The hepatotoxicity of several drugs is increased by mild viral infections. During such infections, death receptor ligands are expressed at low levels, and most parenchymal cells survive. We tested the hypothesis that subliminal death receptor stimulation may aggravate the hepatotoxicity of drugs, which are transformed by cytochrome P-450 cytochrome P-450 into glutathione-depleting reactive metabolites. Twenty-four-hour-fasted mice were pretreated with a subtoxic dose of the agonistic Jo2 anti-Fas antibody (1 μg per mouse) 3 hours before acetaminophen (500 mg/kg) or 1 hour before bromobenzene (400 mg/kg) administration. Administration of Jo2 alone increased hepatic inducible nitric oxide synthase nitric oxide synthase but did not modify serum alanine aminotransferase (ALT), hepatic adenosine triphosphate (ATP), glutathione (GSH), cytochrome P-450, cytosolic cytochrome c, caspase-3 activity or hepatic morphology. However, pretreating mice with Jo2 further decreased both hepatic GSH and ATP by 40% 4 hours after acetaminophen administration, and further increased serum ALT and the area of centrilobular necrosis at 24 hours. In mice pretreated with the Jo2 antibody before bromobenzene administration, hepatic GSH 4 hours after bromobenzene administration was 51% lower than in mice treated with bromobenzene alone, and serum ALT activity at 24 hours was 47-fold higher. In conclusion, administration of a subtoxic dose of an agonistic anti-Fas antibody before acetaminophen or bromobenzene increases metabolite-mediated GSH depletion and hepatotoxicity. Subliminal death receptor stimulation may be one mechanism whereby mild viral infections can increase drug-induced toxicity. (HEPATOLOGY 2004;39:655–666.)
Accumulating evidence indicates that the hepatotoxicity of some drugs is increased in patients with mild viral infections. In patients with tuberculosis treated with isoniazid, rifampin, and/or pyrazinamide, the risk of developing drug-induced hepatitis was increased 5-fold in patients infected with the hepatitis C virus (HCV), 4-fold in patients infected with the human immune deficiency virus (HIV), and 14-fold in patients coinfected with both viruses.1 The risk of antituberculous drug-induced hepatitis was increased 3-fold in hepatitis B virus (HBV) surface antigen carriers.2 Azathioprine-induced liver injury appears to be increased by infections with HCV, HBV, or cytomegalovirus.3, 4 In HIV-infected patients receiving highly active antiretroviral therapy, the risk of developing severe hepatotoxicity was increased several-fold by HCV or HBV coinfections.5, 6 Buprenorphine-induced hepatitis7 and ibuprofen-induced hepatitis8 mostly occurred in HCV-infected patients. Although lethal overdoses of aspirin cause microvesicular steatosis in children, therapeutic doses do not, unless aspirin is administered during viral infections.9 Finally, the hepatotoxicity of acetaminophen—also called paracetamol or N-acetyl-p-aminophenol (APAP)—may be increased in patients with infectious mononucleosis,10 measles,11 or acute hepatitis A or B,12 as well as in mice with influenza B virus infection.13
During viral infections, activated lymphocytes express membrane-bound tumor necrosis factor-α (TNF-α) and Fas ligand on their surface and release granzyme B, interferon-γ, TNF-α, soluble (proteolysed) Fas ligand, and full-length Fas ligand attached to microvesicles.14, 15 When present in large amounts, these effectors can damage mitochondria and activate caspases14, 16 to trigger apoptosis not only in infected parenchymal cells but also in bystander (noninfected) cells.17 In severe cases of acute hepatitis A or B, a major cytotoxic response can kill most hepatocytes and trigger fulminant hepatitis. However, in other viral infections, including chronic hepatitis B or C, most hepatocytes survive, although activated lymphocytes and excess cytokines are present in the liver.18 In these milder forms of viral diseases, hepatocytes are stimulated by death receptor ligands, but only in a subliminal way, which, in most hepatocytes, is insufficient to trigger cell death.
