Acetaminophen (APAP) is safe at therapeutic dosage but can cause severe hepatotoxicity if used at overdose. The mechanisms of injury are not yet fully understood, but previous reports had suggested that the mitochondrial permeability transition (mPT) may be involved in triggering hepatocellular necrosis. We aimed at inhibiting mitochondrial cyclophilin D (CypD), a key regulator of the mPT, as a potential therapeutic target in APAP hepatotoxicity. Wildtype mice treated with a high dose of APAP (600 mg/kg, intraperitoneal) developed typical centrilobular necrosis, which could not, however, be prevented by cotreatment with the selective CypD inhibitor, Debio 025 (alisporivir, DEB025, a nonimmunosuppressive cyclosporin A analog). Similarly, genetic ablation of mitochondrial CypD in Ppif-null mice did not afford protection from APAP hepatotoxicity. To determine whether APAP-induced peroxynitrite stress might directly activate mitochondrial permeabilization, independently of the CypD-regulated mPT, we coadministered the peroxynitrite decomposition catalyst Fe-TMPyP (10 mg/kg, intraperitoneal, 90 minutes prior to APAP) to CypD-deficient mice. Liver injury was greatly attenuated by Fe-TMPyP pretreatment, and mitochondrial 3-nitrotyrosine adduct levels (peroxynitrite marker) were decreased. Acetaminophen treatment increased both the cytosolic and mitochondria-associated P-JNK levels, but the c-jun-N-terminal kinase (JNK) signaling inhibitor SP600125 was hepatoprotective in wildtype mice only, indicating that the JNK pathway may not be critically involved in the absence of CypD.
These data support the concept that an overdose of APAP results in liver injury that is refractory to pharmacological inhibition or genetic depletion of CypD and that peroxynitrite-mediated cell injury predominates in the absence of CypD. (HEPATOLOGY 2011;)
Acetaminophen (APAP) is a widely used analgesic and antipyretic drug that is safe at therapeutic dosage. However, when given at overdose APAP can produce hepatic injury in both humans and mice. Despite extensive research over several decades, the underlying molecular mechanisms of hepatocyte injury are not yet completely understood,1, 2 limiting the development and therapeutic application of novel cytoprotective agents in APAP-induced liver injury. What has become clear is that in both the initial phases of cell injury (interactions of the thiol-reactive intermediate, N-acetyl-p-benzoquinone imine [NAPQI], with glutathione and proteins, accompanied by oxidant and nitrative stress) and the subsequent propagation phase (signaling followed by hepatocellular demise), mitochondria seem to play a key role.3-5 Evidence has been accruing that, following exposure of hepatocytes to APAP in vitro or in vivo, mitochondria readily undergo outer membrane permeabilization, thus inducing necrotic cell death, largely by way of caspase-independent mechanisms.6, 7
How exactly NAPQI and its downstream signaling events lead to mitochondrial permeabilization is currently not known. It has been suggested that the process may involve the mitochondrial permeability transition (mPT).8-11 The mPT is a functional term that involves sustained opening of a megapore that spans both the inner and outer mitochondrial membrane, allowing exchange of solutes ≤1.5 kDa, leading to mitochondrial swelling, rupture of the outer membrane, and release of proapoptotic proteins. Although the physiological properties of the mPT have been well studied, the molecular nature of this pore remains ill-defined. Originally, a crucial role was attributed to the ADP/ATP translocator (ANT) and the voltage-dependent anion carrier (VDAC), but this concept had to be revised recently as it was found that mitochondria from ANT- or VDAC-knockout mice were still able to undergo mPT. On the other hand, the matrix protein cyclophilin D (CypD) seems to be a critical player involved in the regulation of the mPT pore. Studies with isolated mitochondria from mice with a genetic deletion of CypD have clearly demonstrated that these mitochondria were much more resistant to mPT inducers than wildtype mitochondria (although they were not completely protected).12-15 As an alternative to genetic deletion of CypD, the interaction of CypD with the mPT pore can also be disrupted by pharmacologic inhibition, e.g., with cyclosporin A (CsA) or other specific cyclophilin ligands. Hence, the demonstration of protective effects afforded by CsA against toxic drug effects has been widely used to make an argument for the involvement of the mPT.
