The transcription factor c-Jun protects against sustained hepatic endoplasmic reticulum stress thereby promoting hepatocyte survival

Authors


  • Potential conflict of interest: Nothing to report.

Abstract

Endoplasmic reticulum (ER) stress due to accumulation of hepatoviral or misfolded proteins is increasingly recognized as an important step in the pathogenesis of inflammatory, toxic, and metabolic liver diseases. ER stress results in the activation of several intracellular signaling pathways including Jun N-terminal kinase (JNK). The AP-1 (activating protein 1) transcription factor c-Jun is a prototypic JNK target and important regulator of hepatocyte survival, proliferation, and liver tumorigenesis. Because the functions of c-Jun during the ER stress response are poorly understood, we addressed this issue in primary hepatocytes and livers of hepatocyte-specific c-Jun knockout mice. ER stress was induced pharmacologically in vitro and in vivo and resulted in a rapid and robust induction of c-Jun protein expression. Interestingly, ER-stressed hepatocytes lacking c-Jun displayed massive cytoplasmic vacuolization due to ER distension. This phenotype correlated with exacerbated and sustained activation of canonical ER stress signaling pathways. Moreover, sustained ER stress in hepatocytes lacking c-Jun resulted in increased cell damage and apoptosis. ER stress is also a strong inducer of macroautophagy, a cell-protective mechanism of self-degradation of cytoplasmic components and organelles. Interestingly, autophagosome numbers in response to ER stress were reduced in hepatocytes lacking c-Jun. To further validate these findings, macroautophagy was inhibited chemically in ER-stressed wildtype hepatocytes, which resulted in cytoplasmic vacuolization and increased cell damage closely resembling the phenotypes observed in c-Jun-deficient cells. Conclusion: Our findings indicate that c-Jun protects hepatocytes against excessive activation of the ER stress response and subsequent cell death and provide evidence that c-Jun functionally links ER stress responses and macroautophagy. (HEPATOLOGY 2012)

The endoplasmic reticulum (ER) is a large eukaryotic cell organelle involved in the synthesis of proteins and lipids, posttranslational protein modification, calcium storage, and drug metabolism.1 Cells encountering a high burden of protein synthesis such as hepatocytes depend on efficient mechanisms to enable proper protein folding in order to cope with the protein load within the ER. However, ER homeostasis is profoundly disturbed in various liver diseases, including toxic, metabolic, or infectious liver disease. Under these conditions, defective posttranslational protein modifications and accumulation of misfolded or viral proteins all contribute to severe ER stress.2 ER stress activates the unfolded protein response (UPR), a network of conserved intracellular signaling pathways aimed at decreasing the functional protein load by translational attenuation, increased chaperone expression, and protein degradation.1 On the molecular level, binding of the chaperone BiP (GRP78) to misfolded proteins activates the proximal UPR transducer proteins PERK (PKR-like ER-localized kinase), IRE1α (inositol requiring enzyme 1α), and ATF6 (activating transcription factor 6).1 PERK phosphorylates eIF2α (eukaryotic initiation factor 2), thereby inhibiting global protein translation but promoting the expression of the transcription factor ATF4. IRE1α is a dual function protein with endoribonuclease and protein kinase activities and required for splicing and activation of XBP1 (X-box binding protein 1). These ER stress-responsive transcription factors ATF4, ATF6, and XBP1 then cooperate to regulate the expression of various ER stress-responsive genes such as the proapoptotic transcription factor CHOP (C/EBP homologous protein, also known as GADD153) or chaperones including BiP and GRP94.2

Adaptation to ER stress occurs if sufficient amounts of chaperones have been synthesized to bind all misfolded proteins.3 Moreover, ER stress is a potent trigger of macroautophagy (hereafter referred to as autophagy), a conserved catabolic pathway by which cells self-digest intracellular components such as expanded and disorganized ER and the protein cargo therein.4 Autophagy is considered a major cytoprotective mechanism and, among others, tightly regulated by nutrient availability and signaling through insulin, PI(3) kinase, and mTOR (mammalian target of rapamycin).5, 6

