Identification in Rat Bile of Glutathione Conjugates of Fluoromethyl 2,2

Recent studies have indicated that the nephrotoxicity of fluoromethyl 2,2-difluoro-1-(trifluoromethyl)vinyl ether (“Compound A”), a breakdown product ...
5 downloads 14 Views 368KB Size
Download PDF
Chem. Res. Toxicol. 1996, 9, 555-561

555

Identification in Rat Bile of Glutathione Conjugates of Fluoromethyl 2,2-Difluoro-1-(trifluoromethyl)vinyl Ether, a Nephrotoxic Degradate of the Anesthetic Agent Sevoflurane Lixia Jin,† Margaret R. Davis,†,‡ Evan D. Kharasch,†,§ George A. Doss,| and Thomas A. Baillie*,†,‡ Departments of Medicinal Chemistry and Anesthesiology, University of Washington, Seattle, Washington 98195, and Department of Drug Metabolism, Merck Research Laboratories, RY80L-109, Rahway, New Jersey 07065 Received September 19, 1995X

Recent studies have indicated that the nephrotoxicity of fluoromethyl 2,2-difluoro-1(trifluoromethyl)vinyl ether (“Compound A”), a breakdown product of the inhaled anesthetic sevoflurane, may be mediated by a reactive intermediate(s) generated via the cysteine conjugate β-lyase pathway. In order to gain a better understanding of glutathione (GSH)-dependent metabolism of Compound A, the present study was carried out with the primary goal of detecting and characterizing Compound A-GSH conjugates. By means of ionspray LC-MS/ MS and NMR spectroscopy, a total of four GSH conjugates (“A1-A4”) were identified from the bile of rats dosed intraperitoneally with Compound A. A1 and A2 were identified as two diastereomers of S-[1,1-difluoro-2-(fluoromethoxy)-2-(trifluoromethyl)ethyl]glutathione, while A3 and A4 were identified as (E)- and (Z)-S-[1-fluoro-2-(fluoromethoxy)-2-(trifluoromethyl)vinyl]glutathione, respectively. Quantitative analyses indicated that approximately 29% of the administered dose of Compound A was excreted into the bile in the form of the above GSH conjugates over a period of 6 h. Studies conducted in vitro demonstrated that the reaction of Compound A with GSH was catalyzed by both rat liver cytosolic and microsomal glutathione S-transferases (GST), with the two enzyme systems exhibiting different product selectivities. Formation of these GSH conjugates also occurred nonenzymatically at an appreciable rate. These results indicate that spontaneous and enzyme-mediated conjugation with GSH represents a major pathway of metabolism of Compound A in rats. Conjugation of Compound A with GSH in vivo appeared to be catalyzed preferentially by microsomal rather than cytosolic GST, based on comparison of biliary, microsomal, and cytosolic metabolic profiles. By analogy with other haloalkenes, further metabolism of the corresponding cysteine conjugates of Compound A by renal cysteine conjugate β-lyase may lead to the formation of reactive acylating agents, which would be expected to bind covalently to cellular macromolecules and cause organ-selective nephrotoxicity.

Introduction Sevoflurane [fluoromethyl 2,2,2-trifluoro-1-(trifluoromethyl)ethyl ether] is a new fluorinated inhalation anesthetic which was introduced into clinical practice in Japan in 1990 and was approved recently in the United States. Sevoflurane has a nonirritating, pleasant odor, minimal effect on heart rate, and low blood gas partition coefficient which leads to a rapid induction of and recovery from anesthesia (1, 2). However, sevoflurane undergoes partial decomposition in anesthesia machines to a haloalkene derivative, fluoromethyl 2,2-difluoro-1(trifluoromethyl)vinyl ether, commonly referred to as “Compound A” (3-5). Compound A has been found to be nephrotoxic in rats at high inhaled concentrations (57), and this observation has raised concerns about the safety of sevoflurane in humans (4, 8). * Corresponding author. † Department of Medicinal Chemistry, University of Washington. ‡ Present address: Department of Drug Metabolism, Merck Research Laboratories, WP26A-2044, West Point, PA 19486. § Department of Anesthesiology, University of Washington. | Merck Research Laboratories. X Abstract published in Advance ACS Abstracts, February 15, 1996.

0893-228x/96/2709-0555$12.00/0

Sevoflurane reacts with the strong base soda lime [NaOH + KOH + Ca(OH)2] or Baralyme [Ba(OH)2 + Ca(OH)2 + KOH] in the carbon dioxide absorbents of anesthesia machines, resulting in loss of the elements of hydrogen fluoride from the isopropyl moiety and formation of Compound A (3-5, 9, 10). Several investigations have demonstrated that the decomposition of sevoflurane to Compound A by soda lime and Baralyme and in clinical anesthesia circuits is time-, temperature-, and concentration-dependent (3-5, 11-17). Concentrations of Compound A are higher in low-flow and closedcircuit anesthesia systems (4, 13-15) than in semiclosed systems (3, 16), where relatively higher flow rates of fresh gases are employed. Concentrations of Compound A typically observed during low-flow sevoflurane anesthesia are 20-40 ppm (4, 15, 16). During surgery, the inhaled concentrations of Compound A are significantly higher than the exhaled concentrations, indicative of uptake by the patient (4, 16). When rats were exposed to Compound A, the LC50 (lethal concentrations in 50% of animals) were 331, 203, and 127 ppm for 3-, 6-, and 12-h exposure periods, respectively (6, 7). The kidney was found to be the primary target organ for toxicity. Elevations in serum © 1996 American Chemical Society

