Background The procoagulant effect of anionic phospholipid may play a major role in the development of arterial thrombosis.
Methods and Results Annexin V, a calcium-dependent anionic-phospholipid–binding protein, was expressed and isolated from Escherichia coli and its antithrombotic effect examined in a rabbit carotid artery thrombosis model. A partially occlusive thrombus was formed in the left carotid artery by application of electric current to produce an ≈50% occlusion of the lumen. After the current was discontinued, flow ceased completely within 42±12 minutes (n=6) because of continuing platelet/fibrin thrombus formation. When annexin V was given at doses of 2.8 to 16.6 μg · kg−1 · min−1 for a period of 180 minutes, starting at the time the current was stopped, there was a dose-dependent inhibition of thrombus formation. At a dose of 5.6 μg · kg−1 · min−1, blood flow remained patent throughout the infusion and for an additional 60 minutes after the infusion was stopped. In addition, there was a decrease in thrombus weight (16±7.4 versus 2.0±1.0 g), 125I-fibrin deposition (≈45% reduction, P<.001), and 111In-labeled platelet accumulation (≈43% reduction, P<.001). Prior mixing of annexin V with phosphatidylserine micelles abolished the antithrombotic effect of annexin V, whereas mixing with phosphatidylcholine micelles had no effect. The antithrombotic effect of annexin V was not associated with bleeding tendency, as judged by the amount of blood absorbed in a gauze pad placed in a surgical incision extending to the muscle tissue and by the standard template bleeding time.
Conclusions These observations support a potentially important role for anionic phospholipid exposure in platelets in arterial thrombosis, and inhibition of this activity could be a novel target for therapy in coronary thrombosis and stroke and after angioplasty.
Thrombosis of arteries plays a key role in the pathogenesis of a variety of ischemic syndromes, including unstable angina, myocardial infarction, and stroke. Arterial thrombus formation usually involves elements of vessel wall damage and platelet activation and aggregation. In addition to forming occlusive aggregates, activated platelets provide a highly efficient catalytic surface for the activation of prothrombin and factor X. The precise components of these catalytic surfaces are not fully elucidated, but anionic phospholipid constitutes a major component for binding sites for proteases and cofactors.1 2 In platelets and in most other cells examined, anionic phospholipids are present only in the inner leaflet of the lipid bilayer membrane.3 After the activation of platelets by certain agonists, anionic phospholipids move from the inner to the outer leaflet of the lipid bilayer; this rearrangement of membrane phospholipid is thought to be a major mechanism responsible for platelet procoagulant activity.4 5 6 In addition, cultured endothelial cells have also been shown to provide a phospholipid-dependent procoagulant surface in vitro.7 8 Furthermore, cells undergoing apoptosis have been shown to expose anionic phospholipid on their surfaces.9 Thus, the exposure at the sites of vascular injury of anionic phospholipids from activated platelets and other cells may play a major role in the initiation, growth, and extension of the thrombus and may provide a novel target for antithrombotic therapy.
The annexins are a family of calcium-dependent anionic-phospholipid–binding proteins.10 A member of this family, annexin V, was originally isolated from placenta, characterized as placental anticoagulant protein I, and sequenced.11 12 The same protein was also isolated and sequenced under different names.13 14 Annexin V binds anionic phospholipids with a very high affinity in a calcium-dependent manner.15 We have shown previously that annexin V binds to anionic phospholipid on platelets and blocks the binding of factors Xa and Va to platelets.16 17 In addition, annexin V has been shown to be an effective inhibitor of thrombus formation in a venous thrombus model and in vitro perfusion models.18 19 20 In the present study, we examined the potential role of annexin V as an inhibitor of arterial thrombosis in a carotid artery injury model developed in our laboratory. Our results indicate that intravenous infusion of annexin V can significantly inhibit thrombosis in this model without impairing the hemostatic response, even at doses that are three times greater than that required to inhibit thrombus formation.
Preparation of Human Recombinant Annexin V
The polymerase chain reaction was used to amplify the cDNA from the initiator methionine to the stop codon with specific oligonucleotide primers from a human placental cDNA library (Clontec). The forward primer was 5′ACCTGAGTAGTCGCCATGGCACAGGTTCTC-3′ and the reverse primer was 5′-CCCGAATTCACGTTAGTCATCTTCTCCACAGAGCAG-3′. The amplified 1.1-kb fragment was digested with Nco I and EcoRI and ligated into the prokaryotic expression vector pTRC 99A (Pharmacia Biotechnology Inc). The ligation product was used to transform competent Escherichia coli strain JM 105 and sequenced. The sequence of the amplified segment was identical to that published by Funakoshi et al.12
Recombinant annexin V was isolated from the bacterial lysates as described by Berger et al21 with some modification. An overnight culture of E. coli JM 105 transformed with pTRC 99A–annexin V was expanded 50-fold in fresh Luria-Bertani medium containing 100 mg/L ampicillin. After 2 hours, isopropyl β-d-thiogalactopyranoside was added to a final concentration of 1 mmol/L. After 16 hours of induction, the bacteria were pelleted at 3500g for 15 minutes at 4°C. The bacterial pellet was suspended in TBS, pH 7.5, containing 1 mmol/L PMSF, 5 mmol/L EDTA, and 6 mol/L urea. The bacterial suspension was sonicated with an ultrasonic probe (model W185, Heat System-Ultrasonic, Inc) at a setting of 6 on ice for 3 minutes. The lysate was centrifuged at 10 000g for 15 minutes, and the supernatant was dialyzed twice against 50 vol TBS containing 1 mmol/L EDTA and once against 50 vol TBS.
