Antithrombotic Effect of Crotalin, a Platelet Membrane Glycoprotein Ib Antagonist From Venom of Crotalus atrox

PLATELET ADHESION and aggregation play important roles in hemostasis and thrombosis. Physiologically, platelets arrest bleeding from damaged blood vessels by sealing wound and initiating the repairing processes. Thrombin-mediated platelet activation and aggregation is partially related to glycoprotein Ib (GPIb)1,2 and this membrane glycoprotein also plays a role in regulating the interactions of platelets with their environment via linking with membrane-associated cytoskeleton.2Additionally, the interaction of platelet GPIb with von Willebrand factor (vWF) is a critical event, allowing platelet adhesion, aggregation, and subsequent thrombus formation in the vessels either with high shear rates or with damaged endothelium.3-5 The role of GPIb in the adhesion of platelets to endothelial cell matrix has been studied with monoclonal antibodies (MoAbs). An MoAb against GPIb inhibited the platelet adhesion to endothelial cell matrix, whereas an MoAb against vWF which inhibits the interaction of vWF with platelets has a less pronounced effect, indicating that GPIb has another role in platelet adhesion, apart from serving as the binding site for vWF.6 

Many of the currently investigated antithrombotic drugs interfere with the process of platelet aggregation.7-9 However, the inhibition of platelet adhesion may be an important and efficient process for preventing the platelet-rich thrombus formation in the damaged vessels with high shear rate. In addition, thrombin is an important stimulator for platelet aggregation under normal physiological and pathological conditions.10-12 Thrombin is an important mediator of platelet aggregation in stenosed canine coronary arteries.13 Therefore, an inhibitor of thrombin and vWF platelet interaction may be used in the prevention of thrombosis. Several snake venom constituents affecting blood coagulation and platelet aggregation, especially Arg-Gly-Asp (RGD)–containing peptides, have been extensively studied.14,15 In addition, GPIb agonists as well as GPIb antagonists have been isolated and characterized.16-20 We report here the purification of a protein, crotalin, from Crotalus atrox venom that specifically inhibits ristocetin-induced agglutination or aggregation without affecting the aggregation induced by adenosine diphosphate (ADP) or collagen. Of particular interest is that crotalin, when administered intravenously (IV), markedly delays platelet-rich thrombus formation of the irradiated mesenteric venules in a fluorescein sodium-treated mice model. Among the antithrombotic agents tested, crotalin appears not only to be the most efficacious agent in prolonging the time lapse for inducing platelet-rich thrombus formation in this model, but also exhibits this antiplatelet activity with a longer duration.

C atrox venom was obtained from LATOXAN (Rosans, France). All of DEAE-Sephadex A-50, Sephadex G-75, Mono-S, and Superose columns were purchased from Pharmacia (Uppsala, Sweden). Halysin was purified fromAgkistrodon halys venom as previously described.21ADP, U46619, collagen (type I, bovine achilles tendon), and fluorescein sodium were purchased from Sigma Chemical Co (St Louis, MO).

MoAb AP1 raised against platelet GPIb was generously provided by Dr Robert Montgomery (Blood Center of Southeastern Wisconsin Milwaukee) and 7E3, an MoAb raised against platelet GPIIb/IIIa complex, was supplied from Dr Barry Coller (The Mount Sinai Medical Center, New York, NY).

According to the modified method of a previous report,22vWF was purified from human plasma through cryoprecipitation, polyethylene glycol (PEG) sedimentation, and gel filtration with a Sepharose CL-4B column, and the purity of vWF was assessed by 5% sodium dodecyl sulfate (SDS)-gel electrophoresis.

Blood was collected from healthy human volunteers, who did not take any medication within the 2 weeks before the study, and anticoagulated with 3.8% sodium citrate (9:1, vol/vol). Citrated blood was immediately centrifuged for 10 minutes at 120g and 25°C, and the supernatant (platelet-rich plasma) was obtained. Human washed platelet suspension was prepared as previously described.22,41Washed platelets were suspended in modified Tyrodes' solution (pH 7.3) (in mmol/L: NaCl, 136.9; CaCl₂, 2; KCl, 2.7; MgCl₂, 1.0; NaH₂PO₄, 0.4; NaHCO₃, 11.9; glucose, 11.1) containing bovine serum albumin (3.5 mg/mL) and adjusted to about 3 × 10⁸platelets/mL.

