Pretreatment of Mice With Macrophage Inactivators Decreases Acetaminophen Hepatotoxicity and the Formation of Reactive Oxygen and Nitrogen Species

Transcription

1 Pretreatment of Mice With Macrophage Inactivators Decreases Acetaminophen Hepatotoxicity and the Formation of Reactive Oxygen and Nitrogen Species SHERRYLL L. MICHAEL, 1 NEIL R. PUMFORD, 1 PHILIP R. MAYEUX, 1 MICHAEL R. NIESMAN, 2 AND JACK A. HINSON 1 Abbreviations: NAPQI, N-acetyl-P-benzoquinone imine; GSH, glutathione; ALT, alanine transaminase; Ig, immunoglobulin; AST, aspartate transaminase; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis; IV, intravenously; IP, intraperitoneally. From the 1 Department of Pharmacology and Toxicology and the 2 Department of Ophthalmology, University of Arkansas for Medical Sciences, Little Rock, AR. Received September 16, 1998; accepted April 13, Supported by the National Institutes of Health (R01 GM and R01 GM to J.A.H. and DK to P.R.M.). Dr. Michael R. Niesman s present address is: Agouron Pharmaceuticals, 4245 Sorrento Valley Boulevard, San Diego, CA Address reprint requests to: Jack A. Hinson, Ph.D., Division of Toxicology, Slot 638, University of Arkansas for Medical Sciences, Little Rock, AR HinsonJackA@exchange.UAMS.edu; fax: Copyright 1999 by the American Association for the Study of Liver Diseases /99/ $3.00/0 186 Hepatotoxic doses of acetaminophen to mice produce not only acetaminophen-protein adducts in the centrilobular cells of the liver, but nitrotyrosine-protein adducts in the same cells, the site of the necrosis. Nitration of tyrosine occurs with peroxynitrite, a species formed by reaction of nitric oxide (NO ) with superoxide (O 2 ). Because NO and O 2 may be produced by activated Kupffer cells and/or infiltrated macrophages, we pretreated mice with the macrophage inactivators/depeleters gadolinium chloride (7 mg/ kg, intravenously [iv]) or dextran sulfate (10 mg/kg, iv) 24 hours before administration of acetaminophen (300 mg/kg). Mice treated with acetaminophen plus gadolinium chloride, or acetaminophen plus dextran sulfate, had significantly less evidence of hepatotoxicity as evidenced by lower serum alanine transaminase (ALT) levels (28 1 IU/L and IU/L, respectively) at 8 hours compared with acetaminophen (6, IU/L). Analysis of hepatic homogenates for acetaminophen-protein adducts at 2 hours, a time of maximal covalent binding and before hepatocyte lysis, indicated that these pretreatments did not decrease covalent binding. Western blot analysis for the macrophage marker protein F4/80 in homogenates revealed not only the expected decrease by the macrophage inactivators/depleters, but also an apparent increase in acetaminophenonly treated mice. At 8 hours nitrotyrosine-protein adducts were present in the acetaminophen-only treated mice, but not in the acetaminophen plus gadolinium chloride treated mice, or acetaminophen plus dextran sulfate treated mice. High levels of heme-protein adducts, a measure of oxidative stress, were detected in livers of the 8 hour acetaminophenonly treated mice. These data suggest that acetaminophen hepatotoxicity is mediated by an initial metabolic activation and covalent binding, and subsequent activation of macrophages to form O 2,NO, and peroxynitrite. Nitration of tyrosine correlates with toxicity. (HEPATOLOGY 1999;30: ) In overdose, the analgesic/antipyretic acetaminophen (paracetamol) produces hepatic necrosis in the centrilobular cells in the liver. 1 The toxicity has been shown to be mediated by cytochrome P450 metabolism to N-acetyl-p-benzoquinone imine (NAPQI). 2 After a therapeutic dose of acetaminophen the metabolite is preferentially detoxified by glutathione (GSH). However, after an overdose, hepatic GSH is depleted and NAPQI covalently binds to cysteine residues on proteins as 3-(cystein-S-yl)acetaminophen adducts. 3,4 Immunochemical evidence has shown that the histological site of binding as well as the relative amount of covalent binding correlate with the development of the toxicity. 5-8 Even though the prevailing theory of toxicity is that NAPQI covalently binds to critical proteins leading to inactivation of function, 9 a growing body of literature suggests that Kupffer cells (hepatic macrophages) and/or infiltrated macrophages are important in the toxicity. Activation of macrophages may lead to the formation of various potentially toxic species and these species may be important in the toxicity. 10,11 The involvement of macrophages in acetaminophen hepatotoxicity was originally shown by determining the effect of compounds known to inactivate and deplete macrophages. Thus, gadolinium chloride 12 and dichloromethylenediphosphonate 13 have been shown to dramatically decrease acetaminophen toxicity in mice. Also, Laskin et al. 14 reported that gadolinium chloride and dextran sulfate decrease hepatotoxicity in rats. Available data suggest that reactive nitrogen and oxygen species may be important in the hepatotoxicity. In earlier studies, it was reported that administration of encapsulated superoxide dismutase to rats decreased their susceptibility to the toxic effects of acetaminophen. 15 Recently, we showed that positively stained adducts occur in the centrilobular areas of the livers of mice treated with toxic doses of acetaminophen by using an anti-nitrotyrosine antibody in an immunohistochemical assay. 16 We interpreted these data to indicate that nitrotyrosine-protein adducts were formed at the cellular site of toxicity by nitration of the phenolic residue of tyrosine by peroxynitrite, a highly reactive product formed by reaction of nitric oxide and superoxide. This product is not only capable of nitratation reactions, but it also has hydroxyl radical-like activity In these mice there was a direct correlation between nitric oxide synthesis, as measured

