ORIGINAL
ARTICLE
Irisin and its relation to insulin resistance and puberty in obese children: a longitudinal analysis Thomas Reinehr1, Clinton Elfers2, Nina Lass1, and Christian L. Roth2 1: Department of Pediatric Endocrinology, Diabetes and Nutrition Medicine, Vestische Hospital for Children and Adolescents Datteln, University of Witten/Herdecke, Germany; 2: Seattle Children’s Research Institute, University of Washington, USA This study is registered at clinicaltrials.gov (NCT00435734).
Context: Irisin is a recently identified myokine affecting metabolic and glucose homeostasis. However, the role of irisin in obesity and its metabolic consequences is discussed controversially and data in children are scarce. Objective: To study the relationships between irisin, insulin resistance, and puberty before and after weight loss in obese children with and without impaired glucose tolerance. Design: 1-year follow-up study in obese children participating in a lifestyle intervention. Setting: Primary care. Patients: 40 obese and 20 normal weight children of similar age, gender, and pubertal stage. Intervention: Outpatient 1-year intervention program based on exercise, behavior and nutrition therapy. Main Outcomes Measures: Fasting serum irisin, weight status (BMI-SDS), and as parameters of the Metabolic Syndrome (MetS): insulin resistance index (HOMA), blood pressure, and lipids. Results: The irisin levels were the highest in obese children with impaired glucose tolerance, followed by obese children with normal glucose tolerance and lowest in normal weight children (p⬍0.001). In a multiple linear regression analysis, baseline irisin was significantly associated with pubertal stage, HDL- cholesterol, and HOMA, but not to age, gender, BMI or any other parameter of the MetS. The irisin concentrations were significantly (p⫽0.010) lower in the prepubertal compared to the pubertal children. In longitudinal analyses, changes of irisin were significantly associated with entry into puberty and change of fasting glucose and 2h glucose in oGTT, but not to change of BMI or any other parameter. Conclusions: Irisin levels are related to pubertal stage and insulin resistance but not to weight status in childhood. Key words: irisin, obesity, children, weight loss, insulin resistance, blood pressure, lipids
O
besity is a complex disease involving a number of different peptides, transmitters, and their receptors controlling energy homeostasis (1). Both adipose tissue and skeletal muscle have been identified as endocrine organs secreting hormones called adipokines and myokines, respectively (2). It has been proposed that there is a mus-
cle-adipose tissue cross-talk (3), critical for the regulation of body weight and metabolism, but the specific metabolic pathways and mediators remain elusive. Irisin, a newly discovered myokine induced by exercise, may contribute to muscle-adipose tissue cross-talk (4). Circulating irisin results from C-terminal cleavage of the
ISSN Print 0021-972X ISSN Online 1945-7197 Printed in U.S.A. Copyright © 2015 by the Endocrine Society Received January 22, 2015. Accepted March 13, 2015.
Abbreviations:
doi: 10.1210/jc.2015-1208
J Clin Endocrinol Metab
jcem.endojournals.org
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 24 March 2015. at 15:44 For personal use only. No other uses without permission. . All rights reserved.
1
2
Irisin in obese children
fibronectin type III domain containing (FNDC) 5 transmembrane protein, which is the product of FNDC5 gene (4, 5). This process is induced by the peroxisome proliferator-activated receptor (PPAR)-␥ coactivator (PGC)-1␣ (4). Irisin, by acting through a currently unknown receptor, increases thermogenesis and possibly improving glucose homeostasis (5– 8). Irisin is suggested to induce uncoupling protein 1 (UCP1) and subsequently increasing energy expenditure in white adipocytes of rodents, a process called adipocyte browning (4). Furthermore, irisin is discussed to be involved in the pathogenesis of various complications of obesity including dyslipidemia, type 2 diabetes mellitus, and arterial hypertension (9), summarized in the definition of the Metabolic Syndrome (MetS) (1). Accordingly, circulating irisin concentrations have been reported to be increased in insulin resistance states and Metabolic Syndrome (MetS) (8 –11). However, there are inconsistencies regarding the relevance of irisin in humans especially in obesity (12). For example, a positive association of irisin levels and BMI was found in some studies (6, 8, 10, 13), while other studies reported no correlation between irisin and BMI (14 – 16) or a negative correlation (7, 9, 10). Furthermore, the few weight loss studies in obese humans have also demonstrated controversial findings reporting decreasing irisin levels in adults (13, 17, 18) and increasing irisin levels in children (19). Interestingly, there are no studies on the relationship between irisin levels and puberty so far. This is of particular interest since insulin resistance increases with entry into puberty (20, 21). Therefore, one would expect an increase of irisin with entry into puberty if a relevant relationship between irisin and insulin resistance exists in humans. Given this partially unclear situation in obese humans and no studies concerning the relationship between puberty and irisin levels, we analyzed the long-term changes of irisin levels in obese children in respect to pubertal stage. We also compared irisin concentrations between obese and normal-weight children, studied the effects of weight reduction by a lifestyle intervention, and analyzed the relationships between irisin levels and parameters of the MetS such as insulin resistance, lipids, blood pressure (BP), and impaired glucose tolerance in the course of 1 year. Longitudinal studies and studies in obese children seem preferable since (1) cross-sectional studies cannot prove causality and are prone to many confounders, (2) adverse patterns of MetS itself begin in childhood, and (3) studies in children have the advantage that there is no potential confusion with other diseases, medications, or active tobacco smoking. We hypothesized that irisin concentrations increase with entry into puberty and are re-
J Clin Endocrinol Metab
lated to insulin resistance and parameters of the MetS. The novelty of our study is the longitudinal analysis of irisin concentrations in obese children participating in a lifestyle intervention analyzing the effect of puberty on irisin concentrations.
