Bioactivation of Diclofenac in Vitro and in Vivo: Correlation to

Diclofenac is widely used in the treatment of, for example, arthritis and muscle pain. The use of diclofenac has been associated with hepatotoxicity, ...
0 downloads 0 Views 295KB Size
Chem. Res. Toxicol. 2008, 21, 1107–1119

1107

Bioactivation of Diclofenac in Vitro and in ViWo: Correlation to Electrochemical Studies Kim G. Madsen,*,† Christian Skonberg,† Ulrik Jurva,‡ Claus Cornett,† Steen H. Hansen,† Tommy N. Johansen,§ and Jørgen Olsen| Department of Pharmaceutics and Analytical Chemistry, Faculty of Pharmaceutical Sciences, UniVersity of Copenhagen, Denmark, DMPK & Physical Chemistry, Lead Generation, AstraZeneca R&D, Mölndal, Sweden, Department of Medicinal Chemistry, Faculty of Pharmaceutical Sciences, UniVersity of Copenhagen, Denmark, and ADME & Assay Technology, NoVo Nordisk A/S, DK-2760 MåløV, Denmark ReceiVed NoVember 30, 2007

Diclofenac is widely used in the treatment of, for example, arthritis and muscle pain. The use of diclofenac has been associated with hepatotoxicity, which has been linked to the formation of reactive metabolites. Diclofenac can be metabolized to 4′-OH- and 5-OH-diclofenac, both of which are able to form quinone imines capable of reacting with, for example, GSH and nucleophilic groups in proteins. Electrochemistry has been shown to be a suitable tool for mimicking some types of oxidative drug metabolism and for studying the formation of reactive metabolites. In these studies, the electrochemical oxidation of diclofenac to a +16 Da metabolite was shown to be identical to a synthetic standard of 5-OH-diclofenac. Furthermore, two different experimental designs were investigated with respect to the electrochemical oxidation of 4′-OH- and 5-OH-diclofenac. In the first approach, the oxidized sample was collected in an aqueous solution of GSH, whereas in the other approach, GSH was added to the sample before the oxidation was performed. From these electrochemical oxidations, a range of GSH conjugates of 4′-OH- and 5-OH-diclofenac were observed and characterized by MS/MS. This allowed the development of sensitive LC-MS methods in order to detect the GSH conjugates from in ViVo (rat bile) and in Vitro (human liver microsomes (HLM), rat liver microsomes (RLM), and rat hepatocytes) samples. A wide range of mono-, di-, and triglutathionyl conjugates were detected in the in Vitro and in ViVo samples. It was also observed that 5-OH-diclofenac formed GSH conjugates with RLM and HLM without addition of NADPH, whereas GSH conjugate formation of 4′-OH-diclofenac was NADPHdependent. This indicated that 5-OH-diclofenac was prone to auto-oxidation. The oxidation potentials of the two hydroxy metabolites were determined by cyclic voltammetry. A difference of 69 mV was observed between the two oxidation potentials, which in part may explain the extent of auto-oxidation for 5-OHdiclofenac. In conclusion, it was shown that electrochemical oxidation was capable of mimicking the metabolic hydroxylation of diclofenac to 5-OH-diclofenac. Furthermore, electrochemical oxidation was used to generate a range of GSH conjugates of 4′-OH- and 5-OH-diclofenac and a number of these conjugates were also detected in metabolism studies with microsomes (HLM/RLM) and freshly isolated rat hepatocytes, and in ViVo in rat bile. Introduction 1

Diclofenac (DCl) is a nonsteroidal anti-inflammatory drug that is widely used for the treatment of arthritis and muscle pain. DCl undergoes extensive hepatic metabolism, including aromatic hydroxylations, which may play an important role in the formation of reactive metabolites that have been associated with idiosyncratic hepatotoxicity of DCl (1). The mechanism behind the formation of the reactive metabolites of DCl has been the topic of many investigations, and it is generally accepted that the two hydroxy metabolites, 4′OH-diclofenac (4′-OH-DCl) and 5-OH-diclofenac (5-OH-DCl), are involved in the formation of the reactive metabolites. † Department of Pharmaceutics and Analytical Chemistry, University of Copenhagen. ‡ AstraZeneca R&D. § Department of Medicinal Chemistry, University of Copenhagen. | Novo Nordisk A/S. 1 Abbreviations: 4′-OH-DCl, 4′-OH-diclofenac; 5-OH-DCl, 5-OH-diclofenac; COSY, correlation spectroscopy; DCl, diclofenac; GSH, glutathione; HBSS, Hanks’ balanced salt solution; HLM, human liver microsomes; NOE, nuclear Overhauser effect; RLM, rat liver microsomes.

Furthermore, DCl can form O-acyl glucuronides, which may be reactive toward cellular components (1). DCl can form GSH conjugates in incubations with microsomes in the presence of GSH and in ViVo, and it has been shown that the GSH conjugates correspond to the addition of GSH to the two hydroxy metabolites (2–5). It has been proposed that quinone imines are formed as the reactive intermediates, which then react with GSH (4). This has been corroborated by Shen et al., who isolated and characterized the quinone imine of 5-OH-DCl from rat liver microsomes (6). In contrast to this, Yan et al. proposed that an arene oxide was formed as the reactive intermediate (7). The pathways for generating the reactive quinone imine intermediates are shown in Scheme 1. Electrochemistry can be used for mimicking some types of biologic reactions, such as oxidative drug metabolism (8–15), including formation of reactive metabolites (16). In this study, the electrochemical synthesis of 5-OH-DCl and mono-, di-, and triglutathionyl conjugates of 4′-OH-DCl and 5-OH-DCl is reported. These conjugates were characterized by MS and MS/MS and used for identification of GSH conjugates

10.1021/tx700419d CCC: $40.75  2008 American Chemical Society Published on Web 04/18/2008

1108

Chem. Res. Toxicol., Vol. 21, No. 5, 2008

Scheme 1. Hydroxylation of Diclofenac, Which Leads to 4′-OH-Diclofenac and 5-OH-Diclofenaca

a Both compounds are able to form quinone imines as reactive intermediates.

from incubations of DCl, 4′-OH-DCl and 5-OH-DCl with rat liver microsomes (RLM) and human liver microsomes (HLM) in the presence of GSH. In addition, a wide range of the GSH conjugates was also detected from incubation of DCl, 4′-OHDCl and 5-OH-DCl with freshly isolated rat hepatocytes and in the bile from a rat dosed with DCl.

Materials and Methods Chemicals and Reagents. DCl (sodium salt), 3-methoxyphenylacetic acid, 2,6-dichloroaniline, oxalyl chloride, boron tribromide (1 M in dichloromethane), glutathione (GSH), nicotinamide adenine dinucleotide phosphate (NADPH), and phosphoric acid were obtained from Sigma-Aldrich (Steinheim, Germany). Sodium thiosulfate, sodium sulfite, sodium dihydrogenphosphate, and disodium hydrogenphosphate were obtained from Merck (Darmstadt, Germany). N-iodosuccinimide, copper(II) sulfate, copper(I) iodide, zinc, and a 40% w/w solution of dimethylamine in water were obtained from Fluka (Sigma-Aldrich, Steinheim, Germany). 4′-OH-DCl was obtained from BD Biosciences (Woburn, MA). Pooled human liver microsomes (HLM; 27 individuals) and pooled rat liver microsomes (RLM; male, Sprague–Dawley) were obtained from BD Biosciences (Woburn, MA). 5-OH-DCl was prepared as described in the section Synthesis of 5-OH-DCl. Water used for mobile phases and solutions for the electrochemical oxidations was purified with a Milli-Q deionization unit (Millipore, France). Solvents used for mobile phases and electrochemical oxidations were of HPLC-grade. Instrumentation. LC-MS and LC-MS/MS were performed using a Thermo Finnigan Surveyor LC system coupled to a Thermo Finnigan TSQ Quantum Ultra AM triple quadrupole mass spectrometer with an ESI interface (Thermo Finnigan, San José, CA). An Agilent LC/MSD ion-trap mass spectrometer coupled to an Agilent 1100-series LC-system was used to study the product

