Antiviral Activity and Pharmacokinetics of 1-(2,3-Dideoxy-2-Fluoro-{beta}-L-Glyceropent-2-Enofuranosyl)Cytosine

1-(2,3-Dideoxy-2-fluoro-ß-L-glyceropent-2-enofuranosyl)cytosine(L-2'-Fd4C) is an L-nucleoside analogue with both anti-humanimmunodeficiency virus (HIV) and anti-hepatitis B virus (HBV)activity with median effective concentrations of 0.12 µMin peripheral blood mononuclear cells and 0.002 µM inHepG2-2.2.15 cells, respectively. The purpose of this studywas to examine the antihepadnavirus potency and pharmacokineticsof L-2'-Fd4C in vivo. HBV-transgenic mice treated intraperitoneallywith L-2'-Fd4C showed a reduction of HBV levels in their bloodcomparable to that produced by lamivudine. The pharmacokineticsof L-2'-Fd4C in rhesus monkeys was evaluated after intravenousand oral administration. The concentrations in plasma declinedin a biexponential manner after intravenous administration,with a long terminal-phase half-life of 5.02 h. The steady-statevolumes of distribution and systemic clearance were 1.09 liter· kg^-1 and 0.25 liter · h^-1 · kg^-1, respectively,with a renal clearance of 0.16 liter · h^-1 · kg^-1.The oral bioavailability was approximately 44%. About 53% ofthe compound administered intravenously and 19% of that administeredorally were recovered unchanged in the urine within the 24-hurine collection period, and no other metabolite was detected.The compound penetrated the central nervous system at concentrationsthat exceeded the median effective antiviral concentration againstHIV in cell cultures. Based upon these observations, furthertesting to develop this agent for treatment of HIV and HBV infectionsis warranted.

INTRODUCTION

Top

Introduction

Nucleoside analogues continue to play an important role in thetherapy of human immunodeficiency virus type 1 (HIV-1) and hepatitisB virus (HBV) infections. Eight of the 16 Food and Drug Administration-approvedanti-HIV drugs are nucleoside analogues, and one analogue, lamivudine(3TC), is used for the treatment of HIV- and HBV-infected patients.Adefovir (Hepsera), an acyclic nucleotide, was recently approvedfor the treatment of HBV, providing an alternative to 3TC. Theclinical use of the present anti-HIV agents is usually limitedby their toxicity (2, 6, 9, 12, 36) and by the emergence ofdrug-resistant viral strains during long-term therapy (7, 8,17, 26, 31). These deficiencies associated with the clinicallyuseful nucleoside analogues have stimulated the developmentof novel antiviral agents for the treatment of HIV and HBV infections.Pertinent structural modifications of the sugar and nucleosidebase moieties have produced antiviral agents with lower toxicitiesand greater efficacies. Among the nucleoside analogues approvedfor the treatment of HIV infections, five are in the ß-Dconfiguration, one is acyclic (Tenofovir; Viread), and one,3TC, is an L-nucleoside (27). The physical and chemical propertiesof L-nucleosides are identical to those of the D-enantiomersexcept for their optical rotation. Favorable characteristicsof L-nucleosides may include an antiviral activity of the activetriphosphate form that is comparable with, and sometimes greaterthan, that of the D-enantiomers, with increased metabolic stabilityand lower toxicity to uninfected cells (29, 34). However, someL-nucleosides such as ß-L-dioxolane-cytidine (Troxatyl)have demonstrated selective toxicity to cancer cells relativeto nontumor tissue (15). ß-L-Dioxolane-cytidine isundergoing phase III clinical trials as an anticancer agent(16). Therefore, L-nucleosides represent an important new approachin designing chemotherapeutic agents for the treatment of viralinfections and cancer.

1-(2,3-Dideoxy-2-fluoro-ß-L-glyceropent-2-enofuranosyl)cytosine(L-2'-Fd4C) is an L-nucleoside with both anti-HIV and anti-HBVactivity (21). In this study, we evaluated the anti-HBV activityof L-2'-Fd4C in the HepG2-2.2.15 cell system and its toxicityprofile in a number of cell lines. In vivo studies were thenperformed in HBV-transgenic mice, and the single-dose oral andintravenous (i.v.) pharmacokinetics were assessed in rhesusmonkeys.

