A comprehensive view of the structure and expression of the ependymoma genome at presentation and relapse

Abstract

2073 Background: Although pediatric and adult ependymomas are associated with significant mortality and morbidity, little is known about the biology of these tumors. To identify underlying genetic alterations and cellular pathways that drive this disease, we conducted a genomic study of 200 adult and pediatric ependymomas. Methods: Using 500k single nucleotide polymorphism arrays, U133 Affymetrix gene and microRNA (miRNA) expression microarrays, and appropriate bioinformatics, we characterized 56 supratentorial (ST), 104 posterior fossa (PF), and 40 spinal (SP) ependymomas. Real-Time polymerase chain reaction and fluorescence in situ hybridization validated observed genetic events. Results: Gene expression profiles segregated tumors by site and identified disease subgroups within each anatomical region (4 ST, 4 PF, 1 SP). miRNA expression profiles identified these same subgroups, indicating that they are biologically distinct. Subgroup-specific gene expression profiles were dictated partly by developmental regulatory genes and partly by large chromosomal gains (eg. 1q, 5p, 16p) and losses (eg. 9p, 22q). Integrated genetic and expression mapping revealed key candidate tumor suppressor (TSG) and onco- genes, likely drivers of these large alterations. While large chromosomal changes occurred more frequently in SP tumors (p < 0.0001), ST tumors averaged more focal changes (n = 13.2) than PF (n = 6.2) or SP tumors (n = 3.0) (p < 0.0001). A total of 29 and 33 non-random focal amplifications and deletions, respectively, encompassing 402 known genes and miRNA clusters, were validated, of which 80 displayed copy number driven expression. These genetic alterations targeted specific cellular functions (e.g., cell adhesion, cell-cycle, neuronal development) and pathways (e.g., NOTCH, EPHRIN, TP53). Our cohort also included five sample sets consisting of primary tumor and at least two corresponding relapses. Genomic analysis of these tumors identified large chromosomal alterations as well as focal gains and losses associated with disease relapse. Conclusions: We present a highly comprehensive view of the ependymoma genome, including 80 previously unrecognized candidate TSG and oncogenes that may afford diagnostic and therapeutic targets. No significant financial relationships to disclose.