Madsen et al. formed from electrochemical oxidation of DCl and to follow the reaction steps in the synthesis of 5-OH-DCl (Agilent, Germany). Preparative chromatography was performed using an Agilent 1100 series LC-system equipped with a fraction collector. Electrochemical oxidations were performed in an ESA 5011 analytical electrochemical cell equipped with a porous graphite working electrode, a Pd counter electrode and a Pd/H2 reference electrode. The cell was controlled by an ESA Coulochem potentiostat (model 5100A, ESA Bioscience, Inc., Chelmsford, MA). Cyclic voltammograms were recorded by the use of a threeelectrode system controlled by a computer-controlled potentiostat (Autolab, Netherlands). A glassy carbon electrode (2 mm diameter) (Radiometer, Denmark) was used as working electrode, a platinum wire as the counter electrode, and a homemade silver/silver chloride electrode as the reference electrode. 1 H NMR was obtained using a Bruker 400 MHz spectrometer equipped with a 1 mm probe (Bruker, Germany). Chromatography. All LC-MS analyses were performed with a Luna Phenyl-Hexyl, 5 µm, 150 mm × 2 mm column (Phenomenex, Torrance, CA) using gradient elution with a mobile phase A consisting of 5% acetonitrile and 0.2% formic acid in water and a mobile phase B consisting of acetonitrile containing 0.2% formic acid. The gradient was linear from 0% B to 30% B over 15 min, followed by an instant increase to 95% B, which was held for 2 min. Finally, the column was allowed to re-equilibrate at 100% A for 8 min. Preparative chromatography was performed on a Luna C18(2), 3 µm, 150 mm × 4.6 mm column (Phenomenex, Torrance, CA) using gradient elution with a mobile phase A consisting of 5% acetonitrile and 0.2% formic acid in water and a mobile phase B consisting of acetonitrile containing 0.2% formic acid. For purification of 5-OH-DCl from the electrochemical synthesis, the gradient was run from 20% B to 60% B over a period of 20 min, and the column was allowed to re-equilibrate at 20% B for 10 min. For purification of 5-OH-DCl from the organic synthesis, the gradient was run from 20% B to 35% B over a period of 10 min and then kept constant at 35% B for 5 min, followed by an increase from 35% B to 60% for the next 10 min. The column was allowed to re-equilibrate at 20% B for 5 min. Preparative column chromatography was performed using silica gel 60 (35–70 µm from Fluka(Sigma-Aldrich, Steinheim, Germany) as stationary phase. Electrochemical Oxidation of Diclofenac. The electrochemical synthesis of 5-OH-DCl was performed by oxidizing a solution of 100 µM DCl in a 0.1 M phosphate buffer (pH 7.4)/acetonitrile (3: 1) at +1000 mV. The oxidized sample was collected in an aqueous solution with an excess of ascorbic acid. The sample was lyophilized and the remaining solid was reconstituted in mobile phase A and purified by SPE using a C18-SPE-cartridge (Varian, 3 cc, 500 mg) to remove the large amounts of inorganic salts and ascorbic acid. The adsorbed compounds were eluted with methanol, which was subsequently removed under a stream of nitrogen gas. The remaining solid was reconstituted in mobile phase A and purified by preparative chromatography. A solution of 25 µM 4′-OH-DCl or 5-OH-DCl in a 0.1 M phosphate buffer (pH 7.4)/acetonitrile (3:1) was oxidized at +600 mV. Two different approaches for generating GSH conjugates were used. In the first approach, the oxidized sample was collected in an aqueous solution of GSH (final concentration was 5 mM). In the second approach, GSH was added to the sample before it was passed through the electrochemical cell (final concentration was 5 mM). All samples were analyzed by LC-MS. Synthesis of 5-OH-DCl. 5-OH-DCl was synthesized according to the method described by Kenny et al. (17), with a minor modification in the last step. Oxalyl chloride (4.6 mL, 54.3 mmol) was added dropwise to a solution of 3-methoxyphenylacetic acid (4.3 g, 25.9 mmol) in anhydrous dichloromethane (150 mL), which was stirred at 20 °C. Dimethylformamide (13 drops) was added, and the solution was stirred until effervescence ceased (approximately 2 h). The solution

BioactiVation of Diclofenac in Vitro and in ViVo was evaporated to dryness and re-evaporated from anhydrous dichloromethane twice. Without purification, 3-methoxyphenylacetyl chloride was redissolved in anhydrous dichloromethane (70 mL) and added dropwise to a stirred two-phase system of 40% w/w aqueous dimethylamine (170 mL) in water (250 mL) and dichloromethane (250 mL) at 0 °C. After 1 h, the organic layer was separated and washed with 1 M HCl (350 mL), 10% aqueous Na2CO3 (350 mL), and water. The organic layer was dried with MgSO4, filtered, and evaporated to dryness. The remaining yellow oil was pure according to LCMS and was used for the next reaction without further purification. 3-Methoxyphenyl-N,N-dimethylacetamide (1.15 g, 5.9 mmol) and N-iodosuccinimide (3 g, 13.3 mmol) were dissolved in acetonitrile (7.5 mL), refluxed for 24 h, and then evaporated. The residue was dissolved in diethyl ether (10 mL) and washed with 10% aqueous Na2S2O3 (3 × 10 mL) affording 2-iodo-5-methoxyphenyl-N,Ndimethylacetamide, which was used without further purification. 2-Iodo-5-methoxyphenyl-N,N-dimethylacetamide (1.5 g, 4.7 mmol), 2,6-dichloroaniline (1.45 g, 9.0 mmol), anhydrous K2CO3(0.5 g, 3.6 mmol), copper(I) iodide (50 mg, 0.25 mmol), and freshly activated copper (164 mg, 2.6 mmol) was stirred and refluxed in anhydrous toluene (12 mL) for 24 h. After the reaction had completed as observed by LC-MS, the mixture was allowed to cool and was filtered, and the solvent was evaporated under reduced pressure. The crude product was diluted with ethyl acetate, and the organic layer was washed with water (3 × 10 mL). 2-(2′,6′Dichlorophenyl)amino-5-methoxyphenyl-N,N-dimethylacetamide was purified by column chromatography on silica gel using ethyl acetate/ hexane 2:3. Activated copper was obtained from a reduction of copper(II) sulfate with zinc. When the solution was colorless, the copper was filtered off. The copper was washed with 5% HCl until hydrogen evolution ceased. The remaining copper was filtered off and washed with distilled water. Copper was stored in water as a paste. Prior to use, an appropriate amount of copper was filtered off and washed with water, methanol, and diethyl ether. Copper(I) iodide was purified prior to use. Copper(I) iodide was stirred with a saturated aqueous solution of potassium iodide and activated charcoal for 3 h. The solution was filtered through Celite. Water was added to the filtrate in order to precipitate copper(I) iodide. The precipitate was washed with water (3 × 5 mL), ethanol (3 × 5 mL), and diethyl ether (3 × 5 mL). The purified copper(I) iodide was dried under vacuum overnight and stored protected against light. 2-(2′,6′-Dichlorophenyl)amino-5-methoxyphenyl-N,N-dimethylacetamide (250 mg, 0.7 mmol) was dissolved in 1,2dichloroethane (10 mL). A 1 M solution of boron tribromide in dichloromethane (2.5 mL, 2.5 mmol) was added dropwise at 20 °C. After stirring for 1 h, the reaction mixture was added to a halfsaturated aqueous NaHCO3 solution (50 mL) and extracted with ethyl acetate (75 mL). The organic phase was isolated, washed with water, and evaporated to give the corresponding phenol. 2-(2′,6′-Dichlorophenyl)amino-5-hydroxyphenyl-N,N-dimethylacetamide (0.15 g, 0.44 mmol) and sodium sulfite (1.5 g, 11.9 mmol) was suspended in ethanol (1.5 mL) in a Schlenk tube under nitrogen. A solution of 1 M NaOH (3.6 mL, 3.6 mmol) was added. The tube was purged with nitrogen and heated to 80 °C for 48 h. The solution was allowed to cool and washed with ether. The solution was acidified to pH 2 with 3 M HCl and extracted with ethyl acetate and evaporated. 5-OH-DCl was purified using preparative chromatography. Acetonitrile was removed from the collected fractions by evaporation at reduced pressure, and the remaining aqueous fraction was extracted with ethyl acetate. The organic phase was dried with MgSO4, filtered, and evaporated to dryness. MS of 5-OH-DCl: m/z 312.0 ([5-OH-DCl + H]+). MS/MS of 5-OH-DCl: m/z 293.9 ([5-OH-DCl - H2O + H]+); m/z 265.9 ([5OH-DCl - H2CO2 + H]+); m/z 231.0 ([5-OH-DCl - H2CO2 Cl· + H]+); m/z 230.0 ([5-OH-DCl - H2CO2 - HCl + H]+). 1 H NMR of 5-OH-DCl (d6-DMSO): δ 3.63 (2 H, s, Ar-CH2COOH); δ 6.28 (1 H, d, J ) 8.4 Hz, 3-H); δ 6.52 (1 H, dd, J ) 2.8 Hz and J ) 8.4 Hz, 4-H); δ 6.68 (1 H, d, J ) 2.4 Hz, 6-H); δ

Chem. Res. Toxicol., Vol. 21, No. 5, 2008 1109

Figure 1. 1H NMR spectra (in d6-DMSO) of 5-OH-DCl obtained from (A) electrochemical oxidation of diclofenac at +1000 mV (the singlet from the methylene group is not seen due to suppression of the HDO signal) and (B) a synthetic standard.

