Chem. Res. Toxicol. 2003, 16, 1359-1364
1359
Bioactivation of Clozapine by Murine Cardiac Tissue in Vivo and in Vitro Dominic P. Williams,*,†,‡ Charles J. L. O’Donnell,†,‡ James L. Maggs,‡ J. Steven Leeder,§ Jack Uetrecht,| Munir Pirmohamed,‡ and B. Kevin Park‡ Department of Pharmacology and Therapeutics, The University of Liverpool, Liverpool, United Kingdom, Department of Pediatrics and Pharmacology, University of Missouri-Kansas City, Kansas City, Missouri, and Faculty of Pharmacy, University of Toronto, 19 Russell Street, Toronto, Canada, M5S 2S2 Received February 28, 2003
Clozapine, an atypical neuroleptic, undergoes bioactivation to a chemically reactive nitrenium ion. This has been implicated in the pathogenesis of clozapine-induced agranulocytosis. Clozapine also causes myocarditis and cardiomyopathy, the mechanisms of which are unknown. To investigate this, we have evaluated whether clozapine undergoes bioactivation by murine cardiac tissue, in comparison to hepatic tissue. Mice were administered clozapine (5 and 50 mg/kg i.p.), and the extent of covalent binding was assessed by Western blotting. There was an increase in irreversible binding of clozapine to several proteins, ranging in mass from 30 to 250 kDa in both hepatic and cardiac tissue. Bioactivation by hepatic and cardiac microsomes was assessed by LC/MS using glutathione to trap the intermediate. Metabolism of radiolabeled clozapine to a glutathionyl conjugate by liver and cardiac microsomes was 30.5 ( 3.3 and 3.6 ( 0.3% of the initial incubation concentration, respectively. Ketoconazole (20 µM), a P450 inhibitor, significantly reduced binding in both hepatic and cardiac microsomes to 6.2 ( 0.2 and 0.5 ( 0.06%, respectively. These data indicate that clozapine undergoes bioactivation in the heart to a chemically reactive nitrenium metabolite that may be important in the pathogenesis of myocarditis and cardiomyopathy observed in man.
Introduction Clozapine (Clozaril; CLZ) is an atypical and a highly effective antipsychotic used in the treatment of refractory schizophrenia (1). Despite its efficacy, its wider use has been restricted because of its propensity to cause agranulocytosis in 0.8% of patients (2). CLZ also causes numerous other adverse drug reactions, including neuroleptic malignant syndrome (3, 4), myocarditis, cardiomyopathy, hepatotoxicity, and nephritis (4-10). The mechanisms of these serious CLZ-induced adverse reactions are unclear, although there does not appear to be a relationship to the occurrence of agranulocytosis. There has been recent concern from drug regulatory agencies about the potential of CLZ to cause cardiac toxicity. The FDA has reported that between 1989 and 1999, there were 28 cases of myocarditis (18 fatal) and 41 cases (10 fatal) of cardiomyopathy associated with CLZ use in the U.S. (11). The risk of myocarditis in the first month of therapy with CLZ was estimated to be 80-fold higher than the background rate. This has prompted strengthening of warnings on the CLZ product information. Similar data have also been reported from Australia (7); two clinical patterns seem to emerge from the currently available data. A proportion of patients develop * To whom correspondence should be addressed. Tel: +44(0)151 794 5791. Fax: +44(0)151 794 5540. E-mail:
[email protected]. † These authors contributed equally in the production of this manuscript. ‡ The University of Liverpool. § University of Missouri-Kansas City. | University of Toronto.
