| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 280, Issue 10, 9160-9169, March 11, 2005
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
¶¶¶¶¶
From the
Departments of Medicinal Chemistry, Protein Engineering, ||Bioanalytical Research and Development, and **Physiology, Genentech, Inc., South San Francisco, California 94080, F. Hoffmann-La Roche, 4002 Basel, Switzerland, and Emory University, Atlanta, Georgia 30322
Received for publication, August 9, 2004 , and in revised form, December 23, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTS AND DISCUSSIONREFERENCES |
|---|
70% at 0.4 mg/kg + 0.1 mg/kg/min infusion. Therefore, the in vitro potency of G17905
[GenBank]
, characterized by slow binding kinetics, correlated with efficacious antithrombotic activity in vivo. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTS AND DISCUSSIONREFERENCES |
|---|
Recently, the tissue factor (TF)-factor VIIa (FVIIa) complex has emerged as an appealing molecular target for various thrombosis-related disorders, including unstable angina, acute myocardial infarction, myocardial ischemia-reperfusion injury, venous thrombosis, sepsis, and glomerulonephritis (7–12). In addition, there is evidence for a role of TF·FVIIa in tumor growth and metastasis (13, 14) (reviewed in Ref. 15). Compelling results from in vivo studies with protein-based TF·FVIIa inhibitors (16–19) and small molecule FVIIa inhibitors (20, 21) suggest that specific inhibition of TF·FVIIa may achieve anticoagulant efficacy without significantly perturbing the normal hemostatic process. The differential activity by TF·FVIIa inhibitors might be related to the recently discovered role of circulating TF in thrombosis (22). Intravascular "blood-borne" TF circulates in blood at low levels and localizes to the forming mural platelet thrombus through adhesive interactions (22–24). In addition, a circulating alternatively spliced form of human TF may also contribute to thrombus formation by accreting into growing thrombi (25). These findings led to the suggestion that specific inhibitors of TF·FVIIa may cause less bleeding because they inhibit intravascular TF at concentrations that are far below those necessary to block the high amounts of the hemostatic extravascular TF (22).
Tissue factor is the high affinity cellular cofactor for FVIIa, a trypsin-like serine protease. The TF·FVIIa complex initiates the coagulation reactions by activating its substrates factor X (FX), factor IX, and FVII. Recent insight into the structures of TF·FVIIa complex (26), FVIIa (27–29), and zymogen FVII (30) and advancements made in understanding the allosteric regulation of FVIIa (31, 32) offer various strategies to inhibit enzyme activity. For instance, phage-displayed libraries of cyclic peptides yielded potent and specific peptides, such as E-76 (33), which binds to a FVIIa exosite (34). Furthermore, antibodies that bind to the substrate interaction region of TF (35–37) displayed potent anticoagulant activity in vitro and in vivo (38, 39). A soluble TF variant with a mutated substrate binding site also exhibited strong anticoagulant activity in rabbit and guinea pig thrombosis models without appreciably interfering with hemostasis (16, 17).
Whereas such peptidic and protein inhibitors of TF·FVIIa display strong specificity and potency, they lack oral bioavailability and are not suitable for long term prophylaxis or treatment. Therefore, we chose to develop small organic active site inhibitors of FVIIa with potential for oral administration. As a first important step toward attaining this objective, the potent and highly specific inhibitor G17905 [GenBank] was produced. Other small organic active site inhibitors of FVIIa with varying selectivity and potency have been reported and were reviewed recently (40). The molecular basis for understanding the high selectivity of G17905 [GenBank] is discussed within the context of the crystal structure of FVIIa·G17905 complex. Enzyme kinetic studies, which were prompted by the discovery that G17905 [GenBank] bound to an unusual, "flipped" active site conformation of FVIIa, showed that G17905 [GenBank] is a slow binding inhibitor. Its ability to inhibit baboon FVIIa allowed us to study the antithrombotic activity and effects on normal hemostasis in baboon models.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTS AND DISCUSSIONREFERENCES |
|---|
The following synthetic substrates were used to measure enzyme activities: Spectrozyme® fVIIa (American Diagnostica) for hepatocyte growth factor activator and Chromozym tPA (Roche Applied Science) for sTF·FVIIa. The following substrates were from DiaPharma (West Chester, OH): S2765 for matriptase; S2222 and S2675 for FXa; S2302 for plasma kallikrein; S2366 for activated protein C, thrombin, and plasmin; S2444 for urokinase type plasminogen activator; S2288 for factor XIa, factor XIIa, and t-PA; S2314 for complement factor C1s; S2586 for chymotrypsin; and S2251 for trypsin.
