Endothelium

The endothelium is an organ with autocrine, paracrine, and endocrine functions that serve to regulate blood vessel tone and thrombogenicity. The endothelium synthesizes an array of factors that regulate vascular tone (e.g., nitric oxide [NO], hyperpolarizing factors, prostaglandins, endothelin, angiotensin II) and interactions with blood elements (e.g., adhesion molecules, thrombomodulin, plasminogen acti­vators), which all work in balance to maintain vascular homeostasis.

Endothelial dysfunction is regarded as the earliest manifestation of vessel injury and is present before histologic evidence of atherosclerosis.⁶ Hence functional testing of the endothelium is an important tool for investigators studying atherosclerosis. Endothelial function can be detected in vivo by invasive and noninvasive techniques. Catheter-directed intraarterial administration of acetylcholine has been used to detect the presence of endothelium-dependent vasodilatation in the coronary and peripheral vasculature. Paradoxical contraction with low doses of acetylcholine is regarded as positive for endothelial dysfunction in the coronary circulation. An acute increase in blood flow can also stimulate endothelial relaxation. Flow-mediated, endothelium-dependent vasodilatation can be measured noninvasively by high-resolution ultrasound techniques.⁷ The primary mediator of flow mediated vasodilation is NO.⁸ Administration of nitroglycerin can detect the ability of vascular smooth muscle to relax in the presence of exogenous NO and thus detects endothelial-independent relaxation.

Subendothelial Space

The subendothelial space is particularly relevant in atherosclerosis because it is the location where atherogenic particles are retained and modified so that they can be taken up by macrophages and SMCs.⁹ In terms of volume, the major component of the extracellular intimal matrix is proteoglycans, which form a three-dimensional network that fills 60% of the space.⁹ The remaining mass of this compartment is filled with type IV collagen of the basement membrane, type I and III collagen randomly dispersed in this matrix, and occasional elastin fibers. Extending from the proteoglycans are sulfated glycosaminoglycan chains, which serve to interact with positive residues on apolipoproteins. Large versican-like proteoglycans containing chondroitin sulfate readily interacts with apolipoprotein B-100 (apoB-100), the apolipoprotein associated with low-density lipoprotein (LDL).¹⁰ The stoichiometry is such that 1 µg of versican can bind to 50 to 100 µg of LDL at physiologic ionic strength, thus suggesting that several lipoprotein particles are retained on each proteoglycan molecule.^10,11

Smooth Muscle Cells

In children, very few SMCs or macrophages are present in the subendothelial space; however, with age and lesion formation, their concentration increases. In autopsy studies, abundant SMCs are found in the coronary arteries in atherosclerosis-prone areas, but they are relatively sparse in areas resistant to atherosclerosis.¹² Whereas human medial SMCs predominantly express proteins involved in the contractile function of the cell, those found in the intima express lower levels of these proteins, have a higher proliferative index, and exhibit greater synthetic capacity for extracellular matrix, proteases, and inflammatory cytokines.^13,14 These “synthetic” SMCs migrate more readily and can produce up to 25 to 46 times more collagen than “contractile” SMCs can. Thus SMCs directly contribute to intimal thickening as a result of increased proteoglycan production, inflammatory cell recruitment, and retention of atherogenic particles within the subendothelial space.

Vasa Vasorum

The vasa vasorum may contribute to atherosclerosis as well. First-order vasa vasorum run longitudinally to the lumen of the host vessel, whereas second-order vasa vasorum are arranged circumferentially. As plaque volume increases, there is an increase in the vasa vasorum visible on histologic section. Whether this is a response to plaque growth or contributes to plaque growth is unknown. Although the development of vasa vasorum in plaque involves angiogenesis, it is unclear what signals this process. Atherosclerotic vasa vasorum can proliferate and thereby lead to extensive neovascularization directed toward lipid-rich atheroma. The neovessels of plaque may provide an additional portal for trafficking leukocytes to promote the inflammatory process or become friable and contribute to intraplaque hemorrhage.¹⁵

Lipid Hypothesis

Cholesterol has been one of the most studied molecules in biomedical research. Since it was first isolated from gallstones in 1784, organic chemists and biochemists, physiologists and physicists, epidemiologists, and other medical scientists have devoted significant portions of their careers to its study and have collectively produced a body of work culminating in 13 Nobel Prizes.³⁰ The modern era of cholesterol research began in St. Petersburg, Russia, at the turn of the 20th century, when Nikolai Anitschkow produced vascular lesions in rabbits by feeding them purified cholesterol dissolved in sunflower oil.^31,32 These lesions closely resembled those seen in human atherosclerosis. He determined that the earliest lesions consisted of foam cells that were Sudan positive and contained birefringent crystals and hypothesized that they were caused by elevated serum cholesterol. He also noted a distinctive pattern consisting of location of the lesions near arterial branch points and concluded that this was probably determined by hemodynamic factors. Nevertheless, this work was criticized because the levels of cholesterol produced were too high and could not be experimentally reproduced in more conventional animal models such as rats and dogs.^31,32