Multilamellar liposomes were prepared according to the method of Kinsky.22 PS (Sigma Chemical Co), lyophilized bovine brain extract, cholesterol, and diacetylphosphate were dissolved in chloroform in a molar ratio of 10:15:1 and dried in a stream of nitrogen in a conical flask. TBS (5 mL) was added to the flask and agitated vigorously in a vortex mixer for 1 minute. The liposomes were washed by centrifugation at 3500g for 15 minutes, then incubated with the bacterial extract, and calcium chloride was added to a final concentration of 5 mmol/L. After 15 minutes of incubation at 37°C, the liposomes were sedimented by centrifugation at 10 000g for 10 minutes, and the bound annexin V was eluted with 10 mmol/L EDTA. The eluted annexin V was concentrated by Amicon ultrafiltration and loaded onto a Sephacryl S 200 column (5×90 cm). The annexin V was recovered in the included volume, whereas most of the liposomes were in the void volume. Fractions containing annexin V were pooled and dialyzed in 10 mmol/L Tris and 2 mmol/L EDTA, pH 8.1, loaded onto an anion exchange column (Resource Q, Pharmacia Biotechnology Inc), and eluted with a linear gradient of 0 to 200 mmol/L NaCl in the same buffer. The purified preparation showed a single band in SDS-PAGE under reducing conditions (Fig 1⇓). For rabbit experiments, the annexin V was dialyzed against HBS (10 mmol/L HEPES, 0.15 mol/L NaCl, pH 7.4) and sterile filtered with 0.2-μm filters.
Unilamellar micelles of PS were prepared as described before.23 Purified phospholipids were obtained from Avanti Polar Phospholipids, Inc. These micelles contained 20 mol% diheptanoylphosphatidylcholine and 80% dioleylphosphatidylserine (PS micelles) or dipalmitoylphosphatidylcholine (PC micelles). Equal volumes of annexin V (1 mg/mL) and phospholipid micelles (0.5 mg/mL) were mixed immediately before administration to the animals. For radiolabeled studies, annexin V was labeled with [125I]NaI by the Iodo-Gen method to a specific activity of ≈350 cpm/ng as described previously.24
Rabbit Model of Carotid Artery Thrombosis
Carotid artery thrombosis was induced with electric current as described previously.25 26 27 28 Male New Zealand White rabbits weighing 3.2 to 3.6 kg were anesthetized with ketamine (15 mg/kg) and xylazine (15 mg/kg). The right femoral artery was cannulated for recording of arterial blood pressure with a microtransducer (Electromedics). The right marginal ear vein was cannulated for administration of fluids and drugs. The right femoral vein was cannulated for drawing blood samples. The left common carotid artery was exposed by a median longitudinal incision in the neck, and a 2.5-mm Doppler flow probe was placed on the carotid artery without constricting the vessel. Proximal to the Doppler flow probe, a 4-mm-long, 23-gauge stainless steel needle electrode was inserted into the lumen of the carotid artery with minimal trauma. This electrode was positioned within the lumen parallel to the vessel wall. The bleeding was arrested by a piece of gel foam (Upjohn), and the needle was stabilized by a “surround collar” sutured around the vessel, which did not narrow the artery. After instrumentation, a 30-minute control period was allowed. During this time, blood pressure, heart rate, mean and phasic carotid artery blood flow, and ECG were continuously monitored. After this control period, thrombus formation was initiated by 150 μA of anodal current applied to the needle electrode until a 50% increase in flow velocity was recorded by the Doppler flow probe. This corresponds to an ≈50% decrease in cross-sectional area due to thrombus formation in the lumen.25 To assess the degree and the variability of carotid occlusion at this point, a separate group of 16 rabbits were similarly instrumented, and current was applied until the carotid artery flow velocity increased by 50%. At this time, current was stopped and the carotid artery pressure was fixed at arterial pressure by buffered glutaraldehyde infused through a perfusion catheter that was placed in the left common carotid artery proximal to the site of thrombus formation. After fixation, the vessel was paraffin-embedded and sectioned at 0.5-mm intervals. Twelve sections were made from each vessel, starting at the site of needle insertion and moving distally. After the sections were stained, planimetry was used to calculate the mean percentage of vessel occlusion by thrombus from the section that was most narrowed by thrombus in each vessel.
