DECEMBER 1995 VOLUME 8, NUMBER 8 0Copyright 1995 by the American Chemical Society
Communications Covalent Modification of Rat Liver Dipeptidyl Peptidase IV (CD26) by the Nonsteroidal Anti-Inflammatory Drug Diclofenac Sally J. Hargus,*s*sfBrian M. Martin,§ John W. George,' and Lance R. Pohlt Molecular and Cellular Toxicology Section, Laboratory of Molecular Immunology, National Heart, Lung and Blood Institute, NIH, Bethesda, Maryland 20892, Department of Anesthesia, Georgetown University, Washington, DC, and Clinical Neurosciences Branch, National Institute of Mental Health, NIH, Bethesda, MD Received July 19, 1995@ Diclofenac is a nonsteroidal anti-inflammatory drug that has been implicated in several cases of severe hepatotoxicity. Our previous study showed that diclofenac metabolites bound covalently and selectively to rat liver plasma membrane proteins with estimated monomeric masses of 110, 140, and 200 kDa. We report here that we have identified the 110 kDa diclofenac-labeled protein in rat liver as dipeptidyl peptidase IV,also known as CD26. In addition, we found that the activity of dipeptidyl peptidase IV in liver plasma membrane fractions was lowered after diclofenac treatment of rats. These results suggest that the hepatotoxicity associated with diclofenac might be due, in part, to the covalent modification of dipeptidyl peptidase IV.
Introduction Diclofenac is a nonsteroidal anti-inflammatory drug
(NSAID)l that has been associated with severe idiosyncratic hepatotoxicity (1-13). Although the mechanism of diclofenac toxicity is unknown, it is thought that * Author to whom correspondence should be addressed at the NIH, Building 10, Room 8N104, Bethesda, MD 20892-1760. Phone: 301496-4841; FAX: 301-480-4852. ' NHLBI-NIH. + Georgetown University. § NIMH-NIH. Abstract published in Advance ACS Abstracts, October 15, 1995. 'Abbreviations: NSAID, nonsteroidal anti-inflammatory drug; DPP IV, dipeptidyl peptidase IV, SDS-PAGE, sodium dodecyl sulfatepolyacrylamide gel electrophoresis; PBS, phosphate-buffered saline; PMSF, phenylmethanesulfonyl fluoride; OG, n-octyl P-D-glUCOpyranOside; PVDF, poly(viny1idene difluoride).
covalently modified proteins may be important in causing the toxicity either directly or by eliciting an immune response (14). In this regard, our previous studies (15, 16) and those of others (17,181 have shown that metabolites of diclofenac covalently modify mouse and rat liver proteins. The diclofenac-labeledproteins detected in our studies with polyclonal sera raised against diclofenacmodified keyhole limpet hemacyanin were found predominantly in plasma membrane fractions and had estimated monomeric masses of 110, 140, and 200 kDa. We report here that we have identified the 110 kDa diclofenac-labeled protein as dipeptidyl peptidase IV (DPP IV, EC 3.4.14.5). In addition, diclofenac adduct formation appears to inhibit the enzymatic activity of DPP IV, suggesting that the adduct may have a direct role in hepatotoxicity induced by diclofenac.
