Extracellular-Regulated Kinase 1/2, Jun N-Terminal Kinase, and c-Jun Are Involved in NF-κB-Dependent IL-6 Expression in Human Monocytes¹

Expression of IL-6 in monocytes is mainly regulated at the transcriptional level (1, 2, 3, 4, 5, 6, 7, 8). In the promoter region of the IL-6 gene five known functional cis-regulatory elements have been described, namely cAMP-responsive element, activator protein-1 (AP-1),⁴ nuclear factor IL-6 (NF-IL6), Sp1, and NF-κB binding sites (8). The transcription factor NF-κB has been shown to be a critical regulator of IL-6 gene transcription (5, 6, 7). Within the 1.2-kb fragment of the 5′-flanking region of the IL-6 gene the functional NF-κB element is located between positions −73 and −63 (5, 6, 7). NF-κB complexes consist of members of the rel multigene family, which is comprised of five major proteins: p50, p65 (Rel A), c-Rel, p52, and Rel B (reviewed in 8 and 9). The most abundant dimer is the p50/p65 NF-κB heterodimer. NF-κB is located in the cytoplasm in an inducible form, in which the heterodimer is complexed to the inhibitory subunit, IκBα. Upon stimulation of the cell, IκBα is rapidly phosphorylated and degraded; the released nucleophilic heterodimer then moves to the nucleus. Both p50 and p65 contribute to NF-κB DNA binding, but only the p65 subunit is responsible for transactivation. Although simultaneous activity of NF-κB and additional transcription factors such as AP-1 and NF-IL6/CEBP, is required to maximally induce IL-6 gene transcription, the contribution of the κB regulatory element to the transcriptional activation of the IL-6 gene appears most prominent (5, 6, 7, 8, 10, 11).

Much progress has been made with respect to the identification of signal transduction pathways involved in cytokine gene expression. Of great importance are the so-called mitogen-activated protein kinase (MAPK) proteins (reviewed in 12 and 13). Activation of the different MAPK signaling pathways ultimately results in the direct or indirect phosphorylation and activation of various transcription factors and alterations in gene expression. Although the involvement of the MAPK pathways in IL-6 gene regulation in human monocytes is far from elucidated, there are several observations from other cell systems that may indicate a functional role for the individual kinase pathways in activating the transcription factor NF-κB. For instance, in COS cells p90^RSK1, the downstream kinase of Raf-1/MEK1/ERK1/2 pathway, has been shown to phosphorylate the N-terminal regulatory domain of IκBα both in vitro and in vivo only after 12-O-tetraphorbol 12-myristate 13-acetate stimulation, and thus enhance NF-κB DNA-binding (14). In agreement, in the lymphoblastoid cell line CEM, overexpression of either MEK1 or ERK1 demonstrated constitutive nuclear localization of NF-κB, suggesting the involvement of the classical ERK pathway in NF-κB DNA binding activity (15). The c-Jun NH₂-terminal kinase/stress-activated protein kinase (JNK/SAPK) pathway can be activated by small GTP-binding proteins, including cdc42 and Rac1 (16). Subsequently, the downstream kinase MEKK1 is activated, which, in turn, activates MKK4, the upstream kinase of JNK (17, 18). Among the substrates of JNK are c-Jun, activating transcription factor-2, and Elk-1. Recently, it was shown that overexpression in T cells of an active form of MEKK1, a constitutive activator of JNK, results in the direct activation of the IκBα kinase, the degradation of IκBα, and the activation of an NF-κB reporter gene (19). Moreover, in HeLa cells MEKK1 was shown to directly activate the IκBα kinase (IKK) complex in vitro, which, through the phosphorylation of IκB, results in the activation of NF-κB (20). These observations indicate a link between NF-κB and the JNK signaling cascade. The question remains, however, whether the above mentioned signaling pathways are also operational in monocytes with respect to κB-mediated IL-6 gene transcription.

In the present study we wanted to further identify the involvement of the ERK1/2 and JNK pathways in mediating κB-dependent IL-6 gene expression in human monocytes. One tool in studies of the importance of phosphorylation states of signaling molecules is the use of specific pharmacologic inhibitors of protein phosphatases (PPases). Opposite of the protein kinases, the PPases regulate the phosphorylation state of key proteins by removing the phosphate group from the tyrosine, serine, or threonine residues of activated proteins. Okadaic acid (OA), a polyether fatty acid, is a potent inhibitor of the phosphoserine and phosphothreonine phosphatases PP1 and PP2A (21). OA is known to be a potent stimulator of IL-1 and TNF-α production in various cell types, including monocytic cell lines (22, 23). We investigated the effect of OA on the ERK1/2 and JNK signaling pathways and their involvement in NF-κB-driven IL-6 expression in human monocytes. Our results provide evidence for the involvement of the Ras/Raf-1/ERK1/2 pathway, the JNK pathway, and the c-Jun protein in NF-κB-mediated IL-6 gene regulation in human monocytes.

RPMI 1640 medium was purchased from Flow Laboratories (Rockville, MD), FBS was obtained from HyClone (Logan, UT), and Lymphoprep was obtained from Nycomed (Oslo, Norway).

