Decoy plasminogen receptor containing a selective Kunitz-inhibitory domain.

Kunitz domain 1 (KD1) of tissue factor pathway inhibitor-2 in which P2' residue Leu17 (bovine pancreatic trypsin inhibitor numbering) is mutated to Arg selectively inhibits the active site of plasmin with ∼5-fold improved affinity. Thrombin cleavage (24 h extended incubation at a 1:50 enzyme-to-substrate ratio) of the KD1 mutant (Leu17Arg) yielded a smaller molecule containing the intact Kunitz domain with no detectable change in the active-site inhibitory function. The N-terminal sequencing and MALDI-TOF/ESI data revealed that the starting molecule has a C-terminal valine (KD1L17R-VT), whereas the smaller molecule has a C-terminal lysine (KD1L17R-KT). Because KD1L17R-KT has C-terminal lysine, we examined whether it could serve as a decoy receptor for plasminogen/plasmin. Such a molecule might inhibit plasminogen activation as well as the active site of generated plasmin. In surface plasmon resonance experiments, tissue plasminogen activator (tPA) and Glu-plasminogen bound to KD1L17R-KT (Kd ∼ 0.2 to 0.3 μM) but not to KD1L17R-VT. Furthermore, KD1L17R-KT inhibited tPA-induced plasma clot fibrinolysis more efficiently than KD1L17R-VT. Additionally, compared to ε-aminocaproic acid KD1L17R-KT was more effective in reducing blood loss in a mouse liver-laceration injury model, where the fibrinolytic system is activated. In further experiments, the micro(μ)-plasmin-KD1L17R-KT complex inhibited urokinase-induced plasminogen activation on phorbol-12-myristate-13-acetate-stimulated U937 monocyte-like cells, whereas the μ-plasmin-KD1L17R-VT complex failed to inhibit this process. In conclusion, KD1L17R-KT inhibits the active site of plasmin as well as acts as a decoy receptor for the kringle domain(s) of plasminogen/plasmin; hence, it limits both plasmin generation and activity. With its dual function, KD1L17R-KT could serve as a preferred agent for controlling plasminogen activation in pathological processes.

Plasmin is a multifunctional proteolytic enzyme (MW 92 000) that circulates in blood as a single-chain inactive zymogen, plasminogen.[1] Although plasminogen is expressed ubiquitously in the body,[2] hepatocytes represent the predominant site of its synthesis.[3,4] Circulating plasminogen has an N-terminal glutamic acid (Glu–Plg) and contains a PAN/apple domain, five kringle domains, and the C-terminal latent serine protease domain.[5] Native plasminogen exists in a tight conformation in which the PAN/apple domain is bound to the kringle 5 domain.[6,7] Upon cleavage of the Arg561–Val562 scissile bond, single-chain plasminogen is converted to a disulfide-linked two-chain active protease, plasmin.[8,9] The heavy chain of Glu-plasmin consists of the PAN/apple domain and five kringle domains, whereas the light chain contains the C-terminal protease domain.[9] Plasminogen can be activated by either urokinase plasminogen activator (uPA) or tissue plasminogen activator (tPA).[9−11] Through self-catalysis, Lys-plasmin (and Lys-plasminogen) is formed that lacks the PAN/apple domain.[5,8] Because of the absence of interaction(s) between the PAN/apple domain and the kringle 5 domain, Lys-plasminogen exist in a more open conformation than Glu–Plg.[12,13] Importantly, Lys-plasminogen is more readily activated by plasminogen activators as compared to Glu–Plg.[14,15] Plasminogen activators, tPA and uPA, are inhibited by plasminogen activator inhibitor-1 (PAI-1)[16] and PAI-2,[17,18] whereas plasmin is inhibited by α2-antiplasmin[19] and α2-macroglobulin.[20]

Plasmin plays an important role in fibrinolysis and is responsible for clot lysis at the site of thrombus formation.[9] During fibrinolysis, plasminogen is primarily activated by tPA, which is released from the damaged endothelium.[21,22] Starting at the N-terminus, tPA consists of a finger domain, an epidermal growth factor-like domain, two kringle domains, and a C-terminal protease domain.[23,24] The finger domain and the second kringle domain of tPA bind to the C-terminal lysine residues exposed in the thrombus clot.[25,26] Plasminogen is also localized to the fibrin clot via its kringle domains 1 and 4, where tPA locally converts it to plasmin.[27] During degradation of fibrin, plasmin generates additional C-terminal lysine residues, which enhances tPA and plasminogen/plasmin binding to the clot for efficient lysis.[28] Furthermore, as compared to the circulating plasmin, fibrin-bound plasmin is poorly inhibited by α2-antiplasmin.[21] Thus, such a localized mechanism of fibrinolysis prevents degradation of circulating fibrinogen.

In severe trauma[29] and during major surgical procedures, such as cardiac surgery, the fibrinolytic system is hyperactivated.[30,31] In trauma, uncontrolled bleeding is the leading cause of preventable death.[32] Antifibrinolytic agents when used prophylactically can significantly reduce blood loss and the need for extensive blood transfusions.[30,33,34] Aprotinin (bovine pancreatic trypsin inhibitor, BPTI), a potent inhibitor of the plasmin active site, has been the leading antifibrinolytic agent used to prevent blood loss in cardiovascular bypass surgery.[35] However, because of its side effects such as kidney damage, myocardial infarction, and anaphylactic potential, it has been removed from the clinical market.[36−38] The presently used antifibrinolytic agents are tranexamic acid (TE) and ε-aminocaproic acid (εACA);[39] both of these lysine analogues are less effective than aprotinin and are also associated with kidney failure and seizures.[39,40] Thus, a molecule with superior efficacy is desirable for use an antifibrinolytic agent.

