Glycation and the AGE–RAGE axis

The study of the role played by glycation in disease pathogenesis originally relied on measuring fructosamine levels in biological fluids, combined with the characterisation of endogenous AGE in the circulation and tissues⁽ Reference Schmidt, Hori and Brett ^{6 ,} Reference Sell and Monnier ^{7 )}. These measurements were related to glycaemia and the topic very much focused on diabetes⁽ Reference Koenig, Peterson and Jones ^{3 –} Reference Koenig, Blobstein and Cerami ^{5 )}.

In hyperglycaemia (post-prandially or in non-controlled diabetes) and to a lesser extent in normoglycaemia, both circulatory proteins and proteins of the endothelium are exposed to (excess) glucose, leading to the slow formation of AGE⁽ Reference Monami, Lamanna and Lambertucci ^{8 –} Reference Beisswenger, Howell and O'Dell ^{11 )}. During that process, glycation adducts are created on the protein molecule, as a function of glucose levels. Accumulation of glycation adducts on the protein promotes excessive cross-linking with other protein molecules, which, in the case of collagen for example, would inhibit the formation of an ordered and functional polymeric complex. Such changes could lead to the formation of a thick vascular wall with (1) reduced elasticity and (2) a high affinity of collagen to bind other circulating proteins such as IgG, albumin and lipoproteins such as LDL⁽ Reference Brownlee, Cerami and Vlassara ^{12 –} Reference Sensi, Tanzi and Bruno ^{21 )}. In turn, the immobilisation of proteins on the vascular wall will promote further glycation and cross-linking and will act as a signal for chemoattraction of macrophages and monocytes, promoting inflammation and ‘foam’ cell formation in the endothelium⁽ Reference Klein, Laimins and Lopes-Virella ^{22 –} Reference Basta, Schmidt and De Caterina ^{24 )}.

The discovery that AGE can bind on cellular receptors and alter intracellular events was a breakthrough, linking glycation to signalling⁽ Reference Schmidt, Vianna and Gerlach ^{25 )}. Receptors such as AGE-R1, AGE-R2, AGE-R3, MSRII, CD36, LOX-1 and the receptor for AGE (RAGE), the most characterised receptor⁽ Reference Vlassara, Li and Imani ^{26 )}, are multi-ligand cell-surface Ig, with the ability to initiate injury-like intracellular events, mainly expression of genes related with inflammation and oxidative stress⁽ Reference Schmidt, Hori and Chen ^{27 –} Reference Kislinger, Fu and Huber ^{29 )}. Upon activation of RAGE, intracellular reactive oxygen species levels are increased through up-regulation of NAD(P)H oxidase expression. This in turns leads to activation of the Ras-mitogen activated protein kinase pathway, ultimately up-regulating NF-κB and the production of inflammatory molecules (including TNF-α, vascular cell adhesion molecule 1, intercellular adhesion molecule 1 and IL-1β). The up-regulation of NF-κB also initiates a positive feedback loop that sensitises the cell (and hence the tissue) to AGE by promoting RAGE production⁽ Reference Basta, Schmidt and De Caterina ^{24 )}.

Together accumulation of AGE in tissues and AGE–RAGE interactions are the two main pathways of glycation involvement in disease pathogenesis. These two pathways are often acting simultaneously and their individual effects are hard to distinguish; hence they are commonly presented in the same context when discussing glycation-related pathophysiology⁽ Reference Brownlee, Cerami and Vlassara ^{12 ,} Reference Takeuchi and Yamagishi ^{30 –} Reference Creager and Lüscher ^{34 )}.