In initial studies, 4 rabbits were infused with annexin V (5.6 μg · kg−1 · min−1 IV) for 180 minutes without thrombus formation to determine whether annexin V had any inherent hemodynamic effects. After it was determined that annexin V did not any have direct effects, thrombus formation studies were done. At the time of 50% increase in flow velocity (≈50% occlusion of the vessel), the current was discontinued and rabbits were randomly allocated into 1 of 11 different treatment groups (see the Table⇓) as follows: group 1, excipient (HBS), n=6; group 2, annexin V 2.8 μg · kg−1 · min−1 for 180 minutes, n=6; group 3, annexin V 4.2 μg · kg−1 · min−1 for 180 minutes, n=6; group 4, annexin V 5.6 μg · kg−1 · min−1 for 180 minutes, n=6; group 5, a mixture of annexin V and PS micelles (5.6 μg · kg−1 · min−1 of annexin V+2.8 μg · kg−1 · min−1 PS micelle) for 180 minutes, n=7; group 6, PS micelles 2.8 μg · kg−1 · min−1 for 180 minutes, n=7; group 7, a mixture of annexin V and PC micelles (5.6 μg · kg−1 · min−1 of annexin V+2.8 μg · kg−1 · min−1 PC micelle) for 180 minutes, n=6; group 8, PC micelles 2.8 μg · kg−1 · min−1 for 180 minutes, n=4; group 9, heparin 35 U/kg as an IV bolus followed by 0.5 U · kg−1 · min−1 for 180 minutes, n=6; group 10, heparin 35 U/kg as an IV bolus followed by 0.25 U · kg−1 · min−1 for 180 minutes, n=5; and group 11, monoclonal antibody against rabbit platelet IIb/IIIa receptor (0.5 mg/kg IV; AZ-1; gift from Michael Ezekowitz, MD, Yale University Medical School, New Haven, Conn), n=7. Where indicated, animals also received 125I-fibrinogen (3 μCi in 1 mL of saline IV) at the time the current was stopped or 111In-labeled platelets as described below. Fibrinogen was purified and radioiodinated as described previously to a specific radioactivity of 130 μCi/mg.24 29
111In-Labeling of Platelets
Before instrumentation of the animals, blood was collected to label platelets with 111In-8-hydroxyquinoline (oxine) as described previously.28 29 30 Blood (21 mL) was collected into 4 mL of acid citrate dextrose and prostacyclin (10 μg/25 mL of blood), mixed, and centrifuged at 125g for 20 minutes at room temperature. The platelet-rich plasma was centrifuged at 1100g for 5 minutes and the platelet-poor plasma removed. The platelet pellet was suspended in 300 μL of platelet-poor plasma for labeling. [111In]InCl3 (Amersham Corp) was prepared in 0.3 mol/L acetate buffer, pH 5.3, to which 50 μL of oxine in ethanol (1 mg/mL) was added. After 15 minutes, the reaction mixture was extracted twice with 2 mL of methylene chloride and dried, and the residue was dissolved in 30 μL of absolute ethanol. Approximately 80% to 95% of the original radioactivity was recovered. 111In-oxine (250 to 300 μCi) was then added to platelet suspensions for 30 minutes at 37°C, the mixture was centrifuged at 1100g for 5 minutes to remove the supernatant plasma, and the platelet pellet was resuspended in 1 mL of autologous plasma. The labeling efficiency was in the range of 50% to 80%. The in vivo viability of labeled platelets was determined by calculating the percentage of administered radioactivity bound to circulating platelets at different time intervals. At 2, 5, 10, 30, 60, and 120 minutes after the administration of 111In-labeled platelets, a 1-mL blood sample was collected, platelets were isolated as described above, and radioactivity was counted in both the platelet pellet and platelet-poor plasma. The percentage recovery of the radiolabel, calculated as described before,29 was found to be 80% to 90% in the platelet pellet.
Determination of the Accumulation of 125I-Fibrinogen/Fibrin and 111In-Labeled Platelets
Deposition of 125I-fibrinogen/fibrin into carotid vessel segments was quantified as previously described.26 27 28 In animals infused with 125I-fibrinogen, either at the time of vessel occlusion or 60 minutes after the infusion was stopped, the left and the right carotid arteries were carefully removed and freed of all the surrounding fibrous tissue (carotid sheath). Then the left carotid artery was weighed and divided into three 2-cm segments: just proximal to the needle electrode insertion site into the vessel lumen; the site of thrombus formation, which corresponded to the position of the needle; and distal to the thrombus. Each segment was weighed, radioactivity was determined, and the counts were normalized according to the weight of the segment. Then the right carotid artery was weighed and, if necessary, trimmed so that its weight was the same as that of the left carotid artery, and radioactivity was determined. Accumulated radioactivity in the left carotid artery segments was expressed as a ratio of that measured from the uninstrumented right carotid artery.
