Author + information
- Received September 18, 2001
- Revision received April 8, 2002
- Accepted April 18, 2002
- Published online July 17, 2002.
- Tullio Palmerini, MD*,
- Mark A Nedelman, MS†,
- Lesley E Scudder, BSc‡,
- Marian T Nakada, PhD∥,
- Robert E Jordan, PhD∥,
- Susan Smyth, MD, PhD‡,
- Ronald E Gordon, PhD§,
- John T Fallon, MD, PhD§ and
- Barry S Coller, MD‡,* ()
- ↵*Reprint requests and correspondence:
Dr. Barry S. Coller, Head, Laboratory of Blood and Vascular Biology, The Rockefeller University, 1230 York Avenue, New York, New York 10021, USA.
Objectives This study was designed to assess the effect of abciximab on platelet and leukocyte deposition 60 min after stent insertion in nonhuman primates.
Background Although it is well established that abciximab improves both short- and long-term clinical outcomes after stent placement, there have been no studies assessing its effect on early platelet and leukocyte deposition.
Methods Cynomolgus monkeys were pretreated with aspirin and either saline or a 0.4 mg/kg bolus of abciximab, and then subjected to angioplasty and Palmaz-Schatz stent placement in the common iliac artery or abdominal aorta. After 60 min, animals were euthanized and the stented artery was evaluated by immunohistochemistry and morphometry.
Results Complete occlusion of the stented vessel with a thin fibrin(ogen) meshwork and trapped blood occurred in two saline-treated and two abciximab-treated animals. In the four remaining saline-treated animals, a layer of erythrocytes trapped in a network of fibrin(ogen) was noted close to the vessel wall, and this was covered by a layer of large, irregular platelet thrombi. Leukocytes formed a monolayer on top of the platelets and near stent struts. In the four remaining abciximab-treated animals, the mean erythrocyte area was 65% smaller (p = 0.070), the platelet aggregate area was 89% smaller (p = 0.049) and the luminal area was 59% larger (p = 0.004). A monolayer of leukocytes also formed on top of the platelets and near stent struts.
Conclusions In control stented blood vessels in this study, platelet thrombi formed not at the vessel wall, but on top of an erythrocyte-rich layer, and platelets recruited leukocytes. Abciximab decreased the size of platelet thrombi, but did not prevent leukocyte recruitment.
Adding abciximab treatment to coronary artery stent insertion has been demonstrated to have both short- and long-term beneficial effects, including prevention of ischemic complications during the first 30 days (1), improved microcirculatory function and recovery of myocardial function (2), prevention of the development of restenosis in diabetic patients (3)and reduced one-year mortality (4). Proposed mechanisms to explain these benefits include inhibition of platelet aggregation, reduced platelet-mediated thrombin generation, reduced release from platelets of vasoactive agents, reduced platelet-mediated clot retraction and reduced intimal hyperplasia in response to vascular injury (reviewed in Coller , Coller and Smyth et al. ). It is also possible that abciximab’s effects are mediated through inhibition of platelet-leukocyte interactions (8), because such interactions have been implicated in thrombin generation, microcirculatory damage and the development of intimal hyperplasia after vascular injury (reviewed in Coller and Smyth et al. ). Moreover, abciximab has been reported to interact with an activated form of the leukocyte integrin αMβ2, which has been implicated in the development of intimal hyperplasia after vascular injury (9).
To better understand the effect of abciximab on platelet and leukocyte deposition at the site of stent insertion, we performed angioplasty and stent insertion in arteries of nonhuman primates treated with saline or abciximab. We evaluated the arteries 60 min after stent insertion because platelet-leukocyte interactions are prominent at that time point in other models of vascular injury (7), and there have been very few studies of the blood cell interactions that occur within the first hour after stent insertion. We chose not to treat the animals with heparin because: 1) we wished to assess the effect of abciximab on thrombin generation at the site of vascular injury, and heparin’s anticoagulant effects might confound such assessment; 2) heparin has demonstrated inconsistent results in protecting against acute stent thrombosis in in vitro and animal models (10–14); and 3) heparin is known to activate platelets in vitro (15)and in vivo (16)and to enhance in vivo platelet-leukocyte aggregate formation (16).
Platelet function and coagulation assay
Platelet function was measured with a modified version of the rapid platelet function assay (RPFA) (Accumetrics, San Diego, California) (17)using 20 μM adenosine diphosphate as the agonist.
