Genetic variation in prostate-specific antigen-detected prostate cancer and the effect of control selection on genetic association studies

Duleeka W Knipe, David M Evans, John P Kemp, Rosalind Eeles, Douglas F Easton, Zsofia Kote-Jarai, Ali Amin Al Olama, Sara Benlloch, Jenny L Donovan, Freddie C Hamdy, David E Neal, George Davey Smith, Mark Lathrop, Richard M Martin

Research output: Contribution to journalArticle (Academic Journal)peer-review

26 Citations (Scopus)

Abstract

BACKGROUND: Only a minority of the genetic components of prostate cancer risk have been explained. Some observed associations of SNPs with prostate cancer might arise from associations of these SNPs with circulating prostate-specific antigen (PSA) because PSA values are used to select controls.

METHODS: We undertook a genome-wide association study (GWAS) of screen-detected prostate cancer (ProtecT: 1,146 cases and 1,804 controls); meta-analyzed the results with those from the previously published UK Genetic Prostate Cancer Study (1,854 cases and 1,437 controls); investigated associations of SNPs with prostate cancer using either "low" (PSA < 0.5 ng/mL) or "high" (PSA ≥ 3 ng/mL, biopsy negative) PSA controls; and investigated associations of SNPs with PSA.

RESULTS: The ProtecT GWAS confirmed previously reported associations of prostate cancer at three loci: 10q11.23, 17q24.3, and 19q13.33. The meta-analysis confirmed associations of prostate cancer with SNPs near four previously identified loci (8q24.21,10q11.23, 17q24.3, and 19q13.33). When comparing prostate cancer cases with low PSA controls, alleles at genetic markers rs1512268, rs445114, rs10788160, rs11199874, rs17632542, rs266849, and rs2735839 were associated with an increased risk of prostate cancer, but the effect-estimates were attenuated to the null when using high PSA controls (Pheterogeneity in effect-estimates < 0.04). We found a novel inverse association of rs9311171-T with circulating PSA.

CONCLUSIONS: Differences in effect-estimates for prostate cancer observed when comparing low versus high PSA controls may be explained by associations of these SNPs with PSA.

IMPACT: These findings highlight the need for inferences from genetic studies of prostate cancer risk to carefully consider the influence of control selection criteria. Cancer Epidemiol Biomarkers Prev; 23(7); 1356-65. ©2014 AACR.

Original languageEnglish
Pages (from-to)1356-65
Number of pages10
JournalCancer Epidemiology, Biomarkers and Prevention
Volume23
Issue number7
DOIs
Publication statusPublished - Jul 2014

Bibliographical note

©2014 American Association for Cancer Research.

Structured keywords

  • Centre for Surgical Research

Fingerprint Dive into the research topics of 'Genetic variation in prostate-specific antigen-detected prostate cancer and the effect of control selection on genetic association studies'. Together they form a unique fingerprint.

Cite this