Original Article

International Journal of Obesity (2012) 36, 465–473; doi:10.1038/ijo.2011.131; published online 12 July 2011

Implication of European-derived adiposity loci in African Americans

J M Hester1,2,3, M R Wing1,2,3, J Li1,2, N D Palmer1,2,4, J Xu1,2, P J Hicks1,2,4, B H Roh1,2,4, J M Norris5, L E Wagenknecht6, C D Langefeld7, B I Freedman8, D W Bowden1,2,4,8 and M C Y Ng1,2

  1. 1Center for Diabetes Research, Wake Forest University School of Medicine, Winston-Salem, NC, USA
  2. 2Center for Genomics and Personalized Medicine Research, Wake Forest University School of Medicine, Winston-Salem, NC, USA
  3. 3Program in Molecular Genetics and Genomics, Wake Forest University School of Medicine, Winston-Salem, NC, USA
  4. 4Department of Biochemistry, Wake Forest University School of Medicine, Winston-Salem, NC, USA
  5. 5Department of Epidemiology, Colorado School of Public Health, University of Colorado Denver, Aurora, CO, USA
  6. 6Division of Public Health Sciences, Wake Forest University School of Medicine, Winston-Salem, NC, USA
  7. 7Department of Biostatistical Sciences, Wake Forest University School of Medicine, Winston-Salem, NC, USA
  8. 8Department of Internal Medicine, Wake Forest University School of Medicine, Winston-Salem, NC, USA

Correspondence: Dr MCY Ng, Center for Diabetes Research, Center for Genomics and Personalized Medicine Research, Wake Forest University School of Medicine, Medical Center Blvd., Winston-Salem, NC 27157, USA. E-mail: mng@wfubmc.edu

Received 29 November 2010; Revised 19 May 2011; Accepted 21 May 2011
Advance online publication 12 July 2011





Recent genome-wide association studies (GWAS) have identified multiple novel loci associated with adiposity in European-derived study populations. Limited study of these loci has been reported in African Americans. Here we examined the effects of these previously identified adiposity loci in African Americans.



A total of 46 representative single-nucleotide polymorphisms (SNPs) in 19 loci that were previously reported in GWAS in Europeans (including FTO and MC4R) were genotyped in 4992 subjects from six African-American cohorts. These SNPs were tested for association with body mass index (BMI) after adjustment for age, gender, disease status and population structure in each cohort. Meta-analysis was conducted to combine the results.



Meta-analysis of 4992 subjects revealed seven SNPs near four loci, including NEGR1, TMEM18, SH2B1 /ATP2A1 and MC4R, showing significant association at 0.005<P<0.05, and had effect sizes between 0.04 and 0.06s.d. units (or 0.30 to 0.44kgm−2) of BMI for each copy of the BMI-increasing allele. The most significantly associated SNPs (rs9424977, rs3101336 and rs2568958) are located in the NEGR1 gene (P=0.005, 0.020 and 0.019, respectively).



We replicated the association of variants at four loci in six African-American cohorts that demonstrated a consistent direction of association with previous studies of adiposity in Europeans. These loci are all highly expressed in the brain, consistent with an important role for central nervous system processes in weight regulation. However, further comprehensive examination of these regions may be necessary to fine map and elucidate for possible genetic differences between these two populations.


body mass index; genome-wide association; African Americans



Obesity is a major public health problem leading to increased mortality and comorbidities, such as type 2 diabetes (T2DM), metabolic syndrome, coronary heart disease, stroke, cancers, liver and gallbladder disease, sleep disorders and osteoarthritis.1 The World Health Organization defines overweight, obese and morbid obese in adults as body mass index (BMI) between 25 to 30kgm−2, 30 to 40kgm−2 and over 40kgm−2, respectively. The prevalence of obesity in both adults and adolescents has been increasing in the past decades in the United States. In fact, obesity incidence has doubled in adults aged 20 years or older from 1980 to 2002.2, 3

Marked racial and gender differences in the prevalence of obesity have been observed. In the National Health and Nutrition Examination Survey conducted in 2003–2004, 31% of European American adults were obese and 4% were morbidly obese. The corresponding figures in African Americans were even more alarming (45% and 11%, respectively), with higher prevalence of obesity in black women (54%) than in black men (34%).4

The increasing prevalence of obesity is contributed by the excessive caloric intake and diminished physical activity in the modern environment. However, genetic factors modulate the impact of the affluent environment on each individual. Considerable evidence from familial segregation and twin studies suggest that there is a significant genetic contribution to adiposity.5 Heritability estimates for BMI have been reported between 60–90% in African Americans.6, 7, 8

Recently, large-scale genome-wide association studies (GWAS) and meta-analyses in Europeans have revealed over 40 novel adiposity loci associated with BMI, waist circumference and/or waist–hip ratio. The strongest adiposity locus identified to date is FTO.9, 10, 11, 12 Additional loci were subsequently identified to be located at, or near the genes BCDIN3D/FAIM2, CDH12, CHST8/KCTD15, CTNNBL1, GNPDA2, LGR4/LIN7C/BDNF, LYPlAL1, MAF, MC4R, MSRA, MTCH2, NEGR1, NPC1, NRXN3, PARD3B, PCSK1, PRL, PTER, SEC16B/RASAL2, SFRS10/ETV5/DGKG, SH2B1/ATP2A1, TFAP2B and TMEM18.13, 14, 15, 16, 17, 18, 19, 20, 21, 22 Several of these loci have been confirmed in Asian populations by GWAS and replication studies.23, 24, 25, 26 Replication studies in African Americans have revealed inconsistent evidence of association of MC4R27, 28 and FTO29, 30, 31, 32 with adiposity measures, and limited data is available for other loci. In the present study, we examined the influence of single-nucleotide polymorphisms (SNPs) at the loci recently identified in GWAS on adiposity in multiple African-American populations.13, 14, 15, 16, 17, 18, 19, 20


