Genome wide scan of soluble RAGE levels in a diabetic population reveals a novel associated variant in RAGE-HLA-DQA2 region

Abstract

RESULTS: We previously reported that higher levels of total soluble receptor for advanced glycation endproducts (sRAGE) predicts coronary heart disease in the Collaborative Atorvastatin Diabetes Study (CARDS). The extent to which genetics influences sRAGE levels needs to be determined in order for sRAGE to be used as a biomarker in the development of RAGE inhibitors. To identify loci associated with sRAGE levels in samples from CARDS participants (N=594), a genome wide scan using the Perlegen 6 SNP set (n =525,625 SNPs) was carried out followed by imputation of loci not directly typed using IMPUTE 2 and HAPMAP II data. sRAGE was measured using the R&D Systems Quantikine Immunoassay (assay coefficient of variation 4.4%). We tested for an association of both directly typed and imputed SNPS with sRAGE using multiple linear regression implemented in SNPTest adjusted for age, sex and BMI. Conditional haplotype-based testing implemented in PLINK was used to test for the independence of associations found. Results: 12 SNPS reached genome wide significance level of p<10-8 all of which were in the RAGE locus or +/- 150 kb upstream or downstream of it. These markers were tagged by two missense mutations rs2070600 (P =1.5x10-11 beta= -6.8(±0.9)) in the RAGE gene region on chromosome 6 and rs2071800 in the HLA-DQA2 region (P=1.2 x10-8 beta=-4.4(±0.7)) that were associated with sRAGE independently of each other. Together variation at these two loci explained 12% of the variance in circulating sRAGE, a level similar to that previously reported by us for known non-genetic determinants. We note that both these SNPs show ethnic variation in a direction that is consistent with the previously reported much lower total sRAGE in those of African Caribbean origin versus whites. In conclusion we report a novel tagging variant rs2071800 in the RAGE-HLA-DQA2 region independently associated with sRAGE levels and confirm the known association of sRAGE levels with rs2070600 in the RAGE gene. Variation in the frequencies of these tagging markers, as seen in the HAPMAP samples may explain the ethnic differences in circulating sRAGE levels.