Materials and Methods Subjects Written informed consent was obtained from children and their parents. The study was approved by the local ethics committee of the University of Witten/Herdecke in Germany. We examined 40 obese Caucasian children from our obesity cohort (for details see (22)) and 20 normal- weight children. We decided to evaluate 10 obese children with impaired glucose tolerance at baseline and substantial weight loss after lifestyle intervention, 10 obese children with impaired glucose tolerance at baseline and no weight loss, 10 obese children with normal glucose tolerance at baseline and substantial weight loss, and 10 obese children with normal glucose tolerance at baseline and no weight loss to include both children with and without weight loss as well as children with and without disturbed glucose metabolism. Substantial weight loss was defined according to BMI-SDS reduction ⬎ 0.5 according to our previous studies (23, 24). All obese children in the different groups were matched for age, gender, and pubertal stage at baseline. All 40 obese children participated in the lifestyle intervention “Obeldicks”, which has been described in detail elsewhere (25). Briefly, this outpatient intervention program for obese children is based on physical exercise, nutrition education, and behavior therapy including the individual psychological care of the child and his or her family. The nutritional course is based on a fat and sugar-reduced diet as compared to the every-day nutrition of German children. None of the children in the current study suffered from endocrine disorders, premature adrenarche, or syndromal obesity (for details of diagnostic procedures see (22)).
Measurements We analyzed anthropometrics, pubertal stage, irisin, and as parameters of MetS: BP, HDL- cholesterol, LDL- cholesterol, triglycerides, glucose, and insulin. These variables were determined in all children at baseline and 1 year later in all obese children after participating in the lifestyle intervention “Obeldicks”. Furthermore, 2h glucose levels were analyzed at baseline and 1 year later in the obese children.
Clinical parameters Height was measured to the nearest centimeter using a rigid stadiometer. Weight was measured unclothed to the nearest 0.1 kg using a calibrated balance scale. BMI was calculated as weight in kilograms (kg) divided by the square of height in meters (m2). The degree of overweight was quantified using Cole’s least mean square method, which normalized the BMI skewed distribution and expressed BMI as a standard deviation score (BMI-SDS) (26). Reference data for German children were used (27). All children in the intervention study were obese according to the definition of the International Obesity Task Force (28). The pubertal developmental stage was determined according
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 24 March 2015. at 15:44 For personal use only. No other uses without permission. . All rights reserved.
doi: 10.1210/jc.2015-1208
jcem.endojournals.org
to Marshall and Tanner. Pubertal developmental stage was categorized into two groups based on breast and genital stages (prepubertal: boys with genital stage I, girls with breast stage I; pubertal: boys with genital stage ⱖII, girls with breast stage ⱖII). Blood pressure was measured using a validated protocol (29). Briefly, BP was measured at the right arm after a 10-minute rest in the supine position with an oscillometric device (Omron M6). Two repeated recordings were made with 5 minutes in between and the lowest value of the 2 recordings of systolic BP (SBP) and diastolic BP (DBP) measurements was recorded. The cuff size was based on the length and circumference of the upper arm and was as large as possible without having the elbow skin crease obstructing the stethoscope (29).
Biochemical parameters Blood sampling was performed in the fasting state at 8 a.m. After clotting, blood samples were centrifuged for 10 minutes at 8000 rpm. Serum was stored at – 81°C for later determination of irisin and insulin. All samples were thawed only once. Irisin levels were measured with a high- specific commercially available enzyme-linked immunoassay (ELISA) (Phoenix Pharmaceuticals, Cat. no. EK-067–29, CA, USA; sensitivity 1.29 ng/mL, intraassay coefficient of variation (CV) 4%– 8%, interassay CV 8%– 12%). Insulin concentrations were measured by microparticle enhanced immunometric assay (MEIATM, Abbott, Wiesbaden, Germany). Glucose levels were determined by colorimetric test using a VitrosTM analyzer (Ortho Clinical Diagnostics, Neckargmuend, Germany). Triglycerides, LDL- and HDL- cholesterol were determined by commercially available test kits (Roche Diagnostics, Mannheim, Germany). Intra- and interassay CVs were ⬍ 5% in all these methods. Homeostasis model assessment (HOMA) was used to detect the degree of insulin resistance using the formula: resistance (HOMA) ⫽ (insulin [mU/l] x glucose [mmol/l]) / 22.5 (30). An oral glucose tolerance test (OGTT) (oGTT) was performed in all obese children according to current guidelines (31). Impaired glucose tolerance (IGT) was defined by 2-hour serum glucose ⬎ 140 mg/dl in oGTT.
Table 1.
