Inhibition of fibrin(ogen) binding to stimulated platelets by a monoclonal antibody specific for a conformational determinant of GPIIIa

An αIIbβ3 monoclonal antibody traps a semiextended conformation and allosterically inhibits large ligand binding

ABSTRACT

Monoclonal antibodies (mAbs) have provided valuable information regarding the structure and function of platelet αIIbβ3. Protein disulfide isomerase (PDI) has been implicated in αIIbβ3 activation and binds to thrombin-activated αIIbβ3. Using human platelets as the immunogen, we identified a new mAb (R21D10) that inhibits the binding of PDI to platelets activated with thrombin receptor-activating peptide (T6). R21D10 also partially inhibited T6-induced fibrinogen and PAC-1 binding to platelets, as well as T6- and adenosine 5'-diphosphate-induced platelet aggregation. Mutual competition experiments showed that R21D10 does not inhibit the binding of mAbs 10E5 (anti-αIIb cap domain) or 7E3 (anti-β3 β-I domain), and immunoblot studies indicated that R21D10 binds to β3. The dissociation of αIIbβ3 by EDTA had a minimal effect on R21D10 binding. Cryogenic electron microscopy of the αIIbβ3-R21D10 Fab complex revealed that R21D10 binds to the β3 integrin-epidermal growth factor 1 (I-EGF1) domain and traps an intermediate conformation of αIIbβ3 with semiextended leg domains. The binding of R21D10 produces a major structural change in the β3 I-EGF2 domain associated with a new interaction between the β3 I-EGF2 and αIIb thigh domains, which may prevent the swing-out motion of the β3 hybrid domain required for high-affinity ligand binding and protect αIIbβ3 from EDTA-induced dissociation. R21D10 partially reversed the ligand binding priming effect of eptifibatide, suggesting that it could convert the swung-out conformation into a semiextended conformation. We concluded that R21D10 inhibits ligand binding to αIIbβ3 via a unique allosteric mechanism, which may or may not be related to its inhibition of PDI binding.

Integrin clustering mechanisms explored with a soluble αIIbβ3 ectodomain construct

The purpose of this study was to test the hypothesis that residues critical for ligand- and temperature-induced clustering of integrin alphaIIbbeta3 are present on its extracellular domain. Sucrose density gradient sedimentation was used to examine the effects of ligand-mimetic peptides and physiological temperature on the oligomeric state of a soluble recombinant ectodomain variant of the alphaIIbbeta3 integrin, alphaIIbDelta962beta3Delta692, and its full-length counterpart. Both the ectodomain construct, isolated from High Five insect cell culture supernatants, and alphaIIbbeta3, isolated from human blood platelets, exhibited similar weight-average sedimentation coefficients at 23 degrees C, in the absence and presence of the ligand-mimetic peptide eptifibatide. These observations indicate that alphaIIbbeta3's ectodomain exhibits a similar extended conformation in both its free and ligand-bound states. Oligomerization was examined by incubation of both alphaIIbDelta962beta3Delta692 and full-length receptors at 37 degrees C, in the presence or absence of ligand-mimetic. Minimal oligomerization was observed with alphaIIbDelta962beta3Delta692. In contrast, full-length alphaIIbbeta3 exhibited substantial temperature-induced increases in its distribution of sedimenting species, indicative of thermal aggregation. These observations suggest that optimum oligomerization requires the participation of the integrin's transmembrane and cytoplasmic regions. In vivo, clustering of ligand-bound integrins may enhance signaling by increasing the local concentration of intracellular integrin-associated proteins.

