Cell and Tissue Research

, 335:191

Vascular endothelium in atherosclerosis

Authors

    • Institute of Cellular Biology and Pathology “Nicolae Simionescu”
  • Camelia S. Stancu
    • Institute of Cellular Biology and Pathology “Nicolae Simionescu”
  • Maya Simionescu
    • Institute of Cellular Biology and Pathology “Nicolae Simionescu”
Review

DOI: 10.1007/s00441-008-0678-5

Cite this article as:
Sima, A.V., Stancu, C.S. & Simionescu, M. Cell Tissue Res (2009) 335: 191. doi:10.1007/s00441-008-0678-5

Abstract

Their strategic location between blood and tissue and their constitutive properties allow endothelial cells (EC) to monitor the transport of plasma molecules, by employing bidirectional receptor-mediated and receptor-independent transcytosis and endocytosis, and to regulate vascular tone, cellular cholesterol and lipid homeostasis. These cells are also involved in signal transduction, immunity, inflammation and haemostasis. Cardiovascular risk factors, such as hyperlipaemia/dyslipidaemia trigger the molecular machinery of EC to respond to insults by modulation of their constitutive functions followed by dysfunction and ultimately by injury and apoptosis. The gradual activation of EC consists initially in the modulation of two constitutive functions: (1) permeability, i.e. increased transcytosis of lipoproteins, and (2) biosynthetic activity, i.e. enhanced synthesis of the basement membrane and extracellular matrix. The increased transcytosis and the reduced efflux of β-lipoproteins (βLp) lead to their retention within the endothelial hyperplasic basal lamina as modified lipoproteins (MLp) and to their subsequent alteration (oxidation, glycation, enzymatic modifications). MLp generate chemoattractant and inflammatory molecules, triggering EC dysfunction (appearance of new adhesion molecules, secretion of chemokines, cytokines), characterised by monocyte recruitment, adhesion, diapedesis and residence within the subendothelium. In time, EC in the athero-prone areas alter their net negative surface charge, losing their non-thrombogenic ability, become loaded with lipid droplets and turn into foam cells. Prolonged and/or repeated exposure to cardiovascular risk factors can ultimately exhaust the protective effect of the endogenous anti-inflammatory system within EC. As a consequence, EC may progress to senescence, lose their integrity and detach into the circulation.

Keywords

Endothelial dysfunctionLipoprotein retentionOxidative stressInflammationAtherosclerosis

Introduction

The vascular endothelium by virtue of its strategic location between the plasma and the underlying tissue and its constitutive properties is endowed with a large array of functions that are vital for body homeostasis. Under physiological conditions, endothelial cells (EC) monitor the transport of plasma molecules, employing bidirectional receptor-mediated and receptor-independent transcytosis and endocytosis, to regulate vascular tone and to synthesise and secrete a large variety of factors. In addition, EC are involved in the regulation of cholesterol and lipid homeostasis, signal transduction, immunity, inflammation and haemostasis (Simionescu and Antohe 2006; Mehta and Malik 2006). Under pathological conditions such as hyperlipidaemia and/or hyperglycemia, alterations in endothelial function precede the development of atherosclerotic plaques and contribute decisively to their perpetuation and to the clinical manifestations of vascular diseases.

Atherosclerosis is a complex multifactorial disease developing in the arterial wall in response to various forms of injurious stimuli and resulting in excessive inflammatory and fibro-proliferative reactions. Among the main risk factors identified by classical epidemiology are hyperlipidaemia, hypertension, diabetes, smoking and aging, out of which hyperlipidaemia and/or hyperglycemia are prerequisites for the initiation and progression of arterial lesions. EC are involved in all stages of atherogenesis and their dysfunction is a key early event in plaque formation. The development of atheroma includes a sequence of events leading to fatty streak and fibro-fatty plaque formation, in humans and in animal models (Hansson et al. 2006; Schwartz et al. 2007).

Vascular endothelium in lesion-prone areas is involved in the initial events of atherogenesis

Heterogeneity of EC defines lesion-prone areas of increased shear stress

The structural-functional characteristics of EC differ with respect to the frequency of caveolae, channels, fenestrae, expression of surface molecules, secretory capacity, junctional organisation, reaction to changes in shear stress, proliferative capacity and response to vasoactive molecules (for a review, see Simionescu and Antohe 2006). Moreover, EC are permanently exposed to various types of biomechanical forces from their surroundings; these are important factors for initiating atherogenesis because of the differential regulation of endothelial transcriptional programmes. Reported data reveal that the heterogeneity of the vascular endothelium is a consequence of the ability of EC to integrate and transduce individually humoral and biomechanical stimuli from the environment (García-Cardeña and Gimbrone 2006). Athero-prone versus athero-protective shear stress waveforms induce differential patterns of gene expression in EC (Dai et al. 2004). In athero-prone areas, the EC acquire a pro-inflammatory phenotype, expressing several important chemokine receptors, and elicit a dysregulation of the expression and organisation of cytoskeletal and junctional proteins. Reactive oxygen species (ROS) and inflammatory molecules activate the transcriptional pathways of nuclear factor-kB (NF-kB) or activator protein 1 (AP-1) and trigger the cytokine-inducible cell-surface expression of adhesion molecule (CAM). A particular role is played by the endothelium lining the vasa vasorum but this is still not clearly understood, although the relationship between changes in the vasa vasorum and the development of atheromatous plaques is well documented. However, even if the proliferation of these vessels is merely reactive, the newly formed microvessels may be a source of disease progression by virtue of endothelial impairment and as an alternative pathway for monocyte migration to the site of early disease (Ritman and Lerman 2007).

