Journal of Pathology

J Pathol 2011; 224: 147152

Published online in Wiley Online Library
(wileyonlinelibrary.com) DOI: 10.1002/path.2885

INVITED PERSPECTIVE

Compounding artefacts with uncertainty, and an amyloid cascade

hypothesis that is too big to fail
Rudy J Castellani1 * and Mark A Smith2#
1
2

Department of Pathology, University of Maryland, Baltimore, MD, USA

Department of Pathology, Case Western Reserve University, Cleveland, OH, USA

*Correspondence to: Rudy J Castellani, 22 South Greene Street, Baltimore, MD 21201, USA. e-mail: rcastellani@som.umaryland.edu
# Mark

Smith was tragically killed in a road accident in December 2010. He had been awarded the Goudie Medal by the Pathological Society
and was due to give the award lecture in Cambridge, UK in January 2011. This review article had been commissioned to accompany that
lecture and was close to completion prior to Dr Smiths death.

Abstract
With each failure of anti-amyloid- therapy in clinical trials, new trials are initiated with no hint of slowing
down. This may be due, in part, to the fact that the amyloid cascade hypothesis has been so modified over time
that it is now impossible to confirm or deny. The hypothesis now states, in effect, that invisible molecules target
invisible structures. Still relevant, however, are multiple factors that surely cast some doubt but have either been
rationalized or overlooked. Among these are the poor correlation between amyloid- deposits and disease, the
substantial differences between familial and sporadic disease, pathological assessment that indicates the secondary
nature of lesions/proteins/cascades, the fact that soluble species are poorly reproducible laboratory phenomena,
and the irrelevance of synaptic assessment to pathological interpretation. Although not yet dogma, the premature
addition of mild cognitive impairment as the implied in vivo homologue to the soluble toxin-synapse interaction
is also problematic. In either case, the amyloid cascade hypothesis continues to dominate the Alzheimers disease
literature and grant applications. The more the neuroscience community perseverates along these lines in the face
of accumulating outcome data to the contrary, the more one is left to wonder whether the hypothesis is too big
to fail.
Copyright 2011 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.

Keywords: Alzheimers disease; amyloid-; mild cognitive impairment; oligomers

Received 3 January 2011; Revised 3 January 2011; Accepted 26 February 2011

No conflicts of interest were declared.

Introduction
For better or worse, the amyloid cascade hypothesis
continues to be a conclusion in search of data [1,2]. As
the hypothesis and its various permutations continue
to dominate the literature, the extent to which the
scientific community clings to the righteousness of the
idea with the passage of time and confrontation of
contradictory data is striking. In a different context,
such a degree of faith would be the envy of any
religion.
But there is an additional problem: with dozens of
clinical trials targeting amyloid- either under way or
having failed, and with no signs of slowing down
[3], a legitimate concern is that the hypothesis has
become too big to fail. With so much time, money
and, indeed, faith invested in the construct, is a
negative outcome simply intolerable for the scientific
community and society who depends on it? Moreover,
does expansion of trials run the risk of obtaining a
kernel of positive data [3], purely out of randomness
and the expanding denominator, thus perpetuating a
Copyright 2011 Pathological Society of Great Britain and Ireland.
Published by John Wiley & Sons, Ltd. www.pathsoc.org.uk

fundamentally flawed paradigm and diverting attention

from biological processes more worthy of targeting?
In this review, we highlight data and concepts that
are either overlooked or subject to rationalization.
With repeated failures of anti-amyloid therapy and the
often reported outcome as worse than placebo, we
hope that we might cure ourselves, as it were, of
the underlying disease of a priori bias, and return to
objective scientific analysis.

