Structure-based approach to pharmacophore identification, in silico screening, and three-dimensional quantitative structure-activity relationship studies for inhibitors of Trypanosoma cruzi dihydrofolate reductase function.

We have employed a structure-based three-dimensional quantitative structure-activity relationship (3D-QSAR) approach to predict the biochemical activity for inhibitors of T. cruzi dihydrofolate reductase-thymidylate synthase (DHFR-TS). Crystal structures of complexes of the enzyme with eight different inhibitors of the DHFR activity together with the structure in the substrate-free state (DHFR domain) were used to validate and refine docking poses of ligands that constitute likely active conformations. Structural information from these complexes formed the basis for the structure-based alignment used as input for the QSAR study. Contrary to indirect ligand-based approaches the strategy described here employs a direct receptor-based approach. The goal is to generate a library of selective lead inhibitors for further development as antiparasitic agents. 3D-QSAR models were obtained for T. cruzi DHFR-TS (30 inhibitors in learning set) and human DHFR (36 inhibitors in learning set) that show a very good agreement between experimental and predicted enzyme inhibition data. For crossvalidation of the QSAR model(s), we have used the 10% leave-one-out method. The derived 3D-QSAR models were tested against a few selected compounds (a small test set of six inhibitors for each enzyme) with known activity, which were not part of the learning set, and the quality of prediction of the initial 3D-QSAR models demonstrated that such studies are feasible. Further refinement of the models through integration of additional activity data and optimization of reliable docking poses is expected to lead to an improved predictive ability.

Dicyclic and tricyclic diaminopyrimidine derivatives as potent inhibitors of Cryptosporidium parvum dihydrofolate reductase: structure-activity and structure-selectivity correlations.

A structurally diverse library of 93 lipophilic di- and tricyclic diaminopyrimidine derivatives was tested for the ability to inhibit recombinant dihydrofolate reductase (DHFR) cloned from human and bovine isolates of Cryptosporidium parvum (J. R. Vásquez et al., Mol. Biochem. Parasitol. 79:153-165, 1996). In parallel, the library was also tested against human DHFR and, for comparison, the enzyme from Escherichia coli. Fifty percent inhibitory concentrations (IC(50)s) were determined by means of a standard spectrophotometric assay of DHFR activity with dihydrofolate and NADPH as the cosubstrates. Of the compounds tested, 25 had IC(50)s in the 1 to 10 microM range against one or both C. parvum enzymes and thus were not substantially different from trimethoprim (IC(50)s, ca. 4 microM). Another 25 compounds had IC(50)s of <1.0 microM, and 9 of these had IC(50)s of <0.1 microM and thus were at least 40 times more potent than trimethoprim. The remaining 42 compounds were weak inhibitors (IC(50)s, >10 microM) and thus were not considered to be of interest as drugs useful against this organism. A good correlation was generally obtained between the results of the spectrophotometric enzyme inhibition assays and those obtained recently in a yeast complementation assay (V. H. Brophy et al., Antimicrob. Agents Chemother. 44:1019-1028, 2000; H. Lau et al., Antimicrob. Agents Chemother. 45:187-195, 2001). Although many of the compounds in the library were more potent than trimethoprim, none had the degree of selectivity of trimethoprim for C. parvum versus human DHFR. Collectively, the results of these assays comprise the largest available database of lipophilic antifolates as potential anticryptosporidial agents. The compounds in the library were also tested as inhibitors of the proliferation of intracellular C. parvum oocysts in canine kidney epithelial cells cultured in folate-free medium containing thymidine (10 microM) and hypoxanthine (100 microM). After 72 h of drug exposure, the number of parasites inside the cells was quantitated by indirect immunofluorescence microscopy. Sixteen compounds had IC(50)s of <3 microM, and five of these had IC(50)s of <0.3 microM and thus were comparable in potency to trimetrexate. The finding that submicromolar concentrations of several of the compounds in the library could inhibit in vitro growth of C. parvum in host cells in the presence of thymidine (dThd) and hypoxanthine (Hx) suggests that lipophilic DHFR inhibitors, in combination with leucovorin, may find use in the treatment of intractable C. parvum infections.

Relationship between dopamine-stimulated phospholipid methylation and the single-carbon folate pathway.

In a previous study we demonstrated the ability of dopamine (DA) to stimulate phospholipid methylation (PLM) via a novel mechanism involving the D4 dopamine receptor (D4R) in which single-carbon folates appeared to be the primary source of methyl groups. To further understand the relationship between D4R-mediated PLM and folate metabolism, we examined the effect of several folate pathway interventions on the level of basal and DA-stimulated incorporation of [14C]-labeled formate into phospholipids in cultured SH-SY5Y neuroblastoma cells. These interventions included: (i) Overexpression of methenyltetrahydrofolate synthetase (MTHFS). (ii) Treatment with 5-formylTHF. (iii) Treatment with the MTHFS inhibitor 5-formyltetrahydrohomofolic acid (5-formylTHHF). (iv) Growth in nucleoside-free media. 31P-NMR was also used to follow DA-induced changes in cell phospholipid composition. MTHFS overexpression and 5-formylTHHF treatment, both of which lower 5-methylTHF levels, each reduced basal PLM and its stimulation by DA. In contrast, 5-formylTHF, which increases 5-methylTHF, caused a dose-dependent increase in both basal and DA-stimulated PLM. Growth in nucleoside-free media caused time-dependent changes in PLM, which were due to the absence of purine nucleosides. While basal PLM was maintained at a reduced level, DA-stimulated PLM was initially increased followed by a later decrease. Together, these findings indicate a close functional relationship between single-carbon folate metabolism and DA-stimulated PLM, consistent with a role for 5-methylTHF as the methyl donor for the D4R-mediated process.

