Changing concepts of atherogenesis

Authors


Dr Peter Libby, LMRC 307, 221 Longwood Avenue, Boston, Massachusetts, USA (fax: 617 7326961; e-mail: plibby@rics.bwh.harvard.edu).

Abstract

Abstract. Libby P (Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts, USA). Changing concepts of atherogenesis (Minisymposium: Review). J Intern Med 2000; 247: 349–358.

This review discusses three stages in the life history of an atheroma: initiation, progression and complication. Recruitment of mononuclear leucocytes to the intima characterizes initiation of the atherosclerotic lesion. Specific adhesion molecules expressed on the surface of vascular endothelial cells mediate leucocyte adhesion: the selectins and members of the immunoglobulin superfamily such as vascular cell adhesion molecule-1 (VCAM-1). Once adherent, the leucocytes enter the artery wall directed by chemoattractant chemokines such as macrophage chemoattractant protein-1 (MCP-1). Modified lipoproteins contain oxidized phospholipids which can elicit expression of adhesion molecule and cytokines implicated in early atherogenesis. Progression of atheroma involves accumulation of smooth muscle cells which elaborate extracellular matrix macromolecules. These processes appear to result from an eventual net positive balance of growth stimulatory versus growth inhibitory stimuli, including proteins (cytokines and growth factors) and small molecules (e.g. prostanoids and nitric oxide).

The clinically important complications of atheroma usually involve thrombosis. Arterial stenoses by themselves seldom cause acute unstable angina or acute myocardial infarction. Indeed, sizeable atheroma may remain silent for decades or produce only stable symptoms such as angina pectoris precipitated by increased demand. Recent research has furnished new insight into the molecular mechanisms that cause transition from the chronic to the acute phase of atherosclerosis. Thrombus formation usually occurs because of a physical disruption of atherosclerotic plaque. The majority of coronary thromboses result from a rupture of the plaque’s protective fibrous cap, which permits contact between blood and the highly thrombogenic material located in the lesion’s lipid core, e.g. tissue factor. Interstitial collagen accounts for most of the tensile strength of the plaque’s fibrous cap. The amount of collagen in the lesion’s fibrous cap depends upon its rate of biosynthesis stimulated by factors released from platelets (e.g. transforming growth factor beta or platelet-derived growth factor), but inhibited by gamma interferon, a product of activated T cells found in plaques. Degradation by specialized enzymes (matrix metalloproteinases) also influences the level of collagen in the plaque’s fibrous cap.

Such studies illustrate how the application of cellular and molecular approaches has fostered a deeper understanding of the pathogenesis of atherosclerosis. This increased knowledge of the basic mechanisms enables us to understand how current therapies for atherosclerosis may act. Moreover, the insights derived from recent scientific advances should aid the discovery of new therapeutic targets that would stimulate development of novel treatments. Such new treatments could further reduce the considerable burden of morbidity and mortality due to this modern scourge, and reduce reliance on costly technologies that address the symptoms rather than the cause of atherosclerosis.

Introduction

Since the dawn of the 20th century, the development of arteriosclerosis was viewed as an inevitable degenerative process. This outlook on arteries was made clear when Sir William Osler, the pioneer of anglophone internal medicine, stated that ‘the stability of tubing of any sort depends on the structure and on the sort of material used; and so it is with the human being. With the poor variety of elastic and muscular fibers in the blood vessels, some are unable to resist the wear and tear of daily life…in time they become old, in threescore or fourscore years the limit of their endurance is reached and they wear out.’[1]

Currently, our concepts of atherogenesis have evolved from vague ideas of inevitable degeneration to a much better defined scenario of molecular and cellular events. As we enhance our understanding of its fundamental mechanisms, we can begin to approach atherogenesis as a modifiable rather than ineluctable process. Eventually, mastery of the cell and molecular biology of atherosclerosis may permit development of novel strategies for mitigating this prevalent disease.

Atherogenesis in humans usually develops over many years, often reckoned in decades. Early lesion formation may even occur in adolescence. Lesion progression depends on genetic make-up, gender and certain well recognized risk factors as well as a number of non-traditional risk factors that are currently the subject of intense investigation. Complications that precipitate acute manifestations of atherosclerosis often happen suddenly. Some individuals with this disease may never experience symptoms, and some may endure chronic stable manifestations whilst evading complications. Still others may have grave or even fatal acute events with no prior warning, some in early adulthood.

