Genotoxicity of tetrahydrofolic acid to hematopoietic stem and progenitor cells.

Metabolically reactive formaldehyde is a genotoxin and a carcinogen. Mice lacking the main formaldehyde-detoxifying gene Adh5 combined with the loss of the Fanconi anemia (FA) DNA repair pathway rapidly succumbed to bone marrow failure (BMF) primarily due to the extensive ablation of the hematopoietic stem cell (HSC) pool. However, the mechanism by which formaldehyde mediates these toxic effects is still unknown. We uncover a detrimental role of tetrahydrofolic acid (THF) in cells lacking Adh5 or the FA repair pathway. We show that Adh5- or FA-deficient cells are hypersensitive to formaldehyde and to THF, presenting DNA damage and genome instability. THF cytotoxicity involved imbalance of the nucleotide pool by deregulation of the thymidylate synthase (TYMS) enzyme, which stalled replication forks. In mice, THF exposure had widespread effects on hematopoiesis, affecting the frequency and the viability of myeloid- and lymphoid-committed precursor cells. Moreover, the hematopoietic stem and progenitor cells (HSPC) showed genomic instability, reduced colony-forming capacity and increased frequency of cycling and apoptotic HSCs upon THF exposure. Overall, our data reveal that the physiological pool of THF and formaldehyde challenge the stability of the genome of HSPCs that might lead to blood disorders.

Hyperhomocysteinemia in Alzheimer's disease: the hen and the egg?

Hyperhomocysteinemia is associated with Alzheimer's disease (AD). The causality of this association is controversial. In this study we tested the effect of a hyperhomocysteinemia-inducing diet in the ArcAß transgenic AD mouse model. At 14 months of age, the hyperhomocysteinemia-inducing diet yielded higher plasma homocysteine levels in ArcAß mice compared with wild-type mice. Levels of plasma 5-methyltetrahydrofolate (5-MTHF) in 14-month-old mice on hyperhomocysteinemia-inducing diet were lower in the transgenic than in the wild-type mice. The folate derivate 5-MTHF serves as cofactor in homocysteine metabolism. Oxidative stress, which occurs in the course of disease in the ArcAß mice, consumes 5-MTHF. Thus, the transgenic mice may plausibly be more vulnerable to 5-MTHF-depleting effects of hyperhomocysteinemia and more vulnerable to hyperhomocysteinemia-inducing diet. This argues that AD pathology predisposes to hyperhomocysteinemia, i.e., as a facultative consequence of AD. However, we also observed that dietary-induced folate reduction and homocysteine increase was associated with an increase of plasma (young animals) and brain (older animals) amyloid-ß concentrations. This suggests that the hyperhomocysteinemia-inducing diet worsened pathology in the transgenic mice. In conclusion, this data may argue that folate reduction and hyperhomocysteinemia may contribute to neurodegeneration and may also be triggered by neurodegenerative processes, i.e., represent both a cause and a consequence of neurodegeneration. Such a vicious cycle may be breakable by dietary or supplementation strategies increasing the availability of 5-MTHF.

Cellular ATP depletion by LY309887 as a predictor of growth inhibition in human tumor cell lines.

The antifolate LY309887 is a specific glycinamide ribonucleotide formyltransferase inhibitor that blocks de novo purine synthesis and produces a depletion of purine nucleotides. The activity of LY309887 in six human tumor cell lines has been examined by growth inhibition and clonogenic assay after continuous exposure for three cell doubling times and by ATP depletion at 24 h. Three cell lines (CCRF-CEM, MCF7, and GC3) were sensitive to LY309887-induced growth inhibition (IC50: 5.6-8.1 nM), whereas the other cell lines (COR-L23, T-47D, and A549) were comparatively resistant (IC50: 36-55 nM). Sensitivity to LY309887 cytotoxicity was consistent with sensitivity to growth inhibition in four of five cell lines tested (MCF7/GC3: 0.01% survival and COR-L23/T-47D: 1-5% survival at 100 nM LY309887). LY309887-induced ATP depletion was measured by luciferase-based ATP assay and confirmed by high performance liquid chromatography measurements. There was a linear relationship between ATP depletion and growth inhibition when data were analyzed for all six cell lines (r2 = 0.93; P < 0.0001). Depletion of 24-h cellular ATP concentrations to < 1 mM was associated with both cell growth inhibition and cytotoxicity in all cell lines studied. In conclusion, cellular ATP depletion induced by LY309887 can be used to predict growth inhibition and cytotoxicity in human tumor cells.

