An integrated approach for the mechanisms responsible for atherosclerotic plaque regression

Atherosclerosis is the leading cause of death worldwide. Plaque regression – the return of the arterial wall to its initial state – is a complex process involving specific cellular and molecular pathways that are able to mobilize all pathological components of the plaque. High-density lipoproteins, endothelial cells and vascular smooth muscle cells play important roles in the progression and regression of atherosclerosis. This article describes many aspects of the atherosclerotic process and integrates the information to provide a comprehensive review of plaque regression – an understanding of which is essential to developing targeted pharmaceutical treatment strategies.

Abstract

Atherosclerosis was originally considered to be an ongoing process that was inevitably associated with age. However, plaques are highly dynamic, and are able to progress, stabilize or regress depending on their surrounding milieu. A great deal of research attention has been focused on understanding the involvement of high-density lipoprotein in atherosclerotic plaque regression. However, atherosclerotic plaque regression encompasses a variety of processes that can be grouped into three main areas: removal of lipids and necrotic material; restoration of endothelial function and repair of denuded areas; and cessation of vascular smooth muscle cell proliferation and phenotype reversal. In addition to the role of high-density lipoproteins in lipid removal, resident macrophages and foam cells are able to regain motility and rapidly migrate on milieu improvement, moving both lipids and necrotic material to regional lymph nodes. Neighbouring endothelial cells can proliferate and replace dead and dysfunctional cells. Circulating endothelial progenitor cells can similarly restore vessel function. Finally, abrogation of smooth muscle cell proliferation occurs secondarily to these processes. This information is integrated in the current article to present a comprehensive and clear depiction of plaque regression. This integrated view of regression is essential to optimize the pharmaceutical targeting of the many processes and pathways involved in plaque regression.

Atherosclerosis emerged as a major health dilemma during the early 20th century. Nearly a full century later, atherosclerosis has become the leading cause of death worldwide (1,2). It is now understood that atherosclerotic lesions are produced from a complex array of molecular and cellular events that act in concert to form asymmetric focal thickenings of the arterial tunica intima (3,4).

Atherosclerosis is initiated by the response of endothelial cells (ECs) to injury caused by myriad noxious stimuli including hyperglycemia, hypertension, hyperlipidemia, infectious agents, obesity, modified lipoproteins, homocysteine, nicotine, free radicals, altered changes in arterial blood flow shear stress and normal spontaneous metabolic damage (5–10). This initial damage induces a loss of basal endothelial homeostasis causing endothelial dysfunction (11). The resulting increase in endothelial permeability permits the accumulation of low-density lipoprotein (LDL) and cellular debris within the tunica intima of the vessel wall, eventually leading to endothelial activation. Once activated, ECs produce an array of chemoattractant cytokines such as monocyte chemoattractant protein-1, macrophage colony-stimulating factor, interferon-γ, platelet-endothelial cell adhesion molecule-1, interleukin (IL)-1, IL-6 and tumour necrosis factor-α (TNF-α), creating a proinflammatory environment that attracts circulating monocytes and T lymphocytes (12–14). Normally, ECs do not express molecules that facilitate the adhesion of circulating leukocytes. However, activated ECs express vascular cell adhesion molecules such as intercellular adhesion molecules (ICAMs), E-selectin and P-selectin, which mediate leukocyte adhesion and infiltration (12,13). Endothelial-derived cytokines subsequently drive the differentiation of monocytes into macrophages, which use pattern recognition receptors to sequester LDL, modified LDL, free cholesterol (FC) and cholesteryl esters (3). Together, activated leukocytes and ECs continue to produce proinflammatory cytokines and growth factors that promote the transition of smooth muscle cells from a quiescent, contractile phenotype to an active and proliferative synthetic phenotype that deposits extracellular matrix at the site of injury, forming a fibrous cap (15–18). Continued exposure to various noxious stimuli, in conjunction with a proinflammatory environment, further exacerbates plaque severity, perpetuates plaque progression, and promotes plaque destabilization, resulting in plaque rupture which manifest as acute events such as stroke or myocardial infarction (19,20).

