Chemico-Biological Interactions 142 (2002) 25 /41 www.elsevier.

com/locate/chembiont

Idiosyncratic NSAID drug induced oxidative stress

Giuseppe Galati a, Shahrzad Tafazoli b, Omid Sabzevari c, Tom S. Chan d, Peter J. OBrien a,d,*
b

Department of Pharmacology, University of Toronto, Toronto, Ont., Canada M5S 2S2 Department of Chemical Engineering and Applied Chemistry, University of Toronto, Toronto, Ont., Canada M5S 2S2 c Faculty of Pharmacy, Tehran University of Medical Sciences, Tehran 14155/6451, Iran d Faculty of Pharmacy, University of Toronto, 19 Russell Street, Toronto, Ont., Canada M5S 2S2

Abstract Many idiosyncratic non-steroidal anti-inflammatory drugs (NSAIDs) cause GI, liver and bone marrow toxicity in some patients which results in GI bleeding/ulceration/fulminant hepatic failure/hepatitis or agranulocytosis/aplastic anemia. The toxic mechanisms proposed have been reviewed. Evidence is presented showing that idiosyncratic NSAID drugs form prooxidant radicals when metabolised by peroxidases known to be present in these tissues. Thus GSH, NADH and/or ascorbate were cooxidised by catalytic amounts of NSAIDs and hydrogen peroxide in the presence of peroxidase. During GSH and NADH cooxidation, oxygen uptake and activation occurred. Furthermore the formation of NSAID oxidation products was prevented during the cooxidation indicating that the cooxidation involved redox cycling of the first formed NSAID radical product. The order of prooxidant catalytic effectiveness of fenamate and arylacetic acid NSAIDs was mefenamic acid /tolfenamic acid /flufenamic acid, meclofenamic acid or diclofenac. Diphenylamine, a common moiety to all of these NSAIDs was a more active prooxidant for NADH and ascorbate cooxidation than these NSAIDs which suggests that oxidation of the NSAID diphenylamine moiety to a cation and/or nitroxide radical was responsible for the NSAID prooxidant activity. The order of catalytic effectiveness found for sulfonamide derivatives was sulfaphenazole /sulfisoxazole / dapsone /sulfanilic acid /procainamide /sulfamethoxazole /sulfadiazine /sulfadimethox-

* Corresponding author. Tel.: '/1-416-978-2716; fax: '/1-416-978-8511 E-mail address: peter.obrien@utoronto.ca (P.J. OBrien). 0009-2797/02/$ - see front matter # 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 0 0 9 - 2 7 9 7 ( 0 2 ) 0 0 0 5 2 - 2

26

G. Galati et al. / Chemico-Biological Interactions 142 (2002) 25 /41

ine whereas sulfanilamide, sulfapyridine or nimesulide had no prooxidant activity. Although indomethacin had little prooxidant activity, its major in vivo metabolite, N -deschlorobenzoyl indomethacin had significant prooxidant activity. Aminoantipyrine the major in vivo metabolite of aminopyrine or dipyrone was also more prooxidant than the parent drugs. It is hypothesized that the NSAID radicals and/or the resulting oxidative stress initiates the cytotoxic processes leading to idiosyncratic toxicity. # 2002 Elsevier Science Ireland Ltd. All rights reserved.
Keywords: NSAIDs; Prooxidant; Oxidative stress; Idiosyncratic drug reactions

1. Introduction A meta-analysis of reported serious and fatal adverse drug reactions in hospital patients suggests that this may be the fourth /sixth cause of mortality in North America [1]. The real incidence of all adverse drug reactions is not known as it is likely that only 10 /15% of recognized reactions are reported. The non-steroidal antiinflammatory drugs (NSAIDs) are widely used long term for the treatment of rheumatoid and osteoarthritis to relieve the pain and inflammation. The number of patients taking NSAIDs is greater than 30 million people with 1 million taking NSAIDs on a daily basis. This means that the potential for adverse effects is considerable. NSAIDs are responsible for up to 25% of all reported adverse drug reactions for which up to 60 /70% of patients using NSAIDs develop peptic ulceration or an enteropathy associated with intramucosal bleeding and other intestinal toxicities. It has been estimated that as a result of NSAID-associated upper gastrointestinal complications more than 100 000 patients are hospitalized and 16 500 die each year in the US [2]. Most of these drugs are still on the market because of their high therapeutic effectiveness. The meta-analysis ranking of NSAIDs for their ability at therapeutic doses for inducing serious gastrointestinal toxicity, e.g. gastrointestinal bleeding were reported as follows: ketoprofen /piroxicam /indomethacin /naproxen /sulindac /aspirin /fenoprofen /diflunisal /diclofenac /ibuprofen [3 /5]. The molecular mechanisms involved are not known. However diclofenac acyl glucuronide covalent binding to enterocyte proteins may contribute to diclofenac induced intestinal ulceration as ulceration was increased if hepatic glucuronyltransferase was induced and was prevented in Mrp2 deficient rats in which the enterohepatic recirculation was blocked [6]. The most common cause of fulminant hepatic failure in the US has been attributed to drug-induced liver disease of which NSAIDs are also a major class. The vast majority of these drugs are thought to be unpredictable toxins with a latency of atleast 1-month unaccompanied by signs of allergy (rash, eosinophilia, etc.) and are therefore described as idiosyncratic [7]. Many of these drugs exhibited an early

