Vascular Injury, Repair, and Restenosis After Percutaneous Transluminal Angioplasty in the Atherosclerotic Rabbit
Background Several nonatherosclerotic animal models of restenosis exist and are used for the evaluation of the vascular response to angioplasty-induced injury. However, few studies have evaluated the response of an atherosclerotic vessel to angioplasty. The present study examined the radiographic, histological, immunohistochemical, and morphometric responses over time of atherosclerotic rabbit femoral arteries after percutaneous transluminal angioplasty (PTA).
Methods and Results Rabbits (n=94) underwent arterial desiccation and were fed a hypercholesterolemic diet for 3 weeks, and then PTA was performed. Arteries were obtained before PTA and 1, 3, 5, 7, 14, and 28 days after PTA. PTA caused radial stretching of the artery, medial compression, intramural hemorrhage, injury to normal arterial segments, and dissection within the intima and media. Thrombus filled and cellular accumulation repaired the dissection. Peak smooth muscle cell and macrophage DNA synthesis was noted at 3 to 5 days after angioplasty, generally at the dissection but also in normal sections of the artery. Adventitial injury and subsequent adventitial cellular proliferation and collagen production were observed. A rapid decrease in the radiographic minimal luminal diameter was noted at 3 days, resulting from vascular recoil or thrombus filling the dissection. At 7 to 14 days, only 24% to 33% of the luminal loss was accounted for by an increase in the intimal area, and 22% to 28% of the intima was neointima.
Conclusions Restenosis in an atherosclerotic artery results from a variable combination of intimal proliferation, vascular remodeling/wound contraction, and recoil of the normal section of the artery. The variability of an atherosclerotic artery to PTA injury results from variable dissection, thrombus formation, and cellular response to injury as well as variable scar contraction and elastic recoil.
Although new advances in percutaneous transluminal coronary angioplasty catheter design and operator expertise have reduced the procedural complication rate, the angiographic restenosis rate remains at 35% to 40%.1 2 No currently devised pharmacological approach has been shown to reduce the rate of restenosis, and an incomplete understanding of the pathophysiology of restenosis limits all approaches to therapy. At the present time, there are at least seven animal models for the investigation of restenosis,3 and these can be divided into single- or double-injury models, depending on the number of steps performed to obtain an atherosclerotic or restenotic lesion. Balloon endothelial denudation and medial layer damage with or without a subsequent high-cholesterol diet constitutes the single-injury model in which there is de novo intimal growth in a previously normal artery. Interventions designed to evaluate the capacity for a device or medication to reduce the resulting intimal proliferation are performed at initial injury. The single-step approach has been evaluated in rat carotid and aortic arteries, rabbit iliac and ear4 arteries, and porcine coronary arteries (see Reference 3 for review). An advantage of such single-injury models is the homogeneity of the resulting lesion, which consists largely of smooth muscle cells (SMCs). Disadvantages include the lack of a preexisting atherosclerotic lesion and the relatively small degree of intimal proliferation in small animal models. The absence of lesion complexity may explain in part why single-injury models of restenosis respond to various antiproliferative regimens, whereas similar strategies in humans have been without effect.
The double-injury model involves initial injury of the endothelial and/or medial layer by a balloon,5 electrical current,6 or air desiccation.7 The animal is generally placed on a high cholesterol diet until angioplasty is performed. At the time of the angioplasty, medications are administered and the subsequent growth of neointima is evaluated as the measure of restenosis. This approach has been evaluated in the rabbit aortic, iliac, and femoral arteries and porcine coronary and iliac arteries.3 A disadvantage of the double-injury model is the heterogeneous nature of the second lesion and thus the difficulty in differentiating the original from the restenotic tissue. Nonetheless, the two-step approach creates greater lesion complexity and thereby more closely mimics human conditions. The temporal response to injury in atherosclerotic vessels has not been evaluated. Accordingly, in this study we have characterized the response to vessel injury after angioplasty to elucidate the resulting variability of injury and repair mechanism. The studies were performed in atherosclerotic rabbit femoral arteries obtained before and up to 4 weeks after percutaneous transluminal angioplasty.
All procedures followed were in accordance with Indiana University Animal Resource Committee guidelines. New Zealand White rabbits weighing between 8 and 10 pounds each were used for this study. Endothelial denudation and medial damage of a 1-cm section of both deep femoral arteries was achieved by air desiccation in a procedure described previously.7 A metal clip was placed in the periadventitia near the center of the desiccated segment to aid in subsequent balloon placement as well as histological and morphometric evaluation. After desiccation, the rabbits were placed on a 150 g/d rabbit chow diet containing 2% cholesterol and 6% peanut oil (Dyets Inc) for 3 weeks until arteriography and angioplasty.
