Immunogenic neo-antigen discovery for personalized immuno therapy
Recent advances in cancer immunotherapy and genomic sequencing technologies have created promising opportunities for precision cancer medicine. Somatic coding mutation in cancer genomes may lead to amino acid alterations that generate immunogenic peptides, called neoantigens. These novel peptides are tumor-specific. An individual’s own MHC genotypes restricts the presentation of these tumor-specific peptides and thus there are a limited number of candidate epitopes for immune cellular recognition. Current studies have not comprehensively studied how many of these MHC-restricted neoantigens are present across tens of thousands of cancer. Nor have these studies determined how somatic mutations and their associated neoantigens are represented among different subclonal populations existing within individual tumors. To address these limitations, we have mapped the complex immunogenomic topology of neoantigen epitopes across 8,000 tumors representing 18 different cancer types. This neoantigen landscape accounts for how neoepitope candidates are distributed among the various clonal subpopulations existing within any given tumor.
Our study utilized genomic data from the Cancer Genome Atlas (TCGA). Our automated pipeline provides results from 8,000 samples from 18 cancer types that have somatic variant calls, copy number variation, whole exome sequence, and RNA-Seq data. Identification of somatic mutations that lead to highly immunogenic antigens involved five different steps; i) in-silico translation of identified mutations, ii) expression as measured per RNA-Seq, iii) patient’s own major histocompatibility complex genotype, iv) binding affinity specific to the MHC alleles for any given patient and v) occurrence of clonal subpopulations as demonstrated by intratumoral genetic heterogeneity.
In summary, we mapped the neoantigen landscape with quantitative representation of the genetic clonal diversity existing within individual tumors. In average, 2% of missense mutations from a patient are detected in RNA-Seq, with binding affinity to own HLA genotypes, and from the largest clonal subpopulation. Our study demonstrates the various relationships among the number of optimal antigens, the number of clonal subpopulations, the number of mutations, and clinical phenotypes such as tumor stage. Our pipeline facilitates the use of exome and RNA-Seq for precision immunotherapy as a potential diagnostic analytic process for immunotherapy.