SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. References
  8. Supporting Information

The nuclear factor erythroid-derived 2, like 2 (Nrf2) transcription factor is a key regulator of the antioxidant defense system, and pharmacological activation of Nrf2 is a promising strategy for prevention of toxin-induced liver damage. However, the consequences of Nrf2 activation on liver regeneration (LR) have not been determined. To address this question, we generated mice expressing a constitutively active Nrf2 (caNrf2) mutant in hepatocytes. Expression of the transgene did not affect liver homeostasis. Surprisingly, however, there was no beneficial effect of Nrf2 activation on CCl4-induced liver injury and fibrosis. Most important, LR after partial hepatectomy was impaired in caNrf2-transgenic mice as a result of delayed hepatocyte proliferation and enhanced apoptosis of these cells after liver injury. Mechanistically, this involved up-regulation of the cyclin-dependent kinase inhibitor p15 and the proapoptotic protein Bcl2l11 (Bim). Using chromatin immunoprecipitation, we show that the p15 and Bcl2l11 genes are direct targets of Nrf2, which are activated under hyperproliferative conditions in the liver. Conclusion: Activated Nrf2 delays proliferation and induces apoptosis of hepatocytes in the regenerating liver. These negative effects of Nrf2 activation on LR should be considered when Nrf2-activating compounds are used for prevention of liver damage. (Hepatology 2014;60:670–678)

Abbreviations
Abs

antibodies

ALT

alanine aminotransferase

ARE

antioxidant response element

AST

aspartate aminotransferase

Bcl-2

B-cell lymphoma 2

Bcl-xL

B-cell lymphoma extra large

BrdU

5-bromo-2'-deoxyuridine

b.w.

body weight

caNrf2

constitutively active Nrf2

ChIP

chromatin immunoprecipitation

Gas1

growth arrest-specific protein 1

IRS

insulin receptor substrate

kb

kilobase

KO

knockout

LR

liver regeneration

mRNA

messenger RNA

NADPH

nicotinamide adenine dinucleotide phosphate

Nqo1

NADP(H) quinone oxidoreductase 1

Nrf2

nuclear factor erythroid-derived 2, like 2

n.s.

nonspecific

PH

partial hepatectomy

qRT-PCR

quantitative reverse-transcriptase polymerase chain reaction

ROS

reactive oxygen species

RPA

RNase protection assay

Tg

transgenic

TSS

transcription start site

TUNEL

terminal deoxynucleotidyl transferase dUTP nick end labeling

WT

wild type

The use of oxygen as an electron acceptor confronts aerobic organisms with the danger of reactive oxygen species (ROS) being formed as by-products of the respiratory chain. ROS formation can be further enhanced by nicotinamide adenine dinucleotide phosphate (NADPH) oxidases, which are particularly abundant in inflammatory cells.[1] Low levels of ROS are required for intracellular signaling,[2] but excessive levels damage all types of cellular macromolecules. To limit ROS-induced cell damage, aerobes developed strategies for efficient ROS detoxification. Of particular importance is the transcription factor nuclear factor erythroid-derived 2, like 2 (Nrf2), which controls expression of numerous genes encoding antioxidant proteins and ROS-detoxifying enzymes.[3] Under homeostatic conditions, Nrf2 is retained in the cytoplasm by binding to Keap1, which also targets Nrf2 for proteasomal degradation. However, some Nrf2 molecules escape this inhibitory mechanism and translocate to the nucleus, where they bind to antioxidant response elements (AREs) in the promoter/enhancer regions of their target genes. This leads to a basal expression of most Nrf2 target genes. In the presence of enhanced ROS levels and/or of electrophiles, which directly couple to Keap1, binding of Nrf2 to Keap1 is weakened, resulting in nuclear accumulation of Nrf2, followed by overexpression or de novo expression of Nrf2 target genes.[3-5]

In vivo, lack of Nrf2 reduced basal and inducible levels of cytoprotective proteins and enhanced the susceptibility to chemical toxicity and carcinogenesis,[5, 6] whereas activation of Nrf2 signaling by genetic deletion of Keap1 or pharmacological activation protected from ROS damage.[6] Moreover, Nrf2 activation by various compounds mediated chemopreventive and anti-inflammatory effects in animal models of cancer, and some of these components are in clinical trials for cancer prevention.[7]