Radiolabeled platelets were reinjected into the rabbits when the current was discontinued. At the end of the study (60 minutes after total occlusion or when the vessel remained patent for 60 minutes after the infusion of drugs was stopped), the carotid artery was removed and the 111In-labeled platelet accumulation ratio was determined as described for the 125I-labeled fibrinogen/fibrin accumulation ratio.
Rabbit Bleeding Assays
We evaluated the homeostatic parameters by two different methods. Template bleeding times were measured with the Simplate device (Organon Teknika). Uniform incisions 10 mm long and 1 mm deep were made on the ventral surface of the rabbit’s ear in such a way as to avoid the superficial veins. Blood was blotted with filter paper (Whatman No. 4) every 30 seconds, avoiding the incision. Bleeding time was defined as the interval between the time of incision until blood did not stain the paper.
The incisional bleeding assay was a modification of previously published methods.19 20 A surgical incision 4 cm long and 1 cm deep was made in the anterior abdominal wall, which incised the first layer of the anterior abdominal wall muscles. A preweighed gauze pad was placed in the incision for 5 minutes, and the amount of blood absorbed into the gauze was weighed. Both of the bleeding assays were performed before administration of annexin V or heparin (baseline) and then 15, 60, 180, and 300 minutes after administration of the test samples.
Tissue and Plasma Concentrations of Annexin V
To examine the tissue distribution and in vivo clearance of annexin V, 125I-labeled annexin V (1.5 μCi, specific activity ≈350 cpm/ng) was given intravenously to rabbits after the current was stopped, and 1-mL venous blood samples were collected at various times (0 to 120 minutes) (n=3). At the end of the study, the animals were killed by exsanguination under anesthesia, and the amount of radioactivity was counted in liver, spleen, right kidney, left kidney, brain, left ventricle (heart), aorta, left lung, right lung, normal proximal uninvolved segment of the left common carotid artery, the thrombosed segment of the same vessel, and the carotid artery thrombus. The radioactivity was expressed as cpm/mg wet tissue wt.
Assessment of Coagulation and Bleeding Parameters With Annexin V Infusion
To determine the potential of annexin V to induce coagulation and bleeding abnormalities, four incremental doses of annexin V (2.8, 4.2, 8.3, and 16.7 μg · kg−1 · min−1 IV) were given by continuous infusion for 60 minutes (n=4; each animal received all four doses in ascending order over a 4-hour period), and aPTT, bleeding time, and incisional bleeding were assessed at the end of each dose before the next higher dose was begun.
Data were analyzed by one-way ANOVA. In each figure, in the text, and in the Table⇑, mean values±SD are shown.
Hemodynamic Effects of Annexin V
Intravenous infusion of annexin V into the instrumented rabbit model, without the presence of thrombosis, for 180 minutes (5.6 μg · kg−1 · min−1) (n=4) did not alter the blood pressure (pretreatment, 112±11 mm Hg versus 3 hours after infusion, 106±9 mm Hg), heart rate (pretreatment, 94±12 bpm versus 3 hours after infusion, 98±14 bpm), or carotid artery blood flow (pretreatment, 18.4±4.9 mL/min versus 3 hours after infusion, 19.2±3.8 mL/min). Thus, annexin V was devoid of any inherent hemodynamic effects detectable by these methods.
Time to Coronary Occlusion
Application of current to the left carotid artery for 47±21 minutes increased the flow velocity by ≈50%, as observed previously.28 In the separate group of 16 animals, the examination of the carotid artery at this time point showed a partially occlusive thrombus that occupied 48.6±4% of the vessel lumen. The scanning electron microscopic appearance of the partially occlusive thrombus is shown in Fig 2A⇓, which demonstrates a predominantly platelet-covered surface for the formed thrombus and underlying fibrin scaffolding.25 26 When the thrombus was removed, the underlying area of the carotid artery closest to the electrode showed endothelial removal and exposure of the subendothelial matrix (Fig 2B⇓). When the current is discontinued at ≈50% occlusion, there is progressive thrombus growth with complete occlusion of blood flow within 42±12 minutes (Figs 3⇓, top, and 4A and Table⇑).
When annexin V (5.6 μg · kg−1 · min−1) was given, starting at the time when current was discontinued, blood flow remained unchanged throughout the infusion and then for an additional 60 minutes after the infusion was stopped (total duration of maintenance of blood flow, >240 minutes) (Figs 3⇑, bottom, and 4A and Table⇑). Note that the phasic carotid artery blood flow has a normal systolic and diastolic phasic blood flow profile at the end of 240 minutes (Fig 3B⇑). Because in the presence of severe narrowing (high-grade stenosis) the phasic flow pattern is gradually lost (flow is predominantly during systole), the preservation of the normal carotid flow pattern by annexin V indicates that there was no further narrowing of the vessel by ongoing thrombus formation. This observation is further supported by the fact that examination of the vessels from animals infused with annexin V revealed only small, nonocclusive thrombi containing fibrin and platelets compared with occlusive thrombi of the same apparent histological composition in control animals infused with HBS (Fig 4B⇓). Hemodynamic variables (arterial blood pressure, heart rate, and phasic and mean carotid blood flow) were not affected by the infusion of annexin V.
