Cytochrome P4502E1 sensitizes to tumor necrosis factor alpha–induced liver injury through activation of mitogen-activated protein kinases in mice

Authors

  • Defeng Wu,

    1. Department of Pharmacology and Systems Therapeutics, Mount Sinai School of Medicine, New York, NY
    Search for more papers by this author
  • Arthur Cederbaum

    Corresponding author
    1. Department of Pharmacology and Systems Therapeutics, Mount Sinai School of Medicine, New York, NY
    • Department of Pharmacology and Systems Therapeutics, Box 1603, One Gustave L. Levy Place, Mount Sinai School of Medicine, New York, NY 10029
    Search for more papers by this author
    • fax: 212-996-7214


  • Potential conflict of interest: Nothing to report.

Abstract

The goal of this study was to evaluate the role of mitogen-activated protein kinase (MAPK) in cytochrome P4502E1 (CYP2E1) potentiation of lipopolysaccharide or tumor necrosis factor alpha (TNF-α)–induced liver injury. Treatment of C57/BL/6 mice with pyrazole (PY) plus lipopolysaccharide (LPS) induced liver injury compared with mice treated with PY or LPS alone. The c-Jun N-terminal kinase (JNK) inhibitor SP600125 or p38 MAPK inhibitor SB203580 prevented this liver injury. PY plus LPS treatment activated p38 MAPK and JNK but not extracellular signal-regulated kinase (ERK). PY plus LPS treatment triggered oxidative stress in the liver with increases in lipid peroxidation, decrease of glutathione (GSH) levels, and increased production of 3-nitrotyrosine adducts and protein carbonyl formation. This oxidative stress was blocked by SP600125 or SB203580. PY plus LPS treatment elevated TNF-α production, and this was blocked by SP600125 or SB203580. Neither SP600125 nor SB203580 affected CYP2E1 activity or protein levels. Treating C57/BL/6 mice with PY plus TNF-α also induced liver injury and increased lipid peroxidation and decreased GSH levels. Prolonged activation of JNK and p38 MAPK was observed. All of these effects were blocked by SP600125 or SB203580. In contrast to wild-type SV 129 mice, treating CYP2E1 knockout mice with PY plus TNF-α did not induce liver injury, thus validating the role of CYP21E1 in this potentiated liver injury. Liver mitochondria from PY plus LPS or PY plus TNF-α treated mice underwent calcium-dependent, cyclosporine A–sensitive swelling, which was prevented by SB203580 or SP600125. Conclusion: These results show that CYP2E1 sensitizes liver hepatocytes to LPS or TNF-α and that the CYP2E1-enhanced LPS or TNF-α injury, oxidant stress, and mitochondrial injury is JNK or p38 MAPK dependent. (HEPATOLOGY 2008.)

Chronic ethanol consumption reduces intestinal epithelial resistance to bacteria entry by disrupting the epithelial barrier and causing enhanced permeability through this “leaky gut.” These changes elevate endotoxin (lipopolysaccharide; LPS) levels in the portal vein.1 LPS-mediated activation of Kupffer cells is important in early ethanol-induced liver injury.2, 3 Tumor necrosis factor alpha (TNF-α), an inflammatory mediator, induces cell death. In alcoholic liver disease, several lines of evidence support the involvement of TNF-α. In an intragastric alcohol feeding model, liver injury was reduced in TNF-R1 knockout mice,4 and by neutralizing antibodies against TNF-α5 or antibiotic treatment to remove bacteria.6 TNF-α levels increased in a chronic alcohol feeding model in which gut permeability to bacterial-derived LPS is increased.7

Both c-Jun N-terminal kinase (JNK) and p38 mitogen-activated protein kinase (MAPK) are transiently activated by TNF-α but show a more prolonged activation under apoptotic conditions.8, 9 On activation, JNK translocates into the nucleus and enhances the transcriptional activity of transcription factors such as c-jun, activating transcription factor 2, and activator protein-1 (AP-1). JNK also has functions not related to c-jun phosphorylation such as direct effects on mitochondria.10, 11 P38 MAPK may activate various transcription factors including activating transcription factor 2, C/EBP homologous protein (CHOP). cAMP-responsive element binding protein (CREB), EST-like protein-1 (ELK1), enhance TNF-α–induced production of interleukin-1 and interleukin-6, and may be a mediator of nuclear factor kappa B (NF-κB) transactivation.8, 12

Ethanol consumption induces oxidative stress, which plays a major role in mechanisms by which ethanol causes liver injury.13 Induction of cytochrome P4502E1 (CYP2E1) by ethanol is 1 pathway in which ethanol induces oxidative stress.14 CYP2E1 metabolizes many small molecular substrates to more reactive toxic products.15, 16 CYP2E1 is an effective generator of reactive oxygen species.14, 15, 17 We evaluated whether an increase in CYP2E1 could sensitize liver to toxins such as Fas or LPS under conditions in which normal liver is usually resistant to these compounds. Induction of CYP2E1 did potentiate the toxicity of Fas or LPS to the liver.18, 19 The goal of the current study was to evaluate the role of MAPKs in the CYP2E1 potentiation of LPS (TNF-α)-induced liver injury in vivo.

