Received: 10 July 2019 Revised: 9 September 2019 Accepted: 7 October 2019 DOI: 10.1002/osp4.381 ORIGINAL ARTICLE Differential association between S100A4 levels and insulin resistance in prepubertal children and adult subjects with clinically severe obesity Siri D. Taxerås1 | María Galán2,3 | Laura Campderros4,5 | Irene Piquer‐Garcia1 | Silvia Pellitero1,6 | Eva Martínez1 | Rocío Puig1 | Icíar Lucena1 | Jordi Tarascó7 | Pau Moreno7 | José Balibrea8 | Joan Bel9 | Marta Murillo9 | María Martínez9 | Marta Ramon‐Krauel10,11 | Manel Puig‐Domingo1,6 | Francesc Villarroya4,5 | Carles Lerin10,11 | David Sánchez‐Infantes1,5 1 Department of Endocrinology and Nutrition, Germans Trias i Pujol Research Institute, Barcelona, Spain 2 Institut de Recerca del Hospital de la Santa Creu i Sant Pau‐Programa ICCC, Barcelona, Spain 3 CIBER de Enfermedades Cardiovasculares (CIBERCV), ISCIII, Madrid, Spain 4 Department of Biochemistry and Molecular Biology, Institute of Biomedicine, University of Barcelona, Barcelona, Spain 5 Biomedical Research Center (Red Fisiopatología de la Obesidad y Nutrición) (CIBEROBN), ISCIII, Madrid, Spain 6 Biomedical Research Center (Red Fisiopatología de la Diabetes y enfermedades metabólicas) (CIBERDEM), ISCIII, Madrid, Spain 7 Department of Surgery, Germans Trias i Pujol Research Institute, Barcelona, Spain 8 Metabolic and Bariatric Surgery Unit, EAC‐BS Center of Excellence, Vall d'Hebron University Hospital, Barcelona, Spain 9 Department of Pediatric, Germans Trias i Pujol Research Institute, Barcelona, Spain 10 Endocrinology Department, Institut de Recerca Sant Joan de Déu, Barcelona, Spain 11 Hospital Sant Joan de Déu, Barcelona, Spain Correspondence David Sánchez‐Infantes, PhD, Germans Trias i Summary Pujol Research Institute (IGTP) Campus Can Objectives: S100A4 has been recently identified as an adipokine associated with insulin Ruti, Carretera de Can Ruti, Camí de les Escoles s/n 08916 Badalona, Barcelona, Spain. resistance (IR) in adult subjects with obesity. However, no data about its levels in children Email: dsanchez@igtp.cat Carles Lerin, PhD, with obesity and only a few approaches regarding its potential mechanism of action have Endocrinology Department, Institut de Recerca Sant Joan de Déu, Barcelona, Spain, Hospital been reported. To obtain a deeper understanding of the role of S100A4 in obesity, (a) Sant Joan de Déu, Barcelona, Spain. 08950 S100A4 levels were measured in prepubertal children and adult subjects with and without Esplugues, Barcelona, Spain. Email: clerin@fsjd.org obesity and studied the relationship with IR and (b) the effects of S100A4 in cultured human adipocytes and vascular smooth muscle cells (VSMCs) were determined. Funding information Instituto de Salud Carlos III, Grant/Award Methods: Sixty‐five children (50 with obesity, age 9.0 ±1.1 years and 15 normal Numbers: CP15/00106, PI17/01455 and weight, age 8.4 ±0.8 years) and fifty‐nine adults (43 with severe obesity, age 46 FIS17/01455; Norges Teknisk‐ Siri D. Taxerås and María Galán contributed equally to this work. Note. Multivariate linear regression was used to assess potential associations between log‐transformed S100A4 levels and physiologic variables. Model was adjusted for sex, age, and BMI‐SDS for children or BMI for adults. βstd: standardized beta coefficient from the regression model. Bold font indicates P < .05. Abbreviations: HbA1c, glycated haemoglobin; HMW‐Adp, high molecular weight adiponectin; HOMA‐IR, homeostatic model assessment‐insulin resistance. -------------------------------------------------------------------------------------------------------------------------------- This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2019 The Authors. Obesity Science & Practice published by World Obesity and The Obesity Society and John Wiley & Sons Ltd Obes Sci Pract. 2020;6:99–106. wileyonlinelibrary.com/journal/osp4 99 100 TAXERÅS ET AL. Naturvitenskapelige Universitet, Grant/Award Number: K‐sted 651010‐Project number ±11 years and 16 normal weight, age 45 ±9 years) were included. Blood from children 70084900‐analyse numb; Norwegian Univer- and adults and adipose tissue samples from adults were obtained and analysed. sity of Science and Technology, Grant/Award Number: K‐sted 651010‐project number Human adipocytes and VSMC were incubated with S100A4 to evaluate their 70084900‐analyse number M50676; Fondo response to this adipokine. Europeo de Desarrollo Regional (FEDER) Results: Circulating S100A4 levels were increased in both children (P = .002) and