Precision prevention of TMPRSS2:ERG prostate cancer.

Abstract

78 Background: Increased integration of tumor biomarker data into prostate cancer epidemiology studies is needed to identify molecular subtypes that underlie its etiology and progression. We hypothesize that the TMPRSS2:ERG gene fusion is a unique prostate cancer subtype that is etiologically distinct from cancers lacking TMPRSS2:ERG. Methods: We leveraged the Physicians’ Health Study and Health Professionals Follow-up Study cohort data on pre- and post-diagnostic lifestyle factors, inherited genetic variants, circulating biomarkers, and clinical data and follow-up for 30 years. We have a tumor repository of men with prostate cancer and tumor tissue microarrays. Using immunohistochemistry, we characterized TMPRSS2:ERG status for 1,491 incident prostate cancer cases in these cohorts, and also have biomarker data on a range of additional markers from immunohistochemistry and mRNA expression profiling. Results: Fifty percent of prostate cancer cases were ERG-positive. ERG-positive cancers show much higher expression of insulin/IGF signaling, PTEN loss, higher VDR expression, as well as expression of mismatch repair genes. In contrast, ERG-negative prostate cancer is characterized by increased presence of chronic inflammation and atrophy. We found higher pre-diagnostic free testosterone levels, but not other sex hormones, were associated with increased risk of ERG-positive (OR = 1.4, 95% CI = 1.0-1.8) but not ERG-negative disease (OR = 0.9, 95% CI = 0.7-1.2). Of 39 known genetic risk loci, six were significantly associated (p < 0.05) with ERG+ versus ERG- cancer (2 expected by chance). Prostate cancer risk factors such as taller height (an indicator of growth factors in puberty) are uniquely associated with ERG-positive prostate cancer. Moreover, we observe a complex interaction of components of insulin/IGF and ERG-status on prostate cancer mortality. Conclusions: TMPRSS2:ERG is a highly prevalent somatic event in prostate cancer that likely defines a unique molecular subtype of this common disease. Understanding the differences between these two prostate cancer subtypes may enhance opportunities for prevention.