7.03 (1 H, t, J ) 8.0 Hz, 4′-H); δ 7.43 (2 H, d, J ) 8.0 Hz, 3′-H and 5′-H) (Figure 1). NOE was observed between the benzylic protons and the aromatic proton in position six (6-H in 5-OH-DCl). From a longrange COSY experiment, a coupling between the benzylic protons and the aromatic proton in the sixth position (6-H in 5-OH-DCl) was observed (Supporting Information, Figure 1). Incubations in Human and Rat Liver Microsomes. DCl, 4′OH-DCl, or 5-OH-DCl was incubated with human and rat liver microsomes (0.5 mg protein/mL). The final concentration of the tested compound was either 10 or 100 µM, the GSH concentration was 5 mM, and the NADPH concentration was 1 mM. All incubations were performed in a 0.1 M phosphate buffer, pH 7.4, at 37 °C for 30 min. The incubations were terminated by addition of one volume of cold acetonitrile (containing 0.2% formic acid) to the incubation mixture. The samples were centrifuged (Eppendorf Centrifuge 5804R) for 10 min at 4 °C (4500g). The supernatant was diluted 1:1 with water and was analyzed by LC-MS. Incubations in Rat Hepatocytes. DCl, 4′-OH-DCl, or 5-OHDCl (10 or 100 µM) was incubated with freshly isolated rat hepatocytes (male, Sprague–Dawley; 1 × 106 cells/mL) in HBSS buffer. The reaction was terminated by adding one volume of cold acetonitrile (containing 0.2% formic acid). The mixture was centrifuged, and the supernatant was diluted with an equal amount of water prior to analysis by LC-MS. Collection of Bile from Rat Dosed with Diclofenac. Three male Sprague–Dawley rats (∼250 g; anaesthetized with pentobarbital, 50 mg/kg ip) were bile-duct cannulated, using PE-10 tubing, through an incision in the abdomen. After 24 h recovery, the rats were given 200 mg/kg DCl (in H2O) po. The bile was collected from 0 to 4 h into 15 mL plastic centrifuge tubes containing 1–2 mL of glacial acetic acid. The sample tubes were kept on dry ice throughout the sampling period and subsequently stored at –30 °C until analysis. During the experiment, the rats were allowed free access to food and water. Prior to analysis, the sample was thawed and an amount of bile was mixed with acetonitrile (final concentration of acetonitrile was 25%). The sample was centrifuged at 11 000 g for 15 min. The supernatant was analyzed by LC-MS analysis. Auto-oxidation of 4′-OH-DCl and 5-OH-DCl in Rat Bile. Twenty-five micromolar solutions of DCl, 4′-OH-DCl, or 5-OHDCl were incubated with blank rat bile for 30 min at 37 °C. The

1110

Chem. Res. Toxicol., Vol. 21, No. 5, 2008

Scheme 2. The Proposed Mechanism of the Electrochemical Formation of 5-OH-Diclofenac, Which Was Observed from Oxidation of Diclofenac at +1000 mV

incubations were terminated by adding one volume of cold acetonitrile (containing 0.2% formic acid) to the incubation mixture. The samples were centrifuged (Eppendorf Centrifuge 5804R) for 10 min at 4 °C (4500g). The supernatant was diluted 1:1 with water and was analyzed by LC-MS. Cyclic Voltammetry of 4′-OH-DCl and 5-OH-DCl. The electrochemical properties of 4′-OH-DCl and 5-OH-DCl were studied by cyclic voltammetry. Nitrogen gas was bubbled through the solutions prior to the recordings. Cyclic voltammograms were recorded from a solution of 200 µM 4′-OH-DCl or 5-OH-DCl in a 0.1 M phosphate buffer (pH 7.4) with a scan speed of 100 mV/s.

Results Synthesis of 5-OH-DCl. 5-OH-DCl was synthesized according to the method described by Kenny et al. (17) with a minor modification. In the hydrolysis step of the amide, excessive degradation, due to oxidation, was observed, and therefore the last step was modified. It was found that the oxidation could be avoided by adding sodium sulfite as antioxidant to the reaction mixture. Electrochemical Oxidation of Diclofenac. When DCl was oxidized at +1000 mV and the oxidized solution was collected in an aqueous solution of ascorbic acid, a product at m/z 312.0 corresponding to the hydroxylation of DCl was identified by LC-MS. From MS/MS experiments, the following fragment ions were assigned: m/z 293.9 ([5-OH-DCl - H2O + H]+); m/z 266.0 ([5-OH-DCl - H2CO2 + H]+); m/z 231.0 ([5-OH-DCl H2CO2 - Cl· + H]+); m/z 230.0 ([5-OH-DCl - H2CO2 - HCl + H]+). The product was purified by solid-phase extraction and preparative chromatography, and the purified product was analyzed by 1H NMR. The chemical shifts for the observed signals were δ 6.21 (1 H, d, J ) 8.4 Hz, 3-H), δ 6.41 (1 H, dd, J ) 2.4 Hz and J ) 8.4 Hz, 4-H), δ 6.58 (1 H, d, J ) 2.8 Hz, 6-H), δ 6.97 (1 H, t, J ) 8.4 Hz, 4′-H), and δ 7.39 (2 H, d, J ) 8.8 Hz, 3′-H and 5′-H) (Figure 1). The signal for the methylene protons was not observed due to suppression of the HDO signal. The MS/MS data and NMR data (Figure 1) were identical to those observed for a synthesized standard of 5-OHDCl. The proposed mechanism for the electrochemical hydroxylation of DCl is shown in Scheme 2. The initial step involves an electron abstraction, which leads to nitrogen-centered radical cation. The radical cation undergoes rearrangement and deprotonation leading to a carbon-centered radical para to the amino group. The radical undergoes a second electron abstraction leading to a carbocation, which reacts with water, and by aromatization of the intermediate, 5-OH-DCl is formed. In theory, it should also be possible to form 3-OH-diclofenac and 4′-OH-DCl by the same mechanism, but these two

Madsen et al.