OA was obtained from Sigma (St. Louis, MO). Radionucleotides were obtained from Amersham (Aylesbury, U.K.). Abs against ERK1, ERK2, and JNK1 were obtained from Santa Cruz Biotechnology (Santa Cruz, CA), and the anti-CD14 was obtained from Becton Dickinson (Sunnyvale, CA). The MEK1 inhibitor PD098059 was purchased from New England Biolabs (Beverly, MA). cDNA probe for IL-6 was provided by Dr. L. Aarden (Central Laboratory of The Netherlands Red Cross Blood Transfusion Service, Amsterdam, The Netherlands). GST-c-Jun fusion protein was provided by Dr. J. Borst (The Netherlands Cancer Institute, Amsterdam, The Netherlands). ELISA kits for IL-6 (Pelikine Compac kits) were purchased from CLB (Amsterdam, The Netherlands), and anti-TNF-α was obtained from Boehringer Mannheim (Mannheim, Germany). The luciferase detection kit was purchased from Promega (Madison, WI), and IL-3 was received from Genetics Institute (Cambridge, MA). All IL-6 promoter constructs, containing various portions of the human IL-6 promoter cloned into the pGL₂ luciferase reporter vector (Promega), were generated in our laboratory. pGAL4-p65 was provided by Dr. P. A. Baeuerle (Tularik, San Francisco, CA); pGAL4-dbd and pGAL4tkluc were provided by Dr. S. Wissink (NIOB, Utrecht, The Netherlands). pRSV-NΔRaf1 was provided by Dr. P. J. Coffer (Department of Pulmonary Diseases, University Hospital Utrecht, Utrecht, The Netherlands), pcDNA3-Flag-JNK was provided by Dr. R. J. Davis (Howard Hughes Medical Institute, University of Massachusetts Medical School, Worchester, MA). pCMV-TAM67 was provided by Dr. M. J. Birrer (Biomarkers and Prevention Research Branch, National Cancer Institute, National Institutes of Health, Rockville, MD). The plasmid p(NF-κB)3xCAT was provided by Dr. J. Hiscott (The Sir Mortimer B. Davis-Jewish General Hospital, Montreal, Canada).

Peripheral blood cells were obtained from healthy volunteer platelet donors, and mononuclear cell suspensions were prepared by Ficoll-Hypaque density-gradient centrifugation. T lymphocytes were depleted by 2-aminoethylisothiouronium bromide-treated SRBC rosetting. Monocytes were further enriched by plastic adherence (1 h, 37°C) and demonstrated a purity of >95%, detected by FACS analysis with anti-CD14 Ab. Monocytes were cultured at 37°C at a density of 1–2 × 10⁶/ml in RPMI 1640 supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin, 6 ng/ml colistine, and the appropriate amount of FBS. The cell line employed in these studies was TF-1, a human erythroleukemia cell line (24). The TF-1 cell line was cultured in RPMI 1640 supplemented with 5% FBS and 10 ng/ml IL-3. The toxicity of OA was assessed by trypan blue exclusion assays.

Freshly isolated monocytes (1 × 10⁶ cells in 1 ml) were incubated in RPMI 1640 with 2% FBS. Sixteen hours after the adherence step, medium was replaced by 1 ml of fresh RPMI 1640 containing 2% FBS, and cells were subsequently stimulated. Twenty-four hours after treatment with medium or OA (30 ng/ml) the cell-free supernatants were collected and analyzed for secreted IL-6 protein by ELISA. Experiments with the MEK1 inhibitor PD98059 were performed by pretreatment of monocytes for 30 min with 10 μM PD98059 before stimulation with medium or OA (25). Similarly, monocytes were preincubated with TNF-α or IL-1β Ab before OA stimulation (1/1000 diluted polyclonal antiserum against TNF-α or IL-1β was maximally effective, as studied with in vitro dose-response curves).

The run-on analysis was performed as previously described using 40 × 10⁶ monocytes/reaction (26). Freshly isolated monocytes were cultured in RPMI 1640 medium supplemented with 2% FBS for 16 h before stimulation with medium or medium plus OA for 1 or 7 h (30 ng/ml).

Five micrograms of the following DNAs were immobilized on Hybond N⁺ membranes (Amersham): 1) EcoRI-linearized pGEM (negative control), 2) EcoRI-linearized plasmid containing a 1.3-kb fragment of the rat GAPDH cDNA, and 3) EcoRI-linearized plasmid containing a 0.3-kb fragment of the human IL-6 cDNA. Hybridization of labeled RNAs to these membranes was performed at 65°C for 18 h in 0.5 mol/l Na₂HPO₄ (pH 7.2), 1 mM EDTA, and 7% SDS. Membranes were then washed as described previously (26) and exposed to Kodak X-OMAT XAR films (Eastman Kodak, Rochester, NY) at −80°C using an intensifying screen. Quantitation of the newly synthesized RNAs was performed by phosphorimaging (PhosphorImager, Molecular Dynamics, Sunnyvale, CA), after normalization to the GAPDH signal.

