Pharmacokinetics of the thymidine analog 2'-fluoro-5-methyl-1-beta-D-arabinofuranosyluracil (FMAU) in tumor-bearing rats.

The thymidine analog 2'-fluoro-5-methyl-1-beta-D-arabinofuranosyluracil (FMAU) is incorporated into DNA and is resistant to catabolism. We performed pharmacokinetic measurements with [(14)C]FMAU and PET studies with [(11)C]FMAU using rats bearing several different syngeneic tumors. Among normal tissues, FMAU uptake reflected relative cell turnover rates. Among tumors, the highest uptake occurred in a rapidly growing colon carcinoma, but was similarly low in both rapidly and slowly growing prostate tumors. FMAU was not catabolized and was rapidly incorporated into DNA by small intestine and colon tumors. Results indicate that FMAU may be useful for imaging tissue DNA synthesis, although tumor uptake was modest and not well correlated with growth rate among the models examined.

Synthesis and antiviral evaluation of halogenated beta-D- and -L-erythrofuranosylbenzimidazoles.

A series of 2-substituted benzimidazole D- and L-erythrofuranosyl nucleosides were synthesized and tested for activity against herpesviruses and for cytotoxicity. The D-nucleosides 2,5, 6-trichloro-1-(beta-D-erythrofuranosyl)benzimidazole (8a) and 2-bromo-5,6-dichloro-1-(beta-D-erythrofuranosyl)benzimidazole (8b) were prepared by coupling 1,2,3-tri-O-acetyl-beta-D-erythrofuranose (D-6) with the appropriate benzimidazole, followed by removal of the acetyl protecting groups. The 2-isopropylamino (9), 2-cyclopropylamino (10), and 2-mercaptobenzyl (11) derivatives were synthesized by nucleophilic displacements of the C-2 chlorine in the benzimidazole moiety of 8a. The D-nucleoside 4-bromo-5, 6-dichloro-2-isopropylamino-1-(beta-D-erythrofuranosyl)benzimid azo le (17) was prepared by coupling D-6 with the appropriate benzimidazole. The L-erythrofuranosyl derivatives, 5, 6-dichloro-2-isopropylamino-1-(beta-L-erythrofuranosyl)benzimid azo le (21a), its 2-cyclopropylamino analogue (21b), and the 2-isopropylamino analogue (25), were prepared by coupling L-6 with the appropriate benzimidazole. Several of these new derivatives had very good activity against HCMV in plaque and yield reduction assays (IC(50) = 0.05-19 microM against the Towne strain of HCMV) and DNA hybridization assays. Very little activity was observed against other herpesviruses. This pattern is similar to the antiviral activity profile observed for the corresponding ribofuranosides 2,5, 6-trichloro-1-(beta-D-ribofuranosyl)benzimidazole (4a), its 2-bromo analogue (4b), and the 2-cyclopropylamino analogue (4c). In comparison, 8a was 15-fold more active against HCMV than 4a, and 8b was 4-fold more active against HCMV than 4b. The 5, 6-dichloro-2-isopropylamino-1-(beta-L-erythrofuranosyl)benzimid azo le (21a) was less active than 4c, which is now in clinical trials for HCMV infection. Both 8a,b had comparable HCMV activity to 4c. Mode of action studies with the D-erythrose analogues established that 8b acted by inhibition of viral DNA processing whereas 9 and 10 may act via a different mechanism. The lack of a 5'-hydroxymethyl group in all members of this series established that antiviral activity occurred without 5'-phosphorylation, a feature required for the activity of most nucleoside analogues.

Synthesis and evaluation of a series of 2'-deoxy analogues of the antiviral agent 5,6-dichloro-2-isopropylamino-1-(beta-L-ribofuranosyl)-1H-benzimidazole (1263W94).

