A novel mouse model for checkpoint inhibitor-induced adverse events.

Immune checkpoint inhibitors have demonstrated significant efficacy in the treatment of a variety of cancers, however their therapeutic potential is limited by abstruse immune related adverse events. Currently, no robust animal model exists of checkpoint inhibitor-induced adverse events. Establishing such a model will improve our mechanistic understanding of this process, which in turn will inform design of improved therapies. We developed a mouse model to determine inflammatory toxicities in response to dual checkpoint blockade in the presence of syngeneic tumors. Mice from susceptible genetic backgrounds received intraperitoneal injections of anti-mouse PD-1 and CTLA-4 antibodies. The mice were monitored for weight loss and histologic evidence of inflammation. Blood was collected for basic metabolic panels and titers of anti-nuclear antibodies. In parallel, mice were also treated with prednisolone, which is commonly used to treat immune related adverse events among cancer patients. Among all the genetic backgrounds, B6/lpr mice treated with anti-CTLA-4 and anti-PD-1 antibodies developed more substantial hepatitis, pancreatitis, colitis, and pneumonitis characterized by organ infiltration of immune cells. Mice that developed tissue infiltration demonstrated high serum levels of glucose and high titers of anti-nuclear antibodies. Finally, while administration of prednisolone prevented the development of the inflammatory adverse events, it also abrogated the protective anti-tumor effect of the checkout inhibitors. Genetic background and treatment modalities jointly modified the inflammatory adverse events in tumor bearing mice, suggesting a complex mechanism for checkpoint inhibitor-related inflammation. Future studies will assess additional genetic susceptibility factors and will examine possible contributions from the administration of other anti-inflammatory drugs.