Association of CDKAL1, IGF2BP2, CDKN2A/B, HHEX, SLC30A8, and KCNJ11 With Susceptibility to Type 2 Diabetes in a Japanese Population Free

DNA samples were obtained from the peripheral blood samples of 1,630 type 2 diabetic patients recruited from the outpatient clinic of the Shiga University of Medical Science, Kawasaki Medical School (978 men and 652 women; age 61.5 ± 11.6 years; duration of diabetes 11.5 ± 13.9 years; A1C 7.4 ± 1.6%; fasting plasma glucose 9.1 ± 3.5 mmol/l; BMI 23.7 ± 3.9 kg/m² [all values are expressed as means ± SD], Table 1). Diabetes was diagnosed according to the WHO criteria. Type 2 diabetes is clinically defined as a disease with gradual adult onset. Subjects who tested positive for anti-GAD antibodies and those diagnosed as mitochondrial disease (mitochondrial myopathy, encephalopathy, lactic acidosis, and stroke-like episodes [MELAS]) or maturity-onset diabetes of young (MODY) were not included in the case group. We also examined 1,064 control subjects who were enrolled from an annual health check conducted either at the Juntendo University or Keio University (Tokyo, Japan; 638 men and 426 women; age 45.5 ± 9.5 years; A1C 4.7 ± 0.4%; fasting plasma glucose 5.1 ± 0.5 mmol/l; BMI 22.9 ± 3.0 kg/m²; Table 1).

Written informed consent was obtained from all the participants, and DNA was extracted using the standard phenol-chloroform procedure. The protocol was approved by the ethics committee of the Institute of Physical and Chemical Research (RIKEN).

Each SNP genotyping was performed by the TaqMan assay (Applied Biosystems, Foster City, CA) or by the multiplex-PCR invader assay as described previously (13). The success rates of these assays were >95%, and there was almost a 100% agreement between the results of genotyping and direct sequencing.

Statistical methods for determining associations and to calculate linkage disequilibrium (LD) coefficients (r²) were described previously (14). We performed the Hardy-Weinberg equilibrium (HWE) test according to the method described by Nielsen et al. (15). Although the genotype distributions of all the SNPs were within the Hardy-Weinberg equilibrium (P ≥ 0.01), some of them had borderline results for HWE test (rs5219 in control, rs7480010 in case, rs8050136 in case; see supplementary Table 1 at http://dx.doi.org/10.2337db07-0979). Therefore, we performed Wright's F statistics (16) to evaluate the difference in the population structure between our case and control groups using randomly selected 96 SNPs. The result indicated that the population structures of our case and control groups were almost the same in view of a very small F_ST value between the two groups (F_ST = 0.001556).

The differences between the case and control groups in terms of genotype distribution were analyzed using a logistic regression analysis. To test the additive model of each SNP after adjusting for sex, age, and log-transformed BMI, the analysis was performed using StatView software. In addition, χ² test to evaluate the additive, dominant, and recessive models of each SNP were also performed by the method of Sladek et al. (8).

The difference in the BMI according to the genotypes was analyzed using a multiple linear regression with log-transformed BMI as the dependent variable and genotype as the independent variable with sex as a covariate for log-transformed BMI (17).

As shown in Table 2, six SNPs within six distinct loci (rs4402960, rs10811661, rs5219, rs1111875, rs13266634, and rs7756992) were found to be significantly associated with type 2 diabetes in our Japanese population. No significant association was observed between the remaining five loci and type 2 diabetes in this population (P ≥ 0.05; Table 2). We also evaluated the association of the 14 SNP loci with type 2 diabetes using χ² test by the method of Sladek et al. (8) (supplementary Table 2) and identified almost consistent results with those obtained by a logistic regression analysis even after selecting control subjects whose age was >50 years old (n = 382, supplementary Table 3). Since FTO variants have been reported to be associated with BMI, we also examined the association of these polymorphisms with BMI in our control subjects, and we found that the polymorphisms within FTO and HHEX were modestly associated with BMI (Table 3).

In the present study, we identified significant associations of SNPs within the CDKAL1, IGF2BP2, CDKN2A/B, HHEX, SLC30A8 and KCNJ11 genes with the susceptibility to type 2 diabetes in a Japanese population. We also found that the SNPs in the FTO gene and the HHEX gene were significantly associated with BMI in our control group.