A series of 2'-deoxy analogues of the antiviral agent 5,6-dichloro-2-isopropylamino-1-(beta-L-ribofuranosyl)-1H-benzimidazole (1263W94) were synthesized and evaluated for activity against human cytomegalovirus (HCMV) and for cytotoxicity. The 2-substituents in the benzimidazole moiety correspond to those that were used in the 1263W94 series. In general, as was found in the 1263W94 series, cyclic and branched alkylamino groups were needed for potent activity against HCMV. Three analogues 3a, 3b and 3d were as potent as 1263W94. Further evaluation of two analogues, 3a and 3b, suggested that these 2'-deoxy analogues may act via a novel mechanism of action similar to that of 1263W94. These 2'-deoxy analogues generally lacked cytotoxicity in vitro. Pharmacokinetic parameters in mice and protein binding properties of 3a were quite similar to 1263W94. However, the oral bioavailability of 3a was only half of that observed for 1263W94.

Design, synthesis, and antiviral evaluation of 2-deoxy-D-ribosides of substituted benzimidazoles as potential agents for human cytomegalovirus infections.

Stereoselective glycosylation of 2,5,6-trichlorobenzimidazole (1b), 2-bromo-5,6-dichlorobenzimidazole (1c), 5,6-dichlorobenzimidazole (1d), 5,6-dichlorobenzimidazole-2-thione (1e), 5,6-dichloro-2-(methylthio)benzimidazole (1f), 2-(benzylthio)-5,6-dichlorobenzimidazole (1g), and 2-chloro-5,6-dimethylbenzimidazole (1h) with 2-deoxy-3,5-di-O-p-toluoyl-alpha-D-erythro-pentofuranosyl chloride was achieved to give the desired beta nucleosides 2b-h. Subsequent deprotection afforded the corresponding free beta-D-2-deoxyribosides 3b-h. The 2-methoxy derivative 3i was synthesized by the treatment of 2b with methanolic sodium methoxide. Displacement of the 2-chloro group of 2b with lithium azide followed by a removal of the protective groups gave the 2-azido-5,6-dichlorobenzimidazole derivative (5). The 2-amino derivative (6) was obtained by hydrogenolysis of 5 over Raney nickel. 5,6-Dichloro-2-isopropylamino-1-(2-deoxy-beta-D-erythro- pentofuranosyl)benzimidazole (10) was prepared using 2'-deoxyuridine (7), N-deoxyribofuranosyl transferase and 1d followed by functionalization of the C2 position. Antiviral evaluation of target compounds established that compounds 3b and 3c were active against human cytomegalovirus (HCMV) at non-cytotoxic concentrations. The activity of these 2-deoxy ribosides, however, was less than the activity of the parent riboside, 2,5,6-trichloro-1-beta-D-ribofuranosylbenzimidazole (TCRB). Compared to TCRB, 3b and 3c were somewhat more cytotoxic and active against herpes simplex virus type 1. Compounds 3d-i with other substituents in the 2-position were inactive against both viruses and non-cytotoxic. In contrast, compounds with amine substituents in the 2-position (5, 6, 10) were active against HCMV albeit less so than TCRB. These results establish that 2-deoxy-D-ribosyl benzimidazoles are less active against the DNA virus HCMV than are the corresponding D-ribosides.

Development of novel benzimidazole riboside compounds for treatment of cytomegalovirus disease.

Benzimidazole ribosides are a new class of compounds with novel mechanisms of action against CMV. One compound in this series, BDCRB, inhibits CMV DNA processing by the UL89 gene product (putative terminase), but rapid metabolism to an inactive compound makes it unsuitable for development as a medicine. Another benzimidazole analogue, 1263W94, has many characteristics that make it an attractive candidate for development, including high potency in vitro, selectivity, good oral bioavailability, and lower toxicity than therapies currently available for treatment of CMV disease. Initial clinical trials have provided encouraging results, including good tolerability and linear pharmacokinetics over a wide dose range. Ongoing and planned clinical trials that will study the safety and tolerability of repeated dosing and evaluate the in vivo antiviral activity and ocular penetration of 1263W94, will help to determine the potential of this drug as an improved therapy for CMV disease.

Studies designed to increase the stability and antiviral activity (HCMV) of the active benzimidazole nucleoside, TCRB.