Subjects and methods


Six African-American cohorts were used in the present study. Cohorts 1 and 2 were derived from a type 2 diabetic nephropathy GWAS. The community non-diabetic cohort (cohort 1) consisted of 816 subjects who reported no history of diabetes and who were recruited from the community and internal medicine clinics at Wake Forest University School of Medicine. Cohort 2 consisted of 899 subjects with T2DM and end-stage renal disease (T2DM–ESRD), recruited from dialysis facilities in the southeastern US.33 An additional community non-diabetic cohort (cohort 3), including 621 subjects (616 unrelated subjects and 5 related subjects from two nuclear families) who reported no history of diabetes, was recruited from the community and internal medicine clinics similar to that of cohort 1. A second diabetic cohort (cohort 4) consisting of 891 subjects with T2DM and 617 subjects with T2DM–ESRD (1005 unrelated subjects and 503 related subjects from 178 nuclear families) was recruited from the community, churches, health fairs, medical clinics and dialysis facilities. The Diabetes Heart Studies cohort (cohort 5) consisted of subjects recruited from the community and internal medicine clinics in two studies that examined the sub-clinical cardiovascular risk in T2DM. A subset of 211 unrelated subjects from the African American Diabetes Heart Study34 and 81 subjects from the family-based Diabetes Heart Study35 were included in this study. All subjects from cohorts 1 to 5 were recruited in North Carolina, South Carolina, Georgia, Tennessee or Virginia. The Insulin Resistance Atherosclerosis (IRAS) Studies cohort (cohort 6) consisted of subjects recruited from two multi-center community-based cohort studies, the IRAS Study36 and the IRAS Family Study,37 designed to examine the epidemiology and genetics of glucose homeostasis traits, respectively. Included are 575 related subjects from 42 families of the IRAS Family Study, recruited from Los Angeles, CA, USA, and 278 unrelated subjects from the IRAS study, recruited from Los Angeles and Oakland, CA, USA.

The clinical characteristics of all cohorts are summarized in Table 1. Informed consent was obtained from all study participants. Recruitment and sample collection procedures for cohorts 1 to 5 and cohort 6 were approved by the Institutional Review Boards at Wake Forest University School of Medicine and the local institutions, respectively.

Clinical studies

Height and weight was measured in all study subjects. BMI is calculated as weight divided by square of height. BMIgreater than or equal to30kgm−2 is considered obese. Genomic DNA was extracted from blood samples using the PureGene system (Gentra Systems, Minneapolis, MN, USA).

Genotyping and quality control

A total of 46 SNPs at 19 novel adiposity loci, identified through recent GWAS in Europeans, were selected for replication in our African-American cohorts. Selection criteria included key SNPs that reached genome-wide significance (P<5 × 10−8) and nearby reported SNPs showing nominal association in these previous studies.10, 13, 14, 15, 16, 17, 18, 19, 20 For loci with multiple associated SNPs, the pair-wise linkage disequilibrium (LD) D′ and r2 in HapMap Phase II Yoruba were assessed using Haploview (http://www.broadinstitute.org/haploview). Only representative SNPs with r2<0.8 were selected for genotyping (Supplementary Table 1). Cohorts 1 and 2 were genotyped at the Center for Inherited Disease Research, using 1μg of genomic DNA on the Affymetrix Genome-wide Human SNP array 6.0 (Affymetrix, Santa Clara, CA, USA). Genotypes were called using Birdseed version 2, APT 1.10.0 (Broad Institute, Cambridge, MA, USA), and samples were grouped by DNA plate to determine the genotype cluster boundaries. Genotyping of samples from cohorts 3 to 6 was performed using the iPLEX Sequenom MassARRAY platform (San Diego, CA, USA). The minimum and average SNP call rates for all cohorts were 95% and 98.1%, respectively. The average genotype concordance rate of 45 blind duplicates was 99.8%. All SNPs had Hardy–Weinberg P-value greater than or equal to0.001 in the combined unrelated samples. For related samples, genotype data identified with Mendelian inconsistency by PedCheck (v. 1.1)38 were removed.

Imputation of genotypes for cohorts 1 and 2

Imputation was performed for autosomes using MACH (version 1.0.16, http://www.sph.umich.edu/csg/abecasis/MaCH/) to obtain missing genotypes for cohorts 1 and 2 that underwent GWAS. SNPs with minor allele frequency greater than or equal to1%, call rate greater than or equal to95% and Hardy–Weinberg P-value greater than or equal to10−4 were used for imputation. A 1:1 mixture of the HapMap II release 22 (NCBI build 36) CEU:YRI consensus haplotypes (http://mathgen.stats.ox.ac.uk/impute/) were used as a reference panel. Imputation was performed in two steps. For the first step, 484 unrelated African-American samples were randomly selected to calculate recombination and error rate estimates. In the second step, these rates were used to impute all samples across the SNPs in the entire reference panel. Imputation results were filtered at an rsq threshold of greater than or equal to0.3 and a minor allele frequencygreater than or equal to0.05. Among the 46 SNPs examined in this study, 21 SNPs were directly genotyped, whereas 24 SNPs were imputed with a minimum rsq of 0.48, and one SNP (rs4712652) failed for both direct genotyping and imputation and was not analyzed in cohorts 1 and 2.

