1. No category

PDF hosted at the Radboud Repository of the Radboud University
Nijmegen
This full text is a publisher's version.
For additional information about this publication click this link.
http://hdl.handle.net/2066/92727
Please be advised that this information was generated on 2014-10-15 and may be subject to
change.
The ref lection of
Alzheimer disease
in CSF
Petra Elise Spies
Part of the research presented in this thesis was performed within the framework of
the Center for Translational Molecular Medicine (www.ctmm.com).
Financial support for printing of this thesis was kindly provided by Zambon Nederland
B.V., Nutricia Advanced Medical Nutrition, Danone Research – Centre for Specialised
Nutrition, Novartis Pharma B.V., Lundbeck B.V., Luminex B.V., Janssen B.V., Internationale
Stichting Alzheimer Onderzoek, Crossbeta Biosciences (Utrecht) – Developing
Therapeutics for Alzheimer’s Disease, Alzheimer Nederland (Bunnik).
ISBN 978-94-91027-23-9
Design cover and lay out by In Zicht Grafisch Ontwerp, Arnhem
Printed and bound by PrintPartners Ipskamp, Enschede
© 2012 P.E. Spies
All rights reserved. No parts of this publication may be reproduced or transmitted in any form or by any
means, electronic or mechanical, including photocopy, recording or otherwise, without prior written
permission of the author.
The reflection of
Alzheimer disease in CSF
Een wetenschappelijke proeve op het gebied van de
Medische Wetenschappen
Proefschrift
ter verkrijging van de graad van doctor
aan de Radboud Universiteit Nijmegen
op gezag van de rector magnificus prof. mr. S.C.J.J. Kortmann,
volgens besluit van het college van decanen
in het openbaar te verdedigen op donderdag 15 maart 2012
om 10.30 uur precies
door
Petra Elise Spies
geboren op 14 januari 1983
te Tiel
Promotor
Prof. dr. M.G.M. Olde Rikkert
Copromotoren
Dr. ir. M.M. Verbeek
Dr. J.A.H.R. Claassen
Manuscriptcommissie Prof. dr. P. Wesseling
Prof. dr. G.J.M. Martens
Prof. dr. P.M.M. Bossuyt (Academic Medical Center - University of Amsterdam)
The reflection of
Alzheimer disease in CSF
An academic essay in Medical Sciences
Doctoral thesis
to obtain the degree of doctor
from Radboud University Nijmegen
on the authority of the Rector Magnificus prof. dr. S.C.J.J. Kortmann,
according to the decision of the Council of Deans
to be defended in public on Thursday March 15, 2012
at 10.30 hours
by
Petra Elise Spies
born on January 14, 1983
in Tiel
Supervisor
Prof. dr. M.G.M. Olde Rikkert
Co-supervisors
Dr. ir. M.M. Verbeek
Dr. J.A.H.R. Claassen
Doctoral Thesis Committee Prof. dr. P. Wesseling
Prof. dr. G.J.M. Martens
Prof. dr. P.M.M. Bossuyt (Academic Medical Center - University of Amsterdam)
Contents
Prologue
Chapter OneIntroduction and Outline
9
15
Part One | CSF biomarkers in the differential diagnosis of dementia
Chapter TwoThe CSF Aβ 42/Aβ 40 ratio
39
Chapter ThreeCSF α-synuclein
55
Chapter Four
The correlation between CSF α-synuclein and CSF Aβ
69
Chapter Five
Experiences with CSF analysis in memory clinics
77
Chapter SixA prediction model to calculate the probability of AD
91
Part Two | CSF biomarkers in pathophysiological hypotheses
Chapter SevenDynamic hypotheses
113
Chapter Eight
The correlation between CSF biomarkers
and cognitive functioning
119
Chapter Nine
AD and cerebrovascular disease
131
Chapter Ten
Why is CSF Aβ 42 decreased in AD?
145
Chapter Eleven
Summary and general discussion
165
Chapter Twelve Nederlandse samenvatting
185
Dankwoord
195
Curriculum Vitae
202
List of publications 203
Donders Graduate School for Cognitive Neuroscience Series
204
Prologue | On elderly chaps and ferocious Orcs
This chapter has been published as: Elderly: a term to avoid or to embrace?
Petra E. Spies, Jurgen A.H.R. Claassen, J Am Geriatr Soc 2011;59:1986-7
Prologue
Prologue
Prologue
On elderly chaps and ferocious Orcs
With the surge in articles that report research on aging, the question has risen whether
the commonly used ‘elderly’ is an appropriate term to describe old people.1,2 A strong
plea was even launched to ban this term from medical journals, and replace it by ‘older
persons’.3 Three main arguments were put forward: 1) old people reject being called
‘elderly’; 2) ‘elderly’ may be confused with ‘frail’; 3) the term ‘elderly’ is pejorative, and its
use thus reflects negatively on older patients as well as on health care for the old.2,3
We question the validity of these arguments.
Old people reject being called ‘elderly’
This argument was based on a European survey on preferential terminology for older
citizens, conducted in 12 countries.4 The multi-linguistic setting was a surprising choice
for an investigation of semantic preferences. Noteworthy, only preference, and not
rejection, was measured. It remains unclear how linguistic and cultural differences were
taken into account. Were participants addressed in their own language? Were the
intuitive subtleties of the English language correctly brought across? The Danish for
example quite liked being called ‘elderly’ – maybe their translation carries with it a
different meaning than the English ‘elderly’. In the English-speaking countries, ‘elderly’
was indeed not a preferred term, but remarkably, neither was ‘older people’. ‘Senior
(citizen)’ received most votes. In contrast, the American media guide issued by The
International Longevity Center advised against the use of senior. Apparently, semantic
preferences cannot easily be generalized.
‘Elderly’ is easily confused with ‘frail’
We find this hard to believe. For example, in the 2001 cinema hit ‘Lord of the Rings’, the
character Gandalf was described as an ‘elderly chap’. Gandalf, who in this film
subsequently slays a dragon-like creature as well as several ferocious Orcs, can hardly be
considered frail. If the large audience can distinguish elderly from frail, so can and
should people who work in the field of aging research. And they do: even though the
concept of frailty remains ill-defined and disputed, old age in itself is not considered a
criterion by most researchers. Underlining this is the fact that those who work in the
field do not shy from the term. When a number of medical journals were scanned for
the use of ‘elderly people’ and ‘older people’, it was found that most authors and editors,
even in geriatrics, had continued to use ‘elderly’ in spite of the proposed ban.1
11
Prologue
One may argue that the term ‘elderly’ has insufficient categorical validity. Obviously,
the inherent heterogeneity in the older age group is tremendous, making the use of a
single term as a group denominator an easy target for criticism. However, we argue that
‘elderly’ is as unambiguous a term as ‘child’ or ‘adult’, terms that also cover a wide
spectrum and that must be further specified by descriptives such as ‘prepubertal’ or
‘healthy’.
‘Elderly’ is a pejorative term
Is it the term elderly that is responsible for the negative image of older people and of
the health services and professionals devoted to their care? Will frenetic avoidance of
this term remove this stigma? Naturally, it can be defended that people do not want to
be seen as old. Most adults, if told that they look much younger than their date of birth
suggests, glow with happiness. When geriatric patients are asked how they get along at
day care, they complain that it is full of old people. The unpleasantness of being
confronted with one’s advanced age may transpire to how any term describing old age
is perceived, including ‘elderly’. But does that make these terms inappropriate? Calling
an 80-year-old ‘elderly’, ‘senior’ or ‘in later life’ will not change his actual age. Even if one
preferred word were found, it would be a matter of time before this new term would
carry the stigma of the old one. It would be better to change the stigma and not the
word. We should remember that elderly is not the same as frail, that old is not the same
as worthless. This message is far more important than the discussion whether elderly or
older people feel insulted by either term, or whether health care for the elderly will rise
in status if it changes its name to health care for later life. Clinicians and researchers
could therefore embrace ‘elderly’ as the preferred term to describe the old people they
care for and investigate. Stemming from the Old English ‘aeld’ it means ‘somewhat
older’, a clear and gentle way to describe an old person. Let’s stop debating semantics,
and start changing a stigma.
12
On elderly chaps and ferocious Orcs
References
1. Quinlan N, O’Neill D. “Older” or “elderly”--are medical journals sensitive to the wishes of older people? J
Am Geriatr Soc 2008;56:1983-4.
2. Francis RM. Editor’s view. Age Ageing 2010;39:220.
3. Falconer M, O’Neill D. Out with “the old,” elderly, and aged. BMJ 2007;334:316.
4. Walker A. Age and attitudes: main results from a Eurobarometer survey. Brussels: Commission of the
European Communities, 1993.
13
Introduction and Outline
The case report in this chapter has been published as: Cerebrospinal Fluid Biomarkers in Diagnosing
Alzheimer’s Disease in Clinical Practice: An Illustration with 3 Case Reports. Diane Slats, Petra E. Spies,
Magnus J.C. Sjögren, Frans R. Verhey, Marcel M. Verbeek, Marcel G.M. Olde Rikkert, Case Rep Neurol. 2010;2:5-11
Chapter
One
Chapter One
Chapter One
Introduction and Outline
Dementia
Dementia is a cognitive disorder that, by definition, interferes with activities of daily
living.1,2 Different types of dementia exist. Alzheimer disease (AD) is the most prevalent
and therefore the best known cause of dementia. This disease will be described in more
detail below. Both vascular dementia (VaD) and dementia with Lewy bodies (DLB) are
reported to be the second most prevalent types of dementia.3,4 Patients with vascular
dementia present with impairments in multiple cognitive domains, slowness in executive
functioning, bradyphrenia, and sometimes focal neurological deficits. Neuro­imaging
shows vascular lesions such as cerebral infarcts or extensive white matter disease.5 DLB
patients have fluctuating functioning and attention, together with visuospatial dysfunction
on neuropsychological examination. They often suffer from visual hallucinations and
parkinsonism. Neuropathologically, this disorder is characterized by intraneuronal
inclusions composed of α-synuclein, the so-called Lewy bodies.6 A less common cause
of dementia is frontotemporal dementia (FTD). Features are early decline in social
interpersonal conduct, emotional blunting, loss of insight, behavioural disorders and
impairments in speech and language.7 The neuropathological substrate of FTD is
heterogeneous. Most cases display cellular inclusions of tau or TAR DNA-binding
protein-43 (TDP-43), but in some cases, the pathologic protein is unknown.7-11
Alzheimer disease
AD is a progressive disorder that generally affects memory first, but other cognitive
domains such as language, perception and praxis are usually affected as well.2,12 Aloïs
Alzheimer was the first to describe the amyloid β (Aβ) plaques and neurofibrillary
tangles that are seen upon neuropathological examination.13 In addition, severe loss of
neurons is found, and often at least a mild degree of cerebral amyloid angiopathy (CAA):
Aβ deposits in the walls of blood vessels. The Aβ content of amyloid plaques and of
these vascular deposits was discovered in the 1980s, the same decade that tau protein
and its hyperphosphorylated form were identified in neurofibrillary tangles.14-18
Aβ is a small protein derived from sequential cleaving of the Aβ precursor protein
(APP) by β- and γ-secretase. Several isoforms of Aβ have been identified, but Aβ42 and
Aβ 40 have received most attention. Aβ 40, 40 amino acids long, is the most abundant Aβ
peptide, and makes up about 60% of the cerebrospinal fluid (CSF) Aβ concentration.
Aβ 42 accounts for approximately 10% of the CSF Aβ concentration in healthy people.19-21
17
Chapter ONE
Two types of Aβ plaques are found in AD patients: diffuse plaques without an amyloid
core that contain only Aβ 42, and classic plaques that have an amyloid core and consist
of both Aβ 42 and Aβ 40.22,23 In contrast, the vascular deposits in the walls of leptomeningeal
and cortical arteries – and less frequently, in capillaries24 – contain mostly Aβ 40.25
Tau is a protein that stabilizes the microtubules of the neuronal cytoskeleton. In AD,
abnormal phosphorylation of tau leads to destabilizing of microtubules and is suggested
to obstruct axonal transport, resulting in synaptic dysfunction. 26 Hyperphosphorylated
tau accumulates intracellularly and forms neurofibrillary tangles. After deterioration of
the neuron, the tangles remain visible as extraneuronal ghost tangles.26,27
Diagnosing dementia
Several criteria sets have been developed that summarize the core features of AD, VaD,
FTD and DLB.2,5-7,12 The Dutch dementia guideline28 states that a diagnosis of dementia
can be made clinically using these criteria. History taking with the patient and a proxy
should focus on gaining the information that is confined in the criteria sets. Neuropsychological assessment can aid in determining whether or not a cognitive disorder is
present that may indicate dementia. When more diagnostic certainty is desired, for
example to differentiate between the different types of dementia, additional
examinations are advised such as structural magnetic resonance imaging (MRI) of the
brain and CSF analysis.
One may wonder whether differentiating between the dementia types (in other
words, to establish an etiological diagnosis) really adds important information that
extends beyond the knowledge that a dementia is present. As there are no cures
available for any of these causes of dementia, what is the point of knowing the
underlying cause? There are several reasons that make it important to clarify the cause
of dementia. First, AD, VaD, DLB and FTD differ importantly in the types of deficits they
induce in cognition and behaviour. Therefore, an adequate etiological diagnosis allows
appropriate and specific education, guidance and care for patients and their informal
caregivers.29,30 Second, knowledge on the type of dementia can provide information on
a patient’s prognosis. For example, DLB patients deteriorate faster than AD patients,
with shorter time between onset of cognitive symptoms and death,31,32 and in FTD
patients, behavioural problems are far more dominant than in AD patients. Third,
accurately diagnosing the cause of dementia is important for pharmacological
management. Cholinesterase inhibitors may not cure the disease, but they can offer
18
Introduction and Outline
functional improvement in AD. DLB patients also respond well to cholinesterase
inhibitors, but these patients are extremely sensitive to the side effects of neuroleptic
drugs.4,33 Knowing the cause of the dementia therefore helps in deciding which drugs
may be beneficial, and which ones should be avoided. A fourth reason to clarify the
cause of the dementia is to advance research in this area. For example, an accurate
diagnosis is essential to select patients for trials. If a drug is aimed at treating AD, but the
patient group is contaminated by patients with other types of dementia, the drug under
study may show no overall effect, despite an effect in patients with true AD. Recognizing
AD in the very early stages of the disease may help in selecting patients for trials aimed
at prevention of the disease.
Underlined by the ongoing efforts to redefine them,2,34,35 the criteria sets that were
laid out so many years ago are insufficient to reliably differentiate between dementia
types. A comparison of the clinical diagnoses made with these criteria to the neuropathological diagnoses made post-mortem showed that their sensitivity or their
specificity is low, depending on the criteria set used and the comparison made.36-38
Especially mixed pathologies are often not recognized clinically.36,37,39
One strategy to improve the accuracy of clinical diagnosis is by introducing
biomarkers in addition to the clinical criteria. A biomarker, as defined by the National
Institute of Health, is ‘a characteristic that is objectively measured and evaluated as an
indicator of normal biologic processes, pathogenic processes, or pharmacologic
responses to a therapeutic intervention’.40 The Dutch dementia guideline already
suggested the use of biomarkers such as atrophy on MRI or decreased CSF Aβ 42 and
increased CSF total tau (t-tau) to gain more certainty about the cause of dementia. This
reasoning was explicated in the research criteria for AD as published in 2007, in which
positive biomarkers are considered a supportive feature of Alzheimer disease, regardless
of whether or not dementia is clinically present.41,42
Biomarkers for AD
The ideal biomarker for AD should have the following properties: it should be able to
detect a fundamental feature of AD’s neuropathology, it should be validated against
neuropathologically confirmed AD cases, it should be precise, reliable, non-invasive,
simple to perform and inexpensive.43 AD biomarkers discovered thus far can be divided
into biomarkers of Aβ plaque deposition (CSF Aβ 42, Aβ imaging with positron emission
tomography (PET)) and biomarkers of neurodegeneration (structural MRI of the
19
Chapter ONE
hippocampus, CSF t-tau and hyperphosphorylated tau (p-tau), fluorodeoxyglucose
(FDG)-PET).44
Hippocampal atrophy, seen on computed tomography (CT), was the first AD
biomarker to be established.45 The involvement of the hippocampus in AD had been
recognized a few years earlier, when severe neuronal loss in this region in AD cases was
described in a neuropathology study.46 A visual rating scale was developed for use with
CT or MRI that is still in use today.47 Hippocampal volume on MRI correlates well with AD
pathology at neuropathological examination, although hippocampal atrophy can also
be found with increasing age and in other diseases such as FTD and hippocampal
sclerosis.48,49
FGD-PET gives a measure of synaptic activity based on glucose metabolism in the
brain. In AD, bilateral temporo-parietal hypometabolism is observed that cannot only
be explained by brain atrophy.50 A positive scan correlates with the presence of AD
pathology on post-mortem examination.51,52
CSF biomarkers are derived from the fluid that circulates around the brain and
spinal cord and thus can provide information from close to the brain. Measuring Aβ 42,
t-tau and p-tau in the CSF succeeded a few years after their discovery53-57 and it was
shown that CSF Aβ 42 is decreased and CSF t-tau and p-tau are increased in AD.56,58-62
These three biomarkers provide excellent discrimination between AD patients and
healthy controls.63,64 The specificity for discrimination with other dementias is lower,65,66
but the combination of biomarkers may still help in the discrimination between AD and
VaD,67 DLB68 and FTD.69
As with all biochemical analyses, the concentrations of CSF biomarkers vary
between different laboratories.70 This variation makes it difficult to compare the results
between laboratories and makes it impossible to provide generally applicable reference
values for use in clinical practice. The variation is probably a result of variations in
analytical procedures and batch-to-batch variation in the biomarker assays. Laboratories
therefore establish their own reference values.64,71 The Alzheimer’s Association is
currently executing a quality control programme that aims to standardize CSF biomarker
measurements in both research and clinical laboratories.72
Aβ imaging is thus far only advocated for research purposes.2
C-Pittsburgh
11
compound B (PiB) binds to fibrillar Aβ in the brain and can be visualized using nuclear
imaging. Because this technique is relatively new, cases with post-mortem confirmation
of the PiB-PET findings are still limited. Some cases have shown a positive correlation
between PiB-binding and fibrillar Aβ deposition, others showed clear Aβ deposition
upon neuropathological examination but a negative PiB-PET scan during life.73-77
20
Introduction and Outline
Furthermore, it has been shown that 33% of control subjects had positive PiB-PET scans
and that PiB-retention increased in elderly without dementia that were followed
longitudinally.78,79 The clinical relevance of these findings is as yet unclear and longer
follow up is needed to determine whether these are preclinical cases of AD or whether
these cases should be considered as false positive scans in truly healthy controls.
Despite the availability of this variety of biomarkers, differentiating between dementia
disorders in clinical practice can still be a challenge, as is shown in the following case report.
Case report
A 63-year-old woman was referred to our memory clinic for a second opinion. The past
two years she had become increasingly anxious in new situations. She had difficulty
finding words and engaging in conversations. In 2007, a neurologist identified no
cognitive disorder. At that time, she was diagnosed by a psychiatrist as having dysthymic
disorder and post-traumatic stress disorder. She was treated with venlafaxine 75 mg
once a day. When she presented to our memory clinic nine months later, her complaints
had not disappeared. The difficulty in engaging in conversations was still present and
she did not manage to finish daily chores because she kept forgetting what she was
doing. The anxiety had made her give up driving, internet banking and using the phone
and video recorder. There was severe loss of initiative. Her Mini Mental State Examination
(MMSE) score was 21 out of 30.80 The score on the Revised Cambridge Cognitive
Examination (CAMCOG-R), a multiple-domain cognitive assessment, was 68 out of 104
(cut-off adjusted for age and education 83).81,82 She scored 2 out of 15 on the geriatric
depression scale (GDS) (a score above 7 indicates depression).83 These results suggested
cognitive decline, possibly AD, since no other underlying disease was apparent, and
history taking and GDS score indicated that her depressive mood was in remission. An
MRI, made in 2007, was re-evaluated and showed asymmetrical frontotemporal atrophy
without medial temporal lobe atrophy (Figure 1). A lumbar puncture was performed,
because the screening neuropsychological assessment and the MRI were not fully
conclusive and left the possibility of a diagnosis of AD as well as FTD. CSF analysis
showed normal concentrations of all biomarkers [reference values]64: Aβ 42 560 pg/ml
[>500 pg/ml]; t-tau 335 pg/ml [<350 pg/ml]; p-tau 77 pg/ml [<85 pg/ml]. These results
made AD less likely. Specific neuropsychological testing was performed, which showed
below-average memory learning curves with unimpaired delayed recall and recognition,
and a decline in attention and executive functions with reduced information processing
21
Chapter ONE
Figure 1 Transversal T2-FLAIR showing frontal cortical and asymmetrical left
temporal lobe atrophy
speed and apathy. Apart from below-average word fluency, there were no language
disorders. The combination of MRI, CSF biomarker results and neuropsychological
examination at that moment resulted in the diagnosis FTD. Fifteen months later the
patient had deteriorated mildly. Her MMSE was 18 out of 30. No behavioural problems
had occurred and her mood was stable. This clinical follow up is atypical for FTD and AD
cannot be ruled out.
CSF biomarkers in clinical practice
As illustrated by this case report, additional examinations to collect biomarkers do not
always result in a clear-cut diagnosis. CSF analysis showed that biomarker concentrations
were all within the pre-determined normal range, but clinical follow up still left room for
AD. It is worth noticing that all CSF biomarker concentrations were close to the cut-off
values. It has been suggested that in patients with a naturally occurring high Aβ load, a
decrease in Aβ 42 with AD may not be as clear.84 Relating Aβ 42 to Aβ 40, the latter as a
measure of the total Aβ load, was suggested to provide a more precise reflection of the
decrease in Aβ 42. This ratio could be a means to improve the accuracy of the existing
biomarker CSF Aβ 42. Questions like these fuelled us to start this clinical diagnostic
22
Introduction and Outline
research. Whether this Aβ 42/Aβ 40 ratio can help us to differentiate between dementia
disorders, alone and in combination with p-tau and t-tau, was addressed in Chapter Two.
Besides improving the accuracy of existing biomarkers, differentiating between
dementia disorders may also be improved if new, specific biomarkers are found. Most
research has focused on biomarkers for AD, however, if hallmarks of other dementias
could be measured, these could serve as biomarkers for that type of dementia. An
example of such a hallmark is α-synuclein, the protein found to accumulate in DLB. It
was isolated from brain tissue in 199385 and was identified in CSF ten years later.86
Chapter Three and Four describe the properties of CSF α-synuclein as a possible
biomarker for DLB. However, finding new CSF biomarkers is only useful if clinicians are
willing to use lumbar puncture as a diagnostic tool. Therefore, the current use of CSF
analysis in the diagnostic work up of memory clinic patients, both in the Netherlands
and in Europe, was investigated in chapter Five. Finally, in order to facilitate the
interpretation of CSF biomarkers in clinical practice, we developed a prediction model
based on CSF biomarkers which estimates the probability that dementia in a memory
clinic patient is due to AD (Chapter Six).
Hypotheses for the pathophysiological cascade of AD
Although the involvement of plaques and tangles in AD has been known for over a
century, many questions around the pathophysiological cascade of AD remain
unresolved. Several hypotheses have been formulated, the amyloid cascade hypothesis
being the oldest one that still finds support (Figure 2).87-89 It postulated that the critical
event in AD is the deposition of Aβ as a result of APP mismetabolism. Aβ, or APP
cleavage products that contain Aβ, were suggested to be neurotoxic and to lead to
­phosphorylation of tau, possible by increasing the intraneuronal calcium concentration.
Hyperphosphorylated tau polymerized into neurofibrillary tangles which further led to
neuronal death. The temporal sequence of the hypothesis (Aβ plaques precede tau
tangles) was based on cases of familial AD who displayed both Aβ plaques and neuro­
fibrillary tangles and in whom a mutation in APP was discovered. Further support was
found in reports of persons with Down syndrome, in whom it was shown that diffuse
plaques were the first detectable lesion, followed by senile plaques, neurofibrillary
tangles, and eventually severe loss of neurons and atrophy.
A critical reappraisal almost 20 years later concluded that research findings thus far
are compatible with this theory, even though the pathway between Aβ and tau is still
23
Chapter ONE
Figure 2 The amyloid cascade hypothesis as defined in 1991
Reprinted from Trends Pharmacol Sci, Volume 12, Hardy J, Allsop D, Amyloid deposition as the central event
in the aetiology of Alzheimer's disease, Page 387, Copyright (1991), with permission from Elsevier.
Figure 3 Hypothetical model of the AD pathological cascade
Biomarker magnitude
Abnormal
Aβ
Tau-meditated neuronal injury and dysfunction
Brain structure
Memory
Clinical function
Normal
Cognitively normal
MCI
Dementia
Clinical disease stage
Reprinted from Lancet Neurology, Volume 9, Jack CR, Jr., Knopman DS, Jagust WJ, Shaw LM, Aisen PS,
Weiner MW, Petersen RC, Trojanowski JQ, Hypothetical model of dynamic biomarkers of the Alzheimer's
pathological cascade, Page 122, Copyright (2010), with permission from Elsevier.
24
Introduction and Outline
Figure 4 T he vascular hypothesis
Vascular Dementia
Vascular risk factors
Alzheimer’s Dementia
Abeta clearance
Amyloid precursor protein
Secretase activity
Vascular
Insufficiency
Beta-Amyloid
Vascular dysregulation
Brain dysfunction and damage
Dementia
With kind permission from Springer Science+Business Media: Acta Neuropathologica, The overlap between
neurodegenerative and vascular factors in the pathogenesis of dementia, Volume 120, 2010, Page 292,
Iadecola C, Figure 3.
unclear, and Aβ toxicity has not sufficiently been demonstrated in vivo to explain the
massive cell loss as seen in AD.90 Correlations between Aβ plaque load and cognitive
impairment or dementia severity have both been described and disputed91-95 and it has
been suggested that not the Aβ plaques, but the still soluble oligomeric Aβ species are
in fact the culprit.96
Recently, a hypothetical model of the AD pathological cascade was described that
incorporated information from studies on several AD biomarkers.44 In line with the
amyloid cascade hypothesis, it was suggested that the initiating event in AD is related
to abnormal processing of Aβ, resulting in Aβ deposition. Aβ biomarkers (decreased CSF
Aβ 42 concentration and increased uptake of PiB) were hypothesized to reach a plateau
before clinical symptoms are apparent (Figure 3). Biomarkers of neuronal injury and neurodegeneration, such as increased CSF t-tau and p-tau and cerebral atrophy on MRI,
were suggested to become abnormal in a later stage of the disease and to correlate
with clinical symptom severity.
25
Chapter ONE
The vascular hypothesis of AD can be regarded as an extension of the amyloid
hypothesis (Figure 4). It proposes a bidirectional relationship between cerebrovascular
disease and AD pathology.97 Cerebrovascular disease, through chronic or intermittent
hypoperfusion and oxidative stress, is suggested to lead to APP upregulation and
perhaps to Aβ deposition, thus initiating or aggravating AD.98-100 Vice versa, AD may lead
to vascular disease since Aβ exerts detrimental effects on cerebrovascular function.101
The influence of vascular factors on tau pathology is less well understood, although
ischemia might induce hyperphosphorylation of tau.97
Hypotheses serve no purpose if they are not tested and challenged. They often
incorporate information from a multitude of sources, such as animal models,
post-mortem research and explorative studies in selected patients groups with strict
in- and exclusion criteria. It cannot simply be assumed that these data can be
extrapolated to clinical practice. For example, previous studies showed that results from
animal models cannot always be replicated in humans102,103 and neuropathological
studies generally suffer from selection bias. For example, one of the largest and most
influential series of neuropathological studies on AD (the Nun study) consisted fully of
members of a religious order.104 These participants may not be representative for the
general population. Even for other neuropathological studies it is important to realize
that the subset of persons who agree to autopsy may not be representative. Testing
hypotheses in the population of interest is therefore an essential part of research. In
Chapter Seven the proposed separation between the timing of Aβ and tau pathology
was investigated in an AD patient with several years of follow up. Chapter Eight takes a
closer look at the suggestion that Aβ biomarkers reach a plateau before clinical
symptoms are apparent, while tau biomarkers supposedly continue to increase during
the clinical stages and thus correlate with clinical symptom severity. The vascular
hypothesis is subject of Chapter Nine, in which the relation between cerebrovascular
disease and AD pathology was investigated in a sample of the memory clinic population.
Once it is established how pathological factors such as Aβ, tau and vascular factors
interact and how this is reflected in their biomarkers, these biomarkers may be used as
intermediate outcomes in trials that aim to influence the pathology of the disease. Since
AD is a disease that develops over several years or even decades, an intervention that
slows or halts this development cannot be expected to yield obvious clinical results in
a short period of time. A major advantage of biomarkers is that they may provide early
evidence that an intervention is successful in changing a pathological process. With the
search for disease-modifying treatments, CSF biomarkers, and especially Aβ 42, have
been advocated as markers to detect and monitor the biochemical effects of newly
26
Introduction and Outline
developed drugs for AD.105,106 To be able to interpret changes in biomarker concentrations
that occur during a pharmacological intervention, it is important to understand the
reason why these biomarker concentrations were abnormal in the first place. In Chapter
Ten, (theoretical) reasons why the CSF Aβ 42 concentration is decreased in AD have been
systematically examined.
Aims of this thesis
This thesis has two aims:
One, to explore the potential use of CSF biomarkers as diagnostic tools in the differential
diagnosis of dementia.
Two, to investigate whether CSF biomarkers, derived from a sample of memory clinic
patients, support the current hypotheses on the pathophysiological cascade of AD.
Outline of this thesis
This thesis is divided in two parts that address the aims outlined above by means of
several investigations. This is followed by a summary and general discussion of the findings.
Part One | CSF biomarkers in the differential diagnosis of dementia
Chapter Two investigates the diagnostic value of the CSF Aβ 42/Aβ 40 ratio in differentiating patients with AD from controls and from patients with either VaD, DLB or FTD. It is
hypothesized that this ratio provides a more valid measure for reflecting changes in Aβ
levels in dementia patients than CSF Aβ 42 alone. The Aβ 42/Aβ 40 ratio combined with
p-tau and t-tau is also compared to the combination of Aβ42, p-tau and t-tau.
Chapter Three examines the value of CSF α-synuclein as biomarker for DLB, by
comparing its concentration between groups of DLB, AD, VaD and FTD patients. In
addition, a comparison to the CSF α-synuclein concentration in controls is made.
Chapter Four further explores the properties of the CSF α-synuclein concentration in
two ways: the presence of a circadian rhythm in CSF α-synuclein concentrations is
27
Chapter ONE
investigated, and the correlation between CSF α-synuclein and Aβ concentrations in
healthy elderly, AD and DLB patients is investigated.
Chapter Five assesses the current use of CSF analysis in memory clinics in Europe and
in the Netherlands.
Chapter Six describes the development and validation of a prediction model for AD
based on CSF biomarkers, meant to facilitate the use of CSF biomarkers in clinical
practice. A few models have been described previously, but they do not offer a formula
that can be applied in clinical practice for the differentiation between dementia due to
AD and dementia due to another cause.107,108 The prediction model in this chapter
translates the results of CSF analysis into the probability that dementia in a memory
clinic patient is due to AD.
Part Two | CSF biomarkers in pathophysiological hypotheses
Chapter Seven challenges the hypothetical model of the AD pathophysiological
cascade (Figure 3) with empirical data. Especially the separation in the moments at
which Aβ and tau become abnormal is questioned. This will be illustrated by a case of
early-onset AD with the longest follow-up of CSF biomarkers reported in the literature.
Chapter Eight explores the correlation between CSF biomarkers and cognitive
performance in a heterogeneous group of memory clinic patients. It is hypothesized
that changes in the CSF biomarkers reflect neuronal damage and directly underlie
cognitive impairment. According to the suggested pathophysiological cascade of AD,
the strongest correlation should be expected with tau, and only a weak or absent
correlation with Aβ.
Chapter Nine investigates the hypothesis that AD patients have an increased
susceptibility to develop cerebrovascular disease. In addition, possible mechanisms for
this relationship are investigated, using CSF Aβ 42, CSF Aβ 40 and APOE-ε4 status.
Chapter Ten reviews the evidence for several proposed reasons why the CSF Aβ42
concentration is decreased in AD, in order to create a better understanding of the
underlying mechanisms that lead to its recognition as an AD biomarker.
Chapter Eleven provides a summary and discussion of the findings presented in this thesis.
28
Introduction and Outline
References
1. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, Text Revision.
Washington DC, American Psychiatric Association 2000.
2. McKhann GM, Knopman DS, Chertkow H, Hyman BT, Jack CR, Jr., Kawas CH, Klunk WE, Koroshetz WJ, Manly
JJ, Mayeux R, Mohs RC, Morris JC, Rossor MN, Scheltens P, Carrillo MC, Thies B, Weintraub S, Phelps CH. The
diagnosis of dementia due to Alzheimer’s disease: Recommendations from the National Institute on AgingAlzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement
2011;7:263-9.
3. Kalaria R. Similarities between Alzheimer’s disease and vascular dementia. J Neurol Sci 2002;203-204:29-34.
4. McKeith I, Mintzer J, Aarsland D, Burn D, Chiu H, Cohen-Mansfield J, Dickson D, Dubois B, Duda JE, Feldman
H, Gauthier S, Halliday G, Lawlor B, Lippa C, Lopez OL, Carlos MJ, O’Brien J, Playfer J, Reid W. Dementia with
Lewy bodies. Lancet Neurol 2004;3:19-28.
5. Roman GC, Tatemichi TK, Erkinjuntti T, Cummings JL, Masdeu JC, Garcia JH, Amaducci L, Orgogozo JM, Brun
A, Hofman A. Vascular dementia: diagnostic criteria for research studies. Report of the NINDS-AIREN
International Workshop. Neurology 1993;43:250-60.
6. McKeith IG, Galasko D, Kosaka K, Perry EK, Dickson DW, Hansen LA, Salmon DP, Lowe J, Mirra SS, Byrne EJ,
Lennox G, Quinn NP, Edwardson JA, Ince PG, Bergeron C, Burns A, Miller BL, Lovestone S, Collerton D, Jansen
EN, Ballard C, de Vos RA, Wilcock GK, Jellinger KA, Perry RH. Consensus guidelines for the clinical and
pathologic diagnosis of dementia with Lewy bodies (DLB): report of the consortium on DLB international
workshop. Neurology 1996;47:1113-24.
7. Neary D, Snowden JS, Gustafson L, Passant U, Stuss D, Black S, Freedman M, Kertesz A, Robert PH, Albert M,
Boone K, Miller BL, Cummings J, Benson DF. Frontotemporal lobar degeneration: a consensus on clinical
diagnostic criteria. Neurology 1998;51:1546-54.
8. Cairns NJ, Bigio EH, Mackenzie IR, Neumann M, Lee VM, Hatanpaa KJ, White CL, III, Schneider JA, Grinberg LT,
Halliday G, Duyckaerts C, Lowe JS, Holm IE, Tolnay M, Okamoto K, Yokoo H, Murayama S, Woulfe J, Munoz
DG, Dickson DW, Ince PG, Trojanowski JQ, Mann DM. Neuropathologic diagnostic and nosologic criteria for
frontotemporal lobar degeneration: consensus of the Consortium for Frontotemporal Lobar Degeneration.
Acta Neuropathol 2007;114:5-22.
9. Hodges JR, Davies RR, Xuereb JH, Casey B, Broe M, Bak TH, Kril JJ, Halliday GM. Clinicopathological correlates
in frontotemporal dementia. Ann Neurol 2004;56:399-406.
10. Mackenzie IR, Neumann M, Bigio EH, Cairns NJ, Alafuzoff I, Kril J, Kovacs GG, Ghetti B, Halliday G, Holm IE, Ince
PG, Kamphorst W, Revesz T, Rozemuller AJ, Kumar-Singh S, Akiyama H, Baborie A, Spina S, Dickson DW,
Trojanowski JQ, Mann DM. Nomenclature and nosology for neuropathologic subtypes of frontotemporal
lobar degeneration: an update. Acta Neuropathol 2010;119:1-4.
11. Seilhean D, Le B, I, Sarazin M, Lacomblez L, Millecamps S, Salachas F, Pradat PF, Le FN, LeGuern E, Dubois B,
Meininger V, Brice A, Hauw JJ, Duyckaerts C. Fronto-temporal lobar degeneration: neuropathology in 60
cases. J Neural Transm 2011;118:753-64.
12. McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stadlan EM. Clinical diagnosis of Alzheimer’s
disease: report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human
Services Task Force on Alzheimer’s Disease. Neurology 1984;34:939-44.
13. Alzheimer A, Stelzmann RA, Schnitzlein HN, Murtagh FR. An English translation of Alzheimer’s 1907 paper,
“Uber eine eigenartige Erkankung der Hirnrinde”. Clin Anat 1995;8:429-31.