adults (P < .001) with obesity compared with their normal‐weight controls. In subjects with obesity, S100A4 levels were associated with homeostatic model assessment‐ insulin resistance (HOMA‐IR) in adults (βstd = .42, P = .008) but not in children (βstd = .12, P = .356). Human adipocytes were not sensitive to S100A4, while incubation with this adipokine significantly reduced inflammatory markers in VSMC. Conclusions: Our human data demonstrate that higher S100A4 levels are a marker of IR in adults with obesity but not in prepubertal children. Furthermore, the in vitro results suggest that S100A4 might exert an anti‐inflammatory effect. Further studies will be necessary to determine whether S100A4 can be a therapeutic target for obesity. K E Y W OR D S insulin resistance, obesity, prepubertal children, S100A4 1 | I N T RO D U CT I O N (appearance of thermogenic beige/brite adipocytes in WAT depots) has been associated with protection against obesity and improvement According to updated World Health Organization (WHO) reports, of associated metabolic alterations.7 Finally, angiogenesis is a key pro- 1 overweight and obesity worldwide have almost tripled since 1975 . cess required for adipogenesis in vivo and has an important role in In 2016, more than 1.9 billion adults (older than 18 years) were over- inflammation‐induced adipose tissue expansion.8 In the normal weight weight, and out of these, 650 million were obese. The problem is not state, adipose tissue is a well‐vascularized tissue.9 During the early only increasing in adults but also in adolescent and children (5‐18 stages of obesity development, angiogenic factors are overexpressed years), where more than 340 million were either overweight or obese to facilitate healthy adipose tissue expansion and confer protection 1 in 2016. Hallmarks of the obesity state, including visceral fat accumu- from metabolic insults.10 Therefore, a better understanding of adipose lation and low‐grade chronic inflammation, are associated with tissue endocrine, immunological and angiogenic capability, and the increased secretion of pro‐inflammatory cytokines and the develop- identification of novel potential biomarkers and therapeutic targets ment of metabolic diseases including cardiovascular disease and type to prevent and treat T2D is one of the goals in obesity field. 2,3 2 diabetes (T2D). S100A4 is a member of the S100 calcium‐binding protein family The pro‐inflammatory state plays an important role in insulin resis- and has been reported to be involved in controlling cell growth, differ- tance,2 but the link between obesity and the development of T2D is entiation, survival, and migration in different cell types, as well as in not well‐characterized. It is known that white adipose tissue (WAT) energy metabolism.11,12 S100A4 was identified by Ebradlize and col- has an endocrine role through the secretion of signalling molecules leagues as a protein associated with tumour metastasis.13 S100A4 such as adipokines by adipocytes and other cell types present in adi- binds to cytoskeletal proteins including F‐actin14 and nonmuscle myo- pose depots. These molecules participate in a hormonal network com- sin heavy chains.15 S100A4 gene expression is elevated in activated municating several tissues and organs including muscle, liver, and macrophages.16 A recent report has identified S100A4 as a cytokine brain.4 Moreover, adipose tissue expansion occurring during obesity secreted by adipose tissue mainly by precursors and immune cells, also leads to immune cell recruitment. Therefore, adipose tissue may with a potential role as a marker of subcutaneous WAT dysfunction be considered an immunological organ since these immune cells can and insulin resistance in adult individuals with severe obesity.17 More- secrete pro‐inflammatory and anti‐inflammatory cytokines affecting over, a study performed in mice postulates that S100A4 protects adipocyte biology, metabolism, and systemic inflammation, and being against diet‐induced obesity and inflammatory state.18 a key factor in insulin resistance development and obesity‐related In this study, our hypothesis was that S100A4 would be elevated pathologies.2,5 However, some patients with obesity show normal glu- and directly correlated to insulin resistance in children with obesity cose homeostasis,6 and previously unrecognized molecules could be as well as in adults. To address this hypothesis, S100A4 levels in chil- involved in this situation. Moreover, browning of subcutaneous WAT dren and adult patients with and without obesity were quantified and TAXERÅS ET AL. 101 their relationship with insulin resistance was determined. Moreover, to measure circulating levels of our molecule of interest. Plasma sam- the effects of S100A4 in human adipocytes and vascular smooth mus- ples were collected after a 12‐hour fasting and stored at −80°C in the cle cells (VSMCs) were studied. Biobank for the Health Sciences Research Institute Germans Trias i Pujol Foundation. 