In the liver, lack of Nrf2 aggravated damage induced by various toxins, whereas genetic activation of Nrf2, through deletion of Keap1 in hepatocytes or pharmacological activation, had the opposite effect.[8-10] However, much less is known about the role of Nrf2 in liver regeneration (LR). The mammalian liver has the unique capability to completely regenerate upon injury. After tissue loss resulting from toxin-induced necrosis, viral infections, or surgery, quiescent hepatocytes and other liver cells reenter the cell cycle and proliferate to reconstitute the original liver mass.[11] Insufficient liver regeneration may lead to development of chronic liver disease and, ultimately, liver failure. Because deficiencies in LR are frequently associated with excessive ROS levels,[9] it is particularly interesting to study the role of Nrf2 in LR.

We previously showed that liver regeneration in response to partial hepatectomy (PH) was strongly delayed in Nrf2 knockout (KO) mice. Because we did not observe activation of Nrf2 above basal levels in wild-type (WT) mice during LR,[12] we wondered whether further Nrf2 activation could enhance the regeneration process. Here, we describe a surprising deleterious effect of activated Nrf2 in LR that results from activation of Nrf2 target genes involved in cell-cycle progression and apoptosis.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. References
  8. Supporting Information
Mice

Mice were maintained in a pathogen-free facility according to federal guidelines. Animal experiments were approved by the local veterinary authorities of Zurich, Switzerland.

PH

PH was previously described.[13] After surgery, mice were injected subcutaneously with buprenorphine (Temgesic; 0.1 µg/g body weight; Essex Chemie AG, Luzern, Switzerland). At different time points after surgery, they were sacrificed, blood was collected by cardiac puncture, and remaining liver tissue was harvested. For sham operation, mice were anaesthetized, the abdomen was opened, and liver lobes were pulled out from the abdominal cavity and then returned to their original site. Liver removed during surgery and liver from nonoperated mice served as controls.

CCl4-Induced Acute Liver Injury and Fibrosis

Female mice (8-10 weeks old) were injected intraperitoneally with a single dose of CCl4 (0.4 mg/g body weight [b.w.] in olive oil) or every 3 days with 0.2 mg/g b.w. over a period of 45 days. Vehicle-injected animals served as controls.

Detection of Proliferating or Apoptotic Cells

Proliferating cells were identified by incorporation of 5-bromo-2'-deoxyuridine (BrdU).[12] Apoptotic cells were identified using a terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining kit (Roche, Rotkreuz, Switzerland). Stained sections were photographed (four pictures per animal), and positive and negative cells were counted.

Chromatin Immunoprecipitation

Liver tissue was harvested 24 hours post-PH. Chromatin lysate was precleared and incubated overnight with specific antibodies (Abs) or normal rabbit or sheep immunoglobulin G (see Supporting Information). Protein-bound DNA was amplified by quantitative polymerase chain reaction (qPCR). The ARE of the NADP(H) quinone oxidoreductase 1 (Nqo1) promoter served as a positive control, and a nonspecific (n.s.) region 2 kilobases (kb) away from this ARE was used as a negative control. Percentage of input bound was calculated by the discrete cosine transform method and averaged over at least three experiments.

Statistical Analysis

Statistical analysis was performed using GraphPad Prism software (5.0a; GraphPad Software, Inc., San Diego, CA). P values are two-tailed and were calculated using Mann-Whitney's U test, unless stated otherwise (P < 0.05, P < 0.01, and P < 0.001).