2 HEPATOLOGY Vol. 30, No. 1, 1999 MICHAEL ET AL. 187 TABLE 1. Effect of Macrophage Inactivators on Acetaminophen-Induced Hepatotoxicity Treatment ALT (IU/L) AST (IU/L) Acetaminophen 6, * 10, * Saline Acetaminophen gadolinium chloride 28 1* 61 7* Gadolinium chloride Acetaminophen dextran sulfate * * Dextran sulfate NOTE. Mice were treated with either gadolinium chloride (7 mg/kg, iv), dextran sulfate (10 mg/kg, iv), or saline (iv) 24 hours before acetaminophen (300 mg/kg, ip). Mice (n 5) were killed at 8 hours after acetaminophen administration. Data are presented as mean SE. *Significantly different from comparable control (P.05). Significantly different (P.05) from acetaminophen-only treated mice. by serum levels of nitrate plus nitrite, and hepatotoxicity as determined by serum alanine transaminase (ALT) levels. Also, recently it was reported that toxic doses of acetaminophen lead to induction of inducible nitric oxide synthetase in the rat and that pretreatment of mice with the nitric oxide synthetase inhibitor aminoguanidine decreases the toxicity. 20 In this study we have treated mice with compounds that inactivate and deplete macrophages to determine the role of these cells in hepatotoxicity of acetaminophen, the potential effect on metabolic activation, and the formation of nitrotyrosine-protein adducts. These data indicate that treatment of mice with compounds known to inactivate macrophages do not decrease covalent binding but decrease the hepatotoxicity and nitrotyrosine-protein adduct formation. MATERIALS AND METHODS Caution. Formaldehyde was used for tissue preservation. Formaldehyde has been reported to be a carcinogen. Hydrogen peroxide is caustic and was used in immunohistochemical analysis. All operations were performed with caution, and with volatile chemicals the operations were performed in an efficient fume hood. Reagents. Acetaminophen, gadolinium chloride, and dextran sulfate were obtained from Sigma Chemical Company (St. Louis, MO). Rat anti-mouse F4/80 antigen and F(ab ) rabbit anti-rat immuno- A FIG. 1. Western blot analyses of hepatic homogenates for acetaminophen-protein adducts. The mice were pretreated with either saline, gadolinium chloride (GdCl 3 ), or dextran sulfate (DS) and subsequently with acetaminophen (APAP) (300 mg/kg) as described in the Materials and Methods section. The livers from each group of mice were combined, homogenized, and analyzed by Western blot analyses using an anti-acetaminophen antiserum. (A) Acetaminophen covalent binding after 2 hours, a time before hepatocyte lysis and of maximal covalent binding. (B) Acetaminophen covalent binding after 8 hours, a time after hepatocyte lysis and release of cytosolic acetaminophen adducts into serum. The bar graph shows the densitometric analysis of the blots. B

3 188 MICHAEL ET AL. HEPATOLOGY July 1999 FIG 2. Immunohistochemical analyses of liver sections for acetaminophen-protein adducts. The livers from the acetaminophen-treated mice reported in Table 1 (8 hours) were analyzed for acetaminophen-protein adducts as described in the Materials and Methods section. (A) Liver section from a mouse pretreated with saline and subsequently treated with acetaminophen (300 mg/kg, ip). (B) Liver section from a mouse pretreated with gadolinium chloride (7 mg/kg, iv) and subsequently treated after 24 hours with acetaminophen (300 mg/kg, ip). globulin-g (IgG) was purchased from Serotec Inc. (Raleigh, NC). We previously reported the raising of the anti-acetaminophen antiserum and its characterization. 21 Peroxidase-labeled goat antirabbit IgG (H L) was procured from Gibco BRL (Gaithersburg, MD). Universal DAKO LSAB (Labeled Streptavidin-Biotin) Peroxidase Kit and DAKO protein block (serum free) were acquired from DAKO Corporation (Carpinteria, CA). Precast 4% to 20% polyacrylamide mini gels were purchased from Owl Scientific (Weyburn, MA). Enhanced chemiluminescence (ECL) kit and ECL Hyperfilm were obtained from Amersham Life Science, Inc. (Arlington Heights, IL). SuperSignal Substrate, Immunopure Peroxidase Suppresser, and Coomassie Plus Protein Assay Reagent were purchased from Pierce Chemical Company (Rockford, IL). Animals. All animal experiments were approved by the University of Arkansas for Medical Sciences Animal Care and Use Committee. All animals received humane care according to the criteria outline in the Guide for the Care and Use of Laboratory Animals prepared by the National Academy of Sciences and published by the National Institutes of Health (NIH publication revised 1985). Eightweek-old male B6C3F1 mice having an average weight of 28 g were obtained from Harlan Sprague-Dawley (Indianapolis, IN). Animals were housed five per cage in clear plastic cages. Mice were fed ad libitum and were maintained on a 12-hour light/dark cycle. For the experiment to determine the effects of macrophage inactivators on toxicity, mice were randomly placed into six treatment groups (n 5 per group) and administered 0.1 ml of either saline, gadolinium chloride (7 mg/kg intravenously [iv]), or dextran sulfate (10 mg/kg iv), 24 hours before subsequent treatment. The mice were then fasted overnight. Experimental mice received 0.4 ml saline or a hepatotoxic dose of acetaminophen (300 mg/kg, intraperitoneally [ip]) and control mice received only saline. At 8 hours the mice were anesthetized with CO 2 and blood was taken from the retro-orbital sinus. The blood was allowed to coagulate at room temperature and the samples were then centrifuged. The serum was removed and stored at 4 C before analysis. Immediately after bleeding, the mice were killed and their livers were removed. A small section was removed from each liver and placed in 10% neutral buffered formalin to be used in immunohistochemical analysis. The remaining portion of each liver was weighed and homogenized in a 3:1 volume/weight ratio of 0.25 M sucrose, 10 mm HEPES, 1 mmol/l ethylenediaminetetraacetic acid (ph 7.5) buffer. The samples in each of the six treatment groups were then pooled by equal volumes, and protein concentration was determined using Pierce Coomassie Plus Protein Assay Reagent. Other aliquots of the homogenate were stored at 80 C. In a separate experiment, the effect of the macrophage inactivators on acetaminophen covalent binding was