Serine Protease Inhibition Assay—The synthetic substrates used for different enzymes are listed under "Reagents." G17905 [GenBank] was incubated with enzyme in 100 mM Hepes, pH 7.8, 140 mM NaCl, 0.1% (v/v) polyethylene glycol 8000, 0.02% (v/v) Tween 80, 5 mM CaCl2 for 40 min at room temperature. Substrate was added, and the linear rates of the increase in absorbance at 405 nm were measured on a kinetic microplate reader (Molecular Devices, Sunnyvale, CA). The inhibitor concentration that gave a 50% inhibition of enzymatic activity (IC50) was determined by fitting the data to a four-parameter equation (Kaleidagraph, Synergy Software, Reading, PA). The Ki values were calculated according to the equation, Ki = IC50/(1 + [S]/Km) (44) using the experimentally determined IC50 and Km values for each enzyme-substrate pair.
Kinetics of TF·FVIIa Inhibition—A complex of sTF and FVIIa was preformed in 96-well Costar® plates (Corning Glass) in 20 mM Hepes, pH 7.5, 150 mM NaCl, 0.5 mg/ml BSA, 5 mM CaCl2 (HBSA buffer) and incubated with increasing concentrations of G17905
[GenBank]
. After 45 min, various concentrations (0.125–4 mM) of the chromogenic substrate Chromozym t-PA were added to start the reaction. The concentrations of sTF and FVIIa in the reaction mixture were 100 and 1 nM, respectively. The rates of substrate hydrolysis were measured on a kinetic microplate reader at 405 nm (Molecular Devices). A plot of the IC50 value as a function of substrate concentration yielded a linear relationship, indicating competitive inhibition mode (45). In addition, the Kappi value was determined by fitting the data to Equation 1 (46),
 | (Eq. 1) |
For FX activation experiments, a complex of rTF and FVIIa in HBSA buffer was incubated with increasing concentrations of G17905 [GenBank] . After 40 min at room temperature, FX was added to start the reaction. The mixture contained 0.005 nM FVIIa, 0.4 nM rTF, 0.25–1.0 nM G17905 [GenBank] , and increasing concentrations of FX. Aliquots taken at different time points were quenched in 20 mM EDTA, 20 mM Hepes, pH 7.5, 150 mM NaCl, and the change in absorbance at 405 nm, upon the addition of S2765 (0.3 mM), was measured on a kinetic microplate reader. For each sample, the concentration of formed FXa was calculated from standard curves with purified FXa, and the background activities, determined in experiments without the addition of FVIIa, were subtracted from each value. The linear rates of FXa formation were determined and are shown in the double-reciprocal plot format (1/v as a function of 1/[FX]), where v is the linear rate of FXa formation (nM FXa/min).