MATERIALS AND METHODS

Top

Materials and Methods

Chemicals. L-2'-Fd4C (molecular weight, 227) (Fig. 1) was synthesized aspreviously described (21). The internal standard, ß-D-2',3'-didehydro-2',3'-dideoxy-5-fluorocytidine(D-D4FC, DPC-817, RVT, and Reverset), was synthesized as reportedpreviously (35). The chemical purity of each compound was verifiedby high-performance liquid chromatography (HPLC) and spectralanalyses as being greater than 98%. Acetonitrile (HPLC grade)and all the other chemicals (analytical grade) used were obtainedfrom Fisher Scientific (Fair Lawn, N.J.).

View larger version (10K):
[in this window]
[in a new window]
FIG. 1. Chemical structure of L-2'-Fd4C, D-D4FC, and 3TC.

Determination of anti-HBV activity in vitro. HepG2-2.2.15 cells harboring and secreting HBV particles wereused in the study (33). The cells were treated with concentrationsof L-2'-Fd4C ranging from 0.001 to 10 µM for a total periodof 9 days, and analysis of the HBV DNA was conducted as previouslydescribed (24, 30). Briefly, HBV DNA from the supernatants washarvested and Southern analysis was performed. The blots werehybridized to a ³²P-labeled HBV probe. The amounts of HBV DNAin the treated cells relative to that in the untreated controlswere measured by phosphorimaging (24, 30). Dose-response curveswere generated from these values and the 50 and 90% effectiveconcentrations (EC₅₀ and EC₉₀, respectively) were calculated(5).

Cytotoxicity of the nucleosides. The cytotoxicity was determined by using an methyltetrazoliumchloride dye protocol (32) with a panel of cells that includedperipheral blood mononuclear, CEM, HepG2, and Vero cells. Thismethod involved treating the cells with the nucleoside analoguesat concentrations up to 100 µM for 3 days. At the endof the treatment period, the cells were treated with methyltetrazoliumchloride dye and the metabolized formazan reduction productwas colorimetrically measured at 490 nm. The ratio of absorbancereadings relative to those of the cells not exposed to the drug(the controls) was then used as an indicator of cell survival.

Anti-HBV activity in vivo. HBV-transgenic mice were developed by using the infectious,head-to-tail dimer of HBV DNA as the transgene. HBV-transgenicmice replicate HBV stably, as evidenced by the consistent presenceof HBV DNA in their blood (20). Groups of 2- to 3-month-oldyoung adult mice of either gender were used for these studies.Blood was collected from the mice prior to treatment, and onlythose samples with viral DNA concentrations of >10⁴ virusgenome equivalents per ml of blood were used. The mice werematched for age and virus titer, bled prior to treatment, andgiven intraperitoneal injections of either 3TC or L-2'-Fd4C(n = 8 per group) at 100 mg/kg of body weight per day or equalvolumes of phosphate-buffered saline (PBS) (50 µl; n =6 per group) for 7 days and sacrificed 8 days later. In additionto the blood samples obtained prior to treatment, serial bloodsamples were collected on days 3, 5, 8, 12, and 15. All of themice were exsanguinated and sacrificed on day 15 after the startof treatment. The titers of viral DNA in the serum samples fromeach mouse were determined by quantitative PCR by using increasingamounts of an internal competitor template, as described previously(20). These experiments were performed with protocols reviewedand approved by the Institutional Animal Care and Use Committeeat Thomas Jefferson University.

To evaluate the results of the virus DNA titers, the valuesfor the cumulative areas under the curves of the virus loads(percentage of pretreatment value) versus the time between days0 and 5 (the last day the mice were given the drug for whichvirus loads were obtained) (AUC_vir) were calculated. Antiviralefficacies during treatment ( ${Delta}$ AUC) were calculated as follows: ${Delta}$ AUC = (average AUC_vir of the PBS-treated animals) - (the respectiveAUC_vir for each drug-treated mouse). The time needed for thevirus loads to rebound (after day 8) was noted as an indicatorof pharmacological persistence in the virus replication compartments.

