This review attempts to define the early events that lead to lesions of human atherosclerosis based on careful morphological studies in human autopsy specimens. In contrast to most small laboratory animals, diffuse intimal thickening (DIT) is present in human arteries before atherosclerosis develops, particularly in the atherosclerosis-prone arteries such as coronary arteries and abdominal aorta. In the earliest stage of atherosclerosis, lipids deposit eccentrically in the deep layer of DIT to form Type I lesions. These layers are enriched in extracellular matrix (ECM) proteoglycans such as biglycan. Following lipid deposition, macrophages appear in these regions and foam cells are observed (Type II lesions). Such observations support the ‘response-to-retention’ hypothesis that states that a principle early event in the pathogenesis of human atherosclerosis is the trapping and retention of lipoproteins by ECM proteoglycans followed by infiltration and accumulation of macrophages.
Despite the fact that millions of dollars have been spent over the last 50 years on atherosclerosis research, little is known about the development of early human atherosclerosis. There are several reasons why the research on early human atherosclerosis has not advanced. First, human atherosclerosis develops very slowly and at different rates from individual to individual, and it is difficult to distinguish between lesion initiation and progression. Second, a thickened intima is present in human arteries before atherosclerosis develops, but whether this intima forms the precursor for the later more advanced lesion is not fully understood. Third, the relationship between extracellular lipids and macrophages has not been clarified. It is generally believed that extracellular lipids are derived from foam cell death, but there are several examples that show that extracellular lipid occurs independently of macrophage cell death. Finally, there are no good animal models for the study of early atherogenesis. The morphological features of early atherosclerosis are different between humans and laboratory animal models, and it may be somewhat misleading to extrapolate the results obtained from animal models to humans.
A number of studies have been published over the years concerning the importance of lipoprotein interaction with proteoglycans in early atherogenesis, supporting the response-to-retention hypothesis (see below). However, it has been unclear whether extracellular lipid and proteoglycan accumulation precede macrophage infiltration in human atherosclerosis. Recent studies show that intimal thickenings are the precursors for the more advanced atherosclerotic lesions and that intimal proteoglycans play important roles in human atherosclerosis.1–5 This review discusses the specific structure of human arteries and the role of proteoglycans in early atherogenesis in humans.
The response-to-retention hypothesis was proposed by Williams and Tabas in 1995 based on early work by a number of investigators and reinforced recently.6,7 This hypothesis states that the retention of atherogenic lipoproteins associated with the extracellular matrix (ECM) in the arterial intima is an initial event in early atherogenesis. It is further postulated that this process begins with predisposing stimuli (e.g. mechanical strain and cytokines) that enhance local synthesis of proteoglycans with high binding affinity for lipoproteins and that atherogenic lipoproteins enter the arterial intima and are bound and retained by proteoglycans (Figure 1A–C).6–12 The interaction between lipoproteins and proteoglycans is ionic in nature and occurs between the negatively charged glycosaminoglycan (GAG) side chains of proteoglycans and the positively charged residues of apolipoproteins.8–10 In addition, the binding can be mediated by accessory molecules, such as lipoprotein lipase (LPL) (Figure 1D).13,14 The hypothesis further states that lipoprotein–proteoglycan complexes exhibit increased susceptibility to modifications, such as oxidation and aggregation, and that these modifications lead to uptake by macrophages to form foam cells (Figure 1E and F).15–17 Furthermore, oxidized lipoproteins enhance the production of proteoglycans with a high affinity to lipoproteins (Figure 1G).18 Thus, the response-to-retention hypothesis emphasizes that the interaction between ECM molecules and lipoproteins and the modification of these two molecules are the key events in early atherogenesis. This hypothesis is not inconsistent with the chronic inflammation hypothesis because retained and modified lipoproteins can stimulate the recruitment of inflammatory cells, such as macrophages and T-cells.
