Chronic liver disease in murine hereditary tyrosinemia type 1 induces resistance to cell death

Authors


Abstract

The murine model of hereditary tyrosinemia type 1 (HT1) was used to analyze the relationship between chronic liver disease and programmed cell death in vivo. In healthy fumarylacetoacetate hydrolase deficient mice (Fah-/-), protected from liver injury by the drug 2-(2- nitro-4-trifluoromethylbenzoyl)-1,3-cyclohexanedione (NTBC), the tyrosine metabolite homogentisic acid (HGA) caused rapid hepatocyte death. In contrast, all mice survived the same otherwise lethal dose of HGA if they had preexisting liver damage induced by NTBC withdrawal. Similarly, Fah-/- animals with liver injury were also resistant to apoptosis induced by the Fas ligand Jo-2 and to necrosis-like cell death induced by acetaminophen (APAP). Molecular studies revealed a marked up-regulation of the antiapoptotic heat shock proteins (Hsp) 27, 32, and 70 and of c-Jun in hepatocytes of stressed mice. In addition, the p38 and Jun N-terminal kinase (JNK) stress-activated kinase pathways were markedly impaired in the cell-death resistant liver. In conclusion, these results provide evidence that chronic liver disease can paradoxically result in cell death resistance in vivo. Stress-induced failure of cell death programs may lead to an accumulation of damaged cells and therefore enhance the risk for cancer as observed in HT1 and other chronic liver diseases. (HEPATOLOGY 2004;39:433–443.)

Hereditary tyrosinemia type 1 (HT1) is an autosomal recessive disease caused by a deficiency of fumarylacetoacetate hydrolase (FAH), the enzyme that carries out the last step of the tyrosine catabolic pathway (Fig. 1).1 HT type 2 (HT2) and HT type 3 (HT3) are caused by deficiency of tyrosine aminotransferase and 4-hydroxyphenylpyruvate dioxygenase (HPD), respectively, enzymes more proximal in the pathway. In HT2 and HT3, plasma tyrosine levels are highly elevated, but there is no liver disease, thereby showing that tyrosine itself is not a hepatotoxin.2 In contrast, hepatic damage is prominent in HT1. There are 2 major clinical forms of the disease, acute and chronic, which can be distinguished on the basis of the severity and rate of progression. In the acute form, symptoms appear during the first months of life, and death caused by hepatic failure usually occurs during the first year.2 The chronic form results in more gradual liver disease, and hepatocarcinoma (HCC) is a frequent complication. By age 5, about one third of patients develop HCC3; thus, HT1 has the highest risk for primary liver cancer of any human disease. Furthermore, fumarylacetoacetate (FAA), the primary metabolite that accumulates in HT1, has been shown to be highly mutagenic.4, 5 FAA is a potent alkylator, causing oxidative damage by reacting with glutathione and sulfhydryl groups of proteins.6

Figure 1.

The pathway for tyrosine catabolism is shown schematically. The enzymes of the pathway are (1) tyrosine aminotransferase;(2) 4-hydroxyphenylpyruvate dioxygenase; (3) homogentisic acid oxidase; (4) maleylacetoacetate isomerase; and (5) fumarylacetoacetate hydrolase, the enzyme deficient in hereditary tyrosinemia type I. NTBC is the pharmacological inhibitor of HPD used in this study. The hereditary tyrosinemia type 3 model mice with genetic deficiency of HPD (Hpd-/-) have hypertyrosinemia without liver injury.

Treatment with NTBC has become the mainstay for the management of HT1. The efficacy of this therapy in improving the liver and renal disease associated with FAH deficiency has been well documented,7 and the drug was recently FDA approved. NTBC blocks the activity of 4-hydroxyphenylpyruvate dioxygenase (HPD), an enzyme acting upstream of FAH in the tyrosine catabolic pathway, and therefore reduces the accumulation of FAA.8 However, recent preclinical animal data indicate that NTBC provides only incomplete protection from liver disease, in particular HCC.9