Based on this concept of CsA cytoprotection, a number of independent studies have provided experimental evidence that the mPT might indeed be involved in the hepatic toxicity induced by APAP.8, 10, 11 However, one caveat is that CsA, given at high dose, as used in some of the mouse studies,8, 11 can inhibit drug transporters at the canalicular membrane domain and also induce cholestasis.16 This could alter the kinetics of APAP and/or its metabolites. Furthermore, and importantly, CsA not only binds to mitochondrial CypD but also to other cyclophilin forms including the cytosolic CypA. The CypA/CsA complex subsequently binds to and inhibits calcineurin, a Ca2+/calmodulin-activated serine/threonine protein phosphatase that has been mechanistically implicated in the immunosuppressive effects of CsA.17 Finally, CsA has been shown to exert other, calcineurin-independent effects on jun-NH2-terminal kinase (JNK) signaling.18 Thus, the role of the CypD-dependent mPT in APAP hepatotoxicity, based solely on the protective effects afforded by CsA, should be revisited. Indeed, studies in isolated hepatocytes have provided evidence that with increasing time and cellular stress, CsA eventually loses its protective effects towards APAP-induced cell injury.10, 19 However, it is not known whether this occurs in vivo, and, importantly, the “CsA-insensitive” mechanism of APAP toxicity has remained enigmatic.
The aim of this study was to investigate whether APAP exerts mitochondrial permeabilization through the mPT and/or through other, CypD-independent mechanisms, using both pharmacologic inhibitors of CypD in vivo and a genetic approach with CypD-deficient (Ppif−/−) mice. The data suggest that high doses of APAP induce mitochondrial peroxynitrite stress that directly triggers mitochondrial permeabilization without the involvement of CypD.
Materials and Methods
Debio 025 (alisporivir, DEB025) was kindly provided by DebioPharm (Lausanne, Switzerland). Acetaminophen was purchased from Sigma (St. Louis, MO); Solutol HS-15 from BASF (Ludwigshafen, Germany); Fe-TMPyP (5,10,15,20-tetrakis(N-methyl-4′-pyridyl)porphyrinato iron(III) from Cayman Chemicals (Ann Arbor, MI); SP600125 from Enzo Life Sciences (Plymouth Meeting, PA); anti-JNK and anti-P-JNK from Cell Signaling (Danvers, MA); anti-cyclophilin D from MitoSciences (Eugene, OR); and anti-3-NT antibody from Abcam (Cambridge, MA).
Animals and Genotyping.
The study design and all protocols for animal care and handling were approved by the Institutional Animal Care and Use Committee of the University of Connecticut. C57BL/6J mice and heterozygous breeder pairs of cyclophilin D-knockout mice (Ppiftm1Mmos/J) were obtained from The Jackson Laboratory (Bar Harbor, ME). For the latter, a breeding colony was established in our animal facilities over five generations before wildtype and knockout mice were used for the study (originally, the mice had been kept on a mixed C57BL/6J × 129X1/SvJ background and bred to C57BL/6J for one generation). All mice were kept on a 14:10-hour light-dark cycle and under controlled environmental conditions. They received mouse chow (Tekland Global Rodent Diet, Harlan Laboratories, Boston, MA) and water ad libitum. The animals were 10-14 weeks old at the time of experimentation. Tail biopsy genotyping was performed according to standard procedures.
APAP was dissolved in 10% (in phosphate-buffered saline [PBS]) Solutol HS-15 solution and administered intraperitoneally (600 mg/kg) to mice after overnight food deprivation.
The selection of this hepatotoxic dose range was based on previous studies with APAP in C57BL/6 mice.20 Debio 025 was dissolved in 10% Solutol HS-15 solution and administered intraperitoneally (10 mg/kg) 90 minutes post-APAP to minimize interference with APAP bioactivation. Fe-TMPyP was dissolved in 10% Solutol HS-15 and injected intraperitoneally (10 mg/kg) 90 minutes prior to APAP administration. This dose results in effective tissue levels (low μM range) that allow the compound to react with peroxynitrite.21
Assessment of Liver Injury.