556 Chem. Res. Toxicol., Vol. 9, No. 2, 1996

blood urea nitrogen (BUN)1 and excretion of glucose, protein, and ketone bodies were observed in rats exposed to high concentrations of Compound A (5). Histological examinations of tissues from exposed animals revealed degeneration and necrosis of tubules in the outer strip of the outer medulla (corticomedullary junction), which occurred at inhaled concentrations of 25-50 ppm and greater (5-7). The toxic potential of Compound A in humans remains unclear at the present time. In patients receiving sevoflurane anesthesia, no adverse effects on renal function were observed under conditions where Compound A was generated in the anesthetic circuit (4, 14-17). However, further information is required before conclusions can be drawn concerning the safety of low-flow sevoflurane anesthesia in humans. To this end, we have conducted studies on the metabolic fate of Compound A in rats, and we have explored the possible role of metabolites as mediators of Compound A nephrotoxicity. Based upon the structural similarity between Compound A and the known nephrotoxic haloalkenes tetrafluoroethene (18, 19) and chlorotrifluoroethene (20-23), we proposed that Compound A might cause kidney damage via metabolism to reactive intermediates by the renal cysteine conjugate β-lyase system (24). Thus, on account of its electrophilic character, Compound A would be expected to react with GSH in vivo to form one or more GSH adducts. Further metabolism of these conjugates to the corresponding cysteine adducts, followed by cleavage via renal cysteine conjugate β-lyase, would afford highly reactive, electrophilic species which might bind covalently to cellular macromolecules and produce organ-selective renal injury. This general mechanism is now believed to underlie the nephrotoxic effects of several structurally-related haloalkenes (25-27). The results of a preliminary study on the disposition of Compound A in rats provided experimental support for the above hypothesis, in as much as two types of Compound A-GSH conjugate and two types of Compound A-mercapturic acid conjugate were detected in bile and urine, respectively (24). Moreover, (aminooxy)acetic acid, a competitive inhibitor of renal cysteine conjugate β-lyase (28), was shown to provide partial protection against Compound A-induced diuresis and proteinuria in rats. However, the exact nature of the Compound A-GSH conjugates was not determined in this pilot study. In order to gain a better understanding of the role of GSH in the metabolism and renal toxicity of Compound A, the objectives of the present investigation were as follows: (i) to fully characterize the biliary Compound A-GSH conjugates by means of combined liquid chromatography-tandem mass spectrometry (LCMS/MS) and nuclear magnetic resonance (NMR) spectroscopy; (ii) to quantify the excretion of these conjugates in rat bile; and (iii) to study the reaction of Compound A with GSH in vitro in order to establish the relative importance of microsomal and cytosolic GST enzymes as catalysts in the formation of the observed conjugates.

Experimental Procedures Materials. Compound A was obtained from Central Glass, Ube City, Japan (99.92% pure with 500 ppm 2,6-di-tert-butyl1 Abbreviations: BUN, blood urea nitrogen; GST, glutathione Stransferase; SDETG, S-(N,N-diethylthiocarbamoyl)glutathione; TFA, trifluoroacetic acid; LC-MS/MS, liquid chromatography-tandem mass spectrometry; CID, collisionally-induced dissociation; SRM, selected reaction monitoring.

Jin et al. 4-methylphenol added as a preservative). The GSH conjugates “A1” and “A2” {diastereomers of S-[1,1-difluoro-2-(fluoromethoxy)2-(trifluoromethyl)ethyl]glutathione}, “A3” {(E)-S-[1-fluoro-2(fluoromethoxy)-2-(trifluoromethyl)vinyl]glutathione}, and “A4” {(Z)-S-[1,1-difluoro-2-(fluoromethoxy)-2-(trifluoromethyl)ethyl]glutathione} were obtained by synthesis, as outlined below. S-(N,N-Diethylthiocarbamoyl)glutathione (SDETG) was prepared as described previously (29). Other chemicals were purchased from commercial sources and were of analytical grade. Instrumentation and Analytical Methods. 1H and 19F NMR spectra were recorded on either a Varian VXR 300, Unity 400, or Unity 500 spectrometer (Varian Associates, Palo Alto, CA). Samples were dissolved in Me2SO-d6, and proton chemical shifts are expressed in parts per million (δ) downfield from tetramethylsilane. Chemical shifts for fluorines were calculated relative to internal trifluoroacetic acid (δ -78.5 ppm). Signal multiplicities are reported as follows: s, d, t, qn (quintet), dd (doublet of doublets), dqn (doublet of quintets), and m. Liquid chromatography-tandem mass spectrometry (LC-MS/ MS) was carried out on a Perkin-Elmer Sciex API III triple quadrupole mass spectrometer equipped with an atmospheric pressure ion source and an IonSpray interface. Analyses were performed with an ionizing voltage of 5 kV, and high-purity air was used as the nebulizing gas at an operating pressure of 40 psi. Collisionally-induced dissociation (CID) of selected precursor ions was performed in the rf-only quadrupole region where argon was employed as target gas at a thickness of 1.8 × 1014 molecules cm-1. Chromatographic separations of the bile from rats dosed with Compound A were performed on an Hewlett Packard 1090 liquid chromatographic system which includes a photodiode array detector. Compound A-related GSH conjugates were detected by LC-MS/MS using the constant neutral loss scanning technique (loss of 129 Da) (30). Specimens of filtered bile (2 µL) were injected onto a Beckman Ultrasphere narrow-bore C18 column (150 mm × 2.0 mm i.d.) coupled via a splitter to both the mass spectrometer and the photodiode array detector. The split ratio (mass spectrometer:diode array) was 3:1. The mobile phase, which consisted of a mixture of solvent A (MeOH) and solvent B (5 mM aqueous NH4OAc containing 0.045% trifluoroacetic acid (TFA), pH 3.0), was delivered at a constant flow rate of 200 µL min-1. The following gradient was employed for analyses: 5% solvent A for 10 min, followed by a linear increase in solvent A at a rate of 0.5% min-1 to 25% solvent A, and then at 2% min-1 to 45% solvent A. Once candidate GSH conjugates had been detected by the constant neutral loss scanning LC-MS/MS approach, bile samples were reanalyzed in order to record their product ion mass spectra by CID of the respective MH+ species. Finally, as authentic samples of the conjugates-of-interest became available, the identity of each biliary GSH adducts was verified by LC-MS/ MS analyses in which metabolites were demonstrated to possess the same LC retention times and fragmentation behavior as the corresponding reference compounds. Quantitative analyses of Compound A-related GSH conjugates in bile were carried out by selected reaction monitoring (SRM) LC-MS/MS. Specimens of bile (40 µL) were treated with internal standard (SDETG, 40 µg) and diluted (to a final volume of 1 mL) with HPLC solvent B. Aliquots (25 µL) of these samples were injected onto the HPLC column where conjugates were separated using the gradient system described above. The analytes and internal standard were detected by monitoring the transitions m/z 488 f 299 for A1 and A2 (24), m/z 468 f 339 for A3 and A4 (24), and m/z 423 f 277 for SDETG (29). The ratios of the areas of metabolites to that of the internal standard were employed to determine the amounts of conjugates in bile sample with reference to calibration curves which were prepared by adding varying amounts of reference compounds together with a fixed amount of SDETG to specimens of drug-free bile. Analyses of Compound A-GSH conjugate formation in vitro were performed on a Shimadzu LC-10A liquid chromatograph (Shimadzu, Kyoto, Japan) which consisted of two LC-10AS solvent delivery units, an SCL-10A system controller, an SPD-

