The process of liver transplantation necessarily involves a variable period of disruption of blood flow to the donor organ. The resulting ischemia reperfusion injury (IRI) is detectable clinically in almost a quarter of recipients and causes a spectrum of dysfunction from isolated hyperbilirubinemia to primary graft nonfunction (1). Donor risk factors for IRI include steatosis, advanced donor age, extended hospital stay before donation, acidosis before procurement, and lengthy cold ischemic time (1, 2). Organs subject to these risk factors are termed “marginal livers.” Conversely, the development of therapies capable of ameliorating IRI might recruit currently unusable organs back into the donor pool. It is estimated that the potential clinical yield of overcoming IRI could amount to as much as a 10% increase in cadaveric donation (3).
There are three known c-Jun amino terminal kinase (jnk) isoforms. Although jnk-3 is only expressed in neural tissue, jnk-1 and jnk-2 are expressed ubiquitously. Previously known as stress-activated protein kinases, jnks phosphorylate a range of transcription factor subunits including cJun that heterodimerizes to form activator protein (AP)-1 in response to cellular stress (4).The jnk enzymes also have important effects on apoptosis. Jnk-1−/− and jnk-2 −/− murine embryonic fibroblasts were protected from apoptosis when subjected to a dose of UV light, which resulted in almost universal death of wild-type and jnk-1 −/− and jnk-2 −/− cells (5).
The jnk enzymes have been identified as potential pharmacologic targets for the amelioration of hepatic IRI (6, 7). However, commonly used jnk inhibitors like SP600125 and D-JNKI-1 affect targets such as the adenosine triphosphate or c-Jun binding sites that are identical between jnk isoforms. Thus, experiments using these substances are unable to contribute to our understanding of the different physiologic roles of jnk-1 and jnk-2 (8).
Selective gene deletion of jnk-1 or jnk-2 has become the favored means by which the role of individual jnk isoforms could be investigated. Studies using jnk-2 gene deleted fibroblasts demonstrated overexpression of phospho c-Jun, whereas jnk-1 deletion led to the opposite effect on this transcription factor (4). Thus, jnk-1 and jnk-2 seem to exert dichotomous effects on the AP-1 transcription factor system (9), with jnk-2 deletion producing AP-1 upregulation.
Downstream of the JNKs and the AP-1 transcription factor system is a huge variety of gene products, hemeoxygenase-1 (HO-1) is one among them. Conversely, there are other pathways leading to the induction of HO-1, most importantly NF-E2 related factor-2 (Nrf-2) (10) in addition to AP-1. cJun overexpression has been shown to induce HO-1 expression (11), and in turn, jnk-2 deletion has been shown to upregulate cJun. Therefore, it could be hypothesized that jnk-2 deletion would lead to HO-1 induction (Fig. 1).
HO-1 catalyzes the degradation of heme into biliverdin, free iron, and carbon monoxide. In the liver, it is expressed principally by Kupffer cells (12). HO-1 induction protects cells and animals from IRI by modulation of every stage of the immune activation pathway. It confers direct cytoprotection (13), which may limit production of Danger Associated Molecular Patterns (DAMPs), mediates the protected phenotype arising from interruption of toll-like receptor signaling (14), switches macrophages in vitro from “classical” to “alternate” activation (15), modulates T-cell activation, and enhances immunologic tolerance (16).
Here, we hypothesized that the protection previously conferred by nonspecific jnk inhibitors might arise from antagonism of the jnk-2 isoform leading to AP-1 upregulation and HO-1 overexpression, resulting in protection from ischemic injury.
Jnk-2 −/− Animals
All procedures and animal care were in accordance with the U. K. Home Office licensing regulations. C57 Bl6J animals with targeted deletion of the jnk-2 gene by neomycin resistance gene insertion were supplied by Dr. R. Davis (Howard Hughes Medical Institute, Worcester, MA) and bred by homozygote to homozygote mating in accordance with local regulations. Wild-type strain specific controls were obtained from commercial sources (Harlan, UK). All animals were male, 7 to 8 weeks old.
