Prohemostatic proteins for the treatment of bleeding

Abstract

This disclosure relates to recombinant FXa polypeptides that can be used as antidotes to completely or partially reverse an anti-coagulant effect of a coagulation inhibitor in a subject, preferably a direct factor Xa inhibitor. Disclosed herein are recombinant factor Xa proteins and a method of completely or partially reversing an anti-coagulant effect of a coagulation inhibitor in a subject.

Claims

1. A recombinant protein comprising a coagulation factor Xa polypeptide, said coagulation factor Xa polypeptide having an alteration or deletion of an amino acid residue corresponding to amino acid residue Phe-396 as indicated in SEQ ID NO: 1.

2. The protein according to claim 1, wherein the alteration or deletion of an amino acid residue corresponding to amino acid residue Phe-396 is combined with an insertion of 1-50 amino acid residue(s) in a region corresponding to the region of amino acid residues between Gly-289 and Asp-320.

3. The protein according to claim 1, wherein the alteration or deletion of an amino acid residue corresponding to amino acid residue Phe-396 is combined with an insertion of 1-50 amino acid residue(s) in a region corresponding to the region of amino acid residues between His-311 and Asp-320 of SEQ ID NO: 1.

4. The protein according to claim 3, wherein the region of amino acid residues corresponding to amino acid residues between His-311 and Asp-320 of SEQ ID NO: 1 has the amino acid sequence of SEQ ID NO: 4, SEQ ID NO: 5, SEQ ID NO: 9, SEQ ID NO: 10, or SEQ ID NO: 11.

5. A pharmaceutical composition comprising the protein of claim 1, or a pharmaceutically acceptable salt thereof, and a pharmaceutically acceptable carrier.

6. A method of completely or partially reversing an anti-coagulant effect of a direct factor Xa inhibitor in a subject in need thereof, the method comprising: administering to the subject a therapeutically effective amount of the protein of claim 1 so as to completely or partially reverse the anti-coagulant effect of the direct factor Xa inhibitor in the subject.

7. The method according to claim 6, wherein the direct factor Xa inhibitor is rivaroxaban (5-chloro-N-[[(5S)-2-oxo-3-[4-(3-oxo-4-morpholinyl)phenyl]-5-oxazolidinyl]methyl]-2-thiophenecarboxamide), apixaban (1-(4methoxyphenyl)-7-oxo-6-[4-(2-oxopiperidin-1-yl)phenyl]-4,5,6,7-tetrahydro-1H-pyrazolo[3,4-c]pyridine-3-carboxamide), edoxaban (N′-(5-chloropyridin-2-yl)-N-[(1S,2R,4S)-4-(dimethylcarbamoyl)-2-[(5-methyl-6,7-dihydro-4H-[1,3]thiazolo[5,4-c]pyridine-2-carbonyl)amino]cyclohexyl]oxamide; 4-methylbenzenesulfonic acid), or betrixaban (N-(5-chloropyridin-2-yl)-2-[[4-(N,N-dimethylcarbamimidoyl)benzoyl]amino]-5-methoxybenzamide).

8. A method of making the protein of claim 1, the method comprising: expressing a nucleic acid molecule comprising a DNA sequence that encodes the protein in a host cell.

9. The method according to claim 8, wherein the nucleic acid molecule is comprised within an expression vector.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1: Blood coagulation FXa structure. Panel A: Schematic γ-carboxyglutamic (“GLA”) EGF-1 and -2 (“EGF”), and serine protease domain (“SP”) structure of coagulation FXa. Panel B: Crystal structure of the human FXa serine protease (pdb 2W26). Indicated are the catalytic triad His-276, Asp-322, Ser-419, rivaroxaban/apixaban contact residues Try-319 and Phe-396, position of residues 316-317 (in spheres), and residues Gly-289, Glu-297, Val-305, and His-311. Panel C: Alignment of region 311-322 in various plasma FX species with conserved residues (highlighted), contact residue Tyr-319, and catalytic residue Asp-322 indicated (SEQ ID NOS. 1 and 29-36). *Indicates venom coagulation FX with insertion in the region corresponding to the region between Gly-289 and Asp-320 of SEQ ID NO:1.

(2) FIGS. 2A-2C: Inhibition of chromogenic FXa activity by direct FXa inhibitors. FIG. 2A: Peptidyl substrate conversion (SpecFXa, 250 riM) by recombinant human coagulation FXa (hFXa, 2 nM, circles) or venom P. textilis coagulation FXa (vptFXa, 10 nM, triangles) in the presence of increasing concentrations (1 nM-10 μM) rivaroxaban (“riva,” closed symbols) or apixaban (“api,” open symbols). Substrate conversion is plotted as the % of incubations in the absence of inhibitor. FIGS. 2B and 2C: Thrombin generation in coagulation FX-deficient plasma was initiated with 0.5 nM hFXa (FIG. 2B) or vptFXa (FIG. 2C) in the absence (grey line/grey column) or presence of 0.4 μM rivaroxaban (“riva,” black line/black column) or 2 μM apixaban (“api,” dotted line/white column). Thrombin formation was assessed using a fluorogenic substrate and peak thrombin concentrations of the various incubations are shown in the insets.

