Integrative multi-omic profiling of the neoantigen landscape of glioblastoma for the development of therapeutic vaccines reveals vast heterogeneity in immunogenic signatures.

Introduction: Glioblastoma (GBM) is the most common primary brain malignancy. Few neoantigens have been tested in trials as the cancer vaccine against GBM. Methods: To better understand the neoantigen landscape and its associated tumor microenvironment (TME) for the optimized vaccine design of our initiated GBM trial, we apply the integrative multi-omics approach to comprehensively profile the mutation, HLA typing, TCR/BCR repertoire, immune cell components on the tumor tissue and peripheral blood mononuclear cell (PMBC) specimen of 24 GBM patients. Results: On average, 148 mutated genes and 200 mutated sites per patient were identified, with no predominant mutated sites and genes in this cohort. Diversified HLA genotypes and expression rate across A, B, and C alleles, with A30:01&A11:01, B13:02, and C06:02, as the most frequent genotypes at respective alleles. Clustered CDR3 of TCR/BCR existed in tumor tissue with decreased richness compared with PMBC. NK and Th1 cells were revealed as the predominant immune cells within the tumor microenvironment (TME). Neoantigens were feasible predicted and designed for each patient, with an average number of 107. Very few neoantigens were shared by more than two patients and no dominant neoantigen could be identified. A minimum of 11-peptide bulk was required to cover this 24-patient cohort, guaranteeing each patient could have at least one neoantigen. Discussion: In summary, our data reveals a heterogeneous landscape of the neoantigen and its associated immune TME of GBM, based on which a peptide bulk is feasibly developed to cover these patients as a cohort.