Abstract 2191: Identification of functional risk alleles in colorectal cancer-associated regions in African Americans

Abstract

Abstract Background. There are disparities in both colorectal cancer (CRC) incidence and survival between African Americans (AAs) and Non-Hispanic Whites (NHWs). Genetic susceptibility factors are believed to contribute to CRC development but have been inadequately studied in AAs. Our previous work showed both shared and independent CRC-associated risk alleles are present in some gene regions identified in genome-wide association studies (GWAS) in NHWs. In the present work, using a functional bioinformatics approach, we searched for risk alleles in these regions in AAs. Methods. 10 chromosomal regions (8q23.3, 8q24.21, 10p14, 11q23.1, 14q22.2, 15q13.3, 16q22.1, 18q21.1, 19q13.11, 20p12) from COGENT consortium meta-analysis were selected for DNA sequencing. Three additional high-interest regions were included. The region surrounding each associated SNP was analyzed using r2 plots, based on European-ancestry genotype data from HapMap Phase 3. DNAs from 63 individuals (25 AA CRC cases and 38 AA controls) were sequenced using Agilent SureSelect enrichment arrays and Illumina HiSeq platform technology. Variants from the targeted sequencing were combined with 1000 genomes variant data for a total of 16,193 variants. Potentially functional variants were identified based on the following criteria: the variant is (i) in a coding region, (ii) an expression quantitative trait locus (eQTL) or a DNAseI eQTL (DsQTL), (iii) in a miRNA binding site, and (iv) in a transcription factor binding site. Golden Helix software was used to process the data. Using the Sequenom Massarray platform, we genotyped variants in discovery and replication sets for a total of 1177 AA CRC cases and 1264 AA controls. We calculated odds ratios (OR) and 95% confidence intervals using logistic regression, assuming an additive genetic model and adjusting for age, sex, and African ancestry. Results. We identified 81 variants in coding sequences, 82 eQTLs, 5 dsQTLs, 38 variants in putative microRNA binding sites, 33 variants in possible transcription factor binding sites, and 6 variants that are in two or more of the preceding categories, for a total of 245 variants. After assay development and quality control, 163 SNPs were available for genetic analysis. Nominal associations (p<0.05) were detected between CRC and 17 variants. We have completed genetic analysis in the replication set for 8 of the 17 SNPs. The SNPs rs1862748 (OR, 0.73; P = 0.001) in the CDH1 region and 16888699 (OR, 1.39; P = 0.0006) in the EIF3H region were statistically significant in both the discovery and replication series. rs1862748 has previously been associated with AA CRC and was identified as a eQTL for ZFP90. rs16888699 was identified as an eQTL for UTP23. Conclusion. Our results suggest that rs1862748 and rs16888699 are CRC risk alleles, validating the bioinformatics approach. Functional studies will are now needed to characterize the genetic mechanisms underlying risk. Citation Format: Fabio Pibiri, Ricky Rechenmacher, Rosa Xicola, Rick Kittles, Robert Sandler, Temitope Keku, Xavier Llor, John Carpten, Sonia Kupfer, Nathan Ellis. Identification of functional risk alleles in colorectal cancer-associated regions in African Americans. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 2191. doi:10.1158/1538-7445.AM2014-2191