Case-Control Study Using Array-Based Genotyping Produced a Panel of Genes Associated with Risk to Acute Myelogenous Leukemia (AML).

Multiple loci with small genetic effects are thought to be linked to AML pathogenesis. To detect such loci requires systematic screening of large number of single nucleotide polymorphisms (SNPs) within large study population. In this study we tested whether 3104 SNPs in 200 candidate genes were associated with risk to AML. Genes were considered potential candidates for their known or suspected roles in DNA repair system, pharmacogenomics and transcriptional regulation in hematopoiesis or putative pathways related to leukemogenesis. Selection of SNPs was performed using dbSNP and HapMap project databases, with emphasis on non-synonymous SNPs or haplotype tagging SNPs. To discover potential SNPs responsible for AML susceptibility, we conducted a case-control study using Affymetrix targeted genotyping 3K array. We applied this platform to analyze the samples from 309 de novo AML patients and 382 healthy controls. Genotype scorings were processed using GeneChip scanner 3000 TG(Affymetrix) and analyzed with GCOS software(Affymetrix). In total more than 255,000 SNPs were genotyped for this study. Samples with suboptimal call rates were excluded. Statistical testing were carried out using χ², Cochran-Armitage trend, Fisher’s exact, odds ratio, haplotype estimation, LD block definition. Here we report that 23 SNPs in 16 genes are associated with elevated or reduced risk to AML. Some of associated genes were transcription factors such as ZNF23 and ZNF233. While associated genes were distributed evenly among the whole genome, four associated genes were found on chromosome 1. Among those associated SNPs, two were located in coding region, one in exon-intron boundary, while the rest of SNPs were located in introns of associated genes. Among 23 SNPs identified, two intronic SNPs from 1^st intron of BAALC(Brain and acute leukemia, cytoplasmic) gene were associated with the reduced risk to AML. Haplotype estimation and linkage disequilibrium pattern of 9 SNPs including two associated SNPs will be presented. In conclusion, we identified SNPs responsible for AML susceptibility by candidate gene-based SNP array approach. These promising data when supported by further molecular validation would greatly enhance the current understanding of AML predisposition and diseases progression. Implication of polymorphic variants in AML etiopathogenesis will be presented and discussed.