Fibrin-incorporated vitronectin is involved in platelet adhesion and thrombus formation through homotypic interactions with platelet-associated vitronectin

When a blood clot is formed, vitronectin (VN) is incorporated. Here we studied the consequence of VN incorporation for platelet interactions under flow. Perfusion of whole blood over a fibrin network, formed from purified fibrinogen, resulted in approximately 20% surface coverage with blood platelets. Incorporation of purified multimeric VN into the fibrin network resulted in a 2-fold increase in surface coverage with platelets and in enhancement of platelet aggregate formation. A human monoclonal antibody (huMab VN18), directed against the multimeric form of VN, inhibited platelet adhesion to the combined fibrin/VN matrix to the level of adhesion on fibrin alone. This inhibition was also shown when whole blood was perfused over a plasma-derived clot. Surprisingly, the inhibitory action of the antibody was not directed toward VN incorporated into the fibrin network but toward VN released from the platelets. We conclude that VN-potentiated platelet-clot interaction requires VN in the clot and multimeric VN bound to the platelet surface. Our results provide evidence that homotypic VN interactions contribute to platelet adhesion and aggregation to a blood clot. This report demonstrates for the first time that self-assembly of VN may provide a physiologically relevant contribution to platelet aggregation on a blood clot.     

Use of microfluidics to assess the platelet-based control of coagulation

This paper provides an overview of the various types of microfluidic devices that are employed to study the complex processes of platelet activation and blood coagulation in whole blood under flow conditions. We elaborate on how these devices are used to detect impaired platelet-dependent fibrin formation in blood from mice or patients with specific bleeding disorders. We provide a practical guide on how to assess formation of a platelet–fibrin thrombus under flow, using equipment that is present in most laboratories. In addition, we describe current insights on how blood flow and shear rate alter the location of platelet populations, von Willebrand factor, coagulation factors, and fibrin in a growing thrombus. Finally, we discuss possibilities and limitations for the clinical use of microfluidic devices to evaluate a hemostatic or prothrombotic tendency in patient blood samples.     

Rate-limiting roles of the tenase complex of factors VIII and IX in platelet procoagulant activity and formation of platelet-fibrin thrombi under flow

The importance of factor Xa generation in thrombus formation has not been studied extensively so far. Here, we used mice deficient in either factor VIII or factor IX to determine the role of platelet-stimulated tenase activity in the formation of platelet-fibrin thrombi on collagen. With tissue factor present, deficiency in factor VIII or IX markedly suppressed thrombus growth, fibrin formation and platelet procoagulant activity (phosphatidylserine exposure). In either case, residual fibrin formation was eliminated in the absence of tissue factor. Effects of factor deficiencies were antagonized by supplementation of the missing coagulation factor. In wild-type thrombi generated under flow, phosphatidylserine-exposing platelets bound (activated) factor IX and factor X, whereas factor VIII preferentially co-localized at sites of von Willebrand factor binding. Furthermore, proteolytic activity of the generated activated factor X and thrombin was confined to the sites of phosphatidylserine exposure. With blood from a hemophilia A or B patient, the formation of platelet-fibrin thrombi was greatly delayed and reduced, even in the presence of high concentrations of tissue factor. A direct activated factor X inhibitor, rivaroxaban, added to human blood, suppressed both thrombin and fibrin formation. Together, these data point to a potent enforcement loop in thrombus formation due to factor X activation, subsequent thrombin and fibrin generation, causing activated factor X-mediated stimulation of platelet phosphatidylserine exposure. This implies that the factor VIII/factor IX-dependent stimulation of platelet procoagulant activity is a limiting factor for fibrin formation under flow conditions, even at high tissue factor concentrations.     

A non-occlusive model of arterial thrombus formation in the rat and its modification by inhibitors of platelet function, or thrombin activity.

