Transcriptional Analysis Reveals Metabolic and Cellular Dysfunction of Tumour Infiltrating NK Cells in Glioblastoma Patients

Abstract

Glioblastoma (GBM) is the most common primary brain tumour with a poor 5-year patient survival rate of less than 7%. The tumour microenvironment (TME) of GBM is characterised by low infiltration of immune cells and immunosuppression. Natural Killer or NK cells are potent anti-cancer immune cells which classically kill cancer cells and thus, they represent a potential source of immunotherapy for GBM. However, they poorly infiltrate into GBM tumour sites. NK cells are classified into cytokine producing CD56bright NK cells and cytotoxic CD56dim NK cells, with both subsets becoming dysfunctional in tumour. One potential source of immunosuppression is TGF-β which is abundant in GBM and known to inhibit NK cell function. However, further research is required to characterise its dysfunctional effects and influence on the metabolic and physiological state of NK cells in GBM. Single cell RNA sequencing (scRNA-seq) advancements have revealed dysfunctional transcription patterns of NK cells in tumour. This work analysed a scRNA-Seq NK cell dataset published by Netskar et al., 2024, to investigate the differential expression patterns of NK cells from GBM patient tumours (GBM-NK cells) with reference NK cells from peripheral blood of healthy donors (HD-PBMC-NK). Global unbiased analysis showed GBM-NK cells had distinct transcriptional patterns from HD-PBMC-NK cells and an enrichment in CD56bright NK cells. Interestingly, there was an apparent overall higher gene expression level in NK cells from tumour compared to blood. Cellular stress and apoptosis related pathways were upregulated while IFN-γ signalling and other cytotoxic pathways related to CD56dim NK cell function were downregulated in GBM-NK cells. We then employed a hypothesis driven approach and selected candidate gene sets for closer analysis on a gene level, revealing amongst others an upregulation of oxidative stress, cellular response to ROS and cytochrome c associated apoptosis in NK cells in the tumour. The complexity of the TME was indicated by contradictory pathway enrichment such as increased signature for both mitochondrial fission and fusion or an increased expression of oxidative phosphorylation signature while come ETC complexes were significantly downregulated. Examination of TGF-β also identified some contradictory findings; there was evidence of higher TGF-β signalling in tumour NK cells but these also had elevated IFNG and MYC gene expression that would be predicted to be lower in the presence of TGF-β signalling. Overall, this thesis highlights that GBM-NK have a stressed, metabolically dysregulated and impaired mitochondrial phenotypes, at least at the RNA level. Cross-referencing TGF-β analysis in combination with other inhibitory signalling pathways known to exist in GBM such EGFR, adenosine signalling etc., could help illustrate how GBM-NK develop an altered transcriptional expression profile. Combined with functional validation, these data will help inform the best design of NK cell immunotherapy for GBM tumours.