Population structure

To account for the effect of population structure on genetic association in these African-American samples, principal components analysis was computed on cohorts 1 and 2 with GWAS data by using all SNPs that passed quality control standards and after exclusion of regions of high LD and inversions. The first principal component (PC1) explained the largest proportion of genetic variation (22%). DNA samples from all cohorts, as well as 44 Yoruba Nigerians and 39 European Americans were genotyped for 77 ancestry informative markers. The African to European ancestral proportion of each African-American sample was then estimated using the Expectation-Maximization algorithm implemented in the program frequentist estimation of individual ancestry proportion (FRAPPE) under a two-population model.39 PC1 was highly correlated with ancestry informative markers (r2=−0.87), suggesting that PC1 largely reflected the ancestry proportions, and was used as a covariate for association analysis of cohorts 1 and 2. The mean (±s.d.) African ancestry proportions estimated by frequentist estimation of FRAPPE in cohorts 1 to 6 were 0.77±0.12, 0.78±0.12, 0.77±0.12, 0.78±0.13, 0.76±0.11 and 0.69±0.14, respectively.

Statistical analyses

Data are presented as mean±s.d. or percentage, as appropriate. BMI was natural logarithmically transformed to best approximate-conditional normality and homogeneity of variance, conditional on cohort, disease status, age and gender. To account for potential gender and cohort differences in the distribution of BMI, subjects were stratified by gender in each cohort and by disease status (ESRD in cohort 4 and T2DM in cohort 6). Within each strata, an individual was considered an outlier if the data was outside of four s.d. A total of three outlier observations were removed from further analyses. The data were then adjusted for age in a linear model and residuals were standardized to a mean of zero and variance of one (that is, Z-score). These Z-scores are the primary unit of analysis for both within individual cohorts and the meta-analyses.


For cohorts 1 and 2, the associations of BMI Z-scores with SNPs were tested by linear regression under an additive model using the program QSNPGWA (www.phs.wfubmc.edu), with additional covariate adjustment for PC1. For cohorts 3–6, a variance component measured genotype method implemented in SOLAR (Almasy and Blangero 1998, http://txbiomed.org/departments/genetics/) was used for association tests in order to account for familial relationships within each cohort. Associations of BMI Z-scores with SNPs were tested under an additive model with adjustment for proportion of African ancestry. Familial correlation was accounted for using a kinship coefficient matrix, in which a correlation was calculated for each set of related pairs.


Association results from all six cohorts were combined using the inverse variance weighted method implemented in METAL (http://www.sph.umich.edu/csg/abecasis/metal/). In order to account for modest relatedness between the cohorts 1–4, BMI Z-scores of all six cohorts were pooled and analyzed together, using variance component method for comparison with the meta-analysis method. To evaluate the potential confounding effect of disease, 2111 community non-diabetic subjects from cohorts 1, 3 and 6, 1516 T2DM–ESRD subjects from cohorts 2 and 4, and 1362 T2DM subjects from cohorts 4, 5 and 6, were separately analyzed using linear regression and variance component methods, as appropriate. The association results were then combined using METAL to assess the overall SNP association in non-diabetics, T2DM–ESRD and T2DM separately. For loci showing multiple associations, conditional SNP analyses that include multiple SNPs as independent variables were performed, using variance component method in the pooled samples to evaluate the independence of the association signals.

All statistical tests were performed by QSNPGWA, PLINK (http://pngu.mgh.harvard.edu/~purcell/plink/) or SAS v.9.1 (SAS Institute, Cary, NC, USA) unless specified otherwise. Posterior study power was calculated using genetic power calculator.40 A nominal P-value <0.05 in the meta-analysis was considered as evidence of significance. To correct for multiple testing, a P-value <0.0011 (0.05/46 SNPs) was considered significant after Bonferroni correction.



Clinical characteristics of the study samples

The clinical characteristics of the study samples in all six cohorts are shown in Table 1. Due to the nature of sample collection, the cohorts consisted of community non-diabetic (cohorts 1 and 3), community-based (cohort 6) or diabetic subjects (cohorts 2, 4 and 5). The distributions of BMI are comparable among all cohorts, except for higher mean BMI (35.3kgm−2) and higher prevalence of obesity (74.3%) in the Diabetes Heart Studies (cohort 5). In view of the heterogeneous phenotypes among the study cohorts, BMI Z-scores were calculated in sex- and disease-specific strata and then combined in each cohort separately for association analyses.