A technically simple model of arterial thrombosis in the rat, induced by a crush injury to the dorsal aorta is described. The mechanical injury to the artery caused deep medial injury and the formation of a platelet-rich thrombus with associated fibrin formation which was assessed both radiometrically and morphometrically. No significant inclusion of erythrocytes was noted in the thrombus. Administration of the platelet inhibitors aspirin, BM 13505 (a thromboxane receptor antagonist) or CGS 12970 (a thromboxane synthase inhibitor) reduced the extent of platelet deposition on the injured vessel, but no decrease in fibrin(ogen) was observed. In contrast, infusion of prostacyclin resulted in reductions in both these components of the thrombus. In studies involving inhibition of thrombin activity, the direct thrombin inhibitor CGP 39393 (recombinant desulphatohirudin) inhibited both the platelet and fibrin(ogen) deposition. The indirect thrombin inhibitors were less effective; unfractionated heparin and low-molecular-weight heparin inhibited both platelet and fibrin(ogen) deposition but only at doses which rendered the blood uncoagulable, as evaluated by the activated partial thromboplastin time. Dermatan sulphate only inhibited platelet deposition. The results suggest that thrombin plays a key role in the initiation of thrombus formation in this experimental model. The agonist prostaglandins (PGG2, PGH2, and TXA2) would appear to have a supporting role in the platelet deposition onto the thrombotic surface but do not have a role to play with respect to fibrin(ogen) deposition.     

The haemostatic plug in haemophilia A: a morphological study of haemostatic plug formation in bleeding time skin wounds of patients with severe haemophilia A

Haemostatic plug formation in four patients with severe haemophilia A (VIII:C less than 1%) was studied in skin biopsies taken at 3, 10 and 30 min and 2 h after a template bleeding time wound had been made. The primary haemostatic plug showed relatively minor changes, consisting of a delay in platelet degranulation and interdigitation. Some platelet aggregates not attached to vessels were encountered in the wound. Subsequently the primary haemostatic plug changed into a firm stable degranulated mass of interdigitated platelets. The major abnormality occurred during the fibrinous transformation. At 2 h many haemostatic plugs consisted of a thin peripheral layer of fibrin and platelet remnants around a central area containing red and white blood cells with a varying amount of plasma and only relatively few fibrin fibres. These observations suggest that fibrin formation in the periphery of the plug is less dependent of factor VIII than in central areas. The lack of fibrin formation in the centre of the plug compensating for the platelet lysis at 2 h may have caused the central erosion of the plug.     

Key role of glycoprotein Ib/V/IX and von Willebrand factor in platelet activation-dependent fibrin formation at low shear flow

A microscopic method was developed to study the role of platelets in fibrin formation. Perfusion of adhered platelets with plasma under coagulating conditions at a low shear rate (250(-1)) resulted in the assembly of a star-like fibrin network at the platelet surface. The focal fibrin formation on platelets was preceded by rises in cytosolic Ca(2+), morphologic changes, and phosphatidylserine exposure. Fibrin formation was slightly affected by α(IIb)β(3) blockage, but it was greatly delayed and reduced by the following: inhibition of thrombin or platelet activation; interference in the binding of von Willebrand factor (VWF) to glycoprotein Ib/V/IX (GpIb-V-IX); plasma or blood from patients with type 1 von Willebrand disease; and plasma from mice deficient in VWF or the extracellular domain of GpIbα. In this process, the GpIb-binding A1 domain of VWF was similarly effective as full-length VWF. Prestimulation of platelets enhanced the formation of fibrin, which was abrogated by blockage of phosphatidylserine. Together, these results show that, in the presence of thrombin and low shear flow, VWF-induced activation of GpIb-V-IX triggers platelet procoagulant activity and anchorage of a star-like fibrin network. This process can be relevant in hemostasis and the manifestation of von Willebrand disease.     

Impaired platelet adhesion to lysed fibrin, whereas neutrophil adhesion remains intact under conditions of flow.

Vessel wall injury induces the formation of a haemostatic plug. Restoration of vascular integrity should involve cessation of further platelet and fibrin deposition and subsequent removal of these thrombi by both the fibrinolytic system and proteases delivered by infiltrating inflammatory cells. We hypothesized that adhesion of platelets and inflammatory cells [polymorphonuclear leucocyte (PMN)] to fibrin is differently supported after exposure of fibrin during fibrinolysis. Fibrin surfaces were exposed to fibrinolytic agents, and platelet and PMN adhesion was studied under conditions of flow. Specific adhesion of platelets to preformed fibrin was reduced by fibrinolytic treatment of the fibrin. PMN adhesion to fibrin was only slightly affected even after 180 min exposure to plasmin. With fibrin still present after fibrinolytic treatment, the impaired platelet adhesion seems explained by loss of the primary platelet adhesion site gamma400-411 on fibrin. PMN binding to fibrin clearly depends on other sites that are less degraded by fibrinolysis. We have shown that PMN adhesion in flowing blood to lysed fibrin was still present, whereas platelet adhesion was impaired due to the loss of the primary platelet adhesion site gamma400-411. Based on our in-vitro perfusion model, we conclude that fibrinolysis specifically interferes with the thrombogenicity of fibrin in the haemostatic plug, whereas the inflammatory response is preserved. The latter may participate in the long-term removal and restructuration of the plug.     

