Enhanced Antifibrinolytic Efficacy of a Plasmin-Specific Kunitz-Inhibitor (60-Residue Y11T/L17R with C-Terminal IEK) of Human Tissue Factor Pathway Inhibitor Type-2 Domain1

Current antifibrinolytic agents reduce blood loss by inhibiting plasmin active sites (e.g., aprotinin) or by preventing plasminogen/tissue plasminogen activator (tPA) binding to fibrin clots (e.g., ε-aminocaproic acid and tranexamic acid); however, they have adverse side effects. Here, we expressed 60-residue (NH2NAE…IEKCOOH) Kunitz domain1 (KD1) mutants of human tissue factor pathway inhibitor type-2 that inhibit plasmin as well as plasminogen activation. A single (KD1-L17R-KCOOH) and a double mutant (KD1-Y11T/L17R- KCOOH) were expressed in Escherichia coli as His-tagged constructs, each with enterokinase cleavage sites. KD1-Y11T/L17R-KCOOH was also expressed in Pichia pastoris. KD1-Y11T/L17R-KCOOH inhibited plasmin comparably to aprotinin and bound to the kringle domains of plasminogen/plasmin and tPA with Kd of ~50 nM and ~35 nM, respectively. Importantly, compared to aprotinin, KD1-L17R-KCOOH and KD1-Y11T/L17R-KCOOH did not inhibit kallikrein. Moreover, the antifibrinolytic potential of KD1-Y11T/L17R-KCOOH was better than that of KD1-L17R-KCOOH and similar to that of aprotinin in plasma clot-lysis assays. In thromboelastography experiments, KD1-Y11T/L17R-KCOOH was shown to inhibit fibrinolysis in a dose dependent manner and was comparable to aprotinin at a higher concentration. Further, KD1-Y11T/L17R-KCOOH did not induce cytotoxicity in primary human endothelial cells or fibroblasts. We conclude that KD1-Y11T/L17R-KCOOH is comparable to aprotinin, the most potent known inhibitor of plasmin and can be produced in large amounts using Pichia.


Introduction
In severe trauma and during major surgical procedures, such as cardiac surgery, the fibrinolytic system is hyperactivated, resulting in severe hemorrhaging [1][2][3]. Extensive bleeding poses significant mortality risks and costs in battlefields, accidents and hospital settings. Uncontrolled bleeding is 2.2. Expression and Purification of KD1-L17R-K COOH and KD1-Y11T/L17R-K COOH in E. coli The cDNA sequences of KD1-L17R-K COOH and KD1-Y11T/L17R-K COOH with C-terminal IEK were cloned and overexpressed as amino-terminal His 6 -tagged fusion proteins in E. coli strain BL21(DE3) pLysS using the T7 promoter system. The recombinant plasmid derived from pET28a, containing a His 6 leader sequence followed by an enterokinase cleavage site and the cDNA encoding the KD1-L17R-K COOH or KD1-Y11T/L17R-K COOH , was prepared according to standard procedures [35]. The sequences of the constructs expressed are given in Figure 1. The His 6 -tagged KD1-L17R-K COOH and KD1-Y11T/L17R-K COOH were expressed in E. coli grown in Luria broth containing 15 mg/liter kanamycin and induced at 37 • C with 1 mM IPTG at mid-log phase (A 600~0 .9) for 5-6 h at 37 • C. The His 6 -tagged KD1-L17R-K COOH and KD1-Y11T/L17R-K COOH were purified from the inclusion bodies using a nickel-charged His-Trap column. The His-Trap purified proteins were refolded using the reduced and oxidized glutathione system and further purified using Q-Sepharose FF column as described previously [32,36].
KD1-Y11T/L17R-KCOOH were expressed in E. coli grown in Luria broth containing 15 mg/liter kanamycin and induced at 37 °C with 1 mM IPTG at mid-log phase (A600 ~0.9) for 5-6 h at 37 °C. The His6-tagged KD1-L17R-KCOOH and KD1-Y11T/L17R-KCOOH were purified from the inclusion bodies using a nickel-charged His-Trap column. The His-Trap purified proteins were refolded using the reduced and oxidized glutathione system and further purified using Q-Sepharose FF column as described previously [32,36].

KD1-Y11T/L17R-K COOH Clone Construction and Expression in Pichia pastoris
Pichia pastoris strain X-33 and the secretion expression vector pPICZαA were purchased from Invitrogen (San Diego, CA, USA). KD1-Y11T/L17R-K COOH cDNA corresponding to the amino acid sequence ( Figure 1) was synthesized by IDT (Coralville, IA, USA). The cDNA was amplified by PCR, and the product was linearized and subcloned into XhoI and NotI restriction sites of pPICZαA. Further vector amplification was carried out in DH5α competent cells. Extracted cDNA was introduced into P. pastoris X-33 via electroporation with a Bio-Rad Gene Pulser electroporator. The transformants were plated on YPD plates supplemented with 500 µg zeocin/mL. Colonies were evaluated by SDS-PAGE for KD1-Y11T/L17R-K COOH expression in BMM medium. Fermentation inoculation shake flasks were prepared using buffered minimal glycerol medium (BMGY) pH 6.0. First, a single colony expressing KD1-Y11T/L17R-K COOH was inoculated into 50 mL for 12 h, and 0.5 mL of the resulting culture was transferred to 300 mL BMGY pH 6.0 for 20 h. The latter was then inoculated into a 15 L NLF BioEngineering Bioreactor (Wald, Switzerland) containing 3 L Basal Salts Medium (BSM) pH 5.0. Protein expression was induced with methanol and carried out for 48 h at 30 • C, pH 5.0 and 40% dissolved oxygen. Fermentation broth was centrifuged at 7200 rpm at 4 • C. Supernatant containing KD1-Y11T/L17R-K COOH was collected and stored at −30 • C until further processing.

