep.2015.08.033 26365184 59. Fernando MR Reyes JL Iannuzzi J Leung G McKay DM The pro-inflammatory cytokine, interleukin-6, enhances the polarization of alternatively activated macrophages PLoS ONE 2014 9 e94188 10.1371/journal.pone.0094188 24736635 60. Zhao CC TNFSF15 facilitates differentiation and polarization of macrophages toward M1 phenotype to inhibit tumor growth Oncoimmunology 2022 11 2032918 10.1080/2162402X.2022.2032918 35127254 61. Li M CCL5 deficiency promotes liver repair by improving inflammation resolution and liver regeneration through M2 macrophage polarization Cell Mol. Immunol. 2020 17 753 764 10.1038/s41423-019-0279-0 31481754 62. Castro BA Macrophage migration inhibitory factor downregulation: a novel mechanism of resistance to anti-angiogenic therapy Oncogene 2017 36 3749 3759 10.1038/onc.2017.1 28218903 63. Yaddanapudi K Control of tumor-associated macrophage alternative activation by macrophage migration inhibitory factor J. Immunol. 2013 190 2984 2993 10.4049/jimmunol.1201650 23390297 64. Schaper F High mobility group box 1 skews macrophage polarization and negatively influences phagocytosis of apoptotic cells Rheumatology 2016 55 2260 2270 10.1093/rheumatology/kew324 27632996 65. Willingham SB The CD47-signal regulatory protein alpha (SIRPa) interaction is a therapeutic target for human solid tumors Proc. Natl Acad. Sci. USA 2012 109 6662 6667 10.1073/pnas.1121623109 22451913 66. Zhang M Anti-CD47 treatment stimulates phagocytosis of glioblastoma by M1 and M2 polarized macrophages and promotes M1 polarized macrophages in vivo PLoS ONE 2016 11 e0153550 10.1371/journal.pone.0153550 27092773 67. Tamagnone L Emerging role of semaphorins as major regulatory signals and potential therapeutic targets in cancer Cancer Cell 2012 22 145 152 10.1016/j.ccr.2012.06.031 22897846 68. Muller MW Association of axon guidance factor semaphorin 3A with poor outcome in pancreatic cancer Int. J. Cancer 2007 121 2421 2433 10.1002/ijc.22949 17631638 69. Casazza A Impeding macrophage entry into hypoxic tumor areas by Sema3A/Nrp1 signaling blockade inhibits angiogenesis and restores antitumor immunity Cancer Cell 2013 24 695 709 10.1016/j.ccr.2013.11.007 24332039 70. Korner, A. et al. Sema7A is crucial for resolution of severe inflammation. Proc. Natl Acad. Sci. USA10.1073/pnas.2017527118 (2021). 71. Chen X TIGIT negatively regulates inflammation by altering macrophage phenotype Immunobiology 2016 221 48 55 10.1016/j.imbio.2015.08.003 26307002 72. Gorvel, L. & Olive, D. Targeting the “PVR-TIGIT axis” with immune checkpoint therapies. F1000Res10.12688/f1000research.22877.1 (2020). 73. Tian C Proteomic analyses of ECM during pancreatic ductal adenocarcinoma progression reveal different contributions by tumor and stromal cells Proc. Natl Acad. Sci. USA 2019 116 19609 19618 10.1073/pnas.1908626116 31484774 74. Tian C Cancer cell-derived matrisome proteins promote metastasis in pancreatic ductal adenocarcinoma Cancer Res. 2020 80 1461 1474 10.1158/0008-5472.CAN-19-2578 32029550 75. Naba A The matrisome: in silico definition and in vivo characterization by proteomics of normal and tumor extracellular matrices Mol. Cell Proteom. 2012 11 M111 014647 10.1074/mcp.M111.014647 76. Puleo F Stratification of pancreatic ductal adenocarcinomas based on tumor and microenvironment features Gastroenterology 2018 155 1999 2013.e1993 10.1053/j.gastro.2018.08.033 30165049 77. Kagan HM Li W Lysyl oxidase: properties, specificity, and biological roles inside and outside of the cell J. Cell Biochem. 2003 88 660 672 10.1002/jcb.10413 12577300 78. Miller BW Targeting the LOX/hypoxia axis reverses many of the features that make pancreatic cancer deadly: inhibition of LOX abrogates metastasis and enhances drug efficacy EMBO Mol. Med. 2015 7 1063 1076 10.15252/emmm.201404827 26077591 79. Geyer PE Holdt LM Teupser D Mann M Revisiting biomarker discovery by plasma proteomics Mol. Syst. Biol. 2017 13 942 10.15252/msb.20156297 28951502 80. Liu Y Cao X Characteristics and significance of the pre-metastatic niche Cancer Cell 2016 30 668 681 10.1016/j.ccell.2016.09.011 27846389 81. Peinado H Pre-metastatic niches: organ-specific homes for metastases Nat. Rev. Cancer 2017 17 302 317 10.1038/nrc.2017.6 28303905 82. Hoye AM Erler JT Structural ECM components in the premetastatic and metastatic niche Am. J. Physiol. Cell Physiol. 2016 310 C955 C967 10.1152/ajpcell.00326.2015 27053524 83. Gillot L Baudin L Rouaud L Kridelka F Noel A The pre-metastatic niche in lymph nodes: formation and characteristics Cell Mol. Life Sci. 2021 78 5987 6002 10.1007/s00018-021-03873-z 34241649 84. Valastyan S Weinberg RA Tumor metastasis: molecular insights and evolving paradigms Cell 2011 147 275 292 10.1016/j.cell.2011.09.024 22000009 85. Vanharanta S Massague J Origins of metastatic traits Cancer Cell 2013 24 410 421 10.1016/j.ccr.2013.09.007 24135279 86. Ribas A Wolchok JD Cancer immunotherapy using checkpoint blockade Science 2018 359 1350 1355 10.1126/science.aar4060 29567705 87. Schiapparelli LM Activity-induced cortical glutamatergic neuron nascent proteins J. Neurosci. 2022 42 7900 7920 10.1523/JNEUROSCI.0707-22.2022 36261270 88. Rayaprolu S Cell type-specific biotin labeling in vivo resolves regional neuronal and astrocyte proteomic differences in mouse brain Nat. Commun. 2022 13 2927 10.1038/s41467-022-30623-x 35614064 89. Sun X Deep single-cell-type proteome profiling of mouse brain by nonsurgical AAV-mediated proximity labeling Anal. Chem. 2022 94 5325 5334 10.1021/acs.analchem.1c05212 35315655 90. Demichev V Messner CB Vernardis SI Lilley KS Ralser M DIA-NN: neural networks and interference correction enable deep proteome coverage in high throughput Nat. Methods 2020 17 41 44 10.1038/s41592-019-0638-x 31768060 91. Bogdanowicz DR Lu HH Studying cell-cell communication in co-culture Biotechnol. J. 2013 8 395 396 10.1002/biot.201300054 23554248 92. Vis MAM Ito K Hofmann S Impact of culture medium on cellular interactions in in vitro co-culture systems Front. Bioeng. Biotechnol. 2020 8 911 10.3389/fbioe.2020.00911 32850750 93. Gauthier NP Cell-selective labeling using amino acid precursors for proteomic studies of multicellular environments Nat. Methods 2013 10 768 773 10.1038/nmeth.2529 23817070 94. Mertins P Ischemia in tumors induces early and sustained phosphorylation changes in stress kinase pathways but does not affect global protein levels Mol. Cell Proteom. 2014 13 1690 1704 10.1074/mcp.M113.036392 95. Tanzer MC Bludau I Stafford CA Hornung V Mann M Phosphoproteome profiling uncovers a key role for CDKs in TNF signaling Nat. Commun. 2021 12 6053 10.1038/s41467-021-26289-6 34663829 96. Di Martino JS A tumor-derived type III collagen-rich ECM niche regulates tumor cell dormancy Nat. Cancer 2022 3 90 107 10.1038/s43018-021-00291-9 35121989 97. Steins A High-grade mesenchymal pancreatic ductal adenocarcinoma drives stromal deactivation through CSF-1 EMBO Rep. 2020 21 e48780 10.15252/embr.201948780 32173982 98. Maurer C Experimental microdissection enables functional harmonisation of pancreatic cancer subtypes Gut 2019 68 1034 1043 10.1136/gutjnl-2018-317706 30658994 99. Pickup MW Mouw JK Weaver VM The extracellular matrix modulates the hallmarks of cancer EMBO Rep. 2014 15 1243 1253 10.15252/embr.201439246 25381661 100. Cox TR Erler JT Fibrosis and cancer: partners in crime or opposing forces Trends Cancer 2016 2 279 282 10.1016/j.trecan.2016.05.004 28741525 101. Ozdemir BC Depletion of carcinoma-associated fibroblasts and fibrosis induces immunosuppression and accelerates pancreas cancer with reduced survival Cancer Cell 2014 25 719 734 10.1016/j.ccr.2014.04.005 24856586 102. Rhim AD Stromal elements act to restrain, rather than support, pancreatic ductal adenocarcinoma Cancer Cell 2014 25 735 747 10.1016/j.ccr.2014.04.021 24856585 103. Tanaka N Clinical implications of lysyl oxidase-like protein 2 expression in pancreatic cancer Sci. Rep. 2018 8 9846 10.1038/s41598-018-28253-9 29959362 104. Alonso-Nocelo, M. et al. Macrophages direct cancer cells through a LOXL2-mediated metastatic cascade in pancreatic ductal adenocarcinoma. Gut10.1136/gutjnl-2021-325564 (2022). 105. Jiang HL Pancreatic ductal adenocarcinoma progression is restrained by stromal matrix J. Clin. Invest. 2020 130 4704 4709 10.1172/JCI136760 32749238 106. Saatci O Targeting lysyl oxidase (LOX) overcomes chemotherapy resistance in triple negative breast cancer Nat. Commun. 2020 11 2416 10.1038/s41467-020-16199-4 32415208 107. Rachman-Tzemah C Blocking surgically induced lysyl oxidase activity reduces the risk of lung metastases Cell Rep. 2017 19 774 784 10.1016/j.celrep.2017.04.005 28445728 108. Kim KE Dynamic tracking and identification of tissue-specific secretory proteins in the circulation of live mice Nat. Commun. 2021 12 5204 10.1038/s41467-021-25546-y 34471136 109. Droujinine IA Proteomics of protein trafficking by in vivo tissue-specific labeling Nat. Commun. 2021 12 2382 10.1038/s41467-021-22599-x 33888706 110. Wei W Cell type-selective secretome profiling in vivo Nat. Chem. Biol. 2021 17 326 334 10.1038/s41589-020-00698-y 33199915 111. Liu, J., Jang, J. Y., Pirooznia, M., Liu, S. & Finkel, T. The secretome mouse provides a genetic platform to delineate tissue-specific in vivo secretion. Proc Natl Acad Sci USA10.1073/pnas.2005134118 (2021). 112. Brunner AD Ultra-high sensitivity mass spectrometry quantifies single-cell proteome changes upon perturbation Mol. Syst. Biol. 2022 18 e10798 10.15252/msb.202110798 35226415 113. Demichev V dia-PASEF data analysis using FragPipe and DIA-NN for deep proteomics of low sample amounts Nat. Commun. 2022 13 3944 10.1038/s41467-022-31492-0 35803928 114. Allen TM Humanized immune system mouse models: progress, challenges and opportunities Nat. Immunol. 2019 20 770 774 10.1038/s41590-019-0416-z 31160798 115. Guler GD Detection of early stage pancreatic cancer using 5-hydroxymethylcytosine signatures in circulating cell free DNA Nat. Commun. 2020 11 5270 10.1038/s41467-020-18965-w 33077732 116. Hosein AN Dougan SK Aguirre AJ Maitra A Translational advances in pancreatic ductal adenocarcinoma therapy Nat. Cancer 2022 3 272 286 10.1038/s43018-022-00349-2 35352061 117. Hyeon, D. Y. et al. Proteogenomic landscape of human pancreatic ductal adenocarcinoma in an Asian population reveals tumor cell-enriched and immune-rich subtypes. Nat. Cancer10.1038/s43018-022-00479-7 (2022). 118. Clausen BE Burkhardt C Reith W Renkawitz R Forster I Conditional gene targeting in macrophages and granulocytes using LysMcre mice Transgenic Res. 1999 8 265 277 10.1023/A:1008942828960 10621974 119. von Burstin J E-cadherin regulates metastasis of pancreatic cancer in vivo and is suppressed by a SNAIL/HDAC1/HDAC2 repressor complex Gastroenterology 2009 137 371.e361 365 120. Weischenfeldt J Porse B Bone marrow-derived macrophages (BMM): isolation and applications CSH Protoc. 2008 2008 pdb prot5080 21356739 121. Frauenstein A Identification of covalent modifications regulating immune signaling complex composition and phenotype Mol. Syst. Biol. 2021 17 e10125 10.15252/msb.202010125 34318608 122. Landgraf P Antileo ER Schuman EM Dieterich DC BONCAT: metabolic labeling, click chemistry, and affinity purification of newly synthesized proteomes Methods Mol. Biol. 2015 1266 199 215 10.1007/978-1-4939-2272-7_14 25560077 123. Kulak NA Geyer PE Mann M Loss-less nano-fractionator for high sensitivity, high coverage proteomics Mol. Cell Proteom. 2017 16 694 705 10.1074/mcp.O116.065136 124. Cox J Mann M MaxQuant enables high peptide identification rates, individualized p.p.b.-range mass accuracies and proteome-wide protein quantification Nat. Biotechnol. 2008 26 1367 1372 10.1038/nbt.1511 19029910 125. Tyanova S The Perseus computational platform for comprehensive analysis of (prote)omics data Nat. Methods 2016 13 731 740 10.1038/nmeth.3901 27348712 126. Phulphagar K Proteomics reveals distinct mechanisms regulating the release of cytokines and alarmins during pyroptosis Cell Rep. 2021 34 108826 10.1016/j.celrep.2021.108826 33691121 127. Perez-Riverol Y The PRIDE database and related tools and resources in 2019: improving support for quantification data Nucleic Acids Res. 2019 47 D442 D450 10.1093/nar/gky1106 30395289 128. Cox J Mann M 1D and 2D annotation enrichment: a statistical method integrating quantitative proteomics with complementary high-throughput data BMC Bioinforma. 2012 13 S12 10.1186/1471-2105-13-S16-S12
PMC010xxxxxx/PMC10170203.txt
==== Front Open Biol Open Biol RSOB royopenbio Open Biology 2046-2441 The Royal Society 37161289 10.1098/rsob.220121 rsob220121 133Research Research Articles Delay eyeblink conditioning performance and brain-wide c-Fos expression in male and female mice Delay eyeblink conditioning performance and brain-wide c-fos expression in male and female mice http://orcid.org/0000-0003-3079-4713 Oyaga Maria Roa 1 † https://orcid.org/0000-0002-2257-3040 Serra Ines Conceptualization Investigation Methodology Project administration Supervision Visualization Writing – review & editing 1 † http://orcid.org/0000-0001-6782-239X Kurup Devika Investigation Methodology Project administration 1 https://orcid.org/0000-0002-8790-5501 Koekkoek Sebastiaan K. E. Conceptualization Methodology Project administration Resources Software Writing – review & editing 1 http://orcid.org/0000-0002-0119-5108 Badura Aleksandra Conceptualization Funding acquisition Investigation Methodology Project administration Resources Software Supervision Validation Visualization Writing – original draft Writing – review & editing [email protected] 1 2 1 Department of Neuroscience, Erasmus MC, 3000 Rotterdam, The Netherlands 2 Netherlands Institute of Neuroscience, Royal Dutch Academy for Arts and Sciences, Amsterdam 1105 BA, The Netherlands † These authors contributed equally to this work. Electronic supplementary material is available online at https://doi.org/10.6084/m9.figshare.c.6619748. 10 5 2023 May 10, 2023 5 2023 10 5 2023 May 10, 2023 13 5 22012128 4 2022 April 28, 2022 12 4 2023 April 12, 2023 © 2023 The Authors. 2023 https://creativecommons.org/licenses/by/4.0/ Published by the Royal Society under the terms of the Creative Commons Attribution License http://creativecommons.org/licenses/by/4.0/, which permits unrestricted use, provided the original author and source are credited. Delay eyeblink conditioning has been extensively used to study associative learning and the cerebellar circuits underlying this task have been largely identified. However, there is a little knowledge on how factors such as strain, sex and innate behaviour influence performance during this type of learning. In this study, we used male and female mice of C57BL/6J (B6) and B6CBAF1 strains to investigate the effect of sex, strain and locomotion in delay eyeblink conditioning. We performed a short and a long delay eyeblink conditioning paradigm and used a c-Fos immunostaining approach to explore the involvement of different brain areas in this task. We found that both B6 and B6CBAF1 females reach higher learning scores compared to males in the initial stages of learning. This sex-dependent difference was no longer present as the learning progressed. Moreover, we found a strong positive correlation between learning scores and voluntary locomotion irrespective of the training duration. c-Fos immunostainings after the short paradigm showed positive correlations between c-Fos expression and learning scores in the cerebellar cortex and brainstem, as well as previously unreported areas. By contrast, after the long paradigm, c-Fos expression was only significantly elevated in the brainstem. Taken together, we show that differences in voluntary locomotion and activity across brain areas correlate with performance in delay eyeblink conditioning across strains and sexes. eyeblink , learning , locomotion , c-Fos , sex , strain Dutch Research Council Vidi/ZonMw/917.18.380,2018 ==== Body pmc1. Introduction Eyeblink conditioning is a behavioural paradigm commonly used to study cerebellar-dependent associative learning. It has been applied in a variety of animals including mice, cats, rats, rabbits and ferrets [1–5], and has also been used in human research to identify behavioural deficits in Parkinson's disease, multiple sclerosis and autism spectrum disorder. [6–8]. In this task, an initially neutral, conditioned stimulus (CS; a flashing light or sound), becomes predictive of an unconditioned stimulus (US; an air puff to the cornea), which elicits a blink of the eye. The paradigm consists of pairing the CS with the US and, over time, an association is formed where blinking is triggered by the presentation of the CS alone. The newly learned association is called conditioned response (CR) [9]. The most commonly used version of this task is the so-called delay eyeblink conditioning paradigm, in which the presentation of the US immediately follows the CS with both of the stimuli ending at the same time [10]. At the circuit level, studies have shown that the association between stimuli during delay eyeblink conditioning relies on the cerebellum. Here, the CS signals coming from the pons and the US information (puff) from the inferior olive via climbing fibres are precisely timed and processed [11,12]. Consequently, several cerebellar areas modulating eyeblink conditioning have been identified in mice, including lobule VI in the vermal region, and crus 1 and simplex in the hemispheric region [13,14]. Inactivation of lobule VI and crus 1 during development causes deficits in learning, supporting their crucial role in eyeblink conditioning [15]. The CR signal leaves the cerebellum via the interposed nucleus, which ultimately connects to the muscles controlling the eyeblink reflex [2,14]. Beyond the ponto-cerebellar and olivo-cerebellar systems, little is known about the potential involvement of other brain areas in delay eyeblink conditioning [16,17], although the amygdala has also been proposed to have a role in this form of associative learning, given its implication in fear conditioning and arousal [18]. Despite extensive knowledge on the cerebellar substrates involved in eyeblink conditioning, our understanding of how intrinsic factors, such as sex, strain and behaviour, contribute to the variability in performance during this task remains limited. As with the vast majority of behavioural mouse tasks, male C57BL/6 (B6) mice are preferentially used to investigate eyeblink performance, with other strains and female mice remaining considerably under-investigated. While strain limitations are usually associated with the use of inbred strains with intention to decrease variability and increase statistical power [19], the underuse of females relates to the presumption that hormonal fluctuations caused by the estrous cycle introduce non-comparable variability across sexes [20]. This remains a common belief despite meta-analyses of rat and mice studies showing that females and males exhibit comparable behavioural, morphological and physiological trait variability [21,22]. In general, knowledge on the contribution of sex towards performance variability in delay eyeblink conditioning is scarce and often conflicting. In humans, females produce more CRs than males both in children below the age of nine, and in adults [23]. In rats, stress enhances conditioning responses but hinders learning in females [24]. However, in rabbits, both males and females show similar conditioning profiles, with females adapting faster to stress [25]. A recent study of trace eyeblink conditioning, a paradigm in which the US is separated from the CS by a time interval, showed that both sexes reached similar learning scores but female mice showed significantly higher CR percentage compared to males in the first 5 days of learning [26]. Given this heterogeneity of results, gaining a deeper understanding of the effects of sex in eyeblink conditioning performance is crucial to better understand which parameters influence learning during this task. Sex and strain are not the only variables that can potentially influence mice performance during delay eyeblink conditioning. This task is most commonly performed in head-fixed mice freely running on a wheel and studies have shown that locomotion speed on the wheel positively correlates with a higher percentage of CRs [27,28]. However, it is still unknown if locomotion strategy during delay eyeblink conditioning is sex dependent. This is highly plausible, given that sex-specific strategies in mice were previously observed in locomotion adaptation [29], exercise capacity [30] and spatial orientation and learning [31], where female mice learned faster. Of note, brain activity in a number of areas involved in delay eyeblink conditioning is modulated by locomotor activity. In the cerebellum, locomotion changes the baseline neuronal activity in the cortex [32,33], although the relevance of this modulation towards associative learning is still not fully understood. Arousal and locomotion influence spontaneous and evoked activity in the neocortex and thalamus [34,35], and recruit many brainstem regions [36]. How locomotion modulates the activation of distinct brain areas during delay eyeblink conditioning, particularly outside of the cerebellar circuit, remains an open question. One method to evaluate brain activity in areas involved in a learning task is through c-Fos detection. c-fos is an immediate early gene, whose expression in neurons is relatively low at rest, but rapidly increases shortly after depolarization, especially following learning paradigms [37]. Upon neuronal activation, c-fos mRNA reaches a maximum expression level after approximately 30 min, while protein expression peaks within 1 to 2 h post-activation [38]. Because of this precise time window of expression, and its increase after exposure to novel objects, surroundings or stimuli, c-Fos is widely used as a neuronal activity marker [39–41]. In mice, it has been commonly used to investigate seizures [42–44], contextual conditioning and memory [45,46] and social behaviours [47–50]. c-Fos immunohistochemistry has several advantages: it is a non-invasive technique that allows quantification of brain wide neural activity, it detects activation in deep nuclei that would be otherwise hard to reach with electrophysiology, it is fast and relatively easy to perform, it allows cellular resolution, and it can be used in combination with other quantitative techniques, for example, with other cell-type specific immunostainings [51,52]. However, it also has its limitations: it depicts transcriptional activation of neurons that are activated, not inhibited, hence, it is limited to the study of excitatory events and its temporal resolution is low. Finally, the presence of c-Fos expression does not provide any information about the connectivity patterns of the areas where the expression can be visualized [51,52]. How the c-Fos expression correlates with learning in different sexes and strains, and across distinct areas following delay eyeblink conditioning remains elusive. Here, using male and female mice of the B6 and B6CBAF1 backgrounds, we investigated the effects of sex, strain and spontaneous locomotion in delay eyeblink conditioning performance. Furthermore, by using c-Fos expression as a proxy for neural activity during learning, we explored how sex, strain and locomotion activity modulate the engagement of distinct brain regions during delay eyeblink conditioning. Using a short 5-day and a standard, long 10-day paradigm, we found that B6 male and female mice showed comparable variability in the delay eyeblink conditioning with females outperforming males at day 5. This sex difference in the performance was also found in the B6CBAF1 mice exposed to the 5-day paradigm, but is no longer present in B6 mice exposed to 10 days of training. Further, we found a strong positive correlation across sexes between learning scores and voluntary locomotion in the B6 mice. Following the short training protocol, c-Fos immunostaining revealed positive correlations between c-Fos positive cell density and learning scores in the cerebellar cortex, as well as multiple previously unreported extra-cerebellar areas. Examination of the c-Fos signal in mice subjected to the long paradigm showed a significant correlation with learning scores solely in the brainstem. These results are congruous with previous studies, which show that cerebellar cortex is necessary for the acquisition of the CRs but not the execution of the learned responses [16,53–55]. 2. Results 2.1. B6 female and male mice show comparable variability in delay eyeblink conditioning but female learning scores peak faster To study differences in learning profiles between sexes, we performed delay eyeblink conditioning experiments with B6 females and males. First, we habituated the animals to the set-up for increasing periods of time over 5 days to decrease anxiety levels and optimize mouse position on the running wheel. Next, we subjected separate batches of mice to a short or long (5- or 10-day) training paradigm in order to capture behavioural variability in distinct stages of learning. The blue LED light (CS) was triggered 250 ms prior to the puff to the cornea (US) in paired trials and the two stimuli co-terminated (figure 1a). Daily sessions consisted of 20 blocks of 12 trails each (1 US-only, 11 paired and 1 CS-only). Mice learned the association between the stimuli progressively and developed a CR that gradually increased over time (figure 1b). Figure 1. B6 female and male mice show comparable variability in eyeblink conditioning and females reach higher learning scores faster. (a) Experimental set-up. Mouse with an implanted headplate is head-fixed on top of a freely rotating wheel. A blue light (CS) is presented 250 ms before a puff (US) to the same eye. In a trained mouse, the CS produces an anticipatory eyelid closure (CR), followed by a blink reflex triggered by the US (UR). (b) Paired trials average traces in females and males during the short (5-day) paradigm. The CR progressively develops due to the CS–US pairing. (c,d) CR percentage and amplitude in CS-only trials over the short (c) and long (10-day) (d) paradigm. Inset shows average response in CS-only trials on day 5 of training. Purple, females; green, males; shaded area, s.e.m. (e) Mean percentage and amplitude response in CS-only trials, comparing the performance of the cohorts used in the short and long paradigms. Differences between female and male performance are no longer detected on day 10. Purple, females (n = 14 short, n = 8 long); green, males (n = 14 short, n = 7 long); *p < 0.05, **p < 0.01. In the short, 5-day paradigm, females showed an increase in CR percentage in CS-only trials on session four that culminated with a 60% CR responses in session five opposed to 40% in males (two-way ANOVA repeated measures for sex and session × sex effect: F1,26 = 1.461, p = 0.24, interaction sex and session: F4,104 = 4.01, p = 0.02, Cohen's d session five: 0.88) (figure 1c, top panel). The CR amplitude was normalized to maximum eye closure during the US-only trials, which triggered an unconditioned response (UR, i.e. blink). The UR maximum amplitude (i.e. peak amplitude = 1) was therefore defined as the maximal eye closure in response to the puff. The CR responses were normalized for each mouse, within each session as daily adjustments of the mouse position on the wheel translated to small variations in the eye position across sessions (see Material and methods). The CR amplitude in CS trials was significantly higher in females compared to males (two-way ANOVA repeated measures for sex and session × sex effect: F1,26 = 6.109, p = 0.02, interaction sex and session: F4,104 = 5.12, p = 0.0008, Cohen's d session five: 0.93) (figure 1c, bottom panel). On day 5, females reached average amplitude of 0.55 while males reached an average of 0.33 (figure 1c, inset). Next, we performed a standard, 10-day training paradigm in a second cohort of mice (figure 1d). Congruous with the results from the first cohort, females showed higher CR percentage and amplitude after 5 days of training (two-way ANOVA; effect of paradigm on percentage: F1,39 = 0.051, p = 0.82; effect of paradigm on amplitude: F1,39 = 0.01, p = 0.94) (figure 1e). However, at the end of the long paradigm, these sex differences were no longer significant (two-way repeated measures ANOVA; effect of sex on percentage: F1,13 = 2.10, p = 0.17; effect of sex on amplitude: F1,13 = 0.97, p = 0.34) (figure 1e). Overall, these results show that male and female mice display comparable variance in delay eyeblink conditioning and that females acquire higher learning scores faster, although these sex differences dissipate after 10 days of training. 2.2. Learning scores correlate with spontaneous locomotor activity We next asked whether mice showed sex-dependent behavioural differences in voluntary locomotor activity during training, as it has been previously shown that locomotion enhances learning in delay eyeblink conditioning [27]. For this purpose, we added infrared cameras to the eyeblink set-ups to record body movement during eyeblink sessions. The cameras were placed at the right back corner of the box, allowing a wide recording angle to capture whole-body movements (electronic supplementary material, figure S1A). We recorded videos of full training sessions for each mouse, which were later analysed offline. The head bar height (figure 1a) was adjusted accordingly to ensure that every mouse could move comfortably on the wheel. To track different body parts and get a meaningful movement output, we used DeepLabCut (DLC), a software for automated animal pose tracking [56] (see Material and methods, Locomotion analysis). This approach allows movement tracking without using physical markers on the body that can hinder natural movement. We tracked five body parts: tail base, hip, knee, right back paw and nose (electronic supplementary material, figure S1A). Animals were head-fixed on top of the wheel; hence, y position was similar between both sexes (electronic supplementary material, figure S1B,C). We selected the speed of the right back paw as a proxy for general locomotion behaviour for further analysis. We also observed more fluctuations in the hip and tail position along the x-axis in males (electronic supplementary material, figure S1B,C). Animals increased their speed on the wheel during short and long training paradigms, and both females and males had comparable variances (F-test for two-sample variances in speed: F = 0.58, p = 0.305). We found that females moved significantly faster than males during learning, reaching on the last day an average speed of 31 and 27 cm s−1 compared to males that reached 24 and 21 cm s−1 during the short and long paradigms, respectively (two-way ANOVA repeated measures for sex and session × sex effect in short paradigm: F1,26 = 12.17, p = 0.0017, Cohen's d session five: 1.07; effect of sex in long paradigm: F1,13 = 7.81, p = 0.02) (figure 2a,b). Because we observed a similar trend between CR amplitude and running speed across sexes and paradigms, we performed a linear regression between speed of the right back paw and CR amplitude in the last session of each paradigm. This showed a clear correlation between the variables (short paradigm: R2 = 0.75, p = 0.002; long paradigm: R2 = 0.72, p < 0.0001) (figure 2c). These results reveal that mice that spontaneously move faster on the wheel, reach higher learning scores in delay eyeblink conditioning. Figure 2. Learning scores correlate with spontaneous locomotor activity. (a,b) CR amplitude and speed of the right back paw over training sessions for the short (i) and long (ii) training paradigms. Purple, females (n = 14 short, n = 8 long); green, males (n = 14 short, n = 7 long); shaded area, s.e.m. Speed: two-way ANOVA for sex and session × sex effect short paradigm: F1,26 = 12.17, p = 0.0017; sex effect long paradigm F1,13 = 7.81, p = 0.02. (c) Positive correlation between CR amplitude and speed of the right back paw on the last session of training (linear regression: short paradigm R2 = 0.75, p = 0.002; long paradigm R2 = 0.72, p < 0.0001). 2.3. Learning scores correlate with c-Fos expression To explore if differences in eyeblink performance correlate with differences in brain activity, we performed c-Fos immunostainings after the last training session of the short and long paradigm. We imaged sections of whole brains with a fluorescent microscope and developed an image analysis workflow to quantify c-Fos positive neurons and identify their location within the hierarchical structure of the Allen Brian Atlas (cerebrum, brainstem and cerebellum) (electronic supplementary material, figure S2). To identify regions potentially involved in associative learning, we first selected mice that showed a CR amplitude of 0.4 or higher in CS-only trials (n = 14, 5 B6 males and 9 B6 females for the short paradigm; n = 13, 6 B6 males and 7 B6 females for the long paradigm). These mice are further referred to as ‘high learners’. We performed Kendall's correlation between the density of c-Fos positive cells and the CR amplitude reached during the last session. In the cerebellum, c-Fos labelling was localized in the granule cell layer in the mice that underwent the short paradigm training, which we confirmed with colocalization with GABAα6, a granule cell-specific marker (figure 3a; electronic supplementary material, figures S3 and S4A,B). In the cerebellar hemispheric regions, crus 1 and simplex had a significant correlation between c-Fos cell density and CR amplitude (crus 1: tau = 0.42, p = 0.042, simplex: tau = 0.52, p = 0.009). In the cerebellar vermis, lobule VI also had a significant correlation and the highest Tau (lobule VI: tau = 0.8, p = 0.009) (figure 3b,c). No significant correlations were found between c-Fos expression and CR amplitude in cerebral or cerebellar areas following the 10-day training paradigm (electronic supplementary material, table S1). Figure 3. Learning scores correlate with c-Fos expression. (a) c-Fos positive granule cells in lobule VI in the cerebellum. (b) Cerebellar areas with a significant positive correlation between c-Fos positive cell density and CR amplitude (crus 1: tau = 0.42, p = 0.042, simplex: tau = 0.52, p = 0.009, lobule VI: tau = 0.8, p = 0.009). (c) Three-dimensional model with significant areas highlighted. (d) c-Fos positive cells in the inferior olive. (e) Brainstem areas with a significant positive correlation between c-Fos positive cell density and CR amplitude (red nucleus: tau = 0.43, p = 0.041, facial nucleus: tau = 0.57, p = 0.006, inferior olive: tau = 0.76, p = 0.0008, pontine nuclei: tau = 0.48, p = 0.021). (f) Three-dimensional model with significant areas highlighted. (g) Brainstem areas with a significant positive correlation between c-Fos positive cell density and CR amplitude in the long paradigm (red nucleus: tau = 0.46, p = 0.04), facial nucleus: tau = 0.47, p = 0.03, inferior olive: tau = 0.52, p = 0.02, pontine nuclei: tau = 0.52, p = 0.02). (h) c-Fos positive cells in the visual cortex. (i) Cortical areas with a significant positive correlation between c-Fos positive cell density and CR amplitude (visual cortex: tau = 0.51, p = 0.013, motor cortex: tau = 0.69, p = 0.0003, somatosensory cortex: 0.54, p = 0.007, amygdala: tau = 0.63, p: 0.001). (j) Three-dimensional model with significant areas highlighted. In the brain stem, we found a significant correlation in the red nucleus, the facial nucleus, the inferior olive and the pontine nuclei in the short and long paradigm cohorts (short paradigm: red nucleus: tau = 0.43, p = 0.041, facial nucleus: tau = 0.57, p = 0.006, inferior olive: tau = 0.76, p = 0.0008, pontine nuclei: tau = 0.48, p = 0.021; long paradigm: red nucleus tau = 0.46, p = 0.04, facial nucleus: tau = 0.47, p = 0.03, inferior olive: tau = 0.52, p = 0.02, and pontine nuclei: tau = 0.52, p = 0.02; figure 3d–g; electronic supplementary material, figure S3 and table S1). Significant positive correlations in the visual, motor and somatosensory cortices, and the amygdala were detected only in mice that underwent the short training (visual cortex: tau = 0.51, p = 0.013, motor cortex: tau = 0.69, p = 0.0003, somatosensory cortex: 0.54, p = 0.007, amygdala: tau = 0.63, p = 0.001) (figure 3h–j; electronic supplementary material, figure S3 and table S1). Double immunostainings using c-Fos and CUX1, a marker for upper cortical layers (II–IV), or CTIP2, for lower cortical layers (V–VI), revealed c-Fos positive cells in all layers of the cortex following the short training paradigm (electronic supplementary material, figure S4C–F). Together, these results confirm the involvement of previously reported areas associated with delay eyeblink conditioning within the olivo-cerebellar and ponto-cerebellar systems [17,57] and suggest that other areas might be involved in the acquisition of this learning task. Furthermore, the results of the long training are in line with the previous reports that suggest transfer of learned responses away from cerebellar cortex [16,53–55] and neocortex once the learning is consolidated [58]. 2.4. Pseudoconditioned mice show lower c-Fos expression To ensure an appropriate control for the quantification of c-Fos positive cells, we included pseudoconditioned mice (n = 8 for the short paradigm, 4 B6 males and 4 B6 females; n = 3 for the long paradigm, 2 B6 males and 1 B6 female) in our experimental design. These mice went through the same experimental steps as the conditioned mice, with the only exception that they were not trained with paired CS-US trials. Instead, we exposed them to a protocol with CS- and US-only trials, keeping the same structure and duration as the conditioned protocol and tracking their locomotion behaviour throughout the experiment. Pseudoconditioned mice did not acquire an association given that there was no substrate for learning (absence of CS-US pairing) but, similar to the conditioned mice, did increase the locomotion speed over training sessions (short paradigm: F2,733, 19,13 = 49.23, p < 0.0001; long paradigm: F1,749, 3,497 = 19.28, p = 0.01) (figure 4a). We applied the same analysis method to quantify the density of c-Fos positive cells in pseudoconditioned mice as described above. Overall, we observed lower c-Fos expression in pseudoconditioned mice compared to high learners and low learners (effect of learning: short paradigm- F2,363 = 251.5, p < 0.0001; mean c-Fos density high learners = 61.74, low learners = 44.45, pseudoconditioned = 17.83; long paradigm H(2) = 28.41, p < 0.001; median c-Fos density high learners = 18.46, low learners = 15.84, pseudoconditioned = 6.77) (figure 4b; electronic supplementary material, figure S5 and table S2). We then compared the c-Fos density between pseudoconditioned, high and low learners. We specifically focused on the brain areas where we found a significant correlation between CR amplitude and c-Fos density during the short paradigm (figure 3). In the short paradigm, both high and low learners had significantly higher c-Fos density mean compared to pseudoconditioned mice in the VC, COA, PG, SIM, ANcr1 and LVI (figure 4b top panel; figure 4c). Differences between high and low learners were also observed in the VC, PG and LVI. Following the long paradigm, no differences in c-Fos density were found between high and low learners (figure 4b bottom panel), while increased c-Fos density was found across all areas (except for SS) between high-learners and pseudoconditioned mice. This suggests that the observed increased c-Fos expression in a number of brain areas is related to the learning of the CS-US pairing and not just to the locomotion or sensory stimulation (light and puff) alone. Figure 4. Pseudoconditioned mice show lower c-Fos expression compared to high and low learners. (a) Change in CR amplitude and speed of the right back paw of pseudoconditioned mice over training sessions for the short (top; n = 8) and long (bottom; n = 3) paradigms. CR amplitude in the left y-axis and speed in the right y-axis. (b) Boxplots depict c-Fos density in high learners (CR amplitude > 0.4 on last session; n = 14 short paradigm, n = 13 long paradigm), low learners (CR amplitude < 0.4 on last session; n = 14 short paradigm, n = 2 long paradigm) and pseudoconditioned mice on the short (top) and long (bottom) paradigm. (c) Representative immunofluorescent images of c-Fos positive cells in a pseudocontidoned mouse (left) and a high-learner mouse (right). Top to bottom: cerebellar cortex, motor cortex, somatosensory cortex. Scale bar: 100 μm. *p < 0.05, **p < 0.01, ***p < 0.001, #p < 0.0001. 2.5. Correlation between learning scores and c-Fos expression is consistent in B6CBAF1 strain Given the evidence of the possible unwanted effects of highly inbred mouse strains like B6 in replicability and reproducibility [59], we wanted to investigate inter-strain variability in associative learning. We asked whether the results obtained in B6 mice could be observed in a different mouse strain. For this purpose, we subjected B6CBAF1 mice (the F1 hybrids of B6 and CBA strains) to the short delay eyeblink conditioning paradigm. B6CBAF1 mice have significantly less retinal degeneration and hearing loss compared to B6 mice, which makes them favourable models for visual and auditory experiments [60–62]. B6CBAF1 mice learned the association between the stimuli and gradually formed CRs. We observed a trend indicating similar sex differences between B6CBAF1 mice and B6. Females reached 53% CR percentage compared to 35% in males (two-way ANOVA repeated measures for sex and session × sex effect: F1,14 = 2.55, p = 0.237, interaction sex and session: F4,104 = 3.01, p = 0.021, Cohen's d session five: 0.86) (figure 5a). When it comes to the amplitude of these responses, BFCBAF1 females showed a trend towards slightly higher CR amplitude over training sessions compared to males (two-way ANOVA repeated measures for sex and session × sex effect: F1,14 = 2.55, p = 0.132, Cohen's d session five: 0.84) (figure 5b). Figure 5. Correlation between learning scores and c-Fos expression is consistent in B6CBAF1 mice. (a) CR percentage in CS only trials over training sessions. Yellow, females (n = 9); cyan, males (n = 7); shaded area, s.e.m. (b) CR amplitude in CS only trials over training sessions. Shaded area, s.e.m. (c) c-Fos positive cells in lobule VI. (d) Cerebellar areas with a significant positive correlation between c-Fos positive cell density and CR amplitude in high learners (crus 1: tau = 0.82, p = 0.0001, simplex: tau = 0.7, p = 0.005, lobule VI: tau = 0.75, p = 0.0007). (e) c-Fos positive cells in the inferior olive. (f) Brainstem areas with a significant positive correlation between c-Fos positive cell density and CR amplitude in high learners (inferior olive: tau = 0.85, p = 0.0004, pontine nuclei: tau = 0.78, p = 0.0003). (g) c-Fos positive cells in the visual cortex. (h) Cortical areas with a significant positive correlation between c-Fos positive cell density and CR amplitude in high learners (visual cortex: tau = 0.64, p = 0.0057, motor cortex: tau = 0.75, p = 0.0008, somatosensory cortex: tau = 0.6, p = 0.009). Following the last day of learning, the B6CBAF1 mice were perfused and we quantified the whole-brain c-Fos expression (electronic supplementary material, figure S6). We selected mice that showed a CR amplitude of 0.4 or higher in CS-only trials (n = 11; 3 males and 8 females) and performed Kendall's correlation between density of c-Fos positive cells and CR amplitude on the last training session. The granule cell layer also contained c-Fos labelling, and crus 1, the simplex and lobule VI were found to have a significant positive correlation (crus 1: tau = 0.82, p = 0.0001, simplex: tau = 0.7, p = 0.005, lobule VI: tau = 0.75, p = 0.0007) (figure 5c,d; electronic supplementary material, figure S7). In the hindbrain, the correlation between c-Fos cells and learning was also significant in the inferior olive and the pontine nuclei (inferior olive: tau = 0.85, p = 0.0004, pontine nuclei: tau = 0.78, p = 0.0003) (figure 5e,f; electronic supplementary material, figure S7). Additionally, the visual, motor and somatosensory cortices showed significant positive correlations (visual cortex: tau = 0.64, p = 0.0057, motor cortex: tau = 0.75 p = 0.0008, somatosensory cortex: tau = 0.6, p = 0.009) (figure 5g,h; electronic supplementary material, figure S7). These results show that most of the areas where we saw a correlation between learning and c-Fos density in B6 mice are also found in B6CBAF1 mice, which strengthens the idea that these areas are active during delay eyeblink conditioning. 2.6. Variability between sexes and strains To further understand inter-strain and inter-sex variability in our dataset, we calculated the coefficient of variance (CV, s.d./mean) for each of the variables that we quantified in both B6 and B6CBAF1 mice. For mice that were subjected to the short paradigm, we selected the 14 high learners (CR amplitude on session 5 > 0.4) B6 mice (n = 5 males, 9 females) and the 11 high-learners B6CBAF1 mice (n = 3 males, 8 females) and grouped them by strain and sex (figure 6). For each group, we calculated the CV for the common variables acquired and previously reported, which can be grouped in two main categories: eyeblink performance and c-Fos expression. Eyeblink performance includes CR amplitude and percentage on the last day of training. c-Fos expression includes the density of c-Fos positive cells in the brain areas where we have found a positive significant correlation across both strains, during the short paradigm: crus 1, simplex, lobule VI, inferior olive, pontine nuclei, and the visual, motor and somatosensory cortices. When comparing strains, we observed that the variances for each variable were similar, with the exception of the density of c-Fos positive cells in the somatosensory cortex, where B6CBAF1 mice seem to be more variable compared to B6 (figure 6a). B6 female and male mice had similar CVs for most of the variables, although in crus 1 and the somatosensory cortex males seem to have a slightly higher CV (figure 6b). However, this could be due to the difference in sample size. We observed something similar between B6CBAF1 female and male mice; males showed slightly higher CV in c-Fos density in the simplex (figure 6c). Additionally, B6CBAF1 mice had the highest CV in the somatosensory cortex. When we calculated the CR amplitude variance across all days of training in the short paradigm, female mice presented with higher variance compared to male mice (F-test for two sample variances in CR amplitude of paired trials- B6: F = 1.74, p = 0.01; B6CBAF1: F = 1.77, p = 0.04), reflecting the higher number of high learners in the female population in the early stages of training. Notably, both B6 and B6CBAF1 females and males presented with very similar variances (B6 females = 0.052, B6CBAF1 females = 0.058; B6 males = 0.030, B6CBAF1 males = 0.033). In the long paradigm cohort, we analysed CVs for the eyeblink performance variables, as well as for the brain areas where we have found a positive significant correlation: red nucleus, facial nucleus, inferior olive and pontine nuclei (figure 3g). Compared to the short paradigm data, long paradigm B6 males and B6 females present with increasing overlapping CVs in the analysed variables (figure 6d). Furthermore, the analysis of CR amplitude variance across the 10 days of training showed no effect of sex on this variable (F = 1.15, p = 0.28), suggesting that both eyeblink performance and c-Fos expression become more similar between sexes with increasing training time. Figure 6. Strain and sex variability. High learners were selected based on CR amplitude on the last training session > 0.4. CV = s.d./mean. (a) For strains: B6, n = 14, B6CBAF1, n = 11. (b) For B6 mice: males, n = 5; females, n = 9. (c) For B6CBAF1 mice: males, n = 3; females, n = 8. (d) For the B6 long paradigm: males, n = 6; females, n = 7. VC, visual cortex; MC, motor cortex; SSM, somatosensory cortex; CV, coefficient of variation. 3. Discussion Understanding behavioural variability in the context of neuroscience research is a challenge. We are still far from fully comprehending how factors like sex and strain, and differences in baseline activity levels give rise to differences in performance in learning tasks. Here we addressed this question by making use of a well-known learning paradigm to study behavioural variability. We found that B6 and B6CBAF1 mice showed comparable variance in delay eyeblink conditioning and locomotion while being head-fixed on a rotating wheel. The differences between sexes were evident with females reaching higher learning scores and running speeds within 5 days of training, but dissipated after a 10-day paradigm. Importantly, we found a robust correlation between learning scores and running speed, which was consistent across sexes and strains. In a similar way, we found that enriched c-Fos expression across several brain areas positively correlates with learning, which suggests the involvement of these regions in eyeblink conditioning. 3.1. Sex, strain and behaviour: comparable variability but difference in performance Opposed to what is sometimes assumed in behavioural studies, we observed that sexes show comparable variability as the training progressed. However, we found sex-dependent differences in performance during the early stages of the acquisition of delay eyeblink conditioning, and in the locomotion patterns. Past studies on sex differences in rodents showed that many behaviours vary between females and males [63]. This is true for innate behaviours such as running [64,65], where female mice show increased duration and speeds of locomotion compared to males, as well as in learned responses measured in laboratory settings, such as spatial learning [66], where male rats outperform females. However, data on sex differences in research is often conflicting, with several studies not finding sex-dependent behavioural differences [67,68]. These divergent results can be attributed to a variety of causes. However, increasing studies point to differences in behavioural strategies and differential expression of fear behaviour as sources for result variability [69,70]. Behavioural sex differences have also been studied in multiple other species. In addition to differences in mating behaviours, non-reproductive behavioural differences have also been found in fish, including increased discrimination in reversal learning in females [71,72]. In rhesus monkeys, females perform better in a spatial delayed-response task [73], while young female chimpanzees exhibit thermite fishing behaviour earlier than males [74]. Notably, conflicting reports on sex-divergent behaviours can also be found in humans, with some studies pointing to significant sex effects [75–77], where others fail to show any significant differences between sexes [78]. Therefore, generalizations of sex-dependent findings across species should be carefully examined and limited to paradigms that can be applied in the same manner across species. In the context of cerebellar-dependent learning, evidence in mice shows that estradiol increases the density of parallel fibre to Purkinje cell synapse and induces LTP, which improves memory formation [79] in mice. Experiments using the trace eyeblink conditioning paradigm in mice showed that both sexes reach similar learning scores but females reach significantly higher CR percentage compared to males in the first 5 days of learning [26]. These findings together with our results suggest that female mice exhibit faster learning rates during the first stages of learning. These results in mice are in line with studies performed in humans, where females were observed to produce more CRs than males throughout the duration of the delay eyeblink conditioning paradigm [23]. It should be noted however that, unlike mouse experiments, in human studies, all trials are delivered during a single training day. It would be of interest to investigate whether males reach learning scores comparable to females over several training sessions. Given that our data and previous reports [26] show faster acquisition of CRs in females of both B6 and B6CBAF1 mouse strains, we postulate that including female mice in future eyeblink studies could reduce the training time to achieve desired scores. This could be advantageous for certain experiments, particularly time-sensitive ones, such as calcium imaging or electrophysiological measurements. In addition, our results show that male and female mice have similar variability regardless of the strain, which indicates that females can be included in studies without tracking the estrous cycle phase. A common experimental setting in neuroscience involves head-fixing awake mice and placing them on a freely moving wheel. In our experiments, the differences in learning scores between sexes were strongly correlated to the changes in speed on the wheel, which is in line with the previous studies [27]. This suggests that spontaneous locomotion might facilitate cerebellar associative learning and could be predictive of learning scores. Locomotion has long been known to induce theta rhythms in the hippocampus [80–82], which in turn facilitate episodic memory consolidation and hippocampal-dependent learning [83,84], and have been shown to be important for social memory formation [85]. In humans, theta rhythms have been linked to active learning [86] and improved episodic memory recall [87]. It would be of interest to investigate how hippocampal theta rhythms influence cerebellar activity during delay eyeblink conditioning in future research by performing dual cerebellar and hippocampal recordings during eyeblink training. Another factor that should be taken into consideration when studying sex differences in performance is stress. It is known that stress plays an important role in modulating neural activity in the hippocampus. Corticosterone- among other stress hormones- increases CA1/CA3 firing rates shortly after a stressful period and induces molecular cascades that enhance calcium influx, which disrupts hippocampal function [88]. Similar mechanisms have been described in the cerebellum, where calcium-based excitability in the DCN is altered in animals with higher levels of corticosterone [89]. Regarding locomotion, the evidence on the effects of stress on walking and running behaviours is inconsistent. Some studies using head-fixed male mice showed that these increased wheel running speeds after experiencing different forms of induced stress [90,91]. Another study using male mice but a different form of head fixation, a ‘mobile home cage’ (MHC) container, found that corticosterone levels were generally lower with larger numbers of rotations of the MHC container [92]. However, others found that male mice not habituated to a wheel run less after stress exposure [93]. Studies that include mice from both sexes in their experimental design have mostly been limited to freely moving behaviours and showed that females exhibit increased exploratory behaviour but not running upon stress [94,95]. A recent study that investigated adaptation to head fixation in both sexes showed that, over the 7-day habituation period, both males and females increased the velocity and the time spent running. However, female mice habituated to running forward within the initial 2 days, while male mice took up to 4 days to habituate to running forward [29]. Unfortunately, in our current study, we did not track the movement over the habituation period. Thus, comparisons between our data and the Warner and Padmanabhan results are not possible. Future eyeblink experiments in combination with stress monitoring, such as cortisol measurements and quantification of innate running behaviour in homecages using for example ‘smart housing’ solutions, would help elucidate the contributors to the correlation between locomotion and sex found in our study. 3.2. Associative learning networks Our results show that the expression of c-Fos in the cerebellar cortex, following the short paradigm of delay eyeblink conditioning, is localized to the granule cell layer. This is expected, given that multiple forms of plasticity have been shown within the synapses in this layer. For example, the mossy fibre-granule cell synapse undergoes both LTP and LTD [96], and evidence has shown that granule cell activity adapts over time during eyeblink conditioning [97] and other types of learning [98,99]. In addition, the induction of LTP by theta-burst stimulation in acute cerebellar slices activates a cAMP-responsive element-binding protein cascade which, in turn, activates c-Fos expression [100]. Our results are consistent with these findings and, overall, they provide evidence on how plasticity, at the input level in the cerebellar cortex, can evoke transcriptional processes that contribute to learning consolidation. The strongest correlations between c-Fos expression and CR amplitude within the cerebellum were observed in simplex, lobule VI and crus 1, which is consistent with the ‘eyeblink region’, but expands beyond the small area usually recorded using electrophysiological approaches [12,13]. Strong c-Fos expression in crus 1 supports our previous findings showing the importance of this lobule in eyeblink conditioning [15]. Outside of the cerebellum, we identified several brain areas that could play a role in eyeblink conditioning. At the brainstem level, we found a positive correlation between CR amplitude and c-Fos density in the pons, inferior olive, red nucleus and facial nucleus. Activity in these areas can be expected given that the red and facial nuclei, together with the oculomotor nucleus, execute the blink. Similarly, the inferior olive and the pontine nuclei relay the US and CS information to the cerebellar cortex, respectively. The positive correlation between learning scores and c-Fos expression was significant in the short and the long paradigm, which suggests the importance of these regions in acquisition and consolidation of delay eyeblink conditioning. Moreover, we found higher c-Fos expression in the visual, motor and somatosensory cortices in mice with higher learning scores at day 5 of learning. Processing in these cortices could facilitate the CS to become more salient and ultimately predict the US. The somatosensory cortex projects to the lateral amygdala which, in turn, projects to the central amygdala, to ultimately contact the pons. The high c-Fos expression found in the amygdala points towards a two-stage conditioning model; where the amygdala would have an initial role with arousal as a salient feature, and a second phase where the cerebellum would take over to form precisely timed CRs [16]. In the motor cortex, higher c-Fos levels in high-performing mice might be due to locomotor activity rather than learning itself. However, as mentioned above, this could play a role in learning either by directly affecting cerebellar input or indirectly as arousal. The comparison of the short- and long-learning paradigm revealed that the positive correlation between c-Fos density and CR amplitude in brainstem areas remained significant in both paradigms. This suggests that the red nucleus, facial nucleus, inferior olive and pontine nuclei have a similar engagement in early stages as well as in later stages of learning. However, the positive correlations in the cerebellar cortex as well as in the visual, motor, somatosensory cortices and the amygdala were significant in the short paradigm but not significant in the long paradigm. This indicates a distinct involvement of these regions between early and late stages of learning. We have observed that eyeblink scores start to plateau on days 5–6 of training, which marks a transition between early and late learning stages and is consistent with the hypothesized memory transfer from the cerebellar cortex [16,55]. Our results are also congruent with studies which show that the motor cortex is not required for an execution of a learned behaviour once the learning is consolidated [58,101]. Notably, the temporal resolution of the c-Fos staining is very poor and the presence of c-Fos expression alone does not provide us with any information about the activity patterns directly related to the stimuli used in the eyeblink conditioning task [51,52]. Nonetheless, our findings give us a better understanding of the networks underlying eyeblink conditioning and provide candidate brain areas, which can be further studied in the context of associative learning, including methods that measure neuronal activity directly with high temporal resolution (i.e. [102,103]). 4. Materials and methods 4.1. Animals C57BL/6J mice were ordered from Charles River (n = 27 males; n = 27 females), and B6CBAF1 mice from Janvier (n = 7 males; n = 9 females). Mice were group-housed and kept on a 12 h light-dark cycle with ad libitum food and water. All procedures were performed in male and female mice approximately 8–12 weeks of age. 4.2. Eyeblink pedestal placement surgery Mice were anaesthetized with isoflurane and oxygen (4% isoflurane for induction and 2–2.5% for maintenance). Body temperature was monitored during the procedure and maintained at 37°C. Animals were fixed in a stereotaxic device (Model 963, David Kopf Instruments, Tujunga CA, USA). The surgery followed previously described standard procedures for pedestal placement [2,14]. In short, the hair on top of the head was shaved, Betadine and lidocaine were applied on the skin and an incision was done in the scalp to expose the skull. The tissue on top of the skull was removed and the skull was kept dry before applying Optibond prime adhesive (Kerr, Bioggio, Switzerland). A pedestal equipped with a magnet (weight approx. 1 g) was placed on top of the skull with Charisma (Heraeus Kulzer, Armonk NY, USA), which was hardened with UV light. Rimadyl was injected subcutaneously (5 mg per kg). Mice were left under a heating lamp for recovery for at least 3 h. Mice were given 3–4 resting days before starting experiments. 4.3. Eyeblink conditioning Mice were habituated to the set-up (head-fixed to a bar suspended over a cylindrical wheel in a sound and light isolating chamber) for 5 days with increasing exposure (15, 15, 30, 45 and 60 min). Training started after 2 days of rest. Forty-three C57BL/6 mice (n = 21 males; n = 22 females), and 16 B6CBAF1 mice (n = 7 males; n = 9 females) were trained using the standard eyeblink protocol [2,12] (figure 1a). Ten CS-only trials of 30 ms with an inter-trial interval (ITI) of 10 ± 2 s were presented before the first training session to acquire a baseline measurement. Mice were next trained for 5 (short protocol) or 10 (long protocol) consecutive days. Each session consisted of 20 blocks of 12 trails each (1 US only, 11 paired and 1 CS only) with an ITI of 10 ± 2 s. The CS was a 270 ms blue LED light (approx. 450 nm) placed 7 cm in front of the mouse. The US was a 30 ms corneal air puff co-terminating with the CS. The puffer was controlled by a VHS P/P solenoid valve set at 30 psi (Lohm rate, 4750 Lohms; Internal volume, 30 µl, The Lee Company, Westbrook, USA) and delivered via a 27.5 mm gauge needle at 5 mm from the centre of the left cornea. The puff was tested every day to make sure that the startle response is minimal. Eye traces were monitored at the beginning of experiments and, if necessary, the puff position was adjusted to avoid a startle response. The inter-stimulus interval was 250 ms. Eyelid movements were recorded with a camera (Basler aceA640) at 250 frames s−1. Eleven C57BL/6 mice (n = 6 males; n = 5 females) were trained using a pseudoconditioning protocol. Pseudoconditioning protocol consisted of 20 blocks of 12 trails each (1 puff only, 12 LED only) with an ITI of 10 ± 2 s. The puff and LED stimulus had the same characteristics as in the conditioning protocol. Data were analysed with a custom written MATLAB code as previously described [15,97]. Traces were normalized within each session to the UR max amplitude. The CR detection window was set to 650–730 ms and CRs were only classified as such when the amplitude was equal or higher than 5% of the UR median. The CR percentage was calculated as the number of counted CRs (equal or higher than 5% of the UR median) divided by the total CS trials per session. 4.4. Locomotion An infrared camera (ELP 1080P) (sampling frequency 60 frames s−1) was placed in each of the eyeblink boxes and connected to an external computer (independent from the eyeblink system). The cameras were positioned at the right back corner of the chamber on top of a magnet tripod attached to a custom-made metal block which allowed stable fixation. The recording angle was standardized by selecting the same reference in the field of view of each camera. Simultaneous video acquisition from the three cameras was performed in Ipi Recorder software (http://ipisoft.com/download/). Body movement recording was parallel to eyelid recording during the training sessions. The output videos (.avi format) from each mouse and session were approximately 35 min (corresponding to the length of an eyeblink session). 4.5. Locomotion analysis We used DLC to track body parts from videos [56] (electronic supplementary material, figure S1). We extracted 40 frames of four different videos from two males and two females (a total of 160 frames). Next, frames were manually labelled with five body parts (tail base, hip, knee, right back paw and nose). These frames were used for training the pre-trained deep neural network ResNet50 [104,105]. Evaluation of the network was done to confirm a low error in pixels between labelled frames and predictions. Video analysis was done by using the trained network to get the locations of body parts from all mice and sessions [(36 B6 mice short paradigm × 5 days) + (18 B6 mice long paradigm × 10 days) = 360 videos]. DLC output is a matrix with x and y positions in pixels and the likelihood of this position for each body part. We used this matrix to calculate distance covered and speed per body part with a custom written code (https://github.com/BaduraLab/DLC_analysis). After confirming normal distribution of the spatial coordinates per body part over training sessions, we performed the Grubbs's test for outlier removal to discard possible tracking errors. 4.6. Tissue processing Mice were anaesthetized with 0.2 ml pentobarbital (60 mg ml−1) and perfused with 0.9% NaCl followed by 4% paraformaldehyde (PFA). Given the peak time expression of c-Fos [39], animals were perfused 90 min after finishing the last training session. Brains were dissected from the skull and stored in 4% PFA at room temperature (rT) for 1.5 h. They were next changed to a 10% sucrose solution and left overnight at 4°C. Brains were embedded in 12% gelatine and 10% sucrose and left in a solution with 30% sucrose and 4% PFA in PBS at rT for 1.5 h. Next, they were transferred to a 30% sucrose solution in 0.1 PB and kept at 4°C. Whole brains were sliced at 50 µm with a microtome and slices were kept in 0.1 M PB. 4.7. Immunostaining and imaging Sections were incubated in blocking solution (10% NHS, 0.5% Triton in PBS) for an hour at rT. After rinsing, sections were incubated for 48 h at 4°C on a shaker in primary antibody solution with 2% NHS (1:2000 Rabbit anti-c-Fos, ab208942, Abcam; 1:1000 Rat anti-Ctip2, ab18465, Abcam; 1:1000 Rabbit anti-GABAalpha6, G5555, Sigma-Aldrich; 1:1000 Rabbit anti-Cux1) [106]. After rinsing, sections were incubated for 2 h at rT on a shaker with secondary antibody (1:500 Donkey anti-rabbit A594, 711-585-152, Jackson; 1:500 Donkey anti-Rabbit A488, 711-545-152, Jackson; 1:500 Donkey anti-rabbit Cy5, 711-175-152, Jackson; Donkey anti-rat Cy3, 712-165-150, Jackson). Sections were counterstained with DAPI. Finally, sections were rinsed in 0.1 M PB, placed with chroomulin on coverslips and mounted on slide glasses with Mowiol. Sections were imaged with a Zeiss AxioImager 2 (Carl Zeiss, Jena, Germany) at 10 x. A DsRed filter and an exposure time of 300 ms was used for the Alexa 595 channel (c-Fos). The DAPI channel was scanned at 20 or 30 ms exposure time. Tile scans were taken from whole-brain slices. We processed half the sections obtained from slicing, hence, the distance between tile scan images was 100 µm. High-resolution images were taken with a LSM 700 confocal microscope (Carl Zeiss, Jena, Germany). 4.8. Image analysis We developed an image analysis workflow for brain region identification and quantification of c-Fos positive neurons following eyeblink conditioning (electronic supplementary material, figure S2) (https://github.com/BaduraLab/cell-counting). The workflow combines Fiji and a SHARP-Track, a software written in MATLAB initially developed to localize brain regions traversed by electrode tracks [107] (https://github.com/cortex-lab/allenCCF/tree/master/SHARP-Track). Brain slices were preprocessed (rotating, cropping and scaling) with a custom written macro in Fiji [108]. Next, slices were registered to the Allen Brain Atlas using the SHARP-Track user interface. Segmentation was performed on the registered slices in Fiji. Given the characteristic c-Fos staining pattern in the cerebellar granule layer (figure 3a), we used different thresholding algorithms for the cerebellum and for the rest of the brain. Following that step, automated cell counting of c-Fos positive neurons was performed with a custom written macro in Fiji (cerebellum—circularity: 0.5–1, size: 0–20 pixels, rest of the brain—circularity: 0.7–1, size: 0–40 pixels) to get the x and y coordinates of every detected cell. The output matrix of coordinates was used to create a ROI array per slice in SHARP-Track. This step allows one to one matching between the ROI array and the previously registered slice. Finally, the reference-space locations and brain regions of each neuron were obtained by overlapping the registration array with the ROI array. ROI counts were normalized by brain region surface following the hierarchical structure of the Allen Brain Atlas. The surface of each brain area was calculated per slice and cell density was defined as ROI counts/surface. 4.9. Statistics Statistics were performed in MATLAB and GraphPad Prism 6. Data are reported as mean ± s.d. or s.e.m. Normality was tested and followed for both eyeblink CR amplitudes and for speed of the right back paw. The corresponding statistical test for the p-values reported are specified in Results. Time data (training sessions) was analysed using two-way repeated measures ANOVA for sex and session. Data were corrected for multiple comparisons using the Sidak or Tukey tests. Sex effect is reported in Results; session effect is significant in all groups of the short and long paradigm (indicating learning over time) and interaction is reported if significant. For Kendall's correlation on c-Fos data, we report tau and p-values. This correlation method was chosen because it does not make assumptions on the data distribution (unlike Pearson's correlation) [109]. For the analysis of c-Fos density in distinct brain areas, a two-way ANOVA with Tukey's multiple comparisons test was used for the short paradigm data. Given the low number of non-learners in the long paradigm, data weres analysed with a Kruskal–Wallis and Dunn's multiple comparisons test. Acknowledgements We thank Roxanne ter Haar, Elize D. Haasdijk and Stephanie Dijkhuizen for their help with eyeblink experiments. Ethics All experiments were performed in accordance with the European Communities Council Directive. All animal protocols were approved by the Dutch National Experimental Animal Committee (DEC). Data accessibility The code described in the paper is available online at https://github.com/BaduraLab/. The supporting data can be found online at https://github.com/BaduraLab/Eyeblink. The data are provided in the electronic supplementary material [110]. Authors' contributions M.R.O.: conceptualization, data curation, formal analysis, investigation, methodology, project administration, visualization, writing—original draft and writing—review and editing; I.S.: conceptualization, investigation, methodology, project administration, supervision, visualization and writing—review and editing; D.K.: investigation, methodology and project administration; S.K.E.K.: conceptualization, methodology, project administration, resources, software and writing—review and editing; A.B.: conceptualization, funding acquisition, investigation, methodology, project administration, resources, software, supervision, validation, visualization, writing—original draft and writing—review and editing. All authors gave final approval for publication and agreed to be held accountable for the work performed therein. Conflict of interest declaration The authors declare no competing interests. Funding This work was supported by the Dutch Research Council (Vidi/ZonMw/917.18.380,2018). This project has received funding from the European Union's Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement no. 955534. ==== Refs References 1. Krupa DJ, Thompson JK, Thompson RF. 1993 Localization of a memory trace in the mammalian brain. Science 260 , 989-991. (10.1126/science.8493536)8493536 2. Ten Brinke MM, Heiney SA, Wang X, Proietti-Onori M, Boele HJ, Bakermans J, Medina JF, Gao Z, De Zeeuw CI. 2017 Dynamic modulation of activity in cerebellar nuclei neurons during pavlovian eyeblink conditioning in mice. eLife 6 , e28132. (10.7554/eLife.28132)29243588 3. Hesslow G, Ivarsson M. 1996 Inhibition of the inferior olive during conditioned responses in the Decerebrate Ferret. Exp. Brain Res. Experimentelle Hirnforschung. Experimentation Cerebrale 110 , 36-46. (10.1007/BF00241372)8817254 4. Jiménez-Díaz L, Gruart A, Miñano FJ, Delgado-García JM. 2007 Extracellular amino acid levels in the interpositus nucleus during classical eyeblink conditioning in alert cats. Behav. Neurosci. 121 , 1106-1112. (10.1037/0735-7044.121.5.1106)17907842 5. Schmajuk NA, Christiansen BA. 1990 Eyeblink conditioning in rats. Physiol. Behav. 48 , 755-758. (10.1016/0031-9384(90)90221-O)2082376 6. Oristaglio J, Hyman West S, Ghaffari M, Lech MS, Verma BR, Harvey JA, Welsh JP, Malone RP. 2013 Children with autism spectrum disorders show abnormal conditioned response timing on delay, but not trace, eyeblink conditioning. Neuroscience 248 , 708-718. (10.1016/j.neuroscience.2013.06.007)23769889 7. Rampello L, Casolla B, Rampello L, Pignatelli M, Battaglia G, Gradini R, Orzi F, Nicoletti F. 2011 The conditioned eyeblink reflex: a potential tool for the detection of cerebellar dysfunction in multiple sclerosis. Multiple Sclerosis 17 , 1155-1161. (10.1177/1352458511406311)21613334 8. Sommer M, Grafman J, Clark K, Hallett M. 1999 Learning in Parkinson's disease: eyeblink conditioning, declarative learning, and procedural learning. J. Neurol. Neurosurg. Psychiatry 67 , 27-34. (10.1136/jnnp.67.1.27)10369818 9. Gormezano I, Schneiderman N, Deaux E, Fuentes I. 1962 Nictitating membrane: classical conditioning and extinction in the albino rabbit. Science 138 , 33-34. (10.1126/science.138.3536.33)13900561 10. Clark RE, Manns JR, Squire LR. 2001 Trace and delay eyeblink conditioning: contrasting phenomena of declarative and nondeclarative memory. Psychol. Sci. 12 , 304-308. (10.1111/1467-9280.00356)11476097 11. Heiney SA, Wohl MP, Chettih SN, Ruffolo LI, Medina JF. 2014 Cerebellar-dependent expression of motor learning during eyeblink conditioning in head-fixed mice. J. Neurosci. 34 , 14 845-14 853. (10.1523/JNEUROSCI.2820-14.2014) 12. Ten Brinke, MM, Boele HJ, Spanke JK, Potters JW, Kornysheva K, Wulff P, IJpelaar AC, Koekkoek SK, De Zeeuw CI. 2015 Evolving models of pavlovian conditioning: cerebellar cortical dynamics in awake behaving mice. Cell Rep. 13 : 1977-1988. (10.1016/j.celrep.2015.10.057)26655909 13. Heiney SA, Kim J, Augustine GJ, Medina JF. 2014 Precise control of movement kinematics by optogenetic inhibition of Purkinje cell activity. J. Neurosci. Official J. Soc. Neurosci. 34 , 2321-2330. (10.1523/JNEUROSCI.4547-13.2014) 14. Gao Z, Proietti-Onori M, Lin Z, Ten Brinke MM, Boele HJ, Potters JW, Ruigrok TJ, Hoebeek FE, De Zeeuw CI. 2016 Excitatory cerebellar nucleocortical circuit provides internal amplification during associative conditioning. Neuron 89 , 645-657. (10.1016/j.neuron.2016.01.008)26844836 15. Badura A, Verpeut JL, Metzger JW, Pereira TD, Pisano TJ, Deverett B, Bakshinskaya DE, Wang SSH. 2018 Normal cognitive and social development require posterior cerebellar activity. eLife 7 , e36401. (10.7554/eLife.36401)30226467 16. Boele HJ, Koekkoek SK, De Zeeuw CI. 2010 Cerebellar and extracerebellar involvement in mouse eyeblink conditioning: the ACDC model. Front. Cell. Neurosci. 3 , 19. (10.3389/neuro.03.019.2009)20126519 17. Ruigrok TJH. 2011 Ins and outs of cerebellar modules. Cerebellum 10 , 464-474. (10.1007/s12311-010-0164-y)20232190 18. Farley SJ, Radley JJ, Freeman JH. 2016 Amygdala modulation of cerebellar learning. J. Neurosci. Official J. Soc. Neurosci. 36 , 2190-2201. (10.1523/JNEUROSCI.3361-15.2016) 19. Festing MFW. 2010 Inbred strains should replace outbred stocks in toxicology, safety testing, and drug development. Toxicol. Pathol. 38 , 681-690. (10.1177/0192623310373776)20562325 20. Meziane H, Ouagazzal AM, Aubert L, Wietrzych M, Krezel W. 2007 Estrous cycle effects on behavior of C57BL/6 J and BALB/cByJ female mice: implications for phenotyping strategies. Genes Brain Behav. 6 , 192-200. (10.1111/j.1601-183X.2006.00249.x)16827921 21. Simpson J, Kelly JP. 2012 An investigation of whether there are sex differences in certain behavioural and neurochemical parameters in the rat. Behav. Brain Res. 229 , 289-300. (10.1016/j.bbr.2011.12.036)22230114 22. Becker JB, Prendergast BJ, Liang JW. 2016 Female rats are not more variable than male rats: a meta-analysis of neuroscience studies. Biol. Sex Diff. 7 , 34. (10.1186/s13293-016-0087-5) 23. Lowgren K, Baath R, Rasmussen A, Boele HJ, Koekkoek SKE, De Zeeuw CI, Hesslow G. 2017 Performance in eyeblink conditioning is age and sex dependent. PLoS ONE 12 , e0177849. (10.1371/journal.pone.0177849)28542383 24. Wood GE, Shors TJ. 1998 Stress facilitates classical conditioning in males, but impairs classical conditioning in females through activational effects of ovarian hormones. Proc. Natl Acad. Sci. USA 95 , 4066-4071. (10.1073/pnas.95.7.4066)9520494 25. Schreurs BG, Smith-Bell C, Burhans LB. 2018 Sex differences in a rabbit eyeblink conditioning model of PTSD. Neurobiol. Learn. Memory 155 , 519-527. (10.1016/j.nlm.2018.04.015) 26. Rapp AP, Weiss C, Matthew Oh M, Disterhoft JF. 2021 Intact female mice acquire trace eyeblink conditioning faster than male and ovariectomized female mice. eNeuro 8 , 1-9. (10.1523/ENEURO.0199-20.2021) 27. Albergaria C, Tatiana Silva N, Pritchett DL, Carey MR. 2018 Locomotor activity modulates associative learning in mouse cerebellum. Nat. Neurosci. 21 , 725-735. (10.1038/s41593-018-0129-x)29662214 28. Albergaria C, Tatiana Silva N, Darmohray DM, Carey MR. 2020 Cannabinoids modulate associative cerebellar learning via alterations in behavioral state. eLife 9 , e61821. (10.7554/eLife.61821)33077026 29. Warner EJ, Padmanabhan K. 2020 Sex differences in head-fixed voluntary running behavior in C57BL/6 J mice. Europ. J. Neurosci. 51 , 721-730. (10.1111/ejn.14654) 30. Konhilas JP, Maass AH, Luckey SW, Stauffer BL, Olson EN, Leinwand LA. 2004 Sex modifies exercise and cardiac adaptation in mice. Am. J. Physiol. Heart Circ. Physiol. 287 , H2768-H2776. (10.1152/ajpheart.00292.2004)15319208 31. Bettis TJ, Jacobs LF. 2009 Sex-specific strategies in spatial orientation in C57BL/6 J mice. Behav. Process. 82 , 249-255. (10.1016/j.beproc.2009.07.004) 32. Ozden I, Dombeck DA, Hoogland TM, Tank DW, Wang SSH. 2012 Widespread state-dependent shifts in cerebellar activity in locomoting mice. PLoS ONE 7 , e42650. (10.1371/journal.pone.0042650)22880068 33. Sauerbrei BA, Lubenov EV, Siapas AG. 2015 Structured variability in Purkinje cell activity during locomotion. Neuron 87 , 840-852. (10.1016/j.neuron.2015.08.003)26291165 34. Vinck M, Batista-Brito R, Knoblich U, Cardin JA. 2015 Arousal and locomotion make distinct contributions to cortical activity patterns and visual encoding. Neuron 86 , 740-754. (10.1016/j.neuron.2015.03.028)25892300 35. Williamson RS, Hancock KE, Shinn-Cunningham BG, Polley DB. 2015 Locomotion and task demands differentially modulate thalamic audiovisual processing during active search. Curr. Biol. 25 , 1885-1891. (10.1016/j.cub.2015.05.045)26119749 36. Ferreira-Pinto MJ, Ruder L, Capelli P, Arber S. 2018 Connecting circuits for supraspinal control of locomotion. Neuron 100 , 361-374. (10.1016/j.neuron.2018.09.015)30359602 37. Sauvage M, Kitsukawa T, Atucha E. 2019 Single-cell memory trace imaging with immediate-early genes. J. Neurosci. Methods 326 , 108368. (10.1016/j.jneumeth.2019.108368)31356836 38. Barros VN, Mundim M, Galindo LT, Bittencourt S, Porcionatto M, Mello LE. 2015 The pattern of c-Fos expression and its refractory period in the brain of rats and monkeys. Front. Cell. Neurosci. 