However, in the case of unresolved and sustained ER stress, the UPR may also trigger apoptosis mainly through the action of CHOP and IRE1α.7 Prolonged activation of CHOP has been linked to increased oxidative stress and activation of the proapoptotic BH3-only protein Bim.8, 9 In addition, the kinase domain of IRE1α binds to the adaptor molecule TRAF2 and activates mitogen-activated protein (MAP) kinase signaling pathways including ASK1 (apoptosis signal-regulating kinase 1) and JNK (Jun N-terminal kinase), which have both been implicated in ER stress-related apoptosis.7, 10 Besides its various functions in regulating hepatocyte survival and proliferation, JNK is an important regulator of insulin resistance and liver steatosis,11 conditions closely related to hepatic ER stress.12 Interestingly, it was also shown that JNK may promote cell survival during ER stress by inducing autophagy.13 Therefore, the biologic significance of JNK activation and of its putative targets during ER stress requires further analyses.1

AP-1 (activating protein 1) is a transcription factor complex that contains dimers of either Jun (c-Jun, JunB, and JunD), Fos (c-Fos, FosB, FRA-1, and FRA-2), ATF, or MAF (musculoaponeurotic fibrosarcoma) protein families.14 c-Jun is a central JNK target and strongly expressed in the livers of patients with acute hepatitis.15 Moreover, genetic analyses in mice have established essential functions of c-Jun in liver development, hepatocyte proliferation, and liver carcinogenesis.16-18 c-Jun is also an important regulator of hepatocyte survival during acute T-cell-mediated hepatitis.15 Interestingly, c-Jun expression is strongly induced following sustained ER stress in mice lacking either ATF6, IRE1α, or functional eIF2α,12 indicating that its expression during ER stress is not exclusively controlled by the IRE1α/JNK axis. However, the functional impact of c-Jun expression during ER stress is only poorly defined. To address this issue, we studied the functions of c-Jun following hepatic ER stress in vitro and in vivo using hepatocyte-specific c-Jun knockout mice. Our findings indicate that c-Jun protects against sustained ER stress and apoptosis in the liver by linking ER stress and autophagy.

Abbreviations

ALT, alanine aminotransferase; AP-1, activating protein 1; AST, aspartate aminotransferase; BW, body weight; ER, endoplasmic reticulum; HBV, hepatitis B virus; HCV, hepatitis C virus; LDH, lactate dehydrogenase; 3MA, 3-methyladenine; PMH, primary mouse hepatocytes; SERCA, sarcoplasmic/endoplasmic reticulum Ca2+-ATPase; TG, thapsigargin; TU, tunicamycin; UPR, unfolded protein response.

Materials and Methods

Cell Culture.

Huh-7 and HepG2 human hepatoma cells were cultured in Dulbecco's modified Eagle medium (DMEM) with 8% serum. Culture conditions of human osteosarcoma UHCVcon-57.3 cells expressing the hepatitis C virus (HCV) polyprotein and of hepatitis B virus (HBV)-expressing Huh-7.93 cells and their parental cell line TA61.1 have been described.19, 20 Primary mouse hepatocytes (PMH) were obtained by liver perfusion with collagenase (CLS2, Biochrom, Berlin, Germany), plated on collagen, and cultured in Williams' medium containing 10% serum, insulin (1 μM, Sigma, Schnelldorf, Germany), dexamethasone (100 nM, Sigma), glutamine (5 mM, Biochrom), and penicillin/streptomycin. Medium was replaced against insulin-free medium 4 hours after plating. PMH viability was determined by Trypan blue staining and exceeded 75%-80%.

Animals.

