The suitability of glioblastoma cell lines as models for primary glioblastoma cell metabolism
Abstract Background In contrast to most non-malignant cells, cells comprising Glioblastoma multiforme (GBM), a deadly brain tumour with extremely poor prognosis, preferentially utilise glycolysis over oxidative phosphorylation for metabolism in a phenomenon known as the ‘Warburg effect’. As effective treatments for GBM are severely lacking, research into therapeutics targeting the disease’s highly glycolytic state offer a promising avenue to improve patient survival. These studies often employ GBM cell lines for in vitro studies which translate poorly to the in vivo patient context. Methods The metabolic traits of the seven most commonly used GBM cell lines were assessed using a Seahorse Bioscience Metabolic Flux Analyser and compared to primary GBM cells and primary healthy mixed neural cells from the same patients. Results In support of the glycolytic nature of the patient-derived GBM cell lines, basal mitochondrial rate (p = 0.043) and ATP-linked respiration (p < 0.001) were significantly lower than primary adjacent normal cells from the same patient and reserve capacity (p = 0.037) and Krebs Cycle capacity (p = 0.002) were significantly higher for 12 patients. While no cell line was found to accurately replicate all metabolic attributes of primary GBM cells, specific parameters could be modelled by specific lines. Conclusions U251MG, U373MG and D54 lines are recommended for researching mitochondrial metabolism, and the D645 line for researching ATP-linked respiration. The T98G cell line recapitulated glycolysis-related metabolic parameters of the primary GBM cells and is recommended for research relating to glycolysis. These findings can guide preclinical research into the development of novel therapeutics targeting metabolic pathways in GBM.