9 , 72. (10.3389/fncel.2015.00072)25814929 39. Chung L. 2015 A brief introduction to the transduction of neural activity into Fos signal. Dev. Reproduct. 19 , 61-67. (10.12717/DR.2015.19.2.061) 40. Gallo FT, Katche C, Morici JF, Medina JH, Weisstaub NV. 2018 Immediate early genes, memory and psychiatric disorders: focus on c-Fos, Egr1 and Arc. Front. Behav. Neurosci. 12 , 79. (10.3389/fnbeh.2018.00079)29755331 41. Bernstein HL, Lu YL, Botterill JJ, Scharfman HE. 2019 Novelty and novel objects increase c-Fos immunoreactivity in mossy cells in the mouse dentate Gyrus. Neural Plasticity 2019 , 1815371. (10.1155/2019/1815371)31534449 42. Morgan JI, Cohen DR, Hempstead JL, Curran T. 1987 Mapping patterns of c-fos expression in the central nervous system after seizure. Science 237 , 192-197. (10.1126/science.3037702)3037702 43. Nakajima T, Daval JL, Gleiter CH, Deckert J, Post RM, Marangos PJ. 1989 c-Fos mRNA expression following electrical-induced seizure and acute nociceptive stress in mouse brain. Epilepsy Res. 4 , 156-159. (10.1016/0920-1211(89)90020-X)2507311 44. Klein BD, Fu YH, Ptacek LJ, White HS. 2004 c-Fos immunohistochemical mapping of the audiogenic seizure network and tonotopic neuronal hyperexcitability in the inferior colliculus of the Frings mouse. Epilepsy Res. 62 , 13-25. (10.1016/j.eplepsyres.2004.06.007)15519128 45. Strekalova T, Zörner B, Zacher C, Sadovska G, Herdegen T, Gass P. 2003 Memory retrieval after contextual fear conditioning induces c-Fos and JunB expression in CA1 hippocampus. Genes Brain Behav. 2 , 3-10. (10.1034/j.1601-183X.2003.00001.x)12882314 46. Frances Davies M, Tsui J, Flannery JA, Li X, DeLorey TM, Hoffman BB. 2003 Activation of α2 adrenergic receptors suppresses fear conditioning: expression of c-Fos and phosphorylated CREB in mouse amygdala. Neuropsychopharmacology 29 , 229-239. (10.1038/sj.npp.1300324) 47. Martinez M, Calvo-Torrent A, Herbert J. 2002 Mapping brain response to social stress in rodents with c-fos expression: a review. Stress 5 , 3-13. (10.1080/102538902900012369)12171762 48. Haller J, Tóth M, Halasz J, De Boer SF. 2006 Patterns of violent aggression-induced brain c-fos expression in male mice selected for aggressiveness. Physiol. Behav. 88 , 173-182. (10.1016/j.physbeh.2006.03.030)16687160 49. Ago Y, Araki R, Tanaka T, Sasaga A, Nishiyama S, Takuma K, Matsuda T. 2013 Role of social encounter-induced activation of prefrontal serotonergic systems in the abnormal behaviors of isolation-reared mice. Neuropsychopharmacology 38 , 1535-1547. (10.1038/npp.2013.52)23426384 50. Zhong J, Liang M, Akther S, Higashida C, Tsuji T, Higashida H. 2014 c-Fos expression in the paternal mouse brain induced by communicative interaction with maternal mates. Mol. Brain 7 , 66. (10.1186/s13041-014-0066-x)25208928 51. Hudson AE. 2018 Genetic reporters of neuronal activity: c-Fos and G-CaMP6. Methods Enzymol. 603 , 197-220. (10.1016/bs.mie.2018.01.023)29673526 52. Lara Aparicio SY et al. 2022 Current opinion on the use of c-Fos in neuroscience. NeuroSci 3 , 687-702. (10.3390/neurosci3040050) 53. Bracha V, Zbarska S, Parker K, Carrel A, Zenitsky G, Bloedel JR. 2009 The cerebellum and eye-blink conditioning: learning versus network performance hypotheses. Neuroscience 162 , 787-796. (10.1016/j.neuroscience.2008.12.042)19162131 54. Bao S, Chen L, Kim JJ, Thompson RF. 2002 Cerebellar cortical inhibition and classical eyeblink conditioning. Proc. Natl Acad. Sci. USA 99 , 1592-1597. (10.1073/pnas.032655399)11805298 55. Kellett DO, Fukunaga I, Chen-Kubota E, Dean P, Yeo CH. 2010 Memory consolidation in the cerebellar cortex. PLoS ONE 5 , e11737. (10.1371/journal.pone.0011737)20686596 56. Mathis A, Mamidanna P, Cury KM, Abe T, Murthy VN, Mathis MW, Bethge M. 2018 DeepLabCut: markerless pose estimation of user-defined body parts with deep learning. Nat. Neurosci. 21 , 1281-1289. (10.1038/s41593-018-0209-y)30127430 57. D'Angelo E, Galliano E, De Zeeuw CI. 2016 Editorial: the olivo-cerebellar system. Front. Neural Circuits 9 , 66. (10.3389/fncir.2015.00066)26793067 58. Kawai R, Markman T, Poddar R, Ko R, Fantana AL, Dhawale AK, Kampff AR, Ölveczky BP. 2015 Motor cortex is required for learning but not for executing a motor skill. Neuron 86 , 800-812. (10.1016/j.neuron.2015.03.024)25892304 59. Åhlgren J, Voikar V. 2019 Experiments done in black-6 mice: what does it mean? Lab Anim. 48 , 171-180. (10.1038/s41684-019-0288-8) 60. Erway LC, Shiau YW, Davis RR, Krieg EF. 1996 Genetics of age-related hearing loss in mice. III. Susceptibility of inbred and F1 hybrid strains to noise-induced hearing loss. Hearing Res. 93 , 181-187. (10.1016/0378-5955(95)00226-X) 61. Milon B, Mitra S, Song Y, Margulies Z, Casserly R, Drake V, Mong JA, Depireux DA, Hertzano R. 2018 The impact of biological sex on the response to noise and otoprotective therapies against acoustic injury in mice. Biol. Sex Diff. 9 , 12. (10.1186/s13293-018-0171-0) 62. Ohlemiller KK. 2019 Mouse methods and models for studies in hearing. J. Acoust. Soc. Am. 146 , 3668. (10.1121/1.5132550)31795658 63. Yang CF, Shah NM. 2014 Representing sex in the brain, one module at a time. Neuron 82 , 261-278. (10.1016/j.neuron.2014.03.029)24742456 64. Lightfoot J, Timothy MJ, Turner MD, Vordermark A, Kleeberger SR. 2004 Genetic influence on daily wheel running activity level. Physiol. Genomics 19 , 270-276. (10.1152/physiolgenomics.00125.2004)15383638 65. Goh J, Ladiges W. 2015 Voluntary wheel running in mice. Curr. Protocols Mouse Biol. 5 , 283-290. (10.1002/9780470942390.mo140295) 66. Safari S, Ahmadi N, Mohammadkhani R, Ghahremani R, Khajvand-Abedeni M, Shahidi S, Komaki A, Salehi I, Karimi SA. 2021 Sex differences in spatial learning and memory and hippocampal long-term potentiation at perforant pathway-dentate Gyrus (PP-DG) synapses in wistar rats. Behav. Brain Funct. 17 , 9. (10.1186/s12993-021-00184-y)34724971 67. Ramsey LC, Pittenger C. 2010 Cued and spatial learning in the water maze: equivalent learning in male and female mice. Neurosci. Lett. 483 , 148-151. (10.1016/j.neulet.2010.07.082)20691760 68. Fritz AK, Amrein I, Wolfer DP. 2017 Similar reliability and equivalent performance of female and male mice in the open field and water-maze place navigation task. Am. J. Med. Genet. Part C Seminars Med. Genet. 175 , 380-391. (10.1002/ajmg.c.31565) 69. Dalla C, Shors TJ. 2009 Sex differences in learning processes of classical and operant conditioning. Physiol. Behav. 97 , 229-238. (10.1016/j.physbeh.2009.02.035)19272397 70. Gruene TM, Flick K, Stefano A, Shea SD, Shansky RM. 2015 Sexually divergent expression of active and passive conditioned fear responses in rats. eLife 4 , e11352. (10.7554/eLife.11352)26568307 71. Magurran AE, Garcia CM. 2000 Sex differences in behaviour as an indirect consequence of mating system. J. Fish Biol. 57 , 839-857. (10.1111/j.1095-8649.2000.tb02196.x) 72. Lucon-Xiccato T, Bisazza A. 2014 Discrimination reversal learning reveals greater female behavioural flexibility in guppies. Biol. Lett. 10 , 20140206. (10.1098/rsbl.2014.0206) 73. McDowell AA, Brown WL, McTee AC. 1960 Sex as a factor in spatial delayed-response performance by rhesus monkeys. J. Comp. Physiol. Psychol. 53 , 429-432. (10.1037/h0043372) 74. Lonsdorf EV, Eberly LE, Pusey AE. 2004 Sex differences in learning in chimpanzees. Nature 428 , 715-716. (10.1038/428715a)15085121 75. Stark R, Wolf OT, Tabbert K, Kagerer S, Zimmermann M, Kirsch P, Schienle A, Vaitl D. 2006 Influence of the stress hormone cortisol on fear conditioning in humans: evidence for sex differences in the response of the prefrontal cortex. NeuroImage 32 , 1290-1298. (10.1016/j.neuroimage.2006.05.046)16839780 76. Pawlowski B, Atwal R, Dunbar RIM. 2008 Sex differences in everyday risk-taking behavior in humans. Evol. Psychol. 6 , 147470490800600104. 77. Hines M. 2010 Sex-related variation in human behavior and the brain. Trends Cogn. Sci. 14 , 448-456. (10.1016/j.tics.2010.07.005)20724210 78. Tschernegg M, Neuper C, Schmidt R, Wood G, Kronbichler M, Fazekas F, Enzinger C, Koini M. 2017 FMRI to probe sex-related differences in brain function with multitasking. PLoS ONE 12 , e0181554. (10.1371/journal.pone.0181554)28759619 79. Andreescu CE, Milojkovic BA, Haasdijk ED, Kramer P, De Jong FH, Krust A, De Zeeuw CI, De Jeu MTG. 2007 Estradiol improves cerebellar memory formation by activating estrogen receptor β. J. Neurosci. 27 , 10 832-10 839. (10.1523/JNEUROSCI.2588-07.2007) 80. McFarland WL, Teitelbaum H, Hedges EK. 1975 Relationship between hippocampal theta activity and running speed in the rat. J. Comp. Physiol. Psychol. 88 , 324-328. (10.1037/h0076177)1120805 81. Bender F, Gorbati M, Cadavieco MC, Denisova N, Gao X, Holman C, Korotkova T, Ponomarenko A. 2015 Theta oscillations regulate the speed of locomotion via a hippocampus to lateral septum pathway. Nat. Commun. 6 , 8521. (10.1038/ncomms9521)26455912 82. Nuñez A, Buño W. 2021 The theta rhythm of the hippocampus: from neuronal and circuit mechanisms to behavior. Front. Cell. Neurosci. 15 , 649262. (10.3389/fncel.2021.649262)33746716 83. Shin J, Kim D, Bianchi R, Wong RKS, Shin HS. 2005 Genetic dissection of theta rhythm heterogeneity in mice. Proc. Natl Acad. Sci. USA 102 , 18 165-18 170. (10.1073/pnas.0505498102) 84. Tort ABL, Komorowski RW, Manns JR, Kopell NJ, Eichenbaum H. 2009 Theta-gamma coupling increases during the learning of item-context associations. Proc. Natl Acad. Sci. USA 106 , 20 942-20 947. (10.1073/pnas.0911331106) 85. Tendler A, Wagner S. 2015 Different types of theta rhythmicity are induced by social and fearful stimuli in a network associated with social memory. eLife 4 , e03614. (10.7554/eLife.03614)25686218 86. Begus K, Bonawitz E. 2020 The rhythm of learning: theta oscillations as an index of active learning in infancy. Dev. Cogn. Neurosci. 45 , 100810. (10.1016/j.dcn.2020.100810)33040970 87. Guderian S, Schott BH, Richardson-Klavehn A, Düzel E. 2009 Medial Temporal theta state before an event predicts episodic encoding success in humans. Proc. Natl Acad. Sci. USA 106 , 5365-5370. (10.1073/pnas.0900289106)19289818 88. Joëls M. 2009 Stress, the hippocampus, and epilepsy. Epilepsia 50 , 586-597. (10.1111/j.1528-1167.2008.01902.x)19054412 89. Schneider ER, Civillico EF, Wang SS. 2013 Calcium-based dendritic excitability and its regulation in the deep cerebellar nuclei. J. Neurophysiol. 109 , 2282-2292. (10.1152/jn.00925.2012)23427305 90. Sibold JS, Hammack SE, Falls WA. 2011 C57 mice increase wheel-running behavior following stress: preliminary findings. Percept. Motor Skills 113 , 605-618. (10.2466/06.16.20.PMS.113.5.605-618)22185076 91. Malisch JL, deWolski K, Meek TH, Acosta W, Middleton KM, Crino OL, Garland Jr T. 2016 Acute restraint stress alters wheel-running behavior immediately following stress and up to 20 h later in house mice. Physiol. Biochem. Zool. 89 , 546-552. (10.1086/688660)27792529 92. Juczewski K, Koussa JA, Kesner AJ, Lee JO, Lovinger DM. 2020 Stress and behavioral correlates in the head-fixed method: stress measurements, habituation dynamics, locomotion, and motor-skill learning in mice. Scientific Rep. 10 , 12245. (10.1038/s41598-020-69132-6) 93. DeVallance E, Riggs D, Jackson B, Parkulo T, Zaslau S, Chantler PD, Olfert IM, Bryner RW. 2017 Effect of chronic stress on running wheel activity in mice. PLoS ONE 12 , e0184829. (10.1371/journal.pone.0184829)28926614 94. Lyons DM, Buckmaster CL, Schatzberg AF. 2018 Learning to actively cope with stress in female mice. Psychoneuroendocrinology 96 , 78-83. (10.1016/j.psyneuen.2018.06.010)29909293 95. Furman O, Tsoory M, Chen A. 2022 Differential chronic social stress models in male and female mice. Europ. J. Neurosci. 55 , 2777-2793. (10.1111/ejn.15481) 96. Gao Z, van Beugen BJ, De Zeeuw CI. 2012 Distributed synergistic plasticity and cerebellar learning. Nat. Rev. 13 , 619-635. (10.1038/nrn3312) 97. Giovannucci A et al. 2017 Cerebellar granule cells acquire a widespread predictive feedback signal during motor learning. Nat. Neurosci. 20 , 727-734. (10.1038/nn.4531)28319608 98. Knogler LD, Markov DA, Dragomir EI, Štih V, Portugues R. 2017 Sensorimotor representations in cerebellar granule cells in larval zebrafish are dense, spatially organized and nontemporally patterned. Curr. Biol. 27 , 1288-1302.28434864 99. Wagner MJ, Kim TH, Savall J, Schnitzer MJ, Luo L. 2017 Cerebellar granule cells encode the expectation of reward. Nature 544 , 96-100. (10.1038/nature21726)28321129 100. Gandolfi D et al. 2017 Activation of the CREB/c-Fos pathway during long-term synaptic plasticity in the cerebellum granular layer. Front. Cell. Neurosci. 11 , 184. (10.3389/fncel.2017.00184)28701927 101. Hwang EJ, Dahlen JE, Mukundan M, Komiyama T. 2021 Disengagement of motor cortex during long-term learning tracks the performance level of learned movements. J. Neurosci. 41 , 7029-7047. (10.1523/JNEUROSCI.3049-20.2021)34244359 102. Cui G, Jun SB, Jin X, Luo G, Pham MD, Lovinger DM, Vogel SS, Costa RM. 2014 Deep brain optical measurements of cell type-specific neural activity in behaving mice. Nat. Protocols 9 , 1213-1228. (10.1038/nprot.2014.080)24784819 103. Jun JJ et al. 2017 Fully integrated silicon probes for high-density recording of neural activity. Nature 551 , 232-236. (10.1038/nature24636)29120427 104. Insafutdinov E, Pishchulin L, Andres B, Andriluka M, Schiele B. 2016 DeeperCut: a deeper, stronger, and faster multi-person pose estimation model. In Computer vision: ECCV 2016 (eds B Leibe, J Matas, N Sebe, M Welling), pp. 34-50. Berlin, Germany: Springer. 105. He K, Zhang X, Ren S, Sun J. 2016 Deep residual learning for image recognition. In 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , Las Vegas, NV, pp. 770-778. (10.1109/CVPR.2016.90) 106. Ellis T, Gambardella L, Horcher M, Tschanz S, Capol J, Bertram P, Jochum W, Barrandon Y, Busslinger M. 2001 The transcriptional repressor CDP (Cutl1) is essential for epithelial cell differentiation of the lung and the hair follicle. Genes Dev. 15 , 2307-2319. (10.1101/gad.200101)11544187 107. Shamash P, Carandini M, Harris K, Steinmetz N. 2018 A tool for analyzing electrode tracks from slice histology. bioRxiv 447995. (10.1101/447995) 108. Schindelin J et al. 2012 Fiji: an open-source platform for biological-image analysis. Nat. Methods 9 , 676-682. (10.1038/nmeth.2019)22743772 109. Puth MT, Neuhäuser M, Ruxton GD. 2015 Effective use of Spearman's and Kendall's correlation coefficients for association between two measured traits. Anim. Behav. 102 , 77-84. (10.1016/j.anbehav.2015.01.010) 110. Oyaga MR, Serra I, Kurup D, Koekkoek SKE, Badura A. 2023 Delay eyeblink conditioning performance and brain-wide c-Fos expression in male and female mice. Figshare. (10.6084/m9.figshare.c.6619748)
PMC010xxxxxx/PMC10182933.txt
==== Front Environ Sci Pollut Res Int Environ Sci Pollut Res Int Environmental Science and Pollution Research International 0944-1344 1614-7499 Springer Berlin Heidelberg Berlin/Heidelberg 37079234 27052 10.1007/s11356-023-27052-8 Research Article Particle and gas phase sampling of PCDD/Fs and dl-PCBs by activated carbon fiber and GC/MS analysis Cerasa Marina 1 Guerriero Ettore 1 Balducci Catia 1 Bacaloni Alessandro 2 Ciccioli Piero 3 Mosca Silvia [email protected] 1 1 grid.494655.f Italian National Research Council, Institute of Atmospheric Pollution Research, Area Della Ricerca Di Roma 1, 00010 Montelibretti (RM), Italy 2 grid.7841.a Chemistry Department, Mathematics, Physics and Natural Sciences Faculty, Sapienza University Piazzale Aldo Moro, 5, 00185 Rome, Italy 3 grid.518160.b Italian National Research Council, Institute for Biological Systems, Area Della Ricerca Di Roma 1, 00010 Montelibretti (RM), Italy Responsible Editor: Constantini Samara 20 4 2023 20 4 2023 2023 30 24 6519265203 23 9 2022 12 4 2023 © The Author(s) 2023 https://creativecommons.org/licenses/by/4.0/ Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Polychlorodibenzo-p-dioxins (PCDDs), polychlorodibenzofurans (PCDFs), and polychlorobiphenyls (PCBs) are semi-volatile compounds and can be partitioned in the atmosphere between the gas and particulate phase, due to their physicochemical properties. For this reason, the reference standard methods for air sampling include a quartz fiber filter (QFF) for the particulate and a polyurethane foam (PUF) cartridge for the vapor phase, and it is the classical and most popular sampling method in the air. Despite the presence of the two adsorbing media, this method cannot be used for the study of the gas-particulate distribution, but only for a total quantification. This study presents the results and the performance aim to validate an activated carbon fiber (ACF) filter for the sampling of PCDD/Fs and dioxin-like PCBs (dl-PCBs) using laboratory and field tests. The specificity, precision, and accuracy of the ACF in relation to the QFF + PUF were evaluated through the isotopic dilution technique, the recovery rates, and the standard deviations. Then the ACF performance was assessed on real samples, in a naturally contaminated area, through parallel sampling with the reference method (QFF + PUF). The QA/QC was defined according to the standard methods ISO 16000–13 and -14 and EPA TO4A and 9A. Data confirmed that ACF meets the requirements for the quantification of native POPs compounds in atmospheric and indoor samples. In addition, ACF provided accuracy and precision comparable to those offered by standard reference methods using QFF + PUF, but with significant savings in terms of time and costs. Supplementary Information The online version contains supplementary material available at 10.1007/s11356-023-27052-8. Keywords Air sampling Gas/particle partitioning PCDD/Fs Dl-PCBs Activated carbon fiber (ACF) Validation GC/MS Ambient air issue-copyright-statement© Springer-Verlag GmbH Germany, part of Springer Nature 2023 ==== Body pmcIntroduction Persistent organic pollutants (POPs) are ubiquitous contaminants frequently found in sediments, soil, fish, wildlife, human adipose tissue, serum, and milk (Hart and Pankow 1994; Lee and Jones 1999; Rodan et al. 1999; Bergknut et al. 2011). With respect to other environmental compartments, the atmospheric burden of POPs is relatively small, but the air is considered the most important vehicle for their global redistribution, especially considering the low solubility in water (Piazza et al. 2013). POPs have common physicochemical features such as resistance to chemicals and to biodegradation, high lipophilicity, and therefore, a tendency to bioaccumulate in adipose tissue, moreover, exhibits toxic effects on humans and wildlife (Rodan et al. 1999; Bergknut et al. 2011). As semi-volatile organic compounds, POPs are in the atmospheric environments in equilibrium in both the gaseous and particulate phases for temperatures above 0 °C (Lei and Wania 2004). In particulate matter, they are linked to the solid matrix by physical and chemical bonds (Hippelein and McLachlan 2000; Larsson et al. 2013; Wang et al. 2021). In this study, two classes of POPs are considered: dioxin-like polychlorinated biphenyls (dl-PCBs) and polychlorinated dibenzo-p-dioxins/polychlorinated dibenzofurans (PCDD/Fs). Due to the multiple equilibria that these POPs can have, ISO/DIS (ISO 2007a, b) and US-EPA (EPA 1999a, b) methods for the determination of PCDD/Fs and dl-PCBs in the atmosphere and indoor air sampling require a quartz fiber filter (QFF) to collect particle-bound contaminants, followed by a cartridge filled with a solid sorbent, usually polyurethane foam (PUF) or styrene–divinylbenzene polymer (i.e., XAD-2 resin) to collect the vapor phase (Kaupp and Umlauf 1992; Król et al. 2011; Degrendele et al. 2020; Wu et al. 2020; López et al. 2021). Despite the double adsorbent media, this sampling system cannot be used for a reliable estimation of the gas-particle partition of PCDD/Fs and dl-PCBs because it is subject to a number of several sampling artifacts. In particular, there may be an over or underestimation of POPs in particulate matter or gaseous fraction: (i) the gas phase compounds adsorbed on the particulate matter could be stripped from the QFF to the PUF cartridge enriching the gaseous fraction; (ii) the particulate matter on the QFF can adsorb some of the gaseous compounds during sampling. Very common is the loss of part of the samples due to a sampled volume greater than the breakthrough volume. To prevent this problem, backup filters consisting of PUF (Degrendele et al. 2020), a combination of XAD-2/PUF (López et al. 2021), or XAD-2 (Wu et al. 2020) have been added to the line of the sampling train. Although some authors have used data obtained with multiple sampling trains to estimate the amount of PCDD/Fs and/or dl-PCBs in the gas and particulate phase in it, as already mentioned, the error committed in this procedure is high. These system sets can only be used for a total estimation of the compounds in both phases (Kaupp and Umlauf 1992; Hart and Pankow 1994; Lee and Jones 1999; Barbas et al. 2018; Wu et al. 2020). A more correct evaluation of the gas-particle distribution of these analytes was carried out using a denuder upstream of the adsorption train for the gaseous phase only and a second adsorbent (or a mixture of adsorbents) downstream for collecting the particulate (Kaupp and Umlauf 1992; Forbes 2020). In this work, an activated carbon fiber (ACF) is proposed as a suitable single adsorbent for the total collection (both vapor and particle phase) of PCDD/Fs and dl-PCBs in atmospheric and indoor samples, meeting the requirement of international standard methods ISO and EPA. Activated carbon fibers or fabrics (ACFs) are considered an advanced group of porous materials with many advantages over granular or powder-activated carbons. ACFs have an extremely high specific surface area (SSA) characterized by a uniform micropores distribution that is directly exposed to the surface (Lordgooei et al. 2001). An ACF felt already validated as a passive sampler for PCDD/Fs and PCBs in the aqueous matrix (Cerasa et al. 2020) was used, characterized by a high specific surface area (SSA) and microporosity distribution. The felt has a sufficient thickness and mechanical strength to fully retain fine atmospheric particles at the sampling rates normally used for their collection with a high-volume sampler while maintaining almost zero impedance. To date, the ACF has already been used for the sampling of these classes of compounds in the air only as a backup filter, in the queue of the QFF + PUF train (Anezaki and Yamaguchi 2011; Anezaki and Kashiwagi 2021), but not for sampling. In this work, the tests that led to the validation of ACF as the only adsorbent for the evaluation of total PCDD/Fs and dl-PCBs in the atmosphere are presented. First, laboratory tests were carried out, evaluating the sampling efficiency for the gas phase with an ad hoc sampling train. Subsequently, the efficiency of the total sampling and the matrix effect was evaluated through real sampling. The method proposed with the ACF was compared with the reference method QFF + PUF. The tests were performed using isotopically labelled standard solutions, through which the R% was evaluated to consider the precision, repeatability, and selectivity of the method. The validation of ACF as an adsorbent material takes the requirements defined by ISO 16000–13 e 14 and EPA TO-4A and TO-9A as QA/QC parameters. A unified adsorbent method could allow a considerable saving in time and solvent consumption, not to be underestimated a simplification in analysis (sampling, extraction, and clean-up procedures). Material and methods Standards and solvents All 13C-labelled standards of PCDD/Fs (EN1948-ES, EN1948-SS, and EN1948-IS) and dl-PCBs (WP-LCS, P48-SS, and WP-ISS) were purchased from Wellington Laboratories, Canada, (Tables S2 and S3, Supplementary Information). For all tests, three solutions of standard 13C12 (10 pg/µl) containing PCDD/Fs and dl-PCBs congeners were used, combining the previous solutions depending on the tests performed. They are distinguished according to the order of addition in the sample: (i) standard sampling solution (SS solution) added to the adsorbent medium before starting the sampling, (ii) extraction standard solution (ES solution) added to the adsorbents after sampling and before extraction, (iii) internal standard solution (IS solution) added before injection and used for quantification of native compounds and recovery rates (%R) of SS and ES solutions. The composition of the SS, ES, and IS solutions is specified later in the sections corresponding to each test. The GC/MS calibration was performed by the isotopic dilution method, using commercially available calibration curves: EN1948-CVS for PCDD/Fs and P48-W-CVS for dl-PCBs (Wellington Laboratories, Canada). Acetone, toluene, dichloromethane, and hexane used in chemical analysis were purchased from Romil. Activated carbon fiber (ACF) The physico-chemical characterization of ACF (Chemical Research 2000 Srl, Italy) used in this work was already described in a previous study (Cerasa et al. 2020). Briefly, the Brunauner-Emmet-Teller (BET) method and the Langmuir equation were used to define an SSA of about 2500 m2/g with a pores diameter of ~ 1.2 nm (microporosity). The Boehm titration analysis yielded a strong acidic and basic component due to carboxyl and pyrone groups, respectively. All organics present on the ACF either due to material production or prolonged exposure to polluted atmospheres were removed by Soxhlet extraction with toluene for 24 h. The ACF was vacuum-dried at 40 °C prior to the use. It was cut into discs with a diameter of 58 and 102 mm, depending on the experiment. Sampling and collection trains used High-volume samplers from TCR Tecora (Cogliate, Italy) with a PM10 cutting sampling head were used in all experiments. The sampling head includes a 102-mm grid holder to house a filter and a cylindrical glass cartridge holder (58 mm × 125 mm long) to house a polyurethane foam (PUF, density 0.022 g/cm3) plug. For all tests, a flow rate of 200 L/min was adopted and maintained constant based on temperature and atmospheric pressure through an electronic system. Figure 1 shows a scheme of the sampling train set-up, as described in detail in the following sections. Set-up A is the reference method, consisting of a QFF and a PUF. In Set-up B, two 58-mm ACF filters (3A and 3B) were placed in the glass cartridge, above the PUF. In Set-up C, the 102-mm QFF was directly replaced with a 102-mm ACF. Before sampling, the QFF was baked in an oven at 400 °C for 5 h, and the PUF was washed by Soxhlet extraction for 24 h with acetone: ethyl acetate (50:50 v/v). ACF was precleaned by 24 h-Soxhlet extraction with toluene and vacuum-dried at 40 °C prior the use.Fig. 1 Sampling trains for the collection of PCDD/Fs and dl-PCBs used in this work. Set-up A: sampling train compliant with ISO and US-EPA standard methods for the collection of PCDD/Fs and dl-PCBs in the air (reference sampling method). 1, QFF (102 mm); 2, PUF; Set-up B: sampling train for ACF breakthrough test of PCDD/Fs and dl-PCBs. 1, QFF (102 mm); 3A, ACF (58 mm); 3B, ACF (58 mm); 2, PUF; Set-up C: sampling train to validate the use of ACF as a single sorbent for the determination of PCDD/Fs and dl-PCBs in indoor and ambient air samples; 3, ACF (102 mm); 2, PUF Reference method The reference method (Fig. 1, Set-up A) meets the requirements of both ISO (indoor air) and EPA (ambient air) standard methods for PCDD/Fs and dl-PCBs samplings (Table S1, Supplementary material). The sampling head includes a 102-mm QFF for the collection of particles and a PUF to collect the gaseous fraction. The SS solution (EN-1948SS and P48-SS, 100 µl; Tables S2 and S3, Supplementary material) was added to the QFF. After sampling, the QFF and the PUF were taken to the laboratory and extracted together in a cellulose thimble with ~ 3 g of Na2SO4, after adding the ES solution containing EN1948ES and WP-LCS (100 µl) (Tables S2 and S3, Supplementary material). The extraction was performed in 250-mL Soxhlet for 36 h with toluene. The extract was first concentrated with a rotary evaporator (40 ± 2 °C and 49 mbar) up to 10 ml and then with a gentle flow of N2 in a water bath (40 ± 2 °C) up to 1 ml. The clean-up involved a multilayer silica column (extract eluted with hexane) and an alumina microcolumn, to separate the PCDD/Fs from the dl-PCBs, as described in Mosca et al. (2010). The two fractions of the eluates were concentrated, and the corresponding IS solutions were added (EN 1948IS for PCDD/Fs and WP-ISS for dl-PCBs; Tables S2 and S3, Supplementary material). Instrumental analysis was performed using a triple quadrupole gas chromatograph/mass spectrometer (Trace 1310 GC/TSQ 8000 Evo, Thermo), and chromatographic separations were achieved using a DB-XLB column (60 m, 0.25 mm, 0,25 mm I.D., Agilent J&W) (Benedetti et al. 2017). Evaluation of ACF breakthrough A preliminary survey was conducted to investigate whether and to which extent PCDD/Fs and dl-PCBs can be retained in the vapor phase on the ACF sorbent at different sampled volumes, using 13C-labelled compounds as tracers, and the ISO breakthrough limits as a reference (ISO 2007a, b). The sampling train used for these tests consisted of a 102 mm QFF, two 58 mm ACFs (3A and 3B), and a PUF (Fig. 1, Set-up B). The QFF was spiked with a known amount of a mixture containing labelled congeners of both PCDD/Fs and dl-PCBs, used as a SS solution (1000–2000 pg of EN1948ES and 1000 pg of WP-LCS; Tables S2 and S3, Supplementary material), in order to simulate a real atmosphere sampling (ISO 2007b; Cerasa et al. 2020). The samplings were performed at different time extensions 24 h (288 m3), 3 days (864 m3), and 1 week (2016 m3), between March and May 2016 (detailed information in Table S4, Supplementary Material) in triplicate at the “A. Liberti” monitoring station of Montelibretti (Rome, Italy, located in the National Research Council of Italy (CNR)) classified as semiurban area, where the concentrations of native PCDD/Fs and dl-PCBs are usually below the limits of detection (LOD). After sampling, the QFF, the ACFs (3A and 3B), and PUF adsorbents were separately extracted with toluene in a Soxhlet apparatus for 36 h, once spiked each of them with the ES solution, containing1000 pg of P48-SS and 1000–2000 pg of EN-1948SS (Tables S2 and S3, Supplementary material). The efficiency of the extraction of these classes of POPs from the ACF has already been investigated in previous studies (Cerasa et al. 2021). Separated fractions of PCDD/Fs and dl-PCBs were obtained by using the clean-up procedure described in the previous subsection. They were all fortified with IS solutions (1000 pg of WP-ISS and 1000 pg of EN 1948 IS, for dl-PCBs and PCDD/Fs, respectively. Tables S2 and S3, Supplementary material) before the GC–MS determinations. ACF as a single sorbent The suitability of ACF as a single absorbent for the determination of the total content (vapor and particulate phase) of PCDD/Fs and dl-PCBs in the air was assessed by seven parallel samplings collected between May and June 2016, sampling lasting from 24 to 168 h for total volumes between 480 m3 and 824 m3, in a very large indoor public area (> 50,000 m3) where a serious fire occurred causing emissions of black smoke particles, presumably enriched with PCDD/Fs and dl-PCBs, due to the presence of electric material (Colapicchioni et al. 2020). The samples were collected in parallel on 102 mm QFF + PUF (Fig. 1 Set-up A), used as a reference according to the ISO/DIS standard methods (ISO 2007a, b) and US- EPA (EPA 1999c, b), and on 102-mm ACF+PUF (Fig. 1 Set-up C). In this sample train (ACF+PUF), the PUF acts as a backup filter for ACF, to verify the absence of a breakthrough in a contaminated atmosphere (matrix effect). For this reason, ACF and PUF were extracted separately. Before sampling, the SS solution was added to the 102-mm ACF, then ACF and PUF were spiked with the ES Solution and extracted separately. The samples were then purified and the IS solution was added for GC–MS analysis. Standards and quantities added, purification method, and GC–MS analysis are the same as reported in the reference method subsection. Data concentrations of each congener of PCDD/Fs and dl-PCBs, expressed in fg TEQ/Nm3, were compared. Quality assurance/quality control The validation of the proposed method based on ACF was carried out using the parameters defined by the standardized methods ISO 16000 13 and 14 and EPA TO 4A and 9A as QA/QC (Table 1).Table 1 QA/QC acceptance criteria and requirements of ISO and EPA reference methods. % RSS, recovery rates of SS; % RES, recovery rate of ES; 1 if the %RSS is < 50 or > 150, the sampling is invalid % RSS % RES EPA TO 9A (PCDD/Fs) 50–120 50–120 (TCDD/F, PeCDD/F, HxCDD/F) 40–120 (HpCDD/F, OCDD/F) EPA TO 4A (dl-PCBs) - 60–120 (PCB) ISO/DIS 16000–13 (50) 75–125 (150)1 ISO/DIS 16000–14 n.a 50–130 (TCDD/F, PeCDD/F, HxCDD/F) 40–130 (HpCDD/F, OCDD/F) 40–120 (PCB) Since the recovery rate ranges imposed by EPA methods are stricter, they were taken as the QC acceptance criteria. The accuracy achieved for duplicates must be ± 30% (EPA 1999b). Furthermore, the breakthrough of the original sampling train shall be less than 10% for every single congener (ISO 2007a, 2007b). All tests involved the use of isotopically labelled standards during all steps (the SS solution in the sampling step, the ES solution in the extraction step, and the IS solution before injection). The recoveries evaluated for all the analytical phases, allow us to interpret the losses of the compounds during each step and ensure the selectivity of the method. The accuracy was estimated as the mean recovery rate of each labelled compound of the standard solutions (%RSS) spiked on the samples and the relative standard deviation (RSD%). The sampling efficiency is the accuracy during the sampling step evaluated through the recovery rate of the SS solution (%RSS) added before the sampling (Eq. 1).1 %RSS=RRF100∗ASSAES∗QESQSS The relative response factor (RRF) is the response of the mass spectrometer to a known amount of an analyte relative to a known amount of a 13C-labelled internal standard calculated through the calibration kit. ASS and AES are the sums of the integrated ion abundances of the quantitation ions for 13C-labelled SS and ES solution compounds; QSS and QES are the quantities of the 13C-labelled SS and ES solutions injected. Furthermore, the accuracy of the ACF method was evaluated in relation to the reference method (QFF + PUF) by both considering the %Rs of all 7 parallels and comparing the quantitative analyses of each pair of congeners for a single sample. The standard deviation of the %R of the triplicates for the laboratory tests and the seven samples served as a measure of the method’s precision. Linearity is evaluated through tests to verify the ability of the ACF to adsorb the gas phase: sampling was carried out with progressively increased volumes of air while still using the same amount of spiked standards. The 7 real samplings performed in a heavily polluted area affected by a fire are accounted for the matrix effect, which is assessed using the average %R and RSD%. A standard mixture of isotopically labelled PCDD/Fs and dl-PCBs was injected repeatedly throughout the batch to test the stability of the analytical instrument, and solvent blank injections (nonane) for GC analysis were used to track potential carry-over and memory effects. These procedures were done to investigate the instrument’s precision. All tests included a laboratory and a field blank for real samples. Generally, LODs and LOQs were calculated to check the sensitivity of the developed method for target compounds. LODs and LOQs were defined as 3 and 10 times the signal-to-noise ratio (S/N) under the lowest spiked concentration of the calibration curve, respectively. Results and discussion Peters et al. (2000) have shown that release of POPS occurs from particles during sampling, as a function of the partial vapor pressure of the specific POP and the volume sampled. According to Peters, these vapors are transferred to the solid adsorbent placed after the particle filter, where they should be retained. Since the retention of PCDD/Fs and dl-PCBs on the QFF + PUF sampling train depends on their concentration in air, the environmental conditions in which sampling is performed, and the total volume sampled, adequate quality control and quality assurance criteria (QA/QC) were defined to attest that an accurate determination in the air is achieved. It can happen, as one of the most common reasons, when the sampled volume exceeds the breakthrough volume on the PUF adsorbent, and hence, part of the sample is lost during sampling. The following subsection argues the cited QA/QC criteria. Retention of PCDD/Fs and dl-PCBs in the vapor phase on the ACF First of all, the retention of PCDD/Fs and dl-PCBs congeners on the ACF was investigated. The sampling train used consisted of the following: 102-mm QFF; two-58 mm AFCs (3A and 3B) and a PUF (Fig. 1, Set-up B), collecting samples at three different volumes, up to 2016 m3 and by spiking the QFF with appropriate amounts of a SS solution, as described in the corresponding section. The spiking approach is simple and provides results that can be safely extrapolated to a real atmospheric sampling, because the retention volume measured is equal to or smaller than that measured under normal atmospheric sampling conditions. Since the most volatile fraction of the SS solution is rapidly stripped from the QFF, a nearly instantaneous transfer to the ACF 3A adsorbent occurs as soon as the aspirating pump is activated. This effect does not normally occur under atmospheric sampling conditions because the stripping of semi-volatile POPs from particles retained on the QFF is much slower, and larger volumes are required to let POPs vapor them to pass through the ACF 3A adsorbent. Increasing volumes were sampled to check the linearity and to see if the breakthrough volume of PCCD/Fs and dl-PCBs congeners was ever reached on the ACF 3A adsorbent in 168-h samples. Since the atmospheric concentration of the contaminants of interest for this paper (native compounds) in “A. Liberti” monitoring station is below the limit of quantification, and labelled compounds act as the natives, a simulated polluted air with a known amount of 13C labelled SS solution (ISO 2007b; Cerasa et al. 2020) spiked on the QFF was used. All the sorbents were separately extracted, and recovery rates of SS solution (%RSS) were evaluated for each sampling test. The breakthrough volume of the ACF 3A could be considered analytically insignificant since the %RSS in ACF 3B and in PUF were lower than 10% of the initial amount spiked on QFF. Figures 2 and 3 report the average recovery rates of SS solution (%Rss) of triplicate sampling for PCDD/Fs and dl-PCBs on QFF and ACF 3A adsorbent for the sampling volumes of 288, 876, and 2016 m3. The %RSS of QFF, ACF A, and ACF B in every test are presented in Supplementary Material, Tables S5–S7. The fraction collected on PUF was always < LOD.Fig. 2 Comparison of average %RSS on 102 mm QFF and 58 mm ACF 3A for PCDD/Fs (Set-up B), at different sampling times (and corresponding volumes). A 24 h (288 m3); B 72 h (876 m3); C 168 h (2016 m3). The “*” means 13C-labelled Fig. 3 Comparison of average %RSS on 102 mm QFF and 58 mm ACF 3A for dl-PCBs (Set-up B), at different sampling times (and corresponding volumes). A 24 h (288 m3); B 72 h (876 m3); C 168 h (2016 m.3). The individual PCB congeners are reported according to the IUPAC nomenclature; the “L” means “labelled” The analysis of data in Figs. 2 and 3 shows that the partitioning of PCDD/Fs and dl-PCBs congeners between the QFF and the ACF 3A adsorbent is fully coherent with the values of their partial vapor pressure (Peters et al. 2000) that, in the homologous series investigated, is inversely related to the number of chlorine atoms in the molecule and molecular weight. Concerning PCDD/Fs (Fig. 2), the fraction of tetrachloro-substituted PCDD/Fs retained on the QFF was < LOD, whereas that of octachloro-substituted PCDD/F was still ca. 70%, at the maximum sampled volume (2016 m3). Differences in the partial vapor pressure also explain why PCDFs with an increasing content of chlorine atoms in the molecule were less retained on the QFF than the corresponding PCDDs congeners having the same degree of chlorination. Similar considerations apply for dl-PCBs reported in Fig. 3, where the most volatile congeners, such as the tetra- and penta-chlorinated ones, were completely lost from the QFF after 24-h of sampling (Fig. 3 A), whereas ca. 16% of the hepta- congeners was still present in it after 168 h of sampling (Fig. 3 C). As expected, an increase in the sampled volume produced an increasing release of PCDD/Fs and dl-PCBs from the QFF, that were transferred as vapors to the ACF 3A adsorbent. Tables S5–S7 in the supplementary show the concentration of PCDD/Fs and dl-PCBs identified separately on each sorbent. The congeners of PCDD/Fs and dl-PCBs collected in the backup filter ACF 3B and in the PUF were almost all < 5% of the initial amount spiked on the QFF. This confirms the absence of breakthrough for up to 2000 m3 sampled and that one 58-mm ACF filter can retain the analytes investigated. Collection of total PCDD/Fs and dl-PCBs in the air on a single ACF sorbent Since no breakthrough volume was achieved on the ACF by any of the tested PCDD/Fs and dl-PCBs congeners, the adsorbent has been shown to efficiently collect the gas phase. Then, the adsorption/retention efficiency of ACF of the particle-bound POPs, introducing the matrix effect was evaluated. As described in the “ACF as a single sorbent” subsection, air samples were collected on a sampling train consisting of a 102-mm ACF/PUF (Fig. 1, Set-up C) in parallel to the reference 102-mm QFF + PUF system (Fig. 1, Set-up A), in an environment naturally contaminated by PCDD/Fs and dl-PCBs. The PUF was considered only as a backup filter. The first step was to define the validity of the parallel samplings (ACF vs QFF + PUF) according to previously reported QA/QC. For this purpose, the %RES were considered and evaluated: if they fall within the established ranges, it means that none of the laboratory steps (extraction and clean up) affects the sample. Once the losses due to the processes that the sample undergoes in the laboratory had been evaluated, the %RSS were considered. Tables 2 and 3 report the average (n = 7) and the range of %R for ES and SS, obtained for the labelled PCDD/Fs and dl-PCBs, respectively.Table 2 Mean values of %RES and %RSS of PCDD/Fs (n = 7) for samples collected with the reference method (QFF + PUF, Set-up A) and the ACF (Set-up C). The PUF in the Set-up C was added as a backup filter Set-up A Set-up C QFF + PUF ACF Backup PUF Mean Min–max Mean Min–max Mean Min–max %RES 13C-2378-TeCDD 69 52–78 70 54–76 76 51–81 13C-12378-PeCDD 79 51–88 83 50–106 97 70–117 13C-123478-HxCDD 60 60–68 70 62–72 60 59–72 13C-123678-HxCDD 64 52–70 73 59–93 91 59–109 13C-123789-HxCDD 49 50–100 76 58–87 67 63–87 13C-1234678-HpCDD 92 50–111 96 65–113 73 65–81 13C-OCDD 78 53–93 80 59–88 89 59–92 13C-2378-TeCDF 87 53–98 108 53–114 82 79–91 13C-12378-PeCDF 90 57–119 88 67–114 76 69–84 13C-23478-PeCDF 91 62–118 83 64–113 74 70–79 13C-123478-HxCDF 84 60–96 95 69–108 101 73–102 13C-123678-HxCDF 99 64–88 75 69–79 77 72–80 13C-234678-HxCDF 105 62–84 82 61–121 95 65–99 %RSS 13C-12378-PeCDF 92 83–106 102 69–114  < 1 - 13C-123789-HxCDF 74 61–87 76 76–87  < 1 - 13C-1234789-HpCDF 90 84–98 87 81–105  < 1 - Table 3 Mean values of %RES and %RSS of dl-PCBs (n = 7) for samples collected with the reference method (QFF + PUF, Set-up A) and the ACF (Set-up C). The PUF in the Set-up C was added as a backup filter Set-up A Set-up C QFF + PUF ACF Backup PUF Mean Min–max Mean Min–max Mean Min–max %RES 81L 90 88–112 92 91–94 76 70–85 77L 90 84–114 100 75–104 72 64–82 123L 85 71–105 76 77–96 80 77–82 118L 76 67–95 70 72–91 79 77–82 114L 72 64–89 67 66–102 93 63–113 105L 69 67–82 73 67–76 79 72–83 126L 65 64–73 73 66–78 73 70–76 167L 108 79–116 88 68–121 85 71–88 156L 109 82–118 80 72–93 78 69–79 157L 103 91–116 82 78–103 90 68–96 169L 79 86–105 80 76–84 58 63–95 189L 92 87–94 88 79–94 83 71–86 %RSS 60L 134 69–137 103 79–121 16  < 1–23 127L 89 81–92 83 79–91  < 1  < 1–3 159L 53 51–54 54 51–56  < 1  < 1 An analysis of data (Tables 2 and 3) shows that results obtained on the single 102-mm ACF filter were comparable to those obtained by collecting PCDD/Fs and dl-PCBs on the combined QFF + PUF reference sampling train. Recovery rates on the backup PUF showed that only limited amounts of PCDD/Fs and dl-PCBs congeners were released from the ACF filter, with a highest value of 16% reached by the most volatile 2,3,4,4′-tetra-CB (60L, sampling standard). Despite the lower %RSS of 159L (2,3,3′,4,5,5′-hexa-CB), the values fall within the second range of sampling efficiency between 50 and 150% still considered valid (ISO 2007a). %RSS values of 159L were systematically lower than those measured in the experiments performed in the “A. Liberti” monitoring station. Since this effect was independent of the sampling train used, it was most likely caused by the different nature and concentrations of POPs collected in the particle and gas phases in the two experiments. The HxCDD/F congeners likewise had a lower extraction efficiency when compared to earlier studies, showing a similar effect. Since the determination of native compounds was possible in all samples, possible matrix effects arising from changes in the sample composition and POP concentrations were investigated. Comparing the concentrations of native PCDD/PCDF and PCB congeners determined with ACF and the QFF + PUF combination when different volumes were passed to the sampling trains would have allowed for the detection of matrix effects, if they had occurred. The minimum and maximum recoveries of all the samples for each class of PCDD/Fs and dl-PCBs fulfil the extraction and sampling efficiency requirement of ISO 16000–13 and 14 and EPA TO-4A and TO-9A reference methods. These results demonstrate that both adsorption sampling trains, QFF + PUF and ACF, are accurate in sampling micropollutants from both outdoor and indoor air. The recoveries of backup PUF corroborate the validity of the results demonstrating the absence of a breakthrough volume since the %RSS are less than 10% of the total initial amount added on ACF (Tables 4 and 5). The %RES are all within the range, validating the results of %RSS. The determination of native compounds was made possible because all seven parallels met the QA/QC standards (%RES and %RSS) and could be considered valid. Tables 4 and 5 report only the concentrations (fg TEQ/m3) of native PCDD/PCDFs and dl-PCBs congeners, respectively, at two different sampling volumes (480 and 830 m3) on the two set-ups: reference method (QFF + PUF) and the proposed single ACF filter, with a backup PUF.Table 4 Comparison of PCDD/Fs concentrations in fg TEQ/Nm3 between the reference method (QFF + PUF) and the proposed one (ACF and PUF as a backup adsorbent). Sample A = 480 m3 ~ 48 h; Sample B = 830 m3 ~ 72 h PCDD/Fs Sample A (480 m3) Sample B (830m3) fg TEQ/Nm3 QFF + PUF ACF Backup PUF QFF + PUF ACF Backup PUF 2378-TeCDD 24.0 28.7 0.2 49.3 47.0 0.1 12,378-PeCDD 20.1 24.5 0.9 13.3 9.5 0.1 123,478-HxCDD 0.05 1.0 0.05 0.03 0.01 0.02 123,678-HxCDD 0.3 0.6 0.05 0.1 0.7 0.02 123,789-HxCDD 0.06 0.1 0.05 0.08 0.1 0.01 1,234,678-HpCDD 0.04 0.02 0.02 0.07 0.04 0.01 OCDD 0.02 0.01 0.001 0.004 0.03 0.001 2378-TeCDF 1126 1201 17.3 1284 1324 16.2 12,378-PeCDF 96.2 72.6 0.8 106.3 72.5 0.01 23,478-PeCDF 287.4 311.5 0.8 554.2 485.2 9.0 123,478-HxCDF 37.0 23.3 0.6 50.5 66.3 0.8 123,678-HxCDF 35.2 26.3 0.03 29.6 37.1 0.3 234,678-HxCDF 8.3 13.6 0.3 5.0 5.9 1.0 123,789-HxCDF 2.0 1.7 0.04 1.8 1.7 0.02 1,234,678-HpCDF 2.9 2.1 0.3 3.7 1.7 0.1 1,234,789-HpCDF 0.2 0.3 0.003 0.8 0.3 0.001 OCDF 0.001 0.003 0.001 0.02 0.04 0.004 Total PCDD/F 1639.8 1707.3 21.4 2098.8 2052.5 27.6 Table 5 Comparison of dl-PCBs concentrations in fg TEQ/Nm3 between the reference method (QFF + PUF) and the proposed one (ACF and PUF as a backup adsorbent). Sample A = 480 m3 ~ 48 h; Sample B = 830 m3 ~ 72 h dl-PCBs Sample A (480 m3) Sample B (830m3) fg TEQ/Nm3 QFF + PUF ACF Backup PUF QFF + PUF ACF Backup PUF PCB 81 1.2 1.7  < LOD 1.4 1.7  < LOD PCB 77 6.1 6.4  < LOD 7.0 7.0  < LOD PCB 123 4.2 4.4  < LOD 5.4 5.3  < LOD PCB 118 47.2 45.0  < LOD 62.4 59.0  < LOD PCB 114 1.3 1.2  < LOD 1.3 1.4  < LOD PCB 105 13.8 13.1 0.1 18.4 18.1 0.1 PCB 126 8.2 9.7 1.3 1.3 3.2 1.2 PCB 167 0.9 1.1  < LOD 1.1 1.0  < LOD PCB 156 2.1 2.1  < LOD 2.3 2.0  < LOD PCB 157  < LOD  < LOD  < LOD  < LOD  < LOD  < LOD PCB 169 76.2 82.1 1.2 12.3 15.6 0.8 PCB 189  < LOD  < LOD  < LOD  < LOD  < LOD  < LOD Total dl-PCBs 161.2 166.8 2.6 112.9 114.3 2.1 Quantitative analysis of native compounds for the seven parallel samplings confirms what was observed by the %Rss value. The matrix effect and the high concentrations of the contaminants in the air appear to be the key influencing factors in the sampling. For this reason, the higher and the lower air volume samplings were compared in Tables 4 and 5. The results (Tables 4 and 5) show that a close correlation existed between the concentrations of native PCDD/Fs and dl-PCBs congeners measured with the two sampling methods. This implied that matrix effects caused by the ACF were minimal and that, regardless of the volume sampled, a strong correlation between the two data sets was feasible. Pearson’s correlation coefficient, evaluated for ∑ PCDD/Fs and ∑ dl-PCBs in fg TEQ/Nm3 from the seven parallel samplings, resulted in 0.927 and 0.892, for ∑ PCDD/Fs and ∑ dl-PCBs, respectively, confirming a strong correlation between the two sampling systems. The concentration (fg/m3) of each congener in the seven parallel samples was compared. Due to the large extent of concentrations (three orders of magnitude), data were normalized before the correlation. Figure 4 shows data concentrations of each congener in (fg/m3) from seven parallel samplings. In particular, Fig. 4 a reports the linear regression curve obtained by plotting the data of native PCDD/Fs congeners obtained using the ACF matrix vs. those obtained with the reference method (QFF + PUF), and Fig. 4 b reports the linear regression curve obtained by plotting the data sets obtained for PCB congeners. Each figure has a box on the bottom right that displays a zoom of the lower data.Fig. 4 Normalized data concentrations of each congener from seven parallel samplings. Reference method vs ACF: a Fitting of PCDD/Fs; b fitting of dl-PCBs. The box on the bottom right of each figure displays a zoom of the lower data. The axes' measuring units are arbitrary As shown in these figures, a linear slope close to 1 was obtained for both in a wide range of concentrations. The correlation coefficients measured with native dl-PCBs (0.9943) and PCDD/Fs (0.9158) were high enough (Bland and Altman 2010) to let us state that the method using a single ACF matrix performed as well as the reference method using the QFF + PUF combination. The six “outlier” normalized data in each figure correspond to 2,3,7,8-TeCDF and PCB118, the common and dominant homologues in most combustion emissions. Advantages of the ACF method Up to now, the method proposed with the ACF has been validated according to the ISO and EPA methods, and the absolute equivalence in the results concerning the double adsorbent system has been demonstrated. Comparing sampling on a single ACF matrix to the QFF + PUF combination also reveals a number of beneficial advantages. PUF suffers some oxidative degradation at high O3 levels in the environment, similar to many other organic polymeric adsorbents (Melymuk et al. 2017). Products resulting from PUF degradation can reduce the efficiency of the clean-up separation, leading to a lower signal-to-noise ratio in the GC–MS determination of PCDD/Fs and dl-PCBs. To a smaller extent, oxidation by O3 is also possible on some POPs deposited on the QFF thus increasing the uncertainty of PCDD/Fs and dl-PCBs determinations. These effects are largely prevented by the ACF as O3 is so rapidly reduced to O2 over a carbon surface as active carbon filters are commonly used in the O3 monitors to generate the zero levels in these instruments. Another advantage offered by the ACF is the saving of solvent for the extraction of PCDD/Fs and dl-PCBs samples. It has been found that a Soxhlet apparatus with a smaller volume (100 mL) can be used with the ACF compared to the 250 mL one required with the QFF + PUF combination. While it is possible to perform 432 cycles in 36 h with a 250-mL Soxhlet apparatus, it is possible to perform 1080 cycles with a one having a volume of 100 mL. Since the extraction efficiency of a Soxhlet apparatus decreases exponentially as a function of the number of cycles, no substantial recovery of the sample occurs above a certain number of cycles. This means that it is possible to reduce the extraction time if the same number of cycles is used to extract PCDD/Fs and dl-PCBs from the ACF instead of the double system QFF + PUF. Reduction in solvent volumes and extraction times also produces a lower volume of wastes and a shorter exposure of the operator to chemicals making the use of the ACF safer. Since no backup adsorbents are required, high-volume sampling on the ACF is easier to handle, and material costs can be even lower than the QFF + PUF combination. Actually, having more sampling devices, the volume of the extraction solvents increases as well as the materials that must be disposed of and of course the final cost of these analyses. The economic and time wasted tends to increase if each sampling device is extracted and analyzed in GC–MS separately. Even more important is that the use of several adsorption media introduces in the analysis a greater possibility of errors due to contamination and sample losses related to the operator’s ability during the various manipulations due to the sample processing steps that increase and to the matrix interferents coming from the materials themselves. Costs and time can be reduced as well as errors associated with the use of multiple capture media if a single sorbent is used to efficiently retain PCDD/Fs and dl-PCBs simultaneously in both the particulate and vapor phase. Conclusions Briefly, the work was developed according to the following steps. First of all, the ability of ACF to retain both PCDD/Fs and dl-PCBs in the gas phase and the breakthrough limit for the different congeners was verified through dedicated experiments performed at increasing sampling volumes. Demonstrate the linearity of the method through R%s that meet the QA/QC from 3.5 up to 0.49 pg/m3 (1000 pg of SS solution from 288 to 2016 m3 sampled). Based on the results obtained, the sampling system was adapted and tested for the simultaneous determination of PCDD/Fs and dl-PCBs in indoor and ambient air samples. Parallel analyses of polluted air samples collected using the ISO and US-EPA reference techniques proved the effectiveness, robustness, and accuracy of the ACF-based system. The results obtained in this work show unequivocally that a single ACF matrix can be used for the simultaneous determination of PCDD/Fs and dl-PCBs in indoor and atmospheric samples satisfying all the QA/QC required by the ISO and EPA reference standard methods. The methods foresee a double sampling system consisting of a QFF for the particulate matter and a PUF for the gaseous phase; the proposed method is able to collect both phases while maintaining the same efficiency and with considerable advantages. Compared to the more widespread and used combined system, the ACF has no matrix effect and does not undergo atmospheric oxidation. It produces great advantages in terms of time and costs as well as being safer and versatile enough to be adapted to different commercially available samplers. Supplementary Information Below is the link to the electronic supplementary material.Supplementary file1 (DOCX 43 KB) Author contribution Conceptualization: Marina Cerasa; data curation: Marina Cerasa, Silvia Mosca; formal analysis: Marina Cerasa; methodology: Ettore Guerriero; supervision: Catia Balducci, Alessandro Bacaloni; validation: Marina Cerasa; visualization: Marina Cerasa, Silvia Mosca; writing—original draft and review: Marina Cerasa, Silvia Mosca; writing—editing: Piero Ciccioli. All authors read and approved the final manuscript. Funding Open access funding provided by Consiglio Nazionale Delle Ricerche (CNR) within the CRUI-CARE Agreement. Data availability The data used to support the findings of this study are available from the corresponding author upon request. Declarations Ethical approval Not applicable. Consent to participate Not applicable. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Change history 7/15/2023 Missing Funding note for the Open Access. ==== Refs References Anezaki K Kashiwagi N Daily variations and factors of atmospheric PCDD/Fs in post-harvest paddy fields: PCDD/F source estimation using a Bayesian semi-factor model Chemosphere 2021 268 129292 10.1016/j.chemosphere.2020.129292 33360941 Anezaki K Yamaguchi K Monitoring of PCDD/Fs and PCBs in ambient air samples by low-volume air sampler using activated carbon fiber felt J Environ Chem 2011 21 303 311 10.5985/jec.21.303 Barbas B de la Torre A Sanz P Gas/particle partitioning and particle size distribution of PCDD/Fs and PCBs in urban ambient air Sci Total Environ 2018 624 170 179 10.1016/j.scitotenv.2017.12.114 29248706 Benedetti P, Guerriero E, Mosca S, Rotatori M (2017) Analysis of polychlorodibenzo-p-dioxins and polychlorodibenzofurans in stationary source emissions in GC–MS/MS using hydrogen as the carrier gas. J Sep Sci 40. 10.1002/jssc.201700026 Bergknut M Wiberg K Klaminder J Vertical and lateral redistribution of POPs in soils developed along a hydrological gradient Environ Sci Technol 2011 45 10378 10384 10.1021/es200938z 22014290 Bland JM Altman DG Statistical methods for assessing agreement between two methods of clinical measurement Int J Nurs Stud 2010 47 931 936 10.1016/j.ijnurstu.2009.10.001 Cerasa M Guerriero E Mosca S Evaluation of extraction procedure of PCDD/Fs, PCBs and chlorobenzenes from activated carbon fibers (ACFs) Molecules 2021 26 6407 10.3390/molecules26216407 34770816 Cerasa M, Benedetti P, De Stefanis A, et al (2020) Validation studies on activated carbon fiber passive sampler for PCDD/Fs and PCBs in water. Chemosphere 239. 10.1016/j.chemosphere.2019.124666 Colapicchioni V Mosca S Cerasa M Evaluation of the concentration of the toxic 2,3,6,7-tetrachlorobiphenylene in air after an electrical material fire J Hazard Mater 2020 393 122284 10.1016/j.jhazmat.2020.122284 32120222 Degrendele C, Fiedler H, Kočan A, et al (2020) Multiyear levels of PCDD/Fs, dl-PCBs and PAHs in background air in central Europe and implications for deposition. Chemosphere 240. 10.1016/j.chemosphere.2019.124852 EPA (1999a) EPA/625/R-96/010b. Compendium of methods for the determination of toxic organic compounds in ambient air - Second edition. Method TO-4A: Determination of Pesticides and Polychlorinated Biphenyls in Ambient Air Using High Volume Polyurethane Foam (PUF) Sampling Followed by Gas Chromatographic/Multi-Detector Detection (GC/MD) EPA (1999b) EPA/625/R-96/010b. Compendium of methods for the determination of toxic organic compounds in ambient air - Second edition. Method TO-9A: Determination of polychlorinated , polybrominated and brominated/chlorinated dibenzo-p-dioxins and dibenzofurans in ambient air EPA (1999c) EPA/625/R-96/010b. Compendium of methods for the determination of toxic organic compounds in ambient air - Second edition. Method TO-10A: Determination Of Pesticides And Polychlorinated Biphenyls In Ambient Air Using Low Volume Polyurethane Foam (PUF) Sampling Followed By Gas Chromatographic/Multi-Detector Detention (GC/MD) Forbes P Atmospheric Chemistry analysis: a review Anal Chem 2020 92 455 472 10.1021/acs.analchem.9b04623 31639303 Hart KM Pankow JF High-volume air sampler for particle and gas sampling. 2. Use of backup filters to correct for the adsorption of gas-phase polycyclic aromatic hydrocarbons to the front Filter Environ Sci Technol 1994 28 655 661 10.1021/es00053a019 22196549 Hippelein M McLachlan MS Soil/air partitioning of semivolatile organic compounds. 2. Influence of temperature and relative humidity Environ Sci Technol 2000 34 3521 3526 10.1021/es991421n ISO (2008) ISO/DIS 16000–13 Indoor air — Part 13: Determination of total (gas and particle-phase) polychlorinated dioxin-like biphenyls (PCBs) and polychlorinated dibenzo-p-dioxins/dibenzofurans (PCDDs/PCDFs) — Collection on sorbent-backed filters ISO (2009) ISO/DIS 16000–14 Indoor air — Part 14: Determination of total (gas and particle-phase) polychlorinated dioxin-like biphenyls (PCBs) and polychlorinated dibenzo-p-dioxins/dibenzofurans (PCDDs/PCDFs) — Extraction, clean-up and analysis by high-resolution gas chromatography and mass spectrometry Kaupp H Umlauf G Atmospheric gas-particle partitioning of organic compounds: comparison of sampling methods Atmos Environ Part A, Gen Top 1992 26 2259 2267 10.1016/0960-1686(92)90357-Q Król S Zabiegała B Namieśnik J Monitoring and analytics of semivolatile organic compounds (SVOCs) in indoor air Anal Bioanal Chem 2011 400 1751 1769 10.1007/s00216-011-4910-x 21567333 Larsson M Kumar Mishra B Tysklind M On the use of electronic descriptors for QSAR modelling of PCDDs, PCDFs and dioxin-like PCBs SAR QSAR Environ Res 2013 24 461 479 10.1080/1062936X.2013.791719 23724952 Lee RGM Jones KC Gas-particle partitioning of atmospheric PCDD/Fs: measurements and observations on modeling Environ Sci Technol 1999 33 3596 3604 10.1021/es980994h Lei YD Wania F Is rain or snow a more efficient scavenger of organic chemicals? Atmos Environ 2004 38 3557 3571 10.1016/j.atmosenv.2004.03.039 López A, Coscollà C, Hernández CS, et al (2021) Dioxins and dioxin-like PCBs in the ambient air of the Valencian Region (Spain): levels, human exposure, and risk assessment. Chemosphere 267. 10.1016/j.chemosphere.2020.128902 Lordgooei M Rood MJ rostam, Modeling effective diffusivity of volatile organic compounds in activated carbon fiber Environ Sci Technol 2001 35 613 619 10.1016/S1004-9541(11)60207-3 11351737 Melymuk L Bohlin-Nizzetto P Prokeš R Uncertainties in monitoring of SVOCs in air caused by within-sampler degradation during active and passive air sampling Atmos Environ 2017 167 553 565 10.1016/j.atmosenv.2017.08.038 Mosca S Torelli GN Guerriero E Evaluation of a simultaneous sampling method of PAHs, PCDD/Fs and dl-PCBs in ambient air J Environ Monit 2010 12 1092 1099 10.1039/b927004c 21491678 Peters AJ Lane DA Gundel LA A comparison of high volume and diffusion denuder samplers for measuring semivolatile organic compounds in the atmosphere Environ Sci Technol 2000 34 5001 5006 10.1021/es000056t Piazza R Gambaro A Argiriadis E Development of a method for simultaneous analysis of PCDDs, PCDFs, PCBs, PBDEs, PCNs and PAHs in Antarctic air Anal Bioanal Chem 2013 405 917 932 10.1007/s00216-012-6464-y 23073699 Rodan BD Pennington DW Eckley N Boethling RS Screening for persistent organic pollutants: techniques to provide a scientific basis for POPs criteria in international negotiations Environ Sci Technol 1999 33 3482 3488 10.1021/es980060t Wang LC Te Lin JC Di Dong C The sorption of persistent organic pollutants in microplastics from the coastal environment J Hazard Mater 2021 420 126658 10.1016/j.jhazmat.2021.126658 34329087 Wu Y Venier M Salamova A Spatioseasonal variations and partitioning behavior of organophosphate esters in the great lakes atmosphere Environ Sci Technol 2020 54 5400 5408 10.1021/acs.est.9b07755 32289228
PMC010xxxxxx/PMC10184429.txt
==== Front Cancer Cell Int Cancer Cell Int Cancer Cell International 1475-2867 BioMed Central London 37183243 2926 10.1186/s12935-023-02926-6 Research Construction of a CCL20-centered circadian-signature based prognostic model in cervical cancer Yu Yuchong 123 Liu Yao 123 Li Yuhong 123 Yang Xiaoming 123 Han Mi [email protected] 123 Fan Qiong [email protected] 123 1 grid.16821.3c 0000 0004 0368 8293 Department of Gynecologic Oncology, The International Peace Maternity and Child Health Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China 2 grid.16821.3c 0000 0004 0368 8293 Shanghai Municipal Key Clinical Specialty of Gynecologic Oncology, Shanghai, China 3 grid.16821.3c 0000 0004 0368 8293 Shanghai Key Laboratory of Embryo Original Diseases Affifiliated to Shanghai Jiao Tong University School of Medicine, Shanghai, China 15 5 2023 15 5 2023 2023 23 9212 1 2023 13 4 2023 © The Author(s) 2023 https://creativecommons.org/licenses/by/4.0/ Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Background Rather low vaccination rates for Human papillomavirus (HPV) and pre-existing cervical cancer patients with limited therapeutic strategies ask for more precise prognostic model development. On the other side, the clinical significance of circadian clock signatures in cervical cancer lacks investigation. Methods Subtypes classification based upon eight circadian clock core genes were implemented in TCGA-CESC through k-means clustering methods. Afterwards, KEGG, GO and GSEA analysis were conducted upon differentially expressed genes (DEGs) between high and low-risk groups, and tumor microenvironment (TME) investigation by CIBERSORT and ESTIMATE. Furthermore, a prognostic model was developed by cox and lasso regression methods, and verified in GSE44001 by time-dependent receiver-operating characteristic curve (ROC) analysis. Lastly, FISH and IHC were used for validation of CCL20 expression in patients’ specimens and U14 subcutaneous tumor models were built for TME composition. Results We successfully classified cervical patients into high-risk and low-risk groups based upon circadian-oscillation-signatures. Afterwards, we built a prognostic risk model composed of GJB2, CCL20 and KRT24 with excellent predictive value on patients’ overall survival (OS). We then proposed metabolism unbalance, especially for glycolysis, and immune related pathways to be major enriched signatures between the high-risk and low-risk groups. Then, we proposed an ‘immune-desert’-like suppressive myeloid cells infiltration pattern in high-risk group TME and verified its resistance to immunotherapies. Finally, CCL20 was proved positively correlated with real-world patients’ stages and induced significant less CD8+ T cells and more M2 macrophages infiltration in mouse model. Conclusions We unraveled a prognostic risk model based upon circadian oscillation and verified its solidity. Specifically, we unveiled distinct TME immune signatures in high-risk groups. Supplementary Information The online version contains supplementary material available at 10.1186/s12935-023-02926-6. Keywords Cervical cancer Circadian clock Prognostic model Bioinformatics Immunology issue-copyright-statement© BioMed Central Ltd., part of Springer Nature 2023 ==== Body pmcBackground With advancement in HPV preventative vaccines development, HPV positive cervical cancer (HPV+ CC), accounting for about 95% of CC, have been considered as preventable for HPV naïve young women [1]. However, HPV is not the only high-risk factor of CC, with approximately 5.5–11% CC HPV negative accompanied by poor prognosis [1]. Besides, HPV vaccines could not cover all HPV types, even for Gardasil-9 with expanded coverage to HPV types (6, 11, 16, 18, 31, 33, 45, 52, and 58) [2, 3], and large amounts of older women with pre-existing HPV infections remain unsuitable for such vaccines since they are not therapeutic [4]. On the other hand, late-stage CC diagnosed as distant or metastatic have limited treatment approaches and poor 5-year OS of only 20% [5]. Considering the rather low vaccination rates worldwide and large potential risk groups with HPV infection, gene signatures with prognosis and predictive value are in urgent need for CC patients. Circadian clock refers to an oscillation pattern of divergent gene networks that cooperate to adapt to environmental cycles and internal timing system on a 24-h-basis. Its core regulators include an activator comprised of CLOCK and BMAL1 and a repressor comprised of Per1/2 and CRY1/2 [6]. Emerging evidence have indicated circadian rhythm as an indispensable factor for pathophysiological processes, such as in lifespan extension through caloric restriction and tumorigenesis. Recent literatures proved that chronic jetlag and chronic circadian disruption could accelerate breast tumor growth by creating an immune-suppressive TME [7]. Furthermore, CRY1/2-/- and Per1/2-/- knockout mice experienced enhanced hepatocarcinoma metastasis along with disrupted serum bile acids and glycogen metabolism [8]. On the other hand, CLOCK mutation mice exhibited resistance to carcinogen-induced skin carcinoma development [9], and knockdown BMAL1 by shRNAs also interrupted the proliferation of leukemia stem cells [10]. Pan-cancer analysis of TCGA provided transcriptional-level evidence of the heterogeneric effects of circadian clock related gene signatures in caner hallmarks as well. PER2 was found to be highly associated with activation of PI3K/AKT oncogenic pathway and cancer metabolism, while CLOCK was quite the opposite [11]. Although disruption of circadian rhythm has been proved to be strongly correlated with cancer hallmarks including uncontrolled proliferation, genome instability, deregulation of metabolism and immune-suppressive TME, the translational value remained obscure until recently. Jiao Wang et al. designed a time-restricted regimen, that is, giving combination therapy of metformin and trastuzumab at zeitgeber time (ZT) 6 rather than ZT18, which significantly suppressed breast tumor growth compared to trastuzumab alone in murine models [12]. To be noticed, such regimen failed to restrain tumor growth given at any other ZTs. They suggested that the underlying mechanism was that metformin-targeted HK2 fell into a circadian oscillation regulated by PPARγ and PER1 and disruption of HK2 interfered with such circadian rhythm thus leading to reversion of trastuzumab [12]. Recently, another circadian clock core gene, TIMLESS was found to be over-expressed in ovarian cancer (OV) and was negatively correlated with B cells and DC cells infiltration in TCGA data analysis [13]. Knockdown of TIMELESS significantly slowed tumor growth in vivo. While disruptions in circadian clock have been described in other gynecologic neoplasms, there have not been systematic bioinformatics analysis to date on circadian machinery of cervical cancer. Considering heterogeneity effects of circadian rhythm in cancer biology, knowledge of its functioning specifically in CC might benefit a lot. Herein, to facilitate understanding into clinical relevance of circadian clock related signature in cervical cancer and illuminate its impact upon cancer metabolism and TME, we investigated TCGA-CESC dataset from the perspective of circadian oscillation thoroughly on a multi-omics level. Through K-means method, Cox regression and lasso analysis, we constructed a prognostic risk model based upon circadian-oscillation-signature and confirmed its validity by time-dependent ROC analysis. Furthermore, we discovered an ‘immune-desert’-like TME pattern in high-risk group by ESTIMATE and CIBERSORT analysis. Afterwards, to infer differences on genomic level, we conducted investigation upon SNP, CNV and TMB as well. With TIDE score and GDSC database, we then analyzed different drug sensitivity, including to immune checkpoints blockades (ICB), in high and low-risk groups. Then in combination with clinicopathological features, we built a risk prognostic model based upon independent prognostic factors and verified its solidity in GSE44001.Finally, through FISH and IHC analysis of our own real-world patients’ clinical surgery specimens, we validated CCL20 as an independent indicator for worse stages, and overexpression of CCL20 in U14 cell lines could lead to significant more M2 macrophages, Tregs and less M1 macrophages, in accordance with CIBERSORT analysis. Methods Data downloaded and arrangement The cervical cancer dataset TCGA-CESC was obtained from the TCGA-GDC official website (https://portal.gdc.cancer.gov/). Gene expression sequencing data (counts and FPKM values) (n = 307) were then transformed into TPM values separately. Clinical data of patients, including age, gender, TNM stage, survival time, and survival status, were downloaded. After excluding patients who lacked clinical information, 304 samples with clinical information were retained. The clinical data of the patients can be found in Additional file 1: Table S1. Genomics sequencing data of CESC patients (n = 304) were downloaded from GDC, among which masked somatic mutation data were selected and visualized using the maftools R package. Validation data were obtained from GEO (https://www.ncbi.nlm.nih.gov/geo/) under the accession number GSE44001. Normal cervical tissue samples were also collected from GTEx, GSE173097, GSE55940, GSE9750, GSE127265 and GSE20167. For immunotherapy analysis, datasets in GSE78220, GSE91061, GSE93157, GSE94873, GSE111636, GSE123728, GSE165252 and GSE176307 were analysed. In order to deal with the imbalance between normal(13) and tumor sample (304) numbers in TCGA, we adopted ROSE R package [14], which has been widely used, with ovun.sample function with parameters as follows: p = 0.5, method = "under". Chip platform was based on GPL14951 Illumina HumanHT-12 WG-DASL V4.0 R2 expression beadchip. Finally, 300 tumor samples were retained for inclusion in this study, and the chip data were standardized using the R limma [15] package. The circadian rhythm-related gene set was obtained from the previously published literature [1]. which contains a total of 24 genes, including ARNTL, ARNTL2, CLOCK, CRY1, CRY2, PER1, PER2, PER3, TIMELESS, BHLHE41, BHLHE40, CSNK1D, CSNK1E, DBP, FBXL3, HLF, NFIL3, NPAS2, NR1D1, NR1D2, RORA, RORB, RORC, and TEF. The different expression genes (DEGs) between high-risk and low-risk groups were compared by Wilcoxon rank-sum test and visualized through the pheatmap R package. Subtype classification and analysis based on circadian rhythm gene signatures Based on circadian rhythm genes and TCGA-CESC expression data, an unsupervised cluster analysis was performed using the "K-means method" in the "ConsensusClusterPlus" R package [16] with parameter distance of “Pearson” to identify circadian rhythm subtypes in cervical cancer patients. The consensus clustering algorithm was used to determine the number of clusters, and the analysis included 1000 iterations to ensure the stability of the classification. Principal component analysis (PCA) was performed on subgroups to judge the differences between samples. Survival analysis was performed after grouping to determine impact of grouping on prognosis. Differential and prognostic circadian rhythm-related genes screening and risk model construction The Wilcoxon rank-sum test and sleuth [17] R package was used to analyze the differences between different groups, genes with significant differences were defined as the absolute value of Log2 (Fold change) > 1.0 and adj.P-value < 0.05. in order to obtain differential circadian rhythm-related genes.To dig for the prognostic value of circadian rhythm-related genes in CESC patients, univariate Cox analysis was first used to screen for prognostic genes, with a threshold of P-value < 0.1. We used lasso and multivariate Cox proportional hazards regression model to identify independent prognostic factors further and establish a prognostic model. We then used tenfold cross-validation to test the model. The formula for calculating the risk score of the risk model is as follows:riskScore=∑iCoefficientgenei∗mRNAExpressiongenei Patients were divided into high-risk and low-risk groups based on risk scores. ROC curves for 1, 3, and 5 years were plotted by time-dependent ROC curves to determine their accuracy. In addition, the model was tested using the external test set GSE44001 according to the regression coefficients of the genes in the model, and a time-dependent ROC curve was drawn for validation. DEGs and functional enrichment analysis of circadian rhythm-related signatures To identify DEGs correlated with the circadian rhythm-related-risk model built above, sleuth [17] and Wilcoxon rank-sum test was used to decipher DEGs between the separated groups. The significantly DEGs were defined as the absolute value of logFC > 1 and adj.P-value < 0.05.Gene ontology (GO) analysis is a standard strategy for pathways enrichment studies, which include biological process (BP), molecular function (MF), and cellular component (CC) [18]. The Kyoto Encyclopedia of Genes and Genomes (KEGG) is a basic method adopted to induce enrichment pathways, such as those related to genomes, biological pathways, and drug metabolism [19]. ClusterProfiler package[20] was used to perform GO and KEGG analysis, and a cutoff value of FDR < 0.05 was considered statistically significant. To investigate the specific pathway pattern between different groups, we performed GSEA (Gene Set Enrichment Analysis) analysis [21]. "C2.cp.v7.2.symbols.gmt" and "C5.all.v7.2.symbols" were downloaded from MSigDB [22] database (https://www.gsea-msigdb.org/gsea/msigdb/index.jsp) for GSEA analysis. FDR < 0.25 was considered to be included. Protein interaction and regulatory network analysis The STRING protein–protein interaction database was used to analyze the interaction between differential genes. The core hub genes were further explored through the CytoHubba plugin in Cytoscape [23]. The hub gene-ceRNA regulation analysis was performed using the mirTarbsse database (mirtarbase.cuhk.edu.cn). The results were filtered based on the experimentally verified results in the luciferase reporter assay. Results were finally visualized by Cytoscape software. Identification and correlation analysis of tumor infiltrating immune cells The immune landscape of the tumor microenvironment (TME) was assessed using the R ESTIMATE package [24]. ESTIMATE analysis quantifies immune activity (level of immune cells infiltration) in tumor. Differences in infiltration characteristics were further compared between high and low-risk groups. Subsequently, to determine the specific level of immune cells subtypes infiltration, we deconvolved the transcriptome expression matrix using the CIBERSORT algorithm [25] (https://cibersortx.stanford.edu) to estimate the composition and abundance of different immune cells. ggplot2 R package was used to show the distribution of 22 immune cell infiltrations in TCGA-CESC dataset. Single nucleotide polymorphism (SNP) and copy number alteration (CNV) analysis To analyze SNPs we used maftools package to analyze frequently mutated genes in patients in high and low-risk groups. Subsequently, patients' masked copy number segment data were downloaded through GDC and subjected to GISTIC 2.0 analysis [26] by GenePattern [27] with default parameters as follows: t_amp = 0.1, t_del = 0.1, join_segment_size = 4,qv_thresh = 0.1,remove_X = 1,res = 0.05,conf_evel = 0.75,do_gene_gistic = 0,do_arbitration = 1,arm_peeloff = 0,sample_center = median. Tumor mutation burden (TMB), microsatellite instability (MSI) and predictive analysis of tumor immunotherapy We calculated TMB by maftools R package. The MSI-Sensor data of CESC patients were obtained from the cBioportal database (https://www.cbioportal.org). In addition, we predicted the potential response of ICB through the Tumor Immune Dysfunction and Exclusion (TIDE) score (http://tide.dfci.harvard.edu) [28] between the high and low risk groups. Drug sensitivity analysis The Genomics of Drug Sensitivity in Cancer (GDSC) database (www.cancerrxgene.org/) can be used to find tumor drug response data and sensitive markers [29]. We used the pRRophetic algorithm [30] to construct a ridge regression model based on gene expression profiles and predicted the sensitivities of high-risk and low-risk groups to common anticancer drugs through IC50 values. Construction of clinical prediction model based on circadian rhythm risk score To demonstrate the independent predictive value, we used univariate Cox and multivariate Cox to analyze the predictive power of risk score combined with clinicopathological features of patients as regards to overall survival (OS). We constructed a clinical prediction model, resampling it with the bootstrap method for validation. According to the clinical significance and statistical value, the clinical prediction nomogram (Nomogram) was constructed. To quantify discriminative performance, time-dependent ROC curves for 1, 3, and 5 years were plotted. A calibration curve was generated to assess the performance of the nomogram by comparing the predicted values of the nomogram with the observed actual survival data. Fluorescence in situ hybridization (FISH) analysis of formalin-fixed, paraffin-embedded (FFPE) tissue samples 22 patients’ cervical cancer surgery specimens were included in this study in the International Peace Maternity and Child Health Hospital, School of Medicine affiliated to Shanghai Jiao Tong University. All patients had stage I/II CV without disease progression except for one patient. In situ hybridization was carried out using the RNAscope fluorescent multiplex assay (Advanced Cell Diagnostics) After dehydration the sections were incubated with pretreat 4 for 20 min at room temperature and hybridized with probes for CCL20 mRNAs for 2 h. CCL20 mRNA copy numbers were determined by quantification of fluorescent spots using ImageJ software. Immunohistochemistry (IHC) analysis of FFPE tissue samples FFPE sections from the biopsies were subjected to IHC by using multiplex IHC kit (Panovue, Beijing, China, Cat No 0004100100). Briefly, primary antibodies were sequentially incubated followed by horseradish peroxidase (HRP)-conjugated secondary antibody and a tyramidefluorophore (Panovue), Nuclei were stained with 4`-6`-diamidino-2-phenylindole (DAPI) (SIGMA-ALDRICH, MI, USA) before the observation. Overexpression of ccl20 in U14 cell lines and flow cytometry analysis of subcutaneous mouse tumor model U14 cells were purchased from the Shanghai Institute for Biological Sciences Chinese Academy of Sciences (Shanghai, China) and routinely maintained in DMEM with 10% fetal bovine serum (FBS) and 1% Penicillin–Streptomycin (Gibco, grand Island, US, Cat No 15140-122) at 37℃ with 5% CO2. C57BL/6 mice (6–8 weeks old) were purchased from the Shanghai Laboratory Animal Center (Shanghai, China) and maintained under specific-pathogen-free (SPF) conditions in the animal facility of SJTU School of Medicine (SJTUSM). For the sake of animal ethics guidance, the number of mice in each mouse cage did not exceed 5.In order to obtain reliable survival data of mice, up to 33 mice were breeded in multiple cages at different floors of our animal centre. U14 were infected with lentivirus containing plasmid (pcSLenti-EF1-mCherry-P2A-Puro-CMV-ccl20-3xFLAG-WPRE) at the multiplicity of infection (MOI) of 1:70 and 4 μg/mL polybrene (Hanbio, Cat No HB-PB-500) for 24 h. After 48 h, cells were cultured with 2 μg/mL puromycin (Sangon Biotech, Shanghai, China, Cat No A610593) for 14 days. Afterwards, monoclonal cell population was constructed by limited dilution within 21 days and chosen by qRT-PCR and western blot verification. Mice were subcutaneously injected with 0.5 × 106 U14 cells [31]. At day21, mice were euthanasia by Carbon dioxide (CO2) inhalation, and tumors were harvested and digested for flow cytometry analysis with antibodies as follows: CD45-Apc-cy7 (BD, cat: 557659), CD4-Bv786 (BD, cat: 563331), CD3-Bv496 (BD, cat: 564661), CD8-Bv650 (BD, cat: 563234), MHC-II- percp-cy5.5 (BD, cat: 562363), CD11b-Af700 (BD, cat: 557960), F4/80-Bv605 (BD, cat: 743281), live/dead- Bv510 (BD, cat: 564406), Foxp3-pe-cy7 (ebioscience, cat: 25-5773-82), CD25-pe-cf594 (BD, cat: 562694), CD206-Bv421 (Biolegend, cat: 141717), CD86 PE (BD, cat: 564198). 