Bradykinin and Its Metabolite, Arg-Pro-Pro-Gly-Phe, Are Selective Inhibitors of α-Thrombin–Induced Platelet Activation
Background Plasma kininogens are selective inhibitors of α-thrombin activation of platelets and endothelial cells. In the present study, we localized the α-thrombin inhibitory sequence of kininogens and describe its mechanism of action.
Methods and Results Bradykinin and an analogue, MKRPPGFSPFRSSRIG, inhibited α-thrombin–induced platelet aggregation and secretion with an IC50 of 0.25 and 1 mmol/L and of 0.23 and 0.5 mmol/L, respectively. The minimal inhibitory peptide was RPPGF. Bradykinin and its analogues did not inhibit ADP-, collagen-, U46619-, or SFLLRN-induced platelet activation or the ability of α-thrombin to cleave chromogenic substrates, clot fibrinogen, or block α-thrombin binding to platelets. Bradykinin, MKRPPGFSPFRSSRIG, and RPPGF abolished α-thrombin–induced (1 nmol/L) calcium mobilization. On flow cytometry, bradykinin and MKRPPGFSPFRSSRIG blocked α-thrombin from removing the epitope of its cleavage site on the cloned thrombin receptor. Furthermore, peptide RPPGF or high-molecular-weight kininogen prevented α-thrombin from cleaving the thrombin receptor peptide, NATLDPRSFLLR, between arginine and serine.
Conclusions These results indicate that bradykinin and its metabolites are selective antithrombins by preventing α-thrombin cleavage of the cloned thrombin receptor between arginine-41 and serine-42. These newly recognized antithrombin peptides, which are termed thrombostatins, contribute to the cardioprotective nature of kinins.
Bradykinin is a vasoactive peptide that is released from the plasma kininogens.1 It has been described as having multiple physiological functions on endothelium. BK in HK participates in the binding of its parent molecules, kininogens, to endothelium.2 It is a potent stimulator of prostacyclin formation.3 4 BK induces superoxide formation and endothelium-dependent smooth muscle hyperpolarization.5 6 Along with acetylcholine, it is a major inducer of nitric acid formation.7 In an in vivo model, BK was the most potent inducer of tissue-type plasminogen activator release.8 These activities contribute to the ability of BK to modulate local blood flow. BK produces vasodilation in most vascular beds, which in the coronary artery circulation results in increased blood flow.9 The combined activities of BK maintain blood vessel patency. These features of kinins have led to its characterization as a cardioprotective agent that decreases myocardial oxygen consumption and ischemia.9 10 11 12 Its elevation by ACE inhibitors is believed to be one mechanism by which these drugs promote their beneficial effects on patients with heart failure.
In addition to the delivery of BK, the parent proteins of kinins, HK and LK, have the ability to be selective inhibitors of α-thrombin by inhibiting its ability to activate cells without interfering with its ability to cleave chromogenic substrates or to clot fibrinogen.13 14 This activity is a unique function for the kininogens. Most naturally occurring human protein inhibitors of α-thrombin are directed toward its catalytic site and/or its anion binding exosite I. HK and LK inhibit α-thrombin from activating cells by blocking it from binding to platelet and endothelial cell membranes.13 14 15 This activity of the kininogens appears to be localized to D3 because isolated D3 produced by proteolytic cleavage of the parent protein with trypsin retains that activity.16 In the present study, we found that the thrombin-inhibitory region that was ascribed to D3 is actually part of D4, the BK region, that is attached to the carboxy terminus of D3. The results show that BK and its in vivo metabolites, which we call thrombostatins, inhibit α-thrombin–induced platelet activation.
Proteins and Peptides
Human α-thrombin (3250 U/mg) was the generous gift of Dr John W. Fenton III, Division of Laboratories and Research, New York State Department of Health (Albany). Purified α-thrombin was iodinated with full preservation of its activity as previously reported.13 HK and LK were purified from plasma as previously reported.2 13 15 16 17 Briefly, HK was 120 kD and LK was 66 kD on reduced SDS-PAGE. The specific activity of purified HK was 12.5 U/mL as measured through procoagulant assay and antigen.17 D3, a 20-kD fragment, was prepared as previously reported.16 A monoclonal antibody to BK (MBK2) was generously provided by Dr Werner Mu¨ller-Esterl, Johannes Gutenberg University of Mainz (Germany). Immunoblotting of BK in kininogens and D3 was performed as previously reported with an enhanced chemiluminescence system (Amersham).15 18 Human fibrinogen and γ-thrombin were purchased from Enzyme Research Laboratories. The activity of α- and γ-thrombin was determined by their ability to clot fibrinogen as previously reported.13
BK (RPPGFSPFR), RPPGF, and RPPGFS were purchased from Sigma Chemical Co. Underlined sequences represent the native sequence of BK or a portion of it. A number of peptides that encompass all or a portion of the BK sequence were synthesized at the University of Michigan Protein Synthesis Core Laboratory: MKRPPGFSPFRSSRIG (MKBKSSRIG), SPFRSSRIGEIKEETT, MISLMKRPPGFSPFRS, MKRPPGFSPFRSS, GFSPFRSSRIG, PPGFSP, FSPFRSS, GPFPR, FPRPG, SFLLRN, LDCNAEVYVVPWEKKIYPTVNCQPLGM (LDC27), CNAEVYVVPWEKK (CNA13), NATFYFKIDNVKKARVQVVAGKKYFI (NAT26), KICVGCPRDIP (KIC11), KNKGKKNGKH (KNK10), LNAENNA (LNA7), NATLDPRSFLLR (NAT12), and FNQTQPERGDNNLTR (FNQ15). Each peptide was synthesized with an Applied Biosystems model 431 peptide synthesizer, with the carboxy-terminal amino acid covalently attached to a solid-phase support and succeeding amino acids coupled sequentially to the amino terminus. A fluorenylmethyloxycarbonyl moiety was then attached at the amino-terminal end as a blocking group. All peptides were purified with the use of preparative reverse-phase HPLC. Some of the non-BK–containing peptides were named for the first three amino acids with the use of the single letter code for each amino acid followed by the number of amino acids in the peptide. Each of these peptides was colorless, odorless, and water soluble, except NAT26, and was characterized as homogeneous according to reverse-phase HPLC, mass spectrometry, and amino acid sequencing. The peptide NAT26 was hydrophobic, requiring 0.1% DMSO for solubilization.
