Stress-activated Protein Kinase/Jun N-terminal Kinase Is Required for Interleukin (IL)-1-induced IL-6 and IL-8 Gene Expression in the Human Epidermal Carcinoma Cell Line KB*

The cytokine interleukin-1 (IL-1) is a major inflammatory hormone which activates a broad range of genes during inflammation. The signaling mechanisms triggered by IL-1 include activation of several distinct protein kinase systems. The stress-activated protein kinase (SAPK), also termed Jun N-terminal kinase (JNK), is activated particularly strongly by the cytokine. In an attempt to delineate its role in activation of gene expression by IL-1, we inhibited the IL-1-induced SAPK/JNK activity by stable overexpression of either a catalytically inactive mutant of SAPKβ (SAPKβ(K-R)) or antisense RNA to SAPKβ in human epidermal carcinoma cells. A detailed analysis of signal transduction in those cells showed that activation of neither NFκB nor p38 mitogen-activated protein kinase was affected, suggesting that we achieved specific blockade of the SAPK/JNK. In untransfected and vector-transfected KB cells, IL-1 induced a strong increase in expression of IL-6 and IL-8 mRNA, along with the synthesis of high amounts of the proteins. In two KB cell clones stably overexpressing the mutant SAPKβ(K-R), and three clones stably overexpressing antisense RNA to SAPKβ, expression of IL-6 and IL-8 in response to IL-1 was strongly reduced at both the mRNA and protein level. These data indicate that the SAPK/JNK pathway provides an indispensable signal for IL-1-induced expression of IL-6 and IL-8.

Interleukin-1 (IL-1) 1 is the prototype inflammatory cytokine. It is produced in two forms (␣ and ␤) by activated monocytes/ macrophages during acute or chronic inflammatory responses. It acts by inducing many genes, including cytokines (e.g. IL-2 and IL-6), chemokines (e.g. IL-8 and MCP-1), proteases (e.g. collagenase and stromelysin), adhesion molecules (e.g. ICAM-1 and E-selectin), and cyclooxygenase. Strength and duration of the expression of these genes is crucial for the intensity of an inflammatory process (reviewed in Ref. 1). Therefore much interest has focused on molecular mechanisms through which these genes are controlled by IL-1.
Two IL-1 receptors have been cloned (type I and type II) that are expressed on many different cell types (2). Only the type I receptor, heterodimerized to the IL-1 receptor accessory protein, is capable of signal transduction (3). Recently the understanding of IL-1 signaling pathways has been markedly increased by identification of novel molecules. IL-1 treatment of cells can activate at least four protein kinase cascades. One cascade involves association of an IL-1 receptor-associated protein kinase and TRAF6 with the IL-1 receptor complex (4,5), leading to activation of an NFB-inducing kinase (6), which activates the IB kinase complex. Phosphorylated IB is ubiquitinated, then degraded by the proteasome. This releases NFB, a major transcription factor regulating IL-1 responsive genes, and allows it to translocate to the nucleus (7,8).
The other cascades activated by IL-1 are those activating the three best known types of mitogen-activated protein kinase (MAPK), namely p42/p44 extracellular signal-regulated protein kinase (ERK), p38 MAPK, and the stress-activated protein kinase that phosphorylates the N-terminal region of c-Jun (SAPK/JNK) (9). Although IL-1 has been shown to activate ERK in some cells (10,11), it is a much more potent inducer of SAPK/JNK and p38 MAPK in cultured cells (12)(13)(14).
