MeQTL analysis of childhood obesity links epigenetics with a risk SNP rs17782313 near MC4R from meta-analysis

Earlier GWAS has identified that rs17782313 near MC4R was associated with obesity. However, subsequent studies showed conflicting results, especially among childhood. Besides, the mechanisms underlying the association between rs17782313 and childhood obesity remain largely unexplored, and genetic and epigenetic may interact and together affect the development of childhood obesity. We conducted a comprehensive meta-analysis to assess the association between rs17782313 and childhood obesity. MeQTL and eQTL analysis was applied to explore the effect of rs17782313 on DNA methylation and MC4R expression. We found that rs17782313 near MC4R was associated with increased childhood obesity risk and BMI z-score in several inheritable models (P < 0.05). Additionally, the similar trend was observed among subgroups of Asians, Caucasian. Furthermore, meQTL and eQTL analysis indicated that individuals carrying rs17782313 TT genotype were significantly associated with increased methylation level of cg10097150 located in MC4R promoter and decreased expression of MC4R than those with CT/CC genotype (P = 1.7 × 10−4 and P = 1.9 × 10−3 respectively). Our results strongly confirmed that rs17782313 was associated with increased risk of childhood obesity. Furthermore, rs17782313 T allele was correlated with promoter hypermethylation and decreased expression of MC4R, thus involved in the development of childhood obesity.


INTRODUCTION
Globally, obesity is becoming an increasingly serious clinical and public health challenge and is a major risk factor for various chronic diseases, such as cardiovascular diseases, type 2 diabetes, metabolic syndrome and several types of cancers [1,2]. The prevalence of obesity is up to 25% in western country and increasing in children [3,4]. Emerging evidence suggests that childhood obesity resulted in an increased mortality in adults [5]. Environmental factors, especially lifestyle and excessive food intake, are closely related to epidemic obesity. However, multiple twin studies have suggested that genetic factors explain about 65% of the variance in obesity [6,7]. Genome-wide association studies (GWASs) have identified a number of loci associated with body mass index (BMI), which is a conventional measure of obesity [8,9]. However, the biological functions of most well known genetic association with small effect size have not fully explained. The majority of GWAS loci are located within non-coding or intergenic regions and do not alter gene coding sequence [10,11]. It is assumed that GWAS polymorphisms are located in regulatory elements and influence the disease by changing gene expression, which is called expression quantitative trait loci (eQTL) [12,13].
Recently, it is becoming distinct that epigenetics, especial DNA methylation, also play a key role in obesity pathogenesis. Mounting evidences showed that the SNPs can affect methylation level at nearby CpG sites [14]. Such SNPs are known as methylation quantitative trait loci (meQTLs). The identification of meQTLs may assist in identifying novel candidate disease-associated genes and

Research Paper
Oncotarget 2801 www.impactjournals.com/oncotarget providing a new approach for connecting DNA sequence variation with phenotype [10].
Previous GWAS revealed that rs17782313 near the melanocortin 4 receptor (MC4R) gene was associated with increased risk of obesity in Europeans [15]. It is well established that MC4R plays a key role in coordinating energy homeostasis and body weight [16]. However, subsequent studies from different age groups have showed conflicting and inconclusive results, especially among childhood. Meanwhile, little is known of the molecular biological mechanisms underlying the effect of rs17782313 on the development of childhood obesity.
In this study, we applied a comprehensive metaanalysis to make an accurate assessment of the association of the MC4R rs17782313 with childhood obesity risk. Furthermore, most previous studies have investigated genetic and epigenetic mechanisms independently and their interaction was not considered. Our study aimed to combine genetic and epigenetic together to investigate the mechanisms that rs17782313 involved in the development of childhood obesity by using public Gene Expression Omnibus (GEO) and The Cancer Genome Atlas (TCGA) databases.

Meta-analysis result
A total of twelve articles including 5,908 cases of obesity and 19,826 controls were enrolled in our metaanalysis study. The detailed characteristics of studies included in this meta-analysis study are supplied in  Figure 1A). In the stratification analysis, an obvious association was observed between rs17782313 polymorphism and childhood obesity in subgroup of Caucasian and Asian in all inheritable models (P < 0.05). Moreover, the similar significant results were also found among subgroups of Taqman and other genotyping methods and sample size more than 1000, but not in the subgroup of sample size less than 1000 under the heterozygous model and recessive model (Table 2). Besides, we applied metaanalysis to evaluate the relationship between rs17782313 genotype and BMI z-score, and four articles were enrolled in the statistical analysis. The results indicated that rs17782313 T > C polymorphism was notably associated with childhood BMI z-score (WMD = 0.14, 95%CI = 0.06-0.23) ( Figure 1B). In addition, we obtained a boundary significant result in the replication of public GEO (GSE73103) database with limited samples (dominant model: OR = 1.24, 95%CI = 1.01-1.62, P = 0.049).