(3) FIGS. 3A and 3B: FIG. 3A: Fluorescent Western blot of recombinant FX (200 ng) obtained from HEK293 cell lines that stably express either recombinant human FX (r-hFX, lane 1, 5), modified human FX-A (mod A, lane 2, 6) or modified human FX-B (mod B, lane 3, 7), before (lanes 1, 2, 3) or after (lanes 5, 6, 7) incubation with RVV-X activator. The heavy chain of endogenous plasma-derived human FXa migrates at ˜29 kDa (lane 9). Relative weight (kDa) of the protein markers (lanes 4, 8) are indicated. FIG. 3B: Recombinant FX in conditioned media from HEK293 cell lines stably expressing either recombinant human FX (black column), modified human FX-A (white column) or modified human FX-B (grey column) was quantified using an FX-specific ELISA. Each individual bar represents a single stable cell line with the highest attainable expression per FX variant.

(4) FIG. 4: Macromolecular substrate activation. Prothrombin conversion (1.4 μM) in the presence of 50 μM PCPS, 20 nM FV (FV810, recombinant B-domain truncated FV) and 0.1 nM of modified human coagulation FXa type A (m-hFXa A), type B (m-hFXa B), recombinant (r-hFXa), or plasma-derived (pd-hFXa) FXa. The substrate conversion is plotted in nM/min/nM Enzyme and data are the mean value of two independent experiments+S.D.

(5) FIGS. 5A and 5B: Inhibition of FXa chimer type-A by DFXIs. Peptidyl substrate conversion (SpecFXa, 250 μM) of RVV-X activated modified human coagulation FXa type A (m-hFXa A, 1 nM) in comparison to recombinant human coagulation FXa (r-hFXa, 3 nM), plasma-derived human coagulation FXa (pd-FXa, 2 nM), and venom P. textilis (vptFXa, 1 nM) FXa. Conversion rates were determined in the presence of 0.001-100 μM of Rivaroxaban (FIG. 5A) or Apixaban (FIG. 5B). The data represent the mean value of two independent experiments, except for r-hFXa (n=1).

(6) FIGS. 6A and 6B: Inhibition of modified human FX-A and modified human FX-B by DFXIs. Peptidyl substrate conversion (SpecFXa, 250 μM) by RVV-X activated modified human FX-A (m-hFXa A, 1 nM) and modified human FX-B (m-hFXa B, 7 nM, in comparison to RVV-X-activated recombinant human coagulation FXa (r-hFXa, 6 nM). Conversion rates were determined in the presence of 0.001-100 μM of rivaroxaban (FIG. 6A) and apixaban (FIG. 6B). The data are the means of two independent experiments.

(7) FIGS. 7A and 7B: Inhibition of modified human FX-A or modified human FX-B by DFXIs in the presence of cofactor Va and phospholipids. Peptidyl substrate conversion (SpecFXa, 250 μM) by RVV-X activated modified human FX-A (m-hFXa A, 2 nM) and activated modified human FX-B (m-hFXa B, 4 nM) in comparison to by RVV-X activated recombinant human coagulation FXa (r-hFXa, 3 nM), in the presence of 50 μM PCPS and 30 nM FV (FV810, recombinant B-domain truncated). Conversion rates were determined in the presence of 0.001-100 μM of Rivaroxaban (FIG. 7A) or Apixaban (FIG. 7B). The data are the means of two independent experiments.

(8) FIGS. 8A-8C: Multiple alignment of coagulation FX proteins of different species. The amino acid sequence of human coagulation FX (Genbank Accession No.: AAH46125.1) (HUM) (SEQ ID NO:1) is compared to the amino acid sequences of M. musculus coagulation FX (Genbank Accession No.: AAC36345.1) (MUS) (SEQ ID NO:29), X. tropicalis coagulation FX (Genbank Accesion No.: NP_001015728) (Xtr) (SEQ ID NO:30), D. rerio coagulation FX (Genbank Accession No.: AAM88343.1) (Dre) (SEQ ID NO:31), T. rubripes coagulation FX (Genbank Accession No.: NP_001027783.1) (Tru) (SEQ ID NO:32), P. textilis coagulation FX isoform 1 (UniprotKB accession No.: Q1L659) (Pte1) (SEQ ID NO:33), P. textilis coagulation FX isoform 2 (UniprotKB accession No.: Q1L658) (Pte2) (SEQ ID NO:34), P. textilis coagulation FX (pseutarin C catalytic subunit precursor; Genbank Accession No.: AAP86642.1) (Pte3) (SEQ ID NO:35) and N. scutatus coagulation FX (UniProtKB accession No.: P82807.2) (Nsc) (SEQ ID NO:36). In these figures, Gly-289, Asp-320, Tyr-319, Glu-297, Val-305 and His-311 of SEQ ID NO:1 are indicated in bold and are underlined. These figures show that there is variation in the region of amino acid residues corresponding to the region between Gly-289 and Asp-320 of SEQ ID NO:1 between coagulation FX proteins of different species. Amino acid residues that are conserved in all species are indicated in the consensus sequence.

