Regulation of c-Jun-NH2 Terminal Kinase and Extracellular-Signal Regulated Kinase in Human Platelets

MITOGEN-ACTIVATED protein kinases (MAP kinases) are a family of serine-threonine kinases1activated by many stimuli, including growth factors and hormones, in proliferative cells.2 This family consists of three subgroups. The extracellular signal-regulated kinases (ERKs) are involved in proliferation, adhesion, and cell progression.3,4 The p38MAP kinase and the c-Jun amino-terminal kinases (JNKs) or stress-activated protein kinases (SAPKs), which include the 46-kD JNK1 and 55-kD JNK2 isoforms, seem to be involved in apoptosis.5,6 Activation of MAP kinases requires dual phosphorylation at threonine and tyrosine residues.7 Moreover, different kinases activate ERKs, JNKs, and p38MAP kinase. ERKs, JNKs, and p38MAP kinase are phosphorylated by MAP kinase/ERK kinase (MEK)1/2, MEK4, and MEK3/6, respectively.8-10 The JNKs cascade is activated by various stresses11 (UV irradiation, heat shock, x-ray irradiation, osmotic shock), inflammatory cytokines12 (tumor necrosis factor [TNF]-α and interleukin-1), and weakly by growth factors. Components of the JNK pathways are dependent on MAP kinase kinase (MKKs), a serine/threonine kinase, p21-activated kinase (PAK), and Ras-like guanosine triphosphate (GTP)–binding proteins cdc42 and Rac.13-15 

Platelets are nonproliferative cells and are a good model for studying the signal transduction of MAP kinases and new functions. In platelets, two MAP kinases, ERK and p38MAP kinase, have been identified.16,17 The ERK cascade is activated by thrombin, or collagen, and seems to involve MEK1/2 and protein kinase C (PKC).18,19 Recently, we found that α_IIbβ₃ downregulates the ERK2 activation induced by thrombin if engaged in fibrinogen-mediated platelet aggregation.20 This was the first report of an integrin downregulating ERK2 activation during cell adhesion. p38MAP kinase is activated by thrombin and seems to be involved in the phosphorylation of PLA₂.21 The last subgroup of MAP kinases, JNKs, has not been identified in platelets.

We determined the conditions of JNK activation during thrombin-induced platelet aggregation. We also assessed the involvement of α_IIbβ₃ integrin engagement in JNK activation and compared it with that in ERK activation. We found that JNK1 was (1) present and activated after thrombin induction; (2) regulated negatively by α_IIbβ₃ engagement; and (3) regulated positively by stirring, like ERK2 activation. This is the first report showing that JNK1 is present and active in platelets and has a similar regulation to ERK2. To our knowledge, this is also the first report suggesting a direct mechanical induction of MAP kinases in platelets.

Bovine thrombin, synthetic peptides Arg-Gly-Asp-Ser (RGDS), Arg-Gly-Glu-Ser (RGES), leupeptin, and aprotinin were purchased from Sigma (St Louis, MO). Protein-G PLUS-Agarose and the rabbit polyclonal antibody directed against JNK1 (C-17) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA), and rabbit polyclonal antibody directed against the phosphorylated forms of JNKs and ERKs was obtained from Promega (Madison, WI). Donkey antirabbit horseradish peroxidase–conjugated IgG was obtained from Jackson Immuno Research (West Grove, PA) and [γ-³²P]adenosine triphosphate (ATP; 167 tera becquerel [Tbq]/mmol) was obtained from ICN (Irvine, CA). Electrophoresis reagents were obtained from Bio-Rad (Richmond, CA) and Euromedex (Souffelweyersheim, France), and chemical products were obtained from Merck (Darmstadt, Germany) and Prolabo (Paris, France). Nitrocellulose sheets were purchased from Schleicher & Schuell (Dassel, Germany). (Fab′)₂ fragments of the mouse monoclonal IgG AP-2 specific for α_IIbβ₃ were an extremely generous gift from Dr T.J. Kunicki (Scripps Clinic, La Jolla, CA).22 

Normal donors were not treated with any drug during the 2 weeks before blood donation given with informed consent. A patient with Glanzmann’s thrombasthenia is characterized by absence of platelet aggregation and fibrinogen binding, as previously described.23 

Venous blood was obtained from healthy donors with their informed consent. Platelets were isolated and washed by differential centrifugation in citrate buffer, pH 6, containing 10⁻⁴ mmol/L PGE₁, 140 mmol/L NaCl, 5 mmol/L KCl, 12 mmol/L trisodium citrate, 10 mmol/L glucose, and 12.5 mmol/L sucrose, and then in the same buffer but without PGE₁. The platelet pellet was resuspended in 10 mmol/L HEPES pH 7.4, 140 mmol/L NaCl, 3 mmol/L KCl, 0.5 mmol/L MgCl₂, 5 mmol/L NaHCO₃, and 10 mmol/L glucose. Cell concentration was adjusted to 5 × 10⁸/mL. Platelets (0.4 mL) were preincubated, without stirring, at 37°C for various times, with various agonists. Platelet aggregation was then initiated by adding bovine thrombin with or without constant stirring (1,200 rpm) in an aggregometer cuvette (Chronolog dual beam aggregometer). Aggregation was measured and is expressed as a percentage change in the transmission of light, with the blank sample (buffer without platelets) defined as 100%.

