Potential conflict of interest: Dr. Sporn received grants from Reata Pharmaceuticals.
In tyrosinemia type 1 (HT1), accumulation of toxic metabolites results in oxidative stress and DNA damage, leading to a high incidence of hepatocellular carcinomas. Nuclear factor erythroid-2 related factor 2 (Nrf2) is a key transcription factor important for cellular protection against oxidative stress and chemical induced liver damage. To specifically address the role of Nrf2 in HT1, fumarylacetoacetate hydrolase (Fah)/Nrf2−/− mice were generated. In acute HT1, loss of Nrf2 elicited a strong inflammatory response and dramatically increased the mortality of mice. Following low grade injury, Fah/Nrf2−/− mice develop a more severe hepatitis and liver fibrosis. The glutathione and cellular detoxification system was significantly impaired in Fah/Nrf2−/− mice, resulting in increased oxidative stress and DNA damage. Consequently, tumor development was significantly accelerated by loss of Nrf2. Potent pharmacological inducers of Nrf2 such as the triterpenoid analogs 1[2-cyano-3,12-dioxooleana-1,9(11)-dien-28-oyl]imidazole have been developed as cancer chemoprevention agents. Pretreatment with 1[2-cyano-3,12-dioxooleana-1,9(11)-dien-28-oyl]imidazole dramatically protected Fah−/− mice against fumarylacetoacetate (Faa)-induced toxicity. Our data establish a central role for Nrf2 in the protection against Faa-induced liver injury; the Nrf2 regulated cellular defense not only prevents acute Faa-induced liver failure but also delays hepatocarcinogenesis in HT1. (HEPATOLOGY 2008;48:487–496.)
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Hereditary tyrosinemia type-1 (HT1) is an autosomal-recessive disease caused by a genetic inactivation of the enzyme fumarylacetoacetate hydrolase (FAH) that carries out the last step of the tyrosine catabolism pathway (Fig. 1A). FAH is mainly expressed in hepatocytes, and accumulation of toxic metabolites such as fumarylacetoacetate (FAA) lead to acute or chronic liver failure.1In vitro, FAA has been shown to be highly mutagenic2 and to cause oxidative damage by reacting with glutathione and sulfhydryl groups of proteins.3 Human HT1 is characterized by an extremely high susceptibility for liver cancer.4 Currently, the drug nitrisinone (2-[2-nitro-4-(trifluoromethyl)benzoyl] cyclohexane-1,3-dione, or NTBC), which blocks the pathway upstream of the formation of FAA, is used to treat patients with HT1. Although the efficacy of this treatment has been well established, many late complications persist.5 Furthermore some patients on NTBC have persistently high alpha-fetoprotein (AFP) levels and are at risk to develop hepatocellular carcinoma (HCC).6 Therefore, additional therapies are required to completely prevent liver cancer in HT1.
A murine model of Fah deficiency has been developed that represents all phenotypic and biochemical manifestations of the human disease on an accelerated time scale.7, 8 Mice receiving NTBC treatment show improved cancer-free survival, but HCCs still develop after a longer period of time even with maximal therapy.9 We have previously shown that Fah−/− hepatocytes divide poorly and develop a resistance against different cell death triggers during Faa-induced liver injury.10 p21 was identified to be responsible for the cell cycle arrest following NTBC withdrawal. Fah/p21−/− mice do not only show continuous liver regeneration during NTBC withdrawal, but they also display a much accelerated hepatocarcinogenesis, suggesting that p53/p21 are important tumor suppressors in HT1.17 The mechanisms underlying the observed cell death resistance are currently unknown. Recently, it has been suggested that Akt confers cell death resistance in Fah−/− mice, thereby favoring hepatocarcinogenesis.11 However, Akt was mainly activated in mice 5 weeks after NTBC withdrawal, whereas apoptosis resistance is already well established after 2 weeks.10 FAA-induced liver injury occurs immediately after NTBC withdrawal; therefore, pathways other than Akt are likely activated to protect hepatocytes. We and others have previously shown that oxidative stress inducible genes such as Nqo1 are activated very early following NTBC withdrawal in Fah−/− mice.3 Most of these genes contain a particular regulatory DNA motif termed antioxidant response element and are regulated by nuclear factor erythroid-2 p45-related factor 2 (Nrf2). Nrf2 belongs to the family of cap-n-collar transcription factors that share a highly conserved basic leucine zipper structure.12 Several studies have established an important role for Nrf2 in the regulation of basal and inducible expression of a battery of antioxidants and other cytoprotective genes. Disruption of Nrf2 leads to enhanced sensitivity against xenobiotics and carcinogens in various organs including the liver.13 We generated Fah/Nrf2−/− mice to specifically address the role of Nrf2 in HT1. In this study, we report that Nrf2 is a critical modulator of FAA-induced toxicity, which dramatically improves survival in acute HT1 and liver injury in chronic HT1. Importantly, Nrf2-mediated hepatocyte survival does not favor hepatocarcinogenesis, but significantly delays hepatocarcinogenesis.