Bioactivation of Sevoflurane Compound A 10AV UV-VIS detector, and an SIL-10A autoinjector. A Supelco ODS column (15 cm × 4.6 mm i.d., 3 µm) (Supelco, Bellefonte, PA) was used for separations with MeOH/10 mM potassium phosphate buffer, pH 3.0 (19:81 v/v), as mobile phase. The flow rate was 1 mL min-1, and compounds eluting from the column were detected by monitoring the UV absorbance at 214 nm. SDETG was used as internal standard, and quantitation was achieved by reference to calibration curves. Synthesis of GSH Conjugates. Glutathione (3 mmol) was dissolved in H2O (50 mL), and the pH of the solution was adjusted to 8.0 with 10 M NaOH. A solution of Compound A (2 mmol) in MeOH (50 mL) was added, and the resulting mixture was stirred at room temperature overnight. The products were concentrated under reduced pressure and subjected to semipreparative HPLC (C18, 250 mm × 10 mm i.d.), which afforded A3, A4, and a mixture of A1 and A2 [mobile phase MeOH/5 mM aqueous NH4OAc containing 0.045% TFA (3:7 v/v)]. The latter fraction containing A1 and A2 was purified further by HPLC [mobile phase MeOH/5 mM aqueous NH4OAc containing 0.045% TFA (1:9 v/v)] to afford pure samples of the two isomers. The individual GSH conjugates were characterized by 1H and 19F NMR and by MS/MS, as follows: Conjugate A1: [1,1-Difluoro-2-(fluoromethoxy)-2-(trifluoromethyl)ethyl]glutathione. 1H NMR: δ 1.91 (m, 2H, Glu-β,β′), 2.33 (m, 2H, Glu-γ,γ′), 3.04 (dd, 1H, J ) 10.1 and 12.9 Hz, Cys-β), 3.35 (dd, 1H, J ) 4.7 and 13.4 Hz, Cys-β′), 3.39 (t, 1H, J ) 6.7 Hz, Glu-R), 3.71 (d, 2H, J ) 5.4 Hz, Gly-R,R′), 4.50 (m, 1H, Cys-R), 5.43 (m, 1H, CH2FO), 5.59 (m, 2H, CH2FO and CH2FOCH(CF3)CF2), 8.63 (d, 1H, J ) 8.3 Hz, CONH), and 8.75 (t, 1H, J ) 5.5 Hz, CONH). 19F NMR: δ -153.49 (t, 1F, J ) 53.7 Hz, CH2F), -87.30 (dqn, 1F, J ) 224.2 and 10.4 Hz, CF2), -84.94 (dqn, 1F, J ) 224.2 and 8.7 Hz, CF2), and -76.45 (m, 3F, CF3). MS/MS (CID of MH+ at m/z 488): m/z 448 ([M + H - 2HF]+), 471 ([M + H - NH3]+), 413 ([M + H - Gly]+), 373 ([M + H - Gly - 2HF]+), 359 ([M + H - pyroglutamic acid]+), 319 ([M + H - pyroglutamic acid - 2HF]+), 299 ([CH2FOCH(CF3)CF2SCH2C(NH2)CONH2]+), 256 ([CH2FOCH(CF3)CF2SCH2CHdNH2]+), 216 ([CH2FOCH(CF3)CF2SCH2CHdNH2 - 2HF]+), and 130 ([pyroglutamic acid + H]+). Conjugate A2: [1,1-Difluoro-2-(fluoromethoxy)-2-(trifluoromethyl)ethyl]glutathione. 1H NMR: δ 1.95 (m, 2H, Glu-β,β′), 2.34 (m, 2H, Glu-γ,γ′), 3.03 (dd, 1H, J ) 9.4 and 13.1 Hz, Cys-β), 3.32 (dd, 1H, J ) 4.9 and 13.2 Hz, Cys-β′), 3.56 (t, 1H, J ) 6.4 Hz, Glu-R), 3.74 (d, 2H, J ) 5.9 Hz, Gly-R,R′), 4.54 (m, 1H, Cys-R), 5.43 (m, 1H, CH2FO), 5.59 (m, 2H, CH2FO and CH2FOCH(CF3)CF2), 8.54 (d, 1H, J ) 8.4 Hz, CONH), and 8.64 (t, 1H, J ) 5.8 Hz, CONH). 19F NMR: δ -153.56 (t, 1F, J ) 53.6 Hz, CH2F), -87.47 (dqn, 1F, J ) 223.3 and 10.5 Hz, CF2), -84.45 (dqn, 1F, J ) 223.3 and 9.1 Hz, CF2), and -76.43 (m, 3F, CF3). MS/MS (CID of MH+ at m/z 488): m/z 448 ([M + H - 2HF]+), 471 ([M + H - NH3]+), 413 ([M + H - Gly]+), 373 ([M + H - Gly - 2HF]+), 359 ([M + H - pyroglutamic acid]+), 319 ([M + H - pyroglutamic acid - 2HF]+), 299 ([CH2FOCH(CF3)CF2SCH2C(NH2)CONH2]+), 256 ([CH2FOCH(CF3)CF2SCH2CHdNH2]+), 216 ([CH2FOCH(CF3)CF2SCH2CHdNH2 - 2HF]+), and 130 ([pyroglutamic acid + H]+). Conjugate A3: (Z)-[1-Fluoro-2-(fluoromethoxy)-2-(trifluoromethyl)vinyl]glutathione. 1H NMR: δ 1.91 (m, 2H, Glu-β,β′), 2.33 (m, 2H, Glu-γ,γ′), 3.13 (m, 1H, Cys-β), 3.36 (m, 2H, Glu-R and Cys-β′), 3.56 (m, 2H, Gly-R,R′), 4.45 (m, 1H, CysR), 5.64 (d, 2H, J ) 52.7 Hz, CH2FO), 8.36 (s, 1H, CONH), and 8.90 (s, 1H, CONH). 19F NMR: see Table 1. MS/MS (CID of MH+ at m/z 468): m/z 451 ([M + H - NH3]+), 393 ([M + H Gly]+), 339 ([M + H - pyroglutamic acid]+), 236 ([CH2FOC(CF3)dCFSCH2CHdNH2]+), 145 ([pyroglutamic acid + NH3]+), 130 ([pyroglutamic acid + H]+), and 76 ([Gly + H]+). Conjugate A4: (E)-[1-Fluoro-2-(fluoromethoxy)-2-(trifluoromethyl)vinyl]glutathione. 1H NMR: δ 1.83 (m, 1H, Glu-β), 1.92 (m, 1H, Glu-β′), 2.30 (m, 2H, Glu-γ,γ′), 3.10 (dd, 1H, J ) 9.8 and 13.7 Hz, Cys-β), 3.25 (m, 1H, Glu-R), 3.40 (dd, 1H, J ) 4.3 and 13.7 Hz, Cys-β′), 3.63 (d, 2H, J ) 4.6 Hz, GlyR,R′), 4.49 (m, 1H, Cys-R), 5.53 (dd, 1H, JH,F ) 52.8 Hz and JH,H ) 2.7 Hz, CH2FO), 5.58 (dd, 1H, JH,F ) 52.4 Hz and JH,H ) 2.7