Comparing techniques: the use of recalcified plasma in comparison with citrated plasma alone and in combination with thrombin in ultrastructural studies

Fibrin plays a vital role in the coagulation process and fibrin fiber morphology can be studied using ultrastructural techniques. When studying the ultrastructure of fibrin networks, thrombin may be added to the plasma, ensuing fibrin network formation. The question that arises is whether there are differences in morphology when thrombin is added to plasma, versus morphology observed when plasma from citrated or recalcified citrated whole blood, is studied. The current study therefore aimed to compare ultrastructure of platelets and fibrin networks from these three techniques. Results indicated comparable platelet ultrastructure between smears formed from the plasma of citrated blood and that of the citrated recalcified blood. This method might give us further information regarding the 'natural state' fibrin assembly and association with platelets, when studying haemostasis. However, when studying the ultrastructure of fibrin networks, the addition of thrombin is necessary to form an expansive, fully coagulated layer of fibrin fibers.     

Dissociation between fibrinogen and fibrin interaction with platelets in patients with different subtypes of Glanzmann's thrombasthenia: studies in an ex vivo perfusion chamber model

To explore the possible role of a residual or variant alphaIIbbeta3 integrin (alphaIIbbeta3) in thrombogenesis, we used a new ex vivo perfusion chamber model to examine blood from patients with different subtypes of Glanzmann's thrombasthenia (GT). Non-anticoagulated blood was perfused through capillaries coated with type III collagen for 4.5 min (shear rate: 1600/s). Platelet deposition was quantified as platelet adhesion and mean thrombus size volume; fibrin and von Willebrand Factor (VWF) were specifically revealed by immunohistochemistry. In two patients with variant and in one patient with type II GT, platelet adhesion was maximal and we observed an unexpected formation of thrombi that were smaller than normal in size. These thrombi were surrounded by a thick meshwork that displayed a strong staining for fibrin and VWF. In two patients with heterozygous GT, platelet adhesion and thrombogenesis were normal. In two patients with type I GT, there was no thrombus formation, although platelet adhesion was also maximal. These data suggest the existence of a substitute pathway for thrombogenesis mediated by fibrin and possibly alphaIIbbeta3 (alphaIIbbeta3 at a reduced level, as in type II, and/or abnormal) as this fibrin network was not observed in type I GT with no alphaIIbbeta3. These interactions might facilitate haemostasis and even lead to thrombosis under certain favourable conditions. Furthermore, these data might have pharmacological relevance to the development of anti-alphaIIbbeta3 antithrombotic agents.     

Trends in the development of antithrombotic agents

In venous thrombosis and pulmonary embolism, blood solidifies to form a clot, which is a homogenous structure containing fibrin, red cells, white cells and platelets. Heparin and oral anticoagulant agents are effective in prophylaxis and treatment. In the future, fibrinolytic therapy (urokinase) and Arvin may prove more useful.

An arterial thrombus consists of platelet masses separated by fibrin strands; any agent that either prevents platelets from sticking to each other or prevents the formation of fibrin might theoretically be effective in preventing thrombosis. However, anticoagulant drugs are at best only marginally useful; although there are other ways of preventing fibrin formation, it now seems reasonable to investigate clinically agents that affect platelet behavior. Some of the methods of studying platelet behavior are described and their limitations are emphasized. Some of the many compounds known to influence platelets are discussed, and those which seem most worthy of clinical trials are indicated.     

Protease and cyclooxygenase inhibitors synergistically prevent activation of human platelets.

