The β3 integrin cytoplasmic tail: protein scaffold and control freak

Authors


Sanford J. Shattil, Division of Hematology–Oncology, Department of Medicine, UCSD, 9500 Gilman Drive, La Jolla, CA 92093-0726, USA.
Tel.: +1 858 822 6425; fax: +1 858 822 6444.
E-mail: sshattil@ucsd.edu

Abstract

Summary.  Platelet integrin αIIbβ3 plays an essential role in thrombus formation through interactions with adhesive ligands. Successful parenteral blockade of these interactions has validated αIIbβ3 as a therapeutic target in cardiovascular medicine. However, oral αIIbβ3 antagonists have not been successful and there is an unmet need for more effective anti-platelet drugs. Growing evidence points to the cytoplasmic tails of αIIb and β3, and the β3 tail in particular, as scaffolds for intracellular proteins that mediate inside-out signaling and regulate αIIbβ3 affinity for ligands. Intracellular protein interactions with the integrin cytoplasmic tails also regulate outside-in signals to the actin cytoskeleton. Here we focus on recent studies that illustrate the relevance of the β3 cytoplasmic tail as a regulatory scaffold in vivo. We speculate that this scaffold or its interacting proteins may serve as therapeutic targets for the development of future anti-thrombotic drugs.

Integrins, like most transmembrane cell surface receptors, interact with web-like intracellular signaling pathways [1]. Such arrangements are inherently complex. Thus, although bidirectional αIIbβ3 signaling in platelets has been subject to intensive investigation, it is only partially understood. Recent studies have begun to chip away at this complexity by identifying intracellular adapters, enzymes and substrates necessary for αIIbβ3 signaling [2]. A subset of these proteins can interact directly or indirectly with the short, enzymatically incompetent, cytoplasmic tails of αIIb or β3 [2–7], leading to the concepts that αIIbβ3 signaling is regulated by such interactions and that the tails function as regulatory scaffolds [3,4,7]. Here we consider emerging evidence for the biological significance of the scaffold function of β3, although the αIIb tail may also function in this regard [8,9]. Inside-out and outside-in αIIbβ3 signaling are often considered separately, but some β3 cytoplasmic tail-interacting proteins may function bidirectionally [10,11].

The β3 cytoplasmic tail consists of ∼47 residues (716K…T762). Early inklings of a regulatory role for the β3 tail derived from patients with platelet dysfunction due to a particularly disruptive point mutation (β3 S752P) [12] or a substantial deletion (β3 Δ724) [13] of the tail. Studies in model cell systems confirmed that such mutations disrupt bidirectional αIIbβ3 signaling, affecting both αIIbβ3 affinity regulation and cytoskeletal organization [13,14]. Starting with Law et al. [15], studies of gene-targeted mice have identified more subtle β3 tail mutations that predominantly affect inside-out or outside-in signaling. They have also shed light on the scaffold and regulatory functions of the β3 cytoplasmic tail based on interactions with talin, kindlin-3 and Src family kinases (SFKs).

Talin resides in the cytoplasm of resting platelets [16], with binding sites for protein partners, such as integrin β tails, vinculin and actin buried in an ‘auto-inhibited’ conformation [17]. These binding sites are exposed during platelet activation, resulting in talin recruitment to the plasma membrane in proximity to αIIbβ3. According to one model, this recruitment enables the F2–3 subdomain within talin’s N-terminal FERM domain to engage a mid-region of the β3 tail centered at 739Trp-Asp-Thr-Ala-Asn-Asn-Pro-Leu-Tyr747 [18]. Then, a second F2–3 interaction with a membrane-proximal region of the β3 tail disrupts a salt bridge between αIIb and β3, causing separation or re-orientation of the αIIb and β3 cytoplasmic and transmembrane domains, and propagation of conformational changes to the extracellular domains [5,19,20].

One mechanism of talin recruitment to membranes and αIIbβ3 involves Rap1, a membrane-tethered Ras family GTPase implicated in integrin-dependent adhesion of many cell types [21]. Studies in model cell systems suggest that when cell agonists activate Rap1-GDP to Rap1-GTP, it enables the recruitment of a Rap1 effector called RIAM to the plasma membrane [22–24]. As RIAM is an adapter that can bind both Rap1-GTP and talin [24], its recruitment would place activated talin in proximity to αIIbβ3. Studies of mouse and human platelets support this model. Mouse platelets deficient in CalDAG-GEF1, a Rap1 guanine nucleotide exchange factor, exhibit impaired, but not absent, fibrinogen binding and aggregation in response to several agonists. The mice show prolonged tail bleeding times, resistance to thrombosis of mesenteric arterioles in a FeCl3 injury model, and defects in leukocyte β2 integrin activation [25]. A similar platelet phenotype is observed in mice deficient in Rap1b [26]. The combined platelet and leukocyte integrin phenotype of CalDAG-GEF1-deficient mice is similar to that in humans with variant leukocyte adhesion deficiency, which has been attributed to reduced CalDAG-GEF1 levels in some patients [27], but not others [28]. RNAi knockdown of RIAM in megakaryocytes inhibits agonist-induced αIIbβ3 activation [24]. Thus, a Rap1/RIAM/talin complex probably mediates inside-out signaling in platelets. Additional studies are needed to ascertain the role of platelet RIAM in hemostasis and to identify potential Rap1-independent pathways of αIIbβ3 activation.