12 mice every independent experiment for three times were sacrificed. Single cells were acquired with a Fortessa flow cytometer (BD Biosciences). Data were analyzed by using FlowJo software 9.0 (FlowJo LLC, Treestar Inc., OR, USA). The protocols of animal experiments were approved by the Animal Ethics Committee of SJTUSM, and performed under the Guide for the Care and Use of Laboratory Animals. Statistical analysis All data processing and analysis were implemented through R software (version 4.1.3). To compare two groups of continuous variables, the statistical significance of normally distributed variables was estimated by the independent Student t-test. The differences among non-normally distributed variables were analyzed by the Mann–Whitney U test (i.e., the Wilcoxon rank-sum test). The Chi-square test or Fisher's exact test was used to compare and analyze statistical significance between two groups of categorical variables. The survival R package was used for survival analysis. The log-rank test was used to evaluate the significance of the difference in survival time between the two groups. A time-dependent receiver operating characteristic (ROC) curve was drawn using the pROC package for R, and the area under the curve (AUC) was calculated to assess the accuracy of the risk model in predicting prognosis [18]. All statistical P values were two-sided, with P < 0.05 considered statistically significant (Fig. 1).Fig. 1 Flow chart Results Integrated transcriptome and genome analysis of circadian rhythm genes in multiple databases revealed differences between tumor and normal cervical tissues Since Youqiong Ye et al. investigated circadian clock core genes expression and mutation landscape across pan-cancer TCGA database except for CESC, we firstly extracted 24 circadian rhythm genes from the RNA-seq data of TCGA-CESC and compared the expression differences between normal and tumor groups. In comparing the normal and tumor groups, eight circadian rhythm genes were differentially expressed, including HLF, TEF, FBXL3, CRY2, RORB, PER1, NFIL3, ARNTL23, and TIMELESS (Additional file 1: Fig. S1A, B). Since there were only 13 in TCGA-GTEx normal cervival tissues, we collected normal cervical tissues from different GEO databases, and compared these differentially expressed circadian genes between GEO-collected and GTEx normal tissues and there was no difference (Additional file 1: Fig. S1C upper). Then, we compated their expression between all normal tissues and TCGA tumor samples (Fig. 2A and Additional file 1: Fig. S1C lower) and use resampling methods only in TCGA-GTEx samples (Fig. 2B) separately, and the conclusion remained the same. Finally, we used GSE9750 as validation, and the expression differences were in accordance with analyses mentioned above (Additional file 1: Fig. S1D). For example, CLOCK showed no difference between normal and tumor samples while ARNTL2 was significantly up-regulated in tumor samples (Additional file 1: Fig. S1D). Then, we extracted the mutation information of 24 circadian rhythm gene and found that the circadian rhythm genes did not have obvious mutations in CESC patients based on the somatic mutation data of TCGA-CESC patients. The mutation frequency of most of these genes is less than 1%. The mutation frequency of PER3, TIMELESS, and CLOCK genes is only 2%, and they have not been reported to be associated with CESC (Fig. 2C). In addition, we analyzed the CNV alteration patterns of 24 genes (Fig. 2D), the details of which are shown in Additional file 1: Table S2. Although amplification and deletion appeared frequently in TCGA-CESC, CNV pattern varies across different circadian clock core genes. For example, heterozygous amplification is dominant in RORC, while heterozygous deletion in PER1/2 (Fig. 2D).Fig. 2 Differences in circadian gene expression and subtype identification in CESC patients. A heat map of differential expression of circadian rhythm genes in normal group (number = 114) and tumor group (number = 304); B representative dotplot of differential expression of circadian rhythm genes in normal group (number = 13) and tumor group (number = 15) with resampling method in ROSE R package (described in methods in detail); C is the somatic mutational change of circadian rhythm genes in CESC patient group; D pie chart shows the CNVs of 24 circadian rhythm-related genes in CESC; E is the heat map of sample clustering when K = 2 using consistent clustering; F is the PCA map of subgroup 1 and subgroup 2 Due to the low mutation rate and heterogeneous CNV pattern of circadian clock genes in TCGA-CESC, we then focused on the differentially expressed circadian genes between tumor and normal mentioned above and used consensus clustering as the clustering strategy. Since k = 2 is sufficient for well-separated subgroups, we separated patients into subgroup 1 and subgroup 2 (Fig. 2E), the apparent difference between which were implicated by PCA analysis as well (Fig. 2F). Construction of circadian rhythm signature-derived risk model uncovered GJB2, CCL20 and KRT24 as independent prognosis indicators In order to dampen our understanding of these two subgroups, we used the DESeq2 package for differential analysis and finally got 86 differential genes (logFC absolute value > 1, adj. P < 0.05) (Additional file 1: Fig. S1E, F). Then, we integrated these significantly differentially expressed circadian rhythm-related genes to construct a circadian rhythm-related risk scoring model to quantitatively evaluate the prognostic information of each CESC patient by risk score. First, a univariate Cox regression analysis was performed, and eight qualified genes were selected for further research (P < 0.05). Furthermore, through lasso analysis we found that when λ = 3, we could build the most solid model (Fig. 3A, B). Afterwards, by multivariate Cox regression analysis, we found that GJB2 (P < 0.001), CCL20 (P = 0.002), and KRT24 (P = 0.004) were all independent prognostic factors separately (Fig. 3C).Fig. 3 Construction of a circadian rhythm-related risk scoring model. A–B are LASSO regression analysis, and the number of variables corresponding to the optimal λ value is 3. C is the result of multivariate Cox stepwise regression analysis, and finally found that all three genes are independent prognostic factors; D is the expression of KRT24, CCL20 and GJB2 in normal group (number = 114) and tumor group (number = 304); E–G are the survival curves of KRT24, CCL20 and GJB2 with Kaplan–Meier analysis in TCGA-CESC data (number = 304); H are time-dependent ROC curve analysis of KRT24 (left) and CCL20 (middle) in TCGA-CESC data, with the one on the right as the 1-year ROC curve of KRT24, CCL20 and GJB2; I is the prognostic nomogram of pathologic T, N stages and KRT24, CCL20, GJB2 in TCGA-CESC; J is the risk score distribution of CESC patients, the survival status of patients; K are time-dependent ROC curve analysis of the training set TCGA-CESC (left), and the right is the test set GSE44001; L is heat map of KRT24, CCL20 and GJB2 expression in high-risk (number = 124) and low-risk groups (number = 180) GJB2 (Gap Junction Protein Beta 2) (also named CX26) encodes a member of the gap junction protein family. To be noticed, CX26 enhance the self-renewal potency of triple-negative breast cancer stem cells through formation of protein complex with the pluripotency transcription factor NANOG and focal adhesion kinase (FAK) [31]. CCL20 (C–C Motif Chemokine Ligand 20) is exclusively overexpressed by myeloid-derived suppressive cells (MDSC). Since its major receptor CCR6 is highly expressed on T cells, researchers have illuminated its participation in suppressive TME [32]. KRT24 (Keratin 24) belongs to the type I (acidic) keratin family, which involves in the intermediate filament (IF) formation. There still lacks research about KRT24 in cancer biology till now. Afterwards, we checked their expression between normal and tumor samples, and found out that they were all significantly up-regulation in tumor (Fig. 3D). Besides, in line with the Cox regression analysis, GJB2 and CCL20 is indicative of worse overall survival rate (OS), and KRT24 is beneficial (Fig. 3E–G). To construct a multivariate cox regression-based risk model, we examined their saparate time-dependent ROC analysis firstly. GJB2 and CCL20 were all excellent indicator for prognosis prediction while KRT24 alone was rather general for 1-year OS (Fig. 3H, I). On another way, we conducted time-dependent AUC analysis for KRT24, and found out that its protective value for survival was more obvious for 5-year OS (Additional file 1: Fig. S2A) and progression free survival (PFS) overall (Additional file 1: Fig. S2B, C). Considering its significant prognostic value and prediction value in PFS, we also included KRT24 in the riskscore model. Based on the penalty coefficients of feature genes derived in multivariate cox regression analysis, a risk score was calculated by cumulative sum of multiplication of gene expression with corresponding coefficients, and then risk score of each sample was calculated. Time-dependent ROC analysis showed that the risk score had good predictive power for OS in CESC patients, with the areas under the curve (AUC) for 1-, 3-, and 5-year OS of 0.795, 0.649, and 0.650, respectively (Fig. 3K left). For verification, we selected dataset GSE44001, normalized and further tested performance of the risk model. Time-dependent ROC analysis in verification dataset showed that AUC for 1-year, 3-year, and 5-year PFS was 0.693, 0.662, and 0.737, respectively, indicating excellent reliability and stability of our risk model (Fig. 3K right). In addition, significant worse prognosis of CESC patients was observed along with gradually rising risk score (Fig. 3J, Additional file 1: Fig. S3B). Finally, GJB2 and CCL20 showed significant more expression in high-risk group which is in accordance with their prognosis value (Fig. 3L). Immune and metabolism-related pathways were enriched in DEGs analysis between high and low-risk groups derived from circadian rhythm signature-related risk model In order to further explore the impact of circadian rhythm-related gene signatures on CESC, patients were divided into high-risk and low-risk groups based on the median expression risk scores. Subsequently, we performed differential expression analysis and found 43 significantly differentially expressed genes (DEGs), of which 29 genes were significantly up-regulated and 14 genes were significantly down-regulated (Figs. 4A, B).Fig. 4 Analysis of differentially expressed genes and functional enrichment based on circadian rhythm-related risk models. A–B are volcano plots and heat maps showing the expression of DEGs between high and low risk CESC patients in the TCGA dataset; C is GO analysis suggesting differential genes and response to lipopolysaccharide, response to molecule of bacterial origin, epithelial cell differentiation. It is related to biological processes such as antimicrobial humoral response, and is related to molecular functions such as RAGE receptor binding and fatty acid binding, and is related to cell composition such as extracellular region and extracellular region part; D is the result of KEGG analysis showing that these differentially expressed genes are involved in IL-17 signaling pathway, Drug metabolism–cytochrome P450 and other pathways. Pathways in C and D all have qvalues less than 0.05 Afterwards, we performed functional enrichment analysis of 43 DEGs. GO analysis showed enrichment of biological processes such as response to lipopolysaccharide, epithelial cell differentiation, and antimicrobial humoral response, and also molecular functions such as RAGE receptor binding, fatty acid-binding, and extracellular region (Fig. 4C). Interestingly, the independent prognosis factors we discovered above, KRT24 and GBJ2, all participate in formation of cell junction and extracellular matrix. Besides, KEGG analysis indicated that DEGs mainly enriched in IL-17 signaling pathway, drug metabolism-cytochrome P450, and Rheumatoid arthritis (Fig. 4D), together with GO analysis indicating potential immune participation. Then we thoroughly went through GO and KEGG common results and unveiled a significant enrichment of pathways related to immunology, such as IL-17 signaling pathway, chemokine signaling pathway, and also pathways correlated to metabolism, such as glutathione metabolism. Detailed GO and KEGG results can be found in Additional file 1: Tables S3 and S4. To facilitate and verify GO and KEGG analysis, we conducted GSEA analysis based upon DEGs. It appeared that NOD-like receptor signaling pathway, RIG -like signaling pathway, fructose and mannose metabolism, pentose phosphate pathway, galactose metabolism, PPAR signaling pathway, and other pathways are enriched in low-risk group (Fig. 5). The nucleotide oligomerization domain (NOD)-like receptors 1 and 2 (NOD1/2) are intracellular pattern-recognition proteins that activate innate immune signaling pathways [33], and RIG-like receptors cooperate with Toll-like (TLR) receptors to impart innate and adaptive immune response as well [34, 35]. Apparently, GSEA analysis unveiled immune and metabolism centered pathways enrichment either, although specific pathways differ from GO and KEGG analysis. Detailed GSEA enrichment results are shown in Additional file 1: Table S5.Fig. 5 GSEA analysis. The main pathways enriched in patients are shown from the rank of low-risk to high-risk groups (left to right) according to the enrichment score NES value PPI network and non-coding RNA analysis found CCL20-centered hub_genes With enrichment analysis pointing to immune and metabolism related pathways, to further narrow down our investigation of DEGs to a set of core hub genes, we used STRING database to analyze the protein interaction pattern (Additional file 1: Fig.S4). With confidential threshold set to 0.4, the interaction pattern of DEGs could be further divided and clustered into four sub-interaction networks. To further concentrate on the core valuables in the PPI network, we chose the network with the most candidate proteins (Fig. 6A). The number of new PPI nodes (proteins) is 16, and there are 98 connecting lines (edges). The average connection degree of each node is 0.811, and the enrichment statistical P-value of the entire PPI is less than 1.0e-23. Since certain genes centered CCL20 having the most connecting edges, we set out to hypothesis that they might serve important functions.Fig. 6 PPI and regulatory network analysis. A is the PPI control network that is re-analyzed from the most specific co-interacted network in Fig. S4A, which shows the node information, connecting line information and different sub-network information of the network in detail. B is the hub gene regulation network based on CytoHubba calculation. C is the miRNA-lncRNA-ceRNA regulatory network predicted by the mirTarbase database; D is the survival plot of patients with different expression of hsa-mir-302d and has-mir-196a, with the high group representing patients with high mean expression of hsa-mir-302d and has-mir-196a in TCGA-CESC (number = 304) Subsequently, we used the CytoHubba plugin in Cytoscape to further identify which interacting proteins could be core hub genes. After calculation, ten hub proteins with the highest scores, including CCL20, CXCL1, S100A9, and CXCL8, were found (Fig. 6B). Next, we analyzed the miRNAs molecules and lncRNAs potentially regulating these hub genes through the mirTarbase database and used Cytoscape to construct a ceRNA regulatory network (Fig. 6C). Astonishingly, CCL20, CXCL8, S100A9, and IL1A are all tightly correlated with myeloid immune cells chemotaxis and functioning [32, 36–39], and long non-coding RNAs, HAGLR and FGD5-AS1 are related to suppressive immune TME either [40, 41]. To facilitate better understanding into this protein-non-coding RNAs interation network, we checked their prognostic values in TCGA-CESC miRNA database, and found out that hsa-mir-302d and has-mir-196a in this network were indicative of worse prognosis (Fig. 6D). Analysis of immune infiltration revealed a “cold” TME in high-risk gorup As a result, we assessed the effect of circadian rhythm-related risk scores on the overall immune profile in TCGA-CESC patients. It could be noticed that distribution of immune cells in CESC patients is heterogeneous, reflecting the complexity of the TME (Fig. 7A). The correlation analysis showed significant negative correlation between CD8 T cells, M0 macrophages (M0), activated mast cells and activated DCs (Fig. 7B). At the same time, CD8 T cells showed significant positive correlation with M1 macrophages (M1), activated memory CD4 T cells (activated CD4 Tm) and CD4 T follicular helper cells (CD4 Tfh). Afterwards, when comparing the infiltration of immune cells in the high-risk group versus the low-risk group, we also discovered significant more infiltration of CD8 T cells (P = 0.004) and an increasing trend of infiltration of activated CD4 Tm (P = 0.072) in low-risk group. Besides, high-risk group could be characterized as TME with significant more M0 (P = 0.013), more neutrophils (P = 0.02), more activated DCs (P = 0.026) and more activated mast cells (P < 0.01) (Fig. 7C), in accordance with correlation analysis. In conclusion, low-risk TME is tend to be one with ‘hot’ tumor features such as CD8 T cells and activated CD4 Tm, while high-risk TME is more likely to be one with suppressive myeloid cells infiltration pattern such as M0 and neutrophils.Fig. 7 Association of circadian rhythm-related risk scores with different immune cell infiltrations. A–B are the panorama analysis and immune cell correlation analysis of immune cell infiltration in the whole CESC patients. C is the difference analysis of 22 different immune cell infiltration levels between the two groups (number in high-risk group = 124, in low-risk group = 180); D–F is the difference of Estimate score, stromal score and immune score in high and low risk groups. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 Furthermore, according to the ESTIMATE results, compared with the low-risk group, patients in the high-risk group had significantly lower levels of cell infiltration abundance, fewer stromal cells and immune cells as well (Fig. 7D–F). Together with the correlation analysis and immune cells subtypes analysis, we suggested that high-risk group has higher tumor purity and an ‘immune-desert’-like TME with significant more suppressive myeloid cells infiltration than low-risk group. Effect of circadian rhythm-related risk score on genomic changes in CESC patients showed genome instability in high-risk group However, whether genome changes also participated in changes among high and low-risk groups remained to be solved. As has been indicated in 2.1, circadian clock genes themselves did not show significant genomic level changes between subgroups in CESC. We further assessed the effect of circadian rhythm-related risk scores on genomic level in CESC patients, including single nucleotide polymorphisms (SNP) and copy number variations (CNV). Analysis of SNPs in driver oncogene mutations revealed that both groups of patients have similar alterations (Fig. 8A), while CNV of both amplification and deletion were more frequent in high-risk group (Fig. 8C, D). When it comes to TMB and MSI, we did not find substantial differences between high and low-risk groups, suggesting that changes in the genomic level were not evident (Fig. 8E, F).Fig. 8 The effect of circadian rhythm-related risk grouping on genetic variation and immunotherapy in CESC patients. A–B are mutational profiles of common tumorigenesis driver genes in high- and low-risk groups of patients. Mutation information for each gene in each sample is shown in a waterfall plot, with various colors indicating different mutation types; subsections above the legend show mutational load; C–D are changes in copy number levels of different genes in patients in high and low risk groups, in which red is the gene with significantly increased copy number, and blue is the gene with significantly deleted copy number; E–F are the differences in TMB and MSI levels of patients in the high and low risk groups, respectively; G is the TIDE score calculated based on the TIDE database Differences in high and low risk groups In addition, given the critical role of cancer immunotherapy, we evaluated sensitivity to immunotherapy by TIDE scores. Surprisingly, TIDE score was significantly lower in high-risk group than in low-risk group, indicating worse responsiveness of immunotherapy of high-risk group, which was in accordance with its ‘immune-desert’-like suppressive myeloid infiltrated TME mentioned above. High-risk group tended to be heterogeneous and have worse immunotherapy effects Since TIDE score indicated worse immunotherapy outcomes in high-risk group, we set out to examine the prognosis value of our risk-model in different immunotherapy cohorts. Surprisingly, our circadian genes-derived model showed excellent prediction value in different immune cohorts (Fig. 9A), especially in melanoma (P = 0.014) and urathelial cohorts (P = 2.5e-05). Unfortunaly, we did not find any cervical cancer immunotherapy cohorts. However, considering activated DCs (aDCs) were actually enriched in high-risk group while they indicated a somewhat “hot” TME on the opposite and aDCs varied greatly inside the high-risk group, we further separated patients in TCGA according to their activated DCs enrichment score calculated by CIBERSORT, and analysed their immune enrichment pattern by ssGSEA and prognosis (Fig. 9B, C). Astonishingly, patients with fewer aDCs have significant more macrophages and neutrophils, while those with higher aDCs have more gamma-delta T cells and cytotoxic T cells. In the low-aDCs group, our risk model was still solid, while its prediction power extremely vanished in the high-aDCs group (Fig. 9C), indicating therapy stratagies targeting DCs might be a way to overcome the worse prognosis of immunotherapy in the high-risk group.Fig. 9 Immunotherapy prognosis and drug sensitivity analysis of circadian rhythm-related high- and low-risk groups. A are the survival plots of patients separated by the risk model in different immunotherapy cohorts; B is ssGSEA analysis of immune infiltration patterns of different patients separated by CIBERSORT scores of aDC expression (low and high by mean expression); C is survival plot of patients separated by CIBERSORT scores of aDC expression in TCGA-CESC (number = 304); D is according to the wilcox test comparing the drug IC50 of high and low risk groups, the top 10 drugs with p value are MS.275, EHT.1864, IPA.3, AS601245, GW843682X, Elesclomol, Salubrinal, Rapamycin, JW7.52.1, AZ628. These drugs all showed higher sensitivity in the low-risk group than in the high-risk group. Statistical analyses in (D) were calculated by non-parametric student-t test analysis between low and high risk groups’ estimated IC50 values. *: P < 0.05; **: P < 0.01; ***: P < 0.001. *: P < 0.05; **: P < 0.01; ***: P < 0.001; ns: no significance Since we have uncovered certain immunotherapy resistance in the high-risk group, we then set out to assess differences in susceptibility to common antitumor drugs between high- and low-risk groups using the GDSC database (Fig. 9D). First, we use the GDSC data as the training set and the pRRophetic package to build a ridge regression model. Then, we input the TCGA-CESC dataset for testing. The test results found that among the 138 input drugs, 41 drugs were detected to be statistically different between the two groups (Additional file 1: Table. S6). Astonishingly, EHT.1864, IPA.3 and AS601245 are all more sensitive to low-risk groups, as these regimens are related to ATP-related pathways’ inhibition or competition, suggesting possible metabolic unbalance in high-risk groups as aforementioned above (Fig. 9D). Construction of a clinical prognostic model based on circadian rhythm-related risk scores Finally, to further explore the clinical value of circadian rhythm-related risk scores, we integrated patients’ clinical characteristics into our prognostic risk model. Although no difference was found in age (Fig. 10A), in terms of stage, the proportion of advanced patients (stage II, III and IV) in the high-risk group was significantly higher than that in the low-risk group (Fig. 10B). We then constructed a prognostic model based on the circadian rhythm-related risk score and clinicopathological characteristics (age and TNM stage) of CESC patients.We tested the model by resampling 1000 time (bootstrap method). It was found that 1-year, 3-year, and 5-year AUCs were 0.894, 0.689, and 0.688, respectively (Fig. 10C). We then visualized our model through nomogram, drew a calibration curve to evaluate the model's accuracy, and found that its 1-, 3-, and 5-year OS estimates showed excellent consistency with the actual observations (Fig. 10D). Finally, except for KRT24, all features chosen showed significant values in the multivariate analysis both for OS and disease free survial (DSS) (Additional file 1: Fig. S3C, D).Fig. 10 Analysis of the predictive power of circadian rhythm-related risk scores for prognosis in patients with CESC. A–B superimposed histograms show the proportion of age and stage of patients in high and low risk groups in the two groups. Age was similarly represented in both groups, with significantly more early-stage patients in the low-risk group than in the high-risk group. C is the time-dependent ROC curve of the riskscore-based clinical prediction model. D is the nomogram of the model. E is the calibration curve of the nomogram, using the bootstrap method and resampling 1000 times; the abscissa is the survival predicted by the nomogram, and the ordinate is the actual observed survival, repeated 1000 times each time, the curve shows the model pair Patient outcomes at 1, 3 and 5 years had good predictive value CCL20 as independent indicator for worse stages in real-world patients and less functional CD8+ T cells with more M2 infiltration in mouse tumor models Although the prognostic value separately for GJB2, KRT24 and CCL20 have been verified in our bioinformatics analysis of TCGA and GEO data, the exact biological and clinical value of them deserved further research. Through GSVA immune scores independently on GJB2, KRT24 and CCL20 (Fig. 11A), we found that role of GJB2 and KRT24 separately in CESC were rather obscure since they had ImmuneScore indicative of “hot” TME with cytotoxic immune cells and DC infiltration. On the other way, CCL20 tended to favor a “cold” TME with significantly less CD8 T cells, more macrophages and more Tregs, has been deciphered through CIBERSORT analysis mentioned above (Fig. 7). Then we set out to investigate the exact expression levels of CCL20 within real-world clinical patients’ data. We collected 22 early stages (Ia1/IIa2, Ib1/2) CESC patients’ surgical specimens and performed FISH for RNA quantification (Fig. 11B). To be noticed, ccll20-007 is the only patient that has recurrence of disease within 5 years, and CCL20 RNA was significantly up-regulated between ccl20-007 and ccl20-001 and ccl20-004, the other two both showed no recurrence. Besides, through IHC analysis for CCL20 protein expression (Fig. 11C), patients that has later stages (Ib1/2, lower panel) had significant more CCL20 expression compared to those at earlier stages (Ia1/IIa2). Quantification analysis of FISH and IHC (Fig. 11D). supported an enhanced CCL20 expression in later stages patients as well. Finally, we constructed lentivirus inducted stable overexpression of CCL20 in U14 (OE-CCL20), and verified the successful overexpression by qRT-PCR, western blot, flow cytometry and IHC (Additional file 1: Figure S5).Then subcutaneous U14 tumor models were established in C57BL/6 mouse. Although we did not obsereved faster tumor growth in CCL20-OE mice (Fig. 11H), we did find out a significant worse survival in CCL20-OE mice (Fig. 10E). By flow cytometry analysis of TME, elevated ratio of M2/M1 was manifested in CCL20 overexpression (OE) group (Fig. 11E, F), along with less CD8+ T cells, less CD107a expression and more PD-1 expression of CD8+ T cells, with Tregs of no significance, in accordance with CIBERSORT analysis (Fig. 10F, G).Finally, considering the significant worse prognosis in high-risk group in immunotherapy cohort, we conducted anti-PD-L1 therapy in CCL20-OE mice, and found the same immunotherapy resistance as well (Fig. 10H), further supporting the same “immune-dessert” TME in mice characterized by less CD8+ T cells and more M2 macrophages.Fig. 11 CCL20 as an independent indicator for worse prognosis and capable of enhancing M2 macrophages infiltration. A GSVA score of different immune subtype cells infiltration seperately for GJB2, KRT24 and CCL20. B FISH analysis for CCL20, with different rows standing for different patients, as ccl20-007, ccl20-001 and ccl20-004 as three representative images. C IHC analysis for CCL20. D ImageJ quantitative analysis for FISH (up) and IHC (down). E is the survival plot of mice during observation of 55 days, with NC-CCL20 referring to mice inoculated with empty plasmid loaded lentivirus infected U14 cells as control and OE-CCL20 referring to mice inoculated with CCL20 overexpression U14 cells (mice in NC-CCL20 are 14, and OE-CCL20 are 17, which were pooled together in separate cages, with each having less than 5 mice); F FC analysis for Tregs, M2/M1, CD8, CD107a and PD-1 in TME of U14 C57BL/6 J mouse model; G are representative FC gating graphs for CD8, CD107a expression of CD8, PD-1 expression of CD8, and macrophages with CD86+ CD206− as M1 and CD86− CD206+ as M2; H is the tumor growth curve of mice with different treatments, with NC-PD-L1 referring to NC-CCL20 mice with PD-L1 blocking antibody(αPD-L1) treatments, OE-PD-L1 referring to OE-CCL20 mice with αPD-L1 treatments, NC-CCL20 and OE-CCL20 referring to mice with PBS treatments as control groups (mice in each group:NC-PD-L1: 6; OE-PD-L1:6; NC-CCL20:5; OE-CCL20: 5). The data were the representatives of at least three independent experiments as means ± S.E.M. Statistical analyses were calculated by non-parametric student-t test or two-sided ANOVA analysis. *: P < 0.05; **: P < 0.01; ***: P < 0.001. *: P < 0.05; **: P < 0.01; ***: P < 0.001; ns: no significance. Scale bar in B: 100 μm; C: 200 μm Discussion Cervical cancer is the fourth most common female cancer worldwide [5]. Besides, overall low prophylactic vaccination rates among adolescents worldwide and preexisting infections in older women demonstrate the urgent need for developing state-of-the-art early diagnostic strategies and therapeutic regimens [42], for example, immunotherapies. Circadian disruption has been associated with tumorigenesis through effects on tumor proliferation, DNA repair and stemness, and combination therapies targeting circadian disorders with other regimens are emerging for GC [12], NSCLC [43], etc. However, little is known about its effects in tumor metabolism and TME, especially in cervical cancer, as well as its prognostic and therapeutic value [44]. In this research, we systematically investigated circadian rhythm related genes genomic and transcriptional level patterns in TCGA-CESC dataset. We successfully divided CESC patients into low-risk and high-risk groups based upon their circadian clock gene signature expression modes. We further confirmed immune and metabolism related pathways enrichment of differentially expressed genes (DEGs) between high and low-risk groups, which includes response to lipopolysaccharide, fructose and mannose metabolism, and IL-17 signaling pathway. Furthermore, we also uncovered significantly different immune infiltration patterns between them. Through lasso regression analysis of DEGs, we proposed a prognosis model composed of GJB2, CCL20 and KRT24, all with independent prognostic value, and validated its predictive value on patients’ overall survival (OS) in dataset GSE44001. Finally, we illuminated that low-risk group was more sensitive to certain chemotherapy regimens and immunotherapy, providing informational clues for application of circadian gene signatures in clinical settings. Coincidentally, circadian oscillation plays an important role in tumor-immune interaction, including antigen presentation, immunogenicity enhancement or loss and TME biology, during which these pathways might experience upregulation or downregulation depending on tumor type as well. When it comes to melanoma, higher CTL infiltration along with higher PD-1/L1 expression was found to be correlated with higher BMAL1 during anti-PD-1 immunotherapy clinical settings [45]. While in KIRC and breast cancer, upregulation of CLOCK, ARNTL and PER3 promotes TME inflammation via modulating macrophages and neutrophils infiltration, which leads to worse prognosis [46, 47]. However, the specific mechanisms under this myeloid-induced inflammation have not been demonstrated yet. In our research, high-risk patients’ TME is characterized as an immune-suppressed TME with more infiltration of macrophages, neutrophils, activated mast cells and activated DCs, while TME in low-risk groups has more CD8 T cells, activated CD4 Tm and rested masted cells, indicating a myeloid-dominated signature in high-risk group with worse prognosis as well. To be noticed, reasons behind this might be attributed to enhanced expression of ICs. For example, RORγ agonists, which can activate BMAL1 transcription [48], attenuate the expression of PD-1 receptors, and LYC-55716, a RORy inhibitor, is currently under a Phase 1 trial used in combination with pembrolizumab for NSCLC (NCT03396497). Nevertheless, direct and indirect circadian control over specific immune cells subtypes should be investigated thoroughly for their potential translational clinical applications. Besides, PPI interaction built upon hub genes (CCL20, KRT16, GJB2) in our analysis pointed to interaction network among IL1A, CXCL1, CXCL8, S100A7 and S100A9, which are important mediators of myeloid immune cells function. The chemokine CCL20 is notably overexpressed by myeloid cells, as is its cognate CCR6 receptor on T cells. Disruption of the CCL20-CCR6 axis in mice restores CTL activity and significantly prolongs survival [32]. CXCL8 functions as a chemotactic factor by guiding the neutrophils to the site of infection [37]. IL1A is produced by monocytes and macrophages as a proprotein, which is proteolytically processed and released in response to cell injury, and thus induces apoptosis. Upregulated S100A7 could promote tumor proliferation through paracrine interaction with RAGE receptors [39]. To be noticed, through KEGG analysis we also uncovered RAGE receptor pathways enrichment in DEGs between high and low-risk groups. In addition, exocrine S100A7 could promote M2 macrophage infiltration in esophageal squamous carcinoma (ESCC) [36, 39]. SY Lim et al. demonstrated that monocytes/macrophages in the metastatic liver microenvironment induce S100A8 and S100A9 in cancer cells, and that these proteins are essential for tumor cell migration and invasion [36]. Another major cancer hallmark, metabolism disorder, is indispensable for cancer cells proliferation requirements under TME selection pressure as well. Constitutive activation of the PI3K/PDK1/AKT pathway and HIF1a pathway under low level of oxygen contribute to increased glycolysis in tumor [49]. Circadian clock has been proved to be essential regulators of glycolysis and oxidative phosphorylation through AKT and HIF1a pathways either [50]. Conversely, hypoxia and HIF1α affect circadian rhythms through regulation of the circadian clock genes CRYs, RORα, Per2, and Cry1 [51]. We also discovered enhanced expression pattern of fructose and mannose metabolism through KEGG analysis, as well as pentose phosphate and galactose metabolism pathways through GSEA analysis in high-risk groups of cervical cancer patients, further proving the potential link of glycolysis and circadian rhythm in cancer. Besides, PPAR signaling pathway and FA metabolism were also enriched in high-risk group, indicating regulation of FA metabolism by circadian oscillation as well. Recently, T Fedchenko et al. found that PPAR-γ agonist given to mice orally induced disruption of PER1/2 and BMAL1 expression in liver through regulation of NFKB and IL-6 pathways [52]. At the same time, PPAR signaling pathway is the center of de novo synthesis of fatty acids. Sai Ma et al. characterized a mutual activation loop between PPARγ and esophageal adenocarcinoma-specific master regulator transcription factors (MRTF) in upregulation of synthesis of phospholipids [53]. However, whether circadian clock signature could manipulate FA metabolism through PPAR pathways in cancer still lacks research till now. Circadian rhythm interference has been attributed to uncontrolled proliferation and dampened DNA damage response, which are further linked to TMB and MSI. Emerging evidence have proved effects of circadian signature upon TMB and MSI, such as the positive regulation of growth promoter SERPINE1 via BMAL2, which promotes MSI [54]. Bioinformatics analysis of TCGA discovered an index of core circadian genes (PER1/2/3, CRY1/2, CLOCK and BMAL1) that is negatively related to MMR pathway [55]. However, we did not recognize significant TMB and MSI changes between high-risk and low-risk groups based upon circadian signature classification model in TCGA-CESC. Indeed, cancer heterogeneity and tissue origin specificity could make regulation of circadian clock even more complicated, indicating necessity for thorough investigation of circadian rhythm in different cancer type separately. CCL20 was indicated as inducer of dampened anti-tumor ability of CTL when it was secreted by macrophages. We not only uncovered connection between expression of CCL20 and worse stage in clinical settings, but also uncovered disrupted CD8+ T cells function with exhaustion phenotype and M2 infiltration triggered by ccl20 overexpression in U14 cancer cell lines in immune-competent murine tumor model. Although Hirotaka et al. found that dietary consumption of Lactobacillus-derived exopolysaccharide induced CCR6+ CD8+ T cells by CCL20-secreting tumor cells [56], Wang et al. also uncovered FOXO1 promoted the migration of M2 macrophages via CCL20 secretion in esophageal squamous cell carcinoma [57]. Besides, resently Liu et al.revealed a specific M2-like macrophages subtype with high CCL20 expression, which is associated with worse prognosis by single-cell analysis in CESC [58]. Besides, they proposed that CCL20+ macrrophges also expressed high levels of CXCL8, which is in accordance with our PPI analysis in TCGA-CESC as well [59]. All in all, the regulation of CCL20 on the tumor immune microenvironment is currently controversial, our research might have shed light upon an indispensable role for M2 macrophages as well in cervical cancer. Despite the critical role of CCL20 in our circadian-based risk model, KRT24, an OS-favorable protective marker in CESC has also been included into our model. KRT24 has also been proved to be a potential tumor suppressor. Désirée. et al. described suppression of viability and proliferation induced by KRT24 upon human HNSCC cell lines and mouse xenograft model [60]. However, the impact of KRT24 upon TME in CESC still lacks certain research. Based upon a significant up-regulated immune-score calculated by ssGSEA with more KRT24 expression (Fig. 11A), we proposed that in-depth profiling of impacts of KRT24 upon TME in CESC, as well as its correlation with CCL20 is in urgent need. To our knowledge, this is the first comprehensive bioinformatics analysis of circadian rhythm signatures in cervical cancer. Through multi-omics analysis of TCGA-CESC dataset and using GSE44001 as verification, we successfully constructed a prognostic risk model based upon circadian rhythm signature and discovered three independent prognostic factors, GJB2, CCL20 and KRT24, with hints upon metabolism features and suppressive myeloid cells enriched TME as poor prognostic indicators. However, due to intrinsic nature of data-mining, the underlying mechanism of impacts of circadian clocks upon macrophages and neutrophils infiltration as well as PPAR signaling pathways and glycolysis-related pathways enrichment still need to be thoroughly verified by molecular and cell biology experiments further. Nevertheless, we believed that our research could facilitate understanding of clinical value of circadian rhythm in cervical cancer, and unleash probability of digging into the prognostic value and even targetable features of circadian clock gene sets in cancer biology. Conclusions Circadian clock disruption has been proved to relate to cancer progression. However, its relevant significance in cervical cancer still lacks thorough research. Through multi-omics analysis of TCGA and GEO publicly available data, we built up a circadian-clock signature based prognostic model, with GJB2, CCL20 and KRT24 as independent-significant prognostic factor. Further through CIBERSORT analysis, FISH and IHC analysis of clinical specimens and flow cytometry analysis of subcutaneous mouse tumor model, CCL20 was identified as inducer of desert-like TME with more M2 macrophages infiltration and as indicator of worse clinical stages. Supplementary Information Additional file 1: Figure S1. Differential gene analysis of circadian subgroups. (A) is the dotplot of expression of selected circadian genes between GEO-collected normal cervical tissues (101) and GTEx normal cervical tissues (13) (upper), and integration analysis of normal cervical tissues (114) compared with TCGA-CESC tumor tissues (304) (down); (B) is the heatmap of circadian candidate genes from a validation dataset GSE9750 (normal=23,tumor=33); (C). Differential gene heatmap between high and low-risk groups; (D). Differential gene volcano map between high and low-risk groups. Figure S2. Prognostic value of KRT24 in cases of OS or PFS. (A) is time-dependent AUC analysis of KRT24 in OS, with each broken line represents the change of the AUC value at the indicated time point; (B) is time-dependent AUC analysis of KRT24 in PFS; (C) is time-dependent ROC analysis of KRT24 for 1, 3 and 5 years’ PFS. Figure S3. (A) Circadian gene expression between high and low risk groups; (B) is the survival plot of patient separated by the risk model in TCGA-CESC; (C-D) are multivariate cox analysis of OS and disease free survival (DSS) of patients in TCGA-CESC by TNM stages, CCL20, KRT24 and GNB2. Figure S4. PPI network of core genes derived from transcriptome DEG analysis between high-risk and low-risk groups. Figure S5. Validation of CCL20-OE U14 cell lines.(A) Immunocytochemistry (ICC) of CCL20 expression for U14 cell lines of NC (NC-CCL20) (left) or overexpression (OE-CCL20) (right); (B) is western blot (WB) plot of NC (left) or OE (right), with HSP90 as the loading control; (C) is the flow cytometry plot of NC-CCL20 (left) or OE-CCL20 (right); (D) is the qRT-PCR results of CCL20 with blue indicating NC-CCL20 and red indicading OE-CCL20. Table S1. Baseline data of TCGA-CESC patients in TCGA database. Table S2. CNVs of 24 circadian gene signatures in TCGA-CESC. Table S3. Top20 significant GO analysis. Table S4. Top20 significant KEGG analysis. Table S5.GSEA analysis results. Abbreviations ZT Zeitgeber time PCA Principal component analysis GO Gene ontology BP Biological process MF Molecular function CC Cellular component KEGG Kyoto encyclopedia of genes and genomes SNP Single nucleotide polymorphism FISH Fluorescence in situ hybridization FFPE Formalin-fixed, paraffin-embedded IHC Immunohistochemistry HRP Horseradish peroxidase DAPI 4′-6′-Diamidino-2-phenylindole FBS Fetal bovine serum MOI Multiplicity of infection SPF Specific-pathogen-free ROC Receiver operating characteristic AUC Area under the curve CNV Copy number variations SNP Single nucleotide polymorphisms TLR Toll-like receptors Acknowledgements Authors would like to thank the participating patients. Authors would also like to thank and acknowledge Dr. Zhouwenli Meng from Shanghai Chest Hospital, for his valuable suggestions in bioinformatic analysis for this study. Author contributions YY conceived and designed the experiments. YY and YL performed the experiments. QF, YL and XY provided the technical support. YW and YY analyized the data and wrote the manuscript. All authors read and approved the final manuscript. Funding This work was supported by the grants from: The Shanghai Municipal Key Clinical Specialty (No. shslczdzk06302); National Natural Science Foundation of China (No. 82103029); The Project of the Science and Technology Commission of Shanghai Municipality (No. 21ZR1469500);The Shanghai Jiao Tong University Medicine-Engineering Fund (No. YG2021QN137). Availability of data and materials All data associated with this study are present in the paper or the Supplementary Materials. After publication, all reasonable requests for materials, data, and code will be fulfilled after completion of a material transfer agreement between the International Peace Maternity and Child Health Hospital, School of Medicine, Shanghai Jiao Tong University. Declarations Ethics approval and consent to participate The study was performed in accordance with the ethical guidelines of the 1975 Dec-laration of Helsinki and approved by Ethics Committee of Shanghai International Peace Maternity and Child Health Hospital. All animal experiments were performed in compliance with established ethical regulations and were approved by the Animal Ethics Committee of Shanghai Laboratory Animal Center (Shanghai, China). Consent for publication Not applicable. Competing interests All authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential competing interests. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Change history 6/23/2023 A Correction to this paper has been published: 10.1186/s12935-023-02968-w Change history 7/12/2023 A Correction to this paper has been published: 10.1186/s12935-023-02952-4 ==== Refs References 1. Xing B Guo J Sheng Y Wu G Zhao Y Human papillomavirus-negative cervical cancer: a comprehensive review Front Oncol 2020 10 606335 10.3389/fonc.2020.606335 33680928 2. Joura EA Giuliano AR Iversen O-E Bouchard C Mao C Mehlsen J Moreira ED Ngan Y Petersen LK Lazcano-Ponce E A 9-valent HPV vaccine against infection and intraepithelial neoplasia in women N Engl J Med 2015 372 711 723 10.1056/NEJMoa1405044 25693011 3. Signorelli C Odone A Ciorba V Cella P Audisio RA Lombardi A Mariani L Mennini FS Pecorelli S Rezza G Human papillomavirus 9-valent vaccine for cancer prevention: a systematic review of the available evidence Epidemiol Infect 2017 145 1962 1982 10.1017/S0950268817000747 28446260 4. Yang A Jeang J Cheng K Cheng T Yang B Wu T-C Hung C-F Current state in the development of candidate therapeutic HPV vaccines Expert Rev Vaccines 2016 15 989 1007 10.1586/14760584.2016.1157477 26901118 5. Small W Bacon MA Bajaj A Chuang LT Fisher BJ Harkenrider MM Jhingran A Kitchener HC Mileshkin LR Viswanathan AN Cervical cancer: a global health crisis Cancer 2017 123 2404 2412 10.1002/cncr.30467 28464289 6. Takahashi JS Transcriptional architecture of the mammalian circadian clock Nat Rev Genet 2017 18 164 179 10.1038/nrg.2016.150 27990019 7. Hadadi E Taylor W Li X-M Aslan Y Villote M Rivière J Duvallet G Auriau C Dulong S Raymond-Letron I Chronic circadian disruption modulates breast cancer stemness and immune microenvironment to drive metastasis in mice Nat Commun 2020 11 3193 10.1038/s41467-020-16890-6 32581213 8. Kettner NM Voicu H Finegold MJ Coarfa C Sreekumar A Putluri N Katchy CA Lee C Moore DD Fu L Circadian homeostasis of liver metabolism suppresses hepatocarcinogenesis Cancer Cell 2016 30 909 924 10.1016/j.ccell.2016.10.007 27889186 9. Hashikawa K-I Katamune C Kusunose N Matsunaga N Koyanagi S Ohdo S Dysfunction of the circadian tran-scriptional factor CLOCK in mice resists chemical carcinogen-induced tumorigenesis Sci Rep 2017 7 9995 10.1038/s41598-017-10599-1 28855649 10. Puram RV Kowalczyk MS Boer de CG Schneider RK Miller PG McConkey M Tothova Z Tejero H Heckl D Järås M Core circadian clock genes regulate leukemia stem cells in AML Cell 2016 165 303 316 10.1016/j.cell.2016.03.015 27058663 11. Ye Y Xiang Y Ozguc FM Kim Y Liu C-J Park PK Hu Q Diao L Lou Y Lin C The genomic landscape and pharmacogenomic interactions of clock genes in cancer chronotherapy Cell Syst 2018 6 314 328.e2 10.1016/j.cels.2018.01.013 29525205 12. Wang J Huang Q Hu X Zhang S Jiang Y Yao G Hu K Xu X Liang B Wu Q Disrupting circadian rhythm via the PER1-HK2 axis reverses trastuzumab resistance in gastric cancer Cancer Res 2022 82 1503 1517 10.1158/0008-5472.CAN-21-1820 35255118 13. Xing X Gu F Hua L Cui X Li D Wu Z Zhang R TIMELESS promotes tumor progression by enhancing macrophages recruitment in ovarian cancer Front Oncol 2021 11 732058 10.3389/fonc.2021.732058 34490127 14. Lunardon N Menardi G Torelli N ROSE: a package for binary imbalanced learning R J 2014 6 82 92 10.1101/gr.239244.118 15. Ritchie ME Phipson B Di Wu Hu Y Law CW Shi W Smyth GK limma powers differential expression analyses for RNA-sequencing and microarray studies Nucleic Acids Res 2015 43 e47 10.1093/nar/gkv007 25605792 16. Wilkerson MD Hayes DN ConsensusClusterPlus: a class discovery tool with confidence assessments and item tracking Bioinformatics 2010 26 1572 1573 10.1093/bioinformatics/btq170 20427518 17. Pimentel H Bray NL Puente S Melsted P Pachter L Differential analysis of RNA-seq incor-porating quantification uncertainty Nat Methods 2017 14 7 687 690 10.1038/nmeth.4324 28581496 18. Ashburner M Ball CA Blake JA Botstein D Butler H Cherry JM Davis AP Dolinski K Dwight SS Eppig JT Gene ontology: tool for the unification of biology Gene Ontol Consortium Nat Genet 2000 25 25 29 10.1038/75556 19. Ogata H Goto S Sato K Fujibuchi W Bono H Kanehisa M KEGG: kyoto encyclopedia of genes and genomes Nucleic Acids Res 1999 27 29 34 10.1093/nar/27.1.29 9847135 20. Yu G Wang L-G Han Y He Q-Y clusterProfiler: an R package for comparing biological themes among gene clusters OMICS 2012 16 284 304 10.1089/omi.2011.0118 22455463 21. Subramanian A Tamayo P Mootha VK Mukherjee S Ebert BL Gillette MA Paulovich A Pomeroy SL Golub TR Lander ES Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles Proc Natl Acad Sci U S A 2005 102 15545 15550 10.1073/pnas.0506580102 16199517 22. Liberzon A Birger C Thorvaldsdóttir H Ghandi M Mesirov JP Tamayo P The molecular signatures database (MSigDB) hallmark gene set collection Cell Syst 2015 1 417 425 10.1016/j.cels.2015.12.004 26771021 23. Chin C-H Chen S-H Wu H-H Ho C-W Ko M-T Lin C-Y cytoHubba: identifying hub objects and sub-networks from complex interactome BMC Syst Biol 2014 8 Suppl 4 S11 10.1186/1752-0509-8-S4-S11 25521941 24. Yoshihara K Shahmoradgoli M Martínez E Vegesna R Kim H Torres-Garcia W Treviño V Shen H Laird PW Levine DA Inferring tumour purity and stromal and immune cell admixture from expression data Nat Commun 2013 4 2612 10.1038/ncomms3612 24113773 25. Newman AM Liu CL Green MR Gentles AJ Feng W Xu Y Hoang CD Diehn M Alizadeh AA Robust enu-meration of cell subsets from tissue expression profiles Nat Methods 2015 12 453 457 10.1038/nmeth.3337 25822800 26. Mermel CH Schumacher SE Hill B Meyerson ML Beroukhim R Getz G GISTIC2.0 facilitates sensitive and confident localization of the targets of focal somatic copy-number alteration in human cancers Genome Biol 2011 12 R41 10.1186/gb-2011-12-4-r41 21527027 27. Coletta A Molter C Duqué R Steenhoff D Taminau J Schaetzen V Meganck S Lazar C Venet D Detours de V InSilico DB genomic datasets hub: an efficient starting point for analyzing genome-wide studies in genepattern, inte-grative genomics viewer, and R/Bioconductor Genome Biol 2012 13 R104 10.1186/gb-2012-13-11-r104 23158523 28. Fu J Li K Zhang W Wan C Zhang J Jiang P Liu XS Large-scale public data reuse to model immunotherapy response and resistance Genome Med 2020 12 21 10.1186/s13073-020-0721-z 32102694 29. Yang W Soares J Greninger P Edelman EJ Lightfoot H Forbes S Bindal N Beare D Smith JA Thompson IR Genomics of drug sensitivity in cancer (GDSC): a resource for therapeutic biomarker discovery in cancer cells Nucleic Acids Res 2013 41 D955 D961 10.1093/nar/gks1111 23180760 30. Geeleher P Cox N Huang RS pRRophetic: an R package for prediction of clinical chemotherapeutic response from tumor gene expression levels PLoS ONE 2014 9 e107468 10.1371/journal.pone.0107468 25229481 31. Liang Y Wang W Zhu X Yu M Zhou C Inhibition of myeloid-derived suppressive cell function with all-trans retinoic acid enhanced anti-PD-L1 efficacy in cervical cancer In Sci Reports 2022 12 1 9619 10.1038/s41598-022-13855-1 32. Walch-Rückheim B Mavrova R Henning M Vicinus B Kim Y-J Bohle RM Juhasz-Böss I Solomayer E-F Smola S Stromal fibroblasts induce CCL20 through IL6/C/EBPβ to support the recruitment of Th17 cells during cervical cancer progression Cancer Res 2015 75 5248 5259 10.1158/0008-5472.CAN-15-0732 26631268 33. Moossavi M Parsamanesh N Bahrami A Atkin SL Sahebkar A Role of the NLRP3 inflammasome in cancer Mol Cancer 2018 17 158 10.1186/s12943-018-0900-3 30447690 34. Heidegger S Wintges A Stritzke F Bek S Steiger K Koenig P-A Göttert S Engleitner T Öllinger R Nedelko T RIG-I activation is critical for responsiveness to checkpoint blockade Sci Immunol 2019 10.1126/sciimmunol.aau8943 31519811 35. Zheng W Ranoa DRE Huang X Hou Y Yang K Poli EC Beckett MA Fu Y-X Weichselbaum RR RIG-I-like receptor LGP2 is required for tumor control by radiotherapy Cancer Res 2020 80 5633 5641 10.1158/0008-5472.CAN-20-2324 33087322 36. Lim SY Yuzhalin AE Gordon-Weeks AN Muschel RJ Tumor-infiltrating monocytes/macrophages promote tumor invasion and migration by upregulating S100A8 and S100A9 expression in cancer cells Oncogene 2016 10.1038/onc.2016.107 27991927 37. Lin C He H Liu H Li R Chen Y Qi Y Jiang Q Chen L Zhang P Zhang H Tumour-associated macro-phages-derived CXCL8 determines immune evasion through autonomous PD-L1 expression in gastric cancer Gut 2019 68 1764 1773 10.1136/gutjnl-2018-316324 30661053 38. Miao L Zhuo Z Tang J Huang X Liu J Wang H-Y Xia H He J FABP4 deactivates NF-κB-IL1α pathway by ubiq-uitinating ATPB in tumor-associated macrophages and promotes neuroblastoma progression Clin Transl Med 2021 11 e395 10.1002/ctm2.395 33931964 39. Lu Z Zheng S Liu C Wang X Zhang G Wang F Wang S Huang J Mao S Lei Y S100A7 as a potential diagnostic and prognostic biomarker of esophageal squamous cell carcinoma promotes M2 macrophage infiltration and an-giogenesis Clin Transl Med 2021 11 e459 10.1002/ctm2.459 34323409 40. Liu G Du X Xiao L Zeng Q Liu Q Activation of FGD5-AS1 promotes progression of cervical cancer through regulating BST2 to inhibit macrophage M1 polarization J Immunol Res 2021 2021 5857214 10.1155/2021/5857214 34692852 41. Chen D Li X Li H Wang K Tian X Identification of immune-related prognostic mRNA and lncRNA in patients with hepatocellular carcinoma J Oncol 2022 2022 5313149 10.1155/2022/5313149 35027925 42. Ferrall L Lin KY Roden RBS Hung C-F Wu T-C Cervical cancer immunotherapy: facts and hopes Clin Cancer Res 2021 27 4953 4973 10.1158/1078-0432.CCR-20-2833 33888488 43. He L Fan Y Zhang Y Tu T Zhang Q Yuan F Cheng C Single-cell transcriptomic analysis reveals circadian rhythm disruption associated with poor prognosis and drug-resistance in lung adenocarcinoma J Pineal Res 2022 10.1111/jpi.12803 36396897 44. van der Watt PJ Roden LC Davis KT Parker MI Leaner VD Circadian oscillations persist in cervical and esophageal cancer cells displaying decreased expression of tumor-suppressing circadian clock genes Mol Cancer Res 2020 18 1340 1353 10.1158/1541-7786.MCR-19-1074 32503923 45. Assis de LVM Kinker GS Moraes MN Markus RP Fernandes PA Castrucci AMdL Expression of the Circadian Clock Gene BMAL1 Positively Correlates With Antitumor Immunity and Patient Survival in Metastatic Melanoma Front Oncol 2018 10.3389/fonc.2018.00185 46. Ramos CA Ouyang C Qi Y Chung Y Cheng C-T LaBarge MA Seewaldt VL Ann DK A non-canonical function of BMAL1 metabolically limits obesity-promoted triple-negative breast cancer iScience 2020 23 100839 10.1016/j.isci.2020.100839 32058954 47. Zhou L Luo Z Li Z Huang Q Circadian clock is associated with tumor microenvironment in kidney renal clear cell car-cinoma Aging (Albany NY) 2020 12 14620 14632 10.18632/aging.103509 32681792 48. Guillaumond F Dardente H Giguère V Cermakian N Differential control of Bmal1 circadian transcription by REV-ERB and ROR nuclear receptors J Biol Rhythms 2005 20 391 403 10.1177/0748730405277232 16267379 49. Dong S Liang S Cheng Z Zhang X Luo L Li L Zhang W Li S Xu Q Zhong M ROS/PI3K/Akt and Wnt/β-catenin signalings activate HIF-1α-induced metabolic reprogramming to impart 5-fluorouracil resistance in colorectal cancer J Exp Clin Cancer Res 2022 41 15 10.1186/s13046-021-02229-6 34998404 50. Dyar Kenneth A Lutter Dominik Artati Anna Ceglia Nicholas J Liu Yu Armenta Danny Jastroch Martin Schneider Sandra de Mateo Sara Cervantes Marlene Atlas of circadian metabolism reveals system-wide coordination and communication between clocks Cell 2018 174 1571 1585.e11 10.1016/j.cell.2018.08.042 30193114 51. Wu Y Tang D Liu N Xiong W Huang H Li Y Ma Z Zhao H Chen P Qi X Reciprocal regulation between the circadian clock and hypoxia signaling at the genome level in mammals Cell Metab 2017 25 73 85 10.1016/j.cmet.2016.09.009 27773697 52. Fedchenko T Izmailova O Shynkevych V Shlykova O Kaidashev I PPAR-γ agonist pioglitazone restored mouse liver mRNA expression of clock genes and inflammation-related genes disrupted by reversed feeding PPAR Res 2022 2022 7537210 10.1155/2022/7537210 35663475 53. Ma S Zhou B Yang Q Pan Y Yang W Freedland SJ Ding L-W Freeman MR Breunig JJ Bhowmick NA A transcriptional regulatory loop of master regulator transcription factors, PPARG, and fatty acid synthesis promotes esophageal adenocarcinoma Cancer Res 2021 81 1216 1229 10.1158/0008-5472.CAN-20-0652 33402390 54. Mazzoccoli G Pazienza V Panza A Valvano MR Benegiamo G Vinciguerra M Andriulli A Piepoli A ARNTL2 and SERPINE1: potential biomarkers for tumor aggressiveness in colorectal cancer J Cancer Res Clin Oncol 2012 138 501 511 10.1007/s00432-011-1126-6 22198637 55. Escala-Garcia M Abraham J Andrulis IL Anton-Culver H Arndt V Ashworth A Auer PL Auvinen P Beckmann MW Beesley J A network analysis to identify mediators of germline-driven differences in breast cancer prognosis Nat Commun 2020 11 312 10.1038/s41467-019-14100-6 31949161 56. Kawanabe-Matsuda H Takeda K Nakamura M Makino S Karasaki T Kakimi K Dietary lactobacillus-derived exopolysaccharide enhances immune-checkpoint blockade therapy Cancer Disc 2022 12 5 1336 1355 10.1158/2159-8290.CD-21-0929 57. Wang Y Lyu Z Qin Y Wang X Sun L Zhang Yu FOXO1 promotes tumor progres-sion by increased M2 macrophage infiltration in esophageal squamous cell carcinoma Theranostics 2020 10 25 11535 11548 10.7150/thno.45261 33052231 58. Liu C Li X Huang Q Zhang M Lei T Wang F Single-cell RNA-sequencing reveals radiochemotherapy-induced innate immune activation and MHC-II upregulation in cervical cancer Signal Trans Target Ther 2023 8 1 44 10.1038/s41392-022-01264-9 59. Liu Chao Zhang Min Yan Xinlong Ni Yanli Gong Yandong Wang Cong Single-cell dissection of cellular and molecular features underlying human cervical squamous cell carcinoma initiation and progression Sci Adv 2023 9 4 8977 10.1126/sciadv.add8977 60. Gül D Habtemichael N Dietrich D Dietrich J Gößwein D Khamis A Deuss E Künzel J Schneider G Strieth S Identification of cytokeratin24 as a tumor suppressor for the management of head and neck cancer Biol Chem 2022 403 869 890 10.1515/hsz-2021-0287 34450690
PMC010xxxxxx/PMC10192331.txt
==== Front Cell Death Discov Cell Death Discov Cell Death Discovery 2058-7716 Nature Publishing Group UK London 37198170 1465 10.1038/s41420-023-01465-3 Article Mitophagy associated self-degradation of phosphorylated MAP4 guarantees the migration and proliferation responses of keratinocytes to hypoxia Feng Yanhai 1234 http://orcid.org/0000-0003-4505-7257 Li Lingfei 4 Zhang Qiong 12 He Yongqing 4 Huang Yao 123 Zhang Junhui 5 Zhang Dongxia 12 Huang Yuesheng 6 http://orcid.org/0000-0001-9749-3435 Lei Xia 4 Hu Jiongyu 27 Luo Gaoxing [email protected] 12 1 grid.410570.7 0000 0004 1760 6682 Institute of Burn Research, Southwest Hospital, Third Military Medical University (Army Medical University), Chongqing, China 2 grid.410570.7 0000 0004 1760 6682 State Key Laboratory of Trauma, Burns and Combined Injury, Southwest Hospital, Third Military Medical University (Army Medical University), Chongqing, China 3 grid.410570.7 0000 0004 1760 6682 Army 953 Hospital, Shigatse Branch of Xinqiao Hospital, Third Military Medical University (Army Medical University), Shigatse, China 4 grid.410570.7 0000 0004 1760 6682 Department of Dermatology, Daping Hospital, Third Military Medical University (Army Medical University), Chongqing, China 5 grid.190737.b 0000 0001 0154 0904 Department of Geriatric Oncology, Department of Palliative care, Department of Clinical nutrition, Chongqing University Cancer Hospital, Chongqing, China 6 grid.263817.9 0000 0004 1773 1790 Department of Wound Repair, Institute of Wound Repair and Regeneration Medicine, Southern University of Science and Technology Hospital, Southern University of Science and Technology School of Medicine, Shenzhen, China 7 grid.410570.7 0000 0004 1760 6682 Endocrinology Department, Southwest Hospital, Third Military Medical University (Army Medical University), Chongqing, China 17 5 2023 17 5 2023 2023 9 16817 1 2023 19 4 2023 4 5 2023 © The Author(s) 2023, corrected publication 2023 https://creativecommons.org/licenses/by/4.0/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/. Our previous study has announced that phosphorylated microtubule-associated protein 4 (p-MAP4) accelerated keratinocytes migration and proliferation under hypoxia through depolymerizing microtubules. However, p-MAP4 should exhibit inhibitory effects on wound healing, for it also impaired mitochondria. Thus, figuring out the outcome of p-MAP4 after it impaired mitochondria and how the outcome influenced wound healing were far-reaching significance. Herein, the results revealed that p-MAP4 might undergo self-degradation through autophagy in hypoxic keratinocytes. Next, p-MAP4 activated mitophagy which was unobstructed and was also the principal pathway of its self-degradation triggered by hypoxia. Moreover, both Bcl-2 homology 3 (BH3) and LC3 interacting region (LIR) domains had been verified in MAP4, and they endowed MAP4 with the capability to synchronously function as a mitophagy initiator and a mitophagy substrate receptor. And, mutating any one of them ruined hypoxia-induced self-degradation of p-MAP4, resulting in destroyed proliferation and migration responses of keratinocytes to hypoxia. Our findings unviewed that p-MAP4 experienced mitophagy-associated self-degradation through utilizing its BH3 and LIR domains under hypoxia. As a result, the mitophagy-associated self-degradation of p-MAP4 guaranteed the migration and proliferation responses of keratinocytes to hypoxia. Together, this research provided a bran-new pattern of proteins in regulating wound healing, and offered a new direction for intervening wound healing. Subject terms Cell migration Mitophagy https://doi.org/10.13039/501100010909 National Science Foundation of China | Young Scientists Fund 82003323 Hu Jiongyu Military Medical Innovation Program in Daping Hospital (2019CXJSC014)https://doi.org/10.13039/501100001809 National Natural Science Foundation of China (National Science Foundation of China) 81773348 Lei Xia issue-copyright-statement© Cell Death Differentiation Association (ADMC) 2023 ==== Body pmcIntroduction As an essential component of the cytoskeleton, the microtubule is rising as an important structural and functional regulator of wound healing [1]. Microtubule-associated protein 4 (MAP4) is one critical protein regulating microtubule homeostasis, and our previous research has verified that MAP4 promoted hypoxia-induced wound healing through depolymerizing microtubules in a phosphorylation-dependent manner [2]. However, the basic biological functions of phosphorylated MAP4 (p-MAP4) are mainly divided into two aspects. In detail, p-MAP4 induced microtubule depolymerization after dissociating from microtubule in hypoxia cells, and the dissociated p-MAP4 then translocated to the mitochondrial outer membrane and impaired mitochondria [3, 4]. Therefore, p-MAP4 should possess the competency to delay wound healing from the aspect of mitochondrial impairment. Contrary to this inference, p-MAP4 promotes wound healing [2]. Thus, the key problem is what is the outcome of p-MAP4 after it impairs mitochondria and how this outcome influences wound healing. Keratinocyte autophagy has been well-identified to promote wound healing [5]. However, investigators draw this conclusion mainly based on the intervention of key molecules in the autophagy process. In addition, the autophagy-lysosomal system is also a classic way to clear damaged organelles, toxic proteins, etc., to maintain intracellular homeostasis [6]. Therefore, exploring the effects of autophagy on wound healing from the perspective of its scavenging ability is far-reaching significance. Mitophagy, a selective autophagy, is widely known for its ability to clear damaged mitochondria [7]. When mitophagy scavenges damaged mitochondria, the mitochondrial proteins and mitophagy substrate receptors are also degraded theoretically. Many proteins, such as BNIP3, have been reported to both work as autophagy activators and autophagy substrate receptors, indicating these proteins should be capable of undergoing self-degradation through mitophagy [8]. Moreover, BNIP3 has been announced to promote wound healing under hypoxia [9]. These facts raise our interests in investigating whether p-MAP4 suffered mitophagy-associated self-degradation in hypoxic keratinocytes and how this outcome impacts wound healing. Herein, the present study confirmed that p-MAP4 experienced mitophagy-associated self-degradation in keratinocytes, and the process was triggered by hypoxia. In addition, the BH3 and LIR domains made MAP4 synchronously function as a mitophagy initiator and a mitophagy substrate receptor, and were the core reasons for its self-degradation. As a result, the self-degradation of p-MAP4 assured the proliferation and migration responses of keratinocytes to hypoxia. Thus, our results hinted that the self-degradation of p-MAP4 provided a cellular basis for hypoxia-induced wound healing. Together, this research provided a bran-new pattern of proteins in regulating wound healing, and proved that the scavenging ability of autophagy might play a more central role than its other functions in mediating wound healing. Results The self-degradation of p-MAP4 was triggered by hypoxia To figure out the outcome of p-MAP4 under hypoxia, its variation tendency with time was first investigated. As shown in Fig. 1A–D, p-MAP4 peaked at 6 h and decreased at 24 h, which was accompanied by the progressively elevated autophagy level (p < 0.05). Meanwhile, the keratinocytes migration and proliferation increased mainly at 24 h after hypoxia treatment (Fig. 1E–J, p < 0.05). Based on these data, and considering the autophagy-lysosomal system was one of the main ways to degrade proteins, we preliminarily speculated that p-MAP4 might be degraded by autophagy which played material roles in regulating keratinocytes migration and proliferation under hypoxia. However, p-MAP4 could not be completely degraded by hypoxia because the autophagy flux induced by hypoxia was obstructed, which was reflected by augmented p62 expressions and accumulated autophagosomes (Fig.1A, K, L, p < 0.05). More directly, 10 nM Bafilomycin A1 (Baf A1), which was a classical inhibitor of the fusion of lysosome and autophagosome, was utilized to further clarify the effects of obstructed autophagy flux in degrading p-MAP4, and the results demonstrated that obstructed autophagy flux could not degrade p-MAP4 (Fig.1M–P, p < 0.05). Next, to further confirm whether autophagy could degrade p-MAP4, 100 nM rapamycin (Rapa), an autophagy activator, was employed. And, the data demonstrated that unobstructed autophagy induced by Rapa could degrade p-MAP4 (Fig.1Q–T, p < 0.05). Interestingly, during the process, the activating effects of p-MAP4 on autophagy, which was reflected by increased LC3-II/LC3-I ratio, was preliminarily discovered (Fig.1M–S, p < 0.05). Based on these facts, we shared a bold speculation that hypoxia might trigger the autophagy-associated self-degradation of p-MAP4.Fig. 1 The self-degradation of p-MAP4 was triggered by hypoxia. A–D Representative bands and corresponding quantifications of LC3, P62, and p-MAP4 in hypoxic keratinocytes. *p < 0.05, **p < 0.01, and ***p < 0.001 versus Con group. n = 5. E, F Representative images and corresponding quantifications of scratch wound healing assays were shown. Scale bar = 100 μm. **p < 0.01 versus Con group. n = 5. G, H Representative images and corresponding quantifications of single-cell motility assays. ***p < 0.001 versus Con group. n = 5. I, J Representative images and corresponding quantifications of EdU staining (Green) in hypoxic keratinocytes. Scale bar = 50 μm. ***p < 0.001 versus Con group. n = 5. K, L Representative images and corresponding quantifications of autophagosome and autolysosome in keratinocytes transfected with mRFP-GFP-LC3. Scale bar = 10 μm. **p < 0.01 and ***p < 0.001 versus Con group. n = 5. M–P Representative bands and corresponding quantifications of LC3, P62, and p-MAP4 in keratinocytes treated by MAP4 (Glu) adenovirus with or without 10 nM Baf A1 for 1 h. **p < 0.01 and ***p < 0.001 versus Con group. ##p < 0.01 and ###p < 0.001 versus CMV-null+Baf A1. n = 5. Q–T Representative bands and corresponding quantification of LC3, P62, and p-MAP4 in keratinocytes treated by MAP4 (Glu) adenovirus with or without 100 nM Rapa for 24 h. **p < 0.01 and ***p < 0.001 versus Con group. ###p < 0.001 versus MAP4 (Glu). n = 5. Mitophagy played a crucial role in the hypoxia-induced self-degradation of p-MAP4 The above results suggested that p-MAP4 might activate autophagy in keratinocytes, but more efforts were required. With this purpose, the MAP4 knockin (KI) mice were employed because they were an effective hyperphosphorylated MAP4 mice model, manifested by increased p-MAP4 at S737 and S760 sites (Fig. 2A, B, p < 0.05). Next, increased LC3-II/LC3-I ratio and enhanced mitophagosome number were then discovered in the skin tissue of MAP4 KI mice (Fig. 2C–F, p < 0.05). More directly, primary MKs of WT or MAP4 KI mice were isolated, and autophagy was measured once again. As shown in Fig. 2F–J, upregulated LC3-II/LC3-I ratio, augmented number of autophagosome represented by GFP-LC3 dots, and increased quantity of mitophagosome measured by TEM were also viewed in the primary MKs of MAP4 KI mice (p < 0.05). These data demonstrated that p-MAP4 activated mitophagy in keratinocytes. Particularly, the autophagy flux induced by p-MAP4 was unobstructed, which was reflected by mRFP-GFP-LC3 staining (Fig. 2K, L, p < 0.05). And, p-MAP4 also gently alleviated hypoxia-impaired autophagy flux, which was demonstrated by nearly unchanged P62 expression but increased autolysosomes in keratinocytes co-treated by hypoxia and MAP4 (Glu) adenovirus (Fig. S1A–D). More importantly, p-MAP4 could be engulfed by mitophagosomes which was indicated by the co-localization of mitophagosomes with p-MAP4 in keratinocytes treated by hypoxia or from MAP4 KI mice (Fig. 2M). And, 40 μM MDIVI-1, a mitophagy inhibitor, could block the degradation of p-MAP4 induced by hypoxia (Fig. 2N, O). Together, p-MAP4 might undergo mitophagy-associated self-degradation in hypoxic keratinocytes.Fig. 2 Mitophagy played a crucial role in hypoxia-induced self-degradation of p-MAP4. A, B Representative bands and corresponding quantification of p-MAP4, p-MAP4 (S737), and p-MAP4(S760) in the skin tissues of MAP4 KI mice. ***p < 0.001 versus the WT group. n = 5. C Representative bands and corresponding statistical analysis of LC3 in skin tissues. **p < 0.01 versus the WT group. n = 5. D, E Representative images and corresponding quantifications of autophagosomes. Yellow arrows pointed to autophagosomes. Scale bar = 0.5 μm. ***p < 0.001 versus WT group. n = 5. F Representative bands and corresponding statistical analysis of LC3 in primary MKs. **p < 0.01 versus the WT group. n = 5. G, H Representative bands and corresponding quantifications of autophagosomes in primary MKs. Yellow arrows pointed to autophagosomes/mitophagosomes. Scale bar = 0.5 μm. ***p < 0.001 versus WT group. n = 5. I, J Representative images and corresponding quantifications of GFP-LC3 positive keratinocytes. ***p < 0.001 versus WT group. n = 5. K, L Representative images and corresponding quantifications of autophagosome or autolysosome in keratinocytes measured by mRFP-GFP-LC3. Scale bar = 10 μm. *p < 0.05 and ***p < 0.001 versus WT group. n = 5. M Representative images of co-staining p-MAP4 with mitophagosomes. Scale bar = 10 μm. n = 5. N, O Representative bands and corresponding quantifications of p-MAP4 in keratinocytes treated by hypoxia with or without 30 min pre-treatment of 40 μM MDIVI-1. *p < 0.05 and ***p < 0.001 versus Con group. ###p < 0.001 versus Hypoxia (24 h) group. n = 5. BH3 domain was the first requisite gateway of the mitophagy-associated self-degradation of p-MAP4 through functioning as an autophagy activator The above data demonstrated p-MAP4 activated mitophagy, and the first key issue required to be solved was how p-MAP4 activated mitophagy. BH3 domain was a classical method to activate autophagy because it contributed to the release of Beclin1 from the Bcl-2/Beclin1 complex through competitively binding to Bcl-2 [10]. Therefore, the BH3 domain was the next focusing point. After sequence alignment, a potential and conserved BH3 domain was discovered in the N terminal of MAP4 (Fig. 3A, B). To further confirm this observation, yeast two-hybrid (Y2H) screening systems in vitro was utilized. As shown in Fig. 3C, direct interaction existed between MAP4 and Bcl-2. And, mutated MAP4, which inactivated the BH3 domain (knockout L83, G87, and D88), cleared the direct interaction between MAP4/Bcl-2 (Fig. 3D). Together, a BH3 domain was confirmed in MAP4.Fig. 3 BH3 domain was the first requisite gateway of hypoxia-induced self-degradation of p-MAP4 through functioning as an autophagy activator. A The sequence alignment of MAP4 with classical BH3 domain. B The BH3 domain of MAP4 in different species. C, D Yeast two-hybrid (Y2H) screening systems was applied to investigate the direct interaction between Bcl-2 and MAP4 before or after the BH3 domain in MAP4 was mutated. E The Bcl-2/Beclin1 and Bcl-2/MAP4 interactions in HaCaT cells treated by hypoxia. n = 5. F The Bcl-2/MAP4 interaction in HaCaT cells treated by MAP4 (Glu) or MAP4 (Ala) adenovirus which was utilized to mimic the hyperphosphorylation or dephosphorylation condition of MAP4. n = 5. G The Bcl-2/MAP4 interaction in HaCaT cells treated by MAP4 (Glu) or BH3-mutated MAP4 (Glu) adenovirus. n = 5. H Representative bands of Beclin1 in cytoplasmic or mitochondrial proteins in HaCaT cells treated by MAP4 (Glu) or BH3-mutated MAP4 (Glu) adenovirus. n = 5. I Representative band and corresponding quantification of LC3 in HaCaT cells treated by MAP4 (Glu) or BH3-mutated MAP4 (Glu) adenovirus. *p < 0.05 versus CMV-null group. ##p < 0.01 versus MAP4 (Glu) group. n = 5. J, K Representative images and corresponding quantification of autophagosome in HaCaT cells treated by MAP4 (Glu) or BH3-mutated MAP4 (Glu) adenovirus. ***p < 0.001 versus CMV-null group. ###p < 0.001 versus MAP4 (Glu) group. n = 5. Yellow arrows point to autophagosomes. L, M Representative band and corresponding quantification of p-MAP4 in HaCaT cells treated by MAP4 (Glu) or BH3-mutated MAP4 (Glu) adenovirus. **p < 0.01 versus Con group. ##p < 0.01 versus Hypoxia (24 h) group. &&&p < 0.001 versus Hypoxia (24 h) + MAP4 (Glu) group. n = 5. Then, co-IP was conducted to investigate the interactions among MAP4/Bcl-2/Beclin1, and the result indicated that increased MAP4/Bcl-2 interaction and decreased Bcl-2/Beclin1 interaction were founded in hypoxic keratinocytes (Fig. 3E). The MAP4/Bcl-2 interaction in HaCaT cells transfected with MAP4 (Glu) or MAP4 (Ala) adenovirus was then measured, and the data showed MAP4/Bcl-2 interaction was phosphorylation-dependent (Fig. 3F), indicating that the BH3 domain was activated under the phosphorylated state of MAP4. Next, BH3-mutated MAP4 (Glu) (L83/G87/D88A and S737/S760E) adenovirus which mimicked BH3 inactivation in p-MAP4 was constructed, and the adenovirus vanished MAP4/Bcl-2 interaction and downregulated the distribution of Beclin1 in cytoplasm (Fig. 3G, H). Next, BH3 inactivation decreased p-MAP4 elevated LC3-II/LC3-I ratio and autophagosome number (Fig. 3I-K, p < 0.05). Moreover, BH3 inactivation abolished hypoxia-induced self-degradation of p-MAP4 (Fig. 3L, M, p < 0.05). Together, these data demonstrated that BH3 activation was the first step of p-MAP4-activated mitophagy, and was also the first checkpoint of mitophagy-associated self-degradation of p-MAP4 in hypoxic keratinocytes. LIR (54-57) domain was the second requisite gateway of the mitophagy-associated self-degradation of p-MAP4 via functioning as a substrate recognition receptor Another key issue of mitophagy-associated self-degradation of p-MAP4 was whether MAP4 could function as a substrate recognition receptor. LIR and UIM motifs have been verified to be the decisive structures of substrate recognition receptors [11, 12]. Luckily, two potential LIRs (34–37 aa and 47–50 aa) and a UIM (463–482 aa) were discovered in the MAP4 sequence, and the motifs were highly conservative among species (Fig.4A, B). Next, selective exogenous expression of WT MAP4 (MAP4-1:1–170 aa or MAP4-2:398-547 aa) was conducted for in vitro binding assays using GST-tagged LC3. This analysis indicated that MAP4 LIR truncation (1–170 aa), rather than UIM truncation (398-547 aa), mediated the direct interaction between MAP4 and LC3 (Fig. 4C, D). Considering two LIRs were contained in MAP4-1, MAP4-1 mut1 (LIR△34,37) and MAP4-1 mut2 (LIR△47,50) were then constructed, which facilitated the observation that MAP4 LIR (47–50 aa) was the only LIR responsible for MAP4/LC3 direct interaction (Fig. 4E, F). Together, MAP4 directly interacted with LC3 through its LIR (47-50 aa), endowing MAP4 with the capability to work as a substrate recognition receptor.Fig. 4 LIR (54-57) domain was the second requisite gateway of hypoxia-induced self-degradation of p-MAP4 via functioning as a substrate recognition receptor. A, B The potential LIR or UIM domains in MAP4 were discovered and aligned among different species. C The direct interaction of GST-LC3 and HIS-MAP4-1 (1–170 aa) measured by GST pull-down. n = 3. D The direct interaction of GST-LC3 and HIS-MAP4-2 (398–547 aa) detected by GST pull-down. n = 3. E, F The direct interaction of GST-LC3 with MAP4-1 Mut1△34,37 or MAP4-1 Mut2△47,50 detected by GST pull-down. n = 3. G The transfection efficiency of LIR mutated MAP4 (Glu) adenovirus in HaCaT cells. n = 3. H–M Representative bands and corresponding quantifications of TOM20, TIM23, VDAC1, GM130, and Calnexin in HaCaT cells treated by hypoxia with or without LIR mutated MAP4 (Glu) adenovirus. *p < 0.05, **p < 0.01, and ***p < 0.001 versus Con group, ##p < 0.01 and ###p < 0.001 versus Hypo + CMV-null group. n = 5. N, O Representative images and corresponding quantifications of autophagosomes in HaCaT cells treated by hypoxia with or without LIR mutated MAP4 (Glu) adenovirus. Scale bar = 0.5 μm. ***p < 0.001 versus CMV-null group. n = 3. Yellow arrows point to mitophagosomes. P, Q Representative band and corresponding quantification of p-MAP4 in HaCaT cells treated by hypoxia with MAP4 (Glu) or LIR mutated MAP4 (Glu) adenovirus. *p < 0.05 and ***p < 0.001 versus Con group. #p < 0.05 and ##p < 0.01 versus Hypoxia (24 h) group. &&&p < 0.001 versus Hypoxia (24 h) + MAP4 (Glu) group n = 5. Given p-MAP4 translocated to mitochondria and activated mitophagy in hypoxic cells [4], whether MAP4 was a qualified mitophagy receptor was further investigated. To achieve this goal, LIR mutated MAP4 (Glu) adenovirus, which mimicked LIR inactivation in p-MAP4 was then constructed, and the results indicated that LIR inactivation abolished hypoxia-induced degradation of mitochondrial proteins and augmentation of mitophagosome number (Fig. 4G–O, p < 0.05). More importantly, LIR inactivation erased hypoxia-induced self-degradation of p-MAP4 (Fig. 4P, Q, p < 0.05). These data verified that LIR (47-50) made MAP4 a qualified mitophagy cargo receptor, and was the second checkpoint of the mitophagy-associated self-degradation of p-MAP4 in hypoxic keratinocytes. Inactivated BH3 domain blocked the proliferation and migration responses of keratinocytes to hypoxia Next, the responses of keratinocytes which were overexpressed with BH3-mutated p-MAP4 to hypoxia, were then assayed. As shown in Fig. 5A, B, BH3-mutated MAP4 (Glu) adenovirus significantly inhibited hypoxia-induced wound healing (p < 0.05), as measured by scratch wound healing assay. Meanwhile, the velocity of single-cell movement stimulated by hypoxia was suppressed by BH3-mutated MAP4 (Glu) adenovirus in keratinocytes through employing a single-cell motility test (Fig. 5C, D, p < 0.05). Moreover, through EdU staining, the adenovirus downregulated hypoxia-induced proliferation of keratinocytes (Fig. 5E, F, p < 0.05). Together, BH3 inactivation inhibited the proliferation and migration responses of keratinocytes to hypoxia.Fig. 5 Inactivated BH3 domain blocked the proliferation and migration responses of keratinocytes to hypoxia. A, B Representative images and corresponding quantifications of scratch wound healing assays in HaCaT cells treated by hypoxia with or without BH3-mutated MAP4 (Glu) adenovirus. Scale bar = 100 μm.