The technique of angiography and angioplasty used was a modification of a previously published technique.8 After placement of a 5F sheath into the aorta via the right carotid artery, heparin 150 U/kg body wt was administered and a multipurpose catheter was placed at the level of the aortic bifurcation. After administration of 2 mL 1% lidocaine to reduce arterial spasm, angiography of the femoral arteries was performed with a straight, multipurpose catheter (Medi-Tech). Angioplasty, performed in the atherosclerotic segment of arteries with ≥20% but <100% stenosis, was achieved by use of a standard angioplasty balloon (Advanced Cardiovascular Systems), ≤0.5 mm larger than the proximal, normal arterial segment, which was inflated for 1 minute at 10 atm pressure. The balloon was deflated for 1 minute, and the cycle was repeated for a total of three inflations. If the lesion showed >50% improvement at subsequent angiography, the procedure was considered successful. Thereafter, the contralateral artery underwent angioplasty. The sheath was removed and the carotid artery and skin sutured. The rabbit was administered cefazolin 2 g IM and placed on a 150 g/d normal rabbit chow diet until euthanasia.
Two groups of rabbits underwent only angiography, and these served as controls. One group, the original lesion group, was euthanatized 3 weeks after the original air desiccation injury. The other, a sham angioplasty group, underwent original injury and subsequent angiography but did not undergo angioplasty. This group was placed on a normal diet for 14 days until euthanasia 5 weeks after the original air desiccation injury.
The rabbits were euthanatized at six time periods after angioplasty: 1, 3, 5, 7, 14, and 28 days. Most rabbits were given the thymidine analogue bromodeoxyuridine (BrdU).6 Briefly, BrdU 100 mg/kg and deoxycytidine 75 mg/kg were administered in a subcutaneous depot 18 hours before euthanasia in addition to administration of 30 mg/kg IM BrdU and 25 mg/kg IM deoxycytidine. A second intramuscular injection of BrdU and deoxycytidine was administered 12 hours before euthanasia to maintain constant serum concentrations. In principle, all cells entering the S phase during the last 18 hours would stain positive with antibodies for BrdU. Deoxycytidine was given because an administration of BrdU in the absence of deoxycytidine causes an inhibition in the internal production of thymidine and cytidine, resulting in a decrease in DNA synthesis (M. Oberhoff, personal communication, 1995). At the time of euthanasia, the rabbit was anesthetized with ketamine and xylazine and a 16-gauge angiocath was placed in the abdominal aorta via a transverse abdominal incision. Two milliliters of 1% lidocaine was injected into the aorta, and a final angiogram was obtained. The venous effluent was collected via the inferior vena cava. The aorta was infused with zinc formalin solution (Anatech) at a pressure of 100 mm Hg, and the rabbit was euthanatized with an overdose of pentothal.
The arteries were obtained en bloc and placed in a zinc formalin bath. The presence of zinc reduces formalin cross-linking, thereby enhancing antibody staining. The overlying skeletal muscle was dissected away from the artery, and the arterial segment was isolated and cut into pieces 3 to 4 mm in length (5 to 8 segments per artery). The segment nearest the metal clip was identified and noted. The distal end of each segment was marked with a short stripe of india ink, numbered, and placed in an individual canister for embedding and sectioning.
Histological and Immunocytochemical Staining
The arterial sections, embedded in paraffin and sectioned at a thickness of 5 to 6 μm, were stained with hematoxylin and eosin and a combined Gomori’s aldehyde fuchsin with a trichrome counterstain. Other unstained sections were deparaffinized and used for immunohistochemical analysis by the avidin-biotin-peroxidase complex method.9 Primary antisera were used at the following dilutions: factor VIII (Dako Corp) 1:100 and fibrinogen/fibrin10 (Cappel Labs) 1:1200. HHF-3511 (Enzo Laboratories), a monoclonal antibody specific for vascular α-actin at a dilution of 1:400, was used to evaluate for the presence of vascular SMCs. RAM-11 (gift from Dr Allen Gown, University of Washington, Seattle, Wash), a monoclonal antibody specific for rabbit macrophages, was diluted to 1:6500.12 The presence of endothelial cells was determined by antibody staining to von Willebrand’s factor (factor VIII) at a dilution of 1:200.13 Staining for BrdU incorporated into DNA used a monoclonal antibromodeoxyuridine antibody (Amersham) at a dilution of 1:100, counterstained with Gill’s hematoxylin.
Selected arterial segments (n=24) underwent resectioning and double-antibody staining to evaluate the relative proliferative indexes of cells staining positive for the presence of RAM-11 or HHF-35. Sections (5-μm) were mounted, deparaffinized, and stained for the presence of either RAM-11 and BrdU or HHF-35 and BrdU according to protocols included in the Dako LSAB 2-AP and PAP kits.
The arterial cross sections were evaluated by a pathologist blinded to the arterial group. The segments were evaluated for the following: (1) internal elastic lamina (IEL) fracture; (2) intimal fracture; (3) inflammation, defined as increased neutrophil infiltration (compared with nonangioplastied arteries); (4) medial compression, defined as a thinned medial layer in relation to adjacent areas or segments; (5) medial dissection; and (6) medial necrosis defined as a reduction of SMCs and the presence of pycnotic nuclei. The presence of a characteristic was noted, and if present in any segment, the characteristic was considered present in that artery. The pathologist was then unblinded, and the arterial segments were characterized further to determine the effect of time after angioplasty on the histological characteristics.
