Present address: K. Eto, Laboratory of Stem Cell Therapy, The Institute of Medical Science, University of Tokyo, Tokyo 108-8639, Japan.
Alessandra Soriani, Department of Experimental Medicine and Pathology, University of Rome ‘La Sapienza’, Viale Regina Elena, 324, 00161, Rome, Italy. Tel.: 39 06 49970517; fax: 39 06 44340632; e-mail: firstname.lastname@example.org
Summary. Fibrinogen binding to platelets triggers αIIbβ3-dependent outside-in signals that promote actin rearrangements and cell spreading. Studies with chemical inhibitors or activators have implicated protein kinase C (PKC) in αIIbβ3 function. However, the role of individual PKC isoforms is poorly understood. Biochemical and genetic approaches were used to determine whether PKCθ is involved in αIIbβ3 signaling. PKCθ was constitutively associated with αIIbβ3 in human and murine platelets. Fibrinogen binding to αIIbβ3 stimulated the association of PKCθ with tyrosine kinases Btk and Syk, and tyrosine phosphorylation of PKCθ, Btk and the actin regulator, Wiskott-Aldrich syndrome protein (WASP). Mouse platelets deficient in PKCθ or Btk failed to spread on fibrinogen. Furthermore, PKCθ was required for phosphorylation of WASP-interacting protein on Ser-488, an event that has been linked to WASP activation of the Arp2/3 complex and actin polymerization in lymphocytes. Neither PKCθ nor Btk were required for agonist-induced inside-out signaling and fibrinogen binding to αIIbβ3. Thus, PKCθ is a newly identified, essential member of a dynamic outside-in signaling complex that includes Btk and that couples αIIbβ3 to the actin cytoskeleton.
Integrin αIIbβ3 binds fibrinogen or von Willebrand factor in response to platelet activation within damaged blood vessels, and it is required for normal platelet adhesion, spreading and aggregation during hemostasis . Platelet activation by agonists such as thrombin or adenosine diphosphate (ADP) is required to convert αIIbβ3 to a high-affinity receptor for these adhesive ligands, a process referred to as inside-out signaling. In turn, ligand binding to αIIbβ3 triggers outside-in signals that lead to polymerization and rearrangements of actin to insure efficient platelet aggregation and spreading . Because of its clinical importance, much attention has focused on the mechanisms of bidirectional αIIbβ3 signaling in platelets [2,3].
A general concept that has emerged about bidirectional αIIbβ3 signaling is that specific enzymes and adapter molecules can interact directly or indirectly with the cytoplasmic domains of αIIb or β3 to mediate this process. Some of these interactions are constitutive, while others are stimulated or reversed in response to fibrinogen binding to αIIbβ3 [4,5]. For example, talin binding is inducible and required for agonist-induced activation of αIIbβ3 , while the inducible binding of calcium- and integrin-binding protein (CIB), PTP-1B and Syk may promote platelet spreading [7–9]. On the other hand, the interaction of αIIbβ3 with c-Src is constitutive and it is the catalytic activity of c-Src that is regulated by fibrinogen binding .
One important family of signaling proteins that has been implicated in the regulation of integrin function and expression is protein kinase C (PKC) [1,11]. Platelets contain representatives from the conventional, novel and atypical PKC subgroups as well as the related protein kinase D [12–20]. Work with broad-spectrum PKC inhibitors has implicated PKC in both inside-out and outside-in αIIbβ3 signaling. Also, stimulation of platelets with phorbol myristate acetate (PMA), a direct activator of conventional and novel PKCs , leads to activation of αIIbβ3 and platelet spreading . However, interpretation of these results is complicated because most inhibitors and activators are not necessarily specific for PKCs or individual PKC isoforms . Consequently, it has been difficult to definitively assign functions to each isoform and to identify PKC substrates that regulate specific aspects of αIIbβ3 signaling.
