New prognostic markers revealed by evaluation of genes correlated with clinical parameters in Wilms tumors

Current treatment protocols for Wilms tumor achieve 90% cure rates, but relapse risk and side effects from therapy remain challenging. Over the last decade, numerous markers have been proposed for classification and/or prediction of outcome. However, cohort sizes were quite variable and often small. We now provide a large-scale reassessment by real-time RT-PCR of 40 markers in 102 Wilms tumors followed by validation of potentially relevant markers in an independent set of 74 tumors. In the first data set, individual comparison with clinical data combined with adjustment for multiple testing and multivariate analysis revealed potentially relevant alteration of CA9, DKK1, EGR1, HEY2, MYC, MYCN, TERT TOP2A, TRIM22, and VEGF expression in association with CTNNB1 mutation status, histological risk, response to chemotherapy, metastasis, relapse, or mortality. To further validate these data, potentially relevant genes for specific outcomes were reanalyzed in a second, independent tumor set. Here, univariate analysis confirmed the association of HEY2 with high-risk tumors and of TRIM22 with mortality. Even where significance levels could not be reached, the direction and extent of differential expression were generally reproducible. Multivariate analysis verified a weak correlation of TONA expression with metastasis and of TRIM22 with fatal outcome. Although we could corroborate only some of the previously reported associations of expression changes with clinical parameters, our results indicate that real-time RT-PCR analysis can facilitate further classification of Wilms tumor and prediction of outcome to adjust treatment accordingly. This article contains Supplementary Material available at http://www.interscience.wiley. com/jpages/1045-2257/suppmat. (C) 2008 Wiley-Liss, Inc.