Binding of a fibrinogen mimetic stabilizes integrin alphaIIbbeta3's open conformation

The platelet integrin alphaIIbbeta3 is representative of a class of heterodimeric receptors that upon activation bind extracellular macromolecular ligands and form signaling clusters. This study examined how occupancy of alphaIIbbeta3's fibrinogen binding site affected the receptor's solution structure and stability. Eptifibatide, an integrin antagonist developed to treat cardiovascular disease, served as a high-affinity, monovalent model ligand with fibrinogen-like selectivity for alphaIIbbeta3. Eptifibatide binding promptly and reversibly perturbed the conformation of the alphaIIbbeta3 complex. Ligand-specific decreases in its diffusion and sedimentation coefficient were observed at near-stoichiometric eptifibatide concentrations, in contrast to the receptor-perturbing effects of RGD ligands that we previously observed only at a 70-fold molar excess. Eptifibatide promoted alphaIIbbeta3 dimerization 10-fold more effectively than less selective RGD ligands, as determined by sedimentation equilibrium. Eptifibatide-bound integrin receptors displayed an ectodomain separation and enhanced assembly of dimers and larger oligomers linked through their stalk regions, as seen by transmission electron microscopy. Ligation with eptifibatide protected alphaIIbbeta3 from SDS-induced subunit dissociation, an effect on electrophoretic mobility not seen with RGD ligands. Despite its distinct cleft, the open conformer resisted guanidine unfolding as effectively as the ligand-free integrin. Thus, we provide the first demonstration that binding a monovalent ligand to alphaIIbbeta3's extracellular fibrinogen-recognition site stabilizes the receptor's open conformation and enhances self-association through its distant transmembrane and/or cytoplasmic domains. By showing how eptifibatide and RGD peptides, ligands with distinct binding sites, each affects alphaIIbbeta3's conformation, our findings provide new mechanistic insights into ligand-linked integrin activation, clustering and signaling.

AlphaIIb's cytoplasmic domain is not required for ligand-induced clustering of integrin alphaIIbbeta3

The platelet integrin alphaIIbbeta3 exhibits bidirectional signaling, in that intracellular messengers enable adhesive macromolecules to bind to its ectodomain, while ligation promotes the association of cytoskeletal proteins with its cytoplasmic domains. In order to understand the linkage between these distant regions, we investigated the effects of receptor occupancy on the solution structure of both full-length recombinant alphaIIbbeta3 and alphaIIbDelta991beta3, an integrin truncation mutant which lacks one cytoplasmic domain. Lysates of (35)S-labeled human A549 cells expressing either full-length alphaIIbbeta3 or alphaIIbDelta991beta3 were examined by sucrose density gradient sedimentation followed by immunoprecipitation to determine the distributions of integrin protomers and oligomers. Recombinant alphaIIbbeta3 exhibited a weight-average sedimentation coefficient, S(w)=11.3+/-1.4 S with 73% sedimenting as protomers/dimers (9.1+/-1.0 S) and 27% as oligomers (15.4+/-0.4 S). Truncation mutant alphaIIbDelta991beta3 exhibited a similar pattern with 65% sedimenting as protomers/dimers. Upon ligation with eptifibatide, both full-length alphaIIbbeta3 and alphaIIbDelta991beta3 sedimented mainly at >14 S, indicating 2-3-fold increased oligomerization. Thus we have demonstrated that alphaIIb's cytoplasmic region is not required for integrin clustering, a key event in outside-in signaling.

A model of platelet aggregation involving multiple interactions of thrombospondin-1, fibrinogen, and GPIIbIIIa receptor

Thrombospondin-1 (TSP) may, after secretion from platelet alpha granules, participate in platelet aggregation, but its mode of action is poorly understood. We evaluated the capacity of TSP to form inter-platelet cross-bridges through its interaction with fibrinogen (Fg), using either Fg-coated beads or Fg bound to the activated GPIIbIIIa integrin (GPIIbIIIa*) immobilized on beads or on activated fixed platelets (AFP), i.e. in a system free of platelet signaling and secretion mechanisms. Aggregation at physiological shear rates (100-2000 s(-1)) was studied in a microcouette device and monitored by flow cytometry. Soluble TSP bound to and induced aggregation of Fg-coated beads dose-dependently, which could be blocked by the amino-terminal heparin-binding domain of TSP, TSP18. Soluble TSP did not bind to GPIIbIIIa*-coated beads or AFP, unless they were preincubated with Fg. The interaction of soluble TSP with Fg-GPIIbIIIa*-coated beads or Fg-AFP resulted in the formation of aggregates via Fg-TSP-Fg cross-bridges, as demonstrated in a system where direct cross-bridges mediated by GPIIbIIIa*-Fg on one particle and free GPIIbIIIa* on a second particle were blocked by the RGD mimetic Ro 44-9883. Soluble TSP increased the efficiency of Fg-mediated aggregation of AFP by 30-110% over all shear rates and GPIIbIIIa* occupancies evaluated. Surprisingly, TSP binding to Fg already bound to its GPIIbIIIa* receptor appears to block the ability of this occupied Fg to recognize another GPIIbIIIa* receptor, but this TSP can indeed cross-bridge to another Fg molecule on a second platelet. Finally, TSP-coated beads could directly coaggregate at shear rates from 100 to 2000 s(-1). Our studies provide a model for the contribution of secreted TSP in reinforcing inter-platelet interactions in flowing blood, through direct Fg-TSP-Fg and TSP-TSP cross-bridges.

