Protein kinase C signaling dysfunction in von Willebrand disease (p.V1316M) type 2B platelets

von Willebrand disease (VWD) type 2B is a paradoxical bleeding disorder resulting from gain-of-function mutations in the A1 domain of von Willebrand factor (VWF), which is responsible for the binding of the molecule to the platelet receptor glycoprotein (GP)Ibα. VWD type 2B is characterized by reduced VWF antigen levels, lack of high-molecular-weight VWF multimers,¹  circulating platelet aggregates, and variable thrombocytopenia, which are dependent on the causative mutation.^2,3  For many years, the severity of the bleeding tendency in VWD type 2B patients has been associated with the low platelet count and the absence of high-molecular-weight VWF multimers.²  We have recently demonstrated that a severe thrombopathy also aggravates this complex clinical picture. Indeed, we showed that VWF/p.V1316M alters platelet signaling by inhibiting the activation of the small GTPase Rap1B, which is critical for talin recruitment and subsequent integrin αIIbβ3 activation.⁴ 

The 2 Rap1 isoforms, Rap1A and Rap1B, are the most abundant small GTPases expressed in platelets.⁵  Rap1 GTPases switch between a GTP-bound (active) and a GDP-bound (inactive) state. All known platelet agonists stimulate GTP loading of Rap1.^6-9  Our recent work identified key pathways regulating Rap1B activation in platelets: fast, but reversible, activation mediated by a calcium-sensing guanine nucleotide exchange factor (CalDAG–GEF-I) and slow, but sustained, activation mediated by protein kinase C (PKC) and the platelet receptor for adenosine 5′-diphosphate (ADP), P2Y12.^10-12 

PKC has a well-documented role in platelet signaling, where it controls the release of storage granules and, thus, the release of the second-wave mediator of platelet activation, ADP.¹³  Moreover, PKC activation has been shown to induce the proteolytic cleavage (shedding) of various platelet surface receptors, including the GPIbα subunit of the VWF receptor complex (GPIb-V-IX). Shedding of GPIbα is a constitutive process in humans and mice, as confirmed by the presence of basal amounts of soluble GPIbα (glycocalicin) in plasma.^14,15  Constitutive and agonist-induced shedding of GPIbα are strongly dependent on the metalloproteinase ADAM17.¹⁶  However, alternative sheddases may also contribute to the regulation of surface expression levels of GPIbα.^16-18  In cells other than platelets, distinct signaling pathways control shedding of receptors, such as epidermal growth factor receptor¹⁹  and CD44.²⁰  Both PKC-dependent (mainly PKCα and PKCδ) and PKC-independent mechanisms have been described.

In the present study, we investigated the molecular mechanisms leading to the severe thrombopathy recently described in p.V1316M-associated VWD type 2B patients and mice.⁴  We demonstrate that mutant VWF/p.V1316M engagement of the GPIbα receptor at the platelet surface leads to upregulated baseline PKC activity in human and murine VWD (p.V1316M) type 2B platelets and a defect in the PKC/P2Y12/Rap1 signaling response to agonist stimulation. These changes in PKC activity lead to increased shedding of GPIbα in mice, a marked reduction in platelet granule release, and impaired integrin activation in mice and humans. Together, these alterations protect the remaining circulating platelets from clearance, an adaptation that is critical to prevent thrombocytopenia and/or thrombosis.

Detailed information is provided in supplemental Methods.

Eight- to 12-week-old C57BL/6 wild-type (WT) mice were purchased from The Jackson Laboratory (Bar Harbor, ME).

IL4R-Ibα–transgenic (Tg),²¹ Caldaggef1^−/−,²² Stim1^fl/flPF4-Cre⁺,²³ Caldaggef1^−/−P2Y12^−/−,²⁴ talin1^fl/flPF4-Cre⁺,²⁵  and Adam17^fl/flPF4-Cre⁺ (Adam17^tm1.2Bbl/J mice were purchased from The Jackson Laboratory), and Vwf/p.V1316M²⁶  mice were bred and housed in the animal facility at the University of North Carolina at Chapel Hill.