Recent technical advances are facilitating the study of PKCs. These include the availability of knockout mice, RNAi-mediated gene knockdown, heterologous expression systems for wild-type and mutant PKCs, and isoform-selective peptide activators and inhibitors [23–25]. In this context, a recent study of mice deficient in PKCβ found that it is required for platelet spreading on fibrinogen, but not for agonist-induced activation of αIIbβ3 . In normal mouse and human platelets, PKCβ was recruited to αIIbβ3 in a manner dependent on fibrinogen binding and on the adapter, RACK1 [26,27]. In the present study, we have adopted a similar strategy to determine the function of PKCθ in αIIbβ3 signaling. The results establish PKCθ as a new member of an αIIbβ3-based outside-in signaling complex, which includes the Tec family tyrosine kinase, Btk, that couples αIIbβ3 ligation to dynamic changes in the platelet actin cytoskeleton.
Mice and reagents
PKCθ (PKCθ−/−)- , Btk (Btk−/y)- , and Syk (Syk−/−)-  deficient mice have been described. Antibody to PKCθ was from BD Biosciences-Transduction Laboratories (San José, CA, USA); antibody to Btk (mouse ascites) was from Upstate Biotechnology, Inc. (Lake Placid, NY, USA); and antibodies to Btk (M-138, C20), Wiskott-Aldrich syndrome protein (WASP; B9), and Syk (N19, 4D10) were from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). Antibody to c-Src (327) was from Dr Joan Brugge, Harvard Medical School (Boston, MA, USA). Antiphosphotyrosine antibodies 4G10 and PY20 were from Upstate Biotechnology, Inc. and BD Biosciences-Transduction Laboratories, respectively. An antibody specific for phosphotyrosine-90 of PKCθ was obtained from Dr Q. Ge, Cell Signaling Technology, Inc. (Danvers, MA, USA). Antibody C14 [specific for non-phosphorylated WASP-interacting protein (WIP) Ser-488] and C45 (total WIP) were from Dr Narayanaswamy Ramesh, Children's Hospital (Boston, MA, USA). Rabbit antiserum no. 8053 to β3 integrin was from Dr Mark Ginsberg, UCSD (La Jolla, CA, USA). Purified human fibrinogen was from Enzyme Research Laboratories, Inc. (South Bend, IN, USA). Fluorescein isothiocyanate (FITC)-conjugated anti-mouse IgG was from Jackson Immuno Research Laboratories, Inc. (West Grove, PA, USA). Rhodamine-phalloidin and Hoechst 33342 were from Molecular Probes (Eugene, OR, USA); bovine serum albumin (BSA), p-nitrophenylphosphate, and cytochalasin D were from Sigma-Aldrich (St Louis, MO, USA). Sodium orthovanadate and sodium fluoride were from Fisher Scientific (Fairlawn, NJ, USA). Protein A and Protein G Sepharose were from Amersham Pharmacia Biotech (Piscataway, NJ, USA); aprotinin from Roche Molecular Biochemicals (Indianapolis, IN, USA). PP2, PP3, SU6656, bisindolylmaleimide I and bisindolylmaleimide V were from Calbiochem (La Jolla, CA, USA). Mammalian expression vector pEGFP-C1 was from BD Biosciences-Clontech (Mountain View, CA, USA).
Human platelets from drug-free volunteers were drawn by venipuncture, anticoagulated with acid-citrate-dextrose, washed and resuspended to 1 × 108 mL−1 in Walsh incubation buffer containing 137 mmol L−1 NaCl, 2.7 mmol L−1 KCl, 1 mmol L−1 MgCl2, 3.3 mmol L−1 NaH2PO4·H2O, 5.6 mmol L−1 glucose, 1 mg mL−1 BSA and 20 mmol L−1 HEPES, pH 7.4 . Mouse platelets were obtained by cardiac puncture and washed and resuspended to 1 × 108 mL−1 in the incubation buffer .
Adhesion, spreading and fibrinogen binding to mouse platelets
For adhesion assays, 50 μL fibrinogen aliquots in PBS (pH 8.0) were incubated overnight at 4 °C in Immulon 2HB microtiter wells at 10 ng–2 μg per well. After washing and blocking for 2 h with 5 mg mL−1 heat-denatured BSA, 50 μL of washed platelets (3 × 107 mL−1) were plated for 1 h at room temperature. After removing non-adherent cells, wells were washed three times with Walsh buffer. Adherent platelets were quantified with an acid phosphatase assay using 5 mmol L−1p-nitrophenylphosphate as substrate and a readout at 405 nm. Blank values were obtained from wells in which no platelets had been added, and blanks were subtracted to obtain specific adhesion. Adhesion was expressed as a percent of total platelets added to the wells.
