Genome-Wide Association Study of Stroke in Sickle Cell Anemia.

Abstract 1528

Poster Board I-551

Stroke is a potentially lethal complication of sickle cell anemia (SCA) and one marker of sickle vasculopathy. Candidate gene studies conducted have demonstrated that stroke is associated with polymorphisms (SNPs) in several genes whose interactions can be used to build risk prediction models. For an unbiased discovery of the complex genetic basis of this complication, we conducted a genome-wide association study in 1387 SCA patients from the Cooperative Study of Sickle Cell Disease to identify single nucleotide polymorphisms (SNPs) associated with stroke. The data included 145 patients with at least one stroke event (cases), and 1242 stroke free patients (controls). Cases and controls had approximately the same median age (18 years) and similar gender composition (cases: 56% males; controls: 52% males). DNA was genotyped with the IIlumina Human610-Quad that includes approximately 600,000 SNPs and we removed samples with a call rate < 93%, and samples with a mismatch between gender reported in the database and heterozygosity of more than 5% SNPs in chromosome X. Error rate was estimated to be less than 5% based on agreement between known repeated samples and identical samples matched using genome-wide identity by descent using the software PLINK. We examined general, allelic, dominant and recessive associations of each individual SNP using Bayesian tests and scored the evidence of association of each model by its posterior probability. We assumed uniform probability on competitive models, so that the posterior odds of each model of association relative to the model of no association is equivalent to the Bayes factor (BF) and conducted extensive simulations to compute the expected number of false positive associations for different thresholds of the BF. The simulations showed that the false positive rate of the Bayesian decision rule changes with the allele frequency and suggested using a BF > 10,000 to reduce the expected number of false positive associations to less than 1 in 100,000 independent tests. Twenty-six SNPs passed this threshold, 15 SNPs were in intragenic regions and 10 SNPs were in known genes, including one SNP in the brain specific angiogenesis inhibitor BAI1 (rs11167147, odds for stroke in carriers of the AC or CC genotype = 0.25 relative to carriers of the AA genotype, BF>22,000) and one SNP in the regulator of angiogenes AIMP1 (rs7654865: odds for stroke in carriers of the AC or CC genotype = 0.10 relative relative to carriers of he AA genotype, BF>10,000). SNPs in other genes involved in angiogenesis (ANGPT1, ANGPT4 and TEK) were also associated with stroke, although none of the associations reached genome-wide significance. The regulation of angiogenesis is controlled by a balance between stimulators and inhibitors. BAI1 is a p53 target gene specifically expressed in the brain that is a transmembrane protein containing an extracellular domain with thrombospondin type-1 repeats that can exhibit anti-angiogenic activity. BAI1 is a mediator in the p53-signaling pathway; p53 has been shown to result in the decreased expression of VEGF and increased expression of BAI1. The VEGF system is integrated into the p53 transcriptional network and both pathways can be abnormal in SCA vessels. AIMP1 encodes a cytokine induced by apoptosis that is involved in the control of angiogenesis, inflammation and wound healing. It induces the expression of TNFRSF1A in endothelial cells and has anti-angiogenic functions via inhibition of HIF1α activities. HIF1α is involved in mediating angiogenic growth of endothelial cells. None of the SNPs in genes that we found associated with stroke in earlier studies reached genome-wide significance, although several SNPs in BMP6, ADCY9, EDN1, ERG, MET, SELP, TEK and TGFBR3 reached lesser statistical significance. We also looked for replication of SNPs that have been associated with stroke in the general population; rs12229103 (NINJ2) was significantly associated with stroke (BF>10). This SNP is within 20kb from rs12425791 that was found to be associated with stroke in the general population. Also SNPs in IMPA2 and AIM1 were significantly associated. Although confirmation of our genetic studies in an independent sample of individuals is needed and functional studies are warranted, our findings provide suggestive evidence for a major role of genes involved in angiogenesis in the modulation of stroke risk, a finding is in agreement with previous work suggesting that angiogenesis is dysregulated in SCA.