We either genotyped or imputed (for 24 SNPs in cohorts 1 and 2 only) 46 representative SNPs in 19 adiposity loci reported in recent GWAS in six African-American cohorts. Table 2 summarizes the individual cohort and meta-analysis results of these SNPs for association with BMI, using an inverse variance method. From the results of the meta-analysis, seven SNPs showed significant association at 0.005<P<0.05 with effect sizes between 0.04 and 0.06 s.d. units (or 0.30 to 0.44kgm−2) of BMI for each copy of effect allele. These seven SNPs were located at, or near four different loci including, NEGR1, TMEM18, SH2B1/ATP2A1 and MC4R. The most significantly associated SNPs (rs9424977, rs3101336 and rs2568958) are located on chromosome 1 near NEGR1 (0.005<P<0.02). The T-allele of rs9424977 showed a trend of association with increased BMI in four of the six cohorts. The rs3101336 and rs2568958 are highly correlated with each other (r2=0.99), but moderately correlated with rs9424977 (r2=0.55) in our samples. Conditional analysis including both rs9424977 and rs3101336 removed the significant associations (P=0.175 for rs9424977 and 0.676 for rs3101336) and suggested that they represent the same association signal. Two SNPs located near the MC4R gene were associated with BMI in the meta-analysis (P=0.018 for rs477181; P=0.046 for rs4450508). The G-alleles of both rs477181 and rs4450508 showed a consistent trend of increased BMI across most of the cohorts. These two SNPs reside in the same LD block (D=0.97), but are not highly correlated (r2=0.35) in our samples. Conditional analysis including rs477181 and rs4450508 removed the significant associations (P=0.152 for rs477181 and 0.420 for rs4450508), again suggesting that they represent the same association signal. There was no evidence of heterogeneity of the effect sizes of all seven significantly associated SNPs across the six cohorts (Pheterogeneity>0.05). None of these SNPs showed significant association (all P>0.0011) after Bonferroni correction for multiple testing.

We further investigated the association of these 46 SNPs in 2111 community non-diabetic, 1516 T2DM–ESRD and 1362 T2DM subjects from cohorts 1 to 6 (Supplementary Table 2). The effect sizes ranged from 0.0005 to 0.24, 0.003 to 0.11, and 0.002 to 0.16 s.d. units, respectively, for the three groups. Overall, there was no significant heterogeneity of effect sizes among the non-diabetic, T2DM–ESRD and T2DM groups (Pheterogeneity>0.05), except for rs10508503 in PTER. The non-significant results for most SNPs in this subset analysis are likely to be attributed to reduced sample size and study power. Due to modest relatedness between the cohorts, the association analyses of the 46 SNPs were repeated, using variance component method to combine all samples into one single group (Supplementary Table 3). The P-value results are highly correlated with the meta-analysis method (slope=0.965, intercept=0.024 by linear regression model).



We examined 46 SNPs in 19 loci identified in recent GWAS in European-derived populations for association with obesity in six African-American cohorts. Meta-analysis revealed seven SNPs (rs94224977, rs3101336, rs2568958, rs2867125, rs7498665, rs477181 and rs4450508) located at, or near the four loci (NEGR1, TMEM18, SH2B1 and MC4R), showing suggestive association with BMI (0.005<P<0.05) (Table 2). Although none of the SNPs showed significant association across all six cohorts due to modest sample size, the consistent direction of association with the European data may indicate their role in the modulation of BMI levels in African Americans.

The NEGR1 gene is located on chromosome 1p31 and is involved in the regulation of neurite outgrowth in the developing brain.41, 42 In the NEGR1 region, the three most significantly associated SNPs from our study (rs9424977, rs3101336 and rs2568958) are located in the same LD block with variable correlations (D=0.82–1, r2=0.35–0.99). In European-derived populations, the LD pattern of these SNPs is similar to that observed in our samples, and the reported effect sizes are between 0.025 and 0.038 s.d. unit of BMI. The effect sizes of meta-analysis results in our African-American cohorts were substantially larger (0.05–0.06 s.d. unit of BMI). The direction of association is the same between the two populations (Supplementary Table 4).

The TMEM18 gene located at chromosome 2p25, has recently been identified as a modulator of glioma-directed stem cell migration and may be involved in cell movement in general.43 The encoded protein is localized to the nucleus, widely expressed in fetal and adult tissues and well conserved among divergent species. TMEM18 contains one associated SNP from our study, rs2867125. The reported effect size in Europeans for rs2867125 is 0.061 s.d. unit of BMI,17 which is comparable to the effect size of 0.06 s.d. unit of BMI observed in our study, with same direction of the association.

The SH2B1 gene located at 16p11.2 has previously been shown to associate with increased serum leptin, total fat and waist circumference, and is a strong prior candidate for regulating body weight.44 Our study confirms rs7498665, a non-synonymous SNP (T484A), to be associated with BMI in African Americans, with higher effect size and same direction of association, as compared with the Europeans (β=0.06 versus 0.036 s.d. unit of BMI, respectively).17

In humans, multiple rare mutations conferring loss of function in the MC4R gene are associated with hyperphagia, severe childhood obesity and hyperinsulinemia.45 Experimental studies show that MC4R is a key regulator of energy balance, influencing food intake and energy expenditure through functionally divergent central melanocortin neuronal pathways.46 MC4R contains two SNPs that were significantly associated with BMI in our study. The rs477181 and rs4450508 are located in the same LD block (D=0.97) at chromosome 18q21, with effect sizes of 0.05 and 0.04 s.d. unit of BMI, respectively. Although the G-allele of rs477181 was associated with BMI in both the Europeans and our African-American samples, however, the direction of association for rs4450508 differs between the two populations (Supplementary Table 4).

