THE MAILLARD REACTION – A JOURNEY FROM THE DISCOVERY OF ADVANCED GLYCATION ENDPRODUCTS TO THEIR CHIEF CELLULAR RECEPTOR, RAGE: A MECHANISM UNDERLYING DIABETIC COMPLICATIONS AND THE INFLAMMATORY RESPONSE

R, Ramasamy, S.F. Yan and A.M. Schmidt

1 INTRODUCTION

An emerging theme in the biology of chronic disease is that modifications of proteins may play major roles in re-routing biological reactions from preservation of homeostasis to induction of tissue injury. Atop the list of potential modifications in diabetes that may significantly alter cellular properties is the process of glycation. From Maillard's discovery that heating proteins particularly in the presence of sugars resulted in their "browning," was born the field of AGEs. In recent years, the identification of the diverse settings in which AGEs may form has greatly expanded our understanding of how proteins subjected to fundamental posttranslational modifications of proteins may experience both loss- and gain-of function properties.

The products of non-enzymatic glycation and oxidation of proteins, the advanced glycation end-products (AGEs), may be formed by multiple means, such as hyperglycaemia, aging, inflammation, oxidation (such as of oxidized low density lipoprotein, oxLDL), hypoxia or ischaemia/reperfusion (I/R) injury, or by specific fonns of food preparation. AGEs may exert diverse effects in biological systems; the discovery that the Receptor for AGE (RAGE) was a signal transduction receptor for this class of modified species suggested specific means by which modified proteins gain functions such as the ability to modulate transcription, translation and activity of proteins in biological systems. AGEs are a heterogeneous class of substances and therefore a particular effort is to determine which specific form(s) of AGEs may interact with RAGE.

In this review, we discuss the growing list of settings in which AGEs may form and interact with RAGE. Insights into roles for AGE-RAGE action in prototypical complications of diabetes and inflammation are also considered. Taken together, a growing body of evidence suggests that antagonism of RAGE may be a logical target for therapeutic intervention in the chronic diseases in which AGEs accumulate, such as diabetes, aging and tissue-damaging inflammatory responses.

2.1 Different Settings for AGE Production and Accumulation

The formation of AGEs, although accelerated in hyperglycaemia, also occurs over extended periods of time even in essentially euglycaemic states, such as in natural aging. Proteins modified in the presence of reducing sugars such as glucose, glucose-6-phosphate or ribose resulted in species that interact with RAGE. More recently, specific AGEs such as carboxy-methyl lysine (CML) AGEs have been shown by numerous groups to interact with RAGE.

Oxidized LDL (ox LDL) species contain AGE epitopes and evidence that they stimulate signal transduction via RAGE emerged from experiments in which incubation of primary wild-type murine aortic endothelial cells with oxLDL activated signal transduction pathways, but in contrast, exposure of RAGE null endothelial cells to oxLDL failed to stimulate signalling or evoke such changes in gene expression. In addition, further evidence that AGE epitopes were responsible for changes in endothelial gene expression was suggested by studies in which pre-treatment of the endothelial monolayers with anti-AGE IgG prior to exposure to oxLDL suppressed signal transduction. Of note, a recently identified ligand family of RAGE, the advanced oxidation protein products (AOPPs), also activate RAGE, but not via AGE epitopes, as anti-AGE antibodies had no impact on AOPP-stimulated RAGE signalling.

Recent experiments highlighted that exposure of wild-type primary murine aortic endothelial cells to hypoxia generated AGE-immunoreactive epitopes, as assessed by ELISA experiments using anti-AGE IgG. In vitro or in vivo, pre-treatment of endothelial cells or wild-type mice prior to hypoxia with either anti-AGE IgG or aminoguanidine, an inhibitor of AGE formation, suppressed hypoxia-mediated activation of Protein kinase C, phosphorylation of mitogen activated protein kinases (MAP kinases), and up-regulation of egr-1 transcripts and protein. These actions occurred via RAGE, as mice and endothelial cells devoid of RAGE failed to upregulate Egr-1 in hypoxia versus wild-type mice and primary endothelial cells.

