Abstract

Ependymoma is the second most common paediatric malignant brain tumour. Overall survival remains relatively poor, and at relapse is dismal. One potential strategy to improve outcome is exploiting the metabolic differences between normal and tumour cells. Under normal physiological conditions, brain cells metabolise glucose for energy. If ‘starved’ of glucose, ketone bodies are metabolised. Mitochondrial defects in brain tumour cells obviate this metabolic flexibility resulting in a dependence on glycolytic energy (Warburg effect). Thus, a high fat, low carbohydrate ketogenic diet (KD) may control tumour growth. We evaluated the expression of ketolytic enzymes; succinyl-CoA:3-oxoacid CoA transferase 1 (OXCT1) and d-b-hydroxybutyrate dehydrogenase 1 (BDH1), both involved in ketone body utilisation and the glycolytic enzyme; pyruvate kinase M2 (PKM2), essential for aerobic glycolysis, in 75 paediatric ependymomas by immunohistochemistry. Expression was assessed as low (<20%) or high (>20%). 84% showed low expression of OXCT1% and 91% of BDH1. Low expression of either results in an inability to metabolise ketones. 82% showed high expression of PKM2, increasing the Warburg effect and when combined with low expression of BDH1 (69%) or OXCT1 (83%), the optimal immunohistochemistry profile, arrests ketone metabolism. Children aged <36 months were 5 times less likely to show high expression of OXCT1 compared with 36 months (p=0.0321) and 7 times more likely to show high expression of PKM2 (p=0.0097). The KD therefore offers a potential addition to current therapies including at relapse. However, the KD would need to be carefully evaluated prior to acceptance as a useful additional therapy.