See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/7331198

Pathophysiology of Alzheimer's Disease

Article  in  Neuroimaging Clinics of North America · December 2005

DOI: 10.1016/j.nic.2005.09.009 · Source: PubMed

CITATIONS READS

47 16,836

3 authors:

Bruno P. Imbimbo Jay Lombard

Chiesi Farmaceutici S.p.A. Genomind, LLC
295 PUBLICATIONS   8,284 CITATIONS    14 PUBLICATIONS   330 CITATIONS   

SEE PROFILE SEE PROFILE

Nunzio Pomara
Nathan Kline Institute-NYU School of Medicine
237 PUBLICATIONS   4,106 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Mutated form of Human Nerve Growth Factor View project

Pharmacogenetic Testing View project

All content following this page was uploaded by Bruno P. Imbimbo on 13 October 2017.

The user has requested enhancement of the downloaded file.

 Neuroimag Clin N Am 15 (2005) 727 – 753

Pathophysiology of Alzheimer’s Disease

Bruno P. Imbimbo, PhDa, Jay Lombard, DOb, Nunzio Pomara, MDc,d,T
a
Research and Development Department, Chiesi Farmaceutici, Parma, Italy
b
Department of Neurology, Weill Medical College of Cornell University, New York, NY, USA
c
Geriatric Psychiatry Division, Nathan S. Kline Institute for Psychiatric Research, Orangeburg, NY, USA
d
Department of Psychiatry, New York University School of Medicine, New York, NY, USA

Alzheimer’s disease (AD) is the most common AD is a neurodegenerative disorder characterized

form of senile dementia, affecting 10% of individuals clinically by progressive deterioration of cognition,
older than 65 and nearly 50% of those older than 85. behavior, and functionality that impairs activities
The prevalence of AD in the United States is of daily living significantly. Morphologically, the
estimated to be approximately 4.5 million and is disease is characterized by brain atrophy and by en-
predicted to increase to up to 13.2 million in the next larged cerebral ventricles. From a biochemical point
50 years. Although basic research in AD has made of view, the most clear-cut and consistent finding is
remarkable progress in the past 2 decades, currently a deficit of the cholinergic system, decreased levels
available drugs are able only to improve cognitive of choline acetyltransferase, and other cholinergic
symptoms temporarily, and no treatment can reverse, markers. Histologically, AD is characterized by extra-
stop, or even slow the inexorable neurodegenerative cellular deposits, called cerebral plaques, composed
process. The pathophysiology of AD is associated of a dense proteinaceous core containing the Ab pep-
with a variety of factors, including the extracellular tide surrounded by dead and damaged neurons. The
deposition of b-amyloid (Ab) plaques, accumulation other typical histopathologic hallmarks of AD are the
of intracellular neurofibrillary tangles, oxidative neurofibrillar tangles within neurons, formed mainly
neuronal damage, and inflammatory cascades. It is by a filamentous, hyperphosphorylated form of the
widely believed, however, that an increase in the microtubule-associated protein t. Whereas most
production of the Ab peptide, the main component of cases of AD occur sporadically, approximately 5%
amyloid plaques, is central to the pathogenesis of the of patients develop the disease early as a result of
disease. Since the first description of the neurotoxic fully penetrant autosomal dominant gene mutations.
properties of the Ab peptide, a vast number of studies
have investigated the cellular and molecular pa-
thology triggered by Ab. This article reviews the
Epidemiology of Alzheimer’s disease
basic knowledge of the pathophysiology of AD. Prog-
ress in understanding the cellular and molecular
The incidence and prevalence of AD rise with
alterations that are responsible for the neuron’s death
increasing age and are higher in women in part
may help significantly in developing more effective
because of their increased longevity. The incidence of
medications to slow the progression of this devastat-
AD ranges from 1% at ages 65 to 70 to approximately
ing disease.
4% over age 85. In the United States, the number of
new cases per year is expected to triple from ap-
T Corresponding author. Geriatric Psychiatry Division, proximately 420,000 in 2000 to more than 1.3 million
Nathan S. Kline Institute for Psychiatric Research, 140 Old in 2050 [1]. Estimates of prevalence of AD range
Orangeburg Road, Orangeburg, NY 10962. from the lowest figure of 3% of the population at
E-mail address: [email protected] (N. Pomara). 65 years to the highest reported estimate of 47% of

1052-5149/06/$ – see front matter D 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.nic.2005.09.009 neuroimaging.theclinics.com
728 imbimbo et al

people over age 85. The prevalence of AD in the processes, including misprocessing of the t protein.
United States in 2000 was estimated to be 4.5 million This cascade ultimately causes neuronal dysfunction
[2]. By 2050, this number will increase by almost and death and leads to the clinical and pathologic
threefold, to 13.2 million. In the United States, AD features of AD.
currently is the eighth leading cause of death, with
approximately 63,000 deaths per year and a death rate Amyloid precursor protein
of 21.8 deaths per 100,000 population [3]. The death APP is a type I transmembrane glycoprotein
rate of AD is increasing by approximately 6% per produced in many cells and processed through the
year. The median survival from initial diagnosis secretory or endosomal-lysosomal pathways. Alter-
recently was estimated to be 4.2 years for men and native splicing of APP provides a total of 10 iso-
5.7 years for women [4]. forms, with lengths of 304, 639, 677, 695, 696, 714,
733, 751, 752, and 770 amino acids [10]. The major
APP isoforms are those with 695, 751, and 770 amino
Genetics of Alzheimer’s disease acids. The 695 amino-acid form is expressed prefer-
entially in neuronal tissue. Although the physiologic
The presence in some families of AD individuals functions of APP remain unclear, its ubiquitous
who have an autosomal dominant inheritance pattern expression during development and in several
has allowed for the discovery of disease genes. adult tissues suggests a fundamental role in cellular
Mutations on three genes, known as causative genes, physiology. APP seems to be involved in the static
are fully penetrant and cause aggressive forms of cell-substrate adhesion or neurite outgrowth synap-
early-onset AD. The causative genes are those encod- togenesis, synaptic plasticity, and promotion of neu-
ing amyloid precursor protein on chromosome 21q21 ronal cell survival [11]. APP also is involved in the
(APP), presenilin 1 on chromosome 14q24 (PSEN1), regulation of cell movement [12]. Exposure of cor-
and presenilin 2 on chromosome 1q42 (PSEN2). tical neurons to APP monoclonal antibodies leads to
Mutations in these genes account for approximately neurite degeneration followed by caspase-dependent
5% of the total number of AD cases [5]. There also apoptosis [13]. The soluble secreted form of APP
are other genes, known as susceptibility genes, (sAPPa) shows similarities with growth factors and
believed to be involved in the pathogenesis of AD increases the proliferation of embryonic neural stem
through complex interactions with environmental cells [14].
factors. Allele polymorphism of one of these sus- Mutations of APP are connected to a limited
ceptibility genes, that encoding apolipoprotein E on number of early-onset familial AD cases [15]. To
chromosome 19q13 (APOE), is associated definitely date, 22 single – amino-acid and 1 two – amino-acid
with increased risk for late-onset AD in many studies. (the Swedish mutation K670N/M671L) mutations
It is suspected that several other AD-susceptibility of APP are identified (see Fig. 1). Four of these
genes exist, and their identification is the subject of mutations are not pathogenic. Most of the mutations
ongoing research [5]. Ultimately, a variety of genes are located adjacent or in proximity to the cleavage
probably will be determined to confer risk, each one site of g-secretase. The most frequent APP mutation
responsible for only a small fraction of all cases. is V717I (the London mutation) [16]. Carriers of
Overall, the four genes linked to AD (APP, PSEN1, the so-called ‘‘Arctic’’ mutation (E693G), found
PSEN2, and APOE) account for approximately half in a Swedish family, show decreased Ab42 and
of the total genetic risk [6]. Ab40 levels in plasma but increased protofibril
formation [17]. This APP mutation is highly amy-
Causative genes loidogenic in vivo [18]. The most recently discovered
APP mutations are one found in a Japanese family
Historically, the first identified gene causing within the Ab sequence (D678N) [19] and a novel mu-
AD was that encoding the substrate (APP), from tation at the b-secretase (BACE1) cleavage site [20].
which the Ab peptide is generated [7]. The other
two AD causative genes discovered encode the en- Presenilins
zymes (presenilins) that produce Ab from APP [8,9]. Presenilin 1 (PS-1) and presenilin 2 (PS-2) are
Mutations in genes encoding APP or presenilins two similar 8-domain transmembrane proteins that,
cause dysregulation of APP processing with an in- within the g-secretase complex, catalyze the last
crease of the amount of Ab produced (Fig. 1). The step of the APP cleavage generating the Ab40 and
accumulation of Ab seems to be an early and ini- Ab42 peptides. Six alternative splicing forms of
tiating event that triggers a series of downstream PS-1 are known with 184, 374, 378, 409, 463, and
 pathophysiology of alzheimer’s disease 729

Fig. 1. Amino-acid sequence of PS-1 (blue circles) and of the C-terminal portion of APP (green circles) generating Ab.
Mutations in PS-1 and APP known to cause familial AD are depicted in red. The b, a, g, and e cleavage sites are indicated by
small scissors. The two amino acids involved in the endoproteolytic cleavage of PS-1 are shown in dark blue. The two aspartate
residues of the catalytic site of PS-1 are highlighted in yellow. Exon junctions also are shown. CTF, C-terminal fragment of PS-1;
NTF, N-terminal fragment of PS-1. (From Hardy J, Selkoe DJ. The amyloid hypothesis of Alzherimer’s disease: progress and
problems on the road to therapeutics. Science 2002;297:355; with permission.)

467 amino acids. Two isoforms of PS-2 are se- including spastic paraparesis, language deficits and
quenced with 414 and 448 amino acids. To date, frontal behavioral disturbances [23]. A PS-1 mutation
144 missense mutations of PS-1 are identified, two (G183V) was found to cause clinically frontotempo-
of which are not pathogenic (see Fig. 1). Eleven mu- ral dementia (FTD) and histologically neurofibrillary
tations of PS-2 are known, one of which is not tangle pathology without amyloidosis [24]. Cultured
pathogenic. PS-1 mutations are the cause of the most cells transfected with this mutated presenilin pro-
severe form of dominant familial AD, with complete duced normal amounts of Ab42, an observation that
penetrance and an onset occurring as early as 30 years raised questions as to the presumption that all auto-
of age [5]. Studies in cell lines expressing human somal dominant forms of AD could be linked to the
mutated presenilins, transgenic mice bearing mutated accumulation of Ab in the brain.
human presenilins, and plasma samples and brain The last described presenilin mutations include
sections of patients who have presenilin mutations one (L174R) in two members of a Bavarian family
show that PS-1 and PS-2 missense mutations increase [25] and one (V97L) in members of a Chinese
the production of Ab42, the highly selfaggregating family [26].
and neurotoxic form of Ab [21]. These mutations
alter the conformation of presenilins, apparently
enhancing coordination between the two catalytic Susceptibility genes
aspartates of presenilins (respectively, in position 257
and 385) and the peptide bond between amino acids Although mutations in the APP, PS-1, and PS-2
alanine and threonine of APP [22]. Certain PS-1 genes cause less than 5% of cases of AD, 25% of
mutations are associated with atypical phenotypes, persons who have late-onset forms of the disease
730 imbimbo et al

have had a close relative who has had dementia [5]. are of the e4 type [33]. Conversely, the APOE*E2
Indeed, there are many families that have a heavy seems to protect from AD [32]. Interestingly, there
load of late-onset AD but do not have mutations in are few individuals who have mutated APP who also
APP, PS-1, or PS-2 genes. Thus, intense research carry the APOE*E2 who fail to develop dementia
efforts are dedicated to identifying additional genetic [34]. Nevertheless, the putative protective effect of
factors in the more common late-onset form of the APOE*E2 remains debated [35]. Many genetic
disease. Genetic linkage studies involving hundreds epidemiologic studies robustly confirm an enhanced
of families have produced evidence for several risk for developing AD in subjects who have one or
chromosomal regions suspected of harboring genes two APOE*E4s who are in their sixth or seventh
related to AD. These include regions on chromo- decade of life. The effect is modified by gender,
somes 6, 9, 10, 12, and 19 [27]. So far, careful however, and is age specific, with its peak effect
scrutiny of these regions has resulted in the identi- observed at approximately age 70 [36]. There are
fication of APOE as an unequivocal gene associated octogenarians carrying two APOE*E4s who seem
with late-onset AD. More than 100 additional to have escaped the APOE*E4 effect. The purified
genes have been investigated as susceptibility genes apoE4 isoform is reported to have higher affinity
through genetic-association studies, but none of them for Ab than apoE3 [37], suggesting it may act as a
are confirmed fully [28]. It is estimated that there pathologic chaperone stabilizing the b-sheet structure
are at least six additional susceptibility genes influ- of Ab fibrils [38] and impairing Ab clearance [39].
encing the age at onset and the risk for sporadic AD Elegant experiments in APOE knockout and trans-
[29]. The hunt continues and additional loci are ex- genic mouse models show that the presence of apoE4
pected to be discovered as the result of genomewide increases the amount of fibrillar Ab deposits mark-
screening [16]. edly in the brain compared with the effects of apoE3
[40]. The APOE*E3 genotype, however, is neither
Apolipoprotein E necessary nor sufficient for the occurrence of AD.
Apoliprotein E (apoE) is a 299 – amino-acid pro- Thus, there are other genetic or environmental factors
tein involved in transport and metabolism of lipids that alone or in conjunction with APOE*E4 can
[30]. Ninety percent of circulating apoE derives from modify the risk for AD [27].
the liver. apoE also is synthesized in other organs
(lung, ovary, muscle, spleen, and kidneys), including Susceptibility loci on chromosome 10
the central nervous system, where it is made by glia, Genome screenings aimed at identifying genetic
macrophages, and neurons. Three major isoforms susceptibility risk factors for late-onset AD find a
of apoE are known (apoE2, apoE3, and apoE4) locus on chromosome 10 that modifies risk for AD by
and these are encoded by three natural allelic apoE elevating Ab42 plasma levels independently from the
variants (e2 [APOE*E2], e3 [APOE*E3], and e4 APOE genotype [41,42]. Several strong candidate
[APOE*E4]). The structural differences of different genes (VR22, PLAU, and IDE) support the linkage.
apoE isoforms are small. apoE3 has cysteine in posi- The VR22 gene encodes a-T catenin, which is a
tion 112 and arginine in positions 158. In apoE4, ar- binding partner of b catenin, known to interact with
ginine replaces the cysteine in position 112. In apoE2, PS-1. VR22 has variants that influence high plasma
the arginine in position 158 is replaced by cysteine. Ab42 levels [43].
The most common APOE allele is APOE*E3, present The IDE gene encodes insulin-degrading en-
in 40% to 90% of the population. The APOE*E3 is zyme, which has a central role in the degradation
present in 10% to 15% of the general population with and clearance of Ab by microglial cells and neurons
frequency higher in northern than in southern Euro- [44]. Recently, the existence of pathologic variants in
pean regions [31]. APOE*E2 is the least frequent the region harboring IDE that influences late-onset
allele (approximately 10%). AD with high plasma Ab42 levels has been sug-
The polymorphism of the APOE gene is con- gested [45].
sidered the most important, best-documented, genetic The PLAU gene encodes urokinase-type plas-
susceptibility risk factor for late-onset AD [5]. Indi- minogen activator, which converts plasminogen to
viduals who carry the APOE*E4 have an increased plasmin. Ab aggregates induce PLAU expression,
risk for developing AD compared with noncarriers, thereby increasing plasmin, which degrades aggre-
with a threefold to fourfold higher risk for het- gated and nonaggregated forms of Ab. Single
erozygotes and a ninefold to tenfold higher risk for missense mutation (P141L) in urokinase-type plas-
homozygotes [32]. In the population of patients minogen activator recently was suggested to increase
who have AD, approximately 45% of APOE alleles Ab42 in late-onset AD [46].
 pathophysiology of alzheimer’s disease 731