After overnight culture in RPMI 1640 supplemented with 2% FBS monocytes were stimulated, and nuclear extracts were prepared according to the miniscale procedure previously described (27). Nuclear extracts were divided into small aliquots and were stored at −80°C. Double-stranded synthetic oligonucleotide probes containing the NF-κB consensus sequences (NF-κB, 5′-AGCTGCGGGGATTTTCCCTG-3′) was used in the gel retardation assay. The consensus sequence for binding of the nuclear factor is underlined. Fifty nanograms of HPLC-purified single-stranded oligonucleotide was labeled with T4-polynucleotide kinase and [γ-³²P]ATP (3000 Ci/mmol; Amersham), separated from nonincorporated radiolabel by Sephadex G-50 chromatography, ethanol precipitated, dried, and dissolved in 20 μl of 10 mM Tris-HCl (pH 7.5), 50 mM NaCl, 10 mM MgCl₂, 1 mM EDTA, and 1 mM DTT, containing a 4-fold excess of the opposite strand. Annealing of the two strands was performed by heating the mixture for 2 min at 90°C and slow cooling to room temperature. Five micrograms of nuclear extract and 0.1 ng of double-stranded labeled oligonucleotide were incubated in 20 mM HEPES (pH 7.9), 60 mM KCl, 0.06 mM EDTA, 0.6 mM DTT, 2 mM spermidine, and 10% glycerol supplemented with 2 μg poly(dI-dC). The binding reaction was performed at 26°C for 25 min. Competition experiments were performed by adding a 100-fold molar excess of unlabeled self or nonself double-stranded oligonucleotides. Supershift experiments were performed by incubating the nuclear extracts and labeled oligonucleotide with polyclonal Abs (1 μg) against the p50, p65, c-Rel, and p52 subunits of NF-κB (Santa Cruz Biotechnology). The samples were loaded on prerun (30 min, 100 V) 4% (30/1) polyacrylamide gels and run for 1 h at 150 V in 0.5× TBE at room temperature. Gels were dried and exposed to Kodak XAR films at −80°C with an intensifying screen. Quantification of protein binding was performed by densitometry using a PhosphorImager (Molecular Dynamics).

The luciferase reporter plasmids pIL6luc(−602), pIL6luc(−122), and pIL6luc(−60), containing various portions of the human IL-6 promoter cloned into the pGL₂ luciferase reporter vector (Promega), were transfected into TF-1 cell line by means of electroporation. Before transfection, cells were cultured for 16 h at a density of 0.5 × 10⁶ cells/ml in the appropriate medium, washed twice, and resuspended in RPMI 1640 at a density of 10 × 10⁶ in 200 μl. When transfected with a single plasmid 25 μg of DNA was added, and the mixture was left at room temperature for 15 min. Cotransfections were performed with 15 μg of the reporter plasmid (pIL6luc(−602), pIL6luc(−122), or p(NF-κB)3xCAT) together with 15 μg of the expression plasmid (pRSV-NΔRaf1, pcDNA3-Flag-JNK1, pCMV-TAM67, or pcDNA3 empty vector). Cotransfections of pGAL4tkluc (5 μg) with either pGAL4-p65 (5 μg) or pGAL4-dbd (5 μg) were performed under similar conditions. Electroporation, in 0.4-cm electroporation cuvettes, was performed at 240 V and 960 μF with a Gene Pulser electroporator (Bio-Rad, Richmond, VA). After electroporation the cells were replated in RPMI 1640 containing 2% FBS. Six hours after transfection cells were stimulated for 24 h with medium or OA (30 ng/ml). The cells were then harvested and lysed by commercially available luciferase lysis buffer. One hundred microliters of lysis product were added to 100 μl of luciferase assay reagents, and luciferase activity was measured with the Anthos Lucy1 luminometer (Anthos Labtec Instruments, Salzburg, Austria).

Phosphorylation of ERK1 and ERK2 was analyzed by Western blotting. Briefly, monocytes were cultured for 16 h in RPMI 1640 containing 0.1% FBS and subsequently stimulated for various periods of time with medium or OA (30 ng/ml). After harvesting, total cell extracts were prepared by resuspending the cells in 500 μl of 1× sample buffer (containing 2% SDS, 10% glycerol, 2% 2-ME, 60 mM Tris-Cl (pH 6.8), and bromphenol blue) and lysing the cells by passing them through a 23G1 needle (three times). Cell extracts were directly boiled for 5 min and were stored at −20°C. Before loading, samples were again boiled for 5 min, and cell extracts were resolved by running 0.1 vol on a SDS-12.5% PAGE gel (acrylamide/bisacrylamide, 173/1) and transferred to cellulose nitrate membrane (Schleicher & Schuell, Germany). Immunoblotting with anti-ERK2 Ab was performed by standard procedures, and detection was performed according to the manufacturer’s guidelines (ECL, Amersham). After ERK2 band-shift detection the membrane was stripped (strip buffer containing 1× PBS, 0.1% Tween, and 0.1% SDS) and rehybridized with anti-ERK1 Ab.

Bacterial expression of GST-c-Jun has been described previously (28). Briefly, GST fusion proteins were expressed in Escherichia coli DH5α, induced with 1 mM isopropyl-β-d-thiogalactopyranoside and purified with glutathione-Sepharose beads.

