Notice: Wiley Online Library will be unavailable on Saturday 30th July 2016 from 08:00-11:00 BST / 03:00-06:00 EST / 15:00-18:00 SGT for essential maintenance. Apologies for the inconvenience.
Center for Neuroscience and Cell Biology, University of Coimbra, Coimbra, Portugal
Faculty of Medicine, University of Coimbra, Coimbra, Portugal
Laboratory of Physiology, Faculty of Medicine, University of Coimbra, Coimbra, Portugal
Center for Neuroscience and Cell Biology, University of Coimbra and Institute of Physiology, Faculty of Medicine, University of Coimbra, 3004-517 Coimbra, Portugal and College of Sciences, University of Texas at San Antonio, USA
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
The prevalence of neurodegenerative diseases is rising dramatically worldwide due to the population aging, a situation that represents an enormous economic and social burden (1, 2). According to the World Alzheimer Report 2011, it is estimated that 36 million people worldwide live with dementia, a number doubling every 20 years, and reaching 115 million people by 2050 (3). Among these, 50–80% corresponds to Alzheimer's disease (AD) cases, the most prevalent form of dementia in people over 65 years (4).
AD is a slow, progressive, and fatal neurodegenerative disease that can remain asymptomatic for as long as 20 years (2, 5). Despite more than a century of research, and the massive information on AD, especially in the last 20 years, the causes and an early and accurate diagnosis and treatment remain elusive (3, 4). The neuropathological hallmarks of AD include deposits of extracellular senile plaques containing aggregates of the amyloid β protein (Aβ), intracellular neurofibrillary tangles (NFT) composed of hyperphosphorylated tau protein, and a massive neuronal loss, mainly in the hippocampus and cortex (6, 7). AD can assume a familiar form (FAD) with an early onset (<65 years, some cases affecting individuals at the age of 30 years), which constitutes <5% of the cases (8). FAD is caused by known mutations, namely in the genes encoding for amyloid β precursor protein (AβPP) and presenilins 1 and 2 (PS1 and PS2, respectively), which are transmitted to offsprings in an autosomal dominant manner (6, 9). These mutations lead to an overproduction and deposition of Aβ, which culminates in hyperphosphorylated tau protein deposition and neuronal loss, as postulated by the amyloid cascade hypothesis (7, 10). In contrast, sporadic AD (SAD) has late onset (≥65 years) and aging, diabetes, and apolipoprotein E4 (APOE4) are considered main risk factors (1). Although the causes of SAD remain under discussion, evidence shows that impaired glucose/energy metabolism, mitochondrial dysfunction, oxidative stress, and altered insulin-signaling pathways are early events in disease pathogenesis (5, 10). Herewith, we discuss the main functions of insulin in the healthy, aged, and AD brain. Finally, we will give an overview of the relationship between AD and diabetes, particularly type 2 diabetes (T2D), putting focus on insulin signaling.
BRAIN INSULIN SIGNALING
Insulin and its receptors (IRs) are ubiquitously expressed in many tissues, including the brain (11, 12) where insulin can reach levels 10- to 100-fold greater than in plasma, especially in the hippocampus, cortex, hypothalamus, olfactory bulb, and pituitary (11, 13). IRs are largely localized in neurons, being less abundant in glia (12, 14). Insulin-mediated signaling at the central nervous system (CNS) has recently emerged as a new and promising field of research.
Insulin produced by pancreatic β-cells is transported by cerebrospinal fluid (CSF) into the brain and crosses the blood–brain barrier (BBB) by an active and saturable process (15, 16). It has been shown that an increase in circulating insulin is associated with a concomitant increase in CSF insulin levels, which affects brain activity (13). However, studies also revealed the presence of insulin in immature nerve cell bodies and, in rodents, less than 1% of the peripherally administered hormone reached the CNS suggesting a probable local insulin biosynthesis (11, 15).
