© 1995 American Heart Association, Inc.
Articles |
The Intima
Soil for Atherosclerosis and Restenosis
From the Department of Pathology, University of Washington, Seattle.
Correspondence to Dr Stephen M. Schwartz, MD, PhD, Department of Pathology, University of Washington, Vascular Biology/Box 357335, Seattle, WA 98195-7335.
Key Words: intima • atherosclerosis • restenosis
 |
Introduction |
|---|
 Top Introduction Formation of the IntimaNeointimal Formation: A Generic...Pharmacology of Neointimal...Potential Clinical Targets Based...Role of Intima in...Special Properties of...Intimal Cells Overexpress...Mechanisms Controlling Plaque...Pathological Significance of...Mechanism of Lumen Occlusion...The Nature of Restenosis...Relevance of Animal Stenosis...Future DirectionsReferences |
|---|
Thirty years ago, John French wrote a seminal review describing the unique properties of the arterial intima.1 His major point was that the smooth muscle cells of the intima have a unique morphology (Fig 1
). French also pointed out that
intimal formation appeared during normal development and aging as well
as in the response of arteries to almost any imaginable injury,
including atherosclerosis.
|
This article attempts to update French's review. We will discuss the
developmental origins of the intima and suggest that the
arterial intima is a distinct tissue with a long and
rapidly increasing list of differentially expressed genes (Table 1). This pattern of intimal gene
expression may be responsible for the origins of
atherosclerosis.
|
The bulk of the review, however, deals with the posited role of smooth muscle replication in hypertension, atherosclerosis, and restenosis. Topics covered will include a revisiting of the 20-year-old observation that atherosclerosis is monoclonal2 3 4 5 as well as reviews of current knowledge of the pharmacology of smooth muscle proliferation. Perhaps most important, however, we will present a critical discussion of evidence for and against a role for smooth muscle replication in these diseases.
|
Formation of the Intima |
|---|
|
TopIntroductionFormation of the IntimaNeointimal Formation: A Generic...Pharmacology of Neointimal...Potential Clinical Targets Based...Role of Intima in...Special Properties of...Intimal Cells Overexpress...Mechanisms Controlling Plaque...Pathological Significance of...Mechanism of Lumen Occlusion...The Nature of Restenosis...Relevance of Animal Stenosis...Future DirectionsReferences |
|---|
Surprisingly little is known about how vessels acquire a coat of smooth muscle cells. In the rat, the precursors of the carotid arteries lack a smooth muscle layer as late as day 12 of gestation.6 The first smooth muscle cells appear as a condensation around these endothelial tubes over a period of
24 hours. As the mesenchymal cells invest the
endothelial tube, the cells begin to express smooth
muscle 
-actin.7 Smooth muscle -actin, however, is
not a definitive marker of smooth muscle lineage and can be found in
many non–smooth muscle cell types.8 More definitive
smooth muscle–specific proteins are expressed later in development.
These include desmin, calponin, smooth muscle 
-actin, and the smooth
muscle myosins.9 10 11 12 13 Whereas these latter genes seem
largely restricted to smooth muscle cells, their expression is
typically lost in cultured cells or in cells that form the
neointima.
It is important to realize that not all species go on to form an intima
unless their arteries are injured. In humans, the intima develops
spontaneously after birth14 15 16 17 and increases rapidly until
6 months of age.14 17 French suggests that
arterial intimal formation is a function of animal size,
because arteries of smaller animals do not form an intima unless the
vessel first undergoes trauma.1 The most interesting site
of intimal growth is the origin of the left anterior descending
coronary artery, a site with a high probability of developing
atherosclerosis in later life.
The only site where intimal formation has been well studied is the ductus arteriosus.18 19 Contrary to conventional wisdom, this intima forms spontaneously before birth, not as the result of trauma associated with ductus closure. Therefore, it is intriguing to consider ductus closure as a model for diseases of the intima, including atherosclerosis and restenosis.
|
Neointimal Formation: A Generic Response of Vessels to Injury? |
|---|
|
TopIntroductionFormation of the IntimaNeointimal Formation: A Generic...Pharmacology of Neointimal...Potential Clinical Targets Based...Role of Intima in...Special Properties of...Intimal Cells Overexpress...Mechanisms Controlling Plaque...Pathological Significance of...Mechanism of Lumen Occlusion...The Nature of Restenosis...Relevance of Animal Stenosis...Future DirectionsReferences |
|---|
"Neointima" is an intima that forms in response to injury. We recognize a new intima because many arteries, especially smaller ones, either do not form an intima at all or do so only slowly as the animal ages.1 20 Neointimal formation, however, occurs in all arteries as a response to a wide variety of injuries, including radiation, application of turpentine to the adventitia, wrapping the vessel, and electrical stimulation, as well as mechanical injuries, including placement of a suture, scratching with a probe, or dilatation of the common carotid artery with an embolectomy balloon catheter.1 21 22 23 24 25 26 27 28 29 30 These changes are even seen in transplanted veins that undergo arterialization. Put another way, neointimal formation is as characteristic of responses to injury in the smooth muscle–rich vessel wall as fibrosis is characteristic of responses in the fibroblastic dermis of skin or gliosis is characteristic of the brain with its glial cells. The most obvious questions about neointimal formation are as follows: (1) Are intimal or neointimal cells, like glia, a distinct cell type? (2) What molecules control neointimal formation? (3) What does neointimal formation after arterial injury tell us about the critical mechanisms underlying arterial pathology in aging, atherosclerosis, or restenosis?
