n dollar level. Morphologically, the extracellular senile amyloid plaques and intracellular neurofibrillary tangles are the characteristic pathological hallmarks of AD. The former involves aggregated amyloid-beta (Abeta) peptide and the latter, hyperphosphorylated tau protein. Exactly how the two are related to one another is still a matter a debate. However, evidence supports the concept that an imbalance between production and clearance of Abeta peptides is an earlier and initiating factor. Promotion of hyperphosphrylated tau, which can no longer bind and stabilize microtubules, is a subsequent, and possibly, more damaging event. Deregulation of cholesterol/lipid metabolism, mitochondrial dysfunction, oxidative/proteostatic stresses, deregulation of immune/inflammatory responses, calcium homeostasis, intracellular trafficking and epigenetics as well as environmental and life-style factors, play/are thought to play a role in AD. No cure is currently available. In the following, important players and affected pathways, various pertinent aspects and newer players identified in genome-wide association studies (GWAS), are presented in more detail.

The amyloid cascade hypothesis
Degeneration of selective brain regions involved in cognition, such as the hippocampus, the entorhinal and frontal cortices, and areas involved in emotional behavior, such as the amygdala, prefrontal cortex, and hypothalamus are characteristic of AD. Anatomically, depositions of Abeta peptides, which are earlier events, occur in the precuneus and frontal lobes; neuronal death occurs mostly in the entorhinal cortex and hippocampus and correlates with tau pathology. The anatomical and temporal disconnect between Abeta and tau pathologies has led to questioning of the amyloid cascade hypothesis whereby the Abeta/Abeta aggregation is the causative agent prompting the subsequent events - neurofibrillary tangles, vascular damage and dementia. Current approaches point to a more complex, non-linear relationship between plaque deposition, formation of tangles and the clinical symptoms of AD. While tau phosphorylation and fibrillary tangles correlate to cognitive decline and neurotoxicity in AD, tau pathology is also evidenced in the normal aging brain. Yet, in the latter case it is confined to limbic regions, and the neocortical presence is in those people that also have the plaques. Thus, Abeta toxicity - oligomers of which are thought to be the most damaging species, is an earlier event that in some fashion prompts downstream occurrences resulting in tau pathology and neurotoxicity. The APP gene is located on chromosome 21 and in support of the role App plays in AD is the fact that humans with trisomy 21 develop AD pathology; this is not observed in individuals with partial trisomy and no triplication of the APP gene. Disruption of mitochondria and calcium homeostasis, impaired Abeta clearance and cholesterol and lipid metabolism, kinase activation, phosphatase inhibition and tau hyperphosphorylation, alterations in immune/neuroimmune response and endosomal vesicle recycling pathways, along with resulting/damage-producing oxidative and proteostatic stresses, somehow combine with/reinforce each other to prompt synapse dysfunction, neuronal death and the irreversible AD outcome.