The potent activity of 2,5,6-trichloro-1-(beta-D-ribofuranosyl)benzimidazole (TCRB) against Human Cytomegalovirus with the concomitant low cellular toxicity at concentrations that inhibit viral growth prompted considerable interest in this research area. This interest was moderated by the pharmacokinetic studies of TCRB in rats and monkeys that revealed the instability of TCRB in vivo. These studies suggested that the instability was due to a cleavage of the glycosidic bond in vivo which released the heterocycle (2,5,6-trichlorobenzimidazole) into the bloodstream. This prompted us to initiate synthetic studies designed to increase the stability of the glycosidic bond of TCRB and BDCRB. Several synthetic approaches to address this and other problems are presented.

Design, synthesis, and antiviral activity of alpha-nucleosides: D- and L-isomers of lyxofuranosyl- and (5-deoxylyxofuranosyl)benzimidazoles.

Several 2-substituted alpha-D- and alpha-L-lyxofuranosyl and 5-deoxylyxofuranosyl derivatives of 5,6-dicholro-2-(isopropylamino)-1-(beta-L-ribofuranosyl) benzimidazole (1263W94) and 2,5,6-trichloro-1(beta-D-ribofuranosyl)benzimidazole (TCRB) were synthesized and evaluated for activity against two herpesviruses (HSV-1 and HCMV) and for their cytotoxicity against HFF and KB cells. Condensation of 1,2,3,5-tetra-O-acetyl-L-lyxofuranose (2a) with 2,5,6-trichlorobenzimidazole (1) yielded the alpha-nucleoside 3a. The 2-bromo derivative and 2-methylamino derivative were prepared by treatment of 3a with HBr followed by deprotection or from methylamine, respectively. Compound 3a was deprotected and the resultant nucleoside used to prepare the 2-cyclopropylamino and 2-isopropylamino derivatives. The 2-alkylthio nucleosides were prepared by condensing 2a with 5,6-dichlorobenzimidazole-2-thione followed by deprotection. Alkylation of this adduct gave the 2-methylthio and 2-benzylthio derivatives. Condensation of 5-deoxy-1,2,3-tri-O-acetyl-L-lyxofuranosyl, prepared from L-lyxose, with 1 or 2-bromo-5,6-dichlorobenzimidazole (15), followed by deprotection, gave the 2-chloro or 2-bromo-5'-deoxylyxo-furanosyl derivative, respectively. The cyclopropylamino derivative was prepared from the 2-chloro derivative. All D-isomers were prepared in an analogous fashion from D-lyxose. Either compounds were inactive against HSV-1 or weak activity was poorly separated from cytotoxicity. In contrast, the 2-halogen derivatives in both the alpha-lyxose and 5-deoxy-alpha-lyxose series were active against the Towne strain of HCMV. The 5-deoxy alpha-L analogues were the most active, IC50's = 0.2-0.4 microM, plaque assay; IC90's = 0.2-2 microM, yield reduction assay. All of the 2-isopropylamino or 2-cyclopropylamino derivatives were less active (IC50's = 60-100 microM, plaque assay; IC90's = 17-100 microM, yield reduction assay) and were not cytotoxic. The methylamino, thio, and methylthio derivatives were neither active nor cytotoxic. The benzylthio derivatives were weakly active, but this activity was poorly separated from cytotoxicity. The alpha-lyxose L-isomers were more active in a plaque assay against the AD169 strain of HCMV compared to the Towne strain, thereby providing additional evidence of antiviral specificity.

The crystal structure of Escherichia coli purine nucleoside phosphorylase: a comparison with the human enzyme reveals a conserved topology.