This review will discuss three stages in the life history of an atheroma: initiation, progression and complication ( Fig. 1). Rather than aspiring to an encyclopaedic treatment, this brief paper will consider selected aspects of this subject where new findings or concepts illustrate mechanisms that may represent future therapeutic targets.

Figure 1.

Evolution of the atheroma. Atheroma formation begins early in life in many instances. Decades of clinically unapparent or stable disease precede the acute manifestations in the various beds. Recent progress is helping us to understand the temporal dispersion, as well as the spatial dispersion, of atheroma.

Lesion initiation

Recruitment of mononuclear leucocytes to the intima is one of the earliest events in the formation of an atherosclerotic lesion. We now appreciate that specific adhesion molecules expressed on the surface of vascular endothelial cells mediate leucocyte adhesion [23]. Structural features have defined several groups of adhesion molecules. Selectins comprise one such family of endothelial leucocyte adhesion molecules, whilst another family shares structural similarity with immunoglobulin molecules. Selectins mediate rolling or transitory contact of leucocytes with the endothelium. Endothelial cells overlying human atheroma express one member of the selectin family, P-selectin, in contrast to those in normal vessels [4]. Hyperlipidaemic mice deficient in P-selectin exhibit slowed atheroma formation consistent with a role for this specific molecule in this process [5–7].

The other major group of endothelial leucocyte adhesion molecules, the immunoglobulin superfamily, mediates more sustained sticking of leucocytes to the endothelium than do the selectins. One member of the immunoglobulin superfamily has special interest with regard to early atherosclerosis. Vascular cell adhesion molecule-1 (VCAM-1) binds to the cognate ligand VLA-4 that is expressed by the very types of leucocytes recruited to the intima during early atherogenesis: monocytes and lymphocytes [8]. Cholesterol-fed rabbits and mice deficient in apolipoprotein E, and hence prone to development of hyperlipidaemia and atheroma, express endothelial VCAM-1 early on at sites of lesion formation [910]. In rabbits with diet-induced hypercholesterolaemia, focal expression of VCAM-1 on aortic endothelial cells occurs before monocyte adhesion. These data support a role for VCAM-1 in atheroma initiation. Mice deficient in VCAM-1 have major developmental abnormalities, thus frustrating attempts to study atherogenesis in these animals [11].

Once adherent, the leucocytes enter the artery wall. Current evidence suggests that certain chemoattractant chemokines direct the migration of white cells into the intima during atherogenesis. Mice lacking the macrophage chemoattractant protein-1 (MCP-1) molecule exhibit attenuated atherogenesis and mononuclear phagocyte accumulation in the artery [12]. Deletion of the receptor for MCP-1 yields a similar phenotype [13]. Recent work has localized a trio of lymphocyte chemoattractants to the human atheroma. Vascular cells produce these three chemokines (IP-10, MIG, and I-TAC) when exposed to the inflammatory mediator interferon-gamma, a molecule elaborated by activated T cells, and perhaps macrophages as well in the atheromatous lesion.

What factors may signal the focal increase in adhesion molecule and cytokine expression at sites of lesion predilection? Modified lipoproteins contain various oxidized phospholipids such as lysophosphatidyl choline, plamitoyl-oxovaleroyl-glycero-phosphoryl choline (POVPC), palmitoyl-glutaroyl-glycero-phosphoryl choline (PGPC) and epoxyisoprostane E2-glycero-phosphocholine (PEIPC) [1415]. Evidence is mounting that such constituents of oxidatively modified lipoproteins can elicit expression of adhesion molecule and cytokines implicated in early atherogenesis. Local shear stress alterations may also influence adhesion molecules either directly or indirectly [16].

Regulation of the expression of adhesion molecules occurs by negative control as well as at the level of gene transcription. For example, the well known endogenous mediator nitric oxide (·NO), usually thought of as a vasodilator, can reduce leucocyte adhesion to arteries [17]. Cell culture studies have shown that ·NO reduces adhesion of mononuclear leucocytes to human endothelial cells. Additionally ·NO can counteract the induction of VCAM-1 expression by endothelial cells stimulated by such inflammatory cytokines as interleukin-1 or tumour necrosis factor-alpha [18]. At a transcriptional level, nitric oxide inhibits VCAM-1 gene expression in endothelial cells by interfering with the nuclear factor kappa B (NFκB) signalling pathway via a novel, non-cyclic GMP-mediated mechanism [19]. Instead ·NO inhibits NFκB by inducing its endogenous inhibitor IκB alpha. Thus ·NO acts as an anti-inflammatory mediator as well as a vasodilator.