Amyloid precursor protein (APP), APP processing and the Abeta peptides
The amyloid precursor protein (APP) is a type I integral membrane protein of rather elusive function. It is constitutively cleaved by several secretases during its maturation, with gamma-secretase playing a central role. Gamma-secretase is a complex consisting of the presenilins (PSEN1 and 2) forming the catalytic center and the PSEN cofactors nicastrin (Ncstn), PSEN enhancer-2 (Psenen), and anterior-pharynx defective-1 (Aph1a and b). Gamma-secretase belongs to the family of intramembrane cleaving proteases (i-CLIPs) and as such, it requires initial ectodomain shedding of the substrate. This is accomplished by members of the 'alpha-disintegrin and metalloprotease' (ADAM) family, known as alpha-secretases in what is referred to as the non-amyloidogenic route, as it does not lead to generation of Abeta peptides. The aspartyl protease BACE1, or beta-secretase, is involved in ectodomain shedding in the amyloidogenic route. Several Abeta peptides are derived: primarily the 40 amino acid long species, and to a lesser extent, the 42, 43 or longer Abeta peptides. The longer peptides have a propensity to self-aggregate. Several other fragments are produced, including an APP intracellular domain (AICD) generated regardless of which secretase performs the first cleavage. AICD peptides are rather unstable but several interactions and subsequent nuclear localization may affect gene expression; activation of glycogen-synthase kinase 3beta (GSK3B) expression being potentially toxic. Rare mutations in APP, PSEN1 and PSEN2 promote self-aggregation or production of aggregation-prone peptides, and are involved in the early-onset, familial form of AD (EOAD). While aggregated Abeta peptides are found in the senile plaques, it is believed that the soluble Abeta oligomers are the toxic species causing synaptic dysfunction and neuronal injury. Oligomeric Abeta has been shown to inhibit LTP (long-term potentiation) while promoting LTD (long-term depression). The strengthening (LTP) and weakening (LTD) of the synapse underlie synaptic plasticity; the latter is important for 'deleting' old memories and behavioral flexibility, and its mechanisms might be 'hijacked' in AD. Defective clearance of Abeta is thought to be involved in sporadic AD. Several genes, including APOE, represent genetic risks for sporadic AD and are thought to play a role in Abeta clearance, among others. Cholesterol, an abundant lipid particularly in myelin sheaths and glial cells, and lipid metabolism in general, are likely to play a role in the processing of APP and Abeta clearance. Gamma- and beta-secretases localize/function in cholesterol-rich lipid rafts; APOE is involved in cholesterol and lipoprotein metabolism; insulin-degrading enzyme (IDE) and neprilysin (NEP) are among the proteases that degrade Abeta and are associated with lipid rafts.

Tau protein, function, phosphorylation and dysfunction, oxidative stress
Tau, the microtubule-stabilizing protein, is composed of four domains: an acidic N-terminal domain, the proline-rich domain, the microtubule-binding region with four repeat domains R1-R4, known as microtubule-binding domains, and the C-terminal domain. The 16-exon protein is subject to alternative splicing of exons 2, 3, and 10, producing six isoforms with either three (R3) or four (R4) microtubule-binding domains. Of approximately 85 Tau phosphorylation sites, about 28 appear to only be phosphorylated in AD, 16 are phosphorylated in both AD and control brains, 31 are phosphorylated in normal, physiological conditions, and 10 sites are putative as the phosphorylating kinase has not been identified. The many kinases for which Tau is a substrate are of three types: proline-directed protein kinases (PDPK), the non-PDPK, and the tyrosine protein kinases (TPK). PDPKs target serine and threonine sites preceding a proline residue. Of the many, which include mitogen activated protein kinases (MAPK), the glycogen-synthase kinase 3 beta isoform (GSK3B) and CDK5 play a prominent role in AD-related Tau hyperphosphorylation. Hyperphosphorylated Tau can no longer bind and stabilize microtubules, and protein self-assembly generates the tangles of paired helical filaments (PHF). Several other kinases may also play a role in AD-related Tau phosphorylation. Protein phosphorylation is reversible through the action of phosphatases. Of the phosphatases targeting Tau, PP2A, PP1, PP5 and PP2B account for ~70%, 11%, 10% and 7% of the activity, respectively. However, in AD brains the activity of PP2A, PP1 and PP5 is decreased by ~50% for the first and 20% for the other two phosphatases. Oxidative stress and possibly other stresses or metabolic impairments promote GSK3B activity while inhibiting PP2A.

