The action of high-dose factor VIIa (FVIIa) in a cell-based model of hemostasis

The cascade model of coagulation 

In the 1960s, two groups proposed a model of coagulation that incorporated a series of sequential steps whereby activation of one clotting factor led to the activation of another, finally leading to a burst of thrombin generation.3, 7 In this model, the clotting pathways were divided into the “extrinsic” and “intrinsic” systems. Both pathways could activate factor X (FX) which, in complex with its cofactor Va, then converted prothrombin to thrombin (Fig 1).

View Large Image Download to PowerPoint

Fig. 1. The “cascade” model of coagulation. Distinct “extrinsic” (TF/FVIIa-initiated) and “intrinsic” (FXII-initiated) pathways converge into the “common pathway” at the FXa/FVa complex.

The currently accepted coagulation cascade is a reasonably good model to explain the usefulness of the coagulation screening test prothrombin time (PT) and activated partial thromboplastin time (aPTT), which correspond to the extrinsic and intrinsic pathways. However, in its current format this model is clearly inadequate to explain the pathways leading to hemostasis in vivo and is inconsistent with clinical observations in several key respects. If there are separate intrinsic and extrinsic coagulation pathways in vivo, why can the activation of FX by the extrinsic (FVIIa/TF-initiated) pathway not compensate for a lack of factor VIII (FVIII) or IX (FIX) in hemophiliacs? In fact, the activation of hemostasis by an intrinsic pathway in vivo is questionable, since deficiency of factor XII (FXII), high-molecular-weight kininogen (HMWK), or prekallekrein (PK) does not cause a tendency to clinical bleeding. It is, however, clear that some components of the intrinsic pathway, such as FVIII and FIX, are essential for hemostasis, since their deficiency leads to hemophilia. We also know that deficiencies of FX, factor V (FV) and factor VII (FVII) lead to serious clinical bleeding syndromes. In contrast, factor XI (FXI) deficiency is much less predictable in causing a bleeding diathesis and leads to a less severe clinical picture than deficiency of FIX or FVIII. Therefore, we and others concluded that it is highly unlikely that separate extrinsic and intrinsic pathways operate under normal conditions in vivo. As a result, the overall model of coagulation required rethinking.

Key observations previously made by several groups led to the revision of earlier models of coagulation. One major observation was that FVIIa/TF activated not only FX but also FIX.12 Other important observations by several groups led to the conclusion that the major initiating event in hemostasis in vivo was the formation of a FVIIa/TF complex at the site of injury.9, 10, 13 This led to the belief that FVIII and FIX deficiency, which resulted in hemophilia A and B, respectively were, in fact, abnormalities of the FVIIa/TF pathway, even though FIX and FVIII were considered to be components of the intrinsic system. Also of importance was the finding that thrombin could directly activate FXI on a charged surface.4 The fact that activated platelets could provide a surface for activation of FXI by thrombin under physiologic conditions2, 11 made it clear why FXII, HMWK, and PK might not be required for hemostasis.

A cell-based model of coagulation 

We have developed a conceptual model of hemostasis that focuses on the roles of specific cell surfaces in controlling the coagulation process. We view hemostasis as occurring in three overlapping phases. The initiation of coagulation takes place on TF-bearing cells (Fig 2).

View Large Image Download to PowerPoint

Fig. 2. A cell-based conceptual model of coagulation. The TF/FVIIa (“extrinsic”) pathway initiates coagulation on the surface of TF-bearing cells. The small amount of thrombin formed on the TF-bearing cell amplifies the procoagulant signal by activating cofactors (FV and FVIII) and FXI, and enhances platelet activation. In the propagation phase of hemostasis the VIIIa/IXa complex (old “intrinsic” FX-activating complex) activates FX on the platelet surface, and that FXa is responsible for the large-scale thrombin production that leads to fibrin clot formation.

An emphasis on localization explains some clinical problems that an enzymatic approach does not 

An emphasis on the role of the coagulation proteins in controlling coagulation suggests that we should be able to understand the physiology and pathophysiology of hemostasis if we had a sufficiently good understanding of the enzymology and kinetics of the individual reactions. So far, this has not been the case. Conversely, studies on the cellular control of coagulation have helped us to understand the pathophysiology of hemophilia and the hemostatic effects of high-dose FVIIa therapy.

Why do hemophiliacs bleed? 

As detailed above, the cascade model does not explain why the extrinsic pathway cannot produce sufficient FXa to, at least partially, compensate for a lack of FVIII or FIX. This means that it is not known why FXa generated by FVIIa/TF is unable to substitute for FXa generated by FIXa/FVIIIa. It has been hypothesized that this is because tissue factor pathway inhibitor (TFPI) shuts off the FVIIa/TF pathway before it can make enough FXa to support generation of hemostatic amounts of thrombin.