|
Pharmacology of Neointimal Formation |
|---|
|
TopIntroductionFormation of the IntimaNeointimal Formation: A Generic...Pharmacology of Neointimal...Potential Clinical Targets Based...Role of Intima in...Special Properties of...Intimal Cells Overexpress...Mechanisms Controlling Plaque...Pathological Significance of...Mechanism of Lumen Occlusion...The Nature of Restenosis...Relevance of Animal Stenosis...Future DirectionsReferences |
|---|
The best-studied model of neointimal formation is the response of the rat carotid artery to balloon angioplasty.23 26 31 32 This simplicity of the rat model has facilitated kinetic analyses of the smooth muscle response, leading to the definition of four waves describing the smooth muscle cell response to injury and to the identification of molecules playing a role in, or interacting with, each of these waves in rats.21 33 34
The use of the rat arterial response to injury as a model for human atherosclerosis or restenosis, however, should only be considered with caution. First, unlike arteries in larger mammals, the rat carotid artery has only rare intimal cells.35 36 Formation of new intima may poorly replicate the response to injury when an intima or even an atherosclerotic lesion already exists.37 Second, although platelet adhesion occurs in injured rat arteries, thrombus formation and leukocyte infiltration are minimal. These important events are prominent after vascular injury in rabbits, swine, and nonhuman primates.38 39 40 41 In pig coronary arteries, injury caused by distension with an oversized stent initially causes thrombus formation, followed within a few days by significant macrophage and lymphocyte infiltration and finally by smooth muscle cell migration and proliferation in the intima.42 Leukocyte infiltration is also common in rabbit arterial lesions43 44 but not in rat lesions.40 Unfortunately, little is known about the molecules controlling response to injury in these more complex systems.
In the rat, the balloon injury model begins with complete destruction
of the endothelium as well as extensive death of medial
smooth muscle cells.32 The first response to balloon
injury, called "first wave," consists of medial smooth muscle
cell proliferation and begins 24 hours after the injury. In elegant
studies, Reidy and colleagues45 46 have shown that this
wave of replication can be completely accounted for by release of bFGF
from dying smooth muscle cells. Despite extensive studies of PDGF as an
in vitro mitogen,47 studies with infused PDGF, as well as
studies with anti-PDGF antibodies, have shown that this molecule is
only a weak mitogen for smooth muscle cells in the media of injured
arteries.48 49 50 However, more limited data suggest that
-adrenergic antagonists and Ang II may also control
medial replication.51 52 53
Although the first wave has no obvious relation to the later migration and intimal proliferation, several studies with antisense agents directed at cell-cycle genes have shown that inhibition of these initial proliferative events in some as-yet-undefined way leads to a diminution of the final extent of neointimal proliferation. Since the antisense is likely only available in the first few days after injury, these studies suggest long-lasting effects on later events, including migration or intimal proliferation.54 55 56 57 58 59 60 The migration of smooth muscle cells across the internal elastic lamina to form the intima constitutes the "second wave." Smooth muscle cells are readily observed on the luminal side of the internal elastic lamina 4 days after injury to rat arteries.32 The duration of migration is not known. Several molecules other than PDGF may contribute to smooth muscle migration, including TGF-ß, bFGF, and, as discussed below, Ang II. The relative contributions of these different molecules are not known, nor do we know whether other molecules are involved.46 53 61 PDGF has received attention because of many studies implicating this molecule as a growth factor for cultured smooth muscle.47 The main effect of PDGF in vivo, however, is on smooth muscle cell migration and not proliferation.49 50 62 Immunoneutralization of endogenous PDGF or depletion of circulating platelets inhibits the development of the intimal lesion in response to balloon injury by inhibiting smooth muscle cell migration toward the intima without a reduction in the frequency of proliferating smooth muscle cells in the media or the intima. In the gentle denudation model, administration of PDGF-BB, which recognizes all isoforms of PDGF receptors, causes a 10- to 20-fold increase in smooth muscle cell migration but no more than a 2-fold increase in smooth muscle cell proliferation.48 In vivo dose-response studies show that discrepancies between these in vivo data and the potent mitogenic effects of PDGF on smooth muscle cells in vitro do not reflect inadequate dosage.48 As discussed below, we do not know whether PDGF is a mitogen in species other than the rat, as is suggested by studies of PDGF transfected into swine arteries. For that matter, we do not know whether PDGF is mitogenic in the rat at other sites of vascular injury, as is suggested by studies of PDGF infused into the kidney.63
Once smooth muscle cells arrive in the intima of the rat artery, neointimal cells closest to the lumen may replicate for weeks to months.64 This replication is called the "third wave." Although we cannot say that any specific molecule has definitively been identified as a third-wave mitogen, a few candidates appear to be present in the intima. For example, PDGF-A chain is overexpressed in the intima and colocalizes with replicating cells. Moreover, when PDGF-B was transfected into swine arteries, there was a marked elevation of replication.65 Unfortunately, we do not know whether this PDGF was acting as a growth factor in the usual sense or was mimicking the transforming mechanism of the viral homologue of PDGF, v-sis.66 67 68 As already noted, in the rat model PDGF-B will not stimulate replication if infused, nor do antibodies to PDGF suppress third-wave replication.49 50 69 Other growth control molecules that appear to be overexpressed in the rat neointima include the Ang II receptor, AT-1, and TGF-ß.61 70 IGF-1 is also overexpressed after injury; however, it is overexpressed in the media.71 The neointima can be stimulated to show a further increase of replication by infusion of several of these molecules. This increased responsiveness to mitogens can be called a "fourth wave" and involves at least TGF-ß, bFGF, and Ang II as agonists.45 51 61 Again, PDGF does not appear to be mitogenic in the rat carotid artery model, even at doses that markedly stimulate migration of smooth muscle cells in the second wave.48
As the time this review was written, no specific molecular antagonist had been shown to inhibit this replication. Even antibodies to bFGF, which are so effective in inhibiting the first wave, have been shown to be impotent against third-wave replication.46 In the absence of inhibitors, Koch's postulate cannot be fulfilled. Therefore, we cannot identify the critical molecules that sustain elevated replication in the third wave.