Animals had free access to food and water. All procedures were in accordance with the U. K. Home Office regulations and conformed with the NIH guidelines for the care and use of laboratory animals. Animals were culled if they exhibited signs of “moderate severity.” Two surgical models were used. In the total hepatic ischemia model (intended to simulate the Pringle maneuver of human surgery), a microserrefine clamp was applied to the portal pedicle at the free edge of the lesser omentum, resulting in ischemia of the whole liver and upstream mesenteric congestion. Preliminary experiments identified that 30-min total hepatic ischemic insults resulted in alanine aminotransferase (ALT) rise and histologic injury, whereas 60-min insults led to considerable mortality.
To avoid mesenteric congestion, which carries a theoretical risk of bacterial translocation, experiments were repeated in a partial hepatic ischemia model. This technique involves selective occlusion of the portal pedicle supplying the left hepatic lobe (leaving the right lobe as a portosystemic shunt; Fig. 2). After ischemic insults, animals were recovered and killed 24 hr postoperatively. Detailed experimental procedures were as described previously (17). Preliminary experiments demonstrated that the optimal duration of pedicle occlusion in this model was 50 min.
Animals were subjected to 30-min (wild-type n=8, jnk-2 −/− n=8) and 60-min (wild-type n=8, jnk-2 −/− n=4) total hepatic ischemic insults. To test the importance of HO-1, animals (wild-type n=8, jnk-2 −/− n=6) were treated with chromium mesoporphyrin IX (CrMP) before 30 min of total hepatic ischemia. In the partial hepatic ischemia model, animals (wild-type n=6, jnk-2 −/− n=6) were subjected to 50-min insults. To test the importance of HO-1, two groups of animals (wild-type n=13, jnk-2 −/− n=13) were treated with CrMP 1 hr before ischemia. To test the importance of Kupffer cells, two groups of animals (wild-type n=8, jnk-2 −/− n=8) were treated with liposomal clodronate 24 hr before ischemia (Fig. 3).
Preparation of Bone Marrow-Derived Monocytes
Femurs and tibias from female, 8 to 10 week old, wild-type and jnk-2 −/− mice were harvested. Bone marrow was aspirated, and cells were grown in Teflon culture dishes for 7 days in L929 conditioned media with 10% FCS (18).
The CrMP was obtained from Frontier Scientific Inc., Logan, UT. It was chosen over alternative compounds for its relative specificity and lack of effects on the inducible nitric oxide synthetase system (19). To further ensure its specificity, dosage was minimized to 2.5 μM/kg. Stock drug was diluted in PBS such that total drug administration volume was 5 μL/g body weight. CrMP was administered intraperitoneally. Clodronate liposomes were provided by Dr. Nico van Rooijen, Amsterdam, NL. Animals received 100 μL clodronate liposomes (intravenously) 24 hr before the surgery, which resulted in effective Kupffer-cell ablation (20).
Western blotting was performed as described previously (12). JNK-1 and JNK-2 were detected using rabbit anti-total JNK #9252 Cell Signaling (Upstate UK, Hampshire, UK), and positive control was #9253 Cell signaling 293 UV-treated cell extracts. HO-1 was detected using rabbit polyclonal anti-HO-1 (Stressgen SPA-896, Nventa, San Diego, CA), and positive control was recombinant HO-1 (Stressgen). Phospho cJun (ser 73) was detected using rabbit polyclonal anti-phospho-cJun (Upstate 06-659). Loading control was Abcam mouse anti-β-Actin (ab6276, Abcam, Cambridge, UK).
HO activity was measured as previously described (21). Briefly, tissue homogenates were incubated with a mixture containing G6P, G6P dehydrogenase, hemin, liver cytosol, and NADPH for 1 hr at 37°C. At the end of the reaction period, bilirubin was solubilized in chloroform and differential absorbance at 480 and 530 nm measured on a Biomate 2 spectrophotometer. Data are presented as picomoles of bilirubin produced per gram of tissue per hour.
