Archives of Neurology and Neurosurgery (ANN)

Transgenic Mice are not Valid for Late Onset Alzheimer’s Disease



Jordan L Holtzman 1*


1Division of Environmental Health Sciences, School of Public Health, University of Minnesota, Minneapolis, Minnesota.


*Corresponding Author: Jordan L Holtzman, Professor, Division of Environmental Health Sciences, School of Public Health, University of Minnesota, Minneapolis, Minnesota, TEL:+1 612-626-9822 ; FAX:+1 612-626-9822 ;E-mail:holtz003@umn.edu


Citation: Jordan L Holtzman (2018) Transgenic Mice are not Valid for Late Onset Alzheimer’s Disease. Archiv Neurol Neurosurgery 2:107.


Copyright: : © 2018 Jordan L Holtzman, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited


Received date:March 14, 2018; Accepted date:April 02, 2018; Published date:April 05, 2018


Abstract

Alzheimer’s disease is a major cause of morbidity and mortality in the elderly. It is characterized by a cognitive decline and the presence in the cerebral cortex of extracellular deposits of β-amyloid (Aβ) in plaque and intracellular deposits of an hyperphosporylated microskeleton protein, tau, in the form of tangles. It is currently thought by a majority of workers in the field that the cognitive decline is due to either the neurotoxicity of the Aβ or the tangles. The early onset familial forms of the condition have a similar pathology and clinical course but are due to mutations in three proteins: amyloid precursor protein and presenilins 1 and 2. None of these mutations are found in the elderly. Yet in spite of their absence most workers believe that the cognitive decline seen in the early and late onset forms have the same underlying biochemical basis. As a result, they have developed mouse models transfected with mutant forms of the proteins associated with the early onset disease. These mice have been used to screen for potential therapeutic agents. Yet all 244 drugs that were found to be effective in these mice failed to show any benefit in clinical trials in elderly patients, including antibodies to the Aβ and inhibitors of the β- and γ-secretases which are thought to be critical for plaque formation. These failures raise the question of whether the mouse models are valid test systems for drug discovery for treatment of the late onset disease. In this review I have outlined the biochemistry of Aβ production and studies from my laboratory that suggest that the cognitive decline is due to a decrease in the posttranslational processing in the endoplasmic reticulum of the synaptic membrane proteins that are necessary for a functioning memory. These findings suggest alternative models for the development of test systems for drug discovery to treat the late onset disease.


Keywords

: Familial Alzheimer's disease, Late Onset Alzheimer's disease, Drug discovery, Clinical trials, Protein folding, Endoplasmic reticulum, N-glycosylation, CSF


As is widely recognized with the aging of the world's population, late onset Alzheimer's disease has become a major cause of morbidity and mortality in the elderly. In fact, in the United States it is now the third leading cause of mortality after cardiovascular disease and cancer. It also represents one of the leading financial burdens on our health care system. This burden, along with the stress which this disease places on both the patients and their families, has led to a consensus that there is a great need for research and development in the search for effective therapies. Recognizing the market potential of such treatments, the pharmaceutical industry has poured massive sums into drug development to fill this growing need. Yet, in spite of these vast sums, they have yet to develop any robust therapies to treat this condition. The magnitude of this failure is illustrated by the report by Cummings et al. (2011). These authors searched http://www.c1inicaltrials.gov/for all phase I-III trials in late onset Alzheimer's disease patients. None of the 244 agents they identified slowed the progression of the disease in patients with mild to moderate Alzheimer's disease. This consistent failure raises the question of whether the animal models currently used for drug screening are actually appropriate for seeking therapies for the late onset disease. He has recently updated his review with similar findings [1].


Currently the gold standards for drug development in Alzheimer's disease are mouse models transfected with genes associated with the familial forms of the disease. The construction of these models is based on the proposition that the cognitive decline is due to the same biochemical mechanism in the early and late onset diseases. In spite of the failure of this approach to lead to the development of therapies that have a robust effect on the cognitive decline, both the academic and industrial communities have shown little interest in taking new approaches to develop screening systems for drug discovery. It is not surprising that the academic community is taking a conservative approach since their funding depends upon a variety of peer review processes. Reviewers tend to be very conservative and deny funding to programs that stray too far from the currently accepted paradigm. On the other hand, the pharmaceutical industry faces the opposite problem. Their funding depends entirely upon their ability to patent and produce marketable products. At the present time rather than directly funding innovated approaches, they have generally used their substantial funds to buy small biotech companies that have potentially marketable products. This approach has worked well in those areas where the cell biology of diseases, such as autoimmune, cancer or cardiovascular diseases, are well understood. But in a field like the treatment of late onset Alzheimer's disease, it has yet to yield any major breakthroughs. Part of the problem is that most of the startups are extensions of well-known academic laboratories which have been well funded through research grants. As a result, their approach is based on the current paradigm, which in this case has yet to produce any agents which halt the progression of the disease in patients with mild to moderate cognitive loss. This narrow focus is a result of the historical development of our understanding of Alzheimer's disease.


Although presenile dementia has been recognized since ancient times, it was not until 1906 when Alois Alzheimer, a German psychiatrist, described a patient, Auguste Deter, who was a 55-year-old woman who had had been institutionalized for several years as a result of disruptive behavior and cognitive decline. At autopsy Alzheimer employed staining techniques which he had developed with Franz Nissl. He described two neuropathological lesions characteristic of what was then called presenile dementia: extracellular deposits and intracellular tangles. In 1910 Emil Kraepelin, a German psychiatrist who worked with Dr. Alzheimer, first named this condition "Alzheimer's Disease" in the eighth edition of his book Psychiatrie.


Concurrent with Alzheimer's work Oskar Fisher, an investigator at the German University of Prague correlated the clinical course in a large number of patients with a variety of neurological diseases with the clinical findings and the observed neuropathology at autopsy. He concluded that the dementia of late onset disease was a result of plaque deposition [2]. Interestingly Alzheimer believed that these deposits were merely a biomarker for this condition. Their difference of opinion on the role of plaque in the development of dementia is a question that is still relevant today.


Robert Katzman [3], a well-known neurologist from the Einstein School of Medicine in the Bronx, NY, published an editorial in the Archives of Neurology in which he contended that based on the neuropathology and the clinical course, both presenile and senile dementia were the same disease described by Alzheimer. He further argued that since the late 1940's this was a widely held opinion among neurologists.


Beginning in the 1980's the core protein of the plaque, β-amyloid (Aβ) was purified and sequenced from the brains of patients with both late onset Alzheimer's disease and Down's syndrome [4,5]. Employing this sequence data both Tanzi et al. [6] and Goldgaber et al. [7] mapped the Aβ sequence to chromosome 21. Based on these reports Kang et al. [8] published a full sequence of the amyloid precursor protein (APP). APP is a highly conserved, single transmembrane protein which is a member of a family of proteins found in all animal species from C. elegans to man. These proteins, including Amyloid Precursor Like Proteins (APLP) 1 and 2, are growth factors that are necessary for normal neuronal development [9]. But it is unlikely that the Aβ portion of the APP sequence serves any important physiological function since it is not found in the APP of most species. The APLP proteins rather than APP are more generally found in various animal species. For example, Drosophilia expresses a full-length member of the APP family protein which is required for normal neuronal development [10].


Furthermore, knocking out APP alone leads to some loss of cognitive skills, but is not lethal [11]. These mild deficits are well tolerated since other members of the APP family can promote neuronal growth. Knocking out any two of the three members of the family is a neonatal lethal [12].


The Metabolism of APP

In order to understand the current studies in Alzheimer’s disease, it is first necessary to understand the origin of Aβ. As indicate below our studies indicate that it is produced during the posttranslational processing in the endoplasmic reticulum (ER) of a single transmembrane protein the amylod prcursor prtoein (APP) [13].


Figure 1
Figure 1: The aminoacid sequence of the amyloid precursor protein.

In vertebrate species the full-length APP is converted to one of two active soluble forms, sαAPP and sβAPP through cleavage by one of two proteases, α-secretase or β- secretase. The cleavage site for the α-secretase is 16 amino acids down from the cleavage site of the β-secretase. Since the sequence for a third of Aβ is between these two sites, the full-length peptide is only formed when APP is cleaved by the β-secretase to give sβAPP. The β-secretase is a membrane bound aspartyl protease, BASEl (Haas et al. 2012). If the cognitive decline is due to the neurotoxicity of Aβ, then inhibiting, BACEI would seem to be a logical treatment for Alzheimer’s disease. In support the role of β-secretase activity in the development of the cognitive decline Jonsson et al. [14] found that in a genome wide sequencing study of 1795 Icelanders that a (A973T) mutation in APP, which inhibited the β-secretase activity by 40%, appeared to be protective against the cognitive decline. Although this mutation was found in a scattering of other Scandinavian subjects, it was not observed in a North American cohort [15]. By inhibiting the β-secretase, these individuals would produce more of the αsAPP. The decreased incidence of cognitive decline could be due to the sift in the production from sβAPP to sαAPP. In light of the possible role of Aβ and sβAPP in the cognitive decline, several pharmaceutical firms are developing small molecular inhibitors of BACEl. There currently are three phase III trials of BACEl small molecules inhibitors in clinical trials in progress and several others in phase I [16]. Merck has reported that their BACE1 inhibitor decreased plaque deposits and was well tolerated.