KD1-Y11T/L17R-K COOH Purification from Pichia pastoris
One hundred milliliters of fermentation supernatant was adjusted to pH 8.5 and centrifuged for 5 min at 1500 RFC at room temperature. Then, the supernatant was collected and adjusted to pH 3.0 and mixed with 0.5 mL Titron X-100 for 30 min. Urea was added up to a final concentration of 4 M and incubated for 2.5 h at room temperature. After incubation, the solution was diluted to a final conductivity of 12 mS. Purification was carried out using Biocad Vision workstation at constant flow rate of 120 cm/h. Sample was loaded on to a SP-Sepharose (GE Healthcare Bio-Sciences Pittsburgh, PA) column previously equilibrated with 50 mM Phosphate buffer pH 2.8 (wash buffer). Then, the column was washed with two column volume wash buffer, and protein was eluted with 50 mM phosphate buffer, pH 2.8, containing 1.0 M NaCl. The fractions containing KD1-Y11T/L17R-K COOH were pooled and supplemented with L-Arginine to 0.5 M and with mannitol to 7%. After 1 h of incubation at room temperature, it was dialyzed against 10 mM phosphate buffer, pH 8.0, using 3.5 kDa MW cutoff membrane dialysis tubing (Spectrum, Palisades Park, NJ, USA).

SDS-PAGE
SDS-PAGE was performed using the Laemmli buffer system [37]. The acrylamide concentration used was 15%, and the gels were stained with Coomassie Brilliant Blue dye.
K i values were obtained by correcting for the effect of substrate according to Beith [38], using Equation (2), where [S] is substrate concentration and K M is specific for each enzyme.

Preparation of DIP-δplasmin
Active-site blocked δplasmin was generated by treating δplasmin with equal volumes of 1 M Tris-HCl, pH 8.0 and 1 M DFP (final concentration of 1 mM DFP) at room temperature for 20 min, followed by incubation on ice for several hours. Additional equal volumes of 1M Tris-HCl, pH 8.0 and 1M DFP (final concentration of 2 mM) were added and the reaction was incubated at room temperature for 20 min and then over night at 4 • C. The DFP inhibited δplamsin (DIP-δplasmin) was dialyzed against 20 mM HEPES pH 7.5, containing 150 mM NaCl and assayed for residual activity using S-2251 synthetic substrate hydrolysis. Based upon the residual activity, >99% of the δplasmin was inactivated. DIP-δplasmin, when analyzed using SDS-PAGE, revealed no protein degradation.
Flow cell surfaces were activated with a mixture of 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide and N-hydroxysulfosuccinimide for 5 min (flow rate 10 µL/min), after which the protein (20 µg/mL in 10 mM sodium acetate, pH 5.5) was placed upon the surface. Unreacted sites were blocked for 5 min with 1 M ethanolamine. The analyte KD1-Y11T/L17R-K COOH (100 to 2000 nM) was perfused through flow cells in HBS-P buffer (20 mM HEPES, pH 7.4, 100 mM NaCl, 0.005% (v/v) P20) at 10 µL/minute for 6 min. After changing to HBS-P buffer without the protein, analyte dissociation was monitored for 10 min. Flow cells were regenerated with HBS-P containing 20 mM EACA. Data were corrected for nonspecific binding by subtracting signals obtained with the analyte infused through a flow cell without the coupled protein. Binding was analyzed with BIAevaluation software (Biacore) using a 1:1 binding model. K d values were calculated from the quotient of the derived dissociation (k d ) and association (k a ) rate constants.

Fibrinolysis (Clot Lysis) Assay
The method of Sperzel and Huetter [40] was followed with minor modifications as outlined earlier [32,41]. Briefly, IIa was used to initiate fibrin formation in NPP and the lysis of the formed clot (fibrinolysis) was induced by simultaneous addition of tPA. Clot formation and lysis were monitored with a Molecular Devices microplate reader (SPECTRAmax 190) measuring the optical density at 405 nm. Briefly, 10 µL of each test compound (KD1-L17R-K COOH , KD1-Y11T/L17R-K COOH , aprotinin) or saline control was added to 240 µL of NPP. Two hundred twenty-five microliters of this mixture was then added to 25 µL IIa and tPA in TBS/BSA containing 25 mM CaCl 2 . In the 250 µL final volume, the concentration of IIa was 0.15 µg/mL and that of tPA was 1 µg/mL. Under control conditions (zero tPA and zero test compound), OD 405 increased immediately indicating clotting followed by an extremely slow decrease, representing fibrinolysis. As clotting was almost complete after 5 min, fibrinolysis induced by tPA was evaluated as a relative decrease of OD 405 up to 60 min. KD1-L17R-K COOH was tested at final concentrations from 0.5 µM to 5 µM, while KD1-Y11T/L17R-K COOH and aprotinin were tested at final concentrations from 0.5 µM to 3 µM.