Abbreviations

3-NT, 3-nitrotyrosine adducts; ALT, aminotransferase; AP-1, activator protein-1; AST, aspartate aminotransferase; CYP2E1,cytochrome P450 2E1; ELISA, enzyme-linked immunosorbent assay; ERK, extracellular signal-regulated kinase; GSH, glutathione; JNK, c-Jun N-terminal kinase; LPS, lipopolysaccharide; MAPK, mitogen-activated protein kinase; MDA, malondialdehyde; NF-κB, nuclear factor kappa B; PY, pyrazole; ROS, reactive oxygen species; TNF-α, tumor necrosis factor alpha.

Materials and Methods

Materials.

Most antibodies were from Santa Cruz Biotechnology Inc. (Santa Cruz, CA). Lipopolysaccharide serotype 055:B5 and pyrazole were from Sigma Chemical Co. (St. Louis, MO). Anti-Nitrotysosine (3-NT) polyclonal antibody was from Upstate (Lake Placid, NY). DNA Fragmentation Detection Kit, SP600125, and SB203580 were from Calbiochem (La Jolla, CA). Oxyblot Protein Oxidation Detection Kit was from Chemicon International (Temecula, CA). Mouse TNF-α enzyme-linked immunosorbent assay (ELISA) kit was from Pierce (Rockford, IL). C-jun Activation Assay Kit and Nuclear Extract Kit were from Active Motif (Carlsbad, CA). Caspase 3 and Caspase 8 substrates were from Calbiochem (La Jolla, CA). Alanine aminotransferase (ALT), aspartate aminotransferase (AST) measuring kit was from Thermo Fisher Scientific Inc. (Waltham, MA). Murine TNF-α was from Fitzgerald (Concord, MA).

Animal Models.

Wild-type C57/BL/6 mice (Charles River Laboratory) received humane care, and studies were carried out according to criteria outlined in the National Institutes of Health Guide for the Care and Use of Laboratory Animals and with approval of the Mount Sinai Animal Care and Use Committee. Mice, 19 to 21 g body weight, were injected intraperitoneally with pyrazole (PY), 150 mg/kg, once per day for 2 days to induce CYP2E1 activity; the control mice were injected with 0.9% saline. On the third day, some mice were pre-injected (intraperitoneally) with either the JNK inhibitor SP600125 or the p38 MAPK inhibitor SB203580, 15 mg/kg body weight, respectively, or with saline. After 1 to 2 hours, mice were injected intraperitoneally with LPS, 4 mg/kg body weight for 24 hours to induce liver injury. In some experiments, mice were injected with murine TNF-α 50 μg/kg, instead of LPS. The control mice were injected with 0.9% saline instead of LPS. Twenty-four hours after LPS or saline injection, blood was collected from the retroorbital venous sinus under anesthesia, and the mouse livers were removed.

Liver Pathology.

Liver samples were fixed in 10% formalin solution, embedded in paraffin, and stained with hematoxylin-eosin for pathological evaluation. Unstained sections were used for immunohistochemistry analysis. Serum was used to determine ALT and AST levels.

Nuclear DNA Fragmentation and Caspase 3 and 8 Activity Determination.

Liver sections were used to evaluate nuclear DNA fragmentation as outlined in the Calbiochem DNA Fragmentation Detection Kit. Positive cell nuclei were stained brown, and the percentage of stained cells was calculated. Activities of caspase 3 and caspase 8 were determined with 50 μmol/L caspase 3-substrate II, AC-DEVD-AMC, or caspase 8 substrate, Granzyme B substrate II, Z-IETD-AFC, and 200 μg supernatant protein. Fluorescence associated with the released aminomethylcoumarin (AMC) (excitation at 380 nm, emission at 460 nm) and aminofluoromethylcoumarin (AFC) (excitation at 400 nm, emission at 505 nm) was expressed as arbitrary units per milligram liver protein.

Activation of MAPKs.

Liver lysates were prepared by homogenizing in phosphate-buffered saline buffer containing a cocktail of proteinase inhibitors (Sigma). Twenty micrograms lysate protein was subjected to an 8% sodium dodecyl sulfate polyacrylamide gel electrophoresis. The blotted membranes were incubated with p-JNK or pp38 MAPK or p-extracellular signal-regulated kinase (ERK) antibodies to determine phosphorylated (activated) JNK, p38, and ERK. The membranes were rinsed with Re-Blot Plus Strong Solution (Chemicon International) and re-incubated with JNK, p38, and ERK antibodies to determine total levels of MAPK. The activation levels were determined by calculating the ratio of phosphorylated MAPK over nonphosphorylated MAPK.