Thrombin induces platelet aggregation and formation of a fibrin clot in platelet-rich plasma; leupeptin, a protease inhibitor, partially inhibits platelet aggregation, but it does not inhibit fibrin clot formation. Indomethacin does not inhibit either thrombin-induced platelet aggregation or fibrin clot formation. However, when the two drugs are given together, a synergistic inhibition of thrombin-induced platelet aggregation occurs, while fibrin clot formation remains unaffected. Thrombin-induced stimulation of the release of serotonin in washed human platelets is also synergistically inhibited by the combined actions of leupeptin and indomethacin. Thrombin and collagen, added simultaneously, induce full platelet aggregation and release of serotonin. Neither leupeptin nor indomethacin inhibits platelet responses elicited by both agonists; however, when leupeptin and indomethacin are given together, a synergistic inhibition of thrombin- and collagen-induced response is observed. These findings might be relevant in prophylaxis and treatment of thromboembolic disease.     

Comparing techniques: the use of recalcified plasma in comparison with citrated plasma alone and in combination with thrombin in ultrastructural studies

Abstract

Fibrin plays a vital role in the coagulation process and fibrin fiber morphology can be studied using ultrastructural techniques. When studying the ultrastructure of fibrin networks, thrombin may be added to the plasma, ensuing fibrin network formation. The question that arises is whether there are differences in morphology when thrombin is added to plasma, versus morphology observed when plasma from citrated or recalcified citrated whole blood, is studied. The current study therefore aimed to compare ultrastructure of platelets and fibrin networks from these three techniques. Results indicated comparable platelet ultrastructure between smears formed from the plasma of citrated blood and that of the citrated recalcified blood. This method might give us further information regarding the ‘natural state’ fibrin assembly and association with platelets, when studying haemostasis. However, when studying the ultrastructure of fibrin networks, the addition of thrombin is necessary to form an expansive, fully coagulated layer of fibrin fibers.     

Exercise-induced myocardial ischemia in patients with coronary artery disease: lack of evidence for platelet activation or fibrin formation in peripheral venous blood

The hypothesis that exercise-induced myocardial ischemia is associated with abnormal platelet activation and fibrin formation or dissolution was tested in patients with coronary artery disease undergoing upright bicycle stress testing. In vivo platelet activation was assessed by radioimmunoassay of platelet factor 4, beta-thrombo-globulin and thromboxane B2. In vivo fibrin formation was assessed by radioimmunoassay of fibrinopeptide A, and fibrinolysis was assessed by radioimmunoassay of thrombin-increasable fibrinopeptide B which reflects plasmin cleavage of fibrin I. Peripheral venous concentrations of these substances were measured in 10 normal subjects and 13 patients with coronary artery disease at rest and during symptom-limited peak exercise. Platelet factor 4, beta-thromboglobulin and thromboxane B2 concentrations were correlated with rest and exercise catecholamine concentrations to determine if exercise-induced elevation of norepinephrine and epinephrine enhances platelet activation. Left ventricular end-diastolic and end-systolic volumes, ejection fraction and segmental wall motion were measured at rest and during peak exercise by first pass radionuclide angiography. All patients with coronary artery disease had documented exercise-induced myocardial ischemia manifested by angina pectoris, ischemic electrocardiographic changes, left ventricular segmental dyssynergy and a reduction in ejection fraction. Rest and peak exercise plasma concentrations were not significantly different for platelet factor 4, beta-thromboglobulin, thromboxane B2, fibrinopeptide A and thrombin-increasable fibrinopeptide B. Peripheral venous concentrations of norepinephrine and epinephrine increased significantly (p less than 0.001) in both groups of patients. The elevated catecholamine levels did not lead to detectable platelet activation. This study demonstrates that enhanced platelet activation, thromboxane release and fibrin formation or dissolution are not detectable in peripheral venous blood of patients with coronary disease during exercise-induced myocardial ischemia.     

Role of collagen-adherent platelets in mediating fibrin formation in flowing whole blood