The impact of rs17782313 on MC4R expression
To investigate eQTL, we analyzed data from The Cancer Genome Atlas (TCGA) for whom both germline genotype from Affymetrix Genome-Wide Human SNP Array 6.0 and mRNA expression in intestines normal tissues from Illumina HiSeq 2000 RNA Sequencing Version 2 were available. ANOVA result revealed that individuals carrying rs17782313 TT genotypes were associated with decreased the expression of MC4R than those with CT/CC genotypes (P = 1.9 × 10 −3 ). (Figure 2).

Association between rs17782313 polymorphism with DNA methylation
Genotype data and correspondent methylation level data were available, which were requested from public GEO database (GSE73103). We extract CpG methylation probes located in 500 kb upstream and downstream from rs17782313. Associations between rs17782313 and methylation levels were tested using a linear model adjusted for age, sex, weight category and a proxy for blood cell type counts. Interestingly, we found that rs17782313 was significantly correlated with methylation level of cg10097150 which is located in the promoter of MC4R (P = 1.7 × 10 −4 ). Here, Individuals with the rs17782313 TT genotype have an obvious hypermethylation level in MC4R promoter than those with CT/CC genotype ( Figure 2).

DISCUSSION
Recently, the GWAS have identified rs17782313 was associated with BMI and obesity in Caucasians. However, the relationship of rs17782313 with obesity among other ethnicities are not fully explored, especially in children. In this study, the results of meta-analysis strongly confirmed that rs17782313 near MC4R gene was significantly associated with increased risk of childhood obesity in four inheritable models ( Further stratification analyses showed that the similar trend association was also observed in subgroup of Asians, Caucasian, sample size more than 1000. Besides, MC4R rs17782313 was boundary correlated with childhood obesity in the replication of available public GEO database (dominant model: OR = 1.24, 95%CI = 1.01-1.62, P = 0.049). In addition, another meta-analysis indicated that rs17782313 T > C polymorphism was notably associated with childhood BMI z-score (WMD = 0.14, 95%CI = 0.06-0.23).
SNP rs17782313 mapped 188kb downstream of MC4R, which is well established as a candidate gene for obesity. MC4R encoded a membrane-bound receptor www.impactjournals.com/oncotarget and member of the melanocortin receptor family protein.
Multiple lines of studies suggested that MC4R might be involved in regulating energy homeostasis. Activation of MC4R can decrease food intake and increase energy expenditure, thereby leading to reduce fat stores [17]. The MC4R signaling was identified to regulate the sympathetic nerve activity and glucose tolerance [18]. In addition, the SNPs in MC4R play an important role in the susceptibility to obesity. Rare functional mutations in MC4R are known to develop childhood-onset obesity [19][20][21][22]. Furthermore, mutations leading to complete loss of function are correlated with a more severe phenotype [23], and targeted disruption of the melanocortin-4 receptor also results in obesity in mice [24]. One study has revealed that common genetic variations near or in the MC4R contributes to obesity in American Indians [25]. However, most of GWAS loci are common variants, which located in intergenic and non-coding regions and do not alter change coding sequence [26,27]. Earlier studies showed that GWAS-identified genetic risk for complex traits often affect phenotype by changing the mount of protein production, rather than altering the type of protein production [28]. We hypothesis the common SNP rs17782313 may influence transcriptional level of MC4R, even though the association signal located 188 kb downstream of the MC4R coding sequence.
MC4R expression are available from public TCGA database, and we obtained evidence that rs17782313 was notably associated with the expression level of MC4R.Our result showed that the expression of MC4R in individuals carrying rs17782313 TT genotype are significantly lower than that in CT/CC genotype individuals (P = 1.9 × 10 −3 ).
Given lack of direct functional evidence relating rs17782313 to MC4R expression, we assume it alter  the methylation level contributing to gene expression. Epigenetic, particular DNA methylation, is a crucial mechanisms of transcriptional regulation that occur without an alteration of the involving DNA sequence [29,30]. Previous study indicated that gene silencing in cancer are associated with promoter Hypermethylation [31], and strongly correlations were observed between many obesity-associated SNPs and DNA methylation  Oncotarget 2804 www.impactjournals.com/oncotarget alternation at proximal promoters and enhancers [32]. It was reported that genotype-specific methylation in the promoter of CHRNB4 affected by SNPs that were remarkably corrected with lung cancer risk [33]. Another study found that rs11636753 G allele was significantly associated with decreased CHRNA5 DNA methylation, lower CHRNA5 expression and increased the risk of nicotine dependence [34].
In this study, we apply a method of integrated experiments to explore the effect of common genetic variant rs17782313 on DNA methylation from available GEO database. Interestingly, linear regression revealed the prominent association between rs17782313 and methylation level of cg10097150 which is located in the promoter of MC4R (P = 1.7 × 10 −4 ). Or rather, individuals carrying rs17782313 TT have a remarkably hypermethylation level in MC4R promoter than those with CT/CC genotype. Altogether, our results provide evidence that rs17782313 may affect the methylation level of MC4R, which contribute to regulate MC4R expression, thus involved in the development of childhood obesity.
There are some limitations in our study. First, our analysis is based on public database and studies with greater sample size are needed to validate our findings.
Additionally, the interaction between environmental exposures and genetic and epigenetic mechanisms should be considered in the following research.
In summary, our meta-analysis of data from current and published studies confirms the significant association between MC4R polymorphism and the risk of childhood obesity and BMI z-score. Furthermore, eQTL and meQTL results strongly indicate that obesity-associated rs17782313 T allele was significantly associated with promoter hypermethylation and decreased expression of MC4R, thus involved in the childhood obesity. This study highlights that genetic and epigenetic mechanism in MC4R promoter may interact and together affect functional relevance for MC4R expression and development of childhood obesity.