myocarditis in the first month of therapy, while another group who have been taking CLZ for longer periods present with cardiomyopathy. The risk of myocarditis and cardiomyopathy in CLZ users is 2000- and 5-fold higher, respectively, than that seen in the general population (7). The relationship between the two clinical presentations is unclear. Initiation of CLZ therapy results in an asymptomatic increase in liver transaminases in 37% of patients; the liver function normalizes in 60% within the first 13 weeks of treatment (5). However, more severe liver injury has also been reported; for example, fulminant hepatic failure is seen in 0.06% of patients (6). CLZ is extensively metabolized in vivo, with the demethyl and N-oxide being the major stable metabolites (12). CLZ also undergoes bioactivation in vivo, in hepatic microsomal incubations, and in peripheral blood neutrophils and bone marrow cells to form a nitrenium ion (1214). This has been elucidated mainly from GSH trapping studies through the formation of glutathionyl-CLZ conjugates. CLZ-modified peptides have also been identified in neutrophils taken from patients on CLZ (15), indicating that bioactivation is occurring in vivo in man. Bioactivation to the nitrenium ion has been implicated as a mechanism for CLZ agranuocytosis, possibly through the initiation of neutrophil apoptosis (16). Bioactivation of CLZ can be catalyzed by P450 enzymes, myeloperoxidase, and horseradish peroxidase (13, 14). Given that the liver contains high concentrations of P450 enzymes, and the recent discovery that the heart also expresses several P450 isoforms (17). It is possibile that bioactivation
10.1021/tx034035z CCC: $25.00 © 2003 American Chemical Society Published on Web 09/17/2003
1360 Chem. Res. Toxicol., Vol. 16, No. 10, 2003
within these organs to the nitrenium ion may be important in the pathogenesis of CLZ-induced hepatotoxicity and cardiotoxicity. Such a mechanism is clearly wellestablished for liver damage observed with other therapeutically used drugs (18) but has to date not been investigated with respect to cardiotoxicity, with the exception of anthracyclines (19). The purpose of this study therefore was to investigate whether murine heart, in comparison to the liver, was able to bioactivate CLZ. Specifically, we have used a Western blotting method to detect CLZ-protein adducts in the liver and heart of mice administered CLZ. Additionally, microsomes prepared from these tissues were assessed for their ability to form a glutathionyl-CLZ conjugate in vitro.
Materials and Methods Materials. CLZ (8-chloro-11-(4-methyl-1-piperazinyl)-5Hdibenzo[b,e][1,4]diazepine) and [14C]CLZ (164 µCi/mg, radiochemical purity 98% by HPLC) were a kind gift from Novartis Pharmaceuticals (Basle, Switzerland); anti-CLZ-NAC-KLH antiserum was generated as previously described (15). Tris HCL, Tris Base, glycine, SDS, NaCl, KCl, glycerol, dimethyl sulfoxide, potassium dihydrogen orthophosphate, dithiothreitol, Triton-X 100, EDTA, Tween 20, Ponceau stain, nitrocellulose membrane, ammonium persulfate, Bromophenol blue, and Hanks’ Balanced Salt Solution were all obtained from Sigma Chemical Co. (Poole, U.K.). Peroxidase-conjugated anti-rabbit antibody, an ECL Western blotting detection system, and prestained molecular mass markers were all obtained from Amersham (Bucks, U.K.). Dosing of Animals and Sample Preparation. The protocols described were undertaken in accordance with a license granted under the Animals (Scientific Procedures) Act (1986) and approved by the University of Liverpool Animal Ethics Committee. Male CD-1 mice (30-40 g; Charles River, U.K.) were administered CLZ (5 [therapeutic] or 50 mg/kg [supratherapeutic], i.p.) or vehicle control (PEG/saline; 75/25%; 50-100 µL). The heart and liver were removed 1 and 24 h after dosing and frozen at -80 °C. A portion (ca. 100 mg) of each organ was homogenized and centifuged at 9000g for 20 min to obtain the S9 