Pharmacokinetics, mass balance, metabolism, and excretion of the liver-targeted acetyl-CoA carboxylase inhibitor PF-05221304 (clesacostat) in humans.

The disposition of the hepatoselective ACC inhibitor PF-05221304 (Clesacostat) was studied after a single 50-mg oral dose of [14C]-PF-05221304 to healthy human subjects.Mass balance was achieved with 89.9% of the administered dose recovered in urine and faeces, over the 11-day study period. The total administered radioactivity excreted in faeces and urine was 81.7 and 8.2%, respectively. Unchanged PF-05221304 accounted for 35.6% of the radioactive dose in faeces, suggesting ∼64% of the administered dose was absorbed.PF-05221304 was principally metabolised via oxidative and reductive pathways involving: (a) N-dealkylation, (b) isopropyl group monohydroxylation to yield enantiomeric metabolites (M2a and M2b), (c) hydroxylation on the 3-azaspiro[5.5]undecan-8-one moiety to metabolites M5 and 519c, and (d) carbonyl group reduction to enantiomeric alcohol metabolites M3, and M4. Secondary metabolites (521a, 521b, and 533), derived from a combination of oxidation and reduction of the primary metabolites accounted for ∼14.8% of the dose. In plasma, unchanged PF-05221304 represented 96.1% circulating radioactivity. Metabolites M1, M2b, and M2a represented 1.94, 1.76, and 0.18% of circulating radioactivity, respectively.Overall, these data suggest that PF-05221304 is well absorbed in humans and eliminated largely via phase I metabolism.

Role of Molybdenum-Containing Enzymes in the Biotransformation of the Novel Ghrelin Receptor Inverse Agonist PF-5190457: A Reverse Translational Bed-to-Bench Approach.

(R)-2-(2-methylimidazo[2,1-b]thiazol-6-yl)-1-(2-(5-(6-methylpyrimidin-4-yl)-2,3-dihydro-1H-inden-1-yl)-2,7-diazaspiro[3.5]nonan-7-yl)ethan-1-one (PF-5190457) was identified as a potent and selective inverse agonist of the ghrelin receptor [growth hormone secretagogue receptor 1a (GHS-R1a)]. The present translational bed-to-bench work characterizes the biotransformation of this compound in vivo and then further explores in vitro metabolism in fractions of human liver and primary hepatocytes. Following oral administration of PF-5190457 in a phase 1b clinical study, hydroxyl metabolites of the compound were observed, including one that had not been observed in previously performed human liver microsomal incubations. PF-6870961 was biosynthesized using liver cytosol, and the site of hydroxylation was shown to be on the pyrimidine using nuclear magnetic resonance spectroscopy. The aldehyde oxidase (AO) inhibitor raloxifene and the xanthine oxidase inhibitor febuxostat inhibited the formation of PF-6870961 in human liver cytosol, suggesting both enzymes were involved in the metabolism of the drug. However, greater inhibition was observed with raloxifene, indicating AO is a dominant enzyme in the biotransformation. The intrinsic clearance of the drug in human liver cytosol was estimated to be 0.002 ml/min per milligram protein. This study provides important novel information at three levels: 1) it provides additional new information on the recently developed novel compound PF-5190457, the first GHS-R1a blocker that has moved to development in humans; 2) it provides an example of a reverse translational approach where a discovery in humans was brought back, validated, and further investigated at the bench level; and 3) it demonstrates the importance of considering the molybdenum-containing oxidases during the development of new drug entities. SIGNIFICANCE STATEMENT: PF-5190457 is a novel ghrelin receptor inverse agonist that is currently undergoing clinical development for treatment of alcohol use disorder. PF-6870961, a major hydroxyl metabolite of the compound, was observed in human plasma, but was absent in human liver microsomal incubations. PF-6870961 was biosynthesized using liver cytosol, and the site of hydroxylation on the pyrimidine ring was characterized. Inhibitors of aldehyde oxidase and xanthine oxidase inhibited the formation of PF-6870961 in human liver cytosol, suggesting both enzymes were involved in the metabolism of the drug. This information is important for patient selection in subsequent clinical studies.