One problem with inhibiting BACEI is that this protease has other targets besides APP [16-19] (Laird et al. 2005; De Strooper et al. 2010; Prox et al. 2012; Vassar 2014). BACEI knockout mice have stunted growth and elderly knockout mice have decrease cognitive skills [17,20]. Hu et al. [21] reported significantly greater deleterious effects of knocking out BACEI, including seizures in up to 20% of the mice which was associated with an increase in the Nav1.2 sodium channels. Approximately 30% of the mice died neonatally. These data clearly indicate that care must be taken in dosing patients to prevent this type of toxicity. As noted above, Merck has reported that in a phase I study their BACE1 inhibitor, verubecestat, has no significant adverse side-effects. Unfortunately, more recently they halted a phase III trial of the drug on advice of the Statistical and Safety Monitoring Board because of futility.


The α-secretase activity is catalyzed by a family of proteases A Disintegrin And Metelloprotease (ADAM). Three members of the ADAM family have been identified by inhibitor and transfection studies as the active proteases, ADAM 9, 10 and 17. Even though it is generally thought that the activation of the α-secretase blocks Aβ production, there are only a few studies seeking drugs that increase the activity of this secretase [22-24] and none are in clinical trials.


The Familial Forms of Alzheimer’s Disease.


The familial forms of early onset Alzheimer’s disease are associated with mutations in three proteins: APP, presilin1 and presenilin2. The first of these, APP, which is encoded on chromosome 21 was found to be associated with 29 variants of familial Alzheimer's disease [25]. Furthermore, linkage genetic studies demonstrated that the early onset disease resulted from a variety of mutations in APP. Two other proteins were also associated with early onset disease, presenilin I (psenl) on chromosome 14 and presenilin 2 (psen2) on chromosome 1. The presenilins are components of the protease, γ-secretase. This a plasma membrane bound protease that is involved in the excision of the membrane portion of Aβ, It is also involved in cell signaling. In particular the γ-secretase is involved in notch signally. Notch is a membrane-bound receptor for integrans [26]. When activated it activates the γ-secretase which clips off the intracellular portion of the receptor that then acts as a transcription factor.


The γ-secretase consists of four components including the two presenilins and two structural proteins [20]. The presenilins are the protease catalytic sites. The associated proteins are nicastrin and APH-1. Subsequent genetic studies have identified a large number of variants in each of the mutant genes. There have been reported to be 166 of the psenl gene and ten in the psen2 gene [27]. The development of small molecule inhibitors of γ-secretase blocked the cognitive decline in transgenic mice, but not in patients with late onset disease. In fact, in one trial the experimental group exhibited decreased cognitive function compared to the placebo group [28]. This finding is not surprising since this protease catalyzes the activation of Notch an important plasma membrane receptor and inhibitors of Notch have been studied as potential cancer therapeutic agents.


Over half the cases of early onset disease are due to the various mutations in the psen1 gene. These genes are dominant negatives, Hence a haploid mutant lead to the development of an early version of the familial form of Alzheimer's disease.


In the case of the three genes associated with the early onset disease all of them lead to an increase production of Aβ. As a result of this association Selkoe [29] has proposed that the dementia found in Alzheimer's disease was due to the neurotoxicity of the Aβ. He proposed that the deposits and soluble aggregates of this peptide were either directly toxic or led to an inflammatory response that, in either case, caused a progressive neuronal injury which leads to the cognitive decline. It is not clear why these three mutations lead to an increased production of APP.


Several workers have identified a number of other problems with the β-amyloid model. These have included a poor correlation between plaque burden and the development of dementia; the most dramatic issue concerning the toxiticty of Aβ [30,31] In spite of the questions which have been raised the majority of the workers in the field still accept the Aβ paradigm. They have raised the question of whether Aβ toxicity is due to the deposits or soluble aggregates. Irrespective of which of the forms of Aβ is toxic they still claim that transgenic mice are valid models for the disease.


The most widely accepted suggestion for the failure of monoclonal antibodies which clear plaque but fail to decrease cognitive decline is because these trials were initiated after the patients’ disease had progressed to a point where they could no longer benefit from therapy [32-36]. In line with this concept Biogen and Roche are currently conducting clinical trials with monoclonal antibodies to examine the question of whether beginning therapy early in the disease can improve outcomes.


The Role of Apolipoprotein E4 in Late Onset Disease

Another area of concern is that until recently the only genetic variant which has been consistently associated with an increased incidence and earlier onset of late onset cognitive decline is the presence of at least one allele of apoE4 [37,38]. The odds ratios for the development of dementia for the various genotypes range from 0.6 for the E2/E2 to 14.9 for the E4/E4 genotype [39]. Yet, the estimated attributable risk ranges from 0 to 95%, with the usual estimates being between 30% and 50% [37,40,41]. This increased risk differs markedly among various populations [39,42].


Since apoE is a component of all of the major plasma lipoprotein fractions and the apoE4 genotype has also been associated with an increased incidence of cardiovascular disease, one proposal for this effect of the apoE gene-isotype has been that the increased incidence of dementia is due to increased cerebral vascular disease [43]. Other suggestions such as decreased clearance of plaque [44] or binding of apoE4 to Aβ [45] would seem to be unlikely because they do not indicate how this leads to the initial accumulation of the deposits. Furthermore, apoE4 does not appear to be a primary binding protein for the naked peptide in the CSF, but rather it has been reported in immunoprecipitation studies that Aβ added to CSF primarily binds to apolipoprotein J, also known as clusterin [46]. It is of interest that in several Genome Wide Association Studies (GWAS) SNP’s in this gene appear to be another gene, along with apoE4, that has been frequently associated with an increased incidence of Alzheimer’s disease [27]. Furthermore, in a recent publication from the Rotterdam study, it has been reported that clusterin levels are higher in individuals with Alzheimer’s disease than in normal subjects [47] suggesting that this binding protein may be induced by the presence of the naked Aβ- peptide and serves to blunt any possible Aβ toxicity.


Yet, new data from Roses’ laboratory has suggested an alternative mechanism whereby the gene load of apoE4 could affect the development of dementia [48]. In a GWAS they reported that the best correlation between the early onset of the late onset dementia was not with the apoE4 allele, but rather with a tightly linked gene for TOMM40/TOM40 which is about 30 kb upstream from the apoE gene. In these studies, it was found that the incidence and time of onset of the cognitive decline was associated with the length of poly-T runs in intron 6. TOMM40/TOM40 is a component of a protein transport pore in the outer membrane of the mitochondria (Humphries et al. 2005). The bulk of mitochondrial proteins are synthesized in the cytosol and transported into the mitochondria through this pore.


Valla et al. [49] reported that mitochondria from individuals with the apoE4 allele have decreased cytochrome oxidase activity, suggesting that defects in this pore lead to compromised mitochondrial function. This lower activity may be due to decrease content of TOMM40/TOM40. The lower TOMM40/TOM40 content is most likely not due to a decrease in the life span of the mRNA since these poly-T’s do not code for any miRNA’s which might regulate its half-life. It is possible that the poly-T runs could lead to alternative splicing of the mRNA’s which in turn would lead to truncation of the TOMM40/TOM40 protein and possibly decreased function. Alternatively, it could just decrease the transcription of the gene.


Figure 2
Figure 2: The TOM complex for the transport of proteins from the cytosol to the mitochondria.

In G W AS studies, another gene, clusterin, which appears to be associated with late onset disease. This is a plasma chaperone. Low levels of this protein in blood are associated with early onset of late onset Alzheimer's disease [50-52]. Low levels of clusterin is a potential biomarker for Late Onset Alzheimer's disease.