Efficacies of lipophilic inhibitors of dihydrofolate reductase against parasitic protozoa.

Competitive inhibitors of dihydrofolate reductase (DHFR) are used in chemotherapy or prophylaxis of many microbial pathogens, including the eukaryotic parasites Plasmodium falciparum and Toxoplasma gondii. Unfortunately, point mutations in the DHFR gene can confer resistance to inhibitors specific to these pathogens. We have developed a rapid system for testing inhibitors of DHFRs from a variety of parasites. We replaced the DHFR gene from the budding yeast Saccharomyces cerevisiae with the DHFR-coding region from humans, P. falciparum, T. gondii, Pneumocystis carinii, and bovine or human-derived Cryptosporidium parvum. We studied 84 dicyclic and tricyclic 2,4-diaminopyrimidine derivatives in this heterologous system and identified those most effective against the DHFR enzymes from each of the pathogens. Among these compounds, six tetrahydroquinazolines were effective inhibitors of every strain tested, but they also inhibited the human DHFR and were not selective for the parasites. However, two quinazolines and four tetrahydroquinazolines were both potent and selective inhibitors of the P. falciparum DHFR. These compounds show promise for development as antimalarial drugs.

Efficient utilization of the reduced folate carrier in CCRF-CEM human leukemic lymphoblasts by the potent antifolate N(alpha)-(4-amino-4-deoxypteroyl)-N(delta)-hemiphthaloyl-L- ornithine (PT523) and its B-ring analogues.

The potent nonpolyglutamatable dihydrofolate reductase inhibitor N(alpha)-(4-amino-4-deoxypteroyl)-N(delta)-hemiphthaloyl-L-o rnithine (PT523) and six of its B-ring (5-deaza, 8-deaza, and 5,8-dideaza) analogues were compared in terms of their ability to: (a) inhibit the growth of CCRF-CEM human leukemic lymphoblasts, and (b) utilize the reduced folate carrier (RFC) in these cells as measured in a competition assay of [(3)H]methotrexate ([(3)H]MTX) influx. The IC(50) values of the hemiphthaloylornithine derivatives against CCRF-CEM cells after 72 hr of drug exposure varied from 0.64 to 1.3 nM as compared with 14 nM for MTX and 4.4 nM for aminopterin (AMT). The K(i) values of these compounds in the [(3)H]MTX influx assay were in the 0.3 to 0.7 microM range as compared with a K(i) of 5.4 microM for AMT and a K(t) of 7.1 microM for MTX. As a group, the affinities of these compounds for the RFC were approximately 10-fold greater than those of their respective glutamate analogues. These results indicate that, in addition to their previously reported tight binding to dihydrofolate reductase, a property contributing to the high potency of PT523 and its B-ring analogs as inhibitors of tumor cell growth is their strong affinity for the RFC.

Analogues of the potent nonpolyglutamatable antifolate N(alpha)-(4-amino-4-deoxypteroyl)-N(delta)-hemiphthaloyl-L-ornithine (PT523) with modifications in the side chain, p-aminobenzoyl moiety, or 9,10-bridge: synthesis and in vitro antitumor activity.

Seven N(alpha)-(4-amino-4-deoxypteroyl)-N(delta)-hemiphthaloyl-L-o rnithine (2, PT523) analogues were synthesized by modifications of the literature synthesis of the corresponding AMT (1) analogues and were tested as inhibitors of tumor cell growth. In growth assays against cultured CCRF-CEM human leukemic cells exposed to drug for 72 h, the IC(50) values of analogues in which N(10) was replaced by CH(2) and CHMe were found to be 0.55 +/- 0.07 and 0.63 +/- 0.08 nM, and thus these analogues are more potent than 1 (IC(50) = 4.4 +/- 1.0 nM) or 2 (IC(50) = 1.5 +/- 0.39 nM). The 10-ethyl-10-deaza analogue of 2 (IC(50) = 1.2 +/- 0.25 nM) was not statistically different from 2 but was more potent than edatrexate, the 10-ethyl-10-deaza analogue of 1, which had an IC(50) of 3.3 +/- 0.36 nM. In contrast, the analogue of 2 with both an ethyl and a CO(2)Me group at the 10-position had an IC(50) of 54 +/- 4.9 nM, showing this modification to be unfavorable. The 4-amino-1-naphthoic acid analogue of 2 had an IC(50) of 1.2 +/- 0.22 nM, indicating that replacement of the p-aminobenzoic acid (pABA) moiety does not diminish cytotoxicity. The analogues in which the (CH(2))(3) side chain was replaced by slightly longer CH(2)SCH(2) and (CH(2))(2)SCH(2) groups gave IC(50) values of 4.4 +/- 1.1 and 5.0 +/- 0.56 nM and thus were somewhat less potent than the parent molecule. However the analogues in which the aromatic COOH group was at the meta and para positions of the phthaloyl ring had IC(50) values of 7.5 +/- 0.47 and 55 +/- 0.07 nM, confirming the low potency we had previously observed with these compounds against other cell lines. Overall, the results in this study support the conclusion that, while the position of the phthaloyl COOH group and the length of the amino acid side chain in 2 are important determinants of cytotoxic potency, changes in the pABA region and 9, 10-bridge are well-tolerated and can even increase potency.

