The increasing prevalence of obesity in urban Indian children is indicative of an impending crisis of metabolic disorders. Although perturbations in the secretion of adipokines and inflammatory molecules in childhood obesity are well documented, the contribution of common variants of genes encoding them is not well investigated. We assessed the association of 125 common variants from 21 genes, encoding adipocytokines and inflammatory markers in 1,325 urban Indian children (862 normal weight [NW group] and 463 overweight/obese [OW/OB group]) and replicated top loci in 1,843 Indian children (1,399 NW children and 444 OW/OB children). Variants of four genes (PBEF1 [rs3801266] [P = 4.5 × 10−4], IL6 [rs2069845] [P = 8.7 × 10−4], LEPR [rs1137100] [P = 1.8 × 10−3], and IL6R [rs7514452] [P = 2.1 × 10−3]) were top signals in the discovery sample. Associations of rs2069845, rs1137100, and rs3801266 were replicated (P = 7.9 × 10−4, 8.3 × 10−3, and 0.036, respectively) and corroborated in meta-analysis (P = 2.3 × 10−6, 3.9 × 10−5, and 4.3 × 10−4, respectively) that remained significant after multiple testing corrections. These variants also were associated with quantitative measures of adiposity (weight, BMI, and waist and hip circumferences). Allele dosage analysis of rs2069845, rs1137100, and rs3801266 revealed that children with five to six risk alleles had an approximately four times increased risk of obesity than children with less than two risk alleles (P = 1.2 × 10−7). In conclusion, our results demonstrate the association of the common variants of IL6, LEPR, and PBEF1 with obesity in Indian children.

Manifestation of chronic metabolic disorders, including type 2 diabetes, cardiovascular diseases, and metabolic syndrome starts in early life, marked by a rapid increase in BMI during childhood (1). The increasing prevalence of obesity in Indian children (2) is indicative of an impending crisis of metabolic disorders that already has reached epidemic proportions in India. In this scenario, there is an urgent need to identify the underlying factors involved in the predisposition of childhood obesity and linking childhood obesity and adult chronic disorders.

Childhood obesity is a state of inflammation, as evident from the dysregulated secretions of adipokines and inflammatory molecules in the obese condition (3). In addition to inflammatory properties, these adipokines and inflammatory molecules also play an important role in energy homeostasis, metabolic processes, and regulation of body fat (3). Although perturbations in the secretion of adipokines and inflammatory molecules in childhood obesity are well documented, studies investigating the contribution of variants in genes encoding them to susceptibility of childhood obesity are few. Here, we examined 125 common variants from 21 genes encoding adipokines and inflammatory markers for association with overweight/obesity, markers of adiposity, and inflammation in urban Indian children.

The study involved the participation of 3,168 children (aged 11–17 years), including 2,261 normal-weight (NW group) and 907 overweight/obese (OW/OB group) children, recruited from school health surveys from four different zones of Delhi, as described previously (4). Subjects were classified as normal weight and overweight/obese according to the age- and sex-specific cutoffs provided by Cole et al. (5), which gives BMI cutoffs for overweight and obesity by sex for children between the age of 2 and 18 years corresponding to the cutoff points of 25 and 30 kg/m2 for adults. Prior permission from school authorities, informed consent from parents/guardians, and verbal assent from the participants themselves was obtained. The study protocol was approved by ethics committees of the participating institutes, and the study was conducted according to principles of the Declaration of Helsinki.

Anthropometric measurements including height, weight, waist circumference (WC), and hip circumference (HC) were taken using standard methods. BMI and waist-to-hip ratio (WHR) were calculated. Blood samples were drawn from the subjects after overnight fasting. Plasma levels of glucose and high-sensitivity C-reactive protein were measured using Cobas Integra 400 Plus (Roche Diagnostic, Mannheim, Germany). Levels of insulin and C-peptide were estimated using Elecsys 2010 (Roche Diagnostics). Plasma levels of leptin, resistin, and adiponectin were measured using commercial enzyme-linked immunosorbent assay kits (R&D Systems, Minneapolis, MN). Anthropometric and clinical characteristics of subjects are provided in Table 1.