Platelet Aggregation and Secretion Studies
Platelet-rich plasma from fresh whole blood in 0.013 mol/L sodium citrate was prepared as previously reported.19 Washed platelets were prepared through gel filtration over Sepharose 2B columns in HEPES-Tyrode's buffer (0.137 mol/L NaCl, 3 mmol/L KCl, 0.4 mmol/L Na2H2PO4, 12 mmol/L NaHCO3, 1 mmol/L MgCl2, and 14.7 mmol/L HEPES containing 20 mmol/L glucose and 0.2 g/dL BSA, pH 7.35) as previously reported.13 Platelets for aggregation and secretion studies were incubated with [14C]-5-hydroxytryptamine for 30 minutes at 37°C.13 Washed platelets (2×108 platelets/mL final concentration) radiolabeled with [14C]-5-hydroxytryptamine were added to a cuvette of an aggregometer (Chronolog Corp) standardized as previously reported.13 After the addition of 50 μmol/L ZnCl2, purified HK (1 μmol/L) or various concentrations of the peptides (0.1 to 3 mmol/L) or buffer alone was added to the cuvette. Once the baseline stabilized, 1 nmol/L α-thrombin (0.125 U/mL) was added to initiate platelet activation. Stirred platelets were allowed to incubate with α-thrombin and additions for 1 minute. In other experiments, platelets were stimulated with 0.625 to 2.5 μmol/L SFLLRN, 10 μmol/L ADP (Sigma) in the presence of 0.5 mg/mL human fibrinogen, 1.25 μg/mL collagen (Horm), or 1 μmol/L U-46619 (Calbiochem Behring). Additional experiments were performed with washed platelets stimulated with γ-thrombin (1 nmol/L) in the presence of human fibrinogen (100 mg/dL). At the conclusion of the incubation, the entire platelet sample was centrifuged at 10 900g (model E, Beckman Instruments) over a solution of 0.135 mmol/L formaldehyde and 5 mmol/L EDTA (one part of formaldehyde/EDTA to four parts of platelet suspension) and stored on ice until an aliquot of the supernatant was assayed for [14C]-5-hydroxytryptamine secretion, as previously reported.13 Percent secretion was determined by the ratio of the supernatant of the agonist-treated specimen to the supernatant of the platelet lysate after the value of the supernatant from unstimulated platelets was subtracted from both. Platelet aggregation was measured in arbitrary units as the initial rate of change in light transmittance in the first minute after introduction of agonist.
In all experiments, gel-filtered platelets (2×108 platelets/mL) were placed into polypropylene tubes and diluted with HEPES-Tyrode's buffer containing 2 mmol/L CaCl2, 50 μmol/L ZnCl2, and additions. The reaction was started by the addition of 1 nmol/L [125I]-α-thrombin. [125I]-α-Thrombin was prepared according to the Iodogen technique as previously reported.13 Incubations were performed at 37°C for specified times with various additions. After incubation, 50-μL aliquots were removed in triplicate for each experimental point and placed into polypropylene microcentrifuge tubes with an extended tip containing 200 μL of an oil mixture that consisted of one part of Apiezon A oil to nine parts of n-butylphthalate and centrifuged at room temperature for 2 minutes at 10 900g in a microcentrifuge. The supernatant was removed, and the tips were amputated to allow placement in the gamma counter. Nonspecific binding was measured in the presence of a 100-fold molar excess in α-thrombin.
Ca2+ Mobilization Studies
The cytoplasmic free Ca2+ concentration ([Ca2+]i) was measured with the use of the fluorescent Ca2+ indicator fura-2 (Molecular Probes, Inc). Gel-filtered platelets in HEPES-Tyrode's buffer were loaded with 1 μmol/L fura-2 through incubation at 37°C for 45 minutes.20 The labeled platelets were then re–gel-filtered to remove excess probe. Aliquots of the labeled platelet suspension were transferred into a quartz cuvette with a magnetic stirrer, which was placed in a thermostatically controlled chamber at 37°C in a spectrofluorophotometer (model RE5000, Dual Wave Length Shimazdu Spectrofluorometer). Reagents were added directly to the cuvette. The excitation wavelengths varied between 340 and 380 nm. The fluorescence was measured by recording emitted light at 510 nm.21 Minimum emission was determined on a 20-μmol/L digitonin-treated platelet sample containing 10 mmol/L EGTA; maximum emission was determined on a 20-μmol/L digitonin-treated platelet sample containing 10 mmol/L EGTA and 10 mmol/L CaCl2.21 The ratio of the fluorescence readings was calculated as R=340/380 nm and processed according to the equation: [Ca2+]i=KD[(R−Rmin)/Rmax−R)](Sf2/Sb2) to calculate the intraplatelet free Ca2+ concentration.22 The KD for fura-2 was assumed to be 224 nmol/L. Rmax and Rmin are the maximum and minimum fluorescence ratios measured at the end of the experiment, respectively; and Sf2 and Sb2 are the fluorescence values at 380 nm in the absence and presence of saturating [Ca2+]i, respectively.