Acyl Glucuronide Metabolites of 6-Chloro-5-[4-(1-hydroxycyclobutyl)phenyl]-1 H-indole-3-carboxylic Acid (PF-06409577) and Related Indole-3-carboxylic Acid Derivatives are Direct Activators of Adenosine Monophosphate-Activated Protein Kinase (AMPK).

Studies on indole-3-carboxylic acid derivatives as direct activators of human adenosine monophosphate-activated protein kinase (AMPK) α1ß1γ1 isoform have culminated in the identification of PF-06409577 (1), PF-06885249 (2), and PF-06679142 (3) as potential clinical candidates. Compounds 1-3 are primarily cleared in animals and humans via glucuronidation. Herein, we describe the biosynthetic preparation, purification, and structural characterization of the glucuronide conjugates of 1-3. Spectral characterization of the purified glucuronides M1, M2, and M3 indicated that they were acyl glucuronide derivatives. In vitro pharmacological evaluation revealed that all three acyl glucuronides retained selective activation of ß1-containing AMPK isoforms. Inhibition of de novo lipogenesis with representative parent carboxylic acids and their respective acyl glucuronide conjugates in human hepatocytes demonstrated their propensity to activate cellular AMPK. Cocrystallization of the AMPK α1ß1γ1 isoform with 1-3 and M1-M3 provided molecular insights into the structural basis for AMPK activation by the glucuronide conjugates.

Biosynthesis of drug metabolites and quantitation using NMR spectroscopy for use in pharmacologic and drug metabolism studies.

The contribution of drug metabolites to the pharmacologic and toxicologic activity of a drug can be important; however, for a variety of reasons metabolites can frequently be difficult to synthesize. To meet the need of having samples of drug metabolites for further study, we have developed biosynthetic methods coupled with quantitative NMR spectroscopy (qNMR) to generate solutions of metabolites of known structure and concentration. These quantitative samples can be used in a variety of ways when a synthetic sample is unavailable, including pharmacologic assays, standards for in vitro work to help establish clearance pathways, and/or as analytical standards for bioanalytical work to ascertain exposure, among others. We illustrate five examples of metabolite biosynthesis and qNMR. The types of metabolites include one glucuronide and four oxidative products. Concentrations of the isolated metabolite stock solutions ranged from 0.048 to 8.3 mM, with volumes from approximately 0.04 to 0.150 ml in hexadeutarated dimethylsulfoxide. These specific quantified isolates were used as standards in the drug discovery setting as substrates in pharmacology assays, for bioanalytical assays to establish exposure, and in variety of routine absorption, distribution, metabolism, and excretion assays, such as protein binding and determining blood-to-plasma ratios. The methods used to generate these materials are described in detail with the objective that these methods can be generally used for metabolite biosynthesis and isolation.

The pharmacokinetics, metabolism, and clearance mechanisms of tofacitinib, a janus kinase inhibitor, in humans.

Tofacitinib is a novel, oral Janus kinase inhibitor. The objectives of this study were to summarize the pharmacokinetics and metabolism of tofacitinib in humans, including clearance mechanisms. Following administration of a single 50-mg (14)C-labeled tofacitinib dose to healthy male subjects, the mean (standard deviation) total percentage of administered radioactive dose recovered was 93.9% (±3.6), with 80.1% (±3.6) in the urine (28.8% parent), and 13.8% (±1.9) in feces (0.9% parent). Tofacitinib was rapidly absorbed, with plasma concentrations and total radioactivity peaking at around 1 hour after oral administration. The mean terminal phase half-life was approximately 3.2 hours for both parent drug and total radioactivity. Most (69.4%) circulating radioactivity in plasma was parent drug, with all metabolites representing less than 10% each of total circulating radioactivity. Hepatic clearance made up around 70% of total clearance, while renal clearance made up the remaining 30%. The predominant metabolic pathways of tofacitinib included oxidation of the pyrrolopyrimidine and piperidine rings, oxidation of the piperidine ring side-chain, N-demethylation and glucuronidation. Cytochrome P450 (P450) profiling indicated that tofacitinib was mainly metabolized by CYP3A4, with a smaller contribution from CYP2C19. This pharmacokinetic characterization of tofacitinib has been consistent with its clinical experience in drug-drug interaction studies.