P-c-jun Translocation and AP-1 Binding.

The activation of c-jun to p-c-jun and its translocation to the nucleus were determined in liver sections using a p-c-jun monoclonal antibody and the Immunoperoxidase Secondary Detection System (Chemicon, Temecula, CA). AP-1 binding was detected using an ELISA kit: Trans AM AP-1c-Fos/FosB/C-Jun/Jun D from Active Motif.

Intracellular Glutathione, Lipid Peroxidation Nitrotyrosine Adducts, and Protein Carbonyl Formation.

Glutathione(GSH) levels and lipid peroxidation analysis were determined as described previously20 and results expressed as nanomoles per milligram of protein. Nitrotyrosine protein adducts were determined with a slot blot method using a polyclonal antibody against nitrotyrosine (3-NT, Upstate, Lake Placid, NY). Bands were scanned using NIH ImageJ version 1.36b software, and results were expressed as arbitrary units. Liver protein oxidation was determined by measurement of protein carbonyl formation using the Oxyblot Protein Oxidation Detection Kit. Bands were quantified using NIH ImagJ software and results expressed as arbitrary units.

Mitochondrial Membrane Swelling.

Mitochondria were prepared by differential centrifugation, and freshly prepared mitochondria were used for the membrane swelling measurement. One milligram mitochondrial protein was mixed with 2 mL incubation solution containing 150 mmol/L KCl, 10 mmol/L Tris-mops, 5 mmol/L glutamate, 2.5 mmol/L malate, 1 mmol/L KPi. Swelling was determined in the absence or presence of 50 μmol/L or 100 μmol/L CaCl2. The absorbance decrease at 540 nm over 5 minutes was detected in a Shimadzu UV160 spectrophotometer.

CYP2E1 and TNF-α Content.

CYP2E1 activity was measured by the rate of p-nitrophenol oxidation to p-nitrocatechol.21 CYP2E1 protein content was determined by immunoblot analysis. Serum TNF-α levels were determined using a mouse TNF-α ELISA kit (Pierce).

Statistical Analysis.

Statistical analysis was performed using one-way analysis of variance with subsequent post hoc comparisons by Scheffe. Values reflect means ± standard error, and the numbers of experiments are given in the figure legends.

Results

Effect of SP600125 and SB203580 on Pyrazole Plus LPS–Induced Liver Injury.

ALT and AST were elevated in mice treated with PY plus LPS as compared with saline, PY, or LPS alone groups (Fig. 1A). The JNK inhibitor SP600125 or the p38 MAPK inhibitor SB203580 lowered the PY plus LPS–elevated ALT and AST (Fig. 1A). Liver morphology evaluation indicated that PY plus LPS induced extensive liver damage including large areas of degeneration and infiltration of inflammatory cells in both the periportal and pericentral areas (Fig. 1B, arrows). Liver morphology was significantly improved in mice treated with SP600125 or SB 203580 (Fig. 1B). The mice treated with PY plus LPS had higher caspase 3 activity compared with mice treated with PY or LPS alone. Caspase 8 activity was similar in the PY plus LPS and LPS groups. SB203580 or SP600125 lowered caspase 3 and caspase 8 activities in the PY plus LPS group (Fig. 1C). PY plus LPS treatment induced DNA fragmentation compared with the PY or LPS alone groups, and SP600125 or SB203580 lowered this (Fig. 1D,E).

Figure 1.

Effect of SP600125 and SB203580 on PY plus LPS–induced liver injury: C57/BL/6 mice were injected (intraperitoneally) with PY once per day for 2 days. On the third day, the mice were pretreated with either saline or the JNK inhibitor SP600125 or the p38 MAPK inhibitor SB203580 1 hour before administration of LPS. Mice were killed after 24 hours. Other mice received either saline or pyrazole alone or LPS alone. (A) Blood was collected from the retroorbital venous sinus under anesthesia and ALT and AST determined. Experiments were repeated 4 times with 3 mice in each group. (#, ##P < 0.05, compared with saline, PY, or LPS alone; *, **P < 0.05, compared with PY + LPS). (B) Liver sections were stained with hematoxylin-eosin for pathological morphology evaluation. The arrows indicate the zone undergoing degeneration changes. (C) Caspase 3 and caspase 8 activities. Experiments were repeated 4 times with triplicate samples in each group. #P < 0.05, compared with saline, PY, or LPS alone; *, **P < 0.05 compared with PY + LPS group. (D) Nuclear DNA fragmentation. Arrows indicate positive stained nuclei and the percentage of positive hepatocytes are shown in the bottom left of each panel. (E) DNA fragmentation was counted from 5 different fields, and results from 3 experiments are summarized. **P < 0.05, compared with saline, PY, or LPS alone; *P < 0.05, compared with PY + LPS.

