C.L. van der Wilt

PURPOSE: To determine the time-dependence of fluorouracil (5FU)-induced thymidylate synthase (TS) inhibition in colon cancer patients, the effect of leucovorin (LV), and the relation to response. PATIENTS AND METHODS: A 5FU injection (500 mg/m2) was given to 47 patients with advanced colorectal cancer; tumor biopsy specimens were obtained 1 to 72 hours after laparotomy. Eleven patients received LV (2-hour infusion of 500 mg/m2) with 5FU midinfusion; biopsies were obtained after 45 hours. TS inhibition was evaluated by comparing the number of total and free 5-fluoro-2'-deoxy-uridine-5'- monophosphate (UMP) (FdUMP) binding sites and the total and residual catalytic activity of TS. RESULTS: The total catalytic TS activity varied from 0 to 621 pmol/h/mg protein and the total number of FdUMP binding sites varied from 0 to 976 fmol/mg protein. The residual catalytic TS activity after 2, 23, and 45 hours was 41%, 65%, and 74% of the total catalytic activity; the number of free FdUMP binding sites was 12%, 27%, and 49% of the total number, respectively. LV enhanced TS inhibition after 45 hours; the residual catalytic activity decreased from 74% to 49%, and the number of free FdUMP binding sites from 49% to 24%. Eleven of 19 patients treated with hepatic arterial infusion of 5FU had a partial response (PR). In the nonresponding patients, total TS activity was significantly higher (P < .05) than in responding patients. A high TS activity with a poor inhibition correlated with no response. CONCLUSION: Residual and total TS activity are predictive for response to 5FU. The findings may be applicable for treatment of patients with advanced disease and TS should be evaluated as a prognostic factor in adjuvant chemotherapy studies.

Deoxycytidine kinase (dCK) is essential for the phosphorylation of gemcitabine (2',2'-difluorodeoxycytidine), a deoxycytidine analogue active against various solid tumors. Cytidine deaminase (CDA) catalyzes the degradation of gemcitabine. We determined whether dCK and/or CDA levels would predict response to gemcitabine. Activities of dCK and CDA were measured in a panel of eight gemcitabine-sensitive and -resistant tumors of a different origin (pancreas, lung, colon, ovary, and head and neck) grown as s.c. tumors in mice. Sensitivity to gemcitabine was expressed as treated versus control (tumor volume treated mice/control mice). Gemcitabine was given on days 0, 3, 6, and 9 (q3dx4) at its maximum tolerated dose. In addition, we measured the mRNA expression and protein levels of dCK in seven human tumor xenografts. dCK activity (mean +/- SE) ranged from 3.3+/-0.3 to 18.4+/-1.2 nmol/h/mg protein. Sensitivity to gemcitabine, expressed as treated versus control, ranged from 0.98 to 0.02, and the activity of CDA varied from 2+/-2 to 411+/-4 nmol/h/mg protein. In contrast to CDA, dCK activity was clearly related to gemcitabine sensitivity (p = -0.93; P < 0.001). This indicates that dCK might be an important prognostic marker for gemcitabine sensitivity. Protein levels were significantly related to both dCK activity (r = 0.96; P < 0.001) and gemcitabine sensitivity (rho = -0.96; P < 0.001). dCK expression as determined by competitive template reverse transcriptase PCR was significantly related with the dCK activity (r = 0.88; P = 0.025) and protein levels (p = 0.80; P = 0.052) but not with gemcitabine sensitivity, suggesting a post-translational regulation of dCK. In conclusion, the clear correlation between dCK levels and gemcitabine sensitivity in various murine tumors and human tumor xenografts may be a prognostic parameter when considering gemcitabine therapy.