During the past few decades, an improved understanding of atherosclerotic pathophysiology has enabled researchers to consider the possibility of inducing plaque regression in addition to the more conventional therapeutic goal of reducing plaque progression. Traditionally, atherosclerosis was viewed as an inevitable unidirectional process that begins in childhood and manifests during adulthood (20). However, during the 1980s, Badimon et al (21) showed that plaque development is not simply a permanent process associated with age, but rather a dynamic process that can be slowed, stopped or reversed. In spite of the evidence supporting the existence of plaque regression, the concept of ‘plaque regression’ was resisted for decades. Critics of regression asserted that the physical attributes of an atherosclerotic plaque, including calcification, necrosis and fibrosis, indicated stability, and the molecular processes associated with plaques such as oxidative damage, cell proliferation and transformation seemed to imply that regression would be very difficult or impossible to achieve. Furthermore, at that time, a majority of animal studies attempting to induce regression in advanced plaques had failed (22–26). However, the ability of even advanced lesions to regress on robust reductions in LDL levels and increases in high-density lipoprotein (HDLs) levels has established plaque regression as a well-accepted process (27,28). Atherosclerosis is now known as a highly dynamic process in which plaques are able to either progress or regress, depending on the surrounding milieu (29,30). However, despite the plethora of research, a comprehensive discussion that incorporates and integrates the current knowledge of the various processes known to occur during plaque regression remains to be completed.

What is atherosclerotic plaque regression? Ideally, plaque regression would be the return of the arterial wall to its initial state. However, with the exception of fatty streaks, even under ideal conditions, the complete reversal of plaque formation does not occur (30). It is important to understand that plaque regression is not simply plaque progression in reverse. Rather, the mechanisms of plaque regression are distinct and are composed of three important processes: the reduction or clearance of necrotic and extracellular material from the tunica intima; endothelial repair, regeneration and return to homeostasis; and the cessation of smooth muscle cell proliferation (Figure 1). Although a reduction of luminal area, plaque volume or intimal medial thickness are indexes commonly used in diagnostic medicine to assess plaque regression, and are possible estimations of regression, they are currently unable to capture the true complexity of this process.

Figure 1).

The integrated mechanism of atherosclerotic plaque regression. The mechanism of regression is complex and involves lipoproteins, endothelial progenitor cells (EPCs), macrophages and vascular smooth muscle cell (SMC). The transport of cholesterol back to the liver begins with the synthesis and secretion of lipid-poor apolipoprotein (apo) AI by the liver and intestine. The first step of reverse cholesterol transport (RCT) is initiated by the ATP-binding cassette transporter (ABCA) A1-mediated efflux of cholesterol from peripheral cells to apoAI or preβ-high-density lipoprotein (HDL). Preβ-HDL matures to HDL₃ after lecithin:cholesterol acyltransferase (LCAT) esterifies its cholesterol producing cholesterol ester. Cholesteryl ester transferase (CEPT) can further mature HDL₃ into HDL₂ through the transfer of its cholesterol esters to apoB-containing lipoproteins (low-density lipoprotein [LDL], very low LDL [VLDL] and intermediate-density lipoprotein [IDL]). Mature HDL particles may then be cleared from the circulation by the liver or intestine via the scavenger receptor class B1 (SR-B1) receptor, which mediates holoparticle uptake. Lipases can release apoAI from HDL, which are filtered and degraded by the kidneys. HDL can also unload cholesterol into the liver or intestine after it has acquired apoE, which enables HDL to interact with the LDL receptor. Finally, ABCA1 located on enterocytes causes the unloading of cholesterol from HDL particles. To complete reverse cholesterol transport, cholesterol is secreted into bile by the liver or secreted directly into the intestinal lumen by the apical member transporters scavenger receptor class B type 1 (SR-B1), ABCG5 and ABCG8. Both the intestine and liver remove cholesterol from the body as feces. During regression, damaged, denuded areas of the vessels’ endothelial lining or damaged endothelial cells can be replaced by circulating bone marrow derived EPCs. Additionally, the amelioration of the pathological components of the plaque will abrogate vascular SMC migration and proliferation, and induces migration of macrophages away from the plaque. HDL production, maturation and degradation are indicated by black arrows and lines. Removal of cholesterol from peripheral tissue is indicated by red arrows and lines. Clearance of cholesterol within circulation is indicated by blue arrows and lines. Movement of EPCs is indicated by purple arrows and lines. oxLDL Oxidized LDL

REMOVAL OF EXTRACELLULAR AND INTRACELLULAR LIPIDS AND NECROTIC MATERIAL

Although the mechanisms of regression are beginning to be understood, there has been considerable progress in isolating the important players in regression and understanding their biological role. Attention, however, within the past few decades has been focused on reverse cholesterol transport (RCT) and its major player, HDL.