α-Thrombin Functional Studies
The ability of α-thrombin (1 nmol/L) to hydrolyze the chromogenic substrate H-d-Phe-Pro-Arg-pNA (S2238) (Pharmacia-Kabi) (0.7 mmol/L) in the absence or presence of additions was measured in 0.05 mol/L Tris-HCl and 0.15 mol/L NaCl, pH 7.4, containing 2 mmol/L Ca2+. The ability of α-thrombin (1 nmol/L) to clot plasma fibrinogen in pooled normal human plasma (George King Biochemicals, Inc) was determined in the absence or presence of additions in 0.01 mol/L Tris-HCl and 0.15 mol/L NaCl, pH 7.4, containing 5 mmol/L Ca2+.
Flow Cytometry Studies
Platelets for flow cytometry studies were prepared from fresh blood anticoagulated with acid citrate dextrose (10 mmol/L trisodium citrate, 66 mmol/L citric acid, 111 mmol/L glucose, pH 4.6) (8.7 mL in 53.3 mL blood). Washed platelets for flow cytometry were prepared according to a modified procedure of Molino et al.23 Briefly, platelet-rich plasma was prepared through centrifugation at 180g for 15 minutes at room temperature. The platelet-rich plasma was made with 2.8 μmol/L with PGE1 (Sigma) and 1:25 (vol/vol) with 1 mol/L sodium citrate. After a 5-minute incubation at room temperature, the platelet-rich plasma was centrifuged at 1200g for 10 minutes at room temperature. The platelet pellet was resuspended in 10 mL of platelet wash buffer (128 mmol/L NaCl, 4.26 mmol/L NaH2PO4, 7.46 mmol/L Na2HPO4, 4.77 mmol/L sodium citrate, 2.35 mmol/L citric acid, 5.5 mmol/L glucose, and 3.5 mg/mL BSA, pH 6.5) followed by centrifugation at 1200g for 5 minutes at room temperature. After resuspension in 5 mL of platelet suspension buffer (137 mmol/L NaCl, 2.6 mmol/L KCl, 13.8 mmol/L NaHCO3, 5.5 mmol/L glucose, 1 mmol/L MgCl2, 0.36 mmol/L NaH2PO4, 10 mmol/L HEPES, and 3.5 mg/mL BSA, pH 7.35), the platelet count was adjusted to 400 000/μL. One hundred microliters of washed platelets were placed in a 5-mL round-bottomed polystyrene tube and treated with various peptides or left untreated, followed by incubation with the platelet agonist α-thrombin (0.125 U/mL or 1 nmol/L) or by no incubation for 5 minutes at room temperature. Primary antibodies were added at a final concentration of 2 μg/mL and were incubated with the platelets for 30 minutes at 4°C. After incubation, the platelets were diluted with 500 μL of platelet suspension buffer and centrifuged at 1200g for 5 minutes at room temperature. The platelet pellets then were resuspended in 100 μL of platelet suspension buffer and incubated with a 1:40 dilution of an anti-mouse IgG conjugated with FITC (code No. 4350, BioSource International). After this incubation for an additional 30 minutes at 4°C, the platelets were recentrifuged at 1200g for 5 minutes, followed by resuspension in 500 μL of platelet suspension buffer.
Monoclonal antibodies to the cloned thrombin receptor, SPAN12 and ATAP138, were generously provided by Dr Lawrence F. Brass of the University of Pennsylvania.23 24 Monoclonal antibody SPAN12 was reared to the 12 amino acids of the cloned thrombin receptor that bridges its α-thrombin cleavage site (N35ATLDPRSFLLR46).23 Monoclonal antibody ATAP138, which mostly recognizes the epitope N47PNDKYEPF55 on the thrombin receptor, was reared to a mixture of peptides of SFLLRNPND and NPNDKYEPF from the cloned thrombin receptor.24 This epitope is preserved on platelets after cleavage by α-thrombin. Monoclonal antibodies LJIb10 to the α-thrombin binding site and LJIb1 to the von Willebrand factor binding site on GPIbα were generously provided by Dr Zaverio Ruggeri, Scripps Institute and Research Foundation, La Jolla, Calif.25 26 An antibody to CD62 (P-selectin) was purchased from Becton-Dickinson (catalog No. 550014). Normal mouse IgG was purchased from BioSource International. The fluorescence of bound FITC-anti-IgG to platelets was monitored with an Epics-C flow cytometer (Coulter Electronics). Light scatter and fluorescence channels were set at logarithmic gain. Laser excitation was at 488 nm. Green fluorescence was observed through a 525-nm band-pass filter. The relative fluorescence intensity of ≥15 000 platelets was analyzed in each sample.
Thrombin Receptor Peptide Cleavage Product Analysis
A thrombin receptor peptide, N55ATLDPRSFLLR46 (NAT12), which spans the α-thrombin cleavage site on the cloned thrombin receptor, was dissolved in 0.01 mol/L NaH2PO4 and 0.15 mol/L NaCl, pH 7.4, according to the procedure of Molino et al.23 It was then incubated with α-thrombin (8 nmol/L) for 1 hour at 37°C in the absence or presence of RPPGF (1 mmol/L), LNA7 (1 mmol/L), or HK (300 nmol/L). Afterward, the reaction mixture was applied to a Vyadec C-18 HPLC column in 0.1% trifluoroacetic acid and eluted with a gradient from 0% to 100% of 80% MeCN/0.1% trifluoroacetic acid. Mass spectroscopy and HPLC (to identify the intact and cleaved products of NAT12) were performed in the University of Michigan Protein Synthesis Core Laboratory.