Although we have emphasized smooth muscle cell mitogens, pathological
loss of smooth muscle quiescence could also represent the loss
of growth inhibitors. Regeneration of the
endothelium inhibits proliferation of the underlying
neointimal cells.32 64 72 At this time,
however, the mechanisms for this inhibition remain incompletely
understood. Suppression of neointimal cell growth might be
due, for example, to endothelial production of
NO. In vitro studies have shown that NO can inhibit replication of
cultured smooth muscle cells.73 Administration of large
doses of L-arginine aimed at chronically raising NO
levels in animals immediately after balloon catheter–induced
endothelial denudation have resulted in significant
inhibition of neointimal thickening.74 75 76
Interestingly, Farhy et al77 have reported that L-NMMA, a
blocker of NO generation, can partially counteract the suppressive
effect of ramipril, an ACE inhibitor, on
neointimal thickening after balloon injury in rat
arteries.77 They speculate that inhibition of
neointimal formation by ACE inhibitors may be a
result of elevation in the level of bradykinin. Bradykinin is a potent
stimulator of NO production and is itself normally degraded by
ACE. When administered alone, however, the same dose of L-NMMA did not
affect neointimal thickening.77 Moreover,
Hansson et al78 have reported that the
cytokine-inducible isoform of NO synthase is chronically
expressed at high levels by smooth muscle cells at the denuded surface
of the neointimal lesion in the rat injured carotid
artery,78 an area in which smooth muscle DNA synthesis
remains chronically elevated.64 In addition, in vitro
studies suggest that in the presence of bFGF, NO can actually stimulate
DNA synthesis and mitogenesis in smooth muscle cells from the rat aorta
in primary culture, whereas a similar treatment causes inhibition of
DNA synthesis in the passaged smooth muscle cells.79 In
summary, we do not have definitive evidence for an
endogenous role of NO as an inhibitor of
replication of neointimal smooth muscle. Table 2 summarizes the molecules that control the three
waves.
|
|
Potential Clinical Targets Based on the Rat Model |
|---|
|
TopIntroductionFormation of the IntimaNeointimal Formation: A Generic...Pharmacology of Neointimal...Potential Clinical Targets Based...Role of Intima in...Special Properties of...Intimal Cells Overexpress...Mechanisms Controlling Plaque...Pathological Significance of...Mechanism of Lumen Occlusion...The Nature of Restenosis...Relevance of Animal Stenosis...Future DirectionsReferences |
|---|
Angiotensin is of special interest as a mediator of neointimal formation because ACE inhibitors that block neointimal formation in the rat model53 80 failed to inhibit restenosis in humans.81 Exploring the reasons for the failure of clinical trials using an ACE inhibitor requires a brief review of some of the pharmacology of the angiotensin system. The ultimate precursor of Ang II is the plasma protein angiotensinogen, from which renin, a kidney-derived endopeptidase, releases the inactive decapeptide Ang I. The carboxypeptidase ACE then metabolizes Ang I to release active Ang II. As already noted, ACE also contributes to the degradation of bradykinin, a potent endogenous inhibitor of neointimal formation.77 Although enzymes other than ACE appear to be able to activate Ang I,82 the lack of specificity of ACE inhibitors can be overcome by use of specific inhibitors directed at the angiotensin receptors.83 84
Components of the angiotensin pathway may be found together in the intima or neointima.85 86 Although both AT-1 and AT-2 receptors are also present in the normal wall, the principal Ang II receptor, AT-1, is elevated in the neointima.70 87 Angiotensinogen, which is available from the circulation, may also be produced by the vessel wall. Rakugi et al88 have reported an increase in angiotensinogen gene expression in both the media and the neointima during the first 2 weeks after balloon injury to the rat abdominal aorta. Although the validity of these data may be questioned in view of contradictory in situ hybridization studies from other authors,89 90 angiotensinogen mRNA is available locally from perivascular fat. The vessel wall may also contain reninlike activity.91 Some of this activity may be chymase, a protease found in human mast cells, especially around blood vessels.82 ACE itself is made by the medial smooth muscle cells and by the endothelium, the major source of this enzyme.92 Rakugi et al88 have reported that despite loss of the endothelium, 2 weeks after balloon injury the rat abdominal aorta expresses increased levels of ACE activity as well as increased immunoreactivity for ACE in the neointima. Fishel et al93 also have reported increased ACE activity in the rat injured carotid artery as well as a tissue gradient of ACE expression, with the greatest immunoreactivity in the most luminal cells of the neointima. Upregulation of ACE activity also occurs in dog arteries 30 days after balloon injury.94 On the other hand, Viswanathan and colleagues70 87 used quantitative autoradiography with iodinated ligands to study ACE expression up to 30 days after injury in the rat carotid artery and thoracic aorta. In these studies, levels of binding to ACE levels per milligram of tissue were similar in neointimal lesions devoid of endothelium 15 or 30 days after injury compared with sham arteries with endothelium. Thus, after balloon denudation, there may be no loss of ACE activity, since neointimal ACE replaces the ACE normally produced by the endothelium.