14. Grundke-Iqbal I, Iqbal K, Tung YC, Quinlan M, Wisniewski HM, Binder LI. Abnormal phosphorylation of the
microtubule-associated protein tau (tau) in Alzheimer cytoskeletal pathology. Proc Natl Acad Sci U S A
1986;83:4913-7.
15. Grundke-Iqbal I, Iqbal K, Quinlan M, Tung YC, Zaidi MS, Wisniewski HM. Microtubule-associated protein tau.
A component of Alzheimer paired helical filaments. J Biol Chem 1986;261:6084-9.
16. Grundke-Iqbal I, Vorbrodt AW, Iqbal K, Tung YC, Wang GP, Wisniewski HM. Microtubule-associated
polypeptides tau are altered in Alzheimer paired helical filaments. Brain Res 1988;464:43-52.
29
Chapter ONE
17. Iqbal K, Grundke-Iqbal I, Zaidi T, Merz PA, Wen GY, Shaikh SS, Wisniewski HM, Alafuzoff I, Winblad B. Defective
brain microtubule assembly in Alzheimer’s disease. Lancet 1986;2:421-6.
18. Iqbal K, Grundke-Iqbal I, Smith AJ, George L, Tung YC, Zaidi T. Identification and localization of a tau peptide
to paired helical filaments of Alzheimer disease. Proc Natl Acad Sci U S A 1989;86:5646-50.
19. Hansson O, Zetterberg H, Buchhave P, Andreasson U, Londos E, Minthon L, Blennow K. Prediction of
Alzheimer’s disease using the CSF Abeta42/Abeta40 ratio in patients with mild cognitive impairment.
Dement Geriatr Cogn Disord 2007;23:316-20.
20. Kanai M, Matsubara E, Isoe K, Urakami K, Nakashima K, Arai H, Sasaki H, Abe K, Iwatsubo T, Kosaka T, Watanabe
M, Tomidokoro Y, Shizuka M, Mizushima K, Nakamura T, Igeta Y, Ikeda Y, Amari M, Kawarabayashi T, Ishiguro
K, Harigaya Y, Wakabayashi K, Okamoto K, Hirai S, Shoji M. Longitudinal study of cerebrospinal fluid levels of
tau, A beta1-40, and A beta1-42(43) in Alzheimer’s disease: a study in Japan. Ann Neurol 1998;44:17-26.
21. Lewczuk P, Esselmann H, Otto M, Maler JM, Henkel AW, Henkel MK, Eikenberg O, Antz C, Krause WR, Reulbach
U, Kornhuber J, Wiltfang J. Neurochemical diagnosis of Alzheimer’s dementia by CSF Abeta42, Abeta42/
Abeta40 ratio and total tau. Neurobiol Aging 2004;25:273-81.
22. Iwatsubo T, Odaka A, Suzuki N, Mizusawa H, Nukina N, Ihara Y. Visualization of A beta 42(43) and A beta 40 in
senile plaques with end-specific A beta monoclonals: evidence that an initially deposited species is A beta
42(43). Neuron 1994;13:45-53.
23. Verbeek MM, Eikelenboom P, De Waal RM. Differences between the pathogenesis of senile plaques and
congophilic angiopathy in Alzheimer disease. J Neuropathol Exp Neurol 1997;56:751-61.
24. Rensink AA, De Waal RM, Kremer B, Verbeek MM. Pathogenesis of cerebral amyloid angiopathy. Brain Res
Brain Res Rev 2003;43:207-23.
25. Gravina SA, Ho L, Eckman CB, Long KE, Otvos L, Jr., Younkin LH, Suzuki N, Younkin SG. Amyloid beta protein
(A beta) in Alzheimer’s disease brain. Biochemical and immunocytochemical analysis with antibodies
specific for forms ending at A beta 40 or A beta 42(43). J Biol Chem 1995;270:7013-6.
26. Braak E, Griffing K, Arai K, Bohl J, Bratzke H, Braak H. Neuropathology of Alzheimer’s disease: what is new
since A. Alzheimer? Eur Arch Psychiatry Clin Neurosci 1999;249 Suppl 3:14-22.
27. Crowther RA. Structural aspects of pathology in Alzheimer’s disease. Biochim Biophys Acta 1990;1096:1-9.
28. www.cbo.nl/Downloads/387/rl_dement_2005.pdf. CBO 2005.
29. Riedijk S, Duivenvoorden H, Van SJ, Niermeijer M, Tibben A. Sense of competence in a Dutch sample of
informal caregivers of frontotemporal dementia patients. Dement Geriatr Cogn Disord 2009;27:337-43.
30. de Rotrou J, Cantegreil I, Faucounau V, Wenisch E, Chausson C, Jegou D, Grabar S, Rigaud AS. Do patients
diagnosed with Alzheimer’s disease benefit from a psycho-educational programme for family caregivers?
A randomised controlled study. Int J Geriatr Psychiatry 2011;26:833-42.
31. Aarsland D, Kurz M, Beyer M, Bronnick K, Piepenstock NS, Ballard C. Early discriminatory diagnosis of
dementia with Lewy bodies. The emerging role of CSF and imaging biomarkers. Dement Geriatr Cogn
Disord 2008;25:195-205.
32. Olichney JM, Galasko D, Salmon DP, Hofstetter CR, Hansen LA, Katzman R, Thal LJ. Cognitive decline is faster
in Lewy body variant than in Alzheimer’s disease. Neurology 1998;51:351-7.
33. McKeith I, Fairbairn A, Perry R, Thompson P, Perry E. Neuroleptic sensitivity in patients with senile dementia
of Lewy body type. BMJ 1992;305:673-8.
34. Albert MS, DeKosky ST, Dickson D, Dubois B, Feldman HH, Fox NC, Gamst A, Holtzman DM, Jagust WJ,
Petersen RC, Snyder PJ, Carrillo MC, Thies B, Phelps CH. The diagnosis of mild cognitive impairment due to
Alzheimer’s disease: recommendations from the National Institute on Aging-Alzheimer’s Association
workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 2011;7:270-9.
35. Sperling RA, Aisen PS, Beckett LA, Bennett DA, Craft S, Fagan AM, Iwatsubo T, Jack CR, Jr., Kaye J, Montine TJ,
Park DC, Reiman EM, Rowe CC, Siemers E, Stern Y, Yaffe K, Carrillo MC, Thies B, Morrison-Bogorad M, Wagster
MV, Phelps CH. Toward defining the preclinical stages of Alzheimer’s disease: recommendations from the
National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s
disease. Alzheimers Dement 2011;7:280-92.
36. Kazee AM, Eskin TA, Lapham LW, Gabriel KR, McDaniel KD, Hamill RW. Clinicopathologic correlates in
Alzheimer disease: assessment of clinical and pathologic diagnostic criteria. Alzheimer Dis Assoc Disord
1993;7:152-64.
30
Introduction and Outline
37. Mendez MF, Mastri AR, Sung JH, Frey WH. Clinically diagnosed Alzheimer disease: neuropathologic findings
in 650 cases. Alzheimer Dis Assoc Disord 1992;6:35-43.
38. Nelson PT, Jicha GA, Kryscio RJ, Abner EL, Schmitt FA, Cooper G, Xu LO, Smith CD, Markesbery WR. Low
sensitivity in clinical diagnoses of dementia with Lewy bodies. J Neurol 2010;257:359-66.
39. Molsa PK, Paljarvi L, Rinne JO, Rinne UK, Sako E. Validity of clinical diagnosis in dementia: a prospective
clinicopathological study. J Neurol Neurosurg Psychiatry 1985;48:1085-90.
40. Biomarkers and surrogate endpoints: preferred definitions and conceptual framework. Clin Pharmacol Ther
2001;69:89-95.
41. Dubois B, Feldman HH, Jacova C, DeKosky ST, Barberger-Gateau P, Cummings J, Delacourte A, Galasko D,
Gauthier S, Jicha G, Meguro K, O’Brien J, Pasquier F, Robert P, Rossor M, Salloway S, Stern Y, Visser PJ, Scheltens
P. Research criteria for the diagnosis of Alzheimer’s disease: revising the NINCDS-ADRDA criteria. Lancet
Neurol 2007;6:734-46.
42. Dubois B, Feldman HH, Jacova C, Cummings JL, DeKosky ST, Barberger-Gateau P, Delacourte A, Frisoni G, Fox
NC, Galasko D, Gauthier S, Hampel H, Jicha GA, Meguro K, O’Brien J, Pasquier F, Robert P, Rossor M, Salloway
S, Sarazin M, de Souza LC, Stern Y, Visser PJ, Scheltens P. Revising the definition of Alzheimer’s disease: a new
lexicon. Lancet Neurol 2010;9:1118-27.
43. Consensus report of the Working Group on: “Molecular and Biochemical Markers of Alzheimer’s Disease”.
The Ronald and Nancy Reagan Research Institute of the Alzheimer’s Association and the National Institute
on Aging Working Group. Neurobiol Aging 1998;19:109-16.
44. Jack CR, Jr., Knopman DS, Jagust WJ, Shaw LM, Aisen PS, Weiner MW, Petersen RC, Trojanowski JQ.
Hypothetical model of dynamic biomarkers of the Alzheimer’s pathological cascade. Lancet Neurol
2010;9:119-28.
45. de Leon MJ, George AE, Stylopoulos LA, Smith G, Miller DC. Early marker for Alzheimer’s disease: the atrophic
hippocampus. Lancet 1989;2:672-3.
46. Ball MJ, Fisman M, Hachinski V, Blume W, Fox A, Kral VA, Kirshen AJ, Fox H, Merskey H. A new definition of
Alzheimer’s disease: a hippocampal dementia. Lancet 1985;1:14-6.
47. Scheltens P, Leys D, Barkhof F, Huglo D, Weinstein HC, Vermersch P, Kuiper M, Steinling M, Wolters EC, Valk J.
Atrophy of medial temporal lobes on MRI in “probable” Alzheimer’s disease and normal ageing: diagnostic
value and neuropsychological correlates. J Neurol Neurosurg Psychiatry 1992;55:967-72.
48. de Leon MJ, George AE, Golomb J, Tarshish C, Convit A, Kluger A, De SS, McRae T, Ferris SH, Reisberg B, Ince
C, Rusinek H, Bobinski M, Quinn B, Miller DC, Wisniewski HM. Frequency of hippocampal formation atrophy
in normal aging and Alzheimer’s disease. Neurobiol Aging 1997;18:1-11.
49. Jack CR, Jr., Dickson DW, Parisi JE, Xu YC, Cha RH, O’Brien PC, Edland SD, Smith GE, Boeve BF, Tangalos EG,
Kokmen E, Petersen RC. Antemortem MRI findings correlate with hippocampal neuropathology in typical
aging and dementia. Neurology 2002;58:750-7.
50. Ibanez V, Pietrini P, Alexander GE, Furey ML, Teichberg D, Rajapakse JC, Rapoport SI, Schapiro MB, Horwitz B.
Regional glucose metabolic abnormalities are not the result of atrophy in Alzheimer’s disease. Neurology
1998;50:1585-93.
51. Hoffman JM, Welsh-Bohmer KA, Hanson M, Crain B, Hulette C, Earl N, Coleman RE. FDG PET imaging in
patients with pathologically verified dementia. J Nucl Med 2000;41:1920-8.
52. Jagust W, Reed B, Mungas D, Ellis W, DeCarli C. What does fluorodeoxyglucose PET imaging add to a clinical
diagnosis of dementia? Neurology 2007;69:871-7.
53. Palmert MR, Podlisny MB, Witker DS, Oltersdorf T, Younkin LH, Selkoe DJ, Younkin SG. The beta-amyloid
protein precursor of Alzheimer disease has soluble derivatives found in human brain and cerebrospinal
fluid. Proc Natl Acad Sci U S A 1989;86:6338-42.
54. Seubert P, Vigo-Pelfrey C, Esch F, Lee M, Dovey H, Davis D, Sinha S, Schlossmacher M, Whaley J, Swindlehurst C,
. Isolation and quantification of soluble Alzheimer’s beta-peptide from biological fluids. Nature 1992;359:325-7.
55. Shoji M, Golde TE, Ghiso J, Cheung TT, Estus S, Shaffer LM, Cai XD, McKay DM, Tintner R, Frangione B, .
Production of the Alzheimer amyloid beta protein by normal proteolytic processing. Science 1992;258:126-9.
56. Vandermeeren M, Mercken M, Vanmechelen E, Six J, van d, V, Martin JJ, Cras P. Detection of tau proteins in
normal and Alzheimer’s disease cerebrospinal fluid with a sensitive sandwich enzyme-linked
immunosorbent assay. J Neurochem 1993;61:1828-34.
31
Chapter ONE
57. Weidemann A, Konig G, Bunke D, Fischer P, Salbaum JM, Masters CL, Beyreuther K. Identification, biogenesis,
and localization of precursors of Alzheimer’s disease A4 amyloid protein. Cell 1989;57:115-26.
58. Galasko D, Chang L, Motter R, Clark CM, Kaye J, Knopman D, Thomas R, Kholodenko D, Schenk D, Lieberburg
I, Miller B, Green R, Basherad R, Kertiles L, Boss MA, Seubert P. High cerebrospinal fluid tau and low amyloid
beta42 levels in the clinical diagnosis of Alzheimer disease and relation to apolipoprotein E genotype. Arch
Neurol 1998;55:937-45.
59. Motter R, Vigo-Pelfrey C, Kholodenko D, Barbour R, Johnson-Wood K, Galasko D, Chang L, Miller B, Clark C,
Green R, . Reduction of beta-amyloid peptide42 in the cerebrospinal fluid of patients with Alzheimer’s
disease. Ann Neurol 1995;38:643-8.
60. Shoji M, Matsubara E, Kanai M, Watanabe M, Nakamura T, Tomidokoro Y, Shizuka M, Wakabayashi K, Igeta Y,
Ikeda Y, Mizushima K, Amari M, Ishiguro K, Kawarabayashi T, Harigaya Y, Okamoto K, Hirai S. Combination
assay of CSF tau, A beta 1-40 and A beta 1-42(43) as a biochemical marker of Alzheimer’s disease. J Neurol
Sci 1998;158:134-40.
61. Tamaoka A, Sawamura N, Fukushima T, Shoji S, Matsubara E, Shoji M, Hirai S, Furiya Y, Endoh R, Mori H. Amyloid
beta protein 42(43) in cerebrospinal fluid of patients with Alzheimer’s disease. J Neurol Sci 1997;148:41-5.
62. Vigo-Pelfrey C, Seubert P, Barbour R, Blomquist C, Lee M, Lee D, Coria F, Chang L, Miller B, Lieberburg I, .
Elevation of microtubule-associated protein tau in the cerebrospinal fluid of patients with Alzheimer’s
disease. Neurology 1995;45:788-93.
63. Blennow K. Cerebrospinal fluid protein biomarkers for Alzheimer’s disease. NeuroRx 2004;1:213-25.
64. de Jong D, Kremer BP, Olde Rikkert MG, Verbeek MM. Current state and future directions of neurochemical
biomarkers for Alzheimer’s disease. Clin Chem Lab Med 2007;45:1421-34.
65. Vanmechelen E, Vanderstichele H, Hulstaert F, Andreasen N, Minthon L, Winblad B, Davidsson P, Blennow K.
Cerebrospinal fluid tau and beta-amyloid(1-42) in dementia disorders. Mech Ageing Dev 2001;122:2005-11.
66. Verbeek MM, de Jong D, Kremer HP. Brain-specific proteins in cerebrospinal fluid for the diagnosis of neurodegenerative diseases. Ann Clin Biochem 2003;40:25-40.
67. de Jong D, Jansen RW, Kremer BP, Verbeek MM. Cerebrospinal fluid amyloid beta42/phosphorylated tau
ratio discriminates between Alzheimer’s disease and vascular dementia. J Gerontol A Biol Sci Med Sci
2006;61:755-8.
68. Mukaetova-Ladinska EB, Monteith R, Perry EK. Cerebrospinal fluid biomarkers for dementia with lewy
bodies. Int J Alzheimers Dis 2010;2010:536538.
69. Schoonenboom NS, Pijnenburg YA, Mulder C, Rosso SM, van Elk EJ, Van Kamp GJ, Van Swieten JC, Scheltens
P. Amyloid beta(1-42) and phosphorylated tau in CSF as markers for early-onset Alzheimer disease.
Neurology 2004;62:1580-4.
70. Verwey NA, van der Flier WM, Blennow K, Clark C, Sokolow S, De Deyn PP, Galasko D, Hampel H, Hartmann T,
Kapaki E, Lannfelt L, Mehta PD, Parnetti L, Petzold A, Pirttila T, Saleh L, Skinningsrud A, Swieten JC, Verbeek
MM, Wiltfang J, Younkin S, Scheltens P, Blankenstein MA. A worldwide multicentre comparison of assays for
cerebrospinal fluid biomarkers in Alzheimer’s disease. Ann Clin Biochem 2009;46:235-40.
71. Mulder C, Verwey NA, van der Flier WM, Bouwman FH, Kok A, van Elk EJ, Scheltens P, Blankenstein MA.
Amyloid-beta(1-42), total tau, and phosphorylated tau as cerebrospinal fluid biomarkers for the diagnosis of
Alzheimer disease. Clin Chem 2010;56:248-53.
72. Mattsson N, Andreasson U, Persson S, Arai H, Batish SD, Bernardini S, Bocchio-Chiavetto L, Blankenstein MA,
Carrillo MC, Chalbot S, Coart E, Chiasserini D, Cutler N, Dahlfors G, Duller S, Fagan AM, Forlenza O, Frisoni GB,
Galasko D, Galimberti D, Hampel H, Handberg A, Heneka MT, Herskovits AZ, Herukka SK, Holtzman DM,
Humpel C, Hyman BT, Iqbal K, Jucker M, Kaeser SA, Kaiser E, Kapaki E, Kidd D, Klivenyi P, Knudsen CS, Kummer
MP, Lui J, Llado A, Lewczuk P, Li QX, Martins R, Masters C, McAuliffe J, Mercken M, Moghekar A, Molinuevo JL,
Montine TJ, Nowatzke W, O’Brien R, Otto M, Paraskevas GP, Parnetti L, Petersen RC, Prvulovic D, de Reus HP,
Rissman RA, Scarpini E, Stefani A, Soininen H, Schroder J, Shaw LM, Skinningsrud A, Skrogstad B, Spreer A,
Talib L, Teunissen C, Trojanowski JQ, Tumani H, Umek RM, Van BB, Vanderstichele H, Vecsei L, Verbeek MM,
Windisch M, Zhang J, Zetterberg H, Blennow K. The Alzheimer’s Association external quality control
program for cerebrospinal fluid biomarkers. Alzheimers Dement 2011;7:386-95.
73. Sojkova J, Driscoll I, Iacono D, Zhou Y, Codispoti KE, Kraut MA, Ferrucci L, Pletnikova O, Mathis CA, Klunk WE,
O’Brien RJ, Wong DF, Troncoso JC, Resnick SM. In vivo fibrillar beta-amyloid detected using [11C]PiB positron
emission tomography and neuropathologic assessment in older adults. Arch Neurol 2011;68:232-40.
32
Introduction and Outline
74. Ikonomovic MD, Klunk WE, Abrahamson EE, Mathis CA, Price JC, Tsopelas ND, Lopresti BJ, Ziolko S, Bi W,
Paljug WR, Debnath ML, Hope CE, Isanski BA, Hamilton RL, DeKosky ST. Post-mortem correlates of in vivo
PiB-PET amyloid imaging in a typical case of Alzheimer’s disease. Brain 2008;131:1630-45.
75. Bacskai BJ, Frosch MP, Freeman SH, Raymond SB, Augustinack JC, Johnson KA, Irizarry MC, Klunk WE, Mathis
CA, DeKosky ST, Greenberg SM, Hyman BT, Growdon JH. Molecular imaging with Pittsburgh Compound B
confirmed at autopsy: a case report. Arch Neurol 2007;64:431-4.
76. Rosen RF, Ciliax BJ, Wingo TS, Gearing M, Dooyema J, Lah JJ, Ghiso JA, LeVine H, III, Walker LC. Deficient high-affinity
binding of Pittsburgh compound B in a case of Alzheimer’s disease. Acta Neuropathol 2010;119:221-33.
77. Cairns NJ, Ikonomovic MD, Benzinger T, Storandt M, Fagan AM, Shah AR, Reinwald LT, Carter D, Felton A,
Holtzman DM, Mintun MA, Klunk WE, Morris JC. Absence of Pittsburgh compound B detection of cerebral
amyloid beta in a patient with clinical, cognitive, and cerebrospinal fluid markers of Alzheimer disease: a
case report. Arch Neurol 2009;66:1557-62.
78. Sojkova J, Zhou Y, An Y, Kraut MA, Ferrucci L, Wong DF, Resnick SM. Longitudinal patterns of beta-amyloid
deposition in nondemented older adults. Arch Neurol 2011;68:644-9.
79. Rowe CC, Ellis KA, Rimajova M, Bourgeat P, Pike KE, Jones G, Fripp J, Tochon-Danguy H, Morandeau L, O’Keefe
G, Price R, Raniga P, Robins P, Acosta O, Lenzo N, Szoeke C, Salvado O, Head R, Martins R, Masters CL, Ames D,
Villemagne VL. Amyloid imaging results from the Australian Imaging, Biomarkers and Lifestyle (AIBL) study
of aging. Neurobiol Aging 2010;31:1275-83.
80. Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”. A practical method for grading the cognitive state
of patients for the clinician. J Psychiatr Res 1975;12:189-98.
81. Roth M, Tym E, Mountjoy CQ, Huppert FA, Hendrie H, Verma S, Goddard R. CAMDEX. A standardised
instrument for the diagnosis of mental disorder in the elderly with special reference to the early detection
of dementia. Br J Psychiatry 1986;149:698-709.
82. Roth M, Huppert FA, Mountjoy CQ, Tym E. The Revised Cambridge Examination for Mental Disorders of the
Elderly. Cambridge University Press 1999.
83. D’Ath P, Katona P, Mullan E, Evans S, Katona C. Screening, detection and management of depression in
elderly primary care attenders. I: The acceptability and performance of the 15 item Geriatric Depression
Scale (GDS15) and the development of short versions. Fam Pract 1994;11:260-6.
84. Wiltfang J, Esselmann H, Bibl M, Hull M, Hampel H, Kessler H, Frolich L, Schroder J, Peters O, Jessen F,
Luckhaus C, Perneczky R, Jahn H, Fiszer M, Maler JM, Zimmermann R, Bruckmoser R, Kornhuber J, Lewczuk
P. Amyloid beta peptide ratio 42/40 but not A beta 42 correlates with phospho-Tau in patients with low- and
high-CSF A beta 40 load. J Neurochem 2007;101:1053-9.
85. Ueda K, Fukushima H, Masliah E, Xia Y, Iwai A, Yoshimoto M, Otero DA, Kondo J, Ihara Y, Saitoh T. Molecular
cloning of cDNA encoding an unrecognized component of amyloid in Alzheimer disease. Proc Natl Acad
Sci U S A 1993;90:11282-6.
86. El-Agnaf OM, Salem SA, Paleologou KE, Cooper LJ, Fullwood NJ, Gibson MJ, Curran MD, Court JA, Mann DM,
Ikeda S, Cookson MR, Hardy J, Allsop D. Alpha-synuclein implicated in Parkinson’s disease is present in
extracellular biological fluids, including human plasma. FASEB J 2003;17:1945-7.
87. Hardy J, Allsop D. Amyloid deposition as the central event in the aetiology of Alzheimer’s disease. Trends
Pharmacol Sci 1991;12:383-8.
88. Hardy JA, Higgins GA. Alzheimer’s disease: the amyloid cascade hypothesis. Science 1992;256:184-5.
89. Selkoe DJ. The molecular pathology of Alzheimer’s disease. Neuron 1991;6:487-98.
90. Hardy J. The amyloid hypothesis for Alzheimer’s disease: a critical reappraisal. J Neurochem 2009;110:1129-34.
91. Forman MS, Mufson EJ, Leurgans S, Pratico D, Joyce S, Leight S, Lee VM, Trojanowski JQ. Cortical biochemistry
in MCI and Alzheimer disease: lack of correlation with clinical diagnosis. Neurology 2007;68:757-63.
92. Naslund J, Haroutunian V, Mohs R, Davis KL, Davies P, Greengard P, Buxbaum JD. Correlation between
elevated levels of amyloid beta-peptide in the brain and cognitive decline. JAMA 2000;283:1571-7.
93. Parvathy S, Davies P, Haroutunian V, Purohit DP, Davis KL, Mohs RC, Park H, Moran TM, Chan JY, Buxbaum JD.
Correlation between Abetax-40-, Abetax-42-, and Abetax-43-containing amyloid plaques and cognitive
decline. Arch Neurol 2001;58:2025-32.
94. Terry RD, Masliah E, Salmon DP, Butters N, DeTeresa R, Hill R, Hansen LA, Katzman R. Physical basis of
cognitive alterations in Alzheimer’s disease: synapse loss is the major correlate of cognitive impairment.
Ann Neurol 1991;30:572-80.
33
Chapter ONE
95. Villemagne VL, Pike KE, Chetelat G, Ellis KA, Mulligan RS, Bourgeat P, Ackermann U, Jones G, Szoeke C,
Salvado O, Martins R, O’Keefe G, Mathis CA, Klunk WE, Ames D, Masters CL, Rowe CC. Longitudinal
assessment of Abeta and cognition in aging and Alzheimer disease. Ann Neurol 2011;69:181-92.
96. Tomic JL, Pensalfini A, Head E, Glabe CG. Soluble fibrillar oligomer levels are elevated in Alzheimer’s disease
brain and correlate with cognitive dysfunction. Neurobiol Dis 2009;35:352-8.
97. Iadecola C. The overlap between neurodegenerative and vascular factors in the pathogenesis of dementia.
Acta Neuropathol 2010;120:287-96.
98. Claassen JA. Alzheimer’s disease: trade-off for increased survival with atherosclerosis? Ann Neurol
2008;64:475.
99. Kalaria RN. The role of cerebral ischemia in Alzheimer’s disease. Neurobiol Aging 2000;21:321-30.
100. Kalaria RN. Linking cerebrovascular defense mechanisms in brain ageing and Alzheimer’s disease. Neurobiol
Aging 2009;30:1512-4.
101. Thomas T, Thomas G, McLendon C, Sutton T, Mullan M. beta-Amyloid-mediated vasoactivity and vascular
endothelial damage. Nature 1996;380:168-71.
102. Claassen JA, Zhang R. Cerebral autoregulation in Alzheimer’s disease. J Cereb Blood Flow Metab 2011;31:1572-7.
103. van der Worp HB, Howells DW, Sena ES, Porritt MJ, Rewell S, O’Collins V, Macleod MR. Can animal models of
disease reliably inform human studies? PLoS Med 2010;7:e1000245.
104. Snowdon DA, Greiner LH, Mortimer JA, Riley KP, Greiner PA, Markesbery WR. Brain infarction and the clinical
expression of Alzheimer disease. The Nun Study. JAMA 1997;277:813-7.
105. Hampel H, Frank R, Broich K, Teipel SJ, Katz RG, Hardy J, Herholz K, Bokde AL, Jessen F, Hoessler YC, Sanhai
WR, Zetterberg H, Woodcock J, Blennow K. Biomarkers for Alzheimer’s disease: academic, industry and
regulatory perspectives. Nat Rev Drug Discov 2010;9:560-74.
106. Blennow K, Hampel H, Weiner M, Zetterberg H. Cerebrospinal fluid and plasma biomarkers in Alzheimer
disease. Nat Rev Neurol 2010;6:131-44.
107. Hulstaert F, Blennow K, Ivanoiu A, Schoonderwaldt HC, Riemenschneider M, De Deyn PP, Bancher C, Cras P,
Wiltfang J, Mehta PD, Iqbal K, Pottel H, Vanmechelen E, Vanderstichele H. Improved discrimination of AD
patients using beta-amyloid(1-42) and tau levels in CSF. Neurology 1999;52:1555-62.
108. De Meyer G, Shapiro F, Vanderstichele H, Vanmechelen E, Engelborghs S, De Deyn PP, Coart E, Hansson O,
Minthon L, Zetterberg H, Blennow K, Shaw L, Trojanowski JQ. Diagnosis-independent Alzheimer disease
biomarker signature in cognitively normal elderly people. Arch Neurol 2010;67:949-56.
34
Introduction and Outline
35
CSF biomarkers in the differential
diagnosis of dementia
Part
One
Part One
Part One
The CSF Aβ42/Aβ40 ratio
This chapter has been published as: The cerebrospinal fluid amyloid β 42/40 ratio in the differentiation of
Alzheimer’s disease from non-Alzheimer’s dementia. Petra E. Spies, Diane Slats, Magnus J.C. Sjögren,
Berry P.H. Kremer, Frans R.J. Verhey, Marcel G.M. Olde Rikkert, Marcel M. Verbeek, Curr Alzheimer Res 2010;7:470-6
Chapter
Two
Chapter Two
Chapter Two
Chapter TWO
Abstract
Background: Amyloid β 40 (Aβ 40) is the most abundant Aβ peptide in the brain. The
cerebrospinal fluid (CSF) level of Aβ40 might therefore be considered to most closely
reflect the total Aβ load in the brain. Both in Alzheimer disease (AD) and in normal aging
the Aβ load in the brain has a large inter-individual variability. Relating Aβ42 to Aβ 40
levels might consequently provide a more valid measure for reflecting the change in Aβ
metabolism in dementia patients than the CSF Aβ42 concentrations alone. This measure
may also improve differential diagnosis between AD and other dementia syndromes,
such as vascular dementia (VaD), dementia with Lewy bodies (DLB), and frontotemporal
dementia (FTD).
Objective: To investigate the diagnostic value of the CSF Aβ 42/Aβ 40 ratio in differentiating AD from controls, VaD, DLB and FTD.
Methods: We analysed the CSF Aβ 42/Aβ 40 ratio, phosphorylated tau181 and total tau in
69 patients with AD, 26 patients with VaD, 16 patients with DLB, 27 patients with FTD,
and 47 controls.
Results: Mean Aβ 40 levels were 2850 pg/ml in VaD and 2830 pg/ml in DLB patients,
both significantly lower than in AD patients (3698 pg/ml; p<0.01). Aβ 40 levels in AD
patients were not significantly different from those in controls (4035 pg/ml; p=0.384).
The Aβ 42/Aβ 40 ratio was significantly lower in AD patients than in all other groups (p
<0.001, ANCOVA). Differentiating AD from VaD, DLB and non-AD dementia improved
when the Aβ 42/Aβ 40 ratio was used instead of Aβ 42 concentrations alone (p<0.01) The
Aβ 42/Aβ 40 ratio performed equally well as the combination of Aβ 42, phosphorylated
tau181 and total tau in differentiating AD from FTD and non-AD dementia. The diagnostic
performance of the latter combination was not improved when the Aβ 42/Aβ 40 ratio was
used instead of Aβ 42 alone.
Conclusion: The CSF Aβ 42/Aβ 40 ratio improves differentiation of AD patients from VaD,
DLB and non-AD dementia patients, when compared to Aβ 42 alone, and is a more easily
interpretable alternative to the combination of Aβ 42, p-tau and t-tau when differentiating AD from either FTD or non- AD dementia.
40
The CSF Aβ42/Aβ40 ratio
Introduction
While making an accurate diagnosis is important for management of dementia, differentiating between Alzheimer disease (AD) and other types of dementia using clinical
criteria is the true clinical challenge: difficult and prone to inaccuracy. The NINCDS-ADRDA
criteria for AD1 have been validated against neuropathological standards and turned
out to have a specificity of 23-69% for differentiating AD from other dementias.2,3
Analysis of cerebrospinal fluid (CSF) has been used increasingly to differentiate the
various dementia disorders.4 Although CSF analysis of amyloid β 42 (Aβ 42), phosphorylated
tau181 (p-tau) and total tau (t-tau) is efficacious in distinguishing between AD dementia
and non-demented controls, it is inaccurate in separating AD from other types of
dementia.5-7 CSF biomarker profiles of AD, vascular dementia (VaD), frontotemporal
dementia (FTD), and dementia with Lewy bodies (DLB) overlap, which makes
discrimination between these various dementias based on CSF analysis alone difficult.
It has been suggested that relating Aβ 42 to the concentration of Aβ 40 may improve
diagnostic accuracy of CSF analysis.8 Aβ 40 is the most abundant Aβ peptide in human
CSF, whereas Aβ 42 accounts for only 10% of the total Aβ population.6,9-11 As such, CSF
Aβ 40 concentrations could be considered as the closest possible reflection of the total
Aβ load in the brain. This total CSF Aβ load shows a large inter-individual variability.12
Patients with a low total Aβ load and therefore a low CSF Aβ 42 concentration11, might
receive an inappropriate diagnosis of AD if only CSF Aβ 42 is considered and an absolute
cut-off value is used. Similarly, in patients with AD but with a high total Aβ load, CSF
Aβ 42 concentrations may not decrease below the cut-off value and in those cases
rejection of the diagnosis AD might be incorrect. Since the total Aβ concentration was
found not to change in various dementia disorders, 8,10,13 and Aβ 40 concentrations were
shown to be similar in groups of AD patients, controls and non-AD dementia patients,13-17
hypothetically the ratio of Aβ 42 to Aβ 40 reflects the Aβ changes in the brain more
accurately than Aβ 42 alone.
It was found in several studies that differentiation between AD and healthy controls,
subjects with non-AD dementia, or subjects with other neurological disorders, improved
when using the CSF Aβ 42/Aβ 40 ratio compared to Aβ42 alone,15,17,18 although these
results were not confirmed in other studies.19,20 The Aβ 42/Aβ 40 ratio has previously been
studied in patients with FTD and AD, and a significantly decreased Aβ40 was found in
FTD patients, resulting in an increased Aβ42/Aβ 40 ratio compared to AD patients.21 In a
recent review it was concluded that levels of CSF Aβ 40 or Aβ 42 cannot usefully
discriminate between AD and DLB,22 although one study found that the Aβ 42/Aβ 40 ratio
41
Chapter TWO
improved diagnostic accuracy of AD versus DLB relative to Aβ 42 alone.23 To our
knowledge, the Aβ 42/Aβ 40 ratio has not been investigated in relation to VaD.
We hypothesized that differentiation between dementia syndromes would improve
by using the CSF Aβ 42/Aβ 40 ratio since it may eliminate inter-individual differences in
total Aβ load. Therefore, we explored the CSF Aβ40 concentration and Aβ42/Aβ 40 ratio
in patients with AD, VaD, DLB, FTD and in controls, and we investigated the diagnostic
value of the Aβ 42/Aβ 40 ratio in differentiating AD from each of these groups individually
as well as from the total group of non-AD dementia patients.
Materials and methods
Patients
Sixty-nine AD, 26 VaD, 27 FTD, and 16 DLB patients were included in this study that was
based on the CSF database of the Alzheimer Centre of the Radboud University Nijmegen
Medical Centre. This database contains clinical data as well as biobanked CSF and serum
of consecutive patients. All patients or their legal representative had given informed
consent for lumbar puncture. We included all patients with a clinically clear-cut
dementia diagnosis, whose CSF was available for Aβ40 and Aβ 42 analysis. Data on t-tau
and p-tau were also obtained. A diagnosis was made by a multidisciplinary team, which
consisted of a geriatrician, neurologist, neuropsychologist, and – if needed – an old-age
psychiatrist. The panel used the accepted clinical diagnostic criteria, i.e. the NINCDS-­
ADRDA criteria for probable AD, the NINDS-AIREN criteria for probable VaD, the 1998
consensus on clinical diagnostic criteria for FTD and the 1996 consensus guidelines for
the clinical and pathologic diagnosis of DLB.1,24-26 In less than 10% of the cases, the
clinicians were aware of the CSF results for Aβ 42, p-tau and t-tau when establishing the
clinical diagnosis, but they were never aware of Aβ 40 levels.
The control group consisted of 47 non-demented patients who underwent lumbar
puncture for various complaints. Most frequently diagnosed were headache (n=10),
polyneuropathy (n=5), radiculopathy (n=4), vertigo (n=3), and delirium (n=2). The
remaining controls had a variety of complaints but turned out not to have central
neurological problems. CSF cell count, glucose, lactate, haemoglobin, bilirubin, total
protein and oligoclonal IgG bands were normal in all controls. Patient characteristics are
listed in Table 1.
