Pretreatment of Rabbits With Either Hirudin, Ancrod, or Warfarin Significantly Reduces the Immediate Uptake of Fibrinogen and Platelets by the Deendothelialized Aorta Wall After Balloon-Catheter Injury In Vivo
Abstract—Fibrinogen and platelets rapidly saturate the exposed subendothelium of a freshly deendothelialized aorta in vivo. As thrombin generated within the site of injury is largely responsible for fibrin(ogen) deposition, we questioned whether various anticoagulant treatments would inhibit uptake of both fibrinogen and platelets in vivo. Rabbits were anticoagulated by pretreatment with either Warfarin, Ancrod, or recombinant hirudin. Each anesthetized, anticoagulated (or saline-injected control) rabbit was injected IV with rabbit 51Cr-platelets and 125I-fibrinogen before a balloon-catheter deendothelializing (or sham) injury of the thoracic aorta. At 10 minutes after injury, the rabbit was exsanguinated and the aorta excised. Platelet adsorption by the deendothelialized aorta surface was substantially reduced in anticoagulated rabbits (controls, 2.2×105/mm2; Warfarin-treated, 1.2×105/mm2; Ancrod-treated, 5.3×104/mm2; r-hirudin–treated [5 mg/kg], 5.3×104/mm2), and a significant reduction of fibrinogen associated with the platelet layer (from 5.3 to 1 to 2 pmol/cm2) and within the underlying intima-media layer (from 16.9 to 5 to 6 pmol/cm2) was observed in the r-hirudin–and Warfarin-treated rabbits. The pattern of aorta-deposited 51Cr-platelets and 125I-fibrin in the anticoagulated rabbits corresponded well with an assessment by transmission electron microscopy of aortic tissue samples. We conclude that ≈70% of fibrinogen uptake is thrombin dependent and that ≈80% of platelet adsorption depends on codeposited fibrin(ogen) during the 10-minute interval after balloon injury. Pretreatment with an agent that interferes with either thrombin or fibrin production will inhibit the immediate interaction of fibrinogen and platelets with the freshly exposed subendothelium.
- Received September 23, 1997.
- Accepted December 19, 1997.
In 1973, Baumgartner1 showed that circulating platelets rapidly carpeted the deendothelialized surface of a rabbit aorta after an inflated balloon catheter had been drawn along the vessel lumen. The surface of the ballooned aorta was essentially saturated with platelets by 10 minutes after injury, although deposition of fibrin was not obvious. Later, Groves et al2 quantified the deposition of 51Cr-labeled platelets on the deendothelialized aortas of rabbits that at various times after balloon-catheter injury had been preperfused, first with a heparin-containing physiological solution and then with 4% glutaraldehyde to fix the tissue. These authors found that the exposed subendothelium was saturated by ≈45 000 platelets per square millimeter, which, when viewed by scanning electron microscopy, appeared as a monolayer, with most platelets spread on the vessel surface. Later, these authors3 reported that heparin, administered (500 U/kg IV) 10 minutes before inducing a balloon-catheter injury, decreased slightly the accumulation of platelets at the exposed subendothelium in vivo, and Dejana et al4 showed that heparin significantly prevented platelets from binding to the exposed subendothelium of the everted rabbit aorta segments rotated in vitro. In the latter study, it was concluded that thrombin was involved in the accumulation of platelets at the surface of the exposed subendothelium. Since then, several reports5 6 have identified thrombin as a primary agent responsible for platelets associating with the ballooned rabbit aorta surface and promoting platelet thrombus formation in vitro. These researchers have studied platelet behavior over a wide range of wall shear rates using a perfusion chamber technique. Recently, however, Gast et al7 have concluded, from perfusion chamber experiments, that thrombin is not involved in platelet thrombus formation immediately after injury but is involved later in thrombogenesis on the exposed aortic subendothelium.
The behavior of plasma 125I-fibrinogen has also been studied in rabbits after a balloon-catheter injury to the aorta.8 9 Fibrinogen saturated the deendothelialized aorta surface in vivo too rapidly (<5 minutes after injury) for a precise time curve to be measured. Subsequently, we demonstrated from studies in vitro that the quantity of fibrinogen bound by the exposed subendothelium was dependent directly on the amount of active thrombin associated with that surface.10 Indeed, an enhanced uptake of fibrinogen was maintained throughout the reendothelialization process, an event that endured for at least 20 months in the rabbit aorta.11
As fibrinogen supports the aggregation of thrombin-activated platelets, we questioned to what extent the deposition of plasma fibrinogen and the deposition of platelets at the ballooned aorta surface in vivo were interdependent. Is a single factor, namely thrombin, responsible for the deposition of both fibrinogen and platelets? In this study, we compare the uptake of fibrinogen with that of platelets over the first 10 minutes after balloon-catheter injury to the rabbit aorta in vivo using a quantitative and a morphological approach. An interval of 10 minutes between balloon injury and exsanguination was chosen because it allows sufficient time for saturation of the ballooned aorta by both fibrin(ogen) and platelets. As we are uncertain of the ratio of adsorbed fibrinogen and deposited fibrin by the aorta wall, we have referred to this material as fibrin(ogen). In addition, we have compared the adsorption of fibrin(ogen) and platelets to the deendothelialized aorta by using rabbits that had either not received an anticoagulant treatment or had been anticoagulated using one of four different procedures to discover any relationship that directly links these responses to injury. The chosen anticoagulants were recombinant hirudin, a highly specific thrombin inhibitor12 ; Ancrod, which acts by effectively removing fibrinogen from the circulatory system; and Warfarin, which selectively inhibits production of active vitamin K–dependent proteins by the liver.