The role of dietary folate in modulation of folate receptor expression, folylpolyglutamate synthetase activity and the efficacy and toxicity of lometrexol.

We have studied the molecular effects of a LFD in a murine model in order to better define the biochemical changes associated with folate deficiency. In addition, we have demonstrated the effect of a LFD on the pharmacokinetic profile and therapeutic activity and toxicity of lometrexol. These studies showed increased density of FR in tumors implanted in LFD mice and a decrease in the affinity of these receptors for folic acid. The results suggest that tumors can compensate for low folate bioavailability by up-regulation of a second FR with slightly lower affinity for folic acid. The higher density of this FR would provide greater capacity for garnering serum folate. FPGS activity increased in several tumors and liver and kidney of LFD mice. The increase in this enzyme activity would result in enhanced polyglutamation of folates and classical antifolates and thus increased cellular retention. Consistent with these changes in liver FPGS, mice injected i.v. with a single dose of lometrexol accumulated significantly more drug in liver and tumors of LFD animals compared to SD mice. Also, higher liver concentrations of lometrexol persisted longer in LFD mice. Polyglutamate analysis showed that longer polyglutamate forms appeared earlier in liver of LFD mice. After 7 days, longer polyglutamyl forms were recovered from liver of LFD mice (octa- and hepta-glutamyl lometrexol) compared to those on SD. A comparison of the efficacy and toxicity of lometrexol in C3H mammary tumor-bearing mice showed that in mice on LFD, lometrexol treatment produced a delayed toxicity with an LD50 of 0.1-0.3 mg/kg, a 3000-fold increase in lethality compared to SD mice. Supplementation of mice with folic acid restored anti-tumor activity and increased the therapeutic dose-range over which efficacy could be assessed. These studies support the use of folic acid supplementation for cancer patients treated with antifolate therapy in order to prevent the biochemical changes in FR and FPGS associated with folate deficiency, prevent delayed toxicity to GARFT inhibitors and enhance the therapeutic potential of this class of drugs.

Functional aspects of membrane folate receptors in human breast cancer cells with transport-related resistance to methotrexate.

Two methotrexate (MTX)-resistant human breast-cancer cell lines with impaired transport via the reduced folate carrier (RFC), one established in vitro (MTX(R)-ZR-75-1) and another inherently resistant (MDA-231), were adapted to grow in medium containing 2 nM folic acid. This induced the expression of previously undetectable membrane folate receptors (MFR) to levels of 8.2 and 2.3 pmol/10(7) cells, respectively. Polymerase chain reaction (PCR) quantitation revealed that MFR messenger-RNA levels of the isoform first described in human nasopharyngeal carcinoma KB cells (MFR-alpha) were increased in low-folate-adapted MTX(R)-ZR-75-1 cells, whereas placental transcripts (MFR-beta) coincided with MFR-alpha expression in low-folate (LF)-adapted MDA-231 cells. These cell lines were used to study the role of MFR in the uptake and growth-inhibitory effects of five different antifolates with varying affinities for MFR: N10-propargyl-5, 8-dideazafolic acid (CB3717) > 5,10-dideazatetra-hydrofolic acid (DDATHF) > N-5-[N-(3,4-dihydro-2-methyl-4-oxoquinazolin-6-methyl) -N-methyl-amino]-2-theonyl}-glutamic acid (ZD1694) >> MTX > edatrexate (EDX). Expression of MFR only slightly decreased the resistant phenotype for MTX, EDX, and ZD1694, suggesting that these drugs are not transported intracellularly to cytotoxic concentrations at these levels of MFR expression. On the other hand, both cell lines became from at least 180- to 400-fold more sensitive to growth inhibition by CB3717 and DDATHF, which may be correlated with their high affinity for MFR. These sensitivity/resistance profiles were largely similar following cell culture in medium containing 1 nM L-leucovorin, a folate with an affinity for MFR 10-fold lower than that of folic acid, the one exception being the increased sensitivity for ZD1694 seen in the LF-adapted cells with the highest level of MFR expression (MTX(R)-ZR-75-1). These results illustrate that the efficacy of MFR in mediating antifolate transport and cytotoxicity depends on their affinity for the folate antagonist, their degree of expression, and the levels of competing folates.

