Cytokines increase CRE binding but decrease CRE-mediated reporter activity in rat hepatocytes by increasing c-Jun



The cyclic AMP response element (CRE) has been implicated in the regulation of the expression of many genes and cellular processes important in hepatocyte function. CRE sites exist in the promoter regions of several genes expressed during inflammation. Numerous studies on the role of CRE in hepatocyte gene expression have been performed in resting hepatocytes, but the role of CRE during inflammation is unknown. To evaluate the regulation of CRE-mediated transcription during sepsis, cultured hepatocytes were exposed to proinflammatory cytokines and lipopolysaccharide (LPS) was injected into rats. Nuclear proteins were collected and CRE binding activity measured by electromobility shift assay (EMSA) using a consensus CRE oligonucleotide. CRE binding activity was increased in vitro by cytokines and in vivo by LPS administration but CRE-dependent reporter activity was decreased by cytokine stimulation. A c-jun N-terminal kinase (JNK) inhibitor reversed the cytokine-induced increase in CRE binding and increased CRE-dependent reporter activity. Supershift assays indicated that cyclic AMP response element binding protein (CREB) and c-Jun proteins were included in the CRE binding complex. CREB induced and c-Jun suppressed reporter activity using a CRE-dependent construct transfected into cultured primary hepatocytes. In conclusion, these data demonstrate that proinflammatory cytokines regulate CRE binding and activity in cultured hepatocytes and suggest that sepsis-induced changes in CRE binding may participate in the cellular response to inflammation. (HEPATOLOGY 2004;39:1343–1352.)

The cAMP response element (CRE) transcription factor complex is a pleiotropic activator that regulates the expression of many genes important in hepatocyte metabolism, such as somatostatin, glycoprotein hormone, tyrosine amino-transferase,1 and phosphoenolpyruvate carboxykinase (PEPCK).2–4 Studies demonstrate that CRE-mediated regulation of hepatocyte gene expression may be quite complex. With the exception of cAMP and selected hormones, the mechanisms that regulate CRE activation and its effects on hepatocyte gene expression have been incompletely characterized.

Septic shock is a major cause of death in critically ill patients, and proinflammatory cytokines play a central role in the pathophysiological sequelae of sepsis.5 One of the pathophysiological consequences of severe sepsis in animal models and human patients is an alteration in glucose metabolism.6, 7 PEPCK expression is regulated by a CRE motif in its promoter, and gluconeogenesis is altered by septic stimuli. These findings suggest that CRE-mediated gene regulation may also be altered by sepsis. Studies of CRE-mediated gene expression have been performed in resting, unstimulated hepatocytes,8 but the role of CRE in hepatocyte gene expression during inflammation and sepsis is unknown. To evaluate the function of CRE during sepsis, we exposed cultured hepatocytes to proinflammatory cytokines as an in vitro model of sepsis and treated rats with lipopolysaccharide (LPS) in vivo to evaluate the effects of cytokines on CRE-dependent gene expression.


CRE, cyclic AMP response element; LPS, lipopolysaccharide; EMSA, electrophoretic mobility shift assay; JNK, c-jun N-terminal kinase; CREB, cyclic AMP response element binding protein; PEPCK, phosphoenolpyruvate; IL-1β, interlukine-1 beta; TNFα, tumor necrosis factor alpha; IFNγ, interferon gamma; AP-1, activator protein 1; CM, cytokine mix.

Experimental Procedures

Reagents and Plasmids.

The following agents were used in this study: human recombinant interleukin (IL)-1β (DuPont, Boston, MA), murine recombinant tumor necrosis factor (TNF)α (Genzyme, Cambridge, MA), and murine recombinant interferon (IFN)γ (Invitrogen Life Technologies, Carlsbad, CA). The polyclonal antibody of phospho-cyclic AMP response element binding protein (CREB) was purchased from Upstate Biotechnology (Lake Placid, NY) and polyclonal antibodies against c-Jun, p-c-Jun(KM-1), JunB, JunD, c-fos, fosB, CREM, ATF-1, ATF-2, ATF-3, and ATF-4 were purchased from Santa Cruz Biotechnology Inc.(Santa Cruz, CA). Cell permeable protein kinase A inhibitor 14-22 amide (PKAI) and c-jun N-terminal kinase (JNK) inhibitor, SP600125, were purchased from Calbiochem-Novabiochem Corporation (San Diego, CA). All other chemicals were purchased from Sigma (Sigma Chemical Inc., St. Louis, MO). pcDNA3.1-c-Jun was the kind gift from Dr. Michael J. Birrer MD, PhD, at the National Cancer Institute.9 pCRE-luc and pLuc-MCS were purchased from Stratagene. pRC/RSV-CREBDIEDML is a mutant type of CREB expression plasmid in which Ser 133 region of wild-type CREB was replaced with DIEDML, which constitutively interacts with CREB binding protein (CBP).10, 11

Cell Treatment and Animal Surgical Procedure.