Animal preparation and surgical procedure
Twelve male cynomolgus monkeys (Maccaca fasicularis) weighing 3.3 to 8.5 kg were treated with aspirin (10 mg/kg/day by mouth) starting three or four days before the procedure. Animals were sedated with 10 mg/kg ketamine, intubated and mechanically ventilated with 2% isoflurane-98% o2at a minute ventilation of 1.5 l/min. 111Indium-labeled autologous platelets from donor monkeys (18)were injected through a sheath in the jugular vein 75 to 110 min before the injury. An arterial sheath (6 to 7 French) was placed in the carotid artery to facilitate the placement of the guide catheter and the angioplasty catheter. An angiogram was performed to measure the vessel diameter. Animals were treated with either saline or a 0.4 mg/kg intravenous bolus of abciximab.
Vascular injury was performed with a 3 to 4 mm semicompliant angioplasty balloon (Olympix, Cordis, Miami, Florida; Valor, Cordis; ACS lifestream, Advanced Cardiovascular Systems, Santa Clara, California). The vessel was stretched to 1.4 times its initial diameter by inflating the balloon three times for 1 min. Mean (±SD) inflation pressures were 7.3 ± 2.1 atms in the abciximab group and 6.5 ± 1.1 atms in controls (p = 0.4). Immediately thereafter, a Palmaz Schatz stent was deployed by stretching the vessel to 1.5 times its initial diameter by balloon inflation three times for 30 s. Mean inflation pressures were 12.7 ± 1.5 atms and 13.2 ± 1.3 atms in the abciximab and control groups (p = 0.6). One hour after stent deployment, animals were euthanized and perfusion-fixed with saline and 4% paraformaldehyde. The vessels containing the stents were harvested as 1.2-cm-long segments, weighed and analyzed for 111Indium in a gamma spectrometer; they were then further fixed in 4% paraformaldehyde for histology.
Blood was drawn through the jugular vein immediately after jugular catheter placement, after angiography, 10 min after drug injection and 60 min after stent deployment.
Histology and immunohistochemistry
Specimens were dehydrated and then infiltrated and embedded in methylmethacrylate. Stents were cut into three 4-mm-long segments using a diamond-edged saw (Isomet, Buehler, Lake Bluff, Illinois). The central portion (containing the hinge region) was not studied, and sections were cut from each of the two other segments for each stain. Sections (5 to 8 μm) for leukocyte analysis were stained with hematoxylin and eosin, whereas sections for histochemistry were soaked in 3% H2o2and then blocked with 2% ovalbumin in phosphate-buffered saline (PBS). After washing in PBS, sections were incubated with a primary antibody for 2 h at 37°C, and then with a biotinylated species-specific secondary antibody for 20 min at 22°C. The slides were reacted with horseradish peroxidase-conjugated streptavidin, developed with diaminobenzidine (Biogenex, San Ramon, California) and counterstained with hematoxylin.
Antibodies to human P-selectin and PSGL-1 were from Pharmingen (San Diego, California); fibrin(ogen) from Behring (King of Prussia, Pennsylvania); platelet factor-4 and plasminogen activating inhibitor-1 (PAI-1) from American Diagnostica (Greenwich, Connecticut); tissue plasminogen activator (t-PA) and tumor necrosis factor-alpha from Accurate Chemical (Westbury, New York); platelet-derived growth factor A chain from Santa Cruz Biotechnology (Santa Cruz, California); interleukin-1-beta from Harlan Sera-Lab (Loughborough, United Kingdom); thrombin/prothrombin from Dr. Peter Harpel (19); and tissue factor (TF) from Dr. Yale Nemerson (20).
Vessel injury produced by the stent was determined using an injury score modified from Schwartz et al. (21). An observer blind to the treatment groups performed the histomorphometric analysis on one representative section from each animal. Using Image Pro plus (Media Cybernetics, Silver Spring, Maryland), the following areas were determined: 1) the lumen; 2) inside the internal elastic lamin (IEL); 3) the erythrocyte-rich layer; and 4) the platelet thrombi. The number of leukocytes in two sections from each animal were counted and normalized for the circumference of the vessel.
Segments were cut from blood vessels of one control animal and one animal treated with abciximab (Numbers 1 and 5, Table 1) and treated with methyl acetate to dissolve the methacrylate. After treatment with 2% glutaraldehyde and 1% osmium tetroxide, the tissue was dehydrated and embedded in EMbed 812 (EMS, Fort Washington, Pennsylvania). Ultrathin sections were stained with uranyl acetate and lead citrate.
Data are described as mean ± SD. Individual values of hematologic and coagulation measurements are shown in Table 1. Statistical analyses of injury score and histomorphometric data were performed using two-tailed ttests of differences in means for unequal variances. The exact p values for each test are reported as a guide to interpreting the results, but no claims of statistical significance are made because these are exploratory analyses.
