Proteinase-activated receptor-2-mediated activation of stress-activated protein kinases and inhibitory kappa B kinases in NCTC 2544 keratinocytes.

In this study we examined the regulation of the stress-activated protein (SAP) kinases and inhibitory kappa B kinases (IKKs) through stimulation of the novel G-protein-coupled receptor proteinase-activated receptor-2 in the human keratinocyte cell line NCTC2544. Trypsin and the peptide SLIGKV stimulated a time-dependent increase in both c-Jun N-terminal kinase and p38 mitogen-activated protein kinase activity. Trypsin also stimulated NF kappa B-DNA binding and the activation of the upstream kinases IKK alpha and -beta. Phorbol 12-myristate 13-acetate also strongly activated both SAP kinases and IKK isoforms, suggesting the potential for a protein kinase C-mediated regulatory mechanism underlying the effects of trypsin. Pre-incubation with selective protein kinase C (PKC) inhibitors GF109203X and Gö6983, or transfection of dominant negative (DN)-PKC alpha, abolished phorbol 12-myristate 13-acetate-mediated c-Jun N-terminal kinase activity, although it only partially inhibited the response to trypsin. In contrast, Gö6983 reduced trypsin-stimulated p38 mitogen-activated protein kinase activity to a greater extent than GF109203X, although DN-PKC alpha or PKC zeta had no substantial effect. Additionally, inhibitors of PKC partially reduced trypsin-stimulated IKK alpha activity but abolished that of IKK beta, whereas DN-PKC alpha but not DN-PKC zeta substantially reduced trypsin-stimulated Flag-IKK beta activity. This study shows for the first time proteinase-activated receptor-2-mediated stimulation of both SAP kinase and IKK signaling and differing roles for PKC isoforms in the regulation of each pathway.

Proteinase-activated receptor-2 (PAR-2) 1 is a recently described member of the seven-transmembrane, G-protein-linked receptor family (1,2) exemplified by the thrombin receptor (PAR-1) (3). It is activated by serine protease-mediated cleavage of the receptor to generate a new N terminus, which then interacts with the second exofacial loop of the receptor (see review in Ref. 4). PAR-2 is strongly activated by trypsin, tryptase, and potentially other unidentified enzymes (5) and is unique among the PAR family, as the newly described PAR-3 and PAR-4 are also thrombin-sensitive (6,7).
PAR-2 is strongly expressed in smooth muscle cells of the airways, vasculature, and intestine, and in cells of epithelial origin such as endothelial cells and enterocytes (8,9). It is also highly expressed in keratinocytes (10), where it regulates a number of inflammatory linked responses such as the expression of interleukin-6 and -8 and granulocyte macrophage-colony stimulating factor (11,12). However, at present there is little information available regarding the role of specific cellular signaling pathways in regulating PAR-2-mediated cellular events. Studies have been limited principally to effects upon the generation of inositol trisphosphate and mobilization of intracellular Ca 2ϩ , both events associated with a receptor coupled to phospholipase C activation (13,14). Two significant intracellular signaling pathways potentially involved in mediating inflammatory responses in a number of cell types are the mitogen-activated protein (MAP) kinases, and the nuclear factor B (NFB) pathway. The MAP kinases consist of the classical isoforms of MAP kinase (extracellular signal-regulated kinases) and the stress-activated protein (SAP) kinases c-Jun N-terminal kinase (JNK) and p38 MAP kinase (15). All three kinases have been implicated in the regulation of a number of processes in keratinocytes such as involucrin expression, protein kinase C (PKC)-mediated hyperproliferation, and UV-mediated apoptosis (16 -18). Previous studies have shown that, in rat aortic smooth muscle cells, trypsin stimulates p42/44 MAP kinase but does not activate JNK and p38 MAP kinase (19). However, different coupling mechanisms may exist in different target cells, particularly in keratinocytes where PAR-2 is highly expressed.
The NFB family of transcription factors is well recognized as being involved in the regulation of a number of pro-inflammatory genes (20). In cytokine-stimulated cells, NFB is regulated by the activation of isoforms of inhibitory B kinase (IKK) that regulate phosphorylation of inhibitory B (21)(22)(23). IKK itself is regulated by a number of upstream regulatory kinases, including NFB-inducing kinase (NIK) and other potential intermediates such as protein kinase B and PKC (24 -26). Although some studies have demonstrated activation of NFB in response to thrombin, bradykinin, and other G-protein-coupled receptor agonists including trypsin (27)(28)(29)(30)(31), very little informa-tion is available regarding the role and regulation of IKK isoforms in NFB signaling following activation of a G-proteincoupled receptors.