Thromboelastography
The effect of different concentrations of KD1-WT, KD1-L17R-K COOH , KD1-Y11T/L17R-K COOH , aprotinin or EACA on fibrinolysis was evaluated with thromboelastography (TEG) using a TEG 5000 Thrombelastograph (Haemonetics Corp, Braintree, MA, USA). Each clot formation/lysis assay contained 300 µL of citrated whole blood, plasmin (1.5 µM final concentration), CaCl 2 (10 mM final concentration) and various concentrations of each antifibrinolytic agent in Ringer's solution to make the final volume to 360 µL. Plasmin and CaCl 2 were added last to initiate simultaneous clotting and fibrinolysis. A 1.5 µM plasmin concentration was chosen based on the plasmin effect on the clot strength and lysis. Each experiment was performed for 180 min to establish the LY60 value. The thromboelastograph was calibrated each day, and each inhibitor concentration was tested in duplicate. TEG Analytical Software (version 4.2.2; Haemonetics Corporation, Braintree, MA, USA) was used to calculate the time to clot initiation (R), maximal clot strength (maximal amplitude (MA), which was directly related to the shear elastic modulus strength, G), and percent lysis 60 min after MA (LY60) [42]. . All cells were maintained in a humidified 5% CO 2 atmosphere at 37 • C and were passaged once they reached 80% confluence. All experiments were performed with cells in the logarithmic growth phase.

Resazurin Reduction Assay
Resazurin reduction assay (Fisher Scientific) was used to evaluate the potential cytotoxicity of antifibrinolytic agents toward primary human endothelial cells and skin fibroblasts. The phosphate buffer that was used to dissolve the samples was included as negative control. The assay was based on the reduction of the nonfluorescent dye resazurin to the highly fluorescent resorufin by viable cells. The fluorescent signal is proportional to the number of live cells, since nonviable cells are unable to reduce the dye and do not produce fluorescent signals. Briefly, cells in 96-well cell culture plates were treated with different concentrations of antifibrinolytic compounds (as described above). After 24 h, resazurin reagent was added to each well and the plates were incubated at 37 • C for 4 h. Fluorescence was measured by the FLUOstar Omega Microplate Reader (BMG Labtech, Offenburg, Germany) using an excitation wavelength of 544 nm and an emission wavelength of 590 nm. Each assay was done in duplicate, with three replicates each. The viability was evaluated based on comparison with untreated cells.

Caspase 3/7 Assay
The influence of antifibrinolytic agents on apoptosis in cells was detected using the Caspase-Glo 3/7 Assay kit (Promega). Caspases 3 and 7 were activated in cells that undergo apoptosis. The assay provided a luminogenic substrate for caspase 3 and 7. Enzymatic activity leads to luminescence, which is proportional to the amount of caspase activity present. Cells were seeded in 96-well plates and treated with antifibrinolytic agents or phosphate buffer (solvent control). Taxol was included as a positive control. After 24 h of treatment, caspase reagent was added to each well, mixed and incubated for 1 h at room temperature. Luminescence was measured using the FLUOstar Omega Microplate Reader (BMG Labtech).

Cell Toxicity Assay
Cell toxicity and cell death were evaluated with the CellTox™ Green Cytotoxicity Assay (Promega). This assay measures changes in membrane integrity that occur as a result of cell death. The dye used in the system is excluded from viable cells but binds to DNA in compromised cells, which results in a fluorescent signal. We measured cell death in HUVEC and primary fibroblasts treated with antifibrinolytic agents at four concentrations (as indicated above) with triplicates per concentration in 24-well plates after 24 h of exposure. Hoechst (Thermo Fisher Scientific, Waltham, MA, USA) was used to stain all nuclei. Images of cells were captured using an inverted microscope (Nicon; Edipse T2000 TE). Green fluorescent cells (FITC filter) and Hoechst stained cells (DAPI filter) were counted using Image J software. Fluorescent cells were displayed as a percentage of all cells.

Statistical Methods
One-way analysis of variance (ANOVA) was used to compare the effect of antifibrinolytic agents in inhibiting fibrinolysis (KD1-WT, KD1-L17R-K COOH , KD1-Y11T/L17R-K COOH , aprotinin) in the plasma clot lysis assay. The p values for comparing any two means were computed using post hoc tests and adjusted for multiple comparisons using Tukey's adjustment. For the TEG data, Levene's F-test revealed that the homogeneity of variance was not met. As such, the Welch's F-test was used and Games-Howell post hoc procedure was conducted to determine which pairs of the mean MA and mean LY60% levels differed significantly. For the cell toxicity assays, collected data sets were analyzed by ANOVA and individual groups were compared using the Student's t-test. All experiments were replicated two or three times, with similar results. Quantitative values are reported as mean ± standard deviation (SD) or standard error of the mean (SEM), as indicated in the figure legends. Differences were considered statistically significant at p values of 0.05 or lower. All statistical analyses were performed using SPSS V27 (IBM Corp., Armonk, NY, USA).

Molecular Modeling
The crystal structures of µ-plasmin [43], plasminogen kringle domain1 [14] and wild-type KD1 [36] were used as templates to model the complexes of KD1-Y11T/L17R-K COOH with µ-plasmin and with plasmin kringle domain1. The protocols for modeling these complexes have been described elsewhere [41,44]. Since the C-terminus residues are disordered in the wild-type KD1 crystal structure, we used the MODELLER program [45] to build this part of the KD1-Y11T/L17R-K COOH molecule. The built models were further refined by subjecting to 1000-step minimization with the harmonic constraints of 10 kcal.mol −1 . Å −2 using the AMBER program [46].