compounds were not observed. 3-OH-Diclofenac may not be formed due to steric hindrance at position C-3. The carbon radical, which would be involved in the formation of 4′-OHDCl, is destabilized by the electron-withdrawing effect of the chloro atoms, which might explain the lack of formation of 4′OH-DCl. When DCl was oxidized at +1000 mV and the oxidized sample was collected in an aqueous solution of GSH, two conjugates were identified at m/z 617.2 at 14.7 min (5-OH-D1) and 15.9 min (5-OH-D2), which corresponded to the GSH conjugates of 5-OH-DCl (see later section). Fragmentation of 5-OH-D1 at m/z 617.2 by MS/MS gave the following fragment ions: m/z 542.2 ([5-OH-D1 - glycine + H]+); m/z 487.8 ([5OH-D1 - glutamic acid + H]+); m/z 384.9 ([5-OH-D1 glycine - glutamic acid - CO + H]+); m/z 341.9 ([5-OH-D1 - glycine - glutamic acid - C3H3NO - 2H + H]+); m/z 324.0 ([5-OH-D1 - glycine - glutamic acid - C3H3NO - 2H H2O + H]+) (Figure 2 (MS/MS of 5-OH-D1) and Supporting Information, Figure 2A (MS/MS of 5-OH-D2)). Both conjugates had retention times and fragmentation patterns identical to the electrochemically generated GSH conjugates of 5-OH-DCl (Supporting Information, Figure 2C,D). When analyzing for 5-OH-DCl-(SG)2 by SRM, it was possible to detect this conjugate at 12.4 min when the oxidation of DCl at +1000 mV was performed in the presence of GSH (data not shown). Characterization of GSH Conjugates Formed from Electrochemical Oxidation of 4′-OH-DCl. When 4′-OH-DCl was oxidized at +600 mV and the formed reactive intermediates were trapped with GSH, eight different GSH conjugates were detected, as shown in Scheme 3. The electrochemical oxidation of 4′-OH-DCl was explored by two approaches, (i) collection of the oxidized sample in an aqueous solution of GSH and (ii) addition of GSH to the sample before the electrochemical oxidation. Two of the conjugates could only be identified if the oxidized sample was collected in an aqueous solution of GSH (4′-OH-D1 and 4′-OH-D2), whereas the remaining GSH conjugates could only be identified if GSH was added to the sample before electrochemical oxidation was performed. Some of the conjugates were only identified as the singly charged ion and others were identified as singly and doubly charged ions. The structure elucidation was based on MS/MS experiments of the singly charged conjugate. The GSH conjugate eluting at 14.4 min was observed at m/z 583.1, which corresponds to 2′-glutathionyl-4′-hydroxy-deschloro-diclofenac (4′-OH-D1). Fragmentation of m/z 583.1 by MS/MS gave the following fragment ions: m/z 507.8 ([4′-OHD1 - glycine + H]+); m/z 454.2 ([4′-OH-D1 - glutamic acid + H]+); m/z 436.1 ([4′-OH-D1 - glutamic acid - H2O + H]+); m/z 419.1 ([4′-OH-D1 - glutamic acid - H2O - NH3 + H]+); m/z 334.1 ([4′-OH-D1 - glutamic acid - glycine - CO - NH3 + H]+); m/z 332.7 ([4′-OH-D1 - glutamic acid - glycine CO - H2O + H]+); m/z 315.0 ([4′-OH-D1 - glutamic acid glycine - CO - 2 H2O + H]+); m/z 130.0 ([glutamic acid H2O + H]+) (Supporting Information, Figure 3A). A second conjugate with a retention time of 15.0 min was detected at m/z 617.2, which corresponds to 3′-glutathionyl-4′hydroxy-diclofenac (4′-OH-D2), assuming that the GSH conjugate formation took place via quinone imine. Fragmentation of m/z 617.2 by MS/MS gave the following fragment ions: m/z 542.1 ([4′-OH-D2 - glycine + H]+); m/z 488.1 ([4′-OH-D2 glutamic acid + H]+); m/z 470.8 ([4′-OH-D2 - glutamic acid - NH3 + H]+); m/z 462.3 ([4′-OH-D2 - glycine - CO - Cl· - NH3 + H]+); m/z 452.0 ([4′-OH-D2 - glutamic acid - 2

BioactiVation of Diclofenac in Vitro and in ViVo

Chem. Res. Toxicol., Vol. 21, No. 5, 2008 1111 Scheme 3. The GSH Conjugates of 4′-OH-Diclofenac, Which Can Be Formed from Reaction of the Quinone Imine Intermediates with GSH (See Table 1)

Figure 2. MS/MS spectra obtained from fragmentation of (A) m/z 617.2 (5-OH-D1) from electrochemical oxidation of diclofenac at a potential of +1000 mV followed by collection of the sample in an aqueous solution of GSH, (B) m/z 888.0 (4′-OH-D4) from electrochemical oxidation of 4′OH-DCl in the presence of GSH at +600 mV, (C) m/z 922.4 (5-OH-D3) from electrochemical oxidation of 5-OH-DCl in the presence of GSH at +600 mV, and (D) m/z 1159.4 (4′-OH-D6) from electrochemical oxidation of 4′-OH-DCl in the presence of GSH at +600 mV.

H2O + H]+); m/z 415.9 ([4′-OH-D2 - glycine - CO - Cl· NH3 - H2CO2 + H]+); m/z 366.9 ([4′-OH-D2 - glutamic acid - glycine - H2O - CO + H]+); m/z 350.0 ([4′-OH-D2 glutamic acid - glycine - H2O - CO - NH3 + H]+); m/z 341.9 ([4′-OH-D2 - glycine - glutamic acid - C3H3NO 2H + H]+); m/z 330.9 ([4′-OH-D2 - glutamic acid - glycine - H2O - CO - HCl + H]+); m/z 309.9 ([4′-OH-D2 - GSH - 2H + H]+); m/z 296.0 ([4′-OH-D2 - glycine - glutamic acid - C3H4NO - 2H - H2CO2 + H]+); m/z 130.5 ([glutamic acid - H2O + H]+) (Supporting Information, Figure 2B). A third conjugate was observed at m/z 854.3 (tR 10.0 min), which corresponds to 2′,6′-diglutathionyl-4′-hydroxy-dideschlorodiclofenac (4′-OH-D3). By MS/MS fragmentation of m/z 854.3, the following fragment ions were detected: m/z 836.3 ([4′-OHD3 - H2O + H]+); m/z 725.0 ([4′-OH-D3 - glutamic acid + H]+); m/z 706.7 ([4′-OH-D3 - glutamic acid - H2O + H]+);

1112

Chem. Res. Toxicol., Vol. 21, No. 5, 2008

m/z 650.4 ([4′-OH-D3 - glutamic acid - glycine + H]+); m/z 596.4 ([4′-OH-D3 - 2glutamic acid + H]+); m/z 452.3 ([4′OH-D3 - 2glutamic acid - glycine - C3H3NO + H]+); m/z 311.8 ([4′-OH-D3 - 2glycine - 2glutamic acid - C3H3NO CO - NH3 - H2O - 2H]+) (Supporting Information, Figure 3B). Two other di-GSH conjugates were detected at m/z 888.0 (tR 11.2 and 11.5 min), which corresponded to 2′,3′-diglutathionyl4′-hydroxy-deschloro-diclofenac or 2′,5′-diglutathionyl-4′-hydroxy-deschloro-diclofenac (4′-OH-D4). By MS/MS fragmentation of m/z 888.0, the following fragment ions were detected: m/z 870.2 ([4′-OH-D4 - H2O + H]+); m/z 813.4 ([4′-OH-D4 - glycine + H]+); m/z 795.4 ([4′-OH-D4 - glycine - H2O + H]+); m/z 759.1 ([4′-OH-D4 - glutamic acid + H]+); m/z 684.2 ([4′-OH-D4 - glycine - glutamic acid + H]+); m/z 630.3 ([4′OH-D4 - 2glutamic acid + H]+); m/z 611.9 ([4′-OH-D4 2glutamic acid - H2O + H]+); m/z 492.2 ([4′-OH-D4 2glutamic acid - glycine - CO - NH3 - H2O]+); m/z 464.3 ([4′-OH-D4 - 2glutamic acid - glycine - CO - NH3 CH2O2]+); m/z 456.3 ([4′-OH-D4 - 2glutamic acid - glycine - CO - NH3 - H2O - HCl]+); m/z 296.8 ([4′-OH-D4 - GSH - glycine - glutamic acid - CO - HCl - H2O + H]+) (Figure 2B). A fifth conjugate had a retention time of 12.7 min and an m/z of 922.4, which corresponded to 3′,5′-diglutathionyl-4′hydroxy-diclofenac (4′-OH-D5). By MS/MS fragmentation of m/z 922.4, the following fragment ions were detected: m/z 847.1 ([4′-OH-D5 - glycine + H]+); m/z 829.6 ([4′-OH-D5 - glycine - H2O + H]+); m/z 793.2 ([4′-OH-D5 - glutamic acid + H]+); m/z 775.0 ([4′-OH-D5 - glutamic acid - H2O + H]+); m/z 664.1 ([4′-OH-D5 - 2glutamic acid + H]+); m/z 631.1 ([4′OH-D5 - glutamic acid - glycine - C3H3NO - H2O + H]+); m/z 613.1 ([4′-OH-D5 - glutamic acid - glycine - C3H3NO - 2H2O + H]+); m/z 500.1 ([4′-OH-D5 - 2glutamic acid glycine - C3H3NO - H2O - 2H + H ]+); m/z 199.2 ([GSH SH2 - glycine]+) (Supporting Information, Figure 3C). A tri-GSH conjugate with a retention time of 7.2 min was observed at m/z 1159.4, which corresponded to 2′,3′,6′-triglutathionyl-4′-hydroxy-dideschloro-diclofenac (4′-OH-D6). By MS/MS fragmentation of m/z 1159.4, the following fragment ions were detected: m/z 1029.8 ([4′-OH-D6 - glutamic acid + H]+); m/z 901.1 ([4′-OH-D6 - 2glutamic acid + H]+); m/z 772.3 ([4′-OH-D6 - 3glutamic acid + H]+); m/z 754.3 ([4′OH-D6 - 3glutamic acid - H2O + H]+); m/z 725.2 ([4′-OHD6 - GSH - glutamic acid + H]+); m/z 652.4 ([4′-OH-D6 3glutamic acid - glycine - CO - NH3]+); m/z 610.1 ([4′OH-D6 - 3glutamic acid - glycine - C3H3NO - H2O + H]+); m/z 519.2 ([4′-OH-D6 - 3glutamic acid - 3glycine - CO + H]+); m/z 432.2 ([4′-OH-D6 - 3glutamic acid - 3glycine C3H3NO - H2CO2 + H]+) (Figure 2D). Another tri-GSH conjugate was seen at 9.3 min and detected at m/z 1193.4, which corresponded to 3′,5′,6′-triglutathionyl4′-hydroxy-deschlorodiclofenac (4′-OH-D7). By MS/MS fragmentation of m/z 1193.4, the following ions were detected: m/z 1064.8 ([4′-OH-D7 - glutamic acid + H]+); m/z 935.1 ([4′OH-D7 - 2glutamic acid + H]2+) (Supporting Information, Figure 3D). The last GSH conjugate was detected at m/z 1464.9, which corresponded to a conjugate of 4′-OH-DCl with four molecules of GSH (2′,3′,5′,6′-tetraglutathionyl-4′-hydroxy-dideschlorodiclofenac) (4′-OH-D8). Only a minor fraction was seen as the singly charged ion, with more intense ions observed at m/z 733.0 ([4′-OH-D8 + 2H]2+) and m/z 489.2 ([4′-OH-D8 + 3H]3+). From MS/MS fragmentation of m/z 1464.9, the following fragment ions were observed: m/z 1317.3 ([4′-OH-D8 -

