Summaryvon Willebrand factor (vWF) mediates platelet adhesion at sites of vascular damage. It acts as a bridge between receptors on platelets and collagens present in the connective tissue. Two collagen binding sites have been identified on the A1 and A3 domain of the vWF subunit. To study the functional importance of these binding sites, we have made two deletion mutants that lack the A1 domain (residues 478-716; ΔA1-vWF; Sixma et al. Eur. J. Biochem. 196,369,1991 [1]) or the A3 domain (residues 910-1113; ΔA3-vWF). After transfection in baby hamster kidney cells overexpressing furin, the mutants were processed and secreted efficiently. Ristocetin or botrocetin induced platelet binding was normal for ΔA3-vWF as was binding to heparin and factor VIII. As reported by Sixma et al. (1) ΔAl-vWF still binds to collagen type III, indicating that the A3 domain is sufficient for the interaction. In the current study, we investigated the binding of ΔA3-vWF to collagen type III. When preincubated on collagen type III it did not support platelet adhesion under flow conditions, whereas it was able to support platelet adhesion when coated directly to a glass surface. The binding of 125I-ΔA3-vWF to collagen was specific but maximal binding was about 40 times less compared to 125I-vWF. When added at 25 times excess, ΔA3-vWF did not compete with 125I-vWF for binding to collagen type III, whereas ΔAl-vWF did. The binding of 125I-ΔA3-vWF could be blocked by excess unlabeled vWF but not by ΔA1-vWF. In conclusion, we demonstrate that the A3 domain in vWF contains the major collagen binding site. The major binding site present on the A3 domain and the minor site present on A1 bind to different sites on collagen.

When the endothelial cell layer Is damaged, a thrombotic reaction starts on the cells' subendothelium and on the connective tissue deposited by smooth muscle cells in the deeper layers. When more severe vascular damage occurs, hemostasis will Involve the vessel adventitia In which fibroblasts are found. In this article, the influence of in vitro cultured endothelial cells, smooth muscle cells, and fibroblasts on the hemostatic balance was studied. To do so, perfusions were performed with low molecular weight heparln antlcoagulated blood over the extracellular matrix of the cells. This method allowed the study of tissue factor-dependent thrombln generation and its influence on formation of fibrin and platelet aggregates. The experiments described in this article show that endothelial cells isolated from different human organs Interfere differently In the hemostatic response. Endothelial cells Isolated from umbilical veins are nonthrombogenic; they do not synthesize tissue factor under unstlmulated conditions. On their extracellular matrix, only adherent platelets are found, but no aggregates and no fibrin. Endothelial cells Isolated from omentum and atrium contain tissue factor activity under unstimulated conditions. As a consequence, thrombln Is generated on their surfaces, and platelet aggregates and fibrin deposition are found on the extracellular matrices after perfusions with whole blood. The matrix of smooth muscle cells and fibroblasts behaved similarly. Increase In shear rate and perfuslon time resulted In an Increase in platelet aggregate formation. Polymerized fibrin deposition decreased when perfusions were performed at higher shear. Both platelet aggregation and fibrin deposition were tissue factor dependent and could be blocked more than 70% by an antibody against tissue factor. Based on these results, we conclude that endothelial cells isolated from umbilical veins form the best nonthrombogenic surface In vitro. Moreover, coagulation-dependent hemostasis should be Included when thrombogenicity of subendothelium Is discussed, especially when it concerns matrix derived from cells present In the deeper layer of the vessel wall.

We describe glycoprotein (GP) Ib as a mediator of adhesion to fibronectin, specifically in flow. A monoclonal antibody (MoAb) directed to the von Willebrand factor (vWF)-binding site on this receptor or the absence of this receptor on the platelet membrane, in the case of a patient with the Bernard-Soulier syndrome, reduced platelet coverage to fibronectin to approximately 30% of the control value. A MoAb directed to the GP Ib-binding site on vWF showed a similar effect. With washed platelets in the absence of plasma vWF, the inhibitory effect of the anti-GP Ib antibody was the same as with whole blood. No inhibition with the anti-GP Ib antibody was observed when we used blood from patients with severe von Willebrand disease (vWD) or from a patient with vWD type I (platelet low). Addition of vWF to vWD blood resulted in restoration of adhesion. Immunoelectron microscopy on platelets adhering to fibronectin showed that GP Ib was homogeneously distributed over the entire surface of the platelet. vWF was present at the central zone and the edges of the platelet and at the basal interface between the platelet and the fibronectin surface. No direct binding of vWF to fibronectin could be demonstrated. These data indicate that GP Ib-mediated adhesion to fibronectin fully depends on vWF and that normal levels of plasma or platelet vWF are sufficient for optimal adhesion to fibronectin. The data suggest that the presence of platelets during perfusion is a prerequisite for vWF to support platelet adhesion to fibronectin.

