Abstract Aims Glioblastoma Multiforme (GBM) is the most aggressive form of primary brain tumour, with a median survival of 12-14 months after diagnosis. Although GBM has been extensively characterised on the molecular level during the past decades, many targeted therapies have been proved ineffective due, in part, to high heterogeneity of GBM. Thus, novel therapies targeting the altered metabolism which is exhibited by all cancer cells have gained much attention. The therapeutic ketogenic diet (KD) is a high fat, low carbohydrate and adequate protein diet. It has been recognized as a treatment for refractory paediatric epilepsy for decades. Recent studies have shown that a KD reduced tumour growth and potentiated the effects of therapy in some glioma animal models. However, the underlying mechanism(s) is still unclear. Thus, the aim of this study was to understand the mechanism of action behind the KD’s effects in inhibiting tumour growth and potentiating chemotherapy and radiotherapy. Method To unravel the mechanism of action, we analyzed the expression of genes encoding chromatin modifying enzymes in brain tumour samples from mice fed either a KD or standard diet (SD), using the Mouse Epigenetic Chromatin Modification Enzyme PCR Array (Qiagen, Germany). The expression of genes of interest selected from the array were validated by qRT-PCR. Human GBM cell lines and primary cells from GBMs were used to validate the results of the GBM mouse model. Beta-hydroxybutyrate, the main physiological ketone body found in the circulation of patients during KD, was used in in vitro experiments to mimic the in vivo physiological effect of a KD. The effect of protein arginine methyltransferase 8 (PRMT8) overexpression in GBM cells was studied using a lentiviral system. Cell proliferation was measured by Sulforhodamine B assay (Sigma, USA). Spheroid growth and invasion was measured in GBM spheroids cultured in Matrigel matrix (Corning, USA). Results Our results highlighted changes in the expression of a number of key chromatin modifying enzymes in mice fed a KD compared to those fed a SD. PRMT8, a gene highly downregulated in GBM, was upregulated in tumors from mice fed a KD, with corresponding downregulation of its target genes, dihydrofolate reductase (DHFR) and C-X-C chemokine receptor type 4 (CXCR4). Our results also showed that overexpression of PRMT8 in GBM cells reduced cell proliferation and invasiveness. Conclusion PRMT8, DHFR and CXCR4 have been shown to play key roles in tumour growth, invasion, migration and chemo/radio-resistance. Moreover, therapeutic strategies to downregulate these genes have been investigated in the form of methotrexate for DHFR inhibition and small molecule inhibitors of CXCR4. Thus, our results suggest that one mechanism through which the KD exerts its therapeutic effects may be through altering the expression of chromatin modifying enzymes. This provides additional support for the use of a KD as an adjuvant in combination with existing therapeutic approaches.