Abstract
Efavirenz (Sustiva, Fig. 1) is a potent and specific inhibitor of HIV-1 reverse transcriptase approved for the treatment of HIV infection. To examine the potential differences in the metabolism among species, liquid chromatography/mass spectrometry profiles of efavirenz metabolites in urine of rats, guinea pigs, hamsters, cynomolgus monkeys, and humans were obtained and compared. The metabolites of efavirenz were isolated, and structures were determined unequivocally by mass spectral and NMR analyses. Efavirenz was metabolized extensively by all the species as evidenced by the excretion of none or trace quantities of parent compound in urine. Significant species differences in the metabolism of efavirenz were observed. The major metabolite excreted in the urine of all species was the O-glucuronide conjugate (M1) of the 8-hydroxylated metabolite. Efavirenz was also metabolized by direct conjugation with glucuronic acid, forming theN-glucuronide (M2) in all five species. The sulfate conjugate of 8-OH efavirenz (M3) was found in the urine of rats and cynomolgus monkeys but not in humans. In addition to the aromatic ring-hydroxylated products, metabolites with a hydroxylated cyclopropane ring (at C14) were also isolated. GSH-related products of efavirenz were identified in rats and guinea pigs. The cysteinylglycine adduct (M10), formed from the GSH adduct (M9), was found in significant quantities in only rat and guinea pig urine and was not detected in other species. In vitro metabolism studies were conducted to show that the GSH adduct was produced from the cyclopropanol intermediate (M11) in the presence of only rat liver and kidney subcellular fractions and was not formed by similar preparations from humans or cynomolgus monkeys. These studies indicated the existence of a specific glutathione-S-transferase in rats capable of metabolizing the cyclopropanol metabolite (M11) to the GSH adduct, M9. The biotransformation pathways of efavirenz in different species were proposed based on some of the in vitro results.
The effective treatment of HIV infection and AIDS is still difficult despite tremendous advances in our understanding of the pathogenesis of the disease and the arrival of potent drugs aimed at different, critical targets in the life cycle of the virus (Havlir and Richman, 1996). Clearly optimal treatment involves multiple drug therapy designed to decrease viral burden as low as possible. New agents with convenient dosing regimens are needed to ensure compliance. Efavirenz (Sustiva, Fig. 1) is a potent non-nucleoside inhibitor of the HIV-1 reverse transcriptase. Clinical trials have demonstrated a durable, long-lasting reduction in HIV RNA after once-a-day dosing in combination with other drugs (Staszewski et al., 1998). To more fully understand the disposition of this agent in relevant species for safety assessment, the metabolism of efavirenz has been described (Mutlib et al., 1998a,b, 1999).
To examine the potential differences in metabolism among species, liquid chromatography/mass spectrometry (LC/MS)1 profiles of efavirenz metabolites in urine of rats, guinea pigs, hamsters, cynomolgus monkeys, and humans were obtained and compared. Efavirenz was metabolized extensively by all species with the major urinary metabolite being the glucuronide conjugate, M1 (Christ et al., 1997). Significant species differences in the metabolism of efavirenz were observed. An attempt was made to determine differences in the nature of metabolites present in different species by isolating sufficient quantities of compounds from urine of rats, guinea pigs, and cynomolgus monkeys for full spectroscopic characterization.
It was shown earlier (Mutlib et al., 1998a), that efavirenz was metabolized to several diconjugates including GSH adducts. The formation and subsequent renal processing of the GSH adduct (M9) was postulated to be responsible for the species-specific renal toxicity observed in rats. GSH-related adducts were not produced by cynomolgus monkeys and humans, species without nephrotoxicity. Interestingly, the major GSH-related adduct found in urine of rats was the cysteinylglycine conjugate (M10). This metabolite has been shown to be derived from metabolite M9 by the action of γ-glutamyltranspeptidase (GGT) present in the kidney of rats. The presence of cysteinylglycine metabolite in urine is unique because it has not been previously isolated and characterized as a urinary metabolite of a GSH conjugate. However, it has been postulated that cysteinylglycines are intermediates in the catabolism of GSH conjugates to mercapturic acids (cysteine and N-acetylcysteine conjugates) (Commandeur et al., 1995). One of the objectives of this study was to demonstrate the existence or the absence of metabolite M10 in different species.
In vitro metabolism of efavirenz and its metabolites was conducted to confirm the formation of GSH conjugate, M9, and the pathways involved in the catabolism of this metabolite. Because species differences in the urinary metabolite profiles were observed, in vitro metabolism studies were conducted with several different isolated metabolites as substrates in the presence of liver and kidney subcellular fractions from rats, cynomolgus monkeys, and humans.
Materials and Methods
Chemicals and Supplies.
Efavirenz ((S)-6-chloro-4-(cyclopropylethynyl)-1,4-dihydro-4-(trifluoromethyl)-2H-3,1-benzoxazin-2-one), 7-OH efavirenz (M5), 8-OH efavirenz (M4), and the sulfate (M3) and glucuronide (M1) conjugates of 8-OH efavirenz were synthesized and characterized by DuPont Pharmaceuticals Company. Mega Bond-Elut C18 cartridges (10 g/60 cc, 500 mg/10 ml, and 500 mg/3 ml) were obtained from Varian Sample Preparation Products (Harbor City, CA). A C18 semipreparative column (10 × 300 mm, 5 μm, ultrasphere) and a C18 analytical column (4.6 × 250 mm, 5 μm, ultrasphere) were obtained from Beckman Instruments Inc. (Fullerton, CA). GSH, β-glucuronidase, p-nitrobenzyl chloride,S-(p-nitrobenzyl) GSH, and 1,2-dichloronitrobenzene were purchased from Sigma Chemical Co. (St. Louis, MO). All general solvents and reagents were the highest grade available commercially.
Liquid Chromatography/Mass Spectrometry.
LC/MS was carried out by coupling a Waters HPLC system to a Sciex API 300 mass spectrometer. The HPLC eluent was introduced into the source using a Turbo ionspray interface (PE Sciex, Thornhill, Ontario, Canada) held at 450–480°C. The electrospray needle was maintained at −3500 V with the orifice potential set at −50 to −60 V. The nebulizer gas was ultrapure nitrogen set at 40 psi. The Turbo ionspray gas flow rate was 6 to 8 liters/min. The mass spectrometer was operated in the negative ion mode (Q1 and MS/MS) to detect the polar metabolites of efavirenz. MS/MS was carried out using nitrogen as the collision gas. The collision energy was kept at 25 to 35 eV.
For the analyses of urine samples from all species, HPLC was carried out using a Waters quaternary pump coupled in sequence to a Waters WISP (model 717 plus) and to a Beckman C18 column (250 × 4.6 mm, 5 μm). Unless otherwise stated, the metabolites were separated by gradient solvent system consisting of a mixture of acetonitrile and 10 mM ammonium formate, pH 3.8. The percentage of acetonitrile in the mobile phase was increased from 25 to 80% in 20 min. The solvent flow rate was 0.8 to 1.0 ml/min. Aliquots of urine samples were injected directly onto the columns when carrying out LC/MS analyses. A postcolumn split (1:1) introduced approximately 0.4 to 0.5 ml/min of eluent to the mass spectrometer.
To obtain the electrospray ionization-liquid chromatography/mass spectrometry (ESI-LC/MS) of the isolated metabolites, aliquots of dissolved metabolites (in methanol or water) were introduced to the mass spectrometer using flow injection analyses method. The mobile phase consisted of approximately 1:1 v/v mixture of acetonitrile and 10 mM ammonium formate (pH 3.8) delivered at a rate of 0.3 ml/min.
High Field NMR.