Samples were subjected to immunoblotting as described previously.24 Briefly, platelet aggregation was stopped by addition of 40 μL of denaturing buffer (10% [wt/vol] sodium dodecyl sulfate [SDS], 100 mmol/L NaCl, 50 mmol/L Tris, 50 mmol/L NaF, 5 mmol/L EDTA, 40 mmol/L β-glycerophosphate, 200 mmol/L sodium orthovanadate, 5 μg/mL leupeptin, 10 μg/mL aprotinin, pH 7.4), and the samples were heated at 95°C for 5 minutes. Proteins were separated by SDS–polyacrylamide gel electrophoresis (SDS-PAGE), in a 12% polyacrylamide gel. Proteins were transferred to nitrocellulose filters by semidry transfer (Enprotech, Natick, MA). Filters were then incubated for 1 hour at room temperature with the polyclonal primary antibody anti-JNK1-P (1:5,000) or anti-JNKs (1:2,000) or anti-ERK2-P (1:20,000) or anti-ERKs (1:20,000). The membranes were washed five times with Tris saline buffer without milk and were then incubated with horseradish peroxidase–conjugated goat antirabbit IgG (1:20,000) for 45 minutes at room temperature. Immunoreactive bands were detected by chemiluminescence using the Amersham ECL enhanced chemiluminescence system (Amersham, Arlington Heights, IL).

Platelet lysates were diluted in 50 mmol/L HEPES pH 7.4, 150 mmol/L NaCl, 10% (vol/vol) glycerol, 1% (vol/vol) Triton X-100 (Sigma), 1.5 mmol/L MgCl₂, 10 μg/mL of leupeptin, 0.5 mmol/L sodium vanadate, and 2 mmol/L EGTA, and were incubated overnight at 4°C with anti-JNK1 antibody (2 μg/sample). They were then incubated with protein-G PLUS agarose (40 μL, vol/vol) for 1 hour ar 4°C. Immune complexes were collected by centrifugation. The sample was then resuspended in 20 μL of kinase buffer (20 mmol/L HEPES, pH 7.4, 10 mmol/L MgCl₂, 10 mmol/L p-nitrophenyl phosphate, 1 mmol/L dithiothreitol) containing GST-Jun (2 μg), 50 μmol/L unlabeled ATP, and 5 μCi of [γ-³²P]ATP per sample.24After 10 minutes at 37°C, samples were subjected to SDS-PAGE, and the gel was stained with Coomassie Blue (Sigma), dried, and autoradiographed.

We first determined the conditions for thrombin-induced JNK1 activation by incubating platelets with various concentrations of thrombin (0 to 1 U/mL) for 2 minutes under stirring conditions (Fig 1A). Phosphorylation was investigated by Western blotting using an antibody that recognized phosphorylated JNK1 and JNK2 (JNK1-P and JNK2-P). In resting platelets and after induction with thrombin (0.01 U/mL), no phosphorylated JNK1 and/or JNK2 were found. JNK1-P was detected only after treatment with 0.02 U/mL thrombin, reaching a maximum intensity at 0.2 U/mL. Moreover, we used an antibody that recognized phosphorylated and nonphosphorylated JNKs as internal control to ensure equal JNKs loading. We assessed JNK1 kinase activity in the same thrombin conditions, after JNK1 immunoprecipitation, by in vitro phosphorylation of GST-cJUN (see Materials and Methods). No GST-cJUN phosphorylation was observed in resting platelets. In contrast, a significant increase in JNK1 activity was detectable with 0.01 U/mL thrombin and reached a maximum at 0.05 U/mL thrombin. The fact that JNK1 activity appears at lower concentrations of thrombin (0.01 U/mL) relative to its state of phosphorylation is probably a result of the difference in the sensitivity of the methods used, and/or in the antibodies used for phosphorylation and kinase activity. Thus, JNK1 activation occurred during thrombin-induced platelet aggregation. No JNK2 activation was detected in the same conditions.

