Human Kininogens Regulate Thrombin Binding to Platelets Through the Glycoprotein Ib-IX-V Complex

TWO FORMS OF purified human plasma kininogen (kinin precursors) were described in plasma by Jacobsen and Kriz¹: high molecular mass kininogen (HK; molecular weight [Mr] = 120 kD) and low molecular mass kininogen (LK; Mr = 68 kD). HK is an α-globulin with a plasma concentration of 80 μg/mL.2 LK is a β-globulin with a plasma concentration of approximately 220 μg/mL.3 Both HK and LK are encoded by the same gene4 through alternate splicing mechanisms. Both HK and LK are identical for the sequences encoded by the first 9 exons, the heavy chain (62 kD), and part of exon 10 that codes for bradykinin (BK) and 12 additional amino acids. However, the mRNA for HK continues through the complete exon 10, whereas the mRNA for LK is spliced to exon 11. Both HK and LK bind to unstimulated platelets in a zinc-dependent fashion with 3,300 and 650 sites per cell and kds of 9.9 nmol/L⁵ and 27 nmol/L,6 respectively. The number of binding sites for HK on platelets is increased by exposure to thrombin, but the kd is unchanged. Puri et al7 first demonstrated that HK inhibits thrombin-induced platelet shape change and increase in intracellular Ca²⁺ by inhibiting the agonist's binding to platelets. The presence of HK and LK in plasma resulted in a 10-fold increase in the concentration of thrombin needed to stimulate platelets. LK also inhibits thrombin binding to platelets and blocks thrombin-induced secretion.6 The action of both HK and LK to inhibit thrombin binding to platelets implies that the common heavy chain is involved.

The heavy chain is composed of three domains with similar disulfide bonding patterns, each highly homologous to the crystalline cysteine protein inhibitor, cystatin.8 Jiang et al9 have isolated domains 1, 2, and 3 of kininogens (D1, D2, and D3) by limited proteolysis of purified LK and shown that D3, but not D1 and D2, inhibited ¹²⁵I-HK binding to platelets, indicating that this domain contains a cell-binding region. Herwald et al10 have defined a 13-residue segment in the region encoded by exon 9 that inhibits the binding of kininogens to endothelial cells. These investigators, using monoclonal antibodies (MoAbs), showed that there was an independent region of D3 responsible for inhibiting thrombin binding to platelets.

Because D3 is encoded by exons 7, 8, and 9 of the kininogen gene, we expressed D3 and each of these exons and all of their combinations as glutathione S-transferase (GST) fusion proteins in Escherichia coli and evaluated each for its ability to inhibit thrombin-induced platelet aggregation.11 D3 folds, as does the native protein, and is equally potent as HK. Only products containing exon 7 inhibited platelet aggregation induced by thrombin, with an IC₅₀ of less than 30 μmol/L. A deletion mutant of exon 7 product (K²³⁶-K²⁷⁰) did not block thrombin-induced platelet aggregation, whereas L²⁷¹-Q²⁹² inhibited with an IC₅₀ of 19 μmol/L. A heptapeptide L²⁷¹-A²⁷⁷ inhibited thrombin-induced aggregation of platelets with an IC₅₀ of 65 μmol/L. The effect is specific for the activation of platelets by thrombin, but not by ADP or collagen.

We considered three possibilities to account for the inhibition of thrombin binding to platelets by kininogens. (1) Kininogens might form a complex with thrombin. However, HK did not inhibit the activity of thrombin on chromogenic substrates or on fibrinogen-to-fibrin conversion. Moreover, enzyme-linked immunosorbent assay (ELISA) techniques failed to demonstrate binding to immobilized HK, and a heterobifunctional noncleavable cross-linking agent failed to produce a complex.11 Furthermore, Hasan et al12 also failed to show an HK-thrombin complex. (2) HK could bind to the hepta-spanning G-protein–linked thrombin receptor. However, HK does not inhibit the activation produced by SFLLRN, the newly exposed peptide after cleavage of this receptor by thrombin. Although this finding does not rule out the G protein-linked receptor, we propose a third hypothesis. Harmon and Jamieson13 have shown that glycoprotein (GP) Ib is the high-affinity binding site for thrombin with B_max = 50 sites/platelet and a kd of 0.3 nmol/L. Previous studies have shown that considerably more thrombin is needed to stimulate Bernard-Soulier platelets14 lacking GP Ib-IX-V, similar to the finding with kininogens.6 7 Using MoAbs to and ligands of GP Ib-IX-V and cells transfected with this platelet receptor complex, as well as recombinant and synthetic peptides derived from kininogen D3, we tested the hypothesis that kininogen inhibits thrombin binding to GP Ib, thereby modulating the ability of thrombin to aggregate platelets.