G. Galati et al. / Chemico-Biological Interactions 142 (2002) 25 /41

27

asymptomatic and reversible mild increase in serum alanine aminotransferase abnormalities which then disappeared. The risk of hepatitis was highest for sulindac which was reported as 148 per 100 000 patients. Hepatitis was much more rare for diclofenac but 15% of diclofenac users experienced an early increase in serum aminotransferase levels. With both NSAIDs there was also evidence that in some cases the NSAID caused liver disease on rechallenge [8]. Diclofenac induced hepatotoxicity and agranulocytosis has been described as idiosyncratic in nature but may involve metabolic activation to reactive intermediates. Diclofenac has been shown to bind covalently to various liver proteins when metabolically activated by CYP2C9 to form the quinoneimines which bind to the CYP2C enzyme [9] or by glucuronyltransferases to form acyl glucuronides which bind covalently to the plasma membrane protein dipeptidyl peptidase IV as well as a 60 kDa microsomal protein [10]. The anti-inflammatory, antipyretic, analgesic drug aminopyrine or its N desmethyl metabolite (metamizol, dipyrone) was removed from US over-the-counter sales in 1938 due to 45 cases of agranulocytosis caused by these drugs [11]. Another outbreak of agranulocytosis in the US occurred as a result of the ingestion of Chinese herbal medicines containing aminopyrine. However, these drugs are still used in Latin America, Asia and several European countries because of their low intestinal toxicity compared with other NSAIDs. The agranulocytosis induced by aminopyrine or diclofenac may involve metabolic activation to reactive intermediates by hypochlorite formed by myeloperoxidase in activated neutrophils [12,13]. Susceptibility of the individual to aminopyrine could result from the large variation (200-fold) in the individuals hepatic metabolism of aminopyrine [14]. The ranking of NSAIDs for inducing bone marrow toxicity, e.g. agranulocytosis/ neutropenia/aplastic anemia were reported as dipyrone /sulphasalazine /phenylbutazone, indomethacin /most other NSAIDs [15,16]. Desmethyl indomethacin, a metabolite of indomethacin formed by hepatic N -deacylation and O -demethylation, was oxidised by activated neutrophils to form a reactive iminoquinone metabolite which was proposed to cause agranulocytosis [17]. The NSAID ranking for estimated excess mortality per 108 persons due to community-acquired agranulocytosis, aplastic anemia, anaphylaxis and serious upper gastrointestinal complications was diclofenac (592) /aspirin (192) /dipyrone (25), acetaminophen (20). This suggests that the current regulatory ranking of NSAID drugs appears inappropriate [18]. Previously we showed that the aminopyrine radical cations formed by peroxidase readily oxidised GSH, NADH or arachidonate which in the process formed reactive oxygen species [19,20]. Myeloperoxidase in activated neutrophils was also shown to metabolise various idiosyncratic drugs that induce agranulocytosis or lupus to reactive metabolites [12,21]. In the following it is shown that most NSAIDs (Fig. 1) are also oxidised by peroxidase to form radical cations which also cooxidise GSH or NADH to form reactive oxygen species. This suggests that NSAID induced GI and idiosyncratic toxicity could also be initiated by the radicals and/or oxidative stress caused by the peroxidase catalysed metabolism of NSAIDs.

28

G. Galati et al. / Chemico-Biological Interactions 142 (2002) 25 /41

Fig. 1. Possible prooxidant NSAIDs.

G. Galati et al. / Chemico-Biological Interactions 142 (2002) 25 /41

29

Fig. 1 (Continued )

30

G. Galati et al. / Chemico-Biological Interactions 142 (2002) 25 /41

2. Materials and methods 2.1. Materials All chemicals of the highest grade available were purchased from Sigma-Aldrich Canada (Oakville, Ont.). 2.2. NADH cooxidation To a cuvette containing 0.1 M Tris /HCl (pH 7.4) buffer (supplemented with 2 mM DETAPAC to prevent trace-metal catalysed autooxidation), 10 mM, 100 mM or 1 mM NSAID, 100 mM NADH, 50 mM or 100 mM H2O2 were added. The reaction was initiated by the addition of 0.1 mM or 2 mM peroxidase (from horseradish, donor:hydrogen peroxide oxidoreductase; EC 1.11.1.7 (HRP)). Kinetic scans were recorded using a Pharmacia ultrospec spectrophotometer at the maximum for NADH (l 0/340 nm) to monitor its rate of oxidation. 2.3. Ascorbate cooxidation To a cuvette containing 0.1 M Tris /HCl (pH 7.4) buffer (supplemented with 2 mM DETAPAC), 10 mM, 100 mM or 1 mM NSAID, 50 mM ascorbate, 50 mM H2O2 were added. The reaction was initiated by the addition of 0.1 mM peroxidase. Kinetic scans were recorded using a Pharmacia ultrospec spectrophotometer at the maximum for ascorbate (l 0/266 nm) to monitor its rate of oxidation. 2.4. GSH cooxidation The reaction mixtures contained 0.1 M Tris /HCl (pH 7.4) buffer (supplemented with 2 mM DETAPAC), 100 mM NSAID, 400 mM GSH and 20 mM H2O2. The reaction was initiated by the addition of 0.25 mM peroxidase and incubated for 30 min. The total amount of GSH and GSSG were measured by the HPLC analysis of deproteinized samples (25% meta-phosphoric acid) after derivatization with iodoacetic acid and 2,4-dinitro-fluorobenzene [22] using a waters HPLC system (Model 510 pumps, WISP 710B auto injector and model 410 UV/VIS detector) equipped with a Waters mBondapak NH2 (10 mM) 3.9 )/300 mm column. 2.5. GSH oxygen activation Oxygen consumption was measured in a 2 ml reaction chamber using a Clarke type electrode (Yellow Springs, CA) at room temperature (20 8C). 10 mM, 100 mM or 1 mM NSAID, 400 mM GSH, 200 mM H2O2 were added to the chamber containing Tris /HCl buffer as described above. The reaction was initiated by the addition of 2 mM peroxidase.