Irisin [ng/ml]
Statistics Statistical analyses were performed using the Winstat® software package (R. Fitch Software, Bad Krozingen, Germany). Normal distribution was tested by the Kolmogorov-Smirnov test. Baseline irisin levels were correlated to anthropometric variables, parameters of the MetS, and insulin resistance index HOMA by Spearman correlation. Changes of irisin levels were correlated to changes of the above mentioned variables in the one-year follow-up by Spearman correlation. Changes of puberty were set as 0 for remaining prepubertal or pubertal and 1 for entry into puberty. Furthermore, multiple backwards linear regressions analyses were calculated with irisin as dependent variable adjusted to age, gender and BMI and the independent variables BP, triglycerides, HDL- cholesterol, LDL- cholesterol, fasting glucose, insulin, HOMA, and pubertal stage. Since baseline irisin levels were not normally distributed irisin levels were log transformed in this analysis. Gender and pubertal stage were used as categorical variables in this model. To compare variables at baseline or in the course of 1 year, Fisher exact test, and Student’s t test for paired and unpaired observations, Wilcoxon test, Mann-Whitney U test, and Kruskal-Wallis test were used as appropriate. A p- value ⬍ 0.05 was considered as significant. Data were presented as mean and standard deviation or median and interquartile range (IQR).
Results The characteristics of the study cohort are demonstrated in Table 1. The irisin concentrations differed significantly between obese children with and without impaired glucose tolerance and normal weight children: The irisin levels were highest in obese children with impaired glucose tolerance, followed by obese children with normal glucose tolerance and lowest in normal weight children (figure 1). At baseline, the 30 boys did not differ significantly (P ⫽
Characteristics of the study population at baseline
number Age [years] Gender Pubertal stage BMI BMI-SDS Systolic BP [mmHg] Diastolic BP [mmHg] LDL-cholesterol [mg/dl] HDL- cholesterol [mg/dl] Triglycerides [mg/dl] Fasting glucose [mg/dl] 2 h glucose [mg/dl] Insulin [mU/liter] HOMA
3
Normal weight children
Children with NGT and no weight reduction
Children with NGT and weight reduction
Children with IGT and no weight reduction
Children with IGT and weight reduction
20 12.3 ⫾ 1.9 50% boys 50% prepubertal 18.9 ⫾ 2.5 0.12 ⫾ 0.69 112 ⫾ 8 52 ⫾ 5 83 ⫾ 7 54 ⫾ 4 76 ⫾ 12 84 ⫾ 3 n.d. 5 (IQR 3–7) 1.2 (IQR 0.7–1.5) 6.4 (IQR 5.7–7.7)
10 13.4 ⫾ 1.7 50% boys 30% prepubertal 29.6 ⫾ 3.6 2.31 ⫾ 0.37 122 ⫾ 15 68 ⫾ 11 87 ⫾ 31 54 ⫾ 13 94 ⫾ 44 85 ⫾ 4 105 ⫾ 5 19 (IQR 16 –23) 4.0 (IQR 3.4 – 4.4) 8.3 (IQR 6.8 –11.2)
10 12.8 ⫾ 1.0 50% boys 30% prepubertal 30.0 ⫾ 3.8 2.38 ⫾ 0.43 125 ⫾ 18 74 ⫾ 12 118 ⫾ 46 52 ⫾ 7 133 ⫾ 52 83 ⫾ 6 108 ⫾ 14 16 (IQR 10 –26) 3.3 (IQR 2.1–5.2) 10.2 (IQR 8.0 –11.7)
10 12.5 ⫾ 1.7 50% boys 30% prepubertal 31.5 ⫾ 5.1 2.56 ⫾ 0.45 123 ⫾ 10 75 ⫾ 8 98 ⫾ 27 50 ⫾ 14 134 ⫾ 59 89 ⫾ 7 156 ⫾ 13 38 (IQR 20 –70) 8.6 (4.3–13.7) 10.7 (IQR 7.2–12.4)
10 12.3 ⫾ 1.9 50% boys 30% prepubertal 29.7 ⫾ 2.7 2.50 ⫾ 0.20 127 ⫾ 14 72 ⫾ 8 104 ⫾ 23 47 ⫾ 12 135 ⫾ 37 88 ⫾ 6 155 ⫾ 8 34 (IQR 20 – 46) 7.1 (IQR 4.3–9.5) 11.2 (IQR 9.5–13.1)
p-value 0.875 0.999 0.688 ⬍0.001 ⬍0.001 0.015 ⬍0.001 0.025 0.208 ⬍0.001 0.095 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001
Data as mean and standard deviation or median and interquartile (IQR); n.d.: not determined, NGT: normal glucose tolerance, IGT: impaired glucose tolerance, BP: blood pressure, p-values derived from Kruskal Wallis test or Fisher exact test
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 24 March 2015. at 15:44 For personal use only. No other uses without permission. . All rights reserved.
