Replication of Genetic Effects of FTO Polymorphisms on BMI in a Korean Population





It has been newly reported in recent studies that single-nucleotide polymorphisms (SNPs) in the first intron of the FTO gene have been associated with BMI in whites. To determine whether the gene is associated with BMI in Asians also, we performed a replication study of the association of the gene with BMI in a Korean population. Two SNPs in the FTO gene (rs1421085 and rs17817449) were genotyped using the TaqMan method in a Korean population (n = 1,733). The two SNPs were then used for an association study with BMI through statistical analyses. The rs1421085 C allele (P = 0.0015, effect size = 0.0056) and rs17817449 G allele (P = 0.0019, effect size = 0.0053) were found to be significantly associated with increased BMI. Our results suggest that FTO may be one of the worldwide obesity-risk genes.

Obesity is one of the most common disorders in clinical practice and is closely related to a number of pathological disorders such as noninsulin-dependent diabetes, hypertension, cancer, gallbladder disease, and atherosclerosis. A large number of genetic association studies suggest that polymorphic variants in genes are associated with BMI (1).

Recently, three independent studies have found that common variants in intron 1 of the FTO gene (fat mass and obesity associated; #MIM:610966) were found to be associated with BMI (2,3,4), and those findings have been replicated in several white populations (3,5,6). However, among other ethnic populations (Asians and Africans), the genetic effects of polymorphisms in intron 1 of FTO have been controversial. The rs9930506 was not associated with obesity-related phenotypes including BMI, weight, and hip circumference in 1,101 African Americans (4). In addition, among oceanic populations (Melanesians, Micronesians, and Polynesians), rs9939609 was not associated with BMI (7). In a Japanese population (864 type 2 diabetes patients and 864 controls), rs8050136 and rs9939690 were not found to be associated with BMI either, although one variant, rs8050136, was associated with the risk of type 2 diabetes (8).

A number of highly linked single-nucleotide polymorphisms (SNPs) have been reported to be associated with BMI in previous studies. In this study, in an effort to replicate the association with BMI among Asians, we examined genetic effects in a Korean cohort (n = 1,733) with two SNPs (rs1421085 and rs17817449) that had been reported in white populations (3). To compare the genetic effects directly, we selected the two SNPs (listed earlier) in intron 1 of FTO as tagging SNPs (see Supplementary Figure S1 online for linkage disequilibrium among SNPs in intron 1 of FTO in an Asian population).

The genotype distributions of the two SNPs (rs1421085 and rs17817449) in high-BMI individuals (>25 (kg/m2)) were in Hardy-Weinberg equilibrium (P = 0.84 and 0.97, respectively). However, genotype distributions among low-BMI individuals (≤25 (kg/m2)) were significantly deviated from Hardy-Weinberg equilibrium (P = 0.011 for rs1421085 and P = 0.011 for rs17817449). Those deviations of FTO variants might be indirect evidence for association of those variants with BMI. The minor allele frequencies of rs1421085 and rs17817449 were 0.149 and 0.147, respectively, which were much lower than those shown in white populations (0.448 and 0.447 among 270 HapMap white samples, respectively). The much lower minor allele frequencies in Asian populations (Korean (this study), Japanese (8), and Chinese (9)) than those in white populations might explain the lower strength of association with BMI among Asian populations.

Genetic associations of the two FTO polymorphisms were analyzed with total cholesterol, triglyceride, high-density lipoprotein cholesterol, fasting blood glucose, and BMI. No significant associations were detected with total cholesterol, triglyceride, high-density lipoprotein cholesterol, and fasting blood glucose (Table 1). However, both SNPs (rs1421085 and rs17817449) were significantly associated with BMI in the Korean population (Table 2) (same direction of effects as observed among white populations). The rs1421085 C allele (P = 0.0015, effect size = 0.0056) and rs17817449 G allele (P = 0.0019, effect size = 0.0053) were found to be significantly associated with increased BMI (Table 2).

Table 1.  Association analysis of FTO gene polymorphisms with obesity-related phenotypes
inline image
Table 2.  Subgroup analysis of association of FTO intronic polymorphisms with BMI according to gender
inline image

The genetic effects were further analyzed according to gender (male and female). Interestingly, the genetic effects of both SNPs were more apparent among male subjects (Table 2).

Genetic association studies provide a potentially powerful tool for identifying genetic variations that influence susceptibility to common diseases. However, there are numerous cases of associations that cannot be replicated afterward, which has led to skepticism about genetic epidemiology studies of complex diseases (10,11). In the case of FTO association with BMI, there have been robust evidences and numerous replications in white populations. However, additional evidence from other major ethnic groups may be needed, such as from African and Asian populations.

