An Increased Uptake of Prothrombin, Antithrombin, and Fibrinogen by the Rabbit Balloon-Deendothelialized Aorta Surface In Vivo Is Maintained Until Reendothelialization Is Complete
The ability of the rabbit aorta intima-media (IM) layer to adsorb certain plasma proteins was measured for up to 20 months after a deendothelializing injury in vivo. Purified radioiodinated rabbit fibrinogen, antithrombin, or prothrombin was injected intravenously into either uninjured or sham-injured rabbits (controls) or rabbits at various times (5 minutes to 20 months) after a balloon-catheter injury to the aorta. After a 10-minute circulation time, a blood sample was taken, and the rabbit was exsanguinated rapidly (via a carotid cannula) and the aorta excised. Uptake of each radiolabeled protein was measured as bound radioactivity per square centimeter of platelet- or endothelium-free aorta IM and was compared with the radioactivity (ergo concentration) in blood at exsanguination. Fibrinogen adsorption by the IM was maximal at 5 minutes after injury (10.9±2.3 pmol/cm2 IM) and declined slowly to 4 to 6 pmol/cm2 at 12 months (controls: 0.8±0.1 pmol/cm2). Uptake of prothrombin (3.7±0.5 pmol/cm2 at 5 minutes) decreased to ≈2 pmol/cm2 at 12 months (controls: 0.3 pmol/cm2). Antithrombin adsorption by the IM (3.3±0.4 pmol/cm2 at 5 minutes) paralleled that of prothrombin over 12 months (controls: 0.3 to 0.4 pmol/cm2), the molar ratio ranging from 0.8 to 1.2. At 20 months, the ballooned aorta had a significantly thickened intima and was ≈90% reendothelialized. Injection of horseradish peroxidase (HRP) into rabbits at 1 or 12 months after balloon injury showed clearly that HRP activity was present throughout the entire depth of the deendothelialized, but not the reendothelialized, thickened intima. These results may indicate that an elevated turnover of hemostatic proteins continues within the deendothelialized intima after injury, conceivably until reendothelialization is complete.
- aortic intima
- balloon-catheter deendothelializing injury
Reprint requests to Dr Mark W.C. Hatton, Department of Pathology (HSC-4N67), McMaster University Health Sciences Centre, 1200 Main St W, Hamilton, ON L8N 3Z5. E-mail firstname.lastname@example.org.
- Received November 1, 1995.
- Revision received May 30, 1996.
Percutaneous transluminal coronary angioplasty by a balloon-tipped catheter is being used increasingly as a means to expand the lumen of obstructed coronary and other arteries.1 In addition, balloon catheters are used in arterial embolectomy.2 However, arteries frequently become occluded immediately after balloon-catheter treatment, due to vasoconstriction and/or an increased deposition of platelets and fibrin amid thrombi.3 This occurrence may be explained by the presence of active thrombin within the extracellular matrix of the arterial wall at the site of angioplasty,4 which may contribute to the release of endothelin, a potent vasoconstrictor.5 A diminished local production of nitric oxide may also be implicated, as an infusion of nitroglycerin can counter vasoconstriction at this time.6
In ≈30% of patients who undergo coronary angioplasty, late restenosis occurs within 6 months of treatment. This is due partly to smooth muscle cell proliferation, which results in progressive hyperplasia within the tunica intima of the artery wall.7 The hyperplastic response may be an indicator of the extent of vascular injury caused by balloon-catheter treatment. The extent of such an injury is assessed either postmortem or by ultrasound imaging.8
Ip et al7 and Schwartz9 consider thrombin and the presence of mural thrombi at the site of injury to be the major effectors of vasoconstriction and intimal hyperplasia. Indeed, McNamara et al10 have identified thrombin as a potent mitogen of smooth muscle growth in vitro. Numerous attempts to control intimal hyperplasia after balloon-catheter treatment of arteries by the administration of thrombin inhibitors in small mammals have been reported. For example, the use of heparin,11 hirudin,12 factor Xa inhibitors,13 antithrombin,14 and arginine15 have been described. All have claimed success to various degrees, but no study has followed the response throughout the vascular healing process.
The permeability properties of the regenerated endothelium have received little attention. The regenerated endothelium after balloon-catheter injury is known to exclude Evans blue dye in contrast to deendothelialized areas of the artery wall, which are readily permeable to the dye.16 17 However, it is not certain whether the regenerated endothelium fully regains the permeability properties of the original endothelium.