Tissue sections were stained for HO-1 as described previously (12). For F4/80 tissue staining, antigen retrieval was with Proteinase-K (Dako S3020), primary antibody was rat anti-mouse F4/80 (Serotec MCA497GA, Serotec, Oxford, UK), and secondary antibody was rabbit anti-rat biotinylated (Dako E0468). In all cases, the chromogen used was DAB.
Histologic Injury Scoring
Each animal was scored for necrosis and cytoplasmic injury on standardized scales by an experienced liver pathologist (C.O.B.) who was blinded to the experimental group and conditions. The two scores were added to give a single index of liver injury (maximum score 8). Degree of severity for the two parameters was determined as follows. For necrosis, 1=increased single cell death (apoptosis, necrosis) without confluent necrosis; 2=microfoci of confluent necrosis in a minority of lobules, not encircling the full circumference of the central vein; 3=confluent necrosis in a majority of lobules, or confluent necrosis around the full thickness of the central vein more than four hepatocyte layers deep; and 4=pan-lobular necrosis. For cytoplasmic injury, 1=sparse and minor nonspecific changes (minor microvesicular change); 2=confluent perivenular hepatocyte cytoplasmic injury in a small minority of lobules (<25%); 3=perivenular hepatocyte injury in 25% to 50% lobules; and 4=diffuse severe cytoplasmic injury in a majority (>50%) of lobules.
Bone marrow-derived monocytes (BMDMs) were stained for four color flowcytometry with biotinylated anti-mouse Gr1 (eBioscience RB6–8D5, eBioscience, San Diego), PE-conjugated anti-F4/80 (eBioscience clone BM8), PerCP-Cy5.5 labeled anti-CD11b (eBioscience M1/70), and FITC- conjugated anti-CD68 FITC (Acris SM1550F, Acris Antibodies GmbH, Germany) after permeabilization with Cytoperm plus (BD 554715, Becton Dickinson, New Jersey). Experiments were conducted on cells from paired, age- and sex-matched wild-type and jnk-2 −/− animals at 7 days of maturation, four animals per group. Samples were analyzed on a BD FACScalibur flowcytometer and using Flow-Jo software (Tree Star, Oregon).
Macrophage Functional Studies
Day 7 BMDMs were plated at a density of 105 cells per well in 24 well plates, 1 day before stimulus. Preliminary experiments established that the optimal dose of lipopolysacharride (LPS) was 500 ng/mL. Cells were preincubated with medium containing CrMP (2.5 μM) or vehicle for 1 hr before changing to fresh medium containing CrMP or vehicle plus LPS. Samples were collected 2 hr after LPS stimulus. ELISA for mouse tumour necrosis factor (TNF)-α was performed using a “Duoset” kit obtained from R&D systems (Minneapolis, MN) according to the manufacturer’s instructions. Experiments were repeated three to five times with cells derived from separate animals.
Cell Survival Studies
Day 7 BMDMs were plated at density of 2×104 cells per well in 96 well plates 1 day before stimulus. Experimental treatments and controls were duplicated in columns of six wells each and repeated with cells grown from four to six separate animals per group. Cells were preincubated with medium containing CrMP (2.5 μM) or vehicle for 1 hr before a 4-hr incubation in medium containing both CrMP or vehicle and ascending concentrations of hydrogen peroxide (0, 1, and 2 mM). Cells were rested overnight in plain medium before incubation with MTT (final concentration 1 mg/mL in medium). Cells were lyzed in dimethyl sulphoxide before colorimetric measurement of Formazan concentration.
Measurement of Apoptosis
Apoptosis was evaluated subjectively by light microscopy of hematoxylin-eosin stained slides. Tissue homogenates were analyzed by ELISA for histones and DNA oligonucleosomes using the “Roche cell death detection plus ELISA” kit in accordance with the manufacturer’s instructions. Caspase 3/7 activity was measured using the Promega Caspase-Glo luminetric caspase 3/7 assay (Promega Inc., Cat G8090; Promega UK Ltd., Southampton, UK) in accordance with the manufacturer’s instructions following modified preparation as described previously (22).