Preparation and Labeling of Rabbit Fibrinogen
Fibrinogen was isolated from ACD-anticoagulated13 plasma from NZW rabbits using first a β-alanine precipitation method14 and purified further by DEAE cellulose chromatography with a Ca2+-containing buffer system.15 By these means, rabbit fibrinogen with >97% thrombin clottability was obtained.8
Fibrinogen was radiolabeled using an Iodo-gen-coated glass vial16 as follows: ≈100 μg of protein (in 300 μL 0.1 mol/L sodium phosphate, pH 7.4) was placed in a flat-bottomed glass vial previously coated with 5 μg Iodo-Gen (Pierce Chemical Co) and containing a small (5 mm) stir bar. A 10-μL vol (1 mCi) of [125I]NaI (ICN Pharmaceuticals) was added and the reaction was allowed to progress for 2 minutes at room temperature. 125I-fibrinogen was dialyzed for 18 hours against 4×250 mL 0.01 mol/L sodium phosphate–buffered 0.15 mol/L NaCl, pH 7.4, at 4°C, stored at 4°C, and used for experiment within 4 days of labeling.
Preparations of rabbit fibrinogen and 125I-fibrinogen were tested for purity before and after reduction by β-mercaptoethanol using PAGE in the presence of 0.1% SDS.17 Plasma samples from some Ancrod-treated rabbits were analyzed for fibrinogen as follows. Unreduced diluted plasmas were electrophoresed using SDS-PAGE and the gel content was electroblotted to Immobilon (Millipore Corp). The blot was developed by using first a polyspecific anti-rabbit fibrinogen antibody (raised in a laying hen and isolated from egg yolks18 ), followed by an alkaline phosphatase–linked anti-chicken IgG (Zymed Inc), and then an appropriate substrate, BCIP, and complexing agent, NBT (both obtained from BioRad). Reactive bands were scanned by using a Howteck Scanmaster “3+” (Howteck Inc) and analyzed using BioImage Whole Band Analysis software (Millipore Corp). The fibrinogen content of rabbit plasma samples was determined quantitatively using an ELISA described previously.19
Warfarin (3-[α-acetonylbenzyl]-4-hydroxycoumarin, Na salt) was purchased from Sigma Chemical Company. Ancrod (also known as “Arvin”) was obtained from Knoll Pharma Inc as a sterile solution (70 IU/mL). Recombinant desulfato-hirudin (r-hirudin; CGP39393) was a gift from Ciba Pharmaceuticals (Horsham, Sussex, UK).
Preparation and 51Cr Labeling of Rabbit Platelets
For each experiment that required 51Cr-platelets, the platelet fraction, obtained from exsanguinated blood of a donor NZW rabbit (weight range 2.3 to 3.0 kg), was radiolabeled with 51Cr before injecting into a second NZW rabbit (designated the experimental rabbit). Procedures used for isolating, washing, and radiolabeling platelets from freshly drawn, ACD-anticoagulated rabbit blood were described previously.20 21 After a fourth wash, the radiolabeled platelets were suspended in 10 mL of platelet-poor rabbit plasma. For Ancrod-treated rabbits, 51Cr-labeled platelets were suspended in 10 mL of freshly prepared, Ancrod-treated rabbit plasma (after removal of the fibrin clot). Measurement of the 51Cr-platelet count (before injection) was determined using a Coulter counter, and radioactivity content using a Minaxi-Autogamma 5000 counter (Canberra-Packard Ltd). The specific radioactivity (ie, number of platelets per 51Cr count per minute) ranged between 2000 and 8000 platelets per cpm (mean: 4900±1385 platelets per cpm; n=40).
Anticoagulation of Rabbit Circulation
All proposals for experiments using NZW rabbits were first approved by the Animal Research Ethics Board (McMaster University) and were conducted within the guidelines recommended for recovery surgery by the Canadian Council on Animal Care.22 Rabbits were injected (ear vein) with one of five anticoagulant treatments, ie, Warfarin, Ancrod, r-hirudin (low or high dose), or Ancrod followed by r-hirudin (low dose); control rabbits were injected with sterile saline. No serious bleeding complications were seen during or after any of the above treatments or during or after sample injections or balloon-catheter treatment.
Each rabbit was injected with Warfarin at 40 mg/kg, dissolved in sterile saline immediately before injection. The same dose was repeated 2 days later. Warfarin-treated rabbits were taken for balloon or sham injury on either day 4 or 5 after the start of Warfarin treatment. Blood samples (each ≈1 mL) were taken from an ear artery into 0.25 mL ACD before Warfarin was injected and again on day 4 (or 5) after the start of Warfarin treatment for measurement of the prothrombin time. The ratio of prothrombin clotting times, post-Warfarin plasma (mean, 24.6 seconds)/pre-Warfarin plasma (10.7 seconds), amounted to 2.3 (±0.3).