A central role of SAPK/JNK for IL-1 signaling is suggested by our finding that IL-1 activates it and its activator, MAPK kinase 7 (MKK7), in rabbit liver in vivo, without activating either p38 MAPK or ERK (15). Ten different SAPK isoforms derived from three different genes (called SAPK ␣, ␤, and ␥ in rat and JNK 2, 3, and 1 in man, respectively), which are highly conserved across species, have been cloned from vertebrate tissues (16,17). The homology among SAPK/JNK isoforms is 80 -90% on the protein level (18). We purified a 46-kDa form of SAPK/JNK activated by IL-1 from KB cells, and 50-and 55-kDa forms from rabbit liver. They accounted for essentially all of the biochemically detectable IL-1-activated JNK activity in liver and were both identified as SAPK␣ (JNK2) by amino acid sequencing a number of peptides. This suggested that SAPK/ JNK isoforms might be expressed or activated in a tissuespecific manner (19). The functional consequence of general activation of SAPK/JNK, or of the individual isoforms, on the regulation of IL-1 responsive genes is unclear (20). SAPK/JNK have been shown to phosphorylate the proteins c-Jun and ATF-2, which are components of the dimeric transcription factor AP-1. These phosphorylations result in enhanced transcription of AP-1-dependent reporter genes (21)(22)(23)(24)(25)(26). Besides phos-phorylating the activation domains of the transcription factors, SAPK/JNK isoforms bind Jun and ATF-2 proteins with different affinities, a mechanism which could result in targeting them to different intracellular substrates and to exert distinct functions (18,25,27). Although AP-1 can be considered to be a principal target of SAPK/JNK, other transcription factors, like the ternary complex factors ELK-1 and SAP-1, are also substrates (28 -31).
In an attempt to elucidate the function of IL-1-induced activation of SAPK/JNK in the human keratinocyte line KB, we stably overexpressed either an inactive mutant of the enzyme or antisense RNA. This allowed us to study the role of SAPK/ JNK on endogenous gene expression. We investigated the role of SAPK/JNK in the regulation of IL-6 and IL-8, two cytokines which are highly inducible by IL-1 in keratinocytes. IL-6 is a multifunctional cytokine that promotes B-cell growth and differentiation and stimulates acute-phase protein synthesis in liver (32). IL-8 is a chemokine attracting and stimulating leukocytes at sites of inflammation (33). We report here that inhibition of SAPK/JNK activation in cells stimulated by IL-1 results in inhibition of IL-1-induced IL-6 and IL-8 gene expression.

EXPERIMENTAL PROCEDURES
Cells and Materials-KB cells were obtained from the American Type Culture Collection and cultured in Dulbecco's modified Eagle's medium in the presence of 10% fetal calf serum.
[␥-32 P]ATP and [␣-32 P]dCTP were purchased from Hartmann Analytics, Braunschweig, Germany. Expression plasmids for GST-Jun (amino acids 1-135) and GST p54 SAPK␤ were kind gifts of Dr. J. R. Woodgett, The Ontario Cancer Research Institute, Toronto, Canada. GST fusion proteins were expressed and purified from E. coli by standard methods. Recombinant bacterially expressed histidine (His) epitope-tagged MAPK activated protein kinase-2 (MAPKAPK-2) was a kind gift of Dr. M. Gaestel (Max-Delbrueck-Center for Molecular Medicine, Berlin, Germany). The chicken anti-p54 SAPK␤ antibodies were produced by immunizing chicken with recombinant GST p54 SAPK␤ and immunoglobulins were purified from eggs. Besides the p54 SAPK␤ (which is identical to human JNK3), the antibodies also recognize purified rabbit SAPK␣ and the human JNK1 and JNK2 isoforms in cells transiently transfected with the respective cDNA, 2 and proteins of corresponding sizes in human mesangial cells (34). Rabbit antiserum to p38 MAPK was to synthetic peptide ISFVPPPLDQEEMES (amino acids 346 -360). Rabbit anti-chicken IgG coupled to horseradish peroxidase was from Sigma. cDNAs for IL-6 (1100 bp), IL-8 (700 bp), and glyceraldehyde-3phosphate dehydrogenase (1400 bp) were amplified by reverse transcription-polymerase chain reaction. Fast Flow S Sepharose, Protein A, and glutathione (GSH)-Sepharose were from Pharmacia, Uppsala, Sweden.
Plasmids and Transfections-The cDNA encoding the open reading frame of p54 SAPK␤ was amplified by polymerase chain reaction with the following primer pairs (sense, 5Ј-gcgcggatccagcaaaagcaaggtagataa-3Ј; antisense, 5Ј-gcgcggatcccctgcaacaacccagcg-3Ј) using the GST p54 SAPK␤ plasmid as template. Introduction of a point mutation in subdomain II (L55R) was done by the overlapping primer method using Pfu polymerase (Stratagene, Heidelberg, Germany) and the primer (5Јgctgagcttcctaatcgcgacatttctgtcgaggacagcg-3Ј). The resulting p54 SAPK␤(K-R) cDNA was subcloned into the BamHI site of peVRF0HA (35) in sense or antisense orientation. Sequence was confirmed by automated DNA sequencing.