4 HEPATOLOGY Vol. 30, No. 1, 1999 MICHAEL ET AL. 189 FIG. 2 (Cont d.) (C) Liver section from a mouse pretreated with dextran sulfate (10 mg/kg, IV) and subsequently treated after 24 hours with acetaminophen (300 mg/kg, IP). (D) Liver section from a saline-treated control mouse. The brown stain in the central vein (A) indicates the presence of acetaminophen-protein adducts. determined. This experiment was similar to the one described earlier except the mice (n 3) were sacrificed at 2 hours. The serum ALT and aspartate transaminase (AST) levels of these mice were the same as the levels of the saline control mice, which indicated that hepatocyte lysis had not occurred. Hepatotoxicity Assay. Serum ALT levels and serum AST levels were used as indicators of hepatotoxicity. These assays were performed by using a diagnostic kit obtained from Sigma Chemical Company in accordance with the method specified in the kit. Immunohistochemistry. Paraffin-embedded tissue sections were deparafinized with xylene (2 5 minutes, 25 C) then rehydrated in a series of graded ethanol washes and deionized H 2 O. The sections were then placed in Pierce Immunopure Peroxidase suppresser for 30 minutes at room temperature to quench endogenous peroxidase activity. Cells were then permeabilized by the addition of 0.15% Triton-X 100 in phosphate-buffered saline for 20 minutes. Next, a DAKO protein block was added to each tissue section for 30 minutes to block nonspecific binding. After washing in phosphate buffered saline, the sections were incubated with primary antibody (either antiacetaminophen antisera 1:1,000 or anti-nitrotyrosine 1:100) for 60 minutes at room temperature. Color development followed the suggested protocol described in the DAKO LSAB kit. The slides were counterstained with Gills Hematoxylin II (Fisher Scientific, Inc., Pittsburgh, PA) for 2 minutes and after rinsing in deionized H 2 O were immersed in ammonia blue for 2 minutes. The slides were dehydrated and mounted with Permount (Fisher Scientific, Inc.). Western Blot and Densitometric Analysis. Liver homogenate (150 µg) was separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) under reducing conditions and transferred to nitrocellulose membranes. Membranes were blocked overnight at 4 C. Membranes were then incubated in either anti-acetaminophen 21 antisera (1:2,000) for 90 minutes or F4/80 antibody (1:100) for 120 minutes. Membranes being stained for the presence of acetaminophen-protein adducts were next incubated with peroxidaselabeled goat anti-rabbit IgG (1:3,000) for 90 minutes. The membranes being stained for the presence of the F4/80 epitope were first incubated in rabbit anti-rat IgG (1:1,000) for 90 minutes followed by incubation in peroxidase-labeled goat anti-rabbit IgG (1:3,000) for 90 minutes. All membranes were visualized using ECL and exposure to ECL Hyperfilm. Densitometric analysis of the film was performed using a Model GS-670 imaging densitometer (Bio-Rad Laboratories, Hercules, CA) in transmittance mode and analyzed using Bio-Rad Molecular Analyst image analysis software. Peak area was determined using volume integration of equal band areas. Heme Bound-Protein Adduct Analysis. Liver homogenate (100 µg) was separated by SDS-PAGE under nonreducing conditions and transferred to nitrocellulose membranes following the method of

5 190 MICHAEL ET AL. HEPATOLOGY July 1999 Vuletich and Osawa. 22 Membranes were then incubated with Super- Signal Substrate and exposed to ECL Hyperfilm. RESULTS To determine the role of macrophages in the hepatotoxicity of acetaminophen, mice were pretreated with either gadolinium chloride (7 mg/kg, iv), dextran sulfate (10 mg/kg, iv), or saline (iv). After 24 hours, the experimental groups were treated with a toxic dose of acetaminophen (300 mg/kg, ip). The effect of acetaminophen on hepatotoxicity was determined at 8 hours. We have previously shown that in this animal model maximal hepatotoxicity occurs by 8 hours. 6,23 Acetaminophen hepatotoxicity was determined by measurement of serum ALT and AST levels. As shown in Table 1, mice that were pretreated with saline and subsequently treated with acetaminophen developed significant evidence of hepatotoxicity. However, mice that were pretreated with the macrophage inactivators dextran sulfate or gadolinium chloride and subsequently treated with acetaminophen showed either significantly less evidence of hepatotoxicity (dextran sulfate) or no evidence of hepatotoxicity (gadolinium chloride) (Table 1). Another experiment was performed to determine if the mechanism by which dextran sulfate or gadolinium chloride decreased the hepatotoxicity of acetaminophen was by an alteration of covalent binding. This experiment was similar to the one described in Table 1, except the mice were killed at 2 hours, a time of maximal covalent binding and before significant hepatocyte lysis. 