Chem. Res. Toxicol., Vol. 9, No. 2, 1996 557 Hz, CH2FO), 8.66 (s, 1H, CONH), and 8.71 (s, 1H, CONH). 19F NMR: see Table 1. MS/MS (CID of MH+ at m/z 468): m/z 451 ([M - NH3]+), 393 ([M + H - Gly]+), 339 ([M + H pyroglutamic acid]+), 236 ([CH2FOC(CF3)dCFSCH2CHdNH2]+), 145 ([pyroglutamic acid + NH3]+), 130 ([pyroglutamic acid + H]+), and 76 ([Gly + H]+). In Vivo Biological Studies. Male Sprague-Dawley rats (250-280 g), obtained from Taconic Farm (Germantown, NY), were anesthetized with an ip injection of ketamine (68.2 mg kg-1) and xylazine (4.4 mg kg-1) simultaneously. A bile duct cannula made of PE-10 tubing (1 cm) connected to silastic tubing (50 cm) was inserted into the common bile duct between the liver and the duodenum of each rat. The cannula was introduced under the skin with an eye probe and externalized on the upper back between the shoulder blades where the cannula was secured to a stainless steel button tether assembly (Instech Laboratories, PA). When the animal recovered from anesthesia, Compound A (36 mg kg-1; 0.2 mmol kg-1) dissolved in corn oil (0.6 mL) was administered intraperitoneally, and bile was collected over ascorbic acid for 6 h post-dose. Specimens of bile were filtered and analyzed without further treatment by LCMS/MS, as described above. In Vitro Biological Studies. Rat liver microsomal and cytosolic fractions were prepared as described previously (31). The cytosol was dialyzed against 100 mM potassium phosphate buffer (pH 7.4) containing 50 mM potassium chloride for 24 h at 4 °C. Protein concentrations were determined by the method of Lowry et al. (32), using bovine serum albumin as standard. Incubations were carried out in 4 mL screw-capped glass vials equipped with silicon rubber septa. The incubation mixtures contained GSH (10 mM), EDTA (1 mM), microsomal or cytosolic protein (0.5 mg mL-1), and potassium phosphate buffer (100 mM, pH 7.4) in a final volume of 3 mL. After 3 min preincubation at 37 °C, the reaction was initiated by adding Compound A (3 µL). At appropriate time points, aliquots (200 µL) of the reaction mixture were removed with a syringe to a centrifuge tube containing 50% phosphoric acid (10 µL) to terminate metabolic activity. Each of the samples was treated with a stock solution of internal standard (SDETG, 1 mg mL-1, 20 µL) and extracted with ethyl acetate (0.6 mL) to remove residual Compound A. These mixtures then were centrifuged to sediment the protein, and the resulting clear aqueous phase (20 µL) was analyzed by HPLC to determine the amount of conjugates present. Samples for the construction of calibration curves were prepared by adding known amounts of analytes to tubes containing 50% phosphoric acid (10 µL) and SDETG (1 mg mL-1, 20 µL) and processed as outlined above.

Results Identification of Glutathione Conjugates in Bile. Following administration of Compound A to rats by intraperitoneal injection, specimens of bile were filtered and analyzed directly by LC-MS/MS. Constant neutral loss scanning was employed to detect candidate GSH conjugates in bile, based on the characteristic loss from the parent (MH+) ion of 129 Da [elimination of the elements of pyroglutamic acid (30)]. Under the LC conditions optimized for the separation of Compound A-GSH conjugates, a total of four GSH adducts (denoted as “A1”-“A4” in Figure 1) were detected. Inspection of the data from this neutral loss scanning experiment revealed that metabolites A1 and A2 both exhibited an MH+ ion at m/z 488, while A3 and A4 both gave an MH+ ion at m/z 468, suggesting that two pairs of isomeric GSH conjugates were excreted in bile. Based on the above mass spectrometric data, the molecular weights of metabolites A1 and A2 corresponded to the products of addition of GSH across the double bond of Compound A. [The mixture of A1 and A2 was referred to previously as “G1” (24).] The product ion spectra of

558 Chem. Res. Toxicol., Vol. 9, No. 2, 1996

Jin et al.

