Chemical and Immunochemical Comparison of Protein Adduct

inflammatory drugs (NSAIDs) zomepirac (ZP) and diflunisal (DF), the hypolipidemic agent clofibric acid (CA), and the anti-epileptic agent valproic aci...
11 downloads 0 Views 219KB Size
Download PDF
Chem. Res. Toxicol. 1996, 9, 659-666

659

Chemical and Immunochemical Comparison of Protein Adduct Formation of Four Carboxylate Drugs in Rat Liver and Plasma Mark J. Bailey and Ronald G. Dickinson* Department of Medicine, The University of Queensland, Brisbane, Queensland 4029, Australia Received January 30, 1996X

Carboxylate drugs usually form acyl glucuronide conjugates as major metabolites. These electrophilic metabolites are reactive, capable of undergoing hydrolysis, rearrangement, and covalent binding reactions to proteins. The last-mentioned property has the potential to initiate immune and other toxic responses in vivo. In this study, we compared the extent and pattern of covalent adduct formation in plasma and livers of rats dosed with the nonsteroidal antiinflammatory drugs (NSAIDs) zomepirac (ZP) and diflunisal (DF), the hypolipidemic agent clofibric acid (CA), and the anti-epileptic agent valproic acid (VPA). These drugs form acyl glucuronides with diverse intrinsic reactivities (apparent first order degradation t1/2 values of 0.5, 0.6, 3, and 60 h, respectively). Rats were dosed iv twice daily for 2 days (50 mg/kg for ZP, DF, and CA, 150 mg/kg for VPA). Chemical analysis of tissues obtained 6 h after the last dose revealed adduct concentrations of 0.31, 0.44, 0.28, and 0.05 µg of drug equivalents/mL of plasma and 2.21, 2.31, 0.96, and 0.96 µg of drug equivalents/g of liver for ZP, DF, CA and VPA treatments, respectively. For both plasma and liver, the higher concentrations of adducts were found with ZP and DF, which have the more reactive glucuronides. The low concentrations of VPA adducts found in plasma were in keeping with the very low reactivity of its glucuronide. In liver, however, VPA adducts achieved concentrations of the same order of magnitude as the other drugs and were accompanied by adducts of the (E)-2-en metabolite of VPA at 0.38 µg of VPA equivalents/g of liver. The liver data for VPA can be explained by an acyl CoA/βoxidation pathway of adduct formation in addition to that from acyl glucuronidation. Immunoblotting using rabbit polyclonal antisera raised against synthetic drug-protein adducts revealed major bands at 110, 140, and ∼200 kDa in livers of ZP- and DF-treated rats. A fourth major band at 70 kDa in ZP-treated liver had the same apparent molecular weight as the only major band detected in CA-treated liver. A 140 kDa band was detected in liver tissue from VPA-treated rats, as well as several lower molecular weight bands. In plasma, the antisera specifically detected drug-modified serum albumin in samples from rats treated with ZP, DF, and CA, but not VPA. The results with this small series of carboxylate drugs suggested that (a) adduct concentrations in plasma but not liver could be related to acyl glucuronide reactivity, (b) while some modified proteins detected were common, the pattern of modification varied from drug to drug, and (c) caution should be exercised in attributing adduct formation exclusively to the acyl glucuronidation pathway.

Introduction Conjugation with glucuronic acid is a major metabolic pathway for most drugs containing a carboxylic acid functional group. These acyl glucuronide metabolites are intrinsically reactive molecules both in vitro and in vivo, able to undergo a number of reactions including hydrolysis, rearrangement, and covalent binding to proteins (14). Each of these reactions can have pharmacological or toxicological implications. Hydrolysis, catalyzed by hydroxide ion, β-glucuronidases, esterases, or serum albumin, leads to regeneration of the pharmacologically active drug. Rearrangement (Figure 1) occurs by hydroxide ioncatalyzed acyl migration to yield the β-glucuronidaseresistant 2-, 3-, and 4-O positional isomers. The isomers, unlike the acyl glucuronide itself, can exist transiently in the open chain form of the sugar ring and thereby react * Correspondence should be addressed to this author at the Department of Medicine, Clinical Sciences Building, Royal Brisbane Hospital, Queensland 4029, Australia. Telephone: 61-7-3365 5337; FAX: 61-73365 5444; E-mail: [email protected]. X Abstract published in Advance ACS Abstracts, March 15, 1996.

0893-228x/96/2709-0659$12.00/0

with protein amino groups via the exposed aldehyde group (Figure 1). In this mechanism of covalent drugprotein adduct formation, the glucuronic acid moiety is retained in the adduct (i.e., rearrangement/glycation mechanism). Alternatively, the acyl glucuronide itself can directly interact with nucleophilic -SH, -OH, and -NH2 groups on protein (Figure 1), with loss of the glucuronic acid moiety (i.e., transacylation mechanism). Mechanistic work to date has been essentially limited to serum albumin in vitro, and suggests (on balance) that both pathways are operative (e.g., refs 5-7). Irrespective of the mechanism, such covalent modification of native proteins by foreign compounds has attracted much attention as a possible mechanism for explaining hypersensitivity, hepatotoxicity, and certain other toxic responses to acidic drugs. Covalent adduct formation with plasma protein, notably serum albumin, has now been documented for the acyl glucuronides of many acidic drugs, including zomepirac (ZP)1 (8), tolmetin (9), diflunisal (DF) (10), probenecid (10), carprofen (11), fenoprofen (12), clofibric acid (CA) (13), and valproic acid (VPA) (14). For six drugs inves© 1996 American Chemical Society

660 Chem. Res. Toxicol., Vol. 9, No. 3, 1996

Bailey and Dickinson

Figure 1. Scheme showing rearrangement of an acyl glucuronide (R ) drug moiety) by acyl migration, and possible pathways of covalent binding to proteins. The transacylation pathway should be strongly preferred by the glucuronide itself as compared to its isomers, whereas the glycation pathway (exemplified with the 2-isomer) requires prior acyl migration.