BACKGROUND: Purine nucleoside phosphorylase (PNP) from Escherichia coli is a hexameric enzyme that catalyzes the reversible phosphorolysis of 6-amino and 6-oxopurine (2'-deoxy)ribonucleosides to the free base and (2'-deoxy)ribose-1-phosphate. In contrast, human and bovine PNPs are trimeric and accept only 6-oxopurine nucleosides as substrates. The difference in the specificities of these two enzymes has been utilized in gene therapy treatments in which certain prodrugs are cleaved by E. coli PNP but not the human enzyme. The trimeric and hexameric PNPs show no similarity in amino acid sequence, even though they catalyze the same basic chemical reaction. Structural comparison of the active sites of mammalian and E. coli PNPs would provide an improved basis for the design of potential prodrugs that are specific for E. coli PNP. RESULTS: The crystal structure of E. coli PNP at 2.0 A resolution shows that the overall subunit topology and active-site location within the subunit are similar to those of the subunits from human PNP and E. coli uridine phosphorylase. Nevertheless, even though the overall geometry of the E. coli PNP active site is similar to human PNP, the active-site residues and subunit interactions are strikingly different. In E. coli PNP, the purine- and ribose-binding sites are generally hydrophobic, although a histidine residue from an adjacent subunit probably forms a hydrogen bond with a hydroxyl group of the sugar. The phosphate-binding site probably consists of two main-chain nitrogen atoms and three arginine residues. In addition, the active site in hexameric PNP is much more accessible than in trimeric PNP. CONCLUSIONS: The structures of human and E. coli PNP define two possible classes of nucleoside phosphorylase, and help to explain the differences in specificity and efficiency between trimeric and hexameric PNPs. This structural data may be useful in designing prodrugs that can be activated by E. coli PNP but not the human enzyme.

Synthesis and antiviral activity of 2'-deoxy-4'-thio purine nucleosides.

A series of 2'-deoxy-4'-thioribo purine nucleosides was prepared by trans-N-deoxyribosylase-catalyzed reaction of 2'-deoxy-4'-thiouridine with a variety of purine bases. This synthetic procedure is an improvement over methods previously used to prepare purine 4'-thio nucleosides. The compounds were tested against hepatitis B virus (HBV), human cytomegalovirus (HCMV), herpes simplex virus (HSV-1 and HSV-2), varicella zoster virus (VZV), and human immunodeficiency virus (HIV-1). Cytotoxicity was determined in a number of cell lines. Several compounds were extremely potent against HBV and HCMV and had moderate to severe cytotoxicity in vitro. The lead compound from the series, 2-amino-6-(cyclopropylamino)purine 2'-deoxy-4'-thioriboside, was the most potent and selective agent against HCMV and HBV replication in vitro; however, this analogue was nephrotoxic when tested in vivo.

Studies of inhibitor binding to Escherichia coli purine nucleoside phosphorylase using the transferred nuclear Overhauser effect and rotating-frame nuclear Overhauser enhancement.

NMR studies of the adenosine analog tubercidin have been carried out in the presence of Escherichia coli purine nucleoside phosphorylase (PNP) in order to characterize the conformation of the enzyme-complexed nucleoside. Although analysis of transferred NOE data at various enzyme/inhibitor ratios indicated a predominantly syn nucleoside conformation in the enzyme-complexed state, the results, particularly the 8(1') and 8(3') NOE interactions, were not quantitatively consistent with any single bound conformation. Dissociation rate constants for the tubercidin-PNP complex were determined based on analysis of chemical shift and line width data as a function of enzyme/inhibitor ratio, Carr-Purcell-Meiboom-Gill measurements of the transverse relaxation rate as a function of pulse rate, and T1 rho experiments as a function of the spin-lock field strength. Dissociation rate constants of 2100 s-1 at 20 degrees C and 1400 s-1 at 10 degrees C were determined using the latter two methods. These rates are sufficiently high to justify the validity of the transferred NOE method for an enzyme as large as PNP. The possible significance of spin diffusion was investigated by the use of the deuterated analog [2'-2H]tubercidin, for which many of the intraligand spin diffusion pathways are eliminated, and by performing a series of transferred ROE experiments. A comparison of data obtained using transferred NOE and ROE measurements provides a basis for separating direct and indirect relaxation pathways. Both approaches indicated that the relatively significant 8(3') NOE interaction was not dominated by spin diffusion. Furthermore, analysis of chemical shift and transverse relaxation data for the tubercidin H-2 resonance gave inconsistent results for the chemical shift of the bound species and was inconsistent with the assumption of a single, bound conformation. These results were interpreted in terms of a 2:1 ratio of a syn, 3'-exo:anti, 3'-endo geometry for bound tubercidin. Ligand competition experiments using 9-deazainosine show that all of the tubercidin TRNOE effects are reversed by addition of the second nucleoside, suggesting that the TRNOE data for tubercidin arise due to interactions at the active sites of PNP rather than as a consequence of nonspecific binding to the enzyme.

