OR34-3 The Androgen Receptor Malignancy Shift Can Be Caused by Mutations Commonly Found in Primary Prostate Cancer

Prostate cancer (PC) is the second most diagnosed cancer and sixth leading cause of cancer mortality in men in the world. The androgen receptor (AR) plays key roles in the development of prostate tissue as well as in the development and progression of PC. Following male development, AR guides cytodifferentiation and homeostasis in benign luminal epithelial cells. However, in PC, AR instead drives the uncontrolled proliferation of these cells; this "AR malignancy shift" is a central event in tumorigenesis, as the AR becomes a primary driver of growth in malignant neoplastic cells. Despite this change in AR behavior having been recognized for decades, the details of the AR malignancy shift remain unclear. Using a ChIP-seq approach in primary human tissues, cell lines, and mouse models, we demonstrate that the AR malignancy shift occurs in every sample analyzed, suggesting that it is necessary for PC development. Using molecular and genetic techniques, we demonstrate that FOXA1, HOXB13, GATA2, JUN, and CEBPA are involved in the regulation of the AR malignancy shift. AR binding sites (ARBS) are enriched for FOX, HOX, and GATA motifs in PC cells but for JUN and CEBP motifs in benign cells. We show that mutations commonly found in localized PCs, including ERG, SPOP and FOXA1 alterations, can CAUSE the AR malignancy shift. Furthermore, these alterations are not transformative on their own but must be coupled to other mutations to transform cells, suggesting that the AR malignancy shift is not sufficient for tumorigenesis. We show that the AR malignancy shift occurs in mouse models of PC as well and that chronic low T, which is associated with increased PC risk and aggressiveness in humans, also causes the AR malignancy shift in mouse prostate tissue without the development of frank carcinoma. In conclusion, we have discovered a previously-unrecognized, fundamental tenet of PC, one which explains how and why AR signaling is different in cancer and benign cells. Our work has the potential to create genomic classifiers that accurately predict clinical outcomes and could be used to stratify patients with localized PC for specific treatments. Furthermore, our work suggests that the AR malignancy shift is a novel target for the treatment and/or prevention of PC.