Despite the known mutagenicity of FAA,4, 5 the details of cancer evolution in HT1 have remained unclear. There are several general ways to think about cancer predisposition in the face of ongoing cellular damage. First, an increased tumor risk may derive from chronic injury and increased mitotic rate, resulting in a higher chance for mutation. Second, malignant transformation can result from the lack of death in genomically damaged cells.10 With regard to HT1, previous studies have shown that homogentisic acid (HGA) acutely induces apoptosis in hepatocytes of mice that are doubly mutant in Hpd and Fah.11, 12 HGA is a metabolite distal to the HPD enzyme and therefore can be converted efficiently to the toxic FAA (Fig. 1). This finding of HGA-induced apoptosis suggested the hypothesis that a high rate of cell turnover triggered by increased cell death underlies the increased tumor risk in HT1. However, markers of hepatocyte death, such as serum transaminases, are not highly elevated in human or murine HT1,2, 13 and the failure of synthetic liver functions (clotting factor deficiency, hyperbilirubinemia) is much more pronounced. In addition, it has been shown that acute cell stress or injuries, including those from oxidants, can “precondition” hepatocytes and induce resistance against subsequent cell death.14 Therefore, the aim of this study was to determine whether increased cell death and cell turnover are indeed present in the cancer-prone HT1 liver during NTBC withdrawal. Alternatively, we wished to consider the hypothesis that hepatocytes in HT1 may display increased cell death resistance. We show that chronic liver injury in HT1, associated with oxidative stress, results in resistance against apoptosis and necrosis-like cell death in vivo.

Abbreviations

HT1, hereditary tyrosinemia type 1; HT2, hereditary tyrosinemia type 2; HT3, hereditary tyrosinemia type 3; FaH, fumarylacetoacetate hydrolase; HCC, hepatocellular carcinoma; FAA, fumarylacetoacetate; NTBC, 2-(2-nitro-4-trifluoromethylbenzoyl)-1,3-cyclohexanedione; HPD, 4-hydroxyphenylpyruvate dioxygenase; APAP, acetaminophen; HGA, homogentisic acid; IAP, inhibitor of apoptosis; HPLC-ECD, high performance liquid chromatography with electrochemical detection; dGua, deoxyguanosine; Ac-DEVD-AMC, Acetyl-Asp-Glu-Val-Asp-7-amino-4-methylcoumarin; BrDU, Bromodeoxyuridine; APAF-1, apoptosis protease activating factor-1.

Materials and Methods

Mice.

We used the Fah-/- strain mice previously described.15 All Fah mutant animals were treated with NTBC-containing drinking water at a concentration of 7.5 mg/L unless otherwise indicated (a gift from S. Lindstedt, Gøtheborg, Sweden). Animal care and experiments were all in accordance with the guidelines of the department of animal care at Oregon Health and Science University. For the ethanol feeding experiments, mice were divided into 2 groups: one group was fed an ethanol-containing liquid diet (1.0 kcal/mL of which 35% was fat-derived, 11% was derived from carbohydrate, 18% was derived from protein, and 36% was derived from ethanol) for 2 weeks; the other was pair-fed a control diet in which ethanol was substituted isocalorically with dextrin maltose (Dyets, Bethlehem, PA).

At the end of each pretreatment, the animals were injected intraperitoneally (i.p.) with HGA (Sigma Chemical Co., St. Louis, MO; 500 mg/kg), Fas antibody (Jo-2; Pharmingen, San Diego, CA; 0.3 μg/g), or acetaminophen (APAP; Sigma Chemical Co.; 800 mg/kg) and sacrificed at indicated time points. At the time of sacrifice, a small fragment of each liver was fixed in buffered formalin. The remaining liver tissues were flash-frozen in liquid nitrogen and stored at −80°C until analysis.

Histology and Terminal Deoxynucleotid Transferase-Mediated dUDP Nick-End Labeling (TUNEL) Assay.

Liver issues were fixed in 10% phosphate-buffered formalin (pH 7.4), dehydrated in 100% ethanol, and embedded in paraffin wax at 58°C. Five-μm sections were rehydrated and stained with hematoxyli-eosin. The TUNEL assay (ApopTag, Serological Corporation, Norcross, GA) was performed according to the manufacturer's recommendations.

DNA Fragmentation Analysis.

Genomic DNA was isolated and purified from mouse livers using a genomic DNA isolation kit (Qiagen, Valencia, CA). DNA samples (1μg each) were electrophoretically separated on 2% agarose gel containing ethidium bromide (0.5 g/L).

Liver Protein Isolation.