Arterial injury and stenting
The study was conducted on two separate groups of animals. The first group (eight animals, 5.2 to 8.5 kg) had iliac arteries that were large enough to accommodate stents (2.5 to 3 mm). The second group (four animals) was smaller (3.3 to 4.7 kg) and had smaller iliac arteries (<2 mm), and so stents were placed in the aortas (2.5 to 3.0 mm). The mean (±SD) diameters of the stented vessels were 2.8 ± 0.3 mm in the controls and 2.8 ± 0.2 mm in the abciximab group.
The first four animals from the first group (two controls and two abciximab-treated) were found to have occluded stents at the time of harvest. Histologic analysis of multiple sections revealed blood trapped in a loose fibrin mesh occluding the vessel. Platelet thrombi were prominent near the stent in one of the control animals, but not in the other three. Stent struts were not properly apposed to the blood vessel wall in both abciximab-treated animals (for the total circumference in one animal and for half the circumference in the other), indicating poor stent deployment (data not shown). These animals were not included in the statistical assessment because the age of the occlusions could not be determined, and the histomorphometric pattern was fundamentally different from that of the other animals (as described later).
The mean injury scores in the eight animals included in the statistical assessment were similar in control and abciximab groups (2.15 ± 0.31 vs. 2.33 ± 0.17, p = 0.25) (Table 1).
Hematology, coagulation and platelet function assays
Hematology and coagulation values were similar among animals within both groups before and after the experiments (Table 1).
Rapid platelet function assay slope values increased at the end of the experiment in all of the control animals (mean 38%). Ten minutes after abciximab administration, the RPFA values in all abciximab-treated animals decreased to 0; at 60 min, the RPFA values in animals 5 to 7 remained 0, but the value in animal 8 was 26% of the baseline value.
Figure 1shows representative sections from one animal in each group. In all four animals in the control group there was a layer of erythrocytes within a fine network of fibrin close to the vessel wall along the entire, or nearly entire, circumference of the vessel (Fig. 1A,left panel). On top of this layer was a variably thick layer of fibrin, and on top of the fibrin layer were large, irregular platelet thrombi (Fig. 1A,center panel). In some animals, platelet thrombi were also identified near the stent struts. Numerous leukocytes deposited on top of the outermost layer of platelets making up the thrombi (Fig. 1A,right panel). In addition, small numbers of leukocytes were present in irregular clusters and in isolated linear arrays near the stent struts and within the platelet thrombi (Fig. 1,A,center and right panels). In abciximab-treated animals (Fig. 1B) the erythrocyte-rich areas and platelet thrombi appeared smaller. Leukocytes also lined the vessel wall of the abciximab-treated animals, and leukocytes were also present in small thrombi adjacent to the stent struts.
By histomorphometry, the areas inside the IEL were similar between the control and abciximab-treated groups (9.51 ± 1.56 mm2vs. 10.7 ± 1.89 mm2; p = 0.36) (Table 1), indicating that the blood vessels in the two groups were similar in size. In the abciximab-treated animals, however, the average erythrocyte-rich area was 65% smaller, the average area of platelet aggregates was 89% smaller and the average luminal area was 59% larger. The differences in erythrocyte-rich areas in the abciximab-treated animals and controls were not as great as the differences in platelet thrombi areas and luminal areas (erythrocyte areas 1.05 ± 0.84 mm2vs. 3.01 ± 1.46 mm2[p = 0.070]; platelet areas 0.05 ± 0.04 mm2vs. 0.45 ± 0.25 mm2[p = 0.049]; luminal areas 9.64 ± 1.21 mm2vs. 6.06 ± 1.01 mm2[p = 0.004]). The mean number of leukocytes deposited was 55% lower in the abciximab-treated group, but the variability between animals in the control group was considerable (33 ± 12 per mm vs. 74 ± 76 per mm; p = 0.36).
Transmission electron microscopy performed on control animal 1 confirmed the presence of exuberant platelet thrombi, with variable loss of platelet granules and discernible borders between individual platelets (Figs. 2A and 2B). Electron micrographs from animal 5, which was treated with abciximab, revealed a single layer of variably spread platelets adherent to the damaged blood vessel, just above the IEL (Figs. 2C and 2D).
In the sections from the control animals, the outermost surface of the platelet aggregates, as well as the majority of the areas within the aggregates, stained positive for platelet factor 4, P-selectin (Fig. 3A), CD40 ligand (Fig. 3C), von Willebrand factor, PAI-1, interleukin-1-beta, tumor necrosis factor-alpha and platelet-derived growth factor A chain.
