61 Background: Immune checkpoint blockade with anti-CTLA-4 has emerged as a successful, novel form of cancer immunotherapy that acts directly on activated T cells instead of on cancer cells. CTLA-4 blockade can result in durable, long-term responses, but only a subset of patients treated benefit from this approach. Androgen deprivation therapy (ADT) is the first-line treatment for advanced prostate cancer, and it produces some of the most dramatic responses in clinical oncology. Clinically, the combination of ADT and anti-CTLA-4 was found to be more effective than ADT alone. To optimize the development of combination regimens, we examined the kinetics of the immune effects of ADT ± anti-CTLA-4 in a murine model of hormone-sensitive prostate cancer. Methods: Mice with established Myc-CaP prostate tumors were divided into 4 treatment groups: 1) untreated, 2) degarelix, 3) anti-CTLA-4, or 4) degarelix plus anti-CTLA-4. Tumor growth was measured biweekly. Tumor-infiltrating lymphocytes and T cell cytokine production was determined by flow cytometric analysis and serum IFN-gamma levels were assessed by ELISA. Results: We found that degarelix rapidly induces castrate levels of testosterone within 24 hours in mice and maintains these levels for at least 35 days. The combination of degarelix and anti-CTLA-4 improved median overall survival compared to degarelix alone (P<0.030, Mantel-Cox log-rank test). Tumor regression was associated with the production of the Th1-cytokines, IFN-gamma and TNF-alpha by tumor-infiltrating CD4+ and CD8+ T cells within 1 week of treatment initiation. This response peaked at 2 weeks after treatment initiation and was associated with elevated levels of serum IFN-gamma levels. In this model, anti-CTLA-4 monotherapy had no effect on tumor growth rates or intratumoral/sera cytokine levels. Conclusions: The combination of immune checkpoint blockade with anti-CTLA-4 and medical castration with degarelix is therapeutically effective in a murine model of hormone-sensitive prostate cancer, likely through an IFN-gamma-dependent manner. This model can be used to better identify potential post-treatment biomarkers that can be studied for associations with clinical outcomes.
Read full abstract