Chem. Res. Toxicol. 1993,6, 459-469
459
Identification of Multi-S-Substituted Conjugates of Hydroquinone by HPLC-Coulometric Electrode Array Analysis and Mass Spectroscopy Barbara A. Hill,tJ Heather E. Kleiner,? Elizabeth A. Ryan,%Deanne M. Dulik,ll Terrence J. Monks,? and Serrine S. Lau*9t Division of Pharmacology and Toxicology, College of Pharmacy, University of Texas at Austin, Austin, Texas 78712, ESA Inc., 45 Wiggins Avenue, Bedford, Massachusetts 01 730, and Drug Metabolism and Pharmacokinetics, SmithKline Beecham Pharmaceuticals, P.O. Box 1539, King of Prussia, Pennsylvania 19406-0939 Received November 17, 1992
Chemical reaction of 1,4-benzoquinone with GSH gives rise to several multisubstituted hydroquinone (HQ)-GSH conjugates, each of which causes renal proximal tubular necrosis when administered to male Sprague-Dawley rats. In addition, HQ has recently been reported to be nephrocarcinogenic following long-term exposure in male rats. Since neither the mechanism nor the extent of HQ oxidation and thioether formation in vivo is known, we have assessed both the qualitative and quantitative significance of HQ-thioether formation in vivo and in vitro. HQ (1.8mmol/kg, ip) was administered t o AT-125-pretreated male Sprague-Dawley rats, and bile and urine samples were analyzed with a HPLC-coulometric electrode array system (CEAS) and by liquid chromatography (LC)/continuous-flow fast atom bombardment (CF-FAB) mass spectroscopy. Five S-conjugates of hydroquinone were identified in bile, and one S-conjugate was identified in urine. The major biliary S-conjugate identified was 2-glutathion-Sylhydroquinone [2-(GSyl)HQl (18.9 f 2.7 pmol). Additional biliary metabolites were 2,5diglutathion-S-ylhydroquinone [2,5-(diGSyl)HQl (2.2 f 0.6 pmol), 2,6-diglutathionS-ylhydroquinone [2,6-(diGSyl)HQl(O.7f 0.3 pmol), 2,3,5-triglutathion-S-ylhydroquinone [2,3,5(triGSy1)HQI (1.2 f 0.1 pmol), and 2-(cystein-S-ylglycyl)hydroquinone.2-(N-AcetylcysteinS-y1)HQ was the only urinary thioether metabolite (11.4 f 3.6 pmol) identified. The quantity of S-conjugates excreted in urine and bile within 4 h of HQ administration [34.3 f 4.5 pmol (4.3 f 1.1% of dose)] appears sufficient to propose a role for such metabolites in HQ-mediated nephrotoxicity and nephrocarcinogenicity. Rat liver microsomes catalyzed the NADPHdependent oxidation of HQ (300 pM),in the presence of GSH, to form 2-(GSyl)HQ, 2,5-(diGSy1)HQ, and 2,6-(diGSy1)HQ. A fraction of the microsomal oxidation of HQ appears to be catalyzed by cytochrome(s) P450, although the exact amount remains unclear. 2-(GSyl)HQ, 2,5-(diGSy1)HQ, and 2,6-(diGSyl)HQ (300 pM) also underwent NADPH-dependent oxidation and GSH conjugation in liver microsomes. The extent of the nonenzymatic oxidation of HQ and its GSH conjugates correlated, approximately, with their half-wave oxidation potentials.
Introduction
studies addressing the metabolism of HQ in humans and experimental animals have shown that it is readily In 1984,annual US. production of hydroquinone (HQ)l absorbed from the gastrointestinal tract and is excreted was estimated to be 34 million pounds. HQ has several in urine primarily in the form of sulfate and glucuronide applications including its use as a developer in photogconjugates (2,3). Divincenzo et al. (3)reported that [%Iraphy, as an antioxidant in the rubber industry, and as an HQ administered either as a single dose or repeated doses intermediate in the manufacture of other antioxidants. to Sprague-Dawley rats was found mainly in the urine HQ, and products containing HQ, are used as depigmenting (91.9%) as free HQ (1.1-8.6’36 of the dose), HQ-monoagents for the treatment of a variety of skin disorders. HQ sulfate (25-42 % 1, and HQ-monoglucuronide (56-66 % ) has also been identified and quantified in the smoke of within 2-4 days; 3.8% was excreted in the feces, about nonfiltered cigarettes (up to 80 pg/cigarette) ( 11. Previous 0.4% was excreted in expired air, and 1.2% remained in the carcass. Radioactivity was widely distributed through* Author to whom all correspondence should be addressed. out the tissues with highest concentrations in the liver t University of Texas at Austin. and kidneys. t Present address: NIH-NCI,Building 37, Room 3C28, Bethesda, MD 20892. Nerland and Pierce ( 4 )subsequently demonstrated that I ESA Inc. 2-(N-acetylcystein-S-yl) hydroquinone was formed in vivo 1 SmithKline Beecham Pharmaceuticals. 1 Abbreviations: AT-125, acivicin,~-[a(S),S(S)]-a-amino-3-chloro-4,5when male Sprague-Dawley rats were treated with either isoxazoleaceticacid; LC/CF-FABMS, liquid chromatography/continuousbenzene (600 mg/kg body wt, ip), phenol (75 mg/kg, body flow-fast atom bombardment mars spectroscopy;EC, electrochemical; wt, ip), or HQ (75 mg/kg body wt, ip). Detection of 2-(NHEPES, 4-(2-hydroxyethyl)-l-piperazineethanesulfonicacid; HPLCCEAS, HPLC-coulometricelectrode array system;HQ, hydroquinone; acetylcystein-S-y1)HQ in the urine of rats treated with 2-(GSyl)HQ,2-gluta~on-S-ylhyeoquinone;2,5-(diGSyl)HQ,2,5-digluHQ implies that HQ undergoes oxidation and thioether tathon-S-ylhydroquinone;2,6-(diGSyl)HQ,2,6-diglutathion-S-ylhydroquinone; 2,3,5-(triGSyl)HQ,2,3,5-triglutathion-S-ylhydroquinone. formation in vivo, although direct evidence for GSH 0893-228~/93/2706-0459$04.00/0 0 1993 American Chemical Society
460 Chem. Res. Toxicol., Vol. 6, No. 4, 1993
conjugation of HQ in vivo remains to be demonstrated. In addition, the mechanism by which HQ is oxidized in biological systems is obscure. Sawahata and Neal (5) speculated that the final step@)in the conversion of HQ to reactive metabolites responsible for its covalent binding to microsomal protein was (were) mediated by an enzyme other than cytochrome(s) P450. Support for this suggestion came from the finding that the molecular mass of the protein (72 kDa) selectively labeled by metabolites of phenol differed substantially from the molecular mass of the various cytochrome P450 enzymes (48-54 kDa) (6). Moreover, these workers also provided convincing evidence that neither air oxidation of HQ nor its oxidation by superoxide anion was involved in the generation of reactive HQ metabolites. Additional evidence for the involvement of a NADPH-dependent, cytochrome P450-independent mechanism of HQ oxidation came from studies reported by Lunte and Kissinger (7). These workers monitored HQ oxidation in mouse liver microsomes by measuring the formation of the GSH conjugate of 1,4-benzoquinone, 2-glutathion-S-ylhydroquinone[2-(GSyl)HQl. The formation of 24GSyl)HQ was dependent upon the presence of both liver microsomes and NADPH. Omission of NADPH caused a 90 % reduction in 2-(GSyl)HQ formation whereas the addition of metyrapone, an inhibitor of cytochrome(s) P450, had no effect on 2-(GSyl)HQ production. Consistent with the data of Sawahata and Neal (5), superoxide dismutase had no effect on 24GSyl)HQ formation from HQ. Although cytosolic peroxidases could also catalyze the conversion of HQ to 2-(GSyl)HQ, this activity represented less than 5 % of the microsomal activity (7). In the present paper, we have investigated rat liver microsomal HQ oxidation in detail, and the data suggest that a significant portion of this reaction is catalyzed by the cytochrome(s) P450. Recently, the National Toxicology Program reported that HQ (1.82 mmol/kg; 13-week gavage studies) caused tubular cell degeneration in the renal cortex of male and female Fischer 344 rats and long-term (2-year) exposure to HQ (1.82 mmol/kg, gavage) caused increases in tubular cell adenomas of the kidney in male Fischer 344 rats (8). Shibata et al. (9) also demonstrated that HQ (0.8% in the diet) induced renal tubular hyperplasia as well as adenomas, and chronic nephropathy in male Fischer 344 rats. Although the mechanism of HQ-induced nephrotoxicity and nephrocarcinogenicity remains unclear, a role for the involvement of GSH conjugation can be proposed. Chemical oxidation of HQ in the presence of GSH results in the formation of several GSH conjugates that exhibit increasing degrees of GSH substitution (IO). Administration of these conjugates to male Sprague-Dawley rats caused varying degrees of renal proximal tubular necrosis. In particular, 2,3,5-triglutathion-S-ylhydroquinone [2,3,5(triGSy1)-HQI was a potent nephrotoxicant, causing elevations in blood urea nitrogen, enzymuria, and glucosuria at a dose of 20 rmollkg (iv) (10). The specific aims of the present work were therefore to determine the extent to which cytochrome(s) P450 contribute to HQ and HQ-thioether oxidation in vitro; whether multisubstituted GSH conjugates of HQ are formed in vivo; and if so, whether the quantity of S-conjugates formed are sufficient to support a role for HQ-thioethers in HQmediated nephrotoxicity and nephrocarcinogenicity. We show for the first time that multi-S-substituted conjugates of HQ are formed as in vivo metabolites of HQ.