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. References
  8. Supporting Information
Generation of Mice Expressing caNrf2 in Hepatocytes

To selectively activate Nrf2 in the liver and avoid potential off-target effects of pharmacological Nrf2 activators, we generated transgenic (Tg) mice expressing a caNrf2 mutant in hepatocytes. caNrf2 lacks the Neh2 domain of Nrf2, which is responsible for binding to Keap1 (Fig. 1A). We generated Tg mice expressing caNrf2 under control of a cytomegalovirus enhancer and a β-actin promoter. To avoid constitutive expression of the transgene, it is preceded by a transcription/translation STOP cassette flanked by loxP sites (Fig. 1B).[14] These mice were mated with mice expressing Cre recombinase under control of the albumin promoter (AlbCre mice).[15] In the double-Tg offspring (AlbCre_caNrf2 mice), the STOP cassette is excised after onset of Cre expression, leading to expression of caNrf2 in hepatocytes postnatally. Expression of the transgene in the liver of two AlbCre_caNrf2 mouse lines was verified by RNase protection assay (RPA) and western blotting. In both lines, the transgene was expressed at similar levels as the endogenous Nrf2 gene (Fig. 1C,D). Nrf2 and caNrf2 were also detected in primary hepatocytes from AlbCre_caNrf2 mice, but not in cells from Nrf2 KO mice (Fig. 1D). Quantitative reverse-transcriptase (RT)-PCR (qRT-PCR) revealed strong up-regulation of well-established Nrf2 target genes in the liver of AlbCre_caNrf2 mice (Fig. 1E). Thus, the caNrf2 transgene is expressed in hepatocytes and functionally active in vivo. Because we previously excluded effects of transgene integration into the genome,[14] all follow-up studies were performed with mice from one line.

image

Figure 1. Characterization of AlbCre_caNrf2 mice. (A) Functional domains of Nrf2. caNrf2 lacks the Neh2 domain responsible for binding to Keap1. (B) Scheme of transgenes used for generation of AlbCre_caNrf2 Tg mice. (C) RNA (20 µg) from the liver of AlbCre mice and of two different lines of AlbCre_caNrf2 mice (lines 10 and 14; pools from 3 mice per genotype) was analyzed by RPA using a probe that distinguishes between endogenous Nrf2 and transgene-derived caNrf2 and a Gapdh probe (loading control). tRNA (20 µg) was used as a negative control. The hybridization probes (probe) were used as a size marker. (D) Lysates from total liver or cultured primary hepatocytes of AlbCre, AlbCre_caNrf2, or Nrf2 KO mice were analyzed by western blotting using an Ab against Nrf2. Ponceau S staining was used as a loading control. (E) qRT-PCR analysis of liver RNA (N = 6 mice per genotype) for expression of Nrf2 target genes (normalized to 18s rRNA). Expression in AlbCre mice was set as 1 (dashed line). Error bars represent mean ± standard deviation. Abcc4, ATP binding cassette, subfamily C4; Cbr3, carbonyl reductase 3; Gclc, glutamate-cysteine ligase, catalytic subunit; Gclm, glutamate-cysteine ligase, modifier subunit: Gst, glutathione S-transferase; hGH, human growth hormone; Hmox-1, Heme oxygenase-1; tRNA, transfer RNA.

Download figure to PowerPoint

caNrf2 Expression Does Not Affect Liver Homeostasis

AlbCre_caNrf2 mice lacked obvious phenotypic abnormalities, and their body size, body weight, and liver-to-body weight ratio were unaltered, compared to control mice (Supporting Fig. 1A). Histological analysis (Supporting Fig. 1B) and immunohistochemical stainings using Abs against Ly6G and CD68 did not reveal obvious abnormalities and demonstrated the absence of an inflammatory infiltrate (data not shown). The number of proliferating and apoptotic cells was extremely low in mice of both genotypes (Supporting Fig. 1C,D), and the low activities of aspartate and alanine aminotransferases (AST and ALT) and lactate dehydrogenase in serum (Supporting Fig. 1E) confirmed the lack of liver damage.