Nfr2 Is Required for the Survival of Fah−/− Mice Following NTBC Withdrawal.
Several known Nrf2 target genes are induced in Fah−/− mice following complete NTBC withdrawal (Supplementary Table 1). Additionally, Nrf2 messenger RNA levels increased significantly in Fah−/− mice in which NTBC treatment had been stopped (Fig. 1A). To determine the significance of the Nrf2-regulated stress response in HT1, Fah/Nrf2−/− mice were generated.
Following complete NTBC withdrawal, Fah−/− mice survived for more than 4 weeks (n = 12). In contrast, Fah/Nrf2−/− mice died within 5 days (n = 13; P < 0.0001; Fig. 1B). NTBC withdrawal induced acute liver failure with histologically-evident hepatocellular necrosis and multiple terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick-end labeling (TUNEL) positive hepatocytes. Compared to the severe damage in double knockout mice, lesions in Fah−/− mice exposed to the same treatment were only moderate (Fig. 1C). To further characterize the Faa-induced cell death, cleavage of caspase-9 and caspase-3 as markers for apoptosis were measured. Processing of the precursor forms of caspase-9 and caspase-3 was clearly evident in Fah/Nrf2−/− mice in contrast to Fah−/− controls (Fig. 1D). Concomitant to histological liver injury, alanine aminotransferase levels were significantly elevated in Fah/Nrf2−/− mice (P = 0.00015; Fig. 1E). Activated Kupffer cells produce multiple inflammatory cytokines, including tumor necrosis factor–α. Interestingly, Fah/Nrf2−/− mice displayed significantly higher tumor necrosis factor–α levels than Fah−/− mice (P = 0.00043; Fig. 1F). The stronger acute phase reaction was further reflected by significantly higher serum interleukin-6 levels (P = 0.015; Fig. 1F) and a strong activation of signal transducer and activator of transcription 3 and c-Jun N-terminal kinase (Jnk) in livers of critically sick Fah/Nrf2−/− mice (Fig. 1D).
Accumulation of Faa leads to negative feedback regulation of upstream enzymes in the tyrosine catabolism pathway.14 To determine whether loss of Nrf2 modulated enzymes of the tyrosine catabolism pathway, expression of Tat, Hpd, Hgd, and Maai was analyzed by microarrays. A similar down-regulation of these genes was evident in Fah−/− and Fah/Nrf2−/− mice following NTBC withdrawal, indicating that Nrf2 did not affect expression of enzymes in the pathway (Fig. 1G).
Fah/Nrf2−/− Mice Develop More Severe Hepatitis and Liver Fibrosis on Low-Dose NTBC Treatment.
Next, we were interested in the role of Nrf2 in chronic HT1 liver disease and its potential involvement in cancer formation. For this experiment, mice were exposed to a reduced treatment regimen of NTBC for 5 to 6 months. Overall, Fah/Nrf2−/− mice displayed more severe hepatitis, on both 100% NTBC and 10% NTBC. Histological examination revealed multiple abnormal hepatic features with diffuse necroinflammation, bile duct proliferation, hepatocyte size variations, and focal microsteatosis, which was most evident in the low-dose NTBC treatment group (Fig. 2A). In contrast to Fah/Nrf2−/− mice following complete NTBC withdrawal, no TUNEL-positive cells were evident in any group on the reduced NTBC treatment.