HDL and its influence on atherosclerotic plaque regression though RCT

HDL classification:

Since the early 1970s, HDL cholesterol (HDL-C) has been associated with a decrease in cardiovascular disease (CVD) (31). Gordon et al (32) demonstrated that there is an inverse relationship between systemic levels of HDL-C and the prospective risk for coronary artery disease; this correlation was maintained even at very low levels of LDL cholesterol. Animal and human studies have shown that infusions of synthetic HDL particles can induce regression of atherosclerotic plaque (21,33,34). Moreover, increases in HDL-C by even 1 mg/dL are associated with a 3% to 4% decrease in cardiovascular-related mortality (35). Since the discovery of the inverse relationship of HDL-C with CVD, a large amount of information has been obtained about its structure and function. Although the full mechanism of HDL’s anti-atherosclerotic effects are not well known, it is becoming clear that HDL participates in ameliorating nearly all aspects of plaque pathogenicity by restoring endothelial cell basal function, decreasing smooth muscle cell proliferation and migration, and decreasing intraplaque inflammation and cellular debris retention (36). It has also become clear that not only is the concentration of HDL within the circulation important for its anti-inflammatory effect, but also its quality (37). Nevertheless, what has been well established is HDL’s contribution to atherosclerotic plaque regression by promoting RCT. However, to understand the RCT pathway, it is essential to become aware of the various structures of HDL.

HDLs were first discovered by Gofman et al (38) in experimental studies using analytical ultracentrifugation to separate and categorize distinct serum lipoprotein families according to their hydrated density. HDLs were defined as lipoproteins of small diameter (5 nm to 17 nm) that sediment in the density region (1.063 g/mL to 1.21 g/mL) (Table 1). It is now understood that HDL is a highly heterogeneous group of particles, and composed of many subclasses. The difference between HDL subclasses is ultimately dependent on the particles’ physical properties including shape, size, density and charge; however, these subclasses can be reclassified according to differences in lipid transfer proteins, enzymes, lipid content and apolipoproteins (39,40). Many methods have been developed to separate HDL into its various subclasses. Gofman et al (38) developed an analytical ultracentrifugation method for identifying the distinct subclasses of HDL particles, which is currently the gold standard technique for HDL subclass determination (41,42). Using this technique and other ultracentrifugation methods, such as rate-zonal or single vertical spin ultracentrifugation, HDLs can be separated into two main subfractions based on their respective densities: HDL₂ (1.063 g/mL to 1.125 g/mL) and HDL₃ (>1.125 g/mL to 1.21 g/mL) (43–45). These two HDL subfractions can also be further subdivided based on their size (diameter) using nondenaturing polyacrylamide gel electrophoresis, yielding HDL_2b (9.7 nm to 12.0 nm), HDL_2a (8.8 nm to 9.7 nm), HDL_3a (8.2 nm to 8.8 nm), HDL_3b (7.8 nm to 8.2 nm) and HDL_3c (7.2 nm to 7.8 nm) (46). Another classification system of HDL arose with the use of agarose gel electrophoresis, separating HDL based on relative negative surface charge density. α-HDL, which comprises the majority of plasma HDL, has a high negative charge density compared with preβ-HDL, also known as nascent HDL (47–49). Two-dimensional electrophoresis, which combines both agarose gel and gradient gel electrophoresis, can further resolve α-HDL and preβ-HDL into 12 distinct HDL subclasses based on their size and motilities with respect to albumin: preβ (preβ1, preβ2 and preβ3), α (α1, α2 and α3) and preα (preα1, preα2 and preα3) (48,50). Finally, HDL can be classified by immunoaffinity methods according to its apolipoprotein (apo) A composition. The majority of proteins found in HDL are either apoAI (65% to 70%) and apoAII (20% to 25%). Furthermore, two major apoA-containing HDL subclasses exist: HDL particles may contain both apoAI and apoAII in a 2:1 molar ratio, or they can contain only apoAI (50,51). Adding to HDL’s complexity, several proteins have been shown to circulate in the plasma bound to HDL including apoA-IV, apoC, apoE, apoF, lecithin:cholesterol acyltransferase, cholesteryl ester transferase (CETP), phospholipid transferase protein, paraoxonase and platelet-activating factor acetylhydrolase. Moreover, proteomic analyses have shown that HDL may contain more than 48 proteins. It is generally accepted that the majority of plasma HDL is in the form of preβ-HDL, as well as HDL₂ and HDL₃ – the mature forms of HDL that are equivalent to α-HDL (52,53) (Table 1). Despite the many different classifications of HDL, all HDL particles have in common their ability to mediate the process of RCT.

TABLE 1.

