Glycation & Insulin Resistance: Novel Mechanisms and Unique Targets?

Abstract

Objectives

Multiple biochemical, metabolic and signal transduction pathways contribute to insulin resistance. In this review, we present the evidence that the post-translational process of protein glycation may play role in insulin resistance. The post-translational modifications, the advanced glycation endproducts (AGEs), are formed and accumulate by endogenous and exogenous mechanisms.

Methods and Results

AGEs may contribute to insulin resistance by a variety of mechanisms, including generation of tumor necrosis factor-alpha, direct modification of the insulin molecule thereby leading to its impaired action, generation of oxidative stress, and impairment of mitochondrial function, as examples. AGEs may stimulate signal transduction via engagement of cellular receptors, such as RAGE, or receptor for AGE. AGE-RAGE interaction perpetuates AGE formation and cellular stress via induction of inflammation, oxidative stress and reduction in the expression and activity of the enzyme, glyoxalase I that detoxifies the AGE precursor, methylglyoxal, or MG.

Conclusions

Once set in motion, glycation-promoting mechanisms may stimulate ongoing AGE production and target tissue stresses that reduce insulin responsiveness. Strategies to limit AGE accumulation and action may contribute to prevention of insulin resistance and its consequences.

Glycation: Formation and Detoxification

The process of nonenzymatic glycation and oxidation of proteins yields a diverse array of modified products. AGEs are a heterogeneous group of compounds formed via endogenous and exogenous mechanisms. Increase levels of glucose drive formation of Schiff bases; Schiff bases subsequently rearrange to form the more stable Amadori products. These reactions are reversible, and the levels of Amadori products are directly related to levels of glucose. Once Amadori products form, further oxidative modifications of these molecules may occur, resulting in the formation of the irreversible advanced glycation endproducts, or AGEs. In settings in which oxidative processes are involved in the formation of these AGEs, so-called “glycoxidative” species such as NΣ-(carboxymethyl)lysine (CML) and pentosidine form^1–2.

Glycolytic intermediates are intracellular sources of AGEs. In these more rapid Maillard-type reactions, three classes of dicarbonyl compounds may form, which are potent glycating agents, including glyoxal, methylglyoxal (MG) and 3-deoxyglucosone. Among these, the most potent glycating agent is likely MG. MG forms via the conversion of glyceraldehyde-3-phosphate and dihydroxyacetone phosphate and principally reacts with arginine groups to form such AGEs as 5-hydro-5-methylimidazolone and the fluorescent AGE, argpyrimidine³. MG levels can be regulated via the activity of two enzymes, glyoxalase I and II; these enzymes “detoxify” MG and result in MG conversion to D-lactate⁴. When the level of activity of glyoxalases is impaired, MG levels rise, thereby favoring further AGE production⁵.

In addition to glucose or glycolytic intermediate derived AGEs, lipid peroxidation may result in the formation of reactive carbonyl compounds capable of interacting with proteins, thereby leading to the production of advanced lipoxidation endproducts, or ALEs. Of note, NΣ-(carboxymethyl)lysine (CML) is one of the ALEs; CML-modified adducts may form both by glycoxidation and lipid peroxidation pathways. Furthermore, when MG reacts with lysine residues of proteins, a homologue of CML-AGE may be produced, known as NΣ-(carboxyethyl)lysine, or CEL⁶. CML-AGE may also be formed by inflammatory mechanisms driven by the myeloperoxidase pathway⁷. Also, natural aging and renal failure^8–10, in the absence of other risk factors such as diabetes, are also associated with endogenous AGE formation and accumulation.

Exogenous sources of AGEs may contribute to AGE formation in vivo. Foods high in fat and those cooked at high temperatures contain AGEs, the most prominent of which is CML-AGE, and the AGE precursor MG^11–12. Furthermore, air pollution-associated fly ash was shown to lead to the generation of AGEs on proteins in fibroblasts exposed to this agent for long periods of time¹³. Hence, dietary and environmental sources of AGEs may represent additional key depots for formation of these products. When detoxification mechanisms (such as glyoxalases) or clearance mechanisms, such as in advanced renal disease are suppressed, AGE levels and their consequences may be amplified.