Literature and search strategy
We search relevant articles published in English from following databases, including PubMed and Embase. The search strategy to identify all possible studies involved the use of the following keywords: (MC4R) and (polymorphism OR variant OR genotype) and (obesity). References of the retrieved articles were hand-searched.

Inclusion criteria and data extraction
The enrolled studies met the following inclusion criteria: (1) Case-control, cross-sectional, cohort studies which evaluated the association of MC4R rs17782313 with obesity. Cases diagnosed by the WHO standards of childhood obesity patients. Height-standard weight method was applied among age less than 10-year-old and BMI (Body mass index) was used for 10 ~18 years old children to assess obesity levels. The control group was composed of normal healthy children; (2) Providing an odds ratio (OR) with 95% confidence interval under various inheritable models or sufficient raw data to calculate these estimates; (3) Reliable and high-quality data by using appropriate statistical analysis method; (4) If data were repetitively reported, the study with biggest sample size was included. Two authors independently assessed the articles for compliance with the inclusion criteria, and inconsistent were resolved by agreeing with the third reviewer.

The cancer genome atlas (TCGA) and gene expression omnibus (GEO) databases
Gene expression profiles were downloaded from TCGA project by RNA-Seq (level 3). In total, 434 colon adenocarcinoma tissues and 41 normal colon samples were included. To control for potential batch effects of mRNA expressions, a series of normalizations and corrections were applied. Briefly, level 3 mRNA expression of each gene was log2 transformed if it was not normally distributed, and genes with zero values were removed. We also accessed TCGA individual level 2 SNP data from tissues and blood, which were genotyped with an Affymetrix Human Genome Wide SNP 6.0 array. The genotype of rs17782313 and corresponding methylation levels were available from GEO database (GSE73103). A total of 355 healthy individuals were genotyped and obtained DNA methylation levels in their blood using the Illuminal 450K BeadChip.

Statistical analysis
Meta-analysis was conducted by STATA version 11 (StataCorp LP, College Station, TX, USA). Heterogeneity was evaluated by Cochran'Q and I 2 statistics. P het ≤ 0.10 and I 2 values > 75% were considered to be significant heterogeneity and we used randomeffect model to calculate combined effect, otherwise we applied fixed-effect model. P < 0.05 was used as threshold to exclude genotype data which did not meet Hardy-Weinberg equivalent. Data were pooled across studies, and OR (odds ratio) and 95%CI (confidence interval) for categorical variables and WMD (weighted mean difference) for continuous outcomes, weighted according to study sample size, were calculated. We applied linear regression to evaluate the relationship between rs17782313 risk allele and gene epigenetic and expression alterations from public GEO and TCGA database by using R software. All statistical analyses were two-sided, and a P value < 0.05 was considered statistically significant.