KB cells were transfected with the cationic lipid Dosper (Boehringer Mannheim, Germany) and 1 g of peVRFOHA-SAPK␤(K-R) and 1 g of the pCDNA3 vector encoding the neomycin resistance gene (Invitrogen, Leek, Netherlands). Vector-transfected cells were obtained by co-transfection with peVRF0HA and pCDNA3. Stably transfected cells were selected by incubation with 600 g/ml G418 (Calbiochem, La Jolla, CA) and pooled. Single cell clones from those pools were generated by limited dilution (0.1 cells/well) in 96-well plates and cultured in conditioned medium. SAPK␤(K-R) sense and antisense RNA overexpressing cell clones were maintained without G418. Vector-transfected cells were maintained in 300 g/ml G418.
Preparation of Whole Cell Extracts-Confluent KB cells were stimulated with IL-1 (10 ng/ml) added to the culture medium. After 15 min at 37°C, the medium was removed, cells placed on ice, washed once in PBS, and scraped in PBS. Cells were collected at 500 ϫ g for 5 min and lysed in whole cell lysis buffer (10 mM Tris, pH 7.05, 30 mM NaPP i , 50 mM NaCl, 1% Triton X-100, 2 mM Na 3 VO 4 , 50 mM sodium fluoride, 20 mM ␤-glycerophosphate and freshly added 0.5 mM PMSF, 0.5 g/ml leupeptin, 0.5 g/ml pepstatin, 10 mM para-nitrophenyl phosphate, 400 nM okadaic acid). After 10 min on ice, lysates were cleared by centrifugation at 10,000 ϫ g for 15 min at 4°C. Protein concentration of supernatants was determined by the method of Bradford and samples stored at Ϫ80°C.
SAPK/JNK Assay-The assay was performed as described (36). Briefly, 10 l containing 30 g of whole cell extract or 10 g of cytosolic extract protein were incubated with 10 l of GST-Jun (1 g) and 10 l of kinase buffer (150 mM Tris, pH 7.4, 30 mM MgCl 2 , 60 M ATP, 4 Ci of [ 32 P]ATP). After 15 min at room temperature 20 l of GSH beads, equilibrated in lysis buffer ϩ 1 mM DTT, were added. Samples were shaken for 30 min at room temperature. Beads were recovered at 10,000 ϫ g for 5 min and washed twice in 200 l of whole cell lysis buffer. Bound GST-Jun was eluted from the beads by boiling for 5 min in SDS-PAGE sample buffer (8% SDS, 100 mM Tris, pH 6.8, 4% ␤-mercaptoethanol, 24% glycerol, 0.02% bromphenol blue). After centrifugation at 10,000 ϫ g for 5 min supernatants were separated on 10% SDS-PAGE. Equal recovery of GST-Jun was confirmed by Coomassie staining. Phosphorylated GST-Jun was visualized by autoradiography and quantified using a PhosphorImager and the Molecular Analyst Program (Bio-Rad).
p38 MAPK Assay-500 g of whole cell extract protein was diluted in 500 l of immunoprecipitation (IP) buffer (20 mM Tris, pH 7.3, 154 mM NaCl, 50 mM sodium fluoride, 1 mM Na 3 VO 4 , 1% Triton X-100). 2 l of rabbit anti-p38 MAPK antiserum were added for 2 h at 4°C. Then 20 l of protein A-Sepharose, equilibrated in IP buffer was added. After 1 h at 4°C beads were spun down at 10,000 ϫ g for 5 min, washed three times in 500 l of IP buffer, and resuspended in 10 l of IP buffer. Then 10 l of HIS-MAPKAPK-2 (750 ng) and 10 l of kinase buffer (see above) were added. After 20 min at room temperature assays were stopped by adding SDS-PAGE sample buffer, boiled for 5 min, centrifuged, and proteins separated on 7.5% SDS-PAGE.
