P2Y receptor-mediated inhibition of tumor necrosis factor alpha -stimulated stress-activated protein kinase activity in EAhy926 endothelial cells.

In the EAhy926 endothelial cell line, UTP, ATP, and forskolin, but not UDP and epidermal growth factor, inhibited tumor necrosis factor alpha (TNFalpha)- and sorbitol stimulation of the stress-activated protein kinases, JNK, and p38 mitogen-activated protein (MAP) kinase, and MAPKAP kinase-2, the downstream target of p38 MAP kinase. In NCT2544 keratinocytes, UTP and a proteinase-activated receptor-2 agonist caused similar inhibition, but in 13121N1 cells, transfected with the human P2Y(2) or P2Y(4) receptor, UTP stimulated JNK and p38 MAP kinase activities. This suggests that the effects mediated by P2Y receptors are cell-specific. The inhibitory effects of UTP were not due to induction of MAP kinase phosphatase-1, but were manifest upstream in the pathway at the level of MEK-4. The inhibitory effect of UTP was insensitive to the MEK-1 inhibitor PD 098059, changes in intracellular Ca(2+) levels, or pertussis toxin. Acute phorbol 12-myristate 13-acetate pretreatment also inhibited TNFalpha-stimulated SAP kinase activity, while chronic pretreatment reversed the effects of UTP. Furthermore, the protein kinase C inhibitors Ro318220 and Go6983 reversed the inhibitory action of UTP, but GF109203X was ineffective. These results indicate a novel mechanism of cross-talk regulation between P2Y receptors and TNFalpha-stimulated SAP kinase pathways in endothelial cells, mediated by Ca(2+)-independent isoforms of protein kinase C.

protein kinases (p42/44 MAP kinase and c-Jun N-terminal kinase (JNK)), phosphatidylinositol 3-kinase, focal adhesion kinase pp125 fak , and related focal adhesion tyrosine kinase (7)(8)(9)(10)(11)(12)(13). Activation of these kinases may underlie the sustained effects of ATP and UTP in smooth muscle and endothelial cells, such as increased cell proliferation (3, 14 -16). However, another potential pro-mitogenic effect of UTP and ATP may be to prevent the pro-apoptotic actions of cytokines in cells where P2Y and cytokine receptors are co-expressed. A number of studies have shown that growth factors, in particular insulinlike growth factor-1, can negatively regulate the stress-activated protein (SAP) kinases JNK and p38 MAP kinase through activation of p42/44 MAP kinase (17)(18)(19). However, to date no study has identified a similar role for G-protein-coupled receptors, which in many cells are linked to increases in JNK and p38 MAP kinase activity (20 -22).
In this study we examined the effect of P2Y receptor stimulation upon tumor necrosis factor ␣ (TNF␣)-stimulated SAP kinase activity in endothelial cells. Rather than activating SAP kinases, as shown in other cell types (9), UTP and ATP caused a marked inhibition of TNF␣-stimulated JNK and p38 MAP kinase activity. This effect, which was cell type-specific, was not mediated by p42/44 MAP kinase, nor intracellular Ca 2ϩ , but required activation of atypical Ca 2ϩ -independent isoforms of PKC. This is the first study to identify such an action of a P2Y receptor.
Products, Hertfordshire, UK) were all of the highest commercial grade available.
Western Blotting and MEK-4 Immunoprecipitation-Western blotting of p42/44 MAP kinase using SDS-polyacrylamide gel electrophoresis was conducted as outlined previously (25). All antibodies were titered for optimum conditions. For immunoprecipitation of MEK-4, equal amounts of precleared cell lysates were incubated in Tris-HCl buffer, pH 7.4, containing 1 mM EDTA, 1 mM EGTA, 1% (w/v) Triton X-100, and 0.5% (w/v) Nonidet P-40 and with 1 g/ml MEK-4 antibody, precoupled to protein G, for 3 h. Lysates were recovered by sequential washes in solubilization buffer and in the same buffer lacking detergents. Precipitates were assessed for either MEK-4 or phospho-MEK-4 content using Western blotting.
