Acetaminophen-induced Oxidation of Protein Thiols CONTRIBUTION OF IMPAIRED THIOL-METABOLIZING ENZYMES AND THE BREAKDOWN OF ADENINE NUCLEOTIDES*

the of Our portion of was to protein thiol

The administration of a hepatotoxic dose of acetaminophen (250 mg/kg) to mice induced the loss of protein thiols in mouse liver. Our data suggest that a significant portion of this loss was due to protein thiol oxidation.
The administration of the nonhepatotoxic regioisomer, 3'-hydroxyacetanilide (600 mg/kg) did not produce a similar decrease in liver protein thiols despite producing similar levels of covalent binding. Mice treated with acetaminophen exhibited decreased glutathione peroxidase activity, decreased thioltransferase activity, and decreased adenine nucleotide concentrations in the liver. The increase in urinary allantoin after the administration of acetaminophen suggests that the decrease in adenine nucleotides was due to their degradation in the liver. Acetaminophen also promoted the conversion of the enzyme xanthine dehydrogenase to the oxidase form, and pretreatment of mice with allopurinol, an inhibitor of xanthine oxidase, significantly decreased acetaminophen-mediated hepatotoxicity.
The conversion of xanthine dehydrogenase to the oxidase form may lead to a transient increase in the production of activated oxygen species. The increase in activated oxygen species coupled with decreases in glutathione peroxidase and thioltransferase activity may be responsible in part for the increased levels of oxidized protein thiols observed following acetaminophen administration.
Acetaminophen (4'-hydroxyacetanilide, APAP)' is commonly administered for the treatment of pain and fever. Overdoses of the drug can produce acute hepatic necrosis in humans (1,2) and experimental animals (3). Acetaminophen is metabolized to a reactive metabolite, N-acetyl-p-benzoquinone imine (NAPQI), by the enzyme cytochrome P-450 (4, 5). Once formed, this metabolite conjugates and depletes cellular glutathione (GSH) levels and then binds extensively to the sulfhydryl groups of cellular proteins (6). Despite considerable research, the mechanism responsible for APAPinduced hepatotoxicity remains unknown. Both the binding of reactive metabolites to critical cellular macromolecules (7) and the production of oxidative stress (8,9) have been proposed to account for the hepatotoxicity of APAP. The regioisomer of APAP, 3'-hydroxyacetanilide (AMAP), is metabolized by cytochrome P-450 to reactive metabolites.
* This work was supported by National Institutes of Health Grant GM25418.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ' The abbreviations used are: APAP, acetaminophen; AMAP, 3'. hydroxyacetanilide; ALT, alanine aminotransferase; NAPQI, N-acetyl-p-benzoquinone imine; DTT, dithiothreitol.
These metabolites also react with GSH and the sulfhydryl groups of proteins (10). However, AMAP, in contrast to APAP, is not hepatotoxic to mice (11). In experiments conducted in this laboratory (12), the administration of AMAP produced similar levels of covalent binding to mouse liver homogenates as did a hepatotoxic dose of APAP, yet no hepatotoxicity occurred. In addition, both AMAP and APAP depleted cytosolic GSH levels in the liver at these doses (12). An analysis of subcellular binding profiles indicates that APAP arylates mitochondrial proteins and depletes mitochondrial GSH to a greater extent than AMAP. APAP administration disrupted calcium homeostasis in the liver while no disruption was observed with AMAP (12). For example, plasma membrane calcium-ATPase activity was lower in APAP-treated animals and their ability to sequester calcium in mitochondria was impaired. These changes apparently lead to mitochondrial calcium cycling 1 h after APAP administration.
In contrast, the nonhepatotoxic regioisomer, AMAP, did not significantly affect either the ability of mitochondria to sequester calcium or the activity of the plasma membrane calcium-ATPases.
The mechanism responsible for the hepatotoxicity of APAP is examined further in the present paper. We report that APAP administration depletes cellular protein thiols by both arylation and oxidation. In addition, APAP leads to the conversion of the enzyme xanthine dehydrogenase to xanthine oxidase. This conversion, in conjunction with decreases in the activities of enzymes that affect cellular thiol homeostasis and the degradation of adenine nucleotides, may lead to a transient production of activated oxygen species and promote the oxidation of protein thiols. for protein thiols as before.
Urine allantoin was measured spectrophotometrically as described by Borchers (27).