Glycation Products & Receptors & mechanisms to sustain and enhance AGE levels

One of the chief mechanisms by which AGEs exert their pathogenic effects is via interaction with cell surface receptors. There are multiple receptors for AGEs such as receptor for AGE (RAGE)¹⁴, CD36¹⁵, Lox-1¹⁶, macrophage scavenger receptors¹⁷ or AGERs -1-2-3¹⁸. RAGE is a signal transduction receptor for AGEs¹⁴. These receptors may contribute to AGE formation and accumulation. For example, when AGEs bind to RAGE, signal transduction mechanisms are instigated in a process requiring the cytoplasmic domain of the receptor. A chief outcome of AGE-RAGE signaling is the generation of oxidative stress, largely through the NADPH oxidase system¹⁹. As oxidative stress facilitates AGE formation, one consequence of the interaction of AGE with RAGE is generation of an environment that favors further AGE production. Thus, it was not surprising that in mice devoid of RAGE, levels of oxidative stress and AGEs were lower compared to wild-type RAGE-expressing animals²⁰. It has been shown that in the kidney of diabetic OVE26 mice devoid of RAGE, levels of glyoxalase I mRNA and protein were higher compared to those of RAGE-expressing OVE26 mice. In parallel, levels of MG and AGEs were lower in the RAGE deficient mice despite equal degrees of hyperglycemia and glycosylated hemoglobin levels²⁰.

The potential roles for RAGE in AGE generation have a further layer of complexity, as RAGE is not solely a receptor for AGEs, but a multi-ligand member of the immunoglobulin superfamily. Thus, beyond AGEs, RAGE also transduces the signals emitted by multiple members of the pro-inflammatory S100/calgranulin family^21–23, and high mobility group box 1 (HMGB1)²⁴. These molecules, at least in part via RAGE, upregulate cytokines, matrix metalloproteinases (MMPs) and gene expression programs in monocytes/macrophages and lymphocytes that amplify tissue-damaging inflammation in settings such as autoimmunity and extensive cellular/tissue stress²⁵. Inflammation itself may generate oxidative stress; for example, one key consequence of activation of the myeloperoxidase pathway is the generation of CML-AGEs⁷. Hence, ligand-RAGE may trigger oxidative stress and inflammation; once S100/calgranulin- and HMGB1-bearing inflammatory cells are recruited to sites of tissue stress, release of these non-AGE ligands may exacerbate oxidative stress and stimulate further AGE formation. Figure 1 summarizes the proposed hypotheses regarding triggers to glycation, consequences of glycation, and amplifying pathways further exacerbate generation and accumulation of AGEs.

Figure 1. Sources of AGEs, Interaction with Cellular Receptors and Perpetuation of Inflammation.

The advanced glycation endproducts (AGEs) may form via multiple exogenous and endogenous mechanisms as indicated in the figure. Exogenously, high fat- and high AGE diets may increase accumulation of AGEs; pollutants such as fly ash have been shown to cause glycation of fibroblast proteins (blue boxes). Endogenously, there are multiple mechanisms of AGE formation, such as transient bouts of elevated glucose and impaired glucose tolerance, via glycolytic intermediates (such as MG [methylglyoxal], glyoxal and 3-DG [3-deoxyglucosone]), in renal failure, natural aging and inflammation (green boxes). These processes may ultimately lead to the production and accumulation of AGEs (illustrated in purple); AGE interaction with cellular receptors such as RAGE, may have multiple consequences that sustain production of AGEs. For example, AGE-RAGE is a potent generator of reactive oxidant species (ROS), which exacerbate AGE formation. AGE-RAGE begets inflammation and in inflammatory milieus, release of distinct RAGE ligands, S100/calgranulins and HMGB1, by infiltrating inflammatory cells, lead to further production of AGEs via inflammation and oxidative stress pathways. Finally, RAGE action downregulates Glyoxalase (Glo) 1 (red inhibitory arrow); downregulation of Glo1 suppresses MG detoxification, and, therefore, leads to further MG-driven AGE production and accumulation (red stimulatory arrows). Hence, from initial AGE production to interaction with receptors such as RAGE, a reinforcement mechanism to sustain AGE generation may occur in settings of chronic cellular stress. [Abbreviations & Symbols: MG, methylglyoxal; 3-DG, 3-deoxyglucosone; AGEs, advanced glycation endproducts; Glo1, glyoxalase I].