42
The CSF Aβ42/Aβ40 ratio
CSF
CSF was collected by lumbar puncture in polypropylene tubes, transported within 30
minutes to the adjacent laboratory at room temperature, centrifuged after routine
investigations, and immediately aliquoted and stored at -80 °C until analysis. Levels of
Aβ 42, t-tau, and p-tau in CSF were measured using enzyme linked immunosorbent
assays (Innogenetics NV, Ghent, Belgium). Inter-assay coefficients of variation for these
assays were 3.8 – 8.4%.27 CSF Aβ 40 was analysed by using a commercial assay based on
the Luminex technology (BioSource, Invitrogen Ltd, Paisley, UK). Inter-assay coefficient
of variation for Aβ40 was 5.4% (at 2875 pg/ml; n=32). The lower limit of detection was for
Aβ 42 50 pg/ml and for Aβ 40 22 pg/ml.
Statistical analysis
Since p-tau, t-tau, Aβ 42, Aβ 40, and the Aβ 42/Aβ 40 ratio were not all normally distributed,
they were log transformed. Differences in p-tau, t-tau, Aβ42, Aβ 40 and the Aβ 42/Aβ 40
ratio were analysed using 2-way ANCOVA. Since sex and age did not contribute to the
differences in Aβ 42, Aβ 40 and the Aβ 42/Aβ 40 ratio between dementia groups and
controls, both (sex and age) were removed from the model and the ANCOVA was
thereafter rerun. Differences between AD and the other groups were further explored
using Dunnett’s post-hoc test.
We used logistic regression analysis to assess the value of the Aβ 42/Aβ 40 ratio
compared to other combinations of CSF biomarkers in differentiating AD from controls,
from VaD, FTD, DLB, respectively, and from these three groups combined (‘non-AD
dementia’). Sex and age contributed significantly to the logistic regression models and
therefore the models were corrected for these factors. Receiver operating characteristic
(ROC) curves were calculated and the areas under the ROC curves were compared using
MedCalc (Mariakerke, Belgium). Sensitivity and specificity were determined based on the
highest Youden index, i.e. the point on the ROC curve at which sensitivity + specificity-1 is
maximized. All other statistical analyses were carried out using SPSS, version 16.0.
Results
The mean level of CSF Aβ 40 was significantly lower in VaD patients than in AD patients
(p<0.001), as well as in DLB patients compared to AD patients (p=0.008) (see Table 1 and
Figure 1. There was no significant difference in Aβ 40 levels between AD patients and
FTD patients (p=0.981) or controls (p=0.384). In AD patients, the concentration of CSF
43
Chapter TWO
Table 1 Clinical characteristics and results of CSF analysis
controls
males/females (n)
23/24
age (y)
61 ± 8
disease duration (mo) n.a.
t-tau (pg/ml)*
p-tau (pg/ml)*
Aβ42 (pg/ml)*
Aβ40 (pg/ml)*
Aβ42/Aβ40 ratio*
AD
34/35
69 ± 8
29 ± 23
n=60
221 ± 80
710 ± 390
49 ± 16
111 ± 52
840 ± 260
428 ± 127
4035 ± 1177 3698 ± 1096
0.21 ± 0.04 0.12 ± 0.04
VaD
17/9
72 ± 9
35 ± 29
n=20
391 ± 526
46 ± 16
631 ± 270
2850 ± 1031
0.23 ± 0.08
FTD
19/8
65 ± 7
34 ± 21
n=26
448 ± 315
70 ± 36
786 ± 229
3590 ± 1038
0.23 ± 0.06
DLB
12/4
76 ± 8
34 ± 27
n=8
257 ± 98
53 ± 17
499 ± 179
2830 ± 697
0.18 ± 0.06
Values are expressed as means ± SD.
*, p-value <0.001 by ANOVA.
AD, Alzheimer disease; VaD, vascular dementia; FTD, frontotemporal dementia; DLB, dementia with Lewy
bodies; t-tau, total tau; p-tau, phosphorylated tau181; n.a., not applicable
Aβ 42 was significantly lower than in VaD and FTD patients and controls (p<0.001),
whereas the concentration of p-tau and t-tau was significantly higher than in all other
groups (p<0.001). The Aβ 42/Aβ 40 ratio was significantly lower in AD patients than in all
other groups (p<0.001) (see Table 1 and Figure 1).
We investigated the diagnostic value of the Aβ 42/Aβ 40 ratio in several comparisons.
First, we investigated if the use of the Aβ 42/Aβ 40 ratio would be superior to the use of
Aβ 42 alone. Differentiation of AD from either VaD, DLB or non-AD dementia significantly
improved when using the Aβ 42/Aβ 40 ratio (p<0.01). In contrast, differentiation of AD
from either FTD or controls did not improve (p>0.05). Second, we analysed the
diagnostic value of the Aβ 42/Aβ 40 ratio compared to the combination of Aβ42, p-tau
and t-tau, which is a frequently used combination of CSF biomarkers in AD research.
Differentiation of AD from either FTD or non-AD dementia was equally good for the
Aβ 42/Aβ 40 ratio or the combination of Aβ 42, p-tau and t-tau (p>0.05). Differentiation of
AD from either controls, VaD or DLB was better using the combination of Aβ42, p-tau
and t-tau than using the Aβ42/Aβ 40 ratio (p<0.05). Third, we analysed if replacement of
Aβ 42 by the Aβ 42/Aβ 40 ratio in the combination with p-tau and t-tau would improve
results. Differentiation of AD from both controls and from all dementia groups was
equally good for both combinations (p>0.05). ROC curves are shown in Figure 2 and
Figure 3. Sensitivity and specificity, AUC and likelihood ratios of each parameter are
shown in Table 2.
44
The CSF Aβ42/Aβ40 ratio
Figure 1 C
erebrospinal fluid Aβ 42 and Aβ 40 concentrations and Aβ 42/Aβ 40 ratio
in controls and patients with various types of dementia
See text for statistical analyses.
Bold horizontal bar represents the median. Central box: 25th to 75th percentile (interquartile range). Vertical
bar bordered by highest and lowest value within 1.5 * interquartile range.
AD, Alzheimer disease; VaD, vascular dementia; FTD, frontotemporal dementia; DLB, dementia with Lewy
bodies
45
Chapter TWO
Figure 2 Receiver Operating Characteristic curves comparing Aβ42, and the
Aβ 42/Aβ 40 ratio in AD versus controls and versus various types of
dementia
AD, Alzheimer disease; VaD, vascular dementia; FTD, frontotemporal dementia; DLB, dementia with Lewy
bodies.
46
The CSF Aβ42/Aβ40 ratio
Figure 3 R eceiver Operating Characteristic curves comparing the Aβ42/Aβ40
ratio, the Aβ 42/Aβ 40 ratio combined with p-tau and t-tau, and the
combination of Aβ 42, p-tau and t-tau in AD versus controls and versus
various types of dementia
AD, Alzheimer disease; VaD, vascular dementia; FTD, frontotemporal dementia; DLB, dementia with Lewy
bodies.
47
48
93%
87%
93%
87%
97%
98%
93%
98%
45.6
43.5
0.980
7.3
7.3
LR
0.994‡
0.947
0.949
AUC
83%
69%
93%
84%
93%
96%
91%
96%
sens.
spec.
0.957
0.963‡
0.900*
0.785
AUC
AD vs VaD
22.8
23.1
5.8
2.7
LR
94%
85%
99%
83%
93%
88%
99%
79%
sens.
spec.
0.936
0.953
0.936
0.941
AUC
4.7
7.4
5.9
6.4
LR
AD vs FTD
65%
75%
85%
85%
90%
100%
90%
100%
sens.
spec.
0.970
0.966‡
0.893*
0.755
AUC
AD vs DLB
n.c.
n.c.
5.5
2.6
LR
83%
74%
90%
82%
93%
84%
87%
90%
sens.
spec.
0.923
0.924
0.903*
0.811
AUC
9.0
5.7
5.1
3.2
LR
AD vs non-AD
AD, Alzheimer disease; VaD, vascular dementia; FTD, frontotemporal dementia; DLB, dementia with Lewy bodies; non-AD, combined group of patients with VaD, FTD or DLB.
Sens., sensitivity; spec., specificity; AUC, area under the ROC curve; LR, likelihood ratio; n.c., not computable.
* p<0.01 compared to AUC for Aβ 42 level
‡ p<0.05 compared to AUC for Aβ 42 /Aβ 40 ratio
Aβ42, p-tau,
t-tau
Aβ42/Aβ40 ratio,
p-tau, t-tau
Aβ42/Aβ40 ratio
Aβ42
sens.
spec.
AD vs controls
types of dementia, using different combinations of CSF biomarkers
Table 2 Sensitivities and specificities, AUC and likelihood ratios for discriminating Alzheimer disease from controls and from other
Chapter TWO
The CSF Aβ42/Aβ40 ratio
Discussion
We explored the CSF Aβ40 concentrations in patients with different types of dementia
and in controls, and examined the value the Aβ 42/Aβ 40 ratio in differentiating AD from
controls, from FTD, DLB, VaD, and from these latter three groups combined. The
observed patterns of Aβ 40 and Aβ 42 in the various dementia groups are remarkable. As
expected, the AD group was characterized by normal Aβ40 and low Aβ42 levels, but
contrary to our expectations, the VaD group had low Aβ 40 and intermediate Aβ 42 levels,
whereas the DLB group was characterized by low concentrations of both Aβ40 and Aβ 42
levels. The FTD group had normal Aβ 40 and Aβ 42 levels.
Only one study has reported the CSF Aβ 40 concentration in patients with VaD. In
accordance with our results, a lower Aβ40 concentration in VaD patients compared to
AD patients was found.28 Multiple explanations for this low Aβ 40 concentration are
possible. For example, it may be caused by the ischemic damage that has occurred in
VaD patients, since cerebral ischemia has been shown to result in plaque formation and
accumulation of Aβ 40 and Aβ 4229-31 and, consequently, leads to a reduced CSF
concentration. In addition, atherosclerosis, one of the risk factors for VaD, may inhibit the
clearance of Aβ 40 across the blood-brain barrier, leading to vascular deposits of Aβ 40.32
Our finding of a lower Aβ 40 concentration in DLB patients compared to AD is in
accordance with the literature.12 Possibly, the intraneuronal α-synuclein aggregates that
are found in DLB affect Aβ metabolism, leading to an overall decrease in Aβ synthesis.
Results from in vitro research indicate an interaction between α-synuclein and Aβ.33,34
Neuropathological studies found a positive correlation between Aβ plaque burden and
α-synuclein load.35,36 Aβ 40 plaque level was greater in cases with α-synuclein aggregates
compared to cases lacking α-synuclein aggregates, while no difference was found in
Aβ 42 plaque burden.35 A greater Aβ 40 plaque burden in DLB might explain the decreased
Aβ 40 concentration in CSF that we found.
The few studies that measured Aβ 40 concentration in FTD patients report varying
results. Our results are supported by previous findings of similar concentrations of Aβ 40
in FTD patients and AD patients.37 In contrast with this, another group reported lower
Aβ 40 concentrations in FTD patients compared to AD patients.21 The varying results
might be explained by the inclusion of heterogeneous FTD patient groups in research,
comprising patients with clinically diagnosed FTD, semantic dementia and primary
progressive aphasia, presumably with different neuropathological substrates.
We investigated the Aβ 42/Aβ 40 ratio under the assumption that Aβ 40 closely
represents the total cerebral Aβ load. Since we found different Aβ 40 concentrations in
49
Chapter TWO
the various dementia groups, it can be debated if the Aβ 42/Aβ 40 ratio is a good
representation of the Aβ 42 fraction of the total Aβ load and thus eliminates inter-­
individual differences in total Aβ concentrations. Nevertheless, differentiating AD from
VaD, DLB and non-AD dementia improved when we used this ratio compared to Aβ 42
alone, probably due to the differences in Aβ 40 concentrations. Results obtained by
applying the Aβ 42/Aβ 40 ratio fulfilled the criteria for biomarkers as established by
the Working Group on molecular and biochemical markers of Alzheimer’s disease38 in
differentiating AD from VaD, FTD, DLB, and non-AD, since both sensitivity and specificity
were >80%.
The Aβ 42/Aβ 40 ratio seems to be a good alternative for the combination of Aβ42,
p-tau and t-tau, to distinguish AD from either FTD or non-AD dementia. The Aβ42/Aβ 40
ratio has the advantage that only two analyses are needed instead of three, and,
consequently, the results are easier to interpret. The differentiation of AD from non-AD
dementia is a frequently encountered problem in clinical practice, when dementia can
be established, but the differential diagnosis encompasses AD and another type of
dementia. Thus, the use of the CSF Aβ 42/Aβ 40 ratio may help in this clinical decision
making. However, the use of four biomarkers, i.e. the Aβ 42/Aβ 40 ratio combined with
p-tau and t-tau, did not improve differentiation when compared to the combination of
the three currently used biomarkers Aβ 42, p-tau and t-tau.
Some bias may have occurred in our analysis since in less than 10% of the patients
the results of the CSF analysis for Aβ42, t-tau and p-tau were known to the clinicians.
However, given the small fraction of patients this applies to, we believe that this did not
affect our results, also since our sensitivity and specificity results for discriminating AD
from other dementias using Aβ 42 are comparable with the literature.18,39,40 The number
of patients in our non-AD groups were moderate, yet not very different from the
numbers mentioned in other studies.12,21,28,37 Besides, the significant results that we
found had a p-value <0.01, suggesting that these results are not expected to change
with an increasing number of patients.
In summary, the Aβ 42/Aβ 40 ratio is more accurate for differentiating AD from other
types of dementia than Aβ 42 alone, and can be a good and easier interpretable
alternative for the combination of the established biomarkers Aβ 42, p-tau and t-tau for
the differentiation of AD from non-AD dementia, which may help to accept or refute the
clinical diagnosis of AD.
50
The CSF Aβ42/Aβ40 ratio
References
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stadlan EM. Clinical diagnosis of Alzheimer’s
disease: report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human
Services Task Force on Alzheimer’s Disease. Neurology 1984;34:939-44.
Kazee AM, Eskin TA, Lapham LW, Gabriel KR, McDaniel KD, Hamill RW. Clinicopathologic correlates in
Alzheimer disease: assessment of clinical and pathologic diagnostic criteria. Alzheimer Dis Assoc Disord
1993;7:152-64.
Varma AR, Snowden JS, Lloyd JJ, Talbot PR, Mann DM, Neary D. Evaluation of the NINCDS-ADRDA criteria in
the differentiation of Alzheimer’s disease and frontotemporal dementia. J Neurol Neurosurg Psychiatry
1999;66:184-8.
de Jong D, Kremer BP, Olde Rikkert MG, Verbeek MM. Current state and future directions of neurochemical
biomarkers for Alzheimer’s disease. Clin Chem Lab Med 2007;45:1421-34.
Blennow K. Cerebrospinal fluid protein biomarkers for Alzheimer’s disease. NeuroRx 2004;1:213-25.
Otto M, Lewczuk P, Wiltfang J. Neurochemical approaches of cerebrospinal fluid diagnostics in neurodegenerative diseases. Methods 2008;44:289-98.
Sjögren M, Andreasen N, Blennow K. Advances in the detection of Alzheimer’s disease-use of cerebrospinal
fluid biomarkers. Clin Chim Acta 2003;332:1-10.
Wiltfang J, Esselmann H, Bibl M, Hull M, Hampel H, Kessler H, Frolich L, Schroder J, Peters O, Jessen F,
Luckhaus C, Perneczky R, Jahn H, Fiszer M, Maler JM, Zimmermann R, Bruckmoser R, Kornhuber J, Lewczuk
P. Amyloid beta peptide ratio 42/40 but not A beta 42 correlates with phospho-Tau in patients with low- and
high-CSF A beta 40 load. J Neurochem 2007;101:1053-9.
Gravina SA, Ho L, Eckman CB, Long KE, Otvos L, Jr., Younkin LH, Suzuki N, Younkin SG. Amyloid beta protein
(A beta) in Alzheimer’s disease brain. Biochemical and immunocytochemical analysis with antibodies
specific for forms ending at A beta 40 or A beta 42(43). J Biol Chem 1995;270:7013-6.
Lewczuk P, Wiltfang J. Neurochemical dementia diagnostics: State of the art and research perspectives.
Proteomics 2008;8:1292-301.
Wiltfang J, Esselmann H, Bibl M, Smirnov A, Otto M, Paul S, Schmidt B, Klafki HW, Maler M, Dyrks T, Bienert M,
Beyermann M, Ruther E, Kornhuber J. Highly conserved and disease-specific patterns of carboxyterminally
truncated Abeta peptides 1-37/38/39 in addition to 1-40/42 in Alzheimer’s disease and in patients with
chronic neuroinflammation. J Neurochem 2002;81:481-96.
Bibl M, Mollenhauer B, Esselmann H, Lewczuk P, Klafki HW, Sparbier K, Smirnov A, Cepek L, Trenkwalder C,
Ruther E, Kornhuber J, Otto M, Wiltfang J. CSF amyloid-beta-peptides in Alzheimer’s disease, dementia with
Lewy bodies and Parkinson’s disease dementia. Brain 2006;129:1177-87.
Verbeek MM, de Jong D, Kremer HP. Brain-specific proteins in cerebrospinal fluid for the diagnosis of neurodegenerative diseases. Ann Clin Biochem 2003;40:25-40.
Hansson O, Zetterberg H, Buchhave P, Andreasson U, Londos E, Minthon L, Blennow K. Prediction of
Alzheimer’s disease using the CSF Abeta42/Abeta40 ratio in patients with mild cognitive impairment.
Dement Geriatr Cogn Disord 2007;23:316-20.
Kanai M, Matsubara E, Isoe K, Urakami K, Nakashima K, Arai H, Sasaki H, Abe K, Iwatsubo T, Kosaka T, Watanabe
M, Tomidokoro Y, Shizuka M, Mizushima K, Nakamura T, Igeta Y, Ikeda Y, Amari M, Kawarabayashi T, Ishiguro
K, Harigaya Y, Wakabayashi K, Okamoto K, Hirai S, Shoji M. Longitudinal study of cerebrospinal fluid levels of
tau, A beta1-40, and A beta1-42(43) in Alzheimer’s disease: a study in Japan. Ann Neurol 1998;44:17-26.
Mehta PD, Pirttila T, Mehta SP, Sersen EA, Aisen PS, Wisniewski HM. Plasma and cerebrospinal fluid levels of
amyloid beta proteins 1-40 and 1-42 in Alzheimer disease. Arch Neurol 2000;57:100-5.
Shoji M, Matsubara E, Kanai M, Watanabe M, Nakamura T, Tomidokoro Y, Shizuka M, Wakabayashi K, Igeta Y,
Ikeda Y, Mizushima K, Amari M, Ishiguro K, Kawarabayashi T, Harigaya Y, Okamoto K, Hirai S. Combination
assay of CSF tau, A beta 1-40 and A beta 1-42(43) as a biochemical marker of Alzheimer’s disease. J Neurol
Sci 1998;158:134-40.
Lewczuk P, Esselmann H, Otto M, Maler JM, Henkel AW, Henkel MK, Eikenberg O, Antz C, Krause WR, Reulbach
U, Kornhuber J, Wiltfang J. Neurochemical diagnosis of Alzheimer’s dementia by CSF Abeta42, Abeta42/
Abeta40 ratio and total tau. Neurobiol Aging 2004;25:273-81.
51
Chapter TWO
19. Giedraitis V, Sundelof J, Irizarry MC, Garevik N, Hyman BT, Wahlund LO, Ingelsson M, Lannfelt L. The normal
equilibrium between CSF and plasma amyloid beta levels is disrupted in Alzheimer’s disease. Neurosci Lett
2007;427:127-31.
20. Schoonenboom NS, Mulder C, Van Kamp GJ, Mehta SP, Scheltens P, Blankenstein MA, Mehta PD. Amyloid
beta 38, 40, and 42 species in cerebrospinal fluid: more of the same? Ann Neurol 2005;58:139-42.
21. Pijnenburg YA, Schoonenboom SN, Mehta PD, Mehta SP, Mulder C, Veerhuis R, Blankenstein MA, Scheltens
P. Decreased cerebrospinal fluid amyloid beta (1-40) levels in frontotemporal lobar degeneration. J Neurol
Neurosurg Psychiatry 2007;78:735-7.
22. Aarsland D, Kurz M, Beyer M, Bronnick K, Piepenstock NS, Ballard C. Early discriminatory diagnosis of
dementia with Lewy bodies. The emerging role of CSF and imaging biomarkers. Dement Geriatr Cogn
Disord 2008;25:195-205.
23. Bibl M, Mollenhauer B, Esselmann H, Lewczuk P, Trenkwalder C, Brechlin P, Ruther E, Kornhuber J, Otto M,
Wiltfang J. CSF diagnosis of Alzheimer’s disease and dementia with Lewy bodies. J Neural Transm
2006;113:1771-8.
24. McKeith IG, Galasko D, Kosaka K, Perry EK, Dickson DW, Hansen LA, Salmon DP, Lowe J, Mirra SS, Byrne EJ,
Lennox G, Quinn NP, Edwardson JA, Ince PG, Bergeron C, Burns A, Miller BL, Lovestone S, Collerton D, Jansen
EN, Ballard C, de Vos RA, Wilcock GK, Jellinger KA, Perry RH. Consensus guidelines for the clinical and
pathologic diagnosis of dementia with Lewy bodies (DLB): report of the consortium on DLB international
workshop. Neurology 1996;47:1113-24.
25. Neary D, Snowden JS, Gustafson L, Passant U, Stuss D, Black S, Freedman M, Kertesz A, Robert PH, Albert M,
Boone K, Miller BL, Cummings J, Benson DF. Frontotemporal lobar degeneration: a consensus on clinical
diagnostic criteria. Neurology 1998;51:1546-54.
26. Roman GC, Tatemichi TK, Erkinjuntti T, Cummings JL, Masdeu JC, Garcia JH, Amaducci L, Orgogozo JM, Brun
A, Hofman A, . Vascular dementia: diagnostic criteria for research studies. Report of the NINDS-AIREN
International Workshop. Neurology 1993;43:250-60.
27. Reijn TS, Olde Rikkert MG, van Geel WJ, de Jong D, Verbeek MM. Diagnostic accuracy of ELISA and xMAP
technology for analysis of amyloid beta(42) and tau proteins. Clin Chem 2007;53:859-65.
28. Bibl M, Mollenhauer B, Esselmann H, Schneider M, Lewczuk P, Welge V, Gross M, Falkai P, Kornhuber J,
Wiltfang J. Cerebrospinal fluid neurochemical phenotypes in vascular dementias: original data and
mini-review. Dement Geriatr Cogn Disord 2008;25:256-65.
29. Bennett SA, Pappas BA, Stevens WD, Davidson CM, Fortin T, Chen J. Cleavage of amyloid precursor protein
elicited by chronic cerebral hypoperfusion. Neurobiol Aging 2000;21:207-14.
30. Li L, Zhang X, Yang D, Luo G, Chen S, Le W. Hypoxia increases Abeta generation by altering beta- and
gamma-cleavage of APP. Neurobiol Aging 2007.
31. Qi JP, Wu H, Yang Y, Wang DD, Chen YX, Gu YH, Liu T. Cerebral ischemia and Alzheimer’s disease: the
expression of amyloid-beta and apolipoprotein E in human hippocampus. J Alzheimers Dis 2007;12:335-41.
32. Kumar-Singh S. Cerebral amyloid angiopathy: pathogenetic mechanisms and link to dense amyloid plaques.
Genes Brain Behav 2008;7 Suppl 1:67-82.
33. Kallhoff V, Peethumnongsin E, Zheng H. Lack of alpha-synuclein increases amyloid plaque accumulation in
a transgenic mouse model of Alzheimer’s disease. Mol Neurodegener 2007;2:6.
34. Mandal PK, Pettegrew JW, Masliah E, Hamilton RL, Mandal R. Interaction between Abeta peptide and alpha
synuclein: molecular mechanisms in overlapping pathology of Alzheimer’s and Parkinson’s in dementia
with Lewy body disease. Neurochem Res 2006;31:1153-62.
35. Lippa SM, Lippa CF, Mori H. Alpha-Synuclein aggregation in pathological aging and Alzheimer’s disease: the
impact of beta-amyloid plaque level. Am J Alzheimers Dis Other Demen 2005;20:315-8.
36. Pletnikova O, West N, Lee MK, Rudow GL, Skolasky RL, Dawson TM, Marsh L, Troncoso JC. Abeta deposition
is associated with enhanced cortical alpha-synuclein lesions in Lewy body diseases. Neurobiol Aging
2005;26:1183-92.
37. Bibl M, Mollenhauer B, Wolf S, Esselmann H, Lewczuk P, Kornhuber J, Wiltfang J. Reduced CSF carboxyterminally truncated Abeta peptides in frontotemporal lobe degenerations. J Neural Transm 2007;114:621-8.
38. Consensus report of the Working Group on: “Molecular and Biochemical Markers of Alzheimer’s Disease”.
The Ronald and Nancy Reagan Research Institute of the Alzheimer’s Association and the National Institute
on Aging Working Group. Neurobiol Aging 1998;19:109-16.
52
The CSF Aβ42/Aβ40 ratio
39. Andreasen N, Minthon L, Davidsson P, Vanmechelen E, Vanderstichele H, Winblad B, Blennow K. Evaluation
of CSF-tau and CSF-Abeta42 as diagnostic markers for Alzheimer disease in clinical practice. Arch Neurol
2001;58:373-9.
40. Riemenschneider M, Wagenpfeil S, Diehl J, Lautenschlager N, Theml T, Heldmann B, Drzezga A, Jahn T, Forstl
H, Kurz A. Tau and Abeta42 protein in CSF of patients with frontotemporal degeneration. Neurology
2002;58:1622-8.
53
CSF α-synuclein
This chapter has been published as: Cerebrospinal fluid α-synuclein does not discriminate between dementia
disorders. Petra E. Spies, René J.F. Melis, Magnus J.C. Sjögren, Marcel G.M. Olde Rikkert, Marcel M. Verbeek,
J Alzheimers Dis 2009;16:363-9
Chapter
Three
Chapter Three
Chapter Three
Chapter Three
Abstract
Alpha-synuclein is the major constituent of Lewy bodies found in neurons in dementia
with Lewy bodies (DLB) and might be of diagnostic value as a biomarker for DLB. We
hypothesized that, as a consequence of increased accumulation of α-synuclein intraneuronally in DLB, the levels of α-synuclein in cerebrospinal fluid (CSF) of DLB patients
would be lower than in other dementias. Our objective was to investigate the CSF levels
of α-synuclein in several dementia disorders compared to control levels and to
investigate the diagnostic value of CSF α-synuclein as a marker to discriminate between
DLB and other types of dementia. We analysed the levels of α-synuclein in CSF of
40 DLB patients, 131 patients with Alzheimer disease (AD), 28 patients with vascular
dementia (VaD), and 39 patients with frontotemporal dementia (FTD). We did not find
any significant differences in CSF α-synuclein levels between DLB patients and patients
with AD, VaD or FTD. We conclude that in clinically diagnosed patients, α-synuclein
does not appear to be a useful biomarker for the differentiation between DLB and other
types of dementia.
56
CSF α-synuclein
Introduction
Dementia with Lewy bodies (DLB) is the second most common type of dementia,1-3
accounting for 15-20% of cases at autopsy.4 It is difficult to differentiate between DLB and
other dementias such as Alzheimer disease (AD) using clinical examination.5 Although
specificity has improved with the introduction of the consensus guidelines for DLB in
1996,1,6,7 sensitivity is low,1,3 resulting in many misclassified cases. Differentiation from
other dementia syndromes is difficult since DLB patients can present with symptoms that
resemble those of other dementias. Extrapyramidal signs and hallucinations, which are
characteristic for DLB patients, may also occur in advanced AD. On neuropsychological
assessment, most DLB patients present with language and memory deficits similar to
those with AD.8 In one study, most patients misdiagnosed with DLB had AD pathology at
autopsy.5 Accurately diagnosing DLB is important for its pharmacological management.
Patient with DLB respond well to cholinesterase inhibitors, but are extremely sensitive to
the side effects of neuroleptic drugs.3,4 Furthermore, prognosis of patients with DLB may
be different, with faster deterioration and shorter time between onset of cognitive
symptoms and death compared to patients with AD.1,5,9
Analysis of the cerebrospinal fluid (CSF) has been increasingly applied to differentiate
the various dementia disorders. Especially in AD, a typical CSF pattern is often observed,
with increased concentrations of total tau protein (t-tau) and phosphorylated tau
protein (p-tau) and decreased concentrations of the amyloid β42 (Aβ 42) protein.10,11 In
contrast, there are no generally accepted biomarkers to distinguish DLB from other
types of dementia. Some studies found equally high CSF t-tau concentrations in AD and
DLB.12,13 In most studies, however, CSF t-tau concentrations were normal to moderately
elevated in DLB,2,8,12,14-17 with normal concentrations of p-tau.18,19 In contrast, Aβ 42
concentrations were decreased to a similar degree as in AD,2,8,14,16 resulting in a low
discriminative value of Aβ 42 / t-tau analysis to differentiate AD from DLB.
Little research has been done with respect to CSF α-synuclein as a diagnostic
marker. α-Synuclein is the major constituent of Lewy bodies, which are characteristic
cytoplasmic inclusions in cortical neurons of DLB patients. Initially α-synuclein was
thought to be an intracellular protein, but recently it was found in CSF and plasma.20
Tokuda et al found that patients with Parkinson disease (PD) had significantly lower
α-synuclein levels in their CSF compared to controls.21 This might be explained by
accumulation of α-synuclein within affected neurons.21 We hypothesized that levels of
α-synuclein in synucleinopathies such as DLB would be decreased in comparison with
dementias not characterized by α-synuclein inclusions, such as AD vascular dementia
57
Chapter Three
(VaD) and frontotemporal dementia (FTD) and with control levels. Therefore, we
investigated the CSF levels of α-synuclein in these dementia disorders compared to control
levels, and studied if CSF α-synuclein may serve as a biomarker for the ­differentiation of
DLB from other dementias.
Materials and Methods
Patients
We screened 526 patients of the CSF database of the Alzheimer Centre of the Radboud
University Nijmegen Medical Centre for inclusion in this study (Figure 1). This database
consists of clinical data and biobank materials of consecutive patients, in whom informed
consent for lumbar puncture was obtained. We excluded 69 patients because CSF was
not available for α-synuclein analysis. Patients whose diagnosis was uncertain or with a
diagnosis other than DLB, AD, VaD or FTD (e.g. mild cognitive impairment) (n=98) were not
included either, as were patients with mixed dementia (n=9). In total, we included 40 DLB,
131 AD, 39 FTD and 28 VaD patients whose diagnosis was established by a multi­disciplinary
panel, which consisted of a geriatrician, neurologist, neuropsychologist, and – if needed
– an old-age psychiatrist. The panel used the accepted clinical diagnostic criteria, i.e. the
1996 consensus guidelines for the clinical and pathologic diagnosis of dementia with
Lewy bodies, the NINCDS-ADRDA criteria for AD, the 1998 consensus on clinical diagnostic
criteria for frontotemporal lobar degeneration, and the NINDS-AIREN criteria for VaD.6,22-24
If possible, information on MMSE was obtained and Aβ42, p-tau and t-tau were analysed.
The panel was aware of the CSF results for Aβ42, p-tau and t-tau, but not for α-synuclein,
when establishing the clinical diagnosis.
We compared CSF α-synuclein levels of the dementia patients with those of 57
control persons aged >50 years (control group A). Data on Aβ 42, p-tau and t-tau was not
available for this group. In order to have reference levels of Aβ 42, p-tau and t-tau levels
to allow comparison of these variables, we included another 55 control persons (control
group B). Both control groups underwent lumbar puncture for various reasons. However,
for inclusion, intensive diagnostic investigations was not allowed to reveal any
neurological disorder. Moreover, the CSF findings of both control groups for cell count,
glucose, lactate, haemoglobin, bilirubin, total protein and oligoclonal IgG bands had to
be normal.
A total of 238 patients with dementia and 112 controls were included (Figure 1).
Patient characteristics are listed in Table 1. Characteristics of controls are listed in Table 2.
58
CSF α-synuclein
Figure 1 O
verview of the patients included in this study
Database
N = 526
Excluded patients:
- 69 CSF not available for
α-synuclein analysis
- 98 insecure/other diagnosis
- 9 mixed dementia
Controls
N = 112
Dementia
N = 238
AD
N = 131
DLB
N = 40
FTD
N = 39
VaD
N = 28
Control group A
N = 57
Control group B
N = 55
AD, Alzheimer disease; DLB, dementia with Lewy bodies; FTD, frontotemporal dementia; VaD,
vascular dementia.
Mean age (± SD) of patients and controls together was 68 (± 10) years and 51% was
male. AD patients were more often female. FTD patients were significantly younger
than patients in the other three dementia groups.
CSF
CSF samples were collected in polypropylene tubes by lumbar puncture and transported
to the laboratory within 30 minutes after withdrawal. The CSF was centrifuged and
aliquots of each sample were immediately frozen at -80 °C until analysis. α-Synuclein in
CSF was measured using a sandwich Elisa system based on a previously described
procedure.25 A disposable flat-bottom microtiterplate (Nunc Maxisorp F96, Roskilde,
Denmark) was coated with 100 µl antibody 211 (0.2 µg/ml in 0.20 M carbonate buffer, pH
9.6) overnight at 4°C. A plate washer (BioTek, Beun de Ronde, Abcoude, The Netherlands)
was used to wash the plate five times with 250 µl phosphate buffered saline containing
0.05% Tween-20 (PBS washing buffer). All further incubations were performed at 37°C,
unless stated otherwise, and all measurements were performed in duplicate. 250 µl of
blocking buffer (2.5% gelatin in PBS washing buffer) was added and incubated for 2
hours. The plate was subsequently washed five times with PBS washing buffer. Next, 100
µl α-synuclein solution (from 0 to 500 ng/ml diluted in PBS) or CSF was added to each
59
Chapter Three
Table 1 Clinical characteristics and results of CSF analysis of dementia patients
DLB
n=40
AD
n=131
FTD
n=39
VaD
n=28
p-value*
No. of males/females 27/13
54/77
24/15
18/10
0.005‡
Age (years)
74.0 ± 8.3
71.7 ± 8.8
66.4 ± 8.6
75.4 ± 8.4
<0.0001
MMSE score
18.9 ± 8.2
n=7
19.6 ± 3.9
n=71
20.3 ± 6.1
n=16
18.5 ± 2.9
n=6
0.843
Amyloid β42 (pg/ml)
533.0 ± 231.0 451.3 ± 147.2 702.1 ± 257.8 650.1 ± 244.6 <0.0001
n=33
n=110
n=33
n=24
T-tau (pg/ml)
318.7 ± 161.6 716.4 ± 408.9 407.3 ± 285.9 400.1 ± 543.1 <0.0001
n=33
n=109
n=33
n=23
P-tau (pg/ml)
56.9 ± 22.2
n=31
108.7 ± 47.0
n=109
68.5 ± 34.6
n=33
52.9 ± 18.6
n=23
<0.0001
α-synuclein (ng/ml)
38.0 ± 29.0
37.0 ± 36.1
49.1 ± 41.5
44.3 ± 40.3
0.187
Values are expressed as means ± SD. *, p-value by ANOVA; ‡, p-value by chi square test.
DLB, dementia with Lewy bodies; AD, Alzheimer disease; FTD, frontotemporal dementia; VaD, vascular
dementia; MMSE, mini mental state examination (0-30, with higher scores indicating better cognition);
t-tau, total tau; p-tau, phosphorylated tau.
Table 2 Clinical characteristics and results of CSF analysis of controls
Control group A
n=57
Control group B
n=55
No. of males/females
30/27
26/29
Age (years)
61.3 ± 8.8
61.1 ± 8.9
Amyloid β42 (pg/ml)
n.a.
822.9 ± 256.3
T-tau (pgm/l)
n.a.
212.4 ± 81.6
P-tau (pg/ml)
n.a.
47.5 ± 15.1
α-synuclein (ng/ml)
30.4 ± 19.1
n.a.
Values are expressed as means ± SD.
T-tau, total tau; p-tau, phosphorylated tau; n.a., not available.
60
CSF α-synuclein
well and incubated for 2.5 hours. After washing the plate five times with PBS washing
buffer, 100 µl of antibody FL-140, diluted 1:1,000 in blocking buffer, was added for 1.5
hours. The plate was washed five times and 100 µl of the peroxidase labeled goat-­
anti-rabbit antibody (dilution 1:5,000 in blocking buffer) was added and incubated for
1 hour. After a final washing step 100 µl of a freshly prepared solution of tetramethyl
benzidine (TMB) was applied and incubated for 15 minutes in the dark at room
temperature. The reaction was stopped with 50 µl 2N H2SO4 and the absorbance was
measured at 450 nm in an ELISA plate reader (Tecan Sunrise, Salzburg, Austria). The
technicians performing the analyses did not have access to the clinical data.
Statistical analysis
Differences in α-synuclein, age, MMSE score, Aβ42, p-tau, and t-tau between dementia
groups were analysed by ANOVA. Levels of α-synuclein, Aβ 42, p-tau, and t-tau did not
follow a normal distribution. We used log transformation to analyse these variables.
