Dextran sulfate triggers platelet aggregation via direct activation of PEAR1.

UNLABELLED
Dextran sulfate (DxS; Mr 500 kD) induces fibrinogen receptor (αIIbβ3) activation via CLEC-2/Syk signaling and via a Syk-independent SFK/PI3K/Akt-dependent tyrosine kinase pathway in human and murine platelets. The platelet surface receptor, responsible for the DxS-induced Syk-independent Akt-activation, has hitherto not been identified. We found that DxS elicited a concentration-dependent aggregation of human platelets resulting from direct PEAR1 activation by DxS. Blocking the PEAR1 receptor, in combination with a selective Syk-inhibitor, completely abrogated the DxS-driven platelet aggregation. The DxS-induced Syk-phosphorylation was not affected in Pear1(-/-) platelets, but Akt-phosphorylation was largely abolished. As a result, the aggregation of Pear1(-/-) platelets was reduced and reversible, i.e. aggregates were less stable compared to wild-type platelet aggregates. Moreover, DxS-induced Pear1(-/-) platelet aggregation was fully abrogated by Syk inhibition, indicating that the remaining platelet aggregation of Pear1(-/-) platelets was Syk dependent. Hence, the Pear1/c-Src/PI3K/Akt- and CLEC-2/Syk-signaling pathways are independently and additively activated during platelet aggregation by DxS.


CONCLUSION
The DxS-induced aggregation of human and murine platelets is the result of activation of PI3K/Akt through direct PEAR1 phosphorylation and parallel Syk-signaling through CLEC-2.

Introduction PEAR1 (platelet endothelial aggregation receptor-1), a type-1 transmembrane protein of the multiple epidermal growth factor (EGF)-like domain protein family, is mainly expressed in platelets and endothelial cells [1,2]. PEAR1 comprises an extracellular EMI domain (protein-protein interaction domain), 15 extracellular EGF-like repeats, and multiple cytoplasmic tyrosines and prolines [1]. Nanda et al. identified PEAR1 as a plateletplatelet contact receptor. During platelet aggregation, PEAR1 becomes phosphorylated at Tyr-925 and Ser-953/1029, which is mediated by the Src family kinases (SFK) c-Src and Fyn, but not Syk. [1]. We reported that PEAR1, c-Src and Fyn, and the p85/ phosphatidylinositol 3-kinase (PI3K) subunit constitute a signaling complex in platelets that sustains the activation of α IIb β 3 in aggregating platelets, favoring the formation of stable platelet aggregates [3], and we identified the high-affinity immunoglobulin E (IgE) receptor subunit α (FcεR1α) as a platelet PEAR1 ligand [4].
Since both DxS and PEAR1 induce sustained platelet aggregation via the SFK/PI3K/Akt pathway in platelets, and since activation of PEAR1 in human platelets is independent from Syk [3], we hypothesized that the Syk-independent DxS-induced platelet aggregation involves PEAR1. The aim of this study was to investigate whether DxS-induced SFK/PI3K/Akt-signaling occurs via direct phosphorylation of PEAR1 ( Figure S1) and whether the activation of CLEC-2 and PEAR1 by DxS suffices to explain DxS-induced platelet aggregation (Figure 1).
sFcεR1α, s5FcεR1α, sPEAR1, and s5PEAR1 Monomeric recombinant FcεR1α protein, pentameric recombinant FcεR1α protein (s5FcεR1α), and pentameric recombinant PEAR1 (s5PEAR1) constructs were designed, produced, and validated as recently published by our group [4]. A human recombinant extracellular PEAR1 domain was made by cloning the corresponding mRNA fragment, encoding amino-acid residues Met1-Ser-754 into the pSecTag2/Hygro A-vector (Life Technologies, Ghent, Belgium). The resulting construct was transfected in COS-7 cells using a jetPRIME transfection kit (Polyplus, Leuven, Belgium) according to the manufacturer's protocol. Conditioned medium containing the extracellular PEAR1 recombinant protein was collected after 48 h. The presence of homogeneous recombinant PEAR1 protein in the medium was confirmed by Western blot (Mr 125 kDa).