The aim of the present study was to determine the thrombin inhibitory domain on kininogens. Previous studies from our laboratory suggested that this region was associated with trypsin-liberated D3 from LK.16 In recent studies, we determined that three surface-exposed portions of D3 were cell binding sites.27 We investigated whether these same sequences also were the α-thrombin–inhibitory regions. Peptide LDC27, which inhibited biotin-HK binding to endothelial cells with an IC50 of 60 μmol/L,27 did not inhibit α-thrombin (1 nmol/L) from inducing platelet aggregation and secretion (data not shown). Furthermore, two other peptides from the other surface-exposed regions on D3, NAT26 and KIC11, which were weaker inhibitors of biotin-HK binding to endothelial cells (IC50=257 μmol/L and >1000 μmol/L, respectively), also did not block α-thrombin–induced (1 nmol/L) platelet activation (data not shown).27 Because peptides LDC27, NAT26, and KIC11 completely cover the surface-exposed portions of D3,27 the α-thrombin inhibitory region must be associated with regions other than trypsin-liberated D3. The amino terminus of D3 isolated after tryptic cleavage of LK is well characterized; the carboxy terminus of liberated D3 has not been characterized. We investigated whether BK, which is D4, remained attached to the carboxy terminus of D3 after LK was cleaved with trypsin. On immunoblotting with a monoclonal antibody to BK, trypsin-cleaved LK28 and isolated D3 (data not shown) had BK antigen associated with them. These data suggested that BK associated with D3 could be an α-thrombin inhibitory region on kininogens.
Because BK or portions of it are associated with isolated D3 and trypsin-cleaved LK, investigations were performed to determine whether BK, the D4, could inhibit α-thrombin–induced platelet activation. Preliminary experiments revealed that peptide MKRPPGFSPFRSSRIG at 1 mmol/L completely inhibited α-thrombin–induced (1 nmol/L) platelet aggregation (Fig 1⇓). A scrambled peptide of MKRPPGFSPFRSSRIG, FSGPKRSPIMGRPSFR at 1 mmol/L, produced only 26% inhibition of aggregation and 8% inhibition of secretion after α-thrombin activation (1 nmol/L) (data not shown). Furthermore, an overlapping peptide, SPFRSSRIGEIKEETT (1 mmol/L), produced only 22% inhibition of aggregation and 4% inhibition of secretion (Fig 1). Next, a series of platelet aggregation and secretion studies were performed to define the region and minimal sequence of BK that retained the ability to inhibit α-thrombin–induced platelet activation (Fig 1). Each peptide that contained the full BK sequence produced concentration-dependent inhibition of α-thrombin–induced platelet aggregation and secretion. In all cases, the degree of inhibition of platelet aggregation was greater at a given concentration of peptide than was the degree of inhibition of platelet secretion (Fig 1). The most potent peptide inhibitor was MKRPPGFSPFRSSRIG, which inhibited platelet aggregation and secretion with an IC50 of 0.23 and 0.5 mmol/L, respectively (Table 1⇓). BK was also a potent inhibitor of α-thrombin–induced platelet activation, with an IC50 of 0.25 and 1.0 mmol/L for aggregation and secretion, respectively. Unlike the peptides that contained full-length BK, the peptide SPFRSSRIGEIKEETT, which contained only the last four amino acids of BK, produced little inhibition of α-thrombin–induced platelet activation with an IC50≥3 mmol/L. Alternatively, the first five amino acids of BK, RPPGF, produced inhibition of α-thrombin–induced platelet aggregation with an IC50 of 0.5 mmol/L. At 1 mmol/L, RPPGF produced 95% and 25% inhibition of platelet aggregation and secretion, respectively. Scrambled peptides of RPPGF (GPFPR, FPRPG) produced no inhibition of α-thrombin–induced platelet aggregation and secretion at 1 mmol/L (data not shown). BK analogues of the mid- or carboxy-terminal regions (FSPFRSS, PPGFSP) were very poor inhibitors of α-thrombin–induced platelet activation with IC50≥2 mmol/L. These data indicated that the first five amino acids of BK was the minimal sequence for inhibition. The ability of these peptides to block α-thrombin–induced platelet activation was specific for this enzyme. These BK analogue peptides did not inhibit ADP- (10 μmol/L) (in the presence of 0.5 mg/mL fibrinogen), collagen- (1.25 μg/mL), or U46619- (1 μmol/L) induced platelet aggregation and secretion. Furthermore, these peptides inhibited γ-thrombin–induced (1 nmol/L) platelet activation in the presence of 100 mg/dL human fibrinogen (data not shown). On the other hand, increasing concentrations of α-thrombin (2 to 10 nmol/L) overcame the inhibition of 1 mmol/L MKRPPGFSPFRSSRIG.
