The Transcription Factor c-Jun Protects against Liver Damage following Activated β-Catenin Signaling

Background The Wnt/β-Catenin signaling pathway is central for liver functions and frequently deregulated in hepatocellular carcinoma (HCC). Analysis of the early phenotypes and molecular events following β-Catenin activation is therefore essential for better understanding HCC pathogenesis. The AP-1 transcription factor c-Jun is a putative β-Catenin target gene and promotes hepatocyte survival, proliferation, and liver tumorigenesis, suggesting that c-Jun may be a key target of β-Catenin signaling in the liver. Methodology/Principal Findings To address this issue, the immediate hepatic phenotypes following deletion of the tumor suppressor Apc and subsequent β-Catenin activation were analyzed in mice. The contribution of c-Jun to these phenotypes was dissected in double mutant animals lacking both, Apc and c-Jun. β-Catenin was rapidly activated in virtually all Apc mutant hepatocytes while c-Jun was induced only after several days, suggesting that its expression was rather a secondary event following Apc deletion in the liver. Loss of Apc resulted in increased hepatocyte proliferation, hepatomegaly, deregulated protein metabolism, and premature death. Interestingly, additional deletion of c-Jun did not affect hepatocyte proliferation but resulted in increased liver damage and mortality. This phenotype correlated with impaired expression of hepatoprotective genes such as Birc5, Egfr Igf1 and subsequently deregulated Akt signaling. Conclusions/Significance These data indicate that c-Jun is not a primary target of β-Catenin signaling in the liver, but rather protects against liver damage, which in turn may promote liver tumorigenesis.


Introduction
The Wnt/b-Catenin signal transduction cascade is a major regulator of liver development and hepatocyte function [1]. Moreover, b-Catenin is an established oncogene and defines a genetically distinct subset of hepatocellular carcinoma (HCC) [2]. Its activity therefore needs to be tightly controlled. Under resting conditions, b-Catenin is bound by a multiprotein complex consisting of Adenomatosis polyposis coli (Apc), Axin and Glycogen synthase kinase 3b and is targeted for proteasomedependent degradation. Disruption of this complex, e.g., upon binding of Wnt ligands to their respective receptors or inactivating mutations of Apc results in stabilization and nuclear translocation of b-Catenin, where it cooperates with ternary complex factors (Tcf) to form a ternary complex with DNA responsive elements, thereby regulating expression of its target genes [3].
Genetic mouse models have been particularly useful to determine the functions of Apc and b-Catenin in the liver. Constitutive loss of Apc or gain of b-Catenin function results in early lethality [1,4,5]. In contrast, conditional and liver-specific knockout of Apc after birth results in increased hepatocyte proliferation, hepatomegaly and premature mortality within days, possibly due to perturbations in protein and ammonia metabolism [6,7]. In addition, Apc acts as tumor suppressor in the liver since its mosaic deletion in only a subset of hepatocytes is sufficient to cause HCC [6]. Therefore, a detailed analysis of the molecular pathways following b-Catenin activation is essential to better understand HCC pathogenesis with the potential to identify novel candidates for targeted therapies.
The immediate phenotypes of Apc loss are b-Catenin-dependent and rescued upon additional knockout of Ctnnb1 (which encodes b-Catenin) [8]. Apc mutant mice are therefore valuable to study the functions of b-Catenin and its target genes in the liver [9]. Interestingly, it was previously shown that the phenotypes of Apc loss in the intestine, but not in the liver, are rescued by additional knockout of c-Myc [8,10]. However, the b-Catenin-dependent signaling network in the liver is still incompletely understood.
Several lines of evidence suggest that the transcription factor c-Jun may be an important mediator of b-Catenin functions in the liver. c-Jun is a member of the AP-1 (activator protein 1) complex that contains dimers of either Jun (c-Jun, JunB and JunD), Fos (c-Fos, FosB, Fra-1 and Fra-2), Atf (activating transcription factor) or Maf (musculoaponeurotic fibrosarcoma) protein families [11]. c-Jun transcription is directly regulated by b-Catenin, at least in the intestine [12,13]. Moreover, c-Jun is an important regulator of liver development and regulates hepatocyte proliferation after partial hepatectomy [14,15,16,17]. c-Jun is also frequently expressed in human HCCs and acts as an oncogene in the liver [18,19]. In addition, c-Jun promotes hepatocyte survival under stress conditions such as acute hepatitis and endoplasmic reticulum (ER) stress [20,21]. We therefore addressed the functions of c-Jun in response to hepatic Apc loss using double mutant mice lacking both, Apc and c-Jun. Our findings indicate that c-Jun is not essential for hepatocyte proliferation in this setting, but rather protects against liver damage, presumably by interacting with growth factor signaling and subsequent activation of Akt .