Datasets were prepared using Microsoft Excel. Statistical analysis was performed using Minitab 14 for Windows XP. Comparison of means of in vivo data was performed using unpaired two-tailed t tests on log-transformed data. In vitro data means were compared using unpaired two-tailed t tests (between-plate data) or paired two-tailed t tests (same-plate data). P values of 0.05 or less were taken as significant. Kaplan-Meier plots were prepared using Minitab 14 for Windows XP. Data are presented as mean±standard deviation.
Hepatocellular Survival of Ischemic Insults
Experiments in two different models of hepatic IRI showed jnk-2 −/− animals to be protected in terms of mortality, ALT, and histologic injury score.
In the total hepatic ischemia model, 75% (six of eight) animals died after 60-min insults (Fig. 4a) and mortality was reduced to 25% (one of four) in jnk-2 −/− animals (Fig. 4b). The 30-min ischemic insults were found to produce minimal mortality, but reliably elevated ALT and histologic injury scores in wild-type but not jnk-2 −/− animals, indicating that jnk-2 deletion conferred protection from injury (Fig. 4c, d). Notably, as further evidence of their protected phenotype, jnk-2 −/− animals subjected to 60-min ischemia had mean ALT release and histologic injury scores lower than the 30-min insult wild-type group. Although because of the severity of the 60-min injury, there were insufficient wild-type survivors to form a comparison group (data not shown).
In the partial hepatic ischemia model, mortality was negligible. After 50 min of ischemia, ALT rises were significantly lower in jnk-2 −/− animals (Fig. 4e), whereas there were no significant differences in histologic injury scores in this model (Fig. 4f).
Apoptosis in Wild-Type and Jnk-2 −/− Livers
A possible mechanism through which jnk-2 −/− animals may have been protected was a defect in the apoptosis pathways (5). Although we confirmed that resting caspase activity was reduced in jnk-2 −/− animals (wild-type 101±10.3 AU, jnk-2 −/− 61±5.6 AU; P=0.001), no increase in apoptosis was seen in wild-type or jnk-2 −/− animals after ischemic insults on histologic evaluation of postischemic specimens, caspase luminometric assay, or Roche “cell death ELISA.” Previously published work has shown hepatic ischemia to induce oncotic necrosis as opposed to apoptosis as a primary mode of cell death (23).
HO-1 Expression in Wild-Type and Jnk-2 −/− Livers
Western blotting was performed on liver lysates using antibodies to jnk (nonisotype specific) and HO-1. Pan-jnk blots confirmed jnk-2 to be deleted in jnk-2 −/− animals (Fig. 5a). Although signal intensity was variable in both groups, Western blotting demonstrated HO-1 induction in jnk-2 −/− animals (Fig. 5b), a finding confirmed on densitometry (P=0.045; Fig. 5c).
Hemeoxygenase activity was measured in wild-type and jnk-2 −/− liver lysates. It was 2.2× greater in jnk-2 −/− animals compared with wild-types (P=0.006; Fig. 5d). Immunohistochemistry confirmed the location of HO-1 expression to be within the Kupffer cells of both wild-type and jnk-2 −/− livers (Fig. 5e). There were no significant differences in hepatocyte or Kupffer cell morphology between resting wild-type and jnk-2 −/− mice.
The Effect of HO-1 Inhibition and Kupffer-Cell Ablation in Wild-Type and Jnk-2 −/− Mice
After 30-min ischemic insults in the total hepatic ischemia model, one of eight wild-type animals died. After HO-1 inhibition with CrMP, mortality rose sharply to seven of eight animals (Fig. 6a), indicating the importance of HO-1 in the survival of ischemic insults. Jnk-2 −/− animals were also dependent on HO-1 for survival of hepatic ischemia with mortality rising from one of eight in HO-1 intact animals to three of six in those subject to HO-1 inhibition with CrMP (Fig. 6b). Surviving jnk-2 −/− animals had much higher ALT release (Fig. 6c) and histologic injury score (Fig. 6d) compared with untreated animals subjected to the same 30-min total hepatic insult.