Madsen et al. Scheme 4. The GSH Conjugates of 5-OH-Diclofenac, Which Can Be Formed from Reaction of the Quinone Imine Intermediates with GSHa

a 5-OH-D4 was not identified from electrochemical oxidation of 5-OHDCl or from the in Vitro and in ViVo samples (See Table 3).

glutamic acid - H2O + H]+); m/z 1206.0 ([4′-OH-D8 2glutamic acid + H]+) (Supporting Information, Figure 3E). MS/MS fragmentation of the doubly and triply charged GSH conjugate showed some characteristic losses of GSH conjugates and thus further supports the structure elucidation (data not shown). Characterization of GSH Conjugates Formed from Electrochemical Oxidation of 5-OH-DCl. When 5-OH-DCl was oxidized at +600 mV and the reactive intermediates were trapped with GSH, three different glutathione containing conjugates were detected, which are shown in Scheme 4. The structure elucidations were based on MS/MS experiments of the singly charged conjugates. Two conjugates eluting at 14.7 min (5-OH-D1) and 15.9 min (5-OH-D2), respectively, were observed at m/z 617.2 ([5-OHDCl-SG + H]+), which corresponded to glutathionyl-5-hydroxydiclofenac. Fragmentation of 5-OH-D1 at m/z 617.2 by MS/ MS gave the following fragment ions: m/z 542.0 ([5-OH-D1 glycine + H]+); m/z 488.0 ([5-OH-D1 - glutamic acid + H]+); m/z 385.0 ([5-OH-D1 - glycine - glutamic acid - CO + H]+); m/z 341.9 ([5-OH-D1 - glycine - glutamic acid - C3H3NO - 2H + H]+); m/z 324.0 ([5-OH-D1 - glycine - glutamic acid - C3H3NO - H2O - 2H + H]+). Similar fragment ions were observed from fragmentation of 5-OH-D2 (Supporting Information, Figure 2C (5-OH-D1) and 2D (5-OH-D2)). A GSH conjugate with a retention time of 12.4 min and a m/z of 922.4 was identified, which corresponded to diglutathionyl-5-hydroxy-diclofenac (5-OH-D3). By MS/MS fragmentation of m/z 922.4, the following fragment ions were detected: m/z 904.3 ([5-OH-D3 - H2O + H]+); m/z 847.3 ([5-OH-D3 glycine + H]+); m/z 829.3 ([5-OH-D3 - glycine - H2O + H]+); m/z 793.3 ([5-OH-D3 - glutamic acid + H]+); m/z 774.9 ([5-OH-D3 - glutamic acid - H2O + H]+); m/z 718.5 ([5OH-D3 - glutamic acid - glycine + H]+); m/z 664.3 ([5-OHD3 - 2glutamic acid + H]+); m/z 502.4 ([5-OH-D3 - 2glutamic acid - glycine - C3H3NO - H2O + H]+) (Figure 2C). From the electrochemically generated GSH conjugates, specific SRM transitions were constructed for 4′-OH-D1 through 4′-OH-D8 (Table 1 ) and 5-OH-D1 through 5-OH-D3 (Table 2). Identification of GSH Conjugates of Diclofenac, 4′-OHDCl, and 5-OH-DCl from in Vitro Incubations. DCl, 4′-OHDCl, or 5-OH-DCl were incubated with human and rat liver microsomes at concentrations of 10 and 100 µM. Furthermore, 10 and 100 µM of the compounds were incubated with freshly

BioactiVation of Diclofenac in Vitro and in ViVo Table 1. The GSH Conjugates Identified from Electrochemical Oxidation of 4′-OH-DCl and Their m/z Values, Retention Times, and SRM Transitions Used for the Identification of the Conjugates in the in Vitro and in ViWo Samples

* The triply charged ion of 4′-OH-D8 gave the highest sensitivity.

Chem. Res. Toxicol., Vol. 21, No. 5, 2008 1113 Table 2. The GSH Conjugates indentified from Electrochemical Oxidation of 5-OH-DCl and Their m/z Values, Retention Times, and SRM Transitions Used for the Identification of the Conjugates in the in Vitro and in ViWo Samples

isolated rat hepatocytes for 0, 30, 60, and 120 min. A general summary of the identified conjugates in the in Vitro incubations is listed in Table 3. 4′-OH-D1 was found from incubation of DCl with HLM at both concentrations. From incubation of 100 µM DCl with RLM, 4′-OH-D1 was also observed. The GSH conjugate could not be detected from incubation of DCl with rat hepatocytes. 4′-OH-D1 was readily detected from incubations of 4′-OH-DCl with HLM, RLM, and rat hepatocytes (selected chromatograms are shown in Supporting Information, Figure 4). 4′-OH-D2 and 5-OH-D1 were detected from incubation of 10 µM DCl with HLM, but at 100 µM, it was only possible to detect 5-OH-D1 and 5-OH-D2. In RLM and rat hepatocytes, 5-OH-D1 and 5-OH-D2 were detected. 4′-OH-D2 was detected in all the samples from incubation of 4′-OH-DCl with HLM, RLM, and rat hepatocytes. 5-OH-D1 and 5-OH-D2 were detected in all samples from incubations of 5-OH-DCl with HLM, RLM, and rat hepatocytes. However, these conjugates were also detected in the control incubations without NADPH added, which indicated that 5-OH-DCl can undergo autooxidation (selected chromatograms are shown in Figure 3 and in the Supporting Information, Figures 5-7). 4′-OH-D3 was found in incubations of DCl with HLM and RLM but not in incubations of DCl with rat hepatocytes. 4′OH-D3 was detected from incubation of 4′-OH-DCl with HLM and RLM and from incubations of 100 µM 4′-OH-DCl with rat hepatocytes (selected chromatograms are shown in the Supporting Information, Figure 8). 4′-OH-D4 was found from incubations of DCl with HLM and RLM but not in incubations of DCl with rat hepatocytes. 4′-OH-D4 was detected in incubations of 4′-OH-DCl with HLM and RLM but not from incubation of 4′-OH-DCl with rat hepatocytes (selected chromatograms are shown in Figure 4 and in the Supporting Information, Figure 9).

