|Home | About | Journals | Submit | Contact Us | Français|
Atherosclerosis, a disease of the large arteries, is the primary cause of heart disease and stroke. In westernized societies, it is the underlying cause of about 50% of all deaths. Epidemiological studies have revealed several important environmental and genetic risk factors associated with atherosclerosis. Progress in defining the cellular and molecular interactions involved, however, has been hindered by the disease’s aetiological complexity. Over the past decade, the availability of new investigative tools, including genetically modified mouse models of disease, has resulted in a clearer understanding of the molecular mechanisms that connect altered cholesterol metabolism and other risk factors to the development of atherosclerotic plaque. It is now clear that atherosclerosis is not simply an inevitable degenerative consequence of ageing, but rather a chronic inflammatory condition that can be converted into an acute clinical event by plaque rupture and thrombosis.
Atherosclerosis is a progressive disease characterized by the accumulation of lipids and fibrous elements in the large arteries. The anatomy of a normal artery is shown in Fig. 1. The early lesions of atherosclerosis consist of subendothelial accumulations of cholesterol-engorged macrophages, called ‘foam cells’. In humans, such ‘fatty streak’ lesions can usually be found in the aorta in the first decade of life, the coronary arteries in the second decade, and the cerebral arteries in the third or fourth decades. Because of differences in blood flow dynamics, there are preferred sites of lesion formation within the arteries. Fatty streaks are not clinically significant, but they are the precursors of more advanced lesions characterized by the accumulation of lipid-rich necrotic debris and smooth muscle cells (SMCs). Such ‘fibrous lesions’ typically have a ‘fibrous cap’ consisting of SMCs and extracellular matrix that encloses a lipid-rich ‘necrotic core’. Plaques can become increasingly complex, with calcification, ulceration at the luminal surface, and haemorrhage from small vessels that grow into the lesion from the media of the blood vessel wall. Although advanced lesions can grow sufficiently large to block blood flow, the most important clinical complication is an acute occlusion due to the formation of a thrombus or blood clot, resulting in myocardial infarction or stroke. Usually, the thrombosis is associated with rupture or erosion of the lesion.
The events of atherosclerosis have been greatly clarified by studies in animal models, including rabbits, pigs, non-human primates and rodents. Mice deficient in apolipoprotein E (apoE) or the low-density lipoprotein (LDL) receptor develop advanced lesions and are the models most used in genetic and physiological studies1. Figure 2 shows stages in the development of atherosclerotic plaques in experimental animals. The first observable change in the artery wall following the feeding of a high-fat, high-cholesterol diet is the accumulation of lipoprotein particles and their aggregates in the intima at sites of lesion predilection (Fig. 2a, b). Within days or weeks, monocytes can be observed adhering to the surface of the endothelium. The monocytes then transmigrate across the endothelial monolayer into the intima, where they proliferate, differentiate into macrophages and take up the lipoproteins, forming foam cells (Fig. 2c, d)2. With time, the foam cells die, contributing their lipid-filled contents to the necrotic core of the lesion. Some fatty streaks subsequently accumulate SMCs, which migrate from the medial layer. With the secretion of fibrous elements by the smooth muscle cells, occlusive fibrous plaques develop and increase in size. Initially, the lesions grow towards the adventitia until a critical point is reached, after which they begin to expand outwards and encroach on the lumen. The lesions continue to grow by the migration of new mononuclear cells from the blood, which enter at the shoulder of the vessel; this is accompanied by cell proliferation, extracellular matrix production and the accumulation of extracellular lipid (Fig. 2e). Atherogenesis can be viewed as a ‘response to injury’, with lipoproteins or other risk factors as the injurious agents2,3.
Epidemiological studies over the past 50 years have revealed numerous risk factors for atherosclerosis (Table 1). These can be grouped into factors with an important genetic component, and those that are largely environmental. The relative abundance of the different plasma lipoproteins appears to be of primary importance, as raised levels of atherogenic lipoproteins are a prerequisite for most forms of the disease. With the exception of gender, and the level of lipoprotein(a), each of the genetic risk factors involves multiple genes. This complexity can be clearly observed in genetic crosses in animals maintained under similar environmental conditions; such studies in rodents have revealed dozens of genetic loci that contribute to lipoprotein levels, body fat and other risk factors4. Another level of complexity involves the interactions between risk factors. Frequently, these are not simply additive; for example, the effects of hypertension on coronary heart disease (CHD) are considerably amplified if cholesterol levels are high5.