In summary, a possible role for the renin-angiotensin system is supported by evidence for the expression of ACE and other components of the angiotensin system in the injured vessel wall. First-wave replication, migration into the intima, and intimal thickening can be prevented by angiotensin receptor agonists directed against the AT-1 receptor.83 84 ACE inhibitors and receptor antagonists seem to block migration as well.83 84 These effects are likely mediated by the AT-1 receptor. One report of inhibition by an AT-2 receptor ligand molecule required that the drug be given locally at very high doses.95 Thus, it is quite reasonable to interpret the effects of ACE inhibitors in blocking neointimal formation as being in large part the result of blocking formation of Ang II.80
The clinical trials may have failed for several reasons. First, the effects of ACE inhibitors may be determined by genetic variation in the population. Ohishi et al96 recently found a striking correlation between restenosis and an isoform of ACE. This isoform, a deletion, had previous been identified by Cambien et al97 as a risk factor for myocardial infarction. Second, it is important to note that ACE inhibitor doses used in the clinical trials were ineffective in the rat studies as well as in trials after arterial injury in swine and nonhuman primates.98 99 100 Effective doses in animal studies were hypotensive,80 possibly because of the elevated levels of bradykinin discussed above. Such doses would not be acceptable in humans. Third, as already noted, ACE is not the only enzyme able to generate Ang II in humans. Chymase is present in human plaques and cannot be inhibited by the converting enzyme inhibitors.82 101 Chymase activity is also dramatically upregulated 30 days after balloon injury in dog arteries.94 Other potential alternative pathways for the activation of Ang II also include serine proteases, such as leukocyte-derived tonin,102 and carboxypeptidases from human platelets103 or from the tunica media.104
The PDGFs are also of special clinical interest because of the emphasis
on these molecules in the "response to injury"
hypothesis.47 We have already noted that in contrast to
FGF or Ang II, infused PDGF-BB seems to function primarily in promoting
migration.48 49 50 The absence of a mitogenic
effect in the rat model, of course, does not rule out an effect in
humans. As already discussed, transfected PDGF-BB was
mitogenic in swine. Moreover, again in the rat, the
production of PDGF-B by cultured smooth muscle cells does
correlate with an enhanced ability to grow without serum-derived growth
factors.105 106 We have recently found that 10% of the
most superficial cells in the neointima express PDGF-B
chain, as demonstrated by in situ hybridization.107 It is
intriguing to think that the subset B chain–positive cells in vitro
are the ancestors of the characteristic B chain–positive cells seen
after passage in vitro. We do not know whether similar cells might
exist in proliferative regions of human restenotic tissue.
PDGF-A chain is also of interest, despite the observation that PDGF-A is itself a nonmitogen or a weak mitogen in vitro.48 This molecule is chronically overexpressed by smooth muscle cells in the neointima in the rat model69 and in atherosclerotic plaques in humans.108 109 A recent immunocytochemical study colocalized PCNA, a marker for cell replication (see below), with PDGF-A chain.110 The mitogenic effect of a number of mild mitogens, including bFGF, TGF-ß, and Ang II, has been shown to be substantially blocked in vitro by antibodies to PDGF-A.49 Thus, we need to consider the possibility that the localized overexpression of PDGF-A seen in the neointima may act as a cofactor that increases the fourth-wave responsiveness to other growth factors.
TGF-ß may also play an important role in third- and fourth-wave replication. As already noted, there is an accumulation of TGF-ß in the neointima,61 and we found that infused TGF-ß was a neointimal mitogen. More recently, Nabel et al111 found augmented neointimal hyperplasia in a swine model after injury and transfection with a TGF-ß plasmid. Similarly, Wolf et al112 found that neointimal formation in the rat model was inhibited by antibodies to TGF-ß. Finally, a recent study of atherectomy specimens found elevated levels of TGF-ß in restenotic lesions.113
These in vivo data contrast with the recent proposal by Grainger and colleagues114 that TGF-ß in humans is an inhibitor of atherosclerosis. The Grainger hypothesis is based on studies in vitro in which TGF-ß is sometimes seen as a growth stimulant and at other times as an inhibitor. This variability may reflect the strain of smooth muscle cells studied and their state of confluence.115 116 117 Grainger et al114 emphasize the inhibitory effects and attempt to link endogenously produced TGF-ß with elevated atherosclerosis risk due to elevated Lp(a). Although Lp(a) is well known as a risk factor for atherosclerosis, a recent study suggests that Lp(a) levels also correlate with an increased incidence of restenosis.118 119 As a homologue of plasminogen, the apoprotein of Lp(a), called Apo(a), is believed to inhibit the formation of plasmin.120 121 Plasmin, in addition to its role in fibrinolysis, is essential to the activation of TGF-ß.122 Grainger and colleagues114 123 124 125 found that Apo(a) is mitogenic for smooth muscle cells in culture and were able to attribute this mitogenesis to inhibition of the formation of plasmin and the subsequent inability of the cells to activate TGF-ß. In this view, the predominant role of TGF-ß would be as an endogenous inhibitor of lesion formation at sites of active coagulation, and Apo(a) would enhance lesion formation by diminishing the production of activated TGF-ß.
Thrombin is another neointimal molecule that is attracting renewed interest because of the recent cloning of a thrombin receptor and the recognition that thrombin can have a direct mitogenic effect on cells independent of the coagulation cascade.126 127 Particularly intriguing is the evidence that thrombin activity is chronically elevated in the neointima and that the neointima, as well as human atherosclerotic plaque, overexpresses the thrombin receptor.128 129 130 In addition, studies involving animal model systems more complex than the rat, especially the pig coronary artery stent model, suggest that thrombosis may well play a key role in the early events leading to restenosis.42
The last three molecules for this discussion are somatostatin, heparin, and the integrins associated with thrombosis and vascular responses to injury. Somatostatin analogues and heparin-like compounds are of special interest because of the extensive studies of their ability to inhibit neointimal formation in animal models. Somatostatin analogues, in particular, angiopeptin, have been found to be effective in a range of animals species and different models of arterial injury, including fat feeding, balloon angioplasty, neointimal formation in vein grafts, and transplant atherosclerosis.131 132 133 134 135 Clinical trials are under way, but preliminary results appear to be equivocal.136 The pharmacological rationale for this approach is unclear; however, it is important to note that somatostatin receptors are widespread in the body and may act as vasodilators or vasoconstrictors in blood vessels.137 138 139 140 Failure or success of angiopeptin trials to prevent restenosis will be especially important, because the dosages of the drug that were used in humans were comparable to those used in animals.