Expression and Purification of KD1-L17R-K COOH and KD1-Y11T/L17R-K COOH in E. coli
The 60-residue His 6 -tagged KD1-L17R-K COOH and KD1-Y11T/L17R-K COOH were expressed in E. coli strain BL21 (DE3) pLysS with an enterokinase cleavage site ( Figure 1). These constructs are 9-residues shorter at the N-terminus and 3-residues shorter at the C-terminus ending with IEK-COOH ( Figure 1) as compared to the previously expressed KD1-L17R with IEKVPK at the C-terminus (designated KD1-L17R-K T ) [41]. The fusion proteins were refolded and purified using Q-Sepharose FF column. The purified KD1 mutant proteins were incubated with enterokinase to remove the His 6 -tag; however, the cleavage was unsuccessful at 1:50 ratio of enzyme to substrate. The reason for the unsuccessful His 6 -tag removal could be the inhibition of enterokinase by KD1 mutants, similar to that described for the inhibition of enterokinase by aprotinin [47]. The SDS-PAGE analysis of purified KD1-L17R-K COOH and KD1-Y11T/L17R-K COOH , each containing the enterokinase cleavage site and His 6 -tag at the NH 2 -terminus, is shown in Figure 2. [41]. The fusion proteins were refolded and purified using Q-Sepharose FF column. The purified KD1 mutant proteins were incubated with enterokinase to remove the His6tag; however, the cleavage was unsuccessful at 1:50 ratio of enzyme to substrate. The reason for the unsuccessful His6-tag removal could be the inhibition of enterokinase by KD1 mutants, similar to that described for the inhibition of enterokinase by aprotinin [47]. The SDS-PAGE analysis of purified KD1-L17R-KCOOH and KD1-Y11T/L17R-KCOOH, each containing the enterokinase cleavage site and His6-tag at the NH2-terminus, is shown in Figure 2.

Expression and Purification of KD1-Y11T/L17R-KCOOH in P. pastoris
Since the His6-tag could not be removed by enterokinase in the E. coli expressed mutants, we expressed the 60-residue double mutant KD1-Y11T/L17R-KCOOH using P. pastoris and purified to homogeneity, as described in the Experimental section. Approximately 50 mg of KD1-Y11T/L17R-KCOOH was purified from 100 mL of culture media. The SDS-PAGE analysis of purified P. pastoris KD1-Y11T/L17R-KCOOH is shown in Figure 2. Note that the P. pastoris expressed KD1-Y11T/L17R-KCOOH was of slightly lower MW compared to the corresponding E. coli expressed KD1-Y11T/L17R-KCOOH containing the His6-tag and the enterokinase cleavage sequence (Figure 1).

Expression and Purification of KD1-Y11T/L17R-K COOH in P. pastoris
Since the His 6 -tag could not be removed by enterokinase in the E. coli expressed mutants, we expressed the 60-residue double mutant KD1-Y11T/L17R-K COOH using P. pastoris and purified to homogeneity, as described in the Experimental section. Approximately 50 mg of KD1-Y11T/L17R-K COOH was purified from 100 mL of culture media. The SDS-PAGE analysis of purified P. pastoris KD1-Y11T/L17R-K COOH is shown in Figure 2. Note that the P. pastoris expressed KD1-Y11T/L17R-K COOH was of slightly lower MW compared to the corresponding E. coli expressed KD1-Y11T/L17R-K COOH containing the His 6 -tag and the enterokinase cleavage sequence ( Figure 1).
0.59 ± 0.1 Aprotinin 0.49 ± 0.1 * K i values represent an average ± SD of three independent measurements.

Fibrinolysis (Clot Lysis) Assay
These experiments were performed to compare the effectiveness of KD1-L17R-KCOOH, KD1-Y11T/L17R-KCOOH, and aprotinin at inhibiting tPA-induced plasma clot fibrinolysis. The addition of IIa to NPP caused fibrin formation, which was reflected by an increase in OD405 (curve IIa, Zero tPA, Figure 5A−C). The simultaneous addition of tPA caused initial clot formation followed by the dissolution of fibrin induced by tPA-mediated conversion of plasminogen to plasmin (curve IIa, tPA; Figure 5A−C); the midpoint of fibrinolysis was between 6 and 7 min in each case in the absence of a fibrinolytic inhibitor. All three agents inhibited fibrinolysis in a dose-dependent manner. Max OD405, OD405 at 60 min and the time to reach fibrinolysis midpoint at each concentration of the inhibitor used are provided in Table 2. Max OD405 did not differ between the inhibitors or with different concentrations of inhibitor. Max OD405 reflected the IIa-induced strength of the fibrin clot formed, which was achieved rapidly before subsequent lysis commenced by tPA generated plasmin at the clot site. Thus, it was anticipated that max OD405 at different concentrations of each inhibitor used would be similar. Further, OD405 at 60 min indicated the extent of fibrinolysis, which was relatively similar for each inhibitor at lower concentrations; however, it was more reduced for KD1-L17R-KCOOH and moderately reduced for KD1-Y11T/L17R-KCOOH as compared to aprotinin at higher concentrations ( Figure 5, Table 2).