Amyloid proteins were derived from pathological

lesions in vivo and in situ
Glenner and Wong in 1984 reported the purification of
a peptide derived from vascular amyloid deposits in
cerebrovascular amyloid of Alzheimers disease (AD)
brains by Sephadex G-100 column chromatography
and high-performance liquid chromatography [4]. They
found a similar peptide in the cerebrovascular amyloid
in Downs syndrome brains, thus initiating the search
for a chromosome 21 precursor species [5]. Masters
J Pathol 2011; 224: 147152
www.thejournalofpathology.com

148

et al found proteinaceous material in amyloid plaque

cores, indicating that poorly soluble pathological hallmark lesions, both vascular and parenchymal, in the
AD and Downs syndrome brain, shared a common
precursor and common pathogenesis [6]. The conclusion by Glenner and Wong at the time was appropriate: This protein may be derived from a unique
serum precursor which may provide a diagnostic test
for Alzheimers disease and a means to understand its
pathogenesis.
It is worth noting that the peptide was derived
exclusively from insoluble pathological lesions
congophilic angiopathy and amyloid plaque cores
rather than soluble fractions. This was strategic, given
the widely held belief, although unproven, that the
insoluble lesions themselves were of aetiological significance. Data not surprisingly accumulated indicating
the toxic nature of fibrillar amyloid- (A), by numerous mechanisms encompassing oxidative stress, inflammatory mediators, induction of apoptosis and so forth
[79]. The most compelling data were derived from
synthetic A species and their effects on cultured neurons; nevertheless, the relevance was ostensibly clear
and the concept embedded in the literature was that
insoluble fibrillar A was toxic. Further details accumulated regarding specific fragments, particularly A42
[10] (more specifically the A42 : A40 ratio, as recent
studies have shown increased ratio due to reduced
A40 , rather than increased A42 [1113]), as the
pathogenic species, as it was overrepresented in neuritic plaque pathology, the more relevant lesion in
diagnosing AD at autopsy [14].
Consistently ignored, however, was and is the poor
correlation with A of whatever species in the human
brain and clinical disease, and the diffuse, non-specific
neocortical deposition of A, irrespective of functional
circuits [1520]. Indeed, amyloid deposits are less
abundant in the medial temporal allocortical areas that
subserve declarative memory [21], where neural function is typically impaired at the onset of clinical disease. Such an obvious flaw might have prompted a
wholesale re-examination of the aetiological significance of A. Instead, the non-specific nature of A
deposition in the elderly brain continued to be widely
viewed as an upstream event, leading to more proximate, but still secondary, accumulation of phosphorylated tau and cell death [22].
Ignored in turn was the well-known regional distribution of phosphorylated tau that correlates well
with disease severity and functional deficits according to the neuroanatomical region, or, in other words,
ignored was the disconnect, both in terms of regional
distribution and temporal deposition, between A and
phosphorylated tau [23]. Indeed, the only certainties
leading up to the elaboration of low n-mer A species
were that A lesions were present in diseased brains,
but that these correlated poorly in terms of presence
or absence of dementia, cognitive domain and disease
progression. Moreover, the very lesions which generated the amyloid cascade hypothesis, and which were
Copyright 2011 Pathological Society of Great Britain and Ireland.
Published by John Wiley & Sons, Ltd. www.pathsoc.org.uk

RJ Castellani and MA Smith

once regarded as hallmark lesions and insoluble, toxic

material by any of several molecular mechanisms, are
now marked for discard as insoluble, non-toxic epiphenomena in the pellet of the ultracentrifuge tube [24].