(9) FIG. 9: Amino acid composition of endogenous hFX and chimeric FX variants. Serine protease domain residues Histidine91 and Tyrosine99 (chymotrypsin numbering; corresponding to His 311 and Tyrosine 319, respectively, of FX as depicted in SEQ ID NO: 1) of endogenous human (hFX) in alignment with chimeric FX type A (c-FX A, middle; sequence between His 311 and Asp 320 corresponds to SEQ ID NO:9), type B (c-FX B; sequence between His 311 and Asp 320 corresponds to SEQ ID NO:10), and type C (c-FX C; sequence between His 311 and Asp 320 corresponds to SEQ ID NO: 11).

(10) FIG. 10: Characterization of FXa: Panel A: Coomassie staining of 5 μg FXa variants on 4-12% Bis-Tris gels. From left to right: plasma-derived Factor Xa (pd-FXa), r-hFXa, chimeric factor Xa type A, B and C (-A, -B, -C). Panel B: Prothrombin conversion (1.4 μM) in the presence of 50 μM PCPS (75% phosphatidylcholine, 25% phosphatidylserine) and 20 nM FV (FV810, recombinant B-domain truncated FV) and 0.1 nM of pd-FXa, r-hFXa, c-FXa-A, c-FXa-B and c-FXa-C. Data points are the mean value of two independent experiments.

(11) FIG. 11: Inhibition of FXa variants by DOACs. Normalized prothrombin conversion by 1 nM of pd-FXa (triangles), r-hFXa (circles), chimeric FXa-A (squares), -B (diamonds) and -C (crosses) was assessed in the presence of 0.001-100 μM of Apixaban (left, closed symbols) or Edoxaban (right, open symbols). Inhibitory constants (determined with Graphpad Prism 6 software suite) of Apixaban for pd-FXa: 2 nM, r-hFXa: 4 nM, c-FXa-A: 130 nM, -B: 760 nM-C: 1270 nM and of Edoxaban for r-hFXa: 0.5 nM, c-FXa-A: 3 nM, -B: 140 nM-C: 270 nM.

(12) FIGS. 12A and 12B: FXa-initiated thrombin generation (TG) profiles for FXa variants. Plasma TG in the absence (FIG. 12A) and presence (FIG. 12B) of DOAC Apixaban (2 μM). Initiation of TG by pd-FXa, r-hFXa, c-FXa-A, c-FXa-B and c-FXa-C in FX-depleted plasma. Curves are the average of at least three independent experiments.

(13) FIGS. 13A and 13B: Tissue factor (TF)-initiated TG profile for r-hFX and c-FX-C. FIG. 13A: Plasma TG at low TF (2 μM) in the absence and presence of 2 μM DOAC Apixaban (Apixa) by 1 unit r-hFX, r-hFX plus Apixaban, c-FXa-C or c-FXa-C plus Apixaban. One unit of r-hFX (7 μg/ml) or c-FXa-C (16 μg/ml) was defined by a prothrombin time-based clotting assay using normal human plasma as reference. Curves represent the average of at least three independent experiments. FIG. 13B: Plasma TG at high TF (20 μM).

(14) FIGS. 14A and 14B: TF-initiated TG profile for r-hFX and c-FX-C. (Upper graph): Plasma TG at low TF (2 μM) in the absence (dotted line) and presence of 200 nM (light grey), 600 nM (dark grey) and 2000 nM (black) DOAC Edoxaban by 1 unit r-hFX (7 μg/ml). (Lower graph): Plasma TG at low TF (2 μM) with similar concentrations of Edoxaban by 1 unit of c-FXa-C (16 μg/ml). Curves represent the average of two independent experiments.

(15) TABLE-US-00002 TABLE 1 Side-chain Side-chain charge Hydropathy Absorbance ε at λmax (×10.sup.−3 M.sup.−1 Amino Acid 3-Letter.sup.(114) 1-Letter.sup.(114) polarity.sup.(114) (pH 7.4).sup.(114) Index.sup.(115) λmax(nm).sup.(116) cm.sup.−1).sup.(116) Alanine Ala A nonpolar neutral 1.8 Arginine Arg R Basic polar positive −4.5 Asparagine Asn N polar neutral −3.5 Aspartic acid Asp D acidic polar negative −3.5 Cysteine Cys C nonpolar neutral 2.5 250 0.3 Glutamic acid Glu E acidic polar negative −3.5 Glutamine Gln Q polar neutral −3.5 Glycine Gly G nonpolar neutral −0.4 Histidine His H Basic polar positive (10%) −3.2 211 5.9 neutral (90%) Isoleucine Ile I nonpolar neutral 4.5 Leucine Leu L nonpolar neutral 3.8 Lysine Lys K Basic polar positive −3.9 Methionine Met M nonpolar neutral 1.9 Phenylalanine Phe F nonpolar neutral 2.8 257, 200, 188 0.2, 9.3, 00.0 Proline Pro P nonpolar neutral −1.6 Serine Ser S polar neutral −0.8 Threonine Thr T polar neutral −0.7 Tryptophan Trp W nonpolar neutral −0.9 280, 219 5.6, 47.0 Tyrosine Tyr Y polar neutral −1.3 274, 222, 193 1.4, 8.0, 48.0 Valine Val V nonpolar neutral 4.2