While inhibition of the TF pathway by TFPI might contribute to the bleeding tendency in hemophilia, we believe that the mechanism underlying hemophilia can be better understood by looking elsewhere. Our data in a cell-based model suggest that the problem in hemophilia is not that insufficient FXa is made, but that it is made on the wrong cell surface. FXa made on a TF-bearing cell has difficulty in making its way to the platelet surface without being inhibited by either antithrombin III (ATIII) or TFPI. At their normal plasma levels, ATIII and TFPI inhibit FXa so effectively that its half-life is 1 minute or less in the solution phase. In order for FXa to be incorporated efficiently into prothrombinase complexes, we believe that it must be formed on the platelet surface (ie, by FIXa/FVIIIa) in close proximity to factor Va (FVa). Thus, we propose that in hemophilia there is a failure specifically of platelet-surface FX activation leading to a failure of platelet-surface thrombin generation.

Why does FVIIa provide hemostasis in hemophilia? 

We believe that the perception of hemophilia as a failure of platelet-surface FXa generation also allows us to understand why high-dose FVIIa (NovoSeven®, Novo Nordisk, Copenhagen, Denmark) is effective in promoting hemostasis in hemophiliacs. It is well documented that FVIIa can lead to hemostasis in hemophiliacs; however, the mechanism by which it acts has been a matter of ongoing debate. The doses of FVIIa required for efficacy produce plasma levels that are several orders of magnitude greater than the Kd for binding of FVIIa to TF. This led many researchers to conclude that the mode of action of high-dose FVIIa is unlikely to be via TF. However, it is also well recognized that FVIIa has very little proteolytic activity in the absence of its cofactor. We have shown that FVIIa can bind to activated platelets (but not unactivated platelets) with an affinity similar to its affinity for synthetic phosphatidylserine (PS)-containing lipid vesicles.6 Once bound to the activated platelet surface, FVIIa can activate FX.6 The amount of FXa generated by this mechanism is low compared to the FVIIa/TF complex, but is sufficient to significantly enhance thrombin generation in experimental models of FIX and FVIII deficiency.1, 6, 8 This finding is consistent with our conceptual model of coagulation in which platelet surface FXa generation is necessary to support platelet surface prothrombinase assembly. Since we believe that platelet surface thrombin generation fails in hemophilia, restoration of platelet surface thrombin generation by high levels of FVIIa is a plausible mechanism of action. In further support of a platelet surface mechanism of action for high-dose FVIIa is our finding that the levels of FVIIa that must be attained for efficacy in vivo correlate very well with the levels needed for restoration of platelet surface thrombin generation in vitro. A platelet surface mechanism would also explain why high-dose FVIIa is also effective in establishing hemostasis in patients with thrombocytopenia and platelet function defects. We have found that in models of these conditions, the addition of high levels of FVIIa enhances the amount of thrombin generated on each platelet.

Recently, FVIIa was studied in a coagulation model using plasma concentrations of purified coagulation factors with reactions initiated by TF relipidated into synthetic phospholipid vesicles.14 This study showed that plasma levels of zymogen FVII prolonged the lag before initiation of coagulation. The lag was shortened by the addition of rFVIIa. The author hypothesized that this mechanism could account for the clinical efficacy of high-dose FVIIa in hemophilia. However, we find that when cell-associated TF is used to initiate coagulation, the presence of zymogen FVII does not extend the lag period, nor does FVIIa shorten the lag before initiation of thrombin generation (Fig 3).

View Large Image Download to PowerPoint

Fig. 3. The presence of plasma levels of zymogen FVII does not delay or reduce thrombin generation in a cell-based experimental model of coagulation. The experimental model contains TF-expressing monocytes plated in microtiter wells, purified unactivated platelets and plasma levels of purified coagulation factors II, V, VIII, IX, X, XI and antithrombin III, and tissue factor pathway inhibitor. The data shown in triangles represent replicates of the thrombin-generation curve when the coagulation reactions are initiated with 0.2 nmol/L FVIIa in the absence of any zymogen FVII. The data shown as circles represent the pattern of thrombin generation when the reactions are initiated by 0.2 nmol/L FVIIa in the presence of plasma levels (10 nmol/L) of FVII. It is clear that there is no significant difference in the thrombin-generation curves.

While our findings suggest that high-dose FVIIa therapy has its primary effect on the activated platelet surface, this does not preclude any contribution from a TF-mediated mechanism. However, a platelet-dependent mechanism provides a better explanation of the clinical efficacy of high-dose FVIIa therapy. In addition, localization of FVIIa to activated platelets may account for the relative lack of thrombotic complications during high-dose FVIIa therapy. Localization to the activated platelet surface would tend to localize procoagulant activity to the site where it is appropriate.