6,23 Analysis of serum for ALT and AST levels indicated no evidence of hepatotoxicity (hepatocyte lysis) at 2 hours in any of the treatment groups. Thus, all serum ALT levels in treated mice at 2 hours were not significantly different from comparable controls. Covalent binding of acetaminophen to proteins was determined in combined hepatic homogenates of the mice in each group by Western blot analysis using an anti-acetaminophen antiserum. The covalent binding in the 2 hour treated data are shown in Fig. 1A, along with the densitometric analysis of the blot. As shown, the mice treated with acetaminophen, acetaminophen plus dextran sulfate, and acetaminophen plus gadolinium chloride had similar levels of acetaminophen covalently bound to the proteins. Thus, the mechanism by which dextran sulfate and gadolinium chloride pretreatment decreased the hepatic toxicity of acetaminophen was not by decreasing metabolic activation or increasing detoxification. If these treatments had decreased metabolic activation or increased detoxification, then an increased level of covalent binding would have been observed in the acetaminophenonly treated mice compared with the mice pretreated with the macrophage inactivators and acetaminophen. Figure 2 shows the immunohistochemical analysis of the 8-hour homogenates for acetaminophen-protein adducts. Figure 2A is a representative liver section from an acetaminophen-only treated mouse. As shown in this section, the majority of the hepatocytes in the centrilobular area were lysed and significant levels of acetaminophen-protein adducts were observed in a few nonlysed hepatocytes in the midzonal regions. These data are in agreement with our previous findings. 6 Figure 2B is a representative liver section from the acetaminophen plus gadolinium chloride treated mice. Consistent with the low ALT levels, hepatocyte necrolysis was not observed and acetaminophen-protein adduct levels were high. Figure 2C is from the acetaminophen plus dextran sulfate treated mice. Only a small number of lysed hepatocytes were observed and acetaminophen-protein adduct levels were high in the nonlysed hepatocytes. The 8-hour hepatic homogenates from the acetaminophen, acetaminophen plus gadolinium chloride, and acetaminophen plus dextran sulfate treated groups were analyzed by Western blot analyses for acetaminophen-protein adducts. Figure 1B shows that the relative amount of acetaminophenprotein adducts in hepatic homogenates from the acetaminophen-only treated mice was much less than that observed in the hepatic homogenates from either the acetaminophen plus gadolinium chloride or the acetaminophen plus dextran sulfate treated mice. These data are consistent with the immunohistochemical finding in Fig. 2. Thus, hepatocyte lysis, as occurs in the acetaminophen-only treated mice at 8 hours, results in loss of hepatic acetaminophen-protein adducts and a large increase in serum ALT levels. In the acetaminophen plus gadolinium chloride treated mice and the acetaminophen plus dextran sulfate treated mice the livers have high levels of acetaminophen-protein adducts and a minor degree of hepatocyte lysis. In previous reports we showed that hepatocyte lysis results in release of hepatic acetaminophen-protein adducts into serum. 23 Hepatic homogenates were also analyzed by Western blot analysis and densitometric analysis for the macrophage/ Kupffer cell epitope F4/80 (Fig. 3). F4/80 is a macrophagespecific protein (160 kd) of unknown function that has epidermal growth factor like domains and shows homology to the seven transmembrane-spanning family of hormone receptors As shown in Fig. 3, pretreatment of mice with gadolinium chloride or dextran sulfate alone decreased the relative amount of this epitope in the liver compared with saline pretreated mice by approximately 50%. Thus, as anticipated, these treatments decreased the hepatic macrophage (Kupffer cell) population. Interestingly, at 2 hours the relative amount of the F4/80 epitope in the acetaminophenonly treated mice appeared to increase by approximately 40% compared with saline-treated mice (Fig. 3A). Also, a similar increase was observed in the hepatic homogenates from acetaminophen plus dextran sulfate treated mice, a treatment group in which toxicity was observed, compared with dextran sulfate only. Two hours is a time immediately before hepatocyte lysis occurs and is a time of engorgement of the liver by blood. Thus, in the mice in which toxicity is subsequently observed, there is an increased macrophage, or possibly monocyte, population in the liver at 2 hours. In recent studies, we immunochemically analyzed liver sections from acetaminophen-treated mice using an antibody raised against nitrotyrosine. A positive stain was observed in the centrilobular cells of the liver after a toxic dose of acetaminophen in mice. The localization of these adducts was the same as the acetaminophen-protein adducts. 16 Nitrotyrosine-protein adducts are formed via nitration by peroxynitrite, a reactive substance formed by a very rapid reaction between nitric oxide and superoxide Because both nitric oxide and superoxide may be formed by activated macrophages, 27 we determined whether the gadolinium chloride plus acetaminophen treated mice had nitrotyrosine-protein adducts at 8 hours using immunohistochemical analyses. As shown earlier, gadolinium chloride pretreated mice did not cause an increase in serum levels of ALT or AST (Table 1) or alter metabolic activation (Fig. 1). Acetaminophen-only treated mice had high levels of nitrotyrosine-protein adducts