Fibrinolysis (Clot Lysis) Assay
These experiments were performed to compare the effectiveness of KD1-L17R-K COOH , KD1-Y11T/L17R-K COOH , and aprotinin at inhibiting tPA-induced plasma clot fibrinolysis. The addition of IIa to NPP caused fibrin formation, which was reflected by an increase in OD405 (curve IIa, Zero tPA, Figure 5A−C). The simultaneous addition of tPA caused initial clot formation followed by the dissolution of fibrin induced by tPA-mediated conversion of plasminogen to plasmin (curve IIa, tPA; Figure 5A−C); the midpoint of fibrinolysis was between 6 and 7 min in each case in the absence of a fibrinolytic inhibitor. All three agents inhibited fibrinolysis in a dose-dependent manner. Max OD405, OD405 at 60 min and the time to reach fibrinolysis midpoint at each concentration of the inhibitor used are provided in Table 2. Max OD405 did not differ between the inhibitors or with different concentrations of inhibitor. Max OD405 reflected the IIa-induced strength of the fibrin clot formed, which was achieved rapidly before subsequent lysis commenced by tPA generated plasmin at the clot site. Thus, it was anticipated that max OD405 at different concentrations of each inhibitor used would be similar. Further, OD405 at 60 min indicated the extent of fibrinolysis, which was relatively similar for each inhibitor at lower concentrations; however, it was more reduced for KD1-L17R-K COOH and moderately reduced for KD1-Y11T/L17R-K COOH as compared to aprotinin at higher concentrations ( Figure 5, Table 2  ). Importantly, KD1-L17R-K COOH increased the fibrinolysis midpoint from~7 min to~10 min at 0.5 µM,~13 min at 1 µM,~17 min at 1.5 µM,~31 min at 3 µM,~43 min at 4 µM and~55 min at 5 µM, respectively ( Figure 5A, Table 2). KD1-Y11T/L17R-K COOH increased the fibrinolysis midpoint from 7 min to~12 min at 0.5 µM,~28 min at 1 µM,~43 min at 1.5 µM and > 60 min at 2 µM, as well as at 3 µM, respectively ( Figure 5B, Table 2). Aprotinin increased the midpoint of fibrinolysis from~7 min to~13 min at 0.5 µM,~40 min at 1 µM, and > 60 min at 1.5 µM as well as at >1.5 µM concentration, respectively ( Figure 5C, Table 2). Cumulatively, the statistical analyses presented in Figure 6 reveal that KD1-Y11T/L17R-K COOH was more effective in increasing the fibrinolysis midpoint as compared to KD1-L17R-K COOH , and aprotinin was slightly more effective than KD1-Y11T/L17R-K COOH . Importantly, KD1-L17R-KCOOH increased the fibrinolysis midpoint from ~7 min to ~10 min at 0.5 µM, ~13 min at 1 µM, ~17 min at 1.5 µM, ~31 min at 3 µM, ~43 min at 4 µM and ~55 min at 5 µM, respectively ( Figure 5A, Table 2). KD1-Y11T/L17R-KCOOH increased the fibrinolysis midpoint from ~7 min to ~12 min at 0.5 µM, ~28 min at 1 µM, ~43 min at 1.5 µM and > 60 min at 2 µM, as well as at 3 µM, respectively ( Figure 5B, Table 2). Aprotinin increased the midpoint of fibrinolysis from ~7 min to ~13 min at 0.5 µM, ~40 min at 1 µM, and > 60 min at 1.5 µM as well as at >1.5 µM concentration, respectively ( Figure 5C, Table 2). Cumulatively, the statistical analyses presented in Figure 6 reveal that KD1-Y11T/L17R-KCOOH was more effective in increasing the fibrinolysis midpoint as compared to KD1-L17R-KCOOH, and aprotinin was slightly more effective than KD1-Y11T/L17R-KCOOH.

Thromboelastography
Thromboelastography experiments were performed to evaluate the effect of KD1-WT [31], KD1-L17R-KCOOH, KD1-Y11T/L17R-KCOOH, aprotinin and EACA on the plasmin induced lysis of clot formed in whole blood by the addition of CaCl2. These data are presented in Figure 7 and summarized in