Metabolism of a serotonin-4 receptor partial agonist 4-{4-[4-tetrahydrofuran-3-yloxy)-benzo[d]isoxazol-3-yloxymethyl]-piperidin-1-ylmethyl}-tetrahydropyran-4-ol (TBPT): identification of an unusual pharmacologically active cyclized oxazolidine metabolite in human.

4-{4-[4-Tetrahydrofuran-3-yloxy)-benzo[d]isoxazol-3-yloxymethyl]-piperidin-1-ylmethyl}-tetrahydropyran-4-ol (PF-4995274, TBPT) is a new agent that is a partial agonist of the human serotonin-4 (5-HT4) receptor and is under investigation for neurological disorders. Metabolism of TBPT was examined in vitro in human liver microsomes and human hepatocytes. Metabolites were also identified in the plasma of healthy human subjects in a phase 1 clinical study. Human-derived metabolite profiles were compared with corresponding profiles obtained in laboratory animal species. There were two major routes of metabolism in vitro: N-dealkylation of the methyltetrahydropyran moiety (M1) and hydroxylation at the seven position of the benzisoxazole moiety (M4). These were also observed in human plasma; however, in that matrix, the major metabolite was an unusual cyclized oxazolidine entity (M2). M2 was proposed to be formed via generation of an intermediate 4° iminium ion on the piperidine ring followed by spontaneous cyclization by attack of the ß-hydroxyl substituent of the tetrahydropyran ring to form a cyclized oxazolidine product. An authentic standard of the metabolite was generated using a methylene-blue-sensitized photochemical oxidation reaction as well as microbial transformation. Further investigation of this metabolite showed that it also possessed 5-HT4 agonism activity similar to the parent. The metabolite was 150-fold more highly protein bound in human plasma than TBPT, which is consistent with its presence as a major circulating metabolite while being only a minor metabolite in in vitro systems. Overall, this illustrates the importance of understanding the complex dispositional properties of a pharmacologically active metabolite.

Reactive metabolite trapping studies on imidazo- and 2-methylimidazo[2,1-b]thiazole-based inverse agonists of the ghrelin receptor.

The current study examined the bioactivation potential of ghrelin receptor inverse agonists, 1-{2-[2-chloro-4-(2H-1,2,3-triazol-2-yl)benzyl]-2,7-diazaspiro[3.5]nonan-7-yl}-2-(imidazo[2,1-b]thiazol-6-yl)ethanone (1) and 1-{2-[2-chloro-4-(2H-1,2,3-triazol-2-yl)benzyl]-2,7-diazaspiro[3.5]nonan-7-yl}-2-(2-methylimidazo[2,1-b]thiazol-6-yl)ethanone (2), containing a fused imidazo[2,1-b]thiazole motif in the core structure. Both compounds underwent oxidative metabolism in NADPH- and glutathione-supplemented human liver microsomes to yield glutathione conjugates, which was consistent with their bioactivation to reactive species. Mass spectral fragmentation and NMR analysis indicated that the site of attachment of the glutathionyl moiety in the thiol conjugates was on the thiazole ring within the bicycle. Two glutathione conjugates were discerned with the imidazo[2,1-b]thiazole derivative 1. One adduct was derived from the Michael addition of glutathione to a putative S-oxide metabolite of 1, whereas, the second adduct was formed via the reaction of a second glutathione molecule with the initial glutathione-S-oxide adduct. In the case of the 2-methylimidazo[2,1-b]thiazole analog 2, glutathione conjugation occurred via an oxidative desulfation mechanism, possibly involving thiazole ring epoxidation as the rate-limiting step. Additional insights into the mechanism were obtained via ¹8O exchange and trapping studies with potassium cyanide. The mechanistic insights into the bioactivation pathways of 1 and 2 allowed the deployment of a rational chemical intervention strategy that involved replacement of the thiazole ring with a 1,2,4-thiadiazole group to yield 2-[2-chloro-4-(2H-1,2,3-triazol-2-yl)benzyl]-2,7-diazaspiro[3.5]nonan-7-yl)-2-(2-methylimidazo[2,1-b][1,3,4]thiadiazol-6-yl)ethanone (3). These structural changes not only abrogated the bioactivation liability but also retained the attractive pharmacological attributes of the prototype agents.

