Effect of tetrahydrouridine and deoxytetrahydrouridine on the interaction between 2'-deoxycytidine and 1-beta-D-arabinofuranosylcytosine in human leukemia cells

Modulation of gemcitabine (2',2'-difluoro-2'-deoxycytidine) pharmacokinetics, metabolism, and bioavailability in mice by 3,4,5,6-tetrahydrouridine

PURPOSE

In vivo, 2',2'-difluoro-2'-deoxycytidine (dFdC) is rapidly inactivated by gut and liver cytidine deaminase (CD) to 2',2'-difluoro-2'-deoxyuridine (dFdU). Consequently, dFdC has poor oral bioavailability and is administered i.v., with associated costs and limitations in administration schedules. 3,4,5,6-Tetrahydrouridine (THU) is a potent CD inhibitor with a 20% oral bioavailability. We investigated the ability of THU to decrease elimination and first-pass effect by CD, thereby enabling oral dosing of dFdC.

EXPERIMENTAL DESIGN

A liquid chromatography-tandem mass spectrometry assay was developed for plasma dFdC and dFdU. Mice were dosed with 100 mg/kg dFdC i.v. or orally with or without 100 mg/kg THU i.v. or orally. At specified times between 5 and 1,440 min, mice (n = 3) were euthanized. dFdC, dFdU, and THU concentrations were quantitated in plasma and urine.

RESULTS

THU i.v. and orally produced concentrations >4 microg/mL for 3 and 2 h, respectively, whereas concentrations of >1 microg/mL have been associated with near-complete inhibition of CD in vitro. THU i.v. decreased plasma dFdU concentrations but had no effect on dFdC plasma area under the plasma concentration versus time curve after i.v. dFdC dosing. Both THU i.v. and orally substantially increased oral bioavailability of dFdC. Absorption of dFdC orally was 59%, but only 10% passed liver and gut CD and eventually reached the systemic circulation. Coadministration of THU orally increased dFdC oral bioavailability from 10% to 40%.

CONCLUSIONS

Coadministration of THU enables oral dosing of dFdC and warrants clinical testing. Oral dFdC treatment would be easier and cheaper, potentially prolong dFdC exposure, and enable exploration of administration schedules considered impractical by the i.v. route.

Plasma pharmacokinetics and oral bioavailability of 3,4,5,6-tetrahydrouridine, a cytidine deaminase inhibitor, in mice

Cytidine analogues such as cytosine arabinoside, gemcitabine, decitabine, 5-azacytidine, 5-fluoro-2'-deoxycytidine and 5-chloro-2'-deoxycytidine undergo rapid catabolism by cytidine deaminase (CD). 3,4,5,6-tetrahydrouridine (THU) is a potent CD inhibitor that has been applied preclinically and clinically as a modulator of cytidine analogue metabolism. However, THU pharmacokinetics has not been fully characterized, which has impaired the optimal preclinical evaluation and clinical use of THU. Therefore, we characterized the THU pharmacokinetics and bioavailability in mice. Mice were dosed with THU iv (100 mg/kg) or po (30, 100, or 300 mg/kg). Plasma and urine THU concentrations were quantitated with a validated LC-MS/MS assay. Plasma pharmacokinetic parameters were calculated compartmentally and non-compartmentally. THU, at 100 mg/kg iv had a 73 min terminal half-life and produced plasma THU concentrations >1 microg/ml, the concentration shown to effectively block deamination, for 4 h. Clearance was 9.1 ml/min/kg, and the distribution volume was 0.95 l/kg. Renal excretion accounted for 36-55% of the THU dose. A three-compartment model fit the iv THU data best. THU, at 100 mg/kg po, produced a concentration versus time profile with a plateau of approximately 10 mug/ml from 0.5-3 h, followed by a decline with an 85 min half-life. The oral bioavailability of THU was approximately 20%. The 20% oral bioavailability of THU is sufficient to produce and sustain, for several hours, plasma concentrations that inhibit CD. This suggests the feasibility of using THU to decrease elimination and first-pass metabolism of cytidine analogues by CD. THU pharmacokinetics are now being evaluated in humans.