Hydrodynamic gene transfer (50-100 μg of pLIVE-murineVWF/p.V1316M plasmid) was performed in mice expressing endogenous VWF/WT, as previously reported for Vwf^−/− mice^1,27,28  (supplemental Figure 1). WT, IL4R-IbαTg, Caldaggef1^−/−, Stim1^fl/flPF4-Cre⁺, Caldaggef1^−/−P2Y12^−/−, talin1^fl/flPF4-Cre⁺, ADAM17^fl/flPF4-Cre⁺, and Vwf^−/− mice (as control) were subjected to the procedure. VWF antigen plasma levels were measured before and after hydrodynamic injections to confirm VWF expression (supplemental Table 1). All experiments were conducted 3-7 days after gene transfer. Mice expressing mutant VWF due to hydrodynamic gene transfer, together with endogenous wild-type VWF (ie, WT-VWF/p.V1316M for [WT-] wild-type mice expressing mutant VWF and endogenous VWF/WT), were compared to mice with the same genotype expressing only endogenous VWF/WT (ie, WT-VWF/WT for [WT-] wild-type mice expressing endogenous VWF/WT only). Following hydrodynamic gene transfer, all mice (with the exception of Vwf^−/−) will express endogenous VWF/WT and high levels of mutant VWF (p.V1316M).

Proteins in mouse platelet lysates or plasma, prepared as described in supplemental Methods, were separated by sodium dodecyl sulfate gel electrophoresis (4%-20% gels; Bio-Rad, Hercules, CA) and transferred to a polyvinylidene difluoride membrane. Phosphorylated PKC substrates, β-actin, or glycocalicin were detected using appropriate secondary antibodies and an Odyssey Infrared Imaging System (LI-COR Biosciences, Lincoln, NE). Quantification was performed with Image Studio Lite software (LI-COR Biosciences).

Phosphorylated PKC substrates in human platelet lysates from 3 VWD (p.V1316M) type 2B patients were detected as previously described.⁴ 

Mice anesthetized with ketamine/xylazine (100 mg/kg and 10 mg/kg, respectively) received Alexa Fluor 488–conjugated antibodies to GPIX (2.5 µg) to label circulating platelets. The saphenous vein was exposed and injured as described previously.²⁹  Platelet accumulation at the site of laser injury was assessed by intravital microscopy.

All animal experiments and protocols were reviewed, approved by, and performed in accordance with the Animal Care and Use Committee of the University of North Carolina at Chapel Hill. All patients were recruited in France, informed about the anonymous use of their data, and gave written informed consent in accordance with the Declaration of Helsinki. The French VWD cohort (Centre de Référence Maladie de Willebrand) was reported to the French data protection authority.

Statistical significance was assessed using Prism 7 software (GraphPad, La Jolla, CA). Two- or 3-way analysis of variance (Bonferroni post hoc analysis) was applied for multiple- and 3-group comparison, respectively. A 2-tailed Student t test was performed for 2-group comparison of parametric data. P ≤ .05 was considered significant. All data are presented as mean ± standard deviation (SD).

We applied hydrodynamic gene transfer, which was previously used to express VWF/p.V1316M in Vwf^−/− mice,^1,27  to induce VWD type 2B–like phenotypes in mice with normal levels of endogenous VWF. Expression of VWF/p.V1316M in WT mice resulted in an ∼90% decrease in platelet count, reducing the peripheral platelet count (PPC) from 9.51 ± 1.76 × 10⁸/mL to 1.11 ± 0.21 × 10⁸/mL (Figure 1A). This reduction is consistent with that observed in Vwf^−/− mice expressing VWF/p.V1316M, in which the PPC dropped from 12.01 ± 2.23 × 10⁸/mL to 1.35 ± 0.55 × 10⁸/mL (supplemental Figure 1A).

