INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

PROSTACYCLIN (PGI₂), a potent vasodilator and inhibitor of platelet aggregation, is synthesized by vascular endothelial cells (ECs) in response to a variety of stimuli. G protein-coupled receptor agonists (e.g., thrombin) generally elicit a very rapid, short-lived PGI₂ release (8, 44, 46). In contrast, early studies in ECs revealed that PGI₂ generation in response to proinflammatory cytokines only became apparent after continuous exposure to agonist for several hours (24, 31). Three key enzymatic steps are involved in PGI₂ generation, all of which are potentially important regulatory sites. Phospholipase A₂ (PLA₂) enzymes catalyze the cleavage of arachidonic acid (AA) from the sn-2 position of phospholipids (e.g., Ref. 21) with subsequent conversion of free AA into prostaglandin (PG) H₂ by two distinct isoforms of cyclooxygenase (COX; Ref. 15); newly formed PGH₂ is then converted to PGI₂ by terminal PGI₂ synthase (26). The protracted PGI₂ generation in response to interleukin-1 (IL-1) has been attributed to enhanced expression of COX-2, the inducible form of COX, and to increased cytosolic PLA₂ (cPLA₂) expression (11, 15, 24, 31). More recently it was suggested that IL-1 may elicit an acute prostanoid release from ECs (6), but the early signaling events responsible for this synthesis remain undefined.

The post-receptor signaling pathways regulating cytokine-driven prostanoid synthesis in vascular endothelium are poorly understood. However, the pleiotropic effects of IL-1 are known to result in part from the triggering of signaling cascades involving activation of families of serine/threonine protein kinases collectively known as the mitogen-activated protein kinases (MAPKs). MAPKs are activated by phosphorylation of specific threonine and tyrosine residues within the signature sequence Thr-X-Tyr catalyzed by upstream, dual-specificity MAPK kinases (MEKs) (reviewed in Ref. 41). Of the three major MAPK subgroups, the c-Jun NH₂-terminal kinases (JNK/stress-activated protein kinases) and the p38 MAPKs (p38^mapk) are thought to be important in mediating cellular responses to extracellular stress. Thus ultraviolet radiation, proinflammatory cytokines, and lipopolysaccharide strongly activate these MAPK families (37), and p38^mapk has been implicated in the regulation of cytokine-stimulated E-selectin expression in endothelium (38). In contrast, the extracellular signal-regulated kinases (ERK1/2; also known as p42/p44^mapk) are typically activated in response to mitogenic stimuli and are generally poorly stimulated by exposure to proinflammatory cytokines (reviewed in Ref. 25). We have recently shown that IL-1 and tumor necrosis factor- (TNF-) transiently activate p42/p44^mapk in human ECs (28). Because this activation does not appear to be obligatory for IL-1-induced E-selection expression (45), and because its involvement in other IL-1-mediated responses has yet to be explored, the functional significance of activation of the ERK pathway by proinflammatory cytokines in ECs is unclear.

The 85-kDa cPLA₂ is one of a growing family of PLA₂ enzymes (35, 40, 42) implicated in stimulus-evoked AA mobilization and is regulated by at least two major posttranslational mechanisms, both of which are thought to be essential for the initiation of AA release: 1) calcium-stimulated membrane association via its calcium-binding domain (39), and 2) phosphorylation of Ser⁵⁰⁵ within the sequence Pro-Leu-Ser-Pro typically recognized by proline-directed protein kinases, such as members of the MAPK families. p42^mapk-mediated phosphorylation of Ser⁵⁰⁵ results in an increase in the specific activity of cPLA₂ (13) and a characteristic shift in the electrophoretic mobility of the phosphorylated, active form of the enzyme (33). Evidence from studies carried out in platelets and neutrophils also implicates the p38^mapk pathway in regulating the phosphorylation state of cPLA₂ (14, 18, 43). Our recent studies have shown that p42^mapk is likely to be an important regulator of PGI₂ generation in thrombin- and vascular endothelial growth factor (VEGF)-stimulated human umbilical vein endothelial cells (HUVEC) via its effects on cPLA₂ phosphorylation (44, 45), but the roles of the ERK and p38^mapk pathways in controlling cytokine-stimulated prostanoid production are unknown. In the present study we establish that IL-1 acutely releases PGI₂ from HUVEC and examine the involvement of cPLA₂ and the p42/44^mapk and p38^mapk pathways in this early synthesis. Our results demonstrate the importance of early ERK-mediated activation of cPLA₂ as a regulator of acute cytokine-stimulated PGI₂ release and identify a novel, receptor-specific cross talk between the p38^mapk and ERK pathways that may regulate the extent of prostanoid synthesis in agonist-stimulated human endothelial cells.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Materials. AA, human -thrombin, bovine serum albumin (BSA; fraction V), pepstatin, aprotinin, leupeptin, sodium fluoride, sodium pyrophosphate, sodium orthovanadate, and sodium glycerophosphate were all purchased from Sigma (Poole, Dorset, UK). 4-(2-Aminoethyl)benzenesulfonyl fluoride (AEBSF), myelin basic protein, and protein G agarose were obtained from Calbiochem (Nottingham, UK). Human recombinant IL-1 was from R&D Systems (Oxford, UK). PD-98059, SB-203580, and U-0126 were generously provided by Dr. A. Saltiel (Parke Davis), Dr. J. Lee (Smith Kline Beecham), and Dr. J. M. Trzaskos (DuPont Merck), respectively. Anti-ACTIVE MAPK antibody recognizing the active, dually phosphorylated forms of p42/p44^mapk, was from Promega (Southampton, UK); anti-p42/p44^mapk antibody was from Affiniti Research Products (Nottingham, UK). Phosphospecific anti-p38^mapk and anti-phospho-MEK1/2 antibodies were purchased from New England Biolabs (Beverly, MA). Anti-p38^mapk antibody, protein A/G PLUS agarose, and glutathione-S-transferase (GST)-activating transcription factor (ATF)-2^1-96 were from Santa Cruz (Santa Cruz, CA). The polyclonal anti-cPLA₂ antibody was a kind gift from Dr. R. Kramer (Eli Lilly, Indianapolis, IN). Horseradish peroxidase-conjugated goat anti-mouse/rabbit immunoglobulins were from Pierce and Warriner (Cheater, Cheshire, UK). Reagents for SDS-PAGE were from Bio-Rad (Hemel Hempstead, Hertfordshire, UK) and National Diagnostics (Hessle, Hull, UK). Polyvinylidene difluoride (PVDF) membranes (Immobilon-P) and I-Block were from Sigma and Tropix (Warrington, UK), respectively. Enhanced chemiluminescence (ECL) Western blotting detection reagents, Hyperfilm-ECL, and [5,6,8,9,11,12,14,15-³H]AA were all obtained from Amersham International (Amersham, UK). ¹²⁵I-labeled 6-ketoprostaglandin F₁ (6-keto-PGF₁) was purchased from Metachem Diagnostics (Piddington, Northampton, UK). [-³²P]ATP and Renaissance 4CN Plus were from DuPont (Dreieich, Germany). Raf-1 immunoprecipitation kinase cascade assay kits were from Upstate Biotechnology (Lake Placid, NY). Culture media were purchased from Sigma or Life Technologies (Paisley, UK). All other reagents were obtained from Sigma or BDH (Poole, Dorset, UK) at the equivalent of analytical reagent grade.