Antitumor activity of antifolate inhibitors of thymidylate and purine synthesis in human soft tissue sarcoma cell lines with intrinsic resistance to methotrexate.

We examined the antitumor effects of two antifolate inhibitors of thymidylate synthesis, N-(5-[N-(3, 4-dihydro-2-methyl-4-oxoquinazolin-6-ylmethyl)-N-methylamino ]-2-theno yl-L-glutamic acid (D1694; Tomudex) and 1843U89 as well as a folate-based inhibitor of purine synthesis, 5, 10-dideazatetrahydrofolic acid (DDATHF) on human soft tissue sarcoma cell lines having intrinsic resistance to methotrexate (MTX) due to impaired accumulation of polyglutamates of MTX (HS-16 and HS-42 cells) and to increased levels of dihydrofolate reductase and thymidylate synthase activity (HS-18 cells). Growth inhibition studies showed that ED50 values for D1694 and 1843U89 after a 24-h exposure were 11-19-fold and 22-222-fold lower, respectively, than those for MTX in HT-1080, a MTX-sensitive cell line, and the three MTX-resistant cell lines. In contrast, DDATHF was less cytotoxic than MTX in both the MTX-sensitive and the three resistant sarcoma cell lines. Uptake of D1694, 1843U89, or DDATHF was 2.5-4.5-fold higher than MTX in these sarcoma cell lines. However, D1694 and 1843U89, unlike MTX, accumulate in HS-16 and HS-42 cells as polyglutamate forms, reaching 70% of the total intracellular drug level after 24 h. DDATHF polyglutamates (9.4-24%) were less in the same cell lines. Much lower Km values for D1694 and 1843U89 as compared to MTX for folylpolyglutamate synthase were measured in the sarcoma cell lines, with Vmax values equal to or slightly higher than those obtained with MTX. D1694 and 1843U89 are significantly more cytotoxic than MTX in intrinsically MTX-resistant sarcoma cell lines as a result of extensive formation of polyglutamates. These two thymidylate synthase inhibitors should be evaluated in patients with soft tissue sarcomas.

The stereospecific cytotoxic potency of (6R) and (6S)-5,10- dideazatetrahydrofolate correlates with cellular folylpolyglutamate synthetase levels.