The purified metabolites were dissolved in methanol-d4 and filtered to remove particulate matter. The structures of these metabolites were determined from proton and carbon 1-dimensional NMR as well as proton-proton two dimensional correlated spectroscopy, proton-carbon heteronuclear multiple quantum correlation (HMQC), and long range proton-carbon heteronuclear multiple bond correlation (HMBC) 2-dimensional NMR using a 400 MHz Varian VXR4S instrument.
Liquid Chromatography/NMR.
1H-LC/NMR was performed using a Bruker AMX-500 MHz NMR spectrometer equipped with a dedicated 1H flow-probe (probe flow cell of 4 mm i.d. with a volume of 120 μl). Stopped-flow 1H-NMR spectra were obtained at 500 MHz using a modified one dimensional version of the nuclear Overhauser effect Spectroscopy (NOESY) pulse sequence for solvent peak presaturation, which produced conditions for ensuring double solvent suppression. Stopped-flow spectra were acquired using 256 or 512 transients with 64K data points and a spectral width of 12,000 Hz. HPLC was performed using a Bruker LC22C pump and LC313 variable wavelength detector. The outlet of the UV detector was connected to the HPLC-NMR probehead via an inert polyethylether ketone (PEEK) capillary (0.25 mm i.d.). HPLC was performed on a C18 column (250 × 4.6 mm, 5 μm) using a gradient elution consisting of two components: (A) 0.1% TFA (deuterated) in D2O and (B) acetonitrile, (Pestnal/analytical grade, Riedel deHaen, Germany). The gradient consisted of increasing the percentage of B from 18 to 30 in 10 min followed by another increase to 40% B in 5 min. The flow rate was set at 1.0 ml/min.
The urine extract from solid phase extraction (SPE) cartridges containing the metabolite of interest, M1, was injected onto a HPLC column and 1H-LC/NMR data was obtained.
In Vitro Studies
Preparation of Liver and Kidney Subcellular Fractions.
The S9 and cytosolic fractions of livers and kidney were prepared from Sprague-Dawley rats (280–290 g b.wt.). The liver microsomes were prepared from rats treated with either phenobarbital (PB) or efavirenz. The rats were allowed food and water ad libitum before sacrifice. For the preparation of subcellular fractions, rats were treated with PB once a day for 3 days (75 mg/kg i.p.) or with efavirenz twice a day for 7 days (50 mg/kg p.o.). Rats treated with saline served as controls. The animals were sacrificed by cervical dislocation 24 h after the last administration of either drug or saline and the livers and kidneys were removed. The livers and kidneys of cynomolgus monkeys were obtained from Hazelton Texas Primate Center (Alice, TX). The 9000g (S9), cytosol, and microsomal fractions of rats and cynomolgus monkeys were prepared as described by Lake (1987). The concentrations of proteins of the subcellular fractions were measured by the Lowry method (Lowry et al., 1951). The microsomal cytochrome P-450 activity was assessed by testosterone hydroxylation assay (Sonderfan et al., 1987). The S9 fractions of human livers and kidneys were purchased from either XenoTech (Kansas City, KS), or from the International Institute for the Advancement of Medicine (Exton, PA). The activity of glutathione-S-transferases (GSTs) in the subcellular fractions was evaluated by either spectrophotometric assay of the reaction of GSH with 1,2-dichloronitrobenzene (Habig et al., 1974) or by LC/MS monitoring of the formation ofS-(p-nitrobenzyl)GSH usingp-nitrobenzyl chloride as the substrate (see below).
LC/MS Assay for GST Activity.
A LC/MS technique was developed to assay GST activity usingp-nitrobenzyl chloride as the substrate. The incubation consisted of 1 mM p-nitrobenzyl chloride and 5 mM GSH in a final volume of 1 ml of 0.1 M phosphate buffer (pH 6.5). The reaction was initiated by the addition of 1 mg of S9 proteins. The incubation was performed at 37°C for 15 min. An incubation with heat-deactivated S9 proteins was included as a control. The incubation was quenched by the addition of 2 ml of ice-cold acetonitrile, and the mixture was centrifuged at 1450g for 10 min. The supernatant was removed and diluted with 1 ml of water and loaded onto a C18 cartridge (500 mg/3 ml) previously conditioned with methanol and water. After the sample had eluted under gravity, the cartridge was sequentially washed with 1 ml of water and 1 ml of 5% methanol in water. The products were eluted from the cartridge with 2 ml of 70% methanol in water. The volume of the eluent was reduced to half under nitrogen at 30°C and aliquots were injected onto the LC/MS to detect the presence of GSH adducts. HPLC was performed on a small bore column (YMC C18-AQ, 2 × 100 mm, 3 μm) with an isocratic mobile phase consisting of a mixture of acetonitrile and 10 mM ammonium formate (pH 7.5) (3:7, v/v) delivered at 0.2 ml/min. The mass spectrometer was operated in the negative ion mode with full scans done from 100 to 600 amu. The conjugate, S-(p-nitrobenzyl)GSH was detected with [M-H]− atm/z 443 as the parent ion.
In Vitro Formation of M11 from M3.
Metabolite M3 [(S)-6-chloro-4-cyclopropylethynyl-8-sulfo-4-trifluoromethyl-1,4-dihydro-2H-3,1-benzoxazin-2-one] and M4 [(S)-6-chloro-4-cyclopropylethynyl-8-hydroxy-4-trifluoromethyl-1,4-dihydro-2H-3,1-benzoxazin-2-one] were incubated with either cynomolgus monkey or rat liver microsomes (untreated as well as PB- and efavirenz-treated) at 37°C for 60 to 90 min. The final incubation mixtures consisted of 10 or 25 μM either M3 or M4, 2 mg of microsomal protein, 3 mM MgCl2, and 2 mM NADPH in a 1-ml volume of 0.1 M phosphate buffer (pH 7.4). Control incubations included samples from which NADPH was excluded. The incubations were stopped by the addition of 2 ml of ice-cold acetonitrile followed by extraction of samples on SPE cartridges as described above. The metabolites were eluted with 2 ml of methanol from the C18 cartridges. After drying the organic phase, the residues were reconstituted in methanol/water (1:4, v/v) and analyzed by LC/MS as described in Liquid Chromatography/Mass Spectrometry. Standards of metabolites M8 and M11 (isolated in vivo) were analyzed for the purpose of comparing of HPLC retention times and mass spectral fragmentation patterns.
In Vitro Formation of M9 (GSH Conjugate) from M11.
Metabolite M11 was incubated at concentrations of 10, 25, 50, and 100 μM in the presence of liver or kidney subcellular fractions (S9 or cytosol) obtained from rats, cynomolgus monkeys, or humans. The incubations, in duplicates, consisted of 5 mg of protein, MgCl2 (3 mM), and GSH (5 mM) in a final volume of 1 ml of 0.1 M sodium phosphate buffer (pH 7.4). Control incubations were carried out with boiled proteins. The incubation mixtures were agitated gently in a metabolic shaker at 37°C for 4 h after which the reaction was quenched by the addition of 2 ml of ice-cold acetonitrile followed by centrifugation at 1450g for 15 min. The supernatant was removed, evaporated to dryness under nitrogen, and the residue was diluted with 1 ml of water. The sample was then extracted on a C18 cartridge (500 mg/3 ml) as described above. The eluent from the cartridge (2 ml of methanol) was evaporated under nitrogen and the residue was analyzed by LC/MS. The GST activities in each subcellular fraction were evaluated using 1,2-dichloronitrobenzene or p-nitrobenzyl chloride. LC/MS analyses of the incubation extracts were performed on an analytical column as described inLiquid Chromatography/Mass Spectrometry. LC/MS comparisons were made with previously identified standards M9, M10, and M17.
In Vitro Formation of M9 from M11 in the Presence of Rat Liver GSTs.