Before left lobar hepatic ischemia, animals were subject to HO-1 inhibition with CrMP or Kupffer-cell depletion with liposomal clodronate (LC; Fig. 6e). In accordance with the observation that HO-1 expression was focused in Kupffer cells, depletion of this cell type led to abolition of HO-1 expression in both wild-type (Fig. 6f) and jnk-2 −/− (Fig. 6g) livers even in the ischemic lobes in which HO-1 is normally potently induced.
In the left lobar hepatic ischemia model, wild-type animals subject to hemeoxygenase inhibition with CrMP demonstrated a trend toward increased susceptibility to IRI, whereas this did not reach statistical significance (P=0.158). They were rendered susceptible by Kupffer-cell ablation (P=0.002; Fig. 6h). Jnk-2 −/− animals subjected to hepatic ischemia after hemeoxygenase inhibition with CrMP became equally susceptible to IRI as HO-1 inhibited wild-type animals (P=0.0009). As in wild-type animals, Kupffer-cell ablation with liposomal clodronate resulted in increased susceptibility to injury (P=0.007; Fig. 6i).
Jnk-2 −/− Macrophage Phenotype
In view of the overexpression of HO-1 by jnk-2 −/− mice, which appeared on immunohistochemistry to be localized to the Kupffer cells, we asked whether jnk-2 −/− macrophages exhibited phenotypic differences from wild-types. Because the yield of Kupffer cells from mouse liver is poor, and the purification process is capable of increasing cellular stress and altering phenotype, it was decided to use BMDMs as a model of in vivo macrophage differentiation. First, the overexpression of HO-1 previously demonstrated in jnk-2 −/− livers was confirmed in jnk-2 −/− macrophages. This was coupled to enhanced expression of phospho c-Jun in accordance with previously published data (4, 11) (Fig. 7a). No phenotypic differences were observed between wild-type and jnk-2 −/− BMDMs on flowcytometry (Fig. 7b,c) or confocal microscopy (data not shown) for F4/80, CD11b, CD68, and Gr1.
Jnk-2 −/− BMDMs were slightly more resistant to killing by hydrogen peroxide at concentrations of 1 and 2 mM compared with wild-types (Fig. 7d), although the effect was relatively modest, and as such unlikely to account alone for the markedly protected phenotype of jnk-2 −/− mice.
After LPS stimulation, jnk-2 −/− BMDMs secreted significantly less TNF-α than wild-type macrophages (P=0.045). This effect was lost after HO-1 inhibition with CrMP (P=0.006) in jnk-2 −/− cells. There was no significant difference between CrMP-treated and -untreated wild-type cells (P=0.32; Fig. 7e).
We show here that animals deficient in jnk-2 are powerfully protected from hepatic IRI, confirming the suggestion from previous inhibitor studies that jnk may have potential as a pharmacologic target for the amelioration of IRI (6, 7, 24).
Although the effects of jnk-2 deletion are likely to be complex, one key observation may explain the protected phenotype of these animals. We found HO-1 to be overexpressed in jnk-2 −/− livers and macrophages, as would be predicted from the observed higher levels of phospho c-Jun seen in these animals (4, 11). HO-1 overexpression has been shown both to confer cytoprotection and modulate inflammation in a wide range of disorders. We used CrMP to inhibit HO-1, which in contrast with tin and zinc protoporphyrins has been shown not to have major effects on inducible nitric oxide synthetase (19). We demonstrated that CrMP-treated jnk-2 −/− mice lose their protection from hepatic IRI. Furthermore, the localization of HO-1 within Kupffer cells prompted us to examine the role of this cell type in protecting the liver from IRI. We found that macrophage ablation resulted in loss of hepatic HO-1 expression and increased susceptibility to hepatic IRI, thus, confirming the importance of Kupffer cells to hepatic HO-1 expression and survival of ischemic insults.
In macrophages, HO-1 induction has been shown to decrease proinflammatory cytokine secretion. Experiments using immortalized “RAW” macrophage cell lines demonstrated reduction in LPS-induced TNF-α and IL-1β secretion along with enhancement of antiinflammatory IL-10 secretion after transfection with HO-1 expression vectors, or exogenous administration of CO (25). Data from our own laboratory confirm a powerful effect of HO-1 on macrophage differentiation and cytokine secretion (15). Here, we have shown jnk-2 deletion to reduce TNF-α secretion as would be expected from our finding of HO-1 induction in these animals.