In areas of normal arterial blood flow, laminar shear stress augments the activity of endothelial ·NO synthase, the enzyme that produces endogenous ·NO. Thus, the endogenous anti-inflammatory action of ·NO should operate at sites of undisturbed arterial flow. Local formation of ·NO should limit the ability of atherogenic stimuli, many of which signal through NFκB, to augment expression of adhesion molecules such as VCAM-1 and chemokines such as MCP-1. Disturbed flow at sites prone to early atheroma formation, such as branches and bifurcations, probably attenuate this endogenous anti-inflammatory pathway. This example illustrates how fundamental studies at the molecular level may shed new light on crucial unanswered questions surrounding atherogenesis, including why lesions tend to form in regions of disturbed blood flow such as branch points or near flow dividers in arteries [16–20].

Once mononuclear leucocytes collect in the intima, they typically accumulate lipid and become foam cells, the hallmark of the early atheromatous precursor, the fatty streak. Recent discoveries have afforded considerable insight into the molecular basis of lipid accumulation by macrophages. For example, we now recognize a number of so-called ‘scavenger receptors’, cell surface molecules that preferentially bind the modified forms of lipoproteins associated with atherogenesis [21]. The scavenger receptor A family is now joined by CD36 [22], macrosialin or CD68 [23–25], and a putative HDL receptor known as SRB-1 [26–28]. Studies with genetically engineered mice that have mutations that inactivate the functions of these various receptors are currently helping to define the functional roles of these various candidate receptors for lipids during foam cell formation in vivo. For example, interruption of scavenger receptor A function reduces lesion formation in hyperlipidaemic mice [29].

Progression of atheroma

Accumulation of macrophage foam cells, the hallmark of fatty streaks, may be reversible and does not itself cause clinical consequences. However, macrophage accumulation within the arterial intima sets the stage for progression of atheroma and evolution into more fibrous and eventually more complicated plaque that can indeed cause clinical disease. Accumulation of smooth muscle cells and their elaboration of extracellular matrix macromolecules may contribute importantly to formation of fibrous lesions during this phase of atheroma progression.

Desquamation of the endothelium due to injury was previously deemed responsible for causing adherence and degranulation of platelets, and the release of fibrogenic substances that could promote smooth muscle cell proliferation and extracellular matrix accumulation, including platelet-derived growth factor (PDGF) [3031]. Currently, we appreciate that atheroma can form in the absence of actual sloughing of endothelial cells. Mononuclear phagocytes, precursors of the atheroma’s characteristic foam cells, can insinuate themselves between intact endothelial cells and enter the intima by diapedesis [32–35]. Indigenous vascular wall cells or infiltrating leucocytes may themselves elaborate mediators such as PDGF. Endothelial cells can produce both isoforms of PDGF, a heterodimeric molecule comprising combinations of two polypeptide chains, denoted the A and B chains [36–38]. Human vascular smooth muscle cells can express the A chain of PDGF in vitro, but they cannot express the B chain [39–42]. Macrophages may also express PDGF genes [43] and those encoding other growth factors such as, inter alia, heparin-binding epidermal growth factor [4445], forms of fibroblast growth factor, and insulin-like growth factors. At later phases of lesion formation, platelets can indeed release fibrogenic mediators at sites of uni- or oligocellular desquamation of endothelium causing mural microthrombi.

The above discussion of lesion initiation highlights the possible involvement of inflammatory cytokines. These protein mediators may also regulate growth factor expression by vascular cells and leucocytes. For example, interleukin-1, a prototypic cytokine, increases production of PDGF A chain by human vascular smooth muscle cells [46]. Interleukin-1 can also augment basic fibroblast growth factor expression by human smooth muscle cells [47]. These examples illustrate the concept that cytokines can elicit secondary expression of a variety of growth-promoting genes by vascular cells and leucocytes.

As in many biological control pathways, the balance of opposing forces determines the outcome. Smooth muscle cells receive growth stimulatory signals as well as those that promote their proliferation. Transforming growth factor beta can inhibit smooth muscle cell proliferation whilst at the same time stimulating their production of extracellular matrix. Interferon gamma, a cytokine derived from activated T lymphocytes, can inhibit smooth muscle cell proliferation and matrix synthesis. Endogenous heparin sulphate glycosaminoglycans can also limit smooth muscle cell division. Thus, smooth muscle cell accumulation depends on the equilibrium between growth-stimulatory and growth-inhibitory stimuli, both limbs of control that are tightly regulated themselves.