The following report describes measurements of the relative uptakes in vivo of plasma fibrinogen, prothrombin, and antithrombin by the IM of rabbit aortas that had been balloon deendothelialized up to 20 months previously. These observations are related to the extent of intimal hyperplasia and reendothelialization of the same ballooned aortas over the same time interval. In addition, we have used Evans blue dye to define the areas of endothelial regeneration and HRP to define the transendothelial transport and distribution of plasma proteins within the neointima that develops after balloon injury.
Preparation of Rabbit Plasma Proteins
All proteins were isolated from rabbit acid-citrate-dextrose18 plasma.
Antithrombin-α (the principal glycoform of antithrombin) was isolated from whole plasma, starting with (NH4)2SO4 precipitation, followed by heparin-Sepharose chromatography, and finally DEAE Sephadex chromatography.19 The purified protein appeared as a single Coomassie blue–stained band (Mr, 60 kD) after PAGE in the presence of 0.1% (wt/vol) SDS.20
The procedure has been described in detail previously.21 The protease inhibitors d-phenylalanyl-l-prolyl-l-arginyl-CHCl2 and 1,5-dansyl-l-glutamyl-l-glycyl-l-arginyl-CHCl2 (both from Calbiochem-Behring Corp) were used throughout the procedure. Briefly, a protein precipitate rich in vitamin K–dependent coagulation factors was produced by adding 1 mol/L BaCl2 dropwise to plasma. The precipitate was resolubilized and reprecipitated twice using 1 mol/L BaCl2. The resulting resolubilized protein was chromatographed, first on DEAE-Sephacel and then on dextran sulfate–Sepharose. A single prothrombin peak was finally rechromatographed on dextran sulfate–Sepharose. Prothrombin was dialyzed against 0.1 mol/L sodium phosphate in the presence of 5×10−6 mol/L d-phenylalanyl-l-prolyl-l-arginyl-CHCl2 and 1,5-dansyl-l-glutamyl-l-glycyl-l-arginyl-CHCl2 and finally snap-frozen and stored at −70°C.
In the presence of factor Xa, Ca2+, thromboplastin, and a trace of human plasma (as a source of factor V), rabbit prothrombin yielded 1300 IU of thrombin per milligram.21 On SDS-PAGE, prothrombin appeared as a single band of Mr 72 kD.
The purification method has been described before.22 Plasma was treated with Al(OH)3 (to remove vitamin K–dependent clotting factors), and the resulting supernatant was subjected to a series of precipitations using 6 mol/L β-alanine,23 yielding a fibrinogen-rich fraction (80% to 90% clottable with thrombin). This fraction was chromatographed on DEAE-cellulose using a Ca2+-containing buffer;24 fibrinogen, which was >95% clottable by thrombin, was eluted as a sharp peak by 0.05 mol/L Tris–0.075 mol/L NaCl–2 mmol/L CaCl2, pH 8.5. Using SDS-PAGE, the fibrinogen peak contained only a single high-molecular-weight band, which hardly entered the 7.5% (wt/vol) acrylamide gel. After reduction by β-mercaptoethanol, fibrinogen appeared entirely as α (three subforms at 68, 65, and 62 kD), β (57 kD), and γ (54 kD) chains as described previously.25
The general procedure26 27 for all proteins was as follows: 50 to 150 μg of protein (in 100 to 200 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 volume (100 to 200 μL) of 0.2 mol/L sodium phosphate, pH 7.4, was added followed by 10 μL (1 mCi) of [125I]NaI or [131I]NaI (ICN Pharmaceuticals). The reaction was carried out at room temperature for 2 minutes. To prevent aggregation of prothrombin during iodination, ε-aminohexanoic acid (final concentration, 0.025 mol/L) was added to the reaction mixture.21 The radiolabeled protein was then removed from the vial to a dialysis sac and dialyzed for 20 hours at 4°C against 250 mL of 0.01 mol/L sodium phosphate–buffered 0.14 mol/L NaCl, pH 7.4, which was changed four times. ε-Aminohexanoic acid (final concentration, 0.025 mol/L) was included in the dialysate, and d-phenylalanyl-l-prolyl-l-arginyl-CHCl2 (final concentration, ≈4×10−4 mol/L) was added to the dialyzing 131I-prothrombin at each change of dialysate (to suppress thrombin formation). All radiolabeled proteins were stored at 4°C and used for experiments within 4 days of iodination.
