LAUSR

Abstract Ependymoma (EPN) is the second most common malignant paediatric brain tumour, which despite extensive genomic sequencing, no novel therapeutic options have been discovered. Multi-omics are anticipated to reveal dysregulated pathways that may be predictive of patient-specific biomarkers. Given the close association between gene expression, active biochemical signaling and metabolism, there is an unmet scientific challenge to determine whether EPN gene expression correlates with aberrant metabolic pathways, thus presenting therapeutic vulnerabilities. We first compared two distinct subgroups of EPN, PF-A and ST-RELA, identifying 115 metabolites and 1580 upregulated genes between the two subgroups, therefore validating previously reported genetic clustering of these two subtypes. We next integrated transcriptomics and metabolomics, comparing 28 intra-tumour tissue regions from eight primary PF-A EPN patients. Polar metabolites and RNA were simultaneously extracted from the same population of cells. RNAseq identified dysregulated genes and liquid chromatography-mass spectrometry (LC-MS) detected 98 significantly altered metabolites between 18 multi-regions, the majority mapping onto the arginine and proline pathways. Integration of genes and metabolites using pathway-based network analysis revealed 124 aberrant gene-metabolite interactions between intra-tumour regions, with large numbers occurring in the glucogenesis and glycine metabolic pathways in 6/8 patients. These may represent ubiquitous and therapeutically relevant metabolic pathways critical for EPN survival. Additionally, patients presented at least one unique intra-tumour genomic-metabolomic interaction, applicable for patient-tailored therapy. This is the first exploration of EPN multi-omic integration and intra-tumour heterogeneity. Selected drug targets predicated on aberrant gene-metabolite networks will be validated in multi-region patient-derived cell lines and orthotopic models using repurposed therapeutics.