Characterizing the combined role of rare and common variants for prostate cancer risk

Abstract

Prostate cancer (PCa) remains a public health problem as it is one of the most frequently diagnosed and most heritable non-skin cancers in men globally. Unfortunately, the disease etiology of PCa is not yet fully understood, but age, family history, race and ethnicity, and genetic factors are widely reported risk factors. Previous studies have shown that risk for PCa depends on both rare and common variant risk separately, but the combined effect of them has not been explored in depth. Here, we used logistic regression models to validate a common variant polygenic risk score (PRS) of 451 PCa risk variants and REGENIE to conduct targeted gene-based analyses of rare, pathogenic coding variants, defined as having a ‘HIGH’ VEP Impact or a ‘Pathogenic/Likely Pathogenic’ ClinVar annotation. Subsequently, we conducted a combined PRS and rare variant carrier status interaction analysis in 82,802 participants from the All of Us (AoU) Research Program through logistic regression models. Our results validated the PRS’ ability to distinguish between PCa cases and controls and confirmed associations for 6 previously reported PCa risk genes (BIK, FAM111A, HOXB13, BRCA2, CHEK2 and ATM). We found evidence of an additive interaction between the PRS and rare variant carrier status, with carriers defined as those with 1 pathogenic coding variant in any of the 6 genes mentioned above. We observe significantly higher risk estimates in carriers in the 80-100% PRS (OR = 4.61; 95% CI: 3.86,5.50) than non-carriers in the 80-100% PRS (OR = 3.11; 95% CI: 2.89, 3.36). Taken together, our results demonstrate the importance of jointly considering the effect of common and rare variants in PCa risk estimation to improve risk stratification.