JNK kinase activity was determined by the ability of this enzyme to phosphorylate its substrate proteins c-Jun in the presence of [γ-³²P]ATP. Before stimulation monocytes were cultured for 16 h in RPMI 1640 plus 0.1% FBS. Cells (15 × 10⁶) were then treated for various periods of time with medium or OA (30 ng/ml). After stimulation, cells were harvested and resuspended in 400 μl of lysis buffer (20 mM HEPES (pH 7.4), 2 mM EGTA, 50 mM β-glycerophosphate, 1 mM DTT, 1 mM Na₃VO₄, 1% Triton X-100, 10% glycerol, 10 μg/ml soybean trypsin inhibitor, 10 μg/ml leupeptin, and 0.4 mM PMSF) and lysed on ice for 15 min. After centrifugation at 3000 rpm and normalization for protein content in the supernatant, lysates were incubated with anti-JNK Ab in a total volume of 500 μl and rotated for 30 min at 4°C. Twenty-five microliters of a slurry of 50% protein A-Sepharose beads was then added to the lysate/Ab mixture and left rotating overnight at 4°C. Subsequently, the Ab/beads/conjugate mixture was washed with lysis buffer (three times), twice with LiCl buffer (500 mM LiCl, 100 mM Tris-Cl (pH 7.6), 0.1% Triton X-100, and 1 mM DTT), and finally with buffer A (three times; 20 mM MOPS (pH 7.2), 2 mM EGTA, 10 mM MgCl₂, 1 mM DTT, and 0.1% Triton X-100). After supernatant was completely removed, the reaction was initiated by adding 50 μl of reaction mix (43.5 μl of buffer A, 20 mM MgCl₂, 25 μM ATP, 3.5 mg/ml GST-c-Jun, and 10 μCi of [γ-³²P]ATP), and incubation took place at 20°C for 30 min. The reaction was terminated by adding 15 μl of sample buffer/2-ME (four times). Before resolving samples on SDS-PAGE gel, samples were boiled for 5 min. A m.w. marker was used to assess the correct protein size. After running, the gel was washed in water for 15 min, in 5% TCA/1% sodium pyrophosphate (twice, 2 h each time), and again in water, and subsequently the gel was dried. Phosphorylated substrate was visualized by autoradiography and was quantified by densitometric scanning.

Freshly isolated human monocytes were stimulated with the serine/threonine-specific PPase inhibitor OA and after 24 h of stimulation IL-6 protein levels were measured in cell-free supernatants. Treatment of monocytes with OA (30 ng/ml) resulted in distinct increases in the secretion of IL-6 protein (1898 ± 1164 vs 4 ± 4 pg/ml for unstimulated control; p < 0.005; Fig. 1). Treatment of the cells with higher concentrations of OA, up to 200 ng/ml, induced even higher levels of IL-6 protein. These concentrations, however, appeared toxic after 24 h of stimulation. Additional experiments were thus performed with 30 ng/ml OA.

OA has been demonstrated to induce the expression of TNF-α from human monocytes (23). To examine whether the effects of OA on IL-6 secretion are mediated directly or indirectly through the release of TNF-α, Ab against this cytokine was added to the cell culture. Fig. 1 illustrates that treatment of monocytes with OA plus anti-TNF-α resulted in a partial down-regulation of IL-6 secretion compared with that after treatment with OA alone (1104 ± 628 pg/ml for OA plus anti-TNF-α vs 1898 ± 1164 pg/ml for OA alone; p < 0.005), indicating that the secretion of IL-6 protein by OA is only partly an indirect phenomenon mediated by the release of TNF-α. Addition of an anti-IL-1β Ab had no effect (data not shown).

To determine whether the OA-mediated increase in IL-6 protein secretion was established at the transcriptional level, nuclear run-on studies were performed. Monocytes were treated with medium or medium plus OA (30 ng/ml) for 7 h, after which the IL-6 in vitro transcription rate was determined. Results demonstrated that after normalization for the GAPDH signal, OA enhanced the IL-6 transcription rate 3.2-fold after 7 h of stimulation (Fig. 2). The increase in the in vitro IL-6 transcription rate was quantified by phosphorimaging. The up-regulated IL-6 transcription rate was reflected at the level of mRNA. Treatment of monocytes with OA (30 ng/ml) for various periods of time (0–10 h) resulted in enhanced IL-6 mRNA levels, as determined by Northern blot analysis (data not shown). IL-6 mRNA was first observed after 6 h of stimulation with OA. The slow activation kinetics of OA, as observed in the nuclear run-on and Northern analysis, are in agreement with previous reports describing slow activation of both TNF-α and IL-1β gene expression after OA stimulation (29, 30).

To delineate the role of the transcription factor NF-κB in OA-mediated IL-6 expression we performed EMSAs using nuclear extracts from OA-stimulated monocytes. In unstimulated monocytes binding of the NF-κB complex was low (Fig. 3, lane 1), whereas stimulation with OA (30 ng/ml) for 4 or 6 h resulted in a distinct increase in NF-κB binding activity (Fig. 3, lanes 2 and 4, respectively). Exposure to OA for shorter periods did not enhance NF-κB binding activity (data not shown). Supershift experiments performed with Abs specific for the p50 and p65 subunits of the NF-κB family indicated that in monocytes the NF-κB complex, which binds to the κB-specific binding sequence, indeed consisted of the p50 and p65 subunits, both before and after OA exposure (Fig. 4,A, lanes 1–3 and 4–6, respectively). It is well established that NF-κB consisting of p50 and p65 can be transcriptionally active (8, 9). Two other members of the NF-κB family, c-Rel and p52, were not identified in the OA-induced NF-κB complex (Fig. 5).

