Definition

A neurodegenerative disease of the brain characterized clinically by insidious, chronic, and progressive cognitive decline and histologically by cerebral accumulations of the proteins beta amyloid (plaques) and tau (tangles).

Historical Background

In 1902, a woman called Auguste D. came under the care of Dr. Alois Alzheimer and then at the University of Frankfurt. The patient manifested changes in behavior and cognition. Her clinical course was characterized by progressive paranoia, delusional thinking, disorientation, and poor memory. She was institutionalized for the last 3 years of her life. Upon her death, Alzheimer analyzed her brain using a silver stain and described both extracellular and intracellular protein accumulations. The extracellular protein accumulations were termed plaques, and the intraneuronal protein accumulations were called tangles. Alzheimer presented the results of this autopsy in 1906. Several other similar cases of relatively “presenile” (i.e., arbitrarily defined as an onset prior to age 55–65) clinical dementia associated with plaques and tangles were noted by Alzheimer and others over the next 4 years. In 1910, Alzheimer’s departmental chair, Emil Kraepelin, published a textbook covering the fields of neurology and psychiatry and referred to patients with presenile dementia, plaques, and tangles as having “Alzheimer’s disease.”

Concurrently, other investigators, such as Oscar Fischer, also reported plaque presence in elderly demented individuals. These individuals were older than those with “presenile” dementia (i.e., generally older than age 55–65). As the commonality of progressive dementia in the elderly was well recognized, the presence of plaques in elderly demented individuals was felt to represent a normal phenomenon. Such individuals were not diagnosed with Alzheimer’s disease. Instead, cognitive decline in elderly adults was attributed to normal aging or other poorly described conditions, such as “hardening of the arteries.” As a result, Alzheimer’s disease remained relatively uncommon for a number of subsequent decades.

In the 1960s, investigators began comparing elderly demented subjects to those diagnosed with “presenile” Alzheimer’s disease. Notable similarities were observed regarding the clinical course (chronic and progressive), the clinical features (cognitive decline that featured evolution of an amnestic state, followed by behavioral changes), and histopathology (plaques and tangles). By the 1970s, the number of demented elderly was growing fast as demographic shifts in the aging population combined with increased recognition of the syndrome. At this point, the original definition of Alzheimer’s disease (as described by Alzheimer and named by Kraepelin) was expanded to account for all dementing individuals with plaques and tangles, although some separation of these groups was envisioned. Those meeting the original criteria of plaque and tangle dementia in presenile adults were designated as having dementia of the Alzheimer type (DAT), while the previously unconsidered elderly cases were designated as having senile dementia of the Alzheimer type (SDAT). With increasing recognition of the problem, Alzheimer’s disease very quickly became incredibly common, as well as a Western civilization health priority.

In the USA, the 1980s saw the establishment of federally funded Alzheimer’s disease research centers, which began to systematically study the clinical course of this progressive dementia, mostly in the common SDAT form. Academic research began to unravel the chemical makeup of plaques and tangles. Investigations into patterns and causes of neurodegeneration were performed. This advancing knowledge enhanced the ability of clinicians to diagnose Alzheimer’s disease at increasingly subtle stages, as well as the ability to pharmacologically intervene to achieve partial, temporary symptomatic benefits in at least some individuals.

Current Knowledge

Scientific Perspective

The plaques seen in persons with Alzheimer’s disease contain several aggregated proteins. The major constituent is a protein called amyloid beta (Aβ). “Beta” is a chemical term that specifies a certain pattern of protein folding. “Amyloid” is a general term that refers to proteins that give a particular appearance when exposed to a particular type of stain, Congo red. The beta amyloid, or Aβ, found in the brains of Alzheimer’s disease patients derives from a particular protein called the amyloid precursor protein (APP).

In the human brain, the APP is 695 amino acids long. It is a transmembrane protein. One end (the carboxyl end) is found inside neurons, in the cytoplasm. The other end (the amino end) extends outside the cell. In between the cytoplasmic and extracellular portions is a stretch that runs through the membrane. The normal function of APP is not well known. APP is digested by different enzymes, which cut the protein at different points. An enzyme complex called the beta secretase (BACE) cuts APP in its extracellular portion. An enzyme or group of enzymes referred to as the alpha secretase cuts APP in its intramembrane segment. The gamma secretase cuts APP twice, both times in its intramembrane segment. Both of the gamma secretase cuts occur closer to the carboxyl end of the APP than the alpha secretase cut.