Pyrazole Plus LPS Activate JNK and p38 MAPK.

PY alone did not induce JNK (Fig. 2A), p38 MAPK (Fig. 2B), or ERK (Fig. 2C) phosphorylation. LPS alone slightly increased JNK phosphorylation but not p38 MAPK or ERK. PY plus LPS treatment resulted in an increase in phosphorylation of JNK and p38 MAPK but not ERK (Fig. 2A-C). P-JNK began to increase as early as 3 hours after injection of LPS into PY-treated mice compared with injection of LPS into saline-treated mice, and JNK remained activated 24 hours after LPS injection into the PY-treated mice. P38 MAPK was only activated 24 hours after LPS administration into PY-treated mice as compared with saline-treated mice (Fig. 2D,E). SP600125 but not SB203580 blocked PY plus LPS–induced phosphorylation of JNK (Fig. 2F). Conversely, the p38 MAPK inhibitor but not the JNK inhibitor blocked the PY plus LPS–induced phosphorylation of p38 MAPK (Fig. 2G).

Figure 2.

Activation of JNK and p38 MAPK. (A, B, and C) Levels of phosphorylated JNK, p38 MAPK, or ERK and total levels of JNK, p38 MAPK, or ERK. Numbers below the blots refer to the phosphorylated/total ratio. Time course activation of JNK (D) and p38 MAPK (E) by PY plus LPS or by LPS alone. Mice were killed 3, 8, or 24 hours after addition of LPS to pyrazole-treated or saline-treated animals. Inhibition effect of SP600125 on activation of JNK (F) or of SB203580 on activation of p38 MAPK (G). Mice were killed 24 hours after addition of LPS to pyrazole-treated animals.

C-jun Activation and AP-1 Binding.

JNK may play a role in cell toxicity through activation of c-jun to p-c-jun.22 LPS plus PY treatment induced p-c-jun nuclear translocation as reflected by the higher content of stained nuclei compared with the PY or LPS alone (Fig. 3A). This enhanced c-jun activation was lowered by SP600125 but not by SB203580 (Fig. 3A). PY plus LPS treatment increased AP-1 binding compared with saline, PY, or LPS alone, and SP600125 but not SB203580 blocked this AP-1 binding (Fig. 3B). These results suggest that the activation of JNK by PY plus LPS activates a JNK downstream target c-jun to increase AP-1 binding activity.

Figure 3.

P-c-Jun and AP-1 binding. (A) Immunohistochemistry with p-c-Jun monoclonal antibody. Arrows indicate the positively stained nuclei. (B) p-c-Jun and AP-1 binding was determined by an ELISA method. *P < 0.05 compared with saline, PY, or LPS alone. **P < 0.05 compared with LPS + PY group. Phorbol-12-myristate-13-acetate–treated nuclear extract from K-562 cells were used as a positive control. Results were from 3 independent experiments.

Oxidative Stress.

Increased oxidative stress is important in the liver injury produced by PY plus LPS.18, 19 Experiments were carried out to evaluate the role of JNK and p38 MAPK in this oxidative stress. In PY plus LPS–treated mice, malondialdehyde (MDA) levels were higher than in the saline, PY, or LPS alone groups (Fig. 4A). Both SB203580 and SP600125 lowered MDA levels. PY plus LPS treatment decreased liver GSH levels compared with controls, PY, or LPS alone (Fig. 4B). Treating the PY plus LPS mice with SB203580 or SP600125 restored liver GSH levels to control values (Fig. 4B). LPS alone also lowered GSH levels although to a lesser extent than the PY plus LPS treatment did. Protein carbonyl and nitrotyrosine protein adduct (3-NT) formation are markers of protein oxidation. PY plus LPS increased 3-NT formation 2.3-fold to 5.7-fold when compared with saline, PY, or LPS alone groups (Fig. 4C,D). The LPS alone treatment also elevated 3-NT formation but to a lesser extent than did the PY plus LPS treatment. SP600125 or SB203580 lowered the increased 3-NT levels in the PY plus LPS group (Fig. 4C,D). Highest levels of dinitrophenylhydrazine–protein carbonyl adducts were found in livers from the PY plus LPS–treated mice, and SB203580 and especially SP600125 decreased these elevated levels (Fig. 4E,F). These results suggest that the activation of JNK or p38 MAPK by PY plus LPS triggered oxidative stress in the liver.

Figure 4.

