IDENTIFICATION OF GENETIC DIFFERENCES BETWEEN TUMORS ARISING IN DIFFERENT BRAIN REGIONS. Meredith G. Albin1, Mukesh K. Sharma2, David H. Gutmann2, Biology Department, Washington University, St. Louis, MO1, Neurology Department, Washington University School of Medicine, St. Louis, MO2
The most common brain tumor in adults is the glioma, a tumor composed of glial cells. In children, gliomas represent the second most common brain tumor. In an effort to identify genes involved in the molecular pathogenesis of glioma, studies in our laboratory have focused on the low-grade pilocytic astrocytoma (PA). These tumors are most commonly located in the optic pathway/hypothalamus, cerebellum cortex, and brainstem. Using high-density microarrays, we demonstrated that PAs arising in different brain regions have distinct patterns of gene expression. These initial findings were next validated on an independent set of PA tumors arising supratentorially (e.g., cortex) or in the posterior fossa (e.g., cerebellum). To determine if this differential brain region-specific pattern of gene expression was also observed in another glial cell tumor type, we studied ependymomas, and identified eight genes with patterns of differential gene expression shared with PAs.
Based on our finding that gene expression patterns might distinguish tumors from different brain regions, high-density gene expression profiling was employed to distinguish normal glial cells arising from these different brain regions. We chose to compare normal cells (astrocytes) and tumors arising supratentorially with those arising in the posterior fossa. Using Significance Analysis of Microarrays (SAM) analysis, a set of genes with differential brain region-specific expression in astrocytes was identified. These findings were validated using quantitative real time reverse transcriptase polymerase chain reaction (RT-PCR) on independent cultures of cortical and cerebellar astrocytes from postnatal day 1 mice.
To investigate whether these differential patterns of gene expression identified region-specific cells of origin for both normal glial cells and glioma, we used bioinformatics algorithms to identify genes whose expression could distinguish both tumors (ependymomas and PAs) and normal astrocytes by brain location. Five genes were identified with common patterns of differential expression between astrocytes, ependymomas, and PAs. Additionally, the gene expression pattern of neural stem cells from these different brain regions was analyzed, and we found that this shared pattern of differential gene expression was retained in these progenitor cells. These findings are important because both normal astrocytes and glial cell tumors are thought to arise from neural stem cells.
Our observations and further studies currently ongoing in the laboratory will enhance our understanding of the genetic differences between glial tumors arising in different regions of the brain. Defining differential gene expression signatures retained in glial cell tumors, normal differentiated glial cells, and neural stem cells may be a useful tool for identifying specific progenitor cells. This may enable us to identify specific subpopulations of progenitor cells that confer unique biological properties on both normal astrocytes and glial cell tumors.