DETAILED DESCRIPTION

Examples

Example 1

(16) Materials and Methods

(17) Rivaroxaban and Apixaban were obtained from Alsachim (Illkirch, France) and dissolved in DMSO (˜30 mg/ml). The peptidyl substrate methoxycarbonylcyclohexylglycylglycylarginine-p-nitroanilide (Spec-Xa) was obtained from Sekisui Diagnostics (Stamford, Conn., USA). All tissue culture reagents were from Life Technologies (Carlsbad, Calif.), except insulin-transferrin-sodium selenite (ITS), which was from Roche (Basel, Switzerland). Small unilamellar phospholipid vesicles (PCPS) composed of 75% (w/w) hen egg L-phosphatidylcholine and 25% (w/w) porcine brain L-phosphatidylserine (Avanti Polar Lipids, Alabaster, Ala.) were prepared and characterized as described previously (Higgins et al., J. Biol. Chem. 1983, 258:6503-6508). FX-depleted human plasma was obtained from Diagnostica Stago (Paris, France). All functional assays were performed in HEPES buffered Saline (20 mM Hepes, 0.15 M NaCl, pH 7.5) supplemented with 5 mM CaCl2 and 0.1% polyethylene glycol 8000 (assay buffer). Mammalian expression vector pCMV4 (Andersson et al., J. Biol. Chem. 1989, 264:8222-8229, carrying recombinant human FX (r-hFX) was a generous gift from Rodney M. Camire (Camire et al., Biochemistry 2000, 39:14322-14329). The pcDNA3 vector was obtained from Invitrogen and the PACE cDNA was a generous gift from Genetics Institute, Boston, Mass. A vector carrying Furin proprotein convertase has been described (U.S. Pat. No. 5,460,950).

(18) Human recombinant Factor V (FV) was prepared, purified, and characterized as described previously (Bos et al., Blood 2009, 114:686-692). Recombinant P. textilis venom FXa (vpt-FXa) was prepared, purified, and characterized as described previously (Verhoef et al., Toxin Reviews (2013) (doi:10.3109/15569543.2013.844712). Plasma-derived human Factor Xa (pd-hFXa), DAPA, human prothrombin and Anti-Human Factor X monoclonal mouse IgG (AHX-5050) were from Haematologic Technologies (Essex Junction, Vt., USA). FX antigen paired antibodies for ELISA were obtained from Cedarlane (Burlington, Canada). RVV-X activator was obtained from Diagnostica Stago (Paris, France), or Haematologic Technologies. Restriction endonuclease Apal was obtained from New England Biolabs (Ipswich, Mass., USA). T4-DNA ligase was obtained from Roche (Roche Applied Science, Indianapolis, Ind., USA).

(19) The DNA sequence encoding modified human FX-A is provided as SEQ ID NO:7. The DNA sequence encoding modified human FX-B is provided as SEQ ID NO:8. Nucleotides encoding SEQ ID NO:4 (to generate modified human FX-A) or SEQ ID NO:5 (to generate modified human FX-B) sequences flanked by Apal restriction sites were synthesized by Genscript (Piscataway, N.J., USA), subcloned into pCMV4 mammalian expression vector using Apal and T4-DNA ligase and sequenced for consistency. Modified human FX-A and modified human FX-B are also referred to as mod-hFX-A and mod-hFX-B, respectively. Stable HEK293 cell lines expressing r-hFX or modified hFX were obtained as described previously (Larson et al., Biochemistry 1998, 37:5029-5038). HEK293 cells were cotransfected with pCMV4 and pcDNA-PACE vectors using Lipofectamine2000 according to the manufacturer's instructions. FX expression of transfectants was assessed by a modified one-step clotting assay using FX-depleted human plasma. Transfectants with the highest expression levels were expanded into T175 culture flasks and conditioned for 24 hours on expression media (DMEM-F12 nutrient mixture without Phenol-red supplemented with: Penicillin/Streptomycin/Fungizone, 2 mM L-glutamine, 10 μg/ml ITS, 100 μg/ml Geneticin-418 sulphate and 6 μg/ml vitamine K). Conditioned media was collected, centrifuged at 10,000 g to remove cellular debris, concentrated in a 10-kDa cut-off filter (Millipore, Darmstadt, Germany), washed with HEPES-buffered saline and stored in 50% glycerol at −20° C. FX antigen levels of glycerol stocks were assessed by sandwich ELISA according to the manufacturer's instructions using human pooled plasma as reference, assuming a plasma FX concentration of 10 μg/ml.

