Bernhard Nieswandt, Chair of Experimental Biomedicine, Rudolf Virchow Center, University of Würzburg, Zinklesweg 10, 97078 Würzburg, Germany. Tel.: +49 931 201 44060; fax: +49 931 201 44068. E-mail: firstname.lastname@example.org
Summary. Heterodimeric receptors of the β1 and β3 integrin families mediate platelet adhesion and aggregation in hemostasis and thrombosis. In resting platelets, integrins are expressed in a low-affinity state but they shift to a high-affinity state and efficiently bind their ligands in response to cellular activation. This review summarizes recent advances in understanding the functional regulation and (patho-) physiological significance of individual platelet integrins with a special focus on studies in genetically modified mice. It is now recognized that β1 and β3 integrins have partially redundant roles in the adhesion process and that their activation is regulated by similar mechanisms, involving Ca2+-dependent and -independent signaling events and essential functions of talin-1 and kindlin-3 in the terminal activation step.
Platelet adhesion and aggregation on the exposed extracellular matrix (ECM) requires the coordinated interaction of different platelet surface receptors with adhesive macromolecules. Under conditions of high shear, the initial recruitment of platelets to the ‘reactive surface’ is mediated by the reversible interaction between the platelet receptor glycoprotein (GP)Ib and collagen-bound von Willebrand factor (vWF). This interaction allows the cells to establish contacts with collagen through the immunoglobulin-like GPVI, which triggers intracellular signals that lead to cellular activation and the release of the secondary mediators adenosine diphosphate (ADP) and thromboxane A2 (TxA2). These agonists together with locally produced thrombin contribute to cellular activation by stimulating receptors that couple to heterotrimeric G proteins, which induce different signaling events and act synergistically to induce full platelet activation . The principal cellular response to these signaling events is the functional upregulation of integrin family adhesion receptors enabling the platelet to establish firm adhesion contacts followed by thrombus growth [1,2]. Integrins are a ubiquitously expressed family of heterodimeric transmembrane receptors that connect the ECM to the cytoskeleton and serve as bidirectional signaling molecules . In this brief review, we will focus mainly on the mechanisms of ‘inside-out’ integrin activation and not discuss ‘outside-in’ signal transduction.
Platelets express integrins of the β1 and the β3 family that are present on the membrane in a low-affinity state and shift to a high-affinity state during platelet activation to bind their ligands efficiently. Three different β1-integrins are found in platelets, namely α2β1 (collagen receptor), α5β1 (fibronectin receptor), and α6β1 (laminin receptor). Among them, α2β1 has been most intensively studied and, in contrast to a previous hypothesis, it is now recognized that it undergoes activation-dependent conformational changes to bind efficiently to collagen [3,4]. This process has been recently proposed to depend strictly on the activation of αIIbβ3  but the underlying mechanisms remain to be determined. Integrin α2β1 plays a significant, but not essential role for adhesion to collagen in vitro  and to the injured vessel wall in vivo . Similarly, the other two β1 integrins play a supportive but non-essential role for platelet attachment and thrombus formation at the exposed ECM as the lack of all β1 integrins does not have a major effect on bleeding times or the ability to form occlusive thrombi in vivo .
The dominant integrin on the platelet surface, αIIbβ3, mediates platelet aggregation through binding of plasma fibrinogen and serves as the principal receptor for platelet adhesion to the ECM in vivo . The mandatory role of αIIbβ3 for hemostasis and thrombosis is well documented in humans and in experimental animals . The shift of integrin αIIbβ3 from a low- to a high-affinity state is considered the ‘final common pathway’ of platelet activation. Therefore, the molecular machinery regulating this process represents a potential target for antithrombotic agents and has been intensively studied for more than two decades, mostly in heterologous cells or genetically modified mice.
Cellular control of integrin activation requires transmission of a signal from the small cytoplasmic tails to the large extracellular domains . To date, more than 20 proteins have been identified that interact with either one or both cytoplasmic tails participating in inside-out and outside-in signaling through the integrin, but only for a few, the functional significance is understood (Fig. 1).
The cytoplasmic tails of integrin β subunits are evolutionary conserved and contain NPXY motifs serving as recognition sequences for phosphotyrosine binding (PTB) proteins, for example, talin-1. Talin-1, a 270-kDa cytoplasmic protein, binds to β cytoplasmic tails of integrins thereby connecting the integrin with the actin cytoskeleton through its actin binding site. The head region of talin-1 contains a FERM domain comprising three subdomains namely F1, F2, and F3. The F3 subdomain contains a PTB domain, which interacts with the conserved NPXY motif of β tails . Talin-1 binding to both cytoplasmic tails (αIIb and β3 tail) is proposed to induce αIIbβ3 activation by disrupting the salt bridge between the two tails, the crucial final step in the activation of several classes of integrins .