Pharmacokinetic studies of rhesus monkeys. Two male rhesus monkeys (Macaca mulatta) weighing 4.6 and 5.6kg were used for the pharmacokinetic studies, and one servedas an untreated control. The animals were maintained at theYerkes National Primate Research Center at Emory University,which is fully accredited by the American Association for Accreditationof Laboratory Animal Care in accordance with guidelines establishedby the Animal Welfare Act and the Guide for the Care and Useof Laboratory Animals (23a) of the National Institutes of Health.The monkeys were given 33.3 mg of L-2'-Fd4C per kg of body weighti.v. (in 10 ml of PBS) so that comparisons could be made withresults of previously reported studies (1, 22, 25, 28). A washoutperiod of at least 4 weeks was used between doses. The oraldose, administered by gastric intubation, was 33.3 mg/kg ina total volume of 10 ml of water, followed by a another 3 mlof water. A control animal was given a sham dose of water withoutthe drug. The monkeys were maintained under anesthesia for 4h after dosing by using a mixture of ketamine HCl (60 mg) andtiletamine HCl plus zolazepam HCl (Telazol; 20 mg) intramuscularly.The animals were monitored for alertness and given additionalanesthesia (30 to 60 mg of ketamine HCl) as necessary. The animalswere maintained most of the time on their sides on a warm heatingpad and were covered with a blanket. Blood samples were takenprior to, and at 0.25, 0.5, 1, 1.5, 2, 3, 4, 6, 8, and 24 hafter, administration of the drug through the femoral vein whilethe animals were lying on their backs. Cerebrospinal fluid (CSF)samples were taken from all of the treated monkeys at 1 h afteradministration of the drug by cisternal or lumbar tap with a22-gauge needle. The monkeys were catheterized for urine collection.Urine samples were collected at 0 to 0.25, 0.25 to 0.5, 0.5to 1, 1 to 1.5, 1.5 to 2, 2 to 3, 3 to 4, 4 to 6, 6 to 8, and8 to 24 h after dosing. The plasma, CSF, and urine samples werefrozen at -70°C until they were assayed.

HPLC analysis of L-2'-Fd4C in monkey plasma, urine, and CSF samples. HPLC analysis was performed with an HP1050 HPLC system (Hewlett-PackardCo., Wilmington, Del.) equipped with a model 1050 photodiodearray detector and a model 1050 autosampler with a PhenomenexC₁₈ column (4.6 by 250 mm; particle diameter, 5 µm). Themobile phase was 3% acetonitrile in 50 mM phosphate buffer (pH7.0).

Preparation of standard working solutions. Working solutions of L-2'-Fd4C were prepared by using 3% acetonitrilein 50 mM potassium phosphate buffer at pH 7.0. Calibration plotsfor L-2'-Fd4C in plasma were prepared by adding working solutionsto blank plasma (plasma from animals that did not receive L-2'-Fd4C)at concentrations ranging from 0.075 to 50 µg/ml. Standardcurves for the analysis of urine and CSF samples were preparedby using urine and water, respectively, over the same concentrationrange. Unweighted linear regression was used to relate concentrationsto areas under the chromatogram curve.

Extraction procedure. For plasma samples, 500 µl of acetonitrile was added toa 50-µl plasma sample in a microcentrifuge tube, and then50 µl of D-D4FC (10 µg/ml) was added as an internalstandard before the solution was thoroughly mixed with a vortexmixer and centrifuged at 11,200 x g for 5 min. The supernatantwas evaporated to dryness, and the samples were reconstitutedwith 200 µl of 3% acetonitrile in potassium phosphatebuffer (pH 7) and injected onto the HPLC column. For CSF samples,20 µl of the internal standard (D-D4FC; 10 µg/ml)was added to 50 µl of the CSF samples and the mixturewas injected into the HPLC column for analysis. Similarly, urinesamples were diluted 50- to 100-fold, 20 µl of the sameinternal standard was added, and the mixture was analyzed byHPLC.