Activation of Nrf2-Mediated Gene Expression in Hepatocytes Does Not Affect Repair of CCl4-Induced Acute Liver Damage or Development of Fibrosis

To determine the consequences of Nrf2 activation for liver injury, repair, and fibrosis, we first used a model of liver damage that is accompanied by a strong inflammatory response. Under these conditions, ROS detoxification is particularly important because of the high levels of ROS that are produced during CCl4 metabolism[16] and by leukocytes.[17] Histomorphometric analysis of hematoxylin and eosin–stained liver sections revealed a similar extent of necrotic damage in AlbCre_caNrf2 and AlbCre mice 24 and 48 hours after a single CCl4-injection (Supporting Fig. 2A). At 120 hours, AlbCre_caNrf2 mice had slightly more residual necrotic tissue, but most of the damage had been repaired. Serum ALT activities were similar in mice of both genotypes (Supporting Fig. 2B). Hepatocellular proliferation and immune cell infiltration were not affected by caNrf2, and there was no difference in the number of apoptotic cells 24 and 48 hours after CCl4-injection (data not shown).

image

Figure 2. Delayed hepatocyte proliferation and enhanced apoptosis in AlbCre_caNrf2 mice after PH. Cell proliferation (A) and apoptosis (B) were assessed by BrdU incorporation and TUNEL staining, respectively, at different time points after PH. Representative liver sections at 48 hours after PH are shown. Bars, 50 µm. Percentage of proliferating or apoptotic cells was determined by counting four to five microscopic fields per liver at 200x magnification (N = 3-7 mice per genotype and time point). (C and D) Serum activities of AST and ALT before and at different time points after PH. (N = 3-6 mice per genotype and time point.) Error bars represent mean ± standard deviation.

Download figure to PowerPoint

After chronic CCl4 treatment, the area of fibrotic tissue, as determined by morphometric analysis of Sirius Red–stained liver sections was similar in mice of both genotypes (Supporting. Fig. 2C). There was also no difference in the number of apoptotic cells (data not shown) and in ALT serum activity (Supporting Fig. 2D).

Impaired LR After PH in AlbCre_caNrf2 Mice

Next, we determined whether activation of Nrf2-mediated gene expression is beneficial after PH, where liver regeneration occurs in the absence of a strong inflammatory response. Surprisingly, however, we observed a delay in hepatocyte proliferation in caNrf2-expressing mice: Whereas 20% of hepatocytes from AlbCre mice had incorporated BrdU at 48 hours after PH, only 12% of hepatocytes from AlbCre_caNrf2 were BrdU positive at this time point. This reduction was compensated by increased proliferation at 72 hours after PH, and the number of proliferating cells returned to almost basal levels 168 hours after PH (Fig. 2A). TUNEL-positive (apoptotic) cells never exceeded 0.9% of all cells in liver of AlbCre mice, but up to 4% of the cells were TUNEL positive in AlbCre_caNrf2 mice between 24 and 72 hours after PH. No difference was observed at earlier or later time points, suggesting that the apoptotic phenotype is restricted to the proliferation phase (Fig. 2B). ALT and AST levels strongly increased upon PH in mice of both genotypes, but enhanced apoptosis in AlbCre_caNrf2 mice did not lead to a further increase (Fig. 2C,D) and there was no obvious necrosis (data not shown). Numbers of neutrophils and total immune cells were similar in caNrf2-Tg and control mice (Supporting Fig. 3A,B). The liver-to-body weight ratio was not significantly different at any stage of the repair process between AlbCre_caNrf2 and AlbCre mice (Supporting Fig. 3C), most likely as a result of a transiently increased size of hepatocytes in AlbCre_caNrf2 mice (Supporting Fig. 3D). Seven days after injury, ∼70% of the original mass was restored in mice of both genotypes.

image

Figure 3. Enhanced expression of p15 and Bcl2l11 in AlbCre_caNrf2 mice after PH. (A and D) RNA from the liver of AlbCre and AlbCre_caNrf2 mice at different time points after PH was analyzed for p15 (A) and Bcl2l11 (D) using qRT-PCR. Expression levels were normalized to 18s rRNA (N = 4-6). (C and F) Liver lysates (30 µg of protein) from AlbCre and AlbCre_caNrf2 mice at different time points after PH were analyzed by western blotting for p15 (C), Bcl2l11 (F), and GAPDH (C and F). Band intensities were analyzed by densitometry and normalized to GAPDH (B and E; N = 2-3 per genotype). One representative of at least four independent western blotting experiments is shown. Error bars represent mean ± standard deviation.

