ToxSci Advance Access originally published online on September 21, 2005
Toxicological Sciences 2006 89(1):31-41; doi:10.1093/toxsci/kfi336
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
REVIEW |
Intracellular Signaling Mechanisms of Acetaminophen-Induced Liver Cell Death
Liver Research Institute, University of Arizona, Tucson, Arizona 85737
1 To whom correspondence should be addressed at Liver Research Institute, University of Arizona, College of Medicine, 1501 N. Campbell Ave, Room 6309, Tucson, AZ 85724. Fax: (520) 626-5975. E-mail: jaeschke{at}email.arizona.edu.
Received August 22, 2005; accepted September 16, 2005
 |
ABSTRACT |
|---|
TOP
ABSTRACT
INTRODUCTIONAcetaminophen hepatotoxicity is the leading cause of drug-induced liver failure. Despite substantial efforts in the past, the mechanisms of acetaminophen-induced liver cell injury are still incompletely understood. Recent advances suggest that reactive metabolite formation, glutathione depletion, and alkylation of proteins, especially mitochondrial proteins, are critical initiating events for the toxicity. Bcl-2 family members Bax and Bid then form pores in the outer mitochondrial membrane and release intermembrane proteins, e.g., apoptosis-inducing factor (AIF) and endonuclease G, which then translocate to the nucleus and initiate chromatin condensation and DNA fragmentation, respectively. Mitochondrial dysfunction, due to covalent binding, leads to formation of reactive oxygen and peroxynitrite, which trigger the membrane permeability transition and the collapse of the mitochondrial membrane potential. In addition to the diminishing capacity to synthesize ATP, endonuclease G and AIF are further released. Endonuclease G, together with an activated nuclear Ca2+,Mg2+-dependent endonuclease, cause DNA degradation, thereby preventing cell recovery and regeneration. Disruption of the Ca2+ homeostasis also leads to activation of intracellular proteases, e.g., calpains, which can proteolytically cleave structural proteins. Thus, multiple events including massive mitochondrial dysfunction and ATP depletion, extensive DNA fragmentation, and modification of intracellular proteins contribute to the development of oncotic necrotic cell death in the liver after acetaminophen overdose. Based on the recognition of the temporal sequence and interdependency of these mechanisms, it appears most promising to therapeutically target either the initiating event (metabolic activation) or the central propagating event (mitochondrial dysfunction and peroxynitrite formation) to prevent acetaminophen-induced liver cell death.
Key Words: acetaminophen hepatotoxicity; oncotic necrosis; apoptosis; endonucleases; DNA fragmentation; oxidant stress; peroxynitrite; covalent binding; reactive metabolites.
|
INTRODUCTION |
|---|
Acetaminophen (AAP) is a safe and effective analgesic/anti-pyretic drug when used at therapeutic levels (Rumack, 2004
). However, an acute or cumulative overdose can cause severe liver injury with the potential to progress to liver failure (Lee, 2004). In fact, AAP overdose is the most frequent cause of drug-induced liver failure in the United States and in Great Britain (Lee, 2004). In addition to its clinical relevance, AAP is extensively used as a model toxin to evaluate novel hepatoprotective agents. Based on the mechanistic insight gained from early preclinical studies (Jollow et al., 1973; Mitchell et al., 1973a,b), N-acetylcysteine was introduced in the 1970s and still is the only clinical antidote against AAP-induced liver injury (Polson and Lee, 2005). Despite substantial efforts to investigate AAP-induced liver cell death during the last 30 years, there are many details of the mechanism that are still unknown. This review will focus on more recent advances in our understanding of intracellular signaling events after AAP overdose. As discussed elsewhere (Jaeschke, 2005), additional aspects of the pathophysiology, e.g., inflammatory mediators and leukocytes, are also important for the overall outcome. However, since the inflammatory response is dependent on the initial injury to hepatocytes, intracellular events may offer the most promising and effective targets for therapeutic interventions. |
INITIATION OF AAP-INDUCED CELL DEATH |
|---|
Among the most extensively studied and least controversial issues is the metabolic activation of AAP. A large portion of a therapeutic dose of AAP is directly conjugated with glucuronic acid or sulfate through glucuronyltransferases or sulfotransferases, respectively (Nelson, 1990
). The AAP-glucuronide is excreted into bile through the canalicular multidrug resistance-associated protein 2 (Mrp2, Abcc2) and into blood through Mrp3 (Abcc3) (Xiong et al., 2000). The biliary elimination of AAP-sulfate is mediated mainly by Mrp2 and to a lesser degree by breast cancer resistance protein (BCRP, ABCG2) (Zamek-Gliszczynski et al., 2005). On the other hand, AAP-sulfate transport into the blood appears not to be mediated by Mrp3 but by other, not fully identified organic anion transporters (Zamek-Gliszczynski et al., 2005). The remaining part of the dose, which is not directly conjugated with hydrophilic groups, is metabolized by the P450 system to a reactive metabolite (Mitchell et al., 1973a), presumably N-acetyl-p-benzoquinone imine (NAPQI) (Nelson, 1990). Although a number of P450 enzymes can metabolize AAP, the most relevant isoenzyme is CYP2E1, especially in humans (Gonzalez, 2005; Raucy et al., 1989). NAPQI reacts with glutathione (GSH) spontaneously or catalyzed by glutathione-S-transferases to form a GSH-adduct, which is mainly excreted into bile through Mrp2 (Chen et al., 2003). Thus, the earliest effect of AAP metabolism is a profound depletion of hepatocellular glutathione (Mitchell et al., 1973b), which affects both the cytosolic and the mitochondrial compartments. Once GSH is