4
Irisin in obese children
J Clin Endocrinol Metab
.404) with respect to their irisin levels as compared to 30 girls (boys: median irisin levels 7.7 (IQR 6.1–11.6) ng/ml; girls: median irisin levels 9.5 (IQR 7.1–11.7) ng/ml). The irisin concentrations differed significantly (P ⫽ .010) between the 22 prepubertal (median irisin concentration 6.8 (IQR 5.7–11.5) ng/ml) and 38 pubertal children (median irisin concentrations 9.5 (IQR 7.3–12.2) ng/ ml). Analyzing only obese children demonstrated the same findings: The irisin concentrations differed significantly (P ⫽ .010) between the 12 obese prepubertal (median irisin concentration 7.5 (IQR 5.8 –11.5) ng/ml) and 28 obese pubertal children (median irisin concentrations 10.9 (IQR 9.4 –12.8) ng/ml). Irisin was significantly correlated at baseline with BMI (r ⫽ 0.51, P ⬍ .001), BMI-SDS (r ⫽ 0.53, P ⬍ .001) and many parameters of the MetS such as insulin (r ⫽ 0.65, P ⬍ .001), 2h glucose in oGTT (r ⫽ 0.33, P ⫽ .020), HOMA (r ⫽ 0.65, P ⬍ .001), HDL- cholesterol (r⫽-0.398, P ⬍ .001), LDL- cholesterol (r ⫽ 0.36, P ⫽ .002), triglycerides (r ⫽ 0.53, P ⬍ .001), diastolic BP (r ⫽ 0.38, P ⫽ .001), but not significantly with age (r ⫽ 0.09), systolic BP (r⫽-0.16), and fasting glucose (r ⫽ 0.06). In a multiple linear backwards regression analysis (r2⫽0.28), baseline log-transformed irisin was significantly associated with pubertal stage (-coefficient 0.09 ⫾ 0.08, P ⫽ .048), HDL- cholesterol (-coefficient – 0.004 ⫾ 0.003, P ⫽ .027), and HOMA (-coefficient 0.014 ⫾ 0.012, P ⫽ .005), but not to age, gender, BMI or any other parameter of the MetS. Excluding the parameters of the MetS except HOMA in this model demonstrated also a significant association to pubertal stage (-coefficient 0.08 ⫾ 0.07, P ⫽ .016) and HOMA (-coefficient 0.011 ⫾ 0.009, P ⫽ .033), but not to age, gender, or BMI.
**
30
***
In the one-year follow-up period, the changes of irisin in the obese children correlated significantly with entry into puberty (r ⫽ 0.38, P ⫽ .009) and change of fasting glucose (r ⫽ 0.28, P ⫽ .043), but not to changes of BMI, BMI-SDS, or any other parameter of the MetS. Five obese children entered into puberty during the study period. Their irisin levels increased significantly (P ⫽ .043) from median 9.1 (IQR 6.1–10.1) ng/ml to median 12.2 (IQR 8.9 –13.6) ng/ml. In the 35 obese children without change of pubertal stage, the irisin levels remained stable (P ⫽ .110) in the study period (median 10.8 (IQR 9.2–12.5) ng/ml vs median 10.0 (IQR 7.7–11.8) ng/ml). In a multiple linear backwards regression analysis (r2⫽0.40), changes of irisin in the one-year follow-up period were significantly associated with entry into puberty (-coefficient 5.2 ⫾ 1.4, P ⬍ .001), changes of fasting glucose (-coefficient 0.17 ⫾ 0.06, P ⫽ .008) and changes of 2h glucose in oGTT (-coefficient 0.03 ⫾ 0.01, P ⫽ .023), but not to age, gender, BMI, changes of BMI, or any other parameter of the MetS. Excluding the parameters of the MetS except changes of HOMA in this model demonstrated also a significant association to pubertal stage (-coefficient 0.16 ⫾ 0.07, P ⫽ .008), but not to age, gender, BMI, or changes of HOMA or changes of BMI. In the 20 obese children with substantial reduction of BMI-SDS in the intervention period, most cardiovascular risk factors improved significantly, while irisin concentrations did not change significantly (Table 2 and figure 2). Analyzing only the 10 children with IGT at baseline and substantial weight loss also demonstrated no significant changes of irisin (P ⫽ .910). In contrast, the irisin levels increased significantly in the children without weight loss (Table 2) and in the children with IGT and without weight loss (figure 2).
ns
Discussion 20
10
0 normal w eight
obese and normal obese and impaired glucose tolerance glucose tolerance
Figure 1. Irisin levels (median, interquartile range, minimum and maximum) in 20 normal weight, 20 obese children with normal glucose tolerance, and 20 obese children with impaired glucose tolerance at baseline (median, interquartile range and range are shown; p-value derived from Kruskal Wallis test with Dunn’s Multiple Comparison Post Test).
To the best of our knowledge, this is the first study analyzing the longitudinal relationships between irisin, pubertal stage, obesity, and the parameters of MetS including insulin resistance index HOMA in obese children participating in a lifestyle intervention. Accordingly to most previous studies (8 –11;17;18;32– 34), irisin was significantly and positively associated to insulin resistance index HOMA and other parameters of the MetS both in cross-sectional and longitudinal analyses of our study. In contrast, one pediatric study reported (35) a negative relationship between HOMA and irisin levels. The reason for this discrepancy remains unclear. Since irisin has been reported to induce glucose and fatty acid uptake in human muscle (7, 36) the increase of irisin in
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 24 March 2015. at 15:44 For personal use only. No other uses without permission. . All rights reserved.
doi: 10.1210/jc.2015-1208
jcem.endojournals.org
5
creasing irisin levels were based short-term hypocaloric diet (8, 13, 17, 18) in contrast to our interven20 tion. It is well known that hypocaloric diet leads to a decrease of lean body mass. Interestingly, lean body 10 mass but not fat mass was a strong positive predictor of irisin levels in a previous study (14). Therefore, in 0 humans irisin seems to be mainly Baseline 1 y later Baseline 1 y later Baseline 1 y later Baseline 1 y later produced by muscle even if a producImpaired Glucose Normal Glucose Impaired Glucose Normal Glucose tion of irisin in the adipose tissue has Tolerance Tolerance Tolerance Tolerance been reported (3, 7). No Weight Loss Substantial Weight Loss One striking finding of our study was that irisin concentrations were B 5 significantly higher in pubertal obese p=0.002 p=0.002 p=0.275 p=0.846 p
ARTICLE
Irisin and its relation to insulin resistance and puberty in obese children: a longitudinal analysis Thomas Reinehr1, Clinton Elfers2, Nina Lass1, and Christian L. Roth2 1: Department of Pediatric Endocrinology, Diabetes and Nutrition Medicine, Vestische Hospital for Children and Adolescents Datteln, University of Witten/Herdecke, Germany; 2: Seattle Children’s Research Institute, University of Washington, USA This study is registered at clinicaltrials.gov (NCT00435734).