DxS binding to PEAR1 (ELISA)
Microtiter plates were coated (O/N, 4°C) with 20 μg/ml pentameric s5FcεR1α in 100 μl TBS, washed with TBS + 0.1% BSA and blocked with 1% BSA in TBS for 1 h at room temperature (RT). Conditioned medium (200 μl) containing the extracellular domain of PEAR1 (see below) was loaded in duplicates and incubated for 2 h at RT, followed by the addition of 100 μl anti-PEAR1 antibody (1 μg/ml) for an additional hour at RT. Binding of DxS was investigated by co-incubation of PEAR1 with DxS during the first binding step. Bound anti-PEAR1 antibodies were detected via HRP-conjugated rabbit anti-goat IgG (1/2000, in TBS + 0.1% BSA for 1 h, RT). Binding was visualized by the addition of 100 μl chromogenic substrate (TMB) for 30 min. The reaction was stopped with 50 μl H 2 SO 4 (2M) and absorbance at 450 nm was measured using a PowerWave X-340 plate reader (Bio-Tek Instruments, Winooski, VT, USA).

Platelet aggregation
Venous blood was collected from healthy donors. Washed platelets were prepared as previously published [3], in the presence of 0.1 IU/ml of apyrase (Sigma-Aldrich). Platelet aggregation was monitored by measuring light transmission through the stirred suspension of washed platelets (3 × 10 5 /μL) at 37°C with a Chronolog dual-beam aggregometer, in the absence of added fibrinogen. In some experiments, platelets were preincubated with LY294002 (5 min, 50 μM), PP1 (10 min, 10 μM), DMSO (5 min), eptifibatide (10 min, 10 µg/ml), or BAY 61-3606 (10 min, 10 μM) at 37°C. The DMSO concentration never exceeded 0.2% (vol/vol). Platelet aggregation was triggered by DxS after adding Ca 2+ (2 mM) and measured and expressed as the percentage of change in light transmission, with the value for the blank sample (buffer without platelets) set at 100%. Each aggregation plot is the representative image of at least three independent experiments. Each aggregation was stopped by adding sodium dodecyl sulfate (SDS) denaturing buffer or ice-cold Triton lysis butter at 30% aggregation to avoid displacement of PEAR1 to the insoluble cytoskeleton fraction (as previously published [3]).

DxS induces human and murine platelet aggregation
DxS induced a concentration-dependent aggregation of stirred washed human ( Figure 2A) and murine ( Figure 2B) platelets in the concentration range previously shown to provoke platelet activation and ATP secretion [5]. Getz et al. performed murine platelet aggregations in the presence of DxS 5 nM. We did not observe substantially altered aggregation of murine platelets for DxS 0.25 nM compared to DxS 5 nM ( Figure 2B). The DxSinduced human platelet aggregation was abrogated in the presence of the α IIb β 3 -receptor blocker eptifibatide, confirming that DxS induced proper platelet aggregation and not platelet agglutination ( Figure 2A).

DxS induces direct PEAR1 activation
As previously published by our group [3], PEAR1 phosphorylation can be the result of direct PEAR1 activation via a PEAR1-ligand interaction or it can be triggered indirectly following plateletplatelet contact as part of platelet amplification induced by various classical platelet agonists (e.g., thrombin, collagen). This is schematically shown in Figure S1; Western blots show corresponding pPEAR1 (PEAR1 immunoprecipitation and Western blot for P-Tyr; Figure S1G).
Activation of platelets with a specific PEAR1 ligand (e.g., soluble recombinant pentameric FcεR1α (s5FcεR1α) or anti-PEAR1 extracellular antibodies, Figure S1A) and activation of platelets with traditional platelet agonist (e.g., collagen or thrombin, binding to their own classical receptors; Figure S1B) result both in platelet aggregation and in PEAR1 phosphorylation under stirring conditions. However, blocking the α IIb β 3 -receptor under stirring conditions in the direct PEAR1-activation pathway ( Figure S1C) preserves the phosphorylation of PEAR1, although aggregation is absent. In contrast, blocking the α IIb β 3 -receptor in the indirect PEAR1-activation pathway ( Figure S1D) results in abrogation of the PEAR1 phosphorylation.
Therefore, to illustrate that DxS binds directly to PEAR1, we analyzed DxS-induced PEAR1 phosphorylation under stirring conditions in the presence of an α IIb β 3 -receptor blocker ( Figure S1C) or, alternatively, under static conditions ( Figure S1E and S1F) to avoid indirect PEAR1 phosphorylation by platelet-platelet contacts. Next, we performed competition experiments (ELISA and platelet aggregation) between DxS and the PEAR1 ligand FcεR1α to confirm the direct DxS-PEAR1 interaction.

Static incubation of washed human platelets with DxS
Static incubation of washed human platelets with DxS (0-10 nM) for 15 min resulted in a dose-dependent PEAR1 phosphorylation comparable to that induced by anti-PEAR1 antibodies (PEAR1 immunoprecipitation and Western blot for P-Tyr; Figure 3A). In agreement with Figure S1E and S1F, this phosphorylation is compatible with direct binding of DxS to PEAR1.