To investigate slow binding inhibition, rTF·FVIIa complex was preformed by incubating FVIIa with rTF in HBSA buffer for 15 min in 96-well Costar® plates (Corning Glass). A pre-equilibrated mixture of G17905
[GenBank]
and FX was added to start the reaction. The concentrations of the reactants were as follows: 0.004 nM FVIIa, 0.4 nM rTF, 400 nM FX, and increasing concentrations (0.25–8 nM) of G17905
[GenBank]
. At various time points during the 20-min reaction period, aliquots were removed and quenched in 20 mM EDTA, 20 mM Hepes, pH 7.5, 150 mM NaCl, and the change in absorbance at 405 nm, upon the addition of S2765 (0.3 mM), was measured on a kinetic microplate reader. The concentration of newly formed FXa was calculated for each time point using a standard curve with purified FXa. For each experiment, the background activities in the absence of added FVIIa were determined and subtracted from each value. The data obtained were best described by a time-dependent inhibitory mechanism (45, 47, 48) and were fitted to Equation 2.
 | (Eq. 2) |
Molecular Dynamics Simulations—A protocol for molecular dynamics simulations was used for all compounds, using Amber 4.1 (49). Simulations were run using the catalytic domain from either a reference protein structure (26) or one similar to that for the complex with G17905
[GenBank]
. Parameters for the ligand were developed using standard RESP protocols as described (50), using grids calculated at the 6–31-G level carried out using Gaussian 94 (Gaussian, Inc., Pittsburgh, PA). Simulations were run with explicit solvent, using a solvent cap of 21 Å. The initial water structure was equilibrated through a multistep process as follows. The starting configuration for the simulation was minimized and then subjected to 25-ps dynamics at 200 K, with heavy atom constraints (3–4 kcal/mol Å2) on the protein and ligand. The system was then reminimized in two steps with decreasing protein and ligand constraints (1.5, 1.0, and 0.5 kcal/mol Å2 in step 1 and 1.0, 0.5, and 0.0 kcal/mol Å2 in step 2 for protein backbone, side chains, and ligand, respectively). Subsequent dynamics of the processed system were carried out using constraints only on the protein backbone 
-carbons (1.0 kcal/mol Å2) and a 250-ps equilibration phase with a 50-ps ramp to 300 K followed by a 500-ps production phase.
Synthesis of G17905 [GenBank] —Synthesis of substituted amidine component G17905 [GenBank] from the commercially available 2-methoxy-4-nitrobenzonitrile 2 (Fig. 1A) was as follows. The demethylation of 2 was done by reaction of 2 with lithium chloride at 150 °C. The resulting phenol was then protected with a benzyl group by treatment with benzyl chloride to produce compound 3. Reduction of the nitro group with hydrogen and palladium on carbon resulted in compound 4. The fluorodiethoxybenzene portion of the molecule could be obtained from aldehyde 8. This aldehyde was prepared in four steps from the commercially available 4-fluorophenol 5. Protection of the phenol with tert-butyldimethylsilyl chloride, lithiation, and reaction with B(OMe)3 resulted in the borate, which was oxidized with hydrogen peroxide to introduce the second OH group (compound 6). Dialkylation of the bis-phenol with iodoethane produced 7, and subsequent lithiation/formylation provided the aldehyde 8 (Fig. 1A).
|
Coagulation Assays—G17905 was diluted in citrated human plasma (Stanford Medical Center, Palo Alto, CA) or in pooled citrated baboon plasma. Prothrombin time (PT), activated partial thromboplastin time (APTT), thrombin time (TT), and Russell viper venom time (RVVT) were determined using standard techniques for an ACL 300 coagulometer (Beckman-Coulter, Brea, CA). Plasma containing G17905 [GenBank] was mixed with appropriate volumes of human TF reagent Innovin® (Dade Behring, Marburg, Germany) for PT, actin FS (Dade Behring) for APTT, human thrombin (10 units/ml) (Sigma) for TT, or rabbit brain cephalin plus Russell viper venom (Sigma) for RVVT. Clotting times were expressed as -fold prolongation of each test sample versus that of a plasma-containing buffer (control).