Studies on talin-1-depleted murine embryonic stem cell-derived megakaryocytes provided the first direct evidence that talin is required for integrin αIIbβ3 activation in response to different agonists . This concept was further supported by the demonstration that mice expressing the L746A mutation of β3 integrin, which is believed to result in abolished talin binding, display impaired inside-out activation of αIIbβ3 . Definitive evidence for an essential function of talin-1 in platelet integrin activation was shortly after provided by the analysis of platelets from conditional talin-1-deficient mice (Tln−/−, Cre-loxP) [12,13]. These platelets were unable to activate integrin αIIbβ3 in response to any tested agonist and consequently did not adhere to αIIbβ3 ligands or aggregate. Importantly, lack of talin-1 also resulted in defective adhesion to collagen, implying that α2β1 also was unable to bind its ligand. This was confirmed by flow cytometry using an antibody that recognizes an activation-dependent epitope on the β1 integrin subunit (9EG7), suggesting that talin-1 is required for activation-dependent conformational changes in all platelet β1 integrins . Together, these results indicated that the inside-out activation of β1 and β3 integrins in platelets are mediated by similar pathways and strongly supported the concept that also the binding of α2β1 to its ligand collagen requires prior inside-out activation of the receptor [3,4]. In contrast to previous reports, talin-1-deficient platelets fail to spread on immobilized fibrinogen in the presence and absence of Mn2+, suggesting that talin-1 is also required for αIIbß3-dependent outside-in signaling . This complete lack of integrin function in Tln−/− platelets resulted in defective hemostasis and abrogated platelet adhesion, and thrombus formation in injured vessels in vivo [12,13].
The above-described findings impressively confirmed that talin-1 is essential for platelet integrin activation. However, the long-standing concept that its binding to β cytoplasmic tails is also sufficient for this process to occur was challenged by the identification of a hitherto unknown protein, the FERM-domain containing kindlin-3, as an essential component in the integrin activation machinery in platelets  (Fig. 1). Kindlin-3 belongs to the kindlin family, which consists of three mammalian isoforms (Kindlin-1, -2, -3), all of which are highly concentrated at sites of cell-ECM adhesion. Binding studies with recombinant protein and mutational approaches showed that kindlins directly interact with the β1- and β3-integrin cytoplasmic tails via their F3 subdomains with the same affinity as talin-1 .
The germ line knockout of kindlin-3 results in perinatal lethality that is associated with anemia and diffuse hemorrhages in different regions of the body. Platelets from fetal liver cell chimeric kindlin-3−/− mice were unable to activate integrin αIIbβ3 in response to any tested agonist, despite unaltered expression of talin-1 . Activation of β1 integrins was also abrogated in the absence of kindlin-3 as revealed by defective adhesion to collagen and flow cytometric analysis using the 9EG7 antibody. Similar to Tln−/− platelets, kindlin-3−/− platelets did not spread on fibrinogen, suggesting that both talin-1 and kindlin-3 may cooperate also in outside-in signaling processes. This complete loss of integrin function resulted in a severe hemostatic defect and resistance to arterial thrombosis in kindlin-3−/− chimeric mice.
At present it remains unclear how talin-1 and kindlin-3 cooperate in integrin activation, but Moser et al.  proposed a mechanism involving direct interaction of the PTB site of the F3 domain in kindlin-3 with the NPXY motif of integrin β-tails. The binding region in integrin β-tails for kindlin-3 appears to be distinct from those of talin-1, allowing either simultaneous or sequential binding of the two proteins to activate the integrin. Further studies will be required to address this fundamental question.
During platelet activation, a variety of different agonists and downstream signaling pathways act in concert and cumulate in talin-1/kindlin-3-dependent integrin activation. These processes are not fully understood but some important molecular checkpoints in this cascade have been identified in recent years (Fig. 1).