This raises an important issue. In view of the recent finding that the treatment of early disease with monoclonal antibodies to Aβ slows progression of the cognitive skills several pharmaceutical firms are involve in clinical trials to ascertain whether anti-Aβ monoclonal antibodies will slow the cognitive decline in the treatment of patients with relatively mild cognitive impairment. A major concern in testing these antibodies is to identify patients who are likely to benefit from this therapy. Currently potential patients are identified by radioactive Aβ brain scans this is both expensive and has a low sensitivity. A more efficient method would be to identify a biomarker that might be less expensive and more sensitive. The recent observational findings would suggest that clusterin may be a more robust biomarker that might improve identification of potential patients who are at an early stage of the decline. This is not consistent with the findings of the Rotterdam study [47]. One approach has been to look for genetic markers that could classify patients who could benefit from antibody therapy. Finally, Lilly has recently announced that it has halted its phase III trial of the Aβ monoclonal solenazumab for a lack of efficacy (Doody et al. 2014).


In line with this approach four large QWAS studies of patients with late onset Alzheimer's disease have identified several genes that may be associated with an increased incidence of the disease. Yet, most of the genes which were reported to have possible SNP's in all of these trials the only two that were identified in more than one trial were apol E4 and clusterin. Early purification and mass spectrographic studies reported by Ghiso et al. [46] found that when Aβ was added to the CSF, it is specifically and tightly bound to clusterin. Since then a number of groups have examined the association between clusterin and Aβ. Furthermore Jongbloed [51] reported that low plasma levels of clusterin were associated with an increased incidence of late onset cognitive decline. This relationship between clusterin levels and cognitive decline is consistent with the Aβ model since clusterin inhibits the nucleation of soluble Aβ to form aggregates. The high affinity of clusterin for Aβ has been reported by several groups in more recent studies.


A surprising finding is that clusterin infused into the right ventricle of the brain prevented traumatic brain injury in mice [53].


As an aside it is of interest that Muller et al [54] were able to obtain the blocks from Alzheimer's original patient and extract her DNA. They reported a single mutation in the psenlinl gene. Rupp et al. [55] attempted to replicate their finding, but reported that the psenlinl gene was normaL


Neurofibular Tangles


The other deposits found in the brains of Alzheimer's patients are tangles. These are aggregates of a neuronal cytoskeleton protein, tau. [56-58]. The protein kinase GSK3β phosphorylates several serines and threonines on tau. The hyperposphorylated tau then aggregates to form intracellular deposits termed tangles. Knocking down GSK3β reduces plaque formation [57]. Furthermore, activation of this kinase increases the levels of BRACE 1 [58] and thereby increasing the level of Aβ production. Of interest studies by Professor Wischik at the University of Edinburgh have suggested that the urinary analgesic, methylene blue clears tangles in the transgenic mice. In an early phase I trial, he found that a low dosage of this agent completely prevented cognitive decline. But unfortunately, in an unpublished phase III trial in patients it has reported that the drug had no effect on the progression of Alzheimer’s disease [59].


Our Findings on the Biochemistry of AB.

The deposition of plaque in the elderly presents an interesting question: If everyone produces Aβ; why are deposits only seen in the elderly? Aβ is the major component of the plaque observed in the brains of patients with Alzheimer's disease. The conundrum is that although Aβ is produced in everyone during the posttranslational processing in the endoplasmic reticulum (ER) of the Aβ, deposits are only observed in the brains of the elderly.


Immunoblotting of the Aβ Complex in the CSF


Our work suggests that normals have a carrier protein(s) that keeps Aβ in solution. Based on immunoblotting studies of cerebrospinal fluid (CSF) from cognitively intact subjects, we find that the bulk of the Aβ is N-glycosylated and bound to two ER chaperones, ERp57 and calreticulin, suggesting that these may be the carrier proteins which prevent aggregation of Aβ and that the deposits are due to faulty ER posttranslational processing of APP with the failure to form the soluble complex. If, as is likely, membrane protein synthesis is similarly affected, it could explain the neuronal dysfunction characteristic of Alzheimer's disease.


Since a portion of the CSF is derived from the extracellular fluid, it is the best approach to determine whether there is increased Aβ production in Alzheimer's patients is to determine their CSF concentrations. Yet, when this has been done, it has been found that the total concentrations of Aβ in the CSF of late onset Alzheimer's patients was equal to or less than that found in normals [60] indicating that factors other than the rate of Aβ production lead to plaque formation. This is unlike early onset familial disease in which the Aβ in the CSF is actually elevated.


For example, one factor could be that Alzheimer's patients have a nucleation factor which enhances plaque deposition. This may be the basis for the increased plaque seen in individuals with at least one apoE4 allele [61] since under oxidizing conditions apoE4 enhances Aβ precipitation more than do the other two isoforms.


Since APP undergoes posttranslational processing in the ER, we posited that Aβ could be bound to one or more ER chaperones. To examine this possibility, we examined CSF samples from a number of cognitively normal patients obtained at the time of surgery for prostate problems.


On western blotting we found that all the Aβ in the CSF was found in a band at 62 Kda (Figure 3b). Similarly, one ER chaperone ERp57 appeared as a doublet with the upper band that also had an Mr=62 Kda which we presumed to be an ERp57-Aβ complex (Figure 3a)


Figure 3
Figure 3: Immunoblots with A) chicken polyclonal antibodies to ERp57 and B) rabbit antibodies to Aβ.

We previously observed this same doublet when we initially purified ERp57 from rat liver [62]. Unfortunately, we never went on to separate the two bands, obviously even though Aβ and ERp57 have similar Mr’s doesn’t prove that they are present as a complex. Hence, we went on to imunopurify the complex and found that the two proteins copurify the identity of the abeta band in the CSF has been confirmed with three monoclonal antibodies, Mab1-17, Mab1-42a and Mab1-42b, and our chicken antibody. On the other hand, we could not detect the complex when we used the monoclonal antibodies, Mab17-24, (4G8) even with very high concentrations of antibody (≥ 1 mg/mL) suggesting that the binding of Mab17-24 to the abetas was sterically hindered. This hindrance may be due to either N-glycosylation at asparagine 27 or O-glycosylation at serine 26 of the abetas. The failure of Mab17-24 to detect the abetas in the complex is consistent with the observations first reported from High's laboratory indicating that ERp57 only binds to N-glycosylated proteins [63]. Unfortunately, for reasons which are unclear at the present time, N- and O-glycosidases (H- and G-) did not hydrolyze the complex suggesting that these enzymes do not have access to the asparagine27. On the other hand, as noted below, the complex could be hydrolyzed in base.


In light of the failure of others to report that the bulk of the abetas appears to be bound to a carrier protein, it has been repeatedly suggested that our observations are an artifact of the antibodies we have used. Although this is a legitimate concern, we feel that it is very unlikely that our results are due to cross-reactivity with other proteins. In response to this criticism we have employed a number of antibodies, many of which have been widely used by other workers in published studies. In particular, it should be noted that it is extremely unlikely that the reaction of the monoclonal antibody, Mab1-17, is due to an artifact of the assay procedure. This antibody is reactive to an epitope in the first 17 amino acids of the abetas. In a search of the NCBI protein data base, we have found that the sequence of this segment of the peptide is unique and highly conserved in vertebrates. Hence, there is very little likelihood that this antibody is nonspecifically cross-reacting with other proteins. Furthermore, as a result of its generally recognized specificity, this antibody has been widely used by many workers to identify the abetas in samples containing a heterogeneous collection of proteins, such as those found in plaque. If there were a problem with cross-reactivity with this antibody, it would raise the question of the validity of these other studies. Similarly, none of the antibodies we have employed for the detection of either the abetas or ERp57 reacted with either keratin, a major contaminant of tissue preparations, or clusterin, a previously identified carrier protein [12,13].


Figure 4
Figure 4: Immunoblots with A) polyclonal antibodies to ERp57 and B) to Aβ. Channel 1 - prestained standards; channels 2 - CSF; channel 3 - CSF Separated by affinity chromatography with anti-Aβ antibodies; channel 4 - CSF separated by affinity chromatography with anti-ERp57 antibodies. The bands were detected with the alkaline phosphotase reaction.

Work from High’s laboratory [63] has suggested that ERp57 only binds to proteins at N-glycosylation sites along with another ER chaperone, calrectulin. And indeed, when we examined the complex by a western blot run on a native gel without DTT, the Aβ, ERp57 and calreticulin all had Mrs of 130 daltons (Fig.5). This is the theorectical combined sum of the molecular weights of the three proteins [13].


Figure 5
Figure 5: Immunoblot of CSF run on a native Gel. Channel 1 - anti-calreticulin; channel 2 - anti-Aβ; channel 3 - anti-ERp57. The bands were detected with the alkaline phosphotase reaction.
Figure 6
Figure 6: Immunoblots with rabbit polyclonal antibodies to Aβ (1282) of normal human CSF after treatments to dissociate the Aβ-ERp57 complex. Channel 1 - molecular weight markers; channel 2 - untreated CSF; channel 3 - CSF treated with 70% formic acid; channel 4 - CSF treated with 80% trifluoroacetic acid; channel 5 - CSF treated with 6 M guanidine isothiocynate; channel 6 - CSF treated with 6 M urea; Channel 7 - CSF proteins which did not bind to a boronate column; Channel 8 - CSF proteins which bound to a boronate column eluded with pH 9 glycine buffer; channel 9 - treatment of the complex with glycine buffer, pH 9.0 Channel 9 – complex stored over night in pH 9 glycine buffer at 4o C.