Serum activity of alanine aminotransferase (ALT) was determined using the Infinity ALT test kit (Fisher Scientific, Waltham, MA). For histopathological analysis, small pieces of liver were fixed in 10% buffered formalin. The tissue was then processed and embedded in paraffin blocks. Tissue sections (5 μm) were stained with hematoxylin-eosin (H&E) and analyzed. The degree of liver injury was scored as described.20
Determination of GSH.
Hepatic glutathione (GSH) was determined in whole-liver homogenates (15-25 mg each of frozen liver tissue) using 5,5′-dithio-bis(2-nitrobenzoic acid, DTNB) and a kinetic assay.22 Absorbance at 412 nm was recorded every 30 seconds for 2 minutes for a total of five readings, and GSH levels were calculated from a standard curve.
Isolation of Hepatic Mitochondria and Cytosol.
Liver mitochondria and cytosol were isolated according to standard protocols23 and the degree of purification of the subcellular fractions was assessed by the inclusion of organelle markers in Western blots. A protease inhibitor cocktail (Sigma, St. Louis, MO) was added to the isolation buffer. Protein content was determined with the Bradford assay using albumin as the reference protein. Both mitochondria and cytosolic fractions were kept at −80°C until analysis.
Cytosolic or mitochondrial protein fractions were reduced, denatured, and loaded on a 10% or 15% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gel. The proteins were transferred to polyvinylidene fluoride membranes (Bio-Rad Laboratories, Hercules, CA). Membranes were then probed with anti-JNK (1:1,000), anti-P-JNK (1:1,000), anti-CypD (1:1,000), or anti 3-nitrotyrosine (3-NT) (1:1,400) antibodies. The antigen-antibody complexes were visualized after incubation with horseradish peroxidase-conjugated secondary antibody using enhanced chemiluminescence detection system (Millipore, Billerica, MA). Anti-β-actin or anti-VDAC antibodies were used to control for equal protein loading for cytosol or mitochondria, respectively.
All measurements were performed at least in duplicate and the mean value was used as one data point. Data were expressed as mean ± standard deviation (SD) with P ≤ 0.05 indicating significance. When normality of distribution passed, Student's t test (for comparison of two groups) or a standard analysis of variance (ANOVA) (for comparison of three or more groups) was used, followed by Dunnett's test for multiple comparisons versus control group. If normality failed, a nonparametric Kruskal-Wallis ANOVA on ranks was used, followed by Dunn's test for multiple comparison versus controls. Data were analyzed with InStat v. 3.06 software (GraphPad).
Pharmacologic Inhibition or Genetic Depletion of Mitochondrial CypD Does Not Protect from APAP Hepatotoxicity.
To investigate the mechanistic role of the CypD-regulated mPT versus other modes of cell death in APAP-induced liver injury, we used a previously characterized mouse model.20 Male wildtype (Ppif+/+) mice were administered APAP (600 mg/kg, intraperitoneal). As expected, APAP caused typical centrilobular necrosis that was evident at 8 hours postdose and became more severe at 24 hours, paralleled by highly increased plasma ALT activity (Fig. 1A,B,D). Because the choice of the solvent can have significant effects on the bioactivation of APAP and/or the subsequent recruitment of immune cells24 and, hence, on the extent of liver injury, we first determined the effects of Solutol HS-15, a widely used solvent for parenteral administration of lipophilic compounds, and compared it with warm saline used to dissolve APAP. We found that Solutol HS-15, in contrast to dimethyl sulfoxide, had no apparent effects on plasma ALT activity (Table 1). Therefore, Solutol HS-15 was used as vehicle for all subsequent experiments.