One possible reason for increased drug-induced hepatotoxicity in patients with mild viral infections could be that subliminal death receptor stimulation, albeit insufficient to cause cell death alone, could synergistically increase hepatotoxic drug effects. A frequent mechanism for drug-induced hepatotoxicity is the formation of reactive drug metabolites by cytochrome P-450 (CYP).19 This is the mechanism for the hepatotoxicity of APAP, an analgesic drug, which can cause liver failure not only after intentional overdoses but also after excessive therapeutic doses.20
APAP is oxidized by CYP1A, CYP2E1, and CYP3A into the reactive N-acetyl-p-benzoquinoneimine (NAPQI), which forms a conjugate with glutathione (GSH), and can deplete hepatic GSH.21 Hepatic GSH depletion plays a critical role in the hepatotoxicity of APAP and other compounds transformed into reactive metabolites, such as bromobenzene.22 No toxicity occurs with APAP or bromobenzene as long as sufficient hepatic GSH remains (>15%-20%), while centrilobular necrosis develops when GSH is severely depleted.22 GSH depletion increases the covalent binding of NAPQI to proteins,22 increases cell calcium,23 and triggers nitrotyrosine formation in proteins and mitochondrial dysfunction.24
In the present study, we tested the possibility that subliminal Fas stimulation may sensitize hepatocytes to the hepatotoxicity of APAP, and we also performed some experiments with bromobenzene. Pretreatment with a small dose of an agonistic anti-Fas antibody markedly increased APAP- or bromobenzene-induced GSH depletion and hepatotoxicity, even though Fas stimulation alone, at this very small dose of the antibody, had no discernible toxicity.
Male Swiss OF1mice, weighing from 28 to 30 g, were purchased from Charles River (L'Arbresle, France). Mice were fed ad libitum with a standard diet (A04-biscuits, UAR, Villemoisson-Sur-Orge, France) and then deprived of food for 24 hours before treatments. All treatments were administered intraperitoneally at fixed times to avoid interference from possible circadian variations.
In preliminary experiments, diverse doses of the agonistic Jo2 purified hamster anti-mouse Fas monoclonal antibody (BD Pharmingen, San Diego, CA) were used to assess mortality and serum alanine aminotransferase (ALT), in order to select a subtoxic dose of the antibody.
In the selected standard protocol, mice were treated at 7 AM with either saline (100 μL) or a nontoxic dose (1 μg per mouse) of the Jo2 antibody. Three hours later, mice received either saline or a toxic dose (500 mg/kg) of APAP (Sigma Chemical Co., St Louis, MO) dissolved in warm saline at a concentration of 25 mg/mL. Some mice were also treated with diethylmaleate (25 μL/kg) in 100 μL of corn oil, or N-acetyl-L-cysteine (120 mg/kg) in 100 μL of saline, 1 hour before APAP administration. Finally, other mice were treated with bromobenzene (400 mg/kg) in 200 μL of corn oil, 1 hour after Jo2 administration (1 μg per mouse).
Mortality Rates, Serum ALT Activity, and Morphologic Studies.
Mortality rates were assessed by counting dead animals 24 hours after APAP administration. In other experiments, mice were sacrificed 2, 5, 10, 24, or 48 hours after APAP administration, and serum ALT activity was measured with a commercial kit (ALT Infinity, Sigma).
Light microscopy studies were performed 4 and 24 hours after APAP administration. Immediately after sacrifice, livers were cut into small fragments. A fragment was fixed in PBS-buffered 10% formalin, dehydrated in an alcoholic series, placed in toluene baths and embedded in paraffin. Three-micrometer-thick sections were prepared for hematoxylin-eosin staining. The surface occupied by centrilobular hepatic necrosis was quantified 24 hours after APAP administration. For each of 7 mice treated with APAP alone and 7 mice treated with both Jo2 and APAP, 15 pericentral areas were randomly selected at ×3.2 magnification. The surface occupied by liver cell necrosis was analyzed with the image analyzer and Histolab Software from Microvision Instuments (Evry, France). The surface of the centrilobular vein was also computed and subtracted.
Electron microscopy studies were performed 4 hours after APAP administration. A liver fragment was cut into 1 mm3-blocks fixed in PBS-buffered 2.5% glutaraldehyde (Sigma) for 2 hours at 4°C. After dehydration in an alcoholic series, blocks were placed in propylene oxide solutions and embedded in epoxy resin. Ten blocks per animal were prepared, and semithin sections were stained with toluidine blue and examined on light microscopy for orientation. Ultrathin sections of at least 3 blocks per animal were stained with uranyl acetate and lead citrate, and examined with a JEOL 1010 electron microscope (JEOL, Tokyo, Japan).
Preparation of Hepatic Fractions and Assessment of GSH.
