Lipoprotein receptor signalling in atherosclerosis

Abstract

The founding member of the lipoprotein receptor family, low-density lipoprotein receptor (LDLR) plays a major role in the atherogenesis through the receptor-mediated endocytosis of LDL particles and regulation of cholesterol homeostasis. Since the discovery of the LDLR, many other structurally and functionally related receptors have been identified, which include low-density lipoprotein receptor-related protein (LRP)1, LRP5, LRP6, very low-density lipoprotein receptor, and apolipoprotein E receptor 2. The scavenger receptor family members, on the other hand, constitute a family of pattern recognition proteins that are structurally diverse and recognize a wide array of ligands, including oxidized LDL. Among these are cluster of differentiation 36, scavenger receptor class B type I and lectin-like oxidized low-density lipoprotein receptor-1. In addition to the initially assigned role as a mediator of the uptake of macromolecules into the cell, a large number of studies in cultured cells and in in vivo animal models have revealed that these lipoprotein receptors participate in signal transduction to modulate cellular functions. This review highlights the signalling pathways by which these receptors influence the process of atherosclerosis development, focusing on their roles in the vascular cells, such as macrophages, endothelial cells, smooth muscle cells, and platelets. Human genetics of the receptors is also discussed to further provide the relevance to cardiovascular disease risks in humans. Further knowledge of the vascular biology of the lipoprotein receptors and their ligands will potentially enhance our ability to harness the mechanism to develop novel prophylactic and therapeutic strategies against cardiovascular diseases.

1. Introduction

Atherosclerosis is the underlying process of cardiovascular and cerebrovascular diseases, which are leading causes of mortality and morbidity worldwide.¹ It is a chronic vascular disorder of large and medium-sized arteries that progresses from fatty streaks in the early stage to advanced lesions with complicated and unstable plaques prone to rupture resulting in athero-thrombotic events.^2–4 Researches encompassing several decades have established that circulating lipoproteins and their receptors play a major role in atherogenesis. The Nobel Prize winning work by Goldstein and Brown revealed for the first time that the low-density lipoprotein (LDL) receptor regulates cholesterol homeostasis by the receptor-mediated endocytosis of LDL particles, and that the mutations in the LDL receptor gene cause elevated serum LDL cholesterol and coronary atherosclerosis in familial hypercholesterolaemia patients.⁵ Subsequent success of cholesterol-lowering statins and a discovery of the scavenger receptors that bind to modified LDL and contribute to foam cell formation in the atheroma further expanded the critical importance of lipoproteins and their receptors in atherogenesis.^5,6 Although elevations of circulating cholesterol carrying lipoproteins, stemming from excessive dietary lipid intake and inadequate removal of cholesterol in the liver, play a key causal role in atherosclerosis, it has become increasingly apparent that dysfunction of vascular cells also contributes to atherogenesis. These cells include monocytes/macrophages, platelets, blood and lymphatic endothelium, and vascular smooth muscle cells (SMCs).² During the decades-long process, diverse events occur in the vascular wall, including endothelial dysfunction, leucocyte/monocyte adhesion and infiltration, lipid-laden foam cell formation, abnormal vascular SMC growth, and platelet activation.^2,7 Manifestation of endothelial dysfunction is a relatively early occurrence that precedes clinical complications of atherosclerosis.^8,9 Following activation of the innate immunity response by proinflammatory cytokines and oxidized lipids, endothelial cells upregulate adhesion molecules, provoking leucocyte-endothelial cell adhesion, and recruitment of inflammatory cells into the lesions.^10,11 Upon infiltration into the arterial wall, these inflammatory monocytes/macrophages in turn generate oxidative stress and accumulate lipids, contributing to perpetuation of proinflammatory processes and foam cell formation.¹² In addition, there is an increasing recognition for a complex role of vascular SMCs. Aberrant SMC proliferation, infiltration, and transition to non-contractile phenotypes likely contribute to plaque progression.^2,13 In the later stage, SMCs produce the extracellular matrix which provides a stable fibrous cap, preventing a rupture of plaques.¹³ The plaque rupture further leads to exposure of prothrombotic factors to activate platelets and thrombus formation, causing major cardiovascular episodes such as myocardial infarction (MI) and stroke.

Since the early discovery of the LDL receptor, many other related receptors that bind to native or modified lipoproteins have been identified. In addition to the initially assigned role as cholesterol transporters, these LDL receptor-related proteins have a wide range of additional functions that influence diverse physiological and pathological processes, including embryonic development, inflammation, haemodynamics, thrombosis, neointima hyperplasia, and atherosclerosis.^14–18 This review will focus on the LDL-receptor family members, low-density lipoprotein receptor-related protein (LRP)1, LRP5, LRP6, apolipoprotein E receptor 2 (ApoER2 or LRP8), very-low density lipoprotein receptor (VLDLR), and the scavenger receptor family members, cluster of differentiation 36 (CD36), scavenger receptor class B type I (SR-BI), and lectin-like oxidized low-density lipoprotein receptor-1 (LOX-1). It will provide a concise view on how these lipoprotein receptors influence atherosclerosis progression through a network of signal transduction pathways in the vascular cells. There are multiple additional receptors and binding proteins for lipoproteins and lipids that also play key roles in atherosclerosis and cardiovascular diseases. The scavenger receptor-A (SR-A), also known as the macrophage scavenger receptor or CD204, has major functions in lipid metabolism and atherogenesis through its interaction with a diverse array of ligands, including oxidized LDL and advanced glycation end-product modified proteins.^5,6,19–21 Various proteoglycans also bind to LDL, and they are abundantly expressed in human atherosclerotic lesions.^22,23 Both chondroitin sulfate proteoglycans and heparin sulfate proteoglycans, including versican and perlecan, have been shown to contribute to development of atherosclerosis in hyperlipidaemic mouse models.^24–26 A group of receptors that binds to lipid molecules such as sphingosines and ceramides also mediates cellular signalling to influence the process of atherogenesis and other vascular disorders.^27–30 A number of excellent reviews provide the updates on functions of these molecules related to atherosclerosis.

2. Low-density lipoprotein receptor family members

The LDL receptor family consists of multiple structurally related transmembrane proteins that participate in a wide range of biological processes, including lipid metabolism, brain development, nutrients delivery, thrombosis, and atherogenesis. These receptors share common structural features, with a large extracellular domain with ligand-binding motifs, a single transmembrane domain, and a cytoplasmic tail with multiple adaptor-binding sites¹⁶ (Figure 1, Table 1). Whereas the role of the LDL receptors is likely limited to lipid metabolism, growing evidence indicates that the other members of the family have additional functions as signal transducing receptors.

Figure 1.

Structure and function of LDL receptor family members, LRP1, LRP5/6, ApoER2, and VLDLR in atherosclerosis. LRP1, LRP5/6, ApoER2, and VLDLR belong to the LDL receptor family that shares common structural features, with a large extracellular domain with ligand-binding motifs, EGF-like repeats, and YWTD β-propeller domains, a single transmembrane domain, and a cytoplasmic tail with multiple adaptor-binding sites. LRP1 in vascular smooth muscle cells and macrophages shows a potent protective role against atherosclerosis. LRP5 and LRP6, the coreceptors for the Wnts signalling receptor Frizzled, also exert anti-atherogenic actions in smooth muscle cells and macrophages. ApoER2 and VLDLR, which does not play a major role in systemic lipid homeostasis, have both anti-and pro-atherosclerotic actions. ApoE binding to ApoER2 in vascular cells leads to decreased inflammation and platelet activation, whereas reelin binding to the receptor causes opposite effects.

Table 1.

Anti- and proatherogenic functions of the LDL receptor family members

		Antiatherogenic functions			Proatherogenic functions
	Lipoprotein receptor	Cell type	Cellular function	Mechanism	Cell type	Cellular function	Mechanism	References
LDLR family members	LRP1	SMC	Proliferation↓	SHP-2-PDGF signalling↓	SMC			31–45
				TGFβ signalling ↓
			Phenotype Switching↓	Protenases ↓
				Connective tissue growth factor ↓
		Macrophages	Inflammation ↓	Inflammatory genes ↓	Macrophages			46–49
	LRP6	SMC	Proliferation ↓	PDGF-ERK1/2-JAK1/STAT1 ↓	SMC			50–52
				PDGFR degradation ↑
			Osteochondrogenic program ↓	Non canonical Wnt signalling↓
	LRP5	Macrophages	Cell migration ↑	Wnt signalling ↑	Macrophages			53 , 54
	ApoER2	Endothelial cells	Inflammation ↓	eNOS activity (apoE) ↑	Endothelial cells	Inflammation ↑	eNOS activity (Reelin) ↓	55 , 56
			Cell migration↑
		Macrophages	Apoptosis ↓	Akt ↑, p53 ↓	Macrophages			57–59
			Cholesterol efflux↑	PI3 kinase/PKC/Sp1-ABCA1 ↑
			Inflammation ↓	Src/p38 MAPK ↑, NF-kB ↓
		Platelets	Activation ↓	Agonist actions ↓	Platelets	Activation ↑	Inhibitory ApoE3 actions ↓	60–64
	VLDLR	SMC	Proliferation/migration ↓?		SMC			65–69
		Macrophages	Cholesterol efflux↑	ABCA1 ↑	Macrophages	Lipid uptake ↑?		58 , 65 , 69 , 70
			Inflammation ↓	Src/p38 MAPK ↑, NF-kB ↓

2.1 Lipoprotein receptor-related protein 1

The low-density lipoprotein receptor-related protein 1 (LRP1) consists of a 515 kDa heavy chain containing four clusters of ligand binding domains and a non-covalently associated 85 kDa light chain containing EGF-like repeats, YWTD β-propeller domains, a transmembrane, and intracellular cytoplasmic domain (ICD)^16,71 (Figure 1). LRP1 was initially cloned as a protein with multiple LDL receptor type A repeats, which mediates chylomicron remnant uptake.^72,73 It was also identified as a receptor in the liver responsible for catabolism of α2-macroglobulin-proteinase complexes.^74,75 To date, LRP1 is known to bind to more than 40 distinct ligands that are structurally unrelated.