HPLC assay specifications. The HPLC method was validated in accordance with the Food andDrug Administration's Guidelines for Industry Bioanalytic MethodValidation protocol (33a). Separate quality control sampleswere used to assess recovery, precision, and accuracy (six samplesper concentration). The limit of quantification for the assaywas 0.075 µg of L-2'-Fd4C/ml. Low (0.075-µg/ml),medium (5-µg/ml), and high (50-µg/ml) concentrationswere selected for the validation procedure. The recovery ofanylate in the assay was taken as the detector response, obtainedfrom the known amount of anylate added to and extracted fromthe biological matrix (monkey plasma). The extraction recoveryof L-2'-Fd4C and the internal standard (D-D4FC) for each concentrationwas calculated by comparing the mean peak areas for six extractedplasma samples with those of the six unextracted samples containingthe same amount of nucleoside. The percent extraction recoverywas calculated from the equation 100 x (peak area for extractedsamples/peak area for unextracted_. samples). The recovery ofL-2'-Fd4C was greater than 90%. The intraday accuracy and precisionof the assay methodologies for plasma samples were determinedby assaying three samples per concentration on the same day.Samples were analyzed on three separate days to determine theinterday accuracy and precision of the assay methodologies.Precision was reported in terms of relative standard deviations,and accuracy was calculated by comparing the measured concentrationsto the known values. The intraday and interday precision valueswere less than 10%, and the intraday and interday accuracieswere greater than 90%. The nonweighted linear regression curveswere linear over a concentration range of 0.075 to 50 µg/ml.

Pharmacokinetic analysis. Since i.v. and oral plasma samples were available from the sameanimal, the i.v. and oral plasma data for each animal were fitsimultaneously to a two-compartment open pharmacokinetic model(14) by means of a nonlinear regression curve-fitting program(WinNonlin, version 1.5, 1997; Scientific Consulting, Inc.,Apex, N.C.) to avoid reporting two different sets of dispositionrate constants for the same animal (10). In addition to estimatingintercompartmental disposition rate constants, the fit alsoestimated the oral absorption rates (K_a) and the fraction ofthe drug absorbed (F) (10). There was no evidence of an appreciablelag time before the onset of oral absorption. Initial estimatesfor the model parameters were obtained by using the parametersfor similar compounds tested in rhesus monkeys. The adequacyof the model fit was assessed by examining the overall dispersionof data over the predicted curves and the predicted standarderrors of the fitted parameters. The fitted model parametersand the fit of the model to the plasma data are shown below(see Table 1 and Fig. 3). Accurate urine sample volumes werenot available for the oral-dose interval between 8 and 24 h.Therefore, the fraction of the total dose excreted in the urine(F_e) was not directly measured but was derived together withthe renal clearance (CL_R) by simultaneously fitting the urinaryexcretion rate (dX_u/dt) concomitantly to the model's predictedconcentrations in plasma (C_p) by using the equation dX_u/dt =CL_RC_p. In this relationship, C_p is the predicted concentrationin plasma based on the plasma data (described above) and coincidingwith the midpoint of the urine collection interval and X_u isthe amount of compound recovered from the urine during thatcollection interval. Fitting i.v. and oral sample data simultaneouslyassumes that there is a constant renal clearance for both routesof administration. This method is considered robust providedthat data are available over at least 1 half-life (t_1/2) anddoes not require measurement of the total amount excreted inthe urine, complete emptying of bladders within all sampledintervals, or urine collection over short intervals relativeto the t_1/2 of the drug (10). Peak concentrations (C_max) inplasma and the corresponding times (T_max) were reported as theobserved concentrations and times of maximal concentration followingoral administration. The AUC from time zero to infinity afterboth administrations (AUC_0-) was calculated as AUC_t plus C_t/ß.AUC_t was calculated by using the linear trapezoidal rule, andC_t/ß is the extrapolation to infinity. The fractionof oral dose absorbed was also estimated as the ratio of AUC_0-for the oral dose to the AUC_0-00 for the i.v. dose and was foundto be similar to the model-fitted value of F.

View this table:
[in this window]
[in a new window]
TABLE 1. Pharmacokinetic parameters of L-2'-Fd4C after oral and i.v. administration to rhesus monkeys^a

View larger version (11K):
[in this window]
[in a new window]
FIG. 3. Model fitted to pooled data (lines) and the observed concentrations of L-2'-Fd4C in the plasma of two rhesus monkeys (filled and open symbols) after the administration of 33.3 mg of L-2'-Fd4C per kg by the i.v. (triangles) and oral (circles) routes.

