Syntheses and antifolate activity of 5-methyl-5-deaza analogues of aminopterin, methotrexate, folic acid, and N10-methylfolic acid

Evidence indicating that modifications at the 5- and 10-positions of classical folic acid antimetabolites lead to compounds with favorable differential membrane transport in tumor vs. normal proliferative tissue prompted an investigation of 5-alkyl-5-deaza analogues. 2-Amino-4-methyl-3,5-pyridinedicarbonitrile, prepared by hydrogenolysis of its known 6-chloro precursor, was treated with guanidine to give 2,4-diamino-5-methylpyrido[2,3-d]pyrimidine-6-carbonitrile which was converted via the corresponding aldehyde and hydroxymethyl compound to 6-(bromomethyl)-2,4-diamino-5-methylpyrido[2,3-d]pyrimidine. Reductive condensation of the nitrile 8 with diethyl N-(4-amino-benzoyl)-L-glutamate followed by ester hydrolysis gave 5-methyl-5-deazaaminopterin. Treatment of 12 with formaldehyde and Na(CN)BH3 afforded 5-methyl-5-deazamethotrexate, which was also prepared from 15 and dimethyl N-[(4-methylamino)benzoyl]-L-glutamate followed by ester hydrolysis. 5-Methyl-10-ethyl-5-deazaaminopterin was similarly prepared from 15. Biological evaluation of the 5-methyl-5-deaza analogues together with previously reported 5-deazaaminopterin and 5-deazamethotrexate for inhibition of dihydrofolate reductase (DHFR) isolated from L1210 cells and for their effect on cell growth inhibition, transport characteristics, and net accumulation of polyglutamate forms in L1210 cells revealed the analogues to have essentially the same properties as the appropriate parent compound, aminopterin or methotrexate (MTX), except that 20 and 21 were approximately 10 times more growth inhibitory than MTX. In in vivo tests against P388/0 and P388/MTX leukemia in mice, the analogues showed activity comparable to that of MTX, with the more potent 20 producing the same response in the P388/0 test as MTX but at one-fourth the dose; none showed activity against P388/MTX. Hydrolytic deamination of 12 and 20 produced 5-methyl-5-deazafolic acid and 5,10-dimethyl-5-deazafolic acid, respectively. In bacterial studies on the 2-amino-4-oxo analogues, 5-deazafolic acid proved to be a potent inhibitor of Lactobacillus casei DHFR and also the growth of both L. casei ATCC 7469 and Streptococcus faecium ATCC 8043. Its 5-methyl congener 22 is also inhibitory toward L. casei, but its IC50 for growth inhibition is much lower than its IC50 values for inhibition of DHFR or thymidylate synthase from L. casei, suggesting an alternate site of action.

Synthesis and antifolate properties of 10-alkyl-8,10-dideazaminopterins

The synthesis of 10-alkyl analogues of the potent antitumor agent 8,10-dideazaminopterin is described. Alkylation of appropriate alpha-alkyl homoterephthalate esters with 2,4-diamino-6-(bromomethyl)-8-deazapteridine afforded 10-alkyl-10-carboxy-4-amino-4-deoxy-8,10-dideazapteroic acid diesters. Ester cleavage and decarboxylation at C-10 were accomplished by heating with sodium cyanide in Me2SO at 170-180 degrees C to afford the 2,4-diamino-10-alkyl-8,10-dideazapteroic acids. The acids were coupled with diethyl glutamate, followed by saponification, to give the 10-alkyl-8,10-dideazaminopterins. The compounds were potent inhibitors of growth in folate-dependent bacteria, Streptococcus faecium and Lactobacillus casei. The 10-methyl and 10-ethyl analogues gave the highest percent increases in life span for mice infected with L1210 leukemia with ILS values of +203 and +235%, respectively.

