Biochemical and biological studies have been carried out with 2-desamino-2-methylaminopterin (dmAMT), which inhibits tumor cell growth in culture but is only a weak inhibitor of dihydrofolate reductase (DHFR). Since it was possible that the species responsible for growth inhibition are polyglutamylated metabolites, the di-, tri-, and tetraglutamates of dmAMT were synthesized and tested as inhibitors of purified recombinant human DHFR, murine L1210 leukemia thymidylate synthase (TS), chicken liver glycinamide ribonucleotide formyltransferase (GARFT), and murine L1210 leukemia aminoimidazolecarboxamide ribonucleotide formyltransferase (AICARFT). The compounds with three and four γ-glutamyl residues were found to bind two orders of magnitude better than dmAMT itself to DHFR, TS, and AICARFT, with 50% inhibitory concentration values in the 200 to 300 nm range against all three enzymes. In contrast, at a concentration of 10 µmm, dmAMT polyglutamates had no appreciable effect on GARFT activity. These findings support the hypothesis that dmAMT requires intracellular polyglutamylation for activity and indicate that replacement of the 2-amino group by 2-methyl is as acceptable a structural modification in antifolates targeted against DHFR as it is in antifolates targeted against TS. In growth assays against methotrexate (MTX)-sensitive H35 rat hepatoma cells and MTX-resistant H35 sublines with a transport defect, dmAMT was highly cross-resistant with MTX, but not with the TS inhibitors N10-propargyl-5,8-dideazafolic acid and N-{5-[N-(3,4-dihydro-2-methyl-4-oxoquinazolin-6-yl)-N-methylamino]thenoyl}-l-glutamic acid, implicating DHFR rather than TS as the principal target for dmAMT polyglutamates in intact cells. On the other hand, an H35 subline resistant to 2′-deoxy-5-fluorouridine by virtue of increased TS activity was highly cross-resistant to N10-propargyl-5,8-dideazafolic acid and not cross-resistant to MTX, but showed partial cross-resistance to dmAMT. Both thymidine and hypoxanthine were required to protect H35 cells treated with concentrations of dmAMT and MTX that inhibited growth by >90% relative to unprotected controls. In contrast, N10-propargyl-5,8-dideazafolic acid and N-{5-[N-(3,4-dihydro-2-methyl-4-oxoquinazolin-6-yl)-N-methylamino]thenoyl}-l-glutamic acid required only thymidine for protection. Like MTX, therefore, dmAMT appears to inhibit purine as well as pyrimidine de novo synthesis, and its effect on cell growth probably reflects the ability of dmAMT polyglutamates to not only block dihydrofolate reduction but also interfere with other steps of folate metabolism, either directly or indirectly via alteration of reduced folate pools. A similar protection pattern was obtained with mouse L1210 leukemia cells as with H35 cells, in that both thymidine and hypoxanthine were required for normal growth in the presence of dmAMT. Although folinic acid alone afforded full protection, 5-aminoimidazole-4-carboxamide could not be used instead of hypoxanthine, suggesting that de novo purine synthesis inhibition by dmAMT probably occurs at the level of AICARFT rather than GARFT. In antitumor assays against L1210 leukemia in mice, comparable lifespan increases were achieved with dmAMT and MTX, but more dmAMT than MTX had to be used to produce the same therapeutic effect. The results of this study suggest that dmAMT may be a promising lead for the development of other, more potent, 2-desamino analogues of classical 2,4-diamino antifolates.


This work was supported by NIH Grants CA19589 and CA25394 (A. R.), CA25933 (J. G.), CA57320 (G. P. B.), and CA41461 (J. H. F.). G. P. B. is a recipient of an American Cancer Society Faculty Research Award.

This content is only available via PDF.