Partial Purification and Assay for an Activator of SAPK/JNK-KB cells were stimulated with IL-1 (10 ng/ml) for 15 min, washed, and scraped in ice-cold PBS. Cells were resuspended in lysis buffer (20 mM Tris, pH 7.4, 1 mM EGTA, 1 mM EDTA, 2 mM DTT, 0.2 mM Na 3 VO 4 , 50 mM sodium fluoride, 1 mM PMSF, 10 M E64, and 1 M pepstatin). Cells were broken by passaging them three times through a 26-gauge needle. Cytosolic extracts were prepared by centrifugation at 100,000 ϫ g for 1 h at 4°C. 1 mg of cytosolic protein was brought to pH 6.0 by adding 50 mM MES, pH 6.0, and loaded on a 0. tion, washed twice in 1 ml of lysis buffer, and resuspended in 10 l of lysis buffer. 10 l of GST-Jun (1 g) and 10 l kinase buffer including 4 Ci of [ 32 P]ATP were added for 15 min. Assays were stopped with SDS sample buffer, boiled, microcentrifuged, and proteins separated on 10% SDS-PAGE. GST-SAPK␤ and GST-Jun were visualized by Coomassie staining of gels and phosphorylation analyzed by autoradiography.
Electrophoretic Mobility Shift Assay-Extracts of nuclear proteins were prepared as described above. Protein DNA-binding reactions, preparation, and labeling of the oligonucleotide used for measurement of NFB activity were carried out exactly as described (37). Protein-DNA complexes were resolved on 4% PAGE and visualized by autoradiography.
Western Blotting-Proteins were separated on SDS-PAGE and electrophoretically transferred to polyvinylidine difluoride membranes (Immobilon R , Millipore, Bedford, MA). The membranes were blocked with 5% dried milk in Tris-buffered saline (TBS) overnight. Membranes were then incubated for 4 -24 h with chicken anti-p54 SAPK␤ antibodies, washed in TBS, and incubated for 2-4 h with the peroxidase-coupled second antibody. Proteins were detected using the Amersham enhanced chemiluminescence system. Autoradiographs were scanned with the GelDoc100system and quantitated with the Molecular Analyst program (Bio-Rad).
Enzyme-linked Immunosorbent Assay-2.5 ϫ 10 6 KB cells were trypsinized and seeded onto six-well plates with 1 ml of culture medium. One day later medium was replaced and cells were stimulated with 10 ng/ml IL-1 for 24 h or left untreated. The medium was recovered, cells collected by a 500 ϫ g centrifugation and the medium stored at Ϫ20°C. Cells were resuspended in PBS and counted. Specific enzyme-linked immunosorbent assays for IL-6 and IL-8 were performed in duplicates exactly to the manufacturer's instructions (Amersham) and cytokine release normalized for cell number.

Stable Overexpression of SAPK␤(K-R) Mutant or Antisense
RNA to SAPK␤ in KB Cells-In our efforts to identify a functional role of SAPK/JNK in IL-1-induced gene expression, we followed approaches that have been used successfully to demonstrate a role for the p42/p44 MAPK cascade in growth factorregulated fibroblast proliferation; i.e. stable overexpression of MAPKK mutants (38) and overexpression of ERK mutants and ERK antisense RNA (39). We constructed a mutant rat p54 SAPK␤ cDNA, in which a lysine residue critical for kinase activity was replaced by arginine. This lysine residue is highly conserved in the protein kinase family (40) and is involved in ATP binding of the p42 MAPK (ERK2) as shown by structural studies (41). The open reading frame of the SAPK␤(K-R) mutant was placed 5Ј to the hemagglutinin (HA)-tag sequence under the control of a cytomegalovirus promotor. KB cells were stably transfected with the construct or the vector alone. Cell clones were analyzed by Western blotting using chicken egg antibodies that were raised against bacterially expressed fulllength GST-p54 SAPK␤ protein. Thereby we identified six KB cell clones overexpressing the SAPK␤ mutant. Western blot detection of two of these is shown in Fig. 1A. Both clones strongly overexpressed the transfected protein, compared with the endogenous SAPK/JNK. Identity of the HA-SAPK␤(K-R) band was confirmed by detection of a product of identical size with anti-HA tag antibodies (not shown).