Rabbits were injected with Ancrod at 1 IU/kg. One hour after the first dose, each rabbit was given a second dose (1 IU/kg), and 2 hours after the first dose, a third dose (2 IU/kg) was given. The rabbit was then housed for 4 hours. At 6 hours after the start of Ancrod treatment, a blood sample was taken into a known volume of ACD and the rabbit was anesthetized in preparation for balloon or sham injury (see below). Platelet levels, taken before and after 6 hours of Ancrod treatment, were measured by CBC analysis of weighed blood samples (to determine the dilution by ACD) by the Section of Laboratory Medicine (Chedoke-McMaster Hospitals). Also, plasmas from the same blood samples were analyzed for fibrinogen content as described above.
r-Hirudin (used either as a low dose of 1.2±0.2 mg/kg, or a high dose of 5.0±0.2 mg/kg) was injected ≈3 minutes before the anesthetized rabbit was subjected to either a balloon or sham injury. Activated partial thromboplastin times were measured in a parallel series of experiments23 on plasma samples taken from r-hirudin–treated NZW rabbits (1 mg/kg) at 5 to 20 minutes after injection; activated partial thromboplastin times for control (ie, prehirudin) plasmas and posthirudin plasmas were 41.8±2.9 and 84.0±19.5 seconds, respectively.
Rabbits were treated with Ancrod as described above. Each rabbit was anesthetized at 6 hours after the start of Ancrod treatment. r-Hirudin (≈1 mg/kg) was injected ≈3 minutes before the rabbit was subjected to a balloon-catheter injury.
Rabbits, after receiving a dose of anesthetic, were injected with 1 mL saline before balloon-catheter or sham injury.
Uptake of Platelets and Fibrinogen by the Aorta Wall Immediately After a Balloon-Deendothelializing Injury In Vivo
A blood sample was taken from the ear artery (≈1 mL) of each NZW rabbit (range: 2.1 to 3.3 kg; mean: 2.6±0.3 kg) into 0.25 mL ACD, weighed, and measured for platelet content.
Rabbits were anesthetized by intravenous injection of sodium pentobarbital (up to 35 mg/kg). After shaving the front of the neck and inner thigh areas, a carotid artery was cannulated (PE 200 tubing) and a femoral artery isolated. At ≈15 minutes after inducing anesthesia, each r-hirudin–or Warfarin-treated rabbit was injected (through the carotid cannula) with ≈10 mL of 51Cr-labeled platelets (containing 4 to 10×106 cpm; mean: 6.0±1.5×106 cpm). The line was then immediately flushed with 1 mL of sterile saline before injecting 1 mL of 125I-fibrinogen (10 to 20×106 cpm; 14.9±3.3×106 cpm; 18.7±9.4 μg) in sterile saline. Again, the line was flushed with 1 mL of sterile saline. All Ancrod-treated rabbits were injected only with 51Cr-platelets followed by 1 mL of sterile saline.
In all recipient rabbits, the proportion of 51Cr-platelets added to the circulation was calculated from the specific radioactivity of the preparation of 51Cr-platelets, using the known quantity of 51Cr-platelets injected and the blood platelet count and weight (and hence the blood volume) of the recipient rabbit; the proportion of donor 51Cr-platelets after injection amounted to 33.5±1.4% (n=40) of the total platelet population.
At 5 minutes after injection, a blood sample (carotid artery; 1 mL) was taken into 0.25 mL ACD and the rabbit was given either a deendothelializing injury or a sham injury to the thoracic aorta. For the balloon treatment, a Fogarty balloon catheter (Baxter Healthcare Corp; type 12-040-4F) was inserted into an exposed femoral artery and maneuvered into the aorta. The thoracic aorta was subjected to two passes with an expanded balloon volume (containing 1 mL of sterile saline) and the artery ligated as described previously.9 For the sham-injury treatment, a femoral artery was ligated but not penetrated. A timer was started as the femoral artery was ligated. At 10 minutes after femoral ligation, a second blood sample was taken before the rabbit was rapidly exsanguinated through the carotid cannula. The chest was opened and the thoracic aorta isolated, flushed with 10 mL of BSA-minimum essential medium (with Earle’s salts and L-glutamine), and then excised.
Adventitial fat was removed and the thoracic aorta divided into eight 1-cm-long segments (segment 1, arch end; segment 8, diaphragm end). Segment 4 from each aorta was placed into 4% (wt/vol) paraformaldehyde for TEM studies. The remaining segments were processed as follows: After measuring the surface area of each segment using a digital planimeter (Lasico), the luminal surface, which for sham-injured aortas included the endothelium and associated basement membrane24 and for deendothelialized aortas included the platelet layer,8 was stripped cleanly from the aorta using cellulose acetate paper as a Häutchen preparation.25 The underlying subendothelium, referred to as the intima-media layer, was then stripped from the vessel wall using Bergh forceps.26
The vessel layers of seven individual segments from each aorta were each placed in a plastic-capped vial (Biovial, Beckman) and measured for radioactivity content. The 125I counts were corrected for crossover of 51Cr into the 125I channel. The corrected 125I radioactivity per unit area, ie, cpm/cm2 of aorta layer, was compared with the respective radioactivity content of 1 mL of blood at exsanguination. The quantity of bound fibrinogen per square centimeter was calculated assuming a hematocrit of 42% and a blood concentration of 3.58 μmol/L fibrinogen19 for a healthy NZW rabbit. By using the known proportion of 51Cr-platelets in the circulation for each 51Cr-platelet preparation and their specific radioactivity (ie, number of platelets per cpm), the mean number of adsorbed platelets per square millimeter could be calculated from the 51Cr content of the cellulose acetate layer obtained from each of the seven 1-cm-length segments of each thoracic aorta.