The turbidimetric method,23 using a Lumi-Aggregometer (Chrono-Log, Havertown, PA), was used to measure platelet aggregation. The extent of aggregation was expressed in light transmission unit.

EDTA (2 mmol/L) and indomethacin (50 μmol/L) were added to platelet suspension at 6 minutes after the addition of thrombin (0.03 U/mL). After centrifugation with an Eppendorf Centrifuge (Model 5414; Hamburg, Germany) at 14,000 cpm for 2 minutes, thromboxane B₂ level of the supernatant was determined by thromboxane B₂ EIA Kit (Amersham, Buckinghamshire, UK).

The vial of enzymobead reagent (Bio-Rad, San Francisco, CA) was hydrated with 50 μL distilled water at 4°C for 1 hour, followed by adding 50 μL phosphate buffer, 50 μg crotalin, 0.5 mCi Na¹²⁵I, and 25 μL β-D-glucose (1%). The mixture was incubated at room temperature for 25 minutes.¹²⁵I-conjugated crotalin was eluted through a Sephadex G-10 column (0.7 × 23 cm; bed volume, 10 mL) with phosphate buffer. The radioactivity of ¹²⁵I-labeled crotalin was detected with radiocounter (Capintec, CRC-7) continuously.

It was performed according to previously described methods24,25 with slight modification. Aliquot of 0.4 mL of platelet suspension (4.5 × 10⁸/mL) were incubated in the presence of Tyrode solution (control), AP1, or excess unlabeled crotalin at room temperature for 2 minutes. Then,¹²⁵I-crotalin in various concentrations was added, and the mixture was gently shaken and incubated for 6 minutes. Platelet suspension mixture, 0.4 mL, was overlayed on sucrose solution (20%, wt/vol) and centrifuged for 5 minutes at 14,000 rpm (Eppendorf centrifuge, 5415). The radioactivities of supernatant (200 μL) and the cut-off tips containing platelet pellet were separately measured by using a γ-counter (Beckman, Fullerton, CA). The specific ¹²⁵I-crotalin binding to the platelet pellet was calculated by subtracting the nonspecific radioactivity bound from the total radioactivity bound. The number of crotalin binding sites per platelet and dissociation constant (kd) were calculated by the method of Scatchard.26 

Bleeding time of mice was measured by a modification of the method described by Dejana et al.27 Saline or various concentrations of crotalin was injected intravenously through a lateral vein of the mouse (ICR, with an average body weight of 20 g). A sharp cut of 2 to 3 mm from tail tip of mouse was made 10 minutes (except as noted) after injection. The tail was then immediately placed into a tube filled with saline, and kept at 37°C for measuring bleeding time.

Normal male mice (ICR) were used. The mice weighing between 18 and 20 g were anesthetized with sodium pentobarbital (50 mg/kg) by intraperitoneal injection. The method of this thrombogenic animal model has been described in detail.21,28,29 In brief, an external jugular vein was cannulated for the administration of dye and crotalin, and a mesenteric preparation was then mounted on a transilluminator and continuously rinsed with warm saline. The room temperature around microscope was kept at 37°C. Mesenteric venules with a diameter around 20 μm were selected for thrombogenic assay.

Fluorescein sodium (100, 150, or 200 μg/mL) was first injected through a cannula, and microvessels were irradiated by filtered light 10 minutes later. Simultaneously, a video timer was started. Platelet adhesion and aggregation were observed through a television monitor and time lapse for inducing the formation of platelet-rich thrombus, as reflected by cessation of blood flow, was taken. In the epi-illumination system, the light from a 100-W mercury lamp was excited by a filter (B-2A; Nikon) with a dichronic mirror (DM510; Nikon, Tokyo, Japan). The filtered light then irradiated a microvessel through an objective lens (20×). In this experimental model, platelet thrombus formation was induced and found to consist exclusively of platelets with pseudopod formation and degranulated appearance, accompanied with moderately damaged endothelial cells.21 

All data are presented as mean ± SEM (n). Student's t-test was used to assess the statistical differences.