exhausted, any remaining NAPQI formed will react with alternative targets, in particular cellular proteins (Jollow et al., 1973). The covalent modification of cellular proteins was originally thought to cause necrotic cell death (Jollow et al., 1973; Mitchell et al., 1973a,b). This hypothesis was supported by the fact that protein binding preceded cell death and occurred in cells, which became later necrotic (Roberts et al., 1991). In addition, any intervention that prevented covalent binding also prevented cell death (Jollow et al., 1973; Mitchell et al., 1973a,b). A major criticism of the protein-binding hypothesis was an apparent lack of dose responsiveness. Part of this effect could be explained by the influx of protein into the liver due to severe hemorrhage, which diluted the covalently modified protein pool (Corcoran et al., 1985). Overall, there is no AAP hepatotoxicity without protein binding. However, there are examples of AAP-mediated protein binding without hepatotoxicity. One explanation for this observation could be that the total binding to cellular proteins may be less relevant than selective modifications of specific critical targets. Therefore, a considerable effort was made to identify covalently modified proteins (Cohen et al., 1997; Qiu et al., 1998). However, as one would expect from a spontaneous chemical reaction of a reactive metabolite with numerous potential intracellular targets, the activity of most modified enzymes was only modestly affected (Halmes et al., 1996; Pumford et al., 1997). The exceptions are glutamine synthetase and carbamyl phosphate synthetase-1, which are inhibited by more than 50% after AAP treatment (Gupta et al., 1997). The significant loss of the function of these enzymes may be responsible for the hyperammonemia observed after AAP overdose but is less likely to cause acute cell death. Thus, covalent binding to cellular proteins in general or the selective modification of one or more critical enzymes does not appear to be the direct cause of AAP-induced liver cell necrosis.
The regioisomer of AAP, 3'-hydroxyacetanilide, can cause glutathione depletion and a similar degree of covalent binding to cellular proteins as AAP, but does not cause liver injury (Tirmenstein and Nelson, 1989). Despite the same overall protein binding, AAP induces more prominent covalent modifications of mitochondrial proteins than 3'-hydroxyacetanilide (Qiu et al., 2001; Tirmenstein and Nelson, 1989). Since the binding of NAPQI to mitochondrial proteins correlates with the potential to cause liver injury, these findings suggest that metabolic activation of AAP and protein binding of the reactive metabolite is a critical initiating event in the toxicity, which needs to be amplified and propagated in order to cause cell death (Jaeschke et al., 2003). This concept is further supported by many pharmacological interventions that protected but did not affect reactive metabolite formation and protein binding (e.g., Birge et al., 1988; Corcoran et al., 1985; Jaeschke, 1990; James et al., 2003c; Salminen et al., 1998; Slitt et al., 2004).
|
MITOCHONDRIAL DYSFUNCTION AND PROPAGATION OF AAP-INDUCED CELL INJURY |
|---|
AAP overdose triggers mitochondrial dysfunction as indicated by inhibition of mitochondrial respiration (Burcham and Harman, 1991
; Meyers et al., 1988; Ramsay et al., 1989) and declining ATP levels (Jaeschke, 1990; Tirmenstein and Nelson, 1990). Since the mitochondrial effects of AAP in vivo occur immediately after GSH depletion (Donnelly et al., 1994) and can be directly reproduced by NAPQI in isolated mitochondria (Burcham and Harman, 1991; Ramsay et al., 1989), covalent binding to mitochondrial proteins may be to a significant degree responsible for the initial mitochondrial dysfunction. However, other contributing factors cannot be excluded. In addition to the inhibition of the mitochondrial respiration, it was recognized that when hepatic GSH levels are beginning to recover after the initial depletion by NAPQI, hepatic concentrations of glutathione disulfide (GSSG), a marker of intracellular reactive oxygen formation, increase substantially above baseline (Jaeschke, 1990; Tirmenstein and Nelson, 1990). Interestingly, no release of GSSG into bile was detected (Jaeschke, 1990). This finding suggested that GSSG was formed in an intracellular compartment, which does not release GSSG, i.e., mitochondria. In fact, when mitochondria were isolated from AAP-treated animals, GSSG levels were extremely high and could account for most if not all GSSG within the cell (Jaeschke, 1990; James et al., 2003b; Knight et al., 2001). These data support the conclusion that AAP induces a mitochondrial oxidant stress. Although it was hypothesized that this oxidant stress may be a consequence rather than the cause of cell injury (Rogers et al., 2000), detailed time-course experiments of the oxidation of 2',7'-dichlorodihydrofluorescein, another marker of cellular oxidant stress, showed that reactive oxygen formation starts immediately after GSH depletion and precedes cell death by several hours (Bajt et al., 2004). Although these results are consistent with an important role of reactive oxygen in the pathophysiology, GSH peroxidase-1-deficient (GPx1–/–) mice were not more susceptible to AAP-induced liver injury than wildtype animals (Knight et al., 2002). However, GPx1–/– mice showed more extensive liver injury in response to oxidant stress induced by neutrophils (Jaeschke et al., 1999, 2002) or diquat (Fu et al., 1999). In addition, AAP treatment did not cause a relevant increase in lipid peroxidation, and loading cells with vitamin E did not protect against AAP overdose (Knight et al., 2003). Taken together, these data indicate that reactive oxygen species such as hydrogen peroxide and hydroxyl radicals are of limited relevance for the mechanism of AAP-induced cell death.