One question which arises is whether the ERp57 is covalently bound to the Aβ? The standard procedure for purifying Aβ from plaque is to solubilize it in either concentrated formic or trifluoroacetic acids [64]. Yet when CSF was treated with either acid employing a standard protocol, the complex remained intact (Figure 6: channels 3,4). This observation would suggest that the ERp57 and Aβ are covalently bound. Similarly, two chaotropic agents, 6 M urea and guanidine isothiocyanate had no effect on the complex (Figure 6; channels 5, 6). On the other hand, the complex partially bound to a polyboronate column (Channel 7) indicating that it is glycosylated. It was released from the column with a glycine buffer (Channel 8). When the complex was exposed to a mild alkaline buffer overnight, it was hydrolyzed and the Aβ was seen at a Mr of around 10 Kda suggesting it was partially hydrolyzed and the resulting naked peptide aggregated to form oligomers (Figure 6. Channel 9).


Our findings would suggest that plaque formation is due to either a decline in the ER chaperone content or in the activity of the N-glycosylation pathway, the only glycosylation pathway in the ER. Both mechanisms seem possible. In a separate study we found that the hepatic content of ERp57 showed ca. 50% decline with age.


Figure 7
Figure 7: The effect of age on the ERp57 content of rat liver. (Erickson et al. 2006).

The strange thing about this study is that the ERp57 and several other ER chaperones showed peaks in January and July. This had nothing to do with the temperature of the animal quarters, since the room was windowless and the temperature was held to within 1o C year-round. Most likely this is an epigenetic phenomenon that was established when the animals were in the wild and had to survive the hash environmental conditions.


Table 1: The effect of age on some ER chaperoenes.


Chaperone Peak mg/g-liver Concentration @874 days mg/g-liver Constitutive Show Cyclic Variation
Bip 80.0 48.5 39 Yes
Calnexin 57.4 40.5 70 No
Calreticulin 76 4.8 8 No
ERp55 34.8 13.1 51 Yes
ERp57 15.4 8.7 39 Yes
ERp72 144 100 30 No
Total 336.2 215.1 37 --

The ERp57-Aβ complex appears to be N-glycosylated at ASN27. In line with this hypothesis all the antibodies to Aβ that we used only Mab17-24 (with an epitope of 17-24) could not detect the Aβ, even with very high concentrations of antibody (≥ 1 mg/mL) suggesting that the binding of Mab17-24 to the Aβ was sterically hindered. This hindrance may be due to either N-glycosylation at ASN27 or O-glycosylation at serine 26 of the Aβ. The failure of Mab17-24 to detect the Aβ in the complex is consistent with the observations first reported from High's laboratory indicating that ERp57 only binds to N-glycosylated proteins [63].


Unfortunately, for reasons which are unclear at the present time, N- and O-glycosidases (H- and G-) did not hydrolyze the complex suggesting that after passage of the complex through the golgi these enzymes do not have access to the ASN27. On the other hand, as noted above, the complex could be hydrolyzed in base. Finally, in support of the existence of this complex, we have in preliminary studies isolated and tentatively identified it from tissue preparations (Erickson et al. unpublished data). One problem with this paradigm is that the conical amino acid sequence for N-glycosylation is ASM-X-SER whereas the Aβ sequence is SER-ASN. This sequence may also guide the glycosylation pathway.


The N-glycosylation Pathway

The first step in the N-glycosylation pathway is the synthesis of a 15-saccharide carbohydrate complex attached to a high molecular weight lipid cofactor, dolichol [65-67] (Fig. 8). The construction of this carbohydrate complex is catalyzed by a series of monsaccharide transferases. The first sugar added to the dolichol phosphate is N-acetyl glucoseamine (Fig. 8). The strange thing about dolichol plus dolichol phophate is that a number of laboratories have found that in the liver it increases 5-10-fold with age [68]. Since there is no increase in dolichol synthesis, it would appear that the accumulation of dolichol is due to a metabolic block at the level of the addition of the first sugar to the dolichol phosphate (Fig. 8). The addition of this first sugar to the dolichol is catalyzed by a specific monosaccharide transferase, MST, ALG7. These data suggest that this MST declines with age and that plaque formation is due to the inability of the ER to synthesize this complex and thereby releasing the naked Aβ. This pathway is critical for the synthesis of most membrane and secretory proteins.


If proteins are not correctly folded they are then degraded by the ER associated degradation (ERAD) pathway [69]. The failure to properly fold the nascent protein also activates the unfolded protein response (UPR) [70]. The UPR is controlled by three receptors in the ER membrane. In order to prevent clogging of the lumen of the ER with aggregates, these receptors signal the nucleus to turn off the transcription of proteins that are not critical for cell survival.


It has been found by a number of investigators that in plaque Aβ is present only as the naked peptide [71]. In view of our findings that Aβ in the CSF is only present as a complex\would suggest that plaque accumulation in the elderly is due to a decreased capacity of the ER to catalyze the posttranslational processing of nascent APP. Since 40% of the proteins in the cell, including many of the synaptic membrane proteins that are required for a functioning memory, are like APP, processed in the ER, these observations would suggest that the cognitive decline seen in the elderly is due in part to a decreased capacity of the ER to catalyze the processing of the key secretory and plasma membrane proteins that are necessary for a functioning memory. Compounding this problem, the decreased activity of the ER also leads to a decline in the golgi apparatus which is necessary for the final modifications of proteins that are processed in the ER [72]. An interesting aside is that since insulin receptors are also plasma membrane proteins, these observations could explain the observed insulin resistance which has been reported in the brains of Alzheimer’s disease patients [73].


This model predicts that those individuals with plaque deposits should have lower levels of Aβ in their CSF because on secretion from the cell the naked peptide precipitates in the brain. A large number of studies have reported that such is the case [60]. Hence, the lower levels of Aβ in the CSF and the deposits in the brain are both due to the failure of the ER to form the soluble complex.


The binding of ERp57 to APP is not unique. Two plasma membrane receptors, the vasopressin receptor [74] (Aiyar et al. 1989) and the 1,25 dihydroxy vitamin D plasma membrane receptor (Nemere 2005), have also been shown to be bound to ERp57. The latter receptor is present in the plasma membrane of enterocytes and modulates the uptake of Ca++ from the intestines. The ERp57 would appear to have a role in signal transduction since knocking it down or inhibiting it with an antibody to the terminal carboxy end of ERp57 inhibits the 1,25 dihydroxy vitamin D stimulated Ca++ uptake (Nemere 2005). It should be noted that, as indicated by the title of their report, Aiyar et al. [74] erroneously identified ERp57 as a phophospholipid phospholipase C. Yet, we found that ERp57 is an ER chaperone that has no phospholipase activity [62].


Similarly, data from other laboratories would suggest that there is also a decline in the N-glycosylation pathway with age. Firstly, there is a reduced content of N-glycosylated proteins in the tissues of the elderly [75]. Yet, this decline could be due to decreases in gene transcription. On the other hand, a number of studies have suggested that the decline may also be due to changes in the activity of the N-glycosylation pathway. This concept is based on the effect of age on the content of dolichol in various tissues, a necessary cofactor for N-glycosylation.


The initial step in the N-glycosylation pathway is the synthesis of an oligosaccharide bound to dolichol phosphate [65-67] (Fig. 8). Dolichol is a high molecular weight terpine which serves as a cofactor for the cytosolic synthesis of the oligosaccharide. Once the complex is fully synthesized, it is transferred to the lumen of the ER and then is attached to an ξ-amino group of asparagine by an oligosaccharide transferase (OST) [66]. The addition of the individual sugars to the carbohydrate complex is catalyzed by a family of specific monosaccharide transferases (MST) [65-67]. The first sugars added to the dolichol phosphate are a pair of N-acetylglucosamines, followed by nine mannoses and three glucoses. A number of studies have shown that as animals age there is a 5 to 10-fold increase in the cellular content of both dolichol and dolichol phosphate [68]. Since there is no increase in the synthesis of dolichol, this finding is consistent with a blockage of the addition of the first N-acetylglucosamine. And indeed, it has been reported that the activity of ALG7, the MST that adds the first N-acetylglucosamine to the dolichol phosphate, does decline with age [76].


