(AS1) Evaluation of GL261 and CT2A Murine Gliomas as Translation Models of Human Glioblastoma Using Functional Genomics and Single Cell Transcriptomics

Introduction: Murine models are critical for the discovery of disease mechanisms and therapeutic vulnerabilities in human glioblastoma, however findings in animals often fail to translate to humans. Here we employed functional genomics and single cell transcriptomics to characterize the similarities and differences between human and murine gliomas.

Methods: Single cell RNA-sequencing (scRNA-seq) was performed on in vivo CT2A and GL261 murine syngeneic glioma models and compared to published human GBM scRNA-seq profiles using non-negative matrix factorization (NMF) and functional annotation analyses. We also performed pooled loss-of-function CRISPR screens in CT2A and GL261 cell lines and compared essential fitness genes to those identified across 41 human GBM lines from the Cancer Dependency Map (Sanger).

Results: Using scRNA-seq profiles of >250,000 cells from CT2A and GL261 orthotopic tumors, we found that murine tumors resembled human GBMs, and recapitulated a similar degree of heterogeneity as seen in patient tumors. Specifically, GL261 tumors were biased towards a developmental-like phenotype whereas CT2A tumors resembled a mesenchymal-like phenotype. Characterization of immune-intrinsic gene programs showed that tumor microenvironments in mice and humans were associated with profound induction of T-cell differentiation, hypoxia, and IFN and TNF-alpha signaling, and that both GL261 and CT2A tumors led to recruitment of immunosuppressive CTLA-positive T-regulatory cells. We also compared CRISPR screen-derived genetic dependencies of murine glioma to human GBM. We found that CT2A cells recapitulated human GBM-specific essential genes with an AUROC of 0.71, including genes related to RNA processing and epigenetic regulation, but not proteoglycan biosynthesis or protein UFMylation.

Conclusion : Our study demonstrates the translational utility of GL261 and CT2A as models of human GBM, and reveals GBM-specific biology that can reliably be modeled and studied in mice. These results will serve as an invaluable resource for future translational studies.