Similarly, heparin has been widely studied as an inhibitor of intimal formation in animal models.141 As with somatostatin, the mechanism of action of heparin as an inhibitor of intimal hyperplasia is poorly understood. Although much attention has been focused on the potential role of heparin on c-myb, this molecule is expressed in the late G1 phase of the cell cycle and likely represents only one of several defects when growth is inhibited.142 Equally intriguing are the role of heparin in inhibiting migration and suggestions from Lindner and Reidy45 that a major action of heparin may be to wash bFGF out of the injured vessel wall after the mitogen is released from dying cells. Unfortunately, clinical trials using heparin to prevent restenosis have been disappointing.143 144 Ellis et al143 administered intravenous heparin to patients over the first 18 to 24 hours after angioplasty and found no difference in the restenosis rates of patients treated with heparin compared with those given a dextrose infusion (41% versus 37%, respectively; P=NS). Similarly, a subsequent attempt to limit restenosis with a single daily subcutaneous injection of 10 000 IU heparin was halted prematurely because of the higher rates of restenosis and clinical events in the heparin treatment group compared with the usual care group.144 This lack of clinical benefit is difficult to explain but may be related to a rebound coagulation phenomenon associated with interrupted anticoagulant treatments.145 Edelman and Karnovsky146 suggest that differences in heparin dose scheduling may be critical. For example, the antiproliferative effect of heparin requires that the drug be administered for at least 4 to 7 days after injury in the rat carotid balloon injury model.32 Furthermore, cell proliferation and the ratio of the intimal to medial area are made worse when rats are treated with heparin dosages and administration schedules similar to those used clinically.146 Low-molecular-weight heparins also appear to be ineffective for the prevention of restenosis.147
The failure of antiproliferative therapies based on
angiotensin, somatostatin, and heparin may be telling us
that we should consider other mechanisms as being critical to
restenosis. A clue as to such mechanisms may have arisen
serendipitously from a recent clinical trial directed at the thrombotic
events occurring immediately after angioplasty. As part of the EPIC
study, the chimeric monoclonal antibody Fab fragment (c7E3), which is
directed against the platelet glycoprotein IIb/IIIa
receptor, was administered to patients undergoing angioplasty or
atherectomy who were at high risk for ischemic
complications.148 149 150 This therapy resulted in a reduction
in acute ischemic complications (eg, nonfatal myocardial
infarction and emergency revascularization
procedures), although at the risk of increased bleeding complications.
Although it is easy to imagine that these trials are affecting only
platelets, these antagonists to
glycoprotein IIb/ IIIa, a platelet adhesive protein,
may also affect a closely related integrin,
vß3. Antagonists to
vß3 have recently been shown to block
smooth muscle migration in vitro and intimal formation in
vivo.151 Moreover, we have found that this same receptor
is required for movement of smooth muscle cells in response to
osteopontin, an abundant and specific marker of intimal smooth muscle
cells in the rat152 and in human
plaque.153 154 Therapies targeted against integrin
receptors involved in platelet aggregation and cell migration
represent an intriguing new direction.
|
Role of Intima in the Formation of an Atherosclerotic Lesion |
|---|
|
TopIntroductionFormation of the IntimaNeointimal Formation: A Generic...Pharmacology of Neointimal...Potential Clinical Targets Based...Role of Intima in...Special Properties of...Intimal Cells Overexpress...Mechanisms Controlling Plaque...Pathological Significance of...Mechanism of Lumen Occlusion...The Nature of Restenosis...Relevance of Animal Stenosis...Future DirectionsReferences |
|---|
Stary et al155 note that the initial lipid accumulation occurs deep in the wall, within the preexisting intimal mass. The intimal mass is not defined as atherosclerosis, since it has not yet accumulated lipid and is seen as a normal feature of vessels in our species. Aside from such semantic issues, localization and initiation of atherosclerosis in intimal masses are reminiscent of experiments in balloon-injured animals in which lesions develop selectively at sites of previous injury and intimal formation.37 Thus, atherosclerosis may begin in the "soil" of the intima. The peculiar properties of the intima that lead to such a phenomenon have not been identified. However, there have been suggestions that local accumulations of specific glycosaminoglycans related to chondroitin sulfate may promote local lipid accumulation.156 157
The time course of intimal smooth muscle proliferation in relation to lipid accumulation in humans is not known. Ross47 158 has suggested that proliferation of smooth muscle cells is important in advanced lesions. This idea, however, is not supported by cell kinetic studies measuring either the frequency of cells able to incorporate labeled thymidine ex vivo or the number of cells identified as replicative on the basis of staining for a cell cycle–specific marker.159 160 The low replicative rates seen in advanced lesions correlate with studies of cells cultured from lesions. Moss and Benditt161 were the first to culture these cells. They found that plaque smooth muscle cells have a greatly shortened life span relative to normal medial cells. This observation has been reproduced by others.162 163 More recently, we have extended this observation by looking at apoptosis. Plaque smooth muscle cells, at least in vitro, have a very high rate of spontaneous cell death. This high level of apoptosis may account for the apparent short replicative life span and suggests that plaque cells have undergone some as-yet-undefined injury in vivo.164
Interestingly, the only actual evidence that smooth muscle proliferation occurs in atherosclerosis depends on evidence that the lesions are monoclonal or oligoclonal. The original observation of clonality by Benditt and Benditt5 has been reproduced by two other groups.2 3 4 It is difficult to imagine how clonality could arise unless proliferation plays a critical role in an event at some time in the ontogeny of the lesions.
Hansson et al165 have recently shown that lymphocytes in lesions are polyclonal. The remaining cells in these lesions are macrophage, smooth muscle, and endothelial cells. Although plaque macrophage and endothelial cells do replicate,159 166 there are no known examples of these cell types forming monoclonal growths other than in neoplasms.47 159 Thus, monoclonality implies that smooth muscle proliferation must occur during the formation of the lesion and that the initial group of cells giving rise to the lesion must be very small. Such an early expansion of intimal smooth muscle mass has been described exactly to occur in the proximal left anterior descending coronary artery, a common site for occlusive coronary artery lesions in adults.14 17 If the increase in mass described in the newborn left anterior descending coronary artery correlates with smooth muscle replication, then very few further doublings may be required to account for the mass of smooth muscle seen in an adult lesion.