6 HEPATOLOGY Vol. 30, No. 1, 1999 MICHAEL ET AL. 191 FIG 3. Western blot analyses of liver homogenates for the macrophage epitope F4/80. The mice were pretreated with either saline, gadolinium chloride (GdCl 3 ), or dextran sulfate (DS) and subsequently treated with acetaminophen (APAP) (300 mg/kg) as described in the Materials and Methods section and as reported in Table 1. (A) Western blot analysis of liver homogenates from the mice sacrificed at 2 hours after acetaminophen administration. The serum ALT levels of these mice were the same as the ALT levels in the saline-treated control mice. (Fig. 4A) whereas acetaminophen plus gadolinium chloride treated mice did not have detectable levels of nitrotyrosineprotein adducts (Fig. 4B); however, tinctorial changes in the relative amount of background stain are evident in the centrilobular areas (Fig. 4B). Similarly, nitrotyrosine-protein adducts were not detected in the acetaminophen plus dextran sulfate treated mice (Fig. 4C) even though these mice had minor evidence of hepatotoxicity. Macrophage activation may lead to increases in peroxynitrite formation (superoxide reacting with nitric oxide), as well as hydrogen peroxide formation (superoxide dismutation). To assess peroxide formation during acetaminophen hepatotoxicity we used a recently developed assay for hemeprotein adducts, a reaction that occurs by oxidation of some heme proteins with peroxides. 22 These heme-protein adducts have peroxidase activity and this activity may be detected by ECL assay after SDS-PAGE separation under nonreducing conditions. 22 As shown in Fig. 5, hepatic homogenates of acetaminophen-only treated mice had much greater levels of heme-protein adducts at 8 hours than hepatic homogenates in the saline-treated mice. The livers from the acetaminophen plus gadolinium chloride treated mice as well as the acetaminophen plus dextran sulfate treated mice had fewer hemeprotein adducts. Thus, acetaminophen hepatotoxicity is accompanied by oxidative stress and conditions that decrease macrophage function decrease toxicity and oxidative stress. DISCUSSION Even though metabolic activation of acetaminophen has been shown to be a correlate of acetaminophen-induced hepatotoxicity, 4-7 recent studies suggest that Kupffer cells (resident hepatic macrophages) and/or infiltrated macrophages may also be important ,16 It has been shown that treatment of rats or mice with inactivators of macrophages dramatically decrease the hepatotoxicity Macrophages may be activated to synthesize increased levels of superoxide and nitric oxide. 10,27 Whereas dismutation of superoxide forms hydrogen peroxide, reaction of superoxide with nitric oxide forms peroxynitrite, a reactive product that nitrates tyrosine and leads to the hydroxyl radical We recently presented evidence that these adducts are formed in the livers of acetaminophen-treated mice by using an antibody raised against nitrotyrosine in an immunohistochemical analysis. 16 The histological localization of these nitrotyrosine-protein adducts correlated exactly with the cellular localization of acetaminophen-protein adducts and toxicity. Thus, there was an apparent increased synthesis of nitric oxide and superoxide at the site of the metabolic activation and the toxicity. These increased syntheses are presumably caused by activation of Kupffer cells and/or infiltrated macrophages. Moreover, we observed increased serum levels of nitrate plus nitrite, another indicator of increased nitric oxide synthesis in acetaminophen hepatotoxicity. Most importantly, the increased serum levels of nitrate plus nitrite in individually treated mice significantly correlated with the relative hepatotoxicity as measured by serum ALT levels. 16 In this report we have extended these findings and examined the relationship among nitrotyrosine-protein adducts, oxidative events in the liver as measured by formation of heme-protein adducts, and the effect of macrophage inactivators on the formation of these adducts and toxicity. As shown in Table 1, treatment of mice with the macrophage inactivators gadolinium chloride or dextran sulfate dramatically decreased the hepatotoxicity of acetaminophen. Analysis of serum ALT levels indicated that dextran sulfate decreased the hepatotoxicity by approximately 90%, whereas gadolinium chloride pretreatment of mice completely protected the mice from the hepatotoxicity as measured by serum ALT levels. Because the relative levels of covalent binding to protein at 2 hours in the hepatic homogenates were not lower in the acetaminophen plus gadolinium chloride treated mice, or in the acetaminophen plus dextran sulfate treated mice compared with the acetaminophen-only treated mice (Fig. 1), the mechanism of decreased toxicity is not by decreased metabolic activation or increased detoxification. If the decrease in toxicity was a result of inhibition of P-450 metabolic activation to form NAPQI, or an increase in glutathione detoxification of NAPQI, then a decreased amount of covalent binding would have been observed. However, as shown in Fig. 4, even though the mice treated with gadolinium chloride plus acetaminophen did not have nitrotyrosineprotein adducts, there were changes in the background dye staining in the centrilobular hepatocytes of these treated mice. These tinctorial changes indicate that alterations have occurred in the proteins that affect the staining and these alterations may be a result of covalent binding to nucleophilic sites on the proteins. Gadolinium chloride and dextran sulfate are believed to exert their effects by altering macrophages. It has been shown that gadolinium chloride treatment not only decreases the ability of macrophages to produce such substances as nitric oxide, 28 but selectively eliminates macrophages from the liver. 29 The relative amount of the macrophage epitope F4/80 in the hepatic homogenate was investigated. F4/80 is a plasma membrane glycoprotein with a molecular weight of approximately 160 kd that is expressed by mature mouse macrophages. 24,25 It has been cloned and shown to be a novel member of the G-protein linked transmembrane seven hor-