***p < 0.001 versus CMV-null group. ###p < 0.001 versus CMV-null + Hypoxia group. &p < 0.05 versus BH3-mutated MAP4 (Glu) group. n = 5. C, D Representative images and corresponding quantifications of single-cell motility assays in HaCaT cells treated by hypoxia with or without BH3-mutated MAP4 (Glu) adenovirus. *p < 0.05 and **p < 0.01 versus CMV-null group. ##p < 0.01 versus CMV-null+Hypoxia group. &p < 0.05 versus BH3-mutated MAP4 (Glu) group. n = 5. E, F Representative images and corresponding quantifications of EdU staining (Green) in HaCaT cells treated by hypoxia with or without BH3-mutated MAP4 (Glu) adenovirus. Scale bar = 50 μm. ***p < 0.001 versus CMV-null group. ###p < 0.001 versus CMV-null+Hypoxia group. &&p < 0.05 versus BH3-mutated MAP4 (Glu) group. n = 5. Inactivated LIR (47–50) domain abolished the proliferation and migration responses of keratinocytes to hypoxia In this section, the influences of the LIR (47–50) domain on the proliferation and migration responses of keratinocytes to hypoxia were directly explored, and LIR (47–50) mutated MAP4 (Glu) adenovirus was used. As shown in Fig. 6A, B, LIR (47–50), mutated MAP4 (Glu) adenovirus showed an inhibitory effect on wound healing in scratch wound healing assay under hypoxia (p < 0.001). In addition, through single-cell mobility, LIR (47–50) mutated MAP4 (Glu) adenovirus was discovered to downregulate the elevated velocity of keratinocytes induced by hypoxia (Fig. 6C, D, p < 0.001). Furthermore, EdU staining was conducted, and the results demonstrated that LIR (47–50) mutated MAP4 (Glu) adenovirus and suppressed hypoxia augmented proliferation of keratinocytes (Fig. 6E, F, p < 0.05). Taking together, LIR inactivation disrupted the proliferation and migration responses of keratinocytes to hypoxia.Fig. 6 Inactivated LIR (47–50) domain abolished the proliferation and migration responses of keratinocytes to hypoxia. A, B Representative images and corresponding quantification of scratch wound healing assays in HaCaT cells treated by hypoxia with or without LIR mutated MAP4 (Glu) adenovirus. Scale bar = 100 μm.***p < 0.001 versus Con group. ###p < 0.001 versus Hypoxia group. n = 5. C, D Representative images and corresponding quantification of single-cell motility assays in HaCaT cells treated by hypoxia with or without LIR mutated MAP4 (Glu) adenovirus. **p < 0.01 versus Con group. ##p < 0.01 versus Hypoxia group. n = 5. E, F Representative images and corresponding quantifications of EdU staining (Green) in HaCaT cells treated by hypoxia with or without LIR mutated MAP4 (Glu) adenovirus. Scale bar = 50 μm. ***p < 0.001 versus Con group. ###p < 0.001 versus Hypoxia group. n = 5. Discussion Previous studies have identified that p-MAP4 promoted wound healing under hypoxia through depolymerizing microtubules [2]. However, p-MAP4 should inhibit wound healing, for it translocated to and impairs mitochondria [4]. Thus, figuring out the outcome of p-MAP4 after it impaired mitochondria and how the outcome influenced wound healing were meaningful. With these purposes, current findings revealed that p-MAP4 experienced mitophagy-associated self-degradation, which was triggered by hypoxia in keratinocytes. And, this process was entirely dependent on the newly confirmed BH3 and LIR domains in the MAP4 sequence. As a result, mitophagy-associated self-degradation of p-MAP4 assured the migration and proliferation responses of keratinocytes to hypoxia. Together, our discoveries provided novel insights into investigating the effects of specific proteins on wound healing, and proved that the scavenging ability of autophagy might play a more central role than its other roles in mediating wound healing. Previous article unveiled that both p-MAP4 and keratinocyte autophagy accelerated hypoxia-induced wound healing [2, 5]. And, the current data demonstrated that p-MAP4 augmented at 6 h and decreased at 24 h in hypoxic keratinocytes, but increased migration of keratinocytes presented at 24 h after hypoxia treatment. Therefore, the influences of p-MAP4 on the migration of keratinocytes might be indirect. Since the decrease in p-MAP4 expression was accompanied by a sustained increase in autophagy levels, whether p-MAP4 could be degraded by autophagy in hypoxic keratinocytes was speculated. However, autophagy flux induced by hypoxia was obstructed, and unsmoothed autophagy flux could not degrade p-MAP4, which has been proved through utilizing Baf A1. And, autophagy does consume p-MAP4, which has been confirmed by incubating keratinocytes with Rapa. During the process, p-MAP4 was discovered to activate autophagy. Further data showed that p-MAP4 stimulated mitophagy in keratinocytes, and mitophagy took a crucial position in hypoxia-induced degradation of p-MAP4. Thus, it was basically certain that p-MAP4 experienced mitophagy-associated self-degradation in keratinocytes, and this process was triggered by hypoxia. To completely elucidate whether p-MAP4 could be self-degraded via mitophagy, the first key problem required to be solved was whether it activated mitophagy by itself. Beclin1 was a classical activator of autophagy and was also an important way of Bcl-2 homology domains (BH3 domains)-containing proteins induced autophagy activation [10]. In detail, BH3-containing proteins competitively bound to Bcl-2 and then released Beclin1 from the Bcl-2/Beclin1 complex, resulting in autophagy activation [10]. Based on these facts, a bold conjecture about whether MAP4 contained a classical BH3 domain was emerged. After careful alignment and detailed experiment, a typical BH3 domain in the N terminal of the MAP4 sequence was obtained and confirmed by yeast two-hybrid technique. To further prove whether the BH3 domain was the reason for p-MAP4-activated mitophagy, BH3 inactivated MAP4 (Glu) adenovirus was constructed. Exhaustive experiments showed that the BH3 domain was the mechanism of p-MAP4-induced autophagy activation. More importantly, inactivated BH3 domain abolished the self-degradation of p-MAP4 induced by hypoxia. Thus, the BH3 domain worked as the first checkpoint of hypoxia-triggered mitophagy-associated self-degradation of p-MAP4 through working as a mitophagy activator. In addition, the confirmed BH3 domain made MAP4 a potential BH3-only protein family member. Even through, certain similarities and differences between MAP4 and traditional BH3-only proteins were existed. Firstly, most BH3-only proteins induced mitochondrial apoptosis [13], which was consistent with our previous findings about MAP4 [4]. However, their mechanisms were different, BH3-only proteins induced apoptosis was independent of mPTP and cytochrome c, but p-MAP4 was dependent [4, 14]. Secondly, the type of cell death induced by BH3-only proteins and MAP4 was different to some extent. Most BH3-only proteins stimulated apoptotic cell death [14], and p-MAP4 motivated pyroptosis [15]. Similar to MAP4, BNIP3, a well-known BH3-only protein, produced a different type of cell death which was described as necrotic cell death [14]. Therefore, MAP4 could work as a novel BH3-only protein with characteristics. Even so, to completely identify MAP4 as a BH3-only protein, more investigations should be conducted. To implement the mitophagy-associated self-degradation, p-MAP4 should have the potential to work as a mitophagy substrate receptor. The LIR and UIM domains have been regarded as the structural basis of mitophagy substrate receptors [11, 12], and both of them have been discovered in the N terminal of MAP4. And, LIR (47–50 aa) has been confirmed to be the only reason for MAP4/LC3 interaction. This result lays a structural foundation for MAP4 to work as a mitophagy cargo receptor. In addition, p-MAP4 activated mitophagy in keratinocytes which has been first described in current research. These results conferred the potential of MAP4 as a mitophagy substrate receptor. And, the process of p-MAP4-mediated mitophagy could be described as follows. LC3-bound isolation membranes selectively recognized mitochondria which was tagged by MAP4, leading to mitochondrial incorporation, mitophagosome formation, and subsequent mitochondrial clearance. Furthermore, inactivated LIR (47–50) domain eliminated p-MAP4-induced mitophagy in keratinocytes and dispelled the self-degradation of p-MAP4 induced by hypoxia. Together, the LIR (47–50) functioned as the second checkpoint of hypoxia-triggered mitophagy-associated self-degradation of p-MAP4 through functioning as a mitophagy cargo receptor. Based on the above evidences, MAP4 was thought to be a bran-new cargo receptor of mitophagy in eukaryotic cells because of its extensive distribution and conservative property [3, 16]. However, some differences existed between MAP4 and other well-known mitophagy cargo receptors. Traditionally, NIX and FUNDC1 were involved in hypoxia-induced mitophagy and were indispensable for the programmed elimination of mitochondria during reticulocyte maturation [17]. However, based on our findings, the mechanism of p-MAP4-induced mitophagy was different. Firstly, the expression of NIX or FUNDC1 was altered, whereas MAP4 remained unchanged. Besides, unlike FUNDC1, both NIX and MAP4 were distributed in other organelles under biological conditions and translocated to and impaired mitochondria under hypoxia [18, 19]. Moreover, the interaction of LC3/MAP4 was phosphorylation-dependent, which was consistent with LC3/Bnip3L interaction [20]. Therefore, MAP4 could work as a qualified mitophagy cargo receptor with specific characteristics. Even so, to completely identify MAP4 as a mitophagy cargo receptor, more efforts were required. In the entire course of the present experiment, we discovered that every aspect of mitophagy-associated self-degradation of p-MAP4 was phosphorylation-dependent. This observation explained why hypoxia was the stimulator during the process because it phosphorylated MAP4 at S737 and S760 sites [3]. In line with our findings, post-translational modifications have been reported as stimulating factors under various pathological conditions. For example, phosphorylation impacted LC3/Bnip3L or LC3/Fundc1 interactions [20]. However, MAP4 was special and its particularity was that the phosphorylation sites (S737/760) and functional regions (LIR domain (Y47/L50) and BH3 domain (L83/G87/D88A)) were located in the C-terminal and N terminal, respectively. Consistently, Mariya et al. demonstrated that O-GlcNAcylation at the C-terminus of histone deacetylase 4 activated its N-terminus and then exhibited cardiac protection effects in diabetes models [21]. These evidences hinted that MAP4 might be a self-regulation protein, and post-translational modification was the mediator. Since MAP4 was divided into microtubule binding (MTB, the C-terminal) and projection (PJ, the N terminal) segments [3], its self-regulation could be understood as follows. Under biological conditions, MAP4 mainly showed its microtubule binding function through its MTB domain, and its PJ domain was inactivated. In contrast, under pathological conditions, post-translational modification at the MTB domain activated its PJ domain which contained BH3 and LIR motifs, and then participated in pathological responses, including mitophagy. Based on these facts, the mitophagy-associated self-degradation of MAP4 was affected by its self-regulation, and post-translational modification was the key. Researchers have announced epidermal keratinocyte autophagy promotes wound healing through intervening crucial molecules of autophagy [5]. Given scavenging ability was the primary function of an autophagy-lysosome system [8], exploring the effects of autophagy on wound healing from its scavenging ability might be more significant. The current data has verified that MAP4 was a mitophagy-associated self-degradation protein. Therefore, MAP4 was a suitable molecule to elucidate the roles of autophagy on wound healing from the aspect of its scavenging ability. Our results showed that p-MAP4 with either LIR or BH3 inactivation delayed hypoxia and promoted wound healing. These facts indicated that p-MAP4 could be viewed as a toxic protein when its LIR or BH3 domain was inactivated. And, its toxicity could be cleared under specific conditions because wild-type p-MAP4 could undergo mitophagy-associated self-degradation. As a result, it guaranteed the migration and proliferation responses of keratinocytes to hypoxia. Together, scavenging toxic proteins might be the primary mechanism of autophagy-promoted wound healing. Taken together, current research has identified p-MAP4 undergone mitophagy-associated self-degradation, which was triggered by hypoxia in keratinocytes. And, the self-degradation of p-MAP4 guaranteed the migration and proliferation responses of keratinocytes to hypoxia. From another aspect, scavenging toxic proteins might be the primary mechanism of autophagy-promoted wound healing. Therefore, our discoveries provided novel insights into investigating the effects of specific proteins on wound healing, and proved that the scavenging ability of autophagy might play a more central role than its other roles in mediating wound healing. Even so, more direct evidences were required. Materials and methods Ethical statement All animal experiments were carried out according to the guidelines of the Care and Use of Laboratory Animals (NIH Pub. 8th edition, 2011). The Animal Experiment Ethics Committee of the Army Medical University (the Third Military Medical University) has approved the present study (approval number: AMUWEC2020413). Animal study The MAP4 (S667A, S737E, and S760E)-knockin (KI) mice were generated and bred as previously described [18]; WT (C57BL6/J) mice were purchased from the Animal Centre, Army Medical University (Third Military Medical University). Male mice, 6–8 weeks old and weighing 18–22 g, were used for the experiments. The mice were fed a standard rodent chow diet, allowed access to water ad libitum, and housed under 12 h light/dark cycles. Prior to the experiments, all animals were allowed to acclimate to the facility for one week. The animals were randomly divided into two groups, WT group and MAP4-KI group. Afterward, the mice were immediately euthanized, and their skin tissues were collected for transmission electron microscopy (TEM) or western blot (WB). Each animal-related test was repeated five times, and the sample size each time was 3. The sample size was estimated according to the ARRIVE Guidelines (Animal Research: Reporting of In Vivo Experiments). HaCaT cells study Human epidermal keratinocyte cells (HaCaT) were the cell line routinely used [2, 9]. And, they were purchased from and identified by the cell Bank of the Chinese Academy of Sciences. The HaCaT cells were cultured in RPMI 1640 medium (HyClone, USA) supplemented with 10% fetal bovine serum (Gibco, USA), 100 U/ml penicillin, and 100 ug/ml streptomycin (Beyotime, China), and were incubated at 37 °C, 5% CO2, and 95% humidity. With different experimental purpose, the HaCaT cells were randomly treated with specific adenoviruses or chemical compounds. Each HaCaT cells-related test was repeated 5 times, and the sample size of each time was 3. Primary mouse epidermal keratinocytes (MKs) isolation and culture Primary mouse epidermal keratinocytes (MKs) were isolated and cultured as previously described [2]. Briefly, the newborn (1–3 days) C57BL/6 or MAP4 KI mice were collected, and their skins were used for MKs isolation. Briefly, the separated skins were incubated with dispase treatment (4 °C) overnight and then were digested with 0.25% trypsin/0.04% EDTA solution (Invitrogen, USA) to obtain the MKs. The MKs were cultured in keratinocyte serum-free medium (K-SFM medium) (Gibco, USA) and were incubated at 37 °C, 5% CO2, and 95% humidity. With different experimental goals, the MKs suffered from specific treatments. Each MKs-related measurement was repeated five times, and the sample size for each time was 3. Hypoxia treatment The HaCaT cells were exposed to hypoxic treatment when necessary, and the detailed information about hypoxia was as follows. A constant flow of nitrogen using a Forma Series II Water Jacket CO2 incubator (model:3131; Thermo Scientific) was used to create a hypoxic condition which was composed of 2% O2, 5% CO2, and 93% N2. The Forma Series II Water Jacket CO2 incubator could accurately maintain the desired temperature (37 °C) and O2 level. After hypoxia, the cells were further used to conduct a single-cell mobility assay, scratch wound healing assay or other experiments. Each hypoxia-related experiment was repeated five times, and three samples each time. HA-MAP4 (Ala), HA-MAP4 (Glu), HA-LIR-mutated MAP4 (Glu), and HA-BH3-mutated MAP4 (Glu) recombinant adenovirus construction and transduction The HA-MAP4 (S737/760 A, Ala), HA-MAP4 (S737/760E, Glu), HA-LIR-mutated MAP4 Glu (F47/L50A and S737/S760E), and HA-BH3-mutated MAP4 Glu (L83/G87/D88A and S737/S760E, Glu) recombinant adenovirus were constructed by and purchased from Shanghai GeneChem, Co. Ltd. (Shanghai, CHN). The corresponding CMV-null adenoviruses were used as negative controls. Especially, MAP4 (Glu) and MAP4 (Ala) are used to mimic the hyperphosphorylation and dephosphorylation of MAP4. Each treatment related to these adenoviruses was repeated 5 times, and the sample size of each time was 3. Western blotting (WB) analysis The protein samples of mouse skin or epidermal keratinocyte extracts were obtained using a total protein extraction kit (Beyotime, CHN). And then, the protein concentrations were detected using the Bradford Protein Quantification Kit (500-0205, Bio-Rad Laboratories). The samples were gone through the SDS-PAGE system and transferred to PVDF membranes (Millipore, USA). The membranes were incubated with specific primary antibodies overnight at 4 °C. The primary antibodies utilized were as follows. MAP4 (1:1000, Affinity, USA), p-MAP4 (1:1000, GL Biochem, CHN), p-MAP4 (S760) (1:1000, GL Biochem, CHN), p-MAP4 (S737) (1:1000, GL Biochem, CHN), LC3 (1:1000, sigma, USA), p62 (1:1000, sigma, USA), Beclin1 (1:1000, CST, USA), Bcl-2 (1:1000, CST, USA), VDAC1 (1:1000, Millipore, USA), Tomm20 (1:1000, Affinity, USA), TIM23 (1:1000, Affinity, USA), GM130 (1:1000, Affinity, USA), Calnexin (1:1000, Affinity, USA), HA-tag (1:1000, proteintech, USA) and β-Actin (1:1000, proteintech, USA). Then, corresponding secondary antibodies were used and the results were visualized using the ChemiDoc XRS System (Bio-Rad Laboratories). Each treatment was repeated 5 times, and the sample size for each time was 3. Immunofluorescence analysis To detect autophagy/mitophagy status and cell proliferation in keratinocytes, the cells cultured on glass coverslips were transfected with GFP-LC3, Dsred-mito, or mRFP-GFP-LC3 adenovirus for 24 h or BeyoClickTM EdU-488 for 2 h, and then the cells were fixed in 4% paraformaldehyde for 25 min which was followed by blocking with 10% goat serum for 1 h. Next, the cells were stained with DAPI for 1 h at room temperature. Finally, cells were imaged using confocal microscopy (Leica Microsystems, Wetzlar, Germany). Each treatment was repeated 5 times, and 3 samples each time. Transmission electron microscope (TEM) The procedure of this part was according to the previous experiment [18]. After fixed in 2.5% glutaraldehyde, HaCaT cells, primary MKs, and skin tissues were successively subjected to dehydration, vibratome sliced and recut on a microtome and stained with uranyl acetate and lead citrate overnight. The sections were visualized through a TEM (JEM-1400 plus, Japan). Each treatment was repeated five times, and the sample size of each time was 3. Co-immunoprecipitation To address the interaction between MAP4/Bcl-2 and Bcl-2/Beclin1, co-immunoprecipitation (co-IP) was conducted. Cells were lysed in RIPA buffer with protease inhibitor tablets. The MAP4, Bcl-2, or Beclin1 antibodies were incubated with cell lysate for 24 h at 4 °C, then the samples were precipitated with protein A/G-Sepharose (Santa Cruz, USA) overnight at 4 °C. The precipitates were washed five times with PBS at 0 °C and separated by SDS-PAGE and probed by rabbit anti-MAP4 (1:1000, Affinity, USA), anti-Bcl-2 (1:1000, CST, USA), and anti-Beclin1 (1:1000, CST, USA) using WB. Each treatment was repeated five times, and three samples each time. Yeast two-hybrid (Y2H) screening systems Y2H systems were performed according to the Matchmaker Gold Yeast Two-Hybrid System User Manual. In brief, the entire MAP4, mutated MAP4 with knocking out L83/G87/D88 and entire Bcl-2 coding regions were successively amplified by PCR and respectively inserted in the pPR3-prey or pBT3-bait to generate a fusion with the GAL4 DNA binding domain (BD) or DNA activation domain (AD). The direct interaction of two proteins was investigated by co-transformation of the pBT3-STE-Bcl-2 and pPR3-N-MAP4 plasmids into the yeast strain NMY51, followed by dilution and selection of transformants on DDO, TDO or QDO plates at 30 °C for 3 days for growth selection. DDO, TDO, and QDO are used to detect bait plasmid toxicity and self-activation. Each treatment was repeated 3 times. Mitochondria fractions preparation Mitochondrial fractions were prepared and validated from HaCaT cells according to the manual instruction of the Cell Fractionation Kit (Abcam, USA). Briefly, HaCaT cells were collected, counted, and diluted to 6.6 × 106 cell/mL. Then, after being centrifuged at 500×g at 4 °C for 1 min, the supernatant was collected and again centrifuged at 10,000×g at 4 °C for 1 min. Then, the pellets were collected and resuspended, and after the same centrifugation process, the supernatants containing mitochondrial proteins were collected. Scratch wound healing assay Monolayers of cells plated in six-well plates were wounded by a 10-μl plastic pipette tip after being incubated at 37 °C for 2 h and then rinsed with medium to remove cell debris [2]. The wound healing process was pictured with an inverted light microscope (Olympus, Japan) at 0 and 24 h. Cell migratory capacity was defined as the wound closure rate (%), which was analyzed using NIH ImageJ software. Each treatment was repeated five times, and the sample size of each time was 3. Single-cell motility assay and quantitative analysis HaCaT or MKs at a density of 0.5 × 104/cm2 were seeded into 24-well plates. Then, time-lapse imaging was performed with a Zeiss imaging system (Carl Zeiss Meditec, Jena, Germany), which maintained 37 °C and 5% CO2. The images were captured every 5 min for 3 h. Later, cell trajectories were analyzed by tracing the position of the cell nucleus at frame intervals of 10 min using NIH ImageJ software. The velocity (μm/min) of each cell was defined as the total length (μm) of the trajectory divided by time (min), which reflected cell motility. Each treatment was repeated five times, and the sample size of each time was 3. GST pull-down assay GST pull-down assays were performed by GeneCreate Biological Engineering CO., Ltd (Wuhan, China). In brief, the LC3, MAP4-1 (1–170 aa), MAP4-2 (398–547 aa), MAP4-1 LIRΔ34,37, or MAP4-1 LIRΔ47,50 genes were first inserted into PGEX-6P-1 and pet-sumo vectors. The five recombinant plasmids were subsequently expressed in Escherichia coli, followed by the purification of GST-LC3, HIS-MAP4-1, HIS-MAP4-2, HIS-MAP4-1 LIRΔ34,37, and HIS-MAP4-1 LIRΔ47,50. The GST pull-down was conducted according to a previously published protocol [22], and measured via WB. Each treatment was repeated three times. Statistical analysis All data were presented using the mean ± SEM. The statistical analysis was performed using SPSS, v. 22.0. The independent-sample t-test and the one-way analysis of variance were applied for comparisons between two groups or more than two groups, respectively. P < 0.05 was used as the threshold to define statistical significance. The error bars are generated according to SEM values. Supplementary information Supplementary Figure Legends Figure S1 Figure S2 Figure S3 Figure S4 Figure S5 Figure S6 Supplementary information The online version contains supplementary material available at 10.1038/s41420-023-01465-3. Author contributions Conceptualization, YHF, XL, and LFL; Methodology, QZ and DXZ; Software, DXZ, YQH, and YSH; Validation, JHZ and YHF; Formal analysis, YHF, JHZ, and LFL; Investigation, LFL; Resources, YHF; Data curation, YHF and QZ; Writing—original draft preparation, YHF; Writing—review and editing, LFL and GXL; Acquisition, XL and LFL. Visualization, JYH and YSH; Supervision, JYH and GXL; Project administration, JYH and GL; Funding acquisition, LFL and XL. Funding This study was supported by the Young Scientists Fund of the National Natural Science Foundation of China (No. 82003323), the Military Medical Innovation Program in Daping Hospital (2019CXJSC014), and the National Natural Science Foundation of China (No. 81773348). Data availability All data we used were contained in the manuscript. Competing interests The authors declare no competing interests. Ethics In animal and cell experiments, the random grouping was adopted without deliberate selection of cell or mouse from growth status, body weight, age/culture time, or other aspects which will influence experimental results. Besides, double-blind was also utilized. In detail, experimental operators Yanhai Feng and Qiong Zhang were respectively responsible for the experimental grouping and the detection of experimental indicators. In animal experiments, the mice which were with disqualified age, weight, gender or living condition, etc. were excluded. And in cell experiments, the cells which were polluted, or with abnormal morphology or growth condition, etc. were excluded. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. These authors contributed equally: Yanhai Feng, Lingfei Li, Qiong Zhang. These authors contributed equally: Xia Lei, Jiongyu Hu, Gaoxing Luo. Change history 7/12/2023 A Correction to this paper has been published: 10.1038/s41420-023-01498-8 ==== Refs References 1. Charafeddine RA Nosanchuk JD Sharp DJ Targeting microtubules for wound repair Adv Wound Care 2016 5 444 54 10.1089/wound.2015.0658 2. Zhang JH Li LF Zhang Q Wang WS Zhang DX Jia JZ Microtubule-associated protein 4 phosphorylation regulates epidermal keratinocyte migration and proliferation Int J Biol Sci 2019 15 1962 76 10.7150/ijbs.35440 31523197 3. Hu JY Chu ZG Dang YM Yan H Zhang Q Liang GP The p38/MAPK pathway regulates microtubule polymerization through phosphorylation of MAP4 and Op18 in hypoxic cells Cell Mol Life Sci 2010 67 321 33 10.1007/s00018-009-0187-z 19915797 4. Hu JH Chu ZG Zhang Q Zhang D Dang Y Ren J Phosphorylation-dependent mitochondrial translocation of MAP4 is an early step in hypoxia-induced apoptosis in cardiomyocytes Cell Death Dis 2014 5 e1424 e1424 10.1038/cddis.2014.369 25232678 5. Qiang L Yang SG Cui YH He YY Keratinocyte autophagy enables the activation of keratinocytes and fibroblasts and facilitates wound healing Autophagy. 2021 17 2128 43 10.1080/15548627.2020.1816342 32866426 6. Klionsky DJ Autophagy: from phenomenology to molecular understanding in less than a decade Nat Rev Mol Cell Biol 2007 8 931 7 10.1038/nrm2245 17712358 7. Terešak P Lapao A Subic N Boya P Elazar Z Simonsen A Regulation of PRKN-independent mitophagy Autophagy. 2022 18 24 39 10.1080/15548627.2021.1888244 33570005 8. Choe SC Anne HB Brady NR Autophagy capacity and sub-mitochondrial heterogeneity shape Bnip3-induced mitophagy regulation of apoptosis Cell Commun Signal 2015 13 37 10.1186/s12964-015-0115-9 26253153 9. Zhang JH Zhang C Jiang XP Li LF Zhang DX Tang D Involvement of autophagy in hypoxia-BNIP3 signaling to promote epidermal keratinocyte migration Cell Death Dis 2019 10 234 10.1038/s41419-019-1473-9 30850584 10. Bellot G Medina RG Gounon P Chiche J Roux D Pouyssegur J Hypoxia-induced autophagy is mediated through hypoxia-inducible factor induction of BNIP3 and BNIP3L via their BH3 domains Mol Cell Biol 2009 29 2570 81 10.1128/MCB.00166-09 19273585 11. Chu CT Ji J Dagda RK Jiang JF Tyurina YL Kapralov AA Cardiolipin externalization to the outer mitochondrial membrane acts as an elimination signal for mitophagy in neuronal cells Nat Cell Biol. 2013 15 1197 205 10.1038/ncb2837 24036476 12. Marshall RS Hua ZH Mali SJ McLoughlin F Vierstra RD ATG8-binding UIM proteins define a new class of autophagy adaptors and receptors Cell 2019 177 766 81 10.1016/j.cell.2019.02.009 30955882 13. Velde CV Cizeau J Dubik D Alimonti J Brown T Israels S BNIP3 and genetic control of necrosis-like cell death through the mitochondrial permeability transition pore Mol Cell Biol 2000 20 5454 68 10.1128/MCB.20.15.5454-5468.2000 10891486 14. Kim HJ Shah MR Tu HC Jeffers JR Zambetti GP Hsieh JJD Hierarchical regulation of mitochondrion-dependent apoptosis by BCL-2 subfamilies Nat Cell Biol 2006 8 1348 1358 10.1038/ncb1499 17115033 15. Feng YH Li LF Zhang Q Zhang JH Huang Y Lv YL Microtubule associated protein 4 (MAP4) phosphorylation reduces cardiac microvascular density through NLRP3-related pyroptosis Cell Death Disco 2021 7 213 213 10.1038/s41420-021-00606-w 16. Li LF Hu JH He T Zhang Q Yang X Lan XD P38/MAPK contributes to endothelial barrier dysfunction via p-MAP4-dependent microtubule disassembly in inflammation-induced acute lung injury Sci Rep 2015 5 8895 10.1038/srep08895 25746230 17. Liu L Feng D Chen G Chen M Zheng QX Song PP Mitochondrial outer-membrane protein FUNDC1 mediates hypoxia-induced mitophagy in mammalian cells Nat Cell Biol 2012 14 177 85 10.1038/ncb2422 22267086 18. Li LF Zhang Q Zhang XY Zhang JH Wang XF Ren J Microtubule associated protein 4 phosphorylation leads to pathological cardiac remodeling in mice EBioMedicine 2018 37 221 235 10.1016/j.ebiom.2018.10.017 30327268 19. Doblado L Lueck C Rey C Arias AKS Prieto I Stacchiotti A Mitophagy in human diseases Int J Mol Sci 2021 22 3903 10.3390/ijms22083903 33918863 20. Liu L Sakakibara K Chen Q Okamoto K Receptor-mediated mitophagy in yeast and mammalian systems Cell Res 2014 24 787 95 10.1038/cr.2014.75 24903109 21. Kronlage M Dewenter M Grosso J Fleming T Oehl U Lehmann LH O-GlcNAcylation of histone deacetylase 4 protects the diabetic heart from failure Circulation. 2019 140 580 94 10.1161/CIRCULATIONAHA.117.031942 31195810 22. Huang Y Feng YH Cui L Yang L Zhang Q Zhang JH Autophagy-related LC3 accumulation interacted directly with LIR containing RIPK1 and RIPK3, stimulating necroptosis in hypoxic cardiomyocytes Front Cell Dev Biol 2021 9 679637 10.3389/fcell.2021.679637 34368130
PMC010xxxxxx/PMC10209928.txt
==== Front BMC Health Serv Res BMC Health Serv Res BMC Health Services Research 1472-6963 BioMed Central London 37231399 9349 10.1186/s12913-023-09349-z Research From idea to systems solution: enhancing access to primary care in Malawi van Niekerk L. [email protected] 12 Fosiko N. 34 Likaka A. 3 Blauvelt C. P. 5 Msiska B. 4 Manderson L. 6 1 grid.8991.9 0000 0004 0425 469X London School of Hygiene and Tropical Medicine, London, UK 2 Chembe Collaborative, Los Angeles, USA 3 grid.415722.7 0000 0004 0598 3405 The Malawi Ministry of Health, Lilongwe, Malawi 4 grid.517969.5 Kamuzu University of Health Sciences, Blantyre, Malawi 5 VillageReach Malawi, Lilongwe, Malawi 6 grid.11951.3d 0000 0004 1937 1135 University of the Witwatersrand, Johannesburg, South Africa 25 5 2023 25 5 2023 2023 23 5474 8 2022 28 3 2023 © The Author(s) 2023, corrected publication 2023 https://creativecommons.org/licenses/by/4.0/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Malawi, like many other countries, faces challenges in providing accessible, affordable, and quality health services to all people. The Malawian policy framework recognizes the value of communities and citizens, as co-creators of health and leaders of localized and innovative initiatives, such as social innovations. Social innovations involve and include communities and citizens, as well as bring about changes in the institutions responsible for care delivery. In this article, we describe the institutionalization process of a citizen-initiated primary care social innovation, named Chipatala Cha Pa Foni, focused on extending access to health information and appropriate service-seeking behavior. An interdisciplinary multi-method qualitative case study design was adopted, drawing on data collected from key informant interviews, observations, and documents over an 18-month period. A composite social innovation framework, informed by institutional theory and positive organizational scholarship, guided the thematic content analysis. Institutional-level changes were analyzed in five key dimensions as well as the role of actors, operating as institutional entrepreneurs, in this process. A subset of actors matched the definition of operating as Institutional Entrepreneurs. They worked in close collaboration to bring about changes in five institutional dimensions: roles, resource flows, authority flows, social identities and meanings. We highlight the changing role of nurses; redistribution and decentralization of health information; shared decision-making, and greater integration of different technical service areas. From this study, the social innovation brought about key institutional and socio-cultural changes in the Malawi health system. These changes supported strengthening the system’s integrity for achieving Universal Health Coverage by unlocking and cultivating dormant human-based resources. As a fully institutionalized social innovation, Chipatala Cha Pa Foni has enhanced access to primary care and especially as part of the Covid-19 response. Keywords Social innovation Primary care Health systems Institutionalization Africa issue-copyright-statement© BioMed Central Ltd., part of Springer Nature 2023 ==== Body pmcBackground The Astana Declaration on Primary Health Care [1] provides a blueprint for partnerships in health with non-state actors and is built on the renewed and growing interest in innovative approaches and strong and inclusive systems responses to achieve Universal Health Coverage (UHC). To realize the goal of UHC and equitable primary health services for all, innovative localized approaches, sensitive to the realities of different social and cultural contexts, are needed [2, 3]. The WHO Framework on Integrated, People-Centered Health Services [4] emphasizes the importance of innovative primary care models, and of engaging all individuals, communities, and civil society in co-producing and being accountable for health. However, despite communities being noted as essential for the achievement of UHC, the full engagement of communities in the design, implementation, management, and monitoring of interventions remains limited [5, 6]. Contrary to the rhetoric, programs aimed at enhancing community and public engagement are mostly top-down prescriptions, with care at best ‘co-produced’ under the guidance of an external expert, without community leadership and ownership [7]. To achieve true co-production of health, a shift is required from services being delivered to communities and citizens, to communities and citizens being recognized for their inherent strengths, given mutual responsibility, and capacitated to co-deliver care [7]. As emphasized by Odugleh-Kolve et al. [8], not only is it a fundamental responsibility for health systems to strengthen the dynamic interrelationship with patients, communities, and stakeholders, but also recognition needs to be given to their emotional, mental, and social capital which can contribute to care delivery. Like many other African nations, Malawi faces challenges in providing accessible, affordable, and quality health services amidst a dual burden of infectious and non-communicable diseases [9]. Three key documents underpin the country’s approach to health care. First, the constitution enshrines the public provision of healthcare as a right for all Malawians [10]. Second, two health policies, the Health Sector Strategic Plan II (2017–2022) [11] and the National Community Health Strategy (2017–2022) [12], express the country’s commitment to achieving UHC, and these policies guide the implementation of health interventions within the context of the decentralized health system. Although the country has no social health insurance fund, a basic health package has been adopted [11], comprising a select number of services provided free of charge at all government health facilities. However, multiple challenges still hinder this package from being accessible and available to all Malawians. Geographic access to care is limited in rural areas although 76% of the population lives within an 8 km radius of a health facility (2016 data), and 15% of Malawians report that they are still unable to attend to their medical health needs, often due to the cost of care [13]. Women are most limited in access to care due to cost, geographic distance and associated travel time, and low levels of education [14, 15]. Facility-level issues including poor attitudes of health workers, limited availability of medicines, and long waiting times [16–18], also affect people’s use of health services. These issues are exacerbated by Malawi’s acute shortage of health professionals, which meets only 48% of national targets [19]. This equates to 1.48 health workers per 1000 population [9], far below the World Health Organization’s recommendation of 4.45 health workers per 1000 population [20]. Malawi depends on non-state actors (within both non-profit and for-profit organizations) to complement the government delivery of health services in the country, with the Christian Health Medical Association being the largest non-state provider (29%) [11]. Within this arrangement, the Malawi Ministry of Health (MoH) takes on a stewarding role, providing policy guidance and technical support. Service delivery, planning, budgeting, and monitoring are devolved to district-level structures and partners. The MoH coordinates partners through the Health Sector Working Group (HSWG), comprised of all cross-sectoral stakeholders including donors and universities. Technical Working Groups (cross-departmental and cross-sectoral thematic groups) provide technical input on various health focus areas. A key goal behind the decentralization of the Malawi health system was to achieve greater citizen participation in public affairs and policy. The National Community Health Strategy (2017–2022) [12] highlights the important role of communities as users, providers, and accountability-holders of health services. Thus, the policy framework is in place to support citizen and community-led innovative initiatives in health. Social innovation to extend access to primary care information Social innovation has been presented, in both academic and policy discourse, as a promising alternate and complementary approach to achieving transformation, especially in systems plagued with complex challenges, convoluted overlaps in authority, and multiple players operating at different scales [21]. Social innovation is described as an ‘agentic, relational, situated, and multi-level process to develop, promote and implement novel solutions to social problems’ [22]. Social innovation results in institutional and systemic change [23], in five dimensions [24]: roles (creating new roles or breaking down role functions); resource flows (leveraging hidden resources or decentralizing resource distribution); authority flows (opening up the decision-making space); social identities (forming new identities and promoting interaction) and meanings (encouraging more ‘whole person’ or whole system purposes). However, for social innovation initiatives to achieve sustained systems strengthening and systems change, these initiatives need to be embedded or institutionalized across all levels of the system: at a micro level, accepted by individuals or groups; at a meso-level incorporated into organizational structures; and at a macro-level, accepted as part of the taken-for-granted structures of the health system [22, 25]. In 2009, Concern Worldwide International, supported by the MoH, turned to the citizens of Malawi to crowdsource, or solicit new ideas from the public to improve maternal healthcare. At the time, Malawi had one of the highest maternal mortality ratios in Africa. The crowdsourcing competition, ‘Share an Idea, Save a Life’, led to 6047 submissions for a grand prize of $10,000. Following several stages of selection, two young Malawians with a background in information technology, Soyapi Mumba and Clement Mwazambumba, were awarded for their winning idea to provide Malawian mothers and others in their communities with health information via their mobile phones. Their intention was to bridge the gap between the community and health facilities, caused by significant distances, long waiting times, and a lack of personalized care. These citizen innovators hoped that by connecting health providers and mothers through mobile technology (voice and text messages), available accurate and timely health information would help women make better decisions about their own and their baby’s health. This was a radical idea: by 2020, mobile phone coverage had risen to 52.3 per 100 inhabitants, but in 2008, mobile phone subscription coverage was still only 10.9 per 100 inhabitants, and even less in rural areas [26]. In 2011, this idea, named ‘Chipatala Cha Pa Foni (CCPF)’ (Health Centre by Phone) was pilot tested and implemented by an international non-governmental organization (NGO), initially in a single district in Southern Malawi, where a central call center was established within the local government district hospital. Following a positive impact evaluation in 2013 [27], CCPF’s geographic reach was extended, and by 2016, it covered eight districts [28]. Although the initial focus was on maternal and child health, by this time, several additional health focus areas were added. In 2017 the initiative was officially adopted by the MoH, through a signed memorandum of understanding, outlining a vision to institutionalize CCPF within the public health system and serve Malawians across all 28 districts (See Fig. 1).Fig. 1 CCPF over time In this article, we present findings from studying how CCPF was institutionalized, and the institutional level changes that were required by this citizen-created initiative to embed it within the Malawi public health system. Methodology A multi-method qualitative case study design was selected. This methodology is well-recognized within health policy and systems research [29] as it enables the study of ‘open systems’ where variables are not linear, and neither the phenomena nor changes over time can be controlled [30]. Social innovation is an evolving process and is highly context-bound, making a case study methodology an ideal approach [31]. This case was purposefully selected from a database on social innovations [32], because it had been scaled up and institutionalized within a low-and middle-income country. Data collection A mapping exercise was undertaken to identify all relevant actors involved in the institutionalization process, and they were approached for semi-structured in-depth interviews. Interviewee categories included: implementers (creators, implementers, community collaborators); government (national and district level representatives); contributors (project partners); and other national actors who have been involved in innovation. As data collection progressed, additional participants, mainly government actors, were invited, with attention given to seeking out participants who may hold contradictory or critical opinions on the institutionalization process. A total of 68 interviews were conducted with 54 participants. Supplementary data included organizational documents (evaluations, presentations, promotional materials), monthly progress reports, and health system documents (policies). Further, observations paid attention to group processes, day-to-day management processes, actor roles, reactions, and contributions. Data collection took place in three rounds over 18 months, with 1 year of intensive engagement (July 2018 – July 2019). Interviewing and other data collection ceased when saturation was reached, and no new information was shared. Data analysis Data collection commenced in the field and interim analyses were conducted between fieldwork intervals to enable iteration and triangulation. All interviews, except one, were audio-recorded and transcribed. The majority of interviews were conducted in English and selected interviews conducted in Chichewa (the local language widely spoken) with the help of a translator were transcribed and then translated. Field notes were typed up. All data were imported into NVivo 12, qualitative analysis software [33]. Based on the understanding of social innovation as a process occurring across multiple levels [22], a composite framework was developed that examined the micro-individual level, the meso-organizational level, and the macro-institutional level (see Fig. 2). This framework accounting for these levels of action and change was further informed by different theories from neo-institutionalism, including institutional entrepreneurship [34] and institutional work [35, 36] (purposeful actions to bring about change), and merged with work from positive organizational scholarship [37, 38]. By combining these theories, it was possible to understand how the agentic efforts of individuals, within a resource-constrained institutional context, can lead to the creation, disruption, and maintenance in five performative institutional dimensions: roles, resource flows, authority flows, social identity, and meanings). A thematic content analysis was conducted using deductive (guided by the conceptual framework) and inductive approaches – allowing for the recognition of patterns, whereby themes (or codes) that emerge from the data subsequently become the categories for analysis [39].Fig. 2 Social Innovation Framework [22] Ethics This study received ethical clearance from the Ethics Review Board at the London School of Hygiene and Tropical Medicine and the National Council for Science and Technology in Malawi. Informed consent was obtained from all the study participants. Consent forms were made available in English (the government and business language) and Chichewa (the local language widely spoken). All methods were conducted in accordance with the approved ethical consent received for this study. Results The role of actors, namely institutional entrepreneurs, their efforts, and the resulting institutional changes will be discussed below. Institutional entrepreneurs Social innovations redefine boundaries and blur lines across different sectors and society [40, 41], as evidenced by CCPF. To institutionalize CCPF, organizations, and entities from across the government, the non-state sector, and the community were involved (Fig. 3). The implementing NGO served as a ‘bridging organization’ [42], a broker between the diverse interests, visions, and resources of the different constituent organizations involved. The implementing NGO was constant throughout the CCPF’s lifecycle from the pilot to its institutionalization, even though the level of involvement of the other organizations fluctuated over time.Fig. 3 Organizations involved in the CCPF institutionalization process Of the 54 actors interviewed, a subset of individual actors played a particularly important role. These were institutional entrepreneurs (IEs), defined by Battilana [43] as ‘‘actors who initiate divergent changes in the institutional context and who actively participate in the implementation’. Nine actors met this definition: three from the non-governmental sector (implementing NGO), five from the government sector (MoH), and one from the community sector (Traditional Authority). Each, in their distinct capacity, played an agentic role between 2012 and 2019, either in facilitating the adoption or in institutionalizing CCPF. Three of the nine IEs were female and all, but one were Malawian nationals. Their backgrounds included nursing, clinical medicine, project management, and tourism. All except the community leader, held mid-level positions within their organization and they were able to make decisions or influence decisions on resources and policy adoption. Government IEs had similar career trajectories, starting at the frontlines of care delivery, holding management positions at the district level, and then being promoted to a national-level position. Compared to the NGO IEs, government IEs had a longer organizational tenure (14–26 years), which provided them with strong professional connections and alliances. Other factors which influenced their receptivity to innovation and their willingness to take on risks were attributed to their experiences gained outside of Malawi, and their prior engagement in ‘innovative’ projects (projects deviating from the norm or standard practice) in previous work settings. Despite diverse career backgrounds and trajectories, IEs had a set of personal characteristics in common. All nine IEs displayed the projective dimensions of agency [44], that is, they were future-oriented or visionary. They were able to see possibilities beyond the reality of resource constraints and challenges, and this ability aligned with the positive emotion of hope. They believed in their own power to realize an innovative possibility in which they held an interest.“Quite often people talk about the Ministry of Health or the health system in Malawi operating in a resource-limiting environment. Though I see that sometimes, I don’t believe it is resource-limited. It just requires innovation to see how best you can do with the little you have. Some people when they see problems, they just look at them and cry. Others look at problems as opportunities, they can do something.” (Interviewee 015 – Government) “There are people who would not agree and there are people who would even discourage you but when I see that there is a positive side to things, I always carry on. I am never discouraged by people that just make arguments for argument’s sake.” (Interviewee 010 – Government). The IEs displayed and articulated their deep passion for social change and positive impact on the lives of others around them. From observations and interviews, it appeared that this passion was not linked to specific projects but was rather something they considered as a personal characteristic. This was complemented by considerable energy and enthusiasm for life in general, and this translated to their work. They attributed their passion to childhood experiences or their Christian faith, which gave their lives a unique purpose and vision.“I have a passion to serve life…not serve life at a small scale but my scope is the biggest impact, that’s why I made an oath, as a medical doctor, I will work in the public service. I want to die empty…the potential in me…technically, whatever I am…God enacted in me, God who created me…. Whatever potential is in me, I will make use of it to the best of my capability. I don’t believe in what I get in my pocket but what I can contribute. That’s my passion.” (Interviewee 038 – Government) The IEs did not view themselves as lone heroes, but they took pride in their relational skills, collaborating with others, and building strong personal relationships. They were willing to engage in discussion, question, and debate with their colleagues. They regarded their colleagues, irrespective of their social position or status, as sources of knowledge, and believed that broad participation in processes was of great value. Each IE openly embraced the opportunity to gain new experiences and knowledge:“I respect each and every other person’s views. When you listen attentively there is always something you can learn, and I always view it that there is nobody who is a blank slate. We are born with some knowledge, we have some information, and you should be objective and not judgmental. I have seen those innovations indeed work, so I am positive about constructive suggestions from colleagues. Maybe that is by nature who I am.” (Interviewee 015 – Government) These actors initiated a series of changes in five institutional dimensions that supported the Chipatala Cha Pa Foni initiative to be embedded in the national public health system. These were: Roles, Resource Flows, Authority Flows, Social Identities and Meanings. Roles For the innovation to work, a change in the institutionalized role of nurses was required. During the innovation pilot in the catchment implementation district, Health Surveillance Assistants (HSAs—community health workers) were initially appointed to attend to incoming calls. In the Malawian health system, HSAs are responsible for health promotion, disease surveillance, and basic service delivery. As part of CCPF, HSAs would provide health information and advice via mobile phones. The Nursing and Midwives Council of Malawi, initially resisted this, believing that a virtual patient-provider interaction held a higher risk than a face-to-face interaction and that the function should only be provided by a registered nurse. To keep the innovation operational, the implementing NGO appointed nurses to replace the HSAs that operated the hotline. However, even with nurses providing this service, there remained significant institutional resistance and opposition, at the time of data collection in 2019. Several interviewees felt that moving the nurse from the bedside to the telephone would compromise nurse clinical skills, knowledge, and attitudes. As one interviewee explained:“Yeah, I had so many concerns at first because you know, we associate nurses with only the clinical aspect, so people started having worries, what about their clinical skills, how will they aah be up to date in terms of skills.” (Interviewee 020- Regulator) No clear policy framework governed m-health or telemedicine initiatives in Malawi when the CCPF pilot commenced. Regulators were concerned that misinformation and incorrect decision-making would follow from the inability of the nurse to see a patient face-to-face to observe physical signs of illness, and their reliance only on dialogue. To mitigate this, it was deemed that CCPF nurses could not provide formal consultations but only provide health information and refer the caller to a health facility. A final and ongoing concern expressed by interviewees regarding the role transformation was that health facilities in Malawi already experienced acute staff shortages and that nurses involved in the innovation were lost from the frontlines. Other interviewees however regarded this as a strength of the innovation, arguing that nurses positioned in a hotline, providing information telephonically, could serve more people and do so more efficiently. Despite these objections, two factors enabled this institutional change to happen. First, the implementing team drew on the existing legitimate function of nurses to provide health promotion and education. Second, they emphasized the limitations hindering access to health care experienced by Malawians: geographic distances, the over-congestion of health facilities, and limited (if any) patient-provider consultation time. The CCPF theory of change was that access to primary care can be improved by providing timely health information, thus ensuring that care is sought appropriately while avoiding unnecessary consultations. One IE explained how he advocated for this change:“I thought they (CCPF) were reaching more people, without people having to walk to the hospital because we haven’t yet reached these people. You know maybe some have 15 km, others have 10km (to walk), so I find it very interesting. At least we are assisting many people and we are trying to minimize congestion at the hospital as they will be in the queue to be told it is just minor.” (Interviewee 014 – Government) Although both the Nursing and Midwives Council and the Ministry of Health (MoH) approved the plan to transform the role of nurses in 2011/12, the challenge associated with this institutional change surfaced again later. As part of the institutionalization process (2017–2020), the MoH needed to incorporate the hotline nurses as government health workers, rather than having them employed directly by the implementing NGO. Within the government staff delineations, there was no staff establishment for nurses operating in a virtual advisory capacity. While this was overcome in December 2020, it was a major barrier, to institutionalization:“To train a health worker here is expensive. Should they be answering phones when we have a high vacancy rate in our hospitals? My thinking said ‘ah, are we ready for this?!”. It is a very good initiative, yes, but should government actually start creating positions for this when we are failing to address the challenges in our hospitals?” (Interviewee 052 - Government) Resource flows Eighty-four percent of Malawians live in rural areas and face multiple socio-economic challenges including poverty, a lack of income-generating opportunities, electricity shortages, limited mobile phone coverage, internet access, and low literacy. Deeply embedded traditional and patriarchal structures influence women’s access to health services. Before the implementation of CCPF, health information, especially in rural areas, was regarded as a privileged resource, owned by health professionals and only accessible to people visiting a health center. With the implementation of CCPF, access to health information became more widely distributed. First, information was sent to users (via text messages) and users could call the hotline free of charge with specific health questions. Second, the hotline was later open to all Malawians, not just mothers or caregivers. One of the main Malawian private telecommunications providers established a dedicated phone number to be used free of charge by people calling the hotline and enabled text messages, customized to the user’s specified health interest, to be sent to thousands of people simultaneously. Both interview and observation data, as well as impact evaluation data, provided evidence of the effect of the enhanced distribution of health information on community actions and agency. In the traditional authority (TA) area visited during fieldwork, the traditional leader explained that the innovation had become part of his strategy to improve the health of the 70,000 people over whom he presided, including to reduce maternal mortality. In this area, few mobile phones existed within the community, and these were mainly owned by men. To overcome this, the leader raised funds to purchase several mobile phones, and he provided these to selected women in the community called ‘secret mothers’. These women had a role to identify pregnancies in the community and provide a mobile phone to any woman who wanted to call the hotline. In addition, the health information communicated via text messages was written onto the walls of buildings all over the village. In this way, the traditional leader was able to disseminate information to raise broad awareness about antenatal care, and so address selected cultural beliefs and practices that were responsible for poor maternal outcomes. The leader simultaneously adopted bylaws to hold the community accountable when they were not adhering to these healthy practices.“CCPF simplified the role of traditional leaders. Instead of us moving around in the community, telling pregnant women to go and deliver in the hands of skilled personnel, to start ANC in the first trimester, telling them whatever message that we want to pass on to the women, was now simplified because most of the women could call the CCPF, then listen to the messages. It eased the workload for the personnel at the health facility, instead of attending a long queue of pregnant women who were just going to the facility just to ask for advice, now, those women could get the advice from the phones. It also gave a kind of empowerment to the pregnant women, to say, ‘when I get to the facility, what I wish to get is A,B,C,D just because they have already heard from the CCPF.” (Interviewee 019 - Traditional Leader) CCPF was one of the first initiatives in Malawi to leverage technology in support of primary health care provision, thus challenging the dominant community institutional logic on care provision. Communities were skeptical that phone calls to the hotline were free of charge, and that remotely located health professionals could provide women with information, such as their expected date of delivery. Community members associated this type of ‘future knowing’ with occultic fortune-telling practices and witchcraft. In response, community mobilization campaigns were strategically hosted by the implementing NGO to establish the innovation’s legitimacy:“The reaction was like, this is new. Globally m-Health is also very new, of course maybe now it can be 10-15 years, but we considered it new, and there were some questions from the communities for example: ‘how can a phone call be free?’. As you know Malawi phone calling tariffs are just very high. I remember facing a lot of challenges from the communities that this might be linked to these satanic things and all that. It was a bit challenging to make people fully understand what Chipatala cha pa Foni is, and we could see that people need to be oriented on what m-Health is about.” (Interviewee 007 - Implementer) The decentralization and re-distribution of health information, as a resource, led to greater distributed agency among community members, who when using the innovation, felt more competent to make informed decisions and improve health-seeking behavior. An independent impact evaluation [45] conducted in 2018 found that CCPF had a statistically significant impact on sexual reproductive health, HIV, maternal health, child health and nutrition indicators; that 87% of users reported following the advice provided; and that 87% appropriately sought care at a health facility when referred. Ninety-eight percent of users reported satisfaction with the service and half indicated that they would still access the service, even if they had to pay for use. Authority flows, social identity, and meanings Institutional changes occurred not only due to the innovation implementation but also due to the institutionalization process. CCPF did not follow a linear implementation and institutionalization process; rather, these processes occurred concurrently, evolving according to circumstance and opportunity. An institutionalization structure of key significance was the steering committee meeting, launched in 2016. It was led by several IEs and this structure supported institutional changes in three key dimensions: authority flows (who decides what), social identities (who belongs to what), and meanings (who signifies what). CCPF was one of the first technology-enabled cross-departmental, cross-boundary initiatives institutionalized in the country. The IEs driving the process were aware from the beginning that financial resources and authority flows are distributed across various health system actors, and thus they considered it important to move away from a top-down command-control approach of leadership and decision-making. They adopted a more collaborative approach with distributed decision-making. As one partner of the project explained:“Definitely the stakeholder collaboration is key. I think that there is a steering committee, and the NGO has been the glue that has kept all the stakeholders together, making sure the communication lines are open and make sure everyone is on the same page. I think that has helped to bring it (CCPF) to where it is now.” (Interviewee 026- Partner) The steering committee meeting structure unlocked the creativity needed to support a process of ‘innovating to institutionalize’. Observational and interview data showed that this second process of innovation involved less institutional disruption than occurred during the pilot. Rather, the institutionalization process focused on developing creative embedded strategies to support the introduction of CCPF within existing institutional boundaries and health system constraints. Within the public sector, system integrity had to be maintained and could not undergo a total redesign, as one government participant explained:“I think it is both: the innovation cannot be rigid, it has to be flexible, for example now, we are taking up Chipatala Cha Pa Foni as part of the Ministry of Health or the health systems in Malawi. There are policies, so the innovation cannot be exactly the way it was, it (CCPF) has to be modified a bit to suit the system but at the same time the system has to respond to welcome this, it (the system) has to adjust in some areas to make sure that this is set. So, it’s both ways.” (Interviewee 015 – Government) Specific practices were important in ‘facilitating changes in authority flows, social identities, and meanings. These practices, discussed further below, included: hosting a space that allowed for creative engagement [46, 47]; vertical and horizontal inclusivity [48] that gave all actors an equal opportunity for contribution; and shared leadership which allowed new IEs to play a role [49–51]. The IE who chaired this meeting was key to facilitating a shared space for discussion, collaboration, creativity, and relational engagement among various partners. Vertical and horizontal inclusivity was achieved by including all actors: representatives from 10 different technical departments within the MoH, from district to national level; different levels of project implementers (from the director to the hotline nurses); partners representing different sectors – private companies, bilateral development agencies, NGOs, academia, and professional associations; and beneficiaries. Any given meeting would be attended by 20 – 40 people. Meeting structures such as these were not unique in the Malawian health system, as multiple multisectoral technical working groups gather at regular intervals around topic areas. However, from observation data, the vertical and horizontal inclusivity across levels of hierarchy and power was unique:“But I learned from one of our steering committee meetings some people in the village have been using CCPF. It was very amazing to see that this person with his phone. He was able to give health messages to almost an entire village and the entire village was able to ask about anything that is related to health, by using his cell phone.” (Interviewee 018 – Government) The leadership style of the steering committee meeting chair did not project downward influence on followers (vertical leadership), but rather demonstrated shared leadership [49–51], allowing for two emerging IEs from the MoH to gain confidence and play a leading role in institutionalization. All steering committee meeting participants were given transparent access to information related to the innovation, such as actions implemented, the latest monitoring data, and challenges faced. Steering committee meeting participants at any level could propose ideas, suggestions, and solutions for how the innovation might continue to develop and adapt to become institutionalized. All were part of the decision-making process on the course of action. Steering committee meeting participants also all contributed resources within their realm of influence – personal or professional networks, financial or in-kind resources, and opportunities for advocacy and promotion.“The collaboration that is there between our partners, [the NGO] and Ministry, has strengthened because we can discuss together. If there are issues, we sit and discuss them among ourselves. Like I said, even the steering committee has incorporated various departments, so it is like everybody is aware. They are left out, and then they will be surprised ‘I don’t know that this is happening’ so there is such transparency and the involvement of different people.” (Interviewee 010 - Government) These practices supported the creation of a shared social identity and emotional solidarity among steering committee meeting participants, often talking about CCPF as ‘this is ours’. This new identity created by the innovation overcame the diversity and distinction between individual institutional identities and assisted with addressing implementation and institutionalization challenges.“It is through the meetings we had; I think we had the first meeting. I don’t know where I was first involved but I found it already there. My colleague was the first one to be involved in one of the meetings and I also have been to attend one of them. Ja. It was quite good. And after that, it is like we are moving together, with the clinical department. Ja, so we share ideas and when there is a meeting we go and then see how we are moving forward.” (Interviewee 009 – Government) The generation of positive emotions also supported the creation of a shared social identity. These positive emotions, such as hope and pride, were first identified within IEs, and they played a role in catalyzing and amplifying these emotions in the broader group. By March 2019, during the second data collection visit, it was clear that the MoH faced two major challenges in fully institutionalizing the initiative by the intended date of 30 June 2019. The first was difficulties in adding 29 additional nurses into the government establishment (i.e., having the necessary funding to create new positions), especially with the resource constraints of the Malawian government and budget cuts in the 2019 financial year. Second, the ongoing operational running costs of the innovation exceeded the funding available to the managing department”. During steering committee meetings, participants critically engaged to identify obstacles and express their own doubts. They were also able to reflect upon decisions that may have had negative and unintended consequences on the process. Yet, as illustrated below, all institutional entrepreneurs and the steering committee participants were resolute that institutionalization would eventually be possible:“There are two scenarios I am seeing because, I believe come July [the institutionalization deadline], we will not have a human resource ready. Come July we will not have financial resources. Come July we will not have good monitoring and evaluation. So, if handover indeed happens in July, then it will crash, and they will start picking it up from the ground. But for me, I think it’s a wonderful innovation and there is a lot of things that can be done. I just hope the government will give it all the support that is needed, I know some people are very interested and have ideas to take it to another level. I think we should expect that we will have challenges, but the good thing is to believe that if we fall, we will try again.” (Interviewee Government 045– District) This future, possibility-oriented and visionary view despite significant challenges conferred the positive emotion of hope. Unlike other positive emotions, hope does not arise in settings where circumstances are perceived as safe. Rather, it comes alive in circumstances when people might fear the worst but yearn for better [38]:“To me it is a mark for Malawi. It is something people said couldn’t work, it was a waste of time, a waste of resources but there it has worked. So, to me, it is possible, once you have enough information, recommendations, or findings that this can make a difference in society. Believe in yourself, believe in what you’re doing.” (Interviewee 051 - Government) “I think when we get to have a steering committee meeting and he (steering committee meeting chair) says for instance: ‘[name], I am not worried about money. I know this thing must run and it will run. Money will be made available, that is the least of my concerns’. At one point I was concerned that the budget that was passed represents, maybe 10% or somewhat of what it takes to run CCPF. But if he is quite confident to say that is not his problem, then why should it give me a headache. So, when you have those discussions with Dr, you go into the room nervous but as you are leaving the room, you go, I think I need to rethink it all and get to the level of confidence that he has.” (Interviewee 002- Implementer) The broader group of steering committee meeting participants rotated, and government representatives were often assigned to attend. Thus, it is unlikely that this group attracted high-hope individuals, such as IEs, by self-selection. Rather, hope was likely generated in the context of the steering committee meetings. In the broader group, a sense of hope and hopefulness was observed, as exemplified by this frontline nurse who attended the meeting and whose job security depended on successful institutionalization:“Yes! We have invested a lot and it will be bad to just give up now. We still have hope that it, though it will not go as we wanted, as we anticipated, we still have hope that it will go through” (Interviewee 003 - Implementer). Pride was also widespread among steering committee meeting participants at the group level. Members viewed the initiative as a way to positively impact their country and they felt that, as Malawians, they could then take credit for implementing an initiative that few other African countries have done. All actors engaged in the institutionalization of CCPF did so on a voluntary, non-renumerated basis, contributing over and above their daily work tasks and functions. Pride is an intangible resource, in the perseverance of ongoing efforts by actors [52].“To contribute something towards the development of my country, that is the most important thing, because when you see people that are using something to which you contributed, you feel good. You know, people are healthy because of my contribution. My country is developing because of what I did at the meeting. I would be very happy if the initiative is continuing because a lot of people can benefit from the service.” (Interviewee 051 – Government District) “I think seeing this project come from just being on paper and then we get feedback, ‘oh my gosh! we got 20-something thousand phone calls from this.’ And then the testimonies, meeting the people that this initiative has impacted is just so heart-warming and so fulfilling for me. I mean, come on! We can see the impact from this woman who was able to be assisted by just a phone call. So yeah, it is really awesome!” (Interviewee 026- Private partner) The final institutional change facilitated by CCPF was in terms of meanings—the explanations for the way things are [53]. Government health institutions are organized according to functional roles and disciplines e.g., a nursing department oversees functions relating to nurses. Yet, findings suggest that CCPF cut across siloed operational, organizational, disciplinary, and sectoral boundaries. By 2018, it had grown to become a comprehensive technology-enabled service delivered by nurses who provided information on all health topics, 24-h a day, 7-days a week, for all people (children, adolescents, working adults, elderly). The institutionalization of CCPF did not come without tension in meanings. This occurred in line with that which is expected from social innovation — approaches which are more whole-person and holistically oriented, contrasted with the traditional siloed (vertical) healthcare approach, organized according to functional areas. As noted, CCPF required the involvement of multiple departments within the MoH, to ensure that the service was implemented according to existing health policy standards and guidelines. A situation that caused significant tension in meanings, related to which department should organizationally ‘house’ this initiative such that leadership could be provided. CCPF was moved from one department to another within the ministry, but frequently respondents questioned whether it would not be better placed in a technical department dealing with health promotion. Several steering committee meeting participants proposed options such as creating a dedicated unit within the government structure. This was not resolved at the time of data collection completion, but as observed, it did stimulate government actors to reflect on an appropriate structure to support this social innovation:“To me, I think, integration is key because in a health facility, there are different players that are interested in their services that are being delivered. So, bringing those service providers together as what we have done, people from different departments coming together as one team to discuss on how we run this, it’s something that I feel is an innovation on its own.” (Interviewee 051B – Government) “So, from the way I understand that, at first, I think there were discussions around where CCPF should be positioned, should it be positioned at the district level, in this case, the Lilongwe DHO. Or should it be attached at the hospital level, at the Kamuzu Central Hospital? Or, if it should fall within the Clinical Services Directorate. So, I think members who were present at the transition meetings suggested that it shouldn’t be tied to any of these three, it should be an individual entity of its own. It is not going to be a department, but it is rather going to be like a unit.” (Interviewee 002 – Implementer). Discussion We have highlighted three contributions of social innovation initiatives to health systems: supporting greater access to primary care for all people; strengthening health system integrity; and nurturing human-based resources in health systems. We discuss these further below. The Chipatala Cha Pa Foni (CCPF) was a citizen-led idea that, through the support of an implementing NGO, improved access to health information for all Malawians. On completion of data collection in July 2019, the social innovation had not yet succeeded to be integrated into the Malawi Ministry of Health (MoH), although it was scaled to all 28 districts in Malawi. Full institutionalization was achieved in December 2020, and henceforth, the social innovation is fully owned and managed by the Malawi MoH. This primary care social innovation proved especially valuable during the Covid-19 pandemic when, now part of a newly created Emergency Medicine Unit, it became a key component of Malawi’s pandemic response, at a time when access to accurate health information was critical. The government adapted and utilized CCPF’s capabilities to triage people with Covid-19 symptoms, do contact tracing, patient monitoring, mapping patient journeys, educate and dispel Covid-19 myths and create referral linkages between the health system and social services. CCPF resulted in changes in five key institutional dimensions. These changes make it possible to classify this initiative as social innovation. The institutional changes did not seek to disrupt the health system; rather the initiative supported strengthening the system integrity. It did so in numerous ways: bringing new health system leaders to the forefront (institutional entrepreneurs); facilitating greater collaboration between government, citizens, and non-state actors; and building capacity for collective creativity. Health systems could thus approach social innovation not as a risk but as a way to strengthen systems integrity, as all the socio-cultural capacities generated have transferrable benefits by application in other existing or future health system initiatives. The socio-cultural systems change which occurred by actors collaborating on a tangible social innovation initiative can provide a subtle and arguably subversive way by which human and other dimensions of health systems can be nurtured, beyond only enhancing programmatic care delivery and health outcomes. This reduced concern by decision-makers that social innovation institutionalization would hinder and distract from ongoing agendas and efforts to build stronger health systems. Rather it provided direction to social innovation’s complementary contribution in institutional strengthening and supporting the achievement of people-centered health systems. Social innovation also cultivated human-based resources in health systems such as positive emotions. Institutional entrepreneurs matched up with Snyder’s understanding of hope as “the sum of the mental willpower and waypower that you have for your goals” [54]. As above, all had mental willpower—the perception that they could initiate actions to achieve the goal—and waypower—the ability to find alternate routes towards achieving their goal if obstacles presented. These qualities identified in the data also differentiate hope from optimism. Optimistic individuals hold fast to the belief of a positive outcome but lack critical thinking about how to arrive there. However, the CCPF actors, especially the nine institutional entrepreneurs, can better be described as high-hope individuals [54, 55], who could influence a larger group, as they unlocked and raised hope levels dormant within a larger group. Raising collective hope levels (future-oriented agency) subsequently became a catalyst towards further action (habitual and practical evaluative agency) and relentless determination, as required to institutionalize a social innovation [44]. The nurturing ‘ground’ for hope came from shared interactions and experiences created by the steering committee meetings. In pressured health system contexts, many meetings may occur, but the components of vertical and horizontal inclusivity, transparency, shared leadership, and creative participation gave way to relational bonds strengthened between members of the steering committee. Through high-hope individuals, in relationship with other members, hope was injected into the larger group. This sustained the ongoing efforts to institutionalize the social innovation, despite the challenges. This created a more positive experience for Malawi nationals, in contrast to their usual experience when innovative initiatives are implemented from outside the system and by non-nationals. The social innovation process, if applied well and with sensitivity, could be an alternate way to overcome some of the past and present colonizing implementation practices. Conclusion Social innovations are initiatives that bring about changes within the institutional and social-cultural dimensions of health systems. In terms of institutional change, the Chipatala Cha Pa Foni, as a social innovation led to a change in the roles of nurses and a change in the flow of health information. In terms of social-cultural change, the role of actors, operating as institutional entrepreneurs, within the social innovation institutionalization process, played a key role. These actors strategically drove the social-cultural change through engaging in shared leadership, creative collaboration, and drawing on their inherent human-based resources such as hope. Institutionalized citizen-led social innovation receiving strong government support can be a practical way to improve primary care access for the achievement of Universal Health Coverage. Acknowledgements We thank the following people: from the Malawi Ministry of Health, Dr Fanny Kachale, Dr Rabson Kachala, and Mr Isaac Dambula; from Kamuzu University of Health Sciences, Prof Don Mathanga and Dr Vincent Jumbe; from VillageReach Malawi, Upile Kachala, Lucky Gondwe and Dr Alinafe Kasiya; and from the London School of Hygiene and Tropical Medicine, Prof Lucy Gilson and Prof Dina Balabanova Authors’ contributions LvN conceptualized this research.LvN and LM prepared the first draft of the manuscript. NF, AL, CPB and BM each provided input and contributions to the various manuscript drafts. All authors approved the manuscript. Funding This research was made possible through funding from the Commonwealth Scholarship Fund and the London School of Hygiene and Tropical Medicine Travel Scholarship. Availability of data and materials The datasets generated and analyzed during the current study are not publicly available due to the confidentiality risk posed to participants. The datasets generated and analyzed during the current study are not publicly available due to the confidentiality risk posed to participants. The datasets used and/or analysed during the current study available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate Research ethics approval for this study was obtained from the London School of Hygiene and Tropical Medicine’s Research Ethics Committee – Reference 15476, Date: 29 June 2018. In-country research ethics approval was obtained from the Malawi National Commission for Science and Technology – Reference NCST/RTT/2/6, Date: 25 May 2018. This study was conducted in partnership with researchers from the University of Malawi’s College of Medicine Social Innovation in Health Initiative, under the Malaria Alert Centre. Informed consent was obtained from all study participants. Consent forms were made available in English and Chichewa. All methods were conducted in accordance with the approved ethical consent received for this study. Consent for publication Not applicable. Competing interests The authors declare no competing interests. The original version of this article was revised: an error in the name of the fourth author was corrected. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Change history 7/12/2023 A Correction to this paper has been published: 10.1186/s12913-023-09798-6 ==== Refs References 1. World Health Organization. Declaration of Astana. Geneva: World Health Organization; 2018. Available: https://www.who.int/primary-health/conference-phc/declaration. 2. Olu O Drameh-Avognon P Asamoah-Odei E Kasolo F Valdez T Kabaniha G Karamagi H Good S O’Malley H Yoti Z Community participation and private sector engagement are fundamental to achieving universal health coverage and health security in Africa: reflections from the second Africa health forum BMC Proc 2019 13 9 7 10.1186/s12919-019-0170-0 31737089 3. Agyepong IA Sewankambo N Binagwaho A Coll-Seck AM Corrah T Ezeh A Fekadu A Kilonzo N Lamptey P Masiye F The path to longer and healthier lives for all Africans by 2030: the Lancet Commission on the future of health in sub-Saharan Africa Lancet 2017 390 10114 2803 2859 10.1016/S0140-6736(17)31509-X 28917958 4. World Health Organization. Framework on integrated people-centered health Services. Geneva: World Health Organization. 2015. Available: https://apps.who.int/gb/ebwha/pdf_files/WHA69/A69_39-en.pdf. 5. George AS Mehra V Scott K Sriram V Chunharas S McIntyre D Community participation in health systems research: a systematic review assessing the state of research, the nature of interventions involved and the features of engagement with communities PLoS One 2015 10 10 e0141091 10.1371/journal.pone.0141091 26496124 6. Allotey P Tan DT Kirby T Tan LH Community engagement in support of moving toward universal health coverage Health Syst Reform 2019 5 1 66 77 10.1080/23288604.2018.1541497 30924744 7. Ocloo J Matthews R From tokenism to empowerment: progressing patient and public involvement in healthcare improvement BMJ Qual Saf 2016 25 8 626 10.1136/bmjqs-2015-004839 26993640 8. Odugleh-Kolev A Parrish-Sprowl J Universal health coverage and community engagement Bull World Health Organ 2018 96 660 661 10.2471/BLT.17.202382 30262948 9. World Health Organization: Malawi. Country cooperation strategy at a glance. 2018. 10. Republic of Malawi. Constitution Act. Malawi; 1996. 11. Government of the Republic of Malawi. Health sector strategic plan II. Malawi; 2017. 12. Ministry of Health: National Community Health Strategy 2017 - 2022 Integrating health services and engaging communities for the next generation 2017 Lilongwe Ministry of Health 13. The Program on Governance and Local Development. The Local Governance Perfomance Index (LGPI) Malawi: selected findings on livelihood. Gothenburg: Institute of Public Opinion and Research University of Gothenburg; 2016. 14. Azad AD Charles AG Ding Q Trickey AW Wren SM The gender gap and healthcare: associations between gender roles and factors affecting healthcare access in Central Malawi, June–August 2017 Arch Public Health 2020 78 1 119 10.1186/s13690-020-00497-w 33292511 15. Yeatman S Chamberlin S Dovel K Women’s (health) work: a population-based, cross-sectional study of gender differences in time spent seeking health care in Malawi PLoS One 2018 13 12 e0209586 10.1371/journal.pone.0209586 30576388 16. Munthali AC Mannan H MacLachlan M Swartz L Makupe CM Chilimampunga C Non-use of formal health services in Malawi: perceptions from non-users Malawi Med J 2014 26 4 126 132 26167263 17. Jafry MA Jenny AM Lubinga SJ Larsen-Cooper E Crawford J Matemba C Babigumira JB Examination of patient flow in a rural health center in Malawi BMC Res Notes 2016 9 363 363 10.1186/s13104-016-2144-x 27456090 18. Dullie L Meland E Hetlevik Ø Mildestvedt T Kasenda S Kantema C Gjesdal S Performance of primary care in different healthcare facilities: a cross-sectional study of patients’ experiences in Southern Malawi BMJ Open 2019 9 7 e029579 10.1136/bmjopen-2019-029579 31324683 19. Ministry of Health of Malawi. Human resources for health strategic plan 2017–2022. Lilongwe: Government of Malawi; 2018. 20. Scheffler R, Cometto G, Tulenko K, Bruckner T. Health workforce requirements for universal health coverage and the Sustainable Development Goals–Background paper N.1 to the WHO Global Strategy on Human Resources for Health: Workforce 2030. In: Human resources for health observer series No 17. Geneva: World Health Organization; 2016. 21. Riddell D Tjornbo O Westley F Dutton JE Golden-Biddle K Agency and innovation in a phase of turbulent change: conservation in the Great Bear Rainforest Using a positive lens to explore social change 2012 New York Routledge 155 180 22. van Wijk J Zietsma C Dorado S de Bakker FGA Martí I Social innovation: integrating micro, meso, and macro level insights from institutional theory Bus Soc 2019 58 5 887 918 10.1177/0007650318789104 23. Westley F Antadze N Making a difference: strategies for scaling social innovation for greater impact Innov J 2010 15 2 1 19 24. Nilsson W George G Baker T Joshi H Social innovation as institutional work Handbook of inclusive innovation, the role of organizations, markets and communities in social innovation 2019 Cheltenham Edward Elgar Publishing 25. Moulaert F MacCullum D Moulaert J Hiller J Vicari Haddock S Social innovation: institutionally embedded, terrritorially (re) produced Social innovation and territorial development 2009 Farnham Ashgate 11 24 26. The World Bank . "Mobile cellular subscriptions (per 100 people) - Malawi". 2021. https://data.worldbank.org/indicator/IT.CEL.SETS.P2?locations=MW. Accessed 20 Jan 2022. 27. Watkins S Robinson A Dalious M Evaluation of the information and communications technology for maternal, newborn and child health project 2013 Malawi Invest in Knowledge Initiative 28. Blauvelt C West M Maxim L Kasiya A Dambula I Kachila U Ngwira H Armstrong CE Scaling up a health and nutrition hotline in Malawi: the benefits of multisectoral collaboration BMJ 2018 363 k4590 10.1136/bmj.k4590 30530659 29. Gilson L Health policy and systems research: a methodology reader 2012 Geneva World Health Organization 30. Yin RK Case study research: design and methods 2014 5 Thousand Oaks Sage Publications 31. Moulaert F Van Dyck B Moulaert F MacCallum D Mehmood A Hamdouch A Framing social innovation research: a sociology of knowledge perspective The international handbook on social innovation 2013 Cheltenham and Massachusetts Edward Elgar 466 479 32. Social Innovation in Health Initiative. 2021. https://socialinnovationinhealth.org. Accessed 20 Jan 2023. 33. NVivo (version 12). https://www.qsrinternational.com/nvivo-qualitative-data-analysis-software/home. 34. Battilana J Leca B Boxenbaum E How actors change institutions: towards a theory of institutional entrepreneurship Acad Manag Ann 2009 3 1 65 107 10.5465/19416520903053598 35. Lawrence TB Suddaby R Clegg SR Hardy C Lawrence TB Nord WR Institutions and institutional work Sage handbook of organizational studies 2006 London Sage 215 254 36. Nilsson W Positive institutional work: exploring institutional work through the lens of positive organizational scholarship Acad Manag Rev 2014 40 3 370 398 10.5465/amr.2013.0188 37. Cameron KS Dutton JE Quinn RE Positive Organizational Scholarship: foundations of a new discipline 2003 San Francisco Berrett-Koehler 38. Fredrickson BL Devine P Plant A Positive emotions broaden and build Advances in experimental social psychology 2013 Burlington Academic 1 53 39. Roberts K, Dowell A, Nie J-B. Attempting rigour and replicability in thematic analysis of qualitative research data; a case study of codebook development. BMC Med Res Methodol. 2019;19(1):66. 40. Murray R Caulier-Grice J Mulgan G The open book of social innovation 2010 London The Young Foundation 41. Murdock A Nicholls A Social innovation: blurring boundaries to reconfigure markets 2012 London Palgrave Macmillan UK: Imprint: Palgrave Macmillan 42. Brown LD Bridging organizations and sustainable development Hum Relat 1991 44 8 807 831 10.1177/001872679104400804 43. Battilana J D’Aunno T Lawrence T Suddaby R Leca B Institutional work and the paradox of embedded agency Institutional work - actors and agency in institutional studies of organizations 2009 Cambridge Cambridge University Press 31 50 44. Emirbayer M Mische A What is agency? Am J Sociol 1998 103 4 962 1023 10.1086/231294 45. VillageReach. Impact evaluation of Chipatala cha pa Foni (CCPF), Malawi’s health and nutrition hotline. Lilongwe: VillageReach; 2019. 46. Furnari S Interstitial spaces: microinteraction settings and the genesis of new practices between institutional fields Acad Manag Rev 2014 39 4 439 462 10.5465/amr.2012.0045 47. Zhou J, Ren R. Striving for creativity: building positive contexts in the workplace. In: Spreitzer GM, Cameron KS, editors. Oxford handbook of positive organizational scholarship. New York; 2011. p. 97–109. 48. Perry-Smith JE Shalley CE The social side of creativity: a static and dynamic social network perspective Acad Manag Rev 2003 28 1 89 106 10.2307/30040691 49. Pearce CL The future of leadership: combining vertical and shared leadership to transform knowledge work Acad Manag Exec 2004 18 1 47 57 50. Pearce CL Manz CC Sims HP Jr Where do we go from here?: Is shared leadership the key to team success? Organ Dyn 2009 38 3 234 238 10.1016/j.orgdyn.2009.04.008 51. Denis J-L Langley A Sergi V Leadership in the plural Acad Manag Ann 2012 6 1 211 283 10.5465/19416520.2012.667612 52. Williams LA DeSteno D Pride and perseverance: the motivational role of pride J Pers Soc Psychol 2008 94 6 1007 1017 10.1037/0022-3514.94.6.1007 18505314 53. Berger PL Luckmann T The social construction of reality: a treatise in the sociology of knowledge 1967 New York Anchor 54. Snyder CR The psychology of hope: you can get there from here 1994 New York Free Press 55. Snyder CR Hope theory: rainbows in the mind Psychol Inq 2002 13 4 249 275 10.1207/S15327965PLI1304_01
PMC010xxxxxx/PMC10224842.txt
==== Front Mol Neurobiol Mol Neurobiol Molecular Neurobiology 0893-7648 1559-1182 Springer US New York 37022587 3314 10.1007/s12035-023-03314-x Article Selective Disruption of Perineuronal Nets in Mice Lacking Crtl1 is Sufficient to Make Fear Memories Susceptible to Erasure Poli Andrea 1 Viglione Aurelia 1 Mazziotti Raffaele 2 Totaro Valentino 1 Morea Silvia 2 Melani Riccardo 3 Silingardi Davide 4 Putignano Elena 2 Berardi Nicoletta 24 http://orcid.org/0000-0001-5614-0668 Pizzorusso Tommaso [email protected] 12 1 grid.6093.c BIO@SNS Lab, Scuola Normale Superiore Via G, Moruzzi 1, 56124 Pisa, Italy 2 grid.5326.2 0000 0001 1940 4177 Institute of Neuroscience, National Research Council, Via Moruzzi, 1, 56124 Pisa, Italy 3 grid.137628.9 0000 0004 1936 8753 Neuroscience Institute, New York University Grossman School of Medicine, New York, NY 10016 USA 4 grid.8404.8 0000 0004 1757 2304 Department of Neuroscience, Psychology, Drug Research, and Child Health NEUROFARBA, University of Florence, 50134 Florence, Italy 6 4 2023 6 4 2023 2023 60 7 41054119 27 9 2022 9 3 2023 © The Author(s) 2023, corrected publication 2023 https://creativecommons.org/licenses/by/4.0/ Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The ability to store, retrieve, and extinguish memories of adverse experiences is an essential skill for animals’ survival. The cellular and molecular factors that underlie such processes are only partially known. Using chondroitinase ABC treatment targeting chondroitin sulfate proteoglycans (CSPGs), previous studies showed that the maturation of the extracellular matrix makes fear memory resistant to deletion. Mice lacking the cartilage link protein Crtl1 (Crtl1-KO mice) display normal CSPG levels but impaired CSPG condensation in perineuronal nets (PNNs). Thus, we asked whether the presence of PNNs in the adult brain is responsible for the appearance of persistent fear memories by investigating fear extinction in Crtl1-KO mice. We found that mutant mice displayed fear memory erasure after an extinction protocol as revealed by analysis of freezing and pupil dynamics. Fear memory erasure did not depend on passive loss of retention; moreover, we demonstrated that, after extinction training, conditioned Crtl1-KO mice display no neural activation in the amygdala (Zif268 staining) in comparison to control animals. Taken together, our findings suggest that the aggregation of CSPGs into PNNs regulates the boundaries of the critical period for fear extinction. Supplementary Information The online version contains supplementary material available at 10.1007/s12035-023-03314-x. Keywords Perineuronal nets (PNNs) Chondroitin sulfate proteoglycans (CSPGs) Fear conditioning Fear extinction Pupillometry issue-copyright-statement© Springer Science+Business Media, LLC, part of Springer Nature 2023 ==== Body pmcIntroduction The ability to extinguish fear memories when threats are no longer present is critical for adaptive behavior. During fear conditioning, the repeated pairing of an initially neutral stimulus (conditioned stimulus; CS) with an aversive stimulus (unconditioned stimulus; US) induces a strong and persistent fear memory [1] that can be inhibited by repeated exposure to the CS in the absence of the US, a process called fear extinction [2]. There is compelling behavioral evidence that extinction training does not erase or reverse the original CS-US association but rather leads to the formation of a new inhibitory memory that competes with the initial fear memory for the control of behavior [3, 4]. Fear memory extinction in adult animals is not permanent but decays with time, a process known as spontaneous fear recovery [5]. Moreover, conditioned fear responses can be restored by presenting the US alone in the context in which extinction training occurred (reinstatement) [4, 6] or may re-emerge following a shift in context (renewal) [7, 8]. The extinction of conditioned fear memories in adults relies on a network of structures, such as the basolateral amygdala (BLA), the lateral amygdala (LA), and the ventromedial prefrontal cortex (vmPFC) [9–12]. Previous studies have provided numerous lines of evidence showing that extinction training of adult animals produces a new memory that inhibits the original fear memory stored in the lateral amygdala (LA) [13, 14]. In contrast, extinction seems to produce a permanent erasure of fear memory in juvenile animals, which do not exhibit reinstatement or context-dependent renewal of conditioned fear responses following an extinction protocol [15–17]. The transition from a fear memory that can be erased in juvenile mice to a persistent fear memory in the adult have been suggested to rely on the maturation of the circuits involved in conditioned fear extinction [7]. For example, removal of chondroitin sulfate proteoglycans (CSPGs) from the amygdala extracellular matrix (ECM) by enzymatic digestion allowed juvenile-like erasure of conditioned fear memories in adult animals [7]. CSPGs are diffusely present in the ECM of the adult brain [18] and condense around some cells forming perineuronal nets (PNNs) [19]. This process is triggered by neuronal production of the cartilage link protein Crtl1 (also known as HAPLN1), which is upregulated during development [20, 21]. The developmental condensation of CSPGs in PNNs of the visual cortex, rather than their sheer presence, play a crucial role in protecting adult visual cortical circuits from being modified by experience [20]. However, it is unknown whether the condensation of CSPGs in PNNs, taking place during development, is also involved in the transition from a conditioned fear memory that can be erased by extinction to a fear memory that is no more susceptible to erasure. Also unknown is the mechanism through which PNNs make adult amygdala circuits resilient to extinction effects, protecting fear memories from erasure. Here, we exploited mice lacking the Crtl1 protein (Crtl1-KO) which have attenuated PNNs but unchanged overall levels of CSPGs [20], to investigate whether preventing the aggregation of CSPGs into PNNs is sufficient to induce fear memory susceptible to erasure, and assessing the associated pattern of activation. In particular, we assessed fear responses through freezing and pupil size, two well-established behavioral and physiological markers of fear memories. Freezing is a defensive response commonly used to evaluate associative fear memory in rodents, while pupil dilation has been widely used to objectively assess fear learning in humans [22–24] and provides valuable information about the role of arousal in modulating fear circuits [25]. We found that Crtl1-KO animals retain the juvenile feature of erasing a specific conditioned fear memory following a protocol of extinction. In particular, Crtl1-KO mice exhibited a stronger reduction in both pupillary and freezing response to the CS with respect to Crtl1-WT mice. This persistent reduction of fear in Crtl1-KO mice did not depend on passive loss of memory, since fear memories assessed 9 days after learning without intervening extinction protocol is comparable in Crtl1-KO and Crtl1 wild-type mice (Crtl1-WT). To assess the mechanisms through which PNN disruption leads to permanent erasure of a fear memory, we analyzed neuronal activation in the amygdala and in the infralimbic cortex (IL) at the end of the extinction protocol via immunostaining for Zif268. We found that, following extinction, there was no neural activation in the amygdala of conditioned Crtl1-KO mice in response to CS, in accordance with the erasure of the conditioned fear memory. On the contrary, Crtl1-WT mice showed a clear activation of these regions. Results Lack of Crtl1 Accelerates Extinction of Fear Memories in Adult Crtl1-KO Mice To investigate whether the condensation of CSPGs in PNNs is crucial in the transition from a conditioned fear memory that can be erased by extinction to a conditioned fear memory which is not erasable, we performed a classical auditory cued fear conditioning and extinction protocol in Crtl1-KO mice and their WT littermates (Fig. 1A). Crtl1-KO mice exhibit a marked decrease in the amount of PNNs in the amygdala and IL cortex (Suppl. Figure 1), key regions for the extinction of adult conditioned fear memories. During the habituation, mice freely explored the chamber (day 0, context A) showing low levels of freezing (below 4%), consistently with a normal habituation to the context, with no difference between genotypes (Fig. 1B). We also found comparable freezing levels between conditioned Crtl1-KO and Crtl1-WT mice during the learning phase of the test (day 1, Fig. 1C). However, we found that Crtl1-KO mice exhibited a significantly accelerated pattern of freezing reduction with respect to Crtl1-WT mice (Fig. 1D) during the first day of extinction (early extinction, day 2). In particular, freezing levels in Crtl1-KO become significantly lower than in Crtl1-WT mice as early as the third block of 2 conditioned stimuli (CS), reaching significantly lower levels of freezing with respect to the beginning of CS presentations (Fig. 1D) at the end of the first day of extinction. During the second day of the extinction protocol (late extinction, day 3), Crtl1-KO mice maintain the low freezing levels achieved during the first day of extinction (Fig. 1E), while Crtl1-WT mice began to decrease freezing, reaching levels comparable to Crtl1-KO mice from the fifth block of 2 CS (Fig. 1E). These results demonstrate that CSPG condensation in PNNs due to cartilage link protein Crtl1 has an important role in promoting an accelerated fear memory extinction but not in learning.Fig. 1 Fear extinction in Crtl1-KO mice. A Diagram showing the fear conditioning and extinction paradigm. B Habituation in Crtl1-KO and Crtl1-WT mice show low freezing levels and no differences between genotypes during 3-min exposition to the conditioned context. C Freezing levels in Crtl1-KO and Crtl1-WT mice during conditioning. Both genotypes exhibit a comparable pattern of freezing increase, coherent with a normal pattern of fear learning and no deficits of fear acquisition (two-way RM ANOVA: genotype p = 0.523, CS-US p < 0.001, interaction genotype x CS-US p = 0.346; post hoc Sidak multiple comparisons, CS-US within Crtl1-WT: 1 vs. 5 p < 0.01; CS-US within Crtl1-KO: 1 vs. 5 p < 0.001). D During early extinction, Crtl1-KO mice but not Crtl1-WT mice exhibited a significantly accelerated pattern of freezing reduction, as early as third block of 2 CS (two-way RM ANOVA: genotype p < 0.001; blocks of 2 CS p < 0.001; interaction genotype x blocks of 2 CS p < 0.001; post hoc Sidak multiple comparisons, genotype within blocks of 2 CS, 1: p = 0.721, 2: p = 0.251, 3: p < 0.001, 4: p < 0.001, 5: p < 0.001, 6: p < 0.001), and significantly reduced their freezing levels at the end of the extinction protocol (post hoc Sidak multiple comparisons, Crtl1-WT 1 vs. 6: p = 0.836; Crtl1-KO 1 vs. 6: p < 0.001). E During late extinction, Crtl1-KO mice kept showing significantly lower freezing from first to fourth block of 2 CS (two-way RM ANOVA: genotype p < 0.001, blocks of 2 CS p < 0.001, interaction genotype x blocks of 2 CS p < 0.001; post hoc Sidak multiple comparisons, genotype within blocks of 2 CS, 1: p < 0.001; 2: p < 0.001; 3: p < 0.001; 4: p < 0.05; 5: p = 0.980; 6: p = 0.369) and significantly reduced their freezing levels (post hoc Sidak multiple comparisons, Crtl1-KO 1 vs. 6: p < 0.001). From the fifth block of 2 CS, Crtl1-WT mice reached freezing levels comparable to Crtl1-KO and significantly reduced their freezing levels (post hoc Sidak multiple comparisons, Crtl1-WT 1 vs. 6: p < 0.001). n = 10 Crtl1-KO, n = 10 Crtl1-WT. *P-value between genotypes, #P-value between CS. CS = conditioned stimulus; US = unconditioned stimulus Pupillometry as a Physiological Readout of Fear Extinction in Crtl1-KO Mice Due to its sensitivity to arousal [25–27], the pupil responds with dilation to salient or threatening stimuli. For this reason, pupil dilations have gained interest as a measure of the conditioned response [23]. We used pupillometry as a physiological readout of fear learning and extinction in Crtl1-KO mice and their WT littermates. We designed a virtual cued fear conditioning protocol, using a visual cue as CS, paired with a tail shock (US). During fear conditioning, mice were head-fixed and free to run on a circular treadmill. An infrared webcam was used to record the pupil and the MEYE Deep Learning tool was employed to perform pupillometry [28] (Fig. 2A). To assess the efficacy of our virtual fear conditioning protocol, first, we tested C57BL6/J wild-type mice receiving the CS alone (sham) or the CS paired with the US (shock) (Suppl. Figure 2A). As shown in Suppl Fig. 2, we observed a stronger pupil dilations in response to CS in shocked mice compared to sham mice the day after conditioning (recall, Suppl. Figure 2C, D) validating the use of pupil size measurement to reveal learned fear. Then, we evaluated fear learning in Crtl1-KO and Crtl1-WT mice (Fig. 2) in a cohort of mice different from the one used for freezing assessment. As assessed with freezing levels, we observed comparable pupillary response between genotypes to the CS after learning (Fig. 2B-D). Interestingly, we found that Crtl1-KO mice exhibited a stronger reduction in the pupillary response to CS with respect to Crtl1-WT mice during the first day of extinction (early extinction) (Fig. 2 E, F). Pupillary responses in Crtl1-KO become significantly lower than in Crtl1-WT mice as early as the second block of 5 CS (Fig. 2F). During the second day of the extinction protocol (late extinction), Crtl1-KO mice still showed significantly lower pupillary responses compared to Crtl1-WT mice (Fig. 2G). Crtl1-WT mice reached pupillary responses comparable to Crtl1-KO mice from the second block of 5 CS of late extinction (Fig. 2H). To exclude possible defects in the pupillary light response present in mutant mice, we evaluated the pupillary light reflex (PLR). The results revealed in Crtl1-KO mice an unaltered PLR during both constriction and pupil re-dilation (Suppl. Figure 3).Fig. 2 Pupillometry assessment of fear extinction in Crtl1-KO mice. A Diagram showing the pupillometry setup and the virtual fear conditioning timeline. B Average of the pupillary responses of Crtl1-WT (top) and Crtl1-KO (bottom) mice during the virtual fear conditioning. The gray area represents the presentation of the CS stimulus and the red area the presentation of the US stimulus. C Average of the pupillary responses of Crtl1-WT (top) and Crtl1-KO (bottom) mice during the virtual fear recall. The purple area represents the presentation of the CS stimulus. D During fear recall, we observed comparable pupillary responses between genotypes to the CS stimulus (unpaired T-test p = 0.566). E On the left, the average fluctuation of pupil size for the first block of 5 CS during early extinction. On the right, the pupil peaks during the presentation of the first 5 CS. We found no differences between genotypes (unpaired T-test p = 0.208). F On the left, the average fluctuation of pupil size for the second block of 5 CS during early extinction. On the right, the pupil peaks during the presentation of the second block of 5 CS. We found a lower pupillary response in Crtl1-KO mice compared to Crtl1-WT mice (unpaired T-test p = 0.010). G On the left, the average fluctuation of pupil size for the first block of 5 CS during late extinction. On the right, the pupil peaks during the presentation of the first 5 CS. We still found a lower pupillary response in Crtl1-KO mice compared to Crtl1-WT mice (unpaired T-test p < 0.01). H On the left, the average fluctuation of pupil size for the second block of 5 CS during late extinction. On the right, the pupil peaks during the presentation of the second 5 CS. We found no differences between genotypes (unpaired T-test p = 0.527). n = 10 Crtl1-KO, n = 11 Crtl1-WT. CS = conditioned stimulus; US = unconditioned stimulus; HAB = habituation; ITI = inter-trial; ns = not significant Thus, we observed a faster extinction of fear memories in Crtl1-KO mice also using a physiological measure. Abnormal Spontaneous Recovery and Fear Renewal in Adult Crtl1-KO Mice Seven days after late extinction (day 10), we assessed spontaneous recovery and fear renewal (Fig. 3A). The results clearly showed that Crtl1-KO mice still displayed attenuation of fear response caused by the extinction protocol while Crtl1-WT mice exhibited a higher fear response retrieval, both for spontaneous recovery (Fig. 3B) and for fear renewal (Fig. 3D). Moreover, the lower freezing level of Crtl1-KO mice remained evident throughout repetition of CS, both for spontaneous recovery and for fear renewal protocols. Interestingly, we also found that both Crtl1-KO and Crtl1-WT mice showed the same freezing levels when placed in the unconditioned context (context B, spontaneous recovery) (Fig. 3C), while Crtl1-WT mice showed a higher context-dependent freezing behavior compared to Crtl1-KO mice when placed in the conditioned context (fear renewal, context A) (Fig. 3E).Fig. 3 Spontaneous recovery and fear renewal in Crtl1-KO mice. A Diagram showing the spontaneous recovery and fear renewal paradigm. B Spontaneous recovery 7 days after extinction (day 10). Crtl1-KO mice kept showing significantly lower freezing during all four CS presentations with respect to Crtl1-WT mice (two-way RM ANOVA, interaction genotype x CS p < 0.001, genotype p < 0.001, CS p < 0.001; post hoc Sidak multiple comparisons, genotype within CS, 1: p < 0.001, 2: p < 0.001, 3: p < 0.001, 4: p < 0.001). C Crtl1-KO and Crtl1-WT mice showed the same freezing levels when placed in the unconditioned context B, 7 days after extinction (day 10) (unpaired T-test, p = 0.408). D Fear renewal 7 days after extinction (day 10). Crtl1-KO mice kept showing significantly lower freezing during all four CS presentations with respect to Crtl1-WT mice (two-way RM ANOVA, interaction genotype x CS p < 0.001, genotype p < 0.001, CS p < 0.001; post hoc Sidak multiple comparisons, genotype within CS, 1: p < 0.001, 2: p < 0.001, 3: p < 0.001, 4: p < 0.001). E Crtl1-WT mice showed a higher context-dependent freezing behavior compared to Crtl1-KO mice when placed in the conditioned context A, 7 days after extinction (day 10) (unpaired T-test, p < 0.001). F Spontaneous recovery 42 days after extinction (day 45). Crtl1-KO mice kept showing significantly lower freezing during all four CS presentations with respect to Crtl1-WT mice (two-way RM ANOVA, interaction genotype x CS p < 0.001, genotype p < 0.001, CS p < 0.001; post hoc Sidak multiple comparisons, genotype within CS, 1: p < 0.001, 2: p < 0.001, 3: p < 0.001, 4: p < 0.001). G Crtl1-KO and Crtl1-WT mice showed the same freezing levels when placed in the unconditioned context B, 42 days after extinction (day 45) (unpaired T-test, p = 0.928). H Fear renewal 42 days after extinction (day 45). Crtl1-KO mice kept showing significantly lower freezing during the first two CS presentations with respect to Crtl1-WT mice (two-way RM ANOVA, interaction genotype x CS p < 0.001, genotype p < 0.001, CS p < 0.001; post hoc Sidak multiple comparisons, genotype within CS, 1: p < 0.001, 2: p < 0.001, 3: p = 0.127, 4: p = 0.092). I Crtl1-WT mice showed a higher context-dependent freezing behavior compared to Crtl1-KO mice when placed in the conditioned context A, 42 days after extinction (day 45) (unpaired T-test, p < 0.001). n = 10 Crtl1-KO, n = 10 Crtl1-WT. *P-value between genotypes, #P-value between CS. CS = conditioned stimulus; ns = not significant; LE = late extinction Assessment of spontaneous recovery and fear renewal 42 days after the end of the late extinction (day 45, Fig. 3A) showed that Crtl1-KO mice still displayed lower fear response caused by the extinction protocol compared to Crtl1-WT (Fig. 3F and Fig. 3H). The lower freezing level of Crtl1-KO mice with respect to Crtl1-WT mice during spontaneous recovery remained evident throughout repetition of CS (Fig. 3F). When the fear renewal protocol was employed, we found that Crtl1-KO mice showed significantly lower freezing during the 1 and 2 CS with respect to Crtl1-WT mice, while, during the 3 and 4 CS, Crtl1-WT mice reached freezing levels comparable to Crtl1-KO mice (Fig. 3H). Again we found that both genotypes showed comparable freezing levels in the unconditioned context (context B, spontaneous recovery) (Fig. 3G), while Crtl1-WT mice showed a higher context-dependent freezing behavior compared to Crtl1-KO mice when placed in the conditioned context (fear renewal, context A) (Fig. 3I). Taken together, these results suggest that PNN disruption in Crtl1-KO mice is sufficient to determine a juvenile-like fear extinction: Crtl1-KO mice show a persistent reduction of fear both in spontaneous recovery and in context-dependent renewal, 7 and 42 days after extinction; Crtl1-WT mice show, instead, the lack of long-term effects of extinction typical of adults. Reduction of Fear in Crtl1-KO Mice Did Not Depend on Passive Loss of Memory To ensure that the reduction in fear memory observed in Crtl1-KO mice after extinction was not due to a weakened consolidation and retention of the memory, we conducted a fear memory extinction test 9 days after fear conditioning (Fig. 4A). We found that freezing levels in the 1 and 2 blocks of 2 CS did not differ between Crtl1-WT and Crtl1-KO mice (Fig. 4B). This result confirms that Crtl1-KO mice do not show deficits in fear memory consolidation and retention. Moreover, we observed a faster decrease in fear response in Crtl1-KO mice compared to Crtl1-WT mice (Fig. 4B). Also in this case, Crtl1-KO, but not Crtl1-WT mice, significantly reduced their freezing levels at the end of the early extinction protocol (Fig. 4B). These results suggest that the higher efficacy of the extinction protocol in Crtl1-KO mice is present for both recent and older memories.Fig. 4 Freezing levels in conditioned Crtl1-KO and Crtl1-WT mice during early extinction starting 9 days after fear conditioning. A Diagram showing the fear conditioning and extinction paradigm. B Even if the extinction procedure started 9 days after fear learning, Crtl1-KO mice, but not Crtl1-WT mice, exhibited a significantly accelerated pattern of freezing reduction, as early as third block of 2 CS (two-way RM ANOVA: genotype, p < 0.01; blocks of 2 CS: p < 0.001; interaction genotype x blocks of 2 CS: p < 0.001; post hoc Sidak multiple comparisons, genotype within blocks of 2 CS, 1: p = 0.059, 2: p = 0.794, 3: p < 0.001, 4: p < 0.001, 5: p < 0.001), and significantly reduced their freezing levels during early extinction (post hoc Sidak multiple comparisons, blocks of 2 CS Crtl1-WT, 1 vs. 5: 0.520, Crtl1-KO, 1 vs. 5: p < 0.001). n = 10 Crtl1-KO, n = 10 Crtl1-WT. *P-value between genotypes, #P-value between CS. CS = conditioned stimulus; US = unconditioned stimulus Fear Extinction in Adult Crtl1-KO Mice is Accompanied by a Complete Loss of Amygdala Activation in Response to the CS To assess neuronal activation in the amygdala and IL cortex of Crtl1-KO mice, we performed an immunostaining for Zif268. Zif268 is an immediate early gene known to be implicated in neuronal plasticity and memory formation [29]. We focused the analysis on the early extinction stage, corresponding to the maximal difference between Crtl1-KO and Crtl1-WT mice. To isolate the specific effect of associative learning on amygdala and IL activation from the effect of exposure to the CS and to the US, we compared conditioned mice to animals that received tone and electrical stimulation in an unpaired pattern (pseudo-conditioned mice) [30] (Fig. 5A).Fig. 5 Amygdala and IL cortex activation after extinction in Crtl1-KO mice. A Diagram showing the experimental design. B Representative diagram of the areas considered for the Zif268 immunohistochemistry: BLA, basolateral amygdala; LA, lateral amigdala; CeM, medial part of the central amygdala; CeL, lateral part of the central amygdala; IL, infralimbic cortex. C Freezing levels in conditioned and pseudo-conditioned Crtl1-KO and Crtl1-WT mice at the end of the conditioning protocol. After 5 presentations of US, pseudo-conditioned Crtl1-KO and Crtl1-WT mice reached freezing levels comparable to conditioned mice (one-way ANOVA, p = 0.216). D Freezing levels in conditioned and pseudo-conditioned mice during the early extinction. Cond-Crtl1-KO mice exhibited a significantly accelerated pattern of freezing reduction (three-way ANOVA, genotype x condition x blocks of 2 CS p < 0.001. Genotype p < 0.001, condition: p < 0.001, blocks of 2 CS p < 0.001; genotype x condition p < 0.001, genotype x blocks of 2 CS: p < 0.001, condition x block of 2 CS, p < 0.001; post hoc Sidak multiple comparisons, difference between Cond-Crtl1-KO and Cond-Crtl1-WT, CS-1: p = 0.218, CS-2: p < 0.001, CS-3: p < 0.001, CS-4: p < 0.001, CS-5: p < 0.001) and significantly reduced their freezing levels during early extinction (post hoc Sidak multiple comparisons, difference between freezing levels for the 1 and 5 blocks of 2 CS Crtl1-WT, p = 0.200; Crtl1-KO, p < 0.001). Pseudo-Crtl1-WT and pseudo-Crtl1-KO mice did not show any significant difference of freezing levels between genotypes but significantly lower freezing levels during all five blocks of 2 CS with respect to conditioned mice (post hoc Sidak multiple comparisons, condition within blocks of 2 CS, Cond-Crtl1-WT, 1 to 5: p < 0.001; Cond-Crtl1-KO, 1 to 5: p < 0.001). E Analysis of Zif268-positive cells in the LA (top), CeM (bottom left), and CeL (bottom right) of conditioned and pseudo-conditioned mice. Cond-Crtl1-KO mice showed significantly reduced levels of LA and CeM activation with respect to Cond-Crtl1-WT mice (LA, two-way ANOVA: genotype p < 0.01; condition p = 0.051; interaction genotype x condition p < 0.001. Post hoc Sidak multiple comparisons: Cond-Crtl1-WT vs. Cond-Crtl1-KO p < 0.001, Cond-Crtl1-WT vs. pseudo-Crtl1-KO p < 0.01) (CeM, two-way ANOVA: genotype p < 0.001; condition p < 0.001; interaction genotype x condition p < 0.001. Post hoc Sidak multiple comparisons: Cond-Crtl1-WT vs. Cond-Crtl1-KO p < 0.001, Cond-Crtl1-WT vs. pseudo-Crtl1-KO p < 0.001) and indistinguishable from those shown by pseudo-Crtl1-WT and pseudo-Crtl1-KO mice (LA, post hoc Sidak multiple comparisons: Cond-Crtl1-KO vs. pseudo-Crtl1-WT p = 0.858, Cond-Crtl1-KO vs. pseudo-Crtl1-KO p = 0.349) (CeM, post hoc Sidak multiple comparisons: Cond-Crtl1-KO vs. pseudo-Crtl1-WT p > 0.99, Cond-Crtl1-KO vs. pseudo-Crtl1-KO p = 0.430). Conditioned mice showed higher levels of CeL activation compared to pseudo-conditioned mice (two-way ANOVA: genotype p = 0.630; condition p < 0.001; interaction genotype x condition p < 0.05. Post hoc Sidak multiple comparisons: condition within WT mice p < 0.001, condition within KO mice p < 0.001). F Representative images of Zif268-positive cells in LA, CeM, and CeL of conditioned and pseudo-conditioned Crtl1-WT and Crtl1-KO mice after the fifth block of 2 CS during early extinction. Scale bar, 200 μm. A larger version of this image is present in Suppl. Figure 5. G Analysis of Zif268-positive cells in the BLA (top) and IL (bottom) of conditioned and pseudo-conditioned Crtl1-WT and Crtl1-KO mice. Cond-Crtl1-KO mice showed significantly reduced levels of BLA activation with respect to Cond-Crtl1-WT mice (two-way ANOVA: genotype, p < 0.001; condition, p < 0.001; interaction genotype x condition, p < 0.001. Post hoc Sidak multiple comparisons: Cond-Crtl1-WT vs. Cond-Crtl1-KO p < 0.001, Cond-Crtl1-WT vs. pseudo-Crtl1-KO p < 0.001) and indistinguishable from those shown by pseudo-Crtl1-WT and pseudo-Crtl1-KO mice (post hoc Sidak multiple comparisons: Cond-Crtl1-KO vs. pseudo-Crtl1-WT p = 0.997, Cond-Crtl1-KO vs. pseudo-Crtl1-KO p = 0.965). Cond-Crtl1-KO mice showed comparable levels of IL activation such as that shown by Cond-Crtl1-WT mice (two-way ANOVA: genotype p < 0.05; condition p < 0.001; interaction genotype x condition p = 0.634. Post hoc Sidak multiple comparisons: genotype within conditioned mice p = 0.124, genotype within pseudo-conditioned mice p = 0.401), while pseudo-Crtl1-WT but not pseudo-Crtl1-KO mice showed reduced Zif268-positive cell density with respect to conditioned mice (post hoc Sidak multiple comparisons: condition within WT, p < 0.05; condition within KO, p = 0.058). H Representative images of Zif268-positive cells in BLA and IL of conditioned and pseudo-conditioned Crtl1-WT and Crtl1-KO mice after the fifth block of 2 CS during early extinction. Scale bar, 200 μm. A larger version of this image is present in Suppl. Figure 5. n = 10 in each group. *P-value between genotypes, #P-value between CS, $P-value between conditions. CS = conditioned stimulus; US = unconditioned stimulus The behavioral results show that both conditioned and pseudo-conditioned Crtl1-KO and Crtl1-WT mice showed a comparable response to the US, reaching a freezing level around 60% at the end of conditioning (Fig. 5C). However, conditioned mice developed a strong response to the CS (Fig. 5D). The results confirmed the enhanced extinction of Crtl1-KO mice by revealing a faster reduction of fear response in Crtl1-KO compared to Crtl1-WT mice (Fig. 5D). As expected, pseudo-conditioned mice showed significantly lower levels of freezing with respect to conditioned mice, and did not show any significant reduction of freezing levels with extinction (Fig. 5D). Therefore, this group of animals could be used to test Zif268 activation. It has been shown that neurons in lateral amygdala (LA) maintain high levels of response to CS even after extinction protocols [31]. However, we found that there was no significant neuronal activation in the LA of conditioned Crtl1-KO mice after extinction, while neuronal activation was clearly present in conditioned Crtl1-WT mice compared to pseudo-conditioned Crtl1-WT mice (Fig. 5E). Density for Zif268 + cells in the main output nucleus of the amygdala, the medial part of the central amygdala (CeM), resulted significantly higher in conditioned Crtl1-WT mice than in pseudo-conditioned Crtl1-WT mice, showing significant neuronal activation in CeM (Fig. 5E). In conditioned Crtl1-KO mice, there was no significant difference in Zif268 + cell density with respect to Crtl1-KO pseudo-conditioned mice, showing absence of neuronal activation in CeM (Fig. 5E), in accordance with behavioral results of very low fear response at the end of early extinction (Fig. 5D). Neuronal activation in conditioned Crtl1-KO mice was significantly lower than in conditioned Crtl1-WT mice, while no difference in neuronal activation was found between the two pseudo-conditioned groups (Fig. 5E). We also found that density for Zif268 + cells in the lateral part of the central amygdala (CeL) resulted significantly higher in both conditioned Crtl1-WT and Crtl1-KO mice than in pseudo-conditioned mice, showing significant neuronal activation in CeL after extinction (Fig. 5E). Thus, the lack of CeM activation in Crtl1-KO mice seems to be due to lack of LA activation (Fig. 5E-F). BLA is a crucial part of the intra-amygdala circuitry, providing a major point of control in the transmission of information between LA and CeM, and in particular is an important site of modulation of LA-CeM transmission during extinction [12, 32, 33]. Density of Zif268 + cells in the BLA showed the same pattern found in CeM, with only conditioned Crtl1-WT mice showing significant neuronal activation in response to CS (Fig. 5G). Thus, the absence of neuronal activation in Crtl1-KO mice is already present in BLA, one of the main inputs to CeM (Fig. 5G-H). We also assessed neuronal activation in IL during early extinction as it seems to play a relevant role in reducing fear response during extinction protocols [11, 34]. We found that both conditioned Crtl1-WT and Crtl1-KO mice showed significant IL activation at the end of early extinction although there was a strong trend for lower IL activation in KO with respect to WT mice (Fig. 5G-H). Intriguingly, Zif268 induction in conditioned Crtl1-WT mice was observed also restricting Zif268 analysis to cells positive for WFA (Suppl. Figure 4). To analyze neuronal activation in Crtl1-KO and WT immediately after memory recall, we performed immunostaining for Zif268 in a different group of conditioned Crtl1-KO and WT mice perfused immediately after presenting one block of two CS (Fig. 6A). We found that Zif268 + cells were higher in LA, BLA, CeM, CeL, and IL of Crtl1-KO than in WT mice (Fig. 6B-E). This result is in line with the behavioral data shown in Fig. 1D, 4B, and 5D, in which we found that Ctrl1-KO mice showed a consistent trend for higher freezing levels in response to the first block of two CS presented during early extinction. Thus, both behavioral and molecular data suggest in Crtl1-KO mice a more flexible circuit compared to WT, characterized by not only stronger memory recall but also faster and more persistent fear extinction (Fig. 5D-H).Fig. 6 Amygdala and IL activation after memory recall in Crtl1-KO mice. A Diagram showing the experimental design. B Analysis of Zif268+ cells in the LA, CeM, and CeL of Crtl1-WT and Crtl1-KO mice. Crtl1-KO mice showed significantly higher levels of LA, CeM, and CeL activation with respect to Crtl1-WT mice (unpaired T-test, LA: p = 0.001, CeM: p < 0.0001, CeL: p < 0.0001). C Representative images of Zif268+ cells in LA, CeM, and CeL of Crtl1-WT and Crtl1-KO mice after memory recall. Scale bar, 200 μm. D Analysis of Zif268+ cells in the BLA and IL of Crtl1-WT and Crtl1-KO mice. Crtl1-KO mice showed significantly higher levels of BLA and IL activation with respect to Crtl1-WT mice (unpaired T-test, BLA: p = 0.006, IL: p < 0.0001). E Representative images of Zif268+ cells in BLA and IL of Crtl1-WT and Crtl1-KO mice after memory recall. Scale bar, 200 μm. n = 8 Crtl1-KO, n = 8 Crtl1-WT. CS = conditioned stimulus; US = unconditioned stimulus Discussion Our results show that the removal of the Crtl1 protein that selectively disrupts the aggregation of CSPGs in PNNs is sufficient to promote an accelerated and persistent fear memory erasure after administration of an extinction protocol. Indeed, during early extinction, Crtl1-KO mice showed reduced freezing levels as soon as the 3rd block of 2 CS while Crtl1-WT mice started showing freezing reduction at the end of the second day of extinction. These results were confirmed also testing mutant and WT mice in a virtual fear conditioning protocol, and using pupil dynamics as a physiological readout of fear learning and extinction. In humans, fear conditioning is often probed by measuring autonomic responses [35], such as skin conductance responses or startle responses [23, 36]. In recent years, these physiological measures have been complemented by pupil dilations [22–24]. Pupillometry offers a promising, complementary method for the quantification of the conditioned response. Indeed, pupil size assessment is not aversive, and it can be easily combined with other measurements and provides information about the activity of the autonomic nervous system [25] and locus coeruleus noradrenergic system [37], which plays an important role in modulating fear responses and extinction [38]. Coherent with the behavioral results, we found in Crtl1-KO mice an accelerated reduction in the pupillary response to CS with respect to Crtl1-WT mice as soon as the first day of extinction. We have also demonstrated that the condensation of CSPGs in PNNs triggered by neuronal production of the Crtl1 protein is needed for the transition from a conditioned fear memory that can be erased by extinction to a conditioned fear memory which is not. In order to assess if the observed acceleration in fear extinction translates in lower fear response at later stages, we tested freezing levels of Crtl1-KO and Crtl1-WT mice in the extinction context (spontaneous recovery) and in the training context (fear renewal) 7 and 42 days after late extinction. Interestingly, for both time points, we found a reduced spontaneous recovery and fear renewal in Crtl1-KO mice with respect to Crtl1-WT. Since early and late extinction were performed immediately after learning, the observed accelerated fear extinction pattern and the reduced freezing shown during spontaneous recovery and fear renewal may be the effect of an impairment in memory consolidation and a manifestation of oblivion. To rule out this possibility, we tested another group of Crtl1-KO and Crtl1-WT mice with a delay of 9 days between the end of the learning phase and the beginning of early extinction. We found that Crtl1-KO mice recapitulated the previous results, revealing that mutant mice do not show an impairment in memory consolidation and retrieval with respect to Crtl1-WT mice. Our behavioral and physiological data are in accordance with the results of Gogolla et al. [7], in which the injection of the chondroitinase ABC enzyme in the BLA is able to induce the acquired fear memories susceptible to erasure. In adult animals, the organization of CSPGs in PNNs is a key event in the control of central nervous system plasticity and in the closure of critical periods in many brain regions [7, 39, 40]. In the visual cortex, it has been shown that the developmental condensation of CSPGs in PNNs, rather than their sheer presence, play a crucial role in protecting adult visual cortical circuits from being modified by experience. The response of visual cortical circuits to monocular deprivation, which can be reinstated in the adult visual cortex by enzymatic removal of CSPGs [39], is also present in mice lacking Crtl1, which have attenuated PNNs but unchanged overall levels of CSPGs [20]. Overall, our data suggest the possibility that common general mechanisms of critical period closure exist in different brain circuits. The exact mechanism of action of PNNs in plasticity is only partially known [41]. Recently, superresolution microscopy provided a detailed description of the tight relationship between the PNN and the synaptic microstructure [42]. It would be of great interest to analyze if these high-resolution features are affected by Crtl1 mutation and if they correlate with memory changes. The extinction of conditioned fear memories in adults relies on a network of structures, such as the amygdala, the vmPFC, and the hippocampus [9, 43]. In particular, when the CS is present, the LA excite glutamatergic neurons in the BLA and GABAergic neurons in the lateral and medial intercalated cells (ITCs) [44] which separates the BLA from the central nucleus (CeA) [45, 46]. LA and BLA project dense glutamatergic synapses onto CeA, with the LA projecting only to its lateral sector (CeL) [47], and the BLA projecting to both the lateral and the medial (CeM) sectors [48–50]. In addition, convergent evidence suggests that the vmPFC, and in particular the IL [51], is necessary for the retention and recall of extinction [52–55]. To elucidate the functional activation of amygdala and IL cortex in Crtl1-KO mice, we replicated our previous extinction experiment trying to isolate the effect of the association between the CS and the US. We compared the neuronal activation of a group of conditioned Crtl1-KO and Crtl1-WT mice with a group of pseudo-conditioned Crtl1-KO and Crtl1-WT mice, in which the CS and US were unpaired. In particular, we performed immunohistochemistry for the immediate early gene Zif268 after early extinction. Zif268 is a member of the zinc finger transcription factor family. It regulates the expression of various late-response genes involved in different neuronal processes, including synaptic plasticity [29]. The role of Zif268 and other immediate early genes, such as c-fos and Arc, in learning and memory has been well described [56]. Zif268 expression is known to increase shortly after fear conditioning [57], suggesting its importance in fear memory formation [58]. Studies in mutant mice have also shown that overexpression of Zif268 enhanced resistance to extinction of aversive memories [59], while failure to induce Zif268 allowed spontaneous recovery [59, 60]. Here we provide further evidence of Zif268’s role as a marker of activity and plasticity changes in the amygdala and IL cortex after fear extinction. Indeed, the density of Zif268+ cells revealed a significant reduction in BLA, LA, and CeM activation of Crtl1-KO mice with respect to Crtl1-WT mice. Notably, Crtl1-KO mice activation was not significantly different from pseudo-conditioned mice. The reduced activation of CeM, as the output of the amygdala, is in line with the reduced behavioral freezing shown by Crtl1-KO mice. Remarkably, the reduced activation of LA is coherent with the possibility that the behavioral reduction of freezing shown in Crtl1-KO mice may be implemented as early as in LA. Regarding IL, we did not find significant differences between Crtl1-KO and Crtl1-WT mice; however, Zif268+ cell density of Crtl1-KO mice showed a reduced variability and the levels are at the significance threshold. Since it has been shown that optogenetic activation of IL for 30 s is able to accelerate and reduce freezing levels [34], it could be hypothesized that in Crtl1-KO mice IL may be precociously (as soon as the first block of 2 CS) and increasingly recruited, with respect to Crtl1-WT mice, contributing to the accelerated reduction of fear. During the fifth block of 2 CS, IL activation would be decreased (or comparable to the Crtl1-WT mice), since fear reduction has already been achieved. We also evaluated the activation of amygdala and IL neurons in Crtl1-KO and WT mice immediately after conditioning (memory recall). Our results revealed a higher number of Zif268+ neurons in Crtl1-KO mice compared to WT mice in both the amygdala nuclei and IL cortex. This observation aligns with previous research demonstrating enhanced memory following degradation of PNNs [61, 62]. It is plausible that the more flexible circuits present in Crtl1-KO mice not only lead to stronger fear conditioning but also facilitate faster and longer-lasting extinction. Taken together, our behavioral and molecular findings suggest that the aggregation of CSPGs into PNNs is a key factor in regulating the boundaries of the critical period for fear memories. Moreover, we suggest a possible activation pathway responsible for the accelerated reduced behavioral freezing shown by Crtl1-KO mice after extinction. We also propose pupillometry as a complementary physiological readout of fear learnings, which can be easily coupled with other physiological measures that require head-fixation procedures. Materials and Methods Animals For this study, we used adult mice (P75-P120) lacking the Crtl1/Hapln1 gene in the CNS, but not cartilage, which leads to attenuated PNNs in the adult brain (Crlt1-KO) [63] and their wild-type littermates control (Crtl1-WT) mice. In our experiments, both male and female mice were utilized as study subjects. To control for the potential influence of sex on our results, we conducted a thorough evaluation of any sex-specific effects across all experiments. Our findings indicated that there were no significant differences observed that could be attributed solely to sex (data not shown). Because the Crtl1 product is essential for cartilage, Crtl1 was disrupted globally (Crtl1−/−) and then reintroduced under the control of the type II collagen-cartilage-specific promoter by crossbreeding with a second transgenic mouse line (Crtl1-Tg), as better described in Czipri et al. The resulting Crtl1−/−/Crtl1-Tg+/+ mice were on a BALB/C background; thus, for this study, they were backcrossed into a C57BL/6 J background for seven generations as described in Carulli et al. [20]. Animals’ genotypes were identified through PCR on tail tissue (P10-P12), with primers for wild-type Crtl1, disrupted Crtl1, and Crtl1 transgene expressed in cartilage [20]. Mice were housed in groups, from two to five animals per cage (60 cm × 40 cm × 20 cm), and maintained in rooms at 22 °C with a standard 12-h light–dark cycle. Food (standard diet, 4RF25 GLP Certificate, Mucedola) and water were available ad libitum and changed weekly. Open-top cages with wooden dust-free bedding were used. All the experiments were carried out according to the directives of the European Community Council (2011/63/EU) and approved by the Italian Ministry of Health. All necessary efforts were made to minimize both stress and the number of animals used. One week before the start of behavioral experiments, all mice were handled daily by the experimenter for 5 min using an open hand approach to minimize anxiety and stress response induced by experimenter manipulation [64]. Fear Conditioning Mice were subjected to an auditory fear conditioning and extinction procedure using a custom-made PVC fear conditioning setup (50 cm × 15 cm × 21 cm). During the test days, the mice were transported in their home cages to a room adjacent to the testing room and left for 2 h before behavioral testing. We used two different contexts: In context A, the walls and the floor were completely black; in context B, the walls had white vertical plastic strips 2.5 cm long, 0.5 cm wide, and 18 cm deep, interposed every 2.5 cm, and white floor. Both chambers were covered with transparent plexiglass lids with a loudspeaker in its center point. The shock grid on the floor was made of stainless steel. Only the grid of context A was electrified by a shock generator (World Precision Instruments, Sarasota, FL) guided by an automated program for CS and US parameter control and footshock delivery. Mice behavior was recorded by a camera controlled by the EthoVision XT 8 software (Noldus Information Technology, The Netherlands). The two chambers were cleaned with 70% ethanol before and after each animal. After each session, mice were housed separately until the end of the test to avoid possible observational fear learning. Conditioning and Extinctions During the habituation day (day 0), mice were placed in context A for 3 min. On day 1 (conditioning), mice were placed again in the context A and conditioned using 5 pairings of the CS (total CS duration 10 s, 7.5 kHz, 80 dB) co-terminating with a US (1 s footshock, 0.6 mA, inter-trial interval: 20 s). On day 2 (early extinction) and day 3 (late extinction), conditioned mice were subjected to the extinction training in context B during which they received 12 unreinforced presentations of the CS on each day. We also tested fear spontaneous recovery and context-dependent fear renewal 7 and 42 days after late extinction using 4 unreinforced presentations of the CS in context B and A respectively. Fear memory retention was tested by submitting an additional group of mice to extinction training 9 days after conditioning, during which they received 10 unreinforced presentations of the CS. Amygdala and Infralimbic Cortex Activation After Early Extinction To assess amygdala and IL activation, after early extinction, a separate group of conditioned Crtl1-KO and Crtl1-WT mice was used. Mice were subjected to the early extinction protocol during which they received 10 unreinforced presentations of the CS. At the end of the extinction training, mice were sacrificed for immunohistochemistry analysis. To isolate the specific effect of associative learning on amygdala and IL activation, we used pseudo-conditioned mice. For pseudoconditioning, on the day of conditioning, mice received the same number of CS as conditioned mice, administered at 1 s interstimulus intervals in context A, without US. Then, mice were placed back in their home cage. After 45 min, the animals were placed again in the context A where they immediately received 5 US at 1 s intervals (1 s footshock, 0.6 mA). This procedure has been designed to make it difficult for animals to associate the US with the CS and the context [30]. On day 2, pseudo-conditioned mice were subjected to the same early extinction protocol of conditioned mice and were then sacrificed for immunohistochemistry. Analysis of Freezing Behavior Recorded videos were manually scored for freezing behavior by two separate experimenters blind to genotype and experimental conditions. Mice were considered to be freezing if no movement was detected for 2 s (defined as the complete absence of movement except for respiratory movements). For all fear conditioning and extinction paradigms, cue evoked freezing behavior was analyzed by calculating the percentage time an animal spent freezing during a given CS presentation, and averages were calculated by pooling freezing across 2 CS presentations if not indicated otherwise. Virtual Fear Conditioning During the virtual fear conditioning, mice were head-fixed. We employed a custom-made apparatus equipped with a 3D printed circular treadmill (diameter: 18 cm) as described in Mazziotti et al. [28]. During each head-fixation session, a curved monitor (24 inches Samsung, CF390) was placed in front of the animal (at a distance of 13 cm). We designed two different virtual environments composed of a γ-linearized procedural virtual corridors written in C# and Unity. The two environments presented sine-wave gratings (context A) or plaid-wave gratings (context B) at different orientations (wall at 0°; floor at 90°), and spatial frequencies (from 0.06 to 0.1 cycles/°). The apparatus was cleaned before and after each animal with 70% ethanol or 1% acetic acid for context A and context B respectively, since the mice may associate the smell with the context. The position of the animal in the virtual corridor was updated using an optical computer mouse, positioned below the circular treadmill, that interfaced with the virtual environment software. As a CS, we used a visual stimulus consisting of a square wave grating patch of 55° (in width and height) of visual space in the binocular portion of the visual field. The grating parameters were as follows: luminance, 8.5 cd/m2; orientation, 0°; contrast, 90%; spatial frequency, 0.1 cycles/°; and drifting, 0.5 cycle/s. A custom-made electrode connected to a shock generator (World Precision Instruments, Sarasota, FL) and controlled by the virtual environment software was positioned on the mouse tail for tail shock delivery (US). For fear conditioning in the virtual environment, mice were introduced gradually to head fixation and to the tail electrode for 5 days (habituation). During the habituation, we performed two sessions of head fixation in which mice were exposed to the two contexts. Each session consisted of 2 min of a uniform gray (luminance, 8.5 cd/m2), to assess the pupil diameter in baseline, and 10 min of the isoluminant virtual environment. After habituation, mice underwent a 4-day fear conditioning training and extinction protocol. On day 1, a group of mice were exposed to context A and a second group to context B and conditioned using 5 pairings of the CS (total CS duration 20 s) co-terminating with a US (2 s tail shock, 0.6 mA, inter-trial interval: 120 s). Fear memory was tested on day 2 in a recall session, presenting 5 unreinforced presentations of the CS in the context opposite of the conditioning one. On day 3 (early extinction) and day 4 (late extinction), conditioned mice were subjected to the extinction training again in the context opposite of the conditioning one during which they received on each day 10 unreinforced presentations of the CS. To test the efficacy of the virtual fear conditioning protocol, we compared the responses of a group of conditioned C57BL/6 J WT mice (shock) with a group of mice that underwent the same conditioning protocol without receiving the tail shock (sham). After each session, mice were housed separately until the end of the test to avoid possible observational fear learning. Pupillometry During the virtual fear conditioning paradigm, we analyzed pupil diameter as a physiological readout of fear response. To record the pupil, we used a USB camera (oCam-5CRO-U, Withrobot Lens: M12 25 mm) connected to a Jetson AGX Xavier Developer Kit (NVIDIA) running a custom Python3 script (30 fps). Real-time pupillometry was performed using MEYE, a convolutional neural network that performs online pupillometry in mice and humans. Pupillometry data has been analyzed using Python 3. All tracks were loaded, and blink removal was applied using the blink detector embedded in MEYE. Blink epochs were filled using linear interpolation and median filtering (0.5 s). The z-score was obtained for each trial using the formula z = (x − x¯baseline)/sbaseline, where x¯baseline and sbaseline are respectively the average and the SD of the baseline. Pupillary Light Reflex For the evaluation of the PLR, we presented 10 s of a white screen (luminance: 30 cd/m2) repeated ten times, interspersed with 50 s of a uniform gray (luminance, 8.5 cd/m2). Each session started with 2 min of a uniform gray (luminance, 8.5 cd/m2) for pupil adaptation. We evaluated the pupil constriction latency (the time needed to reach the minimal pupil size during pupil constriction), the pupil constriction amplitude (maximal relative change in pupil area during constriction), the pupil re-dilation latency (the time needed to reach the maximal pupil size during pupil re-dilation), and the pupil re-dilation amplitude (maximal relative change in pupil area to recover the constriction). Surgery Mice were deeply anesthetized using isoflurane (3% induction, 1.5% maintenance), placed on a stereotaxic frame and head-fixed using ear bars. Prilocaine was used as a local anesthetic for the acoustic meatus. Body temperature was maintained at 37° using a heating pad, monitored by a rectal probe. The eyes were treated with a dexamethasone-based ophthalmic ointment (Tobradex, Alcon Novartis) to prevent cataract formation and keep the cornea moist. Respiration rate and response to toe pinch were checked periodically to maintain an optimal level of anesthesia. A subcutaneous injection of Lidocaine (2%) was performed prior to scalp removal. The skull surface was carefully cleaned and dried, and a thin layer of cyanoacrylate was poured over the exposed skull to attach a custom-made head post that was composed of a 3D printed base equipped with a glued set screw (12 mm long, M4 thread, Thorlabs: SS4MS12). The implant was secured to the skull using cyanoacrylate and UV curing dental cement (Fill Dent, Bludental). At the end of the surgical procedure, the mice recovered in a heated cage. After 1 h, mice were returned to their home cage. Paracetamol was used in the water as antalgic therapy for 3 days. We waited for 7 days before performing head-fixed pupillometry to provide sufficient time for the animals to recover. Immunohistochemistry For immunofluorescence labeling, following the early extinction protocol, conditioned and pseudo-conditioned mice were transcardially perfused with 4% paraformaldehyde in phosphate buffer 1 h after behavioral testing, when Zif268 peaks in its expression [65]. Brains were removed and post-fixed for 24 h at 4 °C in 4% paraformaldehyde and then cryoprotected for 72 h at 4 °C in 20% sucrose and 0.05% sodium azide in PBS, pH 7.4. Brains were then snap-frozen in 2-methylbutane and cryosectioned in OCT using a cryostat (Leica Biosystems, CM 3050S) to obtain 40-μm-thick sections collected in PBS. Free-floating sections were incubated for 2 h at room temperature (RT) in a blocking solution composed of 10% bovine serum albumin (BSA), 0.5% Triton X-100, in PBS. For Zif268 staining, sections were incubated overnight at 4 °C in a solution composed of 10% BSA, 0.3% Triton X-100, and 1:500 rabbit polyclonal anti-Zif268 primary antibody (SantaCruz), in PBS. Sections were then washed for 3 times, 10 min each time. Primary antibody was revealed by incubating sections for 2 h at RT in a solution composed of 1% BSA, 0.1% Triton X-100, and 1:400 goat anti-rabbit AlexaFluor 488 secondary antibody (Invitrogen), in PBS. Sections were then washed for 3 times, 10 min each time and mounted on glass slides and covered with VectaShield mounting medium (Vector). For the quantification of PNNs, mice were transcardially perfused and the 40-µm coronal sections were cut on a freezing microtome (Leica). Slices were incubated for 2 h RT in a blocking solution composed of 3% BSA in PBS. Then, slices were incubated overnight at 4 °C with a solution containing biotinylated Wisteria floribunda Lectin (WFA, B-1355–2, Vector Laboratories, 1:200) and 3% BSA in PBS. On the following day, sections were washed for 3 times in PBS (10 min each) and incubated with a solution of red fluorescent streptavidin (Streptavidin, Alexa Fluor™ 555 con488 jugate, S21381, Thermo Fisher, 1:400) and 3% BSA in PBS for 2 h at RT, and washed again 3 times in PBS and subsequently mounted on glass slides and covered with VectaShield mounting medium. Sections were acquired at 16X using a confocal laser scanning microscope (Leica Biosystems, BM 6000) and digitized with Leica confocal software. From 3 to 5 sections were acquired for each amygdala nucleus and IL for each animal. Zif268-positive cells and PNNs were manually counted using the MetaMorph software and ImageJ software by two separate experimenters blind to genotype and experimental conditions. Statistical Analysis All statistical analyses were performed using GraphPad Prism 7 and Python custom scripts (Pupillometry). Parametric t-test, analysis of variance (ANOVA), and repeated measure-ANOVA (RM-ANOVA) were used. ANOVA was followed by appropriate post hoc tests. Significance was set at P < 0.05 for all tests. Error bars represent s.e.m. in all figures. Supplementary Information Below is the link to the electronic supplementary material.Supplementary file1 (DOCX 3.34 MB) Acknowledgements We gratefully acknowledge NVIDIA Corporation’s support with the Jetson AGX Xavier Developer Kit’s donation for this research. Raffaele Mazziotti was supported by Fondazione Umberto Veronesi. Author Contribution NB and TP designed the research; AP, AV, SM, and DS performed the experiments; AP, AV, VT, and RM performed data analysis; AP, AV, and TP wrote the manuscript. All authors read and approved the final manuscript. Funding Open access funding provided by Consiglio Nazionale Delle Ricerche (CNR) within the CRUI-CARE Agreement. This work was supported by intramural funds of the Scuola Normale Superiore, the "CRONOLAB" project of the PRO3 joint program and the Tuscany Health Ecosystem (THE) Project (CUP I53C22000780001), funded by the National Recovery and Resilience Plan (NRPP), within the NextGeneration Europe (NGEU) Program. Data Availability The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request. Declarations Ethics Approval All the experiments were carried out according to the directives of the European Community Council (2011/63/EU) and approved by the Italian Ministry of Health. Consent to Participate Not applicable. This study does not involve human subjects. Consent for Publication Not applicable. This study does not involve human subjects. Competing Interests The authors declare no competing interests. Andrea Poli and Aurelia Viglione contributed equally to this work. Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References 1. LeDoux JE Emotion circuits in the brain Annu Rev Neurosci 2000 23 155 184 10.1146/annurev.neuro.23.1.155 10845062 2. An B, Kim J, Park K, et al (2017) Amount of fear extinction changes its underlying mechanisms Elife 6. 10.7554/eLife.25224 3. Myers KM Davis M Mechanisms of fear extinction Mol Psychiatry 2007 12 120 150 10.1038/sj.mp.4001939 17160066 4. Bouton ME Context and behavioral processes in extinction Learn Mem 2004 11 485 494 10.1101/lm.78804 15466298 5. Quirk GJ Mueller D Neural mechanisms of extinction learning and retrieval Neuropsychopharmacology 2008 33 56 72 10.1038/sj.npp.1301555 17882236 6. Vouimba R-M Maroun M Learning-induced changes in mPFC–BLA connections after fear conditioning, extinction, and reinstatement of fear Neuropsychopharmacology 2011 36 2276 2285 10.1038/npp.2011.115 21750582 7. Gogolla N Caroni P Lüthi A Herry C Perineuronal nets protect fear memories from erasure Science 2009 325 1258 1261 10.1126/science.1174146 19729657 8. Baldi E Bucherelli C Brain sites involved in fear memory reconsolidation and extinction of rodents Neurosci Biobehav Rev 2015 53 160 190 10.1016/j.neubiorev.2015.04.003 25887284 9. Merz CJ Hamacher-Dang TC Stark R Neural underpinnings of cortisol effects on fear extinction Neuropsychopharmacology 2018 43 384 392 10.1038/npp.2017.227 28948980 10. Amano T Unal CT Paré D Synaptic correlates of fear extinction in the amygdala Nat Neurosci 2010 13 489 494 10.1038/nn.2499 20208529 11. Milad MR Quirk GJ Neurons in medial prefrontal cortex signal memory for fear extinction Nature 2002 420 70 74 10.1038/nature01138 12422216 12. Herry C Ciocchi S Senn V Switching on and off fear by distinct neuronal circuits Nature 2008 454 600 606 10.1038/nature07166 18615015 13. LeDoux JE Coming to terms with fear Proc Natl Acad Sci U S A 2014 111 2871 2878 10.1073/pnas.1400335111 24501122 14. Maren S Quirk GJ Neuronal signalling of fear memory Nat Rev Neurosci 2004 5 844 852 10.1038/nrn1535 15496862 15. Quirk GJ Paré D Richardson R Erasing fear memories with extinction training J Neurosci 2010 30 14993 14997 10.1523/JNEUROSCI.4268-10.2010 21068303 16. Kim JH Richardson R A developmental dissociation in reinstatement of an extinguished fear response in rats Neurobiol Learn Mem 2007 88 48 57 10.1016/j.nlm.2007.03.004 17459734 17. Kim JH Hamlin AS Richardson R Fear extinction across development: the involvement of the medial prefrontal cortex as assessed by temporary inactivation and immunohistochemistry J Neurosci 2009 29 10802 10808 10.1523/JNEUROSCI.0596-09.2009 19726637 18. Mouw JK Ou G Weaver VM Extracellular matrix assembly: a multiscale deconstruction Nat Rev Mol Cell Biol 2014 15 771 785 10.1038/nrm3902 25370693 19. Brückner G Grosche J Schmidt S Postnatal development of perineuronal nets in wild-type mice and in a mutant deficient in tenascin-R J Comp Neurol 2000 428 616 629 10.1002/1096-9861(20001225)428:4<616::AID-CNE3>3.0.CO;2-K 11077416 20. Carulli D Pizzorusso T Kwok JCF Animals lacking link protein have attenuated perineuronal nets and persistent plasticity Brain 2010 133 2331 2347 10.1093/brain/awq145 20566484 21. Carulli D Rhodes KE Fawcett JW Upregulation of aggrecan, link protein 1, and hyaluronan synthases during formation of perineuronal nets in the rat cerebellum J Comp Neurol 2007 501 83 94 10.1002/cne.21231 17206619 22. Leuchs L Schneider M Czisch M Spoormaker VI Neural correlates of pupil dilation during human fear learning Neuroimage 2017 147 186 197 10.1016/j.neuroimage.2016.11.072 27915119 23. Leuchs L Schneider M Spoormaker VI Measuring the conditioned response: a comparison of pupillometry, skin conductance, and startle electromyography Psychophysiology 2019 56 e13283 10.1111/psyp.13283 30259985 24. Korn CW Staib M Tzovara A A pupil size response model to assess fear learning Psychophysiology 2017 54 330 343 10.1111/psyp.12801 27925650 25. Bradley MM Miccoli L Escrig MA Lang PJ The pupil as a measure of emotional arousal and autonomic activation Psychophysiology 2008 45 602 607 10.1111/j.1469-8986.2008.00654.x 18282202 26. Bradshaw J Pupil size as a measure of arousal during information processing Nature 1967 216 515 516 10.1038/216515a0 6057275 27. Nassar MR Rumsey KM Wilson RC Rational regulation of learning dynamics by pupil-linked arousal systems Nat Neurosci 2012 15 1040 1046 10.1038/nn.3130 22660479 28. Mazziotti R, Carrara F, Viglione A, et al (2021) MEYE web app for translational and real-time pupillometry eneuro 8 ENEURO 0122–21 2021 29. Veyrac A Besnard A Caboche J The transcription factor Zif268/Egr1, brain plasticity, and memory Prog Mol Biol Transl Sci 2014 122 89 129 10.1016/B978-0-12-420170-5.00004-0 24484699 30. Sacchetti B Scelfo B Tempia F Strata P Long-term synaptic changes induced in the cerebellar cortex by fear conditioning Neuron 2004 42 973 982 10.1016/j.neuron.2004.05.012 15207241 31. Repa JC Muller J Apergis J Two different lateral amygdala cell populations contribute to the initiation and storage of memory Nat Neurosci 2001 4 724 731 10.1038/89512 11426229 32. Bocchio M Nabavi S Capogna M Synaptic plasticity, engrams, and network oscillations in amygdala circuits for storage and retrieval of emotional memories Neuron 2017 94 731 743 10.1016/j.neuron.2017.03.022 28521127 33. Davis P Zaki Y Maguire J Reijmers LG Cellular and oscillatory substrates of fear extinction learning Nat Neurosci 2017 20 1624 1633 10.1038/nn.4651 28967909 34. Do-Monte FH Quiñones-Laracuente K Quirk GJ A temporal shift in the circuits mediating retrieval of fear memory Nature 2015 519 460 463 10.1038/nature14030 25600268 35. Critchley HD Mathias CJ Dolan RJ Fear conditioning in humans: the influence of awareness and autonomic arousal on functional neuroanatomy Neuron 2002 33 653 663 10.1016/S0896-6273(02)00588-3 11856537 36. Hamm AO Stark R Sensitization and aversive conditioning: effects on the startle reflex and electrodermal responding Integr Physiol Behav Sci 1993 28 171 176 10.1007/BF02691223 8318444 37. Reimer J McGinley MJ Liu Y Pupil fluctuations track rapid changes in adrenergic and cholinergic activity in cortex Nat Commun 2016 7 13289 10.1038/ncomms13289 27824036 38. Giustino TF Maren S Noradrenergic modulation of fear conditioning and extinction Front Behav Neurosci 2018 12 43 10.3389/fnbeh.2018.00043 29593511 39. Pizzorusso T Medini P Berardi N Reactivation of ocular dominance plasticity in the adult visual cortex Science 2002 298 1248 1251 10.1126/science.1072699 12424383 40. Nowicka D Soulsby S Skangiel-Kramska J Glazewski S Parvalbumin-containing neurons, perineuronal nets and experience-dependent plasticity in murine barrel cortex Eur J Neurosci 2009 30 2053 2063 10.1111/j.1460-9568.2009.06996.x 20128844 41. Fawcett JW, Oohashi T, Pizzorusso T (2019) The roles of perineuronal nets and the perinodal extracellular matrix in neuronal function. Nat Rev Neurosci 20(8):451–465. 10.1038/s41583-019-0196-3 42. Sigal YM Bae H Bogart LJ Structural maturation of cortical perineuronal nets and their perforating synapses revealed by superresolution imaging Proc Natl Acad Sci U S A 2019 116 7071 7076 10.1073/pnas.1817222116 30890637 43. Duvarci S Pare D Amygdala microcircuits controlling learned fear Neuron 2014 82 966 980 10.1016/j.neuron.2014.04.042 24908482 44. Ehrlich I Humeau Y Grenier F Amygdala inhibitory circuits and the control of fear memory Neuron 2009 62 757 771 10.1016/j.neuron.2009.05.026 19555645 45. McDonald AJ Cytoarchitecture of the central amygdaloid nucleus of the rat J Comp Neurol 1982 208 401 418 10.1002/cne.902080409 7119168 46. Nitecka L Ben-Ari Y Distribution of GABA-like immunoreactivity in the rat amygdaloid complex J Comp Neurol 1987 266 45 55 10.1002/cne.902660105 3429715 47. Li H Penzo MA Taniguchi H Experience-dependent modification of a central amygdala fear circuit Nat Neurosci 2013 16 332 339 10.1038/nn.3322 23354330 48. Krettek JE Price JL Amygdaloid projections to subcortical structures within the basal forebrain and brainstem in the rat and cat J Comp Neurol 1978 178 225 254 10.1002/cne.901780204 627625 49. Pitkänen A Stefanacci L Farb CR Intrinsic connections of the rat amygdaloid complex: projections originating in the lateral nucleus J Comp Neurol 1995 356 288 310 10.1002/cne.903560211 7629320 50. Savander V Go CG LeDoux JE Pitkänen A Intrinsic connections of the rat amygdaloid complex: projections originating in the basal nucleus J Comp Neurol 1995 361 345 368 10.1002/cne.903610211 8543667 51. Sotres-Bayon F Quirk GJ Prefrontal control of fear: more than just extinction Curr Opin Neurobiol 2010 20 231 235 10.1016/j.conb.2010.02.005 20303254 52. Quirk GJ Russo GK Barron JL Lebron K The role of ventromedial prefrontal cortex in the recovery of extinguished fear J Neurosci 2000 20 6225 6231 10.1523/JNEUROSCI.20-16-06225.2000 10934272 53. Hugues S Chessel A Lena I Prefrontal infusion of PD098059 immediately after fear extinction training blocks extinction-associated prefrontal synaptic plasticity and decreases prefrontal ERK2 … Synapse 2006 10.1002/syn.20291 16786530 54. Burgos-Robles A Vidal-Gonzalez I Santini E Quirk GJ Consolidation of fear extinction requires NMDA receptor-dependent bursting in the ventromedial prefrontal cortex Neuron 2007 53 871 880 10.1016/j.neuron.2007.02.021 17359921 55. Laurent V Westbrook RF Inactivation of the infralimbic but not the prelimbic cortex impairs consolidation and retrieval of fear extinction Learn Mem 2009 16 520 529 10.1101/lm.1474609 19706835 56. Gallo FT Katche C Morici JF Immediate early genes, memory and psychiatric disorders: focus on c-Fos, Egr1 and Arc Front Behav Neurosci 2018 12 79 10.3389/fnbeh.2018.00079 29755331 57. Ressler KJ Paschall G Zhou X-L Davis M Regulation of synaptic plasticity genes during consolidation of fear conditioning J Neurosci 2002 22 7892 7902 10.1523/JNEUROSCI.22-18-07892.2002 12223542 58. Malkani S Rosen JB Specific induction of early growth response gene 1 in the lateral nucleus of the amygdala following contextual fear conditioning in rats Neuroscience 2000 97 693 702 10.1016/S0306-4522(00)00058-0 10842014 59. Baumgärtel K Genoux D Welzl H Control of the establishment of aversive memory by calcineurin and Zif268 Nat Neurosci 2008 11 572 578 10.1038/nn.2113 18425121 60. Herry C Mons N Resistance to extinction is associated with impaired immediate early gene induction in medial prefrontal cortex and amygdala Eur J Neurosci 2004 20 781 790 10.1111/j.1460-9568.2004.03542.x 15255988 61. Morellini F Sivukhina E Stoenica L Improved reversal learning and working memory and enhanced reactivity to novelty in mice with enhanced GABAergic innervation in the dentate gyrus Cereb Cortex 2010 20 2712 2727 10.1093/cercor/bhq017 20194688 62. Romberg C Yang S Melani R Depletion of perineuronal nets enhances recognition memory and long-term depression in the perirhinal cortex J Neurosci 2013 33 7057 7065 10.1523/JNEUROSCI.6267-11.2013 23595763 63. Czipri M Otto JM Cs-Szabó G Genetic rescue of chondrodysplasia and the perinatal lethal effect of cartilage link protein deficiency J Biol Chem 2003 278 39214 39223 10.1074/jbc.M303329200 12732630 64. Hurst JL West RS Taming anxiety in laboratory mice Nat Methods 2010 7 825 826 10.1038/nmeth.1500 20835246 65. Lonergan ME Gafford GM Jarome TJ Helmstetter FJ Time-dependent expression of Arc and zif268 after acquisition of fear conditioning Neural Plast 2010 2010 139891 10.1155/2010/139891 20592749