In summary, the available direct evidence suggests that smooth muscle replication occurs very early in the formation of atherosclerotic lesions.14 17 We cannot rule out the possibility that smooth muscle replication occurs at a very low rate over several years or that replication occurs at a high rate in an episodic fashion.
|
Special Properties of Atherosclerotic Intima |
|---|
|
TopIntroductionFormation of the IntimaNeointimal Formation: A Generic...Pharmacology of Neointimal...Potential Clinical Targets Based...Role of Intima in...Special Properties of...Intimal Cells Overexpress...Mechanisms Controlling Plaque...Pathological Significance of...Mechanism of Lumen Occlusion...The Nature of Restenosis...Relevance of Animal Stenosis...Future DirectionsReferences |
|---|
Whether or not the atherosclerotic plaque is hyperproliferative, the lesions are progressive, and it seems obvious that progression must depend on unique properties either of the intima or of the plaque itself.167 Table 1
is an attempt to
collate those properties with special regard to genes that are
underexpressed or overexpressed by the intimal or plaque smooth muscle
cell relative to medial smooth muscle cells. This table emphasizes rat
and human data, the species about which the most is known. Some rabbit
data are also included. In addition, the table includes data from
cultured smooth muscle cells, since the preservation of properties in
vitro identifies intimal smooth muscle cells that may be differently
differentiated from medial smooth muscle cells.
Table 1 also lists a large number of molecules whose expression is
characteristically decreased by intimal cells compared with medial
cells. Loss of expression of the genes associated with the fully
differentiated, "contractile" phenotype was first
described by Campbell and Campbell168 and has been
confirmed by others.169 170 171 Campbell and Campbell have
given the change from a contractile to a "synthetic"
phenotype the name "modulation." The term synthetic is
used because cultured cells lose much of their microfilaments and, as
seen by transmission electron microscopy, acquire an extensive rough
endoplasmic reticulum. Because modulation coincides with the onset of
replication in vitro, several workers have proposed that special
properties of the modulated smooth muscle cell may include
proliferation and synthesis of matrix molecules that contribute to
formation of the intima and the plaque.168 169 170 171 Molecules
lost in modulated cells include smooth muscle myosin, desmin,
caldesmon, and -actin.10 172
Unfortunately, the term synthetic has been loosely applied to cells
lacking a smooth muscle phenotype, because intimal cells as
well as many cells in the media may lack the contractile
phenotype without showing evidence of an active synthetic
apparatus169 173 (D. Gordon and S.M. Schwartz,
unpublished data involving the internal mammary artery, 1994).
Inversely, smooth muscle cells can rapidly leave quiescence in the
vascular wall after balloon injury, enter S phase, and still express
abundant -actin mRNA.174 Thus, the equation of
synthetic phenotype with replication is unconvincing.
Similarly, the media of avian aortas contains a second cell type that,
while apparently quiescent, lacks morphological features of the
contractile smooth muscle cell.175 176 177 Thus, loss of the
contractile phenotype should be distinguished from the onset of
proliferation or change to a phenotype associated with high
levels of protein synthesis.
An excellent example of the difference between a vessel with only a media and one with an atherosclerotic intima is the issue of vascular contractility. Normally, the vessel wall exists in a relaxed state due to the endogenous production of NO by endothelium. NO-dependent relaxation, however, is greatly impaired in atherosclerotic vessels.178 This loss of function is not due to a change in the smooth muscle of the media. Loss of NO function has been attributed to endothelial injury or to the inactivation of NO by free radicals produced in the plaque.179 In contrast, Joly et al180 described the appearance of the inducible form of NO synthase that follows balloon injury. Similarly, as described above, the neointima overexpresses angiotensin receptors, and the plaque has been shown to overexpress endothelin.181 Although we have not yet distinguished the effects of atherosclerosis from those of the more general appearance of the neointima, our point is simply that we might expect the vasomotor activity of a diseased artery with an intima to be very different from vasomotor regulation in a normal artery. The contribution of this altered pharmacology to the ability of the vessel wall to maintain a normal lumen caliber is largely unexplored.
In summary, Table 1 points out that intima and media are quite
different tissues. At least some genes are overexpressed in the intima,
even in the absence of atherosclerosis; for example,
the human intima overexpresses the ß1 chain of laminin and the ED-A
splice form of fibronectin.182 183 In the rat, a large
number of genes are overexpressed after injury, but so far only
tenascin has been shown to be overexpressed in the few intimal cells
seen in the normal rat artery.184 185 Tenascin, ß1
laminin, and ED fibronectin are important, since their expression
suggests that intimal smooth muscle cells rather than merely lacking
the properties of medial smooth muscle may have their own unique
properties. We have already discussed the possible significance of
overexpression of NO synthase, AT-1 angiotensin receptors,
proteoglycans, PDGF, or TGF-ß in the ontogeny of smooth muscle
proliferation and lipid accumulation. Differences in NO
metabolism186 or AT-1
receptors,87 to cite just two markers, are likely to mean
that vessels with an intima have very different pharmacology than
vessels with only a media. A critical issue, as we will see below, is
whether the unique properties of intimal cells represent a
unique cell type with heritable properties that distinguish intimal
smooth muscle from medial smooth muscle.
|
Intimal Cells Overexpress Certain Genes In Vitro as Well as In Vivo |
|---|
|
TopIntroductionFormation of the IntimaNeointimal Formation: A Generic...Pharmacology of Neointimal...Potential Clinical Targets Based...Role of Intima in...Special Properties of...Intimal Cells Overexpress...Mechanisms Controlling Plaque...Pathological Significance of...Mechanism of Lumen Occlusion...The Nature of Restenosis...Relevance of Animal Stenosis...Future DirectionsReferences |
|---|
Cell culture studies suggest that intimal cells may be differentiated from the typical medial smooth muscle cell. Smooth muscle cell lines cultured from 2-week-old rat pups have several unique properties. These include an epithelioid morphology, the ability to grow without platelet-released growth factors, and secretion of PDGF.187 188 These same properties appear when neointimal cells, isolated 2 weeks after arterial injury, are placed in culture.105 When the same cells were studied for gene expression, a group of genes was found that shows overexpression or even unique expression in pup and intimal cells compared with normal medial cells cultured from adult arteries.187 189 190 This collection of properties and genes led us to suggest that the vessel wall, at least in the rat, contains two types of cells:
and µ cells. Genes overexpressed by
cells include PDGF-B chain, CYPIA1, elastin, and
osteopontin.187 189 190
Unlike the distinction between contractile and synthetic
phenotypes, which is lost when all cells put in culture become
synthetic, the differences between the µ phenotype and
phenotype seem to be maintained in passage, and cells with
or µ properties can be isolated by cloning cells from mixed
cultures.191 192 Moreover, some of these genes, although
first identified in vitro, are also overexpressed or uniquely expressed
in the neointima in vivo. At least for the rat, we suggest
that smooth muscle cells contain two distinct "cell types," one
of which takes a special part in the formation of the
neointima.