Rap1b, a small GTP binding protein of the Ras family, is a critical regulator of integrin activation downstream of second messengers thereby regulating integrins of the β1, β2, and β3 families . Deficiency of Rap1b in platelets leads to defective αIIbβ3 activation, prolonged bleeding times, and protection against arterial thrombosis . In platelets, activation of Rap1b is controlled by Ca2+ and diacylglycerol-regulated guanine-nucleotide-exchange factor I (CalDAG-GEFI) as revealed by the analysis of mice lacking the protein. CalDAG-GEFI deficiency resulted in impaired platelet aggregation responses to ADP or TxA2ex vivo and prolonged bleeding times and protection from arterial thrombosis in vivo. Interestingly, however, responses to other agonists, most notably thrombin and collagen, were only mildly affected demonstrating the existence of a CalDAG-GEFI-independent mechanism of αIIbβ3 activation . The activation of protein kinase C (PKC) represents an independent pathway leading to Rap1 and αIIbβ3 activation in murine platelets. It has been proposed that rapid but reversible activation of Rap1 is mediated by CalDAG-GEFI, whereas sustained Rap1 activation is a consequence of PKC activation . Among the different PKC isoforms expressed in platelets, PKCα appears to play a dominant role here as murine platelets lacking this enzyme display impaired αIIbβ3 activation in response to a collagen-related peptide and thrombin .
Agonist-induced elevations in cytosolic Ca2+ concentrations occur through transient mobilization of limited Ca2+ amounts from the endoplasmic/sarcoplasmic reticulum (ER/SR) followed by sustained store-operated Ca2+ entry (SOCE) through the plasma membrane . Recently, stromal interaction molecule 1 (STIM1) was identified as the ER/SR-resident Ca2+ sensor that detects declines in ER/SR Ca2+ concentrations and allows SOCE by opening the plasma membrane Ca2+ channel Orai1 (Fig. 1). Lack of either STIM1 or Orai1 virtually abrogates agonist-induced Ca2+ entry, while the mobilization of Ca2+ from the ER/SR is only moderately or not affected, respectively [21,22]. Despite the dramatically reduced Ca2+ response, Stim1−/− and Orai1−/− platelets display largely intact agonist-induced integrin activation demonstrating that this process can be initiated by relatively small rises in cytosolic Ca2+ concentrations. Under flow conditions in vitro and in vivo, however, lack of SOCE results in the formation of unstable thrombi indicating that sustained activation of αIIbβ3 (through CalDAG-GEFI) may require SOCE.
RIAM and other MRL proteins
The Rap1 effector molecule RIAM (Rap1-GTP-interacting adaptor molecule) belongs to the MRL family of adaptor molecules that include Mig-10, RIAM, and lamellipodin (MRL). MRL proteins interact with both Rap1-GTP and talin-1 by direct binding via short amino-terminal sequences (Fig. 1). This interaction mediates Rap1-induced adhesion of β1 and β2 integrins by providing a scaffold to connect the membrane targeting sequences in Ras GTPases to talin-1. Recent evidence indicates that also lamellipodin contains a talin-1 binding sequence, which allows talin-1 recruitment to the membrane and subsequent integrin activation . Knockdown of RIAM blocks talin-1 recruitment to αIIbβ3 and integrin activation , whereas its overexpression induced integrin activation and enhanced cell adhesion . Together, these results implicate Rap1-induced formation of an ‘integrin-activation complex’ consisting of the effector molecule RIAM and talin-1 that leads to the unmasking of the integrin-binding site on talin-1 . Further studies will be required to assess the in vivo relevance of RIAM and other MRL proteins in platelet integrin activation.
Taken together, major advances have been made during the last years in understanding the functional regulation and (patho-) physiological significance of individual platelet integrins, many of them by the analysis of genetically modified mice. From these studies, it is becoming increasingly clear that the regulation of integrin function is a highly integrated process that involves the concerted action of multiple proteins and it can be anticipated that additional players will be identified in the future. Proteins like the integrin αIIb binding chloride channel regulatory protein ICln or CIB (a calcium and integrin binding protein) appear to play important roles here, but their exact role has not been definitively identified by in vivo studies in platelets. Another remaining important question is whether or not different platelet integrins are regulated by different signaling pathways. Although the majority of studies in the past focused on the functional regulation of αIIbβ3, multiple lines of evidence now indicate that β1 integrins in platelets are regulated by very similar or even identical mechanisms with essential roles for talin-1 and kindlin-3 in the terminal activation steps [12,14]. Nevertheless, it will be important to further dissect the signaling pathways that control the function of individual integrins as the specific targeting of certain adhesive mechanisms in platelets, while leaving others intact, might open new avenues towards safer antithrombotic protection.
The work of the authors has been supported by the Deutsche Forschungsgemeinschaft (SFB 688).
Disclosure of Conflict of Interests
The authors state that they have no conflicts of interest.