Platelets stabilize factor VIIIa against loss of activity through subunit dissociation

Abstract Activated platelets have binding sites for factor VIII (FVIII) that include soluble fibrin bound to the αIIbβ3 integrin (Phillips et al 2004. J Thromb Haemost; Gilbert et al 2015. Blood). Platelets support FVIII activity at 100-fold lower concentrations than standard FVIII activity assays with phospholipid vesicles (PLV). The mechanism(s) through which platelet binding sites protect or increase FVIII activity have been only partially explored. Factor VIIIa is a semi-stable heterotrimer that can lose activity through dissociation of the A2 domain or through cleavage of scissile bonds susceptible to activated Protein C (APC) or other proteases. We wished to determine whether FVIII activity on platelet binding sites is enhanced by stabilization against A2 dissociation and/or protease cleavage. Like FVIII, the bispecific antibody emicizumab (Emi) serves as a cofactor that facilitates activation of FX by FIXa on either vesicles or platelets. Thus, Emi is an ideal control reagent to determine whether platelet binding sites act primarily through FVIII(a) vs. factor IXa or factor X. Site-directed mutagenesis of recombinant B-domain-deleted human factor VIII (FVIII-WT) replaced the primary APC cleavage sites R336/R562 with Q (FVIII-QQ) or stabilized the A2 domain against spontaneous dissociation by replacing D519/E665 with V (FVIII-VV) or both (FVIII-QQVV). Full-length recombinant factor VIII (octocog alfa (Antihemophilic Factor), Advate) was also used as a control (FVIII). Plasma clotting assays were performed using delipidated factor VIII-deficient plasma with corn trypsin inhibitor to inhibit contact pathway activation. FVIII was added to plasma in the presence of either PAR1/PAR4 activated cryopreserved platelets (apheresis platelets in 5% DMSO stored at -80C and then purified via density gradient, 3x107/mL) or 10 μM 20%PS PLV. Clotting was started by adding 5 mM Ca++ and 1 pM FXIa (intrinsic pathway) and clot time measured by change in turbidity. The impact of platelets vs. PLV on FVIII activity was compared to emicizumab activity. FVIII and Emi were also tested using a defined two-stage Xase assay with PLV or IIa-activated platelets. FIXa, FX, and Ca++ were added and FXa generated was measured using a chromogenic substrate. To test the dissociation of A2, FVIII with platelets or vesicles was preactivated with IIa and then incubated for 0-14 min before Xase assay start. Intrinsic pathway activated FVIII clotting was faster on platelets than on PLV. The relative platelet:PLV FVIII activity varied from approximately 40:1 to 2:1 with the highest ratios achieved with lower fVIII concentrations. In contrast, activity of Emi was similar on platelets and PLV. This indicated that FVIII binding sites primarily protect or enhance FVIII(a) rather than FIXa or FX. Experiments with protein C-deficient plasma demonstrated that the enhanced activity on platelets is not primarily due to preferential cleavage by APC on PLV vs. platelets. We tested FVIII mutants to determine whether platelets protect FVIIIa activity through cleavage of susceptible residues or by preventing dissociation of the A2 domain. FVIII-VV and FVIII-QQVV supported clotting equivalently on platelets and PLV. Whereas FVIII-QQ and FVIII-WT activity remained much greater on platelets vs. PLV. These data suggest that platelet binding sites may protect FVIIIa from spontaneous dissociation over the time course of clotting. In purified FXase assays, activated FVIII-WT activity decay was slower with platelets vs. PLV reaching 50% activity after 6.6 ± 0.3 min versus 3.7 ± 0.2 min. Similar results were seen using FVIII-QQ. Activity of FVIIIa derived from full-length FVIII decayed at the slower rate on both platelets and vesicles, suggesting that the B-domain acts to stabilize FVIIIa. We are currently investigating the effect of the reported slower IIa cleavage of full-length FVIII at residue 1689 as a possible explanation. Further experiments will probe the role(s) of FIXa and FX on decay of B domain deleted FVIII activity on platelets vs PLV.Our data shows that platelet binding sites provide enhanced factor VIII activity compared to phospholipid vesicles and suggests that stabilization of FVIIIa against dissociation contributes to the enhanced activity. These findings are relevant to assays of FVIII activity, particularly in the presence of inhibitory antibodies and to the relative activity of FVIII vs. FVIII mimetics that partially compensate for lack of FVIII activity.