Capecitabine, is an oral fluoropyrimidine carbamate converted sequentially by carboxylesterases (CES) to DFCR, by cytidine deaminase (CDD) to DFUR, and by thymidine phosphorylase (TP) to form 5FU. 5FU is inactivated by dihydropyrimidine dehydrogenase (DPD) to dihydro-5FU or interacts with its target thymidylate synthase (TS) to exert its cytotoxic effect. Metabolism differs between normal tissue and tumors, with 5FU being preferentially formed in the tumor due to higher expression of TP. In this study, the metabolism of capecitabine was investigated in both mouse liver and human xenografts to determine whether a particular step was important for tumor response. The gene expression in mouse liver and fresh human hepatocytes was also compared to assess the relevance of preclinical models for optimisation of capecitabine administration. Nude mice bearing HCT-116 human colon cancer xenografts were treated with 2.1 mmol/kg/d of capecitabine. Liver and tumor samples were collected on days 0 and 7 of treatment. The concentrations of capecitabine and its metabolites were determined in both tissues by HPLC/MS. The expression of the genes of interest (GoI), CES, CDD, TP, TS and DPD was determined by qRT-PCR in tumor xenografts, mouse liver and fresh human hepatocytes and results normalised with reference genes. On day 7, xenograft tumors contained high concentrations of 5FU (2385 nM) while 5FU was undetectable in liver tissue. Similar levels of DFUR were detected in liver and xenografts (42 versus 100 nM, respectively). Lower levels of DFCR were detected in tumor xenografts as compared to mouse liver (82 and 609 nM, respectively). CES1 was highly expressed in mouse liver and absent in HCT-116 xenografts. CDD was higher in tumor than in mouse liver (GoI/Ref = 1.4 and 0.02, respectively) while similar levels of TP were expressed in both tissues (GoI/Ref = 0.49 and 0.6). DPD expression was 10-fold lower in tumor compared to mouse liver (GoI/Ref = 0.12 vs 1.24). Taken together, mouse liver seems to preferentially form DFCR which is taken up by tumor cells and rapidly converted to DFUR and 5FU. Whilst the liver metabolises 5FU completely, due to the high expression of DPD, detectable levels of 5FU are maintained in tumor tissue due to the reduced inactivation by DPD. In fresh human hepatocytes, CDD was higher than in mouse liver (GoI/Ref = 3.8 and 0.02, respectively) but similar to HCT-116 xenografts while TP was 3-fold lower. Human liver may therefore form more DFUR than mouse liver. DPD levels were similar in hepatocytes and mouse liver but significantly higher than in xenografts. DPD as well as TP may potentially contribute in maintaining significant concentrations of 5FU in tumor tissue.

[Proc Amer Assoc Cancer Res, Volume 47, 2006]