LRP1 is expressed in multiple cell types including hepatocytes, vascular SMCs, neurons, macrophages, fibroblasts, and endothelial cells.⁷¹ LRP1 in the liver cooperates with the LDL receptor in the endocytosis and clearance of cholesterol-rich chylomicron remnants from the circulation. However, the large number and functionally diverse ligands for LRP1 and the embryonic lethality in global LRP1^−/− mice suggest that LRP1 is essential in various physiological and pathological processes in addition to lipid metabolism.^{31–34,46,47,76–83} First, LRP1 acts as an endocytic receptor for extracellular ligands, which escorts the ligands to lysosomes for degradation. Several extracellular molecules involved in signalling pathways have been identified to bind to LRP1, and LRP1 modulates these pathways by facilitating the removal of the ligand from cell surface. Second, in some cell types, upon binding to a ligand LRP1 initiates a signal transduction through the recruitment of adaptor proteins to its ICD of the receptor. For example, tissue-type plasminogen activator stimulates the N-methyl-D-aspartate receptor cascade via an LRP1-mediated interaction with the adaptor protein postsynaptic density protein 95.⁸⁴ Third, LRP1 is known to undergo a regulated intramembrane proteolysis in which shedding of the LRP1 generates a substrate for γ-secretase-mediated cleavage, releasing the ICD of LRP1. The ICD domain then translocates to the nucleus to modulate gene expression.^85,86 Furthermore, genome-wide association studies (GWAS) in humans have revealed that LRP1 gene represents a susceptibility locus for elevated plasma lipid, coronary heart disease, and abdominal aortic aneurysms.^87–89

2.1.1 SMC LRP1

Under pathological conditions, SMCs are induced to proliferate and migrate, contributing to the development of atherosclerosis and restenosis.^2,13 Platelet-derived growth factor (PDGF) is a potent mitogen for SMCs and numerous studies have revealed a critical role for this pathway in atherosclerosis and neointima formation.^90–93 PDGF binding to PDGF receptor β (PDGFRβ), a tyrosine kinase receptor, initiates a signalling pathway that promotes SMC growth and migration, and LRP1 has been shown to potently suppress the process.^{31–33,35–37} This was initially demonstrated by Boucher et al. using mice with SMC-specific deletion of LRP1 (smLRP1^−/−) on hypercholesterolaemic LDLR^−/− background.³¹ The study found that smLRP1^−/−;LDLR^−/− mice were more susceptible to atherosclerosis compare to LDLR^−/− mice, and that these mice showed upregulation of PDGFRβ expression and the downstream signalling, leading to excessive proliferation of vascular SMCs and markedly enlarged aortas. These effects were inhibited by a treatment of the mice with Gleevec (Imatinib Mesylate), an inhibitor of tyrosine kinases including the PDGFRβ. These studies revealed that LRP1 in SMCs plays a major role in protecting the integrity of the vascular wall and reducing atherosclerosis by suppressing PDGFRβ activation. Mechanistically, LRP1 modulates PDGF signalling through recruitment of various adaptor molecules to its ICD. Upon activation of PDGFRβ, LRP1 is tyrosine phosphorylated at NPxY motif within the ICD, which creates a docking site for phosphotyrosine-binding domain and Src homology 2 domain containing adaptor proteins.^36,38,39 One such protein is SHP-2, a tyrosine phosphatase that is required for activation of the extracellular signal-regulated kinase (ERK) pathway by PDGFR.⁹⁴ SHP-2 binds with high affinity to the phosphorylated form of the LRP1-ICD and its association with LRP1 seems to attenuate PDGF-mediated signalling events.³⁹ Another study found that LRP1-ICD can associate with c-Cbl, a ubiquitin E3-ligase that regulates turnover of receptor tyrosine kinases such as PDGFRβ, and that LRP1 deficiency in mouse fibroblasts results in an accelerated turnover of PDGFRβ, suggesting a role of LRP1 as an anchor protein that controls the traffic of PDGFRβ from the cell surface to intracellular compartments.⁴⁰ However, whether or how these molecular mechanisms contribute to the anti-atheroprotective action of LRP1 in vivo is yet to be elucidated.

In addition to atherogenesis, studies in mice with genetic modifications revealed that LRP1 expressed in SMCs plays a protective role against aneurysm formation.^31,33–35 In healthy arteries, SMCs maintain a contractile phenotype, which allows them to contract in response to changes in vascular pulse pressure. Disruptions of SMC contractile function, which switch the cells to a synthetic phenotype, are known to predispose the aorta to aneurysms.^95,96 SmLRP1^−/− mice have exaggerated aortic dilatation that is attributed to extensive disruption of elastic lamina with numerous breaks.^31,34 Histological analyses of the aortic vessel wall demonstrated that SMCs from smLRP1^−/− mice have appearance of a synthetic phenotype, containing more synthetic organelles and fewer focal adhesions, compared to those from control mice.³⁴ This effect is unlikely to arise from excessive PDGF signalling, as the PDGF signalling pathway is not upregulated in smLRP1^−/− mice in the absence of LDLR deficiency.³⁴ Multiple potential mechanisms by which LRP1 regulates SMC phenotype-switching are proposed. First, the extensive disruption of the elastic lamina in smLRP1^−/− mice seems to be resulted from excessive expression of several proteases. One of these proteases, high-temperature requirement factor A1 (HtrA1), is an LRP1 ligand and it was detected in excess in the aortic wall of smLRP1^−/− mice.³⁴ In addition to HtrA1, increased expression of metalloproteinases, including MMP-9, MMP-2 and membrane associated type 1-matrix metalloproteinase, was also found in the vessel wall of smLRP1^−/− mice.^41,42 Degradation of extracellular matrix proteins and elastic fibres by these proteases may contribute to the development of aneurysms in the mouse model. Secondly, an increased matrix deposition in the vessel walls observed in smLRP1^−/− mice may be due to excessive connective tissue growth factor accumulation.^31,32,34 Connective tissue growth factor is a secreted matricellular protein that is a key mediator of fibrosis and plays an important role in vascular development.⁴³ Together, these studies highlight the important role for LRP1 in vascular homeostasis, in which it protects the integrity and function of elastic lamina by regulating protease activity, as well as the pathways involved in vascular extracellular matrix deposition. A small study conducted on human tissues revealed that LRP1 expression was significantly attenuated in abdominal aortic tissues with aneurysms, compared with normal aortic tissue.^44,71

In addition to atherogenesis and aneurysm, LRP1 in SMCs is also implicated in development of pulmonary arterial hypertension (PAH) through modulation of TGFβ signalling that is known to regulate vascular remodelling.^97–99 PAH is characterized by remodelling of pulmonary arterioles, leading to increased vascular resistance, right ventricular hypertrophy, and heart failure. SMC proliferation, deposition of extracellular matrix, and vascular remodelling are hallmarks of PAH. A recent study by Calvier et al. has demonstrated that smLRP1^−/− mice display exaggerated TGFβ1 signalling in SMCs and that these mice develop spontaneous PAH.⁴⁵ Earlier studies in vivo and in cultured cells showed that LRP1 directly interacts with TGFβ, and that the lipoprotein receptor is required for growth inhibition by TGFβ.^33,81,100 Furthermore, in pulmonary artery SMCs derived from PAH patients LRP1 protein expression was downregulated, which was associated with enhanced TGFβ1 signalling.⁴⁵

2.1.2 Macrophage LRP1

Lipoprotein receptors in macrophages are generally implicated in promotion of foam cell formation and vascular inflammation.^2,10 However, macrophage LRP1 has been shown to exert atheroprotective effects. Hu et al.⁴⁶ reported that mice with selective deletion of LRP1 in macrophages (macLRP1^−/−) on apoE^−/− and LDLR^−/− background display an increase in total atherosclerotic lesion area with a higher prevalence of advanced lesions. Another study also found that LDLR^−/− mice receiving bone marrow transplantation from macLRP1^−/− mice fed on high-fat diet have an increased lipid deposition and macrophage infiltration in the aorta compared to those receiving control bone marrow.^47,48 Macrophages from macLRP1^−/− mice showed an elevated production of proinflammatory markers, such as macrophage chemoattractant protein-1 (MCP1) and tumour necrosis factor α (TNFα).^47,48 Increased production of TNFα by lipopolysaccharides (LPS) treatment was also observed in macrophages isolated from mice carrying a knock-in mutation in the LRP1-ICD, revealing that interaction with adaptor molecules via LRP1-ICD may be important for this activity.⁴⁹ A potential mechanism in macrophages by which LRP1 influences the inflammatory response may be through transcriptional regulation of inflammatory genes. In cultured macrophages, it has been shown that inflammatory mediators, such as LPS and interferon-γ, can induce the LRP1 shedding and release of LRP1-ICD from the plasma membrane, which translocates to the nucleus where it suppresses activation of the genes responsive to inflammatory stimuli.^85,86

These studies in cultured cells and in genetically modified mice collectively point to a strong protective role of LRP1 in SMCs and macrophages against development of multiple vascular disorders.

