Ian V.J. Murray

Vascular and metabolic dysfunction in Alzheimer's disease: a review.

Alzheimers disease (AD) is thought to start years or decades prior to clinical diagnosis. Overt pathology such as protein misfolding and plaque formation occur at later stages, and factors other than amyloid misfolding contribute to the initiation of the disease. Vascular and metabolic dysfunctions are excellent candidates, as they are well-known features of AD that precede pathology or clinical dementia. While the general notion that vascular and metabolic dysfunctions contribute to the etiology of AD is becoming accepted, recent research suggests novel mechanisms by which these/such processes could possibly contribute to AD pathogenesis. Vascular dysfunction includes reduced cerebrovascular flow and cerebral amyloid angiopathy. Indeed, there appears to be an interaction between amyloid β (Aβ) and vascular pathology, where Aβ production and vascular pathology both contribute to and are affected by oxidative stress. One major player in the vascular pathology is NAD(P)H oxidase, which generates vasoactive superoxide. Metabolic dysfunction has only recently regained popularity in relation to its potential role in AD. The role of metabolic dysfunction in AD is supported by the increased epidemiological risk of AD associated with several metabolic diseases such as diabetes, dyslipidemia and hypertension, in which there is elevated oxidative damage and insulin resistance. Metabolic dysfunction is further implicated in AD as pharmacological inhibition of metabolism exacerbates pathology, and several metabolic enzymes of the glycolytic, tricarboxylic acid cycle (TCA) and oxidative phosphorylation pathways are damaged in AD. Recent studies have highlighted the role of insulin resistance, in contributing to AD. Thus, vascular and metabolic dysfunctions are key components in the AD pathology throughout the course of disease. The common denominator between vascular and metabolic dysfunction emerging from this review appears to be oxidative stress and Aβ. This review also provides a framework for evaluation of current and future therapeutics for AD.

Amyloid-β Metabolite Sensing: Biochemical Linking of Glycation Modification and Misfolding

Glycation is the reaction of a reducing sugar with proteins and lipids, resulting in myriads of glycation products, protein modifications, cross-linking, and oxidative stress. Glycation reactions are also elevated during metabolic dysfunction such as in Alzheimers disease (AD) and Downs syndrome. These reactions increase the misfolding of the proteins such as tau and amyloid-β (Aβ), and colocalize with amyloid plaques in AD. Thus, glycation links metabolic dysfunction and AD and may have a causal role in AD. We have characterized the reaction of Aβ with reactive metabolites that are elevated during metabolic dysfunction. One metabolite, glyceraldehyde-3-phosphate, is a normal product of glycolysis, while the others are associated with pathology. Our data demonstrates that lipid oxidation products malondialdehyde, hydroxynonenal, and glycation metabolites (methylglyoxal, glyceraldehyde, and glyceraldehyde-3-phosphate) modify Aβ42 and increase misfolding. Using mass spectrometry, modifications primarily occurred at the amino terminus. However, the metabolite methylglyoxal modified Arg5 in the Aβ sequence. 4-Hydroxy-2-nonenal modifications were similar to our previous publication. To place such modifications into an in vivo context, we stained AD brain tissue for endproducts of glycation, or advanced glycation endproducts (AGE). Similar to previous findings, AGE colocalized with amyloid plaques. In summary, we demonstrate the glycation of Aβ and plaques by metabolic compounds. Thus, glycation potentially links metabolic dysfunction and Aβ misfolding in AD, and may contribute to the AD pathogenesis. This association can further be expanded to raise the tantalizing concept that such Aβ modification and misfolding can function as a sensor of metabolic dysfunction.