Stirring conditions in the presence of an α IIb β 3 -receptor blocker
To further investigate a direct interaction between DxS and PEAR1, we investigated the phosphorylation state of PEAR1 in DxS-induced platelet aggregation in the presence of eptifibatide ( Figure 3B; anti-PEAR1-Ab served as a positive control). The phosphorylation state of PEAR1 was unaffected by eptifibatide ( Figure 3B, right panel), further supporting direct binding of DxS to PEAR1 (in agreement with Figure S1C).

Competition experiments
Next, we confirmed the direct binding of DxS to PEAR1 by competition experiments for PEAR1 with its physiologic ligand FcεR1α. We previously reported that phosphorylation of PEAR1 requires clustering of the PEAR1 receptor [3,4]. As summarized in Figure 3C, static incubation of platelets with a (divalent) anti-PEAR1-extracellular antibody or with pentameric recombinant FcεR1α (s5FcεR1α) resulted in PEAR1 activation, whereas incubation with Fab-fragments of the anti-PEAR1 antibody [8] or with a monomeric form of recombinant FcεR1α (sFcεR1) prevents clustering of PEAR1 [3,4], abrogating its phosphorylation.
As shown in Figure 3D and E, DxS-induced PEAR1 phosphorylation is blocked when platelets were preincubated with an excess of recombinant pentameric PEAR1 (s5PEAR1), compatible with  the direct binding of pentameric s5PEAR1 to DxS and thus competing with DxS binding to platelet PEAR1. Similarly, DxSinduced PEAR1 phosphorylation is blocked when platelets were preincubated with an excess of monomeric sFcεR1α, shielding the PEAR1 receptor from DxS-mediated PEAR1 clustering. Since we were interested in direct phosphorylation of PEAR1, these competition experiments were performed under static conditions. We also confirmed the direct interaction of DxS to PEAR1 through ELISA; the binding of recombinant PEAR1 to its coated ligand, i.e., recombinant pentameric FcεR1α (s5FcεR1α), was dose dependently inhibited by DxS (500 kD; 1-100 nM) with an IC 50 = 25 ± 3.4 nM ( Figure 3F). All of these results confirmed direct phosphorylation of PEAR1 by DxS.
The PEAR1 antagonist sFcεR1α alone incompletely blocked aggregation, without affecting the degree of Syk phosphorylation ( Figure 4B). Correspondingly, inhibition of the Syk/PLCγ2 pathway after preincubation with the selective Syk-inhibitor BAY 61-3606 only partially reduced DxS-mediated human platelet aggregation ( Figure 4C). The phosphorylation of PEAR1 and Akt remained unaffected in the presence of BAY 61-3606 ( Figure 4A; right panel), because of the lack of Syk involvement in PEAR1 signaling, as previously reported [3].
To further support these findings in human platelets, we will extrapolate our results to murine platelets and confirm our observations in Pear1 −/− platelets.

DxS and Pear1 signaling in murine platelets
This is the first manuscript using Pear1 −/− platelets. Therefore, we first investigated whether the Pear1-signaling pathway in murine platelets is comparable to that in human platelets and confirmed the direct activation of Pear1 in murine platelets by DxS as well. Pear1 was absent in washed Pear1 −/− platelets ( Figure 5A). Similar to human platelets, incubation of washed murine WT platelets with a polyclonal murine anti-Pear1-Ab resulted in phosphorylation of Pear1 and similar phosphorylation was observed after incubation with DxS 0.25 nM ( Figure 5B; P-Tyr after immunoprecipitation for Pear1; Western blot). Also similar to human platelets, we found that the phosphorylation of Pear1 by DxS was blocked by the SFK-inhibitor PP1 ( Figure 5C; P-Tyr after immunoprecipitation for Pear1; Western blot) and that DxS induced the phosphorylation of Akt downstream of Pear1, which was blocked by the PI3K-inhibitor LY294002 ( Figure 5D; Western blot). These findings identify a similar signaling pathway for PEAR1 in human and murine platelets. In order to avoid indirect Pear1 activation, as previously explained, these signaling experiments were performed under static conditions, again confirming direct binding of DxS to Pear1 (see Figure S1E).