Normal intima and diffuse intimal thickening
Arterial intima is defined as the innermost layer of the artery composed of endothelial cells, subendothelial tissue, and internal elastic lamina. It is generally believed that the subendothelial tissue is thin and consists mainly of ECM with few cells. However, thickened intima composed of fibrocellular tissue is present in human coronary arteries and aorta and increases with age. There are two types of intimal thickenings, i.e. eccentric intimal thickening and diffuse intimal thickening (DIT).19 Eccentric intimal thickening, also known as intimal cushion, is a focal increase in the thickness of the intima associated with branches and orifices. DIT, also known as musculoelastic intimal thickening, is localized to the non-branching long segments of arteries and spreads out circumferentially and longitudinally. These two types of intimal thickenings are contiguous and cannot always be clearly distinguished, and are usually just referred to as ‘DIT’. DIT has been considered by some as an early stage of atherosclerosis. This is probably because similar fibrocellular thickening is observed in some pathologic situations, such as restenosis after percutaneous transluminal coronary angioplasty, cardiac allograft arteriopathy, and balloon injury in experimental animal models. However, we agree with the Council of American Heart Association (AHA) who defined DIT as normal intima, because DIT is universally present in human arteries independent of race, and shows a well-organized structure composed of smooth muscle cells (SMCs), elastin, and proteoglycans (Figure 2A).2,19,20 The SMCs in DIT exhibit very low proliferative activity and widely express cellular FLICE-inhibitory protein that is believed to play an anti-apoptotic role, maintaining a stable SMC phenotype.21,22 No lipid deposits are present and only small numbers of macrophages are seen in the superficial layer of DIT (Figure 2B and C). Neovascularization is not found and angiogenic factors, such as vascular endothelial growth factor and its receptors (flt-1 and Flk-1), are not detected.23,24 No advanced glycation end products are present in DIT less than 40 years of age.25
DIT is believed to be related to atherosclerosis, partly because DIT is consistently present in atherosclerosis-prone arteries.19,26,27 For example, using human autopsy subjects who died between 36 weeks of gestation and 30 years of age, we examined the distribution of DIT in systemic arteries, including atherosclerosis-prone arteries (abdominal and descending aorta and coronary, carotid, and iliac arteries) and atherosclerosis-resistant arteries (ascending aorta and hepatic, mesenteric, splenic, renal, and intracranial cerebral arteries) and found that DIT was expressed specifically in the atherosclerosis-prone arteries and not in the resistant arteries.2 These findings strongly suggested that DIT plays some important role(s) in atherogenesis. DIT is already found in foetuses and infants. In the proximal left anterior descending artery, a thickened intima is present in 33% of the cases 1 month before birth and in all cases by 3 months after birth.28 The intima/media ratio gradually increases with age and to 1.0 or more in young adults.2,28
DIT is composed of two layers. The inner layer is called the proteoglycan layer because it contains ECM with abundant proteoglycans in addition to SMCs. The outer layer is called the musculoelastic layer because of the abundance of SMCs and elastic fibres with smaller amount of proteoglycans. The distinction between the two layers is not clear in some cases. Although the precise mechanisms underlying the formation of DIT are unknown, Stary et al.19 suggested that DIT represents physiological adaptations to mechanical stresses. For example, SMCs and elastic fibres in DIT are oriented longitudinally, in contrast to the medial SMCs and elastic fibres that are arranged in a circumferential manner.29 The thickness of DIT in hypertensive arteries is greater than that in normotensive arteries.30 In a vascular graft model in baboons and in a flow-cessation model in mice, the thickness of the neointima is inversely correlated with the grade of blood flow.31,32 Furthermore, thickened intima is found in the region where shear stress is low in the human abdominal aorta.33 However, this correlation is not always observed for all arteries, and no association between shear stress and intimal thickness has been found in human coronary arteries.34
Early atherogenesis in humans
Extracellular lipid deposition in the deep layer of diffuse intimal thickening