Comparable evidence has not yet been discovered for subsets in human smooth muscle. Cultured human fetal smooth muscle cells also differ from adult smooth muscle cells in their expression of two homeobox genes, HoxB7 and HoxC9. These genes are also seen in pup cells but not in rat cells, however. HoxB7 and HoxC9 are not seen in human intima or atherosclerotic plaques (J.M. Miano, A.B. Firulli, E.N. Olson, P. Hara, C.M. Giachelli, S.M. Schwartz, unpublished data, 1994). However, as already noted, human intima preserves certain genes that are also seen in the fetal vessel.10 Moreover, if atherosclerotic lesions are monoclonal, the origin of lesions in the intima suggests that lesions are derived from a unique subset of vessel wall smooth muscle cells localized to this layer.5
|
Mechanisms Controlling Plaque-Specific Gene Expression by Smooth Muscle Cells |
|---|
|
TopIntroductionFormation of the IntimaNeointimal Formation: A Generic...Pharmacology of Neointimal...Potential Clinical Targets Based...Role of Intima in...Special Properties of...Intimal Cells Overexpress...Mechanisms Controlling Plaque...Pathological Significance of...Mechanism of Lumen Occlusion...The Nature of Restenosis...Relevance of Animal Stenosis...Future DirectionsReferences |
|---|
Overexpression of certain molecules by intimal cells might be the result of two quite different, but not necessarily independent, mechanisms. First, as suggested above, intimal cells might belong to a unique lineage either derived from a common precursor or differentiated from typical smooth muscle cells. The detection of stable
and µ
phenotypes in vitro supports this hypothesis. Alternatively,
and more obviously, plaque cells may show overexpression or repression
of certain molecules as a result of mediators present in the plaque
environment. Particularly important among such mediators are oxidation
products or more traditional inflammatory
mediators.193 194 Collins195 has suggested
that a common factor linking inflammation and oxidation is the role of
nuclear factor-
B as a trans-acting factor induced by
oxidized radicals and by many cytokines.195 Table 1 includes plaque-specific genes whose overexpression might be
explained as an inflammatory response. In some cases, however, this may
be misleading, since overexpression in the plaque is permanent, whereas
overexpression in response to cytokines is often only
transient. Distinguishing between extrinsic and endogenous
controls of gene expression is likely to depend on cell culture results
showing stable expression, as is evident for some genes in Table 1, as
well as promoter analyses taking place in the individual
labs. |
Pathological Significance of Plaque-Specific or Intimal-Specific Genes |
|---|
|
TopIntroductionFormation of the IntimaNeointimal Formation: A Generic...Pharmacology of Neointimal...Potential Clinical Targets Based...Role of Intima in...Special Properties of...Intimal Cells Overexpress...Mechanisms Controlling Plaque...Pathological Significance of...Mechanism of Lumen Occlusion...The Nature of Restenosis...Relevance of Animal Stenosis...Future DirectionsReferences |
|---|
Among the more interesting genes shown to be overexpressed in plaques are tissue factor, MCSF, GMCSF, osteopontin, and bone morphogenic protein.
Tissue Factor
The normal arterial wall contains little or no tissue
factor. The time course for the appearance of tissue factor in plaques
is not known, however. Tissue factor is prominent in cells of the
plaque, including smooth muscle cells. Davies and
Thomas196 and others197 198 199 200 have
suggested that thrombosis is a critical event in plaque progression.
Thus, it is likely that accumulation of tissue factor is critical to
the morbidity of lesions. Moreover, Taubman201 has shown
that tissue factor mRNA expression was elevated after balloon injury in
rats. This was not surprising, since tissue factor was known from in
vitro studies to be a component of the early phases of the cell cycle.
Perhaps more surprisingly, he found an elevation of tissue factor
activity, implying that a procoagulant response is part of the
formation of neointima.
MCSF/GMCSF
The presence of leukocyte growth factors MCSF and GMCSF as well as
receptors for leukocyte factors (see Table 1) is of special importance
because of growing evidence that plaque macrophage, rather than
plaque smooth muscle cells, may constitute the major unique
proliferative element in the plaque.159 160 The unique
properties of the proliferative plaque macrophage have yet to
be explored.
Osteopontin and Bone Morphogenic Protein
Osteopontin and bone morphogenetic protein 2a have both been found
in smooth muscle cells of atherosclerotic
plaques.154 202 203 The presence of these molecules is
supportive of the hypothesis that plaques are derived from unique
subsets of smooth muscle cells, because osteopontin was first found in
neointimal cells and in the rat pup artery as part of the
phenotype.190 Bone morphogenetic protein 2a,
on the other hand, is also seen only in plaques. Intriguingly, cells
found to express bone morphogenetic protein 2a also express
immunocytochemical markers associated with a special form of smooth
muscle cell, the pericytes seen around small vessels, again suggesting
a unique lineage for plaque smooth muscle compared with medial smooth
muscle.203 The association of osteopontin and bone
morphogenetic protein 2a with areas of calcification in the plaque also
suggests that the mechanisms of bone mineralization may also play a
role in vessel wall calcification.