Mice were killed 2, 4, and 6 hours after APAP administration. One liver fragment was frozen at −80°C to later prepare a liver homogenate for the assay of total hepatic GSH. Another fragment was used to prepare cytosolic and mitochondrial fractions, as previously reported.25 Livers were homogenized in 220 mmol/L mannitol, 70 mmol/L sucrose, 2 mmol/L HEPES, and 0.1 mmol/L ethylenediaminetetraacetic acid (EDTA, pH 7.4), and centrifuged at 600g for 15 minutes at 4°C. The supernatant was centrifuged at 8,000g for 15 minutes. The 8,000g supernatant (cytosolic fraction) was stored at −80°C. The pellet was washed with the same buffer and centrifuged again at 8,000g. The final mitochondrial pellet was resuspended in buffer, and stored at −80°C. Hepatic GSH was determined by measuring nonprotein sulfhydryls as previously described.26
Total Microsomal CYP and CYP Proteins.
Hepatic microsomes were isolated by differential centrifugation and kept at −80°C. Total CYP was measured by the CO difference spectrum of dithionite-reduced microsomes.27 To assess individual CYPs, microsomal proteins were fractionated on sodium dodecyl sulfate (SDS)-polyacrylamide (10%) gels and transferred to nitrocellulose sheets. Immunoreactive CYP1A, CYP2E1, and CYP3A proteins were revealed as previously described.27
Hepatic Adenosine Triphosphate (ATP) and Caspase-3 Activity.
To assess hepatic ATP, the liver was quickly excised and immediately ground in liquid nitrogen. The liver powder was transferred into 500 μL of ice-cold 1 mol/L perchloric acid. After centrifugation at 4°C, 400-μL aliquots were neutralized with 5 mol/L K2CO3 and centrifuged again at 4°C. The pellet was resuspended in PBS containing 0.2 N NaOH, and used to determine protein content, and the supernatant was used to measure ATP with a luciferine-luciferase kit (Roche Diagnostics, Mannheim, Germany).28
To assess caspase-3 activity, livers were homogenized in 50 mmol/L HEPES buffer (pH 7.4) containing 5 mmol/L dithiothreitol, 1 mmol/L EDTA, and 0.1% 3-[(3-cholamidopropyl)dimethylammonio]-J-propanesulfonate. After centrifugation at 14,000g, the supernatant was recovered and protein concentration was measured. Cytosolic proteins (100 μg) were assayed for caspase-3 activity using the Caspase-3 Assay Kit (Biomol, Plymouth Meeting, PA). The liberation of p-nitroanilide (pNA) was determined at 405 nm with 200 μmol/L Ac-DEVD-pNA as substrate.
γ-Glutamylcysteine Synthetase (γ−GCS) Activity.
γ-GCS (an enzyme also called glutamate-L-cysteine ligase) catalyzes the rate-limiting step in the synthesis of GSH.29 As described by Seelig and Meister,29 the activity of γ-GCS was determined using L-α-aminobutyrate as an alternative substrate to L-cysteine, which may auto-oxidize. Livers were homogenized in 100 mmol/L Tris (pH 8) containing 5 mmol/L MgCl2 and 2 mmol/L dithiothreitol. Homogenates were centrifuged at 13,550g for 30 minutes at 4°C. The supernatants were used to determine γ-GCS activity following the oxidation of NADH at 340 nm in 100 mmol/L Tris buffer (pH 8) containing 150 mmol/L KCl, 2 or 5 mmol/L disodium ATP, 10 mmol/L L-glutamate, 10 mmol/L L-α-aminobutyrate, 2 mmol/L phosphoenolpyruvate, 2 mmol/L disodium EDTA, 20 mmol/L MgCl2, 0.2 mmol/L NADH, and 17 μg pyruvate kinase/lactate dehydrogenase. Results were expressed as nmol of NADH oxidized per mg of protein and per minute.
Western Blot Analysis of Bax and Cytochrome c.
Cytosolic and mitochondrial proteins were prepared as described above. Proteins were submitted to SDS-polyacrylamide gel electrophoresis (12% polyacrylamide for Bax and 15% for cytochrome c) and transferred to nitrocellulose membranes. Membranes were blocked with 1% polyvinyl pyrrolidone in PBS-Tween (0.1%) buffer, and incubated with 1 μg/mL of mouse anti-cytochrome c monoclonal antibody (Biomol)) or mouse anti-Bax monoclonal antibody (BD Pharmingen). Blots were exposed to the corresponding (isotype-specific) peroxidase–conjugated anti-immunoglobulin and revealed by an enhanced chemiluminescence detection system (Amersham Pharmacia, Orsay, France). To assess equal protein loading, blots were stripped, and cytosolic blots were incubated with an anti-β-actin antibody (Sigma), while mitochondrial blots were incubated with an antibody against subunit 1 of cytochrome c oxidase (Molecular Probes, Eugene, OR).