The plasminogen system also plays an essential role in pericellular proteolysis-dependent degradation of the extracellular matrix (ECM)[41−46] and activation of cytokines.[47] Plasmin activates several pro-matrix metalloproteases (proMMPs), including proMMPs 3, 9, 12, and 13,[48,49] that degrade other matrix components such as collagens.[50,51] Furthermore, plasmin plays a significant role in angiogenesis by releasing certain matrix-associated growth factors such as fibroblast growth factor and vascular endothelial growth factor[52,53] as well as in the activation of latent transforming growth factor-β.[54] Thus, plasmin is involved in inflammation, wound healing, cell migration, tumor growth, and apoptosis.[9] All of these functions are attributed to uPA (in association with uPA receptor)-mediated activation of plasminogen bound to the cell surface receptors present on monocytes/macrophages, smooth muscle cells, endothelial and epithelial cells, keratinocytes, and fibroblasts as well as platelets.[55] Plasminogen receptors with C-terminal lysine residues are abundantly expressed by different cell types,[56−58] and invasive properties of tumor cells are dependent on plasmin-mediated proteolysis of ECM.[59,60] Thus, attenuating plasminogen activation by preventing its binding to the cell surface receptors and inhibiting the formed plasmin could be an attractive target for controlling these pathological processes.

Tissue factor pathway inhibitor-2 (TFPI-2), also known as placental protein 5 or matrix serine protease inhibitor, contains three Kunitz-type inhibitory domains in tandem.[61−63] Kunitz domain 1 (KD1) of TFPI-2 is the only inhibitory domain, and it inhibits plasmin, factor (F) XIa, plasma kallikrein (pKLK), and FVIIa/tissue factor.[64] Using serine protease subsite S2′/P2′ profiling,[65] we designed a variant KD1-L17R molecule that is a much more potent inhibitor of plasmin; furthermore, it does not inhibit pKLK or FXIa.[66] (For KD1 and BPTI residue numbering, KD1 and BPTI residues P2, P1, P1′, and P2′ are numbered according to Schechter and Berger.[65] The corresponding subsites are S2, S1, S1′, and S2′ in the enzyme, respectively. The residue numbering corresponds to BPTI such that residue 15 is at the P1 position and residue 17 is at the P2′ position. For comparison, KD1 is numbered using the BPTI numbering system such that residues 15 and 17 in BPTI correspond to residues 24 and 26 in KD1, respectively. Thus, the residues in KD1 differ by nine from BPTI.) The Escherichia coli expressed KD1-L17R[66] has a C-terminal valine and nine residues at the N-terminus that are not part of the Kunitz domain; it is termed here as KD1L17R-VT. (For the KD1L17R-VT designation, KD1L17R-VT contains residues 1–73 of human TFPI-2; it also has four additional residues (Gly-Ser-His-Met) at the N-terminus that were introduced during cloning. It has a C-terminal valine residue.[66]) In this article, we obtained a smaller molecular species that has C-terminal lysine and lacks the first 7–9 residues; it is referred to as KD1L17R-KT. (For the KD1L17R-KT designation, approximately 80% of the KD1L17R-KT contains residues 8–72, and the remaining 20% has residues 10 (residue 1 in BPTI)–72 of human TFPI-2; it has a C-terminal lysine residue. Here, these two species are collectively referred to as KD1L17R-KT.) Structural and functional characterization of KD1L17R-KT, including its inhibition of plasmin and binding to the kringle domains of plasminogen, are presented here. KD1L17R-KT with its dual function might be a preferred agent for controlling plasmin generation and its inhibition in pathological processes.

Experimental Procedures

Materials

E. coli strain BL21(DE3) pLysS and the pET28a expression vector were obtained from Novagen Inc. (Madison, WI). Amicon centrifugal filter devices (3000 Mr cutoff) were purchased from Millipore (Bedford, MA). Q-Sepharose FF, SP-Sepharose FF, Superdex-200, and His-Trap HP columns were obtained from Amersham Biosciences. Kanamycin, isopropyl thiogalactopyranoside (IPTG), and 12-O-tetradecanoylphorbol-13-acetate (PMA) were obtained from Sigma. Human α-thrombin (IIa), FXIa, pKLK, and plasmin were purchased from Haematologic Technologies Inc. tPA (Alteplase) was purchased from Genentech (South San Francisco, CA). uPA was obtained from Calbiochem, EMD Biosciences Inc. (San Diego, CA). Glu–Plg was obtained from Enzyme Research Laboratories (South Bend, IN). εACA (Amicar) was obtained from ICN Biomedicals Inc. (Aurora, OH). Normal pooled plasma (NPP) was purchased from George King Bio-Medical Inc. (Overland Park, KS). Plasmin substrate S-2251 (H-d-Val-Leu-Lys-p-nitroanilide) and pKLK and FXIa substrate S-2366 (pyroGlu-Pro-Arg-p-nitroanilide) were obtained from Diapharma Inc. All other reagents were of the highest purity commercially available.

Expression and Purification of KD1L17R-VT

Residues 1–73 of human TFPI-2 containing the KD1 cDNA sequence were cloned and overexpressed as an N-terminal His6-tagged fusion protein in E. coli strain BL21(DE3) pLysS using the T7 promoter system.[66−69] Point mutant KD1L17R-VT was generated using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). His6-tagged KD1L17R-VT was expressed in E. coli grown in Luria broth containing 15 mg/L of kanamycin and induced at 37 °C with 1 mM IPTG. His6-tagged KD1L17R-VT was purified from the inclusion bodies and refolded as described.[66,68−70]

Purified His6-tagged KD1L17R-VT was digested with IIa at a 1:1000 enzyme/substrate molar ratio for 2 h at 37 °C. Complete digestion of KD1 by IIa was confirmed by SDS-PAGE analysis of temporal aliquots.[69] His6-tag-free protein was separated from the His6 tag and IIa using Superdex-200 gel-filtration chromatography equilibrated with 50 mM Tris-HCl, pH 7.5, containing 100 mM NaCl (TBS).[66] The preparation was characterized with respect to protein concentration using an extinction coefficient (A280) of 2.45 at 1 mg/mL, purity (SDS-PAGE), N-terminal sequence, and mass spectrometry prior to its use in biochemical experiments.

Preparation and Isolation of KD1L17R-KT

Incubation of KD1L17R-VT (1 mg/mL in TBS at pH 7.5) with IIa at a 1:50 enzyme/substrate molar ratio for 24 h at 37 °C resulted in a smaller KD1L17R-KT molecular species. KD1L17R-KT was separated from IIa using Superdex-200 gel-filtration chromatography equilibrated with TBS, pH 7.5.[66] Purified KD1L17R-KT was characterized with respect to protein concentration using an extinction coefficient (A280) of 2.84 at 1 mg/mL, purity (SDS-PAGE), N-terminal sequence, and mass spectrometry prior to its use in biochemical experiments.