Cell culture. HUVEC were isolated by collagenase digestion using a modification of a procedure originally described by Jaffe et al. (16). Primary HUVEC cultures were grown in medium 199 (M199) supplemented with 20% (vol/vol) fetal calf serum, L-glutamine (5 mM), NaHCO₃ (25 mM), penicillin (50 U/ml), and streptomycin (50 U/ml) and maintained at 37°C in 5% CO₂-95% air. All tissue culture plastic was precoated with gelatin (1%) before the cells were plated. On reaching confluence (~1 × 10⁶ cells/25-mm² flask), primary cultures were trypsinized by brief exposure to a trypsin/EDTA solution (0.1%/0.025% in phosphate-buffered saline; PBS), plated onto 75-mm² flasks, and cultured in M199 additionally supplemented with heparin (90 µg/ml) and endothelial cell growth supplement (ECGS; 20 µg/ml). ECGS was prepared from porcine brain as described by Maciag et al. (23). When confluent (~3.5 × 10⁶ cells/75-mm² flask), cells were plated onto either 24-well tissue culture trays (~1 × 10⁵ cells/well) or 60-mm-diameter tissue culture dishes (~1 × 10⁶ cells/dish). Confluent passage 2 cells were subsequently used for experimentation 4-5 days after plating; all experiments were routinely conducted at 37°C.

Measurement of PGI₂ release. Confluent cultures of HUVEC in 24-well tissue culture trays were washed with serum-free M199 supplemented with HEPES (25 mM) and glutamine (5 mM) (H-M199, pH 7.4). Cell monolayers were treated as described in legends to Figs. 1A and 4, and the PGI₂ levels in the cell supernatants were quantified using a specific radioimmunoassay for 6-keto-PGF₁, the stable hydrolysis product of PGI₂, as previously described (46). Supernatants from HUVEC monolayers used in immunoblotting studies were also routinely assayed for 6-keto-PGF₁ content.

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Interleukin-1 (IL-1) induces early arachidonic acid (AA) mobilization and prostacyclin (PGI2) release from cultured endothelial cells. A: confluent human umbilical vein endothelial cells (HUVEC) in 24-well tissue-culture trays were exposed to human recombinant IL-1 (; 100 U/ml) or vehicle control () for the times indicated. Supernatants were sampled and assayed for 6-ketoprostaglandin F1 (6-keto-PGF1) as described in METHODS. B and C: confluent monolayers of HUVEC in 24-well tissue culture trays were prelabeled with [3H]AA for 24 h and subsequently incubated with varying concentrations of IL-1 for 30 min (B) or with vehicle () or IL-1 (; 100 U/ml) for times ranging between 0 and 60 min (C). Radioactivity released into the supernatant was quantified as described in METHODS. Data are expressed as means ± SE of 3-4 observations per treatment from typical experiments taken from a series of 3-4 experiments for A-C (* P < 0.05, ** P < 0.01 vs. time-matched controls).

Measurement of [³H]AA release. Release of AA from endothelial cells was quantified as previously described (44). Briefly, confluent HUVEC monolayers in 24-well trays were incubated for 24 h with [5,6,8,9,11,12,14,15-³H]AA (1 µCi/ml) in serum-containing M199. After incubation, cells were washed twice in serum-free H-M199 and subsequently exposed to the same medium supplemented with 0.3% fatty acid-free BSA and other additions as indicated. At the end of the incubations the medium above the monolayers was collected and centrifuged for 5 min at 13,000 g, and the radioactivity in aliquots of supernatant was determined by beta-scintillation counting.

