GPVI and the not so eager cleaver

Platelets control their responsiveness, in part, by shedding adhesion and signaling receptors from their surface. The molecular mechanism by which this occurs, however, is incompletely understood. In this issue of Blood, Bender and colleagues make judicious use of mice genetically deficient in selected candidate proteases to shed new light on the unexpected complexity of ectodomain shedding.¹ 

Hematopoietic cells control their adhesive phenotype through several mechanisms, including de novo synthesis of adhesion receptors (eg, VCAM-1, ICAM-1), compartmentalization within the cell (eg, P-selectin, LAMP-3), conformational change (eg, integrins), receptor clustering (eg, growth factor and cytokine receptors), and proteolytic removal of the extracellular domain of transmembrane adhesion and signaling receptors (eg, L-Selectin, TNF-α). The latter, commonly referred to as ectodomain shedding, has been shown to regulate such diverse processes as cell adhesion and migration, development, cellular signaling, and apoptosis.² 

Ectodomain shedding plays an important role in platelet biology as well. An array of adhesion and signaling receptors reside on the surface of platelets that, upon exposure to their ligands, initiate the activation of a complex network of signaling pathways leading to platelet activation, adhesion, and thrombus formation. Several of these receptors are known to be removed from the plasma membrane after platelet activation as part of a naturally occurring negative feedback loop intended to limit platelet accumulation and thrombus growth. The first example of a shed platelet plasma membrane–derived adhesion receptor was provided by the laboratory of the recently deceased Graham Jamieson, who published a series of studies in the 1970s characterizing glycocalicin³ : a soluble, circulating plasma protein that was later found to be derived from glycoprotein (GP) Ibα, the major subunit of the platelet receptor for von Willebrand factor. The importance of ectodomain shedding in passivating platelet responsiveness has been extended by findings that the extracellular domains of GPV and GPVI are also rapidly lost from the platelet surface after various, seemingly unrelated, forms of cellular activation.

GPVI is a 62-kDa platelet-specific type I transmembrane glycoprotein expressed on the surface of human and murine platelets in a noncovalent complex with the immunoreceptor tyrosine-based activation motif (ITAM)–containing subunit, the FcRγ chain.^4,5  GPVI is composed of 2 extracellular immunoglobulin (Ig)–homology domains, a transmembrane domain, and a 51-amino acid cytoplasmic domain,^6,7  and serves as the major platelet-activating receptor for collagen, signaling via the Syk/SLP-76/PLCγ2 pathway to activate the integrins α₂β₁ and α_IIbβ₃ leading to platelet activation and thrombus formation.^8,9  Studies performed by Nieswandt and colleagues nearly a decade ago revealed the surprising finding that injection into mice of rat anti–mouse GPVI monoclonal antibodies results in loss of GPVI from the surface of circulating murine platelets,¹⁰  with a corresponding reduction in platelet responsiveness to collagen.¹¹  Antibody-induced shedding (sometimes termed immunodepletion) of GPVI appears to be operable in humans as well, as platelets from a patient with a circulating autoantibody specific for GPVI also become devoid of cell-surface GPVI and fail to form thrombi over collagen-coated surfaces, while remaining responsive to other platelet agonists.¹²  Taken together with the observation that GPVI can be down-regulated in human platelets in vivo using monoclonal antibody therapy,¹³  there is growing interest in the development of GPVI-based therapeutics, in part because the use of more broadly acting antiplatelet agents like ADP receptor or antifibrinogen receptor antagonists is associated with a low, but still problematic, incidence of bleeding. Understanding how potential anti-GPVI–based agents lead to removal of this adhesion and signaling receptor from the platelet surface, and to corresponding platelet passivation, is therefore of paramount importance before widespread adoption of this novel therapeutic modality can be considered.