The de novo purine synthesis inhibitor 5,10-dideazatetrahydrofolate (DDATHF) has previously been shown to inhibit the growth of mouse L1210 and human CCRF-CEM leukemia cells. The present study demonstrates that both the 6R and 6S diastereomers of DDATHF are also cytotoxic to mammalian cells in a stereospecific manner. The cytotoxic potency of (6R)-DDATHF (also known as Lometrexol) towards different cell lines varied by approximately 14-fold and that of (6S)-DDATHF by as much as 156-fold. Compared to (6R)-DDATHF, (6S)-DDATHF was 6.0- and 7.2-fold more cytotoxic to human WiDr colon adenocarcinoma and Chinese hamster ovary (CHO) cells, respectively, and only 1.5- and 2.0-fold more cytotoxic to human T24 bladder carcinoma and mouse L1210 leukemia cells, respectively. However, compared to (6S)-DDATHF, (6R)-DDATHF was 8.7- and 6.9-fold more cytotoxic to C3H/10T1/2 clone 8 and clone 16 mouse fibroblasts, respectively. Weak inhibition of aminoimidazolecarboximide ribonucleotide formyltransferase (AICARFT, EC 2.1.2.3) appeared to have little role in the cytotoxicity of DDATHF diastereomers to WiDr cells during a 24-h exposure. Although glycinamide ribonucleotide formyltransferase (GARFT, EC 2.1.21) is the main biochemical target of DDATHF, DDATHF stereoisomers' cytotoxic potency showed no clear negative correlation with cellular GARFT levels. However, cellular folylpolyglutamate synthetase (FPGS, EC 6.3.2.17) levels correlated with cytotoxic potency in a positive manner. Surprisingly, two enzyme-dose/DDATHF LD90-response curves were observed for FPGS corresponding to differences in (6R) and (6S)-DDATHF cytotoxic potency among the six cell lines studied.

The reduced unsubstituted pteroate moiety is required for folate toxicity of cultured cerebellar granule neurons.

The proconvulsive and putative excitotoxic properties of folates led us to study folate toxicity in cultured rat cerebellar granule neurons. Exposure of these neurons to tetrahydrofolate (THF) (EC50 = 263 +/- 12 microM) and dihydrofolate (EC50 = 709 +/- 29 microM), but not to folate, formyl-THF, methyl-THF or methotrexate, resulted in a concentration-dependent neurotoxicity. THF and dihydrofolate toxicity were characterized by prominent neuronal swelling and early loss of neurite networking and could not be blocked by glutamate receptor antagonists, L-channel calcium blockers or anticonvulsants. Cleavage of THF into glutamate and tetrahydropteroate by glutamate carboxypeptidase in the presence of 1 microM MK-801 and 5 microM 6-cyano-7-nitroquinoxaline-2,3,dione shifted the concentration-toxicity curve to the left (EC50 = 47 +/- 4 microM). These results suggest (1) that the glutamate moiety is not necessary for folate toxicity, (2) that the unsubstituted reduced pteroate moiety is required for folate toxicity and (3) that unsubstituted reduced pteroates are previously unrecognized neurotoxins. Because there is no endogenous glutamate carboxypeptidase activity in cerebellar granule neuron cultures and no increase of glutamate levels in the medium of cultures exposed to THF or dihydrofolate, pteroate metabolites are unlikely to mediate the THF or dihydrofolate toxicity of cultured cerebellar granule neurons.

Functional effects of a naturally occurring amino acid substitution in human thymidylate synthase.

A major mechanism underlying the cytotoxicity of fluoropyrimidine analogs such as 5-fluorouracil and 5-fluoro-2'-deoxyuridine (FdUrd) occurs via the formation of 5-fluoro-2'-deoxyuridylate (FdUMP), a tight-binding inhibitor of thymidylate synthase (TS). Genetic variation in the structure of the TS molecule is an important determinant of response to fluoropyrimidines, because such variation may affect the binding of FdUMP to the enzyme. Previous studies have shown that the colonic tumor cell line HCT116 expresses two structurally distinct TS polypeptides that differ by the presence of tyrosine or histidine at residue 33. Compared with the Tyr-33 form, the His-33 form confers a 3-4-fold level of FdUrd resistance to cells; this was postulated to be derived from the reduced affinity of the enzyme for FdUMP and N5,N10-methylenetetrahydrofolate, ligands required for the formation of a stable inhibitory complex. In the present study, the Tyr-33 and His-33 forms have been purified to homogeneity, and their properties have been compared in detail. The Km values for dUMP and N5,N10-methylenetetrahydrofolate in the TS reaction were not significantly different between the two enzymes. In contrast, the catalytic efficiency (kcat) was 8-fold lower for the His-33 form. Kinetic and equilibrium binding measurements demonstrated that the dissociation constant for FdUMP binding into the ternary complex was 3-4-fold higher for the His-33 form; this was shown to be due to both a decrease in the rate of FdUMP association with the enzyme and an increase in the rate of FdUMP dissociation from the ternary complex. A TS form containing phenylalanine at residue 33 was created by site-directed mutagenesis and was shown to be very similar to the Tyr-33 enzyme with regard to kcat, pH/activity profile, and effect on FdUrd response. Thus, it is the presence of histidine at residue 33, rather than the absence of tyrosine, that is responsible for the alterations in catalytic and ligand-binding functions exhibited by the His-33 form. Possible mechanisms by which the histidine residue perturbs the structure of the TS active site are discussed.

