Abstract AIMS Infiltrative margin GBM, as a proxy for residual disease post-surgery, can be identified by 5-aminolevulininc acid (5ALA) to facilitate the discovery of uniquely expressed or upregulated molecular pathways. Complementary bioinformatics approaches offer reproducible ways to enable interoperability of heterogenous data sources towards identification of therapeutic targets associated with GBM infiltration. METHOD RNA-seq primary tumour data was first categorised as: 5ALA+ (infiltrative margin) vs tumour core and 5ALA+ vs 5ALA- (reactive brain). Two levels of filtration (genes with adjusted p-value <0.05; degree of fold change with log2>1.5) were applied to genes, creating two separate gene lists. Cross-referencing gene lists identified transcripts upregulated in both 5ALA+ vs core and 5ALA+ vs 5ALA-. This candidate gene list was inputted into STRING to generate an uncurated protein-protein interaction (PPI) network, which returned a significant PPI enrichment p-value to confirm network connectivity, and next inputted into Metascape to elucidate biological processes associated with tumour infiltration. RESULTS Of 141 upregulated 5ALA+ genes TCP10L3, HELT, MGAM, ALOX5 and FOXI1 exhibited the greatest fold change. The most prevalent 5ALA+ biological processes were primitive streak formation, disaccharide metabolic process and sodium ion transmembrane transport. LOC157273, LINC00200, LINC00656, LOC285626 and LINC01568 were identified as long non-coding RNAs (lncRNA) most abundant in the infiltrative margin of GBM relative to tumour core and reactive brain. Regulatory associations between candidate 5ALA+ lncRNA and genes, and in silico drug repurposing via molecular docking, are being investigated. CONCLUSION The plasticity of GBM warrants analysis of the GBM infiltrative margin via computational toolkits to enhance understanding of upregulated molecular pathways, functionally associated with an invasive phenotype.
Read full abstract