Glycation and health throughout the lifecycle

Glycation is relevant to all stages in the lifecycle, including conception and early gestation. The reproductive tract is a known site for AGE accumulation both in men⁽ Reference Mallidis, Agbaje and Rogers ^{35 )} and women⁽ Reference Diamanti-Kandarakis, Piperi and Patsouris ^{36 )}. AGE accumulation is followed by changes in the distribution of RAGE in reproductive tissues⁽ Reference Mallidis, Agbaje and Rogers ^{37 )}, and the soluble isoform of RAGE in seminal/follicular fluid⁽ Reference Karimi, Goodarzi and Tavilani ^{38 ,} Reference Bonetti, Borges and Braga ^{39 )}, which may lead to lower sperm quality⁽ Reference Karimi, Goodarzi and Tavilani ^{38 )}, lower likelihood of success following assisted reproduction⁽ Reference Malickova, Jarosova and Rezabek ^{40 ,} Reference Jinno, Takeuchi and Watanabe ^{41 )} and reduced embryonal quality and development⁽ Reference Bonetti, Borges and Braga ^{39 ,} Reference Jinno, Takeuchi and Watanabe ^{41 ,} Reference Hao, Noguchi and Kamada ^{42 )}. During the course of pregnancy, activation of the AGE–RAGE axis may be involved in the pathogenesis of preeclampsia⁽ Reference Oliver, Buhimschi and Dulay ^{43 –} Reference Cooke, Brockelsby and Baker ^{45 )}. So far, evidence on the involvement of AGE and/or RAGE in fetal development is limited and based on animal studies. For example, a study on transgenic mice showed that overexpression of RAGE was associated with impairments in alveolar morphogenesis. The degree of RAGE overexpression was related to the magnitude of the abnormality with homozygous mice having histological changes similar to human bronchopulmonary dysplasia. The study also found that these early life changes could lead to increased risk of ‘destructive’ emphysema⁽ Reference Fineschi, De Cunto and Facchinetti ^{46 )}. Glycation has also been proposed as a mechanism of ageing⁽ Reference Gul, Rahman and Salim ^{47 ,} Reference Gkogkolou and Bohm ^{48 )}. Evidence from animal models suggest that a diet low in AGE (50 % reduction in AGE intake) was associated with amelioration of insulin resistance, lower AGE accumulation (both indications of the ageing process) and ultimately increased lifespan compared with the controls⁽ Reference Cai, He and Zhu ^{49 )}. Similarly, mice on caloric restriction, a popular model of lifespan expansion in animal models, have lower levels of collagen cross-linking and lower levels of lens cataract, suggesting lower AGE accumulation in the vitreous and the extracellular matrix⁽ Reference Taylor, Lipman and Jahngen-Hodge ^{50 ,} Reference Reiser ^{51 )} as well as in the brain⁽ Reference Mouton, Chachich and Quigley ^{52 )}. In fact, mice fed high AGE diets while on caloric restriction did not show any increase in their lifespan and the authors of the report suggested that lower AGE intake may be one of the mechanisms behind the caloric restriction model⁽ Reference Cai, He and Zhu ^{49 ,} Reference Cai, He and Zhu ^{53 )}. An interesting observation linking the effect of AGE in ageing and as early in life as in conception comes from a study showing the active involvement of AGE accumulation in ovarian ageing and ovarian function in human subjects⁽ Reference Stensen, Tanbo and Storeng ^{54 )}.

HbA1c and risk of chronic diseases

Even though the exact mechanisms of disease pathogenesis remain elusive, extensive evidence is available to associate glycation with disease risk. Glycation has a particular relevance for age-related diseases, including Alzheimer's disease⁽ Reference Smith, Sayre and Monnier ^{55 ,} Reference Srikanth, Maczurek and Phan ^{56 )}, skin ageing⁽ Reference Gkogkolou and Bohm ^{48 )} and cataract⁽ Reference Gul, Rahman and Salim ^{47 )}. These conditions are characterised by increased, possibly lifelong, deposition of AGE in the affected tissue⁽ Reference Vitek, Bhattacharya and Glendening ^{57 –} Reference Nicholl, Stitt and Moore ^{59 )}.

As in vivo glycation is believed to be mainly driven by plasma glucose concentrations, the most established relationship is between glycation and diabetes. HbA1c is the gold standard method for diabetes diagnosis and monitoring⁽ Reference American Diabetes ^{60 )}. According to the American Diabetes Association, individuals with HbA1c levels between 5·7 and 6·5 % are considered at high risk of developing diabetes. Those with HbA1c > 6·5 % are classified as having diabetes⁽ Reference American Diabetes ^{61 )}. Among patients with diabetes, higher HbA1c levels are associated with increased risk of retinopathy⁽ Reference Porta, Sjoelie and Chaturvedi ^{62 –} Reference McCarter, Hempe and Gomez ^{66 )}, neuropathy⁽ Reference El-Salem, Ammari and Khader ^{67 )} and nephropathy⁽ Reference McCarter, Hempe and Gomez ^{66 )}.