Effect of SP600125 and SB203580 on PY plus LPS induced oxidative stress. (A) Lipid peroxidation and (B) GSH levels were determined as described in Materials and Methods. **P < 0.05, compared with saline, PY, or LPS alone; *P < 0.05 compared with PY plus LPS group; experiments were repeated 4 times. (C, D). 3-Nitrotyrosine protein adducts (3-NT). *P < 0.05, compared with saline, PY, or LPS alone; **P < 0.05, compared with PY + LPS group. Experiments were repeated 3 times with triplicates in each group. (E) Protein carbonyl formation. Two of the major bands (labeled A, B) were scanned, with results shown in the bar graph (F). #, ##P < 0.05 compared with saline, PY, or LPS alone; *, **P < 0.05, compared with PY + LPS group. Experiments were repeated 3 times.

TNF-α and CYP2E1 Levels.

To study whether PY enhances production of TNF-α by LPS, serum TNF-α levels were determined. PY alone did not elevate TNF-α over the control levels. As expected, LPS alone increased TNF-α production by 50% to 100% after treatment for 3 or 24 hours. PY plus LPS increased TNF-α levels 3-fold and 4-fold after 3 and 24 hours' LPS treatment (Fig. 5A). SP600125 and SB203580 blocked the formation of TNF-α, suggesting that TNF-α production was JNK or p38 MAPK dependent (Fig. 5A).

Figure 5.

TNF-α and CYP2E1 levels. (A) Serum TNF-α levels were determined by ELISA. #, ##P < 0.05, compared with saline, PY, or LPS alone groups; *, **P < 0.05, compared with PY plus LPS group. This experiment was repeated 4 times with duplicate mice in each group. (B) CYP2E1 activity was determined by assaying microsomal oxidation of P-nitrophenol. There was no significant difference when PY + LPS, PY + LPS + SP, or PY + LPS + SB groups were compared with the PY group, P > 0.05. *P < 0.05 compared with PY alone. (C) Liver CYP2E1 protein content was determined by western blot. Bands were scanned with a computer soft program ImagJ from NIH, and the CYP2E1/β-actin ratio is shown below the blots.

PY alone or LPS plus PY induced a 2-fold increase in CYP2E1 activity compared with control or LPS-treated mice. (Fig. 5B). SP600125 or SB203580 did not affect this induced CYP2E1 activity (Fig. 5B). PY alone produced a 4-fold increase in CYP2E1 protein, whereas LPS alone had no effect (Fig. 5C). Pyrazole plus LPS treatment did not further elevate CYP2E1 protein or activity over that produced by PY alone. SP600125 or SB203580 did not affect CYP2E1 levels (Fig. 5C). These results suggest that SP600125 or SB203580 do not protect against the PY plus LPS liver injury by modifying CYP2E1 activity or content.

CYP2E1 Sensitizes the Liver to TNF-α Toxicity.

One obvious explanation for the protection against liver injury produced by PY plus LPS by SP600125 or SB203580 is that their inhibition of JNK or p38 MAPK decreases TNF-α production induced by LPS. We therefore bypassed the TNF-α production step by directly injecting TNF-α instead of LPS. TNF-α alone usually does not induce liver injury; only when administered together with adenosine triphosphate–depleting agents such as galactosamine or with protein or DNA synthesis inhibitors does TNF-α cause cell toxicity.10, 11 We studied whether the induction of CYP2E1 by PY could sensitize the liver cells to TNF-α toxicity. C57/BL/6 wild-type mice were treated for 2 days with PY followed by injection on day 3 with TNF-α. Liver injury was evaluated 24 hours after the TNF-α administration. PY alone did not elevate ALT or AST levels compared with saline control (Fig. 6A). TNF-α alone increased ALT and AST approximately 2-fold to 3-fold. PY plus TNF-α treatment increased ALT and AST approximately 10-fold to 12-fold (Fig. 6A). SP600125 or SB203580 lowered the pyrazole plus TNF-α increase in transaminases (Fig. 6A). Saline, PY, or TNF-α alone did not cause liver pathology changes. Pyrazole plus TNF-α treatment induced severe liver damage (Fig. 6B). SB203580 or SP600125 prevent this necrosis change (Fig. 6B). Lipid peroxidation was determined to evaluate whether CYP2E1 induction sensitizes the liver to TNF-α–induced oxidative stress. PY alone did not elevate cell lysate or mitochondrial MDA levels. MDA levels were elevated in the total cellular lysate and mitochondria in the TNF-α alone group, and further elevated after treatment with PY plus TNF-α (Fig. 6C). SP600125 or SB203580 lowered the elevated MDA levels in lysates and mitochondria (Fig. 6C). PY plus TNF-α lowered total as well as mitochondrial GSH levels compared with saline, PY, or TNF-α alone treatment. SP600125 or SB203580 restored total and mitochondrial GSH to the TNF-α alone values (Fig. 6D).

Figure 6.