TABLE 1

Anthropometric and clinical characteristics of study subjects

Anthropometric and clinical characteristics of study subjects
Anthropometric and clinical characteristics of study subjects

Genotyping.

In stage 1, we genotyped 1,325 children (862 NW and 463 OW/OB children) for 125 single nucleotide polymorphisms (SNPs) from 21 genes, encoding adipocytokines and inflammatory markers (Supplementary Table 1), using Illumina GoldenGate assay (Illumina, San Diego, CA). The SNPs were selected on the basis of their location in functionally important regions of genes, previous reports of association with metabolic disorders, minor allele frequency >0.05, and tag SNPs. Genotyping data were subjected to extensive quality control that included genotype confidence score >0.25, SNP call rate >0.9, GenTrans score >0.6, cluster separation score >0.4, Hardy-Weinberg equilibrium (P > 0.01 in NW, OW/OB, and all subjects], and minor allele frequency >0.05. Quality control passed SNPs (n = 83), had a call rate >98%, and a concordance rate of 99.97% based on 5% duplicate samples. Genotype frequencies for all the SNPs are provided in Supplementary Table 1.

In stage 2, genotyping for the replication of four loci (rs2069845 [IL6], rs1137100 [LEPR], rs3801266 [PBEF1], and rs7514452 [IL6R]) was performed in 1,843 children (1,399 NW and 444 OW/OB children) using iPLEX (Sequenom, San Deigo, CA). The average genotyping success rate was 93.5%, with >99.7% consistency in genotyping based on 5% duplicates.

Statistical analyses.

Statistical analyses were performed using PLINK version 1.07 (http://pngu.mgh.harvard.edu/∼purcell/plink) and SPSS version 17.0 (SPSS, Chicago, IL). Genotype frequencies were checked for Hardy-Weinberg equilibrium using the χ2 test. Prior to analysis, continuous variables were inverse normal transformed to achieve normal distribution, and internal age- and sex-specific z scores (SDs from the study group mean) were generated by dividing the difference between the study group mean and the individual’s value by SD for that age and sex. The association of SNPs with overweight/obesity was tested using logistic regression analysis under additive model adjusting for age and sex. Linear regression analysis was performed to assess the association of SNPs with quantitative traits adjusting for age and sex for obesity-related traits and age, sex, and z BMI for other traits. Meta-analysis was performed by combining summary estimates of stage 1 and stage 2. Bonferroni correction was applied in the meta-analysis stage to account for multiple testing, and a P value of <7.6 × 10−4 was considered significant after correcting for the number of independent SNPs (r2 < 0.80; n = 66). For quantitative clinical traits, a P value of <5.4 × 10−5 was considered significant after accounting for multiple comparison (0.05/[66*14]). The statistical power of the study was determined using Quanto software (http://hydra.usc.edu/gxe/), assuming a log additive model of inheritance and 24% prevalence of overweight/obesity (1) at α = 0.05. Sample size in stage 1 and combined analysis provided 44–76% and 76–97% power, respectively, to detect the association of the variant with an allele frequency of 0.20 and an effect size of 1.20–1.30.

The cumulative effect of associated SNPs was determined through allele dosage analysis by calculating effective risk scores (ERSs), as provided earlier (6). Subjects were classified into different groups by ERS (<2, 2–3, 3–4, 4–5, and >5–6), and odds ratios (ORs) were calculated considering an ERS of <2 as the reference.

Stage 1 analysis revealed the association of 17 variants in LEPR, IL6, IL6R, PBEF1,CES2, and TNFRSF1B with overweight/obesity at P < 0.05 (Supplementary Table 1), with the strongest signal at rs3801266 of PBEF1 (P = 4.5 × 10−4). The IL6 variants (rs2069845 [P = 8.7 × 10−4] and rs2069849 [P = 1.6 × 10−3]) and the LEPR nonsynonymous variants (rs1137100 [K109R] [P = 1.4 × 10−3], rs1137101 [Q223R] [P = 2.5 × 10−3], and rs8179183 [K656N] [P = 2.6 × 10−3]) were associated with the risk of overweight/obesity. Among the IL6R variants, rs7514452 and rs10752641 were found to be associated (P = 2.3 × 10−3 and 0.03, respectively). Association analysis by considering BMI as a continuous trait provided similar results (Supplementary Table 1).