FTO is expressed in a range of tissues including the brain, adipose tissue, pancreas, and hypothalamus. However, the way by which the FTO gene product acts on energy metabolism should be elucidated by additional investigation. The one relevant role with BMI of this gene is the potential involvement in hypothalamus-pituitary axis implicated in body weight regulation (12). Recently, putative function of FTO has been reported. Studies of wild-type mice indicate that FTO mRNA is most abundant in the brain, particularly in hypothalamic nuclei governing energy balance, and that FTO mRNA levels in the arcuate nucleus are regulated by feeding and fasting. Studies can now be directed toward determining the physiologically relevant FTO substrate and how nucleic acid methylation status is linked to increased fat mass (13).

In this study, first positive associations with BMI were detected in an Asian population. Our results are concordant with previous studies in white populations. However, our results do not agree with previous studies in Asian populations, e.g., no associations were found among Oceanic (Melanesians, Micronesians, and Polynesians), Japanese, and Chinese populations (7,8,9). The first referenced study of Oceanic populations might be unreliable because of weak statistical power due to limited sample sizes (n = 39–116) (ref. 7). In the other study, of a Japanese population, the association was performed with 864 type 2 diabetes patients and 864 controls using two FTO variants, rs8050136 and rs9939690. Although one variant, rs8050136, was associated with type 2 diabetes, neither SNP was found to be associated with BMI (8). In addition, three FTO SNPs (rs8050136, rs9939609, and rs9930506) were not associated with obesity-related phenotypes among Chinese Han population (9). The direct comparison would not be possible because they had tested different SNPs (Supplementary Table S1 online). However, negative association in their study among Chinese population could be expected when considering the high linkage disequilibrium (r 2 = 0.63–1, data not shown) among FTO SNPs in Asians (Supplementary Figure S1 online). Although it is hard to decipher the discrepancies with our study on the effect of FTO variants on BMI, the possible difference in diets, differing genetic background of the subjects (due to differing ethnicities), and/or low sample sizes in both studies could be plausible explanations. Therefore, further study with larger samples of other Asian populations is needed to confirm the identified genetic effects on BMI in a Korean population.

In summary, we have replicated the genetic association of two SNPs in intron 1 of FTO (rs1421085 and rs17817449) in a Korean population (n = 1,733) and, to our knowledge, this is the first such replication in an Asian population. Although further studies with larger samples of other Asian populations are needed, FTO polymorphisms show potential for being one of the global BMI-associated genetic factors.


Subjects and measures

The Korean study population consisted of 996 subjects recruited from an obesity clinic in Kirin Oriental Medical Hospital and 737 subjects recruited from an obesity clinic in Yeungnam University Medical Center. All subjects provided written informed consent, and this study was approved by the Institutional Review Board of the Korea Institute of Oriental Medicine and Yeungnam University. All study subjects were examined in the morning after overnight fasting. Blood samples were drawn for biochemical measurements (total cholesterol, triglyceride, and high-density lipoprotein cholesterol) and DNA extraction. The characteristics of the recruited individuals are shown in Table 3.

Table 3.  Clinical characteristics of study subjects (n = 1,733)
inline image

Genotyping with fluorescence polarization detection

Two SNPs (rs1421085 and rs17817449) of the FTO gene were genotyped with TaqMan (14). Amplifying primers and the TaqMan probes were purchased as Assay on Demand (Applied Biosystems, Foster City, CA). The probes were labeled with two dyes—the FAM dye on one allelic probe and the VIC dye on the other. The genomic templates were amplified using PCRs in a 384-well format plate of a total reaction volume of 5 μl. These were performed in solutions with the TaqMan Universal Master mix (Applied Biosystems), PCR primer of concentration 900 nmol/l, TaqMan probe of concentration 200 nmol/l, and 20 ng of genomic DNA. The well plates were moved into a thermal cycler (PE 9700, Applied Biosystems) and amplified with the condition of heating (at 50 °C for 2 min and 95 °C for 10 min) followed by amplification reaction (40 cycles of 95 °C for 15 s, 60 °C for 1 min). Assay products were transferred to a Prism 7900HT instrument (Applied Biosystems) read the intensity of the fluorescence in each well of the plate.


The χ2-tests were used to determine whether individual variants were in equilibrium at each locus in the population (Hardy-Weinberg equilibrium). We examined the linkage disequilibrium expressed as Lewontin's D′ (|D′|), where D′ = D/|Dmax|, and the linkage disequilibrium coefficient r 2 between all pairs of biallelic loci (15). The independent effects of genetic polymorphisms on BMI were assessed by multiple linear regression analyses with adjustments for age and/or sex as covariates. All statistical analyses were performed using SAS, version 8.02 (SAS, Cary, NC).

Supplementary Material

Supplementary material is linked to the online version of the paper at http:www.nature.comoby


This work was supported by a Korea Research Foundation Grant (M10643020001-06N4302-00100) and the National Research Laboratory Program as part of the National Research and Development Program of the Ministry of Commerce, Industry, and Energy (No. 10020202).


The authors declared no conflict of interest.