2 | M A T E RI A L S , S U BJ E CT S A ND M E TH O D S 2.3 | Circulating parameters 2.1 | Children study Glucose, insulin, and glycated haemoglobin (HbA1c) levels were mea- sured in the certified core clinical laboratory at the hospital. The Data from this cohort have been previously published, reporting the homeostatic model assessment‐insulin resistance (HOMA‐IR) was effects of lifestyle intervention on the metabolomic profiles in blood calculated. and urine.19,20 In the present study, we quantified S100A4 levels in S100A4 circulating levels were measured using a human protein the blood samples to evaluate potential associations with glucose S100A4 ELISA kit (Cusabio Biotech Co, LTD, China). High molecular homeostasis. Briefly, children attending the Obesity Unit at the Sant weight (HMW) adiponectin levels were measured using a human pro- Joan de Déu Hospital were recruited between January 2013 and tein HMW‐adiponectin ELISA kit (R&D Systems, China). December 2014. All parents signed the informed consent form. Inclu- sion criteria were (a) age from 7 to 10 years of age; (b) with obesity, 2.4 | Cell culture defined as body mass index‐standard deviation score (BMI‐SDS) greater than 2 standard deviations for a given age and sex, using the 2.4.1 | Human aortic VSMCs WHO standards; and (c) prepubertal, defined as Tanner stage I breast development for girls and testicular volume less than 4 mL in boys. Human aortic VSMCs were obtained by using a modification of the Exclusion criteria included any form of endocrine obesity, major con- explant technique and cultured as previously described.22 Human genital or chronic disease, drug‐induced obesity, use of drugs for VSMC were obtained from nonatherosclerotic aortas of hearts weight loss, or involvement in another weight‐loss programme. Fifty‐ removed on occasion of transplant surgery at the Hospital de la Santa three children with obesity were initially enrolled in the study; plasma Creu i Sant Pau (Barcelona, Spain) and cultured in M199 (Gibco, Carls- samples were available from 50 of these subjects at the hospital's bad, CA, USA) supplemented with 20% fetal calf serum (FCS), 2% Biobank. Fifteen prepubertal normal‐weight children (BMI‐SDS human serum, 2 mmol L−1 L‐glutamine, and antibiotics as previously between −1 and 1 standard deviations) between 7 and 10 years of described (22). Cells from four different donors between passages 3 age were also recruited for the study. Plasma samples from all subjects and 6 were seeded in six‐multiwell plates. Subconfluent cells were were obtained in the morning after an overnight fast by qualified per- starved in medium supplemented with 1% fetal bovine serum (FBS) sonnel at the hospital. for 24 hours prior to the addition of S100A4 (37.5, 75, or 150 ng mL−1) recombinant protein (R&D Systems) or vehicle (phosphate‐buff- 2.2 | Adult subjects study ered saline [PBS] 1×) for 24 hours in the presence or absence of TNF‐α (50 ng mL−1) (Sigma Aldrich) used as positive controls of inflam- All participants gave their written informed consent before collecting matory stimuli. clinical data. Forty‐three patients (11 males and 32 females) with severe obesity were included in this study. Ten of them were diag- 2.4.2 | Human adipocytes nosed with T2D, and eight had pre‐diabetes. All the patients were evaluated by the same endocrinologist (S. P.) following the institutional Human SGBS23 were used as a model of human adipocyte cells in cul- protocol for bariatric surgery between September 2012 and October ture. Cells were maintained with DMEM/F12 supplemented with 10% 2015, according to criteria for bariatric surgery formulated in Spanish heat‐inactivated FBS, 32μM biotin and 16μM pantothenic acid. For dif- Position Statement between Obesity, Endocrinology, Diabetes and ferentiation, cells were incubated with DMEM/F12 supplemented with Surgery Societies.21 Inclusion criteria were (a) adults with BMI greater 32μM biotin, 16μM pantothenic acid, 25nM dexamethasone, 500μM −2 −2 than 35 kg m and (b) adults with BMI less than 26 kg m and