Identification of anti-TIM-3 based checkpoint inhibitor combinations with activity in immunotherapy refractory melanoma models.

BACKGROUND: A significant percentage of melanomas are refractory to immune checkpoint inhibitor (ICI) monotherapies and combinations. As there are currently no effective second-line therapies available for ICI-resistant patients, we sought to identify novel checkpoint inhibitor combinations for future clinical evaluation. METHODS: We used responding and resistant SM1, SW1 and B16-10 melanoma models to explore the mechanism of action of anti-programmed cell death protein 1 (PD-1)+lymphocyte activation gene (LAG-3), anti-PD-1+cytotoxic lymphocyte associated protein 4 (CTLA-4), the anti-PD-1+LAG-3+CTLA-4 triplet, the anti-PD-1+T-cell immunoglobin and mucin domain 3 (TIM-3) doublet and the anti-PD-1+LAG-3+TIM-3 triplet. Flow cytometry, tetramer binding assays, T-cell depletion and analysis of the tumor-draining lymph nodes were used to determine the mechanisms of action of the ICI combinations. Single-cell RNA sequencing (scRNA-Seq) data from patients with melanoma was additionally analyzed to determine immune checkpoint expression across T-cell subsets and their association with ICI resistance. RESULTS: The PD-1+CTLA-4+LAG-3 triplet combination had good efficacy in the SM1 and B16-F10 melanoma models and led to a more favorable immune microenvironment (increased CD69+CD8+ T cells and gp-100-tumor specific T cells, reduced exhausted T cells) than either doublet. The ICI-resistant SW1 model did not respond to the PD-1+CTLA-4+LAG-3 triplet and was characterized by an accumulation of TIM-3+terminally exhausted CD8+T cells. Targeting of TIM-3 using the anti-PD-1+LAG-3+TIM-3 triplet led to complete tumor regressions in the SW1 and SM1 models. The TIM-3-based triplet decreased exhausted CD8+T cells and regulatory T cells, increased numbers of gp-100 reactive CD8+T cells and increased progenitor-exhausted CD8+T cells (TCF7+Ki67+ TIM-3-) numbers in tumor-draining lymph nodes. scRNA-Seq analysis demonstrated that TIM-3 levels were higher in the CD8+T cells of patients with melanoma who were non-responsive to immunotherapy. PD-1 and TIM-3 expression was strongly correlated in the CD8+T cells from patients with melanoma, providing the rationale for co-targeting these checkpoints. CONCLUSION: Melanoma models characterized by high levels of terminally exhausted CD8+T cells (TIM-3+) can be successfully treated with the anti-PD-1+LAG-3+TIM-3 combination. The observation that high TIM-3 expression in CD8+T cells is associated with ICI resistance suggests the potential utility of anti-TIM-3-based combinations as a second-line treatment strategy in patients with advanced melanoma.