Résumé
Prostate cancer is one of the leading causes of death from cancer in men. The prognosis of patients is defined according to staging, PSA levels and Gleason score, which differentiates low grade (Gleason < 4) and high grade (Gleason ≥ 4) cancers. High-grade tumors are associated with high metastatic potential and poor prognosis. They are treated with hormone therapy, to which microtubule poisons are added when they become resistant to the anti-androgen. Despite many clinical trials with other chemotherapeutic agents, response rates remain low. Moreover, none of these trials took into account the tumor grade. We therefore envisaged a new rational in silico approach to screen for drug candidates that could be used as an alternative to docetaxel, based on an expression signature of 86 genes that could distinguish low-grade and high-grade tumors, with a reliability of 81% (PNAS 2006 103:10991-6). We explored the NCI databases, which allow access to both gene expression profiles of 60 human tumor cell lines and their in vitro sensitivity to thousands of anticancer drugs and extracted the expression profiles of the 86 genes’ signature. We calculated for each gene the Pearson coefficients of correlation (r) between their expression level in the 60 cell lines and cell sensitivity to 152 core anticancer compounds. We found that the expression of 11 genes was associated with sensitivity to oxaliplatin. They include DPM1, PCCB, ATP5G3, and SHMT2 genes involved in metabolism; RHOT2, CD59 and JUN genes involved in signal transduction, the CDKN2C cyclin-dependent kinase inhibitor gene, RPL13 and EIF4A1 genes involved in translation, and the unknown FLJ35093 gene. This signature seems specific to DACH platinum derivatives since no correlation was found with the sensitivity to cisplatin. Functional validation of this signature was performed in vitro using the prostate cancer cell lines DU145 and LNCaP and the benign prostate hyperplasia (BPH) cells as “normal” cells. We measured the effect of siRNA-mediated downregulation of each of the 11 genes on the sensitivity to oxaliplatin or cisplatin by clonogenic or MTT assays. Preliminary data confirmed our in silico results for 7 out of the 11 genes, for which donwregulation induced a significant change in IC50 for oxaliplatin (but not cisplatin) in DU145 and/or LNCaP cells. For EIF4A1 we found a correlation opposite to that obtained from in silico results. Also, inhibition of CDKN2C could affect the sensitivity to cisplatin in LNCaP cells. Conversely, no effect of each siRNA tested could be observed in BPH cells, suggesting that functional validation should be tested in a tumoral background. Our new rational approach allowed the identification of oxaliplatin as an alternative therapy for high grade prostate cancers. It also identified a potential gene expression signature that could be used to predict tumor response to oxaliplatin (and potentially DACH platinum derivatives) in the clinic.
Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 5064. doi:10.1158/1538-7445.AM2011-5064