Metabolism, excretion, and pharmacokinetics of ((3,3-difluoropyrrolidin-1-yl)((2S,4S)-4-(4-(pyrimidin-2-yl)piperazin-1-yl)pyrrolidin-2-yl)methanone, a dipeptidyl peptidase inhibitor, in rat, dog and human.

The disposition of 3,3-difluoropyrrolidin-1-yl{(2S,4S)-4-[4-(pyrimidin-2-yl)piperazin-1-yl]pyrrolidin-2-yl}methanone (PF-00734200), a dipeptidyl peptidase IV inhibitor that progressed to phase 3 for the treatment of type 2 diabetes, was examined in rats, dogs, and humans after oral administration of a single dose of [(14)C]PF-00734200. Mean recoveries of administered radioactivity were 97.1, 92.2, and 87.2% in rats, dogs, and humans, respectively. The majority of radioactive dose was detected in the urine of dogs and humans and in the feces of rats. Absorption of PF-00734200 was rapid in all species, with maximal plasma concentrations of radioactivity achieved within 1 h after the dose. Circulating radioactivity was primarily composed of the parent drug (79.9, 80.2, and 94.4% in rat, dog, and human, respectively). The major route of metabolism was due to hydroxylation at the 5' position of the pyrimidine ring (M5) in all species. In vitro experiments with recombinant cytochrome P450 isoforms suggested that the formation of M5 was catalyzed both by CYP2D6 and CYP3A4. Molecular docking simulations showed that the 5' position of the pyrimidine moiety of PF-00734200 can access the heme iron-oxo of both CYP3A4 and CYP2D6 in an energetically favored orientation. Other metabolic pathways included amide hydrolysis (M2), N-dealkylation at the piperazine nitrogen (M3) and an unusual metabolite resulting from scission of the pyrimidine ring (M1). Phase II metabolic pathways included the following: carbamoyl glucuronidation (M9), glucosidation (M15) on the pyrrolidine nitrogen, and conjugation with creatinine to form an unusual metabolite/metabonate (M16). The data from these studies suggest that PF-00734200 is eliminated by both metabolism and renal clearance.

Optimized assays for human UDP-glucuronosyltransferase (UGT) activities: altered alamethicin concentration and utility to screen for UGT inhibitors.

The measurement of the effect of new chemical entities on human UDP-glucuronosyltransferase (UGT) marker activities using in vitro experimentation represents an important experimental approach in drug development to guide clinical drug-interaction study designs or support claims that no in vivo interaction will occur. Selective high-performance liquid chromatography-tandem mass spectrometry functional assays of authentic glucuronides for five major hepatic UGT probe substrates were developed: ß-estradiol-3-glucuronide (UGT1A1), trifluoperazine-N-glucuronide (UGT1A4), 5-hydroxytryptophol-O-glucuronide (UGT1A6), propofol-O-glucuronide (UGT1A9), and zidovudine-5'-glucuronide (UGT2B7). High analytical sensitivity permitted characterization of enzyme kinetic parameters at low human liver microsomal and recombinant UGT protein concentration (0.025 mg/ml), which led to a new recommended optimal universal alamethicin activation concentration of 10 µg/ml for microsomes. Alamethicin was not required for recombinant UGT incubations. Apparent enzyme kinetic parameters, particularly for UGT1A1 and UGT1A4, were affected by nonspecific binding. Unbound intrinsic clearance for UGT1A9 and UGT2B7 increased significantly after addition of 2% bovine serum albumin, with minimal changes for UGT1A1, UGT1A4, and UGT1A6. Eleven potential UGT and cytochrome P450 inhibitors were evaluated as UGT inhibitors, resulting in observation of nonselective UGT inhibition by chrysin, mefenamic acid, silibinin, tangeretin, ketoconazole, itraconazole, ritonavir, and verapamil. The pan-cytochrome P450 inhibitor, 1-aminobenzotriazole, minimally inhibited UGT activities and may be useful in reaction phenotyping of mixed UGT and cytochrome P450 substrates. These methods should prove useful in the routine assessments of the potential for new drug candidates to elicit pharmacokinetic drug interactions via inhibition of human UGT activities and the identification of UGT enzyme-selective chemical inhibitors.