There are only a few studies that examined the pathology of early human atherosclerosis, and most of them were based on gross observations and did not address the relationships between ECM lipid deposits and the occurrence of macrophages. Using coronary arteries of autopsy subjects who died between 7 and 49 years of age, we examined the process of the early phase of human atherosclerosis with light microscopy (Figure 2D–L) and found that human atherosclerosis began with extracellular deposition of apolipoprotein B (apoB)-containing lipids in the outer layer of the pre-existing DIT (Figure 2D and E and Figure 3).1 The deposition occurred eccentrically. This initial stage of atherosclerosis corresponds to the earliest phase of Type I lesion as defined by the AHA classification.35 Recently, extracellular deposits of lipoproteins in the intima of early human lesions were also detected with immunohistochemistry by other investigators.3,4,36
It is generally believed that extracellular lipids originate from dead foam cell macrophages and accumulate to form a lipid core in the advanced lesion. However, there are several studies suggesting that the accumulation of lipids precedes macrophage infiltration in the early phase of atherosclerosis. For example, the accumulation of apoB before the infiltration of macrophages is found in the aortic intima of foetuses and young adults.37,38 Investigating human aorta of young and middle-aged adults with electron microscopy, Guyton and Klemp39 found that an early lipid core arose from lipids accumulating gradually in ECM of the deep intima. Our results confirmed that the extracellular lipid deposition occurs independently of macrophages in the initial stage of atherosclerosis, because no macrophages were present around the extracellular lipids and only a small number of macrophages were present in the superficial layer of the intima as shown in Figure 2E and F.1 The origin of the extracellular lipids is not precisely known, but thought to be derived from plasma. Lipoprotein particles extracted from grossly normal human aortic intima show the same size and lipid composition as those of plasma low-density lipoprotein (LDL).40 Smith and Slater41 found a high percentage of unesterified cholesterol in the deep layers of the amorphous lipid pool in fibrous plaques and concluded that most of the cholesterol in the lipid pool was derived directly from plasma LDL. Interestingly, studies have been recently published which demonstrate that the accumulation of erythrocyte membranes derived from intraplaque haemorrhage and thrombosis contributes to the enlargement of the necrotic core of advanced atherosclerotic plaques.42,43 However, it seems unlikely that degradation of erythrocytes is involved in the formation of lipid pool in early lesions because intraplaque haemorrhage and thrombosis rarely occur in the early phase of atherosclerosis.
Infiltration of macrophages
Our study demonstrated that as the lesion progressed, lipids accumulated in the deep layer of the intima forming a Type I lesion (Figure 2G and H).1 CD68-positive macrophages were marginally increased in number and infiltrated deeper into the forming lesion (Figure 2I).1 Finally, accumulation of foam cell macrophages occurred in the interface between extracellular lipids and infiltrating macrophages to form a Type II lesion (Figure 2J–L).1 These findings strongly suggest that macrophages enter the intima from the circulating blood, infiltrate deeper toward the deposited lipids, and phagocytize the deposited lipids to become foam cells. It is unlikely that macrophages penetrate the base of intima, because macrophages are rarely seen in the underlying inner media.1 In Type II lesions, accumulation of macrophages is found in the adventitia and outer media, suggesting that macrophages also infiltrate the arterial wall from outside (Figure 2L).1
The infiltration of monocyte/macrophages into the arterial wall is regulated by various chemoattractant factors. Monocyte chemoattractant protein-1 (MCP-1) is one of the most potent chemoattractant factors that promote monocyte infiltration. According to in vitro studies, cultured human endothelial cells, SMCs, and macrophages that are stimulated by modified LDL produce MCP-1 and enhance monocyte migration into the subendothelial space.44,45 These events are expected to occur in vivo as well, since these molecules have been identified in lesions by immunohistochemical studies. Enzymatically modified LDL and oxidized LDL (ox-LDL) are found in the extracellular space of the intima of early human lesions and MCP-1 is localized in endothelial cells and subendothelial macrophages in atherosclerotic lesions.1,3,36,46 However, the temporal and spatial relationship between the retention of lipids and the expression of chemoattractant factors needs further scrutiny since it is not clear how the retained lipids in the deep layer of the intima stimulate the cells in the superficial layer, such as endothelial cells, to express chemoattractant factors and enhance macrophage infiltration from the circulating blood.