Metabolite M11 was incubated in the presence of partially purified GSTs (Sigma Chemical Co., St. Louis, MO) isolated from rat livers. The incubation consisted of 100 μM M11 and 2 mg of protein reconstituted in a final volume of 1 ml of 0.1 M phosphate buffer. The mixture was incubated in a metabolic shaker for 1 h at 37°C after which it was extracted on a C18 cartridge (500 mg/10 ml) as described above. The final eluting solvent was 2 ml of 30% methanol in water. The sample was reduced in volume (0.5 ml) and an aliquot injected onto LC/MS column. The formation of GSH conjugate (M9) was monitored using conditions described above.
In Vitro Formation of M10 from M9.
Metabolite M9 was incubated at a concentration of 100 μM in the presence of liver or kidney subcellular fractions (S9 or cytosol) obtained from rats. The incubations, done in duplicate, consisted of 5 mg of protein and MgCl2 (3 mM) in a final volume of 1 ml of 0.1 M sodium phosphate buffer (pH 7.4). Control incubations were carried out with boiled proteins. The incubation mixtures were agitated gently in a metabolic shaker at 37°C for 1 h. At the end of the incubation, the mixtures were extracted by SPE and the extracts were analyzed by LC/MS using conditions described inLiquid Chromatography/Mass Spectrometry.
Hydrolysis of Metabolite M10 using Rat Kidney Microsomes.
Metabolite M10, which was isolated previously from the urine of rats, was incubated in the presence of kidney microsomes obtained from rats treated with either efavirenz or with saline. The incubation was done in 1 ml of 0.1 M phosphate buffer (pH 7.4) using either 10 or 100 μM M10 in the presence of 2 mg of microsomal protein. Multiple incubations of M10 in the presence of rat kidney microsomes (2 mg/ml) were performed to generate sufficient quantities of metabolite for spectroscopic characterization. After incubating for 2 h at 37°C in a metabolic shaker, the reaction was quenched with 2 ml of ice-cold acetonitrile. The samples were centrifuged at 1450g for 10 min and the supernatants removed and concentrated under nitrogen at 30°C. The residues were diluted with 1 ml of water and loaded onto C18 cartridges (500 mg/3 ml), which were preconditioned with methanol and water. After the samples had eluted under gravity, the columns were eluted subsequently with 2 ml of water followed by 2 ml of 30% methanol in water. The eluent containing the product was dried under vacuum and subsequently purified on a C18 semipreparative column (Beckman C18, 250 × 10 mm, 5 μm) using acetonitrile and 10 mM ammonium formate (pH 4.1) as the mobile phase. The separation of metabolites was done by a gradient consisting of ramping acetonitrile from 25 to 80% over 20 min at a flow rate of 4 ml/min. Metabolite M17 (the product from enzymatic hydrolysis of M10) eluting at 6 min was collected from several injections, dried, and submitted for LC/MS and NMR analyses.
In Vivo Studies
Urine Samples.
Rat study
Urine samples from a number of studies were analyzed by LC/MS. These included samples from groups of rats given increasing doses of efavirenz. The rats were dosed either with 10, 50, 250, or 500 mg/kg once daily for 3 days. In another study 10 male rats (weighing approx. 250–400 g, CD strain) were dosed orally with efavirenz suspension (50 mg/ml made in 0.5% methylcellulose) once daily at 250 mg/kg for 3 days, then twice daily for the next 7 days. The dosing volume was 5 ml/kg. Bile and urine samples were collected over ice and pooled from all the animals on a daily basis and stored frozen at −20°C until analyzed. These two sets of studies involved bile duct-cannulated rats. Urine samples from noncannulated rats were also analyzed from studies in which the animals were given single oral doses of efavirenz at 30, 100, 300, 500, or 700 mg/kg after a period of induction (30 mg/kg for 3 days before the final dose). The urine samples were either collected over a 0- to 24-h period or at intervals of 0- to 4-, 4- to 8-, and 8- to 24-h. Aliquots of urine (20–50 μl) were injected directly onto the HPLC column for analyses by LC/MS.
Guinea pig study.
To isolate sufficient quantities of metabolites a total of six guinea pigs (3 males and 3 females) were given single oral dose of efavirenz (250 mg/kg/day for 3 days) and urine samples (0–24, 24–48, and 48–72 h) collected over ice in metabolic cages.
Hamster study.
Hamsters (6/sex/group) were administered efavirenz (either 50 or 250 mg/kg/day) once daily for 7 days followed by twice daily dosing for the subsequent 7 days. Urine samples were collected on day 12 (0–24 h) and stored frozen until analyzed.
Cynomolgus monkey study.
Urine samples were obtained from cynomolgus monkeys given 75 mg/kg twice daily of efavirenz over 7 days. Samples were collected from these monkeys by catheterizing the urinary bladder at the termination of the experiment. Urine samples were also obtained from safety assessment studies in which monkeys were administered oral doses of efavirenz at 75 mg/kg, twice daily.
Human study.
Urine samples were obtained from clinical studies in which subjects were given daily doses of either 2 × 200, 2 × 400, or 1600 mg. LC/MS analyses were done either by direct injection of the urine onto the HPLC column or by concentrating 5 ml of urine on C18 cartridges and then subsequently injecting an aliquot (30 of 100 μl) of the reconstituted sample.
Plasma Samples.
Plasma samples were obtained from rats given 800, 900, and 1,000 mg/kg of efavirenz (single oral dose) 2-h postdose. To 0.3 ml of plasma, 2 ml of methanol was added to precipitate the protein. After centrifuging the samples at 1450g for 5 min, the clear supernatant was removed and dried under a stream of nitrogen at 37°C. The extract was reconstituted in 130 μl of methanol/water (3:1, v/v), centrifuged, and aliquots (40 μl) were injected onto an HPLC column. LC/MS analyses of the extracts were carried out using conditions described above.
Plasma samples from at least four human subjects were analyzed by LC/MS. The dose was 1200 mg and plasma collected at 4- to 6-h postdose. To 0.5 ml of plasma, 1 ml of water was added and the sample loaded onto a preconditioned C18 cartridge (500 mg/3 ml). The cartridges were sequentially conditioned with 2 ml of methanol and 2 ml of water. After the sample had eluted under gravity, the cartridge was washed with 1 ml of water, followed by 1 ml of methanol. The organic phase was collected and dried under nitrogen. The residue was reconstituted in 100 μl of methanol of which 30 μl was injected onto the HPLC/MS.
Isolation and Characterization of In Vivo Metabolites
Initial Urine Cleanup.
Urine samples from rats, guinea pigs, and cynomolgus monkeys were cleaned initially on MegaBond Elut C18 cartridges (10 g/60 ml, Varian Sample Preparation Products). The cartridges were conditioned with 20 ml of methanol followed by 20 ml of water. The urine samples were loaded onto the cartridges and eluted under gravity. The cartridges were subsequently washed with 20-ml aliquots of varying percentages of methanol in water (0, 5, 15, 25, 35, 45, 55, 65, 75, and 100% methanol). Each fraction was analyzed by LC/MS for the presence of efavirenz metabolites. The fractions containing the appropriate metabolite(s) were pooled and dried under vacuum. Additional purification of individual metabolites was carried out on HPLC and C18 cartridges as described below.
Metabolite M1 (average molecular weight of 507 with [M-H]− at m/z 506).
This metabolite was isolated previously from rat urine (Christ et al., 1997). Additional confirmation of this metabolite included characterization by on-line LC/NMR (stopped-flow) of rat urine extract (as described above). The identity of this metabolite was confirmed by comparing its retention time and mass spectral fragmentation with synthetic standard of this conjugate.
Metabolite M2 (average molecular weight of 491 with [M-H]− at m/z 490).
