Stabilizing Factor VIIIa By Combining Mutations That Modulate Inter-Subunit Interactions and Proteolytic Inactivation

Factor (F) VIIIa serves as a cofactor for FIXa forming intrinsic FXase complex. The activity of FVIIIa is labile due to the tendency for A2 subunit of the cofactor to dissociate, thereby inactivating FVIIIa and dampening FXase activity. Recently we have shown that a combination of hydrophobic point mutations generated FVIII(a) variants with improved FVIII thermal stability and enhanced A2 subunit retention in FVIIIa (Wakabayashi et al. J. Thromb. Haemost. 2012, 10: 492-495). Furthermore, FVIIIa is a target for proteolytic inactivation catalyzed by FXa, the product of FXase complex. We have shown this proteolysis further contributes to the down-regulation of FXase and that altering the sequence flanking the primary FXa cleavage site in FVIIIa (at Arg336) yields marked reductions in the rates of proteolytic inactivation of FVIIIa catalyzed by FXa (DeAngelis et al. J. Biol. Chem. 2012, 287: 15409-15417). In this study we prepared various combinations of the above mutations to obtain a panel of novel FVIII molecules and examined the attributes of these reagents in assays monitoring FVIII thermal stability, FVIIIa spontaneous decay rates, rates of proteolytic inactivation by FXa, and thrombin generation potential.

The mutants we prepared include 336(P4-P3’)562 [where residues 333-339 (PQLRMKN) that flank the fast FXa-cleavage site at Arg336 are replaced with residues flanking the slow cleavage site at Arg562 (residues 559-565, VDQRGNQ)] are combined with D519V/E665V (336-V) or D519/VE665V/A108I (336-VI). Specific activity values for the combined mutation variants ranged from 85% to 170% the WT value using a one-stage clotting assay and 95% to 110% using a two-stage chromogenic assay. FVIII thermal stability was tested by monitoring FVIII activity remaining during incubation at 57ºC over a 20 min time course. Thermal decay rates for 336-V and 336-VI variants were reduced by ∼1.7 and ∼5-fold as compared with the WT FVIII value. These rate values reflected additive effects of the individual mutations since rate values for controls FVIII D519V/E665V and FVIII A108I were reduced ∼1.7, and ∼3-fold, respectively, relative to WT, while the thermal decay rate for FVIII 336(P4-P3’)562 was WT-like. FVIIIa spontaneous decay rates were determined following activation of FVIII by thrombin and these values were reduced by ∼25-fold for the 336-V and 336-VI variants as compared with the WT FVIIIa value. Interestingly, this magnitude of rate reduction suggested a synergistic effect since rate values were reduced ∼14-fold for the D519V/E665V control and were essentially unaffected in the 336(P4-P3’)562 and A108I controls. FVIIIa inactivation by FXa was monitored by a one-stage clotting assay after FVIIIa was incubated with 5 nM FXa at 37ºC for a 30 min time course. FVIII 336-V and 336-VI variants showed similar resistance to inactivation by FXa (∼10-fold reduced rate compared with WT FVIIIa) as the 336(P4-P3’)562 control. FVIII D519V/E665V and A108I variants showed slightly reduced inactivation rates (∼1.1 and 1.6 fold) as compared with WT FVIII. Thrombin generation assays were performed using FVIII deficient plasma. Assays were run using 0.25 nM FVIII and 4 µM phospholipid vesicles, initiated with 0.25 pM tissue factor, and the amount of generated thrombin was calculated overtime by monitoring the development of fluorescent thrombin substrate peptide. FVIII 336-V and 336-VI variants generated comparable amounts of thrombin as FVIII D519V/E665V, showing ∼70% increases in endogenous thrombin potential (ETP), while the A108I and 336(P4-P3’)562 controls showed WT-like ETP values. Taken together, these results indicate that it is possible to combine the above gain-of-function FVIII mutations to yield FVIII variants such as the 336-VI form in order to generate a more stable procofactor as judged by improved thermal stability (∼5-fold relative to WT), enhanced retention of A2 subunit increasing FVIIIa stability (∼25 fold) and increased resistance to proteolytic inactivation (∼10 fold). The latter two attributes would potentially prolong FXase activity during clotting and this effect is suggested by the improved thrombin generation parameters for this variant.