Although the earliest studies on AGEs involved their mechanistic links to the pathogenesis of diabetic complications, recent intriguing studies have suggested that glycated products and AGEs may contribute to the pathogenesis of insulin resistance as well. In the section to follow, we will present a summary of data in human subjects suggesting that AGEs may be linked to insulin resistance.

Glycation and links to insulin resistance: studies in human subjects

A number of studies have reported interesting associations between AGE levels and insulin resistance, even in the absence of diabetes. Tan and colleagues measured AGE levels, associated inflammatory markers and insulin resistance (the latter by the homeostatic model assessment index or HOMA-IR) in 207 healthy subjects without diabetes. Serum levels of AGEs correlated in a statistically significant manner with HOMA-IR in both male and female subjects. Even after adjustment for age, gender, body mass index, waist circumference, cigarette smoking, adiponectin levels, and markers of oxidative stress and inflammation, AGE levels were an independent correlate of HOMA-IR on multiple regression analysis²⁶. Tahara and colleagues reported on 322 nondiabetic Japanese subjects; serum AGE levels correlated with HOMA-IR in these subjects and after multiple regression analysis, AGEs, along with waist circumference, glycosylated hemoglobin, and triglycerides, were correlated with the degree of insulin resistance. When age-adjusted HOMA-IR levels stratified by AGE tertiles were compared using statistical tests, the authors found trends in both male and female subjects²⁷. Sarkar and colleagues tested the potential association between levels of total carbonyl compounds in serum with HOMA-IR levels in type 2 diabetic subjects and found a highly significant correlation between the degree of insulin resistance and carbonyl levels, whereas levels of lipid peroxidation end products and thiobarbituric reactive substances (TBARS) were not significantly correlated with HOMA-IR, perhaps suggesting that pre-AGE compounds rather than the pathophysiological state of oxidative stress alone was required to impact insulin sensitivity²⁸. Diamanti-Kandarakis and coworkers studied AGE levels in a group of subjects with polycystic ovary syndrome (PCOS). Subjects with PCOS often display insulin resistance. Among the 193 subjects studied, 100 had PCOS and the remaining subjects were age- and body mass index-matched control individuals. Subjects with PCOS had significantly higher AGE levels versus subjects with isolated hyperandrogenaemia, anovulation, ultrasound diagnosed PCO and control subjects²⁹.

Taken together, these findings in human subjects suggest potential links of AGEs to insulin resistance. Evidence suggestive of mechanistic links between AGEs and insulin resistance was provided by the results of in vitro and in vivo experimentation. In the sections to follow, we discuss these findings and suggest that AGEs might be new therapeutic targets for mitigating insulin resistance.

Glycation and Insulin Resistance: In Vitro Evidence

Direct effects of modification of insulin

Recent work has suggested that glycation of insulin is a distinct possibility and that such glycation may change the properties of insulin, impacting on its action. Although insulin has a very short half-life of 5–10 minutes, it has been shown that glycation of insulin and proinsulin may occur in the pancreas during the periods of insulin synthesis and storage. Boyd and coworkers prepared monoglycated insulin to correspond to precisely what has been found in vivo, that is, single glycation modification in the phenylalanine¹ position in the amino-terminus of the insulin B-chain. When compared to non-glycated control insulin, infusion of the glycated form to mice undergoing a glucose tolerance test revealed a 20% reduction in glucose-lowering potency. In isolated abdominal muscle, monoglycated insulin was 20% less effective in mediating glucose update and in glucose oxidation and glycogen production compared to the non-glycated form of insulin³⁰. In plasma pooled from four male human subjects with type 2 diabetes (mean glycosylated hemoglobin levels of 8.1%), Hunter and colleagues showed that monoglycated insulin comprised approximately 9% of the total insulin. When pure monoglycated insulin was tested in hyperinsulinemic euglycemic clamp studies, administration of glycated insulin resulted in a reduced requirement for exogenous glucose infusion to maintain euglycemia and revealed that 70% more glycated insulin was required to induce a similar amount of insulin-mediated uptake of glucose³¹. Hence, in animal models and human subjects, evidence suggests that glycation of insulin may impair its action.