Figure 8
Figure 8: Synthesis of the N-glycosylation carbohydrate complex (Taken from Helenius and Aebi 2002)

In summary our data along with data from other laboratories suggest that the aging process is associated with a decline in the capacity of the ER to catalyze the posttranslational processing of nascent proteins, including most membrane and secretory proteins that are needed for a functioning memory. This paradigm would imply that the cognitive decline is not due to the toxicity of Aβ, but rather to a decline in the synthesis of the synaptic membrane proteins that are necessary for a functioning memory [13]


The ER and the Mitochondria: The effect of aging on Mitochondrial Function

It is well known that as animal age their mitochondria acquire structural defects [77]. The damaged mitochondria are cleared from the cell by autophagy. Many of the proteins which make up the autophagic vacuoles are processed in the ER and golgi [78].Hence, with a decline in ER function there is also a decline in autophagy. Such declines have been associated with the onset of some age related neurodegenerative diseases [79].


Second, mitochondria are directly bound to the ER through the mitochondrial associated membrane (MAM) [80-82]. This structure serves three, major functions. The first is to transport critical phospholipids from the ER, where they are synthesized, into the mitochondria where they serve vital roles in the maintenance of the structure and function of the mitochondria. Further, it regulates Ca++ homeostatsis in the mitochondria and thereby oxidative phosphorylation [83-90].


A final role for the ER in maintaining mitochondrial structure and function is the control of mitochondrial fission and fusion. As defective mitochondria are cleared by autophagy, they are replaced through a process involving both mitochondrial fission and fusion [91-94]. Recent studies have indicated that several proteins in mitochondrial associated membrane (MAM) play major roles in both of these processes [80] (Friedman et al. 2011). In particular two critical proteins which tether the ER to the mitochondrial outer membrane are mitofusin 1 and 2 (mfn1 & 2) [80] (Fig. 9).


Figure 9
Figure 9: The tethering of the ER to the mitochondria (Taken from Merkwirth and Langer 2008).

These are transmembrane proteins that are located in both the ER membrane and the outer membrane of the mitochondria. Both mfn1 and 2 undergo posttranslational processing in the ER. Hence, a decline in the capacity of the ER to catalyze their posttranslational processing would be expected to have profound effects on mitochondrial structure and function [95-99]. Such declines in ER function could compromise the ability of the cell to maintain metabolically active mitochondria. Together these findings suggest that one aspect of the decline in neuronal function seen with age is a decline in the capacity of the compromised mitochondria to maintain normal ATP production resulting from declines in the function of the ER.


It should be noted that even those individuals who have only the favorable forms apoE, apoE2 and apoE3, are still at risk for the development of Alzheimer’s disease [40,41]. Other factors which have been proposed to cause the dementia, such as oxidative stress, inflammation, heavy metals, hyperphosphorylation of tau leading to tangles and advanced glycation end products, may enhance the progression of the disease, but I would contend that the decline in the synthesis of synaptic proteins is still the primary underlying defect that leads to the cognitive decline seen in the elderly.


Why is there a loss of Myelin with Age?

Finally, one well known clinical phenomenon is that it has been routinely observed in imaging studies that there is a decline in the volume of myelin with age. Myelin is associated with an onion like structure of the plasma membranes of oligodenrocytes in the brain and Shwann cells in the periphery [100-102].


There is no obvious connection between the deposition of plaque and the decreased volume of myelin seen in the elderly. This loss is seen even in areas which do not show plaque deposits. An alternative model for this decline is that it is due to the reduced capacity of the ER and golgi to process the plasma membrane proteins that are needed to maintain this structure [103]. And indeed, Jung et al. (2011) found that knocking out ERp57 and another ER chaperone, calnexin, led to faulty folding of critical myelin proteins. Since the ER content of both of these chaperones decline with age [13] (Table 1), this loss could account for the decrease in myelin seen in the elderly.


Figure 10
Figure 10: Electron microscopic Image of a Schwann Cell.

Conclusions

In summary, published data suggest that the dementia seen in the elderly is not simply due to the neurotoxicity of the Aβ in the extracellular space of the cerebral cortex. The failure of the phase III clinical trials of the monoclonal antibodies bapineuzumab from Elan/J&J/Pfizer and solanezumab from Lilly, both of which have been shown to clear plaque, suggests that Aβ has does not have a major role in the cognitive decline seen in late onset Alzheimer’s disease [32-34]. A similar failure was seen in a vaccine trial [104]. These failures make it difficult to continue to accept the possible major role of the neurotoxicity of Aβ in the etiology of the dementia associated with Alzheimer’s disease [32-34]. These failures would suggest that we should develop new cellular and animal’s models in the search for effective therapeutic agents.


On the other hand, it is quite possible that the antibodies to Aβ may prove to be effective in individuals with Down’s syndrome or the familial forms of Alzheimer’s disease. In these conditions the dementia may be due to the excess production of sβAPP leading to continued apoptosis of neurons through Death Receptor 6 [105]. Although αsAPP appears to be present at much lower concentrations than βsAPP, it is thought to be active at these concentrations (Herzog et al. 2004). Since the 17 carboxy terminal amino acids of αsAPP are the same as the amino terminus of Aβ any antibody which is directed to an epitope in this region of Aβ will also react with and possibly clear αsAPP. This could therefore prevent the continued apoptosis of neurons and the progression of the dementia.


Proposed Future Directions: New Animals and Cellular Models for Drug Discovery

The total failure of the pharmaceutical industry to develop robust therapies for the treatment of late onset disease raises the question of whether they are using the wrong animal model to screen for potential drugs? Many government agencies and pharmaceutical firms have large libraries of compounds which can be screened for their potential activity in enhancing or blocking the specific pathways which have been identified as being important in a disease process. Once these pathways have been identified, it is necessary to develop a reporting system to facilitate the rapid identification of potentially effective therapeutic agents. In the case of aging in general and Alzheimer's disease in particular, our basic biochemical studies have suggested that the physiological functional declines that are seen are due to a decrease in N-glycosylation and/or a decline in the chaperone content of the ER. As a result of these losses there is a decline in the capacity of the ER pathway to fold proteins into their active configurations. Hence, the logical approach is to identify agents which can enhance the production of the ER chaperones and the enzymes that are involved in the N-glycosylation pathway.


Classically workers have constructed pseudo genes in which the promoter region that regulates the transcription of the gene encoding the target protein is coupled to a reporter containing the genes for proteins such as luciferase, galactocidases or various fluorescent proteins. The usual fluorescent proteins are various derivatives of the green fluorescent protein (GFP) such as red (RFP), cyan (CFP), yellow (YFP) and mCherry fluorescent proteins. Each of these indicator systems can be readily quantitated by either standard chemical and/or physical techniques. When the construct is activated by a test compound, the cell synthesizes the reporter protein. In the case of luciferase, the stimulated synthesis of this enzyme can be detected by bioluminescence after the addition of luciferin and ATP. Similarly, the increased production of galactocidases can be determined by the addition of X-gal to the medium which is then hydrolyzed to give a blue color that can then be quantitated by standard spectrophotometric techniques. Finally, the increase in fluorescent proteins can be can be determined by standard fluorometry. All of these assays can be performed in microtiter plate readers and thereby facilitate high throughput screening for drug discovery.


Unfortunately, as we have become more knowledgeable about the control of gene expression, it has become clear that both transcription and translation are regulated by other portions of the gene and not just by the promoter region alone. As is well known the gene for a specific protein has numerous components. These include the exons which contain the coding for the protein sequence, the introns which are inserted between the exons and various control regions including the promoter region which precedes the first exon and the region which follows the last exon, the 3’untranslated region (3'-UTR). As a result, the activation of a gene and the translation of the mRNA frequently is regulated by the RNA coded in sequences in all of these various regions of the gene.


Furthermore, in all eukaryotic cells the DNA is rapped around a set of nuclear proteins, the histones, to form chromatin. This compacts the DNA and makes it possible for a large amount of bulky DNA to fit within the nucleus. Hence, the first step is for the cell to unwind the portion of the DNA containing the relevant gene. This process is mediated though the activation of various enzymes which modify the histones by the addition or removal of various modifications, such as phosphate, acetyl or methyl groups. These modifications are catalyzed by a set of enzymes in the nucleus which in turn are controlled by a number of factors.