Changes in EC constitutive properties are the initial event in atherogenesis

The induction of vascular lesions in humans and in most animal models is dependent on hypercholesterolaemia. Plasma cholesterol elevation can either be induced by dietary supplementation, by hepatic overproduction of lipoproteins (Lp) or by genetic mutations of receptors and/or receptor ligands responsible for Lp clearance.

In hyperlipidaemia and/or hyperglycemia, the initial response of EC is the synchronised modulation of two constitutive functions: permeability and biosynthesis. A characteristic sequence of events has been observed in all lesion-prone areas of atherosclerosis, such as the cardiac valves, aortic arch and coronary and carotid arteries. These are the focal sites where atherosclerosis develops in animal models of saturated fat-induced atherosclerosis, such as the hypercholesterolaemic rabbit (Simionescu et al, 1986) and the hyperlipidaemic (HL) hamster (Nistor et al. 1987). The same distribution of athero-prone sites is encountered in humans (Chatzizisis et al. 2007).

In all lesion-prone areas, the EC are the first cells to experience the impact of hyperlipidaemia. Data from the HL hamster show that they modulate towards a secretory phenotype, with a characteristic increase in the number of biosynthetic organelles (endoplasmic reticulum and Golgi apparatus) and the appearance of an abundance of intracellular microfilaments and microtubules (Fig. 1). This secretory shift is accompanied by significant hyperplasia of the basement membrane, which appears as multiple interconnected layers (Filip et al. 1987; Sima et al. 1990; for a review, see Simionescu 2007). Plasma hypercholesterolaemia is associated with increased transcytosis of βLp, leading to their accumulation within and outside the EC hyperplasic basement membrane, against the fragmented internal elastic lamina (Figs. 2, 3a). At this location, Lp interact with proteoglycans (PG) and other matrix proteins and carry on their conversion to oxidatively modified and reassembled Lp (MLp; Simionescu et al. 1986; Nistor et al. 1987). MLp have been identified in early intimal thickenings of human aorta and in the late atheroma (Tirziu et al. 1995). Accumulation of Lp in the subendothelium depends on EC properties and Lp characteristics, such as their oxidative modification (Steinberg 2005), and leads to MLp retention (Tabas et al. 2007). MLp, together with the cytokines produced by the activated endothelium, induce p21-activated kinase (PAK); in cultured bovine aortic EC, the inhibition of PAK has been demonstrated to reduce permeability in athero-prone areas (Orr et al. 2007). Recently, the Ser/Thr kinase Akt and protein kinase G have been found to phosphorylate PAK at Ser21 within the Nck-binding sequence, thus inhibiting the interaction between PAK and Nck (Fryer et al. 2006).
https://static-content.springer.com/image/art%3A10.1007%2Fs00441-008-0678-5/MediaObjects/441_2008_678_Fig1_HTML.gif
Fig. 1

Structural modulation of endothelial cells (EC) towards a secretory phenotype in the initial stage of atherogenesis (l vascular lumen). Early (2 weeks) ultrastructural changes of the aortic valve EC of a hyperlipemic hamster. Note the presence of numerous copies of biosynthetic organelles such as rough endoplasmic reticulum (RER), Golgi apparatus (G) and centrioles (C). Bar = 200nm

https://static-content.springer.com/image/art%3A10.1007%2Fs00441-008-0678-5/MediaObjects/441_2008_678_Fig2_HTML.gif
Fig. 2

Intima of the aortic arch of a hyperlipidaemic hamster after 2 weeks of diet (SMC smooth muscle cells, l vascular lumen). Numerous modified lipoproteins (MLp) that appear as variously sized vesicles are present in the meshwork of the hyperplasic endothelial basement membrane (bm) and lie against and in-between the fragmented internal elastic lamina (iel). Bar = 800nm. Inset: Enlarged view of modified lipoproteins (liposome-like structures) originating from transcytosed βLp; incubation with filipin reveals the presence of free cholesterol in the rim of the vesicles. Bar = 500nm

https://static-content.springer.com/image/art%3A10.1007%2Fs00441-008-0678-5/MediaObjects/441_2008_678_Fig3_HTML.gif
Fig. 3

a Electron micrograph showing a segment of the intima of an aortic arch from a hyperlipidaemic hamster perfused in situ with low-density lipoprotein (LDL) bound to 5-nm gold particles (LDL-Au). Note the accumulation of LDL-Au in the subendothelial space (EC endothelial cell). Bar = 200nm. b Representation of the relative mass distribution of the four major lipoprotein (Lp) constituents (cholesteryl esters, unesterified cholesterol, phospholipids and triglycerides), as determined in plasma low-density lipoprotein (LDL) and β very-low-density lipoprotein (βVLDL), compared with the modified lipoproteins (MLp) obtained from extracts of subendothelial lipid deposits from rabbit aortas after 2 weeks of hypercholesterolaemic diet