Figure 1. Detection of GSH conjugates in the bile of a rat which had been treated with Compound A (36 mg kg-1, ip). The upper chromatogram was obtained from constant neutral loss scanning LC-MS/MS analysis of a bile specimen collected between 0 and 6 h post-dose and depicts all constituents of the sample which eliminated 129 Da upon CID. The center and lower chromatograms depict the detection of those metabolites whose parent [M + H]+ ions appear at m/z 488 and 468, respectively, and which eliminate 129 Da upon CID.

Figure 3. Spectra of product ions obtained by CID of the [M + H]+ ions (m/z 468) of (A) metabolite A3 and (B) metabolite A4. The origins of characteristic product ions are as shown. “GS” denotes the glutathion-S-yl moiety.

Figure 2. Spectra of product ions obtained by CID of the [M + H]+ ions (m/z 488) of (A) metabolite A1 and (B) metabolite A2. The origins of characteristic product ions are as shown. “GS” denotes the glutathion-S-yl moiety.

A1 and A2 (Figure 2), obtained from on-line LC-MS/MS analysis of bile, were similar to each other and to that of G1 (24). In principle, GSH would be expected to attack Compound A preferentially at the terminal carbon, which is more electron-deficient and less sterically hindered than the internal position, to form [1,1-difluoro-2-(fluoromethoxy)-2-(trifluoromethyl)ethyl]glutathione. Since this addition of GSH to Compound A leads to the

formation of a new chiral center in the molecule, it seemed likely that A1 and A2 represented the two diastereomers of this structure. This proposal was shown to be correct when authentic samples of the diastereomers of [1,1-difluoro-2-(fluoromethoxy)-2-(trifluoromethyl)ethyl]glutathione were prepared by synthesis and found to exhibit HPLC and MS/MS characteristics identical to those of the biological conjugates. NMR analyses of A1 and A2 indicated clearly that the addition of GSH to Compound A was regiospecific, resulting in the formation of the diastereomers of the conjugate in which GSH was linked exclusively via the terminal carbon of Compound A. The absolute configuration of the new chiral center in A1 and A2 remains to be defined. Metabolites A3 and A4 each exhibited an MH+ ion at m/z 468, which was consistent with the molecular weight of [1-fluoro-2-(fluoromethoxy)-2-(trifluoromethyl)vinyl]glutathione, a dehydrofluorinated analog of conjugates A1 and A2. [The mixture of A3 and A4 was referred to previously as “G2” (24).] The product ion spectra of A3 and A4 (Figure 3), which were similar to each other, closely resembled that of G2 (24). Since [1-fluoro-2(fluoromethoxy)-2-(trifluoromethyl)vinyl]glutathione contains a single CdC double bond, more than one geometric isomer could be formed biologically and it was concluded, therefore, that metabolites A3 and A4 most likely corresponded to the (E)- and (Z)-isomers of this olefin. This hypothesis was shown to be correct when the corresponding synthetic materials were found to possess LC-MS/ MS properties identical to those of metabolites A3 and A4. The stereochemistry of the individual isomers was assigned on the basis of their 19F NMR properties,

Bioactivation of Sevoflurane Compound A Table 1.

19F

Chem. Res. Toxicol., Vol. 9, No. 2, 1996 559

NMR Data (at 375 MHz) for Conjugates A3 and A4a

a Footnotes: †Samples were dissolved in Me SO-d and chemical 2 6 shifts are expressed in parts per million (δ) relative to internal trifluoroacetic acid (δ -78.5 ppm). ‡Signal multiplicities: q ) quartet, m ) mutiplet, dd ) doublet of doublets, tq ) triplet of quartets, tm ) triplet of mutiplets.

especially the 4-bond F-F couplings (4JFF) between the vinyl fluorine and the trifluoromethyl group. It has been well documented that a cis disposition of the two groups results in a larger coupling constant than for a trans relationship (33). Therefore, conjugate A4, which exhibited a larger coupling constant (4JFaFb ) 21.8 Hz) (Table 1), was identified as (Z)-[1-fluoro-2-(fluoromethoxy)-2(trifluoromethyl)vinyl]glutathione, in which the vinyl fluorine is cis to the CF3 group. Metabolite A3, with a smaller coupling constant (4JFaFb ) 11.1 Hz) (Table 1), was identified as (E)-[1-fluoro-2-(fluoromethoxy)-2-(trifluoromethyl)vinyl]glutathione, in which the vinyl fluorine is trans to the CF3 moiety. Quantitative Analysis of Compound A-GSH Conjugates in Rat Bile. In order to assess the quantitative significance of GSH-dependent metabolism of Compound A in vivo, the fraction of a dose of Compound A excreted as GSH conjugates was determined in bile collected over 6 h. Since isomeric conjugates A1 and A2, as well as A3 and A4, exhibited very similar mass spectrometric properties, a chromatographic separation of these metabolites became essential so that each conjugate could be quantified individually. An LC-MS/MS assay was developed in which metabolites were separated on a narrow-bore reverse phase HPLC column and quantified by SRM procedures. By this approach, it was found that a total of 29.2 ( 2.7% (mean ( SD, N ) 4) of the dose of Compound A (36 mg kg-1; 0.2 mmol kg-1) was excreted into the bile as GSH conjugates over 6 h. Conjugates A2 and A4 were the two major biliary metabolites, accounting for 12.6 ( 2.5% and 11.7 ( 1.1%, respectively, of the administered dose. Metabolites A1 and A3, on the other hand, accounted for 1.6 ( 0.3% and 3.3 ( 0.2%, respectively, of the administered dose. Conjugation of Compound A with GSH in Vitro. Incubation of Compound A and GSH with rat liver microsomes, boiled rat liver microsomes and rat liver cytosol resulted in time-dependent formation of GSH adducts A1-A4 (Figure 4). Incubations conducted either in the presence of boiled rat liver cytosol or in the absence of cellular subfractions exhibited similar time course profiles as that obtained from boiled rat liver microsomes (data not shown). The two saturated GSH adducts (A1 and A2), which were formed in similar amounts, were the major products of such nonenzymatic conjugation. In the case of two minor products, the formation of the (Z)isomer A4 was favored over that of the (E)-isomer A3 by a factor of 4:1.