Results
Hepatic loss of Apc results in rapid activation of b-Catenin but delayed induction of c-Jun expression Compound mutant mice carrying conditional alleles of Apc and c-Jun as well as an inducible Cre recombinase under control of the interferon-responsive Mx1 promoter were generated to analyze the functional impact of c-Jun on the immediate phenotypes of Apc loss in the liver. Mice were injected with double stranded RNA (poly(I?C)) in order to induce Cre-mediated recombination and to generate single Apc knockout (SKO) or Apc/c-Jun double knockout mice (DKO).
Since deletion of Apc leads to stabilization and subsequent nuclear translocation of b-Catenin, immunohistochemistry for b-Catenin was performed to functionally assess the Apc recombination efficiency. b-Catenin was predominantly localized on cell membranes in control livers, but translocated to the nucleus in most hepatocytes in SKO and DKO mice as early as 2 days after poly(I?C) injection (Fig. 1A, B). c-Jun expression was almost absent in control livers. In SKO livers, nuclear c-Jun expression was detected by immunohistochemistry only in a small subset of hepatocytes 2 days after loss of Apc and further increased thereafter (Fig. 1C, D). Immunoblotting confirmed that c-Jun expression in SKO livers mainly occurred 6 days after Cre induction (Fig. 1E). In contrast, c-Jun expression was absent in DKO hepatocytes ( Fig. 1.C-E) and efficient recombination was also evident by PCR analysis of the respective loci (data not shown). Quantitative PCR analysis was performed to analyze the expression kinetics of b-Catenin target genes in more detail. Nuclear translocation of b-Catenin in SKO and DKO livers at 2 days after Cre induction correlated with increased expression of established b-Catenin target genes such as Axin2, the transcription factor Sp5 and the intestinal stem cell marker Lgr5. In contrast, expression of c-Jun and c-Myc was delayed and only apparent after 6 days following Cre induction. This was also consistent with previous observations that c-Myc may not be a direct transcriptional target of b-Catenin in the liver [9,22]. In conclusion, our findings indicate that Mx1Cre mediates efficient recombination of each, the Apc and c-Jun locus. Moreover, the delayed and more restricted expression pattern of c-Jun suggests that c-Jun is not a primary target of b-Catenin in the liver following loss of Apc. c-Jun is not required for b-Catenin-dependent hepatocyte proliferation and liver zonation Loss of Apc in the liver results in increased hepatocyte proliferation, subsequent hepatomegaly and mortality within days [8]. Consistent with these findings, SKO animals died after a median of 11 days following poly(I?C) injection. Interestingly, median survival was significantly reduced to 8 days in DKO animals, while survival of control animals was not affected ( Fig. 2A). Moderate hepatomegaly was evident in all Apc mutants irrespective of c-Jun expression, indicating that this phenotype was independent of c-Jun (Fig. 2B). Hepatocyte proliferation was determined by immunohistochemistry for the proliferation marker Ki67. Interestingly, the hepatocyte proliferation index increased to the same extent in SKO and DKO livers (Fig. 3A, B), indicating that c-Jun is not an essential mediator of b-Catenin-dependent hepatocyte proliferation.
Apc and b-Catenin also regulate the spatial expression pattern of genes involved in hepatic glutamine and ammonia metabolism thereby controlling a process called liver zonation, which is profoundly disturbed upon loss of Apc [7] [23]. However, expression of the established b-Catenin target gene and zonation marker glutamine synthetase (GS, encoded by the Glul gene), as well as ornithine aminotransferase (Oat) was comparable in SKO and DKO livers as determined by qPCR (Fig. 3D). Immunohistochemistry further revealed that GS expression was confined to pericentral hepatocytes in Apc f/f control animals (Fig. 3C), although it should be noted that GS staining is more restricted to the most pericentral hepatocyte layer in Apc +/+ livers (small inset in Fig. 3C), consistent with the previous finding that conditional Apc alleles are hypomorphic [24]. In contrast, GS was similarly expressed throughout the liver in SKO and DKO mutants, suggesting that c-Jun is not essential for regulating liver zonation.
c-Jun attenuates liver damage following Apc loss c-Jun is a stress-responsive immediate early gene and its delayed expression may reflect increased cell stress occurring at later timepoints following Apc loss. Histological analyses revealed that deletion of Apc resulted in moderate hypertrophy of periportal hepatocytes in SKO mice, which was more pronounced in DKO livers. In contrast, hepatocytes within the pericentral compartment of DKO livers were considerably smaller and frequently contained vacuoles, which may reflect increased hepatic stress (Fig. 4A, white arrows). Moreover, scattered apoptotic hepatocytes were evident by H&E staining (Fig. 4A, black arrowheads at higher magnification in the lower panel). Although hepatocyte apoptosis as determined by TUNEL assay was significantly induced in DKO livers, it was a rather rare event (Fig. 4B). However, serum transaminase concentrations were strongly increased in DKO animals, reflecting increased liver damage following Apc loss (Fig. 4C). Interestingly, increased liver damage was also observed in Apc mutant mice heterozygous for c-Jun, in which c-Jun expression was reduced by 60% as determined by immunohistochemistry (data not shown). These findings indicate that c-Jun protects against liver damage following loss of Apc and that these functions are likely affected by c-Jun gene dosage.
Loss of c-Jun affects the expression of genes related to growth factor signaling and Akt phosphorylation Expression analysis of candidate genes related to cell survival revealed that expression of the hepatoprotective genes Birc5 (also known as Survivin), insulin like growth factor 1 (Igf1) and epidermal growth factor receptor (Egfr) was reduced in DKO livers, while expression of the apoptosis-related BH3 genes Bim, Bax and Bcl-2 was not affected (Fig. 5A). Importantly, Igf1 and Egfr expression was also reduced in livers lacking c-Jun specifically in hepatocytes ( Figure S1). Igf1 and other signaling cascades such as the insulin pathway converge at the serine/threonine kinase Akt, a central regulator of cell survival and metabolism. Importantly, Akt phosphorylation was strongly reduced in DKO livers (Fig. 5B). Reduced Akt phosphorylation in DKO livers was likely not mediated by insulin, since serum insulin concentrations and hepatic insulin receptor expression were reduced to the same extent in SKO and DKO mice (Fig. 5B,C). Increased hepatocyte vacuolization observed in DKO livers resembled the phenotype of sustained chemically-induced ER stress in livers lacking c-Jun [21]. In keeping with this notion, expression of the chaperone and ER stress marker BiP was strongly induced in DKO livers and predominantly occurred in vacuolated hepatocytes (Fig. 5B,D). However, expression of other ER stress marker genes including Gadd153, Gadd34 and XBP-1 splicing was not altered in DKO as compared to SKO livers (data not shown). However, BiP (also known as glucose regulated protein 78, Grp78) expression also occurs during other stress responses such as starvation. In keeping with this notion, hepatic glycogen content was specifically reduced in DKO livers (Fig. 5C). These findings suggest that the composite loss of Apc and c-Jun results in deregulated signaling through growth factors and Akt, which in turn may exacerbate liver damage in the absence of c-Jun.