Folate analogues. 22. Synthesis and biological evaluation of two analogues of dihydrofolic acid possessing a 7,8-dihydro-8-oxapterin ring system

Two analogues of dihydrofolic acid possessing a 7,8-dihydro-8-oxapterin ring system have been synthesized and evaluated for their antifolate activities. These compounds, N-[(2-amino-4-hydroxy-7,8-dihydro-8-oxa-6-pteridinyl)benzoyl]-L-glutamic acid (3) and N-[[(2-amino-4-hydroxy-7,8-dihydro-8-oxa-6-pteridinyl) methyl]benzoyl]-L-glutamic acid (4), were synthesized by reacting the appropriately substituted alpha-halo ketones with 2,5-diamino-4,6-dihydroxypyrimidine (2). Elaboration of p-carbomethoxybenzaldehyde (5) to p-carbomethoxyphenacyl bromide (7) was accomplished by its oxidation with Jones reagent and the successive treatment of the oxidation product with SOCl2, CH2N2, and HBr. Commercially available p-vinylbenzoic acid (11) was converted to its glutamate conjugate 12 and was further converted to the bromo ketone, diethyl N-[p-(1-bromo-2-oxopropyl)benzoyl]-L-glutamate (17), by a series of reactions involving epoxidation, oxirane ring opening with HBr, Jones oxidation, Zn/HOAc reduction, and successive treatment of the reduction product 16 with SOCl2, CH2N2, and HBr. These bromo ketones, 7 and 17, upon reaction with pyrimidine 2, gave the diethyl esters of the target compounds, which were hydrolyzed to 3 and 4 with NaOH. Compound 4 underwent an interesting acid-catalyzed isomerization where the double bond of 4 was shifted from the 5,6-position to the 6,9-position to give the isomer 19. Both compounds 3 and 4 were inactive against Lactobacillus casei (ATCC 7469) and did not serve as synthetic substrates of L. casei dihydrofolate reductase. Compound 4 showed activity against Streptococcus faecium (ATCC 8043), but 3 was inactive against this organism.

Folate analogues. 21. Synthesis and antifolate and antitumor activities of N10-(cyanomethyl)-5,8-dideazafolic acid

A close analogue of the antileukemic agent 5,8-dideaza-N10 propargylfolic acid (2) was synthesized by replacing the propargyl moiety of 2 with a cyanomethyl group. This compound, N10-(cyanomethyl)-5,8-dideazafolic acid (3), was evaluated for its antifolate and antitumor activities in several biological test systems. Alkylation of diethyl N-(4-aminobenzoyl)-L-glutamate with bromoacetonitrile gave diethyl N-[4-[(cyanomethyl)amino]benzoyl]-L-glutamate (7). Reaction of 7 with 2 amino-6-(bromomethyl)-4-hydroxyquinazoline (9) in dimethylacetamide gave the corresponding diethyl ester 11, which was hydrolyzed to the target compound 3. The known antileukemic agent 2 was also synthesized for comparative studies by employing a modified procedure, which resulted in a better yield of this product. Both compounds 2 and 3 were evaluated for their antifolate activities by using two folate-requiring microorganisms, Streptococcus faecium and Lactobacillus casei. They were further evaluated as inhibitors of thymidylate synthase and dihydrofolate reductase derived from the above organisms, as well as for their antitumor activity by using selected tumor cells in culture. Compound 2 was found to be as equally potent as methotrexate (MTX) against S. faecium, and it was an excellent inhibitor of L. casei thymidylate synthase. The cyanomethyl analogue 3 was less active than 2 in all the test systems, except the inhibition of dihydrofolate reductase.