CYP2E1 sensitizes liver to TNF-α–induced liver injury and oxidative stress: Effect of SP600125: C57/BL/6 mice were treated with TNF-α alone or TNF-α plus PY or TNF-α plus PY plus SP600125 or SB203580 for 24 hours. (A) Serum ALT and AST levels. #, ##P < 0.05, compared with saline, PY, or TNF-α alone; *, **P < 0.05, compared with TNF-α + PY groups. (B) Pathological morphology evaluation. The arrows indicate the necrosis zone in the liver from TNF-α plus PY mice. (C) Lipid peroxidation and (D) GSH were determined in liver lysates or isolated mitochondria. #, ##P < 0.05, compared with saline, PY, or TNF-α treated alone; *, **P < 0.05, compared with TNF-α + PY group. Experiments were repeated 4 times with duplicate mice in each groups.

Role of CYP2E1 in PY Plus TNF-α–Induced Liver Injury.

Pyrazole treatment alone increased P-nitrophenol (PNP) oxidation approximately 2-fold and CYP2E1 protein 2.6-fold. Similar increases were found after PY plus TNF-α treatment (data not shown). To provide evidence that induction of CYP2E1 is a key factor that sensitizes the liver to TNF-α toxicity in PY-treated mice, SV 129 CYP2E1 knockout mice and wild-type SV 129 mice were treated with PY for 2 days followed by injection with TNF-α. Liver toxicity was determined 24 hours later. Wild-type SV 129 mice treated with TNF-α alone had ALT and AST levels similar to saline control values. PY plus TNF-α treatment elevated ALT and AST levels and SP600125 lowered these levels (Fig. 7A). In SV 129 CYP2E1 knockout mice, treatment with PY plus TNF-α did not induce liver injury, because ALT and AST levels were the same as found for TNF-α alone treatment (Fig. 7A). In SV 129 wild-type mice, TNF-α plus PY induced liver necrosis (Fig. 7B) compared with the TNF-α alone–treated mice. SP600125 significantly prevented TNF-α plus PY–induced necrosis. However, in SV 129 CYP2E1−/− mice, TNF-α plus PY failed to induce liver injury (Fig. 7B). Western blot and PNP oxidation results confirmed the absence of CYP2E1 in the CYP2E1 knockout mice after the various treatments (data not shown).

Figure 7.

Effect of TNF-α plus PY in SV 129 wild-type and SV 129 CYP2E1 knockout mice. Wild-type SV 129 mice or SV 129 CYP2E1−/− mice were treated with TNF-α alone or TNF-α plus PY or TNF-α plus PY plus SP600125 for 24 hours and liver injury was evaluated. (A) Log scale of ALT and AST levels. *P < 0.01, compared with TNF-α alone; **P < 0.05, compared with TNF-α + PY group. (B) Morphology. Arrow indicates a zone of necrosis.

Activation of JNK and p38 MAPK.

TNF-α alone did not activate JNK or p38 MAPK; however, TNF-α plus PY treatment activated both JNK and p38 MAPK (Fig. 8A,B). SP600125 blocked the activation of JNK (Fig. 8C), whereas SB203580 blocked the activation of p38 MAPK (Fig. 8D). These results suggest that induction of CYP2E1 sensitizes liver cells to TNF-α toxicity and that activation of JNK and p38 MAPK play a role in the enhanced toxicity.

Figure 8.

Activation of JNK and p38 MAPK. (A, B) C57/BL/6 mice were treated with TNF-α alone or TNF-α plus PY, or (C), TNF-α plus PY plus SP600125 or (D) TNF-α plus PY plus SB203580. The activation of JNK (A) and p38 MAPK (B) or the inhibition of JNK (C) and p38 MAPK (D) were determined by western blots. Results are from 4 experiments with 2 mice in each group.

Mitochondrial Swelling.

Mitochondria from mice treated with PY plus LPS (Fig. 9A,B) or PY plus TNF-α (Fig. 9C,D) when incubated with 100 μM CaCl2underwent swelling to a greater extent than did mitochondria from saline, PY, or LPS/TNF-α alone (Fig. 9A-D). Mitochondrial swelling was decreased after treatment with SB203580 or SP600125. As a control, mitochondrial swelling induced by PY plus LPS or TNF-α plus PY was decreased by the classic inhibitor of the mitochondrial membrane permeability transition, cyclosporine A (Fig. 9A-D).

Figure 9.

Mitochondrial membrane swelling. Mitochondrial swelling was induced by adding 50 or 100 μM CaCl2. Some samples contained 2 μmol/L CsA (a mitochondrial membrane swelling inhibitor). Swelling was assayed as the decrease in absorbance at 540 nm. (B) **P < 0.05, compared with saline, PY, or LPS alone; *P < 0.05, compared with PY + LPS group. (D), #, ##P < 0.05, compared with saline, PY, or TNF-α treated alone. *, **P < 0.05 compared with TNF-α plus PY group.