The MAPK signaling cascade involving Raf-1, MEK1, and ERK1/2 phosphorylation is probably an important signal transduction pathway leading to cytokine gene expression (12, 13). To study the effects of OA on activity of this MAPK pathway, we measured the phosphorylation of ERK1 (p44^MAPK) and ERK2 (p42^MAPK). Monocytes were treated with OA (30 ng/ml) for various periods of time and were subsequently analyzed for phosphorylated ERK1 and ERK2 by means of immunodetection. As shown in Fig. 6, OA induced a shift in the mobility of both ERK1 and ERK2, indicating phosphorylation of the respective proteins. The observed shift, however, demonstrated slow kinetics. The ERK1 shift was first seen after 30 min of stimulation with OA (data not shown), and phosphorylation peaked at 90 min of stimulation. The ERK2 shift was first observed after 60 min of stimulation, and an optimal phosphorylation state was found at 90 min. In contrast, PMA-mediated ERK phosphorylation peaked as early as 5 min after addition of the stimulus. In view of the slow activation kinetics of the ERKs after OA treatment, we treated monocytes with OA and anti-TNF-α Ab simultaneously and subsequently analyzed for ERK phosphorylation. Anti-TNF-α coincubation did not affect the OA-induced ERK shifts (data not shown). In this respect, delayed onset of OA-induced activity was described previously for OA-mediated ERK activity in Jurkat T cells and B lymphocytes (31, 32). A relevant explanation for the delayed onset is that OA elicits its effect by disrupting the balance between low, basal protein phosphorylation and basal protein dephosphorylation in favor of protein phosphorylation. Signaling modules will be subsequently activated only after a relatively long incubation time.

The JNKs are, like the ERKs, members of the MAPK family and are also known as SAPK (12, 13). We determined activation of the JNK pathway by measuring JNK in vitro kinase activity as a result of OA treatment. After monocytes were treated with OA (30 ng/ml) for various periods of time, JNK1 was immunoprecipitated, and equal amounts of protein were subsequently assayed for in vitro kinase activity toward its specific substrate (c-Jun) in the presence of radiolabeled ATP. As depicted in Fig. 7, OA enhanced JNK1 activity as early as 15 min after stimulation. JNK1 kinase activity peaked at 90 min of stimulation with OA.

To summarize, we demonstrated that OA up-regulated IL-6 protein and mRNA levels in monocytes, which coincided with enhanced binding activity of NF-κB. Furthermore, the MAP kinases ERK1, ERK2, and JNK1 were activated in response to OA treatment. The question remains, however, whether the OA-induced phosphorylation of ERK1, ERK2, and JNK1 is directly involved in the up-regulation of IL-6 gene expression by OA. To answer this question with regard to the ERK1/2 pathway we treated monocytes with the Raf-1/MEK1/ERK pathway-specific pharmacological inhibitor PD98059. At a concentration of 10 μM PD98059 the activity of MEK1, the kinase upstream from ERK1 and ERK2, was almost totally inhibited (25). In our experiments PD98059 dramatically down-regulates OA-induced IL-6 secretion (100% for OA-induced IL-6 secretion vs 7.3 ± 4.1% for PD98059- plus OA-mediated IL-6 secretion; p < 0.005; Fig. 8), indicating that OA-enhanced phosphorylation of ERK1 and ERK2 is an important step in the elevation of IL-6 protein levels. PD98059 at this concentration was not toxic for the monocytes, as determined by trypan blue exclusion assays and lactate dehydrogenase levels in culture supernatants (data not shown) (33).

Band-shift assays using nuclear extracts derived from monocytes pretreated with PD98059 for 30 min before OA stimulation for 6 h demonstrated no considerable change in the binding activity of NF-κB (or of AP-1 or NF-IL6; data not shown), as determined after phosphorimaging quantification (Fig. 3, lanes 3 and 5). Thus, the ERK pathway did not play a role in DNA binding activity of this transcription factor. We, subsequently, demonstrated that the observed difference between inhibited IL-6 protein secretion and the lack of effect regarding NF-κB binding activity by PD98059 could not be explained by a change in composition of the NF-κB complex. The NF-κB complex consisted of p50 and p65 subunits both with and without PD98059 pretreatment (Fig. 4 A, lanes 4–9).

IL-6 transcription was investigated in more detail by transiently transfecting the TF-1 cell line with three 5′-end deletion constructs derived from the human IL-6 promoter linked to the luciferase reporter gene, pIL6luc(−602), pIL6luc(−122), and pIL6luc(−60) (Fig. 9,A). Results from the transfection experiments demonstrated that, in concordance with the observations in monocytes, OA induced activity of the entire IL-6 promoter in TF-1 cells (Fig. 9 B). After transfection with the largest promoter construct pIL6luc(−602), which contains the binding sites for both AP-1 and NF-κB, OA (30 ng/ml) up-regulated IL-6 promoter activity 7.8 ± 4.8-fold (p < 0.05). When pIL6luc(−112), which merely contains the NF-κB binding site, was introduced into TF-1 cells, OA elicited a 14.3 ± 10.0-fold (p < 0.05) induction in IL-6 promoter activity. This would imply possible negative regulatory cis elements in the promoter region downstream from the NF-κB site, within the region −122 through −602. Using a promoter construct lacking the binding sites for any of the transcription factors (pIL6luc(−60)), a minimal residual induction of the promoter activity was observed after OA treatment (1.5 ± 0.3-fold). These results indicated that OA is a potent inducer of IL-6 gene transcription, and furthermore, that transactivation by the NF-κB complex greatly contributes to OA-induced IL-6 gene transcription.