We performed a replication analysis for rs3801266 (PBEF1), rs2069845 (IL6), rs1137100 (LEPR), and rs7514452 (IL6R) in stage 2. Associations of rs2069845, rs1137100, and rs3801266 were replicated (P = 7.9 × 10−4, 8.3 × 10−3, and 0.036, respectively) and corroborated in a subsequent meta-analysis (P = 2.3 × 10−6, 3.9 × 10−5, and 4.3 × 10−4, respectively) that remained significant after multiple testing correction (Table 2). Variants rs2069845, rs1137100, and rs3801266 also were significantly associated with BMI in stage 2 and meta-analysis and remained significant after multiple testing correction (Table 2). All three variants (rs2069845, rs1137100, and rs3801266) showed association with weight, WC, and HC (β range 0.05–0.14 z score units and P range 2.0 × 10−3 to 8.3 × 10−7) (Fig. 1). The association of rs1137100, rs2069845, and rs3801266 with weight, and of rs3801266 with HC remained significant after correcting for multiple comparisons (P < 5.4 × 10−5).

TABLE 2

SNPs showing the association with obesity and BMI in urban Indian children

SNPs showing the association with obesity and BMI in urban Indian children
SNPs showing the association with obesity and BMI in urban Indian children
FIG. 1.

Effect of overweight/obesity-associated SNPs on anthropometric measures (height, weight, BMI, WC, and HC). A: Effect of genotypes of associated SNPs on measures of adiposity. Mean z scores adjusted for age and sex are plotted on the y-axis, for the corresponding genotypes on the x-axis. B: z Score change per risk allele of associated SNPs (95% CI). Change in z scores was determined using regression analysis adjusted for age and sex.

FIG. 1.

Effect of overweight/obesity-associated SNPs on anthropometric measures (height, weight, BMI, WC, and HC). A: Effect of genotypes of associated SNPs on measures of adiposity. Mean z scores adjusted for age and sex are plotted on the y-axis, for the corresponding genotypes on the x-axis. B: z Score change per risk allele of associated SNPs (95% CI). Change in z scores was determined using regression analysis adjusted for age and sex.

Allele dosage analysis of rs2069845, rs1137100, and rs3801266 revealed a significant trend in the increase of risk for overweight/obesity by 1.32-fold, with an increase in each ERS (PTrend = 8.1 × 10−11) (Fig. 1). The distribution of NW and OW/OB children in different ERS groups is provided in Supplementary Fig. 1. Children with ERSs 5–6 had an OR of ~4 for being overweight/obese compared with individuals with an ERS <2 (P = 1.2 × 10−7). A significant increase in the measures of adiposity with the increase in ERS was observed, with the highest influence on weight and BMI, which both increased by 0.13 z score units per increase in ERS (P = 1.2 × 10−12 and 5.0 × 10−12) (Fig. 2).

FIG. 2.

Allele dosage analysis showing the association of combined risk alleles on childhood obesity and anthropometric measures. A: Obesity. B: Height. C: Weight. D: BMI. E: Waist circumference. F: Hip circumference. ORs or mean z scores with 95% CIs for combined effect are plotted on the y-axis, for the corresponding effective risk score on the x-axis.

FIG. 2.

Allele dosage analysis showing the association of combined risk alleles on childhood obesity and anthropometric measures. A: Obesity. B: Height. C: Weight. D: BMI. E: Waist circumference. F: Hip circumference. ORs or mean z scores with 95% CIs for combined effect are plotted on the y-axis, for the corresponding effective risk score on the x-axis.

Variants rs2069845, rs1137100, and rs3801266 did not show a significant association with the plasma levels of adipokines and markers of glucose homeostasis. In addition, we performed exploratory analysis to investigate the association of variants of these adipokines and inflammatory marker genes with quantitative clinical traits using stage 1 data (Supplementary Table 2). Variants in RETN (rs3745367 and rs1862513) and ADIPOQ (rs822395 and rs1063538) showed an association with the plasma levels of resistin (P = 2.7 × 10−5 and 2.0 × 10−3) and adiponectin (P = 2.0 × 10−3 for both), respectively. Only a nominal association of SNPs with markers of glucose homeostasis was observed.