Western Blot Analysis of Inducible Nitric Oxide Synthase (iNOS).
Livers were homogenized in 0.1% 3-[(3-cholamidopropyl)dimethylammonio]-J-propanesulfonate, 5 mmol/L EDTA, 2 mmol/L dithiothreitol, and 25 mmol/L N-(hydroxy-2-ethyl)-piperazine N′-2-ethane sulfonic acid buffer (pH 7.5), containing antiproteases (Complete Protease Inhibitor Cocktail Tablets, Roche Diagnostics). The homogenate was centrifuged at 14,000g for 20 minutes at 4°C. Proteins of the supernatant were submitted to SDS-polyacrylamide (7.5%) gel electrophoresis, and transferred to nitrocellulose membranes. Membranes were blocked with 5% albumin in PBS-Tween (0.1%) buffer, and incubated first with 0.4 μg/mL of rabbit anti-iNOS polyclonal antibody (Biomol), then with a peroxidase-conjugated anti-rabbit immunoglobulin, and finally revealed by an enhanced chemiluminescence detection system (Amersham Pharmacia). Blots were stripped and incubated with an anti-β-actin antibody (Sigma). The 130-kd iNOS protein band and the β-actin band were quantified by laser densitometry, and the iNOS/β-actin densitometric ratio was calculated.
Differences between one control group and several treated groups were assessed by ANOVA, followed by Dunnett tests. Differences in the surfaces of necrotic areas in mice treated with APAP alone and mice treated with both Jo2 and APAP were assessed by a Student's t test.
Mortality Rates and Selection of a Subtoxic Dose of Jo2.
We first studied the effects of pretreating the animals with various doses of the Jo2 anti-Fas antibody on mortality rates 24 hours after APAP administration (500 mg/kg) (Fig. 1). The death rate was 10% with APAP alone and 20% with 6 μg per mouse of the Jo2 antibody alone. Although no death occurred with 1, 2, or 4 μg of the Jo2 antibody alone, pretreating mice with these Jo2 doses 3 hours before APAP administration dose-dependently increased the percentage of dead animals, from 10% in mice receiving APAP alone, to 16%, 20%, and 36% in mice first pretreated with 1, 2, or 4 μg per mouse of the Jo2 antibody.
In a second series of experiments, we looked for a subtoxic dose of the Jo2 antibody administered alone (Fig. 2). Serum ALT activity was markedly increased 8 hours after administration of 4 or 6 μg of the Jo2 anti-Fas antibody per mouse and was still moderately increased with 2 μg. However, ALT was not significantly increased with 1 μg of the Jo2 antibody (mean ± SEM: 62 ± 7 U/L in control mice and 82 ± 9 in Jo2-treated mice; Fig. 2). This subtoxic dose was selected for further studies.
Effect of a Subtoxic Jo2 Dose on APAP Toxicity.
Mice were pretreated with 1 μg per mouse of the Jo2 anti-Fas antibody 3 hours before APAP (500 mg/kg) administration, and serum ALT activity was measured 2, 5, 10, 24, and 48 hours after APAP administration (Fig. 3). At 5, 10, or 24 hours after APAP administration, ALT activity was higher with the combined Jo2/APAP treatment than with APAP alone, even though Jo2 alone had no significant effect on ALT.
Because apoptotic cells, unlike necrotic cells, are quickly removed by phagocytosis, we first looked for apoptosis and/or necrosis 4 hours after APAP (or saline) administration, using both light and electron microscopy. After administration of Jo2 alone (1 μg per mouse), hepatic morphology appeared normal on both light and electron microscopy. After administration of either APAP alone, or first Jo2 and then APAP, centrilobular hepatocytes exhibited the typical ultrastructural features of necrosis (Fig. 4), without any apoptotic hepatocytes. On light microscopy, there was no difference at that time in the surface of centrilobular hepatic necrosis in mice treated with APAP alone or those treated with both Jo2 and APAP. However, although mediolobular hepatocytes were normal in APAP-treated mice, their cytoplasm appeared abnormally clear in Jo2/APAP-treated mice (results not shown).