Since PAR-2 activation displays many features similar to cytokines regarding the activation of inflammatory mediator release, we sought to determine if the signaling events mediated by PAR-2 were similar. Thus, we examined SAP kinase and NFB activation, pathways that are known to be strongly activated by cytokines (15,20). Here we show for the first time in a transfected epithelial cell line (32) strong activation of both SAP kinase and IKK/NFB signaling pathways following PAR-2 activation, an effect mimicked by activation of protein kinase C. Using isoform-selective protein kinase C inhibitors (33,34) and pleotropic dominant negative mutants of PKC (35), we demonstrate the involvement of PKC-dependent and -independent mechanisms in the regulation of SAP kinase activity. We also present evidence supporting an important role for typical PKC isoforms in mediating PAR-2-stimulated IKK/ NFB signaling.

EXPERIMENTAL PROCEDURES
Materials-All chemicals and reagents were obtained from appropriate commercial sources. Escherichia coli expression plasmids for GST-MAPKAP kinase-2 and GST-c-Jun (5-89) were kind gifts of J. Woodgett (Ontario Cancer Institute, Toronto, Canada) and C. J. Marshall (Chester Beatty Laboratories, London, United Kingdom (UK)). The plasmids encoding wild type and DN-IKK␣ and IKK␤ were kind gifts from D. Goeddel (Tularik Inc.), whereas those encoding dominant-inhibitory PKC mutants PKC␣ (T/A) 3  PCR Cloning of the Human PAR-2 cDNA and Cellular Expression-Human PAR-2 cDNA was cloned by PCR amplification from a human umbilical vein endothelial cell cDNA library. PCR primer design was based on the published human PAR-2 sequence (2). The primer sequences were as follows: forward primer, 5Ј AACCAAGCTTTCTCGG-TGCGTCCAGT-3Ј; reverse primer, 5ЈGCTCTAGACTGCAATTCCCAT-CTGAGG-3Ј. The design also incorporated unique HindIII and XbaI sites in the forward and reverse primers respectively, to allow directional cloning. The PCR product was subcloned into a similarly digested pRc/RSV vector using T4 DNA ligase to generate the pRC/RSV-PAR-2 expression vector. The human keratinocyte cell line NCTC2544 was then transfected with pRC/RSV-PAR-2 using Lipofectin, and clonal cell lines were isolated in 800 g/ml Geneticin. Several clones were tested for PAR-2 expression using PCR and assay of [ 3 H]inositol phosphate ([ 3 H]IP) accumulation (see below). One of these clones (G) was used in the study.
Cell Culture-Human skin epithelial cells NCTC2544 were maintained in M199 medium with Earl's salt supplement, 10% (v/v) fetal calf serum, 100 units of penicillin/ml, and 100 g of streptomycin/ml in a humidified atmosphere containing 5% CO 2 at 37°C. NCTC2544 cells stably expressing human PAR-2 (G) were maintained in complete M199 medium containing 400 g/ml Geneticin for selection pressure and passaged using Versene. Primary human keratinocytes were cultured from neonatal foreskins obtained from the Royal Hospital for Sick Children (Edinburgh, UK) as outlined previously (36). Cells were grown in supplemented serum-free media and used at passages 3-4.
RT-PCR-Cells were grown to confluence in 100-mm 2 culture dishes and the total RNA extracted by cell lysis using the RNeasy kit (Qiagen Ltd.). cDNA was produced by using oligo(dT) primers and Superscript kit (Life Technologies, Inc.). The PAR-2 sequence was then amplified from the cDNA by PCR using KlenTaq polymerase mix (CLONTECH Ltd.) and the PAR-2 primers indicated above at a concentration of 10 pmol/ ml. A control reaction was set up using actin forward (5Ј-CGTGGGC-CGCCGCCCTAGGCACCA-3Ј) and reverse (5Ј-TTGGCCTTAGGGT-TCAGGGGG-3Ј) primers (19). The reaction was carried out in a Techne Genius automatic cycler using 5 min denaturing at 96°C followed by a step cycle. The step cycle consisted of a denaturing step at 96°C for 1.5 min, annealing at 60°C for 1 min, and then a product extension step at 68°C for 2 min. A final extension step, 65°C for 5 min, was then included. Samples were exposed to 40 cycles of the PCR conditions. Samples were resolved on a 2% (w/v) agarose gel, containing 10 mg/ml ethidium bromide. The products were run against a 1-kilobase pair DNA ladder (Life Technologies, Inc.) and were imaged under UV light.
