PO32GENE CO-EXPRESSION NETWORK ANALYSIS REVEALS UNIQUE MOLECULAR CHARACTERISTICS GOVERNING THE MARGINS OF GLIOBLASTOMA

Abstract

INTRODUCTION: Glioblastoma (GBM) is the most common malignant brain tumor in adults with a poor outcome due to recurrence from residual disease. Tumour cell invasion into brain parenchyma is likely a major contributor to multimodal treatment failure. Understanding the mechanisms that govern invasion will permit the identification of molecular targets that are functionally important for recurrence disease progression. METHOD: We analysed RNA-sequencing data consisting of tissues within the contrast-enhancing (CE) region, core (C) and non-enhancing (NE) margins of tumors from 27 GBM, and 17 normal brain (NB) samples. To identify NE-specific alterations in transcriptome organization, we conducted a genome-wide gene co-expression network analysis. RESULTS: Twenty modules of co-expressed genes were identified and revealed that expression patterns of NE are more similar to NB than CE for the majority of the modules. Two discrete modules (M1 and M2) were significantly associated with NE and are highly enriched with oligodendocyte (M1: 13 genes, FDR < 3.11 × 10–09; M2: 36 genes, FDR < 2.63 × 10–32) and white matter-associated (M1: 8 genes, FDR < 1.16 × 10–09; M2: 10 genes, FDR < 1.01 × 10–05) genes. M1 was enriched for integrin mediated signalling (DOCK1, FUT8, GAB2), vacuole organization (MAN2A1, PSEN1, TFEB), and platelet activation (RHOG, TGFB3). In contrast, M2 was enriched for lipid biosynthetic processes (GPAT2, GPD1, PIP4K2A, UGT8), ion transport (BEST1, SGK2, SGK3, TMC6), and cell shape regulation (ERMN, FGD4, S100B, SEMA4D). CONCLUSION: Unique molecular GBM abnormalities i implicate transcriptional dysregulation of distinct biological processes that are associated with invasion.