For platelet spreading assays, fibrinogen-coated coverslips were prepared (100 μg mL−1 coating concentration), placed matrix side up in wells of a 12-well plate, and 107 platelets in 0.5 mL were added for 45 min at room temperature. Adherent cells were fixed in 3.7% formaldehyde, permeabilized with 0.1% Triton X-100, and stained with primary and FITC-conjugated secondary antibodies as indicated. Rhodamine-phalloidin was used to stain F-actin. Single images were acquired with a Leica fluorescence microscope equipped with a laser scanning confocal system (MRC 1024; BioRad Laboratories, Hercules, CA, USA) [9,31]. The surface area of individual platelets was measured using Image-Pro Plus Software (Media Cybernetics, Inc., Silver Spring, MD, USA).
Binding of FITC-fibrinogen to platelets was quantified by flow cytometry . Non-specific binding was determined in the presence of 5 mmol L−1 EDTA. Specific binding was defined as total minus non-specific binding. Statistical analysis was performed using Student's t-test.
Immunoprecipitation and Western blotting analyses of human and mouse platelets
To initiate outside-in αIIbβ3 signaling, washed platelets were incubated for the indicated periods of time with 250 μg mL−1 fibrinogen and 0.5 mmol L−1 MnCl2 at room temperature [10,20]. Alternatively, platelets were allowed to adhere to fibrinogen-coated plates or maintained in suspension for 45 min . Then, platelets were lysed in buffer containing 1% NP-40, 150 mmol L−1 NaCl, 1 mmol L−1 sodium vanadate, 0.5 mmol L−1 sodium fluoride, 1 mmol L−1 leupeptin, complete protease inhibitor cocktail (Roche Applied Science, Indianapolis, IN, USA) and 50 mmol L−1 Tris (pH 7.4). Lysates were clarified by centrifugation at 13 000 g for 10 min at 4 °C and 500–800 μg of protein from the soluble fraction were immunoprecipitated and processed for Western blot analysis [10,20].
Results and Discussion
PKCθ is constitutively associated with αIIbβ3 and involved in outside-in αIIbβ3signaling
PKCβ associates with αIIbβ3 in fibrinogen-adherent platelets through the adapter molecule RACK1, and it is required for cytoskeletal reorganization and full platelet spreading . As PKCθ may be involved in cytoskeletal remodeling in activated T lymphocytes  and in signaling downstream of non-integrin adhesion receptors, glycoprotein (GP) Ib-IX-V and GP VI, in platelets , we asked if there was a relationship between PKCθ and αIIbβ3. Human platelets were maintained in suspension or plated on fibrinogen for 40 min. Cells were lysed in a buffer containing NP-40 detergent and αIIbβ3 was immunoprecipitated with a rabbit anti-β3 antiserum. Equal amounts of PKCθ were recovered from the αIIbβ3 immunoprecipitates of non-adherent and adherent platelets (Fig. 1A). Identical results were obtained with platelets in suspension to which soluble fibrinogen had bound in response to 0.5 mmol L−1 MnCl2 (data not shown). MnCl2 was utilized in this and subsequent experiments because it induces fibrinogen binding to αIIbβ3 directly without causing global platelet activation through other receptors . In contrast to the results with PKCθ, the association of protein kinase Syk with αIIbβ3 increased upon platelet adhesion to fibrinogen, as reported previously  (Fig. 1A). Thus a pool of PKCθ, like c-Src , is constitutively associated with αIIbβ3 in platelets. However, these results do not establish whether the interaction of PKCθ with αIIbβ3 is direct, as in the cases of c-Src and Syk [10,36], or indirect, as in the case of PKCβ .