Effects of APAP and AMAP on Protein
Thiol Levels- Table  I   returned to control levels in the AMAP-treated animals at 6 h.
Significant decreases were observed in protein thiol levels at 1 and 6 h following APAP administration.
At 1 h, protein thiol levels were 78.5% of controls. These levels remained depressed at the 6-h time point.
In a previous report (In), we quantified the levels of covalent binding to liver homogenates obtained from animals treated with these same doses of AMAP and APAP. Values of about 0.9 and 0.8 nmol bound per mg protein were detected in liver homogenates at 1 and 6 h, respectively, after APAP administration.
Similar values were obtained with AMAP. Levels of 1.0 and 0.6 nmol bound per mg protein were measured at 1 and 6 h, respectively, with this drug. Based on protein thiol determinations, a loss of 14.9 nmol of sulfhydryl groups/mg protein occurred 1 h after APAP administration. This value is about 15 times the loss expected from covalent binding alone.
Further experiments were conducted to determine the contribution of sulfhydryl group oxidation to the observed loss of protein thiols. To assess this contribution, liver homogenates were either sonicated or treated with sodium dodecyl sulfate in the presence of the sulfhydryl-reducing reagent DTT. These treatments were used to disrupt cellular membranes and expose disulfide groups. DTT was removed prior to protein thiol determinations.
Following these procedures, no significant decreases from control values were observed in the levels of protein thiols in the liver homogenates of either APAP-or AMAP-treated animals at 1 or 6 h (data not shown). This experiment suggests that oxidation accounts for a large fraction of the observed decrease seen in protein thiol levels following APAP administration.

Effects
of APAP and AMAP on Cytosolic GSH-metabolizing Enzymes-We have shown previously that both AMAP and APAP deplete cytosolic GSH levels 1 h after administration and that these levels rebound by 6 h (12). However, little is known about the effects of these agents on those GSHmetabolizing enzymes involved in the protection of the cell against oxidative stress. The effects of AMAP and APAP on cytosolic glutathione peroxidase activity are shown in Table  II. Some inhibition of the enzyme was detected after AMAP administration, but much greater decreases in activity occurred in APAP-treated animals. One h after APAP administration, glutathione peroxidase activity decreased to about 60% of control values. At 6 h, activity levels returned to about 70% of controls.  Effects of AMAP and APAP on Conversion of Xanthine Dehydrogenase to Xanthine Oxidase-Xanthine dehydrogenase is a widely distributed enzyme which is involved in the degradation of all purines. Studies indicate that sulfhydryl reagents and the oxidation of disulfides can convert this enzyme into an oxidase form (28). Following this conversion, the enzyme utilizes oxygen instead of NAD' as an electron acceptor during the metabolism of purines. The conversion of the enzyme to the oxidase form with the subsequent produc-tion of superoxide anion and hydrogen peroxide has been implicated in the pathogenesis of ischemia-reperfusion injury to tissues (28).
The effects of AMAP and APAP on the conversion of xanthine dehydrogenase to the oxidase form are shown in Table IV. It is widely accepted that under normal circumstances the enzyme exists almost entirely in the NAD+dependent dehydrogenase form (29). Our results indicate that only about 10% of the total xanthine dehydrogenase-oxidase activity exists in the oxidase form in control animals. This value agrees with the values obtained by other researchers (29). AMAP produced no significant alterations in the relative ratio of the two forms of the enzyme. However, there was a significant increase in the amount of the oxidase form present 1 h after APAP administration.
As much as 42% of the total enzyme activity was detected as the oxidase form. The magnitude of this conversion exceeds that reported to occur in an ischemia-reperfusion injury (29). After 6 h, the fraction of the enzyme present as the oxidase form returned to control levels.
The total amounts of the xanthine oxidase-dehydrogenase enzyme activity remained fairly constant between treatments. There was a slight rise in the total activity in the livers of APAP-treated animals, but this rise was not found to be significant.
Effects of AMAP and APAP on Hepatic Adenine Nucleotide and Inorganic Phosphate Pools-The main endogenous substrates of the enzyme xanthine dehydrogenase-oxidase are the breakdown products of purine metabolism. Since adenine nucleotides comprise the major labile purine pool of the cell, the status of this pool was examined. The effects of AMAP and APAP on the hepatic levels of adenine nucleotides and inorganic phosphate are reported in Table V. In control animals, the bulk of the adenine nucleotide pool exists as ATP. ADP was also measured as well as minor amounts of AMP. The total adenine nucleotide pool was about 3.4 Fmol/ g of liver, wet weight.
AMAP treatment moderately increased this total adenine nucleotide pool and lowered hepatic inorganic phosphate levels.
A much different effect was seen following APAP administration In APAP-treated animals, adenine nucleotide pools were significantly decreased below control levels. This decrease was greatest at 1 h following the administration of the  decreased to 53.4% of controls. The major nucleotide which contributed to this loss was ATP. At 6 b, the total adenine nucleotide pool recovered to some degree, but it was still significantly depressed. Hepatic inorganic phosphate levels were also affected by APAP administration, and these levels reflected the changes seen in the adenine nucleotide pool in that inorganic phosphate levels were elevated at 1 and 6 h after the APAP treatment.