7 192 MICHAEL ET AL. HEPATOLOGY July 1999

8 HEPATOLOGY Vol. 30, No. 1, 1999 MICHAEL ET AL. 193 mone receptor family and has sequences related to those in epidermal growth factor. 26 Western blot analyses of this epitope indicated that mice pretreated with only gadolinium chloride or only dextran sulfate had less F4/80 epitope than mice treated with saline (Fig. 3). Thus, these data are consistent with gadolinium chloride and dextran sulfate decreasing toxicity by decreasing Kupffer cells. Interestingly, in the acetaminophen-treated mice, the amount of F4/80 epitope appeared to increase compared with saline at 2 hours (Fig. 3). Activation of Kupffer cells does not increase the F4/80 epitope. 24 It is possible that this indicates infiltrated macrophages. If this is correct, then infiltrated macrophages may play a major role in acetaminophen hepatotoxicity; however, in an earlier study we did not observe histologically recruited macrophages until 24 hours. 6 Another possible factor relative to understanding this increase is that the increase may be associated with hepatic engorgement of blood that occurs before hepatocyte lysis. 30 Also, monocytes in blood have the F4/80 epitope and thus the increase may be reflective of increased levels of monocytes in the liver. 31 Kupffer cells may produce toxicity by activation leading to increased formation of a variety of bioactive substances including cytokines, chemokines, growth factors, cell adhesion proteins, and cytokine receptors. In addition, Kupffer cell activation can lead to increased levels of nitric oxide and superoxide. 10,27,32 Because tumor necrosis factor null mice are equally susceptible to the toxic effects of acetaminophen as the wild type mice, presumably this factor is not involved. 33 In our present report we show that the relative hepatotoxicity correlated with nitration and oxidation events occurring in the liver; however, other factors that are formed simultaneously may contribute to the toxicity. Figure 4 shows that nitrotyrosine-protein adducts correlate with development of the toxicity. Nitrotyrosine-protein adducts were present in the centrilobular areas of livers of mice treated with acetaminophen only (Fig. 4A); however, in mice pretreated with gadolinium chloride and subsequently with acetaminophen, nitrotyrosine-protein adducts were not observed (Fig. 4B). Similarly, in mice pretreated with dextran sulfate and subsequently with acetaminophen, nitrotyrosine-protein adducts were not detectable (Fig. 4C). Thus, nitration of tyrosine on proteins correlated with toxicity. Also, the formation of heme-protein adducts correlated with the development of the toxicity. Heme-protein adducts were previously shown to be formed when myoglobin was incubated with hydrogen peroxide and thus these adducts may be a result of the increased levels of superoxide leading to increased levels of hydrogen peroxide. 22 Because NAPQI formation leads to glutathione depletion, 3 and mammalian GSH peroxidase has been reported to have a relatively high K m ( 3 mmol/l), 34 an FIG. 5. SDS-PAGE analyses of hepatic homogenates for hepatic hemeprotein adducts. The mice were pretreated as described in the Materials and Methods section and in the legend to Table 1. Liver homogenate (100 µg) were separated by SDS-PAGE under nonreducing conditions and transferred to nitrocellulose membranes. Peroxidase activity was detected using the ECL assay. The most intense staining band has a molecular weight of approximately 16 kd and the less intense staining band has a molecular weight of 31 kd. Densitometric analysis of each band was performed and the data are presented as percent of control activity. impairment of hydrogen peroxide detoxification may occur in the affected cells in acetaminophen hepatotoxicity. Hemeprotein catalyzed peroxidation may be important in the toxicity or at least a marker of oxidative stress. Also, it is ; FIG. 4. Immunohistochemical analyses of liver sections for hepatic nitrotyrosine-protein adducts. Mice were killed 8 hours after treatment. Liver sections were stained for nitrotyrosine as described in the Materials and Methods section and counter stained with Gills Hematoxylin II. (A) Liver section from a mouse pretreated with saline and subsequently treated with acetaminophen (300 mg/kg, ip). (B) Liver section from a mouse pretreated with gadolinium chloride (7 mg/kg, iv) and subsequently treated after 24 hours with acetaminophen (300 mg/kg, ip). (C) Liver section from a mouse pretreated with dextran sulfate (10 mg/kg, iv) and subsequently treated after 24 hours with acetaminophen (300 mg/kg, ip). (D) Liver section from a saline-treated control mouse. The brown stain in the central vein (A) indicates the presence of nitrotyrosine-protein adducts. FIG. 6. Postulated mechanism of acetaminophen-induced hepatotoxicity. Metabolic activation includes all of the events occurring by acetaminophen metabolism to NAPQI. These events include GSH depletion, covalent binding of NAPQI to protein as acetaminophen-protein adducts, and superoxide formation by cytochrome P4502E1.