Roles of intimal smooth muscle cells in early atherogenesis
DIT contains abundant SMCs which are thought to be a principle cell type involved in the early stages of atherogenesis. Aikawa et al.47 examined three types of smooth muscle myosin heavy chain isoforms in human coronary arteries and found that phenotypic modulation of SMCs had already occurred in DIT and was more advanced in the atherogenic plaque. Murry et al.48 found that monoclonal groups of SMCs were present in DIT and suggested that the monoclonality of SMCs found in atherosclerotic plaques arose by expanding the pre-existing clone in DIT. Furthermore, intimal SMCs produce or express various molecules that possibly play roles in cell migration, cell proliferation, and ECM production in atherogenesis. For example, αvβ3 integrin, heparin-binding epidermal growth factor-like growth factor, and transforming growth factor (TGF)-β are produced by and colocalized with SMCs in DIT.49–51 In addition, SMCs express thrombomodulin that may contribute to maintaining an anti-coagulative state of the intima.52 Interestingly, a potent coagulative agent, tissue factor (TF), is also expressed by SMCs, but the expression is weak and confined to the cytoplasm in DIT, in contrast to the advanced lesion in which TF is widely and strongly expressed not only in the cytoplasm of SMCs, but also in the extracellular space.53 This localization in advanced lesions suggests a role of TF in thrombosis in case of the plaque rupture. Furthermore, since SMCs are a major source of proteoglycans and proteoglycans are enriched in DIT, it may be that this layer is pre-disposed to lipid entrapment due to the interaction of the lipoproteins with the proteoglycans.
SMCs as well as macrophages transform to foam cells in the process of atherosclerosis. In early human lesions, macrophage-derived foam cells are present predominantly in the upper layer of the intima, whereas SMC-derived foam cells are found in the deeper musculoelastic layer.39,54 Examining DIT and early atherosclerotic lesions in humans, Kockx et al.55 found that the SMC-derived foam cells in the early lesion showed a strong expression of BAX, a proapoptotic protein, and suggested that these SMCs became susceptible to apoptosis. Apoptosis and loss of SMCs are thought to contribute to the progression of atherosclerosis from early to more advanced lesions. Guyton and Klemp39 examined human early lesions with electron microscopy and found loss of cells in the deep musculoelastic layer when there was an early lipid core, a transitional feature from early to more advanced lesions. The apoptosis of SMCs may be regulated by ox-LDL, because BAX is expressed in ox-LDL positive SMCs,56 and there is a positive correlation between the number of ox-LDL positive cells and apoptotic SMCs in human atherosclerotic lesions.57
Similarities and differences between humans and animal models
Some laboratory mammals are useful models for the study of human atherosclerosis, as they show morphologic similarities in atherosclerotic lesions to those of humans.58,59 However, it is also true that there are some decisive differences between mammalian models and humans as illustrated in Figure 4. First of all, most mammals do not develop prominent DIT-like humans (Figure 4A and B). No thickened intima is present in the aorta and coronary arteries in small laboratory mammals, such as mice, rats, and rabbits, except small thickenings in the aortic arch and branching portion of coronary arteries in rabbits.60 Intimal cushions have been observed in birds. For example, intimal cushions are present at the celiac bifurcation of the aorta in atherosclerosis-susceptible White Carneau and atherosclerosis-resistant Show Racer pigeons but only the White Carneau breed goes on to develop atherosclerotic lesions at these branch sites.61 Such findings suggest that these smooth muscle cushions are necessary, but not sufficient to cause the formation of the atherosclerotic lesion. It is of interest that subsequent studies in this animal model demonstrated differences in the metabolism of the vascular proteoglycans between the susceptible and resistant breeds.62,63 There are different opinions regarding larger mammals. Sims found no intimal thickening in coronary arteries in the dog, sheep, pig, baboon, and cow,64 whereas French described small thickenings of the intima in the thoracic aorta and coronary arteries in the cat, dog, cow, and horse.60 Non-human primates may be akin to humans in terms of the distribution of intimal thickenings. Both eccentric- and diffuse-type intimal thickenings are found in the coronary artery of rhesus monkeys, but the intimal thickness is not well known in this species.65
Minimal deposition of lipids prior to the macrophage infiltration in the intima is found with electron microscopy in the cholesterol-fed rabbits and apolipoprotein E (apoE)-deficient mice, suggesting that the same molecular interactions occur both in experimental animals and humans in the earliest stage of atherosclerosis.66,67 However, it is noteworthy that the environment where lipids deposit is considerably different. In small experimental animals, lipids deposit in the narrow subendothelial space that consists almost exclusively of ECM and contains few SMCs (Figure 4C); whereas, in humans, the lipid deposition occurs initially in the deep layer of the thickened intima of DIT where there are abundant ECM and SMCs (Figure 4D and E). Accumulation of foam cell macrophages in the intima is widely recognized as the initial stage of atherosclerosis in small experimental animals and usually called ‘fatty streak’ according to the AHA definition.58,68 In animal models, foam cell macrophages occupy, almost, the whole thickness of the intima as a predominant component of the lesion (Figure 4F). Very little ECM and few SMCs are present in these lesions; whereas, in human lesions, foam cell accumulation is found in the middle layer of the thickened intima, together with overlying non-foamy macrophages, underlying extracellular lipid deposits, SMCs, and ECM (Figure 4G).27,35 The formation and accumulation of foamy macrophages occur faster in animal models than humans. However, the more important difference is that proliferation of SMCs, production of ECM, and accumulation of extracellular lipids occur after the accumulation of foamy macrophages in animal models while the opposite is true for humans. These differences in early lesions may lead to different outcomes as the lesion progress to more advanced lesions.
