Jump to
Abstract
Among the many effects of heparin independent of its effects on coagulation are inhibition of vascular smooth muscle cell proliferation and regulation of leukocyte–blood vessel interactions. The potential link between these effects was examined in an animal model of vascular injury rich in inflammatory cells: the placement of endovascular metal stents in rabbit iliac arteries. Monocyte adhesion stimulated by early focal thrombus was maximal after 3 days, with infiltrating monocytes and intimal cell proliferation maximal after 7 days. Tissue monocyte number dictated cell proliferation at each time point (R2=.92, P<.0001). Heparin reduced both early monocyte adhesion as well as monocyte infiltration within the neointima 7 and 14 days after stent placement. Reductions in adherent and tissue monocytes were commensurate with reductions in intimal cell proliferation and intimal thickening. At 14 days, heparin's inhibition of mononuclear cell adhesion was correlated with its suppression of intimal thickening (R2=.82, P<.0001). Monocytes have been hypothesized to serve as markers, initiators, and promoters of arterial occlusive diseases. Heparin's ability to inhibit mononuclear cell adhesion and penetration and reduce neointimal size and cell proliferation after vascular injury may further implicate monocytes in the pathogenesis of neointimal hyperplasia after mechanical arterial injury.
- Received January 11, 1995.
- Revision received March 14, 1996.
Among the many effects of heparin independent of those on coagulation include its inhibition of VSMC proliferation1 2 and regulation of leukocyte–blood vessel interactions.3 4 5 6 In vivo mechanisms that underlie heparin's actions remain incompletely defined. There is also increasing evidence that monocytes control VSMC proliferation and migration, lipid metabolism, and inflammation within the blood vessel wall. Circulating monocytes are among the earliest cells recruited into experimentally induced vascular lesions in animals7 8 9 10 11 12 13 14 and spontaneous atherosclerosis in human arteries.15 16 As a result, monocytes have been hypothesized to serve as markers, initiators, and promoters of arterial occlusive diseases. We wondered whether the effects of heparin on monocyte adhesion and neointimal hyperplasia in vivo after vascular injury might be linked. To answer this question, we examined the vascular biology of endovascular stent placement in rabbit iliac arteries, because this model of vascular injury is rich in inflammatory cells and responsive to heparin in a dose-dependent fashion.17 18 19 20 21 22
We report herein that monocytes adhere to the luminal surface of stent-injured arteries and then penetrate the healing arterial wall, with resultant neointimal hyperplasia. The number of monocytes adherent to and infiltrating injured arteries was reduced by more than 70% by continuous heparin administration and in direct concert with heparin's inhibition of cell proliferation and intimal thickening. These data demonstrate for the first time that heparin inhibits monocyte adhesion and infiltration into injured large arteries. They also suggest that modulation of monocyte adhesion may contribute to heparin's inhibition of intimal thickening after vascular injury. Inhibition of monocyte adhesion and migration into mechanically injured arteries may alter vascular repair, limit the extent of neointimal hyperplasia, and perhaps provide novel approaches to the control of accelerated arteriopathies that follow vascular intervention or injury.