As expected from their high homology to each other, different SAPK/JNK isoforms are recognized by the chicken anti-SAPK␤ antibodies (see "Experimental Procedures"). In the KB cells they detected several endogenous JNK isoforms which appear as two doublet bands of about 45 to 48 kDa and 54 to 57 kDa. As detailed by Gupta et al. (18), the JNK 1 and JNK 2 genes each can give rise to differently spliced transcripts which encode proteins of 46 and 55 kDa. The JNK 3 gene transcripts apparently are translated into two doublets of 45-48 kDa and 54 -57 kDa, due to partial usage of an additional upstreamlocated start codon (18). Thus while it is not possible to unequivocally assign each product to a specific isoform, the observed pattern is in close accordance with that reported in Ref. 18. All four bands were confirmed to represent different SAPK isoforms, since we observed protein kinases of corresponding sizes in in-gel kinase assays with GST-Jun as substrate. 3 Three proteins of higher M r on SDS-PAGE were not related to SAPK activity and represent proteins reacting nonspecifically with the chicken egg immunoglobulin.
On the cDNA level sequence comparison between the 1.3kilobase open reading frames of SAPK␤ and those of JNK1, JNK2, and JNK3 revealed a homology of 78, 74, and 91%, respectively. Therefore, overexpression of the full-length open reading frame in antisense orientation was likely to hybridize to all JNK mRNAs present in KB cells and inhibit their translation. In fact, suppression of similarly related proteins has been demonstrated in a study where a 1.75-kilobase antisense fragment of the ERK1 gene inhibited both, the expression of ERK1 and ERK2 isoforms (39). The homology of ERK1 and ERK2 is comparable to that between the JNK isoforms. We isolated several clones stably overexpressing antisense RNA. Endogenous SAPK/JNK protein levels in three of those clones and in untransfected KB cells were compared by Western blot. The bands corresponding to endogenous JNK isoforms were all significantly reduced in all three clones (Fig. 1B). The abundance of three cross-reacting proteins of higher molecular M r on SDS-PAGE was not reduced. According to densitometric quantitation the amounts JNK proteins in the three different clones were 24, 54, and 28% of the amount in control cells.
Stable Overexpression of SAPK␤(K-R) and SAPK␤ Antisense RNA Inhibits IL-1-induced SAPK/JNK Activation-We next investigated whether overexpression of the kinase-inactive SAPK␤ had a dominant negative effect and interfered with SAPK/JNK activity induced by IL-1. Kinase activity was determined in vitro using GST-Jun (amino acids 1-135) as substrate. The Jun protein was then purified from the reaction mixture by adsorption to GSH beads. This assay allowed quantitative measurement of SAPK/JNK activity in the presence of both endogenous SAPK isoforms and the SAPK␤ mutant. In whole cell lysates of clones overexpressing SAPK␤(K-R), IL-1mediated stimulation of total SAPK/JNK activity was decreased by about 50% (Fig. 2A). A comparable degree of inhibition was observed when kinase activity was analyzed in cytosolic extracts of the cells (Fig. 2B). No apparent differences were observed in the kinetics of the IL-1 induced transient SAPK/JNK activation in clones 2 and 11 compared with untransfected or vector-transfected KB cells. After 60 min of IL-1 treatment Jun kinase activity was down-regulated irrespective of the presence of SAPK␤(K-R) (Fig. 2B). In agreement with this, the expression of MKP-1, a highly specific MAPK phosphatase implicated in down-regulation of SAPK/JNK (42), was induced by IL-1 within 15 min and sustained over several hours in KB cells as well as in the cells overexpressing SAPK␤(K-R) (data not shown). SAPK/JNK activity was also impaired in cytosolic extracts of SAPK␤ antisense RNA expressing clones. Impairment was even stronger than in the SAPK␤(K-R) expressing clones (Fig.  3, see also Table I Procedures." The fold increase in SAPK/JNK activation induced by IL-1 obtained from at least six independent experiments is shown. Error bars show standard error of the mean. B, the same cells as in A were stimulated for the indicated times with 10 ng/ml IL-1 or left untreated. Cytosolic extracts were prepared as described in detail under "Experimental Procedures" and SAPK/JNK activity determined. The GST-Jun phosphorylation was analyzed by autoradiography, and quantified with a PhosphorImager (C). by interfering with the ERK MAPK cascade using the same approaches (38,39,43).