M2 was not isolated from biological samples because it was present in small quantities and was difficult to purify in sufficient quantities for NMR analyses.
Metabolite M3 (average molecular weight of 411 with [M-H]− at m/z 410).
Metabolite M3 was isolated from guinea pig urine. The 65% methanol fraction from the initial cleanup of urine showed the presence of this conjugate along with other metabolites. The combined extract was rechromatographed on C18 cartridge and the fractions containing metabolite M3 were pooled (55–60% methanol in water). After removing the solvents under vacuum, the dried powder was reconstituted in methanol and 100-μl aliquots were injected onto a semipreparative C18 HPLC column (250 × 10 mm; Beckman). The mobile phase consisted of a mixture of acetonitrile and 10 mM ammonium formate (pH 3.8) (45:55, v/v) delivered at a rate of 4.5 ml/min. Metabolite M3 was eluted with a retention time of 8 to 10 min. The fractions corresponding to this metabolite were pooled, dried, and rechromatographed on a C18 cartridge to remove any salts. The metabolite was eluted from the cartridge using 70% methanol in water. An aliquot of the extract (20 μl) was injected onto the LC/MS to check purity before NMR analyses. The structure of this metabolite was also confirmed by comparing its retention time and mass spectral fragmentation pattern with that of a synthetic standard.
Metabolite M4 (average molecular weight of 331 with [M-H]− at m/z 330).
The identity of this metabolite present in urine of various species was confirmed by comparing its retention time and mass spectral fragmentation pattern with the synthetic standard.
Metabolite M5 (average molecular weight of 331 with [M-H]− at m/z 330).
The identity of this metabolite was confirmed by comparing its retention time and mass spectral fragmentation pattern with synthetic standard.
Metabolite M6 (average molecular weight of 507 with [M-H]− at m/z 506).
Metabolite M6 was isolated from guinea pig urine. The 65% methanol extract from the initial chromatography on C18 was dried and rechromatographed on a MegaBond C18 cartridge. Additional chromatography of this fraction on Bond-Elut C18 cartridge using increasing percentages of methanol in water was done. The 7-OH efavirenz glucuronide conjugate was eluted with 60% methanol in water. LC/MS analyses showed the presence of 7-OH efavirenz glucuronide and the free aglycone (7-OH efavirenz) indicating the possibility of cleavage of glucuronic acid from the conjugate during the chromatography. The fraction containing this metabolite was dried under vacuum and the extract was subjected to additional purification by semipreparative HPLC (Beckman C18, 250 × 10 mm) using a mixture of acetonitrile and 10 mM ammonium formate, pH 3.5 (3:2, v/v) as the mobile phase. The flow rate was 4.5 ml/min. Metabolite M6 eluted with a retention time of 7.5 min. The fractions containing this metabolite were pooled, dried, and rechromatographed on a C18 cartridge (eluted using 30, 50, and 70% methanol in water, with the metabolite M6 eluting in the last fraction). The fraction containing the metabolite was dried and analyzed by LC/MS and NMR.
Metabolite M7 (average molecular weight of 411 with [M-H]− at m/z 410).
Metabolite M7 was isolated from guinea pig urine (55% methanol extract). The extract was reconstituted in 1:1 (v/v) methanol/water and 100-μl aliquots were injected onto a semipreparative HPLC column (Waters Symmetry C18, 300 × 7.8 mm). The mobile phase consisted of 1:1 (v/v) mixture of acetonitrile/10 mM ammonium formate (pH 3.8). The flow rate was 3.3 ml/min. The fractions corresponding to metabolite M7 were pooled and dried under vacuum. The dried extract was reconstituted in water and loaded onto a C18 cartridge and eluted with 45% methanol in water. The solvents were removed under vacuum and the dried sample was analyzed by LC/MS before NMR analyses.
Metabolite M8 (average molecular weight of 347 with [M-H]− at m/z 346).
Metabolite M8 was isolated from guinea pig urine. After the initial cleanup of the urine, the 45% methanol fraction was further purified on C18 cartridges using different percentages of methanol in water as the eluent. Metabolite M8 was eluted with 40% methanol in water. The solvent was removed and the extract chromatographed on a semipreparative column (Beckman C18, 250 × 10 mm, 5 μm) using a mobile phase consisting of acetonitrile and 0.1% acetic acid. The percentage of acetonitrile was increased from 25 to 80% over 20 min at a flow rate of 5 ml/min. Metabolite M8 was eluted from the column with a retention time of 11.5 min. After removing the solvent under vacuum, the dried residue was submitted for mass spectral and NMR analyses.
Metabolite M11 (average molecular weight of 427 [M-H]− at m/z 426).
Metabolite M11 was isolated from guinea pig and cynomolgus monkey urine and was also obtained from β-glucuronidase hydrolyses of a previously isolated metabolite (the sulfate/glucuronide diconjugate, Metabolite M12 (Mutlib et al., 1998). Metabolite M11 was purified on a Beckman C18 column using identical HPLC conditions as described for Metabolite M8. This compound was eluted from the column with a retention time of 5.5 min. After removing the solvent under vacuum, the dried extract was analyzed by LC/MS and NMR.
Metabolite M14 (average molecular weight of 523 [M-H]− at m/z 522).
This metabolite was isolated from guinea pig urine. The 35% methanol extract (obtained from the initial cleanup of the guinea pig urine) was rechromatographed on a C18 cartridge and the fractions corresponding to this metabolite were pooled and dried under vacuum. Additional chromatography of this extract was done on a semipreparative column (Beckman C18, 250 × 10 mm) using a mixture of acetonitrile and water (27:73, v/v). The pH of the aqueous phase was adjusted to 3.3 with acetic acid. The flow rate was 5 ml/min. Metabolite M14 was eluted with a retention time of approximately 10 min. The fractions corresponding to this metabolite were pooled, dried, and rechromatographed on the same HPLC column using a mixture of methanol and water (60:40, v/v) with the aqueous pH adjusted to 4.7 with acetic acid. The metabolite was eluted from the column with a retention time of 4.3 min with the solvent delivered at a rate of 4.5 ml/min.
Metabolite M15 (average molecular weight of 427 [M-H]− at m/z 426).
Metabolite M15 was isolated from guinea pig urine using the same procedure as described for Metabolites M8 and M11. This compound was eluted from the C18 (Beckman, 250 × 10 mm) column with a retention time of 5.0 min. The HPLC solvents were subsequently removed under vacuum and the residue was analyzed by LC/MS and NMR.
Metabolite M16 (average molecular weight of 523 [M-H]− at m/z 522).
This metabolite was isolated from guinea pig urine. The 45% methanol extract was rechromatographed on a C18 cartridge and the fractions containing this metabolite were pooled and dried under vacuum. Additional chromatography of this extract was done on a semipreparative C18 column using acetonitrile and 10 mM ammonium formate (pH 4.1) as the solvent. The gradient consisted of increasing the percentage of acetonitrile from 25 to 80 over 20 min at a flow rate of 5 ml/min. The metabolites were monitored with a variable wavelength UV detector set at λ = 254 nm. The fraction containing Metabolite M16 was further purified on an analytical column (Beckman C18, 250 × 4.6 mm) using a mixture of acetonitrile and 0.1% acetic acid as the mobile phase. The percentage of acetonitrile was increased from 10 to 90 over 40 min at a flow rate of 1 ml/min. Metabolite M16 was eluted from the analytical column with a retention time of 16.3 min. The HPLC solvent was removed and the residue was analyzed by LC/MS and NMR.
Metabolite M17 (average molecular weight of 548 [M-H]− at m/z 547).
This metabolite was isolated from guinea pig urine and LC/MS data were compared with a standard obtained from GGT-catalyzed hydrolysis of M10.