An enantiospecific synthesis of the human cytomegalovirus antiviral agent [(R)-3-((2-amino-1,6-dihydro-6-oxo-9H-purin-9-yl)methoxy)-4- hydroxybutyl]phosphonic acid.

The racemic isosteric phosphonate of ganciclovir monophosphate (BW2482U89, SR3745, [3-((2-amino-1,6-dihydro-6-oxo-9H-purin-9-yl)methoxy)-4- hydroxybutyl]phosphonic acid, 1) has potent and selective in vitro activity against human cytomegalovirus. An enantiospecific synthesis of the R-enantiomer of compound 1 starting from L-arabinose was developed. The synthesis involved (1) the preparation of a chiral acyclic moiety, (2) the coupling of the chiral acyclic moiety to diacetylguanine, (3) the introduction of phosphorus, and (4) the final deprotection. The R-enantiomer, which has stereochemistry analogous to the natural compound GMP, was tested against human cytomegalovirus and had an IC50 of 1.7 microM, which was approximately 2-fold more active than the racemic material. Both racemic and chiral compounds were less toxic than ganciclovir to bone marrow progenitor cells in an in vitro assay.

Influence of stereochemistry on antiviral activities and resistance profiles of dideoxycytidine nucleosides.

beta-L-2',3'-Dideoxycytidine (beta-L-ddC) and beta-L-5-fluoro-2',3'-dideoxycytidine (5-F-beta-L-ddC) were prepared and shown to have potent activity against human immunodeficiency virus type 1 (HIV-1) and hepatitis B virus (HBV). These compounds were compared with beta-D-2',3'-dideoxycytidine (beta-D-ddC) and two beta-L-oxathiolane nucleosides (beta-L-3'-thio-2',3'-dideoxycytidine and beta-L-5-fluoro-3'-thio-2',3'-dideoxycytidine) in terms of anti-HIV and anti-HBV activity, cytotoxicity, and development of HIV-1 resistance. Compared with beta-D-ddC, the beta-L-dideoxycytidine nucleosides had similar anti-HIV-1 activities, significantly greater anti-HBV activities, and decreased toxicities to a B-cell line, T-cell lines, and human bone marrow progenitor cells. HIV-1 strains resistant to beta-D-ddC were susceptible to the beta-L-ddC analogs. Compared with the beta-L-oxathiolane nucleosides, beta-L-ddC and 5-F-beta-L-ddC had similar anti-HIV-1 activities, decreased anti-HBV activities, and greater toxicities to B- and T-cell lines and bone marrow progenitor cells. There were similarities between the beta-L-ddC and beta-L-oxathiolane nucleosides in the rate of development and pattern of resistant HIV-1 selection. While the in vitro activity and cytotoxicity profiles of the beta-L-ddC nucleosides differed from those of the beta-D-ddC and beta-L-oxathiolane nucleosides, the data presented herein suggest that the sugar configuration of a dideoxynucleoside analog may play a major role in the rate of development and the pattern of HIV-1 resistance.

Varicella-zoster virus thymidine kinase. Characterization and substrate specificity.