Methods
Surgical Procedure and Tissue Processing
New Zealand White rabbits (Millbrook Farm Breeding Labs, Amherst, Mass) weighing 3 to 4 kg were housed individually in steel mesh cages and fed rabbit chow and water ad libitum. To reduce the significant incidence of early occlusive thrombosis that is seen in this model,17 18 aspirin (Sigma Chemical Co, 0.07 mg/mL) was added to the drinking water starting 1 day before surgery to achieve an approximate dose of 5 mg·kg−1·d−1. Under anesthesia with ketamine (Aveco Co) 35 mg/kg and sodium pentobarbital (Nembutal, Abbott Laboratories) 4 mg/kg IV, both femoral arteries were exposed and ligated. The iliac artery endothelium was denuded by use of a 3F balloon embolectomy catheter (Baxter HealthCare Corp) passed retrogradely via the arteriotomy into the abdominal aorta and withdrawn in the inflated state three times. After balloon denudation an endovascular metal stent mounted coaxially on a 3-mm angioplasty balloon (Advanced Cardiovascular Systems) was passed retrogradely via the arteriotomy into each iliac artery and expanded with a 15-second inflation at a pressure of 8 to 10 atm. All rabbits received a single intravenous bolus of standard anticoagulant heparin (100 U/kg, Elkins-Sinn Inc) at the time of stent deployment to limit the stent thrombosis that has been reported in this model.17 18
Two major series of experiments were performed, including one that examined the time courses of heparin's effects on monocyte recruitment and neointimal hyperplasia after arterial injury, and the other that examined how varying the dose and mode of heparin delivery affected these actions. This protocol allowed us to explore heparin's actions on monocyte recruitment and intimal thickening, with systemic and locally controlled means of heparin delivery used to ensure accurate and reproducible local pharmacokinetics directly at sites of injury and avoid confounding variations in systemic drug exposure and effects on circulating monocytes. For the dose/mode-of-delivery studies, heparin (Choay heparin 1453, 12 000 to 18 000 Da, USP 160 U/mg, Choay Institute, Sanofi Inc) was administered in a sustained fashion via one of three routes: perivascular, intravenous, or intra-arterial. Controlled local perivascular release was provided via ethylene vinyl acetate copolymer matrices,18 23 33% loaded with native heparin (8 arteries). The matrices were fabricated as previously described18 23 to provide either continuous heparin delivery for 2 weeks or a burst of drug release early after implantation with little thereafter and were incubated for 48 hours at 37°C before implantation to achieve immediate steady rates of heparin delivery in vivo. Subcutaneously implanted osmotic minipumps (Alza Corp) provided continuous femoral intravenous delivery (3 arteries, 0.3 mg·kg−1·h−1). Metal stents to which heparin had been ionically bound provided intra-arterial delivery of the drug (8 arteries).18
For each method of heparin delivery, in vitro and in vivo release kinetics studies have been reported previously.18 23 Four stented arteries in animals that received no postprocedure heparin served as controls. Four normal rabbit iliac arteries were also examined. aPTT was measured with a desktop analyzer (Ciba-Corning Diagnostics Corp) at the time of the procedure and 7 and 14 days later. Arteries were harvested 14 days after surgery. After anesthesia was induced with intravenous sodium pentobarbital, the inferior vena cava opened, and perfusion performed with lactated Ringer's solution via left ventricular puncture, both iliac arteries were excised and immersed in Carnoy's fixative (60% methanol, 30% chloroform, 10% glacial acetic acid, vol/vol/vol). After their proximal and distal ends were marked, nonthrombosed arterial segments with metal stents were oriented along flow lines, embedded in DDK-plast (Delaware Diamond Knives, Inc), and cut with a tungsten carbide knife (Delaware Diamond Knives, Inc). Five-micron-thick cross sections were taken from three sites along each stent, including one from each end and the middle. Segments from normal arteries were embedded in paraffin and also cut into 5-μm-thick cross sections.
For time-course studies heparin was administered via femoral intravenous infusion (0.3 mg·kg−1·h−1) for 3 (4 arteries), 7 (6 arteries), or 14 (4 arteries) days. Control arteries with implanted endovascular stents but no continuous heparin were also examined after 3 (4 arteries), 7 (4 arteries), or 14 (6 arteries) days. To allow for immunocytochemical staining and examination, all animals received BrdU (50 mg/kg IV, Sigma Chemical Co) 1 hour before they were killed. The arteries were harvested as described above, fixed in Carnoy's solution or 10% p-formaldehyde, and embedded at −20°C in methyl methacrylate mixed with n-butyl methacrylate (Sigma Co).24
All animal care and procedures were in accordance with the guidelines of the American Association for Accreditation of Laboratory Animal Care and the National Institutes of Health.