Uptake of 125I-Fibrinogen by the Aorta Wall During Exsanguination
Rabbits that had not received an anticoagulant treatment were anesthetized and a carotid cannula was inserted as described above. A balloon or sham injury was performed, and the clock was started at the moment of femoral ligation as the balloon (or sham) injury was completed. An intravenous dose of 125I-fibrinogen (≈19 μg in 1 mL of sterile saline) was injected at either 7 minutes, 8 minutes, or 9 minutes after the clock was started. An intra-arterial blood sample (≈2 mL into 0.5 mL ACD) was taken at 10 minutes, and the rabbit was exsanguinated immediately through the carotid cannula. The aorta was isolated, flushed with BSA-minimum essential medium Eagle, and excised as described above. The surface area and radioactivity content of the Häutchen preparation of the platelet layer (of balloon-injured aortas) or endothelium (of the sham-injured aortas) and of the underlying intima-media were measured, and the fibrinogen content per square centimeter of aorta wall was calculated relative to the radioactivity content of blood at exsanguination as described above.
Measurement of Prothrombin Time
After collection, blood samples were weighed (to determine the dilution due to ACD) and centrifuged (10 000 rpm, 2 minutes) to separate the plasma fraction. An automated plasma coagulation procedure described by Organon Technica, the supplier of Simplastin Excel (a form of thromboplastin), was used with a coagulation analyzer (Behring) to determine the prothrombin time for the ACD-plasma samples.
Morphology of the Platelet Layer
One-millimeter-thick, full-circumference rings of paraformaldehyde-fixed aorta wall were prepared for TEM as described previously.2 3 The samples were postfixed in buffered 1% osmium tetroxide for 1 hour, stained in 2% aqueous uranyl acetate for 1 hour, dehydrated through graded ethanol, and embedded in Spurr’s resin. One-micrometer-thick sections, stained with toluidine blue, were examined by light microscopy, and areas showing maximal platelet accumulation with the luminal surface were selected for examination by TEM. Thin sections (60 to 90 nm), stained with lead citrate, were viewed on a Philips TEM 301 transmission electron microscope (Eindhoven, the Netherlands).
Properties of 125I-Fibrinogen
Rabbit 125I-fibrinogen, as shown by SDS-PAGE (Fig 1⇓), consisted essentially of one single band, which barely entered the 7.5% polyacrylamide gel. After reduction by β-mercaptoethanol, the three component chains of fibrinogen, Aα, Bβ, and γ, appeared wholly separated as bands of Mr estimated at 65 kD, 58 kD, and 56 kD, respectively. The Coomassie-stained bands matched well with those of the autoradiograph (compare lane 1 with 2 and lane 3 with 4 in Fig 1⇓).
Effect of Ancrod on the Clearance of 125I-Fibrinogen From the Rabbit Circulation
A three-dose regimen was used in the study of the effect of Ancrod pretreatment of rabbits used for studying the behavior of 51Cr-platelets after balloon-catheter injury (see below). Blood samples were taken before and at 6 hours after the start of Ancrod treatment. ELISAs of the fibrinogen content of the plasmas were performed in the presence of appropriate quantities of purified rabbit fibrinogen on the same 96-well plates, and corrections were made for ACD content (Fig 2a⇓). From the ELISAs, plasma fibrinogen levels in the Ancrod-treated rabbits was ≈37% of that in normal rabbits. However, as shown by the immunoblot in Fig 2b⇓, the polyspecific anti-fibrinogen antibody could not distinguish between intact fibrinogen and fibrinogen-related degradation products. Densitometric measurements of similar immunoblots of plasmas taken from five rabbits before and at 6 hours after Ancrod treatment showed that the content of intact fibrinogen in plasma before Ancrod treatment ranged from 91.0% to 99.8% (mean: 95.0±1.7%) of all stained bands, whereas “intact” fibrinogen in plasma at 6 hours after Ancrod treatment amounted to 38.0±0.6%. Thus, the plasma level of intact fibrinogen in plasma at 6 hours after starting the three-dose regimen of Ancrod treatment is calculated to be ≈14% of normal plasma levels (ie, 38% of 37% of normal plasma level), equal to ≈1 μmol/L.
Measurements of platelet levels indicated that at 6 hours after the start of Ancrod treatment, the platelet level (3.3±0.2×1011/L) was insignificantly different from that before treatment (3.7±0.3×1011/L).
Uptake of 51Cr-Platelets and 125I-Fibrinogen by the Surface of the Aorta Wall After a Deendothelializing Injury to Saline-Treated Rabbits
At 10 minutes after a balloon-catheter injury to the aorta of a saline-treated rabbit, the aortic surface contained an abundance of adsorbed platelets and fibrinogen, as determined by the 51Cr- and 125I radioactivity contents, relative to the aortic surface of the sham-operated rabbit. The platelets adsorbed by the deendothelialized aortas of eight rabbits amounted to a mean of 222 800/mm2 (Table 1⇑). By contrast, platelets adsorbed by the aorta surface in the sham-injured rabbit (ie, the endothelium) amounted to 768/mm2. The fibrin(ogen) content of the platelet layer was calculated to be 5.3 pmol/cm2 in aortas from balloon-injured rabbits, and 0.7 pmol/cm2 with the aortic endothelium of sham-injured rabbits (Table 2⇑).