|
Mechanism of Lumen Occlusion in Atherosclerosis |
|---|
|
TopIntroductionFormation of the IntimaNeointimal Formation: A Generic...Pharmacology of Neointimal...Potential Clinical Targets Based...Role of Intima in...Special Properties of...Intimal Cells Overexpress...Mechanisms Controlling Plaque...Pathological Significance of...Mechanism of Lumen Occlusion...The Nature of Restenosis...Relevance of Animal Stenosis...Future DirectionsReferences |
|---|
Vascular lumen size can be regulated either by rapid changes in smooth muscle tone or by structural rearrangement of cells and extracellular matrix components of the wall. The latter mechanisms include "remodeling" and "elastic recoil." For example, Glagov et al204 noted that human vessels can undergo massive accumulations of atherosclerotic mass without lumen narrowing. The vessel compensates for the new mass by undergoing a compensatory redistribution of vessel wall mass so that lumen size is maintained. This remodeling permits a normal level of blood flow until an adaptational limit is exceeded. This limit appears to occur when
40% or more of the area bounded by the internal elastic lamina is
occupied by intimal mass. Decreased lumen size in
atherosclerosis may depend on pharmacological
mechanisms associated with the neointima or the plaque that
cause a failure of normal remodeling events. In contrast, elastic
recoil, an expression often used to describe the rapid loss of lumen
gain after angioplasty, describes the resistance of the vessel wall
component to the mechanical stretch imposed by the inflated balloon
catheter. Remodeling is a normal response that allows the vessel to maintain normal levels of blood flow and wall stress and can be seen in small muscular arteries as well as in large elastic arteries.205 206 Branching patterns of conduit vessels seem to be genetically determined.207 Thus, the only way arteries can respond to a demand for an increase or a decrease in blood flow is by changing vessel caliber. These changes, in some cases, occur by rearrangement of existing vessel wall components rather than synthesis of new mass or cell replication.208
Despite the observations of Glagov et al204 of arterial remodeling during atherogenesis in human coronary arteries (above), remodeling has been largely ignored in animal models of arterial injury and repair. The extent of remodeling versus loss of lumen due to intimal formation is illustrated by an important experiment by Jamal et al.206 When Jamal et al balloon-injured the rabbit carotid artery, they found the expected neointimal hyperplasia. However, the animals showed no narrowing of the lumen despite a 100% increase in total wall thickness attributable to the neointima. In contrast, the same study showed a significant (14%) narrowing in response to an experimental restriction of flow in the vicinity of the thyroid artery where endothelium had regenerated. This effect of the endothelium in an injured vessel is consistent with observations that structural adaptation to changes in flow requires an endothelium.206 Since reendothelialization generally correlates with a diminution of intimal thickening, these data may even imply a negative correlation of intimal mass with luminal narrowing.72
The role of the endothelium in controlling lumen size may be especially important given the often neglected fact that plaques have a prominent microcirculation that comprises capillaries arising from the adventitia.209 210 211 The role of these small vessels in regulating structural change or, for that matter, contractile properties of the atherosclerotic vessel remains unexplored.
If the correlation of intimal (or plaque) mass with lumen caliber is not a simple one, how do we account for angiographic changes seen after aggressive lipid-lowering therapy?212 213 214 It is important to realize that the reported clinical benefit of these studies far exceeds the improvement of stenosis diameter. The improvement in lumen size in these studies is very small (eg, 0.7% to 5.3%, or an increase of 0.003 to 0.117 mm in minimum absolute diameter).212 214 To put these results in perspective, one should note that 6 months after percutaneous coronary angioplasty, there is an average improvement of 16% diameter stenosis units and a 0.47-mm increase in minimum absolute diameter.215 Moreover, even these modest changes in lumen diameter could reflect changes in adherent thrombotic material or state of vasospasm rather than a decrease in mass of the atherosclerotic intima.216 Rather than opening the lumen, lipid-lowering therapies may improve clinical outcome by stabilizing the lesion.217 Davies and colleagues196 218 219 and others199 200 have suggested that the formation of fissures in plaques is the critical step leading to vascular occlusion. Davies' hypothesis would imply that we may be able to develop diagnostic tests, based on plaque composition as assessed by magnetic resonance imaging, IVUS, or even gene expression patterns in atherectomy specimens, that might indicate lesion prognosis or the effectiveness of drugs targeted at stabilizing the lesion.
These same tests may allow us to estimate changes in vessel wall mass much as the advent of roentgenography led to estimates of tumor cell growth. The ability to study the progress of human lesions may be critical to our understanding of how lesions progress and eventually kill. Animal models with advanced disease are rare, with the exception of the spontaneously hypercholesterolemic swine.220 Autopsy data in humans are also confusing. Serial angiographic studies in humans suggest that the majority of myocardial infarctions may occur because of thrombotic occlusion of arteries that previously did not contain hemodynamically significant stenoses (eg, <50%).221 222 However, postmortem studies do not bear this out. For example, Qiao and colleagues223 224 have reported that in both native coronary arteries and saphenous venous bypass grafts, atherosclerotic plaque rupture with thrombosis most commonly occurred at sites with severe narrowing (eg, >90% area stenosis).
|
The Nature of Restenosis Following Angioplasty |
|---|
|
TopIntroductionFormation of the IntimaNeointimal Formation: A Generic...Pharmacology of Neointimal...Potential Clinical Targets Based...Role of Intima in...Special Properties of...Intimal Cells Overexpress...Mechanisms Controlling Plaque...Pathological Significance of...Mechanism of Lumen Occlusion...The Nature of Restenosis...Relevance of Animal Stenosis...Future DirectionsReferences |
|---|
Fig 2
shows the tissue that characterizes
coronary artery restenosis as seen in atherectomy
specimens or at autopsy. Despite use of the word "proliferative"
by pathologists to describe this tissue,225 as we have
already suggested, it may be entirely incorrect to assume that the
formation of a neointima in a balloon-injured rat carotid
artery is a model for the loss of arterial caliber seen
after angioplasty, ie, "restenosis."
|
