Abstract 1880: Modulation of ANRIL Expression is a Possible Mechanism Mediating the Association Between Chromosome 9p21 Polymorphisms and Coronary Atherosclerosis Risk
Background: Several recent studies have shown an association between single nucleotide polymorphisms (SNPs) on chromosome 9p21 and coronary artery disease (CAD) risk. These SNPs are outside any known coding region and the mechanisms mediating the association are still unknown. One possibility is that it reflects polymorphisms in factors affecting expression of adjacent genes in cis. We therefore investigate whether 9p21 SNPs are associated with allelic expression in cis of genes in this region, namely the two cell-cycle inhibitors (CDKN2A and 2B) and an antisense non-coding RNA of unknown function (ANRIL).
Methods: We genotyped 54 SNPs in a 270kb region in two cohorts of healthy volunteers (178 Caucasians and 309 mixed ancestry South Africans). Cis-acting effects were assessed by allelic expression ratios in cDNA from peripheral blood in individuals heterozygous for at least one of two transcribed SNPs tested for each gene. Allelic ratios were quantified by mass spectrometry and the association between the SNPs and allelic expression assessed using novel methods that allow combining data from several transcribed SNPs. 192, 185 and 202 individuals were informative for CDKN2A, 2B and ANRIL respectively.
Results: Allelic expression of CDKN2A, 2B and ANRIL is correlated (p<0.05) and influenced by cis-acting factors. Patterns in the two cohorts were similar. For each gene several SNPs were associated with expression (p<10−6 in the combined data set); some of the functional SNPs are shared whilst others are specific for particular genes. For example CDKN2A promoter SNPs are independently associated with CDKN2A and 2B expression (p<10−4) but not with ANRIL expression. Lead CAD SNPs (rs1333049, rs10757278 and rs2383206) are associated with ANRIL expression (p<10−9), but not with CDKN2A/2B. Other SNPs within the CAD risk haplotype (rs496892 and rs7044859) are associated with allelic expression of all three genes (p<10−3). CAD risk SNPs correlate with ANRIL under-expression.
Conclusions: CDKN2A, 2B and ANRIL expression is influenced by multiple cis-acting SNPs. CAD risk SNPs are highly associated with expression of ANRIL, but less consistently with CDKN2A/2B. ANRIL down-regulation is a potential mechanism by which these SNPs confer disease susceptibility.