To determine the cellular proliferation index, arterial segments were photographed at ×40 to ×100 magnification and all nuclei were counted in the intimal and medial layers. A proliferation index was obtained by dividing the number of BrdU-positive nuclei by the total nuclei in that cell layer. The presence of BrdU staining and the location of positive cells and their focality within the arterial layer were noted. Because it was difficult to determine the external border of the adventitia, adventitial BrdU uptake was reported as either focal or generalized. Cellular proliferation indexes were also determined for populations staining positive for BrdU and either HHF-35 or RAM-11 within the intimal and medial layers.
The luminal, (neo)intimal, and medial areas were evaluated by morphometric analysis of arterial cross sections obtained 1, 2, and 4 weeks after angioplasty and the cross sections of the sham angioplasty and original groups. Morphometric analysis was done on arterial cross sections imaged on a Macintosh computer with an NIH image software package. All arterial segments exhibiting an intimal layer cut in an axis perpendicular to the long axis of the artery and exhibiting no fixation or sectioning-related artifacts were evaluated. The endoluminal border, the circumference bounded by the IEL, and the external elastic lamina (EEL) were traced by hand, and the luminal, intimal, and medial areas were calculated. Areas of mural thrombus were excluded from measurement. To qualitatively evaluate the longitudinal variability of these parameters within given arteries, four arteries were sectioned at 200-μm distances and each section was evaluated for the aforementioned areas.
Radiograms from all time points were evaluated by measuring the minimal luminal diameter (MLD) of the greatest stenosis. If arterial spasm or aneurysm formation was noted, the radiogram was excluded from further evaluation.
The segment exhibiting the greatest degree of intimal area, regardless of the luminal area, was considered the representative sample for that artery and was used for future comparisons between groups. For determination of the proliferation index, the results in all segments within a single artery were averaged and the artery was treated as an independent observation. Analysis of the morphometrically derived luminal, intimal, and medial areas from the different groups was compared by use of a one-way ANOVA and a protected least-significant-difference test for multiple comparisons. In addition, the EEL circumference was compared between groups by ANOVA and a protected least-significant-difference test. The protected least significant difference identifies differences in groups (as determined by ANOVA) while protecting for the performance of multiple comparisons, with the resultant increased chance of finding a nonmeaningful difference. Data are expressed as mean±SEM. A value of P<.05 was considered significant.
In total, 94 arteries were evaluated. The average cholesterol level after 3 weeks of diet was 1535±76 mg/dL. Despite discontinuation of the diet at the time of percutaneous transluminal angioplasty, serum cholesterol levels remained elevated over the duration of the study, with mean levels between 1091 and 1945 mg/dL. For example, in the 28-day group, the mean cholesterol level was 1400±214 mg/dL before angioplasty and was 1235±398 mg/dL at euthanasia.
The intima and media of the arterial segments obtained before angioplasty were composed of SMCs, foam cells, collagen, and elastin (Fig 1⇓). The relative number of foam cells within the intimal and medial layers was heterogeneous within arteries of the original lesion group and between arteries from the same rabbit. The intima was intact, and the media exhibited no compression, necrosis, or fibrosis. IEL fracture was seen in approximately half (7 of 13) of the arteries. No inflammation was observed, but increased fibrinogen/fibrin staining within the intima, indicating arterial leakiness, was noted. Sham angioplasty arteries were similar in composition to original lesion arteries. The histology is summarized in Table 1⇓.
IEL fracture was noted in 70 of 75 (93%) of arteries obtained between 1 and 28 days after angioplasty. In addition, neointimal fracture and medial damage, as manifested by compression, necrosis, and eventual fibrosis, were observed. The arterial response to angioplasty-induced injury could be differentiated into three phases: stage 1, the acute phase of mural thrombus formation and inflammation; stage 2, the subacute phase, consisting of cellular accumulation presumably as a result of proliferation and migration; and stage 3, the chronic phase of connective tissue production.
Stage 1: Mural Thrombus/Inflammatory Phase (1 to 5 Days)
Arteries obtained early after angioplasty (1 to 5 days) showed angioplasty-induced dissection within the intima and often deep dissections extending into the medial and adventitial layers (Fig 2⇓). The dissection planes generally contained thrombus, which stained strongly for fibrinogen/fibrin. Within the mural thrombus, platelets, neutrophils, lymphocytes, and monocytes were observed.
Inflammation was noted in all three arterial layers the day after angioplasty (Table 1⇑) and decreased thereafter. Intramural hemorrhage was apparent. Focal but occasionally general medial compression and necrosis without fibrosis were noted in most arteries. Focal injury of the “normal” nonatherosclerotic areas of the artery consisting of medial compression and necrosis (Fig 3⇓) was observed. Endothelial denudation was observed in the majority of arterial segments between days 1 and 5, although slight regeneration was seen on days 3 and 5 (Table 2⇓).