Context: Irisin is a recently identified myokine affecting metabolic and glucose homeostasis. However, the role of irisin in obesity and its metabolic consequences is discussed controversially and data in children are scarce. Objective: To study the relationships between irisin, insulin resistance, and puberty before and after weight loss in obese children with and without impaired glucose tolerance. Design: 1-year follow-up study in obese children participating in a lifestyle intervention. Setting: Primary care. Patients: 40 obese and 20 normal weight children of similar age, gender, and pubertal stage. Intervention: Outpatient 1-year intervention program based on exercise, behavior and nutrition therapy. Main Outcomes Measures: Fasting serum irisin, weight status (BMI-SDS), and as parameters of the Metabolic Syndrome (MetS): insulin resistance index (HOMA), blood pressure, and lipids. Results: The irisin levels were the highest in obese children with impaired glucose tolerance, followed by obese children with normal glucose tolerance and lowest in normal weight children (p⬍0.001). In a multiple linear regression analysis, baseline irisin was significantly associated with pubertal stage, HDL- cholesterol, and HOMA, but not to age, gender, BMI or any other parameter of the MetS. The irisin concentrations were significantly (p⫽0.010) lower in the prepubertal compared to the pubertal children. In longitudinal analyses, changes of irisin were significantly associated with entry into puberty and change of fasting glucose and 2h glucose in oGTT, but not to change of BMI or any other parameter. Conclusions: Irisin levels are related to pubertal stage and insulin resistance but not to weight status in childhood. Key words: irisin, obesity, children, weight loss, insulin resistance, blood pressure, lipids
O
besity is a complex disease involving a number of different peptides, transmitters, and their receptors controlling energy homeostasis (1). Both adipose tissue and skeletal muscle have been identified as endocrine organs secreting hormones called adipokines and myokines, respectively (2). It has been proposed that there is a mus-
cle-adipose tissue cross-talk (3), critical for the regulation of body weight and metabolism, but the specific metabolic pathways and mediators remain elusive. Irisin, a newly discovered myokine induced by exercise, may contribute to muscle-adipose tissue cross-talk (4). Circulating irisin results from C-terminal cleavage of the
ISSN Print 0021-972X ISSN Online 1945-7197 Printed in U.S.A. Copyright © 2015 by the Endocrine Society Received January 22, 2015. Accepted March 13, 2015.
Abbreviations:
doi: 10.1210/jc.2015-1208
J Clin Endocrinol Metab
jcem.endojournals.org
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 24 March 2015. at 15:44 For personal use only. No other uses without permission. . All rights reserved.
1
2
Irisin in obese children
fibronectin type III domain containing (FNDC) 5 transmembrane protein, which is the product of FNDC5 gene (4, 5). This process is induced by the peroxisome proliferator-activated receptor (PPAR)-␥ coactivator (PGC)-1␣ (4). Irisin, by acting through a currently unknown receptor, increases thermogenesis and possibly improving glucose homeostasis (5– 8). Irisin is suggested to induce uncoupling protein 1 (UCP1) and subsequently increasing energy expenditure in white adipocytes of rodents, a process called adipocyte browning (4). Furthermore, irisin is discussed to be involved in the pathogenesis of various complications of obesity including dyslipidemia, type 2 diabetes mellitus, and arterial hypertension (9), summarized in the definition of the Metabolic Syndrome (MetS) (1). Accordingly, circulating irisin concentrations have been reported to be increased in insulin resistance states and Metabolic Syndrome (MetS) (8 –11). However, there are inconsistencies regarding the relevance of irisin in humans especially in obesity (12). For example, a positive association of irisin levels and BMI was found in some studies (6, 8, 10, 13), while other studies reported no correlation between irisin and BMI (14 – 16) or a negative correlation (7, 9, 10). Furthermore, the few weight loss studies in obese humans have also demonstrated controversial findings reporting decreasing irisin levels in adults (13, 17, 18) and increasing irisin levels in children (19). Interestingly, there are no studies on the relationship between irisin levels and puberty so far. This is of particular interest since insulin resistance increases with entry into puberty (20, 21). Therefore, one would expect an increase of irisin with entry into puberty if a relevant relationship between irisin and insulin resistance exists in humans. Given this partially unclear situation in obese humans and no studies concerning the relationship between puberty and irisin levels, we analyzed the long-term changes of irisin levels in obese children in respect to pubertal stage. We also compared irisin concentrations between obese and normal-weight children, studied the effects of weight reduction by a lifestyle intervention, and analyzed the relationships between irisin levels and parameters of the MetS such as insulin resistance, lipids, blood pressure (BP), and impaired glucose tolerance in the course of 1 year. Longitudinal studies and studies in obese children seem preferable since (1) cross-sectional studies cannot prove causality and are prone to many confounders, (2) adverse patterns of MetS itself begin in childhood, and (3) studies in children have the advantage that there is no potential confusion with other diseases, medications, or active tobacco smoking. We hypothesized that irisin concentrations increase with entry into puberty and are re-
J Clin Endocrinol Metab
lated to insulin resistance and parameters of the MetS. The novelty of our study is the longitudinal analysis of irisin concentrations in obese children participating in a lifestyle intervention analyzing the effect of puberty on irisin concentrations.