Male Sprague-Dawley rats weighing 200 to 250 g (Harlan Sprague-Dawley, Madison, WI) were used in all experiments. All animal care was in accordance with the University of Pittsburgh's Animal Care and Use committee and followed guidelines prescribed by the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals. Rat primary hepatocytes were isolated as previously described12 using a modification of the in situ collagenase (type IV, Sigma Chemical Co., St. Louis, MO) perfusion technique of Seglen.13 Hepatocyte purity assessed by microscopy was typically greater than 98% and viability consistently exceeded 95% by Trypan blue exclusion.

Hepatocytes (106/cc) were plated onto 100-mm gelatin-coated Petri-dishes (Corning Co., Corning, NY) with 5 cc of Williams E Media (Life Technologies, Grand Island, NY) with L-arginine (0.50 mM), insulin (10–6 M), HEPES (15 mM), L-glutamine, penicillin, streptomycin, and 10% low endotoxin calf serum (CS, HyClone Laboratories, Logan, UT). After 16 hours of incubation, medium was changed to fresh insulin-free medium with 5% CS and experimental conditions were established. After stimulation, the hepatocytes were harvested at the indicated time points for protein, RNA, or nuclear extract preparation. Hepatocytes were plated in duplicate or triplicate culture for each condition. Experiments were performed a minimum of three times to ensure reproducibility.

In the in vivo experiments, LPS (10 mg/kg body weight) was injected intraperitoneally in 1 mL of saline. Either 6 or 24 hours later, the rats were anesthetized and samples of liver tissue obtained. Liver samples were frozen in liquid nitrogen and stored at −80°C until use. The same volume of normal saline was injected intraperitoneally into control rats.

Preparation of Nuclear Extract.

A modification of the method by Schreiber et al.14 was used for preparation of primary hepatocyte nuclear extracts. Primary hepatocytes were washed with cold phosphate-buffered saline (PBS). After centrifugation, the cell pellets were frozen at −80°C for at least 1 hour and resuspended in 500 μL buffer A (10 mM HEPES [pH 7.9], 1.5 mM MgCl2, 10 mM KCl, 0.5% NP40). After 10 minutes of incubation on ice, cells were centrifuged for 10 minutes at 6,000 rpm. Nuclei were then washed with 500 μL buffer A without NP40 and centrifuged for 10 minutes at 6,000 rpm. The pellet was then resuspended and salt extracted in 200 μL buffer B (20 mM HEPES [pH 7.9], 25% glycerol [v/v], 1.5 mM MgCl2, 0.5 mM ethylenediaminetetraacetic acid [EDTA], 0.5 M NaCl) on ice for 30 minutes. All buffers included 0.5 mM DTT (dithiothreitol), 0.2 mM PMSF, 1 μg/mL leupeptin, 1 μg/mL pepstatin A, and 1 μg/mL chymostatin. After centrifugation at 20,000g for 30 minutes, supernatants were collected, aliquoted, and stored at −80°C. Protein concentration was measured using bicinchoninic acid (BCA) protein assay reagent (Pierce Chemical Company, Rockford, IL).

Electrophoretic Mobility Shift Assay (EMSA).

Consensus CRE nucleotides (5′-AGA GAT TGC CTG ACG TCAGAG AGC TAG-3′) and consensus activator protein 1 (AP-1) nucleotides (5′-CGC TTG ATG ACT CAG CCG GAA-3′) (Santa Cruz Biotechnology Inc., Santa Cruz, CA) were end-labeled with [γ-32P]ATP by T4 polynucleotide kinase (Boehringer Mannheim Corp., Indianapolis, IN). For binding reactions, 5 × 104 cpm of labeled oligonucleotide probes were incubated with 5 μg of nuclear extract and 1 μg of poly(dI-dC) in binding buffer (4% [v/v] glycerol, 1 mM MgCl2, 0.5 mM EDTA, 0.5 mM DTT, 50 mM NaCl, 10 mM Tris.Cl, pH 7.5) at room temperature for 20 minutes. Protein-DNA complexes were separated by electrophoresis in a 5% non-denaturing polyacrylamide gel in 0.5× TBE buffer and visualized by autoradiography. For competition experiments, a 100-fold molar excess of cold and mutant oligonucleotides (CRE 5′-AGA GAT TGC CTG ATA TCAGAG AGC TAG-3′ and AP-1 5′-CGC TTG ATG ACT TGGCCG GAA-3′, Santa Cruz Biotechnology Inc.) were included in the mixture. When supershift analysis was performed, polyclonal antibodies against CREB, CREM, C/EBPα, c-Jun(D), c-fos, and ATF-1 were added to the extracts 15 minutes before the addition of radio-labeled probes.

Western Blot.

Nuclear extracts were electrophoresed in a 10%–12% polyacrylamide gel containing 0.1% sodium dodecyl sulfate (SDS) and transferred to a nitrocellulose membrane (Schleicher & Schvell, Inc. Keene, NH). Detection of immunoreactive proteins was performed with the enhanced chemiluminescence system (Amersham Pharmacia Biotech, Piscataway, NJ).

DNA Transfection and Luciferase Assay.

Hepatocytes (4 × 105/well) were plated in 6-well plates and co-transfected with plasmids at the indicated concentrations and 0.2 μg/well pIETLacZ (β-galactosidase expression plasmid with CMV promoter) using LipofectAME for 6 hours. Cells were allowed to recover 24 hours and were stimulated as described. At indicated time pints, the cells were washed two times with PBS, lysed in 100 μL of lysis buffer obtained from Promega Luciferase Assay Kit, and centrifuged to remove cell debris. The luciferase activity was measured in relative light units using a Berthold Autolumat LB953 luminometer (Nashua, NH). The β-galactosidase activity was measured as described previously.15 Briefly, an aliquot of the cell extract from transfected cells was incubated at 37°C for 1 hour with 4.86 mg/mL chlorophenol red-β-D-galactopyranoside (Boehringer Mannheim), 62.3 mM MgCl2, and 45 mM β-mercaptoethanol. The reaction was stopped with 0.5 mL of 1 M Na2CO3, and the amount of chlorophenol red formed was measured spectrophotometrically at 575 nm. To normalize the transfection efficiencies, the luciferase activity was expressed as a ratio of relative light units to the β-galactosidase activity obtained from the same extract.

Treatment of Hepatocytes With CRE Oligonucleotides.

A 24-mer CRE palindrome oligonucleotide TGACGTCATGACGTCATGACGTCA (phosphorothoiate) and correspondent mutant oligonucleotide TGTGGTCATGTGGTCATGTGGTCA (phosphorothoiate), as described previously,16 were obtained from Invitrogen (Carlsbad, CA). These oligonucleotides (200 nM/well or 1,000 nM/dish) were transfected using Lipofectamine as described. Cytokines were added 16 hours after transfection and cells were harvested at the indicated times for total RNA or nuclear extract preparation.

RNA Preparation and Northern Blotting.

Homogenized frozen liver samples and cultured hepatocytes were harvested and total cellular RNA preparation and Northern blot analysis were performed as previously described.12

Data and Statistical Analysis.

Data are presented as the mean ± SD. The ANOVA analysis from SigmaStat program (SPSS Science, Chicago, IL) was used to determine statistical significance (P < .05).


Cytokines Induce CRE Binding Activity In Vitro and In Vivo but Inhibit CRE-Reporter Activity in Hepatocytes.

To determine if cytokines altered protein binding to CRE motifs in primary hepatocytes, a time course analysis was performed using a consensus CRE oligonucleotide in EMSA. Low levels of CRE binding activity were detected in unstimulated hepatocytes, whereas CRE binding was increased as early as 30 minutes after stimulating cultured hepatocytes with a combination of proinflammatory cytokines (cytokine mix [CM]) including 200 U/mL IL-1β, 500 U/mL TNFα, and 100 U/mL IFNγ (Fig. 1A). This complex peaked in intensity at 4 to 6 hours and decreased at 24 hours. One predominant DNA-protein complex was identified, and the specificity of protein binding was assessed by the addition of competitor oligonucleotides. A 100 molar excess of unlabeled CRE consensus oligonucleotide abolished CRE binding but a two-base mutant CRE consensus oligonucleotide did not alter the binding activity. To evaluate the effect of proinflammatory cytokines produced during sepsis on in vivo CRE binding activity, an animal model of endotoxemia was used. In rats injected with LPS (10 mg/kg body weight), hepatic CRE binding was significantly increased at both 6 hours and 24 hours compared to rats injected with saline alone (Fig. 1B). Unlike cultured primary hepatocytes, three sequence specific bands appeared in whole liver extracts subjected to gel shift assay (Fig. 1B).

Figure 1.

Time course of cytokine or LPS induced CRE binding in rat primary hepatocytes or liver. (A) The primary hepatocytes were stimulated with cytokines (IL-1β, 200 U/mL; TNFα, 500 U/mL; IFNγ, 100 U/mL, CM), and (B) the rats were injected with LPS and nuclear extracts were prepared at the indicated time point. A total of 12 animals were used in the experiment with 3 animals in each group shown in numbers. The nuclear extracts were incubated with 32P-labeled CRE consensus oligonucleotide in the presence or absence of 100-fold excess of unlabeled or mutant CRE oligonucleotide. Wt, 100 molar excess of unlabeled CRE consensus oligonucleotide; Mt, 100 molar excess of two base mutant CRE consensus oligonucleotide. LPS, lipopolysaccharide; CRE, cyclic AMP response element; IL-1β, interleukin-1β; TNF, tumor necrosis factor; IFN, interferon; CM, cytokine mix; Comp, Competition.

To determine which individual cytokines were predominantly responsible for the increased CRE binding activity demonstrated in the experiments described above, single cytokines at different concentrations were used to stimulate cultured hepatocytes for 90 minutes (Fig. 2A). When used at a high concentration, both IL-1β and TNFα induced CRE binding activity at a level similar to that of the combination of proinflammatory cytokines. IFNγ also induced CRE binding activity when used as a single stimulus but the binding activity did not increase at higher concentrations and was not as great as that produced by IL-1β and TNFα alone. To determine if time-dependent changes in CRE binding activity existed for individual cytokines, the CRE binding activity at 90 minutes and 16 hours was measured by EMSA (Fig. 2B). IL-1β and TNFα induced increased CRE binding at both 90 minutes and 16 hours of stimulation but the binding activity in hepatocytes stimulated by the single cytokines for 16 hours was reduced compared to hepatocytes stimulated with the combination of proinflammatory cytokines.

Figure 2.

Cytokines induce CRE binding activity but inhibit CRE-reporter activity in primary rat hepatocytes. Nuclear extracts from unstimulated or cytokine-stimulated primary hepatocytes were incubated with 32P-labeled CRE consensus oligonucleotides in the presence or absence of 100-fold excess of unlabeled (Wt) or mutant (Mt) CRE oligonucleotides at 90 minutes and 16 hours after stimulation. (A) Hepatocytes were treated with different concentration of cytokines for 90 minutes. The lower panel shows fold induction calculated by scanning densitometry. *P < .05 vs. media. (B) Nuclear extracts from unstimulated or cytokine-stimulated primary hepatocytes 90 minutes and 16 hours after stimulation. (C) Cultured primary rat hepatocytes were transfected by 1 μg/well pCRE-Luc and then stimulated with IL-1β, 200 U/mL; TNFα, 500 U/mL; or IFNγ, 100 U/ml for 6 hours and then luciferase assays were performed. Indicated is the fold induction of relative luciferase activity vs. unstimulated group (n = 12, *P < .05 vs. media. CRE, cyclic AMP response element; IL-1β, interleukin-1β; TNF, tumor necrosis factor; IFN, interferon; CM, cytokine mix; pCRE-Luc, 5× consensus CRE-driven promoter.

To determine if the cytokine-induced increase in CRE binding activity was associated with increased CRE-dependent transcriptional activity, we utilized a CRE-dependent reporter construct. Primary rat hepatocytes were transfected (1 μg/well) with pCRE-luc, which harbors a 5× consensus CRE sequence in its luciferase enhancer region, and stimulated by IL-1β, TNFα, and IFNγ for 6 hours. As a basal transcriptional control, 1 μg pLucMCS, in which only a TATA box is included in its enhancer region to drive luciferase expression, was transfected (Fig. 2C, Control). Unlike some cell types, in which basal CRE transcriptional activity is almost undetectable,17 there is a relatively high level of CRE transcriptional activity in unstimulated primary hepatocytes. TNFα and IFNγ significantly depressed luciferase activity while IL-1β caused less inhibition of CRE-dependent luciferase activity. These data suggest that cytokine-mediated CRE binding decreases CRE transcriptional activity and CRE-mediated gene expression.

In Vivo Competition of CRE Binding Alters the Effect of Cytokines on PEPCK Gene Expression.

It has been shown that CRE-palindromic oligonucleotides can penetrate into cells, compete with transcription factors for CRE binding sites, and specifically interfere with CRE- and AP-1-directed transcription in vitro.16, 18 To determine the effect of cytokine-mediated CRE binding on CRE-dependent gene expression, we evaluated PEPCK gene expression, which has been shown to be CRE-dependent,3 and used a CRE palindromic oligonucleotide (CREP) decoy in primary hepatocytes. Cultured primary hepatocytes were treated by 24-mer CREP for 24 hours, and then the cells were stimulated with cytokines. The cells were harvested and nuclear extracts were prepared for EMSA assay after 4 hours of cytokine stimulation. As shown in Fig. 3A, CRE binding was completely diminished by CRE decoy palindrome oligonucleotide treatment while a two-base mutant control oligonucleotide had no effect on cytokine-induced CRE binding activity.

Figure 3.

CRE-decoy oligonucleotide competes with CRE in vitro and increases PEPCK mRNA in hepatocytes. Primary hepatocytes were treated with 24-mer CRE palindrome (CREP) or two base mutant CRE palindrome (mCREP) for 24 hours, then the cells were stimulated with cytokine mix for 4 hours. (A) Nuclear extracts from hepatocytes stimulated with cytokine mix were reacted in EMSA with 32P-labeled CRE consensus oligonucleotides in the presence of 100-fold excess cold CRE and AP-1, or mutant CRE competitor. (B) Total RNA was prepared and subjected to Northern blot using a specific rat PEPCK cDNA probe (upper panel). Bottom panel shows RNA gel photography for equal loading. CRE, cyclic AMP response element; PEPCK, phosphoenolpyruvate; mRNA, messenger RNA; rRNA, ribosomal RNA; EMSA, electrophoretic mobility shift assay; AP-1, activator protein; CM, cytokine mix; cDNA, complementary DNA.

To determine the effect of cytokine-induced CRE binding on PEPCK mRNA expression, primary hepatocytes were cultured for 16 hours with CRE palindrome decoy oligonucleotides (or mutated control oligonucleotides), stimulated with cytokines for 6 hours, and Northern blot analysis performed. Cytokines decreased PEPCK mRNA expression and this decrease was reversed by transfection with the CRE decoy oligonucleotide but not the mutant decoy (Fig. 3B).

CREB and c-Jun Family Proteins Bind to a CRE Consensus Nucleotide in Primary Hepatocytes.

In order to identify proteins that may be binding to CRE sites after cytokine stimulation and producing decreased CRE-mediated transcriptional activation and RNA expression, supershift analysis of CRE complexes was performed. We performed EMSA supershift analysis using extracts from hepatocytes stimulated with cytokines for 90 minutes (Fig. 4A, upper) and liver tissues of rats injected with LPS for 6 hours (Fig. 4A, lower). We found that when extracts from cultured hepatocytes were used, the CRE complex was shifted by antibodies to CREB, ATF-3, c-Jun, and JunD (Fig. 4A, upper). When extracts from liver tissue were used, CREB, c-Jun, and JunD were binding to the consensus CRE oligonucleotide (Fig. 4A, lower).

Figure 4.

(A) CREB and Jun Family bind to the CRE consensus oligonucleotide. Supershift assay was performed using nuclear extracts from cytokine- stimulated hepatocytes (IL-1β, 200 U/mL; TNFα, 500 U/mL; IFNγ, 100 U/mL) at 90 minutes (upper) or from rat liver treated with LPS for 6 hours (lower). Specific polyclonal antibodies indicated were added to EMSA reactions including the 32P-labled CRE consensus oligonucleotide. The migration of the super-shifted complex is indicated. (B) Effect of cytokines on CREB and c-Jun protein phosphorylation and expression in nuclear extracts. Western blots were carried out using nuclear extracts from cytokine stimulated primary hepatocytes at indicated time points of stimulation. Western blots were performed using polyclonal antibodies against CREB and Ser133 phosphorylated CREB (upper) or polyclonal antibodies against c-Jun and phosphorylated c-Jun. (C) Cytokines increase c-jun binding to consensus CRE sequence. Primary hepatocytes were stimulated with IL-1β, 200 U/mL; TNFα, 500 U/mL; IFNγ, 100 U/ml, or cytokine mix for 4 hours. The nuclear extracts were incubated with 32P-labeled CRE consensus oligonucleotide in the presence of a polyclonal antibody against c-Jun. No antibody control was unstimulated hepatocytes with normal serum. (D) Effect of overexpression of CREB and c-Jun on CRE-reporter activity in primary rat hepatocyte. Primary hepatocytes were co-transfected by pCRE-Luc 1ug and pRc/RSV-CREBDIEDML, pcDNA3.1-c-Jun or the combination for 6 hours at the indicated doses. Luciferase assays were performed after recovery for 24 hours (n = 10 per group). Relative luciferase activity is indicated by fold induction. *P < .05 vs. vector. CREB; cyclic AMP response element binding protein; CRE, cyclic AMP response element; IL-1β, interleukin-1β; TNF, tumor necrosis factor; INF, interferon; LPS, lipopolysaccharide; EMSA, electrophoretic mobility shift assay; CM, cytokine mix; pCRE, 5× consensus CRE-driven promoter; pRc/RSV-CREB, CREB expression vector, pcDNA 3.1-c-Jun, c-Jun expression vector; ATF, activating transcription factor; pCREB, phosphorylated CREB; CREM, cAMP response element modulator.

CRE-dependent trans-activation can occur either by increased binding of activator proteins to a CRE regulatory element or by phosphorylation-induced trans-activation of constitutively binding proteins (principally CREB).19 In addition, CREB protein is itself inducible by many of the conditions that increase CRE-dependent gene expression because the promoter region of the CREB gene contains CRE regulatory elements.20 To evaluate the mechanism for cytokine-induced CRE binding activity, we performed Western blot analysis of nuclear extracts to evaluate the levels of CREB and c-Jun proteins (Fig. 4B). The cytokines that increased CRE binding activity increased CREB phosphorylation but had no effect on the levels of total CREB protein in nuclear extracts (Fig. 4B). Cytokines induced c-Jun phosphorylation from 1 hour to 16 hours as shown in Fig. 4B. Additionally, cytokines also increased total c-Jun accumulation after 4 hours of stimulation.

To test the role of c-Jun in cytokine-induced inhibition of CRE transcriptional activity, EMSA supershift assay was performed using nuclear extracts from hepatocytes stimulated with cytokine mix or IL-1β, TNFα, and IFNγ individually for 4 hours. As indicated in Fig. 4C, c-Jun was significantly increased in CRE complexes after cytokine stimulation.

To test our hypothesis that one or more of the proteins bound to CRE sequences acts as a transcriptional suppressor, expression vectors of CREBDIEDML and c-Jun were co-transfected with pCRE-Luc and luciferase assay performed 24 hours after transfection (Fig. 4D). Overexpression of CREB significantly increased CRE transcriptional activity. Overexpression of c-Jun decreased cytokine-induced and CREB-induced CRE dependent reporter activity suggesting that c-Jun acts as a suppressor of CRE-dependent transcription in a dose-dependent fashion.

Cytokines Induced AP-1 Binding Activity in Primary Hepatocytes.

Because similarities exist between the CRE target sequence and AP-1 target sequences, we also used an AP-1 consensus oligonucleotide as a competitor of the CRE binding complex. DNA-protein complexes were specifically reduced by either 100-fold molar excess unlabeled CRE or unlabeled AP-1 consensus oligonucleotides but not by a mutated CRE oligonucleotide (Fig. 3A). The ability of CRE-decoy oligonucleotides to compete with AP-1 binding elements in hepatocytes was also examined using 32P labeled AP-1 consensus oligonucleotide. As shown in Fig. 5, cytokines induced AP-1 binding in primary hepatocytes at 4 hours and CRE-decoy oligonucleotide treatment diminished AP-1 protein binding. Both unlabeled CRE consensus oligonucleotides and unlabeled AP-1 consensus oligonucleotides compete with cytokine-induced AP-1-protein binding.

Figure 5.

CRE-decoy oligonucleotide competes with AP-1 binding in vivo. Primary hepatocytes were treated with 24-mer CRE palindrome (CREP) or two base mutant CRE palindrome (mCREP) for 24 hours. Then, the cells were stimulated with CM for 4 hours and nuclear extracts prepared. Nuclear extracts from hepatocytes stimulated with CM were reacted in EMSA with 32P-labeled AP-1 consensus oligonucleotide in the presence of 100-fold excess cold CRE and AP-1, or mutant AP-1 competitor. CRE, Cyclic AMP response element; AP-1, activator protein 1; CREP, CRE palindrome; mCREP, mutant CRE palindrome; CM, cytokine mix; EMSA, electrophoretic mobility shift assay.

PKA and JNK Pathways Are Partially Involved in Cytokine-Induced Changes in CRE Binding Activity.

Phosphorylation of CREB at serine 133 by the cyclic AMP-dependent protein kinase (protein kinase A, PKA) is required to trigger CREB transactivation activity.21, 22 To evaluate whether phosphorylation was necessary for the increased CRE binding activity seen in these studies, hepatocytes were stimulated in the presence of cAMP dependent PKA inhibitor (PKAI) and a specific c-Jun N terminal kinase inhibitor, SP600125.23–25 The addition of the PKAI slightly reduced cytokine-stimulated hepatocyte CRE binding activity (Fig. 6A) but had no significant effect on CRE-dependent luciferase activity (Fig. 6B) at the concentration that can block at least 80% of PKA activity26 in hepatocytes. SP600125 inhibited cytokine-induced CRE binding in a dose dependent manner and significantly increased CRE-dependent luciferase activity (Fig. 6B).

Figure 6.

Effect of PKA and JNK inhibitors on CRE binding activity and PEPCK mRNA expression in primary hepatocytes. (A) Primary hepatocytes were pretreated PKA inhibitor (PKAI) or SP600125 at the indicated concentration for 30 minutes before cytokine stimulation. Nuclear extracts were collected 90 minutes after cytokine stimulation. EMSA was carried out using 5 μg nuclear extract and 32P-labeled CRE consensus oligonucleotide in the presence or absence of 100-fold excess of unlabeled (Wt) or two base mutant (Mt) CRE oligonucleotide. (B) Primary hepatocytes were transfected with pCREluc and PIETlacZ. After 24 hours of recovery, cells were treated with cytokine mix (CM) with different concentrations of PKA or JNK inhibitor for 6 hours, and cells were lysed and subject to luciferase assay. The relative luciferase activity vs. control is indicated. *P < .05 vs. media, **P < .05 vs. CM. (C) Northern blot. Primary hepatocytes were pretreated with SP600125 for 30 minutes, and then stimulated with CM for 8 hours. Total RNA was prepared and subjected to Northern blot with PEPCK or GAPDH probes. PKA, protein kinase A; JNK, c-jun N-terminal kinase; CRE, cyclic AMP response element; PEPCK, phosphoenolpyruvate carboxykinase; mRNA, messenger RNA; PKAI, PKA inhibitor; EMSA, electrophoretic mobility shift assay; CM, cytokine mix; GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

As shown above, the effects of SP600125 on CRE binding activity and CRE dependent luciferase transcription suggest that phosphorylation of c-Jun by JNK may play an important role in the effects of cytokines on CRE dependent transcriptional activity. We therefore tested SP600125 on cytokine-induced inhibition of PEPCK mRNA expression. Figure 6C indicates that SP600125 partially reversed the cytokine-induced suppression of PEPCK mRNA expression but a dose-dependent effect was not seen at the concentration above 5 μM.


CRE-mediated regulation of gene expression is important in hepatocyte gluconeogenesis, cell growth, and differentiation. CRE binding sites are present in many genes that are hypothesized to be important in the host response to inflammation and infection but the role of CRE-mediated gene regulation in sepsis has received little attention. We used both in vitro and in vivo models of sepsis using pro-inflammatory cytokines and LPS to study the role of CRE motifs in hepatocytes during sepsis. We demonstrate that CRE binding activity is increased by cytokine stimulation in cultured primary rat hepatocytes and hepatocytes from rats treated with LPS in vivo. In cultured hepatocytes, CRE binding is rapidly induced by cytokines (30 minutes), peaks at 4–6 hours after stimulation, and remains elevated compared to unstimulated hepatocytes for at least 16 hours before declining after 24 hours. Cytokine-induced CRE binding activity was associated with decreased CRE-dependent transcriptional activity as measured by luciferase activity with a CRE-dependent reporter construct. Expression of mRNA of PEPCK, a CRE-mediated gene, was also suppressed by cytokines. These data suggest that, in hepatocytes subjected to pro-inflammatory cytokines, increased binding of proteins to CRE sites decreases subsequent CRE-mediated gene expression.

Both CREB and c-Jun proteins were components of the CRE binding complex in cytokine-stimulated hepatocytes. Cytokine-stimulated CRE binding at 90 minutes was not associated with increased CREB and c-Jun protein expression, but was associated with elevated levels of phosphorylated CREB protein. CREB is known to form heterodimers with a variety of other transcription factors especially AP-1 family proteins.27 The products of the proto-oncogenes Jun and Fos typically bind as a heterodimeric complex to the AP-1 binding site;28 however, Jun/Jun homodimers bind to both CRE and AP-1 sequences,29 and CREB-1/(ATF-2)/Jun heterodimers bind CRE.27, 30 Although the binding patterns at CRE sites are different in different cell types, the CRE-1 site in the PEPCK promoter contains a CRE consensus sequence and binds both CREB and AP-1 proteins from rat liver nuclei.31, 32 Our data in primary cultured hepatocytes treated with pro-inflammatory cytokines revealed a similar binding pattern using a consensus CRE oligonucleotide. Our data show that cytokines inhibited PEPCK gene expression in primary hepatocytes. Competition of CRE binding by 24-mer CRE palindromic oligonucleotide increased PEPCK mRNA expression. These findings suggest that the protein(s) responsible for the increased CRE binding after cytokine stimulation are also suppressing CRE-dependent gene expression. However, we cannot conclude that cytokines inhibit PEPCK expression only through CRE motifs because the CRE palindrome also competes with AP-1 binding activity.

Our studies have shown that the CRE binding complex induced by cytokines consists, in part, of CREB and c-Jun proteins. Transcriptional activity utilizing a CRE-reporter co-transfected with c-Jun or CREB indicated that CREB activated CRE reporter activity and c-Jun acted as a CRE repressor in hepatocytes. The inhibition in CRE-mediated transcriptional activation by cytokines may occur through increased c-Jun binding at CRE sites in cultured primary hepatocytes. A specific c-Jun N-terminal kinase inhibitor reversed cytokine-induced CRE binding as well as CRE-dependent reporter activity, indicating that this activity is associated with c-Jun phosphorylation. Several studies show that c-Jun down regulates33–37 or up regulates38–41 gene expression through CRE binding sites but the mechanism remains unclear. CREB/c-Jun heterodimers may form in vivo and contribute to c-Jun-mediated changes in CRE activity and may contribute to the results seen here.34, 36, 42 Heterodimer formation between CRE-BP1 and c-Jun can be extensive.43 However, making correlations between c-Jun and CREB and their functional activity, particularly reaching definite conclusions regarding which factors drive gene expression under certain pathological circumstance, has been difficult. We have previously demonstrated that glucagon and cAMP, potent activators of PKA and CRE-mediated gene expression in hepatocytes, decreased cytokine-stimulated expression of the inducible nitric oxide synthase gene in cultured hepatocytes.44, 45 The present study demonstrated that CRE decoy nucleotides reverse cytokine inhibition of PEPCK mRNA expression. Whether increased CRE binding increases or decreases the expression of specific hepatocyte genes may depend on the specific proteins binding to specific CRE sites in the promoter region of the gene of interest as opposed to a specific, generalizable property of the CRE site itself. More complete identification of all proteins contributing to CRE binding in cytokine-stimulated hepatocytes will help elucidate the role of CRE in hepatocyte function during sepsis.

In the present study, inhibition of c-Jun phosphorylation with the JNK inhibitor, SP600125, decreased cytokine induced CRE binding and partially reversed cytokine induced inhibitory effects on CRE dependent reporter activity. This result indicates that the phosphorylation status of c-Jun may mediate the effects of cytokines on CRE binding and CRE dependent reporter activity. Two possible mechanisms may account for this finding. In most cell types, c-Jun is a labile protein with a half-life of approximately 2 hours.46 Phosphorylation by JNK inhibits degradation of c-Jun, leading to accumulation of the protein.47 As shown in Fig. 4B, cytokines not only phosphorylated c-Jun but also increased c-Jun expression at 4 and 16 hours. Whether increased phosphorylation or increased expression of c-Jun is primarily responsible for the results seen in this study will require further analysis. We cannot exclude the possibility of the interaction between the c-Jun activation domain with a co-activator contributing to the findings described above.19, 48–50

In summary, hepatocytes stimulated with a mixture of pro-inflammatory cytokines increased CRE binding activity in vitro and in vivo. The individual cytokines TNFα and IL-1β also increase CRE binding activity. Cytokine-induced CRE binding was associated with decreased CRE-dependent transcriptional activation. The cytokine induced CRE binding complex consists, in part, of CREB and c-Jun proteins, and the cytokine-induced suppression of CRE transcriptional activation was mediated, in part, by c-Jun. These data suggest that cytokine-induced alterations in transcription factor binding exert complex effects on CRE-regulated genes in hepatocytes. These complex protein/DNA interactions may have profound effects on hepatic function during sepsis and inflammation and may contribute to the alterations in hepatic metabolism seen after infection.


We thank Dr. Michael J. Birrer for providing pcDNA3.1-c-Jun plasmid and Dr. Jean-Rene Cardinaux for providing pRC/RSV-CREBDIEDML plasmid.