Meta-analysis of genome-wide association study of homeostasis model assessment β cell function and insulin resistance in an East Asian population and the European results

Compared with Western populations, Asians develop diabetes at younger ages, at lower degrees of obesity. Because diabetes and the related traits are influenced by the interplay between genetic and environmental factors, it is important to understand the genetic differences between Asian and Western populations. Recently, a large-scale meta-analysis of genome-wide association studies for beta cell function and insulin resistance in the European ancestry was reported by the Meta-Analyses of Glucose and Insulin-related traits Consortium (MAGIC). The MAGIC study reported 17 SNPs for homeostasis model assessments (HOMA-%B: beta cell function and HOMA-IR: insulin resistance). In this study, we tried to replicate the effects of reported SNPs by MAGIC study, which were influencing HOMAs in two Korean populations. HOMA-IR and HOMA-B were computed using two HOMA models (HOMA1 and HOMA2). The HOMA2 model has recently been updated with physiological adjustments into a computer version, providing a more accurate index. Dupuis et al. (Nat Genet 42: 105-116, 2010). In this study, we examined the reported SNPs in two Korean community-based cohorts (Ansung and Ansan). The Korean genotypes and glucose and insulin traits for 5,974 non-diabetic subjects were obtained from a previous genome-wide association study. Although we expected the HOMA2 to be suitable to replicate the results of different ethnics, our results revealed that the HOMA1 was more significantly replicated. As a result, 5 SNPs (rs10830963 in MTNR1B, rs4607517 in GCK, rs2191349 in DGKB/TMEM195, rs174550 in FADS1, rs7034200 in GLIS3) were significantly replicated with HOMA-%B, but no SNP was replicated with HOMA-IR. Two SNPs (rs560887 in G6PC, rs13266634 in SLC30A8) and one SNP (rs35767 in IGF1) showed the weak association p values (unadjusted p values lower than 0.05) for HOMA-%B and HOMA-IR, respectively. The replicated SNPs and the weakly associated SNPs were also significantly associated with the fasting glucose levels. They revealed the same direction of the effect sizes in both studies, but the effect sizes were stronger in Koreans than in MAGIC. Conclusively, our results indicated that SNPs from MTNR1B, GCK, DGKB, FADS1, and GLIS3 were consistently associated with HOMA-%B in both Korean and MAGIC populations.