2.2 LRP5 and LRP6

LRP5 and LRP6 are closely related type I membrane proteins, and their extracellular domains are comprised of LDLR type A repeats, EGF-like domain, and YWTD-type β-propeller domain^101,102 (Figure 1). These receptors act as coreceptors for the frizzled, the seven-transmembrane receptors for Wnts.^103,104 The Wnt signalling pathway is a ubiquitous signalling cascade that regulates a wide range of physiologic processes, and aberrant Wnt signalling has been implicated in cancer, developmental disorders, pulmonary diseases, renal disorders, and cardiovascular diseases in humans.^105–107 In the canonical Wnt pathway, Wnt ligands bind to the frizzled receptor and LRP5/6 coreceptors, resulting in formation of multiprotein complexes in the plasma membrane which contain axin and GSK3β.^108–111 The complex formation promotes stabilization of β-catenin and its translocation into the nucleus, where it regulates transcription of various genes.^{53,109,110,112,113}

2.2.1 Lipoprotein receptor-related protein 6

LRP6 gene maps to human chromosome 12p11-p13, which translates to the transmembrane protein containing 1613 amino acid residues.¹⁰¹ Importance of LRP6 in atherogenesis came to light by a discovery of a missense mutation in LRP6 gene in humans.¹¹⁴ The index patient presented with MI at 48 years of age with a history of hypertension, hyperlipidaemia, and diabetes without obesity. The mutation resulted in substitution of cysteine for arginine at the amino acid 611 in the second EGF-like domain (LRP6-R611C), and it cosegregated with the metabolic phenotypes. The crystal structure of LRP6 revealed that this substitution disrupts a salt bridge between arginine at 611 and glutamate at 477 in the E2 propeller domain, altering the relative orientation between YWTD and the EGF domains.¹¹⁵ Through screening of individuals with early onset familial coronary artery disease and metabolic syndrome, three additional mutations in LRP6 gene were identified, which cosegregated with the metabolic traits in the kindred of the affected subjects.¹¹⁶ All three mutations reside in the second propeller domain responsible for ligand binding. Patients with the LRP6-R611C mutation had extremely high levels of plasma cholesterol,¹¹⁴ and common variants in LRP6 gene have also been linked to a mild increase of plasma LDL cholesterol in the general population.¹¹⁷ Thus the function of LRP6 in LDL clearance was investigated in cells expressing wild-type LRP6 or LRP6-R611C mutant.¹¹⁸ Overexpression of the mutant receptor resulted in lower LDL uptake, compared to the wild-type receptor. Similarly LDL uptake in lymphocytes isolated from heterozygote R611C mutation carriers was significantly lower compared to those from unaffected family members.¹¹⁸ In cells expressing wild-type LRP6, the receptor was mainly localized at the cell surface and early endocytic compartments. In contrast, LRP6-R611C mutant cells had reduced cell surface expression with marked accumulation of the receptor in late endocytic compartments.¹¹⁸ Further examinations also revealed a lower rate of LDL/apoB dissociation from LRP6-R611C at an acidic pH, compared to wild-type receptor. These studies suggest that LRP6 may serve as a receptor for LDL endocytosis for LDL clearance in the liver.

In addition to its role in the lipid metabolism, an athero-protective function of LRP6 has been demonstrated in SMCs through suppression of PDGF signalling. In human atherosclerotic coronary arteries, colocalization of LRP6 expression with PDGFRβ was detected in the subintimal layers.⁵⁰ In cultured SMCs, the antiproliferative effect of LRP6 was associated with decrease in PDGF-induced activation of ERK1/2 and JAK1/STAT1 pathway, and these inhibitory effects on PDGF signalling were not observed in SMCs expressing LRP6-R611C mutant.⁵⁰In vivo studies in the knock-in mice harbouring LRP6-R611C mutation showed that the mutant mice have medial hyperplasia, and that the effect was further augmented by an arterial injury, resulting in partial to total vascular obstruction.⁵¹ Furthermore, LRP6-R611C mice on LDLR^−/− background fed high-cholesterol diet displayed markedly increased arterial neointima formation and an accelerated atherosclerosis. Enhanced expressions of PDGFR and PDGF ligands were detected in the aortic media and in SMCs in the knock-in mice, compared with wild-type controls. Mechanistically, excessive PDGF signalling in LRP-R611C mice has been linked to activation of non-canonical Wnt signalling and upregulation of Sp1, a transcription factor known to target PDGF and PDGFRβ gene expression.⁵¹ Another mechanism has been also proposed in cultured SMCs that LRP6 suppresses PDGF-induced SMC proliferation through promotion of lysosome degradation of PDGFRβ.⁵⁰ These findings suggest that LRP6 serves as a negative regulator of PDGF-dependent SMC proliferation and that a loss of the function may contribute to development of atherosclerosis in humans with the LRP6 mutation. A specific role of SMC LRP6 in atherogenesis was further investigated in a study using mice lacking LRP6 selectively in SMCs.⁵² Compared to control LRP6^fl/fl mice on LDLR^−/− background, mice lacking LRP6 selectively in vascular SMCs (LRP6^fl/fl; SM22-Cre; LDLR^−/−, designated LRP6-VKO) on atherogenic diet exhibited increased osteochondrogenic calcification and arterial stiffening without affecting plasma lipid profiles. Vascular calcification paralleled upregulation of aortic osteochondrogenic programs and circulating osteopontin, a matricellular regulator of arteriosclerosis. In LRP6-VKO mice multiple canonical and non-canonical Wnt ligands along with Frizzled 10, a receptor for non-canonical Wnt signalling, were upregulated in aortas. Mechanistically, the study proposed that LRP6 suppresses non-canonical Wnt signalling that drives vascular SMC osteochondrogenic programs through a pathway involving upstream stimulatory factor 1 and protein arginine methylation. In summary, the strong proatherogenic phenotypes in humans with the LRP6 gene variation and in the genetic loss-of-function mouse models indicate that LRP6 provides a key protection against dyslipidaemia and atherosclerosis.

2.2.2 Lipoprotein receptor-related protein 5

LRP5 gene is located on chromosome 11q12-13, and it encodes 1615 amino acid protein, which is 71% homologous to LRP6.¹¹⁹ Polymorphisms of LRP5 gene have been associated with a variety of human diseases, including osteoporosis, obesity, and diabetes.^119–123 LRP5 is widely expressed in multiple tissues, including the liver, where the receptor is implicated in LDL and chylomicron clearance.^54,124 In human atherosclerotic lesions, LRP5 has been detected in endothelial cells, vascular SMC, monocytes, and macrophages.⁵³ Studies in cultured human macrophages showed that treatment with aggregated LDL, a mimic of modified LDL, increases LRP5 expression, resulting in upregulation of Wnt pathway proteins, including β-catenin and its downstream targets, and that silencing LRP5 by siRNA attenuates macrophage migration.⁵³ Involvement of LRP5 in atherosclerosis has been further demonstrated in mice deficient in LRP5 globally (LRP5^−/−).⁵⁴ These mice fed high-cholesterol diet displayed downregulation of β-catenin in the aorta compared to wild-type mice, indicating that LRP5 is required for Wnt signalling activation in the vascular wall. Compare to control mice on high-cholesterol diet, LRP5^−/− mice displayed increased plasma cholesterol levels and enhanced formation of atherosclerotic lesions, with elevated levels of circulating inflammatory cytokines and macrophage infiltration in the aorta. In sum, LRP5 deficiency and resulting downregulation of Wnt pathway contributes to exaggeration atherosclerosis, likely due to hypercholesterolaemia, systemic, or local inflammation and macrophage migration in the arterial wall.