fraction. The microsomal fraction was prepared and assessed for P450 and protein content, as previously described (13). SDS-Page and Immunoblotting of Mouse Homogenates. CLZ-modified proteins were detected by Western blot analysis with a polyclonal anti-CLZ-NAC-KLH antibody using a modification of the methods reported previously (15). Briefly, proteins (10 µg total protein/lane) were separated by electrophoresis on SDS-PAGE gels using 4 and 10% acrylamide as stacking and resolving gels, respectively. After they were boiled, samples were subjected to electrophoresis using a MiniPROTEAN 3 Cell (Biorad) at 170 V. After electrophoresis, proteins were transferred (1 h at 30 V) to a nitrocellulose membrane (0.45 µm) with a Mini Trans-Blot cell (Biorad) and stained with Ponceau stain to assess the efficiency of the transfer. The Ponceau-stained blot was scanned and subjected to whole lane densitometry. Blots were repeated if the arbitrary net density of a lane differed by more than 10% from the mean net density of all of the lanes. The membrane was washed with tris-buffered saline (pH 7.4) containing 0.1% (v/v) Tween 20 (0.1% T-TBS). The nitrocellulose membrane was then blocked for 2 h at room temperature (10% (w/v) powdered skimmed milk in 0.1% T-TBS), washed, and then incubated with the anti-CLZ antibody (1:3000 dilution) for 15 h at room temperature. Membranes were washed three times with 0.1% T-TBS and then incubated with HRP-conjugated donkey anti-rabbit IgG (diluted 1:5000). The membranes were washed, and CLZ-modified proteins were visualized using enhanced chemiluminescence detection (Amersham, Buckinghamshire, U.K.). Band and lane volumes were quantified by UVISoft software (UVITech, U.K.). Arbitrary net
Williams et al. density was calculated by comparing the lane densities of CLZdosed mice to saline-dosed mice. In experiments to determine the specificity of the anti-CLZ antibody, the antibody (1:3000 final dilution) was incubated with CLZ (10 µM) for 30 min before the addition of the primary antiserum to the nitrocellulose membrane. The washing and subsequent development of the immunoblots were performed as outlined above. Hepatic and Cardiac Microsomal Incubations. Unlabeled CLZ (200 µM; for LC/MS analysis) or [14C]CLZ (2 µM, 0.1 µCi; for radiometric analysis) was incubated with hepatic and cardiac microsomes (1 mg of protein) as described previously (13, 20) for 1 h at 37 °C; the reaction was initiated by the addition of NADPH (1 mM). Some incubations also contained GSH (1 mM) or ketoconazole (20 µM). Some incubations were performed with boiled microsomes (10 min at 95 °C). The reactions were terminated using ice-cold methanol (2 mL), the protein was precipitated, and the supernatant was evaporated to dryness. Incubation samples containing [14C]CLZ were reconstituted in water-methanol and analyzed by radiometric HPLC. Incubation samples containing unlabeled CLZ (200 µM) were reconstituted in distilled water, repeatedly extracted with ether (to remove nonpolar metabolites and excess parent compound), evaporated to dryness, and finally reconstituted in water-methanol (200 µL, 1:1, v/v) prior to analysis by LC/MS. HPLC and LC/MS Analysis of GSH Conjugates. The conditions for HPLC and LC/MS have been described previously (12). Briefly, samples (50 µL) were eluted from a 5 µm Columbus C8 column (25 cm × 0.32 cm; Phenomonex, Macclesfield, Cheshire) with a gradient of acetonitrile in 6 mM ammonium formate, pH 3.5: 10-25% over 15 min and 25-55% over 20 min. The flow rate was 0.9 mL/min. The mobile phase was delivered by two Jasco PU980 pumps (Jasco Corporation, Great Dunmow, Essex, U.K.) via a HG-980-30 mixing module. Statistical Analysis. Results are presented as means ( SD. Statistical analysis was performed by the unpaired t-test after log transformation of the densitometric data and testing for normal distribution of the metabolic data. A p < 0.05 was accepted as being significant.