Annexin V inhibited thrombus formation in a dose-dependent manner, being effective in all animals receiving 5.6 μg · kg−1 · min−1 but ineffective at a dose of 2.8 μg · kg−1 · min−1 (Fig 4A⇑ and Table⇑). When animals were anticoagulated with heparin (35 U/kg as an IV bolus followed by 0.5 U · kg−1 · min−1), the left carotid artery also remained patent (5 of 6 animals), although lower levels of heparin (35 U/kg as an IV bolus followed by 0.25 U · kg−1 · min−1) did not prevent carotid occlusion (Fig 4A⇑ and Table⇑). Similarly, the administration of monoclonal antibody to rabbit IIb/IIIa receptor (0.5 mg/kg IV) prevented carotid artery thrombosis (5 of 7 animals) (Table⇑). However, there was a significant reduction in platelet counts after the administration of the antibody (before, 264±32×105/mm3 versus 4 hours after, 92±18×105/mm3, P<.001; 4 hours after without antibody, 214±46×105/mm3, P=NS). After the administration of monoclonal antibody to IIb/IIIa, there was no significant increase in bleeding for the first 1 to 2 hours. After 3 to 4 hours, however, a marked increase in bleeding at incisional sites, as well as significant blood loss by the incisional method, was observed (data not shown). Lower doses of the monoclonal antibody were ineffective in preventing carotid artery thrombosis.
We investigated the effect of PS micelles on the antithrombotic activity of annexin V. When annexin V at a dose equivalent to 5.6 μg · kg−1 · min−1 for 180 minutes was mixed with PS micelle at a dose equivalent to 2.8 μg · kg−1 · min−1 for 180 minutes and then infused, the effect of annexin V was markedly diminished (Fig 4A⇑ and Table⇑). Infusion of the same dose of PS or PC micelles (2.8 μg · kg−1 · min−1 for 180 minutes) had no significant effect compared with animals receiving HBS (Fig 4A⇑ and Table⇑).
Radiolabeled Fibrinogen Accumulation in Thrombus
Because annexin V would be expected to interfere primarily with the platelet procoagulant mechanism ultimately leading to the generation of fibrin, it was important to assess its effect on the deposition of radiolabeled fibrinogen/fibrin in the thrombosed carotid segment (Fig 5A⇓). Animals were infused with 125I-fibrinogen at the time the current was stopped, and 125I-fibrinogen/fibrin accumulation was measured. The accumulated fibrinogen/fibrin was expressed as the ratio of radioactivity in the segment from the instrumented artery to that of the radioactivity in the corresponding segment of the contralateral nonmanipulated artery. In the presence of annexin V, 125I-fibrinogen/fibrin accumulation ratios decreased in the thrombosed segments by ≈47% (from 42±4.2 [untreated] to 22.1±4.7 [treated], P<.001), whereas there was no difference in the segment proximal or distal to the lesion.
Platelet Accumulation in Thrombus
Because platelets are an important component of the thrombus in the electrically induced carotid thrombosis model and thrombin-induced platelet aggregation is thought to contribute significantly to this process, it was important to determine the effect of annexin V on platelet contribution to intravascular clotting in this setting (Fig 5B⇑). To compare platelet deposition in the thrombosed, proximal, and distal segments of the left carotid artery, a platelet accumulation index using 111In-labeled platelets was used. The counts in the left and right carotid segments were compared as for labeled fibrinogen. In animals treated with annexin V, the platelet deposition indices decreased by ≈42% (from 76.1±6.9 [untreated] to 44.0±9.7 [treated], P<.001) in the thrombosed segment of the carotid artery. There were no significant changes in the platelet deposition indexes observed in the proximal or distal carotid artery segments. These data suggest that annexin V also caused a significant decrease in platelet deposition into the thrombus. To verify the viability of labeled platelets, we determined the 111In counts in blood as a function of time after injection of labeled platelets into the rabbits. Two minutes after injection, 95% of the injected counts were present; at 5 minutes, 78%; at 10 minutes, 64%; and at 30 minutes, 62%. No significant decrease in 111In activity was noted after 30 minutes (80% to 90% of this circulating radioactivity was associated with the platelets). This strongly suggests that the damaged platelets are rapidly removed by spleen or other reticular-endothelial cells and that the remaining circulating 111In was localized in normal platelets.