Platelet aggregation induced by ADP or epinephrine is enhanced in habitual smokers

A novel system has been developed to evaluate biochemically induced platelet aggregation by means of a particle-counting technique that uses laser light scattering. Using this system, we compared the differences in platelet aggregability between 90 smoking healthy males after 10 hours of smoking abstinence and 141 age-matched nonsmoking healthy males. Smokers had more small spontaneous platelet aggregates and more medium and large aggregates induced by 1 or 5 microM of epinephrine than nonsmokers. No large aggregates with 5-microM epinephrine-induced platelet aggregation were seen in 10% of smokers and 24% of nonsmokers; these subjects showed no small aggregates in spontaneous aggregation at all. Smokers had significantly more small, medium, and large aggregates induced by 1 microM of adenosine diphosphate (ADP) than nonsmokers. Smokers showed a positive correlation between age and 1-microM epinephrine-induced large platelet aggregates, percent reduction of optical density, and 1-microM ADP-induced medium and large aggregates. Smokers also showed a positive correlation between fibrinogen concentration in plasma and small spontaneous aggregates. On the other hand, nonsmokers showed a significant positive correlation between age and small spontaneous aggregates, and a positive correlation between fibrinogen and 1- or 5-microM epinephrine-induced large aggregates, and between 1 microM ADP-induced large aggregates and percent reduction of optical density. These results confirmed that platelet aggregability is enhanced in smokers, and we speculate that long-term smoking might enhance the sensitivity of platelets to epinephrine or ADP.

Time and force dependence of the rupture of glycoprotein IIb-IIIa-fibrinogen bonds between latex spheres

We studied the shear-induced breakup of doublets of aldehyde/sulfate (A/S) latex spheres covalently linked with purified platelet GPIIb-IIIa receptor, and cross-linked by fibrinogen. Flow cytometry with fluorescein isothiocyanate-fibrinogen showed than an average of 22,500 molecules of active GPIIb-IIIa were captured per sphere, with a mean K(d) = 56 nM for fibrinogen binding. The spheres, suspended in buffered 19% Ficoll 400 containing 120 or 240 pM fibrinogen, were subjected to Couette flow in a counter-rotating cone-plate rheoscope. Doublets, formed by two-body collisions at low shear rate (G = 8 s(-1)) for < or =15 min, were subjected to shear stress from 0.6 to 2.9 Nm(-2), their rotations recorded until they broke up or were lost to view. Although breakup was time dependent, occurring mostly in the first 2 rotations after the onset of shear, the percentage of doublets broken up after 10 rotations were almost independent of normal hydrodynamic force, F(n): at 240 pN, 15.6, 16.0, and 17.0% broke up in the force range 70-150 pN, 150-230 pN, and 230-310 pN. Unexpectedly, at both [fibrinogen], the initial rate of breakup was highest in the lowest force range, and computer simulation using a stochastic model of breakup was unable to simulate the time course of breakup. When pre-sheared at low G for >15 min, no doublets broke up within 10 rotations at 70 < F(n) < 310 pN; it required >3 min shear (>1110 rotations) at F(n) = 210 pN for significant breakup to occur. Other published work has shown that binding of fibrinogen to GPIIb-IIIa immobilized on plane surfaces exhibits an initial fast reversible process with relative low affinity succeeded by transformation of GPIIb-IIIa to a stable high-affinity complex. We postulate that most doublet breakups observed within 10 rotations were from a population of young doublets having low numbers of bonds, by dissociation of the initial receptor complex relatively unresponsive to force. The remaining, older doublets with GPIIb-IIIa in the high-affinity complex were not broken up in the time or range of forces studied.