Flow cytometry forward scatter (FSC) values were determined as an estimation of platelet size. Compared with controls, platelet size was increased approximately twofold in WT mice expressing VWF/p.V1316M (Figure 1B). A similar twofold increase in platelet size was observed in Vwf^−/−-VWF/p.V1316M mice compared with their respective controls (supplemental Figure 1B). Furthermore, platelet aggregates were present in blood smears prepared from WT-VWF/p.V1316M mice (Figure 1C) similarly to Vwf^−/−-VWF/p.V1316M mice (supplemental Figure 1C). Thus, VWD type 2B phenotypes can be induced in WT mice expressing VWF/p.V1316M by hydrodynamic gene transfer.

We and other investigators have demonstrated that VWF is present at the platelet surface in VWD type 2B patients and mice.^27,30  VWD type 2B mutations alter VWF conformation and increase its binding affinity for platelet receptor GPIbα.^31,32  To test whether VWF/p.V1316M affects platelet count and size via binding to GPIbα, we performed hydrodynamic gene transfer in Tg mice expressing a chimeric human interleukin-4 (IL-4) receptor α/GPIbα protein (IL4R-IbαTg) (ie, mice that lack the extracellular domain of GPIbα required for binding of VWF).²¹  At baseline, the platelet count in IL4R-IbαTg mice is ∼35% lower than in WT controls (6.15 ± 0.30 × 10⁸/mL and 9.51 ± 1.76 × 10⁸/mL, respectively; Figures 1A and 2A). Importantly, platelet counts in IL4R-IbαTg mice were not significantly altered 4 days after VWF/p.V1316M overexpression (Figure 2A). Moreover, platelet size, which was increased twofold in WT mice expressing VWF/p.V1316M, did not change in IL4R-IbαTg mice after hydrodynamic gene transfer (Figure 2B). Finally, platelet aggregates were not detected in blood smears from IL4R-IbαTg-VWF/p.V1316M mice (Figure 2C). These results demonstrate that the onset of the VWD (p.V1316M) type 2B–associated macrothrombocytopenia is highly dependent on the physical interaction between VWF and the extracellular portion of GPIbα at the platelet surface.

We next investigated whether signaling events downstream of GPIbα engagement, which are important for αIIbβ3 activation, may contribute to abnormal platelet clearance and thrombocytopenia in VWD (p.V1316M) type 2B. To address this question, we expressed VWF/p.V1316M in mice with select signaling defects, such as mice defective in store-operated calcium entry (Stim1^fl/flPF4-Cre⁺), calcium-dependent and -independent Rap1 GTPase signaling (Caldaggef1^−/−P2Y12^−/−), or talin-mediated integrin signaling (talin1^fl/flPF4-Cre⁺). GPIbα engagement was shown by different groups to cause calcium mobilization and integrin activation in platelets.^33-35  We and other investigators demonstrated a key role for Rap GTPase signaling in talin-dependent integrin activation in platelets.^8,22,36  Unexpectedly, platelet aggregate formation and macrothrombocytopenia induced by VWF/p.V1316M (80-90% reduction in platelet count and twofold increase in platelet size) were not different among WT (Figure 1), Stim1^fl/flPF4-Cre⁺, Caldaggef1^−/−P2Y12^−/−, and talin1^fl/flPF4-Cre⁺ mice (Figure 2D-E). These data suggest that platelet aggregate formation and macrothrombocytopenia depend on the binding of VWF/p.V1316M to GPIbα but are independent of downstream platelet signaling events important for αIIbβ3 integrin activation.

