INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The discoveries in 1990 (25) that nonnucleoside reverse transcriptase inhibitors (NNRTIs) could inhibit human immunodeficiency virus type 1 (HIV-1) replication in vitro and that their activities were potent and selective for the RT of HIV-1 (2, 11, 14) signaled an important advance for the treatment of HIV infection. In contrast to the nucleoside inhibitors of HIV-1 RT (NRTIs), the NNRTIs inhibited HIV-1 RT in a noncompetitive fashion and at a site on the RT distinct from the site inhibited by NRTIs (12). Clinical experience with NRTIs demonstrated that these compounds were able to reduce the morbidity associated with progression of HIV-1 infection. Nonetheless, due to acute and chronic toxicities, various degrees of antiviral activity, and, most significantly, the development of resistance when used in monotherapy, the NRTIs were limited in their ability to provide durable suppression of HIV-1 replication (22). The identification of NNRTIs, and likewise protease inhibitors, as potent inhibitors of HIV-1 replication that acted in a manner distinct from that of the NRTIs provided the basis for combination or "coactive" therapy (11). By inhibiting HIV-1 through distinct mechanisms, various combinations of NRTIs, protease inhibitors, and NNRTIs have yielded potent coactive regimens that have demonstrated durable suppression of HIV-1 replication (5, 34) and greatly improved clinical benefit.

The NNRTIs of the HIV-1-encoded RT are a chemically diverse set of structures that show extremely potent and selective antiviral activities (2, 3, 5, 12, 14). Despite the structural diversity seen in this group of inhibitors, X-ray crystallography studies of HIV-1 RT complexed with a number of different NNRTIs have shown all of these compounds to bind to a single allosteric site, approximately 10 Å from the polymerase catalytic site (13, 21). This binding site, which is predominantly hydrophobic, is located in the p66 palm domain of the p66/p51 heterodimer between the -sheet comprising 4, 7, and 8 and the sheet comprising 9, 10, and 11 (21).

Emivirine (EMV) [6-benzyl-1-(ethoxymethyl)-5-isopropyl-uracil] belongs to the hydroxyethyl phenyl thymine (HEPT) series of NNRTIs (Fig. 1). This HEPT derivative is in a different class of compounds from nevirapine (26, 29), delaviridine (12), and efavirenz (17), the three NNRTIs currently approved for HIV-1 treatment. EMV is structurally similar to nucleoside analogs, but it has been shown to be a potent, noncompetitive inhibitor of HIV-1 RT (3, 16), with K_i values of 0.20 and 0.01 µM for dTTP- and dGTP-dependent DNA or RNA polymerase activity, respectively (37). Fifty percent inhibitory concentrations (IC₅₀) and IC₉₀ against laboratory-adapted strains of HIV-1 ranged from 1.6 to 19 nM and from 7.9 to 98 nM, respectively (3, 7). Against clinical isolates, the IC₅₀ for EMV ranged from 2 to 40 nM (3, 7, 32). EMV was also effective against HIV-1 when it was combined with other therapeutic agents in vitro (16, 27, 31). EMV has been shown to bind strongly to human plasma proteins, with the extent of binding ranging from 78 to 96% at concentrations in human serum from 10 to 100% (4). The addition of 30% human serum to the extracellular medium resulted in a twofold increase in IC₅₀, while the addition of 50% human serum resulted in a fivefold increase. However, protein binding is typically a saturable process and EMV has demonstrated activity in clinical trials (C. P. Moxham, K. Borroto-Esoda, D. Noel, P. A. Furman, G. M. Szczech, and D. W. Barry, Abstr. 10th Int. Conf. Antivir. Res., abstr. A44, 1997). EMV is also specific for HIV-1 RT and was without effect on HIV-2 (3, 5, 32). In this paper, we present the results of a preclinical safety and pharmacokinetic assessment of EMV. The results are in support of the rapid development of EMV as a potential agent for the treatment of HIV infection.