normal IBMX, 2μM rosiglitazone, 10 μg mL−1 transferrin, 20μM insulin, −1 glucose levels (less than 100 mg dL ). Exclusion criteria included any 100nM cortisol, and 0.2nM T3 for 7 days. Cells were then incubated form of endocrine obesity, major congenital or chronic disease, drug‐ with DMEM/F12 supplemented with 32μM biotin, 16μM panthotenic induced obesity, use of drugs for weight loss, or involvement in acid, 10 μg mL−1 transferrin, 20μM insulin, 100nM cortisol, and 0.2nM another weight‐loss programme. Demographic and clinical data, T3 for seven more days or until they reach more than 90% of differen- including age, history of diabetes, and hypertension were prospec- tiation, where indicated cells were treated with 1mM cAMP (Sigma‐ tively recorded in all subjects. Visceral adipose tissue was obtained Aldrich) and S100A4 (150 ng mL−1 and 750 ng mL−1) for 6 and 24 hours from a subset of these patients (n = 16). These samples were collected or with 1μM CL316.243 (Cayman Chemicals) for 24 hours for acute during bariatric surgery and frozen at −80°C until they were used. Six- response. Cells were treated with S100A4 (150 and 750 ng mL−1) during teen normal‐weight healthy subjects were included as a control group all the differentiation process for chronic response. 102 TAXERÅS ET AL. 2.4.3 | Gene expression TABLE 1 Anthropometric and metabolic parameters Normal weight(n = Obesity (n = P Visceral adipose tissue or cells were collected in liquid nitrogen. Total Children 15) 50) Value RNA was extracted using a column‐affinity–based methodology Age, y 8.4 (0.8) 9.0 (1.1) .060 (NucleoSpin RNA II; Macherey‐Nagel, Duren, Germany). RNA yield Sex (F/M) 7/8 28/22 n.a. was determined by spectrophotometry (NanoDrop). Five hundred nanograms of total RNA was retrotranscribed into cDNA using ran- BMI‐SDS 0.04 (0.86) 3.41 (0.75) <.001 dom hexamer primers and Multiscribe reverse transcriptase (TaqMan Glucose, mg dL−1 86 (6) 85 (8) .620 reverse transcription reagents, Thermo Fisher Scientific), following Insulin, m.u.int/L 3.1 (1.6) 13.4 (6.6) <.001 the manufacturer's instructions. Platinum Quantitative PCR HbA1c, % 5.3 (0.2) 5.3 (0.2) .891 SuperMix‐UDG with ROX reagent (Themo Fisher) was used as master HOMA‐IR 0.67 (0.36) 2.82 (1.44) <.001 mix reagent and expression levels of each gene of interest were a Triglyceride, mg 43 [15] 70 [36] <.001 assessed with the specific TaqMan probes (Thermo Fisher). mRNA dL−1 levels of browning markers including uncoupling protein 1 (UCP1), per- LDL‐Cho, mg dL−1 91 (19) 103 (23) .054 oxisome proliferator‐activated receptor gamma coactivator 1‐alpha −1 HDL‐Cho mg dL 62 (11) 44 (10) <.001 (PGC‐1α) and inflammatory markers including monocyte Cholesterol mg dL−1 162 (21) 164 (24) .770 chemoattractant protein‐1 (MCP‐1/CCL2), interleukin‐6 (IL‐6), or adi- a −1 S100A4, ng mL 48 [28] 74 [39] .002 pogenesis markers such as adiponectin (ADIPOQ) were assessed. HMW‐Adp,a ng mL−1 5208 [3902] 3496 [2477] .006 Adults Normal weight (n = 16) Obesity (n = 43) 2.5 | Statistical analysis Age, y 45 (9) 46 (11) 0.933 Unless otherwise stated, data are presented as mean ± SEM. Normal- Sex (F/M) 13/3 32/11 n.a. −2 ity was tested using the Shapiro‐Wilk test. To analyse potential asso- BMI, kg m 24.2 (2.3) 45.7 (7.4) <.001 −1 ciations with other variables, standard least‐squares linear regression Glucose, mg dL 89 (8) 109 (29) .002 in JMP v14.1 was performed (SAS Institute Inc, Cary, NC) with log‐ Insulin, m.u.int/L 6.9 (4.9) 14.0 (15.6) .008 transformed circulating S100A4 levels as the independent variable, HbA1c, % 5.2 (0.2) 5.8 (0.8) <.001 sex, age, and BMI (BMI‐SDS for children) as covariates, and either HOMA‐IR 1.6 (1.3) 4.0 (5.2) .003 HOMA‐IR, insulin, HbA1c, or HMW‐adiponectin (HMW‐Adp) as the a Triglyceride, mg 62 [20] 127 [53] <.001 independent variable. Two‐tailed test, Mann‐Whitney U test for dL−1 non‐normal distributed data, or one‐way ANOVA with Tukey's post LDL‐Cho, mg dL−1 103 (23) 98 (34) .516 hoc test for more than two groups were applied to determine statisti- −1 HDL‐Cho, mg dL 67 (15) 42 (9) <.001 cal significance. P < .05 was considered significant for all analyses. −1 Cholesterol, mg dL 188 (30) 164 (38) .021 The Institutional Ethics Committee, in accordance with the Decla- ration of Helsinki, approved the study with adult subjects (CP15/ S100A4,a ng mL−1 52.5 [33.8] 114.8 [86.6] <.001 00106 and PI17/01455). The children cohort study was approved by HMW‐Adp,a ng mL−1 5917 [4651] 3042 [1984] .008 the Sant Joan de Déu Hospital's Ethic Committee (Comité