The problem with familial disease

The extent to which neuroscientists cling to the amyloid cascade can be placed squarely at the feet of
familial disease and Downs syndrome. The existence
of rare, autosomal dominant mendelian conditions that
mimic AD, albeit imperfectly, and the early onset of
AD-like changes in chromosome 21 over-expressing
Downs syndrome, is tempting evidence in favour of
aetiology. Germline mutations in and around the A
coding regions, as well as within the notch signalling
domain of the putative -secretase complex, and APP
overexpression in Downs syndrome, indicate unambiguously that such abnormalities leads to A deposition, phosphorylated tau pathology and early-onset
clinical disease. In addition to the linear geneprotein
concept per central dogma of molecular biology, there
is evidence of enzymesubstrate involvement [1].
Here again, however, a reductionist small molecule
construct tends to ignore several empirical facts. First,
autosomal dominant, early-onset AD, is exceedingly
rare. Less than 100 families with APP mutations
have been described worldwide, and only several hundred families with presenilin mutations have been
described (http://www.molgen.ua.ac.be/ADMutations/
default.cfm). The commonly cited 5% figure for familial AD cases is likely an overestimate for genuine
familial disease. When one considers that roughly
30 million patients have AD worldwide, several hundred families would comprise less than a fraction of
1% of AD cases.
Second, the clinical phenotype varies considerably among familial cases. In addition to the early
age at onset, some patients present with congophilic
angiopathy-related haemorrhage, others present with
long tract signs, such as spastic paraparesis, still others
present with seizures [25,26]. The pathological phenotype also variesmany have a prodigious amyloid
accumulation, morphologically unusual depositions (eg
cotton wool plaques), extensive white matter, cerebellar and basal ganglia deposition, -synuclein pathology, etc.
Third, the issue of singular cascades and germline
mutations is, or should be, trivialized by the human
proteome, comprised of some 25 000 protein transcripts, and nearly two million protein species of
diverse function [27]. The idea that, in sporadic disease, such diversity within the neuron is governed
by physiological alterations in a single small peptide
seems extremely remote. Even in the setting of familial disease, pathogenic mutations that declare metabolic
derangements very quickly in pulse-labelling experiments have no discernible clinical or pathological
J Pathol 2011; 224: 147152
www.thejournalofpathology.com

Compounding artefacts with uncertainty, and an amyloid cascade hypothesis

consequences in vivo for decades. It remains an open

question, therefore, whether familial disease is truly
representative of the sporadic condition, and an even
greater question whether experimental models with one
to several mutations based on familial disease are an
appropriate model to study the relevance of treatment
for the sporadic condition. The outcome of clinical trials seems to indicate, at the absolute minimum, that
sporadic and familial diseases are sufficiently different
to warrant new approaches.

The problem with pathology

Autopsy examination as the gold standard for diagnosis is a misnomer [1]. It has never been shown, for
example, that one can examine the brains of 100 centenarians who were cognitively intact during life, and
100 centenarians who had AD dementia during life,
and separate them with any degree of accuracy, based
on neuropathological examination, whereas the converse has been shown [28]. Such would be the true
test of the aetiological nature of pathological lesions
and, by extension, the proteins derived from them.
If lesions were truly aetiological, the identification of
lesions would diagnose disease. Instead, neuroscientists rely on crude and imprecise pathological criteria,
such as CERAD [14], or criteria derived irrespective
of the presence or degree of dementia, such as Braak
[15]. Moreover, in the case of CERAD, and its precursor Khachaturian, criteria [29], increased pathology is
required with increased age. Taken to its logical conclusion, one of two possibilities exists that are mutually
exclusive: either older patients tolerate pathology better than younger ones, or lesions and their constituent
proteins are secondary phenomena. If one agrees that
the elderly brain does not tolerate toxic phenomena
better than the middle-aged brain, then the toxicity
of lesions becomes an impossibility and the adaptive
or productive nature of the lesions becomes infinitely
more likely.
In the case of Braak stageing, the likelihood that
someone was demented during life increases with
stage, but does not reach 100% even for Braak stage
VI. In a recent analysis by Nelson et al [30], it was
shown that within the Braak stage VI group, minimental status data varied according to the count of
neurofibrillary tangles, such that high tangle counts
in the setting of Braak stage VI were synonymous
with dementia. In essence, these data indicate that the
diagnosis of dementia is only certain with end-stage
pathology, a further indication that the pathology, the
proteins that comprise the pathology and the cascades
that lead to protein accumulation are late or end-stage
events in the context of clinical disease. Interestingly,
this runs counter to non-neoplastic pathology in all
other organ systems, where diagnosis becomes less
specific and less clear with advanced disease (e.g.
end-stage kidney disease, interstitial lung disease, liver
Copyright 2011 Pathological Society of Great Britain and Ireland.
Published by John Wiley & Sons, Ltd. www.pathsoc.org.uk

149

disease, neuromuscular disease), which is again a clear

indication of the secondary nature of pathological
lesions in AD.