Crotalin was purified from C atrox venom through columns of DEAE-Sephadex A-50 and Sephadex G-75 and refractionated by fast protein liquid chromatography using Mono-S (Fig 1A)and Superose columns (Fig 1B). The purified fraction migrated as a single band and the apparent molecular weight was estimated to be 30 kD under nonreducing or reducing conditions by SDS-polyacrylamide (15%) gel electrophoresis (Fig 1B, inlet) and named crotalin.

Amino acid analysis showed that crotalin is a polypeptide, consisting of about 260 amino acid residues (Asp/Asn 36, Glu/Gln 25, Ser 14, Gly 21, His 9, Thr 8, Ala 14, Arg 22, Pro 10, Tyr 6, Val 16, Ile 18, Leu 28, Cys 7, Phe 10, and Lys 16). However, the N-terminal amino acid sequence of crotalin was found to be blocked.

Crotalin showed a marked inhibitory effect on ristocetin (1.0 mg/mL)-induced human washed platelet aggregation with IC₅₀of 2.4 μg/mL (Fig 2). This inhibitory effect was independent of the incubation time of crotalin with platelets. Furthermore, crotalin was specific for ristocetin-induced platelet aggregation because it had little effect on the platelet aggregations caused by ADP (20 μmol/L) plus 200 μg/mL of fibrinogen, U46619 (1 μmol/L), or collagen (10 μg/mL). Crotalin also inhibited ristocetin (1 mg/mL)-induced platelet aggregation of platelet-rich plasma with IC₅₀ value of 6.3 μg/mL (Fig2), indicating that it might have an antithrombotic effect in vivo.

Similarly, crotalin (10 μg/mL) apparently did not affect platelet aggregation caused by high concentration of thrombin (>0.05 U/mL). However, crotalin (20 to 100 μg/mL) prolonged the latent period in triggering platelet aggregation in a dose-dependent manner, with a slight inhibition on the maximal aggregation (<30%) inhibition). The MoAb AP1 (40 to 80 μg/mL) showed a similar effect to that of crotalin (data not shown).

As shown in Table 1, crotalin at a lower concentration (5 to 10 μg/mL) inhibited thromboxane B₂formation of platelets stimulated by ristocetin and vWF. At higher concentration (20 μg/mL), crotalin significantly inhibited thromboxane B₂ formation of platelets caused by thrombin (0.03 U/mL), whereas it did not affect thromboxane B₂formation stimulated by collagen.

¹²⁵I-crotalin bound to platelets in a dose-dependent manner, reaching a saturated binding at 10 μg/mL (0.33 μmol/L) (Fig3). The Scatchard analysis of¹²⁵I-crotalin binding data showed that the binding sites of crotalin were 58,632 ± 3,152 per platelet with a kd value of 3.2 ± 0.1 × 10⁻⁷ mol/L (Fig 4).¹²⁵I-crotalin binding to platelets was blocked either by unlabeled crotalin (100 μg/mL, >85%) or AP1 (20 μg/mL, >90%), but not by the MoAb against GPIIb/IIIa, 7E3 (40 μg/mL), or 5 mmol/L EDTA (data not shown).

IV administration of crotalin to mice significantly prolonged the bleeding time in a dose-dependent manner (Fig5). Crotalin pronouncedly prolonged the bleeding time (>10 minutes) as measured 1 hour after a bolus injection of crotalin (300 μg/kg), and this effect slowly ran down thereafter (Fig 6). The bleeding time almost returned to the baseline level within 4 hours after the administration of crotalin. However, the platelet count, as measured at 10 minutes after the administration of crotalin (300 μg/kg), did not show a major change, although about 20% decreasement was observed (128 ± 10 × 10⁴, n = 4, control v104 ± 13 × 10⁴ platelets/μL, n = 4, experimental; P > .05). Even when the dose of crotalin was increased to 600 μg/kg, a transient slight decrease of platelet count (about 20%) was observed at the first 5 minutes after the administration of crotalin, and the platelet count returned to control level within 20 minutes.