electrochemical oxidation of diclofenac (+1000 mV) electrochemical oxidation of 4′-OH-diclofenac (+600 mV) electrochemical oxidation of 5-OH-diclofenac (+600 mV) diclofenac, 10 µM, HLM, NADPH diclofenac, 100 µM, HLM, NADPH diclofenac, 10 µM, RLM, NADPH diclofenac, 100 µM, RLM, NADPH 4′-OH-diclofenac, 10 µM, HLM, NADPH 4′-OH-diclofenac, 100 µM, HLM, NADPH 4′-OH-diclofenac, 10 µM, RLM, NADPH 4′-OH-diclofenac, 100 µM, RLM, NADPH 5-OH-diclofenac, 10 µM, HLM, NADPH 5-OH-diclofenac, 100 µM, HLM, NADPH 5-OH-diclofenac, 10 µM, RLM, NADPH 5-OH-diclofenac, 100 µM, RLM, NADPH diclofenac, 10 µM, rat hepatocytes diclofenac, 100 µM, rat hepatocytes 4′-OH-diclofenac, 10 µM, rat hepatocytes 4′-OH-diclofenac, 100 µM, rat hepatocytes 5-OH-diclofenac, 10 µM, rat hepatocytes 5-OH-diclofenac, 100 µM, rat hepatocytes rat bile, diclofenac, 200 mg/kg

nd +

+ nd nd nd + + + +

nd nd + + +

+

+ + nd + + + + +

nd nd + + +

4′-OH-D2

nd

4′-OH-D1

+

nd nd nd +

+ + + + + + + +

+

nd

4′-OH-D3

+

nd nd nd nd

+ + + + + + + +

+

nd

4′-OH-D4

nd

nd nd nd nd

nd nd nd nd nd nd nd nd

+

nd

4′-OH-D5

nd

nd nd nd nd

+ + nd nd nd nd nd nd

+

nd

4′-OH-D6

nd

nd nd nd nd

nd nd nd nd nd nd nd nd

+

nd

4′-OH-D7

nd

nd nd nd nd

nd nd nd nd nd nd nd nd

+

nd

4′-OH-D8

+ + + +

+ + + + + + + + +

+ + + +

+ + + + + + + + +

+

+

+

+

5-OH-D2

5-OH-D1

+ + +

+ + + + + +

+ + + +

+

+

5-OH-D3

nd nd nd

nd nd nd nd nd nd

nd nd nd nd

nd

nd

5-OH-D4

Table 3. Overview of the Identified GSH Conjugates in the Samples from Incubation of Diclofenac, 4′-OH-Diclofenac, and 5-OH-Diclofenac with Microsomes (RLM or HLM) in the Presence of GSH and Rat Hepatocytes and in the Bile from a Rat Dosed with Diclofenac

1114 Chem. Res. Toxicol., Vol. 21, No. 5, 2008 Madsen et al.

BioactiVation of Diclofenac in Vitro and in ViVo

Chem. Res. Toxicol., Vol. 21, No. 5, 2008 1115

Figure 3. Chromatograms of 4′-OH-D2 (tR 15.0 min), 5-OH-D1 (tR 14.7 min) and 5-OH-D2 (tR 15.9 min) from the samples of (A) electrochemical oxidation of 4′-OH-DCl at +600 mV followed by collection in an aqueous solution of GSH, (B) electrochemical oxidation of 5-OH-DCl at +600 mV followed by collection in an aqueous solution of GSH, (C) bile from a rat dosed with diclofenac (200 mg/kg), (D) incubation of 10 µM diclofenac with HLM, GSH, and NADPH, (E) incubation of 10 µM diclofenac with RLM, GSH, and NADPH, (F) incubation of 10 µM diclofenac with rat hepatocytes.

4′-OH-D5 could not be detected in any of the incubation samples. However, 5-OH-D3 was detected from incubation of DCl with HLM, RLM, and rat hepatocytes. Furthermore, 5-OHD3 was also detected from incubation of 5-OH-DCl with HLM, RLM, and rat hepatocytes (selected chromatograms are shown in Figure 4 and in the Supporting Information, Figure 10). Another peak was observed in the chromatograms of the samples from incubation with rat hepatocytes with a retention time 11.3 min. From MS/MS fragmentation of m/z 922.4, it was not possible to assign any fragment ions, which corresponded to a GSH-containing compound. 4′-OH-D6 was previously identified in low concentration from incubation of DCl with HLM, but could not be shown in these experiments. 4′-OH-D7, 4′-OH-D8, and 5-OH-D4 could not be detected in any of the samples. In ViWo Identification of GSH Conjugates of Diclofenac. Rats were dosed with DCl (200 mg/kg), and bile was collected from 0 to 4 h. The bile was analyzed by LC-MS/MS (SRM) for the eight 4′-OH-DCl GSH conjugates (4′-OH-D1-D8) and for 5-OH-D1, 5-OH-D2, and 5-OH-D3 by the use of the same SRM transitions as used for the analysis of the samples from incubation with liver microsomes and hepatocytes. The following GSH conjugates were identified in the rat bile: 4′-OH-D1

to 4′-OH-D4 and 5-OH-D1 to 5-OH-D3 (chromatograms are shown in Figure 3, Figure 4 and Supporting Information, Figure 11). Furthermore, other DCl metabolites were identified in the collected bile. Two compounds were detected at m/z 633.0, which corresponds to GSH conjugates of dihydroxylated DCl. The identity of the compounds was confirmed by MS/MS fragmentation and the following fragment ions were detected: m/z 557.9 ([(OH)2-DCl-SG - glycine + H]+); m/z 503.8 ([(OH)2-DCl-SG - glutamic acid + H]+); m/z 486.0 ([(OH)2DCl-SG - glutamic acid - H2O + H]+); m/z 400.9 ([(OH)2DCl-SG - glutamic acid - glycine - CO + H]+); m/z 382.9 ([(OH)2-DCl-SG - glutamic acid - glycine - H2O - CO + H]+); m/z 358.0 ([(OH)2-DCl-SG - glycine - glutamic acid - C3H3NO - 2H + H]+); m/z 326.2 ([(OH)2-DCl-SG - GSH - 2H + H]+); m/z 145.3 ([(OH)2-DCl-SG - (OH)2-DCl - SH - glutamic acid + H]+) (Figure 5A). Two DCl metabolites were detected at m/z 661.3. The m/z value corresponded to the GSH conjugate of the dimethylated dihydroxydiclofenac. By MS/MS fragmentation of m/z 661.3, the following fragment ions were detected: m/z 615.3 ([(OCH3)2DCl-SG - CH2O2 + H]+); m/z 585.8 ([(OCH3)2-DCl-SG glycine + H]+); m/z 531.8 ([(OCH3)2-DCl-SG - glutamic acid

1116

Chem. Res. Toxicol., Vol. 21, No. 5, 2008

Madsen et al.

Figure 4. Chromatograms of 4′-OH-D4 (tR 11.2 and 11.5 min) and 5-OH-D3 (tR 12.6 min): (A) 4′-OH-D4 obtained from electrochemical oxidation of 4′-OH-DCl at +600 mV in the presence of GSH; (B) 4′-OH-D4 obtained from the bile from a rat dosed with diclofenac (200 mg/kg); (C) 4′-OH-D4 obtained from incubation of 10 µM diclofenac with HLM, GSH, and NADPH; (D) 4′-OH-D4 obtained from the incubation of 10 µM 4′-OH-DCl with RLM, GSH, and NADPH; (E) 5-OH-D3 obtained from electrochemical oxidation of 5-OH-DCl at +600 mV in the presence of GSH; (F) 5-OH-D3 obtained from the bile from a rat dosed with diclofenac (200 mg/kg); (G) 5-OH-D3 obtained from the incubation of 100 µM diclofenac with HLM, GSH, and NADPH; (H) 5-OH-D3 obtained from the incubation of 100 µM diclofenac with RLM, GSH, and NADPH.