Folate analogues. 20. Synthesis and antifolate activity of 1',2',3',4',5',6'-hexahydrohomofolic acid

The synthesis of 1',2',3',4',5',6'-hexahydrohomofolic acid (3), a close analogue of homofolic acid (2), has been carried out by replacement of the benzene ring of 2 with a cyclohexane ring. The synthetic methods employed here were based on the Boon-Leigh strategy to obtain products with unambiguous structures. Based on a number of chemical and spectral observations, a tentative cis stereochemistry was assigned to the 1,4-substituents of the cyclohexane ring of both the homopteroate analogue 13 and the target compound 3. We investigated hexahydrohomopteroic acid (13), hexahydrohomofolic acid (3), and their 7,8-dihydro and d,l-5,6,7,8-tetrahydro derivatives for antifolate activity employing several biological test systems. The dihydro and tetrahydro derivatives of both 13 and 3 were active against Streptococcus faecium, whereas they were inactive against Lactobacillus casei. These compounds were neither substrates nor inhibitors of L. casei dihydrofolate reductase or thymidylate synthase.

An hypothesis on the role of pteroylpolyglutamate derivatives as coenzymes

Evidence is assembled supporting the view that the poly-gamma-Glu chain of pteroyl poly-gamma-Glu derivatives plays a role in coordinating the activities of enzymes involved in the sequential metabolism of single carbon units.

Synthesis and antitumor activity of 10-alkyl-10-deazaminopterins. A convenient synthesis of 10-deazaminopterin

Requirements for large-scale synthesis of the potent antitumor drug 10-deazaminopterin have led to development of a facile synthesis of this compound and its 10-alkyl analogues. The lithium diisopropyl amide generated dianions of appropriate p-alkylbenzoic acids were alkylated with 3-methoxyallyl chloride. The resulting 4-(p-carboxyphenyl)-1-methoxy-1-butenes were brominated at pH 7-8 to afford the 2-bromo-4-(p-carboxyphenyl)butyraldehydes. Condensation with 2,4,5,6-tetraminopyrimidine and subsequent in situ oxidation of the resulting dihydropteridines yielded crystalline 10-alkyl-10-deaza-4-amino-4-deoxypteroic acids. The pteroic acids were coupled with diethyl glutamate via the mixed anhydride method, followed by saponification at room temperature, to give the target 10-deazaminopterins. The 10-alkyl compounds were approximately equipotent to 10-deazaminopterin as growth inhibitors of folate-dependent bacteria. Their abilities to inhibit Lactobacillus casei and L1210 derived dihydrofolate reductases were also similar. Transport properties in vitro were suggestive of an improved therapeutic index for the 10-alkyl analogues. Against L1210 in mice, the percent increase in life span at the LD10 dosage was +151% (methotrexate), +178% (10-deazaminopterin), +235% (10-methyl analogue), and +211% (10-ethyl analogue). 10,10-Dimethyl-10-deazaminopterin was less effective at an equimolar dosage, but the ILS at the maximum dose tested (72 mg/kg) was +135%. It was far less toxic than the other analogues possibly because of enhanced clearance.

10-Propargylaminopterin and alkyl homologues of methotrexate as inhibitors of folate metabolism

Reported antifolate activity against leukemia L1210 by N-[14-[[(2-amino-4-hydroxy-6-quinazolinyl)methyl]-propargylamino]benzoyl]]-L-glu tamic acid through potent inhibition of thymidylate synthase (EC 2.1.1.45) prompted us to include the propargyl group in a study of the effect on folate metabolism and membrane transport of replacing the 10-methyl group of methotrexate with other groups. Along with the propyl (8a) and octyl (8b) homologues of methotrexate, the propargyl compound 8c was prepared for evaluation. Syntheses of 8a,b were achieved by a standard multistep sequence involving preparation of the side-chain precursors via tosylated intermediates and then their alkylation with 6-(bromomethyl)-2,4-pteridinediamine hydrobromide. The side-chain precursor to 8c was prepared by direct alkylation of diethyl N-(4-aminobenzoyl)-L-glutamate with propargyl bromide and was separated from unchanged amine and dipropargyl coproduct by a combination of methods, including dry-column chromatography and recrystallization. Subsequent steps leading to 8c were like those used to prepare 8a,b. Biological evaluations of the three compounds consisted of studies of their effects on enzyme inhibition [(dihydrofolate reductase (EC 1.5.1.3) and thymidylate synthase)], L1210 cell growth inhibition, cellular membrane transport with various murine cell types (L210, S180, Ehrlich, and epithelial), in vivo (mice) activity vs. L1210 leukemia and S180 ascites, and plasma clearance in mice. The in vivo results vs. S180 ascites offered evidence that 8c might have a better therapeutic index against this tumor than methotrexate, but no other result from either of these compounds suggested significant superiority over methotrexate.