G. Galati et al. / Chemico-Biological Interactions 142 (2002) 25 /41

31

2.6. UV/VIS spectra of peroxidase oxidation products Changes in the absorption spectra during the oxidation of NSAIDs by peroxidase/ H2O2 was determined by scanning the spectra from 230 /600 nm using a Pharmacia ultrospec spectrophotometer to scan every 90 s a solution containing 2 ml 0.1 M Tris /HCl (pH 7.4) buffer (supplemented with 2 mM DETAPAC), 50 mM NSAID, 50 mM H2O2 and 0.1 mM HRP. Peroxidase oxidation products had absorbance maxima as follows: diclofenac 455 nm [23], N -deschlorobenzoyl indomethacin 415 nm, 5-Aminosalicylate (5-AS) 255 nm, diphenylamine 485 nm, dapsone 420 nm, procainamide 415 nm, sulfadiazine 420 nm. Oxidation of NSAIDs was prevented with 100 mM NADH, 300 mM ascorbate or 200 mM GSH. 2.7. Isolation and incubation of hepatocytes Hepatocytes were obtained by collagenase perfusion of the liver of male Sprague / Dawley rats as described [24]. Approximately 85 /90% of the hepatocytes excluded trypan blue. Cells were suspended at a density of 106 cells/ml in round bottomed flasks rotating in a water bath maintained at 37 8C in Krebs-Henseleit buffer pH 7.4, supplemented with 12.5 mM Hepes under an atmosphere of 10% O2: 85% N2: 5% CO2. Cell viability was determined by measuring the exclusion of trypan blue (final concentration 0.1% w/v). Hepatocytes were preincubated for 30 min prior to the addition of chemicals. Stock solutions were prepared fresh prior to use. Hepatocyte cytotoxicity was determined by the trypan blue exclusion test [24] as the percentage of cells that take up trypan blue (0.1% w/v) at 2 h following the addition of chemicals to the hepatocytes. 2.8. HPLC analysis of hepatocyte GSH/GSSG The total amount of hepatocyte GSH and GSSG were measured by the HPLC analysis of deproteinized samples (25% meta-phosphoric acid) after derivatization with iodoacetic acid and 2,4-dinitro-fluorobenzene [22] using a Waters HPLC system (Model 510 pumps, WISP 710B auto injector and model 410UV/VIS detector) equipped with a Waters mBondapak NH2(10 mM) 3.9 )/300 mm column.

3. Results Table 1 compares the prooxidant activity of the radicals formed by the peroxidase catalysed oxidation of aminopyrine, indomethacin and aspirin with their major in vivo metabolites. The rate of NADH or ascorbate cooxidation and GSH oxygen activation catalysed by aminoantipyrine (a metabolite of aminopyrine or dipyrone) was much more effective than aminopyrine or dipyrone when oxidised by peroxidase. The order of catalytic effectiveness found for the pyrazolone NSAIDs (Group I in Fig. 1) was aminoantipyrine /aminopyrine /dipyrone. Aminopyrine

32

G. Galati et al. / Chemico-Biological Interactions 142 (2002) 25 /41

(10 mM) and its metabolite 4-aminoantipyrine (10 mM) increased GSSG formation by greater than 10% in isolated rat hepatocytes after a 90 min incubation. Table 1 also shows that NADH or ascorbate were rapidly oxidised in the presence of catalytic amounts of N -deschlorobenzoyl indomethacin, the major metabolite of indomethacin (Group II in Fig. 1). Indomethacin was less effective as a catalyst. NADH, ascorbate or GSH markedly inhibited the formation of N -deschlorobenzoyl indomethacin oxidation products as described in materials and methods (results not shown). Salicylicate, the major metabolite of aspirin (Group III in Fig. 1), also catalysed NADH or ascorbate cooxidation unlike aspirin. 5-AS, an agent widely used in the treatment of inflammatory bowel diseases and a major metabolite of sulfasalazine [25] was a much more effective catalyst of NADH or ascorbate cooxidation than salicylate. The NADH cooxidation was also accompanied by oxygen uptake for all

Table 1 Possible NSAID prooxidant radicals: aminopyrine/indomethacin/aspirin and their metabolites NSAID or NSAID metabolite NADH oxidationa (k )/103 min ( 1) Ascorbate oxidationb (k )/103 min ( 1) GSH:O2c (mM min ( 1)

I. Pyrazolone derivatives None Aminopyrine 1 mM Dipyrone 1 mM Aminopyrine 100 mM Aminopyrine 10 mM 4-Aminoantipyrine 100 mM 4-Aminoantipyrine 10 mM II. Indoleacetic acid derivatives Indomethacin 100 mM N -Deschlorobenzoyl indomethacin 100 mM III. Salicylate derivatives Aspirin 1 mM Salicylate 1 mM 5-Aminosalicylate 100 mM 5-Aminosalicylate 10 mM

99/2 4299/43 329/4 609/6 199/3 1779/17 469/5 119/2 919/9

1139/14 Interference Interference 1749/18 1289/12 3699/38 2279/23 1239/14 1309/15

B/2 4029/41 739/7 739/7 129/2 1529/15 249/3 B/2 NDd

99/2 199/3 18439/178 899/8

1149/13 1159/13 Interference 48909/476

B/2 B/2 B/2 B/2

The results shown represent the average of three separate experiments9/S.D. a NADH oxidation was monitored at 340 nm using a Pharmacia ultrospec spectrophotometer with the following reagents 0.1 M Tris /HCl (pH 7.4) buffer (supplemented with 2 mM DETAPAC), 100 mM NADH, 50 mM H2O2, 0.1 mM HRP and NSAID (concentrations as shown). b Ascorbate oxidation was monitored at 266 nm using a Pharmacia ultrospec spectrophotometer with the following reagents 0.1 M Tris /HCl (pH 7.4) buffer (supplemented with 2 mM DETAPAC), 50 mM ascorbate, 50 mM H2O2, 0.1 mM HRP and NSAID (concentrations as shown). c GSH dependent O2 uptake was measured using a Clarke type electrode in the presence of 0.1 M Tris / HCl (pH 7.4) buffer (supplemented with 2 mM DETAPAC), 400 mM GSH, 200 mM H2O2, 2 mM HRP and NSAID (concentrations as shown). d ND: not determined.

G. Galati et al. / Chemico-Biological Interactions 142 (2002) 25 /41 Table 2 Possible NSAID prooxidant radicals: diphenylamine containing NSAIDs NSAID or moieties NADH oxidationa (k )/103 min ( 1) GSSG formedc Ascorbate GSH:O2d oxidationb (mM GSH equiv.) (mM min( 1) (k )/103 min( 1)

33

IV. Anthranilic acid derivatives i) Fenamate NSAIDs None 99/2 Mefenamic acid 1149/11 Tolfenamic acid 539/5 Flufenamic acid 109/2 Meclofenamic acid 99/2 ii) Fenamate moieties Diphenylamine 5439/54 Anthranilic acid 99/2 N -phenylanthranilic acid 659/6 2,3(CH3)2 aniline 47709/471 2CH3, 3Cl aniline 539/5 Trifluoromethylaniline 99/2 (V) Arylacetic acid derivatives Diclofenac 1 mM 539/5 (2 mM HRP) 2,6Cl2 aniline 99/2

1139/14 1649/19 1259/15 1139/14 1139/14 3969/38 1239/15 1319/16 30169/303 1269/15 1149/14 1209/14 1149/14