Different cutting combinations generate various APP by-products. Cutting of an APP by beta and gamma secretases generates a 38–43 amino acid stretch, and this stretch tends to assume a beta folding conformation and has the features of an amyloid protein (i.e., birefringence under the microscope when stained with Congo red). The 40 and 42 amino acid-long variants of Aβ predominate in plaques and are often designated Aβ40 and Aβ42. Aβ42 seems to be particularly important to the formation of the amyloid plaques of Alzheimer’s disease, probably because this version of the protein is quite insoluble. When Aβ accumulations begin to form in brain, they are not associated with disrupted cell elements and are called “diffuse plaques.” Another type of more evolved plaque can also be found in Alzheimer’s disease patients, in which Aβ becomes condensed at the center of the plaque, and the vicinity of the plaque is associated with disrupted cell elements such as degenerating axons and dendrites. As axons and dendrites are collectively called “neurites,” this type of plaque is called a “neuritic plaque.”

The tangles of Alzheimer’s disease are found primarily in neurons. Under the microscope tangles have a fibrous quality to them, and hence tangles in Alzheimer’s disease are referred to as “neurofibrillary tangles.” Neurofibrillary tangles consist of a protein called tau. Normally, tau is found in association with microtubules, which act as a skeleton, or “cytoskeleton” supporting the cellular structure. The function of tau appears to be the stabilization of these microtubules. Like many proteins, after its production tau is modified by the addition and subtraction of phosphate groups on certain amino acids, especially serine and threonine. During embryonic development, tau is heavily phosphorylated, but during youth and early adulthood, this heavily phosphorylated pattern is rare if at all seen. In Alzheimer’s disease, though, tau again takes on a heavily phosphorylated pattern, which is felt to reflect an abnormal physiologic event and is referred to as tau “hyperphosphorylation.” Hyperphosphorylated tau molecules begin to pair off, a process called “dimerization.” Hyperphosphorylated tau dimers, also called “paired helical filaments,” are quite insoluble and begin to aggregate with each other. This aggregation, typically visible extending from cell bodies into axons, comprises the neurofibrillary tangle.

As impressive as this advancing understanding of plaque and tangle composition is, recognizing what constitutes these aggregations does not address why they form. In this regard, genetic studies of DAT subjects who inherit the disorder in an autosomal-dominant fashion have had a large impact. Several hundred such families have been documented. In these families the disease affects about 50% of each generation, with typical onset occurring in the third, fourth, fifth, or sixth decades. A small number of these families have demonstrable mutations in the gene that encodes the APP. This gene is located on chromosome 21, the same chromosome that is present in excess in Down’s syndrome. Down’s syndrome patients invariably accumulate Aβ plaques in their fifth decade. A somewhat larger number of these families have mutations in the gene that encodes a protein called presenilin 1. This gene is found in chromosome 14. Presenilin 1 protein constitutes part of the gamma secretase complex. A smaller number of families have mutation of a related gene on chromosome 1, which encodes a related protein, presenilin 2. Presenilin 2 can also participate in formation of the gamma secretase. Mutations in the genes that encode APP, presenilin 1, and presenilin 2 all enhance the production of Aβ42. This has lent support to the “amyloid cascade hypothesis,” which posits as Aβ42 is generated it begins to interfere with neuronal function, kill neurons, and generate the other histologic features seen in Alzheimer’s disease. While the logic underlying this hypothesis is obvious, it is important to keep in mind it assumes the very small subset of early-onset, autosomal-dominant Alzheimer’s disease (which accounts for far less than 1% of those affected) have a similar if not identical etiology to the common sporadic, late-onset cases that constitute the vast majority. In those subjects, what initiates Aβ42 production remains an open area of debate. Conceivably, population diversity in genes that contribute to APP production or processing could cause Aβ42 to appear. Environmental factors could lead to Aβ42 formation. Also, a variety of age-related factors promote Aβ42 formation.