Free iodide content was determined by trichloroacetic acid precipitation (final concentration, 10%, wt/vol) as described before.28 Specific radioactivities of 131I-prothrombin, 125I-fibrinogen, and 125I-antithrombin measured ≈9×106, 1.2×107, and 9×106 disintegrations per minute per microgram, respectively, and free iodide (ie, nonprecipitable by trichloroacetic acid) contents amounted to 1.0% to 1.5% of the total radioactivity content. The counting efficiency of the gamma counter (Minaxi AutoGamma 5000; Canberra-Packard Ltd) was determined to be 39.8% and 75.2% for 131I and 125I, respectively; from these data and knowing the specific radioactivity, the extent of protein iodination was calculated as 1 labeled protein molecule in ≈50, 2, and 9 molecules of 131I-prothrombin, 125I-fibrinogen, and 125I-antithrombin, respectively.
All preparations of radiolabeled proteins were assessed relative to their respective unlabeled proteins by SDS-PAGE and by ND-PAGE,29 which was modified for slab gels. ND-PAGE was particularly important as a method to test whether the iodination procedure had caused aggregation of radiolabeled prothrombin because the presence of SDS destroys any evidence of aggregation.21 After electrophoresis, the gels were stained (Coomassie), dried, and then exposed to Kodak X-AR5 film for autoradiography.
In Vivo Studies
All proposals for animal experiments 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.30
For the experiments involving radiolabeled proteins in vivo, anesthetized New Zealand White male rabbits (2.8±0.1 kg; n=60) were subjected either to a deendothelializing injury (by balloon catheter; n=48) to the thoracic aorta or to a sham injury (controls; n=12). In addition, three rabbits were taken as untreated controls; ie, they received no treatment. To maintain as consistent a procedure as possible, one person undertook all the surgical duties to achieve balloon and sham injury. The rabbits were held for periods of time ranging from 5 minutes to 20 months (to allow the injured aorta wall to heal to various extents) before experimentation. The experimental details are as follows.
Balloon Deendothelialization of the Thoracic Aorta
Each rabbit was first anesthetized (sodium pentobarbital, 35 mg/kg) and an inner thigh area shaved. Under aseptic conditions, a femoral artery was exposed and ligated. A balloon catheter (Fogarty 4F; Baxter Healthcare Corp) was passed along the femoral artery up to the aortic arch. The balloon, containing sterile saline, was inflated to 1.0 mL and drawn along the thoracic aorta to deendothelialize the aorta wall as described before.31 The procedure was repeated, the catheter withdrawn, and the femoral artery tied off. For the sham-injured controls, a femoral artery was exposed and ligated but not penetrated. The leg wound was closed.
Some anesthetized rabbits were taken for experiment at 5 minutes after injury or sham injury, whereas others were allowed to recover (2 to 4 hours) in a recovery room. When conscious, the latter rabbits were transferred to regular housing and supplied with standard rabbit chow and unlimited drinking water for a period of 1 day to 20 months before experiment. No evidence of sepsis was seen during healing; consequently, treatment with antibiotic was not required.
Uptake of Radiolabeled Proteins
At a selected time after balloon or sham injury, the rabbit was anesthetized (sodium pentobarbital, up to 35 mg/kg). The neck area was shaved and a carotid artery exposed. A cannula was placed in the carotid artery (for rapid exsanguination later). The rabbit was injected (ear vein) with either 125I-fibrinogen and 131I-prothrombin or 125I-antithrombin and 131I-prothrombin. The two radiolabeled proteins were injected separately (≈23 μg 125I-fibrinogen or 15 μg 125I-antithrombin, followed by 10 μg 131I-prothrombin, each in 1 mL sterile saline). At 9 minutes after injection, a blood sample (≈0.75 mL) was taken from an ear artery into 0.25 mL acid-citrate-dextrose, and at 10 minutes, the rabbit was exsanguinated rapidly through the carotid cannula. No anticoagulant was administered to the rabbit either before or during surgery.