As IRs, insulin-like growth factor-1 receptors (IGF-1Rs) are widely distributed throughout the brain (16, 17). These membrane-bound receptors belong to the superfamily of tyrosine kinase receptors and are homologous, triggering similar intracellular signaling events (12). Binding of insulin or IGF-1 promotes the receptor autophosphorylation, stimulating its tyrosine kinase activity and, subsequently, phosphorylating insulin receptor substrate (IRS) proteins on tyrosine residues as well as the Src homology collagen (Shc) peptide. As a consequence, two main signaling cascades, mediated by phosphatidylinositol 3-kinase (PI3K) and mitogen-activated protein kinase (MAPK), are activated (14, 18, 19). After PI3K activation, downstream signaling proteins, such as serine (Ser)/threonine (Thr) kinase Akt, are recruited to the plasma membrane, being then translocated to the cytosol and nucleus, thereby phosphorylating target proteins (e.g., glycogen synthase kinase-3β; GSK-3β) (20, 21). The Ser/Thr protein kinase GSK-3β contains two distinct forms: an active form (Ser9 dephosphorylated) that is mostly found in nuclei, mitochondria, and membrane lipid rafts and the cytosolic inactive form (14).
The insulin/IGF-1-mediated activation of Akt, protein kinase C, or c-AMP-dependent protein kinase leads to the inactivation of GSK-3β, which triggers multiple effects, including synthesis of proteins involved in neuronal glucose metabolism, antiapoptotic mechanisms, and antioxidant defense (13, 22). It was also described that overexpression of a constitutively active GSK-3β promoted cell death, while its inhibition prevented apoptosis (11). Other known molecules targeted by PI3K/Akt signaling are FoxO3, nuclear factor-kB (NF-kB), and cAMP response element-binding (CREB). In fact, Akt may also phosphorylate and inhibit FoxO3, preventing the disruption of mitochondrial membrane potential and cytochrome c release, thus promoting neuronal survival (13, 14). NF-kB phosphorylation by Akt has been shown to protect against oxidative stress and apoptosis by increasing Cu/Zn superoxide dismutase (Cu/Zn SOD) expression and manganese SOD (MnSOD) levels (11, 14). Moreover, the CREB target of Akt has been involved in the stimulation of neuronal glucose metabolism and enhancement of mitochondrial membrane potential, ATP levels, nicotinamide adenine dinucleotide phosphate (NADPH) redox state, and hexokinase activity (23). The activation of MAPK pathway seems to promote the expression of genes involved in cell and synapse growth as well as in cell repair and maintenance (19, 20). Interestingly, several studies suggest a crosstalk between both insulin signaling pathways that, by converging at Bad phosphorylation, may play an antiapoptotic role (24). Thus, both PI3K/Akt and MAPK pathways appear to underlie both neurotrophic and neuroprotective actions of insulin (Fig. 1).
The classical effects of insulin include glucose uptake, regulation of cell proliferation, gene expression, and the suppression of hepatic glucose production (15, 21). However, the IRs present in the CNS are slightly different from their peripheral counterparts (12). Insulin-mediated neuronal IRs and/or IGF-1Rs activation regulate a multitude of physiological functions, such as food intake, inhibition of hepatic gluconeogenesis, counter-regulation of hypoglycemia, reproduction, modulation of tau protein phosphorylation, AβPP metabolism and Aβ clearance, neuronal survival, and memory (14, 15, 18). Food intake and energy homeostasis seem to be regulated by hypothalamic glucosensing neurons, in which insulin signaling yields an anorexigenic effect by activating ATP-sensitive potassium (K+ATP) channels, leading to neuronal hyperpolarization. The anorexigenic effect may be due to the inhibition of neuropeptide Y and agouti-related peptide expression and induction of proopiomelanocortin and cocaine- and amphetamine-regulated transcript production (14, 15). Regarding brain glucose metabolism, although insulin is not a major regulator, recent studies suggest that changes in circulating insulin levels may modulate glucose transporters (GLUTs) expression (5, 20). Indeed, it was described that not only the increase in insulin levels enhanced GLUT4 expression, but also, in the human brain, fasting insulin levels stimulated glucose metabolism (10, 12). Hypothalamic insulin was also associated with a reduction in hepatic glucose production (15). Apart from these roles, cerebral IRs and IGF-1Rs have been also suggested to be involved in cortical and hippocampal synaptic plasticity, thus affecting memory and learning (14, 18). Insulin-mediated PI3K signaling cascade affects both long-term potentiation (LTP) and long-termdepression (LTD) by modulating glutamate, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid, N-methyl-D-aspartate (NMDA), and gamma-aminobutyric acid (GABA) receptors (13, 18). MAPK activity is also essential for LPT induction and memory consolidation (25). Moreover, among other neuronal roles, insulin also promotes neurite outgrowth and axonal regeneration (14, 18, 25). Under this perspective, it is not surprising that the impairment of brain insulin signaling will affect cellular function and survival potentiating brain aging and age-related diseases such as AD.