Roles of proteoglycans in early atherogenesis
A number of studies indicate that the lipid-binding capacity of proteoglycans contributes to retaining atherogenic lipoproteins in the intima.6,8–10 For example, mice expressing proteoglycan-binding-defective LDL develop significantly less atherosclerosis than mice expressing wild-type control LDL.69 Chondroitin sulfate-rich proteoglycans from cultured human aortic SMCs have a high affinity to lipoprotein(a) and LDL.70 There are over 20 different proteoglycans identified in the vascular wall that are categorized into three groups according to their cellular location, i.e. extracellular proteoglycans, cell surface proteoglycans, and intracellular proteoglycans.71 Versican is a predominant vascular ECM proteoglycan, and biglycan and decorin are the second most quantitatively significant group of extracellular proteoglycans.71 Atherosclerosis-prone arteries, such as the coronary artery, have a thickened intima (DIT) enriched in versican and biglycan, and atherosclerosis-resistant arteries, such as the internal thoracic artery, have a thin intima enriched in decorin.72
Biglycan is a member of chondroitin sulfate/dermatan sulfate containing small leucine-rich proteoglycans and has a binding ability to apoB-containing lipoproteins, i.e. LDL, very-low-density lipoprotein (VLDL), and intermediate-density lipoprotein (IDL).73 Overexpression of human biglycan by rat SMCs results in production of an ECM with increased high-affinity lipoprotein binding.74 Biglycan is present in the intima of normal human blood vessels, including coronary arteries, other muscular arteries, and saphenous veins.72,75 Our immunohistochemical study revealed that biglycan was localized in the outer layer of the DIT in human coronary arteries, as shown in Figure 5.1 This pre-lesional distribution of biglycan is similar to the distribution of lipids in the early phase of Type I lesions (Figure 2E), suggesting that biglycan exists in the intima before lipids deposit and plays a pivotal role in initial deposition of lipids in the arterial intima. Indeed, colocalization of biglycan and apolipoproteins, such as apoB and apoE, are found in early and advanced human and murine lesions.1,5,76 ApoE is an important determinant of high-density lipoproteins (HDL) binding to biglycan, suggesting that biglycan plays a role in retaining not only apoB-containing lipoproteins, but also apoE-containing HDL.77
It is of interest that lipids deposit eccentrically, whereas biglycan is localized concentrically (Figure 5). There are three possibilities that explain this regional difference in the amount of lipid deposits. The first and most important possibility is the regional difference in the proteoglycans. For example, mechanical strain, which is unevenly distributed in the arterial wall, increases mRNA for specific vascular proteoglycans, such as versican, biglycan, and perlecan in cultured SMCs.11 Proteoglycans produced by TGF-β1-treated cultured SMCs show longer GAG chains and greater binding affinity to LDL than those produced by control SMCs.12 Interestingly, TGF-β1 shows patchy distribution in DIT.51 Although experimental data, especially those obtained from mouse models, suggest that TGF-β promotes plaque stabilization by stimulating collagen synthesis, inhibiting foam cell formation by macrophages, and inhibiting T-cell activation,78 there are several human studies showing that TGF-β plays a pro-atherogenic role in early atherosclerosis. Merrilees’ group examined the distribution of TGF-β and synthesis of proteoglycans in organ culture of atherosclerosis-susceptible and resistant blood vessels and concluded that TGF-β is a key factor in upregulation of subendothelial proteoglycan synthesis and is correlated with susceptibility to atherosclerosis.79 Examining the expression of TGF-β and its receptors in human atherosclerotic lesions, Bobik et al.80 suggested that TGF-β is active in lipid-rich lesions and contributes to the pathogenesis of early atherosclerosis. Two other mechanisms that may cause regional difference in lipid distribution in the intima are uneven plasma lipoprotein concentration and the permeability of the arterial wall. Luminal surface concentration of LDL is increased in areas where wall shear stress is low, suggesting that increased surface LDL concentration results in an increased lipid infiltration rate into the intima.81 However, the relationship between permeability of the arterial wall and susceptibility of atherosclerosis is debatable. The permeability of LDL is greater in the atherosclerosis-susceptible areas than in the atherosclerosis-resistant areas of the rabbit aorta,82 but opposite results are obtained in the pigeon aorta.83