tigated in one laboratory, an excellent correlation (r2 ) 0.995) was observed between the extent of covalent binding to serum albumin in vitro and the apparent firstorder degradation rates (representing intramolecular rearrangement and hydrolysis) of the acyl glucuronides (15). A good correlation (r2 ) 0.873) was also obtained for adduct concentrations in plasma of humans given five of these drugs, after correction for the measured plasma glucuronide concentrations. These results suggested that the extent of modification of plasma protein in vitro and possibly in vivo may be predictable from the degradation t1/2 of the acyl glucuronide. Covalent adduct formation in vivo is not restricted to plasma protein: in rats dosed with DF, adducts were found in liver, intestine, kidney, bladder, and skeletal muscle in addition to plasma (16, 17). Liver proteins covalently modified by acidic drugs have been investigated by immunoblotting, though in the case of diclofenac, different patterns of selective modification were reported from two different laboratories (18, 19). Recently, it has been shown that covalent modification of mouse and rat serum albumin by incubation with the acyl glucuronides of the nonsteroidal anti-inflammatory drugs (NSAIDs) tolmetin (20) and DF (21), respectively, conferred an immunogenic capacity to the native protein in the relevant rodent species. If an immune reaction is responsible for the damage sometimes observed with the use of acidic drugs, the identity and location of modified proteins will be significant in understanding the mechanisms of toxicity. The disposition of the modified protein will be important in its availability to the immune system as an immunogen and a major determinant in the type of hypersensitivity observed. Similarly, the nature of any antigen-immune system interactions will determine the type of tissue damage observed. If functional changes to macromol1 Abbreviations: ZP, zomepirac; DF, diflunisal; CA, clofibric acid; VPA, valproic acid; NSAID, nonsteroidal anti-inflammatory drug; BSA, bovine serum albumin; HSA, human serum albumin; KLH, keyhole limpet hemocyanin; ECL, enhanced chemiluminescence; ELISA, enzymelinked immunosorbent assay; SDS-PAGE, sodium dodecyl sulfatepolyacrylamide gel electrophoresis; TBS, Tris buffered saline; PBS, phosphate buffered saline; AUC, area under the concentration-time curve.

Figure 2. Chemical structures of the drugs.

ecules are responsible for the damage, the identity of modified molecules will obviously be critical to the understanding of disease progression. In the present study, four carboxylate drugs (Figure 2) were chosen to represent a broad range of intrinsic reactivities of acyl glucuronide metabolites. ZP (a NSAID withdrawn from the market because of several unexplained anaphylactic reactions leading to several deaths (8)), DF (another NSAID occasionally associated with hypersensitivity responses (22)), CA (a hypolipidemic agent sometimes associated with hepatotoxicity (23)), and VPA (an anti-epileptic agent infrequently associated with severe hepatotoxicity (24)) have degradation t1/2 values for their acyl glucuronides of 0.5 (25), 0.6 (26), 3,2 and 60 (14) h, respectively. The aim of this study was twofold: (a) to compare the extent and pattern of in vivo protein modification in plasma and liver from rats given carboxylate drugs with diverse in vitro degradation t1/2 values for their acyl glucuronide metabolites, and (b) to determine whether representative data on adduct formation could be reliably obtained from study of a single model carboxylate drug or whether study of several drugs was always prudent.

Experimental Section Materials. ZP, DF, VPA, Tween 20, 1-cyclohexyl-3-(2morpholinoethyl)carbodiimide metho-p-toluenesulfonate, bovine 2

King and Dickinson, unpublished results.

Carboxylate Drug-Protein Adducts serum albumin (BSA), essentially fatty acid free human serum albumin (HSA), and nonanoic acid were purchased from Sigma (St. Louis, MO). Samples of (E)-2-en-VPA and 4-en-VPA were kindly provided by Reckitt and Colman Pharmaceutical Division (Sydney, Australia). N-Hydroxysulfosuccinimide was purchased from Fluka (Buchs, Switzerland). CA was supplied by ICI Pharmaceuticals Division (Macclesfield, U.K.). Keyhole limpet hemocyanin (KLH) was supplied by Pierce Imject (Rockford, IL). Enhanced chemiluminescence (ECL) reagents, nitrocellulose (ECL grade), and hyperfilm ECL were purchased from Amersham (Amersham, U.K.). Goat anti-rabbit IgG (H and L chains)-horseradish peroxidase complex was purchased from Rockland (Gilbertsville, PA). All other chemicals used were of at least analytical grade. Rabbits (New Zealand White) were obtained from the Central Animal Breeding House of The University of Queensland. Male Sprague-Dawley-derived rats (300-350 g) were obtained from The University of Queensland Medical Faculty Animal House. Experiments were approved by the University’s Animal Experimentation Ethics Committee. Preparation of Immunogens ZP-KLH, DF-KLH, CAKLH, and VPA-KLH. This method was adapted from that used to couple caproyl pyrraline to BSA (27). DF (11 mg), CA (9.6 mg) and VPA (7.3 mg), were each dissolved with 10 mg of KLH in 2 mL of water, and the pH was adjusted to 7.0 with 0.05 M NaOH. N Hydroxysulfosuccinimide (5.6 mg) was dissolved in 600 µL of water (solution A). 1-Cyclohexyl-3-(2morpholinoethyl)carbodiimide metho-p-toluenesulfonate (243 mg) was dissolved in 2.3 mL of water (solution B). To the DF, CA, and VPA solutions were added 150 µL of solution A and then 400 µL of solution B. The mixtures were stirred overnight at room temperature and then dialyzed against 3 changes of 1000 volumes of phosphate buffered saline (PBS; 8 mM phosphate, pH 7.4, containing 145 mM NaCl) for 24 h at room temperature. Because of the very limited solubility of ZP in water, ZP-KLH was prepared by a different method. KLH (20 mg) in PBS was dissolved in 2 mL of water and dialyzed against 2 changes of 100 volumes of water over 3 h. ZP (14 mg) was dissolved in 2 mL of H2O and the solution added dropwise with stirring to the KLH solution. Solution A (150 µL) and solution B (400 µL) were added, and the mixture was stirred at room temperature overnight. The solution was dialyzed against 3 changes of 1000 volumes of water for 24 h and finally against 1000 volumes of 10 mM sodium phosphate buffer (pH 7.4) for 12 h at room temperature. These procedures led to the coupling of 1.50 µg of ZP, 4.30 µg of DF, 0.51 µg of CA, and 2.77 µg of VPA per mg of KLH. Similar procedures were used to obtain drug-modified HSA for enzyme-linked immunosorbent assay (ELISA) and inhibition immunoblotting. Coupling of drug to carrier protein was assessed by the chemical cleavage method, described in detail below for drug covalently bound to liver and plasma proteins. Production of Antisera against ZP-KLH, DF-KLH, CAKLH, and VPA-KLH. For each of DF and CA, a set of multisite sc injections consisting of 1.5 mg of modified KLH in Freund’s complete adjuvant was administered to a rabbit. This was followed by three im injections of approximately 800 µg of modified protein in Freund’s incomplete adjuvant given at 2 week intervals. Antibody titer against drug-modified HSA was monitored by ELISA (28) using blood samples taken immediately before each im injection. Two weeks after the final im injection, the rabbits were exsanguinated by cardiac puncture under ketamine anesthesia. The blood was heparinized and the separated plasma stored at -70 °C. Titers of the final antisera (technically antiplasmas) were checked by ELISA. ZP-KLH and VPA-KLH antisera were produced similarly, except that only two im injections were made, and the rabbits were exsanguinated 1 week after the second im injection. Antibody specificity was assessed by inhibition immunoblotting, with each drug antiserum being preincubated with (a) HSA modified by that drug, (b) HSA modified by the other drugs, and (c) blank medium, prior to blotting against liver samples from rats treated with the specific drug and from control rats. As well, direct immunoblotting with each antiserum was