For those rabbits chosen for experiment at 5 minutes after balloon-catheter injury, a cannula was placed into a carotid artery before inducing the balloon-catheter injury. At 5 minutes after injury, the intravenous dose of radiolabeled proteins was injected; the radiolabeled proteins were allowed to circulate before blood sampling at 9 minutes and exsanguination at 10 minutes after injection of the dose. Again, no anticoagulant was administered either before or during surgery. The thoracic aorta was exposed surgically, flushed with 10 mL of minimum essential medium (GIBCO) containing 0.35% (wt/vol) BSA, and excised. The aorta was cut into eight 1-cm-long segments. Two 1-mm rings were removed from one segment and placed in 1 mL of 4% (vol/vol) paraformaldehyde in 0.1 mol/L sodium cacodylate to assess the extent of intimal hyperplasia and reendothelialization (see below). Surface area and radioactivity measurements were made on the isolated IM taken from seven or eight 1-cm aorta segments.31 To isolate the IM, cellulose-acetate paper (Gelman Sciences) was used en face either to strip endothelial cells cleanly from the intact aortas32 or to strip the platelet-leukocyte layer from the luminal surface of balloon-injured aortas.22 Having removed the overlying endothelium or platelet-leukocyte layer, Bergh forceps were used to separate the residual IM from the adventitia.31 The time interval between exsanguination of the rabbit and completion of the separation of aorta layers varied from 40 to 60 minutes. Radioactivity bound by the isolated IM (ie, counts per minute per square centimeter of IM) was related to the quantity of radioactivity in the circulation (ie, counts per minute per milliliter of blood) at the time of euthanitization. A hematocrit of 42% and blood concentrations of antithrombin, fibrinogen, and prothrombin equivalent to 1.47,19 3.58,34 and 1.03 μmol/L,21 respectively, were assumed for a healthy rabbit.
Uptake of HRP
Rabbits at 1 month and 12 months after a balloon-catheter deendothelialization (n=5), together with age-matched control rabbits (n=3), were injected (4.52 mg/mL in sterile saline; 1 mL/kg; ear vein) with Evans blue dye17 (Allied Chemical) 30 minutes before inducing anesthesia (sodium pentobarbital). A carotid artery and a femoral artery were cannulated with polyethylene tubing. HRP35 (type II; Sigma Chemical Company) was injected (20 mg/mL in sterile saline; 5 mL/kg) into the carotid cannula over a 2-minute interval. Two minutes after the injection was complete, oxygenated Krebs-Henseleit solution at room temperature was injected into the carotid cannula and simultaneously the femoral cannula was opened. Infusion with Krebs-Henseleit solution was continued (pressure, 100 mm Hg) for 1 minute and was then replaced by infusion of a glutaraldehyde solution (2.5%, vol/vol, in 0.1 mol/L sodium cacodylate buffer, pH 7.4). Perfusion with glutaraldehyde was continued for 3 minutes with the femoral cannula open and for a further 1 minute after the femoral cannula was closed.
The thoracic aorta was excised and opened longitudinally along the ventral surface. The areas stained blue were identified and mapped. Samples between the third and fifth pairs of intercostal arteries that were stained blue or unstained (white) or that included the junction of blue and white areas were removed and processed for TEM.
For some tissue samples, the tissue-associated HRP was reacted with DAB35 (Sigma) to form an electron-dense end product before processing the tissue for TEM. To do this, samples were rinsed for 18 hours in 0.05 mol/L Tris-HCl buffer, pH 7.6, and then incubated in DAB solution (0.5 mg/mL in 0.05 mol/L Tris-HCl buffer, pH 7.6, with fresh H2O2 [30%] added, to a final concentration of 0.01%, vol/vol) in the dark with agitation for 20 minutes. The tissues were rinsed three times in distilled water prior to postfixation in 1% (wt/vol) OsO4 in Tris buffer for 1 hour at 4°C. The samples were then dehydrated and embedded in Spurr resin by using a standard technique.17 Some of the tissue samples were snap-frozen and used to prepare frozen sections to further identify the distribution of HRP. HRP activity in the sections was identified by using DAB as described above for tissue samples. Sections were counterstained with Mayer's hematoxylin.
Sample sections (0.5 μm thick), taken from plastic-embedded tissues, were mounted on glass slides, stained with toluidine blue, and examined by LM. Thin sections for TEM were cut from areas of interest identified in the 0.5-μm sections, mounted on copper grids, stained with lead citrate, and examined in a Philips TEM 301 (Eindhoven).
Morphology of the Neointima
Sections from the aortas of 10 sham-injured or uninjured control rabbits and 15 of the balloon-injured animals were examined by LM to determine the extent of intimal hyperplasia and endothelial regeneration. All LM evaluations were made, independently, by two people who had no knowledge of the previous history of the tissues.
Two 1-mm rings were cut from each aorta; one ring included an opening to one of the fourth pairs of intercostal arteries, and the other ring was taken from midway between the fourth and fifth pairs of intercostals (see above). These samples, stored in paraformaldehyde, were processed for paraffin-wax embedding. Serial whole-circumference sections (5 μm thick) were cut from each aorta ring and mounted on glass slides. One section was stained with hematoxylin/eosin. Adjacent sections were stained for the presence of vWF to identify the endothelium. These sections were incubated in 1% (wt/vol) BSA in TBS (0.05 mol/L Tris-0.14 mol/L NaCl, pH 7.4) (as a blocking step), rinsed, and then exposed to goat anti-human vWF (Atlantic Antibodies) in BSA-TBS for 1 hour, rinsed, and exposed to biotinylated rabbit anti-goat IgG (Vector Labs) in TBS for 1 hour. After further rinsing, biotinylated anti-vWF was visualized by using the Elite Vectastain ABC kit (Vector Labs). The sections were counterstained with Mayer's hematoxylin. The proportion of luminal surface that stained positive for vWF was measured relative to the entire luminal surface by LM.
