Abstract 2728: Characterization of Splicing-Derived Neoantigens Using Splicemutr Shows Their Independence from TMB and Potential As a Biomarker for Immunotherapy Response

Alternative-splicing dysregulation has been shown to play a role in cancer oncogenesis, conferring survival benefits to tumors. Yet, similar to mutation-caused tumor immunity, alternative-splicing dysregulation in cancer has also been shown to create neoepitopes capable of causing an immune response. In this work, we develop a new bioinformatics pipeline, splicemutr, to characterize alternative splicing as a missing source of antigens in cancer from a splice-junction perspective. The splicemutr pipeline identifies antigenic, splice variants by kmerizing differentially spliced transcripts and performing mhc-binding affinity prediction using mhcnuggets. To benchmark this algorithm on established splicing alterations, we first apply our approach to HPV-positive head and neck tumors and find that splicemutr shows enrichment of immunogenic trans-splicing events and evidence for selection against immunogen-causing splice-junctions when compared to normal. Further expanding this to pan-cancer analysis in The Cancer Genome Analysis (TCGA) enables us to evaluate the extent to which alternative splicing relates to common immunogenic biomarkers, including notably tumor mutation burden. Using splicemutr, we show that within 14 cancer types, splicing antigenicity does not correlate with the tumor mutation burden, showing splicing antigenicity to be independent from a common biomarker of tumor response to treatment. To further relate splicing antigens to the efficacy of immunotherapy, we apply our pipeline to compare splicing-based antigenicity in PD1 resistant 4T1MSI implanted Balb/cJ mice sensitized to nivolumab immunotherapy through combination therapy with a BET inhibitor (BETi) and in post-treament samples from an advanced melanoma clinical trial cohort treated with combination nivolumab and ipimulimab. In the mouse experiment, the nivolumab-BETi combination arm showed marked tumor shrinkage compared to the control and nivolumab alone arms. Using splicemutr, we show that the splicing antigenicity increases from combination therapy to nivolumab alone to the untreated arm, suggesting depletion of splicing-based antigens in those samples that respond well to treatment. In the advanced melanoma clinical trial cohort treated with combination nivolumab and ipimulimab, we show that the splicing-based antigenicity is higher in the non-responders than responders. Based upon our mouse data, we hypothesize that the post-treatment human data has decreased splicing-based antigens as the tumor clones containing them may be targeted by the immune system and depleted during successful immunotherapy response. Altogether, these data suggest the potential role of splicing as candidate antigens in cancer and use of our splicemutr pipeline for biomarker discovery based upon splicing antigenicity. Citation Format: Theron Palmer, Michael Kessler, Ludmila Danilova, Archana Balan, Mark Yarchoan, Neeha Zaidi, Tamara Lopez-Vidal, Ali Saeed, Jessica Gore, Elizabeth M. Jaffee, Alexander Favorov, Valsamo Anagnostou, Daria Gaykalova, Rachel Karchin, Elana Fertig. Characterization of splicing-derived neoantigens using splicemutr shows their independence from TMB and potential as a biomarker for immunotherapy response [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 2728.