A polygenic two-hit hypothesis for prostate cancer