The numbers of layers of intimal smooth muscle were counted at eight predetermined points around the circumference of each hematoxylin/eosin–stained aorta ring and from these values the mean number of smooth muscle layers was calculated.
Calculation of Data
When appropriate, data are given as the mean±SEM.
Properties of Purified Prothrombin, Antithrombin, and Fibrinogen After Radiolabeling
Fig 1⇓ shows the PAGE patterns of these proteins before and after radiolabeling. As demonstrated by ND-PAGE (Fig 1a⇓), 131I-prothrombin (lane 2) migrated as a single component with a similar mobility to unlabeled prothrombin (lane 1) and contained essentially no aggregated products after iodination. Both radiolabeled antithrombin (Fig 1b⇓, lane 2) and fibrinogen (Fig 1c⇓, lane 2) migrated as single-band components that were coincident with their respective unlabeled proteins on SDS-PAGE. After reduction by β-mercaptoethanol, 125I-fibrinogen appeared entirely as α, β, and γ chains (Fig 1c⇓; compare lanes 3 and 4).
Uptake of Radiolabeled Proteins by the Aorta Wall In Vivo Before and After Injury
At various times ranging from 5 minutes to 20 months after balloon or sham injury, anesthetized rabbits were injected with two radiolabeled proteins, either 131I-prothrombin and 125I-antithrombin or 131I-prothrombin and 125I-fibrinogen. The labeled proteins were allowed to circulate for only 10 minutes. The quantity of each protein associated with the platelet-leukocyte–free (or endothelium)-free IM layer was measured relative to that in the circulation. Fig 2⇓ shows the quantities (expressed as picomoles per square centimeter) of these proteins associated with the aortic IM. The concentration of prothrombin (0.1 to 0.3 pmol/cm2), antithrombin (0.3 to 0.4 pmol/cm2), and fibrinogen (0.6 to 0.9 pmol/cm2) within the IM of uninjured or sham-injured aortas did not change significantly as these groups of rabbits aged over 12 months or more.
In the ballooned-injured rabbits, the concentration of IM-bound fibrinogen increased to 10.9±2.3 pmol/cm2 by 5 minutes after injury, falling significantly to 5 to 6 pmol/cm2 by 1 day (Fig 2⇑). This concentration did not change significantly over the following 20 months. The concentrations of IM-bound prothrombin (3.7±0.5 pmol/cm2) and antithrombin (3.3±0.4 pmol/cm2) at 5 minutes remained at elevated levels for up to 60 to 70 days after injury and then decreased to 2.2±0.4 pmol/cm2 and 1.5±0.7 pmol/cm2 (P<.05), respectively, over the following 10 months. The concentration of prothrombin and antithrombin contained by one balloon-injured aorta IM at 20 months after injury resembled closely that of an uninjured aorta IM.
In general, the quantity of IM-bound 125I-prothrombin contained by individual aorta segments after balloon injury was matched closely by the quantity of IM-bound antithrombin. Fig 3a⇓ shows that a linear relationship exists between the molar ratio of IM-bound antithrombin/prothrombin and the concentration of IM-bound prothrombin for rabbits injected at 5 minutes after balloon injury. However, as the rate of uptake of prothrombin increased, the molar ratio of antithrombin/prothrombin decreased.
A plot of the molar ratio of IM-bound antithrombin/IM-bound prothrombin for all aortas against time after balloon injury revealed an approximately equimolar relationship that extended beyond the immediate period after injury (Fig 3b⇑). We conclude that the rate of insudation of antithrombin largely matches that of prothrombin at the site of injury.
Intimal Thickening and Reendothelialization of the Ballooned Aorta In Vivo
A smooth muscle cell–rich neointima developed after balloon-catheter injury. Fig 4⇓ shows the increase in intimal smooth muscle layers with respect to time after injury compared with sham-injured control aortas. The number of layers increased rapidly from 0 up to 6 to 8 layers over the first 60 days, after which intimal growth slowed markedly.