Antibody to fibrin(ogen) stained the periphery of the platelet aggregates, the thin fibrin strands throughout the erythrocyte layer, and the thick fibrin layer. Antibodies to prothrombin/thrombin, and to t-PA gave patterns similar to those found with the antibody to fibrin(ogen). Staining for TF also gave a pattern that was very similar to that with the antibody to fibrin(ogen), but in addition, granulocytes gave weak staining and monocytes gave more intense staining (Fig. 3E). Leukocytes, but not platelets, stained positive for P-selectin glycoprotein ligand-1 (PSGL-1).
In abciximab-treated animals, the antibodies to platelet alpha-granule and membrane proteins stained a discontinuous layer of platelets on the blood vessel surface (Figs. 3B and 3D). Antibodies to fibrin(ogen) and prothrombin/thrombin also stained very weakly. Tissue factor antigen was present as a discrete but thin layer on the vessel wall; in addition, individual neutrophils stained variably positive and monocytes stained strongly positive (Fig. 3F). Leukocytes stained positive for PSGL-1.
In the control group and three of the four abciximab-treated animals, the platelet deposition values correlated well with the platelet thrombus area by histomorphometry. In animal 8, however, platelet deposition was high relative to the platelet thrombus area. This was the only animal to sustain loss of a segment of the vascular wall in response to the injury, and thus radiolabeled platelets may have been trapped outside the lumen of the blood vessel.
Acute occlusions of stented vessels
The first two animals in both the saline- and abciximab-treated groups had occluded stents at 60 min, making it impossible to establish the time of occlusion. Histomorphometry demonstrated that the majority of the occluding thrombus consisted of a thin fibrin mesh entrapping blood. Stent deployment was good in both control animals, but was suboptimal in both abciximab-treated animals. It is notable that abciximab did not prevent the occlusion, despite preventing platelet aggregate formation.
Erythrocyte-rich area near blood vessel wall
In control animals, the presence of a layer of blood trapped in a fine fibrin network adjacent to the vascular wall was both striking and unexpected. The vascular wall itself showed little evidence of thrombus formation, although there was some leukocyte deposition and platelet thrombus formation near the stent struts. We could not find previous descriptions of studies in which blood vessels were analyzed 60 min after stent placement, but Carter et al. reported mural thrombus, with morphology very similar to what we observed, 24 h after stent placement in porcine coronary arteries (22). Of note, Carter et al. treated their animals with heparin, making it unlikely that our results are due to our not using heparin. We propose that stent insertion leads to a reduction in blood flow immediately adjacent to the wall, which favors thrombin generation and the development of a fibrin network that can trap the nearby layer of stagnant blood. Platelet thrombi can then form on top of that surface. The trapped layer of blood may ultimately serve as a scaffold for subsequent smooth muscle cell migration, leading to intimal hyperplasia (22).
There was a trend toward a smaller mean area of trapped erythrocytes with abciximab treatment, and in one animal there was no layer at all. The most plausible explanation for abciximab’s effect is that it decreases platelet thrombus formation at the sites where the stent struts contact the blood vessel wall and thus decreases thrombin generation and fibrin formation.
The large platelet thrombi originated not at the blood vessel wall, but rather on top of the trapped blood cell layer. Electron microscopy confirmed that the thrombi were composed almost exclusively of platelets, and that the platelets in the deeper regions of the thrombi had coalesced.
Abciximab treatment reduced platelet thrombus formation area by 89%, but did not prevent the deposition of a discontinuous monolayer of platelets. These results are consistent with in vitro and in vivo evidence in other systems, demonstrating that abciximab blocks platelet-platelet interactions, but not platelet adhesion (5,6).
Tissue factor antigen was present on the outer surface of the platelet aggregates in the control animals and on the surface of the platelet monolayer in the abciximab-treated animals. These observations are consistent with the data of Giesen et al. (23)and Rauch et al. (24), who demonstrated that TF circulates in blood in association with small lipid vesicles and that these TF-rich vesicles can bind to platelet aggregates that form on collagen-coated surfaces in vitro. Thus, platelet aggregates can recruit blood-borne TF, providing a mechanism to initiate thrombin generation even in the absence of exposing tissue factor in the lipid-rich core of an atherosclerotic lesion or in the adventitia.