Differences in gender distribution among the dementia groups were determined by the
chi square test. The associations between α-synuclein and age, MMSE score, Aβ 42, p-tau
and t-tau were examined by linear regression. Statistical analyses were carried out using
SAS, version 8.2.
Results
The mean CSF level of α-synuclein was not significantly different in the four dementia
groups; nor did we find a significant difference between the dementia groups and the
control group A. We found no association of CSF α-synuclein with MMSE score, sex,
Aβ 42 or p-tau among the dementia patients. The level of α-synuclein decreased with
increasing t-tau (B -0.17, p 0.030) and age (B -0.02, p 0.001) among the dementia cases
(Figure 2). Taking these variables into account did not change our initial findings about
the absence of an association between dementia type and α-synuclein.
In comparison with the univariable associations with α-synuclein, the association
with age remained stable and significant (B -0.02, p 0.002), while the association with
t-tau diminished and lost significance (B -0.14, p 0.113) in a linear regression model that
included both variables as well as dementia type.
In DLB patients, the concentration of Aβ42 was significantly lower than in the control
group B, although still above the cut-off value of 500 pg/ml for normal values. P-tau and
t-tau concentrations were in the range of normal although significantly increased
61
Chapter Three
Figure 2 Scatter plot of CSF α-synuclein versus age in the dementia groups
compared to these controls. For AD patients, Aβ 42 concentration was significantly
decreased, while p-tau and t-tau concentrations were significantly higher than in the
control group B.
Discussion
We investigated CSF α-synuclein in patients with different types of dementia. To our
knowledge, this is one of the first studies comparing α-synuclein levels in DLB to other
dementias. Since we did not observe differences in the CSF levels of α-synuclein
between the various clinically diagnosed dementia disorders, α-synuclein is not a useful
biomarker for the differentiation between DLB, AD, FTD or VaD.
What may have accounted for the lack of difference in α-synuclein levels that we
observed between the different types of dementia, is that α-synucleinopathies, such as
DLB, and tauopathies, such as AD and part of the FTD cases, may be disorders with
considerable overlap, instead of distinct entities.26,27 Neuropathological findings of the
various disorders have been reported to be overlapping. The presence of Lewy bodies
in brains of AD patients was higher than expected.26 More than 60% of cases harboured
62
CSF α-synuclein
α-synuclein positive Lewy bodies,28 whereas 87% of brains with a neuropathological
diagnosis of DLB also met CERAD criteria for definite or probable AD. 26 When Iqbal et al
divided AD patients into subgroups based on their CSF biomarker profile, a subgroup
with low Aβ 42 and only a slight increase in t-tau levels consisted mainly of cases with
AD of Lewy body type, their CSF biomarker profile comparable to our DLB group.29
This supports the hypothesis of overlapping etiology of α-synucleinopathies and
tauopathies. Interestingly, Lewy body pathology has also been found in brain tissue of
elderly individuals without cognitive impairment.26,27
Another explanation for the lack of difference in α-synuclein levels might be
heterogeneity within the DLB group. Possibly, there are subgroups within the DLB
group. For example, patients who display parkinsonism early in the disease could have
lower levels of α-synuclein than patients whose parkinsonism develops later in the
course of the disease. We are uncertain about such differences within our DLB group.
Studies addressing the association of α-synuclein in CSF with PD report contradictory
findings. Tokuda et al found lower CSF α-synuclein in PD patients compared to controls,21
which led to the assumption that α-synuclein could be used as a marker for
α-synucleinopathies. However, others did not find significant differences between PD
patients and controls.20,30 In Lewy bodies, α-synuclein is aggregated and insoluble.3
However, under normal conditions α-synuclein is a soluble synaptic protein, that can
also be found in human CSF and blood plasma of healthy humans,3,16,20 and that may
be secreted into the CSF by neurons.20,21,31 Possibly, the CSF α-synuclein levels that we
and others measured reflect a physiological level of soluble α-synuclein, independent
of α-synuclein accumulation. The levels of α-synuclein that we measured were in
the same order of magnitude as found in controls by Tokuda et al, using the same
antibodies.21
We found that age had a weak influence on the levels of α-synuclein, with
α-synuclein decreasing with age. This is consistent with the study by Tokuda et al.21 With
aging, axonal transport of α-synuclein slows down and the amount of soluble
α-synuclein decreases.27 This may account for the decreasing levels of α-synuclein found
in CSF. Our initial finding that the concentration of t-tau increased with decreasing
α-synuclein lost significance after correcting for age and dementia type.
Our study population is representative for the population seen at a memory clinic
and the CSF data indicate that our study group was very well comparable to the study
groups used by other research centres. We observed the typical pattern of decreased
Aβ 42 and increased t-tau /p-tau levels in AD patients.8,10 In DLB we observed comparable
decreased Aβ 42 concentrations, although not as low as in AD, and normal to slightly
63
Chapter Three
increased t-tau and p-tau concentrations, as has been reported before.2,12,14-19 Also, our
observations in the VaD and FTD groups are in accord with previous studies.8,17,32-36
Although the clinical diagnoses in our study populations were made according to
accepted clinical criteria, some misclassification of DLB patients as AD patients probably
has occurred. The presence of AD pathology in DLB modifies its clinical presentation
with a lower rate of visual hallucinations and parkinsonism.3 This contributes to the
complicated clinical differentiation between these two types of dementia and once
more illustrates the urgent need for a discriminative biomarker. Our findings suggest
that α-synuclein cannot be used as such a diagnostic marker. Future studies, e.g. with
post-mortem verification of the clinical diagnosis, are warranted to confirm our findings,
but based on the findings of our study, α-synuclein cannot be advocated as a biomarker
for DLB.
64
CSF α-synuclein
References
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. Aarsland D, Kurz M, Beyer M, Bronnick K, Piepenstock NS, Ballard C. Early discriminatory diagnosis of dementia with
Lewy bodies. The emerging role of CSF and imaging biomarkers. Dement Geriatr Cogn Disord 2008;25:195-205.
Kanemaru K, Kameda N, Yamanouchi H. Decreased CSF amyloid beta42 and normal tau levels in dementia
with Lewy bodies. Neurology 2000;54:1875-6.
McKeith I, Mintzer J, Aarsland D, Burn D, Chiu H, Cohen-Mansfield J, Dickson D, Dubois B, Duda JE, Feldman
H, Gauthier S, Halliday G, Lawlor B, Lippa C, Lopez OL, Carlos MJ, O’Brien J, Playfer J, Reid W. Dementia with
Lewy bodies. Lancet Neurol 2004;3:19-28.
McKeith IG, Ballard CG, Perry RH, Ince PG, O’Brien JT, Neill D, Lowery K, Jaros E, Barber R, Thompson P, Swann
A, Fairbairn AF, Perry EK. Prospective validation of consensus criteria for the diagnosis of dementia with
Lewy bodies. Neurology 2000;54:1050-8.
Hohl U, Tiraboschi P, Hansen LA, Thal LJ, Corey-Bloom J. Diagnostic accuracy of dementia with Lewy bodies.
Arch Neurol 2000;57:347-51.
McKeith IG, Galasko D, Kosaka K, Perry EK, Dickson DW, Hansen LA, Salmon DP, Lowe J, Mirra SS, Byrne EJ,
Lennox G, Quinn NP, Edwardson JA, Ince PG, Bergeron C, Burns A, Miller BL, Lovestone S, Collerton D, Jansen
EN, Ballard C, de Vos RA, Wilcock GK, Jellinger KA, Perry RH. Consensus guidelines for the clinical and
pathologic diagnosis of dementia with Lewy bodies (DLB): report of the consortium on DLB international
workshop. Neurology 1996;47:1113-24.
McKeith IG, Dickson DW, Lowe J, Emre M, O’Brien JT, Feldman H, Cummings J, Duda JE, Lippa C, Perry EK,
Aarsland D, Arai H, Ballard CG, Boeve B, Burn DJ, Costa D, Del ST, Dubois B, Galasko D, Gauthier S, Goetz CG,
Gomez-Tortosa E, Halliday G, Hansen LA, Hardy J, Iwatsubo T, Kalaria RN, Kaufer D, Kenny RA, Korczyn A,
Kosaka K, Lee VM, Lees A, Litvan I, Londos E, Lopez OL, Minoshima S, Mizuno Y, Molina JA, Mukaetova-Ladinska EB, Pasquier F, Perry RH, Schulz JB, Trojanowski JQ, Yamada M. Diagnosis and management of dementia
with Lewy bodies: third report of the DLB Consortium. Neurology 2005;65:1863-72.
Otto M, Lewczuk P, Wiltfang J. Neurochemical approaches of cerebrospinal fluid diagnostics in neurodegenerative diseases. Methods 2008;44:289-98.
Olichney JM, Galasko D, Salmon DP, Hofstetter CR, Hansen LA, Katzman R, Thal LJ. Cognitive decline is faster
in Lewy body variant than in Alzheimer’s disease. Neurology 1998;51:351-7.
de Jong D, Kremer BP, Olde Rikkert MG, Verbeek MM. Current state and future directions of neurochemical
biomarkers for Alzheimer’s disease. Clin Chem Lab Med 2007;45:1421-34.
Sunderland T, Linker G, Mirza N, Putnam KT, Friedman DL, Kimmel LH, Bergeson J, Manetti GJ, Zimmermann
M, Tang B, Bartko JJ, Cohen RM. Decreased beta-amyloid1-42 and increased tau levels in cerebrospinal fluid
of patients with Alzheimer disease. JAMA 2003;289:2094-103.
Blennow K, Vanmechelen E, Hampel H. CSF total tau, Abeta42 and phosphorylated tau protein as biomarkers
for Alzheimer’s disease. Mol Neurobiol 2001;24:87-97.
Tschampa HJ, Schulz-Schaeffer W, Wiltfang J, Poser S, Otto M, Neumann M, Kretzschmar HA. Decreased CSF
amyloid beta42 and normal tau levels in dementia with Lewy bodies. Neurology 2001;56:576.
Andreasen N, Minthon L, Davidsson P, Vanmechelen E, Vanderstichele H, Winblad B, Blennow K. Evaluation
of CSF-tau and CSF-Abeta42 as diagnostic markers for Alzheimer disease in clinical practice. Arch Neurol
2001;58:373-9.
Gomez-Tortosa E, Gonzalo I, Fanjul S, Sainz MJ, Cantarero S, Cemillan C, Yebenes JG, Del ST. Cerebrospinal
fluid markers in dementia with lewy bodies compared with Alzheimer disease. Arch Neurol 2003;60:1218-22.
Mollenhauer B, Bibl M, Wiltfang J, Steinacker P, Ciesielczyk B, Neubert K, Trenkwalder C, Otto M. Total tau
protein, phosphorylated tau (181p) protein, beta-amyloid(1-42), and beta-amyloid(1-40) in cerebrospinal
fluid of patients with dementia with Lewy bodies. Clin Chem Lab Med 2006;44:192-5.
Vanmechelen E, Vanderstichele H, Hulstaert F, Andreasen N, Minthon L, Winblad B, Davidsson P, Blennow K.
Cerebrospinal fluid tau and beta-amyloid(1-42) in dementia disorders. Mech Ageing Dev 2001;122:2005-11.
Buerger K, Zinkowski R, Teipel SJ, Tapiola T, Arai H, Blennow K, Andreasen N, Hofmann-Kiefer K, DeBernardis
J, Kerkman D, McCulloch C, Kohnken R, Padberg F, Pirttila T, Schapiro MB, Rapoport SI, Moller HJ, Davies P,
Hampel H. Differential diagnosis of Alzheimer disease with cerebrospinal fluid levels of tau protein
phosphorylated at threonine 231. Arch Neurol 2002;59:1267-72.
65
Chapter Three
19. Hampel H, Buerger K, Zinkowski R, Teipel SJ, Goernitz A, Andreasen N, Sjoegren M, DeBernardis J, Kerkman
D, Ishiguro K, Ohno H, Vanmechelen E, Vanderstichele H, McCulloch C, Moller HJ, Davies P, Blennow K.
Measurement of phosphorylated tau epitopes in the differential diagnosis of Alzheimer disease: a
comparative cerebrospinal fluid study. Arch Gen Psychiatry 2004;61:95-102.
20. El-Agnaf OM, Salem SA, Paleologou KE, Cooper LJ, Fullwood NJ, Gibson MJ, Curran MD, Court JA, Mann DM,
Ikeda S, Cookson MR, Hardy J, Allsop D. Alpha-synuclein implicated in Parkinson’s disease is present in
extracellular biological fluids, including human plasma. FASEB J 2003;17:1945-7.
21. Tokuda T, Salem SA, Allsop D, Mizuno T, Nakagawa M, Qureshi MM, Locascio JJ, Schlossmacher MG, El-Agnaf
OM. Decreased alpha-synuclein in cerebrospinal fluid of aged individuals and subjects with Parkinson’s
disease. Biochem Biophys Res Commun 2006;349:162-6.
22. McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stadlan EM. Clinical diagnosis of Alzheimer’s
disease: report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human
Services Task Force on Alzheimer’s Disease. Neurology 1984;34:939-44.
23. Neary D, Snowden JS, Gustafson L, Passant U, Stuss D, Black S, Freedman M, Kertesz A, Robert PH, Albert M,
Boone K, Miller BL, Cummings J, Benson DF. Frontotemporal lobar degeneration: a consensus on clinical
diagnostic criteria. Neurology 1998;51:1546-54.
24. Roman GC, Tatemichi TK, Erkinjuntti T, Cummings JL, Masdeu JC, Garcia JH, Amaducci L, Orgogozo JM, Brun
A, Hofman A, . Vascular dementia: diagnostic criteria for research studies. Report of the NINDS-AIREN
International Workshop. Neurology 1993;43:250-60.
25. van Geel WJ, Abdo WF, Melis R, Williams S, Bloem BR, Verbeek MM. A more efficient enzyme-linked
immunosorbent assay for measurement of alpha-synuclein in cerebrospinal fluid. J Neurosci Methods
2008;168:182-5.
26. Galpern WR, Lang AE. Interface between tauopathies and synucleinopathies: a tale of two proteins. Ann
Neurol 2006;59:449-58.
27. Mukaetova-Ladinska EB, McKeith IG. Pathophysiology of synuclein aggregation in Lewy body disease.
Mech Ageing Dev 2006;127:188-202.
28. Wirths O, Bayer TA. Alpha-synuclein, Abeta and Alzheimer’s disease. Prog Neuropsychopharmacol Biol
Psychiatry 2003;27:103-8.
29. Iqbal K, Flory M, Khatoon S, Soininen H, Pirttila T, Lehtovirta M, Alafuzoff I, Blennow K, Andreasen N,
Vanmechelen E, Grundke-Iqbal I. Subgroups of Alzheimer’s disease based on cerebrospinal fluid molecular
markers. Ann Neurol 2005;58:748-57.
30. Borghi R, Marchese R, Negro A, Marinelli L, Forloni G, Zaccheo D, Abbruzzese G, Tabaton M. Full length alpha-synuclein is present in cerebrospinal fluid from Parkinson’s disease and normal subjects. Neurosci Lett
2000;287:65-7.
31. Lee HJ, Patel S, Lee SJ. Intravesicular localization and exocytosis of alpha-synuclein and its aggregates. J
Neurosci 2005;25:6016-24.
32. Bian H, Van Swieten JC, Leight S, Massimo L, Wood E, Forman M, Moore P, de K, I, Clark CM, Rosso S,
Trojanowski J, Lee VM, Grossman M. CSF biomarkers in frontotemporal lobar degeneration with known
pathology. Neurology 2008;70:1827-35.
33. de Jong D, Jansen RW, Kremer BP, Verbeek MM. Cerebrospinal fluid amyloid beta42/phosphorylated tau
ratio discriminates between Alzheimer’s disease and vascular dementia. J Gerontol A Biol Sci Med Sci
2006;61:755-8.
34. Kapaki E, Paraskevas GP, Papageorgiou SG, Bonakis A, Kalfakis N, Zalonis I, Vassilopoulos D. Diagnostic value
of CSF biomarker profile in frontotemporal lobar degeneration. Alzheimer Dis Assoc Disord 2008;22:47-53.
35. Sjögren M, Minthon L, Davidsson P, Granerus A-K, Clarberg A, Vanderstichele H, Vanmechelen E, Wallin A,
Blennow K. CSF levels of tau, beta-amyloid(1-42) and GAP-43 in frontotemporal dementia, other types of
dementia and normal aging. J Neural Transm 2000;107:563-79.
36. Verbeek MM, Pijnenburg YA, Schoonenboom NS, Kremer BP, Scheltens P. Cerebrospinal fluid tau levels in
frontotemporal dementia. Ann Neurol 2005;58:656-7.
66
CSF α-synuclein
67
The correlation between
CSF α-synuclein and CSF Aβ
This chapter has been published as: CSF α-synuclein concentrations do not fluctuate over hours and are
not correlated to amyloid β in humans. Petra E. Spies, Diane Slats, Marcel G.M. Olde Rikkert, Jack Tseng,
Jurgen A.H.R. Claassen, Marcel M. Verbeek, Neurosci Lett 2011;504:336-8
Chapter
Four
Chapter Four
Chapter Four
Chapter Four
Abstract
Reports on the value of cerebrospinal fluid (CSF) α-synuclein as a biomarker for dementia
with Lewy bodies and Parkinson disease are contradicting. This may be explained by
fluctuating CSF α-synuclein concentrations over time. Such fluctuations have been
suggested for CSF amyloid β concentrations. Furthermore, a physiological relationship
between α-synuclein and amyloid β has been suggested based on in vitro research. We
performed repeated CSF sampling in healthy elderly and Alzheimer disease (AD)
patients and showed that sinusoidal fluctuations in CSF α-synuclein concentrations
were not present. Furthermore, we did not find evidence for an interaction between
amyloid β and α-synuclein concentrations in CSF.
70
The correlation between CSF α-synuclein and CSF Aβ
Alpha-synuclein is an important component of the intraneuronal Lewy bodies that
accumulate in Parkinson disease (PD) and dementia with Lewy bodies (DLB).
Cerebrospinal fluid (CSF) concentrations of α-synuclein have been advocated as a
biomarker for PD and DLB, although this suggestion may be challenged.1-4 In vitro
research has shown that α-synuclein increases cellular amyloid β (Aβ) secretion and that
Aβ may stimulate the aggregation of α-synuclein, suggesting a physiological relationship
between these proteins.5,6 It is therefore conceivable that the CSF concentrations of
these proteins correlate. Furthermore, recent data demonstrated sinusoidal fluctuations
in CSF Aβ 42 and Aβ 40 concentrations that may result from a circadian rhythm,7,8 which
implies that CSF α-synuclein concentrations may follow a circadian or sinusoidal rhythm
as well. Such fluctuations in CSF α-synuclein concentrations could explain the
contradicting results regarding the value of CSF α-synuclein as a biomarker for PD and
DLB. We aimed to investigate 1) the correlation between CSF α-synuclein and Aβ
concentrations in healthy elderly, Alzheimer disease (AD) patients and DLB patients; 2)
the occurrence of sinusoidal fluctuations in CSF α-synuclein concentrations.
To test for a correlation between CSF α-synuclein and Aβ42 or Aβ 40, we performed
a cross-sectional analysis in 164 persons (n=17 neurological controls, n=113 AD patients
and n=34 DLB patients) diagnosed in our memory clinic. CSF was obtained by lumbar
puncture. In this part of the study, CSF α-synuclein was measured using a sandwich
ELISA based on a previously described procedure.9 Aβ 42 was measured using ELISA
(Innogenetics NV, Ghent, Belgium) and Aβ 40 was analyzed using xMAP technology
(BioSource, Invitrogen Ltd, Paisley, UK). The biomarkers were not normally distributed
and were log transformed. Pearson’s correlation coefficient was calculated. We did not
observe a correlation between α-synuclein and Aβ 42 in the entire group (r=-0.089,
p=0.256) or in any of the subgroups (controls: r=0.070, p=0.789; AD: r=-0.069, p=0.468;
DLB: r=-0.195, p=0.268). Aβ 40 concentrations were known of 65 persons (n=6 controls,
n=45 AD patients, n=14 DLB patients). Again, we did not observe a correlation between
α-synuclein and Aβ40 in the entire group (r=0.082, p=0.515) or in the AD and DLB
subgroups (AD: r=0.206, p=0.174; DLB: r=-0.225, p=0.439). The control group was
considered too small to reliably perform a subgroup analysis.
To investigate fluctuations in CSF α-synuclein concentrations, we analysed CSF
samples of 6 healthy controls (59-85 years old) and 6 AD patients (64-77 years old) who
underwent repeated CSF sampling from an indwelling intrathecal catheter with 1 hour
intervals.8 The study was approved by the institutional review board and informed
consent was obtained from all participants. 26-33 samples were available per participant.
In this study, CSF α-synuclein concentrations were measured by aforementioned ELISA.9
71
Chapter Four
CSF Aβ 40 was analyzed by ELISA (the Genetics Company, Schlieren, Switzerland) and
CSF Aβ 42 was measured using xMAP-based technology (Innogenetics NV, Ghent,
Belgium). As previously reported, 8 the mean coefficients of variation for Aβ 42 and Aβ 40
were 11.1% and 10.9% for AD patients and 14.4% and 11.2% for healthy controls.
Sinusoidal fluctuations in Aβ 42 and Aβ 40 concentrations were observed, although much
less pronounced compared to observations in younger subjects.7 The continuous rise in
Aβ levels that has been reported in some studies10 was not identified in this study.8 In
the present study, the mean coefficient of variation for α-synuclein was 21.5% for AD
patients and 23.6% for healthy controls, while the intra-assay coefficient of variation was
3.5-14.9%, depending on the α-synuclein concentration.9 Time trends in α-synuclein
concentrations were analysed with linear mixed model analysis with subject as random
factor. To account for the longitudinal relationship of samples a first-order autoregressive
covariance structure was used. No linear trend in α-synuclein concentrations over 33
hours was observed. To investigate the presence of a sinusoidal pattern we included a
cosinar analysis in the model. This analysis did not identify sinusoidal variation in CSF
α-synuclein concentrations (Figure 1). These results did not change when AD patients
and healthy controls were analysed separately.
To detect a potential temporal delay in the interaction between CSF α-synuclein
and Aβ, which cannot be assessed by cross-sectional measurements, we also
investigated the correlation between α-synuclein and Aβ in the repeated CSF samples
of these persons (Figure 2). We examined Pearson’s correlation coefficients per individual
between α-synuclein and Aβ 42 at different time lags, by correlating the α-synuclein
concentration of each time point to the Aβ42 concentration 0 to 6 hours from that time
point. This time period was arbitrarily chosen as a period in which an effect should be
detectable.
We also performed these analyses for α-synuclein and Aβ40, for Aβ 42 and α-synuclein
and for Aβ 40 and α-synuclein. A significance level of 0.01 was used in order to maintain
sensitivity despite multiple comparisons. Twenty-five out of 336 tested correlations
were significant at this level, 23 in four AD patients and 2 in one healthy control.
Coefficients ranged between -0.66 and -0.46 except for 3 in one AD patient that ranged
between 0.54 and 0.66. These correlations were found between all combinations of
proteins and at all time lags and no consistent pattern in any combination of two
proteins or at any time lag could be identified. Therefore, these findings were interpreted
to result from chance.
In summary, this study demonstrated that CSF α-synuclein concentrations do not
show sinusoidal fluctuations over a 33-hours period. Therefore, sinusoidal fluctuations
72
The correlation between CSF α-synuclein and CSF Aβ
Figure 1 M
ean concentrations of CSF α-synuclein over 33 hours in AD patients
(n=6) and healthy elderly controls (n=6)
Closed circles: AD patients; open squares: healthy controls; bars represent 95% confidence interval.
Figure 2 M
ean concentrations of CSF α-synuclein, Aβ 42 and Aβ 40 over 33 hours
in 12 persons (n=6 AD patients and n=6 healthy elderly controls)
Closed triangles: Aβ 40; open triangles: Aβ 42; closed squares: α-synuclein; bars represent 95% confidence
interval.
73
Chapter Four
in the CSF α-synuclein concentrations are unlikely to be the explanation for the
conflicting reports regarding the value of CSF α-synuclein as biomarker for PD or DLB.1-4
Other causes that may explain the discrepancies between these reports may comprise
differences in assays and patient characteristics. Furthermore, this study revealed that
the interaction between Aβ and α-synuclein as observed in vitro is not reflected in a
correlation in their CSF concentrations. However, our observations are limited to soluble
α-synuclein and Aβ 42 and Aβ 40 as measured in CSF and leave open the possibility that
intraneuronal or aggregated forms of these proteins interact.
74
The correlation between CSF α-synuclein and CSF Aβ
References
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. Mollenhauer B, Locascio JJ, Schulz-Schaeffer W, Sixel-Doring F, Trenkwalder C, Schlossmacher MG. alphaSynuclein and tau concentrations in cerebrospinal fluid of patients presenting with parkinsonism: a cohort
study. Lancet Neurol 2011;10:230-40.
Hong Z, Shi M, Chung KA, Quinn JF, Peskind ER, Galasko D, Jankovic J, Zabetian CP, Leverenz JB, Baird G,
Montine TJ, Hancock AM, Hwang H, Pan C, Bradner J, Kang UJ, Jensen PH, Zhang J. DJ-1 and alpha-synuclein
in human cerebrospinal fluid as biomarkers of Parkinson’s disease. Brain 2010;133:713-26.
Spies PE, Melis RJ, Sjogren MJ, Rikkert MG, Verbeek MM. Cerebrospinal fluid alpha-synuclein does not
discriminate between dementia disorders. J Alzheimers Dis 2009;16:363-9.
Aerts MB, Esselink RA, Abdo WF, Bloem BR, Verbeek MM. CSF alpha-synuclein does not differentiate
between parkinsonian disorders. Neurobiol Aging 2011.
Mandal PK, Pettegrew JW, Masliah E, Hamilton RL, Mandal R. Interaction between Abeta peptide and alpha
synuclein: molecular mechanisms in overlapping pathology of Alzheimer’s and Parkinson’s in dementia
with Lewy body disease. Neurochem Res 2006;31:1153-62.
Kazmierczak A, Strosznajder JB, Adamczyk A. alpha-Synuclein enhances secretion and toxicity of amyloid
beta peptides in PC12 cells. Neurochem Int 2008;53:263-9.
Bateman RJ, Wen G, Morris JC, Holtzman DM. Fluctuations of CSF amyloid-beta levels: implications for a
diagnostic and therapeutic biomarker. Neurology 2007;68:666-9.
Slats D, Claassen JA, Spies PE, Borm GF, Besse KTC, van Aalst W, Tseng J, Sjogren JM, Olde Rikkert MG, Verbeek
M. Hourly variability of cerebrospinal fluid biomarkers in Alzheimer Disease and healthy older volunteers.
Neurobiol Aging 2011;in press.
van Geel WJ, Abdo WF, Melis R, Williams S, Bloem BR, Verbeek MM. A more efficient enzyme-linked
immunosorbent assay for measurement of alpha-synuclein in cerebrospinal fluid. J Neurosci Methods
2008;168:182-5.
Waring J, Slats D, Gonzales C, Dean R, Lee D, Siemers E, Claassen JA, Li J, LeBlond D, Slemmon R, Bateman RJ,
Tong G, Fox G. An assessment of variability in CSF biomarkers in clinical experimental models: a
meta-analysis. Alzheimers Dement 2011;7:S100.
75
Experiences with CSF analysis in
memory clinics
Part A. of this chapter has been published as: CSF biomarker utilisation and ethical considerations
of biomarker assisted diagnosis and research in dementia: perspectives from within the European
Alzheimer’s Disease Consortium (EADC). Diane Slats, Petra E. Spies, Magnus J.C. Sjögren, Pieter-Jelle Visser,
Marcel M. Verbeek, Marcel G.M. Olde Rikkert, Patrick G. Kehoe, J Neurol Neurosurg Psychiatry 2010;81:124-5
Part B. of this chapter has been published as: Experiences with cerebrospinal fluid analysis in Dutch
memory clinics. Petra E. Spies, Diane Slats, Inez Ramakers, Frans R.J. Verhey, Marcel G.M. Olde Rikkert,
Eur J Neurol 2011;18:1014-6
Chapter
Five
Chapter Five
Chapter Five
Experiences with CSF analysis in memory clinics
A. CSF biomarker utilisation in Europe
Cerebrospinal fluid (CSF) biomarker analysis for dementia diagnostics (i.e., the analysis of
amyloid β42, total tau and phosphorylated tau) is increasingly used in clinical practice.1
However, there is still debate among researchers and clinicians about the sensitivity and
specificity of various biomarker analyses, especially when comparing dementia
subtypes.2 The lack of consensus and heterogeneity for evidence on CSF biomarker use
and validity is likely to have resulted in variable practices: belief in the utility of biomarker
measurement has likely stimulated the use of CSF analysis in clinical practice in some
places whereas elsewhere this practice probably has not been adopted at all. The extent
of this variability in clinical practice is currently unknown but hypothesised to be
considerable.3
Aim
To investigate the extent to which CSF biomarkers are collected for clinical practice and
for research purposes in state of the art memory clinics across Europe, with assessment
of related consenting approaches.
Methods
We asked all 54 memory research centres from the European Alzheimer’s Disease
Consortium (EADC) (http://eadc.alzheimer-europe.org/introduction.html) about practical
and ethical aspects of CSF biomarker collection in dementia diagnostics as part of an
online survey on medical and research practice in these centres. The EADC includes
centres from 18 countries. Centres have been selected as EADC members because of
their expertise in clinical and basic research on Alzheimer disease and related disorders.
Most centres are leading centres on dementia care and research in their countries.
Results
The survey had a 63% response rate. CSF was collected for routine clinical use in 56%
and for scientific use in 87% of responding centres but while 56% of the centres took
CSF routinely, only 44% of all centres used the measurement of CSF biomarkers as part
of the diagnostic process. Although all centres reported that they obtained informed
consent before collecting CSF biomarkers, the scope of consent taking was variable:
79
Chapter Five
44% of the centres obtained separate consent for each biomarker to be collected
whereas 85% obtained informed consent for all possible future research use.
Furthermore, in the case of biomarker research, 65% of the centres retained the ability
to relate biomarkers to patients. More than half of the centres (59%) reported sending
materials abroad for diagnostic and/or scientific reasons, but no centres commercially
sold their biomaterials.
Conclusion
From the centres surveyed here it seems that there is wide acceptance and practice of
obtaining informed consent for collecting CSF biomarkers across Europe. However,
there seem to be variable practices with regard to the acquisition of these biomarkers
for research and clinical use. Also, ethical procedures with respect to handling of these
materials, such as asking consent for future use, shipment of materials and anonymisation
of samples, varied between centres. Despite the growing application of biomarkers in
diagnostic criteria,3,4 we are only at the beginning of a potentially biomarker dominated
era in dementia research and memory clinics. The successful maturation of this era in
Europe would be facilitated by the development of practical and ethical guidelines for
application of CSF biomarkers in dementia care practice as well as in research.5
80
Experiences with CSF analysis in memory clinics
References
1. 2. 3. 4. 5. Frankfort SV, Tulner LR, van Campen JP, Verbeek MM, Jansen RW, Beijnen JH. Amyloid beta protein and tau
in cerebrospinal fluid and plasma as biomarkers for dementia: a review of recent literature. Curr Clin
Pharmacol 2008;3:123-31.
Le Bastard N, Martin JJ, Vanmechelen E, Vanderstichele H, De Deyn PP, Engelborghs S. Added diagnostic
value of CSF biomarkers in differential dementia diagnosis. Neurobiol Aging 2009 doi:10.1016/j.
neurobiolaging.2008.10.017.
Olde Rikkert MG, Lauque S, Frolich L, Vellas B, Dekkers W. The practice of obtaining approval from medical
research ethics committees: a comparison within 12 European countries for a descriptive study on acetylcholinesterase inhibitors in Alzheimer’s dementia. Eur J Neurol 2005;12:212-7.
Dubois B, Feldman HH, Jacova C, DeKosky ST, Barberger-Gateau P, Cummings J, Delacourte A, Galasko D,
Gauthier S, Jicha G, Meguro K, O’Brien J, Pasquier F, Robert P, Rossor M, Salloway S, Stern Y, Visser PJ, Scheltens
P. Research criteria for the diagnosis of Alzheimer’s disease: revising the NINCDS-ADRDA criteria. Lancet
Neurol 2007;6:734-46.
Olde Rikkert MG, van der Vorm A, Burns A, Dekkers W, Robert P, Sartorius N, Selmes J, Stoppe G, VernooijDassen M, Waldemar G. Consensus statement on genetic research in dementia. Am J Alzheimers Dis Other
Demen 2008;23:262-6.
81
Chapter Five
B. CSF analysis in the Netherlands
Abstract
Background: Evidence on cerebrospinal fluid (CSF) analysis to demonstrate Alzheimer
disease has not yet been implemented in diagnostic guidelines.
Methods: We investigated the use of CSF analysis in a survey among all known memory
clinics in the Netherlands, of which 85 of 113 (75.2%) responded.
Results: Sixty percent of respondents used CSF analysis, in 5% (median) of patients. The
analysis almost always confirmed the working diagnosis in 68.4% and sometimes
changed it in 28.2%. Complications occurred very infrequently (0%, median) and were
mild. Reasons not to perform CSF analysis included the lack of clear recommendations
in diagnostic guidelines.
Conclusions: These results ask for a guideline update to clarify the use of CSF analysis
as an add-on diagnostic method.
82
Experiences with CSF analysis in memory clinics
Introduction
Recent hypothetical models on Alzheimer disease (AD) biomarker dynamics regard
cerebrospinal fluid (CSF) biomarker analysis of amyloid β 42 (Aβ 42), phosphorylated tau
(p-tau) and total tau (t-tau) to demonstrate AD as evidence based.1 Most diagnostic
guidelines, however, have not yet implemented this evidence. In contrast to the latest
AD research criteria,2 the most frequently used AD-criteria, NINCDS-ADRDA and DSM-IV-TR,
do not include CSF biomarkers.
Triggered by the gap between biomarker evidence and guidelines, we aimed to
investigate the current use of CSF analysis in the diagnostic work up of Memory Clinic
(MC) patients and the clinician’s opinion on reasons to use CSF analysis.
Methods
A questionnaire was sent to all known MCs in the Netherlands, i.e. all centres known to
perform early dementia diagnostics in a multidisciplinary context. MCs that were
approached were requested to check our list of MCs to see if any centres were missing.
The questionnaire was distributed by e-mail and could be returned by e-mail or on
paper. Participants not responding within two weeks were approached by phone or in
writing. Approval of this study was not needed in accordance with local regulations.
With respect to CSF analysis, questions were asked about frequency of use; type of
patients in whom lumbar puncture was performed; influence of the results on diagnosis;
reasons to apply or not to apply this method; and frequency and type of complications
(Appendix).
One hundred and thirteen MCs were approached, of which 85 responded (75.2%),
and 84 completed at least part of the questions on CSF analysis. Sixty-two MCs of these
84 were based in general (54) or academic (8) hospitals.
Results
Fifty MCs (59.5% of respondents, 46 hospital based) used CSF analysis, in 5% (median,
range 0.01-75%) of patients, most often in patients suspected of mild dementia (Table 1).
Most of the MCs found that the results almost always confirmed the working diagnosis
(68.4%) and hardly ever changed it (53.8%), although change was reported to occur
sometimes by 28.2% of MCs. The results hardly ever contradicted the working diagnosis,
according to 61.5% of MCs.
83
Chapter Five
Table 1 Characteristics of the type of patients and results of using CSF
biomarkers in dementia diagnostics in Dutch memory clinics
Median (range)
No. of MCsa that responded (%b)
Subjective complaints
Cognitive disorder, no dementia
Mild dementia
Moderate-severe dementia
0% (0-75%)
5% (0-95%)
15% (0-100%)
2% (0-80%)
26 (52%)
32 (64%)
33 (66%)
29 (58%)
Failing procedure
Refusal patient
Complications
0% (0-10%)
5% (0-50%)
0% (0-10%)
25 (50%)
30 (60%)
23 (46%)
CSF confirms diagnosis
Never
Hardly ever
Sometimes
Almost always
Always
N (%c)
1 (2.6%)
0 (0%)
11 (28.9%)
26 (68.4%)
0 (0%)
38 (76%)
CSF changes diagnosis
Never
Hardly ever
Sometimes
Almost always
Always
N (%c)
6 (15,4%)
21 (53,8%)
11 (28,2%)
1 (2,6%)
0 (0%)
39 (78%)
CSF contradicts diagnosis
Never
Hardly ever
Sometimes
Almost always
Always
N (%c)
5 (12,8%)
24 (61,5%)
10 (25,6%)
0 (0%)
0 (0%)
39 (78%)
Type of patient
a
MC = memory clinic
percentage of the MCs responding to the specific question
c
percentage of the 50 MCs using CSF analysis
b
84
Experiences with CSF analysis in memory clinics
Reasons to perform CSF analysis were diagnostic uncertainty; to differentiate between
dementia types; or to exclude certain diseases such as Creutzfeldt-Jakob disease or
neurolues; young age of onset; positive family history; the desire for more diagnostic
certainty, e.g. in case of implications of the diagnosis for the patient’s employment.