Mice with conditional alleles of c-Jun (c-Junf/f) were crossed with transgenic Alfp-Cre mice to obtain animals with hepatocyte-specific knockout of c-Jun (c-JunΔli).21 Mice were bred on a mixed genetic background (C57BL/6 × 129/Sv × FVB/N) and housed under specific pathogen-free conditions. Either c-Junf/f or AlfpCre c-Junf/+ littermates were used as controls. Transgenic mice expressing a GFP-LC3 fusion protein22 were bred on a C57BL/6 background. All animals received humane care and experiments were performed in accordance with local and institutional regulations.

ER Stress Models.

ER stress was induced by incubation of cells with either thapsigargin (TG, 1-2 μM, Enzo Life Sciences, Lörrach, Germany) or tunicamycin (TU, 1.5 μg/mL, Sigma) in vitro or in vivo upon intraperitoneal injection of TU dissolved in 150 mM dextrose (1 μg/g body weight [BW]).23

Cytotoxicity Assays.

Hepatocyte damage was determined by analysis of either lactate dehydrogenase (LDH) or transaminases alanine aminotransferase (ALT) and aspartate aminotransferase (AST) in cell culture supernatant or mouse sera using semiautomated clinical routine methods.

Histology, Immunohistochemistry, and Immunofluorescence.

For histology, livers were fixed in 3.7% neutral buffered formaldehyde at 4°C and embedded in paraffin. Immunohistochemistry was performed using the Envision kit (Dako, Hamburg, Germany) and antibodies for c-Jun (#9165; Cell Signaling, NEB, Frankfurt, Germany) and ATF6α (H-280, Santa Cruz, Heidelberg, Germany). Immunofluorescence of PMH was performed with two independent antibodies for LC3B (ab87349-100, Abcam, Cambridge, UK, and #3868, Cell Signaling) which yielded similar results.

Transmission Electron Microscopy.

PMH were grown on 12 mm Aclar plastic discs and fixed for 1 hour in 2.5% glutaraldehyde in 0.1 mol/l sodium phosphate buffer, pH 7.4. The samples were then rinsed with the same buffer, postfixed in 1%-2% osmium tetroxide in phosphate-buffered saline (PBS), washed in ddH2O, dehydrated in a graded series of ethanol, and embedded in Agar 100 resin. The 70-nm sections (nominal thickness) were cut parallel to the substrate, poststained with uranyl acetate and lead citrate, and examined with an FEI Morgagni 268D (FEI, Eindhoven, The Netherlands) operated at 80 kV. Images were acquired using an 11 megapixel Morada CCD camera (Olympus-SIS).

Quantitative Polymerase Chain Reaction (qPCR).

Upon isolation of total RNA using Qiazol (Qiagen, Hilden, Germany), complementary DNA (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, Mannheim, Germany). Loading was normalized to Hprt, Actin messenger RNA (mRNA) and 18S rRNA. Primer sequences are available upon request.

Western Blot Analysis.

Hepatocyte and total liver lysates were analyzed by immunoblot using antibodies for NS5A (kindly provided by D. Moradpour), BiP (#3177, Cell Signaling), c-Jun (#9165; Cell Signaling), phospho-c-Jun (Ser63, #9261, Cell Signaling), JNK (#9258, Cell Signaling), phospho-JNK (pT183/pY185 #700031, Invitrogen and #9251, Cell Signaling), CHOP (#2895, Cell Signaling and MA1-250, Thermo, Dreieich, Germany), actin (Sigma), cleaved caspase 3, 7 and 12 (#9661, #9491, #2002, all Cell Signaling), and LC3B (ab87349-100, Abcam).

Statistics.

Data in bar graphs represent mean ± standard deviation (SD). Statistical analysis was performed by using the nonparametric Mann-Whitney test or nondirectional two-tailed Student's t test as appropriate.

Results

c-Jun Expression Is Induced upon Acute ER Stress In Vitro and In Vivo.