Frozen tissue was homogenized in phosphate-buffered saline containing 1% wt/vol Nonidet P-40, 0.5% wt/vol sodium deoxycholate, 0.15% wt/vol sodium dodecyl sulfate 10%, and Complete Protease Inhibitor tablets (Roche, Mannheim, Germany), then sonicated for 20 seconds on ice and centrifuged at 14,000g for 10 minutes. Protein concentration was determined with the BioRad Protein assay kit as recommended by the manufacturer (BioRad, Hercules, CA).

Caspase Activity.

Liver lysates were prepared by homogenization in hypotonic buffer (25 mmol HEPES, pH 7.5; 5 mM MgCl2; 1 mM EGTA; 1 mM phenylmethylsulfonyl fluoride; and 1 mg/ml leupeptin and aprotinin). Homogenates were centrifuged at 15,000g for 15 minutes, and extracted proteins (50 μg) were tested in triplicate experiments by measuring the proteolytic cleavage of specific fluorogenic substrate of Acetyl-Asp-Glu-Val-Asp-7-amino-4-methylcoumarin (Ac-DEVD-AMC) for caspase-3 and of Acetyl-Leu-Glu-His-Asp-7-amino-4-trifluoromethylcoumarin (Ac-LEHD-AFC) for caspase-9 (CaspACE Assay System; Promega, Madison, WI) and of Ac-LEHD-AFC for caspase-9.

Antibodies.

Antibodies against caspase-9, phospho-Jun N-terminal kinase (JNK), phospho-p38 mitogen-activated protein kinase (p38), phospho-stress-activated protein kinase/extracellular signal-regulated kinase kinase (SEK), phospho-mitogen-activated protein kinase kinase MKK3/6, c-Jun and BH3 interacting domain death agonist (Bid) were obtained from Cell Signaling Technology (Beverly, MA); Antibodies against heat shock protein (Hsp)27, HSp70, inhibitor of apoptosis protein 2 (cIAP2), Flice inhibitory protein (FLIP) and nuclear factor kappaB inhibitors alpha (IkBa) and beta (IkBb) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA), X-chromosome-linked inhibitor of apoptosis protein (XIAP) from PharMingen (San Diego, CA) and Hsp32 from Stressgen (Victoria, Canada).

Western Blot Analysis.

Protein extracts were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred onto polyvinylidene difluoride membrane (Millipore, Bedford, MA). Coomassie staining was used to demonstrate equal protein loading. After the transfer, the nonspecific antibody-binding activities of the membranes were blocked in Tris-buffered saline (10 mM Tris-HCl, pH 8.0; 150 mM NaCl) containing 0.3% (vol/vol) Tween 20 and 10% (wt/vol) nonfat dry milk for 30 minutes at room temperature. Incubations with appropriate dilutions of primary and secondary antibodies were done overnight to specific antisera at 4°C and for 45 minutes in the same buffer, followed by extensive washes with Tris-buffered saline containing 0.3% Tween 20. Detection of immunolabeled proteins was done using a chemiluminescence kit (BioRad) and Hyperfilm-enhanced chemiluminescence film (Amersham Biosciences, Piscataway, NJ). The films were scanned; signal intensity was quantified using the Multi-Analyst software program (BioRad, Hercules, CA).

8-Hydroxydeoxyguanosine (8OHdGua) Analysis.

Isolation and purification of DNA was performed as described above. Genomic DNA samples were then treated with nuclease P1 and alkaline phosphatase and analyzed for the presence of 8OHdGua by high performance liquid chromatography with electrochemical detection (HPLC-ECD). 8OHdGua and deoxyguanosine (dGua) were separated on a Waters 510 HPLC and quantified by using a Bioanalytical Systems LC-4C electrochemical detector (West Lafayette, IN) and a Beckman 160 UV detector (Beckman Coulter, Fullerton, CA), respectively. Levels of 8OHdGua/106 dGua were calculated by converting the peak areas 8OHdGua and dGua to amounts based on standard curves of the authentic standards 8OHdGua (Cayman Chemicals, Ann Arbor, MI) and dGua (Sigma Chemical Co.).

Hepatic GSH Determination.

Liver tissues were perfused with 0.9% NaCl containing 0.16 mg/ml heparin as anticoagulant. A portion of liver was homogenized in ice-cold 5% metaphosphoric acid at a 1:20 ratio (wt/vol), followed by centrifugation at 3,000g for 10 minutes. GSH in the supernatant was assayed using a colorimetric kit from Calbiochem (San Diego, CA).

Results

Fah Mutant Mice Off NTBC Are Protected From HGA-Induced Acute Liver Failure.