Results Immunoblotting of Homogenates Prepared from Mice Administered CLZ. There was a dose-dependent increase in the irreversible binding of CLZ to the S9 fraction prepared from liver at both 1 and 24 h (Figure 1). This was identified as a significant increase in the mean arbitrary net density from vehicle-treated mice as compared with CLZ-administered mice (n ) 4). The protein binding seen with high dose CLZ produced bands with molecular masses ranging from 30 to >160 kDa. At a lower dose, three main bands with approximate masses of 50, 105, and 180 kDa were visualized. With cardiac tissue, there was an increase in CLZ protein conjugate formation (Figure 2), which decreased with time. There were numerous bands in cardiac tissue. Hapten inhibition experiments demonstrated the specificity of binding of the anti-CLZ antibody (Figures 1C,D and 2C,D). In each case, 10 µg of total protein was loaded onto each lane. Metabolism and Bioactivation of CLZ by Microsomes from Liver and Heart. The criteria used for identification of glutathionyl-CLZ conjugates were defined in our previous publications, namely, coelution with positive control standards and mass spectrometry identification of m/z values (13-15). After incubation of CLZ with liver microsomes, selective ion monitoring at m/z 632, the [M + 1]+ ion for glutathionyl-CLZ conjugates, revealed the presence of a glutathionyl CLZ metabolite
Protein Modification by Clozapine in Vivo
Figure 1. Western blot (B) detection of CLZ-modified hepatic proteins from mice dosed 5 or 50 mg/kg CLZ and sacrificed at either 1 or 24 h. Ten micrograms of total protein was loaded per lane. The graph (A) shows the results of densitometric analysis from four animals per group; *p < 0.05. Representative Western blot showing detection of drug-modified polypeptides is below the graph. There was 10 µg of total protein loaded onto each lane. Western blot of hepatic protein from mice dosed CLZ. In each case, 10 µg of total protein was loaded per lane. (C) Detection of drug-modified polypeptides in hepatic tissue from mice dosed 5 or 50 mg/kg CLZ and sacrificed at 24 h. (D) Shows an identical experiment to panel C except that the primary antiserum was incubated with CLZ (10 µM) for 30 min before the addition to the nitrocellulose membrane.
(Rt, 18 min; Figures 3A and 4A,B). The peaks were absent in incubations that lacked GSH (data not shown) and in the incubations containing boiled microsomes. Coincubation with ketoconazole decreased the formation of the glutathionyl conjugate from 31 to 6% (p < 0.0001) (Table 1). The formation of demethyl CLZ, which is dependent upon CYP1A2 (13) in human microsomes, was not affected by coincubation with ketoconazole (a CYP inhibitor that is not specific at this concentration), which is in accordance with our previous data with human tissue (13). There was an additional peak observed, particularly in the incubations containing ketoconazole, but the identity of this peak could not be ascertained. With cardiac microsomes, LC/MS (m/z 632; Figure 3B) and radiometric (Figure 5A) analyses again demonstrated the formation of the glutathionyl-CLZ conjugate, although this was almost 30-fold less than that seen with hepatic microsomes (p < 0.0001). Coincubation with ketoconazole significantly (p < 0.0001) reduced the formation of the CLZ-GSH conjugate formed by cardiac microsomes (Figures 3B and 5B; Table 1). The amount of conjugate formed was also largely abolished when boiled microsomes were used (p < 0.0001) (Figures 3B and 5C).
Discussion Bioactivation of CLZ to a reactive nitrenium ion has been implicated in the pathogenesis of CLZ-induced agranulocytosis (21, 22). This reactive intermediate can be formed by human neutrophils and bone marrow
Chem. Res. Toxicol., Vol. 16, No. 10, 2003 1361
Figure 2. Western blot (B) detection of CLZ-modified cardiac proteins from mice dosed 5 or 50 mg/kg CLZ and sacrificed at 1 or 24 h afterward. Ten micrograms of total protein was loaded per lane. The graph (A) shows the results of densitometric analysis from four animals per group; *p < 0.05. (C) Detection of drug-modified polypeptides in cardiac tissue from mice dosed 5 or 50 mg/kg CLZ and sacrificed at 24 h. (D) Shows an identical experiment to panel C except that the primary antiserum was incubated with CLZ (10 µM) for 30 min before the addition to the nitrocellulose membrane. Above the Western blots for C and D, Ponceau blots have been included to ensure equal protein loading. There was 10 µg of total protein loaded onto each lane.