View larger version (11K): [in this window] [in a new window]   FIG. 1.   Chemical structure of EMV.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Cytotoxicity. Human bone marrow cells collected from normal healthy volunteers were used to assess bone marrow stem cell cytotoxicity. The mononuclear cells were isolated from whole heparinized marrow by Ficoll-Hypaque gradient centrifugation as described previously (33). Cells were washed twice with Hanks' balanced salt solution and counted with a hemocytometer, and viability was assessed by Trypan blue dye exclusion. The cells were then plated in a bilayer of soft agar or methylcellulose (10⁵/plate), and treated with a 0, 0.1, 1, 10, or 100 µM concentration of either EMV or zidovudine (AZT). After 14 days of incubation at 37°C in a humidified atmosphere of 5% CO₂, colonies (50 cells) were counted with an inverted microscope.

Mitochondrial toxicity. Toxicity towards mitochondria was evaluated with HepG2 cells by measuring cell growth, the production of lactic acid in extracellular medium, mitochondrial DNA content, and structural changes in the mitochondria as previously described (10). HepG2 cells (2.5 × 10⁴ cells/ml), grown in minimal medium with nonessential amino acids and supplemented with 10% serum, 1% sodium pyruvate, and 1% penicillin-streptomycin, were plated in 12-well culture dishes and treated with various concentrations (0, 0.1, 1, and 10 µM) of EMV. After 4 days of incubation, cell growth was assessed by counting the number of cells. The medium was collected, and lactic acid was measured with an assay kit purchased from Boehringer Mannheim Corp. (Mannheim, Germany).

EMV measurement. A high-performance liquid chromatography (HPLC) method for the quantification of EMV in plasma was developed and validated. The HPLC system consisted of a Waters Millennium System, Waters 717 WISP with refrigerated autosampler, Varian 9010 ternary pump, HPLC column heater, and Applied Biosystems model 783A programmable absorbance detector. EMV was extracted from plasma with diethyl ether after addition of 0.1 M boric acid and sodium hydroxide buffer, pH 10. The organic layer was transferred to a glass tube and evaporated to dryness under a stream of nitrogen at room temperature. The residue was reconstituted in 250 µl methanol-acetonitrile-H₂O (7:45:48, vol/vol/vol) and transferred to HPLC vials. A structural analog of EMV was used as an internal standard, and it was added to the plasma samples before extraction to correct for any variations in recovery. EMV was separated from the internal standard with a TSK-GEL ODS-80TM, 150- by 4.6-mm column (Tosohaas, Montgomeryville, Pa.) with an in-line frit filter under gradient conditions. The retention times for EMV and the internal standard were 8 to 10 and 12 to 15 min, respectively. The peaks were detected by UV absorbance at a 268-nm wavelength and integrated to quantify the EMV levels against the known internal standard. With an injection volume of 100 µl and a sample volume of 1.0 ml, EMV concentrations in plasma samples from mouse, rat, rabbit, and monkey ranged from 5 to 800 ng/ml. The standard curve was linear in the range of 2.5 to 1,000 ng/ml.

Absorption, distribution, metabolism, and excretion. To guide the selection of an appropriate nonrodent animal model for EMV toxicology experiments, male Sprague-Dawley rats, beagle dogs, and cynomolgus (Macaca fascicularis) monkeys were given oral doses of the compound. EMV was suspended in 0.5% tragacanth gum, and a single dose of 50 mg/kg of body weight was given by gavage to seven groups of five male rats (body weight, 128 to 200 g) and to a group of four fasted male cynomolgus monkeys weighing 3.1 to 3.9 kg. Blood samples were collected from the rats at 0.25, 0.5, 1, 2, 4, 6, and 8 h postdose and from the monkeys at 0.5, 1, 4, 8, 24, and 48 h postdose. For the dogs, EMV was placed in gelatin capsules and a single dose was given to four fasted males. Blood samples were collected at 5 min and 0.25, 0.5, 1, 2, 4, 8, 12, and 24 h postdose. The blood samples were spun to separate the plasma, which was then analyzed for concentrations of EMV by the HPLC method described above. In this experiment, additional groups of male rats were given a single dose of EMV by injection into the caudal vein (0.88 mg/kg), by gavage (5 mg/kg), by intrarectal infusion (5 mg/kg) while the rats were anesthetized with urethane, and by infusion into the portal vein (0.25 mg/kg), again while the rats were anesthetized. The vehicles were 0.5% tragacanth gum in the cases of the oral and intrarectal doses and plasma (filtrate collected from untreated animals) in the cases of the intravenous and intraportal vein doses. Blood samples were collected at 5 min and 0.25, 0.5, 1, 2, 4, 6, and 8 h after the intravenous dose; at 5 min and 0.25, 0.5, and 1 h after the intraportal vein dose; and at 0.25, 0.5, 1, 2, 4, 6, and 8 h after the oral or intrarectal dose. Concentrations of EMV in plasma harvested from the blood samples were measured as described above to study absorption and first-pass metabolism in rats.