Research on new biomarkers was not mentioned in this sample of memory clinics.
Reasons not to perform CSF analysis included the lack of clear recommendations in
guidelines; not being convinced of the diagnostic advantage; other diagnostic methods
being more accessible; the procedure being too invasive; the presence of contra-indications for lumbar puncture such as an intracerebral tumor or use of anticoagulants.
Lumbar puncture failed in 0% (median, range 0-10%). Five percent (median, range
0-50%) of patients refused to undergo lumbar puncture. Complications occurred in 0%
(median, range 0-10%). Post-puncture headache was the only complication reported.
Discussion
This survey shows that CSF analysis to diagnose dementia has found its way into clinical
practice, although there is a wide range in the frequency of use. The proportion of
centres using CSF analysis in the Netherlands is largely the same as the proportion of
expert centres across Europe using CSF analysis routinely.3 The analysis was most often
found to confirm the diagnosis, and not to contradict it, although it sometimes led to a
change in diagnosis. We may conclude that in clinical practice, CSF analysis is perceived
as sufficiently valid to change the pre-test probability of AD and of inflammatory
diseases, and thus to change the diagnostic hypothesis. The biomarkers Aβ 42, p-tau and
t-tau are analysed in two core laboratories in the Netherlands (located in Nijmegen and
Amsterdam), of which the Nijmegen laboratory is certified as the national reference
laboratory for CSF analyses for AD, which diminishes the analytical variance that is still
seen in AD biomarker analyses among various laboratories. Newsletters with reference
standards are regularly sent out from the reference centre to reduce heterogeneity in
interpretation of the biomarker values. However, the interpretation of the data itself
could not be controlled for in this study.
We did not ask MCs to specify the exact analyses that were performed. In an
unknown proportion of MCs the standard AD biomarkers Aβ 42, p-tau and t-tau may
have been used, but in some MCs other unspecified CSF analyses may have been
applied. Chosen analyses depend on the clinical context. As shown by the stated
reasons to perform CSF analysis, this is both used to exclude diseases and to gain more
85
Chapter Five
certainty about the presence of AD pathology. The percentages of confirmation and
change may therefore reflect test characteristics of the analysis, but also the selection of
patients in whom lumbar puncture is performed. Complications occurred infrequently
and were mild, which confirms previous reports suggesting high safety of lumbar
puncture.4
The lack of clear recommendations in dementia guidelines, the uncertainty perceived
in clinical practice and the heterogeneity of the application of CSF diagnostics found in
this survey all ask for a dementia guideline update, in which the use of CSF analysis as an
add-on diagnostic method should be clarified. Although relatively high inter-assay
variability of CSF biomarker analyses will make it impossible to provide a single reference
value in a guideline, a high diagnostic performance of CSF biomarkers in diagnosing the
neurodegenerative diseases causing dementia can be reached when laboratories
establish their own reference values.5,6 Studies on cost-effectiveness of CSF analysis and
other AD biomarkers are urgently needed to elucidate the most efficient diagnostic
pathway, considering all currently available diagnostic biomarkers.
86
Experiences with CSF analysis in memory clinics
References
1. Jack CR, Jr., Knopman DS, Jagust WJ, Shaw LM, Aisen PS, Weiner MW, Petersen RC, Trojanowski JQ.
Hypothetical model of dynamic biomarkers of the Alzheimer’s pathological cascade. Lancet Neurol
2010;9:119-28.
2. Dubois B, Feldman HH, Jacova C, DeKosky ST, Barberger-Gateau P, Cummings J, Delacourte A, Galasko D,
Gauthier S, Jicha G, Meguro K, O’Brien J, Pasquier F, Robert P, Rossor M, Salloway S, Stern Y, Visser PJ,
Scheltens P. Research criteria for the diagnosis of Alzheimer’s disease: revising the NINCDS-ADRDA
criteria. Lancet Neurol 2007;6:734-46.
3. Slats D, Spies PE, Sjogren MJ, Visser PJ, Verbeek MM, Rikkert MG, Kehoe PG. CSF biomarker utilisation and
ethical considerations of biomarker assisted diagnosis and research in dementia: perspectives from
within the European Alzheimer’s Disease Consortium (EADC). J Neurol Neurosurg Psychiatry 2010;81:124-5.
4. Peskind ER, Riekse R, Quinn JF, Kaye J, Clark CM, Farlow MR, DeCarli C, Chabal C, Vavrek D, Raskind MA,
Galasko D. Safety and acceptability of the research lumbar puncture. Alzheimer Dis Assoc Disord
2005;19:220-5.
5. de Jong D, Kremer BP, Olde Rikkert MG, Verbeek MM. Current state and future directions of neurochemical
biomarkers for Alzheimer’s disease. Clin Chem Lab Med 2007;45:1421-34.
6. Mattsson N, Zetterberg H, Hansson O, Andreasen N, Parnetti L, Jonsson M, Herukka SK, van der Flier WM,
Blankenstein MA, Ewers M, Rich K, Kaiser E, Verbeek M, Tsolaki M, Mulugeta E, Rosen E, Aarsland D, Visser
PJ, Schroder J, Marcusson J, de LM, Hampel H, Scheltens P, Pirttila T, Wallin A, Jonhagen ME, Minthon L,
Winblad B, Blennow K. CSF biomarkers and incipient Alzheimer disease in patients with mild cognitive
impairment. JAMA 2009;302:385-93.
87
Chapter Five
Appendix
If CSF analysis is not applied for dementia diagnostics within your memory clinic, please
continue to question 42.
34.If CSF analysis is applied for dementia diagnostics within your memory
clinic, what are your reasons to do so?
1.........................................................................................................................
2.........................................................................................................................
3.........................................................................................................................
...........................................................................................................................
...........................................................................................................................
35.
How often does the result of CSF analysis confirm your working diagnosis?
 Never (0-20%)
 Hardly ever (20-40%)
 Sometimes (40-60%)
 Almost always (60-80%)
 Always (80-100%)
36. How often does the result of CSF analysis change your working diagnosis?
 Never (0-20%)
 Hardly ever (20-40%)
 Sometimes (40-60%)
 Almost always (60-80%)
 Always (80-100%)
37.
How often does the result of CSF analysis contradict your working diagnosis?
 Never (0-20%)
 Hardly ever (20-40%)
 Sometimes (40-60%)
 Almost always (60-80%)
 Always (80-100%)
38.In what percentage of patients in the categories below did you perform
lumbar puncture in 2008?
 Persons with subjective complaints
...........%
 Persons with cognitive complaints, no dementia
...........%
 Persons with mild dementia
...........%
 Persons with moderate to severe dementia
...........%
88
Experiences with CSF analysis in memory clinics
39. In what percentage of these patients did lumbar puncture fail?
...........
40.a) In how many patients did complications occur as a result of lumbar puncture?
...........
b) Which were these complications?
.................................................................................................................
.................................................................................................................
41.
What percentage of patients refuses lumbar puncture for diagnostic purposes
when proposed to them?
...........%
42.What are your reasons to decide NOT to use lumbar puncture? (please mark
your answer, more than one answer possible)
 The procedure is too invasive
 The risk of complications is too large
 Other additional examinations are more accessible
 Time to results is too long
 The patient does not fulfill the criteria as formulated in the CBO guidelines*
 I am not convinced of the diagnostic advantage
 The guideline dementia does not give clear advice on CSF analysis
 Other:
...........................................................................................................................
...........................................................................................................................
...........................................................................................................................
* Dutch guidelines: http://www.cbo.nl/thema/Richtlijnen/Overzicht-richtlijnen/Geriatrie/
43.Do you feel you are sufficiently informed about the application of CSF
analysis in clinical practice?
 Yes
 No
89
A prediction model to calculate
the probability of AD
A revised version of this chapter has been accepted for publication as: A prediction model to calculate
probability of Alzheimer’s disease using cerebrospinal fluid biomarkers. Petra E. Spies, Jurgen A.H.R. Claassen,
Petronella G.M. Peer, Marinus A. Blankenstein, Charlotte E. Teunissen, Philip Scheltens, Wiesje M. van der Flier,
Marcel G.M. Olde Rikkert, Marcel M. Verbeek, Alzheimers Dement 2012, DOI: 10.1016/j.jalz.2012.01.010
Chapter
Six
Chapter Six
Chapter Six
Chapter Six
Abstract
Background: We aimed to develop a prediction model based on CSF biomarkers, that
would yield a single estimate representing the probability that dementia in a memory
clinic patient is due to Alzheimer disease (AD).
Methods: All patients suspected of dementia in whom the CSF biomarkers had been
analysed were selected from our memory clinic database. Clinical diagnosis was AD
(n=272) or non-AD (n=289). The prediction model was developed with logistic regression
analysis and included CSF amyloid β 42, phosphorylated tau181 and sex. Validation was
performed on an independent dataset, containing 334 AD and 157 non-AD patients.
Results: The prediction model estimates the probability that AD is present as follows:
p(AD) = 1/(1+e -(-0.3315 + score)), where score is calculated from -1.9486*ln(amyloid β42) +
2.7915*ln(phosphorylated tau181) + 0.9178*sex (male = 0, female = 1). When applied to
the validation dataset, the discriminative ability of the model was very good, area under
the receiver operating characteristic curve 0.85. The agreement between the probability
of AD predicted by the model and the observed frequency of AD diagnoses was very
good after taking into account the difference in AD prevalence between the two
centres.
Conclusions: We developed a prediction model that can accurately predict the
probability of AD in a memory clinic population suspected of dementia based on CSF
amyloid β42, phosphorylated tau181 and sex.
92
A prediction model to calculate the probability of AD
Background
Cerebrospinal fluid (CSF) biomarkers amyloid β 42 (Aβ 42), phosphorylated tau181 (p-tau)
and total tau (t-tau) are increasingly used in the clinical diagnosis of Alzheimer disease
(AD). A guideline for their interpretation is, however, still lacking. The NINCDS-ADRDA
clinical criteria for dementia due to AD were developed before the biomarker era,1 and
while the newly proposed research criteria for AD are more recent, they do not provide
a ready-made guideline for implementation.2,3 A combination of decreased Aβ 42 and
increased t-tau and p-tau concentrations is considered a supportive feature of AD.
However, minimal deviations from the reference ranges are intuitively less supportive
than concentrations that deviate strongly from these reference ranges. Clinicians who
frequently use CSF biomarkers may have developed such an intuitive interpretation of
the results, but for clinicians with less experience with CSF biomarkers, it can be more
difficult to draw equivocal conclusions from the results provided by the lab.
CSF biomarkers provide excellent differentiation between healthy controls and AD
patients,4,5 but this is not a real challenge in clinical practice. The question that the
clinician is mostly faced with, is how to differentiate AD from other dementia disorders
or from a psychiatric disorder that mimics dementia. Although the reported specificities
of CSF biomarkers for other dementias are lower than the specificity for healthy
controls,6-9 they are often used for differential diagnostic purposes in memory clinics.10
Previously reported models based on CSF biomarkers do however not provide a
formula that can be applied in clinical practice for the differentiation of AD from non-­
AD dementia.11,12
We aimed to facilitate the use of CSF biomarkers in clinical practice by developing
a prediction model for AD that yields a single estimate that would represent the
probability that AD is the cause of the dementia in a memory clinic patient. The prediction
model was validated on an independent database.
Methods
Development of the prediction model
All patients with a presumptive diagnosis of dementia who had visited the Radboud
University Nijmegen Medical Centre (RUNMC) memory clinic between 1993 and 2008
and of whom CSF was available were selected. The clinical diagnosis was made by a
multidisciplinary panel that used all available clinical data including neuropsychological
93
Chapter Six
assessment and MRI, but not the CSF results. Diagnosis was AD (n=272) or non-AD
(n=289). AD patients fulfilled the NINCDS-ADRDA criteria and DSM-IV-TR criteria for
dementia. Non-AD patients were defined as patients in whom AD was part of the
differential diagnosis, but who were eventually diagnosed with another diagnosis, such
as another type of dementia or a psychiatric disorder causing the cognitive disorder. We
did not include patients diagnosed with subjective cognitive complaints or mild
cognitive impairment (MCI), since only part of these patients will progress to develop
dementia due to AD, and hence they cannot be classified as AD or non-AD. Levels of
Aβ 42, t-tau, and p-tau were measured at the RUNMC laboratory using ELISA (Innogenetics,
Ghent, Belgium). Patient characteristics are shown in Table 1.
Statistical analysis
The CSF biomarkers were log transformed to reduce strong influence of outliers. The
prediction model was developed with logistic regression analysis, using group
membership (AD or non-AD) as dependent variable. Sex and age were included in the
analysis as readily available patient characteristics. The three log transformed CSF
biomarkers, sex and age were entered with backward stepwise selection. Significant
predictors were retained in the model. A significance level of 0.10 was used. The ability
of the model to discriminate between AD and non-AD was quantified by the area under
the receiver operating characteristic curve (AUC). Calibration of the model, i.e. the
agreement between the probability of AD predicted by the model and the observed
frequency of AD diagnoses, was assessed by plotting the mean predicted probability
against the observed frequency of AD for each probability class (0-0.1, 0.1-0.2, …, 0.9-1.0).
Validation of the prediction model
The dataset used for validation of the prediction model consisted of all patients with
a presumptive diagnosis of dementia who had visited the VUmc Alzheimer Centre
between 2005 and 2008. Clinical diagnosis was made by a multidisciplinary panel
and was AD (n=334) or non-AD (n=157). CSF results had not been used for diagnosis;
all other available data including neuropsychological assessment and MRI had been.
Levels of Aβ42, t-tau, and p-tau of these patients were measured at the VUmc laboratory
using ELISA (Innogenetics, Ghent, Belgium) (Table 1).
Before applying the prediction model on this validation dataset, a shrinkage factor
was applied, which is a methodological procedure in model validation that accounts for
the expected overfitting of the developed model.13 The amount of shrinkage was based
on the p-value of the model and the number of variables analysed. Another adjustment
94
A prediction model to calculate the probability of AD
Table 1 C
haracteristics of patients in the development and validation datasets
M/F (n)
Age (y)
MMSE
CSF Aβ42 (pg/ml)
CSF p-tau (pg/ml)
CSF t-tau (pg/ml)
Diagnosis (n)
VaD
FTLD
DLB
Other dementia †
Psychiatric
Other ‡
Development dataset
Validation dataset
AD
(n=272)
Non-AD
(n=289)
AD
(n=334)
Non-AD
(n=157)
111/161
69.4 ± 9.2
19 ± 5
410 [128-1162]
95 [23-370]
558 [75-2497]
180/109
68.9 ± 11.1
20 ± 7
553 [118-1362]
53 [15-378]
264 [74-13440]
162/172
67.3 ± 9.2
21 ± 5
455 [213-1335]
82 [16-224]
619 [91-3150]
97/60
66.1 ± 9.4
23 ± 5
739 [201-1419]
46 [17-133]
347 [83-1285]
69 (24%)
76 (26%)
50 (17%)
42 (15%)
12 (4%)
40 (14%)
20 (13%)
79 (50%)
32 (20%)
26 (17%)
0
0
Age and MMSE are expressed as mean ± SD, CSF biomarkers are expressed as median [range].
† Other dementia comprised progressive supranuclear palsy, corticobasal degeneration, Parkinson dementia,
dementia due to alcohol abuse
‡
O ther diagnoses comprised, amongst others, delirium, normal pressure hydrocephalus, vitamin B12
deficiency
AD, dementia due to Alzheimer disease; non-AD, non-Alzheimer disease; Aβ 42, amyloid β 42; t-tau, total tau;
p-tau, phosphorylated tau181; VaD, vascular dementia; FTLD, frontotemporal lobe degeneration; DLB,
dementia with Lewy bodies
of the prediction model was based on the need to correct for the difference in AD
prevalence between the centres. Prevalence was defined as the number of AD patients
among the total number of patients in the dataset. If the prevalence in a new population
is higher than the prevalence in the population that was used to develop the model, the
predicted probabilities will underestimate the actual probabilities. A lower prevalence
in the new population will lead to overestimation of the actual probabilities. Ideally, the
mean probability predicted by the model should be equal to the observed prevalence.
The model is adjusted accordingly by correcting the intercept of the logistic regression
formula.14 We examined two types of adjustment. One is a straightforward adjustment
based on the discrepancy between the prevalence of AD in the validation and in the
development datasets: correction factor = ln((prevalence in validation dataset/(1 prevalence in validation dataset))/(prevalence in development dataset/(1 - prevalence
95
Chapter Six
in development dataset))). This correction factor can be used under the condition that
the difference in prevalence between the datasets is not caused by any of the variables
included in the model. It is added to the intercept of the original model.15,16 This resulted
in adjusted model A. The other type of adjustment requires information of individual
cases in the new population on CSF Aβ42, p-tau and sex. The original prediction model
without the intercept is used to calculate a score per individual, by multiplying the
coefficients of the model by the individual’s values of CSF Aβ42, p-tau and sex. A logistic
regression model is then fitted with these individual scores as linear predictor and their
diagnosis (AD or non-AD) as dependent variable. The slope of this model is fixed at 1.
The intercept that is obtained is the correction factor, which is added to the intercept of
the original model.17,18 This resulted in adjusted model B.
To be able to take into account inter-laboratory variation, both laboratories (that
participate actively in the Alzheimer’s Association CSF Quality Control program19,20)
analysed Aβ 42, t-tau, and p-tau in 32 randomly chosen CSF samples from the RUNMC
biobank. For this purpose, the CSF biomarkers were measured in two aliquots of the
same CSF sample, that were both kept frozen at -80 °C and were thawed only once
for analysis in either centre. With linear regression a correction factor was calculated
(Supplementary Figure 1) that could be applied to the CSF values in the validation
dataset.
Discrimination of adjusted model A and B was quantified by the AUC. The AUC is
based on the ranking of the predicted probabilities of the subjects, and is not influenced
by correction factors. Therefore it is the same for both adjusted models. Calibration was
assessed by plotting the predicted probabilities against the observed frequency of AD
in the validation dataset. Statistical analyses were carried out using SAS version 9.2.
Results
Development of the prediction model
The prevalence of AD in the development dataset was 48%. Ln(Aβ 42), ln(p-tau) and
female sex were significant predictors of a diagnosis of AD (Table 2). Age had no influence.
Due to a strong correlation with p-tau, the addition of t-tau was not significant. The
prediction model estimates the probability that AD is present as follows: p(AD) = 1/(1+
e-(intercept + score)), where the intercept is -0.3315 and score is calculated from -1.9486*ln(Aβ42) +
2.7915*ln(p-tau) + 0.9178*sex (Table 2). The AUC of this model was 0.87, indicating very
good discrimination between AD and non-AD. The agreement between the probability
96
A prediction model to calculate the probability of AD
Supplementary Figure 1
Scatterplots of inter-laboratory variation
Aβ42 RUNMC = 6.6 + 1.4*Aβ42 VUmc
p-tau RUNMC = 7.7 + 1.1*p-tau VUmc
t-tau RUNMC = 39.6 + 0.80*t-tau VUmc
Diagonal line represents line of perfect agreement between analysis of RUNMC and VUmc.
Aβ42, CSF amyloid β 42; p-tau, CSF phosphorylated tau181; t-tau, CSF total tau; RUNMC, Radboud University
Nijmegen Medical Centre; VUmc, VU University Medical Center
97
Chapter Six
Table 2 Results of logistic regression analysis for development of the prediction
model
intercept
ln(Aβ42)
ln(p-tau)
sex †
Coefficient
-0.3315
-1.9486
2.7915
0.9178
p-value
<.0001
<.0001
<.0001
†
male=0, female=1
p(AD) = 1 / (1+e -(-0.3315 + score))
Score = -1.9486 * ln(Aβ 42) + 2.7915 * ln(p-tau) + 0.9178 * sex
Aβ 42, amyloid β 42; p-tau, phosphorylated tau181
of AD predicted by the model and the observed frequency of AD diagnoses was also
good (Figure 1): for example of the cases for whom the predicted probability of AD was
0.1 or less (n=86, mean predicted probability 0.056, i.e. AD expected in 5.6%), 3.5%
actually had a diagnosis of AD. Likewise, of the cases for whom the predicted probability
of AD was over 0.9 (n=67, mean predicted probability 0.943, i.e. AD expected in 94.3%),
91.0% actually had a diagnosis of AD.
Validation of the prediction model
A shrinkage factor of 0.9813 was applied by multiplying the coefficients in the prediction
model by this factor (Table 3). The prevalence of AD in the validation dataset was 68%,
whereas it was 48% in the development dataset. Two types of correction for this difference
in prevalence were examined, resulting in two adjusted models: model A with simple
prevalence correction and model B with prevalence correction based on individual data.
When using the simple correction factor based only on the prevalence of AD in the
validation dataset, the intercept was adjusted according to the formula: corrected
intercept = -0.3315 + ln((0.68/(1 - 0.68))/(0.48/(1 - 0.48))) (Table 3; please note that the
prevalences mentioned in the text are rounded and that the aforementioned calculation
thus results in an intercept that is slightly different from the one mentioned in the Table).
This resulted in adjusted model A: p(AD) = 1/(1+e -(intercept + score)), where the intercept is
0.4840 and score is calculated from -1.9122*ln(Aβ42) + 2.7393*ln(p-tau) + 0.9006*sex. The
agreement between the probabilities predicted by adjusted model A and the observed
frequency of AD in the validation dataset was very good in the low and high prediction
classes (Figure 2a and Supplementary Table 1). For example, of the cases for whom the
98
A prediction model to calculate the probability of AD
Figure 1 C
alibration plot of predicted probability against observed frequency of
AD in the development dataset using the original prediction model
Horizontal axis: probability of AD as estimated by the prediction model (range 0-1). Patients are divided
into classes of probabilities (0.0-0.1, 0.1-0.2 etc.). Each dot represents the mean probability of patients
within one class. Vertical axis: actual percentage of AD patients. Each dot represents the percentage of AD
patients within one probability class. Dotted line represents line of perfect agreement between predicted
probability and observed percentage of diagnoses.
Table 3 Intercept and coefficients of the original prediction model and of the
adjusted models A and B
intercept
ln(Aβ42)
ln(p-tau)
sex †
Original model A: Shrinkage + simple B: Shrinkage + prevalence
prevalence correction correction based on individual data
-0.3315
0.4840
1.1528
-1.9486
-1.9122
-1.9122
2.7915
2.7393
2.7393
0.9178
0.9006
0.9006
†
male=0, female=1
p(AD) = 1 / (1+e -(intercept + score))
Score = coefficient * ln(Aβ 42) + coefficient * ln(p-tau) + coefficient * sex
Aβ 42, amyloid β 42; p-tau, phosphorylated tau181
99
Chapter Six
Figure 2 Calibration plot of predicted probability against observed frequency of
AD in the validation dataset
A: adjusted model A with simple correction for difference in prevalence
B: adjusted model B with extensive correction for difference in prevalence
Horizontal axis: probability of AD as estimated by the prediction model (range 0-1). Patients are divided
into classes of probabilities (0.0-0.1, 0.1-0.2 etc.). Each dot represents the mean probability of patients
within one class. Vertical axis: actual percentage of AD patients. Each dot represents the percentage of AD
patients within one probability class. Dotted line represents line of perfect agreement between predicted
probability and observed percentage of diagnoses.
100
A prediction model to calculate the probability of AD
Supplementary Table 1
Observed number of cases per class of predicted probability in the validation
dataset
a: adjusted model A with simple correction for difference in prevalence
Class
0-0.1
0.1-0.2
0.2-0.3
0.3-0.4
0.4-0.5
0.5-0.6
0.6-0.7
0.7-0.8
0.8-0.9
0.9-1.0
Mean predicted
probability of AD
5.96%
14.22%
25.08%
34.25%
43.98%
55.92%
64.77%
75.33%
85.82%
93.97%
n per class
52
44
26
29
22
45
36
57
81
99
AD, n (%)
5 (9.62%)
9 (20.45%)
14 (53.85%)
11 (37.93%)
15 (68.18%)
35 (77.78%)
30 (83.33%)
50 (87.72%)
74 (91.36%)
91 (91.92%)
Non-AD, n (%)
47 (90.38%)
35 (79.55%)
12 (46.15%)
18 (62.07%)
7 (31.82%)
10 (22.22%)
6 (16.67%)
7 (12.28%)
7 (8.64%)
8 (8.08%)
b: adjusted model B with extensive correction for difference in prevalence
Class
0-0.1
0.1-0.2
0.2-0.3
0.3-0.4
0.4-0.5
0.5-0.6
0.6-0.7
0.7-0.8
0.8-0.9
0.9-1.0
Mean predicted
probability of AD
6.72%
14.92%
24.06%
34.84%
45.26%
56.24%
66.74%
74.92%
85.48%
95.02%
n per class
21
39
29
21
27
28
25
54
77
170
AD, n (%)
3 (14.29%)
3 (7.69%)
7 (24.14%)
7 (33.33%)
14 (51.85%)
14 (50.00%)
17 (68.00%)
46 (85.19%)
66 (85.71)
157 (92.35)
Non-AD, n (%)
18 (85.71%)
36 (92.31%)
22 (75.86%)
14 (66.67%)
13 (48.15%)
14 (50.00%)
8 (32.00%)
8 (14.81%)
11 (4.29)
13 (7.65)
predicted probability of AD was 0.1 or less (n=52, mean predicted probability 0.060, i.e.
AD expected in 6.0%), 9.6% actually had a diagnosis of AD; of the cases for whom the
predicted probability of AD was over 0.9 (n=99, mean predicted probability 0.940, i.e.
AD expected in 94.0%), 91.9% had a diagnosis of AD. In the middle classes the observed
frequency of AD was higher than predicted, for example, of the cases for whom the
predicted probability of AD was 0.4-0.5 (n=22, mean predicted probability 0.440, i.e. AD
101
Chapter Six
expected in 44.0%), 68.1% actually had a diagnosis of AD. Mean predicted probability,
which translates to the predicted prevalence, was 0.58.
For the second, more extensive correction resulting in adjusted model B, the scores
of the individual cases in the validation dataset, as obtained with the original prediction
model, were fitted in a logistic regression model with a slope fixed at 1. The intercept of
this logistic regression model was 1.4843 and this was added to the intercept of our
original prediction model, resulting in an adjusted intercept of 1.1528 (Table 3). Adjusted
model B was: p(AD) = 1/(1+e -(1.1528 + score)), score = -1.9122*ln(Aβ42) + 2.7393*ln(p-tau) +
0.9006*sex (Table 3).
The agreement between the probabilities of AD predicted by adjusted model B and
the observed frequency of AD was very good in all classes (Figure 2b and Supplementary
Table 1). For example, of the cases for whom the predicted probability of AD was 0.1 or
less (n=21, mean predicted probability 0.067, i.e. AD expected in 6.7%), 14.3% actually
had a diagnosis of AD; of the cases for whom the predicted probability of AD was
0.4-0.5 (n=27, mean predicted probability 0.453, i.e. AD expected in 45.3%), 51.9% had a
diagnosis of AD; and of the cases for whom the predicted probability of AD was over 0.9
(n=170, mean predicted probability 0.950, i.e. AD expected in 95%), 92.4% had a diagnosis
of AD. Mean predicted probability was 0.68, equal to the prevalence of AD in this
dataset.
When correction for inter-laboratory variation was applied to the CSF biomarker
values in the validation dataset, the performances of both adjusted models (A and B)
remained largely the same; mean predicted probability was 0.57 and 0.68, respectively
(Supplementary Figure 2).
The AUC of adjusted model A and B was 0.85, indicating that the ability of the
model to discriminate between AD and non-AD was very good.
Discussion
We developed a prediction model that yields the probability that a patient from the
memory clinic population suspected of dementia has AD. The probability is calculated
from the patient’s sex combined with the levels of CSF Aβ42 and CSF p-tau. The model
was built using a dataset containing over 500 patients and was validated on an
independent and comparably large dataset. An important feature of this model is that
it takes into account the differential diagnostic dilemma that is often the reality of
clinical practice. When a previously described model based on CSF Aβ 42 and tau aimed
102
A prediction model to calculate the probability of AD
Supplementary Figure 2
Calibration plot of predicted probability against observed frequency of AD in the
validation dataset, corrected for inter-laboratory differences
A: adjusted model A with simple correction for difference in prevalence
B: adjusted model B with extensive correction for difference in prevalence
Horizontal axis: probability of AD as estimated by the prediction model (range 0-1). Patients are divided
into classes of probabilities (0.0-0.1, 0.1-0.2 etc.). Each dot represents the mean probability of patients
within one class. Vertical axis: actual percentage of AD patients. Each dot represents the percentage of AD
patients within one probability class. Dotted line represents line of perfect agreement between predicted
probability and observed percentage of diagnoses.
103
Chapter Six
to discriminate AD from non-AD dementia, almost half of the non-AD patients were
incorrectly classified as AD.12 Another model based on Aβ 42 and p-tau could very well
discriminate between AD and controls and correctly identified MCI patients that would
progress to AD, but did not offer a formula that can be used by others.11 We have
provided the details of our model to allow others to use it and adjust it, and in doing so
retain the information of over 1,000 patients. The complex formula of the prediction
model can be programmed in an application that will only require the clinician to enter
the three parameters sex, Aβ 42 and p-tau.
The variables that were selected in our model are in line with previous research.
Epidemiological studies indicate that AD is more frequent in women.21,22 CSF Aβ 42 and
p-tau were found to identify AD in a group of controls and MCI patients.11 In addition, it
has been shown that this combination of biomarkers may help to differentiate AD from
VaD23, DLB24 and FTD.25 Total tau on the other hand is considered a general marker of
neurodegeneration and the least specific of the three CSF biomarkers.7
An important step in creating a prediction model is its validation. We validated our
model using an independent dataset and by examining two different types of
adjustment for a difference in prevalence. Model A can be calculated and applied by
other centres to adjust the model to their local situation. The minimum requirement for
this correction is to know the prevalence of AD in that centre.14 The possibility to adjust
our model is an important advantage of our approach since existing knowledge can be
incorporated into the model without loss of information and without the need to
develop a new model. Calibration can further be fine-tuned with individual data (model
B), but in that case a complete dataset is required, which might be more difficult to
acquire. However, the prediction model can be used regardless of which of these two
adjustments is applied, as with both adjustments the discrimination between AD and
non-AD was very good (AUC = 0.85) and agreement between predicted probability and
observed frequency of AD was good to excellent, especially when the predicted
probabilities for AD were low or high. The low and high probabilities are also clinically
the most meaningful. For example, even with perfect agreement between predicted
probability of AD and observed frequency of AD, a calculated probability of AD of 50%
will be of no aid in the differential diagnosis between AD or non-AD dementia. In
contrast, calculated probabilities of, for instance, less than 20% or over 80%, will truly
inform decision making, and in our models provide an accurate prediction of the
probability of AD.
Remarkably, correction for inter-laboratory variation did not influence our results.
Differences between the laboratories may have been diminished by inclusion of large
104
A prediction model to calculate the probability of AD
groups of patients and because the laboratories used the same enzyme linked
immunosorbent assays. Inter-laboratory variation was substantial in a multicentre
analysis, but decreased with the use of the same assay.26 Apparently, the quality of the
CSF analyses performed by the two laboratories in this study is comparable. However,
substantial inter-laboratory variation may still exist when the results of our laboratories
are compared to other laboratories, especially if different assays are used. To take this
inter-laboratory variation into account, a correction factor can be calculated and applied
to the values of CSF Aβ 42 and p-tau. Such a correction factor may be derived from the
ongoing international Quality Control program by the Alzheimer’s Association.19,20
Our model provides a continuous outcome. We decided not to calculate an optimal
cut-off point that yields the best ratio between true positives and false negatives.
Defining the ‘best’ ratio is arbitrary, and although a dichotomous outcome would
appear to make clinical diagnosing easier, it would be at the cost of loss of information.
As an example, let us consider a 70% or higher probability of AD as supporting a
diagnosis of AD and a probability of less than 70% as supportive of a non-AD diagnosis.
Sensitivity would be 80.5%, specificity 79.6%, positive predictive value 89.4%, negative
predictive value 65.8%, and 80.2% of cases would be correctly classified (data not
shown). Although these percentages appear very promising, such a dichotomous
cut-off value means that when a patient has a 69.9% probability of AD and another has
a 70.1% probability of AD, the former patient would be classified as non-AD and the
latter as AD, whereas objectively their chance of having AD would be equal.
How should the probability of AD as derived from our model be interpreted and
implemented? For a clinically meaningful interpretation, this estimate should not be
interpreted on its own, but should be weighed against the prior probability of AD,
which the clinician derives from interpretation of a combination of history taking,
physical examination, neuropsychological evaluation and neuroimaging. Theoretically, all
these factors could be included in a prediction model. It is, however, difficult to specify
which items of, for example, history taking or neuropsychological testing should be
included, especially since neither has reached any form of standardization across
centres.
It is important to emphasize that the prediction model is applicable in the memory
clinic population suspected of dementia. The model is not yet validated for MCI patients
or patients for whom AD is not part of the differential diagnosis. However, both in
multicentre studies and in meta-analyses it has repeatedly been shown that
combinations of CSF biomarkers can identify incipient AD in MCI patients.27-30 We
therefore have reason to speculate that our prediction model will be of value in MCI
105
Chapter Six
patients as well, to discriminate MCI patients with AD pathology from those without AD
pathology, but to confirm this assumption the prediction model should first be validated
in a group of MCI patients.
A limitation of our model is one that is common in dementia research: we used the
clinical diagnosis as a reference standard, and some misclassifications are bound to have
occurred. This limitation is not easily overcome. Not only is it difficult to obtain neuropathological information of this many patients, but even a neuropathological diagnosis
is not as unequivocal as it is sometimes presented, especially when the clinical picture
is unclear.31 More recent developments such as amyloid imaging by 11C- Pittsburgh
Compound B cannot serve as a reference standard, since positive scans are found in
33% of healthy controls32 and it is as yet unknown whether these represent very early AD
or false-positives. Diagnoses in our datasets were made by multidisciplinary panels
highly experienced in diagnosing dementia and although some misclassifications are
not ruled out, our model reflects current clinical practice in the best possible way.
To conclude, we developed a prediction model that can serve as a first step in the
implementation of CSF biomarker analysis in diagnosing dementia due to AD in the
memory clinic population suspected of dementia.
106
A prediction model to calculate the probability of AD
References
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stadlan EM. Clinical diagnosis of Alzheimer’s
disease: report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and
Human Services Task Force on Alzheimer’s Disease. Neurology 1984;34:939-44.
Dubois B, Feldman HH, Jacova C, DeKosky ST, Barberger-Gateau P, Cummings J, Delacourte A, Galasko D,
Gauthier S, Jicha G, Meguro K, O’Brien J, Pasquier F, Robert P, Rossor M, Salloway S, Stern Y, Visser PJ,
Scheltens P. Research criteria for the diagnosis of Alzheimer’s disease: revising the NINCDS-ADRDA
criteria. Lancet Neurol 2007;6:734-46.
McKhann GM, Knopman DS, Chertkow H, Hyman BT, Jack CR, Jr., Kawas CH, Klunk WE, Koroshetz WJ, Manly
JJ, Mayeux R, Mohs RC, Morris JC, Rossor MN, Scheltens P, Carrillo MC, Thies B, Weintraub S, Phelps CH. The
diagnosis of dementia due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers
Dement 2011;7:263-9.
Blennow K. Cerebrospinal fluid protein biomarkers for Alzheimer’s disease. NeuroRx 2004;1:213-25.
de Jong D, Kremer BP, Olde Rikkert MG, Verbeek MM. Current state and future directions of neurochemical
biomarkers for Alzheimer’s disease. Clin Chem Lab Med 2007;45:1421-34.
Vanmechelen E, Vanderstichele H, Hulstaert F, Andreasen N, Minthon L, Winblad B, Davidsson P, Blennow
K. Cerebrospinal fluid tau and beta-amyloid(1-42) in dementia disorders. Mech Ageing Dev 2001;122:200511.
Verbeek MM, de Jong D, Kremer HP. Brain-specific proteins in cerebrospinal fluid for the diagnosis of
neurodegenerative diseases. Ann Clin Biochem 2003;40:25-40.
Spies PE, Slats D, Sjogren JM, Kremer BP, Verhey FR, Rikkert MG, Verbeek MM. The cerebrospinal fluid
amyloid beta42/40 ratio in the differentiation of Alzheimer’s disease from non-Alzheimer’s dementia.
Curr Alzheimer Res 2010;7:470-6.
