Preview

PDF, English Download (53MB) | Terms of use

Abstract

Glioblastoma (GB), the most common brain malignancy in adults, remains incurable despite being extensively characterized genetically and epigenetically. Current multimodal treatment regimens only marginally extend patients’ survival, and patients almost invariably succumb to the disease rapidly due to tumor recurrence, underscoring the unmet clinical needs in GB treatment. Central to our understanding of GB biology is the brain tumor stem cell hypothesis, which postulates the existence of a population of self-renewing tumor cells. These cells co-opt transcriptional circuitry critical for maintaining the neural stem cell (NSC) state, thereby maintaining plasticity, tumor heterogeneity and potentiating aggressive phenotypes.

Our previous work identified the oligodendroglial transcription factor SOX10 as a master regulator of a subtype of GB. In this thesis, I strive to understand the role of SOX10 in mediating tumor cell phenotypic plasticity and its potential clinical implications in GB. The analysis of public datasets suggested that SOX10-high and SOX10-low human samples occupy different cellular states on the developmental spectrum, indicating an inverse correlation between SOX10 expression and NSC-related cell states in GB. Using a mouse syngraft model, I demonstrated that the knockdown (KD) of Sox10 in tumor cells leads to increased tumorigenicity and aggressiveness, corresponding to the phenotypic transitions recently published in mouse GB models derived from different neural progenitor cells. Single-cell transcriptomic profiling uncovered the cell state plasticity of Sox10-KD tumors, wherein a population of quiescent founder cells appears to drive the developmental-like phenotype of KD tumor cells, mirroring the spectrum of normal NSC development. Further in vitro investigations showed that the downregulation of Sox10 leads to the emergence of slow-cycling stem-like cells within both mouse and human glioma stem-like cells, exemplified by an increase in Notch pathway activity, limited proliferative capacity upon low growth factor culture conditions, and increased resilience to differentiation cues. Furthermore, I showed that temozolomide, the chemotherapeutic agent used in first-line clinical therapy, downregulated SOX10 expression in human brain tumor stem cells. Upon release from therapeutic pressure, recovered cells retained lower expression of SOX10 and maintained a pool of quiescent stem-like tumor cells, suggesting that treatment pressure might similarly induce a latent pool of quiescent founder tumor cells, which potentially fuel tumor regrowth. Finally, I showed that the quiescent state observed in the Sox10-KD model depends on the activity of the Notch pathway. Its pharmacological inhibition drives KD cell transition to a less quiescent, activated state, increasing their vulnerability to anti-proliferative treatment induced by Fimepinostat, a dual inhibitor of the PI3K and HDAC pathways.

In summary, my doctoral study unraveled a transition to quiescent cell states mediated by the reduction of SOX10 expression in brain tumor stem cells. Using SOX10-KD cells as a model, I provided data indicating the feasibility of combination therapies depleting the quiescence founder state and simultaneously targeting resulting more proliferative cell states. My study adds to a growing body of evidence on the plastic differentiation hierarchy of brain tumor stem cells, the importance of quiescent stem cells in mediating phenotypic disparity, and the feasibility of a “state-inducing” therapeutic approach in glioblastoma.