SpliceIO: A robust platform for the discovery and validation of splicing-derived neoantigens.

Abstract

2581 Background: Splicing dysregulation is a major hallmark of cancer, affecting tumor progression, metastasis, and therapy resistance. Aberrant alternative splicing plays a critical role in generating tumor-specific neoantigens targetable through an arsenal of immunotherapeutic modalities. Such neoantigens cannot be discovered using conventional DNA-based tools; they require specialized software for alternative splicing analysis using RNA-seq data. Methods: Envisagenics has developed SpliceIO, a proprietary software platform for splicing-derived neoantigen discovery. SpliceIO uses RNA-seq data to uncover extracellular neoantigens encoded by transmembrane proteins or presented to immune cells through the MHC complex. This predictive ensemble maximizes the value of RNA-seq data to predict fundamental aspects of neoantigen activity, such as their protein localization, MHC binding affinity, epitope accessibility and cancer pathway relevance. Here, we summarize the discovery and validation of novel splicing-derived neoantigens identified using SpliceIO in acute myeloid leukemia and colorectal cancer. Results: We analyzed >10,000 RNA-seq samples from the four tumor types and controls derived from 32 normal tissues. To maximize the prevalence of neoantigen target candidates in patient subpopulations, RNA-seq samples were pre-stratified using SpliceSlice, a proprietary software for spliceosomal profiling. As a result, patients sharing common spliceosome alterations were grouped into 11 coherent subsets that are distinct from normal tissue. Using this approach, we have discovered several promising candidates that we prioritized for experimental validations. SpliceIO target candidates were validated experimentally in cancer cell lines and primary tumor tissues to confirm RNA expression, translation, and membrane localization. Additionally, we performed proteomics to evaluate the presence of neopeptides generated from the unique splicing events. Using this approach, we have discovered and validated several tumor specific splicing derived neoantigens that are also recurrent in >15 % of the population. Conclusions: Together, our results strongly support the role of aberrant splicing as a major source of tumor-specific neoantigens. Our combined computational and experimental approach promotes novel target discovery, enabling the development of novel immunotherapeutics for cancer treatment.