Gene expression analysis to identify genes whose expression profiles correlate with sensitivity of human cancer xenografts to anticancer drugs.

3

Purpose: The establishment of a diagnostic method for predicting the efficacy of against anticancer drugs for individual patients is important. In this study, we performed genome-wide gene expression analysis to identify genes whose expression profiles correlated with the sensitivity of 30 human cancer xenografts to 4 fluorouracil (FU) anticancer drugs (UFT, TS-1, 5’-DFUR, capecitabine), CDDP, CPT-11, and Taxol. Methods: Efficacy of anticancer drugs were assayed using human tumor xenografts in nude mice. RNA isolated from each of these xenografts was used to prepare biotinylated cRNA probes and carry out real time RT-PCR. These probes were hybridized to Affymetrix Human Focus oligonucleotide arrays. Expression values for each gene on each chip were then imported into GeneSpring software for further analysis. A multivariate analysis was performed by using JMP and SAS software. Real time RT-PCR was carried out using the ABI Prism®7700 Sequence Detection System. Results: By the gene expression profiling correlated significantly with sensitivity to FU derivatives, we identified common genes associated with cell proliferation (MYB etc) and cell adhesion (CLDN3 etc). On the other hand, many genes that showed specific correlations with individual drugs were also seen. Some candidate genes associated with drug sensitivity were validated by real time RT-PCR. In the candidate genes, DPD (dihydropyrimidine dehydrogenase) mRNA expression profiles of the tumors showed a significant negative correlation with chemosensitivity to the FU derivatives (UFT, 5’-DFUR, and capecitabine) except for TS-1. The variable fold-difference in DPD gene expression in human tumor xenografts was about 5-fold as against that in TS (thymidylate synthase) gene expression. Conclusion: Our results demonstrated that the expression level of DPD may affect the antitumor effects of most FU derivatives but not of TS-1, which composed of tegafur (a prodrug of FU), 5-chloro-2,4-dihydroxypyridine that is about 200-fold more potent DPD inhibitor than uracil, and potassium oxonate that reduces gastrointestinal toxicity. Therefore clinical application of TS-1 might be an effective strategy for treatment of high DPD expressed tumors. In this study we identified several candidate chemosensitivity-related genes for each drug, and further study will be needed to verify our finding. These results should provide useful information for applied research to predict tumor responsiveness to anticancer drugs and may lead to tailor-made chemotherapy in the future.