Apoptosis (programmed cell death) of vascular smooth muscle cells (VSMCs) has recently been identified as an important process in a variety of human vascular diseases, including atherosclerosis, arterial injury, and restenosis after angioplasty. VSMC apoptosis is regulated by interactions between the local cell–cell and cytokine environment within the arterial wall, and the expression of pro- and anti-apoptotic proteins by the cell, including death receptors, proto-oncogenes and tumour suppressor genes. This review summarises our current knowledge of the occurrence and mechanisms underlying VSMC apoptosis in atherosclerosis and arterial remodelling.
Time for primary review 29 days.
Apoptosis (programmed cell death) of vascular smooth muscle cells (VSMCs) has recently been identified in physiological remodelling of the vasculature, and also in disease states such as atherosclerosis and restenosis after angioplasty. The implication of this observation is that cell death, together with cell proliferation, migration and matrix turnover, may contribute to changes in vascular architecture in development and disease. Furthermore, cell death in the vessel wall may be tightly regulated by both specific cytokines and gene products. This review focuses on the role of VSMC apoptosis in two areas of vascular disease, that of arterial remodelling and of atherosclerotic plaque rupture.
Remodelling can be defined as a change in calibre of the vessel, often with little or no change in overall tissue mass. Remodelling can thus be viewed as net movement of tissue towards or away from the lumen, which can decrease or increase luminal diameter respectively . The contribution remodelling makes to vessel calibre changes can be assessed by measuring the circumference of a defined portion of the vessel, for example the area circumscribed by the internal or external elastic lamina, by either post mortem studies or by intravascular ultrasound. Thus, if a reduction in circumference of the vessel occurs, without an increase in intimal area, then the luminal reduction can be wholly ascribed to constrictive remodelling of the vessel wall.
Remodelling is both a physiological and pathological process in vessel wall dynamics. Thus, compensatory enlargement of an artery occurs as the atherosclerotic plaque within the vessel enlarges [2, 3]. Only when the ability of the vessel to remodel is exceeded does the atherosclerotic plaque narrow the lumen [2, 3]. Similarly, a reduction in blood flow through a vessel is accompanied by constrictive remodelling of that vessel, whether the reduction in blood flow is due to surgery or physiological changes in flow after birth.
Constrictive remodelling of an artery has recently been shown to be the major process involved in restenosis after angioplasty (Fig. 1). In many animal models of injury to normal or atherosclerotic vessels, up to 80% of lumen stenosis is due to remodelling, with little increase in intimal tissue mass or area [4–6]. Indeed, intimal mass does not correlate with angiographic restenosis in many studies and vessels that underwent restenosis had similar intimal areas to those that did not [5, 7, 8]. Restenosis following angioplasty to diseased vessels in non-human primates has also been shown to be due to remodelling rather than neointima formation . In human disease, necropsy studies indicate that 40% of restenotic lesions show no increase in neointimal mass . Furthermore, intravascular ultrasound has allowed direct quantification of the contributions made by neointima formation and remodelling of the vessel after angioplasty. In a landmark study, Mintz et al demonstrated that after conventional angioplasty, 73% of the loss of lumen is due to constrictive remodelling, and only 23% due to neointima formation . Profound changes in vessel calibre and vessel wall mass also occur in closure of the ductus arteriosus after birth, hypertrophy of the artery wall in hypertension and regression of hypertrophy after treatment of hypertension. Whilst there are significant changes in absolute vessel wall mass in all of these conditions, significant remodelling of the arteries also occurs.
3 VSMC apoptosis in remodelling
The potential biological processes involved in remodelling are many, including VSMC proliferation, adventitial fibrosis and migration of adventitial myofibroblasts, [12–14], net production of extracellular matrix at one surface accompanied by matrix degradation further away [15, 16], reorganisation of mural thrombus, and resolution of inflammation after injury. However, apoptosis of VSMCs is also a major determinant of VSMC number in remodelling. For example, profound reductions in VSMC number and vessel calibre are found in physiological remodelling after birth, or after ligation of a distal artery [17–19]. In these vessels, there may be no difference in rates of VSMC proliferation in vessels that remodel compared with vessels that do not; the reduction in VSMC number in remodelling arteries is entirely due to differences in apoptotic rates. Thus, VSMC apoptosis is seen as the major process governing VSMC number in these models of remodelling. Of note, these studies [17–19]also define the time taken for apoptosis of VSMCs to occur in the vessel wall. Pulse-chase labelling indicates that VSMCs undergo apoptosis and are cleared within 4 h of markers of apoptosis occurring. Thus, VSMC apoptosis is a rapid method of reducing VSMC mass in the vessel wall. The remodelling seen in closure of the ductus arteriosus is accompanied by VSMC apoptosis, in co-ordination with changes in VSMC proliferation and matrix synthesis . Medial VSMC apoptosis also occurs in the regression of vascular hypertrophy during drug treatment of hypertension in spontaneously hypertensive rats . Clearly, VSMC apoptosis can exert major changes in arterial architecture, either alone, or in co-ordination with VSMC proliferation and matrix turnover.