precursors in vitro (12) and can lead to neutrophil cell death (14), by both necrotic and apoptotic processes (16). Furthermore, the identification of CLZ-haptenated proteins in neutrophils in patients on CLZ therapy (15) is consistent with the bioactivation of CLZ in vivo. The mechanism by which CLZ bioactivation leads to agranulocytosis is unknown, but various mechanisms including apoptosis, hypersensitivity, and genetic susceptibility have been postulated (22, 23). Our previous studies (12-14, 22) detailed the metabolism and bioactivation of CLZ in human, rat, and mouse hepatic microsomes and in human neutrophils. The reactive metabolite of CLZ could be detoxified by extensive GSH conjugate formation. In the present study, we provide evidence that CLZ bioactivation occurs in vivo in a murine model through demonstration of covalent binding to cardiac and liver proteins. Interestingly, tissue distribution studies in mice administered 15 µmol/kg of [14C]CLZ i.v. have shown that 5.5 and 3.2% of the dose remains in the liver and kidneys, respectively, while less than 3% of the dose remains in the other organs, including the brain and heart (12). In accordance with our previous findings (12, 13), mouse liver microsomes were able to bioactivate CLZ to reactive intermediates that could be trapped as GSH conjugates. Bioactivation was inhibited by ketoconazole indicating that it was P450 mediated (13). In this study, we have extended our observations by using an anti-CLZ antibody, which demonstrates that the reactive metabolite of CLZ binds to several proteins in murine liver. CLZ can cause myocarditis and cardiomyopathy: our data show that CLZ is bioactivated by microsomes
1362 Chem. Res. Toxicol., Vol. 16, No. 10, 2003
Williams et al.
Figure 3. LC/MS ion current (m/z 632) chromatograms of GSH conjugate of CLZ formed by the incubation of CLZ (200 µM) with GSH (1 mM) in the presence of either (A) hepatic or (B) cardiac microsomes (1 mg). Overlaid on each chromatogram is the trace obtained from incubations using boiled microsomes (95 °C, 10 min).
prepared from mouse heart leading to the formation of a reactive intermediate that can bind to GSH. Furthermore, a single dose of CLZ administered in vivo led to binding of the reactive metabolite to several cardiac proteins, the nature of which needs to be elucidated. Our data suggest that CLZ bioactivation is mediated by P450 enzymes since the formation of the GSH conjugate was reduced by heat inactivation of the microsomes and by the use of the P450 inhibitor ketoconazole. The concentration of ketoconazole was chosen from our previous investigations that demonstrated ketoconazole had an IC50 of 9.7 µM for the inhibition of protein reactive metabolite formation in human liver microsomes (13). Additionally, ketoconazole has an IC50 of 42 µM for the inhibition of demethyl CLZ formation, indicating that the concentration used in the current study was not so high as to be completely nonspecific (13). There is evidence that ketoconazole can inhibit peroxidase enzymes, but the IC50 (200 µM) is 10 times greater than the concentration used in this study (24). The recent finding that the heart expresses various P450 isoforms both in man and
in mice (17, 25, 26) supports the hypothesis that cardiac P450 enzymes are involved in CLZ bioactivation. How does the finding that the heart can bioactivate CLZ relate to the clinical features of CLZ-mediated cardiac muscle damage? There are two clinical syndromes, one of which is acute presenting as myocarditis (7), with eosinophilic infiltration. The other is more chronic presenting as cardiomyopathy. The former may represent an immune-mediated reaction to CLZ binding of cardiac proteins, while the latter may, at least partly, be due to apoptosis of cardiomyocytes, which is known to be important in the pathogenesis of patients with severe heart failure of various aetiologies (27). It is clear from in vitro studies that neutrophils that become covalently bound by CLZ subsequently die by apoptosis (16). However, the potential contribution of P450 expression and the relative content of GSH within the various cardiac regions will play important roles in determining the particular cell type to which CLZ binds, and more importantly, the fate of that cell.