Adenosine Triphosphate (ATP) Reduces Amyloid-β Protein Misfolding in vitro

Alzheimers disease (AD) is a devastating disease of aging that initiates decades prior to clinical manifestation and represents an impending epidemic. Two early features of AD are metabolic dysfunction and changes in amyloid-β protein (Aβ) levels. Since levels of ATP decrease over the course of the disease and Aβ is an early biomarker of AD, we sought to uncover novel linkages between the two. First and remarkably, a GxxxG motif is common between both Aβ (oligomerization motif) and nucleotide binding proteins (Rossmann fold). Second, ATP was demonstrated to protect against Aβ mediated cytotoxicity. Last, there is structural similarity between ATP and amyloid binding/inhibitory compounds such as ThioT, melatonin, and indoles. Thus, we investigated whether ATP alters misfolding of the pathologically relevant Aβ42. To test this hypothesis, we performed computational and biochemical studies. Our computational studies demonstrate that ATP interacts strongly with Tyr10 and Ser26 of Aβ fibrils in solution. Experimentally, both ATP and ADP reduced Aβ misfolding at physiological intracellular concentrations, with thresholds at ~500 μM and 1 mM respectively. This inhibition of Aβ misfolding is specific; requiring Tyr10 of Aβ and is enhanced by magnesium. Last, cerebrospinal fluid ATP levels are in the nanomolar range and decreased with AD pathology. This initial and novel finding regarding the ATP interaction with Aβ and reduction of Aβ misfolding has potential significance to the AD field. It provides an underlying mechanism for published links between metabolic dysfunction and AD. It also suggests a potential role of ATP in AD pathology, as the occurrence of misfolded extracellular Aβ mirrors lowered extracellular ATP levels. Last, the findings suggest that Aβ conformation change may be a sensor of metabolic dysfunction.

Probing and trapping a sensitive conformation: Amyloid-β fibrils, oligomers, and dimers

Alzheimers disease (AD) is a devastating neurodegenerative disease with pathological misfolding of amyloid-β protein (Aβ). The recent interest in Aβ misfolding intermediates necessitates development of novel detection methods and ability to trap these intermediates. We speculated that two regions of Aβ may allow for detection of specific Aβ species: the N-terminal and 22-35, both likely important in oligomer interaction and formation. We determined via epitomics, proteomic assays, and electron microscopy that the Aβ(42) species (wild type, ΔE22, and MetOx) predominantly formed fibrils, oligomers, or dimers, respectively. The 2H4 antibody to the N-terminal of Aβ, in the presence of 2% SDS, primarily detected fibrils, and an antibody to the 22-35 region detected low molecular weight Aβ species. Simulated molecular modeling provided insight into these SDS-induced structural changes. We next determined if these methods could be used to screen anti-Aβ drugs as well as identify compounds that trap Aβ in various conformations. Immunoblot assays determined that taurine, homotaurine (Tramiprosate), myoinositol, methylene blue, and curcumin did not prevent Aβ aggregation. However, calmidazolium chloride trapped Aβ at oligomers, and berberine reduced oligomer formation. Finally, pretreatment of AD brain tissues with SDS enhanced 2H4 antibody immunostaining of fibrillar Aβ. Thus we identified and characterized Aβs that adopt specific predominant conformations (modified Aβ or via interactions with compounds), developed a novel assay for aggregated Aβ, and applied it to drug screening and immunohistochemistry. In summary, our novel approach facilitates drug screening, increases the probability of success of antibody therapeutics, and improves antibody-based detection and identification of different conformations of Aβ.

Small Molecules and Alzheimer’s Disease: Misfolding, Metabolism and Imaging

Small molecule interactions with amyloid proteins have had a huge impact in Alzheimers disease (AD), especially in three specific areas: amyloid folding, metabolism and brain imaging. Amyloid plaque amelioration or prevention have, until recently, driven drug development, and only a few drugs have been advanced for use in AD. Amyloid proteins undergo misfolding and oligomerization via intermediates, eventually forming protease resistant amyloid fibrils. These fibrils accumulate to form the hallmark amyloid plaques and tangles of AD. Amyloid binding compounds can be grouped into three categories, those that: i) prevent or reverse misfolding, ii) halt misfolding or trap intermediates, and iii) accelerate the formation of stable and inert amyloid fibrils. Such compounds include hydralazine, glycosaminoglycans, curcumin, beta sheet breakers, catecholamines, and ATP. The versatility of amyloid binding compounds suggests that the amyloid structure may serve as a scaffold for the future development of sensors to detect such compounds. Metabolic dysfunction is one of the earliest pathological features of AD. In fact, AD is often referred to as type 3 diabetes due to the presence of insulin resistance in the brain. A recent study indicates that altering metabolism improves cognitive function. While metabolic reprogramming is one therapeutic avenue for AD, it is more widely used in some cancer therapies. FDA approved drugs such as metformin, dichloroacetic acid (DCA), and methylene blue can alter metabolism. These drugs can therefore be potentially applied in alleviating metabolic dysfunction in AD. Brain imaging has made enormous strides over the past decade, offering a new window to the mind. Recently, there has been remarkable development of compounds that have the ability to image both types of pathological amyloids: tau and amyloid beta. We have focused on the low cost, simple to use, near infrared fluorescence (NIRF) imaging probes for amyloid beta (Aβ), with specific attention on recent developments to further improve contrast, specificity, and sensitivity. With advances in imaging technologies, such fluorescent imaging probes will open new diagnostic avenues.