The importance of genetics and environment in human CHD has been examined in many family and twin studies6. Within a population, the heritability of atherosclerosis (the fraction of disease explained by genetics) has been high in most studies, frequently exceeding 50%. Population migration studies, on the other hand, clearly show that the environment explains much of the variation in disease incidence between populations. Thus, the common forms of CHD result from the combination of an unhealthy environment, genetic susceptibility and our increased lifespan5.
Pathological studies have revealed a defined series of changes in the vessel during atherogenesis (Fig. 2) and showed that blood-derived inflammatory cells, particularly monocytes/macrophages, have a key role. Tissue culture studies with vascular cells and monocytes/macrophages suggested possible pathways of disease initiation and progression. They provided evidence for the central role of the endothelium in mediating inflammation, and suggested that accumulation of oxidatively modified LDL in the intima contributes significantly to monocyte recruitment and foam-cell formation. During the past decade, understanding of the molecular mechanisms in atherogenesis has been revolutionized by studies in transgenic and gene-targeted mice7. These have allowed in vivo testing of hypotheses, although it should be noted that studies in mice are limited by significant species differences compared with humans, and that reliable mouse models for thrombosis involving lesion rupture have not been developed.
The endothelium, with its intercellular tight junctional complexes, functions as a selectively permeable barrier between blood and tissues. It has both sensory and executive functions, and can generate effector molecules that regulate thrombosis, inflammation, vascular tone and vascular remodelling. For example, removal of the endothelium results in a burst of SMC migration and proliferation, which subsides when the endothelium regenerates8. Among the important physical forces acting on endothelial cells (ECs) is fluid shear stress, which has effects on EC morphology. Cells in the tubular regions of arteries, where blood flow is uniform and laminar, are ellipsoid in shape and aligned in the direction of flow. Cells in regions of arterial branching or curvature, where flow is disturbed, have polygonal shapes and no particular orientation. These latter areas show increased permeability to macromolecules such as LDL and are preferential sites for lesion formation8.
As shown in Fig. 3, a primary initiating event in atherosclerosis is the accumulation of LDL in the subendothelial matrix. Accumulation is greater when levels of circulating LDL are raised, and both the transport and retention of LDL are increased in the preferred sites for lesion formation. LDL diffuses passively through EC junctions, and its retention in the vessel wall seems to involve interactions between the LDL constituent apolipoprotein B (apoB) and matrix proteoglycans9. In addition to LDL, other apoB-containing lipoproteins, namely lipoprotein(a) and remnants, can accumulate in the intima and promote atherosclerosis. Lipoprotein(a), a particle resembling LDL but containing an additional polypeptide termed apolipoprotein(a) that is linked to apoB by a disulphide bridge, seems to be particularly atherogenic owing to its additional effects on fibrinolysis and SMC growth10.
Native LDL is not taken up by macrophages rapidly enough to generate foam cells, and so it was proposed that LDL is somehow ‘modified’ in the vessel wall11. It has subsequently been shown that trapped LDL does indeed undergo modification, including oxidation, lipolysis, proteolysis and aggregation, and that such modifications contribute to inflammation as well as to foam-cell formation. One of the modifications most significant for early lesion formation is lipid oxidation as a result of exposure to the oxidative waste of vascular cells. Such modifications initially give rise to ‘minimally oxidized’ LDL species that have pro-inflammatory activity but may not be sufficiently modified to be recognized by macrophage scavenger receptors. Mice lacking 12/15-lipoxygenase have considerably diminished atherosclerosis, suggesting that this enzyme may be an important source of reactive oxygen species in LDL oxidation12. Lipoxygenases insert molecular oxygen into polyenoic fatty acids, producing molecules such as hydroperoxyeicosatetraenoic acid (HPETE), which are likely to be transferred across the cell membrane to ‘seed’ the extracellular LDL.
High-density lipoprotein (HDL) is strongly protective against atherosclerosis. An important mechanism underlying this protective effect is the role of HDL in the removal of excess cholesterol from peripheral tissues. But in addition, HDL also protects by inhibiting lipoprotein oxidation. The antioxidant properties of HDL are due in part to serum paraoxonase, an esterase carried on HDL that can degrade certain biologically active oxidized phospholipids13,14.