Download figure to PowerPoint

Insulin/Insulin-Like Growth Factor 1 Receptor and Notch1 Signaling Are Not Affected in AlbCre_caNrf2 Mice

Because the delayed hepatocyte proliferation that occurs in Nrf2 KO mice after PH results, at least in part, from ROS-induced impairment of insulin/insulin-like growth factor 1 receptor signaling,[12] we analyzed the activity of this pathway during LR. However, the activating tyrosine phosphorylation of insulin receptor substrates (IRS)−1 and −2, as well as the inhibitory phosphorylation at serine 307, were similar between AlbCre and AlbCre_caNrf2 mice (Supporting Fig. 4A). Consistent with this finding, there was no difference in phosphorylation of protein kinase B, which occurs in response to IRS-1/IRS-2 phosphorylation at tyrosine 608 (Supporting Fig. 4B).

image

Figure 4. Bcl2l11 and p15 genes are targets of Nrf2 in the regenerating liver. (A) AREs upstream of the Bcl2l11 and p15 TSS. The consensus ARE sequence, as defined by Rushmore et al.,[33] and the rat NQO1 (rNQO1) ARE are shown for comparison. (B) Region upstream of the Bcl2l11 TSS. Arrows indicate binding sites of the primers used for amplification of the two AREs (s1 and s2). (C) ChIP from liver lysates of AlbCre_caNrf2 mice 24 hours after PH using an Nrf2 Ab. Binding of Nrf2 to the Nqo1 ARE or to a nonspecific region of the Nqo1 promoter (n.s.) served as positive or negative controls. Average of the results from five independent ChIP experiments is shown as percentage of the input bound by the Ab. (D) ChIP using liver lysates of AlbCre_caNrf2 mice 24 hours after PH and Abs against total histone H3 and the modified histone, H3K4me2. The average of the values from three independent ChIP experiments is calculated as fold of H3K4me2 over total histone H3 occupancy. Comparisons to the control group (Nqo1 n.s.) were assessed using one-way analysis of variance, followed by Dunnett's test. Error bars represent mean ± standard deviation.

Download figure to PowerPoint

Notch1 had previously been identified as an Nrf2 target gene, and activation of Notch signaling rescued the reduced liver-to-body weight ratio observed after PH in Nrf2 KO, compared to WT, mice.[18] However, we did not observe significant differences in messenger RNA (mRNA) levels of Notch1 or its target (Hes1) before or after PH between AlbCre_caNrf2 and AlbCre mice (Supporting Fig. 4C).

Nrf2 Activation Affects Cell Cycle and Apoptosis Regulators in the Regenerating Liver

Next, we used an unbiased approach to unravel the mechanisms underlying the regeneration phenotype of AlbCre_caNrf2 mice. Expression profiling was performed using RNA from liver tissue removed upon surgery (0 hours) and 24 hours after PH. A large number of previously identified cytoprotective Nrf2 target genes, in particular, genes with a function in ROS or drug detoxification, were up-regulated in AlbCre_caNrf2 mice, demonstrating the validity of the data (Supporting Table 1). Because Nrf2 is usually a positive regulator of gene expression, we analyzed the data for genes that are up-regulated by caNrf2 and encode cell-cycle inhibitors. This analysis revealed up-regulation of the genes encoding the cyclin-dependent kinase inhibitor p15 and growth arrest-specific protein 1 (Gas1). Their differential expression was verified by qRT-PCR (Fig. 3A and Supporting Fig. 5A). The delay in cell-cycle progression was also reflected by a mild reduction in expression of cyclins D1 and B1 as well as of aurora B kinase (Supporting Fig. 5B-D). By contrast, there was no significant difference in expression of p21 and p27 (Supporting Fig. 5E,F).

Because of the known inhibitory effect of p15 on hepatocyte proliferation,[19] we focused on this protein. p15 mRNA and protein levels were strongly increased in AlbCre_caNrf2 versus AlbCre mice 24 hours after PH, whereas no difference was noted in noninjured liver. A second, but less pronounced, increase was observed 72 hours after PH, most likely corresponding to the second round of hepatocyte proliferation (Fig. 3A,B).