Brain aging results from an accumulation of intracellular damage caused by, among other factors, an excess of mitochondrial free radicals generation (due to an imbalance in cells oxidative status) that potentiate cells degeneration and, eventually, death (26). Further, epigenetic factors modulated by aging and a sedentary lifestyle have been described to induce an oxidative redox shift by attenuating mitochondrial metabolism and increasing reliance on glycolysis (27). This alteration starts a damaging cycle involving oxidized membrane receptors, signaling molecules, transcription factors, and epigenetic transcriptional regulators (11, 27). Accordingly, a decrease in IRs levels and impaired insulin signaling, mainly in the hippocampus, cortex, and choroid plexus have been observed (11). Under such oxidizing environment, IRs and IGF-1Rs may suffer oxidative modifications, resulting in the blockage of insulin stimulation, a situation reinforced by the failure of oxidized protein tyrosine phosphatase-1b to reactivate the receptors. As a consequence, despite increased glucose levels and stimulation of insulin production, glucose metabolism becomes impaired, thus hampering insulin signaling (27, 28). Impaired ligand–receptor binding and modifications in IRs/IGF-1Rs internalization can also promote chronic insulin resistance, which may occur due to an age-related increase in cholesterol and the subsequent decrease in membrane fluidity (28). Furthermore, insulin resistance and high levels of glucose potentiate oxidative stress, mitochondrial dysfunction, and advanced glycation end products (AGEs) formation (15) potentiating age-related alterations. Recently, Muller et al. (29) reported that IGF-I signaling was deteriorated in the brain of aged mice. The authors observed that basal as well as IGF-I-induced activation of the brain IGF-1R/Akt/GSK3 pathway was markedly reduced even though old mice have higher levels of brain IGF-1R (29). It was also reported that insulin resistance is associated with cognitive decline in nondemented aged individuals [age, years 73.3 (6.7)] (30).
Interestingly, it was shown that a reduction in insulin/IGF-1 signaling was associated with increased longevity both in model organisms and in aged mammal brain (31). More specifically, reduced central insulin/IGF-1 signaling was shown to increase stress resistance and extend lifespan in worms and flies. However, in the mammalian brain, reduction of insulin/IGF-1 signaling (e.g., by neuron-specific knockout of IRS2) extended lifespan and ameliorated AD pathology (31). Harries et al. (32) also found that genes involved in insulin signaling [phosphatase and tensin homolog (PTEN), PI3K, and 3-phosphoinositide-dependent protein kinase 1 (PDK1)] and insulin production and sensitivity [forkhead transcription factors (FOXO)] were inversely correlated with age. In contrast, other studies showed that increased insulin/IGF-1 signaling has neuroprotective effects (31, 33).
Although further work is needed to clarify this controversial findings, accumulating evidence shows that impaired insulin (and IGF-1) signaling is closely associated with neurodegenerative diseases, namely AD.
Diseased Brain: AD
Insulin signaling impairment has been increasingly associated with cognitive decline and increased risk for dementia (1, 19, 34). Evidence shows that AD is associated with a decrease in the levels of insulin in the CSF and/or CSF/plasma insulin ratio, a decline in the expression of IRs and an increase in fasting plasma insulin levels (11, 35). The decrease in insulin (as well as IGF-1) levels suggests an impaired insulin transport into the brain (34). Alternatively, this may also occur due to deregulated BBB function induced by prolonged peripheral hyperinsulinemia (17, 19). Other authors reported lower expression of IRs and IGF-1Rs as well as a reduction in their tyrosine kinase activity and IRS expression, a situation potentiated with the progression of the disease (17, 19, 34). Impaired insulin/IGF-1 ligand–receptor binding can also arise from changes in membrane cholesterol levels, thus affecting membrane composition upon aging and/or APOE4 genotype (19, 34).