Further studies stereospecificity at carbon 6 for membrane transport of tetrahydrofolates. Diastereoisomers of 5-methyltetrahydrofolates as competitive inhibitors of transport of methotrexate in L1210 cells

The unnatural d diastereoisomer at carbon 6 of 5-methyltetrahydrofolate was only slightly less effective than the natural 1 diastereoisomer as a competitive inhibitor of the carrier-mediated membrane transport of [3H]methotrexate into L1210 murine leukemia cells. The apparent Ki for a mixture containing equal amounts of both natural and unnatural diastereoisomers was not significantly different from that found for the unnatural form. These results show that the reduced folate carrier system in these cells has a strong affinity for the unnatural stereoisomer, a finding in contrast to that obtained with the corresponding diastereoisomer of 5-formyltetrahydrofolate.

Folate analogues altered in the C9-N10 bridge region. 18. Synthesis and antitumor evaluation of 11-oxahomoaminopterin and related compounds

The chemical synthesis of 11-oxahomoaminopterin (1) has been carried out using procedures which were also found to be applicable to the synthesis of 11-oxahomofolic acid (2). Reaction of 1-bromo-4-[p-(caarbomethoxy)phenoxy]-2-butanone (10) with sodium azide gave 1-azido-4-[p-(carbomethoxy)phenoxy]-2-butanone (11). Protection of the carbonyl group of 11 as the ethylene ketal and subsequent base hydrolysis of the product gave 1-azido-4-(p-carboxyphenoxy)-2-butanone ketal (13). The glutamate conjugate 14 was prepared from 13 by the isobutyl chloroformate method and was hydrogenated to diethyl N-[(alpha-amino-2-oxo-4-butanoyl)-p-anisoyl]-L-glutamate ketal (15). Reaction of 15 with 6-chloro-2,4-diamino-5-nitropyrimidine (16) and 2-amino-6-chloro-4-hydroxy-5-nitropyrimidine (17) and deprotection of the corresponding products gave the intermediates 18 and 19, which were elaborated to 1 and 2 using a series of steps involving deprotection, dithionite reduction, cyclization, oxidation, and hydrolysis. Although 11-oxahomoaminopterin showed antifolate activity against two folate-requiring microorganisms and inhibited Lactobacillus casei DHFR, it was inactive against L-1210 leukemia in mice at a maximum dose of 48 mg/kg. Compound Lactobacillus casei DHFR, it was inactive against L-1210 leukemia in mice at a maximum dose of 48 mg/kg. Compound 1 was also tested for its ability to be transported via the methotrexate transport system using the L-1210 and Ehrlich tumor cell lines, and these results are compared with those of related analogues. The growth inhibitory activity of 1 in the L-1210 cell lines in culture was found to be 15 times weaker than that of methotrexate.

Polyglutamyl derivatives of tetrahydrofolate as substrates for Lactobacillus casei thymidylate synthase

Tetrahydropteroylpolyglutamates containing up to seven Glu residues were tested as substrates for Lactobacillus casei thymidylate synthase. The Km values decreased from 24 microM for the monoglutamate to 1.8 microM for the triglutamate. Addition of residues 4, 5, 6, and 7 did not decrease the Km further. When monoglutamate and polyglutamate substrates were simultaneously incubated with the enzyme, the rate observed was characteristic of the polyglutamate even when the monoglutamate concentration was 44 times that of the polyglutamate. Iodoacetamide treatment inhibited the enzyme to the same extent with monoglutamate and polyglutamate substrates. Addition of 0.3 M NaCl doubled the rate obtained with the polyglutamate substrate whereas the rate with the monoglutamate was inhibited 25%. MgCl2 stimulated the reaction only 10% with the polyglutamate substrate compared with 80% stimulation obtained with the monoglutamate. Inhibition by fluorodeoxyuridylate was similar with both mono- and polyglutamate substrates; however, with the phosphonate derivative of fluorodeoxyuridine, the polyglutamate substrate enhanced inhibition 5- to 8-fold.