1

1

1

1

In vivo metabolism. To study the effects of EMV on hepatic drug-metabolizing enzymes, groups of five male Sprague-Dawley rats were given 0 (tragacanth vehicle), 15, 50, or 150 mg of EMV kg¹ day¹ by gavage for 14 days. The daily dose was divided and given in two equal installments separated by approximately 6 h, similar to the procedure in the rat toxicology experiments described below. Another group of four rats was given phenobarbital once a day at 80 mg kg¹ day¹ for 14 days as a positive control. Following necropsy, livers from the rats were perfused with 10 ml of saline via the portal vein, removed, and weighed. An aliquot (2.5 g) from the left lateral lobe was homogenized in 0.25 M sucrose on ice, and microsomes were isolated (20). Protein content, cytochrome b5 and P450 levels, and the activities of NADPH-cytochrome c reductase, aniline hydroxylase, aminopyrine N-demethylase, UDP-glucuronyltransferase, and 7-ethoxycoumarin O-deethylase were measured by standard spectrophotometric assays. Statistical differences were calculated by one-way analysis of variance and Dunnett's critical difference test (8) or, in the case of phenobarbital, Student's t test.

In vitro metabolism. Three experiments designed to determine species-related differences in the microsomal oxidative metabolism of EMV and to identify the principal (human) cytochrome P450 enzymes involved were performed in vitro. First, liver microsomes from four humans, four cynomolgus monkeys, and four Wistar rats, all males, were isolated (20) and incubated for 25 min at 37°C with seven concentrations of EMV, ranging from 3 to 300 µM. Reactions were stopped by adding acetonitrile to the mixture. The samples were then extracted and analyzed by HPLC for EMV and its dealkylated product. K_m values and maximum rates of metabolism (V_max) were determined by the Gauss Newton method from Lineweaver-Burk plots. Second, pooled liver microsomes (0.2 mg of protein) isolated from male rats, male cynomolgus monkeys, and male humans were incubated for 0 to 60 min at 37°C with 100 µM EMV. The incubation times were selected to be higher and lower than the 25 min used in the first experiment. Reactions were stopped with cold perchloric acid (30%), and protein was precipitated by centrifugation. The supernatants were then analyzed by the HPLC-mass spectrometry (MS) method described below. Third, liver microsomes from 10 to 14 individual human donors were incubated for 0 or 15 min at 37°C with EMV at concentrations of 10 and 100 µM. Samples were prepared and analyzed by HPLC-MS. Some of the human samples were incubated with troleandomycin (20 or 100 µM) or nifedipine (20 µM), inhibitors of many cytochrome P450 3A-catalyzed reactions, or 5 µM furafylline, a cytochrome P450 1A2 inhibitor, to obtain chemical inhibition data for correlation analyses. The experiments were controlled by using microsomes with certified activity, by running samples in duplicate or triplicate, by using zero protein blanks, by comparing interindividual variation in the rates of EMV metabolism among samples of liver microsomes from 14 humans, and by using a Pearson product-moment correlation (8) to calculate regression coefficients between the metabolic data for EMV and several standard probe drugs.

