Endogenous T-cell responses to cancer-specific mutations (neoantigens) are a common denominator of successful immunotherapies. However, endogenous neoantigen-specific T cell immunity in immunotherapy refractory tumors remains poorly characterized. Pancreatic ductal adenocarcinoma (PDAC) is the prototypical immunotherapy refractory cancer: response rates to single-agent checkpoint blockade are <2%, and 5-year survival with conventional therapies remains <7%. In these rare 7% of PDAC patients that survive > 5 years, T cell immunity has been linked to their exceptional outcome however the relevant antigens remained unknown. We sought to identify the landscape, antigenic potential, and T cell functional states induced by neoantigens in rare long-term pancreatic cancer survivors. Using genetic, immunologic, and computational techniques, we recently reported that tumors with the highest predicted neoantigen number and the greatest density of CD3+CD8+ infiltrates together, but neither parameter alone, could identify long-term survivors, suggesting differential inherent neoantigen immunogenicity. To investigate the specific neoantigen qualities promoting differential immunogenicity, we developed a fitness model that quantifies neoantigen immunogenicity based on estimations of the relative MHC binding affinity of each neoantigen to its wild type counterpart, as well as a nonlinear dependence on sequence similarity of neoantigens to known antigens. Our neoantigen quality fitness model identified long-term survivors in two independent PDAC datasets, as well as predicted survival in anti-CTLA4 treated melanoma patients, and anti-PD-1 treated lung cancer patients. In long-term PDAC survivors, we detected high quality neoantigen-specific intratumoral T cell clones persisting in the blood up to 12 years after primary tumor removal, with both unique clonal and phenotypic profiles. Our results identify neoantigens with unique qualities as T cell targets in both endogenous and immunotherapy treated cancers. Neoantigen quality can therefore inform rational neoantigen immunogenicity predictions that may guide patient and target selection for immunotherapies.

Citation Format: Marta Luksza, Joanne P. Leung, Alexander Solovyov, David Redmond, Miriam Merad, Sacha Gnjatic, Christine Iacubuzio-Donohue, Ronald P. DeMatteo, Timothy A. Chan, Jedd Wolchock, Steven D. Leach, Benjamin D. Greenbaum, Taha Merghoub, Vinod P. Balachandran. Mapping immune recognition of non-self neoantigens in human pancreatic cancer [abstract]. In: Proceedings of the Fourth CRI-CIMT-EATI-AACR International Cancer Immunotherapy Conference: Translating Science into Survival; Sept 30-Oct 3, 2018; New York, NY. Philadelphia (PA): AACR; Cancer Immunol Res 2019;7(2 Suppl):Abstract nr IA34.