Chem. Res. Toxicol., Vol. 9, No. 3, 1996 661 performed against liver and plasma samples from control rats and from rats treated with each of the four drugs. Administration of Drugs to Rats. The rats were each prepared with a catheter in the right external jugular vein as described previously (29) and then placed unrestrained in metabolism cages. Food and water were supplied ad libitum. After a recovery period of at least 2 h, iv dosing with 50 mg/kg (ZP, DF, and CA, 10 mg/mL in 0.1 M NaHCO3) or 150 mg/kg (VPA, 30 mg/mL) was given twice daily for two days at approximately 12 h intervals. Rats were killed under pentobarbitone anesthesia by exsanguination via the aorta 6 h after the final dose. The livers were perfused with cold normal saline via the portal vein, excised, snap frozen, and stored at -70 °C. Arterial blood was centrifuged, and the plasma stored at -70 °C. Control rats were killed by exsanguination without prior catheterization and their livers and plasma obtained as described above. Analysis of Drug Covalently Bound to Liver and Plasma Proteins. For ZP, DF, and CA, liver tissue was thawed, chopped, and homogenized in 2 volumes of 0.1 M sodium phosphate buffer (pH 4.5). Duplicate 1.0 g homogenate or 0.5 mL plasma samples were precipitated with 3 mL of acetonitrile containing 4% (v/v) acetic acid. The samples were vortexed, ultrasonicated in a bath for 15 min, and vortexed again before centrifugation. Pellets were resuspended in 0.5 mL of 0.01 M sodium phosphate buffer (pH 4.5) before reprecipitation with 3 mL of acetonitrile. This precipitation/resuspension step was repeated nine times to remove noncovalently bound drug and metabolites from the protein pellets. The pellets were gently dried under air and then digested in 0.75 mL of 2 M NaOH solution overnight at 70 °C. After cooling and acidification (1.0 mL of 2 M HCl), 25 µL of internal standard solution (100 µg of CA/mL of methanol for ZP and DF pellets and 100 µg of DF/mL of methanol for CA pellets) was added, followed by extraction with 3.5 mL of ether/hexane (1:1 v/v). After equilibration and separation by centrifugation, the organic layer was removed and evaporated to dryness under a stream of air. After reconstitution of the residue in 100 µL of HPLC mobile phase, 40 µL was analyzed essentially as previously described for DF using isocratic reverse-phase HPLC (16, 30). In brief, the mobile phase was prepared by mixing 530 mL of methanol with an aqueous solution (0.01 M Na2HPO4 adjusted to pH 2.7 with H3PO4, and Na2SO4‚10H2O was then dissolved to a concentration of 4% w/v) to a final volume of 1 L. The flow was 2 mL/min, with column eluent being monitored at 226 nm. Under these conditions, CA, DF, and ZP eluted at approximately 12, 17, and 20 min, respectively. Plasma and liver samples from VPA-treated rats were processed, up to the alkaline digestion step, as for samples from the other drug treatments. An earlier method for analysis of VPA and certain of its metabolites (31) was modified for the current work. After acidification of the alkaline digests of the pellets (above), 2 mL of 1-chlorobutane was added. After equilibration and centrifugation, the organic layer was transferred and equilibrated with 100 µL of 0.5 M NaOH. After centrifugation, the organic layer was aspirated and the remnants were evaporated under a stream of air. Aqueous 1 M HCl/4 M NaCl solution (0.5 mL) and internal standard solution (100 µg/mL nonanoic acid in chloroform, 100 µL) were added. After equilibration, 0.5 µL of the chloroform phase was analyzed by direct injection into the GC/MS column. The GC/MS system was comprised of a Hewlett-Packard Model 7673 autoinjector, a Model 5890 Series II GC, and a Model 4971 MS. System control and data acquisition were performed using the PC based HP-MS Chemstation G1034C software. Analyte separation was achieved on a HP-FFAP capillary column (25 m × 0.22 mm × 0.3 µm film) using helium as the carrier gas at a flow rate of 0.8 mL/min. The GC oven was programmed to hold at 80 °C for 1 min, ramp to 130 °C at 50 °C/min and hold for 0.5 min, and then ramp to 240 °C at 15 °C/min and hold for 12 min. Retention times for VPA, 4-en-VPA, (E)-2-en-VPA, and nonanoic acid were 6.8, 7.3, 7.5, and 8.3 min, respectively. The MS was operated in selected ion mode, and the ions monitored were m/z

662 Chem. Res. Toxicol., Vol. 9, No. 3, 1996

Bailey and Dickinson

Table 1. Reactivity of Acyl Glucuronides and Covalent Binding to Rat Plasma and Livera acyl glucuronide degradation rate constantb (h-1) zomepirac diflunisal clofibric acid valproic acid (E)-2-en-valproic acid

1.39 1.16 0.23 0.012

covalent binding to plasma (µg of drug/mL of plasma)

covalent binding to liver (µg of drug/g of liver)

0.31 ( 0.14 (n ) 4) 0.44 ( 0.15 (n ) 4) 0.28 ( 0.05 (n ) 3) 0.05 ( 0.02 (n ) 5) not detected

2.21 ( 0.78 (n ) 4) 2.31 ( 0.57 (n ) 4) 0.96 ( 0.10 (n ) 4) 0.96 ( 0.25 (n ) 8) 0.38 ( 0.11 (n ) 8)

a For the covalent binding studies, rats were dosed iv at 50 mg/kg (except valproic acid 150 mg/kg) twice daily for 2 days and tissues taken 6 h after the last dose. Results for covalent binding are means ( SD of analyses of duplicate samples, where n ) the number of animals studied. b Apparent first-order degradation rate constants for the acyl glucuronides in buffer at pH 7.4 and 37 °C.