0893-228x/95/2708-0993$09.00/00 1995 American Chemical Society
Communications
994 Chem. Res. Toxicol., Vol. 8, No. 8, 1995
Experimental Section Materials were obtained from the following commercial sources: monoclonal anti-phosphotyrosine (clone PY 20) from ICN (Aurora, OH); n-octyl P-D-glucopyranoside (OG) and hydrogenated Triton X-100 from Calbiochem (LaJolla, CA); wheat germ agglutinin-Sepharose 6MB from Pharmacia (Piscataway, NJ); N-acetylglucosamine from Schweizerhall (South Plainfield, NJ); Microcon 30 from Amicon (Beverly, MA); nitrocellulose membranes from Bio-Rad; enhanced chemiluminescence substrates from Amersham (Arlington Heights, IL); X-OMAT AR film from Kodak (Rochester, NY);BCA protein assay reagent kit from Pierce Chemical (Rockford, IL); Blotto from Advanced Biotechnologies Inc. (Columbia, MD); Immobilon poly(viny1idene difluoride) (PVDF) transfer membranes from Millipore (Bedford, MA); monoclonal mouse anti-rat DPP IV (clone 236.3) from Endogen (Boston, MA); goat anti-rabbit IgG and goat anti-mouse IgG (peroxidase conjugates) and protein A-agarose from Boehringer Mannheim (Indianapolis, IN); diclofenac (sodium salt), Gly-Pro-p-nitroanilide, and p-nitroaniline from Sigma (Indianapolis, IN). Diclofenac antisera was prepared a s described previously (15). SDS-PAGE and Immunoblotting. Proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), stained with Coomassie brilliant blue, or were transferred to a nitrocellulose membrane and immunoblotted with diclofenac antiserum according to published procedures (16 ) . Immunoblotting with DPP IV monoclonal antibody was done in a similar manner, except that DPP IV monoclonal antibody was diluted 1:100, and goat anti-mouse IgG (peroxidase conjugate) was used as the second antibody. Purification of the 110 kDa Protein. Twelve male Sprague-Dawley rats (175-200 g, Taconic Farms, Germantown, NY) were given intraperitoneal injections of saline (2 mL) or diclofenac (200 mgkg, from a 20 mg/mL stock solution in water) and were killed by decapitation 16 h later. Plasma membrane fractions (400 mg total protein, 4.7 mg/mL) from saline-treated controls or diclofenac-treated rats were isolated as described previously (19). Plasma membranes were extracted with phosphate-buffered saline (PBS, pH 7.4) containing 1 mM phenylmethanesulfonyl fluoride (PMSF), 1mM EDTA, 1pM leupeptin, 1pM pepstatin, and 25 mM OG for 3 h at 4 "C, with stirring. After centrifugation at lOOOOOg for 1.5 h, the resulting pellet was extracted with an equal volume of buffer A (PBS containing 1mM PMSF, 1mM EDTA, 1pM leupeptin, 1pM pepstatin and 50 mM OG) for 16 h at 4 "C. The mixture was centrifuged at 1OOOOOg for 1.5 h, and the supernatant (107 mg of protein) was loaded onto a chromatography column (1 x 25 cm) of wheat germ agglutinin-Sepharose 6MB (10 mL) that was equilibrated with buffer A. The column was washed with 5 volumes of buffer A, and bound proteins were eluted with buffer A containing 0.5 M N-acetylglucosamine. Fractions were collected, analyzed by SDS-PAGE, and immunoblotted with diclofenac antisera. Amino Acid Sequence Analyses. Wheat germ agglutininpurified fractions containing the 110 kDa protein were concentrated (Microcon 30), and proteins were separated by SDSPAGE in order to further purify the 110 kDa protein and transferred to Immobilon PVDF (20). Approximately 20 pg of the 110 kDa protein was used for the N-terminal amino acid sequence analysis by automated Edman degradation (20). To obtain internal peptides of the protein, approximately 40 p g of the 110 kDa protein on PVDF (2 lanes of 20 pg/lane) was reduced in 100 mM Tris (pH 8.6) containing 6 M guanidine hydrochloride and 10 mM dithiothreitol at 55 "C for 90 min and then was alkylated in 1 M Tris (pH 7.5) containing 67 mM iodoacetic acid at room temperature for 30 min (20). The membrane was washed successively with water (1mL, twice), methanol (1 mL), and then water (1 mL, 3 times). The membrane was minced with a clean razor blade, transferred to a 1-mL microfuge tube, and then incubated in 100 pL of digestion buffer [lo0 mM Tris (pH 8) containing 10% (v/v) acetonitrile, 1% (v/v) hydrogenated Triton X-100, and 2 p g of trypsin] overnight at 37 "C (21). The digested protein fragments were separated by reversed-phase HPLC using a Reliasil C4
A
B kDa
200 140
110
L
1 2 3 4
I
I
1 2 3 4