Thromboelastography
Thromboelastography experiments were performed to evaluate the effect of KD1-WT [31], KD1-L17R-K COOH , KD1-Y11T/L17R-K COOH , aprotinin and EACA on the plasmin induced lysis of clot formed in whole blood by the addition of CaCl 2 . These data are presented in Figure 7 and summarized in Table 3. Figure 7A shows the TEG traces at different concentrations of plasmin on the clot formation initiated with CaCl 2 . In the absence of plasmin, the average maximal amplitude (MA) achieved was 47 mm with a shear elastic modulus strength G of~4620 dyn/cm 2 , and no clot lysis could be detected at 60 min (LY60 < 0.1%). At 1.5 µM plasmin, the MA reached was~7 mm with a G value of~401 dyn/cm 2 and 100% clot lysis occurred within 30 min (Table 3). At >1.5 µM plasmin, no clot formation was observed. Figure 7B-F illustrate the average TEG traces at different concentrations (1 µM to 7.5 µM) of KD1-WT, KD1-L17R-K COOH , KD1-Y11T/L17R-K COOH and aprotinin on clot formation and lysis in the presence of 1.5 µM plasmin. The data indicate that all antifibrinolytics tested improved the clot firmness (MA) and shear strength (G) and inhibited fibrinolysis in a concentration dependent manner (Table 3). Notably, at inhibitor concentrations of 5 µM (corresponding to the high dose of the Hammersmith regime, which is the established clinical administration regimen for aprotinin) [33], KD1-Y11T/L17R-K COOH improved the clot strength MA to~80% (37.5 mm) and G to~65% (~3004 dyn/cm 2 , whereas aprotinin improved the MA to~69% (32.9 mm) and G to~53% (~2453 dyn/cm 2 ). However, LY60 of~12% was observed with KD1-Y11T/L17R-K COOH compared to 0.2% with aprotinin. At 7.5 µM concentration, both KD1-Y11T/L17R-K COOH and aprotinin had similar MA (~83% and~80%) and G (~70% and~65%), as well as LY60 (each 0.2%). EACA also improved the MA, G and LY60 in a dose dependent manner ( Figure 7G). However, at 3-mM concentration of EACA, i.e., the dose used in the clinical setting, it improved the MA and G only up to~67% and~50% respectively. Cumulatively, the TEG data indicate that EACA is not as effective as KD1-Y11T/L17R-K COOH or aprotinin in restoring the MA and G. Furthermore, KD1-WT and KD1-L17R-K COOH were also not as effective as KD1-Y11T/L17R-K COOH or aprotinin. Importantly, at higher concentrations (≥7.5 µM), KD1-Y11T/L17R-K COOH restored the MA and G, and inhibited fibrinolysis similar to aprotinin.
Multiple comparison analyses performed on the concentration-dependent enhancement of maximal amplitude (MA), shear elastic modulus strength (G) and LY60 by KD1-WT, KD1-L17R-K COOH , KD1-Y11T/L17R-K COOH and aprotinin in TEG experiments are presented in Figures 8-10. At 1 µM inhibitor concentration, the MA was not significantly different between the control and each Kunitz inhibitor except the KD1-WT ( Figure 8A). At 2 or 3 µM inhibitor concentrations, the MA enhancement by aprotinin was statistically significant (p < 0.05) as compared to KD1-WT, KD1-L17R-K COOH and KD1-Y11T/L17R-K COOH ( Figure 8B,C). Above 3 µM, enhancement in MA was statistically not different for KD1-L17R-K COOH , KD1-Y11T/L17R-K COOH and aprotinin; however, it was significantly lower for KD1-WT ( Figure 8D,E). For EACA, Student's t-test was performed to compare MA between control and each EACA concentration tested (Figure 8F). At 500 µM, 1000 µM or 3000 µM EACA concentration, the MA enhancement was statistically significant as compared to the control. dose dependent manner ( Figure 7G). However, at 3-mM concentration of EACA, i.e., the dose used in the clinical setting, it improved the MA and G only up to ~67% and ~50% respectively. Cumulatively, the TEG data indicate that EACA is not as effective as KD1-Y11T/L17R-KCOOH or aprotinin in restoring the MA and G. Furthermore, KD1-WT and KD1-L17R-KCOOH were also not as effective as KD1-Y11T/L17R-KCOOH or aprotinin. Importantly, at higher concentrations (≥7.5 µM), KD1-Y11T/L17R-KCOOH restored the MA and G, and inhibited fibrinolysis similar to aprotinin.   KD1-Y11T/L17R-KCOOH. Additionally, multiple comparison analyses of LY60 for each inhibitor at selected concentrations are presented in Figure 10. At 1 or 5 µM, aprotinin was significantly better at preventing fibrinolysis compared to each KD1 inhibitor, whereas KD1-WT was inferior to each inhibitor at all concentrations tested. At 2 or 3 µM, KD1-Y11T/L17R-KCOOH and aprotinin were superior to KD1-L17R-KCOOH and no LY60 was observed with any inhibitor at 7.5 µM concentration. Overall, it would appear that aprotinin and KD1-Y11T/L17R-KCOOH were superior to other inhibitors at inhibiting fibrinolysis in the TEG experiments.  Notably up to 3 µM, aprotinin enhanced G significantly compared to the KD1-based inhibitors ( Figure 9A-C). Surprisingly, at 5 and 7.5 µM inhibitor concentrations, the enhancement of G by KD1-Y11T/L17R-K COOH was significantly higher as compared to the other inhibitors ( Figure 9D,E). This observed improvement in clot shear strength G for KD1-Y11T/L17R-K COOH versus aprotinin might possibly have been to FXIa and kallikrein inhibition by aprotinin versus essentially no inhibition by KD1-Y11T/L17R-K COOH . Additionally, multiple comparison analyses of LY60 for each inhibitor at selected concentrations are presented in Figure 10. At 1 or 5 µM, aprotinin was significantly better at preventing fibrinolysis compared to each KD1 inhibitor, whereas KD1-WT was inferior to each inhibitor at all concentrations tested. At 2 or 3 µM, KD1-Y11T/L17R-K COOH and aprotinin were superior to KD1-L17R-K COOH and no LY60 was observed with any inhibitor at 7.5 µM concentration. Overall, it would appear that aprotinin and KD1-Y11T/L17R-K COOH were superior to other inhibitors at inhibiting fibrinolysis in the TEG experiments.