To confirm that OA can indeed directly enhance the transcriptional activity of NF-κB, TF-1 cells were transfected with the pGAL4-p65 or pGAL4-dbd constructs in combination with the reporter plasmid pGAL4tkluc (34). The GAL4-transactivator fusion proteins are exclusively nuclear and are regulated independently of IκB. The reporter plasmid pGAL4tkluc is under the control of multiple GAL4 binding sites. This GAL4 one-hybrid technique allows analysis of the κB p65 transactivation mechanism. The results demonstrated that OA enhanced p65-mediated transactivation by 2.2 ± 1.2-fold (p < 0.05), indicating a direct effect of OA on transcriptional activity of the p65 subunit of the NF-κB complex (Fig. 9 C). Transfection with the negative control plasmid pGAL4-dbd showed no enhanced promoter activity after OA treatment.

Since previous reports on the transcriptional activation of the IL-6 gene have emphasized the indispensability of the NF-κB binding site for the activity of the IL-6 promoter, we focussed on the involvement of the ERK1/2 and JNK1 pathways in NF-κB-dependent IL-6 promoter activity (5, 6, 7). For this purpose TF-1 cells were transiently transfected with pIL6luc(−122), comprising the NF-κB binding site only, together with either pRSV-NΔRaf1 or pcDNA3-Flag-JNK1, before stimulation with OA. Previously, pRSV-NΔRaf1, encoding a dominant negative mutant of Raf-1, was shown to block mitogen-activated protein kinase activity by growth factors (35). pcDNA3-Flag-JNK1 encodes a catalytically inactive mutant of JNK1 (36). The results of these experiments are depicted in Fig. 10 A. Both pRSV-NΔRaf1 (10 ± 8% for pRSV-NΔRaf1 vs 100% for pcDNA3; p < 0.05) and pcDNA3-Flag-JNK1 (54 ± 32% for pcDNA3-Flag-JNK1 vs 100% for pcDNA3; p < 0.05) inhibited NF-κB-mediated IL-6 promoter activity. Moreover, cotransfections of the pCMVluc plasmid, containing a constitutive promoter, with either pRSV-NΔraf1 or pcDNA3-Flag-JNK1 did not affect constitutive promoter activity. These observations emphasize the specificity of the dominant negative MAP kinase constructs in inhibiting IL-6 promoter activity.

To confirm that the findings were indeed correlated with the transactivation capacity of NF-κB itself and not with possible regulatory elements in the vicinity of the NF-κB binding site, TF-1 cells were simultaneously transfected with pRSV-NΔRaf1 or pcDNA3-Flag-JNK1 and p(NF-κB)3xCAT. This CAT reporter plasmid solely contains three consecutive GGGAAAATCC κB binding sites adjacent to the SV40 promoter. Again, it was demonstrated that both pRSV-NΔRaf1 and pcDNA3-Flag-JNK1 impaired OA-induced κB-mediated gene transcription (58 ± 11% for pRSV-NΔRaf1 vs 100% for pcDNA3 (p < 0.05) and 54 ± 24% for pcDNA3-Flag-JNK1 vs 100% for pcDNA3 (p < 0.05); Fig. 10 B). Taken together, these data would implicate a role for both the Raf-1/MEK1/ERK and the MEKK1/JNK pathways in enhancing the NF-κB transactivation capacity.

With the described results we demonstrated that both the MEKK1/JNK and Raf-1/MEK1/ERK pathways are involved in enhancing NF-κB-mediated promoter activity, but the latter pathway does not affect OA-induced NF-κB DNA binding. This observed difference would imply an activation signal different from mere binding of NF-κB to its consensus sequence. To resolve whether serine kinase activity may be involved in this process, transiently transfected TF-1 cells with pIL6luc(−122) were treated with the serine kinase-specific inhibitor H7 30 min before OA stimulation. Fig. 11 is a representative of two independent experiments. H7 was capable of reducing basal pIL6luc(−122) promoter activity. More importantly, however, OA-induced pIL6luc(−122) promoter activity was blocked 10- to 100-fold by H7, which coincided with dramatically reduced IL-6 protein secretion levels in monocytes (Fig. 8). In contrast, band-shift assays demonstrated that the binding of NF-κB to its consensus sequence was not altered due to H7 pretreatment, implying an essential role for serine kinase activity in regulating the transactivation potential of NF-κB (Fig. 3, lane 6). H7 pretreatment did inhibit AP-1 binding activity, indicating differential signaling modules involved in binding of NF-κB and AP-1 (data not shown).

