FVIII half-life extension by coadministration of a D'D3 albumin fusion protein in mice, rabbits, rats, and monkeys

Abstract

A novel mechanism for extending the circulatory half-life of coagulation factor VIII (FVIII) has been established and evaluated preclinically. The FVIII binding domain of von Willebrand factor (D'D3) fused to human albumin (rD'D3-FP) dose dependently improved pharmacokinetics parameters of coadministered FVIII in all animal species tested, from mouse to cynomolgus monkey, after IV injection. At higher doses, the half-life of recombinant FVIII (rVIII-SingleChain) was calculated to be increased 2.6-fold to fivefold compared with rVIII-SingleChain administered alone in rats, rabbits, and cynomolgus monkeys, and it was increased 3.1-fold to 9.1-fold in mice. Sustained pharmacodynamics effects were observed (ie, activated partial thromboplastin time and thrombin generation measured ex vivo). No increased risk of thrombosis was observed with coadministration of rVIII-SingleChain and rD'D3-FP compared with rVIII-SingleChain alone. At concentrations beyond the anticipated therapeutic range, rD'D3-FP reduced the hemostatic efficacy of coadministered rVIII-SingleChain. This finding might be due to scavenging of activated FVIII by the excessive amount of rD'D3-FP which, in turn, might result in a reduced probability of the formation of the tenase complex. This observation underlines the importance of a fine-tuned balance between FVIII and its binding partner, von Willebrand factor, for hemostasis in general.

Graphical abstract

Figure 1.

Domain structure of full-length VWF and rD′D3-FP. (A) Domain structure of native full-length VWF residues 1 to 2813, including propeptide D1D2 (blue) and the D′D3 domain (orange). Cysteines responsible for intra- and intermolecular dimerization are shown as yellow stars (C1099 and C1142 in the D3 domain; C2771, C2773, and C2811 in the CK domain). (B) rD′D3-FP monomer consisting of the propeptide domain D1D2 and the D′D3 domains of VWF N-terminally fused to human serum albumin (HSA; purple). This monomer dimerizes via its D3 domain to form the pro-rD′D3-FP dimer, which subsequently is processed by furin-catalyzed cleavage of the propeptide D1D2 to yield the final product, mature dimeric rD′D3-FP, as shown in panel C.

Figure 2.

Time course of rVIII-SingleChain plasma levels after coadministration with rD′D3-FP in rats and rabbits and dependency of FVIII half-life extension on molar excess of rD′D3-FP over endogenous VWF. Plasma levels of rVIII-SingleChain (antigen, dosed at 200 IU/kg in rats [A] and 150 IU/kg in rabbits [B]), alone or coadministered with increasing doses of rD′D3-FP and pdVWF (400 IU/kg in rats and 300 IU/kg in rabbits) as control. The data show the dose-dependent increase in rVIII-SingleChain plasma exposure with increased dosing of rD′D3-FP. The shorter half-life of rVIII-SingleChain coadministered with pdVWF might be explained by a shorter half-life of human VWF in a rat and rabbit model. The horizontal dashed line represents the detection limit of the assay; data are mean ± standard deviation (SD) (n = 2-3). (C) Fold increase in the terminal plasma half-life of rVIII-SingleChain as a function of the molar excess of rD′D3-FP over endogenous VWF for rats (●) and for rabbits (○). The dotted gray line represents a ratio of 1; ie, the terminal plasma half-life of rVIII-SingleChain alone.

Figure 3.

Impact of rD′D3-FP coadministration on aPTT and thrombin-generation assays in rVIII-SingleChain–treated FVIII-ko mice. The effect of rVIII-SingleChain alone (red lines) or coadministered with rD′D3-FP (blue lines) was assessed as aPTT (A) or in a thrombin-generation assay (B). A vehicle control (gray lines) was tested as negative control. (A) For aPTT, rVIII-SingleChain combined with rD′D3-FP resulted in a reduction of the clotting time even at 96 hours postadministration, whereas rVIII-SingleChain alone resulted in an aPTT similar to that in vehicle-treated animals after 72 hours. (B) Thrombin generation by rVIII-SingleChain in the presence of rD′D3-FP was increased from 48 to 96 hours compared with rVIII-SingleChain alone. Data are mean ± SD of 3 to 10 animals. Dashed lines represent mean vehicle values and dotted lines represent mean ± SD vehicle values.