Lipid oxidation and modification of amyloid-β (Aβ) in vitro and in vivo.

Oxidative damage and amyloid-β (Aβ) protein misfolding are prominent features of Alzheimers disease (AD). In vitro studies indicated a direct linkage between these two features, where lipid oxidation products augmented Aβ misfolding. We tested this linkage further, mimicking specific conditions present in amyloid plaques. In vitro lipid oxidation and lipid modification of Aβ were thus performed with elevated levels of copper or physiological levels of calcium. These in vitro experiments were then confirmed by in vivo immunohistochemical and chemical tagging of oxidative damage in brains from the PSAPP mouse model of AD. Our in vitro findings indicate that: 1) high levels of copper prevent lipid oxidation; 2) physiological concentrations of calcium reduce 4 hydroxy-2-nonenal (HNE) modification of Aβ; and 3) anti-Aβ and HNE antibody epitopes are differentially masked. In vivo we demonstrated increased lipid oxidation around plaques but 4) a lack of immunological colocalization of HNE-adducts with Aβ. Thus, the lack of colocalization of Aβ and HNE-adduct immunostaining is most likely due to a combination of metals inhibiting HNE modification of Aβ, quenching lipid oxidation and a masking of HNE-Aβ histopathology. However, other forms of oxidative damage colocalize with Aβ in plaques, as demonstrated using a chemical method for identifying oxidative damage. Additionally, these findings suggest that HNE modification of Aβ may affect therapeutic antibodies targeting the amino terminal of Aβ and that metals effect on lipid oxidation and lipid modification of Aβ could raise concerns on emerging anti-AD treatments with metal chelators.

Ruthenium red colorimetric and birefringent staining of amyloid-β aggregates in vitro and in Tg2576 mice.

Alzheimers disease (AD) is a devastating neurodegenerative disease most notably characterized by the misfolding of amyloid-β (Aβ) into fibrils and its accumulation into plaques. In this Article, we utilize the affinity of Aβ fibrils to bind metal cations and subsequently imprint their chirality to bound molecules to develop novel imaging compounds for staining Aβ aggregates. Here, we investigate the cationic dye ruthenium red (ammoniated ruthenium oxychloride) that binds calcium-binding proteins, as a labeling agent for Aβ deposits. Ruthenium red stained amyloid plaques red under light microscopy, and exhibited birefringence under crossed polarizers when bound to Aβ plaques in brain tissue sections from the Tg2576 mouse model of AD. Staining of Aβ plaques was confirmed via staining of the same sections with the fluorescent amyloid binding dye Thioflavin S. In addition, it was confirmed that divalent cations such as calcium displace ruthenium red, consistent with a mechanism of binding by electrostatic interaction. We further characterized the interaction of ruthenium red with synthetic Aβ fibrils using independent biophysical techniques. Ruthenium red exhibited birefringence and induced circular dichroic bands at 540 nm upon binding to Aβ fibrils due to induced chirality. Thus, the chirality and cation binding properties of Aβ aggregates could be capitalized for the development of novel amyloid labeling methods, adding to the arsenal of AD imaging techniques and diagnostic tools.

