A Genome-wide association study of metastatic prostate cancer.

Abstract

160 Background: There are no biomarkers currently used in clinical practice that can predict which men will develop metastatic prostate cancer (mPrCa). Germline genetic variants that are highly associated with metastatic disease could be used for such prediction and may be identifiable from genome-wide association analyses (GWAS) conducted in appropriately scaled mPrCa case-control datasets. Methods: 451 mPrCa cases sequentially sampled from University of Utah/Huntsman Cancer Institute clinics were genotyped on the Illumina OmniExpress SNP platform and population-matched via principal components analysis to 1043 controls from dbGaP study phs000733 genotyped on the Illumina 5M SNP platform. The mPrCa cases and controls were analyzed via genome-wide association to identify SNPs highly associated with mPrCa. The dataset was then merged and re-analyzed with 176 lethal prostate cancer cases (identified in the Utah SEER Tumor Registry with a linked Utah death certificate including prostate cancer as a cause of death) in an attempt to identify a larger number of potential candidate SNPs associated with metastatic or lethal prostate disease. Results: The GWAS of mPrCa cases identified 4 significant SNPs at the 8q24.21 locus (p < 5e-8), as well as 9 SNPs in 6 previously unreported regions (5q14.1, 6p24.1, 8q21.3, 15q13.3, 15q22.2, and 17q25.1) with suggestive evidence for association (p < 1e-6). In addition, 30 SNPs with nominal evidence (p < 0.001) in the mPrCa-only GWAS appeared with more extreme p-values when the 176 lethal subjects were included in the analysis, and may be appropriate candidate variants to test in external datasets. Conclusions: In this case-control GWAS study of 627 men with mPrCa or lethal PrCa, germline SNPs at the 8q24.21 locus and 6 previously unreported regions may be associated with mPrCa. These findings should be validated in external datasets of men with mPrCa.