SDS-PAGE

SDS-PAGE was performed using the Laemmli buffer system.[71] The acrylamide concentration used was between 10 and 15%, and the gels were stained with Coomassie Brilliant Blue dye.

Amino Acid Sequence Analysis and Mass Spectrometry

For N-terminal sequence analysis, proteins were transferred to the polyvinylidene difluoride membrane from the reduced SDS-PAGE gel. Automated N-terminal protein sequencing was performed using a pulsed-liquid solid-phase sequenator with online phenylthiohydantoin (PTH)-amino acid analysis (Applied Biosystems cLC system) at the UCLA Biopolymer Core facility. Mass spectrometry (MS) data were collected by matrix-assisted laser desorption ionization (MALDI)-MS and electrospray ionization (ESI)-MS. The protein solution was desalted and concentrated prior to MS measurements using reversed-phase media packed into pipet tips (C4-resin ZipTip, Millipore, Billerica, MA). A Voyager DE-STR MALDI time-of-flight mass spectrometer (Applied Biosystems, Foster City, CA) operating in the linear mode was used for the MALDI-MS measurements. The protein sample solution (0.5 μL) was spotted onto a stainless steel sample plate followed by 0.5 μL of the MALDI matrix solution, α-cyano-4-hydroxycinnamic acid (CHCA). The spotted samples were allowed to dry at room temperature prior to data acquisition. ESI-MS data was collected using a nanospray source and a solariX hybrid 15-T Fourier transform ion cyclotron resonance (FT-ICR) mass spectrometer (Bruker Daltonics, Billerica, MA).

KD1L17R-KT Binding to tPA and Glu–Plg Using Surface Plasmon Resonance (SPR)

Binding studies were performed on a Biacore T100 flow biosensor (Biacore, Uppsala, Sweden) at 25 °C. Glu–Plg (∼98% purity using SDS-PAGE) or tPA (>98% purity using SDS-PAGE) was immobilized on a carboxymethyl-dextran flow cell (CM5 sensor chips, GE Healthcare) using amine-coupling chemistry. Flow-cell surfaces were activated with a mixture of 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide and N-hydroxysulfosuccinimide for 5 min (flow rate of 10 μL/min), after which the protein (20 μg/mL in 10 mM sodium acetate, pH 5.5) was injected onto the surface. Unreacted sites were blocked for 5 min with 1 M ethanolamine. Each analyte (KD1L17R-VT and KD1L17R-KT, 100–2000 nM) was perfused through flow cells in HBS-P buffer (20 mM HEPES, pH 7.4, 100 mM NaCl, and 0.005% (v/v) P20) at 10 μL/min 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 εACA. 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. Kd values were calculated from the quotient of the derived dissociation (kd) and association (ka) rate constants.

Protease Inhibition Assay

All reactions were carried out in TBS, pH 7.5, containing 0.1 mg/mL BSA (TBS/BSA) and 2 mM Ca2+ (TBS/BSA/Ca2+, pH 7.5) as described.[66,68] Each enzyme (plasmin, pKLK, and FXIa) was incubated with various concentrations (10–1 to 2 × 103 nM) of KD1L17R-KT for 1 h at rt in a 96-well microtitration plate (total volume of 100 μL). A synthetic substrate (5 μL) appropriate for each enzyme was then added to a final concentration of 1 Km, and residual amidolytic activity was measured in a Molecular Devices Vmax kinetic microplate reader. The inhibition constant, Ki, was determined using the nonlinear regression data analysis program GraFit. Data for KD1L17R-KT was analyzed with an equation for a tight-binding inhibitor (eq 1) where vi and v0 are the inhibited and uninhibited rates, respectively, and [I]0 and [E]0 are the total concentrations of inhibitor and enzyme, respectively.[72,73]Ki values were obtained by correcting for the effect of substrate according to Beith[72] using eq 2, where [S] is the substrate concentration and Km is specific for each enzyme.

Fibrinolysis (Clot Lysis) Assay

The method of Sperzel and Huetter[74] was followed with minor modifications as outlined earlier.[66] 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 (KD1L17R-VT, KD1L17R-KT, and εACA) or saline control was added to 240 μL of NPP. Two-hundred twenty-five microliters of this mixture was then added to 25 μL of IIa and tPA in TBS/BSA containing 25 mM CaCl2. 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), OD405 increased immediately, indicating clotting, followed by an extremely slow decrease, representing fibrinolysis. Because clotting was almost complete after 5 min, fibrinolysis induced by tPA was evaluated as a relative decrease of OD405 up to 60 min. KD1L17R-VT (±εACA) and KD1L17R-KT were tested at final concentrations from 2 to 5 μM and εACA (without KD1L17R-VT), from 0.1 to 5 mM.

Mouse Liver-Laceration Model

A protocol using this mouse model to study the efficacy of KD1L17R-VT and of KD1L17R-KT was approved by the UCLA Chancellor’s Animal Research Committee. C57/BL6 mice (8–10 weeks old) weighing ∼20 g were used. Animals were fed standard rodent chow and water and were housed in a vivarium room with normal room temperature and 12 h light–dark cycles. The dosage of KD1L17R-VT or KD1L17R-KT was based on the aprotinin dose used in humans (i.e., 4 μg/g) adjusted for mouse weight.[75] A calculated blood level achieved by this dose was ∼8 μg/mL (∼1 μM). Animals injected with εACA (100 μg/g, calculated blood level of 200 μg/mL or ∼1.3 mM) were used as positive controls, and those receiving saline were used as negative controls. The dosage of εACA was based on the dose used in human protocols.[39,75] Polymixin B-immobilized resin (Detoxi-Gel endotoxin removing gel, Pierce, Rockford, IL) was used to remove lipopolysaccharide from saline, KD1L17R-VT, and KD1L17R-KT. εACA was dissolved at 20 mg/mL in lipopolysaccharide-free saline. Both antifibrinolytic agents and saline were free of lipopolysaccharide, as tested using the Limulus Amebocyte lysates kit (Biowhittaker Inc., Walkersville, MD).