Discussion

Because CYP2E1 and LPS/TNF-α are each considered risk factors for alcohol-induced liver injury, we recently developed models in which pyrazole treatment increases the toxicity produced by LPS in rats and mice.18, 19 Studies were carried out to evaluate the possible role of MAPK in this pyrazole plus LPS toxicity. MAPK are important cellular signaling molecules that convert various extracellular signals into intracellular responses.12, 22–24 In general, prolonged activation of either p38 MAPK or JNK promotes cell death with an associated decrease in mitochondrial membrane potential.10, 11, 25 We found a role for p38 MAPK in CYP2E1-dependent toxicity in HepG2 cells and hepatocytes.20, 26 Pastorino and colleagues27 reported that ethanol increased TNF-α toxicity in CYP2E1 expressing liver cells via a p38 MAPK-dependent pathway. Liu and colleagues28, 29 showed that increased CYP2E1 expression sensitized hepatocytes to TNF-α toxicity via prolonged activation of JNK.

Activation of JNK and p38 MAPK is important in the overall toxicity produced by LPS in mice with elevated levels of CYP2E1 because an inhibitor of JNK or of p38 MAPK can each blunt the liver toxicity. JNK and p38 MAPK are activated in mice treated with LPS plus pyrazole to a greater extent than mice treated with pyrazole or LPS alone. JNK activation occurred earlier (3 hours) than did the activation of p38 MAPK (24 hours), and this activation of JNK was sustained for the 24 hours of the experiment. SP600125 also prevented downstream events frequently associated with JNK activation such as translocation of activated c-jun into the nucleus and enhanced AP-1 binding activity. Whether transcription of “death genes” plays a role in the JNK mediation of toxicity in the LPS plus pyrazole model is not known. The role of c-jun activation in TNF-α–induced toxicity is not clear.29, 30 Inhibition of JNK activation prevented hepatocyte apoptosis induced by TNF-α plus actinomycin D10, 25, 30; therefore, the proapoptotic action of JNK can be independent of transcription. Similarly, whether changes in upstream factors associated with JNK activation such as MAPK kinases, MAPK kinase kinases, apoptosis signal–regulating kinase, or phosphatases that promote dephosphorylation of JNK occur in this model is not yet known. For example, oxidation of thioredoxin by CYP2E1/LPS(TNF-α)–derived reactive oxygen species (ROS) may liberate and activate apoptosis signal–regulating kinase to initiate the sequence of phosphorylations leading to JNK or p38 MAPK activation.31 Inhibition of MAPK phosphatases by CYP2E1/LPS (TNF-α)–derived ROS such as MAP kinase–MAPK phosphatase 132 may prolong the activated phosphorylated state of JNK or p38 MAPK. Studies to evaluate such possibilities are planned.

Is there a role for MAPK in the LPS plus PY elevated oxidative/nitrosative stress? The increase in MDA levels, protein carbonyl adducts, 3-nitrotyrosine protein adducts, or the decrease in GSH levels were all prevented by SP600125 and by SB203580. This suggests that the elevated oxidative/nitrosative stress is downstream of the activated MAPK. Generally, increased ROS production is associated with activation of MAPK. Recent studies have suggested a possible reciprocal relation between ROS production and MAPK activation in which ROS accumulation activates the MAPK and activated MAPK further promotes ROS production.25 For example, TNF-α–induced accumulation of ROS is decreased in cells devoid of JNK1 and JNK 2.33 Thus, the coupling of ROS and JNK signaling may be bidirectional, leading to a feedback cycle.25 In any event, inhibition of JNK and p38 MAPK lowers the oxidative/nitrosative stress, which likely plays a central role in the blunting of the liver injury by SP600125 and SB203580.

The liver injury produced in the LPS plus pyrazole model depends on the combined actions of CYP2E1 and LPS (in other words, LPS stimulation of TNF-α production) as evident by the decrease in injury in CYP2E1 knockout mice19 and in TNFR1 knockout mice (Y. Lu, unpublished results). CYP2E1was elevated to the same extent by pyrazole alone as by pyrazole plus LPS treatment. Hence, the liver injury is not attributable to just CYP2E1 alone; rather, a second “hit” is required to cause the injury. We suggest that CYP2E1 primes or sensitizes the liver to this second hit by LPS. SP600125 or SB203580 had no effect on CYP2E1 content or activity, but each inhibitor lowered the elevated levels of TNF-α. Clearly, 1 critical role for JNK and p38 MAPK in this liver injury model is to be involved in the LPS Kupffer cell interactions that ultimately stimulate TNF-α production.