After the platelet layer was removed, the underlying subendothelium (isolated as the intima-media layer) was found to contain significant quantities of 51Cr and 125I radioactivities. As the platelet layer had been cleanly removed by cellulose acetate paper8 and the specific radioactivity of the 51Cr-labeled platelet preparation was known, the 51Cr data were calculated as “platelet equivalents” per square millimeter, ie, the number of platelets equivalent to the measured 51Cr radioactivity per square millimeter. The intima-media layer of the eight balloon-injured aortas contained ≈237 000 platelet equivalents per square millimeter in contrast to 616/mm2 determined for that of the sham-injured aortas (Table 1⇑). The mean fibrin(ogen) content of the balloon-injured aortas amounted to 16.9 pmol/cm2, in contrast to 0.4 pmol/cm2 of sham-injured aorta (Table 2⇑).
Effect of Pretreatment with r-Hirudin on Uptake of 51Cr-Platelets and 125I-Fibrinogen by the Aorta Wall After a Deendothelializing Injury
Deendothelialized aortas from r-hirudin–treated rabbits contained significantly fewer adsorbed platelets compared to those from saline-treated rabbits, decreasing to ≈121 000/mm2 (ie, 55% of the value measured for saline-injected, ballooned aortas) in rabbits pretreated with low dose, and to 53 000/mm2 (24%) in rabbits pretreated with high dose of r-hirudin (Table 1⇑). Fibrin(ogen) deposition in the platelet layer was similarly decreased to 2.0 pmol/cm2 (38% of the saline-injected ballooned aorta) and 0.9 pmol/cm2 (17%) in the low- and high-dose r-hirudin–treated rabbits, respectively (Table 2⇑).
Compared with aortas from saline-treated rabbits, the platelet equivalents per square millimeter of intima-media from ballooned aortas of low- and high-dose rabbits was decreased significantly to ≈41 000 (17% of the saline-injected, ballooned aorta) and 16 000 (7%), respectively (Table 1⇑). Fibrin(ogen) associated with the intima-media layer (6.7 and 6.4 pmol/cm2) of low- and high-dose rabbits was decreased significantly to 40% and 38%, respectively, of the values measured for the intima-media from saline-treated rabbits (Table 2⇑). Compared with the low dose, the use of a high dose of r-hirudin did not significantly decrease further the uptake of fibrinogen by the deendothelialized aorta.
Effect of Pretreatment With Ancrod on the Uptake of 51Cr-Platelets by the Aorta Wall After a Deendothelializing Injury
The plasma levels of fibrinogen were reduced by Ancrod treatment to ≈14% of normal plasma fibrinogen at the time of balloon-catheter treatment. In parallel, the adsorption of platelets by the exposed subendothelium was decreased significantly to 24% of the value measured for the saline-treated, ballooned rabbits, ie, from 222 800/mm2 (saline-treated) to 52 700/mm2 (Table 1⇑). After removal of the platelet layer, the platelet equivalents associated with the underlying intima-media had decreased to ≈15 400/mm2, ≈7% of the value for saline-injected, ballooned aortas.
The effect of injecting r-hirudin into Ancrod-treated rabbits decreased the deposition of platelets on the ballooned aorta subendothelium to ≈42 000/mm2, which is 19% of the mean value determined for saline-injected ballooned aortas (Table 1⇑). This decrease was similar to that observed with either Ancrod alone or high-dose r-hirudin alone. However, the platelet equivalents associated with the intima-media layer decreased to ≈8800/mm2 (3.7%) in the Ancrod/r-hirudin–treated rabbits, a value substantially less than that in either Ancrod or low-dose r-hirudin pretreatments.
Effect of Pretreatment With Warfarin on the Uptake of 51Cr-Platelets and 125I-Fibrinogen by the Aorta Wall After a Deendothelializing Injury
In four Warfarin-treated balloon-injured rabbits, platelet deposition on the surface of the deendothelialized aorta was decreased to 121 600/mm2 (ie, 55%) and fibrin(ogen) to 1.4 pmol/cm2 (26%) compared with aortas from saline-treated, balloon-injured rabbits (Tables 1⇑ and 2⇑).
From the 51Cr radioactivity, the platelet equivalents associated with the intima-media were decreased significantly in the Warfarin-pretreated rabbit to ≈30 000/mm2 (ie, 12.5% of the saline-treated; Table 1⇑). Fibrin(ogen) associated with the intima-media was decreased to 5.5 pmol/cm2 (30%; Table 2⇑).
Uptake of 125I-Fibrinogen by the Deendothelialized Aorta Wall During Exsanguination
An intravenous dose of 125I-fibrinogen was injected into nonanticoagulated rabbits at either 7, 8, or 9 minutes after inflicting a balloon (or sham) injury to quantify the fibrin(ogen) associated with the aorta surface (cellulose acetate layer) and underlying intima-media at 3 minutes, 2 minutes, and 1 minute before exsanguination (Fig 3⇓). These data allowed extrapolation to the time of the start of exsanguination so that the quantity of fibrin(ogen) deposited on and within the aorta wall during the exsanguination process could be assessed. Compared with the fibrin(ogen) content at 10 minutes after balloon injury (ie, 5.3 pmol/cm2 of platelet layer; 16.9 pmol/cm2 of intima-media; Table 2⇑), the accumulation of fibrin(ogen) by the platelet layer (amounting to <5% of the total fibrinogen) and by the intima-media (<10%) of the aorta wall during exsanguination was considered to be negligible.
From the slope of each curve (Fig 3⇑) and the mean level of fibrin(ogen) saturation from the data in Table 2⇑, the rates of turnover of fibrin(ogen) at the platelet layer and the intima-media were estimated to be ≈6 minutes and ≈10 minutes, respectively, during the 10-minute interval after injury.