Histological Analysis
Tissue and cell structures were highlighted in histological sections by staining with Verhoeff's tissue elastin stain or hematoxylin and eosin. Neointimal hyperplasia was quantified as intimal cross-sectional area, as measured in elastin-stained sections by using computer-assisted digital planimetry. Postmortem contraction of the luminal surface was prevented by the “scaffolding” of the metal stent, thus keeping the outer diameter of the specimen at a fixed and constant diameter. This procedure also provided a flat, taut luminal surface whereon leukocytes adherent to the surface could be individually examined. The nuclear morphology of these cells appeared to be either monocytoid or polymorphonuclear. To minimize sampling error, three equally spaced cross sections along the length of each stent were analyzed and the results averaged. In each cross section, the total number of adherent cells was counted and adherent cell nuclear morphology characterized under ×600 magnification.
Rabbit macrophages were identified immunocytochemically with a species-specific antibody that labels tissue macrophages but not circulating or adherent monocytes25 (monoclonal mouse RAM 11 IgG, DAKO Co). The number of proliferating cells was quantified immunocytochemically on the basis of their uptake of BrdU (anti-BrdU, DAKO Co). Standard immunocytochemical protocols were followed.26 Sections were incubated with the primary antibody and then with a biotinylated species-specific secondary antibody (Vector Laboratories Inc), and cells were “stained” by avidin-biotin peroxidase or avidin-biotin–alkaline phosphatase (kits from Vector Laboratories Inc) followed by 3,3-diaminobenzidine (Sigma) or alkaline phosphatase (Vector Laboratories Inc). In these sections, overall intimal cell density was calculated by dividing the number of nuclei by the intimal area. Numbers of immunologically identified monocytes/macrophages or proliferating cells were then counted and the densities of these cell types calculated.
Statistics
All data are presented as mean±SE. Comparisons between treatment groups used an unpaired t test. Values of P<.05 were considered significant.
Results
To explore whether a reduction in monocyte adhesion to and infiltration of mechanically injured arteries might underlie heparin's inhibition of intimal thickening and cell proliferation, we identified luminally adherent cells of monocyte lineage by nuclear morphology and infiltrating cells of monocyte lineage by immunocytochemical techniques in both untreated and heparin-treated arteries.
Endovascular stents induced an intimal lesion initially composed of thrombus and soon thereafter replaced by organized SMC hyperplasia.27 Monocytes played a critical role in this transformation. At baseline, no monocytes appeared to have adhered to rabbit iliac arteries; 3 days after endovascular stent implantation, however, 175.3±29.1 monocytes were noted on the luminal surface of each stented cross section (Fig 1A⇓), whereas <0.1% of cells within the vessel wall were of monocyte lineage (Table⇓). After 4 more days, much of the intima had changed to primarily an SMC composition. Although the number of adherent monocytes had dropped by 48.5% at this time, the number of immunocytochemically identified infiltrating monocytes had increased >10-fold. At this time cell proliferation within the vessel wall was maximal: 1.60±0.62% of all cells were proliferating. Seven days later (14 days after initial injury), the number of adherent cells had fallen by an additional 68.9%, to ≈10% of initial levels. Infiltrating monocyte density had fallen by 43%, and intimal proliferation had decreased as well. Progressive intimal thickening was not seen between 7 and 14 days (Table⇓).
Early monocyte adhesion: Photomicrographs of rabbit iliac arteries 3 days after balloon withdrawal denudation followed by endovascular stent implantation, stained with Verhoeff's tissue elastin stain. In an untreated artery (A) numerous adherent leukocytes, primarily monocytes, are at the luminal surface. With continuous intravenous heparin treatment (B) there is a 90% reduction in the number of adherent cells. Arrows mark the internal elastic lamina. Original magnification ×400.
Temporal Effects of Heparin on Vascular Repair After Balloon Denudation and Stent Implantation in Rabbit Arteries
Heparin treatment reduced the number of cells adherent to the luminal surface as early as 3 days after injury (Table⇑, Fig 1B⇑), leading to markedly reduced numbers of infiltrating RAM 11–positive macrophages 7 and 14 days after injury (Fig 2⇓). Similar reductions in intimal cell proliferation and area accompanied this inhibition, directly commensurate at each point with heparin's inhibition of monocyte infiltration (Table⇑). For heparin-treated and control groups, intimal cell proliferation was positively correlated with tissue RAM 11–positive macrophage density (R2=.916, P<.0001).