689/7 2609/28 2209/21 929/10 809/9 1129/12 1239/13 2329/24 2609/29 2689/29 809/9 769/7 689/7

B/2 4219/42 2629/28 619/6 129/3 439/5 1109/12 1229/14 4899/48 6409/65 249/3 49/2 B/2

The results shown represent the average of three separate experiments9/S.D. a NADH oxidation was monitored at 340 nm using a Pharmacia ultrospec spectrophotometer with the following reagents 0.1 M Tris /HCl (pH 7.4) buffer (supplemented with 2 mM DETAPAC), 100 mM NADH, 50 mM H2O2, 0.1 mM HRP (unless shown otherwise) and 100 mM NSAID (unless shown otherwise). b Ascorbate oxidation was monitored at 266 nm using a Pharmacia ultrospec spectrophotometer with the following reagents 0.1 M Tris /HCl (pH 7.4) buffer (supplemented with 2 mM DETAPAC), 50 mM ascorbate, 50 mM H2O2, 0.1 mM HRP and 100 mM NSAID. c GSH oxidation was measured by determining GSSG levels in the reaction as detected by HPLC. The reaction mixture contained 0.1 M Tris /HCl (pH 7.4) buffer (supplemented with 2 mM DETAPAC), 100 mm NSAID, 400 mm GSH, 20 mm H2O2 and 0.25 mm HRP for 30 min. d GSH dependent O2 uptake was measured using a Clarke type electrode in the presence of 0.1 M Tris / HCl (pH 7.4) buffer (supplemented with 2 mM DETAPAC), 400 mM GSH, 200 mM H2O2, 2 mM HRP and 1 mM NSAID.

of the above NSAIDs. GSH prevented the formation of 5-AS peroxidase oxidation products at 255 nm (results not shown). Table 2 shows that anthranilic or arylacetic acid NSAIDs (Group IV and V in Fig. 1) cooxidised NADH, ascorbate or GSH when oxidised by peroxidase/H2O2. The cooxidation of GSH was accompanied by rapid oxygen uptake. This also occurred with NADH (results not shown). Furthermore only catalytic amounts of NSAIDs or H2O2 were required for the complete oxidation of NADH or GSH. Negligible oxidation of NADH or GSH occurred in the absence of either NSAIDs, peroxidase or H2O2. The order of effectiveness of fenamate and NSAIDs for catalysing NADH, ascorbate of GSH cooxidation was mefenamic acid /tolfenamic acid /flufenamic

34

G. Galati et al. / Chemico-Biological Interactions 142 (2002) 25 /41

acid, meclofenamic acid or diclofenac. Furthermore as shown in Fig. 2 the diclofenac peroxidase oxidation products (455 nm) were not formed in the presence of NADH, ascorbate or GSH. The bright yellow oxidation product formed when diclofenac is oxidised by peroxidase/H2O2 immediately disappears upon addition of NADH, ascorbate or GSH. As shown in Table 2 N -phenylanthranilic acid, a diphenylamine derivative (Fig. 1), is a common moiety of all fenamate NSAIDs and was a less effective prooxidant catalyst towards NADH, ascorbate and GSH cooxidation than mefenamic acid but was more effective than mefenamic acid and N -penylanthranilic acid for NADH and ascorbate cooxidation but less effective GSH cooxidation. Spectral studies showed that the diphenylamine peroxidase oxidation products (485 nm) were not formed in the presence of NADH, ascorbate or GSH (results not shown). Anthranilic acid was effective at catalysing GSH and ascorbate cooxidation but not at catalysing NADH cooxidation. The order of prooxidant catalytic activities of other moieties of the fenamate NSAIDs were 2,3(CH3)2 aniline (mefenamic acid) /2CH33Cl aniline (tolfenamic acid) /trifluoromethyl aniline (flufenamic acid) whereas 2,6Cl2 aniline (diclofenac) had little prooxidant activity.

Fig. 2. Peroxidase-catalysed oxidation of diclofenac is prevented by NADH, ascorbate or GSH. The reaction mixture contained 2 ml of 0.1 M Tris /HCl (pH 7.4) buffer (supplemented with 2 mM DETAPAC), 50 mM diclofenac (a); in addition 50 mM H2O2 and 1 mM HRP were added (b); the reaction mixture contained 2 ml of 0.1 M Tris /HCl (pH 7.4) buffer (supplemented with 2 mM DETAPAC), 50 mM diclofenac, 50 mM H2O2, 1 mM HRP and 100 mM NADH (c) or 300 mM ascorbate (d) or 200 mM GSH (e). The UV spectra were recorded using a Pharmacia ultrospec spectrophotometer.

G. Galati et al. / Chemico-Biological Interactions 142 (2002) 25 /41 Table 3 Possible NSAID prooxidant radicals: sulfonamide derivatives NSAIDs VI. Sulfonamide derivatives None Sulfaphenazole 1 mM Sulfisoxazole 1 mM Dapsone 1 mM Sulfanilic acid 1 mM Procainamide 1 mM Sulfamethoxazole 1 mM Sulfadiazine 1 mM Sulfadimethoxine 1 mM Sulfanilamide 1 mM Sulfapyridine 100 mM Pyridine 100 mM Nimesulide 100 mM NADH oxidationa (k )/103 min ( 1) GSH:O2b (mM min( 1)

35

119/2 40459/407 10309/102 1349/16 1179/11 619/6 359/5 209/4 159/3 129/2 119/2 119/2 119/2

B/2 249/3 189/3 B/2 B/2 B/2 B/2 B/2 B/2 B/2 B/2 B/2 B/2

The results shown represent the average of three separate experiments9/S.D. a NADH oxidation was monitored at 340 nm using a Pharmacia ultrospec spectrophotometer with the following reagents 0.1 M Tris /HCl (pH 7.4) buffer (supplemented with 2 mM DETAPAC), 100 mM NADH, 2 mM HRP, 100 mM H2O2 and 100 mM or 1 mM NSAID as shown. b GSH dependent O2 uptake was measured using a Clarke type electrode in the presence of 0.1 M Tris / HCl (pH 7.4) buffer (supplemented with 2 mM DETAPAC), 400 mM GSH, 200 mM H2O2, 2 mM HRP and 1 mM NSAID.