2.3 ApoER2 and VLDLR

ApoER2, also known as LRP8, and its closely related receptor VLDLR belong to LDL receptor family, and they share common structural features with other members, with a large extracellular domain containing ligand-binding motifs, a single transmembrane domain, and a cytoplasmic tail with multiple adaptor-binding sites¹⁶ (Figure 1). However, unlike other members of the family, which play a major role in maintaining lipid homeostasis by mediating internalization of lipoproteins, ApoER2 or VLDLR does not contribute to the lipid metabolism. Global deletion of ApoER2 or VLDLR gene in mice did not result in alterations of plasma lipid profiles.^55,125 Instead, earlier works established the receptors as a signalling receptor for reelin, and they showed that the ligand–receptor complex is critically involved in neuronal migration during brain development.^126–128 In the neuronal cells, reelin binding to ApoER2 or VLDLR initiates recruitment of its adaptor Disabled-1 (Dab-1) to the cytoplasmic NPxY motif, which in turn activates a series of kinases to regulate cellular function required for normal brain development.^126–128 Subsequently it has become apparent that ApoER2 is abundantly expressed in cells in the circulation and in the vascular walls, including platelets, monocytes/macrophages, endothelial cells, and SMCs.^{55,57,60,129–132} Similar to the other members of the LDLR, ApoER2 binds to discreet sets of ligands, including reelin, apolipoprotein E (apo E), activated protein C, and β2 glycoprotein I.^{55,56,126,131,132} Studies in both cultured cells and in mouse models have revealed that ApoER2 in the vascular cells has critical functions in regulation of coagulation, neointima hyperplasia, and atherosclerosis. In humans, GWAS have linked polymorphisms of ApoER2 gene to premature coronary artery disease and MI.^133–136 In particular, homozygous carriers of the ApoER2-R952Q variant have a two-fold increased risk of these conditions.^133,135,136

2.3.1 Endothelial cell ApoER2

Reelin is an extracellular matrix glycoprotein that was originally found in the developing brain where it is secreted by Cajal–Retzius neurons, and reelin binding to ApoER2 and VLDLR in the neuron is required for normal brain development and neuronal survival in the adult brain.¹³⁷ Further studies found that reelin is also abundantly present in the circulation and in the hepatic stellate cells.¹³⁸ Extraneuronal actions of reelin have been initially reported, which include promotion of platelet spreading, thrombin generation, and clot formation.¹³⁹ A proatherogenic function of reelin-ApoER2 tandem has been recently revealed.⁵⁶ Using hyperlipidaemic LDLR^−/− mice with inducible deletion of reelin either ubiquitously or selectively in the liver, the study demonstrated that these mice lacking circulating reelin have attenuated atherosclerosis progression. In the atherosclerotic lesion, reelin deficient mice showed reduced macrophage infiltration and endothelial inflammation, with lower expression of adhesion molecules vascular cell adhesion molecule-1 (VCAM-1) and intercellular adhesion molecule-1 (ICAM-1). Analysis by intravital microscopy further confirmed decreased leucocyte-endothelial interaction in the reelin-deficient mice in vivo. Mechanistically, the study in cultured human endothelial cells showed that reelin treatment increases expression of ICAM-1, VCAM-1, and E-selectin by suppressing endothelial nitric oxide synthase (eNOS) activity and increasing NF-kB activity in an ApoER2-dependent manner.

ApoER2 in endothelial cells also binds to apolipoprotein E (apoE), and in contrast to reelin, interaction with apoE3, a common apoE allele in humans, with the receptor has been shown to result in anti-atherogenic actions.⁵⁵ In humans, cardiovascular disease risk is modified by genetic variants of apoE. Compared with the most common allele apoE3, individuals with apoE4 allele have an increased risk of atherosclerosis and coronary heart disease.^140,141 An ApoER2 variant ApoER2-R952Q, which is associated with early onset of MI, also has an additive effect with apoE4, with the combined genotype R952Q/E4 showing a 3.9-fold greater susceptibility to cardiovascular disease.¹³³ Ulrich et al.⁵⁵ demonstrated that in endothelial cells apoE3 binding to ApoER2 stimulates eNOS and endothelial cell migration, and that it also attenuates monocyte-endothelial cell interaction. Interestingly, apoE4 did not alter eNOS activation, endothelial migration or monocyte-endothelial interaction; instead it antagonized apoE3-ApoER2 actions. Reconstitution experiments in culture further revealed that ApoER2-R952Q is incapable of mediating apoE3 actions, and in vivo studies further indicated that endothelial repair followed by arterial injury is attenuated by deletion of ApoER2 or overexpression of apoE4. Furthermore, in a model of neointima formation invoked by carotid artery endothelial denudation, ApoER2^−/− mice displayed exaggerated neointima hyperplasia.⁵⁵ Thus, the apoE3-ApoER2 tandem promotes endothelial NO production, endothelial repair, and endothelial anti-inflammatory properties, and it prevents neointima formation. These findings collectively indicate that ApoER2 in endothelium exerts ligand-dependent signalling that influences the processes related to atherogenesis.

2.3.2 Macrophage ApoER2

A role of macrophage ApoER2 in atherosclerosis development has been implicated in a study using mice with global ApoER2 deficiency.⁵⁷ ApoER2^−/− mice on hypercholesterolaemic LDLR^−/− background showed accelerated atherosclerosis with more complex lesions and extensive necrosis cores compared to single LDLR^−/− mice. The atherosclerotic plaques of the double knockout mice contained higher numbers of macrophages expressing the pro-apoptotic phospho-p53. In vitro studies using cultured mouse primary macrophages or cell lines showed that oxidized LDL (oxLDL) treatment induces more lipid accumulation and cell death in ApoER2-deficient cells, compared with control cells.⁵⁷ Furthermore, ApoER2-null macrophages displayed attenuated Akt activation, increased levels of the phospho-p53 and increased activation of peroxisome proliferator-activated receptor γ (PPARγ). Additional anti-atherogenic mechanisms of macrophage ApoER2 have been found in studies using cultured macrophages.⁵⁸ One study showed that treatment with reelin or human apoE3 upregulates ABCA1 mRNA and protein levels, which was associated with increased apoAI-mediated cholesterol efflux. It further demonstrated that the upregulation of ABCA1 requires ApoER2 and Dab-1, and that pharmacological or genetic inhibitions of PI3 kinase, Protein Kinase Cζ and Sp1 DNA binding abrogates the effect on ABCA1.⁵⁸ Another study using cultured mouse macrophages overexpressing human ApoER2 demonstrated that apoE treatment promotes conversion into anti-inflammatory M2 phenotypes with reduced NFkB activation, cell motility and proinflammatory cytokine production.⁵⁹ These effects of apoE on macrophage polarization were inhibited by inhibitors of Src family kinases or p38 MAP kinase. These studies suggest an anti-atherogenic function of ApoER2 in macrophages, likely through promotion of cholesterol efflux, prevention of apoptosis, and/or conversion to anti-inflammatory phenotypes.

2.3.3 Platelet ApoER2

Three splice variants of ApoER2 are expressed in platelets,^61,62 and in isolated human platelets, lipidated apoE has been shown to attenuate agonist-induced platelet activation by stimulating intracellular NOS and cGMP likely through ApoER2.^60–63 Studies by Robertson et al.⁶⁴ also showed that platelets isolated from ApoER2^+/− or ApoER2^−/− mice had reduced activation in response to ADP or thrombin, compared to those from wild-type mice. Ex vivo treatment of the isolated wild-type mouse platelets with lipidated apoE inhibited the platelet activation and this effect was markedly attenuated in ApoER2^−/− mouse platelets. In vivo, both ApoER2^+/− and ApoER2^−/− mice demonstrated increased time for vessel occlusion in response to vascular injury; however, significantly smaller increase in occlusion time was observed in homozygous ApoER2^−/− mice compared to ApoER2^+/−. These in vivo and ex vivo studies suggest that platelet ApoER2 may play two conflicting functions, one as a receptor that augments platelet activation by agonists and the other as a mediator for inhibitory actions of apoE3 on platelets.

2.3.4 Very-low density lipoprotein receptor

As mentioned above, global deletion of the VLDLR gene in mice did not alter the plasma lipoprotein profiles, indicating VLDLR has a minimum contribution to regulation of systemic lipid metabolism.¹²⁵ However, when crossed with LDLR^−/− mice, lack of the VLDLR resulted in an increase in serum triglyceride levels on a high-fat diet or after prolonged fasting, suggesting that the VLDLR may be involved in peripheral triglyceride uptake.^142,143 Furthermore, VLDLR^−/− mice fed a high-fat diet or placed on ob/ob background showed decreased weight gain, possibly due to a reduction in whole-body free fatty acid uptake.¹⁴⁴ These studies suggest that VLDLR may contribute to delivery of triglyceride or free fatty acid to peripheral cells.

A role of VLDLR in atherosclerotic process is yet to be clarified. Abundant VLDLR expression was observed in endothelial cells on capillaries and small arteries as well as in coronary arteries, but it was minimally detected in aortic endothelium.¹⁴⁵ The receptor was also expressed in neointimal SMCs and macrophage-derived foam cells in the lesion.^65–68 When placed on LDLR^−/− background, VLDLR deficiency alone or in combination with selective overexpression of VLDLR in endothelium had minimum impact on atherosclerotic development compared to LDLR^−/− mice.⁶⁹ In contrast, VLDLR^−/− mice displayed increased intimal thickening induced by vascular injury, indicating that VLDLR in SMC likely has a protective role against neointima hyperplasia.⁶⁹ In macrophages, both pro- and anti-atherogenic roles of VLDLR have been reported. Transplantation of VLDLR^+/+ macrophages into VLDLR^−/− mice accelerated the development of atherosclerotic lesions, suggesting a pro-atherogenic role of macrophage VLDLR.⁷⁰ In contrast, studies in cultured macrophages have demonstrated that activation of VLDLR by reelin or ApoE3 increases ABCA1 transporter expression to promote macrophage cholesterol efflux.⁵⁸ In addition, the suppression of anti-inflammatory M2 phenotype was observed in bone marrow macrophages derived from VLDLR^−/− mice but not wild-type or LDLR^−/− mice, and treatment with ApoE promoted conversion of cultured macrophages from pro-inflammatory M1 to M2 phenotype through a process dependent on p38MAPK and Src family kinase.⁵⁹ Collectively, these observations in the experimental models suggest that the proatherogenic effect of VLDLR arising from the lipid uptake function may be counter-balanced by their antiatherogenic effects, such as elevation of ABCA1 transporter and conversion into an anti-inflammatory phenotype.