Clearance and Tissue Distribution of Annexin V
Clearance of annexin V was rapid, with an α-phase of ≈5 minutes, at the end of which 10% to 12% radioactivity remains in circulation (Fig 6A⇓). The majority of the annexin V was cleared through the kidneys (data not shown). Among the vascular structures, the highest ratio of blood to tissue was found in the thrombus (Fig 6B⇓). After the thrombus was removed, the thrombosed left common carotid artery segment containing the thrombus did not show a significantly higher count compared with other vascular structures, such as right carotid artery or aorta. We also infused 125I-labeled annexin V starting at the time of initiation of current (initiation of thrombus formation) (n=3). The objective was to determine whether annexin V would bind to the carotid arterial wall if there was no thrombus to cover the area of injury. After 180 minutes of infusion, the left carotid artery was carefully removed and counted, and it did not show a significantly higher count compared with the uninstrumented right carotid artery or the aorta. These results indicate that annexin V concentrates preferentially in the thrombus and not in the vessel wall.
Alterations in Hemostatic and Bleeding Parameters
Infusion of annexin V results in the prolongation of aPTT in rabbit plasma (Fig 7⇓). The control aPTT was 26.9±4.6 seconds; with 2.8 μg · kg−1 · min−1 of annexin V infusion, it was 34.7±2.7 seconds; with 4.2 μg · kg−1 · min−1, 37.5±7.0 seconds; with 8.4 μg · kg−1 · min−1, 37.7±4.1 seconds; and with 16.8 μg · kg−1 · min−1, 36.0±3.3 seconds. Although the prolongation of aPTT was statistically significant (P<.05), the increasing doses of annexin V infusion did not cause a significant further prolongation in aPTT, unlike heparin infusion (27.4±2.2 seconds with 35 U/kg bolus+0.25 U · kg−1 · min−1 IV and 55.0±8 seconds with 35 U/kg bolus+0.5 U · kg−1 · min−1 IV) (Fig 7⇓). The effect of annexin V on hemostatic response to a cutaneous abdominal wound was assessed after a standardized incision by weighing the amount of blood absorbed by a sponge placed in the wound for 5 minutes (Fig 8A⇓). Animals infused with either saline, annexin V, or a mixture of annexin V and PC or PS micelles did not bleed excessively compared with untreated controls. Consistent with this observation, when abdominal incision bleeding was measured, no difference in accumulated radioactivity in the gauze pad was noted between annexin V–treated and control animals that had been infused with 125I-fibrinogen. In contrast, animals receiving heparin at levels required to prevent occlusive carotid thrombosis (0.5 U · kg−1 · min−1) had markedly increased bleeding. The dose of heparin used was then steadily decreased until a level was reached at which there was only a moderate increase in bleeding from the abdominal wound (0.25 U · kg−1 · min−1) (Fig 8A⇓). At this concentration of heparin, however, occlusion of the left carotid artery occurred at time intervals similar to those noted in saline-treated control animals (Table⇑). The contrast between the increased bleeding tendency of animals receiving heparin (35 U/kg bolus followed by 0.5 U · kg−1 · min−1, the amount of heparin required to prevent formation of an occlusive thrombus) and the apparent lack of bleeding in animals treated with annexin V was also qualitatively evident throughout the experimental manipulations with regard to blood loss at incisions in the chest wall and catheter insertion sites. Neither annexin V and nor heparin produced a significant change in template bleeding times (Fig 8B⇓).
After vessel wall injury, platelets rapidly adhere to the damaged vessel wall and to one another to form the primary hemostatic plug. In addition, after activation, platelets accelerate the generation of thrombin by providing an effective catalytic surface for the conversion of prothrombin by factor Xa (in the presence of factor VIIIa and Ca2+) and factor X by factor IXa (in the presence of factor VIIIa and Ca2+).1 2 3 Both factor Va and factor VIIIa were shown to bind with high affinity to a limited number of binding sites in platelets, ≈1000 sites for factor Va and ≈400 sites for factor VIIIa.31 32 33 34 The occupancy of Va receptor is essential for the subsequent binding of factor Xa, whereas binding of factor VIIIa, though not essential for factor IXa binding, enhances the binding affinity of IXa by fivefold.33 Platelet-bound factor Xa catalyzes the activation of prothrombin 300 000 times (Kcat/Km) faster than in solution, and platelet-bound factor IXa catalyzes the activation of factor X 17 000 000 times faster than in solution.1 2 31 32 33 34 The nature of the binding site has not been fully elucidated, but anionic phospholipid is required for formation of this binding site.
The importance of the exposure of anionic phospholipid for hemostasis in vivo is demonstrated in Scott syndrome, a rare bleeding disorder associated with deficiency of platelet procoagulant activity.35 The platelets in this disorder have decreased exposure of anionic phospholipid after platelet activation and have reduced factor Va-Xa and factor IXa binding sites.36 37 Whether the exposure of anionic phospholipid is also important in the pathogenesis of thrombosis associated with vessel wall injury is not known. The expression of platelet anionic phospholipids in vitro requires a stronger stimulus than that for the induction of aggregation and secretion.4 6 16 17 Such strong stimuli may be provided in vivo in arterial injury by the exposure of subendothelial components to the adhering platelets. Furthermore, the high fluid shear stress typically seen at these sites may be an additional stimulus, resulting in a strong activation. The exposure of anionic phospholipid by these stimuli may play a significant role in the growth of platelet-rich thrombi in arteries as their growth is stabilized by a fibrin scaffold.