Acute and chronic ER stress induces the UPR that is associated with activation of the MAP kinase JNK. Because JNK and its prototypic target c-Jun are important regulators of hepatocyte fate, we addressed the question of whether expression of c-Jun is induced during the UPR in the liver. Human HepG2 and HuH-7 hepatoma cells were incubated with thapsigargin (TG), a commonly used inducer of ER stress that inhibits the endoplasmic Ca2+-ATPase SERCA. Expression of c-Jun RNA and protein was increased 4 hours after TG incubation and remained elevated thereafter (Fig. 1A,B). Accumulation of hepatoviral proteins is another important cause of ER stress in the liver and was therefore studied in respective human cell culture models. c-Jun expression as well as expression of the chaperone and ER stress marker BiP was indeed induced in two cell lines stably expressing infectious HBV (HuH-7.93) or the entire HCV polyprotein (UHCVcon-57.3; Fig. 1A,C and data not shown).

Figure 1.

c-Jun expression is induced upon acute ER stress in vitro and in vivo. (A) c-Jun mRNA expression was determined by qPCR in human HepG2 and HuH-7 hepatoma cells after incubation with TG as well as in HuH-7.93 cells stably expressing HBV and their parental cell line TA61.1; *P ≤ 0.05; §P ≤ 0.01; (n ≥ 4/genotype). (B) c-Jun expression in TG treated Huh-7 cells at the indicated timepoints. BiP is shown as a marker of ER stress. Actin was used as loading control. (C) Immunoblot of c-Jun and BiP in human UHCVcon-57.3 cells inducibly expressing the HCV polyprotein in a tetracyclin (Tet)-regulated manner as determined by NS5A expression. (D) Immunoblot of PMH with the indicated genotypes after administration of TG or TU for 5 hours. (E) Hepatic c-Jun expression was determined by immunohistochemistry in untreated and TU treated control mice (1 μg/g BW, 8 hours).

PMH from control mice (c-Junf/f) and mice with a hepatocyte-specific c-Jun knockout (AlfpCre c-Junf/f, i.e., c-JunΔli) were used to further analyze the functional impact of c-Jun on the ER stress response. Unlike in human hepatoma cells, c-Jun mRNA expression was not significantly altered in ER-stressed PMH (data not shown). However, a robust induction of c-Jun protein expression and phosphorylation as well as JNK activation was apparent in control PMH upon incubation with TG or TU, an inhibitor of N-terminal protein glycosylation, whereas there was no c-Jun expression in c-JunΔli cells (Fig. 1D; Supporting Fig. 1). These findings suggest that ER stress-dependent c-Jun expression in mice is mainly regulated by a posttranscriptional mechanism. Moreover, c-Jun protein expression was strongly induced in the liver after induction of ER stress in vivo by intraperitoneal injection of TU (Fig. 1E). These findings indicate that c-Jun expression is induced following various types of ER stress in the liver.

Profound ER Distension in the Absence of c-Jun.

The ER stress-related phenotypes of c-Jun-deficient PMH were then analyzed in more detail. Untreated PMH isolated from c-JunΔli mice did not display gross differences in cell morphology compared to controls (Fig. 2A, small inserts left). Control PMH displayed moderate shrinkage with retracted cytoplasm upon incubation with TG for 24 hours, whereas massive cytoplasmic vacuolization occurred in PMH lacking c-Jun (Fig. 2A). Dysregulated ER stress responses are linked to hepatic steatosis; however, the vacuoles observed in c-Jun-deficient PMH did not contain lipids as determined by Oil Red O staining (Fig. 2A, right insert). Interestingly, profound hepatocyte vacuolization and ballooning was also observed in c-JunΔli livers in vivo upon induction of ER stress by TU (Fig. 2B), which also correlated with mildly increased steatosis (Supporting Fig. 2).

Figure 2.

ER stress causes massive cytoplasmic vacuolization in hepatocytes lacking c-Jun. (A) Phase contrast images of untreated (small inserts left) and TG-treated (2 μM, 24 hours) PMH with the indicated genotypes. Some vesicles containing lipids as determined by Oil red O staining are indicated by white arrowheads (right insert). (B) Liver histology of mice with the indicated genotypes after treatment with TU (1 μg/g for 72 hours) was determined by hematoxylin and eosin (H&E) staining.