To evaluate possible functions of PKCθ in αIIbβ3 signaling, platelets from PKCθ-deficient mice (PKCθ−/−) were compared with those from PKCθ+/+ littermates. Note that these PKCθ−/− mice exhibited no obvious bleeding defects (data not shown). Soluble fibrinogen binding to platelets requires agonist-induced inside-out activation of αIIbβ3, whereas immobilized fibrinogen can support the initial adhesion of unstimulated platelets . PKCθ−/− platelets stimulated with ADP, epinephrine or PAR4 receptor-activating peptide (AYPGKF, single letter amino acid code)  bound soluble fibrinogen to the same extent as did PKCθ+/+ platelets, whether these agonists were added individually or in combination (Fig. 1B). Surface expression of αIIbβ3 was normal in PKCθ−/− platelets, as determined by flow cytometry (data not shown). Thus, PKCθ is not required for the expression or activation of αIIbβ3. However, platelet responses that are dependent on outside-in αIIbβ3 signaling-filopodial extension and spreading on fibrinogen, were reduced in PKCθ−/− platelets, and the reduction in spreading was statistically significant (P < 0.05) [Fig. 1C(a),C(b),D]. The spreading defect exhibited by PKCθ−/− platelets could be rescued by co-stimulation of fibrinogen-adherent cells with AYPGKF, ADP or PMA (Fig. 1C,D), the latter a direct activator of conventional and novel PKCs and other proteins containing typical C1 domains [21,22]. The adhesion of PKCθ−/− platelets was modestly reduced at high concentrations of immobilized fibrinogen (P < 0.05) (Fig. 1E), possibly secondary to reduced integrin-matrix contacts engendered by the defective platelet spreading.
Several conclusions can be drawn from these results. First, there is a selective requirement for PKCθ in outside-in αIIbβ3 signaling pathways that regulate platelet spreading. Our preliminary studies using a CHO cell model system and constitutively active or kinase-inactive PKCθ mutants suggest that the catalytic activity of PKCθ is required for αIIbβ3-dependent cell spreading on fibrinogen. Second, the requirement for PKCθ can be eliminated by co-stimulation of platelets with traditional G protein-coupled agonists, perhaps in part because they activate PKC isoforms other than PKCθ [12–19]. Third, although work with pharmacological inhibitors has implicated PKC in inside-out αIIbβ3 signaling , PKCθ is not essential for this process. Interestingly, neither is PKCβ , consistent with the suggestion that PKCα is required for inside-out αIIbβ3 signaling . A patient with an αIIbβ3 activation defect and a partial deficiency of PKCθ has been described [39,40]. As the primary defect appears to be a mutation in the CBFA2 transcription factor, expression of many proteins may be abnormal in this individual's platelets, making it difficult to ascribe the αIIbβ3 activation defect solely to PKCθ deficiency. Fourth, deficiency of either PKCθ or PKCβ  leads to a platelet spreading defect and one isoform cannot substitute for the other in support of spreading, even though both isoforms associate with αIIbβ3. Therefore, PKCθ and PKCβ may operate in parallel pathways that link αIIbβ3 to actin rearrangements, and their relevant substrates may differ.
PKCθ is tyrosine-phosphorylated during outside-in αIIbβ3 signaling
Fibrinogen binding to platelets induces tyrosine phosphorylation of multiple proteins [2,5]. To investigate whether PKCθ is one of them, platelets were incubated for 15 min in the presence or absence of fibrinogen and MnCl2, and PKCθ immunoprecipitates were probed on Western blots with antibodies to phosphotyrosine. Tyrosine phosphorylation of PKCθ increased in a fibrinogen-dependent manner (Fig. 2A). When T lymphocytes are activated through the T-cell receptor/CD3 complex, PKCθ becomes phosphorylated at Tyr-90, a modification that appears to influence cell proliferation . Using a phospho-specific antibody to Tyr-90, we found that fibrinogen binding to platelet αIIbβ3 did not induce phosphorylation of this tyrosine residue; however, stimulation of platelets with collagen did (Fig. 2B), possibly through collagen interaction with GP VI [18,42,43]. Thus, fibrinogen binding to αIIbβ3 leads to tyrosine phosphorylation of PKCθ, but not to phosphorylation of Tyr-90. Further work will be needed to identify the sites of phosphorylation in PKCθ and their role in outside-in αIIbβ3 signaling.