Transient Transfection Procedures-NCTC2544 clones expressing PAR-2 were grown to 60 -70% confluence on six-well plates and transiently co-transfected with various plasmids using the LipofectAMINE Plus ® transfection system, according to the manufacturer's protocol. For measuring kinase activities, 200 ng of the plasmid encoding HA-JNK1 was transfected together with 600 ng of DN-PKC␣ (T/A) 3 , PKC (T/A) 4 , or blank vector pcDNA3. For each transfection, the total amount of DNA was adjusted to 800 ng/well with blank vector. A similar procedure was carried out for FLAG-p38 MAP kinase except that 100 ng of the construct was employed. For epitope-IKK assays, wild type Myc-IKK␣ and FLAG-IKK␤ were transfected at 900 and 100 ng/well, respectively, in the presence of either 300 ng of DN-PKC (as above) or blank vector. For the NFB luciferase assays, 250 ng of a plasmid containing 3ϫ NFB DNA enhancer, linked to the firefly luciferase reporter gene luc, or blank vector were transfected together, or in conjunction with 500 ng of DN-IKK␣ and/or ␤, to give 1.25 g/well (37). After a 6-h incubation period with the DNA mixture in media free of serum and antibiotics, cells were transferred into complete M199 medium for another 30 h. Cells were then rendered quiescent by serum deprivation for 24 h before stimulation.
JNK and p38 MAP Kinase Assays-For assay of JNK, pre-cleared supernatants were added to a 20-l slurry of GST-c-Jun-(5-89)/GSH-Sepharose beads and mixed for 3 h at 4°C, whereas for p38 MAP kinase, full-length GST-MAPKAP kinase-2 was used. The precipitates were resuspended in 25 l of kinase buffer and the kinase reaction started by the addition of [␥-32 P]ATP (1-2 Ci, 25 M) and incubated for 20 min at 30°C. The reaction was terminated by addition of 10 l of 4ϫ Laemmli sample buffer. Samples were boiled for 5 min and resolved by SDS-polyacrylamide gel electrophoresis (11% (w/v) gel). Gels were dried and subjected to autoradiography overnight (39).
For epitope-tagged JNK assays, agonist-stimulated cells were lysed in 400 l of lysis buffer and supernatants incubated for 3 h at 4°C with 10 l of protein G-Sepharose beads pre-coupled with anti-HA antibody (Y-11; Santa Cruz, 0.2 g) for 1 h at 4°C. HA-JNK1 activity was determined in vitro in 30 l of kinase buffer containing 1 M ATP, 2 Ci of [␥-32 P]ATP, and 2 g of affinity-purified GST-c-Jun-(5-89) as a substrate for 20 min at 30°C. A similar protocol was employed for FLAG-p38 MAP kinase activity except that 1 g of anti-FLAG antibody (anti-Oct A; Santa Cruz) was employed in the immunoprecipitation step and 1 g/tube GST-ATF-2 was used in the kinase assay.
Electrophoretic Mobility Shift Assay-Following termination by washing in ice-cold PBS, cells were harvested and pelleted and crude nuclear extracts made as described previously (40).
NFB-DNA binding was assessed by electrophoretic mobility shift assay, according to the kit manufacturer's instructions (Promega).