Our results show that in a model of arterial thrombosis, anionic phospholipids play a significant role in thrombus formation. The intravenous infusion of annexin V significantly inhibited thrombus formation in a dose-dependent manner, as judged by maintenance of blood flow with normal systolic and diastolic phasic profiles and inhibition of fibrin deposition and platelet accumulation at the site of arterial injury. Most significantly, the prior mixing of annexin V with PS micelles attenuated the antithrombotic effect of annexin V, whereas prior mixing with PC micelles had no effect. These results suggest that mechanistically, the anticoagulant effect of annexin V is mediated by its interaction with PS, presumably on activated platelets. Annexin V has been shown to be selectively taken up in the thrombi in vivo.38 It is also possible that annexin V may have similar effects on endothelial or other cell types at the site of vascular injury.
One intriguing finding to emerge from these studies was the lack of hemostatic compromise observed in rabbits treated with annexin V, even at three times the concentration required to induce an antithrombotic effect. In contrast, the dose of heparin required to maintain vascular patency induced significant bleeding at extravascular sites. Although one must be careful not to overinterpret these results, if this finding is verified in other models of arterial thrombosis, it may provide important clues to the inherent differences between the mechanisms involved in hemostasis and those involved in thrombosis. The two methods used to evaluate hemostasis, the template bleeding time and the abdominal incision method, both evaluate bleeding from capillary sources. Occlusion by an arterial thrombus requires the continued recruitment and incorporation of platelets into the developing thrombus after the first wave of platelets has recognized the site of injury. These platelets are laid down on a scaffold of fibrin, which requires activation of a soluble coagulation system. In small capillaries, on the other hand, thrombus growth is much less crucial to achieve a thrombus mass sufficient to occlude the vessel. Thus, the procoagulant activity associated with platelets may not be required. Another possible explanation for the disparity between the antithrombotic effect and the antihemostatic effect of fibrin is the observation that annexin V does not inhibit the activity of the tissue factor VIIa complex as effectively as it inhibits prothrombinase activity.39 In extravascular tissues, the tissue factor–initiated clotting mechanisms may be less amenable to inhibition. Finally, the differential effect can also be due to the most obvious differences between capillaries and arteries: the ratio of endothelial surface to blood volume. In capillaries, where this ratio is large, the altered endothelial surface may suffice to bring about the arrest of bleeding, whereas in large arteries, where this ratio is small, platelet procoagulant activity may have a predominant role. In any event, our data suggest that the platelet procoagulant activity is not as important for capillary hemostasis as for arterial thrombosis. Of interest, the Ms Scott for whom the Scott syndrome was named has a normal bleeding time.35
In conclusion, our experiments show that the inhibition of procoagulant activity of platelets by annexin V can attenuate experimentally induced arterial thrombosis in the absence of excessive bleeding. These observations lend support to the notion that anionic phospholipid exposure in platelets has an important role in arterial thrombosis. Inhibition of this activity may be a novel target for antithrombotic therapy in conditions associated with arterial injury, such as acute myocardial infarction and stroke, and after angioplasty.
Selected Abbreviations and Acronyms
|aPTT||=||activated partial thromboplastin time|
This work was supported by NIH grants HL-50653, HL-40860, and HL-50100 and by a Grant-in-Aid (93-14960) from the American Heart Association. We are grateful for the technical assistance of Gerald Todd and Anhquyen Le. Special thanks go to Dr Jose Lopez for the critical review of the manuscript and suggestions.
Guest editor for this article was James H. Chesebro, MD, The Cardiovascular Institute, New York, NY.
Presented in part at the 68th Scientific Sessions of the American Heart Association, Anaheim, Calif, November 13-16, 1995, and published in abstract form (Circulation. 1995;92[suppl I]:I-805).
- Received December 10, 1996.
- Revision received April 14, 1997.
- Accepted May 19, 1997.
- Copyright © 1997 by American Heart Association
Op den Camp JAF. Lipid asymmetry in membranes. Annu Rev Med. 1979;48:47-71.
Rosing J, van Rijn JLML, Bevers EM, van Dieijen G, Comfurius P, Zwaal RFA. The role of activated platelets in prothrombin and factor X activation. Blood. 1985;65:319-326.
Rosing J, Bevers EM, Comfurius P, Hemker HC, van Dieijen G, Zwaal RFA. Impaired factor X and prothrombin activation associated with decreased anionic phospholipid exposure in platelets from a patient with a bleeding disorder. Blood. 1985;65:1557-1562.