Susceptibility loci on chromosome 12 (409 amino acids), t-D (382 amino acids), t-E
A genomic screen of families who have late-onset (411 amino acids), and t-F (440 amino acid). Isoform
AD reveals a susceptibility locus on chromosome 12 PNS-t is expressed in the peripheral nervous system
not linked to the presence of the APOE*E4 [47]. (PNS), whereas the others are expressed in the central
The genes encoding a2-macroglobulin (a2M) nervous system. Protein t is a major component of
and low-density lipoprotein receptor-related pro- the intracellular filamentous deposits (neurofibrillar
tein (LRP) present on this chromosome initially tangles) that histologically define not only AD but
were considered suspects. Later, it was proposed also other neurodegenerative diseases, including pro-
that a genetic variation in a transcriptional factor, gressive supranuclear palsy, corticobasal degenera-
LBP-1c/CP2/LSF, is the susceptibility factor for late- tion, Pick’s disease, and argyrophilic grain disease.
onset AD [48]. a2M is a serum protease inhibitor, Neurofibrillar tangles consist mostly of paired helical
also expressed in the brain implicated in AD on the filaments of protein t in the hyperphosphorylated,
basis of its ability to mediate the clearance and deg- insoluble form [58]. For a long time, it was unclear
radation of Ab [49]. a2M polymorphisms resulting if the dysfunction of protein t follows disease or if
from a deletion allele or an amino-acid substitution disease follows t dysfunction. This was resolved
(V1000I) are hypothesized to explain its association when mutations in gene encoding t on chromosome
with AD. More recently, new population-based stud- 17q21 (MAPT) were found to cause the inherited
ies are published in which a2M deletion/insertion FTD and parkinsonism (FTDP-17), historically also
polymorphisms are not associated with risk for devel- termed Pick’s disease [59]. Currently, 32 different
oping AD [50,51] . Thus, it may be possible that the mutations are identified in more than 100 families
locus on chromosome 12 responsible for AD sus- [60]. Approximately half of the known mutations
ceptibility does not correspond to a2M but to an ad- reduce the ability of protein t to interact with mi-
jacent gene. crotubules and increase its propensity to assemble
The low-density LRP is a multifunctional recep- into abnormal filaments. The other mutations perturb
tor that is highly expressed in the brain. LRP is the the normal ratio of protein t isoforms [61]. Although
main apoE receptor expressed in neurons, mediates no mutations in protein t are yet associated with AD,
neurite outgrowth, is responsible for the endocytosis the abundance of neurofibrillar tangles correlates with
of secreted APP, and is detected in senile plaques, the degree of neurodegeneration. The presence of
dystrophic neuritis, and reactive astrocytes in AD neurofibrillar tangles in AD may manifest a ‘‘down-
brain [52]. Recently, it was shown that LRP interacts stream’’ response to the pathologic events initiated
with b-secretase on the cell surface and acts as a by this disease [6]. In a triple transgenic mouse model
b-secretase substrate [53]. Genetic association be- of AD harboring mutant genes for APP (K670N/
tween two different LRP polymorphisms (a TTTC M671L), PS-1 (M146V), and protein t (P301L), Ab
repeat and a cytosine to thymine silent substitution) deposition develops before the tangle pathology [62].
and AD is reported. These associations, however, are Similarly, studies in double transgenic mouse express-
not confirmed in other studies and the link remains ing the Swedish mutation of APP and the t protein
controversial [54]. containing three of the mutations present in FTDP-17
Several members of the glyceraldehyde-3-phos- show that Ab facilitates the phosphorylation of t and
phate dehydrogenase (GAPD) gene family are found, its aberrant aggregation [63].
including one within the 12p candidate region
associated with late-onset AD [55]. GAPD is a key
enzyme in cellular energy production and also plays Other susceptibility genes
an important role in several other cellular processes, Many other genes emerging from studies on
including neuronal apoptosis and neurodegenerative cohorts of sporadic AD cases and age- and sex-
diseases, including AD. It is known to bind APP [56] matched controls are hypothesized as associated to
and Ab [57]. sporadic AD.
Neprilysin is a zinc metalloproteinase reported
Microtubule-associated protein s as a major Ab-degrading enzyme expressed in the
Microtubule-associated protein t promotes micro- brain [64]. The decreased expression and activity of
tubule assembly and stability and is involved in the neprilysin may contribute to the development of AD
establishment and maintenance of neuronal polar- by promoting the accumulation of Ab. The gene
ity. Eight isoforms of protein t are known: PNS-t encoding neprilysin (NEP) is located on chromosome
(757 amino acids), fetal-t (315 amino acids), t-A 3q25. At least 10 single nucleotide polymorphisms
(351 amino acids), t-B (380 amino acid), t-C in the NEP are found to be associated to sporadic
732 imbimbo et al

AD, suggesting that genetic variations within or ex- Nongenetic risk factors for Alzheimer’s disease
tremely close to the NEP might influence the sus-
ceptibility to AD [65,66]. AD is caused by a combination of genetic and
Ubiquilin 1 (UBQLN1) is a highly expressed environmental factors, and the role of environment
protein that promotes the accumulation of uncleaved cannot be ignored. A lot of effort has been devoted
PS-1 and PS-2 by stimulating their biosynthesis. The to identifying those specific genetic and environ-
gene encoding UBQLN1 is one of several candidate mental factors, determining their relative importance,
genes for AD located near a well-established linkage understanding their interactions, and capitalizing on
peak on chromosome 9q22. Variants in UBQLN1 this knowledge to treat and prevent the disease.
substantially increase the risk for AD, possibly by
influencing alternative splicing of this gene in the
Traumatic head injury
brain [67].
b-secretase is the enzyme the starts the first
Several studies suggest a link between risk for
cleavage of APP to generate Ab. The gene encoding
developing AD and traumatic head injury, but the
b-secretase (BACE1) is located on chromosome
association is controversial [75]. Human postmortem
11q23 near a region with increased lod scores for
and experimental studies show Ab deposition and
AD. The biologic functional and genetic associa-
t pathology after head injury. A pig model of brain
tion studies indicate that the BACE1 gene may be a
injury shows that even a mild trauma with full re-
genetic risk factor for late-onset AD. BACE1 gene
covery leads to the accumulation of Ab and formation
polymorphism C786G may act as an APOE*E4-
of neurofilament inclusions [76]. An overexpression
dependent risk factor for developing late-onset AD in
of APP is observed in the brains of gunshot survivors
Chinese populations [68].
[77] and in head-injured sheep [78]. Ab levels in
Tumor necrosis factor-a (TNF-a) is a potent im-
the cerebrospinal fluid also are increased in patients
munomodulator that may mediate neurotoxicity [69],
who have traumatic head injury and in animal models
which makes it an appropriate AD candidate gene.
of brain trauma. Studies involving large groups of par-
The gene encoding TNF-a is located near a chro-
ticipants, however, such as the Multi-Institutional Re-
mosomal region, 6p21, which shows genetic linkage
search in Alzheimer Genetic Epidemiology (MIRAGE)
association with AD in several of the full-genome
study [79] and Rotterdam [80] studies, provide con-
screens in AD [28]. Recent studies, however, suggest
tradictory results. A recent analysis of 15 case-control
that TNF-a may act as a risk factor only for vascular
studies supports an association between a history of
dementia [70] and that 308 adenine/guanine (A/G) poly-
previous head injury and the risk for developing AD,
morphism is not associated with late-onset AD [51].
but this association was restricted to male patients
a1-antichymotrypsin (ACT) is a serine protease
[81]. The presence of APOE*E4 is associated with
inhibitor and an acute phase protein. ACT is
worse neurologic impairment in head injury [82] but
synthesized predominantly in the liver and other
the influence of the APOE genotype on the prognosis
tissues, including the brain. ACT is a major compo-
of traumatic brain injury is under discussion [75].
nent of the senile plaques and may act as a pathologic
chaperone, promoting Ab assembly into neurotoxic
fibrils. Increased levels of ACT are found in the brain Hypercholesterolemia and high-density lipoprotein
and peripheral blood of patients who have AD [71]. cholesterol
ACT gene is located on chromosome 14q32. It con-
tains common bi-allelic polymorphisms, adenine/ The relationship between total serum cholesterol,
thymine (A/T) transversion, resulting in either an ala- its high- and low-density lipoprotein fractions, and
nine or a threonine at codon 15 in the signal peptide AD is the subject of several epidemiologic studies.
sequence. The alanine allele is associated with an The results of these studies sometimes are contrast-
increased risk for AD only in APOE*E4 carriers [72]. ing. A small Finnish study of 444 elderly men sug-
Other studies, however, do not confirm these findings gests that regardless of genetic status, high serum
[73]. More recently, reports suggest that the T allele of cholesterol represents an independent risk factor
a guanine/thymine (G/T) polymorphism at position [83]. Similarly, another small French study involving
51 of ACT is associated with an increased risk for 334 elderly subjects suggests that regardless of APOE
early-onset of AD independent of the presence of the status and other potential confounding variables,
APOE*E4. After manifestation of the disease the ACT elevated high-density lipoprotein cholesterol reduces
TT genotype also is associated with faster cognitive the risk for AD significantly [84]. Alternatively,
decline in patients who have the APOE*E4 [74]. other epidemiologic studies show that lipid levels
 pathophysiology of alzheimer’s disease 733

are lower in persons who develop AD, not higher. For Homocysteine, folate, and vitamin B12
example, a study of 987 elderly subjects, most of
them of Caribbean Hispanic origin, concludes that Elevated serum homocysteine levels are consid-
total cholesterol has a weak but significant inverse ered a risk factor for cognitive decline and AD, but
association with incident AD, regardless of APOE data from prospective studies are controversial.
genotype [85]. A recent study of 392 subjects shows Serum concentration of homocysteine increases with
that high cholesterol in late life is associated with age. In healthy elderly subjects, plasma total homo-
decreased dementia risk [86]. The conflicting results cysteine concentrations are correlated negatively with
may be explained by the timing of the cholesterol cognitive performance [101,102] and high plasma
measurements in relationship to age and clinical onset homocysteine levels are associated with brain atrophy
of dementia. Studies in patients taking cholesterol- [103]. High homocysteine levels are associated with
lowering drugs, such as 3-hydroxy-3-methylglutaryl low vitamin B12 and folate levels but it is unknown if
coenzyme A reductase inhibitors (statins), also are it is homocysteine toxicity or vitamin insufficiency
contradictory. Some studies conclude that statin use that is responsible for the observed cognitive dys-
lowers the risk for dementia [87], whereas others function. In healthy elderly subjects, low folate levels
conclude that this association is artifactual [88]. appear to be a risk factor for cognitive decline [104].
Results of epidemiologic studies prompted many Vitamin B12 and folic acid supplementation normalize
laboratory studies. In vitro, APP751-transfected cells plasma total homocysteine concentrations but do not
with elevated cholesterol levels show a dramatic affect cognitive performance [105]. In patients who
reduction in secretion of sAPPa, the nonamyloido- have AD, plasma homocysteine levels are described
genic metabolite of APP [89]. In cultured neurons, as significantly higher than in controls [106] but
cholesterol depletion obtained by combination of this observation is not confirmed in other studies
lovastatin treatment and methyl-b-cyclodextrin [107,108]. No correlations between homocysteine
extraction, reduces production of Ab below detect- levels and cognitive functioning are found [109] and
able levels [90] by shifting the APP processing high homocysteine levels are not found associated
to a-secretase [91]. At least some of the actions with a decrease in memory scores over time [107].
of statins on a-secretase seem to occur via the Hence, there is little evidence to justify treating cog-
rho/ROCK pathway, which can modulate sAPPa nitive impairment with vitamin B12 or folate supple-
generation [92]. In contrast, recently it was shown mentation. This is consistent with the findings from
that statins may determine an accumulation of intra- recent systematic reviews of randomized double-blind
cellular amyloidogenic metabolite of APP (sAPPb) trials, which have not found any evidence of poten-
and Ab via an isoprenoid-dependent mechanism tial benefit of vitamin supplementation. A systematic
[93]. Adding complexity to the picture, recently it review of the literature establishes that folic acid plus
was shown that in human neuroblastoma cells, treat- vitamin B12 is effective in reducing the serum homo-
ment with statins decreased the association of the cysteine concentrations, but there is no beneficial
g-secretase complex significantly with lipid rafts effect of folic acid on cognition in older healthy
[94]. In vivo, rabbits on a 2% cholesterol diet devel- women and in patients who have mild to moderate
oped increased Ab levels within the neurons of their cognitive decline and different forms of dementia
brains [95]. Studies in transgenic animal models of [109]. A recent study shows that homocysteic acid, an
AD show that fat feeding increases brain plaque oxidized metabolite of homocysteine, induces intra-
load and that hypocholesterolemic agents, such as neuronal accumulation of Ab42 that is associated with
statins, lower plaque burden [96,97]. cytotoxicity [110]. But further research is required to
Definitive answers on the role of cholesterol in AD establish if raised serum total homocysteine is a cause
can be obtained only from clinical studies. Indeed, or consequence of disease [111].
quite a few prospective clinical trials of cholesterol-
lowering statins in AD patients are available. Most Alcohol
of these studies are negative. In addition, two large
randomized, placebo-controlled clinical trials fail to Light to moderate alcohol, in particular wine, in-
show any effect of statins (simvastatin and prava- take may be related to a low risk for AD [112]. This is
statin) on cognitive performance or on dementia de- apparent in two major studies. A French population-
velopment [98,99]. Perhaps the only encouraging based prospective study involving 3777 subjects
evidence, albeit preliminary, is the apparent ability finds a protective effect of wine consumption on AD
of atorvastatin to slow progression of dementia in with adjusted odds ratios of 0.55 and 0.28 for sub-
AD [100]. jects consuming 1 to 2 and 3 to 4 glasses of wine per
734 imbimbo et al

day, respectively [113]. These results are confirmed Prevalent pathophysiologic hypothesis:
from the analysis of the Rotterdam Study involving the B-amyloid cascade
5395 subjects. Light to moderate drinking (one to
three drinks per day) is associated significantly with a Neuritic plaques, the characteristic lesions found
lower risk for any dementia, with a hazard ratio of in the brain of patients who have AD, are composed
0.58 [114]. A third, smaller study involving 402 sub- mainly of aggregates of a peptide with 40 or
jects again finds that light to moderate drinking is 42 amino-acid residues, known as Ab. The Ab pep-
associated significantly with a decreased risk for AD tide is the result of the metabolic processing of a
compared with nondrinking, with an adjusted relative complex transmembrane glycoprotein, known as
risk 0.50 [115]. A further study involving 980 sub- APP. APP may be processed metabolically according
jects reveals that the association between wine to two pathways (Fig. 2). In the so-called ‘‘nonamy-
consumption and lower risk for AD is confined to loidogenic pathway,’’ the a-secretase enzyme cleaves
individuals who do not have the APOE*E4 [116]. APP within the Ab sequence and releases its trans-
Intake of liquor, beer, and total alcohol is not membrane fragment, sAPPa, which seems to exert
associated with a lower risk for AD. Other studies neuroprotective activity. In the amyloidogenic path-
find a modest, although significant, protective effect way, the b-secretase enzyme releases sAPPb plus a
of alcohol intake, with an odds ratio of 0.83 [117]. 12-kd protein fragment (C99), which in turn is
cleaved by the g-secretase complex giving way to
Ab. The g-secretase complex consists in a catalytic
Other dietary risk factors component, presenilin, and three cofactors, nicastrin,
anterior pharynx defective 1 (Aph-1), and presenilin
Several other dietary elements and foods are enhancer 2 (Pen-2) (Fig. 3). Presenilin is activated
reported to be either risk or protective factors for the biologically by endoproteolysis that generates
development of AD [118]. Antioxidant vitamins E two heterodimers that each present the key aspar-
and C have received particular attention, because tate residue.
animal and in vitro studies suggest that oxidative The correlation among Ab histopathologic le-
stress has a role in the pathogenesis of AD. Analysis sions, brain cell death, and cognitive deficiency in AD
of the Rotterdam Study finds that high intake of represents the so-called ‘‘Ab hypothesis’’ of the dis-
vitamin C and vitamin E is associated with lower ease. This hypothesis was conceptualized in 1991 by
risk for AD with rate ratios of 0.82 and 0.82, respec- Hardy and Allsop [125] and updated in 2002 by
tively [119]. Another prospective study conducted in Hardy and Selkoe [126] (Fig. 4).
815 individuals suggests that vitamin E from food,
but not other antioxidants, may be associated with The generation of the b-amyloid hypothesis
a reduced risk for AD [120]. This association was
observed only among individuals who did not have The Ab peptide first was sequenced from the
the APOE*E4. A major 2-year, double-blind, placebo- meningeal blood vessels of patients who had AD in
controlled study of 341 patients who had moderate 1984 [127]. A year later, Ab was recognized as the
AD detected a significant delay in clinical worsening primary component of the senile plaques of AD brain
for patients treated with vitamin E (2000 IU/day), [128]. In 1987, it was realized that the gene encoding
although statistical analysis applied to this study APP is located on chromosome 21 [129], the same
is questioned [121]. A recent 3-year, double-blind, gene found in Down syndrome (trisomy 21) that de-
placebo-controlled trial in 769 subjects who had velops the same Ab neuropathology as AD. In 1990,
mild cognitive impairment shows that vitamin E it was found that a mutation in the APP gene causes
supplementation (2000 IU/day) does not reduce the Ab deposition in the cerebral vessel walls of Dutch
progression to AD [122]. Thus, a recent review patients who had hereditary cerebral hemorrhage and
concludes that there are no clear-cut answers as to amyloidosis [130]. These findings provided a genetic
whether or not vitamin E is worth prescribing for framework for the Ab hypothesis (see Fig. 4) [125].
AD [123].
Reports on other dietary factors are inconsis- The development of the b-amyloid hypothesis
tent and there are few large epidemiologic studies
exploring the associations between nutrients and AD. Soon after the generation of the Ab hypothesis, a
Available data do not permit definitive conclusions series of studies supported it. The 28 residue synthetic
regarding diet and AD or specific recommendations peptide homologous to the Dutch APP variant was
for diet modification for the prevention of AD [124]. found to accelerate fibril formation in vitro [131].
 pathophysiology of alzheimer’s disease 735