Apoliprotein E (APOE), APOE isoforms, cholesterol trafficking, Abeta clearance
Apolipoprotein E (APOE) has long been known as a risk factor in the etiology of sporadic, late-onset AD (LOAD) and continues to be one of the most important susceptibility genes. There are three major isoforms of APOE that differ in their 112/158 polymorphisms. E2 has cysteines at both sites and is the least prevalent in humans (~8%); E3 has 112Cys and 158Arg and is the most prevalent variant in humans (78%); E4 has arginines at both sites and relatively low prevalence in humans (~14%) but is the only form present in the great apes. It is the ApoE4 variant that is implicated in some 50%-60% of sporadic AD cases. APOE is primarily expressed in the liver, followed by the brain where it is synthesized predominantly by astrocytes. The N-terminal portion of the protein contains the lipoprotein receptor binding site and the two polymorphic positions; the C-terminal portion contains the lipid binding domain. APOE plays important roles in cholesterol transport and apolipoprotein metabolism; it is the major cholesterol carrier in the brain. APOE is also involved in the clearance of Abeta peptides, and reduced clearance of the peptide is a feature of LOAD. APOE is lipidated by ABCA1, and the lipoprotein binds to one of its receptors which belongs to the low-density lipoprotein receptor (LDLR) family, such as LRP1, abundantly expressed in neurons. The E4 isoform is less efficient in Abeta clearance. APOE, in an isoform-specific manner, is involved in a range of other perturbations, such as synaptic and mitochondrial dysfunction, neuroinflammation and cognitive impairment, among others.

Altered cholesterol/lipid metabolism, vesicle endocytosis and susceptibility genes
The brain is the most cholesterol-rich organ. Cholesterol is abundant in the myelin sheath followed by the plasma membrane of astrocytes, microglia and neurons; it plays a pivotal role in synaptic and neuronal function. As a major carrier of cholesterol and chaperone for Abeta clearance, APOE and its isoforms can profoundly impact the function/dysfunction of synapses and neurons. Other genes important for cholesterol metabolism have been identified as LOAD susceptibility genes in GWAS studies. ApoJ, which encodes clusterin, carries out functions similar to APOE. The LRP2 receptor appears to be involved in internalizing ApoJ-Abeta complexes. The risk allele appears to be associated with alternative splicing of the Apoj gene. Sorl1 functions as a receptor for lipoproteins, including Apoe, promoting their endocytic uptake. Another AD susceptibility gene that may be associated with cholesterol metabolism is Abca7, but its exact function remains to be established. Though not with a direct role in cholesterol metabolism, but dependent upon membrane association, are the endocytic and synaptic vesicle endocytosis pathways, the latter proceeding in a clathrin-dependent and independent manner. Among the GWAS-identified susceptibility genes are Picalm and Bin1, both involved in clathrin-dependent endocytic processes. The Sorl1 gene is also thought to play a role in endocytosis and sorting.

Neurovascular coupling, astrocytes, microglia, neuroinflammation and susceptibility genes
Neurovascular coupling, also known as functional hyperemia, is the provision of an adequate supply of oxygen and nutrients in response to neuronal activity; it relies on an inter-cellular signaling network of neurons and astrocytes, smooth muscle and endothelial cells of cerebral microvessels - with astrocytes providing a bridge. The astrocyte end-foot covers more than 99% of cerebral capillaries. Astrocytes are also involved in potassium ion buffering, maintenance of the blood brain barrier (BBB), provision of structural support to synapses and a protective role at tripartite synapses. They release antioxidants, e.g. glutathione, and take up excess glutamate. A model of neurovascular coupling posits that glutamate activates metabotropic glutamate receptors on astrocytes, prompting release of calcium and downstream events, resulting in cyclooxygenase-mediated arachidonic acid metabolism and prostanoid biosynthesis. PGI2, known as prostacyclin, is a primary prostaglandin of endothelial cells; signaling through its G-protein-coupled receptor elicits a potent vasodilation effect. Of note is the fact that astrocytes are also the site of the de novo biosynthesis of cholesterol; they are also an important site for the expression of Apoe, the cholesterol carrier involved in Abeta clearance, and of Clu/ApoJ - a GWAS-identified AD susceptibility gene with roles similar to Apoe. Amyloid plaques are surrounded by reactive astrocytes and activated microglia. Microglia - the immune cells of the nervous system, degrade soluble beta amyloids and phagocytize amyloid fibrils. IDE and NEP, the Abeta-degrading enzymes, are present in microglia. Microglia can show enhanced sensitivity to inflammatory stimuli, known as priming, conducive to increased production of cytokines and ROS. This can result in activation of astrocytes which then secrete specific cytokines in what can become a vicious neuroinflammatory cycle. Neuroinflammation is a hallmark of AD; among the GWAS-identified susceptibility genes are the microglial TREM2, CR1 and CD33; astrogliosis is evident in AD and increases with cognitive decline; neurovascular coupling is impaired in aging and in AD; hypoperfusion in AD is reported. In addition to hypoperfusion, hypometabolism of glucose - the main fuel for the brain, is also observed in AD. Studies show that important glycolytic enzymes are modified in ways indicative as being ROS-mediated.

