Establishing A Mouse Model of Checkpoint Inhibitor-Induced Colitis: Pilot Experiments and Future Directions.

RATIONALE: Checkpoint inhibitors show significant promise for treatment of a variety of cancers, however, their efficacy is limited by variable anti-tumor responses and autoimmune side effects, both of which are unpredictable and poorly understood. Currently, no robust animal model of checkpoint inhibitor-induced colitis exists. Establishing such a model will improve our mechanistic understanding of this process, which in turn should inform design of improved therapies.

METHODS: Wild-type C57/B6 mice received intraperitoneal injections of anti-mouse CTLA-4 and/or anti-PD-1 antibodies every three days for a total of five doses. The mice were monitored for weight loss and histologic evidence of colitis. Similar experiments were conducted with additional treatments of either broad-spectrum antibiotics or dextran sodium sulfate (DSS) in WT mice, or in three different genetic backgrounds: IL-10^-/-, IL-10rb^-/- or CD5^-/-/CD6^-/-.

RESULTS: Anti-CTLA-4 and/or anti-PD-1 antibodies administered to WT mice, even at doses four times the maximum equivalent treatment dose in humans, did not elicit weight loss or signs of colitis. Similarly, perturbing gut epithelial barrier and immune functions with broad-spectrum antibiotics or sub-clinical doses of DSS did not induce colitis in treated mice. Finally, genetic defects in treated mice known to contribute to inflammatory bowel disease and loss of immune tolerance did not elicit colitis.

CONCLUSIONS: Multiple susceptibility factors utilized in other models of colitis failed to cause disease in treated mice, suggesting a unique mechanism for checkpoint inhibitor-induced colitis. Future studies will assess additional genetic susceptibility factors, and will also examine possible contributions from alterations in the gut microbiota.