¹²⁵I was purchased from ICN Pharmaceuticals, Inc (Irvine, CA). Avidin-fluorescein isothiocyanate (FITC) was purchased from Sigma (St Louis, MO). All other reagents were of the best purity available. Human α-thrombin was a generous gift of Dr John W. Fenton II (Division of Laboratories and Research, New York State Department of Health, Albany, NY) and had an activity of 2400 NIH U/mg and a purity of 99.24%. Human HK (1.6 mg/mL), migrating as a single band of 120,000 daltons (reduced) on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), was purchased from Enzyme Research Laboratories (South Bend, IN). Echicetin was a gift of Dr Manling Peng (Department of Biochemistry, Temple University School of Medicine, Philadelphia, PA). All statistical probabilities were assessed using the Student's t-test.

Platelet isolation and aggregation.Platelet isolation and aggregation were performed as described previously.11 Briefly, an aliquot of platelets (>100,000/μL) in a 0.5-mL cuvette at 37°C was incubated with stirring and added inhibitors for 2 minutes; then, aggregation was monitored for 5 minutes after the addition of 2 nmol/L thrombin. Platelets deficient in GP Ib-IX-V were obtained from a patient with Bernard-Soulier disease15 and prepared similarly, except that the final platelet concentration was 36,000/μL. Higher concentrations of platelets were not possible due to the very low concentration of platelets in the patient's plasma.

Iodination of thrombin and HK.Iodination of thrombin and HK was performed by the iodogen method.16 The resulting preparation of ¹²⁵I-α-thrombin had a specific radioactivity of 1.9 × 10⁶ cpm/μg. HK was prepared similarly, yielding a specific radioactivity of 1.3 × 10⁸ cpm/μg. ¹²⁵I-HK retained greater than 90% of its coagulant activity compared with the unlabeled protein.

MoAbs.The MoAbs AK-1, which binds to GP IX close to its site of interaction with GP Ib; AK-2, which recognizes the 45-kD ligand-binding peptide tail of GP Ib; and AK-3, which binds to the macroglycopeptide of GP Ib,17 were provided by Dr William J Booth (University of Sydney, Westmead, Australia). AK-2 blocks thrombin binding to GP Ib, but does not inhibit thrombin-induced platelet aggregation, whereas AK-1 and AK-3 have no effect on thrombin binding. The MoAb to GP IX, FMC-25 (which does not block thrombin binding), was supplied by Dr Manling Peng and developed by Berndt et al.18 SZ 2, which only blocks thrombin binding at low concentrations but can inhibit thrombin-induced platelet aggregation, was purchased from Biodesign International (Kennebunk, ME) and recognized GP Ib (extracellular domain).19 The biotinylated antibody TM-60 (260 μg/mL) that binds to the N-terminal 45-kD peptide of GP Ib and also inhibits thrombin-induced platelet aggregation20 was a generous gift of Dr Naomasa Yamamoto (The Tokyo Metropolitan Institute of Medical Science, Tokyo, Japan). The antihuman α-thrombin MoAb EST-7 was purchased from American Diagnostica Inc (Greenwich, CT).

Recombinant proteins and synthetic peptides.Exon products from the D3 portion of HK were expressed as GST-fusion proteins in E coli, cleaved by thrombin, and purified by high-pressure liquid chromatography (HPLC). Peptides were synthesized, purified, and characterized as described.11 