Clot Lysis Assay—Pooled citrated human plasma containing t-PA (0.4 µg/ml) and G17905 [GenBank] (0–500 µM) were added by the sampling arm of an ACL 300 coagulometer to one well of a coagulometer rotor. Human thrombin (10 units/ml in 120 mM CaCl2) (Sigma) was added to the companion well. After sample loading, clot formation and subsequent dissolution were initiated when the reaction components were mixed by centrifugation. Clot lysis was measured as the time-dependent change in light scatter, and the lysis time (i.e. the time for 50% decrease of maximal light scatter signal) was calculated. Without inhibitor the clot lysis time (with 0.4 µg/ml t-PA) was about 500 s (equal to normal clot lysis time). The concentration of G17905 [GenBank] to prolong the normal clot lysis time by 2-fold was determined.
APC Resistance Assay—G17905 at various concentrations (0–500 µM) was added to freshly thawed citrated human plasma. The plasma was mixed with the reagents of a Coatest APC resistance kit (DiaPharma Group Inc., West Chester, OH), according to the manufacturer's instructions. APTTs (with or without exogenous APC) were determined in an ACL 300 coagulometer, and the APC ratio was calculated. When screening human samples, a ratio of 
0.75 is considered the threshold for further evaluation. We found that the APC ratio could be converted to percentage of normal APC activity by determining the APC ratios of standards generated by mixing normal human plasma with APC-depleted plasma. This standard curve was linear over the ratio range 1 to 0.75, with a ratio of 0.75 equivalent to 50% normal plasma activity. Therefore, the IC50 of G17905
[GenBank]
in this assay was defined as the concentration that resulted in a 25% reduction of the APC ratio.
Baboon Arterio-venous Shunt Thrombosis Model—The animal experiments were approved by the Institutional Animal Care and Use Committee of Emory University in accordance with United States federal guidelines. The ability of G17905
[GenBank]
to inhibit thrombus formation following acute tissue factor exposure in vivo was assessed in a baboon arterio-venous shunt model that used a TF-coated graft instead of denuded baboon aorta (51). Baboons were fitted with an exteriorized silicone rubber shunt surgically placed between the femoral artery and vein. A TF-coated Gore-Tex vascular graft (2-cm length, 4-mm inner diameter) was inserted into the shunt. The graft had been coated with TF by infusing the pores of the graft with undiluted human TF reagent (Innovin®) for 5 min followed by drying overnight. The blood flow of 100 ml/min, controlled by a clamp placed distal to the graft, was monitored continuously by use of an ultrasonic flow meter (Transonic Systems Inc.). The compound G17905
[GenBank]
dissolved in phosphate-buffered saline or phosphate-buffered saline alone (control) was delivered via a venous catheter (t = 0). Before initiation of the study, blood was withdrawn, and the platelets were labeled with 111In oxine. The labeled platelets were injected at least 1 h before graft insertion. The accumulation of 111In-labeled platelets onto the graft was measured using a 
scintillation camera (General Electric 400T). Data were acquired at 5-min intervals and analyzed using a computer-assisted image processing system interfaced with the camera. The total number of platelets deposited was calculated by dividing the deposited platelet radioactivity (cpm) by the whole blood 111In-platelet radioactivity (cpm/ml) and by multiplying with the circulating platelet count (platelets/ml) measured for each experiment.
For fibrin deposition measurements, 5 µCi of 125I-labeled autologous fibrinogen was injected into the baboon at least 15 min before graft insertion. At the end of the experiment (1 h), the graft was rinsed with isotonic saline. Because of the overlapping 111In and 125I emission spectra, 125I radioactivity was measured after a delay of 30 days (11 half-lives of 111In) in order to minimize interference by 111In. Total fibrin accumulation was then calculated as the ratio of deposited 125I activity (cpm) and clottable fibrinogen radioactivity (cpm/ml) multiplied by the circulating fibrinogen concentration (mg/ml) measured for each experiment.
Baboon Blood Loss Model—Surgical bleeding in baboons receiving intravenous infusions of saline, G17905 [GenBank] , or Lovenox® was measured following femoral artery dissection and isolation as described previously (52). Briefly, after a midthigh incision, the muscle layers were separated, and the superficial femoral artery was dissected free of the surrounding tissue over a length of 10 cm. The isolated artery was not otherwise manipulated, and the incision was sutured shut. During the procedure and for 30 min after vessel isolation, the gauze sponges used to absorb lost blood were collected, and the total volume of shed blood determined from the hemoglobin content extracted from the sponges.