Validation of isolated metabolites from drug metabolism studies as analytical standards by quantitative NMR.

In discovery and development, having a qualified metabolite standard is advantageous. Chemical synthesis of metabolite standards is often difficult and expensive. As an alternative, biological generation and isolation of metabolites in the nanomole range are readily feasible. However, without an accurately defined concentration, these isolates have limited utility as standards. There is a significant history of NMR as both a qualitative and a quantitative technique, and these concepts have been merged recently to provide both structural and quantitative information on biologically generated isolates from drug metabolism studies. Previous methodologies relied on either specialized equipment or the use of an internal standard to the isolate. We have developed a technique in which a mathematically generated signal can be inserted into a spectrum postacquisition and used as a quantitative reference: artificial signal insertion for calculation of concentration observed (aSICCO). This technique has several advantages over previous methodologies. Any region in the analyte spectra, free from interference, can be chosen for the reference signal. In addition, the magnitude of the inserted signal can be modified to appropriately match the intensity of the sample resonances. Because this is postacquisition quantification, no special equipment or pulse sequence is needed. Compared with quantitation via the addition of an internal standard (10 mM maleic acid), the signal insertion method produced similar results. For each method, precision and accuracy were within ± 5%, stability of signal response over 8 days was ± 5%, and the dynamic range was more than 3 orders of magnitude: 10 to 0.01 mM.

Metabolism of ramelteon in human liver microsomes and correlation with the effect of fluvoxamine on ramelteon pharmacokinetics.

Ramelteon is a melatonin receptor agonist used as a treatment for insomnia. It is subject to a remarkably large drug-drug interaction (DDI) caused by fluvoxamine coadministration, resulting in a more than 100-fold increase in exposure. The objective of this study was to determine whether the DDI could be estimated using in vitro metabolism data. Ramelteon was shown to undergo hydroxylation in human liver microsomes to eight metabolites via six pathways. The main routes of metabolism included hydroxylation on the ethyl side chain and the benzylic position of the cyclopentyl ring, as assessed through enzyme kinetic measurements. Hydroxylation at the other benzylic position was observed in human intestinal microsomes. Ramelteon metabolism was catalyzed by CYP1A2, CYP2C19, and CYP3A4 as shown through the use of recombinant human cytochrome P450 enzymes and specific inhibitors. In liver, CYP1A2, CYP2C19, and CYP3A4 were estimated to contribute 49, 42, and 8.6%, respectively, whereas in intestine only CYP3A4 contributes. The in vitro data were used to estimate the magnitudes of DDI caused by ketoconazole, fluconazole, and fluvoxamine. The DDIs caused by the former were reliably estimated (1.82-fold estimated versus 1.82-fold actual for ketoconazole; 2.99-fold estimated versus 2.36-fold actual for fluconazole), whereas for fluvoxamine it was underestimated (11.4-fold estimated versus 128-fold actual). This suggests that there may be a limit on the magnitude of DDI that can be estimated from in vitro data. Nevertheless, the example of the fluvoxamine-ramelteon DDI offers a unique example wherein one drug can simultaneously inhibit multiple enzymatic pathways of a second drug.