EC may modulate the accumulation of apolipoprotein-B (apoB)-containing Lp in the arterial wall through other mechanisms, such as influencing the composition of the subendothelial matrix that is responsible for the Lp retention and/or the nature of the transcytosed Lp. In order to demonstrate the active role played by EC in this process, Brown et al. (2004) have explored the variation in the retention of apoB-containing Lp in the arterial wall of athero-susceptible B6 mice, compared with athero-resistant C3H mice, and have demonstrated that the former have significantly more apoB in their aortic wall than the latter. They have also found that, although the uptake of native low-density lipoproteins (LDL) is comparable between the two strains, the aortic EC of C3H mice have an increased ability to transform native LDL to oxidised LDL (oxLDL) and to take up oxLDL, in comparison with B6 mice, and that this difference leads to reduced deposition of LDL in C3H mice (Brown et al. 2004).

Persistent dyslipidaemia leads to EC dysfunction

Subendothelial accumulation and retention of MLp

Transcytosis, deposition and accumulation of MLp within the subendothelium start the atherogenic process in hyperlipidaemic and/or hyperglycemic hamsters (for a review, see Simionescu 2007) and in humans (Mehrabi et al. 2000; Williams and Tabas 2005). Studies by Schwenke and StClair (1993) with circulating labeled Lp have implicated subendothelial retention, rather than increased intimal influx, as a major determinant of Lp accumulation in athero-prone regions of white Carneau pigeons with naturally occurring and cholesterol-aggravated aortic atherosclerosis.

We can safely assume that the accumulation of MLp within the subendothelium is the end result of a complex process that includes an augmented influx of Lp, a reduced efflux from the vessels intima and an enhanced interaction with matrix proteins.

Within the subendothelium, Lp transported through the EC, either LDL (major cholesterol carrier Lp in humans) or β-very low density lipoproteins (βVLDL) as is the case for hypercholesterolaemic rabbits and hamsters, exhibit changes in their structure and composition: decreased lipid and increased apoB content (Mora et al. 1989). The increased ratio of unesterified/total cholesterol of the aortic MLp (Fig. 3b) could be attributed to the hydrolysis of the cholesteryl esters in the intima, a process that may contribute to the appearance of unesterified cholesterol (UC)-rich phospholipid vesicles. Electron-microscopic examination of accumulated MLp shows their various dimensions and electron opacity, whereas incubation with filipin reveals the presence of increased UC in the rim of the vesicles (Fig. 2, inset).

In vitro and in vivo studies have provided strong evidence that certain non-matrix molecules play important roles in Lp retention. The most extensively studied are lipoprotein lipase, sphingomyelinase (SMase) and lipoprotein-associated phospholipaseA(2), which is also called sPLA 2. SMase, an enzyme secreted by EC into the extracellular milieu of the arterial wall causes the aggregation of Lp particles and increases their affinity for PG and their capacity to load macrophages, smooth muscle cells (SMC) and EC with lipids (Öörni et al. 2005). sPLA2 that circulates bound mainly to LDL in the blood can convert LDL into small dense particles that have an enhanced tendency to interact with PG and are associated with increased risk of cardiovascular disease (Hurt-Camejo et al. 2000). Moreover, sPLA2 is also present in the human artery wall, where it may act locally to promote the development of atheroma (Lavi et al. 2007). In vitro proteolysis of apoB100 by exposure to sPLA2 increases the binding of the modified LDL to PG (Flood et al. 2004).

In the arterial intima, the accumulation of PG, the alteration of pericellular glycoproteins and the modulation of collagen turnover influence Lp retention, cell behaviour and calcinosis. An important role is played by the matrix metalloproteinases (MMPs) rather than by their inhibiting factors (tissue inhibitors of metalloproteinases), which are expressed in EC constitutively and are not increased by oxLDL. Conversely, oxLDL increases MMP-1 and -3 expression in humans (Li et al. 2003) via the lectin-like oxidised LDL receptor-1 (LOX-1) and activation of protein kinase C. MMP-2 expression is constitutive in EC and further up-regulated by ROS. Therefore, inflammatory and immune mechanisms seem to be the main processes in matrix remodelling by EC.