Figure 1. Purification of 110 kDa diclofenac-labeled protein from liver plasma membranes of diclofenac-treated rats. (A) Anti-diclofenac immunoblot of detergent extracts of rat liver plasma membrane fractions. Lane 1: plasma membrane fraction before extraction; lane 2: supernatant from 25 mM OG extraction; lane 3: supernatant from 50 mM OG extraction; lane 4: pellet remaining after the. 50 mM OG extraction. (B) Wheat germ agglutinin purification of diclofenac-labeled proteins. Lane 1: Coomassie blue-stained gel of plasma membrane proteins extracted in 50 mM OG; lane 2: Coomassie blue-stained gel of proteins eluted from column in presence of 0.5 M N-acetylglucosamine; lane 3: anti-diclofenac immunoblot of plasma membrane proteins extracted in 50 mM OG before chromatography; lane 4: anti-diclofenac immunoblot of proteins eluted from column in the presence of 0.5 M N-acetylglucosamine. Each lane contained 100 p g of protein. column (2 mm x 15 cm), and a 60 min linear gradient of 9 5 5 (v/v) solvent A (0.12% trifluoroacetic acid)/solvent B (0.12% trifluoroacetic acid in acetonitrile) to 40:60 (v/v) solvent A/solvent B at a flow rate of 0.2 m u m i n . Peaks were collected and sequenced by automated Edman degradation. Amino acid sequences from the 110 kDa protein were compared to sequences of proteins in SWISS-PROT database. Immunoprecipitation. Wheat germ agglutinin chromatography fractions that contained approximately 50 p g of the 110 kDa protein were incubated with DPP IV monoclonal antibody (80 pg) or phosphotyrosine monoclonal antibody (80 pg) a s a control for 4 h at 4 "C in a rotating mixer. Protein A-agarose was added [300 pL, 2.6 mg/mL commercial solution diluted 1 : l O in 10 mM Tris (pH 7.6) containing 0.5% (w/v) casein, 0.15 M NaC1, and 0.02% (w/v) thimerosal], and the mixture was incubated overnight at 4 "C in a rotating mixer. Samples then were centrifuged (lOOOOg, 10 min), and supernatants were removed. The pellets were washed three times with 1 mL of PBS containing 1% (w/v) Nonidet P-40 and once in 1mL of Trisbuffered saline (50 mM Tris, 200 mM NaCl, pH 7.4) and then were boiled in SDS-PAGE sample buffer [0.125 M Tris (pH 6.8) containing 5% (w/v) SDS, 20% (v/v) glycerol, 0.02% (w/v) bromophenol blue, and 40 mM dithiothreitol]. Immunoprecipitated proteins were separated by SDS-PAGE and then were analyzed by immunoblotting with DPP IV monoclonal antibody or diclofenac antisera as described above. DPP IV Assay. DPP IV activity in liver plasma membranes was determined a s described (22)with the following modification: assay buffer containing 25 mM OG was used in all incubations, and the assay was done in a total volume of 1mL. Liver plasma membrane fractions were prepared a s described above, using plasma membrane fractions from rats given diclofenac (200 m g k g ) or saline, and killed 16 h later. The plasma membrane proteins were extracted in PBS containing 50 mM OG, and the resulting extracts were adjusted to 1pg/pL protein with assay buffer. Assays were conducted with 10 or 50 pg of the extract by measuring absorbance at 384 nm of p-nitroaniline cleaved from Gly-Pro-p-nitroanilide during 10 min at 37 "C, using a Hewlett Packard 8452A diode array spectrophotometer. Other Methods. Protein concentrations were determined using the BCA protein assay reagent kit with bovine serum
Communications
Chem. Res. Toxicol., Vol. 8, No. 8, 1995 995
Table 1. Identity of the Amino Acid Sequences of the N-Terminus and of Three Internal Peptides from the 110 kDa Diclofenac-LabeledProtein with Portions of the Deduced Amino Acid Sequence of a Rat Liver cDNA Encoding DPP IV sequence of 110 kDa diclofenac-labeled protein
position in DPP N sequence
MKTPWKVLLGLLGVAALVTI LGTLEVEDQI WEYYDSVYTE YMGLPTPEDNLDH
1-20 599-608 660-669 671-683 kDa
200
I
I
1
2
3
4
Figure 2. Immunoprecipitation of 110 kDa diclofenac-labeled protein with DPP IV monoclonal antibody. Immunoprecipitations were done from wheat germ agglutinin-purified fractions of DPP IV from either control or diclofenac-treated rats. Lanes 1 and 2: Anti-DPP IV immunoblots of immunoprecipitated proteins from control rats (lane 1)and diclofenac-treated rats (lane 2). Lanes 3 and 4: Anti-diclofenac immunoblots of immunoprecipitated proteins from control rats (lane 3) and diclofenactreated rats (lane 4). albumin a s the standard, according to the manufacturer’s instructions. Statistical analysis was done by employing Student’s t-test for paired samples (N = 3).