While Fah-deficient mice display liver injury,15 mice doubly mutant in Hpd and Fah are protected from liver disease because HPD deficiency prevents the generation of FAA, the hepatotoxic metabolite12 (Fig. 1). However, acute administration of HGA, a tyrosine metabolite downstream of the HPD enzyme, can evoke a massive apoptotic response in hepatocytes of these animals.12, 16 FAA itself can also induce apoptosis of cultured cells.11 To confirm these findings, HGA was injected i.p. into 10 Fah-deficient mice kept healthy by NTBC, which blocks HPD pharmacologically.8 As expected, all mice became ill and died within approximately 70 hours (Fig. 2A). Liver sections examined at various time points after treatment showed a massive increase of TUNEL-positive cells (Fig. 3B). However, in contrast to the expected distinct nuclear TUNEL staining, we observed diffuse staining in the cytosol and the nucleus consistent with extensive DNA fragmentation and diffusion of these fragments into the cytosol. This pattern is indicative of necrosis-like cell death.17 To further characterize the HGA-induced cell death, transaminase levels, DNA laddering, cleavage of caspase-9, and activities of caspase-9 and caspase-3 were measured. Serum transaminases were significantly increased after HGA administration (Fig. 4C). DNA laddering was visible after 6 hours (Fig. 4A), and cleavage of the 49 kd precursor form of caspase-9 was evident 3 hours after treatment with HGA (Fig. 4B). However, both caspase-9 and caspase-3 enzyme assays revealed only moderate activity increases (Fig. 4C). Together, these data indicate that HGA induces not only classical caspase-dependent apoptosis in Fah-/- mice but also a necrosis-like form of cell death; this is in agreement with previous findings showing that caspase inhibitors do not completely protect Fah/Hpd double-mutant mice from HGA-induced death.12

Figure 2.

Survival effects of NTBC withdrawal and chronic ethanol feeding in Fah-/- mice. (A) Fah-/- mice on NTBC (▪) died because of acute liver failure within 70 hours following a challenge with HGA (500 mg/kg; n = 10); all Fah-/- mice off NTBC (▴), and 6 of 10 mice after 2 weeks of ethanol feeding (▾), survived the same dose of HGA (n = 10 in each group). (B) Fah-/- mice on NTBC (▪) died rapidly following a high dose of the FAS agonist Jo-2 (0.3 μg/g) (n = 10); Fah-/- mice off NTBC (▴) showed a significant survival advantage (n = 10). (C) A similar survival advantage was seen in mice off NTBC challenged with acetaminophen (800 mg/kg).

Figure 3.

Histological analysis of hepatic apoptosis in Fah deficient mice. TUNEL staining (brown color/ arrows) of liver sections is shown. Original magnification ×200. Insets have a ×400 magnification. (A and B) Healthy Fah-/- mice on NTBC (A) before and (B) 6 hours after challenge with HGA. (C and D) Fah-/- mice after 2 weeks off NTBC, (C) before and (D) 6 hours after challenge with HGA. (E and G) Fah-/- mice on NTBC and (F and H) Fah-/- mice off NTBC following challenge with Jo-2 (E and F) and acetaminophen (G and H). (I and J) NTBC-treated mutants after 2 weeks of ethanol feeding, (I) before and (J) 6 hours after HGA administration. (B, E, and G) TUNEL-positive cells are clearly visible after 6 hours in healthy, NTBC treated mice. Stress induced by NTBC withdrawal (D, F, and H) or ethanol diet (J) prevented cell death.

Figure 4.

(A) DNA fragmentation in mice on and off NTBC following challenge with HGA, APAP, and Jo-2. (B) Detection of uncleaved/inactive and cleaved/active caspase-9 before, 3 hours after, and 6 hours after induction of apoptosis with HGA and Jo-2 in Fah-/- mice on and off NTBC. (C) Caspase-3 and caspase-9 activity and ALT levels in mice challenged with HGA, APAP, and Jo-2 (filled columns: mice on NTBC; open columns: mice off NTBC). (D) Increased levels of Fas receptor and Bid in mice off NTBC.