IKK Assay-Cell lysates were incubated in 300 l of solubilization buffer (20 mM Tris-HCl (pH 7.6), 1 mM EDTA, 0.5 mM EGTA, 10% glycerol (v/v), 0.1% Brij 35, 150 mM NaCl, 1% Triton X-100 (w/v), 20 mM NaF, 20 mM ␤-glycerophosphate, 0.5 mM Na 3 VO 4 , 1 mM phenylmethylsulfonyl fluoride, 0.5 mg/ml leupeptin, 0.5 mg/ml aprotinin) on ice for 30 min. Pre-cleared samples were then added to either IKK␣ or IKK␤ polyclonal antibody coupled to Protein G-Sepharose beads, and mixed for 2 h at 4°C. Immunoprecipitates were recovered by centrifugation and washed twice in solubilization buffer and once in kinase buffer (25 mM HEPES (pH 7.6), 20 mM MgCl 2 , 5 mM ␤-glycerophosphate, 0.1 mM Na 3 VO 4 , 2 mM dithiothreitol). Precipitates were then resuspended in 25 l of kinase buffer and the kinase reaction initiated by addition of [␥-32 P]ATP (5 Ci), 25 M ATP, and 2 g of GST-IBN (1-70, Nterminal truncated IB). Samples were incubated in a final volume of 30 l, with shaking, at 30°C for 30 min before termination of the kinase reaction by addition of 4ϫ Laemmli sample buffer. Samples were then resolved on 11% (w/v) acrylamide SDS-polyacrylamide gel electrophoresis and the phosphorylated protein identified by autoradiography. For Flag-IKK␤ assays the same procedure was followed except 1 g/tube anti-FLAG antibody was used in the immunoprecipitation step.
NFB Reporter Activity Assay-NCTC2544 cells grown on six-well culture plates were transfected as described previously and were quiesced in serum-free M199 overnight. Cells were treated with appropriate agonists for 5 h and assayed as outlined previously (37).

RESULTS
In preliminary experiments we found that in NCTC2544 cells expressing PAR-2, trypsin, and the PAR-2-activating hexapeptide (SLIGKV) stimulated strong activation of [ 3 H]IP accumulation (-fold stimulation; 30 nM trypsin ϭ 12.6 Ϯ 2.1), which was not observed in wild type NCTC2544 cells or in cells transfected with empty vector. This response was comparable with the stimulation observed by 50 M UTP, a P2Y2 receptor agonist (-fold stimulation: 8.75 Ϯ 0.76), suggesting a level of PAR-2 expression comparable with other endogenous receptors. Trypsin and SLIGKV also stimulated p42/44 MAP kinase activation, but only in cells expressing hPAR-2. The level of MAP kinase activity was also comparable to stimulation with UTP. Furthermore, a comparison of PAR-2 mRNA levels in clone G and human primary keratinocytes by RT-PCR indicated that the level of receptor mRNA expression was similar ( Fig. 1). This suggests that, in clone G cells, the level of PAR-2 expression is not substantially greater than might be expected for endogenous expression in other cell types.
However, in contrast to the moderate activation of p42/44 MAP kinase, trypsin strongly stimulated both the known SAP kinases JNK and p38 MAP kinase (Figs. 2 and 3). The activation of both kinases was rapid in onset and reached a peak between 25 and 30 min before returning to basal values by 90 min. Maximum activation was 22.3 Ϯ 6.6-and 24.9 Ϯ 9.3-fold for JNK and p38 MAP kinase, respectively, a level of stimulation comparable with TNF-␣ (ϳ20-fold). The PAR-2 peptide mimetic SLIGKV also stimulated both JNK and p38 MAP kinase (Figs. 2 and 3B). However, the degree of activation of both kinases was less than that observed for trypsin and the duration of activation was much shorter. This was not due to any nonspecific effect of trypsin, since no activation of SAP kinase activity was observed in response to trypsin in nontransfected cells (Fig. 2C). In additional experiments, trypsin was also found to stimulate strongly MAPKAP kinase-2 activity ( Fig. 3D) but not in vector control cells.
In addition to activation of SAP kinase signaling, trypsin (30 nM) or SLIGKV (200 M) stimulated a strong increase in NFB DNA binding in NCTC2544 cells. This was observed as early as 30 min and was maximal by 60 min (Fig. 4, A and B). Supershift analysis using anti-NFB antibodies indicated that the stimulated DNA-binding complex consisted of p50 and p65 (data not shown). Activation of NFB DNA binding was also accompanied by stimulation of NFB luciferase reporter activity (Table I) and the loss in expression of both ␣ and ␤ isoforms of IB. Furthermore, we observed strong activation of both isoforms of IKK by trypsin (Fig. 4, C and D). An increase in activity of both IKK␣ and IKK␤ was observed as early as 10 min, with maximum stimulation by 30 min (3.7 Ϯ 0.2-and 5.4 Ϯ 0.1-fold, respectively; n ϭ 3, Fig. 4E), which was again comparable to that observed with the cytokine TNF-␣.