Immunoblotting procedures. Immunoblotting studies were performed essentially as previously described (45, 46). Briefly, confluent HUVEC in 60-mm-diameter dishes (~1 × 10⁶ cells/dish) were serum deprived for 16 h in serum- and ECGS-free M199 supplemented with glutamine (5 mM). Those cultures that exhibited significant cell detachment on serum starvation were discarded. Monolayers were subsequently washed twice in H-M199 and challenged as detailed in legends to Figs. 2, 3, and 5-8. Incubations were terminated by washing with ice-cold PBS containing Na₃VO₄ (0.4 mM) and whole cell lysates prepared in lysis buffer [63.5 mM Tris · HCl (pH 6.8), 10% glycerol, 2% SDS, 1 mM Na₃VO₄, 1 mM AEBSF, 50 µg/ml leupeptin, 5% -mercaptoethanol, and 0.02% bromphenol blue]. The protein content of cell lysates was measured using a bicinchoninic acid (BCA) protein assay (Pierce and Warriner), and equal quantities of protein (100 µg/lane) were resolved by SDS-PAGE (10%). Gels were cast using a Protean II XI (20 cm) electrophoresis system (Bio-Rad). Samples were subjected to prolonged electrophoresis overnight and then transferred onto PVDF (Immobilon-P) membrane. Membranes were blocked for 3 h in TBST [50 mM Tris, 150 mM NaCl, and 0.02% (vol/vol) Tween 20, pH 7.4] containing 3% (wt/vol) BSA. For immunodetection of cPLA₂, blots were blocked for 2 h in 0.2% I-Block. Membranes were incubated overnight in TBST/0.2% BSA containing anti-p42/44^mapk (1:25,000), anti-ACTIVE p42/p44 (1:20,000), anti-phospho-MEK1/2 (1:1,000), anti-cPLA₂ (1:5,000), anti-p38^mapk (1:1,000), or phosphospecific anti-p38^mapk (1:500) antibody. Blots were then washed in TBST (8 × 15 min) and incubated with horseradish peroxidase-conjugated goat anti-rabbit/mouse IgG as appropriate (1:10,000) for 1 h. After further washing (8 × 15 min), immunoreactive bands were visualized either colorimetrically with Renaissance 4CN Plus or by enhanced chemiluminescence (ECL) according to the manufacturer's instructions. In experiments employing phosphospecific antisera, equal loading was verified by reprobing with antibody recognizing total protein. Enhanced phosphorylation/activation of either p42^mapk or cPLA₂ is reflected in a decreased electrophoretic mobility such that the active, phosphorylated forms of the enzymes are retarded in the gel. Where indicated, densitometric analysis of cPLA₂ or p42^mapk gel shifts was performed using a Bio-Rad Gel Doc 1000 system and Molecular Analyst software.

View larger version (32K): [in this window] [in a new window]   Fig. 2.   IL-1 activates mitogen-activated protein kinase (MAPK) kinase (MEK) and p42mapk and enhances cytosolic phospholipase A2 (cPLA2) phosphorylation in a time-dependent manner. Confluent HUVEC in 60-mm-diameter tissue culture dishes were serum starved for 16 h. Washed monolayers were exposed to medium alone or to IL-1 (100 U/ml) for the times indicated, and whole cell lysates were prepared as described in METHODS. Cell lysates were subjected to SDS-PAGE followed by immunoblotting. Blots were probed with antisera recognizing either phosphorylated MEK1/2 (A), p42mapk (B), or cPLA2 (C). Immunoreactive proteins were visualized using enhanced chemiluminescence (ECL). Phosphorylated forms of the proteins are indicated (-P). The mobility shifts of phosphorylated vs. nonphosphorylated forms of p42mapk and cPLA2 are indicated. The positions of the molecular mass markers are given at left of each immunoblot. Immunoblots are from a representative experiment from a series of 4 with similar results.

View larger version (33K): [in this window] [in a new window]   Fig. 3.   MEK inhibitors attenuate IL-1-induced extracellular signal-regulated kinase (ERK) activation and cPLA2 phosphorylation. Quiescent HUVEC were pretreated for 30 min with vehicle alone or with the indicated concentrations of PD-98059 (A, top and B, top) or U-0126 (A, bottom and B, bottom) and subsequently exposed to buffer alone or to IL-1 (100 U/ml) for a further 30 min in the continued presence or absence of inhibitor. Immunoblots from whole cell lysates were probed with anti-p42mapk (A, top), anti-phospho-p42/44mapk (A, bottom), or anti-cPLA2 (B) antibodies, and immunoreactive proteins were visualized using ECL. Positions of phosphorylated and nonphosphorylated forms of p42/44mapk and cPLA2 are indicated. Blots are representative of results obtained in 4 separate experiments. C: densitometric analysis of the immunoblots shown in B. Open bars, nonphosphorylated cPLA2; hatched bars, phosphorylated cPLA2; solid bars, total (phosphorylated plus nonphosphorylated) cPLA2.