Liver histology was also assessed 24 hours after APAP administration (Fig. 5). In 3 mice treated with Jo2 alone (Fig. 5B), liver morphology was similar to that in control mice (Fig. 5A). In 7 mice treated with APAP alone, there were centrilobular sinusoidal dilation and pericentral hepatocyte necrosis (Fig. 5C). In 7 mice treated with both Jo2 and APAP, extensive centrilobular and mediolobular necrosis was present, together with centrilobular sinusoidal dilation (Fig. 5D). The area occupied by centrilobular necrosis (mean ± SEM for 7 mice) was significantly larger (P < .0001) in mice treated with both Jo2 and APAP (79 ± 15 × 103 μm2) than in mice treated with APAP alone (57 ± 15 × 103 μm2). Again, no apoptotic hepatocytes were observed on light microscopy, whatever the treatment.
Taken together, these results indicate that pretreatment with a subtoxic dose of the Jo2 antibody (1 μg per mouse) increases the extent of hepatic necrosis 24 hours after APAP administration.
Effects of Jo2 and/or APAP on Hepatic GSH.
A critical factor in APAP toxicity is the extent of GSH depletion.22 We therefore assessed total hepatic GSH, and cytosolic and mitochondrial GSH, 2, 4, and 6 hours after APAP administration (Fig. 6). Although hepatic GSH was unchanged after administration of JO2 alone, GSH decreased 2 and 4 hours after APAP administration, to partially recover at 6 hours. Four hours after APAP administration, total hepatic GSH was 42% lower—and cytosolic GSH 55% lower—with the combined Jo2/APAP treatment than with APAP alone. Mitochondrial GSH was less depleted by APAP than cytosolic GSH and was not significantly lower with the combined Jo2/APAP treatment than with APAP alone.
Effects of Jo2 and/or APAP on Hepatic CYP.
CYP1A, CYP2E1, and CYP3A oxidize APAP into NAPQI, which depletes GSH.21 One mechanism that can increase APAP-mediated GSH depletion is CYP induction.30 We therefore assessed total hepatic microsomal CYP and immunoreactive CYP1A, CYP2E1 and CYP3A after Jo2 and/or APAP administration (Table 1). While Jo2 alone had no effect on hepatic CYPs, APAP alone slightly decreased total CYP, probably as a consequence of centrilobular necrosis. The Jo2/APAP treatment tended to cause greater decreases in immunoreactive CYPs, although differences from APAP alone were not statistically significant (Table 1). Thus, no CYP induction, but rather a slight CYP decrease, occurred with the Jo2/APAP treatment.
Table 1. Hepatic CYPs, ATP, and Caspase-3 Activity After Administration of the Jo2 Anti-Fas Antibody and/or APAP
Total CYP (nmol/mg protein)
CYP1A (% control)
CYP2E1 (% control)
CYP3A (% control)
ATP (nmol/mg protein)
Caspase-3 (pmol/min/mg protein)
NOTE. 24-hour-fasted mice were pretreated with saline or the Jo2 anti-Fas antibody (1 μg per mouse), then treated with saline or APAP (500 mg/kg) 3 hours later, and finally killed 4 hours after APAP (or saline) administration. Hepatic microsomes were prepared, and total CYP (mean ± SEM for 5 mice) was measured from the absorbance of its carbon monoxide complex in dithionite-reduced microsomes. Immunoreactive CYPs (mean ± SEM for 4 mice) were quantified on microsomal immunoblots, and protein bands were expressed as the percentage of the mean value in control mice. After snap-freezing, total hepatic ATP (mean ± SEM for 5–7 mice) was measured by a luciferine-luciferase assay. Caspase-3 activity (mean ± SEM for 5 mice) was assessed in a cytosolic fraction by measuring the release of p-nitroaniline (pNA) from 200 μmol/L Ac-DEVD-pNA.
Significantly different from control mice (P < .05).
Significantly different from APAP-treated mice (P < .05).
Effects on ATP, Caspase-3 Activity and γ-GCS Activity.
Rather than an increased metabolic activation, a decreased GSH synthesis could account for the greater GSH depletion after the dual Jo2/APAP treatment than after APAP alone. The rate-limiting step in GSH synthesis is the ATP-driven synthesis of γ-glutamyl-L-cysteine by γ-GCS.31 Therefore, this synthesis is dependent first on the availability of ATP and substrates31 and on the amount of γ-GCS, which can be cut and inactivated by caspase-3.32
We therefore measured hepatic ATP, caspase-3 and γ-GCS activity in mice treated with Jo2 and/or APAP (Table 1). Hepatic ATP was not significantly modified by Jo2 alone or APAP alone but was decreased by 60% by the combined Jo2/APAP treatment. Caspase-3 activity was not significantly modified by Jo2 alone but was increased to a similar extent by APAP alone or by the combined Jo2/APAP treatment.