Folate analogues altered in the C9-N10 bridge region. 16. Synthesis and antifolate activity of 11-thiohomoaminopterin

The synthesis of 11-thiohomoaminopterin (1), which is a close analogue of 11-thiohomofolic acid (2), has been carried out by modification of the Boon-Leigh procedure. Treatment of 1-chloro-4-[p-(carbomethoxy)thiopenoxy]-2-butanone (5) with sodium azide gave 1-azido-4-[p-(carbomethoxy)thiophenoxy]-2-butanone (6). After protection of the carbonyl group of 6, the product 7 was catalytically hydrogenated to 1-amino-4-[p-(carbomethoxy)thiophenoxy]-2-butanone ketal (3). Reaction of 32 with 6-chloro-2,4-diaminmo-5-nitropyrimidine gave the desired pyrimidine intermediate, which was elaborated to 4-amino-4-deoxy-11-thiohomopteroic acid (20) by standard procedures. Alternately, 1-azido-4-[p-(carbomethoxy)thiophenoxy]-2-butanone ketal (7) was hydrolyzed to the corresponding acid (8) and coupled with diethyl L-glutamate to obtain diethyl N-[p-(1-azido-2-oxo-4-thiobutanoyl)benzoyl]-L-glutamate ketal (10), which was used for the large-scale preparation of 11-thiohomoaminopterin (1). Although 11-thiohomoaminopterin showed antifolate activity against two folate-requiring microorganisms and inhibited Lactobacillus casei dihydrogolate reductase, it did not exhibit any antitumor activity against L-1210 lymphoid leukemia in mice at a maximum dose of 48 mg/kg.

Folate analogues altered in the C9-N10 bridge region: 11-thiohomofolic acid

The synthesis of 11-thiohomofolic acid (2) has been accomplished by an unambiguous procedure. Reaction of 1-chloro-4-[p-(carbomethoxy)thiophenoxy]-2-butanone (10) with hydroxylamine under carefully controlled conditions gave the corresponding oxime 33. Conversion of this oxime to 1-phthalimido-4-[p-(carbomethoxy)thiophenoxy]-2-butanone oxime (4) was carried out by its reaction with potassium phthalimide using crown 18 ether as a catalyst. Hydrazinolysis of compound 4 gave 1-amino-4-[p-(carbomethoxy)thiophenoxy]-2-butanone oxime (5), which was used for the construction of the title compound 2 by modification of the Boon and Leigh procedure. An alternate synthesis utilizing 1-hydroxy-4-[p-(carbomethoxy)thiophenoxy]-2-butanone (11) and 4-hydroxy-2,5,6-triaminopyrimidine has also been carried out. Compound 2 did not exhibit any antifolate activity against Lactobacillus casei or Streptococcus faecium. The dithionite reduction product, 7,8-dihydro-11-thiohomofolic acid, was able to function as a substrate of L. casei dihydrofolate reductase. The catalytic reduction product of 2, consisting of a mixture of diastereomers, exhibited powerful antifolate activity against both these organisms.

Methotrexate analogues. 12. Synthesis and biological properties of some aza homologues

Methotrexate analogues, in which an additional nitrogen atom is inserted between the phenyl ring and the carbonyl group of the side chain, were prepared by photochemical methods. The compounds were less inhibitory toward dihydrofolate reductase and thymidylate synthetase derived from Lactobacillus casei than was methotrexate. They were also less cytotoxic against human lymphoblastic leukemia cells (CCRF-CEM). In vivo against L-1210 leukemia in mice, the aza homologue of methotrexate showed significant antitumor activity (%ILS = 55%) compared to methotrexate (%ILS = 88%).