Mitochondrial homeostasis
Mitochondrial dysfunction is associated with the pathogenesis of AD. Evidence of structural changes such as fragmented mitochondria with abnormal cristae, impairment of bioenergetics, diminished enzyme activities (cytochrome oxidase, α-ketodehydrogenase, pyruvate dehydrogenase), decreased adenosine triphosphate (ATP) synthesis, mitochondrial membrane depolarization, increased ROS production and altered mitochondrial biogenesis have been reported. Mitochondria move along cytoskeletal tracks, fuse and divide. Proper mitochondrial distribution is particularly important in cells, such as neurons, and is maintained by large GTPases: dynamin related protein 1 (Drp1) and optical atrophy 1 (Opa1). Drp1 and Opa1 together with mitochondrial fission factor (Mff), mitochondrial fission protein 1 (Fis1), mitofusin-1 and mitofusin-2 (Mfn1, Mfn2) proteins regulate the assembly and stability of the respiratory chain supercomplexes which affect the proper formation and function of synapses. Synaptic degeneration processes like excessive division, defective axonal transport of mitochondria and increased DRP1-mediated mitochondrial fission have been observed in AD patients¿ neurons and animal models of AD. A mitochondrial motor/adaptor complex consisting of kinesin, dynein, proteins Miro and Milton enables mitochondrial movement. Kinesin-1 (KIF5) is the major driving force behind the transport of neuronal mitochondria in the anterograde direction, so the loss of KIF5A and KIF5B isoforms promotes mitochondrial mislocalization. Defects in mitochondrial transmission at Miro may contribute to the pathogenesis of Alzheimer¿s disease, as knockdown of Milton or Miro enhances Tau-induced neurodegeneration. The Miro level is downregulated in the presenilin 1 E280A mutation that is associated with familial AD. Interestingly, overexpression of the kinesin/Miro/Trak2 interactor Armc10 can prevent Abeta-induced mitochondrial fragmentation. The mitochondrial cascade hypothesis proposes that inherited mutations in mitochondrial DNA (mtDNA) determine the ability of mitochondria to respond to and recover from stress signaling that is mediated by molecules such as reactive oxygen species (ROS). The mtDNA is particularly vulnerable to ROS-induced mutations and lesions. Progressive accumulation of mtDNA mutations with age leads to a loss of efficiency of the Electron Transport Chain (ETC), decline of bioenergetic capacity, pathologies and death. The histopathology of AD develops after reaching a critical threshold of mitochondrial dysfunction, including neuronal apoptosis, Abeta deposition, and neurofibrillary tangles (NFT). Most ATP is produced in the mitochondrial ETC. Electrons are transported through the protein complexes to molecular oxygen, and protons are pumped by complexes I, III, and IV across the inner mitochondrial membrane to generate an electrochemical gradient. Abeta and abnormally hyperphosphorylated Tau protein may act synergistically to trigger mitochondrial toxicity in AD. Abeta impairs the mitochondrial membrane potential, primarily decreases complex IV and later complex I activity, decreases the production of ATP, increases the levels of ROS, whereas the abnormally hyperphosphorylated Tau impairs complex I activity in the first place. Abeta and Tau proteins disturb mitochondrial dynamics. Abeta elicits mitochondrial transport defects and fragmentation. Tau hyperphosphorylation induces dissociation of microtubules and impairs the transport of distinct cargoes along neuronal axons, including mitochondria. Hyperphosphorylated Tau reduces mitochondrial fission, which induces an elongation of the mitochondrial network. Oxidative stress is considered as a primary factor for phosphorylation and formation of NFT early in the pathogenesis of AD. Oxidative stress may trigger a circular chain of events including Abeta generation, Tau hyperphosphorylation and neuroinflammation leading to neurodegeneration. The mouse model of AD shows