Statistics—Differences in platelet deposition, fibrin deposition, and blood loss between G17905
[GenBank]
and saline-treated baboons or Lovenox® and saline-treated baboons were evaluated for statistical significance by Student's t test. A p value of 0.05 was considered significant.
Determination of Plasma Concentrations of G17905 [GenBank] —At various time points, blood samples were collected by a separate venous catheter, and platelet-poor plasma was prepared by centrifugation and stored at –70 °C until further analysis. PT and APTT measurements were carried out as described above. Prolongation of clotting time was expressed as the clotting time of the sample divided by the clotting time of plasma collected prior to drug administration (post/pre). For measurements of G17905 [GenBank] plasma concentrations, an internal standard was added to the plasma samples, and proteins were precipitated with 80% acetonitrile. After centrifugation at 3,000 rpm for 10 min, the supernatant was transferred to round bottom 96-well plates and dried under nitrogen, followed by reconstitution in 100 µl of 20% (v/v) acetonitrile, 0.1% (v/v) formic acid. The samples were injected into an liquid chromatography/MS-MS (Agilent 1100/Sciex API 3000), and G17905 [GenBank] was eluted at a flow rate of 0.25 ml/min with a gradient of 25 mM ammonium acetate, 0.2% (v/v) formic acid to 90% (v/v) acetonitrile, 10% (v/v) ammonium acetate buffer with 0.2% formic acid. Analysis of G17905 [GenBank] was performed by liquid chromatography/MS-MS mass spectrometry (API 4000; Applied Biosystems, Foster City, CA) using multiple reaction monitoring (546.2 to 152.1).
X-ray Crystallography—A short form of FVIIa (rFVIIa), comprising the epidermal growth factor-2 and protease domains, was produced as described previously (30) and concentrated to 10 mg/ml in a buffer containing 150 mM NaCl and 20 mM benzamidine/HCl in 20 mM bis-Tris, pH 6.5. Crystals formed at 4 °C over 2 weeks in hanging drops made as a 50/50 (v/v) mixture of protein stock and reservoir containing 85 mM Hepes, pH 7.5, 1.7 M ammonium sulfate, 15% (v/v) glycerol, 1.7% (v/v) polyethylene glycol 400 and 5 mM CaCl2. Crystals were transferred to a synthetic mother liquor containing 2.0 M ammonium sulfate, 15% (v/v) glycerol, 2% (v/v) polyethylene glycol 400, 5 mM CaCl2, and 100 mM Hepes, pH 7.5, to which excess G17905 [GenBank] was added in 100% (v/v) Me2SO, and incubated overnight. A crystal was prepared for x-ray data collection by immersion in liquid nitrogen and intensity data extending to 2.0 Å were collected on a Quantum 4 ccd detector (ADSC; San Diego, CA) at beam line 9-2 at the Stanford Synchrotron Radiation Laboratory in space group P41212 with cell parameters a = 95.22 Å and c = 116.3 Å. Data processing and reduction were performed using HKL (53) (HKL Research, Charlottesville, VA) and ccp4 (54). The structure was solved by molecular replacement (AMoRe) (55) using the relevant domains from the TF·FVIIa complex (26). 878 reflections (2.5%) were sequestered from refinement for calculation of Rfree. Refinement was performed using XPLOR98 (Accelrys, San Diego, CA). Overall anisotropic and bulk solvent corrections were applied to the data, restrained individual isotropic atomic thermal factors were used, and the final R and Rfree for data from 30 to 2.0 Å are 19.0 and 20.8%, respectively. Data reduction statistics and final model metrics appear in Table I.