One consequence of inflammatory conditions is the generation of AG Es. At least in part via the action of the myeloperoxidase pathway, or via generation of reactive oxygen species (ROS), AGE formation may occur in non-diabetic wounds or atherosclerosis, for example. Of note, increased ROS may support AGE formation via depletion of anti-oxidant defenses. In the presence of decreased levels of glutathione, the enzyme glyoxalase-1 is less efficient, thereby reducing detoxification of the pre-AGE methylglyoxal (MG0).

A burgeoning literature suggests that specific forms of food preparation, such as in heating procedures, may lead to excessive amounts of AGEs in the diet. In animal models, low AGE diets were associated with less cellular perturbation and decreased manifestations of vascular disease compared to high AGE diets. A recent study suggested that a food-derived AGE, pronyl glycine, was a novel ligand for RAGE. Treatment of Caco-2 intestinal epithelial cells with pronyl glycine AGE activated MAP kinase signalling via RAGE. Taken together, these considerations indicate that stimuli to AGE generation are diverse and not limited to hyperglycaemic states. Even in euglycaemia, such factors as aging, hyperlipidaemia, inflammation, oxidative stress and modifications of food ingredients by preparation techniques may produce AGE species. Although the full range of these species capable of binding RAGE is likely yet to be fully elucidated, heterogeneous AGE mixtures, CML-AGEs and prolyl glycine, for example, may stimulate signalling mechanisms via RAGE. Apart from distinct AGEs, however, the repertoire of RAGE ligands is more broad and includes ligands specifically linked to inflammation and cellular migration.

2.2 Multi-ligand Nature of RAGE

At least four other classes of RAGE ligands have been identified. S100/calgranulins are members of a family of molecules which have distinct intracellular functions, such as calcium binding. In the extracellular space, multiple S100/calgranulins may bind RAGE, such as S100A12, S1008, S100P, S100A6, and S100A4. Release of S100/calgranulins to the extracellular space, we predict, is a key control mechanism by which activated, dead or injured cells may release these molecules. Although initially most probably intended as a cellular defense strategy, if sustained, the release and paracrine/autocrine action of these molecules may intensify inflammation and tissue damage, thereby facilitating chronic disease.

High mobility group box-1 (HMGB-1) ligands also bind to RAGE. Although these molecules may also bind certain toll receptors (TLRs), such as TLRs 2 and 4, HMGB I acts on RAGE-expressing cells to stimulate signalling. HMGB1 is a member of the non-histone DNA binding family of molecules normally present in the nucleus. Akin to S100/calgranulins, HMGBI molecules may be released by necrosis, thereby freeing them to act in the extracellular space. Once available in the extracellular space, these molecules may act on a diverse array of inflammatory cells, vascular cells, cardiomyocytes, and neurons as examples, thereby propagating inflammatory mechanisms, at least in part via cell surface receptors such as RAGE. In the context of inflammation, Mac-1 has been shown to be a ligand for RAGE. Further, amyloid-β peptide and β-sheet fibrils in general interact with RAGE. As amyloid often forms and accumulates in response to chronic inflammation, it is plausible that this class of ligands represents yet an additional means by which RAGE signalling may be amplified and sustained in chronic inflammation. In the section to follow, we review the evidence that RAGE signalling importantly contributes to the pathogenesis of the inflammatory response.

2.3 RAGE and the Inflammatory Response – its Natural Function?

We and others had long speculated that AGEs were not the only ligands of RAGE and indeed, the discovery that S100/calgranulins in particular were RAGE ligands evoked the initial suggestion that AGE was involved in inflammatory reponses. Work of Tracey and colleagues implicated HMGB1 in late responses to sepsis; hence, the finding that RAGE transduced the signals of these classes of ligands firmly linked RAGE to inflammation. RAGE is expressed on monocytes/macrophages and ligand-RAGE interaction contributes to their biology, including migration, activation and delayed apoptosis in the face of ligand challenge. RAGE is also expressed on T and B lymphocytes, and dendritic cells, thereby suggesting important roles for ligand-RAGE interaction in adaptive immune responses.

