Gene expression profiling of grade II oligodendrogliomas and ependymomas

Grade II gliomas are morphologically and clinically heterogeneous tumors for which histopathological typing remains the major tool for clinical classification. To what extent the major histological subtypes (astrocytic, oligodendroglial, oligoastrocytic and ependymal tumors) constitute is largely unresolved. Morphological classification can often be ambiguous and would be facilitated by specific subtype markers. Gene expression profiling is becoming an established method to characterize different types of cancers, and it has been used extensively for glioblastoma analysis and for low-grade glioma to glioblastoma comparison. However, few studies of low-grade oligodendrogliomas and ependymomas have been reported. Using oligonucleotide-based microarray analysis, we compared the transcriptional profiles of grade II oligodendrogliomas and grade II ependymomas (two types of grade II gliomas). An essential initial step toward more objective tumor classification is the establishment of taxonomy for tumors based on their gene expression profiles. We aimed to identify expression profiles that differentiate these two types of grade II gliomas. After comparing the expression signature of a tumor to that of normal tissue, differentially expressed genes were used to perform hierarchal clustering, principal component analysis and prediction analysis. Hierarchal clustering and principal component analysis divided nine different tumor tissue samples and four normal brain tissue samples into categories that corresponded well with clinical pathology analysis, grade II oligodendrogliomas, grade II ependymomas and normal tissues. By using expression data from the most informative gene clusters, we constructed a tumor-clustering model. Class prediction using nearest shrunken centroid analysis withcross-validation identified 53 differentially expressed genes between the two types of grade II gliomas.   Key words: Grade II ependymoma, grade II oligodendrogliomas, microarray, differentially expressed genes, clustering, classification, prediction.