Atherogenesis and Advanced Glycation: Promotion, Progression, and Preventiona

The nonenzymatic glycation of proteins by aldehyde or keto-groups of reducing sugars, in a process known as the Maillard reaction, eventually leads to the formation of advanced glycation endproducts (AGEs).1–3 These highly reactive and crosslinking adducts have been implicated as causal factors in many vascular diseases, especially diabetic macro- and microvasculopathy, in many cases acting through a specific AGE receptor-mediated mechanism.4 Recently, AGEs are beginning to be recognized as playing a role in the formation and acceleration of atherosclerotic lesions, not only in diabetics5 but also in patients without diabetes.6–9 Atherosclerosis is marked not only by the complex nature of its pathogenesis but also by its involvement in a number of different pathologies; most notably myocardial and cerebral infarction. Formation of the atherosclerotic plaque is the pathophysiological hallmark of this process. The interactions between blood-borne proteins and the many distinct cellular components of the vessel wall that lead to the formation of dense plaques have been shown to be extremely complex and multifactorial in nature. The following review will discuss the role of advanced glycation in large-vessel vasculopathy and how AGEs and AGE receptors may play a significant role in the transformation of vessel wall architecture and performance, as well as the initiation and progression of plaque formation. Furthermore, we highlight the potential for assessment of patients at risk by measuring serum AGE levels and a novel atherosclerosis intervention therapy through the AGE inhibitor, aminoguanidine. In the early stages of the Maillard reaction, the spontaneous and nonenzymatic reaction of reducing sugars with proteins forms a variety of “early” products that are largely reversible and reach equilibrium with ambient glucose over a short period of time.10 The intermediate and late products of this reaction consist of a heterogeneous group of sugar-amino adducts that can be transformed, through cyclical condensation, dehydration, and oxidation reactions, into irreversible advanced glycation endproducts (AGEs) (Fig. 1). AGE moieties persist for the lifetime of the protein or lipid they are attached to. can display distinctive fluorescence, and have the ability to crosslink to each other. Structure of advanced glycation endproducts. “R” represents the lysine or arginine group on protein. Furoyl-furanyl imidazole (FFI), 1-alkyl-2-formyl-3,4-diglycosyl pyrrole (AFGP), N-carboxymethyl-lysine (CML). Due to the heterogeneous nature of the precursor adducts, instability during isolation, and microenvironmental variation in reaction conditions, the precise structural elucidation of AGEs as they form in vivo remains largely unknown. However, several AGE adducts have been identified as model AGEs, such as 2-(2-furoyl)-4(5)-furanyl-1H-imidazole (FFI), 1-alkyl-2-formyl-3,4-glycosyl-pyrrole (AFGP), and pyrraline.11,12 Although their in vivo significance remains controversial,13,14 these model AGEs have been useful models for the investigation of crosslinking mechanisms. AGE formation on a number of proteins has been extensively characterized, such as human serum albumin, in which at least 10 of the 59 ɛ-amino lysine residues are subject to glycation in vivo.15 It is also known that Amadori product or AGE attachment occurs at distinct lysine residues, which may or may not influence the enzymatic activity of proteins and may depend on the inherent differences in the reactivity of amino residues.16 Of paramount importance in initiation of atherosclerotic lesion formation is the increased accumulation of low density lipoprotein (LDL) species in the vessel wall leading to the formation of “fatty streaks.” It has been widely speculated that oxidative modification of LDL (ox-LDL) in vivo results in the reduced recognition by the LDL receptor, in increases in serum LDL levels, and in an increased uptake of the modified protein by scavenger receptors on macrophages and vascular smooth muscle cells. However, studies in our own laboratory suggest that advanced glycation of both the lipid and apoB components of the LDL molecule (lipid-AGE/apoB-AGE) represents a physiologically significant modification of lipoproteins17 that occurs concomitantly with LDL oxidation in vitro.18 Indeed, the presence of amino groups on phospholipids, such as phosphatidylethanolamine and phosphatidylserine, suggests that glucose reacts with lipid amines to form AGEs.18 Oxidation-reduction reactions occurring during advanced glycation may oxidize fatty acid residues independently of transition metals or exogenous, free radical-generating systems that are frequently added in experimental studies. When free amine-containing phosphatidylethanolamine was incubated with glucose and metal chelators, AGEs, as well as oxidation products, formed as a function of glucose concentration.18 Significantly, LDL oxidation and AGE-LDL formation are inhibited by co-incubation with the AGE inhibitor, aminoguanidine. Therefore it is apparent that amine-glucose interactions leading to fatty acid oxidation products are a possible in vivo pathway leading to atherogenic lipid formation. Advanced glycation of LDL modification is found to be markedly elevated in diabetics and non-diabetic patients with atherosclerosis.19 Chemical modification of apoB has been shown to interfere with its recognition by the LDL receptor18 and could