Synthesis of aza homologues of folic acid

Folic acid analogues containing an additional nitrogen atom between the phenyl ring and the carbonyl group of the side chain were synthesized. None of the compounds showed significant inhibitory activity against human lymphoblastic leukemia cells (CCRF-CEM) in culture or against Lactobacillus casei (ATCC 7469) growth. Against L1210 leukemia in mice, the aza homologue of folic acid, 4, and the aspartic acid analogue, 14, showed no increase in life span over control animals. These compounds were more toxic in vivo than the corresponding methotrexate analogues. Compound 4 supported the growth of Streptococcus faecium (ATCC 8043), and its tetrahydro derivative supported the growth of Pediococcus cerevisiae (ATCC 8081). These results strongly suggest that 4 can substitute for folate derivatives as cofactors for serine transhydroxymethylase, thymidylate synthetase, and dihydrofolate reductase.

Phosphonate analogue of 2'-deoxy-5-fluorouridylic acid

The phosphonate analogue (6) of 2'-deoxy-5-fluorouridylic acid has been prepared via a Pfitzner--Moffatt oxidation and Witting reaction. This compound was found to inhibit thymidylate synthetase from three sources and to be cytotoxic to H.Ep.-2 cells in culture.

Nalpha-(pteroyltetra (gamma-glutamyl))-lysine as a ligand for the purification of thymidylate synthetase by affinity chromatography

Nalpha-(pteroyltetra (gamma-glutamyl))-lysine Sepharose was synthesized and shown to be a stable high capacity affinity matrix capable of bringing about the purification of Lactobacillus casei thymidylate synthetase to maximum specific activity from crude extracts in high yield. Under conditions optimal for binding of thymidylate synthetase, dihydrofolate reductase was not bound.

Diastereoisomers of 5,10-methylene-5,6,7,8-tetrahydropteroyl-D-glutamic acid

The diastereoisomers of 5,10-methylene 5,6,7,8-tetrahydropteroyl-D-glutamate were resolved and tested as substrates and inhibitors of Lactobacillus casei thymidylate synthetase. No activity was observed. The compounds were neither growth factors nor inhibitors for Lactobacillus casei, Streptococcus faecium, or Pediococcus cerevisiae. 7,8-Dihydropteroyl-D-glutamate is 50% as active as 7,8-dihydropteroyl-L-glutamate (dihydrofolate) as a substrate for L. casei dihydrofolate reductase.

Aspects of the reversal of methotrexate toxicity in rodents

An ip injection of a Lactobacillus casei dihydrofolate reductase preparation into rats and mice given a single lethal dose of methotrexate (MTX) caused a marked lowering of free MTX in the blood. Alternatives to citrovorum factor as agents for reversing MTX toxicity were explored in mice. dl, L-5-Methyltetrahydrofolate, dl,L-5,10-methylenetetrahydrofolate, l,L-5,10-methylenetetrahydrofolate, and dihydrofolate were also effective MTX antagonists; d,L-5,10-methylenetetrahydrofolate was inert.

Synthesis and biological activity of 10-thia-10-deaza analogs of folic acid, pteroic acid, and related compounds

The 10-thia analogs of pteroic acid, folic acid, their esters, and their 4-amino analogs were synthesized through a reaction sequence involving, as a key step, the condensation of 2-amino-3-cyano-5-chloromethylpyrazine with appropriately substituted thiols. The abilities of the products to inhibit the growth of methotrexate (MTX)-sensitive and MTX-resistant microorganisms were investigated as were their abilities to inhibit dihydrofolic acid reductase and thymidylic acid synthetase. Several compounds had high activity.