In humans, a few genetic variants of VLDLR have been reported that are associated with lipid traits and cardiovascular disease risk.^146–148 One variant (rs1454626) located in the 5' flanking region of VLDLR was associated with coronary artery disease, BMI, and levels of LDL-associated apolipoprotein B.¹⁴⁷ The study also identified receptor–ligand genetic interactions between the common VLDLR SNP and an apoE genotype for predicting coronary artery disease case status. More recently, another association signal was located in the intron of VLDLR; minor A allele of rs3780181 was associated with increased plasma total cholesterol and LDL-associated cholesterol.^146,148 Using a functional genetic approach, the study demonstrated that deletion of the intronic enhancer domain containing the risk allele affects VLDLR expression, potentially influencing the circulating levels of total cholesterol.¹⁴⁸

3. Scavenger receptor family members

The scavenger receptors belong to a large family of pattern recognition receptors, which show capacity to interact with a wide range of ligands, and they are expressed in many cell types related to lipid metabolism and atherogenesis, including hepatocytes, macrophages, endothelial cells, SMCs, and platelets.^149,150 Interacting with the circulating native or modified lipoproteins, such as oxLDL, these receptors modulate vascular inflammation, lipid accumulation, and plaque formation (Figure 2, Table 2).

Figure 2.

Structure and function of scavenger receptor family members, CD36, SR-BI, and LOX-1 in atherosclerosis. The scavenger receptors belong to a large family of pattern recognition receptors, which can interact with a wide range of ligands. Interacting with the circulating native or modified lipoproteins, such as oxLDL, these receptors modulate vascular inflammation, lipid accumulation, and plaque formation. Both CD36 and SR-BI have two short N- and C-terminal cytoplasmic tails, two transmembrane domains, and a large extracellular domain that binds to the ligands. CD36, through binding to modified lipoproteins and lipids, plays a major role in atherogenesis, by promoting endothelial dysfunction, macrophage foam cell formation, vascular inflammation, and atherothrombosis. Hepatic SR-BI protects against atherosclerosis by facilitating the reverse cholesterol transport and maintaining anti-atherogenic characteristics of HDL in the vascular cells; however, recent findings indicate that the receptor has a key pro-atherogenic function in endothelium by mediating transport of LDL into the artery wall. LOX-1 contains a short N-terminal cytoplasmic domain, a single transmembrane domain, a NECK or stalk domain, and a C-terminal extracellular domain that binds to oxLDL. In a variety of vascular cells, LOX-1 mediates a majority of pro-atherogenic actions of oxLDL, contributing inflammation, smooth muscle cell proliferation, and platelet activation.

Table 2.

Anti- and proatherogenic functions of the scavenger receptor family members

		Antiatherogenic functions			Proatherogenic functions
	Lipoprotein receptor	Cell type	Cellular function	Mechanism	Cell type	Cellular function	Mechanism	References
Scavenger receptors	CD36	Macrophages			Macrophages	Inflammation ↑	NLRP3 inflammasome ↑	151–159
						Spreading ↓	FAK, SHP-2 ↓
						Cellular polarity ↓	Rac, non-muslce myosin II ↓
						oxLDL uptake ↑	Vav ↑
		Endothelial cells			Endothelial cells	Apoptosis ↑	TSP-Fyn-p38 ↑	160–165
						Cell migration/tube formation ↓	TSP-VEGFR signalling ↓
							Fyn-NADPH oxydase-ROS ↑
		SMC			SMC	Proliferation ↑
		Platelets			Platelets	Activation ↑	Fyn-PKC-NADPH oxidase-ROS ↑	166–173
							JNK/ERK5 ↑
							NO-cGMP ↓
	SR-BI	Macrophages	Cholesterol efflux ↑		Macrophages			174–185
			Inflammation ↓	NF-kB, p38, JNK ↓
			Cholesterol hydrolysis ↑
			Efferocytosis ↑	Src, PI3 kinase Rac1 ↑
		Endothelial cells	eNOS activity ↑	Src, Akt ↑	Endothelial cells	LDL transcytosis ↑	DOCK4-Rac1 ↑	186–196
				Sphingosine-1-phosphate receptor signalling ↑
			HDL transport ↑
	LOX1	Endothelial cells			Endothelial cells	Inflammation ↑	MCP1 ↑	197–201
						Apoptosis ↑	Caspase ↑
							Anti-apoptotic proteins ↓
						eNOS activity ↓	RhoA-ROCK-arginase II ↑
							Dysfuncitonal HDL action ↑
		SMC			SMC	Proliferation ↑		202–205
						Apoptosis ↑
		Macrophages			Macrophages	Oxidative stress ↑		206–209
						Cell migration ↓
		Platelets			Platelets	Aggregation/activation ↑		210–213

3.1 Cluster of differentiation 36

CD36 (also referred as SR-B2) belongs to the class B scavenger receptor family, and it is highly expressed in vascular and hematopoietic cells, including vascular SMCs, endothelial cells, macrophages, and platelets.^166,167,214 CD36 has two short N- and C-terminal cytoplasmic tails, two transmembrane domains, and a large extracellular domain that is heavily glycosylated^166,214 (Figure 2). A homology modelling based on published extracellular structure of the related protein Lysosomal Integral Membrane Protein II (LIMP-II) indicates that there are distinct hydrophobic pockets potentially responsible for lipid recognition and fatty acid trafficking.²¹⁵ Through discrete sites on the extracellular loop, CD36 interacts with multiple ligands, including oxidized phospholipid (oxPC_CD36),²¹⁶ free fatty acids,²¹⁷ the matricellular protein thrombospondins (TSP),¹⁶⁰ and proteins containing the thrombospondin type 1 repeat (TSR) domain²¹⁸ and cell-derived microparticles.²¹⁹ CD36 has at least three well-established, cell-type specific functions. First, it serves as a membrane receptor for TSP in endothelium to modify cellular functions, such as angiogenesis.¹⁶⁰ Second, CD36 facilitates translocation of free fatty acids from the extracellular milieu into the cytoplasm, mainly in adipocytes and myocardium.^217,220 Third, it functions as a scavenger receptor that recognizes and internalizes specific exogenous and endogenous danger signals, including modified LDL and HDL.^166,167,214 This function plays a pivotal role in the pathogenesis of atherosclerosis, by promoting endothelial dysfunction, macrophage foam cell formation, vascular inflammation, and atherothrombosis.

3.1.1 Macrophage CD36

Under hyperlipidaemic condition, cholesterol derived from LDL particles accumulates in the subendothelial space in the vascular wall and triggers inflammation, leading to foam cell formation.^2,12 Numerous studies in cultured cells and in mouse models have shown that CD36 plays a major part in the process through multiple mechanisms. First, in response to oxLDL CD36 promoted lipid accumulation in cultured macrophages, and CD36 deficiency in mice (CD36^−/−) was associated with attenuated foam cell formation and protection from atherosclerosis.^151,152 However, the in vivo consequence of CD36 deletion in foam cell formation and atherogenesis in mice has been disputed by another study, which did not find significant reduction in foam cell formation or atherosclerosis.¹⁵³ Secondly, CD36 has been shown to promote a pro-inflammatory process in macrophages by oxLDL internalization and subsequent conversion into cholesterol crystals, potent inducers of the NLRP3 inflammasome, and production of proinflammatory cytokine Interleukin 1β (IL-1β).¹⁵⁴ Third, CD36 promotes macrophage dysfunction by initiating a signalling pathway in macrophages in response to oxLDL. Park et al.¹⁵⁵ revealed that CD36 mediates oxLDL-induced inhibition of macrophage spreading and migration via sustained activation of focal adhesion kinase and inactivation of Src homology 2-containing phosphotyrosine phosphatase (SHP-2), leading to disruption of actin cytoskeletal rearrangements. Another study showed that oxLDL-CD36 signalling leads to a loss of macrophage cellular polarity due to dysregulation of Rac and non-muscle myosin II, key determinants of cellular symmetry and lamellipodia formation.¹⁵⁶ Additionally, a mechanism in macrophages has been reported that involves CD36-mediated activation of Vav family guanine nucleotide exchange factors.^157,158 Vav is highly regulated by Src-mediated tyrosine phosphorylation and it acts as a central signalling hub to several signalling pathways,¹⁵⁹ and CD36 has been shown to activate Vav via Src family kinases, which in turn promotes uptake of oxLDL.^157,158In vivo, analysis of aortas from hyperlipidaemic mice indicated that Vav proteins are activated in a CD36-dependent manner. In sum, CD36 mediates pro-inflammatory and proatherogenic actions of oxLDL in macrophages.