These same vessel samples were stained using anti-vWF antibody (Fig 5⇓) to indicate the extent of endothelial covering (Table⇓). The reendothelialization process, as shown by vWF-positive luminal cells in adjacent sections, continued progressively after balloon injury. Ballooned aortas were ≈50% reendothelialized by about 60 to 100 days36 and 80% to 100% reendothelialized by 12 to 20 months after balloon injury.
Evaluation of the Permeability of Deendothelialized and Reendothelialized Areas of the Aortic Intima by HRP Activity
For these studies, rabbits that had received balloon-catheter treatment to the thoracic aorta were housed for 1 or 12 months after injury. The luminal surface of all ballooned aortas appeared blue stained to various extents by Evans blue, although no blue staining was seen in aortas from two control rabbits. At 1 month after balloon injury to two rabbits, ≈70% of the surface was stained blue, with the white areas surrounding the vessel orifices. The junction between the blue and white areas could be clearly identified. By LM and TEM, the white areas corresponded to the reendothelialized regions and the blue areas to the deendothelialized regions where smooth muscle cells formed the luminal layer (Figs 6 and 7⇓⇓). At 12 months after injury (two rabbits), only small areas, amounting to ≈10% to 20% of the luminal surface, appeared blue; these aortas were assumed, therefore, to be ≈80% to 90% reendothelialized.
Regarding HRP activity, minimal staining by DAB was observed by LM of sections from the uninjured vessel of either a 1-month or a 12-month control rabbit. By contrast, in aortas at 1 and 12 months after balloon injury, reendothelialized areas that were remote from the junction with deendothelialized areas showed no evidence of HRP reaction product (Fig 6a⇑), whereas deendothelialized areas were heavily stained with DAB (indicating penetration by HRP) to the full depth of the neointima to the internal elastic lamina (Fig 6b⇑). At the junction between reendothelialized and deendothelialized areas, the reaction product extended for some distance beneath the leading edge of the endothelial layer (Fig 6c⇑), and this was also evident by TEM (Fig 7a⇑).
By TEM, there was minimal evidence of HRP reaction product within the subendothelial space of the normal aorta, although some endothelial cells included vesicles or vacuoles that contained reaction product (Fig 7b⇑). The regenerated endothelial cells of balloon-injured aortas showed similar intercellular staining, but there was no evidence of increased accumulation of HRP in the subendothelium (Fig 7c⇑).
Over the past 10 years, the rabbit aorta has been the model used in our laboratory for investigating the short- and long-term behavior of certain plasma proteins in their hemostatic response to balloon-catheter–induced vascular injury. Fibrinogen, prothrombin, and antithrombin have been shown to be taken up at various rates by the deendothelialized aorta immediately after injury in vivo.19 21 22 Fibrinogen saturated the ballooned aorta wall within 5 minutes, whereas maximal uptake of prothrombin was not observed until 15 to 20 minutes and of antithrombin until 45 to 60 minutes after injury. Moreover, measurable thrombin activity was produced and released from the damaged aortic surface for at least 10 days after injury.22
In the present studies, the relatively elevated rates of uptake of prothrombin, fibrinogen, and antithrombin by the aortic IM observed immediately after a deendothelializing injury are seen to fall slowly over the ensuing 20 months. At 12 months, the rates of uptake were still significantly greater than those of respective proteins in the IM of the uninjured aorta. The principal reasons for an enhanced uptake of plasma protein, particularly those involved directly in hemostasis, by the injured aorta IM are an increased permeability of the luminal surface during the healing process and an increased sequestration within the neointima coupled with an increased metabolic turnover.
After a balloon-catheter injury to the aorta, the luminal surface is essentially denuded of endothelium (Table⇑). Almost immediately, smooth muscle cells migrate from the media into the residual intima and start to proliferate. This neointima, consisting of smooth muscle cells and fibrous and nonfibrous connective tissues,37 grows progressively thicker with time (Fig 4⇑). Simultaneously, progressive reendothelialization is observed (Table⇑), originating largely from the intact endothelium of the intercostal arteries.38 39
From the results of the permeability study (Figs 6 and 7⇑⇑), we observed that the newly generated endothelium is relatively impermeable to Evans blue (ergo albumin) and to HRP. From this, we assume that the radiolabeled plasma proteins are also entering the aorta wall through deendothelialized rather than reendothelialized areas. At 12 months, the reendothelialization process appeared to be 80% to 90% complete (Table⇑), and although the elevated rates of uptake of fibrinogen, prothrombin, and antithrombin by the ballooned IM were declining, each was significantly greater (P<.01) than that associated with the uninjured IM.