PMC010xxxxxx/PMC10236376.txt
==== Front J Nematol J Nematol jofnem jofnem Journal of Nematology 0022-300X 2640-396X Sciendo 37274092 jofnem-2023-0014 10.2478/jofnem-2023-0014 Research Paper Reaction of Winter Cover Crops to Meloidogyne enterolobii and Glasshouse Bioassay for Evaluating Utility in Managing M. enterolobii in Soybeans Saha Neel Schwarz Tanner Mowery Samantha Gorny Adrienne M. [email protected] Department of Entomology and Plant Pathology, North Carolina State University, Raleigh, NC 27695, USA This paper was edited by Peter DiGennaro. 2 2023 2 5 2023 55 1 2023001411 10 2022 © 2023 Saha et al., published by Sciendo 2023 Saha et al., published by Sciendo https://creativecommons.org/licenses/by/4.0/ This work is licensed under the Creative Commons Attribution 4.0 International License. Abstract Meloidogyne enterolobii is an invasive and highly aggressive root-knot nematode pathogen impacting the Southeastern United States. Winter cover cropping may be a cost-effective method for reducing populations of M. enterolobii in between summer cash crops, yet a gap in the knowledge remains about the response of these cover crops to M. enterolobii and their utility in suppressing nematode populations prior to a cash crop. A “two-step” glasshouse bioassay was performed to evaluate eight winter cover crops popular in North Carolina for their direct response to M. enterolobii infection, and to quantify their effect in reducing nematode populations for the following soybean plants. Data on cover crop root galling, soybean root galling, soybean shoot fresh weight, soybean root fresh weight, eggs per gram of soybean root, and a modified reproductive factor were collected. Cereal cover crops did not display root galling, and there was significantly less root galling in those soybean plants following cereal winter cover crops when compared to those following broadleaf winter cover crops. Broadleaf winter cover crops resulted in significantly higher eggs per gram of soybean root and modified reproductive factor in the soybean plants, compared to cereal cover crops and non-inoculated controls. Results from this study suggest that cereal winter cover crops may be poor-hosts to M. enterolobii and may significantly reduce M. enterolobii populations before a soybean crop, compared to broadleaf winter cover crops. This study lays the groundwork for management recommendations and future field trials to assess management of M. enterolobii through winter cover cropping. Keywords biomass Guava root-knot nematode interaction management reproductive factor ==== Body pmcThe plant-parasitic nematode Meloidogyne enterolobii Yang and Eisenback, 1983 (syn. M. mayaguensis) is an invasive species to the Southeastern region of the contiguous United States (U.S.), having first been reported in Florida in 2001, and subsequently in North Carolina and South Carolina (Brito et al., 2004; Ye et al., 2013; Rutter et al., 2019). The species has been described as highly virulent, even reproducing on crop cultivars possessing root-knot nematode resistance genes such as tomato (Mi-1 gene), soybean (Mir1 gene), and bell pepper (N gene) (Williamson and Roberts, 2009; Castagnone-Sereno, 2012). Because of the lack of genetic host resistance available in commercial crop varieties, the current management options in the Southeastern U.S. rely heavily on cultural and chemical tactics, including the use of fumigant and non-fumigant nematicides, rotation to non-host crops, and sanitation measures to prevent further spread of the nematode. The use of non-host cover crops has recently gained attention by many producers, agricultural extension personnel, and crop consultants, as it offers a possible economically viable alternative to costly fumigants. Further, the use of cover crops provides pathogen management options for organic production or low input systems. Yet the utility and impact of cultural tactics such as cover cropping for suppression of M. enterolobii populations remains understudied. Cover crops are grown between cash crop cycles or incorporated within cash crop cycles to cover land that is not utilized for primary production. Cover cropping has been implemented to improve soil structure, reduce soil erosion, increase organic matter, increase available nitrogen, suppress weed populations, and suppress pathogens and pests like nematodes within the context of various cropping systems (Sainju and Singh, 1997; Lu et al., 2000; Hartwig and Ammon, 2002). The use of cover crops for plant-parasitic nematode suppression mainly focuses on the host status of the cover crop, yet certain cover crops (e.g., brassica species and some sudangrasses) may also be leveraged for their biofumigant properties as “green manures” (Kruger et al., 2013). Non-host or poor-host cover crops serve to reduce nematode populations and reproduction between cash crop cycles, resulting in lower initial nematode populations and reduced pathogen pressure on the main cash crop. Previous work has reported the host status of several summer cover crop plant species to Meloidogyne enterolobii. Sunn hemp (Crotalaria juncea) and sorghum-Sudangrass (Sorghum biocolor x S. sudanense) have been identified as non-host to M. enterolobii (Bui and Desaeger, 2021; de Brida et al., 2018). Reported host summer cover crop species include sunflower (Helianthus annuus) and cowpea (Vigna unguiculata) (Bui and Desaeger, 2021). Further, Brito et al. (2007) reported the mustard ‘Florida Broad Leaf’ (Brassica oleracae) to be a good host to M. enterolobii. Khanal and Harshman (2022) corroborated sunflower and cowpea as host and sunn hemp and sorghum sudangrass as non-host to M. enterolobii. They further reported that buckwheat (Fagopyrum esculentum) and sesame (Sesamum indicum) were poor-host, while pearl millet (Pennisetum glaucum) and grain sorghum (Sorghum bicolor) were non-host (Khanal and Harshman, 2022). In contrast to summer cover crops, winter cover crops are grown in rotation with summer cash crops. Winter cover crops are generally planted during the fall after harvest of the main cash crops, allowed to grow over the winter months, and are then either harvested or terminated in early spring before the planting of another main cash crop. Winter cover crops serve to utilize residual NO3, with some species of winter cover crop serving to increase the organic matter content in the soil and, particularly with leguminous species, enhancing available nitrogen (Sainju and Singh, 1997). Winter cover crops are generally species that can survive cooler and cold temperatures, including cereals such as oat, wheat, rye, and barley, as well as leguminous species such as crimson clover, winter pea, and some vetches. Several species of brassica, including mustard and canola (rapeseed), are also utilized as winter cover crops by some farmers. In the Southeastern U.S., oats (Avena sativa) and wheat (Triticum aestivum) are commonly used as winter cover crops, and these have been demonstrated to be non-host to M. enterolobii (de Brida et al., 2018). Although several summer cover crops have been identified as host or non-host to M. enterolobii, the response of winter cover crops to M. enterolobii remains less explored. Further, a gap remains in the understanding of how the use of a winter cover crop may translate into reducing M. enterolobii initial populations for a following cash crop. The objectives of this study were (1) to evaluate winter cover crops common in agricultural systems of the southeastern U.S. for their response to Meloidogyne enterolobii and (2) to quantify through a greenhouse bioassay the utility of these winter cover crops in reducing initial populations of the nematode for a successive soybean crop. This study was intended to inform future field-based experiments to explore the impact of winter cover crops on M. enterolobii management in soybean under field conditions. Materials and Methods Nematode culture and extraction: A culture of Meloidogyne enterolobii was established from infected soil collected from a soybean field in Wilson County, North Carolina in 2017, and has been continuously maintained on tomato plants (cv. ‘Rutgers’) in the glasshouse at North Carolina State University. Identity of this isolate is periodically checked using species-specific amplification primer sets MeF/MeR (Long et al., 2006) and Ment17F/Ment17R (Kiewnick et al., 2015). The potting media used for maintaining cultures was a steam-sterilized 1:1 soil-to-sand mixture. To extract M. enterolobii eggs for inoculating cover crop plants, tomato culture plant root systems were destructively harvested and rinsed in tap water to remove excess soil, then soaked in a 10% household bleach solution (6% available NaOCl) for 45 sec to break down the gelatinous matrix surrounding mature egg sacs, aided by hand through gentle agitation (Hussey and Barker, 1973). Nematode eggs released from the culture plant root system were recovered by passing the extraction solution through stacked 250 μm, 75 μm, and 25 µm mesh sieves and rinsing with fresh tap water to remove residual NaOCl. Eggs were collected from the 25 μm mesh sieve in a final volume of 35 ml. To remove additional soil and plant tissue debris, 15 ml of a 2.05 M sucrose solution was added to each sample, gently inverted, then centrifuged at 1,000 rcf for 5 min. Nematode eggs suspended in the sucrose solution were recovered by passing the supernatant through a 25 µm mesh sieve, rinsing with fresh tap water, collecting, and suspending in tap water. Extracted egg samples were counted under a compound inverted microscope (Nikon Instruments Inc., Melville, NY, USA) at 100x magnification. Extracted eggs were stored at 4°C for 24 h prior to inoculation. Plant growth, inoculation, and assessment: Seeds of eight different winter cover crop species were obtained (Johnny’s Select Seeds, Winslow, ME, USA) (Table 1). Seeds (approx. 1 g per pot) were sewn into 15.2 cm diameter clay pots filled with steam-sterilized 1:1 soil-to-sand mix. Pots were amended with slow-release fertilizer (Osmocote Smart-Release Plant Food, The Scotts Company, Marysville, OH, USA) and watered daily. Three weeks after sowing, pots were inoculated with 10,000 M. enterolobii eggs per pot by creating three small holes, approx. 0.5 cm wide by 3 cm deep equidistant from each other and pipetting inoculum into the holes. Eight pots of each cover crop species were included, with each pot considered as a replicate. Non-inoculated pots of crimson clover (Trifolium incarnatum) and ryegrass (Lolium multiflorum) were included as non-inoculated control. Cover crops were maintained in the glasshouse for 60 days post inoculation at 26– 28°C with no supplemental lighting. At 60 days post inoculation, cover crop plants were destructively harvested and the presence or absence of root galling was recorded, whereas presence was defined as root galling severity of ≥ 1 on a severity index scale of 0–10 (Bridge and Page, 1980). From each pot, 200 g of soil was removed and transferred to a new 15.2 cm diameter clay pot; cover crops’ roots were discarded after root galling evaluation and were not transferred with the soil. These new pots were topped with additional, fresh soil mixture (steam-sterilized 1:1 soil-to-sand mixture) and three soybean seeds (Glycine max; variety not stated; Sustain Seed+Soil, Minster, OH, USA) were seeded into each pot. Soybean plants were germinated, thinned to one plant per pot, and maintained in the greenhouse at 26–28°C with no supplemental lighting. At 60 days post germination, soybean plants were destructively harvested. Plants were gently removed from the pots, preserving integrity of the root system, and roots rinsed with tap water to remove excess soil. Soybean fresh shoot biomass and fresh root biomass were measured by cutting the plants at the crown and weighing foliage and roots separately. Soybean root galling severity was estimated on a scale from 0–100% using a scale modified from Bridge and Page (1980). Nematode eggs were then extracted from soybean roots following the procedure described above and counted under a compound inverted microscope. The total number of eggs per plant was estimated by calculating the total eggs per ml of extraction solution, then multiplying by the total volume of extraction solution (50 mL). The number of eggs per gram of soybean root tissue was calculated by dividing the total number of eggs in each sample by the weight of the soybean fresh root system. A modified reproductive factor (RFmod) was calculated for nematode populations on each soybean plant by dividing the total number of eggs extracted from the soybean plant (final population; Pf) by the initial number of eggs inoculated to the preceding cover crop (initial population; Pi, n = 10,000). The experiment was performed twice (with Trial 1 being performed from April to September 2021, and Trial 2 being performed from November 2021 to April 2022). Table 1. Host status of winter cover crops and response of soybean after inoculation of the preceding cover crop with the root-knot nematode Meloidogyne enterolobii, Trial 1. Means (n = 8) followed by the same letter within each column are not significantly different at the 0.05 level. Plant species Common name Root galling in cover cropa Soybean shoot weight (g) Soybean root weight (g) Soybean root galling severity Hordeum vulgare Barley −         41.4 bc         52.0 abc             0 c Trifolium incarnatum Crimson clover +         24.8 d         38.8 c         59.4 ab Guillenia flavescens Yellow mustard +         31.2 cd         35.9 c         48.1 b Avena nuda Streaker hulless oats −         42.6 bc         51.1 abc             0 c Lolium multiflorum Ryegrass −         50.8 b         57.2 ab             0 c Secale cereale Winter rye −         53.0 ab         49.9 abc             0 c Vinca villosa Hairy vetch +         26.4 cd         51.4 abc         67.9 a Triticum aestivum Spring wheat −         43.3 bc         44.1 bc             0 c Lolium multiflorum Ryegrass (non-inoculated) −         65.9 a         67.0 a             0 c Trifolium incarnatum Crimson clover (non-inoculated) −         65.6 a         63.7 a             0 c P-value b           <0.0001             0.0369           <0.0001 MSEc         166.4         295.7         169.2 a Presence or absence of root galling was evaluated whereas presence was defined as root galling severity of ≥1 on a severity index scale of 0–10 (Bridge and Page 1980). ‘-’ indicates absence of galling, ‘+’ indicates presence of galling. b Significance value of Fisher’s least significant difference. c Mean squared error. Statistical analysis: Data sets from Trial 1 and Trial 2 were assessed to determine if they could be combined by comparing means using a Wilcoxon rank sum test (Mann and Whitney, 1947) and evaluating homogeneity of variances using Fisher’s F-Test (Markowski and Markowski, 1990). Evidence of variability was observed; therefore, the data sets from each trial were analyzed separately. The main effects of the preceding winter cover crop on soybean fresh shoot and root biomass, soybean root galling severity, the number of eggs per gram of soybean root, and RFmod were evaluated using analysis of variance within RStudio (v. 3.6.2; R Core Team, 2019) with replicate as a random effect, and removing a single outlier with high leverage in Trial 1. For each trial dataset, the assumptions of normality and homogeneity of variances in the data were verified by visualization of the data and performing the Shapio-Wilks (function ‘shapio.test’) and Levene’s (function ‘leveneTest’ within the ‘car’ package) tests. Where significant differences were identified, Fisher’s Least Significant Difference test was used to separate means (function ‘LSD.test’ within the ‘agricolae’ package). Results Winter cover crops as host to Meloidogyne enterolobii: In both Trial 1 and Trial 2, root galling symptoms typical of infection by Meloidogyne enterolobii were observed on roots of crimson clover (Trifolium incarnatum), yellow mustard (Guillenia flavenscens), and hairy vetch (Vinca villosa). No root galling was observed on the roots of barley (Hordeum vulgare), Streaker hulless oats (Avena nuda), ryegrass (Lolium multiflorum), winter rye (Secale cereale), or spring wheat (Triticum aestivum) (Tables 1,3; Figure 1). No root galling was observed in the non-inoculated control plants in either trial (Tables 1,3). Figure 1: Effect of preceding winter cover crop on soybean roots following (A) yellow mustard, (B) barley, and (C) ryegrass. Significant root galling was observed in soybean roots following yellow mustard, crimson clover, and hairy vetch, while no root galling was observed on the roots of barley, Streaker hulless oats, ryegrass, winter rye, or spring wheat. Table 3. Host status of winter cover crops and response of soybean after inoculation of the preceding cover crop with the root-knot nematode Meloidogyne enterolobii, Trial 2. Means (n = 8) followed by the same letter within each column are not significantly different at the 0.05 level. Plant species Common name Root galling in cover cropa Soybean shoot weight (g) Soybean root weight (g) Soybean root galling severity Hordeum vulgare Barley −         14.92         31.59           0.3 d Trifolium incarnatum Crimson clover +         10.09         16.20         26.7 b Guillenia flavescens Yellow mustard +         10.09         18.83         14.6 c Avena nuda Streaker hulless oats −         11.39         20.32             0 d Lolium multiflorum Ryegrass −           6.35         17.71           0.3 d Secale cereale Winter rye −         12.65         22.15           0.3 d Vinca villosa Hairy vetch +           5.03         13.58         46.2 a Triticum aestivum Spring wheat −           8.31         20.01           0.3 d Lolium multiflorum Ryegrass (non-inoculated) −         11.99         20.12              0 d Trifolium incarnatum Crimson clover (non-inoculated) −         11.47         17.41              0 d P-valueb           0.070 (ns)           0.063 (ns)           <0.0001 MSEc         33.1         88.0         121.5 a Presence or absence of root galling was evaluated whereas presence was defined as root galling severity of ≥ 1 on a severity index scale of 0–10 (Bridge and Page 1980). ‘-’ indicates absence of galling, ‘+’ indicates presence of galling. b Significance value of Fisher’s least significant difference c Mean squared error Response of soybean to inoculation of preceding winter cover crop: In Trial 1, there were significant differences in soybean shoot fresh weight among treatment groups of preceding winter cover crops, with the non-inoculated control treatments having the greatest soybean shoot fresh weight, and crimson clover having the lowest soybean shoot weight (Table 1). In Trial 2, there were no significant differences in soybean shoot weight among treatment groups (Table 3). In Trial 1, there were significant differences in soybean root fresh weight among treatment groups, with crimson clover, yellow mustard, and spring wheat having significantly lower soybean root weights than the non-inoculated control plants (Table 1). In Trial 2, there were no significant differences in soybean root fresh weight among treatment groups (Table 3). In Trial 1, there were significant differences in soybean root galling severity, with crimson clover, yellow mustard, and hairy vetch having significantly greater severity of root galling than all other treatment groups (Table 1). In Trial 2, an identical trend was observed, with crimson clover, yellow mustard, and hairy vetch having significantly greater severity of root galling than all other treatment groups (Table 3). Nematode reproduction on soybean plants after inoculation of preceding winter cover crop: In Trial 1, there were significant differences in the number of eggs per gram of soybean root, with soybean plants following crimson clover, yellow mustard, and hairy vetch having significantly more eggs per gram of root than all other treatment groups (Table 2). An identical trend was observed in Trial 2, with soybean plants following crimson clover, yellow mustard, and hairy vetch having significantly more eggs per gram of root than all other treatment groups (Table 4). In both Trial 1 and Trial 2, there were significant differences in RFmod, with soybean plants following crimson clover, yellow mustard, and hairy vetch having a significantly higher RFmod than all other treatment groups (Tables 2,4). There was a low level of eggs per gram of soybean root in non-inoculated controls (Table 2,4). Table 2. Reproduction of Meloidogyne enterolobii on soybean after inoculation of preceding winter cover crop plants, Trial 1. Means (n = 8) followed by the same letter within each column are not significantly different at the 0.05 level. Plant species Common name Eggs per gram of soybean root tissue RF moda Hordeum vulgare Barley             10.1 b         0.05 b Trifolium incarnatum Crimson clover         1914.2 a         7.84 a Guillenia flavescens Yellow mustard         3104.3 a         7.19 a Avena nuda Streaker hulless oats             15.2 b         0.07 b Lolium multiflorum Ryegrass               6.8 b         0.03 b Secale cereale Winter rye               6.9 b         0.04 b Vinca villosa Hairy vetch         2195.3 a       11.33 a Triticum aestivum Spring wheat               6.5 b         0.03 b Lolium multiflorum Ryegrass (non-inoculated)               3.5 b         0.02 b Trifolium incarnatum Crimson clover (non-inoculated)               4.3 b         0.02 b P-value b               0.0001     < 0.0001 MSEc             2380380     19.5 a Modified Reproductive Factor, calculated by dividing the total number of eggs extracted from soybean root systems (final population), by the initial number of eggs inoculated to the preceding cover crop (initial population; herein, 10,000). b Significance value of Fisher’s least significant difference. c Mean squared error. Table 4. Reproduction of Meloidogyne enterolobii on soybean after inoculation of different proceeding winter cover crop plants, Trial 2. Means (n = 8) followed by the same letter within each column are not significantly different at the 0.05 level. Plant species Common name Eggs per gram of soybean root tissue RF moda Hordeum vulgare Barley             13.7 c         0.06 c Trifolium incarnatum Crimson clover         6041.5 a         7.13 a Guillenia flavescens Yellow mustard         1452.4 bc         3.71 b Avena nuda Streaker hulless oats               3.7 c         0.01 c Lolium multiflorum Ryegrass             10.8 c         0.02 c Secale cereale Winter rye               5.0 c         0.01 c Vinca villosa Hairy vetch         2476.6 b         2.15 bc Triticum aestivum Spring wheat               9.2 c         0.02 c Lolium multiflorum Ryegrass (non-inoculated)             16.4 c         0.02 c Trifolium incarnatum Crimson clover (non-inoculated)                   0 c             0 c P-valueb           < 0.0001           0.0003 MSEc         2396198         11.1 a Modified Reproductive Factor, calculated by dividing the total number of eggs extracted from soy bean root systems (final population), by the initial number of eggs inoculated to the preceding cover crop (Initial population; herein, 10,000). b Significance value of Fisher’s least significant difference. c Mean squared error. Discussion Meloidogyne enterolobii is an emerging nematode pathogen in the Southeastern U.S., where it poses the risk of significantly affecting agricultural production. Soybean (Glycine max) is susceptible to M. enterolobii, and the nematode species is able to overcome and infect even those cultivars possessing the Mir1 root-knot nematode resistance gene (Castagnone-Sereno, 2012). In the authors’ experience, growers’ awareness of nematode pathogens and the interest in using cultural pathogen management tactics (such as winter cover cropping) has increased in North Carolina in recent years. However, gaps in knowledge remain, specifically regarding the utility of winter cover crops in reducing M. enterolobii populations in between cash crops. The objectives of the present study were to evaluate several winter cover crops common in North Carolina for their response to M. enterolobii, and to quantify through a glasshouse bioassay the suppressive effect of these winter cover crops on M. enterolobii initial populations to a following soybean crop. In this study, all of the cereal winter cover crops assessed, including barley, Streaker hulless oats, ryegrass, winter rye, and spring wheat, did not produce galling symptoms (all less than 0.3 severity across both trials). Furthermore, the eggs per gram of soybean root and RFmod were also significantly lower following cereal cover crops than the three broadleaf cover crops assessed. The broadleaf cover crops assessed, including yellow mustard, hairy vetch, and crimson clover, all produced significantly more severe soybean root galling, eggs per gram of soybean root, and increased RFmod. In addition, the broadleaf cover crops had directly observable root galling. This response supports the conclusion that the broadleaf winter cover crops evaluated here may be good-host to M. enterolobii, while the cereal winter cover crop species may be poor-host to M. enterolobii. In both Trial 1 and Trial 2, there was a small number of eggs recovered from the roots of soybean plants from non-inoculated control cover crop plants. This may have been due to splash dispersal of infested soil from neighboring inoculated pots, or may have been the result of possible contamination of shared extraction tools such as sieves during the harvesting and evaluation step of the experiments. Due to this result in the control plants, we refrain from concluding that any of the cover crop species tested are complete or strong non-host to M. enterolobii, but rather conclude that they are likely poor-host to the nematode. De Brida et al. (2018) reported that several cultivars of wheat (Triticum aestivum), oat (Avena sativa), and sorghum (Sorghum bicolor) were non-host to M. enterolobii, and reported a significantly lower galling index, egg mass index, and reproductive factor compared to a tomato control. Bui and Desaeger (2021) evaluated several summer cover crops for their host status to M. enterolobii. They found that sunn hemp (Crotalaria juncea) and sorghum-Sudangrass (Sorghum bicolor x S. sudanense) were poor-host to the nematode, while sunflower (Helianthus annuus) and cowpea (Vigna unguiculata) were good-host (Bui and Desaeger, 2021). More recently, Khanal and Harshman (2022) also performed an experiment in which the host statuses of several summer cover crops were evaluated to M. enterolobii. They reported that warm-season grass cover crops, including three types of millet, sorghum, and sorghum-Sudangrass. were non-host to the nematode (Khanal and Harshman, 2022). Similarly to the results of this study, they identified broadleaf cover crops that were host to the nematode, including sunflower, cowpea, and buckwheat (Fagopyrum esculentum) (Khanal and Harshman, 2022). Cetintas et al. (2007) found common vetch (Vinca sativa) to be a host to M. enterolobii, consistent with the finding in this study that the related species hairy vetch (Vinca villosa) resulted in a high degree of root-galling due to M. enterolobii. The results of these previous studies taken together with the results of the present study strongly suggest that cereal crops are poor-to non-host to M. enterolobii and producers may consider them as a good rotational or cover crop choice for fields impacted by M. enterolobii, while broadleaf cover crops may support M. enterolobii reproduction and exacerbate field infestation. In Trial 1 and Trial 2, a similar trend in the results was observed in the response variables related to soybean growth among the treatment groups consisting of the preceding cover crop. For example, in both trials, soybeans following hairy vetch had the highest average root galling severity score, followed by those of crimson clover and then yellow mustard (Tables 1,2). Of the cereal winter cover crops, no galling was observed in Trial 1, and only a minute degree of galling was observed in Trial 2. In contrast, a higher degree of variability was observed in response variables related to nematode reproduction. For example, there was no consistent trend between trials when considering RFmod among treatment groups. Overall, soybean plant growth and nematode reproduction across all treatment groups was numerically lower in Trial 2 when compared to Trial 1. This may be because Trial 2 was performed during the winter months in North Carolina. Glasshouse bioassays remain an important tool for nematologists, and bioassays have been used to evaluate treatments such as nematicides (e.g., de Oliveira Silva et al., 2019), or to assess crop genotypic response or pathogen virulence (e.g., Thies, 2011; Hallmann and Kiewnick, 2018), or in diagnostic capacities (e.g., Gugino et al., 2008; Gorny et al., 2021). Bioassays that use sensitive indicator crops may have a lower limit of detection compared to direct extraction, due to second-stage juveniles present infecting the plant promptly, and additional time in the assay for eggs to hatch and infect (Gugino et al., 2008; Duncan, 1991). Thus, bioassay results may support robust management decisions. Here, we demonstrate the effectiveness of a “two-step” bioassay for evaluating both the direct response of the cover crop and its ability to support or suppress nematode populations that may impact a following cash crop. However, there remain certain shortcomings of the glasshouse bioassay. For example, the temperature at which the cover crops were grown and the length of time for which they were maintained may not fully replicate field conditions, and thus, field populations of M. enterolobii may behave and reproduce differently than what was observed in this study. Furthermore, the design of this experiment only evaluated root galling in the cover crop; a direct measurement of nematode reproduction (such as quantifying egg production) in the cover crop was not obtained. Marquez et al. (2022) noted that simply evaluating the production of root galls is not sufficient to fully determine host status. However, we believe the complimentary data of soybean growth and nematode reproduction under the soybean crop supplement our evaluation of cover crop host status. Additionally, once cover crop roots were removed from the soil to evaluate root galling, these roots were not placed back into the soil that was then transferred to a new pot for soybean growth. This may have effectively removed a portion of the nematode population contained within egg masses attached to the cover crop roots. This is in contrast to a field setting, where winter cover crop roots would likely be incorporated through tillage and any egg masses on the roots would remain in the soil. A broad purview should be maintained, and evaluating these winter cover crop species under field conditions is necessary to support management recommendations. Prior to this work, the response of several winter cover crops to M. enterolobii was not known. To the best of our knowledge, this is the first data to begin evaluating the reaction of the winter cover crops barley, crimson clover, yellow mustard, Streaker hulless oats, ryegrass, winter rye, spring wheat, and hairy vetch to M. enterolobii. In summary, this study identified several winter cover crop species with little galling in response to M. enterolobii and provided a quantitative estimate of the suppressive effect of these winter cover crops on M. enterolobii populations in a following soybean crop. This study will provide preliminary management recommendations for agricultural producers for managing this nematode with winter cover cropping and will also provide direction for future field studies on managing M. enterolobii with winter cover cropping. Acknowledgements The authors thank P. Bennett Jeffreys, Kelci Cox, and Grace Artus for excellent technical assistance. Funding for this work was provided by a USDA NIFA Specialty Crops Research Initiative grants (Award No. 2021-51181-35865) and by the NC Soybean Producers Association (Award Nos. 21-111 and 22-101). ==== Refs Literature Cited Bridge J. Page S. L. J. 1980 . Estimation of root-knot nematode infestation levels on roots using a rating chart . Tropical Pest Management 26 : 296 - 298 . Brito J. A. Powers T. O. Mullin P. G. Inserra R. N. Dickson D. W. 2004 . Morphological and molecular characterization of Meloidogyne mayaguensis isolates from Florida . Journal of Nematology 36 : 232 - 240 . 19262811 Brito J. A. Stanley J. D. Mendes M. L. Cetintas R. Dickson D. W. 2007 . Host status of selected cultivated plants to Meloidogyne mayaguensis in Florida . Nematropica 37 : 65 - 71 . Bui H. X. Desaeger J. A. 2021 . Host suitability of summer cover crops to Meloidogyne arenaria, M. enterolobii, M. incognita and M. javanica . Nematology 24 : 171 - 179 . Castagnone-Sereno P. 2012 . Meloidogyne enterolobii (= M. mayaguensis): profile of an emerging, highly pathogenic, root-knot nematode species . Nematology 14 : 133 - 138 . Cetintas R. Kaur R. Brito J. A. Mendes M. L. Nyczepir A. P. Dickson D. W. 2007 . Pathogenicity and reproductive potential of Meloidogyne mayaguensis and M. floridensis compared with three common Meloidogyne spp. Nematropica 37 : 21 - 31 . de Brida A. L. Castro B. M. de C. Zanuncio J. C. Serrão J. E. Wilcken S. R. S. 2018 . Oat, wheat and sorghum cultivars for the management of Meloidogyne enterolobii . Nematology 20 : 169 - 173 . de Oliveira Silva J. Loffredo A. da Rocha M. R. Becker J. O. 2019 . Efficacy of new nematicides for managing Meloidogyne incognita in tomato crop . Journal of Phytopathology 167 : 295 - 298 . Duncan L. W. 1991 . Current options for nematode management . Annual Review of Phytopathology 29 : 469 - 490 . Gorny A. M. Hay F. S. Pethybridge S. J. 2021 . Response of potato cultivars to the northern root-knot nematode, Meloidogyne hapla, under field conditions in New York State, USA . Nematology 23 : 425 - 433 . Gugino B. K. Ludwig J. W. Abawi G. S. 2008 . An on-farm bioassay for assessing Meloidogyne hapla infestations as a decision management tool . Crop Protection 27 : 785 - 791 . Hallmann J. Kiewnick S. 2018 . Virulence of Meloidogyne incognita populations and Meloidogyne enterolobii on resistance cucurbitaceous and solanaceous plant genotypes . Journal of Plant Diseases and Protection 125 : 415 - 425 . Hartwig N. L. Ammon H. U. 2002 . Cover crops and living mulches . Weed Science 50 : 688 - 699 . Hussey R. S. Barker K. R. 1973 . A comparison of methods of collecting inocula of Meloidogyne spp., including a new technique . Plant Disease 57 : 1025 - 1028 . Khanal C. Harshman D. 2022 . Evaluation of summer cover crops for host suitability of Meloidogyne enterolobii . Crop Protection 151 : 105821 . Kiewnick S. Frey J. E. Braun-Kiewnick A. 2015 . Development and validation of LNA-based quantitative real-time PCR assays for detection and identification of the root-knot nematode Meloidogyne enterolobii in complex DNA backgrounds . Phytopathology 105 : 1245 - 1249 . 25775103 Kruger D. H. M. Fourie J. C. Malan A. P. 2013 . Cover crops with biofumigation properties for the suppression of plant-parasitic nematodes: A review . South African Journal of Enology and Viticulture 34 : 287 - 295 . Long H. Liu H. Xu J. H. 2006 . Development of a PCR diagnostic for the root-knot nematode Meloidogyne enterolobii . Acta Phytopathologica Sinica 36 : 109 - 115 . Lu Y.-C. Watkins K. B. Teasdale J. R. Abdul-Baki A. A. 2000 . Cover crops in sustainable food production . Food Reviews International 16 : 121 - 157 . Mann H. B. Whitney D. R. 1947 . On a test of whether one of two random variables is stochastically larger than the other . Annals of Mathematical Statistics 18 : 50 - 60 . Markowski C. A. Markowski E. P. 1990 . Conditions for the effectiveness of a preliminary test of variance . The American Statistician 44 : 322 - 326 . Marquez J. Hajihassani A. Davis R. F. 2022 . Evaluation of summer and winter cover crops for variations in host suitability for Meloidogyne incognita, M. arenaria, and M. javanica . Nematology 24 : 841 - 854 . Rutter W. B. Skantar A. M. Handoo Z. A. Mueller J. D. Aultman S. P. Agudelo P. 2019 . Meloidogyne enterolobii found infecting root-knot nematode resistant sweetpotato in South Carolina, United States . Plant Disease 103 : 775 . Sainju U. M. Singh B. P. 1997 . Winter cover crops for sustainable agricultural systems: influence on soil properties, water quality, and crop yields . HortScience 32 : 21 - 28 . Thies J. A. 2011 . Virulence of Meloidogyne incognita to expression of N gene in pepper . Journal of Nematology 43 : 90 - 94 . 22791917 Williamson V. M. Roberts P. A. 2009 . Mechanisms and genetics of resistance . Pp. 301 - 325 Perry R. N. Moens M. Starr J. Root-knot nematodes . Wallingford, UK : CABI Publishing . Yang B. Eisenback J. D. 1983 . Meloidogyne enterolobii n. sp. (Meloidogynidae), a root-knot nematode parasitizing pacara earpod tree in China . Journal of Nematology 15 : 381 - 391 . 19295822 Ye W. M. Koenning S. R. Zhuo K. Liao J. L. 2013 . First Report of Meloidogyne enterolobii on cotton and soybean in North Carolina, United States . Plant Disease 97 : 1262 .
PMC010xxxxxx/PMC10236533.txt
==== Front RSC Adv RSC Adv RA RSCACL RSC Advances 2046-2069 The Royal Society of Chemistry 37274392 d2ra07959c 10.1039/d2ra07959c Chemistry Comment on “Which fraction of stone wool fibre surface remains uncoated by binder? A detailed analysis by time-of-flight secondary ion mass spectrometry and X-ray photoelectron spectroscopy” by Hirth et al., 2021, RSC Adv., 11, 39545, DOI: 10.1039/d1ra06251d https://orcid.org/0000-0001-7279-1292 Okhrimenko Denis V. a Ceccato Marcel b Tougaard Sven c Foss Morten b Pezennec Eric d Solvang Mette a a ROCKWOOL A/S Denmark b Interdisciplinary Nanoscience Center (iNANO), Aarhus University Denmark c University of Southern Denmark Denmark d Knauf Insulation Belgium 2 6 2023 30 5 2023 2 6 2023 13 24 1668816692 13 12 2022 25 5 2023 This journal is © The Royal Society of Chemistry 2023 The Royal Society of Chemistry https://creativecommons.org/licenses/by/3.0/ The article mentioned in the title of this comment paper reports on an investigation of the organic binder presence and distribution on stone wool fibres with surface sensitive techniques (X-ray photoelectron spectroscopy (XPS), QUASES XPS modelling, time-of-flight secondary ion mass spectrometry (ToF-SIMS) mapping) and attempts to correlate the results with fibre performance in in vitro acellular biosolubility tests. However, the study has assumptions, hypothesis and results that do not take into account the recognised science and regulations on biopersistence of stone wool fibres, limitations of the utilized surface sensitive techniques and modelling approach and it contains a contradiction with biosolubility experiments. In this comment article, we discuss these points, propose improved QUASES XPS modelling and present recent ToF-SIMS mapping results that reflect biosolubility behaviour of the stone wool fibres. A thin organic layer (binder, oil, adventitious carbon) on a stone wool fibre surface is incomplete and inhomogeneous and leaves the bare fibre surface available for dissolution in lung fluid. pubstatusPaginated Article ==== Body pmcIntroduction Hirth and colleagues1 have recently investigated the distribution of organic material (binder and mineral oil) on stone wool fibres. The work follows up previous publications by the authors: Wohlleben et al., 2017 2 and Sauer et al., 2021.3 The starting point for their publications is the authors' view that hazard assessment of man-made vitreous fibres (MMVF) is solely based on biodurability measurements of naked fibres (i.e. without binder). Similar to the previous discussion,4 we would like to bring attention to the fact that in vitro acellular biodurability tests either on fibres with or without binder are not relevant for the hazard assessment and regulations on mineral wool fibres. Actually, MMVF hazard assessment includes investigation of fibre biopersistence via in vivo animal studies with typically nasal inhalation or intratracheal installations of fibres produced without binder5,6 (Note Q of the European Regulation (EC) No. 1272/2008 (CLP) (EC2008)) and epidemiological studies on workers, where the impact of fibres produced with binder is studied, both recognised at international and European level.7–13 Despite this, the papers1–3 attempted to find differences in the in vitro acellular behaviour of fibres with and without binder, using binder removal techniques that modify fibre chemistry,14,15 wettability and thus likely solubility.16,17 The paper1 explored the distribution of binder (presumably phenol-urea-formaldehyde, PUF) on stone wool fibres and tried to find a correlation between dissolution rate of stone wool measured in a simulated lung fluid (phagolysosomal simulant fluid, PSF) and the amount and thickness of organic material on the fibre surface. The article1 reports the use of surface sensitive techniques, such as X-ray photoelectron spectroscopy (XPS), time-of-flight secondary ion mass spectroscopy (ToF-SIMS) mapping and modelled XPS data with QUASES software.18 The general findings of the paper about organic matter obtained with XPS and QUASES XPS modelling are in line with previously published results on stone wool samples.19,20 However, we would like to stress several points concerning the assumptions and hypotheses in the publication,1 analytical techniques limitations in spatial resolutions and interpretation of results and as well present the newest results using QUASES XPS and ToF-SIMS on stone wool samples with PUF binder. Discussion The assumptions on biopersistence assessment Hirth et al.1 state that fibre biodurability is currently assessed on “naked” fibres (i.e. produced without binder) because there is an assumption that fibres produced with organic matter (binder) would not have a completely coating of the fibre, and that this would rather be localised in the areas where fibres enter into contact and thus leaving large fraction of the fibre surface uncoated. It has to be mentioned that for biopersistence tests, fibres produced without binder are traditionally used also for other reasons. In in vivo studies,5,6 fibres without binder are recommended because aerosol sizing, fibre diameter measurements and sterilization of the test material are impaired by the presence of binder.† Binder presence also causes fibres to agglomerate, which may result in suffocation of the animals after intratracheal instillation. As in real inhalation scenario, respirable fibres are present as single fibres, while larger agglomerates are not able to reach the alveolar region of the lung, this should also be avoided in the in vivo tests by using fibres without binder.† However, in an earlier in vivo study it is shown that stone wool fibres produced with and without binder perform similarly (Wagner et al., 1982,21 Experiment 1 for stone wool fibres injected intrapleurally). Fibre safety is also largely explored by epidemiological studies7–13 at manufacturing sites, where no adverse effects of stone wool fibres as produced, meaning possibly with binder,31 are found on workers. Epidemiology is the first type of studies that IARC22 is using to investigate carcinogenicity of substances, including stone wool fibres, followed by in vivo investigations. Thus, today, fibres' biosolubility in vitro (acellular and cellular) is not the key indicator to assess the stone wool fibres hazard assessment. Incomplete information on composition of test material and organic matter No details are provided in the paper regarding composition of the stone wool fibres, unlike in previous publications by the authors.2,3 The lack of information about fibre composition makes it difficult to follow the dependence of the fibre dissolution rate on the inorganic composition of the fibres, which authors concluded to be the main factor. The authors1 state that they expect phenolic resin to be commonly used as a binder based on their finding of traces of nitrogen but no further information on the organic binder chemical composition is provided. In the paper the binder appears to be treated as a classic organic molecule without further differentiation of other binder components (such as oil, coupling agent etc.). We further note that providing SEM images (a standard technique for the study of microstructure on fibre surfaces) of the stone wool fibres would have been beneficial and would have enabled the distinction between micrometre size areas with binder and the rest of the surface. Resolution of ToF-SIMS and XPS results We would like to highlight that the used low ToF-SIMS mapping resolution could give the impression that the signal coming from the fibre surface is dominated by carbon from oil and binder (as the oil and binder are on the top of the fibres). We do not think that this is sufficient documentation in the paper to conclude, that fibres are coated almost at 100%. In another recent study,16 yet with different ToF-SIMS resolution (300 × 300 μm; 128 × 128 pixels), ion source (Bi1+) and binder applied to the fibres (sugar-based binder, SBB), it was possible to observe a signal from the fibre substrate itself (Al+), indicating that binder does not completely coat the fibre surface. In Fig. 1 we present recent ToF-SIMS imaging results on stone wool fibres with PUF binder from Barly et al., 2019 23 (F3 sample, 3.6 wt% PUF binder, 0.1 wt% mineral oil) performed with the same ToF-SIMS settings as in Okhrimenko et al., 2022.16 The signal from fibre substrate (Al+, Fig. 1a, a1) is dominating over the signal from organic layer originated likely from PUF binder (C7H7+) and oil (C3H7+) in many areas on the fibre surface (Fig. 1b, b1, c and c1), indicating that PUF binder and mineral oil coverage is neither uniform nor complete. Fig. 1 ToF-SIMS imaging results of F3 stone wool fibres (3.6 wt% PUF binder, 0.1 wt% mineral oil) from Barly et al., 2019:23 (a-a1) Al+; (b-b1) overlay of Al+ (red) and C7H7+ (green); (c-c1) overlay of Al+ (red) and C3H7+ (green) and (d-d1) overlay Al+ (red) and total intensity (green). The images (a–d) are obtained with Bi1+ ion source and other settings similar to those in Okhrimenko et al., 2022.16 The images (a1–d1) are obtained with Bi3++ ion source similar to Hirth et al., 2021.1 It is important to consider the influence of the ToF-SIMS mapping spatial resolution and other settings on the conclusions about binder distribution on the fibres. This would help to understand why fibres with different binders (PUF and SBB) and fibres without any binder perform similar in dissolution tests as found in.16,23 We also note that the journal number and year for the work by Barly et al., 2019 23 in the Notes and references of Hirth et al., 2021 1 are not correct. For the XPS results, no survey (wide-scan) spectra of the studied stone wool samples are presented. This does not allow to check the presence of additional chemical elements in the different stone wool samples that were compared. Complete survey spectra would have enabled the authors and the readers to get a first qualitative view of comparison between the samples. Moreover, it would be beneficial for the readers if it was acknowledged that both methods, XPS and ToF-SIMS, are extremely sensitive towards contaminations by adventitious carbon, which can originate from fibre storage and handling, as well as from apparatus in situ, and interfere with the performed analysis, reducing its representativity. Limitations of the modelling of surface layer thickness We would like to note that the results of the QUASES XPS modelling to support the hypothesis that binder and mineral oil completely shield the surface of the fibres should be interpreted with greatest caution. QUASES XPS modelling works the best if reference spectra are available, i.e. in this case this would be a spectrum of “naked” fibres without organic matter on their surface. In the absence of reference spectra, several models describing experimental XPS spectra are possible. The authors1 chose to simulate the surface layer in a similar way as in the study by Okhrimenko et al., 2018,20i.e. as a uniform carbon layer with thickness 1–3 nm on top of the fibres. While the approach of using the background of the Si and Al XPS peaks to determine the binder distribution can be relevant at the considered thicknesses (≤10 nm), we note that binder droplets can be thicker (30–50 nm) and they are “blind” spots for QUASES analysis. With QUASES software version 7.5, the results from Okhrimenko et al., 2018 20 can be re-evaluated using the automation option. QUASES v.7.5 uses the simplex method to determine the combination of all structure parameters which gives the minimum root mean square, RMS, between the spectrum and the background in the desired energy range. Using the automated structure determination facility that varies the structure until the RMS deviation in the 1270 to 1310 eV energy range reaches a minimum, we observed an improved fit of the XPS spectra when the software applies the model where 20% of the fibre surface remained uncoated (Fig. 2a, RMS 8.3 × 10−4), compared to the previous fitting with a 4 nm thick uniform layer presented in Okhrimenko et al., 2018 20 (Fig. 2b, RMS 20.3 × 10−4). This reduction in RMS is substantial but it should also be supplemented by a visual inspection of the spectra: the fit in Fig. 2a is seen to be virtually perfect in the full energy range from 1270 to 1295 eV, whereas there are clear deviations in Fig. 2b in this energy range. Any other structural model also gives substantially worse fits to the background. Fig. 2 XPS spectra QUASES fitting using model: (a) with 80% of the surface covered with 7.1 nm organic layer and 20% of the uncoated surface; (b) uniform organic layer of 4.0 nm thick (Okhrimenko et al., 2018).20 Presence of the organics-free areas correlates with ToF-SIMS mapping results presented in our work for PUF treated fibres and by Okhrimenko et al., 2022 16 for SBB treated fibres and also with the similar dissolution behaviour of the fibre material irrespective if binder was applied or not23 and which type of the binder was applied.16 To sum up, the realization in the paper1 analyses without reference to neither the substrate nor the binders and without proof of the goodness of the modelling (only one example provided with narrow energy range, 1225–1345 eV) and the fact that their XPS spectrum was recorded with a rather low signal to noise level does not allow to ensure the solidity of the results and the findings. In addition, we note that information about how the general background is accounted for in the analysed energy region would have been very helpful. In QUASES XPS this is done by subtracting a straight which is fitted to the spectrum on the high energy side of the region. However, if peaks are present in this region (which is clearly the case here as seen in Fig. 2), the slope of this line can be uncertain and this adds to the uncertainty of the analysis. We avoid this problem by including in the analysis all peaks on the high energy side in the full energy range. Fibre dissolution and dissolution rate evaluation The paper1 acknowledges that binder thickness is not a predictor for dissolution rates. This is confirmed by Fig. 7a in the paper,1 showing no correlation between dissolution rates and total binder content determined with thermogravimetry (TGA). It even shows that there is a reverse dependence of the dissolution rate on organic layer thickness in Fig. 7b of the paper1 (i.e. the higher the thickness of the organic layer, the faster fibres dissolve). The results contradict with the authors'1 hypothesis that the fibres are completely coated with the binder. The same authors previously demonstrated that mass loss of fibre with binder can reach up to 10% after 30 days of dissolution in PSF2 and higher in liquids with citrate (up to 100% within few days3). Taking into account such mass losses during dissolution and no time delay of the dissolution in the beginning of the tests, one can hardly expect any surface shielding effects by binder/organic layer shortly after beginning the dissolution test. The observed fibre (with binder) dissolution can be explained by the fact that the organic layer on the fibre surface is incomplete and inhomogeneous in reality and leaves bare surface available for dissolution, as we have just shown with the newest QUASES XPS modelling and ToF-SIMS results presented here for PUF- and recently for SBB-treated fibers.16 Therefore, Hirth et al., 2021 1 results confirmed that binder presence cannot affect the dissolution of the stone wool fibres. It was shown that it is the inorganic chemical composition of the fibre that is among of the prime factors in in vivo pathogenicity5,6 and for in vitro cellular24,25 and acellular26–28 dissolution rates. Besides that, experimental conditions16,28–30 (e.g. fluid flow rate to sample surface area ratio, fluid composition, temperature, pH, dynamic or batch experiment and sample preparation) are crucial for the determination of the dissolution rates in in vitro acellular studies. Conclusions In conclusion, we find that there are several methodological limitations in the article, which might provide an incorrect image of the dissolution and biosolubility of stone wool fibres. The conclusions made by Hirth et al., 2021 1 are in contrast with the existing science and regulations on biopersistence of stone wool fibres and other MMVF fibres. The present authors hope that provided comments, the additional examples of QUASES XPS modelling approach and application of ToF-SIMS mapping technique would support a better understanding of the biosolubility of the stone wool fibres, accepted terminology and existing regulations on MMVF biopersistence. Conflicts of interest The authors declare following competing financial interest(s): D. V. O. and M. S. are employees of ROCKWOOL A/S, a company producing stone wool fibres. E. P. is employed by Knauf Insulation, a company producing stone and glass wool fibres. Supplementary Material The current authors would like to thank M. Rosborg from ROCKWOOL A/S; E. Chaudan, Q. Hérault, A. de Reydellet, S. Brossard, J. Voronkoff, R. Faure and T. Cretin, from Saint-Gobain; N. Drnovšek from Knauf Insulation; N. Mascaraque and A. Aznar from Ursa Insulation S.A., R. Suovanen from Paroc; J. de Cruz from European Insulation Manufacturers Association and other EURIMA members for discussion and help with the article preparation. † Personal Communication 2019, F. Schulz, Fraunhofer Institute for Toxicology and Experimental Medicine ITEM Hannover, Germany. ==== Refs Notes and references Hirth S. Waindok H. Wohlleben W. RSC Adv. 2021 11 39545 39552 10.1039/D1RA06251D 35492464 Wohlleben W. Waindok H. Daumann B. Werle K. Drum M. Egenolf H. Part. Fibre Toxicol. 2017 14 29 10.1186/s12989-017-0210-8 28784145 Sauer U. G. Werle K. Waindok H. Hirth S. Hachmöller O. Wohlleben W. Chem. Res. Toxicol. 2021 34 3 780 792 33464877 Koch C. Okhrimenko D. V. Solvang M. Aznar A. Pezennec E. Chaudan E. Magrane Francesch J. Lindberg P. Herault Q. Alami Badissi A. Chem. Res. Toxicol. 2021 34 7 1697 1698 34236839 Kamstrup O. Davis J. M. G. Ellehauge A. Guldberg M. Ann. Occup. Hyg. 1998 42 3 191 199 10.1016/S0003-4878(98)00019-2 9684559 Kudo Y. Shibata K. Miki T. Ishibashi M. Hosoi K. Sato T. Kohyama N. Aizawa Y. Environ. Health Prev. Med. 2005 10 239 248 10.1007/BF02897697 21432126 Baan R. A. Grosse Y. Mutat. Res., Fundam. Mol. Mech. Mutagen. 2004 553 1–2 43 58 10.1016/j.mrfmmm.2004.06.019 15288532 Boffetta P. Saracci R. Andersen A. Bertazzi P. A. Chang-Claude J. Cherrie J. Ferro G. Frentzel-Beyme R. Hansen J. Olsen J. Plato N. Teppo L. Westerholm P. Winter P. D. Zocchetti C. Epidemiology 1997 8 3 259 268 10.1097/00001648-199705000-00006 9115020 Enterline P. E. Marsh G. M. Henderson V. Callahan C. Ann. Occup. Hyg. 1987 31 4B 625 656 3450231 Marsh G. M. Buchanich J. M. Youk A. O. J. Occup. Environ. Med. 2001 43 9 803 808 10.1097/00043764-200109000-00009 11561363 Marsh G. M. Enterline P. E. Stone R. A. Henderson V. L. J. Occup. Med. 1990 32 7 594 604 10.1097/00043764-199007000-00009 2167956 Simonato L. Fletcher A. C. Cherrie J. W. Andersen A. Bertazzi P. Charnay N. Claude J. Dodgson J. Esteve J. Frentzel-Beyme R. Gardner M. J. Jensen O. Olsen J. Teppo L. Winkelmann R. Westerholm P. Winter P. D. Zocchetti C. Saracci R. Ann. Occup. Hyg. 1987 31 4B 603 623 3450230 Egnot N. S. Benson S. M. Vater M. F. Hazan R. Patel O. Marsh G. M. Regul. Toxicol. Pharmacol. 2020 112 104585 10.1016/j.yrtph.2020.104585 31991162 Kirkegaard L. F. Korsgaard M. Yue Y. Mørup S. Glass Sci. Technol. 2005 78 1 1 6 Smedskjaer M. M. Yue Y. Deubener J. Gunnlaugsson H. P. Mørup S. J. Non-Cryst. Solids 2010 356 6–8 290 298 10.1016/j.jnoncrysol.2009.12.004 Okhrimenko D. V. Bøtner J. A. Riis H. K. Ceccato M. Foss M. Solvang M. Toxicol. in Vitro 2022 78 105270 10.1016/j.tiv.2021.105270 34757181 Stranghoener M. Dultz S. Behrens H. Schippers A. Geochim. Cosmochim. Acta 2020 273 85 98 10.1016/j.gca.2020.01.005 Tougaard S. , QUASES-Tougaard Inc., Odense, Denmark, 2022, http://www.quases.com Okhrimenko D. V. Budi A. Ceccato M. Johansson D. B. Lybye D. Bechgaard K. Stipp S. L. S. Polym. Degrad. Stab. 2021 183 109431 10.1016/j.polymdegradstab.2020.109431 Okhrimenko D. V. Thomsen A. B. Ceccato M. Johansson D. B. Lybye D. Bechgaard K. Tougaard S. Stipp S. L. S. Polym. Degrad. Stab. 2018 152 86 94 10.1016/j.polymdegradstab.2018.04.001 Wagner J. C. , Berry G. B. , Hill R. J. , Munday D. E. and Skidmore J. W. , in Biological Effects of Man made Mineral Fibres. Proceedings of a symposium, 1982, World Health Organization, Copenhagen, 1984, vol. 2 , pp. 207–233 Andersen A. , Axten C. , Bernstein D. M. , Brochard P. , Castranova V. , Donaldson K. , Dumortier P. , Everitt J. I. , Gustavsson P. , Hesterberg T. W. , Jaurand M. C. , Kane A. B. , Marsh G. M. , Morimoto Y. , Muhle H. , Oberdörster G. , Olin S. , Savolainen K. M. and Schneider T. , Man-made Vitreous Fibres – IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, IARC Press, Lyon, France, 2002, pp. 1–433 Barly S. H. Q. Okhrimenko D. V. Solvang M. Yue Y. Stipp S. L. S. Chem. Res. Toxicol. 2019 32 12 2398 2410 31682107 Luoto K. Holopainen M. Karppinen K. Perander M. Savolainen K. Environ. Health Perspect. 1994 102 5 103 107 Luoto K. Holopainen M. Savolainen K. Ann. Occup. Hyg. 1995 39 6 855 867 10.1016/0003-4878(95)00052-6 Guldberg M. de Meringo A. Kamstrup O. Furtak H. Rossiter C. Regul. Toxicol. Pharmacol. 2000 32 184 189 10.1006/rtph.2000.1418 11067774 Guldberg M. Jensen S. L. Knudsen T. Steenberg T. Kamstrup O. Regul. Toxicol. Pharmacol. 2002 35 217 226 10.1006/rtph.2001.1523 12052006 Potter R. M. Hoffman J. W. Hadley J. G. Inhalation Toxicol. 2023 35 40 47 10.1080/08958378.2023.2166167 36648029 Guldberg M. Christensen V. R. Perander M. Zoitos B. Koenig A. R. Sebastian K. Ann. Occup. Hyg. 1998 42 233 243 10.1016/S0003-4878(98)00026-X Guldberg M. Madsen A. L. Sebastian K. Fellman J. Potter R. Corning O. Bauer J. Manville J. Searl A. Maquin B. Glass Sci. Technol. 2003 76 199 205 Solvang M. Okhrimenko D.V. Koch C. J. Occup. Environ. Hyg. 2023 10.1080/15459624.2023.2205470
PMC010xxxxxx/PMC10237333.txt