Superoxide can react with nitric oxide (NO) to form peroxynitrite, a potent oxidant and nitrating species (Denicola and Radi, 2005). This reaction is diffusion limited and is actually several times faster even than the superoxide dismutase-catalyzed dismutation reaction (Denicola and Radi, 2005). Immunohistochemical staining for nitrotyrosine protein adducts in cells undergoing necrosis provided evidence for peroxynitrite formation after AAP overdose (Hinson et al., 1998; Knight et al., 2001). Subcellular fractionation showed that peroxynitrite is predominantly generated in mitochondria (Cover et al., 2005b) (Fig. 1), which is consistent with the increased formation of superoxide in these cell organelles (Jaeschke, 1990; Tirmenstein and Nelson, 1990). The source(s) of NO after AAP treatment remain somewhat unclear. On the one hand, nitrotyrosine staining is reduced in inducible nitric oxide synthase–deficient (iNOS–/–) mice (Gardner et al., 2002; Michael et al., 2001), which would suggest that iNOS is a major source of NO for peroxynitrite formation after AAP treatment. In addition, animals lacking the anti-inflammatory gene interleukin-10 (IL-10) responded to an AAP overdose with higher iNOS induction and increased liver injury compared to wildtype animals (Bourdi et al., 2002). However, nitrotyrosine protein adducts were detected as early as 0.5 to 1 h after AAP exposure in the absence of iNOS induction (Cover et al., 2005b; Knight et al., 2001). These results indicate that iNOS is an important, but not the only, possible source for NO in AAP hepatotoxicity.
|
Although there is extensive evidence for peroxynitrite formation before tissue injury, the important question remains if this reactive nitrogen species is a relevant mediator for the overall cell death. Experiments with iNOS–/– mice (Gardner et al., 2002
; Michael et al., 2001) or iNOS inhibitors (Gardner et al., 1998; Hinson et al., 2002; Kamanaka et al., 2003) yielded conflicting results. Interventions that enhanced iNOS induction increased AAP-induced liver injury (Bourdi et al., 2002; Tinel et al., 2004). We used a different approach to address this critical question. Since GSH is a potent scavenger of peroxynitrite (Kirsch et al., 2001), we hypothesized that delayed treatment with GSH after AAP may accelerate the recovery of cellular and, in particular, mitochondrial GSH levels, which then may be able to scavenge peroxynitrite in vivo and may attenuate cell injury (Knight et al., 2002). Indeed, intravenous administration of GSH, which is rapidly degraded in the kidney and provides amino acids for the resynthesis of GSH in the liver (Wendel and Jaeschke, 1982), effectively prevented nitrotyrosine staining, strongly attenuated AAP-induced liver injury, and promoted cell cycle activation and regeneration (Bajt et al., 2003; Knight et al., 2002). Since this treatment did not prevent mitochondrial dysfunction and the mitochondrial oxidant stress in wildtype as well as in GPx1–/– mice, these results strongly supported the hypothesis that peroxynitrite is a relevant mediator of AAP hepatotoxicity (Knight et al., 2002). These findings were later confirmed by using N-acetylcysteine (James et al., 2003c). |
CONSEQUENCES OF MITOCHONDRIAL OXIDANT STRESS AND PEROXYNITRITE FORMATION |
|---|
Oxidant stress caused by reactive oxygen and nitrogen species and increased Ca2+ levels are well-known inducers of the mitochondrial membrane permeability transition (MPT) in many cell types including hepatocytes (Kim et al., 2003
). The MPT is characterized by mitochondrial swelling, uncoupling of the oxidative phosphorylation, and formation of pores in the inner mitochondrial membrane. Opening of these pores allows the passage of solutes of 
1500 Dalton molecular weight (Kim et al., 2003). A consequence of the MPT is the breakdown of the mitochondrial membrane potential (
m), the inability to synthesize ATP, and finally, necrotic cell death (Kim et al., 2003). If the insult is milder and only a limited number of mitochondria undergo the MPT, the resulting release of cytochrome c can promote apoptotic cell death (Kim et al., 2003). In general, the MPT is dependent on the presence of Ca2+ and can be inhibited by cyclosporin A (CsA). However, a more severe insult may trigger a Ca2+-independent, CsA-insensitive MPT (He and Lemasters, 2002).
Since AAP overdose depletes mitochondrial GSH levels, induces formation of peroxynitrite in mitochondria (Cover et al., 2005b) (Fig. 1), and causes cellular Ca2+ accumulation (Corcoran et al., 1988) through inhibition of the Ca2+-Mg2+-ATPase (Tsokos-Kuhn et al., 1988), it is not surprising that AAP induced the mitochondrial MPT in primary cultured mouse hepatocytes 4–6 h after exposure to AAP (Kon et al., 2004a). The MPT occurred after GSH depletion and after initiation of reactive oxygen formation, but preceded necrotic cell death (Bajt et al., 2004; Kon et al., 2004a). The AAP-induced MPT and cell necrosis could be substantially delayed with CsA but not prevented at later time points (Kon et al., 2004a). These data suggest that AAP initially triggers a CsA-sensitive, regulated MPT, which is followed by a CsA-insensitive, unregulated MPT (Kon et al., 2004a). CsA treatment also protected against AAP hepatotoxicity in vivo (Haouzi et al., 2002; Masubuchi et al., 2005). More recent observations indicate that ferrous iron mobilized from lysosomes and oxidant stress appears to be involved in triggering the MPT after AAP (Kon et al., 2004b). This mechanism explains the previously reported protective effect of iron chelation against AAP toxicity in vitro (Adamson and Harman, 1993). In addition, N-acetylcysteine treatment 2 h after AAP attenuated the oxidant stress, the loss of the mitochondrial membrane potential, and cell injury (Bajt et al., 2004; Reid et al., 2005). Taken together, these findings support the hypothesis that reactive oxygen and Fenton-type reaction products may be important in the induction of the MPT and cell necrosis in vitro. Consistent with this conclusion, antioxidants such as vitamin E have been shown to protect against AAP-induced cell injury in culture (Nagai et al., 2002). However, in contrast to the findings in vitro, the iron chelator desferoxamine is either not at all protective (Smith et al., 1986) or only delays toxicity in vivo (Schnellmann et al., 1999). Treatment with 
- or 
-tocopherol, which clearly prevented lipid peroxidation and cell injury after iron/allyl alcohol administration, did not protect against AAP-induced liver injury (Knight et al., 2003). Thus, together these observations indicate that there might be a higher emphasis on oxidant stress mechanisms in vitro compared to the in vivo situation where peroxynitrite appears to dominate (Knight et al., 2002). The reason for this effect could be an artificially enhanced mitochondrial oxidant stress due to the generally used hyperoxic cell culture conditions (Halliwell, 2003).