Results Immunoblot analyses with diclofenac antiserum revealed that very little of the 110 kDa diclofenac-protein adduct was extracted from rat liver plasma membranes with 25 mM OG (Figure lA, lane 2). In contrast, 50 mM OG solubilized the majority of the 110 kDa adduct that was present in the plasma membranes (Figure lA, lane 3). When the 50 mM detergent extract was chromatographed on a column of wheat germ agglutinin-Sepharose 6MB, the 110 kDa protein bound to the column and was eluted with 0.5 M N-acetylglucosamine(Figure lB, lanes 2 and 4), indicating that the 110 kDa protein was glycosylated (23). The 200 kDa diclofenac-protein adduct and smaller amounts of the 140 kDa adduct also were eluted from the column with 0.5 M N-acetylglucosamine (Figure lB, lanes 2 and 4). After transblotting the purified 110 kDa protein to PVDF, amino acid sequences of the N-terminal and of three tryptic peptides were determined. All of the sequences were identical to regions of the deduced amino acid sequence of a rat liver cDNA encoding rat liver DPP IV (Table 1) (24, 25). To confirm that the 110 kDa diclofenac-labeled protein was DPP IV, immunoprecipitations with DPP IV monoclonal antibody were done from control or diclofenac-labeledrat liver plasma membrane fractions that had been purified by wheat germ agglutinin chromatography. Analyses of the immunoprecipitates by immunblotting with either DPP IV monoclonal antibody (Figure 2, lanes 1and 2) or with diclofenac antisera (Figure 2, lanes 3 and 4) revealed that the immunoprecipitatedDPP IV from diclofenac-treatedrats contained the diclofenac moiety. In addition, immuno-
precipitates of fractions from diclofenac-treated rats contained a 200 kDa protein that was detected in both anti-DPP IV and anti-diclofenacimmunoblots, suggesting that the 200 kDa adduct was a dimer of the 110 kDa adduct. To rule out the possibility that the observed immunoprecipitation results were due to nonspecific interactions of proteins, the immunoprecipitation experiment was repeated under identical conditions with phosphotyrosine monoclonal antibody in place of DPP IV monoclonal antibody. The 110 kDa protein was not immunoprecipitated by the monoclonal antibody against phosphotyrosine (data not shown). To determine if diclofenac adduct formation might affect the activity of DPP IV, a comparison was made of DPP IV activity in 50 mM OG extracts of rat liver plasma membranes from control and diclofenac-treated rats. Reaction rates were linear for 10 min, when either 10 or 50 pg of protein extract was used in the assay (data not shown). It was found that diclofenac treatment significantly decreased the activity of DPP IV by 22%,from 3.76 f 0.08 to 2.92 f 0.04 nmol-min-l*(mgof protein-l) ( p < 0.05).
