A germline variant in the TP53 polyadenylation signal confers cancer susceptibility

Abstract

To identify new risk variants for cutaneous basal cell carcinoma, we performed a genome-wide association study of 16 million SNPs identified through whole-genome sequencing of 457 Icelanders. We imputed genotypes for 41,675 Illumina SNP chip-typed Icelanders and their relatives. In the discovery phase, the strongest signal came from rs78378222[C] (odds ratio (OR) = 2.36, P = 5.2 × 10(-17)), which has a frequency of 0.0192 in the Icelandic population. We then confirmed this association in non-Icelandic samples (OR = 1.75, P = 0.0060; overall OR = 2.16, P = 2.2 × 10(-20)). rs78378222 is in the 3' untranslated region of TP53 and changes the AATAAA polyadenylation signal to AATACA, resulting in impaired 3'-end processing of TP53 mRNA. Investigation of other tumor types identified associations of this SNP with prostate cancer (OR = 1.44, P = 2.4 × 10(-6)), glioma (OR = 2.35, P = 1.0 × 10(-5)) and colorectal adenoma (OR = 1.39, P = 1.6 × 10(-4)). However, we observed no effect for breast cancer, a common Li-Fraumeni syndrome tumor (OR = 1.06, P = 0.57, 95% confidence interval 0.88-1.27).

Figure 1

Overview of single-point SNP association data obtained from genomic sequencing in the 17p13 region covered by haplotypes A6 and A8. The region shown is chr17:7,186,095–7,680,389 (HG18 Build 36). The upper panel shows BCC association P values for SNPs in the region identified by whole-genome sequencing of 457 individuals. We determined association by two-way imputation (see the text for details); only SNPs with P < 0.01 are plotted. The positions of the TP53 SNP rs78378222 and the newly discovered SNP giving the second-highest signal in the region (chr17:7,640,788) are indicated. The locations of UCSC genes in the region are shown in the middle panel. The lower panel shows recombination rates calculated as described previously from HapMap2 release 22 data.