Animals were injected intravenously via the tail vein with the drug in a volume of 100 μL of sterile balanced salt solution immediately preceding anesthesia. Anesthesia was induced and maintained with Isoflurane. The animals were placed on a heating pad in supine position throughout the procedure to maintain the rectal temperature at 37 °C. The abdomen was shaved and prepped. The liver was exposed via an anterior right subcostal incision. The left lobe of the liver was exteriorized by gentle continuous pressure on the left flank posteriorly. A 5 mm transverse incision was made on the left lobe at 1 cm from its inferior edge using a sharp sterile no. 10 scalpel blade. All blood oozing from the site of the left-lobe incision was collected in a sterile, preweighed 3 × 3 cm2 plastic tray for a total duration of 30 min. Animals were euthanized at the end of the 30 min period.

A factorial one-way analysis of variance (ANOVA) method was used to compare the mean total blood loss in milligrams from the mouse liver across the four treatment groups (saline, εACA, KD1L17R-VT, and KD1L17R-KT). The p value for comparing any two means was computed using a posthoc test under this ANOVA model. Examination of the mean total blood loss error distribution pooled across all four treatments showed that it was normal (Gaussian) after controlling for the effect of treatment. The Shapiro–Wilk goodness of fit test was used to compare formally this distribution to a Gaussian. Thus, no transformations of total blood loss were needed.

Reduction in Glu–Plg Binding to U937 Cell Surface Receptors by Micro(μ)-Plasmin–KD1L17R-KT

Recombinant N-terminal His6-tagged μ-plasminogen, containing residues 542–791 of the human plasminogen protease domain,[76] was expressed as inclusion bodies using the pET28a vector in E. coli strain BL21(DE3). The His6-tagged μ-plasminogen was isolated using nickel-charged His-Trap column, refolded, and purified using SP-Sepharose FF column as outlined.[76] Isolated μ-plasminogen (1 mg/mL) was digested with IIa to remove the His6 tag[66] and was activated to μ-plasmin by uPA at a 1:1000 enzyme/substrate molar ratio for 2 h at 37 °C. Removal of the His6 tag and activation to μ-plasmin was confirmed by reduced SDS-PAGE (15% acrylamide). IIa, uPA, and the His6 tag were separated from μ-plasmin using a Superdex-200 gel-filtration column.[76] KD1L17R-VT and KD1L17R-KT each was incubated with purified μ-plasmin at a molar ratio of 1.2:1 for 2 h at room temperature. Each complex was then separated from the free KD1L17R-VT or KD1L17R-KT using a Superdex-200 gel-filtration column.[76] Details of the procedure for obtaining the μ-plasmin–KD1L17R-VT or μ-plasmin–KD1L17R-KT complex are described elsewhere (with the crystallization and structure determination of these complexes).

U937 monocyte-like cells were maintained at 37 °C with 5% CO2 using RPMI-1640 medium supplemented with 10% fetal calf serum as outlined earlier.[77,78] The cells were stimulated with 40 nM PMA for 18 h under sterile conditions as described.[78] The nonadherent cells were recovered by gentle decanting, washed, and suspended in 50 mM Hepes containing 100 mM NaCl and 0.1 mg/mL BSA, pH 7.5 (HBS/BSA, pH 7.5). The cells (106/mL, 200 μL) were then incubated at 37 °C for 1 h with 2 μM Glu–Plg containing varying concentrations of μ-plasmin–KD1L17R-KT or μ-plasmin–KD1L17R-VT. The cells were washed and resuspended in 200 μL of HBS/BSA, pH 7.5, and the cell-bound Glu–Plg was activated with 10 nM uPA for 20 min at 37 °C. Each reaction mixture was diluted 10-fold, and the plasmin activity (reflective of bound Glu–Plg) was measured by S-2251 synthetic-substrate hydrolysis using the Molecular Devices kinetic microplate reader.[58,78] The IC50 (μ-plasmin–KD1L17R-KT concentration required for 50% decrease in Glu–Plg binding to the cell surface receptors) was determined by fitting the data to the following the IC50 four-parameter logistic equation from Halfman[79] given belowwhere y is the rate of S-2251 hydrolysis by plasmin generated from the Glu–Plg bound to the U937 cell receptors in the presence of a given concentration of KD1L17R-KT or KD1L17R-VT represented by x, a is the maximum rate of S-2251 hydrolysis in the absence of KD1L17R-KT or KD1L17R-VT, and s is the slope factor. Each point was weighted equally, and the data were fitted to eq 3 using the nonlinear regression analysis program GraFit from Erithcus Software. To obatian the Kd value for the interaction of KD1L17R-KT with Glu–Plg, we used the following equation, as described by Cheng and Prusoff[80] and further elaborated by Craig.[81]where [Glu–Plg] concentration employed was 2 μM (∼physiologic concentration).[9] To obtain the Kd (KD1L17R-KT) for Glu–Plg, we used a value of 1 μM for Kd (Glu–Plg/cell surface receptors).[9,78,82]

Molecular Modeling

The crystal structures of μ-plasmin,[83] plasminogen kringle domain 1,[84] and wild-type KD1[69] were used as templates to model both the binary complex of KD1L17R-KT with plasmin kringle domain 1 and the ternary complex of KD1L17R-KT with the plasmin protease domain and kringle domain 1. The protocols for modeling these complexes have been described earlier.[66,85] Because the C-terminus residues are disordered in the wild-type KD1 crystal structure, we used the MODELLER program[86] to build this part of the KD1 molecule. The built models were further refined by energy minimization using the program CHARMM with the CHARMM19 force field[87] consisting of 50 steps of the steepest descent followed by 500 steps of adopted basis Newton–Raphson. Harmonic restraints of 10 kcal/mol/Å2 were applied on the Cα atoms of the protein during the entire minimization.

Results and Discussion

Proteolysis of KD1L17R-VT by IIa

In addition to the 58 residues corresponding to the BPTI-Kunitz domain, the 73-residue KD1L17R-VT construct contains additional nine residues at the N-terminus. The C-terminal tail residues (59EKVPKV64, BPTI numbering) in KD1L17R-VT contain two lysine residues.[66] We investigated whether incubation of KD1L17R-VT with higher concentrations of IIa, as compared to those used during the removal of His6 tag, could cleave between the K60–V61 and/or between K63–V64 residues (59EK↓VPK↓V64). If so, then the resulting molecule(s), in addition to inhibiting the active site of plasmin, could also serve as a decoy receptor to attenuate pathological plasminogen activation. Incubation of KD1L17R-VT (1 mg/mL in TBS, pH 7.5) with IIa at a 1:50 enzyme/substrate molar ratio for 24 h at 37 °C resulted in the formation of a smaller molecular weight species referred to as KD1L17R-KT. SDS-PAGE data of KD1L17-VT and KD1L17R-KT are given in Figure 1. KD1L17R-KT was separated from IIa using Superdex-200 gel-filtration chromatography for further studies.