In addition to the MPT, peroxynitrite may cause a number of other adverse effects in the mitochondria. Recently, we showed a decline of mitochondrial DNA (mtDNA) during AAP hepatotoxicity as assessed by a slot blot hybridization assay (Cover et al., 2005b). When nitrotyrosine formation in mitochondria was eliminated by treatment with GSH, which accelerated the recovery of mitochondrial GSH levels, the loss of mtDNA was only partially prevented (Cover et al., 2005b). This suggests that peroxynitrite is only one of several causes of mtDNA modifications (Rogers et al., 1997). Although the effect of mtDNA loss on acute cell toxicity may be minimal, if the cell survives the initial insult, long-term survival may be jeopardized.
Mitochondrial release of Ca2+, together with the inhibition of the Ca2+-Mg2+-ATPase in the plasma membrane (Tsokos-Kuhn et al., 1988), may lead to an increase of cytosolic Ca2+ levels sufficient to activate Ca2+-dependent intracellular proteases such as calpains. These enzymes can proteolytically cleave structural proteins within the cell and contribute to oncotic necrosis (Liu et al., 2004). In addition, release of calpains from necrotic cells can affect neighboring cells and expand the injury within the liver (Limaye et al., 2003). Another consequence of the disrupted intracellular Ca2+ homeostasis is the accumulation of Ca2+ in the nucleus and activation of a Ca2+-dependent endonuclease (Ray et al., 1990).
|
MITOCHONDRIAL DYSFUNCTION AND NUCLEAR DNA FRAGMENTATION |
|---|
AAP overdose causes fragmentation of nuclear DNA and karyolysis both in vivo and in primary cultured hepatocytes, as first recognized by Corcoran and coworkers (Ray et al., 1990
; Shen et al., 1991). A characteristic DNA ladder on agarose gel provided evidence for internucleosomal DNA cleavage (Ray et al., 1990; Shen et al., 1991). DNA fragmentation was later confirmed through other assays such as the antihistone ELISA (Lawson et al., 1999), which measures the presence of small DNA fragments in the cytosol, and the terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) assay (Lawson et al., 1999), which indicates DNA strandbreaks. Together these assays clearly demonstrate nuclear DNA damage, which occurs in parallel to AAP-induced liver cell injury (Lawson et al., 1999; Ray et al., 1990). The fact that DNA damage and cell injury was prevented by a general endonuclease inhibitor (Shen et al., 1992) suggested that endonuclease-mediated DNA fragmentation is a critical event in the pathophysiology. The finding that no nitrotyrosine protein adducts could be detected in the nucleus supports the hypothesis that peroxynitrite was not directly responsible for the DNA damage after AAP overdose (Cover et al., 2005b) (Fig. 1). However, delayed treatment with GSH effectively scavenged mitochondrial peroxynitrite and prevented nuclear DNA fragmentation (Cover et al., 2005b). Thus, there appears to be a connection between mitochondrial dysfunction and nuclear DNA damage.
The most studied endonuclease is the DNA fragmentation factor (DFF40) or caspase-activated DNase (CAD), which is activated through cleavage of its inhibitor (DFF45/ICAD) by caspase-3 (Nagata et al., 2003). In the liver, caspase-3 is mostly activated through cytochrome c release from mitochondria and caspase-9 activation by the apoptosome (Hill et al., 2003). Although mitochondrial cytochrome c release has been shown during AAP hepatotoxicity (Adams et al., 2001; El-Hassan et al., 2003; Knight and Jaeschke, 2002) (Fig. 2), a number of laboratories consistently reported either no or at best a minor activation of caspases (Adams et al., 2001; El-Hassan et al., 2003; Ferret et al., 2001; Gujral et al., 2002; Lawson et al., 1999; Nagai et al., 2002; Tinel et al., 2004). In addition, the staining pattern of the TUNEL assay and the amount of low-molecular-weight DNA fragments in the cytosol or plasma are clearly different between AAP-induced cell injury and death receptor-mediated apoptosis, where activation of DFF/CAD is undisputed (Cover et al., 2005b; Gujral et al., 2002; Jahr et al., 2001; Lawson et al., 1999). Thus, despite DNA ladders undistinguishable from apoptosis, there is no evidence for a relevant activation of DFF/CAD during AAP hepatotoxicity. Nevertheless, until the responsible endonucleases are definitively identified, a contribution of DFF/CAD to nuclear DNA fragmentation cannot be completely ruled out.
