Identification of novel human glioblastoma-specific transcripts by serial analysis of gene expression data mining

BACKGROUND: Glioblastoma multiforme (GBM) remains the most common and aggressive primary brain tumor in adults with a poor median survival, and molecular biomarkers for GBM pathogenesis are in need. PURPOSE: The objective of this study is to identify potential novel genes for GBM pathogenesis by gene expression data mining. MATERIALS AND METHODS: Available SAGE libraries of GBM, astrocytoma, and normal brain tissues were collected from the Cancer Genome Anatomy Project (CGAP). Significance analysis for microarray (SAM) and CGAP-SAGE-Genie-DGED were used to identify differentially expressed tags, and specific tags that were differentially expressed only in GBM were further selected. Tags to genes association was performed by CGAP-SAGE-Genie-SAV. Immunohistochemistry was used to investigate distribution and validate expression of the interested gene. RESULTS: Three genes were significantly differentially expressed just in brain. up-regulated expression of STAB1 and down-regulated expression of SH3GL2 and DNM3. Immunohistochemistry assay indicated that STAB1 mainly expressed in vascular endothelial cells and over-expressed in GBM samples compared to normal samples. CONCLUSIONS: Our study shows that data mining of public sources of gene expression is an effective way to identify novel tumor-associated genes, and this work may contribute to the identification of candidate genes for GBM angiogenesis.