Figure 4. Time course of formation of Compound A GSH conjugates in the presence of (A) boiled rat liver microsomes, (B) rat liver cytosol, and (C) untreated rat liver microsomes. The incubation mixtures contained Compound A (3 mL; d ) 1.48 g mL-1), GSH (10 mM), EDTA (1 mM), protein (0.5 mg mL-1), and potassium phosphate buffer (100 mM, pH 7.4) in a final volume of 3 mL. Data represent means ( SD (N ) 3).

Both rat liver cytosolic and microsomal glutathione S-transferases (GST) were shown to catalyze the addition of GSH to Compound A. However, these two enzyme systems differed markedly with respect to the composition of the resulting conjugates. With cytosolic GSTs, the relative amounts of conjugates A1-A4 were 31%, 46%, 8%, and 15%, respectively. With microsomal GST, the relative amounts of conjugates A1-A4 were 15%, 37%, 7%, and 41%, respectively.

Discussion The present investigation, which focused on the GSHdependent metabolism of sevoflurane Compound A in

560 Chem. Res. Toxicol., Vol. 9, No. 2, 1996

Figure 5. Proposed scheme for the formation of GSH conjugates A1-A4 from Compound A. See text for details.

rats, led to the identification of four GSH conjugates (A1A4) which were excreted in bile. By means of ionspray LC-MS/MS and NMR techniques, these four adducts are identified as two pairs of stereoisomers. The first pair of metabolites, A1 and A2, were characterized as the two diastereomers of [1,1-difluoro-2-(fluoromethoxy)-2-(trifluoromethyl)ethyl]glutathione, corresponding to the products of addition of GSH across the activated double bond of Compound A. The second pair, A3 and A4, proved to be the two geometric isomers of [1-fluoro-2(fluoromethoxy)-2-(trifluoromethyl)vinyl]glutathione, corresponding to dehydrofluorinated analogs of A1 and A2. The formation of both S-(haloalkyl)- and S-(haloalkenyl)-GSH adducts is believed to reflect attack on the R-carbon of 1,1-dihaloalkenes by the thiol group of GSH to generate an intermediate β-carbanion (34-36). In the case of Compound A, this carbanion reacts either by abstracting a proton from the medium to form the saturated metabolites A1 and A2 (pathway a) or by eliminating a fluoride ion from the R-carbon to yield the unsaturated adducts A3 and A4 (pathway b) (Figure 5). It should be noted that the proton attached to the chiral center in the Compound A-derived portion of metabolite A1 and A2 is expected to be acidic, due to the presence of adjacent electron-withdrawing groups. As a result, conjugates A1 and A2 theoretically may undergo basecatalyzed dehydrofluorination to yield the unsaturated derivatives A3 and A4. This possibility was ruled out, however, based on the observation that incubation of A1 and A2 at 37 °C in aqueous buffers (pH 7.4-11) resulted in less than 0.1% of the substrates being converted to A3 and A4 over 24 h (data not shown). It follows, therefore, that metabolites A3 and A4 most likely were formed directly during conjugation of Compound A with GSH, as outlined in Figure 5, and were not secondary metabolites resulting from the dehydrofluorination of A1 and A2. This conclusion also was supported by the time course of formation of these conjugates in vitro, which demonstrated that metabolites A3 and A4 were formed simultaneously with A1 and A2 in incubation mixtures (Figure 4). Conjugation of Compound A with GSH in vitro also proceeded nonenzymatically, in a similar fashion to that reported for the highly reactive dichloroethyne (37, 38). The two major products of the chemical reaction with Compound A, namely, diastereomers A1 and A2, were formed in approximately equal amounts, consistent with non-stereoselective addition of GSH to the double bond.

Jin et al.

Of the two less abundant conjugates, the (Z)-isomer A4 was favored over the (E)-isomer A3, most likely due to the fact that A4 is thermodynamically more stable since the glutathionyl moiety is bound trans to the trifluoromethyl group which is bulkier than the oxygen in the fluoromethoxy residue. Compound A also underwent conjugation with GSH in a GST-mediated reaction, as do other haloalkenes (25-27). Both rat liver cytosolic and microsomal GSTs catalyzed the conjugation of Compound A with GSH, but these enzymes exhibited clear differences with respect to the distribution of the resulting conjugates. Similar differences in product stereoselectivity for cytosolic and microsomal GSTs have been reported in the formation S-(2-chloro-1,1,2-trifluoroethyl)glutathione from chlorotrifluoroethene (39, 40) and have been interpreted to indicate that microsomal and cytosolic GSTs exhibit either different active-site topographies or different chemical mechanisms, or both (40). The results of the present study with Compound A support this view. A comparison of the profile of biliary Compound A-GSH conjugates with those obtained from incubations with liver cytosolic and microsomal fractions suggests that the conjugation of Compound A with GSH in rats in vivo may be catalyzed preferentially by microsomal GST with a modest contribution from cytosolic enzymes. However, A3 was found to be present in the bile in higher concentrations than A1, despite the fact that it was the least abundant product among the four conjugates in all in vitro incubations. This indicates that other factors, such as stereoselective transport and further metabolism of the conjugates in liver, also may affect the biliary metabolite profile. Additional studies will be required to determine the relative contributions of the cytosolic and microsomal GSTs to the conjugation of Compound A with GSH in vivo. Quantitative studies indicated that a total of 29% of the administered dose of Compound A (36 mg kg-1) was excreted into the bile over 6 h in the form of GSH adducts A1-A4, indicating that conjugation with GSH plays an important role in the disposition of Compound A in rats. As a result, GSH-dependent bioactivation of Compound A may be a significant factor mediating the renal toxicity in rats. In addition, fluoride ion, another known renal toxin, is liberated during the metabolism of Compound A to the vinyl GSH conjugates A3 and A4. It has been demonstrated that inorganic fluoride, released during the metabolism of certain fluorinated anesthetics, may cause renal toxicity at high concentrations (41). In the case of Compound A, approximately 14% of the administered dose was excreted into bile over 6 h as conjugates A3 and A4, raising the possibility that high circulating fluoride ion levels (not measured in this study) also may contribute to the renal toxicity of Compound A. Although Compound A-induced nephrotoxicity has been documented in rats, the phenomenon has not been observed in clinical studies with sevoflurane. Therefore, an understanding of the mechanism by which Compound A elicits renal toxicity in rats may serve as a basis for further studies on the potential for Compound A-mediated toxicity in humans. For example, it will be important to document whether Compound A follows a similar metabolic pathway in humans to that in rats and, if so, to determine the relative importance of the GSH-dependent versus alternative routes of biotransformation for this drug. Studies currently are underway to address some of these issues.