JNK and p38 MAP Kinase Activity Assay-To measure JNK activity, cells were stimulated as appropriate and the reaction terminated by rapid aspiration and the addition of ice-cold PBS. The cells were solubilized in 20 mM HEPES buffer, pH 7.7, containing 50 mM NaCl, 0.1 mM EDTA, 0.1 mM Na 3 VO 4 , 0.1 mM PMSF, 10 g/ml aprotinin, 10 g/ml leupeptin, and 1% (w/v) Triton X-100. Lysates were clarified by centrifugation for 5 min at 13,000 rpm and supernatants incubated with 5 g of GST-c-Jun-(5-89) immobilized on glutathione-Sepharose at 4°C for 3 h. Beads were then washed three times in solubilization buffer and twice in 25 mM HEPES buffer, pH 7.6, containing 20 mM ␤-glycerophosphate, 0.1 mM NaV 3 O 4 , 2 mM dithiothreitol. Precipitates were then incubated with the same buffer containing 25 M/1 Ci of ATP/[␥-32 P]ATP in a final volume of 30 l for 30°C for 30 min. The reactions were terminated by the addition of 4 ϫ SDS-sample buffer and aliquots of each sample subjected to electrophoresis on 11% SDS-polyacrylamide gel electrophoresis. Phosphorylation of GST-c-Jun was then determined by autoradiography.
For p38 MAP kinase activity, a procedure identical to that outlined above was employed except that the pH of the solubilization buffer was modified to pH 7.4. Full-length MAPKAP kinase-2 was used in the precipitation step, and the kinase buffer employed for the phosphorylation reaction was pH 7.4.
Indirect Immunofluoresence-Cells grown on coverslips were stimulated as appropriate, and the reaction terminated by rapid aspiration and the addition of PBS. Cells were fixed with paraformaldehyde (20 min). After washing three times with PBS, the reaction was quenched (50 mM NH 4 Cl/PBS, 10 min). Following a further three washes, cells were incubated in 0.1% (w/v) Triton X-100 for 4 min. Nonspecific binding was blocked by washing three times in PBS, goat serum, 0.2% (w/v) gelatin for 5 min and three times in PBS. The relevant antibody was diluted as appropriate and placed on coverslips for 60 min. Coverslips were washed as outlined above and then incubated with a secondary fluorescein isothiocyanate-labeled antibody, washed, and then fixed. Coverslips of fixed cells were imaged using a Zeiss 4 Laser Scanning Confocal microscope operating in confocal mode using either 40ϫ, 63ϫ, or 100ϫ Plan-APOCHROMAT 1.4 NA oil-immersion objectives. Images were collected using a 488-nm laser and a long 510-nm filter set. Data files were saved in TIF format and analyzed using MetaMorph software (Universal Imaging, West Chester, PA).

RESULTS
Initially, we examined the effects of UTP (30 M) and forskolin (10 M) upon TNF␣-stimulated signaling events in EAhy926 cells. Preincubation with either agent for 60 min inhibited TNF␣ (20 ng/ml)-induced activation of both JNK and p38 MAP kinase by about 80% (Fig. 1). When sorbitol (0.5 M) was used as the stimulant, forskolin and UTP were both less effective, only producing about 50% inhibition. ATP (30 M) also inhibited TNF␣-induced SAP kinase activity, but UDP (30 M) and EGF (100 nM) were ineffective. UTP had no effect upon the increase in IKK␣ activity or loss in IB␣ expression evoked by TNF␣ (not shown), indicating that the site of inhibition of SAP kinase activity by UTP is downstream of the NFB signaling cascade.
UTP also inhibited enzyme activity downstream of p38 MAP kinase. Both sorbitol and TNF␣ increased MAPKAP kinase-2 activity in EAhy926 cells by 3-5-fold (Table I). Preincubation with UTP or ATP (30 M) reduced the increase by over 80% (Table I). Forskolin (10 M) only inhibited the TNF␣-induced increase in MAPKAP kinase-2 activity by about 30% and had little effect on the sorbitol-induced activity. This is much less than the inhibition of p38 MAP kinase by forskolin and may be due to an indirect stimulation of MAPKAP kinase-2 by cAMPraising agents. 2 The similar effects of UTP and ATP, but not UDP, on TNF␣stimulated SAP kinase activity suggested that P2Y 2 and/or P2Y 4 receptors (5, 6) could be present in EAhy926 cells. To characterize the interaction of these receptors with SAP kinases, we used 1321N1 human astrocytoma cells expressing recombinant human P2Y 2 or P2Y 4 receptors. However, in these cells, UTP (1-30 M) stimulated both JNK and p38 MAP kinase activity by 3-5-fold and did not inhibit TNF␣-mediated increases in SAP kinase activity over the micromolar range (  (200 M) (Fig. 2, panel C). This suggests that the inhibitory effect of UTP may be cell type-specific.