+ H]+); m/z 513.6 ([(OCH3)2-DCl-SG - glutamic acid - H2O + H]+); m/z 445.0 ([(OCH3)2-DCl-SG - (OCH3)2-DCl - H2O - C8H9NO3 (2-(2-aminophenyl)acetic acid moiety) - OCH3]+); m/z 410.9 ([(OCH3)2-DCl-SG - glutamic acid - glycine CH2O2 + H]+) (Figure 5B). Stability of 4′-OH-DCl and 5-OH-DCl. Two aspects of the stability of 4′-OH-DCl and 5-OH-DCl were investigated, which were the possibilities of (i) auto-oxidation and (ii) intramolecular ring cyclization. To investigate whether DCl, 4′-OH-DCl, or 5-OH-DCl underwent auto-oxidation, the compounds were incubated with

0.1 M phosphate buffer (pH 7.4), 5 mM GSH, and blank rat bile at 37 °C for 30 min. From the incubation of DCl and 4′-OH-DCl, it was not possible to detect any formation of GSH conjugates. However, incubation of 5-OH-DCl with GSH in buffer and blank rat bile resulted in the formation of 5-OH-D1, 5-OH-D2, and 5-OHD3. Since 5-OH-DCl formed GSH conjugates without metabolic activation, 5-OH-DCl seems to undergo auto-oxidation, which may be due to a lower oxidation potential of this compound. The electrochemical properties of 4′-OH-DCl, 5-OH-DCl, and

BioactiVation of Diclofenac in Vitro and in ViVo

Chem. Res. Toxicol., Vol. 21, No. 5, 2008 1117

by MS/MS fragmentation of m/z 599.0, and the following fragment ions were detected: m/z 524.0 ([OH-indolinone-DClSG - glycine + H]+); m/z 470.0 ([OH-indolinone-DCl-SG glutamic acid + H]+); m/z 453.0 ([OH-indolinone-DCl-SG glutamic acid - NH3 + H]+); m/z 367.0 ([OH-indolinone-DClSG - glycine - glutamic acid - CO]+); m/z 349.9 ([OHindolinone-DCl-SG - glycine - glutamic acid - CO NH3]+); m/z 323.9 ([OH-indolinone-DCl-SG - glycine glutamic acid - C3H3NO - 2H + H]+) (Supporting Information, Figure 12A,B). Furthermore, a metabolite was identified at m/z 904.4, which corresponds to a diglutathionyl conjugate of the indolinone derivative. This conjugate was also detected as a doubly charged ion at m/z 452.8. By MS/MS fragmentation of m/z 904.4, the following fragment ions were detected: m/z 885.7 ([OH-indolinone-DCl-(SG)2 - H2O + H]+); m/z 828.8 ([OH-indolinone-DCl-(SG)2 - glycine + H]+); m/z 775.1 ([OH-indolinone-DCl-(SG)2 - glutamic acid + H]+); m/z 699.8 ([OH-indolinone-DCl-(SG)2 - glycine glutamic acid + H]+); m/z 645.9 ([OH-indolinone-DCl-(SG)22glutamic acid + H]+); m/z 631.2 ([OH-indolinone-DCl-(SG)2 glycine - glutamic acid - C3H3NO + H]+); m/z 481.8 ([OHindolinone-DCl-(SG)2 - 2glutamic acid - glycine - C3H3NO H2O - 2 H + H]+); m/z 380.9 ([OH-indolinone-DCl-(SG)2 2glutamic acid - 2glycine - C3H3NO - H2O - CO + H]+) (Supporting Information, Figure 12C).

Discussion

Figure 5. MS/MS spectra of two novel GSH conjugates identified from rat bile: (A) m/z 633.0 ((OH)2DCl-SG); (B) m/z 661.3 ((OCH3)2-DClSG).

Figure 6. Cyclic voltammograms of 0.2 mM 4′-OH-DCl (black), 5-OHDCl (thick gray), and paracetamol (thin gray) in 0.1 M phosphate buffer (pH 7.4), recorded with a scan speed of 100 mV/s.

paracetamol (as a reference compound) were therefore investigated by cyclic voltammetry. The cyclic voltammograms are shown in Figure 6. The oxidation potential was determined to +167 mV for 5-OH-DCl and +236 mV for 4′-OH-DCl. The reduction potentials were –439 and –131 mV for 5-OH-DCl and 4′-OH-DCl, respectively. Paracetamol had an oxidation potential of +392 mV and a reduction potential of –50 mV. Two metabolites were identified in the samples from electrochemical oxidation and bile at m/z 599.0 with retention times at 12.5 and 13.6 min. This m/z value corresponds to a GSH conjugate of an indolinone derivative formed from intramolecular cyclization. The identity of the conjugate was confirmed

The use of electrochemistry for mimicking some types of metabolic reactions, including the formation of reactive metabolic intermediates, has been reported by several authors (8–16). In the present study, the electrochemical oxidation of DCl, 4′OH-DCl, and 5-OH-DCl was studied. 4′-OH-DCl and 5-OHDCl are believed to be metabolized to quinone imines, which can react with GSH and other biological relevant nucleophiles. In the first part of the study, it was possible to mimic the CYP3A4 catalyzed hydroxylation of DCl to 5-OH-DCl by electrochemical oxidation. For comparison, a standard of 5-OHDCl was synthesized, and the MS/MS and NMR data were in agreement with the data obtained from the compound purified from electrochemical oxidation. This illustrates the advantage of using electrochemical oxidation as an alternative method for generating standards of metabolites. 5-OH-DCl was prepared by a single oxidation in the electrochemical cell. Contrary to this, the traditional organic synthesis of 5-OH-DCl involved six different steps, which is a time-consuming process. The second aim of the study was focused on the electrochemical oxidation of 4′-OH-DCl and 5-OH-DCl to their corresponding quinone imines followed by reaction with GSH. Two different experimental designs were used, and from these experiments, it was possible to identify and characterize a wide range GSH conjugates of 4′-OH-DCl and 5-OH-DCl by LCMS and LC-MS/MS. The mono-GSH conjugates were formed from a reaction between the electrochemically generated quinone imine of both 4′-OH-DCl and 5-OH-DCl and GSH. The di- and tri-GSH conjugates were either formed from a simultaneous attack of GSH or more likely, from a reoxidation of the monoGSH conjugate, which then reacted with another molecule of GSH. The di- and tri-GSH conjugates were only formed when the oxidation of the 4′-OH-DCl and 5-OH-DCl was performed with GSH added to the sample during the electrochemical oxidation. From the electrochemically generated GSH conjugates of 4′OH-DCl and 5-OH-DCl, it was possible to develop sensitive SRM transitions for each GSH conjugate, which facilitated

1118

Chem. Res. Toxicol., Vol. 21, No. 5, 2008

detection of the conjugates in the in Vitro and in ViVo experiments. A range of GSH conjugates were identified, and the results are summarized in Table 3. This is, to our knowledge, the first time that it has been reported that DCl can form diand triglutathionyl conjugates. The proposed pathway for the GSH conjugates observed in the in Vitro microsomal samples and in ViVo samples is shown in Schemes 3 and 4. It was not possible to deduce whether the formation of an arene oxide or a benzoquinone imine was the initial reactive intermediate in the enzymatically catalyzed oxidation. If the arene oxide was involved in the first bioactivation step, the formed GSH conjugate could possibly be enzymatically reoxidized to the benzoquinone imine intermediate, which would be able to react with another molecule of GSH. The reactive quinone imine of 5-OH-DCl would in theory be able to form three different mono-GSH conjugates, but only two of these were identified. It is possible that two of the conjugates coeluted in one of the identified peaks. However, Tang and co-workers identified and isolated only two GSH conjugates from 5-OH-DCl. The two GSH conjugates were characterized by NMR, and it was shown that GSH reacted selectively at the carbon ortho to the hydroxy group (position C-4 and C-6) (2). It has also been shown that GSH reacts with the quinone imine of paracetamol selectively at the carbon ortho to the hydroxy group (16). This would indicate that the carbon ortho to the amino group is less reactive than the carbon ortho to the hydroxy group. These findings are further supported by the fact that only one diglutathionyl conjugate was identified from 5-OH-DCl, indicating that position C-3 is less reactive than position C-4 and C-6. It was possible to identify GSH conjugates of 4′-OH-DCl in which GSH had reacted ortho to the amino-group (position C-2′). This may be explained by the electronegativity of the chloro atom, which makes the position C-2′ more electrophilic, thus increasing the reactivity toward GSH. It was observed that GSH conjugates of 5-OH-DCl were formed without adding NADPH as a cofactor in the microsomal incubations. In addition, GSH conjugates were formed from incubation of 5-OH-DCl with GSH in a 0.1 M phosphate buffer, pH 7.4, and in blank bile. This was also observed by Kenny et al. and Tang et al., who reported that GSH conjugates were formed from auto-oxidation of 5-OH-DCl but not of 4′-OHDCl (3, 17). From experiments with cyclic voltammetry, it was possible to estimate the oxidation potential of 5-OH-DCl and 4′-OH-DCl, but surprisingly there was only a minor difference in the recorded oxidation potentials. Therefore, other factors may contribute to the observed selective auto-oxidation of 5-OHDCl. These results are interesting, since both 4′-OH-DCl and 5-OH-DCl are formed in Vitro and in ViVo, but evidently 5-OHDCl does not need to be enzymatically reoxidized before it can react with a nucleophile, in contrast to 4′-OH-DCl, which has to be enzymatically reoxidized before it can react with a nucleophile. If reactive metabolites are involved in the hepatotoxicity of DCl, 5-OH-DCl appears to be an important metabolite, since 5-OH-DCl is prone to auto-oxidation and thus capable of reacting with nucleophiles multiple times even without enzymatic oxidation. This means that 5-OH-DCl can react with proteins, and without enzymatic reoxidation, it may react with another functional group in the protein leading to cross-links in the protein. From these results and our previous studies (16), we have shown that electrochemistry is a suitable tool for studying some