Laboratory artefacts and invisible substrates

The A oligomer or toxic intermediate concept is
the most substantial revision of the amyloid cascade hypothesis since its inception. In brief, the concept states that soluble A species, specifically nonmonomeric species up to dodecamers, are early, toxic
phenomena [31]. This is significant for two reasons.
First, it renders moot the issue of poor correlation
between A lesions and disease. Second, it provides
a bridge to discussion of the synapse.
In both instances, issues continue to be overlooked.
A oligomers are first and foremost in vitro elaborations, identified in the soluble fraction after high-speed
centrifugation and impossible to study in vivo. Indeed,
the most compelling studies involve injection of conditioned medium, derived from oligomer-secreting APP
V717F Chinese hamster ovary cells, into rat lateral ventricle, followed by outcome measure assessmenta
nakedly artificial environment in comparison with sporadic AD [32,33]. Also unaddressed is the diffuse
nature of A deposits once fibrillized, poor correlation
with brain regions involved in early symptomatology,
and exactly how oligomers may be so geographically
eloquent as to produce preclinical disease (e.g. related
to medial temporal lobe dysfunction) and then lose that
eloquence once fibrillized.
The addition of synapse studies to in vitro oligomer
biology only compounds the uncertainty. The common
refrains in the literature that synaptic damage (more
recently synaptic dysfunction) is the true pathological
substrate for AD or the best correlate for cognitive
impairment, are non-statements. As the synapse provides for neurotransmission, it can be claimed that any
non-neoplastic disease of the nervous system correlates with synaptic dysfunction. It comes as no surprise,
therefore, that the synapse has been shown to be deficient or damaged in a condition such as AD, where
there is macroscopic tissue loss and/or overt clinical
signs. Omitted from the discussion, however, is that
synapse assessments, either by electron microscopy or
immunohistochemistry, are error-prone investigational
exercises that have no diagnostic relevance, no specificity to AD, and otherwise do not provide an independent method of disease characterization in human
brains. As problematic as the diagnostic criteria are,
as indicated above, adding synaptic damage to the
pathological assessment, either by light microscopy or
ultrastructural examination, is a non-starter, as it would
add a hopeless layer of confusion to the interpretation. In the end, the best we can conclude is that A
oligomers, or in vitro laboratory artefacts, affect the
synapse by indirect outcome measures in highly experimental and likely irrelevant constructs. The assessment
J Pathol 2011; 224: 147152
www.thejournalofpathology.com

150

of the synapse, in turn, is investigational by definition,

and further implicated in virtually all nervous system
diseases.