|Treatment||Plasma ALT activity (U/L)|
|Saline alone||18 ± 2|
|Solutol HS-15 (9% in PBS) alone||20 ± 12|
|DMSO (5% in PBS) alone||22 ± 3|
|APAP in saline||1911 ± 1939|
|APAP in Solutol HS-15||2092 ± 1374|
|APAP in DMSO (0.5%)||289 ± 223*|
|APAP in DMSO (5%)||61 ± 24*|
Previous reports from several laboratories have demonstrated that CsA can effectively protect mouse hepatocytes from APAP-induced injury both in vitro10 and in vivo.8, 11 However, CsA can have a number of off-target effects including those unrelated to CypD. To avoid these confounding factors we used the CsA analog, Debio 025, which is a more selective mitochondrial CypD inhibitor and whose potency to inhibit the immune system (through calcineurin-mediated pathways) is >3,000-fold lower than that of CsA.25 Debio 025 (10 mg/kg, intraperitoneal) was injected 1.5 hours post-APAP (when APAP bioactivation was largely completed and the bulk of hepatic GSH already consumed by NAPQI),26 thereby minimizing drug-drug interactions. Surprisingly, we found that Debio 025 did not protect from APAP-induced hepatotoxicity (Fig. 1C,D). A pilot study had revealed that there was a similar lack of protective effects when Debio 025 was given simultaneously with APAP (data not shown), indicating that the lack of protection was not simply due to the delayed administration of the CypD inhibitor. These findings suggest that, besides the CypD-dependent mode of the mPT, there may be another mode of mitochondrial permeabilization induced by high-dose APAP.
To further substantiate these findings and to fully exclude any possible drug-drug interactions due to the presence of the pharmacologic inhibitors, we next determined the extent of liver injury induced by APAP in a genetic mouse model of CypD deletion (Ppif−/− mice) (Fig. 2A). We first had to ascertain that these CypD-deficient mice exhibited similar rates of APAP bioactivation as their wildtype controls. Therefore, we measured hepatic GSH consumption after a hepatotoxic dose of APAP in Ppif−/− mice and their wildtype littermates over the first 90 minutes (an established marker for the extent of NAPQI formation).26 Although the Ppif−/− mice had initially higher (+30%) hepatic GSH levels, we found no significant differences in the extent of GSH depletion between the two genotypes (Fig. 2B). Next, we assessed the degree of liver injury after 4, 8, and 24 hours in both Ppif+/+ and Ppif−/− mice injected with APAP (600 mg/kg, intraperitoneal). In line with the results from the experiments with Debio 025, Ppif−/− mice were not protected from APAP toxicity at this high dose, but instead developed typical centrilobular necrosis after 24 hours, the manifestation of which was not different from that in wildtype controls (Fig. 2D). Taken together, these data indicate that the mitochondrial signaling involved in APAP hepatotoxicity includes a CypD-independent mode, at least at this high dose. In contrast, at a much lower dose inhibition of the CypD pathway may still afford cytoprotection, as shown in a recent report.27
Mitochondrial CypD-Independent Mode of APAP-Induced Cell Death Is Mediated by Mitochondrial Peroxynitrite Stress and Can Be Attenuated by the Metalloporphyrin, Fe-TMPyP.
It has been demonstrated previously that in isolated mitochondria from Ppif−/− mice the sensitivity to thiol oxidation remained unchanged despite the lack of CypD.12 This oxidative activation of the mPT could be due to the well-documented generation of peroxynitrite in mitochondria following APAP administration to mice3, 28 or cultured mouse hepatocytes.29 Commensurate with these earlier data, we confirmed the presence of 3-nitrotyrosine (3-NT) adducts (a biomarker for peroxynitrite) in mitochondrial fractions of both wildtype and Ppif-null mice (Fig. 3). Next, to address the question of whether peroxynitrite is causally involved in the pathogenesis, we coadministered the cell-permeable peroxynitrite scavenger, Fe-TMPyP (10 mg/kg, intraperitoneal) 90 minutes prior to APAP. We found that, in Ppif−/− mice, Fe-TMPyP not only decreased the extent of 3-NT adduction (Fig. 3B), but also attenuated the degree of APAP hepatotoxicity (Fig. 2E; Table 2). In contrast, administration of Fe-TMPyP to wildtype mice was not hepatoprotective. This indicates that the metalloporphyrin did not simply interfere with APAP bioactivation. Moreover, these findings suggest that peroxynitrite stress is overruled by the CypD-regulated mPT in normal animals, but becomes critical in the absence of CypD.