In arterial injury, where the long term response may be accompanied by remodelling, VSMC apoptosis is detectable in two phases. Medial VSMC apoptosis is a rapid event after injury, with loss of VSMCs occurring within 30 min of injury [22, 23]. Apoptosis of intimal VSMCs (and medial VSMCs to a lesser extent) also occurs in the later response to injury, from about 8–21 days in the rat carotid artery model [24, 25]. Injury to a pre-existing intima will also induce medial VSMC apoptosis, over the same 30 min to 4 h time course seen for primary injury . Low levels of apoptosis may also occur chronically in this model, as there is continuing proliferation in injured vessels . In humans, apoptotic VSMCs are seen in the restenosis lesion after angioplasty at a greater frequency than that seen in the normal vessel wall . However, it is not known if VSMC apoptosis occurs in humans immediately after arterial injury, or if it occurs during the remodelling process prior to the patient presenting with restenosis.
4 Regulators of VSMC apoptosis in remodelling
Apoptosis is regulated by a complex interplay between cell surface signals, such as those from death receptors and survival cytokines, and the expression of specific intracellular gene products (for instance, those encoded by specific proto-oncogenes, tumour suppressor genes and the Bcl-2 family of genes) (Fig. 2). The TNF-receptor (TNF-R) family of death receptors are type I membrane proteins comprising TNF-R1 (p55), Fas (CD95) and death receptors (DR)-3, 4 (TRAIL R1) and 5 (TRAIL R2). Family members all consist of an extracellular domain, a hydrophobic transmembrane domain, and a cytoplasmic domain, containing the death domain, a protein motif responsible for protein: protein interactions with adapter molecules. Ligand binding to receptor recruits an adapter molecule, (FADD (to Fas), TRADD (to TNF-R1) or RIP (to both)), to the receptor, which then activates a cascade of cysteine proteases (caspases) leading to cell death. The induction of apoptosis via death receptors is rapid, occurring within a few hours of ligand binding, and does not require new RNA or protein synthesis [28, 29]. Some ligands for these receptors are widely expressed (TNF-α, TRAIL). However, Fas-L is more restricted, particularly to lymphocytes, monocyte/macrophages, endothelial cells or VSMCs.
Apoptosis is also regulated by the expression of specific intracellular proteins, such as the Bcl-2 family of proteins, of which there are an increasing number. The prototype family member, Bcl-2, is a membrane protein which protects against a wide variety of stimuli which induce apoptosis. Bcl-2 family members can homodimerise or heterodimerise with other family members, such as the pro-apoptotic proteins Bax, Bad or Bik, and the balance of expression between family members can predispose or protect against apoptosis. In particular, Bcl-2 family members play an important role in the release of cytochrome C, a potent pro-apoptotic stimulus, from the mitochondria of cells undergoing apoptosis. Apaf-1, a human homologue of a nematode death gene, ced 4, binds to cytochrome c to induce activation of caspases . In this way, mitochondrial release of cytochrome c during apoptosis is linked to caspase activation.
In general, apoptosis proceeds in 2 separate phases, the ‘decision’ phase and the ‘execution’ phase. The decision phase involves the integration of pro and anti-apoptotic signals from the cell surface and from within the cell. If the balance of these signals favours apoptosis, the execution phase is triggered by the activation of the caspase cascade which result in the cleavage of intracellular proteins and the final disintegration of the cell. VSMCs express many (if not all) of the known caspase family members required to execute apoptosis [31, 32]. However, the importance of any particular pathway regulating apoptosis depends upon the cell type and the stimulus for apoptosis. Thus, different pathways may be responsible for inducing apoptosis in VSMCs in different disease states, in VSMCs at different stages of the disease, or in VSMCs compared with endothelial cells.
