Response, Resistance Mechanisms, and Biomarkers for Radiation and Immune Checkpoint Blockade

Abstract

Immune checkpoint inhibitors result in impressive clinical responses in multiple cancer types but optimal results may require combination therapy and approaches to enhance the anti-tumor T cell repertoire. We hypothesize that radiation (RT) combined with immune checkpoint inhibitors can safely improve response by enhancing the T cell repertoire available for re-activation by checkpoint blockade. Biomarkers and immune mechanisms of response and resistance can be identified to further improve outcome. A single-institution phase one trial of RT and an anti-CTLA4 antibody (anti-CTLA4) for metastatic melanoma using escalating doses of RT (6–8 Gy x 2–3) given to a single-index lesion stratified by anatomical site (lung/bone, liver/subcutaneous). This was conducted in parallel with pre-clinical mouse studies that utilized genomic and immune profiling to discover markers and determinants of response and resistance. Findings in mice were corroborated using patient biopsy and blood. In patients, no dose-limiting toxicities occurred and major tumor regressions were observed. The overall survival was 35% (95% CI: 19–64%, median follow-up 21.3 mo). In mouse models, addition of RT to anti-CTLA4 improved responses in irradiated and unirradiated tumors; however, resistance was common as it was in patients. Unbiased analysis revealed that resistance was due to upregulation of PD-L1 on melanoma cells and associated with T cell dysfunction (exhaustion). Accordingly, optimal response, survival, and immunity for melanoma and other cancer types required RT, anti-CTLA4, and anti-PD-L1/PD-1. Even with dual checkpoint blockade, omission of RT resulted in high rates of relapse. Investigation into how each treatment impacted immune mechanisms revealed that anti-CTLA4 inhibits T regulatory cells, RT enhances T cell receptor (TCR) repertoire diversity of intratumoral T cells, and anti-PD-L1 reverses T cell exhaustion. Together, dual checkpoint blockade expands T cells, while RT shapes the TCR repertoire of expanded peripheral clones. Similar to mice, low melanoma PD-L1 expression in clinical trial patients treated with RT + anti-CTLA4 predicts markedly longer overall (HR = 0.11, 95% CI: 0.02–0.62, P = 0.013) and progression-free (HR = 0.08, 95% CI: 0.01–0.76, P = 0.028) survival, while high PD-L1 levels predict rapid progression and persistent T cell exhaustion. RT can be safely combined with anti-CTLA4 in patients with metastatic melanoma. Elevated melanoma PD-L1 is a dominant resistance mechanism. In mice, the combination of RT, anti-CTLA4, and anti-PD-L1 promotes response through distinct mechanisms. These results will inform future trials and suggest a means of converting RT from a local treatment to one with systemic implications.