Materials and Methods Subjects Written informed consent was obtained from children and their parents. The study was approved by the local ethics committee of the University of Witten/Herdecke in Germany. We examined 40 obese Caucasian children from our obesity cohort (for details see (22)) and 20 normal- weight children. We decided to evaluate 10 obese children with impaired glucose tolerance at baseline and substantial weight loss after lifestyle intervention, 10 obese children with impaired glucose tolerance at baseline and no weight loss, 10 obese children with normal glucose tolerance at baseline and substantial weight loss, and 10 obese children with normal glucose tolerance at baseline and no weight loss to include both children with and without weight loss as well as children with and without disturbed glucose metabolism. Substantial weight loss was defined according to BMI-SDS reduction ⬎ 0.5 according to our previous studies (23, 24). All obese children in the different groups were matched for age, gender, and pubertal stage at baseline. All 40 obese children participated in the lifestyle intervention “Obeldicks”, which has been described in detail elsewhere (25). Briefly, this outpatient intervention program for obese children is based on physical exercise, nutrition education, and behavior therapy including the individual psychological care of the child and his or her family. The nutritional course is based on a fat and sugar-reduced diet as compared to the every-day nutrition of German children. None of the children in the current study suffered from endocrine disorders, premature adrenarche, or syndromal obesity (for details of diagnostic procedures see (22)).
Measurements We analyzed anthropometrics, pubertal stage, irisin, and as parameters of MetS: BP, HDL- cholesterol, LDL- cholesterol, triglycerides, glucose, and insulin. These variables were determined in all children at baseline and 1 year later in all obese children after participating in the lifestyle intervention “Obeldicks”. Furthermore, 2h glucose levels were analyzed at baseline and 1 year later in the obese children.
Clinical parameters Height was measured to the nearest centimeter using a rigid stadiometer. Weight was measured unclothed to the nearest 0.1 kg using a calibrated balance scale. BMI was calculated as weight in kilograms (kg) divided by the square of height in meters (m2). The degree of overweight was quantified using Cole’s least mean square method, which normalized the BMI skewed distribution and expressed BMI as a standard deviation score (BMI-SDS) (26). Reference data for German children were used (27). All children in the intervention study were obese according to the definition of the International Obesity Task Force (28). The pubertal developmental stage was determined according
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 24 March 2015. at 15:44 For personal use only. No other uses without permission. . All rights reserved.
doi: 10.1210/jc.2015-1208
jcem.endojournals.org
to Marshall and Tanner. Pubertal developmental stage was categorized into two groups based on breast and genital stages (prepubertal: boys with genital stage I, girls with breast stage I; pubertal: boys with genital stage ⱖII, girls with breast stage ⱖII). Blood pressure was measured using a validated protocol (29). Briefly, BP was measured at the right arm after a 10-minute rest in the supine position with an oscillometric device (Omron M6). Two repeated recordings were made with 5 minutes in between and the lowest value of the 2 recordings of systolic BP (SBP) and diastolic BP (DBP) measurements was recorded. The cuff size was based on the length and circumference of the upper arm and was as large as possible without having the elbow skin crease obstructing the stethoscope (29).
Biochemical parameters Blood sampling was performed in the fasting state at 8 a.m. After clotting, blood samples were centrifuged for 10 minutes at 8000 rpm. Serum was stored at – 81°C for later determination of irisin and insulin. All samples were thawed only once. Irisin levels were measured with a high- specific commercially available enzyme-linked immunoassay (ELISA) (Phoenix Pharmaceuticals, Cat. no. EK-067–29, CA, USA; sensitivity 1.29 ng/mL, intraassay coefficient of variation (CV) 4%– 8%, interassay CV 8%– 12%). Insulin concentrations were measured by microparticle enhanced immunometric assay (MEIATM, Abbott, Wiesbaden, Germany). Glucose levels were determined by colorimetric test using a VitrosTM analyzer (Ortho Clinical Diagnostics, Neckargmuend, Germany). Triglycerides, LDL- and HDL- cholesterol were determined by commercially available test kits (Roche Diagnostics, Mannheim, Germany). Intra- and interassay CVs were ⬍ 5% in all these methods. Homeostasis model assessment (HOMA) was used to detect the degree of insulin resistance using the formula: resistance (HOMA) ⫽ (insulin [mU/l] x glucose [mmol/l]) / 22.5 (30). An oral glucose tolerance test (OGTT) (oGTT) was performed in all obese children according to current guidelines (31). Impaired glucose tolerance (IGT) was defined by 2-hour serum glucose ⬎ 140 mg/dl in oGTT.
Table 1.