MCI: a fictitious intermediate stage to parallel

fictitious intermediate toxins
In recent years, mild cognitive impairment (MCI) has
become the most widely used clinical model to describe
the boundary between normal ageing and early AD.
While the existence of degrees of cognitive impairment with uncertain outcome predates the introduction of the MCI terminology, Petersen and co-workers
raised the discussion to a new level by suggesting that
patients with mild cognitive impairment can be clinically defined [34]. Later studies described MCI as an
intermediate stage of cognitive decline between normal and pathological brain ageing [35], while others
indicate that MCI is a heterogeneous clinical entity
[36]. Additional MCI subtypes, such as amnestic MCI,
multidomain MCI, MCI plus, etc. [37], complicate
the issue and suggest still further nosological entities
whose underlying pathology and clinical outcome are
variable.
It is perhaps worth noting that recognition of subtle levels of cognitive dysfunction dates as far back as
the seventh century BC [38], although in truth what
we now know as MCI has most likely been recognized since the dawn of the human species. The Global
Deterioration Scale (GDS), for instance, which comprised the means of assessment in the original introduction of MCI into modern literature in 1988 [39],
assesses such qualities as orientation, work performance, word finding, reading retention, remembering
names, and tendencies to misplace valuable objects.
Similarly, the Clinical Dementia Rating (CDR) scale
[40], which has found some utility in the MCI determination, assesses memory, orientation, judgement, problem solving, community affairs, home and hobbies, and
personal care. Formal tests of cognitive dysfunction,
including neuropsychological testing, add to the pool
of data, and each is individually debated and disputed
[41]. The ultimate determinant, however, is clinical
judgement and experience [42], which, while in keeping with many syndromic diagnoses, is prima facie
evidence of the uncertainty inherent in the putative
entity.
Much of the confusion over MCI may reside in a lack
of understanding of what constitutes genuine neurodegenerative disease, and the premature identification of
a condition or entity that in reality does not exist. The
introduction of a term, and indeed an acronym, tends
to create the impression of an entity, where a simple
non-specific phrase, such as possible preclinical disease, may be more appropriate and less misleading.
MCI is a concept which, while legitimate as a concept,
falls far short of a clinicopathological entity and therefore a target for diagnosis and therapy. The instability
Copyright 2011 Pathological Society of Great Britain and Ireland.
Published by John Wiley & Sons, Ltd. www.pathsoc.org.uk

RJ Castellani and MA Smith

of outcome [4350], lack of genetic predispositions

[51] and absence of treatment implications [50,5254],
aside from a marginal benefit with treatment of depression [55], are all evidence of the fundamental nature
of MCI, which is diagnostic uncertainty.
Nevertheless, the putative entity dovetails remarkably well with the synaptic-A oligomer concept, as
though the elaboration of the two in parallel were preordained [56]. As the oligomer concept continues, perhaps with more emphasis on the intracellular compartment, an increasing association with MCI is a virtual
certainty. On the one hand, there is a clinical syndrome with subtle cognitive complaints and no clear
pathological substrate; on the other is a putative soluble
lesion that might produce subtle cognitive complaints
in the absence of clear pathological lesions. The fact
that both MCI and synaptic attack of soluble A are
highly conceptual, one being an acronym applied to
a subjective assessment with unstable outcome, and
the other being a laboratory artefact applied to immeasurable ultrastructure, is once again overlooked in the
desire to advance the frontiers of knowledge.

Conclusion
In the excitement over the prospect of small molecule
intervention, the data now indicate undue optimism. In
retrospect this was predictable, yet prospectively the
temptation of a narrowly circumscribed cascade as the
explanation for AD is still evident. The neuroscience
community nevertheless might confront the factors that
have been overlooked and/or rationalized to date, so
that other avenues may be explored with equal rigor. A
partial list of such factors include: the poor correlation
between A deposits and disease; the spatial and temporal disconnect between A and phosphorylated tau
deposition; the substantial differences between familial
and sporadic disease; the irrelevance of experimental
models that contain one to several germline mutations;
the empirical evidence that autopsy diagnosis indicates the secondary nature of pathological lesions, their
derivative proteins and cascades; the fact that soluble
species are poorly reproducible laboratory phenomena;
the inability to study synaptic structure in disease; and
the irrevelevance of the synapse to pathological diagnosis. Further problematic is the momentum that exists
to apply toxic intermediates to the MCI concept, given
the fact that MCI is a diagnosis of uncertainty with
unstable outcome, and is subjective by definition.
The amyloid cascade hypothesis has been invaluable
in elucidating the physiology of A metabolism, but it
has failed in equal measure to produce any tangible
treatment benefit. In point of fact, outcome data to
date tend to be worse than control. More importantly,
the failure was predictable and had been predicted on
many levels. Continued undaunted efforts at this point
are an indication more of the marketing of the idea and
the substantial investment in its ultimate success than
J Pathol 2011; 224: 147152
www.thejournalofpathology.com

Compounding artefacts with uncertainty, and an amyloid cascade hypothesis

a confrontation of the reality. Only with a paradigm

shift and investment in alternative promising constructs
will we avoid the unsavory outcome that a flawed,
fundamentally misinterpreted set of data was too big
to fail.