Changes in EC glycocalyx

The role of the endothelial glycocalyx in atherogenesis is, as yet, not well-established but some interesting observations point to its involvement. In healthy vessels, the EC glycocalyx regulates vascular permeability, blood cell–vessel wall interactions, shear stress sensing and balanced signalling and fulfils a vasculo-protective role. Loss of glycocalyx results in the shedding of endogenous protective enzymes, such as extracellular superoxide dismutase (SOD), and increases oxidative stress in EC (Van den Berg et al. 2006). An inverse relationship between the glycocalyx thickness and the intima–media ratio has been reported, reflecting a reduction of its vasculoprotective capacity at sites with high atherogenic risk. Together, these data suggest that the endothelial glycocalyx is involved in the initiation and progression of the atherosclerotic process (Nieuwdorp et al. 2005).

The EC plasmalemma exhibits, under physiological conditions, a net negative charge because of the highly sulphated glycosaminoglycan chains, which contribute to the characteristic non-thrombogenic surface of the endothelium (Simionescu et al. 1991). An alteration of the net negative surface charge of EC has been noted in HL hamster arteries, concomitantly with a similar alteration found in circulating cells (monocytes and platelets; Sima et al. 1990). The loss of the anionic sites may account for the increased permeability and augmented adhesiveness of the vessel wall in athero-prone areas.

Modulation of intercellular junctions

Intercellular tight junctions (zonula occludens) are a constant and important structure of EC, their number and complexity varying according to EC type, e.g. brain and large arteries contain many tight junctions, whereas post-capillary EC contain few or none. In the late stages of atherosclerosis, EC tight junctions may open and allow the passage of plasma molecules (Dejana et al. 1997).

Gap junction (GJ) channels result from the docking of two hemichannels or connexons, formed by the hexameric assembly of connexins that directly connect the adjacent cells. The expression pattern of vascular connexins is altered during atherosclerotic plaque formation in mouse (Chadjichristos and Kwak 2007). GJ, apart from their role in cell-to-cell signalling, regulate communication in nitric oxide (NO)-induced vasodilation of the hamster mesenteric resistance arteries, through both cGMP-dependent and -independent pathways. Uncoupling of the GJ by heptanol causes a reduced effect on artery wall contraction, which is most prominent when hyperlipidaemia is associated with diabetes, thus suppressing both endothelium-dependent and independent relaxation (Georgescu et al. 2006).

Role of oxidative stress and control of leucocyte traffic

Persistent dyslipidaemia generates ROS, which activate EC, a pivotal early event in atherogenesis. Existing evidence implicates ROS as a common denominator in the development of most cardiovascular diseases. Superoxide (O2·−) and hydrogen peroxide (H2O2) are two of the most biologically important ROS in the cardiovascular system; they are produced by the vascular cells via several oxidases, including the NADPH oxidases (NADPHox), xanthine oxidase, lipoxygenases and cytochrome P450, or via the uncoupling of the mitochondrial respiratory chain and uncoupling of endothelial nitric oxide synthase (eNOS). Production of ROS is counterbalanced by antioxidant enzymes such as SOD, catalase, glutathione peroxidase, the thioredoxins and the peroxiredoxins. ROS production in vascular injury is positively regulated by many of the cytokines whose expression is increased after EC activation. Excess production or impaired ROS removal in a pro-inflammatory environment regulates virtually all the EC responses, including monocyte adhesion, platelet aggregation, inflammatory gene induction and impaired endothelium-dependent relaxation (Papaharalambus and Griendling 2007).

EC activation is characterised by the synthesis and secretion of cytokines such as interleukins (IL), tumor necrosis factor-α (TNFα), angiotensin II, vascular endothelial growth factor and the expression of CAMs: intercellular adhesion molecule-1 (ICAM-1), vascular cell adhesion molecule-1 (VCAM-1), E and P selectin and chemokines, such as the monocyte chemoattractant protein-1 (MCP-1) and fractalkine.

Another important stimulus for CAM expression is the fluid shear stress, which exerts both pro-inflammatory and protective effects, depending on the type of shear. Disturbed flow conditions, which occur at branch points or curvatures, enhance ICAM-1 and VCAM-1 expression, whereas laminar shear stress prevents apoptosis and monocyte adhesion by producing NO· and inducing expression of antioxidant genes. Oscillatory shear stress leads to continuous O2·− production in an NADPH-oxidase-dependent manner, resulting in NF-κB-mediated monocyte adhesion (Papaharalambus and Griendling 2007).

The endothelial expression of CAM, secretion of growth factors, cytokines and chemokines, trigger T-lymphocytes and monocyte recruitment, diapedesis and homing within the subendothelium (for a review, see Simionescu 2007). Platelets represent another source of chemokines and their precursors involved in the process of atherogenic recruitment of leucocytes. From their alpha-granules, platelets secrete chemokines, such as PF4 or ENA-78, and precursors for the CXCR2 ligand NAP-2; in addition, they deposit chemokines, such as RANTES or PF4, on the EC lining of early atherosclerotic lesions, where RANTES oligomers trigger the CCR-1–dependent arrest of monocytes via a mechanism involving platelet P-selectin (Weber et al. 2004). However, little evidence exists for priming and activation of peripheral blood polymorphonuclear leucocytes or monocytes consistent with a former platelet-bound population. Recent data suggest that P-selectin glycoprotein ligand-1 P-selectin–dependent platelet binding to monocytes represents a normal physiological process with little impact on the potential of monocytes to cause vascular injury in humans (Bournazos et al. 2008). Postprandial Lp can also activate leucocytes in the blood and up-regulate the expression of CAM on the endothelium, facilitating adhesion and migration of inflammatory cells into the subendothelial space (Alipour et al. 2007).