Transmission electron microscopy was used to analyze the PMH vacuolization phenotype in more detail. Again, no gross phenotypic differences were apparent in untreated PMH of either genotype, whereas distension of the nuclear envelope and the ER was observed after incubation with TG. These alterations were most pronounced in the absence of c-Jun (Fig. 3). Many vacuolar membranes were lined by ribosomes, suggesting that these structures were indeed derived from the ER (Fig. 3, lower panel) and vacuoles contained electron-dense material most likely corresponding to misfolded protein aggregates. Therefore, ER stress in the absence of c-Jun resulted in profound cytoplasmic vacuolization due to ER distension.

Figure 3.

Cytoplasmic vacuolization in the absence of c-Jun is due to ER expansion. Transmission electron microscopy was used to study the morphology of untreated and TG-treated PMH (2 μM for 24 hours). Ribosomes are indicated with black arrowheads. N, nucleus; M, mitochondrium.

Exacerbated and Sustained ER Stress in the Absence of c-Jun.

Expansion of the ER is, among others, controlled by the transcription factors XBP1 and ATF6, raising the question whether ER distension in c-JunΔli hepatocytes correlated with increased ER stress. A strong increase in XBP1 splicing and expression of CHOP and BiP was observed in control PMH after incubation with TG for 4 hours, which was fully reversible after 24 hours (Fig. 4A), consistent with previous findings in mouse embryonic fibroblasts.24 In contrast, expression of these ER stress-related genes was strikingly induced in c-JunΔli PMH after incubation with TG for 4 hours and sustained at later timepoints (Fig. 4A). Consistently, expression of other ER stress-related genes such as ATF4, the CHOP target genes GADD34 and ERO1L (ER oxidase 1), the chaperone GRP94, and genes involved in ER-associated protein degradation such as EDEM was increased in the absence of c-Jun (Fig. 4B). Increased XBP1 splicing and BiP expression was also evident in c-JunΔli PMH after incubation with TU in vitro (data not shown). TU-mediated activation of the three canonical UPR branches also occurred in vivo in c-JunΔli livers, in which IRE1 activity as determined by XBP1 splicing was profoundly induced at early and late timepoints (Fig. 4C). TU treatment also resulted in increased phosphorylation of JNK1 and JNK2, which was not profoundly affected by loss of c-Jun, suggesting that JNK is not hyperactivated in c-JunΔli livers (Supporting Fig. 3). Moreover, more prominent nuclear translocation of ATF6 (Fig. 4D) and induction of BiP, GRP54, GRP94, GRP170, CHOP, GADD34, calreticulin, and calnexin was evident in these livers, particularly at later timepoints (Fig. 4C,E). These findings indicate that ER stress is exacerbated and sustained in the absence of c-Jun.

Figure 4.

Exacerbated and sustained ER stress in the absence of c-Jun. (A) Relative expression of ER stress-related genes in PMH at the indicated timepoints after administration of TG (1 μM) was determined by qPCR. Expression was normalized to the expression in untreated control cells; *P ≤ 0.05; §P ≤ 0.01; (n = 6-20/genotype and timepoint). (B) Relative expression of the indicated genes in hepatocytes 4 hours after TG administration (1 μM). Expression was normalized to the expression in untreated control cells; *P ≤ 0.05; §P ≤ 0.01; (n ≥ 5/genotype). (C) Relative hepatic expression of the indicated genes in mice treated with TU. Expression was normalized to untreated control livers. *P ≤ 0.05; (n = 3-6/genotype and timepoint). (D) Active (i.e., nuclear) ATF6α expression was determined by immunohistochemistry of livers from mice with the indicated genotypes upon TU treatment for 8 hours. Some positive nuclei are indicated with arrowheads. (E) Immunoblot analysis of BiP, CHOP, phosphorylated c-Jun, and c-Jun at the indicated timepoints following TU treatment. CHOP is indicated by an arrow.

c-Jun Promotes Hepatocyte Survival During ER Stress.