Ètic d' Note. Data for normally distributed variables are shown as mean (SD), and Investigació Clínica—CEIC PIC‐21‐12). two‐tail Student's t test was applied to compare groups. Abbreviations: BMI‐SDS, standardized body mass index; HbA1c, glycated haemoglobin; HDL‐Cho, high‐density lipoprotein; HMW‐Adp, high molec- 3 | RESULTS ular weight adiponectin; HOMA‐IR, homeostatic model assessment‐insulin resistance; LDL‐Cho, low‐density lipoprotein. a Non‐normally distributed variables are shown as median [IQR], and 3.1 | Circulating S100A4 levels are elevated in both groups were compared using the two‐tail Mann‐Whitney U test. Bold font prepubertal children and adult subjects with obesity indicates P < .05. Both prepubertal children and adults with obesity showed higher insulin children (1.5‐fold, P = .002) and adult subjects with obesity (2.2‐fold, resistance (estimated by HOMA‐IR) compared with their normal‐weight P < .001, Table 1) were found compared with normal‐weight controls. counterparts (Table 1). However, glucose and HbA1c levels in children with obesity were not altered compared with their normal‐weight con- trols, while they were significantly elevated in adults with obesity 3.2 | S100A4 levels are associated with HOMA‐IR in (Table 1). Both children and adults with obesity showed higher triglycer- adults but not in prepubertal children ide levels, lower HDL‐cholesterol, and lower HMW‐Adp concentration than normal‐weight subjects (Table 1). Circulating S100A4 concentra- Then the association between circulating S100A4 levels and insulin tion was measured and significantly elevated levels in both prepubertal resistance in the two different cohorts was tested. Given that obesity TAXERÅS ET AL. 103 FIGURE 1 S100A4 is expressed in visceral white adipose tissue (vWAT) and correlates with circulating levels and homeostatic model assessment‐insulin resistance (HOMA‐IR). (A‐ B) Bivariate correlation plots between log‐ transformed circulating S100A4 levels and HOMA‐IR in (A) adults and (B) children with obesity. (C‐D) Bivariate plot between S100A4 mRNA in vWAT and (C) circulating levels and (D) HOMA‐IR. S100A4 gene expression is expressed as arbitrary units of S100A4 mRNA levels relative to the expression of the housekeeping transcript PPIA. *P < .05 has a major impact on both S100A4 levels and insulin resistance, the 3.4 | S100A4 exerts anti‐inflammatory effects in focus of the analysis was on subjects with obesity to avoid this con- vascular smooth‐muscle cells in vitro founding factor. Circulating S100A4 levels were strongly correlated to HOMA‐IR in adult subjects with obesity (Figure 1A) and were Since WAT seems to play a role in the secretion and function of maintained after adjusting for sex, age, and BMI (βstd = .42, P = S100A4, the aim was to evaluate the effects of this adipokine in two .008, Table 2). However, no correlation between S100A4 and different cell types with a relevant role in adipose tissue expandability HOMA‐IR was found in prepubertal children with obesity (Figure 1B) during obesity: adipocytes and VSMC. For these experiments, human before or after adjusting for sex, age, and BMI‐SDS (βstd = .12, P = adipocytes and VSMC were incubated with S100A4 for 24 hours .356, Table 2). A modest trend towards a correlation with HMW‐ (acute) or during 1 week of differentiation for human adipocytes Adp was observed in children, without reaching statistical significance (chronic). No effects were found in UCP1, PGC‐1α, CCL2, IL‐6, or (βstd = .27, P = .061, Table 2). ADIPOQ in human adipocytes (Figure 2). In human VSMC, a dose‐ response curve with 37.5, 75, and 150 ng mL−1 of S100A4 for 24 hours was performed, showing that only the incubation with 75 3.3 |S100A4 gene expression in WAT is associated and 150 ng mL−1 significantly decreased CCl2, IL‐6, and IL‐1β mRNA with S100A4 circulating levels and HOMA‐IR in adult levels. No effect was observed with the lowest dose of S100A4 subjects with obesity (37.5 ng mL−1). Furthermore, only the dose of 150 ng mL−1 partially blocked the increase of CCl2 and IL‐6 (but not IL‐1β) expression To evaluate the potential role of WAT in S100A4 secretion, S100A4 induced by incubation with TNF‐α, used as a positive control for mRNA levels were measured in visceral WAT (vWAT) samples from inflammation (Figure 3). adult patients with obesity. S100A4 expression in this tissue showed a positive correlation with its circulating levels and HOMA‐IR (Figure 1C,D). 