possibly lead to abnormally high levels of circulating LDL.20,21 ApoB is a relatively large protein with many potential lysine and arginine AGE modification sites (Fig. 2), but the predominant site of AGE modification of apoB has now been confirmed to be at ∼ 1,791 amino acid residues distal to the putative binding domain recognized by the LDL receptor.17 Schematic diagram of the low density lipoprotein (LDL) particle showing the AGE-modified domain. A striking correlation has recently been demonstrated between serum apoB-AGE and AGE tissue levels in the vessel wall of endarterectomized carotid arteries from non-diabetic patients.9 The vascular endothelium expresses receptors for AGEs,22–24 and it is likely that circulating AGE-LDL can be endocytosed by the endothelium, leading directly to AGE-modified LDL transport into the vessel wall. This is consistent with recent evidence obtained from human carotid artery atheromatous lesions, where serum LDL-AGE levels and histologically observed AGE immunoreactivity in vessel walls appeared to correlate with the severity of lesion (Fig. 3).9 AGE localization in sections of human carotid artery from non-diabetic patients. (A) Normal artery showing AGE immunoreactivity within the extracellular matrix of the vessel wall. (B) IgG control showing only autofluorescence. (C) Smooth muscle cells within a fatty streak from an atheromatous carotid artery showing AGE immunoreactivity. These cells also show extensive lipid accumulation. (D) IgG control of section (C) showing only autofluorescence of intracellular lipid droplets. It was long speculated that a cellular receptor-based system may exist as part of a mechanism whereby AGEs could be removed from tissues and sera in vivo thereby limiting their potentially toxic and harmful effects. Furthermore, such a system could contribute to dysfunctional responses of cells and tissues in pathological situations, such as atherosclerosis, hyperlipidemia, or diabetes. Work in this laboratory demonstrated that AGE proteins formed in vivo or synthesized in vitro are recognized by macrophage AGE-receptors, which are distinct from previously described scavenger receptors.25,26 Two AGE-binding proteins isolated from rat liver membranes were initially identified.27 The first, now referred to as AGE-R1, was shown to be homologous to a ∼50 kD component of the oligosaccharyltransferase complex (OST-48), while the 80 kD component, termed AGE-R2, is homologous to the protein kinase substrate 80K-H.28,29 Both these proteins are expressed on the surface of many human cells, including endothelial cells, monocytes/macrophages, T-lymphocytes, neural tissue, renal cells, and smooth muscle cells.22 While AGE-R1 likely serves as both a cell-surface and intracellular AGE-binding and transporting protein, AGE-R2 in its capacity as a substrate for protein kinase C may play a role in receptor-mediated intracellular signalling—leading to activation and cytokine and growth factor secretion associated with AGE-receptor binding.30,31 Significantly, both AGE-R1 and AGE-R2 are expressed in human atherosclerotic plaques in a distribution pattern similar to intracellular AGE accumulation (Fig. 4). Consistent with this, endocytic uptake of AGEs has been described in the endothelium,23 smooth muscle cells, and macrophage/monocytes in vitro, and their expression in vivo implies an inherent capacity of these cells for binding and endocytosing of AGEs.23 It also suggests that the process begins early and persists during atheromatous plaque formation. Also, AGE-LDL in serum, although unrecognized by the native LDL receptor, is likely to be quickly scavenged by macrophages, while endothelial cells have the potential to endocytose and pass active AGE macromolecules through to the vessel wall.23 Macrophages are obligate phagocytes, but even normally quiescent smooth muscle cells are known to be phenotypically transformed to phagocytose thus-modified LDL. It follows that this may be an important causal factor for AGE/lipid accumulation in fatty streaks and foam cells. AGE-receptor localization in atheromatous carotid arteries from non-diabetic patients. (A) Area of intimal thickening showing anti-AGE-R1 staining localized to the endothelium and also invading monocytes. (B) Section of similar area to (A) showing AGE-R2 localization. (C) Arteriole within carotid artery wall showing AGE-R1 localization to smooth muscle cells and monocytes/macrophages. (D) AGE-R2 localization, showing an identical pattern to AGE-R1 (C). Another 32 kD AGE-recognizing protein has recently been described with putative links to the AGE receptor complex. This protein, known as Galectin-3, Mac-2, or carbohydrate binding protein-35 (CBP-35), exhibits high-affinity binding for AGE ligands (Kd 3.5 × 107 M−1)32 and is present in most cells, though on the cell surface only after activation. AGE binding to Galectin-3 (in the present context referred to as AGE-R3) occurs within the 18 kD C-terminal peptide and promotes high molecular weight complex formation with the AGE ligand and with other membrane-associated receptor molecules on the cell surface. This complex formation is likely to involve the attack on a thiol ester of AGE-R3 by nucleophilic groups present on AGE proteins. The significance of the participation of high energy thiol ester bonds may reflect the need for efficient attachment and subsequent trafficking to intracellular degradative pathways of a massive and heterogeneous class of AGE-modified structures. Additional