IMMU-66. PHYSIOLOGICALLY RELEVANT MURINE GLIOMA MODELS THAT REPRESENT GBM HETEROGENEITY AND ENABLE THE NEXT GENERATION OF DISCOVERY AND THERAPY EFFICACY STUDIES

Abstract To transform cancer drug development pipelines and gain a deeper mechanistic understanding of cancer: immune interactions, we desperately need better experimental models that faithfully recapitulate the cellular heterogeneity, complex and dynamic cell:cell interactions, and therapy responses in humans. To address this need, we developed and characterized a cohort of five new mouse glioma models in the C57BL6/j background. These transplantable tumorsphere models were derived from spontaneous gliomas that formed in genetically engineered mouse models with different oncogenic drivers, including EGFRviii, PDGFA, PTEN, p53, and NF1 mutations. They show a range of responses to standard of care (TMZ+IR) similar to GBM patients and all are resistant, to immune checkpoint inhibitors, anti-PD1+CTLA4, treatment, unlike the GL261 model and similar to GBM patients. We deeply characterized each model, including immune phenotyping and single-cell RNA-sequencing analyses. We also performed a cross-species comparison of human and mouse GBM and stromal cell phenotypes at the single-cell level. Our results show that new mouse models recapitulate the heterogeneity of glioma cell states observed in individual patients and, importantly, also the complex and immune suppressive TME. We demonstrate at the single-cell level that glioma-associated T and myeloid cell types are molecularly similar and conserved across species. Leveraging these data sets and experimental models, we are testing now novel immunotherapy targets for GBM.