To determine whether increased liver injury in Fah/Nrf2−/− mice on low-dose NTBC results in more liver fibrosis, deposition of collagen was analyzed by immunohistochemistry. Representative micrographs of Masson's trichrome and Sirius red staining of liver tissue sections are shown in Fig. 2A. In Fah−/− mice on 10% NTBC and Fah/Nrf2−/− mice on 10% NTBC and 100% NTBC an increased collagen accumulation was evident. A computer-aided morphometric analysis revealed that the area of fibrosis in sections of Fah/Nrf2−/− mice on low-dose NTBC was increased two-fold in comparison to Fah−/− controls (n = 4; P < 0.0001; Fig. 2B). Because α-smooth muscle actin (α-Sma)-positive myofibroblasts are mainly responsible for the production of extracellular matrix in the liver, activation of hepatic myofibroblasts was determined in mice on 100% NTBC and 10% NTBC. Livers from Fah/Nrf2−/− mice on low-dose NTBC showed the strongest α-Sma staining. In line with the immunohistochemical analysis, several collagen genes were more strongly induced in Fah/Nrf2−/− mice compared to Fah−/− controls (Fig. 2C).
Cellular Detoxification System Is Impaired in Fah/Nrf2−/− Mice.
There is increasing evidence that Nrf2 plays a central role in the regulation of phase 2 detoxification enzymes.15 Additionally, Nrf2 regulates the synthesis of nonenzymatic scavengers such as glutathione. Thus, we wished to determine how loss of Nrf2 affects the regulation of these genes in almost healthy mice (100% NTBC) and in mice with moderate (10% NTBC) and severe (off NTBC) liver injury. Glutathione (GSH) and oxidized glutathione (GSSG) levels were similar in mice of both groups on NTBC (n = 4; Fig. 3A). Complete NTBC withdrawal caused a significant drop of GSH in livers of Fah/Nrf2−/− mice (P < 0.00001); in contrast, GSH levels were significantly increased Fah−/− controls (n = 4; P < 0.00001; Fig. 3A). GSH/GSSG levels were not significantly changed in either group, suggesting that glutathione primarily acts as a nonenzymatic scavenger in HT1. The biosynthesis of glutathione is dependent on the availability of glutamate-cysteine ligase, which consists of a heavy catalytic subunit (Gclc) and a light regulatory subunit (Gclm). In line with increased glutathione levels, protein levels of Gclc and Gclm were induced in Fah−/− mice and remained unchanged in Fah/Nrf2−/− mice (Fig. 3B). Previously, it has been shown that supplementation of N-acetylcysteine (NAC), a precursor of GSH, ameliorates acetaminophen induced liver injury in Nrf2−/− mice.13 Therefore, NAC or GSH monoethylester, a cell-permeable analog of GSH, were given to Fah/Nrf2−/− mice during NTBC withdrawal. However, supplementation of NAC was not sufficient to reduce the mortality of Fah/Nrf2−/− mice (n = 6; Fig. 3C). For this reason, we wondered which other genes involved in xenobiotic metabolism were affected by loss of Nrf2 in HT1. Microarray analysis identified a broad array of genes, which were differently regulated in Fah/Nrf2−/− mice on low-dose NTBC in comparison to Fah−/− controls (Supplementary Table 2). Several of these genes have previously been shown to be dependent on Nrf2 in different organs and cell types. Lower expression of glutathione S-transferases (GSTs), Gst3, Gst4, and Gst6, was further confirmed by semiquantitative real-time polymerase chain reaction (Fig. 3D).
Fah/Nrf2−/− Livers Show Slightly More Oxidative Damage.