Once the DNA is freed from the histones the specific gene is activated by protein complexes which are termed transcription factors which bind to gene-specific sequences on the DNA. The transcription factors are mobilized by a variety of reactions, most of which occur in the cytosol. The activated transcription factor complex is then translocated into the nucleus and bind to the appropriate sites on the DNA activating or repressing the gene. Furthermore, there are enzymes that bind the DNA, unwind its secondary and tertiary structure and separate the activated gene from its bound complementary strand of DNA. This then initiates the transcription of the gene to form a RNA complimentary copy of the exons and introns. The copy is then spliced to give the message RNA (mRNA) which then translocates into the cytosol where it binds to a set of particles, the ribosomes which catalyze the translation of the mRNA. Other factors bind the mRNA to initiate and inhibit translation of the message to form the gene product protein. After the message is translated a number of times, the mRNA is destroyed. The number of times that the message is translated is determined in part by sequences found in the 3’-UTR of the mRNA.


A further complication is that ironically, when introns were first discovered, they were thought to be "junk DNA." That is they seemed to serve no obvious function, but a vast number of studies have clearly demonstrated that they play vital roles in the control of gene expression and hence protein synthesis. Similarly, the 3’-UTR region contains sequences which seem to be involved in the control of transcription. In view of the complexity of this process and the large number of sites which play a role in the control of transcription and translation, it would seem that the potentially most fruitful approach to identifying effective therapeutic agents would be to develop screening techniques which leave the gene and the associated control systems as much intact as is possible. Based on these considerations, the construction of pseudo genes containing only the promoter region does not permit the widest search for drugs affecting the various control points in the initiation and completion of the synthetic process.


I feel that the body of evidence indicates that a more likely paradigm for the etiology of the dementia associated with Alzheimer’s disease is due to declines in the chaperone content of the ER or the activity of the N-glycosylation pathway. This proposal would suggest new cellular models for high through-put drug screening in which the exons of ERp57, and other chaperones and the various MST’s are labeled with a fluorescent protein, such as GFP. The cells could then be incubated with the test compounds and screened in a microtiter plate reader for increased fluorescence compared to those incubated with the vehicle alone. Studies from my laboratory would suggest that a good initial candidate for such trials would be the insecticide, methoxychlor, since in toxicity studies we incidentally found that feeding low doses of this agent specifically increased the ERp57 content of hepatic microsomes (Morrell et al. 2000).


Similarly, it would also be possible to create animal models based on this paradigm. In these studies, the animals could be transfected with a plasmid containing a conditional promoter and the sequence for small, antisense RNA’s directed to the mRNA’s of the chaperones or the MST’s. Since both the chaperone and MST knockouts are lethal mutants, for cognitive testing the transcript would have to be activated in mature animals by the administration of some agent such as doxycycline. The administration of antisense reagents directed to these various targets would cause problems with the psychomotor testing since the loss of function could be due to a loss of muscle mass.


These animal constructs would also serve as proof of principle model systems. For example, in validating the hypothesis that the loss of cognitive function is due to a decline in the posttranslational processing of proteins in the ER, the animals’ cognitive and motor skills can be tested by standard psychomotor procedures both before and after the antisense is activated. Similarly, it could also be used to determine whether the test proteins have a role in the loss of immunological function, cardiac function, kidney and insulin production and the decrease in other tissue functions seen with aging.