Pear1 −/− platelets confirm that DxS-induced platelet aggregation is mediated via Pear1 and CLEC-2/Syk
The DxS-induced platelet aggregation in Pear1 −/− platelets was reduced and reversible compared to WT platelet aggregation ( Figure 6A; left panel), compatible with a weaker Pear1-induced α IIb β 3 -integrin stabilization and thus decreased aggregate stability. Corresponding Western blots of Pear1 −/− platelets compared to WT platelets resulted in strongly reduced phosphorylation of Akt in Pear1 −/− platelets, whereas phosphorylation of Syk and PLCγ2 was not affected ( Figure 6A; right panel). Finally, we showed that the residual DxS-induced platelet aggregation of Pear1 −/− platelets could be fully abrogated by the Syk-inhibitor BAY 61-3606 (compared to aggregation of WT platelets with DxS; Figure 6B), confirming that both Pear1-and Syk-coupled receptor signaling are sufficient to explain the DxS-induced platelet aggregation.

Discussion
DxS was originally evaluated as an anticoagulant with heparinlike properties [9]. The present study was undertaken based on recent findings that high molecular DxS initiated human and murine platelet activation both via a Syk-dependent [5,10] and a Syk-independent pathway, the latter involving PI3K-mediated phosphorylation of Akt [5]. Alsheri et al. recently showed that the Syk-dependent pathway through which DxS induces platelet aggregation is predominantly mediated via CLEC-2 [7], but the platelet surface receptor(s) responsible for the DxS-induced platelet aggregation through the Syk-independent pathway remained to be identified [5,7].
Both DxS and PEAR1 induce sustained platelet aggregation via SFK/PI3K/Akt and share the same underlying signaling pathway in human platelets [3,5]. Therefore, we hypothesized that PEAR1 is responsible for the activation of the Syk-independent pathway upon DxS-induced platelet aggregation, compatible with our previous findings, showing that PEAR1 signaling does not recruit Syk (Figure 1). To support this hypothesis, it was crucial to show that DxS directly induced phosphorylation of PEAR1 or in other words to show that DxS functions as an activating ligand for PEAR1. Therefore, as schematically shown in Figure S1, we analyzed whether DxS induces phosphorylation of human platelet PEAR1 under static incubations and under stirring conditions in the presence of an α IIb β 3 -receptor blocker, in order to avoid indirect PEAR1 phosphorylation induced by PEAR1-FcεR1α interactions during platelet-platelet contacts. We also performed competition experiments with an excess of the PEAR1-blocking sFcεR1α-monomer [4] (shielding the receptor and preventing PEAR1-multimerization) and by an excess of soluble pentameric PEAR1 (s5PEAR1), inhibiting the binding of DxS to platelet PEAR1. All these results identified DxS as a direct activating ligand for PEAR1 and excluded indirect activation of PEAR1 through DxS, as seen during platelet aggregation by classical agonist, e.g., collagen and thrombin.
Next, we showed that the activation of Syk and PEAR1 by DxS suffices to explain the DxS-induced platelet aggregation. Although preincubation with monomeric sFcεR1α eliminated the DxS-induced PEAR1 activity, the remaining DxS-induced human platelet aggregation was substantial. Correspondingly, shielding the PEAR1 receptor during aggregation did not affect the phosphorylation state of Syk or PLCγ2. In counterpart, blocking the phosphorylation of Syk by a selective Syk inhibitor [11,12] also mildly reduced the DxS-induced human platelet aggregation, whereas it did not affect phosphorylation of PEAR1. Furthermore, the combined preincubation with selective blockers of both the PEAR1 and Syk pathway completely abrogated DxS-induced platelet aggregation as well as PEAR1 and Syk phosphorylation. All these results suggested that both receptors independently but additively initiate DxS α IIb β 3 -mediated human platelet aggregation. To further support our findings, we extrapolated our analyses to murine platelets and confirmed our findings in Pear1 −/− platelets.  Figure S1).
Similarly as previously reported [5], we confirmed that DxS (5 nM) also induces murine platelet aggregation, although we observed no significant differences in aggregation intensity between low-dose (0.25 nM) and high-dose (5 nM) DxS. We identified the presence of Pear1 on murine platelets and showed for the first time that murine Pear1 signaling is coupled to the SFK/PI3K/Akt pathway, similarly as for human PEAR1. We showed that murine platelet Pear1 was also directly phosphorylated by DxS under static conditions, to the same extent as reached with a polyclonal anti-mPear1-Ab (positive control). Interestingly, DxS-induced Pear1 −/− platelet aggregation was reduced and reversible, compared to WT platelet