Mulder C, Verwey NA, van der Flier WM, Bouwman FH, Kok A, van Elk EJ, Scheltens P, Blankenstein MA.
Amyloid-beta(1-42), total tau, and phosphorylated tau as cerebrospinal fluid biomarkers for the diagnosis
of Alzheimer disease. Clin Chem 2010;56:248-53.
Spies PE, Slats D, Ramakers I, Verhey FR, Olde Rikkert MG. Experiences with cerebrospinal fluid analysis in
Dutch memory clinics. Eur J Neurol 2011;18:1014-6.
De Meyer G, Shapiro F, Vanderstichele H, Vanmechelen E, Engelborghs S, De Deyn PP, Coart E, Hansson O,
Minthon L, Zetterberg H, Blennow K, Shaw L, Trojanowski JQ. Diagnosis-independent Alzheimer disease
biomarker signature in cognitively normal elderly people. Arch Neurol 2010;67:949-56.
Hulstaert F, Blennow K, Ivanoiu A, Schoonderwaldt HC, Riemenschneider M, De Deyn PP, Bancher C, Cras
P, Wiltfang J, Mehta PD, Iqbal K, Pottel H, Vanmechelen E, Vanderstichele H. Improved discrimination of
AD patients using beta-amyloid(1-42) and tau levels in CSF. Neurology 1999;52:1555-62.
van Houwelingen JC, le CS. Predictive value of statistical models. Stat Med 1990;9:1303-25.
Janssen KJ, Vergouwe Y, Kalkman CJ, Grobbee DE, Moons KG. A simple method to adjust clinical
prediction models to local circumstances. Can J Anaesth 2009;56:194-201.
Smits M, Dippel DW, Steyerberg EW, de Haan GG, Dekker HM, Vos PE, Kool DR, Nederkoorn PJ, Hofman PA,
Twijnstra A, Tanghe HL, Hunink MG. Predicting intracranial traumatic findings on computed tomography
in patients with minor head injury: the CHIP prediction rule. Ann Intern Med 2007;146:397-405.
Steyerberg EW. Clinical prediction models, 1st edition Berlin: Springer, 2009.
Steyerberg EW, Borsboom GJ, van Houwelingen HC, Eijkemans MJ, Habbema JD. Validation and updating
of predictive logistic regression models: a study on sample size and shrinkage. Stat Med 2004;23:2567-86.
Janssen KJ, Moons KG, Kalkman CJ, Grobbee DE, Vergouwe Y. Updating methods improved the
performance of a clinical prediction model in new patients. J Clin Epidemiol 2008;61:76-86.
Alzheimer’s Association Cerebrospinal Fluid Quality Control Program. http://www.neurophys.gu.se/
sektioner/psykiatri_och_neurokemi/neurokem/thealzassqcprogram. Accessed September 1, 2011.
Mattsson N, Andreasson U, Persson S, Arai H, Batish SD, Bernardini S, Bocchio-Chiavetto L, Blankenstein
MA, Carrillo MC, Chalbot S, Coart E, Chiasserini D, Cutler N, Dahlfors G, Duller S, Fagan AM, Forlenza O,
107
Chapter Six
21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 108
Frisoni GB, Galasko D, Galimberti D, Hampel H, Handberg A, Heneka MT, Herskovits AZ, Herukka SK,
Holtzman DM, Humpel C, Hyman BT, Iqbal K, Jucker M, Kaeser SA, Kaiser E, Kapaki E, Kidd D, Klivenyi P,
Knudsen CS, Kummer MP, Lui J, Llado A, Lewczuk P, Li QX, Martins R, Masters C, McAuliffe J, Mercken M,
Moghekar A, Molinuevo JL, Montine TJ, Nowatzke W, O’Brien R, Otto M, Paraskevas GP, Parnetti L, Petersen
RC, Prvulovic D, de Reus HP, Rissman RA, Scarpini E, Stefani A, Soininen H, Schroder J, Shaw LM,
Skinningsrud A, Skrogstad B, Spreer A, Talib L, Teunissen C, Trojanowski JQ, Tumani H, Umek RM, Van BB,
Vanderstichele H, Vecsei L, Verbeek MM, Windisch M, Zhang J, Zetterberg H, Blennow K. The Alzheimer’s
Association external quality control program for cerebrospinal fluid biomarkers. Alzheimers Dement
2011;7:386-95.
Vina J, Lloret A. Why women have more Alzheimer’s disease than men: gender and mitochondrial toxicity
of amyloid-beta peptide. J Alzheimers Dis 2010;20 Suppl 2:S527-S533.
Plassman BL, Langa KM, Fisher GG, Heeringa SG, Weir DR, Ofstedal MB, Burke JR, Hurd MD, Potter GG,
Rodgers WL, Steffens DC, Willis RJ, Wallace RB. Prevalence of dementia in the United States: the aging,
demographics, and memory study. Neuroepidemiology 2007;29:125-32.
de Jong D, Jansen RW, Kremer BP, Verbeek MM. Cerebrospinal fluid amyloid beta42/phosphorylated tau
ratio discriminates between Alzheimer’s disease and vascular dementia. J Gerontol A Biol Sci Med Sci
2006;61:755-8.
Mukaetova-Ladinska EB, Monteith R, Perry EK. Cerebrospinal fluid biomarkers for dementia with lewy
bodies. Int J Alzheimers Dis 2010;2010:536538.
Schoonenboom NS, Pijnenburg YA, Mulder C, Rosso SM, van Elk EJ, Van Kamp GJ, Van Swieten JC,
Scheltens P. Amyloid beta(1-42) and phosphorylated tau in CSF as markers for early-onset Alzheimer
disease. Neurology 2004;62:1580-4.
Verwey NA, van der Flier WM, Blennow K, Clark C, Sokolow S, De Deyn PP, Galasko D, Hampel H, Hartmann
T, Kapaki E, Lannfelt L, Mehta PD, Parnetti L, Petzold A, Pirttila T, Saleh L, Skinningsrud A, Swieten JC,
Verbeek MM, Wiltfang J, Younkin S, Scheltens P, Blankenstein MA. A worldwide multicentre comparison
of assays for cerebrospinal fluid biomarkers in Alzheimer’s disease. Ann Clin Biochem 2009;46:235-40.
Mattsson N, Zetterberg H, Hansson O, Andreasen N, Parnetti L, Jonsson M, Herukka SK, van der Flier WM,
Blankenstein MA, Ewers M, Rich K, Kaiser E, Verbeek M, Tsolaki M, Mulugeta E, Rosen E, Aarsland D, Visser
PJ, Schroder J, Marcusson J, de LM, Hampel H, Scheltens P, Pirttila T, Wallin A, Jonhagen ME, Minthon L,
Winblad B, Blennow K. CSF biomarkers and incipient Alzheimer disease in patients with mild cognitive
impairment. JAMA 2009;302:385-93.
Schmand B, Huizenga HM, van Gool WA. Meta-analysis of CSF and MRI biomarkers for detecting
preclinical Alzheimer’s disease. Psychol Med 2010;40:135-45.
van Rossum I, Vos S, Handels R, Visser PJ. Biomarkers as predictors for conversion from mild cognitive
impairment to Alzheimer-type dementia: implications for trial design. J Alzheimers Dis 2010;20:881-91.
Visser PJ, Verhey F, Knol DL, Scheltens P, Wahlund LO, Freund-Levi Y, Tsolaki M, Minthon L, Wallin AK,
Hampel H, Burger K, Pirttila T, Soininen H, Olde Rikkert MG, Verbeek MM, Spiru L, Blennow K. Prevalence
and prognostic value of CSF markers of Alzheimer’s disease pathology in patients with subjective
cognitive impairment or mild cognitive impairment in the DESCRIPA study: a prospective cohort study.
Lancet Neurol 2009;8:619-27.
Scheltens P, Rockwood K. How golden is the gold standard of neuropathology in dementia? Alzheimers
Dement 2011;7:486-9.
Rowe CC, Ellis KA, Rimajova M, Bourgeat P, Pike KE, Jones G, Fripp J, Tochon-Danguy H, Morandeau L,
O’Keefe G, Price R, Raniga P, Robins P, Acosta O, Lenzo N, Szoeke C, Salvado O, Head R, Martins R, Masters
CL, Ames D, Villemagne VL. Amyloid imaging results from the Australian Imaging, Biomarkers and
Lifestyle (AIBL) study of aging. Neurobiol Aging 2010;31:1275-83.
A prediction model to calculate the probability of AD
109
CSF biomarkers in pathophysiological
hypotheses
Part
Two
Part Two
Part Two
Dynamic hypotheses
This chapter has been published as: CSF biomarkers in Alzheimer’s disease: are the hypotheses more dynamic
than the biomarkers? Petra E. Spies, Marcel M. Verbeek, Marcel G.M. Olde Rikkert, Jurgen A.H.R. Claassen,
J Am Geriatr Soc 2010;58:1619-20
Chapter
Seven
Chapter Seven
Chapter Seven
Dynamic hypotheses
Biomarkers for Alzheimer disease (AD) such as hippocampal atrophy and abnormal
concentrations of cerebrospinal fluid (CSF) amyloid β42 (Aβ 42), phosphorylated tau181
(p-tau) and total tau (t-tau) have become established tools in the diagnostic work-up of
patients suspected of dementia. Recently, AD research has seen a shift in focus towards
the prodromal and preclinical stages of this disease, at the same time creating a shift in
emphasis from clinical criteria towards biomarkers that represent core features of AD.
The CSF biomarkers Aβ 42, p-tau and t-tau are promising candidates, as they reflect the
neuropathologic features of AD and research indicates they become abnormal early in
the disease process. However, data are currently lacking regarding the exact timing of
these biomarkers becoming abnormal, and the order in which these events happen.
Longitudinal studies to elucidate the dynamics of this process are still ongoing, but
several hypothetical models have been proposed. These models are relevant in
providing a framework for further research, for example to establish a paradigm in
which biomarker changes over time could become surrogate end-points in disease
modification trials. In addition, biomarker dynamics might provide information on
disease stage and progression once these models have been confirmed.
An interesting feature that these models have in common is that they make a clear
distinction between the moment that CSF Aβ 42 becomes abnormal and the moment
that CSF tau becomes abnormal. Amyloid deposition (resulting in reduced levels of CSF
Aβ 42) is suggested to be the earliest event, occurring long before cognitive symptoms
are apparent, and plateauing early in the disease. Tau on the other hand is suggested to
become abnormal much later, when clinical symptoms are already present, and to
continue to increase in patients who progress from mild cognitive impairment (MCI) to
moderate and severe dementia.1,2
We question this distinction in timing between amyloid and tau. To illustrate that
these models can be challenged by empirical data, we present a case of early-onset AD
with the longest follow-up of CSF biomarkers reported in the literature. During the nine
years between the earliest clinical manifestation and advanced dementia, we noted no
essential changes in CSF Aβ 42, p-tau and t-tau. At age 53, this man presented with a mild
memory disorder objectified by neuropsychological testing. Neuroimaging (CT
cerebrum) was normal, but CSF analysis showed abnormal results: Aβ42 294 pg/ml
(reference value >500 pg/ml), p-tau 141 pg/ml (<85 pg/ml), and t-tau 660 pg/ml (<350
pg/ml).3 No interference with activities of daily living was established and MCI was
diagnosed. Over the next four years he gradually progressed – confirmed by repeated
neuropsychological evaluation – to a diagnosis of probable AD, Clinical Dementia
Rating (CDR) 1. Nine years after first contact, we re-evaluated this patient, now
115
Chapter Seven
progressed to CDR 2, with marked brain atrophy on CT. Lumbar puncture showed Aβ 42
399 pg/ml, p-tau 108 pg/ml, t-tau 655 pg/ml.
This case suggests that either there is no delay between changes in CSF Aβ 42, p-tau
and t-tau, or that the change in tau occurs much earlier than suggested, that is, already
in the preclinical stage. Studies in MCI patients support this by showing that both CSF
Aβ 42 and tau are already abnormal in those patients who progressed to AD at follow
up.4 We observed no further decrease in Aβ 42 or increase in p-tau or t-tau during disease
progression in this case. This observation is supported by the few longitudinal studies
performed so far in MCI and AD patients – albeit with a much shorter follow up of 2
years – that show a remarkable stability of CSF Aβ 42 and t-tau concentrations.4,5 This
argues against the assumption that a change in pathological direction of these
biomarkers could be of use as surrogate markers to reflect progression of disease.
This case questions models built largely on cross-sectional data. Inter-individual
differences in progression rate and disease duration at diagnosis make cross-sectional
studies unsuitable for inferences about the timing and order of events. Longitudinal
studies covering the full spectrum of the disease, with measurement of imaging and
biochemical markers, will be needed before a valid temporal model can be constructed.
We speculate, however, that such studies will fail to demonstrate a clear distinction
between the timing of Aβ 42 and tau becoming abnormal. Likely, they will demonstrate
a lack of relevant CSF Aβ42, p-tau and t-tau changes over time in the transition from
early MCI to AD.
116
Dynamic hypotheses
References
1. Jack CR, Jr., Knopman DS, Jagust WJ, Shaw LM, Aisen PS, Weiner MW, Petersen RC, Trojanowski JQ. Hypothetical
model of dynamic biomarkers of the Alzheimer’s pathological cascade. Lancet Neurol 2010;9:119-28.
2. Perrin RJ, Fagan AM, Holtzman DM. Multimodal techniques for diagnosis and prognosis of Alzheimer’s
disease. Nature 2009;461:916-22.
3. de Jong D, Kremer BP, Olde Rikkert MG, Verbeek MM. Current state and future directions of neurochemical
biomarkers for Alzheimer’s disease. Clin Chem Lab Med 2007;45:1421-34.
4. Zetterberg H, Pedersen M, Lind K, Svensson M, Rolstad S, Eckerstrom C, Syversen S, Mattsson UB, Ysander C,
Mattsson N, Nordlund A, Vanderstichele H, Vanmechelen E, Jonsson M, Edman A, Blennow K, Wallin A. Intraindividual stability of CSF biomarkers for Alzheimer’s disease over two years. J Alzheimers Dis 2007;12:255-60.
5. Bouwman FH, van der Flier WM, Schoonenboom NS, van Elk EJ, Kok A, Rijmen F, Blankenstein MA, Scheltens
P. Longitudinal changes of CSF biomarkers in memory clinic patients. Neurology 2007;69:1006-11.
117
The correlation between CSF biomarkers
and cognitive functioning
This chapter has been published as: Cerebrospinal fluid tau and amyloid β proteins do not correlate with
cognitive functioning in cognitively impaired memory clinic patients. Petra E. Spies, Jurgen A.H.R. Claassen,
Diane Slats, Marcel G.M. Olde Rikkert, Marcel M. Verbeek, Roy P.C. Kessels, CNS Spectr 2010;15:588-9
Chapter
Eight
Chapter Eight
Chapter Eight
Chapter Eight
Abstract
Objective: Aim of the study was to investigate the influence of CSF Aβ 42, p-tau and
t-tau on cognitive functioning.
Methods: We analyzed the ability of the CSF biomarkers amyloid β 42, phosphorylated
tau181 and total tau to predict the results on a cognitive screening test that assesses
multiple cognitive domains (CAMCOG-R) in 65 memory clinic patients (73.1 ± 8.2 years)
(n=30 probable Alzheimer disease (AD), n=7 possible AD, n=12 non-AD dementia, n=16
mild cognitive impairment).
Results: We found no correlations between CSF biomarkers and CAMCOG-R performance
in the whole group, nor in subgroups based on aberrant biomarker concentrations.
Discussion: Changed concentrations of CSF amyloid β 42, p-tau and t-tau cannot be
directly linked to cognitive function in our sample of patients with cognitive impairment.
Possibly, compensatory mechanisms such as cognitive reserve determine cognitive
performance, rather than the absolute amount of damage caused by Aβ deposition and
tangle formation. In addition, abnormal CSF biomarker concentrations may not be a
direct reflection of the amount of neuronal damage, but merely serve as an indicator of
AD pathology.
Conclusion: While CSF biomarkers are valuable in establishing AD pathology, they
cannot be used to predict severity of cognitive impairment.
120
The correlation between CSF biomarkers and cognitive functioning
Introduction
Cerebrospinal fluid (CSF) biomarkers amyloid β 42 (Aβ 42), phosphorylated tau181 (p-tau)
and total tau (t-tau) concentrations are altered in Alzheimer disease (AD) patients
compared to controls. In other types of dementia and also in part of the patients with
mild cognitive impairment (MCI), changes in these biomarker concentrations, although
less specific, have been reported.1-3 Aβ is said to be neurotoxic,4 while CSF tau and p-tau
are suggested to reflect neurodegeneration.5 Likely, pathological concentrations of
these proteins induce or signal neuronal damage, resulting in cognitive impairment.
To date, research on the correlation between CSF biomarkers and cognitive
functions, however, shows contradicting results. In AD patients, inconsistent correlations
between the Mini Mental State Examination (MMSE) and Aβ42 have been reported.6-9 A
correlation between the MMSE and tau was often lacking,6,7,9 but sometimes present.9
The MMSE is a cognitive screening test that is easily applicable in clinical practice, but
lacks sensitivity and specificity and may be biased by lack of correction for educational
level.10,11 This possibly explains these inconsistent results. A more extensive assessment
of cognitive functioning may be a more valid approach to study the relation between
CSF biomarkers and cognitive functioning. The few studies that have used such an
extensive assessment again showed inconsistent results. In AD patients, a weak
correlation between a paired-associate memory test and CSF p-tau and t-tau, but
not Aβ 42, has been reported,12 while others failed to find a correlation between the
Cambridge Cognitive Examination (CAMCOG), a cognitive measure that covers multiple
domains, and CSF tau.13 In a heterogeneous group of MCI and dementia patients (both
AD and non-AD), a correlation was found between false recognition and levels of CSF
Aβ 42, but not t-tau.14
We aimed to investigate the influence of CSF Aβ 42, p-tau and t-tau concentrations
on cognitive functioning. We hypothesized that changes in these CSF biomarkers
reflect neuronal damage and directly underlie cognitive impairment. Therefore, we
expect a correlation between CSF biomarkers and cognitive performance, irrespective
of the clinical diagnosis. To test this hypothesis, we retrospectively investigated a
heterogeneous group of memory clinic outpatients, using the CAMCOG-R as a multiple-­
domain cognitive assessment.
121
Chapter Eight
Methods
This retrospective study used the database of the memory clinic of the Radboud University
Nijmegen Medical Centre/Alzheimer Centre Nijmegen. We included all outpatients who
visited our memory clinic between 2005 and October 2009 and underwent lumbar
puncture as part of their diagnostic work up. All patients had a diagnosis of either mild
cognitive impairment (MCI) or a type of dementia (note that no lumbar punctures have
been performed in participants without impairment on cognitive screening tests at
admission to the memory clinic). We used the Revised Cambridge Cognitive Examination
(CAMCOG-R) as a measure of cognitive functioning.15 Its score ranges from 0-104, with a
higher score indicating a better performance. Cut-off scores adjusted for age and
education level have been established. Performance can be divided into a memory and a
non-memory section (maximum score 37 and 67, respectively). These sections have been
shown to discriminate between normal aging and early AD.16 Time between lumbar
puncture and CAMCOG-R had to be less than 3 months. CSF was collected in polypropylene
tubes, transported at room temperature to the adjacent laboratory within 30 minutes,
centrifuged after routine investigations, and immediately aliquoted and stored at -80 °C
until analysis. Levels of Aβ42, t-tau, and p-tau in CSF were measured using enzyme linked
immunosorbent assays (Innogenetics, Ghent, Belgium).
Sixty-five patients (32 males) were included. Of these, 30 were eventually diagnosed
with probable AD and 7 with possible AD according to the NINCDS-ADRDA criteria,17 16
were diagnosed with MCI according to the criteria of the International Working Group
on Mild Cognitive Impairment,18 and 12 with another type of dementia (5 vascular
dementia, 2 frontotemporal dementia, 3 dementia with Lewy bodies, 2 dementia of
unknown cause) (see Table 1). Mean age (± SD) was 73.1 (± 8.2) years.
We used linear regression to examine the relation between CSF biomarkers and the
performance on the CAMCOG-R. As dependent variable we used the CAMCOG-R score,
related to the cut-off value, since this cut-off is adjusted for age and education. This relative
score was calculated using the following formula: total score/cut-off * 100, and: (non-)
memory score * (cut-off/maximum score). Aβ42, p-tau181 and t-tau concentrations were
not normally distributed and were log transformed. Analyses were corrected for sex.
Primary analyses included the total sample of 65 patients, predicting the CAMCOG-R total
score and the scores on the memory and non-memory sections using CSF biomarkers.
For secondary analyses, we divided the sample into subgroups based on their biomarker
results to increase the likelihood of a neurodegenerative disorder being present. We used
previously established cut-off values2 (see Table 2) and reran our linear regression models.
122
The correlation between CSF biomarkers and cognitive functioning
Table 1 C
haracteristics of total sample and of subgroups based on clinical
diagnosis
M/F
Age (y)
Years of education
MMSE
CAMCOG-R total
CAMCOG-R memory
CAMCOG-R
non-memory
CSF Aβ42 (pg/ml)
CSF t-tau (pg/ml)
CSF p-tau181 (pg/ml)
Total
sample
Probable
AD
Possible
AD
Other types
of dementia
MCI
32/33
73.1 ± 8.2
10.6 ± 3.8
21.3 ± 5.2
67.9 ± 15.8
20.2 ± 6.8
47.7 ± 10.7
11/19
71.5 ± 8.8
10.4 ± 3.9
19.2 ± 3.9
60.5 ± 11.7
17.0 ± 4.8
43.5 ± 8.9
4/3
74.3 ± 8.5
12.0 ± 4.2
23.6 ± 6.5
73.6 ± 18.4
22.4 ± 8.4
51.1 ± 11.6
6/6
76.1 ± 7.5
11.9 ± 4.3
19.7 ± 6.2
64.3 ± 18.8
20.4 ± 7.7
43.8 ± 12.4
11/5
73.1 ± 7.5
9.6 ± 2.8
25.7 ± 2.9
82.0 ± 7.7
25.2 ± 5.5
56.8 ± 5.2
571 ± 225
570 ± 399
91 ± 47
514 ± 210
706 ± 464
110 ± 52
506 ± 66
739 ± 340
110 ± 27
636 ± 232
334 ± 214
55 ± 30
657 ± 264
416 ± 250
75 ± 33
Values are expressed as means ± SD.
AD, Alzheimer disease; MCI, mild cognitive impairment; Aβ 42, amyloid β 42; t-tau, total tau; p-tau181,
phosphorylated tau181.
We did not use the clinical diagnoses for subgroup analyses, since both CAMCOG-R and
results of CSF analyses may have influenced the diagnostic decision making, whereas CSF
biomarker concentrations are an objective measure and therefore subgroup analyses
based on these concentrations are more likely to yield valid results.
123
Chapter Eight
Table 2 Characteristics of subgroups based on CSF biomarker results
Aberrant Aβ42*
Aberrant
p-tau181*
Aberrant t-tau*
All 3
biomarkers
aberrant*
19
3
4
7
14/19
71.3 ± 8.0
11.3 ± 3.9
20.8 ± 5.5
65.6 ± 16.4
19.5 ± 6.4
46.4 ± 11.5
17
6
2
6
16/15
71.8 ± 8.4
10.8 ± 3.9
21.3 ± 6.0
68.3 ± 17.5
20.2 ± 7.0
48.2 ± 11.8
23
6
4
8
21/20
71.7 ± 8.0
10.9 ± 3.8
21.2 ± 5.7
68.2 ± 16.6
19.6 ± 7.0
48.6 ± 11.0
13
3
1
4
10/11
70.5 ± 8.8
11 ± 3.9
20.9 ± 6.1
67.1 ± 17.6
19.7 ± 7.0
47.4 ± 11.9
410 ± 73
659 ± 415
103 ± 50
478 ± 132
881 ± 366
131 ± 36
515 ± 185
771 ± 374
116 ± 42
412 ± 70
876 ± 359
132 ± 36
Diagnoses
Probable AD
Possible AD
Other dementia
MCI
M/F
Age (y)
Years of education
MMSE
CAMCOG-R total
CAMCOG-R memory
CAMCOG-R
non-memory
CSF Aβ42 (pg/ml)
CSF t-tau (pg/ml)
CSF p-tau181 (pg/ml)
Values are expressed as means ± SD.
AD, Alzheimer disease; MCI, mild cognitive impairment; Aβ 42, amyloid β 42; t-tau, total tau; p-tau181,
phosphorylated tau181.
* Aberrant Aβ 42: < 500 pg/ml; aberrant p-tau181: >85 pg/ml; aberrant t-tau: >350 pg/ml. 2
Results
Figure 1 shows the correlation between each of the CSF biomarkers and the total
CAMCOG-R score. In the total sample including all 65 patients, neither CSF Aβ 42, p-tau181,
t-tau, nor the p-tau/Aβ 42 ratio was a significant predictor of the total score on the
CAMCOG-R. Using the memory score and non-memory score as dependent variables in
the linear regression model did not change these results.
With respect to the secondary analyses, only three models of the 36 models tested
showed a statistically significant outcome. First, in the group with an aberrant Aβ42
concentration, a higher Aβ 42 was related to a higher total CAMCOG-R score (B 34.8,
p=0.040). Second, in the group with an aberrant p-tau181 concentration, a higher t-tau
124
The correlation between CSF biomarkers and cognitive functioning
was related to a lower total CAMCOG-R score (B -24.9, p=0.015). Last, in the group with
an aberrant t-tau concentration, a higher t-tau was related to a lower non-memory
score (B -5.9, p=0.046). All other secondary analyses (k=33) were not significant (data not
shown).
Figure 1 S catter plots of CSF biomarkers vs total CAMCOG-R score (adjusted for
age and education) in patients with probable Alzheimer disease (AD),
possible AD, non-AD dementia and mild cognitive impairment (MCI)
125
Chapter Eight
Discussion
We investigated the correlation between CSF biomarkers and cognitive functioning as
measured with an extensive cognitive screening instrument that covers all major
cognitive domains. We did not find such a correlation in a heterogeneous group of
patients with cognitive impairment. Secondary analyses in subgroups characterized by
abnormal biomarker profiles did not convincingly change this finding.
Our results are in agreement with previous research that reported no correlation
between the CAMCOG and CSF tau in AD patients.13 However, we did not replicate
previously reported findings between CSF biomarkers and memory function.12,12,14
Memory function is validly assessed by the memory section of the CAMCOG-R,16
indicating that a correlation, if present, should have been detected in our study.
Moreover, it is remarkable that previous studies found correlations in the same cognitive
domain but with different CSF biomarkers. That is, one found a correlation with p-tau
and t-tau, but not with Aβ 42,12 while another found a correlation with Aβ 42, but not with
tau.14 It is difficult to explain this discrepancy; it is unlikely that this discrepancy can be
attributed to differences in patient samples – only AD patients, versus a heterogeneous
group of MCI and dementia patients, among which AD patients. Furthermore, this
cannot explain why we did not find a correlation with Aβ 42 in our heterogeneous group.
In addition, a large number of tests and analyses have been performed in these previous
studies that had not been corrected for multiple comparisons. The reported correlations
were weak to moderate at best and in one study possibly driven by a cluster of patients
with extreme values of p-tau and t-tau.12 Consequently, it is possible that these previous
findings are the result of a Type I error (i.e. false positive results).
We hypothesized that changes in Aβ 42 and p-tau and t-tau, as measured in CSF,
would reflect neuronal damage leading to impaired cognitive functioning. The present
findings indicate that these changed concentrations in CSF cannot be directly linked to
cognitive functioning. There are several explanations for this. First, it may not only be
the absolute amount of damage caused by Aβ deposition and tangle formation that
determines cognitive performance, but also compensatory mechanisms, such as interindividual differences in cognitive reserve. Plaques and tangles have indeed been found
in individuals with no signs of cognitive impairment during life,19 suggesting that these
individuals could somehow compensate for the neuronal damage induced by the
plaques and tangles. Moreover, it has been shown that plaque burden is a poor correlate
of cognitive status at the time of death.20
126
The correlation between CSF biomarkers and cognitive functioning
Alternatively, abnormal CSF biomarker concentrations may not be a direct reflection
of neuronal damage, but merely serve as an indicator of AD pathology. For example, in
prodromal AD, i.e. patients with MCI who developed AD at follow-up, aberrant
concentrations of CSF biomarkers have been found.3,21 Furthermore, in patients with
established dementia, concentrations of CSF biomarkers remained stable during
follow-up, while cognitive performance deteriorated.22-24 A correlation between CSF
Aβ 42 and post-mortem plaque burden has been reported to be present25 as well as
absent.26
In AD, it has been hypothesized that amyloid deposition reaches a plateau before
clinical symptoms are apparent.5,27 Increased tau is suggested to develop later and to
continue to increase during disease progression.5 According to this model, one would
not expect a correlation between Aβ 42 and cognitive functioning in AD patients, but
one would expect a correlation between p-tau and t-tau and cognitive functioning in
AD patients, regardless of causality. To prevent ourselves from merely finding a
correlation between two features of one disease, we examined a heterogeneous group
of patients with cognitive impairment. We did not find a correlation between CSF
biomarkers and cognitive functioning, showing that CSF biomarkers do not directly
underlie cognitive impairment in a concentration-dependent way.
Conclusion
Concentrations of CSF amyloid β42, p-tau and t-tau have no clear relationship with
cognitive functioning once cognitive decline has started. Severe cognitive impairment
can be present in the absence of biomarker abnormality, and vice versa. Cognitive
performance may be determined by other factors than the absolute amount of damage
the brain has suffered, such as cognitive reserve. In addition, abnormal CSF biomarker
concentrations may not reflect the amount of neuronal damage, but merely serve as an
indicator of Alzheimer disease. Clinicians should be cautious that CSF biomarkers can be
valuable in establishing AD pathology, but they cannot be used to predict severity of
cognitive impairment. Future studies on the relation between cognitive function and
CSF biomarkers should focus on the rate of cognitive decline over time.
127
Chapter Eight
References
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 128
Blennow K. CSF biomarkers for mild cognitive impairment. J Intern Med 2004;256:224-34.
de Jong D, Kremer BP, Olde Rikkert MG, Verbeek MM. Current state and future directions of neurochemical
biomarkers for Alzheimer’s disease. Clin Chem Lab Med 2007;45:1421-34.
Visser PJ, Verhey F, Knol DL, Scheltens P, Wahlund LO, Freund-Levi Y, Tsolaki M, Minthon L, Wallin AK,
Hampel H, Burger K, Pirttila T, Soininen H, Olde Rikkert MG, Verbeek MM, Spiru L, Blennow K. Prevalence
and prognostic value of CSF markers of Alzheimer’s disease pathology in patients with subjective
cognitive impairment or mild cognitive impairment in the DESCRIPA study: a prospective cohort study.
Lancet Neurol 2009;8:619-27.
Suh YH, Checler F. Amyloid precursor protein, presenilins, and alpha-synuclein: molecular pathogenesis
and pharmacological applications in Alzheimer’s disease. Pharmacol Rev 2002;54:469-525.
Jack CR, Jr., Knopman DS, Jagust WJ, Shaw LM, Aisen PS, Weiner MW, Petersen RC, Trojanowski JQ.
Hypothetical model of dynamic biomarkers of the Alzheimer’s pathological cascade. Lancet Neurol
2010;9:119-28.
Engelborghs S, Maertens K, Vloeberghs E, Aerts T, Somers N, Marien P, De Deyn PP. Neuropsychological
and behavioural correlates of CSF biomarkers in dementia. Neurochem Int 2006;48:286-95.
Ivanoiu A, Sindic CJ. Cerebrospinal fluid TAU protein and amyloid beta42 in mild cognitive impairment:
prediction of progression to Alzheimer’s disease and correlation with the neuropsychological
examination. Neurocase 2005;11:32-9.
Riemenschneider M, Schmolke M, Lautenschlager N, Vanderstichele H, Vanmechelen E, Guder WG, Kurz
A. Association of CSF apolipoprotein E, Abeta42 and cognition in Alzheimer’s disease. Neurobiol Aging
2002;23:205-11.
Sunderland T, Linker G, Mirza N, Putnam KT, Friedman DL, Kimmel LH, Bergeson J, Manetti GJ, Zimmermann
M, Tang B, Bartko JJ, Cohen RM. Decreased beta-amyloid1-42 and increased tau levels in cerebrospinal
fluid of patients with Alzheimer disease. JAMA 2003;289:2094-103.
Olin JT, Zelinski EM. The 12-month reliability of the Mini-Mental State Examination. Psychol Assessment
1991;3:427-32.
Schmand B, Lindeboom J, Hooijer C, Jonker C. Relation between education and dementia: the role of test
bias revisited. J Neurol Neurosurg Psychiatry 1995;59:170-4.
van der Vlies AE, Verwey NA, Bouwman FH, Blankenstein MA, Klein M, Scheltens P, van der Flier WM. CSF
biomarkers in relationship to cognitive profiles in Alzheimer disease. Neurology 2009;72:1056-61.
Tsolaki M, Sakka V, Gerasimou G, Dimacopoulos N, Chatzizisi O, Fountoulakis KN, Kyriazis G, Papanastasiou
J, Kazis A. Correlation of rCBF (SPECT), CSF tau, and cognitive function in patients with dementia of the
Alzheimer’s type, other types of dementia, and control subjects. Am J Alzheimers Dis Other Demen
2001;16:21-31.
Hildebrandt H, Haldenwanger A, Eling P. False recognition correlates with amyloid-beta (1-42) but not
with total tau in cerebrospinal fluid of patients with dementia and mild cognitive impairment. J
Alzheimers Dis 2009;16:157-65.
Roth M, Huppert FA, Mountjoy CQ, Tym E. The Revised Cambridge Examination for Mental Disorders of
the Elderly. Cambridge University Press 1999.
Schmand B, Walstra G, Lindeboom J, Teunisse S, Jonker C. Early detection of Alzheimer’s disease using the
Cambridge Cognitive Examination (CAMCOG). Psychol Med 2000;30:619-27.
McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stadlan EM. Clinical diagnosis of Alzheimer’s
disease: report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and
Human Services Task Force on Alzheimer’s Disease. Neurology 1984;34:939-44.
Winblad B, Palmer K, Kivipelto M, Jelic V, Fratiglioni L, Wahlund LO, Nordberg A, Backman L, Albert M,
Almkvist O, Arai H, Basun H, Blennow K, de LM, DeCarli C, Erkinjuntti T, Giacobini E, Graff C, Hardy J, Jack C,
Jorm A, Ritchie K, van DC, Visser P, Petersen RC. Mild cognitive impairment--beyond controversies,
towards a consensus: report of the International Working Group on Mild Cognitive Impairment. J Intern
Med 2004;256:240-6.
The correlation between CSF biomarkers and cognitive functioning
19. Savva GM, Wharton SB, Ince PG, Forster G, Matthews FE, Brayne C. Age, neuropathology, and dementia.
N Engl J Med 2009;360:2302-9.
20. Gandy S. The role of cerebral amyloid beta accumulation in common forms of Alzheimer disease. J Clin
Invest 2005;115:1121-9.
21. Schmand B, Huizenga HM, van Gool WA. Meta-analysis of CSF and MRI biomarkers for detecting
preclinical Alzheimer’s disease. Psychol Med 2010;40:135-45.
22. Andreasen N, Hesse C, Davidsson P, Minthon L, Wallin A, Winblad B, Vanderstichele H, Vanmechelen E,
Blennow K. Cerebrospinal fluid beta-amyloid(1-42) in Alzheimer disease: differences between early- and
late-onset Alzheimer disease and stability during the course of disease. Arch Neurol 1999;56:673-80.
23. Sluimer JD, Bouwman FH, Vrenken H, Blankenstein MA, Barkhof F, van der Flier WM, Scheltens P.
Whole-brain atrophy rate and CSF biomarker levels in MCI and AD: A longitudinal study. Neurobiol Aging
2008.
24. Zetterberg H, Pedersen M, Lind K, Svensson M, Rolstad S, Eckerstrom C, Syversen S, Mattsson UB, Ysander
C, Mattsson N, Nordlund A, Vanderstichele H, Vanmechelen E, Jonsson M, Edman A, Blennow K, Wallin A.
Intra-individual stability of CSF biomarkers for Alzheimer’s disease over two years. J Alzheimers Dis
2007;12:255-60.
25. Strozyk D, Blennow K, White LR, Launer LJ. CSF Abeta 42 levels correlate with amyloid-neuropathology in
a population-based autopsy study. Neurology 2003;60:652-6.
26. Engelborghs S, Sleegers K, Cras P, Brouwers N, Serneels S, De LE, Martin JJ, Vanmechelen E, Van BC, De
Deyn PP. No association of CSF biomarkers with APOEepsilon4, plaque and tangle burden in definite
Alzheimer’s disease. Brain 2007;130:2320-6.