Species differences in the biotransformation of an alpha 4 beta 2 nicotinic acetylcholine receptor partial agonist: the effects of distinct glucuronide metabolites on overall compound disposition.

The metabolism and disposition of (1R,5S)-2,3,4,5-tetrahydro-7-(trifluoromethyl)-1,5-methano-1H-3-benzazepine (1), an alpha(4)beta(2) nicotinic acetylcholine receptor partial agonist, was investigated in Sprague-Dawley rats and cynomolgus monkeys receiving (1R,5S)-2,3,4,5-tetrahydro-7-(trifluoromethyl)-1,5-methano-1H-4[(14)C]-3- benzazepine hydrochloride ([(14)C]1) orally. Although both species chiefly (>or=62%) cleared 1 metabolically, species-specific dispositional profiles were observed for both 1 and total radioactivity. Radioactivity was excreted equally in the urine and feces of intact rats but largely (72%) in bile in bile duct-cannulated animals. In monkeys, radioactivity recoveries were 50-fold greater in urine than feces and minimal (<5%) in bile. Both species metabolized 1 similarly: four-electron oxidation to one of four amino acids or two lactams (minor) and glucuronide formation (major). In rats, the latter pathway predominantly formed an N-carbamoyl glucuronide (M6), exclusively present in bile (69% of dose), whereas in monkeys it afforded an N-O-glucuronide (M5), a minor biliary component (4%) but the major plasma (62%) and urinary (42%) entity. In rats, first-pass hepatic conversion of 1 to M6, which was confirmed in rat hepatocytes, and its biliary secretion resulted in the indirect enterohepatic cycling of 1 via M6 and manifested in double-humped plasma concentration-time curves and long t(1/2) for both 1 and total radioactivity. In monkeys, in which only M5 was formed, double-humped plasma concentration-time curves were absent, and moderate t(1/2) for both 1 and total radioactivity were observed. A seemingly subtle, yet critical, difference in the chemical structures of these two glucuronide metabolites considerably affected the overall disposition of 1 in rats versus monkeys.

Quantitative interference by cysteine and N-acetylcysteine metabolites during the LC-MS/MS bioanalysis of a small molecule.

During LC-MS/MS quantification of a small molecule in human urine samples from a clinical study, an unexpected peak was observed to nearly co-elute with the analyte of interest in many study samples. Improved chromatographic resolution revealed the presence of at least 3 non-analyte peaks, which were identified as cysteine metabolites and N-acetyl (mercapturic acid) derivatives thereof. These metabolites produced artifact responses in the parent compound MRM channel due to decomposition in the ionization source of the mass spectrometer. Quantitative comparison of the analyte concentrations in study samples using the original chromatographic method and the improved chromatographic separation method demonstrated that the original method substantially over-estimated the analyte concentration in many cases. The substitution of electrospray ionization (ESI) for atmospheric pressure chemical ionization (APCI) nearly eliminated the source instability of these metabolites, which would have mitigated their interference in the quantification of the analyte, even without chromatographic separation. These results 1) demonstrate the potential for thiol metabolite interferences during the quantification of small molecules in pharmacokinetic samples, and 2) underscore the need to carefully evaluate LC-MS/MS methods for molecules that can undergo metabolism to thiol adducts to ensure that they are not susceptible to such interferences during quantification.

Biotransformation of an alpha4beta2 nicotinic acetylcholine receptor partial agonist in sprague-dawley rats and the dispositional characterization of its N-carbamoyl glucuronide metabolite.