cell surface binding proteins for AGE determinants have been identified and cloned from various tissues and cells, including a 35 kD member of the immunoglobulin superfamily of receptors, named RAGE (also shown to serve as a receptor for amphoterin33), an 80 kD protein homologous to lactoferrin,34,35 and lysozyme.36 Based on protein sequence, ligand specificity, and functional differences between the AGE-binding polypeptides described above, it is now believed that additional AGE-binding proteins, with or without “receptor” characteristics may exist, which may contribute to AGE removal, transport, and/or processing as parts of a well-conserved mechanism that may also serve in diverse cellular functions. AGEs form irreversible crosslinks with long-lived proteins and have been shown to accumulate on tissues with increasing age and at an accelerated rate in diabetes.37,38 AGEs have also been demonstrated morphologically in the extracellular matrix (ECM) of vessels walls and atherosclerotic lesions using AGE-specific antibodies.5,8,9 Functionally, AGE crosslinks in the ECM are known to cause tissue rigidity, reduced solubility, and a decrease in the susceptibility of proteins to enzymatic digestion. Moreover, AGEs can act as trapping sites for a variety of soluble proteins, notably immunoglobulins39 and LDL species within the walls of large or small vessel walls.40 Also of significance to atherosclerosis, it has been the observation that AGE adducts, abundant in the vessel wall, can chemically interfere with the vasodilatory action of endothelium-derived nitric oxide (NO.), reacting directly to inactivate its bioactivity.41 These important AGE modulatory effects on NO. have been further confirmed in diabetic and aging animals.41,42 Furthermore, the markedly abnormal vasodilatory responses induced by the exogenous administration of AGE-modified proteins to healthy animals were inhibited by aminoguanidine.42 In addition, it has been shown that the NO-mediated antiproliferative effects are inhibited in vitro by AGEs.43 One leading hypothesis of fatty streak formation describes early endothelial dysfunction, one manifestation of which is the increased accumulation of lipid within the vessel wall. Altered endothelial function is also associated with increased adherence/migration of monocytes/macrophages and T-lymphocytes into the intima and initiation of a prolonged sub-inflammatory response. The receptor-mediated interaction of native AGEs with the endothelium may be of critical importance in initiation of atherosclerosis. Using cultured endothelial cells, it was shown that AGEs promote transendothelial migration of human monocytes,44 increase vascular permeability,23 and cause downregulation of the surface anticoagulant thrombomodulin.23 We then reported an increased vascular permeability in AGE-infused rats,42,45 and subsequently a marked acceleration of atherogenesis in AGE-infused rabbits, associated with a significant increase in expression of vascular cell adhesion molecule-1 (VCAM-1) and intercellular adhesion molecule-1 (ICAM-1) on the endothelial surface of the aorta followed by mononuclear cell adhesion and subendothelial infiltration.7 These AGE-induced changes were significantly accelerated in these otherwise normal animals placed briefly (two weeks) on a cholesterol-rich diet, manifested by multifocal atheromas containing foam cells and extensive lipid deposition at lesion sites. Taken together, these findings indicate that AGEs actually cause significant dysfunctional changes to the vascular endothelium in ways that render the vessel wall susceptible to atherosclerosis lesion formation. A pathological imbalance in cytokines and growth factors has long been recognized in atherosclerosis.46 Here too, AGE binding to the AGE receptor has been shown to elicit cytokine-producing responses from monocytes, macrophages, T-lymphocytes, and fibroblasts, which in non-pathological systems may be involved in normal tissue remodelling.30,31 Of relevance to atherosclerosis, TNFα and interleukin-1β (IL-1β) are released by AGE-stimulated monocytes,30 which may promote localized cellular recruitment and further inflammatory responses. In addition, the concomitant increase in AGE receptor-mediated expression of platelet-derived growth factor (PDGF)31 and insulin-like growth factor-IA (IGF-IA)31 by monocytes has implications for the smooth muscle proliferative response characteristic of atheroma formation. It is of particular interest that resting human and rat CD4+ and CD8+ T-cells were found to constitutively express such receptors and to bind AGE proteins, while PHA-pre-stimulation of T-cells resulted in AGE receptor upregulation and IFNγ synthesis and release.47 This suggests that AGEs in cooperation with primed tissue macrophages can also activate IFNγ-producing T-cells, which may contribute to tissue damage. This is consistent with evidence that activated IFNγ-expressing T-lymphocytes found in atherosclerotic lesions may be important participants in the chronic inflammatory response underlying atherogenesis.42 Amadori product formation represents an important step in advanced glycosylation chemistry because progression from this point to protein-to-protein crosslinking requires multiple rearrangement reactions to create intermediates that can react with amino groups to form irreversible AGEs. An important pharmacological strategy for inhibition of this process utilizes the nucleophilic compound aminoguanidine, which was shown to be a potent