|Histopathology score (24 hours post-APAP)|
|Vehicle control (n = 3)||3||0||0||0||0||0|
|APAP (n = 4)||0||0||0||3||1||0|
|APAP + Fe-TMPyP (n = 6)||1||1||3||1||0||0|
|Fe-TMPyP (n = 3)||3||0||0||0||0||0|
|Vehicle control (n = 3)||3||0||0||0||0||0|
|APAP (n = 5)||0||0||1||3||1||0|
|APAP + Fe-TMPyP (n = 6)||0||1||0||2||3||0|
Deletion of CypD Dissociates APAP-Induced Hepatic Injury from the JNK Signaling Pathway in Ppif−/− Mice.
It has been previously established that the JNK pathway plays a critical role in mediating APAP hepatotoxicity in mice.3, 20, 30, 31 To ascertain whether JNK signaling is also involved in the peroxynitrite-mediated, CypD-independent mode of mitochondrial permeabilization, we analyzed the levels of hepatic JNK and phospho-JNK in both wildtype and Ppif−/− mice treated with APAP. As shown in Fig. 4A, both the cytosolic and mitochondrial P-JNK levels were increased over vehicle controls in both genotypes. The relative abundance of cytosolic and mitochondria-associated P-JNK was similar in wildtype and Ppif−/− mice, indicating that the loss of CypD was not coupled with impaired mitochondrial translocation of P-JNK. We also did not find any significant differences in the total JNK levels between the two genotypes. The absence of immunoreactive β-actin (a marker for cytosol) and VDAC (a marker for mitochondria) in isolated mitochondria and cytosol, respectively, was taken as an indicator of a high degree of purification of the subcellular fractions (Fig. 4B). Next, to explore the functional role of JNK signaling downstream of its phosphorylation, we cotreated wildtype and CypD-deficient mice with SP600125, a cell-permeable, ATP-competitive JNK inhibitor (30 mg/kg, intraperitoneal, 90 minutes post-APAP).32 Although SP600125 completely protected wildtype mice from APAP hepatotoxicity, which is commensurate with previous reports, Ppif−/− mice were refractory to the effects of SP600125, as reflected both in the histopathologic analysis (Fig. 5A) and serum ALT activity (Fig. 5B). This suggests that the JNK signaling pathway is not critically involved in the CypD-independent mode of the mPT.
The aim of this study was to investigate whether pharmacologic inhibition or genetic deletion of mitochondrial CypD, a key regulator of the mPT pore, would render mice resistant to a high dose of APAP (the dose used in this mouse study, 600 mg/kg, is clinically relevant and actually corresponds to a human dose of 3.5 g, if corrected for interspecies differences using a dose scaling factor33). Targeting CypD has been successfully used in other disease models where the mPT has been implicated, e.g., drug enteropathy,34 renal ischemia,35 or myopathy.36 However, here we found that neither inhibition of mitochondrial CypD with the nonimmunosuppressant cyclosporin analog, Debio 025, nor genetic homozygous deletion of Ppif (coding for mitochondrial CypD) protected from a high hepatotoxic dose of APAP. Therefore, other, CypD-independent mechanisms must have been activated to cause liver injury. The findings that APAP-induced hepatic toxicity in Ppif−/− mice was greatly attenuated by cotreating the mice with the peroxynitrite decomposition catalyst Fe-TMPyP indicate that the peroxynitrite stress-mediated mode of cell death induction was sufficient to trigger overt liver injury, independently of CypD. These conclusions are in line with earlier reports on mitochondrial swelling in isolated mitochondria from Ppif−/− mice that had demonstrated that the CypD-dependent mode of the mPT is primarily activated by increased mitochondrial Ca2+ levels, whereas the CypD-independent mode(s) was activated by thiol oxidation.12 This study shows for the first time that the CypD-independent mode of hepatocellular injury is relevant in vivo.
