Advanced glycation end products, dementia, and diabetes.

It is becoming abundantly clear that the insight into the pathological process of Alzheimer’s disease (AD) provided through autosomal dominant variants of the condition is only a partial one. The formation and aggregation of Aβ and the phosphorylation and aggregation of tau are clearly part of the core pathogenesis. However, although these may be necessary processes, and indeed in familial forms of the condition possibly also sufficient processes, they are not the whole story in the common late-onset forms of the disease. Here, mixed pathologies are the norm and at post mortem the plaques and tangles formed by Aβ and tau are accompanied also by inflammation and by vascular disease. This finding is congruent with the epidemiology that has long pointed to only three substantial factors that alter risk of dementia other than age: head injury, anti-inflammatory drugs, and diabetes. It is reasonably clear why head injury and anti-inflammatory drugs might affect risk but the relationship between diabetes and dementia has been far less clear. It could be that diabetes simply increases risk of vascular and related damage to the brain as it does to limbs, kidneys, and other organs. More interesting to molecular scientists are the observations that insulin signaling modifies amyloid precursor protein (APP) metabolism and tau phosphorylation in cells and in animal models, suggesting a possible influence of metabolic pathways on the canonical pathway of AD. Most intriguing of all is the observation that such pathways and processes are related not only to metabolic disease and to AD, but also to aging or longevity itself. In PNAS, Cai et al. (1) shed light on these complex interactions and point to possible clinical implications, including both biomarkers and potential therapeutics. In line with the considerable evidence for an oxidant-mediated pathogenic effect, Cai et al. show that a pro-oxidant diet in mice induces β-cleaved APP and Aβ generation. At the same time these mice, fed methyl-glyoxyl (MG) derivatives, had increased weight and systemic insulin resistance. In both the mice and also in a human cohort, dietaryMG levels and accompanying advanced glycation end products (AGEs) correlated positively with cognitive deficits or decline and inversely with survival factor sirtuin-1 (SIRT1) levels and other markers of insulin sensitivity. Themost striking, and sometimesneglected, observation in relation to risk of AD is that age increases it. This might be a simple inconsequential coincidence; it takes a lifetime of insult to result in neuronal dysfunction. However, the results of Cai et al. (1) hint at a more substantial possibility. The sirtuins, of which there are seven in mammals, have long been implicated as factors involved in longevity (2, 3). The most studied—SIRT1, -3 and -6—are all increased as a consequence of caloric restriction (CR), and CR increases both longevity and insulin sensitivity and also reduces certain features of AD pathology (4). It may indeed be the effect of CR on SIRT expression or function that is the critical factor, as mice lacking SIRT1 had a shortened lifespan even in the context of CR (5). These data suggest that the positive effects of CR are, at least in part, mediated by the induction of sirtuins. Cai et al. (1) add another element to this growing story in showing that an oxidative diet of MG derivatives (MG+) decreases SIRT1 protein in mouse brain and, in man, there is evidence of increased MG in blood correlated with decreased SIRT1 mRNA levels in circulating monocytes. There is an apparent anomaly here though: in Cai et al. increased MG is associated with cognitive decline and decreased SIRT1, whereas in mouse models of neurodegeneration SIRT1 levels are generally increased (6). A plausible explanation for this otherwise puzzling observation