The effect of side-chain, para-aminobenzoyl region, and B-ring modifications on dihydrofolate reductase binding, influx via the reduced folate carrier, and cytotoxicity of the potent nonpolyglutamatable antifolate N(alpha)-(4-amino-4-deoxypteroyl)-N(delta)-hemiphthaloyl-L- ornithine.

N(alpha)-(4-Amino-4-deoxypteroyl)-N(delta)-hemiphthaloyl-L-o rnithine (PT523) is an unusually tight-binding dihydrofolate reductase (DHFR) inhibitor and is efficiently taken up into cells via the reduced folate carrier (RFC). Unlike classical DHFR inhibitors with a glutamate side chain, such as methotrexate and aminopterin, PT523 cannot form polyglutamates. Thus, it resembles lipophilic antifolates such as trimetrexate in not requiring metabolic activation by folylpolyglutamate synthetase in order to produce its antifolate effect. However, in contrast to trimetrexate, PT523 retains growth inhibitory activity in cells with the multidrug resistance phenotype. As part of the preclinical development of this drug, we have performed systematic modification of several regions of the PT523 molecule, with the aim of defining the optimal structural features for DHFR binding, influx into cells via the RFC, and the ability to inhibit cell growth. The following structure-activity correlations have emerged from this ongoing investigation, and are discussed: (1) the hemiphthaloylornithine side chain has the optimal length; (2) the preferred location of the aromatic carboxyl group is the ortho position; and (3) replacement of the phenyl ring of the para-aminobenzoic acid moiety by naphthalene, of nitrogen at the 10-position of the bridge by carbon, and of nitrogen at the 5- and/or 8-position of the B-ring by carbon are all well tolerated. Several of the second generation analogs of PT523 are more potent DHFR inhibitors and better RFC substrates than PT523 itself, and are more potent inhibitors of tumor cell growth in culture.

Identification of Cryptosporidium parvum dihydrofolate reductase inhibitors by complementation in Saccharomyces cerevisiae.

There is a pressing need for drugs effective against the opportunistic protozoan pathogen Cryptosporidium parvum. Folate metabolic enzymes and enzymes of the thymidylate cycle, particularly dihydrofolate reductase (DHFR), have been widely exploited as chemotherapeutic targets. Although many DHFR inhibitors have been synthesized, only a few have been tested against C. parvum. To expedite and facilitate the discovery of effective anti-Cryptosporidium antifolates, we have developed a rapid and facile method to screen potential inhibitors of C. parvum DHFR using the model eukaryote, Saccharomyces cerevisiae. We expressed the DHFR genes of C. parvum, Plasmodium falciparum, Toxoplasma gondii, Pneumocystis carinii, and humans in the same DHFR-deficient yeast strain and observed that each heterologous enzyme complemented the yeast DHFR deficiency. In this work we describe our use of the complementation system to screen known DHFR inhibitors and our discovery of several compounds that inhibited the growth of yeast reliant on the C. parvum enzyme. These same compounds were also potent or selective inhibitors of the purified recombinant C. parvum DHFR enzyme. Six novel lipophilic DHFR inhibitors potently inhibited the growth of yeast expressing C. parvum DHFR. However, the inhibition was nonselective, as these compounds also strongly inhibited the growth of yeast dependent on the human enzyme. Conversely, the antibacterial DHFR inhibitor trimethoprim and two close structural analogs were highly selective, but weak, inhibitors of yeast complemented by the C. parvum enzyme. Future chemical refinement of the potent and selective lead compounds identified in this study may allow the design of an efficacious antifolate drug for the treatment of cryptosporidiosis.

Structure-based design of selective inhibitors of dihydrofolate reductase: synthesis and antiparasitic activity of 2, 4-diaminopteridine analogues with a bridged diarylamine side chain.