As shown in Fig 6b⇑, HRP staining was observed to the full depth of the deendothelialized intima. Presumably, adsorbed fibrinogen, prothrombin, and antithrombin would also occupy a similar space. One may speculate that as the intimal volume increases over the first 70 days after injury, a larger “sink” becomes available for plasma proteins to occupy as the surface area of the deendothelialized lumen decreases. The net effect is that the uptake of prothrombin (Fig 2b⇑) and antithrombin (Fig 2c⇑) remained relatively unchanged for the first 70 days after injury. In contrast, the rate of uptake of bound fibrinogen decreased significantly over the first 24 hours after injury, declining more slowly over the next 10 months or more. From experiments in vitro, most of the fibrinogen bound by the deendothelialized IM is dependent on the presence of active thrombin within the deendothelialized IM.22 25 Thus, from the uptake of fibrinogen (Fig 2a⇑) and from previous observations on thrombin released from the luminal surface of the balloon-injured aortic wall,22 we conclude that the concentration of active thrombin within the intima falls rapidly over the first 24 hours and continues to decrease steadily over the first 70 days after balloon injury, presumably because control of prothrombin metabolism by endogenous antithrombins is intensifying within the deendothelialized intima. Thus, as the time after injury increases, the extent of fibrinogen uptake may depend less on the presence of active thrombin within the site of injury and more on passive diffusion.
The process of healing may also be dependent on the severity of the injury produced.40 In the present experiments, despite the care taken to standardize the injury, considerable variation of response was observed. A good example of this variation is found in Fig 3a⇑ in which, within a group of 10 rabbits injected with 131I-prothrombin at 5 minutes after balloon injury, the amount of prothrombin bound over a 10-minute interval ranged from 1.5 to 6.1 pmol/cm2. We believe that this reflects the different degree of injury produced in each aorta. Fibrinogen uptake by the deendothelialized aortic IM is also known to vary widely.21 22 As mentioned above, the extent of fibrinogen deposition in vitro has been found to be directly proportional to the content of active thrombin within the deendothelialized IM.25 The extent of fibrinogen uptake by the IM, therefore, may be an indicator of thrombin presence on or within the denuded surface and consequently a direct measure of the extent of injury. The rate of antithrombin uptake by the deendothelialized rabbit aorta over the initial 10-day period after injury in vivo correlated closely with thrombin activity released from this damaged tissue ex vivo22 and with prothrombin uptake by the ballooned aorta.21 Thus, any continuity of thrombin production after injury would require a continuous prothrombin supply and would be matched by a proportionate uptake of antithrombin and fibrinogen.
We cannot ignore the likelihood that at 12 or even 20 months after injury, the thickened deendothelialized intima is relatively thrombogenic. Several reports7 9 41 have pointed to thrombin as an important mitogen for smooth muscle growth. The duration of greatest uptake of prothrombin and antithrombin, ie, the initial 70-day period after injury (Fig 4⇑), was commensurate with the most active phase of smooth muscle proliferation, the rates of uptake of both proteins decreasing as smooth muscle cell proliferation slowed down. Comparing the results of Fig 4⇑ with those of the Table⇑, it is evident that the rate of smooth muscle cell hyperplasia appeared to slow as the endothelialization process approached the 50% reendothelialized stage. Whether the endothelium directly or indirectly suppressed the growth of intimal smooth muscle cells is not fully understood, although endothelial cells are known to inhibit smooth muscle cell mitosis when both cell types are grown in coculture.39
The chronically increased uptakes of prothrombin, fibrinogen, and antithrombin do not preclude an increased generation of thrombin within the wound site over the entire reendothelializing period. The extent to which IM-bound prothrombin is converted to thrombin or IM-bound fibrinogen to fibrin during the period of reendothelialization after balloon injury is not yet known, although, as shown in Fig 3b⇑, an approximately equimolar ratio of IM-bound antithrombin/prothrombin was sustained for up to 20 months after injury.
Selected Abbreviations and Acronyms
|PAGE||=||polyacrylamide gel electrophoresis|
|PTCA||=||percutaneous transluminal coronary angioplasty|
|TEM||=||transmission electron microscopy|
|vWF||=||von Willebrand factor|
This study was supported by a grant-in-aid from the Heart and Stroke Foundation of Ontario. We thank Shelley Serebrin and Myron Kulczycky for technical assistance.
Chesebro JH, Badimon L, Fuster V. Importance of antithrombin therapy during coronary angioplasty. J Am Coll Cardiol. 1991;17:96B-100B.
