Investigating oscillatory gene expression in glioblastoma

Abstract

Oscillatory gene expression can be defined as gene expression at the mRNA and/or protein level that demonstrates peaks and troughs which occur with some periodicity. Previous studies have demonstrated that expression dynamics of oscillatory genes, particular those that oscillate at the ultradian scale of less than 24 hours, play a key role in controlling cell state transitions in neurogenesis and somitogenesis. For example, changes in the dynamics of expression between oscillatory and sustained expression can govern fate decisions between a stem cell and differentiated state, or between a quiescent and actively proliferating state. This underpins the tissue patterning and phenotypic heterogeneity observed in the developing nervous and musculoskeletal system. Whether oscillatory gene expression dynamics might similarly be responsible for cell-state plasticity and heterogeneity in cancer remains unanswered. This thesis lays the foundation for exploring this question in Glioblastoma (GBM), an incurable human brain cancer defined by remarkable inter and intra-tumoural heterogeneity, driven in part by underlying treatment and microenvironment induced cancer stem cell state plasticity. In my thesis, I infer novel, previously unstudied gene oscillators in glioblastoma that might be particularly important in governing entry into and exit from quiescence. This was achieved by applying a genome-wide bioinformatic inference approach to interrogate single-cell (sc) RNA-seq GBM tumour data. 5 tumours from the Neftel et al. (2019) dataset were analysed (GSE131928). OscoNet, a computational algorithm optimised within the Papalopulu lab to infer gene oscillators from static scRNA-seq data was used to examine the neoplastic cell population within each tumour. The median number of inferred gene oscillators per tumour was 838. Genes consistently inferred to oscillate across all 5 tumours were identified, giving rise to a core list of 117 genes. Given that only a handful of oscillators have been studied in the literature to date, the considerably large number of inferred oscillators from my analyses suggests that oscillatory gene expression might be more widespread than previously appreciated. Interestingly, inferred oscillators were enriched for cancer-related pathways whereas non-oscillators were conversely enriched for more generic housekeeping functions, highlighting the important role that oscillatory expression might play in cancer biology. A shortlist of candidate oscillatory genes that could have an important role in the regulation of glioblastoma cancer stem cell (GSC) quiescence were identified, including the transcription factors: NFIB, JUNB, FOS, TCF4, TSC22D1 and SOX2. SOX2 was selected for further in-vitro validation of oscillatory expression. Using the CRISPR/Cas9 system, I established a pipeline for generating fluorescent knock-in fusion reporters in a patient-derived GSC line to track endogenous protein expression over time. By undertaking live single-cell confocal imaging of cells harbouring a SOX2 reporter in proliferative and BMP4 mediated quiescent experimental conditions, I found that SOX2 protein oscillations were detected in both conditions at a periodicity of approximately 17hrs. This finding highlights a previously unknown aspect of SOX2 expression that could shed mechanistic insight into how it regulates its downstream targets, whereby not just the presence or absence of SOX2 protein, but also its pattern of expression over time might serve to dictate diverse cell-fate decisions. Future studies on how oscillations in SOX2 (and other inferred genes) might regulate GSC quiescence could lead to important therapeutic breakthroughs in this cancer of unmet need. Importantly, the methodological approach presented in this thesis can be applied to any other cancer model or biological context.

Date of Award	1 Aug 2024
Original language	English
Awarding Institution	The University of Manchester
Supervisor	David Coope (Supervisor), Nancy Papalopulu (Supervisor) & Elli Marinopoulou (Supervisor)