The activity of γ-GCS was measured ex vivo, in the presence of high concentrations of ATP (5 mmol/L) and substrates (10 mmol/L) (Fig. 7). γ-GCS activity was not modified by the administration of Jo2 alone. It was decreased to a similar extent by APAP alone or the combined Jo2/APAP treatment (24% and 27%, respectively). Administration of z-VAD-fmk, a caspase inhibitor, 30 minutes before APAP totally prevented the decrease in γ-GCS activity. To reproduce the 60% decrease in liver ATP observed in Jo2/APAP-treated mice (Table 1), the γ-GCS activity of Jo2/APAP-treated mice was also measured with a 60% lower ATP concentration in the assay (2 mmol/L instead of 5 mmol/L). This low ATP concentration further decreased γ-GCS activity by 30% (Fig. 7). Overall, the activity measured with Jo2/APAP-homogenates in the presence of a low ATP concentration was decreased by 47% compared to the activity of control homogenates in the presence of a high ATP concentration (Fig. 7). Thus, both a severe decrease in ATP and a moderate, caspase-3-mediated decrease in γ-GCS may combine their effects to partially hamper hepatic GSH synthesis and aggravate GSH depletion with the dual Jo2/APAP treatment.
Bax, Cytochrome c, and iNOS.
No changes were detected in cytosolic or mitochondrial Bax or in cytosolic or mitochondrial cytochrome c (Fig. 8).
Hepatic iNOS was significantly increased by the Jo2 treatment alone (Fig. 8). Similar average increases were observed after APAP alone or after the combined Jo2/APAP treatment, although differences from control mice were not statistically significant because of higher inter-mouse variability in APAP-treated mice (Fig. 8).
Role of GSH Depletion in Toxicity.
Because GSH depletion plays a critical role in APAP toxicity,22 the more severe GSH depletion after the combined Jo2/APAP treatment than after treatment with APAP alone could account, at least in part, for the higher toxicity of the dual treatment. To confirm this relationship, we used small doses of N-acetyl-L-cysteine (a GSH precursor) or diethylmaleate (a GSH-depleting agent) to modulate GSH depletion (Fig. 9). In mice receiving the combined Jo2/APAP treatment, the administration of a small dose of N-acetyl-L-cysteine (120 mg/kg), 1 hour before APAP administration, restored hepatic GSH to levels similar to those observed with APAP alone, and concomitantly decreased serum ALT activity.
Conversely, in mice not treated with Jo2, pretreatment with a small dose of diethylmaleate (25 μl/kg) 1 hour before APAP administration reproduced both the severe GSH depletion and the high serum ALT observed after the combined Jo2/APAP treatment.
Subliminal Fas Stimulation Also Increases the Toxicity of Bromobenzene.
To see whether these observations can be extended to other compounds forming toxic metabolites, we also studied the effects of a Jo2 pretreatment on the hepatotoxicity of bromobenzene. Like APAP, bromobenzene is transformed by CYP into reactive electrophilic metabolites that deplete hepatic GSH.22 As with APAP, the extent of bromobenzene-induced hepatic necrosis is critically dependent on the extent of hepatic GSH depletion.22 Pretreatment with the Jo2 antibody (1 μg per mouse), 1 hour before bromobenzene administration (400 mg/kg), aggravated GSH depletion and considerably increased serum ALT activity compared to the administration of bromobenzene alone (Fig. 10).
The hepatotoxicity of several drugs is increased by mild viral infections,1–13 which may cause subliminal stimulation of death receptors.14–18 Because drug-induced toxicity is often due to the formation of reactive metabolites,19 we tested the hypothesis that subliminal death receptor stimulation may increase metabolite-mediated hepatotoxicity.
Subliminal Fas stimulation was achieved by pretreating mice with a small dose of the agonistic Jo2 anti-Fas antibody (1 μg per mouse). This subtoxic Jo2 dose did not significantly increase serum ALT activity (Fig. 3), hepatic cytosolic cytochrome c (Fig. 8) or hepatic caspase 3 activity (Table 1), and caused no morphological changes on light microscopy (Fig. 5) or electron microscopy. However, pretreating mice with this subtoxic Jo2 dose increased both serum ALT activity (Fig. 3) and the area of necrosis (Fig. 5) 24 hours after APAP administration (500 mg/kg).