PKCθ interacts with Btk and Syk protein tyrosine kinases during outside-in αIIbβ3 signaling
In light of the interactions that occur between PKCθ and the Tec family protein tyrosine kinase, Btk, when platelets are stimulated through GP Ib-IX-V and GP VI , we asked whether a similar interaction is promoted by αIIbβ3. Fibrinogen binding to platelets in response to MnCl2 caused increased association of PKCθ with Btk, as determined by co-immunoprecipitation (Fig. 2C). Furthermore, fibrinogen binding led to an increased association of PKCθ (Fig. 2D) or Btk (Fig. 2E) with Syk. Altogether, these results are consistent with the notion that αIIbβ3 functions as a molecular scaffold upon which specific signaling proteins mediate outside-in signaling [2,4]. Some of these proteins are constitutively associated with the integrin (e.g. PKCθ (Fig. 1A), c-Src  and RACK1 ), others are recruited in response to fibrinogen binding (e.g. Syk, PTP-1B, PKCβ, integrin-linked kinase and CIB) [7–9,20,44,45], and still others dissociate from αIIbβ3 upon fibrinogen binding (e.g. Csk, PP1c) [9,46].
As gene knockout studies have shown that PKCθ (Fig. 1C,D), c-Src and Syk  are all required for platelet spreading on fibrinogen, and as Btk may be drawn into a signaling complex with PKCθ and Syk in response to fibrinogen binding (Fig. 2C–E), the potential role of Btk was studied in more detail. Binding of soluble fibrinogen to human platelets induced tyrosine phosphorylation of Btk (Fig. 3A), a response similar to that observed in fibrinogen-adherent platelets . Phosphorylation was dependent on an Src kinase because preincubation of platelets with an Src inhibitor (SU6656 or PP2) eliminated or reduced this response, while an inactive congener (PP3) had no effect (Fig. 3A). However, αIIbβ3-dependent tyrosine phosphorylation of Btk occurred normally in the presence of bisindolylmaleimide I, a broad-spectrum PKC inhibitor (Fig. 3A), and in a single study fibrinogen-dependent Btk phosphorylation was normal in PKCθ−/− platelets (data not shown). These results indicate that although PKCθ associates with Btk, it may not be required for Btk phosphorylation during outside-in αIIbβ3 signaling. Tyrosine phosphorylation of Btk was also observed upon fibrinogen binding to normal mouse platelets, to platelets deficient in Syk (Fig. 3B), and to platelets pretreated with 10 μmol L−1 cytochalasin D to inhibit actin polymerization (Fig. 3C). Thus, PKCθ may help to recruit Btk to an αIIbβ3 signaling complex, where it becomes phosphorylated by Src in a manner independent of Syk or actin polymerization.
To determine whether Btk plays a role in signaling from αIIbβ3 to the actin cytoskeleton, Btk-deficient mouse platelets were plated on fibrinogen. As Btk is an X-linked gene, male mice were used to generate the knockout, which were designated Btk−/y . Compared with control Btk+/− platelets, Btk−/y platelets spread poorly on fibrinogen (Fig. 4A), and this difference was significant (P < 0.05) (Fig. 4B). However, similar to PKCθ−/− platelets (Fig. 1C,D), Btk-deficient platelets spread normally when co-stimulated with AYPGKF, ADP or PMA (Fig. 4A,B), and they bound soluble fibrinogen normally in response to agonists (Fig. 4C). This result appears at variance with the reported normal spreading of Btk-deficient mouse platelets on fibrinogen . We do not know the reason for this difference, although our conclusion was based on computer-based image analysis of large numbers of platelets. In addition, we studied washed platelets instead of platelet-rich plasma ; consequently, there might have been differences between the two platelet preparations in the amount of residual ADP or other factors that could facilitate platelet spreading .