Discussion In a previous study, we concluded that the diclofenacprotein adducts found in rat liver plasma membrane fractions that had molecular masses of 110,140, and 200 kDa were formed from an acyl glucuronide metabolite of diclofenac (16). In the present study we have identified the 110 kDa diclofenac-labeled protein as DPP IV. This conclusion is based on the amino acid sequence identity of the 110 kDa diclofenac-protein adduct with the sequence of DPP IV (Table l),the recognition of the adduct by a DPP IV monoclonal antibody (Figure 2), and the localization in the liver of both DPP IV (26) and the diclofenac-protein adducts (16) primarily in the bile canalicular domain of the plasma membrane, with the majority of the protein located on the extracellular surface (27). DPP IV (reviewed in ref 28>,also known as CD26,1F7, Tp 103, THAM, and adenosine deaminase binding protein, is a homodimeric integral membrane glycoprotein comprised of 110 kDa subunits. It is a serine exopeptidase that cleaves dipeptides from the N-terminus of proteins and peptides preferentially containing proline as the penultimate amino acid residue. DPP IV is present in many tissues includingliver, kidney, pancreas, spleen, small intestine, and placenta and on activated T lymphocytes. The glycoprotein is thought to be involved in the processing of proline-containing peptides and proteins, cell adhesion, and the activation of T lymphocytes and has been implicated in HIV-induced immunosuppression (29) and apoptosis (30). Although the physiological role of DPP IV in liver plasma membrane remains to be determined, the finding that treatment of rats with diclofenac led to a decrease in the peptidase activity of DPP IV raises the possibility that the hepatotoxicity caused by diclofenac is a direct result of DPP IV inhibition. Alternatively, since diclofenac is bound to DPP IV on the outer surface of rat hepatocytes, perhaps in susceptible human patients the adducts would be recognized as nonself and lead to an immune-mediated hepatotoxicity (31). Similar mechanisms may contribute to the hepatotoxicities caused by other acidic NSAIDs (14), which also form reactive acyl glucuronide metabolites in a manner analogous to that
996 Chem. Res. Toxicol., Vol. 8, No. 8, 1995
of diclofenac (32). Moreover, since DPP IV is present in many extrahepatic tissues, it is possible that extrahepatic toxicites associated with NSAIDs (33) also could be due to the covalent modification of DPP IV.
References (1) Dunk, A. A., Walt, R. P., Jenkins, W. J., and Sherlock, S. S. (1982) Diclofenac hepatitis. Br. Med. J . (Clin. Res. Ed.) 284, 1605-1606. (2) Schapira, D., Bassan, L., Nahir, A. M., and Scharf, Y. (1986) Diclofenac-induced hepatotoxicity. Postgrad. Med. J . 62, 63-65. (3) Breen, E. G., McNicholl, J., Cosgrove, E., McCabe, J., and Stevens, F. M. (1986) Fatal hepatitis associated with diclofenac. Gut 2, 1390-1393. (4) Salama, A., Schutz, B., Kiefel, V., Breithaupt, H., and MuellerEckhardt, C. (1989)Immune-mediated agranulocytosis related to drugs and their metabolites: mode of sensitization and heterogeneity of antibodies. Br. J . Haematol. 72, 127-132. (5) Sallie, R. (1990) Diclofenac hepatitis [letter]. J. Hepatol. 11,281. (6) Epstein, M., Vickars, L., and Stein, H. (1990) Diclofenac induced immune thrombocytopenia. J . Rheumatol. 17, 1403-1404. (7) Mueller-Eckhardt, C., and Salama, A. (1990) Drug-induced immune cytopenias: a unifylng pathogenetic concept with special emphasis on the role of drug metabolites. Transfus. Med. Rev. 4, 69-77. (8) Iveson, T. J., Ryley, N. G., Kelly, P. M., Trowell, J . M., McGee, J. 0.. and Chauman. R. W. (1990) Diclofenac associated heDatitis. J. Hepatol. io, 85-89. (9) Helfeott. S. M.. Sandberg-Cook, J.. Zakim. D.. and Nestler, J. (1996) Diclofenac-associagd hepatotoxicity [see comments]. JAMA, J . Am. Med. Assoc. 264, 2660-2662. (10) Purcell, P., Henry, D., and Melville, G. (1991)Diclofenac hepatitis. Gut 32, 1381-1385. (11) Sallie, R. W., McKenzie, T., Reed, W. D., Quinlan, M. F., and Shilkin, K. B. (1991) Diclofenac hepatitis. Aust. N . 