Cholesterol is absolutely insoluble in water, and the early work done in Russia provided no clues as to how it was transported to the arterial wall and formed lesions. The discovery that cholesterol was associated with proteins that allowed it to be transported in the aqueous environment led to investigations into lipoproteins. The first investigation in which the lipoprotein content in whole serum was accurately quantified involved analytic ultracentrifugation and was led by John Gofman of the Donner Laboratory at Berkley.³³ Lipoprotein fractions were isolated and characterized by their densities and floatation characteristics.³³ Of importance, this group noted that it was not simply the total cholesterol that was important but the species of lipoprotein contained within the cholesterol.³⁴

Cholesterol is transported in the aqueous environment by esterification of the sterol of long-chain fatty acids and packaging of these esters with the hydrophobic cores of plasma lipoproteins. With its polar hydroxyl group esterified, cholesterol remains sequestered within this core, which is essentially an oil droplet composed of cholesteryl esters and triglycerides, solubilized by a surface monolayer of phospholipid and unesterified cholesterol and stabilized by protein. In persons who are fasting, lipids circulate in plasma as lipoprotein particles that are defined on the basis of their density as very-low-density lipoprotein (VLDL), intermediate-density lipoprotein (IDL), LDL, and high-density lipoprotein (HDL). The protein of the VLDL, IDL, and LDL molecule is apoB-100, which has β-electrophoretic mobility.

The major milestone in the lipoprotein field was discovery of the defective gene associated with familial hypercholesterolemia by Goldstein and Brown. In fibroblasts cultured from normal human subjects, the activity of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase, the rate-controlling enzyme in cholesterol biosynthesis, is regulated by the content of LDL (but not HDL) in the culture medium. Goldstein and Brown found that 3-hydroxy-3-methylglutaryl coenzyme A reductase was not sensitive to normal feedback regulation by LDL in cultured skin fibroblasts from homozygous familial hypercholesterolemia patients.^35,36 They then determined that the defect was related to deficient binding of LDL to the cells of patients with familial hypercholesterolemia and that suppression of 3-hydroxy-3-methylglutaryl coenzyme A reductase activity was related to the amount of LDL bound.³⁷ Therefore these patients had overproduction of cholesterol because they lacked the appropriate receptor. They demonstrated that cellular uptake of LDL absolutely requires the LDL receptor, without which the LDL cholesterol concentration builds up to 800 to 1000 mg/dL (20.72 to 25.90 mmol/L). Recognition of the importance of this work resulted in Goldstein and Brown sharing the Nobel Prize in Physiology and Medicine in 1985.

Epidemiologic evidence from numerous studies, including the Japanese migration studies^38,39 and the Framingham Heart Study,⁴⁰ demonstrated the association between cholesterol and incident CHD. Other studies showed that diets rich in unsaturated fats resulted in lowered cholesterol and reduced cardiovascular events.⁴¹

Widespread acceptance of the causative role of cholesterol began with publication of the Coronary Primary Prevention Trial, sponsored by the National Institutes of Health.^42,43 This trial demonstrated that reducing cholesterol with the bile acid–binding resin cholestyramine would reduce cardiovascular events. A decade later, this trial was reinforced with powerful data emerging from the statin era.⁴⁴ Although the primacy of the “cholesterol hypothesis” has been called into question, there is no doubt that lipids play a critical role in the pathogenesis of atherosclerosis.

Response-to-Injury Hypothesis

The basis of this hypothesis lies in the similarities between atherosclerosis and the response of arteries to experimental injury. Although the origins of the hypothesis date back to Virchow, the modern version of the hypothesis builds on earlier work by Duguid (1949),⁴⁵ Poole and Florey (1958),⁴⁶ and French (1966)⁴⁷ and was formally advanced in 1973 by Ross and Glomset,⁴⁸ who emphasized the importance of injury to the endothelium as a seminal event initiating the development of atherosclerosis lesions. Even though the definition of the term injury has been modified over the last three decades,⁴⁹ all response-to-injury hypotheses have emphasized the primacy of the endothelium in stimulating the cascade of events leading to lesion formation.^48–53