Atherosclerosis is characterized by the recruitment of monocytes and lymphocytes, but not neutrophils, to the artery wall (Fig. 4). A triggering event for this process is the accumulation of minimally oxidized LDL, which stimulates the overlying ECs to produce a number of pro-inflammatory molecules, including adhesion molecules and growth factors such as macrophage colony-stimulating factor (M-CSF). The biological activity of minimally oxidized LDL is contained primarily in its phospholipid fraction, and three active oxidation products resulting from the scission or rearrangement of unsaturated fatty acids have been identified15. Oxidized LDL can also inhibit the production of nitric oxide (NO), a chemical mediator with multiple anti-atherogenic properties, including vasorelaxation. Mice lacking endothelial NO synthase showed enhanced atherosclerosis, due in part to raised blood pressure16. In addition to oxidized LDL, a number of other factors are likely to modulate inflammation, including haemodynamic forces, homocysteine levels, sex hormones, and infection. Diabetes may promote inflammation in part by the formation of advanced endproducts of glycation that interact with endothelial receptors17.
The entry of particular types of leukocytes into the artery wall is mediated by adhesion molecules and chemotactic factors. After cultured ECs are exposed to oxidized LDL, they will bind monocytes but not neutrophils. The first step in adhesion, the ‘rolling’ of leukocytes along the endothelial surface, is mediated by selectins which bind to carbohydrate ligands on leukocytes. Studies of mice deficient in P- and E-selectins or the cell adhesion molecule ICAM, revealed the role of these adhesion molecules in atherosclerosis18,19. The firm adhesion of monocytes and T cells to endothelium can be mediated by the integrin VLA-4 on these cells, which interacts with both VCAM-1 on the endothelium and the CS-1 splice variant of fibronectin. Both in vitro and in vivo studies suggested that these interactions have a role in atherosclerosis20. Finally, mice deficient in monocyte chemotactic protein (MCP-1) or its receptor CCR2 had significantly reduced atherosclerotic lesions, suggesting that MCP-1/CCR2 interaction has a role in monocyte recruitment in atherosclerosis21,22.
The cytokine M-CSF stimulates the proliferation and differentiation of macrophages, and influences various macrophage functions such as expression of scavenger receptors. Mice with a spontaneous null mutation of M-CSF had dramatically reduced lesions, suggesting an obligatory role for macrophages in lesion formation23.
LDL must be extensively modified (‘highly oxidized’) before it can be taken up sufficiently rapidly by macrophages to form foam cells (Fig. 5). This modification presumably involves reactive oxygen species produced by ECs and macrophages, but several enzymes are also thought to be involved, including myeloperoxidase, sphingomyelinase and a secretory phospholipase, all of which occur in human atherosclerotic lesions. Myeloperoxidase generates highly reactive species such as hypochlorous acid and tyrosyl radical, and myeloperoxidase-modified LDL binds to macrophage scavenger receptors24. Sphingomyelinase may promote lipoprotein aggregation, leading to increased retention and enhanced uptake by macrophages25. Finally, a secretory phospholipase (group II sPLA2) can promote LDL oxidation, and transgenic mice overexpressing the enzyme show increased atherosclerosis26.
The rapid uptake of highly oxidized (and otherwise modified) LDL particles by macrophages, leading to foam-cell formation, is mediated by a group of receptors that recognize a wide array of ligands. Two such ‘scavenger’ receptors, SR-A and CD36, appear to be of primary importance, and mice lacking either receptor show a modest reduction in atherosclerotic lesions27,28. The expression of scavenger receptors is regulated by peroxisome proliferator-activated receptor-γ, a transcription factor whose ligands include oxidized fatty acids, and by cytokines such as tumour necrosis factor-α and interferon-γ (IFN-γ)29.
Macrophages actively secrete apoE, and this may promote cholesterol efflux to HDL, thereby inhibiting the transformation of macrophages to foam cells. Evidence for this role of apoE comes from bone marrow transplantation studies showing that mice transplanted with marrow from apoE-null mice develop much larger lesions than mice receiving marrow from control mice30. Interestingly, mice deficient in ACAT1, the enzyme responsible for cholesterol esterification in macrophages, are still able to develop significant lesions31.
Fibrous plaques are characterized by a growing mass of extracellular lipid, mostly cholesterol and its ester, and by the accumulation of SMCs and SMC-derived extracellular matrix (Fig. 6). Cytokines and growth factors secreted by macrophages and T cells are important for SMC migration and proliferation and extracellular matrix production.