==== Front FEMS Yeast Res FEMS Yeast Res femsyr FEMS Yeast Research 1567-1356 1567-1364 Oxford University Press 37173282 10.1093/femsyr/foad029 foad029 Research Article AcademicSubjects/SCI01150 Editor's Choice Using the AKAR3-EV biosensor to assess Sch9p- and PKA-signalling in budding yeast https://orcid.org/0000-0002-4679-3013 Botman Dennis Systems Biology Lab, AIMMS/A-LIFE, Vrije Universiteit Amsterdam, 1081 HV Amsterdam, The Netherlands Kanagasabapathi Sineka Systems Biology Lab, AIMMS/A-LIFE, Vrije Universiteit Amsterdam, 1081 HV Amsterdam, The Netherlands Savakis Philipp Systems Biology Lab, AIMMS/A-LIFE, Vrije Universiteit Amsterdam, 1081 HV Amsterdam, The Netherlands https://orcid.org/0000-0003-3929-0423 Teusink Bas Systems Biology Lab, AIMMS/A-LIFE, Vrije Universiteit Amsterdam, 1081 HV Amsterdam, The Netherlands Corresponding author. Systems Biology Lab, AIMMS/A-LIFE, Vrije Universiteit Amsterdam, De Boelelaan 1108, 1081 HV Amsterdam, The Netherlands. E-mail: [email protected] These authors contributed equally to this work. 2023 12 5 2023 12 5 2023 23 foad02928 10 2022 01 5 2023 10 5 2023 02 6 2023 © The Author(s) 2023. Published by Oxford University Press on behalf of FEMS. 2023 https://creativecommons.org/licenses/by/4.0/ This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited. Abstract Budding yeast uses the TORC1-Sch9p and cAMP-PKA signalling pathways to regulate adaptations to changing nutrient environments. Dynamic and single-cell measurements of the activity of these cascades will improve our understanding of the cellular adaptation of yeast. Here, we employed the AKAR3-EV biosensor developed for mammalian cells to measure the cellular phosphorylation status determined by Sch9p and PKA activity in budding yeast. Using various mutant strains and inhibitors, we show that AKAR3-EV measures the Sch9p- and PKA-dependent phosphorylation status in intact yeast cells. At the single-cell level, we found that the phosphorylation responses are homogenous for glucose, sucrose, and fructose, but heterogeneous for mannose. Cells that start to grow after a transition to mannose correspond to higher normalized Förster resonance energy transfer (FRET) levels, in line with the involvement of Sch9p and PKA pathways to stimulate growth-related processes. The Sch9p and PKA pathways have a relatively high affinity for glucose (K0.5 of 0.24 mM) under glucose-derepressed conditions. Lastly, steady-state FRET levels of AKAR3-EV seem to be independent of growth rates, suggesting that Sch9p- and PKA-dependent phosphorylation activities are transient responses to nutrient transitions. We believe that the AKAR3-EV sensor is an excellent addition to the biosensor arsenal for illuminating cellular adaptation in single yeast cells. The AKAR3-EV biosensor measures the signalling activity of two major kinases -PKA and Sch9- in single living budding yeast cells. Saccharomyces cerevisiae FRET biosensor intracellular signalling PKA Sch9 single-cell Kyoto University 10.13039/501100005683 KU Leuven 10.13039/501100004040 ==== Body pmcIntroduction One universal aspect of life is change, and the ability to adapt to it is a major determinant of reproductive success. For the unicellular organism Saccharomyces cerevisiae (budding or baker’s yeast), there is no exception. In the wild, this yeast lives on fruit and tree bark, where it endures feast-famine cycles (Liti 2015, Jouhten et al. 2016). In industry, domesticated yeast also experiences changing conditions when used in large-scale fermenters with inoculation transitions and poor stirring (Lara et al. 2006, Wehrs et al. 2019). For budding yeast, nutrient availability is a major environmental parameter that sets the investment in metabolism, stress resistance, and proliferation (Broach 2012, Conrad et al. 2014). Understanding the logic of these circuitries is a key challenge in cell biology. Also for industry, control over this response, e.g. by selection of certain preferred subpopulation or removal of undesired populations, could increase production efficiencies (Xiao et al. 2016). Nutrient adaptations can be captured on population level using bulk assays, but these methods may mask the true responses produced by the intracellular signalling circuits inside single cells. Thus, a more profound characterization of the cellular adaptation of budding yeast to nutrient changes at the single-cell level is highly desired. Regarding carbon sources, fruit sugars are preferred by yeast, and therefore it has developed various pathways to sense and adapt to changes in the availability of these substrates (Rolland et al. 2002, Santangelo 2006, Rødkaer and Faergeman 2014). The cAMP-PKA pathway is one of the major signalling cascades that get activated when cells encounter fermentable sugars (Eraso and Gancedo 1985, Beullens et al. 1988, Botman et al. 2021). cAMP production is activated via two routes (Casperson et al. 1985, Kataoka et al. 1985, Broek et al. 1987, Beullens et al. 1988, Mbonyi et al. 1988, Munder and Küntzel 1989, Engelberg et al. 1990, van Aelst et al. 1990, van Aelst et al. 1991, Pardo et al. 1993, Colombo et al. 1998, Yun et al. 1998, Kraakman et al. 1999, Rolland et al. 2000, Rolland et al. 2001, Lemaire et al. 2004, Kim et al. 2013): via import and metabolism of sugars and via extracellular sensing of glucose and sucrose by the G-protein coupled receptor Gpr1p. These two inputs give a transient increase in cAMP, which causes the dissociation of Bcy1p (a PKA regulator) from the PKA subunits Tpk1-3. This finally results in an increase of PKA activity; PKA is a major effector kinase in yeast, accounting for 75%–90% of the cellular changes during a transition from a glucose-derepressed (respiratory) state to a fermentable glucose-repressed (fermentative) state (Thevelein 1994, Rolland et al. 2002, Winderickx et al. 2003, Santangelo 2006, Zaman et al. 2009, Broach 2012, Kunkel et al. 2019). The evoked transition gives a radical change in yeast physiology; cells change their metabolism from respiratory to fully fermentative, repress metabolic pathways for other carbon sources, decrease their stress resistance, and make large investments in ribosomal biogenesis. The TORC1-Sch9 cascade is a second major signalling cascade in yeast cells (Crauwels et al. 1997, Roosen et al. 2004, Slattery et al. 2008, van Zeebroeck et al. 2021). In contrast to PKA, which gets activated by mostly fermentable sugars, Sch9p is activated by TORC1p when a complete palette of nutrients for growth (such as amino acids, nitrogen, phosphate, and a fermentable carbon source) is available (Crauwels et al. 1997, Conrad et al. 2017, van Zeebroeck et al. 2021). Although the two pathways can operate independently (Zurita-Martinez and Cardenas 2005), they have many positive interactions and can rescue each other’s activities, probably via the large overlap in their targets: Sch9 and PKA proteins both phosphorylate the RRxT motif (Reinders et al. 1998, Ma et al. 1999, Plank et al. 2020, Plank 2022). Activation of Sch9p activates 90% of the genes that PKA also activates (Zaman et al. 2009), making Sch9p as important as PKA for proper cellular decision making in yeast. Currently, PKA and Sch9p activities are difficult to measure in (single) cells: the most common method is measuring the activity of trehalase, a PKA and Sch9p target, or using kemptide as a substrate (Crauwels et al. 1997, Roosen et al. 2004). However, these bulk assays lack single-cell information and show only static activity levels of PKA and Sch9p activity. Studies suggest that the PKA activity in yeast is more a transient phenomenon activated during (mostly) sugar transitions and that the TORC1-Sch9p axis dictates the steady-state growth mode of a yeast cell (Crauwels et al. 1997, Kunkel et al. 2019). Dynamic readouts are needed to substantiate these interesting hypotheses. Moreover, single-cell dynamics allows to test for heterogeneity during nutrient transitions—a phenomenon that we did not find for cAMP dynamics (Botman et al. 2021), but it is unknown if heterogeneity exists more downstream. Furthermore, the sensitivity of Sch9p and PKA activity with respect to glucose remains to be characterized. Finally, the basal RRxT phosphorylation status of the cell at various growth rates is also poorly characterized. Here, we implemented and tested the mammalian PKA sensor AKAR3-EV in yeast to provide a tool that can help to enlarge our understanding of PKA and Sch9 signalling. The sensor allowed us to measure the single-cell dynamics of the PKA and Sch9p phosphorylation status in a robust and accurate manner. We found large heterogeneous responses of yeast cells for some nutrient transitions, which we did not previously find in cAMP dynamics. The detected heterogeneity potentially affects the overall cellular state since these two kinases constitute the vast majority of the cellular transition during a transition to a fermentable carbon source. Furthermore, our data implies that the phosphorylation status is not related to growth rate. How the two kinases regulate the cellular transitions at a single-cell level can be studied in more depth using this sensor. Material and methods Sensor construction AKAR3EV with YPET-eCFP as a FRET pair was kindly provided by Dr. Aoki (Komatsu et al. 2011). The sensor was amplified using KOD One™ PCR Master Mix (Toyobo, Osaka, Japan) with 5′-ATGCTAGCACGGAGCTCACTGAATTCGGCATGGTGAG-3′ and 5′-ATGGATCCACGGTCGACACTTTAATCCAGAGTCAGGCG-3′ as forward and reverse primers, respectively. Next, the PCR product and the pDRF1-GW plasmid were digested using BamHI-HF and NheI-HF (New England Biolabs, Ipswich, MA, USA), and the PCR product was ligated into the plasmid using T4 DNA ligase (New England Biolabs), which yield pDRF1-GW AKAR3-EV, containing the AKAR3-EV sensor under PMA1 promotor expression. AKAR3-EV-NR was constructed by performing a PCR with KOD One™ PCR Master Mix on pDRF1-GW AKAR3-EV with 5′- TATTCCGGATTGAGGCGCGCGGCCCTGGTTGACGGCGGCCGCATGGTGAGCAAGGGC-3′ as a forward primer and 5′-ATGGATCCACGGTCGACACTTTAATCCAGAGTCAGGCG-3′ as a reverse primer. The PCR product and pDRF1-GW AKAR3-EV were digested using Kpn2I and XbaI (Thermo Fisher Scientific, Waltham, MA, USA). Afterwards, the PCR product was ligated in the digested pDRF1-GW AKAR3-EV, replacing the sensor domain and eCFP with the non-responding (T506A) sensor domain and eCFP. pDRF1-GW eCFP was made by performing a PCR with KOD One™ PCR Master Mix on AKAR3-EV with FW primer 5′-ATGCTAGCATGGTGAGCAAGGGCG-3′ and RV primer 5′- TAGCGGCCGCTTACTTGTACAGCTCGTCCATGCCG -3′, after which the PCR product and pDRF1-GW were digested using NheI-HF and NotI-HF (New England Biolabs). Finally, the PCR product was ligated into pDRF1-GW using T4 DNA ligase. Yeast transformation Strains used in this study are listed in Table 1. Strains were transformed by resuspending yeast cells from either a YPAD plate or a selective plate in a transformation mixture containing 240 µL PEG 3350 (50% w/v), 40 µL 1 M LiAC, 10 µL Salmon Sperm DNA (10 mg/mL, Sigma–Aldrich, Stl. Louis, MO, USA), and 500–1000 ng plasmid DNA. Next, cells were incubated for 15–20 minutes at 42°C. Afterwards, the cells were centrifuged at 13 000 g for 30 seconds, the supernatant was removed, and 150 µL water was added. Finally, the cells were plated on selective plates. Table 1. Used strains in this study. Strain Characteristics Source W303-1A WT MATa, leu2-3112, trp1-1, can1-100, ura3-1, ade2-1, his3-11,15 In-house W303-1A + pDRF1-GW W303-1A MATa + pDRF1-GW In-house W303-1A + pYX212 AKAR3 W303-1A MATa + pYX212 AKAR3 This work, pYX212 AKAR3 was provided by Sonia Colombo (Colombo et al. 2017, Colombo et al. 2022) W303-1A + pDRF1-GW AKAR3-EV W303-1A MATa + pDRF1-GW AKAR3-EV This work W303-1A + pDRF1-GW AKAR3-EV-NR W303-1A MATa + pDRF1-GW AKAR3-EV-NR This work W303-1A + pDRF1-GW eCFP W303-1A MATa + pDRF1-GW eCFP This work W303-1A + pYES2ACT1 pHluorin W303-1A MATa + pYES2ACT1 pHluorin In-house, pYES2 pHluorin was provided by Gertien Smits (Orij et al. 2009, Orij et al. 2012) W303-1A CYR1K1876M W303-1A MATa + CYR1K1876M (Vanhalewyn et al. 1999) W303-1A CYR1K1876M + pDRF1-GW AKAR3-EV W303-1A MATa CYR1K1876M + pDRF1-GW AKAR3-EV This work W303-1A sch9Δ + pDRF1-GW AKAR3EV W303-1A MATa sch9::TRP1 + pDRF1-GW AKAR3EV (Roosen et al. 2004) W303-1A sch9Δ W303-1A MATa sch9::TRP1 In-house W303-1A ypk1Δ W303-1A ypk1::TRP1 (Niles and Powers 2014) W303-1A ypk1Δ + pDRF1-GW AKAR3-EV W303-1A ypk1::TRP1 + pDRF1-GW AKAR3-EV This work SP1 WT Matα his3, leu2, ura3, trp1, ade8, can1 (Ma et al. 1999) S18-1D (SP1 TPK1wimp) SP1 Matα TPK1wimp, tpk2::HIS3, tpk3::TRP1 (Nikawa et al. 1987) SP1 + pDRF1-GW AKAR3EV SP1 Matα + pDRF1-GW AKAR3EV This work SP1 + pDRF1-GW AKAR3 SP1 Matα + pYX212 AKAR3 This work, pYX212 AKAR3 was provided by Sonia Colombo (Colombo et al. 2017, Colombo et al. 2022) SP1 TPK1wimp + pDRF1-GW AKAR3EV S18-1D Matα + pDRF1-GW AKAR3EV This work SP1 + pDRF1-GW AKAR3EV-NR SP1 Matα + pDRF1-GW AKAR3EV-NR This work S25-31C MATa his3, leu2, ura3, trp1, ade8, tpk2::HIS3, tpk3::TRP1, bcy1::LEU2, Sch9::ADE8 (Toda et al. 1988, Crauwels et al. 1997) S25-31C + pDRF1-GW AKAR3EV MATa his3, leu2, ura3, trp1, ade8, tpk2::HIS3, tpk3::TRP1, bcy1::LEU2, Sch9::ADE8 + pDRF1-GW AKAR3EV This work Yeast growth Cells expressing URA3 plasmids (pDRF1-GW, pYES2, or pYX212) were grown overnight in 1x YNB medium (Sigma–Aldrich, St. Louis, MO, USA), containing 1% ethanol (Boom BV, Meppel, The Netherlands), 20 mg/L adenine hemisulfate (Sigma–Aldrich), 20 mg/L L-tryptophan (Sigma–Aldrich), 20 mg/L L-histidine (Sigma–Aldrich), and 60 mg/L L-leucine (SERVA Electrophoresis GmbH, Heidelberg, Germany). For WT strains, uracil (Sigma–Aldrich) was added to a final concentration of 20 mg/L. The cells were subsequently diluted and grown overnight to an OD600 of 0.1–1.5, with at least 5 divisions. For the experiments that involved S25-31C, cells were grown on 1x YNB medium containing 1% ethanol, 5 mM glucose (Boom BV, Meppel, The Netherlands), 20 mg/L adenine hemisulfate, 20 mg/L L-tryptophan, 20 mg/L L-histidine, and 60 mg/L L-leucine until glucose was exhausted. Next, cells were kept on this medium for 2 more days, after which they were visualized under the microscope. Concanavalin (ConA) coverslips ConA coverslips were made as described by Hansen et al. (2015). To prepare the coverslips, the ConA was diluted to 200 µg/mL and put on coverslips. The coverslips were dried overnight in a 6-well plate. Microscopy Cells were grown as described and transferred to the 6-well plates containing the ConA-coated coverslips. Next, the coverslip was placed in a Attofluor cell chamber (ThermoFisher Scientific, Waltham, MA, USA), and 1 mL of medium was added to the cell chamber. The coverslips were visualized using a Nikon Ti-eclipse microscope (Nikon, Minato, Tokio, Japan) at 30°C equipped with a TuCam system (Andor, Belfast, Northern Ireland, UK) and 2 Andor Zyla 5.5 sCMOS Cameras (Andor) and a SOLA 6-LCR-SB light source (Lumencor, Beaverton, OR, USA). Cells expressing the sensors, except pHluorin, were excited via a 438/24 nm excitation filter (Semrock, Lake Forest, IL, USA), and emission was passed through a 458 nm long-pass (LP) dichroic mirror. The acceptor and donor emissions were filtered by a 542/27 and 483/32 nm filter (Semrock). Direct acceptor fluorescence was recorded with a 500/24 nm excitation filter, a 520 LP dichroic filter, and a 542/27 nm emission filter. pHluorin was excited at 460–500 and 380–420 nm, and emission was recorded at 510–560 nm. For all perturbations, a baseline was recorded, after which YNB medium containing the compound of choice in a 10x concentration was added and diluted in the cell chamber to a 1x concentration. Rapamycin experiments Cells were grown as described and incubated with 200 nM rapamycin or a solvent (100% ethanol) for at least 2 h, after which the perturbations (addition of 10 mM glucose) were performed. Microscopy analysis Cells were segmented using an in-house script. In brief, this script stabilizes any drift using the image stabilizer plugin (Li 2008). Next, background correction was performed, and cells were segmented using the Weka Segmentation plugin (Arganda-Carreras et al. 2017b, Arganda-Carreras et al. 2017a), and the mean fluorescence for each cell was calculated for each frame. The resultant text files were analysed using R 4.1.3. For all cells, 40% bleedtrough correction was performed and the FRET ratio (i.e. bleedtrough-corrected YFP divided by the CFP signal) was calculated. Finally, baseline normalization was performed for time-lapse data. pHluorin ratios were calculated by dividing the fluorescence at 380–420 nm excitation over the fluorescence at 460–500 nm. For the dose-response fit, the final FRET levels (i.e. the mean FRET value of the last 3 frames) after the glucose additions were fitted (using the nls function in R) against the final glucose concentration according to equation 1, with [glucose] as the glucose concentration in mM, max as the maximal change in normalized FRET, and K0.5 as the glucose concentration giving 50% of the maximal response. (1) \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} }{}\begin{eqnarray*} {\rm{normalized\,\, FRET\,\, response}} - 1{\rm{\ }} = \frac{{\max {\rm{\ }} \cdot [{\rm{glucose}}]}}{{{{\rm{K}}}_{0.5} + \left[ {{\rm{glucose}}} \right]}}\ \end{eqnarray*}\end{document} Clustering was performed using the factoextra package in R. The optimal amount of clusters was determined by eye. For the long-term ethanol to mannose transition, microscopy images were segmented with a convolutional neural network with customized weights (Pachitariu and Stringer 2022). Frame-to-frame association of segmented objects was done through maximum matching based on inverse centroid distance. Fluorescence background was estimated from fluorescence images masked with the dilated segmentation images using the Background2D class from the photutils python package (Bradley et al. 2022). FRET ratios were calculated as the bleedthrough-corrected YFP signal, divided by the CFP signal. For growth type classification, cells that were in or that reached G2-phase (i.e. cells that had or did get a bud) during the experiment (between 2 and 6 h after the transition) were manually identified and classified as growing if there was a perceptible increase in bud volume, and as non-growing if there was no such perceptible increase. Analysis was restricted to cells that were present at the start of the experiment. Flow cytometry Cells were grown as described with YNB medium containing either 1% ethanol, 100 mM glucose, 100 mM galactose, or 100 mM mannose. Next, samples were measured using a CytoFLEX S Flow Cytometer (Beckman Coulter, Brea, CA, USA). Cells were excited using a 405 and a 488 nm laser, and emission fluorescence was passed through a 470/20 and 525/40 nm filter and recorded by avalanche photodiodes. Events with a saturating forward or side scatter were filtered, after which the median fluorescence signal of the cells expressing the empty pDRF1-GW plasmid was subtracted from all samples. Next, bleedtrough was calculated using the eCFP-expressing strain, and cells with at least a acceptor fluorescence signal of 2500 (arbitrary units) were kept. Lastly, FRET ratios were calculated for all remaining cells. Growth assays Yeast strains were grown on 1% ethanol medium as described. Cells were washed twice by centrifuging at 3500 g for 3 minutes and resuspending in YNB medium without a carbon source. Afterwards, cells were resuspended in YNB medium without carbon source to an OD600 of 1. Next, 20 μL of cells were put in a well of a 48-well microtiter plate having 480 μL YNB medium containing either 10 mM glucose, 10 mM galactose, or 0.1% ethanol. OD600 was measured every 5 minutes with a CLARIOstar plate reader (BMG LABTECH, Ortenberg, Germany) at 30°C and 700 rpm orbital shaking. Growth rates were calculated by calculating a moving average over each growth curve. Next, a sliding window was used in which a linear regression was fitted for each window, which gives the growth rate during this window. Next, to reduce the effect of outlier growth rates of the sliding window fits, the tenth fastest found slope was selected as the determined growth rate. Results AKAR3-EV shows a FRET response in budding yeast. To visualize the cellular kinase activities of Sch9p and PKA, we tested the use of the AKAR3-EV sensor, which consists of YPET-FHA1-EVlinker-RRATmotif-eCFP (Zhang et al. 2001, Allen and Zhang 2006, Komatsu et al. 2011). This sensor, developed for mammalian PKA assays, should also work in yeast as PKA and Sch9p also have RRxT as recognition site (Reinders et al. 1998, Ma et al. 1999, Plank et al. 2020). We constructed a non-responsive AKAR3-EV-NR sensor as a control by mutating the threonine in the RRxT motif to alanine (T506A), and expressed both AKAR3-EV and AKAR3-EV-NR in the W303-1A and SP1 strains. Both W303-1A and SP1 were used as some mutant W303-1A strains have difficulties to express sensors. We previously also observed this for the yEPAC cAMP sensor or even single fluorescent proteins, which indicates this is a more general problem for some W303-1A mutants and not a sensor-specific issue. In these two strains, we assessed the FRET response from a 1% ethanol to a 100 or a 10 mM glucose transition (Figs. 1A, B, and Supplementary Fig. S1). Furthermore, we tested the older generation AKAR3 sensor for performance comparison (Colombo et al. 2017, Colombo et al. 2022). The AKAR3-EV sensor gave a clear increase in FRET after glucose addition, which was 3–4 times higher compared to the AKAR3-EV-NR. Furthermore, we found only a marginal response of the original AKAR3 sensor, showing that the AKAR3-EV sensor has an improved ability to visualize the phosphorylation activity of PKA and Sch9p in yeast. Cellular expression was more than sufficient, and the sensor showed a uniform distribution in cells (Fig. 1C). Of notice, the AKAR3-EV-NR sensor shows a significant increase shortly after glucose addition in the W303-1A strain. This increase is not caused by osmotic changes since 100 mM sorbitol addition did not increase FRET levels (Supplementary Fig. S2A). Conclusions about the FRET responses of AKAR3-EV directly after a transition should therefore be taken carefully. Furthermore, the AKAR3-EV shows no basal drift in FRET, which we found for AKAR3. The original AKAR3 sensor also shows a small dip in FRET response after the glucose addition in W303-1A, which could come from the differential pH sensitivity of the fluorescent proteins in this sensor. AKAR3-EV uses YPET-eCFP as a FRET pair, which shows more pH robustness compared to the eCFP-Venus FRET pair used in AKAR3 (Supplementary Fig. S2B) (Botman et al. 2019). Lastly, expression levels of AKAR3-EV did not affect basal FRET levels, and growth was not affected on various carbon sources (Supplementary Figs. S2C and S2D). In conclusion, the AKAR3-EV can be used in yeast to assess cellular phosphorylation of the RRxT motif. Figure 1. AKAR3-EV responds to glucose addition. (A) Response of W303-1A WT cells expressing AKAR3, AKAR3-EV NR, or AKAR3-EV grown on 1% ethanol to 100 mM glucose addition. (B) Response of SP1 WT cells expressing either AKAR3, AKAR3-EV NR, or AKAR3-EV grown on 1% ethanol to 10 mM glucose addition. (C) Expression and distribution of W303-1A cells expressing AKAR3EV. Lines show the mean FRET value, normalized to the baseline, shades indicate SD, colours indicate the expressed sensor. All data are obtained from at least two biological replicates. AKAR3-EV responses are dependent on sch9 and PKA signalling, but not ypk1 To assess the whether the AKAR3-EV sensor is influenced by cAMP-PKA and TORC1-Sch9 signalling, we compared normalized FRET responses of W303-1A and SP1 WT to strains carrying the TPK1wimp mutation (which has low PKA activity) and the sch9∆ strain upon a shift from 1% ethanol to 10 mM glucose (Figs. 2A, B and Supplementary Figs. S3A, S3B). We also included non-normalized FRET responses in figure S3, as the normalized FRET responses do not show the starting phosphorylation status of a cell but only report relative changes. The TPK1wimp shows a lower initial response but does gradually obtained a similar change in normalized FRET at the end of the timelapse (normalized FRET levels of 1.16 and 1.14 for WT and TPK1wimp, respectively). In contrast, sch9 deletion gave a reduction in the reached plateau of 25% (normalized FRET levels of 1.24 and 1.18 for WT and sch9Δ, respectively). To note, for both the TPK1wimp and the sch9Δ, we found increased absolute FRET levels during the complete transition, indicative for a higher phosphorylation state. We also tested whether W303-1A CYR1K1876M, which lacks the classical cAMP peak upon glucose addition shows an altered AKAR3-EV response. This strain showed a normalized FRET response that reached a plateau similar to the sch9Δ strain (Figs. 2C and Supplementary Fig. S2C), but with different dynamics. Its absolute FRET levels were comparable to WT (Supplementary Fig. S3C). Finally, the RRxT motif can potentially also be phosphorylated by YPK1p and YPK2p (Chen et al. 1993) with YPK1p being the dominant protein. This kinase is regulated by TORC2, and this signalling branch is involved in lipid metabolism and can potentially also be activated by sugars (Jacquier and Schneiter 2010, Plank et al. 2020). We tested the FRET response of AKAR3-EV in a W303-1A ypk1Δ strain and found a negligible decrease in the maximal response (1.28 versus 1.26 normalized FRET levels for WT and ypk1Δ, respectively, Fig. 2C and Supplementary Fig. S3C). Since both the cAMP-PKA and TORC1-Sch9p cascades affect the AKAR3-EV response, we assessed whether a strain mutated in both cascades showed any increase in FRET levels upon glucose addition. For this, we used the S25-31C strain (tpk2Δ, tpk3Δ, bcy1Δ, and Sch9Δ), which is known to have impaired phosphorylation activity assessed by trehalase activity (Crauwels et al. 1997). As expected, 100 mM glucose addition to this strain showed no increase in FRET levels of AKAR3-EV, indicative that AKAR3-EV FRET (and hence RRxT phosphorylation) levels rely on PKA and Sch9p signalling (Fig. 2D and Supplementary Fig. S2D). In contrast, there is a slight decrease in FRET signal, which may be explained by glucose-induced phosphatase activity on the RRxT motif. The phosphorylation state of the sensor is in the end a steady-state balance between phosphorylation and dephosphorylation activity. Figure 2. AKAR3EV responds to both Sch9 and PKA activity. (A) FRET response of SP1 WT and SP1 TPK1wimp cells expressing AKAR3EV grown on 1% ethanol to 10 mM glucose addition. (B) FRET response W303-1A WT and W303-1A sch9Δ cells expressing AKAR3EV grown on 1% ethanol to 10 mM glucose addition. (C) FRET response W303-1A WT, W303-1A sch9Δ, W303-1A CYR1K1876M, or W303-1A ypk1Δ cells expressing AKAR3EV grown on 1% ethanol to 100 mM glucose addition. (D) FRET response of S25-31C cells expressing AKAR3EV grown on 1% ethanol to a 100 mM glucose pulse. (E) FRET response of SP1 WT and SP1 TPK1wimp cells expressing AKAR3EV or AKAR3-EV NR to 10 mM glucose addition. Cells were grown on 1% ethanol, incubated for at least 2 h with rapamycin, mock (ethanol), or without any addition and pulsed with 10 mM glucose. Lines show the mean FRET value, normalized to the baseline, shades indicate SD, colours indicate the expressed sensor, strain, and/or treatment. All data are obtained from at least two biological replicates. Finally, we confirmed the dependency of RRxT phosphorylation on Sch9p and PKA by the simultaneous suppression of both pathways. For this, we used the TORC1 inhibitor rapamycin to inhibit this signalling cascade and combine this with the SP1 WT and TPK1wimp strains, which have decreased (but not completely absent) PKA activity. In wild-type SP1, we found a decrease of 27% in the FRET response of rapamycin-treated cells compared to untreated cells after a 10 mM glucose pulse (Fig. 2E and Supplementary Fig. S2E). SP1 TPK1wimp cells already showed a 27% decreased maximal response (compared to WT cells), which further decreased to 60% when treated with rapamycin. This response was still higher compared to the response of the non-responsive sensor, which most likely is attributed to the remaining activity of TPK1wimp. In conclusion, we show that the AKAR3-EV sensor can be used to measure TORC1-Sch9p and cAMP-PKA activity by visualizing the cellular RRxT phosphorylation status. Deletion of SCH9 or removal of the classical cAMP peak results in a decreased phosphorylation status of the cell directly after the carbon source transition. In contrast, decreased PKA activity gives a decreased initial response, but eventually reaches the same plateau as wild-type cells within the short timeframe measured. Thus, the individual influence of Sch9p and PKA on the phosphorylation dynamics can be further elucidated by the AKAR3-EV sensor. Sugar transitions shows distinct phosphorylation dynamics with single-cell heterogeneity In our previous study, we found that cAMP signalling in yeast is different for different sugars, and it was homogeneous with no clear subpopulations or non-responders (Botman et al. 2021). We used the AKAR3-EV sensor to measure the downstream response to different sugars. W303-1A cells grown on 1% ethanol medium were pulsed with either 100 mM glucose or sucrose (both Gpr1p agonists), 100 mM fructose (no Gpr1p agonist or antagonist), or 100 mM mannose (an antagonist of Gpr1p) (Lemaire et al. 2004, Botman et al. 2021). Glucose, sucrose, and fructose gave a clear transient increase in the FRET ratio after addition, where sucrose and glucose gave a slightly faster and more sustained response. In contrast, mannose showed a decline in FRET signal after which the FRET response increased to higher levels, which was sustained until the end of the timelapse recording (Fig. 3A, Supplementary Fig. S4 for SP1 responses). For fructose, glucose, and sucrose, we found no clear and significant subpopulations. The most striking response at the single-cell level, however, was found for mannose addition. Mannose addition gave a highly heterogenous response (Figs 3B–E, Supplementary movie S1), which was not found for the non-responsive sensor (Supplementary Fig. S5) and was not caused by sensor expression levels, as the sum of the donor and acceptor emission (used as a proxy for sensor expression) was similar among the clusters (Fig. 3D, second panel). After the dip in FRET levels (which all cells do seem to have), we identified three discrete responses, which we clustered using k-means clustering. The first cluster consisted of cells that have a broad timeframe in which cells increase in FRET levels. Furthermore, final FRET levels are slightly increased (mean normalized FRET value of 1.08), and the slope of the FRET increase was high (mean normalized FRET increase of 0.26 per minute). The second cluster showed clear switchers, which obtained a final normalized FRET value of 1.25. Finally, the third cluster consisted of cells that did not increase but decreased their RRxT phosphorylation levels (a mean normalized FRET value of 0.90 at the end of the timeframe). Interestingly, for the mannose transitions, cells that had a lower baseline FRET value also seem to end in a lower FRET state, and vice versa (Fig. 3D). Figure 3. Single-cell assessment of AKAR3EV responses shows heterogeneity for mannose, but not for fructose, sucrose, and glucose additions. (A) Response of W303-1A WT cells expressing AKAR3EV grown on 1% ethanol to the various sugar additions, added at t = 0 minutes. Lines show mean FRET value, normalized to the baseline, shades indicate SD. Facet titles show the final concentration of the added sugar. (B) Baseline-normalized single-cell traces of the transitions depicted in A. (C) The three identified clusters found for the heterogenic response after 100 mM mannose addition. Lines depict single-cell traces of the normalized FRET value. (D) Violin plots of basal FRET level (not baseline normalized), the sum of the donor and acceptor emission fluorescence as a proxy for sensor expression, the change in normalized FRET values after 50 minutes (Δnormalized FRET and the maximal slope obtained (normalized FRET change per minute). Each dot depicts a single cell, horizontal bar presents the mean. All data are obtained from at least two biological replicates. (E) Static pictures of a 1% ethanol to 100 mM mannose or 100 mM glucose transition. Colours indicate FRET ratio (normalized to the first baseline frame), mannose was added at 0 minutes. Scale bar indicates 10 µm. We further explored the heterogeneous FRET response after mannose addition by evaluating the cellular growth of these yeast for at least 2 h after mannose addition. For accurate growth assessment, analysis was restricted to cells that had a bud at the start of the experiment or developed a bud during the experiment. (Fig. 4). Single-cell FRET curves were also determined (Fig. 4A), and k-means clustering was performed (Fig. 4B). We found again the 3 clusters, and when we overlayed growth versus non-growth in these clusters (Fig. 4B and C), we found one cluster with mostly growers, one cluster with non-growers, and a mixed cluster. Growing cells had a significant lower basal FRET level (0.84 ± 0.16 versus 0.94 ± 0.16 for growing and non-growing, respectively. Student’s t-test, P = 0.006, Fig. 4D), obtained a larger normalized FRET change (0.20 ± 0.09 versus 0.07 ± 0.20 for growing and non-growing, respectively. Student’s t-test, P < 0.001, Fig. 4D), and showed a faster normalized FRET increase per minute (0.12 ± 0.04 versus 0.09 ± 0.05, P = 0.001). Figure 4. AKAR3-EV shows differential FRET responses based on the growth state upon mannose addition. (A) Response of W303-1A WT cells expressing AKAR3EV grown on 1% ethanol to 100 mM mannose, added at t = 0 minutes. Lines show the normalized FRET value per cell. Facet titles show whether the cell was a confirmed grower or a non- grower. (B) The three identified clusters found after 100 mM mannose addition. Lines depict single-cell traces of the normalized FRET value, colours depict whether cells grow or do not grow (C) Histogram depicting which fractions of cells grow (blue) or not grow (red) for each cluster. (D) Violin plot of the basal FRET level (not baseline normalized), the change in normalized FRET values at the end of the recorded timelapse (Δnormalized FRET), and the maximal slope obtained (normalized FRET change per minute). Each dot depicts a single cell. White bar represents the mean. P values from student’s t-test are shown. All data obtained from three biological replicates. In conclusion, the AKAR3-EV sensor shows multifarious responses after sugar addition, dependent on the specific sugar. For mannose, we found a highly heterogenous response, and its clustering suggests that the rapid dynamics of the RRxT phosphorylation status during a carbon-source transition correlates with the onset of rapid growth. The nutrient-induced phosphorylation system has a high affinity for glucose When we tested the response of the AKAR3-EV sensor in W303-1A to different levels of glucose (Fig. 5A), we found a graded response that could be fitted by a single binding curve with a relatively high affinity, with a K0.5 of 0.27 mM for glucose (Fig. 5B). A dose-response curve of strain SP1 gave similar parameters (K0.5 of 0.26 mM), indicating that this high affinity is not a strain-specific feature (Supplementary Fig. S6). In addition, no clear subpopulations or non-responders were detected (Supplementary Fig. S7). This is in line with previous results found for cAMP (Botman et al. 2021). Figure 5. Dose-response of AKAR3-EV to glucose addition shows a high affinity. (A) dose-response curves of 1% ethanol-grown W303-1A cells expressing AKAR3-EV to various concentrations of glucose additions, added at t = 0 minutes. Lines show the mean baseline-normalized FRET response. Colours indicate the final glucose concentration after glucose addition. (B) Final FRET response of AKAR3-EV-expressing cells to the glucose concentration added, dots show mean response, error bars indicate SD, and grey line indicates fit (from equation 1). Fitting of the final AKAR3-EV FRET levels to the glucose concentration pulsed shows saturation kinetics with a K0.5 of 0.24 and a maximal normalized FRET value of 1.24. Inset shows the zoom-in of the AKAR3-EV responses from 0 to 1 mM of glucose additions. All data obtained from at least two biological replicate. In conclusion, we found that the affinity of the RRxT phosphorylation system (i.e. TORC1-Sch9p and cAMP-PKA) has a high affinity for glucose—in comparison, the affinity we found for cAMP peak height was 3.0 mM, ten times higher, therefore. Furthermore, this system appears to be homogeneous since no clear non-responders or subpopulations were found. AKAR3-EV shows no growth-rate dependent FRET status After characterizing the TORC1-Sch9p and cAMP-PKA signalling response during nutrient transitions, we studied the steady state phosphorylation levels during (balanced) growth on different carbon sources. W303-1A cells expressing either the AKAR3-EV or the AKAR3-EV-NR sensor were grown to mid-exponential phase on 1% ethanol (growth rate of 0.16 h−1), galactose (growth rate of 0.25 h−1), glucose (growth rate of 0.36 h−1), or mannose (growth rate of 0.34 h−1), and the FRET level distributions across a population were measured using a flow cytometer (Fig. 6A). We found significant differences between the conditions tested (Kruskal–Wallis, P < 0.01), except between galactose and mannose (Wilcoxon signed-rank test, P = 0.6). However, we also found such differences in the non-responsive sensor, and no clear relation was found between growth rate and the FRET levels of AKAR3-EV after correction for aspecificity (Figs. 6B and C). Figure 6. Growth rate does not relate with AKAR3EV basal levels. (A) Violin plot of basal AKAR3EV and AKAR3EV NR FRET levels measured using flow cytometry in W303-1A WT on various carbon sources, normalized to the 100 mM glucose condition. White dot depicts median FRET value, errorbars indicate SD. (B) Violin plot of AKAR3-EV FRET levels after correcting for aspecific signal on the four carbon sources. White dot depicts the median FRET value, error bars indicate SD. (C) AKAR3-EV FRET levels (corrected by AKAR3-EV NR signal) plotted against the growth rate of W303-1A WT on the carbon sources. Each dot depicts the population average FRET level on a specific carbon source (indicates by the dot colour), error bars indicate SD. Representative dataset shown from two biological replicates. Discussion In the present study, we implemented the mammalian-optimized AKAR3-EV sensor in yeast and tested whether this sensor can also be applied to study Sch9p and PKA signalling in yeast. In yeast, the sensor has sufficient expression and a homogenous distribution in cells. Furthermore, growth rates were not affected by the sensor, indicating the sensor is harmless to cells (Supplementary Fig. S1C). The AKAR3-EV sensor showed the expected response to an ethanol-to-glucose transition and outperformed a previously published AKAR sensor in yeast (Fig. 1A). The baseline drift of the original AKAR sensor can be caused by the photochromic behaviour of one of the FPs (Botman et al. 2019), or differential FP responses to changes in the intracellular composition. We also made a non-responsive version of AKAR3-EV by mutating the RRxT motif to RRXA (Plank et al. 2022). This AKAR3-EV NR sensor indeed showed significantly lower responses, although we found that some transitions evoked a rather large response immediately after the transition. The origin of this response is not known and may be caused by nonspecific phosphorylation of the sensor domain, although this domain does not contain any other known phosphorylation sites of PKA or Sch9p. This response can also originate from changes of the intracellular composition (e.g. changes in redox potential, pH, and ion concentration). If needed, the AKAR3-EV NR can be used to correct for these aspecific responses. This sensor can be potentially improved for use in yeast by using phosphorylation sites that are more common for yeast (i.e. RRXS) (Plank et al. 2022), although this could also increase the basal phosphorylation status of the sensor, decreasing its dynamic FRET range. In addition, (y)mTurquoise2 could be used instead of eCFP since this fluorescent protein is a better FRET donor (Goedhart et al. 2012, Mastop et al. 2017). The AKAR3-EV FRET response is indeed dependent on both TORC1-Sch9p and cAMP-PKA signalling (Fig. 2). Glucose addition to ethanol-grown cells showed that impaired PKA signalling (using the TPK1wimp strain) decreased the rate of phosphorylation, although the maximal response remains similar. An Sch9 deletion, on the other hand, resulted in a lower normalized response after a glucose transition. Yet, these transitions were performed in different strains (SP1 and W303-1A, respectively). Interestingly, we also found that the W303-1A CYR1K1876M mutation in which the transient cAMP peak is absent (Botman et al. 2021) showed a decreased maximal response. Further research should clarify whether and why missing the transient cAMP peak has a long-term effect on cell signalling status and fitness. The S25-31C strain with impaired signalling in both Sch9 as PKA, showed a decrease in RRxT phosphorylation, proving that the AKAR3-EV sensor indeed measures solely PKA and Sch9p activity. This was further confirmed by adding rapamycin in SP1 WT and SP1 TPK1wimp cells, where the rapamycin-treated TPK1wimp strain showed a severe reduction in FRET response after glucose addition. Lastly, we found no significant effect of ypk1 deletion on the FRET response of AKAR3-EV in W303-1A, indicating that the AKAR3-EV sensor measures specifically PKA and Sch9p activity. As mentioned, some mutant strains had higher absolute FRET levels, suggesting that the signalling architecture can compensate for a decreased kinase activity. This can occur, for example, via negative feedback patterns known for PKA (Nikawa et al. 1987, Nikawa et al. 1987, Dong and Bai 2011) and TORC1 (Péli-Gulli et al. 2017) or altered activity of RRxT phosphatases. Another possibility is that an altered cellular composition confounds FRET levels (Moussa et al. 2014), for which the non-responsive sensor can function as a control. Thus, as for every FRET sensor, quantitative conclusions about normalized and absolute FRET levels should be taken with care. In this study, we did not wish to make quantitive conclusions about signalling mutants on the absolute phosphorylation status, but rather use these mutants to show that the AKAR3-EV sensor indeed measures the activity of these two kinases. A major strength of biosensors is the ability to measure single-cell responses. Therefore, we assessed single-cell responses during transitions from ethanol-grown cells to glucose, sucrose, fructose, and mannose (Fig. 3). As previously found for cAMP signalling (Botman et al. 2021), hardly any heterogeneity or subpopulations were found for the glucose, sucrose, and fructose transitions. In contrast, we did find a heterogenic response upon mannose addition. Mannose is known as an antagonist of cAMP signalling (Lemaire et al. 2004, Botman et al. 2021), but is metabolized at rates similar to glucose as cells grow comparable on these sugars (growth rate of 0.34 h−1 for mannose and 0.36 h−1 for glucose). Glycolytic startup, determined by pH measurements, also indicates that mannose is transported and metabolized at least as fast as glucose (Supplementary Fig. S8). The conflicting signals between the signalling and metabolism of mannose may be the reason for the heterogenic response. Cells obtaining a higher RRxT phosphorylation state, compared to the pre-transition state, seem to start growth whereas cells with a (s)lower response seem to halt growth. This confirms that PKA and Sch9p activity, and the RRxT phosphatases have an effect on the cellular decision to start growth, in line with previous studies showing that cAMP-PKA signalling and Sch9p are involved in cell cycle progression (Hubler et al. 1993, Müller et al. 2003, Jorgensen et al. 2004, Futcher 2006, Cocklin and Goebl 2011). The heterogenous Sch9p and PKA signalling dynamics potentially transmit further downstream, where it is shown that different temporal nuclear localization patterns of transcription factors (such as msn2p, a phosphorylation target of Sch9p and PKA) can result in differential transcription and translation responses potentially resulting in an altered growth response (Hao and O'Shea 2012, Zid and O'Shea 2014, Hansen and O'Shea 2016). The AKAR3-EV sensor revealed a relatively high affinity (K0.5 = 0.24 mM) of the phosphorylation system for glucose (Fig. 4 ). This is far below the affinity of the high-affinity hexose transporters in yeast with a lowest Km of ∼1–2 mM, for HXT6 and 7 (Reifenberger et al. 1997, Maier et al. 2002), which are expressed in these ethanol-grown cells. Furthermore, the affinity is lower compared to our previously obtained glucose affinity of cAMP peak levels in yeast (Botman et al. 2021), which confirms that the RRxT phosphorylation status of the cell is not solely determined by cAMP-PKA signalling. The high affinity of the RRxT phosphorylation system is, on the other hand, not too far from the Monod constant for glucose-limited growth, which is around 0.5 mM, admittedly for another strain (CEN.PK) (Canelas et al. 2011). Moreover, these values are in line with previous observations for Mig1 translocation after glucose additions (Bendrioua et al. 2014). The fact that we did not find clear subpopulations at any glucose concentrations shows that, at least for glucose, yeast cells sense its concentration in a highly accurate and robust manner. One large unanswered question is whether the basal signalling status of the Sch9p and PKA signalling pathways depends on the specific growth rate. This may be expected as transcription of ribosomal genes is an important target of the pathway, and ribosomal content does scale with growth rate (Metzl-Raz et al. 2017). The obtained flow cytometry data are statistically significantly different between most conditions, but given the small size of the difference relative to the spread of the distributions (Figs. 6B and C), their biological relevance is disputable. The small differences could be caused by a maximal phosphorylation status of the cell under the conditions tested, especially since the RRxT is a preferred motif of Sch9p and PKA (Plank et al. 2020). Yet, we showed that the FRET levels can increase for >20% in cells grown in 1% ethanol medium, and the basal FRET level of this condition is comparable with the other conditions tested. This implies that the sensor is not maximally phosphorylated. Therefore, we believe that the basal phosphorylation status of the RRxT motif is unchanged and not maximally phosphorylated between various growth rates. Unchanged RRxT phosphorylation levels do not necessarily imply that PKA and Sch9p activities are also unchanged across growth conditions, as the phosphorylation level is a result of both phosphorylation and dephosphorylation. Furthermore, PKA and Sch9p phosphorylate also other motifs, potentially with different rates across conditions (Plank et al. 2022). We hypothesize that the RRxT phosphorylation response upon sugar transitions steers cells to the right cellular physiology, after which this signalling system returns to its basal level again. The dynamics of Fig. 3A also suggest a temporal impact of sugars on the phosphorylation state. In summary, the AKAR3-EV proved to be a robust sensor to measure the nutrient-induced RRxT phosphorylation status in yeast cells. However, since the phosphorylation status is an integrated output of the activity of both Sch9p and PKA, but also on phosphatases, its interpretation is more challenging than for other sensors. Nonetheless, we believe that the AKAR3-EV sensor is a useful addition to the toolbox that can help to elucidate how yeast cells respond and adapt to nutrient changes. Data and resources AKAR3-EV and AKAR3-EV NR in pDRF1-GW can be acquired via Addgene (https://www.addgene.org/182533/and https://www.addgene.org/182534/). Data can be found via at DOI:10.17632/w655db7rj9.1 Supplementary Material foad029_Supplemental_Files Click here for additional data file. Acknowledgements We are grateful to Dr. Kazuhiro Aoki (Kyoto University) for sharing AKAR3-EV with us. We thank Dr. Sonia Colombo (Università Milano Bicocca) for providing AKAR3 in pYX212. We also thank Dr. Johan Thevelein (KU Leuven) for sharing various yeast strains. Lastly, we acknowledge Dr. Ted Powers for sharing the W303-1A ypk1Δ strain. Competing interests Authors declare to have no competing interests. Funding This work was supported by Amsterdam Institute for Life and Environment (A-LIFE). ==== Refs References Allen MD , ZhangJ. Subcellular dynamics of protein kinase A activity visualized by FRET-based reporters. Biochem Biophys Res Commun. 2006;348 :716–21.16895723 Arganda-Carreras I , KaynigV, RuedenCet al. Trainable Weka Segmentation: a machine learning tool for microscopy pixel classification. Bioinformatics. 2017a;33 :2424–6.28369169 Arganda-Carreras I , KaynigV, SchindelinJet al. Trainable Weka segmentation: a machine learning tool for microscopy image segmentation. Bioinformatics. 2017b;33 (15 ):2424–6.28369169 Bendrioua L , SmedhM, AlmquistJet al. Yeast AMP-activated protein kinase monitors glucose concentration changes and absolute glucose levels. J Biol Chem. 2014;289 :12863–75.24627493 Beullens M , MbonyiK, GeertsLet al. Studies on the mechanism of the glucose-induced cAMP signal in glycolysis and glucose repression mutants of the yeast Saccharomyces cerevisiae. Eur J Biochem. 1988;172 :227–31.2831059 Botman D , de GrootDH, SchmidtPet al. In vivo characterisation of fluorescent proteins in budding yeast. Sci Rep. 2019;9 :2234.30783202 Botman D , O'TooleTG, GoedhartJet al. A yeast FRET biosensor enlightens cAMP signaling. Mol Biol Cell. 2021;32 :1229–40.33881352 Bradley L ,SipoczB,RobitailleTet al. astropy/photutils: 1.6.0. 2022. https://zenodo.org/record/7419741 Accessed date March 6th 2023. Broach JR . Nutritional control of growth and development in yeast. Genetics. 2012;192 :73–105.22964838 Broek D , TodaT, MichaeliTet al. The S. cerevisiae CDC25 gene product regulates the RAS/adenylate cyclase pathway. Cell. 1987;48 :789–99.3545497 Canelas AB , RasC, ten PierickAet al. An in vivo data-driven framework for classification and quantification of enzyme kinetics and determination of apparent thermodynamic data. Metab Eng. 2011;13 :294–306.21354323 Casperson GF , WalkerN, BourneHR. Isolation of the gene encoding adenylate cyclase in Saccharomyces cerevisiae. Proc Natl Acad Sci U S A. 1985;82 :5060–3.2991907 Chen P , LeeKS, LevinD E. A pair of putative protein kinase genes (YPK1 and YPK2) is required for cell growth in Saccharomyces cerevisiae. Mol Gen Genet. 1993;236–236 :443–7. Cocklin R , GoeblM. Nutrient sensing kinases PKA and Sch9 phosphorylate the catalytic domain of the ubiquitin-conjugating enzyme Cdc34. PLoS One. 2011;6 :e27099.22087249 Colombo S , BroggiS, ColliniMet al. Detection of cAMP and of PKA activity in Saccharomyces cerevisiae single cells using Fluorescence Resonance Energy Transfer (FRET) probes. Biochem Biophys Res Commun. 2017;487 :594–9.28433631 Colombo S , LongoniE, GnugnoliMet al. Fast detection of PKA activity in Saccharomyces cerevisiae cell population using AKAR fluorescence resonance energy transfer probes. Cell Signal. 2022;92 . Colombo S , MaP, CauwenbergLet al. Involvement of distinct G-proteins, Gpa2 and Ras, in glucose- and intracellular acidification-induced cAMP signalling in the yeast Saccharomyces cerevisiae. EMBO J. 1998;17 :3326–41.9628870 Conrad M , SchothorstJ, KankipatiHNet al. Nutrient sensing and signaling in the yeast Saccharomyces cerevisiae. FEMS Microbiol Rev. 2014;38 :254–99.24483210 Conrad M , KankipatiHN, KimpeMet al. The nutrient transceptor/PKA pathway functions independently of TOR and responds to leucine and Gcn2 in a TOR-independent manner. FEMS Yeast Res. 2017;17 :1–12. Crauwels M , DonatonMCV, PernambucoMBet al. The Sch9 protein kinase in the yeast Saccharomyces cerevisiae controls cAPK activity and is required for nitrogen activation of the fermentable-growth-medium-induced (FGM) pathway. Microbiology (Reading). 1997;143 :2627–37.9274016 Dong J , BaiX. The membrane localization of Ras2p and the association between Cdc25p and Ras2-GTP are regulated by protein kinase A (PKA) in the yeast Saccharomyces cerevisiae. FEBS Lett. 2011;585 :1127–34.21457714 Engelberg D , SimchenG, LevitzkiA. In vitro reconstitution of cdc25 regulated S. cerevisiae adenylyl cyclase and its kinetic properties. EMBO J. 1990;9 :641–51.2155776 Eraso P , GancedoJM. Use of glucose analogues to study the mechanism of glucose-mediated cAMP increase in yeast. FEBS Lett. 1985;191 :51–54. Futcher B . Metabolic cycle, cell cycle, and the finishing kick to start. Genome Biol. 2006;7 :107.16677426 Goedhart J , von StettenD, Noirclerc-SavoyeMet al. Structure-guided evolution of cyan fluorescent proteins towards a quantum yield of 93%. Nat Commun. 2012;3 :751.22434194 Hansen AS , HaoN, O'SheaEK. High-throughput microfluidics to control and measure signaling dynamics in single yeast cells. Nat Protoc. 2015;10 :1181–97.26158443 Hansen AS , O'SheaEK. Encoding four gene expression programs in the activation dynamics of a single transcription factor. Curr Biol. 2016;26 :R269–71.27046808 Hao N , O'SheaEK. Signal-dependent dynamics of transcription factor translocation controls gene expression. Nat Struct Mol Biol. 2012;19 :31–40. Hubler L , Bradshaw-RouseJ, HeidemanW. Connections between the Ras-cyclic AMP pathway and G1 cyclin expression in the budding yeast Saccharomyces cerevisiae. Mol Cell Biol. 1993;13 :6274–82.8413227 Jacquier N , SchneiterR. Ypk1, the yeast orthologue of the human serum- and glucocorticoid-induced kinase, is required for efficient uptake of fatty acids. J Cell Sci. 2010;123 :2218–27.20516150 Jorgensen P , RupesI, SharomJRet al. A dynamic transcriptional network communicates growth potential to ribosome synthesis and critical cell size. Genes Dev. 2004;18 :2491–505.15466158 Jouhten P , PonomarovaO, GonzalezRet al. Saccharomyces cerevisiae metabolism in ecological context. FEMS Yeast Res. 2016;16 :fow080.27634775 Kataoka T , BroekD, WiglerM. DNA sequence and characterization of the S. cerevisiae gene encoding adenylate cyclase. Cell. 1985;43 :493–505.2934138 Kim J , RoyA, IiDJ. Biochimica et biophysica acta the glucose signaling network in yeast. BBA - General Subjects. 2013;1830 :5204–10.23911748 Komatsu N , AokiK, YamadaMet al. Development of an optimized backbone of FRET biosensors for kinases and gtpases. Mol Biol Cell. 2011;22 :4647–56.21976697 Kraakman L , LemaireK, MaPet al. A Saccharomyces cerevisiae G-protein coupled receptor, Gpr1, is specifically required for glucose activation of the cAMP pathway during the transition to growth on glucose. Mol Microbiol. 1999;32 :1002–12.10361302 Kunkel J , LuoX, CapaldiAP. Integrated TORC1 and PKA signaling control the temporal activation of glucose-induced gene expression in yeast. Nat Commun. 2019;10 :3558.31395866 Kunkel J , LuoX, CapaldiAP. Integrated TORC1 and PKA signaling control the temporal activation of glucose-induced gene expression in yeast. Nat Commun. 2019;10 ,. Lara AR , GalindoE, RamírezOTet al. Living with heterogeneities in bioreactors: understanding the effects of environmental gradients on cells. Mol Biotechnol. 2006;34 :355–82.17284782 Lemaire K , van de VeldeS, van DijckPet al. Glucose and sucrose act as agonist and mannose as antagonist ligands of the G protein-coupled receptor Gpr1 in the yeast Saccharomyces cerevisiae. Mol Cell. 2004;16 :293–9.15494315 Li K . The image stabilizer plugin for ImageJ. ImageJ. Net. 2008. Liti GK . The fascinating and secret wild life of the budding yeast S. cerevisiae. Elife. 2015;4 :1–9. Ma P , WeraS, van DijckPet al. The PDE1-encoded low-affinity phosphodiesterase in the yeast Saccharomyces cerevisiae has a specific function in controlling agonist-induced cAMP signaling. Mol Biol Cell. 1999;10 :91–104.9880329 Maier A , VölkerB, BolesEet al. Characterisation of glucose transport in Saccharomyces cerevisiae with plasma membrane vesicles (countertransport) and intact cells (initial uptake) with single Hxt1, Hxt2, Hxt3, Hxt4, Hxt6, Hxt7 or Gal2 transporters. FEMS Yeast Res. 2002;2 :539–50.12702270 Mastop M , BindelsDS, ShanerNCet al. Characterization of a spectrally diverse set of fluorescent proteins as FRET acceptors for mTurquoise2. Sci Rep. 2017;7 :11999.28931898 Mbonyi K , BeullensM, DetremerieKet al. Requirement of one functional RAS gene and inability of an oncogenic ras variant to mediate the glucose-induced cyclic AMP signal in the yeast Saccharomyces cerevisiae. Mol Cell Biol. 1988;8 :3051–7.2850478 Metzl-Raz E , KafriM, YaakovGet al. Principles of cellular resource allocation revealed by condition-dependent proteome profiling. Elife. 2017;6 :e28034.doi: 10.7554/eLife.28034.001.28857745 Moussa R , Papa MzeN, ArrehHYet al. An evaluation of genetically encoded FRET-based biosensors for quantitative metabolite analyses in vivo. J Biotechnol. 2014;191 :250–9.25107505 Müller D , ExlerS, Aguilera-VázquezLet al. Cyclic AMP mediates the cell cycle dynamics of energy metabolism in Saccharomyces cerevisiae. Yeast. 2003;20 :351–67.12627401 Munder T , KüntzelHK Glucose-induced CAMP signaling in Saccharomyces cerevisiae is mediated by the CDC25 protein. FEBS Lett. 1989;242 :341–5.2536619 Nikawa J , CameronS, TodaTet al. Rigorous feedback control of cAMP levels in Saccharomyces cerevisiae. Genes Dev. 1 :931–7. 1987.2828175 Nikawa J , SassP, WiglerMK Cloning and characterization of the low-affinity cyclic AMP phosphodiesterase gene of Saccharomyces cerevisiae. Mol Cell Biol. 1987;7 :3629–36.2824992 Niles BJ , PowersTK TOR complex 2–Ypk1 signaling regulates actin polarization via reactive oxygen species. Mol Biol Cell. 2014;25 :3962–72.25253719 Orij R , PostmusJ, ter BeekAet al. In vivo measurement of cytosolic and mitochondrial pH using a pH-sensitive GFP derivative in Saccharomyces cerevisiae reveals a relation between intracellular pH and growth. Microbiology (Reading). 2009;155 :268–78.19118367 Orij R , UrbanusML, VizeacoumarFJet al. Genome-wide analysis of intracellular pH reveals quantitative control of cell division rate by pH(c) in Saccharomyces cerevisiae. Genome Biol. 2012;13 :R80.23021432 Pachitariu M , StringerC. Cellpose 2.0: how to train your own model. Nat Methods. 2022;19 :1634–41.36344832 Pardo LA , LazoPS, RamosS. Activation of adenylate cyclase in cdc25 mutants of Saccharomyces cerevisiae. FEBS Lett. 1993;319 :237–43.8458416 Péli-Gulli MP , RaucciS, HuZet al. Feedback inhibition of the rag gtpase GAP complex Lst4-Lst7 safeguards TORC1 from hyperactivation by amino acid signals. Cell Rep. 2017;20 :281–8.28700931 Plank M . Interaction of TOR and PKA signaling in S. cerevisiae. Biomolecules. 2022;12 :210.35204711 Plank M , CarmiolN, MitriBet al. Systems level analysis of time and stimuli specific signaling through PKA. Biorxiv. 2022, 2022.03.10.483795. Plank M , PerepelkinaM, MüllerMet al. Chemical genetics of AGC-kinases reveals shared targets of Ypk1, protein kinase A and Sch9. Mol Cell Proteomics. 2020;19 :655–71.32102971 Reifenberger E , BolesE, CiriacyM. Kinetic characterization of individual hexose transporters of Saccharomyces cerevisiae and their relation to the triggering mechanisms of glucose repression. Eur J Biochem. 1997;245 :324–33.9151960 Reinders A , BürckertN, BollerTet al. Saccharomyces cerevisiae cAMP-dependent protein kinase controls entry into stationary phase through the Rim15p protein kinase. Genes Dev. 1998;12 :2943–55.9744870 Rødkaer Sv , FaergemanNJ. Glucose- and nitrogen sensing and regulatory mechanisms in Saccharomyces cerevisiae. FEMS Yeast Res. 2014;14 :683–96.24738657 Rolland F , De WindeJH, LemaireKet al. Glucose-induced cAMP signalling in yeast requires both a G-protein coupled receptor system for extracellular glucose detection and a separable hexose kinase-dependent sensing process. Mol Microbiol. 2000;38 :348–58.11069660 Rolland F , WankeV, CauwenbergLet al. The role of hexose transport and phosphorylation in cAMP signalling in the yeast Saccharomyces cerevisiae. FEMS Yeast Res. 2001;1 :33–45.12702461 Rolland F , WinderickxJ, TheveleinJM. Glucose-sensing and -signalling mechanisms in yeast. FEMS Yeast Res. 2002;2 :183–201.12702307 Roosen J , EngelenK, MarchalKet al. PKA and Sch9 control a molecular switch important for the proper adaptation to nutrient availability. Mol Microbiol. 2004;55 :862–80. Santangelo GM . Glucose signaling in Saccharomyces cerevisiae Glucose signaling in Saccharomyces cerevisiae. Microbiol Mol Biol Rev. 2006;70 :253–82.16524925 Slattery MG , LikoD, HeidemanW. Protein kinase A, TOR, and glucose transport control the response to nutrient repletion in Saccharomyces cerevisiae. Eukaryot Cell. 2008;7 :358–67.18156291 Thevelein JM . Signal transduction in yeast. Yeast. 1994;10 :1753–90.7747517 Toda T , CameronS, SassPet al. SCH9, a gene of Saccharomyces cerevisiae that encodes a protein distinct from, but functionally and structurally related to, cAMP-dependent protein kinase catalytic subunits. Genes Dev. 1988;2 :517–27.3290050 van Aelst L , Boy-marcotteE, CamonisJHet al. The C-terminal part of the CDC25 gene product plays a key role in signal transduction in the glucose-induced modulation of CAMP level in Saccharomyces cerevisiae. Eur J Biochem. 1990;6680 :675–80. van Aelst L , JansWH, TheveleinM. Involvement of the CDC25 gene product in the signal transmission pathway of the glucose-induced RAS-mediated cAMP signal in the yeast Saccharomyces cerevisiae. J Gen Microbiol. 1991;137 :341–9.1849965 Vanhalewyn M , DumortierF, DebastGet al. A mutation in Saccharomyces cerevisiae adenylate cyclase, Cyr1(K1876M), specifically affects glucose- and acidification-induced cAMP signalling and not the basal cAMP level. Mol Microbiol. 1999;33 :363–76.10411752 van Zeebroeck G , DemuyserL, ZhangZet al. Nutrient sensing and cAMP signaling in yeast: g-protein coupled receptor versus transceptor activation of PKA. Microbial Cell. 2021;8 :17–27. Wehrs M , TanjoreD, EngTet al. Engineering robust production microbes for large-scale cultivation. Trends Microbiol. 2019;27 :524–37.30819548 Winderickx J , HolsbeeksI, LagatieOet al. From feast to famine; adaptation to nutrient availability in yeast. In: Topics in Current Genetics. Vol. 1 : Berlin Heidelberg: Springer, 2003, 305–86. Xiao Y , BowenCH, LiuDet al. Exploiting nongenetic cell-to-cell variation for enhanced biosynthesis. Nat Chem Biol. 2016;12 :339–44.26999780 Yun C , TamakiH, NakayamaRet al. Gpr1p, a putative G-protein coupled receptor, regulates glucose-dependent cellular cAMP level in yeast Saccharomyces cerevisiae. Biochem Biophys Res Commun. 1998;252 :29–33.9813141 Zaman S , LippmanSI, SchneperLet al. Glucose regulates transcription in yeast through a network of signaling pathways. Mol Syst Biol. 2009;5 :1–14. Zhang J , MaY, TaylorSSet al. Genetically encoded reporters of protein kinase A activity reveal impact of substrate tethering. Proc Natl Acad Sci. 2001;98 :14997–5002.11752448 Zid BM , O'SheaEK. Promoter sequences direct cytoplasmic localization and translation of mRNAs during starvation in yeast. Nature. 2014;514 :117–21.25119046 Zurita-Martinez SA , CardenasME. Tor and cyclic AMP-protein kinase A: two parallel pathways regulating expression of genes required for cell growth. Eukaryot Cell. 2005;4 :63–71.15643061
PMC010xxxxxx/PMC10238306.txt
==== Front Synthese Synthese Synthese 0039-7857 1573-0964 Springer Netherlands Dordrecht 37274612 4178 10.1007/s11229-023-04178-3 Original Research Model templates: transdisciplinary application and entanglement http://orcid.org/0000-0003-0194-6837 Knuuttila Tarja [email protected] http://orcid.org/0000-0003-1870-2345 Loettgers Andrea [email protected] Department of Philosophy, University of Vienna, Universitätsstraße 7, A-1010, Vienna, Australia 2 6 2023 2 6 2023 2023 201 6 20020 4 2022 28 4 2023 © The Author(s) 2023 https://creativecommons.org/licenses/by/4.0/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The omnipresence of the same basic equations, function forms, algorithms, and quantitative methods is one of the most spectacular characteristics of contemporary modeling practice. Recently, the emergence of the discussion of templates and template transfer has addressed this striking cross-disciplinary reach of certain mathematical forms and computational algorithms. In this paper, we develop a notion of a model template, consisting of its mathematical structure, ontology, prototypical properties and behaviors, focal conceptualizations, and the paradigmatic questions it addresses. We apply this notion to three widely disseminated and powerful model templates: the Sherrington-Kirkpatrick model of spin glasses, scale-free networks, and the Kuramoto model of synchronization. We argue that what appears to be an interdisciplinary model transfer between different domains turns out, from a broader perspective, to be the application of transdisciplinary model templates across a multitude of domains. We also point out a further feature of template-based modeling that so far has not been discussed: template entanglement. Such entanglement enhances and makes manifest the conceptual side of model templates. Keywords Models Template transfer Model templates Interdisciplinarity Transdisciplinarity Network science Statistical mechanics http://dx.doi.org/10.13039/501100000781 European Research Council 818772 Knuuttila Tarja University of ViennaOpen access funding provided by University of Vienna. issue-copyright-statement© Springer Nature B.V. 2023 ==== Body pmcIntroduction Much philosophical attention has been paid to the fact that different models may be used to represent the same natural or social systems. That a seemingly identical model may be applied to different domains, often studied by different disciplines, has gotten less attention. Yet such transdisciplinary application of models is pervasive in contemporary model-based theoretical practice. Several models that have also featured prominently in the philosophy of science discussions, such as the Lotka-Volterra model, the Ising model, and scale-free networks, can be found across mathematically and computationally oriented scientific domains, from the natural to the social sciences. While philosophers have certainly noted this dissemination of certain mathematical forms, methods, and the models in which they are embedded, they have mostly left it at that. One problem has been that of finding suitable analytic tools for addressing the cross-disciplinary reach of mathematical forms and computational algorithms. The recent discussion of templates and template transfer has made such analytic tools available. The discussion of templates originates from Paul Humphreys’ notions of a theoretical template and a computational template (Humphreys, 2002, 2004). Subsequently, several case studies used these notions to examine interdisciplinary transfers of the Lotka-Volterra model (Knuuttila & Loettgers 2011, 2017; Houkes & Zwart, 2019), the Ising model (Knuuttila & Loettgers, 2014, 2016b), the virial theorem and the ideal gas law (Price, 2019), and the Chomsky hierarchy (Lin, 2022). Meanwhile, many other philosophers also addressed the interdisciplinary transfers of models, knowledge, or facts, though not using the notion of a template as their main unit of analysis (Herfeld & Doehne, 2019; Herfeld & Lisciandra, 2019; Howlett & Morgan, 2010; Jhun et al., 2018; Zuchowski, 2019). While these analyses of model transfer and migration provide a fresh and valuable perspective on contemporary modeling practice, we argue, as anticipated in Humphreys (2002, 2004), that there is a bigger picture yet to be targeted. The focus on model/template transfer thus far has tended to focus on the exchange of particular models or templates between two or more disciplines. Such a starting point attends to the questions concerning the underlying motivation, representational success, and creativity of particular model/template transfers (Houkes & Zwart, 2019). However, the bigger picture we are after here is to gain a better understanding of the contemporary modeling practice in mathematically and computationally oriented fields that is largely template-based, in both the natural and social sciences. We argue that what appears to be an interdisciplinary model transfer between different domains turns out, from a broader perspective, to be the application of transdisciplinary model templates across a multitude of domains (Knuuttila & Loettgers, 2014, 2016b). With transdisciplinarity, we refer to such research strategies that cross disciplinary boundaries aiming for a more unified approach. The notion of a model template better captures such strategies than the various template notions offered by Humphreys, due to the model template containing both a mathematical structure and a generic conceptualization that makes it able to address a particular kind of phenomenon potentially encountered in various domains. We suggest that it is precisely the entwinement of conceptual, mathematical, and computational resources from which model templates derive their transdisciplinary prowess. In order to articulate further the notion of a model template, we study three powerful model templates: The Sherrington-Kirkpatrick model (Sherrington & Kirkpatrick, 1975), scale-free networks (Barabási & Albert, 1999), and the Kuramoto model (Kuramoto, 1975a). Based on these case studies and earlier literature on templates, we further articulate the notion of a model template, consisting of its mathematical structure, ontology, prototypical properties and behaviors, focal conceptualizations, and the paradigmatic questions it addresses. We also discuss the entanglement of the three model templates, a phenomenon that has not so far been addressed in the discussion of templates. Scale-free networks and the Kuramoto model have been grounded in statistical mechanics; the origin of the Sherrington-Kirkpatrick model, and the topologies and the dynamics of the three models have been connected in subsequent research. Computational templates Whenever you have a sudden increase in usable mathematics, there will be a sudden, concomitant increase in scientific progress in the area affected. (Humphreys, 2004, p. 55). Humphreys (2002, 2004) introduced the notion of a computational template for the analysis of computational science, focusing on how computer simulations are related to traditional mathematical modeling, in particular, to solve equations (see also Humphreys, 2019a, p. 114). He pointed out that the deployment of tractable mathematics drives much of the progress in physical and other sciences, a fact that “ought to be uncontroversial, but is not emphasized enough” (Humphreys, 2004, p. 58). The prominence of a relatively small number of computational templates in the quantitatively oriented sciences is mainly due, according to Humphreys, to their computational tractability (p. 68). Moreover, the use and reuse of the same computational templates across different domains and disciplines is epistemically advantageous, since “science would be vastly more difficult if each distinct phenomenon had a different mathematical expression” (ibid.). Examples of computational templates include the Poisson distribution, the wave equation, harmonic oscillators, the Ising model, the Lotka-Volterra equations, and agent-based models. For Humphreys, such templates offer a new unit of analysis that enables envisioning a different kind of organization of quantitively oriented sciences than the conventional disciplinary matrix. Humphreys (2004, 2019a) contrasts the notion of a computational template with theoretical templates. The latter are general representational devices that occur within an interpreted theory that is about a particular subject matter (Humphreys, 2019a, p. 114). Theoretical templates constrain the substitution of schematic variables according to the intended interpretation of the theory, and they are often part of the fundamental principles of the theories in question (ibid.). Newton’s Second Law, Schrödinger’s equation, and the Lotka-Volterra equations provide, for Humphreys, examples of theoretical templates. However, the Lotka-Volterra equations do not have the same theoretical status as the two former ones, as these equations can already be considered a computational template. Indeed, Humphreys (2019a) mentions another type of theoretical template that, instead of being fundamental to its respective theory like Newton’s Second Law or Schrödinger’s equation, is itself derived from “fundamental principles of the domain and can be explained entirely in terms of the subject matter of that field” (2019a, p. 114). Such theoretical templates are derived from construction assumptions, involving also a correction set that is not necessarily explicitly formulated, giving guidance to the process of de-idealization, or correcting for abstractions, approximations, and other assumptions. Presumably, the Lotka-Volterra model is an example of the latter kind of theoretical template, though its origins are heterogeneous: both Volterra and Lotka creatively employed resources from other fields, such as mechanics, physical chemistry and even demography (Knuuttila & Loettgers, 2017). Humphreys does not explicitly discuss examples of these less fundamental kinds of theoretical templates, therefore it is unclear what templates he had in mind. To characterize theoretical and computational templates and distinguish them from each other, Humphreys studies Newton’s Second Law. He shows how computational templates can be drawn from theoretical templates by specifying parameter values as well as general features, for instance, various forces in the case of Newton’s Second Law. If one achieves a mathematically tractable equation through such substitutions, with possible uses across different fields, one has arrived at a computational template. Not all theoretical templates, such as Schrödinger’s equation, can be detached from their original theoretical context and turned into computational templates. But some can, and such computational templates can be found at different levels of abstraction. Many computational templates also trace their origin back to formal fields of inquiry, such as the Poisson distribution and scale-free networks. In his notion of a computational template, Humphreys (2004) combines two things: tractability and cross-disciplinary applicability, the latter boiling down to the detachability of a template from its original theoretical domain. The two features do not necessarily align, though they often do, since what is tractable may be usable in another context. It is tractable mathematics, related to the inability to analytically solve theoretical templates, rather than interdisciplinarity that is central to Humphreys’ treatment of computational science in his book, Extending Ourselves (2004). In rendering non-linear equations more tractable, and allowing for de-idealization, computer simulations enlarge the scope of tractable computational templates. The Lotka-Volterra model provides a good example of the expansion of tractable mathematics that computational science has made possible. Due to its non-linearity, it became a fully-fledged computational template only after the introduction of computers and computational methods, which led to its renaissance through the work of May (May, 1974), for example. Among the targets of Humphreys’ discussion are the semantic approaches that abstract from the linguistic particularities of syntactic representation. Humphreys aims to show that it is precisely the tractability of a syntactic form that matters in “bringing mathematically formulated theories to bear on natural systems” (Humphreys, 2019b, p. 68). It is important to keep in mind that for Humphreys, computational templates are not yet models but can be turned into such. According to him, computational models consist of six components.1 They are based on a computational template, such as differential or other kinds of equations, or other formal apparatuses and methods. The other components are construction assumptions, the correction set, interpretation, the initial justification of the template, and its output representation. The construction assumptions cover many things: ontology, idealizations, abstractions, and constraints. The correction set indicates, in advance, and often only implicitly, how the template could be adjusted to better fit the empirical data. Consequently, Humphreys views computational models as results of elaborate construction. Template transfer and template application While Humphreys originally addressed computational science with his template notions, other philosophers later picked up the notion of a computational template and applied it to interdisciplinary model transfer. Knuuttila and Loettgers applied Humphreys’ discussion of computational templates to study the templates and concepts that Volterra and Lotka transferred from physics and other domains to construct their respective versions of the Lotka-Volterra model (2011, 2017), which itself turned into a successful template. Furthermore, they examined the transfer of the Ising model via the Sherrington-Kirkpatrick model to neural networks and socio-economic systems (Knuuttila & Loettgers, 2014, 2016b). Several other studies of model transfer followed suit, highlighting different aspects of model transfer, and employing template notions. Houkes and Zwart (2019) studied the application of the Lotka-Volterra model to technology transfer, showing how some of these transfers were more conformist, and some more creative in nature. Price (2019) argued, using the quantum theory of atoms in molecules (QTAIM) as an example, that the application of a template to a new domain requires the construction of a “landing zone” to be successful. Bradley and Thébault (Bradley & Thébault, 2019) distinguished between imperialism and migration in model transfer, of which the latter explicitly requires “re-sanctioning” to allow for the application to a new domain. Lin (2022) examined the transfer of the Chomsky hierarchy to computer science and cognitive biology, finding that the spillover of knowledge claims from former applications could be crucial for the further reapplication of a mathematical construct. Several of these studies address one particular ambiguity in Humpreys’ original template account (e.g., Houkes & Zwart, 2019; Knuuttila & Loettgers, 2014; Lin, 2022). While discussing computational models, Humphreys appears to consider computational templates as merely tractable formal templates to be complemented by other components of the computational model (2004, p. 103, see above). Yet he approaches elsewhere the construction of computational templates in terms of construction assumptions similar to those of computational models, including an assumed ontology covering “mechanisms that operate within and between systems” (2004, pp. 78–79). Clearly, computational templates understood in this latter way do not boil down to formal templates, and consequently, it is difficult to distinguish computational templates from computational models (Houkes & Zwart, 2019). In order to distinguish between these two interpretations of computational templates, we call them thin and thick notions of a computational template. The notion of a model template comes closer to the thick notion of a computational template, though it underlines the conceptual side of templates, to which Humphreys did not pay too much attention until his final work (Humphreys, 2023). Humphreys in Humphreys (2019a) turned explicitly to template and model transfer, launching two novel template notions, a formal template and a transdomain template, leaving computational templates aside. Humphreys (2019b, 2023) explains, in line with the thin notion of computational templates, that his “focus was originally on computational aspects of representations and so a primarily formal approach to transfer seemed suitable […]” given that “with only formal application conditions, the tractability of a given model is independent of the application” (Humphreys, 2023). However, he acknowledges that in model transfer, one also needs to pay attention to other constraints related to the empirical adequacy of the model. Humphreys (2019a) uses network models as examples of formal templates as they have their origin in formal fields of inquiry. He claims that in their application, all empirical content is contained in the (often complicated) mappings of the formal template to a particular target system (2019a, pp. 116–117). However, not all cases in which a seemingly identical formal template is applied in different domains qualify as instances of template transfer. Humphreys argues that there are cases in which an identical formal template is arrived at by different, independently justified construction assumptions. Volterra’s population biological and Goodwin’s economic construal of the Lotka-Volterra equations serves as an example of such a case for Humphreys, and he also refers to Lotka’s and Volterra’s different construals of Lotka-Volterra equations (see Knuuttila & Loettgers 2011, 2017). Humphreys calls such templates transdomain templates. It is important to note that in such cases, according to him, there is no template transfer, as the templates are arrived at independently of each other despite their similar formal structure. Finally, Humphreys recognizes that there are “stylized formal templates”—or “off-the-shelf models”—that “are opportunistically justified at the system level by analogical reasoning from their previous successful applications to systems that are recognized as similar” (2019a, p. 115). He does not further contemplate these cases, however, arguing instead that construction assumptions of formal and transdomain templates can be made explicit, and empirically verifiable (at least in principle). In Humphreys’ (2019a) discussion of template transfer, the theoretical and ontological aspects of templates, as well as the tractability considerations, recede into the background. His focus is on formal and empirical aspects of template transfer which is especially clear in his discussion of formal templates. Since Humphreys (2019a) does not consider the cases of independently derived transdomain templates as instances of model transfer, the earlier cross-disciplinary applicability of computational templates is lost, apart from formal templates. As a result, it becomes unclear why scientists would engage in template transfer and application in the first place. The notion of a model template (Knuuttila & Loettgers, 2014, 2020) seeks precisely to address this question. A model template is “a formal platform for minimal model construction coupled with very general conceptualization without yet any subject-specific interpretation or adjustment (Knuuttila & Loettgers, 2016a, 2016b, p. 382). The notion aims “to capture the intertwinement of a mathematical structure and associated computational tools with theoretical concepts that, taken together, depict a general mechanism that is potentially applicable to any subject or field displaying particular patterns of interaction” (Knuuttila & Loettgers, 2014, p. 295). The point Knuuttila and Loettgers make is that even in the case of templates originating from formal fields, such as scale-free networks, there is a generic ontology and conceptualization that suggests the kinds of systems and problems it is applicable to, and thus motivates and guides template transfer and application. The shift from formal and (thin) computational templates to model templates as units of analysis takes into account the theoretical and conceptual resources that templates make available and that facilitate their applications to new domains, providing material for further theoretical development. Furthermore, the notion of a model template also addresses the aforementioned worry that several interpreters of Humphreys’ work on templates have voiced concerning the difficulty of distinguishing computational templates from computational models. The broader notion of a model template appears fruitful because it is often indeed difficult to tell apart templates and models—and in fact templates and methods as well, as we will later show in the case of the Sherrington-Kirkpatrick model. One can discern cases, in which the model itself is the primary object of study, and, on the other hand, cases in which a model is applied to some specific system in a particular domain. In the former cases, the model comes closer to a template, yet it still embodies much more conceptual and theoretical content than a mere formal template does. Consequently, we consider the notion of a model template as a promising alternative to Humphreys’ computational and formal templates in the analysis of cross-disciplinary model application, though being simultaneously in line with his original vision. In what follows, we seek to articulate the notion of a model template further by developing a categorization of the different constituents of model templates by studying three successful and widely applied model templates: the Sherrington-Kirkpatrick model, the scale-free networks model, and the Kuramoto model.2 The respective origins of the three templates in physics, applied mathematics, and the study of synchronization and self-organization, display the heterogeneity of the cross-disciplinary trajectories of model templates. However, despite their different origins, the three templates have also come together, becoming entangled with one another. Model templates The Sherrington-Kirkpatrick model was originally introduced in physics to study spin glasses, later finding applications in as disparate fields as statistical physics, computer science, neural networks, and financial markets (Knuuttila & Loettgers, 2016b, 2020). The scale-free networks and the Kuramoto model can be located in the more formal fields of network science and the study of synchronized oscillators, neither of them tied to any particular discipline. Scale-free networks originated in mathematics, in the investigation of random networks (Erdös & Rényi, 2011). The study of networks only entered physics once the scale-free networks were grounded in statistical mechanics and critical phenomena, in an attempt to assign scale-free networks a universal status (Barabasi, 2018). The Kuramoto model addresses the general phenomenon of synchronized oscillation. The interest in such oscillations goes back to the seventeenth century when the Dutch mathematician and physicist Christiaan Huygens observed the synchronization of two pendulum clocks. The omnipresence of synchronized oscillations in nature made them a subject of increasing interest, especially in mathematically oriented domains (Strogatz, 1994). The introduction of the Kuramoto model in the mid-1970s drew many of these different attempts together, making them mathematically tractable (Kuramoto, 1975a, 1975b). Like scale-free networks, the Kuramoto model was linked to statistical mechanics. Sherrington-Kirkpatrick (SK) model The Sherrington-Kirkpatrick (SK) model (Sherrington & Kirkpatrick, 1975) was constructed to study the properties and behavior of spin glasses. The SK model has been a subject of intense mathematical and theoretical research within statistical mechanics for nearly 50 years (Panchenko, 2012). The model was constructed to study spin glasses that are made of particular kinds of magnetic alloys. It models the competing ferromagnetic and antiferromagnetic interactions taking place in said alloys. These interactions lead to exceptional properties and behaviors, which explains the interest physicists have taken in spin glasses. But the SK model did not remain in physics, it became a model template as it was applied to a large number of different phenomena such as, for example, pattern recognition in neural networks (Hopfield, 1982), social networks, and peer effects (Bertrand et al., 2000; Bramoullé et al., 2009). The SK model has its origin in the Ising model (Ising, 1925), which also has become a model template in itself (Knuuttila & Loettgers, 2014, 2016b, 2017, 2020). In the Ising model, magnetic moments are described as binary variables located on a grid exhibiting only the nearest neighbor interactions. At high temperatures, the magnetic moments are subject to thermal fluctuations, but with a decreasing temperature, they start to align—the material performing a phase transition from antiferromagnetic to ferromagnetic. The phase transition takes place at a critical temperature TC and is accompanied by a pronounced increase in magnetic susceptibility χ,(i.e., a measure of how much a material will become magnetized in a magnetic field), and in specific heat c (i.e., the amount of heat needed to raise the temperature by one degree Celsius per unit mass). Given the interactions among the magnetic moments, the behavior of spin glasses can be analyzed as a particular kind of collective phenomena. Collective phenomena are the results of the interaction between constituents of a system, such as ferromagnetism being caused by the interaction between magnetic moments. The notion of collective phenomena aligns ferromagnetic material and spin glasses, enabling them to be modeled using the same mathematical tools. In constructing the SK model, Sherrington and Kirkpatrick made use of the Ising model. They formalized the phenomenon of disorder resulting from competing interactions between magnetic moments in spin glasses in which neighboring magnetic moments try to align a magnetic moment in different directions, but modified it in such a way that it would allow for the exploration of disorder and frustration in spin glasses. In this instance, frustration is understood as resulting from not being able to satisfy the competing interactions (Parisi, 1986). In the mathematical rendering of the SK model, the interactions are modeled as a function of the distance between the magnetic moments Si and Sj, which reads as Jij=J(Ri-Rj), with Ri and Rj as the positions of the magnetic moments on, for example, a two-dimensional grid. The positive values of Jij correspond to ferromagnetic, and negative values to antiferromagnetic couplings. The overall energy of the system is described by the following equation:1 E=-∑i,jJijSiSj The disorder and competition among magnetic moments result in a highly structured energy landscape, consisting of a large number of local energy minima. As Mézard et al. write: “The picture which has emerged is that the main characteristic of the glassy phase is the existence of a large number (infinite when the number of spins N→∞) of equilibrium states α=1,2,⋯⋯” (Mezard et al., 1987, p. 13). Due to the disorder and frustration, no ground state can be assigned to spin glasses. For the SK model to function as a model template, the topology of its rugged energy landscape, with its large number of local energy minima, is crucial. As a model template, the SK model can be applied both to various kinds of empirical phenomena as well as used as an optimization method. An example of the first kind of application is the Hopfield model (Hopfield, 1982), which is an artificial neural network modeling auto-associative memory. In this case, the large number of local energy minima allows for the storage of a large number of patterns. The application of the SK model to neuroscience brought along concepts from statistical mechanics such as phase transitions, order parameters, critical exponents, and symmetry breaking, as well as specific methods like the method of mean field approximation and the replica symmetry breaking that became detached from their original context (Knuuttila & Loettgers, 2014, 2020, see also discussion below, in Sect. 5). On a more general level, both the SK and Hopfield models can be approached as networks having specific topologies and dynamics. The topology of the energy landscape of the SK model is crucial also for its application to optimization. The research on the SK model has led to computational methods that can be applied to many combinatorial optimization problems, such as NP complete problems3 like the traveling salesman problem, matching problem, or protein folding. According to Parisi, “it is quite possible that on the long run the applications of these ideas beyond solid state physics will be the most interesting ones.” (Parisi, 1986, p. 2). The traveling salesman problem consists of calculating the shortest route through different cities, such that the salesman passes every city only once (Fig. 1). It is a combinatorial optimization problem, in which different routes have different lengths and travel costs. In the traveling salesman problem as in the case of the SK model, the goal is to find the true minimum in a landscape of a large number of local minima. What makes the problem belong to the class of NP problems is the fact that the number of possible tours increases rapidly with the number of cities n, which is given by n-1!