Results
Characterization of Metabolites.
Metabolite M1
This metabolite has been characterized previously (Christ et al., 1997). Metabolite M1 was found to have identical LC/MS retention time and mass spectral fragmentation pattern as the synthetic standard of 8-OH efavirenz glucuronide. The mass spectral data from the ESI-LC/MS of metabolite M1 is given in Table 1. The major fragment ion (m/z 330) was formed by the characteristic loss of glucuronic acid (−176 amu) from the parent ion (Mutlib and Abbott, 1992). Metabolite M1 was subsequently isolated from rat urine and 1H-NMR obtained and compared with that of synthetic standard. Metabolite M1 was identified as [(S)-6-chloro-4-(cyclopropylethynyl)-1,4-dihydro-2-oxo-4-(trifluoromethyl)-2H-3,1-benzoxazin-8-yl]β-d-glucopyranose.
Metabolite M2.
Metabolite M2 was found to produce [M-H]− atm/z 490. The ESI-LC/MS spectrum (Table 1) showed fragment ions at m/z 314 (aglycone formed by the loss of glucuronic acid −176 amu) and 244. Because this metabolite was formed by direct conjugation of efavirenz with glucuronic acid, it was postulated that this was an N-glucuronide (only position available for conjugation with glucuronic acid).N-glucuronides of compounds have been reported before (Maguire et al., 1982; Mohri et al., 1985; Mutlib and Nelson, 1990). Because this metabolite was produced in low quantities and was found to coelute with M1, no attempts were made to isolate M2 for NMR analyses. M2 was identified as [(S)-6-chloro-4-(cyclopropylethynyl)-1,4-dihydro-2-oxo-4-(trifluoromethyl)-2H-3,1-benzoxazin-1-yl]β-d-glucopyranose.
Metabolite M3.
The sulfate conjugate of 8-OH efavirenz (M3) was isolated from guinea pig urine. The mass spectral data showed [M-H]− at m/z 410. The ESI-LC/MS showed fragment ions at m/z 330, 286, 258, 249, 246, 230, 210, 182, and 162 (Table 1). The loss of 80 amu from the parent ion suggested a sulfate conjugate (Weidolf et al., 1988). The 1H-NMR (Table2) showed an intact cyclopropyl group at δ 0.80 and 0.96 (-CH2-, 4H), and at δ1.48 (CH, m, 1H). The aromatic protons [δ 7.25 (d) and 7.60 (d)] showed a characteristic small m-coupling (1 Hz). The identity of this metabolite was confirmed by comparing its LC/MS retention time and mass spectral fragmentation pattern with that of a synthetic standard. Hence, this metabolite was confirmed as (S)-6-chloro-4-(cyclopropylethynyl)-8-sulfo-4-(trifluoromethyl)-1,4-dihydro-2H-3,1-benzoxazin-2-one.
Metabolite M4.
The identity of metabolite M4 was confirmed by comparing its LC/MS retention time and mass spectral fragmentation pattern with a synthetic standard of 8-OH efavirenz. The ESI-LC/MS showed [M-H]− at m/z 330 with major fragment ions at m/z 286, 266, 258, 250, 246, 230, and 210 (Table 1). The origin of some of these fragment ions was attributed to neutral loss of CO2 from the molecule with subsequent losses of HCl and HF. For example, ion atm/z 286 is formed by the loss of CO2 from m/z 330. An additional neutral loss of HCl or HF leads to ions atm/z 250 and 266, respectively. The origin of some of these fragment ions is postulated in Fig.2. This metabolite was confirmed as (S)-6-chloro-4-(cyclopropylethynyl)-8-hydroxy-4-(trifluoromethyl)-1,4-dihydro-2H-3,1-benzoxazin-2-one.
Metabolite M5.
This metabolite was found to produce a mass spectral fragmentation pattern similar to 8-OH efavirenz (metabolite M4) except for the distinct reduction in the intensity of the molecular ion atm/z 330 (see Table 1). The compound showed MH+ at m/z 332 when the mass spectrometer was operated in the positive ion mode. The identity of this metabolite was confirmed by comparing its retention time and mass spectral fragmentation pattern with that of a synthetic standard. Metabolite M5 was confirmed as (S)-6-chloro-4-(cyclopropylethynyl)-7-hydroxy-4-(trifluoromethyl)-1,4-dihydro-2H-3,1-benzoxazin-2-one.
Metabolite M6.
The glucuronide conjugate of 7-OH efavirenz, metabolite M6, was isolated from guinea pig urine. The mass spectral data (Table 1) showed a fragmentation pattern similar to that of metabolite M1 except that the ion at m/z 330 (corresponding to the loss of glucuronic acid from the parent ion) was very weak. This observation was consistent with the mass spectral data of other 7-OH efavirenz metabolites where the fragment ion at m/z 330 was very weak or totally absent. The 1H-NMR data (Table 2) of this metabolite showed that substitution on the aromatic ring was at C7 (as evidenced by lack of coupling between the C5 and C8 aromatic protons). The cyclopropyl protons were intact at δ 0.86, 0.98 ppm (-CH2-, m, 4H) and at δ 1.51 (CH, m, 1H). The anomeric proton (H-1") of the glucuronic acid was at δ 4.91 as a doublet. The other protons of the glucuronic acid moiety were at δ 3.93 (H-5", d, 1H), 3.72 (H-4", t, 1H), 3.60 (H-3", t, 1H), and at 3.48 (H-2", t, 1H). Metabolite M6 was unambiguously identified as the glucuronide conjugate of 7-OH efavirenz, [(S)-6-chloro-4-(cyclopropylethynyl)-1,4-dihydro-2-oxo-4-(trifluoromethyl)-2H-3,1-benzoxazin-7-yl]β-d-glucopyranose.
Metabolite M7.
The sulfate conjugate of 7-OH efavirenz, metabolite M7, was also isolated from guinea pig urine. The ESI-LC/MS showed [M-H]− at m/z 410 (Table1) with fragment ions at m/z 286 and 258 (base peak). The expected m/z for the aglycone (m/z 330) was missing as expected for 7-hydroxylated analogs of efavirenz. The structure of this metabolite was confirmed by 1D 1H-NMR, HMQC, and HMBC 2D NMR. Proton NMR (Table 2) showed that the cyclopropyl group was intact as seen by signals at δ 0.80, 0.96(-CH2-, m, 4H), and at δ1.50 (CH, m, 1H). The 2D experiments showed that the acetylene group was also intact. The lack of coupling between the two aromatic protons indicated substitution at C7 producing a 1H-NMR spectrum with aromatic signals as singlets at δ 7.28 and 7.47 ppm. Metabolite M7 was confirmed as (S)-6-chloro-4-(cyclopropylethynyl)-7-sulfo-4-(trifluoromethyl)-1,4-dihydro-2H-3,1-benzoxazin-2-one.
Metabolite M8.
The isolated metabolite showed characteristic ion fragment atm/z 262 (see Table 1) with the [M-H]− at m/z 346. Other characteristic ion fragments included ions at m/z302, 282, 274, 264, 262, 254, 246, 242, and 226. The presence of only one hydroxyl group on the aromatic ring was confirmed by the mass spectral fragmentation pattern. The characteristic ion atm/z 302 is formed by the loss of CO2 from the parent ion. Subsequent losses of HF (−20 amu) from the ion m/z 302 leads to fragment ions at m/z 282, 262, and 242. The structure of this metabolite was confirmed by 1H-NMR data (Table 2), which showed the existence of two aromatic proton signals at δ 7.0 and 6.93 ppm. The protons at C5 and C7 showed a small coupling (<1 Hz) as expected from m-coupled aromatic protons. The presence of hydroxyl group at C14 was confirmed by1H-NMR data, which showed a characteristic absence of the multiplet at δ 1.5 ppm normally present in the spectrum of efavirenz (the signal corresponds to the methine coupled to the -CH2- groups of the cyclopropyl side chain. Metabolite M8 was confirmed as (S)-6-chloro-4-(1-hydroxycycloprop-1-yl)ethynyl-8-hydroxy-4-(trifluoromethyl)-1,4-dihydro-2H-3,1-benzoxazin-2-one.