Next we wished to determine how preexisting liver disease would affect cell death pathways in HT1 mice. The presence of marked oxidative stress in FAH deficiency has been previously demonstrated by up-regulation of known response genes such as glutamate cysteine ligase, NAD(P)H:quinone oxidoreductase, and Hsp32/heme oxigenase-1 (Fig. 5C and D).8, 18–20 Importantly, these markers of oxidative stress are present in mice 2 weeks off NTBC, the time point chosen for our studies. Furthermore, direct measurement in 2 Fah-/- mice of 8-oxoguanine in hepatocyte DNA after 4 weeks off NTBC showed levels of 80 and 109 8-OHdGua residues/106 dGua, respectively; wild-type controls (n = 4) had only 2.65 ± 1.2 8-OHdGua residues/106 dGua, indicating an approximately 35-fold increase in oxidatively damaged bases.

Figure 5.

(A) Unchanged levels of IκBα and -β in mice on and off NTBC. (B) Expression of XIAP, clAP2 and FLIP in Fah-/- mice on NTBC and off NTBC. (C) Expression of Hsp27, 32, and 70 proteins before and 3 and 6 hours after injection of HGA in Fah-/- mice on and off NTBC, and (D) during NTBC withdrawal. (E) Phosphorylation of SEK and JNK and (F) MKK 3/6 and p38 in mice on and off NTBC following challenge with HGA, APAP, and Jo-2, and during NTBC withdrawal. (G) Expression of c-Jun in mice on and off NTBC and (H) during NTBC withdrawal.

Therefore, in order to determine whether HGA also causes rapid hepatocyte death in mice with preexisting liver disease, the drug was injected i.p. into 10 Fah-/- mice that had been taken off NTBC for 14 days. This time point was chosen because mice off NTBC still had the same weight and overall health as the controls did, despite hepatic dysfunction. We have previously shown that Fah-/- mice have clearly abnormal liver functions at this time point, including marked hyperbilirubinemia, elevation of transaminases, and histological abnormalities.8 Unexpectedly, all 10 animals survived the otherwise 100% lethal dose of 500 mg/kg HGA (Fig. 2A). Histology, TUNEL staining (Fig. 3D), and DNA laddering (Fig. 4A) confirmed a significant reduction of hepatocyte death compared to Fah-/- mice without liver disease at all time points. Additionally, transaminase levels were only slightly elevated, and caspase-9 and caspase-3 activities were not detectable (Fig. 4C).

Fah Mutant Mice Off NTBC Are Protected From Cell Death Induced by Fas or APAP.

To determine whether this protection from cell death was specific to HGA or represented a general phenomenon, two additional triggers of cell death were tested: Fas ligand, an inducer of classical caspase-dependent apoptosis, and the drug APAP, which induces caspase-independent cell death.21 First, Fah-/- mice either on or off NTBC (n = 10 mice in each group) were injected with the Fas agonist Jo-2. Eighty percent of the mice on NTBC died rapidly after a single 0.3 μg/g intraperitoneal dose of Jo-2 (Fig. 2B). In contrast, the animals off NTBC showed a statistically significant survival advantage (P = .002) (Fig. 2B). TUNEL staining showed massive apoptotic injury in NTBC-treated mice (Fig. 3E), which was confirmed by DNA laddering (Fig. 4A), cleavage of caspase-9 (Fig. 4B), and activation of caspase-9 and caspase-3 (Fig. 4C). All markers of caspase-dependent apoptosis and transaminase levels were significantly lower in animals off NTBC (Figs. 3F, 4A, 4B, and 4C). Importantly, this protective effect was seen despite elevation of the protein levels of Fas receptor (3-fold elevated) and Bid (2-fold elevated), which connects the death receptor with the mitochondrial death pathway (Fig. 4D).

Next, we determined whether mice off NTBC were also protected from liver failure induced by APAP. Mice on and off NTBC (n = 10 mice in each group) were challenged with APAP (800mg/kg i.p.) to induce acute liver failure; again, a significant survival advantage in mice off NTBC was found. In agreement with previous publications, diffuse TUNEL staining (Fig. 3G) and DNA laddering (Fig. 4A), but no activation of caspase-9 and caspase-3 (Fig. 4C) were seen after APAP administration in Fah-/- mice on NTBC.21 Transaminase levels, DNA laddering, and TUNEL staining were significantly reduced in mice off NTBC (Figs. 3H, 4A, and 4C). Together, our data showed that NTBC withdrawal in Fah-/- mice caused resistance to multiple inducers of hepatocyte death.

Ethanol Induced Resistance to HGA-Induced Cell Death Similar to That in Fah Mutant Mice Off NTBC.