As part of a larger search for potent as well as selective inhibitors of dihydrofolate reductase (DHFR) enzymes from opportunistic pathogens found in patients with AIDS and other immune disorders, N-[(2,4-diaminopteridin-6-yl)methyl]dibenz[b,f]azepine (4a) and the corresponding dihydrodibenz[b,f]azepine, dihydroacridine, phenoxazine, phenothiazine, carbazole, and diphenylamine analogues were synthesized from 2, 4-diamino-6-(bromomethyl)pteridine in 50-75% yield by reaction with the sodium salts of the amines in dry tetrahydrofuran at room temperature. The products were tested for the ability to inhibit DHFR from Pneumocystis carinii (pcDHFR), Toxoplasma gondii (tgDHFR), Mycobacterium avium (maDHFR), and rat liver (rlDHFR). The member of the series with the best combination of potency and species selectivity was 4a, with IC(50) values against the four enzymes of 0. 21, 0.043, 0.012, and 4.4 microM, respectively. The dihydroacridine, phenothiazine, and carbazole analogues were also potent, but nonselective. Of the compounds tested, 4a was the only one to successfully combine the potency of trimetrexate with the selectivity of trimethoprim. Molecular docking simulations using published 3D structural coordinates for the crystalline ternary complexes of pcDHFR and hDHFR suggested a possible structural interpretation for the binding selectivity of 4a and the lack of selectivity of the other compounds. According to this model, 4a is selective because of a unique propensity of the seven-membered ring in the dibenz[b,f]azepine moiety to adopt a puckered orientation that allows it to fit more comfortably into the active site of the P. carinii enzyme than into the active site of the human enzyme. Compound 4a was also evaluated for the ability to be taken up into, and retard the growth of, P. carinii and T. gondii in culture. The IC(50) of 4a against P. carinii trophozoites after 7 days of continuous drug treatment was 1.9 microM as compared with previously observed IC(50) values of >340 microM for trimethoprim and 0.27 microM for trimetrexate. In an assay involving [(3)H]uracil incorporation into the nuclear DNA of T. gondii tachyzoites as the surrogate endpoint for growth, the IC(50) of 4a after 5 h of drug exposure was 0.077 microM. The favorable combination of potency and enzyme selectivity shown by 4a suggests that this novel structure may be an interesting lead for structure-activity optimization.

5-deazafolate analogues with a rotationally restricted glutamate or ornithine side chain: synthesis and binding interaction with folylpolyglutamate synthetase.

Rotationally restricted analogues of 5-deazapteroyl-L-glutamate and (6R,6S)-5-deaza-5,6,7,8-tetrahydropteroyl-L-glutamate with a one-carbon bridge between the amide nitrogen and the 6'-position of the p-aminobenzoyl moiety were synthesized and tested as substrates for folylpolyglutamate synthetase (FPGS), a key enzyme in folate metabolism and an important determinant of the therapeutic potency and selectivity of classical antifolates. The corresponding bridged analogues of 5-deazapteroyl-L-ornithine and (6R,6S)-5-deaza-5,6,7, 8-tetrahydropteroyl-L-ornithine were also synthesized as potential inhibitors. Condensation of diethyl L-glutamate with methyl 2-bromomethyl-4-nitrobenzoate followed by catalytic reduction of the nitro group, reductive coupling with 2-acetamido-6-formylpyrido[2, 3-d]pyrimidin-4(3H)-one in the presence of dimethylaminoborane, and acidolysis with HBr/AcOH yielded 2-L-[5-[N-(2-acetamido-4(3H)-oxopyrido[2, 3-d]pyrimidin-6-yl)methylamino]-2, 3-dihydro-1-oxo-2(1H)-isoindolyl]glutaric acid (1). When acidolysis was preceded by catalytic hydrogenation, the final product was the corresponding (6R,6S)-tetrahydro derivative 2. A similar sequence starting from methyl N(delta)-benzyloxycarbonyl-L-ornithine led to 2-L-[5-[N-(2-amino-4(3H)-oxopyrido[2, 3-d]pyrimidin-6-yl)methylamino]-2, 3-dihydro-1-oxo-2(1H)-isoindolyl]-5-aminopentanoic acid (3) and the (6R,6S)-tetrahydro derivative 4. Compounds 3 and 4 were powerful inhibitors of recombinant human FPGS, whereas 1 and 2 were exceptionally efficient FPGS substrates, with the reduced compound 2 giving a K(m) (0.018 microM) lower than that of any other substrate identified to date. (6R,6S)-5-Deazatetrahydrofolate, in which the side chain is free to rotate, was rapidly converted to long-chain polyglutamates. In contrast, the reaction of 1 and 2 was limited to the addition of a single molecule of glutamic acid. Hence rotational restriction of the side chain did not interfere with the initial FPGS-catalyzed reaction and indeed seemed to facilitate it, but the ensuing gamma-glutamyl adduct was no longer an efficient substrate for the enzyme.

Inhibition of dihydrofolate reductases from Toxoplasma gondii, Pneumocystis carinii, and rat liver by rotationally restricted analogues of pyrimethamine and metoprine.

