Engineered resistance to camptothecin and antifolates by retroviral coexpression of tyrosyl DNA phosphodiesterase-I and thymidylate synthase.

PURPOSE: Gene transfer of cDNA sequences that confer drug resistance can be used (1) to protect hematopoietic cells against the toxic effects of chemotherapy, (2) for in vivo enrichment of genetically engineered cells and (3) to protect cytotoxic T lymphocytes in drug-resistant immunotherapy approaches for the treatment of cancer. We have previously developed strategies to confer resistance to agents targeting thymidylate synthase (TS) and have now expanded our drug resistance strategies to include retroviral expression of tyrosyl-DNA phosphodiesterase (TDP-I), an enzyme recently implicated in the repair of topoisomerase-I (Top-I)/DNA lesions induced by camptothecin (CPT). The combination of TS and Top-I inhibition has been shown to be an effective treatment for several types of cancer. MATERIALS AND METHODS: Retroviral vectors were generated that individually encoded TS and TDP-I or that coexpressed both enzymes. Murine fibroblast and Chinese hamster lung transfectants were generated with the vectors and resistance to TS- and Top-I-directed inhibitors was tested. Murine bone marrow progenitor cells were also transduced using recombinant retroviruses encoding TS and TDP-I and the degree of drug resistance conferred to gene-modified cells was tested. RESULTS: Enforced expression of TDP-I increased TDP-I activity in gene-modified cells and conferred up to threefold resistance to CPT. The degree of resistance was dependent on the duration of drug treatment. Simultaneous expression of the TS gene encoding E. coli TS optimized for expression in mammalian cells (optecTS) and TDP-I conferred extremely high-level resistance to concurrent treatment with the TS-inhibitor BW1843U89 and CPT. Furthermore, by direct analysis of DNA fragmentation using the comet assay, substantial protection was conferred (fourfold) against DNA fragmentation associated with combination drug treatments by dual enzyme expression compared to non-modified cells. Hematopoietic progenitor assays of murine bone marrow cells transduced with retroviral vectors encoding TS and TDP-I showed that bone marrow cells could be protected from the cytotoxic effects of TS and Top-I inhibition. CONCLUSIONS: Enforced expression of optecTS and TDP-I conferred antifolate and CPT resistance to genetically modified cells. Additionally, this work further illustrated a role for TDP-I in the repair of dead-end Top-I complexes and implied that TDP-I expression analysis may aid in predicting the therapeutic effectiveness of the CPT class of compounds.

Differences in natural ligand and fluoropyrimidine binding to human thymidylate synthase identified by transient-state spectroscopic and continuous variation methods.

Thymidylate synthase (TS) is a central target for the design of chemotherapeutic agents due to its vital role in DNA synthesis. Structural studies of binary complexes between Escherichia coli TS and various nucleotides suggest the chemotherapeutic agent FdUMP and the natural ligand dUMP bind similarly. We show, however, that FdUMP binding to human TS yields a substantially greater decrease in fluorescence than does dUMP. Because the difference in quenching due to ligand binding was approximately two-fold and this difference was not seen when using ecTS, the intriguing result indicated a significant difference in the mode of FdUMP binding to the human enzyme. We compared the binding affinities of dUMP, FdUMP, and TMP to TS from both species and found no significant differences for the individual ligands. Because binding affinities were not different among the ligands, the method of continuous variation was employed to determine binding stoichiometry. Similar to that found for dUMP binding to human and ecTS, FdUMP displayed single site occupancy with both enzymes. These results show that nucleotide binding differences exist for FdUMP and dUMP binding to the human enzyme. The observed differences are not due to differences in stoichiometry or ligand affinity. Therefore, although the crystal structure of human TS with various nucleotide ligands has not been solved, these results show that the differences observed using fluorescence methods result from as yet unidentified differential interactions between the human enzyme and nucleotide ligands.