Recent studies have shown that the interaction of CD40 with its ligand CD40L (CD154) makes an important contribution to the development of advanced lesions32. This interaction was first recognized as being essential to major immune reactions involving T and B cells, but it is now clear that CD40 is also expressed on macrophages, ECs and SMCs. The engagement of CD40 and CD40L results in the production of inflammatory cytokines, matrix-degrading proteases and adhesion molecules. Studies using CD40L-null mice or neutralizing antibodies to CD40L have shown that disruption of the interaction results in smaller lesions that are less inflammatory and more fibrous32. Although studies with immunodeficient mice originally indicated a modest role of lymphocytes in atherogenesis33, studies of CD40–CD40L32, of antibodies to oxidized LDL epitopes34, and of the T-lymphocyte product IFN-γ35 are consistent with a major role for lymphocytes.
Several risk factors seem to contribute to the development of fibrous lesions, including elevated homocysteine, hypertension and hormones. Elevated homocysteine levels appear to injure ECs and to stimulate proliferation of vascular SMCs36. Some of the effects of raised blood pressure on atherosclerosis seem to be mediated by components of the renin–angiotensin pathway. For example, angiotensin II directly stimulates SMC growth and the production of extracellular matrix. Studies with spontaneously hypertensive rats (SHR) indicate that raised blood pressure stimulates expression of platelet-derived growth factor, a potent mitogen for SMCs37. Oestrogen has multiple anti-atherogenic properties, including effects on plasma lipoprotein levels and stimulation of prostacyclin and NO production38.
Infection by cytomegalovirus has been linked to atherosclerosis and arterial restenosis (a narrowing of the vessel lumen due to vascular remodelling following angioplasty)35. On the basis of in vitro studies, a plausible mechanism for this link is stimulation of SMC migration by the virus-coded chemokine receptor US28 (ref. 39). Cytomegalovirus infection is also associated with inactivation of the p53 protein, and p53-null mice exhibited increased SMC proliferation and accelerated atherosclerosis40.
The monoclonal patchiness of atherosclerotic lesions originally suggested that the disease may involve a nonmalignant transformation of SMCs, but this patchiness has now been shown to result from normal development41. Nevertheless, evidence consistent with oncogene activation, loss of heterozygosity and microsatellite instability in human lesions has been reported5.
Pathological studies suggest that the development of thrombus-mediated acute coronary events depends principally on the composition and vulnerability of a plaque rather than the severity of stenosis (Fig. 7). Vulnerable plaques generally have thin fibrous caps and increased numbers of inflammatory cells. Maintenance of the fibrous cap reflects matrix production and degradation, and products of inflammatory cells are likely to influence both processes. For example, T cells produce IFN-γ, which inhibits the production of matrix by SMCs, and macrophages produce various proteases that degrade extracellular matrix, including interstitial collagenase, gelatinases and stromolysin3. Rupture frequently occurs at the lesion edges, which are rich in foam cells, suggesting that factors contributing to inflammation may also influence thrombosis. In this regard, it is notable that the incidence of myocardial infarction and stroke increases during acute infections.
The stability of atherosclerotic lesions may also be influenced by calcification and neovascularization, common features of advanced lesions. Intimal calcification is an active process in which pericyte-like cells secrete a matrix scaffold which subsequently becomes calcified, akin to bone formation. The process is regulated by oxysterols and cytokines42. The growth of small vessels from the media may provide a conduit for entry of inflammatory cells43.
The thrombogenicity of the lesion core is likely to depend on the presence of tissue factor, a key protein in the initiation of the coagulation cascade. The production of tissue factor by ECs and macrophages is enhanced by oxidized LDL, infection or the ligation of CD40 on ECs to CD40L on inflammatory cells44. The expression of other molecules mediating thrombosis, such as plasminogen activator, may also be important.
Although the common forms of atherosclerosis are multifactorial, studies of rare mendelian forms have provided the most important insights into the disease (Table 2). Studies of familial hypercholesterolaemia helped unravel the pathways that regulate plasma cholesterol metabolism, knowledge of which was important for the development of cholesterol-lowering drugs. In the past year, Tangier disease, a rare recessive disorder characterized by the virtual absence of circulating HDL, was shown to be due to mutations in the gene for the ATP-binding-cassette (ABC) transporter 1, providing an excellent candidate cause for more common forms of HDL deficiency45,46. Recently found mutations in the mineralocorticoid receptor, a kidney protein that is involved in the body’s handling of salt, explain why some women have a sharp rise in blood pressure during pregnancy47.