Platelet binding assays.Platelets were adjusted to yield 1 × 10⁵ cells in 1 μL in a total volume of 50 μL, including all test reagents. HK, thrombin, or antibodies were incubated for 30 minutes. To measure specific binding, the total ¹²⁵I-thrombin binding was determined and the binding in the presence of a 100-fold molar excess unlabeled ligand (nonspecific binding) was subtracted. All platelet binding assays were performed in 96-well MultiScreen DV 0.65 μmol/L hydrophilic plates and washed with 10 mg/mL bovine serum albumin (BSA) in HEPES/Tyrodes buffer, pH 7.3, on a MultiScreen Vacuum Filtration Manifold system (Millipore Corp, Bedford, MA). After washing and removal of the excess radioactive reagent by vacuum extraction, the filters were separated from the plate with a Millipore Multi-punch apparatus into individual tubes and counted for 1 minute in an LKB-Wallac 1274 RiaGamma counter (LKB-Wallac, Turku, Finland). In the fluorescence assays, biotinylated antibody binding was detected with FITC-labeled avidin (10 μL of a 1-20 dilution of the stock reagent) and incubated for 30 minutes at 23°C. After washing and removal of the supernatant by vacuum, plates were analyzed for fluorescent antibody bound to the platelets at an excitation wavelength of 485 nm and detected at an emission wavelength of 530 nm in a Millipore fluorescence microtiter plate reader.

Cell lines.L2H/V cells are transfected mouse L cells that stably express the full GP Ib-IX-V complex (GP Ibα, GP Ibβ, GP IX, and GP V).21 To make this cell line, we transfected the GP V gene22 into a previously isolated clone (L2H)23 that only expressed three components of the complex (GP Iba, GP Ibβ, and GP IX). The cells were cultured in a mixture of Dulbecco's modified Eagle's medium and F12 medium (DMEM/F12, 1:1; Life Technologies Inc, Grand Island, NY) supplemented with 10% heat-inactivated fetal bovine serum (FBS). The medium contained 1 mmol/L sodium hypoxanthine, 4 μmol/L aminopterin, 160 μmol/L thymidine (HAT; Life Technologies, Inc) and 500 μg/mL hygromycin. The cells were maintained in an atmosphere of 5% CO₂ and 99% humidity.

Flow cytometry ligand binding assay.Thrombin binding to L2H/V cells was evaluated with flow cytometry using a FACScan flow cytometer (Becton Dickinson, San Jose, CA). Confluent L2H/V and parental L cells (negative control) were washed with 0.1 mol/L Na phosphate, pH 7.4, 0.15 mol/L NaCl (phosphate-buffered saline [PBS]) and incubated overnight with DMEM/F12 medium containing 1% BSA to eliminate any potential effects from von Willebrand factor (vWF ) or protease inhibitors that may have been present in the FBS. The cells were detached with 0.25 mmol/L EDTA, washed with PBS, and then suspended in PBS containing 1% BSA to a concentration of 2 to 2.5 × 10⁶ cells/mL. HK, peptide, or buffer was then added and incubated for 30 minutes at 37°C. Thrombin was added to the cell suspension to a final concentration of 0.5 nmol/L and the cells were incubated at 37°C for 20 minutes. Unbound thrombin was removed by washing the cells twice with PBS. The thrombin that remained bound to the cells was recognized by staining with 0.15 μg/mL of the monoclonal antithrombin antibody EST-7 (American Diagnostica) for 60 minutes at room temperature followed by 1 mg/mL of FITC-conjugated rabbit antimouse IgG (Zymed, South San Francisco, CA) for 30 minutes at room temperature. At the end of the incubation, unbound antibodies were removed by washing the cells twice in PBS and binding was analyzed by flow cytometry. An argon ion laser was used to excite the fluorophor at 488 nm; a 520-nm bandpass filter was used to detect the emitted light.

Two methods were used to control for nonspecific binding. First, the binding of thrombin to L2H/V cells was compared with thrombin binding to parental L cells, which served as negative controls. A second control was to omit thrombin and incubate the cells only with the primary and secondary antibodies.

LNAENNA-induced shift of thrombin-induced platelet aggregation.The heptapeptide L²⁷¹-A²⁷⁷ (LNAENNA) coded for by part of exon 7 of HK has been shown to inhibit thrombin-induced platelet aggregation in a concentration-dependent manner (IC₅₀ = 65 μmol/L)11 at fixed thrombin concentrations. In this experiment, a fixed concentration of LNAENNA shifted the dose response of thrombin to induce platelet aggregation more than sevenfold, from an EC₅₀ of 2.5 nmol/L in the presence of LNAENNA (400 μmol/L) compared with 0.35 nmol/L for platelets in its absence of LNAENNA (Fig 1). This finding is similar to the increased requirement for thrombin to activate platelets previously shown in the presence of plasma kininogens.7,11 These data recall the findings of reduced thrombin aggregation and a shift of concentration dependence seen in Bernard-Soulier platelets lacking the GP Ib-IX-V receptor complex.24 These results raised the possibility that the kininogen inhibition of thrombin-induced platelet aggregation may be due to interaction of kininogens with GP Ib-IX-V on the platelet membrane, modulating the high-affinity binding of thrombin to this receptor.