Interruption of TNF-α signaling has previously been shown to protect animals from hepatic IRI (26–30). The finding that jnk-2 −/− macrophages secrete less TNF-α in response to LPS stimulus provides an additional possible explanation for the protected phenotype of jnk-2 −/− mice. This finding complements recent data demonstrating the dependence of TNF-α induction of hepatocyte apoptosis on caspase-8 activation by jnk-2 (31, 32), and that pharmacologic inhibition of jnk reduced TNF-α secretion after paracetamol poisoning (33).
The HO-1 has a powerful role in immunomodulation and drives macrophages down antiinflammatory differentiation pathway (15). Because jnk-2 deletion results in HO-1 induction, it is likely that this contributes to the protection seen in jnk-2 −/− animals.
The authors thank Kathryn Sangster and Spike Clay for technical assistance; Drs. Roger Davis and Richard Flavell for the use of jnk-2 −/− mice; and Dr. Nico van Rooijens for the kind gift of clodronate liposomes.
1. Deschênes M, Belle SH, Krom RA, et al. Early allograft dysfunction after liver transplantation: A definition and predictors of outcome. National Institute of Diabetes and Digestive and Kidney Diseases Liver Transplantation Database. Transplantation
1998; 66: 302.
2. Ploeg RJ, D’Alessandro AM, Knechtle SJ, et al. Risk factors for primary dysfunction after liver transplantation—A multivariate analysis. Transplantation
1993; 55: 807.
3. Devey LR, Friend PJ, Forsythe JL, et al. The use of marginal heart beating donor livers for transplantation in the United Kingdom. Transplantation
2007; 84: 70.
4. Sabapathy K, Hochedlinger K, Nam SY, et al. Distinct roles for JNK1 and JNK2 in regulating JNK activity and c-Jun-dependent cell proliferation. Mol Cell
2004; 15: 713.
5. Tournier C, Hess P, Yang DD, et al. Requirement of JNK for stress-induced activation of the cytochrome c-mediated death pathway. Science
2000; 288: 870.
6. Uehara T, Bennett B, Sakata ST, et al. JNK mediates hepatic ischemia reperfusion injury. J Hepatol
2005; 42: 850.
7. Uehara T, Xi Peng X, Bennett B, et al. c-Jun N-terminal kinase mediates hepatic injury after rat liver transplantation. Transplantation
2004; 78: 324.
8. Bogoyevitch MA, Boehm I, Oakley A, et al. Targeting the JNK MAPK cascade for inhibition: Basic science and therapeutic potential. Biochim Biophys Acta
2004; 1697: 89.
9. Conze D, Krahl T, Kennedy N, et al. c-Jun NH(2)-terminal kinase (JNK)1 and JNK2 have distinct roles in CD8(+) T cell activation. J Exp Med
2002; 195: 811.
10. Alam J, Stewart D, Touchard C, et al. Nrf2, a Cap‘n’Collar transcription factor, regulates induction of the heme oxygenase-1 gene. J Biol Chem
1999; 274: 26071.
11. Hock TD, Liby K, Wright MM, et al. JunB and JunD regulate human heme oxygenase-1 gene expression in renal epithelial cells. J Biol Chem
2007; 282: 6875.
12. Patel A, van de Poll MC, Greve JW, et al. Early stress protein gene expression in a human model of ischemic preconditioning. Transplantation
2004; 78: 1479.
13. Vile GF, Basu-Modak S, Waltner C, et al. Heme oxygenase 1 mediates an adaptive response to oxidative stress in human skin fibroblasts. Proc Natl Acad Sci USA
1994; 91: 2607.
14. Shen X-D, Ke B, Zhai Y, et al. Absence of toll-like receptor 4 (TLR4) signaling in the donor organ reduces ischemia and reperfusion injury in a murine liver transplantation model. Liver Transpl
2007; 13: 1435.
15. Devey L, Ferenbach D, Mohr E, et al. Tissue-resident macrophages protect the liver from ischemia reperfusion injury via a heme oxygenase-1- dependent mechanism. Mol Ther
2009; 17: 65.