27. Ingelsson M, Fukumoto H, Newell KL, Growdon JH, Hedley-Whyte ET, Frosch MP, Albert MS, Hyman BT,
Irizarry MC. Early Abeta accumulation and progressive synaptic loss, gliosis, and tangle formation in AD
brain. Neurology 2004;62:925-31.
129
AD and cerebrovascular disease
This chapter has been submitted as: Alzheimer disease is not associated with cerebrovascular disease.
Petra E. Spies, Marcel M. Verbeek, Magnus J.C. Sjögren, Frank-Erik de Leeuw, Jurgen A.H.R. Claassen.
Chapter
Nine
Chapter Nine
Chapter Nine
Chapter Nine
Abstract
Objective: Preclinical and post-mortem studies suggest that Alzheimer disease (AD)
enhances the susceptibility to develop cerebrovascular disease (CVD). The objective of
this study was to investigate this association in a memory clinic population.
Methods: In 81 memory clinic patients (n=42 with AD) the cross-sectional association
between AD and the presence and severity of CVD was investigated with logistic and
linear regression analysis. Additionally, the influence of AD on the relation between
vascular risk factors and CVD was studied. Analyses were adjusted for age. AD was
defined as impairment in episodic memory, hippocampal atrophy and an aberrant
concentration of cerebrospinal fluid (CSF) biomarkers. CVD was assessed on MRI through
a visual rating scale. CSF amyloid β42, amyloid β40 and APOE-ε4 status are all suggested
mechanisms for increased CVD risk in AD, and their associations with CVD were analyzed.
Results: AD was neither associated with the presence (p=0.332) nor with severity of
CVD (p=0.926). Expectedly, patients with more vascular risk factors had more CVD, but
this relationship was not influenced by presence or absence of AD (p=0.104). An
association between either CSF amyloid β42, amyloid β40 or APOE-ε4 status and CVD
was not found, nor did these variables change the relation between vascular risk factors
and CVD.
Conclusions: In this memory clinic population, CVD in patients with AD was fully
accounted for by vascular risk factors and age and was comparable to CVD in patients
without AD. These results do not support the notion that AD pathology leads to CVD.
132
AD and cerebrovascular disease
Introduction
A growing body of research suggests that Alzheimer disease (AD) increases the
susceptibility to develop cerebrovascular disease (CVD), including post-mortem findings
of more severe CVD in AD patients compared to controls.1-6 However, most studies were
performed in animals and convincing translational evidence in living humans is scarce.
Importantly, no research has been performed in a representative clinical setting (e.g. a
memory clinic population). This study aimed to explore evidence for this increased risk
for CVD in a clinical sample of patients with AD, by investigating whether these AD
patients demonstrate more CVD on MRI, compared to patients without AD, and whether
the relation between risk factors for vascular disease and CVD is influenced by the
presence of AD. In addition, we investigated candidate mechanisms for this relationship,
using cerebrospinal fluid (CSF) amyloid β 42 (Aβ 42), CSF Aβ 40 and APOE-ε4 status.
Observations in families harboring amyloid precursor protein (APP) mutations suggest
that Aβ deposition, a hallmark of AD, may lead to CVD. Affected family members suffered
from dementia, cerebral hemorrhage and white matter damage, attributed to the
deposition of Aβ in cerebral blood vessels known as cerebral amyloid angiopathy
(CAA).7,8 Furthermore, severe vascular dysfunction and a distorted vascular architecture
were noted in transgenic mice overexpressing APP and were attributed to Aβ. 2,3 Based
on these findings, CSF Aβ 42 and Aβ 40, as markers of amyloid pathology,9,10 are expected
to be related to vascular dysfunction and thus to be related with the extent of CVD: a
lower CSF Aβ 42 or Aβ 40 (reflecting more severe Aβ pathology) is associated with more
severe CVD. Next, APOE-ε4 is reported to be a risk factor both for vascular disease and
for AD.11,12 Therefore, an association between AD and CVD may be explained by APOE-ε4
status.
Material and Methods
Study population
We included consecutive patients who visited our memory clinic and who underwent
both MRI and a lumbar puncture, n=81. Baseline characteristics are summarized in
Table 1. History of vascular events (present or absent), diabetes mellitus (present or
absent), hypertension (present or absent) and smoking status (current or non-smoker)
were registered from patient records. Vascular events comprised myocardial infarction,
stroke, TIA, percutaneous transluminal coronary angioplasty or coronary artery bypass
133
Chapter Nine
Table 1 Baseline characteristics of the patients included in this study
Total group
M/F, n
Age, mean ± SD (y)
MMSE, mean ± SD
Vascular events present, n
Hypertension, n
Current smoker, n
Diabetes, n
Vascular risk factor score, n
48/33
70.9 ± 8.3
24.0 ± 3.7
17 (21%)
68 (84%)
14 (17%)
13 (16%)
0 8
1 56
2 17
ARWMC, median [range]
1.5 [0-20]
MTA, median [range]
2.0 [0.5-4]
CSF Aβ40, median [range] (pg/ml) 2576 [413-6918]
CSF Aβ42, median [range] (pg/ml) 496 [248-1474]
CSF p-tau, median [range] (pg/ml) 80 [29-201]
CSF t-tau, median [range] (pg/ml) 425 [108-1529]
APOE genotype, n APOE-ε4+
38
APOE-ε427
Clinical diagnosis, n SCI
10 (12%)
MCI
24 (30%)
ADem.
36 (44%)
Other dem.
11 (14%)
AD
Non-AD
22/20
73.5 ± 6.3
22.9 ± 4.0
12 (71%)
37 (88%)
8 (22%)
6 (14%)
3
27
12
2.5 [0-14]
2.5 [2.0-4]
2458 [413-6918]
469 [248-1166]
87 [36-201]
446 [146-1516]
23
12
0
12 (29%)
24 (57%)
6 (14%)
26/13
68.1 ± 9.2
25.2 ± 3.0
5 (13%)
31 (80%)
6 (18%)
7 (18%)
5
29
5
1.5 [0-20]
1.5 [0.5-4]
2985 [647-6042]
583 [329-1474]
70 [29-186]
346 [108-1529]
15
15
10 (26%)
12 (31%)
12 (31%)
5 (13%)
AD, Alzheimer disease (based on research criteria); non-AD, non-Alzheimer disease; CVD, cerebrovascular
disease; MTA, medial temporal lobe atrophy; Aβ 40, amyloid β 40; Aβ 42, amyloid β 42; p-tau181, phosphorylated
tau181; t-tau, total tau; SCI, subjective cognitive impairment; MCI, mild cognitive impairment; ADem.,
clinically established dementia due to Alzheimer disease; Other dem., other types of dementia.
grafting. Hypertension was defined as use of any antihypertensive medication or systolic
blood pressure ≥ 140 mmHg.
MRI assessment
Two clinical researchers (PS, JC) independently rated severity of CVD on MRI FLAIR
transverse sections, using the age-related white matter changes (ARWMC) visual rating
scale.13 This scale assesses 5 brain areas and takes into account lacunar infarcts as well
as white matter lesions. The scale ranges from 0 to 30, in this study scores ranged from
0 to 20. Medial temporal lobe atrophy (MTA) was assessed visually on a scale from
134
AD and cerebrovascular disease
0 to 4, using T1 coronal sections.14 The intraclass correlation coefficient15 (95% CI) was
0.88 (0.82 to 0.92) for the ARWMC, 0.75 (0.63 to 0.83) for right MTA and 0.67 (0.52 to 0.78)
for left MTA. If total score differed 2 points or more between the raters on either scale,
the MRI was reviewed to reach consensus. ARWMC scores and MTA scores of both raters
were averaged for statistical analysis. When left and right MTA score differed, the highest
score was used.
CSF analysis
CSF was collected in polypropylene tubes, transported at room temperature to the
adjacent laboratory within 30 minutes, centrifuged after routine investigations, and
immediately aliquoted and stored at -80 °C until analysis. Concentrations of Aβ 42,
phosphorylated tau181 (p-tau) and total tau (t-tau) were measured using ELISAs
(Innogenetics NV, Ghent, Belgium). CSF Aβ40 was analyzed by a commercial assay based
on the Luminex technology (BioSource, Invitrogen Ltd, Paisley, UK). APOE genotype of
65 patients was available.
Classification of patients
Patients were classified as ‘AD’ (AD present) or ‘non-AD’ (AD absent) (Table 1), according
to the new research criteria for Alzheimer disease:16 impairment in episodic memory
(Rey Auditory Verbal Learning Test), an MTA-score of at least 2, and an aberrant
concentration of at least one of the three CSF biomarkers (i.e. Aβ42 below, or p-tau or
t-tau above the reference value established in our own laboratory17). Classification into
AD or non-AD was based solely on these parameters and did not take into account the
severity of CVD on MRI. The clinical diagnosis was not used because it could have been
influenced by CVD on MRI, with the effect that patients with more severe CVD may have
been less likely to receive a diagnosis of AD, which would lead to an underestimation of
the association between AD and CVD. An additional benefit of using this definition
instead of the clinical diagnosis is that it will result in inclusion of all patients with
Alzheimer disease (including those in mild cognitive impairment stage) and not only of
patients with a clinically established dementia due to Alzheimer disease.
Statistical analysis
The association between AD (present vs. absent) and CVD (present vs. absent, with
present defined as an ARWMC score of at least 1) was investigated with logistic regression
analysis to allow correction for age. The association between AD (dichotomous) and
CVD severity (continuous, using ARWMC score) was investigated with linear regression
135
Chapter Nine
analysis. This analysis, and all further analyses, were also adjusted for age, since age is
a strong independent predictor for CVD.18
To study whether AD influenced the association between vascular risk factors and
CVD, vascular risk factors were summarized into one score as follows: 0, no history of
vascular events and no other risk factors (diabetes mellitus, hypertension or current
smoker); 1, no history of vascular events but at least one risk factor present; 2, history of
vascular events and at least one other risk factor present. An interaction term (AD *
vascular risk factor score) was then introduced in the linear regression model.
To explore the possible factors explaining an association between AD and CVD, we
investigated the association between, subsequently, CSF Aβ42, CSF Aβ 40, and APOE
status (presence of an APOE-ε4 allele) and CVD. In addition, we studied whether each of
these variables influenced the association between vascular risk factors and CVD by
introducing interaction terms. CSF Aβ 42 and Aβ 40 were log transformed to approach
normal distribution. All statistical analyses were performed with SPSS 16.0.
Results
To investigate the hypothesized increased susceptibility to develop CVD in AD, first the
extent of CVD on MRI was compared between memory clinic patients who did and
those who did not have evidence for AD based on the research criteria for AD. AD was
neither associated with the presence of CVD (p=0.332) nor with severity of CVD
(p=0.926). The 1.0 point higher median ARWMC in AD patients was explained by the
higher mean age in this group (Beta for age 0.18, p=0.001). The association between
vascular risk factors and ARWMC (Beta=2.23, p=0.004) was not influenced by the
presence or absence of AD (p=0.104).
Next, we investigated whether biomarkers associated with AD (CSF Aβ42, Aβ40, and
APOE status) were related to CVD, because in animal studies each of these biomarkers
has been associated with cerebrovascular dysfunction. However, neither CSF Aβ 42 nor
CSF Aβ 40 concentrations were associated with the presence (p=0.81 and p=0.87) or
severity (p=0.11 and p=0.53) of CVD (Figure 1), nor did they change the association
between vascular risk factors and CVD severity (p=0.36 and p=0.79). Carrying an
APOE-ε4 allele was not associated with CVD (p=0.42). Unexpectedly even, in contrast
with the hypothesis that APOE-ε4 carriers are more susceptible to cerebrovascular
disease,11 the ARWMC score was 1.5 points lower in patients with an APOE-ε4 allele
(p=0.038). However, when the model was adjusted for vascular risk factors, the
136
AD and cerebrovascular disease
Figure 1 S catterplot of log transformed CSF Aβ42 vs ARWMC score
Open circles: patients with Alzheimer disease (based on research criteria, see text for explanation); closed
squares: non-Alzheimer disease
association lost significance (p=0.111). The association between vascular risk factors and
CVD severity was not different between patients with or without an APOE-­ε4 allele
(p=0.424). Thus, APOE-ε4 status was associated with vascular risk factors, but did not
increase susceptibility to develop CVD.
Discussion
We have investigated the hypothesis that AD patients are more prone to CVD in 81
patients who attended our memory clinic, and who had all undergone a routine but
comprehensive diagnostic workup. Contrary to the hypothesis, the presence and severity
of CVD were not related to whether or not patients had AD. More precisely, the extent of
CVD was fully accounted for by classic vascular risk factors and age. These results were not
attributable to confounding by a clinical diagnosis of AD (that may have been influenced
by MRI findings of CVD), because presence of AD was defined by AD research criteria,
using cognitive, neuroimaging and CSF biomarkers, blinded for ARWMC scores.
137
Chapter Nine
Our results do not support the notion that having AD increases susceptibility to
develop CVD. Previous studies, however, strongly suggested an etiological relationship
between AD and CVD. Most of that research was performed in transgenic mice with
mutated APP genes that abundantly produce Aβ. Vascular function and vascular
architecture in these mice were severely affected by Aβ,2,3 as was their ability to adapt to
changes in their systemic blood pressure in order to stabilize cerebral perfusion, which
led to an increased risk of cerebral ischemia during systemic hypotension.4 Assuming
that cerebrovascular function is similarly affected in AD patients, it was hypothesized
that AD patients – compared to non-AD patients – have an increased prevalence and
severity of CVD. We were unable to confirm this, as presence and severity of CVD were
comparable between AD and non-AD patients in our population.
Since previous studies with transgenic mice attributed vascular effects to Aβ, we
also investigated the possible association of CSF Aβ 42 and Aβ 40 with CVD in our clinical
population. Under the prevailing assumption that low concentrations of Aβ42 and Aβ 40
in CSF reflect increased deposition of these proteins in the brain parenchyma and
vasculature, respectively, we hypothesized an inverse relation between CSF Aβ
concentrations and severity of CVD: lower CSF Aβ concentrations reflect more severe
Aβ pathology and are associated with more severe CVD. However, no such relationship
between severity of Aβ pathology and CVD was found.
The apparent discrepancy between transgenic mice studies and our results may
lead to the question whether these animal studies are truly a valid translational model
for human AD. In general, translational failures are not uncommon, which may partly be
explained by a lack of external validity of some animal models.19 In the case of AD, there
are obvious differences between the various APP mice models and sporadic AD. For
example, the engineered overproduction of Aβ in transgenic APP mice may be
representative for the rare cases of familial AD, but such overproduction is not evident
in sporadic AD (although it is not definitely ruled out either20). The properties of Aβ in
transgenic APP mice are also different: it aggregates in a much shorter time-frame than
in human AD and, when in plaques, dissolves more easily.21 Likewise, proteins that
co-deposit with Aβ in human AD brains, are rarely found in plaques in mice.21 In line
with these differences, it is not surprising that some findings in animals cannot be
replicated in human AD. The Aβ 42 immunization trials, that were very successful in mice
but failed in human studies, are a well-known example.22 Similarly, preservation of
cerebral blood flow during systemic blood pressure fluctuations was severely impaired
in transgenic APP mice, but was intact in AD patients.23
138
Aside from questioning whether mouse models are representative for human
AD and cerebrovascular disease
disease, it should be stressed that signs of CVD such as infarcts or white matter lesions
have not been investigated in transgenic APP mice. This leaves open the possibility that
the Aβ-induced distorted cerebral vasculature and endothelial dysfunction in these
mice do not lead to infarcts or white matter lesions, and would thus go unnoticed on
MRI.2,3 It is therefore conceivable that Aβ affects the cerebral vasculature in AD patients
as it does in transgenic APP mice, but that the resulting vascular dysfunction does not
lead to lesions that can be detected as CVD on MRI. If this were the case, however, the
question rises what the clinical relevance of such vascular dysfunction would be.
Human data suggesting that AD induces CVD stem mostly from autopsy studies.
Patients fulfilling the criteria for definite AD at autopsy had more watershed infarcts,
which led to the suggestion that AD lowers the threshold for development of CVD.1 This
lowered threshold implies that for the same vascular risk factors, CVD would be more
severe when AD pathology is present. This was not confirmed in our in vivo study,
where AD had no influence on the relation between vascular risk factors and CVD. An
important side note to the autopsy study is that patients with hypertension or lacunar
infarcts were excluded, resulting in a highly selected population.1 Therefore, the results
of that study do not apply to a memory clinic population, where a large proportion has
hypertension.24 In our sample, for example, over 80% of patients had hypertension.
We repeated our analyses with APOE, since the presence of an APOE-ε4 allele is a
risk factor for AD but has also been associated with vascular disease.11 The previously
suggested link between AD and CVD (increased expression of CVD in AD) may thus be
explained because APOE-ε4 promotes both diseases. Contrary to previous findings,
however, we found that CVD severity was less in APOE-ε4 carriers compared to
non-carriers. When adjusted for vascular risk factors, this association lost significance. It
appears that APOE-ε4 status does not increase susceptibility to CVD and does not
account for an association between AD and CVD.
Our results are supported by cohort studies with autopsy data that showed a
co-occurrence of AD pathology and vascular lesions,25,26 without evidence for a
synergistic effect between AD pathology and cerebral infarcts.25 Since selection bias is
inherent to these autopsy studies – these studies are often performed in cohorts of
religious orders, and subjects who agree to autopsy may be different from those who
do not agree – the authors concluded that in vivo confirmation of their findings was
desired.25 A major strength of our study is therefore that we investigated a representative
sample – a memory clinic population – with data obtained in vivo. Furthermore, we
used an unbiased definition of AD that was based on positive biomarkers, consistent
with the latest research criteria for AD.16
139
Chapter Nine
A limitation of our study is its cross-sectional design, which is not suitable to study
causality. However, if two factors are causally related – in this study AD and CVD – these
factors cluster more frequently than is expected by chance, and therefore a cross-sectional design should find an association between these factors. The purpose of this
study was to investigate the existence of an association between AD and CVD, with the
premise that causality between these factors had already been inferred from previous
work. Therefore, the absence of this association in this cross-sectional design advocates
against the existence of a causal relationship between AD and CVD.
Based on the comparable distribution of ARWMC scores in the two groups, it is not
expected that a larger sample size would have led to different findings. Much larger
samples may provide sufficient power to detect small (≤1.0) differences in ARWMC
score, but such a finding would have no clinical significance.
Although this study used FLAIR sections to assess white matter lesions, subtle
vascular damage may have been missed that would have been visible with a more
sensitive method such as diffusion tensor imaging.27 However, this limitation applies to
both the AD and the non-AD patients: there is no reason to assume that lesions were
selectively underestimated in AD. A visual rating scale was used to evaluate white
matter lesions. More detailed analysis of white matter lesions can be obtained by
volumetric methods, which express the extent of white matter lesions in ml. However,
the visual scale is widely used in clinical studies and shows excellent correlation with
these volumetric measurements.28
Finally, both AD and CVD are slowly progressive diseases, with little variation in their
biomarkers that were studied here. Therefore, even with a cross-sectional design with a
single time point measurement, it is possible to reliably study associations between
these two diseases.
To conclude, data obtained in vivo from a sample of memory clinic patients showed
that the severity of CVD on MRI was not influenced by AD, nor by surrogate markers of
AD pathology, but was fully accounted for by vascular risk factors and age. CVD in
patients with AD was comparable to CVD in patients without AD. The postulated
increased susceptibility of AD patients to CVD was not confirmed.
140
AD and cerebrovascular disease
References
1. Suter OC, Sunthorn T, Kraftsik R, Straubel J, Darekar P, Khalili K, Miklossy J. Cerebral hypoperfusion
generates cortical watershed microinfarcts in Alzheimer disease. Stroke 2002;33:1986-92.
2. Iadecola C, Zhang F, Niwa K, Eckman C, Turner SK, Fischer E, Younkin S, Borchelt DR, Hsiao KK, Carlson GA.
SOD1 rescues cerebral endothelial dysfunction in mice overexpressing amyloid precursor protein. Nat
Neurosci 1999;2:157-61.
3. Meyer EP, Ulmann-Schuler A, Staufenbiel M, Krucker T. Altered morphology and 3D architecture of brain
vasculature in a mouse model for Alzheimer’s disease. Proc Natl Acad Sci U S A 2008;105:3587-92.
4. Niwa K, Kazama K, Younkin L, Younkin SG, Carlson GA, Iadecola C. Cerebrovascular autoregulation is
profoundly impaired in mice overexpressing amyloid precursor protein. Am J Physiol Heart Circ Physiol
2002;283:H315-H323.
5. Thomas T, Thomas G, McLendon C, Sutton T, Mullan M. beta-Amyloid-mediated vasoactivity and vascular
endothelial damage. Nature 1996;380:168-71.
6. Smith EE, Greenberg SM. Beta-amyloid, blood vessels, and brain function. Stroke 2009;40:2601-6.
7. Maia LF, Mackenzie IR, Feldman HH. Clinical phenotypes of Cerebral Amyloid Angiopathy. J Neurol Sci
2007;257:23-30.
8. Roks G, van Harskamp F, de Koning I, Cruts M, De JC, Kumar-Singh S, Tibben A, Tanghe H, Niermeijer MF,
Hofman A, Van Swieten JC, Van BC, van Duijn CM. Presentation of amyloidosis in carriers of the codon 692
mutation in the amyloid precursor protein gene (APP692). Brain 2000;123 ( Pt 10):2130-40.
9. Blennow K, Hampel H, Weiner M, Zetterberg H. Cerebrospinal fluid and plasma biomarkers in Alzheimer
disease. Nat Rev Neurol 2010;6:131-44.
10. Verbeek MM, Kremer BP, Rikkert MO, Van Domburg PH, Skehan ME, Greenberg SM. Cerebrospinal fluid
amyloid beta(40) is decreased in cerebral amyloid angiopathy. Ann Neurol 2009;66:245-9.
11. Schneider JA, Bienias JL, Wilson RS, Berry-Kravis E, Evans DA, Bennett DA. The apolipoprotein E epsilon4
allele increases the odds of chronic cerebral infarction [corrected] detected at autopsy in older persons.
Stroke 2005;36:954-9.
12. Mayeux R, Stern Y, Ottman R, Tatemichi TK, Tang MX, Maestre G, Ngai C, Tycko B, Ginsberg H. The
apolipoprotein epsilon 4 allele in patients with Alzheimer’s disease. Ann Neurol 1993;34:752-4.
13. Wahlund LO, Barkhof F, Fazekas F, Bronge L, Augustin M, Sjogren M, Wallin A, Ader H, Leys D, Pantoni L,
Pasquier F, Erkinjuntti T, Scheltens P. A new rating scale for age-related white matter changes applicable
to MRI and CT. Stroke 2001;32:1318-22.
14. Scheltens P, Leys D, Barkhof F, Huglo D, Weinstein HC, Vermersch P, Kuiper M, Steinling M, Wolters EC, Valk
J. Atrophy of medial temporal lobes on MRI in “probable” Alzheimer’s disease and normal ageing:
diagnostic value and neuropsychological correlates. J Neurol Neurosurg Psychiatry 1992;55:967-72.
15. Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull 1979;86:420-8.
16. Dubois B, Feldman HH, Jacova C, DeKosky ST, Barberger-Gateau P, Cummings J, Delacourte A, Galasko D,
Gauthier S, Jicha G, Meguro K, O’Brien J, Pasquier F, Robert P, Rossor M, Salloway S, Stern Y, Visser PJ,
Scheltens P. Research criteria for the diagnosis of Alzheimer’s disease: revising the NINCDS-ADRDA
criteria. Lancet Neurol 2007;6:734-46.
17. de Jong D, Kremer BP, Olde Rikkert MG, Verbeek MM. Current state and future directions of neurochemical
biomarkers for Alzheimer’s disease. Clin Chem Lab Med 2007;45:1421-34.
18. Pantoni L, Garcia JH. The significance of cerebral white matter abnormalities 100 years after Binswanger’s
report. A review. Stroke 1995;26:1293-301.
19. van der Worp HB, Howells DW, Sena ES, Porritt MJ, Rewell S, O’Collins V, Macleod MR. Can animal models
of disease reliably inform human studies? PLoS Med 2010;7:e1000245.
20. Johnston JA, Liu WW, Todd SA, Coulson DT, Murphy S, Irvine GB, Passmore AP. Expression and activity of
beta-site amyloid precursor protein cleaving enzyme in Alzheimer’s disease. Biochem Soc Trans
2005;33:1096-100.
21. Timmer NM, Kuiperij HB, De Waal RM, Verbeek MM. Do amyloid beta-associated factors co-deposit with
Abeta in mouse models for Alzheimer’s disease? J Alzheimers Dis 2010;22:345-55.
141
Chapter Nine
22. Orgogozo JM, Gilman S, Dartigues JF, Laurent B, Puel M, Kirby LC, Jouanny P, Dubois B, Eisner L, Flitman S,
Michel BF, Boada M, Frank A, Hock C. Subacute meningoencephalitis in a subset of patients with AD after
Abeta42 immunization. Neurology 2003;61:46-54.
23. Claassen JA, Zhang R. Cerebral autoregulation in Alzheimer’s disease. J Cereb Blood Flow Metab 2011;
31:1572-7.
24. Wong ND, Lopez VA, L’Italien G, Chen R, Kline SE, Franklin SS. Inadequate control of hypertension in US
adults with cardiovascular disease comorbidities in 2003-2004. Arch Intern Med 2007;167:2431-6.
25. Schneider JA, Bennett DA. Where vascular meets neurodegenerative disease. Stroke 2010;41:S144-S146.
26. Snowdon DA, Greiner LH, Mortimer JA, Riley KP, Greiner PA, Markesbery WR. Brain infarction and the
clinical expression of Alzheimer disease. The Nun Study. JAMA 1997;277:813-7.
27. Huang L, Ling XY, Liu SR. Diffusion tensor imaging on white matter in normal adults and elderly patients
with hypertension. Chin Med J (Engl ) 2006;119:1304-7.
28. Kapeller P, Barber R, Vermeulen RJ, Ader H, Scheltens P, Freidl W, Almkvist O, Moretti M, Del ST, Vaghfeldt
P, Enzinger C, Barkhof F, Inzitari D, Erkinjunti T, Schmidt R, Fazekas F. Visual rating of age-related white
matter changes on magnetic resonance imaging: scale comparison, interrater agreement, and
correlations with quantitative measurements. Stroke 2003;34:441-5.
142
AD and cerebrovascular disease
143
Why is CSF Aβ42 decreased in AD?
This chapter has been accepted for publication as: Reviewing reasons for the decreased CSF
Abeta42 concentration in Alzheimer disease. Petra E. Spies, Marcel M. Verbeek, Thomas van Groen,
Jurgen A.H.R. Claassen, Front Biosci 2012
Chapter
Ten
Chapter Ten
Chapter Ten
Chapter Ten
Abstract
Cerebrospinal fluid (CSF) amyloid β42 (Aβ 42) concentrations are decreased in patients
with Alzheimer disease (AD). Consequently, low Aβ42 is considered a positive biomarker
for AD and use of this biomarker is increasingly proposed for early detection of the
disease. Surprisingly, the mechanisms that underlie the decrease in CSF Aβ 42 remain
speculative. Better understanding of this biomarker is an essential step to unravel AD
pathophysiology and to develop and evaluate treatment. Therefore, we systematically
examined the possible reasons for the decreased CSF Aβ 42 concentration in AD. Under
normal conditions, Aβ42, after its production, is secreted into the brain interstitial fluid
(ISF). It can be degraded by proteases, taken up by microglia, or cleared from the ISF
across the blood-brain barrier. Alternatively, it can be transported to the CSF and be
cleared from there. Aggregation of Aβ 42 appears the most likely cause for the decreased
CSF Aβ 42 concentration in AD: the aggregated state inhibits Aβ 42 from being transported
from the ISF to the CSF. Evidence for other possibilities such as a decreased production
of Aβ 42, an increased proteolytic breakdown or microglial uptake of Aβ 42, or an increased
clearance of Aβ 42 to the blood, is – at best – scarce or even absent.
146
Why is CSF Aβ42 decreased in AD?
Introduction
The characteristic plaques of Alzheimer disease (AD) were discovered by Aloïs Alzheimer
in 1906, but it was not until the 1980s that it was demonstrated that these plaques
consist of the amyloid β protein (Aβ).1-3 The realization that this protein was the same as
the protein accumulating in brains of patients with Down syndrome led to the discovery
of the amyloid precursor protein (APP), located on chromosome 21, as the precursor
protein of Aβ.4 Subsequently, pathogenic mutations in the APP gene were discovered in
familial AD.5-8 In the 1990s, Aβ could be measured in cerebrospinal fluid (CSF) for the first
time.9,10 It was thereafter repeatedly shown that CSF Aβ 42 was decreased in AD.11-14
Nowadays, it is an accepted biomarker for AD15,16 that is frequently used in clinical
practice.17,18 Surprisingly, the cause of its decrease has not yet been fully elucidated,
although several explanations have been offered: changes in Aβ generation or
degradation may affect the CSF concentration, or an alteration in the solubility of Aβ 42
may diminish clearance of Aβ 42 from the interstitial fluid (ISF) to CSF.11-14 Determination
of the CSF Aβ 42 concentration in AD may be masked by its interaction with binding
proteins, such as apolipoprotein J or E, or an increased clearance of Aβ42 from CSF might
explain the diminished levels of Aβ 42 in the CSF of AD patients.14
In spite of all these suggestions, the actual explanation for the decreased CSF Aβ 42
in AD remains largely unidentified. Here, we will review the evidence for each of the
proposed explanations to create a better understanding of the underlying mechanisms
that lead to its recognition as an important AD biomarker.
Aβ42 metabolism
1. Production of Aβ 42
Aβ42 is produced by sequential cleavage of the amyloid precursor protein (APP)
(Figure 1). APP is a membrane-bound protein whose function remains unclear, although a
role in cell adhesion, cell growth and synaptogenesis has been suggested.19,20 APP can be
cleaved by either α- or β-secretase, and subsequently by γ-secretase. Whether Aβ is
produced from APP depends on which of these enzymes first cleaves APP.
Cleavage of APP by α-secretase at the plasma membrane or in the trans-Golgi
network
generates N-terminal fragment soluble APP-α (sAPP-α) and C-terminal
fragment C83.21 sAPP-α is released into intracellular vesicles or extracellularly. C83 remains
membrane-bound and is cleaved by γ-secretase to produce p3 (Aβ17-40 and Aβ17-42/43)
147
Chapter Ten
Figure 1 The non-amyloidogenic and amyloidogenic pathways of APP
processing
APP is cleaved by either α- or β-secretase. Cleavage by α-secretase (the non-amyloidogenic pathway)
generates sAPP-α and C83 (left). Cleavage by β-secretase (the amyloidogenic pathway) generates sAPP-β
and C99 (right). C83 is cleaved by γ-secretase, generating AICD and p3 (not shown). C99 is also cleaved by
γ-secretase, generating AICD and Aβ (right). (See text for abbreviations.)
Republished with permission from the American Society for Clinical Investigation, from The role of cerebral
amyloid β accumulation in common forms of Alzheimer disease, S. Gandy, J Clin Invest 2005;115:1121-9;
permission conveyed through Copyright Clearance Center, Inc.
and APP intracellular domain (AICD) (CT57-59).22-25 The latter is released into the cytoplasm.
This pathway is called the non-amyloidogenic pathway since no Aβ40 or Aβ42 is produced.
Cleavage of APP by β-secretase, the amyloidogenic pathway, takes place in the
trans-Golgi network and in endosomes 20,26 and generates N-terminal fragment soluble
APP-β and C-terminal fragment C99. sAPP-β – just like sAPP-α – can be released into
intracellular vesicles or extracellularly. C99 remains membrane-bound and is subsequently
cleaved by γ-secretase to produce Aβ42 or Aβ40 and AICD.22,23 Cleavage by γ-secretase
takes place in the endoplasmatic reticulum, trans-Golgi network, endosomes, and for a
small part at the plasma membrane.26 Other isoforms of Aβ have been identified as well,
with heterogeneity either at the C- or N-terminus of the peptide, but for clarity this
review will be limited to Aβ 42 and Aβ 40.
148
Why is CSF Aβ42 decreased in AD?
The amyloidogenic and non-amyloidogenic pathways are mutually exclusive.22 APP
molecules that are not cleaved at the cell membrane by α-secretase can be internalized by
endocytosis and subsequently be cleaved by β-secretase.27 Redistribution of APP
towards either the cell membrane or the cell interior can therefore influence the
proteolytic pathway that APP undergoes.22
Aβ42 and Aβ40 are either retained in cells or secreted into the ISF that surrounds the
brain cells.23 The equilibrium between intra- and extracellular Aβ 42 and Aβ 40 is in part
regulated by active transport via the receptor for advanced glycation end products
(RAGE),28 a multiligand receptor of the immunoglobulin family.29
2. Degradation of Aβ 42
2.1. Proteases
Several proteases can break down Aβ (both Aβ42 and Aβ 40) in vitro; in vivo evidence
stems mostly from animal studies. Insulin-degrading enzyme (IDE) and neprilysin have
received most attention. IDE is a zinc-metalloproteinase present in the cytosol, in
intracellular membranes and at the cell surface.25,30-32 It is secreted by neurons32 and has
also been found in CSF.25 It can therefore degrade both intracellular and extracellular Aβ.
IDE only degrades soluble, monomeric Aβ. 25,30,31 The cleavage products are not
neurotoxic and are not prone to accumulate in plaques.25 Neprilysin is an axonal and
synaptic membrane-anchored protein with its catalytic site facing the cell exterior31,32
and is therefore primarily involved in the breakdown of extracellular Aβ 42. It can degrade
monomeric as well as oligomeric Aβ.32
Other proteases that have been reported to degrade Aβ in vitro include matrix metalloproteinase-9, angiotensin converting enzyme, cathepsin D, plasmin and endothelin
converting enzyme-1.32,33 Aβ aggregates can stimulate tissue-type and urokinase-type
plasminogen activator to generate plasmin. Plasmin can degrade both monomeric Aβ
and Aβ fibrils in vitro, although degradation of fibrils is far less efficient. Its role in
degrading Aβ in vivo is unclear.33
2.2. Microglia
Microglia are the macrophages of the brain. They take up soluble Aβ by macropinocytosis and possibly by binding of Aβ to the low-density lipoprotein receptor-related
protein (LRP).34 Fibrillar forms of Aβ interact with the microglia cell surface and bind to
the CD36 receptor expressed by microglia. CD36, a major pattern recognition receptor,
mediates the microglial response to Aβ, leading to an intracellular signaling cascade
that stimulates phagocytosis.35 Ablation of microglia in mice led to an increase in soluble
149
Chapter Ten
Aβ 40 and Aβ 42, but had no influence on number and size of Aβ plaques.36 This would
support the idea that microglia are inefficient in degrading fibrillar Aβ.34 However, others
have found that upon activation, for example by cerebral ischemia or after vaccination
with Aβ specific antibodies, microglia can degrade fibrillar Aβ.37-39
3. Clearance of Aβ 42
Aβ 42 and Aβ 40 can also be cleared from the extracellular compartment. It can be cleared
from the ISF across the blood-brain barrier towards the circulation, or it can be
transported from the ISF to the CSF, and be cleared from there. Alternatively, Aβ can be
transported with the ISF or CSF to the peripheral lymphatic system (Figure 2).
3.1. Active transport of Aβ from ISF to systemic circulation
Both Aβ 42 and Aβ 40 can be removed from ISF to the blood by crossing the blood-brain
barrier. The blood-brain barrier is primarily formed by endothelial cells in brain capillaries
that are closely connected via tight junctions which exclude the transfer of many
macromolecules from ISF into blood.40 Astrocytes and pericytes may also play a role in
the control of transport across the blood-brain barrier. Aβ 42 and Aβ 40 are primarily
transported across the blood-brain barrier by LRP.29 LRP is a multiligand lipoprotein
receptor that mediates endocytosis of secreted proteins. It provides a rapid transport of
brain-derived Aβ into the blood.41 LRP-mediated transport is more efficient for Aβ 40
when compared to Aβ42,29,31 probably related to Aβ 40’s lower propensity to aggregate.
Aβ 42 and Aβ 40 that is produced in the periphery can enter the brain ISF from the
systemic circulation by active transport across the blood-brain barrier via RAGE, which
is expressed on the luminal surface of brain vessels.29
3.2. Drainage of Aβ with ISF to lymphatics
One of the routes that ISF can follow is drainage along axonal tracts and through
perivascular spaces. ISF travels along these perivascular spaces towards the surface of the
brain and exits at the base of the skull, where it drains to regional lymph nodes in the
neck.42 The idea that Aβ is transported together with ISF along this pathway is supported
by the fact that Aβ has been found in the basement membranes of capillary and arterial
walls, but was not detectable further downstream in the walls of the carotid arteries.43
3.3. Transport of Aβ from ISF to CSF
Another route for ISF is towards the CSF, where it constitutes 10-30% to the total CSF
production.44 ISF mixes with CSF at the ventricles and presumably also in the sub­-
150
Why is CSF Aβ42 decreased in AD?