102 and 115 for VPA, 99, 113, and 142 for 4-en-VPA and (E)2-en-VPA, and 115 and 129 for nonanoic acid. Standard curves were prepared from liver and plasma of untreated rats by spiking washed pellets prior to alkaline digestion with calculated amounts of the appropriate drug. Standard curves for each tissue were derived from at least four concentrations. Coefficients of determination (r2) were at least 0.99 in each case. The adequacy of the washing procedure to remove noncovalently bound drug was verified by analysis of liver homogenates and plasma (from untreated rats) which had been spiked with DF at 100 µg/g of tissue or /mL of plasma. Immunoblotting of Drug Modified Proteins. Separation of proteins from plasma and liver homogenates from control and drug-treated rats was carried out on 7.5% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) minigels under denaturing conditions. The gels were then equilibrated with transfer buffer for 20 min before electroblotting onto nitrocellulose (32). Nonspecific binding was minimized by blocking in 5% skim milk powder in Tris buffered saline (TBS; 20 mM Tris base, pH 7.5, containing 150 mM NaCl) for 1 h at room temperature on a shaking platform. Drug antisera were diluted (anti-ZP and anti-DF, 1/5000; anti-CA and anti-VPA, 1/2000) in diluent (TBS containing 1% skim milk powder, 1% BSA, 3% control liver homogenate, and 0.05% Tween 20). The antisera dilutions were determined from preliminary experiments designed to optimize signal to noise ratio. Blots were incubated in antisera overnight at 4 °C with shaking. The blots were then rinsed for 5 min followed by three 10 min washes in TBS containing 0.05% Tween 20. Anti-rabbit IgG-horseradish peroxidase complex was diluted 1/2000 in the diluent described above. Blots were incubated in this solution for 1 h at room temperature with shaking. After washing as before, modified proteins were visualized with the ECL system according to the manufacturer’s instructions. The identity of the major band modified in plasma of ZP-treated rats was investigated using protein staining and immunoblot analysis. In brief, plasma samples from control and ZP-treated rats were run on SDS-PAGE as described above and immunoblotted with ZP antiserum. The major band at ∼70 kDa, not detected in the control samples, had the classic wide band appearance of serum albumin. Confirmation of identification was achieved by demonstrating that this band (a) was detected in plasma from both control and ZP-treated rats using a serum albumin antiserum and (b) was the major protein in plasma using Coomassie protein staining.

Results Rats were given moderately high iv doses of the acidic drugs (50 mg/kg for ZP, DF, and CA, 150 mg/kg for VPA) twice daily for 2 days, and tissues were taken 6 h after the last dose. The concentrations of drug covalently bound to proteins in plasma and liver homogenates, as revealed by alkaline digestion of tissue and measurement of liberated drug, are shown in Table 1. For both plasma and liver, the higher concentrations of adducts were found for ZP and DF. VPA treatment induced a much lower level of modification in plasma, whereas its level of modification of liver tissue was of a similar order of

magnitude to those of the other drugs. A second VPArelated peak appeared in chromatograms of livers from all animals treated with this drug. This peak was identified as the (E)-2-en metabolite of VPA by GC/MS using the authentic reference material. Covalent adducts of (E)-2-en-VPA were not detected in plasma samples from VPA-dosed rats. Adducts of 4-en-VPA, a minor hepatotoxic metabolite of VPA, were not detected in either tissue. Polyclonal antisera were raised in rabbits against each drug coupled through its carboxyl function to the carrier protein KLH. Antibody titer was checked at each stage of the immunization regimen using ELISA with drugmodified or unmodified HSA as the coating antigen. In each case, a considerable elevation in immunoreactivity with the drug-modified HSA was observed after two im injections. For DF and CA, which were studied first, the titers were not substantially increased by a third injection. A third im injection was thus not used in generation of the ZP and VPA antisera. The specificity of each antisera was assessed by inhibition immunoblotting. For example, liver tissue blots from ZP-dosed animals were incubated with (a) ZP antiserum, (b) ZP antiserum containing ZP-modified HSA, and (c) ZP antiserum containing CA-modified HSA. Only when the ZP-modified HSA was included in the incubation was binding of the antisera to specific bands inhibited. Similarly, ZP antiserum detected unique proteins in the livers of ZPtreated rats, but not in the livers of control animals or those treated with the other drugs. These results (not shown) established the presence of antibodies specific for the modifications caused by the individual drug. The antisera were used to identify drug-modified proteins from liver and plasma samples from drugtreated rats, after separation by SDS-PAGE and blotting onto nitrocellulose. Figure 3 shows the results of the immunoblots. Several modified proteins were detected in the livers of ZP- and DF-treated rats, although three major bands with apparent molecular weights of 110, 140, and ∼200 kDa were strongly stained with both drug treatments. A fourth major band at 70 kDa, detected in livers from ZP-treated rats, had the same apparent molecular weight as the only major band detected in livers from CA-treated animals. Several bands were detected in liver tissue from VPA-treated rats, including a major band at 140 kDa and several others between 40 and 55 kDa. A major modified band at 70 kDa was detected in the plasma of rats treated with ZP and DF. This band was just detectable in the plasma of rats dosed with CA, but not in those dosed with VPA. In plasma from control and ZP-treated rats, protein staining and immunoblot analysis using a serum albumin antiserum as well as the drug antiserum confirmed that this band was serum albumin. Several other bands were modified

Carboxylate Drug-Protein Adducts

Figure 3. Immunoblots of liver homogenates and plasma from control (C) and test (T) rats dosed with the acidic drugs twice daily for 2 days. Tissues were removed 6 h after the last dose, and proteins were separated by SDS-PAGE and immunoblotted with drug antisera.

in plasma from ZP- and DF-treated animals. Indeed, ZP had modified many proteins in both plasma and liver.