Previously, oxidative and endoplasmic reticulum stress has been implicated in the pathogenesis of Faa-induced liver disease.3, 16 We therefore sought to determine whether loss of the Nrf2-regulated stress response increases oxidative and endoplasmic reticulum stress in Fah−/− mice. The 8-oxoguanine/2′-deoxyguanosine ratio, a marker for oxidative damage to DNA, was not significantly different in Fah/Nrf2−/− and Fah−/− mice on 100% NTBC (n = 4); in contrast, Fah/Nrf2−/− mice on low-dose NTBC displayed significantly higher 8-oxoguanine/2′-deoxyguanosine ratio than Fah−/− controls (P = 0.006; Fig. 3E). Similarly, for the same amount protein loaded, more carbonyl groups, which are a hallmark of the oxidation status of proteins, were detected in Fah/Nrf2−/− mice on low-dose NTBC (n = 4; Fig. 3F).
Identification of Functional Gene Groups and Transcriptional Networks Differently Regulated in Fah/Nrf2−/− and Fah−/− Mice Off NTBC.
To better understand the molecular changes occurring in Fah−/− and Fah/Nrf2−/− mice during liver injury, Ingenuity Pathway Analysis was performed with genes differently regulated between both groups. The most significant categories modified in Fah/Nrf2−/− mice were those related to cancer. This group includes 137 genes, which are involved in various aspects of carcinogenesis (Supplementary Table 3). Next, we were interested how the differently regulated genes were integrated into specific regulatory and signaling pathways. Biological relevant networks were drawn through the use of Ingenuity Pathway Analysis and several major pathways were identified. The analysis identified p53, cyclin D1, and c-jun as important participants in the response to Faa-induced liver injury based on the differential expression of other genes in the same pathway (Fig. 4A). In line with the microarray data, increased levels of p53, p21, c-jun, and cyclin D1 were evident in Fah/Nrf2−/− mice on low-dose NTBC compared to Fah−/− mice (Fig. 4B,C).
We have previously compared proliferation of hepatocytes following partial hepatectomy in wild-type and Fah−/− mice off NTBC and observed a markedly impaired cell division. Subsequently, p21 was identified as critical regulator for the cell cycle arrest in Fah−/− hepatocytes.17 We were therefore interested in whether activation of the p53/p21 pathway affects liver regeneration in Fah/Nrf2−/− mice. Fah−/− mice on 100% NTBC and 10% NTBC displayed only a few Ki67-positive hepatocytes (n = 4). In contrast, significantly more Ki67-positive cells were detectable Fah/Nrf2−/− mice on 100% NTBC (P = 0.01). Interestingly, however, Ki67 labeling remained low in mice with the most severe damage, Fah/Nrf2−/− mice on low-dose NTBC (Fig. 4D). Along with the activation of the p53/p21 pathway, livers of Fah/Nrf2−/− mice on low-dose NTBC displayed enormous quantities of AFP, which were almost undetectable in livers of Fah−/− mice (Fig. 4C). For situations in which hepatocyte regeneration is compromised after liver damage, bipotential hepatic progenitor cells, oval cells, are activated, which are capable of differentiating into either hepatocytes or cholangiocytes. Immunohistochemical staining revealed a strong activation of oval cells in livers of Fah/Nrf2−/− mice (n = 4; Fig. 4C).
Nrf2 Delays Liver Tumor Development in Fah−/− Mice.
To analyze the role of Nrf2 in Faa-induced liver tumor development, mice were kept on the regular-dose NTBC or on low-dose NTBC for 9 months. None of the Fah−/− or Fah/Nrf2−/− mice on 100% NTBC and only 10% of the Fah−/− mice on 10% NTBC developed tumors within this time frame. In contrast, enlarged livers with multiple tumors were evident in all Fah/Nrf2−/− mice on low-dose NTBC (n = 12; P < 0.0001; Fig. 5A,C). The average number of tumors was 13 per Fah/Nrf2−/− mouse and one in the Fah−/− mouse (Fig. 5B). Histological analysis of liver tumors in Fah/Nrf2−/− mice revealed increased proliferation as shown by bromodeoxyuridine staining (n = 8; Fig. 5D). Interestingly, p21 expression was lost during liver tumor development in livers of Fah/Nrf2−/− mice on 10% NTBC (Fig. 5E,F).