Thrombin-induced aggregation of Bernard-Soulier platelets.Blood was collected in ACD from a patient with well-characterized Bernard-Soulier disease whose platelets contained less than 3% of normal GP Ib-IX-V.15 The platelets were separated from plasma by gel filtration. These platelets are present at $110$ the normal concentration in plasma; therefore, the final platelet count was significantly reduced in the aggregation cuvette (36,000 platelets/μL). Preincubation with 2 μmol/L HK followed by stimulation with thrombin (4 nmol/L) resulted in an inhibition of the rate of platelet aggregation of only 14% compared with these platelets in the absence of HK. In contrast, as previously reported,7 11 thrombin aggregation of normal gel-filtered platelets was inhibited by the presence of HK. Under similar conditions (4 nmol/L thrombin and 2 μmol/L HK), the inhibition was 52%. Full inhibition was achieved when thrombin was reduced to 2 nmol/L in the normal platelets.

Effects of MoAbs and ligands of the GP Ib-IX-V complex on ¹²⁵I-HK binding to platelets.We postulated that HK might bind to the GP Ib-IX-V complex on unstimulated platelets. To test this hypothesis, we evaluated the binding of ¹²⁵I-HK to normal gel-filtered platelets in the presence of several MoAbs that recognize various regions of the GP Ib-IX-V complex (Fig 2). HK binding was reduced by more than 50% by 1.68 μmol/L of AK-2, which recognizes the 45-kD peptide tail region of the extracellular domain containing the thrombin and vWF binding sites of GP Ib, and 1.68 μmol/L of AK-1, which binds to GP IX near the ligand binding sites on GP Ib. In contrast, HK binding was minimally inhibited (<20%) by AK-3 (1.68 μmol/L), which recognizes the macro-glycopeptide core of the α-chain of GP Ib, which does not contain the 45-kD peptide tail.17 An MoAb directed to GP IX, FMC-25 (1.68 μmol/L), strongly (>75%) blocked HK binding. Echicetin (1.68 μmol/L), a snake venom protein that binds tightly to both the thrombin and vWF binding regions on the N-terminal portion of GP Ib,25 inhibited binding of HK to platelets greater than 50%. Bovine vWF (1.68 μmol/L), which binds to the N-terminal region of GP Ib in a region close to the thrombin binding site,26 blocks HK binding (40%). These results show that ¹²⁵I-HK binding to platelets could be inhibited by antibodies directed to the GP Ib-IX-V complex or protein ligands that have been previously shown to bind to this high-affinity platelet receptor for thrombin. All antibodies and ligand means were significantly lower compared with ¹²⁵I-HK alone (P < .0005), with the exception of AK-3.

Effect of HK on the ability of ¹²⁵I-SZ 2 MoAb binding to platelets.The MoAb SZ 2, directed to the α-subunit of GP Ib and able to precipitate the GP Ib-IX-V complex, has been shown to partially inhibit thrombin binding to platelets.19 To determine if HK could directly block access to the GP Ib-IX-V complex by the MoAb, ¹²⁵I-SZ 2 (53 nmol/L) was incubated in the absence or presence of increasing amounts of HK (0.55 to 1.01 μmol/L) and compared with the effect of excess unlabeled SZ 2 (0.88 μmol/L; Fig 3). A 10-fold molar excess of HK significantly (P < .002) inhibited the binding of SZ 2 to platelets (44% ± 11.7% standard error of the mean [SEM]) and could further reduce SZ 2 binding at increasing concentrations of HK with 84% ± 3.2% SEM inhibition found at 1.01 μmol/L. A 20-fold excess of unlabeled SZ 2 was capable of reducing labeled binding to 67% ± 7.9% SEM compared with incubation with ¹²⁵I-SZ 2 alone. This experiment suggests that HK may bind directly to GP Ib and compete with the antibody for the high-affinity binding site for thrombin on the platelet surface.

