Lyn and PECAM-1 function as interdependent inhibitors of platelet aggregation

Rapid formation of platelet thrombi at sites of vessel injury is important to prevent blood loss, but thrombus formation must be regulated to prevent vessel ischemia and occlusion. Thrombus formation depends on platelet adhesion to, and activation on, exposed subendothelial collagen. Platelets adhere to collagen via the integrin α₂β₁, and activation depends on signals transduced by the platelet-specific glycoprotein VI (GPVI) and Fc receptor γ-chain (FcRγ) collagen receptor complex (GPVI/FcRγ) as previously reviewed.^1-3  Signal transduction by GPVI is initiated by Src family kinase (SFK)–mediated phosphorylation of 2 tyrosine residues within the FcRγ immunoreceptor tyrosine-based activation motif (ITAM), which supports binding and activation of the tyrosine kinase p72^syk, assembly of a signaling complex that activates phospholipase C γ2, mobilization of calcium and activation of protein kinase C to trigger granule release, and integrin activation to enable platelet aggregation as reviewed previously.^1,2 

Among the SFKs expressed by platelets, Fyn and Lyn physically associate with, and participate in, signal transduction by the GPVI/FcRγ complex.^4-6  Whereas Fyn plays only a stimulatory role, Lyn has been reported to initially stimulate and subsequently inhibit platelet responses to GPVI-specific agonists.⁷  The stimulatory function of Lyn has recently been found to rely on its constitutive interaction with a unique proline-rich domain within the GPVI cytoplasmic tail that activates the enzyme,⁸  and to be important for activating a phosphatidylinositol 3 kinase–Akt pathway that induces secretion of platelet granule contents, which amplify platelet activation.⁹  The mechanism underlying Lyn-mediated inhibition of GPVI-specific responses is, however, not yet defined. In hematopoietic cells other than platelets, Lyn has been shown to inhibit the activity of ITAM-coupled receptors by phosphorylating immunotyrosine-based inhibitory motif (ITIM)–containing receptors that bind tyrosine or inositol phosphatases, thereby dampening cellular activation.^10,11  PECAM-1 is a well-characterized ITIM-containing receptor that inhibits platelet responses to stimulation via GPVI.^12-14  In the present study, we demonstrate that Lyn inhibitory function in platelets depends on its ability to phosphorylate PECAM-1 ITIMs, which enables recruitment of the Src homology 2 (SH2) domain–containing phosphatase-2 (SHP-2) tyrosine phosphatase. These results indicate that PECAM-1 and Lyn act cooperatively to inhibit collagen-related peptide (CRP)–induced platelet activation.

CRP (GKP*[GPP*]₁₀GKP*G, in which P* represents hydroxyproline) was synthesized and cross-linked in the Protein Chemistry Core of the Blood Research Institute as previously described.¹⁵  The PECAM-1 (M-20), SH-PTP2 (C-18), SH-PTP2 (B-1), and goat anti–rabbit IgC-HRP (SC-2004) antibodies were purchased from Santa Cruz Biotechnology. The affinity-purified rabbit polyclonal antibody specific for PECAM-1 phosphorylated on the tyrosine residue at position 686, anti–PECAM-1 pY686, was developed and characterized in our laboratory at the Blood Research Institute (C. P. Paddock, B. L. Lytle, F. C. Peterson, T. Holyse, P.J.N., B. F. Volkman, and D.K.N., manuscript in revision).

Mice were maintained in a facility free of well-defined pathogens under the supervision of the Biomedical Resource Center at the Medical College of Wisconsin. Animal protocols were approved by the Medical College of Wisconsin Institutional Animal Care and Use Committee. PECAM-1–deficient mice (pecam-1^−/−),¹⁶  backcrossed for more than 10 generations onto a C57BL/6J background, were cross-bred with Lyn-deficient mice (lyn^−/−),¹⁷  kindly supplied by Dr C. Lowell (University of California, San Francisco) to obtain double-heterozygous mice (pecam-1^+/−/lyn^+/−) that were then bred to establish pecam-1^+/−/lyn^+/+, pecam-1^+/+/lyn^+/−, pecam-1^+/−/lyn^−/−, and pecam-1^−/−/lyn^+/− breeding colonies. Matings of pecam-1^+/−/lyn^+/+ or pecam-1^+/+/lyn^+/− breeders yielded single-deficient offspring (pecam-1^−/−/lyn^+/+ or pecam-1^+/+/lyn^−/−) and age- and sex-matched wild-type controls. Matings of pecam-1^+/−/lyn^−/− or pecam-1^−/−/lyn^+/− breeders yielded double-deficient offspring (pecam-1^−/−/lyn^−/−) and age- and sex-matched single-deficient controls (pecam-1^+/+/lyn^−/− or pecam-1^−/−/lyn^+/+). Mouse genotypes were determined by polymerase chain reaction amplification of genomic tail DNA, and lack of PECAM-1 or Lyn expression was confirmed by Western blot analysis of platelet lysates.