Discussion
Although Apc may exert functions independent of b-Catenin [25], Apc mutant mice are an established model to study the immediate liver phenotypes and changes in gene expression following b-Catenin activation [4,7,8,9]. Here, we studied the impact of the AP-1 transcription factor c-Jun on these phenotypes using compound mutant mice lacking both Apc and c-Jun.
A key finding of our study is the observation that c-Jun is likely not a direct b-Catenin target gene and not an essential mediator of the primary phenotypes of Apc loss in the liver such as deregulated liver zonation, hepatocyte proliferation and subsequent hepatomegaly. c-Jun has previously been identified as a b-Catenin target gene in the intestine and mediates b-Catenin-dependent intestinal tumorigenesis in Apc Min mice [12,13]. However, these functions are likely dependent on the experimental system since c-Jun expression was also not required for proliferation of small intestinal epithelial cells upon knockout of Apc or an activating mutation of Ctnnb1 as well as during colitis-associated colorectal cancer [26].
By contrast, hepatic loss of Apc and c-Jun rather resulted in increased liver damage and mortality. It should be noted that these phenotypes may also be influenced by extrahepatic functions of c-  Jun, since Mx1Cre-mediated recombination also occurred in extrahepatic tissues such as bone marrow and spleen. It is therefore uncertain whether the increased mortality of DKO mice can be attributed to increased liver damage alone and the extrahepatic functions of c-Jun and causes of death following Apc loss need to be analyzed in more detail in the future. Moreover, it has been shown previously that Mx1Cre-mediated deletion of IKKb in non-parenchymal liver cells and hepatocytes results in opposite tumorigenesis phenotypes as compared to hepatocyte-specific deletion alone [27]. However, comparable results have not been reported upon Mx1Cre-mediated deletion of c-Jun, at least during chemically induced liver carcinogenesis and acute T-cell-mediated hepatitis [18,20]. On the other hand, inducible recombination of Apc and c-Jun was highly efficient by using Mx1Cre and the hepatic phenotypes observed in these mice closely resembled previous reports of hepatic Apc loss by using different recombination approaches [6,8]. Our mouse model can therefore be considered to correctly recapitulate the phenotypic hallmarks of hepatic Apc loss.
The hepatoprotective functions of c-Jun observed here are consistent with the emerging concept that c-Jun promotes hepatocyte survival during diverse stress conditions such as acute hepatitis, hepatic ER stress responses and liver tumorigenesis [18,20,21]. On the molecular level, increased liver damage in DKO livers correlated with impaired expression of genes implicated in the regulation of hepatocyte survival such as Birc5 and Egfr, which have both been reported as putative direct c-Jun and b-Catenin target genes in other experimental systems [8,28,29,30]. Moreover, composite loss of Apc and c-Jun resulted in impaired Igf1 expression, which also occurred in livers specifically lacking c-Jun in hepatocytes, suggesting that Igf1 expression is indeed regulated by c-Jun in hepatocytes.
Nutrient availability, insulin and related growth factors regulate cell survival by activation of the serine/threonine kinase Akt, which is a master regulator of cell metabolism, survival, growth and cancer and frequently activated in HCC [31,32,33]. Moreover, Akt substantially contributes to regeneration and tumorigenesis following loss of Apc in the intestine [34]. Akt phosphorylation was substantially reduced in DKO livers, suggesting that c-Jun may directly or indirectly interact with hepatoprotective Akt/mTor signaling in this context. Impaired Akt phosphorylation also correlated with reduced hepatic glycogen content and it is therefore tempting to speculate that c-Jun may be involved in the regulation of metabolic responses in the liver. In addition, Birc5, Egfr, Igf as well as downstream Akt signaling all promote liver tumorigenesis [32,35], suggesting that c-Jun may promote the formation of b-Catenin-dependent HCC by directly or indirectly regulating the expression of these genes.
In conclusion, we provide evidence that c-Jun is not a primary target of b-Catenin signaling in the liver, but rather protects against liver damage in this context. These findings provide a rationale to further explore the interactions of c-Jun with hepatoprotective Akt signaling in metabolic liver disease, hepatocarcinogenesis and tumor cell metabolism.