|
During apoptosis, DFF/CAD is thought to cleave DNA first into 50 kb fragments (Nagata et al., 2003
). These large fragments are then degraded to smaller internucleosomal repeats by a Ca2+/Mg2+-dependent endonuclease (DNAS1L3) located in the nucleus (Yakovlev et al., 1999). Indirect support for a role of DNAS1L3 comes from earlier work by Corcoran's group showing a correlation between DNA ladders and nuclear Ca2+ accumulation (Ray et al., 1990). In addition, treatment with Ca2+ channel blockers inhibited nuclear Ca2+ accumulation, attenuated DNA fragmentation and liver injury (Ray et al., 1993). Interestingly, DNAS1L3 is inhibited by poly(ADP-ribose) polymerase (PARP) (Yakovlev et al., 2000). During apoptosis, this inhibitory effect is generally overcome by caspase-3 activation, which inactivates PARP (Yakovlev et al., 2000). However, PARP activation occurs predominantly after major DNA fragmentation during AAP-induced liver injury (Cover et al., 2005a), and although there is no caspase activation during AAP hepatotoxicity, there is still PARP cleavage (Zhang et al., 2000), presumably through other intracellular proteases. Thus, activation of DNAS1L3 appears to be involved in AAP-induced DNA fragmentation. However, it remains unclear what endonuclease replaces DFF/CAD.
The pro-apoptotic Bcl-2 family member Bax resides in the cytosol but can translocate to mitochondria and can form pores in the outer mitochondrial membrane alone or in combination with other Bcl-2 proteins, e.g., Bad and the truncated form of Bid (Chao and Korsmeyer, 1998). Formation of these pores together with formation of the MPT pores in the inner membrane can release proteins from the intermembrane space of mitochondria (Jaeschke and Lemasters, 2003; Scorrano and Korsmeyer, 2003). Apoptosis-inducing factor (AIF) (Susin et al., 2000) and endonuclease G (van Loo et al., 2001) are proteins that have been implicated in chromatin condensation and nuclear DNA fragmentation, respectively. Therefore, we assessed the release of endonuclease G and AIF from mitochondria and their translocation to the nucleus in response to AAP treatment. In primary cultured mouse hepatocytes, nuclear translocation of endonuclease G and AIF was observed between 3 and 6 h after AAP exposure, i.e., after GSH depletion but before cell death (Bajt et al., unpublished observation). Nuclear translocation of endonuclease G and AIF and DNA fragmentation was inhibited by treatment with N-acetylcysteine (Bajt et al., unpublished observation), which also attenuated cell death (Bajt et al., 2004). In vivo, mitochondrial translocation of Bax and truncated Bid (tBid) precedes nuclear DNA fragmentation and mitochondrial oxidant stress (Adams et al., 2001; Bajt et al., 2005; El-Hassan et al., 2003) (Fig. 2). However, in Bax-deficient (Bax–/–) mice, only DNA fragmentation and cell injury but not peroxynitrite formation was delayed (Bajt et al., 2005). On the other hand, DNA damage and injury was similar in Bax–/– and in wild-type mice at later time points (
12 h) (Bajt et al., 2005; Bajt et al., unpublished results). These findings indicate that AAP-induced Bax translocation to the mitochondria may induce the early release of intermembrane proteins, e.g., endonuclease G, AIF, the second mitochondria-derived activator of caspases (Smac), and cytochrome c. Endonuclease G and AIF translocate to the nucleus and cause DNA damage, which correlates with cellular necrosis. Therefore, endonuclease G and AIF may be the link between Bax-mediated pore formation in the outer mitochondrial membrane and nuclear DNA fragmentation. Furthermore, endonuclease G may be the DNase that generates the large DNA fragments for the activation of DNAS1L3. However, since the oxidant stress and peroxynitrite formation occur independent of Bax, the resulting MPT will eventually cause endonuclease G and AIF release, which leads to further DNA fragmentation and cell injury. Thus, the Bax-mediated signaling mechanism is overridden by peroxynitrite and the mitochondrial MPT at later time points.
How DNA damage contributes to cell necrosis remains unclear. One possible link could be the activation of PARP, which is critical for DNA repair (Meyer-Ficca et al., 2005). However, excessive activation of PARP, which depletes cellular levels of NAD+ and ATP, leads to an energy crisis within the cell, causing cell necrosis (Ha and Snyder, 1999). Although there is clear evidence for PARP activation after AAP overdose, the activation of this enzyme occurs after the onset of DNA fragmentation and cell injury (Cover et al., 2005a). In addition, no protective effect was observed in PARP-deficient (PARP–/–) mice or with specific PARP inhibitors (Cover et al., 2005a). In fact, PARP–/– mice showed a slightly increased injury (Cover et al., 2005a). This finding is similar to a previous in vitro study using a PARP inhibitor (Shen et al., 1992). In contrast, it was reported that high doses of the PARP inhibitor 3-aminobenzamide protected against AAP hepatotoxicity in vivo (Ray et al., 2001). However, this effect appears to be independent of PARP, because the chemical was as effective in wildtype as in PARP–/– mice (Cover et al., 2005a). Together, these findings suggest that PARP activation is a response to DNA damage after AAP, but is not actively involved in the process of cell injury. On the other hand, regeneration was suppressed in PARP–/– mice at later time points after AAP overdose (Bajt et al., unpublished observations).