Cell Toxicity Studies
Here, we wanted to gain insights into the potential toxicity of KD1-Y11T/L17R-K COOH compared with aprotinin and the currently used antifibrinolytic agents EACA and TXA. Patients are typically treated with antifibrinolytic agents via intravenous injections while undergoing major surgery, or via external use in trauma situations. We therefore tested cytotoxicity in endothelial cells and skin fibroblasts, the cells most likely to be exposed to therapeutic doses of KD1-Y11T/L17R-K COOH . The plasma half-life of TXA in humans, rats and dogs is~120 min [48]. The half-life each of the two KD1 variant homologs (aprotinin and Ecallantide) is also~120 min [49,50] in humans, whereas the half-life of aprotinin in mice, rats or dogs is~70 min [51]. The half-life of each KD1 variant is not known, but might be short, and will be determined in future research. Since the half-life of each of the antifibrinolytic agents in vivo is short, infusion is usually continuous throughout the duration of surgery. Treatment duration was therefore set at 24 h and the chosen dose range included the equivalent of~3x the clinical dose for each of the reagents tested.
A resazurin assay of HUVEC treated with KD1-Y11T/L17R-K COOH or aprotinin for 24 h did not result in any significant change in cell viability compared to cells treated with phosphate buffer control over the entire dose range from 0.1-30 µM ( Figure 11A). The same result was obtained after treatment with EACA (dose range 1-60 mM) and TXA (dose range 0.2-30 mM). Cell viability was equally unchanged in primary human skin fibroblasts ( Figure 11B), indicating that none of the antifibrinolytic agents tested caused measurable cytotoxicity within the 24 h duration of treatment.
Viability is the endpoint of cytotoxicity. Thus, we examined the induction of apoptosis resulting from caspase activation. Caspase 3/7 assays were performed in HUVEC cells following treatment with antifibrinolytic agents ( Figure 11C). Caspase 3/7 activities significantly increased when the cells were treated with the two higher concentrations of TXA (10 mM and 30 mM) and, to a lesser extent, after exposure with EACA (20 mM and 60 mM). In contrast to TXA and EACA, KD1-Y11T/L17R-K COOH and aprotinin did not induce caspase activity above baseline at all concentrations. Taxol was included as a positive control. None of the antifibrinolytics increased caspase activity above baseline in primary fibroblasts across all doses.
To confirm the above results using a different assay, we performed CellTox green cytotoxicity assays in HUVEC cells and primary fibroblasts. The CellTox green dye binds DNA, resulting in fluorescent staining only when membrane integrity has been compromised. No significant increase in the percentage of fluorescent cells could be detected 24 h after treatment with the highest dose (30 µM) of KD1-Y11T/L17R-K COOH in the endothelial cells ( Figure 11D) or in the fibroblasts ( Figure 11E). Cell cytotoxicity also did not increase significantly over baseline when cells (HUVEC or fibroblasts) were treated with other antifibrinolytic agents. For brevity, the data for aprotinin, TXA and EACA are not shown.
In summary, 24 h treatment of HUVEC cells and primary human fibroblasts with 0.1-30 µM KD1-Y11T/L17R-K COOH or aprotinin did not decrease viability, induce apoptosis or show any sign of cytotoxicity. However, TXA and EACA induced apoptosis (cell death) at higher concentrations in HUVEC cells, as inferred from an increase in caspase 3/7 activity. Fluorescence intensity is plotted against concentration of each antifibrinolytic inhibitor. Note that, fluorescence intensity is proportional to relative cell number. Data points represent means from three independent experiments ± SEM. In each case, cell viability appears to be not significantly different from the untreated cells (p > 0.05). (C) Apoptosis in HUVEC cells. HUVEC cells were either untreated or treated with KD1-Y11T/L17R-K COOH , aprotinin, EACA or TXA for 24 h at increasing concentrations (C1-C4). Taxol was included as a positive control. Luminescence, displayed as relative light units (RLU), is proportional to caspase-3/7 activity. EACA and TXA, but not KD1-Y11T/L17R-K COOH or aprotinin show significantly increased caspase activity at concentrations used in C3 and C4 compared to the untreated cells. Data are mean ± SD from three experiments. (* p < 0.05, ** p < 0.01, *** p < 0.001). (D) Absence of cytotoxicity in HUVEC and (E) primary human skin fibroblasts with KD1-Y11T/L17R-K COOH . Cells were either untreated or treated with 30 µM KD1-Y11T/L17R-K COOH for 24 h. Taxol (0.05 µM) was included as positive control. FITC: CellTox green dye binds to DNA when membrane integrity has been compromised. Fluorescent signal indicates cytotoxicity. DAPI: nuclear stain, binds to all nuclei. Representative images of one of three independent experiments performed are shown. Graph depicts quantification of cytotoxicity assay. Percent of green cells out of all cells per field were calculated. Mean ± SD value of 4 fields per treatment group are displayed (** p < 0.01, **** p < 0.0001).