Hill et al.
Materials and Methods Caution: Hydroquinone is a potential carcinogen and must be handled accordingly. In addition, the synthetic thioethers are potent nephrotoxicants in the rat. All these compounds should therefore be handled with protective clothing in a wellventilated fume hood. Chemicals. HQ was obtained from Fluka, A.G. (Buchs,S.G.,
Switerland). GSH, magnesium chloride,sodium azide,rotenone, antimycinA, 8-naphthoflavone,metyrapone,catalase,superoxide dismutase, and arylsulfatase were purchased from Sigma Chemical Co. (St. Louis, MO). Potassium cyanidewas purchased from JT Baker Chemicals (Phillipsburg, NJ) and l-benzylimidazole from Aldrich Chemical Co. (Milwaukee, WI). Sodium phenobarbital was a product of Mallinckrodt (Paris, KY). NADPH, NADH, glucose 6-phosphate, glucose-6-phosphate dehydrogenase, and 8-glucuronidase/arylsulfatase were obtained from Boehringer-Mannheim (Indianapolis, IN). AT-125 (acivicin; NSC 163501) was generously provided by Dr. Thomas Petry of the Upjohn Co. 2-(GSyl)-HQ,2,5-(diGSyl)HQ,2,6-(diGSyl)HQ 2,3,5-(triGSyl)HQ, and 2-(N-acetylcystein-S-yl)hydroquinone were synthesizedaccording to previously published methodology (IO). Aminobenzotriazole and SKF-525A were kind gifts from Dr. Alan Buckpitt (UC Davis) and Dr. Dan Ziegler (UT Austin), respectively. All other reagents were of the highest grade commercially available. Animals. Male Sprague-Dawley rats (350-400 or 160-250 g) obtained from Harlan Sprague-Dawley (Houston,TX) were used for the in vivo and in vitro studies, respectively,and were allowed food and water ad libitum prior to and during experiments. SurgicalTechnique. Rats were anesthetized with equithesin (a mixture of 35 mg/kg pentobarbital and 140 mg/kg chloral hydrate, ip) during the surgical procedure and maintained under anesthesia with equithesin (20 mg/kg sodium pentobarbital and 80 mg/kg chlorate hydrate, im) administered every hour after the surgery. The left jugular vein was cannulated with PE-10 tubing, and mannitol (7%) was perfused through the jugular vein (0.1 mL/min), by means of an adjustable volume syringe pump (Model BSP, Braintree Scientific, Inc., Braintree, MA), throughout the experiment. The abdomen was then sterilized and an incision made, and the right and left ureters and bile duct were cannulatedwithPE-10tubing. The free endsof the cannulae were brought outside the body through two stainless steel hypodermic needles (14 gauge) inserted at the right and left of the anus, respectively. The 14-gauge steel tubing was then removed. After the surgery was complete, rats were pretreated with AT-125 (lOmg/kg,ip) 1hpriortoadministrationofHQ (1.8 mmol/kg, ip). Urine and bile samples were then collected at hourly intervals, on ice, for a period of 4 h. The urine and bile sampleswere analyzed by HPLC with both UV and EC detection. The in vivo metabolism of HQ was assessed by HPLC (Shimadzu LC-6A) coupled to UV (280 nm) (Shimadzu SPD6AV) and EC detection (ESACoulochem, Model 5100A)(detector 1 at -0.20 V and detector 2 at +0.20 V). The EC detector was equipped with two porous graphite test electrodes connected in series with a glassy carbon reference electrode. Aliquots (10 ML of a 1:40 dilution of urine and bile samples with mobile phase) were injected onto a Phenomenex Partisil5-pm ODS-3 reversephase analyticalcolumn. Each sample was eluted with a mixture of 94% citric acid (4 mM)/ammonium acetate (8 mM) and 6% methanol (pH 4.0) containing 20 mg/L EDTA, at a flow rate of 1 mL/min. Under these conditions, authentic samples of 2-(GSyl)HQ,2,5-(diGSyl)HQ,2,6-(diGSyl)HQ,and 2,3,5-(triGSy1)HQ were eluted with retention times of 13.1,19.3,21.8, and 10.4 min, respectively. Thioethers of HQ were identified by comparison of HPLC retention times with authentic standards, by HPLC-CEAS (Coulochem electrode array system),and mass spectroscopic analysis (described below). HQ-thioethers were quantified by HPLC-EC using a ShimadzuC-R4A Chromatopac. Standard curves were prepared from each authentic standard, and quantitation was achieved by linear regression analysis (11). Identificationsof HQ-glucuronide and HQ-sulfatein urine and bile was accomplishedby analyzingsamples after @-glucuronidase/
Glutathione Conjugation of Hydroquinone
arylsulfataseand arylsulfatasehydrolysis. Aliquots (0.5 mL) were diluted 1:1with either of the following: (a) 0.1 M sodium acetate buffer (pH 5.0), followed by hydrolysis with 150 pL of @-glucuronidase, 4.5 IU/mL, and arylsulfatase, 14units/mL, at 37 "C for 16 h; or (b) 0.01 M Tris buffer (pH 7.0), followed by hydrolysis with 50 pL of arylsulfatase, 14.5 IU/mL, at 37 "C for 16h. After incubation, the mixture was extracted twice with 4 mL of ethyl acetate. The organic extracts were combined and evaporated to dryness under nitrogen and reconstituted with 1mL of methanol. Control incubations were performed by omitting either 8-glucuronidase/arylsulfatase or arylsulfatase from the incubation mixtures. Aliquots of both aqueous and ethyl acetate extracts (10 pL of a 1:20 dilution with mobile phase) were analyzed by HPLC-UV-EC as described above. Identification of Hydroquinone-Thioethers by HPLCCEAS. Authentic standards and urine and bile samples were analyzed by HPLC coupled to a Coulochem electrodearray system (CEAS), which permits peak identification both by retention time and by electrochemical oxidative profile (hydrodynamic voltammogram)(12,13). Ion-pair reverse-phase gradient elution chromatography was performed by mixing two mobile phases (A and B). Mobile phase "A" consisted of 0.1 M monobasic sodium phosphate, 10 mg/L sodium dodecyl sulfate, and nitrilotriacetic acid, adjusted to pH 3.35 with phosphoric acid. Mobile phase "B" consisted of 0.1M sodium phosphate, 50 mg/L sodium dodecyl sulfate, and nitrilotriacetic acid in 50% methanol (v/v) at pH 3.45. The analytical apparatus used was a CEAS (ESA, Inc.) coupled to an ESA 460 HPLC autosampler with a Neslab refrigerated circulating water bath to maintain samples at 4 "C during the automated analysis. Gradient elution (1 mL/min) was performed using two ESA 420 dual piston pumps and a Kontron M800 gradient mixer and consisted of a linear gradient from 6% to 40% B over 17 min (initiated 5 min after injection of sample),40% B for 1min, and a linear gradient from 40% to 100%B over 9 min. The electrode potentials