RESULTS

Top

Results

Inhibition of HBV DNA replication by the nucleosides. We evaluated a series of 2'-fluoro-2',3'-unsaturated nucleosidesin the "unnatural" L-configuration (21). Of the various nucleosidestested in this series, the most potent analogue against HBVwas L-2'-Fd4C, with an EC₅₀ of 0.002 µM in the HepG2-2.2.15system. L-2'-Fd4C, like the positive control 3TC,showed no toxicityin any of the various cell lines tested (peripheral blood mononuclear,Vero, CEM, and HepG2) up to 100 µM. L-2'-Fd4C had thesame EC₅₀ as 3TC, but the EC₉₀ of L-2'-Fd4C was 0.05 µMand that of 3TC was 0.01 µM.

Anti-HBV activity in HBV-transgenic mice. Among the mice injected with 3TC, two patterns of responsivenesswere observed (Fig. 2A). In half the mice, the level of HBVDNA became undetectable in serum during the period of treatmentand remained undetectable for up to a week after the end oftreatment. Among the remaining 3TC-treated animals, the levelof viral DNA became undetectable in serum during treatment butpartially rebounded within a week after the end of treatment.All of the PBS-injected mice had little or no change in viralDNA levels during the experiment (Fig. 2A). These observationsverify earlier results from this laboratory that showed thatHBV replication in these transgenic mice is sensitive to 3TCand that, without treatment, consistent virus titers are observed(18, 20). Among the transgenic mice treated with L-2'-Fd4C (Fig.2B), the virus levels became undetectable during the periodof treatment. In these mice, the virus remained undetectableor rebounded to barely detectable levels after the end of treatment.In one of these six mice, the virus levels became undetectablebut partial rebound was observed by day 15, at the end of theexperiment. One mouse was a nonresponder, and the last one inthis group died unexpectedly shortly after the beginning oftreatment. Collectively, these results showed that both 3TCand L-2'-Fd4C are active against HBV in this animal model.

View larger version (12K):
[in this window]
[in a new window]
FIG. 2. (A) HBV DNA levels in mice treated with 3TC ( ${square}$ , ${blacklozenge}$ ) or PBS ( ${blacksquare}$ ). Levels for mice with sustained ( ${blacklozenge}$ ) or transient ( ${square}$ ) responses to 3TC are shown. (B) HBV DNA levels in mice treated with L-2'-Fd4C, showing a sustained ( ${blacklozenge}$ ), transient ( ${square}$ ), or no ( ${blacksquare}$ ) response. Treatments were conducted for the first 7 days, as indicated by horizontal bars.

To estimate the antiviral efficacy of each treatment, the viralAUC were calculated using data shown in Fig. 2 to representthe virus load for each animal through a period of time. Theaverage AUC for individual groups of drug-treated mice (on days0 to 5) was then subtracted from the AUC for PBS-injected animalsduring the same time interval to produce the difference value( ${Delta}$ AUC). The antiviral efficacy of L-2'-Fd4C (average ${Delta}$ AUC, 251%;standard deviation, 46%; coefficient of variation, 18.3%) wasat least as high as that of 3TC (average ${Delta}$ AUC, 235%; standarddeviation, 93%; coefficient of variation, 39.5%), which wasused as a positive control, and no statistical difference wasfound (P > 0.05). This suggested that pharmacokinetic studieswith larger animals are warranted to further develop L-2'-Fd4C.

Pharmacokinetic studies with rhesus monkeys. Pharmacokinetic studies were performed with two rhesus monkeysgiven oral and i.v. doses of 33.3 mg of L-2'-Fd4C/kg. The concentrationsin plasma versus the time data were fitted simultaneously forthe two routes of administration (Fig. 3). The fitted parametersfor the individual monkeys and the pooled data are shown inTable 1. The concentrations in plasma declined in a biexponentialmanner following i.v. administration, having a terminal-phaset_1/2 of 5.02 h. The central-compartment volume of distributionwas 0.43 liter · kg^-1, and the steady-state distributionvolume was 1.09 liter · kg^-1. The respective systemic,renal, and nonrenal clearance values were 0.25, 0.16, and 0.09liter · h^-1 per kg, respectively. L-2'-Fd4C was rapidlyabsorbed after oral administration (K_a = 0.18 h^-1), with peakconcentrations in plasma achieved 2 h after dosing and approximately44% of the oral dose absorbed. The peak concentrations for theoral doses were 6.1 and 6.6 µg/ml (26.7 and 29.2 µM),respectively, for the two animals, and L-2'-Fd4C was detectablein plasma for up to 24 h after dosing.