The inability of the UPR to reestablish ER homeostasis results in sustained activation of the ER stress response, which subsequently triggers cell death. Incubation of control PMH with relatively high TG concentrations only resulted in moderate hepatocyte damage as determined by LDH release to the cell culture medium (Fig. 5A). By comparison, increased cell death as determined by morphological criteria was evident in c-JunΔli PMH (Supporting Fig. 4). This phenotype correlated with increased LDH release and cleavage of caspases 3, 7, and 12 in PMH lacking c-Jun (Fig. 5A,B), in which expression of the chemokines CXCL1 and CXCL2 was also induced, most likely reflecting increased hepatocyte damage (Fig. 5C). Caspase cleavage in c-JunΔli PMH also correlated with a higher expression of the proapoptotic BH3 protein Bim and C/EBPβ, two important regulators of ER-stress-related apoptosis (Fig. 5C).9 Although incubation of PMH with TU in vitro did not cause major LDH release or caspase cleavage in either genotype due to the fact that the kinetics and severity of ER stress was more limited in this model, serum ALT and AST concentrations were strikingly increased in c-JunΔli animals after TU injection in vivo (Fig. 5D). Sustained ER stress has previously been linked to oxidative cell damage. Expression of several genes related to oxidative stress such as ATF3, HO1 (Heme oxygenase-1), SESN2 (Sestrin2), as well as NOX2 (NADPH-oxidase 2) was consistently induced in livers lacking c-Jun after TU treatment (Fig. 5E). Moreover, increased expression of Bim and Bax was observed in these livers (Fig. 5E). These findings indicate that c-Jun protects against ER stress-related hepatocyte damage, oxidative stress, and subsequent cell death.

Figure 5.

c-Jun protects against ER stress-related hepatocyte death in vitro and in vivo. (A) Cell damage of PMH with the indicated genotypes after incubation with different TG concentrations for 48 hours was assessed by LDH release into the supernatant. §P ≤ 0.01; (n ≥ 7/genotype). (B) Cleavage of caspases 3, 7, and 12 in PMH after incubation with TG for 24 hours was determined by immunoblot analysis. Cleaved caspase 12 is indicated by an arrow. (C) Relative expression of the indicated genes in PMH 4 hours after TG administration. Expression was normalized to untreated control cells. *P ≤ 0.05; §P ≤ 0.01 (n ≥ 5/genotype). (D) Serum transaminases of mice with the indicated genotypes treated with TU for 72 hours. §P ≤ 0.01. (n = 7/genotype). (E) Relative expression of the indicated genes in livers at 48 hours after TU treatment. Expression was normalized to the expression in untreated control livers. *P ≤ 0.05; §P ≤ 0.01 (n = 3-6/genotype).

c-Jun Links ER Stress and Autophagy to Promote Hepatocyte Survival.

ER stress is a potent inducer of autophagy and it has been shown previously that IRE1-dependent activation of JNK links ER stress and autophagy in tumor cells and mouse embryonic fibroblasts, thereby promoting cell survival.13 Autophagosome formation is associated with posttranslational modification of the microtubule associated protein LC3. PMH isolated from transgenic mice expressing GFP-labeled LC3 were used in pilot experiments to study the kinetics of ER stress-related autophagosome formation. Numbers of GFP-LC3 labeled puncta corresponding to autophagosomes were increased after 8 hours and most abundant after 24 hours of TG incubation and disappeared thereafter (data not shown). Consistently, a marked increase in LC3-positive puncta as determined by immunofluorescence was apparent in control PMH 24 hours after administration of TG (Fig. 6A,B). This increase was reversible by coadministration of the PI3 kinase inhibitor 3-methyladenine (3MA), a widely used inhibitor of autophagy. Importantly, the ER stress-mediated increase in autophagosome numbers was also absent in PMH lacking c-Jun (Fig. 6A,B), whereas LC3 conversion was not affected as determined by immunoblot analysis (Fig. 6C). This suggests that c-Jun may function at the phagophore expansion rather than at the induction step, which resembles the phenotype observed in ER-stressed human HEK293 cells upon knockdown of BiP.6, 25

