Improving the validity of murine and human models of high-grade glioma

High-grade gliomas (HGG) are devastating brain tumors with few effective therapies and a median overall survival time of 1-5 years, depending on their severity and other factors. Molecular characterization of these tumors has revealed unique genetic, epigenetic, and transcriptomic patterns that correlate with patient age, tumor location, and prognosis, among other clinical variables. However, therapies targeting these subgroup-specific alterations that have shown preclinical promise have largely failed to translate to favorable patient response. One reason for this failure is likely a lack of model systems that faithfully recapitulate HGG features such as their intratumoral heterogeneity, complex mutational landscapes, and germline genetic variation impacting tumor biology. Thus, improvement upon available HGG models is critical for achieving true personalized therapies for this diverse class of tumors. In this work, we first interrogated the importance of host germline genetic variation on HGG growth using the Collaborative Cross. Second, we probed existing and new work on genetic alterations and transcriptomic signatures in HGG including EGFR and IDH, setting the stage for more genetically accurate HGG models. Finally, we presented a novel, human induced pluripotent stem cell model of H3.3K27M diffuse midline glioma (DMG), a particularly devastating subset of HGG. We used this model to show that the combination of H3.3K27M and activation of platelet-derived growth factor receptor alpha (PDGFRA) confers unique phenotypes and therapeutic vulnerabilities including increased amphiregulin production, a metabolic switch, and sensitivity to combination treatment with ONC201 and trametinib. Taken together, these chapters explore novel methods of modeling HGG and underscore the importance of precise molecular stratification for the development of targeted therapies.