Further experiments found that PAR-2 coupling to the NFB pathway was not a result of transfection conditions. In human basal keratinocytes, trypsin stimulated both NFB-DNA binding activity and JNK activation with kinetics similar to that observed in the transfected cells (Fig. 5). The level of trypsinstimulated NFB-DNA binding was again comparable to that observed in response to TNF-␣.
We also found the known activator of protein kinase C, PMA, was also a strong activator of both SAP kinases (Fig. 6). PMA strongly activated JNK and p38 MAP kinase activity with kinetics similar to that observed for trypsin. Maximal activity of ϳ20 -30-fold was observed within 30 min before returning to  PMA also strongly stimulated NFB-DNA binding activity (Fig. 7). Maximal DNA binding was obtained between 30 and 60 min and was accompanied by a strong increase in the activity of both IKK␣ and IKK␤ isoforms with kinetics similar to those observed for trypsin and TNF-␣ (Fig. 7, B-D).
In order to investigate the role of protein kinase C isoforms in the regulation of PAR-2-mediated SAP kinase and NFB signaling, selective inhibitors of PKC isoforms and DN-PKC mutants were utilized. Pre-treatment of cells with GF109203X, which inhibits PKC isoforms ␣, ␤, ␦, and ⑀ but which shows some preference for PKC␣ (33,34), abolished PMA-stimulated JNK and p38 MAP kinase activity (Fig. 8, A-C). However, over a number of experiments, it was found that trypsin-stimulated JNK and p38 MAP kinase activities were both reduced by ϳ30%, respectively (Fig. 8, A-C). In addition, following transfection of the cells with DN-PKC␣ (35), PMA-stimulated activation of HA-JNK1 was reduced by 80%, whereas trypsinstimulated JNK-1 activity was again only partially reduced (Fig. 9). This inhibition was ϳ20% and correlated well with the effects observed with the PKC inhibitor. Furthermore, transfection with DN-PKC did not affect PMA-or trypsin-stimulated HA-JNK1 activity (Fig. 9), suggesting no input from this atypical isoform. Additional control experiments showed that transfection of clone G with either DN-PKC␣ or DN-PKC did not affect JNK activity in response to TNF-␣ (data not shown).
However, pre-treatment with another isoform selective inhibitor of PKC Gö6983, which, in addition to effects similar to GF109203X, is capable of inhibiting the atypical PKC and also PKCs ␥ and ␦ (41), revealed the potential for other isoforms to regulate trypsin-stimulated p38 MAP kinase activity. In addition to abrogating PMA-stimulated p38 MAP kinase activity, increasing concentrations of Gö6983, while being relatively ineffective against trypsin-stimulated JNK activity (Fig. 10 and C), strongly inhibited the p38 MAP kinase response to trypsin (Fig. 10, B and C). Trypsin-stimulated p38 MAP kinase activity was reduced by 75%, whereas JNK activity was reduced by ϳ20%.
However, results obtained using DN-PKC␣ and DN-PKC indicated that trypsin-stimulated p38 MAP kinase activity was not under additional control from PKC (Fig. 11). Furthermore, although DN-PKC␣ abolished PMA-stimulated p38 MAP kinase activity (data not shown), it had only a marginal effect upon trypsin-stimulated FLAG-p38 MAP kinase activity (Fig.   11) even when the transfected ratio of DN-PKC␣ to p38 MAP kinase was increased from 3:1, the ratio used in the JNK assays, to 10:1. Similarly, DN-PKC was without effect on trypsin-stimulated FLAG-p38 MAP kinase activity (Fig. 11).
The effect of PKC inhibition was also examined using transfection of FLAG-and Myc-tagged IKK isoforms (Fig. 14). In preliminary experiments transfection of FLAG-IKK␣ alone generated very low basal and agonist-stimulated activity, whereas transfection with IKK␤ alone generated extremely high basal activity with little additional stimulated activity. However, co-transfection of Myc-IKK␣ with IKK␤ at a ratio of 9:1 generated low basal conditions and robust responses to PMA, trypsin, and TNF-␣. Under these co-transfection conditions, PMA-stimulated FLAG-IKK␤ activity was abolished by the expression of DN-PKC␣, whereas the response to trypsin was also substantially reduced by this mutant (Fig. 14, A and  B). Co-transfection of DN-PKC did not, however, result in any further inhibition of IKK␤ activity. Myc-tagged IKK␣ activity could not be reliably measured under these conditions. The effect of selectively inhibiting PKC isoforms was also assessed at the level of NFB-DNA binding activity (Fig. 15). Pre-incubation of cells with increasing concentrations of GF109203X or Gö6983 abolished PMA-stimulated NFB-DNA binding; however, surprisingly, both interventions only partially inhibited trypsin-stimulated activity (maximal reductions of 67.2 Ϯ 1.7% and 53.2 Ϯ 1.0% for GF109203X and Gö6983, respectively (Fig. 15, B and D).