Metabolites M9 and M10.
Characterizations of M9 and M10 have been reported previously (Mutlib et al., 1998a, 1999). Metabolites 9 and 10 were identified as (S)-6-chloro-8-sulfo-4-(2-(1-hydroxycyloprop-1-yl)-2-(S-(γ-glutamylcysteinylglycinyl))ethen-1-yl)-4-(trifluoromethyl)-1,4-dihydro-2H-3,1-benzoxazin-2-one and (S)-6-chloro-8-sulfo-4-(2-(1-hydroxycyloprop-1-yl)-2-(S-cysteinylglycinyl) ethen-1-yl)-4-(trifluoromethyl)-1,4-dihydro-2H-3,1-benzoxazin-2-one, respectively.
Metabolite M11.
This metabolite was isolated from guinea pig and cynomolgus monkey urine. The mass spectrum of this compound showed [M-H]− at m/z 426, which produced a characteristic loss of 80 amu (-SO3), indicating the presence of a sulfate moiety (Table 1). Other fragment ions included ions at m/z 282, 274, 262 (base peak), 254, 242, 238, and 226. The fragmentation pattern was identical with that obtained from metabolite M8, suggesting that this was a sulfate conjugate of metabolite M8. The nature of the fragment ions is illustrated in Fig. 3. Metabolite M11 was also generated by incubating the previously isolated mixed diconjugate (metabolite M12; see Mutlib et al., 1998a, 1999) with β-glucuronidase (free of sulfatase activity). The LC/MS retention time and mass spectral fragmentation of the released aglycone matched that of the metabolite isolated from urine. The 1H-NMR data (Table 1) was very similar to that obtained for metabolite M8 except for the slight downfield shift of the aromatic proton signals of metabolite M11 (perhaps due to the deshielding by the sulfate group on C8). The aromatic protons showed small coupling (m-coupling of less than 3 Hz) as expected. The cyclopropane C15 and C16 protons were found as multiplets at 1.02 and 1.08 ppm. The C14 proton was distinctly absent from the spectrum. Metabolite M11 was confirmed as (S)-6-chloro-4-(1′-hydroxycyclopropylethynyl)-8-sulfo-4-trifluoromethyl-1,4-dihydro-2H-3,1-benzoxazin-2-one.
Metabolites M12 and M13.
Characterizations of M12 and M13 have been reported previously (Mutlib et al., 1998a, 1999). Metabolites M12 and M13 were identified as (S)-6-chloro-8-sulfo-4-(2-(1-glucopyranosyl-cycloprop-1-yl)ethynyl)-4-(trifluoromethyl)-1,4-dihydro-2H-3,1-benzoxazin-2-one and (S)-6-chloro-8-sulfo-4- (2-(1-oxo-3-hydroxypropyl)-2-(S-(γ-glutamylcysteinylglycinyl) ethen-1-yl)-4-(trifluoromethyl)-1,4-dihydro-2H-3,1-benzoxazin-2-one, respectively.
Metabolite M14.
A metabolite with [M-H]− atm/z 522 was isolated from guinea pig urine. The ESI-LC/MS showed [M-H]− atm/z 522 with fragment ions atm/z 346 (base peak), 302, 282, 262, 254, 242, and 226. The ion at m/z 346 corresponded to a loss of glucuronic acid moiety (−176 amu). The 1H-NMR data (Table 2) showed the presence of two coupled aromatic protons at δ 7.24 and 7.40 (<3 Hz for m-coupled protons), indicating hydroxylation at C8 position. The methine of the cyclopropyl ring was substituted with a hydroxyl group, giving a1H-NMR spectrum that showed the cyclopropyl protons appearing at δ 1.02 and 1.08 (-CH2-, 4H). The glucuronic acid protons were present at δ 5.01 (H-1", s, 1H), 4.10 (H-5", d, 1H), and 3.50 to 3.80 (H-2", H-3", and H-4", m, 3H). Additional NMR experiments, using gradient HMQC and HMBC, showed that the glucuronic acid was attached to the hydroxyl group on C8 instead of C14. Metabolite M14 was identified as (S)-6-chloro-4-[(1-hydroxycycloprop-1-yl)ethynyl]-1–4-dihydro-2-oxo-4-(trifluoromethyl)-2H-3,1-benzoxazin-8-yl]β-d-glucopyranose.
Metabolite M15.
This metabolite was isolated from guinea pig urine. The mass spectrum of this compound showed [M-H]− atm/z 426, which produced a characteristic loss of 80 amu (-SO3), indicating the presence of a sulfate moiety (Table 1). Other fragment ions included ions atm/z 408, 302, 282, 254, 242, 227, 191, 173, and 155. The 1H-NMR showed that a substitution at C7 had taken place as shown by the appearance of singlets in the aromatic region. The cyclopropane C15 and C16 protons were found at 1.02 and 1.08 ppm. The C14 proton was distinctly absent from the spectrum. Metabolite M15 was confirmed as (S)-6-chloro-7-sulfo-4-[(1-hydroxycycloprop-1-yl)ethynyl]-4-trifluoromethyl-1,4-dihydro-2H-3,1-benzoxazin-2-one.
Metabolite 16.
The ESI-LC/MS showed [M-H]− atm/z 522 with fragment ions atm/z 346 (weak), 302, 282, 262, 254, 246, and 226. The ion at m/z 346 corresponded to a loss of glucuronic acid moiety (−176 amu). The 1H-NMR data (Table 2) showed the presence of two aromatic protons appearing as singlets at δ 6.83 and 7.42, indicating hydroxylation at the C7 position. The methine of the cyclopropyl ring was substituted with a hydroxyl group, giving a 1H-NMR spectrum that showed the cyclopropyl protons appearing at δ 1.02 and 1.08 (-CH2-, 4H). The glucuronic acid protons were present at δ 4.98 (H-1", s, 1H), 3.90 (H-5", d, 1H), and 3.50 (H-2", m, 1H), 3.58 to 3.70 (H-3" and H-4", m, 2H). Additional NMR experiments, using gradient HMQC and HMBC, showed that the glucuronic acid was attached to the hydroxyl group on C7 instead of at C14. Metabolite M16 was identified as (S)-6-chloro-4-[(1-hydroxycycloprop-1-yl)ethynyl]-1–4- dihydro-2-oxo-4-(trifluoromethyl)-2H-3,1-benzoxazin-7-yl]β-d-glucopyranose.
Metabolite M17.
This was the cysteine conjugate found in guinea pig urine. The ESI-LC/MS mass spectral data showed [M-H]− atm/z 547 with the major fragment ions atm/z 467, 380, 296, 282, 266, 264, and 246 (see Table 1). This metabolite was confirmed as the cysteine conjugate by incubating metabolite M10 with rat kidney microsomes containing dipeptidase activity and isolating metabolite M17 as a product. Metabolite M17 was tentatively identified as (S)-6-chloro-8-sulfo-4-(2-(1-hydroxycycloprop-1-yl)-2-(S-(cysteinyl)ethen-1-yl)-4-(trifluoromethyl)-1,4-dihydro-2H-3,1-benzoxazin-2-one.
Metabolite Profiles of Efavirenz in Urine of Rats, Guinea Pigs, Hamsters, Cynomolgus Monkeys, and Humans.