|
AAP HEPATOTOXICITY AND REGENERATION |
|---|
The liver is a unique organ in the sense that a significant loss of liver cells due to drug toxicity or other insults can be overcome by regeneration (Mehendale, 2005
). The replacement of necrotic cells occurs predominantly through proliferation of hepatocytes, which need to exit the quiescent state (G0) and enter the cell cycle (Michalopoulos and DeFrances, 2005). Although the entire process is highly complex and has been extensively studied, only a few key aspects will be addressed. The first step involves the priming of hepatocytes by TNF- (Akerman et al., 1992) and interleukin-6 (IL-6) (Cressman et al., 1996), which makes the cells more responsive to growth factors (Fausto, 2000). Various growth factors then induce expression of cell cycle proteins including cyclins, cyclin-dependent kinases, and cell cycle inhibitors. The simultaneous expression of activators and inhibitors of the cell cycle ensures a sensitive regulation of the process (Fausto, 2000). One of the markers of cell cycle activation is the expression of cyclin D1, which signals the commitment of the cell to DNA synthesis (Fausto, 2000). Consistent with the general concept of regeneration, cell proliferation after AAP overdose involves IL-6 (James et al., 2003a) and TNF- (Chiu et al., 2003), as indicated by the impaired regenerative response in IL-6–/– and TNF receptor type 1 (p55)–/– mice. Moreover, scavenging of mitochondrial peroxynitrite with GSH attenuates AAP-induced liver injury and causes expression of cyclin D1 and proliferating cell nuclear antigen (PCNA) (Bajt et al., 2003). The most extensively stained cells are located in the transition zone between necrotic and healthy cells (Bajt et al., 2003). Interestingly, neither pretreatment with growth factors (Francavilla et al., 1993) nor IL-6 (Bajt et al., 2003) prevented AAP-induced liver injury. On the hand, forcing hepatocytes into cell division with a low dose of thioacetamide attenuated AAP hepatotoxicity (Chanda et al., 1995). These data suggest that cell cycle activation and regeneration is a critical event in the pathophysiology of AAP-induced liver failure. A more detailed knowledge of the events may reveal important new therapeutic strategies. |
APOPTOTIC SIGNALING PATHWAYS AND AAP-INDUCED ONCOTIC NECROSIS |
|---|
As discussed, there is a significant overlap between events normally associated with apoptosis and AAP-induced cell death. The first observation of this kind was DNA fragmentation as indicated by DNA ladders after AAP overdose (Ray et al., 1990
). This initial finding lead to the conclusion that about 40% of hepatocytes actually die by apoptosis (Ray et al., 1996). However, a detailed morphological study of AAP-induced cell death in comparison with TNF-induced apoptosis clearly demonstrated that >95% of injured hepatocytes die through oncotic necrosis in vivo (Gujral et al., 2002). Apoptosis is quantitatively at best a minor issue. Similarly, necrosis is the dominant cell death in primary hepatocytes (Bajt et al., 2004; Kon et al., 2004a; Nagai et al., 2002). Nevertheless, the increasing overlap in signaling events, e.g., Bax and tBid translocation to mitochondria (Fig. 2), cytochrome c release from mitochondria (Fig. 2), PARP cleavage (Zhang et al., 2000), internucleosomal DNA cleavage (Ray et al., 1990), between apoptosis and AAP-mediated cell events continuously provides fuel for new speculation regarding the mode of AAP-induced cell death. These observations, together with a recent report showing that the pancaspase inhibitor benzyloxycarbonyl-Val-Ala-Asp-fluoromethylketone (Z-VAD-fmk) protected against AAP-induced hepatotoxicity, lead to the hypothesis that AAP induces hepatocellular apoptosis, which is rapidly aborted due to declining ATP levels (El-Hassan et al., 2003). This study agrees with our previous data that there is no relevant caspase activation during AAP hepatotoxicity (Lawson et al., 1999). In fact, the mitochondrial dysfunction induced by AAP overdose is so severe that it effectively prevents Fas receptor-mediated caspase activation and apoptosis (Knight and Jaeschke, 2002). In addition, treatment with 10 mg/kg Z-VAD-fmk 1.5 and 3 h after AAP administration was ineffective in preventing AAP-induced cell death (Lawson et al., 1999). The same dose is highly effective in eliminating Fas- or TNF-receptor mediated apoptosis in vivo (Bajt et al., 2000; Jaeschke et al., 1998, 2000; Lawson et al., 1999). However, it remained unclear why pretreatment with a suicide substrate of caspases (10 mg/kg Z-VAD-fmk) can prevent AAP hepatotoxicity even in the absence of any caspase activation (El-Hassan et al., 2003). More recent findings provide an explanation for this unexpected effect. Even a small amount of the solvent dimethyl sulfoxide, which is necessary to get these inhibitors in solution, is sufficient to inhibit metabolic activation of AAP and prevent hepatotoxicity (Jaeschke et al., 2005). This explains why only treatment with caspase inhibitors before but not after AAP administration protects (Jaeschke et al., 2005). Thus, there is no reliable evidence in vivo or in primary hepatocytes that caspase inhibitors are able to protect against AAP hepatotoxicity (Jaeschke et al., 2005; Lawson et al., 1999; Nagai et al., 2002). Other arguments, which do not support the hypothesis that AAP-induced cell death is an apoptotic process rapidly deteriorating into secondary necrosis, include the fact that, during secondary necrosis, apoptotic features (e.g., massive caspase activation, morphological characteristics) are still clearly detectable (Bajt et al., 2000; Gujral et al., 2004; Jaeschke et al., 2004). None of these more specific parameters of apoptosis are detectable in quantities relevant to explain the extent of cell death seen during AAP hepatotoxicity.