We previously demonstrated that the activation of Rap1B was decreased in VWD (p.V1316M) type 2B platelets stimulated via G protein–coupled or immunoglobulin-like receptors.⁴  We also demonstrated that Rap1B activation upon agonist stimulation is mediated by a fast, but reversible, pathway (CalDAG–GEF-I dependent) and a slow, but sustained, pathway (PKC/P2Y12 dependent)^10-12  (supplemental Figure 2A). To establish the molecular mechanism underlying this cellular activation defect, we first studied the kinetics of αIIbβ3 activation in platelets from WT and WT-VWF/p.V1316M mice [VWD2B(p.V1316M)-platelets]. To do so, we added JON/A-PE, a probe that irreversibly binds to activated αIIbβ3, to platelets before and 10 minutes after agonist stimulation (supplemental Figure 2B). As expected, JON/A-PE binding in PAR4p-activated WT platelets was similar, irrespective of when the antibody was added (Figure 3A-B, open bars), indicating that both pathways required for full integrin activation were active. In contrast, JON/A-PE binding in PAR4p-activated VWD2B(p.V1316M) platelets was comparable to controls when the antibody was added before the agonist (Figure 3A), but it was markedly reduced if the antibody was added 10 minutes after the agonist (Figure 3B). These results suggested that a defect in PKC/P2Y12 signaling, mediating sustained integrin activation, is responsible for the decreased αIIbβ3 activation response observed in VWD2B(p.V1316M) platelets.

To further test this conclusion, we expressed VWF/p.V1316M in Caldaggef1-deficient mice. Integrin αIIbβ3 activation in platelets collected from these mice depends primarily on the contribution of the PKC/P2Y12 signaling pathway. Expression of VWF/p.V1316M led to severe macrothrombocytopenia in Caldaggef1^−/− mice (supplemental Figure 3). JON/A-PE binding in platelets isolated from Caldaggef1^−/−-VWF/p.V1316M mice did not significantly change upon agonist stimulation (Figure 3C), confirming that αIIbβ3 integrin dysfunction in VWD (p.V1316M) type 2B is a consequence of a defect in the PKC/P2Y12 signaling pathway.

Importantly, no difference in the kinetics of αIIbβ3 activation was observed between platelets from IL4R-IbαTg-VWF/WT and IL4R-IbαTg-VWF/p.V1316M mice (supplemental Figure 4), demonstrating that the abnormal interaction between GPIb and mutant VWF in VWD (p.V1316M) type 2B induces the change in PKC/P2Y12 signaling in these cells.

To investigate whether the alteration in Rap1 signaling that we observed in VWD (p.V1316M) type 2B is due to an impaired signaling function of PKC or P2Y12, we next studied integrin activation under conditions of direct stimulation of P2Y12. Only a small, but significant, defect in JON/A-PE binding was observed between control and VWD2B(p.V1316M) platelets upon costimulation with the second-wave mediators (ADP and U46619) (Figure 4A). These results suggested that VWF/p.V1316M engagement of GPIbα primarily affects PKC function in platelets. Consistent with this conclusion, VWD2B(p.V1316M) platelets exhibited marked defects in α-granule and dense granule secretion, processes that are dependent on PKC signaling. To study dense granule release, platelets were loaded with radiolabeled [³H]serotonin. Compared with controls (90% release), VWD2B(p.V1316M) platelets released only 40% of the incorporated serotonin upon PAR4p agonist stimulation (Figure 4B). As a marker for α-granule release, we quantified P-selectin at the platelet surface. Surface P-selectin levels were highly increased (30-50–fold) in PAR4p-activated WT platelets compared with nonstimulated cells (Figure 4C, red bars). In contrast, P-selectin surface expression only increased approximately fivefold in stimulated VWD2B(p.V1316M) platelets (Figure 4C, blue bars).

