Identification of candidate tumor related genes using comprehensive gene expression analysis of neuroblastic tumors.

Background: Neuroblastic tumors (NBTs) represent a heterogeneous and relatively well defined spectrum of neoplastic diseases. We performed gene expression analysis to molecularly characterize the distinct clinicobiological subtypes of NBTs. Methods: Gene expression analysis of 106 NBTs (10 ganglioneuromas, 10 stage 4s, 29 loco-regional (stages 1, 2, 3), and 57 stage 4) and 12 neuroblastoma cell lines was performed using Affymetrix Genechip Human Genome U95 Set Arrays. Differential expression between predefined groups of biological and clinical relevance was determined by differences >3 standard deviations between the means for groups, step-down permutation and false discovery rate methods. Results: Gene expression analysis of clinically defined groups including stage 4 infants versus children, metastatic versus non metastatic, and stroma poor loco-regional versus stage 4, revealed that many differentially expressed genes mapped to chromosomal regions with well described recurrent abnormalities in NBTs (chromosomes 1, 11, 17, 19 and X). Pairwise comparison analysis of biologically defined groups, including triploid versus diploid/tetraploid NBTs, identified significantly discriminating genes that map to the same chromosomal regions. Gene expression analysis of MYCN-amplified versus MYCN non amplified NBTs, confirmed a functional overrepresentation for genes involved in protein biosynthesis. Conclusions: Gene expression profile analysis of clinically and biologically relevant subgroups of NBTs revealed differential expression of genes at specific chromosomal regions known to have strong association between genetic alterations and NBT clinical features. The identification of altered expression helps to define critical chromosomal regions and identify tumor related candidate genes within each region. No significant financial relationships to disclose.