Next, several possible mechanisms of inhibition for UTP were examined. We have shown previously that UTP stimulates the induction of MAP kinase phosphatase-1 (MKP-1) in EAhy926 cells (10). Therefore, we investigated if JNK and p38 MAP kinase are co-localized with MKP-1. UTP (30 M) and forskolin (10 M) substantially induced MKP-1, and this, as expected, was restricted to the nucleus (Fig. 3). However, JNK was evenly distributed throughout the cell, but p38 MAP kinase was restricted to the cytosol and TNF␣ did not evoke relocation of either to the nuclear compartment (Fig. 4). Thus, dephosphorylation of p38 MAP kinase and JNK by MKP-1 is unlikely to play a role in the inhibitory actions of UTP.
Data showing that the site of inhibition of p38 MAP kinase and JNK in EAhy926 cells is likely to be upstream of JNK are shown in Fig. 5. Immunoblotting with phospho-JNK antibodies (Fig. 5, panel A) showed a similar pattern of UTP mediated inhibition of TNF␣-stimulated JNK activation (p46 and 54 isoforms) and confirmed initial experiments using solid phase assays. In addition, activation of the signaling cascade at the level upstream of JNK was estimated using antibodies specific for phospho-MEK-4 ( Fig. 5, panel B). Using this method we found that preincubation with either UTP or ATP strongly inhibited TNF␣-stimulated MEK-4 phosphorylation. This was not due to effects upon the recovery of MEK-4, which was found to be similar throughout each stimulation. Additional experiments using a phospho-MEK3/6 antibody showed a similar phenomenon (not shown) although the quality of the antibody precluded a more detailed analysis.
The UTP-mediated inhibition of TNF␣-stimulated SAP kinase activity was very rapid (Fig. 6). Preincubation with UTP (30 M) for as little as 5 min prior to TNF␣ addition, or even simultaneous administration, essentially abolished the TNF␣induced increase in JNK activity over 30 min. Addition of UTP after TNF␣ had stimulated JNK still resulted in a substantial inhibition, which was only overcome after 15-25 min. A similar time course of inhibition was observed for p38 MAP kinase (not shown). These results confirm the unlikeliness of an inducible phosphatase playing a role in the negative regulation of SAP kinase activity by UTP.
The next experiments were designed to determine which UTP-mediated signaling events are important for the inhibition of SAP kinase activity. Pertussis toxin pretreatment did not affect the inhibitory actions of UTP on JNK activity (not shown), ruling out a Gi2 protein-dependent mechanism. Preincubation with the MAP kinase kinase-1 (MEK-1) inhibitor PD 098059 (50 M) (27)(28)(29), 30 min prior to addition of UTP, prevented the UTP-stimulated increase in p42/44 MAP kinase activity, as estimated by MAP kinase gel-shift analysis (Fig. 7,  panel A). However, PD 098059 did not substantially reverse the  inhibitory effects of UTP upon TNF␣-stimulated JNK or p38 MAP kinase activity (Fig. 7, panels B and C). This suggests that MAP kinase kinase-1 is not involved in the inhibitory actions of UTP. Consistent with this conclusion, EGF (100 nM) strongly stimulated p42/44 MAP kinase, but did not inhibit either SAP kinase (not shown).
UTP is known to mobilize Ca 2ϩ in endothelial cells and so the role of Ca 2ϩ was investigated. Bathing EAhy926 cells in Ca 2ϩ -free medium and BAPTA-AM (50 M) did not reverse the inhibitory effect of UTP upon TNF␣-stimulated JNK activity (Fig. 8, panel A). Furthermore, the Ca 2ϩ ionophore A23178 (1 M) strongly stimulated rather than inhibited JNK activity, such that JNK activity increased in a time-dependent manner, reaching a peak within 30 min at some 8 -10-fold of basal values (Fig. 8, panel B). A23178 was additive with TNF␣ in stimulating JNK activity.
