Engineering a protein Z-dependent protease inhibitor (ZPI) mutant as a novel antagonist of ZPI anticoagulant function for hemophilia treatment

Abstract

Background:

Protein Z-dependent protease inhibitor (ZPI), is an important anticoagulant protein in plasma that functions in complex with its cofactor, protein Z (PZ) to rapidly inhibit activated factor X (FXa) on a procoagulant membrane surface. Recent studies suggest that the ZPI-PZ anticoagulant complex is a promising target for restoring hemostasis in hemophilia (Girard TJ, et al. JTH 2019, 17(1):149–156).

Objective:

Engineering a ZPI mutant as a novel antagonist of ZPI anticoagulant function.

Methods:

We engineered two alanine mutations in human ZPI, one in the reactive loop P1 Y387 residue to inactivate the FXa/FXIa inhibitory function, and the second in the K239 binding interface residue to enhance the affinity of the inactive ZPI for PZ. The mutant was expressed, purified and characterized by in vitro and plasma assays.

Results:

The mutant, Y387A/K239A (ZPI-2A), bound PZ >20-fold tighter than WT ZPI or a PZ antibody (PZAb). FXa inhibition assays showed that ZPI-2A effectively neutralized ZPI/PZ anti-FXa activity with a ~3-fold molar excess over WT ZPI whether FXa was bound to FVa in prothrombinase or unbound. Thrombin generation assays in a purified system or in normal/hemophilia plasmas showed that ZPI/PZ activity was reversed by ZPI-2A in a dose-dependent manner, with a 3-fold molar excess sufficient to fully reverse ZPI/PZ inhibition of thrombin generation.

Conclusions:

ZPI-2A is a potent antagonist of ZPI/PZ anticoagulant function, capable of fully blocking the anti-FXa activity of plasma levels of ZPI/PZ at significantly lower doses than a PZAb and thus a promising prophylactic agent for treating hemophilia.

INTRODUCTION

Recent efforts to provide an alternative therapy for the treatment of hemophilia that overcomes the problem of inhibitory antibodies associated with factor VIII/factor IX replacement therapy have focused on downregulating natural anticoagulants to restore hemostasis[1–3]. A potential confounding problem in this approach is the critical importance of these anticoagulants for maintaining hemostasis and other essential anti-inflammatory/cytoprotection functions. A safer recent approach has been to target another anticoagulant protein, protein Z-dependent protease inhibitor (ZPI)[4, 5], which makes an important contribution to hemostasis but whose deficiency only mildly impairs hemostasis[6–8].

ZPI is a serpin superfamily protein rapidly inhibiting activated factor X(FXa) in the presence of protein Z(PZ), procoagulant phospholipids and calcium, or inhibiting activated factor XI(FXIa) in the absence of these cofactors. ZPI circulates in plasma as a tight complex with PZ, a vitamin K-dependent protein, structurally homologous to factors II,VII,IX,X, but lacking protease activity[9, 10]. ZPI-PZ complex inhibition of FXa is dependent on PZ anchoring ZPI to a procoagulant membrane surface through the calcium-dependent association of PZ with the membrane via its N-terminal γ-carboxyglutamic acid (Gla) domain. ZPI binding to the C-terminal pseudocatalytic and EGF2 domains of PZ that extend away from the membrane then positions ZPI to present its reactive loop and P1 bait residue (Y387) to the catalytic-site of membrane-associated FXa and form an inhibitory ZPI-FXa complex through the standard serpin mechanism[11]. We recently reported that the ZPI-PZ complex also inhibits membrane-associated FXa bound to factor Va(FVa) in the prothrombinase complex at a physiologically relevant rate[8].

The importance of ZPI/PZ as regulators of hemostasis is clear from the findings that combined deficiencies of ZPI or PZ with the factor V Leiden mutation (FVL) exacerbate the moderate thrombosis phenotype of FVL in mice and humans[6, 7]. The ZPI-PZ anticoagulant complex could thus be a potential target to restore hemostasis in hemophilia. Consistent with this suggestion, a recent study showed that in a hemophilia mouse, knocking down ZPI/PZ ameliorated the hemophilia bleeding phenotype, equivalent to restoring ~15% FVIII activity, an activity high enough to achieve protection from bleeding for hemophilia patients[4]. Here, we describe a novel approach for downregulating ZPI anticoagulant function as a means of restoring hemostasis in hemophilia.