Other studies have examined the influence of MC4R in African-derived populations. Grant et al.27 evaluated the MC4R locus in a cohort of 4688 European American children and 3723 African-American children. The rs571312, rs10871777 and rs476828 (perfect surrogates for rs17782313, an SNP widely replicated in European populations) yielded odds ratios in the European American cohort of 1.137–1.145 (0.042<P<0.054) for obesity, but there was no significant association with these SNPs in the African-American cohort. However, they observed significant association with rs1942880 (P=0.008) and rs12457166 (P=0.013), which are located in the same LD region with rs633265 (r2=0.22, D=0.9, P=0.45) and rs477181 (r2=0.11, D=1, P=0.018) reported in this study. The rs17782313 shows a trend towards the association with BMI in our study (P=0.08); however, there was significant heterogeneity in the effect sizes across the six cohorts (Pheterogeneity=0.007). Additionally, the T-allele was associated with increased BMI in our study as opposed to the C-allele reported in the European populations,14 though allele frequencies are similar between the two populations (Supplementary Table 4). Both sample heterogeneity and the difference in effect allele could contribute to the lack of association for this SNP in our study. Recently Kang et al.28 reported a GWAS and follow-up analysis of anthropometric traits in African-derived populations. No SNPs reached a genome-wide significance level for association with BMI. However, rs6567160 in MC4R showed evidence of replication in their GWA panel (P=0.0035). Rs6567160 is in high LD with rs633265 and rs1942880 (r2=0.93, D0.98). In Europeans, the extent of LD block containing associated SNPs in the MC4R gene region is greater (18kb), compared with that observed in African Americans (13kb). Therefore, the causal variant(s) may be in high LD in European-derived, but not in African-derived populations, due to differences in LD structure. This feature of African-derived genomes may facilitate identification of the true risk variant. We have recently used such an approach to strongly implicate a single SNP as the risk variant in the TCF7L2 diabetes gene.47

In European-derived populations, one of the earliest, and by far the strongest, adiposity genes identified by GWAS is FTO.11 Additional GWAS and replication studies on adiposity, childhood and adult obesity confirmed the association of FTO with adiposity and obesity in Europeans,9, 12, 16, 17 as well as in Asians.23, 24, 25 However, recent studies in African Americans have revealed inconclusive results.12, 29, 30, 31, 32 FTO is a relatively large gene, approximately 430 kilobases (kb) in length, containing nine exons. Association studies have focused primarily on intron 1, around 47kb in length, where it is believed that the causal variant resides. As may be expected, the LD structure of FTO differs between European- and African-derived populations, with a greater number of regions of high LD and larger LD blocks observed in individuals of European descent compared with individuals of African descent. Three SNPs in particular (rs3751812, rs8050136 and rs9939609) have been widely studied in Europeans and included in many replication studies in other populations, including African Americans. These SNPs reside in the same LD block (r2=0.60, HapMap YRI), located within intron 1 of FTO. However, the inconclusive findings in African Americans do not point to a definitive risk variant in this population.

Despite some success in recent replication efforts involving FTO, there is a lack of evidence for association with BMI in the African-American populations from our study. We genotyped 10 index SNPs located across the FTO region in all six cohorts. We were unable to replicate association at any of these loci, though the direction of association was consistent with previous studies in Europeans (Supplementary Table 4). These negative results may partly be due to a lack of study power. The reported effect size for the A-allele of rs8050136 in Europeans is 0.08 s.d. unit of BMI,17 and given the respective allele frequency of 0.45 in our African-American samples, we would have 98 and 75% power to detect association at α-level of 0.05 and 0.0011, respectively, under an additive model. However, given our sample size, allele frequency and the effect size in our study of 0.03 s.d. unit of BMI for rs8050136, we had only 32 and 3.6% power to detect association at α-level of 0.05 and 0.0011, respectively. On the other hand, our strongest association observed was for rs9424977 at NEGR1 (P=0.005 in the meta-analysis, Table 2). The effect size of this SNP in our samples is 0.06 s.d. unit of BMI, and given the BMI-increasing allele frequency of 0.46, we had an estimated 85 and 39% power to detect association at α-level of 0.05 and 0.0011, respectively. In general, we have good power (greater than or equal to80%) to detect an effect size of 0.06 s.d. unit of BMI with an allele frequency of at least 0.40 at α-level of 0.05. However, after correction for multiple testing, we would need more than 10000 samples to have comparable power. Given the prior association of these SNPs in European studies, we primarily presented significance levels without multiple comparisons to enhance study power. Our findings may reflect both allelic and locus heterogeneity in African Americans as compared with the European-derived populations. Larger sample sizes and thorough examination of these loci will be required to determine their effects in African Americans.

Although we were able to replicate moderate association with BMI at four loci in our study, some of the discrepant findings between our study and those performed in European populations could be due to several factors. (1) The study power of our samples may be limited. There are a substantial number of associated SNPs that showed large differences in allele frequencies between the ancestral HapMap European (CEU) and African (YRI) populations (Supplementary Table 4). Some of the SNPs, such as those in PRL, PTER, PCSK1 and MC4R are, or nearly are monomorphic in the YRI population, which will have little, if any, power to be detected. (2) Allelic heterogeneity and different LD patterns such that there may be different causal variants or that different set of SNPs tagging the same causal variant(s) in Europeans versus African Americans. (3) Locus heterogeneity in which there may be novel loci affecting adiposity in African Americans. (4) Complex and/or differential gene-gene and gene-environment interaction that modify the genetic risk of individual locus.

In order to increase sample size and study power, we used both healthy subjects and subjects with T2DM or T2DM–ESRD. As obesity is a risk factor of diabetes, it is unclear if the presence of T2DM or ESRD will in turn influence the genetic effect on adiposity measures. Some of the earlier studies on adiposity have used patients with various metabolic diseases (for example, T2DM, coronary artery disease, hypertension) as part of the gene discovery cohorts to increase study power and did not show strong heterogeneity of effect size between the healthy subjects and the patients.16, 17 In the present study, the effect sizes were comparable in non-diabetic, T2DM–ESRD and T2DM subjects in the meta-analysis, suggesting that the presence of disease has minimal effect on the genetic associations.