Amyloids as Sensors and Protectors (ASAP) Hypothesis

This paper propounds the Amyloids as Sensors and Protectors (ASAP) hypothesis. In this novel hypothesis, we provide evidence that amyloids are capable of sensing dysfunction, and after misfolding, initiate protective cellular responses. Amyloid proteins are initially protective, but chronic stress and overstimulation of the amyloid sensor leads to pathology. This proposed ASAP hypothesis has two sequential stages: (i) sensing, and then (ii) protection. Sensing involves a conformational change of amyloids in response to the cellular environment. The protection aspect translates conformational change into a cellular response via several mechanisms. The most obvious mechanism is that protein misfolding triggers the protective unfolded protein response, and thus downregulates protein translation and increases chaperone proteins. Other documented responses include metabolic pathways and microRNAs. This ASAP hypothesis has precedence, as amyloid sensors exist (evidenced by CPEB and Sup35), and both prion and amyloid-β sensing redox stress and metals. Our hypothesis expands on previous observations to link sensing with inciting protective cellular response. Furthermore, we substantiate the ASAP hypothesis with previously published evidence from several amyloid diseases. This novel hypothesis links disparate findings in amyloid diseases: metabolic dysfunction, unfolding protein response/chaperones, modification of amyloids, and nutrient or caloric sensing. While this hypothesis can be applied to Alzheimers disease, it goes beyond the Alzheimers context. Thus all amyloid proteins can potentially act as sensors of misfolding-causing stress. Finally, this hypothesis will allow for the sensor mechanism and metabolic dysfunction to serve as biomarkers of the diseases as well as therapeutic targets early in disease pathology.

Glycogen and amyloid-beta: key players in the shift from neuronal hyperactivity to hypoactivity observed in Alzheimer's disease?