Ectodomain shedding of transmembrane receptors is largely carried out by members of 2 large families of zinc-dependent metalloproteinases: MMPs (matrix metalloproteinases) and ADAMs (metalloproteinases containing a disintegrin and metalloprotease domain). ADAMs are type I transmembrane glycoproteins, and are thought to cleave their substrates in cis: that is, they eat their neighbors (see figure)! There are 21 distinct genes in humans that encode ADAM proteases, though only 12 (ADAMs 8, 9, 10, 12, 15, 17, 19, 20, 21, 28, 30, and 33) are thought to have actual protease activity.²  Of these, there is compelling evidence that ADAM17 (also known as tumor necrosis factor α converting enzyme, or TACE) is responsible for the cleavage and release of the extracellular domains of GPIbα¹⁴  and GPV.¹⁵  The identity of the ADAM that sheds GPVI from the platelet surface, however, is less clear.

The extracellular domain of GPVI has been shown to be shed from the platelet surface in response to a wide variety of physiologic and nonphysiologic agents. These include binding of GPVI-specific antibodies,^10,12,13,16  binding of GPVI-specific ligands such as collagen, collagen-related peptide (CRP), and convulxin,^16,17  engagement of the ITAM-linked Fc receptor, FcγRIIa,¹⁸  dissociation of calmodulin from the cytoplasmic domain of GPVI induced by the calmodulin inhibitor, W7,¹⁷  and treatment of platelets with phorbol 12-myristate-13-acetate, N-ethylmaleimide (NEM), or the mitochondrial poison, carbonyl cyanide m-chlorophenyl hydrazone (CCCP).^19,20  Regardless of the type of stimulation, GPVI shedding can in all cases be blocked by addition of ethylenediaminetetraacetic acid (EDTA) or the broad-range metalloproteinase inhibitor, GM6001, demonstrating that removal of the GPVI ectodomain is a cation-dependent MMP- or an ADAM-mediated event. Although no direct evidence exists, each of these stimuli presumably induce minor conformational changes in GPVI that expose cryptic sequences on the receptor that make it vulnerable to proteolytic cleavage. This process appears to require subthreshold levels of cellular activation, as inhibition anywhere along the GPVI/FcRγ-chain ITAM→Src→Syk→LAT→PLCγ2 signal transduction circuit blocks GPVI shedding.^17,21 

In this issue, Bender and colleagues¹  use newly generated megakaryocyte-specific ADAM10-deficient mice, mice lacking ADAM17 protease activity, and mice whose platelets lack both ADAM10 and ADAM17 functional activity to examine their relative roles in mediating GPVI ectodomain shedding. The authors make the fascinating and unexpected finding that the ADAM used very much depends on the way that the platelets are stimulated. Thus, although ADAM10 functions as the predominant GPVI sheddase when platelets are stimulated in vitro with the calmodulin inhibitor W7, ADAM17 serves as the predominant GPVI sheddase when platelets are activated, also in vitro, using CCCP (see figure). Perhaps most relevant to the contemplated development of large or small molecule therapeutic agents that target GPVI directly, neither ADAM10 nor ADAM17 appears to function as GPVI sheddases when anti-GPVI antibodies are used to down-regulate the receptor in vivo. The identity of the protease that becomes activated in response to antibody-induced shedding is unknown.

What might account for the differences seen in GPVI ectodomain shedding induced by various agonists? It is possible that chemical agents like NEM, W7, and CCCP act indirectly to activate ADAM proteases, perhaps in addition to their effects on GPVI itself or its downstream signaling components. This would explain why agents, like antibodies, that bind GPVI directly, are unable to initiate ADAM10- or ADAM17-mediated cleavage. Solving the interesting and important puzzle of ADAM10 and ADAM17 activation, however, will not address the mechanism by which molecules that bind GPVI directly induce shedding in vivo. Because targeting the GPVI/FcRγ-chain complex represents a promising therapeutic approach—not only for its potential to selectively limit platelet reactivity to collagen while preserving other platelet-activation pathways,¹³  but also for its potential to dampen inflammatory disease²² —the intriguing observations of Bender et al that GPVI is shed in vivo in an ADAM10/ADAM17-independent manner warrant further investigation. Identification of the protease responsible for mediating in vivo cleavage of this unique platelet adhesion and signaling receptor may very well reveal additional molecular targets that extend beyond GPVI biology for treating thrombotic and inflammatory human disease.