Discussion It is now well established that acyl glucuronides are intrinsically-reactive electrophilic metabolites of carboxylate drugs, manifesting this reactivity along three common, interrelated pathways. One might expect, given the common chemical origin of reactivity of acyl glucuronides, that any pattern of covalent modification of proteins in a tissue in vivo would not vary greatly from drug to drug, but that the extent of modification would depend on the individual reactivity of a specific drug glucuronide. The present investigation aimed to address such questions in rats using four carboxylate drugs forming acyl glucuronides with diverse intrinsic reactivities. Many factors will contribute to adduct concentrations in a particular tissue at a particular time point in vivo, including the size of the dose and the proportion undergoing glucuronidation, the intrinsic reactivity of the glucuronide, the duration of exposure of glucuronide to target proteins, the availability/saturability of protein binding sites, and the stability of the adducts formed. Glucuronidation occurs mainly in the endoplasmic reticulum of the liver. Subsequent transport of the glucuronide preferentially across the sinusoidal membrane into blood or preferentially across the canalicular membrane into bile will obviously influence the particular pathway of intrahepatic exposure and the overall extent of exposure to plasma proteins. Such preferential transport is determined by a number of factors including molecular size (33, 34). In rats, preferential biliary excretion occurs for compounds with molecular masses of >300-350 Da, which includes most drug glucuronides. For humans, however, the molecular weight threshold for preferential biliary excretion is higher (∼500 Da), and many glucuronides are eliminated primarily into blood. The amount

Chem. Res. Toxicol., Vol. 9, No. 3, 1996 663

of acyl glucuronide recovered in urine is the portion which survives chemical and enzymic degradation during systemic circulation prior to renal excretion. Biliary excretion of acyl glucuronide is usually followed by its enzymic (and chemical) hydrolysis in the gut and absorption of liberated parent drug into portal blood (enterohepatic circulation). Subsequent reglucuronidation in the liver thus enhances exposure of the body (at least the enterohepatic organs) to the reactive glucuronide. A good estimate of in vivo exposure of plasma proteins to an acyl glucuronide can clearly be obtained by measuring the area under the concentration-time curve (AUC) of the glucuronide in plasma. However, reliable estimation of intrahepatic exposure is very problematic, as determination of AUC of glucuronide in liver is not usually feasible. Some information can be obtained from the total amount of glucuronide excreted, but this may grossly underestimate exposure to reactive glucuronides because of systemic and enterohepatic cycling, and it ignores considerations of extrahepatic glucuronidation. In the present study, bile-intact rats were given four iv doses of ZP, DF, CA, and VPA at 50, 50, 50, and 150 mg/kg (respectively) over 2 days. Selection of the dosing regimen followed from our earlier work which quantified tissue adducts of DF in rats given twice daily 50 mg/kg iv doses of DF for 7 days. The dosages of DF, CA, and VPA were selected after reference to earlier studies in bile-exteriorized rats given the same or similar doses of these drugs, which showed that biliary recovery of acyl glucuronide was ca. 40%, 65%, and 60%, respectively, and urinary recovery ca. 1%, 20%, and 15%, respectively (35, 36, 29). The dose of VPA selected for the present study was higher than that for the other drugs because of its more rapid metabolic clearance in rats (29). In the case of ZP, the original metabolic work had concluded that the acyl glucuronide was only a minor metabolite in rats, as it was recovered only in trace quantities in urine, though it was the major metabolite in humans (37, 38). However, we have recently shown2 that biliary excretion of ZP acyl glucuronide accounts for ∼25% of 50 mg/kg doses in bile-exteriorized rats. Its poor urinary recovery presumably reflects very limited sinusoidal elimination from rat hepatocytes plus high systemic instability, i.e., a similar dispositional scenario to that of DF acyl glucuronide (35). The apparent first-order degradation t1/2 values of the acyl glucuronides of ZP, DF, CA, and VPA in buffer at pH 7.4 and 37 °C are 0.5 (25), 0.6 (26), 3,2 and 60 (14) h respectively. Covalent binding to plasma protein has been documented from in vitro studies with the four glucuronides (8, 14, 26, 39), though for VPA glucuronide, there was a considerable lag time before covalent adducts became detectable. In the present in vivo study, the degree of protein modification found in plasma of rats was comparable for ZP, DF, and CA (Table 1). The low values found for VPA were in keeping with its acyl glucuronide being some 20-120 times less reactive (in vitro) than the other glucuronides. An earlier report (15) overviewing data from five different panels of human volunteers given five carboxylate drugs found a good correlation (r2 ) 0.873) between in vitro degradation rate constants of the glucuronides and maximum adduct concentrations in plasma, when the latter were normalized for exposure to glucuronide as measured by AUC in plasma. Plasma concentrations of the glucuronides were not monitored over the 4 dose/2 day regimen of the present study in intact rats. However, AUC values for

664 Chem. Res. Toxicol., Vol. 9, No. 3, 1996

the acyl glucuronides following single iv doses of the drugs to intact rats were available for DF and VPA from our earlier studies (35, 29), and were obtained for ZP and CA during the present study.3 When a linear regression analysis was performed on drug bound to plasma protein normalized for exposure by dividing by the AUC of the glucuronide (i.e., ratios of 0.0103, 0.00951, 0.00439, and 0.000933 h-1 for ZP, DF, CA, and VPA, respectively) versus the acyl glucuronide degradation rate constant (i.e., 1.39, 1.16, 0.23, and 0.012 h-1, respectively), a very good correlation (r2 ) 0.960) was obtained. Despite the limitations of using data from two different dosing regimens (four doses for the adduct data, one dose for the AUC data) and from using a small series of compounds, these results in rats linking glucuronide reactivity in vitro with plasma adducts in vivo are in agreement with the earlier observations in humans (15). Nonetheless, the in vivo situation is very complex, as noted earlier, and more work is required to better define other factors influencing plasma adduct concentrations. Unlike plasma, the extent of covalent binding to liver tissue was not grossly dissimilar among the four drugs (Table 1). Although ZP and DF again gave the higher covalent binding, there was no apparent relationship between total adduct concentrations and glucuronide reactivity, and no reliable measure of intrahepatic exposure to glucuronide could be obtained. CA and VPA produced the same adduct concentrations (on a w/w basis). Given the relative stability of VPA glucuronide, this suggested a contribution to VPA covalent binding from another pathway, i.e., one yielding a very reactive metabolite that interacts rapidly with proteins close to the site of activation (the liver), but which does not survive long enough to reach the systemic circulation. Such a pathway is probably β-oxidation, a well documented major route of VPA metabolism (24), which is initiated by formation of the CoA thioester of VPA. Such thioesters are reactive and can acylate hepatic proteins (40). The finding of adducts of (E)-2-en-VPA in liver (but not in plasma, Table 1) lends support to this suggestion, as the CoA thioester of (E)-2-en-VPA is an intermediate in the β-oxidation sequence converting VPA to 3-oxo-VPA. A contribution to covalent binding of (E)-2-en-VPA via its glucuronide conjugate cannot be denied, of course, but would be expected to be minor given the very limited formation of this glucuronide relative to that of VPA glucuronide (31, 41). Support for the existence of these mechanisms of adduct formation also comes from previous studies using radiolabeled VPA and isolated rat hepatocytes (42), which showed that inhibitors of both glucuronidation and β-oxidation decreased the amount of radiolabel covalently associated with hepatic proteins. Other mechanisms of covalent interaction of metabolites of VPA with protein have been suggested (42, 43), but CoA formation and acyl glucuronidation seem most likely to be the activating mechanisms responsible for covalent binding of VPA itself. Interestingly, adducts of 4-en-VPA, a minor VPA metabolite hypothesized to have a role in the pathogenesis of severe VPA hepatotoxicity (42-44), were not detected. β-Oxidation and acylation of hepatic proteins are well known for long chain fatty acids (40), and it is not surprising that VPA, a branched medium chain fatty acid (Figure 2), would undergo such reactions. In this context, it is possible that acylation of hepatic proteins subsequent to CoA formation may have made a 3