Other factors are recognized to play a role in Alzheimer’s disease, and where these factors fit into or what they tell us about the etiologic hierarchy of the disease is unclear. One factor relates to the APOE gene on chromosome 19. The APOE gene shows population variability due to the presence of two polymorphic positions. The common APOE variants are the ε2, ε3, and ε4 forms. The APOE ε4 form is over represented in those with Alzheimer’s disease, where it seems to move up the age of presentation in those destined to develop the disorder. Mitochondrial function is also altered in Alzheimer’s disease, and these alterations are not limited to the brain.

Diagnostic Perspective

Dementia is defined as a cognitive decline that has advanced to that point it interferes with activities of daily living. While dementia has many different etiologies, Alzheimer’s disease is the most common cause of dementia, accounting for 50–60% of dementia verified by neuropathological examination of the brain at autopsy. The clinical diagnosis (i.e., diagnosis in life) of Alzheimer’s disease is made in patients who have progressive dementia with no other systemic or brain diseases that could account for the progressive cognitive decline. A diagnosis of “definite Alzheimer’s disease” is traditionally diagnosed at autopsy by the presence of plaques and tangles (although in some older schemas tangles are not requisite) in an individual with a clinical history suggestive of dementia. The presence of plaques and tangles in typical brain regions (mesial temporal, parietal, and inferior frontal structures) is quite common in elderly persons with the clinical syndrome of Alzheimer’s disease. As a result of the high prevalence of Alzheimer’s disease with advancing age (at least one commonly quoted study estimates approximately half of those over the age of 85 have it), the specificity of the clinical diagnosis is high. Recognition of how common Alzheimer’s disease is in later life has also served to enhance clinician awareness, thus improving sensitivity of the diagnosis. In the hands of an experienced physician, clinical diagnostic accuracy is excellent.

Criteria originally designed to facilitate identification of subjects for clinical trials have helped to standardize clinical diagnostic approaches. These criteria, such as those proposed by the National Institute of Neurologic, Communicative Disorders, and Stroke (NINCDS) and the Alzheimer’s Disease and Related Disorders Association (ADRDA) in the 1980s emphasize the importance of establishing that a progressive dementia exists in a patient. Two basic approaches are commonly used toward this end. One is to demonstrate a pattern of cognitive domain strengths and weaknesses that reliably suggest decline from a previous level of cognitive function has emerged. For example, defective memory retention in the presence of another defective cognitive domain (language, executive function, visuospatial function, and praxis) in an elderly patient with cognitive complaints and an otherwise unremarkable physical exam is strongly suggestive of Alzheimer’s disease. The other approach focuses more on defining the degree and nature of emerging declines in daily living activities. This latter technique focuses extensively on collateral history obtained from family members or friends of the patient.

The diagnosis is made primarily through clinical impression, although that impression is influenced by a small set of recommended laboratory and imaging tests. These tests are serologic (vitamin B12 level, thyroid function tests, electrolytes with renal and hepatic indices, and a blood cell count) and structural (brain imaging by either computed tomography or magnetic resonance imaging) in nature. As currently used, they mostly serve to rule out the presence of concomitant pathologies that can interfere with cognition. Although this has contributed to the view that the Alzheimer’s disease diagnosis is one of exclusion, it should be noted that certain patterns of cognitive decline elicited by clinical history or demonstrable by neuropsychological testing are so typical of Alzheimer’s disease they can be used to support a diagnosis of inclusion. It is important to note, though, that while fluorodeoxyglucose PET and APOE genotyping are occasionally used to address specific questions, such as whether the presence of Alzheimer’s disease versus a frontotemporal dementia is more likely in a patient, these tests are not routinely used in the diagnosis of Alzheimer’s disease and cannot by themselves establish a diagnosis of Alzheimer’s disease.

Another PET application is also now available, which reveals the presence of brain fibrillary Aβ deposition (in essence, the presence of amyloid plaques). At the time of this writing amyloid plaque scanning, which is expensive and not covered by third-party payers, is not routinely used to diagnose Alzheimer’s disease although the utility of this test is under study and this may change in the future. Spinal fluid measurements of Aβ and tau protein levels yield similar information to that which is obtained through an amyloid scan, but this test is currently used more in research settings than it is in the clinic setting.