The fibrinogen RIBS-I epitope (gamma373-385) appears proximate to the gamma408-411 adhesive domain but is not involved in interaction between receptor-bound or surface-adsorbed fibrinogen and platelet GPIIbIIIa

The carboxyl terminus of the fibrinogen (Fg) gamma chain (gamma400-411) is necessary and sufficient to support platelet aggregation and adhesion. However, a monoclonal antibody (mAb) to the Fg RIBS-I epitope (gamma373-385), the anti-Fg-RIBS-I, which binds only to platelet-bound or surface-adsorbed Fg but not soluble Fg, inhibits platelet aggregation. In this study, we showed that this same antibody also inhibits the adhesion of platelets to Fg-coated polystyrene beads. We then investigated the mechanisms by which the anti-Fg-RIBS-I antibody inhibits platelet aggregation and adhesion. The Fg RIBS-I epitope does not interact with platelet GPIIbIIIa, since recombinant Fg missing the last four amino acids, the Ala-Gly-Asp-Val, on the carboxyl terminus of its gamma chains supports neither platelet aggregation nor adhesion to surfaces, nor GPIIbIIIa binding, while it binds anti-Fg-RIBS-I normally. Purified, soluble GPIIbIIIa (265 kDa) inhibits the binding of both the anti-Fg-RIBS-I and 4A5 (a mAb specific to gamma408-411 of Fg), however, peptide G13 (1.5 kDa), corresponding to the Fg gamma chain binding domain on GPIIba (GPIIb300-312), only inhibits the binding of 4A5, and does not affect the binding of the anti-Fg-RIBS-I to Fg. The anti-Fg-RIBS-I reduces the on-rate of the 4A5 binding to Fg with no measurable changes in the dissociation of the Fg-bound 4A5. These data indicate that the inhibition of platelet aggregation and adhesion by the anti-Fg-RIBS-I antibody is due to the steric hindrance of the Fg gamma400-411 to platelet GPIIbIIIa. Thus the Fg RIBS-I epitope (gamma373-385) does not appear to be involved in direct interaction with platelet GPIIbIIIa, leaving the gamma408-411 of Fg as the sole domain mediating platelet aggregation and adhesion.

Role of the gamma chain Ala-Gly-Asp-Val and Aalpha chain Arg-Gly-Asp-Ser sites of fibrinogen in coaggregation of platelets and fibrinogen-coated beads

Fibrinogen (Fg) mediates platelet aggregation and adhesion to artificial surfaces. The carboxyl terminus of the gamma chain of Fg (residues AGDV at gamma408-411) is known to play an exclusive role in platelet aggregation, while there is no known role for the consensus RGD sites in the Aalpha chain. In this study, we used flow cytometry to measure the coaggregation (CA) of platelets with Fg-coated beads, and investigated which domains in surface-immobilized Fg support platelet adhesion. CA of platelets with Fg-beads was nearly abolished in the presence of 4A5, a monoclonal antibody (mAb) whose epitope includes AGDV, while Z69/8, a mAb that also binds to the gamma chain carboxyl terminus but does not cover AGDV, had little effect. When beads were coated with recombinant Fg (rFg) lacking AGDV, CA was similarly abolished. In contrast, beads coated with Fg that lacked the RGDS site, supported platelet CA as did intact Fg. These results were confirmed in experiments that measured the binding of activated soluble glycoprotein IIb and IIIa (GPIIbIIIa), the platelet membrane glycoprotein complex known to be the Fg receptor, to immobilized Fg. This binding was inhibited by mAb 4A5, but not by mAb Z69/8. Binding was totally retained when beads were coated with Fg lacking RGDS, but was completely lost when beads were coated with Fg lacking AGDV. These results demonstrated that the AGDV sequence on the carboxyl terminus of the gamma chain of Fg plays an exclusive role in platelet adhesion to surface-immobilized Fg, while the carboxyl terminus of the Aalpha chain, including a consensus RGD site, is not required.