4 | DISCUSSION To the best of our knowledge, this is the first time that S100A4 circu- TABLE 2 Correlations between circulating S100A4 levels and phys- iologic variables lating levels are reported to be elevated in prepubertal children with obesity compared with normal weight healthy controls. As expected S100A4 Children Adults from previously published data (ref Arner), S100A4 levels were also Variables βstd P value βstd P value elevated in our cohort of adult patients with severe obesity compared HOMA‐IR .12 .356 .42 .008 with normal‐weight controls and showed a correlation with insulin Insulin .06 .650 .43 .008 resistance. Surprisingly, the association between S100A4 and insulin HbA1c .13 .391 .25 .106 resistance was not present in children. S100A4 levels have been reported to be increased in human adults HMW‐Adp .27 .061 .16 .445 with obesity, being positively associated with sWAT dysfunction, 104 TAXERÅS ET AL. FIGURE 2 Treatment with S100A4 does not cause changes in human adipocytes. (A) Relative mRNA levels of different markers of adipogenesis, thermogenic capacity, and inflammation in human adipocytes treated with 150 or 750 ng mL−1 S100A4 for 24 hours or during differentiation, where indicated. Expression was normalized to 18S. (B) Representative optical microscopy images from beige adipocytes treated with 750 ng mL−1 S100A4 since day 1 of differentiation. (*P < .05) relative to control. The bars represent means ± SEM inflammation, and insulin resistance.17 Other researchers have shown correlation between S100A4 and HMW‐Adp, it is plausible that the that S100A4 deficient mice were more susceptible to develop obesity elevated S100A4 levels may play a beneficial role during early states and insulin resistance after a high‐fat diet, suggesting a protective role of obesity in an attempt to balance the hyperinsulinaemia and the for S100A4 in diet‐induced obesity.18 Furthermore, overexpressed potential development of insulin resistance in this population in which S100A4 inhibited adipogenesis and decreased inflammatory markers obesity is only at the beginning of eliciting metabolic dysfunction. in 3T3‐L1 adipocytes.18 Our results suggest a potential explanation Moreover, our in vitro experiments demonstrate a beneficial direct for the controversial role of S100A4 in obesity and insulin resistance. effect of S100A4 on VSMC by reducing inflammatory markers. Our cohort of children with obesity showed hyperinsulinaemia, but S100A4 is secreted by adipocyte precursors and immune cells17 and, this excess of insulin secretion was able to maintain normal glucose during obesity, both an alteration of adipogenesis and the recruitment levels. Researchers have reported an induction of adiponectin by of immune cells occur (reviewed in24). In addition, new blood vessel 18 S100A4 in 3T3‐L1 adipocytes and in mice. Although there was no formation (angiogenesis), including VSMC, is required for a correct FIGURE 3 Treatment with S100A4 reduces the expression of pro‐inflammatory marker genes in human vascular smooth muscle cell (VSMC). Human VSMCs were exposed to S100A4 (37.5, 75, or 150 ng mL−1) or/and TNF‐α for 24 h. (A‐C) Expression of CCL2, IL6, and IL1β assessed by real‐time PCR and normalized to GAPDH 24 h after incubation with S100A4 (37.5, 75, or 150 ng mL−1). (D‐F) Expression of CCL2, IL6, and IL1β assessed by real‐time PCR and normalized to GAPDH 24 h after incubation with TNF‐α or TNF‐α + S100A4 (75 or 150 ng mL−1). Values are shown as mean ± SEM (n = 5); one‐way ANOVA was implemented to compare different treatments (***P < .001, **P < .01, and *P < .05 vs vehicle; #P < .05 vs TNF‐α) TAXERÅS ET AL. 105 adipose tissue expandability.25 Moreover, inflammation and vascular contributed to data analysis and interpretation. Joan Bel, Marta dysfunction contribute to the development of obesity and insulin Murillo, and María Martínez contributed to study design and resistance.26 Therefore, it is tempting to speculate that S100A4 secre- reviewed/edited the manuscript. Marta Ramon‐Krauel and Manel tion by immune cells infiltrated in adipose tissue during obesity acts on Puig‐Domingo contributed to study design and data interpretation. VSMC and thereby regulates the vascular tone, distribution of blood Francesc Villarroya, Carles Lerin, and David Sánchez‐Infantes contrib- flow, as well as angiogenesis, inflammatory processes, and redox sta- uted to study design, researched data, and wrote/reviewed/edited the tus. It is plausible that that this beneficial effect could contribute to manuscript. Each author listed in the manuscript has seen and improving