Modeling and investigations of high-grade glioma to understand tumorigenesis and heterogeneity

Glioblastoma (GBM) is the most common primary malignant brain tumor. As per 2021 WHO classification, the corresponding diagnosis in children is called diffuse pediatric-type high-grade glioma (pHGG). In both cases tumors are histone 3 (H3) and isocitrate dehydrogenase 1/2 (IDH1/2) wildtype (wt). Standard of care includes maximum-safe surgical tumor resection followed by radiotherapy with concomitant and adjuvant temozolomide (TMZ). Prognosis is poor in both adults and children, with a median survival of 8 and 17 months, respectively. Hallmarks of the disease like high inter- and intra-tumor heterogeneity, bidirectional plasticity and tumor cell infiltration pose a formidable challenge for development of targeted therapies. This implicates a great need to develop relevant preclinical models to better understand the disease.In study 1, we used two platelet-derived growth factor (PDGF)-driven neonatal transgenic mouse models to uncover the underpinnings of supratentorial pHGG. By inducing the same oncogenic mutations in two different cell types we observed a significant difference in tumor incidence depending on developmental age and cell of origin. By performing functional, phenotypic and transcriptomic investigations we found that both mouse models were closely related to and representatives of human pHGG.In study 2, we used a cohort of 218 IDH1/2 wt GBM samples that were explanted under both serum-free culture (SFC) and serum-containing culture conditions. We found that SFCs were most successful (58.7%), but that there was a subset of GBM samples that could only be sustainably established in serum-containing medium (14.2%). We called these exclusive serum cultures (ESCs). Phenotypic investigations of ESCs revealed that they displayed all functional characteristics of GBM stem cells (GSC), i.e., persistent proliferation, extended self-renewal capacity and orthotopic tumor formation. On transcriptome level ESCs were distinct but most closely related to the mesenchymal GBM subtype, thus representing unique cell models of GBM. In study 3, we performed a comparative analysis of mouse GSC (mGSC) and human GSC (hGSC) cultures using assay for transposase-accessible chromatin using sequencing (ATAC-seq). Unsupervised cluster analysis of ATAC-seq data separated mGSC cultures according to their cell of origin (COO) and hGSC cultures into three clusters that each was related to one of the mGSC cell of origin groups. There were distinct functional differences between hGSC cultures of the three clusters and a significant difference in patient survival between two of the clusters. We conclude that cross-species epigenetic analysis supports an important role of the cell of origin for GSC malignancy which impacts patient survival