Bailey and Dickinson, unpublished results.

Bailey and Dickinson

contribution to the covalent binding observed with the other three drugs. Conversely, acyl glucuronidation may have made a contribution to, or been responsible for, covalent binding of acidic drugs previously attributed exclusively to CoA formation (40). The major bands observed in the livers of rats treated with the NSAIDs ZP (70, 110, 140, and ∼200 kDa) and DF (110, 140, and ∼200 kDa) are similar to those reported in livers of rats (50, 110, 140, and 200 kDa) (18) and mice (50, 70, 110, and 140 kDa) (45) treated with another NSAID, diclofenac. These workers did not isolate diclofenac acyl glucuronide, but used metabolic cofactors and inhibitors in vitro to conclude that P450 metabolism was responsible for the 50 kDa band (which they found in the microsomal fraction) and that glucuronidation accounted for the formation of the 110, 140, and 200 kDa proteins which were located in the plasma membrane. Recently, the 110 kDA band has been identified as the membrane enzyme dipeptidyl peptidase (46). Work on diclofenac from another laboratory found a somewhat different pattern of hepatic adducts (19, 47,,48). In cultured rat hepatocytes exposed to diclofenac, the major modified band detected radiochemically was a 60 kDa protein located in both microsomes and plasma membranes. Additional modified proteins at 50, 80, and 126 kDa were recognized using a diclofenac antiserum. In liver homogenates from diclofenac-treated rats, immunoblotting revealed protein adducts at 60 and 80 kDa. The reasons for the different patterns found in the different laboratories are unclear at present, but contributing factors could include differences in model systems (in vivo rat/cultured hepatocytes), samples (liver homogenate/subcellular fractions), and specificity of antibodies. Indeed, in the study of radiolabeled diclofenac in cultured rat hepatocytes (19), the radioactivity recovered in microsomes, cytosol, and the plasma membrane fraction accounted for only one quarter of the total protein-associated radioactivity in the hepatocytes. Such factors probably also contribute to differences between the patterns of DF-modified proteins found in liver homogenates in the present study and those found earlier in subcellular fractions (28). The patterns of protein modification detected in the livers of CA- and VPA-treated rats in the present study were different to that of the NSAIDs. The 70 kDa protein modified by CA treatment corresponded with one modified by ZP treatment, and the 140 kDa band in VPAtreated liver corresponded with a band found in the livers of the NSAID-treated animals. Several smaller molecular weight bands (∼40, 43, and 55 kDa) were modified in livers from VPA-treated animals. These proteins are possibly enzymes of the fatty acid β-oxidation complex, some of which are known to have similar molecular weights (49). In plasma, the pattern of modification was relatively simple, with serum albumin being the major protein modified by ZP, DF, and CA (although staining of the CA band was very weak). Modified serum albumin was not detected in plasma from VPA-treated rats, in agreement with the chemical analysis data showing low levels of adducts (Table 1). This reflects the relative stability of VPA glucuronide and affirms that its contribution to total hepatic VPA-protein adducts should be limited. In both plasma and liver, the ZP-KLH antiserum specifically recognized many proteins, whereas the total adduct concentrations of ZP as measured chemically were not greatly different from those of DF and CA (Table 1). This

Carboxylate Drug-Protein Adducts

could be due to several factors, for example, to a higher titer of antibodies (in the polyclonal antiserum raised against KLH chemically modified by ZP) specific for the type of adducts prevalent with in vivo modification by ZP, or to differential stability of adducts from different drugs in the SDS-PAGE/immunoblotting system. Similar effects may explain the very low intensity of the serum albumin band detected in plasma of CA-treated rats by immunoblotting (Figure 3), despite the chemical analysis data showing a substantial level of modification (Table 1). The location of adducts of foreign compounds in the plasma membrane of the liver would make them, like plasma adducts, accessible to the immune system, and therefore they are possible immunogens and targets for immune attack. Covalent modification of plasma membrane proteins could also disrupt normal transport processes into blood or bile. In this regard, the recent identification of the 110 kDa band in diclofenac-treated rat liver as dipeptidyl peptidase, and the observation that the activity of this enzyme in plasma membrane fractions was lowered after diclofenac treatment (46), exemplifies the possible consequences of adduct formation. In conclusion, this is the first study to compare, by both chemical and immunochemical means, adduct formation in both plasma and liver from a series of carboxylate drugs. It is also the first to identify (E)-2-en-VPA adducts in VPA-treated animals. While there are some similarities among the four drugs with respect to adduct formation, there are also differences as to which proteins are modified and to what extent. In plasma, extent of covalent adduct formation may be predictable given knowledge of the exposure of this tissue to an acyl glucuronide and of the intrinsic reactivity of the glucuronide. For liver, however, uncertainties regarding the extent of exposure to glucuronide and possible adduct formation from other mechanisms make prediction of the extent of adduct formation very problematic. Caution should therefore be exercised in attributing adduct formation exclusively to the acyl glucuronide pathway and in assuming that a single drug will typify the carboxylate class. However, the similarities between the patterns of modification found for the NSAIDs ZP, DF (this study), and diclofenac (18) suggest that these results might be more representative of the carboxylate class than those obtained here with CA or VPA.

Acknowledgment. We thank Drs. John de Jersey and Simon Worrall for many helpful discussions, Mr. Michael Franklin for the GC/MS work, and Ms. Adrienne Williams for raising the antisera. This work was supported by the National Health and Medical Research Council of Australia.