Currently, there are only two experimental conditions where AAP unequivocally causes apoptotic cell death. First, exposure of hepatoma cell lines to high concentrations of AAP in cell culture will induce classical caspase-dependent apoptosis (Boulares et al., 2002; Macanas-Pirard et al., 2005). Since these hepatoma cell lines lack the capacity to metabolically activate a relevant amount of AAP and therefore do not properly mimic the initiating events of AAP hepatotoxicity, the relevance of these mechanisms for the pathophysiology in vivo has to be questioned. Second, AAP induces apoptosis in primary hepatocytes when the necrotic cell death is prevented by fructose and glycine treatment (Kon et al., 2004a). However, these findings in cell culture remain to be confirmed in vivo. In general, inhibition of AAP-induced cell necrosis in vivo results in a permanent protection and does not induce apoptosis.
|
GENOMICS AND PROTEOMICS APPROACHES |
|---|
In the recent past, livers from AAP-treated mice were subjected to genomics and proteomics analysis. When mRNA expression of livers from AAP-treated animals (300 mg/kg; 6 h) was compared to controls, numerous genes encoding stress proteins, cell cycle and growth inhibitors, adhesion molecules and structural proteins, inflammatory mediators, and cell signaling proteins were upregulated, and many genes involved in cell metabolism were downregulated (Reilly et al., 2001
). While the relevance for the pathophysiology of AAP-induced liver injury of some of the identified genes is known (e.g., heat shock proteins, heme oxygenase-1), the importance of most of the more than 300 up- or downregulated genes is still unclear. In a related study, which compared toxic and nontoxic doses of AAP at different time points, the authors demonstrated major changes in mitochondrial protein expression consistent with the hypothesis that mitochondria are a critical target for the toxicity (Ruepp et al., 2002). However, most gene expression alterations occurred after the protein changes, suggesting that genomic changes may modulate more downstream events of the toxicity rather than any early events. A more recent proteomic analysis of livers of two strains of mice, one with a higher susceptibility for AAP hepatotoxicity than the other, showed a more pronounced loss of several mitochondrial proteins in the susceptible strain (Welch et al., 2005). In addition, the resistant strain had a higher expression of a number of proteins related to stress response, antioxidant enzymes, and cell proliferation (Welch et al., 2005). Thus, genomic and proteomic analysis confirmed the importance of mitochondria as a target for AAP hepatotoxicity. The pathophysiological importance of many of the altered genes and proteins after AAP overdose remains to be investigated in more detail. |
SUMMARY |
|---|
A fraction of the dose of AAP is metabolically activated to a reactive metabolite (NAPQI), which first depletes cellular glutathione and subsequently covalently binds to cellular proteins (Fig. 3). These initiating events lead to disturbances of the cellular Ca2+ homeostasis, with increase of the cytosolic Ca2+ levels, Bax and Bid translocation to the mitochondria, and a mitochondrial oxidant stress and peroxynitrite formation. The Bcl-2 family members form pores in the outer mitochondrial membrane and release cytochrome c, Smac, AIF, and endonuclease G from the mitochondrial intermembrane space. Reactive oxygen species and peroxynitrite induce the membrane permeability transition, which causes the collapse of the mitochondrial membrane potential, eliminates ATP synthesis, and causes further release of mitochondrial proteins. The declining ATP levels appear to prevent caspase activation by the release of cytochrome c and Smac. AIF and endonuclease G translocate to the nucleus and induce DNA fragmentation, which is further aggravated by the nuclear Ca2+/Mg2+-dependent endonuclease DNAS1L3. The massive nuclear DNA damage and the rapid elimination of functional mitochondria, together with activation of intracellular proteases (calpains), lead to cell membrane failure and oncotic necrosis of the hepatocytes. The postulated intracellular signaling events after AAP overdose can explain the massive cell death and liver failure. However, many aspects are still unclear and require further investigation. In addition, it has to be kept in mind that AAP-induced cell death in vivo can be modulated by changes in the expression levels of P450 and phase II detoxification enzymes, variation in the GSH and antioxidant levels (nutritional status), and preexisting conditions affecting the susceptibility of hepatocytes (steatosis, mitochondrial abnormalities, inflammation). Therefore, to most effectively protect against AAP overdose, it is important to focus on central mechanisms of the pathophysiology. At the present time, this appears to be the metabolic activation as initiating event in the toxicity and mitochondrial dysfunction as the key cellular event that controls the propagation of the injury.
|
superoxide; ONOO–, peroxynitrite; PARP, poly(ADP-ribose) polymerase; Smac, second mitochondria-derived activator of caspases; tBid, truncated from of Bid. |
ACKNOWLEDGMENTS |
|---|
Work from the authors' laboratory was supported by National Institutes of Health grants R01 AA 12916 and R01 DK 070195.
|
REFERENCES |
|---|
Adams, M. L., Pierce, R. H., Vail, M. E., White, C. C., Tonge, R. P., Kavanagh, T. J., Fausto, N., Nelson, S. D., and Bruschi, S. A. (2001). Enhanced acetaminophen hepatotoxicity in transgenic mice overexpressing BCL-2. Mol. Pharmacol. 60, 907–915.
Adamson, G. M., and Harman, A. W. (1993). Oxidative stress in cultured hepatocytes exposed to acetaminophen. Biochem. Pharmacol. 45, 2289–2294.[CrossRef][ISI][Medline]
Akerman, P., Cote, P., Yang, S. Q., McClain, C., Nelson, S., Bagby, G. J., and Diehl, A. M. (1992). Antibodies to tumor necrosis factor-alpha inhibit liver regeneration after partial hepatectomy. Am. J. Physiol. 263, G579–G585.[ISI][Medline]
Bajt, M. L., Lawson, J. A., Vonderfecht, S. L., Gujral, J. S., and Jaeschke, H. (2000). Protection against Fas receptor-mediated apoptosis in hepatocytes and nonparenchymal cells by a caspase-8 inhibitor in vivo: Evidence for a postmitochondrial processing of caspase-8. Toxicol. Sci. 58, 109–117.
Bajt, M. L., Knight, T. R., Farhood, A., and Jaeschke, H. (2003). Scavenging peroxynitrite with glutathione promotes regeneration and enhances survival during acetaminophen-induced liver injury in mice. J. Pharmacol. Exp. Ther. 307, 67–73.
