Examining downstream effects of concizumab in hemophilia A with a mathematical modeling approach

Abstract

Background: Tissue factor (TF) pathway inhibitor (TFPI) is an anticoagulant protein that inhibits factor (F)Xa, the TF-FVIIa-FXa complex, and early forms of the prothrombinase complex. Concizumab is a monoclonal antibody that blocks FXa inhibition by TFPI and reduces bleeding in hemophilia.

Objectives: To examine how concizumab impacts various reactions of TFPI to restore thrombin generation in hemophilia A using mathematical models.

Methods: A compartment model was used to estimate plasma concentrations of free concizumab and its complexes with TFPIα and TFPIβ. Concizumab was integrated into a flow-mediated mathematical model of coagulation, and a small injury was simulated under hemophilia A conditions. Simulations were then analyzed to determine how concizumab's blockade of TFPI anticoagulant activities, specifically the inhibition of FXa in plasma and on platelets, inhibition of TF:FVIIa at the subendothelium, and prior sequestration of plasma TFPIα to the endothelium via TFPIβ, altered thrombin generation.

Results: Concizumab improved simulated thrombin generation in hemophilia A by simultaneously altering all 3 mechanisms of the TFPI anticoagulant blockade examined. Concizumab sequestered ∼75% of plasma TFPIα through the formation of ternary TFPIα-concizumab-TFPIβ-complexes. For all TF levels, reducing the TFPIα plasma concentration had the largest impact on the lag time, followed by blocking TFPIα inhibition of TF:FVIIa:FXa and subsequently by blocking TFPIα inhibition of FXa in plasma and on the platelet surface.

Conclusion: The effectiveness of concizumab is mediated through the blockade of TFPI anticoagulant activities in plasma and on multiple physiological surfaces. An important and previously unrecognized function of concizumab was the sequestration of plasma TFPIα to the endothelium.

Keywords: TFPI-antibody; concizumab; hemophilia A; mathematical modeling.

FIGURE 1

Model schematics. The compartment model (left) predicts levels of concizumab bound to tissue factor pathway inhibitor (TFPI)α and TFPIβ. Output from the compartment model is input to the flow-mediated coagulation model (right), which simulates thrombin generation that results from reactions occurring in a small reaction zone above the exposed tissue factor and subjected to flow.

FIGURE 2

Selected steady-state concentrations estimated with the compartment model and plotted as functions of total intravascular concentrations of concizumab (C). Simulations are shown with tissue factor pathway inhibitor (TFPI)β (solid black lines) and without TFPIβ (dotted black lines). Total plasma C concentrations are shown (A) with an orange dot and blue triangle indicating the total intravascular concizumab level leading to 4 nM total plasma concizumab. Steady-state concentrations of free plasma TFPIα (not bound to C), C:TFPIα, and TFPIα:C:TFPIβ are shown in (B–D), respectively.

FIGURE 3

Thrombin generation time courses with varied tissue factor TF, concizumab, and factor (F)VIII. TF was low (black, 4 fmol/cm²), intermediate (brown, 9 fmol/cm²), or high (copper, 20 fmol/cm²). Simulations were run in the presence (solid curves) and absence (dotted and dash-dotted) of concizumab. FVIII is set to 1% of normal to simulate hemophilia A (A) or to 100% of normal (B). The thin black line represents 1 nM thrombin. The lag time was defined as the time when the thrombin curves crossed this line.

FIGURE 4

Concentration time courses of selected model species during simulated, flow-mediated coagulation. The panels show free tissue factor (TF) pathway inhibitor (TFPI)α in plasma (A), the subendothelial TF:factor (F)VIIa:FXa:TFPI complex (B), active FXa both in plasma and bound to a platelet surface (C), and platelet-bound prothrombinase (FXa:FVa and FXa:FV-h; D). The total intravascular concizumab level was 21.5 nM (4.14 nM free plasma concizumab and complexes with TFPIα and TFPIβ). TF was fixed at 9 fmol/cm², and FVIII was 1% (0.01 nM). Concizumab blocked the inhibitory action of TFPIα (A, B), which led to increases in FXa and prothrombinase (C, D).

FIGURE 5

Thrombin generation as a function of tissue factor (TF) concentration and concizumab’s TF pathway inhibitor (TFPI) inhibitory mechanisms. Lag time (time to 1 nM thrombin) as a function of TF (A) and thrombin time courses (B) for various downstream mechanisms of concizumab at a fixed TF level of 9 fmol/cm². Factor (F)VIII is fixed at 1%, and no other factors/inhibitor levels were varied. A thin gray line represents TF = 9 fmol/cm² in (A) and 1 nM thrombin in (B). SE, subendothelium. Note that the points of intersection are the same data in both plots.

FIGURE 6

Thrombin generation time courses with concizumab (C), tissue factor (TF) pathway inhibitor (TFPI)β, and varied factor (F)V-“half” (FVh) interactions. The solid black curve for TF = 9 fmol/cm² in this figure and Figure 5B is the same and shows simulations with both concizumab and TFPIβ. The orange solid curves are the same scenario but in the absence of TFPIβ. The dashed lines are with TFPIβ, but the binding of C:TFPIα to FVh was disabled. FVIII was set to 1% (0.01 nM).

FIGURE 7

The simulated effects of protein S on thrombin generation. Thrombin time courses for tissue factor (TF) at 9, 13, and 17 fmol/cm² with 1% factor (F)VIII (0.01 nM). Black curves are without protein S (baseline association rate between TF pathway inhibitor [TFPI]α and platelet-bound FXa [plt-FXa]). Orange curves are with protein S (increased associate rate). The dotted black and orange curves are without concizumab. Solid curves (with concizumab) are on top of each other, so they appear dashed.