Selective Alteration of FXa Zymogenicity Provides a Dynamic Range of Variants that Improve Hemostasis in Hemophilia

Abstract 201

There is significant interest in developing safe and effective procoagulant hemostatic molecules to treat bleeding disorders. Recently, our laboratory has generated a novel FXa variant (e.g. FXa^I16L) with zymogen-like properties that shows promise in bypassing deficiencies upstream of the common pathway. FXa^I16L is rendered partially inactive due to a defect in transitioning from zymogen to protease where residue 16 is critical to this process. However, the biological activity of FXa^I16L is rescued once its cofactor FVa is made available following coagulation activation. While FXa^I16L is effective in correcting the hemophilia phenotype both in vitro and in vivo, the mutation site and general mechanism of action indicates that there is inherent flexibility in this new class of bypassing agent. To this end, we extended the mutational framework at positions 16 and 17 and created a series of zymogenic mutants exhibiting a broad range of activities. We surmised that depending on the amino acid at these positions, the transition from zymogen to protease will be altered to varying degrees. Following an initial screen of ten variants, five were pursued which fell into three categories relative to wild-type (wt)-FXa: “zymogen-like 1” (<1% activity), “zymogen-like 2” (∼3% activity; similar to FXa^I16L with ∼5% activity) and “zymogen-like 3” (∼20% activity). Kinetic studies revealed that all mutants have an impaired ability to bind a small chromogenic substrate at the active site with zymogen-like 1 variants having the weakest affinity. Since physiological inhibitors of FXa predominantly target the active site, half-lives of the variants should correlate with zymogenicity. In line with this, zymogen-like 1 variants exhibited the longest half-life (3–4 hr) in hemophilia B (HB) plasma, followed by zymogen-like 2 (1–2 hr) and zymogen-like 3 (0.25 hr); these are all much greater than wt-FXa (<1–2 min). Assessment of biological activity in HB plasma by aPTT revealed that while wt-FXa (0.1 nM) normalized the prolonged clotting time, the zymogen-like variant 3 was ∼75% effective and zymogen-like 2 derivatives were ∼50% effective. Surprisingly, even the zymogen-like 1 variants with almost undetectable chromogenic activity, shortened the prolonged aPTT (<20% effective). However, the apparent activities of each zymogenic mutant were greatly improved when saturating amounts of FVa were added, indicating that suboptimal amounts of the cofactor are generated during the clotting assay. Furthermore, in a thrombin generation assay using HB plasma each zymogenic variant (0.1 nM) yielded a robust IIa signal with endogenous IIa potential (ETP) and peak heights comparable to wt-FXa. Only the zymogen-like 1 variant had a prolonged time to peak thrombin (∼25 min vs. ∼15 for all other variants/wt-FXa) indicating a delay in thrombin generation. We speculate this reflects the time needed to increase the FVa concentration during the assay, as variants with enhanced zymogen-like character require higher concentrations of cofactor to rescue catalytic function. Next, we evaluated whether these in vitro parameters had a correspondingly relevant impact on the hemophilic phenotype in vivo. In a tail clip assay two different experiments were conducted in which proteins were infused into HB mice (n = 3–6; 450 μg/kg) either 2 min after or 5 min prior to injury. In either experiment, total blood loss in HB mice treated with zymogen-like 1 or zymogen-like 2 variants was significantly reduced compared to PBS treated (n = 5) animals. Furthermore, in a third experiment (n = 3–6; 450 μg/kg) in which proteins were infused 30 min prior to injury only zymogen-like 1 variants were effective in reducing blood loss. Counterintuitively, the ‘higher' activity zymogen-like 3 variant was not particularly effective in the tail clip model. Collectively these in vivo data show that the more active the protease the more difficult it is to overcome the protective mechanism of circulating inhibitors (e.g. antithrombin III) to achieve a therapeutic benefit. By evading interactions with inhibitors and imparting a prolonged half-life, the new zymogenic variants may have a sustained impact on improving hemostasis in hemophilia. We conclude that the plasticity of this new class of FXa variants provides a useful and flexible platform to selectively bioengineer activity and half-life.