High-density lipoprotein (HDL) characteristics

HDL particle	Density, g/mL	Preferred efflux pathway	Electrophoretic mobility	Subpopulations	Diameter, nm
Preβ-HDL	1.21	ABCA1	Preβ	Preβ1	5.4–7
		Aqueous diffusion		Preβ2	12–14
				Preβ3
HDL2	1.063–1.125	SR-BI	α	HDL2b	9.7–12.0
		ABCGI		HDL2a	8.8–9.7
		Aqueous diffusion
HDL3	1.125–1.21	SR-BI	α	HDL3a	8.2–8.8
		ABCGI		HDL3b	7.8–8.2
		Aqueous diffusion		HDL3c	7.2–7.8

ABC ATP-binding cassette transporter; SR-BI Scavenger receptor class B type I

RCT:

According to the generally accepted definition of RCT, the first step of RCT is the efflux of extrahepatic cholesterol from intracellular cholesterol pools. This usually focuses on the removal of cholesterol from macrophages. Alternatively, critics of this view attest that assembly of HDL particles and their subsequent secretion into the circulation is the initial step of this process (54–56). In the present review, however, the overview of HDL’s involvement in RCT will begin with the production and release of HDL to provide a clear synopsis of RCT. Despite the disagreement regarding the initial steps in RCT, it is accepted that the acquisition of cholesterol from peripheral stores by HDL is the major mechanism for promoting atherosclerotic plaque regression.

The production of all HDL subclasses begins with the synthesis of apoAI and apoAII. These proteins are primarily generated in the liver and secondarily by the small intestine (56,57). Both hepatocytes and enterocytes can secrete lipid-poor apoAI and AII into the circulation, where it acquires phospholipids and cholesterol to form preβ-HDL (58–60). Regardless, the majority of apoA produced in the intestine is packaged into chylomicrons, which are eventually internalized by the liver. In contrast, hepatocyte apoA secretion into plasma is more complex. Similar to chylomicrons, very low density lipoprotein (VLDL) – another triglyceride-rich lipoprotein – can lose cholesterol. Phospholipids can also lose or gain apolipoproteins during lipolysis, which generates preβ-HDL. VLDL can also be amalgamated into HDL₂ and HDL₃ (61). Finally, hepatocytes can synthesize and secrete both apoA and preβ-HDL directly into the circulation. In spite of the different modes of release, once in the circulation, preβ-HDL can be used for RCT.

The first and rate-limiting steps of RCT is cholesterol efflux from macrophages, and the peripheral tissue to the HDL particle (62). There are currently four proposed mechanisms of cholesterol efflux into lipid-poor apoAI and the various HDL particles: aqueous diffusion, scavenger receptor class B type 1 (SR-BI)-mediated efflux, ATP-binding cassette (ABC) transporter A1-mediated efflux and ABCG1-mediated efflux (54,55).

First, cholesterol efflux can occur passively by adhering to its concentration gradient. As a result, the passive diffusion pathway is bidirectional, and the rate-limiting step of this process is dependent on the ratios and interactions between the FC in the donor cells and the phospholipids within the plasma membrane of the acceptor molecule – the HDL particle (62,63). The kinetics of this interaction are accelerated by higher proportions of unsaturated phospholipids and decreases in the sphingomyelin content of the plasma membrane (64,65). Because cholesterol efflux by this pathway is not affected by HDL particle size, all HDL subclasses are equal in their ability to accept cholesterol through aqueous diffusion (66).

In contrast to aqueous diffusion, the remaining three cholesterol efflux pathways are mediated by protein-protein interactions. Cholesterol efflux can be mediated by the ATP-independent, passive SR-BI pathway. Cholesterol transport is initiated by the binding of HDL to SR-BI, which subsequently forms a hydrophobic channel allowing efflux to occur (67). SR-BI promotes efflux exclusively to larger, more mature HDL particles such as HDL₂ and HDL₃, which can further increase the SR-BI-mediated selective uptake of cholesterol if they contain apoAII (68). Because this process is bidirectional, HDL may be a donor and an acceptor of cholesterol. Thus, its role in aiding RCT in macrophages may not be as important as the remaining energy-dependent cholesterol efflux pathways (69).

Cholesterol and phospholipid efflux to apoAI is promoted by the interaction of ABCA1 with preβ-HDL or with lipid-poor apoAI (49). The activation of ABCA1 induces the removal of cholesterol from the plasma membrane of the donor cell and can also cause efflux from intracellular cholesterol pools (70,71). The rate of lipid efflux is dependent on the concentration of activated ABCA1 and on the amount of unsaturated phospholipids in the donor cell (72). In addition, the concentration of ABCA1 in the plasma membrane is regulated by the interaction of apoA1 and ABCA1, which prevents its intracellular degradation (73).