Bioactivation of Sevoflurane Compound A

Acknowledgment. These studies were supported in part by research grants from the National Institutes of Health (GM48712 and ES05500), a grant from Abbott Laboratories, and a Faculty Development Award from the Pharmaceutical Research and Manufacturers of America Foundation (to E.D.K.), which are gratefully acknowledged.

References (1) Wallin, R. F., Regan, B. M., Napoli, M. D., and Stern, I. J. (1975) Sevoflurane: A new inhalational anesthetic agent. Anesth. Analg. 54, 758-766. (2) Holaday, D. A., and Smith, F. R. (1981) Clinical characteristics and biotransformation of sevoflurane in healthy human volunteers. Anesthesiology 54, 100-106. (3) Hanaki, C., Fujii, K., Morio, M., and Tashima, T. (1987) Decomposition of sevoflurane by soda lime. Hiroshima J. Med. Sci. 36, 61-67. (4) Frink, E. J., Jr., Malan, T. P., Morgan, S. E., Brown, E. A., Malcomson, M., and Brown, B. R., Jr. (1992) Quantification of the degradation products of sevoflurane in two CO2 absorbants during low-flow anesthesia in surgical patients. Anesthesiology 77, 1064-1069. (5) Morio, M., Fujii, K., Satoh, N., Imai, M., Kawakami, U., Mizuno, T., Kawai, Y., Ogasawara, Y., Tamura, T., Negishi, A., Kumagai, Y., and Kawai, T. (1992) Reaction of sevoflurane and its degradation products with soda lime. Toxicity of the byproducts. Anesthesiology 77, 1155-1164. (6) Gonsowski, C. T., Laster, M. J., Eger, E. I., II, Ferrell, L. D., and Kerschmann, R. L. (1994) Toxicity of Compound A in rats. Effect of a 3-hour administration. Anesthesiology 80, 556-565. (7) Gonsowski, C. T., Laster, M. J., Eger, E. I., II, Ferrell, L. D., and Kerschmann, R. L. (1994) Toxicity of Compound A in rats. Effect of increasing duration of administration. Anesthesiology 80, 566573. (8) Mazze, R. I. (1992) The safety of sevoflurane in humans. Anesthesiology 77, 1062-1063. (9) Huang, C., Venturella, V. S., Cholli, A. L., Venutolo, F. M., Silbermann, A. T., and Vernice, G. G. (1989) Detailed investigation of fluoromethyl 1,1,1,3,3,3-hexafluoro-2-propyl ether (sevoflurane) and its degradation products. Part I: Synthesis of fluorinated, soda lime induced degradation products. J. Fluorine Chem. 45, 239-253. (10) Cholli, A. L., Huang, C., Venturella, V., Pennino, D. J., and Vennice, G. G. (1989) Detailed investigation of fluoromethyl 1,1,1,3,3,3-hexafluoro-2-propyl ether (sevoflurane) and its degradation products. Part II: Two-dimensional fluorine-19 NMR characterization of fluoromethyl 1,1,3,3,3-pentafluoro-2-propenyl ether. Appl. Spectrosc. 43, 24-27. (11) Wong, D. T., and Lerman, J. (1992) Factors affecting the rate of disappearance of sevoflurane in Baralyme. Can. J. Anaesth. 39, 366-369. (12) Strum, D. P., Johnson, B. H., and Eger, E. I. II. (1987) Stability of sevoflurane in soda lime. Anesthesiology 67, 779-781. (13) Ruzicka, J. A., Higalgo, J. C., Tinker, J. H., and Baker, T. M. (1994) Inhibition of volatile sevoflurane degradation product formation in an anesthesia circuit by a reduction in soda lime temperature. Anesthesiology 81, 238-244. (14) Bito, H., and Ikeda, K. (1994) Closed-circuit anesthesia with sevoflurane in humans. Effects on renal and hepatic function and concentrations of breakdown products with soda lime in the circuit. Anesthesiology 80, 71-76. (15) Bito, H., and Ikeda, K. (1994) Long-duration, low-flow sevoflurane anesthesia using two carbon dioxide absorbents. Quantification of degradation products in the circuit. Anesthesiology 81, 340345. (16) Frink, E. J., Jr., Isner, R. J., Malan, T. P., Jr., Morgan, S. E., Brown, E. A., and Brown, B. R., Jr. (1994) Sevoflurane degradation product concentrations with soda lime during prolonged anesthesia. J. Clin. Anesth. 6, 239-242. (17) Liu, J., Laster, M. J., Eger, E. I., II, and Taheri, S. (1991) Absorption and degradation of sevoflurane and isoflurane in a conventional anesthetic circuit. Anesth. Analg. 72, 785-789. (18) Odum, J., and Green, T. (1984) The metabolism and nephrotoxicity of tetrafluoroethylene in the rat. Toxicol. Appl. Pharmacol. 76, 306-318. (19) Green, T., and Odum, J. (1985) Structure/activity studies of the nephrotoxic and mutagenic action of cysteine conjugates of chloroand fluoroalkenes. Chem.-Biol. Interact. 54, 15-31. (20) Dohn, D. R., and Anders, M. W. (1982) The enzymatic reaction of chlorotrifluoroethylene with glutathione. Biochem. Biophys. Res. Commun. 109, 1339-1345.