Several other transcription factors, such as specificity protein, AP-1, retinoic acid responsive element and CCAAT/enhancer-binding protein participate in the activation of the ICAM-1 gene, particularly in EC.

Pentraxins (PTX) are a superfamily of evolutionarily conserved proteins characterised by a structural motif, the PTX domain. The C-reactive protein, which together with the serum amyloid P component constitutes the short PTX arm of the superfamily, was the first to be identified. PTX3 exists as a free or extracellular-matrix-associated molecule and binds the complement fraction C1q. EC and macrophages are potent producers of PTX3 in response to inflammatory stimuli. PTX3 immunoreactivity has been observed in EC and macrophages within lesions and in subendothelial SMC and in foam cells within lipid-rich areas of plaques. Enzymatically degraded and oxLDL, but not native LDL, induce PTX3 up-regulation in cholesterol-loaded human primary SMC. Of relevance to this point, atherogenic Lp may induce the local production of proinflammatory cytokines, including IL-1 and IL-6, which, in turn, up-regulate PTX3 expression by infiltrating monocytes, endothelium and SMC (Presta et al. 2007).

Lipid infusion increases the production of ROS and inflammation in humans. In EC, free fatty acids (FFA) stimulate NADPHox induced ROS through a protein-kinase-C-dependent mechanism.

Transcriptional factors and endothelial dysfunction

NF-kB and AP-1 are the most widely studied transcriptional factors to be influenced by the cellular redox state.

NF-kB is an inducible transcription factor present at increased levels in the thickened intima-media of atherosclerotic lesions, whereas little or no activated NF-kB has been detected in healthy vessels. Several of the cytokines and growth factors found in atherosclerotic lesions, such as TNFα, IL-1β, MCP-1 and tissue factor, activate NF-kB in cultured EC. The evidence suggests that oxidative stress induces, whereas antioxidants prevent, the cytoplasmic–nuclear translocation of NF-kB. Although the primary mode of activation of NF-kB by ROS appears to be the release from the inhibitor subunit IkB and translocation to the nucleus, ROS may modulate the activity of NF-kB by regulating post-translational modifications of the NF-kB subunits or of other transcriptional cofactors that influence the transcriptional activity of NF-kB (Pennathur and Heinecke 2007).

AP-1 activity is regulated by both transcriptional and post-translational mechanisms in response to a variety of extracellular stimuli, including mitogens, phorbol esters and differentiation signals. In EC, agents such as H2O2, LDL and oxLDL activate AP-1 DNA-binding activity and regulate the vascular inflammatory genes MCP-1 and ICAM-1 by AP-1-binding elements in the gene promoters. Recently, data regarding the mechanism underlying ROS-mediated AP-1 activation have been reported for cultured human vascular SMC (Manea et al. 2008).

Both NF-kB and AP-1 play a role in the FFA regulation of CAM in EC and SMC. Conjugated linoleic acids and omega-3 long-chain polyunsaturated fatty acids reduce to different extents the TNFα-induced expression of CAMs (ICAM-1, VCAM-1, but not E-selectin) in human umbilical vein EC and SMC as a function of the FFA type and concentration. Reduction of the expression of ICAM-1 and VCAM-1 proteins by n-3 polyunsaturated fatty acids is less dependent on the NF-kB pathway than the reduction by conjugated linoleic acids; this suggests the involvement of other transcription factors (i.e. AP-1) in the FFA regulation of CAM expression in human endothelial and SMCs (Goua et al. 2008).

Nitric oxide, prostacyclin, endothelin and angiotensin II

NO modulates vasomotor tone and has several important anti-atherogenic effects on EC and platelets through downstream signalling pathways (Jones and Bolli 2006). In atherosclerosis, NO bioactivity is diminished, principally because of decreased eNOS expression and/or activity, eNOS uncoupling, enhanced breakdown or scavenging of NO, impaired NO-mediated signalling events and disequilibrium of the NO/redox balance (Braam and Verhaar 2007). In the case of “eNOS uncoupling”, electrons flowing from the reductase domain to the heme are diverted to molecular oxygen, rather than to the substrate L-arginine, thereby resulting in the production of superoxide instead of NO (Cai 2005). Other biochemical mechanisms are proposed to be involved in eNOS uncoupling, such as tetrahydrobiopterin deficiency, increases in endogenous asymmetric dimethylarginine, L-arginine deficiency and oxidative stress (Schmidt and Alp 2007). Interestingly, eNOS protein level is increased rather than decreased in disease states such as atherosclerosis and diabetes, although NO bioavailability is reduced.