Stage 2: Cellular Accumulation Phase (1 to 14 Days)
Arteries obtained 5 to 14 days after angioplasty exhibited less neointimal fracture and little inflammation (Table 1⇑). The dissection was repaired by 7 days after angioplasty, and it was difficult to differentiate between the intima and original neointima. In areas in which the angioplasty dissection appeared to have occurred, cells staining for HHF-35 were present, filling in the original dissection planes, and in some arteries, areas of organized thrombus with interdigitated SMCs were noted. Foam cells generally were seen in the media, but foam cells observed in the intima were located near the internal elastic lamina.
The intima was rich in SMCs that stained for HHF-35. Collagen and elastin staining of the intima was relatively weak. Four of 12 arteries exhibited medial fibrosis at 7 days after angioplasty in areas of previous compression and necrosis. Endothelial coverage of the repaired intima as determined by the presence of continuous layer of factor VIII–positive cells was observed in a majority of arterial segments at 7 days after angioplasty (Table 2⇑).
Stage 3: Connective Tissue Production Phase (>14 Days)
Two weeks after angioplasty, no neointimal fractures were evident, although evidence of previous IEL fracture was observed. Medial fibrosis was seen in 28 of 30 (93%) arteries obtained 2 to 4 weeks after angioplasty. Increased adventitial fibrosis, observed as a thick band around the EEL, was also seen. An increased density of SMCs was observed on the luminal surface of the neointima, whereas increased extracellular matrix formation was noted throughout the rest of the intima. Few foam cells were observed in the neointima. Endothelial regeneration was completed by 28 days after angioplasty. Collagen and elastin increased over the weeks after angioplasty. Neovascularization was first observed at 28 days. In most cases, it was impossible to differentiate the restenotic neointima from the original atherosclerotic intima (Fig 4⇓).
The total number of visible nuclei in the intimal and medial layers (Table 3⇓) per cross section obtained from those arteries that did not undergo angioplasty was 1561±414 and 1288±206, respectively, and in the sham angioplasty group, the intimal nuclei totaled 1182±219 and the medial, 1103±273. By 3 days after angioplasty, the nuclei had decreased to 533±208 in the intimal layer (P<.001) and 763±251 in the medial layer (P<.001). Total cell number increased thereafter to a maximum of 2664±400 and 1749±214 in the intimal and medial layers, respectively, at day 14 after angioplasty (Table 3⇓).
Three weeks after initial air desiccation injury, the original lesion showed BrdU uptake in ≈10% of the cells in the intima and media (Table 3⇑). The sham angioplasty arteries obtained 2 weeks later exhibited a proliferation index of 3.0% to 3.7% in both arterial layers. In the experimental group, increased BrdU uptake could be observed in all three arterial layers as early as 1 day after angioplasty, and approximately one third of all intimal and medial cells stained positive at 3 days after surgery. Areas of dissection exhibited increased DNA synthesis (Fig 5⇓), although increased focal DNA synthesis was also seen in nonatherosclerotic but injured areas of the artery. Arteries obtained 1 to 5 days after angioplasty exhibited statistically significant increases in BrdU uptake both in the intima and the media (P<.001). BrdU labeling showed a heterogeneous uptake of BrdU after angioplasty in that some arteries showed areas of a focal intimal and medial BrdU staining, whereas other arteries showed a more generalized BrdU uptake.
Double-antibody staining for the presence of BrdU and either RAM-11 or HHF-35 revealed that DNA synthesis was present in cells staining positive for RAM-11 as well as those positive for HHF-35 (Fig 6⇓). Comparison of the proliferation indexes of the two cell types indicates that cells staining positive for the presence of RAM-11 exhibited earlier DNA synthesis in the intimal layer than did the cells staining positive for HHF-35. In the media, however, earlier DNA synthesis activity was noted in the HHF-35 cells (34.2% versus 9.3% at day 3 after angioplasty, Table 3⇑). Activated macrophages demonstrated more sustained DNA synthesis over the initial 2 weeks after angioplasty in both the intimal and medial layers. At all time points, cells were noted that were BrdU-positive but stained for neither the presence of RAM-11 nor HHF-35 (Fig 7⇓). In the intima, these cells primarily were near the luminal surface. In the medial layer, the proliferating cells were interspersed within the media or represented discrete intramedial areas. These cells appeared histologically similar to cells staining positive for HHF-35.
An adventitial response was observed after angioplasty. By 24 hours after angioplasty, increased BrdU activity and cellularity were observed in the adventitia. These cells did not stain for HHF-35 or RAM-11 and did not have histological characteristics of lymphocytes or neutrophils. The cells showing increased BrdU positivity were focally localized in 26 of 40 segments obtained 1 to 7 days after angioplasty, with the remaining 14 segments exhibiting diffuse uptake of BrdU positive cellularity. By 14 days, little BrdU uptake was observed in the adventitia. Increased collagen was noted in the adventitia starting at 5 days after surgery and increasing thereafter. In some segments, the EEL was disrupted and it was impossible to differentiate the cellular composition of the damaged media from the adventitia. In those arteries with disrupted EEL and IEL, the adventitia extended into the neointima (Fig 4⇑).