Introduction: Alzheimers disease (AD) begins to develop decades prior to its clinical manifestation (Sperling et al., 2011), and while it is the most common form of dementia, as of yet there is no cure. Two of the most researched pathological features contributing to disease development are the extracellular amyloid plaques composed of amyloid-beta proteins (Aβ) and neurofibrillary tangles of tau proteins. Another feature of AD is the progression of early neuronal excitability/hyperactivity to silencing/hypoactivity (Palop and Mucke, 2010), with hypoactivity explained by the synaptic failure hypothesis (Selkoe, 2002). In vitro and animal model studies have demonstrated that Aβ pathology associates with increased excitability of hippocampal neurons. Furthermore, hippocampal hyperactivity has been reported in human subjects at high risk for developing AD. In an effort to identify possible factors involved in this progression of neuronal activity in AD, we reviewed the literature and identified a novel interaction between Aβ, glycogen, neurons, and astrocytes with a focus on events in the synapse. Neuronal activity: Within the brain, Aβ is released at the presynaptic sites, and increased neuronal activity leads to higher levels of Aβ within the interstitium (Palop and Mucke, 2010). Normal levels of Aβ (pM) augment synaptic activity (Palop and Mucke, 2010; Puzzo and Arancio, 2013; Fogel et al., 2014), but higher levels of Aβ (nM) act as a negative feedback mechanism and inhibit synaptic activity (Palop and Mucke, 2010; Puzzo and Arancio, 2013). Thus, under normal conditions, elevated levels of Aβ would be expected to homeostatically inhibit neuronal activity. However, in AD there is chronic elevation of interstitial Aβ levels because of oversecretion due to neuronal hyperactivity (Cirrito et al., 2005) and reduced clearance at the blood-brain barrier (Bateman et al., 2006). This elevation of Aβ then results in misfolding and accumulation into pathological amyloid plaques (Selkoe, 2002; Palop and Mucke, 2010). These plaques then serve as a reservoir of toxic oligomeric Aβ species (Selkoe, 2002; Sheng et al., 2012; Klein, 2013) that may lead to neuronal and synaptic loss. This synaptic loss has been directly correlated to decreased cognition in rigorous studies by Scheff (Scheff and Price, 2006). Chronic elevation of Aβ leads to negative feedback (Puzzo and Arancio, 2013) and neuronal hypoactivity (Sheng et al., 2012). Synaptic energy: Metabolic dysfunction and neuronal excitability are linked, as synaptic activity is coupled to energy supply, with 97% of the energy for a resting neuron being supplied by the mitochondria. The increased energy demands of active neurons are met with an increased reliance on glycolysis (Belanger 2011) with astrocytes increasing additional nutrient supply via the glycogen-lactate shunt (Belanger et al., 2011; Stobart and Anderson, 2013). Stimulation of astrocytes results in increased intracellular calcium levels, which then result in the following: glucose uptake, vasodilation, astrocyte glycogenolysis, and release of adenosine and adenosine triphosphate (ATP) (Maragakis and Rothstein, 2006; Hertz et al., 2007; Belanger et al., 2011; Obel et al., 2012). Thus, increased synaptic activity results in increased nutrient delivery in the brain (Belanger et al., 2011). Pathologically, astrocyte dysfunction with reduction of nutrient delivery to the neurons could in part contribute to the metabolic dysfunction (decreased nutrient delivery and glucose utilization), leading to AD pathology. Glycogen: The main energy supplies in the brain are glucose and glycogen. Glycogen is primarily located within astrocytes, stimulated by insulin (Brown, 2004), and serving as the largest energy reserve in the brain (Belanger et al., 2011). Astrocytes convert glycogen stores to lactate, which is then supplied to the neurons (Belanger et al., 2011). There is a tight coupling between neuronal activity and glycogen mobilization (Belanger et al., 2011), and as such, glycogen stores correlate with cognition (Belanger et al., 2011). Importantly, astrocyte glycogen is concentrated around synapses and is known to be important for synaptic activity, and associated with greater synapse protection (Brown, 2004; Belanger et al., 2011). While glycogenolysis satisfies increased synaptic energy demands in the short term, it also occurs when glucose levels are normal (Brown, 2004; Hertz et al., 2007; Belanger et al., 2011). This is likely because the energy yield is 50% greater when using glycogen instead of glucose as a substrate for glycolysis (Hertz et al., 2007; Obel et al., 2012). Thus decreased glycogen levels correlate with decreased cognition and synaptic loss. Measurement of glycogen levels throughout the course of AD is not currently documented in the literature, but would be warranted to further understand any possible correlation between glycogen levels and cognition in AD. We propose that glycogen levels may be reduced in AD, with decreased glycogen synthesis and increased glycogenolysis occurring at the same time (Figure 1). First, glycogen synthase kinase 3β (GSK-3β) activity is upregulated in AD (Llorens-Martin et al., 2014), and because GSK-3β inhibits glycogen synthase (Brown, 2004), this will result in diminished glycogen synthesis. Tangentially, GSK-3β is also associated with tau pathology in AD (Llorens-Martin et al., 2014). Second, adenosine binding to adenosine A2 receptors (A2A) increases the rate of glycogenolysis, at least in cultured astrocytes and brain slices (Xu et al., 2014). The A2A are located on synapses where they alter metabolism, neurotransmitter release (acetylcholine and glutamate), and modulate cognitive function (Gomes et al., 2011). Of note, the pathological upregulation of A2A receptors in AD (Gomes et al., 2011) may result in depletion of brain glycogen levels. Thus the loss of glycogen is linked to decreased cognition (Stobart and Anderson, 2013) and synaptic dysfunction (Sheng et al., 2012). Figure 1 Depiction of changes in astrocytes, neurons and blood vessels in the hyper- and hypoactive states in Alzheimers disease (AD). Speculation as to the cause of initial neuronal hyperactivity: We examined possible factors involved in the progression of neuronal hyperactivity to hypoactivity in AD (Figure 1), but have not addressed the initial causation of hyperactivity. One hypothesis is that the hyperactivity compensates (Mormino et al., 2012; Elman et al., 2014) for neuronal and synaptic loss during early AD. The studies by Scheff indicate that synaptic loss in early AD, directly correlated to cognitive loss (Scheff and Price, 2006). These studies also found that although there was a loss of synapse numbers, in the remaining synapses there was an increase in the size of the total contact areas, which the authors suggested to be a synaptic compensatory mechanism in response to AD pathology (Scheff and Price, 2006). The initial neuronal loss may be due to metabolic dysfunction via reduced glucose utilization (e.g., insulin resistance in the brain, also termed Type 3 diabetes) or hypoperfusion (Beach et al., 2007; De La Monte, 2012). Metabolic dysfunction increases Aβ generation (reviewed by Murray et al., 2011), and reduced ATP levels increase formation of Aβ oligomers, at least in vitro (Coskuner and Murray, 2014). Thus the elevation of Aβ will initially lead to neuronal hyperactivity (Puzzo and Arancio, 2013), and any generated toxic Aβ oligomers will cause cell death (Klein, 2013) (Figure 1). Thus, it is quite possible that metabolic dysfunction underlies the early changes in AD, including neuronal hyperactivity and death. This suggestion is bolstered by epidemiological findings where metabolic diseases, such as obesity in middle age and type 2 diabetes, are associated with increased risk of AD (Barnes and Yaffe, 2011). Conclusion: Neuronal hyperactivity may compensate for cognitive declines in AD. Such hyperactivity results in elevated Aβ, A2A and reduced glycogen stores, which would normally feedback to reduce hyperactivity (Figure 1). Pathologically, chronic activation of such feedback would result in hypoactivity and cell death. We have used the current literature to delineate these linkages between Aβ, glycogen and neuronal activity in AD. Thus, excessive Aβ generation, astrocyte stimulation during neuronal hyperactivity, along with reduced glycogen storage, may lead to neuronal hypoactivity. We look forward to direct experimental verification of these links, which we have identified within existing publications. We thank Vern Giammartino (http://www.verngiammartino.com/) for the graphical depiction of this process. We also acknowledge and apologize that due to a limitation of number of references not all relevant publications were cited.