Physiological remodelling after flow reduction requires the presence of the endothelium . Although the endothelium produces a number of agents which may regulate VSMC apoptosis, the most intensely studied agent is nitric oxide (NO)(Fig. 2). NO induces apoptosis in a number of cell types [34–40], most likely acting via peroxynitrite . In addition, cytokines such as interleukin 1-β, interferon gamma and tumour necrosis factor-α can induce apoptosis via induction of NO [34, 35]. The downstream mediators of NO-induced apoptosis are not completely known, but they include cyclic GMP , the tumour suppressor gene p53 [39, 42, 43]and the cell death receptor Fas (CD95) [36, 44, 45]. Thus, cytokines such as IL-1β may induce NO synthesis within VSMCs, which upregulates p53, which in turn transcriptionally activates Fas. Nitric oxide can also down regulate Bcl-2, which normally protects against apoptosis [46, 47]and upregulate Bax, which promotes apoptosis, possibly via induction of p53 . Thus, NO alters the Bcl-2/Bax ratio in favour of apoptosis.
In contrast, NO may directly inhibit apoptosis by inhibiting the activity of components of the caspase cascade [48–50]. This appears to be due to specific S-nitrosylation of Cys 163, a functionally essential amino acid conserved among caspase-like proteases [49, 50]. NO also reduces the level of intracellular glutathione and the activity of NO to regulate apoptosis or to nitrosylate caspases can be reversed by antioxidants such as glutathione [43, 51], indicating that NO-induced apoptosis may be under the influence of cellular redox status. This is not surprising as NO modulates intracellular thiol redox states and the thiol redox state of the cell influences NO production. The fact that the caspase enzymes are the final common path for most (if not all) apoptotic deaths indicate that NO may be an inhibitor of apoptosis whatever the stimulus . The dual roles of nitric oxide to induce or protect against apoptosis is also emphasised by the finding that different concentrations of NO can increase or decrease p53 DNA binding (and thus transcriptional activation of p53-induced apoptosis genes such as Bax) . However, a word of caution is necessary in interpreting many of the published studies on the regulation of apoptosis by NO. At present, the physiological role of NO in mediating apoptosis of VSMCs in vivo is unproved. In addition, many of the published in vitro studies use a combination of NO donors or methods of generating NO which may have effects other than that ascribed to NO.
In addition to NO, a number of other apoptosis genes have been implicated in VSMC apoptosis in remodelling, particularly after arterial injury. Bcl-xL, an anti-apoptotic member of the bcl-2 family is seen only in intimal VSMCs , and injury is associated with a loss of bcl-xL expression . Down regulation of bcl-xL in rabbit atheromatous lesions or ballooned rabbit carotid arteries has been shown both to induce apoptosis and promote lesion regression . The role of anti-oxidants and the redox state of the cells in apoptosis has also been emphasised by studies showing that glutathione levels are rapidly reduced after arterial injury when medial VSMC apoptosis occurs. Furthermore, this loss of glutathione and the consequent medial VSMC apoptosis can be prevented by administration of anti-oxidants .
Finally, VSMC apoptosis in remodelling may occur via signalling through death receptors such as Fas. Endothelial cells express Fas and Fas ligand, and VSMCs express Fas. Exposure of VSMCs to Fas ligand, either soluble Fas-L or Fas-L on the surface of endothelial cells may therefore induce VSMC apoptosis. Indeed, adenovirus delivery of Fas ligand to an injured artery induces medial VSMC apoptosis and profoundly reduces neointima formation . The potency of this effect was explained by the observation that a bystander killing could be demonstrated, and also via the fact that VSMCs expressing ectopic Fas ligand could also induce apoptosis in invading inflammatory cells which express Fas, reducing any inflammatory infiltrate.
5 VSMC apoptosis in plaque rupture
Apoptosis of vascular smooth muscle cells leading to plaque rupture is an attractive concept to account for the relative numbers of VSMCs and macrophages found in unstable plaques. Rupture sites are characterised by their paucity of VSMCs and their high concentration of macrophages and inflammatory cells . Apoptosis of VSMCs has been detected in the shoulder regions of plaques, the sites which appear most likely to rupture , suggesting that apoptosis of VSMCs may predispose to rupture. In addition, preliminary evidence indicates that there may be higher rates of apoptosis in plaques in unstable versus stable angina patients . However, at present we lack conclusive evidence for the role of VSMC apoptosis in plaque rupture, partly because we lack good animal models of plaque rupture and partly because of the technical difficulties in unequivocally identifying apoptotic VSMCs in a complex tissue such as the atherosclerotic plaque.