Studies in other ethnicities, particularly African-derived populations, are potentially valuable as they may help to fine map the signals of association and, because additional variants present at high frequency in African-derived populations may be absent or rare in European samples.48 Additionally, it is not clear whether associations found in the European samples can be consistently replicated in the samples of predominantly recent African ancestry; genetic, environmental or phenotypic heterogeneity, gene by environment interactions, or different recombination histories between populations could all contribute to a lack of replication in African-derived populations. Along with differences in the prevalence of obesity between African-American and European-derived populations, there are also ethnic differences in metabolic risk factors.49 African Americans have different distribution of fat, with lower amounts of visceral fat, compared with the non-Hispanic whites.50

In summary, we replicated association of seven key SNPs at four loci (NEGR1, TMEM18, SH2B1 and MC4R) in six African-American cohorts that demonstrated consistent association with previous studies of adiposity in Europeans. These loci are all highly expressed in the brain (and several especially in the hypothalamus), consistent with an important role for central nervous system processes in weight regulation. However, further comprehensive examination of these regions may be necessary to confirm our findings and elucidate for possible genetic differences in different populations.


Conflict of interest

The authors declare no conflict of interest.



  1. Guh DP, Zhang W, Bansback N, Amarsi Z, Birmingham CL, Anis AH. The incidence of co-morbidities related to obesity and overweight: a systematic review and meta-analysis. BMC Public Health 2009; 9: 88. | Article | PubMed |
  2. Flegal KM, Carroll MD, Ogden CL, Johnson CL. Prevalence and trends in obesity among US adults, 1999–2000. JAMA 2002; 288: 1723–1727. | Article | PubMed | ISI |
  3. Flegal KM, Carroll MD, Ogden CL, Curtin LR. Prevalence and trends in obesity among US adults, 1999–2008. JAMA 2010; 303: 235–241. | Article | PubMed | ISI | CAS |
  4. Ogden CL, Carroll MD, Curtin LR, McDowell MA, Tabak CJ, Flegal KM. Prevalence of overweight and obesity in the United States, 1999–2004. JAMA 2006; 295: 1549–1555. | Article | PubMed | ISI | CAS |
  5. Loos RJ, Bouchard C. Obesity—is it a genetic disorder? J Intern Med 2003; 254: 401–425. | Article | PubMed | ISI | CAS |
  6. Nelson TL, Brandon DT, Wiggins SA, Whitfield KE. Genetic and environmental influences on body-fat measures among African-American twins. Obes Res 2002; 10: 733–739. | Article | PubMed | ISI |
  7. Duncan AE, Agrawal A, Grant JD, Bucholz KK, Madden PA, Heath AC. Genetic and environmental contributions to BMI in adolescent and young adult women. Obesity (Silver Spring) 2009; 17: 1040–1043. | Article | PubMed |
  8. Sale MM, Freedman BI, Hicks PJ, Williams AH, Langefeld CD, Gallagher CJ et al. Loci contributing to adult height and body mass index in African American families ascertained for type 2 diabetes. Ann Hum Genet 2005; 69: 517–527. | Article | PubMed | ISI |
  9. Dina C, Meyre D, Gallina S, Durand E, Korner A, Jacobson P et al. Variation in FTO contributes to childhood obesity and severe adult obesity. Nat Genet 2007; 39: 724–726. | Article | PubMed | ISI | CAS |
  10. Hinney A, Nguyen TT, Scherag A, Friedel S, Bronner G, Muller TD et al. Genome wide association (GWA) study for early onset extreme obesity supports the role of fat mass and obesity associated gene (FTO) variants. PLoS One 2007; 2: e1361. | Article | PubMed | CAS |
  11. Frayling TM, Timpson NJ, Weedon MN, Zeggini E, Freathy RM, Lindgren CM et al. A common variant in the FTO gene is associated with body mass index and predisposes to childhood and adult obesity. Science 2007; 316: 889–894. | Article | PubMed | ISI | CAS |
  12. Scuteri A, Sanna S, Chen WM, Uda M, Albai G, Strait J et al. Genome-wide association scan shows genetic variants in the FTO gene are associated with obesity-related traits. PLoS Genet 2007; 3: e115. | Article | PubMed | CAS |
  13. Benzinou M, Creemers JW, Choquet H, Lobbens S, Dina C, Durand E et al. Common nonsynonymous variants in PCSK1 confer risk of obesity. Nat Genet 2008; 40: 943–945. | Article | PubMed | ISI | CAS |
  14. Chambers JC, Elliott P, Zabaneh D, Zhang W, Li Y, Froguel P et al. Common genetic variation near MC4R is associated with waist circumference and insulin resistance. Nat Genet 2008; 40: 716–718. | Article | PubMed | ISI | CAS |