Our study demonstrates the association of rs2069845 (IL6), rs1137100 (LEPR), and rs3801266 (PBEF1) with the risk of overweight/obesity, BMI, body weight, WC, and HC in urban Indian children. Allele dosage analysis revealed a cumulative effect of these variants on the risk of overweight/obesity and the effect size on measures of adiposity, which greatly increased with the increase in each risk allele.

Studies investigating the association of IL6 variants with adult and childhood obesity have generated contradictory results (79). A study in Greek school children had shown an association of the IL6 variant rs1800795 with parameters related to obesity (10). We did not detect an association of this variant with overweight/obesity; rather, we found an association of rs2069845 in intron 4 of IL6, which is in moderate linkage disequilibrium with rs1800795 (D' = 1.0; r2 = 0.63), with overweight/obesity and measures of adiposity in Indian children.

Mutations in LEPR are associated with a severe form of obesity (11); however, there have been conflicting reports regarding the association of its common variants with a complex form of obesity. LEPR coding variants (rs1137101 [Q223R] and rs8179183 [K656N]) have been inconsistently shown to be associated with adiposity in previous studies as a result of ethnic differences in allele frequencies (12). Here, we obtained a significant association of rs1137101, rs8179183, and rs1137100 (K109R) with childhood obesity. To the best of our knowledge, this is the first report demonstrating an association of K109R with obesity and adiposity measures in children. Although the functional significance of K109R is unknown, it causes a conservative change in the membrane-distal part of the leptin receptor extracellular domain (13) and thus might have a possible role in modulating the binding affinity of the receptor.

Visfatin, encoded by PBEF1, is a recently identified adipokine with a role in inflammation and insulin resistance (3). Only a rare variant in PBEF1 has been reported to be associated with severe obesity in French subjects (14). However, no common variants influencing the risk of obesity or its measures has been identified to date. In our study, the PBEF1 variant rs3801266 provided strong evidence for the association with overweight/obesity and anthropometric parameters in children.

Leptin, adiponectin, and resistin are the major adipocyte-secreted hormones with pleiotropic effects on metabolism, inflammation, and insulin resistance (3), the central factors underlying metabolic disorders. Genetic variants in genes encoding these adipokines are known to contribute to variations in their plasma levels. Consistent with the previous report of an association of rs3745367 of RETN with plasma resistin levels (15), we found strong evidence of an association of this polymorphism with plasma resistin levels. Our study also supports previous observations of an association of ADIPOQ variants with plasma adiponectin levels (1618). The widely studied P12A polymorphism of PPARG has been shown to be associated with various metabolic disorders, including obesity and type 2 diabetes in adults. The attempts to assess the role of P12A on the susceptibility to obesity in children have provided inconsistent reports (19,20). Our study did not detect any association of PPARG variants with childhood obesity. On a similar note, variants from LEP and RETN, which have been suggested to be associated with adult/childhood obesity (2124), were not found to be associated in our study.

Although the sample size of our study is considerable, there is a likelihood of false-negative observations for variants with small effect sizes, because the current study is sufficiently powered to capture only large effects (OR >1.3) of the common variants with frequencies >0.20. In addition, population stratification also can lead to spurious association results. To minimize the effect of population stratification, we have collected samples from a small geographical region that forms a homogenous cluster, as reported by the Indian Genome Variation Consortium (25). Moreover, the multidimensional scaling analysis using the genotype data for 595 unlinked markers (r2 < 0.20) for stage 1 samples shows that the study population belongs to one cluster (Supplementary Fig. 2).

In conclusion, we demonstrate that common variants of IL6, LEPR, and PBEF1 are associated with increased susceptibility to obesity and measures of adiposity in Indian school children. Although the case-control study design limits the potential to identify the causal relationships, our study provides a lead for future investigations toward understanding the contribution of inflammatory genes in genetic predisposition to obesity in childhood. This would help in understanding the molecular mechanisms and exploring of therapeutic options toward prevention of childhood obesity.