To directly monitor PKC activity, we determined the phosphorylation status of PKC substrates in resting and PAR4p-activated platelets. Although mouse WT platelets showed little PKC activity under resting conditions, a significant increase (6.4-fold) in PKC substrate phosphorylation was observed 10 minutes after agonist stimulation (Figure 5A-B). Compared with controls, PKC substrate phosphorylation was markedly elevated in resting VWD2B(p.V1316M) platelets (fourfold increase) (Figure 5A-B). However, upon cellular stimulation, PKC substrate phosphorylation increased only twofold in VWD2B(p.V1316M) platelets (Figure 5A-B). To exclude that the observed phenotype was due to the high circulating levels of mutant VWF associated with the gene-transfer procedure, we performed similar platelet studies in Vwf/p.V1316M–knock-in mice,²⁶  such as mice that express physiologic levels (≈100%) of VWF/p.V1316M. PKC substrate phosphorylation was also significantly increased in resting platelets from the knock-in mice compared with controls (supplemental Figure 5). Importantly, similar to our observations with murine platelets, PKC substrate phosphorylation was increased approximately twofold (1.6-, 2.3-, and 1.8-fold for P1, P2, and P3, respectively) in circulating platelets from 3 patients carrying the VWF/p.V1316M mutation (Figure 5C-D; supplemental Figure 6), despite their low plasmatic VWF antigen levels (supplemental Table 2).

PKC activation provides a powerful signal for the proteolytic cleavage of cell surface receptors,^17,19  including GPIbα on platelets.³⁷  Given the markedly increased baseline levels of PKC substrate phosphorylation observed in VWD2B(p.V1316M) platelets (Figure 5), we evaluated platelet surface expression of GPIbα and GPIX by flow cytometry, before and after expression of the mutant VWF/p.V1316M in mice. Although GPIbα expression was almost completely abolished following expression of VWF/p.V1316M (Figure 6A), no significant difference in expression was observed for GPIX (Figure 6B). Reduced expression of GPIbα was confirmed with a second antibody recognizing a different epitope on its extracellular domain (data not shown).

The main platelet sheddase for GPIbα is ADAM17.¹⁶  We next expressed VWF/p.V1316M in ADAM17^fl/flPF4-Cre⁺ mice. Similar to what we observed in WT-VWF/p.V1316M mice (Figure 1), the PPC was reduced by 84% compared with controls (Figure 6C), platelet size was increased approximately twofold (Figure 6D), and platelet aggregates were found in the blood smears of ADAM17^fl/flPF4-Cre⁺ mice expressing VWF/p.V1316M (Figure 6E). Unexpectedly, GPIbα surface expression was not retained in platelets from ADAM17^fl/flPF4-Cre⁺-VWF/p.V1316M mice (Figure 6F), suggesting that ADAM17 is not critical for GPIbα shedding under these conditions. Consistent with this conclusion, plasma levels of glycocalicin were significantly increased in ADAM17^fl/flPF4-Cre⁺-VWF/p.V1316M and WT-VWF/p.V1316M mice compared with control mice expressing only VWF/WT (Figure 6G).

We next evaluated the impact of these platelet function defects on hemostasis using a recently developed model of laser ablation injury to the saphenous vein.²⁹  In this model, repeated laser ablations to the same site can be performed to test the hemostatic response to injuries of different severity. Compared with controls, platelet adhesion was significantly reduced at sites of laser damage in WT-VWF/p.V1316M mice (Figure 7A-B). Consistent with the platelet-adhesion defect, WT-VWF/p.V1316M mice also exhibited markedly prolonged blood loss from the injury site (Figure 7C). We next determined whether the low platelet count was responsible for the hemostasis defect observed in WT-VWF/p.V1316M mice. Using our recently established method for adoptive platelet transfer,^38,39  we generated mice with a peripheral count <1 × 10⁸ WT platelets per mL of blood, comparable to the PPC of WT-VWF/p.V1316M mice. In agreement with a recent study,⁴⁰  a platelet count of ∼10% of normal was sufficient to prevent prolonged bleeding in our model (Figure 7D). In summary, these studies confirm that defects other than the low platelet count contribute to the hemostasis defect in p.V1316M-associated VWD type 2B.