Another member of the ABC family of transporters, ABCG1, is the last major efflux pathway in RCT. ABCG1, which is involved with regulating intracellular cholesterol homeostasis, interacts with all subclasses of HDL (74–76). It also mediates cholesterol efflux through the reorganization of existing cholesterol pools, and increasing its diffusion to and uptake by HDL particles. Although all HDL subclasses can mediate efflux through the ABCG1 pathway, larger HDL particles are more effective acceptors, as in the SR-BI efflux pathway (77).

Following the removal of cholesterol from peripheral sites, HDL undergoes further maturation characterized by increased particle size due to the progressive accumulation of cholesterol (78). In circulation, cholesterol becomes esterified to preβ-HDL through an apoA1-mediated reaction with lecithin:cholesterol acyltransferase, converting lecithin and cholesterol into lysolecithin and CE (79). Because CE are hydrophobic, they can move to the developing core of the HDL particle, which causes it to enlarge and become spherical. This ultimately transforms preβ-HDL into HDL₃ or HDL₂. This process not only increases the surface area for FC acceptance, it also helps maintain the FC gradient (69,80). Once a mature HDL particle is synthesized, it has three possible fates. First, lipolytic enzymes such as lipoprotein, hepatic and endothelial lipases can hydrolyze triglycerides and phospholipids causing apoAI to dissociate from HDL, which facilitates the clearing of apoAI from circulation by the kidneys and the conversion of mature HDL particles back into nascent preβ-HDL (81–83). Second, constituents of mature HDL particles may be transported to other HDL subclasses or other lipoproteins. For example, certain plasma lipid transfer proteins can transfer triacylglycerol and phospholipids among lipoproteins (84). CETP transfers the majority of CEs to apoB-containing lipoproteins, such as VLDL and LDL, which produces small cholesterol ester-depleted HDL particles that are either cleared by the kidneys or incorporated into another HDL particle forming a large mature HDL particle (85). ApoB-containing lipoproteins can be removed from the circulation by hepatic LDL receptors and catabolized by the liver, and hepatic SR-BI can also bind LDL and VLDL. Thus, the CETP pathway is believed to be an important route of RCT (86–90). However, clinical studies involving humans with CETP deficiencies report conflicting and controversial results, which ultimately has led to questioning of whether CETP is detrimental or instrumental in CVD (19,91–94). It is also likely that cholesterol transferred to LDL or VLDL can be redeposited to peripheral tissue. Third, cholesterol can be returned to the liver through the clearance of HDL from the circulation. SR-BI, located on the surface of liver or adrenal cells, recognizes circulating HDL particles and facilitates the unloading of their FC and cholesterol esters (95). This process does not involve the internalization and degradation of the HDL particle. Thus, removal of cholesterol and CEs produces nascent HDL particles that have the potential to be recycled back into the RCT pathway (85). As opposed to HDL efflux from macrophages, SB-RI is the most abundant receptor on hepatocytes, thus making this pathway an important mode of cholesterol influx. An alternative method of influx involves the acquisition of apoE enabling HDL to deliver cholesterol to hepatic tissue through its interaction with LDL receptors, which mediates the internalization of HDL particles and their subsequent catabolism (96). Novel routes to HDL holoparticle uptake by hepatocytes, such as HDL internalization by the receptor P2Y13, continue to be identified (97).

Cholesterol within hepatocytes must be secreted into the intestinal lumen to complete the RCT pathway. The rate of cholesterol excretion is dependent on several transporters on its apical membrane. Cholesterol efflux into bile is dependent on both SR-BI and the ABC family of transporters. The rate of cholesterol efflux is primarily controlled by ABCG5 and ABCG8 transporters (98,99). Although SR-BI can also facilitate cholesterol secretion into bile, it serves only a minor role on the apical membrane, being secondary to the ABC transporters. However, SR-BI is an important mode of influx on the basolateral membrane of hepatocytes (100,101).