Safety pharmacology experiments. Experiments performed to detect potential pharmacologic effects of EMV are listed in Table 1. EMV was suspended in 0.5% tragacanth gum and administered orally to mice and rats (3 to 10/group) at the concentrations indicated in Table 1. Male and female beagle dogs (five/group), anesthetized with pentobarbital, were given EMV intraduodenally to study a range of cardiovascular parameters. Isolated strips of ileum collected from male Hartley guinea pigs (five/group) were exposed in vitro to concentrations of EMV as large as 10⁶ µM in an effort to detect potential pharmacologic effects on smooth muscle. Other parameters studied are listed in Table 1.

Toxicology experiments. The acute toxicities of EMV and one of its putative metabolites, 6-benzyl-5-isopropyl-uracil (BIU), were assessed in CD male and female rats. Briefly, EMV and BIU were synthesized, suspended in 0.5% tragacanth gum, and administered to groups of five male and five female rats as single doses of 0, 2,083, 2,500, and 3,000 mg/kg. Animals were observed for 14 days, and body weights were recorded. In addition, preclinical safety evaluation experiments were performed with mice, rats, and cynomolgus monkeys as listed in Table 2. In all the experiments, EMV was delivered orally. The vehicle in the 1-month experiments was 0.5% tragacanth gum, and that in the subchronic and chronic experiments was 0.5% methylcellulose. The daily doses shown in Table 2 were given in two equal portions with 6 to 12 h between doses. The animals were bled at five to seven timed intervals to provide toxicokinetic data. In each subchronic experiment, the bleeding was performed on dose day 1 or 2 and repeated at the end of the dose period. In the chronic (6-month) rat study, bleeding was at dose day 1 and weeks 13 and 26. Monkeys in the 1-year experiment were bled at dose day 1 and weeks 4, 13, 26, and 52. Plasma was separated from the blood, extracted, and analyzed by HPLC for concentrations of EMV.

1

1

1

1

Preclinical pharmacokinetics and toxicokinetics. The comparative (rat, dog, and cynomolgus monkey) experiment was used to define pharmacokinetic parameters for laboratory animals. Calculation of the maximum concentration of a drug in serum (C_max) and time to C_max (T_max) was achieved by the use of standard formulas applied to values derived with the concentrations of EMV measured in plasma as described earlier. Values for the area under the concentration-time curve (AUC) were calculated by extrapolating the measured concentrations in plasma to infinity (AUC) and applying the trapeziodal rule. Bioavailability in rats was calculated from the ratios of AUC values to doses for the various dose routes (oral, intrarectal, and intra-portal vein) compared to the AUC for the intravenous dose of EMV and is expressed as percent bioavailability.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Cytotoxicity. Bone marrow toxicity has been associated with certain nucleoside analogs. Since EMV contains a substituted nucleobase, we examined the effect of the compound on bone marrow progenitor cells (Table 3). The results of these experiments demonstrated that EMV did not cause significant cytotoxicity compared with AZT (positive control). EMV was determined to have 50% cytotoxic concentrations of 30 and 50 µM for the erythroid (BFU-E) and granulocyte macrophage (CFU-GM) progenitor cells, respectively. In these experiments AZT gave 50% cytotoxic concentrations of <0.1 µM for the BFU-E progenitor cells and 7 µM for the CFU-GM progenitor cells.

Mitochondrial toxicity. The effect of EMV on mitochondrial functions was examined in exponentially growing HepG2 cells (Table 4). After the HepG2 cells were incubated with EMV at concentrations of 0.1 to 10 µM, no effect on cell growth, lactic acid production, mitochondrial DNA synthesis, or mitochondrial structure was seen compared to what occurred with untreated HepG2 cells.