When platelets are stimulated through GP Ib-IX-V and GP VI, formation of a complex between Btk and PKCθ may facilitate tyrosine phosphorylation of the latter . We found that Btk was neither required for the constitutive association of PKCθ with αIIbβ3 (Fig. 5A) nor for the fibrinogen-dependent tyrosine phosphorylation of PKCθ (Fig. 5B). As Btk−/y platelets express Tec , and as Tec family kinases are involved in polarization of integrins and PKCθ during T-cell activation , a requirement for Tec in αIIbβ3-dependent tyrosine phosphorylation of PKCθ cannot be excluded.
Potential mechanisms of αIIbβ3-dependent actin reorganization
To understand better how PKCθ might relay αIIbβ3 signals to the actin cytoskeleton, we focused on two proteins, WASP and WIP, which appear to participate in the regulation of the F-actin nucleation and branching activity of the Arp2/3 complex in platelets [49–53]. In activated T lymphocytes, PKCθ participates in the phosphorylation of WIP at Ser-488 . This uncouples WIP from WASP, enabling the latter to be activated by Cdc42 and in turn to activate the Arp2/3 complex . Also, WIP interacts with cortactin and actin monomers to promote actin polymerization through the Arp2/3 complex . Therefore, we asked if phosphorylation of WIP Ser-488 is regulated by αIIbβ3. Western blots of platelet lysates were probed with an antiserum (C14) specific for WIP that is not phosphorylated on Ser-488 . Fibrinogen binding to human platelets caused a partial and transient decrease in C14 immuno-reactivity, indicative of a transient increase in WIP Ser-488 phosphorylation (Fig. 6A). Similar results were obtained with PKCθ+/+ murine platelets. In contrast, fibrinogen binding to PKCθ−/− platelets failed to induce phosphorylation of WIP Ser-488 (Fig. 6B). Thus, PKCθ needs to be present for αIIbβ3-dependent serine phosphorylation of WIP, providing a potential explanation for the defective spreading of PKCθ−/− platelets on fibrinogen.
In activated T lymphocytes, WASP is phosphorylated on Tyr-291 in an Src-dependent fashion, and this event may collaborate with Cdc42 to activate WASP [56,57]. WASP is phosphorylated on tyrosine by a collagen-related peptide that binds to the collagen receptor GP VI, but not to the integrin α2β1 . Similarly, we found that human platelets incubated with MnCl2 and fibrinogen exhibited tyrosine phosphorylation of WASP (Fig. 6C). WASP phosphorylation was blocked by 2 mmol L−1 RGDS, which inhibits fibrinogen binding (data not shown) and by 5 μmol L−1 PP2, but not PP3 (Fig. 6D). Phosphorylation was reduced, but not eliminated, in Syk−/− mouse platelets (Fig. 6E). However, αIIbβ3-dependent tyrosine phosphorylation of WASP was normal in PKCθ−/− or Btk−/y mouse platelets (data not shown). As human and mouse platelets deficient in WASP spread normally [59,60], and related Scar/Wave actin-regulatory proteins are present in platelets [60,61], the precise function of WASP downstream of αIIbβ3 requires further study.
Altogether, these results indicate that PKCθ and Btk collaborate with other protein kinases to transmit signals from αIIbβ3 to the platelet actin cytoskeleton. Furthermore, they highlight the fact that physical and functional interactions of PKC with integrins are likely to be isoform-, integrin- and cell-type specific . Consequently, advances in understanding these relationships will require going beyond the traditional use of chemical PKC activators and inhibitors, the selectivity of which is usually insufficient to draw unambiguous conclusions in cells like platelets that express multiple PKCs and integrins. Even the use of gene-targeted mice, as in the present study, and the investigation of patients with inherited bleeding disorders associated with a deficiency of PKCθ , raise new questions concerning the identities and functions of PKCθ substrates that regulate platelet adhesive responses.
We thank Joan Brugge (Harvard Medical School, Boston, MA), Mark Ginsberg (UCSD, San Diego, CA) and Narayanaswamy Ramesh (Childrens Hospital, Boston, MA) for reagents. These studies were supported by grants from the National Institutes of Health.