Evaluation of metabolic and synaptic dysfunction hypotheses of alzheimer’s disease (AD): A meta-analysis of CSF markers

Background: Alzheimer’s disease (AD) is currently incurable and a majority of investigational drugs have failed clinical trials. One explanation for this failure may be the invalidity of hypotheses focus-ing on amyloid to explain AD pathogenesis. Recently, hypotheses which are centered on synaptic and met-abolic dysfunction are increasingly implicated in AD. Objective: Evaluate AD hypotheses by comparing neurotransmitter and metabolite marker concentrations in normal versus AD CSF. Methods: Meta-analysis allows for statistical comparison of pooled, existing cerebrospinal fluid (CSF) marker data extracted from multiple publications, to obtain a more reliable estimate of concentrations. This method also provides a unique opportunity to rapidly validate AD hypotheses using the resulting CSF con-centration data. Hubmed, Pubmed and Google Scholar were comprehensively searched for published Eng-lish articles, without date restrictions, for the keywords “AD”, “CSF”, and “human” plus markers selected for synaptic and metabolic pathways. Synaptic markers were acetylcholine, gamma-aminobutyric acid (GABA), glutamine, and glycine. Metabolic markers were glutathione, glucose, lactate, pyruvate, and 8 other amino acids. Only studies that measured markers in AD and controls (Ctl), provided means, standard er-rors/deviation, and subject numbers were included. Data were extracted by six authors and reviewed by two others for accuracy. Data were pooled using ratio of means (RoM of AD/Ctl) and random effects meta-analysis using Cochrane Collaboration’s Review Manager software. Results: Of the 435 identified publications, after exclusion and removal of duplicates, 35 articles were in-cluded comprising a total of 605 AD patients and 585 controls. The following markers of synaptic and met-abolic pathways were significantly changed in AD/controls: acetylcholine (RoM 0.36, 95% CI 0.24-0.53, p<0.00001), GABA (0.74, 0.58-0.94, p<0.01), pyruvate (0.48, 0.24-0.94, p=0.03), glutathione (1.11, 1.01-1.21, p=0.03), alanine (1.10, 0.98-1.23, p=0.09), and lower levels of significance for lactate (1.2, 1.00-1.47, p=0.05). Of note, CSF glucose and glutamate levels in AD were not significantly different than that of the controls. Conclusion: This study provides proof of concept for the use of meta-analysis validation of AD hypothe-ses, specifically via robust evidence for the cholinergic hypothesis of AD. Our data disagree with the other synaptic hypotheses of glutamate excitotoxicity and GABAergic resistance to neurodegeneration, given ob-served unchanged glutamate levels and decreased GABA levels. With regards to metabolic hypotheses, the data supported upregulation of anaerobic glycolysis, pentose phosphate pathway (glutathione), and anaple-rosis of the tricarboxylic acid cycle using glutamate. Future applications of meta-analysis indicate the pos-sibility of further in silico evaluation and generation of novel hypotheses in the AD field.