AGEs mediate their tissue effects through three main mechanisms: (1) Cross linking extracellular (matrix) proteins thereby affecting tissue mechanical properties; (2) Cross linking intracellular proteins thus altering their physiological functions[5,6]; and (3) Binding to their cell surface receptor RAGE to inducing multiple intracellular signalling cascades. In the following sections these mechanisms will be discussed.
Crosslinking tissue proteins
Cross linking extracellular matrix proteins: AGE for-mation is a process of chronicity usually affecting long-lived proteins. Extracellular matrix proteins especially collagen type IV that is involved in basement membrane structure, are more prone to advanced glycation due to their long turnover[20,78]. Advanced glycation and crosslinking of other extracellular matrix proteins, e.g., collagen I and elastin render them stiffer and less susceptible to proteolytic digestion (Figure 2). This may contribute to the observed increase in vascular stiffness reported in diabetes and old age[20,51,79]. In addition, cross linking myocardial collagen with AGEs has been suggested to cause myocardial stiffness and diastolic dysfunction in diabetic patients[51,80].
Figure 2 Effects of advanced glycation end products on extracellular matrix proteins.
In extracellular matrix, advanced glycation end products (AGEs) form on different molecules as collagen, laminin and elastin. This alters the physiological properties of the matrix and increases its stiffness. AGEs upregulate transforming growth factor (TGF)-β that increases the production of extracellular matrix components by binding to its receptor.
AGEs alter the structure of low density lipoproteins (LDL) through glycation therefore preventing their clearance from the circulation via the normal elimination route, i.e., uptake by the endothelial cells. Instead they are uptaken by the blood monocytes leading to foam cells that contribute to the pathogenesis of atherosclerosis[81,82].
Cross linking intracellular proteins: AGEs have been shown to be implicated in crosslinking of intracellular proteins and hence altering their physiological properties and functions. For instance AGEs were shown to cross link the domains of both the Ryanodine receptor and SERCA2a in cardiomyocytes leading to alterations in calcium homeostasis reported in diabetic cardiomyopathy[83,84].
RAGE dependant effects of AGEs
RAGE receptor: RAGE belongs to the immunoglobulin superfamily of receptors. It has been reported that RAGE gene is located on chromosome 6 in humans between genes coding for class II and class III major histocompatibility complexes. The RAGE promoter has been shown to possess nuclear factor kappa B (NF-κB) binding sites, hence linking RAGE expression to the inflammatory cascade.
RAGE is a multiligand receptor with advanced glycation end products being identified as its first known ligands. Hereafter, multiple other RAGE ligands have been revealed including; high mobility group protein box-1, some members of the S100 protein family amyloid β[88,89] and fibrillar protein aggregates[90,91]. Therefore, RAGE has an important role in the pathogenesis of the diseases induced by such ligands, e.g., inflammation, tumours, neurodegeneration and amyloidoses[21,88,92,93].
RAGE structure: Full length RAGE comprises three domains, an extracellular domain of 332-amino acids arranged as a single “V”-type immunoglobulin-like (variable) domain with subsequent two “C”-type (constant) domains[89,94]. Modern biochemical techniques have revealed evidence suggesting that both the V and C1 domains of RAGE function together as an incorporated single structural unit for the binding of some ligands. In contrast, C2 RAGE domain is suggested to function completely independent from the VC1 complex while remaining attached to it through a flexible hinge (Figure 3). It has been deduced from experimental work that different RAGE ligands interact with one or more of its domains.
Figure 3 Receptor for advanced glycation end product (i.
e., colon) receptor structure and its functional implication in binding different advanced glycation end products. A diagrammatic illustration of the structure of the receptor for advanced glycation end product (RAGE) showing that it is composed of an extracelluar portion, a transmembrane portion and an intracytoplasmic tail. The extracelluar portion comprises three domain V, C1 and C2. The first two are believed to work together as a single functional complex (VC1) whereas the C2 domain remains attached to the VC1 complex but works independently from it. The diagram also illustrates how multiple RAGE receptors polymerise within the cell membrane to facilitate high affinity binding of the positively charged V domain with the negatively charged advanced glycation end products (AGEs) independent of their chemical structure. That is why RAGE is considered one of the pattern recognition receptors.