Pharmacology and phase I trial of high-dose oral leucovorin plus 5-fluorouracil in children with refractory cancer: a report from the Children's Cancer Study Group.

Because of the synergy seen in adult trials when 5-fluorouracil is combined with leucovorin, we initiated a Phase I trial of this combination in children's refractory cancer. Leucovorin, an equal mixture of the (6R,S)-diastereoisomers, was administered p.o. for 6 consecutive days as 4 equal doses at 0, 1, 2, and 3 h totaling 500 mg/m2/day. 5-Fluorouracil was given daily on days 2 to 6 as an i.v. bolus immediately following the last dose of leucovorin. The leucovorin dose was held constant while the 5-fluorouracil dose was escalated in cohorts of patients from 300 mg/m2/day to its maximally tolerated dose. Thirty-five patients (19 with acute leukemia and 16 with solid tumors) were evaluable for toxicity. The maximally tolerated dose of FUra was 450 mg/m2/day for 5 treatments for patients with solid tumors and 650 mg/m2/day for 5 treatments for the children with leukemia. The dose-limiting toxicities were myelosuppression and stomatitis. Other side effects included transient, mild elevations of serum transaminases, mild nausea, vomiting, and diarrhea. The pharmacokinetics of high-dose p.o. leucovorin was studied in 23 children. There was considerable interpatient variability in the plasma concentrations of total bioactive folates (TBAF), (6S)-leucovorin, and (6S)-5-methyltetrahydrofolic acid. The maximum plasma concentration (Cmax) of TBAF was 821 +/- 97 (SE) nM, occurring at a median of 8 h; the Cmax of (6S)-leucovorin was 77 +/- 11 nM, occurring at 4 h. The TBAF concentration fell to 146 +/- 42 nM by 24 h. (6S)-5-Methyltetrahydrofolic acid accounted for 90 +/- 7% of the TBAF at the Cmax. The plasma concentration of (6R)-leucovorin, the unnatural isomer, was equal to that of TBAF. Thus, p.o. leucovorin reduced the 5-fold excess of (6R)-leucovorin over TBAF seen after i.v. doses. The relative amounts of the three major plasma species were approximately the same as in adults, even though the Cmax of each compound was lower.

Synergy between 5,10-dideaza-5,6,7,8-tetrahydrofolic acid and methotrexate in mice bearing L1210 tumors.

In vivo studies with 5,10-dideaza-5,6,7,8-tetrahydrofolic acid (DDATHF), an inhibitor of glycinamide ribonucleotide transformylase, indicate that at doses ranging from 2.5 to 10 mg/kg, it prolongs the survival of mice implanted with L1210 tumors. Lower doses of this agent have no effect. Parallel studies with methotrexate indicate that DDATHF is not as potent or as efficacious as methotrexate in this animal model. Low doses of DDATHF combined with low doses of methotrexate can cause a significant increase in the survival of L1210 tumor-bearing mice, suggesting synergism between these two antifolates.

Synthesis and biological properties of 5,10-dideaza-5,6,7,8-tetrahydrofolic acid.

