Most prostate cancer (PCa) deaths are due to castration-resistant PCa (CRPC), following failure of androgen-deprivation therapy (ADT). ADT is the standard of care for patients with advanced PCa. However, nearly universal progression to castration-resistant prostate cancer (CRPC) occurs 2-3 years after ADT is initiated. Although there have been recent improvements in the treatment of CRPC, even the most promising therapies are still not curative. One approach to this problem is to improve the initial treatment of advanced prostate cancers, by combining complementary therapies with ADT, to prevent progression of such advanced cancers to CRPC. Immunotherapy with checkpoint inhibitors (CPIs) has not been effective in prostate cancers, perhaps because such cancers are “cold” (lacking cytolytic CD8 T-cells). Some cold tumors may be caused by infiltration of myeloid cell populations (tumor associated macrophages and myeloid-derived suppressor cells) into the tumor immune cell microenvironment (TIME). Recently, we found that in a PTEN-deficient mouse PCa model, castration induces an immunosuppressive state within the tumor that is concurrent with tumor recurrence. The response to castration/ADT is tri-phasic: a pro-apoptotic regression phase when tumor shrinks, followed by selection for a residual population of resistant tumor cells and finally recurrent growth as CRPC. Using PCa cell lines to model the first two phases of the response to ADT, we have shown that ADT induces apoptosis, thereby enriching for an ADT-resistant stem/progenitor population that we propose is the in vivo source of TNF. Mechanistically, in our model system the response to ADT is driven by the soluble mediators TNF and CCL2, which facilitate communication within the TIME. Specifically, a TNF-CCL2-CCR2 paracrine loop is induced between prostate cancer cells and non-tumor cells in the microenvironment: TNF produced by tumor cells acts on myofibroblasts to induce CCL2 production, which in turn recruits CCR2+ tumor-associated macrophages (TAMs). To investigate the ADT response within the TIME in an in vivo model of prostate cancer, we employed a prostate-specific PTEN-deficient mouse model (PbCre4 x PTENf/f). Castration caused the tumors to regress, consistent with initial phase of the response that is seen in the human disease. At late times post-castration (5-6 weeks), corresponding to the selection phase, we observed a coordinate increase in the stem/progenitor tumor cell population, as well as TNF and CCL2, within the TIME. Immunohistochemical staining of tumors 5 weeks post-castration revealed an increase in TAMs, and a decrease in CD8 T-cells, consistent with an immunosuppressive or immunoevasive state. This phenotype was reversed by a soluble receptor that binds TNF (etanercept). We also observed increased myeloid-derived suppressor cells (MDSC). Thus, following ADT, TNF derived from an ADT-resistant stem/progenitor epithelial tumor cell population promotes an immunosuppressive state via CCL2 in the TIME. Analysis of public human PCa data sets shows TNF and stem/progenitor marker expression are both increased in CRPC, consistent with our hypothesis that ADT drives the development of an immunosuppressive state via a TNF-CCL2-CCR2 axis. Our results set the stage for the future development of immunotherapies that could improve the efficacy of ADT.

Citation Format: John J. Krolewski, Kai Sha, Michalis Mastri, Dean Tang, Kevin Eng, Kent L. Nastiuk. Towards combining androgen deprivation and immunotherapy to prevent progression to castration-resistant prostate cancer [abstract]. In: Proceedings of the Fourth CRI-CIMT-EATI-AACR International Cancer Immunotherapy Conference: Translating Science into Survival; Sept 30-Oct 3, 2018; New York, NY. Philadelphia (PA): AACR; Cancer Immunol Res 2019;7(2 Suppl):Abstract nr A084.