increased proteolytic cleavage of APP, increased production of Abeta, and impaired Cu/Zn-SOD activity. Mitochondrial SOD2 deficiency, like chronic oxidative stress, increases the levels of Ser396- phosphorylated Tau in the mouse model of AD and in neuroblastoma cells, respectively. The effects of Abeta and Tau on mitochondrial function are synergistic and age-associated, resulting in a decrease in mitochondrial respiratory capacity and reduction of ATP synthesis, which finally lead to synaptic loss and neuronal death. Uncoupling proteins (UCPs), a family of mitochondrial anion carrier proteins, are located on the IMM. UCP2 and UCP3 can be activated by ROS or products of lipid peroxidation to diminish proton motive force and reduce mitochondrial membrane potential and ATP production. The activation of UCPs is considered to be a protective mechanism in response to oxidative stress. In AD brains the expression of UCP2, 4, and 5 is significantly decreased. Most evidence demonstrates that oxidative stress is linked with Abeta-induced neurotoxicity, Tau pathology, mitochondria dysfunction, and metal dyshomeostasis.

Epigenetics
Brain regions affected by AD pathology show global alterations in DNA methylation (5mC) and hydroxymethylation (5hmC) in the early stages of AD, but not in later stages of disease. Analysis of AD-associated CpG sites in 34 patients discovered 479 differentially methylated regions (DMRs), most of which were found to be hypermethylated. Of the 25 most significant DMRs, eight genes (LOC100507547, PRDM16, PPT2, PPT2-EGFL8, PRRT1, C10orf105, CDH23, and RNF39) had been previously reported in the first AD epigenome-wide association studies (EWAS). The most significant AD-associated DNA methylation has been identified in the ANK1 gene in a cohort of 122 individuals. Genetic variation in ANK1 has been previously associated with diabetes and recognized as an increased risk factor for developing dementia and AD in patients with type 2 diabetes. The recent EWAS study determined DNA methylomic profiles associated with amyloid plaques in 708 individuals, and identified 71 differentially methylated CpG sites associated with neuropathology, including RNF34, CDH23, SLC2A1, COQ7 genes and the HOXA gene cluster. In addition, ANK1, CHD23, DIP2A, RHBDF2, RPL13, SERPINF1 and SERPINF2 genes show significant AD-associated gene expression changes. The current EWAS study's limitation is that 5mC and 5hmC are indistinguishable.

Protective/neurotoxic roles of miRNA
Micro RNAs (miRNAs) play a large role in the regulation of genes, that are responsible for Abeta production and phosphorylated Tau, including APP, presenilin-1, and presenilin-2. miRNAs also regulate cellular changes via the ApoE4 genotype and other polymorphisms, including sortilin-related receptor 1, clusterin, complement component receptor 1, CD2AP, CD33, EPHA1, and MS4A4/MS4A6E genes that are involved in AD pathogenesis. RNA sequencing analysis has revealed many miRNAs that are brain-specific, including: miR-134, miR-135, let-7g, miR-101, miR-181a-b, miR-191, miR-124, miR-let-7c, let-7a, miR-29a and miR-107. Most of these miRNAs are responsible for synaptic functions, neurotransmitter release, synapse formation and neurite outgrowth. A review of more than 50 miRNAs expressed in biofluid and brain of AD patients shows that most of miRNAs are downregulated. Reduced levels of miR-9, miR-29, miR-107, miR-124 and miR-339 correlate with elevated BACE1 expression, which promotes the formation of Abeta from APP. Upregulation of miR-29c promotes learning and memory behaviors in a mouse model for age-related dementia (SAMP8) by increasing the activity of the protein kinase A/cAMP response element-binding protein, which is involved in neuroprotection. Overexpression of miR-29c or administration of recombinant pre-miR-29b reduces the levels of BACE1 and Abeta peptides in human neuroblastoma cells (SH-SY5Y) and neuronal cell culture, respectively. miR-29a disrupts the activity of