IL-1-induced Activation of SAPK Kinase (SAPKK) and p38 MAPK Is Unaltered in Cells Overexpressing SAPK␤(K-R) or SAPK␤ Antisense RNA-Kinase-dead mutants have been widely used in transient co-transfection experiments and
shown to act as dominant negative inhibitors. We were interested to know at which point of the SAPK cascade the SAPK␤(K-R) mutant was acting as an inhibitor. The mutated enzyme contains an intact regulatory domain and is phosphorylated by partially purified SAPKK in vitro. 3 By interacting with SAPKK, large amounts of SAPK␤(K-R) might prevent its activation in response to IL-1. This prompted us to investigate the regulation of the SAPKK activity, which has not been unequivocally identified in KB cells but may correspond to MKK7 (15,44,45), in the cells overexpressing SAPK␤(K-R). We measured IL-1-induced SAPKK activation in a two-stage assay. Partially purified SAPKK was first incubated with recombinant GST-SAPK␤; the latter was then adsorbed to GSH-Sepharose beads and its activity measured on GST-Jun (1-135). Fig. 4 shows that in the two clones overexpressing SAPK␤(K-R) activation of the major IL-1-induced SAPKK was normal in comparison to untransfected or vector-transfected cells. Therefore the SAPK␤(K-R) mutant inhibits activation of the SAPK pathway downstream of SAPKK activation, presumably by competition with the endogenous SAPK/JNK for SAPKK.
The p38 MAPK cascade is strongly stimulated by IL-1 in KB cells. P38 MAPK phosphorylates and activates MAPKAPK-2, which phosphorylates the small heat shock protein hsp27 (13,46,47). MAPKAPK-2 is a highly specific substrate for p38 MAPK and we made use of it to assess the activation state of p38 MAPK in the stably transfected cells. IL-1 activation of p38 MAPK in KB cells was not affected by overexpressing SAPK␤(K-R) (Fig. 5A), nor by overexpressing SAPK␤ antisense RNA (Fig. 5B), suggesting that there was no interference with this pathway. The ERK MAPK cascade has been found by us to be activated in response to IL-1 in fibroblasts (11,15) but not in KB cells (12) and was therefore not further investigated in this study.
Activation of NFB in Clones Overexpressing SAPK␤(K-R) or SAPK␤ Antisense RNA-Activation of NFB is an important effect of IL-1 which has been demonstrated to be essential for IL-1-induced transcriptional regulation. This pathway involves activation of a kinase complex that phosphorylates the inhibitor IB and is distinct from the known MAPK pathways (7,8). The results of electrophoretic mobility shift assays showed rapid activation of NFB in KB cells occurring within 15 to 30 min after IL-1 stimulation. That response was not affected by overexpression of the SAPK␤ mutant or antisense RNA (Fig.  6). The data presented in Figs. 2-6 suggest specific inhibition of the SAPK/JNK cascade by expression of mutant SAPK␤ or its antisense RNA without interference with other early signaling events induced by IL-1.

Impaired IL-1-induced Expression of IL-6 and IL-8 in KB Cells Overexpressing SAPK␤(K-R) or SAPK␤ Antisense RNA-
Untransfected and vector-transfected KB cells respond to IL-1 with a rapid induction of IL-6 and IL-8. Their mRNAs are increased from barely detectable levels to high amounts within 1 h, and remain elevated for more than 30 h (Fig. 7). Analysis of their RNAs in the SAPK␤(K-R) transfected clones revealed that this response was markedly impaired, resulting in much lower mRNA levels for both cytokines especially at times following their initial increase. Impairment of IL-6 and IL-8 induction was even more pronounced in the antisense-transfected clones. In those cells the amounts of both transcripts were low already in the early phase of IL-1 stimulation.