Examination of advanced human plaques for apoptosis has revealed remarkably high indices, with some studies reporting up to 40% of cells showing TUNEL positivity [25, 27, 57]. Whilst many of these apoptotic cells are not VSMCs, these figures are remarkable for a tissue with little proliferative activity [59, 60]. Subsequent studies have indicated that the true apoptotic index in the plaque may be of the order of 1% , as much of the TUNEL positivity may be artefactual due to TUNEL identifying non-cellular DNA-containing fragments in the plaque, or RNA synthesis or splicing unrelated to apoptosis [61, 62]. These studies [61, 62]also indicated that the vast majority of apoptosis is concentrated around the lipid core region of the plaque, and may occur in macrophages, not VSMCs. A 1% level of apoptosis may still indicate that apoptosis is important in regulating plaque cellularity. In a tissue with little proliferative activity, 1% apoptosis would markedly reduce lesion cellularity. Furthermore, direct induction of apoptosis has been shown to reduce lesion cellularity and size in rabbit models of atherosclerosis [23, 55]. The consequences of VSMC apoptosis in plaques should also be considered. Although apoptosis may reduce lesion cellularity and promote rupture, apoptosis may also promote thrombosis over the plaque directly. Both apoptotic endothelial cells and VSMCs can act as a substrate for the generation of thrombin [63, 64]. Indeed, plaque VSMCs undergoing apoptosis have the same potency to generate thrombin as platelets . This is due to the fact that apoptotic cells expose phosphatidylserine on the surface early in the process . In the presence of Factors V and VII, exposed PS can then act as a substrate for thrombin generation. At present, it is difficult to know how potent this mechanism of thrombin generation is in vivo. Significant thrombosis has not been seen in animal models where apoptosis has been induced in the intima or media, although thrombin activity has not been directly analysed [23, 55].
6 Regulators of VSMC apoptosis in atherosclerotic plaques
VSMCs in the atherosclerotic plaque reside in a complex milieu of pro and anti-apoptotic agents, with apoptosis being governed by both cell–cell, cell–matrix and cell–cytokine interactions. Macrophage and T lymphocyte cytokines can induce apoptosis of human and rat VSMCs in culture , possibly via NO induction of p53 and Fas . Human VSMCs express Fas, and T lymphocytes, macrophages and endothelial cells express Fas ligand , so that Fas/Fas ligand-mediated apoptosis may be triggered in VSMCs at sites of high inflammatory cell content. VSMCs also express TNF-α , and macrophages can synthesise TNF-α, so that TNF-mediated apoptosis may occur in plaques. T-lymphocytes also express Granzyme B, and can kill via both Granzyme B/perforin and Fas-mediated pathways. Thus, there are multiple pathways which can be activated in VSMCs in plaques which may induce apoptosis. What is harder to ascertain is the relative potencies and importance of these pathways.
The level of surface expression of Fas or TNF-R1 is low in human VSMCs compared with haemopoietic cells (MRB-unpublished observations). Indeed, Fas ligand alone or a cross-linking (agonistic) anti-Fas antibody induces minimal apoptosis in VSMCs in the absence of cycloheximide (MRB, unpubl. data). TNF-α, Il-1β or IFN-γ alone are also unable to induce apoptosis of VSMCs . A combination of TNF-α, IL-1β and IFN-γ can induce apoptosis of VSMCs, albeit at quite high concentrations. However, this combination of cytokines can sensitise VSMCs to Fas-mediated apoptosis, via upregulation of Fas on the cell surface . It is thus highly probable that one pathway leading to apoptosis can prime another pathway, increasing the response to a particular apoptotic stimulus. Equally, the convergence of many pathways leading to apoptosis on a conserved caspase cascade indicates that additive and synergistic effects of many agents to promote apoptosis is likely.
Apoptosis is also regulated both positively and negatively by conventional growth factors. Thus, insulin-like growth factor (IGF)-1, and platelet-derived growth factor (PDGF), are survival factors for subconfluent human VSMCs, from both atherosclerotic plaques or normal vessels [68, 69]. Basic fibroblast growth factor has also been shown to be a survival factor for VSMCs , and delivery of antisense bFGF can inhibit neointima formation after injury in the rabbit . However, the role of these growth factors in the plaque is more difficult to ascertain. Plaques show increased expression of both PDGF and IGF-1 compared with normal vessels [72, 73]which would be predicted to protect against apoptosis and promote cell proliferation. However, as mentioned above, there is relatively little proliferative activity in plaques and apoptosis of VSMCs is present despite high levels of growth factors. One explanation for this observation has come from the finding that serum-starved confluent cells undergo apoptosis upon reintroduction of growth factors, possibly due to an aborted attempt at entering the cell cycle . There is extensive evidence now that VSMCs in advanced atherosclerotic plaques exhibit a ‘senescent’ phenotype, characterised by an inability to proliferate, and markers of senescence [68, 75, 76]. When plaque VSMCs are forced to enter the cell cycle they undergo apoptosis rather than successful replication . Thus, it is possible that the effect of mitogens acting on senescent VSMCs in the plaque is to induce apoptosis rather than cell proliferation.