Recent advances in human genetic research have facilitated the identification of genes conferring susceptibility to common diseases such as type 2 diabetes from across the entire human genome by using a large number of subjects, and genome-wide association studies have been conducted worldwide (8–12). Although the importance of TCF7L2 as a susceptibility gene for type 2 diabetes has been well established, its polymorphism could account for ∼20% of all cases for type 2 diabetes in white populations (2,4,5). The population-attributable risk of the TCF7L2 polymorphism in the Japanese was ∼2% because the risk allelic frequency in a previously studied Japanese population was very low (6,7). Further, many important genes for the disease remain to be identified, especially in East Asian populations.

Several groups have independently performed genome-wide association studies for type 2 diabetes in white populations (8–12). All these studies have demonstrated that the TCF7L2 polymorphism is most strongly associated with the susceptibility to type 2 diabetes, and from a large set of replication studies, they have identified additional candidate loci also associated with the disease. The results of these genome-wide association studies are also considered highly consistent with regard to white populations. However, there are considerable differences in phenotype between Japanese (lean and less hyperinsulinemic Asian type 2 diabetes) and white (European descent) type 2 diabetes, and these differences might affect the genetic contribution of each gene to conferring susceptibility to type 2 diabetes. Therefore, the new candidates should also be evaluated in different ethnic groups because there are clear ethnic differences in terms of genetic contribution to diseases (13,18).

In the present study, we also identified the six SNPs within the CDKAL1, IGF2BP2, CDKN2A/B, HHEX, SLC30A8, and KCNJ11 genes that were significantly associated with type 2 diabetes in our Japanese population. Since the risk alleles for these variations in the tested population were entirely consistent with those in white populations (8–12), the contribution of these polymorphisms to type 2 diabetes susceptibility is highly convincing across the different ethnicities, although there are considerable differences in allele frequencies between the Japanese and white populations (Table 4 and supplementary Table 4). Regarding the HHEX locus, only rs1111875 had significant association with type 2 diabetes in the present study, whereas both rs1111875 and rs7923837 were associated with the disease in white populations. The risk allele frequencies of SNPs in the HHEX gene were significantly different between our population and white populations (rs1111875 28.4 vs. 56.1%, rs7923837 20.6 vs. 60.1% in Japanese and white populations, respectively; supplementary Table 4), and there were some differences in the LD coefficients (r²) between rs1111875 and rs7923837 among those populations (0.5 in the Japanese and 0.698 in the white population). Therefore, these differences might explain the discrepancies in the results for the association of SNPs in the HHEX gene with type 2 diabetes, although the possibility of insufficient study power in the present study could not be excluded.

We also found that the SNPs within FTO (rs9939609 and rs8050136) and HHEX (rs1111875 and rs7923837) were modestly associated with BMI in our control subjects (Table 3). The association of the SNPs in FTO with type 2 diabetes was not significant in our population (Table 2 and supplementary Tables 2 and 3). Therefore, the variations in the FTO gene might directly affect body weight rather than type 2 diabetes itself; this finding is also consistent with that in white populations (17,19).

Among the other four loci, the association of the Pro12Ala polymorphism in the PPARG gene with type 2 diabetes has been well established (20,21), and this association was also observed in Japanese populations (22,23), although we could not determine whether the association of this polymorphism with type 2 diabetes was significant. Because the frequencies of the Ala allele or its carrier (X/Ala) found in the present study (three and 6%, respectively) are consistent with those in previous studies on Japanese populations, the frequencies of the Ala allele can be considered very low in the Japanese as compared to those in white populations. Because an estimated power to detect the association of the SNP with type 2 diabetes in the present study is ∼20%, the results do not appear strong enough to determine the association between the PPARG Pro12Ala polymorphism and type 2 diabetes.

With regard to the remaining three loci, the results for EXT2, LOC387761, and rs930039 were not always in agreement with those of the genome-wide association studies in white populations. In addition, the allele frequencies of those SNPs were also significantly different between the Japanese and the white populations (Table 4 and supplementary Table 4). Because the estimated powers of the present study were >90%, >90%, and >70% for rs1113132 (EXT2), rs7480010 (LOC387761), and rs930039, respectively, the contribution of these three loci in the Japanese populations is considered minor, if present at all; however, more replication studies are required for the precise evaluation of these loci.

In conclusion, we identified significant associations between SNPs within the CDKAL1, IGF2BP2, CDKN2A/B, HHEX, SLC30A8, and KCNJ11 genes and type 2 diabetes in a Japanese population. These loci are considered strong candidates for conferring susceptibility to type 2 diabetes across different ethnicities. However, further studies are required to elucidate the association of other loci with the susceptibility to type 2 diabetes and the biological significance of these genes and gene polymorphisms.

We thank Shuichi Tsukada, Masa-aki Kobayashi, and the technical staff of the Laboratory for Diabetic Nephropathy at the SNP Research Center for providing technical assistance. This work was partly supported by the Japanese Millennium Project.