Recent work has indicated that Rel family members may interact with other transcription factors to elicit transactivation of genes. One such example is the AP-1 complex (8, 37). Since blocking the JNK pathway inhibited transactivation potential of NF-κB, we investigated whether interaction of NF-κB with c-Jun, an important downstream target of JNK1, plays a role in transcriptional activation of the NF-κB complex. For this purpose, we transfected TF-1 cells with the pIL6luc(−122) construct together with pCMV-TAM67, the N-terminally truncated dominant negative c-Jun expression vector (38). The effectivity of pCMV-TAM67 was verified using a reporter containing five consecutive AP-1 binding sites (p(TRE)5xCAT). The AP-1 transcription factor is composed of c-Jun and c-Fos subunits (39). Our experiments confirmed that reporter activity was indeed inhibited when p(TRE)5xCAT was cotransfected with pCMV-TAM67 (data not shown). OA-induced NF-κB-driven IL-6 promoter activity was dramatically down-modulated after introduction of pCMV-TAM67 (100% and 5.9 ± 0.5%; p < 0.05, respectively), implying a role for c-Jun in the transcriptional potential of NF-κB (Fig. 10,A). To confirm that the effect elicited by pCMV-TAM67 indeed involved the NF-κB complex and not possible regulatory proteins in the vicinity of the NF-κB site, TF-1 cells were transfected with pCMV-TAM67 and p(NF-κB)3xCAT simultaneously. Overexpression of dominant negative c-Jun again strongly inhibited NF-κB-mediated promoter activity (28 ± 7% for pCMV-TAM67 vs 100% for pcDNA3; p < 0.05; Fig. 10 B). These results strongly suggest that c-Jun cooperates with NF-κB in activating a NF-κB-driven promoter element.

IL-6 is a pleiotropic cytokine released from activated monocytes that plays a crucial role in the immune response (40). In the present paper we used a chemical inhibitor of serine/threonine protein phosphatases, okadaic acid, to study the role of serine/threonine phosphorylation in intracellular signaling pathways that control IL-6 expression (21). Treatment of monocytes with OA resulted in enhanced phosphorylation of ERK1 (p44^MAPK) and ERK2 (p42^MAPK) and increased kinase activity of JNK, with concomitantly enhanced IL-6 protein secretion, IL-6 mRNA, and binding activity of the transcription factor NF-κB. Both ERK1/2 and JNK belong to the family of MAPKs that are important mediators of signal transduction from cell surface receptors to the nucleus. The three classical parallel MAP kinase cascades have been identified to date are mediated through ERK1/ERK2 (p44^MAPK and p42^MAPK), JNK (SAPK1), and p38 (RK, CSBP, MPK2) (12, 13). In our studies, pretreatment of monocytes with PD98059, a specific inhibitor of MEK1, the upstream kinase from ERK1/2, showed an inhibition of OA-induced IL-6 protein secretion from monocytes, which was not reflected by a change in NF-κB DNA binding activity (25). Although this would implicate involvement of the MEK1/ERK1/2 pathway in IL-6 protein secretion, this inhibition was not accomplished by altering the DNA binding properties of NF-κB. This observation in monocytes is contradictory to reports in other cell systems describing a close association between ERK1 activity and phosphorylation and degradation of IκB protein, thus leading to NF-κB activation. Sonoda et al., for instance, reported in vitro phosphorylation of GST-IκB-α by OA-activated ERK1 (41). Studies by Schouten et al. suggested that the mitogen-activated 90-kDa ribosomal S6 kinase (p90^rsk1) is an IκBα kinase (42). p90^rsk1, which is a serine/threonine kinase downstream from the Raf/MEK1/ERK1/2 cascade, was shown to phosphorylate IκBα on two serine residues, i.e., Ser³² and Ser³⁶. Moreover, in the lymphoblastoid cell line CEM, overexpression of either MEK1 or ERK1 demonstrated constitutive nuclear localization of NF-κB, suggesting involvement of the classical ERK pathway in NF-κB DNA binding activity (15). In our cell system, however, it is unlikely that OA activates ERK1, which, in turn, phosphorylates IκB, followed by the degradation of IκB. Rather, IκBα is phosphorylated due to OA stimulation of a pathway distinct from the Raf-1/MEK1/ERK pathway. Recently, two upstream kinases of IκBα have been cloned, IKKα and IKKβ (IκBα kinase) (43, 44). It was shown that OA inhibited a PP2A, which normally dephosphorylates IKKα, thus leading to enhanced IκBα phosphorylation and degradation and enhanced NF-κB binding activity (45). This signaling route may, independently from the Raf-1/MEK1/ERK1/2 pathway, account for the enhanced NF-κB DNA binding activity as observed after OA stimulation.