Disequilibrium of the NO/redox balance may result in the development of atherosclerosis, myocardial tissue remodelling and hypertrophy. NO reacts with peroxyl radicals, a crucial step in the direct nitrosylation of proteins (Elahi et al. 2007). Peroxynitrite (ONOO-), a potent oxidant formed by the combination of superoxide anion and NO, has been reported to inhibit prostacyclin synthase (PGIS) selectively; sub-micromolar levels of ONOO- inhibit prostaglandin I2 (PGI2)-dependent vasorelaxation and trigger increased PGH2 formation and the ensuing PGH2-dependent vasospasm. Nitration of PGIS might be a new pathogenic mechanism for superoxide-induced endothelium dysfunction in vascular diseases (Zou 2007).

Endothelium modulation of vascular tone is a result of the synthesis and release of vasodilators, such as NO, prostacyclin and EDHF, of vasoconstrictors, such as endothelin-1 (ET-1) and prostanoids, and of the conversion of angiotensin I to angiotensin II at the endothelial surface. The vasoconstrictors predominantly act locally but may also exert some systemic effects and have a role in the regulation of the arterial structure and remodelling (Deanfield et al. 2007). ET-1 exerts its biological activity by binding to two specific ET-1 receptor subtypes: endothelin-A and endothelin-B. Endothelial secretion of ET-1 is polarised, taking place towards the underlying vascular SMC, thus exerting primarily autocrine and paracrine effects and rending ET plasma levels less physiologically relevant. Other endothelium-derived vasoconstriction agents such as superoxide anions, prostaglandins, and thromboxane A2 are documented as accompanying endothelial dysfunction.

Endothelial dysfunction correlates with increased plasma levels of remnant-like lipoprotein particles (RLp), which are potentially highly atherogenic. A recent hypothesis has been proposed whereby RLp impair endothelial function via direct and indirect effects on eNOS. The RLp may affect the autophosphorylation of focal adhesion kinase and its downstream phosphatidylinositol kinase/Akt signalling pathway, resulting in direct eNOS inactivation through the induction of intracellular ROS. In addition, RLp might indirectly affect the expression or activation of eNOS by stimulating secretion of various inflammatory factors (for a review, see Zheng and Liu 2007).

EC-derived foam cells, injury and apoptosis: late events in atherosclerosis

Scavenger receptors

Since the cloning of the first macrophage scavenger receptors (SR) in 1990, the SR family has expanded to include eight different subclasses of structurally unrelated receptors that share the defining feature of being able to bind modified LDL. These receptors have subsequently been found to bind and “scavenge” a broad array of modified self and nonself ligands, including apoptotic cells, anionic phospholipids, amyloid and pathogen components. The SR are thus believed to be members of the group of pattern recognition receptors that mediate the innate immune host response through recognition of highly conserved pathogen-associated molecular patterns (Moore and Freeman 2006).

Class B of SR consists of CD36, SR-BI and lysosomal integral membrane protein–II. CD36, which binds oxLDL, is a type III (multiple transmembrane domain) receptor. Despite the high degree of homology of CD36 and SR-BI, these two receptors appear to play distinct roles in lipid metabolism and atherosclerosis. Endothelial CD36 is a multiple ligand receptor that binds to oxLDL, trombospondin, Plasmodium falciparum-infected erythrocytes, long-chain fatty acids and Gram-negative and Gram-positive bacteria (Adachi and Tsujimoto 2006).

LOX-1 has been identified as the main receptor that mediates the action of oxLDL in the vascular wall. The LOX-1 expression detected in EC in early atherosclerotic lesions of human carotid arteries is increased by ROS, ET-1, TNFα, shear stress, protein-kinase-C activation, angiotensin-II and inflammatory molecules (Chen et al. 2007).

EC-derived foam cells

The formation of foam cells as a result of the lipid loading of EC is a late event in atherosclerosis, as documented in hypercholesterolaemic rabbits and hamsters. Since the process is gradual, we suggest that, in the initial stage of atherogenesis, upon the accumulation and retention of MLp within the intima, the EC lining the plaque take up MLp, which are either degraded within the cell or exocytosed into the lumen; in time, the non-regulated uptake of MLp by the EC-SR is overwhelmed, leading to the accumulation of numerous large lipid droplets within the EC (Fig. 4a, b). However, the EC-derived foam cells appear to maintain some endothelial attributes, such as Weibel-Palade bodies, intercellular junctions and caveolae (Sima et al. 1990; Constantinescu et al. 2000).
https://static-content.springer.com/image/art%3A10.1007%2Fs00441-008-0678-5/MediaObjects/441_2008_678_Fig4_HTML.gif
Fig. 4

Segment of the intima of an aortic valve from a hyperlipidaemic hamster (bm basal lamina, l lumen). a Circulating monocyte (M) infiltrating between two endothelial cells (EC) that contain lipid droplets (ld). Bar = 500nm. b After 6 weeks of hyperlipidaemic diet, the hamster aortic valve displays marked alterations including the presence of EC loaded with numerons, large lipid droplets (Id). Bar = 500nm