As the given dose of fluorescein sodium was increased, the latent period in inducing platelet plug formation was shortened (Table 2). IV administration of crotalin at 300 μg/kg pronouncedly delayed platelet-rich thrombus formation and significantly prolonged the occlusion time in mice receiving different doses of fluoresein sodium (Table 2). The occlusion time was lengthened from 100 ± 14 to 311 ± 25 seconds in fluorescein sodium (150 μg per mouse)-treated mice after IV administration of crotalin. Crotalin dose-dependently prolonged the occlusion time in causing platelet plug formation, and administration of 300 μg/kg of crotalin resulted in the maximal lengthening of occlusion time (Table3). This antithrombotic effect lasted at least for 2 hours (Fig 7). On the other hand, a continuous infusion of prostaglandin I₂(PGI₂) at 0.5 μg/kg/min, as in our previous study,29 showed a maximal lengthening effect on occlusion time. Higher dose (2 μg/kg/min) of PGI₂ did not further increase its antithrombotic activity. Halysin, an RGD-containing peptide purified from venom of A halys, inhibited platelet aggregation via a competitive inhibition of fibrinogen binding to platelet GPIIb/IIIa.30 It completely inhibited ex vivo platelet aggregation of platelet-rich plasma induced by collagen (15 μg/mL) 20 minutes after the IV administration of halysin at the dose of 10 mg/kg (data not shown). In comparing the maximal effect of PGI₂, halysin, ancrod,29 and crotalin on the occlusion time in the same in vivo model, crotalin appears to be the most efficacious agent in prolonging the occlusion time (Table4).

Crotalin, a newly purified protein from the venom of C atrox,specifically inhibited ristocetin-induced platelet aggregation in vitro and exhibited the antithrombotic activity in vivo. Crotalin specifically inhibited platelet aggregation induced by ristocetin either in platelet suspension supplemented with vWF or in platelet-rich plasma with an IC₅₀ of 2.4 and 6.3 μg/mL, respectively. In contrast to RGD-containing peptides, crotalin at 40 μg/mL did not affect collagen- and U46619-induced platelet aggregation. Ristocetin-induced platelet aggregation was mediated through the initial binding of vWF to platelet GPIb, and subsequently resulted in the exposure of the fibriongen receptor.31,32 The¹²⁵I-crotalin binding site was 58,632 per platelet with a kd value of 3.2 × 10⁻⁷ mol/L. ¹²⁵I-crotalin binding to platelets was selectively inhibited by AP1, an MoAb raised against platelet GPIb, but not by 7E3, an MoAb raised against GPIIb/IIIa. The binding was not affected by EDTA, indicating that the binding process is divalent-cation independent. Through the analysis of binding data and its inhibitory activity on ristocetin-induced platelet aggregation, crotalin appears to be a selective antagonist of platelet membrane GPIb. Regarding the binding sites of crotalin, the estimation of 58,632 per platelet is higher than that estimated with GPIb MoAb (around 30,000). It has been reported that the binding sites of other venom GPIb antagonists ranged from 21,500 to 47,440.20 We also obtained a similar value of 60,000 per platelet as probed with another GPIb antagonist purified from A acutus (unpublished data, December 1996). Therefore, the varied number of binding sites among GPIb antagonists and MoAbs may be explained as following: the smaller molecular size of venom GPIb antagonists may get a better access to the surface canalicular system of platelets as compared to the bulky, bivalent structure of the MoAb. Earlier studies indicate that low concentration of thrombin exerts its effect through the binding of platelet membrane GPIb, a high-affinity binding site of thrombin.42 Kroll et al33suggested that the binding of vWF to GPIb leads to the activation of phospholipase A₂ and subsequent formation of thromboxane. In this study, crotalin as well as AP1 prolonged the latent period in triggering platelet aggregation caused by low concentrations of thrombin (0.03 U/mL), with a slight inhibition on platelet aggregation. Crotalin blocked thromboxane B₂ formation of platelets challenged by ristocetin plus vWF or low concentration of thrombin, confirming the hypothesis that the ligation of GPIb may lead to the activation of endogenous phospholipase A₂. Regarding the molecular characterization of crotalin, a platelet GPIb antagonist, the preliminary data show that crotalin is unique as a single chain polypeptide that is quite different from the known venom GPIb-binding proteins because they are heterodimer in nature, sharing a highly homologous sequence with C-type lectins.20 Whether it exists as a monomer or dimer under physiological condition is unknown. However, the detailed characterization of the physiocochemical properties of crotalin is in progress.