were set at 50-mV intervals from -250 mV (channel 1) to +500 mV (channel 16). The detection system was composed of four cell packs, in series, each consisting of four porous graphite working electrodes with associated reference and counter electrodes. Samples were injected onto an ODS (3-pm)reverse-phase column. The column, electrodes, and pulse dampeners were housed in a temperatureregulated compartment (35 "C). All system components were controlled by the CEAS software on an Epson Equity 11+ microcomputer. Each standard was prepared at a concentration of 0.1-1.0 mg/mL in 0.1 M perchloric acid and 10% methanol (v/v) containing 10 pg/mL ascorbate. Each stock solution was sparged with nitrogen and stored at -80 OC. Secondary stock solutions were prepared by dilution of the above with nitrogensparged physiological saline (containing 100 ng/mL ascorbate) to a concentration of 10 pg/mL. Identification of Hydroquinone-Thioethers by LC/CFFAB. Confirmation of the HPLC-CEAS data was achieved by liquid chromatography (LC)/continuous-flow fast atom bombardment (CF-FAB)mass spectroscopy using a Finnigan TSQ70 triple-quadrupole mass spectrometer (San Jose, CA) with a Finnigan Bioprobe flow FAB interface. The LC conditions were as follows: Hewlett-Packard Model 1090 liquid chromatograph; column, Beckman Altex ODS (2.1 X 150 mm, 5 pm); mobile phase: solvent A, 0.1 M ammonium acetate (pH 5.0) containing 77% glycerol; solvent B, methanol containing 7% glycerol; flow rate 0.25 mL/min. Isocratic elution conditions were 5% solvent B. UV detection was done at 280 nm. The flow to the mass spectrometer was split approximately 42:l to provide a flow rate to the ion source of approximately 5-6 pL/min. FAB ionization conditions were as follows: 6-kV accelerating voltage, 1-mA current using xenon as the source of fast atoms. Electron multiplier voltagewas 1400eV. Full scans over the desired mass range were obtained within 2.0 s. Aliquota (1pL) of each synthetic standard (approximately 1mg/mL concentration) were injected directly into the ion source for detection by FAB ionization. Aliquots (50 pL) of urine or bile were injected onto the HPLC column for LC/MS analyses.
Chem. Res. Toxicol., Vol. 6, No. 4, 1993 461
I n Vitro Oxidation of Hydroquinone and its Glutathione Conjugates. The in vitro oxidation of HQ and its GSH conjugates was determined in rat liver microsomes. Rata were euthanized by cervical dislocation. The livers were removed and microsomes isolated as described by Hinson et al. (14). Protein concentrations were determined by the method of Lowry et al. (15)using bovine serum albumin as a standard. To find the optimal protein concentration and time period for which HQ oxidation remained linear, HQ (300 mM) was incubated with various concentrations of microsomal protein, ranging from 0.5 to 4 mg/mL for various periods of time. Subsequent incubations contained 2 mg of rat liver microsomal protein and 20 mM 4-(2hydroxyethy1)-1-piperazineethanesulfonic acid (HEPES) buffer (pH 7.4), in a final volume of 1 mL. Experimental samples contained 100pl of a NADPH generating system, which consisted of 8 mM glucose 6-phosphate,0.52 mM NADP+,15pg of glucose6-phosphate dehydrogenase, and 1 mM magnesium chloride dissolved in 20 mM HEPES buffer (pH 7.4). Unless otherwise noted, the mixtures were incubated with 300 pM HQ (dissolved in 30 pL of methanol) or 300 pM of the GSH conjugates (dissolved in 300 pL of 20 mM HEPES buffer, pH 7.4), at 37 OC for 30 min in a Dubnoff metabolic shaking incubator. For the inhibitor studies, samples were preincubated with the inhibitor for 2 min at 37 "C prior to the addition of substrate. The reactions were stopped by adding 1% perchloric acid to the samples and placing them in an ice bath. After the protein was precipitated (approximately 10min), the samples were centrifuged at l6000g for 2 min and the supernatants collected for HPLC analysis. A series of control incubations were performed to determine the NADPH dependence of HQ oxidation. Samples were incubated in the presence or absence of the following factors as single variables: GSH; NADPH generating system; rat liver microsomes; or HQ. Products formed in incubations containing either freshlyprepared microsomes,previously frozen microsomes (stored at -70 OC), or boiled microsomes were also compared. Microsomal samples were analyzed by HPLC (Shimadzu LC6A with CR5A integrator). Aliquota (100pL) were injected onto a partisil5-pm ODs-3reverse-phase analyticalcolumn (Whatman, Clifton, NJ). The samples were eluted with an isocratic mixture of methanol, water, and acetic acid (8:91:1) at a flow rate of 1 mL/min for 25 min, and the eluate was monitored at 310 nm. Under these conditions, authentic samples of 2-(GSyl)HQ,2,3(diGSyl)HQ, 2,5-(diGSyl)HQ, 2,6-(diGSyl)HQ,and 2,3,5-(triGSy1)HQ were eluted with retention times of 11.1,7.2, 16.3, 17.8, and 9.5 min, respectively. Metabolites were identified by comparison of HPLC retention times with authentic standards. Determination of Oxidation Potentials. Various HQ-Sconjugates were dissolved in the HPLC-EC mobile phase (see above) at a concentration of 10 pM. Half-wave oxidation potentials were determined using HPLC (Shimadzu LC-6A) with EC (coulometric response) detection (ESA Coulochem Model 5100A). An aliquot of each sample (10pL; 50 pmol) was injected onto a Partisil5-pm (Phenomenex,Torrance, CA) ODs-3reversephase analytical column and eluted with the HPLC-EC mobile phase at a flow rate of 1 mL/min. A potential of -0.4 V was applied to the first test electrode to assure complete reduction of the injected hydroquinone-thioethers prior to reaching the second test electrode. The applied potential at the second test electrode was varied from -0.30 to +0.40 V. Peak areas were measured at each potential (Shimadzu C-RIA Chromatopac) and expressed as a percentage of the maximum response obtained. Statistical Analysis. All data (n = 3) were expressed as means h standard error or standard deviation. Mean values were compared by Student's Newman-Kuels test following ANOVA.
Results Identification of Hydroquinone-Thioethers as in Vivo Metabolites of Hydroquinone. The extent to whichHQ (1.8mmol/kg, ip) undergoes oxidation and GSH conjugation in vivo was examined in bile duct- and ureter-
462 Chem. Res. Toxicol., Vol. 6, No. 4 , 1993
A\- A
Hill et al.