Further studies were performed to determine the mechanism by which these peptides inhibit α-thrombin–induced activation of platelets. Investigations were performed to determine whether these peptides inhibit α-thrombin–induced Ca2+ mobilization in platelets. These studies were especially important to determine whether there was a true dichotomy in the ability of BK analogue peptides to inhibit platelet aggregation and platelet secretion. Pre-liminary investigations showed that BK or peptide MKRPPGFSPFRSSRIG alone did not mobilize Ca2+ in human platelets (data not shown). α-Thrombin induced a large change in Ca2+ mobilization that was inhibited by HK (Fig 2⇓). BK and MKRPPGFSPFRSSRIG blocked α-thrombin–induced Ca2+ mobilization like their parent protein HK. Increasing concentrations of BK and peptide MKRPPGFSPFRSSRIG produced decreasing Ca2+ mobilization, with an IC50 of 0.23 and 0.3 mmol/L, respectively (Fig 3⇓). The minimal sequence of BK that produced inhibition of α-thrombin–induced Ca2+ mobilization was the peptide RPPGF (Fig 4⇓). At 1 mmol/L, it produced 80% inhibition of α-thrombin–induced Ca2+ mobilization. With the concentration of RPPGF reduced to 0.5 and 0.125 mmol/L, the level of α-thrombin–induced Ca2+ mobilization returned toward levels that occurred when no inhibitor to α-thrombin was present (Fig 4). These data indicated that these BK analogues interfered with α-thrombin activation of platelets well before the platelet aggregation phenomena.29
Further investigations were performed to determine the mechanism or mechanisms by which BK and its analogues inhibited α-thrombin–induced platelet activation. Because both kininogens and D3 inhibit binding of α-thrombin to human platelets, experiments were performed to determine whether these BK peptides also inhibited iodinated α-thrombin binding to platelets (Fig 5⇓). As previously reported, a 200-fold molar excess in HK to [125I]-α-thrombin produced the same level of nonspecific binding as proteolytically active α-thrombin during binding studies on platelets and endothelial cells.13 14 15 16 Although each of the following peptides was a good inhibitor of α-thrombin–induced platelet aggregation, secretion, and Ca2+ mobilization, peptides MKRPPGFSPFRSSRIG and GFSPFRSSRIG (each at 1 mmol/L) did not block [125I]-α-thrombin binding to washed platelets. These data indicated that the mechanism by which these BK analogues inhibited α-thrombin–induced platelet activation was different than that produced by HK, LK, or D3.13 14 15 16
Additional investigations were performed to determine the mechanism by which these peptides inhibited α-thrombin activation of platelets. Studies showed that BK or MKRPPGFPFRSSRIG from D4 did not inhibit the thrombin receptor activation peptide SFLLRN (20 μmol/L)-induced Ca2+ mobilization (Fig 6⇓). These data indicated that these BK-related peptides probably were not blocking the triggering of the α-thrombin stimulus response coupling pathway in platelets. Other investigations showed that 1 mmol/L RPPGF and MKRPPGFSPFRSSRIG did not block 1 nmol/L α-thrombin from cleaving a chromogenic substrate (Fig 7A⇓). In addition, RPPGF (0.1 to 3 mmol/L) did not inhibit the ability of α-thrombin to clot plasma fibrinogen (Fig 7B). These data suggested that these peptides were not producing inhibition of the activation by α-thrombin of platelets by blocking its active site and/or its anion binding exosite I. Moreover, inhibition of α-thrombin by RPPGF was not substrate inhibition. Based on mass spectroscopy and HPLC, α-thrombin did not cleave RPPGF (data not shown).
Because these peptides did not block α-thrombin binding to platelets and SFLLRN-mediated platelet activation, further studies determined whether these BK-related peptides interfered with α-thrombin proteolyzing of one of its platelet's receptors. GPIbα is a recognized α-thrombin binding site.25 26 Monoclonal antibody LJIb10, which is directed to the α-thrombin binding site on GPIbα, gave a single peak with much rightward projection on a flow cytogram of unstimulated platelets (Fig 8⇓, top left, lighter curve). In platelets that were activated with α-thrombin (1 nmol/L), the platelet epitope to LJIb10, the α-thrombin binding site on GPIbα, was markedly reduced, as indicated by a shift to the left in the flow cytogram (Fig 8, top left, darker curve). The degree of platelet activation from the washing procedure and after 1 nmol/L α-thrombin was such that only a small amount of CD62 became further exposed on the activated platelet surface (Fig 9⇓, bottom right). The presence of 1 mmol/L BK, MKRPPGFSPFRSSRIG, and RPPGF abolished the ability of α-thrombin to attenuate the antigenic expression of the LJIb10 epitope. This inhibition of expression of GPIbα by these peptides was sequence specific. Scrambled peptides of RPPGF (GPFPR and FPRPG) did not inhibit α-thrombin from altering the antigenic expression of this receptor (Fig 8). Additional studies were performed to determine whether the role of these peptides in preserving the epitope of monoclonal antibody LJIb10 was specific for this region on GPIbα or was characteristic of α-thrombin alteration of GPIbα in general. Investigations with the monoclonal antibody LJIb1, the von Willebrand factor epitope on GPIbα, revealed that the peptide MKRPPGFSPFRSSRIG also blocked α-thrombin from changing the antigenic expression of this epitope (Fig 9). These combined data indicated that these BK analogue peptides interfered with the alteration by α-thrombin of expression of GPIbα in general and did not specifically block the interaction by α-thrombin with the LJIb10 platelet epitope.