As trypsin-stimulated IKK activation was largely abolished by PKC inhibitors, whereas NFB-DNA binding was only partially inhibited, we assessed the IKK dependence of trypsin and TNF-␣-stimulated NFB reporter activity using DN forms of IKK␣ and -␤ (Table I). We found that introduction of either mutant alone or in combination reduced the absolute level of trypsin-stimulated reporter activity. However, this was accompanied by a reduction in basal reporter activity such that the -fold stimulation remained approximately the same (4 -7-fold). In contrast, the -fold stimulation with TNF-␣ was markedly reduced by either DN-IKK␣ or ␤ alone (-fold stimulation: TNF-␣ ϭ 7.9-fold, ϩ DN-IKK␣ ϭ 1.5-fold, ϩDN-IKK␤ ϭ 2.0-fold) and abolished in response to both mutants in combination.

DISCUSSION
In this study we examined the actions of trypsin, mediated through the novel G-protein-coupled receptor PAR-2, upon two key signaling pathways relevant to keratinocyte function: the SAP kinase pathway and the NFB pathway. To do this we expressed hPAR-2 in NCTC2544 cells, a cell line with characteristics of keratinocytes (32), which does not endogenously express PAR-2 or PAR-1. Trypsin-and SLIGKV-stimulated accumulation of [ 3 H]IP confirmed the coupling of PAR-2 to the IP 3 /Ca 2ϩ signaling cascade, a phenomenon that has been previously demonstrated in a number of cell types (13,14). This level of [ 3 H]IP accumulation was comparable with that observed with UTP, suggesting that the level of receptor expression is not prohibitively high relative to endogenous receptor expression. This was confirmed by using RT-PCR, which showed that the level of mRNA for PAR-2 in clone G was comparable to human keratinocytes (Fig. 1). However, it should be noted that these results do not provide an absolute estimate as to the level of cell surface receptor expression in the clone G cells. The level of mRNA expression provides only an indication that the level of receptor is not unrealistically high in comparison to endogenous expression of PAR-2 in another cell type.
In initial studies we found moderate activation of p42/44 MAP kinase in response to trypsin as assessed by Western blotting. 2 However, in contrast to this response, we found that both trypsin and SLIGKV strongly stimulated SAP kinase activity in NCTC2544 cells. Both JNK and p38 MAP kinase were activated to a level comparable with the cytokine TNF-␣, demonstrating for the first time a link between PAR-2 and SAP kinase activation.
PAR-2-mediated activation of JNK and p38 MAP kinase was further confirmed using the peptide SLIGKV, which mimics the N-terminal sequence generated by trypsin mediated cleavage of PAR-2. However, it should also be noted that stimulation of SAP kinase activity is more robust with the enzyme relative to the peptide. This phenomenon has been observed previously in relation to thrombin and TRAP stimulation of PAR-1 associated signaling events (42,43) and reflects differences in the efficiency of receptor activation by trypsin relative to the peptide. Indeed, many studies indicate that concentration of 1-200 M SLIGKV may not maximally activate PAR-2 (13,14).