Our study is the first to provide genetic evidence that the macrothrombocytopenia and platelet aggregates observed in p.V1316M-associated VWD type 2B are due to the direct interaction of mutant VWF/p.V1316M and GPIbα. We did not observe evidence for a role of integrin activation downstream of engagement of GPIbα in these processes. However, we did detect a marked upregulation of PKC activity in platelets from mice and 3 patients expressing VWF/p.V1316M. This latter finding is consistent with previous work showing increased PKC activity in platelets stimulated with VWF/ristocetin⁴¹  and more recent work showing limited granule secretion, a process controlled by PKC signaling, immediately after unfolding of the mechanosensory domain of GPIbα.⁴²  Importantly, in the studies by Deng et al,⁴²  limited granule secretion was also observed in platelets incubated with plasma obtained from a patient carrying the VWF/p.V1316M mutation. Thus, the binding of mutant VWF/p.V1316M to GPIbα induces the activation of PKC in mice and in humans. Consistent with this alteration in PKC signaling, VWD (p.V1316M) type 2B platelets exhibit a desensitization of PKC-mediated granule release and integrin activation and proteolytic degradation of GPIbα (at least in mice). Together with the marked thrombocytopenia and changes in VWF multimer composition, these platelet defects lead to marked bleeding complications in VWD (p.V1316M) type 2B.

Most of our studies were performed in mice expressing the VWF/p.V1316M variant by hydrodynamic gene transfer.^1,4,27,28  This model is very powerful, because it allows for the very rapid induction of a severe VWD type 2B phenotype in Vwf^−/− mice, as previously reported, and in WT or genetically modified mice, such as mice with endogenous WT/VWF, as demonstrated here. However, it is important to remember that hydrodynamic gene transfer leads to overexpression of VWF in plasma^26,27  and that mutant VWF generated following hydrodynamic gene transfer is mainly produced in hepatocytes, whereas endogenous VWF/WT is produced in endothelial cells and megakaryocytes. As a result, no heteromultimers composed of mutant and WT monomers are expected to circulate in plasma of these mice. In contrast, heteromultimers would be expected for patients and knock-in mice heterozygous for the VWD type 2B mutation. These differences may explain, at least in part, why the observed platelet phenotype in this murine model of VWD (p.V1316M) type 2B was stronger than that reported for platelets from patients carrying the same mutation. For example, shedding of GPIbα, induced by the VWF–GPIb interaction, is a striking observation in mice expressing VWF/p.V1316M due to hydrodynamic gene transfer. In contrast, shedding of GPIbα is rarely described for VWD type 2B patients,⁴³  even though there is strong evidence that human platelet GPIbα undergoes increased proteolysis in situations of increased platelet turnover.^14,44  Unfortunately, large studies to determine the clinical relevance of GPIbα shedding in VWD type 2B patients are missing. Compared with mice expressing VWF/p.V1316M due to hydrodynamic gene transfer,²⁷  plasma VWF levels are markedly lower in type 2B patients.^2,26  In our studies, we observed an inverse correlation between mutant VWF/p.V1316M levels in plasma and GPIbα surface expression in WT-VWF/p.V1316M mice (supplemental Figure 7). Furthermore, GPIbα levels were only reduced approximately twofold in platelets from knock-in mice carrying the VWF/p.V1316M mutation (data not shown) (ie, mice in which plasma VWF/p.V1316M levels are lower than in our hydrodynamic gene-transfer model).²⁶ 

