IMMU-28. THE INTRA-TUMORAL SPATIAL HETEROGENEITY OF T CELL ANTIGENS IN GLIOBLASTOMA: AN INTEGRATED MULTI-OMICS APPROACH

Abstract

Abstract This study investigates the intratumoral heterogeneity of T cell antigen presentation in 15 newly diagnosed glioblastoma patients. Our multi-omics approach includes mass spectrometry-based immunopeptidome analysis, next-generation sequencing (whole-exome and RNA sequencing), and imaging mass cytometry performed on biopsies deriving from the necrotic core (NEC), the gadolinium contrast-enhancing region (T1), and the peritumoral infiltrating zone (INF, 5-ALA positive). A total of 24493 unique HLA class I and 17394 unique HLA class II peptides were identified. Comparative profiling of peptides from our study and a benign tissue database (in-house; n = 429 donors combined with HLA ligand atlas (https://hla-ligand-atlas.org)) revealed that 20% of HLA class I ligands are glioblastoma exclusive. Of these ligands, 21%, 21%, and 15% were exclusively presented in the INF, T1, or NEC zone, respectively. Here, we focused on the INF-specific antigens as this zone is likely to remain after tumor resection and gives rise to glioblastoma recurrence. Interestingly, two INF-specific HLA ligands showed a frequency presentation of 45% in our patient cohort. One ligand originated from BAALC (Brain and acute leukemia cytoplasmic protein) and one from NCAN (Neurocan core protein), which are both known to be glioblastoma-associated proteins. Immunogenicity of pre-selected candidate antigens was assessed with autologous expanded T cells and revealed a set of novel promising targets for immunotherapy. Integrated RNA/DNA sequencing also enabled the identification of neoepitopes deriving from tumor- and region-specific mutations. Furthermore, imaging mass cytometry contributed to investigating the immune compartment of the tumor microenvironment in high dimensionality and spatial resolution. In conclusion, our approach characterized the intra-tumoral regional heterogeneity of both infiltrating immune cells and tumor antigens. This multi-omics spatial atlas of the immune landscape can be used for the informed design of immunotherapy strategies against glioblastoma.