Twenty-one conformationally restricted tricyclic pyrimethamine and metoprine analogues with one or two chlorine atoms, or other substituents, at different positions of the phenyl ring were tested for potency and species selectivity against dihydrofolate reductase (DHFR) from Toxoplasma gondii, Pneumocystis carinii, and rat liver. Heterocyclic systems studied included indeno[2,1-d]pyrimidines, benzo[f]quinazolines, and benzo[3,4]cyclohepta[1,2-d]pyrimidines. All but one of the analogues were more potent against T. gondii and rat liver DHFR than against P. carinii DHFR, and those with a one-carbon (CH2) bridge were generally less potent than those with a two-carbon (CH2CH2, CH=CH) or three-carbon (CH2CH2CH2) bridge. Although a number of compounds with a two- and three-carbon bridge were more potent than pyrimethamine against P. carinii DHFR, and especially T. gondii DHFR, none of them were selective for the P. carinii versus the mammalian enzyme, and only those with a one-carbon bridge showed selectivity approaching that of pyrimethamine for the T. gondii enzyme. Computer-simulated docking into the active site pocket of P. carinii and human DHFR suggested that, as a group, the rotationally restricted tricyclic structures are at a disadvantage relative to pyrimethamine and metoprine, in that torsional relief of unfavorable steric interactions between the chlorine atoms and two critical serine and threonine residues in the active site is prevented by the bridge.

PT523 and other aminopterin analogs with a hemiphthaloyl-L-ornithine side chain: exceptionally tight-binding inhibitors of dihydrofolate reductase which are transported by the reduced folate carrier but cannot form polyglutamates.

Nonpolyglutamatable antifolates are potentially of therapeutic interest for the treatment of tumors that are inherently refractory, or have become resistant, to classical antifolates as a result of decreased expression of the enzyme folylpolyglutamate synthetase. An interesting class of water-soluble nonpolyglutamatable analogs of aminopterin (AMT) have been developed, which are much more cytotoxic because they bind more tightly to dihydrofolate reductase (DHFR) and also utilize the reduced folate carrier (RFC) pathway more efficiently for influx into the cell. This review summarizes the in vitro and in vivo preclinical data on the initial lead compound, Nalpha-(4-amino-4-deoxypteroyl)-Ndelta- hemiphthaloyl-L-ornithine (PT523). In addition, the synthesis and in vitro biochemical and biological properties of several types of second-generation analogs are discussed. Analogs modified in the B-ring of the pteridine moiety have been found to be of particular interest because their affinity for DHFR and their influx rate into cells via the RFC pathway are even greater than those of PT523. The hemiphthaloylornithine moiety, which is larger and more hydrophobic than the glutamate side chain of classical antifolates, appears to be chiefly responsible for the exceptionally high biological potency of PT523 and its B-ring analogs.

Synthesis and antiparasitic and antitumor activity of 2, 4-diamino-6-(arylmethyl)-5,6,7,8-tetrahydroquinazoline analogues of piritrexim.

Nineteen previously undescribed 2,4-diamino-6-(arylmethyl)-5,6,7, 8-tetrahydroquinazolines (5a-m, 10-12) were synthesized as part of a larger effort to assess the therapeutic potential of lipophilic dihydrofolate reductase (DHFR) inhibitors against opportunistic infections of AIDS. Condensation of appropriately substituted (arylmethyl)triphenylphosphoranes with 4, 4-ethylenedioxycyclohexanone, followed by hydrogenation (H2/Pd-C) and acidolysis, yielded the corresponding 4-(arylmethyl)cyclohexanones, which were then condensed with cyanoguanidine to form the tetrahydroquinazolines. Three simple 2, 4-diamino-6-alkyl-5,6,7,8-tetrahydroquinazoline model compounds (9a-c) were also prepared in one step from commercially available 4-alkylcyclohexanones by this method. Enzyme inhibition assays against rat liver DHFR, Pneumocystis carinii DHFR, and the bifunctional DHFR-TS enzyme from Toxoplasma gondii were carried out, and the selectivity ratios IC50(rat)/IC50(P. carinii) and IC50(rat)/IC50(T. gondii) were compared. The three most potent inhibitors of P. carinii DHFR were the 2,5-dimethoxybenzyl (5j), 3, 4-dimethoxybenzyl (5k), and 3,4,5-trimethoxybenzyl (5l) analogues, with IC50 values of 0.057, 0.10, and 0.091 microM, respectively. The remaining compounds generally had IC50 values in the 0.1-1.0 microM range. However all the compounds were more potent against the rat liver enzyme than the P. carinii enzyme and thus were nonselective. The T. gondii enzyme was always more sensitive than the P. carinii enzyme, with most of the analogues giving IC50 values of 0.01-0.1 microM. Moderate 5-10-fold selectivity for T. gondii versus rat liver DHFR was observed with five compounds, the best combination of potency and selectivity being achieved with the 2-methoxybenzyl analogue 5d, which had an IC50 of 0.014 microM and a selectivity ratio of 8.6. One compound (5l) was tested for antiproliferative activity against P. carinii trophozoites in culture at a concentration of 10 microgram/mL and was found to completely suppress growth over 7 days. The suppressive effect of 5l was the same as that of trimethoprim (10 microgram/mL) + sulfamethoxazole (250 microgram/mL), a standard clinical combination for the treatment of P. carinii pneumonia in AIDS patients. Four compounds (5a,h,k,l) were tested against T. gondii tachyzoites in culture and were found to have a potency (IC50 = 0.1-0.5 microM) similar to that of pyrimethamine (IC50 = 0.69 microM), a standard clinical agent for the treatment of cerebral toxoplasmosis in AIDS patients. Compound 5h was also active against T. gondii infection in mice when given qdx8 by peritoneal injection at doses ranging from 62.5 (initial dose) to 25 mg/kg. Survival was prolonged to the same degree as with 25 mg/kg clindamycin, another widely used drug against toxoplasmosis. Three compounds (5j-l) were tested for antiproliferative activity against human tumor cells in culture. Among the 25 cell lines in the National Cancer Institute panel for which data were confirmed in two independent experiments, the IC50 for at least two of these compounds was <10 microM against 17 cell lines (68%) and in the 0. 1-1 microM range against 13 cell lines (52%). One compound (5j) had an IC50 of <0.01 microM against four of the cell lines. The activity profiles of 5k,l were generally similar to that of 5j except that there were no cells against which the IC50 was <0.01 microM.