Progressing atheroma often accumulate calcium. This aspect of atherogenesis also appears to depend critically on biological functions of lesional cells regulated by specific molecular messengers. Virchow described chondrous metaplasia and bone formation in atheroma in the mid-19th century [48]. Recent work has characterized the expression by vascular smooth muscle cells of proteins involved in bone formation and mineralization [49]. For example, smooth muscle cells can express osteopontin [50–52]. Members of the transforming growth factor family, such as bone morphogenetic proteins, may also act on vascular wall cells to promote mineralization of atheroma [53]. The accumulation of mineral in atheroma may depend upon both degradative and synthetic processes. Osteoclasts, the cell type involved in bone resorption, are actually a special form of mononuclear phagocyte. Macrophages within atheroma probably act like osteoclasts in advancing atheroma. Atherosclerosis-prone mice lacking macrophage colony-stimulating factor (M-CSF), a macrophage activator, exhibit increased calcium accumulation in atheroma [5455]. Far from being a passive or inevitable degenerative process, lesion mineralization also appears to depend upon closely controlled or positive and negative loops, as previously discussed with regard to smooth muscle cell growth and extracellular matrix deposition.

The molecular mechanisms of the thrombotic complications of atheroma

Arterial stenoses by themselves seldom cause acute unstable angina or acute myocardial infarction. Indeed, sizeable atheroma may remain silent for decades or produce only stable symptoms such as angina pectoris precipitated by increased demand such as exertion. However, seemingly without warning, such stable lesions may cause dreaded acute manifestations of atherosclerosis, such as acute myocardial infarction or stroke. Recent research has furnished new insight into the molecular mechanisms that cause transition from the chronic to the acute phase of atherosclerosis [56]. We now appreciate that thrombosis actually causes most of the acute manifestations of this disease. Formerly, we presumed that arteries with critical stenoses tended to thrombose and precipitate acute manifestations of atherosclerosis. We have now learned that the degree of luminal obstruction by an atheroma has little relation to its likelihood of causing thrombosis. The majority of acute myocardial infarctions result from atheroma that cause less than a 50% stenosis of the artery, as assessed by arteriography [57].

Instead, thrombus formation usually occurs because of a physical disruption of atherosclerotic plaque [565859]. Because wall tension varies directly with radius (the Laplace relationship), the biomechanical stresses experienced by non-obstructing atheroma may be greater than that of stenoses, which yield a smaller residual lumen [60]. The physical disruption of the lesion that causes the thrombosis may be a superficial erosion that permits platelets to contact the pro-aggregatory collagen in the intima’s basement membrane [6162]. However, the majority of coronary thromboses result from a rupture of the plaque’s protective fibrous cap, which permits contact between blood and the highly thrombogenic material located in the lesion’s lipid core, e.g. tissue factor. Because of the critical role of plaque rupture in coronary thrombosis, the biomechanical strength of the plaque’s fibrous cap has considerable importance as a determinant of a particular lesion’s stability. Approaching this issue in molecular terms, one must recognize that interstitial forms of collagen account for most of the tensile strength of the plaque’s fibrous cap. Therefore, we must understand the metabolism of the macromolecules of the extracellular matrix to delineate the mechanism of rupture of the atherosclerotic plaque ( Fig. 2).

Figure 2.

Inflammation and the regulation of collagen metabolism in the plaque’s intima (see text for explanation). Figure reproduced with permission from Libby (1995) [56].

The amount of collagen in the lesion’s fibrous cap depends upon its rate of biosynthesis by the arterial smooth muscle cell, the main source of collagens in arteries. Certain factors released from degranulating platelets, including transforming growth factor beta or PDGF, stimulate collagen synthesis by vascular smooth muscle cells [63]. In contrast, gamma interferon, which is produced by activated T cells, markedly inhibits interstitial gene expression and protein synthesis in these cells [63]. This latter finding has particular bearing on the pathophysiology of plaque rupture because T lymphocytes accumulate at sites where plaques rupture and cause fatal thrombosis [61]. Interestingly, cells at places where human plaques actually rupture express class II major histocompatibility antigens, cell surface structures involved in T-cell activation [61]. In human smooth muscle cells, gamma interferon, but not a variety of other stimuli tested, can induce the expression of the class II molecules [6465]. Thus, expression of class II molecules in these regions indicates the presence of interferon gamma, the T-cell product that inhibits collagen synthesis. Compared to regions lacking T-cell infiltrates, pathological studies have shown low levels of interstitial collagen mRNA and protein at sites of T-cell accumulation in human atheromata [66]. Together, these results suggest that release of inflammatory mediators such as interferon gamma by leucocytes in particular places in atherosclerotic plaques may predispose them to rupture by impeding the ability of smooth muscle cells to maintain and repair the collagen crucial to the integrity of the plaque’s fibrous cap [56].