A previous study from our laboratory has shown that, in another cell type, where PAR-2 is endogenously expressed, PAR-2 is specifically linked to p42/44 MAP kinase activity (19), suggesting that coupling to SAP kinase may be due to the levels of expression of PAR-2 in NCTC2544. However, UTP, acting through the G-protein-coupled P2Y2 receptor, also stimulated p42/44 MAP kinase, indicating a similarity in receptor levels and efficiency in coupling to downstream signaling systems between exogenous PAR-2 and endogenous P2Y2. It is thus possible that PAR-2 is more efficiently connected to SAP kinase signaling in certain cell types, and, in preliminary experiments with primary cultures of endothelial cells and human keratinocytes, we have found a similar pattern of JNK activation induced through endogenous PAR-2. 2 In examining the NFB pathway, we found that trypsin and, to a lesser extent, SLIGKV strongly stimulated NFB-DNA binding. This in itself is not unusual, as several agents, including some G-protein-coupled receptor agonists, have been shown previously to induce modest increases in NFB activation through indirect mechanisms, possibly involving pp60 src or p42/44 MAP kinase/p90 s6k (31,44). However, in a number of preliminary studies, we found no evidence of p42/44 MAP kinase involvement in the regulation of PAR-2-mediated NFB activity, as PD098059 did not modify NFB-DNA binding activity. 3 Rather, in this cell type, we have found that activation of PAR-2 results in stimulation of the upstream mediator IKK. A similar result was observed for TNF-␣, raising the possibility that PAR-2 and the TNF-␣ receptor expressed in this cell type could share some common signaling elements relevant to IKK activation. PAR-2 coupling to NFB signaling is not restricted to transfected cell systems, since we have shown that trypsin also stimulates a robust increase in NFB-DNA binding in primary cultures of human keratinocytes (Fig. 5). Indeed, a recent study has shown PAR-2 mediated activation of NFB in coronary smooth muscle cells (29) and the activation of IKK though other G-protein-coupled receptors such as bradykinin (45), indicating the potential relevance this pathway has in the cellular actions mediated through this class of receptor.
We found that activation of PMA strongly stimulated SAP kinase activation and IKK signaling. PMA has been shown previously to be a moderate activator of SAP kinase signaling in some cell types (46,47) while being inactive in others (16) and has also been linked to inhibition of SAP kinase activity (39,48). However, activation of PKC is strongly implicated in keratinocyte cell differentiation (48) through effects upon AP-1 and CREB, events known to be distal to JNK and p38 MAP kinase activation (49). PMA-stimulated NFB DNA binding has similarly been demonstrated in keratinocytes (50,51), and it is possible that, at least in this cell type, PAR-2 may activate both cascades by a common PKC-mediated pathway.
However, our experiments indicated clear differences in the regulation of PAR-2-mediated SAP kinase and IKK signaling by PKC isoforms. Inhibition of PKC isoforms by GF109203X, Gö6983, and introduction of DN-PKC␣ only partially reduced trypsin-stimulated JNK activity, suggesting a small involvement for conventional PKCs. Experiments using GF109203X and Gö6983 also suggest a lack of involvement for PKC⑀ and some other novel or atypical isoforms, as these compounds are able to inhibit isotypes such as PKC⑀, -␦, and - (34). Using the pleotropic DN-PKC␣ and -mutants supported these observations, particularly since these mutants have the ability to inhibit several different isoforms in addition to PKC␣ and PKC. Although these results point to a predominantly PKC-independent pathway for PAR-2-mediated regulation of JNK, some PKC isoforms have been omitted, and it is possible that an isoform highly expressed in keratinocytes, such as PKC (52,53), may be involved.
The finding that p38 MAP kinase activation was much more sensitive to Gö6983 than GF109203X indicates differences in the roles played by PKC isoforms in the regulation of either JNK or p38 MAP kinase following PAR-2 activation. Unexpectedly, experiments using DN mutants of PKC␣ and -also suggested no role for these isoforms, although p38 MAP kinase was inhibited by Gö6893, which has been shown to inhibit PKC (40,54). Preliminary experiments in our laboratory have shown that PAR-2-mediated p38 MAP kinase activation is Ca 2ϩ -independent but is abolished by PMA pre-treatment. This suggests a role for an unidentified Ca 2ϩ -independent, diacylglycerol-dependent isoform of PKC, such as PKC␥ or -␦, which Gö6983 has been also been shown to inhibit (40,54). Apart from this study, little evidence supports differential regulation of p38 MAP kinase and JNK through different PKCs. Rather, other recent studies have indicated a role for both PKC-dependent and -independent inputs into SAP kinase activation mediated through G-protein-coupled receptors (47).