Endothelial injury may result from mechanical disruption, exposure to toxic or infectious agents, or endogenous inflammatory signals. Injury to the endothelium allows adhesion of platelets and an influx of LDL and other serum factors into the subendothelial space. Platelets release their alpha granules and stimulate migration of SMCs into the intima, where they proliferate and form a thickened neointima responsible for narrowing of the arterial lumen. Restoration of a healthy endothelial cell layer abates the process. During work on this hypothesis, Ross et al performed experiments leading to the discovery of platelet-derived growth factor (PDGF).^54,55

Others counter that it is not injury to the endothelium that is the initiating event but rather retention of inflammatory lipids in the subendothelial space, which renders a particular area susceptible to atherosclerosis. By noting that LDL accumulates in the intima within 2 hours after a bolus infusion⁵⁶ and that the accumulation occurs before the formation of fatty streaks,^57,58 they concluded that retention of LDL is the initiating event of atherosclerosis. The so-called response-to-retention theory purports that retained apoB–containing lipoproteins stimulate a macrophage- and T cell–dominated inflammatory response in the arterial wall.^59,60

Monoclonal Hypothesis

During the same year that Ross and Glomset were formulating the response-to-injury hypothesis of atherosclerosis, another group of investigators at the same institution (University of Washington) were writing their own hypothesis of atherosclerosis. This hypothesis suggests that each lesion of atherosclerosis is derived from a single SMC that serves as a precursor for the clonal expansion of proliferating SMCs.⁶¹ The hypothesis put forth by Benditt and Benditt uses the concept that in every female cell there is only one active X chromosome and the progeny of that cell will express the same X chromosome as the parent cell. Glucose-6-phosphate dehydrogenase (G6PD) has two isoforms that can be separated by electrophoresis. Its gene is located on the X chromosome and can therefore be used to identify the progeny of a parent cell. This approach was used to determine that uterine leiomyomas are composed of cells with the same active X chromosome, whereas adjacent normal myometrium was composed of a mixture of cells containing both G6PD isoforms and, therefore, that both X chromosomes were active.⁶²

The Benditts examined a series of atherosclerotic plaque from four black females and compared them with adjacent normal areas of arterial wall. They determined that SMCs from the plaque contained only one G6PD isoform, whereas adjacent control areas contained a mixture of isoforms. This allowed them to conclude that each lesion is a clonal outgrowth derived from a single precursor SMC located in the intima. It is noteworthy that SMCs within individual human neonatal intimal thickenings are monoclonal in origin, whereas cells from the subjacent media are polyclonal.

Others challenged the hypothesis by noting that identification of a single enzyme phenotype does not necessarily imply clonal origin.^63,64 Regardless, the work was importantly heuristic. It hinted toward the role that modern molecular biology would play in unraveling the genetic basis of atherosclerosis.

LDL Retention

Trafficking of circulating LDL into and out of the subendothelial space is probably a function of concentration gradients, as well as endothelial permeability.^9,73 Normal intima does not retain LDL particles, thus suggesting that most particles return to plasma or are degraded in situ. However, subendothelial retention of apoB-100–containing lipoproteins is an early event in atherosclerosis.⁷³ Factors favoring net retention include a balance of uptake and degradation of LDL by macrophages, egress of LDL-containing macrophages into the circulation, and complex interaction with proteoglycans such that the LDL particles are sufficiently modified to maintain the gradient. Once bound to the matrix, the lipoprotein is modified so that it becomes oxidized.^9,74 Oxidized LDL, unlike native LDL, is chemotactic to monocytes and rapidly taken up by macrophages to form foam cells by scavenger receptors.

Monocytes

As early as 1958, Poole and Florey noted that macrophages adhere to the surface of endothelial cells overlying atheroma.⁴⁶ In studies of hypercholesterolemic monkeys, monocytes are seen attached to the endothelium within 12 days of initiation of an atherogenic diet.¹⁹ It is now apparent that monocytes play a very early role in atheroma formation.^75,76 Endothelium, which is rendered activated, expresses adhesion molecules that interact with circulating leukocytes, principally monocytes and lymphocytes. E- and P-selectins slow monocytes, mediate rolling, and loosely tether them to the endothelium.⁷⁷ More permanent fixation is due to members of the immunoglobulin superfamily—vascular cell adhesion molecule-1 (VCAM-1)⁷⁸ and intercellular adhesion molecule-1 (ICAM-1)—which are soon upregulated and firmly fix the leukocytes to the wall. The importance of the selectins, VCAM-1, and ICAM-1 in the pathogenesis of atherosclerosis is evident in either ICAM-1 or VCAM-1 mutant animals exhibiting far less atherosclerosis than do wild type controls.^79–81 The adherent leukocytes are stimulated to migrate into the subendothelial space by a number of chemokines, including MCP-1 and oxidized LDL.^82–84

Macrophages

Once within the subendothelial space, monocytes undergo a series of phenotypic modulations and become resident tissue macrophages that take up oxidized LDL via the scavenger receptor A (SR-A), as well as CD36.⁸⁵ One of the key signals for macrophage activation, macrophage colony-stimulating factor, can enhance scavenger receptor expression and promote replication of macrophages and their production of proinflammatory cytokines. Experiments in mutant mice deficient in macrophage colony-stimulating factor have shown the importance of this factor in atheroma formation.^86,87 Although most foam cells are macrophage in origin, SMCs also take up lipids via scavenger receptors and CD36.