TABLE I Comparison of KB cell clones overexpressing SAPK␤(K-R) or SAPK␤ antisense RNA
Results (mean Ϯ S.E.) of SAPK/JNK protein amounts (determined by Western blot, n ϭ 4) and activity (determined as in Figs. 2 and 3, n ϭ 6 to 9) and of IL-1-induced cytokine production (values from Fig. 8  The secretion of IL-6 and IL-8 by the transfected and control cells was analyzed after 24 h of IL-1 treatment by specific enzyme-linked immunosorbent assay. As shown in Fig. 8, IL-6 and IL-8 production was decreased in clones overexpressing SAPK␤(K-R), and in clones overexpressing SAPK␤ antisense RNA. Of note, suppression was stronger in the latter clones, corresponding to a stronger decrease in their JNK activity (see also Table I). Thus the results obtained on the IL-6 and IL-8 protein level reflect the effects of SAPK inhibition obtained on the mRNA level and confirmed an impaired response to IL-1induced IL-6 and IL-8 expression in cells manipulated to suppress SAPK/JNK activation. DISCUSSION A prominent early effect of IL-1 is activation of SAPK/JNK. Due to the lack of cell permeable specific inhibitors to the enzymes there is still very little information about their func-tion in vivo. Mice bearing null mutations at all three loci could resolve this issue. Unless a particular SAPK/JNK isoform is essential for a particular process, the existence of so many isoforms might compensate for deficits resulting from null mutations at only one or two loci. Consequently attempts have been made to inhibit the SAPK pathway at the level of the upstream activator(s). Partial inhibition of SAPK/JNK activation by stable overexpression of an inactive MKK4/SEK1 mutant in mouse fibroblasts resulted in enhanced thermotolerance and higher resistance to cytotoxic agents (48). Two recent studies (49,50) reported MKK4/SEK1 knockouts by gene disruption in mouse embryonic stem cells. Homozygous mutant cells showed a complete inhibition of SAPK/JNK activation in response to anisomycin and heat shock, whereas osmotic stress and UV light still activated the cascade (49,50), results which are consistent with the existence of additional activators of the SAPK/JNK pathway, such as the recently cloned MKK7 (44,51). These important experiments therefore do not eliminate the need to examine SAPK/JNK function with regard to stimulus and cell type in vivo.
In this paper we present for the first time evidence for a role of SAPK in IL-1-induced expression of endogenous genes. By stably overexpressing an inactive mutant of SAPK␤ in human KB cells we achieved partial inhibition of the SAPK/JNK in vivo. Using an alternative approach we strongly inhibited the SAPK pathway by overexpressing SAPK␤ antisense RNA.
We investigated early IL-1-induced signaling events, because one major concern regarding stable overexpression of MAPK mutants in cells is the possible interference with related signaling pathways. According to our previous results IL-1 did not significantly activate the ERK MAPK pathway in KB cells (12), implying that it is not involved in IL-1 effects in these cells. However, IL-1 strongly activated the p38 MAPK (13). The p38 MAPK was likely to be affected, for following reasons. First, SAPK/JNK and p38 MAPK are activated by the same stressful stimuli and inflammatory cytokines. Second, MKK4/ SEK1 and MKK7, the only enzymes that activate SAPK/JNK, have also been shown to activate p38 MAPK when overexpressed, suggesting that they can act as common activators of the two MAPK cascades (24,44,52). Large amounts of a SAPK mutant may bind to these upstream activators and interfere with their activation by a MAPKK kinase. Third, SAPK/JNK and p38 MAPK phosphorylate common substrates, like ATF-2, a component of AP-1 (22,53). Phosphorylation and activation of such substrates might be prevented by an excess of a SAPK/ JNK mutant.
We found that the SAPK␤(K-R) mutant, although strongly overexpressed, did not interfere with activation of its activator. The most likely explanation for the inhibition of IL-1-induced SAPK/JNK activation is therefore competition of the SAPK␤(K-R) mutant with the endogenous kinases for the activating SAPKK. A competitive inhibitory mechanism would imply that much higher overexpression might be required for full blockade of IL-1 induced SAPK/JNK. We also found that overexpression of SAPK␤(K-R) did not affect p38 MAPK activation, concluding that it did not compete with p38 MAPK for a common activator of both enzymes.
The inhibition of IL-1-induced SAPK/JNK activity by antisense RNA occurs by a mechanism completely different from that of the mutant protein. This approach therefore represented an independent means of confirming the cellular phenotype resulting from overexpression of SAPK␤(K-R). If each approach resulted in inhibition of SAPK/JNK activation they should each cause similiar downstream effects. IL-1 activated NFB and p38 MAPK normally in cells overexpressing either SAPK␤(K-R) or the antisense RNA. Since inhibition of SAPK/ JNK was seemingly specific, we made use of the cell lines to assess the importance of the enzymes for expression of cytokine genes in response to IL-1. We examined gene expression of IL-6 and IL-8, two important mediators generated by IL-1. Both were strongly suppressed at mRNA and protein levels, suggesting that SAPK/JNK activation is crucial for the IL-1-induced formation of IL-6 and IL-8. Interestingly, inhibition of the IL-1-induced SAPK/JNK activity correlated in extent with inhibition of IL-6 and IL-8 gene expression, both being more pronounced with antisense RNA overexpression. Furthermore, the relatively weak SAPK/JNK inhibition upon overexpression of the SAPK␤ mutant resulted in decreased IL-6 and IL-8 mRNA levels at late time points of IL-1 treatment while being insufficient to suppress the initial mRNA increase. This indicates that full induction of mRNA expression at that time can occur with submaximal SAPK/JNK activation, while maximal activation of the pathway is required for inducing a sustained increase in expression of the IL-6 and IL-8 genes.