/2. In the case of n=50 it would take a longer time than that of the existence of our universe to calculate all the different tours.Fig. 1 The graph shows a cut through the energy landscape of a system consisting of a large number of local energy minima The method of simulated annealing provides another example of such a computational application. This method is named after a procedure in metallurgy, which aims to find the state of the lowest energy of a metal. To avoid getting trapped in a local energy minimum, the metal is heated up in irregular time intervals during the process of cooling down. Figure 1 shows a cut through the rugged energy landscape of a spin glass. The objective function in this diagram represents, in the case of spin glasses, the energy function. The local minima of spin glasses can be explored by calculating the probability of a configuration at a specific temperature, or, alternatively in the case of the salesman problem, the probability assigned to a specific route. As shown in the schematic diagram, even when the energy function or the cost function results in a local metastable energy minimum, small perturbations, e.g., heating up in the case of metals, guarantee that the system does not get trapped in them. Larger parts of the energy or cost landscape get explored in this way. Though such a method does not give an exact solution, it can provide a satisfactory approximation (see Hopfield & Tank, 1986). Scale-free (SF) networks According to a common narrative among network scientists, networks provide a natural way of analyzing our highly connected world. This narrative has gained momentum since huge amounts of data have become available from, for example, the human genome project, food webs, the web of human sexual partners, and the spread of infectious diseases. Scale-free networks have become, together with other network models such as small-world (Watts & Strogatz, 1998, p. 440) and random networks (Erdös & Rényi, 2011), central to the field of network science. The differences between these three different network models lie in their respective topologies and dynamics that influence what kind of problems and phenomena they can be applied to. Crucial to the SF network is the assumption that connections k follow a power law distribution P(k)∼k-γ, where the scaling parameter γ is a constant lying in the range 2<γ<3. In Fig. 2, the peculiar feature of the power law distribution is illustrated in comparison with a random distribution. The left-hand side in this figure depicts a probability distribution for randomly distributed k and the right-hand side a probability distribution in which k follows a power law. In the case of the random distribution, the probability distribution of k has the form of a bell-shaped curve with the maximum representing the mean value of k and the width representing the standard deviation. This distribution is the well-known normal distribution.Fig. 2 The two curves illustrate the difference between random and power law distributions and their respective effect on the network structure (Glushnev et al., 2003, p. 53) In the second case, the probability distribution follows a power law—the probability exponentially decreases with increasing k. A comparison of the two distributions shows that for those values of k where the random distribution is already zero, in the case of the power law distribution still some finite probability exists for k. This specific feature is called a fat-tailed distribution. Fat-tailed distributions have the property of decaying very slowly, allowing for more outlier data than is the case with normal distributions. As a result, extreme events such as large earthquakes, blackouts, or stock market crashes are more likely to occur than what would be predicted by normal distributions. In the SF network, a power law distribution allows for the formation of a few highly connected hubs. This is not the case in a random network, where the connections between the nodes follow a normal distribution. Figure 3 shows a comparison between a random and a SF network. In the topology of the SF network, some nodes are more connected (i.e., “hubs”) and others are less connected. Because of the high connectivity of hubs, SF networks are highly robust against the random removal of the connections between the nodes.Fig. 3 The illustrations show the difference in the distributions of nodes in a random and a scale-free network. Some dots in the scale-free network represent the highly connected hubs (Barabási & Oltvai, 2004) When Barabási and Albert introduced the SF network in 1999 (Barabási & Albert, 1999), there were already available data on protein networks of yeast and the World Wide Web (WWW).4 Using a robot to detect which web pages in the WWW are connected to each other, they claimed that they were able to directly infer from the data that the topology of the WWW has a small number of highly connected nodes, the “hubs”. Barabási and Albert claim that for large values of k the degree distribution follows a power law Pk∼k-γ. They extended their empirical findings to cover various kinds of systems, arguing that such unrelated networks as scientific networks, gene regulatory and protein networks, and food webs also show similar topologies that follow a power law and are thus by definition scale-free (Albert & Barabási, 2002). It is important to note that the connection between power law distribution and scale-freeness is central to statistical mechanics and phase transitions. Indeed, in grounding SF networks in statistical mechanics, Barabási and Albert sought to give an ontic, and even a universal interpretation of networks. Such universality would justify the application of SF networks to various kinds of systems, a tendency that can be observed in network science more generally. To justify such a move, Albert and Barabási invoke the earlier work on random networks, which are supposed to produce second order phase transitions (continuous phase transitions) if sufficiently many links have formed in the network (Albert & Barabási, 2002). The phase transition is supposed to result from an ‘explosive-percolation,’ meaning that a large fraction of nodes suddenly become connected. In the case of scale-free networks, the question then becomes whether they also exhibit phase transitions, and on what parameters such transitions might depend. The link between networks and statistical mechanics made physics concepts such as phase transition, symmetry breaking, order parameters, and critical exponents available for network theorists. As a result, these concepts became independent from their specific meaning in physics—as was also the case with the Hopfield model. In contrast to the Hopfield model, network models such as the SF are more general and abstract and do not address any particular system. While Albert and Barabási understand SF networks as an extension of models such as the Ising model and SK spin glass model, they nevertheless point out that models with static grids and equidistant nodes do not apply to networks such as the WWW or chemical networks in cells: “The success of these modeling efforts (Ising and SK model) is based on the simplicity of the interactions between the elements: there is no ambiguity as to what interacts with what, and the interactions are uniquely determined by the physical distance. We are at a loss, however, to describe systems in which physical distance is irrelevant or for which there is ambiguity as to whether two components interact” (Albert & Barabási, 2002, p. 2). Instead of the distance of nodes, it is the distribution of the connections between the nodes that is crucial in the SF network. The main idea of how the nodes become connected in the SF network over time is captured in the saying, ‘the rich are getting richer’. The probability that a new node will be connected to one of the nodes in the network depends on the number of connections the node has to other nodes. This phenomenon is called preferential attachment. While Albert and Barabási seek to accumulate evidence for the omnipresence of SF networks in real-world systems, this claim has been met with criticism. Scientists do not usually have access to the complete network, and it is questionable whether they can infer from the distribution of the sub-network the distribution of the whole network (Stumpf et al., 2005). Moreover, the identification of power law distributions from real data is difficult. For example, such an identification could be complicated by fluctuations, especially in the tail of the distribution (Clauset et al., 2009). Consequently, it does seem that the universality of scale-free networks should be taken with a grain of salt. The Kuramoto model The Kuramoto model (Kuramoto, 1975a, 1975b) is a model of coupled oscillators developed by Yoshiki Kuramoto to study synchronization between different oscillatory systems. Synchronized oscillations belong to the larger and more general class of such phenomena as synchronized behavior and self-organization. There are many different instances of such synchronization in nature as well as in engineered systems. On an abstract level, synchronization describes a specific relatedness in dynamical systems—they become synchronized in case phenomena occur at the same time. In movies, the flow of pictures and sound is synchronized, birds and fish show swarming behavior, and our different electronic devices such as computers and mobiles are often synchronized. Examples of synchronized oscillations, in turn, are the millions of fireflies in Southeast Asia blinking in sync, an audience after a play falling in sync when clapping their hands, or the circadian clock, which is entrained by the environment and regulates the various processes taking place in an organism in a 24-h cycle (Strogatz, 2003). Synchronized oscillations have been characterized by Pikovsky as the “adjustment of rhythms of oscillating objects due to weak interactions” (Pikovsky et al., 2002). The dynamics of oscillatory systems are examined mathematically by the dynamics of the phases of single and coupled oscillators in phase space (Strogatz, 1994). The following example may be helpful. Each of the oscillators can be thought of as a runner on a circular running track. In the unsynchronized case, the runners orbit the circle with different frequencies (i.e., different speeds) and phases (i.e., different positions on the track). In the synchronized case, the frequencies and phases of the runners coincide. The coupling between the oscillators is modeled as a coupling between their phases. The resulting dynamic of the phases of the oscillators—i.e., the change in the phase of the oscillators over time—is described by a trajectory in the high dimensional phase space, which entails all possible uncoupled, coupled, and synchronized states of oscillators. The mathematical tractability of the Kuramoto model is far from obvious. Kuramoto was able to solve exactly a system of infinitely many differential equations, all nonlinear and coupled together. The model consists of a system of coupled non-linear differential equations of the following form:2 dθidt=ωi+KN∑j=1Nsinθj-θii=1,⋯,N In this system, θi is the phase of oscillator i, ωi its frequency, and K, the coupling strength between the oscillators. The non-linearity enters the model by sin function. In the original version, the oscillators are identical and located on a one-dimensional grid (string). The oscillators are coupled by their phases via the sin function. Kuramoto’s derivation of his model of synchronized oscillators was inspired by Arthur Winfree’s earlier model (Winfree, 1967) that addressed synchronized phenomena, such as synchronization in pacemaker cells and the circadian clock, or blinking fireflies. He suggested conceptualizing the transition from an unsynchronized to a synchronized state as a phase transition (Winfree, 1967, 2001). In a remarkable feat, Kuramoto was able to tie the problem of synchronization to statistical mechanics, making the model computationally tractable in one stroke. As such, Kuramoto’s model provides yet another example of transferring the concepts of phase transition, symmetry breaking, order parameter, and critical exponents into the research of complex networks. The tractability of the model depends on what is called the order parameter of the system, a value that indicates how synchronized the system is. In statistical mechanics, this value indicates the occurrence of a phase transition. Kuramoto employed the mean field method, in which a system of interacting entities is described as independent entities coupled to a mean field. The technical details do not concern us here, but it is important to recognize that with the order parameter, the phases and, thereby the oscillators, become decoupled and consequently computationally tractable. Because of its computational tractability and flexibility, the Kuramoto model is a powerful computational template. It can be found in a large spectrum of research areas such as neuroscience, condensed matter physics, cell and molecular biology, sociology, and electrical engineering. Over time, the Kuramoto model has been expanded from a one- to a two-dimensional grid (Wiley et al., 2006), from identical to non-identical oscillators (Brede, 2008), and made to cover different network topologies such as small-world networks (Watts & Strogatz, 1998). Like SF networks, the Kuramoto model has some grounding in statistical physics, getting its model template status from this conceptual framework, in addition to the previous work on synchronized oscillators. A categorization of the constituents of model templates The previous discussion of the three models shows that each model can be considered a model template due to the way their mathematical structure intertwines with computational methods and theoretical concepts. Table 1 provides our categorization of the different constituents of a model template aiming to analyze the notion discussed in the literature so far (Knuuttila & Loettgers, 2014, 2016b). In Table 1, we distinguish between the mathematical structure, ontology, prototypical properties and behaviors of model templates, and their generic and system-specific conceptualizations, as well as the paradigmatic problems the templates are used to address, mentioning also some applications, all of them discussed in our presentation of the cases above. We do not claim this categorization to be exhaustive.Table 1 The constituents of model templates Dimensions of the model template Sherrington-Kirkpatrick model Scale-free networks The Kuramoto model Mathematical structure: The overall energy of the SK model is given by: E=-∑i,jJijSiSj, with Jij=JRi-Rj, depending on the distance between the magnetic moments Network topology evolves following a preferential attachment mechanism: the connection between the nodes is given by a power law distribution P(k)∼k-γ Non-linear coupled differential equations: dθidt=ωi+KN∑j=1Nsin(θj-θi) Ontology: Magnetic moments are modeled as binary variables, taking values + 1 or −1. The magnetic moments are placed on a grid and their coupling is modeled as a function of the distance between the magnetic moments Si and Sj,. Network ontology of nodes and edges, whose distribution follows the power law and is thus scale-free Identical oscillators, which are either placed on a grid or some network, such as small-word networks. The coupling is modeled via the phases of the oscillators Prototypical properties and behaviors: A rugged energy landscape consisting of a huge number of metastable states, which are due to disorder and frustration in the interaction between magnetic moments This specific property finds its expression in the topology, in which the magnetic moments are placed on the edges of a grid in random directions Networks containing hubs, with a high degree of connectivity. These networks are highly robust against the random removal of the connections between the nodes In this case, the rigorous topology of a grid is replaced by a more dynamical topology A phase transition from unsynchronized to synchronized oscillations. The Kuramoto model was the first model to exhibit such transitions The topology of the model is flexible because different kinds of grids, as well as small-world, or scale-free networks, can be implemented within this framework Conceptualizations: Generic System-specific Cooperativity, phase transitions, symmetry breaking, criticality and critical exponents, phase transitions Temperature and its different re-interpretations (e.g., as noise), disorder, and frustration Network topology and dynamics, phase transitions, symmetry breaking, critical exponents, universality Hubs, scale-freeness, power law Nonlinear dynamics, phase transitions, attractors Synchronicity among identical and non-identical oscillators Paradigmatic problems: Clustering, collective phenomena, and combinatorial optimization problems Formation of hubs, modeling the evolution of structure, and robustness in complex networks Interaction leading to synchronization and self-organization Examples of applications: Pattern recognition in neural networks, peer group behavior, welfare participation, and school achievements Optimization problems: traveling salesman problem, matching problem, scheduling Biological, ecological, and social networks such as neural networks, chemical networks in cells, food webs, celebrity and academic networks, and WWW Power systems analysis, brain dynamics, synchronizing Josephson junctions in superconducting materials, synchronization in chemical oscillators Mathematical structure forms the core of the model template. Together with ontology, covering basic objects and relations, and the prototypical properties and behaviors of the model template, it leads to a particular topology that is crucial for the application of the template. Typically, such applications address some paradigmatic general problems across different disciplines, such as collective phenomena, clustering, formation of hubs, synchronization, and self-organization. Generic conceptualizations apply independently from a specific research context, typically deriving from a more general theoretical framework, such as statistical mechanics. The system-specific conceptualizations are those that are crucial for identifying the template in question, and may also need considerable reinterpretation in various applications—showing how any subject-specific application of a template typically involves elaborate model construction. Below we illustrate the categorization of the constituents of model templates, using the SK model as an example. Due to reasons of space, we cannot discuss the other two models here, but do hope that the preceding discussion and Table 1 indicate how they function as model templates. The peculiar topology of the SK model is a key to its twofold function as a template for modeling various empirical phenomena, and a method of solving combinatorial optimization problems. As an example of the former, in constructing his model of auto-associative memory, Hopfield saw an opportunity of making use of the topology of the SK template to model the storage of a large number of patterns. He turned magnetic moments into neurons and rendered them as binary variables, although neurons in biological networks are non-linear units. Neurons can be made binary by introducing a threshold: if the summed-up signals exceed a threshold, the neuron fires an action potential and goes over into a quiet state (see Knuuttila & Loettgers, 2014). The topology of the SK model enables it to address particular kinds of paradigmatic problems such as clustering and optimization. As for conceptualizations, the SK model is firmly rooted in statistical mechanics, exemplified by its generic concepts of collective phenomena and phase transitions. The central concept of collective phenomena allows for different kinds of constituents and interactions and so it did not require any reinterpretation in the case of the Hopfield model. The concept of phase transition in the Hopfield model describes the transition from a state in which the network can recognize a stored pattern to a state in which it loses this ability. The critical parameter, in turn, refers to the storage capacity of networks, e.g., the number of patterns that can be stored in the network. On the other hand, system-specific concepts like temperature needed more thorough and less straightforward reinterpretation. In neural networks, temperature has been interpreted as noise, hindering pattern recognition. In contrast, in its use as a computational method for combinatorial optimization problems, the conceptual side of the SK model template is not so important. The topological similarity between the rugged energy landscape of the SK model and the landscape of metastable states of the optimization problems is crucial. Template entanglement What we find particularly interesting in our study of the three model templates—the SK model, SF networks, and the Kuramoto model—is that they are not isolated templates. Even though they originated from different domains, they also have become entangled with one another. This entanglement has happened both through establishing connections between the topologies and dynamics of these templates and through the adoption of the conceptual framework of phase transitions from statistical mechanics. Albert and Barabási sought to link their networks to statistical mechanics by making them scale-free via possessing a specific topology in which the distribution of the nodes follows a power-law distribution. Scale-freeness is a generic concept from statistical mechanics by which systems, whose patterns are independent from scale, become assigned to universality classes in which the universality corresponds to the behavior of systems as they approach a critical point, such as the phase transition in magnetic systems and liquid–gas transitions at their respective critical temperatures. In the case of these two systems, they are described by the same power law sharing a critical exponent. Different universality classes are distinct from each other by the critical exponent assigned to them. Via the property of scale-freeness, the conceptual framework of statistical mechanics was thus introduced to SF networks, although concepts such as symmetry breaking and phase transition remain vague in this context. Moreover, their properties are still subject to intense debate (Broido & Clauset, 2019; Zhang et al., 2015). SF networks are often applied as “plug-in” templates, being applied to big data collected in, for example, high-throughput omics studies, co-authored works, or the World Wide Web. The topology of the original Kuramoto model agrees at the outset with the topology of Ising-style models. The oscillators are equidistantly located on a grid and interact with each other. The interaction between the oscillators generates collective phenomena, expressed in the Kuramoto model as a phase transition from unsynchronized to synchronized oscillations. This occurrence of a phase transition depends on the coupling constant between the phases of the oscillators. By approaching synchronization as a phase transition, Kuramoto related his model to statistical mechanics. With this move, the Kuramoto model became embedded into the conceptual framework of statistical mechanics, though mainly as a result of making it computable. But the entanglement of the Kuramoto model extends also to network models due to the flexibility of its topology. The constituents, identical or non-identical oscillators, can be placed on a grid, like the magnetic moments in the SK model, or they can become the nodes in a SF or small-world network. In referring to this entangled set of models, Watts and Strogatz point out that:[n]etworks of coupled dynamical systems have been used to model biological oscillators, Josephson junction arrays, excitable media, neural networks, spatial games, genetic control networks and many other self-organizing systems. Ordinarily, the connection topology is assumed to be either completely regular or completely random. But many biological, technological and social networks lie somewhere between these two extremes. (Watts & Strogatz, 1998, p. 441) Watts and Strogatz’s small-world networks inhabit this middle ground. They are arrived at “rewiring” regular networks by introducing increasing amounts of disorder. These developments have had a feedback effect on spin glass research and the SK model as researchers have used small-world or scale-free networks in modeling the properties of such systems. We suggest that the entanglement between model templates during their development as well as in their application appears as an important characteristic of modeling practices making use of templates. Template entanglement is apparent in the development of a novel transdisciplinary research field devoted to the study of dynamical processes in complex networks (e.g., Barrat et al., 2008). Within this field, the properties of existing model templates are explored, and novel model templates are constructed from existing ones and applied to physical, chemical, biological, social, and ecological systems. The entanglement between model templates is an essential part of this transdisciplinary modeling practice in which model templates themselves are subject to changes and extensions. The implications of such entanglement cut deep: what is at stake may come down to a new kind of template-based unification yet to be addressed by philosophers of science. Conclusions The omnipresence of the same basic equations, function forms, algorithms, and quantitative methods is one of the most striking characteristics of the contemporary quantitatively oriented sciences making heavy use of modeling. We have suggested that the notion of a model template provides a suitable unit of analysis for studying the transdisciplinary nature of present modeling practices. By combing theoretical discussions and case studies of three widely applied model templates, the Sherrington-Kirkpatrick model, scale-free networks, and the Kuramoto model, we have articulated the notion of a model template by developing a categorization of the constituents of a model template. This categorization includes the mathematical structure, a generic ontology, the prototypical properties and behaviors, the core conceptualizations of a template as well as the paradigmatic problems to which the template can be applied. Many of these features are immanent in the mathematical structure of model templates, being crucial for their application to different domains. The notion of a model template covers but does not reduce to what Houkes and Zwart (2019) describe as a “thin” intentional interpretation of a template that they studied in the context of the transfer of the Lotka-Volterra template. The “thin” intentional interpretation leaves open the precise mechanism and entities at work, yet makes use of some generic features of the interaction between the entities, such as mutualism, predation, and competition. Moreover, the model template notion draws together and unifies the different features of Humphreys’ notions of computational and formal templates. As discussed in Sects. 2 and 3, Humphreys used the notion of a computational template to approach the organization of computational science, and the notion of a formal template when discussing model transfer. The model template notion recovers some crucial aspects of Humphreys’ thick notion of a computational template such as ontology understood in terms of the generic mechanism in operation, but goes beyond it in including also focal conceptualizations of a template and the paradigmatic questions it is used to address. In his latest paper, Humphreys (2023) appears to accept the usefulness of the notion of a model template in writing that “in order to properly account for model transfer, the vehicle of transfer must be taken to be model templates rather than formal templates.” The constituents of the notion of a model template are articulated at such a level of abstraction that would allow conceiving of model templates as genuinely transdisciplinary devices that can be used for model construction to address particular kinds of problems and phenomena across the disciplinary spectrum. While we acknowledge the importance of analogical reasoning in model and template transfer, the advantage of the template-based approach is precisely due to its transdisciplinary scope. Moreover, the notion of a model template also allows the entanglement of transdisciplinary templates to come into view: addressing the entanglement of templates requires a broader perspective on templates than viewing them as merely mathematical and computational objects able to be transferred from one domain to another. As we have argued, the usefulness and applicability of model templates accrues, apart from their computational benefits, also from the conceptual, theoretical, and computational resources they offer. But what justifies, beyond such resources, the transdisciplinary application of model templates to particular kinds of problems mushrooming across different disciplines? What makes us think that the seemingly similar phenomena are really due to similar kinds of general principles? Many complexity theorists and network scientists tend to suppose that the same basic organizational principles and topologies govern most diverse materially different systems. Such a view motivates, as we have seen, Barabási’s universalist approach to scale-free networks that has nevertheless met a lot of criticism (e.g., Broido & Clauset, 2019). Indeed, different applications of the same model template may vary greatly in terms of their success. While the application of the SK model to auto-associative memory (Hopfield, 1982) has been constitutive for the entire field of neural networks, Hopfield’s further attempt to apply neural network models to earthquakes has not been so fruitful (Herz & Hopfield, 1995). Kuramoto, in turn, did originally consider his model mainly as a mathematical exercise, but the model has since found truly surprising applications, e.g., to the Josephson effect. And the applications have been piling up: from arrays of coupled lasers to power grids and cardiac arrhythmias. Yet, the Kuramoto model cannot be too widely applied given the restrictive conditions that need to be met by any application. What makes it so important is that it provides a well-understood case of spontaneous synchrony. Strogatz (2003) has put this succinctly by pointing out that “[the] Kuramoto model has always been a solution waiting for a problem” (p. 170).5 Perhaps this is precisely what can more broadly be said about all model templates: they provide well-understood and properly defined schematic models of some general phenomena and are as such potentially applicable to numerous problems and targets. Acknowledgements The authors wish to thank the two anonymous reviewers, Chia-Hua Lin, and all the participants of the “Transdisciplinary Model Transfer and its Interfaces” workshop in Vienna, March 12–13, 2021. A special thanks go to the late Paul Humphreys for his kindness, unwavering support, and collegiality throughout the years. Funding Open access funding provided by University of Vienna. This project received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (Grant agreement No 818772). Declarations Conflict of interest The authors declare that there are no conflicts of interest. 1 Computational models do not need to be based on a computational template, but Humphreys does not discuss this possibility. 2 As our examination of these three models draws on the theoretical discussion and earlier case studies discussed above, the categorization should not be taken as only a result of inductive exercise. Our approach is in line with the philosophical discussion of case study methodology, according to which case studies supply both evidential and interpretative resources for philosophical theorizing (e.g., Kinzel 2015; Knuuttila and Loettgers 2016a). 3 With NP complete problems, Parisi refers to a class of computational problems for which there can be no effective computational algorithm or it has not been found yet. 4 These data only became available by the late 1990s and the early 2000s (Barabási & Albert, 1999). 5 See also Gelfert’s article on Turing patterns as “a model in search of targets” (Gelfert 2018). Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. ==== Refs References Albert R Barabási A-L Statistical mechanics of complex networks Reviews of Modern Physics 2002 74 1 47 97 10.1103/RevModPhys.74.47 Barabasi A-L Network science—The scale free property 2018 Cambridge University Press Barabási A-L Albert R Emergence of scaling in random networks Science 1999 286 5439 509 512 10.1126/science.286.5439.509 10521342 Barabási A-L Oltvai ZN Network biology: Understanding the cell’s functional organization Nature Reviews Genetics 2004 5 2 101 113 10.1038/nrg1272 Barrat A Barthélemy M Vespignani A Dynamical processes on complex networks 2008 Cambridge University Press Bertrand M Luttmer EFP Mullainathan S Network effects and welfare cultures* The Quarterly Journal of Economics 2000 115 3 1019 1055 10.1162/003355300554971 Bradley S Thébault KPY Models on the move: Migration and imperialism Studies in History and Philosophy of Science Part A 2019 77 October 81 92 10.1016/j.shpsa.2017.11.008 Bramoullé Y Djebbari H Fortin B Identification of peer effects through social networks Journal of Econometrics 2009 150 1 41 55 10.1016/j.jeconom.2008.12.021 Brede M Synchrony-optimized networks of non-identical Kuramoto oscillators Physics Letters A 2008 372 15 2618 2622 10.1016/j.physleta.2007.11.069 Broido AD Clauset A Scale-free networks are rare Nature Communications 2019 10 1 1017 10.1038/s41467-019-08746-5 Clauset A Shalizi CR Newman MEJ Power-law distributions in empirical data SIAM Review 2009 51 4 661 703 10.1137/070710111 Erdös P Rényi A Newman M Barabási A-L Watts DJ On the evolution of random graphs The structure and dynamics of networks 2011 Princeton University Press 38 82 Gelfert Axel Christian A Hommen D Retzlaff N Schurz G Models in Search of Targets: Exploratory Modelling and the Case of Turing Patterns Philosophy of science: Between the natural sciences, the social sciences, and the humanities. European Studies in Philosophy of Science 2018 Springer International Publishing 245 69 Glushnev, N., O’Donovan, B., Troussov, A. (2003). Efficient implementation of morphological finite-state transition networks employing their statistical properties. In International conference on natural language processing and knowledge engineering, 2003: Proceedings (pp. 50–55). 10.1109/NLPKE.2003.1275868. Herfeld C Doehne M The Diffusion of scientific innovations: A role typology Studies in History and Philosophy of Science Part A 2019 77 October 64 80 10.1016/j.shpsa.2017.12.001 Herfeld C Lisciandra C Knowledge transfer and its contexts Studies in History and Philosophy of Science Part A 2019 77 October 1 10 10.1016/j.shpsa.2019.06.002 Herz AVM Hopfield JJ Earthquake cycles and neural reverberations: Collective oscillations in systems with pulse-coupled threshold elements Physical Review Letters 1995 75 6 1222 1225 10.1103/PhysRevLett.75.1222 10060236 Hopfield JJ Neural networks and physical systems with emergent collective computational abilities Proceedings of the National Academy of Sciences 1982 79 8 2554 2558 10.1073/pnas.79.8.2554 Hopfield JJ Tank DW Computing with neural circuits: A model Science 1986 233 4764 625 633 10.1126/science.3755256 3755256 Houkes W Zwart SD Transfer and templates in scientific modelling Studies in History and Philosophy of Science Part A 2019 77 October 93 100 10.1016/j.shpsa.2017.11.003 Howlett P Morgan MS How well do facts travel? The dissemination of reliable knowledge 2010 Cambridge University Press Humphreys P Computational models Philosophy of Science 2002 69 S3 S1 11 10.1086/341763 Humphreys P Extending ourselves: Computational science, empiricism, and scientific Method 2004 Oxford University Press Humphreys P Knowledge transfer across scientific disciplines Studies in History and Philosophy of Science Part A 2019 77 October 112 119 10.1016/j.shpsa.2017.11.001 Humphreys P Philosophical papers 2019 Oxford University Press Humphreys, P. (2023). Template transfer and model transfer. Synthese, Special Issue Transdisciplinary Model and Template Transfer. Ising E Beitrag zur Theorie des Ferromagnetismus Zeitschrift Für Physik 1925 31 1 253 258 10.1007/BF02980577 Jhun J Palacios P Weatherall JO Market crashes as critical phenomena? Explanation, idealization, and universality in econophysics Synthese 2018 195 4477 4505 10.1007/s11229-017-1415-y Kinzel K Narrative and evidence. How can case studies from the history of science support claims in the philosophy of science? Studies in History and Philosophy of Science Part A 2015 49 February 48 57 10.1016/j.shpsa.2014.12.001 Knuuttila, T., & Loettgers, A. (2011). The productive tension: Mechanisms vs. templates in modeling the phenomena. In Models, simulations, and representations. Routledge. Knuuttila T Loettgers A Magnets, spins, and neurons: The dissemination of model templates across disciplines The Monist 2014 97 3 280 300 10.5840/monist201497319 Knuuttila T Loettgers A Sauer T Scholl R Contrasting cases: The Lotka-Volterra model times three The philosophy of historical case studies. Boston Studies in the Philosophy and History of Science 2016 Springer International Publishing 151 78 Knuuttila T Loettgers A Model templates within and between disciplines: From magnets to gases—and socio-economic systems European Journal for Philosophy of Science 2016 6 3 377 400 10.1007/s13194-016-0145-1 Knuuttila T Loettgers A Modelling as indirect representation? The Lotka-Volterra model revisited The British Journal for the Philosophy of Science 2017 68 4 1007 1036 10.1093/bjps/axv055 Knuuttila T Loettgers A Holm S Magnetized memories: Analogies and templates in model transfer Philosophical perspectives on the engineering approach in biology Living machines? History and philosophy of biology 2020 1 Routledge Kuramoto Y International symposium on mathematical problems in theoretical physics Lecture Notes in Physics 1975 30 420 10.1007/BFb0013365 Kuramoto Y Araki H Self-entrainment of a population of coupled non-linear oscillators International symposium on mathematical problems in theoretical physics. Lecture notes in physics 1975 Springer 420 422 Lin C-H Knowledge transfer, templates, and the spillovers European Journal for Philosophy of Science 2022 12 1 6 10.1007/s13194-021-00426-w May RM Biological populations with nonoverlapping generations: Stable points, stable cycles, and chaos Science 1974 186 4164 645 647 10.1126/science.186.4164.645 4412202 Mezard M Parisi G Virasoro MA Spin glass theory and beyond: An introduction to the replica method and its applications 1987 World Scientific Publishing Company Panchenko D The Sherrington-Kirkpatrick model: An overview Journal of Statistical Physics 2012 149 2 362 383 10.1007/s10955-012-0586-7 Parisi G Spin glass theory Pysica 1986 140A 312 318 Pikovsky A Rosenblum M Kurths J Synchronization: A universal concept in nonlinear science American Journal of Physics 2002 70 6 655 655 10.1119/1.1475332 Price J The landing zone—Ground for model transfer in chemistry Studies in History and Philosophy of Science Part A 2019 77 October 21 28 10.1016/j.shpsa.2018.06.010 Sherrington D Kirkpatrick S Solvable model of a spin-glass Physical Review Letters 1975 35 26 1792 1796 10.1103/PhysRevLett.35.1792 Strogatz, Steven H. 1994. Nonlinear dynamics and chaos with student solutions manual: With applications to physics, biology, chemistry, and engineering, 2nd edn. CRC Press. Retrieved from https://books.google.at/books?hl=de&lr=&id=wUBvDwAAQBAJ&oi=fnd&pg=PT7&ots=ANx9XtjNsY&sig=aYxyFbKFI2hA8wGw_UDr80drjz8&redir_esc=y#v=onepage&q&f=false Strogatz SH Sync: The emerging science of spontaneous order 2003 1 Hyperion Stumpf MPH Wiuf C May RM Subnets of scale-free networks are not scale-free: Sampling properties of networks Proceedings of the National Academy of Sciences 2005 102 12 4221 4224 10.1073/pnas.0501179102 Watts DJ Strogatz SH Collective dynamics of ‘small-world’ networks Nature 1998 393 6684 440 442 10.1038/30918 9623998 Wiley DA Strogatz SH Girvan M The size of the sync basin Chaos: An Interdisciplinary Journal of Nonlinear Science 2006 16 1 015103 10.1063/1.2165594 Winfree AT Biological rhythms and the behavior of populations of coupled oscillators Journal of Theoretical Biology 1967 16 1 15 42 10.1016/0022-5193(67)90051-3 6035757 Winfree AT The geometry of biological time 2001 Springer Science & Business Media Zhang L Small M Judd K Exactly scale-free scale-free networks Physica A: Statistical Mechanics and Its Applications 2015 433 September 182 197 10.1016/j.physa.2015.03.074 Zuchowski L Modelling and knowledge transfer in complexity science Studies in History and Philosophy of Science Part A 2019 77 October 120 129 10.1016/j.shpsa.2017.10.003
PMC010xxxxxx/PMC10240329.txt
==== Front Health Equity Health Equity heq Health Equity 2473-1242 Mary Ann Liebert, Inc., publishers 140 Huguenot Street, 3rd Floor New Rochelle, NY 10801 USA 37284538 10.1089/heq.2023.0011 10.1089/heq.2023.0011 Special Collection: Improving Care for Veterans Through Health Equity Research (#5/14) Predicting Breast Cancer Risk for Women Veterans of African Ancestry in the Million Veteran Program Luoh, et al.; Health Equity 2023, 7.1 http://online.liebertpub.com/doi/10.1089/heq.2023.0011 Luoh Shiuh-Wen 1 2 * https://orcid.org/0000-0001-8355-5332 Minnier Jessica 1 2 3 Zhao Hongyu 4 Gao Lina 1 2 1 VA Portland Health Care System, Portland, Oregon, USA. 2 Knight Cancer Institute, Oregon Health & Science University, Portland, Oregon, USA. 3 OHSU-PSU School of Public Health, Portland, Oregon, USA. 4 Department of Biostatistics, Yale School of Public Health, VA Connecticut Health Care System, New Haven, Connecticut, USA. * Address correspondence to: Shiuh-Wen Luoh, MD, PhD, VA Portland Health Care System, Portland, OR 97239, USA. [email protected] [email protected] iORCID ID (https://orcid.org/0000-0001-8355-5332). 26 5 2023 May 2023 2023 26 5 2023 May 2023 7 1 303306 23 2 2023 Accepted February 23, 2023 © Shiuh-Wen Luoh et al., 2023; Published by Mary Ann Liebert, Inc. 2023 Shiuh-Wen Luoh et al., https://creativecommons.org/licenses/by/4.0/ This Open Access article is distributed under the terms of the Creative Commons License [CC-BY] (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Breast cancer is a leading cause of cancer and, therefore, a major health threat for women in the United States and worldwide. We have seen over the years major advances in breast cancer prevention and care. Breast cancer screening with mammography leads to reduction in breast cancer mortality, and breast cancer prevention treatment with antiestrogens results in reduction in breast cancer incidence. More progress, however, is urgently needed for this common cancer that affects 1 in 11 American women in their lifetime. Not all women have the same breast cancer risk. A personalized approach is highly desirable as women with higher breast cancer risk may benefit from more intense breast cancer screening and/or prevention intervention while lower risk women may be spared with the cost, inconvenience, and emotional burden of these procedures. In addition to age, demographics, family history, lifestyle, and personal health, genetics is an important determinant of an individual's risk for breast cancer. Over the past 10 years, advances in cancer genomics identified multiple common genetic variants from population studies that collectively can contribute significantly to an individual's breast cancer risk. The effects of these genetic variants can be summarized as a “polygenic risk score” (PRS). We are among the first groups to prospectively evaluate the performance of these risk prediction instruments among women veterans of the Million Veteran Program (MVP). A 313-variant PRS (PRS313) predicted incident breast cancer for a prospective cohort of European (EUR) ancestry women veterans with an area under the receiver operating characteristic curve (AUC) of 0.622. The PRS313 performed less well for AFR ancestry however, with an AUC of 0.579. This is not surprising as most genome-wide association studies were conducted in people of European ancestry. This is an important area of health disparity and unmet need. The large population size and diversity of the MVP provide a unique and important opportunity to explore novel approaches to produce accurate and clinically useful genetic risk prediction instruments for minority populations. breast cancer biomarkers risk prediction and African ancestry ==== Body pmcIntroduction Breast cancer is a leading cause of cancer death for American women. It is also a serious issue for women veterans. Our most recent Million Veteran Program (MVP) analysis found that out of 53.5K women veterans from the MVP without prior breast cancer diagnosis (mean age at entry 48.8 and 30.0% of African [AFR] ancestry), there were 818 new breast cancer diagnoses over a median follow-up of 5.7 years, translating into a breast cancer incidence of 2.70 per 1000 per year (unpublished). Breast cancer screening with mammography leads to reduction in mortality from breast cancer.1 However, the current breast cancer screening strategy is largely “one size fits all,” except for very high-risk women, such as those with BRCA mutations or prior chest wall irradiation. Appropriate risk assessment is highly desirable as higher risk women may benefit from more intense screening approaches, whereas lower risk women may do equally well with less intensity, therefore, avoiding the cost and inconvenience of screening. In addition, tamoxifen has been approved for women above the age of 35 years for breast cancer prevention if they have an elevated risk, as seen with an average 65-year-old Caucasian woman.2,3 There are a growing number of characteristics that have been validated to predict a woman's risk of developing breast cancer. This includes age; mutation status for breast cancer genes, such as BRCA1 and BRCA2; family history of cancer; personal history of benign breast biopsy; history of benign breast disease, such as atypia; exposure to hormones; breast density as seen on mammography; and more recently, background parenchymal enhancement on MRI.4 Harnessing these features may enable us to more precisely assess the risk of breast cancer development. Understanding individuals' risks for developing breast cancer may allow us to adopt a more appropriate breast cancer screening and/or prevention strategy. Recent successes from genome-wide association studies (GWAS) of breast cancer (e.g., A study with more than 150,000 breast cancer cases and more than 110,000 controls5) have led to a growing literature on the development and evaluation of breast cancer risk prediction models, for example Refs.6–8 Results In a prospective cohort of 35,130 women veterans without history of breast cancer (median age 49 years and median follow-up 3.9 years) in the MVP,9 we evaluated the performance of clinical breast cancer risk prediction models that comprised demographics, lifestyle, personal health, environmental exposure, and a genetic risk model, 313-variant polygenic risk score (PRS313).10 The performance of PRS313 alone was assessed. Clinical risk models tested included a literature review and a Breast Cancer Risk Assessment Tool,11,12 implemented with or without PRS313. Thirty one percent of this cohort were of non-Hispanic AFR ancestry. Individualized Coherent Absolute Risk Estimator11 literature review12,13 in combination with PRS313 had an area under the receiver operating characteristic curve (AUC) of 0.708 (95% confidence interval, 0.659–0.758) in women with European or non-AFR ancestries and 0.625 (0.539–0.711) in AFR women.9 Breast Cancer Risk Assessment Tool with PRS313 had an AUC of 0.695 (0.662–0.729) in European or non-AFR women and 0.675 (0.626–0.723) in AFR women.9 PRS313 alone without clinical or demographic predictors yielded an AUC of 0.622 (0.580–0.664) in women of European descent and a lower AUC of 0.579 (0.522–0.636) in AFR women. Incorporation of PRS313 with clinical models improved prediction in European, but not in AFR women.9 Models estimated up to 9% of European and 18% of AFR women with absolute lifetime risk >20%. Women with a lifetime breast cancer risk of >20% are generally considered to be high risk and may be appropriate for clinical studies that would examine whether these women would benefit from more intense breast cancer screening. Discussion We found that PRS313 underperformed in a prospective cohort of women veterans of AFR ancestry in the MVP. This is consistent with prior cross section or case–control studies.9,14–16 This is a significant area of unmet need as African Americans have higher risk of developing early-onset breast cancer and about 40% higher breast cancer mortality than other ancestral groups in the United States.17 PRS has shown great promise in improving biomedical outcomes through precision medicine. However, currently available PRS tools are much more accurate in individuals of European descent, as other ancestral groups were under-represented and understudied. In a transethnic breast cancer study,18 the authors performed AFR ancestry GWAS meta-analysis (9241 cases and 10,193 controls), and then meta-analyzed with European ancestry GWAS data (122,977 cases and 105,974 controls) from the Breast Cancer Association Consortium. There was a 13-fold difference between the number of breast cancer subjects between European and AFR ancestries. As a result of sample size and possibly different genetic architecture between the two populations, only six loci were identified from individuals of African ancestry, even after trans-ethnic analysis, in comparison with more than 200 loci reported from individuals of European ancestry. The differential performance of PRS across ancestral groups raises a significant concern about potentials to exacerbate health disparities and evokes ethical controversy surrounding the clinical implementation of PRS in general. It is, therefore, essential to include participants representing diverse populations in genomic medicine studies to ensure equitable benefit from scientific discoveries and to prevent further increase in health disparities. To circumvent this, we perform combined analysis with additional cases and controls from the MVP to increase the gain in statistical power for novel locus discoveries in different ancestral groups. Although the most straightforward solution to increase PRS performance across ancestral groups is to build well-powered GWAS resources for diverse populations, we will entertain an alternative approach by taking advantage of the observation that genetic data from different ancestral groups still share substantial information, although with differences in allele frequency, linkage disequilibrium structure, and genetic architecture. Recently, models have been developed to borrow information from European ancestral groups to help boost the power of predictions in non-European ancestral groups. We plan to conduct a comprehensive simulation and real data analyses in the MVP to evaluate these newly developed cross-population prediction models, including XPASS19 and PRScsx20 (that have been published), MePred and SDPRX (that are being developed by MVP researchers),21 and other methods. We believe this innovative comprehensive evaluation will offer both insights and guidance to improve risk stratifications for both European and non-European populations, expand our knowledge of disease mechanisms, and benefit disease screening and early prevention strategies. Authors Contribution All authors have contributed to the writing, editing, and approval of this publication. Disclaimer The opinions expressed in this article do not reflect those of the VA or U.S. government. Author Disclosure Statement No competing financial interests exist. Funding Information VA Office of Research and Development Merit Review 1I01BX004188-01. Cite this article as: Luoh S-W, Minnier J, Zhao H, Gao L (2023) Predicting breast cancer risk for women veterans of African ancestry in the Million Veteran Program, Health Equity 7:1, 303–306, DOI: 10.1089/heq.2023.0011. Abbreviations Used AFR African AUC area under the receiver operating characteristic curve GWAS genome-wide association studies MVP Million Veteran Program PRS polygenic risk score PRS313 313-variant PRS ==== Refs References 1. Harris R, Yeatts J, Kinsinger L. Breast cancer screening for women ages 50 to 69 years a systematic review of observational evidence. Prev Med 2011;53 (3 ):108–114; doi: 10.1016/j.ypmed.2011.07.004 21820465 2. Fisher B, Costantino JP, Wickerham DL, et al. Tamoxifen for the prevention of breast cancer: Current status of the National Surgical Adjuvant Breast and Bowel Project P-1 study. J Natl Cancer Inst 2005;97 (22 ):1652–1662.16288118 3. Cuzick J, Forbes J, Edwards R, et al. First results from the International Breast Cancer Intervention Study (IBIS-I): A randomised prevention trial. Lancet 2002;360 (9336 ):817–824.12243915 4. Kuhl CK. Predict, then act: Moving toward tailored prevention. J Clin Oncol 2019;37 (12 ):943–945.30844319 5. Zhang H, Ahearn TU, Lecarpentier J, et al. Genome-wide association study identifies 32 novel breast cancer susceptibility loci from overall and subtype-specific analyses. Nat Genet 2020;52 (6 ):572–581.32424353 6. Jia G, Lu Y, Wen W, et al. Evaluating the utility of polygenic risk scores in identifying high-risk individuals for eight common cancers. JNCI Cancer Spectr 2020;4 (3 ); doi: 10.1093/jncics/pkaa021 7. Yanes T, Young M-A, Meiser B, et al. Clinical applications of polygenic breast cancer risk: A critical review and perspectives of an emerging field. Breast Cancer Res 2020;22 (1 ); doi: 10.1186/s13058-020-01260-3 8. Houlahan KE. Do breast cancer risk scores work for you? J Natl Cancer Inst 2021;113 (9 ):1118–1119.33769538 9. Minnier J, Rajeevan N, Gao L, et al. Polygenic breast cancer risk for women veterans in the million veteran program. JCO Prec Oncol 2021;(5 ):1178–1191; doi: 10.1200/po.20.00541 10. Mavaddat N, Michailidou K, Dennis J, et al. Polygenic risk scores for prediction of breast cancer and breast cancer subtypes. Am J Hum Genet 2019;104 (1 ):21–34.30554720 11. Choudhury PP, Maas P, Wilcox A, et al. iCARE: An R package to build, validate and apply absolute risk models. PLoS One 2020;15 (2 ):e0228198; doi: 10.1371/journal.pone.0228198 32023287 12. Choudhury PP, Wilcox AN, Brook MN, et al. Comparative validation of breast cancer risk prediction models and projections for future risk stratification. J Natl Cancer Inst 2020;112 (3 ):278–285; doi: 10.1093/jnci/djz113 31165158 13. Maas P, Barrdahl M, Joshi AD, et al. Breast cancer risk from modifiable and nonmodifiable risk factors among white women in the United States. JAMA Oncol 2016;2 (10 ):1295–1302.27228256 14. Shieh Y, Fejerman L, Lott PC, et al. A polygenic risk score for breast cancer in US Latinas and Latin American women. J Natl Cancer Inst 2020;112 (6 ):590–598; doi: 10.1093/jnci/djz174 31553449 15. Du Z, Gao G, Adedokun B, et al. Evaluating polygenic risk scores for breast cancer in women of African Ancestry. J Natl Cancer Inst 2021;113 (9 ):1168–1176.33769540 16. Liu C, Zeinomar N, Chung WK, et al. Generalizability of polygenic risk scores for breast cancer among women with European, African, and Latinx Ancestry. JAMA Netw Open 2021;4 (8 ):e2119084.34347061 17. DeSantis CE, Ma J, Gaudet MM, et al. Breast cancer statistics, 2019. CA Cancer J Clin 2019;69 (6 ):438–451.31577379 18. Adedokun B, Du Z, Gao G, et al. Cross-ancestry GWAS meta-analysis identifies six breast cancer loci in African and European ancestry women. Nat Commun 2021;12 (1 ):4198–4205.34234117 19. Cai M, Xiao J, Zhang S, et al. A unified framework for cross-population trait prediction by leveraging the genetic correlation of polygenic traits. Am J Hum Genet 2021;108 (4 ):632–655.33770506 20. Ruan Y, Lin Y-F, Feng Y-CA, et al. Improving polygenic prediction in ancestrally diverse populations. Nat Genet 2022;54 (5 ):573–580.35513724 21. Zhou G, Chen T, Zhao H. SDPRX: A statistical method for cross-population prediction of complex traits. Am J Hum Genet 2023;110 (1 ):13–22.36460009
PMC010xxxxxx/PMC10241234.txt
==== Front Endocrinol Diabetes Metab Case Rep Endocrinol Diabetes Metab Case Rep EDM Endocrinology, Diabetes & Metabolism Case Reports 2052-0573 Bioscientifica Ltd Bristol 37042492 10.1530/EDM-22-0362 EDM220362 Adolescent/young adult Male Asian - Japanese Japan Pancreas Diabetes Unique/Unexpected Symptoms or Presentations of a Disease Unique/Unexpected Symptoms or Presentations of a Disease A family with type A insulin resistance syndrome caused by a novel insulin receptor mutation O Horikawa and others Horikawa Osamu 1 http://orcid.org/0000-0001-5848-0666 Ugi Satoshi 12 Takayoshi Tomofumi 3 Omura Yasushi 4 Yonishi Maya 1 Sato Daisuke 1 Fujita Yukihiro 1 Fuke Tomoya 5 Hirota Yushi 3 Ogawa Wataru 3 Maegawa Hiroshi 1 1 Department of Medicine, Shiga University of Medical Science, Otsu, Shiga, Japan 2 Department of Medicine, Omihachiman Community Medical Center, Omihachiman, Shiga, Japan 3 Division of Diabetes and Endocrinology, Kobe University Graduate School of Medicine, Kobe, Japan 4 Department of Internal Medicine, Kohka Public Hospital, Kohka, Shiga, Japan 5 Department of Medicine, Saiseikai Shiga Hospital, Ritto, Shiga, Japan Correspondence should be addressed to S Ugi; Email: [email protected] 16 3 2023 01 4 2023 2023 2 22-036213 9 2022 16 3 2023 © the author(s) 2023 the author(s) https://creativecommons.org/licenses/by/4.0/ This work is licensed under a Creative Commons Attribution 4.0 International License. Summary A 17-year-old boy was referred to our endocrinology clinic for a clinical investigation of hyperinsulinemia. An oral glucose tolerance test showed plasma glucose concentrations in the normal range. However, insulin concentrations were considerably elevated (0 min: 71 μU/mL; 60 min: 953 μU/mL), suggesting severe insulin resistance. An insulin tolerance test confirmed that he had insulin resistance. There was no apparent hormonal or metabolic cause, including obesity. The patient had no outward features of hyperinsulinemia, including acanthosis nigricans or hirsutism. However, his mother and grandfather also had hyperinsulinemia. Genetic testing showed that the patient (proband), his mother, and his grandfather had a novel p.Val1086del heterozygous mutation in exon 17 of the insulin receptor gene (INSR). Although all three family members have the same mutation, their clinical courses have been different. The onset of the mother’s diabetes was estimated at 50 years, whereas the grandfather developed diabetes at 77 years. Learning points Type A insulin resistance syndrome is caused by mutations in the insulin receptor (INSR) gene and results in severe insulin resistance. Genetic evaluation should be considered in adolescents or young adults with dysglycemia when an atypical phenotype, such as severe insulin resistance, or a relevant family history is observed. Clinical courses may differ even if the same genetic mutation is found in a family. Patient Demographics Adolescent/young adult Male Asian - Japanese Japan Clinical Overview Pancreas Diabetes Publication Details Unique/unexpected symptoms or presentations of a disease April 2023 ==== Body pmcBackground Type A insulin resistance syndrome (IRS) is caused by mutations in the insulin receptor (INSR) gene (1, 2). IRS is a rare disorder, but recent findings suggest that at least 0.05% of the population harbors a pathological INSR mutation (3). Because this disorder has a mild phenotype and is not well understood, many of these patients are misdiagnosed and incorrectly treated. We report a family with type A IRS caused by a novel mutation in INSR. The trigger for the diagnosis of the 17-year-old boy in the family was a urinalysis at school. Genetic testing showed that his mother and grandfather had the same mutation, but their clinical courses were different. Case presentation A 17-year-old boy was referred to our endocrinology clinic for investigation of hyperinsulinemia. Glycosuria had been noted during a health checkup at high school, and the boy was found to have extreme hyperinsulinemia with a normal glucose concentration. His height, body weight, and body mass index were 165 cm, 52.6 kg, and 19.3 kg/m2, respectively (Table 1). He was healthy. He had no outward features of hyperinsulinemia, including acanthosis nigricans or hirsutism, a normal birth weight of 3000 g, and no mental retardation. He had a family history of diabetes in relatives on his mother’s side. Table 1 Clinical characteristics of the family. Characteristic Proband Mother Grandfather Age (years) 17 50 87 Height (cm) 165 156 165 Body weight (kg) 52.6 56 45.3 Body mass index (kg/m2) 19.3 23 16.6 FPG (mg/dL) 75 107 117 Insulin (μU/mL) 50 37 125.3* C-peptide (ng/mL) 2.08 2.62 7.6 Proinsulin (pmol/L) 5.4 ND ND HbA1c (%) 5.6 6.5 6.4 Triglycerides (mg/dL) 87 114 109 HDL cholesterol (mg/dL) 70 72 43 LDL cholesterol (mg/dL) 82 125 96 Anti-IR antibody None ND ND Birth weight (g) 3000 3900 Unknown Acanthosis nigricans None None None Hirsutism None None None Diabetes treatment None None DPP4I Estimated diabetes onset No diabetes 50 years 77 years *Postprandial value. Anti-IR, anti-insulin receptor; BMI, body mass index; DPP4I, dipeptidyl peptidase 4 inhibitor; FPG, fasting plasma glucose; HbA1c, hemoglobin A1c; HDL, high-density lipoprotein; LDL, low-density lipoprotein; ND, not determined. Investigation The patient’s fasting glucose concentration was 75 mg/dL and his hemoglobin A1c (HbA1c) level was 5.6%. Plasma concentrations of insulin, C-peptide, and proinsulin were 50 μU/mL, 2.08 ng/mL, and 5.4 pmol/L, respectively (Table 1). The molar ratios of insulin–C peptide and proinsulin–insulin were 0.44 (normal: 0.1–0.2) and 0.02 (normal: 0.04–0.23), respectively. An oral glucose tolerance test (OGTT) was performed (Table 2). Plasma glucose concentrations were in the normal range. However, insulin concentrations were considerably elevated (0 min: 71 μU/mL; 60 min: 953 μU/mL), suggesting severe insulin resistance. We performed the insulin tolerance test (Table 3). Blood samples were collected before and 3, 6, 9, 12, and 15 min after intravenous injection of regular insulin (0.1 U/kg). The plasma glucose disappearance rate (KITT) was calculated from the linear slope of the plasma glucose concentration curve between 3 and 15 min, as described previously (4). KITT has been reported to be strongly correlated with the glucose infusion rate in a glucose clamp study and has been used for the assessment of insulin resistance (4, 5). The calculated KITT value was 1.91%/min, which was considered to be insulin resistance (mean values of normal, obese, and diabetic subjects are 5.65 ± 0.35, 4.14 ± 0.52, and 2.73 ± 0.29, respectively) (4). Anti-insulin receptor antibody was negative. There was no apparent hormonal or metabolic cause, including obesity. Table 2 OGTT results of the family members. Time (min) 0 30 60 90 120 Glucose (mg/dL)  Proband 77 198 193 123 125  Mother 107 221 253 ND 281  Father 98 155 138 ND 91  Brother 94 174 195 143 119 Insulin (μU/mL)  Proband 71 664 953 840 708  Mother 37 122 188 ND 349  Father 2.6 42.5 30 ND 8.1  Brother 5.4 36.9 78.6 45.2 47.5 ND, not determined; OGTT, oral glucose tolerance test. Table 3 ITT results of the proband. Time (min) 0 3 6 9 12 15 Glucose (mg/dL) 74 73 74 73 69 63 KITT (%/min) 1.91 ITT, insulin tolerance test; KITT, plasma glucose disappearance rate. These findings suggested a genetic cause of insulin resistance (6). Therefore, an OGTT was also performed on the patient’s parents and brother (Table 2). His father and brother had normal glucose tolerance and insulin concentrations. However, his mother reached the diagnostic criteria for diabetes and had high insulin concentrations. Her HbA1c level was 6.5%. These findings suggested that the patient had inherited a genetic cause of insulin resistance from his mother. Consistent with this possibility, four of nine of his maternal grandfather’s siblings and his grandfather had diabetes (Fig. 1). The patient’s grandfather was 87 years old. His diabetes was being treated at another clinic by using a dipeptidyl peptidase 4 inhibitor. We found that he also had hyperinsulinemia, with a fasting glucose concentration of 230 mg/dL and a postprandial insulin level of 125 μU/mL. The clinical characteristics of the patient’s mother and grandfather are shown in Table 1. After analysis of these data, type A IRS was highly suspected. Figure 1 Pedigree of the family. The arrow shows the proband. The gray square and gray circle indicate family members found to have the mutation, and the numbers in the figure represent age in years. We received the consent of the patient (proband), his parents, and his grandfather to perform genetic testing. This research was approved by the ethics committee of the Kobe Graduate School of Medicine (approval no. 170105). We sequenced all 22 exons of the insulin receptor gene (INSR, Gene Bank NM_000208.4). Sequencing of the genomic DNA of the patient showed a heterozygous deletion mutation in the 3′ end of exon 17. The specific mutation was c.3256_3258delGTG, p.Val1086del or a deletion of three base pairs in the intron just following them, c.3256+1‗3258+3delGTG (Fig. 2A). The same mutation was identified in the patient’s mother and grandfather, but not in his father who did not have hyperinsulinemia. (Fig. 2A, C and D). We could not distinguish whether the mutation was c.3256_3258delGTG or dc.3256+1‗3258+3delGTG because both are GTG mutations. Therefore, we sequenced the cDNA of the patient and his mother and identified the mutation as p.Val1086del in exon 17 (Fig. 2E and F), which is a novel INSR mutation. Figure 2 Sequencing results of genomic DNA (A–D) and cDNA (E, F) of the insulin receptor gene (INSR). The proband, his mother, and his grandfather have a heterozygous GTG deletion at the 3′ end of exon 17. Sequencing of genomic DNA could not distinguish between this mutation and a GTG deletion next to it at the 5′ end of the intron. Sequencing of cDNA confirmed the mutation as p.Val1086del in exon 17. Treatment The patient did not have diabetes when the genetic test results were confirmed. We planned to follow up twice a year. We advised the patient to maintain a healthy lifestyle and a body mass index within the normal weight range. Six months later, the HbA1c level of the patient’s mother was elevated (6.8%), and we started to administer 250 mg of metformin daily. Outcome and follow-up The HbA1c level of the patient was not elevated 2 years later. Treatment of the patient’s mother with 250 mg of metformin daily was not effective (7.0% at 6 months after starting treatment) and was increased to 500 mg daily. The HbA1c level dropped to 6.5% at 3 months after this increase. The HbA1c level of the grandfather was stable, and information on his clinical course to date was sought to predict the clinical course of his daughter and grandson (the proband). We found that his HbA1c level had been 5.1, 6.6, and 6.9% at 67, 77, and 87 years, respectively. He had begun taking a dipeptidyl peptidase 4 inhibitor at 77 years of age and was still being treated with the same regimen. We estimated that his onset of diabetes was at 77 years of age (Table 1). His HbA1c levels have never exceeded 7.1%. Discussion We describe a family with type A IRS. A 17-year-old boy (the proband) presented with severe insulin resistance possibly caused by a novel deletion mutation of the B subunit of INSR. The trigger for the diagnosis of the proband was glycosuria, which was found at a health checkup at high school. The patient’s mother and grandfather had the same genetic mutation, and we studied this family (spanning three generations) to better understand the natural history of this syndrome. Mutations of the INSR gene result in extreme insulin resistance and dysglycemia (7) because of the dysfunction of INSR. Disorders associated with mutations of INSR have various phenotypes. Rabson–Mendenhall syndrome (RMS) and Donohue syndrome have the most severe insulin resistance, whereas type A IRS displays milder manifestations (1, 8). Homozygous or compound heterozygous mutations are thought to cause severe syndromes (RMS and Donohue syndrome) (9). Heterozygous mutations cause insulin resistance by a dominant negative effect but should lead to the formation of some fully functional wt/wt receptors, which would result in a less severe phenotype (9). Mutations causing type A IRS rather than RMS or Donohue syndrome are more frequently found in the tyrosine kinase (TK) domain (9). Some mutations in the TK domain result in decreased tyrosine activity of INSR (10, 11). The deletion mutation Val1086del found in the family of our patient was a novel mutation. Although we cannot conclude that this mutation was definitely the cause of the insulin resistance in this family, Val1086 is in a TK domain near the cluster of tyrosine phosphorylation sites at positions 1158, 1162, and 1163. Therefore, this deletion mutation may lead to decreased TK activity of INSR. However, a functional analysis is required to prove how the mutation affects the function of INSR. Type A IRS is considered to be rare. However, a nationwide survey conducted in Japan from 2014 to 2016 identified 23 cases of type A IRS and 10 cases of RMS/Donohue syndrome (3). On the basis of a genetic analysis of these cases using the Hardy–Weinberg principle, we estimate that at least 0.05% of the population might harbor a pathological INSR mutation in one allele (3). Owing to the mild phenotype of type A IRS and a lack of understanding of this disorder, many of these patients may be misdiagnosed and incorrectly treated. According to a nationwide survey in Japan, 39% of individuals with type A IRS were identified by urinalysis at school, and it was the largest trigger for diagnosis (3). New classification and diagnostic criteria for IRS have been recently published on the basis of the results of a nationwide survey conducted in Japan (6). According to this survey, the major feature of IRS is the presence of hyperinsulinemia (fasting serum insulin concentration of >30 μU/mL) with no apparent cause of insulin resistance, such as obesity or other conditions. Our patient and his mother met those criteria. Clinicians and members of the public need to be aware of the possibility of such genetic syndromes of insulin resistance. We examined the phenotypes of the patient’s mother and grandfather to gain an understanding of the natural course of type A IRS. The OGTT of the proband’s mother showed hyperinsulinemia, but it was attenuated compared with that of her son (Table 2). The continuation of hyperinsulinemia leads to the failure of β-cell function, which suggests that decreased insulin secretion in the patient’s mother resulted in an abnormal HbA1c level of 6.5%. To support this conclusion, some patients with INSR variants require high-dose insulin therapy in the later stages of the disease process (12). Another patient showed a tendency for a progressive increase in post-load glucose concentrations and a decrease in insulin concentrations during an OGTT over 8 years (13). The onset of diabetes in the proband’s grandfather was estimated to be 77 years, and this is a later onset of diabetes than that in his daughter. Several reports showed that the phenotype, such as the severity of diabetes or insulin resistance with or without acanthosis nigricans, was different within the family, despite them having the same mutation (10, 14, 15). The exact reasons for this finding are unknown. Other genetic and/or environmental factors, such as age and sex, may contribute to the different phenotypes. Additionally, an unknown abnormality in the insulin signaling pathway might have been inherited from the patient’s grandmother. However, we speculate that changes in Japanese lifestyles over the past century affected the diabetes onset time in these two individuals. At the time of writing, the proband does not have diabetes, but he may develop diabetes earlier than his mother because lifestyles are still changing in Japan. Therefore, follow-up is important so that treatment can be started promptly. Moreover, maintaining a healthy lifestyle and a body mass index within the normal weight range is important. The patient’s mother and grandfather have the same mutation, but their diabetes onset times were different, suggesting that lifestyle factors can affect the clinical course. We chose metformin for the treatment of the proband’s mother. This choice was made because the inhibition of hepatic gluconeogenesis by insulin-independent multiple molecular mechanisms is thought to be largely responsible for the glucose-lowering effect of metformin (16). Metformin was the most prescribed medication in a nationwide survey in Japan (3). In this previous study, among 23 cases of identified type A IRS, 13 (57%) patients were treated with metformin, followed by insulin in 8 (35%) patients. In fact, the effectiveness of metformin for the treatment of type A IRS has been reported (17, 18). However, no established treatment strategy, including metformin, has been established for type A IRS. Long-term observation of the patient’s mother needs to be performed. In conclusion, we describe a family with type A IRS possibly caused by a novel mutation in the B subunit of INSR. The prevalence of type A IRS might be more frequent than previously thought. Therefore, we need to be aware of the possibility of such genetic syndromes of insulin resistance. This case report suggests that the clinical course of type A IRS can be affected by lifestyle factors. Declaration of interest The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of this case report. Funding This study did not receive any specific grant from any funding agency in the public, commercial or not-for-profit sector. Patient consent Written informed consent for publication of their clinical details was obtained from the patient, his parents and his grandfather for publication of the case report. Author contribution statement OH and SU prepared the manuscript. OH, MY, DS and YF were endocrinologists involved in the clinical care of the patient. YO was an endocrinologist involved in the clinical care of the grandfather. FT was the endocrinologist who identified the hyperinsulinemia of the proband. TT and YH performed the genetic analysis. YH, Professor HM and Professor WO critically reviewed the manuscript. All authors agree to be accountable for all aspects of the work and each contributed considerably to the work. Acknowledgements The authors thank Carol Wilson, PhD from Edanz (https://jp.edanz.com/ac) for editing a draft of this manuscript. ==== Refs References 1 Semple RK Savage DB Cochran EK Gorden P & O'Rahilly S . Genetic syndromes of severe insulin resistance. Endocrine Reviews 2011 32 498–514. (10.1210/er.2010-0020)21536711 2 Riddle MC Philipson LH Rich SS Carlsson A Franks PW Greeley SAW Nolan JJ Pearson ER Zeitler PS & Hattersley AT . Monogenic diabetes: from genetic insights to population-based precision in care. Reflections from a 2020. Diabetes Care 2020 43 3117–3128. (10.2337/dci20-0065)33560999 3 Takeuchi T Ishigaki Y Hirota Y Hasegawa Y Yorifuji T Kadowaki H Akamizu T Ogawa W & Katagiri H . Clinical characteristics of insulin resistance syndromes: a nationwide survey in Japan. Journal of Diabetes Investigation 2020 11 603–616. (10.1111/jdi.13171)31677333 4 Bonora E Moghetti P Zancanaro C Cigolini M Querena M Cacciatori V Corgnati A & Muggeo M . Estimates of in vivo insulin action in man: comparison of insulin tolerance tests with euglycemic and hyperglycemic glucose clamp studies. Journal of Clinical Endocrinology and Metabolism 1989 68 374–378. (10.1210/jcem-68-2-374)2645308 5 Okita K Iwahashi H Kozawa J Okauchi Y Funahashi T Imagawa A & Shimomura I . Usefulness of the insulin tolerance test in patients with type 2 diabetes receiving insulin therapy. Journal of Diabetes Investigation 2014 5 305–312. (10.1111/jdi.12143)24843779 6 Ogawa W Araki E Ishigaki Y Hirota Y Maegawa H Yamauchi T Yorifuji T & Katagiri H . New classification and diagnostic criteria for insulin resistance syndrome. Diabetology International 2022 13 337–343. (10.1007/s13340-022-00570-5)35463863 7 Kahn CR Flier JS Bar RS Archer JA Gorden P Martin MM & Roth J . The syndromes of insulin resistance and acanthosis nigricans. Insulin-receptor disorders in man. New England Journal of Medicine 1976 294 739–745. (10.1056/NEJM197604012941401)176581 8 Musso C Cochran E Moran SA Skarulis MC Oral EA Taylor S & Gordon P . Clinical course of genetic diseases of the insulin receptor (type A and Rabson-Mendenhall syndromes): a 30-year prospective. Medicine (Baltimore) 2004 83 209–222. (10.1097/01.md.0000133625.73570.54)15232309 9 Hosoe J Kadowaki H Miya F Aizu K Kawamura T Miyata I Satomura K Ito T Hara K Tanaka M , Structural basis and genotype-phenotype correlations of INSR mutations causing severe insulin resistance. Diabetes 2017 66 2713–2723. (10.2337/db17-0301)28765322 10 Haruta T Takata Y Iwanishi M Maegawa H Imamura T Egawa K Itazu T & Kobayashi M . Ala1048-->Asp mutation in the kinase domain of insulin receptor causes defective kinase activity and insulin resistance. Diabetes 1993 42 1837–1844. (10.2337/diab.42.12.1837)8243830 11 Taira M Taira M Hashimoto N Shimada F Suzuki Y Kanatsuka A Nakamura F Ebina Y Tatibana M Makino H , Human diabetes associated with a deletion of the tyrosine kinase domain of the insulin receptor. Science 1989 245 63–66. (10.1126/science.2544997)2544997 12 Wei C & Burren CP . Diagnostic and management challenges from childhood, puberty through to transition in severe insulin resistance due to insulin receptor mutations. Pediatric Diabetes 2017 18 835–838. (10.1111/pedi.12486)28093873 13 Awata T Matsumoto C Momomura K Takahashi Y Odawara M Kasuga M Kadowaki T & Iwamoto Y . A 3-basepair in-frame deletion (delta Leu999) in exon 17 of the insulin receptor gene in a family with insulin resistance. Journal of Clinical Endocrinology and Metabolism 1994 79 1840–1844. (10.1210/jcem.79.6.7989492)7989492 14 Takahashi I Yamada Y Kadowaki H Horikoshi M Kadowaki T Narita T Tsuchida S Noguchi A Koizumi A & Takahashi T . Phenotypical variety of insulin resistance in a family with a novel mutation of the insulin receptor gene. Endocrine Journal 2010 57 509–516. (10.1507/endocrj.k09e-339)20339196 15 Lin L Chen C Fang T Chen D Chen K & Quan H . Type A insulin resistance syndrome misdiagnosed as polycystic ovary syndrome: a case report. Journal of Medical Case Reports 2019 13 347. (10.1186/s13256-019-2304-4) 16 Rena G Hardie DG & Pearson ER . The mechanisms of action of metformin. Diabetologia 2017 60 1577–1585. (10.1007/s00125-017-4342-z)28776086 17 Sanderson EE Shah M Hooper AJ Bell DA & Choong CS . Monogenic diabetes due to an INSR mutation in a child with severe insulin resistance. Endocrinology, Diabetes and Metabolism Case Reports 2022 21–0114. (10.1530/EDM-21-0114) 18 Rique S Ibáñez L Marcos MV Carrascosa A & Potau N . Effects of metformin on androgens and insulin concentrations in type A insulin resistance syndrome. Diabetologia 2000 43 385–386. (10.1007/s001250050059)10768102