Figure 1

SDS-PAGE of KD1L17R-VT and KD1L17R-KT. Lane 1, mixture of reduced EZ-Run protein markers (Fisher BioReagents); lane 2, 5 μg of unreduced KD1L17R-KT; lane 3, 5 μg of unreduced KD1L17R-VT; lane 4, 5 μg of reduced KD1L17R-KT; and lane 5, 5 μg of reduced KD1L17R-VT. The concentration of acrylamide used was 15%.

The N-terminal sequence(s) of KD1L17R-KT is presented in Table 1, and the MALDI-TOF-ESI data are presented in Figure 2A. Combined analysis of the sequence(s) and the mass spectrometry data reveal that ∼80% of KD1L17R-KT has an N-terminal GNNAEI sequence and ∼20% has an N-terminal NAEI sequence; both species have the same C-terminal sequence, EKVPK. Amino acid sequence alignment of BPTI, KD1L17R-VT, and KD1L17R-KT is presented in Figure 2B. We interpret this to mean that IIa cleaved KD1L17R-VT between VPK↓V at the C-terminus; however, the hydrolysis at PT↓GN and GN↓NA at the N-terminus could be by IIa or conceivably by a contaminating protease(s). Notably, no cleavage was observed between EK↓VP at the C-terminus. Thus, KD1L17R-KT is homogeneous at the C-terminus, with ∼20% of the molecules lacking the first two amino acids at the N-terminus (Figure 2A,B).

Table 1

N-Terminal Sequence Analysis of IIa-Cleaved KD1L17R-VTa

cycle number	amino acid (pmol)	amino acid (pmol)
1	Gly (82.2)	Asn (19.2)
2	Asn (82.6)	Ala (21.5)
3	Asn (68.5)	Glu (16.5)
4	Ala (53.3)	Ile (14.7)
5	Glu (44.6)	X (Cys)
6	Ile (42.5)	Leu (12.8)
7	X (Cys)	Leu (11.9)
8	Leu (35.2)	Pro (10.8)

There were two readable sequences one at the ∼80 pmol level and the other at the ∼20 pmol level. Note that the minor sequence is internal to the first sequence starting at cycle 3.

Figure 2

MALDI-TOF-ESI mass spectrometry data and amino acid sequence alignment of KD1L17R-KT with the starting molecule (KD1L17R-VT) and BPTI. (A) Mass spectrometry data of KD1L17R-KT obtained using an Applied Biosystems Voyager DE-STR. The spectra were calibrated; thus, the masses should be within 0.1% of the theoretical values. (B) Amino acid sequences of KD1L17R-VT and the two KD1L17R-KT species: one with N-terminus sequence Gly-Asn-Asn-Ala-Glu-Ile (∼80%) and a second with N-terminus sequence Asn-Ala-Glu-Ile (∼20%). Note that the minor sequence is internal to the major sequence starting at cycle 3 asparagine and that the calculated difference in MW between the two species is 171 Da; this is consistent with the mass spectrometry data shown in panel A. We infer that IIa or a contaminating protease cleaved the peptide bonds (shown by ↓) between amino acids Thr–Gly and Asn–Asn on the N-terminal side and between Lys–Val on the C-terminal side of KD1L17R-VT. The numbering system used is that of BPTI (prototype Kunitz inhibitor). The N-terminal Gly-Ser-His-Met sequence in KD1L17R-VT is derived from the IIa cleavage site introduced during cloning. The P1 and P2′ subsites recognized by plasmin are marked.

IIa cleaves Arg/Lys (P1 subsite)-××× peptide bonds and has a strong preference for proline or a small hydrophobic residue at the P2 position.[88] Furthermore, a bulky or an acidic residue at the P2 position is strongly disfavored by IIa.[88,89] Accordingly, under the conditions of higher concentration of IIa used, cleavage between 62Pro-Lys↓Val64 (Figure 1S) of KD1L17R-VT is expected, whereas cleavage between 59Glu-Lys↓Val-Pro62 is disfavored (Figure 2B). Although threonine or asparagine at the P1 subsite is not preferred by IIa, its S1 pocket residue Asp189 can interact with Thr or Asn through a water molecule (Figure 1S). The proteolysis between Pro-Thr↓Gly-Asn or Gly-Asn↓Asn-Ala by IIa might also be facilitated by the presence of Pro or Gly at the P2 position (Figure 1S). As a result, we obtained a stable KD1L17R-KT protein that has an intact Kunitz domain and contains a C-terminal lysine.

KD1L17R-KT Binding to tPA and Glu–Plg

We used SPR to study the binding of KD1L17R-KT to immobilized tPA (Figure 3A) and Glu–Plg (Figure 3B). The kon for binding of tPA to KD1L17R-KT was 0.91 ± 0.2 × 103 M–1 s–1, koff was 1.9 ± 0.3 × 10–4 s–1, and Kd was 210 ± 20 nM. The kon for binding of Glu–Plg to KD1L17R-KT was 1.1 ± 0.4 × 103 M–1 s–1, koff was 3.1 ± 0.8 × 10–4 s–1, and Kd was 280 ± 30 nM. KD1L17R-VT showed negligible binding to tPA or Glu–Plg in SPR experiments, indicating that it is the C-terminal lysine in KD1L17R-KT that is necessary for this interaction.

Figure 3

Interaction of KD1L17R-KT with tPA and Glu–Plg as measured by SPR. (A) tPA binding to KD1L17R-KT. tPA was coupled to the CM5 chip by the amine coupling method, and an immobilization level of 1193 response units (RU) was attained for the bound protein. Five concentrations (0.1–2 μM) of KD1L17R-KT were used, and 6 min association and 10 min dissociation times (flow rate of 10 μL/min) were employed. Details are provided in the Experimental Procedures. The RU value obtained with KD1L17R-VT at 1 μM was 16 instead of 155 with KD1L17R-KT. (B) Glu–Plg binding to KD1L17R-KT. Glu–Plg was coupled to the CM5 chip, and an immobilization level of 742 RU was attained for the bound protein. The concentrations of KD1L17R-KT used and the analyte association and dissociation protocols were the same as in panel A. The RU value obtained with KD1L17R-VT at 1 μM was 12 instead of 86 with KD1L17R-KT. (C) Modeled complex of KD1L17R-KT with plasminogen kringle domain 1. The electrostatic surface of the plasminogen kringle domain 1 and a cartoon representation of the KD1L17R-KT (yellow) are depicted. The residues that form hydrogen bonds and salt bridges (shown as dashed lines) between the kringle domain and KD1L17R-KT are shown in stick representation. The carbon atoms are shown in green for the kringle domain and yellow for KD1L17R-KT. Oxygen atoms are shown in red and nitrogen atoms, in blue. The KD1L17R-KT residues are labeled with the suffix I. In the electrostatic surface, blue represents positive, red represents negative, and white represents neutral charge.