Comparison of the Original and Restenotic Lesions
Although the initial lesion frequently was eccentric, the restenotic lesion frequently was concentric. The neointima of restenotic arteries observed >14 days after angioplasty varied from the intima of the original, atherosclerotic arteries before angioplasty in that the composition of the intimal plaque stained more strongly for HHF-35 and fewer RAM-11 positive cells were seen. The foam cells that were observed were located near the IEL. The media of the initial lesion was a combination of SMCs and foam cells, whereas after angioplasty, media exhibited areas of acellularity associated with fibrosis in addition to the mixture of predominant SMCs and foam cells. Increased vasa vasorum within the media was observed 4 weeks after angioplasty. The arterial lesions from the sham angioplasty group were similar in composition to the original lesions, although fewer foam cells were present.
Radiographic Changes After Angioplasty
The initial mean radiographic MLD was 0.64±0.05 mm (Table 4⇓). Angioplasty increased the MLD to 1.07±0.05 mm for an acute gain averaging 0.43 mm, whereas the mean MLD was 1.06±0.43 mm 1 day after angioplasty. Between days 1 and 3, a rapid decrease in MLD was noted, followed by a slower decrease in MLD between days 3 and 14. The MLD at 28 days was increased over the 14-day MLD and approximated the postangioplasty MLD. The period between days 14 and 28 coincided with the period noted in histological analysis of decreased SMC proliferation and increased extracellular matrix formation. The radiographic MLD significantly correlated with the lumen area determined by morphometry (r=.44, P=.001) with a slope of 0.63±0.18 (P=.001) and an intercept of 0.35±0.11.
Morphometric analysis of arteries obtained before angioplasty and ≥7 days after angioplasty revealed a significant increase of luminal area after angioplasty (P<.001) but showed no statistically significant difference in the mean intimal and medial areas (Table 5⇓). The summed intima-media area increased after angioplasty, although the change was not statistically significant. The EEL circumference at 7 days after angioplasty was significantly increased (P<.001) but did not change thereafter. To determine whether a sampling bias that minimizes the morphometrically derived intimal area was present, four arteries obtained 14 days after PTA underwent serial sectioning at 200-μm distances. The results of a representative artery are plotted in Fig 8⇓, which indicates the varied intimal and medial areas within the artery as well as ratios of intima to media.
Given the substantial variability within this model and the resulting absence of statistical difference between the average intimal areas found at 1, 2, and 4 weeks after angioplasty, computation of the sample sizes required in this model to detect given changes in area was performed with two data sets. If only the segment of each artery exhibiting the greatest intima area was used, the variability of lesion area resulted in a sample size of 34 arteries necessary to detect a 50% increase in intimal formation, at a significance level of P=.05 (two-tailed test) and a power of 0.8. For a power of 0.9, the sample size is 44 arteries. If the averaged intimal area for all segments of the artery was used, then 64 arteries were necessary to show a statistically significant 50% increase in area with a power of 0.80 and a significance of P=.05, whereas 86 arteries per group were needed for a power of 0.90. This higher number reflects the increase in variability of data when all diseased arterial segments are used, reflecting an overall plaque mass. Conversely, a similar number of arteries per group would be required to show a 50% decrease in intimal formation (ie, 50% reduction in lesion restenosis) resulting from a pharmacological treatment to reduce restenosis.
Angioplasty in this model caused radial stretching of the artery, compression injury of the media, intramural hemorrhage, and dissection of the intima and/or media. The subsequent vascular response to this injury exhibited characteristics common to the acute inflammatory response to injury. Restenosis resulted from a combination of mural thrombus filling the dissection planes and forming a framework for subsequent cellular ingrowth as well as deposition of collagen and elastin in the intima, media, and adventitia.14 The extracellular matrix presumably was produced by fibroblasts in the adventitia and by SMCs in the media and may have been a component of intramural wound contraction. Collagen and elastin may have exerted arterial constriction from the adventitia. Proliferating SMCs and macrophages were noted in the intima and media of regions possessing a preexisting lesion as well as in regions of the media of nonatherosclerotic sections damaged by angioplasty. The vascular reaction to injury may thus be divided into three overlapping stages of vascular repair and restenosis: stage 1, the acute response to injury, consisting of mural thrombus formation, inflammation, intramural hemorrhage, and vessel recoil, which resulted immediately from angioplasty and was noted for 5 days; stage 2, the subacute phase of cellular accumulation in the area of thrombus formation or arterial injury (days 1 through 14); and stage 3, the subacute and chronic phases of extracellular matrix and adventitial connective tissue formation (days 7 through 28).