The metabolism and disposition of (1S,5R)-2,3,4,5-tetrahydro-7-(trifluoromethyl)-1,5-methano-1H-3-benzazepine (1), an alpha(4)beta(2) nicotinic acetylcholine receptor partial agonist, was determined in Sprague-Dawley rats after oral administration of [(14)C]1. In intact animals, mass balance was achieved within 48 h, with 5 times more radioactivity excreted in urine than in feces. Compound 1 underwent renal and metabolic clearance equally and exhibited a very long half-life attributable to a secondary peak occurring 8 h postdose in its serum concentration-time curve. In bile duct-cannulated (BDC) rats, mass balance was also achieved within 48 h with 73.7, 23.4, and 5.5% of the dose detected in bile, urine, and feces, respectively. Rats metabolized 1 by two primary routes: four-electron oxidation to either four amino acids or a lactam and formation of an N-carbamoyl glucuronide (M6), which was only detected in bile. The presence of M6 solely in bile and the double-humped serum concentration-time curve of 1 suggested the indirect enterohepatic cycling of 1 via M6 after oral administration. To explore this mechanistic hypothesis further, intravenous studies were conducted with 1 in both intact and BDC rats to determine the extent of 1 undergoing indirect enterohepatic cycling via M6. Compared with the pharmacokinetics in intact rats, total serum clearance was higher (1.7-fold) and volume of distribution was lower (1.6-fold) in BDC rats, resulting in a correspondingly shorter (2.5-fold) half-life, with 56% of administered 1 undergoing recirculation, an amount consistent with that (68% of dose) of M6 observed in bile from rats dosed orally with [(14)C]1.

In vitro metabolism and covalent binding of enol-carboxamide derivatives and anti-inflammatory agents sudoxicam and meloxicam: insights into the hepatotoxicity of sudoxicam.

Sudoxicam and meloxicam are nonsteroidal anti-inflammatory drugs (NSAIDs) from the enol-carboxamide class. While the only structural difference between the two NSAIDs is the presence of a methyl group on the C5-position of the 2-carboxamidothiazole motif in meloxicam, a marked difference in their toxicological profile in humans has been discerned. In clinical trials, sudoxicam was associated with several cases of severe hepatotoxicity that led to its discontinuation, while meloxicam has been in the market for over a decade and is devoid of hepatotoxicity. In an attempt to understand the biochemical basis for the differences in safety profile, an in vitro investigation of the metabolic pathways and covalent binding of the two NSAIDs was conducted in NADPH-supplemented human liver microsomes. Both compounds demonstrated NADPH-dependent covalent binding to human liver microsomes; however, the extent of binding of [(14)C]-meloxicam was approximately 2-fold greater than that of [(14)C]-sudoxicam. While inclusion of glutathione (GSH) in microsomal incubations resulted in a decrease in covalent binding for both NSAIDs, the reduction in binding was more pronounced for meloxicam. Metabolite identification studies on [(14)C]-sudoxicam in NADPH-supplemented human liver microsomes indicated that the primary route of metabolism involved a P450-mediated thiazole ring scission to the corresponding acylthiourea metabolite (S3), a well-established pro-toxin. The mechanism of formation of S3 presumably proceeds via (a) epoxidation of the C4-C5-thiazole ring double bond, (b) epoxide hydrolysis to the corresponding thiazole-4,5-dihydrodiol derivative, which was observed as a stable metabolite (S2), (c) ring opening of the thiazole-4,5-dihydrodiol to an 2-oxoethylidene thiourea intermediate, and (d) hydrolysis of the imine bond within this intermediate to yield S3. In the case of meloxicam, the corresponding acylthiourea metabolite M3 was also observed, but to a lesser extent; the main route of meloxicam metabolism involved hydroxylation of the 5'-methyl group, a finding that is consistent with the known metabolic fate of this NSAID. Inclusion of GSH led to a decrease in the formation of M3 with the concomitant formation of an unusual two-electron reduction product (metabolite M7). The formation of M7 is proposed to arise via reduction of the imine bond in 2-oxopropylidene thiourea, an intermediate in the thiazole ring scission pathway in meloxicam. In conclusion, the results of our analysis suggest that if the covalent binding of the two NSAIDs is important to the overall hepatotoxicity risk, the differences in metabolism (differential preponderance of formation of the acylthiourea relative to total metabolism), differential effects of GSH on covalent binding, and finally differences in daily doses of the two NSAIDs may serve as a plausible explanation for the marked differences in toxicity.