Metabolite profiles of efavirenz in urine of rats, cynomolgus monkeys, humans, and guinea pigs dosed with efavirenz are depicted in schemesFS1, FS2, FS3, andFS4, respectively. The species differences in the formation of metabolites M1 to M17 are shown in Tables3 and 4.
The glucuronide conjugate of 8-OH efavirenz (M1) was the major metabolite excreted in the urine of all the species studied. This metabolite was easily isolated from urine and characterized unambiguously as the glucuronide conjugate (Christ et al., 1997) by comparing its HPLC retention time and mass spectral fragmentation pattern with the synthetic standard. The glucuronide conjugate of 7-OH efavirenz (M6) was found as a significant metabolite in urine of hamsters, guinea pigs, cynomolgus monkeys, and humans. Interestingly, metabolite M6 was totally absent in rat urine. The direct conjugation of efavirenz with glucuronic acid gave theN-glucuronide (M2) (with [M-H]− atm/z 490), which was present in urine of all the species studied. This was a significant metabolite in urine of animals given the first dose of efavirenz (data not shown). After multiple dosing with efavirenz, the 8-OH efavirenz glucuronide (M1) became the predominant metabolite. In bile duct-cannulated and noncannulated rats, the N-glucuronide was clearly seen as the major metabolite in urine and bile at early time points of collection (day 1, 0–4 h). β-Glucuronidase hydrolysis of the isolated N-glucuronide conjugate resulted in the release of free efavirenz.
The sulfate conjugate, M3, was not found in urine of humans, but was present in other species. Hamsters, guinea pigs, cynomolgus monkeys, and humans showed significant amounts of metabolite M7 (the regioisomer of M3) in urine. Metabolite M7 was totally absent in rat urine.
It appears from Table 3 that all species were capable of hydroxylating at C14 and at C8. However, because efavirenz was rapidly converted to 8-OH efavirenz (metabolite M4), secondary metabolites with a hydroxyl group at both C8 and C14 were commonly found in urine of different species. The unconjugated dihydroxylated metabolite M8 (C8 and C14 hydroxylated efavirenz) was found in urine of guinea pigs and cynomolgus monkeys. The sulfate conjugate (the position of sulfation was deduced to be on the C8 hydroxyl group) of this metabolite (resulting in metabolite M11) was found in significant quantities in urine of hamsters, guinea pigs, and cynomolgus monkeys. This metabolite was not found in humans and was present in only small quantities in rat urine. The glucuronide conjugate of the C8, C14 dihydroxylated metabolite (M14), was detected in urine of cynomolgus monkeys and humans. None of this metabolite was detected in urine of rats.
GSH-Related Adducts of Efavirenz in Urine of Rats.
The GSH adduct (M9) was previously identified and characterized in bile of rats given efavirenz (Mutlib et al., 1998a, 1999). Rats given very high doses of efavirenz (700 mg/kg, p.o.) were found to excrete small quantities of the GSH adduct (M9) in urine. This conjugate was not present in urine of hamsters, cynomolgus monkeys, or humans. Rats and guinea pigs were the only species that excreted cysteinylglycine (M10) and cysteine adducts (M17), which are breakdown products of metabolite M9, in urine.
In Vitro Formation of GSH Adduct, M9, from Metabolite M11.
The in vitro conversion of M11 to M9 was determined in the presence of liver and kidney subcellular fractions obtained from rats, cynomolgus monkeys, and humans. The activities of GSTs present in these preparations were confirmed using 1,2-dichloronitrobenzene orp-nitrobenzyl chloride as the substrates. It was found that the rat liver was capable of transforming M11 to M9 (see Fig.4). Incubation of M11 with S9 fractions from either cynomolgus monkey or human livers did not produce any M9. In the presence of rat kidney S9 fractions, M11 was converted to mostly M10 (cysteinylglycine conjugate), which is a breakdown product of M9. Trace quantities of M9 were found in the extracts. The kidney S9 fraction from either cynomolgus monkeys or humans did not generate any GSH or cysteinylglycine adducts.
In Vitro Formation of M10 from M9.
Metabolite M10 (cysteinylglycine) was produced when M9 (GSH adduct) was incubated with rat kidney cytosol and S9 fractions. The metabolite (M10) generated from M9 was found to have identical retention time and mass spectral fragmentation as the isolated standard.
In Vitro Formation of M17 from M10.
Metabolite M17 was formed when M10 was incubated with rat kidney microsomes. LC/MS analyses of the microsomal extract showed the presence of only one product (M17) in addition to the unchanged substrate. The mass spectral data showed [M-H]− at m/z 547 consistent with the cleavage of glycine (−57 amu) from the cysteinylglycine conjugate (M10).
In Vitro Metabolism of M3 to M11.
Microsomal incubations were done to demonstrate the formation of M11 from M3. Liver microsomes from saline-, PB-, and efavirenz-treated rats were used in this study. The biotransformation of M3 to M11 was found to be NADPH dependent. Greater quantities of M11 were produced from M3 in the presence of PB- and efavirenz-induced microsomes than the control (saline-treated) preparations; however, the overall yield was low as determined by LC/MS analyses.
Metabolites Present in Plasma.
Plasma samples from rats given high doses of efavirenz (800, 900, and 1000 mg/kg) were analyzed by LC/MS. It was found that two metabolites, M1 and M3, were present in the plasma extract. Trace quantities of metabolite M2 were also detected in rat plasma. The identities of these metabolites were confirmed by comparing the LC/MS retention times and mass spectral data with authentic standards.
Analyses of human plasma samples showed mostly unchanged efavirenz. Other metabolites present in trace quantities included metabolites M1, M2, M4, M6, and M7.
Discussion
The profiling and identification of efavirenz metabolites in urine of different species was facilitated by the presence of chlorine isotope (M + 2) in the mass spectra. This enabled us to detect the drug-related products in urine samples with ease. The presence of these metabolites was later confirmed by radiolabeled studies.
Efavirenz was metabolized extensively by all five species. The proposed metabolic pathways for efavirenz in four species are given in schemes FS1to FS4. The metabolic pathways in rats leading to the formation of mono- and diconjugated metabolites are proposed in scheme FS1. Initial hydroxylation on the aromatic ring of efavirenz leads to the formation of M4, which is subsequently conjugated with either glucuronic or sulfuric acids to form M1 and M3, respectively. As supported by in vitro results, M3 could be further hydroxylated to form M11, which is further converted by a specific GST to the GSH conjugate, M9. M11 could also be glucuronidated to form M12, producing a mixed glucuronide-sulfate diconjugate. The GSH adduct, M9, is subsequently degraded by GGT present in rat kidneys to form M10, which was found as a significant metabolite in urine of rats given high doses of efavirenz.
Scheme FS2 shows the formation of efavirenz metabolites in cynomolgus monkeys. Compared with rats, no GSH-related adducts were found in urine of cynomolgus monkeys (Table 4). Significant quantities of M8 and M11 (precursor to the GSH adduct) were found in urine of monkeys given efavirenz. Furthermore, cynomolgus monkeys were capable of producing 7-OH efavirenz (metabolite M5) and its glucuronide and sulfate conjugates (M6 and M7, respectively). Interestingly, these metabolites (M5, M6, and M7) were not found in rats.
Scheme FS3 shows the proposed pathways for the efavirenz metabolites produced by humans. The humans, similar to rats and monkeys, produced M1 as the major metabolite. Humans were also capable of forming the 7-hydroxylated derivatives (M5, M6, and M7), similar to cynomolgus monkeys. Analysis of human urine from several subjects showed total absence of any GSH-related adducts. Interestingly, M14 was found as a metabolite in human urine. The presence of this metabolite indicated that humans were also capable of hydroxylating at C-14 position however, this did not lead to the formation of any GSH products.