Fig. 2. Schematic representation of APP processing and Ab accumulation. Mature APP (center, inside dashed box) is
metabolized by two competing pathways, the a-secretase pathway that generates sAPPa and C83 (also known as CTFa) (left)
and the b-secretase pathway that generates sAPPb and C99 (right). The carboxyterminal fragment C99 is cleaved by g-secretase
to generate Ab and the APP intracellular domain (AICD). The carboxyterminal fragment C83 also is cleaved by g-secretase
for the secreted peptides p3 (not shown). Ab aggregates into small multimers (dimers, trimers, and so forth), known as oligomers.
Oligomers seem to be the most potent neurotoxins, whereas the end-stage senile plaque is relatively inert. (From Gandy S. The
role of cerebral amyloid b accumulation in common forms of Alzheimer disease. J Clin Invest 2005;115:1122; with permission.)

Several APP mutations causing familial AD were components of the g-secretase complex) causing
discovered, most of them concentrated at the cleavage familial AD also enhance the processing of APP
sites of APP [132 – 134]. These mutations promote to form amyloidogenic Ab strongly reinforced the
generation of Ab by favoring proteolytic processing hypothesis [21]. The finding that APP transgenic
of APP by b- or g-secretase. The subsequent dem- mice with APOE knocked out show a marked
onstration that mutations in PS-1 and PS-2 (the major reduction in cerebral Ab deposition suggested a role

Fig. 3. Schematic representation of the four components of the g-secretase complex. The g-secretase complex consists of a
catalytic component, presenilin, and three cofactors, nicastrin, Aph-1, and PEN-2. Presenilin is activated biologically by
endoproteolysis that generates two heterodimers that each present the key aspartate residue. (From Gandy S. The role of cerebral
amyloid b accumulation in common forms of Alzheimer disease. J Clin Invest 2005;115:1124; with permission.)
736 imbimbo et al

Missense mutations in APP, PS1, or PS2 genes specific association between this chromosome 10 lo-
cus and the gene encoding insulin-degrading enzyme,
the enzyme responsible for Ab degradation in neu-
Increased Aβ42 production and accumulation rons [138], suggests that impaired Ab catabolism and
clearance may contribute to the risk for late-onset
AD. Recently, it was found that Ab deposition de-
Aβ42 oligomerization and deposition
as diffuse plaques
velops before the tangle pathology in triple transgenic
mice expressing simultaneous mutant genes for hu-
man APP, PS-1 and t protein [139].
In addition to this genetic evidence, there are
Subtle effects of Aβ oligomers on synapses
several biochemical studies that reinforce the Ab hy-
pothesis. Brain deposition of Ab42 plaques pre-
cedes the formation of neurofibrillary tangles or
Microglial and astrocytic activation
(complement factors, cytokines, etc.) overt neuronal loss and the appearance of clinical
symptoms in familial AD with PS-1 mutation [140].
Brain levels of Ab40 and Ab42 are elevated early in
Progressive synaptic and neuritic injury
patients who have AD, and levels of both peptides,
especially Ab42, were correlated strongly with cog-
nitive decline. In the frontal cortex, Ab was elevated
before the occurrence of significant t pathology
Altered neuronal ionic homeostasis;
oxidative injury [141]. Cerebrospinal fluid concentrations of Ab42
were significantly lower in patients who had AD
compared with controls, suggesting a diminished
Altered kinase/phosphatase activities ➤ tangles clearance of the toxic peptide from central nervous
system [142]. Patient who have incipient AD [143]
and patients who have mild cognitive impairment
Widespread neuronal/neuritic dysfunction [144] have higher plasma levels of Ab42 than
and cell death with transmitter deficits matched controls. Finally, expression and activity of
b-secretase, the enzyme that initiates the amyloido-
genic metabolic pathway, seem to be increased in the
Dementia brain of sporadic AD [145,146]. Taken together, these
findings are consistent with the notion that cerebral
Fig. 4. The sequence of pathogenic events leading to AD Ab accumulation is the primary influence in AD and
proposed by the amyloid cascade hypothesis. The curved that the rest of the disease process, including t tangle
violet arrow indicates that Ab oligomers may injure the formation, results from an imbalance between
synapses and neurites of brain neurons directly, in addition
Ab production and Ab clearance.
to activating microglia and astrocytes. (From Hardy J,
Selkoe DJ. The amyloid hypothesis of Alzherimer’s disease:
progress and problems on the road to therapeutics. Science
2002;297:354; with permission.) The role of b-amyloid oligomers

The specific forms of Ab that cause injury to

for Ab in APOE polymorphism, a risk factor for neurons in vivo are not established definitively but
sporadic AD [135]. In 1998, mutations in the gene soluble oligomeric species of Ab are considered the
encoding the t protein were identified as responsible ‘‘bad guys.’’ These soluble species include small
for FTD, suggesting that the neurofibrillary tangles globular structures 2.7 to 6.0 nm in diameter, called
in AD brains likely are deposited after changes in Ab-derived diffusible ligands [147], and curvilinear
Ab metabolism and initial plaque formation [136]. structures, known as protofibrils [148]. Ab oligomers
In addition, transgenic mice overexpressing mutant are elevated strikingly in AD brain tissue [149]
human APP and mutant human t undergo increased and in the brain tissues of transgenic mouse models
formation of t-positive tangles, again suggesting that of AD [150]. In neuronal cell cultures, Ab oligo-
either APP or Ab influences the formation of neuro- mers are found strongly neurotoxic [151]. In vivo,
fibrillary tangles [137]. The discovery of a locus on Ab oligomers rapidly blocked long-term potentia-
chromosome 10 that determines an increase in Ab in tion, a classic experimental paradigm for synaptic
patients who have late-onset AD [41] and the allele- plasticity [152]. Fibrillar aggregates of Ab cause local
 pathophysiology of alzheimer’s disease 737

structural disruption of synapses and neurite breakage cinoma cell line [157]. Ab peptides seem to be
in transgenic mouse models of AD [153]. generated at different subcellular sites with Ab40
solely in the trans-Golgi network [158] and Ab42 in
The role of cholesterol the endoplasmic reticulum and the Golgi compart-
ments [159]. Intracellular Ab seems to be mainly in
Cholesterol is shown to alter APP processing and oligomeric form, mainly dimeric [160].
Ab deposition in vitro and in vivo. The addition of Autoptic studies on brain tissue of patients who
exogenous cholesterol to APP-transfected cells leads have AD show that Ab42 accumulates in pyramidal
to an increase of Ab40 and Ab42 secretion [154]. neurons of hippocampus and entorhinal cortex [161].
Introduction of high cholesterol diet induces Ab With increasing cognitive dysfunction and plaque de-
deposition in rabbits [94] and mice [95]. Other position, intraneuronal Ab42 immunoreactivity seems
studies find, however, that increased dietary cho- to be attenuated [161]. Intracellular Ab deposition
lesterol reduces brain APP metabolism [155]. High is evident before the appearance of paired helical
cholesterol may affect Ab production through the sta- filaments-positive structures, indicating that it is
bilization of APP, b- secretase, or g-secretase within one of the first neurodegenerative alterations in the
the lipid rafts on the cell membrane of neurons. In- AD brain [162].
creased cholesterol also may change the fluidity of Human neuroblastoma cells stably expressing the
the membrane in a way that maintains APP in close Swedish mutation of APP show increased intra-
proximity with b- and g-secretase (Fig. 5). cellular Ab levels [163]. Several PS-2 and few PS-1
mutations lead to enhanced intracellular Ab42-levels
Intraneuronal b-amyloid accumulation with a concomitant decrease in intracellular Ab40
[164]. The ratio of intracellular Ab42:Ab40 also
In addition to the deposition of Ab into extra- increases markedly in cells coexpressing APP and
cellular plaques, there is increasing evidence that Ab either mutant PS-1 or PS-2 [163].
also accumulates intracellularly as monomeric and Studies in transgenic mouse models of AD also
oligomeric species and that this process may be an support the pathophysiologic role of intraneuronal
early event in AD pathophysiology [156]. Studies in Ab. Transgenic mice expressing mutant human APP
cell cultures, postmortem brain tissue of patients who in combination with mutant PS-1 show intracellular
have AD, and transgenic AD mouse models suggest Ab accumulation in hippocampal and cortical neu-
that intracellular Ab may be neurotoxic by stimulat- rons before extracellular plaque deposition [165,166].
ing oxidative stress and apoptotic cell death. As observed in postmortem brain tissue of pa-
First evidenceof intracellular generation of Ab was tients who have AD, the intraneuronal Ab immuno-
reported in human neurons derived from a teratocar- reactivity declines with increased extracellular plaque

Fig. 5. Schematic representation describing the possible role of cholesterol in the APP processing. Nonamyloidogenic
cleavage of APP by a-secretases (ADAM) requires a membrane domain that is poor in cholesterol and rich in phospholipids.
Amyloidogenic cleavage of APP by the b-secretase (BACE) and g-secretases (PS) requires a membrane domain that is cho-
lesterol rich, such as a lipid raft. (From Casserly I, Topol E. Convergence of atherosclerosis and Alzheimer’s disease: in-
flammation, cholesterol, and misfolded proteins. Lancet 2004;363:1140; with permission.)
738 imbimbo et al

accumulation [167]. The intraneuronal Ab deposition neurotoxic effect and stimulates an increase in
in hippocampus seems to be increased fourfold by voltage-dependent C2+ channel current activity, sug-
high fat and cholesterol diet in transgenic mice gesting that Ab may act as a physiologic regulator of
expressing APP with the Swedish mutation [168]. ion channel function in neurons [178]. In vitro studies
Triple-transgenic mice expressing mutant APP in show that spontaneous neuronal activity stimulates
combination with mutant PS-1 and mutant pro- b-secretase, leading to an enhanced secretion of the
tein t show early intraneuronal Ab immunoreactivity Ab peptide in normal neurons and in neurons
together with synaptic dysfunction before plaque or overexpressing the Swedish mutated APP [179]. In
tangle deposition was evident [139]. turn, secreted Ab depresses neuronal activity via
Besides intraneuronal Ab production, cellular inhibition of the NMDA excitatory pathway. This
uptake of Ab from the environment is a second negative feedback loop, in which neuronal activity
mechanism that contributes to intracellular Ab ac- promotes Ab production and Ab secretion decreases
cumulation. Selective intracellular accumulation of synaptic activity, could maintain the physiologic ho-
Ab42 is reported in cells treated with synthetic meostasis of neurons. In AD, neurons might fail to
Ab-peptides and occurs via endocytosis [169]. This be depressed by Ab, leading to a gradual build-up of
accumulation seemd specific for Ab42 that seems neuronal activity and further Ab secretion. Alterna-
resistant to degradation. Ab40 and shorter pep- tively, the production of Ab might become indepen-
tides are eliminated with a half-life of approximately dent from neuronal activity [176]. Other in vitro
1 hour [170]. studies suggest that physiologic production of Ab is
The mechanisms with which intracellular Ab important for neuronal viability [180]. Incubation
may be neurotoxic are not understood completely. with b- or g-secretase inhibitors induces death in
It is reported that intracellular Ab42 correlates neuronal cells but not in non-neuronal cells. The ad-
with apoptotic cell death in brains of patients who dition of Ab40 prevents the toxicity of secretase
have AD [171]. Transgenic mice expressing Ab42 inhibitors in neuronal cells. The protective effect of
in neurons show extensive neurodegeneration with Ab40 is concentration dependent with significant
morphologic and biochemical evidence that neuro- effects at concentrations as low as 10 ppm. More
nal death occurred via apoptosis [172]. Rat cortical recently, it is proposed that Ab, at physiologic
neurons expressing human APP underwent apoptosis concentrations, may stimulate postsynaptically the
as soon as they accumulated intracellularly Ab42 expression of proteins related to synaptic plasticity
[173]. Treatment of cell cultures with apoptosis- via cyclic adenosine monophosphate – responsive,
inducing substances, such as etoposide or melphalan, element-regulated gene expression [181].
led to an increase in cellular Ab42 in damaged
neurons [174]. In addition to apoptotic mechanisms, Evidence against the b-amyloid hypothesis
it seems that stimulation of oxidative stress may
mediate toxicity of intracellular Ab. Human neuro- The Ab hypothesis of AD also is criticized
blastoma cells treated with hydrogen peroxide shows strongly. The most frequent objection is that the
a significant increase of intracellular Ab levels [175]. number of amyloid deposits in the brain does not
There is accumulating evidence suggesting that correlate well with the degree of cognitive impair-
intraneuronal Ab42 is a major risk factor for neuron ment experienced by patients in life [182]. In ad-
loss and a trigger for the Ab cascade of pathologic dition, the histopathologic work of Braak and Braak
events. Thus, Ab pathology may be seen in AD as shows that the distribution pattern and packing
a continuous process from an initial abnormal Ab density of amyloid deposits are of limited signifi-
intracellular accumulation to the well-established ex- cance for differentiation of neuropathologic stages
tracellular Ab aggregation, culminating in the for- [183]. The disease severity correlates much better
mation of amyloid plaques and dystrophic neuritis. with Ab assayed biochemically than with histologi-
cally determined plaque counts, and the concentra-
Physiologic functions of b-amyloid tion of soluble Ab species seems to correlate with
cognitive impairment [184 – 186]. Recently, it was
It is still unclear if Ab has any normal physiologic shown that soluble oligomers of Ab, but not mono-
role in the brain [176]. Ab peptide is present in the mers or insoluble amyloid fibrils, may be responsible
cerebrospinal fluid and in plasma of healthy indi- for synaptic dysfunction in the brains of patients who
viduals throughout life [177]. Studies in cultured rat have AD and in AD animal models. The so-called
cortical neurons incubated with exogenously added ‘‘AD diffusible ligands’’ [147] or protofibrils [148]
Ab peptides indicate that unaggregated Ab40 has no cause subtle injury to cultured neurons. Injection in
 pathophysiology of alzheimer’s disease 739

rats of Ab oligomers can inhibit long-term potentia- then is internalized to be delivered to lysosomes for
tion in the hippocampus, which is required for subsequent degradation or transported for trans-
memory formation [152]. Thus, it is hypothesized cytosis across the blood-brain barrier into plasma
that large polymeric aggregates (such as the amyloid [191]. Recently it was shown that Ab can be trans-
plaques) represent inactive reservoirs of species that ported across the blood-brain barrier and be cleared
are in equilibrium with smaller, putatively neurotoxic from the brain after directly binding LRP [192]. LRP
assemblies (Ab oligomers). favors clearance of Ab40 over Ab42. Specific anti-
Another concern arises from the fact that some LRP antibodies are shown to reduce efflux of Ab40
presenilin mutations that increase Ab production from mice brain by up to 90% [190]. Cerebral amy-
strongly seem to be associated with atypical symp- loid load is doubled in human APP transgenic mice
toms, such as spastic paraparesis (weakness affecting engineered to possess low levels of LRP at the blood-
the lower extremities), rather than cognitive symp- brain barrier and no expression of the critical LRP
toms typical of AD [187]. chaperone, the receptor-associated protein [193].

