Lifestyle, biochemical, and genetic risk factors for prostate cancer

• Main Author: Byrne, Karl Smith
• Other Authors: Key, Tim ; Wrightson, Ruth
• Published: University of Oxford 2017
• Online Access: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.748906

Despite considerable research there are no established risk factors for prostate cancer beyond age, family history, ethnicity, genetic factors, and insulin-like growth factor I. However, evidence from existing studies of risk factors for prostate cancer suggest that many lifestyle and biochemical exposures, and genetic variants may be risk factors for prostate cancer. Exposures investigated in this thesis are: vasectomy status, microseminoprotein-beta (MSP), human kallikrein 2 (HK2), and the lactase persistence SNP, rs4988235. Additionally, this thesis contains an analysis of how germline polymorphism, including rs4988235, may determine the intake of dairy produce. Vasectomy has been implicated as a risk factor for total and aggressive prostate cancer. However, vasectomy was not associated with risk of prostate cancer overall (hazards ratio (HR): 1.05; [95% CI 0.96-1.15]), with risk for high grade or advanced stage tumours (1.01; [0.84-1.21] & 0.83; [0.64-1.07], respectively), or for death from prostate cancer (0.88; [0.68-1.12]) in 84,753 men in the European Prospective Investigation into Cancer and Nutrition (EPIC). A previous prospective study in the Multi-Ethnic Cohort found a significant protective association of MSP with prostate cancer. In a nested case-control study (1,871 cases & 1,871 controls) plasma MSP concentrations in EPIC were inversely associated with prostate cancer risk after adjusting for total prostate-specific antigen (PSA) concentration (odds ratio (OR) for highest versus lowest fourth of MSP = 0.65; [95% CI 0.51-0.84], p-trend = 0.001). No heterogeneity in this association was observed by tumor stage or histological grade. Mendelian randomisation (MR) analyses suggest a causal protective association of MSP with prostate cancer risk (OR per unit increase in MSP for MR: 0.96; [95% CI 0.95-0.97]). Although HK2 has previously been found to improve discrimination for prostate cancer in predictive models, there has been no large prospective investigation of the association of HK2 with prostate cancer in men without elevated PSA. However, in EPIC plasma HK2 concentrations were not associated with prostate cancer risk independent of circulating concentrations of PSA in a nested case-control study (2,867 cases & 2,867 controls) (OR for highest versus lowest fourth of HK2 before adjustment for PSA = 7.09; [95% CI 5.82-8.65], p-trend = 0.001 vs. after adjustment for PSA = 1.29; [0.98-1.67], p-trend = 0.1). Further, there was no evidence that including HK2 in a model of prostate cancer risk based on PSA and age improved discrimination for prostate cancer overall or for high grade tumours (overall area under the curve (AUC): 0.816 vs. 0.816 or high-grade AUC: 0.752 vs. 0.752). The lactase persistence SNP, rs4988235, has previously been investigated in the estimation of prostate cancer risk as a marker of dairy intake. In these analyses rs4988235 was not associated with prostate cancer risk overall (per-allele OR: 1.01; [95% CI 0.99-1.04]), with risk for high grade or advanced stage tumours (0.99; [0.95-1.04] and 0.99; [0.93-1.04], respectively), or for death from prostate cancer (0.96; [0.90-1.03]) in the PRACTICAL consortium, a large (48,471 cases & 29,866 controls) prostate cancer genetics consortium. Further, in UK Biobank (up to 41,514 men) rs4988235 was only associated with the intake of dairy milk (difference between CC and TT: 19.9 g/d; [12.8-26.9]) and not the intake of other dairy products (yogurt, ice cream, and cheese). It is possible that alternative germline polymorphisms may determine differences in the intake of dairy milk, which may be useful for future MR studies into prostate cancer risk. However, no SNPs were genome-wide significant in relation to the intake of dairy milk in a discovery GWAS in the UK Biobank (22,041 men). Further, putative associations for SNPs significant at the lower tentative genome-wide significance threshold (p < 10^-5) were not replicated when investigated in an independent sample of men from the UK Biobank (50,701 men). In conclusion, there is no strong evidence that vasectomy, HK2, or the lactase persistence SNP as a marker of dairy intake, are associated with prostate cancer risk. Given the null GWAS for dairy milk intake, rs4988235 remains the strongest known genetic determinant of dairy milk intake. Lastly, observational and MR results provide compelling evidence that MSP is a potentially causal protective factor for prostate cancer risk. Future research should focus on replicating the putative causal association of MSP with prostate cancer and better understanding the purported role of MSP in tumour suppression or pathogen defense.