C.L. van der Wilt

Basis for effective combination cancer chemotherapy with antimetabolites.

Most current chemotherapy regimens for cancer consist of empirically designed combinations, based on efficacy and lack of overlapping toxicity. In the development of combinations, several aspects are often overlooked: (1) possible metabolic and biological interactions between drugs, (2) scheduling, and (3) different pharmacokinetic profiles. Antimetabolites are used widely in chemotherapy combinations for treatment of various leukemias and solid tumors. Ideally, the combination of two or more agents should be more effective than each agent separately (synergism), although additive and even antagonistic combinations may result in a higher therapeutic efficacy in the clinic. The median-drug effect analysis method is one of the most widely used methods for in vitro evaluation of combinations. Several examples of classical effective antimetabolite-(anti)metabolite combinations are discussed, such as that of methotrexate with 6-mercaptopurine or leucovorin in (childhood) leukemia and 5-fluorouracil (5FU) with leucovorin in colon cancer. More recent combinations include treatment of acute-myeloid leukemia with fludarabine and arabinosylcytosine. Other combinations, currently frequently used in the treatment of solid malignancies, include an antimetabolite with a DNA-damaging agent, such as gemcitabine with cisplatin and 5FU with the cisplatin analog oxaliplatin. The combination of 5FU and the topoisomerase inhibitor irinotecan is based on decreased repair of irinotecan-induced DNA damage. These combinations may increase induction of apoptosis. The latter combinations have dramatically changed the treatment of incurable cancers, such as lung and colon cancer, and have demonstrated that rationally designed drug combinations offer new possibilities to treat solid malignancies.

Thymidylate synthase and drug resistance

Thymidylate synthase is an important target for both fluorinated pyrimidines and for new folate analogues. Resistance to 5-fluorouracil (5FU) can be related to insufficient inhibition of thymidylate synthase. The 5FU-nucleotide FdUMP induces inhibition of thymidylate synthase which is enhanced and retained for longer in the presence of increased folate pools, for which leucovorin is a precursor. In a murine model system, 5FU treatment caused a 4-fold induction of thymidylate synthase levels which may have contributed to resistance. Addition of leucovorin to this treatment prevented this induction and increased the antitumour effect 2-3-fold. In the clinical setting, 5FU administration to patients resulted in approximately 50% inhibition of TS after 48 h. The combination with leucovorin resulted in a more pronounced inhibition after 48 h (approximately 70%). A significant relationship was observed with outcome of treatment; when thymidylate synthase levels were high and inhibition was low, no response was observed. A separate study showed that low thymidylate synthase levels appeared to be an independent prognostic factor for adjuvant therapy.

Prolonged retention of high concentrations of 5-fluorouracil in human and murine tumors as compared with plasma

SummaryConcentrations of 5-fluorouracil (5-FU) and its active metabolite 5-fluoro-2′-deoxy-5′-monophosphate (FdUMP) were measured in biopsy specimens of tumor tissue, normal mucosa, metastatic liver nodules, and normal liver tissue obtained from 39 patients and in two murine colon tumors (colon 26 and colon 38) after a single injection of 5FU at a therapeutic dose (500 mg/m2 and 100 mg/kg, respectively). These data were compared with plasma concentrations. Peak plasma concentrations (300–500 μm) of 5FU were comparable in human and murine plasma. The half-life of plasma elimination (during the period from 15 to 120 min) in both mouse and man ranged from 10 to 20 min, whereas at between 2 and 8 h, plasma concentrations varied from 0.1 to 1 μm, the half-life being about 100 min. In both species, 5FU could be measured in plasma at concentrations ranging from 0.01 to 1 μm for several days after 5FU treatment. 5FU concentrations in tissue samples obtained from 14 patients were measured during the time range of 1–6 h, those in samples taken from 7 patients, during the interval of 19–27 h; and those in samples obtained from 18 patients, within the interval of 40–48 h after injection. 5FU tumor concentrations varied between 0.78–21.6, 0.44–6.1, and 0.17–10.8 μmol/kg wet wt., respectively. Some of the 48-h samples were obtained from patients who had received leucovorin plus 5FU; coadministration of leucovorin did not alter 5FU tissue concentrations. At between 4 and 48 h, the tissue concentration/plasma concentration ratio was at least 10. 5FU concentrations in murine tumors were measured for up to 10 days after 5FU administration, with plateau 5FU tumor concentrations being about 50 μmol/kg wet wt. in colon 38 and about 200 μmol/kg wet wt. in colon 26 at 2 h after treatment; after 4 days, values of 0.5 and 4.8 μmol/kg, respectively, were obtained and after 10 days, respective concentrations of 0.1 and 0.07 μmol/kg were detected. The FdUMP concentrations measured in colon 26 and colon 38 tumors were 214 and 46 pmol/g, respectively, at 2 h after 5FU administration, and these values subsequently decreased to about 15 pmol/g in both tumors. In human tumors the initial FdUMP concentration ranged from 10 to 1000 pmol/g; at later time points the level of FdUMP was just above the detection limit of the assay. In liver metastases, high 5FU concentrations seemed to be related to high levels of FdUMP, which was likely of importance for the antitumor effect. The prolonged retention of 5FU should be taken into consideration in the design of biochemical modulation studies.