Washed platelets were prepared from the pooled whole blood of 3 to 4 genetically identical 6- to 9-week-old mice, and platelet suspensions (3 × 10⁸ platelets/mL) were stimulated with CRP. For aggregation studies, platelets were stimulated in an aggregometer with varying doses of CRP essentially as described previously.¹⁸  Light transmission was monitored over time, and aggregation was quantified as a function of light transmission, with 100% aggregation corresponding to 100% light transmission.

Platelet lysate preparation, Western blot analyses, and immunoprecipitation (IP) of SHP-2 from platelet lysates were performed essentially as described previously.¹⁸  Western blot analysis for detection of PECAM-1 tyrosine phosphorylation, PECAM-1 antigen, and SHP-2 used the rabbit polyclonal antibodies anti–PECAM-1 pY686, PECAM-1 (M-20), and SH-PTP2 (C-18), respectively, as primary antibodies and goat anti–rabbit IgG-HRP as the secondary antibody. The mouse monoclonal antibody SH-PTP2 (B-1) was used for IP of SHP-2.

To determine whether Lyn and PECAM-1 are required for the inhibitory function of each other in platelets, we measured aggregation in response to various concentrations of the GPVI-specific agonist, CRP, in platelets derived from wild-type, PECAM-1–deficient, Lyn-deficient, and PECAM-1/Lyn double-deficient mice. First, we performed pairwise comparisons of the responses of PECAM-1– or Lyn-deficient platelets with those of PECAM-1/Lyn-positive or PECAM-1/Lyn-negative platelets that were otherwise genetically identical. This approach enabled statistical analysis of quantitative differences in aggregation that were observed in platelets derived from mice that differed from one another only by PECAM-1 and/or Lyn deficiency. The effect of inhibitory molecules can best be observed at agonist concentrations that induce an intermediate level of response in the presence of the inhibitor, so that a higher level of responsiveness can be detected in the absence of the inhibitor. Consequently, platelets were stimulated with doses of CRP that were empirically determined to induce low, intermediate, and high levels of responsiveness in the presence of the putative inhibitor for each pairwise comparison. As shown in Figure 1A, concentrations of CRP that elicited an intermediate level of responsiveness from platelets that contained both PECAM-1 and Lyn elicited a response of significantly greater magnitude from platelets that were missing either PECAM-1 (Figure 1Ai) or Lyn (Figure 1Aii), which indicates that PECAM-1 and Lyn each function as inhibitory molecules in the presence of the other. In contrast, concentrations of CRP that elicited an intermediate response from platelets that contained either PECAM-1 only (Figure 1Aiii) or Lyn only (Figure 1Aiv) failed to elicit a response of greater magnitude from platelets that were missing both PECAM-1 and Lyn, demonstrating that PECAM-1 has no inhibitory function in the absence of Lyn, and that Lyn loses its inhibitory function in the absence of PECAM-1. Note that, because different doses of CRP were used for each pairwise comparison of platelet aggregation, comparisons among the different experiments in Figure 1A are not valid. Secretion was assessed by lumiaggregometry at least once for every pairwise comparison shown in Figure 1A and was similarly affected by PECAM-1 and/or Lyn deficiency (data not shown). To ensure that single- and double-deficient platelets were equally responsive relative to each other and hyperresponsive relative to double-positive platelets, we compared the responsiveness of platelets derived from mice of all 4 genotypes in a single experiment. As shown in Figure 1B, when directly compared, PECAM-1–deficient, Lyn-deficient, and PECAM-1/Lyn double-deficient platelets were all equally hyperresponsive to stimulation with a dose of CRP that induced an intermediate level of responsiveness from wild-type platelets. Together, these results indicate that PECAM-1 inhibits platelet responses in the presence, but not in the absence, of Lyn and that Lyn inhibits platelet responses in the presence, but not in the absence, of PECAM-1. We conclude from these results that PECAM-1 and Lyn depend on one another to dampen platelet responses to weak, GPVI-specific stimuli.