another target gene encoding neuronal navigator 3, a protein that is involved in axonal guidance and is enriched in degenerating pyramidal neurons in AD. Administration of miR-339-5p reduces BACE1 and Abeta peptides in human neuroblastoma and primary brain cultures. Decreased levels of miR-107 and miR-339-5p also correlate with enhanced levels of neuritic plaque and neurofibrillary tangles in the autopsied brain cortex of AD patients. miR-34, miR-107, miR-124, miR-132, miR-137, miR-148a, and miR-339-5p are involved in Tau phosphorylation, APP processing, neuronal differentiation, neuroinflammation, and the cell cycle. Expression of a miR-34c isoform increases in the hippocampus of AD patients and aging mice and suppresses hippocampal sirtuin 1 (SIRT1), which negatively impacts memory formation. Treatment with Abeta peptides increases the expression of miR-34c, that in turn decreases vesicle-associated membrane protein 2 (VAMP2) levels and disrupts synaptic function, learning and memory - effects that can be reversed by silencing miR34c expression. miR-34a targets TREM2 expression and is upregulated in brain tissue and blood mononuclear cells of AD patients. Mutations in TREM2 cause rare, autosomal recessive forms of early onset dementia. miR-34a-mediated down-regulation of TREM2 expression may impair phagocytic responses and contribute to amyloidogenesis and inflammatory degeneration in the AD brain. Elevated levels of miR-26b, miR-125b and miR-138 in the autopsied brain tissue of AD patients correlate with hyperphosphorylation of Tau protein. Upregulated miR-26b decreases the level of its target retinoblastoma protein that leads to activation of a major kinase, CDK5, and causes hyperphosphorylation of Tau protein. Increased expression of miR-138 promotes Tau phosphorylation by targeting the retinoic acid receptor-alpha (EARA)/glycogen synthase kinase-3β pathway. Upregulated miR-125b increases hyperphosphorylation of Tau protein by decreasing phosphatase activities in AD patients. In primary neuron cultures, overexpression of miR-125b decreases the levels of dual-specificity phosphatase 6 (DUSP6) and protein phosphatase1 catalytic subunit alpha (PPP1CA), downregulated also in the brain tissue of AD patients. Elevated miR-125b also downregulates synaptic protein synapsin-2 (SYN-2) and 15-lipoxygenase (15-LOX), thereby potentially causing pathogenic effects of AD such as synaptic and neurotrophic deficits and astrogliosis. miR-125b regulates factors involved in the innate immune system and in the pro-inflammatory response, such as repression of complement factor-H protein (CFH) and downregulation of interferon regulatory factor 4 (IRF4). Another highly upregulated proinflammatory miRNA in AD brains is miR-146a, that has multiple binding sites in CFH mRNA 3′UTR (just like miR-125b) and thus also potentially downregulates CFH. Mutation in presenilin 2 (PS2), that causes autosomal dominant AD, increases the pro-inflammatory responses of microglia cells in culture. Dysfunction of PS2 causes a decrease in miR-146a that results in the increase of its target proteins IL-1 receptor-associated kinase-1 and an increase in microglia cell NF-κB transcriptional activity and pro-inflammatory cytokine expression. The ApoE4 genotype is a major risk factor for late-onset sporadic AD. Overexpression of ApoE4-linked miR-33 impairs cellular cholesterol efflux and increases extracellular Abeta levels by interfering with the clearance of Abeta in neurons. Pharmacological inhibition of miR-33 markedly decreases Abeta levels in the cerebral cortex of APP/PS1 knockout (-) mice. Recent advances in understanding the mechanism of miRNA gene regulation and the role of altered expression of specific miRNAs contribute to the discovery of diagnostic biomarkers for AD and novel therapeutic strategies. To see the ontology report for annotations, GViewer and download, click here. [Click to see KEGG map - hsa05010 and related entry at Reactome - R-HAS-8862803.]...(less)