Immunoprecipitation and immune-complex kinase assay of p38^mapk. Confluent, quiescent HUVEC in 60-mm dishes were washed and treated as described for the immunoblotting studies. The medium was aspirated, and incubations were terminated by washing in ice-cold PBS supplemented with 0.4 mM Na₃VO₄; all subsequent procedures were carried out at 4°C. Monolayers were lysed in a buffer comprising 20 mM Tris · HCl, 137 mM NaCl, 2 mM EDTA, 25 mM -glycerolphosphate, 2 mM sodium pyrophosphate, 10% glycerol, 1% Triton X-100, 1 mM Na₃VO₄, 10 µg/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride (PMSF), and 40 µg/ml aprotinin (pH 7.4) and allowed to stand on ice for 30 min with gentle agitation. Whole cell lysates were centrifuged (4°C, 10 min, 13,000 g), and the supernatants were retrieved and transferred to fresh 1.5-ml microcentrifuge tubes. Equal amounts of lysate protein were subsequently incubated with anti-p38^mapk antibody (10 µl/incubation) for 3 h at 4°C with constant rotation. Immune complexes were captured with protein A/G plus agarose (4°C, overnight). The bead suspensions were pelleted, and immunoprecipitates were washed twice in lysis buffer and twice in kinase buffer (25 mM HEPES, 25 mM -glycerolphosphate, 25 mM MgCl₂, 2 mM dithiothreitol (DTT), 100 µM Na₃VO₄; pH 7.4). Immune-complex kinase activity was assayed by incubation for 30 min at 30°C in kinase buffer (25-µl final volume) containing [-³²P]ATP (5 µCi, 50 µM) and GST-ATF-2 (3 µg). Reactions were terminated by the addition of 2× concentrated Laemmli sample buffer. Proteins were resolved by SDS-PAGE (12%) and transferred to PVDF membrane, and GST-ATF-2 phosphorylation was assessed by autoradiography (70°C).

Raf-1 immunoprecipitation and kinase activity assay. Quiescent HUVEC monolayers in 60-mm dishes were washed and treated as described in the legend to Table 1. Incubations were terminated by washing in ice-cold PBS supplemented with 0.4 mM Na₃VO₄; all subsequent procedures were carried out at 4°C. Raf-1 activity was assessed using a Raf-1 immunoprecipitation kinase cascade assay kit, according to the manufacturer's instructions. Briefly, 200 µl of buffer A [50 mM Tris (pH 7.5), 1 mM EDTA, 1 mM EGTA, 0.5 mM Na₃VO₄, 0.1% -mercaptoethanol, 1% Triton X-100, 50 mM NaF, 5 mM Na pyrophosphate, 10 mM Na glycerophosphate, 0.1 mM PMSF, 1 µg/ml aprotinin, 1 µg/ml leupeptin, 1 µg/ml pepstatin, and 1 µM okadaic acid] was applied to the monolayers, which were then placed on ice for 1 h. Whole cell lysates were collected and centrifuged (15 min, 16,000 g, 4°C). The resulting supernatants were precleared with protein G agarose and incubated (2 h with rotation; 4°C) with anti-Raf-1 antibody previously conjugated to protein G agarose. The bead suspensions were pelleted, and immunoprecipitates were washed twice in buffer A and once in assay dilution buffer (ADB: 20 mM MOPS, pH 7.2, 25 mM -glycerol phosphate, 5 mM EGTA, 1 mM Na₃VO₄, 1 mM DTT). The protein G agarose/enzyme immune complexes were subsequently incubated with inactive MEK1 and inactive ERK2 for 30 min at 30°C. Samples were then briefly centrifuged to pellet the beads, and 4 µl of the supernatant were removed and mixed with myelin basic protein (final concn 0.6 mg/ml) and [-³²P]ATP (10 µCi/tube), made up to a final volume of 34 µl with ADB. Samples were incubated for 10 min at 30°C and then slowly spotted onto P81 phosphocellulose squares. Membranes were washed in 0.75% phosphoric acid (3 × 5 min) and acetone (1 × 5 min) and transferred to scintillation vials containing 5 ml of scintillation cocktail, and radioactivity was quantified on a Beckman beta counter.

View this table: [in this window] [in a new window]   Table 1.   Effects of IL-1, thrombin, and SB-203580 on Raf-1 activity

Statistical analysis. Data were analyzed for statistically significant differences between experimental conditions using ordinary ANOVA (Bonferroni multiple comparison test). Data are expressed as means ± SE; P < 0.05 was considered statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

IL-1 stimulates arachidonate release and PGI₂ synthesis in HUVEC monolayers. Exposure of confluent HUVEC monolayers to IL-1 at a maximally effective concentration (100 U/ml) resulted in significant accumulation of 6-keto-PGF₁ after a 30-min incubation, which continued to increase for the duration of the experiment (Fig. 1A). In seven independent experiments the mean increase in PGI₂ synthesis evoked by a 30-min exposure to IL-1 was 197 ± 27 (% over basal; ± SE). Parallel studies showed that neither COX-2 mRNA nor protein levels were enhanced in cells exposed to IL-1 for 30 min (data not shown); thus these early changes in prostanoid synthesis occurred in the absence of enhanced COX-2 expression. To examine the effects of IL-1 on AA mobilization, HUVEC monolayers were prelabeled with [³H]AA and exposed to IL-1 at various concentrations for different time periods. Consistent with PGI₂ synthesis, early [³H]AA mobilization (30 min) in IL-1-stimulated HUVEC was concentration dependent with maximal release achieved at 100 U/ml IL-1 (Fig. 1B) and occurred in a time-dependent manner (Fig. 1C). All subsequent experiments employed 100 U/ml IL-1. TNF- promoted similar early changes in AA release and PGI₂ synthesis in HUVEC (data not shown). In accordance with our previous findings (44), thrombin at 1 U/ml elicited both a greater and more rapid release of AA compared with a maximal dose of IL-1 (15 min thrombin: 380 ± 48; 30 min thrombin 605 ± 82% increase over basal; mean ± SE from 3 independent experiments).

