GENO-02INTEGRATED ANALYSIS OF MOLECULAR SUBTYPES IN GLIOBLASTOMA MULTIFORME

Abstract

BACKGROUND: The Cancer Genome Atlas (TCGA) project previously identified genetic changes in the DNA sequence and copy number, DNA methylation, and gene expression of glioblastoma multiforme (GBM) tumors, leading to an enthused effort to compile and characterize changes in the genome, transcriptome, and proteome of human GBM tumors, so that aberrantly functioning molecular pathways and tumor subtypes could be identified. METHODS: Agilent GeneSpring and Mass Profiler Professional (MPP) software was used to perform integrated multi-omics analysis of mRNA expression, miRNA expression, protein expression, and copy number aberration associated with GBM. Additionally, analysis was performed on an independent discovery proteomics study of GBM and control (brain fragments from epilepsy surgery) patients measured using a label-free proteomics approach. RESULTS: Correlation analysis performed between mRNA and miRNA datasets from TCGA identified potential candidate miRNAs regulating the known Proneural signature genes OLIG2, NKX2-2, and SOX11. It also highlighted the molecular events leading to suppression of p53 in Proneural tumor signaling. Sample-sample correlation and PCA analysis on patient specimens resulted in clear separation between control and tumor samples. In addition, the tumor samples were segregated into two to three subgroups, suggesting that different TCGA subtypes may be present. Indeed, combined analysis of mRNA expression data from TCGA study and proteomics data from the label-free study identified a core subset of 54 GBM subtype signatures, clearly differentiating the known subtypes in the larger TCGA cohort. CONCLUSION: Our observation suggests that GBM tumor heterogeneity may be best characterized by both mRNA and protein abundance of the signature genes, opening a possibility for follow-up studies with potential diagnostic and clinical significance.