Madsen et al.

of the important metabolic pathways of drugs. Therefore, it would also be an applicable tool to use in drug discovery, for example, for generating large amounts of a metabolite for structure elucidation by NMR as shown for 5-OH-diclofenac and clozapine-SG (16), studying the formation of reactive metabolites (16), or studying the reactivity of reactive metabolites as reported for paracetamol, for example (16). In addition, electrochemistry may also be a suitable tool for studying compounds that are mechanism based inhibitors. It is not always possible to trap the reactive intermediates of mechanism inhibitors, but it may be possible to generate these metabolites by electrochemical reactions and then further study them in detail. Further research in these areas would be beneficial to exploit the use of electrochemistry in the area of drug development. In conclusion, the advantage of using electrochemistry for mimicking the hydroxylation of DCl to 5-OH-DCl was demonstrated. Furthermore, electrochemistry was used to generate mono-, di- and triglutathionyl conjugates of 4′-OH-DCl and 5-OH-DCl, and by using sensitive LC-MS-methods, it was possible to identify these same GSH conjugates in microsomal incubation (HLM/RLM), in rat hepatocytes incubations, and in rat bile from in ViVo studies. These experiments have demonstrated that electrochemical cells can be used for small-scale synthesis of compounds that otherwise can be difficult to purify from complex matrices or synthesize by traditional means. Furthermore, electrochemistry may be used to generate conjugates that can aid in selective search for the same conjugates in biological samples. Acknowledgment. We thank Scott Fauty at Merck & Co. (West Point, PA) for helping with the collection of the bile samples. The Danish Medical Research Council Grant No. 2202-0340, The Danish Technical Research Council Grant No. 56-01-0014, and Lundbeckfonden Grant No. 142/02 are acknowledged for financial support. Supporting Information Available: Selected chromatograms, MS/MS spectra, and NMR data. This material is available free of charge via the Internet at http://pubs.acs.org.

References (1) Tang, W. (2003) The metabolism of diclofenac - enzymology and toxicology perspectives. Curr. Drug Metab. 4, 319–329. (2) Tang, W., Stearns, R. A., Bandiera, S. M., Zhang, Y., Raab, C., Braun, M. P., Dean, D. C., Pang, J., Leung, K. H., Doss, G. A., Strauss, J. R., Kwei, G. Y., Rushmore, T. H., Chiu, S. H. L., and Baille, T. A. (1999) Studies on cytochrome P-450-mediated bioactivation of diclofenac in rats and in human hepatocytes: Identification of glutathione conjugated metabolites. Drug Metab. Dispos. 27 (3), 365–372. (3) Tang, W., Stearns, R. A., Wang, R. W., Chiu, S. H., and Baillie, T. A. (1999) Roles of human hepatic cytochrome P450s 2C9 and 3A4 in the metabolic activation of diclofenac. Chem. Res. Toxicol. 12 (2), 192–199. (4) Poon, G. K., Chen, Q., Teffera, Y., Ngui, J. S., Griffin, P. R., Braun, M. P., Doss, G. A., Freeden, C., Stearns, R. A., Evans, D. C., Baille, T. A., and Tang, W. (2001) Bioactivation of diclofenac via benzoquinone imine intermediates-identification of urinary mercapturic acid derivates in rats and humans. Drug Metab. Dispos. 29 (12), 1608– 1613. (5) Yu, L. J., Chen, Y., DeNinno, M. P., O’Conell, T. N., and Hop, C.E.C.A. (2005) Identification of a novel glutathione adduct of diclofenac, 4′-hydroxy-2′-glutathion-deschloro-diclofenac, upon incubation with human liver microsomes. Drug Metab. Dispos. 33 (4), 484–488. (6) Shen, S., Marchick, M. R., Davis, M. R., Doss, G. A., and Pohl, L. R. (1999) Metabolic activation of diclofenac by human cytochrome P450 3A4: Role of 5-hydroxydiclofenac. Chem. Res. Toxicol. 12, 214–222. (7) Yan, Z., Li, J., Huebert, N., Caldwell, G. W., Du, Y., and Zhong, H. (2005) Detection of a novel reactive metabolite of diclofenac: evidence

BioactiVation of Diclofenac in Vitro and in ViVo

(8) (9)

(10)

(11)

(12)

for CYP2C9-mediated bioactivation via arene oxides. Drug Metab. Dispos. 33 (6), 706–713. Volk, K. J., Yost, R. A., and Brajtertoth, A. (1992) Electrochemistry on Line with Mass-Spectrometry - Insight Into Biological Redox Reactions. Anal. Chem. 64 (1), A21–A33. Deng, H. T., and Van Berkel, G. J. (1999) A thin-layer electrochemical flow cell coupled on-line with electrospray-mass spectrometry for the study of biological redox reactions. Electroanalysis 11 (12), 857–865. Getek, T. A., Korfmacher, W. A., McRae, T. A., and Hinson, J. A. (1989) Utility of solution electrochemistry mass spectrometry for investigating the formation and detection of biologically important conjugates of acetaminophen. J. Chromatogr. 474 (245), 256. Jurva, U., Wikström, H. V., Weidolf, L., and Bruins, A. P. (2003) Comparison between electrochemistry/mass spectrometry and cytochrome P450 catalyzed oxidation reactions. Rapid Commun. Mass Spectrom. 17, 800–810. Jurva, U., Wikstrom, H. V., and Bruins, A. P. (2000) In vitro mimicry of metabolic oxidation reactions by electrochemistry/mass spectrometry. Rapid Commun. Mass Spectrom. 14 (6), 529–533.

Chem. Res. Toxicol., Vol. 21, No. 5, 2008 1119 (13) van Leeuwen, S. M., Blankert, B., Kauffmann, J. M., and Karst, U. (2005) Prediction of clozapine metabolism by on-line electrochemistry/ liquid chromatography/mass spectrometry. Anal. Bioanal. Chem. 382 (3), 742–750. (14) Johansson, T., Weidolf, L., and Jurva, U. (2007) Mimicry of phase I drug metabolism–novel methods for metabolite characterization and synthesis. Rapid Commun. Mass Spectrom. 21 (14), 2323–2331. (15) Lohmann, W., and Karst, U. (2007) Generation and identification of reactive metabolites by electrochemistry and immobilized enzymes coupled on-line to liquid chromatography/mass spectrometry. Anal. Chem. 79 (17), 6831–6839. (16) Madsen, K. G., Olsen, J., Skonberg, C., Hansen, S. H., and Jurva, U. (2007) Development and Evaluation of an Electrochemical Method for Studying Reactive Phase-I Metabolites: Correlation to in Vitro Drug Metabolism. Chem. Res. Toxicol. 20 (5), 821–831. (17) Kenny, J. R., Maggs, J. L., Meng, X., Sinnott, D., Clarke, S. E., Park, B. K., and Stachulski, A. V. (2004) Syntheses and Characterization of the Acyl Glucuronide and Hydroxy Metabolites of Diclofenac. J. Med. Chem. 47, 2816–2825.

TX700419D