Irisin [ng/ml]
Statistics Statistical analyses were performed using the Winstat® software package (R. Fitch Software, Bad Krozingen, Germany). Normal distribution was tested by the Kolmogorov-Smirnov test. Baseline irisin levels were correlated to anthropometric variables, parameters of the MetS, and insulin resistance index HOMA by Spearman correlation. Changes of irisin levels were correlated to changes of the above mentioned variables in the one-year follow-up by Spearman correlation. Changes of puberty were set as 0 for remaining prepubertal or pubertal and 1 for entry into puberty. Furthermore, multiple backwards linear regressions analyses were calculated with irisin as dependent variable adjusted to age, gender and BMI and the independent variables BP, triglycerides, HDL- cholesterol, LDL- cholesterol, fasting glucose, insulin, HOMA, and pubertal stage. Since baseline irisin levels were not normally distributed irisin levels were log transformed in this analysis. Gender and pubertal stage were used as categorical variables in this model. To compare variables at baseline or in the course of 1 year, Fisher exact test, and Student’s t test for paired and unpaired observations, Wilcoxon test, Mann-Whitney U test, and Kruskal-Wallis test were used as appropriate. A p- value ⬍ 0.05 was considered as significant. Data were presented as mean and standard deviation or median and interquartile range (IQR).
Results The characteristics of the study cohort are demonstrated in Table 1. The irisin concentrations differed significantly between obese children with and without impaired glucose tolerance and normal weight children: The irisin levels were highest in obese children with impaired glucose tolerance, followed by obese children with normal glucose tolerance and lowest in normal weight children (figure 1). At baseline, the 30 boys did not differ significantly (P ⫽
Characteristics of the study population at baseline
number Age [years] Gender Pubertal stage BMI BMI-SDS Systolic BP [mmHg] Diastolic BP [mmHg] LDL-cholesterol [mg/dl] HDL- cholesterol [mg/dl] Triglycerides [mg/dl] Fasting glucose [mg/dl] 2 h glucose [mg/dl] Insulin [mU/liter] HOMA
3
Normal weight children
Children with NGT and no weight reduction
Children with NGT and weight reduction
Children with IGT and no weight reduction
Children with IGT and weight reduction
20 12.3 ⫾ 1.9 50% boys 50% prepubertal 18.9 ⫾ 2.5 0.12 ⫾ 0.69 112 ⫾ 8 52 ⫾ 5 83 ⫾ 7 54 ⫾ 4 76 ⫾ 12 84 ⫾ 3 n.d. 5 (IQR 3–7) 1.2 (IQR 0.7–1.5) 6.4 (IQR 5.7–7.7)
10 13.4 ⫾ 1.7 50% boys 30% prepubertal 29.6 ⫾ 3.6 2.31 ⫾ 0.37 122 ⫾ 15 68 ⫾ 11 87 ⫾ 31 54 ⫾ 13 94 ⫾ 44 85 ⫾ 4 105 ⫾ 5 19 (IQR 16 –23) 4.0 (IQR 3.4 – 4.4) 8.3 (IQR 6.8 –11.2)
10 12.8 ⫾ 1.0 50% boys 30% prepubertal 30.0 ⫾ 3.8 2.38 ⫾ 0.43 125 ⫾ 18 74 ⫾ 12 118 ⫾ 46 52 ⫾ 7 133 ⫾ 52 83 ⫾ 6 108 ⫾ 14 16 (IQR 10 –26) 3.3 (IQR 2.1–5.2) 10.2 (IQR 8.0 –11.7)
10 12.5 ⫾ 1.7 50% boys 30% prepubertal 31.5 ⫾ 5.1 2.56 ⫾ 0.45 123 ⫾ 10 75 ⫾ 8 98 ⫾ 27 50 ⫾ 14 134 ⫾ 59 89 ⫾ 7 156 ⫾ 13 38 (IQR 20 –70) 8.6 (4.3–13.7) 10.7 (IQR 7.2–12.4)
10 12.3 ⫾ 1.9 50% boys 30% prepubertal 29.7 ⫾ 2.7 2.50 ⫾ 0.20 127 ⫾ 14 72 ⫾ 8 104 ⫾ 23 47 ⫾ 12 135 ⫾ 37 88 ⫾ 6 155 ⫾ 8 34 (IQR 20 – 46) 7.1 (IQR 4.3–9.5) 11.2 (IQR 9.5–13.1)
p-value 0.875 0.999 0.688 ⬍0.001 ⬍0.001 0.015 ⬍0.001 0.025 0.208 ⬍0.001 0.095 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001
Data as mean and standard deviation or median and interquartile (IQR); n.d.: not determined, NGT: normal glucose tolerance, IGT: impaired glucose tolerance, BP: blood pressure, p-values derived from Kruskal Wallis test or Fisher exact test
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 24 March 2015. at 15:44 For personal use only. No other uses without permission. . All rights reserved.
