Genome Wide Association Study (gwas) for Identification of Single Nucleotide Polymorphisms (snps) Associated with Individuals Presenting Extreme Phenotypes of Tobacco Induced Non-Small Cell Lung Cancer (nsclc) Risk

ABSTRACT Aim: We analyzed the genome of individuals presenting extreme phenotypes of sensitivity and resistance to develop tobacco induced NSCLC to identify SNPs associated with these phenotypes. For this purpose, we used one of the most powerful GWAS platforms available. We hypothesized that SNPs may modulate individual susceptibility to carcinogens and that selection of extreme phenotypes would enrich the frequencies of alleles that contribute to the trait, thus increasing the power to identify them. Methods: From an identification cohort (n=3631) we selected caucasian heavy smokers that either developed NSCLC at an early age (cancer-cohort) or that did not present NSCLC at an advanced age (cancer-free cohort). We analyzed their genomic DNA using the array Illumina HumanOmni 2.5 Quad that includes over 2 million powerful markers selected from the 1000 Genomes Project, targeting genetic variation down to 1% minor allele frequency. Statistical significance of SNPs was calculated using logistic regression and Fisher's test. Results: 96 patients (48 per cohort) were selected. Mean age for the cancer and cancer-free cohorts was 49 years (range 38-55) and 76 years (72-84). Mean tobacco consumption was 41 pack-years (range 18-99) and 68 pack-years (40-120). GWAS identified 8 SNPs differentially expressed by logistic regression and 54 SNP by Fisher's test (p Function Gene Oncogenes MSX2 SOX11 Tumor supressors CSMD1 FOXF1 Tobacco induced NSCLC ABCB5 Regulators of transcription DROSHA HDAC9 KIAA0947 Regulators of inflammation PellinoE3 TRIM9 Related with cancer ABHD6 GRIK1 RAB40B SCN1A SLC24A2 SLC25A26 ZFYVE26 Conclusions: We identified candidate SNPs associated with the risk of developing tobacco-induced NSCLC in individuals with extreme phenotypes. Several identified SNPs were located within or near genes that constitute potentially relevant targets for modulation of cancer risk. Validation of the most significant SNPs in an independent set of individuals with similar phenotypes, selected from the EPIC-Spain project (www.epic-spain.com, n=40,000) is ongoing. Disclosure: All authors have declared no conflicts of interest.