Receptor for advanced glycation end products

Transportation and degradation of B-amyloid RAGE is a cell surface multifunctional receptor
that binds a large array of different ligands [194].
Enormous progress has been achieved in under- Immunocytochemical studies demonstrate that in AD
standing how Ab is generated from APP and huge the expression of advanced glycation end products is
pharmacologic efforts have been made in trying to elevated in neurons close to Ab deposits and in the
interfere with Ab production. In the past few years, cells of Ab containing vessels. Glycation of Ab
interest also has been directed to the fate of Ab once it markedly enhances its aggregation in vitro, and the
is generated from APP and the pharmacologic glycation of protein t, in addition to hyperphosphory-
approaches to boost the processes involved in the lation, seems to enhance the formation of paired
physiologic clearance of Ab from the brain. Soluble helical filaments. Further, RAGE is found to be a
Ab can be removed from the brain via two basic specific cell-surface receptor for Ab with affinity in
mechanisms: favoring its transportation across the the nanomolar range [195]. The binding of Ab to
blood-brain barrier and augmenting its proteolytic RAGE is shown to stimulate expression of proin-
degradation [188]. flammatory cytokines and to decrease cerebral blood
flow [189]. The active participation of RAGE in the
Transportation of Ab across the blood-brain barrier pathogenesis of AD also is confirmed in RAGE-
engineered transgenic mice. Down-regulation of
Ab can be transported across the blood-brain RAGE decreases the influx of Ab from the periphery
barrier and exported out of the brain into the blood into the central nervous system [196]. Double trans-
stream either by direct binding to LRP or by first genic mice overexpressing mutant human APP and
binding apoE and a2M and then transported by LRP. RAGE display exaggerated neuropathologic findings
Once Ab enters the blood stream, it can re-enter the and early abnormalities in spatial learning and mem-
brain via the receptor for advanced glycation end ory [197]. Conversely, transgenic mice bearing a
products (RAGE) or be delivered via chaperone dominant-negative RAGE construct targeted to neu-
molecules, such as apoE or a2M, to peripheral sites rons crossed with animals with mutant human APP
of degradation, such as the liver or kidney [189]. In display preservation of spatial learning and memory
addition, evidence also exists suggesting that 10% to and diminished neuropathologic changes [197].
15% Ab can enter into the cerebral spinal fluid from
the brain interstitial fluid and onward into the Metabolic degradation of Ab
bloodstream [190].
Current evidence in vivo and in vitro suggests that
Low-density lipoprotein receptor-related protein there are three major players in amyloid turnover:
LRP is an approximately 600-kd multifunctional insulin-degrading enzyme, neprilysin-converting
receptor that is highly expressed in the brain. LRP is enzyme, and endothelin-converting enzyme (ECE),
found in plaque deposits in brains of patients who all of which are zinc metallopeptidases. Other
have AD. A key role for LRP in exporting Ab from proteases also are implicated in amyloid metabolism,
brain is proven in several animal studies. Initially, Ab including angiotensin-converting enzyme, urokinase-
forms a complex with a2M or apoE on the luminal type plasminogen activator, tissue plasminogen acti-
side of endothelium. This Ab complex binds LRP and vator, and plasmin, but for these the evidence is less
740 imbimbo et al

compelling [198]. Ab also can be internalized and the metabolic suppression of the endogenous Ab
degraded by activated microglia in the brain. levels in a gene dose-dependent manner [206]. Con-
Neprilysin-converting enzyme and ECE are versely, transgenic overexpression of neprilysin in
homologous membrane proteins of the M13 pep- neurons significantly reduced brain Ab and amyloid
tidase family, which normally play roles in the plaque formation [202]. Neuronal upregulation of
biosynthesis or metabolism of regulatory peptides. neprilysin in young transgenic mice expressing the
Insulin-degrading enzyme is distinct structurally and Swedish mutated form of APP led to reduction of
mechanistically. The regional, cellular, and subcellu- brain Ab levels and delayed Ab plaque deposition in
lar localizations of these enzymes differ, providing young but not in aged animals with pre-existing
an efficient and diverse mechanism for protecting plaque pathology [207]. Unilateral intracerebral injec-
the brain against the normal accumulation of Ab. tion of a lentiviral vector expressing human nepri-
Reduction in expression levels of some of these lysin reduced Ab deposits by half with respect
proteases after insults (eg, hypoxia and ischemia) or to the untreated side in transgenic mouse models
aging might predispose to the development of AD. of amyloidosis [208]. Finally, the recent findings
Conversely, enhancement of their levels by gene that neprilysin immunoreactivity is decreased in AD
delivery or pharmacologic means could be neuro- brains [209] and that single nucleotide polymor-
protective. Even a small enhancement of Ab metabo- phisms of the gene encoding neprilysin is associated
lism could slow the progression of the disease. with patients who have sporadic AD [66] reinforce
the role of neprilysin in the pathogenesis of AD.
Insulin-degrading enzyme
Endothelin-converting enzyme
Insulin-degrading enzyme (IDE) is a 110-kd pro-
ECE, a member of the M13 family, is a key
tein that, in addition to Ab, hydrolyzes several regu-
component in the regulation of blood pressure and
latory peptides, including insulin, glucagons, atrial
electrolyte balance. ECE-1 also is proposed as Ab-
natriuretic factor, transforming growth factor-a,
degrading enzyme. It is shown that treatment of
b-endorphin, amylin, and APP intracellular domain
endogenous ECE-expressing cell lines with the
[199]. Several studies support the role of IDE in
metalloprotease inhibitor phosphoramidon causes a
Ab degradation [200], the most convincing being
twofold to threefold elevation in extracellular Ab
those in transgenic animals. IDE knockout mice
concentration that is believed the result of inhibi-
show increased cerebral accumulation of endogenous
tion of intracellular Ab degradation [210]. Mice defi-
Ab [201]. Double transgenic mice overexpressing
cient for ECE-1 and a closely related enzyme, ECE-2,
IDE and human mutated APP show a 50% decrease
show increased Ab40 and Ab42 brain levels [211].
in cerebral Ab40 and Ab42 levels and Ab load [202].
Further support of a potential role of ECE in the
Further support for the role of IDE in AD patho-
pathophysiology of AD derives from a large case-
genesis derives from recent reports of its genetic
control study that finds that a functional genetic
association with late-onset AD [45]. Hippocampal
variant (A allele) of the gene encoding ECE reduces
IDE mRNA levels are found lower in subjects who
risk for late-onset AD [212].
have an APOE*E4, suggesting that the genetic risk
conferred by APOE*E4 may be mediated in part by
its effect on IDE expression [203].
Mechanisms of B-amyloid toxicity

Neprilysin Various mechanisms are proposed to explain the

Neprilysin is a type II membrane protein that pathway by which Ab induces neuronal cell death,
hydrolyzes circulating biologically active peptides, including intracellular calcium accumulation, reactive
including enkephalin, cholecystokinin, neuropep- oxygen species (ROS) and nitric oxide production,
tide Y, and substance P [204]. The catalytic site of inflammatory processes, and increased sensitivity
neprilysin is exposed extracellularly. Expression of to apoptosis.
neprilysin in primary cortical neurons led to signifi-
cant decreases in Ab40 and Ab42 levels in culture Oxidative stress
media in a dose-dependent manner [205]. The proof
that neprilysin is the major Ab-degrading enzyme in Oxidative stress is a term indicating processes that
vivo derived from studies in knockout mice, in which lead to uncontrolled production of highly ROS.
neprilysin deficiency resulted in defects in the Increased oxidative stress is an early event in AD
degradation of exogenously administered Ab and in [213]. Cells in the brains of AD patients exhibit
 pathophysiology of alzheimer’s disease 741

abnormally high amounts of oxidatively modified radical (OH*), a highly reactive oxyradical and potent
proteins, lipids, and DNA, especially in the areas in inducer of membrane-associated oxidative stress that
which Ab is abundant [214]. Ab aggregation is contributes to the dysfunction of the endoplasmatic
triggered by the presence of the redox-active ions, reticulum (see Fig. 6) [215].
Fe2+ and Cu+, and this interaction generates hydrogen Attack of DNA by ROS, particularly OH*, leads
peroxide. The formation of hydrogen peroxide leads to the formation of approximately 20 major products.
to lipid peroxidation with the consequent production Of the possible base adducts resulting from oxidative
of 4-hydroxynonenal, a neurotoxic aldehyde that stress, one of the most prominent appears to be
covalently modifies proteins on cysteine, lysine, and the result of the hydroxylation of guanine at the
histidine residues (Fig. 6). Some of the proteins C8 position, leading to the formation of 8-oxoguanine
modified oxidatively by this Ab-induced oxidative and the corresponding formamido-pyridine [216].
stress include membrane transporters, guanosine tri- These markers of oxidative stress are localized to
phosphate (GTP)-binding proteins, and ion channels. neurofibrillary tangles and senile plaques in AD
Oxidative modifications of t by 4-hydroxynonenal brains [217]. DNA damage is a feature of neuron
and other ROS can promote its aggregation and cell death that has been detected in AD and in
thereby may induce the formation of neurofibrillar experimental models of AD. Studies demonstrate an
tangles. Ab also can cause mitochondrial oxidative increased oxidative DNA damage in AD brain. DNA
stress and dysregulation of calcium ion (Ca2+) ho- damage induced by free radicals or enzymatic
meostasis, resulting in impairment of the electron modifications can be a trigger that initiates the cell
transport chain, increased production of superoxide death program.
anion radical (O2 ), and decreased production of ATP.
O2 is converted to hydrogen peroxide by the activity
of superoxide dismutase (SOD). O2 also can interact The role of nitric oxide
with nitric oxide (NO) via NO synthase (NOS) to pro-
duce peroxynitrite (ONOO*). Interaction of hydrogen NO is synthesized from L-arginine by NOS. In
peroxide with Fe2+ or Cu+ generates the hydroxyl tissues, NOS occurs in three isoforms: endothelial

Fig. 6. Schematic overview of the neuronal oxidative stress induced by Ab. The neurotoxic action of Ab involves generation of
ROS and disruption of cellular calcium homeostasis. (From Mattson MP. Pathways towards and away from Alzheimer’s disease.
Nature 2004;430:633; with permission).
742 imbimbo et al

(eNOS), neuronal (nNOS), and inducible (iNOS). In Finally, there also is evidence that Ab can bind
the brain, all three isoforms of NOS are present and and activate the classical complement cytolytic path-
iNOS and eNOS are expressed aberrantly in AD, way [226].
especially in association with Ab deposits [218].
Dysregulation of vascular NO production can occur The role of intracellular calcium
from chronic cerebral hypoperfusion during aging
[219]. Increased deposition of Ab results in increased Ca2+ plays a fundamental role in learning and
mRNA and protein expression of iNOS and genera- memory and also is involved in neuron survival
tion of NO [220]. NO can interact rapidly with and death. The inability of neurons to regulate cal-
superoxide anion O2 forming more reactive peroxy- cium homeostasis is an aspect of AD pathogenesis
nitrite (ONOO ). ONOO can induce lipid peroxi- that seems intimately involved in the dysfunction
dation and functional alterations in proteins and and death of neurons [227]. Several studies in cul-
DNA, eventually leading to neuronal death [216]. tured neurons from patients who have AD and from
transgenic mice bearing AD-causing mutations of
PS-1, PS-2, or APP show alterations of cellular cal-
The role of inflammatory mediators cium homeostasis [228]. These alterations include
increased amount of calcium released by the en-
Several studies indicate that Ab deposition acti- doplasmic reticulum, defect in capacitative calcium
vates a series of inflammatory reactions that may entry, and calcium influx pathway activated by
mediate neuronal death. In the AD brain there are depletion of intracellular stores [229]. These effects
elevated levels of a diverse range of proinflammatory seems linked to the increased production of Ab42 and
molecules. These inflammatory molecules are pro- the decreased levels of sAPPa [215]. Ab may perturb
duced principally by activated microglia and astro- calcium regulation by inducing oxidative stress, which
cytes, which are found clustered within and adjacent impairs membrane calcium pumps and enhances cal-
to the senile plaque. cium influx through voltage-dependent channels
The inflammatory products released by Ab- and ionotropic glutamate receptors (see Fig. 6). Ab
activated microglia include cytokines, such as in- can promote calcium influx by forming channels in
terleukin 1 (IL-1), interleukin 6 (IL-6), TNF-a, and membranes or by activating cell surface receptors
transforming growth factor-b (TGF-b, [221]. Al- coupled to calcium influx. Presenilin and APP mu-
though their expression is induced by the presence tations are a rare cause of AD, however, and there
of Ab, these cytokines also are able to promote the is no evidence that a calcium-regulating action
accumulation of Ab in a vicious circle that fuels the of presenilin is involved in sporadic forms of AD.
progression of the disease [222]. Activated microglia Recent studies suggest, however, that wild-type PS-1
also secrete chemokines, a diverse group of small normally may serve a calcium-regulating function
proteins that controls the recruitment of cytotoxic in the endoplasmic reticulum [230].
and helper T lymphocyte to the sites of inflam- Disturbances of calcium regulation in AD may not
mation. These chemokines include interleukin 8 be limited to neurons. Studies in lymphocytes from
(IL-8), interferon-g – inducible protein, macrophage patients who have familial or sporadic AD demon-
inflammatory protein-1a, macrophage inflammatory strate abnormalities in calcium signaling similar to
protein-1b, and monocyte chemoattractant protein-1 those in neurons [231]. Thus, perturbed cellular cal-
(MCP-1) [223]. Additionally, studies in glial cultures cium homeostasis seems to be a widespread abnor-
show that Ab activated nuclear factor-kB, a tran- mality in familial and sporadic forms of AD that may
scription factor for several inflammatory mediators, contribute to the disease process.
including IL-1b and IL-6 [224].
In AD, the number of reactive astrocytes is Disturbed energy metabolism
increased and the expression of phospholipase A2
in these cells is upregulated, leading to increased Neurons are highly dependent on glucose for ATP
arachidonic acid/prostaglandin inflammatory path- generation necessary for ATP’s various functions, in-
way activity. It is shown that Ab deposition is cluding synaptic transmission. ATP production de-
accompanied by activation of astrocyte to secrete pends on a well-preserved mitochondrial structure
chemokines, in particular MCP-1 and regulated and function. When mitochondria fail to provide ade-
on activation, normal T-cell expressed and secreted quate energy, partial neuronal depolarization and loss
(RANTES), which serve as potent macrophage che- of calcium homeostasis occur and may lead to neu-
moattractants [225]. ronal apoptosis. Biochemical studies show that the
 pathophysiology of alzheimer’s disease 743