Effects of AMAP and APAP on Urinary Allantoin Levels-
The final breakdown product of purine metabolism in mice is allantoin. The enzyme urate oxidase metabolizes uric acid to allantoin, and this compound is excreted in the urine. Control urinary allantoin levels were determined to be 16.4 rmol of allantoin excreted per 100 g body weight in 5 h (Table VI). The APAP treatment elevated urinary allantoin levels over control values while the AMAP treatment had no significant effect.

Effects of Altopurinot on APAP-induced Hepatotoxicity-
The xanthine oxidase inhibitor, allopurinol has been shown to inhibit APAP toxicity (14, 30). Fig. 1 indicates that allopurinol is also effective in inhibiting the APAP-induced toxicity observed in the present study.   were given APAP at a dose of 250 mg/kg. After 24 h, serum samples were collected and serum ALT levels were measured. C, APAP + allopurinol-treated animals (n = 5). Allopurinol was administered 3 h prior to APAP at a dose of 125 mg/kg. APAP was administered as before. Serum ALT levels were measured at 24 h after APAP administration.

DISCUSSION
The data in the present study suggest that APAP administration produces a complex sequence of events which may lead to the transient production of activated oxygen, the loss of protein thiols, and eventually to cell death. The proposed sequence of events are outlined in Fig. 2. APAP is first oxidized by cytochrome P-450 to the reactive metabolite NAPQI (4). Following APAP administration, both the mitochondrial and cytosolic pools of GSH are depleted (12). Once GSH is depleted, cellular proteins are arylated, presumably, by the reactive metabolite NAPQI (6). In addition, cytosolic glutathione peroxidase activity is inhibited in APAP-treated animals. The inhibition of this enzyme as well as the depletion of GSH may make the cell very vulnerable to the deleterious effects of activated oxygen species. Even under normal circumstances, hydrogen peroxide and superoxide anion are generated in the cell from a variety of sources (31). This background production of activated oxygen may become significant if the GSH-glutathione peroxidase detoxification pathway is compromised. Under such circumstances, the pro- duction of activated oxygen species may lead to the oxidation of sensitive protein thiols. The inhibition of thiol transferase activity may also contribute to the increased levels of oxidized protein thiols in APAP-treated animals by inhibiting the reduction of oxidized protein thiols.
It is known that NAPQI can directly oxidize as well as arylate thiols (32)(33)(34). In this manner, APAP may lead to arylation and oxidation of critical cellular proteins. Studies with isolated hepatocytes and dimethylated analogues of NAPQI (35, 36) suggest that arylation may be more toxic to isolated cells than protein oxidation. However, 3,5dimethyl-NAPQI is known to generate oxidized glutathione (GSSG), (35), and may thus promote the formation of low molecular weight protein mixed disulfides rather than protein-protein or internal protein disulfides we hypothesize occurs after APAP administration.
These disulfides may differ in their relative abilities to produce toxicity.
Previously, we reported that APAP administration disrupted calcium homeostasis in the liver. APAP-treated mice displayed decreased plasma membrane calcium-ATPase activity and impaired mitochondrial calcium sequestration 1 h after receiving the drug (12). The influx of extracellular calcium as a result of this plasma membrane calcium-ATPase inhibition in addition to the loss of the ability of mitochondria to sequester calcium may lead to large-scale calcium cycling by mitochondria.
This calcium cycling may have several important consequences for cell function. Thomas and Reed (37) suggest that calcium cycling may be involved in the production of oxidative stress in isolated hepatocytes by an unknown mechanism. Calcium has also been shown to increase hydrogen peroxide production by isolated rat heart mitochondria (38). In addition, calcium cycling decreases the rate of ATP synthesis by cell mitochondria (39). The uptake of calcium by mitochondria is known to require respiratory energy and to take precedence over ATP formation (40). This decreased ATP synthesis may lead to the increased breakdown of adenine nucleotides in response to energy requiring cellular metabolism.
Our results agree with this hypothesis. The total liver adenine nucleotide pool declined to about 50% of control levels 1 h after APAP administration, and this decrease coincided with the loss in the ability of mitochondria to sequester calcium (12). The degradation of this pool is also supported by the increase in the amount of the terminal purine breakdown product, allantoin, excreted in the urine in APAP-treated mice. Our results also indicate that APAP administration promotes the conversion of the enzyme xanthine dehydrogenase to the oxidase form. This conversion in conjunction with the degradation of the adenine nucleotide pool may lead to the production of activated oxygen species and the subsequent oxidation of protein thiols as seen in this study. This loss in protein thiols may in turn lead to cell death. Studies with a variety of toxic agents have correlated the loss of protein thiols with the development of cellular toxicity (41)(42)(43).