Figure 6.

c-Jun links ER stress and autophagy to promote hepatocyte survival. (A) Autophagy was analyzed by LC3B immunofluorescence (green) in TG-treated PMH of the indicated genotypes (1 μM for 24 hours) and after administration of the PI(3) kinase inhibitor 3-methyladenine (3MA, 5 mM) an inhibitor of autophagy. Nuclei were counterstained with DAPI (blue). (B) Quantification of LC3B positive puncta/hepatocyte in one representative experiment out of three. *P ≤ 0.05; §P ≤ 0.01. (C) Expression of LC3B at 8 hours after TG was analyzed by immunoblotting. (D) Transaminase and LDH release of c-Junf/f PMH after incubation with TG (2 μM) and 3MA (5 mM) for 24 hours. §P ≤ 0.01. (n = 4). (E) Phase contrast images of c-Junf/f PMH incubated with TG and either JNK inhibitor SP600125 (10 μM), 3MA (5 mM), or chloroquine (CLQ, 10 μM), all resulting in profound cytoplasmic vacuolization reminiscent of the phenotype in c-Jun deficient PMH.

To further validate these findings, control c-Junf/f PMH were incubated with 3MA and TG, which resulted in similar hepatocyte damage as determined by release of LDH and AST as compared to TG-treated c-JunΔli PMH (Figs. 5A, 6D, and data not shown). Moreover, cytoplasmic vacuolization reminiscent of TG-treated c-JunΔli cells was also observed in c-Junf/f PMH after administration of TG and the JNK inhibitor SP600125 as well as after inhibition of autophagy by 3MA or chloroquine (Fig. 6E). These findings provide evidence that c-Jun links ER stress and autophagy thereby promoting hepatocyte survival.

Discussion

ER stress is increasingly recognized as an important step in the pathogenesis of various liver diseases.2 Sustained ER stress results in liver steatosis, which is attenuated by expression of BiP.12, 26 Another mechanistic link between metabolic liver disease and ER stress was provided by the recent finding that aberrant hepatic lipid metabolism in obese mice disrupts calcium homeostasis by inhibition of SERCA and causes severe ER stress.27 Pharmacological inhibition of SERCA by TG is therefore a highly relevant model to study ER stress responses in hepatocytes.

JNK is an important regulator of insulin resistance and liver steatosis11 and activated upon ER stress and subsequent IRE1 signaling.1 However, data concerning the functions of JNK and its targets in the ER stress response are controversial because JNK activation has been linked to both ER stress-dependent apoptosis and cell survival.13, 28

It was previously shown that TU injection into mice lacking either ATF6, IRE1α, or functional eIF2a results in sustained hepatic ER stress and prominent expression of c-Jun.12 Here we provide genetic evidence that c-Jun is an important negative regulator of ER stress responses in the liver, because ER stress was exacerbated and sustained and resulted in increased cell death in hepatocytes lacking c-Jun.

These findings are consistent with previous observations that c-Jun promotes hepatocyte survival during acute hepatitis by regulating transcription of inducible nitric oxide synthase (NOS2).15 Moreover, it has recently been shown that c-Jun protects against TG-mediated cell death of mouse embryonic fibroblasts by regulating transcription of the calcineurin inhibitor RCAN1.29 However, hepatic expression of NOS2 and RCAN1 was not substantially altered in c-Jun-deficient hepatocytes, suggesting a distinct molecular mechanism of c-Jun action.

