Genetic determinants for susceptibility, progression and prognosis of prostate cancer

Abstract

Prostate cancer is the most commonly diagnosed form of non-skin cancer among men in developed countries. Although a large proportion of patients eventually die from the disease, many indolent tumors are found via prostate specific antigen (PSA) testing. However, todays diagnostic tools are unable to distinguish small localized tumors that will have a benign development from early stage aggressive disease. Thus, over-diagnosis and over-treatment are two major concerns in prostate cancer management. Genetics have been shown to play an important role for prostate cancer initiation with an estimated heritability of 58% and over 100 identified single nucleotide polymorphisms (SNPs) associated with prostate cancer risk. However, much less is known about the involvement of genes in the progression and prognosis of the disease.

The overall objective of this thesis is to enhance the understanding of genetic determinants for initiation, progression and prognosis of prostate cancer. The purpose of Study I was to develop a prediction model for prostate cancer susceptibility, based on the current knowledge of genetic risk variants. Furthermore, we aimed to study the potential role of established prostate cancer risk variants in disease progression among men with a localized disease (Study III). In Study II, the heritability of prostate cancer-specific survival among diagnosed men was estimated and a genome-wide search for genetic determinants of the same outcome was performed in Study IV. We found that a polygenic risk score model with 65 established prostate cancer risk SNPs and 68 novel variants optimally separates prostate cancer cases from healthy controls, with a prediction accuracy measured using the area under the curve (AUC) of 0.68. Furthermore, we observed that these 133 SNPs could be used for risk stratification; compared with an intermediate genetic risk score category (40%-60%), men with a low genetic risk score (lowest 5% percentile) had 84% decreased relative risk of prostate cancer and men with 5% highest risk scores had a four-fold increased relative risk.

Using a novel conditional likelihood approach for time-to-event data in brother pairs and father-son pairs, the heritability of prostate cancer survival was estimated to be 10%. We could also observe that common family environment had no effect (estimated to 0%) on prostate cancer survival. However, data simulations suggest that this may be underestimated. Furthermore, we could not find any association between SNPs and prostate cancer prognosis. None of 23 established prostate cancer risk SNPs investigated were found to be associated with disease progression in a cohort of men with localized disease. Moreover, in a genome-wide association study (GWAS) we did not find any association with prostate cancer survival at a genome-wide significant level.

In conclusion, with the current knowledge of prostate cancer genetics it is possible to identify men with high and low prostate cancer susceptibility risk. However, the predictive performance of established SNPs is not yet sufficient to be used alone in a screening program of prostate cancer. Furthermore, the findings in this thesis regarding prostate cancer progression and survival suggest that development of prostate cancer and progression to lethal disease may be two separate biological mechanisms that involve different genes. In order to identify genetic risk variants associated with prostate cancer progression, future studies should be designed to find common variants with very low penetrance or rare variants with moderate to large effect.