Table 3 shows that sulfonamide NSAIDs (Group VI in Fig. 1) were also effective catalysts for the peroxidase catalysed cooxidation of NADH or GSH which was accompanied by oxygen uptake. The order of effectiveness of the sulfonamide NSAIDs for catalysing NADH cooxidation was sulfaphenazole /sulfisoxazole / dapsone /sulfanilic acid /procainamide /sulfamethoxazole /sulfadiazine /sulfadimethoxine whereas sulfanilamide, sulfapyridine (or its pyridine moiety) or nimesulide had no prooxidant activity. NADH also prevented the formation of the peroxidase oxidation products of dapsone (420 nm), procainamide (415 nm) and sulfadiazine (420 nm) (results not shown).

4. Discussion Aminopyrine induced agranulocytosis has been associated with myeloperoxidase catalysed metabolic activation to reactive intermediates [12]. The following evidence suggests that cellular oxidative stress mediated by prooxidant N-cation radical metabolites could contribute to aminopyrine toxicity. Peroxidase catalysed aminopyrine metabolism involves two N -demethylation steps which release HCHO and form N -methylaminoantipyrine and then aminoantipyrine via intermediate N-cation radicals. The intermediate aminopyrine N-cation radical can be measured by ESR and absorbs at 565 nm [26]. Two radicals disproportionate to an iminium cation (Eq.

36

G. Galati et al. / Chemico-Biological Interactions 142 (2002) 25 /41

(1)) which is hydrolysed to N -methylaminoantipyrine and HCHO (Eq. (2)).

' 2R2 NCH' 3 0 R2 N CH2 ' R2 NH2 R2 N' CH2 ' H2 O 0 R2 NH2 ' HCHO
+

(1) (2)

Previously we showed that GSH prevents aminopyrine metabolism to HCHO by reducing the aminopyrine N-cation radical to aminopyrine and that in the process GSH was oxidised to its thiyl radical (Eq. (3)) which formed GSSG and reactive oxygen species presumably via Eq. (4) and Eq. (5), [19,27]. R2 NCH' 3 ' GSH 0 R2 NH2 ' GS + + GS ' GS( X GSSG( + + GSSG ( ' O2 X GSSG ' O( 2
+ +

(3) (4) (5)

Furthermore NADH was also oxidised by aminopyrine radicals to form NAD ' and was accompanied by oxygen uptake and reactive oxygen species formation presumably via Eq. (6) and Eq. (7), [19]. R2 NCH' 3 ' NADH 0 R2 CH3 NH2 ' NAD NAD ' O2 0
+ + ' O( 2 ' NAD + +

(6) (7)

The major aminopyrine metabolite aminoantipyrine (Fig. 1) was found to be more effective than aminopyrine at catalysing GSH or NADH oxidation. This could be explained as the Km for aminopyrine of 8.5 mM was much higher than the Km for the metabolites aminoantipyrine and N -methylaminoantipyrine [26]. The rate of oxygen uptake associated with GSH or NADH oxidation was also dependent on the peroxidase, aminoantipyrine metabolite or H2O2 concentrations whereas total oxygen uptake was not dependent on the drug concentration indicating that aminoantipyrine like aminopyrine acted as mediators of GSH or NADH oxidation catalysed by the peroxidase-H2O2 system. Oxygen consumption was proportional to the concentration of GSH or NADH with up to approximately 0.4 mol oxygen consumed for the oxidation of 1 mol GSH and up to 0.8 mol oxygen consumed for the oxidation of 1 mol NADH [19,27]. The prooxidant activity of aminopyrine N radical cations formed intracellularly has also been demonstrated in the intact liver as perfusion of the rat liver with aminopyrine caused GSSG release into the perfusate [28]. Furthermore the amount of GSSG formed was not diminished if glutathione peroxidase deficient livers (from Selenium deficient rats) were perfused with aminopyrine which suggests that the GSSG was formed from thiyl radicals formed by aminopyrine radical cations rather than as a result of GSH oxidation by H2O2 and glutathione peroxidase [29]. Indomethacin is a very effective NSAID but its use is limited by a high incidence of GI toxicity and thrombocytopenia, aplastic anemia and agranulocytosis or leukopenia. Indomethacin is an indoleacetic acid derivative which readily undergoes N -deacylation in vivo to form N -deschlorobenzoyl indomethacin (5OCH32CH3 indole-3-acetic acid) [30]. This is the major metabolite formed in vivo and we have found it to be a much better catalyst for NADH and ascorbate cooxidation than indomethacin. It was also shown that creatine kinase, an oxidative stress biomarker,