Effect of HK on biotinylated TM-60 binding to platelets.In this experiment, we used a fixed HK concentration (1.3 μmol/L) to measure the concentration dependence of the binding of the biotinylated MoAb, TM-60, which inhibits thrombin- and ristocetin-induced platelet aggregation directed to GP Ib-IX-V. A fluorescence assay was used to measure the amount of MoAb, TM-60, bound to platelets (Fig 4). At low concentrations of TM-60 (10 nmol/L), binding was completely inhibited by HK. As the ratio of TM-60 to HK was increased, additional binding of TM-60 was observed. Thus, HK competed for binding with TM-60 directed to GP Ib-IX-V, presumably by binding to GP Ib, and the inhibition was reversed by high concentrations of TM-60 (60 nmol/L or higher).

Recombinant polypeptide K²⁷⁰-Q²⁹² from kininogen D3 alters the binding of TM-60 to platelets.The recombinant polypeptide, K²⁷⁰-Q²⁹², was previously shown to inhibit thrombin-induced platelet aggregation with an IC₅₀ of 19 μmol/L. When K²⁷⁰-Q²⁹² was preincubated with platelets, TM-60 binding was blocked with increasing concentrations (2 to 100 μmol/L), reaching a plateau at a maximum binding of 60% of control (Fig 5). Similarly, when 100-fold molar excess (to TM-60) of D-Phe-Pro-Arg-chloromethyl ketone (PPACK)-inactivated α-thrombin was used, the antibody binding to platelets was also 60% of control. The polypeptide fragment of kininogens could only achieve partial blocking of TM-60 binding, unlike intact HK, which was a more potent blocker of this antibody directed to GP Ib. GST alone did not alter the binding of TM-60.

Kininogen derivatives alter ¹²⁵I-thrombin binding to the platelet high-affinity site.Direct binding of ¹²⁵I-thrombin to platelets was performed to determine if the number of high-affinity binding sites could be reduced by various fragments derived from kininogen D3 (Fig 6). Because the kd of the high-affinity site is 0.3 nmol/L, we first studied thrombin binding from 0.1 to 0.6 nmol/L as reported previously²⁷; 0.3 nmol/L yielded approximately 50% of maximal binding in that range. Thrombin (0.3 nmol/L) binding to platelets alone yielded an average of 41 ± 13.3 SEM sites per cell, in good agreement with previous studies27 that found that GP Ib-IX-V complex only accounted for the highest affinity, lowest capacity binding site. We found that a peptide (C³³³NAEVYVVPGC³⁴³; 100 μmol/L), derived from exon 9 (with the penultimate W changed to G and terminal amino acid changed from E to C to allow disulfide formation and folding), similar to the binding site for HK to endothelial cells,10 did not significantly alter the number of high-affinity sites for thrombin (58 ± 40 SEM). In contrast, the peptide derived from the exon 7 product of HK, K²⁷⁰-Q²⁹² (100 μmol/L), and the heptapeptide L²⁷¹-A²⁷⁷ (100 μmol/L) could reduce high-affinity thrombin binding to 29 ± 11.5 SEM and 6 ± 6.6 SEM sites per platelet, respectively (P < .05). These results show that a primary site on kininogens for inhibition of the thrombin binding to the high-affinity receptor (GP Ib-IX-V complex) resides within the exon 7 product of kininogen D3.

Thrombin binding to cellular GP Ib-IX-V complex is regulated by HK.The preceding experiments indicate that thrombin stimulation of isolated platelets is partially regulated by HK on the platelet surface, and the GP Ib-IX-V complex may be the primary site for this regulatory effect. However, the hepta-spanning G-protein–coupled receptor represents another potential site. Therefore, L cells, a cell line derived from mouse fibroblasts transfected with GP Ib-IX-V complex, were used to measure the interaction of thrombin and HK with this receptor complex. FITC-labeled thrombin (590 nmol/L) was incubated with L cells or L cells expressing the GP Ib-IX-V complex in the presence or absence of HK (1.3 μmol/L; Fig 7). These cells have been shown to bind thrombin with high affinity.27a Thrombin could be detected on L cells, by flow cytometry, expressing the GP Ib-IX-V complex with a fluorescent intensity of more than four times (19.2 ± 2.4 SEM) the levels of control L cells without GP Ib-IX-V (4.4 ± 1.2 SEM). HK significantly (P < .0001) reduced fluorescent thrombin binding to 12.3 ± 3.8 SEM or by 53% after subtraction of the contribution of the untransfected control cells. This experiment shows that HK competes with thrombin for high-affinity sites to the cellular GP Ib-IX-V complex.