4
0
I
5
10
15 20 Time (min)
5
10
15 20 Time (min)
I 25
I 30
25
30
(C)
0
Figure 1. HPLC-UV (Aand B)and HPLC-EC (C and D) analysis of bile from a rat before (A and C) and after (B and D) administration of hydroquinone (1.8 mmol/kg, ip). The HPLC conditions were as follows: the mobile phase (1 mL/min) contained a mixture of 94 % citric acid (4 mM)/a”onium acetate (8 mM) and 6% methanol (pH 4.0) containing 20 mg/L EDTA; UV detection was monitored at 280 nm; EC detection was monitored at -0.20 V (detector 1) and +0.20 V (detector 2). Individual peaks represent: (1) HQ-glucuronide, 4.4 min; (2) HQ, 8.8 min; (3) 2,3,5-(triGSyl)HQ,10.4 min; (4) 2-(GSyl)HQ, 13.1min; (5) 2,5-(diGSyl)HQ,19.3 min; (6) 2,6-(diGSyl)HQ,21.8 min; and (7) unknown, 25.0 min.
cannulated, male Sprague-Dawley rats pretreated with AT-125 (to inhibit y-glutamyl transpeptidase). Preliminary identification of 2-(GSyl)HQ, 2,5-(diGSyl)HQ, 2,6(diGSyl)HQ, and 2,3,5-(triGSyl)-HQ in the bile was achieved by HPLC analysis utilizing EC detection (Figure 1). This technique identified the presence of six electroactive compounds in the bile of HQ-treated rats. The peaks eluting at 10.4,13.1,19.3,and 21.8min corresponded to the retention times of authentic 2,3,5-(triGSyl)HQ, 2-(GSyl)HQ, 2,5-(diGSyl)HQ, and 2,6-(diGSyl)HQ, respectively. The peak eluting a t 8.8 min corresponded to authentic HQ, and the identity of the peak eluting at 25.0 min is not known. In addition, HQ-glucuronide and HQsulfate were identified by HPLC-UV and enzymatic methodology. Confirmation of the initial assignments of the EC-active metabolites of HQ was achieved by HPLC-CEAS and LC/ CF-FAB. The assignment of unknown peaks of bile and urine was determined both by retention time accuracies and by comparison of electrochemical oxidation profiles (or “signature”) across the electrodes (ratio accuracy). In the 16-channeloxidative array used in the present studies, a compound will usually show oxidative responses across three or four electrodes. The ratio of responses across these channels, due to the efficiency of each electrode, is an excellent qualitative index of the identity and purity of the peak (12,13). The maximum signal is generated by the electrode whose potential lies closest to the half-wave potential, and the ratio of this signal to that generated by the preceeding and/or subsequent electrode is compared to that obtained from known standards (12, 13). In the present study, HPLC-CEAS analysis of bile from HQtreated rats revealed the presence of several metabolites with retention times and accuracy ratios consistent with authentic HQ-thioethers (Figure 2). Thus, 2-(GSyl)HQ
(14.33,0.94 [4/51;retention time, ratio accuracy [electrode pair]), 2,3,5-tricystein-S-yl-HQ (17.2, 0.87 13/41), 2,3,5(triGSy1)HQ (20.53, 0.74 [5/41), 2,5-(diGSyl)HQ (22.23, 0.98 E5/41), 2,6-(diGSyl)HQ (23.71,0.92 [3/41,0.93 [5/41), and 2-(cystein-S-ylglycyl)HQ (24.93, 0.94 [4/51) were all identified by HPLC-CEAS. The only HQ-thioether identified in urine of HQ-treated rats was 2-(N-acetylcystein-S-y1)HQ (10.85, 0.82 [3/41) (Figure 3) although both HQ-sulfate and HQ-glucuronide were readily identified by HPLC-UV and enzymatic methodology. The absence of multisubstituted HQ-thioethers in urine, despite their presence in bile, may indicate that the major fraction of these metabolites are retained in the animal. Following a single ip dose of 1.8 mmol/kg body wt HQ, 2-(N-acetylcystein-Syl)HQ was confirmed as a urinary thioether metabolite in the male Sprague-Dawley rat by LC/CF-FAB. The LC/CF-FAB analysis of a neat urine sample demonstrated the presence of 2-(N-acetylcysteinS-y1)HQat an HPLC retention time of 7.4 min. Molecular ion species a t m/z 272 and 270 were observed in positive and negative ion modes, respectively (Figure 4), which were consistent with the mass spectra obtained from the synthetic standard. In addition, the ether glucuronide conjugate of HQ was also observed in the urine sample, eluting near the HPLC solvent front. A molecular ion was observed for this metabolite a t m/z 285 in the negative ion mode. A bile sample collected from a male rat receiving 1.8 mmol/kg HQ was also analyzed for metabolite composition by LC/CF-FAB MS. The reconstructed single-ion chromatograms for relevant HQ-thioether metabolites are depicted in Figure 5. A molecular ion corresponding to the ether glucuronide conjugate of HQ was observed at m/z 285 (negative ion mode). Molecular ions corresponding to the following thioether metabolites were observed: 24GSyl)HQ (mlz 416 and 414 in positive and negative ion modes, respectively; Figure 6) and two (diGSy1)HQ metabolites (mlz 719 in negative ion modes; Figure 7). In general, positive ion spectra were weaker in intensity than the corresponding negative ion spectra for these glutathionyl conjugates. A characteristic fragment ion was detected in the negative ion mode at m/z 306 for both mono- and disubstituted conjugates, corresponding to loss of intact GSH as its thiolate anion. In addition, C-S bond cleavage to HQ thiolate anion was also detected as a prominent fragment ion in the diGSyl conjugate spectra a t m/z 446 (negative ion mode). Although 2,3,5-(triGSy1)HQ was detected in the bile sample by HPLC-CEAS, no molecular ions were observed corresponding to this conjugate by LC/CF-FAB MS analysis. The synthetic standard of 2,3,5-(triGSyl)HQ produced no molecular ions or characteristic fragment ions upon CF-FAB analysis. I t may be postulated that the lack of spectral data for highly substituted glutathionyl conjugates may be related to their relatively high polarity and subsequent inability to desorb from the very polar liquid matrix, glycerol. Quantitation of Thioether Metabolites of Hydroquinone Formed in Rat Bile and Urine. The S-conjugates of HQ in bile and urine samples were quantitated by HPLC-EC. The two major S-conjugates detected after HQ administration were 2-(GSyl)HQin the bile and 2-(Nacetylcystein-S-y1)HQ in the urine (Figure 8). The maximum rate of 2-(GSyl)HQexcretion into bile was observed within the first hour following HQ administration whereas the maximum rate of excretion of 2-(N-acetylcystein-S-
Chem. Res. Toxicol., Vol. 6, No. 4, 1993 463
Glutathione Conjugation of Hydroquinone
ILI BILE
I I1
k
I
*
3 -
iti
- - .
. . .
'
I
"
"
. . . .
