Date of Award
Doctor of Philosophy (PhD)
Glioblastoma (GBM) is the most common intrinsic central nervous system malignancy in adults, accounting for approximately 45% of such cancers. Despite advances in chemo- and radiotherapeutic approaches for various malignancies over the past decade, GBM remains an incurable disease with a dismal prognosis. Even after treatment with the current standard of care, which consists of maximal safe surgical resection, radiotherapy, and both concomitant and adjuvant chemotherapy with temozolomide, median survival is only approximately 17 months. Both treatment failure and difficulties in developing novel targeted therapies for GBM have partly been attributed to the molecular and cellular inter- and intra-tumoral heterogeneity of these neoplasms. Two major questions arise from these observations: 1) What is the extent and clinical relevance of genetic diversity in GBM? And 2) do any shared mechanisms control the malignant phenotype of GBM cells?We first sought to profile the genetic inter- and intra-tumoral heterogeneity across ten patients harboring isocitrate dehydrogenase 1 (IDH1) wildtype GBMs, which represent 90-95% of all GBMs. Diagnostic workflows for GBM patients increasingly include DNA sequencing-based analysis of a single tumor site following biopsy or resection. We hypothesized that sequencing of multiple sectors within a given tumor would provide a more comprehensive representation of the molecular landscape and potentially inform therapeutic strategies. We demonstrated, using image-guidance directed sampling of two to four sectors of contrast-enhancing areas of IDH1 wildtype GBM tumors, that whole-exome sequencing of individual sectors reveals a spatially divergent mutational landscape. In two extreme cases of regional heterogeneity, we described, for the first time, treatment na ve tumors with region-specific hypermutator phenotypes. In remarkable contrast to the spatial diversity of the overall mutational landscape, we demonstrated that TERT promoter mutations are unique in being recurrent in all analyzed tumors and clonal in all tumor sectors. Finally, we examined the potential therapeutic consequences of multisector sequencing data and found that multi-site analyses may be necessary to accurately characterize individual GBM tumors and identify meaningful therapeutic options. Next, we asked whether shared epigenetic/transcriptional mechanisms might control the malignant phenotypes of genetically diverse GBM cells. GBM tumors are heterogeneous and contain a tumor-initiating pool known as glioblastoma stem-like cells (GSCs). GSCs are therapy resistant and may drive recurrence post-treatment. We reasoned that regulation of the pluripotency-related transcription factor SOX2, which is indispensable for the tumorigenicity of GSCs, may represent one such mechanism. First, we found that the mitotic E3 ubiquitin ligase, CDC20-Anaphase-Promoting Complex (APC), drives the invasiveness, self-renewal, and tumorigenic capacity of multiple, genetically heterogenous primary GSC lines. Mechanistically, we found that CDC20-APC operates through SOX2 to control GSC phenotypes by regulating SOX2 protein stability and transcriptional activity. Second, we used immunoprecipitation followed by mass spectrometry (IP-MS) to identity the E3 ubiquitin ligase TRIM26, previously reported to play a role in immune regulation. We found that TRIM26 directly binds to SOX2 via TRIM26’s C-terminal PRY-SPRY domain. Unexpectedly, we found that TRIM26 knockdown decreased SOX2 protein stability and conversely increased SOX2 polyubiquitination in primary GSCs. Accordingly, TRIM26 knockdown reduced SOX2 transcriptional activity, self-renewal, and in vivo tumorigenicity in genetically divergent GSC lines. Mechanistically, we discovered TRIM26 stabilizes SOX2 protein by competitively reducing the interaction of SOX2 with WWP2, a bonafide SOX2 E3 ligase in GSCs. Consistent with this hypothesis, WWP2 depletion in the setting of TRIM26 knockdown rescued SOX2 protein levels, self-renewal, and in vivo tumorigenicity in GSCs. Taken together, our data provide further evidence of genetic heterogeneity among and within GBM tumors. However, epigenetic and transcriptional mechanisms controlling the GSC state appear to be a more common feature of GBM, raising the intriguing possibility that disruption of the malignant epigenetic/transcriptional landscape of GSCs represents a unifying therapeutic strategy.
Chair and Committee
Albert H. Kim
Mahlokozera, Tatenda, "Overcoming genetic heterogeneity in glioblastoma by targeting transcriptional dependencies" (2022). Arts & Sciences Electronic Theses and Dissertations. 2652.
Available for download on Friday, April 19, 2024