This concept was tested in a murine model of orthotopic allogeneic heart transplantation. Using allo-mismatched donors and recipients, we found that administration of sRAGE prolonged allograft survival in a dose-dependent manner compared to vehicle treatment. Histology of the grafts revealed a marked reduction in T lymphocyte and macrophage influx into the tissue in the presence of sRAGE. To precisely address the specific RAGE-dependent mechanisms in T lymphocytes, OTII T lymphocytes (CD-4 like T lymphocytes expressing T cell receptors recognizing ovalbumin) were employed. OTII mice were bred into the RAGE null background. In both in vitro and in vivo studies, RAGE deficient OTII cells displayed significantly less proliferation and cytokine production in response to ovalbumin vs. RAGE-expressing OTII controls. We concluded from those experiments that RAGE was required for effective T lymphocyte priming responses. Further studies using an islet allograft model suggested that RAGE activation in T cells may be a key factor in the early events linked to Th1 differentiation. Thus, RAGE may have multiple roles in diabetic complications due to its actions in distinct steps in these processes. From transducing the effects of glycated ligands, to amplifying inflammatory responses, RAGE may exert significant impact in diabetes complications. In the sections to follow, we consider roles for RAGE in prototypic complications of diabetes and how AGEs and inflammatory mechanisms may contribute.

2.4 RAGE and Diabetic Nephropathy

Diabetes is a major cause of end-stage renal failure. Multiple publications in both animal models and human diabetes have shown accumulation of broad classes of AG Es in the diabetic kidney. Hence, to test the role of RAGE in the pathogenesis of diabetes-associated nephropathy, we first established that in both human and rodent diabetes, RAGE was principally expressed in glomerular epithelial cells (podocytes) and in glomerular endothelial cells. By immunohistochemistry of in situ tissue preparations of renal cortex, there was no evidence that RAGE was expressed to any appreciable degrees in tubular cells in either the diabetic or non-diabetic state. Given the emerging central role for the podocyte in the earliest initiating mechanisms in diabetic nephropathy, it was logical to test the premise that RAGE contributed to the pathogenesis of diabetes-associated nephropathy.

The first studies to address this concept employed pharmacological antagonists of RAGE, such as soluble RAGE. Soluble RAGE, composed of the extracellular ligand-binding domain of RAGE, binds ligands and blocks their interaction with cell surface receptors. Soluble RAGE is generated in a baculovirus expression system and purified to homogeneity and rendered free of any contaminating lipopolysaccharide. Soluble RAGE was administered to db/db mice from the onset of hyperglycaemia through at least six months. Administration of sRAGE suppressed indices of mesangial expansion, thickening of the glomerular basement membrane and microalbuminuria. In other studies, distinct pharmacological antagonists of ligand-RAGE interaction were tested, such as neutralizing anti-RAGE antibodies. In mouse models of type I and 2 diabetes, anti-RAGE antibodies beneficially impacted the course of renal disease.

More recently, genetic approaches have been employed to test the role of RAGE in diabetes-associated nephropathy. Myint and colleagues studied homozygous RAGE null mice and found that compared to wild-type animals, RAGE deficient animals displayed suppression of the following major endpoints such as kidney enlargement, increased glomerular cell number, mesangial expansion, advanced glomerulosclerosis, increased albuminuria and increased serum creatinine.

As a definite pitfall of most available mouse models of diabetes is the failure to progress to highly advanced and more human-like stages of disease, Inagi and colleagues created transgenic mice that over-expressed megsin (a glomerular-specific serpin), RAGE and iNOS to develop a model system predicted to develop severe nephropathic changes. These authors found that compared with the single- or two- gene transgenic mice, the triple transgenic mice (over-expression of RAGE) developed severe albuminuria, glomerular hypertrophy, diffuse mesangial expansion, inflammatory cell infiltration and interstitial fibrosis, with 30-40% of the glomeruli displaying nodule-like lesions at age 16 weeks of age, a point that was relatively early in terms of other available mouse models. This study reinforced that RAGE appeared to be one of the "multiple hits" responsible for advanced diabetes-associated nephropathy, at least in mouse models.