3.1.2 Endothelial cell and smooth muscle cell CD36

Several studies have linked CD36 to endothelial dysfunction, which plays a key role in the initial stage of cardiovascular diseases.^2,8 CD36 in endothelium has been shown to interact with to the TSR domain in TSP1 and TSP2.¹⁶⁰ A study showed that the interaction between TSP1 and endothelial CD36 increases phosphorylation of Src family kinase Fyn, followed by activation of p38 MAPK, leading to enhanced apoptosis.¹⁶¹ Another study found that TSP1-CD36 interaction promotes dephosphorylation of pro-angiogenic vascular endothelial growth factor receptors (VEGFR), resulting in reduction of endothelial cell migration and formation of tube-like structures.¹⁶² CD36 on microvascular endothelial cells also binds to anionic phosphatidylserine on the surface of extracellular vesicles, leading to inhibition of cell migration.¹⁶³ This is mediated by a pathway that involves activation of Fyn and NADPH oxidase and generation of reactive oxygen species (ROS). Recent studies found that mice lacking CD36 selectively in endothelial cells have reduced uptake of fatty acids in the heart, skeletal muscle, and brown adipose tissue and that these mice are resistant to high-fat diet-induced insulin resistance.¹⁶⁴

CD36 has also been shown to contribute to development of neointima hyperplasia through enhancement of SMC proliferation.¹⁶⁵ Yue et al. found that CD36 is highly expressed in human vessels with intimal hyperplasia and luminal occlusion, and that mice lacking CD36 globally or selectively in SMCs on hyperlipidaemic background have reduction in carotid artery injury-induced neointima thickening. In primary vascular SMCs in culture, the study further demonstrated that CD36 is required for induction of cell proliferation and cell cycle progression by serum or a TSR peptide. These observations indicate that CD36 in the endothelium or SMCs exerts various proatherogenic functions; however, an effect of tissue-specific CD36 deletion on atherogenesis is yet to be tested.

3.1.3 Platelet CD36

Studies in mice and in isolated platelets have demonstrated a critical role of platelet CD36 in linking dyslipidaemia and oxidative stress with prothrombotic phenotypes. CD36^−/− mice on the hyperlipidaemic apoE^−/− background exhibited reduced thrombus formation induced by vascular injury.¹⁶⁷ In isolated human platelets, inhibition of CD36 actions by the blocking antibody or the competing lipid ligand attenuated platelet activation by oxidized lipids without affecting basal platelet functions. In response to oxLDL, platelet CD36 has been shown to activate Src family kinase Fyn, which leads to increased ROS production via NADPH oxidase 2 (NOX2).^166,168 Activation of Fyn upregulated another tyrosine kinase Syk and PLCγ2, linking the CD36 signalling to PKC, which in turn promotes phosphorylation of the p47 subunit of NOX2 to induce NOX complex formation to generate ROS.^166,168 Platelets express a number of redox-sensitive effectors for ROS, including MAP kinase family members JNK and ERK5.^169–172 Inhibition of JNK activity prevented platelet activation by oxLDL, and an activated form of JNK was detected within growing thrombi in wild-type mice, but not in CD36^−/− mice.¹⁶⁹ For ERK5, in isolated platelets activation of the kinase by oxLDL was observed, and platelets isolated from mice exposed to hyperlipidaemia showed elevated activation of ERK5.¹⁷⁰In vivo, apoE^−/− mice harboring ERK5 deficient platelets displayed attenuated thrombus formation compared to those expressing the kinase in platelets.¹⁷⁰ Intervention studies in mice confirmed that oxLDL induction of ERK5 requires CD36, Src family kinases, and NOX. Platelet CD36-ERK5 has been also implicated in a coagulation pathway.¹⁷³ The study reported that CD36 activation of ERK5 upregulates the caspases, resulting in externalization of phosphatidylserine that promotes a formation of coagulation enzyme complexes, and that inhibition of CD36 or ERK5 prevents accelerated fibrin formation.¹⁷³ In addition to the direct activation of platelets or the coagulation pathway, ROS generated by the CD36-NOX pathway has also been shown to suppress an inhibitory pathway in platelets that is mediated by nitric oxide-cGMP signalling.²²¹ Collectively, CD36 promotes platelet activation and coagulation by oxLDL, contributing to prothrombotic phenotypes under hyperlipidaemia.

3.1.4 Human genetics of CD36

The human CD36 gene is located on chromosome 7q11.2, and it has 15 exons encoding a 471 amino acid protein.^222,223 CD36 deficiency has been reported to occur in 3–10% of Asian and African populations.^224–226 Two types of CD36 deficiency have been described; type I lacks CD36 both in monocytes and platelets, and type II lacks the receptor expression in platelets but not monocytes.²²⁷ Type I CD36 deficiency has been associated with insulin resistance, postprandial hypertriglyceridaemia and hypertension in Japanese populations.^228–231 In addition to CD36 deficiency, a number of common genetic variants have been discovered that are associated with metabolic syndrome, hypertension, coronary artery diseases, and ischaemic stroke in multiple populations.^{226,232–238} Ghosh et al. demonstrated that platelet CD36 expression levels vary among individuals and that specific polymorphisms in the CD36 gene are associated with platelet surface expression levels of the receptor and with responsiveness to its ligand oxLDL.²³³ In agreement with the proinflammatory and proatherogenic functions of the receptor shown in the experimental models, SNPs associated with higher CD36 in platelets were linked to obesity, metabolic syndrome, and dyslipidemia.²³³ However, the basis for the contrasting effects of complete vs. a partial loss of CD36 on the metabolic and cardiovascular complications is not yet fully understood.

3.2 Scavenger receptor class B type I

The scavenger receptor Class B type I (SR-BI, SR-B1, or SCARB1) was first identified as a protein with a sequence closely related to CD36 and LIMP-II, and the gene was initially named CLA1, subsequently termed SCARB1^239,240 (Figure 2). The work by Acton et al. identified SR-BI as the high-affinity HDL receptor, which mediates the selective uptake of HDL-associated cholesterol ester (CE) into the liver.²³⁹ SR-BI also facilitates the bidirectional flux of free cholesterol (FC) between cells and HDL.^241,242 A major role of SR-BI in lipoprotein metabolism and atherogenesis was initially established in mice lacking the receptor globally (SR-BI^−/−). SR-BI^−/− mice on hyperlipidaemic backgrounds displayed accelerated atherosclerosis resulting in coronary artery occlusion, MI, and decreased survival rate.^174,243,244 These mice have severe hypercholesterolaemia with abnormally enlarged HDL particles compared to control mice. Subsequent investigations revealed that the accelerated atherosclerosis observed in SR-BI^−/− mice is due primarily to a loss of hepatic SR-BI. Liver-specific SR-BI deficiency in mice resulted in increased atherosclerosis, mimicking the phenotype in global SR-BI knockout mice,^186,245 and overexpression of hepatic SR-BI reduced atherosclerosis in LDLR^−/− mice.²⁴⁶ Multiple mechanisms by which hepatic SR-BI provides atheroprotection have been proposed. First, SR-BI in the liver plays a critical role in the reverse cholesterol transport, in which peripheral cholesterol is delivered to the liver for excretion in both mice and humans.^14,17 In humans, although majority of plasma cholesterol in LDL is internalized by hepatic LDL receptors, it has been estimated that hepatic SR-BI is responsible for at least 33% of HDL CE uptake.²⁴⁷ Consistent with a role for SR-BI in hepatic uptake of HDL in humans, subjects carrying a loss of CE transport function variant in the SR-BI (SCARB1) gene (P297S) have elevated HDL cholesterol in the form of enlarged, apoE-enriched particles, parallel to those observed in SR-BI^−/− mice.^248,249 Additionally, the HDL particles isolated from SR-BI^−/− mice showed altered lipid and protein compositions with increased FC and decreased antioxidative molecules, likely contributing to reduced cholesterol efflux capacity.^243,250,251 Secondly, hepatic SR-BI has been implicated in clearance of proatherogenic lipoproteins VLDL and Lp(a),^252,253 suggesting that impaired removal of these molecules may also contribute to the exaggerated atherogenesis in SR-BI^−/− mice. Third, hepatic SR-BI may also provide atheroprotection by attenuating thrombosis.^254–256 SR-BI^−/− mice displayed increased susceptibility to arterial thrombosis.^254,255 Bone marrow transplantation experiments demonstrated that hyperlipidaemia and resulting platelet FC overload due to loss of hepatic SR-BI, not SR-BI deficiency in platelets, are likely responsible for the prothrombotic phenotype observed in SR-BI^−/− mice.^254,255 Studies in humans also found that platelets isolated from subjects carrying the SR-BI variant P297S are enriched in FC and that they are prothrombotic compared with control platelets.^248,249 Taking together, hepatic SR-BI protects against atherosclerosis by facilitating the reverse cholesterol transport and maintaining anti-atherogenic characteristics of HDL.