I
"
"
1
"
15 /rd A5 :_g 35 Figure 2. HPLC-coulometric electrode array analysis (HPLC-CEAS) of bile and urine collected from a male Sprague-Dawley rat treated with 1.8 mmol/kg hydroquinone. Assignment of the identity of each peak was based upon both the retention time and the ratio accuracy (electrode pair): I, hydroquinone; 11,2-(GSyl)HQ;111,2,3,5-(triGSyl)HQ;IV, 2,5-(diGSy1)HQ;V, 2,6-(diGSy1)HQ;VI, 2-(CysG1y)HQ;and VII, 2-(N-acetylcystein-S-yl)HQ.Each individualchromatogramrepresentsthe signal obtained by applyingincreasing potentials to each electrode, in +50-mV increments. For the bile, channels 1-15 (-250 to +450 mV) are displayed, and for the urine, channels 3-15 (-150 to +450 mV) are displayed, at a full screen current of 1PA. The chromatographic conditions were as described in the Materials and Methods. ~
-
g
5
-
-
-
-
y1)HQ into urine occurred during the first 2 h (Figure 8). This is consistent with the metabolism of 2-(GSy1)HQ via the enzymes of the mercapturic acid pathway to 2-(Nacetylcystein-S-y1)HQ. Small, but significant, amounts of 2,5-(diGSyl)HQ, 2,6-(diGSyl)HQ,and 2,3,5-(triGSy1)HQ were also excreted into bile 2 h following HQ administration (Figure 8, Table I). The excretion of 2,5(di-GSyl)HQ,2,6-(diGSyl)HQ,and 2,3,5-(triGSy1)HQ into bile suggests that 2-(GSyl)HQ can undergo further oxidation and reaction with GSH in vivo. The total urinary and biliary S-conjugates formed within 4 h following in vivo administration of HQ are listed in Table I. The largest
amount of S-conjugate formed was 2-(GSy1)HQ (18.9 f 2.7 pmol), which was followed by 2-(N-acetylcystein-Sy1))HQ (11.4 f 3.6 pmol) (Table I). The total amount of S-conjugates excreted into urine and bile within 4 h of HQ administration was 4.3 f 1.1%of the dose (34.3 f 4.5 pmol) (Table I). Finally, there was a small quantity of free HQ (-0.25 % of the administered dose) detected in the bile and urine (0.87 f 0.39 and 1.11 f 0.26 pmol, respectively). Characterizationof Rat Liver Microsomal Hydroquinone Oxidation. The mechanism of HQ oxidation is unclear, although several lines of evidence suggested that
464 Chem. Res. Toxicol., Vol. 6, No. 4, 1993 (A)
I 0
I
5
I
I
10
15
I
I
I
20
25
30
I
I
I
I
15
20
25
30
Time (min)
I
0
5
10
iI
Time (min)
Figure 3. HPLC-UV (Aand B) and HPLC-EC (Cand D) analysis of urine from a rat before (A and C) and after (Band D) administration of hydroquinone (1.8mmol/kg, ip). The HPLC conditions were as follows: the mobile phase (1 mL/min) contained a mixture of 94% citricacid (4mM)/ammonium acetate (8 mM) and 6% methanol (pH 4.0) containing 20 mg/L EDTA; UV detection was monitored at 280 nm; EC detection was monitored at -0.20 V (detector 1) and +0.20 V (detector 2).
CZ
miz
Individual peaks represent the following: (1)HQ-glucuronide, 4.4 min; (2) HQ-sulfate, 6.4 min; (3) peak associated with the urine, 7.0 min; (4) HQ, 8.8 min; and (5) 2-(N-acetylcystein-Syl)HQ, 18.7 min.
Figure 4. LC/CF-FAB mass spectra of 2-(N-acetylcystein-Sy1)HQ detected in rat urine in (A) positive ion and (B) negative ion modes. Molecular ions were observed at m/z 272 and 270, respectively. The ion at m/z 289 in the positive ion mode
Table I. Quantitation of Urinary and Biliary S-Conjugates of Hydroquinone metabolite pmol/4 h (% dose) 18.9 f 2.7 (2.3 f 0.6) 2-glutathion-S-ylhydroquinone 2.2 f 0.6 (0.3 f 0.1) 2,5-diglutathion-S-ylhydroquinone 0.7 f 0.3 (0.1 f 0.1) 2,6-diglutathion-S-ylhydroquinone 2,3,5-triglutathion-S-ylhydroquinone 1.2 f 0.1 (0.2 f 0.01) 2-(N-acetylcystein-S-yl)hydroquinone 11.4 f 3.6 (1.5 f 0.6) 34.3 f 4.5 (4.3 f 1.1) total (S-Conjugates) Values represent the mean f SE (n= 3). AT-125 (10mg/kg, ip) was given 1 h prior to HQ (1.8 mmol/kg, ip). Urine and bile were collectedfor 4h in 1-hfractions, and a 10-rLaliquot of a 1:40dilution was analyzed by HPLC-EC. Quantitation of each peak consisted of adding the total rmol of product formed at each time point over the 4-h collection period. Numbers in parentheses represent percentage of the administered dose.
at m/z 362 in the negative ion mode corresponds to a glycerol
(I
this oxidation was mediated by a NADPH-dependent microsomal enzyme. However, whether or not this could be attributed to cytochrome P450 was unclear. We therefore investigated microsomal HQ oxidation in more detail, with particular reference to the oxidation not only of HQ but also of the HQ-GSH conjugates. We initially determined the kinetics for the microsomal oxidation of HQ [ K , = 120 pM, V,, = 2.2 nmol/(mg.min)] and subsequently performed all other experiments using substrate concentrations of 300 pM. Rat liver microsomes catalyzed the oxidation of HQ (300 pM) in the presence of GSH (1mM) and a NADPH generating system (Table 11). The major product of HQ oxidation was 2-(GSyl)HQ, but both 2,5-(diGSyl)HQ and 2,6-(diGSyl)HQ were also formed. In contrast, HQ oxidation decreased 92 5% in heatinactivated microsomes, and 97 % in microsomes lacking a NADPH generating system. Thus, the liver microsomal oxidation of HQ required microsomal protein and was
corresponds to an ammonium adduct of the conjugate; the ion
adduct of the conjugate.
Table 11. Effect of Cofactors on Liver Microsomal Oxidation of Hydroauinone (300rcMP nmoll subcellular (mg of proteimmin) (% control) fraction cofactors Experiment 1 microsomes NADPH generating system 2.23 f 0.01 (100) boiled NADPH generating system 0.18 f 0.01 (8 f l ) b microsomes microsomes none 0.07 f 0.01 (3 f l ) b Experiment 2 microsomes NADPH generating system 2.41 f 0.04 (100) microsomes NADPH (1 mM) 1.41 f 0.02 (59 f l ) b microsomes NADH (1 mM) 0.96 f 0.05 (40 f 2 ) b Experiment 3 cytosol NADPH generating system 0.06 f 0.01 0.05 f 0.003 cytosol none Incubations were conducted in HEPES buffer (pH 7.4) at 37 O C in the presence of 300pM HQ, 1mM GSH,and 2 mg/mL microsomal protein. Values represent the mean f SE of 3 determinations. Values significantlydifferent from microsomessupplemented with a NADPH generating system at p < 0.01. NADPH-dependent. Replacement of the NADPH generating system with NADPH also supported HQ oxidation, although as expected, this was less efficient. Interestingly, NADH also supported the microsomal oxidation of HQ. Rat liver cytosolic protein did not support the oxidation of HQ in either the presence or absence of a NADPH generating system. We further delineated the nature of liver microsomal HQ oxidation (Table 111). Metyrapone (50 pM), which inhibits some but not all isoforms of cytochrome P450,
Glutathione Conjugation of Hydroquinone
Chem. Res. Toxicol., Vol. 6, No. 4, 1993 465 OH
m/z:285 100
E+06
2.807
5 0-
Time (min)
Figure 5. Reconstructed single-ion chromatograms of m/z 285,414, and 719, and the total-ion chromatogram (RIC) obtained from LC/CF-FAB analysis of biliary metabolites of HQ in the rat. Corresponding proposed structures are provided with each individual ion chromatogram. The position of addition of the second GSH molecule to HQ cannot be determined from the spectra alone, but is also based on its chromatographic (retention time) properties. iEt04
(M+H)+
I
o-.. 1
416.1
OH
.Et05
I
OH
446.2
380
400
688
598
880
m/z
(M-H)4
2
sE+05
Figure 7. LC/CF-FAB mass spectra of 2,5-diglutathion-S-y1HQ detected in negative ion mode. A molecular ion was observed at m/z 719; thiolate anion formation was observed at m/z 446, and a base peak ion at m/z 306 corresponds to loss of GSH. Table 111. Effects of Various Inhibitors on the Microsomal Oxidation of Hydroquinone (300 pM). ~~~
~
inhibitors 50 pM metyrapone
~~
1mM SKF-525A 1mM 1-benzylimidazole 1mM aminobenzotriazole
~
~~
~
~~
53 f I d
? inhibitors inhibition 1mM sodium azide 26 f 7 c 10 pM rotenone 8 f 4b 38 p M antimycin A 15 f 1"
30 f 6c
1mM KCN
%
inhibition 28 f 2c 82 f Id
9fF
58f.3
0388
4,i"b 560
350
, . .,
.,
558 ,
,
,
,
450
480
mlz
Figure 6. LC/CF-FAB mass spectra of 2-glutathion-S-yl-HQ detected in rat urine in (A) positive ion and (B) negative ion modes. Molecular ions were observed at m/z 416 and 414, respectively. Glycerol adducts were observed at m/z 508 and 506, respectively. decreased microsomal HQ oxidation by 28 % (Table 111). 1-Benzylimidazole (1 mM) and aminobenzotriazole (1 mM), which are relatively selective P450 inhibitors, and inhibit most P450 enzymes, decreased HQ oxidation by 53% and 30%, respectively. SKF-525A (1 mM), a nonselective inhibitor of cytochrome(s) P450, decreased microsomal HQ oxidation by 82 % . These data suggest that a portion of microsomal HQ oxidation is catalyzed by
All data are expressed as the mean i standard deviation of 3 experiments. Treatments that resulted in different responses were foundto be statisticallysignificantat *p< 0.05 and cp< 0.01 [control value = 2.41 f 0.07 nmol/(min-mg)]and d p < 0.01 [control value = 1.74 f 0.02 nmol/(min*mg)l. a
cytochrome(s)P450, although the exact amount remains unclear. Sodium azide (1mM), an inhibitor of peroxidasecatalyzed reactions, decreased HQ oxidation by 26 %. As noted above, NADH supported microsomal HQ oxidation a t rates 60% of that observed in the presence of the NADPH generating system, suggesting possible contamination of the microsomal preparation with mitochondrialderived oxidases. However, contamination by mitochondrial electron transport chain components was ruled out in studies utilizing rotenone (10 pM), antimycin A (38 pM), and KCN (1 mM). None of these inhibitors of
466
Chem. Res. Toxicol., Vol. 6, No. 4, 1993
Hill et al.