When the array data were analyzed for up-regulation of proapoptotic genes, we identified a significant regulation of Bcl2l11 (Bim). This was confirmed by qRT-PCR and western blotting (Fig. 3D-F), and the time course of Bcl2l11 expression correlated with the increase in apoptotic cells during liver regeneration.

p15 and Bcl2l11 Are Targets of Nrf2 in the Regenerating Liver

To determine whether p15 and Bcl2l11 are direct targets of Nrf2, we searched for AREs within 10 kb upstream of the transcriptional start sites (TSS) and indeed identified putative AREs in these regions (Fig. 4A,B). Binding of Nrf2 to an ARE in the murine p15 gene promoter and an ARE in the human BCL2l11 gene promoter had previously been demonstrated by chromatin immunoprecipitation (ChIP) using lysates of mouse embryonic fibroblasts[20] or human lymphoblastoid cell lines,[21] respectively, although the functional consequences had not been explored. To test whether the previously identified p15 ARE[20] (Fig. 4A) and two AREs that we identified upstream of the TSS of the murine Bcl2l11 gene (Fig. 4A,B) are bound by caNrf2/Nrf2 in the regenerating liver, we performed ChIP using lysates from liver of AlbCre_caNrf2 mice 24 hours after PH. We indeed observed strong binding of caNrf2/Nrf2 to these AREs (Fig. 4C). Binding to an ARE in the promoter of the well-characterized Nrf2 target gene, rat Nqo1, served as a positive control (Fig. 4C). ChIP using Abs against histone H3 and dimethylated histone H3 (H3K4me2) revealed that the p15, Bcl2l11, and Nqo1 AREs are located within regions of open chromatin (Fig. 4D), indicating active gene expression. Consistent with the lack of Nrf2 activation in WT mice during LR,[12] only very low enrichment of Nrf2 at the tested AREs was observed using liver lysates obtained from control mice 24 hours post-PH (data not shown). Taken together, these data identified p15 and Bcl2l11 as direct targets of activated Nrf2 in the regenerating liver.

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. References
  8. Supporting Information

Previous studies demonstrated a potent protective effect of Nrf2 in various models of toxin-induced liver injury, and Nrf2-activating compounds are therefore promising candidates for liver protection under stress conditions. However, under these conditions, regeneration of the damaged liver is also required. We and others previously showed that loss of Nrf2 impairs the liver's regenerative capacity,[12, 22] although Nrf2 is not further activated by liver injury.[12] Therefore, we determined whether Nrf2 activation above basal levels enhances LR. Instead of using Nrf2-activating pharmacological compounds, which have additional targets and whose activities are not restricted to hepatocytes, we used a genetic approach to determine the consequences of selective activation of Nrf2-mediated gene expression in those cells. This was achieved using a well-characterized caNrf2 mutant.[14, 23] This approach is also superior to analysis of Keap1-deficient mice, because Keap1 not only binds Nrf2, but also additional signaling molecules.[24]

Surprisingly, activation of Nrf2 in hepatocytes did not improve liver regeneration after CCl4 treatment, although expression of various ROS-detoxifying enzymes and antioxidant proteins was strongly up-regulated by the caNrf2 transgene. Regeneration was even delayed upon PH, indicating that Nrf2 activates expression of genes that interfere with appropriate regeneration. In the case of CCl4-mediated injury, this may be compensated by the positive effect of Nrf2 activation on ROS detoxification, which is particularly important in toxin-mediated liver injury that induces a strong inflammatory response and involves production of ROS through CCl4 metabolism.[16] However, ROS detoxification may be less rate limiting upon PH, which is not accompanied by strong inflammation.