Finally, a possible role of PKC in the inhibitory actions of UTP was investigated. Acute incubation with the PKC activator PMA (100 M) substantially inhibited the TNF␣-induced JNK activation in a manner similar to that observed for UTP (Fig. 9, panel A), while chronic pretreatment with PMA partially reversed the inhibitory effects of UTP upon TNF␣-stimulated JNK activity (Fig. 9, panel B). Consequently, we examined the effect of several PKC inhibitors upon TNF␣stimulated JNK activity (Fig. 10). GF109203X (500 nM) reversed PMA-mediated inhibition of TNF␣-stimulated JNK activity, but did not modify the inhibitory effects of UTP (Fig.  10, panel A). In contrast, preincubation with Ro318220 (1 M) and Go6983 (1 M) resulted in a total reversal of the inhibitory effects of UTP (Fig. 10, panels B and C). DISCUSSION In this study we have identified a novel mechanism of crosstalk regulation between the P2Y receptor and TNF␣ receptor signaling pathways in endothelial cells. UTP exerted a strong inhibitory effect upon TNF␣-and sorbitol-stimulated JNK and p38 MAP kinase activity, an effect that was also seen downstream of p38 MAP kinase, at the level of MAPKAP kinase-2. It has been shown recently that UTP and other G-protein-coupled receptor agonists, such as thrombin and angiotensin II, can stimulate both JNK and p38 MAP kinase in a number of cell types (20 -22). To our knowledge this is the first study to identify an inhibitory action of a G-protein-coupled receptor, the P2Y receptor, upon TNF␣-stimulated SAP kinase activation.
In these experiments, ATP had similar effects to UTP, but UDP was inactive. This rules out the P2Y 6 subtype as the site of action, as UDP is the most potent agonist at this receptor (5,6,24). UTP is not an agonist at the P2Y 1 and P2Y 11 subtypes (5,6), suggesting that the responses seen here are mediated via the P2Y 2 and/or P2Y 4 subtypes, consistent with our previous studies in EAhy926 cells (8,10). (Note that although ATP is not an agonist per se at the human P2Y 4 receptor, in a static culture system as used here, it can donate a phosphate group to endogenous UDP to produce UTP, which will activate the P2Y 4 receptor (see Refs. 30 and 31). At present, it is difficult to differentiate pharmacologically between the P2Y 2 and P2Y 4 receptors due to the lack of selective ligands. Also, we cannot yet rule out the possibility that more than one P2Y receptor subtype is expressed in EAhy926 cells and that UTP and ATP act at separate sites.
UTP also inhibited the TNF␣-induced responses in NCTC2544 keratinocytes. However, in 1321N1 human astrocytoma cells stably expressing the recombinant human P2Y 2 or P2Y 4 receptors, we found that UTP alone significantly increased JNK activity and failed to reverse TNF␣-mediated SAP kinase activation. This suggests that the effects mediated by P2Y receptors are dependent upon the cell type under study and that the P2Y receptor in EAhy926 cells and NCTC2544 keratinocytes couples to additional components. In support of this idea, the inhibitory effect of UTP is seen at lower concentrations (IC 50 ϳ1 M) than we and others have previously observed for P2Y receptor-mediated excitatory responses, including activation of p42/44 MAP kinase (8, 13) and JNK (9) and generation of [ 3 H]inositol phosphates (not shown).

P2Y Receptor Inhibition of Stress-activated Protein Kinases
We sought to identify the site(s) and mechanism(s) responsible for the inhibitory action of UTP. One possibility was a role for an inducible MAP kinase phosphatase, in particular MKP-1, which we have shown previously to be induced by UTP and forskolin in EAhy926 cells (10). However, the data did not support such a role. Agonist-stimulated MKP-1 expression was located exclusively in the nucleus, while p38 MAP kinase was restricted to the cytosolic compartment and JNK was distributed evenly across the cell. TNF␣ and UTP did not evoke translocation of either SAP kinase to the nucleus. This was not due to a general lack of cellular responsiveness, as nuclear translocation of p42/44 MAP kinase was observed following stimulation with PMA 3 and suggests that the normal inducible nuclear phosphates MKP-1 and MKP-2, which regulate JNK and p38 MAP kinase within the nucleus (10,32), are not involved in the inhibitory actions of UTP.
It may be that other cytosolic MAP kinase phosphatases which display substrate specificity for JNK and p38 MAP kinase, such as M3/6 and MKP-5 (33,34), are involved in the negative regulation of SAP kinases. However, we found that the site of action of UTP was upstream of JNK and p38 MAP kinase, at the level of MEK or above. Inhibition of TNF␣stimulated MEK-4 activation by UTP implies that dephosphorylation at serine and threonine of MEKs or MEKKs, possibly by PP2A (33), or uncoupling of the pathway further upstream is a far more likely mechanism of action.