Synthesis and potent antifolate activity and cytotoxicity of B-ring deaza analogues of the nonpolyglutamatable dihydrofolate reductase inhibitor Nalpha-(4-amino-4-deoxypteroyl)-Ndelta-hemiphthaloyl- L-ornithine (PT523).

Six new B-ring analogues of the nonpolyglutamatable antifolate Nalpha-(4-amino-4-deoxypteroyl)-Ndelta-hemiphthaloy l-L-ornithine (PT523, 3) were synthesized with a view to determining the effect of modifications at the 5- and/or 8-position on dihydrofolate reductase (DHFR) binding and tumor cell growth inhibition. The 5- and 8-deaza analogues were prepared from methyl 2-L-amino-5-phthalimidopentanoate and 4-amino-4-deoxy-N10-formyl-5-deaza- and -8-deazapteroic acid, respectively. The 5,8-dideaza analogues were prepared from methyl 2-L-[(4-aminobenzoyl)amino]-5-phthalimidopentanoate and 2, 4-diaminoquinazoline-6-carbonitriles. The Ki for inhibition of human DHFR by the 5-deaza and 5-methyl-5-deaza analogues was about the same as that of 3 (0.35 pM), 11-fold lower than that of aminopterin (AMT, 1), and 15-fold lower than that of methotrexate (MTX, 2). However the Ki of the 8-deaza analogue was 27-fold lower than that of 1, and that of the 5,8-dideaza, 5-methyl-5,8-dideaza, and 5-chloro-5,8-dideaza analogues was approximately 50-fold lower. This trend was consistent with the published literature on the corresponding DHFR inhibitors with a glutamate side chain. In colony formation assays against the human head and neck squamous carcinoma cell line SCC25 after 72 h of treatment, the 5- and 8-deaza analogues were approximately as potent as 3, whereas the 5,8-dideaza analogue was 3 times more potent. 5-Methyl and 5-chloro substitution was also favorable, with the 5-methyl-5-deaza analogue being 2. 5-fold more potent than the 5-deaza analogue. However the effect of 5-methyl substitution was less pronounced in the 5,8-dideaza analogues than in the 5-deaza analogues. The 5-chloro-5,8-dideaza analogue of 3 was the most active member of the series, with an IC50 = 0.33 nM versus 1.8 nM for 3 and 15 nM for MTX. The 5-methyl-5-deaza analogue of 3 was also tested at the National Cancer Institute against a panel of 50 human tumor cell lines in culture and was consistently more potent than 3, with IC50 values in the low-nanomolar to subnanomolar range against most of the tumors. Leukemia and colorectal carcinoma cell lines were generally most sensitive, though good activity was also observed against CNS tumors and carcinomas of the breast and prostate. The results of this study demonstrate that B-ring analogues of 3 inhibit DHFR activity and tumor cell colony formation as well as, or better than, the parent compound. In view of the fact that 3 and its B-ring analogues cannot form polyglutamates, their high cytotoxicity relative to the corresponding B-ring analogues of AMT is noteworthy.

Pharmacokinetics, antifolate activity and tissue distribution of PT523 in SCC VII tumor-bearing mice.

PURPOSE: To monitor the pharmacokinetics of PT523 and methotrexate in C3H mice with transplanted SCC VII tumors; to compare the impact of PT523 and methotrexate on tumor and normal host 5,10-methylenetetrahydrofolate levels; and to synthesize [14C]PT523 and determine its time-dependent tissue distribution in tumor and host tissues. METHODS: C3H mice bearing SCC VII tumors were given i.p. PT523 or methotrexate. Plasma drug levels and tumor, gut and marrow 5,10-methylenetetrahydrofolate were assayed. [14C]PT523 was synthesized and administered i.v. to tumor-bearing mice for tissue distribution analysis. RESULTS: Areas under the curve, mean residence times, whole body clearances, apparent distribution volumes, and plasma protein binding of PT523 vs methotrexate were, respectively, 4311 vs 6472 microM x min(-1); 20 vs 16 min; 0.56 vs 0.36 ml min(-1); 532 vs 325 ml x kg(-1); and 70% vs 30%. Both PT523 and methotrexate caused time-dependent declines in 5,10-methylenetetrahydrofolate in tumor and marrow, but not in gut mucosa [corrected]. Gut levels began to recover within 4 h in the PT523-treated group only. [14C]PT523 distributed mainly into the liver, duodenum, kidneys, lungs, tumor, pancreas and muscle; less into the spleen, blood cells, heart, brain and testicles; and very little into gut [corrected. Only 35% of the dose was excreted, and 2.9-fold more in feces than urine. CONCLUSIONS: Despite its more rapid clearance, accumulation of PT523 in extravascular tissues was greater than that of methotrexate. Consequently, less PT523 was recovered in feces and urine and its apparent volume of distribution was greater. PT523 selectively depleted 5,10-methylenetetrahydrofolate pools in tumor and, less persistently, in marrow, but spared the gut mucosa [corrected]. [14C]PT523 tissue distribution correlated with organ mass and blood supply.