The varicella-zoster virus (VZV) thymidine kinase (TK) EC 2.7.2.21) catalyzes the phosphorylation of many anti-VZV nucleosides. Purified, bacterially expressed VZV TK was characterized with regard to N-terminal amino acid sequence, pI value, pH optimum, metal ion requirement, phosphate donor and acceptor specificity, and inhibition by dTTP. Initial velocities of thymidine phosphorylation with variable MgATP concentrations fit a two-site model with apparent Km values for MgATP of 0.10 and 900 microM. dTTP was a noncompetitive inhibitor of thymidine phosphorylation but was competitive with MgATP. Phosphate donor and acceptor specificities of the bacterially expressed enzyme were indistinguishable from those of VZV TK purified from infected cells. Detailed studies of the nucleoside specificity with the bacterially expressed enzyme showed that, for a given sugar moiety, thymine nucleosides were the most efficient substrates followed by nucleosides of cytosine, uracil, adenine, and with some exceptions, guanine. For a given pyrimidine or purine (except guanine), 2'-deoxyribonucleosides were the most efficient substrates, followed by arabinosides, ribonucleosides, 2',3'-dideoxyribonucleosides, and the acyclic moiety of acyclovir.

Novel 6-alkoxypurine 2',3'-dideoxynucleosides as inhibitors of the cytopathic effect of the human immunodeficiency virus.

Twenty-one 6-alkoxypurine 2',3'-dideoxynucleosides were enzymatically synthesized with nucleoside phosphorylases purified from E. coli. Eighteen analogs exhibited anti-HIV-1 activity in MT4 cells. Two analogs, 6-(hexyloxy)-(17) and 6-(heptyloxy)-(18) purine 2',3'-dideoxynucleoside, were as potent as 2',3'-dideoxyinosine (ddI, didanosine, Videx). Although the antiviral activities of 17 and 18 were equivalent, 18 was more cytotoxic. Analogs containing less than four carbons in the 6-alkoxypurine substituent exhibited weak anti-HIV-1 activity. Analogs containing more than seven carbons in the 6-alkoxypurine substituent were too cytotoxic to be effectively evaluated for antiviral activity. Several 6-alkoxypurine 2',3'-dideoxynucleosides were evaluated for substrate activity with calf intestinal adenosine deaminase (ADA). Increasing the carbon chain length of the 6-alkoxypurine substituent decreased the rate of dealkoxylation. The best substrate in this series was 6-methoxypurine 2',3'-dideoxynucleaside (1); however, the rate of dealkoxylation of 100 microM 1 was 0.17% of the rate of deamination of 100 microM 2',3'-dideoxyadenosine. Compound 17, the most potent anti-HIV-1 analog, was not a substrate for ADA. EHNA (erthro-9-(2-hydroxy-3-nonyl)adenine), a potent inhibitor of ADA, had little effect on the antiviral activities of 17 and ddI. In contrast, coformycin, a potent inhibitor of both ADA and AMP deaminase, dramatically decreased the antiviral activity of 17, but not the antiviral activity of ddI. Thus, AMP deaminase appeared to be involved in the anabolism of 17. The pharmacokinetic profile of 17, the most promising analog in this series, was determined in the rat. At least seventeen metabolites of 17, including ddI, were detected in plasma samples. This analog also had poor oral bioavailability.

6-Dimethylamino-9-(beta-D-arabinofuranosyl)-9H-purine: pharmacokinetics and antiviral activity in simian varicella virus-infected monkeys.

6-Dimethylamino-9-(beta-D-arabinofuranosyl)-9H-purine (ara-DMAP) effectively prevented the development of rash and appreciably reduced viremia in simian varicella virus-infected monkeys. Doses of 100 and 50 mg/kg/day, administered orally, were highly effective. The lowest dose of 20 mg/kg/day was much less effective in preventing moderate viremia. However, the 20 mg/kg/day did prevent the development of rash in two of three monkeys. All three doses of ara-DMAP reduced liver infection as reflected by lower aspartate aminotransferase values in the sera of the African green monkeys. Orally administered ara-DMAP was rapidly absorbed. However, significant variation among individual monkeys in the AUC values, peak plasma levels, and plasma half-lives were observed.

Membrane permeation characteristics of 5'-modified thymidine analogs.

