A B ST R A CT The binding of '251-von Willebrand factor to platelets stimulated by thrombin, ADP, and a combination of ADP + epinephrine (EPI) is specific, saturable, and reversible. Active platelet metabolism and divalent cations are required for binding induced by these stimuli, but not by ristocetin, suggesting the existence of different mechanisms involved in the vWF-platelet interaction. A monoclonal antibody directed against an epitope of membrane glycoprotein (GP) lb had no effect on the binding of '251-vWF to normal platelets stimulated by thrombin or a combination of ADP + EPI, but completely blocked ristocetin-induced binding. Binding induced by thrombin to GPIb-blocked platelets was specific. Moreover, thrombin-induced binding of 1251-vWF was increased, rather than decreased, in two patients with the Bernard-Soulier syndrome whose platelets lacked GPIb. Conversely, monoclonal antibodies directed against the GPIIb/IIIa complex had no effect on ristocetininduced binding of (vWF)' is a large multimeric glycoprotein that circulates in blood complexed with the Factor VIII procoagulant activity protein (1). It plays an essential role in platelet function as shown by the prolonged bleeding time in von Willebrand disease (1). Specific binding sites for vWF are induced on the platelet membrane by the antibiotic ristocetin (2), a nonphysiologic agent, as well as by thrombin (3, 4) and ADP (5). Platelet membrane glycoprotein (GP) lb is considered to function as the surface receptor for vWF (6). In the Bernard-Soulier syndrome, a congenital bleeding disorder, platelets lack GPIb (7) and the ristocetin-induced binding of vWF is decreased (8). On the other hand, in Glanzmann thrombasthenia, another congenital platelet abnormality, there is a marked decrease of the membrane GP complex Ilb/Illa but normal content of GPIb (9). The ristocetin-induced binding of vWF to thrombasthenic platelets has been reported to be normal (8). In contrast, we have recently shown that binding of vWF to thrombasthenic platelets stimulated by thrombin is severely deficient (4). Therefore, we postulated that different sites are involved in ' Abbreviations used Marguerie et al. (12). The final platelet suspensions were all in modified calcium-free Tyrode buffer, containing 137 mM NaCI, 2 mM MgCI2, 0.42 mM NaH2PO4, 11.9 mM NaHCO3, 2.9 mM KCI, 5.5 mM glucose, 10 mM Hepes, pH 7.35, and 20 mg/ml bovine serum albumin (Fraction V, CalbiochemBehring Corp., American Hoechst Corp., San Diego, CA). When experiments of serotonin release were performed, PRP was labeled before washing by incubating 20 ml with 0. (3.3,M). However, ADG-washed platelets usually gave reversible aggregation to ADP and were unresponsive to epinephrine (EPI). On the contrary, gel-filtered platelets always gave irreversible aggregation to ADP, and responded to 20 MM EPI in the presence of fibrinogen (3.3 ,M) with a typical two-wave aggregation. Contamination of plasma vWF in all washed platelet preparations was below 5 X 10' U/dl (1 dl of a normal plasma pool contains ...

We have separated von Willebrand factor (vWF) multimers of different size into several fractions which were characterized by SDS-agarose gel electrophoresis and by measuring the ratio between ristocetin cofactor activity (Ricof) and von Willebrand antigen (vWF:Ag) content. The pooled fractions contained vWF with multimeric structures and Ricof similar to those in plasma. The pool was labelled with 125I and used for inhibition binding studies with individual fractions to calculate the dissociation constants (Kd values expressed in mol/l) of the individual fractions for ristocetin-dependent binding to GP Ib and thrombin-induced binding to GP IIb/IIIa. Direct binding studies of the 125I-vWF pool gave mean Kd values of 2.02 +/- 0.05 x 10(-8) for GP Ib and 1.15 +/- 0.02 x 10(-8) for the GP IIb/IIIa complex. Inhibition binding studies gave Kd mean values one third to one tenth as high for larger multimers and 3-10 times higher for smaller multimers, for both GP Ib and IIb/IIIa complex. Similar results were observed when binding studies were carried out in the presence of platelets from a patient with afibrinogenaemia. These data on binding correlated very well with ristocetin- and thrombin-induced aggregation of afibrinogenaemic platelets, since equal concentrations of the higher molecular weight forms gave significantly higher aggregation rates. Based on these results, we conclude that the affinity of the vWF molecule for its two platelet receptors is greater for the largest multimers.