|
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTS AND DISCUSSIONREFERENCES |
|---|
|
However, these acylsulfonamides still lacked the necessary selectivity toward sTF·FVIIa relative to other serine proteases. It had been observed earlier that incorporation of a hydroxyl group ortho to the amidine could increase sTF/VIIa selectivity for given inhibitors,2 although this usually resulted in a significant loss in potency. When the hydroxyl group was incorporated into very potent inhibitors, such as the acylsulfonamides, the loss in potency was minimal. The most potent and selective of these inhibitors is G17905 [GenBank] , 1 (Fig. 1) (R-4-[2-(3-aminobenzenesulfonylamino)-1-(3,5-diethoxy-2-fluorophenyl)-2-oxo-ethylamino]-2-hydroxy-benzamidine). G17905 [GenBank] inhibited sTF·FVIIa with a Ki of 0.35 ± 0.11 nM.
Selectivity of G17905 [GenBank] Inhibitory Activity—Trypsin-like serine proteases participate in many physiological processes, such as cell migration, differentiation, immunity, tissue regeneration, and wound healing. Thus, interference of G17905 [GenBank] with proteases other than TF·FVIIa could limit its therapeutic usefulness, particularly for long term administration. In addition, highly selective TF·FVIIa inhibitors were shown to have a reduced bleeding liability (16–21). Therefore, we first determined the activity of G17905 [GenBank] toward a panel of 14 serine proteases. Except for activated protein C (Ki = 87 nM) and plasma kallikrein (Ki = 114 nM), G17905 [GenBank] did not appreciably inhibit any of the examined enzymes providing selectivities (Ki (enzyme)/Ki (sTF·FVIIa)) that were equal or greater than 3400-fold (Table II). Underscoring the high selectivity, G17905 [GenBank] did not impair the activities of serine proteases participating in nonhemostatic processes, such as innate immunity (C1s), tissue regeneration (hepatocyte growth factor activator) (58) and epidermal differentiation (matriptase) (59). Although the examined enzymes only constitute a subset of the full complement of chymotrypsin-like serine proteases in the human, the results suggest that G17905 [GenBank] is highly specific for TF·FVIIa. In agreement, G17905 [GenBank] prolonged the PT but was much less active in TF·FVIIa-independent clotting assays, such as APTT, TT, and RVVT (Table III). In addition, G17905 [GenBank] did not interfere with t-PA and plasmin-mediated clot lysis at concentrations of up to 63 µM. However, consistent with the lower selectivity ratio for activated protein C, the IC50 value of G17905 [GenBank] in a plasma-based activated protein C assay was 23.5 µM.
|
|
atoms being 0.35 Å or less (ALIGN (61)). The regions of greatest conformational divergence among these comparisons are the loop regions numbered in the 60s and 170s and the C-terminal region as well as a short section in the active site from Lys192 to Asp194 (see below).
The G17905
[GenBank]
inhibitor is found occupying the substrate binding cleft (Fig. 3A). The 
a-weighted simulated annealing omit electron density for G17905
[GenBank]
(contoured at 6 times root mean square deviation) is complete and continuous for all atoms except the final atoms of the alkyl ether substituents of the central ring. The central planar ring makes interplanar angles of 48° with the aniline ring and 93° with the benzamidine ring. The interplanar angle between the aniline and benzamidine rings is 71°. A total of 940 Å2 of solvent-accessible surface area is buried in the rFVIIa·G17905 interaction. G17905
[GenBank]
partakes in five direct hydrogen-bonding interactions with rFVIIa, one with Ser195, two with the Asp189 side chain, and two with the Gly219 carbonyl oxygen (Fig. 3A). The aniline nitrogen on the upper ring hydrogen-bonds to a water molecule interacting with the Asp60 and Tyr94 side chains and main chain amides of Gly97 and Thr98. There is another water-mediated interaction between an ether oxygen attached to the central ring and the Gly219 amide nitrogen.
|
|
The Lys192-Gly193 peptide flip did not disturb the arrangement of the catalytic triad residues Asp102, His57, and Ser195, which were unchanged from previous structures (26, 27). However, N28 of the inhibitor and the carbonyl oxygen of Lys192 provide additional hydrogen-bonding partners. The Ser195 hydroxyl hydrogen atom must hydrogen-bond with the carbonyl oxygen of Lys192 (Fig. 3C). The hydrogen atom attached to the G17905
[GenBank]
N28 is directed toward an electron lone pair of the Ser195 hydroxyl oxygen, and the His57 N
2 hydrogen atom, if present, would be directed to the remaining lone pair. The protonation state of His57 is not known.