One receptor involved in the ability of HK to modulate thrombin interaction with platelets appears to be the GP Ib-IX-V complex. Several lines of evidence support this conclusion. Platelets from patients with Bernard-Soulier disease require 10-fold the thrombin concentration to cause complete aggregation, and, unlike normal platelets, this thrombin-induced aggregation is not inhibited by HK. The 10-fold increase in thrombin concentration for platelet activation is similar to that observed when kininogen is added to either purified systems or to kininogen-deficient plasma.7 Certain antibodies to GP Ib, such as TM-60, also induced this shift of the thrombin dose-response curve.28 It is notable that the inhibitory effect of HK is only observable at thrombin concentrations less than 2 nmol/L, in the range of the known high-affinity binding site for thrombin to GP Ib. MoAbs to the extracellular portion of GP Ib, AK-2, which recognize the 45-kD peptide tail containing the ligand binding domain,17 and AK-1 to GP IX both block HK binding to platelets. Alternatively, AK-3, directed to the macroglycopeptide lacking the ligand-binding domain, causes minimal inhibition. Conversely, HK inhibits the binding of MoAb SZ 2 (directed to GP Ib) to platelets. The binding of TM-60 (directed to the GP Ib-IX-V complex and known to inhibit thrombin activation of platelets) to platelets is also inhibited by HK. The MoAbs, FMC-25 and AK-1, both to GP IX, inhibit the binding of HK to platelets, suggesting that GP IX and, thus, the heterodimeric complex are also involved in HK binding. Alternatively, antibodies to GP IX might interfere with higher order complexes with more than one copy of each component. Because antibodies can produce effects by steric hindrance, we also used a low molecular mass (28-kD dimer) ligand, echicetin, that is known to bind to GP Ib,25 inhibiting both thrombin and vWF29 interaction with that receptor. This naturally ocurring inhibitor blocks binding of HK to platelets. The vWF and thrombin binding sites, although not identical on GP Ib, are spatially proximate. We have also observed that bovine vWF inhibits HK binding. Bovine vWF was used because it does not require ristocetin for its binding,30 as does human vWF, thus obviating that complication. These studies all indicate that HK binds to the GP Ib-IX-V receptor on platelets.

To show that HK inhibits the high-affinity binding of thrombin to platelets, we have performed direct binding of ¹²⁵I-thrombin to platelets at very low concentrations. We have reproduced the results of Harmon and Jamieson13 and De Marco et al27 and found 41 sites compared with 40 to 50 sites in their studies. We have shown that HK inhibited this binding. Finally, we have shown that HK partially inhibits the binding of thrombin to cells transfected with GP Ib-IX-V. The findings that minimal binding was observed in the untransfected cells and that the binding to the transfected cell was inhibited by HK, support the ability of HK to compete with thrombin for binding to GP Ib-IX-V. However, the inhibition of high-affinity thrombin binding was only 52%. We are uncertain of the explanation for why we were only able to inhibit thrombin binding to the L cells by 50%. The existence of other potential binding sites in mouse cells is not likely to be the reason because the control is an untransfected L cell that is basically equivalent to cells expressing the GP Ib-IX-V complex. The only differences between the two cells other than the presence of the complex is that those expressing the complex are also expressing the genes that mediate resistance to HAT medium (the herpes simplex thymidine gene) and hygromycin are grown in these selective media. It is possible, although unlikely, that expression of these resistance genes or growth in the media somehow increases the nonspecific thrombin binding of the complex-expressing cells. Another possibility is that because the ratio of HK to thrombin was about two to one, transfected L cells may have a higher number of copies of GP Ib-IX-V, resulting in an insufficient ratio for full inhibition.

Recently, Connolly et al31 showed that gene knockout of the G-protein–coupled thrombin receptor failed to abolish thrombin-induced aggregation in murine platelets but did abolish thrombin responses in the murine fibroblasts of that mouse. The discovery that a second protease-activated receptor (PAR2) is part of a thrombin receptor gene family32 suggests that existence of a putative second G-protein–linked thrombin receptor on platelets. However, it appears that murine fibroblasts (L cells) contain the original G-protein–coupled receptor.