In distinct studies, Jia and colleagues showed that methylglyoxal modification of insulin at an arginine residue of the insulin B-chain reduced glucose uptake induced by MG-modified insulin in 3T3-L1 adipocytes and L8 skeletal muscle cells compared to native insulin.³² In other studies, MG modification of insulin affected insulin signaling directly; Fiory and colleagues showed that MG modification of insulin blocked IRS tyrosine phosphorylation and PI3K protein activation in the INS-1 pancreatic beta cell line³³. Oliveira and colleagues showed that MG-modification of insulin reduced insulin fibril formation and lead to the generation of insulin native-like aggregates³⁴. The authors speculated that this would result in diminished insulin action. It is important to note that the extent to which insulin is modified in vivo by MG modification in human subjects is not completely clarified.

Glycation of Distinct Proteins and Insulin Resistance

There are numerous published examples of protein glycation and mechanisms by which they may impact insulin sensitivity. Glycation of albumin has been shown to increase production of tumor necrosis factor-alpha (TNF-alpha); TNF-alpha has been linked to insulin resistance via induction of pro-inflammatory mechanisms that suppress insulin signal transduction^35–36. In L6 skeletal muscle myotubes, Cassese and colleagues showed that human glycated albumin induced Src-mediated activation of PKC alpha and suppressed IRS-1 action in a RAGE-dependent manner³⁷.

Because of the diverse nature of potential targets for AGE- and MG-modification, Gugliucci and colleagues proposed the hypothesis that MG modification of AMP-kinase (AMPK) might contribute to metabolic dysfunction and hepatic insulin resistance. Gugluicci reasoned that the sensing of AMP levels by AMPK is dependent on a domain with three arginine residues. If any or all of those residues were modified by the potent glycating agent MG, then it is plausible that functional modification might ensue as well³⁸. Hence, if AMPK activity were reduced, consequences might included enhanced gluconeogenesis and lipogenesis – both features that characterize hepatic insulin resistance. Although not proved experimentally at this time, the concept nevertheless raises the intriguing possibility that glycated modification of distinct biochemical moieties that contribute to regulation of insulin sensitivity beyond AMPK, such as other key metabolic enzymes, molecules involved in mitochondrial biogenesis and molecules that modulate oxidative stress, for example, might drive further AGE- or MG-mediated microenvironments conducive to the development of insulin resistance. In this context, efforts to reduce glycation in vivo have been tested in experimental animal model systems.

Glycation and Insulin Resistance: In Vivo Evidence

Support for roles for glycation in vivo has been sought with the use of anti-glycating agents to address the question, does reduction or prevention of glycation in vivo block the development of insulin resistance in vulnerable organisms? To test this concept, investigators have tested such agents in an array of animal models.

Unoki-Kubota and colleagues examined KK-A(y) mice, a mouse model of obesity and type 2 diabetes, and found that serum levels of AGEs correlated positively with the levels of insulin in these animals. To address if glycation products played roles in the insulin resistance phenotype, they administered pyridoxamine, an inhibitor of AGE formation, to the mice and found that in a dose-dependent manner, the agent decreased fasting insulin levels and improved insulin sensitivity, but did not affect fasting blood glucose levels (at least over the short time course of administration)³⁹.

Guo and colleagues treated Sprague-Dawley rats with MG in the drinking water to address the question of whether MG would induce insulin resistance. Two different therapeutic approaches were added to the experimental design: MG alone versus MG + N-acetyl cysteine (NAC, an antioxidant) or MG + TM2002, an inhibitor of AGEs. The animals were treated for four weeks and then subjected to hyperinsulinemic euglycemic clamp studies. MG administration induced insulin resistance; compared to MG alone, treatment with either NAC or TM2002 completely improved the insulin resistance induced by MG. In an additional set of animals, co-treatment with MG and high salt diet was tested. Compared to MG or high salt diet alone, co-treatment resulted in higher systolic blood pressure, increased urinary excretion of albumin and urinary TBARS, and renal nitrotyrosine levels. Of note, the AGE derived from MG, CEL AGE, was higher in any groups treated with MG versus groups not given MG in this study⁴⁰. Importantly, although these studies did not detail the precise molecular mechanisms by which MG induced insulin resistance, they nevertheless provided evidence linking MG to insulin resistance in vivo.