Animals
Mice with conditional alleles for c-Jun or Apc (c-Jun f/f and Apc f/f , respectively) were crossed with transgenic mice expressing Cre under the control of the interferon-responsive Mx1 promoter [16,36,37]. Tg(Mx1Cre) Apc f/f c-Jun +/+ and Tg(Mx1Cre) Apc f/f c-Jun f/f animals were used to generate single (SKO) and double knockout mice (DKO), respectively. Cre-mediated recombination was induced by injection of polyinosine-polycytidylic acid (poly (I?C), 15 mg/g BW, Amersham Biosciences, Piscataway, NJ) and mice were analyzed at the indicated time points. Mice were maintained on a mixed genetic background (C57BL/66129/Sv) and housed under specific-pathogen free conditions. Littermates not expressing Cre were used as controls. Floxed Apc alleles are considered to have hypomorphic functions [24]. Pilot experiments comparing Apc f/f c-Jun f/f with Apc +/+ c-Jun f/f mice revealed no differences in liver size, hepatocyte proliferation or RNA expression of the genes studied here, indicating that the phenotypes of the controls used here was not substantially affected by the floxed Apc alleles ( Figure S2). For some control experiments, mice lacking c-Jun specifically in hepatocytes (Tg(AlfpCre) c-Jun f/f , c-Jun Dli ) were used [11]. All animals received humane care and experiments were performed in accordance with local and institutional regulations. All experiments for this study were specifically approved by the local animal ethics committee (Regierungsprä sidium Freiburg, Germany, permit number G08/ 12).

Cytotoxicity assays and liver glycogen analysis
Hepatocyte damage was determined by the level of serum transaminases and serum glucose concentrations were analyzed using semi-automated clinical routine methods. Liver glycogen concentrations were analyzed using a glycogen assay kit (BioCat, Heidelberg, Germany) according to the manufacturer's protocol.

Western blot analysis and ELISA
Total liver lysates were analyzed by immunoblot using antibodies for b-Actin (Sigma), insulin receptor (sc711, Santa Cruz), BiP (#3177, Cell Signaling), p-Akt (Ser473 #4060, Cell Signaling) and Akt (# 4685, Cell Signaling). Serum insulin concentrations were determined using the insulin (rat) ultrasensitive ELISA kit with 75% mouse sensitivity (EIA-2943; DRG Instruments, Marburg, Germany). qPCR Upon isolation of total liver RNA using Qiazol (Qiagen, Hilden, Germany), cDNA synthesis was performed using the first strand cDNA synthesis kit (Fermentas, St. Leon-Rot, Germany). qPCR was performed with SYBR Green (Invitrogen, Karlsruhe, Germany) on a 480 Lightcycler (Roche) and the expression levels of transcripts were calculated with the comparative CT method. The individual RNA levels were normalized for Hprt. Primer sequences are available upon request.

Statistics
Data in bar graphs represent mean 6 S.D. as indicated. Statistical analysis was performed using the nondirectional twotailed Student's t test. Statistical analysis of mouse survival was determined by log rank test.