In late stages of atherosclerosis, all cellular components of the plaque, EC, SMC and macrophages, accumulate considerable number of lipid droplets and exhibit the foam cell characteristics (Fig. 5).
https://static-content.springer.com/image/art%3A10.1007%2Fs00441-008-0678-5/MediaObjects/441_2008_678_Fig5_HTML.gif
Fig. 5

Electron micrograph of an area of the aortic intima from a hyperlipidaemic hamster at 35 weeks of diet (l lumen). Note the presence, within the vessel wall, of numerous lipid-droplet (ld)-filled foam cells derived from endothelial cells (EC), smooth muscle cells (SMC) and macrophages (MAC). The SMC present within the intima that contain numerous lipid droplets are surrounded by dark stained fragments of elastin (arrows). Bar = 1,8 micrometers

Apoptosis

EC apoptosis is assumed to be triggered by local inflammatory mediators or the cytolytic attack of activated killer T cells, cytokines, ROS and oxLDL, which increase EC synthesis of MMPs (for a review, see Simionescu 2007). Apoptosis and injury of EC can also generate endothelial microparticles, which are vesicles of plasmalemmal origin, characterised by the exposure of phosphatidylserine on the outer leaflet as a consequence of the calcium-dependent activation of scramblase and floppase/ABC1 and the inhibition of translocase/flippase activities (Boulanger et al. 2006; Werner et al. 2006). Elevated levels of circulating microparticles have been seen in a variety of conditions associated with endothelial activation, dysfunction or apoptosis. Shear stress is a major determinant of endothelial apoptosis but its role in the release of membrane microparticles by EC is as yet unknown.

The increased level of circulating EC in patients with atherosclerotic disease and vascular inflammation suggests a direct relationship between the number of these cells in the peripheral circulation and the extent of endothelial injury (Goon et al. 2005).

Reversal of endothelial dysfunction

A reliable assessment of the functional state of the endothelium in various pathological conditions can be achieved by determining the level of molecules of endothelial origin circulating in the blood. These include direct products of activated EC, such as NO, inflammatory cytokines, CAM, regulators of thrombosis and markers of endothelial damage and repair. The change in the balance of tissue plasminogen activator and its endogenous inhibitor, plasminogen activation inhibitor-1, and the increased secretion of von Willebrand factor are also signs of endothelial activation and dysfunction.

Attempts to restore EC functions in hyperlipidaemia include plasma cholesterol reduction and the inhibition of receptors for modified Lp clearance. Lowering the plasma lipids is beneficial for stopping or causing the regression of atherosclerosis, in part by reducing vascular inflammation and thus diminishing the cardiometabolic risk in patients (Després et al. 2008).

In hypercholesterolaemic rabbits, atherosclerotic aortas produce high levels of ROS, and oxLDL accumulate in the atheroma underlying the EC, which overexpress VCAM-1; few, if any, EC lining the atheroma express eNOS. Lipid lowering reduces ROS production, oxLDL accumulation and plasma levels of anti-oxLDL IgG. In addition, VCAM-1 and MCP-1 expression decreases, eNOS expression increases, and EC exhibit a relatively normal ultrastructure (Aikawa et al. 2002). The data attest that lipid lowering reduces oxidative stress and EC activation in vivo and contributes to the recovery of EC functions and plaque stabilisation.

Another efficient means of cholesterol reduction is statin treatment, which besides inhibiting endogenous cholesterol synthesis, has numerous pleiotropic effects. Restoration of endothelial function is the result of the reduction of circulating LDL levels and the ensuing transcytosed LDL, restoration of the serum antioxidant potential, transcriptional activation of the eNOS gene and inhibition of O2- formation by EC and, subsequently, of the NO-/O2- balance (for reviews, see Sima and Stancu 2001; Steinberg 2006).

The SR, through their ability to clear potentially deleterious modified Lp that accumulate in the artery wall, are beneficial during the initial stages of atherogenesis. With time, as macrophages pathways for metabolising modified Lp-derived cholesterol become overwhelmed, the unregulated nature of these receptors result in the formation of foam cells and chronic inflammation. Inhibition of SR involved in modified Lp uptake and the formation of foam cells can also reverse the atherosclerotic process. Anti-LOX-1 therapy may effectively reverse critical pathogenic elements of nephropathy in diabetes and dyslipidaemia in hybrid Zucker-spontaneous hypertensive rats (Dominguez et al. 2008). However, the effectiveness of therapies targeted at inhibiting SR pathways and the Lp clearance from the intima remain questionable, because the fate of these pro-inflammatory Lps in the artery wall is unknown.