In addition to the liver, SR-BI is abundantly expressed in macrophages and endothelial cells, and accumulating evidence indicates that the receptor has extrahepatic functions to influence the progress of atherosclerosis.^{175–179,186,245,257}

3.2.1 Macrophage SR-BI

Bone marrow transplantation studies have established that SR-BI in monocyte or macrophages provides atheroprotection. ApoE^−/− or LDLR^−/− mice receiving transplantation of SR-BI deficient bone marrow cells developed accelerated atherosclerosis, compared those receiving control bone marrow cells.^175–178 No differences were observed in plasma cholesterol or lipoprotein profiles, indicating that proatherogenic effect of monocyte/macrophage SR-BI deficiency is not likely due to abnormal lipoproteins. Likewise, transplantation of SR-BI^−/− mice on apoE hypomorphic background with SR-BI^+/+ bone marrow cells markedly reduced coronary atherosclerosis and MI.¹⁷⁹ Multiple mechanisms have been proposed by which macrophage SR-BI affords atheroprotection. One mechanism is through mediating removal of cholesterol from macrophages. Several studies in culture have shown that SR-BI, along with ABCA1, is responsible for cholesterol efflux to HDL from cholesterol-laden monocyte-derived macrophages.^180–182 Macrophage SR-BI has been also implicated in suppression of proinflammatory aspects associated with atherosclerosis. Studies have found that hyperlipidaemic mice receiving SR-BI^−/− bone marrow showed increased serum levels of proinflammatory cytokines including IL-1β, IL-6, and TNF-α, compared to those receiving control bone marrow.¹⁷⁵ Furthermore, both in vitro and in vivo models indicate that SR-BI deficiency promotes macrophage conversion to the pro-inflammatory phenotype likely through activation NF-kB, p38, and JNK signalling pathway.^174,183 Another mechanism by which macrophage SR-BI could reduce atherosclerosis is by targeting cytoplasmic CE via autophagosome for lysosomal hydrolysis.¹⁸⁴ Studies in cultured macrophages demonstrated that autophagic process is impaired in SR-BI-deficient macrophages in the setting of infection. An additional mechanism for macrophage SR-BI in efferocytosis has also been proposed, in which haematopoietic SR-BI deficiency results in severely defective efferocytosis in mouse atherosclerotic lesions.^175,185 The studies further found that efferocytosis of apoptotic cells was reduced in SR-BI^−/− vs. wild-type macrophages and that SR-BI interaction with cell surface phosphatidylserine of the apoptotic cells triggers efferocytosis process via activation of Src, PI3 kinase, and Rac1.¹⁷⁵ Defective efferocytosis in macrophages thus may play a role in exaggerated atherosclerotic phenotypes in SR-BI^−/− mice. Collectively, a number of studies both in vitro and in vivo strongly support atheroprotective functions of macrophage SR-BI.

3.2.2 Endothelial cell SR-BI

Disruption of vascular endothelial function is a key initial step in atherosclerotic lesion progression, facilitating monocyte infiltration into the arterial intima.^2,8 A series of studies have established that SR-BI interaction with HDL prevents endothelial dysfunction by stimulating eNOS activity and increasing bioavailable NO.^187,188 NO is a key signalling molecule for the maintenance of normal vascular function, through regulation of vascular tone, prevention of endothelial inflammation and platelet activation, and reduction of SMC growth.^258,259 Studies in cultured endothelial cells indicate that HDL-SR-BI activation of eNOS requires recruitment of the adaptor molecule PDZK1 to C-terminal cytoplasmic domain of the receptor, which in turn activates Akt kinase, resulting in the phosphorylation of eNOS at Ser1177.^188,189In vivo, mice lacking SR-BI or PDZK1 show attenuated endothelial repair.^189,190 Furthermore, both in vivo and in vitro studies revealed that anti-inflammatory and atheroprotective effects of HDL-associated sphingosine 1-phosphate (S1P) are mediated by endothelial SR-BI along with S1P receptors.^191,192In vivo overexpression of SR-BI selectively in endothelial cell has been shown to decrease atherosclerosis in wild-type, apoE^−/− or SR-BI^−/− mice on atherogenic diet.¹⁹³ Although the atheroprotective phenotype in these mice is likely in part due to decreased plasma cholesterol and increased HDL cholesterol, endothelial SR-BI may also contribute to the reduced atherosclerosis by facilitating HDL transport between circulation and arterial wall to promote cholesterol efflux from macrophages.^193,194 An additional atheroprotective mechanism has been proposed by which lymphatic endothelial SR-BI mediates the transport of HDL in lymphatic vessels to effectively remove cholesterol from peripheral tissue.^195,196

In contrast to these atheroprotective roles of endothelial SR-BI described above, a strong proatherogenic function of endothelial SR-BI has been recently revealed.^186,260 In cultured endothelia cells and in isolated mouse aortas ex vivo, Armstrong et al.²⁶⁰ first demonstrated using total internal reflection fluorescence microscopy that LDL is transported across endothelial cells, and that the LDL transport is mediate by SR-BI, but not LDL receptor. More recently, using mice lacking SR-BI selectively in endothelial cells, Huang et al.¹⁸⁶ have demonstrated in vivo that endothelial SR-BI promotes atherosclerosis by facilitating delivery of LDL into arteries and its accumulation in macrophages. The study discovered that internalization of LDL-SR-BI requires recruitment of the guanine nucleotide exchange factor dedicator of cytokinesis 4 (DOCK4) to the C-terminal domain of SR-BI and subsequent activation of Rac1. It further revealed that the expressions of SR-BI and DOCK4 transcripts are elevated in human atherosclerotic arteries, compared with normal arteries. In summary, current evidence indicates a dichotomous role of endothelial SR-BI in atherogenesis. The receptor mediates anti-inflammatory effects of HDL, contributing to increased bioavailable NO and protection against atherosclerosis. On the other hand, in conditions of high plasma LDL, SR-BI facilitates LDL transport across the endothelial layer into the artery wall, which leads to foam cell formation and progression of atherosclerosis.

3.2.3 Human genetics of SR-BI

The discovery of SR-BI encoded by SCARB1 gene as the high affinity HDL receptor triggered numerous studies to identify human variants at the locus for their relationship with plasma lipid phenotypes and cardiovascular diseases.^261,262 Earlier works using targeted resequencing and genotyping discovered several coding polymorphisms that are associated with high plasma levels of HDL or low levels of LDL cholesterol.^261,262 Some of the studies found the variants that are associated with increased risk for vascular diseases, including MI, peripheral arterial disease, common carotid intimal–medial artery thickness and coronary heart disease.^263–265 GWAS further identified SCARB1 SNPs that are associated with HDL cholesterol levels and cardiovascular disease risks.^88,266,267 More recently, rare nonsynonymous variants in SCARB1 gene have been reported, further indicating the association of SR-BI, HDL metabolism, and risk for cardiovascular diseases. Through targeted sequencing of participants with extremely high HDL cholesterol levels, Vergeer et al. identified a carrier of a novel P297S variant in SCARB1 gene.²⁴⁸ Compared to non-carrier family members, P297S carriers exhibited higher HDL cholesterol levels, and HDL isolated from the carriers displayed reduced cholesterol efflux capacity to macrophages, compared to non-carriers. However, the variant was not associated with differences in carotid intimal-medial thickness. Subsequently, a second non-synonymous coding variant in SCARB1 (P376L) was reported.²⁴⁹ The variant was also identified through targeted sequencing in individuals with high HDL cholesterol levels. The P376L homozygous subject exhibited extremely high levels of circulating HDL cholesterol and a carotid intimal medial thickness that was higher than 75th percentile for age and gender. Genotyping of the P376L variant in multiple cohorts revealed a significant association of the P376L variant with increased coronary heart disease. Biochemical analysis demonstrated that HDL particles in P376L carriers are larger in size with increased apoA-I content, and that the selective HDL cholesterol uptake function is absent in hepatocyte-like cells derived from induced pluripotent stem cells (iPSC) from the homozygous subject. Collectively, these studies in humans offered a strong support for a key role of SR-BI in HDL metabolism and atherosclerosis.

3.3 Lectin-like oxidized low-density lipoprotein receptor-1

Lectin-like oxidized low-density lipoprotein receptor-1 (LOX-1, also known as OLR1, SR-E1) is a 50 kDa transmembrane glycoprotein that belongs to the C-type lectin family.^268,269 LOX-1 contains a short N-terminal cytoplasmic domain, a single transmembrane domain, a NECK or stalk domain, and a C-terminal extracellular domain that binds to oxLDL (Figure 2). In human atherosclerotic lesions, LOX-1 mRNA and proteins were detected in endothelial layers, intimal SMCs and macrophages, whereas they were undetectable in normal human aortas.²⁷⁰ Furthermore, elevations in circulating indicators of LOX-1 expression or activation in humans have been associated with a higher risk for coronary heart diseases.^271,272 A causal role of LOX-1 for atherogenesis was revealed by Mehta et al. using mice lacking LOX-1 globally (LOX-1^−/−).²⁷³ These mice displayed attenuated oxLDL binding to aortic endothelial cells, and on hyperlipidaemic LDLR^−/− background, LOX-1^−/− mice showed a marked reduction in atherosclerotic lesion and intima thickness, compared to LDLR^−/− mice. The study further found that LOX-1^−/− mice have decreased expression of proinflammatory markers NF-κB, CD68, and the phosphorylated p38 MAPK in the vascular wall. In agreement with the proatherogenic role of LOX-1, the receptor is also implicated in pathogenesis of myocardial ischaemia in mice. Using a rat model of coronary artery ischaemia–reperfusion, a study showed that treatment with an anti-LOX-1 neutralizing antibody, which blocks LOX-1-mediated cellular uptake of oxLDL, results in reduction in infarct size.²⁷⁴ In this rat model, expression of LOX-1 mRNA and protein was upregulated in the ischaemic heart, which was associated with elevations of markers of inflammation, oxidative stress, and apoptosis.²⁷⁴ Studies in LOX-1^−/− vs. wild-type mice confirmed the role of LOX-1 in ischaemia, showing that LOX-1 deficiency improved survival and reduced collagen deposition and cardiomyocyte hypertrophy.²⁷⁵