METHODS

Proteins/lipids.

Plasma-derived human FXa, PZ, and prothrombin were from Enzyme Research Laboratories (South Bend, IN). Human FVa and sheep anti-human PZ polyclonal antibody (PZAb) were from Haematologic Technologies (Essex Junction, VT). Recombinant human WT and Y387A/K239A mutant (ZPI-2A) ZPIs (expressed in E. Coli) and NBD fluorophore labeled K239C were prepared as described [10, 12]. Small unilamellar phospholipid vesicles (SUVs) made from porcine brain extract(PBE) (Polar extract, 18.5% phosphatidylserine (PS)/33% phosphatidylethanolamine(PE)/13% phosphatidylcholine (PC)) or from 20% synthetic dioleoyl PS and 20%PE/60%PC from bovine heart (Avanti, Alabaster, AL) were prepared as described[8, 13, 14].

Experimental conditions.

All experiments were conducted in 50mM Tris buffer, pH 7.4, containing 0.1 M NaCl, 0.1% PEG 8000, 1 mg/ml BSA at 25°C or 37°C.

PZ Binding.

Binding of ZPI-2A, ZPI, or PZAb to PZ was measured by competitive binding titrations in which ZPI/ZPI-2A/PZAb competitors were titrated into a solution of NBD-labeled ZPI-PZ complex and the fluorescence decrease due to competitor displacement of PZ from the labeled ZPI-PZ complex was measured. Observed relative fluorescence changes were fit by the competitive binding equation to obtain K_D for ZPI/ZPI-2A/PZAb interactions with PZ assuming a stoichiometry of 1.0 and independently determined K_D for the labeled ZPI-PZ interaction [10].

ZPI-2A effects on ZPI-PZ inhibition of free and prothrombinase-bound FXa.

Reactions of ZPI-PZ complex with membrane-associated free and prothrombinase-bound FXa in the absence and presence of ZPI-2A were done in the presence of 25μM PBE SUVs and 5mM CaCl₂, with or without FVa, under pseudo-first order conditions in which ZPI-PZ was in large molar excess over FXa (>20-fold). Reactions(100μl) were initiated by adding premixed ZPI/PZ or ZPI/PZ/ZPI-2A to FXa pre-incubated with lipid and calcium with or without FVa, and after varying times quenched with 1 ml Pefafluor FXa substrate (40μM) containing 5mM EDTA to measure residual FXa activity from the initial rate of substrate hydrolysis monitored fluorimetrically(4). The observed pseudo-first order inhibition rate constant (k_obs) was obtained by fitting full progress curves to a single exponential with a nonzero endpoint. Apparent second-order association rate constants (k_a,app) were obtained by dividing k_obs by the ZPI-PZ complex concentration calculated from a measured K_D of ~10nM[10].

Thrombin generation assays.

Mixtures containing ZPI, PZ, ZPI-2A, PZAb as indicated, together with 25μM PBE SUVs, 5mM Ca²⁺, and prothrombin were incubated for 2 min at 37°C, and then FXa/FVa pre-incubated with the same lipid and calcium concentrations were added to initiate the reaction. At different time points, samples were withdrawn and diluted into buffer containing 100μM S2238/10mM EDTA and the thrombin generated was measured from the initial rate of absorbance change by comparing with a standard curve constructed with known thrombin concentrations.

The effect of ZPI, PZ and ZPI-2A on thrombin generation in ZPI/PZ deficient plasma (Hyphen Biomed, France) and pooled FVIII deficient plasma from hemophilia patients (George King, Overland Park, KS) activated with TF(Innovin), was measured at 37°C essentially as described [8].TF activator, ZPI/PZ, ZP-2A, corn trypsin inhibitor(CTI) and SUVs (20%PS/20%PE/60%PC) (40μl) were added to 100μl plasma samples and after 2min incubation at 37°C, 10μl buffered CaCl₂ was added to initiate coagulation. Final concentrations were 2pM TF, 35nM ZPI/PZ, 0–500nM ZPI-2A, 50μg/ml CTI, 0–200μg/ml PZAb, 25μM SUV, 16mM CaCl₂. At different time points, samples were withdrawn and diluted into buffer containing 100μM S2238 and 10mM EDTA and the thrombin generated was measured from the rate of substrate hydrolysis as above.