Glycation has recently attracted attention as a risk factor for normoglycaemic individuals. For the purpose of the present paper, we conducted a systematic literature search to identify studies documenting the effect of increased glycation on the risk of non-communicable chronic diseases in normoglycaemic subjects. We identified fifteen reports from eight studies (European Prospective Investigation into Cancer and Nutrition⁽ Reference Khaw, Wareham and Bingham ^{68 ,} Reference Khaw, Wareham and Bingham ^{69 )}, Atherosclerosis Risk in Communities study⁽ Reference Selvin, Steffes and Zhu ^{70 –} Reference Selvin, Rawlings and Grams ^{73 )}, Australian Diabetes, Obesity and Lifestyle study⁽ Reference Barr, Boyko and Zimmet ^{74 )}, the Hoorn Study⁽ Reference de Vegt, Dekker and Ruhe ^{75 ,} Reference van't Riet, Rijkelijkhuizen and Alssema ^{76 )}, Framingham Offspring⁽ Reference Meigs, Nathan and D'Agostino ^{77 )}, Rancho Bernardo⁽ Reference Park, BarrettConnor and Wingard ^{78 )}, Women's Health Study⁽ Reference Pradhan, Rifai and Buring ^{79 –} Reference Blake, Pradhan and Manson ^{81 )} and National Survey of Cardiovascular Disorders 1990⁽ Reference Sakurai, Saitoh and Miura ^{82 )}) analysing data from a total of over 63 000 participants, followed-up for 4–15 years. The outcomes of interest were diabetes risk, CVD, IHD, stroke, CHD and all-cause and CVD mortality. Two reports focused on the association between glycation and cancer risk, especially colorectal⁽ Reference Khaw, Wareham and Bingham ^{69 )} and breast cancer⁽ Reference Lin, Ridker and Rifai ^{80 )}. Overall, the studies showed a positive relationship between higher HbA1c and the risk of stroke and/or CVD and/or mortality ranging between 18 and 55 % higher risks per 1 % increase in HbA1c⁽ Reference Khaw, Wareham and Bingham ^{68 –} Reference Barr, Boyko and Zimmet ^{74 ,} Reference Sakurai, Saitoh and Miura ^{82 )}. As far as cancer incidence is concerned, the results are still inconclusive. Data from the European Prospective Investigation into Cancer and Nutrition cohort suggest a 33 % increase in the incidence of colorectal cancer per every 1 % increase in HbA1c⁽ Reference Khaw, Wareham and Bingham ^{69 )}, but an analysis of the Women's Health Study data did not find any association between HbA1c and breast cancer risk⁽ Reference Lin, Ridker and Rifai ^{80 )}. As the two cancer types differ significantly in aetiology, colorectal cancer has a strong dietary link⁽ Reference Edwards, Ward and Kohler ^{83 )} whereas breast cancer is mainly of genetic aetiology⁽ Reference McPherson, Steel and Dixon ^{84 )}; more research is needed before any conclusion is reached.

Oxidative stress and protein glycation in normoglycaemia

As observed by Selvin et al.⁽ Reference Selvin, Steffes and Zhu ^{70 )}, fasting glucose may fail to explain the positive relationship between HbA1c and CVD and/or mortality. Correction for classical risk factors (including smoking, dyslipidaemia and inflammation) explain the relationship better⁽ Reference de Vegt, Dekker and Ruhe ^{75 ,} Reference van't Riet, Rijkelijkhuizen and Alssema ^{76 ,} Reference Pradhan, Rifai and Buring ^{79 ,} Reference Blake, Pradhan and Manson ^{81 )}, suggesting that a shared mechanism may drive the increase in HbA1c levels. Although indications and potential mechanisms are in place to suggest an active involvement of oxidative stress in protein glycation in normoglycaemia and hence the increase in the risk of chronic diseases, so far little evidence is available to support such a hypothesis.

In our previous work, we hypothesised that oxidative stress could be this shared mechanism, which acts as a glycation driver in normoglycaemia.

Using the Scottish Health Surveys datasets 1993–2010, we have shown that, in individuals without diabetes and HbA1c levels lower than 6·5 %, age–sex adjusted HbA1c levels are positively correlated with smoking status, an association seen even among ex-smokers who used to smoke regularly⁽ Reference Vlassopoulos, Lean and Combet ^{85 )}. Smoking status was used as a proxy for oxidative stress and, in a similar way, fruit and vegetable intake was used as a proxy for antioxidant intake. Smoking was positively associated with HbA1c levels from as few as ten cigarettes per day a finding consistent with previous reports⁽ Reference Sargeant, Khaw and Bingham ^{86 ,} Reference Clair, Bitton and Meigs ^{87 )} (Fig. 1). The likelihood of having an HbA1c level within the prediabetes range (5·7–6·4 %) was double among smokers compared with non-smokers; this was seen even with less than ten cigarettes per day smoked. Interestingly, smoking cessation does not lead to complete reversal to the non-smoking state, as former smokers were found to have lower HbA1c levels than smokers but not as low as never smokers⁽ Reference Sargeant, Khaw and Bingham ^{86 ,} Reference Clair, Bitton and Meigs ^{87 )}. In a linear regression model, smoking was associated with 0·08 % higher HbA1c compared with no smoking, which is equal to 0·25 times the sd. As expected, vegetable intake had the opposite effect being associated with lower age–sex adjusted HbA1c levels with more portions consumed. In fact, for every extra 80 g portion of vegetable consumed there was an associated 0·01 % reduction in HbA1c.