Given the well-documented effect of PKC signaling on sheddase activity,^16,17,19  it is likely that GPIbα shedding is a result of the above-described PKC activation in VWD2B(p.V1316M) platelets. Surprisingly, VWF/p.V1316M-induced cleavage of GPIbα was not inhibited in mice deficient in ADAM17, an enzyme that we previously identified as the main sheddase for GPIbα in the circulation of healthy mice and immediately following PKC stimulation in vitro.¹⁶  Our findings do not exclude the possibility that VWF/p.V1316M–GPIbα engagement induces ADAM17 activity. However, a long-term VWF/p.V1316M–GPIbα engagement, as seen after hydrodynamic gene transfer, may activate GPIbα sheddases other than ADAM17, such as ADAM10¹⁷  or the cysteine protease calpain.⁴⁵  Similar ADAM17-dependent and -independent shedding events were shown for other receptors, such as CD44 on leukocytes.²⁰  Of note, defective calpain activity has been reported in platelets obtained from patients affected by Montreal platelet syndrome⁴⁶  (subsequently identified as p.V1316M-associated VWD type 2B⁴⁷ ), but it was never studied again in patients carrying the aforementioned or other type 2B mutations. Future studies should be directed toward a better understanding of the molecular mechanisms by which engagement of GPIbα in VWD type 2B leads to altered PKC signaling in platelets. Information derived from these studies could then be used to develop strategies to prevent the thrombopathy and, thus, to improve hemostasis in these patients.

Although alterations in PKC signaling and GPIbα surface expression are less pronounced in platelets from VWD type 2B patients carrying the VWF/p.V1316M mutation, there is good phenotypic similarity between the patients and the mouse model with regard to impaired activation of Rap1B and αIIbβ3 integrin in response to agonist stimulation.⁴  Our studies strongly suggest that the intracellular signaling defect also occurs at the level of PKC activity regulation. In mice, compared with the sixfold increase in PKC substrate phosphorylation observed in activated control platelets, only a twofold increase in PKC activity over baseline was observed in activated VWF/p.V1316M platelets. Consistent with this defect, platelets from VWF/p.V1316M mice exhibited markedly impaired granule secretion in response to agonist stimulation, similar to what we described for human platelets.⁴  Considering that integrin activation was only mildly impaired in VWF/p.V1316M platelets stimulated with the second-wave mediators, thromboxane A₂ and ADP, and only the second phase of integrin activation was altered in the mutant platelets, it can be concluded that the defect in Rap1B signaling in VWD (p.V1316M) type 2B is the result of impaired ADP secretion and feedback signaling via P2Y12. Consistent with these results, Akt phosphorylation was unaffected in resting conditions and markedly reduced in PAR4p-stimulated platelets from mice expressing the mutant VWF/p.V1316M (supplemental Figure 8). It is unclear whether the defective PKC signaling response to agonist stimulation in VWF/p.V1316M platelets reflects a desensitization of this pathway or whether engagement of GPIbα by mutant VWF/p.V1316M specifically activates PKC isoforms that inhibit, rather than activate, platelets. The latter was demonstrated for collagen-activated platelets, in which activation of PKC(δ) downstream of the collagen receptor antagonizes signaling by classical PKC isoforms required for granule secretion and platelet adhesion.⁴⁸  Irrespective of the underlying mechanism, however, we provide the first evidence that an inherited alteration in a plasma protein can lead to an acquired platelet function disorder due to a defect in PKC/P2Y12/Rap1 signaling. It will be interesting to see whether similar defects in Rap1 signaling can be detected in VWD patients carrying mutations other than p.V1316M and in patients affected by other diseases, especially those that are caused by nonphysiological engagement of GPIbα, as seen in immune thrombocytopenia purpura or thrombotic thrombocytopenia purpura.

It is reasonable to speculate that the observed downmodulation of platelet reactivity toward mutant VWF/p.V1316M is intended to limit thrombosis. Another reason could be the prevention of near-complete thrombocytopenia, a condition that, in the mouse model, leads to increased embryonic lethality and marked blood–lymphatic mixing.^11,38  Interestingly, the molecular mechanisms by which platelets secure vascular integrity during development differ significantly from those established for classical hemostatic plug formation.³⁸  Adhesion receptors, such as GPIb-V-IX or αIIbβ3 integrin, and signaling pathways, such as the Rap1 GTPase pathway, are critical for plug formation but not vascular integrity during development. Thus, downmodulation of GPIbα expression and PKC/P2Y12/Rap1 signaling would provide an elegant mechanism to increase the peripheral platelet count to levels required for vascular integrity in development.