Kooptiwut and colleagues addressed these concepts directly in pancreatic islets. They retrieved islets from the pancreata of DBA/2 and C57BL/6 mice and exposed them to 11.1 mM or 40 mM glucose concentrations alone or in the presence of the AGE inhibitor aminoguanidine. They found that chronic exposure to hyperglycemia resulted in decreased glucose (20 mM)-stimulated islet secretion of insulin in the DBA/2 not C57BL/6 islets, and that this was associated with reduced glucokinase in the DBA/2 islets. When the islets were co-incubated with aminoguanidine, the levels of DBA/2 glucokinase rose, thereby suggesting that chronic glucose-mediated increases in AGEs could participate in islet dysfunction over this time course, especially in DBA/2 background, and that efforts to reduce AGEs might be protective⁴¹.

Finally, Matsumoto and colleagues reported on a model system in silkworm in which exposure of the silkworm to high glucose resulted in decreased growth, in parallel with increased AGE levels. When the silkworms were treated with high glucose and with aminoguanidine, growth was restored, thereby suggesting that the accumulation of AGEs was contributing to growth suppression. Interestingly, in this model, treatment with AICAR (stimulator of AMPK) or insulin also improved growth, thus identifying positive controls for a novel model system to test the adverse effects of high glucose and AGEs on silkworm growth⁴².

Perspectives, Challenges & Future Directions

Taken together, the data discussed in this review, from both in vitro and in vivo experimentation, provide supporting evidence that AGEs may contribute, at least in part, to the pathogenesis of insulin resistance by a variety of potential mechanisms, as summarized in Figure 2. The finding that AGE inhibitors, although they may have additional mechanism(s) of action beyond suppressing AGE, suppress insulin resistance in animal models, identifies an entirely novel area of therapeutic possibilities for this disorder. It is important to note that in addition to endogenous formation of AGEs, high fat diet-induced AGE introduction in the diet, and environmental pollutants may be linked to AGE generation as well. Hence, dietary modification and reduction in environmental pollutants may also hold promise for therapeutic intervention in this key area of public health concern.

Figure 2. Glycation, AGEs & Cellular Consequences: Implications for the Pathogenesis of Insulin Resistance & Potential Therapeutic Strategies.

There are a number of ways in which glycation products (early or AGEs) might impact insulin resistance states (illustrated in blue). Glycation product-mediated generation of TNF-alpha may directly block insulin signaling. AGE or MG-modification of insulin itself might result in marked reduction of insulin action. In fact, the consequences of impaired insulin action, such as downstream increased production of reactive oxygen species (at least in part via mitochondrial dysfunction) might further augment AGE production and target tissue insulin resistance (illustrated in green). In this review, we discussed some of the means to block AGEs (such as aminoguanidine, pyridoxamine, N-acetyl cysteine (NAC), TM2002 and possibly AGE receptor antagonists, as means to suppress insulin resistance at the target tissue level. [Abbreviations & Symbols: MG, methylglyoxal, AGE, advanced glycation endproduct; AMPK, AMP kinase; and NAC, N-acetyl cysteine]

It is important to note that recent studies have highlighted potentially adaptive roles for MG in the nervous system. Distler and colleagues showed that MG is protective against anxiety; the authors showed that glyoxalase 1 increases anxiety in transgenic mice overexpressing this enzyme by reducing GABA_A receptor activation by MG. Consistent with specific roles for MG, administration of MG assuaged anxiety in the glyoxalase 1 transgenic mice⁴³.

Furthermore, data suggest that blockade of cellular receptors for AGEs might be beneficial in insulin resistance states, in which even fleeting bouts of hyperglycemia might facilitate AGE formation sufficient to trigger activation of signal transduction receptors. A clear challenge in the field will thus be to identify the specific AGEs and their levels in plasma and / or in tissues that mediate insulin resistance vs. those that trigger diabetic complications. It will be important to decipher if differential effects of distinct types of AGEs, their levels in the tissue microenvironment and possible receptor interactions differ in these settings.

Finally, the role of AGE-receptor interaction might be critical to address in therapeutic strategies for additional reasons. Specifically, AGE-receptor interaction has been implicated in the late stage steps in insulin resistance in which accruing damage to the pancreatic beta cell eventually overwhelms the capacity of the islet to compensate for such stress, thereby resulting in hyperglycemia and diabetes. AGEs have been shown to damage pancreatic beta cells via oxidative stress in a process that appears to require RAGE^44–45. To what extent AGE precursors and AGEs are implicated in the development of insulin resistance and type 2 diabetes in vivo is an important question and one worthy of investigation. As the incidence of obesity and insulin resistance continue to rise in adolescents and adults at staggering levels, new approaches to tackling this world-wide epidemic are warranted.