Fig. 1. Age–sex adjusted mean (sd) of %HbA1c according to number of cigarettes per day (adapted from Vlassopoulos et al.⁽ Reference Vlassopoulos, Lean and Combet ^{85 )}).

The hypothesis that glycative and oxidative damage are closely related in vivo is supported by evidence showing that in purified plasma albumin, oxidative damage, measured as a reduction in free thiol groups, was positively related to glycative damage, measured as fructosamine and carbonyl rate⁽ Reference Guerin-Dubourg, Catan and Bourdon ^{88 )}. Moreover, Cys-34, a key site of oxidative damage in albumin in vivo ⁽ Reference Kawakami, Kubota and Yamada ^{89 )}, has also been suggested as a glycation site, especially from α-oxoaldehydes⁽ Reference Rondeau and Bourdon ^{90 )}. Since in vitro models are often removed from physiologically relevant reactions, it is important to set up mechanistic studies with adequate parameters. To test the hypothesis that, in normoglycaemia, oxidative stress promotes glycation, we carried out 4-week albumin incubation studies (albumin has a half-life of 14–28 d). Glucose concentrations 5 and 10 mm were employed to replicate normoglycaemia and (non-controlled) diabetes, respectively, whereas 20 and 30 mm glucose were used as the positive controls (supraphysiological concentrations). There is no consensus on the plasma levels of H₂O₂ (from nearly 0 to 35 μm ⁽ Reference Varma and Devamanoharan ^{91 –} Reference Frei, Yamamoto and Niclas ^{93 )}); we used a low concentration H₂O₂ (10 nm) to simulate physiologically relevant oxidative stress⁽ Reference Mueller, Riedel and Stremmel ^{94 )}. Co-incubation of albumin with glucose and physiological levels of H₂O₂ led to significantly higher glycation at all glucose levels tested, after 2 and 4 weeks incubation, compared to glucose alone. At the physiological glucose level (5 mm), there was no significant glycation (v. negative control) in absence of H₂O₂ (Fig. 2), indicating that oxidative stress plays an important role in glycation in normoglycaemia. Physiologically, in the presence of oxidative stress, proteins can get quickly oxidised and remain in this form in circulation until they are degraded by proteases⁽ Reference Stadtman and Levine ^{95 )}. As extracellular/circulating proteins are more likely to get oxidised first before getting glycated, due to the relative speed of the reactions, the same experiments were repeated using pre-oxidised protein. The pre-oxidised bovine serum albumin led to a higher production of fructosamine when incubated with glucose as compared with the native incubated BSA. Oxidative stress also drove glycation of human plasma proteins, in presence of 5 mm glucose.

Fig. 2. Differences in fructosamine concentration after incubation with glucose alone compared to glucose and constant exposure to oxidation from hydrogen peroxide (10 nm) after 2 and 4 weeks incubation. *P < 0·05 native v. constant oxidation; fructosamine was measured using the nitroblue tetrazolium method with the synthetic fructosamine equivalent deoxy-morpholino-fructose (DMF) as a calibrator (adapted from Vlassopoulos et al.⁽ Reference Vlassopoulos, Lean and Combet ^{96 )}).

Brought together, these results⁽ Reference Vlassopoulos, Lean and Combet ^{85 ,} Reference Vlassopoulos, Lean and Combet ^{96 )} indicate the potential role for oxidative stress as a driver for glycation in normoglycaemic individuals. The increased levels of HbA1c seen in smokers and those consuming low amounts of fruit and vegetables could be partially due to their impaired redox status, as stipulated by the epidemiological data. This interaction between oxidative stress and glycation will be subtle but with potentially sizeable long-term effects. Hence, dietary interventions aiming to restore the antioxidant/pro-oxidant balance in subjects at high risk of oxidative stress could be of value in chronic disease prevention.