Preclinical pharmacokinetics and toxicokinetics. Pharmacokinetic experiments identified rats and monkeys as appropriate for toxicology experiments (Table 5). In dogs, levels of EMV in plasma declined rapidly and were undetectable by 1 h postdose. Autoradiography demonstrated that [¹⁴C]EMV was widely distributed to tissues of rats given 10 mg/kg by gavage, and at 0.5 h postdose, radioactivity was noted in all tissues, including brain and spinal cord. At 96 h postdose, radioactivity was detected only in the contents of the gastrointestinal tract. The total excretion of EMV was 99% of the administered dose, with 38% of the radioactivity being excreted into urine and 61% being excreted into feces. In the rat biliary excretion experiment, 25% of the administered EMV was excreted into bile at 1 h postdose; 75% was excreted into bile after 8 h. At 48 h postdose, 87, 11, and 2% of the labeled compound was found in the bile, urine, and feces, respectively. In a separate experiment, concentrations of EMV in brains from rats given a dose of 250 mg/kg by gavage were the same as those in plasma over the interval of 0.5 to 12 h postdose (Fig. 2).

View larger version (11K): [in this window] [in a new window]   FIG. 2.   Plasma and brain EMV concentrations in rats. Rats were given an oral dose of EMV (250 mg/kg of body weight). Plasma and brain tissues were collected at the time intervals indicated, the latter being homogenized in 1.15% KCl. EMV was extracted, and concentrations were measured by an HPLC assay described in Materials and Methods. Data are expressed as means ± standard deviations (SD) (number of mice/group, 4). Conc., concentration.

1

1

1

1

In vitro metabolism. When liver microsomes from rats, cynomolgus monkeys, and humans were incubated with EMV, the relative affinities of the drug (at 3 µM) for the microsomes were greater for humans than rats but greater for monkeys than humans while the K_m/V_max values were greater for rats than monkeys but significantly greater for humans than rats (Table 8), suggesting a much slower metabolism of EMV in humans. This was confirmed with additional in vitro experiments where human liver microsomes formed only about a third (8 nmol) of the total EMV metabolites measured for microsomes from rats (24 nmol) and monkeys (26 nmol) after 60 min of incubation under identical conditions. These in vitro experiments also showed that three putative metabolites, detected with m/z of 319, 245, and 261 by MS, were produced (although their identities have not yet been confirmed) by the three species, but in differing proportions. In the rat and monkey, 65% of the total metabolites formed was that with the m/z of 245, tentatively identified as BIU. The other two metabolites were approximately equal in amount. In contrast, that with an m/z of 319 was the predominant metabolite in human microsomes, accounting for 54% of the total metabolites measured, followed by the putative metabolite BIU (37%). The three species formed the m/z-319 and -261 metabolites at similar rates, but at 60 min, the rat and monkey microsomes had produced five times more BIU than the human samples. Comparison with results of standard cytochrome P450-associated reactions, such as testosterone 6-hydroxylation for cytochrome P450 3A4 or -5, and inhibition experiments with nifedipine and troleandomycin showed that, in humans, EMV was metabolized by the cytochrome P450 enzymes 3A4 and 3A5. However, at low, pharmacologic concentrations (10 µM) of EMV, 40% of the total BIU formed in human microsomes was not abolished by troleandomycin, a cytochrome P450 3A4- and P3A5-specific inhibitor. Coupled with the results of the correlation analysis of cytochrome P450 1A2 activity (7-ethoxyresorufin O-dealkylase) in the individual human samples, this result suggested that in humans, the formation of BIU was catalyzed, in part, by cytochrome P450 1A2. We then observed that furafylline, a cytochrome P450 1A2-specific inhibitor, inhibited BIU formation in human microsomes treated with 10 µM EMV by approximately 50%, without affecting the formation of the other two major metabolites.

Safety pharmacology experiments. There were no important or consistent pharmacologic effects of EMV in the wide variety of safety pharmacology experiments performed. A significant increase in the duration of anesthesia produced by hexobarbital was observed in mice given an oral dose at 300 mg/kg but not at 100 mg/kg. Similarly, EMV accelerated intestinal transit in mice at 300 mg/kg per os but not at 100 mg/kg. There was no effect on respiration rate, blood pressure, heart rate, or electrocardiographs in anesthetized dogs given EMV at 300 mg/kg of body weight.