mitochondrial electron transport chain is defective in deposition may contribute to the cerebral hypo-
sporadic AD [232]. Mitochondrial dysfunction is perfusion observed early in the AD process.
linked to oxidatively damaged mitochondrial DNA
and eventually may lead to cell death. How exactly
Dysfunction of the axonal transport
this translates into loss of specific neuronal popula-
tions, including cholinergic forebrain, hippocampal
Many of the neurons affected in AD are relatively
pyramidal and cortical neurons, is unclear.
large and have long axons. Damage to axons of
Brain imaging studies demonstrate deficits in glu-
such neurons with alterations in axonal transport is
cose use in living patients who have AD, an abnor-
observed in the early stages of patients who have AD
mality that may occur before the onset of clinical
and also in young AD transgenic mice [243]. Axonal
symptoms [233]. The activities of key enzymes in-
defects consist of swellings that accumulated abnor-
volved in energy metabolism (cytochrome c oxidase,
mal amounts of microtubule-associated and molecu-
pyruvate dehydrogenase complex, and a-ketoglutarate
lar motor proteins, organelles, and vesicles. APP
dehydrogenase complex) are decreased in brain cells
and its proteolytic enzymes, PS-1 and b-secretase,
of patients who have AD. It is proven that altered
are transported axonally to synaptic terminals in
proteolytic processing of APP may contribute to im-
brain regions affected in AD [244]. Reductions in
paired energy metabolism. Transgenic human APP
microtubule-dependent transport of APP may stimu-
mice show age-dependent decrease in glucose me-
late its proteolytic processing with the consequent
tabolism in brain regions involved in cognitive
accumulation of Ab [245]. Ab can have many adverse
processes [234]. Moreover, hypoxic tolerance is
effects on the functions and integrity of pre- and
decreased significantly in young human APP mutant
postsynaptic terminals, including inducing oxidative
mice, suggesting an early role for perturbed energy
stress, impairing calcium homeostasis and perturbing
metabolism [235]. Insulin resistance, a sign of sys-
the functions of mitochondria and the endoplas-
temic abnormality in glucose regulation, is consid-
mic reticulum.
ered a risk factor for AD [236]. Caloric restriction,
which enhances insulin sensitivity, attenuates Ab de-
position in transgenic mouse models of AD [237]. All Activation of apoptotic death
these findings support a role for perturbed glucose
metabolism in AD pathogenesis, although this re- Apoptosis is a form of programmed cell death that
mains to be established in humans. involves changes in the cytoplasm, endoplasmic
reticulum, mitochondria, and nucleus. Typically it
includes the production or activation of proteins, such
Cerebral hypoperfusion as p53, Bax, Bad, and Par-4. These proteins increase
the membrane permeability of mitochondria with the
Cerebral blood flow is impaired abnormally release in cytoplasm of cytochrome c and apoptosis-
in AD. Neuroimaging detection of cerebral hypo- inducing factor. The membrane permeability of
perfusion in selected brain regions seems to predict endoplasmic reticulum also is increased with the
AD at the mild cognitive impairment stage, and pos- consequent release of calcium. The latter events then
sibly even earlier, with consistent accuracy. Brain activate cysteine aspartyl proteases, called caspases,
hypoperfusion initiates oxidative stress and cogni- which cleave various protein substrates. Caspase-
tive decline and neurodegeneration is reinforced fur- mediated cleavage of cytoskeletal proteins and ion
ther [238]. channel proteins causes cell shrinkage and may
Rodent models of ischemic brain show that brain prevent necrosis, whereas cleavage of DNA de-
hypoperfusion enhances APP mRNA expression grades chromosomes.
[239] and increased APP accumulation in surviving Presenilin mutations causing familial AD render
neurons [240]. In vitro experiments show that freshly neurons vulnerable to apoptosis induced by Ab and
solubilized Ab enhances constriction of cerebral other stimuli, apparently by altering calcium regu-
and peripheral vessels. In vivo, intra-arterially infused lation of endoplasmic reticulum [246]. APP mu-
freshly solubilized Ab40 in rats results in decreased tations also are sufficient to trigger apoptosis in
blood flow and increased vascular resistance specifi- cultured cells [247]. Different studies show that
cally in cerebral cortex [241]. A decreased cerebral extracellular amyloid deposits and intracellular Ab
hemodynamic response to electrical stimulation is may activate caspases, leading to cleavage of cyto-
documented in a human APP transgenic mouse skeletal proteins, including protein t [248]. In addi-
model of AD [242]. These data suggest that Ab tion to Ab, C-terminal proteolytic products of APP
744 imbimbo et al

are implicated in neuronal apoptosis [247]. Increased campus of AD brains, suggesting that reduction of
expression of p53, a DNA damage indicator, is ob- this enzyme may be a critical factor in the hyper-
served in neurons exposed to Ab [249]. Indeed, neu- phosphylation of t in AD [258].
rons treated with inhibitors of p53, agents that Several studies correlate the severity of AD with
stabilize mitochondrial and endoplasmic reticulum the accumulation of neurofibrillary tangles rather than
membranes or caspase inhibitors, are resistant to Ab [259]. Coexistence of t and Ab in the majority of
being killed by Ab [250]. Other studies suggest AD brains, however, suggests a pathologic inter-
that Ab increases the expression of the death effector action between these two proteins. This hypothesis
Bax and simultaneously downregulates the antiapo- is supported by a transgenic mouse model in which
ptotic bcl2 [251]. animals are crossbred to express mutant t and APP.
In these mice, neurofibrillary tangle pathology is
enhanced substantially in the limbic cortex [137], sug-
gesting a synergistic neurodegenerative effect be-
Minor pathophysiologic hypothesis: the T-protein tween these two proteins.
cascade Conversely, there are several neurodegenerative
disorders in which t is regarded as the primary
Neurofibrillary tangles consist of paired helical pathologic hallmark feature. These taupathies in-
filaments found in the cytoplasm of neurons primarily clude frontotemporal dementia (FTD), Pick’s disease,
in the hippocampus [252]. The formation of neuro- corticobasal degeneration, and progressive supra-
fibrillary tangles in AD is related directly to abnormal nuclear palsy. FTD, although less common than AD,
function of protein t. t is a microtubule-associated accounts for a significant percentage of patients who
phosphoprotein that plays a critical role in supporting have dementia, especially those who present with
axonal transport and promotion of cell stability. The earlier onset and primary psychiatric symptoms. FTD
major polypeptides of paired helical filaments is linked to mutations in the t gene with subsequent
are microtubule-associated protein t. The level of t aggregation of t and neurodegeneration. In compari-
in AD neocortex is sevenfold higher than in aged son, mutations in t are not implicated in AD.
control brain and this increase is in the form of
abnormally phosphylated protein [253]. The hyper-
phosphorylation of t results in reduced binding of
t to microtubules and subsequent disruption of Summary
axonal transport [254]. It is proposed that abnormal
hyperphosphorylation of t results in neurodegenerate In the past 2 decades, huge progress has been
effects via inhibition of microtubule function and made in understanding the cellular and molecular
impairments of neuronal axonal transportand a toxic processes that initiate and feed the AD cascade. The
gain of function in which hyperphosphorylated enhanced knowledge of the pathophysiology of AD
t forms inert polymers, which become neuro- lays the groundwork for significant improvements in
fibrillary tangles. the diagnosis and pharmacologic treatment of the
Several protein kinases are implicated in the disease. Fundamental advancement has been made in
pathologic hyperphosphylation of t in AD, including understanding the genetic basis of the disease with
glycogen synthase kinase-3 (GSK-3) and cyclin- the identification of causative genes (APP, PS-1, and
dependent kinase. GSK-3 is demonstrated to phos- PS-2) for early-onset familial AD. In addition, several
phorylate endogenous t [255] and plays a role in the susceptibility genes are proposed as associated with
brain by regulating cytoskeletal processes, synaptic the late-onset form of the disease, but only the
plasticity, and modulation of microtubule dynamics. polymorphism of the APOE gene consistently is
Neurons exhibit strong GSK-3 immunoreactivity in confirmed to play a role in the most frequent form of
AD compared with normal controls and is found to the disease. The recognition that causative genes of
colocalize in areas of granulovaculear degenera- familial AD encode the substrate and the enzymes
tion [256]. that generate the Ab peptide and that mutations of
Specific protein phosphatases, including phospha- these genes cause brain accumulation of Ab, the most
tase 2A (PP-2A), are shown to have reduced activity prominent histopathologic hallmark of the disease,
in AD brains. PP-2A is demonstrated in in-vitro represents the most convincing argument for the
studies to dephosphorylate several phosphoserine generation of the Ab hypothesis of AD. Although
residues in t [257]. Furthermore, a recent study other pathophysiolgic pathways are proposed for
demonstrates reduced PP-2A expression in the hippo- AD (protein t hyperphosphorylation, oxidative
 pathophysiology of alzheimer’s disease 745

stress, inflammatory cascade, and so forth) the most [11] Coulson EJ, Paliga K, Beyreuther K, et al. What
prominent one is the Ab hypothesis. Detailed under- the evolution of the amyloid protein precursor super-
standing of cerebral degeneration and accumulation gene family tells us about its function. Neurochem
Int 2000;36:175 – 84.
of Ab has generated the most promising hopes for the
[12] Sabo SL, Ikin AF, Buxbaum JD, et al. The Alzheimer
discovery of disease-modifying treatments. This goal
amyloid precursor protein (APP) and FE65, an
has been facilitated greatly by the development of APP-binding protein, regulate cell movement. J Cell
well-characterized and reproducible transgenic ani- Biol 2001;153:1403 – 14.
mal models of the disease. Recent progress in the [13] Rohn TT, Ivins KJ, Bahr BA, et al. A mono-
refinement of the Ab hypothesis includes the identi- clonal antibody to amyloid precursor protein in-
fication of the potential toxicologic role of intra- duces neuronal apoptosis. J Neurochem 2000;74:
cellular Ab. The recent elucidation of the physiologic 2331 – 42.
pathways of Ab clearance and degradation may [14] Caille I, Allinquant B, Dupont E, et al. Soluble form
generate new biologic targets for drug discovery. of amyloid precursor protein regulates proliferation
of progenitors in the adult subventricular zone. Devel-
Progress needs to be made, however, in achieving a
opment 2004;131:2173 – 81.
clear understanding of the mechanisms that link Ab
[15] Eckman CB, Mehta ND, Crook R, et al. A new patho-
accumulation and neuronal death. Ultimately, the genic mutation in the APP gene (I716V) increases
only way to validate the Ab hypothesis fully is to the relative proportion of Ab42(43). Hum Mol Genet
identify Ab interfering drugs that are effective in 1997;6:2087 – 9.
inhibiting disease progress. We predict that the next [16] Finckh U, Kuschel C, Anagnostouli M, et al. Novel
5 years will be crucial in this respect. mutations and repeated findings of mutations in fa-
milial Alzheimer disease. Neurogenetics 2005;6:85 – 9.
[17] Nilsberth C, Westlind-Danielsson A, Eckman CB,
et al. The ‘Arctic’ APP mutation (E693G) causes Alz-
References heimer’s disease by enhanced Ab protofibril for-
mation. Nat Neurosci 2001;4:887 – 93.
[1] Hebert LE, Beckett LA, Scherr PA, et al. Annual [18] Cheng IH, Palop JJ, Esposito LA, et al. Aggressive
incidence of Alzheimer disease in the United States amyloidosis in mice expressing human amyloid pep-
projected to the years 2000 through 2050. Alzheimer tides with the Arctic mutation. Nat Med 2004;10:
Dis Assoc Disord 2001;15:169 – 73. 1190 – 2.
[2] Hebert LE, Scherr PA, Bienias JL, et al. Alzheimer [19] Wakutani Y, Watanabe K, Adachi Y, et al. Novel
disease in the US population: prevalence estimates amyloid precursor protein gene missense mutation
using the 2000 census. Arch Neurol 2003;60:1119 – 22. (D678N) in probable familial Alzheimer’s disease.
[3] Hoyert DL, Kung HC, Smith BL. Deaths: preliminary J Neurol Neurosurg Psychiatry 2004;75:1039 – 42.
data for 2003. Natl Vital Stat Rep 2005;53:1 – 48. [20] Shi XP, Tugusheva K, Bruce JE, et al. Novel muta-
[4] Larson EB, Shadlen MF, Wang L, et al. Survival after tions introduced at the b-site of amyloid b protein
initial diagnosis of Alzheimer disease. Ann Intern precursor enhance the production of amyloid b pep-
Med 2004;140:501 – 9. peptide by BACE1 in vitro and in cells. J Alzhei-
[5] Bird TD. Genetic factors in Alzheimer’s disease. mers Dis 2005;7:139 – 48.
N Engl J Med 2005;352:862 – 4. [21] Scheuner D, Eckman C, Jensen M, et al. Secreted
[6] St George-Hyslop PH, Petit A. Molecular biology and amyloid b-protein similar to that in the senile plaques
genetics of Alzheimer’s disease. C R Biol 2004;328: of Alzheimer’s disease is increased in vivo by
119 – 30. the presenilin 1 and 2 and APP mutations linked
[7] Goate A, Chartier-Harlin M-C, Mullan M, et al. Seg- to familial Alzheimer’s disease. Nat Med 1996;2:
regation of a missense mutation in the amyloid pre- 864 – 70.
cursor protein gene with familial Alzheimer’s disease. [22] Berezovska O, Lleo A, Herl LD, et al. Familial
Nature 1991;349:704 – 6. Alzheimer’s disease presenilin 1 mutations cause
[8] Sherrington R, Rogaev EI, Liang Y, et al. Cloning of alterations in the conformation of presenilin and
a novel gene bearing missense mutations in early interactions with amyloid precursor protein. J Neuro-
onset familial Alzheimer disease. Nature 1995;375: sci 2005;25:3009 – 17.
754 – 60. [23] Menendez M. Pathological and clinical heterogeneity
[9] Levy-Lahad E, Wasco W, Poorkaj P, et al. Candidate of presenilin 1 gene mutations. J Alzheimers Dis
gene for the chromosome 1 familial Alzheimer’s 2004;6:475 – 82.
disease locus. Science 1995;269:973 – 7. [24] Dermaut B, Kumar-Singh S, Engelborghs S, et al.
[10] Bayer TA, Wirths O, Majtenyi K, et al. Key factors A novel presenilin 1 mutation associated with Pick’s
in Alzheimer’s disease: b-amyloid precursor protein disease but not b-amyloid plaques. Ann Neurol 2004;
processing, metabolism and intraneuronal transport. 55:617 – 26.
Brain Pathol 2001;11:1 – 11. [25] Klunemann HH, Rogaeva E, Neumann M, et al.
746 imbimbo et al