RESULTS AND DISCUSSION

Our previous structural/functional characterization has shown that the ZPI-PZ interaction is mediated by ionic interactions of D74, D238 and D293 on helix G and helix A of ZPI with the pseudocatalytic domain of PZ, and hydrophobic interactions of Y240 of ZPI with a cavity formed between the PZ EGF2 and pseudocatalytic domains of PZ (Fig. 1A)[15]. D293 and Y240 in particular are hot spots for ZPI binding to PZ and account for most of the ZPI-PZ binding energy. Notably, one of the ZPI contact residues, K239, next to Y240, was found to antagonize the ZPI-PZ interaction. Mutating the wild-type Lys239 to alanine or cysteine residues thus resulted in ZPI mutants that bound PZ over 20-fold tighter than WT ZPI[10]. This enhanced binding appeared to result from the smaller neutral sidechains either removing a steric blockade or neutralizing a repulsive interaction of the positively charged K239 sidechain, thereby improving the critical hydrophobic interaction between ZPI Y240 and PZ.

Figure 1. ZPI K239A/Y387A (ZPI-2A) mutant is a potent antagonist of ZPI/PZ anticoagulant function.

A, Crystal structure of the ZPI (cyan)-PZ (green) complex (PDB:3H5C). The mutated amino acids are in blue sphere. ZPI residues that bind PZ are in red stick. B, Competitive binding of ZPI-2A, WT ZPI and PZAb with NBD-ZPI complexed with PZ. ZPI-2A (open circles), WT ZPI (closed circles) or PZAb (triangles) were titrated into 50nM K239C-NBD ZPI/63nM PZ complex in I 0.15 Tris buffer pH 7.4, 1.5mM CaCl₂, 1 mg/ml BSA at 25°C and the decrease in NBD fluorescence due to displacement of PZ from the labeled ZPI monitored. Solid lines are computer fits of titrations by the competitive binding equation which yielded the reported K_Ds in the text for competitor binding. C, ZPI-2A blocks WT ZPI/PZ inhibition of membrane-associated FXa. Reactions contained ~20nM WT ZPI/PZ complex, 25μM PBE SUVs, 5mM Ca²⁺, 0.5nM FXa, and the indicated concentrations of ZPI-2A in I 0.15 Tris buffer, pH 7.4, 25°C. The solid line is a fit by the competitive binding equation. D, ZPI-2A blocks WT ZPI/PZ inhibition of prothrombinase-bound FXa. Reactions contained ~50 nM WT ZPI/PZ complex, 15nM FVa, 0.5nM FXa, 25μM PBE SUVs, 5mM Ca²⁺ and the indicated concentrations of ZPI-2A in I 0.15 Tris buffer, pH 7.4 at 25°C. The solid line is a fit by the competitive binding equation. E, ZPI-2A or the PZAb reverses WT ZPI-PZ inhibition of prothrombinase activation of prothrombin. Reactions contained ~50nM WT ZPI-PZ complex, 15nM FVa, 0.03nM FXa, 25μM PBE SUVs, 5mM Ca²⁺, 1.4μM prothrombin and the indicated concentrations of ZPI-2A or 200μg/ml PZAb in I 0.15 Tris buffer, pH 7.4 at 37°C. F, ZPI-2A reverses WT ZPI-PZ inhibition of thrombin generation in ZPI/PZ-deficient plasma. Reactions were performed in ZPI/PZ-deficient plasma (100μl) primed with plasma levels of WT ZPI and PZ in the absence and presence of the indicated final concentrations of ZPI-2A in I 0.15 Tris buffer, pH 7.4 at 37°C. Coagulation was activated by adding SUVs (20% PS/20% PE/60%PC), Ca²⁺, CTI (to block contact activation) and TF to a total volume of 150μl. Final concentrations were 1.5-fold dilution citrated plasma, 35nM ZPI, 35nM PZ, 25μM SUVs, 16mM Ca²⁺, 50μg/ml CTI, 2pM TF, 0–500nM ZPI-2A. G, ZPI-2A increases thrombin generation in FVIII-deficient hemophilia plasma. Reactions were performed in FVIII deficient plasma (100μl) in the absence and presence of the indicated concentrations of ZPI-2A(0–100nM), PZAb (0–200μg/ml) or control sheep IgG (Sigma-Aldrich, St. Louis, MO) in I 0.15 Tris buffer, pH 7.4 at 37°C. Coagulation was activated by adding SUVs(20% PS/20% PE/60%PC), Ca²⁺, CTI and TF to a total volume of 150μl. Final concentrations were 1.5-fold dilution citrated plasma, 25μM SUVs, 16mM Ca²⁺, 50μg/ml CTI, 2pM TF. All data represent the average of 3–4 independent measurement ± SD for figure 1B–1G.