The high toxicity of the Jo2/APAP treatment could theoretically be due either to the enhancement of Fas toxicity by APAP or to the enhancement of APAP toxicity by Fas. However, all evidence points to the second alternative. First, previous studies have shown that pretreatment with APAP before the administration of a high Jo2 dose instead decreases Jo2-induced apoptosis,33 possibly because APAP-induced ATP depletion prevents apoptosis,34 which requires energy. Second, in the present study, the features of hepatitis after Jo2/APAP administration differed from Jo2-induced liver failure16 but resembled APAP-induced toxicity. Indeed, although a high dose of Jo2 (8 μg per mouse) causes major translocation of mitochondrial cytochrome c to the cytosol, extensive caspase activation, and massive apoptosis in periportal hepatocytes,16 the Jo2/APAP treatment did not change cytosolic cytochrome c (Fig. 8), did not further increase caspase-3 activity above the mild increase observed with APAP alone (Table 1), and did not cause periportal apoptosis (Figs. 4 and 5). Instead, the Jo2/APAP treatment caused necrosis of centrilobular hepatocytes, as did APAP alone (Figs. 4 and 5). Thus, it can be concluded that the Jo2 pretreatment increases the hepatotoxicity of APAP rather than vice versa.
APAP is transformed by CYPs into NAPQI, which forms a conjugate with GSH and can deplete hepatic GSH.21 GSH depletion plays a critical role in APAP-induced hepatotoxicity.21 No toxicity occurs as long as sufficient hepatic GSH remains to inactivate NAPQI, but necrosis develops when GSH is severely depleted.22 GSH depletion increases both the covalent binding of NAPQI to proteins22 and protein thiol depletion,23 which inactivates plasma membrane Ca2+-ATPases to increase cell calcium23 and activate diverse cytotoxic pathways.35 Mitochondrial dysfunction plays a major role in APAP-induced hepatotoxicity. When GSH is depleted, NAPQI covalently binds to and inactivates mitochondrial proteins36–38 to block electron flow in the respiratory chain,36, 38 which may increase superoxide anion formation. Although iNOS was not yet induced 16 hours after APAP administration,39 iNOS was induced at 24 hours.40 The superoxide anion reacts with NO to form peroxynitrite,40 which forms nitrotyrosine adducts in proteins and may inactivate mitochondrial proteins to further decrease hepatic ATP.41 ATP depletion prevents apoptosis and causes necrosis,42 explaining why APAP mainly causes necrosis. Because CYPs are much more expressed in centrilobular than periportal hepatocytes, necrosis selectively affects centrilobular cells.19 APAP-induced centrilobular necrosis can be prevented by administration of either an iNOS inhibitor40 or a superoxide dismutase mimic,43 reinforcing the view that peroxynitrite, which is formed by the reaction of the superoxide anion with NO, plays an important role in APAP-induced toxicity.
Jo2 has several mitochondrial effects, which can aggravate APAP-induced ATP depletion. First, Fas stimulation can cause the outer mitochondrial membrane to leak pro-apoptotic proteins and can trigger mitochondrial permeability transition.16 Although we could not detect an increase in cytosolic cytochrome c with the small Jo2 dose used in the present study, even after the dual Jo2/APAP treatment (Fig. 8), this does not exclude an increased permeability of some mitochondria, especially since necrosis, which ruptures plasma membranes, may allow cytochrome c to leak out of the cytoplasm, while β-actin, which is mostly associated with the cytoskeleton, may be retained within the cell. Second, Jo2 increases hepatic iNOS expression (Fig. 8).44 Induction of iNOS by JO2 could increase peroxynitrite formation and further impair mitochondrial function in APAP-treated mice. Indeed, although APAP alone nonsignificantly decreased hepatic ATP by 30%, the Jo2/APAP treatment decreased ATP by 60% (Table 1).
Jo2 administration also sensitized the liver to APAP-mediated GSH depletion. Four hours after APAP administration, total hepatic GSH was 42% lower with the combined Jo2/APAP treatment than with APAP alone (Fig. 6). Reactive metabolite-mediated GSH depletion is due to an imbalance between GSH synthesis, on the one hand, and CYP-mediated reactive metabolite formation and GSH consumption, on the other hand. CYP proteins were unchanged after administration of Jo2 alone and were slightly decreased after APAP administration (Table 1), possibly as a consequence of centrilobular necrosis. The rate-limiting step in GSH synthesis is the ATP-driven synthesis of γ-glutamyl-L-cysteine by γ-GCS, an enzyme which is cut and inactivated by caspase-3.32 In mice receiving the dual Jo2/APAP treatment, 2 mechanisms may add their effects to decrease γ-GCS activity and GSH synthesis. First, either APAP alone or the Jo2/APAP treatment increased caspase-3 activity (Table 1) and concomitantly decreased the ex vivo activity of γ-GCS (Fig. 7). This decrease was prevented by z-VAD-fmk (Fig. 7), confirming the role of caspases in decreasing γ-GCS activity.32 Second, decreasing ATP by 60% in the assay (to 2 mmol/L instead of 5 mmol/L), in order to mimic the 60% ATP decrease in Jo2/APAP-treated mice (Table 1), further decreased the γ-GCS activity in Jo2/APAP-treated mice (Fig. 7). Thus, the added effects of γ-GCS inactivation due to APAP-mediated caspase activation, and of severe ATP depletion after the dual Jo2/APAP treatment, markedly decrease γ-GCS activity in Jo2/APAP-treated mice, which may decrease GSH synthesis and aggravate APAP-induced GSH depletion in these mice.