Assessment of the Luminal Surface of Aortas using TEM
The quantitative measurements of fibrin(ogen) and platelet adsorption by the sham- and balloon-injured aorta surface were supported by TEM inspection of the fourth segment from each aorta (Fig 4a⇓ through 4d). The luminal surfaces of aortas from sham-injured rabbits were essentially devoid of platelets associated with the endothelium (data not shown).
Platelets accumulated in deposits ranging in depth from <5 to >20 platelets on the deendothelialized aortas from saline-treated rabbits (Fig 4a⇑). The platelets adsorbed directly to the subendothelium were invariably degranulated and spread. As the depth of platelets increased, platelets were more rounded and less shape changed as their distance from the aorta wall increased. Polymerized fibrin with a periodicity of ≈25 nm (see inset, Fig 4a⇑) was seen filling the interplatelet spaces. By contrast, platelets adsorbed to ballooned aortas from hirudin-treated (5 mg/kg) rabbits were comparatively sparse, the deendothelialized aortic surface covered generally with only a monolayer of shape-changed, vacuolated, and spread platelets adhering closely to the exposed subendothelium (Fig 4b⇑). The luminal surfaces of the ballooned aortas of Ancrod-treated rabbits (Fig 4c⇑), Warfarin-treated rabbits (Fig 4d⇑), and Ancrod/r-hirudin–treated rabbits (data not shown) after balloon-catheter injury were similar in appearance to those from r-hirudin–treated rabbits with respect to the platelet layer.
The present data support evidence that thrombin, generated spontaneously at or within the exposed aortic subendothelium after injury, is involved directly with the rapid deposition of fibrin(ogen) and platelets on the exposed surface of the subendothelium in vivo (Fig 4a⇑); evidence of polymerized fibrin with a periodicity of 25 nm27 was frequently observed between aggregated platelets (Fig 4a⇑, inset). We also conclude that pretreatment with various anticoagulants that either inhibited thrombin production or activity or fibrin(ogen) deposition reduced substantially the accumulation of aggregated platelets and associated fibrin(ogen) at the surface of the exposed subendothelium. Thus, pretreatment of the rabbit with r-hirudin led to a parallel decrease in the quantities of adsorbed fibrinogen and platelets in the platelet layer (Tables 1⇑ and 2⇑), the higher dose of r-hirudin (5 mg/kg) reducing the uptake of fibrinogen by ≈70% and platelets by 80%. Kelly et al28 reported a similar dose-dependent effect of r-hirudin on platelet deposition on endarterectomized aortic tissue and on platelet and fibrin deposition on collagen-coated tubing when used as arteriovenous shunts in baboons. We conclude that the deposition of fibrin(ogen) and the deposition of platelets after endothelial injury are not independent events.
The fact that Warfarin treatment also caused a decrease in fibrinogen and platelet uptake after injury further strengthens the concept that thrombin plays a pivotal role in recruiting fibrinogen and platelets to the wound site. Warfarin acts as a vitamin K antagonist in vivo, inhibiting the synthesis of biologically active, vitamin K–dependent coagulation factors, such as factors VII, IX, X, and prothrombin.29 From ELISA measurements of plasma samples taken before treatment and on day 4 or 5 since the start of treatment, Warfarin induced a sharp decline in rabbit plasma prothrombin concentration from 1.8 μmol/L to ≈0.2 to 0.3 μmol/L (A. Parshad and M.W.C. Hatton, unpublished data, 1996). From this information and our knowledge of the fractional distribution of prothrombin in the rabbit,9 we conclude that Warfarin causes the plasma prothrombin concentration, and consequently the resident concentration of prothrombin within the extracellular space of the aortic intima-media, to be decreased by up to 10-fold. This depleted reserve of prothrombin (and, presumably, of factors VII, IX, and X) within the intima-media may be insufficient to provide an adequate concentration of thrombin during the crucial first few minutes after injury, and the compromised plasma levels of these vitamin K–dependent factors would be inadequate to provide a sufficient flux to meet the prothrombinase requirement within the injured intima-media.
The similar effect of hirudin and Warfarin on the behaviors of fibrinogen and platelets toward the aorta wall after balloon injury suggests strongly that most of the platelet and fibrin(ogen) deposition is linked directly with thrombin production at the wound site. Why, then, should pretreatment with Ancrod, a snake-venom protease, which selectively depletes plasma fibrinogen30 and which does not affect the plasma levels of prothrombin or other vitamin K–dependent factors in vivo or interfere with thrombin activation,31 decrease platelet accumulation at the deendothelialized aorta wall (Table 1⇑, Fig 4c⇑) to a monolayer similar to the effect of r-hirudin? Ancrod cleaves the Aα chain of mammalian fibrinogens, releasing only fibrinopeptide A to yield des-A-fibrinogen (or Ancrod-fibrin monomer).32 In vitro, des-A-fibrinogen polymerizes poorly compared with thrombin-derived fibrin monomer,33 34 and, for this reason, Ancrod-fibrin clots are believed to be relatively easily dispersed and cleared from the circulation. As a result, after 6 hours of Ancrod treatment (Fig 2⇑), the intact fibrinogen level measured ≈14% of the normal plasma level, a value similar to that previously reported in rabbits after 6 hours of Ancrod treatment.35 We also observed that the platelet level had not changed significantly after 6 hours of Ancrod treatment.