G. Galati et al. / Chemico-Biological Interactions 142 (2002) 25 /41

37

was readily inactivated by the indomethacin radicals formed by peroxidatic metabolism [31]. Other sulfhydryl enzymes including alcohol dehydrogenase and the glycolytic enzyme glyceraldehyde-3-phosphate dehydrogenase were also readily inactivated [31]. ESR spectroscopy was also used to show that lactoperoxidase at pH 5.5 catalysed indomethacin oxidation to form a nitrogen centered free radical and likely a carbon centered skatole radical which reacted with molecular oxygen. Spectral evidence also suggested indomethacin first bound at the aromatic donor binding site of the peroxidase [32]. Intestinal or gastric peroxidase activity was impaired by indomethacin and it has been suggested that gastric ulceration resulted from the decreased H2O2 detoxification [33,34]. However it is unlikely that such peroxidase inactivation occurs in the presence of the hydrogen donors ascorbate, NADH or GSH. Instead oxidative stress induced by the peroxidase catalysed metabolism of indomethacin or its deacylated metabolite to form prooxidant radicals could explain why indomethacin induced gastric ulceration was prevented by catalase, superoxide dismutase or the iron-chelating agent desferoxamine [35]. Although aspirin was inactive, salicylic acid, the pharmacologically active hydrolysis product and the major metabolite of aspirin underwent a peroxidase catalysed oxidation of NADH and GSH resulting in oxygen activation or a peroxidase catalysed oxidation of ascorbate. Other investigators have shown that salicylate also acted as a catalyst for myeloperoxidase-initiated LDL oxidation which they attributed to the formation of the reactive intermediate salicylate phenoxyl radical [36]. Aspirin was inactive as a peroxidase substrate because the phenolic hydroxy group is blocked by acetylation. Previously we also implicated acetaminophen phenoxyl radicals in GSH cooxidation, oxygen activation and arachidonate peroxidation by peroxidase activated acetaminophen [19]. Acetaminophen also increased myeloperoxidase catalysed LDL oxidation [37] and salicylate increased lipid peroxidation in activated neutrophils [36]. However the prooxidant activity of myeloperoxidase activated salicylate was prevented by trace amounts of the salicylate metabolite gentisic acid, an antioxidant, so that long term aspirin antithrombotic medication to prevent cardiovascular disease may not result in prooxidant effects [36]. 5-AS (Fig. 1) is an effective agent in the treatment of chronic inflammatory bowel disease but is associated with pancreatitis, neutropenia, thrombocytopenia, lupus and interstitial nephritis in some patients. Benzimidazole products formed from iminoquinone metabolites were recently reported in patients taking 5-AS [38]. Hypochlorite, the oxidant released by activated neutrophils also oxidised 5-AS to an iminoquinone reactive metabolite which formed GSH conjugates [39]. Interestingly NADH or ascorbate were rapidly oxidised by catalytic amounts of 5-AS in our peroxidase catalysed reaction system and 5-AS spectral changes that occurred as a result of 5-AS oxidation was prevented. Oxygen uptake also occurred during NADH cooxidation which can be attributed to 5-AS radical formation (Eqs. (6) and (7)). Whether this radical is a phenoxyl or an N-cation radical is not known. Toxicity (particularly GI toxicity) is the major drawback in the use of fenamate NSAIDs. The fenamates (Fig. 1) are potent analgesic anti-inflammatory drugs that likely act as acidic competitive inhibitors of arachidonate binding to cyclooxygenase

38

G. Galati et al. / Chemico-Biological Interactions 142 (2002) 25 /41

(COX) and cause a time-dependent loss of activity. They are N -arylated derivatives of anthranilic acid and have two phenyl rings that in the case of meclofenamic acid are almost perpendicular 758 to each other or 698, 508 and 448 for diclofenac, mefenamic acid and flufenamic acid respectively [40]. Anthranilic acid itself was relatively ineffective at catalysing NADH or GSH oxidation with low concentrations of H2O2 and peroxidase. A common moiety for the fenamate NSAIDs is diphenylamine or N -phenylanthranilic acid (Fig. 1) and both were a more effective catalyst than most fenamate NSAIDs. The order of effectiveness found for the fenamate and arylacetic acid NSAIDs was mefenamic acid /tolfenamic acid / flufenamic acid, meclofenamic acid or diclofenac. The diphenylamine moiety is a pesticide used in the production of fruits and vegetables which can induce kidney adenomas and carcinomas in rats and has been classified as a human carcinogen with intermediate potency. It also induces oxidative liver DNA damage in rats in vivo [41]. The diphenylnitroxide radical has been detected by ESR [42]. Most of the anti-inflammatory effects of NSAIDs are thought to be due to their ability to bind to and inhibit the COX moiety of prostaglandin endoperoxide synthase. However it is generally believed that the NSAIDs causing GI toxicity do so because they inhibit the gastroprotective constitutive COX-1 preferentially than COX-2 which is induced by cytokines, mitogens and endotoxins in inflammatory cells. Nimesulide differs from the other NSAIDs studied in that it is a selective COX2 inhibitor belonging to the diarylpyrazole group of compounds with a sulfonamide group that includes celecoxib which is now in current therapeutic use. As COX-2 inhibitors they would be expected to have minimal gastric and renal disturbances [43]. It is therefore of particular interest that nimesulide was found to be inactive in forming prooxidant radicals when metabolically activated by peroxidase. Nimesulide is also associated with a relatively low occurrence of GI toxicity compared with other sulfonamide NSAIDs [44]. The order of sulfonamide catalytic effectiveness found for NADH cooxidation was sulfaphenazole /sulfisoxazole /dapsone /sulfanilic acid / procainamide / sulfamethoxazole / sulfadiazine / sulfadimethoxine / sulfanilamide /sulfapyridine, nimesulide. Furthermore pyridine, a moiety of sulfapyridine, also had no peroxidase catalysed prooxidant activity. Sulfapyridine was also associated with a relatively low occurrence of GI toxicity compared with 5AS [45]. The negligible prooxidant activity of sulfanilamide, a common moiety of the sulfonamides and the high activity of sulfaphenazole or sulfisoxazole suggests that another moiety contributes to the prooxidant activity, e.g. 5-amino-1-phenylpyrazole (sulfaphenazole) or 5-amino-3,4-dimethylisoxazole (sulfisoxazole). These sulfonamide drugs are also associated with idiosyncratic bone marrow toxicity, e.g. agranulocytosis/neutropenia/aplastic anemia [15,16] likely as a result of metabolic activation by myeloperoxidase [46]. Mitotic and chromosomal abnormalities have also been observed in cells incubated with sulfadiazine and sulfaphenazole [47]. In conclusion many NSAIDs are oxidised by peroxidases to form prooxidant radicals which oxidise GSH, ascorbate or NADH and generate reactive oxygen species. Furthermore the target tissues of idiosyncratic toxicity i.e. GI tract, bone marrow or liver, contain not only various peroxidases but also prostaglandin

G. Galati et al. / Chemico-Biological Interactions 142 (2002) 25 /41

39

endoperoxide synthase. Recently NSAID-enteropathy in the rat has been attributed to both the uncoupling of intestinal mitochondrial oxidative phosphorylation as well as the inhibition of COX [48]. Interestingly whilst NSAIDs inhibit the COX activity of the prostaglandin endoperoxide synthase complex they do not affect the peroxidase moiety of the complex and could still utilize intracellular H2O2 or hydroperoxides to oxidise NSAIDs to radicals [49,50]. Endogenous hydrogen peroxide formation by various intracellular oxidases could initiate the above peroxidase catalysed oxidative stress [51]. It is hypothesized that NSAID prooxidant radicals formed by metabolism of NSAIDs by these peroxidases could contribute to the gastrointestinal mitochondrial toxicity and oxidative stress associated with NSAID induced gastrointestinal toxicity or idiosyncratic bone marrow or hepatic injury.

Acknowledgements This work was supported by a grant from the National Sciences and Engineering Research Council of Canada.