Toxicology experiments. In single-dose experiments, the approximate lethal oral dose of EMV for rats was 3 g/kg for males and 2.5 g/kg for females. BIU, a putative metabolite of EMV, did not produce death in rats given a single oral dose of 3 g/kg, indicating that it was no more toxic than EMV. The 1-month experiment with rats identified 50 mg/kg/day as a no-effect dose where average peak levels of EMV in plasma were 3,090 ng/ml and the average AUC was 5,064 ng · h/ml, as determined in the pharmacokinetic experiment. At the next dose level (150 mg/kg/day 4), decreased body weights, increased blood urea nitrogen (BUN), vacuoles in kidney tubules, increased serum alanine aminotransferase activity, and hepatocellular hypertrophy were observed in the animals. The livers were normal by histopathology. The 1-month experiment with monkeys identified a no-effect dose of 40 mg/kg/day where average peak values for plasma EMV ranged from 5 to 67 ng/ml and the AUC ranged from 30 to 412 ng · h/ml. Inconsistent emesis, mild diarrhea, and liver and kidney effects similar to those in rats were also observed in the monkeys at the next-highest dose, 200 mg/kg/day. At this dose, average peak values for EMV in plasma ranged from 30 to 170 ng/ml and the AUC_0-24 ranged from 324 to 1,912 ng · h/ml.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The chronic treatment of HIV-1 infection requires the availability of active, tolerable, safe, and conveniently administered agents for potent coactive regimens. EMV is an NNRTI derived from the HEPT chemical series of NNRTIs, which are among the most active in inhibiting the replication of HIV-1. Although EMV functions as an NNRTI, structurally it resembles an NRTI (12). In order to address the potential for NRTI-like toxicities (36), in vitro experiments using human bone marrow progenitor cells were performed. When compared with AZT, EMV had no effect on inhibiting growth of human bone marrow progenitor cells at concentrations that exceeded those measured in plasma in animals given doses of EMV that produced the kidney and gastrointestinal toxicities already described. EMV also demonstrated low cytotoxicity when it was compared with other anti-HIV agents such as saquinavir (7, 16). Recent reports have indicated that mitochondrial toxicity plays a major role in the adverse effects related to some nucleoside analogs (10), suggesting that nucleoside analogs under development should be evaluated for potential mitochondrial dysfunction. EMV had no effect on HepG2 cells treated with a 10 µM concentration of the drug for several days. When other nucleoside analogs were tested in HepG2 cells at this concentration, toxic effects such as increased lactic acid production (ranging from 13 to 79%) and mitochondria that were swollen or had loss of cristae were produced (10).

Monkeys were selected as the appropriate nonrodent toxicology model for several reasons. In pharmacokinetic experiments, the disappearance of EMV from plasma in monkeys was more gradual than in rats and much more gradual than in dogs. Thus, compared to dogs, monkeys would be more likely to achieve the exposure to EMV needed for adequate safety assessment. Dose-limiting neurological toxicities have been encountered with the clinical use of nucleoside analogs, most notably dideoxycytidine (6) and dideoxyinosine (24). Monkeys are a good model for studying potential effects of new antiviral drugs on the peripheral nervous system because noninvasive techniques to measure the conduction velocity in peripheral nerves exist. Indeed, these techniques were used in the chronic monkey experiment with EMV. Finally, the enzyme systems in monkeys and humans are very similar, including those that phosphorylate deoxynucleosides (18).

Since HIV-1 replication can occur not only in plasma but also in sanctuary sites such as the lymphoreticular system (15, 28) and the central nervous system (30), the ability of antiretroviral agents to distribute to these sites is desirable. To this aim, the tissue distribution of EMV was examined in rats following oral administration of ¹⁴C-labeled EMV. There was widespread whole-body distribution of radioactivity at 0.5 h postadministration that decreased to negligible levels by 24 and 96 h postadministration. Rats readily absorbed EMV and excreted it into bile. Moreover, EMV was able to cross the blood-brain barrier by 0.5 h after oral dosing. This ability is important because one limitation of many HIV therapies is the failure to distribute to the central nervous and lymphoreticular systems, where replication is known to occur (35). The penetration of EMV into cerebral spinal fluid in HIV-infected volunteers is under examination in clinical trials.