Effects of a cPLA₂ inhibitor on AA release and PGI₂ synthesis. Arachidonyl trifluoromethyl ketone (AACOCF₃), a trimethyl ketone analog of AA, has been shown to inhibit cell-associated 85-kDa cPLA₂, as well as human recombinant 85-kDa cPLA₂ (1, 27). To evaluate further the potential role of cPLA₂ in cytokine-driven PGI₂ generation, we examined the effects of AACOCF₃ on [³H]AA release and PGI₂ synthesis. Pretreatment (30 min) with AACOCF₃ (10 µM) completely suppressed AA release in response to a 30-min exposure to IL-1 (IL-1, 118 ± 4%; IL-1 plus AACOCF₃, 103 ± 5% increase over basal; mean ± SE, n = 3). AA release under basal conditions was also partially, but not significantly, inhibited after AACOCF₃ treatment. In parallel experiments AACOCF₃ (10 µM) also attenuated IL-1-stimulated PGI₂ formation (IL-1, 182 ± 24%; IL-1 plus AACOCF₃, 100 ± 13% increase over basal; mean ± SE, n = 3) without significantly modifying basal PGI₂ synthesis (AACOCF₃ alone, 105 ± 20%). These results demonstrate that IL-1 promotes early AA release in a cPLA₂-dependent manner.

Phosphorylation of MEK, p42^mapk, and cPLA₂ in HUVEC exposed to IL-1. We have previously reported that inhibition of MEK, the upstream activator of p42^mapk, attenuates thrombin- or VEGF-induced PGI₂ release and that this results from inhibition of p42^mapk-mediated cPLA₂ phosphorylation (44, 45). To determine whether these mechanisms contribute to the acute release of PGI₂ from IL-1-stimulated HUVEC, we examined the effects of IL-1 on activation of MEK and p42^mapk and on the phosphorylation state of cPLA₂. Phosphorylated MEK was detected in whole cell HUVEC lysates under basal conditions and increased in a time-dependent manner in response to IL-1 with maximal phosphorylation observed after 20-30 min (Fig. 2A). Consistent with this time course, and confirming our recent observations (28, 45), IL-1 promoted a pronounced, time-dependent increase in the upper, electrophoretically slower band of p42^mapk and a concomitant decrease in the lower band (Fig. 2B). Increased phosphorylation was detected by 15-20 min, reached a maximum at 30 min, and decreased after 60 min of incubation with IL-1. As previously described (44), phosphorylated and nonphosphorylated forms of cPLA₂ are present in nonstimulated HUVEC (Fig. 2C). Enhancement of cPLA₂ phosphorylation by IL-1 followed kinetics similar to those for p42^mapk, with maximally elevated phospho-cPLA₂ evident at 20-30 min (Fig. 2C); in contrast, cPLA₂ phosphorylation was maintained for up to 2 h after IL-1 stimulation (Fig. 2C and data not shown). Thus enhanced activation of MEK, p42^mapk, and cPLA₂ coincides temporally with the onset of AA mobilization and PGI₂ synthesis in IL-1-stimulated HUVEC.

Inhibition of MEK modulates IL-1-induced p42^mapk activation, cPLA₂ phosphorylation, and PGI₂ synthesis. The close correlation between IL-1-induced p42^mapk/cPLA₂ phosphorylation and PGI₂ generation prompted us to investigate further the potential role of the MEK/ERK/cPLA₂ pathway in acute, cytokine-driven PGI₂ formation. Prior treatment of HUVEC (30 min) with the cell-permeant MEK inhibitors PD-98059 (9) or U-0126 (10) dose-dependently attenuated IL-1-induced ERK activation, with complete inhibition observed at 5 or 1 µM, respectively (Fig. 3A). Treatment with IL-1 caused a nearly complete gel shift of cPLA₂ that was partially inhibited by PD-98059 concentrations of 5 µM and above and by 0.1-3 µM U-0126 (Fig. 3B). These findings were confirmed by densitometric analysis of the immunoblots, with a maximal inhibition of ~50% with either inhibitor (Fig. 3C). In accordance with this partial inhibitory effect, IL-1-stimulated [³H]AA release was inhibited by 12 ± 2% (mean ± SE; n = 3 independent experiments) after treatment with 5 µM PD-98059. PD-98059 also caused a dose-dependent reduction of basal and IL-1-induced PGI₂ release (Fig. 4A). Interestingly, PGI₂ synthesis was totally blocked at 1 µM PD-98059, a concentration that partially inhibited IL-1-stimulated p42^mapk activation (Fig. 3A) and did not detectably reduce cPLA₂ phosphorylation (Fig. 3B). In contrast, U-0126, at a concentration (1 µM) that abolished IL-1-stimulated ERK activation and maximally affected cPLA₂ phosphorylation (Fig. 3), reduced IL-1-evoked PGI₂ synthesis by only 59 ± 5% (mean ± SE, n = 3 independent experiments). These results show that ERK-dependent and -independent mechanisms regulated cPLA₂ phosphorylation in response to IL-1 and that inhibition of cPLA₂ activation correlates with PGI₂ synthesis.

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Inhibition of PGI2 synthesis by PD-98059 and SB-203580. HUVEC in 24-well trays were incubated for 30 min with the indicated concentrations of PD-98059 (A) or SB-203580 (B) followed by treatment with vehicle alone (open bars) or with 100 U/ml IL-1 (hatched bars) for a further 30 min in the continued absence or presence of inhibitor. 6-keto-PGF1 formation was quantified by radioimmunoassay as described in METHODS. Data are means ± SE of 3 observations from single experiments representative of 4 with each inhibitor (* P < 0.05, ** P < 0.01, *** P < 0.001).