Versican interacts with hyaluronan, a long chain GAG, to create expanded viscoelastic pericellular matrices that are required for arterial SMC proliferation and migration.84 Immunohistochemical studies, including our preliminary study, revealed that versican was predominantly localized in the inner layer of DIT of coronary arteries,72 and this localization is likely responsible for the histological manifestation of the inner myxoid layer of DIT. Diffuse distribution of versican across the DIT is also seen in some cases.75 Versican is present throughout both early and advanced atherosclerotic lesions. Although in vitro studies indicate that versican is capable of binding to LDL and has a larger number of biding sites to LDL than biglycan, versican is not often colocalized with lipids and lipoproteins in vivo. For example, in monkey and human atherosclerotic lesions, versican is present primarily in the ECM of SMC-rich areas in early lesions and in the fibrous cap and plaque margins in advanced lesions, but absent in the lipid core.5,85 Versican is not colocalized with apoE and apolipoprotein A (apoA)-I in human atherosclerotic lesions.5 It is noteworthy that versican and hyaluronan are present at the plaque/thrombus interface in erosions, suggesting a possible role in thrombosis.86 This is supported by the fact that platelets bind well to bovine and human aorta versican, but weakly to decorin and biglycan.87 In contrast to human lesions, versican is absent or detected in low levels in mouse lesions.76 This may be due to the macrophage-rich nature of murine atherosclerosis, because macrophages can induce versican degradation and modulate its synthesis.76
Like biglycan, decorin is a member of the small leucine-rich proteoglycan family and is present in the intima of normal blood vessels, although in lower levels than biglycan.72,75 The distribution of decorin in DIT was similar to that of biglycan in our study, but the positively-stained area and the number of positive cases were smaller than those that showed positive staining for biglycan.1 Decorin links LDL with collagen type I in vitro,88 and colocalizes with collagens and apoB in atherosclerotic lesions in vivo,1,89 suggesting that decorin plays a role in atherogenesis by linking lipoproteins to collagen fibres.90 Decorin is also related to cellular events and ECM assembly. Decorin colocalizes with TGF-β1 in atherosclerotic lesions,85 and is suggested to inhibit the responsiveness of SMCs to TGF-β1.91 Mechanical stress causes a decrease in the expression of decorin by human SMCs which could lead disorganization and loosening of collagen in the ECM.11
Perlecan is a heparan sulfate proteoglycan and is present in basement membranes throughout the intima and media of the vascular wall.71 Perlecan is almost absent in early lesions and appears in advanced lesions in areas bordering the plaque core in monkey atherosclerosis.85 Perlecan is the most abundant proteoglycan in murine atherosclerosis. It is found in both early and advanced lesions and is colocalized with apoA-I, suggesting that perlecan acts as a retaining factor for HDL in the atherosclerotic lesion.76 Interestingly, a heterozygous deficiency of perlecan leads to reduced atherosclerosis in apoE-deficient mice in the early phase of atherosclerosis development.92 Perlecan also has other role(s) in atherogenesis, such as uptake of lipoproteins by cells and intracellular lipid accumulation.71 For example, cells expressing perlecan, but no other proteoglycans, bind, internalize, and degrade atherogenic lipoproteins enriched in LPL.93
Accessory molecules and modification of lipoproteins in atherogenesis