Molecular downstream events and induction of thymidylate synthase in mutant and wild-type p53 colon cancer cell lines after treatment with 5-fluorouracil and the thymidylate synthase inhibitor raltitrexed

Inhibition of the key enzyme in DNA synthesis, thymidylate synthase (TS), by 5-fluorouracil (5-FU) and the novel antifolate raltitrexed (Tomudex; ZD1694), induces dTTP depletion, resulting in DNA strand breaks, which can initiate pathways leading to an apoptotic mode of cell death. We studied 5-FU- and ZD1694-induced TS inhibition in relation to the expression of p53, p21, Bcl-2 and Bax in six colon carcinoma cell lines, two with a wild-type (wt) p53 (Lovo, LS174T) and four with a mutant (mt) p53 (WiDr, WiDr/F, HT29 and SW948) phenotype. In untreated cells, a reciprocal correlation between p53 and Bcl-2 was found: in cells with a low wt p53, Bcl-2 expression was present; whilst in cells with mt p53, Bcl-2 expression was not detectable. Exposure to 5-FU (50 and 100 microM) and ZD1694 (50 and 100 nM) for 24 and 48 h induced p53 and p21 expression in wt p53 cells, but not in mt p53 cells. TS was induced approximately 2-10-fold in all cell lines. TS induction was highest after ZD1694 exposure in the mt p53 cells HT29 and WiDr/F (6-10-fold). After 5-FU treatment, TS was present both as the free enzyme and in the ternary complex; however, predominantly as the ternary complex between TS, FdUMP and 5,10-methylenetetrahydrofolate. In wt p53 cells, both drugs increased Bax expression up to 5-fold, whereas in mt p53 cells, only a very slight induction was found. In wt p53 cells, Bcl-2 expression hardly changed after drug treatment. These results indicate a p53-independent induction of TS but a regulatory role of wt p53 in the synthesis of Bax in the colon carcinoma cell lines after TS inhibition.

Folate depletion increases sensitivity of solid tumor cell lines to 5-fluorouracil and antifolates

Cancer cell lines in standard cell culture medium or in animal models are surrounded by an environment with relatively high folate (HF) levels, compared with folate levels in human plasma. In the present study we adapted 4 colon cancer (C26‐A, C26‐10, C26‐G and WiDr) and 3 squamous cell carcinoma of the head and neck (HNSCC) cell lines (11B, 14C and 22B) to culture medium with low folate (LF) levels (2.5, 1.0 and 0.5 nM, respectively) and investigated whether folate depletion had an effect on sensitivity to antifolates and which mechanisms were involved. All LF cell lines showed a higher sensitivity to 5‐fluorouracil (5‐FU) alone or in combination with leucovorin (LV) (2–5‐fold), to the thymidylate synthase (TS) inhibitors, AG337 (2–7‐fold), ZD1694 (3–49‐fold), ZD9331 (3–40‐fold), LY231514 (2–21‐fold) or GW1843U89 (4–29‐fold) or to the dihydrofolate reductase (DHFR) inhibitor PT523 (2–50‐fold) compared with their HF variants cultured in standard medium containing up to 8 μM folic acid. LV could only increase sensitivity to 5‐FU in HNSCC cell lines 14C and 14C/F. The differences in sensitivity could partially be explained by a 2–7‐fold increased transport activity of the reduced folate carrier (RFC) in LF cell lines, whereas no significant change in folylpolyglutamate synthetase (FPGS) activity was observed. Furthermore, the protein expression and catalytic activity of the target enzyme TS were up to 7‐fold higher in HF colon cancer cells compared with the LF variants (p < 0.05). Although the TS protein expression in LF HNSCC cells was also lower than in HF variants, the TS catalytic activity and FdUMP binding sites were up to 3‐fold higher (p < 0.05). Thus, changes in TS levels were associated with differences in sensitivity. These results indicate that folate depletion was associated with changes in TS and RFC levels which resulted in an increase in sensitivity to 5‐FU and antifolates. The folate levels in LF medium used in this study are more representative for folate levels in human plasma and therefore these data could be more predictive for the activity of 5‐FU and antifolates in a clinical setting than results obtained from cell lines cultured in HF medium or in animal models. Int. J. Cancer 87:771–778, 2000.