References
[1] J. Lazarou, B.H. Pomeranz, P.N. Corey, Incidence of adverse drug reactions in hospitalized patients, JAMA 279 (1998) 1200 /1205. [2] M.M. Wolfe, D.R. Lichtenstein, G. Singh, Gastrointestinal toxicity of nonsteroidal antiinammatory drugs, N. Engl. J. Med. 340 (1999) 1888 /1899. [3] J.L. Carson, L.R. Willet, Toxicity of nonsteroidal anti-inammatory drugs, Drugs 46 (Suppl. 1) (1993) 243 /248. [4] J.L. Wallace, Mechanisms of nonsteroidal anti-inammatory drug (NSAID) induced gastrointestinal damage-potential for development of gastrointestinal tract safe NSAIDs, Can. J. Physiol. Pharmacol. 72 (1994) 1493 /1498. [5] D. Henry, L.L. Lim, L.A. Garcia Rodriguez, S. Perez Gutthann, J.L. Carson, M. Grifn, R. Savage, R. Logan, Y. Moride, C. Hawkey, S. Hill, J.T. Fries, Variability in risk of gastrointestinal complications with individual non-steroidal anti-inammatory drugs: results of a collaborative metaanalysis, Br. Med. J. 312 (1996) 1563 /1566. [6] S. Seitz, U.A. Boelsterli, Diclofenac acyl glucuronide, a major biliary metabolite, is directly involved in small intestinal injury in rats, Gastroenterology 115 (1998) 1476 /1482. [7] N. Kaplowitz, Avoiding hepatic injury from drugs, Gastroenterology 117 (1999) 759. [8] A.V. Manoukian, J.L. Carson, Nonsteroidal anti-inammatory drug-induced hepatic disorders, Drug Safety 15 (1996) 64 /71. [9] W. Tang, R.A. Stearns, S.M. Bandiera, Y. Zhang, C. Raab, M.P. Braun, T.A. Bailie, Studies on cytochrome P450 mediated bioactivation of diclofenac in rats and in human hepatocytes: identication of glutathione conjugates, Drug Metab. Disp. 27 (1999) 365 /372. [10] S.J. Hargus, B.M. Martin, J.W. George, L.R. Pohl, Covalent modication of rat liver dipeptidyl peptidase IV (CD26) by the nonsteroidal anti-inammatory drug diclofenac, Chem. Res. Toxicol. 8 (1995) 993 /996. [11] F.W. Madison, T.L. Squier, The etiology of primary granulocytopenia (agranulocytic angina), J. Am. Med. Assoc. 102 (1934) 755 /759.

40

G. Galati et al. / Chemico-Biological Interactions 142 (2002) 25 /41

[12] J.P. Uetrecht, H.M. Ma, E. MacKnight, R. McClelland, Oxidation of aminopyrine by hypochlorite to a reactive dication: possible implications for aminopyrine induced agranulocytosis, Chem. Res. Toxicol. 8 (1995) 226 /233. [13] G. Miyamoto, N. Zahid, J.P. Uetrecht, Oxidation of diclofenac to reactive intermediates by neutrophils, myeloperoxidase and hypochlorous acid, Chem. Res. Toxicol. 10 (1997) 414 /419. [14] J.A. Agundez, C. Martinez, J. Benitez, Metabolism of aminopyrine and derivatives in man: in vivo study of monomorphic and polymorphic metabolic pathways, Xenobiotica 25 (1995) 417 /427. [15] M.M. van der Klauw, J.P. Wilson, B.H. Stricker, Drug-associated agranulocytosis: 20 years of reporting in The Netherlands (1974 /1994), Amer. J. Hematology 57 (1998) 206 /211. [16] S. Shapiro, Risks of agranulocytosis and aplastic anemia, JAMA 256 (1986) 1749 /1757. [17] C. Ju, J.P. Uetrecht, Oxidation of a metabolite of indomethacin (desmethyldeschlorobenzoyl indomethacin) to reactive intermediates by activated neutrophils, HOCl and the myeloperoxidase system, Drug Metab. Disp. 26 (1998) 676 /680. [18] S.E. Andrade, C. Martinez, A.M. Walker, Comparative safety evaluation of non-narcotic analgesics, J. Clin. Epidemiol. 51 (1998) 357 /365. [19] V.V. Subrahmanyam, L.G. McGirr, P.J. OBrien, Oxygen activation during drug metabolism, Pharmacol. Ther. 33 (1987) 63 /72. [20] P.J. OBrien, Radical formation during the peroxidase catalysed metabolism of carcinogens and xenobiotics: the reactivity of these radicals with GSH, DNA and unsaturated lipid, Free Radic. Biol. Med. 4 (1988) 169 /183. [21] X. Jiang, G. Khusigara, R.L. Rubin, Transformation of lupus-inducing drugs to cytotoxic products by activated neutrophils, Science 266 (1994) 810 /813. [22] D.J. Reed, J.R. Babson, P.W. Beatty, A.E. Brodie, W.W. Ellis, D.W. Potter, High-performance liquid chromatography analysis of nanomole levels of glutathione, glutathione disulde and related thiols and disuldes, Anal. Biochem. 106 (1980) 55 /62. [23] K.V. Zuurbier, A.R. Bakkenist, R.H. Fokkens, N.M.M. Nibbering, R. Wever, A.O. Muijsers, Interaction of myeloperoxidase with diclofenac: oxidation of diclofenac to dihydroxyazobenzene, Biochem. Pharmacol. 40 (1990) 1801 /1808. [24] P. Moldeus, J. Hogberg, S. Orrenius, Isolation and use of liver cells, Meth. Enzymol. 52 (1978) 60 / 71. [25] H. Allgayer, Sulfasalazine and 5-ASA compounds, Gastroenterol. Clinics North Am. 21 (1992) 643 / 657. [26] B.W. Grifn, P.L. Ting, Mechanisms of N -demethylation of aminopyrine by hydrogen peroxide catalysed by horseradish peroxidase, Biochemistry 17 (1978) 2206 /2211. [27] P. Moldeus, P.J. OBrien, H. Thor, M. Berggren, S. Orrenius, Oxidation of glutathione by free radical intermediates formed during peroxidase-catalysed N -demethylation reactions, FEBS Lett. 162 (1983) 411 /415. [28] N. Oshino, B. Chance, Properties of glutathione release observed during reduction of organic hydroperoxide, demethylation of aminopyrine and oxidation of some substances in perfused rat liver, and their implications for the physiological function of catalase, Biochem. J. 162 (1977) 509 /525. [29] H. Sies, G.M. Bartoli, R.F. Burk, C. Waydhas, Glutathione efux from perfused rat liver after phenobarbital treatment, during drug oxidations, and in selenium deciency, Eur. J. Biochem. 89 (1978) 113 /118. [30] R.E. Harman, M.A.P. Meisinger, G.E. Davis, F.A. Kuehl, Jr., The metabolites of indomethacin, a new antiinammatory drug, J. Pharmacol. Exp. Ther. 143 (1964) 215 /220. [31] T. Muira, S. Muraoka, Y. Fujimoto, Inactivation of creatine kinase during the interaction of indomethacin with horseradish peroxidase and hydrogen peroxide: involvement of indomethacin radicals, Chem.-Biol. Interact. 134 (2001) 13 /25. [32] R. Chatterjee, U. Bandyopathy, A. Mazumdar, R.K. Banerjee, Lactoperoxidase-catalysed oxidation of indomethacin, a nonsteroidal anti-inammatory drug, through the formation of a free radical, Biochem. Pharmacol. 52 (1996) 1169 /1175. [33] M. Valotti, K.F. Tipton, G.P. Sgaragli, Oxidative ring-coupling of tyrosine and its derivatives by intestinal peroxidase, Biochem. Pharmacol. 43 (1992) 945 /951.