Novel PS1 mutation in a Bavarian kindred with fa- on chromosome 10 in late-onset Alzheimer’s disease
milial Alzheimer disease. Alzheimer Dis Assoc Dis- pedigrees. Science 2000;290:2303 – 4.
ord 2004;18:256 – 8. [42] Myers A, Holmans P, Marshall H, et al. Susceptibil-
[26] Jia J, Xu E, Shao Y, et al. One novel presenilin-1 ity locus for Alzheimer’s disease on chromosome 10.
gene mutation in a Chinese pedigree of familial Alz- Science 2000;290:2304 – 5.
heimer’s disease. J Alzheimers Dis 2005;7:119 – 24. [43] Ertekin-Taner N, Ronald J, Asahara H, et al. Fine
[27] Kamboh MI. Molecular genetics of late-onset Alz- mapping of the a-T catenin gene to a quantitative trait
heimer’s disease. Ann Hum Genet 2004;68:381 – 404. locus on chromosome 10 in late-onset Alzheimer’s
[28] Bertram L, Tanzi RE. Alzheimer’s disease: one disease pedigrees. Hum Mol Genet 2003;12:3133 – 43.
disorder, too many genes? Hum Mol Genet 2004; [44] Vekrellis K, Ye Z, Qiu WQ, et al. Neurons regulate
13:R135 – 41. extracellular levels of amyloid b-protein via prote-
[29] Tanzi RE, Bertram L. Twenty years of the Alzhei- olysis by insulin-degrading enzyme. J Neurosci 2000;
mer’s disease amyloid hypothesis: a genetic perspec- 20:1657 – 65.
tive. Cell 2005;120:545 – 55. [45] Ertekin-Taner N, Allen M, Fadale D, et al. Genetic
[30] Mahley RW. Apolipoprotein E: cholesterol transport variants in a haplotype block spanning IDE are sig-
protein with expanding role in cell biology. Science nificantly associated with plasma Ab42 levels and risk
1988;240:622 – 30. for Alzheimer disease. Hum Mutat 2004;23:334 – 42.
[31] Gerdes LU, Klausen IC, Sihm I, et al. Apolipopro- [46] Ertekin-Taner N, Ronald J, Feuk L, et al. Elevated
tein E polymorphism in a Danish population com- amyloid b protein (Ab42) and late onset Alzheimer’s
pared to findings in 45 other study populations disease are associated with single nucleotide poly-
around the world. Genet Epidemiol 1992;9:155 – 67. morphisms in the urokinase-type plasminogen acti-
[32] Corder EH, Saunders AM, Risch NJ, et al. Protective vator gene. Hum Mol Genet 2005;14:447 – 60.
effect of apolipoprotein E type 2 allele for late onset [47] Scott WK, Grubber JM, Conneally PM, et al. Am.
Alzheimer disease. Nat Genet 1994;7:180 – 4. Fine mapping of the chromosome 12 late-onset Alz-
[33] Mayeux R, Stern Y, Ottman R, et al. The apolipo- heimer disease locus: potential genetic and pheno-
protein e4 allele in patients with Alzheimer’s disease. typic heterogeneity. J Hum Genet 2000;66:922 – 32.
Ann Neurol 1993;34:752 – 4. [48] Lambert JC, Goumidi L, Vrieze FW, et al. Hum. The
[34] St George-Hyslop P, McLachlan DC, Tsuda T, et al. transcriptional factor LBP-1c/CP2/LSF gene on chro-
Alzheimer’s disease and possible gene interaction. mosome 12 is a genetic determinant of Alzheimer’s
Science 1994;263:537. disease. Mol Genet 2000;9:2275 – 80.
[35] Selkoe DJ. Defining molecular targets to prevent [49] Qiu Z, Strickland DK, Hyman BT, Rebeck GW.
Alzheimer disease. Arch Neurol 2005;62:192 – 5. a2-Macroglobulin enhances the clearance of endoge-
[36] Farrer LA, Cupples LA, Haines JL, et al. Effects of nous soluble b-amyloid peptide via low-density lipo-
age, sex, and ethnicity on the association between protein receptor-related protein in cortical neurons.
apolipoprotein E genotype and Alzheimer disease. A J Neurochem 1999;73:1393 – 8.
meta-analysis. APOE and Alzheimer Disease Meta [50] Gibson AM, Singleton AB, Smith G, et al. Lack of
Analysis Consortium. JAMA 1997;278:1349 – 56. association of the a2-macroglobulin locus on chro-
[37] Tokuda T, Calero M, Matsubara E, et al. Lipidation of mosome 12 in AD. Neurology 2000;54:433 – 8.
apolipoprotein E influences its isoform-specific inter- [51] Zhang P, Yang Z, Wan CL, et al. Neither the tumor
action with Alzheimer’s amyloid beta peptides. necrosis factor alpha-308 A/G polymorphism nor the
Biochem J 2000;348(Pt. 2):359 – 65. a2-macroglobulin polymorphism was associated with
[38] Wisniewski T, Frangione B. Apolipoprotein E: late-onset Alzheimer’s disease in the Chinese popu-
a pathological chaperone protein in patients with lation. Yi Chuan Xue Bao 2004;31:1 – 6.
cerebral and systemic amyloid. Neurosci Lett 1992; [52] Deane R, Wu Z, Sagare A, et al. LRP/amyloid
135:235 – 8. b-peptide interaction mediates differential brain ef-
[39] Cedazo-Minguez A, Wiehager B, Winblad B, et al. flux of Ab isoforms. Neuron 2004;43:333 – 44.
Effects of apolipoprotein E (apoE) isoforms, [53] von Arnim CA, Kinoshita A, Peltan ID, et al. The low
b-amyloid (Ab) and apoE/Ab complexes on protein density lipoprotein receptor-related protein (LRP) is a
kinase C-a (PKC-a) translocation and amyloid novel b-secretase (BACE1) substrate. J Biol Chem
precursor protein (APP) processing in human 2005;280:17777 – 85.
SH-SY5Y neuroblastoma cells and fibroblasts. Neu- [54] Glaser C, Schulz S, Handschug K, et al. Genetic and
rochem Int 2001;38:615 – 25. functional characteristics of the human in vivo LRP1/
[40] Holtzman DM, Bales KR, Tenkova T, et al. Apoli- A2MR receptor suggested as a risk marker for
poprotein E isoform-dependent amyloid deposition Alzheimer’s disease and other complex (degenera-
and neuritic degeneration in a mouse model of Alz- tive) diseases. Neurosci Res 2004;50:85 – 101.
heimer’s disease. Proc Natl Acad Sci USA 2000; [55] Li Y, Nowotny P, Holmans P, et al. Association of
97:2892 – 7. late-onset Alzheimer’s disease with genetic variation
[41] Ertekin-Taner N, Graff-Radford N, Younkin LH, et al. in multiple members of the GAPD gene family. Proc
Linkage of plasma Ab42 to a quantitative locus Natl Acad Sci USA 2004;101:15688 – 93.
 pathophysiology of alzheimer’s disease 747

[56] Schulze H, Schuler A, Stuber D, et al. Rat brain [71] DeKosky ST, Ikonomovic MD, Wang X, et al. Plasma
glyceraldehyde-3-phosphate dehydrogenase interacts and cerebrospinal fluid a1-antichymotrypsin levels
with the recombinant cytoplasmic domain of Alz- in Alzheimer’s disease: correlation with cognitive im-
heimer’s b-amyloid precursor protein. J Neurochem pairment. Ann Neurol 2003;53:81 – 90.
1993;60:1915 – 22. [72] Kamboh MI, Sanghera DK, Ferrell RE, et al.
[57] Oyama R, Yamamoto H, Titani K. Glutamine syn- APOEE4-associated Alzheimer’s disease risk is modi-
thetase, hemoglobin a-chain, and macrophage migra- fied by a1-antichymotrypsin polymorphism. Nat
tion inhibitory factor binding to amyloid b-protein: Genet 1995;10:486 – 8.
their identification in rat brain by a novel affinity [73] Muller U, Bodeker RH, Gerund T, et al. Lack of
chromatography and in Alzheimer’s disease brain by association between a1-antichymotrypsin polymor-
immunoprecipitation. Biochim Biophys Acta 2000; phism, Alzheimer’s disease, and allele E4 of apolipo-
1479:91 – 102. protein E. Neurology 1996;47:1575 – 7.
[58] Heutink P. Untangling tau-related dementia. Hum [74] Licastro F, Chiappelli M, Grimaldi LM, et al. A new
Mol Genet 2000;9:979 – 86. promoter polymorphism in the a-1-antichymotrypsin
[59] Goedert M, Spillantini MG. Biochim. Tau mutations gene is a disease modifier of Alzheimer’s disease.
in frontotemporal dementia FTDP-17 and their rele- Neurobiol Aging 2005;26:449 – 53.
vance for Alzheimer’s disease. Biophys Acta 2000; [75] Jellinger KA. Head injury and dementia. Curr Opin
1502:110 – 21. Neurol 2004;17:719 – 23.
[60] Goedert M, Jakes R. Mutations causing neurodegen- [76] Smith DH, Chen XH, Nonaka M, et al. Accumulation
erative tauopathies. Biochim Biophys Acta 2005; of amyloid b and t and the formation of neurofila-
1739:240 – 50. ment inclusions following diffuse brain injury in the
[61] D’Souza I, Schellenberg GD. Regulation of tau iso- pig. J Neuropathol Exp Neurol 1999;58:982 – 92.
form expression and dementia. Biochim Biophys [77] Oehmichen M, Meissner C, Konig HG. Brain injury
Acta 2005;1739:104 – 15. after survived gunshot to the head: reactive alterations
[62] Oddo S, Caccamo A, Kitazawa M, et al. Amyloid at sites remote from the missile track. Forensic Sci Int
deposition precedes tangle formation in a triple 2001;115:189 – 97.
transgenic model of Alzheimer’s disease. Neurobiol [78] Van Den Heuvel C, Finnie JW, Blumbergs PC, et al.
Aging 2003;24:1063 – 70. Upregulation of neuronal amyloid precursor protein
[63] Perez M, Ribe E, Rubio A, et al. Characterization of (APP) and APP mRNA following magnesium sul-
a double (amyloid precursor protein-t) transgenic: phate (MgSO4) therapy in traumatic brain injury.
t phosphorylation and aggregation. Neuroscience J Neurotrauma 2000;17:1041 – 53.
2005;130:339 – 47. [79] Guo Z, Cupples LA, Kurz A, et al. Head injury and
[64] Carson JA, Turner AJ. b-Amyloid catabolism: roles the risk of AD in the MIRAGE study. Neurology
for neprilysin (NEP) and other metallopeptidases? 2000;54:1316 – 23.
J Neurochem 2002;81:1 – 8. [80] Mehta KM, Ott A, Kalmijn S, et al. Head trauma and
[65] Helisalmi S, Hiltunen M, Vepsalainen S, et al. risk of dementia and Alzheimer’s disease: The
Polymorphisms in neprilysin gene affect the risk of Rotterdam Study. Neurology 1999;53:1959 – 62.
Alzheimer’s disease in Finnish patients. J Neurol [81] Fleminger S, Oliver DL, Lovestone S, et al. Head
Neurosurg Psychiatry 2004;75:1746 – 8. injury as a risk factor for Alzheimer’s disease: the
[66] Shi J, Zhang S, Tang M, et al. Mutation screening and evidence 10 years on; a partial replication. J Neurol
association study of the neprilysin gene in sporadic Neurosurg Psychiatry 2003;74:857 – 62.
Alzheimer’s disease in chinese persons. J Gerontol A [82] Baum L, Chen L, Ng HK, et al. Apolipoprotein E
Biol Sci Med Sci 2005;60:301 – 6. isoforms in Alzheimer’s disease pathology and eti-
[67] Bertram L, Hiltunen M, Parkinson M, et al. Family- ology. Microsc Res Tech 2000;50:278 – 81.
based association between Alzheimer’s disease and [83] Notkola IL, Sulkava R, Pekkanen J, et al. Serum total
variants in UBQLN1. N Engl J Med 2005;352: cholesterol, apolipoprotein E e4 allele, and Alzhei-
884 – 94. mer’s disease. Neuroepidemiology 1998;17:14 – 20.
[68] Kan R, Wang B, Zhang C, et al. Genetic association [84] Bonarek M, Barberger-Gateau P, Letenneur L, et al.
of BACE1 gene polymorphism C786G with late- Relationships between cholesterol, apolipoprotein E
onset Alzheimer’s disease in Chinese. J Mol Neurosci polymorphism and dementia: a cross-sectional anal-
2005;25:127 – 32. ysis from the PAQUID study. Neuroepidemiology
[69] Floden AM, Li S, Combs CK. b-amyloid-stimulated 2000;19:141 – 8.
microglia induce neuron death via synergistic stimu- [85] Romas SN, Tang MX, Berglund L, et al. APOE ge-
lation of tumor necrosis factor a and NMDA recep- notype, plasma lipids, lipoproteins, and AD in com-
tors. J Neurosci 2005;25:2566 – 75. munity elderly. Neurology 1999;53:517 – 21.
[70] Fung HC, Chen CM, Wu YR, et al. Heat shock [86] Mielke MM, Zandi PP, Sjogren M, et al. High total
protein 70 and tumor necrosis factor-a in Taiwanese cholesterol levels in late life associated with a reduced
patients with dementia. Dement Geriatr Cogn Disord risk of dementia. Neurology 2005;64:1689 – 95.
2005;20:1 – 7. [87] Wolozin B, Kellman W, Ruosseau P, et al. Decreased
748 imbimbo et al

prevalence of Alzheimer disease associated with (OPTIMA). Total plasma homocysteine, age, systolic
3-hydroxy-3-methyglutaryl coenzyme A reductase in- blood pressure, and cognitive performance in older
hibitors. Arch Neurol 2000;57:1439 – 43. people. J Am Geriatr Soc 2002;50:2014 – 8.
[88] Li G, Higdon R, Kukull WA, et al. Statin therapy and [103] Sachdev PS, Valenzuela M, Wang XL, et al. Relation-
risk of dementia in the elderly: a community-based ship between plasma homocysteine levels and brain
prospective cohort study. Neurology 2004;63:1624– 8. atrophy in healthy elderly individuals. Neurology
[89] Bodovitz S, Klein WL. Cholesterol modulates alpha- 2002;58:1539 – 41.
secretase cleavage of amyloid precursor protein. [104] Kado DM, Karlamangla AS, Huang MH, et al.
J Biol Chem 1996;271:4436 – 40. Homocysteine versus the vitamins folate, B6,
[90] Simons M, Keller P, De Strooper B, et al. Cholesterol and B12 as predictors of cognitive function and
depletion inhibits the generation of b-amyloid in decline in older high-functioning adults: MacArthur
hippocampal neurons. Proc Natl Acad Sci USA Studies of Successful Aging. Am J Med 2005;118:
1998;95:6460 – 4. 161 – 7.
[91] Kojro E, Gimpl G, Lammich S, et al. Low cholesterol [105] Lewerin C, Matousek M, Steen G, et al. Significant
stimulates the non amyloidogenic pathway by its ef- correlations of plasma homocysteine and serum
fect on the a-secretase ADAM 10. Proc Natl Sci USA methylmalonic acid with movement and cognitive
2001;98:5815 – 20. performance in elderly subjects but no improvement
[92] Pedrini S, Carter TL, Prendergast G, et al. Modula- from short-term vitamin therapy: a placebo-controlled
tion of statin-activated shedding of Alzheimer APP randomized study. Am J Clin Nutr 2005;81:1155 – 62.
ectodomain by ROCK. PLoS Med 2005;2:69 – 78. [106] Mizrahi EH, Bowirrat A, Jacobsen DW, et al. Plasma
[93] Cole SL, Grudzien A, Manhart IO, et al. Statins cause homocysteine, vitamin B12 and folate in Alzheimer’s
intracellular accumulation of amyloid precursor pro- patients and healthy Arabs in Israel. J Neurol Sci
tein, b-secretase-cleaved fragments, and amyloid 2004;227:109 – 13.
b-peptide via an isoprenoid-dependent mechanism. [107] Luchsinger JA, Tang MX, Shea S, et al. Plasma
J Biol Chem 2005;280:18755 – 70. homocysteine levels and risk of Alzheimer disease.
[94] Urano Y, Hayashi I, Isoo N, et al. Association of ac- Neurology 2004;62:1972 – 6.
tive g-secretase complex with lipid rafts. J Lipid Res [108] Ariogul S, Cankurtaran M, Dagli N, et al. Vita-
2005;46:904 – 12. min B12, folate, homocysteine and dementia: are
[95] Sparks DL, Scheff SW, Hunsaker 3rd JC, et al. they really related? Arch Gerontol Geriatr 2005;40:
Induction of Alzheimer-like b-amyloid immuno- 139 – 46.
reactivity in the brains of rabbits with dietary cho- [109] Malouf M, Grimley EJ, et al. Folic acid with or
lesterol. Exp Neurol 1994;126:88 – 94. without vitamin B12 for cognition and dementia.
[96] Refolo LM, Malester B, LaFrancois J, et al. Hyper- Cochrane Database Syst Rev 2003;4:CD004514.
cholesterolemia accelerates the Alzheimer’s amyloid [110] Hasegawa T, Ukai W, Jo DG, et al. Homocysteic acid
pathology in a transgenic mouse model. Neurobiol induces intraneuronal accumulation of neurotoxic
Dis 2000;7:321 – 31. Ab42: implications for the pathogenesis of Alzhei-
[97] Refolo LM, Pappolla MA, LaFrancois J, et al. A mer’s disease. J Neurosci Res 2005;80:869 – 76.
cholesterol-lowering drug reduces b-amyloid pathol- [111] Ellinson M, Thomas J, Patterson A. A critical
ogy in a transgenic mouse model of Alzheimer’s evaluation of the relationship between serum vita-
disease. Neurobiol Dis 2001;8:890 – 9. min B, folate and total homocysteine with cognitive
[98] Heart Protection Study Collaborative Group. MRC/ impairment in the elderly. J Hum Nutr Diet 2004;17:
BHF Heart Protection Study of cholesterol lowering 371 – 83.
with simvastatin in 20,536 high-risk individuals: [112] Letenneur L. Risk of dementia and alcohol and wine
a randomised placebo-controlled trial. Lancet 2002; consumption: a review of recent results. Biol Res
360:7 – 22. 2004;37:189 – 93.
[99] Shepherd J, Blauw GJ, Murphy MB, et al. [113] Orgogozo JM, Dartigues JF, Lafont S, et al. Wine
PROspective Study of Pravastatin in the Elderly at consumption and dementia in the elderly: a prospec-
Risk. Pravastatin in elderly individuals at risk of tive community study in the Bordeaux area. Rev
vascular disease (PROSPER): a randomised con- Neurol 1997;153:185 – 92.
trolled trial. Lancet 2002;360:1623 – 30. [114] Ruitenberg A, van Swieten JC, Witteman JC, et al.
[100] Sparks DL, Sabbagh MN, Connor DJ, et al. Atorvas- Alcohol consumption and risk of dementia: the
tatin for the treatment of mild to moderate Alzheimer Rotterdam Study. Lancet 2002;359:281 – 6.
disease: preliminary results. Arch Neurol 2005;62: [115] Huang W, Qiu C, Winblad B, et al. Alcohol con-
753 – 7. sumption and incidence of dementia in a community
[101] McCaddon A, Hudson P, Davies G, et al. Homo- sample aged 75 years and older. J Clin Epidemiol
cysteine and cognitive decline in healthy elderly. 2002;5:959 – 64.
Dement Geriatr Cogn Disord 2001;12:309 – 13. [116] Luchsinger JA, Tang MX, Siddiqui M, et al. Alcohol
[102] Budge MM, de Jager C, Hogervorst E, et al, intake and risk of dementia. J Am Geriatr Soc
Oxford Project To Investigate Memory and Ageing 2004;52:540 – 6.
 pathophysiology of alzheimer’s disease 749