the inflammatory state during the early stages of obesity approved the submission of this version of the manuscript and takes development, delaying the development of adipose tissue dysfunction full responsibility for the contents. and insulin resistance. In accordance with other researchers,17 no direct effect of S100A4 ACKNOWLEDGEMENTS on adipocytes was observed possibly due to the lack of specific recep- Samples from patients included in this study were provided by the IGTP tors, suggesting that these cells are not a target for S100A4 action. In HUGTP Biobank and the “Biobanc de l'Hospital Infantil Sant Joan de agreement with Arner et al,17 who postulated that S100A4 is an Déu per a la Investigació,” both integrated in the Spanish National adipokine, our data showed that S100A4 is expressed in WAT. In addi- Biobanks Network of Instituto de Salud Carlos III (PT13/0010/0009), tion, S100A4 vWAT gene expression correlated with circulating levels, and they were processed following standard operating procedures with suggesting that vWAT is a source of circulating S100A4. Moreover, the appropriate approval of the Ethical and Scientific Committees. S100A4 mRNA and circulating levels positively correlated with glu- cose homeostasis parameters, suggesting a role of this adipokine in CONFLIC T OF INT E RE ST the metabolic status of these patients. An intriguing possibility is that The authors declared no conflict of interest. a chronic release of this molecule in the obesity status could lead to an S100A4 resistant situation, in a similar way to what has been shown FUNDING INFORMATION for other molecules including leptin and FGF21.27,28 Further studies D. S. I. and M. G. are investigators of the Miguel Servet Fund from are warranted to determine the occurrence of an S100A4 resistant Carlos III National Institute of Health, Spain. This study was supported state in obesity. by Instituto de Salud Carlos III, by the Fondo Europeo de Desarrollo The limitations of this study include the lack of adipose tissue sam- Regional (FEDER), Madrid, Spain (CP15/00106 and FIS17/01455), ples in children to analyse the molecular mechanisms involved in the and by K‐sted 651010‐project number 70084900‐analyse number elevation of S100A4 and the lack of correlation with insulin resistance. M50676 from Norwegian University of Science and Technology Moreover, HOMA‐IR was used as a marker of insulin resistance due to (NTNU). the lack of a clamp assay. Another important limitation is that our vas- cular cellular model comes from aorta and not from adipose tissue ves- ORCID sels. In addition, longer and chronic S100A4 exposure would be desirable to evaluate the long‐term effect of S100A4 in vitro. The Carles Lerin https://orcid.org/0000-0003-0950-4363 strengths include the large cohort of plasma samples from prepubertal David Sánchez‐Infantes https://orcid.org/0000-0001-6086-7501 children with obesity and the availability of additional physiologic parameters allowing to identify the associations between S100A4 RE FE RE NC ES and other surrogates of metabolic state. 1. World Health Organization. Obesity and overweight. 2017 18 Oct 2017 [cited 2018 5 May]. In conclusion, our human data demonstrate for the first time that 2. Apostolopoulos V, de Courten MP, Stojanovska L, Blatch GL, S100A4 circulating levels are higher in children with obesity compared Tangalakis K, de Courten B. The complex immunological and inflamma- with normal‐weight individuals but do not correlate with insulin resis- tory network of adipose tissue in obesity. Mol Nutr Food Res. tance as in adult subjects. Moreover, our in vitro results suggest a 2016;60:43‐57. potential anti‐inflammatory effect of S100A4 in VSMC. Further stud- 3. Heymsfield SB, Wadden TA. Mechanisms, pathophysiology, and man- ies will be necessary to determine whether S100A4 can be a thera- agement of obesity. N Engl J Med. 2017;376:254‐266. peutic target for obesity and the development of insulin resistance 4. Ronti T, Lupattelli G, Mannarino E. The endocrine function of adipose associated with this condition. tissue: an update. Clin Endocrinol (Oxf). 2006;64:355‐365. 5. Grant RW, Dixit VD. Adipose tissue as an immunological organ. Obesity (Silver Spring). 2015;23:512‐518. AUTHOR CONTRIBUTIONS 6. Sims EA. Are there persons who are obese, but metabolically healthy? Metabolism. 2001;50:1499‐1504. Siri D. Taxerås, María Galán, and Laura Campderros researched and 7. Villarroya J, Cereijo R, Villarroya F. An endocrine role for brown adi- analysed the data, contributed to data interpretation, and wrote the pose tissue? Am J Physiol Endocrinol Metab. 