The Ancrod results suggest that as much as 80% of the platelet deposit on the subendothelium is caused directly by codepositing fibrin(ogen) and therefore only indirectly by thrombin. If all platelet adsorption, including the monolayer, was attributed to thrombin, then further pretreatment of the Ancrod rabbit with r-hirudin would decrease platelet binding further. This result was not observed (see Table 1⇑). We conclude that the initial monolayer of platelets is attracted to the exposed subendothelium independently of thrombin production. Subendothelium-associated collagen and other adhesive proteins (eg, fibronectin, laminin, or von Willebrand factor) have been implicated as binding sites for the monolayer of adherent platelets immediately after injury.36 Furthermore, in the saline-treated rabbit, the mass of platelets that aggregate with platelets of the initial monolayer are associated with obvious fibrin (Fig 4a⇑), and therefore their presence is clearly dependent on thrombin (and hence fibrin) production at the site of injury. Presumably, the production rate of thrombin after injury corresponds directly with the extent of the injury and, in the presence of normal plasma levels of fibrinogen, stimulates an appropriate deposition of fibrin(ogen) and platelets. Ancrod pretreatment causes a significant decrease in the availability of thrombin-derived fibrin for significant platelet aggregation to take place on the monolayer. We note that Chang and Huang35 reported that the depletion of fibrinogen caused by Ancrod leads to a decreased platelet aggregation activity in rabbits. Possibly, subendothelium-adsorbed des-A-fibrinogen or other Ancrod-derived fibrinogen degradation products may directly inhibit platelet aggregation, or the polymerization of thrombin-derived fibrin, which promotes platelet aggregation, with the platelet monolayer. A similar suggestion to explain the effect of Ancrod-induced hypofibrinogenemia on platelet plug formation has been made previously by Daniel et al.37
Previous measurements of the deposition of platelets after a deendothelializing injury to the rabbit thoracic aorta have claimed that 42 000 to 45 000 platelets per square millimeter of subendothelium are equivalent to a saturated monolayer of platelets.2 38 In their perfusion-fixation procedure, Groves et al2 perfused heparin-injected balloon-injured rabbits, first with Locke-Ringer physiological fluid (100 mm Hg) until the perfusate was clear of red cells, followed by a 4% (wt/vol) glutaraldehyde solution for 1 to 2 minutes; the aorta was excised later. In the present study, we have used neither heparin nor a perfusion-fixation technique; rather, the rabbit was rapidly exsanguinated through a carotid cannula, the aorta exposed, briefly rinsed, and then excised and processed. We accept that our procedure to deendothelialize an aorta may conceivably cause greater medial damage and therefore greater thrombin generation than the technique of Groves et al.2 This difference in technique probably accounts, in part, for the greater deposition of platelets on the deendothelialized aorta in this study and in another recent study39 of platelet deposition on the exposed subendothelium surface. Arguably, the process of exsanguination and the accompanying decline in blood pressure, followed by a partial collapse of the thoracic aorta may have encouraged further platelet aggregation at the surface of the artery wall. However, as shown by experiment (Fig 3⇑), only a minor proportion (<5%) of the total fibrin(ogen) in the platelet layer was deposited during exsanguination. From this result, it follows that only a minor fraction of the platelet layer would be deposited during the exsanguination step.
After inducing a deendothelializing injury to the rabbit iliac artery, Goldberg et al40 discovered that a platelet α-granule–specific protein, platelet factor 4, was located in increasing quantity within the intima-media during the first 30 minutes, but was markedly diminished at 4 hours after injury. The authors concluded that platelets in contact with the denuded vessel release the contents of their α-granules into the subendothelial space. Later, in a study of human platelet interaction with 35 S-labeled extracellular matrix proteoglycan, Yahalom et al41 discovered that heparitinase, a lysosomal enzyme,42 was released from adherent platelets and degraded heparan sulfate chains within the matrix. From Table 1⇑, it appears that 51Cr-labeled platelets of the platelet layer release a large proportion of their 51Cr contents into the underlying intima-media layer, possibly as microparticles resulting from the interaction between platelets, thrombin, and exposed collagen.43 We rule out the possibility that the 51Cr in the intima-media is due to free 51Cr, ie, not contained by platelets, as the quantity of 51Cr associated with the intima-media from aortas of sham-injured rabbits or from anticoagulated, balloon-injured rabbits was much less than that of the intima-media of saline-injected, balloon-injured rabbits. Indeed, >90% of 51Cr deposited into the subendothelial space was inhibited by pretreatment with either r-hirudin, Ancrod, or Warfarin and >96% by the administration of r-hirudin to the Ancrod-treated rabbit. We also reject the possibility that a high proportion of residual platelets remain adsorbed to the subendothelium after Häutchen preparation, as a previous study8 using scanning electron microscopy has indicated the platelet layer to be entirely removed by cellulose acetate paper.
We conclude, similar to Gast et al,7 that direct adherence of a monolayer of platelets to the exposed subendothelium takes place rapidly after injury and independently of thrombin generation. However, as thrombin is generated during the first few minutes after injury, fibrinogen is drawn to the site of injury, interacting with subendothelium-bound thrombin and forming fibrin. Subendothelium-associated thrombin and fibrin then stimulate more platelets to interact and aggregate with the platelets that comprise the monolayer. The presence of anticoagulant conditions inhibits in part or completely the secondary stage of platelet deposition, which involves thrombin and fibrin.