9 194 MICHAEL ET AL. HEPATOLOGY July 1999 noted that GSH peroxidase is believed to be important in the detoxification of peroxynitrite 35 and is inhibited by peroxynitrite. 36 Thus, alteration of GSH peroxidase activity may be a critical factor in nitrotyrosine-protein adduct formation, hydrogen peroxide detoxification, and toxicity. Lastly, it is unclear if peroxynitrite formation has any role in hemeprotein adduct formation. At present the exact cells important in reactive oxygen and nitrogen generation is unknown. Kupffer cells are localized in the hepatic sinusoids, and thus production of reactive nitrogen and oxygen intermediates by this cell is anticipated to occur outside the hepatocytes. This may be important in the recent report that sinusoidal endothelial cell toxicity occurs during acetaminophen-induced hepatotoxicity in vivo. 37 Most importantly, it has been shown by a large number of investigators that acetaminophen is toxic to isolated hepatocytes and these hepatocytes presumably do not contain macrophages. Thus, acetaminophen toxicity may occur by more than one mechanism; however, both mechanisms may involve reactive nitrogen and oxygen species. Collectively the data in this report, our previous report, 16 and the previous reports by Blazka et al. 12 and Goldin et al. 13 in the mouse, and the reports by Laskin et al. 14 and Gardner et al. 20 in the rat, strongly suggest that acetaminophen-induced hepatotoxicity is mediated by a two-step process in rodents. In the first step, acetaminophen is metabolically activated to NAPQI. NAPQI preferentially conjugates with GSH at low doses of acetaminophen; however, after large doses, GSH is depleted and NAPQI covalently binds to protein. As a result of metabolic events, macrophage activation is postulated to occur. Macrophage activation leads to formation of increased levels of nitric oxide and superoxide that may react together to produce peroxynitrite. Peroxynitrite nitrates tyrosine resides in the proteins of affected cells and has been reported to have hydroxyl radical-like activity. 17,18 In addition, superoxide dismutation leads to increased levels of hydrogen peroxide and this may be important in the observed increase in heme-protein adducts. Heme-protein adducts may catalyze various peroxidation reactions. Also, macrophage activation may lead to increased synthesis of various inflammatory cytokines and other substances. 10,27 These cumulative events are postulated to be important in the toxicity (Fig. 6). Our previous immunohistochemical analyses for acetaminophenprotein adducts indicated a very high correlation between adduct formation and toxicity. 6 Thus, the cells containing acetaminophen-protein adducts are the most susceptible to cell killing. If hepatocyte lysis were mediated by peroxynitrite it would be anticipated that these cells may be highly susceptible to cell killing by this mechanism because peroxynitrite is detoxified by GSH 35 and GSH is depleted in these cells by NAPQI. 3 This susceptibility may explain why we previously found the very high correlation between acetaminophen adduct formation and the development of toxicity. 6 Additional studies are needed to clarify all of the cell types important in acetaminophen toxicity. Even though data indicate a major role for macrophages, the exact role of Kupffer cells versus infiltrated macrophages is not clear. Previously, recruited macrophages were not observed earlier than 24 hours after a toxic dose of acetaminophen 6 ; however, the F4/80 data in Fig. 3 suggested that the macrophage content of the liver may have increased under toxic conditions and thus it is not clear which type of macrophages are important in the toxicity. Also, it is unclear if other cells may contribute to production of superoxide and nitric oxide. Hepatocytes 38 and sinusoidal endothelial cells 39 may produce nitric oxide, and cytochrome P450s may produce superoxide during acetaminophen metabolism. 40 Thus, to more fully understand acetaminophen hepatotoxicity it will be necessary to determine if endothelial cells and/or hepatocytes are contributors to production of reactive nitrogen or reactive oxygen, the importance of Kupffer cells versus infiltrated macrophages, and the relative importance of reactive oxygen, reactive nitrogen, and specifically the role of peroxynitrite versus other substances in the toxicity. REFERENCES 1. Hinson JA. Biochemical toxicology of acetaminophen. Rev Biochem Toxicol 1980;2: Dahlin DC, Miwa GT, Lu AYH, Nelson SD. N-Acetyl-p-benzoquinone imine: a cytochrome P-450-mediated oxidation product of acetaminophen. Proc Natl Acad Sci U S A 1984;81: Mitchell JR, Thorgeirsson SS, Potter WZ, Hashimoto M, Mitchell JR. Acetaminophen-induced hepatic necrosis. IV. Protective role of glutathione. J Pharmacol Exp Ther 1973;187: Jollow DJ, Mitchell JR, Potter WZ, Davis DC, Gillette JR, Brodie BB. Acetaminophen-induced hepatic necrosis. IV. Role of covalent binding in vivo. J Pharmacol Exp Ther 1973;187: Hinson JA, Pumford NR, Roberts DW. Mechanisms of acetaminophen toxicity: immunochemical detection of drug-protein adducts. Drug Metab Rev 1995;27: Roberts DW, Bucci TJ, Benson RW, Warbritton AR, McRae TA, Pumford NR, Hinson JA. Immunochemical quantitation of 3-(cystein-S-yl)acetaminophen protein adduct in acetaminophen hepatotoxicity. Am J Pathol 1991;138: Cohen SD, Khairallah EA. Selective protein arylation and acetaminopheninduced hepatotoxicity. Drug Metab Rev 1997;29: Bartolone JB, Sparks K, Cohen SD, Khairallah EA. Immunochemical