This study was supported by Diabetes Mellitus, New Drug Discovery Research and Development, Molecular Mechanisms, and Genetic and Epidemiological Factors (NWP0032-OB1), funded by the Council of Scientific and Industrial Research (CSIR), Government of India. R.T. is grateful to the Fogarty International Center and the Eunice Kennedy Shriver National Institute of Child Health and Human Development at the National Institutes of Health for providing the postdoctoral fellowship (grant no. 1-D43-HD-065249).

No potential conflicts of interest relevant to this article were reported.

R.T. researched data, contributed to the discussion, and wrote the manuscript. Y.M. researched data and contributed to the discussion. O.P.D., G.C., and D.B. researched data, contributed to the discussion, and reviewed and edited the manuscript. S.G. researched data. R.K.M. reviewed and edited the manuscript. N.T. contributed to the discussion and reviewed and edited the manuscript. D.B. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

The authors are thankful to all the participating subjects, their parents, and the school authorities for support and cooperation in carrying out the study. The authors thank Kuntal Bhadra from the Institute of Nuclear Medicine and Allied Sciences for his help in the sample collection. The authors also thank Dr. Abhay Sharma and Dr. Chetana Sachidanandan from CSIR–Institute of Genomics and Integrative Biology for their critical evaluation of the manuscript.