Bajt, M. L., Knight, T. R., Lemasters, J. J., and Jaeschke, H. (2004). Acetaminophen-induced oxidant stress and cell injury in cultured mouse hepatocytes: Protection by N-acetyl cysteine. Toxicol. Sci. 80, 343–349.
Bajt, M. L., Lemasters, J. J., and Jaeschke, H. (2005). Role of mitochondrial Bax translocation in acetaminophen-induced hepatic necrosis (abstract). Toxicol. Sci. 84(Suppl. 1), 215.
Birge, R. B., Bartolone, J. B., Nishanian, E. V., Bruno, M. K., Mangold, J. B., Cohen, S. D., and Khairallah, E. A. (1988). Dissociation of covalent binding from the oxidative effects of acetaminophen. Studies using dimethylated acetaminophen derivatives. Biochem. Pharmacol. 37, 3383–3393.[CrossRef][ISI][Medline]
Boulares, A. H., Zoltoski, A. J., Stoica, B. A., Cuvillier, O., and Smulson, M. E. (2002). Acetaminophen induces a caspase-dependent and Bcl-XL sensitive apoptosis in human hepatoma cells and lymphocytes. Pharmacol. Toxicol. 90, 38–50.[CrossRef][ISI][Medline]
Bourdi, M., Masubuchi, Y., Reilly, T. P., Amouzadeh, H. R., Martin, J. L., George, J. W., Shah, A. G., and Pohl, L. R. (2002). Protection against acetaminophen-induced liver injury and lethality by interleukin 10: Role of inducible nitric oxide synthase. Hepatology 35, 289–298.[CrossRef][ISI][Medline]
Burcham, P. C., and Harman, A. W. (1991). Acetaminophen toxicity results in site-specific mitochondrial damage in isolated mouse hepatocytes. J. Biol. Chem. 266, 5049–5054.
Chanda, S., Mangipudy, R. S., Warbritton, A., Bucci, T. J., and Mehendale, H. M. (1995). Stimulated hepatic tissue repair underlies heteroprotection by thioacetamide against acetaminophen-induced lethality. Hepatology 21, 477–486.[CrossRef][ISI][Medline]
Chao, D. T., and Korsmeyer, S. J. (1998). BCL-2 family: Regulators of cell death. Annu. Rev. Immunol. 16, 395–419.[CrossRef][ISI][Medline]
Chen, C., Hennig, G. E., and Manautou, J. E. (2003). Hepatobiliary excretion of acetaminophen glutathione conjugate and its derivatives in transport-deficient (TR-) hyperbilirubinemic rats. Drug Metab. Dispos. 31, 798–804.
Cohen, S. D., Pumford, N. R., Khairallah, E. A., Boekelheide, K., Pohl, L. R., Amouzadeh, H. R., and Hinson, J. A. (1997). Selective protein covalent binding and target organ toxicity. Toxicol. Appl. Pharmacol. 143, 1–12.[CrossRef][ISI][Medline]
Chiu, H., Gardner, C. R., Dambach, D. M., Durham, S. K., Brittingham, J. A., Laskin, J. D., and Laskin, D. L. (2003). Role of tumor necrosis factor receptor 1 (p55) in hepatocyte proliferation during acetaminophen-induced toxicity in mice. Toxicol. Appl. Pharmacol. 193, 218–227.[CrossRef][ISI][Medline]
Corcoran, G. B., Bauer, J. A., and Lau, T. W. (1988). Immediate rise in intracellular calcium and glycogen phosphorylase a activities upon acetaminophen covalent binding leading to hepatotoxicity in mice. Toxicology 50, 157–167.[CrossRef][ISI][Medline]
Corcoran, G. B., Racz, W. J., Smith, C. V., and Mitchell, J. R. (1985). Effects of N-acetylcysteine on acetaminophen covalent binding and hepatic necrosis in mice. J. Pharmacol. Exp. Ther. 232, 864–872.
Cover, C., Fickert, P., Knight, T. R., Fuchsbichler, A., Farhood, A., Trauner, M., and Jaeschke, H. (2005a). Pathophysiological role of poly(ADP-ribose) polymerase (PARP) activation during acetaminophen-induced liver cell necrosis in mice. Toxicol. Sci. 84, 201–208.
Cover, C., Mansouri, A., Knight, T. R., Bajt, M. L., Lemasters, J. J., Pessayre, D., and Jaeschke, H. (2005b). Peroxynitrite-induced mitochondrial and endonuclease-mediated nuclear DNA damage in acetaminophen hepatotoxicity. J. Pharmacol. Exp. Ther. 315, doi: 10.1124/jpet.105.088898.
Cressman, D. E., Greenbaum, L. E., DeAngelis, R. A., Ciliberto, G., Furth, E. E., Poli, V., and Taub, R. (1996). Liver failure and defective hepatocyte regeneration in interleukin-6-deficient mice. Science 274, 1379–1383.
Denicola, A., and Radi, R. (2005). Peroxynitrite and drug-dependent toxicity. Toxicology 208, 273–288.[CrossRef][ISI][Medline]
Donnelly, P. J., Walker, R. M., and Racz, W. J. (1994). Inhibition of mitochondrial respiration in vivo is an early event in acetaminophen-induced hepatotoxicity. Arch. Toxicol. 68, 110–118.[CrossRef][ISI][Medline]
El-Hassan, H., Anwar, K., Macanas-Pirard, P., Crabtree, M., Chow, S. C., Johnson, V. L., Lee, P. C., Hinton, R. H., Price, S. C., and Kass, G. E. (2003). Involve