Next, we modeled the complex of KD1L17R-KT with the kringle domain 1 of plasminogen. The C-terminal lysine binds in an analogous manner as εACA (Figure 3C). In addition, two other interactions were noted: one, a carboxylate of Glu59 of the inhibitor KD1L17R-KT makes a salt bridge with the guanidino group of Arg32 of plasminogen, and two, the carbonyl group of Lys60 of the inhibitor makes a hydrogen bond with the side chain of Arg70 of kringle domain 1 of plasminogen. These additional interactions could reflect the higher affinity of KD1L17R-KT for kringle domain 1 of plasminogen as compared to its affinity for εACA.[84,90] Although not investigated, on the basis of the SPR affinity data, similar interactions between tPA and KD1L17R-KT might exist.

Inhibition Profile of KD1L17R-KT

KD1L17R-KT inhibited plasmin (Figure 4) with a high affinity (Ki =1 ± 0.2 nM) very similar to KD1L17R-VT. Furthermore, both KD1L17R-VT[66] and KD1L17R-KT inhibited FXIa and pKLK with Ki > 10 μM (Figure 4). Moreover, KD1L17R-VT[66] or KD1L17R-KT (data not shown) did not inhibit IIa, activated protein C, tPA, uPA, or tissue kallikreins. Collectively, the data indicate that the Kuntiz domain in KD1L17R-KT is not altered and is fully functional.

Figure 4

Determination of equilibrium inhibition constants (Ki) of KD1L17R-KT with plasmin, FXIa, and pKLK. The enzyme activity is expressed as the percent fractional activity (inhibited rate/uninhibited rate) at increasing inhibitor concentrations. The inhibition constants (Ki) were determined using eqs 1 and 2 as outlined in the Experimental Procedures. The concentration of plasmin used was 3 nM, whereas FXIa and pKLK were 1 nM each.

Fibrinolysis Assay

These experiments were performed to compare the effectiveness of KD1L17R-KT, KD1L17R-VT, and εACA to inhibit tPA-induced plasma clot fibrinolysis. Addition of IIa to NPP caused fibrin formation, which is reflected by the increase in OD405 (curve 1, Figure 5A–C). Simultaneous addition of tPA caused initial clot formation followed by dissolution of fibrin induced by tPA-mediated conversion of plasminogen to plasmin (curve 2, Figure 5A–C); the midpoint of fibrinolysis was between 6 to 7 min in each case. All three antifibrinolytic agents inhibited fibrinolysis in a dose-dependent manner. KD1L17R-VT increased the fibrinolysis midpoint to 22 min at 2 μM, to 29 min at 3 μM, and to 43 min at 5 μM, respectively (Figure 5A). In contrast, KD1L17R-KT increased the fibrinolysis midpoint to 27 min at 2 μM, to 37 min at 3 μM, and to >60 min at 5 μM, respectively (Figure 5A). Thus, at each concentration tested, KD1L17R-KT was more effective as an antifibrinolytic agent as compared to KD1L17R-VT. Because the active-site inhibition of plasmin by both KD1L17R-VT[66] and KD1L17R-KT is similar (Figure 4), the additional increased potency of KD1L17R-KT could be due to its binding to the kringle domains of tPA and/or plasminogen (Figure 3A,B). As a result, localized plasminogen activation at the fibrin clot might also be attenuated.

Figure 5

Effect of KD1L17R-VT, KD1L17R-KT, and εACA on fibrinolysis in human NPP. IIa was added to NPP to initiate clot formation, which is associated with an increase in optical density at 405 nm (curve 1 in panels A, B, and C). Simultaneous addition of tPA converted plasminogen to plasmin, which dissolved the fibrin clot completely within ∼10 min, as indicated by an initial increase followed by a decrease in OD405 (curve 2 in panels A, B, and C). (A) Inhibition of fibrinolysis by KD1L17R-VT and KD1L17R-KT. KD1L17R-VT: 2 (red open triangles), 3 (blue open triangles), and 5 μM (brown open triangles). KD1L17R-KT: 2 (red closed triangles), 3 (blue closed triangles), and 5 μM (brown closed triangles). In each case, IIa was used to initiate clotting and tPA to initiate fibrinolysis. (B) Effect of εACA on the inhibition of fibrinolysis by KD1L17R-VT. KD1L17R-VT: 2 (red open triangles), 3 (blue open triangles), and 5 μM (brown open triangles). KD1L17R-KT plus equimolar εACA: 2 (red closed triangles), 3 (blue closed triangles), and 5 μM (brown closed triangles). As in panel A, in each case, IIa was used to initiate clotting and tPA was used to initiate fibrinolysis. (C) Inhibition of fibrinolysis by εACA alone: 0.1 (magenta closed squares), 0.3 (blue open triangles), 0.5 (green closed triangles), 1.0 (red closed squares) and 5 mM (green open circles). As in panels A and B, in each case, IIa was used to initiate clotting and tPA was used to initiate fibrinolysis.

Higher concentrations of KD1L17R-VT or KD1L17R-KT were needed to inhibit fibrinolysis than those predicted from the Ki values for plasmin inhibition. A most likely explanation for these observations is that antifibrinolytic assay is a kinetic-based assay, whereas the equilibrium inhibition constant, Ki, was obtained upon prolonged incubation of the inhibitor with plasmin.

