Activated protein C has a regulatory role in factor VIII function

Abstract

Mechanisms thought to regulate activated factor VIII (FVIIIa) cofactor function include A2-domain dissociation and activated protein C (APC) cleavage. Unlike A2-domain dissociation, there is no known phenotype associated with altered APC cleavage of FVIII, and biochemical studies have suggested APC plays a marginal role in FVIIIa regulation. However, the in vivo contribution of FVIIIa inactivation by APC is unexplored. Here we compared wild-type B-domainless FVIII (FVIII-WT) recombinant protein with an APC-resistant FVIII variant (FVIII-R336Q/R562Q; FVIII-QQ). FVIII-QQ demonstrated expected APC resistance without other changes in procoagulant function or A2-domain dissociation. In plasma-based studies, FVIII-WT/FVIIIa-WT demonstrated dose-dependent sensitivity to APC with or without protein S, whereas FVIII-QQ/FVIIIa-QQ did not. Importantly, FVIII-QQ demonstrated approximately fivefold increased procoagulant function relative to FVIII-WT in the tail clip and ferric chloride injury models in hemophilia A (HA) mice. To minimize the contribution of FV inactivation by APC in vivo, a tail clip assay was performed in homozygous HA/FV Leiden (FVL) mice infused with FVIII-QQ or FVIII-WT in the presence or absence of monoclonal antibody 1609, an antibody that blocks murine PC/APC hemostatic function. FVIII-QQ again demonstrated enhanced hemostatic function in HA/FVL mice; however, FVIII-QQ and FVIII-WT performed analogously in the presence of the PC/APC inhibitory antibody, indicating the increased hemostatic effect of FVIII-QQ was APC specific. Our data demonstrate APC contributes to the in vivo regulation of FVIIIa, which has the potential to be exploited to develop novel HA therapeutics.

Graphical abstract

Figure 1.

Characterization of FVIII-QQ. (A) FVIII domain structure with thrombin and APC cleavage sites noted. (B) SDS-PAGE analysis of 1.5 µM of FVIII-WT and FVIII-QQ before and after 20-minute incubation with 10 nM of thrombin. The gel was stained with Coomassie blue. (C) Representative tracings of thrombin generation in HA human plasma reconstituted with varying concentrations of either FVIII-WT (blue line) or FVIII-QQ (red dashed line) initiated with 1 pM of FXIa in the presence of 4 µM of PCPS and 7.5 mM CaCl₂. (D) Decline in FVIIIa activity because of A2-domain dissociation determined by intrinsic Xase assay; 5 nM of FVIIIa-WT (blue squares) or FVIIIa-QQ (red squares) was incubated with 100 nM of thrombin for 30 seconds, and residual activity of FVIIIa was assessed over 15-minute incubation as described in “Methods.” Data shown are representative of 3 independent experiments.

Figure 2.

Activated protein C cleavage of FVIII-WT and FVIII-QQ. (A) Western blot analysis of 10 nM of FVIII-WT and FVIII-QQ after 30-minute incubation with 6 nM of APC, 20 µM of PCPS, and 6 nM of hirudin. FVIII fragments were visualized with an anti-A2 antibody (GMA-012). (B) Western blot analysis of 10 nM of FVIII-WT, FVIII-QQ, and FVIII-R372Q after 30-minute incubation with 6 nM of APC, 20 µM of PCPS, and 6 nM of hirudin; 30 ng of purified protein was loaded on the gel, and FVIII fragments were visualized with GMA-012. (C) Inactivation of 10 nM of FVIII-WT (blue squares) and FVIII-QQ (red triangles) by 6 nM of APC in the presence of 20 µM of PCPS and 6 nM of hirudin over time in purified intrinsic Xase assay compared with inactivation of 10 nM of FVIII-WT (open blue squares) and FVIII-QQ (open red triangles) by 6 nM of APC with 100 nM of PS in the presence of 20 µM of PCPS and 6 nM of hirudin. Initial velocities of FXa generation throughout incubation were compared with 0-minute time point to determine residual FVIII activity. Representative plots of duplicate experiments are plotted. Data were fit to exponential decay or linear regression (FVIII-QQ with APC only). (D) Western blot analysis of 10 nM of FVIII-WT, FVIII-QQ, and FVIII-R372Q, 20 µM of PCPS and 6 nM of hirudin after 2- and 10-minute incubations with either 100 nM of PS or 6 nM of APC or 6 nM of APC and 100 nM of PS; 20 ng of purified protein was loaded on the gel, and FVIII fragments were visualized with GMA-012. FVIII-R372Q is resistant to cleavage at Arg372. SC, single chain.

