Variants in ASK1 are associated with skeletal muscle ASK1 expression, in vivo insulin resistance, and type 2 diabetes in Pima Indians

Abstract

Objective: Prior genome-wide association and exon array expression studies both provided suggestive evidence that apoptosis signal regulating kinase 1 (ASK1) may influence in vivo insulin action in Pima Indians. Genetic variants in or near ASK1 were analyzed to assess the role of this gene in insulin action and type 2 diabetes.

Research design and methods: Genotypic data from 31 variants were used to determine the linkage disequilibrium pattern across ASK1 in Pima Indians. Eight tag SNPs were initially genotyped in 3,501 full-heritage Pima Indians. Replication for association with diabetes was assessed in a second population-based sample of 3,723 Native Americans and the published DIAGRAM study. Quantitative traits were analyzed in 536 nondiabetic Native Americans, and ASK1 expression was examined in skeletal muscle of 153 nondiabetic Native Americans.

Results: Three tag SNPs were associated with type 2 diabetes (rs35898099, P = 0.003, odds ratio [95% CI] 1.27 [1.08-1.47]; rs1570056, P = 0.007, 1.19 [1.05-1.36]; rs7775356, P = 0.04, 1.14 [1.01-1.28]) in the full-heritage Pima Indians. The association with rs35898099 was replicated in a second sample of Native Americans (P = 0.04, 1.22 [1.01-1.47]), while that for rs1570056 was replicated in the DIAGRAM study of Caucasians (Z statistic based P = 0.026; fixed-effect model, 1.06 [1.00-1.12]). The diabetes risk allele for rs1570056 was associated with reduced insulin action as assessed by either HOMA-IR in 2,549 nondiabetic full-heritage Pima Indians (P = 0.027) or a hyperinsulinemic-euglycemic clamp among 536 nondiabetic Native Americans (P = 0.02). Real-time PCR identified a positive correlation between ASK1 expression in skeletal muscle biopsies and in vivo insulin action (P = 0.02, r = 0.23), and the risk allele for rs1570056 was associated with lower ASK1 expression (P = 0.003, r = -0.22).

Conclusions: ASK1 variants may increase susceptibility to type 2 diabetes by decreasing insulin sensitivity via reduced ASK1 expression.

FIG. 1.

Relative position and LD pattern of the 31 variants across the ASK1 locus in Pima Indians. LD is shown as r² and the eight tag SNPs (no. 1, 5, 7, 8, 9, 10, 20, and 31) were selected by Tagger algorithm (r² ≥0.8, minor allele frequency >0.05).

FIG. 2.

Association of rs1570056 with ASK1 expression in skeletal muscle from 153 nondiabetic Native Americans. Data were raw means ± SE. P value was calculated under an additive model and adjusted for age, sex, percent body fat, and Pima heritage. Before the analysis, the relative ASK1 expression levels were logarithmically transformed to approximate the normal distribution.

FIG. 3.

Positive correlation between ASK1 expression in skeletal muscle and in vivo insulin action (insulin-stimulated glucose disposal). Before the analysis, both in vivo insulin action (insulin-stimulated glucose disposal) and the relative ASK1 expression levels were logarithmically transformed to approximate the normal distribution. P value was adjusted for age, sex, percent body fat, and Pima heritage. EMBS, estimated metabolic body size.