Sustained expression of CHOP and its target gene GADD34 is considered an important switch between cell adaptation and death in response to ER stress.30 CHOP regulates the expression of several proapoptotic genes such as Bim.9 TU-mediated cell death of renal epithelial cells was consistently rescued in either CHOP−/−, GADD34−/−, or Bim−/− mice.8, 9 Our observation that expression of CHOP, GADD34, and Bim was increased and sustained in ER stressed PMH and livers lacking c-Jun is therefore consistent with CHOP-dependent cell death. Moreover, CHOP has been shown to inhibit Bcl-2 and to perturb the cellular redox state, likely by regulating the expression of ERO1 and NOX2.8, 31, 32 Expression of these and other oxidative stress-related genes including ATF3, HO1, and SESN2 was induced in c-Jun deficient livers, consistent with increased oxidative damage. CHOP transcription is mainly controlled by ATF4. However, CHOP is also an established AP-1 target gene and its expression in pancreatic beta cell-derived INS-1E cells is directly repressed by JunB.33 Increased CHOP expression in c-Jun-deficient hepatocytes may therefore reflect either a response to exacerbated ER stress or a specific inhibitory interaction of c-Jun with the CHOP promoter.

Sustained ER stress in c-Jun-deficient hepatocytes coincided with profound cytoplasmic vacuolization due to massive ER distension. A similar phenotype has been observed following sustained ER stress in several tumor cell lines.34, 35 Degradation of disorganized ER and the protein aggregates therein is mainly mediated by autophagy. Indeed, various studies identified ER stress as a strong inducer of autophagy either by the action of IRE1/JNK or calcium-dependent activation of PKCθ (protein kinase C-theta).13, 36 Inhibition of autophagy consistently resulted in increased cell death of ER stressed mouse embryonic fibroblasts and immortalized hepatocytes.13, 36 Autophagy deficient mice are prone to obesity, exacerbated ER stress in the liver, and insulin resistance.37 Moreover, autophagy was shown to protect against oxidative stress in a rat hepatocyte cell line.38

The profound vacuolization of c-JunΔli hepatocytes was paralleled by reduced numbers of LC3-stained autophagosomes. Interestingly, LC3 conversion in c-Jun-deficient PMH as determined by immunoblot analysis was not affected at earlier timepoints. This finding resembles the phenotype observed in HEK293 cells after knockdown of BiP and suggests that c-Jun may regulate autophagy at the phagophore expansion rather than induction step.6, 25 Massive cytoplasmic vacuolization was also observed in ER stressed wildtype PMH upon inhibition of JNK, consistent with the previous finding that JNK promotes autophagy in response to ER stress.13 Autophagy and apoptosis are mutually activated and related in a complex functional manner.39 It is therefore possible that autophagy in TG-treated hepatocytes lacking c-Jun was impaired due to ongoing apoptosis. However, pharmacological inhibition of autophagy in c-Junf/f PMH resulted in increased cell damage and hepatocyte vacuolization reminiscent of c-Jun mutant cells, strongly suggesting that the interaction of c-Jun with the autophagy machinery is indeed an important regulator of hepatocyte cell fate.

In conclusion, our findings indicate that c-Jun protects hepatocytes against sustained ER stress, thereby promoting cell survival (Fig. 7A). In the absence of c-Jun, exacerbated ER stress and subsequent activation of CHOP, GADD34, and Bim as well as impaired autophagy all contribute to increased cell death (Fig. 7B). These protective functions of c-Jun are very likely to have a strong impact on hepatocyte fate during ER stress-related liver diseases given the broad expression of c-Jun under these conditions. Moreover, these functions may also apply to other cell types in which AP-1 was shown to regulate ER stress responses such as pancreatic beta cells33 and should be carefully considered when targeting the JNK/c-Jun axis therapeutically.

Figure 7.

Schematic working model. (A) c-Jun protects against sustained ER stress and consecutive CHOP expression and promotes hepatocyte survival by inducing autophagy. (B) Upon loss of c-Jun, ER stress is exacerbated and cell death occurs due to impaired autophagy as well as sustained expression of CHOP and its proapoptotic target genes.

Acknowledgements

The authors thank N. Mizushima, D. Moradpour, M. Nassal, and E.F. Wagner for providing mice, cell lines, and antibodies, the animal house staff at Freiburg University Hospital for support, and L. Bakiri, L. Hui, R. Eferl, and C. Thoma for helpful comments and discussions.

Ancillary