Lyn-deficient mast cells exhibit defective tyrosine phosphorylation of, and SHP-2 binding to, PECAM-1 ITIMs.¹⁹  We therefore sought to determine whether Lyn is required for PECAM-1 ITIM phosphorylation and SHP-2 binding in platelets. PECAM-1 failed to become phosphorylated on its C-terminal ITIM tyrosine (Figure 2A) and to recruit SHP-2 (Figure 2B) in PECAM-1–positive, Lyn-deficient platelets that were allowed to aggregate to the same extent as wild-type platelets after stimulation with CRP. These data demonstrate that Lyn is required for both PECAM-1 phosphorylation and SHP-2 binding in GPVI-activated platelets. Platelets possess SFKs other than Lyn that have been shown to associate with PECAM-1 in platelets²⁰  or mediate PECAM-1 phosphorylation in other cells.^21-26  The reasons why these other SFKs do not cooperate with PECAM-1 to inhibit platelet responses to GPVI stimulation are not known, but potentially include their lack of involvement in platelet activation by GPVI-specific agonists, or failure to be active and productively colocalize with PECAM-1 during this crucial early phase of GPVI-dependent activation. Our previous studies have revealed that Lyn becomes inactivated, in an aggregation-dependent manner, by Csk-mediated phosphorylation of its C-terminal inhibitory tyrosine residue.²⁷  Inactivation of Lyn to greater and greater extents as platelet aggregation progresses may account for the decreased levels of PECAM-1 phosphorylation and SHP-2 binding that were observed in fully relative to partially aggregated platelets (Figure 2). Collectively, these findings suggest that continued responsiveness of platelets to GPVI-specific agonists, especially if present in low doses, may require persistent inhibition of Lyn so that the inhibitory effects of PECAM-1/SHP-2 complexes can be overcome.

Our results are consistent with an obligatory cooperative model in which Lyn and PECAM-1 act in an interdependent manner to control the rate and extent of early platelet activation. In resting platelets, active forms of Lyn are associated with paxillin family members²⁷  and with the GPVI cytoplasmic domain,⁸  in which state Lyn is primed to contribute to rapid platelet activation. Observations of delayed responsiveness of Lyn-deficient platelets on stimulation with collagen^7-9  or von Willebrand factor²⁸  provide evidence in support of an important role for Lyn in initial platelet activation to these agonists. Once platelets have been activated, however, Lyn initiates a feedback inhibitory pathway by phosphorylating PECAM-1 ITIMs, which go on to recruit SHP-2 and form PECAM-1/SHP-2 complexes that terminate activation signals by dephosphorylating, among other targets, GPVI-associated Fc receptor γ-chain or FCRγ chain ITAMs.^18,29  The observations that Lyn-deficient platelets exhibit enhanced spreading on immobilized fibrinogen³⁰  and potentiated responses to CRP stimulation⁷  support an important role for Lyn in inhibiting platelet responses at later stages of activation. The extent to which continued platelet responsiveness in the face of Lyn-mediated inhibitory activity requires activation of SFKs other than Lyn or, additionally, active inhibition of Lyn²⁷  remains to be determined.

This work was supported by the National Institutes of Health (RBI HL90883-PKN) and the Blood Center Research Foundation.

National Institutes of Health

Contribution: Z.M. performed research, analyzed data, and wrote the paper; Y.H. and J.X. helped design research and analyze data; P.P. performed research and analyzed data; P.J.N. helped design research and write the paper; and D.K.N. designed research, analyzed data, and wrote the paper.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Debra K. Newman, PhD, Blood Research Institute, BloodCenter of Wisconsin, 8727 Watertown Plank Rd, Milwaukee, WI 53226; e-mail: Debra.Newman@bcw.edu.