We noted substantial variability in the intimal area between and within arteries as well as variability in the histological appearance of the restenotic lesion. The variability is postulated to have resulted from (1) inconsistent formation of dissection, (2) variable thrombus formation within dissection planes, (3) variable cellular proliferation rates, and (4) an unpredictable degree of elastic recoil and scar contraction. In addition, the arterial response to injury was superimposed on variable degrees of preexisting atherosclerosis, rendering the determination of the subsequent neointimal accumulation within individual arteries difficult to assess. In some arteries, the extent of neointimal accumulation may have been small in relation to the preexistent atherosclerotic area.
A rapid decrease in the radiographic MLD was noted in the initial 3 days after angioplasty, which may represent contributions from elastic recoil of the stretched normal segment as well as the formation of thrombus within fracture planes leading to decreased luminal diameter and wound contraction. A significant decrease in the total number of cells in the intima and media was noted on the third and fifth days, suggesting cell loss due to necrosis or apoptosis. Concurrently, a threefold increase in BrdU labeling of intimal and medial SMCs and macrophages was noted at 3 to 7 days. Since the cell cycle for SMCs is 24 to 30 hours,15 entry into S phase by some cells occurred early after injury. Although focal areas of angioplasty-induced injury and DNA synthesis of nonatherosclerotic media were observed and may represent a source of accumulating SMCs, the greatest degree of BrdU uptake was noted in SMCs and macrophages near a dissection (Fig 5⇑) and may herald the subsequent growth of the cells into the thrombotic framework, as some dissections exhibited both thrombus as well as cellular elements. A population of cells exhibited DNA synthesis but did not stain with either RAM-11 or HHF-35 and may represent endothelial cells, fibroblasts, or SMCs in the synthetic phenotype possessing depolymerized actin (Fig 7⇑). Proliferation of a minimum of three cell types was accordingly observed, which may contribute to variability in cellular accumulation.
By 7 days after angioplasty, little evidence of thrombus was observed, the artery was repaired without signs of the original dissection, and the composition of the newly formed neointima was primarily cellular, with little connective tissue noted. By 14 days after angioplasty, the cellular density of the intima decreased (Table 2⇑) and an increased connective tissue matrix consisting of collagen and elastin fibers was observed. SMCs express increased mRNA for collagen when density-arrested,16 and increased collagen synthesis has been observed when proliferating SMCs reach confluence.17 18 By 14 days after angioplasty, the artery showed a decrease in radiographic lumen; however, over the subsequent 14 days, the lumen either remained constant or increased in size, mirroring observations in the rat model of balloon injury.19 Insofar as markedly elevated cholesterol levels were noted during the entire postangioplasty period, it is not anticipated that the histological or radiographic increase at 28 days resulted from lesion regression secondary to a decline in serum cholesterol levels, although this possibility cannot be excluded. The decrease in intimal area may have resulted from a decrease in cellular elements within the lesion and an increase in contracture resulting from the increased connective tissue accumulation.
There was a significant relation between the radiographic MLD and the morphometrically derived luminal area, although the correlation coefficient was low, at r=.44. One possible explanation for the low correlation coefficient includes the inherent limitation of extrapolating a two-dimensional area from a single radiographic image. Thus, an eccentric lesion of a given morphometric MLD may be reflected by either a smaller radiographic MLD if the x-ray beam is parallel to the long axis of the lumen or a larger MLD if the x-ray beam is perpendicular. Second, despite attempts to match radiographic MLD with morphometric luminal area by use of anatomic landmarks such as side branches as well as placement of the surgical clip near the site of induced atherosclerosis, sampling error was possible. Additional sampling error may have been present, because the morphometric luminal area as shown in Table 5⇑ represents the mean area of the sections with the greatest amount of intimal mass. However, intimal mass may not necessarily correlate with luminal area, owing to compensatory arterial enlargement.20 A third possibility is related to the presence of thrombus. Thrombus within dissection planes (Fig 2⇑) or intraluminal thrombus would reduce the radiographic MLD; on morphometric analysis, however, thrombus was excluded from measurement, thereby yielding a larger morphometric MLD. Fourth, fixation of the artery results in tissue shrinkage and may therefore result in a smaller morphometric lumen than observed in vivo.