aggregation, compatible with a weaker Pear1-induced α IIb β 3 -integrin stabilization and thus decreased aggregate stability, in line with our findings in human platelets [3]. The Pear1 −/− platelet aggregation by DxS was reduced compared to WT platelets but not completely abrogated, confirming that Pear1 is not the only surface receptor on mouse platelets activated by DxS during platelet aggregation, in agreement with previous reports [7,10]. Corresponding Western blots showed that DxS only induced minimal phosphorylation of Akt in Pear1 −/− platelets, whereas Syk and PLCγ2 phosphorylation was comparable to that for WT platelets activated by DxS. Ultimately, we found that the DxS-induced platelet aggregation of Pear1 −/− platelets was abrogated after preincubation with a Syk inhibitor. This indicated that the remaining platelet aggregation of Pear1 −/− platelets was fully Syk dependent, confirming that Pear1 and Syk signaling both are sufficient to explain DxS-induced murine platelet aggregation, as seen in human DxS-induced platelet aggregation.
The major platelet receptors that employ SFK and Syk are the platelet receptor for podoplanin CLEC-2, the major collagen receptor GPVI, and the platelet receptor for von Willebrand GPIb. As mentioned above, Alsheri et al. recently found that DxS activates both GPVI and CLEC-2, with a clear preference for CLEC-2 over the collagen receptor. They found reduced platelet aggregation in CLEC-2-deficient and in GPVI/CLEC-2 double-deficient platelets, whereas the response to DxS was not altered in GPVI-deficient platelets [7]. Although the DxS-induced Akt phosphorylation was strongly reduced during aggregation of Pear1 −/− platelets, it was not completely absent. This is in agreement with the recent report of Manne et al. [10], who showed that the PI3K/Akt pathway is a downstream player of CLEC-2, further supporting that CLEC-2 is the Syk-coupled receptor responsible for DxS-induced platelet aggregation. However, our data show that PEAR1 activation is the major trigger for the DxS-induced Akt signaling and that there is only a minor role for Akt in DxSactivated Pear1 −/− platelets, downstream of CLEC-2.
In conclusion, DxS is a direct ligand of PEAR1, causing α IIb β 3 activation via the PEAR1/PI3K/Akt pathway and via the complementary activation of Syk, both in human and in murine platelets. The major effect of DxS on Akt activation results from PEAR1 phosphorylation via c-Src through clustering of the receptor and direct activation of PEAR1, whereas DxS-induced Syk signaling occurs independently from PEAR1-pathway activation. The direct PEAR1 signaling in human and murine platelets is similar and Pear1 stabilizes murine platelet aggregates.
Further research will have to define whether DxS-induced platelet aggregation can be of importance in certain (physio) pathologic processes and further research in Pear1 −/− mice will need to address the contribution of Pear1 to thrombus stability in proper models of flow-dependent thrombosis and whether PEAR1 is activated by other negatively charged and sufficiently long biologically active polymers (e.g., dermatan, heparan sulfate, chondroitin sulfate).  Figure 6. Pear1 −/− platelets confirmed that DxS-induced platelet aggregation is mediated via Pear1 and CLEC-2/Syk. (A) DxS-induced platelet aggregation in Pear1 −/− platelets was weaker and displayed decreased platelet aggregate stability compared to WT platelets, compatible with a weaker Pear1-induced α IIb β 3integrin stabilization (left panel); corresponding Western blot for phosphorylation of Akt, Syk, and PLCγ2 showed a strongly reduced, but not completely abrogated, phosphorylation of Aktin DxS-induced Pear1 −/− platelet aggregation, whereas the Syk/PLCγ2 activation remained unaffected. (B) DxS-induced platelet aggregation was fully eliminated in Pear1 −/− platelets after preincubation with a Syk inhibitor, indicating that Pear1 and Syk signaling is sufficient to explain DxS-induced platelet aggregation. (A-B: washed murine platelets). G0A6514N). The Center for Molecular and Vascular Biology is supported by the "Programmafinanciering KU Leuven (PF/10/014)" and by the "Geconcerteerde Onderzoeksacties" (GOA 2009/13) from the University of Leuven. This research was supported by the Bayer Chair in Cardiovascular Medicine, an unrestricted grant provided to PV.

Ethical committee
Human samples were handled according to the Helsinki Declaration. Formal permission was given by the Ethics Committee of the Leuven University Hospitals to use blood from healthy individuals. All animal experimental procedures were approved by the local Ethics Committee of the KU Leuven.

Supplemental material
Supplemental data for this article can be accessed on the publishers' website.

Declaration of interest
None declared.