Epimer interconversion, isomerization, and hydrolysis of tetrahydrouridine: Implications for cytidine deaminase inhibition

Tetrahydrouridine (THU) is an inhibitor of cytidine deaminase (CDA), the enzyme responsible for the deactivation of ara-C and other cytidine analogues in vivo, and therefore is capable of improving the therapeutic efficacy of these antitumor agents. In aqueous solution formulations, THU exists as a mixture of epimers differing in stereochemistry of the 4-OH substituent. The aims of this study were to investigate the interconversion kinetics of the epimers of THU, the CDA inhibitory effects of these epimers, and the stability and degradation mechanisms of THU epimer mixtures in aqueous solution with the ultimate goal of developing optimal conditions for a parenteral formulation of THU. A stability indicating HPLC assay utilizing a derivatized beta-cyclodextrin column was developed to separate the two epimers of THU and to monitor their reversible isomerization to their beta-ribopyranosyl counterparts and their hydrolysis to form N-glycosidic bond cleavage products. MS and one- and two-dimensional (1)H- and (13)C-NMR measurements were conducted to identify THU epimers and degradation products and to quantitatively model the degradation kinetics. The interconversion reaction between the two THU epimers is acid catalyzed with a first-order rate constant for conversion of epimer 1(1) to epimer 1(2) of (7.4 +/- 0.3) x 10(-3) h(-1) and an equilibrium constant ([1(2)]/[1(1)] of 1.7 +/- 0.1 at pH 7.4 and 25 degrees C. Epimer interconversion was therefore sufficiently slow at pH 7.4 to allow the isolation of each and evaluation of their CDA inhibitory activities utilizing 1% (w/v) mouse kidney homogenates as a source for cytidine deaminase and cytidine as a substrate. Inhibition constants for the two THU epimers (1(1) and 1(2)) were determined to be 8 +/- 1 x 10(-7) M and 6.2 +/- 0.2 x 10(-8) M, respectively. Studies at elevated temperature suggested that THU degradation from epimer mixtures is biphasic with the initial rate of disappearance being acid catalyzed and first order in initial THU concentration, thus ruling out dimerization as a potential reaction mechanism. NMR/MS analyses revealed that the major degradation products included the beta-ribopyranosyl THU isomers (two epimers), the reduced pyrimidinone base (tetrahydrouracil), and various anomers of D-ribose formed through N-glycosidic bond cleavage, and the products of subsequent reactions of the base. Kinetic modeling of the data obtained from both HPLC and NMR measurements indicated that in an acidic solution THU beta-ribofuranosyl --> beta-ribopyranosyl isomerization is a rapid equilibrium reaction, which proceeds through an intermediate observable in 1H-NMR, and is followed by slower N-glycosidic bond hydrolysis. All the reactions between THU, its ribopyranosyl isomers, the intermediate, and the base are acid catalyzed and appear to proceed through the same sugar ring-opened intermediate (carbinolamine), consistent with previous literature.

Cyclopentenyl cytosine increases the phosphorylation and incorporation into DNA of 1-beta-D-arabinofuranosyl cytosine in a human T-lymphoblastic cell line

The cytotoxic effect of 1-beta-D-arabinofuranosyl cytosine (araC) depends on the intracellular phosphorylation into its active compound araCTP, on the degree of degradation of araCTP and on the incorporation of araCTP into DNA. Deoxycytidine triphosphate (dCTP) inhibits the phosphorylation of araC (by feedback inhibition of the enzyme deoxycytidine kinase) and the incorporation of araCTP into DNA (by competition for DNA polymerase). In a T-lymphoblastic cell line, we studied whether the cytotoxicity of araC (2 nM-50 microM) could be enhanced by decreasing the concentration of dCTP, using the nucleoside-analogue cyclopentenyl cytosine (CPEC), an inhibitor of the enzyme CTP synthetase. Preincubation of the cells with CPEC (100-1,600 nM) for 2 hr increased the concentrations of araCMP 1.6-9.5-fold, which was significant for each concentration of CPEC used. The concentration of araCDP remained low, whereas the concentration of araCTP changed depending on the concentration of araC used. With 2-15 microm of araC and a preincubation with 400 nM of CPEC, the araCTP concentration increased by 4-15% (not significant), and the total amount of araC nucleotides increased significantly by 21-45%. When using a concentration of araC of 2 nM after a preincubation with CPEC of 100 nM, the concentration of araCMP increased by 60% (p = 0.015), whereas that of araCTP decreased by 10% (p = 0.008). This was compensated by an increase of 41% (p = 0.005) of araCTP incorporation into DNA, which represented 43% of all araC metabolites. Moreover, by performing pulse/chase experiments with 400 nM of CPEC and 2 microM of araC, the retention of cytosolic araCTP and the incorporated amount of araCTP into DNA were increased by CPEC. The modulation by CPEC of araC metabolism was accompanied by a synergistic increase of araC-induced apoptosis and by an additive effect on the araC-induced growth inhibition.