(20) Expression media was conditioned for 24 hours on stable cell lines expressing either r-hFX, modified human FX-A or modified human FX-B. An aliquot of conditioned media was incubated with RVV-X (10 ng/l; Haematologic Technologies) for 120 minutes at 37° C. After activation, modified human FX-A or modified human FX-B are also referred to as m-hFXa A or m-hFXa B, respectively. Assuming similar substrate affinities for all FXa variants, the concentration of FXa in media was subsequently determined by peptidyl substrate conversion (Spec-Xa, 250 μM) using known concentrations of pd-hFXa as reference. Steady-state initial velocities of macromolecular substrate cleavage were determined discontinuously at 25° C. as described (Camire, J. Biol. Chem. 2002, 277:37863-70). Briefly, progress curves of prothrombin activation were obtained by incubating PCPS (50 μM), DAPA (10 μM), and prothrombin (1.4 μM) with human recombinant FV-810 (B-domain truncated, constitutively active), and the reaction was initiated with either 0.1 nM of pd-hFXa, r-hFXa, m-hFXa B, or 0.033 nM of m-hFXa A. The rate of prothrombin conversion was measured as described (Krishnaswamy et al., Biochemistry 1997, 36:3319-3330).

(21) Recombinant FX and modified human FX-A and modified human FX-B (200 ng) were activated by RVV-X (0.5 U/ml) for 60 minutes at 37° C. and subjected to electrophoresis under reducing (30 mM dithiothreitol) conditions using pre-cast 4-12% gradient gels and the MES buffer system (Life Technologies) and transferred to a nitrocellulose membrane using the Trans-Blot Turbo Transfer System (Bio-Rad Laboratories, Hercules, Calif., USA). The blot was probed with an anti-heavy chain FX antibody and protein bands were visualized using a Dyelight-800 anti-mouse fluorescent antibody (Thermo Scientific, Rockford, Ill. USA). Plasma-derived hFXa (200 ng) was used as a reference.

(22) Thrombin generation was adapted from protocols earlier described (Hemker et al., Pathophysiol. Haemost. Thromb. 2003, 33:4-15). Briefly, FX-depleted plasma was mixed with Corn Trypsin Inhibitor (70 μg/ml), buffer (25 mM HEPES, 175 mM NaCl, 5 mg/ml BSA, pH 7.5) and PCPS (20 μM) and incubated for 10 minutes at 37° C. in a 96-well microplate. Thrombin formation was initiated by addition of pd-hFXa (0.5 nM) or vpt-FXa (0.5 nM) preincubated with Rivaroxaban (0.4 μM) or Apixaban (0.2 μM), supplemented with FluCa and immediately transferred to the plasma mix. The final reaction volume was 120 μl, of which 64 μl was FX-depleted plasma. Thrombin formation was determined every 20 seconds for 30 minutes and corrected for the calibrator using a software suite (Thrombinoscope, version 5.0). The mean endogenous thrombin potential (the area under the thrombin generation curve) was calculated from at least two individual experiments. Calibrator and fluorescent substrate (FluCa) were purchased from Thrombinoscope (Maastricht, The Netherlands).

(23) Peptidyl substrate conversion (Spec-Xa, 250 μM final) of each FXa variant was performed in the absence or presence of direct FXa inhibitors Rivaroxaban and Apixaban (0.001 μM-100 μM final) at ambient temperature. Calcium-free stocks of pd-hFXa (2 nM final) or vpt-FXa (10 nM final) were diluted in assay buffer and incubated in a 96-well microplate in the presence of assay buffer or inhibitor for 2 minutes. Substrate conversion was initiated with Spec-Xa and absorption was monitored for 10 minutes at 405 nM in a SpectraMax M2e microplate reader equipped with the Softmax Pro software suite (Molecular Devices, Sunnyvale, Calif., USA). In order to assay DFXI sensitivity of each recombinant FX variant, glycerol stocks (5-40 μl) of r-hFX, modified human FX-A and modified human FX-B were diluted in assay buffer and incubated with RVV-X (0.5 U/ml) for 60 minutes at 37° C. Activated stocks were subsequently diluted in assay-buffer, incubated for 2 minutes in a 96-well microplate in the presence of assay buffer or inhibitor and assayed for substrate conversion as described. The relative concentration of rhFX, m-hFXa A and m-hFXa B was assessed from the rate of substrate conversion in the absence of inhibitor using known concentrations of pd-hFXa as reference.

(24) Results

(25) Venom-Derived P. textilis (Vpt)-FXa is Resistant to Inhibition by DFXIs

(26) Biochemical characterization of purified recombinant venom-derived P. textilis FXa (vptFXa) revealed that this protease, unlike any other FXa species known to date, is resistant to inhibition by the direct anticoagulants rivaroxaban and apixaban, which have been designed to reversibly block the active site of FXa. Consistent with previous observations, the Ki for human FXa (hFXa) inhibition was approximately 1 nM (Perzborn, J. Thromb. Haemost. 2005, 3:514-521), whereas vptFXa inhibition was at least a 1000-fold reduced (FIG. 2A). These findings were corroborated in a plasma system mimicking in vivo fibrin generation, demonstrating that physiological concentrations of the FXa inhibitors hardly affected vptFXa-initiated thrombin formation, while a significant reduction was observed with hFXa present (FIGS. 2B and 2C).