Discussion
Earlier, based on structural information and S2 -subsite specificity, we designed a 73-residue Kunitz domain plasmin inhibitor from TFPI-2 KD1 [32]. The KD1-WT inhibits plasmin as well as pKLK, FXIa and FVIIa/TF with comparable affinities, whereas KD1-L17R inhibits only plasmin. The change in residue 17 (BPTI numbering) from Leu to Arg made the KD1-L17R specific for plasmin and dramatically reduced pKLK and FXIa inhibition. As compared to the current 60-residue KD1-L17R-K COOH , the previously expressed KD1-L17R had 13 additional residues (9 from the TFPI-2 sequence and 4 from the IIa cleavage site) at the N-terminus and four (VPKV) at the C-terminus, apart from the core Kunitz domain. Although these additional residues do not interfere with KD1-L17R function, they are flexible and could be disordered as inferred from the crystal structure of the KD1-WT [36].
Therefore, a new 60-residue KD1-L17R-K COOH mutant was expressed and its inhibition profile was characterized. Since none of the active site inhibition profiles of 60-residue KD1-L17R-K COOH had changed from the previously expressed 73-residue KD1-L17R, it was predicted that KD1-L17R-K COOH would be very effective in reducing blood loss and could be comparable to aprotinin in the two mouse bleeding models (liver laceration and tail-amputation) tested [32,41,52].
The 73-residue KD1-L17R has IEKVPKV at the C-terminus and valine could be removed by extended incubation with IIa [41]. The removal of Val residue at the C-terminus generated a C-terminal lysine that made the KD1-L17R a dual reactive inhibitor of fibrinolysis by inhibiting the plasmin active site, as well as plasminogen activation [41]. Moreover, extended incubation with IIa resulted in a heterogeneous population of KD1-L17R with different N-terminal residues [41]. The structural analysis of the modeled complex of plasmin and KD1-L17R indicated that changing residue Tyr11 to Thr would be beneficial for plasmin inhibition. Threonine in KD1-Y11T/L17R-K COOH made an additional hydrogen bond with residue Q192 of plasmin ( Figure 12A). Interestingly, 73-residue KD1-L17R contained two lysine residues at the C-terminal segment (IEKVPKV), and either of them could serve as a C-terminal residue. Further, the modeling of 60-residue KD1-L17R-K COOH with the C-terminus IEK sequence showed that it will enhance the interactions with the kringle domains of plasminogen and tPA ( Figure 12B). Compared to the VPK sequence, the IEK sequence had two additional interactions arising from Arg57 and Glu59 of Kunitz domain with plasmin kringle residues Glu151 and Arg153, respectively ( Figure 12B). Similar interactions are predicted to occur with the kringle domain of tPA as well. For these reasons, the 60-residue double mutant (KD1-Y11T/L17R-K COOH ) was expressed with the IEK C-terminus.
The newly E. coli expressed KD1-L17R-K COOH and KD1-Y11T/L17R-K COOH with C-terminal IEK sequence both contained His6-tag and the enterokinase cleavage sequence; however, these additional residues could not be removed by enterokinase. Similar to the 73-residue KD1-L17R construct, the presence of additional residues did not affect the inhibition properties of KD1-L17R-K COOH and KD1-Y11T/L17R-K COOH mutants. Therefore, the 60-residue KD1-Y11T/L17R-K COOH was expressed in P. pastoris. As predicted, KD1-Y11T/L17R-K COOH inhibited plasmin with increased affinity as compared to KD1-L17R-K COOH (0.59 nM vs. 0.9 nM). Further, the 60-residue KD1-Y11T/L17R-K COOH with IEK C-terminal bound to the kringle domains of tPA and plasmin with increased affinity (35 nM to 50 nM) Figure 5 as compared to the KD1-L17R-K T with C-terminal VPK (250 nM to 300 nM) [41]. The modest increase in plasmin active site inhibition and significantly improved affinity for kringle domains of plasminogen and tPA was reflected in strong inhibition of fibrinolysis by KD1-Y11T/L17R-K COOH in plasma clot lysis assay ( Figure 4B) and in restoring MA, G and LY60 in the TEG experiments ( Figure 7B-F).
resulted in a heterogeneous population of KD1-L17R with different N-terminal residues [41]. The structural analysis of the modeled complex of plasmin and KD1-L17R indicated that changing residue Tyr11 to Thr would be beneficial for plasmin inhibition. Threonine in KD1-Y11T/L17R-KCOOH made an additional hydrogen bond with residue Q192 of plasmin ( Figure 12A). Interestingly, 73residue KD1-L17R contained two lysine residues at the C-terminal segment (IEKVPKV), and either of them could serve as a C-terminal residue. Further, the modeling of 60-residue KD1-L17R-KCOOH with the C-terminus IEK sequence showed that it will enhance the interactions with the kringle domains of plasminogen and tPA ( Figure 12B). Compared to the VPK sequence, the IEK sequence had two additional interactions arising from Arg57 and Glu59 of Kunitz domain with plasmin kringle residues Glu151 and Arg153, respectively ( Figure 12B). Similar interactions are predicted to occur with the kringle domain of tPA as well. For these reasons, the 60-residue double mutant (KD1-Y11T/L17R-KCOOH) was expressed with the IEK C-terminus.  The KD1 double mutant (KD1-Y11T/L17R-K COOH ) made in P. pastoris is a compact, homogeneous and an effective specific plasmin inhibitor of human origin. The properties of KD1-Y11T/L17R-K COOH were comparable to aprotinin in plasmin inhibition assay, plasma clot lysis assay and in the TEG experiments. Moreover, KD1-Y11T/L17R-K COOH did not inhibit pKLK, FXIa and FVIIa/sTF. Furthermore, KD1-Y11T/L17R-K COOH did not induce any measurable cytotoxicity in primary endothelial cells or skin fibroblasts ( Figure 11). However, TXA and EACA caused apoptosis in these cells at higher concentrations, which could be achieved during renal clearance of these antifibirinolytics. These results are in agreement with KD1-L17R-K T single mutant, which did not induce renal toxicity, seizures or any detectable histopathologic changes in the mouse kidney [32]. In case of aprotinin, its acidic nature and pKLK inhibition results in altered renal activity, which leads to kidney damage [32,53]. The current antifibrinolytics EACA and TXA cause seizures by inhibiting glycine receptors [54]. Since lysine analogs are not as effective as aprotinin, the higher doses of EACA and TXA increase the risk of renal failure, as these agents reach very high concentrations during clearance by glomerular filtration [55,56]. The KD1Y11T/L17R-K COOH data from the current study are encouraging; however, the compound needs to be evaluated in suitable animal bleeding models before it can be considered for clinical trials.

Patents
S.P. Bajaj has a patent pending on the KD1-L17R and related molecules. Funding: This research was funded by National Heart Lung and Blood Institute, grant number R01HL141850.