To identify coding variation that increases susceptibility to type 2 diabetes (T2D) via its effect on a pre-diabetic trait, we obtained exome sequence data on 177 Pima Indians. 148,616 variants (11.1% novel) were detected that met our quality control criteria. Of these variants, 15,799 and 15,642 were non-synonymous and synonymous SNPs, respectively. Selected variants (N=360) were genotyped in 555 non-diabetic Pima Indians characterized as inpatients in our Clinical Research Center for % body fat (PFAT), central obesity (waist/thigh), glucose disposal rates (M) during a hyperinsulinemic-euglycemic clamp, acute insulin response (AIR) to intravenous glucose, and 2-hour plasma glucose concentrations (2-hr glucose) during an OGTT. Variants were further assessed for association with BMI and T2D in up to 7,667 subjects who had participated in a longitudinal study of T2D in the Gila River Indian Community. Two SNPs (in the genes ASXL3 and CYB5A) were significantly associated with a pre-diabetic trait after correction for multiple testing (360 SNPs analyzed for 5 traits required a p<2.710-5). The SNP in ASXL3 (rs2282632) was associated with insulin sensitivity (M; p= 2.010-5, adjusted for age, sex, PFAT and nuclear family membership) but was not associated with T2D. In contrast, the SNP in CYB5A (rs7238987) was associated with PFAT (p=6.710-6) in 555 non-diabetic subjects, BMI measured at a non-diabetic exam in 5880 subjects (p=6.210-7) and T2D in 7667 subjects (OR=1.14 1.03-1.14; p=9.210-3). CYB5A encodes a membrane bound microsomal hemoprotein which acts as an electron carrier for the stearoylCoA-desaturase (SCD) complex facilitating the conversion of saturated fatty acid to mono unsaturated fatty acid. Prior studies have reported that SCD activity is associated with obesity. We conclude that exome sequencing is a viable technique for identifying new loci that increase risk for T2D via an influence on a pre-diabetic trait. We have also obtained whole genome sequence data on 335 Pima Indians. Sequencing was performed by Illumina (N=301) and Complete Genomics, Inc (N=34). 13 million variants were found, including 11 million SNPs, 1.6 million Indels and 255,802 substitutions. Among all SNPs, 2.7 million were novel. One of our approaches for follow-up on variation detected by whole genome sequencing has been to perform large-scale genotyping of all novel variation that maps within known T2D genes. This approach led our analysis of a R1420H in SUR1 (ABCC8 gene). Missense variants in KCNJ11 and ABCC8, which encode the KIR6.2 and SUR1 subunits of the beta-cell KATP channel, have previously been implicated in type 2 diabetes, neonatal diabetes, and hyperinsulinemic hypoglycemia of infancy (HHI). Genotyping of this variant in all American Indians living in the Gila River Indian Community showed that 3.3% of the population (N = 7,710)carried this variant. R1420H carriers had a higher mean birth weight, and a twofold increased risk for type 2 diabetes with a 7 year earlier onset age despite being leaner than non-carriers. One individual homozygous for R1420H was identified; retrospective review of his medical records was consistent with HHI and a diagnosis of diabetes at age 3.5 years. In vitro studies show that the R1420H substitution decreases KATP channel activity. Identification of this loss-of-function variant in ABCC8 with a carrier frequency of 3.3% impacts clinical care since homozygous inheritance and potential HHI will occur in 1/3,600 births in this American Indian population. In addition to follow-up studies of novel variation in known gene for T2D and obesity, we wanted to explore variation across all loci. Therefore, we designed an entirely custom Affymetrix Axiom array based on all variation detected in the Pima Indian genomes. This array was initially utilized to genotype 3,637 full heritage Pima Indians who had participated in a longitudinal study of T2D (Stage 1) and 548,206 variants were successfully analyzed which tagged 92% of the 4.9M common variants (minor allele frequency >0.05) detected in the Pima genomes. An additional 4,060 non-full heritage Pima Indians from the same longitudinal study were subsequently genotyped with the array (Stage 2). Among the top associations was a new signal for T2D that had not been captured in our prior genome-wide assocation study which used a commerciallly available genotyping array (Affymetrix 6.0). This SNP (rs11564707)independently associated with T2D in the full heritage (risk allele G frequency=0.45) and non-full heritage (G=0.40) Pima Indian samples (full heritage P=8x10-7; OR95%CI=1.331.19-1.49 and non-full heritage P=5x10-3; OR= 1.221.10-1.40, adjusted for age, sex, birth year, family membership, reported heritage and fraction Indian). Combining the two samples provided the most robust association with T2D (2.2x10-8, OR=1.281.18-1.40), but this association was entirely driven by the 4,175 females (adjusted P=2.5x10-9; OR=1.411.26-1.58); no association was observed in 3,314 males (P=0.14; OR= 1.110.97-1.27). The gender x genotype interaction was significant (P=0.001). This female specific association was also observed for measures of early insulin secretion assessed in full heritage Pima Indians who were normal glucose tolerant. Both the 30 minute insulin levels during a 75g oral glucose tolerant test and the insulin response to a 25g intravenous glucose bolus were reduced in females (both p<0.01) but not males. Rs11564707 maps near the microRNA 4686 and the upstream region of the TH gene and tags 11 nearby variants with an r2>0.99 (including a novel T>TTTTGTTTG insertion at chr11:2,204,322). TH encodes tyrosine hydroxylase which is a good physiologic candidate for this association; however, future functional studies are required to establish the causative variant and affected gene.
Showing the most recent 10 out of 14 publications