The hepatobiliary route of HDL-C removal is generally accepted to be the major mode of cholesterol excretion. However, Cheng and Stanley (102) suggested that there is an additional route to fecal cholesterol loss through a nonbiliary fecal route. Nonbiliary sterol loss is believed to be regulated by the liver. Under conditions in which the hepatic cholesterol load exceeds a manageable amount, cholesterol is repackaged into plasma lipoproteins that are bound for secretion by the intestines (103). Although the mechanism of the nonbiliary fecal pathway is not well understood, many aspects are shared with the hepatobiliary route of sterol loss. It has been suggested that lipoproteins responsible for the transport of cholesterol to the intestine are either apoE-rich HDL or apoB-containing lipoproteins (104,105). It is quite likely that most lipoproteins are able to facilitate this process because enterocytes express SB-RI on their basolateral membrane (78). Similar to hepatocytes, enterocytes express ABCG5/G8 as well as SR-BI on their apical membranes, allowing the efflux of cholesterol from the enterocyte to the intestinal lumen (106).

Monocyte/macrophage migration

During regression, cholesterol removal is accompanied by the disappearance of macrophage and foam cells. This is one of the first noticeable indications of plaque regression (30). It was initially believed that the disappearance of monocytes and macrophages from atherosclerotic plaques was only caused by apoptosis, lysis in situ and phagocytosis by new macrophages. However, it is now evident that macrophages within atherosclerotic lesions can also regain motility and migrate to regional lymph nodes when the local environment improves.

Mechanisms of cell motility and immobilization:

Cells facilitate migration through cell motility cycling. This process contains four sequential events beginning with the following: polarization of the cell; extension of lamellipodia at the leading edge; the formation of a focal adhesion attaching the cell to the underlying substrate; and the contraction and rear detachment of the cell resulting in cell movement. The regulation of these events is coordinated by the Rho family of GTPase (107). The exchange of GDP for GTP activates Rho-GTPases, which control cytoskeleton remodelling by activating downstream targets (108). Two members of this family, RhoA – a GTPase associated with actin filament organization and promotion of focal adhesions – and Rac1 – which induces polymerization of actin that forms both lamellipodia and membrane ruffles – have important roles in cell immobilization (109,110).

During atherosclerosis, macrophages within atherosclerotic plaques accumulate FC and CEs (111–113). Accumulation of cholesterol causes apoptosis and secondary necrosis in macrophages and impairs chemotaxis. This immobilization is associated with increased levels of Rac-GTP, the active form of Rac, and reduced levels of RhoA. The translocation of Rac to the plasma membrane facilitates its activation. Cholesterol-rich domains within the inner leaflet of the plasma membrane maintain Rac-GTP in the membrane. The loss of proper cell polarization, cell spreading and plasma membrane ruffling caused by increased Rac activity abrogates forward movement of the cell (114–116).

Mechanisms of macrophage migration during plaque regression:

The migration of macrophage and foam cells from atherosclerotic plaques is complex and is controlled by plaque dynamics. Improvements in the plaque milieu causes the transformation of macrophages from an immobilized to a mobile state by the expression of dendritic cell markers (117,118) such as chemokine (C-C motif) receptor 7 (CCR7), an essential requirement for dendrite cell migration (119). CCR7 control of immune cell emigration is mediated by the activation and upregulation of the liver X receptor (28). The downstream target of the liver X receptor, ABCA1, is also unregulated (118). Macrophage/foam cell migration and morphological transformation can be, in part, facilitated by ABCA1-induced cholesterol redistribution. ABCA1 may decrease membrane cholesterol pools releasing Rac-GTP (120,121). ABCG1 and LDL receptor-related protein 1 may also play a small role in macrophage migration from atherosclerotic plaques (119,122).

ENDOTHELIAL REPAIR, REGENERATION AND HOMEOSTASIS

Vascular ECs serve as a dynamic barrier separating the vessel wall from blood. They also control vascular homeostasis by regulating vascular tone, leukocyte trafficking and vessel permeability. Endothelial dysfunction – an early hallmark of atherosclerosis – is characterized by the conversion of ECs from a normal physiological phenotype to a vasoconstrictive, procoagulant, platelet-activating phenotype that contributes to atherosclerosis (123,124). Prolonged endothelial dysfunction can lead to cell apoptosis, which can ultimately denude the vessel wall (125,126). To achieve atherosclerotic regression, these dysfunctional ECs must be returned to basal homeostasis and dead cells need to be replaced. Currently, there are three known mechanisms that can result in endothelial cell replacement: circulating endothelial progenitor cells, local endothelial cell proliferation and migration, and abrogation of endothelial apoptosis.