Activation of p38^mapk by IL-1 and thrombin. To address the potential role of p38^mapk in cytokine-driven prostanoid synthesis, we initially measured p38^mapk activation by immune-complex kinase assay. Treatment with IL-1 or thrombin enhanced p38^mapk activity, which was partially attenuated by the p38^mapk inhibitor SB-203580 (Fig. 5C). The incomplete inhibition of p38^mapk reflects the rapid reversibility of SB-203580:p38^mapk interaction during immunoprecipitation (43). SB-203580 also dose-dependently suppressed agonist-stimulated p38^mapk activation when added directly to immunoprecipitates (data not shown). IL-1 and thrombin enhanced phosphorylation of p38^mapk (Fig. 5, A and B), and this was unaffected by SB-203580 treatment (data not shown). The kinetics of the responses differed, with transient or sustained activation observed in response to IL-1 or thrombin, respectively. In addition, IL-1-stimulated p38^mapk activation (5 min) occurred more rapidly than p42^mapk phosphorylation (20-30 min; Fig. 2), whereas p42^mapk activation by thrombin (1 min; Refs. 44 and 45) was earlier than p38^mapk activation (Fig. 5B). Neither PD-98059 nor U-0126 affected the phosphorylation state or activity of p38^mapk as assessed by in-gel kinase assay (data not shown). Thus both IL-1 and thrombin activate p38^mapk but with different kinetics.

View larger version (27K): [in this window] [in a new window]   Fig. 5.   Effects of IL-1 and thrombin on p38mapk phosphorylation and activity. A and B: quiescent HUVEC were exposed to medium alone [control (C)], to IL-1 (100 U/ml; A) or to thrombin (1 U/ml; B) for the times indicated. Whole cell lysates were subjected to SDS-PAGE followed by immunoblotting with a phosphospecific p38mapk antibody. Immunoreactive proteins were visualized using ECL. The position of phosphorylated p38mapk (p38-P) is indicated. C: quiescent HUVEC were pretreated for 30 min with SB-203580 (30 µM) and then challenged with vehicle (C), thrombin (T; 1 U/ml, 15 min), or IL-1 (IL; 100 U/ml, 30 min). p38mapk was immunoprecipitated from whole cell lysates, and its activity was measured by assessing the phosphorylation of GST-ATF-2 as described in METHODS. Blots in A-C are each representative of 2 separate experiments with similar results.

Effects of SB-203580 on agonist-stimulated PGI₂ synthesis, p42^mapk activation, and cPLA₂ phosphorylation. In contrast to the inhibitory effects of MEK inhibition on IL-1-induced p42^mapk activation and cPLA₂ phosphorylation, blockade of p38^mapk enhanced the phosphorylation of p42^mapk (1 µM SB-203580; Fig. 6B) and cPLA₂ (Fig. 6C). Consistent with these effects SB-203580 strongly potentiated IL-1-stimulated MEK phosphorylation (Fig. 6A). However, despite potentiation of the MEK/ERK/cPLA₂ pathway, SB-203580 markedly inhibited basal and IL-1 (100 U/ml)-stimulated PGI₂ release (Fig. 4B). In contrast to our findings in IL-1-stimulated HUVEC, SB-203580 dose-dependently reduced activation of MEK, p42^mapk, and cPLA₂ in response to thrombin (Fig. 7).

View larger version (44K): [in this window] [in a new window]   Fig. 6.   SB-203580 potentiates IL-1-stimulated MEK phosphorylation, p42/44mapk activation, and cPLA2 phosphorylation. Quiescent HUVEC were pretreated for 30 min with vehicle or with the indicated concentrations of SB-203580 and then challenged with either medium alone or IL-1 (100 U/ml) for a further 30 min in the continued presence of inhibitor. Cell lysates were analyzed by immunoblotting with either anti-phospho-MEK1/2 (A), anti-phospho-p42/44mapk (B), or anti-cPLA2 (C) antibodies. Immunoreactive proteins were visualized using ECL. Positions of the phosphorylated forms are indicated (-P). Immunoblots are each representative of results obtained from 2-3 separate experiments.

View larger version (46K): [in this window] [in a new window]   Fig. 7.   Inhibition of thrombin-stimulated MEK, p42/44mapk, and cPLA2 phosphorylation by SB-203580. Quiescent HUVEC monolayers were pretreated for 30 min with vehicle or with the indicated concentrations of SB-203580 and subsequently exposed to vehicle or thrombin (1 U/ml) for a further 10 min in the continued presence of inhibitor. Whole cell lysates were separated by SDS-PAGE and analyzed by immunoblotting with anti-phospho-MEK1/2 (A), anti-phospho-p42/44mapk (B), or anti-cPLA2 (C) antibodies. Immunoreactive proteins were visualized using ECL. Positions of the phosphorylated forms are indicated (-P). Immunoblots are each representative of results obtained from 2-3 separate experiments.

Effects of IL-1 and thrombin on Raf-1 activation. To determine whether the differential effects of SB-203580 on IL-1- vs. thrombin-mediated ERK activation reflect differences in the signaling components upstream of ERK, we measured Raf-1 activation in immunoprecipitates from thrombin- and IL-1-stimulated cells (Table 1). At the time points examined, IL-1, but not thrombin, significantly elevated Raf-1 activity, suggesting that thrombin uses alternative pathway(s) to activate MEK. Because Raf-1 activity was enhanced in the presence of SB-203580 alone, SB-203580 treatment resulted in strong potentiation of Raf-1 activity in both IL-1- and thrombin-stimulated HUVEC. Thus the opposing effects of p38^mapk blockade on IL-1- (Fig. 6) vs. thrombin-induced p42/44^mapk activation (Fig. 7) cannot be explained by differential effects of the inhibitor on Raf-1 activation.