(27) Human-Venom P. textilis FXa Chimeras

(28) A striking structural element that is not only limited to vptFXa, but also present in venom FX from the Australian snake Notechis scutatus, is an altered amino acid composition at a position close to the hFXa active site (FIG. 1, Panel C). Given its location, it was hypothesized that this unique helix may not only modulate the interaction with rivaroxaban and/or apixaban, but also with FVa, as the FVa binding site is C-terminal to this helix (Lee et al., J. Thromb. Haemost. 2011, 9:2123-2126). To test this hypothesis, the two-protein coding DNA constructs as listed in SEQ ID NOS:7 and 8 were prepared. The mod-hFX-A chimera as provided in SEQ ID NO:7 comprises the relevant part of the N. scutatus DNA sequence (indicated in bold and underlined) and the mod-hFX-B chimera as provided in SEQ ID NO:8 comprises the relevant part of the P. textilis sequence (indicated in bold and underlined).

(29) Using these DNA constructs, HEK293 cell lines were generated that stably produced both chimeric proteins and subsequently assessed the expression levels of modified human FX from HEK293 cells by conditioning the cells on expression media for 24 hours. Western blot analysis revealed expression of full-length FX for both chimeric variants similar to wild-type FX (FIG. 3A). Incubation with activator from Russell's Viper Venom (RVV-X) resulted in proteolytic activation of approximately 30% of zymogen FX to FXa, indicated by the appearance of the ˜29 kDa heavy chain band. The heavy chain of both modified human FXa-A and modified human FXa-B migrated at a slightly higher molecular weight, which is consistent with the insertion of a snake sequence that is 12 or 13 residues longer as compared to that of human FXa, respectively. Analysis of the FX antigen levels in conditioned media indicated that whereas the expression of mod-hFX-A was approximately seven-fold reduced, that of mod-hFX-B was similar to wild-type human FX (FIG. 3B). The low FX antigen levels of mod-hFX-A correlated with the similarly low FX activity levels observed employing a modified clotting assay. This indicates that while the protein expression of mod-hFX-A is suboptimal as compared to that of the other FX variants, its FX function is not perturbed.

(30) To test zymogen activation of FX, rFX and modified human FX-A and modified human FX-B was converted to FXa using FX activator from Russell's Viper Venom (RVV-X). Both modified human FXa-A and modified human FXa-B displayed protease activity upon RVV-X activation, as assessed by conversion of the small FXa-specific peptidyl substrate SpectroZyme Xa. In addition, the prothrombin conversion rates in the presence of the human cofactor FVa of both chimeras were similar to human FXa (both pd-hFXa and r-hFXa) (FIG. 4). Collectively, these observations suggest that the snake sequence insertions do not severely hamper the enzymatic properties of human FX.

(31) Inhibition of FXa Chimeras by DFXIs

(32) To estimate the inhibitory constant (Ki) of Rivaroxaban and Apixaban for RVV-X activated modified human FX-A, the activated recombinant protein was pre-incubated with 0.001 to 100 μM of inhibitor and subsequently assayed for its catalytic activity toward SpectroZyme Xa. While incubation with 0.5 μM Rivaroxaban resulted in full inhibition of r-hFXa and pd-hFXa, mod-hFXa-A remained fully active under these conditions (FIG. 5A). Moreover, the chimeric variant still displayed partial chromogenic activity following incubation with 100 μM Rivaroxaban, similar to the P. textilis venom FXa. These data indicate that the Ki for inhibition of mod-hFXa-A is at least 100-fold increased as compared to that of human FXa. A similar reduced sensitivity for inhibition by Apixaban was observed (FIG. 5B).

(33) Assessment of the inhibition of mod-hFXa-B by rivaroxaban and apixaban resulted in a Ki similar to that observed for mod-hFXa-A (FIGS. 6A and 6B). Thus, a reduced sensitivity for inhibition by apixaban and rivaroxaban was shown. Finally, DFXI-inhibition of the chimeric FXa variants was not altered in the presence of the cofactor FVa and negatively charged phospholipid vesicles, suggesting that both the free protease as well as that assembled into an FVa-FXa-lipid-bound complex are equally resistant to inhibition by Rivaroxaban and Apixaban (FIGS. 7A and 7B).

Example 2

(34) Materials and Methods

(35) Unless indicated otherwise, materials and methods as used in this example were the same or similar to the materials and methods indicated in Example 1.

(36) Construction and expression of recombinant FX: DNA encoding chimeric FX-A (c-FX A), chimeric FX-B (c-FX B) and chimeric FX-C (c-FX C) were synthesized at Genscript (Piscataway, N.J., USA), subcloned into pCMV4 mammalian expression vector using Apal and T4-DNA ligase and sequenced for consistency. Stabile HEK293 cell lines expressing recombinant human or recombinant chimeric FX were obtained as described previously (Larson et al., Biochemistry 1998, 37:5029-5038). HEK293 cells were cotransfected with pCMV4 and pcDNA-PACE vectors by LIPOFECTAMINE® 2000 according to the manufacturer's instructions.