This finding suggested that an effective antagonist of the ZPI-PZ interaction might be designed by engineering two mutations in ZPI, one in the reactive loop P1 Y387 residue to inactivate the FXa/FXIa inhibitory function of the serpin, and the second in the K239 binding interface residue to enhance the affinity of the inactive ZPI for PZ(Fig. 1A). We thus substituted both the Y387 and K239 residues with alanines. We expected that the resulting Y387A/K239A ZPI (ZPI-2A) would efficiently compete with and displace WT ZPI from its complex with PZ to result in inactive ZPI-2A-PZ complexes and uncomplexed WT ZPI lacking anticoagulant function.

A series of tests were performed. As expected, ZPI-2A was essentially inactive as a FXa inhibitor in the presence of PZ, lipid and calcium cofactors or as a FXIa inhibitor without cofactors (~0.1% wild-type activity). We next measured the ZPI-2A affinity for PZ compared with WT ZPI and PZAb affinities by a competitive binding fluorescence assay that employs a site-specific NBD-labeled ZPI mutant (K239C) which undergoes a 300–400% increase in NBD fluorescence upon PZ binding(Fig. 1B). Mutant and WT ZPIs or PZAb were titrated into the NBD-labeled ZPI-PZ complex and the displacement of PZ from the labeled ZPI by the competitors was followed from accompanying decreases in NBD fluorescence to an endpoint fluorescence characteristic of free NBD-ZPI [10]. Fitting the data by the competitive binding equation provided K_Ds of 0.4±0.1 nM and 10.0±1.0nM for ZPI-2A and WT ZPI binding to PZ, respectively(Fig. 1B), confirming the expected >20-fold higher affinity of the ZPI-2A mutant than WT ZPI as previously reported for the K239A single ZPI mutant[10]. A K_D of 37±11nM obtained for PZAb binding to PZ indicated that the PZAb was a poorer antagonist of the ZPI-PZ interaction.

The ability of ZPI-2A to antagonize ZPI-PZ complex inhibition of membrane-associated FXa either free or bound to factor Va in prothrombinase was next tested. ZPI-2A dose-dependently decreased the apparent second-order rate constant (k_a,app) for ZPI-PZ complex (20nM) inhibition of FXa (0.5nM) in the presence of procoagulant membrane vesicles and calcium(Fig. 1C), consistent with ZPI-2A competitively displacing wild-type ZPI from its complex with PZ and reversing the ~1000-fold PZ enhancement of the rate of ZPI inhibition of FXa. Notably, k_a,app for ZPI/PZ inhibition of FXa (1.8±0.1×10⁷M⁻¹s⁻¹) was decreased by ~90% (0.26±0.05×10⁷M⁻¹s⁻¹) when a 2.8-fold molar excess of ZPI-2A was added. In the presence of physiologically relevant levels of FVa (15nM) sufficient to saturate FXa[16], k_a,app for ZPI-PZ complex (50nM) inhibition of prothrombinase-bound FXa (0.5nM) was decreased ~4-fold (0.49±0.02×10⁷M⁻¹s⁻¹) compared with ZPI/PZ inhibition of unbound FXa, consistent with our recent report that FVa partially protects FXa from ZPI/PZ inhibition [8]. Again, ZPI-2A dose-dependently decreased k_a,app for ZPI-PZ complex inhibition of FXa bound to FVa in prothrombinase at plasma levels of ZPI/PZ (50nM)(Fig. 1D). In this case, a ~3-fold molar excess of ZPI-2A over wild-type ZPI produced a ~90% decrease in k_a,app for the inhibition of prothrombinase-bound FXa (0.49±0.06×10⁶M⁻¹s⁻¹).