The membrane permeation characteristics of 5'-deoxythymidine (5'-ddThd) and 5'-azido-5'-deoxythymidine (5'-N3-5'-ddThd) were investigated in human erythrocytes, with an inhibitor-stop assay, at 20 degrees. Uptake of both nucleoside analogs occurred without metabolism, was nonconcentrative, and was partially inhibited by nucleosides or inhibitors of nucleoside transport at micromolar permeant concentrations. At higher permeant concentrations (greater than 1.0 mM), the influx rate of each analog was linearly dependent on concentration and insensitive to inhibition by nucleosides, inhibitors of nucleoside transport, and nucleobases. Kinetic analyses using nonlinear regression revealed that a saturable component of 5'-ddThd influx (Km = 200 microM) was competitively inhibited by thymidine (dThd) (Ki = 86 microM) or 5-iodo-2'-deoxyuridine (Ki = 84 microM). Similarly, a saturable component of 5'-N3-5'-ddThd influx (Km = 220 microM) was competitively inhibited by 2-chloroadenosine (Ki = 18 microM). The Ki values for these nucleoside inhibitors were similar to their reported Km values as permeants of the nucleoside transporter. Both 5'-ddThd and 5'-N3-5'-ddThd competitively inhibited the influx of dThd (Km = 60 microM), with similar Ki values (150 and 200 microM, respectively). We conclude that these two 5'-modified dThd analogs enter human erythrocytes both by nonfacilitated diffusion and by the nucleoside transporter. The absence of the 5'-hydroxyl group of dThd (5'-ddThd) resulted in a large increase in the octanol/buffer partition coefficient, in an ability to permeate human erythrocytes by nonfacilitated diffusion, and in a 3-fold diminished binding to the nucleoside transporter. The 5'-azido group (5'-N3-5'-ddThd) resulted in an additional 1.4-fold increase in the octanol/buffer partition coefficient and in a 2-fold increase in the rate of nonfacilitated diffusion.

Metabolism and pharmacokinetics of the anti-varicella-zoster virus agent 6-dimethylaminopurine arabinoside.

The metabolism of 6-dimethylaminopurine arabinoside (ara-DMAP), a potent inhibitor of varicella-zoster virus replication in vitro, was studied in rats and cynomolgus monkeys. Rats dosed intraperitoneally or orally with ara-DMAP excreted unchanged ara-DMAP and one major metabolite, 6-methylaminopurine arabinoside (ara-MAP), in the urine. They also excreted allantoin and small amounts (less than 4% of the dose each) of hypoxanthine arabinoside (ara-H) and adenine arabinoside (ara-A). The relative amount of each urinary metabolite excreted remained fairly constant for intraperitoneal ara-DMAP doses of 0.3 to 50 mg/kg of body weight. Rats pretreated with an inhibitor of microsomal N-demethylation, SKF-525-A, excreted more unchanged ara-DMAP and much less ara-MAP than did rats given ara-DMAP alone. Rats pretreated with the adenosine deaminase inhibitor deoxycoformycin excreted more ara-MAP and much less ara-H and allantoin. ara-MAP was shown to be a competitive alternative substrate inhibitor of adenosine deaminase (Ki = 16 microM). Rats given ara-DMAP intravenously rapidly converted it to ara-MAP and purine metabolism end products; however, ara-A generated from ara-DMAP had a half-life that was four times longer than that of ara-A given intravenously. In contrast to rats, cynomolgus monkeys dosed intravenously with ara-DMAP formed ara-H as the major plasma and urinary end metabolite. Rat liver microsomes demethylated ara-DMAP much more rapidly than human liver microsomes did. ara-DMAP is initially N-demethylated by microsomal enzymes to form ara-MAP. This metabolite is further metabolized by either adenosine deaminase, which removes methylamine to form ara-H, or by microsomal enzymes, which remove the second methyl group to form ara-A.

6-N-substituted derivatives of adenine arabinoside as selective inhibitors of varicella-zoster virus.