Author contributions
Drs Smith and Castellani contributed equally to the
ideas contained in the manuscript and to the writing of
the manuscript.

References
1. Castellani RJ, Lee HG, Zhu X, et al . Alzheimer disease pathology
as a host response. J Neuropathol Exp Neurol 2008; 67: 523531.
2. Castellani RJ, Lee HG, Siedlak SL, et al . Reexamining
Alzheimers disease: evidence for a protective role for amyloid-
protein precursor and amyloid-. J Alzheimers Dis 2009; 18:
447452.
3. Lemere CA, Masliah E. Can Alzheimer disease be prevented by
amyloid- immunotherapy? Nat Rev Neurol ; 6: 108119.
4. Glenner GG, Wong CW. Alzheimers disease: initial report of the
purification and characterization of a novel cerebrovascular amyloid protein. Biochem Biophys Res Commun 1984; 120: 885890.
5. Glenner GG, Wong CW. Alzheimers disease and Downs syndrome: sharing of a unique cerebrovascular amyloid fibril protein.
Biochem Biophys Res Commun 1984; 122: 11311135.
6. Masters CL, Simms G, Weinman NA, et al . Amyloid plaque core
protein in Alzheimer disease and Down syndrome. Proc Natl Acad
Sci USA 1985; 82: 42454249.
7. Forloni G, Chiesa R, Smiroldo S, et al . Apoptosis-mediated neurotoxicity induced by chronic application of -amyloid fragment
2535. NeuroReport 1993; 4: 523526.
8. Eikelenboom P, Veerhuis R, Scheper W, et al . The significance of
neuroinflammation in understanding Alzheimers disease. J Neural
Transm 2006; 113: 16851695.
9. Mark RJ, Blanc EM, Mattson MP. Amyloid -peptide and oxidative cellular injury in Alzheimers disease. Mol Neurobiol 1996;
12: 211224.
10. Suzuki N, Cheung TT, Cai XD, et al . An increased percentage
of long amyloid- protein secreted by familial amyloid- protein
precursor (-APP717) mutants. Science 1994; 264: 13361340.
11. Hellstrom-Lindahl E, Viitanen M, Marutle A. Comparison of A
levels in the brain of familial and sporadic Alzheimers disease.
Neurochem Int 2009; 55: 243252.
12. Bentahir M, Nyabi O, Verhamme J, et al . Presenilin clinical mutations can affect -secretase activity by different mechanisms.
J Neurochem 2006; 96: 732742.
13. Kumar-Singh S, Theuns J, Van Broeck B, et al . Mean age-ofonset of familial Alzheimer disease caused by presenilin mutations
correlates with both increased A42 and decreased A40. Hum
Mutat 2006; 27: 686695.
14. Mirra SS, Heyman A, McKeel D, et al . The Consortium to Establish a Registry for Alzheimers Disease (CERAD). Part II. Standardization of the neuropathologic assessment of Alzheimers disease. Neurology 1991; 41: 479486.
15. Braak H, Braak E. Neuropathological stageing of Alzheimerrelated changes. Acta Neuropathol 1991; 82: 239259.
16. Arriagada PV, Growdon JH, Hedley-Whyte ET, et al . Neurofibrillary tangles but not senile plaques parallel duration and severity of
Alzheimers disease. Neurology 1992; 42: 631639.
Copyright 2011 Pathological Society of Great Britain and Ireland.
Published by John Wiley & Sons, Ltd. www.pathsoc.org.uk