(37) Purification of chimeric FX(a): Recombinant chimerix FX products A, B and C were prepared, purified and characterized as described previously (Camire et al., 2000), with the exception that the immunoaffinity purification was replaced by a calcium gradient purification of FX on a POROS HQ20-sepharose column. The typical yield of fully γ-carboxylated recombinant FX was 0.9 mg/liter conditioned medium. Purified recombinant chimeric FX was activated with RVV-X (0.1 U/mg FX), isolated by size-exclusion chromatography on a Sephacryl S200 HR column (Vt 460 ml) and stored at −20° C. in HBS containing 50% vol/vol glycerol. Purified products were visualized by Coomassie staining.

(38) Macromolecular substrate activation: Steady-state initial velocities of macromolecular substrate cleavage were determined discontinuously at 25° C. as described (Camire, 2002). Briefly, progress curves of prothrombin activation were obtained by incubating PCPS (50 μM), DAPA (10 μM), and prothrombin (1.4 μM) with human recombinant FV-810 (20 nM, B-domain truncated, constitutively active FV), and the reaction was initiated with either 0.1 nM of pd-hFXa, r-hFXa, c-FXa A, c-FXa B or c-FXa C. The rate of prothrombin conversion was measured as described (Krishnaswamy et al., 1997). Prothrombin conversion was assayed in absence or presence of direct FXa inhibitors Edoxaban (CAS Registry Number 912273-65-5; manufactured by Daiichi Sankyo, marketed as Savaysa) and Apixaban (0.001 μM-100 μM final) in order to determine DOAC sensitivity of each recombinant FXa variant.

(39) Thrombin generation assays: Thrombin generation was adapted from protocols earlier described (Hemker et al., 2003). Briefly, thrombin generation curves were obtained by supplementing FX-depleted plasma with Tissue Factor (TF, 2 or 20 pM final), Corn Trypsin Inhibitor (70 μg/ml), PCPS (20 μM) and 1 Unit (prothrombin time-specific clotting activity) of r-hFX (7 μg/ml) or chimeric FX-C (16 μg/ml). Thrombin formation was initiated by adding Substrate buffer (Fluca) to the plasma. FXa thrombin generation curves were obtained by supplementing FX-depleted plasma with Corn Trypsin Inhibitor (70 μg/ml), assay buffer and PCPS (20 μM). Thrombin formation was initiated by addition of FXa premixed with Rivaroxaban or Apixaban, assay buffer without calcium and supplemented with Fluca. The final reaction volume was 120 μl, of which 64 μl was FX-depleted plasma. Thrombin formation was determined every 20 seconds for 30 minutes and corrected for the calibrator, using the software of Thrombinoscope. The lag time, mean endogenous thrombin potential (the area under the thrombin generation curve), time to peak and peak thrombin generation, was calculated from at least three individual experiments.

(40) Results

(41) The 9-13 residue insertion in the serine protease domains of P. textilis venom, P. textilis isoform and N. scutatis venom FXa has prompted construction of chimeras of human and snake FX. Three protein coding DNA constructs were made that incorporate each of these insertions in human FXa (FIG. 9). Using these DNA constructs, HEK293 cell lines were generated that stably produce either recombinant normal human FX (r-hFX) or three types of chimeric FX (c-FX A, c-FX B and c-FX C). Expression levels of recombinant human and chimeric FX from HEK293 cells were determined by culturing the cells on expression media for 24 hours after which the clotting activity of conditioned medium was assessed by a modified one-step PT clotting assay in FX-depleted plasma. Recombinant γ-carboxylated FX was purified from conditioned media by successive ion-exchange chromatography steps. A fraction of the FX pool was subsequently activated with the FX activator from Russell's Viper venom, isolated by size-exclusion chromatography and characterized by SDS-PAGE. The heavy chain of purified plasma-derived factor Xa migrates as a 50/50 mixture of FXa-α and FXa-β at ˜34-31 kDa. While autoproteolytic excision of the C-terminal portion of FXa-α (residues 436-447) yields the β form of FXa, both isoforms are functionally similar with respect to prothrombinase assembly, prothrombin activation, antithrombin recognition, and peptidyl substrate conversion (Pryzdial and Kessler, 1996).

(42) The purified products of r-hFXa and chimeric FXa-B and -C migrated predominantly as FXa-β, chimeric FXa-A migrates as a 50/50 mixture of α and β FXa instead (FIG. 10, Panel A). Kinectis of macromolecular substrate activation by r-hFXa and chimeric FXa (A/B/C) on negatively charged phospholipid vesicles (PCPS) in the presence of the cofactor FVa shows that all chimeric variants assemble into the prothrombinase complex. However, the catalytic rate of chimeric FXa variants -A, -B and -C is respectively 8.2-, 6.8-, and 2.3-fold reduced compared to recombinant human FXa. Furthermore, recombinantly prepared human FXa shows a modest decrease in catalytic efficiency compared to plasma-derived FXa (FIG. 10, Panel B).