In contrast, conventional PKC isoforms were found to play a more substantial role in the regulation of IKK signaling. Trypsin-mediated activation of IKK␤ was largely abolished by GF109203X, Gö6983, and chronic treatment with PMA (results not shown), consistent with a role for typical PKC isoforms in this pathway. Indeed, PKC␣ has been implicated in PMA regulation of IKK␤ in HeLa cells (26), and our present study shows that this is also an intermediate involved in PAR-2 stimulation. Trypsin-mediated activation of IKK␣ was only partially affected by GF109203X and chronic PMA pretreatment (results not shown) and was further decreased but not abolished by Gö6983. This may also implicate an additional minor role for atypical PKC isoforms in the regulation of IKK␣; however, in general, PKC-mediated regulation of IKK␣ has not been shown to be a common feature of cytokine receptor stimulation. This may therefore be a feature of G-protein-coupled receptors that strongly activate PKC.
Studies to confirm the role of PKC isoforms in the regulation of PAR-2-mediated IKK activation were only partially successful. Transfection of clone G cells with IKK␤ generated a very high basal activity that was not increased upon agonist stimulation. This high basal activity could be lowered by co-transfection of IKK␣, such that agonist stimulation could be revealed. This confirms a recent study indicating that IKK␣ functions, in part, to regulate the endogenous activity of IKK␤ (55), therefore suggesting that under normal conditions endogenous IKK␣ and IKK␤ levels are comparable or favor IKK␣. This was confirmed by immunoblotting of IKK␣ and IKK␤, which showed comparable levels of the two kinases in clone G cells (data not shown). Using this co-transfection procedure, we were, nevertheless, able to confirm the important role for PKC␣ in the regulation of IKK␤ and the lack of effect of PKC. However, assay of IKK␣ either following transfection alone or with IKK␤ did not give consistent agonist-stimulated activity. Given that IKK␣ and IKK␤ exist as a complex involving IKK␥, NIK, and possibly other intermediates (56 -58), transfection of exogenous IKK is likely to disturb the equilibrium between the kinases such that and this may be reflected in the lack of IKK␣ activity.
Our studies also suggest that PKC-mediated regulation of both IKK␣ and IKK␤ may only play a minor role in the regulation of PAR-2-mediated NFB-DNA binding activity. Treatment with GF109203X only partially affected NFB-DNA binding activity, whereas Gö6983, which is more effective against IKK␣, caused a similar reduction (Fig. 13). This suggests that, in the case of PAR-2, either IKK-independent mechanisms are involved in the regulation of NFB or sufficient IB phosphorylation and degradation can be initiated via the remaining IKK signal. In support of the former hypothesis, we found that introduction of DN-IKK␣ and IKK␤ did not reduce trypsinstimulated NFB reporter activity, although their presence abolished the response to TNF-␣. However, these interpretations are difficult since, as discussed above, IKK␣ can unidirectionally regulate IKK␤ activity in cells both in a negative and positive manner (55). Therefore, introduction of DN mutants, particularly IKK␣, may enhance endogenous IKK activity. Alternatively, it is possible that IKK may be involved in other aspects of signaling that are directly involved in NFB translocation and DNA binding, and these possibilities are currently being investigated in our laboratory. However, the fact that TNF-␣-stimulated IKK activity was not affected by PKC inhibitors, while being abolished by DN-IKKs, indicates differences in the coupling of cytokine and G-protein-coupled receptors to IKK signaling events.
The results in this study do not necessarily exclude the possibility that other PKC-independent events may regulate IKK activity following PAR-2 stimulation. This includes involvement of NIK and protein kinase B (24,25). Furthermore, given the co-activation through PAR-2 of both IKK␣ and SAP kinase signaling, it is also possible that there are upstream intermediates regulating both IKK and SAP kinase signaling in response to PAR-2 activation. For example, it has been shown that MEKK-1 can substitute for NIK in the regulation of IKK by cytokines (59). Other MEKKs believed to be involved in the regulation of JNK and p38 MAP kinase, MEKK-2-4 (60, 61), have also been shown to regulate IKK signaling (62). These possibilities are currently being examined in our laboratory.
PAR-2-mediated regulation of IKK isoform activity may have important consequences for PAR-2 function in epidermal keratinocytes. Abnormal skin development, similar to psoriasis, has been recently reported in IKK␣ knockout mice (63,64), whereas in IKK␤ knockouts epidermal development was unaffected, although inflammatory mediator production was found to be compromised (65). Given that PAR-2 may regulate both normal keratinocyte differentiation and inflammatory mediator production, the differences in the regulation of IKK␣ and -␤ relative to NFB activation in keratinocytes exposed to trypsin, or other proteinases, may provide the basis for specific therapeutic intervention.