Kinetics of rTF·FVIIa Inhibition by G17905 [GenBank] —Enzyme kinetic studies with a substituted benzamidine indicated that FVIIa active site inhibitors are competitive with respect to small peptidyl substrates and noncompetitive with respect to macromolecular substrate activation (70). So far, all published substituted benzamidine inhibitors of FVIIa bind to the standard active site conformation (71–76). However, G17905 [GenBank] binds to the nonstandard active site conformation, raising the possibility that G17905 [GenBank] could follow different inhibition kinetics. In rTF·FVIIa-mediated FX activation assays, we found G17905 [GenBank] to follow noncompetitive inhibition by decreasing the Vmax values in a concentration-dependent fashion without changing the Km (Fig. 4A). On the other hand, in amidolytic assays with Chromozym t-PA, G17905 [GenBank] was a tight binding, competitive inhibitor as indicated by the linear increase of IC50 values as a function of substrate concentration (Fig. 4A, inset). Therefore, active site inhibitors of FVIIa have identical inhibition kinetics independent of whether they occupy the standard or the nonstandard ("flipped") conformation of FVIIa. The results reinforce the view that important contributions to the binding energy in the FX-TF-FVIIa interaction are derived from exosite contacts in the FVIIa protease domain (34) and the substrate interaction region of TF (35–37).
|
Furthermore, crystal structures of rFVIIa in complex with compounds of the bis-sulfonamide series suggested that slow binding inhibition is not strictly associated with binding to the "flipped" active site conformation. Bis-sulfonamide inhibitors, exemplified by compound 15, bound to the standard conformation of the active site, although they were slow binding inhibitors (data not shown). Therefore, it is likely that analogous to the Staphylococcus aureus exfoliative toxins A and B (63–65, 68), the standard and the "flipped" FVIIa active site conformations are separated by a low energy barrier and form an equilibrium that can rapidly be shifted to optimally accommodate inhibitors.
The slow binding kinetics may be related to the observation that compared with the low Ki value, much higher concentrations of G17905 [GenBank] were needed (>20,000-fold) to prolong the PT by 2-fold. A plausible explanation for this finding is that the rapid clot formation in PT assays (<10 s) does not allow for establishing an inhibitor/enzyme equilibrium. In addition, the high plasma protein binding (97%) of G17905 [GenBank] significantly reduced inhibitor concentrations available for FVIIa binding, thus contributing to the relatively poor translation of potency from purified systems to plasma clotting assays.