Selected Abbreviations and Acronyms
|ELISA||=||enzyme-linked immunosorbent assay|
|NZW||=||New Zealand White|
|PAGE||=||polyacrylamide gel electrophoresis|
|TEM||=||transmission electron microscopy|
This study was supported by a grant from the Heart and Stroke Foundation of Ontario. We thank Andrew Parshad for his technical help with the Warfarin experiments and Rena Cornelius for help with the densitometry. We also thank Dr Hallie Groves (Department of Biomedical Sciences, McMaster University) for providing helpful comments and criticisms of the manuscript and Dr Graham Pay (Ciba Pharmaceuticals, Horsham, UK) for generously supplying recombinant desulfato-hirudin.
Inauen W, Baumgartner HR, Bombeli T, Haeberli A, Straub PW. Dose- and shear rate–dependent effects of heparin on thrombogenesis induced by rabbit aorta subendothelium exposed to flowing human blood. Arteriosclerosis. 1990;10:607–615.
Turitto VT, Weiss HJ, Baumgartner HR. Platelet interaction with rabbit subendothelium in von Willebrand’s disease: altered thrombus formation distinct from defective platelet adhesion. J Clin Invest. 1984;74:1730–1741.
Gast A, Tschopp TB, Baumgartner HR. Thrombin plays a key role in late platelet thrombus growth and/or stability: effect of a specific thrombin inhibitor on thrombogenesis induced by aortic subendothelium exposed to flowing rabbit blood. Arterioscler Thromb. 1994;14:1466–1474.
Hatton MWC, Southward SMR, Ross-Ouellet B, DeReske M, Blajchman MA, Richardson M. An increased uptake of prothrombin, antithrombin, and fibrinogen by the rabbit balloon-deendothelialized aorta surface in vivo is maintained until reendothelialization is complete. Arterioscler Thromb Vasc Biol. 1996;16:1147–1155.
Aster RH, Jandl JH. Platelet sequestration in man, I: methods. J Clin Invest. 1964;43:843–854.
Straughn W, Wagner RH. A simple method for preparing fibrinogen. Thromb Diath Haemorrh. 1967;16:198–206.
Lawrie JS, Ross J, Kemp GC. Purification of fibrinogen and the separation of its degradation products in the presence of calcium ions. Biochem Soc Trans. 1979;7:693–694.
Kinlough-Rathbone RL, Packham MA, Mustard JF. Platelet aggregation. In: Harker LA, Zimmerman E, eds. Methods of Hematology: Measurements of Platelet Function. Edinburgh, Scot: Churchill-Livingstone; 1983:64–91.
Olfert ED, Cross BM, McWilliam AA, eds. Guide to the Care and Use of Experimental Animals, I. 2nd ed. Ottawa, Canada: Canadian Council on Animal Care; 1993.
Hatton MWC, Moar SL, Richardson M. On the interaction of rabbit antithrombin III with the luminal surface of the normal and de-endothelialized rabbit thoracic aorta in vitro. Blood. 1986;67:878–886.
Haust MD, Wyllie JC, More RH. Electron microscopy of fibrin in human atherosclerotic lesions: immunohistochemical and morphological identification. Exp Mol Pathol. 1965;4:205–216.
Kelly AB, Marzec UM, Krupski W, Bass A, Cadroy Y, Hanson SR, Harker LA. Hirudin interruption of heparin-resistant arterial thrombus formation in baboons. Blood. 1991;77:1006–1012.
Stirling Y. Warfarin-induced changes in procoagulant and anticoagulant proteins. Blood Coagul Fibrinolysis. 1985;6:361–373.
Reid HA. Defibrination by Agkistrodon rhodostoma venom. In: Russell FE, Saunders PR, eds. Animal Toxins. New York, NY: Pergamon Press; 1967:323–335.
Ewart MR, Hatton MWC, Basford JM, Dodgson KS. The proteolytic action of Arvin on human fibrinogen. Biochem J. 1970;118:603–609.
Spellman GG, Macoviak JA, Gralnick HR. Comparison of polymerization of Ancrod and thrombin fibrin monomers. Blood. 1977;50:619–623.
Shen L, Hermans J, McDonagh J, McDonagh RP. Role of fibrinopeptide B release: comparison of fibrins produced by thrombin and Ancrod. Am J Physiol. 1977;232:H629–H633.
Mruk JS, Chesebro JH, Webster MWI. Platelet aggregation and interaction with the coagulation system: implications for antithrombotic therapy in arterial thrombus. Coron Artery Dis. 1990;1:149–158.
Adelman B, Stemerman MB, Handin RI. Interaction of platelets and fibrin with injured rabbit aortic neointima: effect of prostaglandin I2 and heparin. Arteriosclerosis. 1983;3:141–148.
Goldberg ID, Stemerman MB, Handin RI. Vascular permeation of platelet factor 4 after endothelial injury. Science. 1980;209:611–612.
Yahalom J, Eldor A, Fuks Z, Vlodavsky I. Degradation of sulfated proteoglycans in the subendothelial extracellular matrix by human platelet heparitinase. J Clin Invest. 1984;74:1842–1849.
Oosta GM, Favreau LV, Beeler DL, Rosenberg RD. Purification and properties of human platelet heparitinase. J Biol Chem. 1982;257:11249–11255.
Sims PJ, Wiedmer T, Esmon CT, Weiss HJ, Shattil SJ. Assembly of the platelet prothrombinase complex is linked to vesiculation of the platelet plasma membrane. J Biol Chem. 1989;264:17049–17057.