MEK inhibitors block COX activity in HUVEC. To examine whether inhibition of IL-1-stimulated PGI₂ release by SB-203580 (despite enhancement of MEK/ERK/cPLA₂ phosphorylation) reflects an independent action of the p38^mapk inhibitor on COX activity, we measured PGI₂ synthesis after application of exogenous AA (Table 2). SB-203580 attenuated AA-induced PGI₂ formation. Parallel studies showed that PD-98059 also inhibited AA-stimulated PGI₂ synthesis, whereas U-0126 (1 µM) had no significant effect (Table 2), confirming that SB-203580 and PD-98059 block COX activity at concentrations that maximally effect ERK/p38^mapk activation, whereas U-0126 does not. Because indomethacin did not modulate IL-1-induced p42/44^mapk activation or cPLA₂ phosphorylation (Fig. 8), the effects of the MEK/p38^mapk inhibitors on IL-1-stimulated ERK/cPLA₂ phosphorylation did not result from their ability to block COX activity.

View this table: [in this window] [in a new window]   Table 2.   Effects of MEK inhibitors, SB-203580, and indomethacin on 6-keto-PGF1 formation

View larger version (28K): [in this window] [in a new window]   Fig. 8.   Indomethacin does not modulate IL-1-stimulated p42/44mapk activation or cPLA2 phosphorylation. Quiescent HUVEC monolayers in 60-mm-diameter tissue-culture trays were treated for 30 min with vehicle or with the indicated concentrations of indomethacin. Cells were subsequently challenged with IL-1 for 30 min, and the resulting cell lysates were analyzed by immunoblotting with antisera recognizing the active, dually phosphorylated forms of p42mapk and p44mapk (A) or cPLA2 (B). The phosphorylated forms of p42/p44mapk and cPLA2 are indicated (-P). Results are from a single experiment representative of 3 with similar results.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The cytosolic, AA-selective 85-kDa cPLA₂ is important in mediating agonist-driven AA mobilization in several cell types (21). This release initiates the biosynthesis of PGs, leukotrienes, and platelet-activating factor, all of which are implicated in the modulation of diverse physiological and pathophysiological processes, including vascular homeostasis and inflammation. The proinflammatory cytokines IL-1 and TNF- have been shown to promote PGI₂ generation by ECs, but their effects are generally delayed in onset and require de novo protein synthesis (5, 15). In the present study we have focused on the post-receptor transduction pathways regulating early changes in cytokine-driven prostanoid synthesis and have demonstrated that IL-1 stimulates acute PGI₂ synthesis in part via a MEK/p42^mapk/cPLA₂ pathway. Thus 1) MEK, p42^mapk, and cPLA₂ were phosphorylated by IL-1 with kinetics coinciding with the onset of AA mobilization and PGI₂ synthesis, 2) AACOCF₃, at concentrations previously reported to modify stimulus-induced AA generation, markedly suppressed IL-1-driven AA release and PGI₂ synthesis, consistent with a key role for the 85-kDa cPLA₂, and 3) inhibition of MEK attenuated cytokine-stimulated p42^mapk activation, cPLA₂ phosphorylation, and PGI₂ formation. We (44, 45) and others (7) have previously shown that rapid PGI₂ synthesis in response to Ca²⁺-mobilizing agonists that promote cPLA₂ translocation also requires p42^mapk-mediated cPLA₂ phosphorylation, and several studies have suggested that phosphorylated cPLA₂ fails to mobilize AA in the absence of an increase in intracellular Ca²⁺ concentration ([Ca²⁺]_i) (e.g., Ref. 48). Because acute exposure to IL-1 neither raises [Ca²⁺]_i nor causes translocation of cPLA₂ (Ref. 11; Houliston and Wheeler-Jones, unpublished observations), these studies provide evidence that enhanced cPLA₂ phosphorylation alone may be sufficient to promote AA mobilization in IL-1-stimulated HUVEC. It is possible that cPLA₂ activation by IL-1 may result from phosphorylation of discrete pools of the enzyme that are already localized within the membrane at resting [Ca²⁺]_i (34). Although there are likely to be several alternative cell-specific mechanisms for cPLA₂ regulation (12, 30, 32, 38, 47), our findings are in keeping with recent studies suggesting that cPLA₂ phosphorylation becomes most important for activation when the rise in [Ca²⁺]_i is insufficient to fully translocate the enzyme (14). The contribution, if any, of the novel, Ca²⁺-independent cPLA₂ and - isoforms to IL-1-stimulated AA release in HUVEC remains to be defined (35, 40, 42).

Our findings implicating ERK activation in IL-1-induced AA/PGI₂ release are based on pharmacological inhibition with PD-98059 or U-0126, two structurally and mechanistically distinct inhibitors of MEK. Under conditions where p42/44^mapk activation by IL-1 was abolished, IL-1-evoked cPLA₂ phosphorylation was only partially inhibited by either of these agents, indicating that MEK-dependent and MEK-independent pathway(s) both regulate early changes in cPLA₂ phosphorylation in response to cytokine. The partial reductions in IL-1-stimulated AA mobilization and PGI₂ synthesis after MEK inhibition are also consistent with the effects on cPLA₂ phosphorylation.