Activated platelets provide assembly sites for coagulation enzyme complexes and in this way can mediate coagulation during hemostasis and thrombosis. In this study, we examined the procoagulant activity of platelets adhering directly to fibrillar collagen, a main thrombogenic constituent of subendothelium. For this purpose, we used a human ex- vivo thrombosis model in which collagen-coated coverslips were exposed to flowing nonanticoagulated blood (shear rate, 65/s) for 5.5 minutes, which led to the deposition of adherent platelets, platelet thrombi, and fibrin. To examine the procoagulant activity of adherent platelets only, a selective antagonist of the platelet GPIIb-IIIa complex, Ro 44- 9883, was infused via a mixing device, resulting in a complete abrogation of platelet thrombus formation but leaving the collagen- adherent platelet layer intact. This platelet layer generated increased postchamber fibrinopeptide A (FPA) levels (203 +/- 33 ng/mL) as compared with control experiments without infusion of inhibitor (95 +/- 13 ng/mL). Concomitantly, fibrin deposition measured by morphometric analysis of cross-sections was also increased, as was the platelet adhesion to collagen. An immunochemical staining of fibrin fibers further showed that the adherent platelets formed the nuclei for fibrin fiber formation. This increase in fibrin deposition was mediated by the intrinsic factor X (F.X) activation complex on adherent single platelets, because almost complete inhibition of FPA generation (9 ng/mL) and fibrin deposition (0.4% +/- 0.2% coverage) was achieved upon coinfusion of the GP IIb-IIIa antagonist and active site-inhibited F.IXa. The large platelet thrombi that were deposited in control experiments contained no significant amounts of immunodetectable fibrin except at the thrombus base, where adherent platelets anchored the thrombi to the collagen surface. These results suggest that the collagen-adherent platelets are important promoters of coagulation during the initial phase of thrombogenesis by providing assembly sites for the F.X activation complex.     

Relationship between tissue factor expression and deposition of fibrin, platelets, and leukocytes on cultured endothelial cells under venous blood flow conditions

Endothelial cell-mediated coagulation and leukocyte adhesion are processes that might be connected by the generation of thrombin. To examine the interaction of procoagulant and proadhesive activity, cultures of endothelial cells were stimulated with tumor necrosis factor-alpha, which resulted in the surface expression of tissue factor. Subsequent exposure to human nonanticoagulated blood at a shear rate of 100 s-1 in a parallel plate perfusion device led to the deposition of polymerized fibrin, which covered 63% of the endothelial surface. In addition, numerous platelet aggregates (71 per 10 mm cross- section) and leukocytes (53 +/- 6/mm2) were deposited on stimulated endothelial cells, whereas no fibrin and only a few platelet aggregates (4 +/- 1 per 10 mm cross-section) and leukocytes (6 +/- 1/mm2) were detected on control cells. A significant portion of the adherent leukocytes bound to fibrin and platelets. However, when the deposition of fibrin and platelet aggregates was inhibited with the anti-tissue factor antibody HTFI-7B8 by 100% and 86%, respectively, leukocyte adherence remained unchanged (68 +/- 6/mm2). This indicated that leukocytes could efficiently adhere to endothelial cells through direct cell-cell contact independent of both thrombin and deposited fibrin. Moreover, this direct adhesion of leukocytes to the endothelial surface was reduced twofold to threefold when fibrin deposition occurred. These data suggest a relationship between endothelial procoagulant and proadhesive properties in that tissue factor-initiated coagulation may contribute to leukocyte adhesion through the formation of an adhesive fibrin/platelet meshwork but concurrently prevents the adhesive endothelial surface to bind leukocytes at its full capacity.     

Recombinant factor VIIa enhances platelet deposition from flowing haemophilic blood but requires the contact pathway to promote fibrin deposition

In prior microfluidic studies with haemophilic blood perfused over collagen, we found that a severe deficiency (<1% factor level) reduced platelet and fibrin deposition, while a moderate deficiency (1–5%) only reduced fibrin deposition. We investigated: (i) the differential effect of rFVIIa (0.04–20 nm ) on platelet and fibrin deposition, and (ii) the contribution of the contact pathway to rFVIIa‐induced haemophilic blood clotting. Haemophilic or healthy blood with low and high corn trypsin inhibitor (CTI, 4 or 40 μg mL^?1) was perfused over collagen at an initial venous wall shear rate of 100 s^?1. At 100 s^?1 wall shear rate, where FXIIa leads to thrombin production without added tissue factor, FXI‐deficient blood (3%) or severely FVIII‐deficient blood (<1%) produced no fibrin at either CTI level. Whereas rFVIIa potently enhanced platelet deposition, fibrin generation was not rescued. Distinct from the high CTI condition, engagement of the contact pathway (low CTI) in moderately FVIII‐deficient (3%) or moderately FIX‐deficient blood (5%) resulted in enhanced platelet and fibrin deposition following 4 nm rFVIIa supplementation. In mildly FVIII‐deficient blood (15%) at <24 h since haemostatic therapy, rFVIIa enhanced both platelet and fibrin generation in either CTI condition although fibrin was produced more quickly and abundantly in low CTI. For tissue factor‐free conditions of severe haemophilic blood clotting, we conclude that rFVIIa reliably generates low levels of ‘signaling’ thrombin sufficient to enhance platelet deposition on collagen, but is insufficient to drive fibrin polymerization unless potentiated by the contact pathway.     