Endothelial progenitor cells and vascular repair

The term ‘stem cell’ or ‘progenitor cell’ refers to immature cells that have the ability to self-renew and differentiate into a variety of cell types. Thus, these cells have the potential to restore the function of damaged tissues (127,128). Endothelial progenitor cells (EPCs), similar to all progenitor cells, are lineage specific, and comprise a highly heterogeneous population of cells capable of differentiating exclusively into ECs (129). The majority of progenitor cells mature from hemato-poietic stem cells. These stem cells are mainly isolated from bone marrow, peripheral blood and umbilical cord, but can be obtained from the spleen, intestine, liver, adipose tissue and adventitia (130). Regardless of their source, all hematopoietic stem cells are CD34⁺ and CD133⁺. Hematopoietic stem cells can also produce nonerythroid myeloid and granulocyte-macrophage lineages, as well as EPCs. Therefore, EPCs are characterized by the co-expression of both hematopoietic stem cell markers and endothelial markers such as vascular endothelial growth factor receptor-2, CD31, endothelial nitric oxide (NO) synthase, and vascular endothelial cadherin (130–134). Although hematopoietic stem cells appear to be the main source of EPCs, other resident bone marrow stem cells, such as mesenchymal stem cells, may generate EPCs. In addition, CD14⁺ myeloid subsets express both the hematopoietic and endothelial markers, and CD14⁺ monocytic cells can differentiate into ECs (135–138). EPCs are instrumental in maintaining the integrity of the vascular endothelium, which contribute to the regression of atherosclerotic plaques.

To obtain any benefit from nontissue resident or exogenously administered EPCs, they must be recruited by and migrate to the site of injury. This process is orchestrated by resident cells of the injured area. During atherosclerosis, various cell types within the plaque are capable of mobilizing and homing EPCs to denudated vessels. First, EPCs are released from their source upon stimulation from molecular signals produced by immune cells within the plaque. Activated M2 type macrophages promote vessel healing through the secretion of granulocyte colony-stimulating factor (G-CSF) (139). G-CSF facilitates the release of EPCs into circulation. In addition, cytokine-mediated release of proteases such as elastase, cathepsin G and matrix metalloproteinase (MMP)-9 discharges EPCs by cleaving the adhesive interaction between EPCs and stromal cells (134,140). The released EPCs are subsequently homed to and mobilized by the injured area. An important homing signal is the chemokine stromal cell-derived factor-1 (SDF-1). Under normal conditions, the bone marrow and many other tissues constitutively express SDF-1. The bone marrow, however, produces a gradient that favours the retention of EPCs. During conditions of ischemia, inflammation and hypoxia, this gradient is reversed by the expression of hypoxia-inducible factor-1, which upregulates SDF-1 in injured tissues (141–144). In addition, NO, estrogen, HDL, vascular endothelial growth factor and erythropoietin contribute to the increase in the plasma titre of EPCs and their recruitment to the site of injury by augmenting the phosphatidyl-inositol-3-phosphate (PIP3)/Akt pathway (145–147). Once in the damaged area, cell adhesion molecules, such as P/E-selectin and ICAM-1, mediate the binding of EPCs to the injured endothelium (148). In severely damaged vessels, exposed matrix proteins activate platelets that adhere to the denuded area. Platelet activation causes microthrombi formation and the expression of SDF-1, targeting EPCs to the damaged endothelium. Moreover, EPCs potentially adhere not only to the endothelium, but also to platelets by interacting with P-selectin and GPIIb integrin (149–151). After EPC attachment, they differentiate into ECs under the influence of the laminar shear stress of the blood (128,152).

Several studies have evaluated the beneficial effects of EPCs on CVD. Bone marrow-derived EPCs can differentiate into ECs and induce neovascularization in ischemic mouse tissue (153,154). There is an inverse relationship between circulating (CD34⁺ and CD133⁺) progenitor cells for both modifiable and nonmodifiable atherosclerotic risk factors, such as smoking and age (155). Moreover, low circulating levels of EPCs are found in patients who are at high risk of developing coronary artery disease (156). Statin therapy, known to reduce the incidence of CVD and cardiovascular events, was able to increase EPC titre and engraftment efficiency (157,158). Physical exercise increases NO and HDL production, both of which increase EPC recruitment and EPC levels (159–161).

Despite the mounting evidence implicating progenitor cell involvement in the regression of atherosclerosis, there are concerns and caution has been advised. Bone marrow-derived progenitor cells have been associated with neovascularization and vessel remodelling, which can cause plaque destabilization (130,162). In addition, EPCs that have the osteoblast marker osteocalcin are retained within the lesion for endothelial repair, but are also believed to lead to the induction and progression of coronary calcification rather than normal endothelial repair (163).