151

17. Giannakopoulos P, Herrmann FR, Bussiere T, et al . Tangle and

neuron numbers, but not amyloid load, predict cognitive status in
Alzheimers disease. Neurology 2003; 60: 14951500.
18. Bierer LM, Hof PR, Purohit DP, et al . Neocortical neurofibrillary
tangles correlate with dementia severity in Alzheimers disease.
Arch Neurol 1995; 52: 8188.
19. Dickson DW, Crystal HA, Mattiace LA, et al . Identification of
normal and pathological aging in prospectively studied nondemented elderly humans. Neurobiol Aging 1992; 13: 179189.
20. Berg L, McKeel DW, Jr., Miller JP, et al . Clinicopathologic studies in cognitively healthy aging and Alzheimers disease: relation of
histologic markers to dementia severity, age, sex, and apolipoprotein E genotype. Arch Neurol 1998; 55: 326335.
21. Arnold SE, Hyman BT, Flory J, et al . The topographical and
neuroanatomical distribution of neurofibrillary tangles and neuritic
plaques in the cerebral cortex of patients with Alzheimers disease.
Cereb Cortex 1991; 1: 103116.
22. Hardy J, Duff K, Hardy KG, et al . Genetic dissection of
Alzheimers disease and related dementias: amyloid and its relationship to tau. Nat Neurosci 1998; 1: 355358.
23. Schonheit B, Zarski R, Ohm TG. Spatial and temporal relationships between plaques and tangles in Alzheimer pathology. Neurobiol Aging 2004; 25: 697711.
24. Walsh DM, Selkoe DJ. A oligomersa decade of discovery.
J Neurochem 2007; 101: 11721184.
25. Jimenez-Escrig A, Rabano A, Guerrero C, et al . New V272A
presenilin 1 mutation with very early onset subcortical dementia
and parkinsonism. Eur J Neurol 2004; 11: 663669.
26. Shrimpton AE, Schelper RL, Linke RP, et al . A presenilin 1
mutation (L420R) in a family with early onset Alzheimer disease,
seizures and cotton wool plaques, but not spastic paraparesis.
Neuropathology 2007; 27: 228232.
27. Jensen ON. Modification-specific proteomics: characterization of
post-translational modifications by mass spectrometry. Curr Opin
Chem Biol 2004; 8: 3341.
28. Brayne C, Richardson K, Matthews FE, et al . Neuropathological
correlates of dementia in over-80 year-old brain donors from the
population-based Cambridge city over-75s cohort (CC75C) study.
J Alzheimer Dis 2009; 18: 645658.
29. Khachaturian ZS. Diagnosis of Alzheimers disease. Arch Neurol
1985; 42: 10971105.
30. Nelson PT, Braak H, Markesbery WR. Neuropathology and cognitive impairment in Alzheimer disease: a complex but coherent
relationship. J Neuropathol Exp Neurol 2009; 68: 114.
31. Tanzi RE. The synaptic A hypothesis of Alzheimer disease. Nat
Neurosci 2005; 8: 977979.
32. Walsh DM, Klyubin I, Fadeeva JV, et al . Naturally secreted
oligomers of amyloid- protein potently inhibit hippocampal longterm potentiation in vivo. Nature 2002; 416: 535539.
33. Shankar GM, Bloodgood BL, Townsend M, et al . Natural
oligomers of the Alzheimer amyloid- protein induce reversible
synapse loss by modulating an NMDA-type glutamate receptordependent signaling pathway. J Neurosci 2007; 27: 28662875.
34. Petersen RC, Smith GE, Waring SC, et al . Mild cognitive impairment: clinical characterization and outcome. Arch Neurol 1999; 56:
303308.
35. Matthews FE, Stephan BC, Bond J, et al . Operationalization of
mild cognitive impairment: a graphical approach. PLoS Med 2007;
4: 16151619.
36. Portet F, Ousset PJ, Visser PJ, et al . Mild cognitive impairment
(MCI) in medical practice: a critical review of the concept and
new diagnostic procedure. Report of the MCI Working Group of the
European Consortium on Alzheimers Disease. J Neurol Neurosurg
Psychiat 2006; 77: 714718.
J Pathol 2011; 224: 147152
www.thejournalofpathology.com