A new approach to regenerating damaged endothelium relies on endothelial progenitor cells (EPC), which have the potential to participate in new vessel formation and endothelial regeneration (Asahara et al. 1997). EPC are bone-marrow-derived circulating cells with the ability to differentiate into mature EC. Although a precise definition is lacking, EPC are generally believed to share cell surface markers of haematopoietic and endothelial lineages, are released from the bone marrow and are mobilised to the periphery in response to a variety of stimuli. Under basal conditions, EPC incorporate at low levels in the vascular endothelium layer. Tissue ischaemia, through the release of growth factors and cytokines, is a potent stimulus for the mobilisation of EPC from the bone marrow and recruitment to sites of injury. Direct evidence that EPC contribute to endothelial regeneration following injury derives mainly from animal studies. Murine bone-marrow transplantation models provide convincing evidence that EPC contribute to endothelial regeneration following a range of injuries, such as mechanical denudation, bypass grafting and hyperlipidaemia. Recent studies demonstrating that infusion of EPC reduce the development of restenosis, endothelial dysfunction and atherosclerosis provide further evidence for the important biological role of these cells (Cubbon et al. 2007; Xu 2007).

In humans, EC dysfunction can be assessed biochemically by determination of various EC markers (e.g. adhesion molecules, cytokines and prostanoids) in the blood. Fatty streaks are the first reversible stage of atherosclerosis, as evidenced by the early markers of EC dysfunction: ROS, MCP-1, E and P selectin and PTX. The unstable plaque can be characterised by increased circulating ICAM, VCAM, MMP, TNFα, IL-6, IL-10 and IL-18, whereas the ruptured plaque exibits additionally CD40-L and complement components.

Thick sections of coronary arteries from hyperlipidaemic hamsters at different stages of atherosclerosis are shown in Fig. 6 to illustrate the association with specific biomarkers.
https://static-content.springer.com/image/art%3A10.1007%2Fs00441-008-0678-5/MediaObjects/441_2008_678_Fig6_HTML.gif
Fig. 6

Sections of coronary arteries from hyperlipidaemic hamsters at various stages of atherosclerosis. The biomarkers specific for each stage are shown under each light-microscopic image. The earliest detectable plasma biomarkers (increased βLps, decreased HDL, increased glucose) lead to EC activation and fatty streak formation (ROS reactive oxygen species, MLp modified Lp, MCP-1 monocyte chemoattractant protein-1, E and P selectins, PTX pentraxins), the unstable plaque markers (ICAM intercellular adhesion molecule, VCAM vascular cell adhesion molecule, MMP matrix metalloproteinase, TNFα tumour necrosis factor-alpha, IL-6 interleukin-6, IL-10 interleukin-10, IL-18 interleukin-18) and the rupture plaque markers (CD40-L and complement components)

Concluding remarks and future directions

Hypercholesterolaemia, diabetes, hypertension, hyperhomocysteinaemia, smoking, aging and increased body mass index are major risk factors for the development of atherosclerosis. They are all associated with an increase in oxidative stress and inflammation and all have endothelial dysfunction as a common denominator. The consequences are a reduced bioavailability of NO, an alteration in the production of prostanoids and an increased release of endothelin-1, all being important features of endothelial dysfunction and all playing key roles in atherosclerosis and other cardiovascular diseases. However, the mechanisms underlying endothelial dysfunction are different, depending on the specific pathology, the vascular bed involved and the additional risk factors. Circulating microparticles may be one of the key factors linking inflammation, oxidative stress, apoptosis and atherothrombosis.

Further work is required to better elucidate the mechanisms that regulate the enzymatic generation of ROS, the importance of the subcellular compartmentalisation of ROS and the molecular effects of ROS on individual proteins. New transgenic and knockout animal models are needed to study in detail the effects of individual generators of ROS and to identify their effectors. Novel biomarkers of oxidative stress are neccessary to identify, at early stages, the need for antioxidant therapy and to assess the effectiveness of new antioxidants.

Therapeutic interventions are well known as not necessarily restoring proper endothelial function and, when they do, they may improve only some of these variables. The effect of therapeutic agents on these parameters remains to be properly assessed, both in experimental animal models and in human disease.

Therapeutic strategies that operate by way of removal of inflammatory Lp components from the arterial wall and by promoting regression of the atherogenic responses to modified and retained Lp represent a major area of current drug development. At present, no anti-inflammatory drugs per se are available that have a beneficial effect on cardiovascular disease, specifically through their ability to decrease the inflammatory state associated with atherogenesis.

Modern directions of research in this area may lead to novel ways of suppressing atherogenesis and/or atherosclerotic plaque progression. Future strategies to revert EC dysfunction and the ensuing cardiovascular disease should be directed to designing and developing new drug combinations, accompanied by improved cooperation between physician and patient towards lifestyle changes and the discovery of early biomarkers necessary for identifying young subjects at risk.

Acknowledgments

This review is a tribute to the memory of our eminent mentor, Professor Nicolae Simionescu, who initiated most of the projects reviewed here. The work of many colleagues who over the years have contributed to the presented data is gratefully acknowledged.

Sources of Funding

The work was supported by grants from NIH-USA, the European Community and the COST Action BM0602, the Romanian Academy, and the Romanian Ministry for Education and Research.

Copyright information

© Springer-Verlag 2008