The antifibrinolytic activity of εACA is attributed to its binding to the kringle domains of tPA and plasminogen.[84,90] Therefore, we examined whether εACA could augment the fibrinolytic inhibition observed for KD1L17R-VT. These data are presented in Figure 5B. Equimolar additions of εACA to the KD1L17R-VT at 2, 3, or 5 μM had no observable effect on the extent of inhibition of plasma clot fibrinolysis. The most likely explanation for this observation is the low affinity (Kd ∼ 10 μM)[84,90] of εACA for tPA or plasminogen as compared to the KD1L17R-KT (Kd ∼ 0.2 to 0.3 μM; Figure 3). Consistent with these data, much higher concentrations of εACA were needed to inhibit plasma clot fibrinolysis in the absence of KD1L17R-VT (Figure 5C). Addition of εACA increased the clot lysis midpoint to 10 min at 0.1 mM, 19 min at 0.3 mM, and >60 min at 0.5 mM, respectively. Essentially no fibrinolysis was observed when the concentrations of εACA used were 1 or 5 mM.

Because augmented fibrinolysis is an important cause of bleeding from small vessels in the mouse liver-laceration model,[91] we evaluated the efficacy of KD1L17R-KT in decreasing blood loss and compared it to the efficacy of εACA and KD1L17R-VT in this animal model. The dose of εACA, KD1L17R-KT, or KD1L17R-VT used is comparable to the dose used in a clinical setting. Although the half-life of each of these two antifibrinolytic agents is short (∼2 h), a total duration of bleeding for 30 min should retain a sufficient level of each inhibitor for efficacy studies. The effects of the three antifibrinolytic agents on the quantity of blood loss are depicted in Figure 6. As compared to saline, the blood loss was reduced by ∼74% by KD1L17R-KT (p = 0.001), ∼50% by KD1L17R-VT (p = 0.002), and ∼49% by εACA (p = 0.002). There was essentially no difference in the amount of blood loss between KD1L17R-VT- and εACA-treated animals. Of significance is the observation that KD1L17R-KT is more effective than KD1L17R-VT (p < 0.05) or εACA (p < 0.05) in inhibiting plasmin-induced fibrinolysis and blood loss in this animal model.

Figure 6

Effect of KD1L17R-VT, KD1L17R-KT, and εACA in a mouse liver-laceration bleeding model. Animals were treated with saline, KD1L17R-VT, KD1L17R-KT, or εACA by intravenous injection. A 5 mm transverse incision was made on the left lobe of the liver. All blood oozing from the incision site was collected for a total duration of 30 min. Total blood loss was compared between different groups. Data are presented as the mean ± SE of 10 animals each for saline, KD1L17R-VT, and KD1L17R-KT and 16 animals for εACA. Probability p values between different groups were as follows: saline vs KD1L17R-VT, 0.002; saline vs KD1L17R-KT, 0.001; saline vs εACA, 0.002; KD1L17R-KT vs KD1L17R-VT, < 0.05; and KD1L17R-KT vs εACA, < 0.05. *, p value < 0.01 compared with saline control.

Unexpectedly, our present data with KD1L17R-KT is not statistically different from the previous data obtained with KD1-L17R, which was presumed to have Val at the C-terminus.[66] To address this discrepancy, we performed SDS-PAGE, mass spectrometry, and N-terminal sequence analysis on the actual KD1-L17R preparation used in the previous animal studies. These data are presented in the Supporting InformationFigure 2S. Evidently, the KD1-L17R previously used is composed of 76 amino acids with an N-terminal sequence of Gly-Ser-His and a C-terminal sequence of Val-Pro-Lys (instead of Pro-Lys-Val). Thus, our present data with KD1L17R-KT and the previous data with KD1-L17R (C-terminal lysine) support a concept that these molecular species have improved efficacy than the KD1L17R-VT in preventing blood loss in the mouse liver-laceration model.

Importantly, 4 of the 16 animals (one out of four each day) treated with εACA depicted generalized seizures shortly after injection of the drug. During surgery, these animals had unilateral or bilateral clonic movements on 3–4.5% isoflurane. Abnormal movements were not observed during the induction of anesthesia, but rather 15–20 min after the onset of surgery. Seizures appeared as repetitive clonic movements of one or more extremities while the animal remained unresponsive under anesthesia. Animals were not observed to have any tonic movements. Convulsion and seizures have been previously observed in animals given εACA or TE and represent a major side effect of these drugs as antifibrinolytic agents.[92,93]

Our observation is also consistent with the recent findings that εACA[39,40,94] or TE[39,94,95] cause seizures and convulsions in a significant number of patients. The εACA and TE concentrations associated with human seizures inhibit glycine receptors and could be the mechanism for the observed side effects.[93] In contrast, aprotinin did not inhibit the function of the glycine receptors.[93] These observations are consistent with our animal data, where we did not observe seizures in animals treated with either KD1L17R-VT or KD1L17R-KT. Furthermore, the use of εACA, TE, and aprotinin is associated with renal dysfunction/failure,[36−39,94] whereas two KD1 molecules (with Arg15 → Lys or Leu17 → Arg) of TFPI-2 did not cause renal toxicity in mouse[66] or rat.[96] Thus, KD1L17R-KT could be a preferred agent in attenuating fibrinolysis in vivo.

Effects of μ-Plasmin–KD1L17R-KT on Glu–Plg Binding to PMA-Stimulated Nonadherent U937 Cells

A majority of the cell surface receptors for plasminogen have lysine as their C-terminal residue.[82] We examined whether KD1L17R-KT (and not KD1L17R-VT) binds to the kringle domain(s) of Glu–Plg and prevents its binding to the U937 cell surface receptors. First, we prepared μ-plasmin–KD1L17R-VT and μ-plasmin–KD1L17R-KT to block the Kuntiz domain function. The SDS-PAGE of the Superdex-200 gel-purified complexes of μ-plasmin–KD1L17R-VT and μ-plasmin–KD1L17R-KT are shown in the inset of Figure 7A. Next, we incubated PMA-stimulated U937 cells with Glu–Plg containing varying concentrations of μ-plasmin–KD1L17R-VT or μ-plasmin–KD1L17R-KT (given in legend to Figure 7A). The increasing concentrations of μ-plasmin–KD1L17R-KT (but not of μ-plasmin–KD1L17R-VT) inhibited Glu–Plg binding to the cell surface receptors deduced from the S-2251 hydrolysis data shown in Figure 7A. The percent reduction in plasmin generation by the purified μ-plasmin–KD1L17R-KT or μ-plasmin–KD1L17R-VT is presented in Figure 7B. We interpret these data to mean that the μ-plasmin–KD1L17R-KT complex interferes with Glu–Plg binding to the PMA-stimulated U937 cells. Using eqs 3 and 4, a Kd value for the interaction of KD1L17R-KT with Glu–Plg was calculated to be ∼300 nM, a value consistent with the value (∼280 nM) obtained from the SPR data (Figure 3B).