Chemopreventive Drug 1[2-Cyano-3,12-dioxooleana-1,9(11)-dien-28-oyl]imidazole Prevents Faa-Induced Liver Failure.
Induction of phase 2 enzymes is an effective mechanism of protection against carcinogenesis and other forms of liver toxicity. One of the most potent substances of this group is 1[2-Cyano-3,12-dioxooleana-1,9(11)-dien-28-oyl]imidazole (CDDO-Im), which activates a number of genes regulated by Nrf2.18, 19 To determine the protective effects of CDDO-Im in HT1, Fah−/− mice were gavaged with CDDO-Im on four consecutive days. During the treatment mice were fed a low protein diet; on day 5, NTBC was withdrawn and mice received a high protein diet to induce a rapid accumulation of toxic Faa (Fig. 6A). Untreated Fah−/− mice died within 4 days; in contrast, 80% of the CDDO-Im treated mice survived for more than 8 days (n = 8; P = 0.0007; Fig. 6B). CDDO-Im did not protect Fah/Nrf2−/− mice, which died within 75 hours following NTBC withdrawal. Histological examination revealed severe hepatocellular damage in untreated mice compared to only moderate damage in mice treated with CDDO-Im (Fig. 6C). The protective effects of CDDO-Im were associated with an induction of Gclc and Gclm (Fig. 6D). Expression of both proteins was abrogated during the Faa induced acute liver failure in untreated Fah−/− mice.
The important and critical role of Nrf2 in the coordination of the mammalian cellular response to a variety of harmful stimuli is increasingly established. Activation of Nrf2 results in up-regulation of a battery of antioxidant proteins in several organs including the liver and lung. Here, disruption of Nrf2 dramatically increased the mortality of Fah/Nrf2−/− mice following complete NTBC withdrawal due to an acute liver failure. Chronic HT1 is a devastating disease with a high incidence of HCC.1, 20 A similar picture became evident in Fah/Nrf2−/− mice on low-dose NTBC. Fah/Nrf2−/− mice developed a more severe hepatitis and more liver fibrosis than Fah−/− controls. Additionally liver tumor development was significantly accelerated by loss of Nrf2. Therefore, the Nrf2-regulated cellular defense does not only prevent acute FAA-induced liver failure but also prevents chronic FAA-induced hepatic injury and delays hepatocarcinogenesis in HT1.
The exact molecular mechanisms by which FAA induces acute liver failure are not completely clear, but it most likely results from direct toxic effects triggering intracellular signaling pathways that promote apoptosis and necrosis. Loss of Nrf2 caused a strong activation of the stress kinase Jnk following NTBC withdrawal. Sustained activation of Jnk has been shown to promote drug-induced hepatocellular necrosis.21 In addition to these events occurring intracellularly in parenchymal cells, the progression and severity of Faa-induced liver injury might be aggravated by a deregulated immune response in Fah/Nrf2−/− mice. There is increasing evidence that cytokines and chemokines released by Kupffer and other inflammatory cells contribute to the progression of drug induced liver failure in vivo.22 Recently, it has been shown that Nrf2 is a critical regulator of the innate immune system that dramatically improves survival during experimental sepsis by protecting against deregulated inflammation.23 In Fah−/− mice, loss of Nrf2 elicited a significantly stronger inflammatory response following NTBC withdrawal in Fah/Nrf2−/− mice, likely contributing to liver injury.