Figure 3.

Effect of APC on FVIII-WT/FVIIIa-WT vs FVIII-QQ/FVIIIa-QQ on thrombin generation in reconstituted HA human and mouse plasma. Thrombin generation was evaluated in the presence of increasing APC concentrations in HA plasma reconstituted with FVIII with 4 µM of PCPS and 7.5 mM of CaCl₂. (A) HA human plasma was either reconstituted with 1 nM of FVIII-WT (blue squares) or FVIII-QQ (red triangle), and thrombin generation was initiated with FXIa (1 pM). (B) FVIII (1.5 nM) was activated with thrombin (30 nM) for 30 seconds and quenched with of hirudin (60 nM). HA human plasma was reconstituted with 0.2 nM of FVIIIa-WT or FVIIIa-QQ. Thrombin generation was initiated with FXIa (10 pM). (C) HA mouse plasma was reconstituted with 1 nM of FVIII-WT (blue squares) or FVIII-QQ (red triangles), and thrombin generation was initiated with FXIa (30 pM). (D) FVIII (1.5 nM) was activated with thrombin (30 nM) for 30 seconds and quenched with hirudin (60 nM). HA mouse plasma was reconstituted with 0.2 nM of FVIIIa-WT or FVIIIa-QQ. Thrombin generation was initiated with FXIa (400 pM). In all panels, residual peak thrombin represents peak thrombin relative to the 0-nM APC condition. Means ± standard errors of the mean of 4 independent experiments are plotted.

Figure 4.

FVIII-QQ demonstrates superior in vivo hemostatic function or clot formation compared with FVIII-WT in HA mice. HA mice were infused with phosphate-buffered saline (PBS; open diamonds) or increasing concentrations of FVIII- WT (blue squares) or FVIII-QQ (red triangles) with or without 10 mg/kg of mAb 1609 as indicated before undergoing tail clip injury (A) or 7.5% FeCl₃ injury (C). WT mice infused with PBS (black circles) served as hemostatically normal controls. Each point represents a single mouse, and medians and interquartile ranges are displayed. Kruskal-Wallis test was used to determine significance relative to WT PBS controls, with P values ≤.1 considered significant. Dose-dependent vessel occlusion of FVIII-WT and FVIII-QQ were determined by empirically fitting tail clip (B) and 7.5% FeCl₃ injury (D) data to a logistic function (solid lines). Points represent median values, and error bars represent interquartile ranges. EC₅₀ and EC₈₀ values were determined from logistic fitting. Dotted line indicates median value of hemostatically normal controls. *P ≤ .1, **P ≤ .05, ***P ≤ .01. n.s., not significant.

Figure 5.

Effect of APC on FVIII-WT/FVIIIa-WT vs FVIII-QQ/FVIIIa-QQ on thrombin generation in reconstituted HA/FVL murine plasma. Thrombin generation was evaluated in the presence of increasing APC concentrations in FVIII-reconstituted HA/FVL murine plasma with 4 µM of PCPS and 7.5 mM of CaCl₂. (A) HA/FVL plasma was either reconstituted with 1 nM of FVIII-WT (blue squares) or FVIII-QQ (red triangles), and thrombin generation was initiated with FXIa (30 nM). (B) FVIII (1.5 nM) was activated with thrombin (30 nM) for 30 seconds and quenched with hirudin (60 nM). HA/FVL murine plasma was reconstituted with 0.2 nM of FVIIIa-WT or FVIIIa-QQ. Thrombin generation was initiated with FXIa (400 pM). In both panels, residual peak thrombin represents peak thrombin relative to the 0-nM APC condition. Means ± standard errors of the mean of 4 independent experiments are plotted.

Figure 6.

Enhanced hemostatic effect of FVIII-QQ relative to FVIII-WT is APC dependent. HA/FVL mice were infused with PBS (open diamonds), FVIII-WT (blue squares), or FVIII-QQ (red triangles) at 2 µg/kg with or without 10 mg/kg of mAPC anticoagulatant inhibitory antibody (mAb 1609) as indicated and then underwent tail clip injury. Each point represents 1 mouse, and medians with interquartile ranges are presented. Kruskal-Wallis test was used to determine significance relative to the FVL PBS controls, with P values ≤.1 considered significant. *P ≤ .1, **P ≤ .05, ***P ≤ .01. n.s., not significant.