In contrast to our present findings in human ECs, p42/p44^mapk does not mediate phosphorylation of cPLA₂ in thrombin- or collagen-stimulated platelets (3), and current evidence suggests that p38^mapk plays a central role in these cells (2, 14, 43). Our results suggest that p38^mapk activation is unlikely to account for the MEK-independent phosphorylation since SB-203580, a pyridinyl imidazole inhibitor of p38^mapk activity (49), did not inhibit IL-1-induced cPLA₂ phosphorylation in HUVEC. It is possible that p38^mapk directly or indirectly (14) phosphorylates cPLA₂ on Ser⁷²⁷, which does not alter the mobility of cPLA₂ on SDS-PAGE (36), or alternatively, that an SB-203580-insensitive p38^mapk isoform contributes to Ser⁵⁰⁵ phosphorylation in IL-1-stimulated endothelium (19).

Our results demonstrated that SB-203580 and PD-98059, but not U-0126, reduced PGI₂ generation in response to exogenous AA, suggesting additional inhibitory effects on COX (4, 18). Inhibition of COX activity at low concentrations may explain why PD-98059 had a more potent effect on PGI₂ synthesis compared with its effects on p42^mapk/cPLA₂ activation. However, inhibition of COX activity with indomethacin did not mimic the effects of either SB-203580 or PD-98059 on cytokine- or thrombin-stimulated p42^mapk/cPLA₂ activation, suggesting that the effects of these inhibitors did not result from COX inhibition and are mediated by blockade of p38^mapk and ERK, respectively. More importantly, U-0126, an inhibitor of MEK with no capacity to block COX in HUVEC, mimicked the partial inhibitory effects of PD-98059 on ERK/cPLA₂ phosphorylation and also partially inhibited PGI₂ synthesis. These results confirm that PGI₂ generation is regulated in part via a MEK/ERK/cPLA₂-dependent pathway.

The present studies also confirmed that SB-203580 treatment inhibited agonist-stimulated p38^mapk activity. As discussed above, blocking p38^mapk activity had no inhibitory effect on IL-1-stimulated cPLA₂ phosphorylation but, in contrast, potentiated cytokine-driven cPLA₂ phosphorylation and p42^mapk activation. These results imply that early triggering of the p38^mapk pathway by IL-1 limits the extent of subsequent activation of p42/p44^mapk and thus of the downstream events regulated by these kinases. The earlier (5 min) activation of p38^mapk in cytokine-stimulated HUVEC compared with that of p42/44^mapk (20-30 min; Ref. 28) is also consistent with this hypothesis. The modulatory effects of p38^mapk on p42/44^mapk activation are unlikely to reflect direct association of p38 with ERK (51) because 1) MEK activation in SB-203580-treated cells is modulated in parallel with ERK, and 2) immunoprecipitation of p38^mapk does not coprecipitate p42/p44^mapk (Houliston and Wheeler-Jones, unpublished observations); communication between the p38 and p42/p44^mapk pathways is therefore likely to occur at the level of upstream regulatory kinases. Furthermore, because IL-1-induced p38^mapk activation was not affected by inhibiting the MEK/ERK pathway, the interaction between ERK and p38^mapk pathways occurs in a unidirectional manner. In marked contrast to IL-1, blocking p38^mapk activation had a differential effect on thrombin-mediated signaling, causing inhibition of MEK, p42/44^mapk, and cPLA₂ phosphorylation. These data suggest that p38^mapk activation positively regulates the ERK pathway in thrombin-stimulated cells and may therefore contribute to both the early (45) and sustained phases of p42/44^mapk activation observed in response to thrombin (Houliston and Wheeler-Jones, unpublished observations). Because MEK1/2 and p42/44^mapk were modified in parallel by SB-203580 treatment, it is likely that the level of regulation by p38^mapk is proximal to MEK1/2 activation. However, measurement of Raf-1 activation by immune-complex kinase assay showed that SB-203580 in the absence of agonist strongly activated Raf-1 (17). In addition, IL-1 but not thrombin elevated Raf-1 activity in HUVEC, demonstrating that alternative Raf-1-independent pathways leading to MEK activation are recruited in thrombin-stimulated cells. The ability of SB-203580 alone to activate Raf-1 caused potentiation of activity in cells subsequently exposed to either IL-1 or thrombin, suggesting that the different effects of SB-203580 on signaling by these two agonists cannot be explained by effects at the level of Raf-1. Different types of cross talk between MAPK pathways are emerging, and the functional outcomes of such interactions are likely to be cell-type specific (20, 22, 29, 50). Our results suggest that differences in the onset and duration of p38^mapk vs. p42/44^mapk activation may determine how these pathways are coordinated to produce the physiological prostanoid responses of IL-1- or thrombin-challenged ECs. A putative signaling pathway that could account for our observations is shown in Fig. 9.

View larger version (30K): [in this window] [in a new window]   Fig. 9.   Regulation of AA release and PGI2 synthesis by MAPK pathways in thrombin- vs. IL-1-stimulated HUVEC. MEKK, MEK kinase; MEKKK; MEK kinase kinase; SB, SB-203580.

In summary, we have shown that IL-1 stimulates early cPLA₂ phosphorylation, AA mobilization, and PGI₂ synthesis in HUVEC through activation of MEK-dependent and -independent pathways. SB-203580-sensitive p38^mapk(s) negatively regulate ERK signaling after IL-1 treatment but facilitate thrombin signaling via the MEK/ERK pathway. These effects may reflect differential usage by IL-1 and thrombin of signaling elements upstream of p42/44^mapk activation. These findings underscore the importance of MAPK cascades in regulating acute events in human endothelium.