This study showed that several intracellular signaling pathways activated by UTP via P2Y receptors are not involved in its inhibitory effects. p42/44 MAP kinase is activated strongly by UTP in EAhy926 cells (8,10) and has been implicated in the inhibition of JNK and p38 MAP kinase activity in other cell types. However, the MEK-1 inhibitor PD 098059, at concentrations that abolished MAP kinase activity (27)(28)(29), failed to reverse the effects of UTP upon JNK and p38 MAP kinase activation. Furthermore, EGF, a robust activator of the MAP kinase cascade (35,36), did not mimic the inhibitory effects of UTP. The lack of effect of EGF also argued against a role for growth factor-mediated tyrosine kinases and/or phosphatases in the inhibition of TNF␣-stimulated SAP kinase activation.
Similarly, we found no evidence for the involvement of a Ca 2ϩ -dependent mechanism in the inhibitory effects of UTP. Removal of extracellular Ca 2ϩ and buffering of intracellular Ca 2ϩ did not suppress the UTP inhibition of TNF␣-evoked increased in JNK activity. Also, elevation of intracellular Ca 2ϩ by A23187 gave a strong and rapid stimulation of JNK activity. This suggest that the normal cellular mechanisms by which Ca 2ϩ can increase JNK activity are present in EAhy926 cells. This contrasts with the finding that in cells transfected with the P2Y 2 receptor, UTP evokes large increases in inositol trisphosphate and intracellular Ca 2ϩ levels (37) and increases JNK activity. Furthermore, previous studies have shown JNK activation by G-protein-coupled receptor agonists to be Ca 2ϩdependent (38,39), further arguing against a role for Ca 2ϩ in inhibitory effects of UTP.
While the inhibitory effects of forskolin imply a role for cAMP-dependent signaling events in the inhibitory effects of UTP, our previous study (10) showed that UTP does not increase cAMP in EAhy926 cells. Thus, forskolin and UTP may act by different mechanisms to inhibit SAP kinases. The present studies have revealed one such potential mechanism for UTP, involving PKC. Short term incubation with PMA mimicked the actions of UTP upon TNF␣-stimulated JNK activity, while chronic pretreatment partially reversed the inhibitory action of UTP. This suggests a role for DAG-sensitive, Ca 2ϩindependent isoforms of PKC.
We showed previously that EAhy926 cells express the PKC␣ (Ca 2ϩ -dependent) and PKC⑀ (Ca 2ϩ -independent) isoforms that are down-regulated by chronic PMA treatment (8). As the inhibitory effects of UTP are independent of Ca 2ϩ , this clearly suggests that PKC⑀ is the more likely to be involved or possibly PKC␦, which is rapidly tyrosine phosphorylated and activated by UTP in PC12 cells (13). Both possibilities were supported by the effects of PKC inhibitors. Ro318220, a nonselective PKC inhibitor (40), fully reversed the effect of UTP. GF109203X, at a concentration (500 nM) that is relatively selective for Ca 2ϩdependent PKC isoforms (41,42), had no effect on UTP, but did reverse the inhibitory effect of PMA. Furthermore, GF109203X, at higher concentrations (3-10 M) that are likely to also inhibit PKC⑀ (42), reversed the effects of UTP (not shown).
At present we cannot rule out that other PMA-insensitive PKC isoforms are also involved, in particular atypical isoforms such as PKC, which is reported to be sensitive to Ro318220 in some cell types (43). Supporting this possibility is our finding that Go6893, which also inhibits PKC (44), also reversed the inhibitory effects of UTP. If PKC is present in EAhy926 cells and also involved in the inhibitory actions of UTP, then this may explain why the effect of PMA has a slower onset and is a less effective inhibitor of the SAP kinase responses than UTP. Alternatively, it may be that other PKC-independent mechanisms are also involved and PKC only plays a conditional role in the inhibitory effects of UTP.
The involvement of PKC isoforms in the inhibition of JNK activity distinguishes our findings from recent studies in other cells types. For example PKC␤ stimulates SAP kinase activity in U-937 and HL60 cells (45). However, in the EAhy926 line used in this present study, PKC␤ isoforms are poorly expressed (8). Thus, the cell-specific expression of PKC isoforms may dictate which effects are manifest upon SAP kinase signaling.
In conclusion, we have identified a novel mechanism of crosstalk regulation between the P2Y receptor and TNF␣ receptor signaling pathways in endothelial cells, involving PKC. Activation of the recombinant human PAR-2 receptor expressed in NCTC2544 keratinocytes also suppressed TNF␣ receptor-mediated increases in SAP kinase activity. This suggests that such cross-talk can occur for other G-protein-coupled receptors.