Uremia is associated with a bleeding diathesis. We investigated platelet adhesion as a cause for the impaired primary hemostasis and the role of von Willebrand factor (vWF) in this process in uremic patients. Perfusions with blood with standardized hematocrit, platelet count, and free Ca2+ ions were performed over inverted and deendothelialized artery segments from human umbilical cords in a modified Baumgartner perfusion chamber. Platelet adhesion in patient perfusates was comparable with control adhesion. However, the high vWF levels present in uremic whole blood did not increase adhesion above the adhesion in control blood with lower vWF levels. These results suggested that a relative adhesion defect was present in patient blood. Control blood in which vWF levels were raised to uremic levels showed the high adhesion that uremic whole blood failed to show. Additionally, in perfusions with uremic plasma in which the initially high vWF level was normalized by dilution with vWF-depleted uremic plasma, adhesion was clearly lower than in normal plasma. Washed patient platelets did not differ from normal platelets in their association with purified vWF, via their adhesion receptors glycoprotein Ib and IIb-IIIa. Patient platelets present in patient plasma showed a similar adhesion defect as control platelets, which were resuspended in the uremic plasma. Therefore, primary defects of uremic platelets were of minor importance for the observed adhesion defect in uremic whole blood. The adhesion defect was not dependent on the presence of uremic vWF; plasma of uremic patients depleted of vWF also inhibited adhesion, and the defect remained present when purified control vWF was added to vWF-depleted uremic plasma. The binding of uremic vWF to the vessel wall and its support of subsequent adhesion were not impaired. These results indicate that the observed adhesion defect was not due to abnormal vWF. Our current results suggest an unknown component present in uremic plasma that directly inhibits platelet interaction with artery segments; however, it has no effect on vWF binding to the vessel wall. High vWF levels in uremic plasma are able to compensate for the defect.

Uremia is associated with bleeding diathesis. Platelet adhesion to the subendothelium is inhibited by a factor in uremic plasma that may play a role in the disturbed hemostasis of uremic patients. In the formation of the hemostatic plug, platelet adherence is followed by stimulus-induced platelet aggregation and reinforcement by thrombin-generated fibrin. To study these processes in uremic blood, a newly developed thrombosis model was used. Perfusates anticoagulated with low-molecular-weight heparin were circulated over a matrix of stimulated cultured endothelial cells. By stimulation of the endothelial cells, tissue factor was synthesized and deposited in the matrix. When this tissue factor rich-matrix was exposed to flowing blood, local thrombin was formed via activation of the extrinsic coagulation pathway. With this system, platelet adhesion, thrombin-dependent platelet activation, and fibrin formation can all be studied at the same surface. In addition to an adhesion defect, decreased aggregate formation was also found in uremic perfusates. Normal platelets in uremic plasma showed similar results, which indicates that a factor in uremic plasma caused this adhesion and aggregation defect. Platelet aggregation in the system was dependent on endogeneously formed thrombin. Fibrinopeptide A generation, however, was normal in uremic perfusates; therefore, uremic plasma has a normal capacity to form thrombin. Resuspension of washed uremic platelets in control plasma did not reverse the aggregation defect in perfusions. In contrast, aggregometer studies with isolated uremic platelets could not detect an abnormal response to threshold concentrations of exogenous thrombin. Thus, uremic toxin (s) cause defective aggregate formation in flow, but not necessarily in the aggregometer. This apparent discrepancy may be due to the higher shear forces in the flow system, which may prevent aggregate formation that is allowed in the aggregometer. Another explanation, that uremic platelets are less responsive to locally formed thrombin than they are to exogenously added thrombin, seems less likely. [Arteriosclerosis and Thrombosis 1991;ll:733-744)