Glanzmann thrombasthenia is an inherited bleeding disorder characterized by the failure of platelets to aggregate in response to almost all stimuli. However, thrombasthenic platelets will aggregate with bovine and porcine von Willebrand factor (vWF) and will show normal ristocetin-induced binding and aggregation in the presence of human vWF. In contrast, we now report that the specific binding of vWF to the thrombin-stimulated platelets was <20% ofnormal in three patients with Glanzmann thrombasthenia. Analysis of binding isotherms was based on the assumption of one class of binding sites for vWF on the platelet membrane. Double-reciprocal plots were used to calculate maximal binding at saturation and apparent dissociation constant (Keg). In nine normals, 2.82 ± 0.64 Bag (± SD) of vWF bound to 108 platelets at saturation, with Kd (± SD) = 3.65 ± 1.23 pg/ml. In two patients with thrombasthenia binding was markedly decreased and did not approach saturation. In the third patient, binding at saturation corresponded to 0.21 jig per 108 platelets, with K, = 3.93 pg/ml. These findings suggest that mechanisms underlying the vWF-platelet interaction are incompletely reflected in ristocetin-dependent assay systems. Moreover, these results, in addition to those previously reported for fibronectin, suggest that the platelet defect in Glanzmann thrombasthenia is not limited to decreased binding of fibrinogen but involves several glycoproteins that are known to interact with platelets.von Willebrand factor (vWF) is a large multimeric glycoprotein that circulates in plasma complexed with the factor VIII procoagulant activity protein (1). Interaction ofvWF with platelets is necessary for normal platelet adhesion to the subendothelium and for normal platelet-to-platelet interaction (aggregation) (2, 3). In vitro, receptor-mediated binding ofvWF to platelets can be demonstrated in the presence ofthe antibiotic ristocetin (4). A congenital qualitative platelet disorder, the Bernard-Soulier syndrome, is associated with decreased ristocetin-induced binding of vWF. Platelets from these patients lack membrane glycoprotein lb (5) and, therefore, this glycoprotein has been thought to function as the vWF receptor on the platelet surface. However, the physiological significance of the ristocetin-induced vWF-platelet interaction is not clear. Ristocetin is not a physiological inducer ofplatelet function and an in vivo counterpart to it has not been demonstrated. Moreover, ristocetininduced platelet aggregation does not correlate with vWF function in several clinical situations (1).More recently, thrombin-a physiological platelet activating agent-has been shown to induce specific binding sites for fibrinogen (6, 7), fibronectin (8), and vWF (9, 10) on the platelet surface. In this report we demonstrate that, as with the latter two glycoproteins (11-15), thrombin-stimulated platelets from patients with Glanzmann thrombasthenia are deficient in their ability to bind vWF. PATIENTS AND METHODSThe three patients with Glanzmann thrombast...

We have studied three afibrinogenemic patients, who had only trace amounts of plasma and platelet fibrinogen as measured by radioimmunoassay, and demonstrate here that the residual aggregation observed in their platelet-rich plasma is dependent upon von Willebrand factor (vWF) binding to the platelet membrane glycoprotein (GP)IIb/IIIa complex. The abnormality of aggregation was more pronounced when ADP, rather than thrombin, collagen, or the combination of ADP plus adrenaline was used to stimulate platelets. With all stimuli, nevertheless, the platelet response was completely inhibited by a monoclonal antibody (LJP5) that is known to block vWF, but not fibrinogen binding to GPIIb/IIIa. Addition of purified vWF to the afibrinogenemic plasma resulted in marked increase in the rate and extent ofaggregation, particularly when platelets were stimulated with ADP. This response was also completely blocked by LJP5.Addition of fibrinogen, however, restored normal aggregation even in the presence of IPS, a finding consistent with the knowledge that antibody LJP5 has no effect on platelet aggregation mediated by fibrinogen binding to GPIIb/IIIa. Two patients gave their informed consent to receiving infusion of 1-desamino--D-arginine vasopressin (DDAVP), a vasopressin analogue known to raise the vWF levels in plasma by two-to fourfold. The bleeding time, measured before and 45 min after infusion, shortened from >24 min to 12 min and 50 s in one patient and from 16 min to 9 min and 30 s in the other. Concurrently, the rate and extent of ADP-induced platelet aggregation improved after DDAVP infusion. The pattern, however, reversed to baseline levels within 4 h. The concentration of plasma vWF increased after DDAVP infusion, but that of fibrinogen remained at trace levels. We conclude that vWF interaction with GPIIb/Il~a mediates platelet-platelet interaction and may play a role in primary hemostasis.

No abstract