[117] Bachman DL, Green RC, Benke KS, et al, MIRAGE [133] Cai XD, Golde TE, Younkin SG. Release of excess
Study Group. Comparison of Alzheimer’s disease amyloid b protein from a mutant amyloid b protein
risk factors in white and African American families. precursor. Science 1993;259:514 – 6.
Neurology 2003;60:1372 – 4. [134] Suzuki N, Cheung TT, Cai XD, et al. 1994. An in-
[118] Petot GJ, Friedland RP. Lipids, diet and Alzheimer creased percentage of long amyloid b protein secreted
disease: an extended summary. J Neurol Sci 2004; by familial amyloid b protein precursor (bAPP717)
226:31 – 3. mutants. Science 1994;264:1336 – 40.
[119] Engelhart MJ, Geerlings MI, Ruitenberg A, et al. [135] Bales KR, Verina T, Dodel RC, et al. Lack of
Dietary intake of antioxidants and risk of Alzheimer apolipoprotein E dramatically reduces amyloid
disease. JAMA 2002;287:3223 – 9. b-peptide deposition. Nat Genet 1997;17:263 – 4.
[120] Morris MC, Evans DA, Bienias JL, et al. Dietary [136] Hardy J, Duff K, Hardy KG, et al. Genetic dissection
intake of antioxidant nutrients and the risk of incident of Alzheimer’s disease and related dementias: amy-
Alzheimer disease in a biracial community study. loid and its relationship to t. Nat Neurosci 1998;1:
JAMA 2002;287:3230 – 7. 355 – 8.
[121] Sano M, Ernesto C, Thomas RG, et al. A controlled [137] Lewis J, Dickson DW, Lin WL, et al. En-
trial of selegiline, a-tocopherol, or both as treat- hanced neurofibrillary degeneration in transgenic
ment for Alzheimer’s disease. The Alzheimer’s Dis- mice expressing mutant t and APP. Science 2001;
ease Cooperative Study. N Engl J Med 1997;336: 293:1487 – 91.
1216 – 22. [138] Bertram L, Blacker D, Mullin K, et al. Evidence for
[122] Petersen RC, Thomas RG, Grundman M, et al. genetic linkage of Alzheimer’s disease to chromo-
Vitamin E and donepezil for the treatment of mild some 10q. Science 2000;290:2302 – 3.
cognitive impairment. N Engl J Med 2005;352: [139] Oddo S, Caccamo A, Shepherd JD, et al. Triple-
2379 – 88. transgenic model of Alzheimer’s disease with plaques
[123] Berman K, Brodaty H. Tocopherol (vitamin E) in and tangles: intracellular Ab and synaptic dysfunc-
Alzheimer’s disease and other neurodegenerative tion. Neuron 2003;39:409 – 21.
disorders. CNS Drugs 2004;18:807 – 25. [140] Lippa CF, Nee LE, Mori H, et al. b-42 deposition
[124] Luchsinger JA, Mayeux R. Dietary factors and Alz- precedes other changes in PS-1 Alzheimer’s disease.
heimer’s disease. Lancet Neurol 2004;3:579 – 87. Lancet 1998;352:1117 – 8.
[125] Hardy J, Allsop D. Amyloid deposition as the central [141] Naslund J, Haroutunian V, et al. Correlation between
event in the aetiology of Alzheimer’s disease. Trends elevated levels of amyloid b-peptide in the brain and
Pharmacol Sci 1991;12:383 – 8. cognitive decline. JAMA 2000;283:1571 – 7.
[126] Hardy J, Selkoe DJ. The amyloid hypothesis of [142] Motter R, Vigo-Pelfrey C, Kholodenko D, et al.
Alzheimer’s disease: progress and problems on the Reduction of b-amyloid peptide42 in the cerebro-
road to therapeutics. Science 2002;297:353 – 6. spinal fluid of patients with Alzheimer’s disease.
[127] Glenner GG, Wong CW. Alzheimer’s disease: initial Ann Neurol 1995;38:643 – 8.
report of the purification and characterization of a [143] Mayeux R, Tang MX, Jacobs DM, et al. Plasma
novel cerebrovascular amyloid protein. Biochem amyloid b-peptide 1 – 42 and incipient Alzheimer’s
Biophys Res Commun 1984;120:885 – 990. disease. Ann Neurol 1999;46:412 – 6.
[128] Masters CL, Simms G, Weinman NA, et al. Amyloid [144] Assini A, Cammarata S, Vitali A, et al. Plasma levels
plaque core protein in Alzheimer disease and of amyloid b-protein 42 are increased in women
Down syndrome. Proc Natl Acad Sci USA 1985;82: with mild cognitive impairment. Neurology 2004;63:
4245 – 9. 828 – 31.
[129] Kang J, Lemaire HG, Unterbeck A, et al. The [145] Yang LB, Lindholm K, Yan R, et al. Elevated
precursor of Alzheimer’s disease amyloid A4 protein b-secretase expression and enzymatic activity de-
resembles a cell-surface receptor. Nature 1987;325: tected in sporadic Alzheimer disease. Nat Med 2003;
733 – 6. 9:3 – 4.
[130] Levy E, Carman MD, Fernandez-Madrid IJ, et al. [146] Li R, Lindholm K, Yang LB, et al. Amyloid b pep-
Mutation of the Alzheimer’s disease amyloid gene in tide load is correlated with increased b-secretase
hereditary cerebral hemorrhage. Dutch type. Science activity in sporadic Alzheimer’s disease patients. Proc
1990;248:1124 – 6. Natl Acad Sci USA 2004;101:3632 – 7.
[131] Wisniewski T, Ghiso J, et al. Peptides homologous to [147] Lambert MP, Barlow AK, Chromy BA, et al. Dif-
the amyloid protein of Alzheimer’s disease containing fusible, nonfibrillar ligands derived from Ab1 – 42
a glutamine for glutamic acid substitution have are potent central nervous system neurotoxins. Proc
accelerated amyloid fibril formation. Biochem Bio- Natl Acad Sci USA 1998;95:6448 – 53.
phys Res Commun 1991;179:1247 – 54. [148] Hartley DM, Walsh DM, Ye CP, et al. Protofibrillar
[132] Citron M, Oltersdorf T, Haass C, et al. Mutation of the intermediates of amyloid b-protein induce acute
b-amyloid precursor protein in familial Alzheimer’s electrophysiological changes and progressive neuro-
disease increases b-protein production. Nature 1992; toxicity in cortical neurons. J Neurosci 1999;19:
360:672 – 4. 8876 – 84.
750 imbimbo et al

[149] Kuo YM, Emmerling MR, Vigo-Pelfrey C, et al. [164] Qi Y, Morishima-Kawashima M, Sato T, et al. Dis-
Water-soluble Ab (N-40, N-42) oligomers in normal tinct mechanisms by mutant presenilin 1 and 2 lead-
and Alzheimer disease brains. J Biol Chem 1996;271: ing to increased intracellular levels of amyloid
4077 – 81. b-protein 42 in Chinese hamster ovary cells. Bio-
[150] Takahashi RH, Almeida CG, Kearney PF, et al. chemistry 2003;42:1042 – 52.
Oligomerization of Alzheimer’s b-amyloid within pro- [165] Wirths O, Multhaup G, Czech C, et al. Intra-
cesses and synapses of cultured neurons and brain. neuronal Ab accumulation precedes plaque formation
J Neurosci 2004;24:3592 – 9. in b-amyloid precursor protein and presenilin-1
[151] Kim HJ, Chae SC, Lee DK, Chromy B, et al. Selec- double-transgenic mice. Neurosci Lett 2001;306:
tive neuronal degeneration induced by soluble oligo- 116 – 20.
meric amyloid b protein. FASEB J 2003;17:118 – 20. [166] Blanchard V, Moussaoui S, Czech C, et al. Time
[152] Walsh DM, Klyubin I, Fadeeva JV, et al. Naturally sequence of maturation of dystrophic neurites asso-
secreted oligomers of amyloid b protein potently ciated with Ab deposits in APP/PS1 transgenic mice.
inhibit hippocampal long-term potentiation in vivo. Exp Neurol 2003;184:247 – 63.
Nature 2002;416:535 – 9. [167] Wirths O, Multhaup G, Czech C, et al. Intraneuro-
[153] Tsai J, Grutzendler J, Duff K, et al. Fibrillar amyloid nal APP/Ab trafficking and plaque formation in
deposition leads to local synaptic abnormalities and b-amyloid precursor protein and presenilin-1 trans-
breakage of neuronal branches. Nat Neurosci 2004;7: genic mice. Brain Pathol 2002;12:275 – 86.
1181 – 3. [168] Shie FS, LeBoeuf RC, Jin LW. Early intraneuronal
[154] Frears ER, Stephens DJ, Walters CE, et al. The role of Ab deposition in the hippocampus of APP transgenic
cholesterol in the biosynthesis of b-amyloid. Neuro- mice. Neuroreport 2003;14:123 – 9.
report 1999;10:1699 – 705. [169] Knauer MF, Soreghan B, Burdick D, et al. Intra-
[155] Howland DS, Trusko SP, Savage MJ, Reaume AG, cellular accumulation and resistance to degradation
Lang DM, Hirsch JD, et al. Modulation of secreted of the Alzheimer amyloid A4/b protein. Proc Natl
b-amyloid precursor protein and amyloid b-peptide Acad Sci USA 1992;89:7437 – 41.
in brain by cholesterol. J Biol Chem 1998;273: [170] Burdick D, Kosmoski J, Knauer MF, et al. Prefer-
16576 – 82. ential adsorption, internalization and resistance to
[156] Wirths O, Multhaup G, Bayer TA. A modified degradation of the major isoform of the Alzheimer’s
b-amyloid hypothesis: intraneuronal accumulation amyloid peptide, Ab1 – 42, in differentiated PC12
of the b-amyloid peptide—the first step of a fatal cells. Brain Res 1997;746:275 – 84.
cascade. J Neurochem 2004;91:513 – 20. [171] Chui DH, Dobo E, Makifuchi T, et al. Apoptotic neu-
[157] Wertkin AM, Turner RS, Pleasure SJ, et al. Human rons in Alzheimer’s disease frequently show intracellu-
neurons derived from a teratocarcinoma cell line ex- lar Ab42 labeling. J Alzheimers Dis 2001;3:231–9.
press solely the 695-amino acid amyloid precursor [172] LaFerla FM, Tinkle BT, Bieberich CJ, et al. The
protein and p roduce intracellular b-amyloid or A4 Alzheimer’s Ab peptide induces neurodegeneration
peptides. Proc Natl Acad Sci USA 1993;90:9513 – 7. and apoptotic cell death in transgenic mice. Nat Genet
[158] Hartmann T, Bieger SC, Bruhl B, et al. Distinct 1995;9:21 – 30.
sites of intracellular production for Alzheimer’s [173] Kienlen-Campard P, Miolet S, Tasiaux B, et al.
disease Ab40/42 amyloid peptides. Nat Med 1997; Intracellular amyloid-b1 – 42, but not extracellular
3:1016 – 20. soluble amyloid-b peptides, induces neuronal apo-
[159] Greenfield JP, Tsai J, Gouras GK, et al. Endoplasmic ptosis. J Biol Chem 2002;277:15666 – 70.
reticulum and trans-Golgi network generate distinct [174] Ohyagi Y, Yamada T, Nishioka K, et al. Selective
populations of Alzheimer b-amyloid peptides. Proc increase in cellular Ab42 is related to apoptosis
Natl Acad Sci USA 1999;96:742 – 7. but not necrosis. Neuroreport 2000;11:167 – 71.
[160] Walsh DM, Tseng BP, Rydel RE, et al. The [175] Misonou H, Morishima-Kawashima M, Ihara Y. Oxi-
oligomerization of amyloid b-protein begins intra- dative stress induces intracellular accumulation
cellularly in cells derived from human brain. Bio- of amyloid b-protein (Ab) in human neuroblastoma
chemistry 2000;39:10831 – 9. cells. Biochemistry 2000;39:6951 – 9.
[161] Gouras GK, Tsai J, Naslund J, et al. Intraneuronal [176] Esteban JA. Living with the enemy: a physiological
Ab42 accumulation in human brain. Am J Pathol role for the b-amyloid peptide. Trends Neurosci
2000;156:15 – 20. 2004;27:1 – 3.
[162] Fernandez-Vizarra P, Fernandez AP, Castro-Blanco S, [177] Seubert P, Vigo-Pelfrey C, Esch F, et al. Isolation
et al. Intra- and extracellular Ab and PHF in clini- and quantification of soluble Alzheimer’s b-peptide
cally evaluated cases of Alzheimer’s disease. Histol from biological fluids. Nature 1992;359:325 – 7.
Histopathol 2004;19:823 – 44. [178] Ramsden M, Henderson Z, Pearson HA. Modulation
[163] Takeda K, Araki W, Tabira T. Enhanced generation of Ca2 + channel currents in primary cultures of rat
of intracellular Ab42 amyloid peptide by mutation of cortical neurones by amyloid b protein (1 – 40) is
presenilins PS1 and PS2. Eur J Neurosci 2004;19: dependent on solubility status. Brain Res 2002;956:
258 – 64. 254 – 61.
 pathophysiology of alzheimer’s disease 751