Investigations were next performed to determine whether these BK analogue peptides interfered with the cleaving by α-thrombin of the cloned thrombin receptor. On treatment of washed platelets with 1 nmol/L α-thrombin, there was a decrease in the antigenic expression of the epitope of the monoclonal antibody SPAN12, which was reared to the 12 amino acids (NATLDPRSFLLR) that bridge the α-thrombin cleavage site on the cloned thrombin receptor23 24 (Fig 10⇓). The rightward projection of the SPAN12 epitope seen on unstimulated platelets (Fig 10, top left, lighter curve) was shifted to the left when platelets were activated with 1 nmol/L α-thrombin (Fig 10, top left, darker curve), similar to the flow cytogram of nonimmune mouse IgG (Fig 10, bottom right). In the presence of 1 mmol/L BK and peptide MKRPPGFSPFRSSRIG, the loss of the SPAN12 epitope on the thrombin receptor on 1 nmol/L α-thrombin–activated platelets was prevented. However, an overlapping peptide, SPFRSSRIGEIKEETT, or an irrelevant peptide, FNQ12 (both at 1 mmol/L), did not prevent α-thrombin from altering the epitope of SPAN12 (Fig 10). These studies suggested that the BK analogue peptides actually prevented α-thrombin from cleaving its cloned receptor. Additional studies showed that peptide MKRPPGFSPFRSSRIG also blocked α-thrombin from removing the epitope of monoclonal antibody ATAP138. This antibody is directed to an antigen on the cloned thrombin receptor that should be preserved after the cleavage by α-thrombin of its receptor24 (Fig 11⇓). Because the ATAP138 epitope should be preserved on platelets after α-thrombin activation, its loss must not be due to proteolysis.
Because BK and other D4 peptide analogues prevented the loss of the SPAN12 epitope on the cloned thrombin receptor after α-thrombin treatment of platelets, we examined whether these peptides actually prevented cleavage of the cloned thrombin receptor by α-thrombin. Peptide NATLDPRSFLLR (NAT12) on HPLC had a single peak (Fig 12A⇓). As shown in Fig 12C, when NAT12 (peak 1) was treated with α-thrombin, its peak was reduced by 81% and two new peaks appeared to its left, constituting 44% (peak 3) and 37% (peak 2), respectively (Table 2⇓). In the presence of RPPGF (Fig 12D⇓ and Table 2⇓), peak 1 of NAT12 was reduced by only 57% after α-thrombin treatment, and its products, peaks 3 and 2, constituted only 31% and 26% of the areas of the chromatographs, respectively. The chromatograph of RPPGF alone (Fig 12B⇓) was between the chromatographs of the α-thrombin cleavage products of NAT12, ie, peaks 3 and 2, in the presence of RPPGF (Fig 12D⇓). In the presence of HK at 42% of its plasma concentration, α-thrombin reduced the size of the intact NAT12 peptide in peak 1 by only 32% (Fig 12E⇓ and Table 2⇓). Moreover, the two α-thrombin cleavage fragments, peaks 3 and 2, constituted only 18% and 14% of the areas of the chromatographs, respectively (Fig 12E⇓ and Table 2⇓). The third peak seen in Fig 12E⇓ between peaks 3 and 2 represented a peak from the HK preparation, not a new α-thrombin cleavage fragment. Furthermore, the majority of the HK, which came through the HPLC chromatograph before peak 3, was not shown in Fig 12E⇓. Finally, when an irrelevant kininogen peptide from D3 and not D4, LNA7, was used, no protection was noted of NAT12 cleavage by α-thrombin (Fig 12F⇓ and Table 2⇓). In the presence of LNA7, α-thrombin produced an 86% reduction in the size of peak 1 of intact NAT12, with 54% and 32% of the mass of peptide in peaks 3 and 2, respectively. The majority of peptide LNA7, which was seen before peak 3, was not shown in Fig 12F⇓.
The finding of the present study that BK and related analogues function as selective α-thrombin inhibitors embellishes the notion that plasma kininogens are antithrombins. Previous investigations have revealed that HK, LK, and D3 function as selective α-thrombin inhibitors by blocking [125I]-α-thrombin from binding to platelets.13 14 15 16 Further information suggests that HK can also function as an indirect α-thrombin inhibitor by blocking α-thrombin–induced, membrane-expressed platelet calpain from contributing to platelet aggregation.14 30 The present finding that BK and its analogues inhibit α-thrombin–induced platelet activation suggests a third mechanism by which HK and LK are antithrombins. Each of these antithrombotic activities of kininogens could be contributing to the constituent anticoagulant nature of the intravascular compartment. In fact, the presence of HK in human plasma requires a 2.5-fold increase in the concentration of γ-thrombin to induce platelet aggregation in platelet-rich plasma.14
Initial studies with platelet aggregation and secretion suggest that these peptides may be inhibiting platelet activation via more than one mechanism. The better inhibition of platelet aggregation compared with platelet secretion as measured through [14C]-5-hydroxytryptamine secretion suggested that these peptides may be interfering with fibrinogen binding to platelets. This assessment is not foreign to this field as previous studies have indicated that HK and fibrinogen are noncompetitive inhibitors of each other binding to granulocytes.31 Further HK blocks fibrinogen binding to activated platelets.31 Ca2+ mobilization studies, however, show that there is no dichotomy between platelet aggregation and secretion. BK and its analogues inhibit α-thrombin–induced Ca2+ mobilization, indicating that the level of inhibition of platelet activation by these peptides occurs before platelet aggregation.
Because BK and its analogues do not block α-thrombin binding to platelets or SFLLRN-induced platelet activation, they must be inhibiting the activation by α-thrombin of the thrombin receptor(s). The loss of the epitopes on GPIbα of both LJIb10 and LJIb1 monoclonal antibodies after α-thrombin activation of platelets is consistent with the findings of others that, on platelet activation, the parent receptor of these epitopes is internalized.32 The prevention of loss of these epitopes by BK and its analogues suggests that these peptides either could be binding to GPIbα and thus preventing all α-thrombin–dependent activities or are inhibiting the internalization of these epitopes by blocking α-thrombin at the level of activation of the cloned thrombin receptor.