LOX-1 was originally identified in cultured aortic endothelial cells as a major receptor responsible for binding, internalization, and degradation of oxLDL.²⁶⁸ LOX-1 was subsequently detected in other cell types, including macrophages, platelets, SMCs, and cardiomyocytes.^{202,206,207,210,276,277} The cellular expression of LOX-1 at basal state is normally low, but it is known to be rapidly upregulated by oxidative and proinflammatory stimuli.^210,269,276

3.3.1 Endothelial cell and smooth muscle cell LOX-1

A number of studies in culture and in vivo models demonstrated that activation of LOX-1 promotes endothelial dysfunction through multiple mechanisms. First, in cultured human endothelial cells, oxLDL was shown to increase MCP1 production and monocyte adhesion in an LOX-1 dependent manner.¹⁹⁷ In a rat model of low-dose endotoxin-induced uveitis, the LOX-1 blocking antibody suppressed leucocyte infiltration in vivo and it reduced leucocyte interaction with retinal veins in situ.¹⁹⁸ Second, LOX-1 has been shown to mediate oxLDL-induced endothelial cell apoptosis in human coronary artery endothelial cells through increased activation of the caspases and reduced expression of anti-apoptotic proteins.¹⁹⁹ Third, oxLDL-LOX-1 also contributes to endothelial dysfunction through upregulation of arginase II, which inhibits NOS activity by competing for the substrate L-arginine.²⁰⁰ In cultured endothelial cells, arginase II activation by oxLDL was attenuated by preincubation with the LOX-1 blocking antibody, and in aortas from wild-type mice fed high-cholesterol diet showed increased arginase II activity and reduced NO production, whereas LOX-1^−/− mice were protected from the adverse effects of the atherogenic diet. Mechanistically, the effect of oxLDL-LOX-1 on arginase II activation was shown to be mediated by RhoA and ROCK signalling. In addition to its role as a mediator of proatherogenic actions of oxLDL, LOX-1 has been implicated in the actions of dysfunctional HDL in endothelial cells. Studies by Besler et al. showed that disease-modified dysfunctional HDL engages LOX-1 instead of SR-BI, resulting in endothelial dysfunction and elevated inflammatory responses.²⁰¹ In contrast to HDL from normal subjects that promotes endothelial NO release in cultured human endothelial cells, HDL from patients with acute coronary syndromes or stable coronary artery diseases failed to induce endothelial NO release. Treatment of the cells with the LOX-1 blocking antibody reversed the defect.²⁰¹

Involvement of LOX-1 in SMCs has been shown in neointima hyperplasia. In balloon-injury animal models, expression of LOX-1 mRNA and protein was upregulated in medial and neointimal SMCs in response to injury, and administration of the LOX-1 blocking antibody suppressed injury-induced neointima hyperplasia, along with ROS formation and leucocyte infiltration in the vascular wall.²⁰² LOX-1 expression was also detected in proliferating SMCs in human restenotic lesions after balloon-angioplasty.²⁰³ In addition to neointima formation, LOX-1 in SMC has been implicated in destabilization and rupture of atherosclerotic plaques.^204,205 In cultured SMCs, the studies showed that oxLDL treatment enhances SMC apoptosis through LOX-1 and that the neutralizing LOX-1 antibody prevents the oxLDL-induced apoptosis.^204,205 Furthermore, LOX-1 expression was colocalized with apoptosis proteins in the rupture-prone regions of human atherosclerotic plaques. Collectively, these findings indicate that LOX-1 in endothelial cells and SMCs mediate potent atherogenic effects of oxLDL during various phases of atherogenesis.

3.3.2 Macrophage and platelet LOX-1

In addition to endothelial cells and SMCs, LOX-1 also facilitates binding and uptake of oxLDL in monocytes and macrophages.^206,207 Studies in cultured macrophages have shown that oxLDL treatment increases oxidative stress in an LOX-1 dependent manner, and that oxLDL suppresses cell motility in macrophages isolated from wild-type mice, but not from LOX-1^−/− mice.^208,209

In activated human platelets, LOX-1 expression is upregulated and the receptor mediates the binding and internalization of oxLDL, similar to other cell types.^210,211 A study in isolated human platelets showed that ADP-induced platelet aggregation was inhibited by the LOX-1 blocking antibody through suppression of the inside-out integrin signalling-dependent activation of glycoprotein IIb/IIIa.²¹¹ Additionally, human electronegative LDL (L5) isolated from patients with MI, which is known to serve as a ligand for LOX-1,²¹² induced platelet activation and aggregation.²¹³In vivo, injection of the L5 into wild-type mice resulted in platelet activation and shortening of tail bleeding time, whereas the effect of L5 was attenuated in LOX-1^−/− mice. In sum, there is evidence suggesting that LOX-1 mediates oxLDL actions in macrophages and in platelets that promote oxidative stress, inflammation, and thrombosis.

3.3.3 Human genetics of LOX-1

Human LOX-1 is encoded by the oxidized low density lipoprotein receptor 1 (OLR1) gene located in the p12.3-p13.2 region of human chromosome 12.^268,278 Genetic studies have identified SNPs in the OLR1 gene that are associated with myocardial ischemia and coronary artery diseases. One of the SNPs is G501C variant in the coding region, resulting in a missense mutation substituting lysine (K) for asparagine (N) at amino acid location 167 (K167N). Amino acid residue 167 is located on the surface of the C-terminal domain of LOX-1, suggesting that the mutation may affect its interactions with ligands. In cross-sectional studies, this variation was associated with an increased incidence of MI, as well as with increased carotid intima-media thickness.^279–281 A meta-analysis of seven case–control studies indicated that the C allele is significantly associated with ischemic stroke.²⁸² However, an in vitro study showed that 167N LOX-1, corresponding to C allele, binds to oxLDL less efficiently than 167K (G allele), which likely confers protection against atherosclerosis.²⁸³ This disparity in the biochemical and human genetic association studies needs further clarification.

When SNPs at distinct locations of the genome are more highly associated than expected, they are in the linkage disequilibrium (LD). An LD block comprises six such SNPs (GGAAGC haplotype vs. CTGGTT) has been found within introns 4 and 5, as well as the 3' untranslated region (3'-UTR) of the OLR1. Studies found an association between the SNPs within the LD block and the risk for MI or atherosclerotic cerebral infarction in some cohorts.^284,285In vitro and in vivo studies have shown that LD block SNPs regulate the production of an alternatively spliced version of the LOX-1 mRNA by modulating retention of the LOX-1 mRNA exon 5.²⁸⁶ Exclusion of exon 5 results in the production of a LOX-1 isoform, termed LOXIN, which lacks a part of the C-terminal lectin-like domain involved in ligand binding. Studies in isolated peripheral blood monocytes from risk (GGAAGC) vs. non-risk (CTGGTT) carriers of the LD SNPs showed that the non-risk SNPs correlate with higher expression of LOXIN, which results in a significant reduction of apoptosis induced by oxLDL treatment.²⁸⁶ The splice variant may have a functional role on plaque instability and pathogenesis of MI.^286,287 In summary, studies in human genetics provide supportive evidence that implicates LOX-1 as an important mediator of proatherogenic functions of oxLDL.

4. Conclusions

The lipoprotein receptors are expressed in multiple types of vascular cells, and they mediate actions of a broad range of ligands to accelerate or block atherogenesis. It is of pivotal importance that we harness these functions of the lipoprotein receptors to develop therapeutic modalities that can prevent or lessen atherosclerosis. To achieve the goal, more comprehensive understanding of the complex process will be needed, likely through employment of innovative strategies. In vitro cell culture models, using either established cell lines or primary human or mouse cells, have allowed us to identify the players in particular signal transduction pathways; however, there are well acknowledged limitations in the approach, stemming from inevitable changes in signature characteristics or a partial loss of cellular identity in cells transferred to culture. Although still in an early stage, human iPSC-derived vascular cells may provide an in vitro modelling of human atherosclerosis or vascular disorders.²⁸⁸ Animal models, especially genetically modified mice on hyperlipidaemia background, have made loss- and gain-of-function studies possible. However, because these receptors regulate diverse aspects of the disease process, it has become apparent that a more temporal and cell-type specific manipulation is necessary to distinguish the direct effects from less specific consequences of compensatory or indirect mechanisms. Lastly, an approach of using human genetics to search for variants that show associations with human conditions has yielded a wealth of information, connecting data from experimental models with human diseases. However, due to involvement of multiple biological pathways and cell types, clear associations with cardiovascular disease may not be readily apparent, especially for more common variants. Recent advancement in the area of transcriptome analysis, including a single cell sequencing of normal and diseased human tissues, indicates that such approach will provide much-needed information regarding the processes which occur in unique sets of vascular cells during atherosclerosis development in humans.²⁸⁹

Conflict of interest: none declared.

Funding

The work was supported by NIH (R01HL126795 and R01DK110127 to C.M.).