1.
Bhargava
SK
,
Sachdev
HS
,
Fall
CH
, et al
.
Relation of serial changes in childhood body-mass index to impaired glucose tolerance in young adulthood
.
N Engl J Med
2004
;
350
:
865
875
[PubMed]
2.
Bhardwaj
S
,
Misra
A
,
Khurana
L
,
Gulati
S
,
Shah
P
,
Vikram
NK
.
Childhood obesity in Asian Indians: a burgeoning cause of insulin resistance, diabetes and sub-clinical inflammation
.
Asia Pac J Clin Nutr
2008
;
17
(
Suppl. 1
):
172
175
[PubMed]
3.
Tilg
H
,
Moschen
AR
.
Adipocytokines: mediators linking adipose tissue, inflammation and immunity
.
Nat Rev Immunol
2006
;
6
:
772
783
[PubMed]
4.
Marwaha
RK
,
Tandon
N
,
Singh
Y
,
Aggarwal
R
,
Grewal
K
,
Mani
K
.
A study of growth parameters and prevalence of overweight and obesity in school children from delhi
.
Indian Pediatr
2006
;
43
:
943
952
[PubMed]
5.
Cole
TJ
,
Bellizzi
MC
,
Flegal
KM
,
Dietz
WH
.
Establishing a standard definition for child overweight and obesity worldwide: international survey
.
BMJ
2000
;
320
:
1240
1243
[PubMed]
6.
Chauhan
G
,
Spurgeon
CJ
,
Tabassum
R
, et al
.
Impact of common variants of PPARG, KCNJ11, TCF7L2, SLC30A8, HHEX, CDKN2A, IGF2BP2, and CDKAL1 on the risk of type 2 diabetes in 5,164 Indians
.
Diabetes
2010
;
59
:
2068
2074
[PubMed]
7.
Grallert
H
,
Huth
C
,
Kolz
M
, et al
.
IL-6 promoter polymorphisms and quantitative traits related to the metabolic syndrome in KORA S4
.
Exp Gerontol
2006
;
41
:
737
745
[PubMed]
8.
Huth
C
,
Illig
T
,
Herder
C
, et al
.
Joint analysis of individual participants’ data from 17 studies on the association of the IL6 variant -174G>C with circulating glucose levels, interleukin-6 levels, and body mass index
.
Ann Med
2009
;
41
:
128
138
[PubMed]
9.
Qi
L
,
Zhang
C
,
van Dam
RM
,
Hu
FB
.
Interleukin-6 genetic variability and adiposity: associations in two prospective cohorts and systematic review in 26,944 individuals
.
J Clin Endocrinol Metab
2007
;
92
:
3618
3625
[PubMed]
10.
Dedoussis
GV
,
Manios
Y
,
Choumerianou
DM
, et al
.
The IL-6 gene G-174C polymorphism related to health indices in Greek primary school children
.
Obes Res
2004
;
12
:
1037
1041
[PubMed]
11.
Farooqi
IS
,
Wangensteen
T
,
Collins
S
, et al
.
Clinical and molecular genetic spectrum of congenital deficiency of the leptin receptor
.
N Engl J Med
2007
;
356
:
237
247
[PubMed]
12.
Paracchini
V
,
Pedotti
P
,
Taioli
E
.
Genetics of leptin and obesity: a HuGE review
.
Am J Epidemiol
2005
;
162
:
101
114
[PubMed]
13.
Stratigopoulos
G
,
LeDuc
CA
,
Matsuoka
N
, et al
.
Functional consequences of the human leptin receptor (LEPR) Q223R transversion
.
Obesity (Silver Spring)
2009
;
17
:
126
135
[PubMed]
14.
Blakemore
AI
,
Meyre
D
,
Delplanque
J
, et al
.
A rare variant in the visfatin gene (NAMPT/PBEF1) is associated with protection from obesity
.
Obesity (Silver Spring)
2009
;
17
:
1549
1553
[PubMed]
15.
Menzaghi
C
,
Coco
A
,
Salvemini
L
, et al
.
Heritability of serum resistin and its genetic correlation with insulin resistance-related features in nondiabetic Caucasians
.
J Clin Endocrinol Metab
2006
;
91
:
2792
2795
[PubMed]
16.
Bouatia-Naji
N
,
Meyre
D
,
Lobbens
S
, et al
.
ACDC/adiponectin polymorphisms are associated with severe childhood and adult obesity
.
Diabetes
2006
;
55
:
545
550
[PubMed]
17.
Petrone
A
,
Zavarella
S
,
Caiazzo
A
, et al
.
The promoter region of the adiponectin gene is a determinant in modulating insulin sensitivity in childhood obesity
.
Obesity (Silver Spring)
2006
;
14
:
1498
1504
[PubMed]
18.
Richards
JB
,
Waterworth
D
,
O’Rahilly
S
, et al
;
GIANT Consortium
.
A genome-wide association study reveals variants in ARL15 that influence adiponectin levels
.
PLoS Genet
2009
;
5
:
e1000768
[PubMed]
19.
Lagou
V
,
Scott
RA
,
Manios
Y
, et al
.
Impact of peroxisome proliferator-activated receptors gamma and delta on adiposity in toddlers and preschoolers in the GENESIS Study
.
Obesity (Silver Spring)
2008
;
16
:
913
918
[PubMed]
20.
Ghoussaini
M
,
Meyre
D
,
Lobbens
S
, et al
.
Implication of the Pro12Ala polymorphism of the PPAR-gamma 2 gene in type 2 diabetes and obesity in the French population
.
BMC Med Genet
2005
;
6
:
11
[PubMed]
21.
Hinuy
HM
,
Hirata
MH
,
Sampaio
MF
, et al
.
Relationship between variants of the leptin gene and obesity and metabolic biomarkers in Brazilian individuals
.
Arq Bras Endocrinol Metabol
2010
;
54
:
282
288
[PubMed]
22.
Conneely
KN
,
Silander
K
,
Scott
LJ
, et al
.
Variation in the resistin gene is associated with obesity and insulin-related phenotypes in Finnish subjects
.
Diabetologia
2004
;
47
:
1782
1788
[PubMed]
23.
Engert
JC
,
Vohl
MC
,
Williams
SM
, et al
.
5′ flanking variants of resistin are associated with obesity
.
Diabetes
2002
;
51
:
1629
1634
[PubMed]
24.
Kongmacheep
P
,
Sirikulchayanonta
C
,
Tungtrongchitr
R
,
Hancharoen
K
.
Identification of leptin gene variants in school children with early onset obesity
.
J Med Assoc Thai
2009
;
92
(
Suppl. 7
):
S108
S114
[PubMed]
25.
Indian Genome Variation Consortium
.
Genetic landscape of the people of India: a canvas for disease gene exploration
.
J Genet
2008
;
87
:
3
20
[PubMed]

Supplementary data