The finding that BK and its analogues prevent the loss of the SPAN12 epitope on the cloned thrombin receptor after α-thrombin activation indicates that these peptides are preventing its proteolysis. This assessment was confirmed through the peptide NAT12 cleavage experiments. Peptide RPPGF and HK prevent α-thrombin from cleaving peptide NAT12 between arginine and serine. Inhibition of α-thrombin by these D4 peptides appears to be occurring through a unique mechanism. BK peptides inhibit at the level of the α-thrombin substrate, ie, the cloned thrombin receptor rather than on α-thrombin itself. How RPPGF and HK inhibit α-thrombin cleavage of NAT12 is not completely known. Bradykinin and its analogues could be binding to α-thrombin and preventing this specific activity. However, previous studies indicate that HK and LK do not form a complex with α-thrombin.13 15 Alternatively, BK and its related compounds could be binding to the cloned thrombin receptor. Further investigations are needed to distinguish between these mechanisms.
It is of interest that BK and its analogues prevent the loss of the ATAP138 epitope of the cloned thrombin receptor. Although this epitope is not proteolyzed after α-thrombin treatment, its loss after α-thrombin activation of platelets may be due to its internalization like GPIbα.32 Previous studies have indicated that the cloned thrombin receptor is internalized after platelet activation.33 34 This fact explains the loss of the epitope of monoclonal antibody ATAP138 even though that portion of the receptor was not cleaved by α-thrombin. Finally, although there is a marked change in the expression of GPIbα and the cloned thrombin receptor when platelets were treated with 1 nmol/L α-thrombin, there was relatively little increase in membrane expression of CD62 or P-selectin. Our platelet-washing procedure through centrifugation of uninhibited platelets probably contributed to some expression of CD62 antigen on unstimulated platelet samples. The degree of platelet activation with 1 nmol/L α-thrombin contributed to only a slight increase in platelet α-granule expression of CD62 on the activated platelet membrane. This finding has been previously recognized through flow cytometry of activated platelets.35
It is particularly interesting that RPPGF and other BK analogues that we call thrombostatins are selective inhibitors of the ability of α-thrombin to activate platelets but do not interfere with the ability of α-thrombin to cleave chromogenic substrates or to clot fibrinogen. RPPGF is the major stable breakdown product of BK, with a circulatory half-life of 4.2 hours.36 37 This breakdown product of BK is the result of angiotensin-converting enzyme cleavage of BK.38 RPPGF has no previously known function; it does not bind the B1 or B2 receptors of the kinins. The present ratios of BK and its analogues to α-thrombin to achieve inhibition of platelet activation do not suggest that this interaction occurs in vivo with these free peptides. However, in the context of its parent proteins, HK and LK, platelet activation by α-thrombin is inhibited by levels of 10% or 2% of its plasma concentration (ie, 50 nmol/L), respectively,13 16 and 42% of plasma HK (300 nmol/L) was able to prevent α-thrombin cleavage of NAT12. However, it is possible that the only inhibitory peptides in a solution of synthetic peptides are those that maintain the native conformation of this sequence in the parent protein. Synthetic peptides that have this conformation would constitute a much lower concentration than the total solution of peptides. If such conformational constraints are a requirement for activity, then it is possible that physiological levels of free peptide may be sufficiently high to mainfest the same degree of inhibition seen with this sequence in the parent protein. These data indicate that BK and its metabolites, RPPGF, in the context of their parent proteins may contribute to the constitutive anticoagulant nature of the intravascular compartment.
Selected Abbreviations and Acronyms
|D3||=||domain 3 of the kininogens|
|D4||=||domain 4 of the kininogens|
|GPIbα||=||platelet glycoprotein Ibα|
This work was supported by grant HL-35553 from the National Institutes of Health. We thank Drs Steven K. Fisher and Kyle R. Palmer of the University of Michigan for their assistance in performing the Ca2+ mobilization assays. We also thank Drs Lawrence F. Brass and Zaverio Ruggeri for their helpful discussions.
- Received December 28, 1995.
- Revision received December 28, 1995.
- Accepted January 30, 1996.
- Copyright © 1996 by American Heart Association
Rocha E, Silva M, Beraldo WT, Rosenfeld G. Bradykinin, a hypotensive and smooth muscle stimulating factor released from plasma globulin by snake venoms and by trypsin. Am J Physiol.. 1949;156:261-273.
Hasan AAK, Cines DB, Zhang J, Schmaier AH. The carboxy terminus of bradykinin and amino terminus of the light chain of kininogens comprise an endothelial cell binding domain. J Biol Chem.. 1994;269:31822-31830.
Nakashima M, Mombouli J-V, Taylor AA, Vanhoutte PM. Endothelium-dependent hyperpolarization caused by bradykinin in human coronary arteries. J Clin Invest.. 1993;92:2867-2871.
Smith D, Gilbert M, Owen WG. Tissue plasminogen activator release in vivo in response to vasoactive agents. Blood.. 1983;66:835-839.
Gohlke P, Linz W, Scho¨lkens BA, Kuwer I, Bartenbach S, Schnell A, Unger T. Angiotensin-converting enzyme inhibition improves cardiac function: role of bradykinin. Hypertension.. 1994;23:411-418.
Parratt JR. Cardioprotection by angiotensin converting enzyme inhibitors—the experimental evidence. Cardiovasc Res.. 1994;28:183-189.