Protein Modification by Clozapine in Vivo
Chem. Res. Toxicol., Vol. 16, No. 10, 2003 1363
Figure 4. Radiomatic chromatograms of GSH conjugates of CLZ formed by the incubation of (A) [14C]CLZ (2 µM; 0.1 µCi) with GSH (1 mM) and hepatic microsomes (1 mg). (B) Incubation of [14C]CLZ (2 µM; 0.1 µCi) with GSH (1 mM), ketoconazole (20 µM), and hepatic microsomes (1 mg). (C) Incubation of [14C]CLZ (2 µM; 0.1 µCi) with GSH (1 mM) and boiled hepatic microsomes (1 mg). Table 1. Formation of Glutathionyl CLZ (CLZ-GSH) Conjugate by Liver and Cardiac Microsomesa % conversion to CLZ-GSH conjugate (mean ( SD) liver heart
+NADPH +NADPH + boiled microsomes +NADPH + ketoconazole (20 µM) +NADPH +NADPH + boiled microsomes +NADPH + ketoconazole (20 µM)
30.5 ( 3.3 0.1 ( 0.04* 6.2 ( 0.2* 3.6 ( 0.3 0.3 ( 0.04* 0.5 ( 0.06*
a Data show incubations carried out in the presence of [14C]CLZ (2 µM; 0.1 µCi), NADPH (1 mM), and GSH (1 mM). One milligram of microsomal protein was used for each incubation. Statistical analysis was performed on three experiments by unpaired t-test, comparing incubations within the same organ; *p < 0.0001.
Covalent binding of CLZ was evident 1 h after administration of a single dose but was not present in the later specimens taken at 24 h. This presumably reflects clearance of the damaged protein, for example, by ubiquitination and subsequent proteolysis (28). Mutations in the protein clearance system can increase the sensitivity of mammalian cells to protein damaging agents (29). Whether this may act as a susceptibility factor for CLZmediated idiosyncratic toxicity will need further inves-
Figure 5. Radiomatic chromatograms of GSH conjugates of CLZ formed by the incubation of (A) [14C]CLZ (2 µM; 0.1 µCi) with GSH (1 mM) and cardiac microsomes (1 mg). (B) Incubation of [14C]CLZ (2 µM; 0.1 µCi) with GSH (1 mM), ketoconazole (20 µM), and cardiac microsomes (1 mg). (C) Incubation of [14C]CLZ (2 µM; 0.1 µCi) with GSH (1 mM) and boiled cardiac microsomes (1 mg).
tigation. However, it is also important to note that we have administered single doses of CLZ, while clinically CLZ is used chronically, often for many years. Thus, the consequences of chronic dosing need further investigation. Our results demonstrate that CLZ readily undergoes oxidation to a reactive metabolite in murine heart both in vivo and in vitro. This has been compared to the bioactivation of CLZ in murine liver (Figure 6). CLZ has been associated with idiosyncratic toxicity in both the liver and the heart in man. However, there is very little
1364 Chem. Res. Toxicol., Vol. 16, No. 10, 2003
Figure 6. Scheme illustrating the bioactivation, protein binding, and GSH adduct formation by CLZ within the murine heart and liver.
evidence actually linking the formation of these adducts to observable toxicity. The formation of drug-protein adducts provides evidence of a chemical alert for potential toxicity. However, the relationship between covalent binding of CLZ, and indeed other drugs, and the specific functional consequences, i.e., whether it leads to clinically symptomatic organ damage, need further investigation.
Acknowledgment. This work was funded by The Wellcome Trust, The Medical Research Council, and Pfizer Ltd.
References (1) Baldessarini, R. J., and Frankenburg, F. R. (1991) N. Engl. J. Med. 324, 746-754. (2) Atkin, K., Kendall, F., Gould, D., Freeman, H., Liberman, J., and O’Sullivan, D. (1996) Br. J. Psychiatry 169, 483-488. (3) Caroff, N. C., Mann, S. C., and Campbel, E. C. (2001) Adverse Drug React. Bull. 209, 799-802.