detection of acetaminophen-bound liver proteins. Biochem Pharmacol 1987;36: Cohen SD, Pumford NR, Khairallah EA, Boekelheide K, Pohl LR, Amouzadeh HR, Hinson JA. Selective protein covalent binding and target organ toxicity. Toxicol Appl Pharmacol 1997;143: Laskin DL, Pendino KJ. Macrophages and inflammatory mediators in tissue injury. Ann Rev Pharmacol Toxicol 1995;35: Winwood PJ, Arthur MJP. Kupffer cells: their activation and role in animal models of liver injury and human liver disease. Semin Liver Dis 1993;13: Blazka ME, Germolec DR, Simeonova P, Bruccoleri A, Pennypacker KR, Luster MI. Acetaminophen-induced hepatotoxicity is associated with early changes in NF-kB and NF-IL6 DNA binding activity. J Inflamm 1996;47: Goldin RD, Ratnayaka ID, Breach CS, Brown IN, Wickramasinghe SN. Role of macrophages in acetaminophen (paracetamol)-induced hepatotoxicity. J Pathol 1996;179: Laskin DL, Gardner CR, Price VF, Jollow DJ. Modulation of macrophage functioning abrogates the acute hepatotoxicity of acetaminophen. HEPA- TOLOGY 1995;21: Nakae D, Yamamoto K, Yosiji H, Kinugasa T, Maruyama H, Farber JL, Konishi Y. Liposome-encapsulated superoxide dismutase prevents liver necrosis induced by acetaminophen. Am J Pathol 1990;136: Hinson JA, Pike SL, Pumford NR, Mayeux PR. Nitrotyrosine-protein adducts in hepatic centrilobular areas following toxic doses of acetaminophen in mice. Chem Res Toxicol 1998;11: Beckman JS, Koppenol WH. Nitric oxide, superoxide, and peroxynitrite: the good, the bad, and the ugly. Am J Physiol 1996;271:C1424-C Pryor WA, Squadrito GL. The chemistry of peroxynitrite: a product from the reaction of nitric oxide with superoxide. Am J Physiol 1995;268:L699- L Beckman JS, Ye YZ, Anderson PG, Chen J, Accavitti MA, Tarpey MM, White CR. Extensive nitration of protein tyrosines in human atherosclerosis detected by immunohistochemistry. Biol Chem Hoppe-Seyler 1994;375: Gardner CR, Heck DE, Yang CS, Thomas PE, Zhang XJ, DeGeorge GL, Laskin JD, et al. Role of nitric oxide in acetaminophen-induced hepatotoxicity in the rat. HEPATOLOGY 1998;27: Matthews AM, Roberts DW, Hinson JA, Pumford NR. Acetaminophen-

10 HEPATOLOGY Vol. 30, No. 1, 1999 MICHAEL ET AL. 195 induced hepatotoxicity: analysis of total covalent binding vs. specific binding to cysteine. Drug Metab Dispos 1996;24: Vuletich JL, Osawa Y. Detection of protein-bound heme adducts by enhanced chemiluminescence: a biomarker assay for oxidative damage [Abstract]. Toxicologist 1998;42 (suppl):219s. 23. Pumford NR, Hinson JA, Potter DW, Rowland KL, Benson RW, Roberts DW. Immunochemical quantitation of 3-(cystein-S-yl)acetaminophen adducts in serum and liver proteins of acetaminophen-treated mice. J Pharmacol Exp Ther 1989;248: Hirsch A, Gordon S. The use and limitation of monoclonal antibodies against mononuclear phagocytes. Immunobiology 1982;161: Hume DA, Perry VH, Gordon S. The mononuclear phagocyte system of the mouse defined by immunohistochemical localisation of antigen F4/80: macrophages associated with epithelia. Anat Rec 1984;210: McKnight AJ, MacFarlane AJ, Dri P, Turley L, Willis AC, Gordon S. Molecular cloning of F4/80, a murine macrophage-restricted cell surface glycoprotein with homology to the G-protein-linked transmembrane 7 hormone receptor family. J Biol Chem 1996;271: MacMicking J, Xie Q-W, Nathan C. Nitric oxide and macrophage function. Annu Rev Immunol 1997;15: Roland CR, Naziruddin B, Mohanakumar T, Glye MW. Gadolinium chloride inhibits Kupffer cell nitric oxide synthase (inos) induction. J Leukoc Biol 1996;60: Hardonk MJ, Dijkhuis FW, Hulstaert CE, Koudstaal J. Heterogeneity of rat liver and spleen macrophages in gadolinium chloride-induced elimination and repopulation. J Leukoc Biol 1992;52: Walker RM, Massey RE, McElligott TF, Racz WJ. Acetaminopheninduced hypothermia, hepatic congestion, and modification by N-acetylcysteine in mice. Toxicol Appl Pharmacol 1981;59: Lee SH, Starkey PM, Gordon S. Quantitative analysis of total macrophage content in adult mouse tissues. Immunochemical studies with monoclonal antibody F4/80. J Exp Med 1985;161: Hagen TM, Huang S, Curnutte J, Fowler P, Martinez V, Wehr CM, Ames BN, et al. Extensive oxidative DNA damage in hepatocytes of transgenic mice with chronic active hepatitis destined to develop hepatocellular carcinoma. Proc Natl Acad Sci U S A 1994;91: Boess F, Bopst M, Althaus R, Polsky S, Cohen SD, Eugster HP, Boelsterli UA. Acetaminophen hepatotoxicity in tumor necrosis factor/lymphotoxin-alpha gene knockout mice. HEPATOLOGY 1998;27: Scott DR, Karageuzian LN, Anderson PJ, Epstein DL. Glutathione peroxidase of calf trabecular meshwork. Invest Ophthalmol Vis Sci 1984;25: Sies H, Sharov VS, Klotz LO, Briviba K. Glutathione peroxidase protects against peroxynitrite-mediated oxidations. A new function for selenoproteins as peroxynitrite reductase. J Biol Chem 1997;272: Padmaja S, Squadrito GL, Pryor WA. Inactivation of glutathione peroxidase by peroxynitrite. Arch Biochem Biophys 1998;349: DeLeve LD, Wang X, Kaplowitz N, Shulman HM, Bart JA, van der Hoek A. Sinusoidal endothelial cells as a target for acetaminophen toxicity. Direct action versus requirement for hepatocyte activation in different mouse strains. Biochem Pharmacol 1997;53: Kuo PC, Slivka A. Nitric oxide decreases oxidant mediated hepatocyte injury. J Surg Res 1994;56: Alexander B. The role of nitric oxide in hepatic metabolism. Nutrition 1998;14: Dai Y, Cederbaum AI. Cytotoxicity of acetaminophen in human cytochrome P4502E1-transfected HepG2 Cells. J Pharmacol Exp Ther 1995;273: