Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Cyclooxygenase-2 (COX-2) plays roles in pathophysiological processes, including inflammation, angiogenesis, and tumorigenesis. Lipopolysaccharide (LPS) and many growth factors induce COX-2 (Wadleigh et al., 2000). LPS is an endotoxin that induces septic shock syndrome and stimulates the production of inflammatory mediators such as COX-2, nitric oxide, tumor necrosis factor- (TNF-), interleukins, prostanoids, and leukotrienes (Lee et al., 1992; Kubes and McCafferty, 2000; Hewett and Roth, 1993). Development of COX-2 inhibitors represents a major advance in the therapy of inflammatory processes, and their use includes prevention or treatment of disorders associated with the induction of this enzyme (e.g., colon cancer). Despite the importance of COX-2 as a target for the treatment of human inflammatory disorders, the functional role of the signaling pathways for the induction of COX-2 is largely unsolved or contradictory.

A variety of stimuli increase intracellular ceramide through sphingomyelinase or de novo synthesis (Hannun, 1994). Ceramide is implicated in cell growth, differentiation, apoptosis, inflammation, and immune responses. Several laboratories studied the ceramide signaling pathways. Ceramide induced COX-2 in mammary epithelial cells (Subbaramaiah et al., 1998). The signaling pathway of TNF--induced COX-2 expression was mediated with the formation of ceramide and sequential activation of extracellular signal-regulated kinase 1/2 (ERK1/2), p38 kinase, c-Jun N-terminal kinase (JNK), I-B kinase complex 1/2, and nuclear factor-B (NF-B) in the COX-2 promoter (Chen et al., 2001). Also, ceramide mediates age-associated increase in COX-2 expression (Claycombe et al., 2002). The ceramide-induced up-regulation in prostaglandin E₂ production was mediated through transcriptional up-regulation of COX-2. By contrast, a study from another laboratory showed that C2-ceramide (C2), a cell-permeable ceramide analog, inhibited LPS-elicited COX-2 induction and prostaglandin E₂ formation with the inhibition of NF-B and activator protein-1 (AP-1) activation (Hsu et al., 2001). They proposed the hypothesis that the inhibition of COX-2 induction by C2 resulted from the inhibition of LPS-stimulated I-B kinase, p38 kinase, and protein kinase C. Given these controversial reports on the role of ceramide in the induction of COX-2, we were tempted to study the effect of C2 on LPS-inducible expression of COX-2 and to explore the cell signaling.

Studies have shown that CCAAT/enhancer binding protein (C/EBP), cyclic-AMP response element-binding protein (CREB) and NF-B were commonly or individually involved in the regulation of the COX-2 gene. Evidence has accumulated to show that the C/EBP element plays an important role in the induction of COX-2. In particular, activation of C/EBP leads to the induction of COX-2 (Thomas et al., 2000; Wadleigh et al., 2000). The regulatory region for the COX-2 gene includes cAMP response element/E-box elements. NF-B, which is activated by the inflammatory responses during viral and bacterial infections (Grilli and Memo, 1999; Kim et al., 2000), is involved in the expression of inflammatory genes (e.g., TNF-) (Muller et al., 1993). Nevertheless, the role of NF-B in COX-2 induction is controversial. In the present study, we examined alterations in the activation of these transcription factors by C2 in LPS-treated macrophages.

Nuclear translocation of C/EBP and C/EBP DNA binding were monitored by immunoblotting, immunocytochemical, and gel mobility shift assays in association with the induction of COX-2 by LPS + C2. Now, we report that C2 potentiates LPS-inducible COX-2 expression in macrophages as a result of the increase in C/EBP activation. C2 increases the expression of C/EBP and induces nuclear translocation and DNA binding of C/EBP via the pathway involving JNK1, which leads to enhancement of LPS-stimulated COX-2 induction. Toward the end of this study, we studied the effect of C2 on AP-1 activation and the role of JNK pathway in the enhanced AP-1 activation, which we expect to contribute to the increase in C/EBP expression.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Materials. C2 was obtained from Merck Co. (Darmstadt, Germany). Anti-COX-2 antibody was obtained from Cayman (Ann Arbor, MI). Anti-C/EBP-, -, -, and - form antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Horseradish peroxidase- or fluorescein isothiocyanate-conjugated goat anti-rabbit IgGs were purchased from Zymed Laboratories Inc. (San Francisco, CA). PD98059 was obtained from Calbiochem (San Diego, CA). [-³²P]ATP (3000 mCi/mmol) was obtained from PerkinElmer Life Sciences (Boston, MA). The consensus oligonucleotides of C/EBP, CREB, NF-B, AP-1, and random prime/5'-end labeling kits were supplied from Promega Corporation (Madison, WI). JNK1 dominant-negative mutant (KmJNK1) was kindly provided from Dr. N. Dhanasekaran (Fels Institute for Cancer Research and Molecular Biology, Department of Biochemistry, Temple University, Philadelphia, PA). MKK1 dominant-negative mutant was gifted from Dr. N.G. Ahn (Howard Hughes Medical Institute, University of Colorado, Boulder, CO). C/EBP-specific dominant-negative expression plasmid (AC/EBP) was a gift from Dr. C. Vinson (National Institutes of Health, Besthesda, MD) (Ahn et al., 1998). Most reagents for the molecular studies were obtained from Sigma Chemical (St. Louis, MO).

Cell Culture. Raw264.7 cells, a murine macrophage cell line (American Type Culture Collection, Manassas, VA), were cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, 100 U/ml penicillin, and 100 µg/ml streptomycin. Raw264.7 cells were plated at a density of 2 to 3 × 10⁶/ml and preincubated for 24 h at 37°C. Cells were maintained at 37°C in a humidified atmosphere containing 5% CO₂. For all experiments, cells were grown to 80 to 90% confluence and were subjected to no more than 20 cell passages. Cells were incubated with 0.1 µg/ml LPS (Escherichia coli 026:B6; Difco, Detroit, MI) for the time periods indicated under Results and in the figure legends. C2 (as dissolved in dimethyl sulfoxide) was simultaneously added with LPS to the incubation medium.

Immunoblot Analysis. The expression of COX-2 was immunochemically monitored in lysates of Raw264.7 cells using anti-mouse COX-2 antibody. The secondary antibodies were horseradish peroxidase-conjugated anti-rabbit antibodies. The band of COX-2 protein was developed using enhanced chemiluminescence immunoblot detection system according to the manufacturer's instructions (Amersham Biosciences, Buckinghamshire, UK). C/EBP-, -, -, and - forms in the nuclear or lysate fractions were immunoblotted with the respective form-specific antibodies.

RT-PCR Analysis. Total RNA (2 µg) obtained from the cells was reverse-transcribed using an oligo(dT) adaptor as a primer to produce cDNAs. The specific cDNA probe for the COX-2 gene was amplified by reverse transcription (RT)-polymerase chain reaction (PCR) using the selective primers and cloned in a TA vector (Promega). The primers used were COX-2: sense primer, 5'-TACAAGCAGTGGCAAAGGC-3'; antisense primer, 5'-CAGTATTGAGGAGAACAGATGGG-3' (287 bp). Expression of glyceraldehyde-3-phosphate dehydrogenase gene was assessed by PCR using the sense primer 5'-TCGTGGAGTCTACTGGCGT-3' and the antisense primer 5'-GCCTGCTTCACCACCTTCT-3' (510 bp). PCRs were conducted using the following conditions for 38 cycles: denaturation at 94°C for 0.5 min, annealing at 49°C for 1 min, and elongation at 68°C for 1.5 min. Band intensities of the amplified DNAs were compared after visualization on an UV transilluminator.

Preparation of Nuclear Extracts. Nuclear extracts were prepared essentially according to the previously published method (Schreiber et al., 1990). Briefly, the cells in dishes were washed with ice-cold PBS. Cells were then scraped, transferred to microtubes, and allowed to swell after the addition of 100 µl of hypotonic buffer containing 10 mM HEPES, pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.5% Nonidet P-40, 1 mM dithiothreitol, and 0.5 mM phenylmethylsulfonyl fluoride. The lysates were incubated for 10 min in ice and centrifuged at 7,600g for 6 min at 4°C. Pellets containing crude nuclei were resuspended in 50 µl of extraction buffer containing 20 mM HEPES, pH 7.9, 400 mM NaCl, 1 mM EDTA, 10 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride and then incubated for 1 h in ice. The samples were centrifuged at 15,000g for 10 min to obtain supernatants containing nuclear fractions. Nuclear fractions were stored at 70°C until use.

Gel Retardation Assay. Double stranded DNA probes for the consensus sequences of C/EBP (5'-TGCAGATTGCGCAATCTGCA-3'), CREB (5'-AGAGATTGCCTGACGTCAGAGAGCTAG-3'), NF-B (5'-AGTTGAGGGGACTTTCCCAGGC-3'), and AP-1 (5'-CGCTTGATGAGTCAGCCGGAA-3') were used for gel-shift analyses after end-labeling of each probe with [-³²P]ATP and T₄ polynucleotide kinase. Nuclear extracts were prepared by modification of the procedure published previously (Schreiber et al., 1990). The reaction mixtures contained 2 µl of 5× binding buffer containing 20% glycerol, 5 mM MgCl₂, 250 mM NaCl, 2.5 mM EDTA, 2.5 mM dithiothreitol, 0.25 mg/ml poly(dI-dC), and 50 mM Tris-Cl, pH 7.5, 5 µg of nuclear extracts, and sterile water in a total volume of 10 µl. Reactions were initiated by addition of 1 µl probe (10⁶ cpm) after 10 min of preincubation and continued for 30 min at room temperature. The specificity of the DNA/protein binding was confirmed by competition reactions, in which a 20-fold molar excess of unlabeled C/EBP, CREB, NF-B, or AP-1 oligonucleotide was added to each reaction mixture before the addition of radiolabeled probe. For supershift assay, the antibodies (2 µg each) were added to the reaction mixture and additionally incubated for 1 h at 25°C. Samples were loaded onto 4% polyacrylamide gels at 100 V. The gels were removed, fixed, and dried, followed by autoradiography.

Immunocytochemistry. Raw264.7 cells were grown on Lab-TEK chamber slides (Nalge Nunc International, Naperville, IL) and incubated in medium for 24 h. Standard immunocytochemical methods were used as described previously by Cho et al. (2002). For immunostaining, cells were fixed in 100% methanol for 30 min and washed three times with PBS. After blocking in 5% bovine serum albumin in PBS for 1 h at room temperature or overnight at 4°C, cells were incubated for 2 h with polyclonal rabbit anti-C/EBP antibody in PBS containing 1% bovine serum albumin. Cells were incubated with 1:100 dilution of fluorescein isothiocyanate-conjugated goat anti-rabbit IgG after serial washing with PBS. Counter-staining with propidium iodide (PI) verified the location and integrity of nuclei. Stained cells were washed and examined using a laser scanning confocal microscope (Leica TCS NT; Leica Microsystems, Wetzlar, Germany).

Transient Transfection with Dominant-Negative Mutant of C/EBP. Cells were plated at a density of 0.5 × 10⁶ cells/well in six-well dishes and transfected the following day. Briefly, cells were incubated with 1 µg of C/EBP dominant-negative mutant (A/CEBP) plasmid or pCMV500 plasmid (an empty vector that was used as a control) DNA and 3 µl of LipofectAMINE reagent (Invitrogen, Carlsbad, CA) in 1 ml of antibiotic-free minimal essential medium (MEM) for 3 h. The cells were incubated in Dulbecco's modified Eagles medium for 3 h and then exposed to LPS or LPS + C2 for 20 h.

Decoy Oligodeoxynucleotide Technique. Double-stranded oligodeoxynucleotide (ODN) was prepared from complementary single-stranded phosphorothioate-bonded ODN (Bioneer, Chungbuk, Korea) by melting at 95°C for 5 min followed by a cool-down phase at ambient temperature (Cho et al., 2002). The sequences of the single-stranded ODN were as follows. Underlined letters denote phosphorothioate-bonded bases: AP-1, 5'-CGCTTGATGACTCAGCCGGAA-3' and mutant AP-1 (mAP-1), 5'-CGCTTGATTACTTAGCCGGAA-3' (Cho et al., 2002). Cells were preincubated with 10 µM decoy ODN for 1 h and then further incubated with LPS (0.1 µg/ml) or LPS + C2 (50 µM) for 12 h. Transfection of the decoy ODN was achieved without using a cationic lipid or liposomal complex.

Stable Plasmid Transfection. Cells were transfected using Transfectam according to the manufacturer's instructions (Promega). Cells were replated 24 h before transfection at a density of 2 × 10⁶ cells in a 10-cm² plastic dish. For use in JNK1 dominant-negative mutant [JNK1()] or MKK1 dominant-negative mutant [MKK1()] stable transfection, 20 µl of Transfectam was mixed with 10 µg of a JNK1() or MKK1() plasmid in 2.5 ml of MEM. Cells were transfected by addition of MEM containing each plasmid and Transfectam and then incubated at 37°C in a humidified atmosphere of 5% CO₂ for 6 h. After addition of 6.25 ml MEM with 10% fetal bovine serum, cells were incubated for additional 48 h at 37°C and 50 µg/ml of G-418 was added to select the resistant colonies.

Scanning Densitometry. Scanning densitometry of the immunoblots was performed with Image Scan and Analysis System (-Innotech Corporation, San Leandro, CA). The area of each lane was integrated using the software AlphaEase version 5.5, followed by background subtraction. One-way analysis of variance procedures were used to assess significant differences among treatment groups. For each significant effect of treatment, the Newman-Keuls test was used for comparisons of multiple group means. The criterion for statistical significance was set at p < 0.05 or 0.01.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

C2 Potentiation of COX-2 Induction by LPS. First, we determined whether C2 altered the level of COX-2 protein in macrophages incubated with LPS. Whereas COX-2 was not detected in control cells, LPS (0.1 µg/ml, 20 h) notably increased the COX-2 protein level. The induction of COX-2 by LPS was enhanced by C2 at the concentrations of 30 to 50 µM (20 h), whereas dihydro-C2, an inactive analog of C2, had no effect (Fig. 1A). Cell viability was not affected by C2 at the concentrations employed. The expression of COX-2 was assessed at 50 µM of C2 in subsequent experiments. To determine the time points for C2 enhancement of COX-2 induction by LPS, cells were treated with LPS (0.1 µg/ml) in the presence or absence of C2 for 6, 12, or 20 h (Fig. 1B). C2 enhanced COX-2 induction by LPS at 12 or 20 h. C2 alone did not induce COX-2 in macrophages at 20 h, which was in agreement with the previous report (Hsu et al., 2001). Studies were extended to determine whether the expression of COX-2 protein paralleled that of its mRNA. LPS increased the COX-2 mRNA at 6 to 12 h. RT-PCR analysis showed that C2 notably enhanced LPS-inducible increase in the COX-2 mRNA at 12 h (Fig. 1C).

View larger version (26K): [in this window] [in a new window]   Fig. 1.   The effect of C2 on LPS-inducible COX-2 expression. A, enhancement of COX-2 induction by C2 in Raw264.7 cells. Cells were treated with varying concentrations of C2 dissolved in dimethyl sulfoxide for 20 h in combination with LPS (0.1 µg/ml). Control cells were incubated with vehicle alone. Dihydro-C2 was used for comparison. The level of COX-2 was monitored as described under Materials and Methods. Data represent the mean ± S.D. with at least three separate experiments (*, p < 0.05; **, p < 0.01, significant compared with LPS alone) (LPS alone = 1). B, time course of COX-2 expression in cells treated with LPS in the presence or absence of C2 (50 µM). The relative COX-2 protein levels were measured by scanning densitometry of the immunoblot bands. Data represent the mean ± S.D. with at least three separate experiments *, p < 0.05; **, p < 0.01, significant compared with LPS alone at 20 h (i.e., 1). C, the effects of C2 on the COX-2 mRNA level. Raw264.7 cells were exposed to LPS (0.1 µg/ml) with or without C2 (50 µM). RT-PCR analysis was performed to determine COX-2 mRNA in total RNA fractions (2 µg each) isolated from cells treated with LPS or LPS + C2 for 1 to 12 h. Equal RNA loading was confirmed by RT-PCR of glyceraldehyde-3-phosphate dehydrogenase mRNA.

Effect of C2 on LPS-Inducible C/EBP, CREB, and NF-B Activation. Expression of the COX-2 gene depends on the C/EBP element present in the upstream region of the gene (Thomas et al., 2000). To test whether C2 potentiation of the COX-2 gene induction was mediated by C/EBP activation, electrophoretic mobility shift for C/EBP binding activity was performed with the nuclear extracts of cells exposed to LPS in the presence or absence of C2 using a radiolabeled C/EBP consensus oligonucleotide. Treatment of cells with LPS for 3 to 12 h resulted in an increase in C/EBP DNA binding compared with control, whereas concomitant treatment of cells with C2 notably enhanced LPS-inducible C/EBP binding (Fig. 2A, left). The band intensity of C/EBP DNA binding maximally increased at 6 h and returned toward that of LPS alone at 12 h. C2 alone (3-12 h) had no effect or minimally increased C/EBP binding to the C/EBP binding site (Fig. 2A, middle).

View larger version (77K): [in this window] [in a new window]   Fig. 2.   Gel-shift analyses of C/EBP, CREB, and NF-B transcription complexes. Nuclear extracts were prepared from Raw264.7 cells cultured with LPS (0.1 µg/ml) or LPS + C2 (50 µM). A, gel-shift analyses of C/EBP. All lanes contained 5 µg of nuclear extracts and 5 ng of labeled C/EBP consensus oligonucleotide. Supershift analyses were carried out by incubating the nuclear extracts (LPS or LPS + C2, 6 h) with the specific polyclonal antibodies directed against C/EBP forms (, , , and ) and CBP/p300 (p300) for 1 h. SS, supershift of the retarded C/EBP complex. Competition studies were carried out by adding a 20-fold excess of an unlabeled C/EBP oligonucleotide (20×) to the nuclear extracts from cells treated with LPS or LPS + C2, and the DNA-binding reactions were performed by gel-shift analysis. B, gel-shift analysis of CREB. Gel-shift analysis was performed as described in A. C, gel-shift analysis of nuclear extracts using the consensus oligonucleotide of NF-B. The specificity of CREB or NF-B binding was confirmed by addition of an excess amount of free probe (20 ×). Results were confirmed by repeated experiments.

Cell Signaling for C2 Enhancement of COX-2 Induction. Major signaling pathways for the induction of COX-2 transmit through MAP kinase pathways (Paul et al., 1999; Wadleigh et al., 2000). The role of the MAP kinase signaling pathways in the potentiation of COX-2 induction by C2 was investigated using specific inhibitors and/or stable transfection with dominant-negative mutant vectors. First, the pathway responsible for the induction of COX-2 by LPS was assessed. It has been shown that the p38 kinase pathway was involved in COX-2 induction (Chen et al., 1999). SB203580 (10 µM) completely inhibited the enzyme induction by LPS. In cells treated with LPS + C2, the enhanced COX-2 induction was also completely blocked by treatment with SB203580 (Fig. 3A). Cells transfected with JNK1() were used to test whether blockade of JNK cascade led to change in the COX-2 induction by LPS or LPS + C2. The stable JNK1() transfection experiment revealed that the JNK1 pathway was responsible for C2-potentiated induction of COX-2 but not the induction of COX-2 by LPS alone (Fig. 3B). PD98059, a chemical inhibitor of MKK1, decreased COX-2 induction by LPS (Fig. 3C, top). COX-2 expression was also decreased in cells stably transfected with MKK1(), which was consistent with the result of chemical inhibition. Interestingly, however, C2-potentiated COX-2 induction by LPS was not affected by PD90859 or stable transfection with MKK1() (Fig. 3C, bottom). These data provided evidence that p38 kinase was required for the inducible and potentiated COX-2 expression, whereas JNK1, but not ERK1/2, played a role in the enhanced COX-2 induction by C2.

View larger version (39K): [in this window] [in a new window]   Fig. 3.   The role of MAP kinases in C2-enhanced LPS induction of COX-2. A, the role of p38 kinase pathway. Cells were treated with SB203580 (SB, 10 µM) for 1 h and further incubated with LPS (0.1 µg/ml) or LPS + C2 (50 µM) for 20 h. Data represent the mean ± S.D. with at least three separate experiments (**, p < 0.01, significant compared with LPS alone; ##, p < 0.01, significant compared with LPS + C2). B, the role of JNK pathway. Raw264.7 cells or cells stably expressing JNK1() were used to assess LPS- or LPS + C2-inducible COX-2. C, the role of MKK1/ERK1/2 pathway. Cells were treated with PD98059 (PD, 30 µM) and subsequently exposed to LPS or LPS + C2. Cells stably expressing MKK1() were used to assess the role of MKK1 in C2 enhancement of LPS-inducible COX-2 expression. Data represent the mean ± S.D. of at least three separate experiments (**, p < 0.01, significant compared with LPS alone in Raw264.7 cells; ##, p < 0.01, significant compared with LPS alone in MKK1() cells). D, MAP kinase activation. Phosphorylation of p38 kinase (p38), JNK, or ERK in cell lysates was assessed by immunoblotting with the specific antibodies. Solid and open arrowheads indicate phosphorylated and unphosphorylated forms of MAP kinases, respectively. Results were confirmed by repeated experiments.

Effects of C2 on DNA Binding Transcription Factors in JNK1() Cells. Given the importance of the JNK pathway for C2-potentiated COX-2 induction, studies were extended to determine whether C2 changed activation of the transcription factors C/EBP, CREB, and NF-B in JNK1() cells. Gel-shift retardation analysis revealed that LPS-inducible band intensity of C/EBP transcription complex was not further increased by C2 in JNK1() cells (6 h) (Fig. 4A). By contrast, LPS-inducible CREB transcription complex (3 h) was potentiated by C2 in JNK1() cells (Fig. 4B). In JNK1() cells, C2 did not change LPS-inducible increase in NF-B DNA binding (data not shown). These data indicate that the JNK1 pathway controlled C2 enhancement in C/EBP DNA binding but not that of CREB or NF-B.

View larger version (40K): [in this window] [in a new window]   Fig. 4.   Gel-shift analyses of C/EBP and CREB in JNK1() cells. Raw264.7 cells stably transfected with JNK1() were cultured with LPS (0.1 µg/ml) or LPS + C2 (50 µM). Nuclear extracts prepared at the indicated times were subjected to gel-shift analysis of C/EBP (A) or CREB (B). All lanes contained 5 µg of nuclear extracts and 5 ng of labeled C/EBP or CREB consensus oligonucleotide. Arrowhead indicates C/EBP or CREB DNA binding complex.

Changes in the Nuclear C/EBP and  Levels by LPS + C2. In view of the importance of C/EBPs as the transcriptional factors responsible for COX-2 induction, we sought to determine the levels of nuclear C/EBPs in macrophages exposed to LPS with or without C2. LPS increased nuclear C/EBP isoforms (38, 35, and 19 kDa). In cells treated with LPS + C2, the level of nuclear 38-kDa C/EBP was increased to a greater extent (Fig. 5A). Conversely, the C/EBP isoform, which was increased by LPS at 3 to 12 h (a maximal increase at 6 h), was rather suppressed by concomitant C2 treatment compared with that of LPS alone at the respective time point. The levels of nuclear C/EBP and - forms were not changed (Fig. 5A).

View larger version (52K): [in this window] [in a new window]   Fig. 5.   The effects of C2 on the activation and expression of C/EBPs. A, Western blot analyses of nuclear C/EBP isoforms in Raw264.7 cells. The levels of C/EBP, C/EBP (LAP and LIP), C/EBP, and C/EBP were determined in the nuclear fractions of cells treated with LPS (0.1 µg/ml) or LPS + C2 (50 µM) for 1 to 12 h. B, immunocytochemistry of C/EBP or C/EBP in cells treated with LPS or LPS + C2. Distribution of C/EBP or C/EBP was also monitored in cells incubated with vehicle (dimethyl sulfoxide) or 50 µM C2 for 3 to 12 h. The same fields were counter-stained with PI. C, expression of C/EBP and C/EBP. The levels of C/EBP [LAP (liver activating protein) and LIP (liver inhibitory protein)] and C/EBP isoforms were determined in the lysates of cells treated with LPS (0.1 µg/ml) or LPS + C2 (50 µM) for 3 to 20 h. Each lane was loaded with 30 µg of cell lysates. The results in A, B, and C were confirmed by repeated experiments.

Abrogation of C2-Mediated C/EBP Activation by JNK1() Transfection. Next, we determined whether JNK1() transfection abolished C2-enhanced nuclear translocation of C/EBP. Immunoblot analysis showed that C2 (6-12 h) failed to increase LPS-inducible nuclear translocation of 38-kDa C/EBP in JNK1() cells (Fig. 6A). The decrease in nuclear translocation of C/EBP isoform by C2 was not affected by JNK1() transfection. Immunocytochemistry showed that LPS induced nuclear translocation of C/EBP in JNK1() cells that was slower, however, than in control cells (Fig. 6B). Consistent with the result of immunoblot analysis, C2 did not increase LPS-inducible nuclear translocation of C/EBP in JNK1() cells. These data provide evidence that the JNK1 pathway controls C2-mediated activation of C/EBP in LPS-treated cells.

View larger version (30K): [in this window] [in a new window]   Fig. 6.   The effects of C2 on the levels of nuclear C/EBP [LAP (liver activating protein) and LIP (liver inhibitory protein)] and C/EBP in JNK1() cells. A, the levels of nuclear C/EBP isoforms in JNK1() cells. Western blot analyses were performed with the nuclear fractions prepared from JNK1() cells treated with LPS (0.1 µg/ml) or LPS + C2 (50 µM) for 6 or 12 h. B, immunocytochemistry of C/EBP in JNK1() cells treated with LPS or LPS + C2. Nuclear translocation of cytosolic C/EBP was monitored in JNK1() cells, as described in Fig. 5B. The results were confirmed by repeated experiments. The same fields were counter-stained with PI.

Inhibition of C2-Potentiated COX-2 Induction by AC/EBP. Given the activation of C/EBP by LPS + C2, we assessed the role of C/EBP in the COX-2 gene expression. Constitutively active AC/EBP was expressed in cells before treatment with LPS + C2. Expression of AC/EBP completely inhibited the ability of C2 to enhance LPS induction of COX-2 (Fig. 7, top). Transfection with pCMV500, a control vector, allowed C2 to potentiate the induction of COX2 by LPS. Overexpression of AC/EBP reduced the extent of COX-2 induction in cells exposed to LPS alone (Fig. 7, bottom), confirming the fact that C/EBP was involved in LPS-inducible COX-2 expression.

View larger version (35K): [in this window] [in a new window]   Fig. 7.   Inhibition of LPS + C2-inducible COX-2 expression by AC/EBP. Raw264.7 cells transfected with the plasmid encoding for AC/EBP were exposed to LPS (0.1 µg/ml) or LPS + C2 (50 µM) for 20 h. Control cells were transfected with pCMV500. The relative COX-2 protein levels were measured by scanning densitometry of the immunoblots. Data represent the mean ± S.D. of at least three separate experiments [ **, p < 0.01, significant compared with LPS alone (i.e., 1).

Effect of C2 on LPS-Inducible AP-1 Activation. AP-1 is a heterodimer of Fos and Jun and is activated in cells challenged with LPS and other inflammatory insults (Tengku-Muhammad et al., 2000). The JNK pathway regulates AP-1 activation (Karin, 1995). The promoter region of the C/EBP gene contains the putative binding sites for the transcription factors AP-1, C/EBP, and CREB (Foka et al., 2001). As part of the efforts to determine how the JNK pathway controls C2-mediated enhanced activation of C/EBP in LPS-treated cells, we next assessed AP-1 activation. Gel-shift analyses revealed that LPS or C2 alone increased AP-1 DNA binding activity (6-12 h) (Fig. 8A). AP-1 was activated to a greater extent in cells exposed to LPS + C2 compared with LPS or C2 alone. Addition of a 20-fold excess of an unlabeled AP-1 binding oligonucleotide to the nuclear extract completely abolished the binding activity, whereas excess unlabeled SP-1 oligonucleotide failed to inhibit binding, suggesting that the binding protein is AP-1. We next monitored whether C2 activated AP-1 in JNK1() cells. AP-1 DNA binding activity was not potentiated by C2 in LPS-treated JNK1() cells (Fig. 8B). Hence, AP-1 activation was controlled by JNK1.

View larger version (48K): [in this window] [in a new window]   Fig. 8.   Gel-shift analysis of AP-1. A, gel-shift analysis of AP-1 with the nuclear extracts prepared from cells treated with LPS (0.1 µg/ml) or LPS + C2 (50 µM) for 3 to 12 h. Each lane contained 5 µg of nuclear extracts. The specificity of AP-1 binding was assessed by competition assays with 20-fold molar excess of unlabeled AP-1 or SP-1 oligonucleotide. Arrowhead, the retarded AP-1 band. B, gel-shift analysis of AP-1 with the nuclear extract prepared from JNK1() cells that were treated with LPS or LPS + C2 for 6 h. Results were confirmed by repeated experiments. C, the effect of AP-1 consensus decoy ODN on the expression of C/EBP by LPS or LPS + C2. Cells were incubated with 10 µM of AP-1 decoy ODN or mutant AP-1 decoy ODN for 1 h and further incubated with LPS or LPS + C2 for 12 h. C/EBP in cell lysates was immunoblotted.

C2 Potentiation of COX-2 Induction by TNF- or Type I Collagen. Studies were extended to assess whether C2 increased the expression of COX-2 by other inducers including TNF- and type I collagen. We found that type I collagen induces nitric-oxide synthase and COX-2 in macrophages (Cho et al., 2002; Y. H. Cho and S. G. Kim, unpublished observations). Western blot analysis revealed that the extent of COX-2 induction by TNF- or type I collagen was potentiated by concomitant C2 treatment (Fig. 9). C2 potentiation of COX-2 induction was also abolished in JNK1() cells (data not shown).

View larger version (19K): [in this window] [in a new window]   Fig. 9.   Enhancement of TNF-- or type I collagen-inducible COX-2 expression by C2 in Raw264.7 cells. Cells were treated with C2 (50 µM) in combination with TNF- (10 ng/ml) or type I collagen (10 µg/ml) for 20 h.

View larger version (21K): [in this window] [in a new window]   Fig. 10.   Schematic diagram illustrating the proposed mechanisms by which C2 enhances LPS-inducible COX-2 expression. JNK/AP-1-mediated C/EBP activation by C2 potentiates COX-2 induction by LPS. p38 kinase plays a role in the inducible and potentiated COX-2 expression.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Macrophages secrete inflammatory mediators, including lipid metabolites (e.g., prostaglandins) and cytokines. COX-2 catalyzes the inducible production of prostaglandins, which represents an important step in the inflammatory process (Wadleigh et al., 2000). The production of prostaglandins by LPS in macrophages is primarily caused by the transcriptional activation of the COX-2 gene (Lee et al., 1992; Reddy and Herschman, 1994). The cis-acting elements identified on the promoter region of the murine COX-2 gene include C/EBP, CREB, and NF-B (Caivano and Cohen, 2000; Caivano et al., 2001). The complex consisting of C/EBP homodimers is involved in the activation of C/EBP response element in macrophages exposed to LPS (Granger et al., 2000). If C/EBP is inactivated, the expression of COX-2 by LPS is impaired (Gorgoni et al., 2001). Although the initial phase of COX-2 expression by LPS involved CREB (Caivano et al., 2001), the activation of CREB was not sufficient to trans-activate the gene (Caivano and Cohen, 2000). In contrast, Eliopoulos et al. (2002) recently showed that Tpl2-dependent CREB activation signals regulate the induction of COX-2 by LPS in macrophages and that p38 kinase and ERK contribute to cell signaling. Despite the presence of the NF-B binding site in the regulatory region of the COX-2 gene, the putative NF-B was not required for the induction of COX-2 by LPS, as shown by dominant-negative inhibition of NF-B and COX-2 reporter gene activity (Wadleigh et al., 2000).

Ceramide signaling has been linked to inflammation and tumorigenesis. Because C2 is a cell-permeable analog, the agent has been widely used to study the pathophysiological effects of ceramide. C2 induced COX-2 in human epithelial carcinoma cells (Chen et al., 2001). We observed that C2 enhanced the extent of COX-2 induction by LPS in macrophages, although C2 alone did not induce the enzyme. C/EBP, which was activated by LPS + C2, was apparently involved in the induction of COX-2, whereas CREB might indirectly affect the enzyme expression. The present study showed that the expression of C/EBP in total cell lysates was increased by LPS and to a greater extent by LPS + C2. The expression of C/EBP is regulated by CREB (Belmonte et al., 2001). The increase in CREB DNA binding by LPS + C2 at earlier times may stimulate the C/EBP expression. We found in additional that C2 potentiated LPS-inducible AP-1 DNA binding activity. Because the promoter region of C/EBP contains putative AP-1 binding sites, the enhanced AP-1 activation by C2 is highly likely to be responsible for the increase in C/EBP expression and thus for C/EBP-mediated induction of COX-2. This is supported by blockage of LPS + C2-inducible C/EBP expression in cells treated with AP-1 decoy ODN. In addition to the increase in the C/EBP expression, C2 elicited its nuclear translocation in LPS-treated cells. Competition experiments using the specific antibodies revealed that LPS induced C/EBP DNA binding activity and that the binding complex comprised C/EBP and  forms. The nuclear translocation of C/EBP was increased by C2 in LPS-treated cells, whereas that of C/EBP was rather decreased. This raised the notion that C/EBP form plays an important role in the transcriptional activation of the COX-2 gene. Increase in the nuclear translocation of C/EBP and the altered ratio of nuclear C/EBP to C/EBP by C2 would contribute to C2-potentiated COX-2 induction. This is consistent with the observation that the increase in C/EBP homodimer complex trans-activated the COX-2 gene expression (Granger et al., 2000). In the present study, the crucial role of C/EBP in the enhancement of COX-2 induction was further supported by the experiment using dominant-negative mutant of C/EBP.

NF-B is activated by oxidative stress and/or inflammation. Activation of the NF-B complex is related with the cellular redox state (Hirota et al., 1999). The intracellular thiol level changes the expression of several genes after early activation of NF-B (Parmentier et al., 2000). In a previous study from another laboratory (Hsu et al., 2001), C2, when present with LPS, inhibited LPS-induced NF-B activation. Hsu et al. (2001) proposed the hypothesis that the inhibition of LPS-mediated IKK, p38 kinase, PKC, NF-B, and AP-1 activation by C2 may lead to COX-2 inhibition (2001). However, another study showed that activation of NF-B was not responsible for the induction of COX-2 (Wadleigh et al., 2000). We found that NF-B was not further activated by C2, which was consistent with the notion that the putative NF-B was not required for COX-2 induction by LPS in macrophages (Wadleigh et al., 2000). Because all of the experimental conditions in the present study were identical to those of Chen et al. (2001), the discrepancy implicating the role of ceramide in the induction of COX-2 is mystery.

Studies have shown that ceramide activates MAP kinases, including p38 kinase, JNK, and ERK1/2 (Chen et al., 2001). The pathways of the MAP kinases, in particular p38 kinase and ERK1/2, regulate the expression of COX-2 (Nagano et al., 2002). Hence, the signaling pathway of ceramide is likely to produce cross-talk with that for COX-2 expression. In the present study, we confirmed that the MAP kinases controlled the expression of COX-2 by LPS. p38 kinase played a critical role in LPS-inducible or C2-potentiated COX-2 induction, as evidenced by complete blockage of COX-2 induction by p38 kinase inhibition. The ERK1/2 pathway was involved in the induction of COX-2 by LPS. However, the inhibition of the MKK/ERK1/2 failed to suppress C2 enhancement in COX-2 induction by LPS, as evidenced by the experiments using a specific chemical inhibitor of MKK1 or cells stably expressing MKK1(). Hence, the ERK1/2 pathway was not responsible for the C2 potentiation of COX-2 induction by LPS. It is highly likely that the MAP kinase signaling pathways responsible for COX-2 induction switch in cells exposed to C2 in combination with LPS. In the present study, we showed for the first time that the JNK pathway, which is not involved in the COX-2 induction by LPS alone, contributes to C2 potentiation of COX-2 induction instead of the ERK pathway (Fig. 11).

We found that JNK1() transfection abrogated both enhanced C/EBP DNA binding and C/EBP nuclear translocation by C2. In contrast, the decrease in the translocation of C/EBP by C2 was not restored by JNK1() transfection, which indicated that the JNK1 pathway was not responsible for the reduced translocation of C/EBP. We conclude that activation of C/EBP, but not C/EBP, is dependent on JNK1. It has been reported that c-Jun mediates the induction of COX-2 by ceramide or sphingomyelinase (Subbaramaiah et al., 1998), and enhanced phosphorylation of c-Jun in combination with c-Fos potentiates its ability to activate transcription of the responsible genes (Karin, 1995). C/EBP and other transcription factors (e.g., c-Jun) may be involved in the formation of enhanceosome protein complex for trans-activation of the COX-2 gene.

We observed that the expression of C/EBP was enhanced by LPS + C2, compared with LPS or C2 alone, and that the enhanced C/EBP expression was controlled by JNK. JNK has been proposed to play a role in AP-1-mediated transcriptional activation of the target genes (Karin, 1995). The promoter region of the C/EBP gene contains the AP-1 binding sites (Foka et al., 2001). Thus, JNK-mediated AP-1 activation by C2 in combination with LPS may contribute to the transcriptional activation of the gene. Our observation has an implication for the finding of AP-1 as a putative transcriptional factor responsible for enhanced C/EBP expression (Fig. 11). Hence, the pathway involving JNK may be responsible for the greater induction and activation of C/EBP by C2 in LPS-treated cells and may consequently lead to C2-potentiated induction of COX-2.

Activation of JNK is an early cellular response after exposure to a variety of stressors such as heat, UV irradiation, DNA damaging agents, and osmotic shock (Adler et al., 1995; Rosette and Karin, 1995), which represents separate stress-activated apoptotic pathway (Leppa and Bohmann, 1999). Ceramide signaling regulates the expression of the genes implicated in inflammation. We found that C2 also enhanced TNF--inducible COX-2 expression. We speculate that the induction of COX-2 by inflammatory mediators would be potentiated in cells whose JNK and ceramide pathways are both activated by stressors (e.g., apoptotic cells). A previous study showed that bovine type I collagen induced nitric-oxide synthase (Cho et al., 2002) and COX-2 (Y. H. Cho and S. G. Kim, unpublished observations). We observed that C2 enhanced COX-2 induction by bovine type I collagen. Hence, C2 was capable of enhancing COX-2 induction by other inducers. Thus, the ceramide-signaling pathway significantly contributes to the induction of COX-2 by LPS and other inducer(s) in macrophages.

In summary, the present study demonstrates that C2 potentiates C/EBP activation and COX-2 induction by LPS, that the pathway of JNK1, but not ERK1/2, is responsible for C/EBP-mediated potentiation of COX-2 induction (which involves AP-1-mediated C/EBP expression), and that p38 kinase plays an essential role in the inducible and potentiated COX-2 expression.