Glanzmann thrombasthenia resulting from a single amino acid substitution between the second and third calcium-binding domains of GPIIb. Role of the GPIIb amino terminus in integrin subunit association

To gain insight into region of the platelet GPIIb-IIIa complex involved in receptor biogenesis and function, we examined the biochemical properties of a defective GPIIb-IIIa complex from patient suffering from type II Glanzmann thrombasthenia. Flow cytometric as well as immunoblot analysis of patient platelets showed significantly reduced levels of GPIIb and GPIIIa compared with a normal control. Patient platelets, however, retained the ability to retract a fibrin clot. Sequence analysis of PCR-amplified platelet GPIIb mRNA revealed an Arg327-->His amino acid substitution between the second and third calcium-binding domains of the GPIIb heavy chain, a residue that is highly conserved among integrin alpha-subunits. The recombinant His327 form of GPIIb was found to be fully capable of associating with GPIIIa, therefore the role of the calcium-binding domains in intersubunit association was further examined by constructing amino-terminal segments of GPIIb that ended before the first, second, and third calcium-binding domains. All three fragments were found to associate with GPIIIa, demonstrating that the calcium-binding domains of GPIIb are not necessary for initial complex formation. Regions amino-terminal to the calcium-binding domains of GPIIb may play a heretofore unappreciated role in integrin subunit association.

Modeling the alpha IIb beta 3 integrin solution conformation

The alpha IIb beta 3 platelet integrin is the prototypical member of a widely distributed class of transmembrane receptors formed by the noncovalent association of alpha and beta subunits. Electron microscopic (EM) images of the alpha IIb beta 3 complex show an asymmetric particle with a globular domain from which two extended regions protrude to contact the lipid bilayer. Distance constraints provided by disulfide bond patterns, epitope mapping, and ligand mimetic cross-linking studies rather suggest a somewhat more compact conformation for the alpha IIb beta 3 complex. We have studied the shape of detergent-solubilized alpha IIb beta 3 by employing a low-resolution modeling procedure in which each polypeptide has been represented as an array of interconnected, nonoverlapping spheres (beads) of various sizes. The number, size, and three-dimensional relationships among the beads were defined either solely by dimensions obtained from published EM images of integrin receptors (EM models, 21 beads), or solely by interdomain constraints derived from published biochemical data (biochemical model, 37 beads). Interestingly, although no EM data were employed in its construction, the resulting overall shape of the biochemical model was still compatible with the EM data. Both kinds of models were then evaluated for their calculated solution properties. The more elongated EM models have diffusion and sedimentation coefficients that differ, at best, by +2% and -18% from the experimental values, determined, respectively, in octyl glucoside and Triton X-100. On the other hand, the parameters calculated for the more compact biochemical model showed a more consistent agreement with experimental values, differing by -7% (octyl glucoside) to -6% (Triton X-100). Thus, it appears that using the biochemical constraints as a starting point has resulted in not only a more detailed model of the detergent-solubilized alpha IIb beta 3 complex, where the relative spatial location of specific domains the size of 5-10 kDa can be tentatively mapped, but in a model that can also reconcile the electron microscopy with the biochemical and the solution data.

Dynamic light scattering studies of alpha IIb beta 3 solution conformation

The prototypical integrin receptor, alpha IIb beta 3, isolated from the membrane fraction of human blood platelets by solubilization in Triton X-100 (reduced) and affinity chromatography on lentil lectin-agarose, has been further purified by gel filtration chromatography in octyl glucoside to obtain the intact receptor complex in a form suitable for hydrodynamic measurements. The molecular weight [(6.0 +/- 0.2) x 10(3)] and Stokes radius (2.3 +/- 0.1 nm) of detergent micelles formed in 0.03 M octyl glucoside have been determined by classical light scattering intensity and dynamic light scattering measurements, respectively. An algorithm has been developed which explicitly considers the contribution of detergent micelles to the intensity autocorrelation function of particles suspended in detergent. This procedure has been validated with polystyrene particles of known radius, as well as with the soluble protein fibrinogen. Application of these procedures to dynamic light scattering data obtained with alpha IIb beta 3 resulted in a translational diffusion coefficient (Dto(20,w)) of (2.78 +/- 0.31) x 10(-7) cm2 s-1, corresponding to a Strokes radius (Rs) of 7.67 +/- 0.85 nm for the integrin/octyl glucoside complex. Light scattering intensity measurements gave a molecular weight of (2.26 +/- 0.22) x 10(5) for the polypeptide moiety of the complex, in excellent agreement with the 2.28 x 10(5) value calculated from primary structure data. As a spherical, hydrated alpha IIb beta 3 complex, with bound detergent, would exhibit a Stokes radius of approximately 5 nm, these data indicate considerable asymmetry in the solution conformation of alpha IIb beta 3.