In control animals, there was a dense layer of leukocytes on the outside of the platelet aggregates. Leukocytes were also found in clusters and in linear arrays near the stent struts, in the trapped blood layer and within the platelet aggregates. Abciximab-treated animals had less extensive platelet aggregate formation and thus had fewer leukocytes associated with these aggregates. Abciximab did not, however, prevent the binding of leukocytes to the platelet monolayer on the surface of the damaged blood vessel or to sites adjacent to the stent struts. These data are consistent with our findings in a vascular injury model in mice lacking glycoprotein (GP) IIb/IIIa and αVβ3, where leukocytes still deposited on top of the platelet monolayer 60 min after injury to the femoral artery (7). Thus, other receptors can support platelet-leukocyte interactions, including but not limited to the interaction between P-selectin on activated platelets and PSGL-1 on leukocytes (reviewed in Coller ) and perhaps platelet GP Ib with leukocyte αMβ2 (25). Our data do not, however, exclude the possibility that GP IIb/IIIa- and/or αVβ3-mediated interactions may contribute to firm adhesion of leukocytes on platelets or platelet-dependent leukocyte activation (8).
The platelet thrombi and platelet-leukocyte interactions that occur in the region of stent insertion may serve as an ongoing source of platelet microparticles, platelet emboli and platelet-leukocyte emboli that may damage the distal microcirculation. Similarly, thrombin generated on the platelet surface and serotonin and thromboxane A2released from platelets may further compromise the distal circulation. Both human and animal model data support a role for abciximab in decreasing microcirculatory damage after vascular injury (2,26), which may be explained at least in part by the smaller platelet thrombi we observed at the site of stent insertion. A protective effect of abciximab on the microcirculation may also contribute to abciximab’s long-term mortality advantage (6)if microcirculatory damage ultimately results in patchy fibrosis, electrical instability and increased risk of sudden death. Thus, to understand the full effects of stent insertion it is important to consider not only large-vessel obstruction at the site of injury, but also the distal effects on the microcirculation initiated by the original injury.
We wish to thank Dr. Carol Bodian (Mount Sinai School of Medicine) for her assistance with the statistical analysis of the data; Lori A. Krueger and Dr. Alan Michelson (University of Massachusetts) for their assistance with the design of the study; Veronica Gulle (Mount Sinai School of Medicine) for assistance with the immunohistochemistry; and Suzanne Rivera for outstanding secretarial assistance.
☆ Dr. Smyth is presently affiliated with the Department of Medicine, University of North Carolina at Chapel Hill School of Medicine, Chapel Hill, North Carolina. Dr. Coller is presently affiliated with the Laboratory of Blood and Vascular Biology, Rockefeller University, New York, New York.
Supported in part by grants 19278 and 54469 from the National Heart, Lung and Blood Institute and Centocor, Inc., Malvern, Pennsylvania. Dr. Coller is an inventor of abciximab and in accord with federal law and the policies of the Research Foundation of the State University of New York shares in payments made to the Foundation for the sales of abciximab.
- internal elastic lamina
- plasminogen activating inhibitor-1
- phosphate-buffered saline
- P-selectin glycoprotein ligand-1
- rapid platelet function assay
- tissue factor
- tissue plasminogen activator
- Received September 18, 2001.
- Revision received April 8, 2002.
- Accepted April 18, 2002.
- American College of Cardiology Foundation
- Neumann F.J.,
- Blasini R.,
- Schmitt C.,
- et al.
- Marso S.P.,
- Lincoff A.M.,
- Ellis S.G.,
- et al.
- Coller B.S.
- Smyth S.S.,
- Reis E.D.,
- Zhang W.,
- Fallon J.T.,
- Gordon R.E.,
- Coller B.S.
- Neumann F.J.,
- Zohlnhofer D.,
- Fakhoury L.,
- Ott I.,
- Gawaz M.,
- Schomig A.
- Makkar R.R.,
- Litvack F.,
- Eigler N.L.,
- et al.
- Krupski W.C.,
- Bass A.,
- Kelly A.B.,
- Marzec U.M.,
- Hanson S.R.,
- Harker L.A.
- Lewis D.A.,
- Keelan P.C.,
- Camrud A.R.,
- et al.
- Furman M.I.,
- Kereiakes D.J.,
- Krueger L.A.,
- et al.
- Smith J.W.,
- Steinhubl S.R.,
- Lincoff A.M.,
- et al.
- Schwartz R.S.,
- Huber K.C.,
- Murphy J.G.,
- et al.
- Carter A.J.,
- Laird J.R.,
- Farb A.,
- Kufs W.,
- Wortham D.C.,
- Virmani R.
- Giesen P.L.,
- Rauch U.,
- Bohrmann B.,
- et al.
- Rauch U.,
- Bonderman D.,
- Bohrmann B.,
- et al.
- Simon D.I.,
- Chen Z.,
- Xu H.,
- et al.
- Kupatt C.,
- Habazettl H.,
- Hanusch P.,
- et al.