RAGE has a single transmembrane domain and a highly charged cytosolic tail formed of 43 amino acids (Figure 3). The cytosolic tail is vital for RAGE ligands to activate intracellular signalling cascades. RAGE isoforms in which the cytosolic tail was absent, bind AGEs but fail to elicit intracellular signaling on ligand binding. It has recently been shown that for some RAGE actions, its tail binds directly to a cytoplasmic molecule known as the mammalian diaphanous-1 which is essential for eliciting phosphorylation/activation steps needed for initiation of the signalling cascade.
Despite the wide variation in their chemical structure, AGEs bind to the V domain of the RAGE receptor. The fact that RAGE recognizes a class of biochemically heterogeneous ligands such as AGEs categorizes RAGE as one of the pattern recognition receptors[97-99] that identifies common features or patterns rather than a specific ligand. Though AGEs exhibit diverse chemical structure, however, they have some common general characteristics. First, all AGE modified proteins demonstrate a net negative charge that accumulates during their formation by glycation and oxidation[100,101]. The second main feature is that modifications of proteins by AGEs lead to creation of multiple covalent cross-links resulting in higher molecular mass molecules (multimers). This ligand geometry is thought to be important for RAGE activation.
Recent molecular structure revealing technologies including X-ray crystallographic, NMR together with conventional biochemical data have illustrated that the unusual ability of the RAGE receptor to bind different AGEs lies in its extracellular portion (VC1 ectodomain) where it has been suggested that the ligand binding is triggered mainly by electrostatic interactions between the positively charged surface of this subunit and negatively charged ligands (Figure 3). The V domain has been shown by NMR to exhibit three distinct areas for mediating AGE-V domain interactions. Such areas are situated in the positively charged regions of the V domain. The first interaction surface includes strand C and loop CC’, the second interaction surface comprises strand C’, strand F and loop FG, and the third interaction one consists of strand A’ and loop EF.
The second main point to be considered here is that studies based on using a fluorescence-labelled receptor demonstrated that RAGE does not float as one molecule in the plasma membrane but instead multiple RAGE receptors aggregate to form receptor assemblies[19,101]. Therefore, RAGE exists in constitutive multimers that usually includes four molecules or more within the plasma membrane[101,103]. RAGE multimers are thought to display a parallel orientation with VC1 subunits exhibiting side to side contacts (Figure 3) as revealed in protein crystals. Therefore it has been speculated that the general multimeric structure adopted by AGE modified proteins together with numerous AGE-modified side chains are essential requirements for preserving the receptor assemblies’ stability needed for activation of the receptor probably in a similar way to that previously described for the receptor tyrosine kinases where the intracellular domains of the receptor that possess intrinsic kinase activity must come very close together in a particular orientation that facilitates cross-phosphorylation of the domains and hence commencement of the signalling cascade.
RAGE isoforms: Despite the presence of a single gene coding for RAGE, there are several splice variants of this gene with three main RAGE isoforms having been identified: full length RAGE, dominant-negative RAGE (DN-RAGE) and endogenous soluble RAGE (es-RAGE). Full length RAGE is composed of the whole three domains; extracellular, trans-membrane and intracellular domains. DN-RAGE has extracellular and transmembrane domains, but no cytosolic tail. Endogenous soluble RAGE possesses the extracellular domain only so it is found free in the circulation. The only RAGE isoform that is capable of eliciting intracellular signalling upon interacting with its ligands is the full length RAGE as this is induced through its cytosolic tail. However, the other two isoforms; DN-RAGE and esRAGE help in clearance and neutralisation of circulating AGEs by competing with full length RAGE in binding them[56,105].
RAGE tissue distribution: RAGE expression has been detected in a number of cells including endothelial cells, smooth muscle cells, monocytes/macrophages, T lymphocytes, cardiomyocytes, glomerular podocytes, dendritic cells, neurons of the central and peripheral nervous systems and transformed cells. Generally, there is a low expression of RAGE in tissues. However, it becomes up-regulated in an environment rich with its ligands as in the case of diabetes or aging. RAGE expression was higher in endothelial cells, monocytes and smooth muscles in diabetic vascular tissue.