Chem. Res. Toxicol., Vol. 9, No. 3, 1996 665

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

(14)

(15)

(16)

(17)

(18)

(19)

(20)

(21)

(22)

References (23) (1) Faed, E. M. (1984) Properties of acyl glucuronides: Implications for studies of the pharmacokinetics and metabolism of acidic drugs. Drug Metab. Rev. 15, 1213-1249. (2) Spahn-Langguth, H., and Benet, L. Z. (1992) Acyl glucuronides revisited: Is the glucuronidation process a toxification as well as a detoxification mechanism? Drug Metab. Rev. 24, 5-47. (3) Dickinson, R. G. (1993) Acyl glucuronide conjugates: reactive metabolites of nonsteroidal anti-inflammatory drugs. Proc. Western Pharmacol. Soc. 36, 157-162. (4) Fenselau, C. (1994) Acyl glucuronides as chemically reactive intermediates. Handb. Exp. Pharmacol. 112, 367-389. (5) Ding, A., Ojingwa, J. C., McDonagh, A. F., Burlingame, A. L., and Benet, L. Z. (1993) Evidence for covalent binding of acyl glucuronides to serum albumin via an imine mechanism as

(24)

(25)

(26)

(27)

revealed by tandem mass spectrometry. Proc. Natl. Acad. Sci. U.S.A. 90, 3797-3801. Williams, A. M., and Dickinson, R. G. (1994) Studies on the reactivity of acyl glucuronidessVI. Modulation of reversible and covalent interaction of diflunisal acyl glucuronide and its isomers with human plasma protein in vitro. Biochem. Pharmacol. 47, 457-467. Ding, A., Zia-Amirhosseini, P., McDonagh, A. F., Burlingame, A. L., and Benet, L. Z. (1995) Reactivity of tolmetin glucuronide with human serum albumin: Identification of binding sites and mechanisms of reaction by tandem mass spectrometry. Drug Metab. Dispos. 23, 369-376. Smith, P. C., McDonagh, A. F., and Benet, L. Z. (1986) Irreversible binding of zomepirac to plasma protein in vitro and in vivo. J. Clin. Invest. 77, 934-939. Hyneck, M. L., Smith, P. C., Munafo, A., McDonagh, A. F., and Benet, L. Z. (1988) Disposition and irreversible plasma protein binding of tolmetin in humans. Clin. Pharmacol. Ther. 44, 107114. McKinnon, G. E., and Dickinson, R. G. (1989) Covalent binding of diflunisal and probenecid to plasma protein in humans: persistence of the adducts in the circulation. Res. Commun. Chem. Pathol. Pharmacol. 66, 339-354. Iwakawa, S., Spahn, H., Benet, L. Z., and Lin, E. T. (1990) Stereoselective binding of the glucuronide conjugates of carpofen enantiomers to human serum albumin. Biochem. Pharmacol. 39, 949-953. Volland, C., Sun, H., Dammeyer, J., and Benet, L. Z. (1991) Stereoselective degradation of the fenoprofen acyl glucuronide enantiomers and irreversible binding to plasma protein. Drug Metab. Dispos. 19, 1080-1086. Sallustio, B. C., Knights, K. M., Roberts, B. J., and Zacest, R. (1991) In vivo covalent binding of clofibric acid to human plasma proteins and rat liver proteins. Biochem. Pharmacol. 42, 14211425. Williams, A. M., de Jersey, J., Worrall, S., and Dickinson, R. G. (1992) Studies on the reactivity of acyl glucuronidessIII. Glucuronide-derived adducts of valproic acid and plasma protein and anti-adduct antibodies in humans. Biochem. Pharmacol. 43, 745755. Benet, L. Z., Spahn-Langguth, H., Iwakawa, S., Volland, C., Mizuma, T., Mayer, S., Mutschler, E., and Lin, E. T. (1993) Predictability of the covalent binding of acidic drugs in man. Life Sci. 53, 141-146. King, A. R., and Dickinson, R. G. (1993) Studies on the reactivity of acyl glucuronidessIV. Covalent binding of diflunisal to tissues of the rat. Biochem. Pharmacol. 45, 1043-1047. Dickinson, R. G., and King, A. R. (1993) Studies on the reactivity of acyl glucuronidessV. Glucuronide-derived covalent binding of diflunisal to bladder tissue of rats and its modulation by urinary pH and β-glucuronidase. Biochem. Pharmacol. 46, 1175-1182. Hargus, S. J., Amouzedeh, H. R., Pumford, N. R., Myers, T. G., McCoy, S. C., and Pohl, L. R. (1994) Metabolic activation and immunochemical localization of liver protein adducts of the nonsteroidal anti-inflammatory drug diclofenac. Chem. Res. Toxicol. 7, 575-582. Kretz-Rommel, A., and Boelsterli, U. A. (1994) Selective protein adducts to membrane proteins in cultured rat hepatocytes exposed to diclofenac: Radiochemical and immunochemical analysis. Mol. Pharmacol. 45, 237-244. Zia-Amirhosseini, P., Harris, R. Z., Brodsky, F. M., and Benet, L. Z. (1995) Hypersensitivity to nonsteroidal anti-inflammatory drugs. Nat. Med. 1, 2-4. Worrall, S., and Dickinson, R. G. (1995) Rat serum albumin modified by diflunisal acyl glucuronide is immunogenic in rats. Life Sci. 56, 1921-1930. Cook, D. J., Achong, M. R., and Murphy, F. R. (1988) Three cases of diflunisal hypersensitivity. Can. Med. Assoc. J. 138, 10291030. Zimmerman, H. J. (1990) Update of hepatotoxicity due to classes of drugs in common clinical use: nonsteroidal drugs, antiinflammatory drugs, antibiotics, antihypertensives, and cardiac and psychotropic agents. Semin. Liver Dis. 10, 322-338. Eadie, M. J., Hooper, W. D., and Dickinson, R. G. (1988) Valproate-associated hepatotoxicity and its biochemical mechanisms. Med. Toxicol. 3, 85-106. Smith, P. C., Hasegawa, J., Langendijk, P. N. J., and Benet, L. Z. (1985) Stability of acyl glucuronides in blood, plasma and urine: studies with zomepirac. Drug Metab. Dispos. 13, 110-112. Watt, J. A., and Dickinson, R. G. (1990) Reactivity of diflunisal acyl glucuronide in human and rat plasma and albumin solutions. Biochem. Pharmacol. 39, 1067-1075. Hayase, F., Nagaraj, R. H., Miyata, S., Njoroge, F. G., and Monnier, V. M. (1989) Aging of proteins: Immunological detection