Because metalloporphyrins including Fe-TMPyP are potent peroxynitrite-scavenging species that can catalytically reduce peroxynitrite even in the absence of added reductants,21, 37 the results of this study suggest that peroxynitrite and other prooxidant species are primarily responsible for APAP-induced cell demise. These data are also commensurate with a previous study demonstrating that administration of Mn-TMBAP (another peroxynitrite scavenger) could prevent APAP-induced liver injury in female Balb/c mice38 and that glutathione (a nonselective antioxidant involved in reversing peroxynitrite-mediated secondary effects) protected from APAP liver injury.39 Extensive and sustained peroxynitrite stress is therefore a causal factor in APAP-induced hepatocyte injury that becomes predominant in the absence of CypD. It may even be possible that initially the presence of CypD may be a protective factor against prooxidant stress. Not only is CypD, like other cyclophilins, a molecular chaperone with foldase activity, but CypD also has recently been implicated in being a redox sensor in human neuroblastoma SH-SY5Y cells. Specifically, during oxidant stress the CypD molecule, which features four sulfhydryl groups, can undergo disulfide formation.40 Thus, deletion of CypD could sensitize mitochondria to oxidant stress. According to this scenario, functional CypD (in wildtype mice) would initially reduce the oxidant stress posed by APAP-induced increased peroxynitrite, which would explain why CsA initially protects from APAP hepatotoxicity in hepatocytes but gradually loses its protective effects.10 It has indeed been shown that mitochondrial CypD itself can be a major target of nitration in a murine model of amyotrophic lateral sclerosis.41 Gradually, as CypD is depleted, this antioxidant function in mitochondria is lost, and peroxynitrite or other prooxidants can trigger a CypD-independent mode of cell death.
Our results obtained with Ppif-null mice are supported by previous reports on the response of mitochondria to inducers of the mPT in vitro. For example, isolated liver mitochondria from Ppif−/− mice were CsA-insensitive and required 2-fold greater Ca2+ loads than those from wildtype mice in order for the mPT pore to open, but the sensitivity to thiol oxidants remained unchanged.12 Furthermore, liver mitochondria isolated from Ppif−/− mice underwent permeabilization after apoptotic stimuli induced by proapoptotic BH3-only proteins (such as Bad, Bid, or Bax) but were protected from Ca2+/ROS-induced necrotic cell death.13, 14 This supports our concept that the mode of APAP-induced cell death is dependent on the presence or absence of CypD. In fact, CypD has been reported to exert antiapoptotic effects in certain cell types.42, 43
The putative pathways of APAP-mediated mitochondrial permeabilization are summarized in Fig. 6. However, the exact mechanism of how the CypD-independent mode of cell death is triggered still remains unclear. It has been suggested previously that low-level chemical stress may induce the “regulated” mPT (CsA-sensitive), whereas high-level chemical stress triggers an “unregulated” mode of mPT (CsA-insensitive).10 One possible mediator of this “unregulated” mPT may be the membrane-bound glutathione transferase (mtMGST1) form that is associated with the inner mitochondrial membrane.44 The thiol group of mtMGST1 (Cys 49) can be directly oxidized by peroxynitrite in each subunit of the homotrimer, thereby activating MGST1 and inducing the mPT.45, 46 Alternatively, BH3-only proteins may be activated by oxidant/nitrative stress and induce the mPT in a CsA-independent manner. A novel finding in our study was that the CypD-independent pathway does not involve JNK signaling, as judged from the lack of response to the JNK inhibitor, SP600125. Hence, the JNK pathway seems to be critically involved in the regulated mPT mode only.
In summary, and with a view to potential therapeutic applications of these findings, the results of this study suggest that, unlike for other mitochondria-mediated disease models,34 APAP-induced liver injury cannot simply be antagonized with CypD inhibitors such as Debio 025. A recent study has shown that deletion of CypD (Ppif knockout mice) protected against low doses (200 mg/kg, intraperitoneal) of APAP27; however, when higher and clinically more relevant doses of APAP overdose were used, the protective effects were lost (this study). Therefore, peroxynitrite scavengers such as novel nontoxic metalloporphyrins may be more promising therapeutic candidates; in fact, this class of cytoprotective agents has recently received increased attention in the potential treatment of oxidative/nitrative stress-mediated cardiovascular injury.47
We thank DebioPharm, Lausanne, Switzerland, for kindly providing Debio 025 (alisporivir, DEB025).