To assess the effectiveness of ZPI-2A as an antagonist of ZPI-PZ inhibition of prothrombinase-bound FXa during prothrombin activation, we performed thrombin generation assays in which prothrombinase activation of plasma concentrations of prothrombin(1.4μM) was monitored in the presence of plasma levels of WT ZPI-PZ complex(~50nM) and increasing concentrations of added ZPI-2A(Fig. 1E). In the absence of ZPI-2A, WT ZPI-PZ complex reduced thrombin generation from prothrombinase activation of prothrombin ~70% to ~400nM thrombin. ZPI-2A dose dependently increased thrombin generation, with 150–250nM ZPI-2A increasing thrombin generation to 80–90% of the level of thrombin generation in the absence of ZPI/PZ, or with ZPI/PZ activity fully blocked by a large molar excess of PZAb. Finally, we performed thrombin generation assays in ZPI/PZ-deficient plasma in the absence or presence of plasma levels of ZPI-PZ complex (n=3, Fig. 1F) as well as in FVIII-deficient hemophilia plasma (n=3, Fig. 1G) in which coagulation was activated with TF, procoagulant membrane vesicles and calcium. ZPI/PZ significantly decreased the thrombin peak (~3-fold) and total thrombin generation (>2-fold) and increased the lag time. This profile was reversed by ZPI-2A in a dose-dependent manner, with a 3-fold molar excess of ZPI-2A over added wild-type ZPI sufficient to fully reverse the decrease in the thrombin peak and total thrombin generated as well as the increase in lag time. In the hemophilia plasma, ZPI-2A as well as the PZAb both dose-dependently enhanced the peak and total thrombin generation and reduced the lag time with a 3-fold molar excess of ZPI-2A over the estimated plasma ZPI achieving a maximal effect.

To ensure that the minor residual activity of the ZPI-2A mutant did not compromise its ability to antagonize wild-type ZPI anticoagulant function, we evaluated a second Y387D/K239A mutant by the same set of assays. This mutant was completely devoid of anti-FXa/FXIa activity (<.001 %) and behaved indistinguishably from the ZPI-2A mutant in these assays (not shown).

In conclusion, our data clearly show that ZPI-2A is a potent antagonist of ZPI/PZ anticoagulant function, readily achieving neutralization of ZPI/PZ activity with a ~3 fold molar excess of mutant over WT ZPI. By contrast, the PZAb requires much higher levels (~1μM) to fully block ZPI-PZ activity[8]. A therapeutic level of ZPI-2A could thus be achieved with as low as 150nM ZPI mutant and result in an upregulation of procoagulant activity in hemophilia patients equivalent to administering ~15% FVIII based on the reported effect of knocking out ZPI or PZ in hemophilic mice [4]. Although as a variant of a natural inhibitor, immunogenicity and consequent crossreactivity might be a potential issue, these results nevertheless provide compelling evidence that the ZPI-2A mutant could be an effective anti-hemophilia prophylactic agent that is potentially capable of providing a much safer therapy for hemophilia than therapies that target other more essential anticoagulants.

Essentials.

• ZPI is a newly identified target to suppress anticoagulant activity in hemophilia.

• A novel inactive ZPI double mutant (Y387A/K239A (ZPI-2A)) with high PZ affinity was engineered.

• In vitro and plasma assays showed ZPI-2A is a potent antagonist of ZPI anticoagulant function.