  15. Loos RJ, Lindgren CM, Li S, Wheeler E, Zhao JH, Prokopenko I et al. Common variants near MC4R are associated with fat mass, weight and risk of obesity. Nat Genet 2008; 40: 768–775. | Article | PubMed | ISI | CAS |
  16. Willer CJ, Speliotes EK, Loos RJ, Li S, Lindgren CM, Heid IM et al. Six new loci associated with body mass index highlight a neuronal influence on body weight regulation. Nat Genet 2009; 41: 25–34. | Article | PubMed | ISI | CAS |
  17. Thorleifsson G, Walters GB, Gudbjartsson DF, Steinthorsdottir V, Sulem P, Helgadottir A et al. Genome-wide association yields new sequence variants at seven loci that associate with measures of obesity. Nat Genet 2009; 41: 18–24. | Article | PubMed | ISI | CAS |
  18. Meyre D, Delplanque J, Chevre JC, Lecoeur C, Lobbens S, Gallina S et al. Genome-wide association study for early-onset and morbid adult obesity identifies three new risk loci in European populations. Nat Genet 2009; 41: 157–159. | Article | PubMed | ISI | CAS |
  19. Lindgren CM, Heid IM, Randall JC, Lamina C, Steinthorsdottir V, Qi L et al. Genome-wide association scan meta-analysis identifies three Loci influencing adiposity and fat distribution. PLoS Genet 2009; 5: e1000508. | Article | PubMed | CAS |
  20. Heard-Costa NL, Zillikens MC, Monda KL, Johansson A, Harris TB, Fu M et al. NRXN3 is a novel locus for waist circumference: a genome-wide association study from the CHARGE Consortium. PLoS Genet 2009; 5: e1000539. | Article | PubMed | CAS |
  21. Speliotes EK, Willer CJ, Berndt SI, Monda KL, Thorleifsson G, Jackson AU et al. Association analyses of 249,796 individuals reveal 18 new loci associated with body mass index. Nat Genet 2010; 42: 937–948. | Article | PubMed | ISI | CAS |
  22. Heid IM, Jackson AU, Randall JC, Winkler TW, Qi L, Steinthorsdottir V et al. Meta-analysis identifies 13 new loci associated with waist-hip ratio and reveals sexual dimorphism in the genetic basis of fat distribution. Nat Genet 2010; 42: 949–960. | Article | PubMed | ISI | CAS |
  23. Cho YS, Go MJ, Kim YJ, Heo JY, Oh JH, Ban HJ et al. A large-scale genome-wide association study of Asian populations uncovers genetic factors influencing eight quantitative traits. Nat Genet 2009; 41: 527–534. | Article | PubMed | ISI | CAS |
  24. Tan JT, Dorajoo R, Seielstad M, Sim XL, Ong RT, Chia KS et al. FTO variants are associated with obesity in the Chinese and Malay populations in Singapore. Diabetes 2008; 57: 2851–2857. | Article | PubMed | ISI | CAS |
  25. Hotta K, Nakamura M, Nakamura T, Matsuo T, Nakata Y, Kamohara S et al. Association between obesity and polymorphisms in SEC16B, TMEM18, GNPDA2, BDNF, FAIM2 and MC4R in a Japanese population. J Hum Genet 2009; 54: 727–731. | Article | PubMed | ISI |
  26. Ng MC, Tam CH, So WY, Ho JS, Chan AW, Lee HM et al. Implication of genetic variants near NEGR1, SEC16B, TMEM18, ETV5/DGKG, GNPDA2, LIN7C/BDNF, MTCH2, BCDIN3D/FAIM2, SH2B1, FTO, MC4R, and KCTD15 with obesity and type 2 diabetes in 7705 Chinese. J Clin Endocrinol Metab 2010; 95: 2418–2425. | Article | PubMed | ISI |
  27. Grant SF, Bradfield JP, Zhang H, Wang K, Kim CE, Annaiah K et al. Investigation of the locus near MC4R with childhood obesity in Americans of European and African ancestry. Obesity (Silver Spring) 2009; 17: 1461–1465. | PubMed |
  28. Kang SJ, Chiang CW, Palmer CD, Tayo BO, Lettre G, Butler JL et al. Genome-wide association of anthropometric traits in African- and African-derived populations. Hum Mol Genet 2010; 19: 2725–2738. | Article | PubMed | ISI |
  29. Bollepalli S, Dolan LM, Deka R, Martin LJ. Association of FTO gene variants with adiposity in African-American adolescents. Obesity (Silver Spring) 2010; 18: 1959–1963. | Article | PubMed |
  30. Grant SF, Li M, Bradfield JP, Kim CE, Annaiah K, Santa E et al. Association analysis of the FTO gene with obesity in children of Caucasian and African ancestry reveals a common tagging SNP. PLoS One 2008; 3: e1746. | Article | PubMed | CAS |
  31. Wing MR, Ziegler J, Langefeld CD, Ng MC, Haffner SM, Norris JM et al. Analysis of FTO gene variants with measures of obesity and glucose homeostasis in the IRAS Family Study. Hum Genet 2009; 125: 615–626. | Article | PubMed | ISI |
  32. Hassanein MT, Lyon HN, Nguyen TT, Akylbekova EL, Waters K, Lettre G et al. Fine mapping of the association with obesity at the FTO locus in African-derived populations. Hum Mol Genet 2010; 19: 2907–2916. | Article | PubMed | ISI |
  33. Freedman BI, Hicks PJ, Bostrom MA, Comeau ME, Divers J, Bleyer AJ et al. Non-muscle myosin heavy chain 9 gene MYH9 associations in African Americans with clinically diagnosed type 2 diabetes mellitus-associated ESRD. Nephrol Dial Transplant 2009; 24: 3366–3371. | Article | PubMed | ISI | CAS |