(43) To determine the inhibitory constant (Ki) of DOACs (Apixaban, Edoxaban) for chimeric FXa (A/B/C), the kinetics of prothrombin activation in the presence of 0.001 to 100 μM of DOAC was assayed. While plasma-derived FXa and recombinant human FXa are fully inhibited at near equimolar concentrations of DOAC, all chimeric FXa variants were able to sustain prothrombin conversion at significantly higher FXa-inhibitor concentrations (Ki Apixaban: 130-1270 nM, Ki Edoxaban: 3-270 nM) (FIG. 11). Given that the chimeric FXa variants comprise similarly positioned insertions with a varying length and composition of amino acids, the close proximity of these insertions to the DOAC-coordinating residue Tyr99 and/or the active site was speculated to be of direct consequence to the decreased sensitivity for the DOACs.

(44) In order to assess the potential of chimeric FXa to restore thrombin generation in DOAC-spiked plasma, a thrombin generation (TG) assay was performed. FXa-initiated (5 nM) thrombin generation in FX-depleted human plasma demonstrated a normal TG profile for c-FXa variant C, and near normal profiles for c-FXa variants A and B (FIG. 12A). While Apixaban (2 μM) dramatically prolonged the lag time and reduced peak thrombin generation in pd-FXa- and r-hFXa-initiated TG, these parameters were unperturbed with the chimeric FXa variants present (FIG. 12B) (Table 2). These results show that chimeric FXa variants are able to restore hemostasis in DOAC-inhibited plasma. In addition, the zymogen form of chimeric FX-C is also able to sustain thrombin generation in FX-depleted plasma. Initiation of coagulation by a low Tissue Factor (TF, 2 μM) concentration generates a robust TG curve for chimeric FX-C that is not affected by Apixaban, unlike TG by r-hFX (FIG. 13A). At low TF concentration, chimeric FX-C displays a short delay in the onset of TG and time to peak; in addition, chimeric FX-C has a larger endogenous thrombin potential (ETP) and higher peak thrombin generation (Table 3). However, these values normalize at high TF (20 μM) concentrations (FIG. 13B) (Table 3). Based on the observations made in the FXa-initiated TG assay, it is expected that zymogen forms of chimeric FX variants A and B also sustain TF-initiated TG in DOAC-spiked plasma. FIGS. 14A and 14B, in combination with Table 4, provide further evidence the effect of chimeric FXa variants on restoring hemostasis in DOAC-inhibited plasma. Taken together, these results show that chimeric FX(a) is able to restore hemostasis in DOAC-inhibited plasma, both in zymogen and protease form.

(45) TABLE-US-00003 TABLE 2 Effect of Apixaban on FXa-initiated TG parameters. Values represent experimental TG values obtained in the presence of Apixaban corrected for TG values obtained inthe absence of Apixaban. pd-FXa r-hFXa c-FXa -A c-FXa -B c-FXa -C Lagtime arrest 299 293 61 12 3 (seconds) Delay in time to 515 467 120 30 7 peak (seconds) Peak Thrombin 26 33 78 85 99 Generation (% of no Apixaban) Area under the 67 76 89 92 101 curve (% of no Apixaban)

(46) TABLE-US-00004 TABLE 3 Summary of low and high TF-initiated TG experiments. r-hFX + c-FX -C + r-hFX + c-FX -C + Low TF (2 pM) r-hFX Apixaban c-FX -C Apixaban High TF (20 pM) r-hFX Apixaban c-FX -C Apixaban Lagtime 132 ± 5  696 ± 162 185 ± 12 186 ± 6  Lagtime 48 ± 1 138 ± 12  72 ± 2  78 ± 2 (seconds) (seconds) Time to peak 324 ± 6  no peak 480 ± 24 492 ± 23 Time to peak 114 ± 6  804 ± 36 138 ± 6 144 ± 9 (seconds) (seconds) Peak thrombin 61 ± 4 8 ± 4 78 ± 1 72 ± 4 Peak thrombin 338 ± 8  32 ± 4  334 ± 15  321 ± 15 (nM) (nM) ETP (nM) 567 ± 61 no ETP  830 ± 131 756 ± 38 ETP (nM) 973 ± 18 694 ± 67 1027 ± 19 1012 ± 33

(47) TABLE-US-00005 TABLE 4 Effect of Edoxaban on TF-initiated TG parameters for r-hFX and c-FX-C. Values represent experimental TG values obtained in the presence of increasing concentrations of Edoxaban. Edoxaban (nM) control 50 100 200 400 600 1000 2000 r-hFX Lagtime(s) 115 247 297 397 538 679 874 1180 ETP remaining 100.% 99.3 87.1 no ETP no ETP no ETP no ETP no ETP Peak height % 100.% 41.2 30.8 23.1 15.9 13.0 10.1 7.1 Time to peak(s) 265 618 756 1290 1890 1932 2472 2562 c-FX -C Lagtime(s) 188 161 161 172 182 197 212 232 ETP remaining 100.% 87.9 92.5 88.3 92.7 96.8 92.7 82.0 Peak height % 100.% 109.3 112.3 101.0 94.6 88.7 77.5 65.3 Time to peak(s) 433 382 388 418 448 483 538 578