An IV infusion of crotalin lengthened bleeding time of mice in a dose-dependent manner. Maximal prolongation was observed during a period of 10 to 60 minutes after injection of crotalin (300 μg/kg), and bleeding time progressively returned to control values over a 4-hour period. However, the platelet count after the administration of crotalin did not show a major change. It has been reported that echicetin and jararaca GPIb-BP caused a transient thrombocytopenia in mice,18,19 and therefore it may be an advantage of crotalin when considering its potential use as antithrombotic agent. From its in vivo experiment, we suggest that crotalin may inhibit platelet aggregation as well as platelet adhesion to subendothelium in vivo through blocking the interaction of vWF with platelet GPIb.

As the first step in hemostasis or thrombosis, the binding of vWF to platelet GPIb is essential for platelet adhesion at high-shear blood flow.34 The platelets from patients with Bernard-Soulier Syndrome were defective in expression of functional GPIb-V-IX complex, and poorly adhered to subendothelium at all shear rates.2Much effort has recently been devoted to characterize the interaction of vWF and platelet GPIb at the molecular level with an aim of developing inhibitors that could be useful in the prevention of thrombosis.35-38 Recently, it has been indicated that inhibition of vWF-platelet GPIb interaction is effective in preventing acute restenosis after thrombolytic therapy.39 

In the present study we evaluated the antithrombotic effect of crotalin in a mouse model. Surprisingly, crotalin apparently delayed platelet-rich thrombus formation in mesenteric microvessels and its antithrombotic activity was dose dependent. The time course of its antithrombotic effect was consistent with that of its effect on bleeding time in mice (Figs 6 and 7).

Table 4 shows the minimal dose of PGI₂, halysin, ancrod, and crotalin in causing the maximal prolongation of occlusion time in this platelet-rich thrombus animal model. Halysin, an RGD-containing venom peptide, completely inhibited ex vivo platelet aggregation for 20 minutes after the administration of halysin at 10 mg/kg. Ancrod (1 U/kg), a thrombin-like enzyme, caused defibrinogenation and exhibited antiplatelet activity for 60 minutes.29 Of the compounds tested, crotalin showed the most pronounced effect in prolonging the occlusion time of the irradiated vessels in inducing platelet-rich thrombus formation as compared with PGI₂, halysin, and ancrod, indicating that blockade of the interaction between vWF and platelet GPIb may be a potential strategy in causing a marked antithrombotic effect. In addition, crotalin exhibits a antithrombotic activity with a longer duration as compared with short duration of PGI₂, halysin or another disintegrin, triflavin.40 

Considering the pharmacokinetic of crotalin, crotalin may be more active as an antithrombotic agent in mice than in human beings because the effective dosage of crotalin in mice ranged from 100 to 300 μg/kg, equivalent to 1.3 to 3.8 μg/mL (assuming 2.0 mL plasma per mouse), even if protein binding is neglected. However, the IC₅₀ of crotalin was about 6.3 μg/mL (in human platelet-rich plasma), five times higher than the effective dosage of 100 μg/kg in mice. On the other hand, the in vivo antithrombotic effect of crotalin in mice may result from both the antiplatelet and anticoagulant activities because the preliminary results show that crotalin prolonged the whole blood clotting time and activated partial thromboplastin time but not the prothrombin time as crotalin was administered IV (unpublished data, December 1996). However, its anticoagulant activity is under investigation.

In conclusion, crotalin specifically inhibited ristocetin-induced platelet agglutination in the presence of vWF through a selective binding of platelet membrane GPIb, resulting in a blockade of interaction between vWF and GPIb. Furthermore, crotalin markedly prolonged the bleeding time when administered IV into mice and was efficious in blocking platelet plug formation in vivo experimental model. Therefore, crotalin may be a valuable tool for developing a new class of antithrombotic drugs for clinic use through the study of its structure-activity relationship.

We appreciate the secretarial work of I.S. Peng.