2,4-diamino-5-deaza-6-substituted pyrido[2,3-d]pyrimidine antifolates as potent and selective nonclassical inhibitors of dihydrofolate reductases.

Fifteen novel nonclassical and two classical 2,4-diamino-6-(benzylamino)pyrido[2,3-d]pyrimidine antifolates were synthesized as potential inhibitors of Pneumocystis carinii, (pc) Toxoplasma gondii, (tg) rat liver (rl), and human (h) recombinant dihydrofolate reductases (DHFR). These analogues lack a 5-methyl substitution which has been shown to be important for increased hDHFR inhibitory activity. In addition, they contain a reversal of the C9-N10 bridge present in folates and most antifolates. The synthesis of the compounds involved the reaction of 2,4,6-triaminopyrimidine with the sodium salt of nitromalonaldehyde to afford the key intermediate 2,4-diamino-6-nitropyrido[2,3-d]pyrimidine (7), in a single step. Reduction of 7 to the 2,4,6-triaminopyrido[2,3-d]pyrimidine (8), followed by reductive amination with the appropriate benzaldehydes or phenylacetaldehydes afforded the target compounds. N9 methylation of these analogues was carried out using formaldehyde and sodium cyanoborohydride. The analogues demonstrated significant inhibition of pcDHFR and tgDHFR. N9 methylation significantly increased DHFR inhibitory potency. Compound 11, the 3'4'5'-trimethoxy-substituted analogue with a selectivity ratio of 9.4 for tgDHFR (compared to rlDHFR) was the most selective analogue of the nonclassical series. Compound 22, the N9 methyl 2'5'-dimethoxy-substituted analogue was the most potent analogue against tgDHFR (IC 50 = 6.3 nM) and was the second most selective analogue for tgDHFR (compared to rlDHFR) in the nonclassical series. The naphthyl-substituted analogues 23-25 were generally more potent against rlDHFR than against pcDHFR and tgDHFR. Selected analogues were also evaluated against Streptococcus faecium (sf) DHFR, Escherichia coli (ec) DHFR, Lactobacillus casei (lc) DHFR and tgDHFR with hDHFR as the mammalian reference, under slightly different assay conditions than those employed for rlDHFR. Analogues 11 and 22 had selectivity ratios of greater than 100 for tgDHFR (compared to hDHFR). Analogue 22 in particular, was the most selective analogue of the nonclassical series against tgDHFR (selectivity ratio = 303.5) with excellent potency (28 nM). Analogue 11, also displayed significant selectivity for sfDHFR (selectivity ratio = 4902). Compound 22 was evaluated in vivo for the inhibition of the growth of T.gondii trophozoites in mice, where at 50 mg/kg orally, it demonstrated distinct prolongation of survival without toxicity. Compounds 11, 12 and 21-23 were evaluated as antitumor agents in the National Cancer Institutes preclinical in vitro screening program. Compounds 12, 22, and 23 showed GI50s for tumor growth inhibition in the 10 -6 - 10 -7 M range.