TIME (hr)
Z.S-(diGSyl)HQ Bile
TIME (hr)
Figure 8. Time-dependent excretion of (A) 2-(GSy1)HQ and 2-(N-acetylcystein-S-yl)HQ into bile and urine, respectively,and (B)2,5-(diGSyl)HQ,2,6-(diGSyl)HQ,and 2,3,5-(triGSyl)HQinto bile obtained from male Sprague-Dawley rate treated with 1.8 mmoVkg (ip). Each point represents the mean SD (n = 4). Table IV. Microsomal Oxidation of Hydroquinone and Its Glutathione Conjugates* complete system substrate (300 pM) +NADPH -NADPH (%) (mV) 1643 f 52b 82 f 4b (5) +12 hydroquinone -34 2-glutathion-S-ylhydroquinone 563 45c 136 f 2c (24) 812 f 32d 352 Bd (43) -85 2,5-diglutathion-S-ylhydro quinone 2,6-diglutathion-S-ylhydro 1088 Id 253 6od (23) -79 quinone
*
* *
*
Values expressed as pmol/(mg-min)and represent the mean standard deviation (n= 3). b Quantified as the sum of the formation of 2-glutathion-S-yl-HQ,2,5-diglutathion-S-yl-HQand 2,6-diglutathion-S-yl-HQ. Quantified as the sum of the formation of 2,5diglutathion-8-yl-HQand 2,6-diglutathion-S-yl-HQ. Quantifiedas the formation of 2,3,5-triglutathion-S-yl-HQ. Numbers in parentheses represent the percentage of the substrate that undergoes nonNADPH-dependent oxidation (autoxidation).
mitochondrial respiration significantly affected HQ oxidation (Table 111). Interestingly, pretreatment of animals with either phenobarbital or 8-naphthoflavone did not lead to induction of microsomal HQ oxidation (data not shown) despite induction of other P45O-related activities (e.g., p-nitroanisole 0-demethylase). HQ was more effectively oxidized by liver microsomes than its corresponding GSH conjugates (Table IV). Both 2,5-(diGSyl)HQ and 2,6-(diGSy1)HQwere found as products of HQ (300 pM) oxidation (or more correctly, as products of subsequent 2-(GSy1)HQ oxidation) a t 120 f 7 and 32 f 6 pmol/(mg.min) (X f SE; n = 3). When 2-(GSyl)HQwas used as the substrate, both 2,5-(diGSyl)HQ and 2,6-(diGSyl)HQwere formed (Table IV), with the former being the major product. Rat liver microsomes also effectively catalyzed the oxidation of 2,6-(diGSyl)HQ and 2,5-(diGSyl)HQ,to 2,3,5-(triGSyl)HQ(Table IV). The extent of oxidation and GSH conjugation of 2-(diGSyl)HQ,2,5-(diGSyl)HQ,and 2,6-(diGSyl)HQranged from
34% to 66% of the specific activity of HQ (Table 111). Whereas the non-NADPH-dependentoxidation of HQ was minimal (Table IV), the “autoxidation” (NADPH-independent oxidation) of 2-(GSyl)HQ, 2,5-(diGSyl)HQ, and 2,6-(diGSyl)HQwas a significant component of the overall oxidation (Table IV). The NADPH-independent oxidation correlated approximately with the corresponding halfwave oxidation potentials (Table IV). Thus, 2,5-(diGSyl)HQ and 2,6-(diGSyl)HQunderwent NADPH-independent oxidation more readily than either HQ or 2-(GSyl)HQ. Kidney microsomes supplemented with a NADPH generating system did not catalyze the formation of 2-(GSyl)HQ from HQ. To exclude the possibility that any 2-(GSyl)HQformed in kidney microsomes might undergo further metabolism by y-glutamyl transpeptidase, AT125was also added to the incubations. Again, no evidence for the enzymatic formation of 2-(GSyl)HQ was found. HPLC analysis also failed to reveal the presence of metabolites [either 2-(cystein-S-ylglycyl)HQor 2-(cysteinS-y1)HQI arising from the further metabolism of 2-(GSy1)HQ by y-glutamyl transpeptidase. The effects of scavengers of reactive oxygen species on the oxidation of HQ, 2-(GSyl)HQ, 2,5-(diGSyl)HQ, and 2,6-(diGSyl)HQ were also investigated. Superoxide dismutase (500 units/mL) significantly (p < 0.01) decreased 2,3,5-(triGSyl) formation from 2,6-(diGSyl)HQ by 33 % , but had no effect on the remaining compounds. Catalase (200 units/mL) had no effect on HQ-GSH conjugate oxidation but significantly (p< 0.01) increased the yield of 2-(GSyl)HQ from HQ by 14% whereas methanol (3% v/v), a *OH radical scavenger, significantly (p < 0.01) decreased 2-(GSyl)HQ formation from HQ by 23% and was without effect on the conjugates.
Discussion Previous studies addressing the metabolism of HQ in mammals after oral intake indicated that HQ was largely conjugated with glucuronic acid and sulfate and excreted in the urine (2,3). More recently, Nerland and Pierce (4) demonstrated that 2-(N-acetylcystein-S-yl)HQ was excreted in the urine of rats treated with HQ, indicating that GSH also participates in the metabolism of HQ. Our present studies confirm and extend this work. Using a combination of HPLC-CEAS and LC/CF-FAB mass spectroscopy, we have demonstrated the presence of 2-(GSyl)HQ, 2,5-(diGSyl), 2,6-(diGSyl)HQ, and 2,3,5(triGSy1)HQ in the bile of rats treated with HQ. This is the first time that the multi-S-substituted conjugates have been identified as in vivo products of HQ metabolism. We also confirmed that 2-(N-acetylcystein-S-yl)HQis a major urinary metabolite of HQ. The observation that several mono- and multi-GSH conjugates of HQ were excreted in bile while only 2-(N-acetylcystein-S-yl)HQwas detected in the urine suggests that the disposition of these quinonethioethers is different. Support for this derives from studies utilizing the in situ perfused rat kidney. In this model, perfusion with 2,3,5-(triGSyl)HQresulted in 100% retention by the kidney whereas perfusion with 2-(GSyl)HQ resulted in the urinary excretion of the anticipated mercapturic acid pathway metabolites, 2-(cystein-S-ylglycine)HQ, 2-(cystein-S-yl)HQ, and 2-(N-acetylcysteinS-yl)HQ.2 In addition, the identification of 2-(cystein2
Hill,Davison, Dulik,Monks, and Lau,unpublished data.