In vitro experiments with liver microsomes from rats, cynomolgus monkeys, and humans revealed that EMV was a substrate for metabolism by liver enzymes. These experiments demonstrated that human hepatic microsomes metabolize EMV more slowly than those from rat or monkey, producing, in aggregate, far less of the three putative EMV metabolites detected than was produced by the other two species. The cytochrome P450 isozymes 3A4 and 3A5 appeared to account for the majority of EMV microsomal metabolism, with 1A2 playing a role at therapeutic concentrations. This contribution from another P450 enzyme, depending on the substrate concentration, has been demonstrated with a protease inhibitor (9) and with diazepam (36). EMV induction of the latter enzyme, cytochrome P450 1A2, was supported by in vivo experiments where levels of 7-ethoxycoumarin O-deethylase activity were elevated in liver microsomes isolated from rats treated with EMV. The in vitro findings were also consistent with the observation of substantial first-pass hepatic metabolism in rats, in which a total oral bioavailability of 18% was observed, compared with an oral absorption of 68%. In addition, the observation that plasma EMV levels in the monkey toxicology experiments were higher at dose week 1 than at subsequent times most likely reflects the induction of hepatic drug-metabolizing enzymes.

The results of the acute, chronic, genetic, and reproductive toxicology experiments have demonstrated an attractive therapeutic opportunity for EMV. The only toxicity observed in the toxicology experiments involved the kidney and was reversible and limited to high doses of EMV. The association of this toxicity to EMV or to its metabolites was not determined. Hepatocellular hypertrophy, without increases in diagnostic liver enzymes or additional histopathologic alterations, was judged to be a functional change associated with the induction of drug-metabolizing enzymes in both rats and monkeys. The safety profile of EMV in all of these experiments supported the initial phase I administration of EMV to humans (C. P. Moxham, G. M. Szczech, M. R. Blum, and D. W. Barry, Program Abstr. 4th Conf. Retrovir. Opportunist. Infect., abstr. 573, 1997) and permitted selection of oral doses of EMV for use in that study. Since the no-effect level was 30 mg/kg/day in both the rat and monkey 1-month toxicology experiments, a dose 10-fold lower (3 mg/kg or 150 mg for a 50-kg patient) was selected as the starting dose for the clinical program with EMV. The safety, tolerability and pharmacokinetics from a multiple-dose study (Moxham et al., 10th ICAR) have supported the continued clinical investigation of EMV.

The results from the reproductive toxicology experiments have further demonstrated a positive preclinical safety profile for EMV. With the exception of decreases in maternal and fetal body weights at the highest dose, 160 mg/kg/day, there was no effect of EMV on developmental or reproductive parameters. Given that women represent a growing percentage of the HIV-infected population, the ability to administer to them safe and effective treatments for HIV-1, without potential harm to their unborn children, is important. Pilot perinatal transmission studies to examine the safety, tolerability, and pharmacokinetics of EMV in pregnant women and their infants have been initiated based on data from our reproductive toxicology experiments (T. B. Grizzle, F. S. Rousseau, C. P. Moxham, and G. M. Szczech, Abstr. 38th Intersci. Conf. Antimicrob. Agents Chemother, abstr. I-22, 1998). The toxicological data, as well as the clinical safety profile for adults, also give additional support for the study of pharmacokinetics, safety, tolerability, and activity of EMV as part of coactive therapy in children.

In conclusion, the extensive study of EMV in the preclinical setting and its favorable preclinical profile have supported the continued development of EMV. These data have allowed for development of EMV beyond the initial administration to humans to large-scale clinical trials with adults and to pilot phase II studies of children, women, and pregnant women. EMV is currently under active development to examine its utility in coactive regimens for the treatment of HIV-1 infection.