Clostridial gas gangrene. I. Cellular and molecular mechanisms of microvascular dysfunction induced by exotoxins of Clostridium perfringens

Mechanisms responsible for the rapid tissue destruction in gas gangrene are not well understood. To examine the early effects of Clostridium perfringens exotoxins on tissue perfusion, a rat model of muscle blood flow was developed. Intramuscular injection of a clostridial toxin preparation containing both phospholipase C (PLC) and theta-toxin caused a rapid (1-2 min) and irreversible decrease in blood flow that paralleled formation of activated platelet aggregates in venules and arterioles. Later (20-40 min), aggregates contained fibrin and leukocytes, and neutrophils accumulated along vascular walls. Flow cytometry confirmed that these clostridial toxins or recombinant PLC induced formation of P-selectin-positive platelet aggregates. Neutralization of PLC activity in the clostridial toxin preparation completely abrogated human platelet responses and reduced perfusion deficits. It is concluded that tissue destruction in gas gangrene is related to profound attenuation of blood flow initiated by activation of platelet responses by PLC.     

Effect of blood flow on thrombin generation is dependent on the nature of the thrombogenic surface

We have investigated the influence of blood flow on thrombin generation, fibrin formation, and fibrin deposition on procoagulant and nonprocoagulant surfaces. Nonanticoagulated human blood was drawn for 5 minutes directly from an antecubital vein over stimulated endothelial cells expressing tissue factor and over human type III collagen fibrils, positioned in parallel-plate perfusion chambers. The shear rates at these surfaces were 50, 650, and 2,600 s-1. Deposition of platelets and fibrin was measured by morphometry. Thrombin and fibrin formation was determined by measuring prothrombin fragments 1 + 2 (F 1 + 2), thrombin-antithrombin III complexes, (T-AT) and fibrinopeptide A (FPA) in blood effluent from the perfusion chamber at the end of the 5- minute perfusion period. On procoagulant endothelial cells, the thrombi were primarily composed of fibrin. The fibrin deposition (81%, 21%, and 2% at 50, 650, and 2,600 s-1, respectively) and plasma levels of F 1 + 2, T-AT and FPA were shear rate dependent and highest at 50 s-1. There was a positive correlation between F 1 + 2 and T-AT and the fibrin deposition (P < .01). In contrast, the collagen surface triggered primarily thrombi that were composed of platelets. The platelet thrombi and plasma levels of F 1 + 2 and T-AT were also dependent on the shear rate, but highest at 650 and 2,600 s-1. F 1 + 2 and T-AT reached the same level as observed with procoagulant endothelial cells at the higher shear rates. There was a positive correlation between F 1 + 2 and T-AT and the platelet thrombus formation (P < .05), confirming the predominant role of platelets in thrombin generation. Thus, thrombin formation is strongly influenced by the blood flow, and this effect depends on the composition of the thrombogenic surface.     

Hematopoietic cell-derived microparticle tissue factor contributes to fibrin formation during thrombus propagation

Tissue factor (TF) is expressed on nonvascular cells and cells within the vessel wall and circulates in blood associated with microparticles. Although blood-borne TF accumulates into the developing thrombus during thrombus formation, the contribution of blood-borne TF and vessel wall TF to thrombin generation in vivo following vessel injury is unknown. To determine the source and role of blood-borne microparticle TF, we studied arterial thrombus formation in a living mouse using intravital microscopy. Platelet, TF, and fibrin accumulation in the developing thrombus was compared in wild-type and low TF mice. Compared to wild-type mice, low TF mice formed very small platelet thrombi lacking TF or fibrin. Wild-type and low TF mice received transplants of bone marrow from wild-type and low TF mice. Arterial thrombi in low TF bone marrow/wild-type chimeric mice had decreased size and decreased TF and fibrin levels. Arterial thrombi in wild-type bone marrow/low TF chimeric mice showed decreased platelet thrombus size but normal TF and fibrin levels. This demonstrates that blood-borne TF associated with hematopoietic cell-derived microparticles contributes to thrombus propagation.