The expression of apoptosis-regulatory molecules within plaque VSMCs may also be different from that in VSMCs in normal vessels. Cultured plaque VSMCs show spontaneous apoptosis in high serum-containing medium . These high rates of apoptosis are not lost by subculture, indicating that this is an intrinsic property of the cells. Plaque VSMCs are also very sensitive to p53-mediated apoptosis . Although the absolute expression and transcriptional activity of p53 is similar to that seen in normal VSMCs in culture, plaque VSMCs in vivo may overexpress p53 , possibly due to the extracellular milieu of inflammatory cytokines such as Il-1β and TNF-α. This increased sensitivity to p53 is dependent upon growth status, but is independent of new RNA or protein synthesis. Thus, growth arrest and the inability to phosphorylate the retinoblastoma protein protect plaque VSMCs from apoptosis [76, 77], and p53 can also regulate apoptosis when VSMCs are driven to proliferate . However, it should be noted that overexpression of p53 in normal rat, rabbit or human VSMCs does not induce apoptosis in culture without DNA damage [77, 79, 80]and overexpression of p53 after balloon injury to the rabbit carotid also does not elicit significant VSMC apoptosis . Thus, it appears likely that the effects of p53 to elicit significant apoptosis may be limited to VSMCs from plaques. If so, the mechanism of this effect may provide illuminating information on the biology of plaque versus normal medial VSMCs.
Other potential regulators in plaque VSMC apoptosis include the Bcl-2 family of proteins. Although there is no difference in expression of Bcl-2 in cultured plaque versus normal VSMCs in culture , atherosclerotic plaques also exhibit reduced Bcl-2 to Bax immunocytochemical staining in vivo, suggesting that the balance of these pro and anti-apoptotic gene products may be shifted towards apoptosis [81, 82].
Apoptosis within the atherosclerotic plaque may also be regulated by cell–matrix interactions. The presence of the extracellular matrix prevents apoptosis in many cell types, via specific integrin-mediated signalling [83–86]. Clearly, matrix metalloproteinases (MMPs) which degrade extracellular matrix may disrupt this cell–matrix interaction, and therefore promote apoptosis. MMPs may also directly promote apoptosis by releasing both soluble Fas ligand and membrane-bound TNF-α from cells [88, 89]. MMPs are produced by both VSMCs and macrophages and are detectable at vulnerable regions of the plaque [90, 91]. Thus, MMP production by macrophages at vulnerable regions may induce VSMC apoptosis, predisposing to plaque rupture. However, MMPs are normally inhibited by specific tissue inhibitors (TIMPs), which would be predicted to inhibit MMP-induced apoptosis. In contrast, recent research has indicated that TIMP-3 may directly promote apoptosis in VSMCs, both in the cells expressing TIMP-3 and bystander cells . The mechanism of this effect is not known.
Finally, the role of circulating factors in inducing apoptosis in plaques should not be underestimated. Low density lipoprotein can induce apoptosis in VSMCs, particularly oxidised LDL [93–95], which may localise with apoptotic VSMCs in human atherosclerotic plaques . The effect of oxidised LDL has been shown to occur particularly via ketocholesterol, possibly via the downregulation of Bcl-2 [94, 97]. Local production of reactive oxygen species may also induce VSMC apoptosis, predominantly due to hydrogen peroxide and the hydroxyl radical . Conversely, anti-oxidants such as pyrrolidinedithiocarbamate and n-acetylcysteine can induce apoptosis in both human and rat VSMCs . Clearly, the effect of individual agents may depend upon the redox state of the cell under study.
Recent research has indicated that apoptosis of vascular smooth muscle cells is a widespread phenomenon, responsible for mediating profound changes in arterial architecture. The regulation of apoptosis in the vessel wall is complex, and likely to consist of multiple interacting pathways within the cell, in addition to cell-cell and cell-matrix interactions. The consequences of VSMC apoptosis in vivo are presently unknown, although recent studies suggest that apoptosis may reduce the neointimal response to injury. The important regulators of VSMC apoptosis are also only just emerging, although intervention against these regulators offers the prospect of selective induction or repression of VSMC apoptosis in a variety of disease states.
MRB is supported by a British Heart Foundation Senior Fellowship.