Integration of Genomic Alterations and Expression Profiling in Glioblastoma Multiforme.
Abstract
Purpose: Glioblastomas (GBM) are highly malignant and heterogeneous gliomas with very poor prognosis. The biological and molecular characterization of these tumors is still challenging and impacts their therapeutic management. Previous genomic surveys have revealed the highly rearranged nature of GBM genome and transcriptome. However, the impacts of tumor DNA aberrations on gene expression remain unclear. Methods: We investigated copy number alterations (CNA) and gene expression to identify causal genetic events in GBM. High-resolution maps of somatic chromosomal alterations were obtained for 20 GBM. Gene expression profiling was carried out on the same tumor samples, and compared to those obtained on nonneoplastic brain samples. Concordance between CNA and gene expression was identified by two complementary approaches (correlated or targeted probes). The resulting GBM signature was validated with an independent microarray data set of 81 GBM and 23 normal brains. Results: Loci targeted for high-priority minimal common regions (MCR) of recurrent CNA were defined and combined with gene expression profiles performed on the same tumor samples. Genes with concordant changes in CNA and expression levels were defined as over/underexpressed genes located in amplified/deleted regions, or as MCR genes with expression correlated to the corresponding genomic state. After validation, we found that the expression of 318 genes was significantly affected by CNA. Associated enriched GO process annotations were related to cell cycle disorder, cellular adhesion and angiogenesis. The gene signature included well-known GBM genes such as EGFR, PDGFA, and p16INK4 but also novel candidate genes. Two tumor suppressor genes PCDH9 and STARD13, involved in tumor invasiveness and resistance to etoposide, were validated by qPCR in an independent set of 57 glioblastoma. Conclusion: This study shows the power of combining genomic alterations and gene expression to identify robust transcriptome signature and putative tumor biomarkers in GBM.