Williams et al. (4) Brown, T. M. (1999) Psychosomatics 40, 518-520. (5) Hummer, M., Kurz, M., Kurzthaler, I., Oberbauer, H., Miller, C., and Fleischhacker, W. W. (1997) J. Clin. Psychopharmacol. 17, 314-317. (6) Macfarlane, B., Davies, S., Mannan, K., Sarsam, R., Pariente, D., and Dooley, J. (1997) Gastroenterology 112, 1707-1709. (7) Killian, J. G., Kerr, K., Lawrence, C., and Celermajer, D. S. (1999) Lancet 354, 1841-1845. (8) Elias, T. J., Bannister, K. M., Clarkson, A. R., Faull, D., and Faull, R. J. (1999) Lancet 354, 1180-1181. (9) Fraser, D., and Jibani, M. (2000) Clin. Nephrol. 54, 78-80. (10) Southall, K. E. (2000) Aust. N. Z. J. Psychiatry 34, 697-698. (11) La Grenade, L., Graham, D., and Trontell, A. (2001) N. Engl. J. Med. 345, 224-225. (12) Maggs, J. L., Williams, D., Pirmohamed, M., and Park, B. K. (1995) J. Pharmacol. Exp. Ther. 275, 1463-1475. (13) Pirmohamed, M., Williams, D., Madden, S., Templeton, E., and Park, B. K. (1995) J. Pharmacol. Exp. Ther. 272, 984-990. (14) Williams, D. P., Pirmohamed, M., Naisbitt, D. J., Maggs, J. L., and Park, B. K. (1997) J. Pharmacol. Exp. Ther. 283, 1375-1382. (15) Gardner, I., Leeder, J. S., Chin, T., Zahid, N., and Uetrecht, J. P. (1998) Mol. Pharmacol. 53, 999-1008. (16) Williams, D. P., Pirmohamed, M., Naisbitt, D. J., Uetrecht, J. P., and Park, B. K. (2000) Mol. Pharmacol. 58, 207-216. (17) Thum, T., and Borlak, J. (2000) Lancet 355, 979-983. (18) Park, B. K., Pirmohamed, M., and Kitteringham, N. R. (1995) Pharmacol. Ther. 68, 385-424. (19) Minotti, G., Licata, S., Saponiero, A., Menna, P., Calafiore, A. M., Di Giammarco, G., Liberi, G., Animati, F., Cipollone, A., Manzini, S., and Maggi, C. A. (2000) Chem. Res. Toxicol. 13, 1336-1341. (20) Pirmohamed, M., Williams, D., Madden, S., Templeton, E., and Park, B. K. (1995) J. Pharmacol. Exp. Ther. 272, 984-990. (21) Uetrecht, J. P. (1992) Drug Metab. Rev. 24, 299-366. (22) Pirmohamed, M., and Park, K. (1997) CNS Drugs 7, 139-158. (23) Guest, I., Sokoluk, B., MacCrimmon, J., and Uetrecht, J. (1998) Toxicology 131, 53-65. (24) Comby, F., Lagorce, J. F., Buxeraud, J., and Raby, C. (1994) J. Pharm. Pharmacol. 46, 50-53. (25) Ding, Z., Godecke, A., and Schrader, J. (2002) Br. J. Pharmacol. 135, 631-638. (26) Wang, J. F., Yang, Y., Sullivan, M. F., Min, J., Cai, J., Zeldin, D. C., Xiao, Y. F., and Morgan, J. P. (2002) Exp. Biol. Med. (Maywood) 227, 182-188. (27) Feuerstein, G. Z., and Young, P. R. (2000) Cardiovasc. Res. 45, 560-569. (28) Wilkinson, K. D. (1995) Annu. Rev. Nutr. 15, 161-189. (29) Tsirigotis, M., Thurig, S., Dube, M., Vanderhyden, B. C., Zhang, M., and Gray, D. A. (2001) Biotechniques 31, 120-126, 128, 130.
TX034035Z