Predictive value of thymidylate synthase and dihydropyrimidine dehydrogenase

before injection, no differences in colony numbers were shown. These results suggest that different AML progenitor cells can be defined in the whole AML population, dependent on the test system used. Moreover, the findings indicate a hierarchy in AML progenitors, including cells with short-term and longterm repopulating abilities, as has already been defined in the normal counterpart. This expanding knowledge on the biological behaviour of AML progenitors ita vitro may be used in future clinical trials to optimise new therapeutic strategies, especially in the possible elimination of AML progenitor cells from the graft.

Effect of WR-2721 on the toxicity and antitumor activity of the combination of carboplatin and 5-fluorouracil

SummaryWe evaluated the effects of WR-2721 on the toxicity and antitumor activity of the combination of 5-fluorouracil (5FU) and carboplatin (CBDCA) in BALB/c and C57Bl/6 mice. On a weekly schedule, i.p. injection of 200 mg/kg WR-2721 at 5 min prior to the administration of this combination enabled us to increase the CBDCA dose from a nontoxic level of 45 mg/kg to a normally toxic dose of 60 mg/kg in non-tumor-bearing BALB/c mice while maintaining the 5FU dose at 100 mg/kg. When WR-2721 was given 30 min before this combination, the CBDCA dose could not be increased to 60 mg/kg without producing drug-related deaths. WR-2721 protected against CBDCA- and 5-FU-induced thrombocytopenia but did not prevent leukopenia or anemia in C57Bl/6 mice. The antitumor activity of the combination against colon 26 tumors in BALB/c mice was increased by pretreatment with WR-2721, which facilitated elevation of the CBDCA dose to 60 mg/kg in combination with 100 mg/kg 5FU. These results reveal better therapeutic efficacy for the combination of 5FU and CBDCA following pretreatment with WR-2721.

Protection by WR-2721 of the toxicity induced by the combination of cisplatin and 5-fluorouracil☆☆☆

We evaluated the effects of WR-2721 and its metabolite WR-1065 on in vitro growth inhibition by 5-fluorouracil (5FU) and cisplatin (CDDP) and the effect of WR-2721 on in vivo toxicity and antitumor effect of 5FU and CDDP. In cell culture both WR-2721 and WR-1065 were not able to reverse growth inhibition caused by either 5FU or CDDP. Administration of WR-2721 i.p. at 525 mg/kg to mice resulted in a severe temperature drop to 27 degrees C; at 200 mg/kg hypothermia was less severe. WR-2721 failed to prevent 5FU toxicity, but the maximum tolerated dose of CDDP in the combination with 5FU (at 100 mg/kg) could be increased from 3 to 7 mg/kg. CDDP at 7 mg/kg enhanced leukopenia caused by 5FU at 100 mg/kg to 20% and thrombocytopenia to 40%; WR-2721 reduced leukopenia and prevented thrombocytopenia induced by the combination. Combination of CDDP, 5FU, and WR-2721 resulted in an enhanced antitumor activity against the murine colon tumor Colon 26 compared to 5FU alone and to 5FU combined with CDDP at their maximum tolerated dose.