Insulin/IGF-1 signaling defects predominantly affect PI3K/Akt pathway, triggering several harmful situations. It was recently proposed that decreased PI3K/Akt-mediated GLUTs activation and their decreased expression in AD brain could lead to a brain glucose hypometabolism and the subsequent decrements in mitochondrial metabolism and ATP production (17). Additionally, increased levels of circulating glucose in CNS may increase AGEs formation and ultimately their toxic effects (17, 25).
MAPK has been shown to be stimulated in AD patients (17), which is correlated with increased neuroinflammation, tau protein hyperphosphorylation, and AβPP trafficking (17, 36). Indeed, the aberrant hyperphosphorylation of tau protein probably arises from an exaggerated activation of GSK-3β, MAPK, and cyclin-dependent kinase 5, major tau kinases responsible for the phosphorylation of tau protein (25, 34). Further, the decreased phosphorylation of GSK-3β and the subsequent increase in its activity may potentiate γ-secretase activity and the amyloidogenic AβPP processing, resulting in increased intracellular levels of Aβ (10, 20). Insulin has been also proposed to modulate extracellular degradation of Aβ by interfering with the activity of insulin-degrading enzyme (IDE) (13, 14), a zinc-metalloprotease that degrades several pathophysiologically significant extracellular substrates, including insulin and Aβ. In fact, studies have demonstrated reduced IDE activity and mRNA and protein levels in the hippocampus of severe AD patients, which negatively correlated with brain Aβ1-42 content (37). Recently, Bomfim et al. (38) reported that Aβ oligomers can activate the tumor necrosis factor α/c-Jun N-terminal kinase pathway, induce IRS-1 phosphorylation at multiple serine residues, and inhibit physiological phosphorylated IRS-1 (at Tyr896) in mature cultured hippocampal neurons. The impairment of IRS-1 signaling was also observed in APP/PS1 transgenic mice, an animal model of AD as well as in cynomolgus monkeys intracerebroventricularly injected with Aβ oligomers (38). Similar observations were made by Talbot et al. (39) in human AD brains. In fact, the insulin resistant state that characterizes the brains of AD subjects is associated with IRS-1 dysregulation and IGF-1 resistance (39).
The studies discussed above clearly show that alterations in insulin signaling have a major impact in AD pathophysiology, given support to the idea that AD can be considered as the T2D of the brain (also known as type 3 diabetes).
DIABETES AND AD—IS INSULIN THE MISSING LINK?
Diabetes mellitus is a heterogeneous metabolic disorder characterized by hyperglycemia that results from the impairment in insulin production and/or action (36, 40). In type 1 diabetes (T1D), the autoimmune destruction of pancreatic β cells culminates in the loss of insulin production, whereas T2D is mainly characterized by an impaired insulin action—insulin resistance (13, 41). Globally, estimates pointed to 250 million diabetic people worldwide in 2010, with 90% of the patients being affected by T2D (1, 42). Diabetes is associated with several long-term complications such as cardiovascular disease, nephropathy, retinopathy, peripheral and autonomic neuropathy, and encephalopathy (11, 36).
Nowadays, it is clear that diabetes has a major impact in the brain, and substantial evidence suggests a role for chronic hyperglycemia, repeated episodes of severe hypoglycemia, vascular complications, and insulin resistance as major events contributing to cognitive dysfunction (18, 40). Many other conditions associated to diabetes potentiate cognitive decline, such as stroke, hypertension, dyslipidemia, and obesity (18, 43). In fact, evidence shows that the risk for neurodegeneration is increased in prediabetes and metabolic syndrome patients (42, 43). T1D and T2D patients have been previously shown to present cognitive dysfunction (memory, attention, intelligence, processing speed, and executive function deficits) and brain structure abnormalities (brain atrophy and white matter abnormalities) (1, 20). Data provided by neuroimaging techniques also revealed that brain atrophy in diabetic patients is more pronounced in cortical, subcortical, and hippocampal areas (1, 41). In T1D patients, cognitive deficits appear to be mainly due to inadequate glycemic control, as patients with acute hyperglycemia and/or hypoinsulinemia perform worse on cognitive function tests (18, 42), whereas in T2D, the decline in cognitive function appears to be more relevant in elderly patients and has been mostly related with insulin resistance (11, 18). Interestingly, T2D has been widely shown to accelerate brain aging, exacerbating its harmful effects, thus increasing brain susceptibility and the risk for development of neurodegenerative diseases (42, 44).