References

  1. Cummings J (2017) Lessons Learned from Alzheimer Disease: Clinical Trials with Negative Outcomes Clin Transl Sci 00: 1–6.
  2. Goedert M (2009) Oskar Fischer and the study of dementia. Brain 132: 1102-1111. [crossref]
  3. Katzman R (1976) Editorial: The prevalence and malignancy of Alzheimer disease. A major killer. Arch Neurol 33: 217-218. [crossref]
  4. Glebber GG, Wong CW (1984) Alzheimer’s disease and Down’s syndrome: sharing a unique cerebrovascular fibril protein. Biochem. Biophys. Res. Commun :1131-1135
  5. Masters CL, Simms G, Weinman NA, Multhaup G, McDonald BL, et al. (1985) Amyloid plaque core protein in Alzheimer disease and Down syndrome. Proc Natl Acad Sci U S A 82: 4245-4249. [crossref]
  6. Tanzi R, Gusella JF, Watkins PC, Bruns JA, St. George-Hyslop P, et al. (1987) Amyloid ß protein gene:cDNA, mRNA distribution and genetic linkage near the Alzheimer locus. Science 235: 880-884.
  7. Goldgaber D, Lerman MI, Saffiotti W, Gaidusek DC (1987) Characterization and chromosomal localization of a cDNA encoding brain amyloid of Alzheimer's disease. Science 235: 877-880.
  8. Kang J, Lemaire H-G, Underbeck A, Saulbaum JM, Masters CL, et al. (1987) The precursor of Alzheimer’s disease A4 protein resembles a cell surface protein. Nature 325: 733-736.
  9. Nicolas M, Hassan BA (2014) Amyloid precursor protein and neuronal development. Development 141: 2543-2548.
  10. Bourdet I, Lampin-Saint-Amaux A, Preat T (2015) Amyloid –ß peptide exacerbates the memory deficit caused by amyloid precursor protein loss-of-function in drosophila. Plos One.
  11. Dawson GR, Seabrook GR, Zheng H, Smith DW, Graham S, et al. (1999) Age related cognitive deficits, impaired longterm potentiation and reduction in synaptic marker density in mice lacking the ß-amyloid precursor protein. Neurosci. 90: 1-13.
  12. von Koch CS, Zheng H, Chen H, Trumbauer M, Thinakaran G, et al. (1997) Generation of APLP2 KO mice and early postnatal lethality in APLP2/ APP double KO mice. Neurobiol Aging 18: 661-669.
  13. Erickson RR, Dunning LM, Holtzman JL (2006) The Effect of Aging on the Chaperone Concentrations in the Hepatic, EndoplasmicReticulum of Male Rats: The Possible Role of Protein Misfolding Due to the Loss of Chaperones in the Decline in Physiological Function Seen with Age. J. Gerontol. Biol. Med 61A: 435-443.
  14. Jonsson T, Atwal JK, Steinberg S, Snaedal J, Palmi V Jonsson, et al. (2012) A mutation in APP protects against Alzheimer's disease and age-related cognitive decline. Nature 488: 96-99.
  15. Bamne MN, Demirci FY, Berman S, Snitz BE, Samantha L, et al. (2014) Investigation of an amyloid precursor protein protective mutation (A673T) in a North American case- control sample of late-onset Alzheimer's disease. Neurobiol Aging 35: 177ge15-177ge16
  16. Vassar R (2014) BACEI inhibitor drugs in clinical trials for Alzheimer's disease. Alzheimer's Res. Therap 6:689.
  17. Laird FM, Cai H, Savonenko AV, Farah MH, He KH, et al. (2005) BACE1, a major determinant of selective vulnerability of the brain to amyloid-B amyloidosis, is essential for cognitive, emotional and synaptic function. J NeuroSci 25: 11693-11706.
  18. De Strooper B, Vassar R, Golde T (2010) The secretase enzymes with therapeutic potential in Alzheimer's disease. Nat. Rev. Neurol. 6:99-107 Development 2014;141: 2543-2554.
  19. Prox J, Rittger A, Saftig P (2012) Physiological functions of the amyloid precursor protein secretases ADAM10, BACE1, and presenilin. Exp Brain Res 217: 331-341. [crossref]
  20. De Strooper B, Iwatsubo T, Wolfe MS (2012) Prsenilins and 'Y-secretase structure, function and role in Altzheimer disease. Cold Spring Hab Perspect Med 2: a006304.
  21. Hu X, Zhou X, He W, Yang J, Xiong W, et al. (2010) BACE1 deficiency causes altered neuronal activity and neurodegeneration. J Neurosci 30: 819-829.
  22. Endres K, Farenholtz F, Lotz J, Heimke C, Teipel S, et al. (2014) Increased CSF AAPs-a levels in patients with Alzheimer disease with acitretin Neurology 83:1930-1935.
  23. Bapista F, Henriques AG, Silva AMS, Wiltfang J, da Cruz e Silva OAB, et al. (2014) Favonoids as therapeutic compounds targeting key involved in Alzheimer's disease. Chern. Neurosci 5: 83-92.
  24. Fahrenholz F, Tippmann F, Endres K (2010) Retinoids as a perspective in treatment of Alzheimer's disease. Neurodegenerative Diseases 7: 190-192.
  25. Tanzi RE (2012) The genetics of Alzheimer disease. Cold Spring Harb Perspect Med 2. [crossref]
  26. Campos LS, L Decker, V Taylor, Skarmes W Notch (2006) Epidermal Growth Factor Receptor, and _1-Integrin Pathways Are Coordinated in Neural Stem Cells. J. Biol. Chem 281: 53000-53009.
  27. Bertram L, Tanzi RE (2009) Genome-wide association studies in Alzheimer's disease. Human Molecular Genetics 18: R137-145
  28. Schor NF (2011) What the halted phase III ?-secretase inhibitor trial may (or may not) be telling us. Ann. Neurol 69: 237-239.
  29. Selkoe DJ (1997) Alzheimer’s disease – genotypes, phenotypes and treatments Science 275:621-637.
  30. Donohue MC, Sperling RA, Peterson R, Chung-Kai S, Weiner MW et al. (2017) Association between brain amyloid and subsequent decline among cognitively normal persons. JAMA Neurology 317: 2305-2316.
  31. Maarouf CL, Daugs ID, Kokjohn TA, Walker DG, Hunter JM, et al. (2011) Alzheimer’s disease and non-demented high pathology control nonagenarians: Comparing and contrasting the biochemistry of successful aging aging. PLos One 6: e27291.
  32. Holtzman JL (2009) The Failure of the APP Transgenic Mouse Models to Predict Efficacy of Agents in Clinical Trials of Alzheimer’s Disease: These Findings Suggest That There is a Need for a New Paradigm for Drug Development. Research Progress in Alzheimer’s Disease and Dementia. Ed. Miao-Kun Sun, Nova Science Inc., New York. Chapter 5, Vol. 4 pp. 112-144.
  33. Holtzman JL (2010) Are we prepared to deal with the Alzheimer's disease pandemic? Clin Pharmacol Ther 88: 563-565. [crossref]
  34. Gandy S (2005) The role of cerebral amyloid beta accumulation in common forms of Alzheimer disease. J Clin Invest 115: 1121-1129. [crossref]
  35. Seabrook GR, Ray WJ, Shearman M, Hutton M (2005) Beyond Amyloid: The next generation of Alzhiemer’s Disease Therapeitucs Molec. Interventions 7: 261-270.
  36. Robakis NK (2010) Are Aß and its derivatives causative agents or innocent bystanders in AD? Neurodegenerative Dis 7: 32-37.
  37. Roses AD (1998) Alzhiemer Diseases A model of gene mutations and susceptibility polymorphisms for complex psychiatric diseases. Amer J. Med Genetics 81: 49-57.
  38. Schmechel DE, Saunders AM, Strittmatter WJ, Crain BJ, Hulette CM, et al. (1993) Increased amyloid beta-peptide deposition in cerebral cortex as a consequence of apolipoprotein E genotype in late-onset Alzheimer disease. Proc. Nat. Acad. Science USA 90: 9649-9653.
  39. Farrer LA, Cupples LA, Haines JL, Hyman B, Kukull WA, et al. (1997) for the APOE and Alzheimer Disease Meta Analysis Consortium. Effects of age, sex, and ethnicity on the association between apolipoprotein E genotype and Alzheimer disease. A meta-analysis. JAMAM 278: 1349-1356.
  40. Ashford JW (2004) APOE genotype effects on Alzheimer's disease onset and epidemiology. J Mol Neurosci 23: 157-165. [crossref]
  41. Kamboh MI (2004) Molecular genetics of late-onset Alzheimer's disease. Ann Hum Genet 68: 381-404. [crossref]
  42. Crean S, Ward A, Mercaldi CJ, Collins JM, Cook MN, et al. (2011) Apolipoprotein E4 prevalence in Alzheimer's disease patients varies across global populations: a systematic literature review and meta-analysis. Demen. Geriatr. Cogn. Disord 31: 20-30.
  43. Altman R, Rutledge JC (2010) The vascular contribution to Alzheimer's disease. Clin Sci (Lond) 119: 407-421. [crossref]
  44. Castellano JM, Kim J, Stewart FR, Jiang H, DeMattos RB, et al. (2011) Human apoE isoforms differentially regulate brain amyloid-beta peptide clearance. Sci. Transl. Med 3:89ra57
  45. Strittmatter WJ, Saunders AM, Schmechel D, Pericak-Vance M, Enghild J, et al. (1993) Apolipoprotein E: high-avidity binding to beta-amyloid and increased frequency of type 4 allele in late-onset familial Alzheimer disease. Proc. Nat. Acad. Sci. USA 90: 1977-1981.
  46. Ghiso J, Matsubara E, Koudinov A, Choi-Miura NH, Tomita M, et al. (1993) The cerebrospinal-fluid soluble form of Alzheimer's amyloid beta is complexed to SP-40,40 (apolipoprotein J), an inhibitor of the complement membrane-attack complex. Biochem. J 293: 27.
  47. Schrijvers EMC, Koudstaal PJ, Hofman A, Breteler MMB (2011) Plasma clusterin and the risk of Alzheimer Disease JAMA 305: 1922-1926.
  48. Roses AD, Lutz MW, Amrine-Madsen H, Saunders AM, Crenshaw DG, et al. (2010) A TOMM40 variable-length polymorphism predicts the age of late-onset Alzheimer's disease. Pharmacogenomics J. 10: 375-384.
  49. Valla J, Yaari R, Wolf AB, Kusne Y, Beach TG, et al. (2010) Reduced posterior cingulate mitochondrial activity in expired young adult carriers of the APOE4 allele, the major late-onset Alzheimer's susceptibility gene. J. Alzheimer's Dis 22: 307-213
  50. Bettens K, Vermeuelen S, Van Cauwberghe C, Heeman B, Asselberg Blankensteen MA, et al. (2015) Cluterin levels in plasma predict cognitive decline and progression to Alzheimer's disease. J of Alzhiemer's Disease 46: 1103-1110.
  51. Jongbloed W, van Dijk KD, Mulder SD, van de Berg WDJ, Jonsson T, et al. (2011) Specialization of endoplasmic reticulum chaperones for the folding and function of myelin glycoproteins P0 and PMP22. FASEB J 25: 3929-3937.
  52. Lancaster TM, Brindley LM, Tansey KE (1989) PreAA495Precusor protein of Alzheimer’s disease Ad amyloid is encoded by 16 exons. Nucleic Acid Res 17: 517-522.