Figure 7

Inhibition of Glu–Plg binding to PMA-stimulated nonadherent U937 cells by μ-plasmin–KD1L17R-KT. (A) Proteolytic activity of plasmin generated on the U937 cell surface. PMA-stimulated U937 cells (200 μL, 106 cells/mL) in HBS/BSA, pH 7.5, were incubated at 37 °C for 1 h with 2 μM Glu–Plg containing various concentrations of μ-plasmin–KD1L17R-KT (ranging from 0 to 3 μM) or μ-plasmin–KD1L17R-VT (2 and 4 μM). The cells were washed, and the cell-surface-bound Glu–Plg was activated with 10 nM uPA for 20 min at 37 °C. The samples were diluted 10-fold, and the plasmin formed was measured by S-2251 synthetic-substrate hydrolysis (OD405) as outlined in the Experimental Procedures. μ-Plasmin–KD1L17R-KT reduced the plasmin formed in a dose-dependent manner, whereas μ-plasmin–KD1L17R-VT had no effect (see panel B). Inset, reduced SDS-PAGE (15% acrylamide) of the size-exclusion gel-purified μ-plasmin–KD1L17R-KT and μ-plasmin–KD1L17R-VT. The μ-plasmin expressed consists of residues 542–791 cleaved between R561 (R15 in chymotrypsin numbering) and V562 (V16 in chymotrypsin numbering). (The chymotrypsin amino acid numbering system is used for the protease domain of plasmin and IIa. Residue 195 in chymotrypsin numbering corresponds to 741 in plasmin and 525 in IIa. Similarly, residue 189 in chymotrypsin numbering corresponds to 735 in plasmin and 519 in IIa.) The heavy chain (HC) consists of residues 562–791 and the light chain consists of 20 residues of 542–561 and is not seen on the gel. Lane 1, MW markers; lane 2, μ-plasmin–KD1L17R-KT complex; lane 3, μ-plasmin–KD1L17R-VT complex. (B) Percent reduction of plasmin generation by purified μ-plasmin–KD1L17R-KT and μ-plasmin–KD1L17R-VT. The plasmin activity is expressed as percent fractional activity (inhibited rate/uninhibited rate) at increasing inhibitor concentrations. (○, μ-plasmin–KD1L17R-KT; and ●, μ-plasmin–KD1L17R-VT.)

The data presented in Figure 7 support a conclusion that a single molecule of KD1L17R-KT can bind simultaneously to the kringle domain of plasminogen/plasmin as well as to the active site of plasmin. A modeled ternary complex of KD1L17R-KT with the plasmin protease domain and the plasminogen/plasmin kringle domain 1 is depicted in Figure 8. The KD1L17R-KT via its P5–P5′ residues interacts with the active site of plasmin, a typical mode of Kunitz domain binding to a serine protease domain. Similarly, the interactions of the C-terminal lysine of KD1L17R-KT with the kringle domain 1 of plasminogen/plasmin are analogous to the lysine analogues εACA or TE. However, as compared to the lysine analogues, KD1L17R-KT has additional contacts with the kringle domain 1 of plasminogen/plasmin that makes the binding considerably stronger (Figure 8).

Figure 8

Modeled ternary complex of KD1L17R-KT with the protease domain of plasmin and the kringle domain 1 of plasminogen/plasmin. A cartoon representation of the KD1L17R-KT (yellow) bound to the plasmin protease domain (magenta) and plasminogen/plasmin kringle domain 1 (cyan) is shown. The KD1L17R-KT C-terminal residues that make hydrogen bonds and salt bridges with residues of the kringle domain 1 are shown in stick representation. Similarly, residues Arg15 and Arg17 of the Kunitz domain that interact with residue Asp189 and Glu73 (chymotrypsin numbering) of the plasmin protease domain are shown in stick representation. The carbon atoms in plasminogen/plasmin are in green, whereas they are yellow in KD1L17R-KT. Oxygen and nitrogen atoms are in red and blue, respectively. The KD1L17R-KT residues are labeled with the suffix I.

Conclusions

Previously, we demonstrated that changing residue Leu17 (BPTI/aprotinin numbering) to Arg in KD1 (73-residue construct KD1L17R-VT) of TFPI-2 abolishes its anticoagulant functions and enhances its plasmin inhibition.[66] KD1L17R-VT has nine residues at the N-terminus that are not part of the Kunitz domain. Furthermore, it has valine at the C-terminus, which would prevent its binding to the kringle domains of plasminogen/plasmin. In this article, we obtained a KD1L17R-KT molecule that lacks the N-terminal residues and primarily contains the Kunitz domain with a C-terminal lysine. As expected, both KD1L17R-VT and KD1L17R-KT inhibit plasmin with the same affinity. However, in contrast to KD1L17R-VT, KD1L17R-KT binds to the kringle domain(s) of tPA and plasminogen via its C-terminal lysine residue. Consequently, KD1L17R-KT (and not KD1L17R-VT) inhibits Glu–Plg binding to U937 cells and therefore serves as a decoy plasminogen receptor. Furthermore, KD1L17R-KT inhibits tPA-induced plasma clot fibrinolysis more potently than KD1L17R-VT (Figure 5A). These data represent a dual function of KD1L17R-KT: one, it attenuates tPA and plasminogen binding to the exposed C-terminal lysine residues in the fibrin, which reduces plasmin formation, and two, it inhibits the active site of the generated plasmin. Thus, KD1L17R-KT is a superior antifibrinolytic agent in each of the in vitro experiment performed.

The efficacy of KD1L17R-KT in reducing blood loss in the mouse liver-laceration model is also improved (p < 0.05) as compared to KD1L17R-VT (Figure 6). Importantly, in the mice treated with εACA, we observed generalized seizures in 25% of the animals. We did not observe such adverse effects with KD1L17R-KT. With its dual function, KD1L17R-KT could be a useful agent in situations where the fybrinolytic system needs to be significantly weakened to prevent blood loss as well as to prevent uPA-mediated activation of plasminogen on cell surfaces.