Ara-C: cellular and molecular pharmacology

The antimetabolite cytosine arabinoside (ara-C) represents a prototype of the nucleoside analog class of antineoplastic agents and remains one of the most effective drugs used in the treatment of acute leukemia as well as other hematopoietic malignancies. The ability of ara-C to kill neoplastic cells is regulated at three distinct but interrelated levels. First, the activity of ara-C depends on conversion to its lethal triphosphate derivative, ara-CTP, a process that is influenced by multiple factors, including nucleoside transport, phosphorylation, deamination, and levels of competing metabolites, particularly dCTP. Second, the antiproliferative and lethal effects of ara-C are linked to the ability of ara-CTP to interfere with one or more DNA polymerases as well as the degree to which it is incorporated into elongating DNA strands, leading to DNA fragmentation and chain termination. Finally, the fate of the cell is ultimately determined by whether a threshold level of ara-C-mediated DNA damage is exceeded, thereby inducing apoptosis, or programmed cell death. The latter process is influenced by components of various signal transduction pathways (e.g., PKC) and expression of oncogenes (e.g., bcl-2, c-Jun), perturbations in which may significantly alter ara-C sensitivity. A better understanding of these factors could eventually lead to the development of novel therapeutic strategies capable of overcoming ara-C resistance and improving therapeutic efficacy.

Improved synthesis of zebularine [1-(beta-D-ribofuranosyl)-dihydropyrimidin-2-one] nucleotides as inhibitors of human deoxycytidylate deaminase

The 2'-deoxy (2a) and 2'-ara-fluoro (3a) derivatives of zebularine [1-(beta-D-ribofuranosyl)-dihydropyrimidin-2-one, 1a] were phosphorylated in high yield to the 5'-nucleotides 2b and 3b, respectively, and characterized by HPLC, enzyme degradation, 1H, 13C and 31P NMR, and high resolution mass spectral analysis. Their inhibitory activity against partially purified MOLT-4 deoxycytidylate deaminase (dCMPD) in the presence of the allosteric effector deoxycytidine triphosphate (dCTP) and Mg+2 ion was examined. Compounds 2b and 3b inhibited dCMPD with Ki values of 2.1 x 10(-8) M and 1.2 x 10(-8) M, respectively. The parent nucleotide, zebularine monophosphate 1b was ineffective at concentrations > 100 mumol. The effect of the nucleosides, 1a-3a, as well as tetrahydrouridine (THU) and 2'-deoxy THU (dTHU), on the cellular production of DNA precursors was examined in human MOLT-4 peripheral lymphoblasts. It was shown that 1a, 2a and 3a all elevated intracellular dCTP and TTP levels in whole cells with the most powerful effect elicited by 1a. The 2'-fluoro derivative 3a was chemically phosphorylated much more cleanly and higher yield than 2a, without the formation of diphosphorylated by-products. This compound was found to be infinitely less sensitive to acid-catalyzed degradation than 2a. Since the substitution of fluorine for hydrogen had a slight potentiating effect on the dCMPD inhibitory activity while stabilizing the compound toward acid-catalyzed and enzymatic depyrimidination, compound 3b emerges as a very attractive tool for the pharmacological modulation of pyrimidine deaminase activity.