The synthesis of the antifolate 5,10-dideaza-5,6,7,8-tetrahydrofolic acid (DDATHF) has been modified. It is prepared from 2-acetamido-6-formyl-4(3H)-pyrido[2,3-b]pyrimidone and [P-(N-[1,3-bis(ethoxycarbonyl)propan-1-yl]aminocarbonyl)] phenylmethyl]-triphenylphosphonium bromide. The synthesis proceeds via a sodium hydride promoted Wittig condensation in 1-methyl-2-pyrrolidone followed by catalytic reduction, mild base hydrolysis, and acid precipitation of the product. Synthesis of DDATHF is achieved in a total of seven steps from commercially available reagents. DDATHF is transported effectively into CCRF-CEM cells and inhibits growth of both human (CEM) and murine (L1210) cells in culture. Studies reported here support the view that methotrexate and DDATHF are transported via a shared transport mechanism.

Folic acid derivatives do not reproduce the neurotoxic effects of kainic acid on chicken retina.

Methyltetrahydrofolic acid, formyltetrahydrofolic acid, folic acid and dihydrofolic acid were injected intravitreally into the eyes of chickens. No short-term or long-term signs of neurotoxicity were observed, even when doses 100-200 times those at which kainic acid produces clear neurotoxic effects were injected. The folic acid derivatives neither inhibited nor potentiated the neurotoxic effects. Thus no support is given to the suggestion that folic acid and its derivatives may act as kainic acid agonists or antagonists, even though the receptors involved in kainic acid-induced neurotoxicity appear to be of the kainic acid rather than quisqualic acid or N-methyl-D-aspartic acid type.

Failure of folic acid derivatives to mimic the actions of kainic acid in brain in vitro or in vivo.

The ability of three derivatives of folic acid, N-5-methyltetrahydrofolic acid (MTHF), tetrahydrofolic acid (THF) and dihydrofolic acid (DHF) to mimic the actions of kainic acid (KA) in a number of in vitro and in vivo systems known to be sensitive to KA was examined. None of the three folate derivatives at 100 microM concentration significantly inhibited the specific binding of [3H]-KA to striatal membranes although 2 microM L-glutamate produced a 40% inhibition. None of the three folate derivatives stimulated the formation of cyclic GMP in cerebellar slices incubated in vitro although KA (0.5 mM) increased cyclic GMP levels by 2.5-fold. Whereas intrahippocampal injection of 2.3 nmoles of KA produces prolonged abnormalities of the EEG, limbic-type seizures and a characteristic pattern of neuronal degeneration in the hippocampal formation and related structures, intrahippocampal injection of a 100-fold greater dose of THF caused only minor and transient EEG abnormalities, no overt seizures and a highly restricted lesion. Whereas intrastriatal injection of 5.6 nmoles of KA caused a profound reduction in the specific activities of choline acetyltransferase and glutamate decarboxylase, markers for striatal intrinsic cholinergic and GABAergic neurons, 50-fold greater doses of MTHF did not affect either enzyme although this high dose of THF did cause a significant 33% reduction in choline acetyltransferase activity. These findings support the suggestion that THF may have weak neurotoxic effects in brain but indicate that the actions of this compound and the related MTHF and DHF are not mediated through KA-specific receptors.

Neurotoxicity of folates: implications for vitamin B12 deficiency and Huntington's chorea.

Recent work has shown that several folates interact with excitatory kainic acid receptors in the mammalian brain and appear to have agonist activity at these receptors. Since kainic acid is a potent neurotoxin it is possible that folates share this toxicity and that high levels of folates result in neuronal damage. Levels of methyltetrahydrofolate are markedly elevated in vitamin B12 deficiency, a disease associated with neuronal destruction. We propose that this destruction occurs as a result of a neurotoxic action of methyltetrahydrofolate. Injection of kainic acid into the basal ganglia of experimental animals produces a pattern of damage similar to that found in patients dying of Huntington's chorea. It is possible that the underlying defect in this disease resides in the pathways of folate metabolism such that a neurotoxic excess of folates accumulates in the central nervous system. Such a disease process might be arrested by antifolate drugs.