The finding that the radiographic loss of MLD was not accompanied by a statistically significant increase in intimal area was unexpected but confirms data published by other investigators.20 21 By other indirect measures of cellular accumulation, including increased BrdU labeling, increased SMC composition, increased absolute cell number of the restenotic lesion, and the radiographic appearance, an increase in neointimal area would be expected. By comparing the radiographic luminal loss with the morphometric area at euthanasia, a relative contribution of neointimal formation to luminal loss can be grossly calculated. This comparison is simplified, and it does not take into account the arterial contraction at fixation or the vasoreactivity at the time of euthanasia and assumes a concentric lesion. Thus, it is intended only for illustrative purposes. The radiographic MLD immediately after angioplasty was 1.01±0.08 mm in the 7-day group, with a MLD of 0.81±0.11 mm at day 7. The corresponding minimal luminal areas are, therefore, 0.95 and 0.52 mm2, respectively (Table 4⇑), and the late luminal area loss is 0.43 mm2. The increase in intimal area was 0.142 mm2 (Table 5⇑). Hence, the intimal increase represented only 33% of the loss of luminal area. The neointima (the increase in measured intima after angioplasty) contributed only 28% (0.142 of 0.500 mm2, Table 5⇑) to the 7-day total intimal area of 0.500 mm2. Thus, with this simplified calculation, the relative contribution of neointima at the point in time when it is most cell-rich is approximately one third of the total luminal loss and about one fourth of the resulting total intimal area. For 14 days after angioplasty, the percentage of luminal loss caused by intimal gain is similarly estimated at 24% and the contribution of neointima to the resulting intima is 22%. The remaining loss of MLD must have resulted from scar contraction and vascular remodeling (Table 5⇑). Given the multifactorial cause of radiographic MLD reduction, it may be anticipated that a treatment strategy designed to affect only SMC proliferation may not be effective in reducing restenosis.
The lack of statistical significance in the morphometric determination of the increase in intimal area reflects the large standard deviation of the intimal area. By post hoc statistical analysis of the variability, we determined that a sample size of 34 arteries would be needed to show a statistically significant 50% increase in neointimal area, rendering the sample size in the present study insufficient to show a statistical difference. This is consistent with the observations of Strauss et al,14 in which a two-injury rabbit iliac model showed a trend toward increasing intimal area 1 to 4 weeks after the angioplasty, although statistical significance was not obtained until 12 weeks after angioplasty. The variability in the intimal area as well as the ratio of intima to media within an artery may lead to false low estimates of intimal area due to sampling bias (Fig 8⇑) and indicates that a large number of sections should be evaluated for each artery, with the section exhibiting the greatest amount of intima chosen to be the representative section. The statistical determination also indicates that large sample sizes (ie, >34 arteries per group) are necessary to evaluate the effect of an agent to reduce restenosis in this model.
The presence of adventitial inflammation and increased DNA synthesis shortly after angioplasty suggests that adventitial fibroblasts also may play a role in the restenotic process. An increased density of circumferential adventitial collagen and intimal-medial fibrosis were observed and may have played a role in subacute vascular constriction. In specimens exhibiting severe injury, the EEL was disrupted and connective tissue extended from the adventitia into the intima and media and so contributed to the restenotic intimal cell mass (Fig 4⇑). Thus, it is possible that the restenotic lesion derives not only from medial SMCs, but in cases of extreme injury with adventitial dissection, from adventitial fibroblasts that migrate into the media and/or neointima as myofibroblasts. These cells may then undergo apoptosis when the wound is healed, revert to a quiescent fibroblast form, or become SMCs expressing α-actin.22
Human restenosis is the vascular response of an atherosclerotic artery to angioplasty-induced injury. In the rabbit model characterized here, the initial fibrofatty atherosclerotic lesion is only an approximation of human atherosclerosis, which tends to be more fibrocalcific, contains lipid pools, and has areas of necrosis. The purpose of the present study was to evaluate the response of an atherosclerotic artery to balloon injury, and although the initial lesion may not represent human atherosclerosis in all its complexity, the arterial response to injury exhibits similarities to the human response noted at necropsy23 24 25 26 and in vivo intracoronary ultrasonography,27 28 29 in which plaque fracture, intimal compression, and vessel stretching have been shown after angioplasty.28 29 Tears or dissections involving the intima, media, and/or adventitia have been observed at necropsy in postangioplasty patients.23 24 25 26 Mural thrombus formation has been observed early,24 26 whereas neointimal growth has been observed at 11 days after angioplasty, with the incidence of neointimal proliferation subsequently increasing to 83%.24 The proliferating intima was seen predominantly around the circumference of the lumen and into the angioplasty-induced dissection, similar to the observations in the present study. These investigators24 also observed increased intimal collagen 6 months after angioplasty. Although it would appear that the sequence of vascular changes occurring in response to angioplasty injury in this model parallel the human process as observed in vivo and at necropsy, it is impossible to know with certainty in humans that these events occur in the months after angioplasty. An extension of the present study into a larger animal in which angioplasty of an atherosclerotic artery is used may aid in the determination of the universality of the proposed vascular response to angioplasty-induced injury and restenosis.
This work was supported by a Grant-in-Aid from the American Heart Association, Indiana Affiliate (R.L.W.), and an educational grant from the Krannert Institute of Cardiology and Eli Lilly and Co. (K.L.M., R.L.W., I.G.-P., D.R.H.). Gratitude is hereby expressed to David Mendel and R. Gerald Dreesen, RRT, for technical assistance, Linda Smith in processing the histological samples, and Linda Rohyans for secretarial assistance.
Presented in part at the 64th Scientific Sessions of the American Heart Association, Dallas, Tex, November 11-14, 1991, and the 41st Scientific Sessions of the American College of Cardiology, Dallas, Tex., March 1992.
- Received February 9, 1994.
- Revision received April 10, 1995.