It is well known that diabetes mellitus is associated with approximately 20% of the neurodegenerative disorders, including vascular dementia, AD, Parkinson's, and Huntington's diseases (15, 44). AD and T2D possess several common features, including impaired glucose metabolism, insulin resistance, blood vessel abnormalities, mitochondrial dysfunction, increased oxidative stress, increased inflammatory response, deposition of amyloidogenic proteins, and deregulated protein phosphorylation, among others (1, 9, 19, 45, 46) (Fig. 2). The interrelation between AD and T2D has been also fostered by multiple studies reporting that, on one hand T2D facilitates AD onset and, on the other hand, that patients with AD have an increased risk for developing T2D (47). Liu et al. (48) investigated the brain insulin-PI3K-Akt signaling pathway in the frontal cortices of AD, T2D, T2D with AD, and control cases. The authors found that the deficiency of insulin-PI3K-Akt signaling was more severe in individuals with both T2D and AD. The levels and the activation of the insulin-PI3K-Akt signaling components correlated negatively with the level of tau protein phosphorylation and positively with protein O-GlcNAcylation (48), suggesting that impaired insulin-PI3K-Akt signaling might contribute to neurodegeneration in AD through down-regulation of O-GlcNAcylation and the consequent promotion of abnormal tau protein hyperphosphorylation and neurodegeneration.
The relation between AD and diabetes has been further demonstrated in animal models. For example, in an experimental rat model of streptozotocin (STZ)-induced T1D, increased levels of hyperphosphorylated tau protein as well as cognitive deficits were observed. Moreover, diabetes exaggerated defects in the brain of AβPP transgenic mice (49). A study performed in Bio-Breeding Zucker Diabetic rat (BBZDR)/Wor rats (a model of T2D) revealed that impaired insulin signaling, alterations of AβPP metabolism, and hyperphosphorylated tau protein preceded neurodegenerative events and neuronal loss (19). We recently reported that the brains of T2D mice presented mitochondrial abnormalities, oxidative imbalance, and levels of Aβ similar to those found in triple transgenic mice of AD (3xTg-AD) (50), reinforcing the idea that T2D is a risk factor for AD. Devi et al. (51) observed that in rodents insulin signaling impairment favors Aβ formation via the translational upregulation of AβPP and β-site AβPP cleaving enzyme 1 (BASE-1) (51). In high-fat diet-induced insulin resistance and diabetic db/db mice, an accumulation of Aβ associated with an increased β- and γ-secretases activities and accumulation of autophagosomes (which facilitate Aβ generation) were also observed (52).
These and other studies clearly indicate that abnormalities in insulin signaling are implicated in AD pathogenesis corroborating the idea that insulin is a link between AD and diabetes, namely T2D.
Impaired insulin signaling is an important issue in understanding the pathogenesis of AD. Indeed, insulin sensitivity is decreased with aging, T2D and AD, which could lead to neuronal dysfunction and cognitive decline. Moreover, it is becoming clear that diabetes is a risk factor for AD, and that these two diseases are linked by several common molecular and cellular processes, being insulin signaling one of the main links. Brain insulin signaling has gained an increased interest in neuroscience research. It has been documented that insulin plays several functions including glucose metabolism, mitochondrial function, amyloidogenesis, and cognitive function. The clarification of the precise molecular and cellular mechanisms underlying AD and/or diabetes, and how these mechanisms intersect, is of the outmost importance for the development of future therapies to delay the onset or prevent both AD and diabetes long-term deleterious effects in CNS.
The authors are grateful to Fundação para a Ciência e a Tecnologia (FCT, project PTDC/SAU-TOX/117481/2010; PTDC/SAU-NMC/110990/2009; PTDC/SAU-NEU/103325/2008), Portugal and Programa de Estímulo à Investigação da Faculdade de Medicina, Universidade de Coimbra, Portugal (PMADSC/2011) for financial support.