TMIC-55. CHARACTERIZATION OF TUMOR-MICROENVIRONMENT INTERACTIONS IN GLIOBLASTOMAS AT THE SINGLE-CELL LEVEL

Abstract

Abstract Glioblastomas are malignant brain tumors that carry a poor prognosis. The tumor microenvironment has been identified as an important regulator of tumor growth and may represent a novel target for therapy. Transcriptional subtypes of glioma are a major source of heterogeneity of expression in gliomas. Gliomas are highly heterogeneous diseases and can be classified into different subtypes including proneural, classical and mesenchymal tumors. We hypothesized that different subtypes of glioma will have different microenvironmental composition and exhibit distinct responses to therapies. To understand whether gliomas induced by different oncogenic drivers affect microenvironment composition, we induced mouse gliomas using a PDGFB and dnp53 driver to model proneural glioma and HRasV12 and dnp53 to model mesenchymal glioma, respectively. We performed single cell transcriptomic profiling to characterize the tumor microenvironment in these glioma models. We found that in the PDGFB/dnp53 glioma model had a large microglia contribution with about 30% tumor-associated microglia. In contrast, the HRasV12/dnp53 glioma model had only sparse microenvironmental cells. In addition, microglia in each model displayed subtype-specific gene expression programs, with microglia in the HRasV12/dnp53 tumor model expressing increased antigen presenting genes while increased levels of osteopontin in the PDGFB/dnp53 tumor model. To determine the tumor-microenvironment interactions, we performed receptor-ligand analysis using CellPhoneDb to identify secretory ligands that support tumor cell growth in microenvironmental cells. We found that tumor-associated pericytes are an important source of growth factor ligands in both PDGFB/dnp53 and HRasV12/dnp53 glioma models. We are investigating the contribution of microglia and pericytes to tumor growth and survival through cell depletion and pharmacological inhibition and determining whether the differences in tumor microglia between the models affects sensitivity to therapies including immunotherapy. Understanding how tumor-intrinsic signaling modulates microenvironment niches and tumor-microenvironment communications aids rational design of combinations targeting the malignant brain tumors.