Glutathione Conjugation of Hydroquinone 5-ylglycy1)HQ in bile and 2- (N-acetylcystein-S-yl)HQin urine of AT-125-pretreated rats indicates that despite the ability of this protocol to inhibit y-glutamyl transpeptidase activity by 94% (16),sufficient residual enzyme activity remains to metabolize a t a least a portion of 2-(GSyl)HQ. Moreover, the presence of 2-(N-acetylcystein-S-yl)HQin the urine of bile duct-cannulated rata indicates that a portion of 24GSyl)HQ must be excreted from hepatocytes into the circulation via the sinusoidal membrane. An alternative interpretation is that organs other than the liver, such as the lung, also metabolize HQ to thioether metabolites. Quantitation of the various HQ-thioethers in urine and bile within 4 h of HQ administration showed that 4.3% of the dose was recoverable as products arising from the in vivo oxidation of HQ. Several lines of evidence suggest that this value is a conservative estimate of the actual fraction of HQ that undergoes oxidation and thioether formation in vivo. As noted above, a portion of the HQthioethers formed in the liver appear capable of being excreted from the sinusoidal pole of the hepatocyte. Consistent with this interpretation, considerably more thioether metabolites were excreted into urine of 2-BrHQtreated rats than were excreted into bile suggesting that a fraction of these GSH conjugates are exported into blood in addition to bile. This is important in terms of assessing HQ oxidation and thioether formation in vivo since metabolism of the multi-GSH-substituted conjugates by the kidney results in the formation of reactive intermediates that are preferentially retained by the kidney, rather than excreted. In support of this scenario, infusion of 2,3,5-(tri-GSyl)[l4C1HQ into the in situ perfused rat kidney results in 100% retention of the administered dose (see above). In addition, following administration of radiolabeled 2-hydroxy-1-glutathion-S-ylestradiol to rats, only 15 % of the dose was recovered in the urine, and 5 % in the feces, after several days (18), indicating that the major fraction of this catechol-thioether was also retained in the animal. Thus, an estimation of HQ oxidation and thioether formation in vivo by only quantifying metabolites excreted into urine and bile may underestimate the significance of this pathway. Moreover, our observation that only the mercapturate corresponding to 2-(GSyl)HQ could be identified in urine, despite identification of multiGSH-substituted HQ-thioethers in bile, is consistent with the above interpretation. Interestingly, the N-acetylcysteine derivative of 2-hydroxy-1-glutathion-S-ylestradiol was the onlymetabolite of this thioether identified in urine
(In,
(18).
The National ToxicologyProgram recently reported that HQ was nephrocarcinogenic in male Fischer 344 rats (8), a finding recently confirmed by Shibata et al. (9). In addition, we have previously shown that several of the GSH conjugates now identified as in vivo metabolites of HQ, particularly 2,3,5-(triGSyl)HQ, are relatively potent nephrotoxicants (IO). Whether or not GSH conjugates of HQ play a role in HQ-mediated nephrocarcinogenicity is unclear. However, the total amount of the HQ S-conjugates excreted into urine and bile (34.3 f 4.5 pmol, even assuming this value to be a maximum estimate, see above) appears sufficient to propose a role for such metabolites in HQ-mediated nephrotoxicity and nephrocarcinogenicity. For example, a dose of 20 pmollkg 2,3,5-(triGSy1)HQ causes severe renal proximal tubular necrosis in rats as evidenced histologically and by measuring elevations in
Chem. Res. Toxicol., Vol. 6, No. 4, 1993 467 blood urea nitrogen (IO). This dose represents only8 pmol/ 400 g rat, the weight of animal employed for the metabolism studies reported herein. Thus, the excretion of 1.2 f 0.1 pmol of 2,3,5-(triGSyl)HQ, or a total of 4.1 f 0.6 pmol for the combination of 2,3,5-(triGSyl)HQ, 2,5-(diGSyl)HQ, and 2,6-(diGSyl)HQ,may be sufficient to produce the more subtle effects associated with chronic exposure to HQ, and that culminate in the nephrocarcinogenic response. Moreover, in rats chronically exposed to HQ, concentrations of the cofactors required for HQ conjugation (UDPglucuronic acid and 3’-phosphoadenosine 5’-phosphosulfate) may become depleted, leading to a greater fraction of the dose undergoing oxidation and thioether formation. Further work is required to assess the validity of this scenario. Tajima et al. (19) have also recently reported the identification of several thioether metabolites of 3-tertbutyl-4-hydroxyanisole in rat urine. Incubation of 2-tertbutylhydroquinone, a metabolite of 3-tert-butyl-4-hydroxyanisole in both humans and rats, with liver microsomes in the presence of GSH resulted in the formation of 2-tert-butyl-5-glutathion-S-ylhydroquinone and 2-tertbutyl-6-glutathion-S-ylhydroquinone. Since 3-tert-butyl4-hydroxyanisole increases the formation of preneoplastic and neoplastic foci in the kidney of rats (20), the role of thioethers in these processes warrants further investigation. Catechol estrogens have been implicated in estradiolmediated nephrocarcinogenicity in Syrian hamsters (21). These estradiol metabolites are readily oxidized and, as noted above, form potentially reactive quinone-thioethers. Several aminophenols have also been shown to be nephrotoxicants. For example, 4-aminophenol causes acute renal proximal tubular necrosis following administration to rats (22-24). In reactions analogous to that of HQ oxidation, oxidative metabolism of 4-aminophenol to the quinone imine and reaction with GSH gives rise to several isomeric multisubstituted conjugates (25).Gartland et al. (26) demonstrated that either depletion of hepatic GSH by pretreatment of animals with buthionine sulfoximine or cannulation of the bile duct to decrease the delivery of hepatic metabolites to the kidney afforded protection against 4-aminophenol nephrotoxicity. These data clearly implicated a role for GSH conjugation in 4-aminophenol nephrotoxicity. Subsequently, Fowler et al. (27)investigated the toxicity of 4-amino-3-glutathionS-ylphenol in male Fischer 344 rats and showed that the conjugate was capable of reproducing 4-aminophenol nephrotoxicity a t doses 3-4-fold lower than that of the parent aminophenol. Klos et al. (28) have also reported the identification of 4-amin0-2-glutathion-S-ylpheno1,
4-amino-3-glutathion-S-ylphenol,4-amino-2,5diglutathionS-ylphenol, and 4-amino-2,3,5(or 6)-triglutathion-Sylphenol in the bile of Wistar rata following administration of 4-aminophenol (100 mg/kg, ip). The latter three conjugates were all capable of causing cytotoxicity when incubated with rat kidney cortical cells, and the toxicity could be prevented by inhibition of y-glutamyl transpeptidase. These data support the proposal that conjugation of redox-active compounds with GSH may be a common mechanism of nephrotoxicity. The realization that quinone-thioethers can possess as much, if not more, biological (re)activity than the parent quinones is now becoming apparent (29). The addition of GSH to quinones frequently has little effect on their redox behavior and in some instances may even facilitate oxidation of the quinol(30-38). Quinone-thioethers also
468 Chem. Res. Toricol., Vol. 6, No. 4, 1993
Hill et al.