The role of GSH depletion in extending APAP-induced toxicity was shown by modulating GSH depletion with N-acetyl-L-cysteine, a GSH precursor, or diethylmaleate, a GSH-depleting agent (Fig. 9). In mice receiving the combined Jo2/APAP treatment, the administration of a small dose of N-acetyl-L-cysteine 1 hour before APAP administration restored hepatic GSH to levels similar to those observed with APAP alone and decreased serum ALT activity. Conversely, in mice not treated with Jo2, pretreatment with a small dose of diethylmaleate 1 hour before APAP administration reproduced both the severe GSH depletion and the high serum ALT activity of the combined Jo2/APAP treatment (Fig. 9).
Taken together, our results suggest that iNOS induction by Jo2, severe ATP depletion with the dual Jo2/APAP treatment, and more severe GSH depletion with this dual treatment than with APAP alone could all contribute to the increased hepatotoxicity of APAP in Jo2-treated mice. It is noteworthy that these factors have mutually aggravating effects. The induction of iNOS by Jo2 may increase peroxynitrite formation and aggravate ATP depletion. ATP depletion aggravates GSH depletion by decreasing GSH synthesis. Finally, GSH depletion aggravates ATP depletion by causing more NAPQI covalent binding to proteins and more nitrotyrosine adduct formation in mitochondrial proteins.24 These mutually aggravating effects, together with the existence of critical thresholds in the extent of GSH and/or ATP depletion that switch the cell's fate from survival to APAP-induced necrosis, could explain why even subliminal Fas stimulation can increase the area of APAP-mediated hepatic necrosis. After administration of APAP alone (500 mg/kg), these critical thresholds may only be crossed in pericentral hepatocytes but not in peripheral cells (with less CYPs than pericentral hepatocytes), thus restricting necrosis to pericentral hepatocytes. However, the mild added effects of a Jo2 pretreatment may also cause these critical thresholds to be crossed in more peripheral cells, including mediolobular hepatocytes, thus enlarging the area of APAP-induced necrosis.
Like APAP, bromobenzene is transformed by CYP into reactive electrophilic metabolites that deplete hepatic GSH.22 In mice pretreated with the Jo2 antibody (1 μg per mouse) 1 hour before bromobenzene administration, hepatic GSH 4 hours after bromobenzene administration was 51% lower than in mice treated with bromobenzene alone, and serum ALT activity at 24 hours was 47-fold higher (Fig. 10). These observations suggest that the enhancement of drug-induced hepatitis by subliminal Fas stimulation could be a general property of drugs transformed into electrophilic metabolites causing GSH depletion.
Not all drugs whose hepatotoxicity is increased by viral infections1–13 are transformed into reactive metabolites. APAP,21 isoniazid45 and buprenorphine46 form reactive metabolites (and azathioprine depletes GSH),47 but aspirin, ibuprofen, and nucleoside reverse transcriptase inhibitors are mostly toxic through their mitochondrial effects.48 Because Fas stimulation damages mitochondria,16 it could have additive hepatotoxic effects with drugs that impair mitochondrial function.
Besides Fas, several other death receptors are stimulated during viral infections14, 17 and might also increase drug-induced toxicity. Indeed, additive hepatotoxic effects have been reported in vitro between TNF-α and salicylate.49 Finally, several viral proteins target mitochondria50; this might also aggravate drug-induced toxicity.
In conclusion, the administration of a subtoxic dose of an agonistic anti-Fas antibody before APAP administration increases APAP-induced ATP depletion, GSH depletion, and centrilobular necrosis, and it also increases bromobenzene-induced GSH depletion and toxicity. Subliminal Fas stimulation could be one mechanism whereby minor viral infections can increase drug-induced toxicity.