Poster #16: Darshan Mehta

Personalized antigen discovery for cancer vaccines using long-read RNA sequencing

Darshan Mehta, B.S. (1); Matthew Mouck, B.S. (2); Anthony C. Wang, MD (3), and Yi Xing, PhD (1, 4, 5) (1) Center for Computational and Genomic Medicine, Children’s Hospital of Philadelphia, Philadelphia, PA, USA (2) Graduate Group in Genomics and Computational Biology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA (3) Department of Neurosurgery, David Geffen School of Medicine, University of California, Los Angeles, Los Angeles, CA, USA (4) Department of Biomedical and Health Informatics, Children’s Hospital of Philadelphia, Philadelphia, PA, USA (5) Department of Pathology and Laboratory Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA

Personalized cancer vaccine development requires identification of tumor-restricted antigens. Tumor antigens derived from RNA dysregulation, including aberrant alternative splicing (AS), could represent an extensive but largely unexplored source of immunotherapy targets. Although short-read RNA sequencing (RNA-seq) is commonly used for transcriptomic analyses, long-read RNA-seq can better characterize complex splicing events and identify open-reading frames. We hypothesized that long-read RNA-seq can facilitate discovery of personalized splicing events and of their correspondingly encoded antigens for cancer vaccine treatments. Methods: We developed a cancer vaccine pipeline to discover personalized antigens from AS junctions using long-read RNA-seq and applied it to our vaccine cohort of 6 pediatric high-grade glioma (pHGG) patients. We performed deep nanopore long-read RNA-seq on the respective tumor specimens, generating ~110 million reads. Outlier abnormal AS junctions were identified through a comprehensive statistical framework comparing patient junction usage to usage across 39 tissue types from the short-read RNA-seq GTEx dataset to ensure tumor enrichment and to mitigate off-tumor effects. To account for differences between the long-read and short-read RNA-seq methods, nominal outlier junctions were further selected against an in-house long-read dataset of 28 normal tissues, spanning distinct tissue types and brain subregions. Peptides encoded along outlier junctions were extracted and assessed for compatibility to each patient’s HLA type. Results: A personalized antigen list was curated for each patient. A median of 341 tumor-enriched junctions per patient was identified after applying our pipeline, encoding a median of 574 HLA-specific antigens. Moreover, 62 junctions and 9 antigens were common in at least 4 patients, suggesting the potential for off-the-shelf antigens. Conclusions: Our pipeline for cancer vaccine antigen discovery uses long-read RNA-seq to expand the antigen space, yielding a personalized list of peptides spanning splicing junctions. These antigens are being used to treat patients in a phase 1 clinical trial.