Because we were interested in studying whether MAPK in the hepatocyte, not the Kupffer cell, played a role in the pyrazole plus LPS toxicity model, we attempted to bypass LPS-mediated TNF-α production by the Kupffer cell by directly adding TNF-α. This approach is similar to that of Wang et al.,34 who studied the roles of JNK1 and JNK2 in a galactoseamine/LPS as well as a galactoseamine/TNF-α liver injury model. Pyrazole potentiated the toxicity of TNF-α, and importantly, SP600125 and SB203580 both blunted the toxicity. Both JNK and p38 MAPK were activated by TNF-α plus pyrazole under conditions in which TNF-α alone (or pyrazole alone) had no effect. These experiments suggest that MAPK in the hepatocyte plays a critical role in the liver injury produced by TNF-α plus pyrazole. Additional experiments will be needed to evaluate whether an early and sustained activation of JNK or p38 MAPK occurs in this model.

Evidence for a role for CYP2E1 in the pyrazole promotion of TNF-α toxicity comes from the failure to observe liver injury when CYP2E1 knockout mice are treated with pyrazole plus TNF-α, analogous to the previous results with pyrazole plus LPS.19 Normally, hepatocytes are resistant to TNF-α–induced toxicity, suggesting that alcohol somehow sensitizes or primes the liver to become susceptible to TNF-α.35 Known factors that sensitize the liver to TNF-α are inhibitors of messenger RNA or protein synthesis, which likely prevent the synthesis of protective factors, inhibition of NF-κB activation to lower synthesis of such protective factors, depletion of GSH, especially mitochondrial GSH, lowering of S-adenosyl methionine coupled to elevation of S-adenosyl homocysteine, that is, a decline in the S-adenosyl methionine/S-adenosyl homocysteine ratio, or inhibition of the proteasome.11, 25, 29, 36–38 Our results suggest that increased oxidant stress from CYP2E1 may sensitize hepatocytes to TNF-α–induced toxicity.

Enhanced oxidative stress occurs in the livers of mice treated with TNF-α plus pyrazole, and these effects are blunted by SP600125 and SB203580. Oxidant stress also occurs within the mitochondria, as mitochondrial levels of MDA were elevated, whereas mitochondrial levels of GSH were lowered after the TNF-α plus pyrazole treatment. This increase in mitochondrial oxidant stress is also blunted by SP600125 and SB203580. Mitochondria appear to be a target for the TNF-α plus pyrazole or LPS plus pyrazole–mediated toxicity as swelling was elevated by mitochondria isolated from livers of these mice compared with mitochondria from mice treated with pyrazole alone or LPS/TNF-α alone. In preliminary experiments, we observed a decline in the mitochondrial membrane potential after TNF-α plus pyrazole treatment. These changes in mitochondrial stability or function, like the increases in mitochondrial oxidant stress, were prevented by SP600125 or by SB203580, indicating a role for JNK and p38 MAPK in the impairment of mitochondrial function, and that the latter is downstream of the activated MAPK.

A scheme that may explain some of the pathways resulting in the hepatocyte toxicity produced by pyrazole plus TNF-α is shown in Fig. 10. We hypothesize that CYP2E1-derived ROS are important in promoting sustained activation of JNK in this model. Activation of p38 MAPK also plays a role in the potentiated toxicity. Elevated ROS lower hepatic and mitochondrial GSH levels; a decrease in GSH sensitizes hepatocytes to TNF-α36, 39 via a JNK-dependent mechanism.40 Sustained activation of JNK may also increase accumulation of ROS, which further activates JNK/p38 MAPK. The ensuing mitochondrial oxidative damage ultimately impairs mitochondrial function, leading to hepatocyte loss of viability. With respect to TNF-α–induced complex II formation, we did not observe any t-Bid formation after pyrazole plus TNF-α treatment. Also, caspase 8 activity after LPS plus pyrazole treatment was the same as after LPS alone treatment (Fig. 1C). The overall injury was mainly necrosis, although whether an early increase in apoptosis switched to necrosis, especially after mitochondrial injury, cannot be ruled out. Although further studies to explore aspects of this scheme are required, this study may serve as a basis as to how 2 independent major risk factors in alcohol-induced liver injury, TNF-α and CYP2E1, interact to promote mitochondrial injury and hepatocyte toxicity. In preliminary experiments, LPS increased ALT levels in mice fed the Lieber-DeCarli ethanol liquid diet but had no effect in the dextrose-fed control; ethanol did not promote LPS toxicity in CYP2E1 knockout mice treated with LPS (Lu and Cederbaum, unpublished observations).

Figure 10.

Model for potentiation of LPS/TNF-α toxicity by induction of CYP2E1 and the central role of JNK and p38 MAPK. Please see text for discussion.

Acknowledgements

The authors thank Dr. Frank Gonzalez, NCI/NIH, for the generous provision of CYP2E1 knockout mice.

Ancillary