  34. Divers J, Wagenknecht LE, Bowden DW, Carr JJ, Hightower RC, Ding J et al. Regional adipose tissue associations with calcified atherosclerotic plaque: African American-diabetes heart study. Obesity (Silver Spring) 2010; 18: 2004–2009. | Article | PubMed |
  35. Bowden DW, Rudock M, Ziegler J, Lehtinen AB, Xu J, Wagenknecht LE et al. Coincident linkage of type 2 diabetes, metabolic syndrome, and measures of cardiovascular disease in a genome scan of the diabetes heart study. Diabetes 2006; 55: 1985–1994. | Article | PubMed | ISI |
  36. Henkin L, Bergman RN, Bowden DW, Ellsworth DL, Haffner SM, Langefeld CD et al. Genetic epidemiology of insulin resistance and visceral adiposity. The IRAS Family Study design and methods. Ann Epidemiol 2003; 13: 211–217. | Article | PubMed | ISI |
  37. Wagenknecht LE, Mayer EJ, Rewers M, Haffner S, Selby J, Borok GM et al. The insulin resistance atherosclerosis study (IRAS) objectives, design, and recruitment results. Ann Epidemiol 1995; 5: 464–472. | Article | PubMed | CAS |
  38. O’Connell JR, Weeks DE. PedCheck: a program for identification of genotype incompatibilities in linkage analysis. Am J Hum Genet 1998; 63: 259–266. | Article | PubMed | ISI | CAS |
  39. Keene KL, Mychaleckyj JC, Leak TS, Smith SG, Perlegas PS, Divers J et al. Exploration of the utility of ancestry informative markers for genetic association studies of African Americans with type 2 diabetes and end stage renal disease. Hum Genet 2008; 124: 147–154. | Article | PubMed | ISI |
  40. Purcell S, Cherny SS, Sham PC. Genetic power calculator: design of linkage and association genetic mapping studies of complex traits. Bioinformatics 2003; 19: 149–150. | Article | PubMed | ISI | CAS |
  41. Marg A, Sirim P, Spaltmann F, Plagge A, Kauselmann G, Buck F et al. Neurotractin, a novel neurite outgrowth-promoting Ig-like protein that interacts with CEPU-1 and LAMP. J Cell Biol 1999; 145: 865–876. | Article | PubMed | ISI | CAS |
  42. Schafer M, Brauer AU, Savaskan NE, Rathjen FG, Brummendorf T. Neurotractin/kilon promotes neurite outgrowth and is expressed on reactive astrocytes after entorhinal cortex lesion. Mol Cell Neurosci 2005; 29: 580–590. | Article | PubMed | ISI | CAS |
  43. Jurvansuu J, Zhao Y, Leung DS, Boulaire J, Yu YH, Ahmed S et al. Transmembrane protein 18 enhances the tropism of neural stem cells for glioma cells. Cancer Res 2008; 68: 4614–4622. | Article | PubMed | ISI | CAS |
  44. Jamshidi Y, Snieder H, Ge D, Spector TD, O’Dell SD. The SH2B gene is associated with serum leptin and body fat in normal female twins. Obesity (Silver Spring) 2007; 15: 5–9. | Article | PubMed | CAS |
  45. Farooqi IS, Keogh JM, Yeo GS, Lank EJ, Cheetham T, O’Rahilly S. Clinical spectrum of obesity and mutations in the melanocortin 4 receptor gene. N Engl J Med 2003; 348: 1085–1095. | Article | PubMed | ISI | CAS |
  46. Balthasar N, Dalgaard LT, Lee CE, Yu J, Funahashi H, Williams T et al. Divergence of melanocortin pathways in the control of food intake and energy expenditure. Cell 2005; 123: 493–505. | Article | PubMed | ISI | CAS |
  47. Palmer ND, Hester JM, An SS, Adeyemo A, Rotimi C, Langefeld CD et al. Resequencing and analysis of variation in the TCF7L2 gene in African Americans suggests that SNP rs7903146 is the causal diabetes susceptibility variant. Diabetes 2011; 60: 662–668. | Article | PubMed | ISI |
  48. Cooper RS, Tayo B, Zhu X. Genome-wide association studies: implications for multiethnic samples. Hum Mol Genet 2008; 17: R151–R155. | Article | PubMed | ISI |
  49. Cossrow N, Falkner B. Race/ethnic issues in obesity and obesity-related comorbidities. J Clin Endocrinol Metab 2004; 89: 2590–2594. | Article | PubMed | ISI | CAS |
  50. Bacha F, Saad R, Gungor N, Janosky J, Arslanian SA. Obesity, regional fat distribution, and syndrome X in obese black versus white adolescents: race differential in diabetogenic and atherogenic risk factors. J Clin Endocrinol Metab 2003; 88: 2534–2540. | Article | PubMed | ISI | CAS |


We thank the study subjects for their participation. Genotyping services were provided by the Center for Inherited Disease Research (CIDR). CIDR is fully funded through a federal contract from the National Institutes of Health to The Johns Hopkins University, contract number HHSC268200782096C. This work was supported by NIH Grants R01 DK087914 (MCYN), R01 DK066358 (DWB), R01 DK053591 (DWB), R01 HL56266 (BIF), R01 DK070941 (BIF), R01 HL060944 (LW), R01 HL061210 (MB), K99 DK081350 (NDP) and by the Wake Forest University School of Medicine Grant M01 RR07122 and Venture Fund (MCYN).

Supplementary Information accompanies the paper on International Journal of Obesity website

Extra navigation