An increasing emphasis on Alzheimer’s disease biomarker measurements is apparent in new Alzheimer’s disease diagnostic research criteria that were formulated in 2011. In addition to essentially continuing recognition of existing Alzheimer’s disease criteria, two new largely biomarker-driven Alzheimer’s disease categories were proposed. One new category, “prodromal Alzheimer’s disease,” captures individuals with a mild cognitive impairment (MCI) syndrome and evidence of fibrillary brain amyloid plaque deposition. “Preclinical Alzheimer’s disease” is defined as intact cognitive function in an individual with evidence of fibrillary brain amyloid plaque deposition.

Treatment Perspective

Although Alzheimer’s disease is currently neither reversible nor curable, it is possible to treat its symptoms. The first approved treatment for Alzheimer’s disease was tacrine, a cholinesterase inhibitor. This drug increased levels of brain acetylcholine by antagonizing its synaptic degradation. Increasing brain cholinergic tone was identified as a pharmacologic target because Alzheimer’s disease patients show a profound loss of acetylcholine due to degeneration of cholinergic neurons in the basal forebrain. Safer cholinesterase inhibitors (donepezil, rivastigmine, and galantamine) have since superseded tacrine. In addition to inhibiting acetylcholinesterase, rivastigmine also inhibits butyrylcholinesterases that also hydrolyze acetylcholine, and galantamine is an allosteric modulator of acetylcholine nicotinic receptors. Each agent shows a similar overall degree of efficacy, although the individual with Alzheimer’s disease may respond to or tolerate one drug better than the other. Treatment cohorts followed for 12 weeks to 3 years indicate that as a group, those started on cholinesterase inhibitors tend to perform and appear slightly improved compared to their immediate pretreatment baseline. This improvement appears detectable for 6–12 months. By 12 months, though, treatment groups return to their pretreatment performance as ascertained by cognitive testing, clinical impression, and caregiver impression. Beyond 12 months, patients continuously decline below their pretreatment baseline, although for at least the next several years, patients appear to perform better on cognitive testing than would otherwise be expected. The clinical meaningfulness of this sustained benefit has fueled considerable debate. Benefits have been observed on measures of cognitive ability, functional ability, behavior, and caregiver stress.

At the time of this writing, memantine is the only non-cholinesterase inhibitor specifically approved for the treatment of Alzheimer’s disease. Under in vitro conditions, memantine blocks a cation channel associated with the NMDA type of glutamate-activated ionotropic receptors. Whether or not this is its primary mechanism of action in Alzheimer’s disease has been questioned. In any case, cohorts of patients with moderate or severe Alzheimer’s disease, when randomized to memantine, perform better on measures of cognitive and functional performance than do concurrent placebo treatment groups. In severe Alzheimer’s disease, the magnitude of observed benefit is similar to that obtained with donepezil. Memantine and donepezil have been studied in combination with each other. Subjects with mini-mental state exam scores of 5–14, who were already on donepezil, did better as a group when memantine was added to their treatment regimen than when placebo was added. Demonstrable benefits in mild Alzheimer’s disease are lacking, and thus the role of memantine in the mild stages of Alzheimer’s disease is not clear.

Two studies concluded high-dose vitamin E (2000 IU each day) might slightly slow decline in Alzheimer’s disease patients. Other studies, though, suggest taking more than 400 IU of vitamin E on a daily basis increases overall mortality. The marginality of any vitamin E benefit, in conjunction with safety concerns, has reduced enthusiasm for the use of vitamin E in Alzheimer’s disease. Although a variety of other prescription medications (estrogens, statins), nonprescription medications (nonsteroidal anti-inflammatories), and nutraceuticals (gingko biloba) have been considered for the treatment of Alzheimer’s disease, published data to date on all other treatment options has been at worst negative and at best insufficient to earn regulatory approval.