16. Xia ZW, Xu LQ, Zhong WW, et al. Heme oxygenase-1 attenuates ovalbumin-induced airway inflammation by up-regulation of Foxp3 T-regulatory cells, interleukin-10, and membrane-bound transforming growth factor-β1. Am J Pathol
2007; 171: 1904.
17. Devey L, Festing M, Wigmore S. Effect of temperature control upon a mouse model of partial hepatic ischaemia/reperfusion injury. Lab Anim
2008; 42: 12.
18. Duffield JS, Forbes SJ, Constandinou CM, et al. Selective depletion of macrophages reveals distinct, opposing roles during liver injury and repair. J Clin Invest
2005; 115: 56.
19. Appleton SD, Chretien ML, McLaughlin BE, et al. Selective inhibition of heme oxygenase, without inhibition of nitric oxide synthase or soluble guanylyl cyclase, by metalloporphyrins at low concentrations. Drug Metab Dispos
1999; 27: 1214.
20. Van Rooijen N, Sanders A. Kupffer cell depletion by liposome- delivered drugs: Comparative activity of intracellular clodronate, propamidine, and ethylenediaminetetraacetic acid. Hepatology
1996; 23: 1239.
21. McNally SJ, Ross JA, James GO, et al. Optimization of the paired enzyme assay for heme oxygenase activity. Anal Biochem
2004; 332: 398.
22. Liu D, Li C, Chen Y, et al. Nuclear import of proinflammatory transcription factors is required for massive liver apoptosis induced by bacterial lipopolysaccharide. J Biol Chem
2004; 279: 48434.
23. Gujral JS, Bucci TJ, Farhood A, et al. Mechanism of cell death during warm hepatic ischemia-reperfusion in rats: Apoptosis or necrosis? Hepatology
2001; 33: 397.
24. Ishii M, Suzuki Y, Takeshita K, et al. Inhibition of c-Jun NH2-terminal kinase activity improves ischemia/reperfusion injury in rat lungs. J Immunol
2004; 172: 2569.
25. Otterbein LE, Bach FH, Alam J, et al. Carbon monoxide has anti- inflammatory effects involving the mitogen-activated protein kinase pathway. Nat Med
2000; 6: 422.
26. Teoh N, Leclercq I, Pena AD, et al. Low-dose TNF-alpha protects against hepatic ischemia-reperfusion injury in mice: Implications for preconditioning. Hepatology
2003; 37: 118.
27. Teoh N, Field J, Sutton J, et al. Dual role of tumor necrosis factor-alpha in hepatic ischemia-reperfusion injury: Studies in tumor necrosis factor- alpha gene knockout mice. Hepatology
2004; 39: 412.
28. Bachmann S, Caldwell-Kenkel JC, Currin RT, et al. Protection by pentoxifylline against graft failure from storage injury after orthotopic rat liver transplantation with arterialization. Transpl Int
1992; 5(suppl 1): S345.
29. El-Ghoneimi A, Cursio R, Schmid-Alliana A, et al. Inhibition of tumor necrosis factor alpha gene transcription by pentoxifylline reduces normothermic liver ischemia-reperfusion injury in rats. Transplant Proc
2007; 39: 1761.
30. Tian Y, Jochum W, Georgiev P, et al. Kupffer cell-dependent TNF-alpha signaling mediates injury in the arterialized small-for-size liver transplantation in the mouse. Proc Natl Acad Sci USA
2006; 103: 4598.
31. Liedtke C, Trautwein C. The role of JNK2 in toxic liver injury. J Hepatol
2006; 45: 762.
32. Wang Y, Singh R, Lefkowitch JH, et al. Tumor necrosis factor-induced toxic liver injury results from JNK2-dependent activation of caspase-8 and the mitochondrial death pathway. J Biol Chem
2006; 281: 15258.
33. Henderson NC, Pollock K, Frew J, et al. Critical role of c-jun (NH2) terminal kinase in paracetamol-induced acute liver failure. Gut
2006; 56: 982.