The cumulative urine excretion profile of L-2'-Fd4C versus themidpoint of the urine collection interval for i.v. and oraldoses in the two monkeys is shown in Fig. 4, together with therespective fitted curves. The urinary accumulation approacheda plateau within 24 h for both routes of administration. Thepercentages of the i.v. and oral doses recovered in the urinewere 53 and 19%, respectively. No additional peaks were seenon the HPLC UV chromatograms of the monkey urine samples, suggestingthat no metabolites of L-2'-Fd4C with UV absorption were excretedin the urine. The mean concentrations in CSF 1 h after the administrationof i.v. and oral doses were 1.97 and 0.65 µM, respectively.

View larger version (13K):
[in this window]
[in a new window]
FIG. 4. Model fitted to pooled data (lines) and the observed cumulative levels of L-2'-Fd4C excreted in the urine of two rhesus monkeys (filled and open symbols) after the administration of 33.3 mg of the compound per kg by the i.v. (triangles) and oral (circles) routes.

DISCUSSION

Top

Discussion

L-2'-Fd4C is a recently synthesized cytosine L-nucleoside analoguewith anti-HIV and anti-HBV activities in vitro (21). The purposeof this study was to determine its antiviral efficacy in vivoin HBV-transgenic mice and its single-dose i.v. and oral pharmacokineticsin rhesus monkeys, to further its development as an antiviralagent.

Short-term treatment with L-2'-Fd4C in HBV-carrying transgenicmice demonstrated efficacy comparable to that of 3TC, the onlyapproved L-nucleoside analogue used for the treatment of HIVand HBV infections (Table 1). L-2'-Fd4C continued to suppressHBV replication even after the end of treatment, since the compoundmaintained low or undetectable levels of viral DNA in the majorityof the treated animals.

The pharmacokinetics profile for L-2'-Fd4C in rhesus monkeyssuggests that this nucleoside analogue may have a more prolongedt_1/2 in plasma than most existing nucleosides (see below). Allobserved concentrations in plasma for the i.v. dose closelyfollow the model predictions, indicating that the tissue dispositionof L-2'-Fd4C is well described by a two-compartment model. However,for the oral dose, there was an approximately 20% differencebetween the observed results and the C_max in plasma predictedby the model, which was in the range of variability for thetwo animals during the distribution phase. The data points between6 and 8 h for the oral dose suggest that the model did not mimicthe absorption phase closely, and there were insufficient datapoints between 8 and 24 h to fit a more complex absorption model.The systemic clearance (CL_T) value of 0.25 liter · h^-1· kg^-1 for L-2'-Fd4C was comparatively low, consideringthe reported hepatic plasma and renal flow rates of 1.26 and0.9 liter · h^-1 · kg^-1, respectively, in monkeys(13). The total clearance of L-2'-Fd4C in rhesus monkeys wasmuch lower than that of 3TC in rhesus monkeys (0.76 liter ·h^-1 · kg^-1), with corresponding t_1/2s in plasma of 5.02and 1.40 h, respectively (1). CL_R of L-2'-Fd4C in rhesus monkeysapproached the glomerular filtration rate (0.16 compared to0.2 liter · h^-1 · kg^-1) and suggests that eliminationoccurred by passive renal filtration. The steady-state volumeof distribution of L-2'-Fd4C was similar to that of 3TC (1.09versus 1.16 liters · kg^-1), suggesting that it has amoderate tissue distribution. Nearly half of the oral dose reachedthe systemic circulation. L-2'-Fd4C demonstrated similar dispositionsinto plasma and urine in the two monkeys (Fig. 3 and 4). Consequently,the fitted parameters of L-2'-Fd4C in individual animals wereconsistent with the pooled data (Table 2).

