Recent examination of advanced prostate cancer (PCa) has suggested a major mechanism of progression to castration-resistant disease (CRPC) to be loss of the retinoblastoma (RB) protein. Along with its critical role in controlling cell cycle progression, RB is known to have important tumor-suppressor functions, and has been shown in PCa to be lost exclusively in late-stage disease. Additionally, loss of RB has been shown to correlate with increased E2F1 transcript and protein expression, via E2F-dependent mechanisms. Despite the vital role RB loss has been shown to play in this fatal stage of disease, the molecular underpinnings remain undefined. Thus, in order to elucidate these CRPC specific alterations, the current study utilizes isogenic models of RB loss in combination with genome-wide binding and transcriptional studies. Data presented herein demonstrate that loss of RB is frequent in CRPC, and represents the main mechanism of RB pathways disruption in PCa as detected through analyses of tumor samples and cell-free DNA. However, this phenomenon is not correlated with changes in proliferative indices, suggesting a role for RB loss outside of canonical cell cycle control. Further, RB loss induces significant genome-wide transcriptional alterations, including upregulation in Myc, E2F, and DNA-repair related pathways. Additionally, loss of RB significantly expands E2F1 binding capacity in castrate conditions, while largely maintaining the RB-intact E2F1 cistrome. Strikingly, while the current RB/E2F1 paradigm suggests that E2F1 exclusively occupies promoter regions of DNA in order to regulate transcriptional changes, RB loss induces marked reprogramming of E2F1 occupied regions, with a distinct increase in enhancer-bound E2F1. Further, motif analyses suggest divergence away from canonical E2F1 binding motifs after RB loss, specifically in regions of expanded E2F1 binding, and additionally suggest likely interaction of novel E2F1 cofactors under RB loss conditions. Interestingly, changes in E2F1 binding capacity after RB loss were seen to be distinct from those detected after androgen-induced RB inactivation, suggesting that the molecular alterations underlying RB loss are discrete from those resulting from functional inactivation. With respect to putative mechanism, it is of note that chromatin accessibility was not significantly altered to sufficiently explain the widespread changes in E2F1 cistrome, regardless of RB status, suggesting a mechanism outside simple opportunistic E2F1 binding after RB loss. Finally, interrogation of a CRPC patient tumor cohort showed predictive capacity for an “Expanded E2F1 Signature,” resulting from genes exhibiting gained E2F1 binding and differential expression after RB loss, in predicting loss of RB in patient samples, and indicating a novel E2F1-driven set of targets vital for CRPC transition in human disease. Together, these data present the first insight into E2F1 activity resulting from RB loss, and the role these changes play in progression to CRPC.

Citation Format: Christopher McNair, Kexin Xu, Amy C. Mandigo, Matteo Benelli, Benjamin Leiby, Daniel Rodrigues, Johan Lindberg, Henrik Gronberg, Mateus Crespo, Bram De Laere, Luc Dirix, Tapio Visakorpi, Fugen Li, Felix Y. Feng, Johann de Bono, Francesca Demichelis, Mark A. Rubin, Myles Brown, Karen E. Knudsen. Differential impact of RB status on E2F1 reprogramming in human cancer [abstract]. In: Proceedings of the AACR Special Conference: Prostate Cancer: Advances in Basic, Translational, and Clinical Research; 2017 Dec 2-5; Orlando, Florida. Philadelphia (PA): AACR; Cancer Res 2018;78(16 Suppl):Abstract nr B040.