Macrophage infiltration: Photomicrographs of rabbit iliac arteries 14 days after balloon withdrawal denudation followed by endovascular stent implantation, stained with a cell-specific antibody that identifies rabbit macrophages (monoclonal mouse RAM 11 IgG). Cells identified immunocytochemically as macrophages are stained red, and straight arrows mark the internal elastic lamina. In untreated arteries (A) numerous tissue macrophages are apparent within the neointima, and adherent monocytes (not staining for macrophage marker RAM 11) are present on the luminal surface (curved arrows). With continuous intravenous heparin treatment (B) neointimal thickening is less, the density of RAM 11–positive tissue macrophages is markedly reduced, and adherent monocytes are not seen. Original magnification ×600.
To explore a possible link between heparin's actions on monocyte recruitment and intimal thickening, we examined the effects of varying doses of heparin on these aspects of vascular repair. Systemic and locally controlled means of heparin delivery were used to ensure accurate and reproducible local pharmacokinetics at sites of injury. This procedure avoided possible confounding variations in systemic drug exposure and their effects on circulating monocytes and focused the investigation at the sites of injury. Adherent leukocytes were examined along the circumference of stented arteries that had been treated with various doses of heparin. Nuclear morphology was well preserved in these arteries and allowed differentiation of leukocyte classes. Adherent cells were neither injured nor dying and their nuclei not pyknotic. Most adherent cells had a monocytic nuclear morphology (Fig 3⇓).
Late monocyte adhesion: A, Photomicrograph of a rabbit iliac artery cross section 14 days after balloon withdrawal denudation followed by endovascular stent implantation, stained with hematoxylin and eosin (original magnification ×320). Eight cells adherent to the luminal surface are numbered sequentially 1-8 (B-F). Sequential photomicrographs of each adherent cell (original magnification ×1600) demonstrate nuclear features suggestive of monocytes.
Whereas normal rabbit iliac arteries were free of adherent cells, 14 days after stent placement 31.1±2.8 mononuclear cells17 lined the luminal surface of each arterial cross section. Continuous heparin delivery reduced the number of these cells 3-fold compared with stent-bearing controls, independent of the mode of heparin administration (9.9±0.6 cells, P<.002 for intravenous delivery; 12.2±1.5 cells, P<.003 for perivascular delivery). Heparin that was given only for the first 3 days after surgery was less effective than continuous heparin delivery in inhibiting mononuclear cell adhesion (17.0±1.0 cells, P<.0005 compared with untreated stent-bearing controls). Low-dose heparin release from the endovascular stent at the luminal surface reduced the number of adherent mononuclear cells only slightly (20.4±3.0 cells, P<.04 compared with stent-bearing controls).
Neointimal hyperplasia was also reduced to varying degrees by heparin.18 Heparin that was delivered continuously via intravenous and perivascular routes was equally effective in inhibiting neointimal hyperplasia, reducing the intimal area from 1.09±0.11 mm2 (untreated controls) to 0.34±0.03 mm2 (P<.002) or 0.48±0.10 mm2 (P<.006), respectively. Perivascular heparin delivery restricted to the early postoperative period was less effective (intimal area, 0.75±0.09 mm2; P<.03), and stent-released intra-arterial heparin was ineffective (intimal area, 1.01±0.15 mm2; P=NS). A linear correlation between the number of adherent monocytes and intimal mass was noted, whether the analysis was performed for individual arteries (R2=.82, P<.0001; Fig 4A⇓) or the mean values of different heparin treatment groups (R2=0.92, P<.004; Fig 4B⇓).
Monocyte adhesion and intimal growth: cross-sectional intimal area (y axis) and adherent mononuclear cell number (x axis) 14 days after balloon withdrawal denudation followed by endovascular stent implantation in rabbit iliac arteries. A, Individual arteries; B, treatment group means. Groups received heparin via intravenous infusion (IV), from the stents themselves (Stent), or from matrices adjacent to the adventitia of stent-bearing arteries (PV). Matrices were incubated in vitro for 2 days before placement and provided heparin delivery in vivo either early (PV, 3 days) or continuously (PV, 14 days).