4
Irisin in obese children
J Clin Endocrinol Metab
.404) with respect to their irisin levels as compared to 30 girls (boys: median irisin levels 7.7 (IQR 6.1–11.6) ng/ml; girls: median irisin levels 9.5 (IQR 7.1–11.7) ng/ml). The irisin concentrations differed significantly (P ⫽ .010) between the 22 prepubertal (median irisin concentration 6.8 (IQR 5.7–11.5) ng/ml) and 38 pubertal children (median irisin concentrations 9.5 (IQR 7.3–12.2) ng/ ml). Analyzing only obese children demonstrated the same findings: The irisin concentrations differed significantly (P ⫽ .010) between the 12 obese prepubertal (median irisin concentration 7.5 (IQR 5.8 –11.5) ng/ml) and 28 obese pubertal children (median irisin concentrations 10.9 (IQR 9.4 –12.8) ng/ml). Irisin was significantly correlated at baseline with BMI (r ⫽ 0.51, P ⬍ .001), BMI-SDS (r ⫽ 0.53, P ⬍ .001) and many parameters of the MetS such as insulin (r ⫽ 0.65, P ⬍ .001), 2h glucose in oGTT (r ⫽ 0.33, P ⫽ .020), HOMA (r ⫽ 0.65, P ⬍ .001), HDL- cholesterol (r⫽-0.398, P ⬍ .001), LDL- cholesterol (r ⫽ 0.36, P ⫽ .002), triglycerides (r ⫽ 0.53, P ⬍ .001), diastolic BP (r ⫽ 0.38, P ⫽ .001), but not significantly with age (r ⫽ 0.09), systolic BP (r⫽-0.16), and fasting glucose (r ⫽ 0.06). In a multiple linear backwards regression analysis (r2⫽0.28), baseline log-transformed irisin was significantly associated with pubertal stage (-coefficient 0.09 ⫾ 0.08, P ⫽ .048), HDL- cholesterol (-coefficient – 0.004 ⫾ 0.003, P ⫽ .027), and HOMA (-coefficient 0.014 ⫾ 0.012, P ⫽ .005), but not to age, gender, BMI or any other parameter of the MetS. Excluding the parameters of the MetS except HOMA in this model demonstrated also a significant association to pubertal stage (-coefficient 0.08 ⫾ 0.07, P ⫽ .016) and HOMA (-coefficient 0.011 ⫾ 0.009, P ⫽ .033), but not to age, gender, or BMI.
**
30
***
In the one-year follow-up period, the changes of irisin in the obese children correlated significantly with entry into puberty (r ⫽ 0.38, P ⫽ .009) and change of fasting glucose (r ⫽ 0.28, P ⫽ .043), but not to changes of BMI, BMI-SDS, or any other parameter of the MetS. Five obese children entered into puberty during the study period. Their irisin levels increased significantly (P ⫽ .043) from median 9.1 (IQR 6.1–10.1) ng/ml to median 12.2 (IQR 8.9 –13.6) ng/ml. In the 35 obese children without change of pubertal stage, the irisin levels remained stable (P ⫽ .110) in the study period (median 10.8 (IQR 9.2–12.5) ng/ml vs median 10.0 (IQR 7.7–11.8) ng/ml). In a multiple linear backwards regression analysis (r2⫽0.40), changes of irisin in the one-year follow-up period were significantly associated with entry into puberty (-coefficient 5.2 ⫾ 1.4, P ⬍ .001), changes of fasting glucose (-coefficient 0.17 ⫾ 0.06, P ⫽ .008) and changes of 2h glucose in oGTT (-coefficient 0.03 ⫾ 0.01, P ⫽ .023), but not to age, gender, BMI, changes of BMI, or any other parameter of the MetS. Excluding the parameters of the MetS except changes of HOMA in this model demonstrated also a significant association to pubertal stage (-coefficient 0.16 ⫾ 0.07, P ⫽ .008), but not to age, gender, BMI, or changes of HOMA or changes of BMI. In the 20 obese children with substantial reduction of BMI-SDS in the intervention period, most cardiovascular risk factors improved significantly, while irisin concentrations did not change significantly (Table 2 and figure 2). Analyzing only the 10 children with IGT at baseline and substantial weight loss also demonstrated no significant changes of irisin (P ⫽ .910). In contrast, the irisin levels increased significantly in the children without weight loss (Table 2) and in the children with IGT and without weight loss (figure 2).
ns
Discussion 20
10
0 normal w eight
obese and normal obese and impaired glucose tolerance glucose tolerance
Figure 1. Irisin levels (median, interquartile range, minimum and maximum) in 20 normal weight, 20 obese children with normal glucose tolerance, and 20 obese children with impaired glucose tolerance at baseline (median, interquartile range and range are shown; p-value derived from Kruskal Wallis test with Dunn’s Multiple Comparison Post Test).
To the best of our knowledge, this is the first study analyzing the longitudinal relationships between irisin, pubertal stage, obesity, and the parameters of MetS including insulin resistance index HOMA in obese children participating in a lifestyle intervention. Accordingly to most previous studies (8 –11;17;18;32– 34), irisin was significantly and positively associated to insulin resistance index HOMA and other parameters of the MetS both in cross-sectional and longitudinal analyses of our study. In contrast, one pediatric study reported (35) a negative relationship between HOMA and irisin levels. The reason for this discrepancy remains unclear. Since irisin has been reported to induce glucose and fatty acid uptake in human muscle (7, 36) the increase of irisin in
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 24 March 2015. at 15:44 For personal use only. No other uses without permission. . All rights reserved.
doi: 10.1210/jc.2015-1208
jcem.endojournals.org
5
creasing irisin levels were based short-term hypocaloric diet (8, 13, 17, 18) in contrast to our interven20 tion. It is well known that hypocaloric diet leads to a decrease of lean body mass. Interestingly, lean body 10 mass but not fat mass was a strong positive predictor of irisin levels in a previous study (14). Therefore, in 0 humans irisin seems to be mainly Baseline 1 y later Baseline 1 y later Baseline 1 y later Baseline 1 y later produced by muscle even if a producImpaired Glucose Normal Glucose Impaired Glucose Normal Glucose tion of irisin in the adipose tissue has Tolerance Tolerance Tolerance Tolerance been reported (3, 7). No Weight Loss Substantial Weight Loss One striking finding of our study was that irisin concentrations were B 5 significantly higher in pubertal obese p=0.002 p=0.002 p=0.275 p=0.846 p