[179] Kamenetz F, Tomita T, Hsieh H, et al. APP process- mediates amyloid-b peptide transport across the
ing and synaptic function. Neuron 2003;37:925 – 37. blood-brain barrier and accumulation in brain. Nat
[180] Plant LD, Boyle JP, Smith IF, et al. The production Med 2003;9:907 – 13.
of amyloid b peptide is a critical requirement for [197] Arancio O, Zhang HP, Chen X, et al. RAGE poten-
the viability of central neurons. J Neurosci 2003;23: tiates Ab-induced perturbation of neuronal function
5531 – 5. in transgenic mice. EMBO J 2004;23:4096 – 105.
[181] Cuello AC. Intracellular and extracellular Ab, a [198] Turner AJ, Fisk L, Nalivaeva NN. Targeting amyloid-
tale of two neuropathologies. Brain Pathol 2005;15: degrading enzymes as therapeutic strategies in neuro-
66 – 71. degeneration. Ann N Y Acad Sci 2004;1035:1 – 20.
[182] Hyman BT. The neuropathological diagnosis of Alz- [199] Authier F, Posner BI, Bergeron JJ. Insulin-degrading
heimer’s disease: clinical-pathological studies. Neu- enzyme. Clin Invest Med 1996;19:149 – 60.
robiol Aging 1997;18:S27 – 32. [200] Selkoe DJ. Clearing the brain’s amyloid cobwebs.
[183] Braak H, Braak E. Neuropathological stageing of Neuron 2001;32:177 – 80.
Alzheimer-related changes. Acta Neuropathol (Berl) [201] Farris W, Mansourian S, Chang Y, et al. Insulin-
1991;82:239 – 59. degrading enzyme regulates the levels of insulin,
[184] Lue LF, Kuo YM, Roher AE, et al. Soluble amyloid amyloid b-protein, and the b-amyloid precursor
beta peptide concentration as a predictor of synaptic protein intracellular domain in vivo. Proc Natl Acad
change in Alzheimer’s disease. Am J Pathol 1999; Sci USA 2003;100:4162 – 7.
155:853 – 62. [202] Leissring MA, Farris W, Chang AY, et al. Enhanced
[185] McLean CA, Cherny RA, Fraser FW, et al. Soluble proteolysis of b-amyloid in APP transgenic mice
pool of Ab amyloid as a determinant of severity of prevents plaque formation, secondary pathology, and
neurodegeneration in Alzheimer’s disease. Ann Neu- premature death. Neuron 2003;40:1087 – 93.
rol 1999;46:860 – 6. [203] Edland SD. Insulin-degrading enzyme, apolipopro-
[186] Wang J, Dickson DW, Trojanowski JQ, et al. The tein E, and Alzheimer’s disease. J Mol Neurosci
levels of soluble versus insoluble brain Ab distin- 2004;23:213 – 7.
guish Alzheimer’s disease from normal and patho- [204] Turner AJ, Isaac RE, Coates D. The neprilysin (NEP)
logic aging. Exp Neurol 1999;158:328 – 37. family of zinc metalloendopeptidases: genomics and
[187] Houlden H, Baker M, McGowan E, et al. Variant function. Bioessays 2001;23:261 – 9.
Alzheimer’s disease with spastic paraparesis and cot- [205] Hama E, Shirotani K, Masumoto H, et al. Clearance
ton wool plaques is caused by PS-1 mutations that of extracellular and cell-associated amyloid b peptide
lead to exceptionally high amyloid-beta concentra- through viral expression of neprilysin in primary
tions. Ann Neurol 2000;48:806 – 8. neurons. J Biochem (Tokyo) 2001;30:721 – 6.
[188] Tanzi RE, Moir RD, Wagner SL. Clearance of [206] Iwata N, Tsubuki S, Takaki Y, et al. Metabolic regu-
Alzheimer’s Abeta peptide: the many roads to perdi- lation of brain Ab by neprilysin. Science 2001;292:
tion. Neuron 2004;43:605 – 8. 1550 – 2.
[189] Zlokovic BV. Clearing amyloid through the blood- [207] Mohajeri MH, Kuehnle K, et al. Anti-amyloid activ-
brain barrier. J Neurochem 2004;89:807 – 11. ity of neprilysin in plaque-bearing mouse models of
[190] Shibata M, Yamada S, Kumar SR, et al. Clearance Alzheimer’s disease. FEBS Lett 2004;562:16 – 21.
of Alzheimer’s amyloid-b1 – 40 peptide from brain by [208] Marr RA, Rockenstein E, Mukherjee A, et al. Neprily-
LDL receptor-related protein-1 at the blood-brain sin gene transfer reduces human amyloid pathology
barrier. J Clin Invest 2000;106:1489 – 99. in transgenic mice. J Neurosci 2003;23:1992 – 6.
[191] Herz 2003. [209] Wang DS, Iwata N, Hama E, et al. Oxidized
[192] Deane R, Wu Z, Zlokovic BV. RAGE (yin) versus neprilysin in aging and Alzheimer’s disease brains.
LRP (yang) balance regulates Alzheimer amyloid Biochem Biophys Res Commun 2003;310:236 – 41.
b-peptide clearance through transport across the [210] Eckman EA, Reed DK, Eckman CB. Degradation of
blood-brain barrier. Stroke 2004;35(11, Suppl 1): the Alzheimer’s amyloid b peptide by endothelin-
2628 – 31. converting enzyme. J Biol Chem 2001;276:24540 – 8.
[193] Van Uden E, Mallory M, Veinbergs I, et al. Increased [211] Eckman EA, Watson M, Marlow L, et al. Alzheimer’s
extracellular amyloid deposition and neurodegenera- disease b-amyloid peptide is increased in mice
tion in human amyloid precursor protein trans- deficient in endothelin-converting enzyme. J Biol
genic mice deficient in receptor-associated protein. Chem 2003;278:2081 – 4.
J Neurosci 2002;22:9298 – 304. [212] Funalot B, Ouimet T, Claperon A, et al. Endothelin-
[194] Yan SD, Zhu H, Zhu A, et al. Receptor-dependent cell converting enzyme-1 is expressed in human cerebral
stress and amyloid accumulation in systemic amyloid- cortex and protects against Alzheimer’s disease. Mol
osis. Nat Med 2000;6:643 – 51. Psychiatry 2004;9:1122 – 8.
[195] Yan SD, Stern D, Kane MD, et al. RAGE-Ab [213] Nunomura A, Perry G, Aliev G, et al. Oxidative
interactions in the pathophysiology of Alzheimer’s damage is the earliest event in Alzheimer disease.
disease. Restor Neurol Neurosci 1998;12:167 – 73. J Neuropathol Exp Neurol 2001;60:759 – 67.
[196] Deane R, Du Yan S, Submamaryan RK, et al. RAGE [214] Butterfield DA, Drake J, Pocernich C, et al. Evidence
752 imbimbo et al

of oxidative damage in Alzheimer’s disease brain: [232] Parker Jr WD, Parks J, Filley CM, et al. Electron
central role for amyloid b-peptide. Trends Mol Med transport chain defects in Alzheimer’s disease brain.
2001;7:548 – 54. Neurology 1994;44:1090 – 6.
[215] Mattson MP. Pathways towards and away from [233] Blass JP. Brain metabolism and brain disease: is
Alzheimer’s disease. Nature 2004;430:631 – 9. metabolic deficiency the proximate cause of Alz-
[216] Parihar MS, Hemnani T. Alzheimer’s disease patho- heimer dementia? J Neurosci Res 2001;66:851 – 6.
genesis and therapeutic interventions. J Clin Neurosci [234] Dodart JC, Mathis C, Bales KR, et al. Early re-
2004;11:456 – 67. gional cerebral glucose hypometabolism in trans-
[217] Smith MA, Perry G, Richey PL, et al. Oxidative genic mice overexpressing the V717F b-amyloid
damage in Alzheimer’s. Nature 1996;382:120 – 1. precursor protein. Neurosci Lett 1999;277:49 – 52.
[218] Luth HJ, Holzer M, Gartner U, et al. Expression of [235] Buchner M, Huber R, Sturchler-Pierrat C, et al.
endothelial and inducible NOS-isoforms is increased Impaired hypoxic tolerance and altered protein bind-
in Alzheimer’s disease, in APP23 transgenic mice ing of NADH in presymptomatic APP23 transgenic
and after experimental brain lesion in rat: evidence for mice. Neuroscience 2002;114:285 – 9.
an induction by amyloid pathology. Brain Res 2001; [236] Watson GS, Craft S. The role of insulin resistance
913:57 – 67. in the pathogenesis of Alzheimer’s disease: implica-
[219] de la Torre JC, Stefano GB. Evidence that Alzhei- tions for treatment. CNS Drugs 2003;17:27 – 45.
mer’s disease is a microvascular disorder: the role of [237] Patel NV, Gordon MN, Connor KE, et al. Caloric re-
constitutive nitric oxide. Brain Res Brain Res Rev striction attenuates Abeta-deposition in Alzheimer trans-
2000;34:119 – 36. genic models. Neurobiol Aging 2005;26:995 – 1000.
[220] Jang SH, Surh YJ. AP-1 mediates b-amyloid-induced: [238] de la Torre JC. Vascular basis of Alzheimer’s patho-
iNOS expression in PC12 cells via the ERK2 and p38 genesis. Ann N Y Acad Sci 2002;977:196 – 215.
MAPK signaling pathways. Biochem Biophys Res [239] Shi J, Yang SH, Stubley L, et al. Hypoperfusion
Commun 2005;331:1421 – 8. induces overexpression of beta-amyloid precursor
[221] Streit WJ. Microglia and Alzheimer’s disease patho- protein mRNA in a focal ischemic rodent model.
genesis. J Neurosci Res 2004;77:1 – 8. Brain Res 2000;853:1 – 4.
[222] Cacquevel M, Lebeurrier N, et al. Cytokines in neu- [240] Kalaria RN, Bhatti SU, Palatinsky EA, et al. Accu-
roinflammation and Alzheimer’s disease. Curr Drug mulation of the b amyloid precursor protein at sites
Targets 2004;5:529 – 34. of ischemic injury in rat brain. Neuroreport 1993;4:
[223] Xia MQ, Hyman BT. Chemokines/chemokine recep- 211 – 4.
tors in the central nervous system and Alzheimer’s [241] Suo Z, Su G, Placzek A, et al. b vasoactivity in vivo.
disease. J Neurovirol 1999;5:32 – 41. Ann N Y Acad Sci 2000;903:156 – 63.
[224] Casal C, Serratosa J, et al. Effects of beta-AP peptides [242] Mueggler T, Baumann D, Rausch M, et al. Age-
on activation of the transcription factor NF-kappaB dependent impairment of somatosensory response in
and in cell proliferation in glial cell cultures. Neurosci the amyloid precursor protein 23 transgenic mouse
Res 2004;48:315 – 23. model of Alzheimer’s disease. J Neurosci 2003;23:
[225] Johnstone M, Gearing AJ, et al. A central role for as- 8231 – 6.
trocytes in the inflammatory response to beta-amyloid; [243] Stokin GB, Lillo C, Falzone TL, et al. Axonopathy
chemokines, cytokines and reactive oxygen species are and transport deficits early in the pathogenesis of
produced. J Neuroimmunol 1999;93:182 – 93. Alzheimer’s disease. Science 2005;307:1282 – 8.
[226] Rogers J, Cooper NR, Webster S, et al. Complement [244] Sheng JG, Price DL, Koliatsos VE. The b-amyloid-
activation by b-amyloid in Alzheimer disease. Proc related proteins presenilin 1 and BACE1 are axonally
Natl Acad Sci USA 1992;89:10016 – 20. transported to nerve terminals in the brain. Exp Neu-
[227] Mattson MP, Chan SL. Neuronal and glial calcium rol 2003;184:1053 – 7.
signaling in Alzheimer’s disease. Cell Calcium 2003; [246] Popescu BO, Ankarcrona M. Mechanisms of cell
34:385 – 97. death in Alzheimer’s disease: role of presenilins.
[228] Brzyska M, Elbaum D. Dysregulation of calcium in J Alzheimers Dis 2004;6:123 – 8.
Alzheimer’s disease. Acta Neurobiol Exp (Warsz) [245] Buxbaum JD, Thinakaran G, Koliatsos V, et al. Alz-
2003;63:171 – 83. heimer amyloid protein precursor in the rat hip-
[229] LaFerla FM. Calcium dyshomeostasis and intracel- pocampus: transport and processing through the
lular signalling in Alzheimer’s disease. Nat Rev perforant path. J Neurosci 1998;18:9629 – 37.
Neurosci 2002;3:862 – 72. [247] McPhie DL, Coopersmith R, Hines-Peralta A, et al.
[230] Yang Y, Cook DG. Presenilin-1 deficiency impairs DNA synthesis and neuronal apoptosis caused by
glutamate-evoked intracellular calcium responses in familial Alzheimer disease mutants of the amyloid
neurons. Neuroscience 2004;124:501 – 15. precursor protein are mediated by the p21 activated
[231] Eckert A, Schindowski K, Leutner S, et al. Alzhei- kinase PAK3. J Neurosci 2003;23:6914 – 27.
mer’s disease-like alterations in peripheral cells from [248] Dickson DW. Apoptotic mechanisms in Alzheimer
presenilin-1 transgenic mice. Neurobiol Dis 2001;8: neurofibrillary degeneration: cause or effect? J Clin
331 – 42. Invest 2004;114:23 – 7.
 pathophysiology of alzheimer’s disease 753

[249] Copani A, Sortino MA, Caricasole A, et al. Erratic synthase kinase-3 beta-mediated tau phosphorylation
expression of DNA polymerases by b-amyloid in cultured cell lines. Neuroreport 2003;14:257 – 60.
causes neuronal death. FASEB J 2002;16:2006 – 8. [256] Leroy K, Boutajangout A, Authelet M, et al.
[250] Eckert A, Keil U, Marques CA, et al. Mitochondrial The active form of glycogen synthase kinase-3beta
dysfunction, apoptotic cell death, and Alzheimer’s is associated with granulovacuolar degeneration in
disease. Biochem Pharmacol 2003;66:1627 – 34. neurons in Alzheimer’s disease. Acta Neuropathol
[251] Paradis E, Douillard H, Koutroumanis M, et al. (Berl) 2002;103:91 – 9.
Amyloid beta peptide of Alzheimer’s disease down- [257] Bennecib M, Gong CX, Grundke-Iqbal I, et al. Role
regulates Bcl-2 and upregulates bax expression in of protein phosphatase-2A and -1 in the regulation
human neurons. J Neurosci 1996;16:7533 – 9. of GSK-3, cdk5 and cdc2 and the phosphorylation
[252] Michaelis ML, Dobrowsky RT, Li G. Tau neuro- of tau in rat forebrain. FEBS Lett 2000;485:87 – 93.
fibrillary pathology and microtubule stability. J Mol [258] Sontag E, Hladik C, Montgomery L, et al. Down-
Neurosci 2002;19:289 – 93. regulation of protein phosphatase 2A carboxyl meth-
[253] Iqbal K, Alonso A, Gong C, et al. Molecular pa- ylation and methyltransferase may contribute to
thology of Alzheimer neurofibrillary degeneration. Alzheimer disease pathogenesis. J Neuropathol Exp
Acta Neurobiol Exp (Warsz) 1993;53:325 – 35. Neurol 2004;63:1080 – 91.
[254] Terwel D, Dewachter I, Van Leuven F. Axonal trans- [259] Arriagada PV, Growdon JH, Hedley-Whyte ET, et al.
port, tau protein, and neurodegeneration in Alzhei- Neurofibrillary tangles but not senile plaques parallel
mer’s disease. Neuromolecular Med 2002;2:151 – 65. duration and severity of Alzheimer’s disease. Neu-
[255] Lee CW, Lau KF, Miller CC, Shaw PC. Glycogen rology 1992;42:631 – 9.

View publication stats