Anticoagulation
For animals that received different doses of heparin, aPTTs were measured just before surgery and 7 and 14 days later.18 Continuous intravenous heparin infusion prolonged all aPTTs to at least twice control values at 7 and 14 days. Perivascular or stent-released heparin did not prolong aPTT in any animal at any time. No incisional or other bleeding was observed in any animal.
Discussion
We have demonstrated for the first time that heparin inhibits monocyte adhesion to and infiltration of injured large arteries. At each point in lesion development, the effects of heparin on monocyte recruitment precisely matched the suppression of neointimal hyperplasia and intimal cell proliferation. The correlation between adherent and infiltrating mononuclear cells and neointimal thickening and proliferation after chronic and severe arterial trauma in this model supports a role for infiltrating monocytes in the pathogenesis of neointimal hyperplasia after arterial injury and for the modulation of monocyte recruitment as a central aspect of heparin's actions in vivo.
A model of vascular injury with endovascular stent placement was chosen for several reasons. First, intimal thickening after this type of injury has been reported to contain numerous inflammatory cells and to provoke more prolonged intimal proliferation than simple balloon injury.17 18 20 27 28 29 Second, we have reported that the means of heparin delivery can be altered in a controlled local fashion in this model, with resultant alterations in vascular repair.18 Histological evaluation of stented arteries provides fixed outer diameters and no luminal collapse and allows the ready enumeration of inflammatory cells adherent to the vessel lumen as well as infiltrating cells.
Monocyte Attraction and Adhesion to Injured Vessels
Monocytes have been found in early and late stages of clinical15 16 and experimental7 8 9 10 11 12 13 14 atherosclerosis. Pathological examination of human coronary arteries after angioplasty or stent placement has revealed leukocytes adherent to and within vessel walls.22 Recently, the activation status of circulating leukocytes, including monocytes, at the time of angioplasty has been reported to predict later restenosis.30
Recruitment of circulating monocytes to the arterial wall may depend on EC expression of specific cell adhesion molecules31 32 33 34 ; elaboration of chemotactic peptides, such as monocyte chemoattractant protein-1,35 by vascular cells36 37 38 ; and creation of regions of altered shear stress where circulating monocytes may be exposed to the vessel surface.39 Both balloon catheter inflation within an artery and expansion of an endovascular stent denude the endothelium, disrupt medial architecture, and cause medial cell death. The chronic presence of an indwelling metal stent might additionally alter the extent of endothelial regrowth, change local flow patterns promoting monocyte–vessel wall contact, and induce a foreign-body reaction.
Our data show that the numbers of adherent and infiltrating cells of monocyte/macrophage lineage in stent-injured arteries increase with the extent of neointimal hyperplasia. Adherent cells may migrate to the vessel wall and promote migration and proliferation of SMCs via elaboration of factors chemotactic and mitogenic for SMCs.40 41 42 43 44 Monocyte-derived cells might also produce reactive oxygen species potentially injurious to vascular wall cells45 46 or basement membrane–degrading enzymes47 48 that release growth-modulating compounds from cell surfaces or extracellular stores.49 50
Effects of Heparin
Heparin is the archetypal modulator of vascular repair and an important mediator of leukocyte biology. Heparin inhibits SMC proliferation in vitro51 and is exceptionally effective at limiting neointimal hyperplasia after experimental arterial injury.1 Even when devoid of antithrombotic activity, heparin is a potent regulator of the vascular response to injury2 and inhibits SMC proliferation in tissue culture by markedly and rapidly inhibiting DNA synthesis in growth-arrested SMCs released from the G0 block and limiting transcription of genes necessary for cell passage from G0 through G1 and into the S phase.52 53 54 The biologic relevance of heparin as a participant in vascular homeostasis is confirmed by evidence that ECs and SMCs produce or moderate production or release of heparin-like molecules that regulate VSMC growth.55 56
Heparin's inhibition of mononuclear cell adhesion to damaged arteries as early as 3 days after injury provides strong evidence for a primary effect of heparin on mononuclear cell–vessel wall interactions rather than an indirect effect secondary to heparin's reduction of thrombosis or neointimal hyperplasia. Although such an effect on leukocyte adhesion to large arteries has not been previously reported, effects of heparin on leukocyte adhesion to smaller vessels were first noted by Cohen et al,3 who invoked antithrombin mechanisms to explain heparin's inhibition of delayed-type hypersensitivity reactions, and later by Sy and coworkers,5 who demonstrated a similar efficacy for heparin that lacked antithrombin activity.