Chang CP, Zhao J, Wiedmer T, Sims PJ. Contribution of platelet microparticle formation and granule secretion to the transmembrane migration of phosphatidylserine. J Biol Chem. 1993;268:7171-7178.
Ravanat C, Archipoff G, Beretz A, Freund G, Cazanave JP, Freyssinet JM. Use of annexin V to demonstrate the role of phosphatidylserine exposure by endothelial cells in maintenance of hemostatic balance. Biochem J. 1992;282:7-13.
Van Heerde WL, Poort S, van’t veer C, Reutelingsperger CPM, DeGroot PG. Binding of annexin V to endothelial cells: effects of annexin V binding on endothelial cell mediated thrombin formation. Biochem J. 1994;302:305-312.
Koopman G. Annexin V for flow cytometric detection of phosphatidylserine expression in B cells undergoing apoptosis. Blood. 1994;84:15-120.
Iwasaki A, Suda M, Nakao H, Nagoya T, Saino Y, Arai K, Mizoguchi T, Sato F, Yoshizaki H, Hirata M, Miyata T, Shidra Y, Muarata M, Maki M. Structure and expression of cDNA for an inhibitor of blood coagulation isolated from human placenta: a new lipocortin-like protein. J Biochem Tokyo. 1987;102:1261-1273.
Tait JF, Gibson D, Fujikawa K. Phospholipid binding properties of human placental anticoagulant protein I, a member of the lipocortin family. J Biol Chem. 1989;264:7944-7949.
Thiagarajan P, Tait JF. Binding of annexin V/placental anticoagulant protein I to platelets: evidence for phosphatidylserine exposure in the procoagulant response to activated platelets. J Biol Chem. 1990;265:17420-17423.
Thiagarajan P, Tait JF. Collagen-induced exposure of procoagulant activity in platelets. J Biol Chem. 1991;266:24302-24306.
Van Heerde WL, Sakariassen KS, Hemker HC, Sixma JJ, Reutelingsperger CPM, DeGroot PG. Annexin V inhibits the procoagulant activity of matrices of TNF-stimulated endothelium under flow conditions. Arterioscler Thromb. 1994;14:824-830.
Benedict CR, Refino CJ, Keyt BA, Pakala R, Paoni NF, Thomas GR, Bennet WF. New variant of human tissue plasminogen activator (TPA) with enhanced efficacy and lower incidence of bleeding compared with recombinant human TPA. Circulation. 1995;92:3032-3040.
Benedict CR, Mathew B, Rex KA, Cartwright J, Sordahl LA. Correlation of plasma serotonin changes with platelet aggregation in an in vivo dog model of spontaneous occlusive coronary thrombosis model. Circ Res. 1986;73:58-67.
Benedict CR, Ryan J, Wolitzky B, Ramos R, Gerlach M, Tijburg P, Stern D. Active site-blocked factor IXa prevents intravascular thrombus formation in the coronary vasculature without inhibiting extravascular coagulation in a canine thrombosis model. J Clin Invest. 1991;88:1761-1765.
Benedict CR, Ryan J, Todd J, Kuwabara K, Tijburg O, Cartwright J, Stern D. Active site-blocked Xa prevents thrombus formation in the coronary vasculature in parallel with inhibition of extravascular coagulation in canine thrombosis model. Blood. 1993;81:2059-2066.
Lakobsen E, Kierulf P. A modified β-alanine precipitation procedure to prepare fibrinogen free of antithrombin III and plasminogen. Thromb Res. 1973;3:145-159.
Kane WH, Lindhout MJ, Jackson CM, Majerus P. Factor V-dependent binding of factor Xa to human platelets. J Biol Chem. 1980;254:1170-1174.
Tracy PB, Nesheim ME, Mann KG. Factor Va-dependent binding of factor Xa to platelets. J Biol Chem. 1981;255:1170-1174.
Ahmad SS, Rawala-Sheikh R, Walsh PN. Comparative interactions of factor IX and factor IXa with human platelets. J Biol Chem. 1989;264:3244-3251.
Nesheim M, Pittman DD, Giles AR, Fass DN, Wang JH, Slonosky D, Kaufman RJ. The effect of plasma von Willebrand factor on the binding of human factor VIII to thrombin activated platelets. J Biol Chem. 1991;266:17815-17820.
Miletich JP, Kane WH, Hoffman SL, Stanford N, Majerus PW. Deficiency of factor Xa-Va binding sites on platelets of a patient with a bleeding disorder. Blood. 1979;54:1015-1022.
Ahmad SS, Rawala-Sheik R, Asby B, Walsh PN. Platelet receptor-mediated factor X activation by factor IXa: high affinity factor IXa receptors are deficient in Scott syndrome. J Clin Invest. 1989;84:824-828.
Stratton J, Dewhurst TA, Kasina S, Reno JM, Cerquira MD, Baskin DG, Tait JF. Selective uptake of radiolabeled annexin V on acute porcine left atrial thrombi. Circulation. 1995;92:3113-3121.