2.J . Med. 21, 251-255. (12) Salama, A., and Mueller-Eckhardt, C. (1992) Immune-mediated blood cell dyscrasias related to drugs. Semin. Hematol. 29, 5463. (13) Scully, L. J., Clarke, D., and Barr, R. J. (1993) Diclofenac induced hepatitis. 3 cases with features of autoimmune chronic active hepatitis. Dig, Dis. Sci. 38, 744-751. (14) Zimmerman, H. J. (1994) Hepatic injury associated with nonsteroidal anti-inflammatory drugs. In Nonsteroidal Anti-inflammatory drugs: mechanisms and clinical uses (Lewis, A. J., and Furst, D. E., Eds.) 2nd ed., pp 171-194, Marcel Dekker, New York. (15) 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. (16) 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. (17) 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
Communications rat hepatocytes. Toxicol. Appl. Pharmacol. 120, 155-161. (18) Kretz-Rommel, A,, and Boelsterli, U. A. (1994) Selective protein
adduds to membrane proteins in cultured rat hepatocytes exposed to diclofenac: radiochemical and immunochemical analysis. Mol. Pharmacol. 45, 237-244. (19) Loten, E. G., and Redshaw-Loten, J. C. (1986) Preparation ofrat liver plasma membranes in a high yield. Anal. Biochem. 154, 183-185. (20) Martin, B. M., and Eliason, W. K. (1991) Determination of Glycoproteins Immobilized on PVDF and Its Derivatives. In Techniques in Protein Chemistry ZI (Villafranca, J. J., Ed.) pp 191-196, Academic Press, San Diego. (21) Fernandez, J.,Andrews, L., and Mische, S. M. (1994) An improved procedure for enzymatic digestion of polyvinylidene difluoridebound proteins for internal sequence analysis. Anal. Biochem. 218, 112-117. (22) Nagatsu, T., Hino, M., Fuyamada, H., Hayakawa, T., and Sakakibara, S. (1976) New chromogenic substrates for X-prolyl dipeptidyl-aminopeptidase. Anal. Biochem. 74, 466-476. (23) Stamatoglou, S. C., Ge, R. C., Mills, G., Butters, T. D., Zaidi, F., and Hughes, R. C. (1990) Identification of a novel glycoprotein (AGpllO) involved in interactions of rat liver parenchymal cells with fibronectin. J . Cell Biol. 111,2117-2127. (24) Ogata, S., Misumi, Y., and Ikehara, Y. (1989) Primary structure of rat liver dipeptidyl peptidase IV deduced from its cDNA and identification of the NHz-terminal signal sequence as the membrane-anchoring domain. J. Biol. Chem. 264, 3596-3601. (25) Hong, W., and Doyle, D. (1987) cDNA cloning for a bile canaliculusdomain-specific membrane glycoprotein of rat hepatocytes. Proc. Natl. Acad. Sci. U S A . 84, 7962-7966. (26) McCaughan, G. W., Wickson, J . E., Creswick, P. F., and Gorrell, M. D. (1990) Identification of the bile canalicular cell surface molecule GPllO as the ectopeptidase dipeptidyl peptidase rV: an analysis by tissue distribution, purification and N-terminal amino acid sequence. Hepatology 11,534-544. (27) Hong, W. J., and Doyle, D. (1988) Membrane orientation of rat gpll0 as studied by in vitro translation. J. Biol. Chem. 263, 16892-16898. (28) Fleischer, B. (1994) CD26: a surface protease involved in T-cell activation. Immunol. Today 15, 180-184. (29) Neubert, R., Helge, H., and Neubert, D. (1994) Thalidomide and the immune system. Down-regulation of the CD26 receptor, probably involved in the binding of HIV components to T cells in primates. Life Sci. 56, 407-420. (30) Morimoto, C., Lord, C. I., Zhang, C., Duke-Cohan, J . S., Letvin, N. L., and Schlossman, S. F. (1994) Role of CD26/dipeptidyl peptidase IV in human immunodeficiency virus type 1infection and apoptosis. Proc. Natl. Acad. Sci. U.S.A. 91, 9960-9964. (31) Pohl, L. R., Satoh, H., Christ, D. D., and Kenna, J. G.(1988) The immunologic and metabolic basis of drug hypersensitivities.Annu. Rev. Pharmacol. 28, 367-387. (32) 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. (33) Singh, G., Ramey, D. R., Morfeld, D., and Fries, J. F. (1994) Comparative toxicity of nonsteroidal anti-inflammatory agents. Pharmacol. Ther. 62, 175-191.
TX9501290