While the antisense RNA approach apparently reduces expression of all SAPK/JNK isoforms to similiar extents (Fig. 1B), the SAPK␤(K-R) mutant may preferentially suppress activity of the corresponding human (JNK3) isoforms. If so, the results in Fig. 7 suggest that JNK3 isoforms might be important particularly for the prolonged elevation of IL-6 and IL-8 mRNA. Taken together, the results obtained with both the dominant negative mutant and the antisense RNA overexpression, although not enabling us to relate specifically to certain SAPK/ JNK isoforms, strongly suggest a crucial role for members of this kinase family in IL-1-induced expression of IL-6 and IL-8.
Although we cannot exclude the possibility that the cell lines generated have genetic defects in other signaling pathways enabling them to survive selection in the presence of reduced SAPK activity, our data suggest that the molecules required for IL-1 signal transduction are present in the cells. This is shown by the fact that IL-6 and IL-8 gene expression could still be induced by IL-1 in all cell lines overexpressing SAPK␤(K-R) or antisense RNA, although to a much lesser extent.
We can only speculate about the molecular mechanism of the IL-6 and IL-8 inhibition. Expression of both genes is controlled at several levels (54 -57). In transient transfection studies ev- Nuclear proteins were prepared and analyzed for binding to an NFB oligonucleotide. Protein DNA complexes were resolved on 4% PAGE and visualized by autoradiography. The inducible NFB complex is indicated (I). Formation of the faster migrating uninducible protein DNA complex (II) was not specific, since it was not competed with cold oligonucleotide (not shown). See "Experimental Procedures" for details. idence has been provided that an intact NFB site is essential for IL-6 and IL-8 transcriptional regulation (58 -60). We found that nuclear translocation and DNA binding of NFB induced by IL-1 was unaltered in the cell lines. Our results therefore strongly suggest that a second, SAPK/JNK-dependent activation mechanism, in addition to NFB, is absolutely required for endogenous gene expression of IL-6 and IL-8.
At present, the nature of this second activation mechanism remains elusive. Recent reports suggested that regulation of IL-6 and IL-8 transcription is the result of a complex functional balance between the known transcriptional activators NFB, NF-IL-6/C/EBP, AP-1, and repressors like Oct-1 and RBP (55)(56)(57)(58)(59)(60)(61)(62). Therefore known or novel AP-1 components, factors distinct from AP-1 or as yet unidentified transcription factors, might be targeted directly or indirectly by SAPK/JNK to regulate IL-6 and IL-8 gene expression. Another level of complexity is added by the observation that IL-1 stabilizes the mRNA of both genes (54,(63)(64)(65).
IL-1 induces the synthesis of several proteins with important (patho)physiological functions in inflammatory responses. The current knowledge suggests that different IL-1-induced signaling pathways contribute to these effects. The NFB pathway has been found crucial for the activation of IL-6 and IL-8 promotors by IL-1 (56 -58). The results presented in this paper suggest that, in addition to the effect of NFB, a major stimulatory effect on the expression of these two cytokines is exerted by the SAPK/JNK pathway. In contrast, inhibition of the p38 MAP kinase cascade by the compound SB203580 hardly affected synthesis of IL-6 and IL-8 (66), while in the same study it strongly suppressed expression of matrix metalloproteinases and cyclooxygenase II.
In summary, the results give new insight into the contribution of the SAPK/JNK pathway to IL-1-induced gene regulation in human cells, leading to the conclusion that each of the major signaling pathways activated by IL-1 contributes in a selective manner to the full pattern of gene induction by IL-1.