New developments in thrombolytic therapy

Pharmacologic lysis of occlusive, ischemia-producing thrombi has become widely accepted during the past decade. New developments in this field have centered around increasing the efficacy of the known plasminogen activators while employing methods to minimize the risk of hemorrhage and decrease the incidence of rethrombosis. Such methods have included the use of thrombus-directed antibodies linked to plasminogen activators, increased plasminogen (substrate) concentration at the thrombus site, anticoagulant and antiplatelet therapy to prevent thrombus propagation and reformation following lysis, and combination plasminogen activator therapy designed to increase efficacy and safety. These new strategies have been extensively tested in vitro and in a variety of animal models. As we have indicated, extrapolation of such results to human patients cannot be done with confidence. However, the strategies are based on sound rationale and the reported findings should serve as the basis for controlled human trials.

Evidence that the central region of glycoprotein IIIa participates in integrin receptor function

We have obtained evidence that the ligand-recognition region of the integrin beta-subunit, platelet glycoprotein IIIa (GPIIIa), is discontinuous. Receptor function can be localized to residues near the N-terminus and to the central region of the polypeptide chain. The epitope recognized by our monoclonal antibody, CS-1, which substantially inhibits fibrin(ogen) binding to ADP- and thrombin-stimulated platelets [Ramsamooj, Doellgast & Hantgan (1990) Thromb. Res. 58, 577-592], is contained within residues 349-422 of GPIIIa. This sequence is adjacent to a proteinase-resistant domain of GPIIIa which is linked by disulphide bond(s) to an N-terminal segment near to the putative Arg-Gly-Asp recognition site [D'Souza, Ginsberg, Burke, Lam & Plow (1988) Science 242, 91-93]. Limited trypsin digestion of purified platelet GPIIIa yielded a mixture of two-chain molecules comprised of an N-terminal fragment disulphide-bonded to one of four fragments, which began at residues 299, 303, 353 or 423. Tryptic cleavage of the 300-422 segment correlated with loss of immunoreactivity with anti-GPIIIa monoclonal antibody, CS-1. Chymotrypsin cleavage of GPIIIa resulted in an N-terminal 19 kDa fragment joined by at least one intrachain cystine residue to a 46 kDa polypeptide beginning at residue 349. Partial reduction with dithiothreitol released the larger chymotryptic fragment with its epitope for CS-1 intact. These results have enabled us to localize the epitope recognized by our inhibitory monoclonal antibody, CS-1, to residues 349-422 of GPIIIa. Our data are consistent with a structure in which both the N-terminal and central regions of GPIIIa, which may be in close proximity in the functional GPIIb-IIIa complex, participate in ligand binding.

Characterization of an antibody raised against reduced glycoprotein IIIa of human platelets

A polyclonal antibody against reduced and vinylpyridylethylated human glycoprotein IIIa was raised in rabbits. Its reactivity with reduced GPIIIa was about 500 times higher than that of the antibody against native GPIIIa. The lowest amounts of purified reduced and native GPIIIa recognized by the antibody against reduced GPIIIa were 25 and 400 ng, respectively. The antibody did not recognize native GPIIIa (about 1-2 micrograms) in platelet extracts and chymotryptic degradation products of GPIIIa. It inhibited ADP-induced platelet aggregation but it did not inhibit fibrinogen binding to ADP-stimulated platelets. Our experiments suggest that the antigenicity of GPIIIa (beta 3 integrin) depends on the conformation of the molecule determined by numerous S-S bridges between cysteine residues.