Although the Raf-1/MEK1/ERK1/2 pathway did not affect NF-κB DNA binding activity, we clearly demonstrated that overexpression of an inactive mutant of Raf-1 inhibited NF-κB-mediated IL-6 promoter activity. This indicates that the ERK pathway serves as an additional signaling pathway that is distinct from IκB degradation, NF-κB nuclear translocation, and DNA binding, but is required for NF-κB transcriptional activation by OA. In agreement with this observation are previous reports describing additional regulatory mechanisms, other than modulating NF-κB DNA binding properties, that enhance the transcriptional activity of NF-κB. In this respect, serine phosphorylation of the p65 subunit of NF-κB as a regulatory step in the transcriptional activity of NF-κB has been reported. Indeed, in our setting OA was capable of enhancing p65-mediated transcription from the GAL4 promoter. The p65 subunit contains at least two strong transactivation domains within its C terminus (46). The first domain, TA1, is contained within the last 30 amino acids of p65, whereas TA2 comprises the adjacent 90 amino acids. Schmitz et al. reported that phosphorylation and transcriptional activity of a defined region within the TA2 domain were stimulated by phorbol ester treatment of cells, and thus augmented the transactivation potential of NF-κB (46). In a recent study the catalytic subunit of protein kinase A was shown to phosphorylate p65 at a serine residue 276, thus increasing the transactivational capacity of NF-κB without affecting nuclear translocation (47). Subsequently, it was demonstrated that as a result of protein kinase A-mediated phosphorylation, p65 creates a site for interaction with the transcriptional coactivator CBP/p300, which results in enhanced transcriptional activity (48, 49). In fibroblasts and hepatoma cells IL-1-mediated NF-κB activation was accompanied by casein kinase II-mediated phosphorylation of the p65 subunit on serine residues (50). These data are in accordance with the results of our experiments showing that NF-κB transactivation capacity, but not DNA binding, is dependent on serine kinase activity. In our experimental setting, however, it remains unclear whether the MAPK-dependent phosphorylation effectively takes place on the p65 subunit itself. In this regard, we demonstrated that the JNK pathway also mediated the transcriptional potential of NF-κB, both in context of the IL-6 promoter and when bound to individual NF-κB binding sites. The JNK protein kinases phosphorylate the NH2-terminal activation domain of c-Jun on Ser⁶³ and Ser⁷³, causing increased c-Jun transcriptional activity (17). Results from cotransfection studies applying pIL6luc(−122) or p(NF-κB)3xCAT, both containing binding sites for NF-κB only, in conjunction with pCMV-TAM67, the NH₂-terminally truncated dominant negative c-Jun, suggest c-Jun/NF-κB protein/protein interaction. pCMV-TAM67 specifically binds to leucine zippers of endogenous Jun or Fos family proteins, forming low activity AP-1 complexes (51). pCMV-TAM67 binding would be expected to render the c-Jun and the c-Fos bZip domains unavailable to p65, which requires the bZip domain for interaction. Alternatively, pCMV-TAM67 may interact via its bZip domain with p65, rendering the p65 unavailable to interact with c-Jun or c-Fos. In our experiments pCMV-TAM67 dramatically impaired κB-driven promoter activity, strongly suggesting that either c-Jun or c-Fos may contribute to the transactivation capacity of NF-κB. This finding is in agreement with a previous report that cross-interaction of c-Jun or c-Fos with the NF-κB subunit p65 leads to synergized potential of NF-κB transactivation (37). Furthermore, in a human keratinocyte progression model TAM67 was also found to inhibit the activation of NF-κB (52). Taking into account that 1) the JNK pathway affects transcriptional activity of NF-κB, and 2) JNK is the main activator of c-Jun, it is tempting to speculate that interaction of NF-κB with the c-Jun protein results in maximal transactivation. Since serine kinase activity is required for the activation of c-Jun by JNK, the inhibitory effect mediated by H7 could be the consequence of impaired c-Jun phosphorylation by JNK. In concordance, Berra and co-workers demonstrated that protein kinase Cζ, a downstream substrate of Ras, triggers the activation of a number of kinases and suggested that MEK1 and ERK1/2 may also participate in NF-κB activation by enhancing the AP-1/NF-κB cross-coupling mechanism (53). In contrast, a recent study by Min et al. showed that transient overexpression of an N-terminal-truncated c-Jun or a catalytically inactive JNK inhibited TNF-induced transcription of the E-selectin but not of a κB-promoter reporter gene in endothelial cells (54). Thus, the observation that c-Jun augments κB-dependent transcription may be cell type and stimulus specific.

We cannot exclude that in our cell system, in addition to the ERK and JNK pathways, the p38 signaling route may also be involved in NF-κB activation. In studies by vanden Berghe et al. TNF-induced NF-κB-mediated transcription, but not NF-κB DNA binding, appeared dependent on activation of the p38/RK pathway in the mouse fibrosarcoma cell line L929sA (55). However, the signaling modules involved in IL-6 gene regulation are not only cell type specific, but also depend on the stimulatory signal. More detailed study is required to elucidate the role of the additional MAPK pathways in IL-6 gene expression.

Until recently limited data existed regarding the way activated kinases might be dephosphorylated and inactivated. Since the MAP kinases ERK1/ERK2 and JNK are involved in the regulation of IL-6 expression, the balance of MAPK kinase and MAPK phosphatase activities determines the state of activation of monocytes. Three mammalian MAPK phosphatases have been identified until now, CL100, PAC-1, and Pyst-1 (56, 57, 58, 59). Although their occurrence in monocytes and leukemic cell lines and the substrate specificity of these PPases have not yet been resolved, they may have important functional roles in mediating cytokine gene expression in monocytes.

We conclude that OA-induced IL-6 gene activation in monocytes is mediated through activation of the JNK and ERK1/2 pathways. The ERK1/2 pathway was shown to be specifically involved in enhancing the transactivational capacity of NF-κB. Both pathways may provide for serine kinase activity and c-Jun activation, which were shown to enhance κB-driven reporter gene expression in human monocytes.