Figure 4.

Effect of rD′D3-FP on tail-clip bleeding in FVIII-ko mice and NMRI mice. (A) The effect of rVIII-SingleChain (100 IU/kg, red circles), alone or coadministered with rD′D3-FP (2.86 mg/kg, blue circles), was assessed in a tail-clip bleeding model to rule out potential negative effects of rD′D3-FP on the hemostasis restored by rVIII-SingleChain (n = 10). A vehicle control (gray circles) was tested at 16 hours postadministration (p.a.) as negative control. There was no significant effect of rD′D3-FP on the rVIII-SingleChain–induced reduction of bleeding time at 6 or 16 hours postadministration, as calculated using variance models to the logarithmically transformed data (pairwise comparisons with Welch's t test). Significant differences as compared to vehicle-treated FVIII-ko mice are marked with an asterisk (*). Graphs show mean ± SD and individual results. (B) Bleeding time in vehicle-treated NMRI mice (saline, gray circles) was compared with NMRI mice treated with high-dose rD′D3-FP (250 mg/kg, blue circles) at 15 minutes after tail clip (n = 20). High-dose rD′D3-FP significantly prolonged bleeding time at 15 minutes postadministration as calculated using an unpaired Student t test. Dotted line represents the observation period.

Figure 5.

Effect of rD′D3-FP on FeCl₃-induced arterial occlusion in FVIII-ko and NMRI mice. (A) The effect of rVIII-SingleChain (100 IU/kg, red line), alone or coadministered with rD′D3-FP (2.86 mg/kg, blue line), was assessed 1 hour after FeCl₃-induced damage of the endothelium in an arterial occlusion model in FVIII-ko mice (n = 10). A vehicle control (gray line) was tested as negative control (n = 5), remaining at 0% occlusion rate. There was no significant effect of rD′D3-FP on the coagulation restored by the coadministered rVIII-SingleChain, suggesting no negative impact of rD′D3-FP on hemostasis. (B) Arterial occlusion in vehicle-treated NMRI mice (saline, gray line) was compared with that after administration of high-dose rD′D3-FP (250 mg/kg, blue line) 15 minutes after FeCl₃-induced damage of the endothelium in an arterial occlusion model (n = 8-9). There was a significant effect of high-dose rD′D3-FP on coagulation in NMRI mice. P values were calculated using the log-rank Mantel-Cox test.

Figure 6.

Impact of rD′D3-FP on aPTT in human plasma. The effect of rD′D3-FP (blue line) or rD′D3-His (red line) on aPTT was assessed in human plasma. A vehicle control (saline, gray lines) was tested as negative control (mean [dashed line] ± SD [dotted lines] of n = 12). rD′D3-FP and rD′D3-His prolonged aPTT. Data are mean ± SD of 3 samples.

Figure 7.

Hypothesis for reduced FVIII efficacy at high concentrations of rD′D3-FP in human plasma. (A) FVIII in human plasma (≈0.3 nM) exists primarily bound to its endogenous carrier protein VWF (≈25 nM dimer) as a result of its low K_D (≈0.1 nM). The fast on- and off-rate of this interaction allows switching of FVIII to rD′D3-FP. At micromolar concentrations, most FVIII is expected to be bound to rD′D3-FP. (B) Upon activation of FVIII, its affinity for VWF is drastically reduced, resulting in its dissociation from rD′D3-FP. Some portion of this free FVIIIa is suggested to bind membranes to form the tenase complex. However, at high micromolar concentrations of rD′D3-FP, rebinding of FVIIIa might result in reduced availability of FVIIIa and, in turn, the observed loss of efficacy. HSA, human serum albumin.