The mechanism we have proposed for the development of APAP hepatoxicity is supported by several lines of evidence. Incubation of isolated hepatocytes with DTT after APAP (44,45) or NAPQI (33,45) exposure decreased toxicity. The DTT treatment had no effect on total covalent binding levels, and both studies concluded that the oxidation of protein thiols by APAP and NAPQI may be an important event in the development of toxicity. Also, fructose has been shown to potentiate APAP toxicity (46). Previous studies have shown that fructose promotes the degradation of liver adenine nucleotides (47), as was observed in the present report with APAP. In addition, Beales et al. (48) showed that calcium-EDTA protects isolated hepatocytes against APAP-induced toxicity. One mechanism proposed by these authors to account for this protection involved the buffering of cytosolic calcium. Buffering cytosolic calcium may also prevent calcium cycling.
Finally, we have demonstrated that the xanthine oxidase inhibitor allopurinol protects against the hepatotoxicity of APAP. These data agree with observations of other researchers (14), and in a recent abstract (30) allopurinol was shown to protect against APAP-induced toxicity without altering APAP-reactive metabolite formation. It was concluded that APAP promotes neutrophil accumulation in the liver of APAP-treated mice. The data presented here suggest that allopurinol also may protect against APAP toxicity by inhibiting liver xanthine oxidase. The magnitude and the time course of the conversion of the enzyme from the dehydrogenase to the oxidase form cannot be explained solely by the accumulation of neutrophils in the liver. A major argument against the occurrence of oxidative stress in APAP toxicity in uiuo is the lack of GSSG that is excreted in the bile in response to APAP toxicity (49). Our results suggest a possible explanation for this observation. We propose that oxidative stress occurs shortly after APAP administration in mice. At this time, activated oxygen species are produced in liver cells which are substantially depleted of GSH and that have impaired cytosolic glutathione peroxidase and thioltransferase activities. However, we observed no inhibition of cytosolic glutathione reductase in APAP-treated mice. It is, therefore, likely that the small amounts of GSSG formed in the cytosol during this period will be rapidly reduced back to GSH. Since mitochondria contain their own glutathione peroxidase and reductase activity, these observations may not apply to this organelle. Indeed, high levels of GSSG have been observed in mitochondria after APAP administration (12). It is important to note that the oxidative stress that we suggest occurs in APAP toxicity is transient in nature, and is complete before the onset of the APAP 6-h time point. Both mitochondrial and cytosolic GSH levels have returned to near control levels 6 h after receiving APAP, and the ability of liver mitochondria to sequester calcium has returned to normal (12). In addition, the fraction of the xanthine dehydrogenase-oxidase enzyme present in the oxidase form returned to control levels at this time point. However, hepatic protein thiol levels did not return to control levels despite the increase in cellular GSH and the return to control levels of thioltransferase activity. This may related to the inaccessability of certain protein disulfides to GSH. Studies have indicated that protein disulfide bonds are frequently resistant to reduction by GSH under physiological conditions (50).
AMAP is not hepatotoxic to mice despite considerable covalent binding to cellular proteins by its reactive metabolites (12). In contrast to APAP, AMAP does not induce calcium cycling or decreases in liver adenine nucleotides. Thus, AMAP would not be expected to generate large-scale oxidative stress according to the mechanism proposed for APAP. In agreement with this hypothesis, AMAP did not significantly affect liver protein thiols. APAP inhibited plasma membrane calcium-ATPase activity and promoted the conversion of xanthine dehydrogenase to xanthine oxidase. AMAP did not cause these changes. This disparity may be due to differences either in the cellular proteins that the reactive metabolites of AMAP and APAP arylate or in the relative ability of the two agents to oxidize protein thiols. It is known that both the plasma membrane calcium-ATPase (51) and xanthine dehydrogenase (28) contain critical thiol groups that are sensitive to oxidation.
Both AMAP and APAP deplete cytosolic GSH, but APAP depletes mitochondrial GSH to a greater extent than AMAP. This difference may be important for the development of toxicity. Recently, we observed that AMAP is hepatotoxic if mice are pretreated with the GSH synthesis inhibitor buthionine sulfoximine.' This effect may relate to the combined ability of buthionine sulfoximine and AMAP to deplete mitochondrial GSH. Depletion of mitochondrial GSH may in turn promote the oxidation of critical cellular thiols. Further research is required to examine this hypothesis.
We present results in the present study which suggest that APAP can induce oxidative stress by a novel mechanism. Additional studies are required to examine this mechanism and to determine if it may be applicable for other hepatotoxic agents.