Bar-Shavit R, Eldor A, Vlodavsky I. Binding of thrombin in subendothelial extracellular matrix: protection and expression of functional properties. J Clin Invest. 1989;84:1096-1104.
Kurihara H, Masao Y, Takao S, Kazuhide Y, Ryozo N, Fumimaro T, Hiroyuki S, Jun I, Masashi Y, Tomoh M, Yoshio Y. The possible role of endothelin-1 in the pathogenesis of coronary vasospasm. J Cardiovasc Pharmacol. 1989;13(suppl 5):S132-S137.
Fischell TA, Derby G, Tse TM, Stadius ML. Coronary artery vasoconstriction routinely occurs after percutaneous transluminal coronary angioplasty. Circulation. 1988;78:1323-1334.
Ip JH, Fuster V, Israel D, Badimon L, Badimon J, Chesebro JH. The role of platelets, thrombin and hyperplasia in restenosis after coronary angioplasty. J Am Coll Cardiol. 1991;17:77B-88B.
Schwartz S. Serum-derived growth factor is thrombin? J Clin Invest. 1993;91:4. Editorial.
McNamara CA, Sarembock IJ, Gimple LW, Fenton JW, Coughlin SR, Owens GK. Thrombin stimulates proliferation of cultured rat aortic smooth muscle cells by a proteolytically active receptor. J Clin Invest. 1993;91:94-98.
Guyton JR, Rosenberg RD, Clowes AW, Karnovsky MJ. Inhibition of rat arterial smooth muscle cell proliferation by heparin, I: in vivo studies with anticoagulant and non-anticoagulant heparin. Circ Res. 1980;46:625-634.
Sarembock IJ, Gertz SD, Gimple LW, Owen RM, Powers ER, Roberts WC. Effectiveness of recombinant desulfato-hirudin in reducing restenosis after balloon angioplasty of atherosclerotic femoral arteries in rabbits. Circulation. 1991;84:232-243.
Ragosta M, Gimple LW, Gertz SD, Dunwiddie CT, Vlasuk GP, Haber HL, Powers ER, Roberts WC, Sarembock IJ. Specific factor Xa inhibition reduces restenosis after balloon angioplasty of atherosclerotic femoral arteries in rabbits. Circulation. 1994;89:1262-1271.
Ali NM, Mazur W, Kleiman NS, Rodgers GP, Schulz DG, Abukhali JM, French BA, Raizner AE. Inhibition of coronary restenosis by antithrombin III in atherosclerotic swine. Circulation. 1994;90:I-239. Abstract.
Tarry WC, Makhoul RG. l-Arginine improves endothelium-dependent vasorelaxation and reduces intimal hyperplasia after balloon angioplasty. Arterioscler Thromb. 1994;14:938-943.
Minick CR, Stemerman MB, Insull W. Effect of regenerated endothelium on lipid accumulation in the arterial wall. Proc Natl Acad Sci U S A. 1977;74:1724-1728.
Richardson M, Moore S. Preparation of large, curved biological surfaces for scanning electron microscopy. Artery. 1980;6:409-417.
Aster RH, Jandl JH. Platelet sequestration in man, I: methods. J Clin Invest. 1964;43:843-854.
Witmer MR, Hatton MWC. Antithrombin III-β associates more readily than antithrombin III-α with the uninjured and deendothelialized rabbit aortic wall in vitro and in vivo. Arterioscler Thromb. 1991;11:530-539.
Straughn W, Wagner RH. A simple method for preparing fibrinogen. Thromb Diath Haemorrhagica. 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.
Regoeczi E. Iodine-Labeled Plasma Proteins, I. Boca Raton, Fla: CRC Press; 1984:49-53.
Hatton MWC, Southward SMR, Ross-Ouellet B, Richardson M, Winocour PD. Comparative metabolism of rabbit plasminogen glycoforms I and II in the alloxan-diabetic rabbit. Am J Physiol. 1995;269:E1017-E1023.
Clarke JT. Simplified “disc” (polyacrylamide gel) electrophoresis. Ann N Y Acad Sci. 1964;121:428-436.
Olfert ED, Cross BM, McWilliam AA, eds. Guide to the Care and Use of Experimental Animals, I. 2nd ed. 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 deendothelialized rabbit thoracic aorta in vitro. Blood. 1986;67:878-886.
Stemerman MB. Effects of moderate hypercholesterolemia on rabbit endothelium. Arteriosclerosis. 1981;1:25-32.
Fager G. Thrombin and proliferation of vascular smooth muscle cells. Circ Res. 1995;77:645-650.