Studies in mice have established a requirement for talin and talin binding to β3 in inside-out αIIbβ3 signaling. Although talin deficiency is embryonic lethal, mice have been generated with a megakaryocyte/platelet-specific deficiency of talin [10,29]. Their platelets show an almost complete block of αIIbβ3 activation, fibrinogen binding and aggregation in response to single agonists, but there is a small residual response to a combination of agonists. Like CalDAG-GEFI- and Rap1b-deficient mice, platelet talin-deficient mice have prolonged tail bleeding times and are resistant to FeCl3-induced carotid artery thrombosis. Furthermore, they exhibit a severe hemorrhagic diathesis manifested by perinatal hemorrhage and high mortality, and survivors have gastrointestinal bleeding and anemia. To understand better the role of talin interaction with β3, knock-in mice were generated that contain a mutation in the β3 cytoplasmc domain (L746A) that selectively abrogates talin interaction [30]. As with talin-deficient platelets, β3 (L746A) platelets exhibit profound but not complete impairment of fibrinogen binding and aggregation, and the mice have prolonged tail bleeding times and are protected from arterial thrombosis. However, unlike talin-deficient mice, β3 (L746A) mice do not have an overt bleeding diathesis, possibly because their talin can interact with other platelet integrins.

The small residual fibrinogen binding and aggregation responses of talin-deficient platelets to combinations of agonists suggest that additional integrin-binding proteins may control αIIbβ3 affinity. One candidate is kindlin-3. Vertebrate kindlins include kindlin-1 (also known as Fermt1), kindlin-2 (Fermt2, Mig-2, URP3) and kindlin-3 (Fermt3, URP2), each containing a split FERM domain, with the N- and C-terminal portions of the domain connected by a PH domain [31,32]. Kindlin binding to the β3 cytoplasmic tail requires tail residues (756Asn-Ile-Thr-Tyr759) membrane distal to talin-binding residues [33,34]. Kindlin-3 is hematopoietic cell-specific and mice deficient in this protein die shortly after birth with diffuse hemorrhages and osteopetrosis [35]. Platelets from kindlin-3-deficient mouse chimeras are defective in agonist-induced αIIbβ3 activation [35], as are platelets from certain humans with LAD-I variant syndrome that lack kindlin-3 [28]. In CHO cells, kindlin-1 and kindlin-2 can function as co-activators of talin-dependent αIIbβ3 activation, an effect abrogated by substitution of β3 Tyr759 with Ala [33,34]. These findings raise many new questions. Among them, what is the precise relationship between talin and kindlin-3 in inside-out signaling? Does kindlin-3 function by regulating talin recruitment and binding to β3 or vice-versa? What are the roles of other β3 tail-binding proteins, such as filamin A, which may compete with talin or kindlin-3 for binding to β3 [36,37]?

The network of proteins involved in outside-in αIIbβ3 signaling is impressive [2]. Many parallels are apparent between this process and signaling triggered by immunoreceptors in platelets and leukocytes, including dependence on SFKs, Syk protein tyrosine kinase and proteins with ITAM-motifs [38–40]. The cytoplasmic tail of β3 may serve a nucleating function for some of these proteins [2]. For example, c-Src and several other SFKs can interact directly with the β3 cytoplasmic tail in vitro and in CHO cells through interactions that require the SFK SH3 domain and the β3 C-terminus (760Arg-Gly-Thr762) [41]. In mouse radiation chimeras, substitution of β3 762Thr with Ala leads to decreased platelet spreading on fibrinogen [42], as does incorporation of an Arg-Gly-Thr synthetic peptide into human platelets [43]. The in vivo consequences of deleting β3 (760Arg-Gly-Thr762) have been evaluated in β3 (Δ760–762) knock-in mice [44]. Their platelets exhibit reduced SFK-mediated tyrosine phosphorylation of β3 and other proteins in response to fibrinogen binding, and reduced spreading after plating on fibrinogen. Inside-out signaling and fibrinogen binding are normal in response to strong agonists, but responses to ADP are diminished. β3 (Δ760–762) mice have variably prolonged tail bleeding times, but no spontaneous hemorrhages, gastrointestinal bleeding or anemia, and they are protected from FeCl3-induced carotid artery thrombosis.

Perspective

The studies discussed here have taught us many of the rules of αIIbβ3 signaling, but they are also exposing our still limited knowledge of this complex process. Evidence from studies of gene-targeted mice and certain humans with platelet dysfunction indicate that the integrin β3 cytoplasmic tail is a regulatory scaffold. While the loss of some scaffolding functions of β3 in mice and humans results in severe hemostatic defects, the loss of others in mice appears more subtle, often with greater effects on thrombosis than hemostasis. One can only speculate what future investigations will uncover. Are there more integrin-regulatory proteins to be discovered? Might there be individuals with platelet dysfunction, perhaps subtle, due to mutations in the αIIb or β3 tails or their binding partners? Could the subtle defects in the scaffold function of β3 observed in some gene-targeted mice serve as a conceptual framework for rationale development of new anti-thrombotic drugs [15,30,44]? Answers to these questions will require a better understanding of αIIbβ3 signaling in human platelets.

Acknowledgments

Work from the author’s laboratory was supported by grants from the National Institutes of Health.

Disclosure of Conflict of Interests

The author states that he has no conflict of interest.

Ancillary