LPL is an enzyme that hydrolyzes the triglycerides of chylomicrons, VLDL, and IDL on the luminal surface of the capillary endothelium. LPL is also found on the arterial endothelium and in the intima, and is considered to play a role in atherogenesis. In DIT, LPL is associated with SMCs but not with ECM;94 whereas, in early and advanced atherosclerotic lesions, LPL is not only associated with SMCs and macrophages, but also present in the ECM.13,94 LPL can bind to GAGs on proteoglycans, decorin-coated collagen, and LDL, suggesting that LPL that is bound to proteoglycans and collagens in the arterial intima leads to retention of LDL by acting as a molecular bridge.13 A recent study suggests that retention of LDL is initiated by direct LDL-proteoglycan binding, but shifts to indirect binding with bridging molecules such as LPL.14
Once retained in the intima in association with proteoglycans, atherogenic lipoproteins undergo several modifications with important pathophysiological consequences. Important among these modifications is oxidation. LDL exposed to chondroitin sulfate proteoglycans and GAG is more susceptible to oxidation than native LDL, and macrophages can take up ECM-bound ox-LDL together with GAGs.15,16 Chang et al.18 found that ox-LDL stimulated monkey SMCs to synthesize GAG, increased the molecular size of versican, biglycan, and decorin, and enhanced LDL binding properties. These findings suggest that molecular interactions between extracellular proteoglycans and lipoproteins result in a vicious cycle of atherosclerosis. Another important modification of lipoproteins is aggregation and fusion.17,95 Aggregated and fused lipoprotein particles bind to proteoglycans more tightly than do native lipoproteins.95
In humans, DIT develops in specific arteries, such as coronary arteries and abdominal aorta, and is closely related to susceptibility to developing atherosclerosis. DIT, per se, is not a part of atherosclerosis, but can act as a depot for extracellular lipids in the earliest initial stages of atherosclerosis (Type I lesion). Extracellular proteoglycans localized in DIT, particularly biglycan, are considered to play an important role in early atherogenesis by binding atherogenic lipoproteins. However, it is not known why lipid deposits eccentrically, while proteoglycans are distributed concentrically in DIT. Modifications of proteoglycans may occur unevenly in the arterial wall and cause regional differences in lipid deposition. After the initial deposition of lipids, macrophages infiltrate the deeper layer of the intima and it is postulated that macrophages take up the deposited lipid-proteoglycan complexes to form foam cells. It is further suggested that these changes create the Type II lesion, together with non-foamy macrophages, extracellular lipids, ECM, and SMCs.
These pathological findings are important when considering therapeutic strategies for the treatment of human atherosclerosis. It may be that preventing the entry and subsequent intimal retention of apoB-containing lipoprotein, particularly at an early age is a key to preventing atherosclerosis.7 For example, drugs that lower the plasma concentration of atherogenic lipoproteins, such as statins, are expected to reduce the entry of atherogenic lipoproteins in the intima. This concept is supported by lipid-lowering interventional studies that could reduce the incidence of coronary heart disease.96,97 In this regard, initiation of LDL-lowering therapy for young people with multiple risk factors is expected to show lifelong benefits.7 Lipoprotein properties (e.g. size, electrical charge, and cholesterol enrichment) and endothelial permeability could also be targets for prevention of lipid entry in the intima. Inhibiting the interaction of apoB with intimal proteoglycans is a potential future approach, involving direct blocking of the interaction between apoB and proteoglycans, and/or manipulating the synthesis of key retentive proteoglycans or their GAG side chains. For example, modifying proteoglycan synthesis and structure, such as GAG elongation and sulfation, may represent a target to prevent LDL binding and retention in the intima.98 Certainly, there is enough evidence to warrant targeting proteoglycans of the vascular wall in the treatment of cardiovascular disease.
This review was prepared with grant support from the National Institutes of Health (HL 18645) (to T.N.W.).
The authors would like to thank Dr Virginia Green for editorial assistance.
Conflict of interest: none declared.