In a search for the mechanisms underlying impaired regeneration in AlbCre_caNrf2 mice, we identified p15 and Bcl2l11 as direct target genes of Nrf2 in the regenerating liver. Up-regulation of p15 within a few hours after PH provides a likely explanation for delayed proliferation of caNrf2-expressing hepatocytes, because elevated expression of p15 in hepatocytes had been shown to inhibit their proliferation in vivo.[19] Furthermore, it was suggested that p15 is required to maintain adult hepatocytes in a quiescent state.[25] Although the effect of Gas1 on hepatocytes has, as yet, not been analyzed, up-regulation of this protein, which inhibits proliferation and enhances apoptosis of various normal and malignantly transformed cells,[18, 26] is likely to further contribute to the regeneration phenotype of AlbCre_caNrf2 mice. Reduced expression of p15 and, particularly, of Gas1 at 72 hours compared to 24 hours, which may involve the activity of other transcriptional regulators, provides a likely explanation of why the proliferation block was overcome 72 hours post PH.

Interestingly, caNrf2 also activated expression of Bcl2l11, a proapoptotic gene[27] in the regenerating liver. This provides a likely explanation for the enhanced apoptosis in the liver of these mice after injury through destruction of the appropriate balance between pro- and antiapoptotic B-cell lymphoma 2 (Bcl-2) family members.[28] A recent study identified the antiapoptotic proteins, B-cell lymphoma extra large (Bcl-xL) and Bcl-2, as targets of Nrf2 in liver cancer cells, and this was associated with prevention of apoptosis by chemical Nrf2 activation.[29, 30] By contrast, we observed Nrf2-mediated up-regulation of the proapoptotic protein, Bcl2l11, but not of Bcl-xL or Bcl-2 (data not shown), in hepatocytes of the regenerating liver, indicating differential effects of Nrf2 on regulation of pro- or antiapoptotic genes in cancerous and precancerous versus normal cells. Consistent with this hypothesis, Bcl2l11 expression could not be induced in immortalized murine hepatocytes (AML12 cells) by the Nrf2-activating compound, tert-butylhydroquinone, whereas a mild induction was observed in primary hepatocytes (Supporting Fig. 6). The particularly strong Nrf2-mediated activation of Bcl2l11 and also of p15 in hepatocytes that occurred after PH may require additional transcriptional cofactors. Such factors could be activated by stimuli that induce cell-cycle entry/progression, a hypothesis that is supported by the lack of activation of these genes in the noninjured liver of AlbCre_caNrf2 mice, but induction of these genes in primary hepatocytes (Supporting Fig. 6), in which G0/G1 transition had been triggered by enzymatic liver dissociation.[31] Selective Nrf2 regulation under hyperproliferative conditions has also been shown for genes encoding various metabolic enzymes.[32] In the future, it will be interesting to identify the cofactors that modify the response to Nrf2 in a cell-type- and proliferation-dependent manner.

Identification of p15 and Bcl2l11 as Nrf2 targets in hepatocytes in vivo revealed a previously unknown function of Nrf2 in control of cell-cycle progression and apoptosis during LR. This could be an important mechanism to prevent uncontrolled proliferation and survival of hepatocytes under conditions when Nrf2 is activated (e.g., by toxic chemicals) and could thus be a strategy to prevent malignant transformation in highly proliferating tissues. Whereas such a mechanism would be beneficial, it obviously comes with negative consequences for regeneration. This needs to be considered when Nrf2-activating compounds are used for chemoprevention under conditions where an efficient regenerative response of the liver is required. In the future, it will therefore be important to determine the effect of such compounds on the LR process.

Acknowledgment

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. References
  8. Supporting Information

The authors thank Drs. Jelena Kühn Georgijevic and Hubert Rehrauer (Functional Genomics Center Zurich) for help with the microarray experiments, Dr. Tamara Ramadan (ETH Zurich) for help with the PH experiments, Dr. Yuet-Wai Kan (San Francisco, CA) for Nrf2 KO mice, and Dr. Dennis Roop (Aurora, CA) for the Nrf2 Ab. U.A.K. is a member of the International Research Training Group (IRTG 1331; Konstanz/Zurich).

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgment
  7. References
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article at the publisher's website.

FilenameFormatSizeDescription
hep26964-sup-0001-suppinfo01.pdf1692KSupporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.