G. Galati et al. / Chemico-Biological Interactions 142 (2002) 25 /41

41

[34] D. Das, D. Bandyopadhyay, R.K. Banerjee, Oxidative inactivation of gastric peroxidase by sitespecic generation of hydroxyl radical and its role in stress-induced gastric ulceration, Free Radic. Biol. Med. 24 (1998) 460 /469. [35] P.M. Vaananen, J.B. Meddings, J.L. Wallace, Role of oxygen-derived free radicals in indomethacininduced gastric injury, Am. J. Physiol. 261 (1991) G470 /G475. [36] M. Hermann, S. Kapiotis, R. Hofbauer, C. Seelos, I. Held, B. Gimeiner, Salicylate promotes myeloperoxidase-initiated LDL oxidation, antagonization by its metabolite gentisic acid, Free Radic. Biol. Med. 26 (1999) 1253 /1260. [37] S. Kapiotis, G. Sengoelge, M. Hermann, I. Held, C. Seelos, B.M. Gimeiner, Paracetamol catalyses myeloperoxidase-initiated lipid oxidation in LDL, Arterioscle. Thromb. Vasc. Biol. 17 (1997) 2855 / 2860. [38] J. Jensen, C. Comett, C.F. Ilsen, S. Bondesen, J. Christensen, L.A. Christensen, J. Tjomelund, S.H. Hansen, Identication of oxidation products of 5 aminosalicylic acid in faeces and the study of their formation in vivo, Biochem. Pharmacol. 45 (1993) 1201 /1209. [39] Z. Lui, R.A. McClelland, J.P. Uetrecht, Oxidation of 5-aminosalicylic acid by hypochlorous acid to a reactive iminoquinone, Drug Metab. Disp. 23 (1995) 246 /250. [40] R. Pouplana, C. Perez, J. Sanchez, J.J. Lozano, P. Puig-Parellada, The structural and electronical factors that contribute afnity of the time-dependent inhibition of PGHS-1 by indomethacin, diclofenac and fenamates, J. Computer-Aided Molecular Design 13 (1999) 297 /313. [41] M. Lodovici, C. Casalini, C. Briani, P. Dolara, Oxidative liver DNA damage in rats treated with pesticide mixtures, Toxicology 117 (1997) 55 /60. [42] K.E. Appel, S. Gorsdorf, T. Scheper, H.H. Ruf, C.S. Ruhl, A.G. Hildebrandt, Metabolic denitrosation of diphenylnitrosamine: a possible bioactivation pathway, J. Cancer Res. Clin. Oncol. 113 (1987) 131 /136. [43] D.L. DeWitt, COX-2-selective inhibitors: the new super aspirins, Molec. Pharmacol. 55 (1999) 625 / 631. [44] A.A. Shah, B. Thjodleifsson, F.E. Murray, E. Kay, M. Barry, G. Sigtharsson, H. Gudjonsson, E. Oddsson, A.B. Price, D.J. Fitzgerald, I. Bjarnason, Selective inhibition of COX-2 in humans is associated with less gastrointestinal injury, Gut 48 (2001) 339 /346. [45] J.M. Fardy, D.A. Lloyd, R.P.E. Reynolds, Adverse effects with oral 5-aminosalicylic acid, J. Clin. Gastroenterol. 10 (1988) 635 /637. [46] J.P. Uetrecht, N.H. Shear, N. Zahid, N -chlorination of sulfamethoxazole and dapsone by the myeloperoxidase system, Drug Metab. Disp. 21 (1993) 830 /834. [47] A. Badr, Cytogenetic activities of 3 sulfonamides, Mut. Res. 104 (1982) 95 /100. [48] S. Somasundaram, G. Sigtharssan, R.J. Simpson, J. Watts, M. Jacob, I.A. Tavares, S. Ra, A. Roseth, R. Foster, A.B. Price, J.M. Wrigglesworth, J.M. Bjarnason, Uncoupling of intestinal mitochondrial oxidative phosphorylation and inhibition of cyclooxygenase are required for the development of NSIAD enteropathy in the rat, Alim. Pharmacol. Ther. 14 (2000) 639 /650. [49] P.J. OBrien, A. Rahimtula, The possible involvement of a peroxidase in prostaglandin biosynthesis, Biochem. Biophys. Res. Comm. 70 (1976) 832 /838. [50] P.J. OBrien, Peroxidases, Chem.-Biol. Interact. 129 (2000) 113 /139. [51] A.G. Siraki, J. Pourahmad, T.S. Chan, S. Khan, P.J. OBrien, Endogenous and endobiotic induced reactive oxygen species formation by isolated hepatocytes, Free Radic. Biol. Med. 32 (2002) 2 /10.