and specific inhibitor of glucose-mediated crosslinking and tissue damage in vivo.49 The terminal amino group of aminoguanidine, by virtue of its low pKa, reacts specifically with glucose-derived reactive intermediates and prevents protein-protein or protein-lipid crosslinks from forming. This mechanism of action of aminoguanidine has now been confirmed in a number of studies, in which aminoguanidine was shown to prevent diabetes-related vascular complications in experimental animals.18,50–55 From these studies, it is apparent that aminoguanidine can be used to prevent advanced glycosylation as well as the AGE-mediated tissue damage in animal models of diabetes and aging. In addition, advanced glycation-modified hemoglobin (AGE-Hb) was monitored in diabetic subjects receiving aminoguanidine. AGE-Hb was found to be significantly reduced in the treated patients as a result of aminoguanidine therapy. However, HbAlc was not reduced. This pointed to the specificity of aminoguanidine for inhibiting post-Amadori, advanced glycosylation reactions. Of direct relevance to atherosclerosis, aminoguanidine therapy caused a 28% decrease in LDL-cholesterol, a 19% decrease in total cholesterol, and a 19% decrease in triglycerides in the same group of patients, all of which diabetic and advanced glycosylation may eventually in diabetics or in at risk of vascular of aminoguanidine is now at In recent studies it has been that high levels of circulating AGE-LDL may have a for AGE deposition within their vessels and may be at increased risk of atheroma levels of circulating AGEs, especially in non-diabetic patients, may reflect dysfunctional LDL mechanisms. Therefore it may in the to early for serum AGEs. an may early to patients and early and/or links between reactive AGEs and atherosclerotic lesion have over recent not only in diabetic but also in non-diabetic Work in this laboratory has now confirmed the toxic nature of AGEs and the of AGE receptors, to their role in the and progression of lesion formation in human atherosclerosis, and ways for potential of atherosclerotic the of early of serum AGEs in susceptible and novel the potential to control advanced glycation interactions with the vessel wall now group that is a member of the AGE receptor the endocytic uptake of AGE is by Galectin-3 or exhibits high-affinity binding for AGEs. likely it to the efficient recognition of a large heterogeneous class of AGE by the AGE receptor it is of AGEs is by this molecule The expression of this molecule in the levels is and it is during cellular activation. When this is induced by AGEs, large of this protein and the as well as other receptor on the surface. This for a period of a after which the from the surface and are studies of most cellular with other AGE-binding proteins, express our studies were on using macrophages membrane-associated or We also and AGEs macrophages into foam Although we have not this evidence suggest that AGE-LDL can be scavenger or AGE-receptor pathways by the foam atheromatous lesions significant of AGEs with intracellular within foam specific studies to this are not is AGE-LDL AGE-LDL has not been directly to the by which this molecule is modified and the of that are involved with glycation from oxidation in such studies. We have as is during The of a similar to that of native of AGE-modified is by the liver system but also by smooth muscle vascular endothelium, and and the products are normally in the It has recently been shown that a significant of the are reactive to to other proteins and cellular components AGEs, or AGE and for a the of AGEs in atheromatous Using such a could AGEs in the we have recently an AGE-specific which can be used to circulating we have reported and as demonstrated it be in the to various serum or tissue by most the as of possible tissue vascular or that AGEs are attached to Although the AGE-LDL apoB and lipid are well we have not other as This is a important of LDL is AGE We can only at this point that not of total LDL is This is to was observed on AGE-modified have found many AGEs in atheromatous lesions, which that AGEs from the and to LDL in the AGEs as well as LDL is from the to the tissues and then to the vascular it is likely while in it can be AGE-modified in by our recent studies, patients with display high levels of through the of reactive to LDL. to LDL is not the only for circulating reactive though it is most in the context of suggest that other proteins, such as or and are modified by reactive in with renal This evidence has to the that most of the AGE modification of the proteins not through the action of but through reactive AGEs that are released from tissues into the to be in the in to be the for a large of AGE not a important that has recently to our and that is the significant by AGEs. Based on the of it apparent that reactive AGEs from are into the where they their reactive a of these AGEs are normally in the and only of that are by it that AGEs bind to the apoB or other We not have a direct to this identified or AGEs in small vessel and and from the that we have which to the of serum and tissue AGEs in patients as multiple evidence from animals with experimental diabetes and from normal animals treated with exogenous AGEs. The were shown to renal and vascular groups have reported the effects of the AGE inhibitor, aminoguanidine, in a number of experimental animal models of diabetes with and small vessel renal and that AGEs play a role in small vessel and studies at these are in