2,4-Diamino-6,7-dihydro-5H-cyclopenta[d]pyrimidine analogues of trimethoprim as inhibitors of Pneumocystis carinii and Toxoplasma gondii dihydrofolate reductase.

Three previously unreported (R,S)-2,4-diamino-5-[(3,4,5-trimethoxyphenyl) alkyl]-6,7-dihydro-5H-cyclopenta[d]pyrimidines 15a-c were synthesized as analogues of trimethoprim (TMP) and were tested as inhibitors of Pneumocystis carinii, Toxoplasma gondii, and rat liver dihydrofolate reductase (DHFR). The length of the alkyl bridge between the cyclopenta[d]pyrimidine and trimethoxyphenyl moiety ranged from one in 15a to three carbons in 15c. The products were tested as competitive inhibitors of the reduction of dihydrofolate by Pneumocystis carinii, Toxoplasma gondii, and rat liver DHFR. Compounds 15a-c had IC50 values of > 32, 1.8 and 1.3 microM, respectively, against P. carinii DHFR, as compared to 12 microM for TMP. Against the T. gondii enzyme, 15a-c had IC50 values of 21, 0.14 and 0.14 microM, respectively, as compared to 2.7 microM for TMP. Inhibitors 15b and 15c with two- and three-carbon bridges were significantly more potent than 15a against all three enzymes. Unlike TMP, 15b and 15c were better inhibitors of the rat liver enzyme than of the microbial enzymes. The potency of 15b and 15c against rat liver DHFR was less than has been reported for the corresponding 6,7-dihydro-5H-cyclopenta[d]pyrimidines with a classical p-aminobenzoyl-L-glutamate side chain as inhibitors of bovine, murine, and human DHFR.

Comparison of two independent crystal structures of human dihydrofolate reductase ternary complexes reduced with nicotinamide adenine dinucleotide phosphate and the very tight-binding inhibitor PT523.

Structural data for two independent crystal forms (monoclinic, C2, and orthorhombic, P2(1)2(1)2(1)) of the ternary complex of the potent antitumor agent PT523 [N alpha-(4-amino-4-deoxypteroyl)-N delta-hemiphthaloyl-L-ornithine], reduced nicotinamide adenine dinucleotide phosphate (NADPH), and recombinant human dihydrofolate reductase (hDHFR) reveals multiple binding orientations for the hemiphthaloyl group of the inhibitor. Analysis of these data shows that PT523 binds with its pteridine ring in the same orientation observed for methotrexate (MTX) analogues. However, in each structure, the hemiphthaloyl ring occupies three alternate conformations. In the C2 lattice, the phthaloyl moiety binds in two extended conformations, A and C, with each conformer having a 180 degrees flip of the o-carboxylate group, and a third, lower occupancy conformer B, with the phthaloyl group folded within contact of the active-site pocket. In the orthorhombic lattice, PT523 also has three conformers for the phthaloyl group; however, these differ from those observed in the monoclinic lattice. Two major conformers, A and C, are displaced on either side of the extended position observed in the C2 lattice, one near the folded B conformer of the C2 lattice and the other extended. These conformers form tighter intermolecular contacts than those in the C2 lattice. Conformer B is folded back away from the active site in a unique position. There are also significant differences in the conformation of the adenine-ribose moiety of NADPH in both complexes that differ from that observed for other inhibitor-NADPH-hDHFR ternary complexes. These data suggest that the added intermolecular contacts made by the hemiphaloyl group of PT523 contribute to its tighter binding to hDHFR than MTX, which does not extend as far from the active site and cannot make these contacts. These crystallographic observations of multiple conformations for the hemiphthaloyl group are in general agreement with solution NMR data for the binding of PT523 to hDHFR [Johnson et al. (1997) Biochemistry 36, 4399-4411], which show that the hemiphthaloyl group may adopt more than one conformation. However, the crystallographic data reveal more discretely occupied positions than can be interpreted from the solution data. These results suggest that crystal packing interactions may influence their stability.

2,4-Diaminothieno[2,3-d]pyrimidine lipophilic antifolates as inhibitors of Pneumocystis carinii and Toxoplasma gondii dihydrofolate reductase.