2013;305:E567‐E572. manuscript. Irene Piquer‐Garcia, Silvia Pellitero, Eva Martínez, Rocío 8. Cristancho AG, Lazar MA. Forming functional fat: a growing under- Puig, and Icíar Lucena researched and analysed the data. Jordi Tarascó, standing of adipocyte differentiation. Nat Rev Mol Cell Biol. Pau Moreno, and José Balibrea obtained adipose tissue samples and 2011;12:722‐734. 106 TAXERÅS ET AL. 9. Cao Y. Adipose tissue angiogenesis as a therapeutic target for obesity 20. Leal‐Witt MJ, Llobet M, Samino S, et al. Lifestyle intervention and metabolic diseases. Nat Rev Drug Discov. 2010;9:107‐115. decreases urine trimethylamine N‐oxide levels in prepubertal children 10. Sun K, Kusminski CM, Luby‐Phelps K, et al. Brown adipose tissue with obesity. Obesity (Silver Spring). 2018;26:1603‐1610. derived VEGF‐A modulates cold tolerance and energy expenditure. 21. Sabench Pereferrer F, Domínguez‐Adame Lanuza E, Ibarzabal A, Socas Mol Metab. 2014;3:474‐483. Macias M, Valentí Azárate V. García Ruiz de Gordejuela A, et al. Qual- 11. Gross SR, Sin CG, Barraclough R, Rudland PS. Joining S100 proteins ity criteria in bariatric surgery: consensus review and recommendations and migration: for better or for worse, in sickness and in health. Cell of the Spanish Association of Surgeons and the Spanish Society of Bar- Mol Life Sci. 2014;71:1551‐1579. iatric Surgery. Cir Esp. 2017;95:4‐16. 12. Chen M, Sinha M, Luxon BA, Bresnick AR, O'Connor KL. Integrin α6β4 22. Churchman AT, Siow RC. Isolation, culture and characterisation of vas- controls the expression of genes associated with cell motility, invasion, cular smooth muscle cells. Methods Mol Biol. 2009;467:127‐138. and metastasis, including S100A4/metastasin. J Biol Chem. 23. Wabitsch M, Brenner RE, Melzner I, et al. Characterization of a human 2009;284:1484‐1494. preadipocyte cell strain with high capacity for adipose differentiation. 13. Ebralidze A, Tulchinsky E, Grigorian M, et al. Isolation and characteriza- Int J Obes Relat Metab Disord. 2001;25:8‐15. tion of a gene specifically expressed in different metastatic cells and 24. Maurizi G, Della Guardia L, Maurizi A, Poloni A. Adipocytes properties whose deduced gene product has a high degree of homology to a and crosstalk with immune system in obesity‐related inflammation. J Ca2+‐binding protein family. Genes Dev. 1989;3:1086‐1093. Cell Physiol. 2018;233:88‐97. 14. Flynn AM, Rudland PS, Barraclough R. Protein interactions between 25. Virtue S, Vidal‐Puig A. Adipose tissue expandability, lipotoxicity and S100A4 (p9Ka) and other cellular proteins identified using in vitro the metabolic syndrome—an allostatic perspective. Biochim Biophys methods. Biochem Soc Trans. 1996;24:341S. Acta. 2010;1801:338‐349. 15. Ford HL, Silver DL, Kachar B, Sellers JR, Zain SB. Effect of Mts1 on the 26. Cao Y. Angiogenesis as a therapeutic target for obesity and metabolic structure and activity of nonmuscle myosin II. Biochemistry. diseases. Chem Immunol Allergy. 2014;99:170‐179. 1997;36:16321‐16327. 27. Myers MG Jr, Leibel RL, Seeley RJ, Schwartz MW. Obesity and leptin 16. Takenaga K, Nakamura Y, Sakiyama S. Cellular localization of pEL98 resistance: distinguishing cause from effect. Trends Endocrinol Metab. protein, an S100‐related calcium binding protein, in fibroblasts and 2010;21:643‐651. its tissue distribution analyzed by monoclonal antibodies. Cell Struct 28. Markan KR, Naber MC, Small SM, Peltekian L, Kessler RL, Potthoff MJ. Funct. 1994;19:133‐141. FGF21 resistance is not mediated by downregulation of beta‐klotho 17. Arner P, Petrus P, Esteve D, et al. Screening of potential adipokines expression in white adipose tissue. Mol Metab. 2017;6:602‐610. identifies S100A4 as a marker of pernicious adipose tissue and insulin resistance. Int J Obes (Lond). 2018;42:2047‐2056. 18. Hou S, Jiao Y, Yuan Q, et al. S100A4 protects mice from high‐fat How to cite this article: Taxerås SD, Galán M, Campderros L, diet‐induced obesity and inflammation. Lab Invest. 2018;98: et al. Differential association between S100A4 levels and 1025‐1038. insulin resistance in prepubertal children and adult subjects 19. Leal‐Witt MJ, Ramon‐Krauel M, Samino S, et al. Untargeted metabolo- with clinically severe obesity. Obes Sci Pract. 2020;6:99–106. mics identifies a plasma sphingolipid‐related signature associated with lifestyle intervention in prepubertal children with obesity. Int J Obes https://doi.org/10.1002/osp4.381 (Lond). 2018;42:72‐78.