We hypothesized that chronic oxidative stress may contribute to the cell death resistance phenotype. To further strengthen the correlation between oxidative stress and the cell death block, Fah-/- mice were kept healthy on NTBC, but were given an ethanol diet as a known alternate cause of chronic oxidative stress.22 Ethanol-treated Fah mutants and normal diet controls were then challenged with the normally lethal dose of HGA (n = 10 in each group). In agreement with our hypothesis, ethanol-treated mice survived significantly better (P = .003) than the controls, which were on a control diet and NTBC (Fig. 2A). Additionally, TUNEL-positive cells were barely detectable before or following challenge with HGA (Fig. 3I and J).

NTBC Withdrawal Does Not Change Overall Glutathione Levels or Induce Nuclear Factor-kappaB (NF-κB) Activation.

We next wanted to ascertain which factors contributed to the observed anti-death adaptation. Decreases in cellular glutathione levels have previously been shown to prevent the execution of programmed cell death, but it has also been reported that glutathione depletion enhances the HGA toxicity in Fah-/- mice.23 Importantly, glutathione levels were not reduced in HT1 mice 2 weeks off NTBC: levels on NTBC were 5.1 + 1.9 mmol/μg liver protein (n = 4); levels off NTBC were 6.9 + 1.54 mmol/μg liver protein (n=4). This indicated that decreases in glutathione levels were not the mechanism for cell death resistance. However, the previously reported up-regulation of genes, such as glutamylcysteine synthetase, involved in the synthesis of glutathione20 could potentially result in a decreased rate of glutathione depletion and therefore contribute to prevention of HGA- and APAP-induced acute liver failure.

We then examined whether NF-κB, a ubiquitous transcription factor implicated in both hepatocyte proliferation and survival, might be involved in this adaptive phenotype.24 NF-κB dimers are sequestered in the cytoplasm by binding to inhibitory IκB proteins. NF-κB can be rapidly activated by signals that induce the sequential phosphorylation of IκBs, at serine residues 32 and 36, and their subsequent proteolysis. This in turn unmasks a nuclear target sequence within NF-κB and leads to its translocation to the nucleus.25 Inducers of NF-κB include interleukin-1 (IL-1), tumor necrosis factor, lipopolysaccharide, phorbol myristate acetate, and hypoxia/reoxygenation.26 It has been demonstrated that NF-κB can be activated by at least 3 different pathways, which lead either to degradation of IκB-α and IκB-β or to degradation of IκB-α alone. The third alternative mechanism is independent of IκB degradation and depends on tyrosine phosphorylation of IκB and its subsequent dissociation from NF-κB. Ischemia/reperfusion injury leads to NF-κB activation by this pathway.27 To determine whether NF-κB was activated in stressed hepatocytes, the degradation of IκBs was measured, either after challenge with HGA or during NTBC withdrawal (Fig. 5A). However, no IκB degradation was observed, and no differences were seen between mice on or off NTBC. Furthermore, we did not observe an increase of IκB-α phosphorylation in any setting (data not shown), and no up-regulation of NF-κB target genes, such as the inhibitors of apoptosis (IAP) XIAP and cIAP2, was found (Fig. 5B). The IAPs belong to a group of structurally related proteins that function as direct inhibitors of apoptosis.28 Although the levels of FLIP (formerly described as antiapoptotic) were up-regulated 2-fold, it seems unlikely that these changes could have contributed to the observed apoptosis resistance, considering recent findings showing that FLIP is enriched in the CD95 death-inducing signaling complex and also potently promotes procaspase-8 activation.29 Overall, these data indicate that FAH deficiency does not lead to significant NF-κB activation.

Another important negative result is the absence of increased hepatocyte proliferation in Fah-/- mice off NTBC. Previous studies have shown that the regenerating liver is highly resistant to Fas-induced apoptosis.30, 31 To determine whether increased proliferation could account for the increased cell death resistance, 2 mutants on NTBC and 2 mutants off NTBC were given 2 doses of Bromodeoxyuridine (BrDU) i.p. either 2 and 14 hours before harvest or continuously for 7 days before sacrifice. Liver samples of mice injected with BrDU injection at 24 and 36 hours after partial (two thirds) hepatectomy were used as positive controls. The BrDU labeling index in mice on and off NTBC was <1/1,000, thereby indicating a very low rate of cell division in both groups regardless of liver injury (data not shown). These results are consistent with previous studies from our laboratory showing much higher rates of cell division in Fah+/+ donor cells compared to Fah-/- hepatocytes during liver repopulation.32

NTBC Withdrawal Induces an Up-regulation of Hsps.