Ten previously unreported 2,4-diaminothieno[2,3-d]pyrimidine lipophilic dihydrofolate reductase inhibitors were synthesized as potential inhibitors of Pneumocystis carinii and Toxoplasma gondii dihydrofolate reductase. Pivaloylation of 2,4-diamino-5-methylthieno[2,3-d]pyrimidine followed by dibromination with N-bromosuccinimide in the presence of benzoyl peroxide gave 2,4-bis(pivaloylamino)-6-bromo-5-(bromomethyl)thieno[2,3-d]pyrimid ine, which after condensation with substituted anilines or N-methylanilines and deprotection with base yielded 2,4-diamino-6-bromo-5-[(substituted anilino)methyl]thieno[2,3-d]pyrimidines. Removal of the 6-bromo substituent was accomplished with sodium borohydride and palladium chloride. The reaction yields were generally good to excellent. The products were tested as inhibitors of dihydrofolate reductase (DHFR) from P. carinii, T. gondii, and rat liver. Although the IC50 could not be reached for the 6-unsubstituted compounds because of their extremely poor solubility, three of the five 6-bromo derivatives were soluble enough to allow the IC50 to be determined against all three enzymes. 2,4-Diamino-5-[3,5-dichloro-4-(1-pyrrolo)anilino]methyl]- 6-bromothieno[2,3-d]pyrimidine was the most active of the 6-bromo derivatives, with an IC50 of 7.5 microM against P. carinii DHFR, but showed no selectivity for either P. carinii or T. gondii DHFR relative to the enzyme from rat liver.

Synthesis and in vitro activity of long-chain 5'-O-[(alkoxycarbonyl)phosphinyl]-3'-azido-3'-deoxythymidines against wild-type and AZT- and foscarnet-resistant strains of HIV-1.

Lipophilic esters of 3'-azido-3'-deoxy-5'-O-(carboxyphosphinyl)thymidine (PFA-AZT) were synthesized and tested for antiretroviral activity in CD4+ HT4-6C cells infected with either wild-type HIV-1LAI, a PFA-resistant strain encoding a single-point mutation in reverse transcriptase (E89K), or an AZT-resistant clinical isolate (A018-post). Arbuzov condensation of 1-octadecyl, 1-eicosanyl, and 1-docosanyl chloroformate with trimethyl phosphite yielded the corresponding dimethyl long-chain alkyl triesters of PFA. Selective removal of one methyl group from the triesters with sodium iodide yielded monosodium salts, whereas treatment with bromotrimethylsilane cleaved both methyl groups while leaving the long-chain alkyl group intact. Neutralization of the resulting [(alkyloxy)carbonyl]phosphonic acids with 2 equiv of sodium methoxide afforded disodium salts of the phosphonic acid moiety. Similar chemistry was used to obtain the mono- and disodium salts of the cholesterol ester of PFA. Reaction of the triesters with phosphorous pentachloride, followed by coupling with AZT and O-demethylation with sodium iodide, afforded 3'-azido-3'-deoxy-5'-O-[[(1-octadecyloxy)carbonyl]phosphinyl ]thymidine (9a), 3'-azido-3'-deoxy-5'-O-[[(1-eicosanyloxy)carbonyl]phosphinyl ]thymidine (9b), 3'-azido-3'-deoxy-5'-O-[[(1-docosanyloxy)carbonyl]phosphinyl ]thymidine (9c), and 3'-azido-3'-deoxy-5'-O-[[(3 beta-cholest-5-enyloxy)carbonyl]phosphinyl]thymidine (9d). Concentrations of 9a-d found to inhibit replication of wild-type HIV-1LAI by 50% (EC50 values) as measured in a plaque reduction assay were in the 0.1-0.3 microM range as compared with 0.013 microM for AZT and 133 microM for PFA. The concentration at which toxicity was observed in 50% of the host cells (TC50 values) as measured by a visual grading scale of cellular morphology was 10 microM for 9a and 9d, 32 microM for 9b, and 320 microM for 9c. Thus, the TC50/EC50 ratio or selectivity index (SI) was 100 for 9a, 230 for 9b, and 1000 for 9c but only 33 for 9d, suggesting that the straight-chained fatty alcohol esters were more therapeutically selective. Similar TC50 and SI values were obtained for rapidly dividing CEM lymphoblasts as for HT4-6C cells. In assays against E89K, 9a-c had mean EC50 values of 0.13, 0.009, and 0.17 microM, whereas the EC50 of PFA was > 1000 microM and that of AZT was 0.009 microM; thus, E89K was highly resistant to PFA but not cross-resistant to either AZT or the lipophilic PFA-AZT conjugates. In viral replication assays against the A018C-post isolate, the mean EC50 values of 9a-c were 0.30, 0.53, and 0.77 microM as compared with 2.9 microM for AZT and 65 microM for PFA; thus, the virus recovered from a patient pretreated with AZT was not cross-resistant to either PFA or 9a-c. A notable feature of these results was that, in addition to being > 1000-fold more potent than PFA against the PFA-resistant mutant, the lipophilic PFA-AZT conjugates were more potent than PFA, as well as AZT, against AZT-resistant HIV-1.