EC apoptosis, migration and proliferation

As atherosclerosis persists, ECs become apoptotic, leading to the denudation of the vessel wall. Apoptosis is known to be initiated by the activation of the death receptor and mitochondrial-mediated apoptotic pathways. HDL may have a role in preventing EC apoptosis and promoting endothelial repair (164). HDL can directly or indirectly inhibit endothelial apoptosis either by decreasing the levels of TNF-α, oxidized LDL and growth factors, or by inhibiting apoptotic pathways. These anti-apoptotic properties are mediated by constituents within HDL. ApoA1, an important protein of the RCT pathways, diminishes oxidized LDL and TNF-α induced apoptosis (164–166). Lysosphingolipids within HDL, such as sphingosylphonphorylcholine and lysosulfatide, can inhibit growth factor-induced apoptosis as well as directly interfere with apoptotic signalling within ECs by activating the Akt signalling pathways (167–169).

Endothelial migration and proliferation is also promoted by HDL. Endothelial migration is believed to be stimulated by the interaction between SR-BI and HDL (157). Alternatively, sphingosine-1-phosphate within HDL can induce endothelial cell migration by activating Ras/Raf1-dependent ERK (157,167). The induction of EC proliferation by HDL is mediated by downregulating ADAMTS-1, by activating the protein kinase C pathway and by increasing intracellular Ca²⁺ levels through phospholipase C activation (170–172).

SMOOTH MUSCLE CELL PROLIFERATION AND REGRESSION

Although there are four known smooth muscle cell subpopulations (173–175), focus has been placed on understanding two morphologically distinct subpopulations. The majority of vascular smooth muscle cells (VSMCs) normally found in the arterial media are elongated and spindle shaped, forming the classic ‘hills and valley’ growth pattern. These ‘swirling-type’ VSMCs have a contractile phenotype. VSMCs with the synthetic phenotype, also known as the ‘epithelioid-type’, are only present in small proportions within the arterial media. The synthetic phenotype of VSMC is cuboidal in shape and displays a cobblestone appearance at confluence. These VSMCs are believed to be the major VSMC contributor to atherosclerotic plaque progression (176–178). Akin to macrophages, accumulation of intercellular lipids can change VSMCs into foam cells, which may be considered a third important phenotype of VSMC found in atherosclerotic plaques. Moreover, de-differentiation of VSMC decreases its affinity to HDL, thus effectively decreasing apo-mediated cholesterol efflux and contributing to the early events of foam cell transformation (179).

Despite the large amount of evidence linking VSMC to atherosclerotic plaque progression, little is known about their involvement in atherosclerotic plaque regression. Nevertheless, regression can be induced by the removal of lipids from sterol-loaded VSMC and by the cessation of VSMC proliferation, which is accompanied by the reversal of the pathological phenotypic modulation. Sterol unloading from VSMCs shares many similarities with macrophage sterol removal. SR-BI, a mediator of cholesterol efflux in macrophage and influx in hepatocytes are also expressed in VSMCs (180). Although synthetic VSMCs have a reduced affinity for NO, NO is still capable of inhibiting smooth muscle cell proliferation under cell culture conditions. Supplementation with L-arginine, the precursor of NO, has also been shown to induce plaque regression in cholesterol-fed rabbits (181). Bioactive fatty acids such as prostacyclin I1 and prostaglandin E2 maintain the contractile phenotype of VSMCs and play a major role in reducing migration and proliferation of VSMCs through activation of peroxisome proliferator-activated receptors (PPARs). Activation of both PPAR-γ and PPAR-α leads to the suppression of proinflammatory cytokines by inhibiting the activity of NF-κB (182,183). In addition, PPAR-α inhibits proliferation through a p16/pRb/E2F-mediated suppression of telomerase activity. PPAR-δ activity inhibits VSMC migration and proliferation by blocking the cell cycle (184–186). PPAR activation regulates lipid metabolism and inhibits foam cell formation by augmenting the expression of scavenger receptors (187).

CONCLUSION

Nearly a half century ago, atherosclerotic plaque regression was considered to be an unachievable feat. Now it is recognized that established plaques can rapidly stabilize and regress in animals and humans. Experimental data have shown that plaque regression is not simply a rewinding of the sequences of events that lead to lesion progression, but instead involves specific cellular and molecular pathways that are eventually able to mobilize all pathological components of the plaque. HDL, ECs and VSMCs all play important roles in the progression and regression of atherosclerosis. Targeting of these cells and molecules in the future will ensure pharmaceutical agents are able to elicit even better plaque regression.