Optimizing STING Activity in Prostate Cancer Pre-Clinical Models

Abstract

We previously showed that radiotherapy (RT) combined with anti-PD-L1 induced abscopal response and prolonged overall survival in a pre-clinical mouse model of prostate cancer. We sought to optimize STimulator of Interferon Gene pathway (STING) activation in vitro and to enhance the abscopal response in our in vivo model.We quantified STING-activator cytosolic dsDNA following RT in 5 prostate cancer cell lines - MycCaP, PC3, 22Rv1, DU145, and LN-CaP. We tested the effects of parameters such as dosing, time after RT exposure, and fractionation. We evaluated PD-L1 expression via Western blotting in parallel experiments. We repeated the published RT and ant-PD-L1 regimen that induced an abscopal response in vitro, but we tested whether anti-PD-L1 administered 1 week before or after RT affected survival.With single doses of RT, the cell lines produced the most cytosolic dsDNA with doses between 12 and 15 Gy. Mathematical modeling of the combined cell lines showed peak production at 13.2 Gy. LN-CaP, the only androgen-sensitive cell line, produced 91-97% less cytosolic dsDNA than the other cell lines. Production of cytosolic dsDNA peaked 8-24 hours after RT and declined to baseline 72 hours after RT. RT doses of 8, 13, or 20 Gy were delivered at various frequencies including: once, once daily x 2 days, once daily x 3 days, every other day x 2 doses, or every other day x 3 doses. Cytosolic dsDNA production was not significantly different between the MycCap and PC3 cell lines. However, max production for MycCap cells occurred with 13 Gy delivered every other day x 3 doses whereas max production for PC3 cells occurred with 13 Gy delivered once daily x 3 days. PD-L1 expression was highest in MycCaP cells exposed correlated with production of cytosolic dsDNA throughout parallel experiments. Within the fractionation scheme, PD-L1 expression was consistently the highest with 13 Gy compared to either 8 Gy or 20 Gy, regardless of fractionation. Finally, median overall survival was unchanged in our mouse model with anti-PD-L1 administered either before or after RT. However, survival at 100 days was significantly different, with 40% vs 0% survival in the cohort treated with anti-PD-L1 before vs after RT.We defined optimal dosing, timing and fractionation schedule for generating cytosolic dsDNA in prostate cancer. Cytosolic dsDNA is correlated with expression of PD-L1, adding evidence that STING induces PD-L1 expression in prostate cancer as similarly reported in other malignancies. Though median overall survival was similar regardless of timing of anti-PD-L1, a percentage of mice exhibited prolonged survival with anti-PD-L1 administered prior to RT. These results guide optimal STING activation in prostate cancer-related pre-clinical studies and will provide the basis for future mechanistic studies regarding STING activation in prostate cancer.