Faa is a highly electrophilic compound, which is able to disrupt essential sulfhydryl reactions by complexation with proteins and GSH. Moreover, it is an efficient depletor of GSH; and vice versa, GSH depletion not only enhances FAA-induced cell death but also amplifies its mutagenicity.24 Nrf2 is critically important for the regulation of GSH homeostasis in the liver,13 and a significant drop of glutathione was evident in Fah/Nrf2−/− mice following NTBC withdrawal. Supplementation of NAC or GSH-Me, however, was not sufficient to improve the survival of Fah/Nrf2−/− mice following NTBC withdrawal, suggesting that the coordinated regulation of multiple Nrf2 target genes is required for the protection against Faa-induced liver injury. A marked reduction of multiple other phase 2 detoxifying enzymes, including several GSTs, was evident in livers of Fah/Nrf2−/− mice. GST enzymes are not only critical in the detoxification of xenobiotics, they are also regarded as a major cellular defense against oxidative stress.25
Factors that determine the progression to HCC in human HT1 are not well understood. To better understand the molecular mechanisms occurring in Fah/Nrf2−/− mice contributing to or preventing tumor development, gene expression analysis was performed. Computational analysis identified genes related to xenobiotic metabolism and to cancer as most significant themes of genes differently regulated in Fah/Nrf2−/− mice in comparison to Fah−/− controls. Pathway analysis identified p53/p21 as the most significant network during early FAA-induced liver disease. Activation of the the p53/p21 DNA damage is most likely a direct affect of the alkylating metabolite FAA on DNA. Previously it has been shown that FAA induces mitotic abnormalities and chromosomal instability in vitro and in vivo.26, 27 Together these data suggest that loss of Nrf2-regulated detoxification of FAA causes severe DNA damage and therefore activation of the p53/p21 DNA damage response. Activation of p21 hepatocytes impairs proliferation in Fah/Nrf2−/− mice on low-dose NTBC and mobilization of the hepatic stem cell pool appears to be important to compensate for the loss of hepatocytes in Fah/Nrf2−/− mice. In Fah/Nrf2−/− mice on low-dose NTBC, massive AFP staining was detectable. Subsequently, Fah/Nrf2−/− mice developed much earlier liver tumors than Fah−/− controls. Particularly, c-jun and cyclin D seem to be crucial for initiation of early tumor development in Fah/Nrf2−/− mice.28, 29 Furthermore, loss of the p21 cell cycle checkpoint correlates with the onset of tumorigenesis, which has also been identified as a feature characterizing the transition of premalignant to malignant lesions during multistep hepatocarcinogenesis in humans.30
HT1 patients receiving NTBC treatment are still at risk of developing HCCs. To analyze the potential of Nrf2 as molecular target for chemoprevention in HT1, the triterpenoid CDDO-Im was used. CDDO-Im dramatically protected Fah−/− mice against FAA-induced toxicity, but failed to increase survival of Fah/Nrf2−/− mice. The striking protection of Fah−/− mice in acute HT1 strongly suggests that treatment with CDDO-Im will delay tumor development in the chronic disease.
In conclusion, these results show that Nrf2 is critically important for a coordinated stress and inflammatory response against toxic tyrosine metabolites. Loss of Nrf2 dramatically increased the mortality of mice in acute HT1. Additionally, the Nrf2-regulated defense response significantly ameliorated chronic liver injury and delayed liver tumor development in chronic HT1. Based on these findings, we propose a cross-talk between Nrf2 and p53/p21 to prevent liver cancer in HT1 (Fig. 7). In Fah−/− mice on fully protective doses of NTBC, FAA levels are low and Nrf2 and p53/p21 are not activated. On suboptimal NTBC therapy, FAA levels are high enough to induce Nrf2, which prevents DNA damage and therefore liver cancer. Loss of Nrf2 results in increased DNA damage and activation of the p53/p21 DNA damage response. Subsequent loss of the p21 cell cycle checkpoint leads to tumor development in Fah/Nrf2−/− mice. Complete NTBC withdrawal leads to a massive accumulation of FAA and a strong activation of Nrf2 and p53/p21. In Fah-deficient mice, activation of Nrf2 prevents the cell from crossing the cell death threshold, and activation of p53/p21 induces a sustained cell cycle arrest, allowing DNA damage repair. Without Nrf2, however, Fah−/− hepatocytes undergo massive apoptosis/necrosis and mice die from liver failure. In the murine model, activation of the Nrf2 pathway with CDDO-Im significantly improved survival in the acute disease. Thus, Nrf2 represents a promising target for cancer prevention for HT1 patients, who are still at risk to develop liver cancer under NTBC therapy.