- Accepted April 17, 1995.
- Copyright © 1995 by American Heart Association
Serruys PW, Luijten HE, Beatt KJ, Geuskens R, deFeyter PJ. Incidence of restenosis after successful coronary angioplasty: a time-related phenomenon: a quantitative angiographic study in 342 consecutive patients at 1, 2, 3, and 4 months. Circulation. 1988;77:361-371.
Banai S, Shou M, Correa R, Jaklitsch MT, Douek PC, Bonner RF, Epstein SE, Unger EF. Rabbit ear model of injury-induced arterial smooth muscle cell proliferation, kinetics, reproducibility, and implications. Circ Res. 1991;69:748-756.
Faxon DP, Weber VJ, Handenschild C, Gottsman SB, McGovern WA, Ryan TJ. Acute effects of transluminal angioplasty in three experimental models of atherosclerosis. Arteriosclerosis. 1982;2:125-133.
Hanke H, Strohschneider T, Oberhoff M, Betz E, Karsch KR. Time course of smooth muscle cell proliferation in the intima and media of arteries following experimental angioplasty. Circ Res. 1990;67:651-659.
Sarembock IJ, LaVeau PJ, Sigal SL, Timms I, Sussman J, Haudenschild C, Ezekowitz MD. Influence of inflation pressure and balloon size on the development of intimal hyperplasia after balloon angioplasty: a study in the atherosclerotic rabbit. Circulation. 1989;80:1029-1040.
Hsu S, Raine L, Fanger H. Use of avidin-biotin-peroxidase complex (ABC) in immunoperoxidase techniques: a comparison between ABC and unlabelled antibody (PAP) procedures. J Histochem Cytochem. 1981;29:577-580.
Rasmussen LR, Garbarsch C, Chemnitz J, Christensen BC, Lorenzen I. Injury and repair of smaller muscular and elastic arteries: immunohistochemical demonstration of fibronectin and fibrinogen/fibrin and their degradation products in rabbit femoral and common carotid arteries following a dilatation injury. Virchows Arch A Pathol Anat Histopathol. 1989;415:579-585.
Tsukada T, Rosenfeld M, Ross R, Gown AM. Immunocytochemical analysis of cellular components in atherosclerotic lesions. Arteriosclerosis. 1986;6:601-613.
Kockx MM, DeMeyer GRY, Andries LJ, Bult H, Jacob WM, Herman AG. The endothelium during cuff-induced neointima formation in the rabbit carotid artery. Arterioscler Thromb. 1993;13:1874-1884.
Strauss BH, Chisholm RJ, Keeley FW, Gottlieb AI, Logan RA, Armstrong PW. Extracellular matrix remodeling after balloon angioplasty injury in a rabbit model of restenosis. Circ Res. 1994;75:650-658.
March KL, Wilensky RL, Roeske RW, Hathaway DR. Effects of thiol protease inhibitors on cell cycle and proliferation of vascular smooth muscle cells in culture. Circ Res. 1993;72:413-423.
Liau G, Chan LM. Regulation of extracellular matrix RNA levels in cultured smooth muscle cells: relationship to cellular quiescence. J Biol Chem. 1989;264:10315-10320.
Stepp M, Kindy MS, Franzblau C, Sonenshein GE. Complex regulation of collagen gene expression in cultured bovine aortic smooth muscle cells. J Biol Chem. 1986;261:6542-6547.
Ang AH, Tachas G, Campbell JH, Bateman JF, Campbell GR. Collagen synthesis by cultured rabbit aortic smooth-muscle cells: alteration with phenotype. Biochem J. 1990;265:461-469.
Kakuta T, Currier JW, Haudenschild CC, Ryan TJ, Faxon DP. Differences in compensating vessel enlargement, not intimal formation, account for restenosis after angioplasty in the hypercholesterolemic rabbit model. Circulation. 1994;89:2809-2815.
Post MJ, Borst C, Kuntz RE. The relative importance of arterial remodeling compared with intimal hyperplasia in lumen renarrowing after balloon angioplasty: a study in the normal rabbit and the hypercholesterolemic Yucatan minipig. Circulation. 1994;89:2816-2821.
Hoshino T, Yoshida H, Takayama S, Iwase T, Sakata K, Shingu T, Yohoyama S, Nori N, Kaburagi T. Significance of intimal tears in the mechanism of luminal enlargement in percutaneous transluminal coronary angioplasty: correlation of histologic and angiographic findings in postmortem human hearts. Am Heart J. 1987;114:503-510.
Honye J, Mahon DJ, Jain A, White CJ, Ramee SR, Wallis JB, Al-Zarka A, Tobis JM. Morphological effects of coronary balloon angioplasty in vivo assessed by intravascular ultrasound imaging. Circulation. 1992;85:1012-1025.
Losordo DW, Rosenfield K, Pieczek A, Baker K, Harding M, Isner JM. How does angioplasty work? Serial analysis of human iliac arteries using intravascular ultrasound. Circulation. 1992;86:1845-1858.