Recent recognition that EC proteoglycans may control early events in leukocyte adhesion57 and activation and binding of leukocyte integrin molecules to vessel wall ligands31 58 59 offers a potential mechanism whereby glycosaminoglycans such as heparin might directly affect leukocyte-endothelium adhesion in tissue culture60 or in vivo.6 Alternatively, the biologic activities of chemotactic factors, such as platelet factor 4 or macrophage inflammatory protein 1, that contain heparin-binding domains could be altered by heparin.61 62 Finally, membrane surface charge and cation availability, which help govern leukocyte adhesion,51 might be affected by the high anionic charge of EC-bound heparin.63
Interestingly, our data demonstrate a reduction in adhesion as early as 3 days after injury, long before endothelial regrowth has occurred, suggesting an endothelium-independent mechanism. Monocyte attachment to the luminal surface via integrin-independent mechanisms offers alternative sites for heparin's inhibitory effects. Such mechanisms of leukocyte adhesion that have been defined in vitro include scavenger receptor–mediated binding64 or vitronectin-dependent binding of UPAR on the cell surface.65 66 67 68 Heparin might, for example, inhibit UPAR-vitronectin interactions by binding to vitronectin in the extracellular matrix,69 attenuating UPA or UPAR expression,70 71 72 or interfering with the binding of UPA to UPAR.73
Seven and 14 days after injury, heparin reduced the number of infiltrating macrophages as well as the degree of monocyte adhesion. This temporal sequence of events and similar effects on cell proliferation and intimal thickening suggest that heparin's primary effect on monocyte action lies in its inhibition of adhesion, not interference with mononuclear cell function within the vessel wall. Furthermore, our data demonstrate that local heparin release without detectable systemic dosing affected monocyte adhesion and neointimal hyperplasia in a fashion identical to that seen with systemic treatment and support the notion of a direct, local effect on the vessel wall rather than a systemic effect on circulating cells. These observations are consistent with the absence of any transferable effect on inflammation conveyed by leukocytes from heparin-treated mice when transplanted into non–heparin treated animals.5
Implications
Our data demonstrate that mononuclear cells adhere to and infiltrate mechanically injured arteries in a lipid-independent model of deep vascular injury and that heparin limits the number of these cells just as it inhibits neointimal thickening and cell proliferation. Participation of mononuclear cells in the vascular response to injury and the regulation of these cells by heparin may help explain heparin's in vivo antiproliferative activity.1 The persistence of a direct temporal relationship between heparin's effects on monocytes and neointimal hyperplasia makes it unlikely that these are parallel but unrelated effects on two independent aspects of the vascular healing process. Exploration of the mechanisms that underlie the linkage between mononuclear cell adhesion and intimal thickening in injured arteries and of heparin's effects on this process may be important in understanding vascular repair and perhaps lead to novel approaches for prevention of restenosis.
Selected Abbreviations and Acronyms
| aPTT(s) | = | activated partial thromboplastin time(s) |
| EC(s) | = | endothelial cell(s) |
| UPA(R) | = | urokinase-type plasminogen activator (receptor) |
| (V)SMC(s) | = | (vascular) smooth muscle cell(s) |
Acknowledgments
This work was supported in part by grants from the American Heart Association, National Center (95004400 to Dr Rogers); and the National Institutes of Health, Bethesda, Md (HL 17747 and GM/HL 49039), the Burroughs Wellcome Fund for Experimental Therapeutics, and the Whitaker Foundation (all to Dr Edelman). We are grateful to Dr Marla Steinbeck for critical review of the manuscript and to William Appel and Philip Seifert for technical assistance.
References
This Issue
Jump to
Article Tools
Share this Article
- Monocyte Recruitment and Neointimal Hyperplasia in Rabbits