Modulation of the effect of 1-beta-D-arabinofuranosylcytosine by 6-mercaptopurine in L1210 cells

In L1210 cells incubated with 1-beta-D-arabinofuranosylcytosine (ara-C), 6-mercaptopurine (6-MP) significantly potentiated 1-beta-D-arabinofuranosylcytosine 5'-triphosphate (ara-CTP) accumulation and ara-C incorporation into DNA (ara-C/DNA). The cytotoxicity of these two drugs was assessed to be at least additive by clonogenic assay. 1-beta-D-Arabinofuranosyluracil (ara-U) level in a cell suspension was suppressed by 6-MP in a concentration-dependent fashion, though intracellular cytidine deaminase (CDD) activity was not affected by 6-MP. In addition, extracted CDD activity was not directly inhibited by 6-MP or by its intracellular metabolites in vitro. After preincubation in the presence or absence of 6-MP, the cell suspension was fractionated to obtain the spent medium and cell pellet. Then, each fraction was incubated with ara-C. Ara-U formation in the spent medium was found to increase conspicuously in relation to the time of preincubation in the control and it was suppressed by 6-MP pretreatment. Ara-U formation in the cell compartment increased slightly in relation to the time of preincubation in the control and substantially no suppression of ara-U formation was observed in spite of 6-MP pretreatment. In conclusion, intracellularly synthesized CDD was thought to be rapidly shed into the medium and the released CDD could play an important role in ara-C inactivation. 6-MP interrupted some step between synthesis and shedding of CDD, resulting in a decrease of the ara-C deamination in the medium and enhancement of its antileukemic effect.

Modulation of deoxycytidylate deaminase in intact human leukemia cells. Action of 2',2'-difluorodeoxycytidine

Cellular metabolism studies had demonstrated previously that low cellular concentrations of 2',2'-difluorodeoxycytidine (dFdC) nucleotides are eliminated by deoxycytidylate deaminase (dCMPD), whereas dCMPD activity is inhibited at high cellular dFdC nucleotide levels (Heinemann et al., Cancer Res 52: 533-539, 1992). An assay for measuring dCMPD activity in intact human leukemia cells has now been developed to permit investigations of the interactions of dFdC nucleotides with dCMPD in intact cells in which the regulated nature of this enzyme was not disrupted. Using [14C]dCyd as the substrate, radioactivity that accumulated in dTTP was quantitated after high-pressure liquid chromotography by a radioactive flow detector. The assay was first characterized using either the dCMPD inhibitor tetrahydrodeoxyuridine (H4dUrd) which directly inhibits dCMPD, or thymidine and 5-fluoro-2'-deoxyuridine (FdUrd) which indirectly inhibit and activate dCMPD, respectively, by affecting the cellular dCTP:dTTP value. Measured by this in situ assay, there was a strong correlation between dCMPD activity and dCTP:dTTP levels. Consistent with previous studies using partially purified enzyme, incubation of cells with dFdC resulted in a concentration-dependent inhibition of dCMPD in situ. The mechanism of modulation of dCMPD by dFdC, however, was clearly different from that of thymidine and FdUrd. In addition to the effect of dFdC on cellular dCTP:dTTP, our findings also suggested an additional inhibitory mechanism, possibly a direct interaction between dCMPD and dFdC 5'-triphosphate. Thus, results obtained using this direct assay of dCMPD in intact cells support the hypothesis that dCMPD is inhibited by nucleotides of dFdC.

Potent inhibitors for the deamination of cytosine arabinoside and 5-aza-2'-deoxycytidine by human cytidine deaminase

Deamination of the nucleoside analogues ARA-C and 5-AZA-CdR by CR deaminase results in a loss of antileukemic activity. To prevent the inactivation of these analogues, inhibitors of CR deaminase may prove to be useful agents. In the present study we investigated the effects of the deaminase inhibitors Zebularine, 5-F-Zebularine, and diazepinone riboside on the deamination of CR, ARA-C, and 5-AZA-CdR using highly purified human CR deaminase (EC 3.5.4.5). These inhibitors produced a competitive type of inhibition with each substrate, the potency of which followed the patterns diazepinone riboside greater than 5-F-Zebularine and THU greater than Zebularine. 5-AZA-CdR was more sensitive than ARA-C to the inhibition produced by these deaminase inhibitors. The inhibition constants for diazepinone riboside lay in the range of 5-15 nM, suggesting that this inhibitor could be an excellent candidate for use in combination chemotherapy with either ARA-C or 5-AZA-CdR in patients with leukemia.