Other attractive candidates for mediating the PCD resistance are the Hsps. Hsps behave as molecular chaperones for other cellular proteins and are known to have a pivotal role in protection from various stresses. In mice the Hsp70 family consists of at least 7 members, and environmental stresses induce the expression of two highly related isoforms: Hsp70.1 and Hsp70.3.33 In Fah-/- mice, none of the Hsps were expressed either before or after challenge with HGA. However, after NTBC withdrawal, a 5-fold up-regulation of Hsp70 protein levels was seen (Fig. 5C). Additionally, strong increases of Hsp27 (200-fold) and Hsp32 (58-fold) were found (Fig. 5C).

Activation of Stress-Activated Protein Kinases Is Associated With Cell Death in Fah-/- Mice; Up-regulation of c-Jun Is Associated With Cell Survival.

Members of the MAPKK group of protein kinases activate JNK and p38 MAPK through phosphorylation.34 We observed a strong activation of both stress kinase pathways in healthy Fah-/- mice on NTBC following challenge with HGA, APAP, or Jo-2 (Fig. 5E and F). In contrast, apoptosis resistant mice off NTBC showed a markedly blunted response with reduced phosphorylation of p38 and JNK (Fig. 5E and F). This effect was also seen in ethanol-pretreated mice (data not shown).

Recent studies have identified c-Jun as an important regulator of hepatocyte proliferation and survival during liver development and regeneration.35 In addition, it has been shown that c-Jun-deficient livers are less vulnerable to induced HCC and have markedly increased rates of apoptosis.36 In accordance with this observation, a significantly higher expression of c-Jun was found in Fah-deficient mice off NTBC; these mice were resistant to cell death and survived the different death triggers (Fig. 5G). The up-regulation of antiapoptotic c-Jun was most prominent during NTBC withdrawal after 4 days (Fig. 5H).

Discussion

In this study, we provide evidence that the liver disease HT1, associated with chronic oxidative stress, results in resistance to apoptosis and apoptosis-like cell death in vivo. Unexpectedly, increased cell turnover caused by hepatocyte death is not a prominent finding in this disease.

Cell Death Resistance in HT1 Liver Disease.

In the neonatal period, Fah-deficient mice die from liver dysfunction unless they are treated with NTBC.8 Discontinuation of NTBC in adult FAH mutant mice provides an accurate model of acute HT1 in humans. The animals develop liver dysfunction within 10 days; this progresses to hepatic failure and death within 6 to 8 weeks.8 Alternatively, suboptimal NTBC treatment provides a model of chronic HT1 with low-grade liver damage and HCC after 3 to 4 months.8 Others have shown that HGA can acutely induce cell death in hepatocytes of Hpd-/-, Fah-/- mice.11, 12 Surprisingly, we found that mutant mice with preexisting liver disease survived normally lethal doses of HGA. Indeed, the already injured Fah-/- hepatocytes were resistant to HGA-induced cell death. This fits with our earlier observation that liver damaged by HT1 is characterized by necro-inflammation and not apoptosis.8 Thus, chronic liver disease paradoxically resulted in protection against cell death rather than increased susceptibility to additional injury. Importantly, this adaptation was a not specific for cell death induced by tyrosine metabolites but extended to injury caused by Jo-2 and APAP, which act by distinct mechanisms.

Potential Mechanisms for Stress-Induced Cell Death Resistance.

Programmed cell death is a complex cellular phenomenon, and several distinct biochemical pathways exist. In addition to classic caspase-mediated apoptosis, other mechanisms have recently been described.37 The cell death resistance observed here in liver damaged by HT1 was mediated not only by the classic apoptotic pathway, modeled by activation of the Fas receptor, but also by other mechanisms, such as the necrosis-like cell death triggered by APAP.

As a result, we wondered which potential mechanisms might underlie this generalized cell death resistance. NF-κB has been described as an essential survival factor in embryonic liver development, and it is well established that, under certain physiological conditions, the activation of NF-κB counteracts the apoptotic machinery.38, 39 However, no significant perturbations of NF-κB signaling were observed in Fah-deficient mice off NTBC, indicating that this pathway was not responsible for the observed cell death protection.