666 Chem. Res. Toxicol., Vol. 9, No. 3, 1996

(28)

(29)

(30)

(31)

(32)

(33)

(34)

(35)

(36)

(37)

(38)

of a glucose-derived pyrrole formed during Maillard reaction in vivo. J. Biol. Chem. 264, 3758-3764. Williams, A. M., Worrall, S., De Jersey, J., and Dickinson, R. G. (1995) Studies on the reactivity of acyl glucuronidessVIII. Generation of an antiserum for the detection of diflunisal-modified proteins in diflunisal-dosed rats. Biochem. Pharmacol. 49, 209217. Dickinson, R. G., Harland, R. C., Ilias, A. M., Rodgers, R. M., Kaufman, S. N., Lynn, R. K., and Gerber, N. (1979) Disposition of valproic acid in the rat: dose-dependent metabolism, distribution, enterohepatic recirculation and choleretic effect. J. Pharmacol. Exp. Ther. 211, 583-595. Dickinson, R. G., and King, A. R. (1989) Reactivity considerations in the analysis of glucuronide and sulfate conjugates of diflunisal. Ther. Drug Monit. 11, 712-720. Dickinson, R. G., Hooper, W. D., Dunstan, P. R., and Eadie, M. J. (1989) Urinary excretion of valproate and some metabolites in chronically treated patients. Ther. Drug Monit. 11, 127-133. Tsang, V. C., Peralta, J. M., and Simons, A. R. (1983) Enzymelinked immunotransfer blot techniques (EITB) for studying the specificities of antigens and antibodies. Methods Enzymol. 92, 377-391. Hirom, P. C., Millburn, P., and Smith, R. L. (1976) Bile and urine as complementary pathways for the excretion of foreign organic compounds. Xenobiotica 6, 55-64. Klaassen, C. D., and Watkins, J. B. (1984) Mechanisms of bile formation, hepatic uptake and biliary excretion. Pharmacol. Rev. 36, 1-67. Dickinson, R. G., King, A. R., and Verbeeck, R. K. (1989) Elimination of diflunisal as its acyl glucuronide, phenolic glucuronide and sulfate conjugates in bile-exteriorized and intact rats. Clin. Exp. Pharmacol. Physiol. 16, 913-924. Emudianughe, T. S., Caldwell, J., Sinclair, K. A., and Smith, R. L. (1983) Species differences in the metabolic conjugation of clofibric acid and clofibrate in laboratory animals and man. Drug Metab. Dispos. 11, 97-102. Grindel, J. M., O’Neill, P. J., Yorgey, K. A., Schwartz, M. H., McKown, L. A., Migdalof, B. H., and Wu, W.-N. (1980) The metabolism of zomepirac sodium I. Disposition in laboratory animals and man. Drug Metab. Dispos. 8, 343-348. Wu, W. N., Weaner, L. E., Kalbron, J., O’Neill, P. J., and Grindel, J. M. (1980). The metabolism of zomepirac sodium II. Isolation and identification of the urinary metabolites in rat, mouse, rhesus monkey and man. Drug Metab. Dispos. 8, 349-352.

Bailey and Dickinson (39) Van Breeman, R. B., and Fenselau, C. (1985) Acylation of albumin by 1-O-acyl glucuronides. Drug Metab. Dispos. 13, 318-320. (40) Hertz, R., and Bar-Tana, J. (1988) The acylation of proteins by xenobiotic amphipathic carboxylic acids in cultured rat hepatocytes. Biochem. J. 254, 39-44. (41) Heinemeyer, G., Nau, H., Hildebrandt, A. G., and Roots, I. (1985) Oxidation and glucuronidation of valproic acid in male ratss Influence of phenobarbital, 3-methylchloranthrane, β-naptholflavone and clofibrate. Biochem. Pharmacol. 34, 133-139. (42) Porubek, D. J., Grillo, M. P., and Baillie, T. A. (1989) The covalent binding to protein of valproic acid and its hepatotoxic metabolite, 2-n-propyl-4-pentenoic acid, in rats and in isolated rat hepatocytes. Drug Metab. Dispos. 17, 123-130. (43) Kassahun, K., Hu, P., Grillo, M. P., Davis, M. R., Jin, L., and Baillie, T. A. (1994) Metabolic activation of unsaturated derivatives of valproic acid. Identification of novel glutathione adducts formed through coenzyme A-dependent and -independent processes. Chem.-Biol. Interact. 90, 253-275. (44) Tang, W., Borel, A. G., Fujimiya, T., and Abbott, F. S. (1995) Fluorinated analogues as mechanistic probes in valproic acid hepatotoxicity: hepatic microvesicular steatosis and glutathione status. Chem. Res. Toxicol. 8, 671-682. (45) Pumford, N. R., Myers, T. G., Davila, J. C., Highet, R. J., and Pohl, L. R. (1993) Immunochemical detection of liver protein adducts of the nonsteroidal anti-inflammatory drug diclofenac. Chem. Res. Toxicol. 6, 147-150. (46) Hargus, S. J., Martin, B. M., George, J. W., and Pohl, L. R. (1995) Covalent modification of rat liver dipeptidyl peptidase IV (CD26) by the nonsteroidal anti-inflammatory drug diclofenac. Chem. Res. Toxicol. 8, 993-996. (47) Kretz-Rommel, A., and Boelsterli, U. A. (1993) Diclofenac covalent protein binding is dependent on acyl glucuronide formation and is inversely related to P450-mediated acute cell injury in cultured rat hepatocytes. Toxicol. Appl. Pharmacol. 120, 155-161. (48) Kretz-Rommel, A., and Boelsterli, U. A. (1994) Mechanism of covalent adduct formation of diclofenac to rat hepatic microsomal proteins. Retention of the glucuronic acid moiety in the adduct. Drug Metab. Dispos. 22, 956-961. (49) Matsubara, Y., Kraus, J. P., Ozasa, H., Glassberg, R., Finocchiaro, G., Ikeda, Y., Mole, J., Rosenberg, L. E., and Tanaka, K. (1987) Molecular cloning and nucleotide sequence of cDNA encoding the entire precursor of rat liver medium chain acyl coenzyme A dehydrogenase. J. Biol. Chem. 262, 10104-10108.

TX960017O