A series of 6-alkylaminopurine arabinosides were synthesized and found to inhibit varicella-zoster virus (VZV). The antiviral activities of these nucleosides were limited to VZV. None of the other viruses tested in the herpesvirus family were affected. The in vitro antiviral potencies of the 18 arabinosides correlated with their efficiencies as substrates of the VZV-encoded thymidine kinase in all but one case. The arabinosides of 6-methylaminopurine and 6-dimethylaminopurine were the most potent analogs, with 50% inhibitory concentrations against VZV of 3 and 1 microM, respectively. They were not cytotoxic to uninfected MRC-5 cells, human Detroit 98 cells, or mouse L cells (50% inhibitory concentration, greater than 100 microM). Neither 6-methylaminopurine arabinoside nor 6-dimethylaminopurine arabinoside was detectably phosphorylated by either adenosine kinase or 2'-deoxycytidine kinase. These two alkylaminopurine arabinosides were also resistant to deamination catalyzed by adenosine deaminase. The VZV-dependent phosphorylation of these nucleosides offers the possibility of a potent and highly selective therapy for VZV infection.

Metabolic disposition and pharmacokinetics of the antiviral agent 6-methoxypurine arabinoside in rats and monkeys.

The metabolism and pharmacokinetics of 6-methoxypurine arabinoside (ara-M), a potent and selective inhibitor of varicella-zoster virus, were investigated in rats and monkeys. In Long Evans rats, orally administered [8-14C]ara-M (10 mg/kg) was well absorbed but extensively metabolized to hypoxanthine arabinoside (ara-H), hypoxanthine, xanthine, uric acid, and allantoin. Only 4% of an oral dose was recovered in the urine as unchanged drug, compared with 40% of an intravenous dose, indicating significant presystemic metabolism. Pretreatment of rats with 1-aminobenzotriazole, an inhibitor of cytochrome P-450, did not alter this metabolism. Pretreatment with deoxycoformycin or erythro-9-(2-hydroxy-3-nonyl)adenine hydrochloride, inhibitors of adenosine deaminase, resulted in a marked decrease in ara-M metabolism, indicating that adenosine deaminase plays a major role in the biotransformation of ara-M. In cynomolgus monkeys, [8-14C]ara-M (10 mg/kg) administered intravenously or orally was extensively metabolized to ara-H. Several minor urinary metabolites were detected in both rats and monkeys. However, adenine arabinoside was not found in urine or plasma from either rats or monkeys after administration of ara-M, except for a very low level detected in the urine of rats pretreated with deoxycoformycin. The elimination half-lives of intravenously administered ara-M in rats and monkeys were 29 and 45 min, respectively. The corresponding half-lives of the primary metabolite, ara-H, were 44 min and 2.3 h. Plasma profiles of orally administered ara-M in both rats and monkeys demonstrated the poor oral bioavailability of this arabinoside. The results of these studies indicate that ara-M is not well suited for oral administration because of extensive presystemic metabolism.

6-Methoxypurine arabinoside as a selective and potent inhibitor of varicella-zoster virus.

Seven 6-alkoxypurine arabinosides were synthesized and evaluated for in vitro activity against varicella-zoster virus (VZV). The simplest of the series, 6-methoxypurine arabinoside (ara-M), was the most potent, with 50% inhibitory concentrations ranging from 0.5 to 3 microM against eight strains of VZV. This activity was selective. The ability of ara-M to inhibit the growth of a variety of human cell lines was at least 30-fold less (50% effective concentration, greater than 100 microM) than its ability to inhibit the virus. Enzyme studies suggested the molecular basis for these results. Of the seven 6-alkoxypurine arabinosides, ara-M was the most efficient substrate for VZV-encoded thymidine kinase as well as the most potent antiviral agent. In contrast, it was not detectably phosphorylated by any of the three major mammalian nucleoside kinases. Upon direct comparison, ara-M was appreciably more potent against VZV than either acyclovir or adenine arabinoside (ara-A). However, in the presence of an adenosine deaminase inhibitor, the arabinosides of adenine and 6-methoxypurine were equipotent but not equally selective; the adenine congener had a much less favorable in vitro chemotherapeutic index. Again, this result correlated with data from enzyme studies in that ara-A, unlike ara-M, was a substrate for two mammalian nucleoside kinases. Unlike acyclovir and ara-A, ara-M had no appreciable activity against other viruses of the herpes group. The potency and selectivity of ara-M as an anti-VZV agent in vitro justify its further study.