We used phosphorylation and kinase assays to investigate the time course of JNK1 activation upon treatment with 1 U/mL thrombin under stirring conditions (Fig 1B). In resting platelets, no JNK1 phosphorylation or activity was detected. In contrast, JNK1 phosphorylation and activity that was detected after 30 seconds or 1 minute, depending on the experiment, reached a maximum between 120 and 180 seconds. In these conditions of JNK1 phosphorylation, immunoblots were probed with an antibody recognizing phosphorylated and nonphosphorylated JNKs and confirmed equal levels of total JNKs in the different platelet lysates. JNK1 was activated during platelet aggregation, raising questions about the respective involvement of α_IIbβ₃ integrin engagement or thrombin activation.

We investigated thrombin-induced JNK1 activation in conditions of inhibition of fibrinogen binding to integrin α_IIbβ₃, and consequently, inhibition of aggregation. The involvement of α_IIbβ₃ in JNK1 activation was assessed using the RGDS peptide, a competitive inhibitor of fibrinogen binding to α_IIbβ₃. In resting platelets, preincubation with RGDS alone, in the absence of thrombin, did not activate JNK1 (Fig 2A and B). In thrombin-activated platelets, however, preincubation with RGDS for 30 seconds (0.5 and 1 mmol/L), which completely inhibited platelet aggregation, did not prevent, but rather enhanced JNK1 phosphorylation and activity after 2 minutes of thrombin activation (Fig 2A). Quantification by densitometry demonstrated 471.2% ± 91.0% more JNK1 phosphorylation in the presence of 1 mmol/L RGDS. A control RGES peptide, which did not block platelet aggregation, did not significantly affect (179.8% ± 37.0%) the thrombin-induced phosphorylation of JNK1. We confirmed these results, by blocking fibrinogen binding using (Fab′)₂ fragments of the anti-α_IIbβ₃ monoclonal antibody AP-2 preincubated for 5 minutes before thrombin addition. Thrombin-induced JNK1 phosphorylation and activity were increased dose-dependently by addition of AP-2 (Fab′)₂, reaching a plateau at 10 μg/mL (Fig 2B). At this concentration (10 μg/mL), the values obtained were similar to those for inhibition by RGDS: 409.6% ± 53.3% JNK1 phosphorylation and 644.2% ± 111.0% JNK1 activity. Thus, the results of both experiments (RGDS and [Fab′]₂ fragments) suggest that activation of the JNK1 pathway (1) occurs during thrombin aggregation independently of α_IIbβ₃ engagement, and moreover, (2) is downregulated by fibrinogen binding to α_IIbβ₃. This negative control of JNK1 activation is possibly because of the presence and/or activation of phosphatases in the same compartment of JNK1 in conditions of platelet aggregation. Different phosphatases are associated with cytoskeleton upon thrombin stimulation. PP2A, a serine/threonine phosphatase, which can dephosphorylate MAP kinases in proliferative cells, is present in the platelet cytoskeleton after thrombin induction.25Moreover, tyrosine phosphatases like SHP1 are also present in the cytoskeleton after α_IIbβ₃engagement.26 The other possibility is that the kinases involved in the activation of JNK1 are partitioned in different compartments after thrombin induction in the presence or absence of α_IIbβ₃engagement.

We have previously shown that α_IIbβ₃engagement negatively regulates ERK2 activation.20 So, we compared thrombin-induced JNK and ERK activation in a patient with Glanzmann’s thrombasthenia, a disease that is characterized by a quantitative defect in platelet α_IIbβ₃. In these conditions, JNK1 activation in stirred platelets was higher in the patient than in the control (Fig 3). The level of upregulation in the patient and control with RGDS was similar, strongly suggesting that this upregulation was caused by the absence of α_IIbβ₃ engagement rather than by any effect of RGDS itself. To ensure that the differences in the levels of JNK-P between control and a patient with Glanzmann’s thrombasthenia are not caused by differences in JNK protein levels, immunoblots were probed with an antibody directed against phosphorylated and nonphosphorylated JNKs. In these conditions, no difference in JNK protein levels was observed. Finally, we also demonstrated ERK2 overactivation, as previously described.20 