In contrast to the mendelian disorders, attempts to identify genes for the common, complex forms of atherosclerosis have met with mixed success. Studies of candidate genes have revealed a number that show significant or suggestive association or linkage with traits relevant to atherosclerosis, but our understanding remains incomplete (Table 3). Large-scale sequencing is now underway to identify polymorphisms for many other candidate genes for hypertension, diabetes and other traits relevant to atherosclerosis48. In an attempt to identify atherosclerosis genes, whole-genome scans for loci associated with diabetes, hyperlipidaemia, low HDL levels and hypertension have been performed49. But few loci with significant evidence of linkage have been found, emphasizing the complexity of these traits.
The use of animal models is a potentially powerful way of identifying genes that contribute to common forms of atherosclerosis. Mice and rats—the most useful mammals for genetic studies—have common variations in many traits relevant to atherosclerosis, and orthologous genes frequently contribute to a trait in rodents and humans50. Mapping and identification of genes contributing to complex traits is easier in rodents than in humans, as shown by the recent identification of a diabetes gene in the SHR rat model51. Studies in animal models should be particularly useful for the identification of genetic factors influencing vascular cell functions; for example, differences in susceptibility to atherosclerosis between certain strains of mice seem to be due to variation that affects EC responses to oxidized LDL52. During this decade it is likely that genome-wide approaches, such as expression array studies and large-scale animal mutagenesis studies, will become widely used in atherosclerosis research.
As a result of the genome projects and large-scale sequencing, tens of thousands of single-nucleotide polymorphisms are being identified and a catalogue of all common variations in humans will be generated over the next few years. This raises the possibility of whole-genome association studies. Given the rapid development of DNA chip technology, it should be possible to type large numbers of polymorphisms in many thousands of individuals. There are, however, significant unresolved issues involving linkage disequilibrium and statistical analysis53 in this approach.
Effective drugs for lowering cholesterol and high blood pressure have been developed. In particular, the statins lower levels of atherogenic lipoproteins and dramatically decrease clinical events and mortality from atherosclerosis54. Nevertheless, heart disease and stroke remain by far the most common causes of death in westernized societies, and new weapons, particularly agents that block disease at the level of the vessel wall or that raise anti-atherogenic HDL, are needed.
Over the past decade, a number of promising new targets have been identified, as discussed above and shown in Figs Figs33–7. For example, interruption of the CD40–CD40L system may have clinical benefits for plaque stability32. The identification of the ABC transporter presents exciting new opportunities for treatment of low HDL levels. It has also become clear that HDLs are functionally very heterogeneous55. Thus, rather than attempting to increase levels of HDL, it may be more productive to focus on functional properties such as its antioxidant activity. Preliminary studies in animals suggest that it may be possible not only to block the development of atherosclerosis but also to achieve significant regression56. The most critical clinical aspect of atherosclerosis is plaque rupture and thrombosis. Although useful mouse models for this have not been developed, a transgenic hypertensive and hyperlipidaemic rat model showed evidence of myocardial infarction57.
Catheterization is the gold standard for diagnosis of atherosclerosis, but it is expensive and carries significant risk. Reliable noninvasive methods of diagnosis are urgently needed. Certain biochemical markers for the disease, such as C-reactive protein, and some noninvasive procedures, such as extravascular ultrasound and ultrafast computerized tomography, should prove useful but have limitations.
As our understanding of the genetics of atherosclerosis increases, genetic diagnosis will become increasingly important. The anticipated ‘biallelic map’ of the genome is likely to drive the evolution of new technologies for gene screening, from high-throughput, genome-wide methods to testing for particular gene variants in individuals. One application of screening will be to distinguish different forms of the disease so that pharmacological intervention can be better targeted. Atherosclerosis is heterogeneous, and the most appropriate therapy will depend on the particular variety of disease. Classification is already used clinically, as patients are grouped according to the variety of risk factors they display, but genetic testing should greatly expand the subdivisions of the disease.
Another potential benefit of genetic studies is testing for susceptibility. Because CHD and stroke are disorders of adults, knowledge of a propensity to disease could be available many years before clinical disease develops, permitting early intervention. Testing for LDL, HDL and blood pressure have long been advocated as a way of identifying individuals at increased risk, and other factors have emerged more recently as risk indicators (Table 1). Once the genes contributing to common forms of the disease have been identified, along with the particular mutations involved, DNA-based tests may add greatly to our ability to assess risk. But given the importance of environmental influences and the complex genetic aetiology of atherosclerosis, efficient screening procedures are unlikely to be available in the near future.
I thank R. Chen and K. Wong for help with the preparation of this manuscript and L. Olson for help with the illustrations. Work in my laboratory was supported by NIH grants.