Inhibition of Thrombus Formation in a Baboon Arterio-venous Shunt Model—The relatively slow equilibration of G17905 [GenBank] with TF·FVIIa complex observed in vitro raised the question whether G17905 [GenBank] could achieve effective antithrombotic activity in vivo. To address this issue, we employed a baboon arterio-venous shunt thrombosis model, previously used to study the specific anti-TF antibody D3H44 and LMWH (39). In this model, a Gore-Tex graft coated with human TF was inserted into the arterio-venous shunt, and deposition of 125I-labeled fibrinogen and 111In-labeled platelets was quantified. Clotting assays with baboon plasma indicated that potency and selectivity of G17905 [GenBank] toward baboon FVIIa was similar to human FVIIa (Table III). In agreement with this, administration of a low and high dose of G17905 [GenBank] specifically prolonged the PT but not the APTT of baboon plasma samples (Fig. 5A). The relatively constant PT prolongations achieved with the two doses were consistent with the determined inhibitor plasma levels, which were at 3 µM for the low dose (0.2 mg/kg + 0.05 mg/kg/min infusion) and at a maximum of 11 µM for the high dose (0.4 mg/kg + 0.1 mg/kg/min infusion) (Fig. 5B). G17905 [GenBank] treatment resulted in a dose-dependent reduction of fibrin deposition to 27.7% of control values (Table V). Similarly, platelet deposition was reduced to 32.3% at the high dose, but remained unchanged at the low dose (Table V). These experiments demonstrate that a TF·FVIIa-selective slow-binding inhibitor can achieve antithrombotic efficacy in vivo. Although we did not measure ex vivo APC activity directly, our human in vitro data suggest that we might have impacted APC activity at the concentrations of G17905 [GenBank] achieved in this study. For example, the maximal G17905 [GenBank] plasma concentration measured in the high dose baboon group was 11 µM, and at that concentration G17905 [GenBank] inhibited APC activity by 31% in the human plasma assay. The consequence of partial inhibition of APC activity in the context of factor VIIa inhibition is difficult to predict. If present, it may have partially blunted the antithrombotic efficacy and certainly would be a factor to consider in designing human dosing regimens.
|
|
In conclusion, we demonstrate that G17905 [GenBank] , a selective slow binding inhibitor of TF·FVIIa, effectively attenuates TF-dependent thrombus formation in vivo. The compound exemplifies a new class of promising TF·FVIIa inhibitors, which feature an S1-binding aminobenzamidine moiety amenable to prodrug modifications. As exemplified by fibrinogen receptor antagonists and direct thrombin inhibitors (77, 78), such a strategy may yield orally available derivatives useful for long term administration. Little is known about the safety of extended treatment with FVIIa active site inhibitors. However, in a guinea pig study, the continuous infusion of a FVIIa inhibitor during a 14-day period appeared to be well tolerated (79). Nevertheless, oral FVIIa inhibitors will require careful study in light of the newly emerging biological functions of TF·FVIIa that are unrelated to hemostasis and thrombosis. For instance, low TF levels lead to ventricular dysfunction in mice (80), and TF·FVIIa activity has been functionally linked to the regulation of the G-protein-coupled receptors PAR-1 and PAR-2, which participate in various biological processes (57, 81, 82). Therefore, for prolonged administration of TF·FVIIa inhibitors, pharmacokinetic properties coupled with a favorable dose-response relationship may be key issues for the safe use of this new class of anticoagulants. Only clinical studies will unambiguously decide their therapeutic usefulness.
|
FOOTNOTES |
|---|
¶ Present address: Plexxikon, Inc., Berkeley, CA 94710.
¶¶ Present address: Dept. of Biomedical Engineering, OGI School of Science and Engineering, Oregon Health and Science University, Beaverton, OR 97006-8921.
|||| To whom correspondence should be addressed: Dept. of Physiology, Genentech, Inc., 1 DNA Way, South San Francisco, CA 94080. Tel.: 650-225-2134; Fax: 650-225-6327; E-mail: dak{at}gene.com.
1 The abbreviations used are: LMWH, low molecular weight heparin; TF, tissue factor; sTF, soluble TF comprising residues 1–219; rTF, relipidated recombinant TF containing residues 1–243; rFVIIa, truncated form of FVIIa (epidermal growth factor-2 and protease domain) used for crystallization; t-PA, tissue type plasminogen activator; PT, prothrombin time, APTT, activated partial thromboplastin time, TT, thrombin time; RVVT, Russell viper venom time; FVII, factor VII; FX, factor X; BSA, bovine serum albumin; MS, mass spectrometry; APC, activated protein C.
2 K. Groebke, Y.-H. Ji, S. Wallbaum, and L. Weber (1999) European Patent Application EP 921116 A1.
3 D. Banner, personal communication.
|
ACKNOWLEDGMENTS |
|---|