Thus, JNK1 maximal activation is obtained during thrombin activation of platelets in conditions in which α_IIbβ₃, and thus, aggregation are blocked (RGDS and monoclonal anti-α_IIbβ₃) or absent (thrombasthenia). However, all conditions implied that platelets were stirred. To test whether stirring per se might affect JNK1 activation, we compared thrombin-induced JNK1 activation in the absence or presence of stirring with or without the RGDS peptide. In stirred platelets, in the presence of the RGDS peptide, JNK1 was overactivated (566% ± 29% v100% in the absence of RGDS) 2 minutes after the addition of thrombin (Fig 4A). Interestingly, the absence of stirring, whether in the presence or absence of RGDS, led only to a partial overactivation (278% ± 69%) of JNK1. This is a strong indication that stirring per se, or a factor depending on it, may represent up to 50% of the JNK1 upregulation. The fact that stirring yields maximal JNK1 activation in the presence versus the absence of RGDS suggests that the remaining 50% upregulation is caused by the blocking of α_IIbβ₃ engagement, per se. To test whether positive regulation of JNK1 by stirring was specific of this MAP kinase, we next studied ERK2 activation by Western blotting using an anti-ERK-P antibody, in the same conditions of JNK1 activation (Fig 4A). We showed, as previously described,20 that ERK2 was overactivated in stirred platelets with the RGDS peptide. This ERK2 overactivation that reached 584% ± 21% after 2 minutes of thrombin induction (0.2 U/mL) was greater than that reported in a previous study,20 probably because of the high sensitivity of the anti-ERK-P antibody used in our system. Stirred (584% ± 21%) and unstirred (365% ± 70%) platelets compared in the presence of RGDS after 2 minutes of thrombin induction suggested that ERK2 activation was also partly regulated by a factor dependent on stirring, but this positive regulation may be weaker for ERK2 (219%) than for JNK1 (288%). Thus, (1) ERK2 and JNK1 activation induced by thrombin were both negatively regulated by α_IIbβ₃ engagement in aggregation itself, and (2) stirring, per se, or a factor dependent on stirring is involved, as a positive regulator in JNK1 activation and to a lesser extent in ERK2 activation.

To confirm the positive regulatory effect of stirring on MAP kinases, we compared ERK2 and JNK1 activation in stirred and unstirred platelets from a thrombasthenic patient lacking platelet α_IIbβ₃ (Fig 4B). In the conditions of identical levels of total JNKs proteins loaded, the activation of JNK1 and ERK2 in stirred platelets was higher than that in unstirred platelets, confirming that stirring was a positive regulator of JNK1 and ERK2 activation. We also confirmed that positive regulation by stirring was stronger for JNK1 than for ERK2 as judged by a lower stirred and unstirred ratio (x3) for ERK2 and x5 for JNK1 (Fig4B). Altogether, our data suggest that JNK1 and ERK2, though to a lesser extent, are activated by thrombin activation and stirring, and downregulated by α_IIbβ₃ integrin engagement. It is tempting to correlate these results with data showing that the fluid shearing of endothelial cells increases the activation of JNK much more than that of ERK.27 We investigated the involvement of stirring itself or of a factor released from platelets dependent on stirring in the positive regulation by quantifying serotonin release in condition of stirred and unstirred platelets in the presence or absence of the RGDS peptide. No differences in serotonin release depending on the conditions used were found (results not shown). The thromboxane analogue, U46619, activated ERK2 and JNK1 (results not shown). We therefore tested the effect of thrombin-induced thromboxane A2 formation as a secondary activator of MAP kinase. In our conditions, flurbiprofen, which blocks cyclooxygenase 1, had no effect on thrombin-induced JNK1 and ERK2 activation, suggesting that thromboxane A2 formation was not involved (results not shown). Thus, though our data cannot totally exclude the role of secreted factors, they are consistent with stirring itself being involved, in the upregulation of JNK1 and ERK2 activation.

The role of MAP kinases in platelet physiology is unclear. However, recent work28 has suggested that the Ras-initiated MAP kinase pathway suppresses integrin activation. Although these conclusions stem from work in Chinese hamster ovary cell line, a model system may be very different from platelets. It is tempting to speculate that in the physiopathologic conditions of shear stress and of platelet activation, MAP kinases (JNK1 and ERK2) may serve as a suppressor of α_IIbβ₃ activation and may act as a negative regulator of platelet activation. The partial downregulation of JNK1 and ERK2 would thus be part of a feedback loop of integrin engagement on MAP kinases activation.

We found that JNK1 and ERK2 were similarly regulated, but it is not clear whether both activation pathways involved a common determinant in platelets, and at which level of the signal transduction pathway the MAP kinases were affected by the engagement of α_IIbβ₃ integrin. It is well-known that Ras cascade activation is involved in the activation of ERKs, but the role of Ras in JNKs activation has only been recently shown.27,29,30 Thus, thrombin-induced JNK and ERK activation may involve a common pathway, and their negative regulation may occur upstream of the JNKs and ERKs. In conclusion, this is the first report showing (1) that JNK1 is present and can be activated in platelets, and (2) that mechanical forces induce ERK1 and JNK2 activation in platelets.

The authors thank Dr T.J. Kunicki for generously donating (Fab′)₂ fragments from the mouse monoclonal IgG AP-2 antibody specific for α_IIbβ₃. The authors also thank E. Savariau and R. Nancel for graphic work.