In contrast, marked changes were seen in other pathways/molecules known to play a role in hepatocyte apoptosis. These alterations include the up-regulation of Hsps, and c-Jun, as well as inactivation of the stress-activated signaling pathways. The activation of eukaryotic Hsp gene expression occurs in response to a wide variety of cellular stresses including oxidative stress.40 Hsps behave as molecular chaperones for other cellular proteins. Biochemical and genetic studies have clearly demonstrated the critical role of stress-inducible Hsps in resistance to cell death. The anti-apoptotic effects of Hsps involve protein-protein interactions that seem not to be directly related to their chaperone functions. For example, Hsp70 interacts with apoptosis protease activating factor-1 (APAF-1), thereby preventing APAF-1 from interacting with procaspase-9,41 and Hsp27 has been re ported to specifically interact with cytochrome c in the cytosol, thereby preventing the activation of caspase9.42 Hsp70 also has the capacity to interact with the apoptosis-inducing factor to negatively interfere with caspase-independent apoptosis.43 In addition,, Hsps have been implicated in the protection against APAP-induced necrosis-like liver failure.44 After NTBC withdrawal, we observed a marked increase of Hsp27, 32, and 70. The up-regulation of multiple Hsps, which have been shown to prevent cell death by either apoptosis or necrosis, strongly suggests that they contribute to the acquired phenotype in mice that are off NTBC.

Changes in the protein levels of the stress-activated protein kinases JNK and p38, and in the expression of c-Jun, were also found. JNK and p38 are activated by multiple forms of cellular stress and have been implicated as mediators of stress-induced programmed cell death.34 Accordingly, JNK and p38 were rapidly activated in Fah-/- mice on NTBC that were challenged with HGA, APAP, or Jo-2, and in mice during NTBC withdrawal. However, cell-death-resistant stressed mutant mice showed marked blunting of these responses. This indicates that impairment of JNK and p38 signaling might contribute to cell death resistance. In contrast to the reduced activation of JNK and p38, a strong induction of c-Jun was observed in mice with liver disease. c-Jun belongs to the family of activating protein-1 transcription factors and has been found to be particularly important in the liver. c-Jun-deficient mice are able to produce fetal hepatocytes, but these are progressively lost after birth, presumably by an apoptotic mechanism.35 In addition, c-Jun was identified as a proto-oncogene in the liver. Hepatocyte-specific inactivation of c-Jun resulted in a dramatic increase of chemically induced liver tumor formation.36 Interestingly, c-Jun was particularly required for the survival of tumor cells, and reduced tumor formation strictly correlated with high apoptotic indices in c-Jun-deficient tumors. These data imply that c-Jun not only is required for hepatocyte proliferation, but also has strong anti-apoptotic functions. Our results are in line with these observations and indicate that c-Jun contributes to survival in fetal hepatocytes and liver tumor cells as well as in adult hepatocytes that are not yet transformed.

At this time it is not clear if up-regulation of Hsps, induction of c-Jun, or impairment of stress kinase signaling are mechanistically of primary importance. Future studies in which gain- and loss-of-function approaches to each pathway are used will be needed to dissect the components of this very complex phenotype.

Cell Death Resistance and HCC.

Regardless of the precise molecular mechanism, it is important to emphasize that the stress-induced cell death resistance described here does not evolve in only a few rare cells via genetic inactivation of pro-apoptotic genes. Rather, it represents a general biochemical response that occurs rapidly and affects the majority of hepatocytes.

In terms of evolution, stress-induced cell death resistance can be viewed as an adaptive response, allowing the animal to survive liver injury. However, survival comes at a price, because hepatocytes with DNA damage can live on and acquire mutations. The combination of DNA damage and cell death resistance might contribute to the high incidence of HCC in HT1.3 It might also contribute to the high rate of mutation reversion, which leads to the phenomenon of spontaneous liver repopulation in HT1.5, 45 In addition to HT1, our findings may have general implications for chronic liver disease associated with oxidative stress. We observed that ethanol can also produce a resistance against HGA-induced cell death in HT1 mice, similar to that found following NTBS withdrawal. Cancer proneness is a feature of several common liver diseases associated with oxidative stress, such as hemochromatosis or hepatitis B and C.46 Thus, stress-induced malfunction of the death machinery may play a role in the pathophysiology of these disorders and could be related to the risk of developing HCC.

Ancillary