Tumor size and origin determine the antitumor activity of cisplatin or 5-fluorouracil and its modulation by leucovorin in murine colon carcinomas

Abstract 5-Fluorouracil (FUra) is one of the few effective agents in the treatment of patients with colorectal cancer. Its effects on the target enzyme thymidylate synthase (TS) can be modulated by leucovorin (LV) or cisplatin (CDDP). Tumor size and differentiation of tumor characteristics can influence therapeutic efficacy. We therefore studied the relationship between tumor size (cutoff point 200 mm3) and the antitumor activity of FUra and its modulation by LV in murine Colon 26 and Colon 38 tumors. The doubling time of tumors measuring >200 mm3 was about 160% longer. The antitumor effect of FUra in these large tumors was decreased and could not be modulated by LV. In addition, three subtypes of Colon 26 (Colon 26-A, Colon 26-B, and Colon 26-10) were identified and characterized for tumor-induced weight loss, TS activity, response to chemotherapy, and histological features. Mice bearing Colon 38 and Colon 26-10 did not lose weight as a result of tumor growth. Colon 26-A caused a weight loss of up to 19%, whereas mice with Colon 26-B tumors remained within 10% of their initial weight and tolerated at least 2.5 times more tumor load than did mice bearing Colon 26-A, which induces cachexia. Among untreated tumors, TS catalytic activity was highest in Colon 26-B (5536 pmol mg protein-1 h-1) and lowest in Colon 38 (799 pmol mg protein-1 h-1); Colon 26-A and Colon 26-10 had intermediate activities (about 2500 pmol mg protein-1 h-1). 5-Fluoro-2′-deoxyuridine monophosphate (FdUMP) binding was comparable in the three Colon 26 subtypes but was lower in Colon 38. The antitumor activity of FUra could be modulated by LV in Colon 38, Colon 26-10, and Colon 26-A but could not in Colon 26-B, with complete responses (CR) being obtained in Colon 26-10 and Colon 38. The latter two were highly sensitive to CDDP, followed by Colon 26-A and Colon 26-B (CRs: 50%, 40%, 25%, and 0, respectively). Furthermore, necrosis was noted in Colon 26-B and Colon 38 but not in Colon 26-A. In conclusion, (1) the antitumor activity of FUra in large tumors is decreased and cannot be modulated by LV and (2) characteristics and sensitivity to chemotherapeutics can vary substantially in closely related tumors of the same origin.

Antitumour activity, toxicity and inhibition of thymidylate synthase of prolonged administration of 5-fluorouracil in mice

Continuous infusions of 5-fluorouracil (5-FU) are increasingly used in the treatment of cancer. Their optimal use, however, has still to be determined since the availability of suitable animal models is limited. We studied continuous infusions in mice using subcutaneously implanted pellets that release 5-FU over a period of 3 weeks. At the maximum tolerated dose (MTD) (based on the systemic toxicity in healthy animals) we assessed the antitumour activity, haematological toxicity, inhibition of thymidylate synthase (TS) in tumours and the concentration of 5-FU in plasma during the 3-week period. We also studied the addition of leucovorin in different schedules. The dose-limiting toxicity was weight loss, and at the MTD of 10 mg of 5-FU released in 21 days per mouse myelosuppression was tolerable (nadir for leucocytes and thrombocytes was approximately 40% of pretreatment levels). In several independent experiments using the 5-FU-resistant Colon 26 tumour, a good antitumour activity was observed during the first part of the infusion, but thereafter the growth of the tumours resumed; the overall effect of continuous infusions was thus comparable to that of bolus injections. Coadministration of leucovorin did not enhance the therapeutic results; depending on the schedule used, it proved ineffective or only increased toxicity. Similar results were obtained with head and neck squamous cell carcinomas and with the 5-FU-sensitive tumour Colon 38. In Colon 26 tumours the TS activity (FdUMP-binding assay) initially decreased to 20-30% of controls and returned to normal after 11 days. In the catalytic TS assay a slight inhibition was observed for the continuous infusion, followed after 11 days by a marked (4-fold) increase in activity. 5-FU plasma levels varied from 0.1 to 1 microM following a circadian rhythm (with a peak at 6 h after light onset), and were maintained during the entire period. Subcutaneously implanted pellets represent a suitable model to study prolonged administration of 5-FU in mice and to evaluate the effect of modulating agents in laboratory animals before transferring data obtained in vitro to the clinic.