Other drug categories are commonly used to treat targeted symptoms associated with Alzheimer’s disease. For instance, antipsychotic medications are often used to treat agitated behavior. Some studies do show efficacy in this regard, although other studies have argued the limited behavioral benefits antipsychotics may confer is canceled out by increased morbidity.

Future Directions

Scientific Perspective

In the short term, considerable effort will be directed at additional studies of Aβ dynamics and homeostasis. Research will focus on the toxicities of different degrees of Aβ aggregation (especially oligomers, defined as short, soluble polymers of amyloid), cellular mechanisms of Aβ disposal, and tissue-level mechanisms of Aβ disposal.

Research over the longer term will need to address the fact that the predominant etiologic hypothesis, the amyloid cascade hypothesis, cannot yet explain why Aβ homeostasis changes in most of those affected or how Aβ might give rise to other aspects of Alzheimer’s disease pathology. It is possible the amyloid cascade hypothesis will prove valid in those with early-onset, autosomal-dominant Alzheimer’s disease caused by mutations of the genes encoding APP, presenilin 1, and presenilin 2 proteins, but not the late-onset cases (the vast majority). Disproving the amyloid cascade hypothesis in the late-onset cases will likely require two events. First, interventions that attempt to treat Alzheimer’s disease by targeting Aβ will need to show absent or limited efficacy. Second, other hypotheses better able to explain the overall Alzheimer’s clinical, and pathological big picture will need to demonstrate viability and durability.

Diagnostic Perspective

Because it may prove easier in the future to prevent neurodegeneration rather than reverse it, the ability to render an early, accurate diagnosis is crucial. Also, the ability to treat the disease (either symptomatically or disease modifying) increases the importance of early diagnosis. This has led the field to recently define prodromal and preclinical Alzheimer’s disease categories. A key question is how these new categories, which were proposed for research purposes, should extrapolate to the clinical setting. The use of biomarkers to define how close an asymptomatic or very mildly impaired individual is to clinical Alzheimer’s disease outside of the research setting requires, and is receiving, careful consideration.

Treatment Perspective

None of the treatments approved for use in Alzheimer’s disease are approved for use in MCI/prodromal Alzheimer’s disease, although available data argue cholinesterase inhibition (at least with donepezil) may provide a marginal benefit. Such a benefit would not be surprising, especially if MCI represents very early Alzheimer’s disease in most people.

Over a decade of experience with symptomatic treatment has made it abundantly clear that disease-modifying treatments are required. Most current approaches toward disease modification are targeted to Aβ homeostasis. Inhibition of its production (gamma secretase inhibitors and modifiers), its targeted removal (active and passive immunization approaches), prevention of its aggregation, and enhancement of enzymatic degradation are all under active pursuit. To date, a phase II Aβ vaccination trial (AN1792) was halted when several of the subjects developed encephalitis. Other data obtained through this trial suggest the approach was successful in reducing cerebral amyloid plaques. However, the most extensive published clinical data from AN1792 indicate that 1 year after vaccination, the rate of cognitive decline was similar to (unchanged from or only very slightly reduced from) the rate of decline shown by the placebo group of that trial. Other anti Aβ trials have featured drugs designed to prevent generation of Aβ from APP, drugs designed to prevent the collection of Aβ monomers into oligomers, and the administration of antibodies to Aβ that are intended to remove Aβ from the brain. To date a number of trials of these agents have not succeeded in reaching their primary end point, while other trials are still underway. It is worth noting that anti-Aβ trial design increasingly is focusing on subjects whose brains show Aβ deposition, but who are clinically asymptomatic or only mildly affected from a clinical perspective. The proposed but unproven rationale for this is that if it turns out to be the case that Aβ does initiate neurodegeneration in Alzheimer’s disease, once the neurodegeneration proceeds to a certain point, it may continue to advance whether or not Aβ is present.

If attacking Aβ fails to meaningfully benefit Alzheimer’s disease patients, the validity of the amyloid cascade hypothesis in late-onset, sporadic Alzheimer’s disease will be called into question. If this happens, new models for drug design will be needed. Currently, mice expressing a mutant APP transgene, sometimes in conjunction with other mutant human transgenes, serve as the gold standard for preclinical testing of potential Alzheimer’s disease treatments.

Cross-References