In a recent study,11 we mapped the region on kininogen domain 3 to the N-terminal segment of exon 7, the product K²⁷⁰-Q²⁹² (IC₅₀ = 19 μmol/L), and defined the minimum necessary sequence as LNAENNA (IC₅₀ = 65 μmol/L) for inhibition of thrombin-induced platelet aggregation. In contrast, a peptide encoded by exon 9 fails to modulate thrombin activation of platelets. That polypeptide is similar to a region reported by Hasan et al12 as a binding site on kininogens to endothelial cells. The same investigators found two other discontinuous sites on kininogen that make up a composite binding site. It is conceivable that the exon 9-encoded peptide contains a binding site on platelets distinct from GP Ib-IX-V that does not interfere with thrombin binding or activation of platelets. Our findings suggest that both K²⁷⁰-Q²⁹² and LNAENNA could interact with GP Ib-IX-V. In favor of this theory is the demonstration that these polypeptides can shift the potency of thrombin about 10-fold, similar to HK in platelet-rich plasma.7 The polypeptides also directly inhibit ¹²⁵I-thrombin binding to platelets. LNAENNA is a discrete region on the surface of D3 that probably represents a minimum sequence for kininogens to inhibit thrombin binding.11 

The binding of thrombin to the high-affinity site of GP Ib may allow a local increase in thrombin to a level that can then cleave the intermediate-affinity seven-transmembrane receptor to expose SFFLRN, which then binds to the exosite on the hepta-spanning polypeptide. The precise mechanism by which the GP Ib-IX-V complex is involved in thrombin's action is unclear. One possibility is that the complex sequesters thrombin and increases its local concentration to a point at which it can cleave the intermediate-affinity seven-transmembered-domain receptor. Alternatively, the complex may itself be involved in thrombin-induced signaling, which may act synergistically with the contribution of the seven-transmembered-domain receptor. Kininogen seems to act as a competitor for the receptor, with 10-fold more thrombin being required to stimulate the platelet, a situation similar to that seen in Bernard-Soulier disease platelets.

The HK peptide, LNAENNA, contains an identical tetrapeptide, NAEN, found in GP Ib at residues 223-226 and located within a 24 amino acid sequence that has been shown to inhibit thrombin-induced platelet aggregation.33 This observation strongly suggests that this region of HK may compete with GP Ib, thus blocking thrombin binding. The presence of an identical sequence in GP Ib is seemingly contradictory to the observation that HK blocks thrombin by binding directly to GP Ib and preventing access to thrombin. This apparent contradiction might be reconciled if the HK-homologous sequence in GP Ib is involved in an intramolecular association with another region of the polypeptide. Competition for that region by HK would be expected to change the conformation of GP Ib and, thus, prevent thrombin binding. Thrombin binds to at least one other site distinct from NAEN,34 and the bound HK could interfere sterically with thrombin binding by competing with GP Ib for binding of thrombin. This hypothesis would also explain why HK blocks the binding of the antibody AK-2, which does not inhibit thrombin-induced platelet aggregation.

Similar to the findings of Greco and Jamieson35 with vWF, we have also observed a maximal inhibition of the GP Ib antibody, TM-60, by HK of only 50%. One explanation of this finding would be provided by a two-receptor model for HK binding to platelets. Stimulated platelets bind fivefold as much HK as do unstimulated platelets, and this increased binding is due to thrombospondin expressed on the activated platelet surface.36 The finding that HK blocking of thrombin antibodies or GP Ib antibodies did not demonstrate a requirement for zinc, in contrast to the absolute requirement for HK binding as reported by Greengard et al37 for stimulated platelets, again suggests the existence of two receptors for HK for platelets. The zinc requirement for HK may be unnecessary for the high-affinity receptor (GP Ib-IX-V complex), but essential for binding to stimulated cells.

The ability of HK to inhibit thrombin-induced aggregation of platelets has two important ramifications. First, in conditions in which HK is depleted, such as the adult respiratory distress syndrome (due to septic shock and multiple trauma),38 thrombin may stimulate platelets at lower concentrations, enhancing disseminated intravascular coagulation. Second, polypeptides from HK or GP Ib or peptomimetic derivatives of these sequences may serve to inhibit thrombin-induced platelet aggregation in situations such as reocclusion after angioplasty or thrombolysis39 without interfering with the other hemostatic functions of thrombin and/or its action on fibrinogen to form a fibrin clot.

In a recent abstract, Joseph et al,40 using an HK affinity column, isolated GPIb from platelet membrane lysate, thus confirming our results by a totally different approach.

The authors thank Dr Yingzhang Lin and Leonard Rick for their assistance in producing the recombinant proteins used in this study.