  53. Huang Z, Cheng c, Jiang L, Yu Z, Cao F, et al. (2005) Dissection of the mitochondrial import and assembly pathway for human Tom40. J. Biol. Chem 280: 11535-11543.
  54. Müller U, Winter P, Graeber MB (2013) A presenilin 1 mutation in the first case of Alzheimer's disease. Lancet Neurol 12: 129-130. [crossref]
  55. Rupp C, K Beyreuther, K Maurer, S Kins (2012) A presenilin 1 mutation in the first case of Alzheimr's Disease: revisited Alzheimers & Dementia 2014;10:869- 872 disease and age-related cognitive decline. Nature 488: 96-99.
  56. Balaraman Y, Limaye AR, Levey AL, Srinivasan S (2006) Glycogen synthase kinase.3ß and Alzheimer's disease: pathophysiology and therapeutic significance. Cell Mol. Life Sci 63: 1226-1235.
  57. Hurtado DE, Molina-Porcel L, Carroll JC, MacDonald C, Aboagye AK, et al. (2012) Selectively silencing GSK3 isoforms reduces plaques and tangles in mouse models Alzheimer's Disease. J Neurosci 32: 7392-7402.
  58. Cai Z, Zhao Y, Zhao B (2012) Roles of glycogen synthase 3 in Alzheimer's disease. Cur. Alzheimer Research 9: 894-879
  59. Gauthier S, Feldman HH, Schneider LS, Wilcock GK, Frisoni GB, et al. (2016) Efficacy and safety of tau-aggregation inhibitor therapy in patients with mild or moderate Alzheimer's disease: a randomised, controlled, double-blind, parallel-arm, phase 3 trial Lancet Neuology 15:
  60. Fagan AM, Holtzman DM (2010) Cerebral fluid biomarkers of Alzhiemer’s disease. Biomarkers in Medicine 4.1:51
  61. Lyall DM, Harris SM, Bastin ME, Munoz-Maniega S, Murray C, et al. (2014) Are APOE e and TOMM40 pol-T repeat length associations with cognitive ageing mediated by brain white matter integrity? Transl. Pyshciatry e445
  62. Srivastava SP, Chen N-Q, Liu Y-X, Holtzman JL (1999) Purification and characterization of a new isozyme of thiol:protein disulfide oxidoreductase from rat hepatic microsomes: Relationship of this isozyme to cytosolic, phosphatidylinositol specific phospholipase C form 1A. J. Biol. Chem 266: 20337-20344.
  63. Elliott JG, Oliver JD, High S (1997) The thiol-dependent reductase ERp57 interacts specifically with N-glycosylated integral membrane proteins. J. Biol. Chem 272: 13849-13855.
  64. Orlando R, Kenny PT, Zagorski MG (1992) Covalent modification of Alzheimer's amyloid beta-peptide in formic acid solutions. Biochem. Biophys. Res. Commun 184: 686-691.
  65. Yan Q, Lennarz WJ (1999) Oligosaccharyltransferase: a complex multisubunit enzyme of the endoplasmic reticulum. Biochem. Biophys. Res. Commun 266: 684-689.
  66. Helenius J, Aebi M (2002) Trans- membrane movement of dolichol linked carbohydrates during N- glycoprotein biosynthesis in the endoplasmic reticulum. Semin. Cell Dev. Biol 13: 171–178.
  67. Kelleher DJ, Gilmore R (2006) An evolving view of the eukaryotic oligosaccharyltransferase. Glycobiology 16: 47R–62R.
  68. Marino M, Dolfi C, Paradiso C, Cavallini G, Masini M, et al. (1998) Age-dependent accumulation of dolichol in rat liver: is tissue dolichol a biomarker of aging?. J. Gerontol. A Biol. Sci. Med. Sci 53: B87-B93
  69. Vembar SS, Brodsky JL (2008) One step at a time: endoplasmic reticulum-associated degradation. Nat. Rev. Molec. Biol 9: 944-957.
  70. Ron D, Harding HP (2012) Protein-folding homeostasis and nutritional regulation, Cold Springs Harb. Perspect. in Biol 4: a13177
  71. Miller DL, Papayannopoulos IA, Styles J, Bobin SA, Lin YY, et al. (1993) Peptide compositions of the cerebrovascular and senile plaque core amyloid deposits of Alzheimer's disease. Arch. Biochem. Biophys 301: 41-52.
  72. Hu Z, Zeng L, Huang Z, Zhang J, Li T, et al. (2007) The study of golgi apparatus in Alzheimer’s disease. Neurochem. Res 32: 1265-1277.
  73. Correia SC, Santos RX, Perry G, Zhu X, Moretta PI, et al. (2011) Insulin-resistant brain state: The culprit in sporadic Alzheimer’s disease? Ageing Research Reviews 10: 264-273.
  74. Aiyar N, Bennett F, Nambi P, Valinski W, Angioli M, et al. (1989) Solubilization of rat liver vasopressin receptors as a complex with a guanine-nucleotide-binding protein and phosphoinositide-specific phospholipase C. Biochem. J 261: 63-70.
  75. Kousvelari EE, Banerjee DK, Murty L, Baum BJ (1988) N-linked protein glycosylation in the rat parotid gland during aging. Mechanisms of Ageing & Development 42:173-181.
  76. Mota OM, Huang GT, Kukuruzinska MA (1994) Developmental regulation and tissue-specific expression of hamster dolichol-P-dependent N-acetylglucosamine-1-P transferase (GPT). Biochem. Biophys. Res. Commun 204: 284-291.
  77. Batlevi Y, La Spada AR (2011) Mitochondrial autophagy in neural function, neurodegenerative disease, neuron cell death, and aging. Neurobiology of Disease 43: 46-51.
  78. Militello D, Colombo MI (2011) A membrane is born: Origin of the autophagosomal compartment. Current Molec. Med. 11: 197-203.
  79. Pan T, Kondo S, Weidong L, Jankovic J (2008) The role of autophagy-lysomome pathway in neurodegeneration associated with Parkinon’s disease. Brain 131: 199-1978.
  80. Merkwirth C, Langer T (2008) Mitofusin 2 builds a bridge between ER and mitochondria. Cell 135: 1165-1167. [crossref]
  81. Goldgaber D, Lerman ML, McBride OW, Saffiotti U, Gajdusek DC, et al. (2009) Hayashi, T., Rizzuto, R., Hajnoczky, G. and Su, T.-P. MAM: more than just a housekeeper. Trends in Cell Biol 18: 81-88.
  82. Schon EA, Przedborski S (2011) Mitochondria: the next (neurode)generation. Neuron 70: 1033-1053. [crossref]
  83. Unterbeck A, Beyreuther K, Muller-Hill B (1989) The PreA4695 precursor protein of Alzheimer's disease A4 amyloid is encoded by 16 exons. Nucleic Acid Res 17: 517-522.
  84. Bateman RJ, Holtzman DM (2011) Human apoE isoforms differentially regulate brain amyloid-beta peptide clearance. Sci. Transl. Med 3: 89ra57.
  85. Carnero E, Sutherland JD, Fortes P (2011) Adenovirus and miRNAs. Biochim Biophys Acta 1809: 660-667. [crossref]
  86. Cummings JL, Morstorf T, Zhong K (2014) Alzheimer?s disease drug- development pipeline: few candidates, frequent failures. Alzheimer's Research & Therapy 6: 37.
  87. Dominguez D, Tournoy J, Hartmann D, Huth T, Cryns K, et al. (2005) Phenotypic and biochemical analyses of BACE1- and BACE2-deficient mice. J Biol Chem 280: 30797-30806. [crossref]
  88. Djupedal I, Ekwall K (2009) Epigenetics: heterochromatin meets RNAi. Cell Res 19: 282-295. [crossref]
  89. Holtzman JL (2013) Cellular and animal models for high-throughput screening of therapeutic agents for the treatment of the diseases of the elderly in general and Alzheimer's disease in particular. Frontiers in Pharmacology 4: 59.
  90. Hyslop P, Van Keuren ML, Patterson D, Pagan S, Kurnit DM, et al. (1987) Amyloid ~ protein gene: cDNA, mRNA distribution, and genetic linkage near the Alzheimer locus. Science 235: 880-884.
  91. Johnson SC, La Rue A, Hermann BP, Xu G, Koscik RL, et al. The effect of TOMM40 poly-T length on gray matter volume and cognition in middle-aged persons with APOE e3/e3 genotype.
  92. Johnson SC, La Rue A, Hermann BP, Xu G, Koscik RL, et al. (2011) The effect of TOMM40 poly-T length on gray matter volume and cognition in middle-aged persons with APOE ε3/ε3 genotype. Alzheimers Dement 7: 456-465. [crossref]
  93. Lemaire HG, Salbauml JM, Multhaup G, Kang J, Bayney RM, et al. (2010) Genetic variation in a single locus and age of onset for Alzheimer’s disease. Alzheimer’s & Dementia 6: 125-131.
  94. Mantripargada K, Owen MJ, Williams J, Linden DEJ (2015) Alzheimer's disease risk variant in CLU is associated with neuronal inefficiency in healthy individuals. Alzheimer's and Dementia 11: 1144-1152.
  95. Mokranjac D, Neupert W (2008) Thirty years of protein translocation into mitochondria: Unexpectedly complex and still puzzling. Biochim. Biophys. Acta 1793: 33-41.
  96. Mori H, Takio K, Ogawara M, Selkoe DJ (1992) Mass spectrometry of purified amyloid beta protein in Alzheimer's disease. J Biol Chem 267: 17082-17086. [crossref]
  97. Morrell SL, Fuchs JA, Holtzman JL (2000) The Effect of Methoxychlor Administration to Male Rats on the Hepatic, Microsomal Iodothyronine 5'-Deiodinase, Form I. J. Pharmacol. Expertl. Therap 294: 308-312.
  98. Oliver JD, van der Wal FJ, Bulleid NJ, High S (1997) Interaction of the thiol dependent reductase ERp57 with nascent glycoproteins. Science 275: 86-88.
  99. Reiman EM, Chen K, Alexander GE, Caselli RJ, Bandy D, et al. (2004) Functional brain abnormalities in young adults at genetic risk for late-onset Alzheimer's dementia. Proc. Nat. Acad. Sci. USA 101:284-289.
  100. Robberecht BC, Engleborghs S, Vandenberghe M, De Deyn PP, Cruts M, et al. (2015) Reduced secreted clusterin as a mechanism for Alzheimer-associated CLU mutations Molec Neurodegeneration 10: 30.
  101. Sinclair KD, Corr SA, Gutierrez CG, Fisher PA, Lee J-H, et al. (2016) Healthy aging of cloned sheep. Nature Comn 17: 2-9.
  102. Sun X (2016) Intra ventricular Apolipoprotein J infusion acts protectively in traumatic brain injury. J Neurochemistry 136: 1017-1025.
  103. Tanzi R, Gusella JF, Watkins PC, Bruns JA, St. George-Hyslop P, et al. (2015) Amyloid –ß peptide exacerbates the memory deficit caused by amyloid precursor protein loss-of-function in drosophila. Plos One.
  104. Holmes C, Bache D, Wilkinson D, Yadegarfar G, Hopkins V, et al. (2008) Long-term effects of Aß42 immunisation in Alzheimer’s disease: follow-up of a randomized, placebo-controlled phase 1 trial. Lancet 372: 216-223
  105. Nikolaev A, McLauglin T, O’Oleary DDM, Tessier-Lavigne M (2009) APP binds DR6 to trigger axon pruning and neuron death via distinct caspases. Nature 457: 981-989.