152

37. Winblad B, Palmer K, Kivipelto M, et al . Mild cognitive impairmentbeyond controversies, towards a consensus: report of the
International Working Group on Mild Cognitive Impairment.
J Intern Med 2004; 256: 240246.
38. Berchtold NC, Cotman CW. Evolution in the conceptualization of
dementia and Alzheimers disease: Greco-Roman period to the
1960s. Neurobiol Aging 1998; 19: 173189.
39. Reisberg B, Ferris SH. Brief Cognitive Rating Scale (BCRS).
Psychopharmacol Bull 1988; 24: 629636.
40. Morris JC. The Clinical Dementia Rating (CDR): current version
and scoring rules. Neurology 1993; 43: 24122414.
41. Huntley JD, Howard RJ. Working memory in early Alzheimers
disease: a neuropsychological review. Int J Geriatr Psychiat 25:
121132.
42. Petersen RC. Mild cognitive impairment as a diagnostic entity.
J Intern Med 2004; 256: 183194.
43. Stephan BC, Brayne C, McKeith IG, et al . Mild cognitive impairment in the older population: who is missed and does it matter?
Int J Geriatr Psychiat 2008; 23: 863871.
44. Busse A, Hensel A, Guhne U, et al . Mild cognitive impairment:
long-term course of four clinical subtypes. Neurology 2006; 67:
21762185.
45. Fischer P, Jungwirth S, Zehetmayer S, et al . Conversion from
subtypes of mild cognitive impairment to Alzheimer dementia.
Neurology 2007; 68: 288291.
46. Farias ST, Mungas D, Reed BR, et al . Progression of mild cognitive impairment to dementia in clinic vs community-based cohorts.
Arch Neurol 2009; 66: 11511157.
47. Solfrizzi V, Panza F, Colacicco AM, et al . Vascular risk factors,
incidence of MCI, and rates of progression to dementia. Neurology
2004; 63: 18821891.

Copyright 2011 Pathological Society of Great Britain and Ireland.

Published by John Wiley & Sons, Ltd. www.pathsoc.org.uk

RJ Castellani and MA Smith

48. Tschanz JT, Welsh-Bohmer KA, Lyketsos CG, et al . Conversion

to dementia from mild cognitive disorder: the Cache County Study.
Neurology 2006; 67: 229234.
49. Ravaglia G, Forti P, Montesi F, et al . Mild cognitive impairment:
epidemiology and dementia risk in an elderly Italian population.
J Am Geriatr Soc 2008; 56: 5158.
50. Petersen RC, Thomas RG, Grundman M, et al . Vitamin E and
donepezil for the treatment of mild cognitive impairment. N Engl
J Med 2005; 352: 23792388.
51. Heun R, Guhne U, Luck T, et al . Apolipoprotein E allele 4 is not
a sufficient or a necessary predictor of the development of mild
cognitive impairment. Eur Psychiat 2010; 25: 1518.
52. Thal LJ, Ferris SH, Kirby L, et al . A randomized, double-blind,
study of rofecoxib in patients with mild cognitive impairment.
Neuropsychopharmacology 2005; 30: 12041215.
53. Feldman HH, Ferris S, Winblad B, et al . Effect of rivastigmine
on delay to diagnosis of Alzheimers disease from mild cognitive
impairment: the InDDEx study. Lancet Neurol 2007; 6: 501512.
54. Winblad B, Gauthier S, Scinto L, et al . Safety and efficacy of
galantamine in subjects with mild cognitive impairment. Neurology
2008; 70: 20242035.
55. Doody RS, Ferris SH, Salloway S, et al . Donepezil treatment of
patients with MCI: a 48-week, randomized, placebo-controlled
trial. Neurology 2009; 72: 15551561.
56. Englund H, Degerman Gunnarsson M, Brundin RM, et al .
Oligomerization partially explains the lowering of A42 in
Alzheimers disease cerebrospinal fluid. Neurodegen Dis 2009; 6:
139147.

J Pathol 2011; 224: 147152

www.thejournalofpathology.com