In mice, proinflammatory subsets of macrophages called M1-macrophages, induced by hyperlipidemia, produce the inflammatory cytokines IL-1β and TNF-α. The IL-1 gene family encodes three major proteins. The first two, IL-1α and IL-1β, exert proinflammatory effects by binding to the IL-1 receptor type 1. The third is IL-1 receptor antagonist (IL-1Ra), an endogenous inhibitor that competitively blocks the binding of IL-1α and IL-1β to the IL-1 receptor. Direct evidence for the proatherogenic role of IL-1 was obtained in experiments in which apoE^−/− mice received subcutaneous injection of recombinant IL-1 receptor antagonist. Mice injected with IL-1 receptor antagonist displayed a marked reduction in atherosclerotic lesion size. Similarly IL-18, previously called interferon-γ (IFN-γ)–inducing factor, is a Th1-promoting cytokine and therefore has the capacity to promote inflammation through the innate and the adaptive immune pathways.^88,89 Knockout of IL-18 in atherosclerosis-prone animals reduces aortic atherosclerotic lesion size.

Macrophages also contribute to thrombosis in several pivotal ways. As discussed later, atherosclerotic lesions that have ruptured characteristically contain abundant macrophages underlying a thin and collagen-poor fibrous cap. Specific collagenases MMP-1, MMP-8, and MMP-13, co-localizes with macrophages in human atheroma. Rupture of the fibrous cap may be a balance between collagen synthesis by SMCs and collagen breakdown by interstitial (matrix metalloproteinases) collagenases generated by activated macrophages. Studies in collagenase-resistant mutant “knockin” mice crossbred with apoE^−/− mice demonstrated increased intimal collagen and SMC content.⁹⁰ Similarly, in compound mutant apolipoprotein E–null mice crossed with MMP-13/collagenase-3^−/− mice have increased fibrillar collagen that is thicker and more aligned in aortic atherosclerosis.⁹¹ Hence interstitial collagenases produced by activated macrophages greatly influence the structure and integrity of the fibrous cap overlying atherosclerotic plaques. Macrophages and SMCs within atherosclerotic plaques also overexpress the potent procoagulant tissue factor in response to C-reactive protein or CD40 ligand. Hence inflammation links atherosclerosis and thrombosis, leading some to refer to it as atherothrombosis.

Lymphocytes and Adaptive Immunity

The finding in the early 1980s that macrophages expressed major histocompatibility class II antigens needed for antigen presentation to CD4⁺ T cells suggested that adaptive immunity was involved in the atherosclerotic process.⁹² T lymphocytes, which account for as much as 10% to 20% of the leukocyte population, are found most abundantly in the shoulder and fibrous cap region of the atheroma. In advanced lesions, these T cells display markers of chronic activation and produce the prototypical Th1 cytokine, (IFN-γ) which further stimulates expression of class II major histocompatibility antigens in SMCs and macrophages. The crossbreeding of severe combined immunodeficiency (scid)/scid mice that lack T and B cells with apoE^−/− mice produces offspring that are both hypercholesterolemic and immunodeficient. Similar to what was found with apoE^−/−/Rag2^−/− double mutants, when lesions in these mice were compared with those in the immunocompetent apoE^−/− animals, a dramatic 70% reduction of lesion size was observed. However, transfer of oxidized LDL-reactive T cells to apoE^−/−/scid/scid mice is more efficient at lesion acceleration than the transfer of T cells with no specificity to a plaque-derived antigen, demonstrating the importance of antigen presentation in the pathogenesis of atherosclerosis.⁹³

Because of the participation of all components of the innate and adaptive immune system, atherosclerosis resembles other inflammatory and autoimmune diseases such as rheumatoid arthritis and type 1 diabetes mellitus. Potential endogenous autoantigens to activate T cells include LDL or heat shock protein 60.^93,94 Cellular and humeral immune responses are mounted toward these antigens in humans and mice, and protective immunization strategies in mice have provided encouraging results.

In mice, functionally distinct T cell subsets appear to exist in atheroma. Natural killer T cells appear to accelerate atherosclerosis when recognizing lipid antigens presented through CD1 molecules.⁹⁵