Chem. Res. Toxicol., Vol. 9, No. 2, 1996 561 (21) Hassall, C. D., Gandolfi, A. J., Duhamel, R. C., and Brendel, K. (1984) The formation and biotransformation of cysteine conjugates of halogenated ethylenes by rabbit renal tubules. Chem.-Biol. Interact. 49, 283-297. (22) Dohn, D. R., Leininger, J. R., Lash, L. H., Quebbemann, A. J., and Anders, M. W. (1985) Nephrotoxicity of S-(2-chloro-1,1,2trifluoroethyl)glutathione and S-(2-chloro-1,1,2-trifluoroethyl)-Lcysteine, the glutathione and cysteine conjugates of chlorotrifluoroethene. J. Pharmacol. Exp. Ther. 235, 851-857. (23) Dekant, W., Lash, L. H., and Anders, M. W. (1987) Bioactivation mechanism of the cytotoxic and nephrotoxic S-conjugate S-(2chloro-1,1,2-trifluoroethyl)-L-cysteine. Proc. Natl. Acad. Sci. U.S.A. 84, 7443-7447. (24) Jin, L., Baillie, T. A., Davis, M. R., and Kharasch, E. D. (1995) Nephrotoxicity of sevoflurane Compound A [fluoromethyl-2,2difluoro-1-(trifluoromethyl)vinyl ether] in rats: Evidence for glutathione and cysteine conjugate formation and the role of renal cysteine conjugate β-lyase. Biochem. Biophys. Res. Commun. 210, 498-506. (25) Commandeur, J. N. M., Stijntjes, G. J., and Vermeulen, N. P. E. (1995) Enzymes and transport systems involved in the formation and disposition of glutathione S-conjugates. Role in bioactivation and detoxication mechanisms of xenobiotics. Pharmacol. Rev. 47, 271-330. (26) Dekant, W., Vamvakas, S., and Anders, M. W. (1989) Bioactivation of nephrotoxic haloalkenes by glutathione conjugation: Formation of toxic and mutagenic intermediates by cysteine conjugate β-lyase. Drug Metab. Rev. 20, 43-83. (27) Dekant, W., Vamvakas, S., and Anders., M. W. (1994) Formation and fate of nephrotoxic and cytotoxic glutathione S-conjugates: cysteine β-lyase pathway. In Advances in Pharmacology (Anders, M. W., and Dekant, W., Eds.) Vol. 27, pp 115-162, Academic, San Diego, CA. (28) Elfarra, A. A., Jakobson, I., and Anders, M. W. (1986) Mechanism of S-(1,2-dichlorovinyl)glutathione-induced nephrotoxicity. Biochem. Pharmacol. 35, 283-288. (29) Jin, L., Davis, M. R., Hu, P., and Baillie, T. A. (1994) Identification of novel glutathione conjugates of disulfiram and diethyldithiocarbamate in rat bile by liquid chromatography-tandem mass spectrometry. Evidence for metabolic activation of disulfiram in vivo. Chem. Res. Toxicol. 7, 526-533. (30) Baillie, T. A., and Davis, M. R. (1993) Mass spectrometry in the analysis of glutathione conjugates. Biol. Mass Spectrom. 22, 319325 (31) Rettie, A. E., Boberg, M., Rettenmeier, A. W., and Baillie, T. A. (1988) Cytochrome P-450-catalyzed desaturation of valproic acid in vitro. Species differences, induction effects, and mechanistic studies. J. Biol. Chem. 263, 13733-13738. (32) Lowry, O. H., Rosebrough, N. J., Lewis-Farr, A., and Randall, R. J. (1951) Protein measurement with Folin phenol reagent. J. Biol. Chem. 193, 265-275. (33) Emsley, J. W., Phillips, L., and Wary, V. (1977) In Fluorine Coupling Constants (Emsley, J. W., Feeney, J., and Sutcliffe, L. H., Eds.) pp 399-400, Pergamon, New York. (34) Modena, G. (1971) Reactions of nucleophiles with ethylenic substrates. Acc. Chem. Res. 4, 73. (35) Rappoport, Z. (1985) The rich mechanistic world of nucleophilic vinylic (SNV) substitution. Recl. Trav. Chim. Pays-Bas 104, 309. (36) Bernasconi, C. F. (1989) Nucleophilic addition to olefins. Kinetics and mechanism. Tetrahedron 45, 4017-4090. (37) Kanhai, W., Dekant, W., and Henschler, D. (1989) Metabolism of the nephrotoxin dichloroacetylene by glutathione conjugation. Chem. Res. Toxicol. 2, 51-56. (38) Patel, N., Birner, G., Dekant, W., and Anders, M. W. (1995) Glutathione-dependent biosynthesis and bioactivation of S-(1,2dichlorovinyl)glutathione and S-(1,2-dichlorovinyl)-L-cysteine, the glutathione and cysteine S-conjugates of dichloroacetylene, in rat tissues and subcellular fractions. Drug Metab. Dispos. 22, 143147. (39) Dohn, D. R., Quebbemann, A., Borch, R. F., and Anders, M. W. (1985) Enzymatic reaction of chlorotrifluoroethene with glutathione: 19F NMR evidence for stereochemical control of the reaction. Biochemistry 24, 5137-5143. (40) Hargus, S., Fitzsimmons, M. E., Aniya, Y., and Anders, M. W. (1991) Stereochemistry of the microsomal glutathione S-transferase catalyzed addition of glutathione to chlorotrifluoroethene. Biochemistry 30, 717-721. (41) Cousins, M. J., and Mazze, R. I. (1973) Methoxyflurane nephrotoxicity: A study of dose response in man. J. Am. Med. Assoc. 225, 1611-1616.

TX950162M