Scheme FS4 shows that guinea pigs were capable of producing all the metabolites and was found as an ideal source for obtaining milligram quantities of metabolites for characterization purposes. Hamsters showed a metabolite profile similar to guinea pigs, but hamsters were incapable of producing the GSH-related products. Metabolite M11, which is considered to be the precursor of M9, was found in significant quantities in both hamster and guinea pig urine (as well as in cynomolgus monkey urine).
Although the glucuronide conjugate, M1, was found as the principal metabolite in urine of all species, significant differences in the nature of excreted metabolites were observed. Of particular importance was the presence of metabolite M10 in urine of rats and its distinct absence in urine of cynomolgus monkeys and humans. The pathways leading to its formation were studied in detail. By observing the structural features of M10, it was postulated that M11 was probably a precursor to the GSH conjugate, M9. Metabolite M11 was found in urine of all species except in humans. It was demonstrated in vitro that this compound was indeed a precursor to the GSH conjugate, M9, when incubated in the presence of rat liver and kidney preparations. The GSH adduct was not produced when M11 was incubated with human or cynomolgus monkey liver and kidney subcellular fractions. Consistent with the in vitro data, the GSH-related adducts, M9 and M10, were not found in urine of cynomolgus monkeys and humans. It is postulated that the absence of GSH-related adducts in humans, hamsters, and cynomolgus monkey urine could be attributed to the lack of a specific GST or lack of substrate selectivity of the enzyme responsible for catalyzing the conjugation of M11 with GSH. The differences in the nature of GSTs present in kidneys/livers of different species have been well documented (Commandeur et al., 1995). GST isozymes may be expressed to a very different extent, both qualitatively and quantitatively, in different tissues (Mannervick et al., 1985; Ketterer, 1986; Mannervick and Danielson, 1988; Sundberg et al., 1993). In addition, there are sex and species differences (Igarashi and Satoh, 1989) in the expression of these GST isozymes. These differences in GSTs could account partially for the presence or absence of the cysteinylglycine conjugate (M10) in urine samples from different species. Metabolite M11 was found to be a very weak electrophile that reacted very slowly with GSH in the absence of enzyme. Hence, it appears that catalysis by an enzyme is needed to produce the GSH adduct from metabolite M11. Furthermore, M11 was not found in humans, which may also explain the absence of M9 or M10 in human urine. Analyses of plasma samples from rats given very high doses of efavirenz (>700 mg/kg) did not show any M9 to be present. Nonetheless, with every passage of blood, rat kidneys are capable of filtering and concentrating small amounts of M9 present in plasma. The levels of metabolite M9 in plasma may also have been below the limit of detection by the assay method. It is also possible that rat kidneys are capable of converting M11 to the GSH adduct as demonstrated by the in vitro studies. Metabolite M11 formed in liver could be transported to kidney where it is further metabolized to M9 by a specific GST. To confirm the presence of metabolite M11 in plasma, rats were dosed with its postulated precursor, metabolite M4 (8-OH efavirenz), (Mutlib et al., 1998a, 1999), and plasma samples were analyzed by LC/MS. It was found that, in addition to metabolites M1 and M3, detectable amounts of metabolite M11 were also present. Based on these observations, it is believed that rat kidneys are also capable of transforming metabolite M11 to the GSH adduct (M9).
The cysteinylglycine conjugate, M10, found in the urine of rats and guinea pigs is produced via the hydrolysis of M9 by GGT. This is the first step in the catabolism of GSH-conjugates (Curthoys and Hughey, 1979). The differences in the levels of this enzyme between different tissues in rats has been described (McIntyre and Curthoys, 1980; Commandeur et al., 1995). Rats have far greater GGT activity in kidneys than any other species. Species variation in the levels of GGT in different tissues of a number of species is very well documented (Hinchman and Ballatori, 1990). Surprisingly, metabolite M10 was found to be one of the major metabolites present in the urine of rats given high doses of efavirenz. This was unusual because cysteinylglycine conjugates are usually catabolized further in kidneys by dipeptidases to cysteine conjugates (Hirota et al., 1987; Commandeur et al., 1995). The in vitro conversion of M10 to the cysteine adduct (M17) was demonstrated by incubating M10 with rat kidney microsomes containing dipeptidase activity. TheN-acetylcysteine conjugate (mercapturate), which is normally found in urine of species producing GSH adducts (Barnsley et al., 1969), was not found in urine of any species. Analyses of human urine showed the absence of M10 and M17, suggesting that M9 was not formed in vivo.
Although all the species were quite capable of hydroxylating efavirenz at the three positions (Table 3), it appears that differences in phase II enzymes (e.g., sulfotransferases and GSTs) led to more dramatic differences in metabolite profiles. For example, urine samples from humans showed significant amounts of metabolite M7 (7-OH efavirenz sulfate) whereas metabolite M3 was distinctly absent. In contrast, at a comparable dose administered to rats (10 mg/kg p.o.), metabolite M3 was excreted as a significant metabolite in urine. The regio-isomer of M3 (i.e.the 7-OH efavirenz sulfate) was not detected in rat urine at any of the doses studied. It appears that humans conjugate metabolite M4 (precursor to metabolite M3) preferentially with glucuronic acid rather than with sulfuric acid. This preferential glucuronidation at the C8 position by humans could explain partially the lack of metabolites M8 and M11, which are potential substrates for GSTs. It appears that humans do not produce and excrete the GSH-related adducts for least two reasons: 1) the substrate (i.e., M11) for GST is not produced, and 2) the enzyme responsible for catalytic conversion of M11 to M9 is not present in humans (as demonstrated by in vitro experiments).
In a number of previously reported cases, the GSHS-conjugation has been identified as a bioactivation pathway leading to GSH conjugates that are reactive (van Ommen et al., 1988, 1992; Liebler, 1988) or that must undergo additional metabolism to become reactive (Nash et al., 1984; Vamvakas et al., 1987; Anders et al., 1988; Patel et al., 1994). The bioactivation by GST leading to subsequent nephrotoxicities has been reported for a number of halogenated compounds (Nash et al., 1984;Dekant et al., 1987; Koob and Dekant, 1991; Commandeur et al., 1995 and the literature cited therein). The nephrotoxicities of these compounds were attributed to additional catabolism of the GSHS-conjugates by rat kidneys leading to the reactive intermediates. Rats have shown signs of nephrotoxicity when administered high doses of efavirenz (Gerson et al., 1999). The presence of M10 in urine of rats in a very species-specific manner led us to postulate that the formation and subsequent renal processing of the GSH adduct, M9, could be responsible for this species-specific nephrotoxicity (Gerson et al., 1999). The differences in the metabolite profiles among the species could account for the absence or presence of nephrotoxicities observed after high doses of efavirenz. Humans and cynomolgus monkeys have shown no signs of nephrotoxicity. Urine from these two species do not contain metabolites M9 or M10, which are believed to be responsible for this nephrotoxicity.
Footnotes
-
Send reprint requests to: Dr. A.E. Mutlib, Drug Metabolism and Pharmacokinetics Section, The DuPont Pharmaceuticals Company, P.O. Box 30, 1094 Elkton Road, Newark, DE 19714. E-mail:abdul.mutlib{at}dupontpharma.com
- Abbreviations used are::
- LC/MS
- liquid chromatography/mass spectrometry
- ESI-LC/MS
- electrospray ionization-liquid chromatography mass spectrometry
- GGT
- γ-glutamyltranspeptidase
- GST
- glutathione-S-transferase
- PB
- phenobarbital
- SPE
- solid phase extraction
- HMQC
- heteronuclear multiple quantum correlation
- HMBC
- heteronuclear multiple bond correlation
- Received March 24, 1999.
- Accepted July 26, 1999.
- The American Society for Pharmacology and Experimental Therapeutics