interact with several enzymes that have either the quinone their potent toxicological activity suggests that they may or GSH as their "usual" substrate or cosubstrate, respeccontribute to the adverse effects seen in rats following tively (39-41). In addition to being potent nephrotoxicants chronic HQ administration. (10, 16, 42) quinone-thioethers have been implicated in the formation of cataracts (431,as potent ferrihemoglobinAcknowledgment. The work presented in this paper forming agents (441,and as neurotoxicants (45). Quinonewas supported by an award from the USPHS to S.S.L. (GM 39338). S.S.L. was also the recipient of a Pharmathioethers have been shown to bind to DNA (461,a property ceutical Manufacturers Association (PMA) Foundation that may have implications for HQ-mediated nephrocarFaculty Development award. During the course of this cinogenicity. work T.J.M. was supported in part by award ES 04662. Since we were able to demonstrate formation of multiB.A.H. was supported in part by a PMA Foundation S-substituted conjugates of HQ in vivo, we subsequently Fellowship for Advanced Predoctoral Training in Pharattempted a preliminary characterization of the enzymemacology and presented portions of the research described (e) responsible for the oxidation of HQ. Previous invesabove to the faculty of the University of Texas in partial tigators have suggested that the microsomal enzyme(s) fulfillment toward the Doctor of Philosophy degree. responsible for the oxidation of HQ is NADPH-dependent H.E.K. was supported in part by a NIEHS Toxicology but not cytochrome P450 (5-7). In the present study, we Training Grant (T32 ES 07247). We are indebted to Dr. confirmed that the enzyme(s) responsible for the oxidation Bill Riffee for providing us with the computer-enhanced of HQ is a NADPH-dependent microsomal enzyme (Table copy of Figure 2. We wish to thank Prof. Dan Ziegler for 11). However, according to the results of the inhibitor his advice with some of the in vitro aspects of this study. studies, a significant portion of rat liver microsomal HQ oxidation is catalyzed by cytochrome(s) P450. Thus, References although metyrapone, an isoform-specific P450 inhibitor (1) Ishiguro, S.,Saugawara, H., Kusama, M., Yano, S., Shimojima, N., (47))inhibited the oxidation of HQ by only 28% (Table and Sugawara,S. (1976)Glass capillary column gas chromatographic 111),1-benzylimidazoleand aminobenzotriazole, relatively analysisof tobaccoandcellulose cigarettesmoke. I. Acidic fractions. selective inhibitors of cytochrome(s) P450 (48-50) and Sci. Pap. Cent. Res. Inat. Jpn. Tob. Salt Public Corp. 118,207-211. (2) Dean, B. J. (1978)Genetic toxicology of benzene, toluene, xylenes SKF-525A, anonselective inhibitor of cytochrome(s) P450 and phenols. Mutat. Res. 47,75-97. which also inhibits non P450 microsomal oxidations (501, (3) Divincenzo, G. D., Hamilton, M. L. Reynolds, R. C., and Ziegler, D. inhibited HQ oxidation between 30% and 82% (Table A. (1984)Metabolic fate and disposition of [W]hydroquinone given 111). The exact amount of microsomal HQ oxidation orally to Sprague-Dawley rata. Toricology 33,9-18. catalyzed by cytochrome(s) P450 remains unclear. (4) Nerland, D. E., and Pierce, W. M. (1990)Identification of N-acetylS-(2,5-dihydroxyphenyl)-L-cysteine as a urinary metabolite of Liver microsomes also catalyzed the oxidation of 24Gbenzene, phenol, and hydroquinone. Drug Metab. Dispos. 18,958Sy1)HQ to 2-(GSyl)-1,4-benzoquinone, which in the pres961. (5) Sawahata, T., and Neal, R. A. (1983)Biotransformation of phenol ence of GSH gives rise to 2,5-(diGSyl)HQ and 2,6to hydroquinone and catechol by rat liver microsomes. Mol. (diGSy1)HQ (Table IV). Both 2,5-(diGSyl)HQ and 2,6Pharmacol. 23,453-460. (diGSy1)HQwere also oxidized in liver microsomes to 2,3,5(6) Tunek, A., Schelin, C., and Jergil, B. (1979)Microsomal target (triGSy1)HQ (Table IV). Therefore, the liver exhibits the proteins of metabolically activated aromatic hydrocarbons. Chem.Biol. Interact. 27, 133-144. necessary factors capable of catalyzing the oxidation of (7) Lunte, S.M., and Kissinger,P. T. (1983)Detection and identification HQ to multi-S-substituted conjugates, an important of sulfhydryl Conjugates of p-benzoquinone in microsomal incubacomponent of which appears to be cytochrome P450. tions of benzene and phenol. Chem.-Biol. Interact. 47, 195-212. However, between 23% and 43% of 2-(GSyl)HQ, 2,5(8) Kari, F. W. (1989) Toxicology and carcinogenesis studies of (diGSyl)HQ, and 2,6-(di-GSyl)HQ oxidation is indepenhydroquinone. National Toxicology Program; Technical Report Series 366, 1-86. dent of NADPH. This "autoxidative" component corre(9) Shibata,M.-A., Hirose, M., Tanaka, H., Asakawa, E., Shirai, T.,and lated, approximately, with the E l p values for these Ito, N. (1991)Induction of renal cell tumors in rata and mice, and conjugates (Table IV). Interestingly, however, neither enhancement of hepatocellular tumor development in mice after catalase nor methanol affected the oxidation of 2-(GSyl)long-term hydroquinone treatment. Jpn. J. Cancer Res. 82,12111219. HQ, 2,5-(diGSyl)HQ, or 2,6-(diGSyl)HQ. Consistent with (IO) Lau, S. S.,Hill,B. A., Highet,R. J., andMonks,T. J. (1988)Sequential previous reports ( 5 , 7 ) ,superoxide dismutase had no effect oxidation and glutathione addition to 1,4-benzoquinone: Correlation on HQ oxidation. Only 2,6-(diGSyl)HQ oxidation was of toxicity withincreased glutathione substitution. Mol. Pharmacol. 34,824-836. significantly inhibited by superoxide dismutase. The basis (11) Spence, J. T.,Cotton, J. W., Underwood, B. J., and Duncan, C. P., for such differences is unclear, but Eyer and Kiese (44) Eds. (1983)Elementary Statistics, Prentice-Hall, Englewood Cliffs. also reported that superoxide anions were involved in the (12) Matson, W. R., Langlais, P., Volicer, L., Gamache, P. H., Bird, E., autoxidation of 4,4-dimethyl-p-amino-2,6-diglutathion-Sand Mark, K. A. (1984) n-Electrode three-dimensional liquid ylphenol, but not of 4,4-dimethyl-p-aminophenol.HQ also chromatography with electrochemical detection for determination of neurotransmitters. Clin. Chem. 30, 1477-1488. undergoes peroxidase-mediated oxidation (51-54). The (13) Mataon, W.R., Gamache, P. H., Beal, M. F., and Bird, E. (1987)EC relative amount of P450-catalyzed and peroxidase-cataarray sensor concepts and data. Life Sci. 41,905-908. lyzed HQ oxidation in vivo is not known. However, on the (14) Hinson, J. A., Nelson, S. D., and Mitchell, J. R. (1977)Studies on basis of the studies with sodium azide (Table 1111, an the microsomal formation of arylating metabolites of acetaminophen and phenacetin. Mol. Pharmacol. 13,625-633. inhibitor of peroxidase-catalyzed reactions (55),the frac(15) Lowrv. 0.H.. Rosebrounh. N. J.. Farr, A. L.. and Randall, R. J. tion of liver microsomal peroxidase-mediated HQ oxidation (1956Protein measurement with the Folin phenol reagent. j .Biol. may be small. Liver cytosolic peroxidases have also been Chem. 193,265-275. reported to contribute little to HQ oxidation (7). (16) Monks, T.J., Lau, S. S., Highet, R. J., and Gillette, J. R. (1985) Glutathione conjugata of 2-bromohydroquinone are nephrotoxic. In conclusion, we have shown for the first time that Drug Metab. Dispos. 13,553-559. multi-S-substituted conjugates are formed as metabolites (17)Lau,S.S.,andMonks,T. J. (1990)Theinvivodispitionof2-bromoof HQ in vivo. Although these metabolites are not formed [Wl-hydroquinone and the effect of y-glutamyl transpeptidase inhibition. Toricol. Appl. Pharmacol. 103,121-132. as major quantitative metabolites of HQ (
