Common genetic variation in and near the melanocortin 4 receptor gene (MC4R) is associated with body mass index in American Indian adults and children

Abstract

Six rare functional coding mutations were previously identified in melanocortin 4 receptor (MC4R) in 6,760 American Indians. Individuals heterozygous for one of these mutations become obese while young. We now investigate whether common non-coding variation near MC4R also contributes to obesity. Fifty-six tag single-nucleotide polymorphisms (SNPs) were genotyped in 3,229 full-heritage Pima Indians, and nine of these SNPs which showed evidence for association were genotyped in additional 3,852 mixed-heritage American Indians. Associations of SNPs with maximum body mass index (BMI) in adulthood (n = 5,918), BMI z score in childhood (n = 5,350), percent body fat (n = 864), energy expenditure (n = 358) and ad libitum food intake (n = 178) were assessed. Conditional analyses demonstrated that SNPs, rs74861148 and rs483125, were independently associated with BMI in adulthood (β = 0.68 kg/m(2) per risk allele, p = 5 × 10(-5); β = 0.58 kg/m(2), p = 0.002, respectively) and BMI z score in childhood (β = 0.05, p = 0.02; β = 0.07, p = 0.01, respectively). One haplotype (frequency = 0.35) of the G allele at rs74861148 and the A allele at rs483125 provided the strongest evidence for association with adult BMI (β = 0.89 kg/m(2), p = 5.5 × 10(-7)), and was also associated with childhood BMI z score (β = 0.08, p = 0.001). In addition, a promoter SNP rs11872992 was nominally associated with adult BMI (β = 0.61 kg/m(2), p = 0.05) and childhood BMI z score (β = 0.11, p = 0.01), where the risk allele also modestly decreased transcription in vitro by 12 % (p = 0.005). This risk allele was further associated with increased percent body fat (β = 2.2 %, p = 0.002), increased food intake (β = 676 kcal/day, p = 0.007) and decreased energy expenditure (β = -53.4 kcal/day, p = 0.054). Common and rare variation in MC4R contributes to obesity in American Indians.

Fig. 1

Additive effects on maximum BMI in American Indian adults (a) and maximum BMI z score in children (b) with increasing numbers of risk alleles of rs74861148 and rs483154. Only 3.3 % subjects carried non-risk alleles at both loci (#risk allele = 0), thus these subjects were combined with those who carried one risk allele from either locus (#risk allele = 1). Maximum BMI values (mean ± SE), allelic effects (95 % CI) and p values were adjusted for age, sex, birth year and heritage

Fig. 2

Haplotype effects of rs74861148 and rs483154 on maximum BMI in American Indian adults (a) and maximum BMI z score in children (b). Haplotype frequency is indicated in the parenthesis. Adult BMI and BMI z score are presented as an adjusted mean ± SE for each haplotype group. β represents the effect per copy of the risk haplotype. p values (additive model) and β (95 % CI) were adjusted for age, sex, birth year and heritage

Fig. 3

a Haplotype effects of rs74861148 and rs483154 on the lifetime BMI trajectory (age 5–50 years). b Haplotype effects of rs74861148 and rs483154 on the rate of BMI change in American Indian children and adults. The risk haplotype carriers (at least one copy) were compared with the non-risk haplotype carriers (no copies). BMI is given as an unadjusted mean at each age group. Rates of BMI change (mean ± SE), beta (95 % CI) and p values were adjusted for sex, birth year and heritage

Fig. 4

In vitro functional analyses of promoter activity of rs11872992. Relative luciferase activity (fold change) was expressed as a ratio of firefly luciferase activity to renilla luciferase activity, and further normalized to pGL3-basic luciferase activity. Data are presented as mean ± SD, n = 12 transfections. The statistical difference in the averaged activity was analyzed by an unpaired t test