L-2'-Fd4C and D-D4FC (Reverset) are antiviral cytosine nucleosidesin development by our group (Fig. 1). The fluorine group ofL-2'-Fd4C is on the sugar ring, while that of D-D4FC is on thenucleoside pyrimidine base at the 5 position. Both nucleosideshave similar steady-state volumes of distribution in rhesusmonkeys (1.09 versus 1.24 liters · kg^-1). Similar steady-statedistribution volumes have been reported for other approved D-nucleosides,such as zidovudine, DDC (2',3'-dideoxycytidine), D4T (3'-deoxy-2',3'-didehydrothymidine),and DDI (2',3'-dideoxyinosine), ranging from 0.90 to 1.09 liters· kg^-1 (3, 25, 27). The t_1/2s in plasma of these latternucleosides range from 0.81 to 1.82 h. However the t_1/2s ofL-2'-Fd4C and D-D4FC are 5.02 and 3.57 h, respectively, suggestingthat they may be given with a longer dose interval (22). L-2'-Fd4Cand D-D4FC had similar oral bioavailabilities (F, 0.44 versus0.41), C_maxs (123.1 versus 33.4 µM), and T_maxs (2.00 versus2.67 h). The ability of antiviral agents to penetrate the centralnervous system is a critical preliminary selection criterion,especially for the anti-HIV drug candidates, since AIDS dementiacomplex can occur in infected patients and the central nervoussystem serves as a sanctuary for HIV (11). Zidovudine is theonly approved anti-HIV agent with proven clinical benefits inthe treatment of central nervous system manifestations of thisdisease (4). D-D4FC was not detected in the CSF of two-thirdsof the monkeys until 2 h after oral dosing, while L-2'-Fd4Cwas detected in the CSF of both monkeys 1 h after oral dosingand its level exceeded the EC₅₀ for HIV measured in human peripheralblood mononuclear cells (22). The precise mechanism that controlsthe degree of penetration of the two agents into the CSF hasnot been determined.

It would be desirable for future dosing protocols of L-2'-Fd4Cto consider its pharmacokinetics in plasma and its cellularpharmacology to ensure that average concentrations in plasmain excess of the EC₉₀ and between-dose trough concentrationsof the EC₅₀ value against the target viruses are reached. TheEC₅₀ and EC₉₀ values of the compound against the virus were0.12 and 1.60 µM, respectively, for HIV and 0.002 and0.05 µM, respectively, for HBV. Antiviral nucleosidesgenerally demonstrate longer t_1/2s in humans than in monkeys;e.g., the t_1/2 of 3TC is 1.40 h in monkeys versus 5 to 7 h inhumans (1, 3, 23). However, the biological activity of nucleosidereverse transcriptase inhibitors (NRTIs) is due to the accumulationof intracellular nucleoside triphosphate (NRTI-TP). Since L-2'-Fd4Cis active against HIV and HBV, it should form an NRTI-TP. Theaccumulation and t_1/2s of most NRTI-TPs are usually of the firstorder with respect to the concentration of NRTI in the plasma.Therefore, NRTI-TP levels are a function of the interplay betweenthe t_1/2 of the NRTI in plasma and the intracellular stabilityof the NRTI-TP. Furthermore, the t_1/2 of the nucleoside triphosphateis significantly longer than the t_1/2 of the NRTI in plasma(37). Therefore, L-2'-Fd4C, like lamivudine, may have potentialfor once-daily dosing. Additional studies on the intracellulart_1/2 of L-2'-Fd4C, using a radiolabeled compound, are planned.

Future work will address whether L-2'-Fd4C treatment is alsoeffective against the chronic liver disease that develops inHBV-transgenic SCID mice after the adoptive transfer of naive,syngeneic splenocytes (20). Such experiments will provide amore comprehensive evaluation of L-2'-Fd4C and should also includestudies of chronically infected woodchucks (19).

The favorable pharmacokinetic profile, including the longestterminal-phase t_1/2 of NRTIs evaluated in rhesus macaques, anacceptable oral bioavailability, the absence of cellular toxicityin human-derived cell lines, and the antiviral activity observedin vitro and in vivo, indicate that L-2'-Fd4C warrants furtherdevelopment as a potential antiviral agent.

ACKNOWLEDGMENTS

This work was supported by NIH grants 1RO37-AI-41980 (R.F.S.)and CA79512 (M.F.), the Emory CFAR grant 1P30-AI-42847 (R.F.S.),grant 1RO1-AI-32351 (R.F.S. and C.K.C.), and the U.S. Departmentof Veterans Affairs (R.F.S).

FOOTNOTES

* Corresponding author. Mailing address: Veterans Affairs Medical Center, Medical Research 151H, 1670 Clairmont Rd., Decatur, GA 30033. Phone: (404) 728-7711. Fax: (404) 728-7726. E-mail: rschina{at}emory.edu.

REFERENCES

Top