In addition to synthesis, degradative processes can influence the level of collagen in the plaque’s fibrous cap and thereby affect its tensile strength. Several specialized enzymes can degrade collagen, elastin and other structurally key components of the artery’s extracellular matrix [67]. Enzymes of the matrix metalloproteinase family can attack interstitial collagen fibrils, molecules ordinarily exceedingly resistant to proteolytic degradation. Activated macrophages within atheroma can elaborate a number of these matrix-degrading enzymes [68–70]. Experiments on cultured mononuclear phagocytes and resident cells of the artery wall have shown that inflammatory mediators such as cytokines augment the expression of matrix metalloproteinase genes [71–74]. Recent work has localized the potent elastases, cathepsins S and K, in macrophages and smooth muscle cells located in complicated human atherosclerotic plaques [75]. Thus, members of several proteinase families may participate in degradation of structurally important constituents of the arterial extracellular matrix. As in the case of many protease cascades in biological control, these protease families have endogenous inhibitors. All known members of the tissue inhibitor of matrix metalloproteinase family (TIMPs 1–4) have been localized in human atheroma [6976], as has the endogenous inhibitor of cathepsin S, cystatin.

Pathological study of ruptured human atherosclerotic plaques has defined other features beyond the presence of a thin and collagen-poor fibrous cap that are characteristic of so-called vulnerable plaques. For example, plaques that have actually ruptured and cause thrombosis usually also have large numbers of macrophages and T cells along with a few smooth muscle cells [6177]. We have proposed that smooth muscle cell death, perhaps by apoptosis or programmed cell death, may contribute to the paucity of smooth muscle cells in vulnerable plaques [78]. Indeed, some smooth muscle cells in atheroma have fragmented DNA and other features characteristic of programmed cell death [78–80]. In vitro studies have shown that inflammatory cytokines found in atheroma can trigger the apoptotic programme in human vascular smooth muscle cells [81]. Smooth muscle cells in the atherosclerotic intima can express fas, the receptor for the death-signalling fas ligand found on the surface of certain T cells [8283]. Atheromata contain fas ligand-bearing T cells [8283]. Thus, far from being innocent bystanders, the leucocytes in the atherosclerotic plaque probably participate actively in the acute coronary syndromes as well as in lesion initiation and progression. These findings furnish additional connections between inflammation and the pathophysiology of atherosclerosis and its clinical complications.

Conclusion

The foregoing discussion gives some examples of how recent progress in the molecular mechanisms of atherogenesis has increased our understanding of this disease at several levels. Certain major common themes emerge from this body of work. We have learned how the balance between positive and negative regulation factors can critically influence all stages of atherosclerosis. Induction of leucocyte adhesion molecules by cytokines and inhibition by nitric oxide exemplify this balance in processes pivotal to lesion initiation. Progression of lesions from fatty streaks to fibrous plaques depends upon a balance between smooth muscle growth and death; each of these processes is in turn dependent upon a balance between positive and negative stimuli. An altered balance between extracellular matrix synthesis and degradation, or matrix-degrading proteinases and their inhibitors can weaken the plaque’s fibrous cap or favour endothelial detachment that predisposes to the acute thrombotic complications of atherosclerosis.

The examples discussed above also illustrate how application of cellular and molecular approaches has fostered a deeper understanding of the pathogenesis of atherosclerosis. This increased knowledge of the basic mechanisms enables us to understand how current therapies for atherosclerosis may act. Moreover, the insights derived from recent scientific advances should aid the discovery of new therapeutic targets that would stimulate development of novel treatments [8485]. Such new treatments could further reduce the considerable burden of morbidity and mortality due to this modern scourge, and reduce reliance on costly technologies that address the symptoms rather than the cause of atherosclerosis.

Acknowledgements

Work from the author’s laboratories described above was supported by grants from the National Heart, Lung and Blood Institute to PL (HL 34636). We thank our colleagues Galina Sukhova, Uwe Schoenbeck, Masanori Aikawa, Yong-Jian Geng, Lynne Henderson, James K. Liao, Richard T. Lee, Maria Muszynski and Elissa Simon-Morrissey for their invaluable contributions to this work.

Received 28 October 1999; accepted 25 November 1999.

Ancillary