In summary, we applied the hydrodynamic gene-transfer technique in mice with specific platelet signaling defects to demonstrate that the thrombopathy in VWD (p.V1316M) type 2B is due to impaired PKC/P2Y12/Rap1 signaling, a pathway critical for sustained integrin activation in platelets. We show that engagement of GPIbα by mutant VWF/p.V1316M leads to increased PKC activity in murine platelets, which affects cellular function in at least 2 ways: (1) the increased baseline activation makes platelets refractory to agonist-induced PKC signaling, leading to impaired granule release and reduced Rap1-mediated integrin activation and (2) shedding of the GPIbα receptor, which limits the interaction with mutant VWF/p.V1316M. Both alterations protect platelets from incorporation into intravascular thrombi and clearance due to the interaction with mutant VWF/p.V1316M. Of particular relevance, disturbed PKC signaling was also observed in 3 patients affected by VWD (p.V1316M) type 2B, providing evidence that a PKC-dependent hypofunction may contribute to the severe bleeding phenotype of these patients. Indeed, our data indicate that VWD type 2B patients carrying the VWF/p.V1316M variant suffer from a congenital VWD and an acquired platelet defect, suggesting that treatments should address the 2 deficiencies. Replacement therapy, the mainstay of VWD treatment, does not correct platelet counts, and it is unlikely that it can impact platelet functions. Thus, platelet transfusions should be considered as adjunctive treatment in VWD type 2B patients presenting with platelet defects.^49,50 

The authors thank Jerry Ware (University of Arkansas), Dougald M. Monroe III (University of North Carolina at Chapel Hill), Brian Petrich (Emory University), and Pamela B. Conley (Portola Pharmaceuticals Inc.) for providing mice deficient in extracellular GPIbα, VWF, talin1, and P2Y12, respectively. The authors also thank Delphine Borgel (INSERM, UMR-S1176, Université Paris-Sud, Université Paris-Saclay, Le Kremlin-Bicêtre, France and Assistance Publique–Hôpitaux de Paris, Laboratoire d’Hématologie, Hôpital Universitaire Necker-Enfants Malades) and Jenny Goudemand (Département d'Hématologie et de Transfusion, Centre Hospitalier Universitaire de Lille, Université Lille 2, EA 2693, Faculté de Médecine, Lille, France, and Institut Pasteur de Lille, European Genomic Institute for Diabetes, INSERM UMR 1011, Université Lille 2, Lille, France) for valuable help with enrolling patients and providing clinical laboratory information.

This work was supported by a Research Fellowship award granted by the European Hematology Association (C.C.) and by National Institutes of Health, National Heart, Lung, and Blood Institute grants R01 HL121650 and R01 HL130404 (W.B.).

Contribution: C.C., O.D.C., P.J.L., C.V.D., and W.B. designed the study; C.C. and D.S.P. performed most of the experiments and analyzed data; S.S., A.H., and C.L.-.B. enrolled the VWD type 2B patients; R.P. performed and analyzed laser-injury experiments; R.H.L. assisted with experiments; K.O.P. was responsible for mouse colony maintenance and genotyping; M.B. helped to prepare human platelet samples; C.C. and W.B. wrote the manuscript; and authors contributed to the editing of the final manuscript.

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

Correspondence: Caterina Casari, INSERM UMR-S1176, 80 rue du General Leclerc, 94276 Le Kremlin-Bicêtre, France; e-mail: caterina.casari@inserm.fr; and Wolfgang Bergmeier, University of North Carolina, 120 Mason Farm Rd, Campus Box 7260, Chapel Hill, NC 27599; e-mail: bergmeie@e-mail.unc.edu.