Zanzinger J, Zheng X, Bassenge E. Endothelium dependent vasomotor responses to endogenous agonists are potentiated following ACE inhibition by a bradykinin dependent mechanism. Cardiovasc Res.. 1994;28:209-214.
Meloni FJ, Schmaier AH. Low molecular weight kininogen binds to platelets to modulate thrombin-induced platelet activation. J Biol Chem.. 1991;266:6786-6794.
Puri RN, Zhou F, Hu C-J, Colman RF, Colman RW. High molecular weight kininogen inhibits thrombin-induced platelet aggregation and cleavage of aggregin by inhibiting binding of thrombin to platelets. Blood.. 1991;77:500-507.
Hasan AAK, Cines DB, Ngaiza JR, Jaffe EA, Schmaier AH. High molecular weight kininogen is exclusively membrane bound on endothelial cells to influence activation of vascular endothelium. Blood.. 1995;85:3134-3143.
Jiang YP, Muller-Esterl W, Schmaier AH. Domain 3 of kininogens contain a cell binding site and a site that modifies thrombin activation of platelets. J Biol Chem.. 1992;267:3712-3717.
Schmaier AH, Colman RW. Platelet high molecular weight kininogen: identification and mechanisms of availability. In: Hawiger J, ed. Platelets, Receptors, Adhesion, Secretion, Part A. Methods Enzymol.. 1989;169:276-296.
Schmaier AH, Claypool W, Colman RW. Crotalocytin: recognition and purification of a timber rattlesnake platelet aggregating protein. Blood.. 1980;56:1013-1019.
Rasmussen UB, Gachet C, Schlesinger Y, Hanau D, Ohlmann EP, Obberghen-Schilling EV, Pouyssegur J, Cazenave J-P, Pavirani A. A peptide ligand of the human thrombin receptor antagonizes α-thrombin and partially activates platelets. J Biol Chem.. 1993;268:14322-14328.
Fisher SK, Domask LM, Roland RM. Muscarinic receptor regulation of cytoplasmic Ca2+ concentrations in human SK-N-SH neuroblastoma cells: Ca2+ requirements for phospholipase C activation. Mol Pharmacol.. 1989;35:195-204.
Grykiewicz G, Poenie M, Tsien RY. A new generation of Ca2+ indicators with greatly improved fluorescence properties. J Biol Chem.. 1985;260:3440-3450.
Molino M, Blanchard N, Belmonte E, Tarver A, Abrams C, Hoxie JA, Cerletti C, Brass LF. Proteolysis of the human platelet and endothelial cell thrombin receptor by neutrophil-derived cathepsin G. J Biol Chem.. 1995;270:11168-11175.
Brass LF, Vassallo RR Jr, Belmonte E, Ahuja M, Cichowski K, Hoxie JA. Structure and function of the human platelet thrombin receptor. J Biol Chem.. 1992;267:13795-13798.
DeMarco L, Mazzucato M, Masotti A, Fenton JW III, Ruggeri ZM. Function of glycoprotein Ibα in platelet activation induced by α-thrombin. J Biol Chem.. 1991;266:23776-23783.
DeMarco L, Mazzucato M, Masotti A, Ruggeri ZM. Localization and characterization of an α-thrombin-binding site on platelet glycoprotein Ibα. J Biol Chem.. 1994;269:6478-6484.
Herwald H, Hasan AAK, Godovac-Zimmerman J, Schmaier AH, Mu¨ller-Esterl W. Identification of an endothelial cell binding site on kininogens' domain D3. J Biol Chem.. 1995;270:14634-14642.
Charo IF, Feinman RD, Detwiler TC. Interrelations of platelet aggregation and secretion. J Clin Invest.. 1977;60:866-873.
Schmaier AH, Bradford HN, Lundberg D, Farber A, Colman RW. Membrane expression of platelet calpain. Blood.. 1990;75:1273-1281.
Gustafson EJ, Lukasiewicz H, Wachfogel YT, Norton KJ, Schmaier AH, Niewiarowski S, Colman RW. High molecular weight kininogen inhibits fibrinogen binding to cytoadhesins of neutrophils and platelets. J Cell Biol.. 1989;109:377-387.
Michelson AD, Benoit SE, Kroll MH, Li J-M, Rohrer MJ, Kestin AS, Barnard MR. The activation-induced decrease in the platelet surface expression of the glycoprotein Ib-IX complex is reversible. Blood.. 1994;83:3562-3573.
Hoxie JA, Ahuja M, Belmonte E, Pizarro S, Parton RG, Brass LF. Internalization and recycling of activated thrombin receptors. J Biol Chem.. 1993;268:13756-13763.
Brass LF, Pizarro S, Ahuja M, Belmonte E, Blanchard N, Stadel JM, Hoxie JA. Changes in the structure and function of the human thrombin receptor during receptor activation, internalization, and recycling. J Biol Chem.. 1994;269:2943-2952.
Ault K, Mitchell J. Analysis of platelets by flow cytometry. In: Darzynkiewicz D, Robinson JP, Crissman HA, eds. Flow Cytometry. 2nd ed. San Diego, Calif: Academic Press; 1994:275-294.
Majima M, Sunhara N, Harada Y, Katori M. Detection of the degradation products of bradykinin by enzyme immunoassays as markers for the release of kinin in vivo. Biochem Pharmacol.. 1993;45:559-567.
Graf K, Graf M, Auch-Schwelk W, Baumgarten CR, Bossaller C, Fleck E. Bradykinin degrading activity in cultured human endothelial cells. J Cardiovasc Pharmacol. 1992;20(suppl 9):S16-S20.