Figure 2 S chematic view of Abeta42 and Abeta40 clearance
Aβ can be Aβ 42 or Aβ 40. Solid arrows depict active transport of Aβ; dashed arrows depict passive transport
of Aβ together with ISF or CSF.
BBB = blood-brain barrier; BCB = blood CSF barrier at choroid plexuses; AV = arachnoid villi
arachnoid compartment.40,44 Whether Aβ reaches the CSF via this route is uncertain. The
glia limitans and the pia mater separate ISF from CSF in the subarachnoid space,
however, the extent to which these layers block the passage of solutes and ISF into CSF
is unknown. They may be fully permeable to fluid and small molecules,29 but on the other
hand, the presence of Aβ deposits in the glia limitans in AD suggests that transport of Aβ
from ISF to CSF is limited in the subarachnoid compartment.43
The concentration of Aβ 40 in CSF is about 6-fold greater than the concentration of
Aβ 42;45-47 in AD, the decreased concentration of CSF Aβ 42 makes the CSF Aβ 40
concentration at least 10-fold that of CSF Aβ 42.46-48
151
Chapter Ten
3.4. Absorption of Aβ from CSF to systemic circulation
Labeled Aβ injected into CSF of rats was detectable in blood a few minutes after
injection.49 From cell culture research it was inferred that Aβ can be absorbed from CSF
at the choroid plexuses.50 The endothelium of the blood vessels within the choroid
plexus is leaky, but the epithelium between blood vessels and ventricular CSF contains
tight junctions and forms the blood-CSF barrier (which should not be confused with the
blood-brain barrier). Small molecules can enter or leave the brain.40 Uptake of Aβ
peptides at the choroid plexus was rapid and occurred as a non-diffusional, yet unclear
uptake process. Efflux of Aβ from CSF to blood was favored over influx from blood into
CSF.50 In addition to the blood-CSF barrier, CSF drains to blood in the subarachnoid
compartment, where CSF drains into the major venous sinuses via arachnoid
granulations. Tight junctions in these granulations prevent passage of proteins,51 and it
is therefore unclear whether Aβ can be cleared from CSF to blood in this way.
3.5. Drainage of Aβ with CSF to lymphatics
The perineural sheath along the olfactory nerve represents another pathway of CSF
drainage. The olfactory nerve fibers enter the nasal mucosa in the roof of the nasal
cavity. At this point CSF drains from the perineural subarachnoid space into the
extracellular matrix, where it is absorbed by blind-ended lymphatic capillaries, and
drained to the regional lymph nodes that serve the nasopharynx.44,51 CSF also flows
along the perineuronal sheaths of the cranial and spinal nerves to regional lymphatics
40,51,52
It seems plausible that Aβ is cleared together with CSF via this pathway, but studies
to confirm this are currently lacking.
Why is CSF Aβ42 decreased in Alzheimer disease?
Several morphological changes that are related to the fate of the Aβ42 peptide have been
observed in AD. Upon post-mortem examination, Aβ plaques are found: diffuse plaques
that contain only Aβ42, and classic plaques that consist of both Aβ42 and Aβ40.53,54 In most
AD cases, at least a mild degree of cerebral amyloid angiopathy is found as well, that
consists of Aβ deposits in the walls of leptomeningeal and cortical arteries and less
frequently in capillaries.55 In contrast to the plaques, these vascular deposits contain
mostly Aβ40.56 The concentration of both Aβ42 and Aβ40 in the AD brain is thus
increased.57,58 However, the CSF concentration of Aβ42 is decreased in AD patients, while
the CSF Aβ40 concentration is unaltered.14,48 We will focus on reasons why CSF Aβ42 is
152
Why is CSF Aβ42 decreased in AD?
decreased, while keeping in mind the other changes that occur in AD. Table 1 provides an
overview of these reasons and the arguments supporting or contradicting them.
1. Is Aβ 42 production reduced?
Theoretically, a decreased concentration of Aβ42 in CSF could simply be the result of a
decreased production of Aβ 42, for example by a reduction in the number of viable
neurons that produce Aβ. However, several arguments make this unlikely. First, a
decreased production does not logically connect to an increased Aβ 42 load in the brain.
Second, if the production of Aβ 42 were decreased, the production of Aβ 40 would
probably be decreased as well, since Aβ 42 and Aβ 40 likely share the same production
pathway. This should then lead to a decreased CSF Aβ 40 concentration, which, however,
is not observed in AD. Third, a decreased concentration of CSF Aβ 42 is also found in
familial AD and in Down syndrome, disorders in which overproduction of Aβ 42 has been
clearly demonstrated.59-62 Decreased Aβ 42 thus exists despite overproduction, which
implies that factors other than production play a role in causing the decreased CSF
concentration.
An alternative theory is that Aβ 42 secretion from the cell is decreased, leaving less
Aβ 42 available in the ISF for transport to the CSF. Although intraneuronal Aβ aggregation
has been described in triple transgenic mouse models,63 Aβ predominantly accumulates
extracellularly in human AD. This implies that the secretion of Aβ from the cell to the
exterior is not diminished and that a decreased secretion of Aβ 42 by the cell is unlikely
to be the cause of the decreased CSF Aβ 42 concentration.
Supporting these arguments against a decreased production or secretion of Aβ 42
is a study showing that the production rate of Aβ 42 and Aβ 40, as measured in CSF using
labeled Aβ 42 and Aβ 40, did not differ between controls and AD patients.64
2. Is Aβ 42 degradation increased?
It can be hypothesized that an increase in proteolytic breakdown of Aβ 42 leads to a
decreased CSF Aβ 42 concentration. However, an increase in proteolytic breakdown
would clear the brain from not only Aβ 42, but also from Aβ 4065 and would preclude
formation of plaques. Understandably, research has focused on finding defects in
proteolytic degradation of Aβ rather than overactivity of these proteases to explain the
elevated cerebral levels of Aβ.32,33 It appears that levels of neprilysin decrease with
aging66 and that low levels of neprilysin make certain brain areas more vulnerable to Aβ
deposition,67,68 suggesting that proteolytic activity is decreased rather than increased.
153
Chapter Ten
Table 1 Overview of arguments supporting or contradicting the hypothetical
mechanisms that could explain the decreased CSF Aβ 42 concentration
in AD
Hypothetical mechanism
Supporting arguments
Decreased production of Aβ42 - Could be the result of a
reduction in viable neurons
due to neurodegeneration
Decreased secretion of Aβ42
from neurons
- Intraneuronal Aβ
aggregation has been
described in Tg mice
- An increased brain Aβ42 load
is found post-mortem
- Aβ40 shares the same
production pathway as
Aβ42 but CSF Aβ40 is not
decreased
-D
isorders with an increased
production of Abeta have
decreased concentrations of
CSF Aβ42
- The production of Aβ42
in AD did not differ from
controls in a labeling study
- The typical Aβ plaques of
AD are found extracellularly
- An increased brain Aβ42
load is found post-mortem
- Increased degradation
would probably affect Aβ40
as well, yet the CSF Aβ40
concentration is unchanged
- Levels of neprilysin (one of
the proteases involved in Aβ
degradation) decrease with
aging
Increased degradation of
Aβ42
Increased microglial uptake
of Aβ42
- Activated microglia are
found in the vicinity of Aβ
plaques
Increased clearance of Aβ42
from ISF to blood
- Expression of LRP (which
promotes transport of Aβ
from ISF to blood across
the blood-brain barrier) by
perivascular cells increased
in reaction to Aβ42
154
Contradicting arguments
-A
n increased brain Abeta42
load is found post-mortem
- Microglia become
dysfunctional with aging
- Perivascular cells expressing
LRP degenerated after
uptake of Aβ
- LRP expression is reduced
in AD
- No increase in plasma Aβ42
is found
-A
n increased brain Aβ42
load is found post-mortem
Why is CSF Aβ42 decreased in AD?
Table 1 C
ontinued
Hypothetical mechanism
Supporting arguments
Contradicting arguments
Hampered transport of Aβ42
with ISF to CSF
- Dilated perivascular spaces
may indicate obstruction of
ISF flow
-D
ilated perivascular spaces
are often seen in the elderly
and their clinical relevance
is unclear
-O
ther proteins (including
Aβ40) still reach the CSF
-P
erivascular Abeta deposits
contain mostly Aβ40
-A
β deposits are found
Perivascular deposition of
perivascularly
Aβ42
Parenchymal aggregation and - An increased brain Aβ42
deposition of Aβ42
load is found post-mortem
- Aβ42 is more prone to
aggregation than Aβ40
Increased clearance of Aβ42
from CSF to the blood
-U
ptake of Abeta from CSF
decreased with aging
-N
o increase in plasma Aβ42
is found
3. Is microglial uptake of Aβ 42 increased?
An increased microglial uptake of Aβ 42 could be hypothesized to lead to the decreased
CSF Aβ 42 concentration by leaving less Aβ 42 available in the ISF for transport to the CSF.
Activated microglia are found in the vicinity of compact Aβ plaques,69 however, the very
presence of these plaques shows that microglial uptake of Aβ is insufficient to lead to a
lower-than-normal concentration of Aβ in ISF. On the contrary, it has been suggested
that with age, microglia become dysfunctional and less able to clear Aβ 42.34,70,71 Thus, it
is unlikely that microglial ability to degrade Aβ is increased in AD.
4. Is Aβ 42 clearance affected?
Changes in the clearance of Aβ from either ISF or CSF (the solid arrows in Figure 1) can
influence the Aβ concentration. An increased transport of ISF or CSF, containing Aβ, to
the lymphatics will not affect the Aβ concentration.
4.1. Is clearance of Aβ 42 from ISF increased?
An increased clearance of Aβ 42 from ISF to blood across the blood-brain barrier would
leave a smaller amount of Aβ 42 in the ISF that can be transported to the CSF and could
thus explain the decreased CSF Aβ 42 concentration. The primary transporter of Aβ
155
Chapter Ten
across the blood-brain barrier is LRP. Even though LRP favors clearance of Aβ 40 over
Aβ 42, it can be hypothesized that in AD, clearance of Aβ 42 is upregulated while clearance
of Aβ 40 remains the same. In support of this hypothesis, expression of LRP by perivascular
cells was increased in response to Aβ42 in vitro, while Aβ40 had no such effect; however,
uptake of Aβ resulted in degeneration of these perivascular cells.72 It has also been
reported that LRP expression was reduced in AD, while RAGE expression (which
transports Aβ from blood to brain) was increased.29 These changes do not support an
increased clearance of Aβ 42 from ISF over the blood-brain barrier and could actually
lead to a reduced transport of Aβ42 out of the brain and an increased uptake of
peripheral Aβ 42 into the brain.
Another argument against an increased transport of Aβ 42 across the blood-brain
barrier is the lack of an increase in plasma Aβ 42 concentration. Increased transport of
Aβ 42 to the blood should result in an increase in the plasma concentration of Aβ 42 –
assuming that the peripheral breakdown and elimination of Aβ 42 are not increased in
AD. Reports on the plasma Aβ 42 concentration in AD patients are contradictory, but a
clear increase has not been shown.73,74 Finally, enhanced clearance of Aβ42 from the ISF
is incompatible with the increased Aβ load that is found in AD brains. In summary, it
cannot be concluded that transport across the blood-brain barrier is increased.
4.2. Is transport of Aβ 42 from ISF to CSF hampered?
4.2.1. Hampered flow of ISF to CSF
ISF flows through the perivascular spaces, mixing with CSF in the subarachnoid
compartment at the brain surface and at the ventricles. If Aβ 42-containing ISF cannot
reach the CSF or if Aβ 42 is deposited before ISF reaches the CSF, the Aβ 42 concentration
in CSF will be decreased. Dilated perivascular spaces may be interpreted as a sign that
ISF cannot, or has difficulty to, reach the CSF.75 However, dilated perivascular spaces are
often seen in the elderly and are not AD-specific.76 Their clinical relevance is unclear. The
fact that Aβ 40 and other proteins can still reach the CSF in AD argues against an
obstruction of transport between ISF and CSF.
4.2.2. Perivascular deposition of Aβ
Signs that Aβ 42 is deposited before it reaches the CSF are present: Aβ deposits are found
in the walls of the cerebral vasculature and in the glia limitans in the subarachnoid
compartment.43,77,78 The deposits in the cerebral vasculature consist mostly of Aβ40, and
to a lesser extent of Aβ42.56 In clinically diagnosed cerebral amyloid angiopathy, a
decreased CSF Aβ 40 concentration was found.79 This suggests that perivascular Aβ
156
Why is CSF Aβ42 decreased in AD?
deposition can indeed lead to a decreased CSF Aβ concentration. However, cerebral
amyloid angiopathy found in AD is mostly mild to moderate,55 and the amount of Aβ 42
that is deposited in the cerebral vasculature is presumably insufficient to be the sole
explanation for the decreased CSF Aβ 42 concentration.
4.2.3. Parenchymal aggregation of Aβ
Aggregation of Aβ 42 into insoluble deposits may be a reason why Aβ 42 is decreased in
CSF: less soluble Aβ 42 is left for transport to the CSF, while the insoluble aggregates
cannot be transported from ISF to the CSF. Aβ 42 is much more prone to aggregation
than Aβ 40.56,80 Both diffuse and classic plaques contain Aβ 42, but only the classic plaques,
which are fewer in number, contain Aβ 40.54 Thus, much more Aβ 42 than Aβ 40 is
deposited in AD. Considering that Aβ 40 is the most abundant Aβ peptide in CSF,81 the
relative amount that is deposited in AD may be too little to induce a clear decrease in
the CSF Aβ 40 concentration. In contrast, the relative amount of Aβ 42 that is deposited is
much larger, which may explain why the CSF Aβ 42 concentration is decreased.
4.3. Is CSF Aβ 42 decreased because more Aβ 42 is cleared from CSF?
Uptake of Aβ 42 from the CSF was inferred from research in cell culture that used Aβ40 as
a model compound for all Aβ species.50 Theoretically, this process could be upregulated
in AD, resulting in more Aβ 42 to be transported from the CSF into the blood. Since the
CSF Aβ 40 concentration is unaltered, this upregulation should then selectively affect
Aβ 42 and not Aβ 40. Research to confirm this is currently lacking. In contrast, research in
rats suggests that uptake of Aβ from the CSF decreases with aging.82 Furthermore, the
increase in the plasma concentration of Aβ 42 that would be expected if upregulated
efflux was the case (and peripheral elimination of Aβ was unaltered), has not been
shown convincingly.73,74 Therefore it seems unlikely that an increased clearance of Aβ 42
from the CSF is the cause of the decreased CSF Aβ 42 concentration in AD.
Perspective
We have reviewed a number of possible mechanisms that could theoretically explain
the typical observation of a decreased CSF Aβ42 concentration in AD. It appears very
difficult to understand the decrease in CSF Aβ 42 found in AD. Based on simple but
systematic reasoning, a number of theoretical mechanisms can be formulated, however,
most of these do not concur with the premise that there is an increased Aβ load in the
157
Chapter Ten
AD brain, and that the decreased CSF Aβ 42 is accompanied by an unaltered CSF Aβ 40
concentration. Furthermore, many of these theoretical mechanisms are contradicted by
(often fragmented) findings from research. For example, a decreased production of
Aβ 42 has not been shown: production rates were comparable between controls and AD
patients. Increased proteolytic breakdown of Aβ is unlikely, since available evidence
suggests a decrease in proteolytic activity rather than an increase. An increased
microglial uptake of Aβ 42 is not supported, instead, a decrease in activity with aging is
found. Increased clearance of Aβ 42 from either ISF or CSF has not been reported; rather,
decreases in clearance have been suggested.
What, in our opinion, remains as the most plausible cause of the decreased Aβ42
concentration in CSF is Aβ 42 aggregation. Aβ 42 is known to aggregate easily and once
aggregates have been formed, transport of these aggregates from the ISF to the CSF
may be seriously impaired. Aβ 40 has different properties and is not as prone to
aggregation, which may explain why less Aβ 40 is found deposited in the AD brain, and
why CSF concentrations of Aβ 40 are unaltered.
Until novel mechanistic insights with respect to the different steps in Aβ metabolism
are obtained, this remains currently the most satisfying explanation for the observed Aβ
changes in AD. Our systematic approach may have provided a framework against which
the currently available knowledge from research can be laid out, as pieces of a puzzle.
In this way, CSF Aβ 42 can be much more than a diagnostic test with an empirical cut-off
value for use in clinical settings. Better understanding of this biomarker may be the key
to a better understanding of the disease.
158
Why is CSF Aβ42 decreased in AD?
References
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. Gorevic PD, Goni F, Pons-Estel B, Alvarez F, Peress NS, Frangione B. Isolation and partial characterization of
neurofibrillary tangles and amyloid plaque core in Alzheimer’s disease: immunohistological studies. J
Neuropathol Exp Neurol 1986;45:647-64.
Masters CL, Simms G, Weinman NA, Multhaup G, McDonald BL, Beyreuther K. Amyloid plaque core protein
in Alzheimer disease and Down syndrome. Proc Natl Acad Sci U S A 1985;82:4245-9.
Roher A, Wolfe D, Palutke M, KuKuruga D. Purification, ultrastructure, and chemical analysis of Alzheimer
disease amyloid plaque core protein. Proc Natl Acad Sci U S A 1986;83:2662-6.
Kang J, Lemaire HG, Unterbeck A, Salbaum JM, Masters CL, Grzeschik KH, Multhaup G, Beyreuther K,
Muller-Hill B. The precursor of Alzheimer’s disease amyloid A4 protein resembles a cell-surface receptor.
Nature 1987;325:733-6.
Chartier-Harlin MC, Crawford F, Houlden H, Warren A, Hughes D, Fidani L, Goate A, Rossor M, Roques P, Hardy
J. Early-onset Alzheimer’s disease caused by mutations at codon 717 of the β-amyloid precursor protein
gene. Nature 1991;353:844-6.
Goate A, Chartier-Harlin MC, Mullan M, Brown J, Crawford F, Fidani L, Giuffra L, Haynes A, Irving N, James L.
Segregation of a missense mutation in the amyloid precursor protein gene with familial Alzheimer’s disease.
Nature 1991;349:704-6.
Mullan M, Crawford F, Axelman K, Houlden H, Lilius L, Winblad B, Lannfelt L. A pathogenic mutation for
probable Alzheimer’s disease in the APP gene at the N-terminus of β-amyloid. Nat Genet 1992;1:345-7.
Murrell J, Farlow M, Ghetti B, Benson MD. A mutation in the amyloid precursor protein associated with
hereditary Alzheimer’s disease. Science 1991;254:97-9.
Seubert P, Vigo-Pelfrey C, Esch F, Lee M, Dovey H, Davis D, Sinha S, Schlossmacher M, Whaley J, Swindlehurst
C. Isolation and quantification of soluble Alzheimer’s β-peptide from biological fluids. Nature 1992;359:3257.
Shoji M, Golde TE, Ghiso J, Cheung TT, Estus S, Shaffer LM, Cai XD, McKay DM, Tintner R, Frangione B. Production
of the Alzheimer amyloid β protein by normal proteolytic processing. Science 1992;258:126-9.
Galasko D, Chang L, Motter R, Clark CM, Kaye J, Knopman D, Thomas R, Kholodenko D, Schenk D, Lieberburg
I, Miller B, Green R, Basherad R, Kertiles L, Boss MA, Seubert P. High cerebrospinal fluid tau and low amyloid
β42 levels in the clinical diagnosis of Alzheimer disease and relation to apolipoprotein E genotype. Arch
Neurol 1998;55:937-45.
Motter R, Vigo-Pelfrey C, Kholodenko D, Barbour R, Johnson-Wood K, Galasko D, Chang L, Miller B, Clark C,
Green R. Reduction of β-amyloid peptide42 in the cerebrospinal fluid of patients with Alzheimer’s disease.
Ann Neurol 1995;38:643-8.
Shoji M, Matsubara E, Kanai M, Watanabe M, Nakamura T, Tomidokoro Y, Shizuka M, Wakabayashi K, Igeta Y,
Ikeda Y, Mizushima K, Amari M, Ishiguro K, Kawarabayashi T, Harigaya Y, Okamoto K, Hirai S. Combination
assay of CSF tau, A β 1-40 and A β 1-42(43) as a biochemical marker of Alzheimer’s disease. J Neurol Sci
1998;158:134-40.
Tamaoka A, Sawamura N, Fukushima T, Shoji S, Matsubara E, Shoji M, Hirai S, Furiya Y, Endoh R, Mori H. Amyloid
β protein 42(43) in cerebrospinal fluid of patients with Alzheimer’s disease. J Neurol Sci 1997;148:41-5.
Dubois B, Feldman HH, Jacova C, DeKosky ST, Barberger-Gateau P, Cummings J, Delacourte A, Galasko D,
Gauthier S, Jicha G, Meguro K, O’Brien J, Pasquier F, Robert P, Rossor M, Salloway S, Stern Y, Visser PJ, Scheltens
P. Research criteria for the diagnosis of Alzheimer’s disease: revising the NINCDS-ADRDA criteria. Lancet
Neurol 2007;6:734-46.
McKhann GM, Knopman DS, Chertkow H, Hyman BT, Jack CR, Jr., Kawas CH, Klunk WE, Koroshetz WJ, Manly
JJ, Mayeux R, Mohs RC, Morris JC, Rossor MN, Scheltens P, Carrillo MC, Thies B, Weintraub S, Phelps CH. The
diagnosis of dementia due to Alzheimer’s disease: Recommendations from the National Institute on AgingAlzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement
2011;7:263-9.
Slats D, Spies PE, Sjogren MJ, Visser PJ, Verbeek MM, Olde Rikkert MG, Kehoe PG. CSF biomarker utilisation
and ethical considerations of biomarker assisted diagnosis and research in dementia: perspectives from
within the European Alzheimer’s Disease Consortium (EADC). J Neurol Neurosurg Psychiatry 2010;81:124-5.
159
Chapter Ten
18. Spies PE, Slats D, Ramakers I, Verhey F, Olde Rikkert MG. Experiences with cerebrospinal fluid analysis in
Dutch memory clinics. Eur J Neurol 2011;DOI: 10.1111/j.1468-1331.2010.03222.x.
19. Marzolo MP, Bu G. Lipoprotein receptors and cholesterol in APP trafficking and proteolytic processing,
implications for Alzheimer’s disease. Semin Cell Dev Biol 2009;20:191-200.
20. Thinakaran G, Koo EH. Amyloid precursor protein trafficking, processing, and function. J Biol Chem
2008;283:29615-9.
21. Skovronsky DM, Moore DB, Milla ME, Doms RW, Lee VM. Protein kinase C-dependent α-secretase competes
with β-secretase for cleavage of amyloid-β precursor protein in the trans-golgi network. J Biol Chem
2000;275:2568-75.
22. Gandy S. The role of cerebral amyloid β accumulation in common forms of Alzheimer disease. J Clin Invest
2005;115:1121-9.
23. Kumar-Singh S. Cerebral amyloid angiopathy: pathogenetic mechanisms and link to dense amyloid plaques.
Genes Brain Behav 2008;7 Suppl 1:67-82.
24. Li QX, Fuller SJ, Beyreuther K, Masters CL. The amyloid precursor protein of Alzheimer disease in human
brain and blood. J Leukoc Biol 1999;66:567-74.
25. Suh YH, Checler F. Amyloid precursor protein, presenilins, and α-synuclein: molecular pathogenesis and
pharmacological applications in Alzheimer’s disease. Pharmacol Rev 2002;54:469-525.
26. Vetrivel KS, Thinakaran G. Amyloidogenic processing of β-amyloid precursor protein in intracellular
compartments. Neurology 2006;66:S69-S73.
27. Kamenetz F, Tomita T, Hsieh H, Seabrook G, Borchelt D, Iwatsubo T, Sisodia S, Malinow R. APP processing and
synaptic function. Neuron 2003;37:925-37.
28. Rauk A. The chemistry of Alzheimer’s disease. Chem Soc Rev 2009;38:2698-715.
29. Deane R, Bell RD, Sagare A, Zlokovic BV. Clearance of amyloid-β peptide across the blood-brain barrier:
implication for therapies in Alzheimer’s disease. CNS Neurol Disord Drug Targets 2009;8:16-30.
30. Farris W, Mansourian S, Chang Y, Lindsley L, Eckman EA, Frosch MP, Eckman CB, Tanzi RE, Selkoe DJ, Guenette
S. Insulin-degrading enzyme regulates the levels of insulin, amyloid β-protein, and the β-amyloid precursor
protein intracellular domain in vivo. Proc Natl Acad Sci U S A 2003;100:4162-7.
31. Tanzi RE, Moir RD, Wagner SL. Clearance of Alzheimer’s Aβ peptide: the many roads to perdition. Neuron
2004;43:605-8.
32. Turner AJ, Fisk L, Nalivaeva NN. Targeting amyloid-degrading enzymes as therapeutic strategies in neurodegeneration. Ann N Y Acad Sci 2004;1035:1-20.
33. Selkoe DJ. Clearing the brain’s amyloid cobwebs. Neuron 2001;32:177-80.
34. Lee CY, Landreth GE. The role of microglia in amyloid clearance from the AD brain. J Neural Transm
2010;117:949-60.
35. El Khoury JB, Moore KJ, Means TK, Leung J, Terada K, Toft M, Freeman MW, Luster AD. CD36 mediates the
innate host response to β-amyloid. J Exp Med 2003;197:1657-66.
36. Grathwohl SA, Kalin RE, Bolmont T, Prokop S, Winkelmann G, Kaeser SA, Odenthal J, Radde R, Eldh T, Gandy
S, Aguzzi A, Staufenbiel M, Mathews PM, Wolburg H, Heppner FL, Jucker M. Formation and maintenance of
Alzheimer’s disease β-amyloid plaques in the absence of microglia. Nat Neurosci 2009;12:1361-3.
37. Akiyama H, McGeer PL. Specificity of mechanisms for plaque removal after A β immunotherapy for
Alzheimer disease. Nat Med 2004;10:117-8.
38. Majumdar A, Cruz D, Asamoah N, Buxbaum A, Sohar I, Lobel P, Maxfield FR. Activation of microglia acidifies
lysosomes and leads to degradation of Alzheimer amyloid fibrils. Mol Biol Cell 2007;18:1490-6.
39. Nicoll JA, Barton E, Boche D, Neal JW, Ferrer I, Thompson P, Vlachouli C, Wilkinson D, Bayer A, Games D,
Seubert P, Schenk D, Holmes C. Aβ species removal after aβ42 immunization. J Neuropathol Exp Neurol
2006;65:1040-8.
40. Abbott NJ. Evidence for bulk flow of brain interstitial fluid: significance for physiology and pathology.
Neurochem Int 2004;45:545-52.
41. Deane R, Wu Z, Sagare A, Davis J, Du YS, Hamm K, Xu F, Parisi M, LaRue B, Hu HW, Spijkers P, Guo H, Song X,
Lenting PJ, Van Nostrand WE, Zlokovic BV. LRP/amyloid β-peptide interaction mediates differential brain
efflux of Aβ isoforms. Neuron 2004;43:333-44.
42. Weller RO, Subash M, Preston SD, Mazanti I, Carare RO. Perivascular drainage of amyloid-β peptides from the
brain and its failure in cerebral amyloid angiopathy and Alzheimer’s disease. Brain Pathol 2008;18:253-66.
160
Why is CSF Aβ42 decreased in AD?
43. Weller RO, Djuanda E, Yow HY, Carare RO. Lymphatic drainage of the brain and the pathophysiology of
neurological disease. Acta Neuropathol 2009;117:1-14.
44. Johnston M. The importance of lymphatics in cerebrospinal fluid transport. Lymphat Res Biol 2003;1:41-4.
45. Hansson O, Zetterberg H, Buchhave P, Andreasson U, Londos E, Minthon L, Blennow K. Prediction of
Alzheimer’s disease using the CSF Aβ42/Aβ40 ratio in patients with mild cognitive impairment. Dement
Geriatr Cogn Disord 2007;23:316-20.
46. Lewczuk P, Esselmann H, Otto M, Maler JM, Henkel AW, Henkel MK, Eikenberg O, Antz C, Krause WR, Reulbach
U, Kornhuber J, Wiltfang J. Neurochemical diagnosis of Alzheimer’s dementia by CSF Aβ42, Aβ42/Aβ40 ratio
and total tau. Neurobiol Aging 2004;25:273-81.
47. Kanai M, Matsubara E, Isoe K, Urakami K, Nakashima K, Arai H, Sasaki H, Abe K, Iwatsubo T, Kosaka T, Watanabe
M, Tomidokoro Y, Shizuka M, Mizushima K, Nakamura T, Igeta Y, Ikeda Y, Amari M, Kawarabayashi T, Ishiguro
K, Harigaya Y, Wakabayashi K, Okamoto K, Hirai S, Shoji M. Longitudinal study of cerebrospinal fluid levels of
tau, A β1-40, and A β1-42(43) in Alzheimer’s disease: a study in Japan. Ann Neurol 1998;44:17-26.
48. Spies PE, Slats D, Sjogren JM, Kremer BP, Verhey FR, Rikkert MG, Verbeek MM. The cerebrospinal fluid amyloid
β42/40 ratio in the differentiation of Alzheimer’s disease from non-Alzheimer’s dementia. Curr Alzheimer
Res 2010;7:470-6.
49. Ghersi-Egea JF, Gorevic PD, Ghiso J, Frangione B, Patlak CS, Fenstermacher JD. Fate of cerebrospinal
fluid-borne amyloid β-peptide: rapid clearance into blood and appreciable accumulation by cerebral
arteries. J Neurochem 1996;67:880-3.
50. Crossgrove JS, Li GJ, Zheng W. The choroid plexus removes β-amyloid from brain cerebrospinal fluid. Exp
Biol Med (Maywood ) 2005;230:771-6.
51. Kapoor KG, Katz SE, Grzybowski DM, Lubow M. Cerebrospinal fluid outflow: an evolving perspective. Brain
Res Bull 2008;77:327-34.
52. Johnston M, Papaiconomou C. Cerebrospinal fluid transport: a lymphatic perspective. News Physiol Sci
2002;17:227-30.
53. Iwatsubo T, Odaka A, Suzuki N, Mizusawa H, Nukina N, Ihara Y. Visualization of A β 42(43) and A β 40 in senile
plaques with end-specific A β monoclonals: evidence that an initially deposited species is A β 42(43).
Neuron 1994;13:45-53.
54. Verbeek MM, Eikelenboom P, De Waal RM. Differences between the pathogenesis of senile plaques and
congophilic angiopathy in Alzheimer disease. J Neuropathol Exp Neurol 1997;56:751-61.
55. Rensink AA, De Waal RM, Kremer B, Verbeek MM. Pathogenesis of cerebral amyloid angiopathy. Brain Res
Brain Res Rev 2003;43:207-23.
56. Gravina SA, Ho L, Eckman CB, Long KE, Otvos L, Jr., Younkin LH, Suzuki N, Younkin SG. Amyloid β protein (A
β) in Alzheimer’s disease brain. Biochemical and immunocytochemical analysis with antibodies specific for
forms ending at A β 40 or A β 42(43). J Biol Chem 1995;270:7013-6.
57. Funato H, Yoshimura M, Kusui K, Tamaoka A, Ishikawa K, Ohkoshi N, Namekata K, Okeda R, Ihara Y.
Quantitation of amyloid β-protein (A β) in the cortex during aging and in Alzheimer’s disease. Am J Pathol
1998;152:1633-40.
58. Harigaya Y, Shoji M, Kawarabayashi T, Kanai M, Nakamura T, Iizuka T, Igeta Y, Saido TC, Sahara N, Mori H.
Modified amyloid β protein ending at 42 or 40 with different solubility accumulates in the brain of
Alzheimer’s disease. Biochem Biophys Res Commun 1995;211:1015-22.
59. Moonis M, Swearer JM, Dayaw MP, St George-Hyslop P, Rogaeva E, Kawarai T, Pollen DA. Familial Alzheimer
disease: decreases in CSF Aβ42 levels precede cognitive decline. Neurology 2005;65:323-5.
60. Portelius E, Andreasson U, Ringman JM, Buerger K, Daborg J, Buchhave P, Hansson O, Harmsen A, Gustavsson
MK, Hanse E, Galasko D, Hampel H, Blennow K, Zetterberg H. Distinct cerebrospinal fluid amyloid β peptide
signatures in sporadic and PSEN1 A431E-associated familial Alzheimer’s disease. Mol Neurodegener 2010;5:2.
61. Tamaoka A, Sekijima Y, Matsuno S, Tokuda T, Shoji S, Ikeda SI. Amyloid β protein species in cerebrospinal fluid
and in brain from patients with Down’s syndrome. Ann Neurol 1999;46:933.
62. Tapiola T, Soininen H, Pirttila T. CSF tau and Aβ42 levels in patients with Down’s syndrome. Neurology
2001;56:979-80.
63. LaFerla FM, Green KN, Oddo S. Intracellular amyloid-β in Alzheimer’s disease. Nat Rev Neurosci 2007;8:499509.
64. Mawuenyega KG, Sigurdson W, Ovod V, Munsell L, Kasten T, Morris JC, Yarasheski KE, Bateman RJ. Decreased
161
Chapter Ten
clearance of CNS β-amyloid in Alzheimer’s disease. Science 2010;330:1774.
65. Leissring MA, Farris W, Chang AY, Walsh DM, Wu X, Sun X, Frosch MP, Selkoe DJ. Enhanced proteolysis of
β-amyloid in APP transgenic mice prevents plaque formation, secondary pathology, and premature death.
Neuron 2003;40:1087-93.
66. Apelt J, Ach K, Schliebs R. Aging-related down-regulation of neprilysin, a putative β-amyloid-degrading
enzyme, in transgenic Tg2576 Alzheimer-like mouse brain is accompanied by an astroglial upregulation in
the vicinity of β-amyloid plaques. Neurosci Lett 2003;339:183-6.
67. Akiyama H, Kondo H, Ikeda K, Kato M, McGeer PL. Immunohistochemical localization of neprilysin in the
human cerebral cortex: inverse association with vulnerability to amyloid β-protein (Aβ) deposition. Brain
Res 2001;902:277-81.
68. Yasojima K, Akiyama H, McGeer EG, McGeer PL. Reduced neprilysin in high plaque areas of Alzheimer brain:
a possible relationship to deficient degradation of β-amyloid peptide. Neurosci Lett 2001;297:97-100.
69. Ohgami T, Kitamoto T, Shin RW, Kaneko Y, Ogomori K, Tateishi J. Increased senile plaques without microglia
in Alzheimer’s disease. Acta Neuropathol 1991;81:242-7.
70. Streit WJ, Conde JR, Fendrick SE, Flanary BE, Mariani CL. Role of microglia in the central nervous system’s
immune response. Neurol Res 2005;27:685-91.
71. Streit WJ. Microglial senescence: does the brain’s immune system have an expiration date? Trends Neurosci
2006;29:506-10.
72. Wilhelmus MM, Otte-Holler I, van Triel JJ, Veerhuis R, Maat-Schieman ML, Bu G, De Waal RM, Verbeek MM.
Lipoprotein receptor-related protein-1 mediates amyloid-β-mediated cell death of cerebrovascular cells.
Am J Pathol 2007;171:1989-99.
73. Giedraitis V, Sundelof J, Irizarry MC, Garevik N, Hyman BT, Wahlund LO, Ingelsson M, Lannfelt L. The normal
equilibrium between CSF and plasma amyloid β levels is disrupted in Alzheimer’s disease. Neurosci Lett
2007;427:127-31.
74. Kawarabayashi T, Shoji M. Plasma biomarkers of Alzheimer’s disease. Curr Opin Psychiatry 2008;21:260-7.
75. Roher AE, Kuo YM, Esh C, Knebel C, Weiss N, Kalback W, Luehrs DC, Childress JL, Beach TG, Weller RO, Kokjohn
TA. Cortical and leptomeningeal cerebrovascular amyloid and white matter pathology in Alzheimer’s
disease. Mol Med 2003;9:112-22.
76. Zhu YC, Dufouil C, Mazoyer B, Soumare A, Ricolfi F, Tzourio C, Chabriat H. Frequency and Location of Dilated
Virchow-Robin Spaces in Elderly People: A Population-Based 3D MR Imaging Study. AJNR Am J Neuroradiol
2011;32:709-13.
77. Weller RO. How well does the CSF inform upon pathology in the brain in Creutzfeldt-Jakob and Alzheimer’s
diseases? J Pathol 2001;194:1-3.
78. Attems J, Lintner F, Jellinger KA. Amyloid β peptide 1-42 highly correlates with capillary cerebral amyloid
angiopathy and Alzheimer disease pathology. Acta Neuropathol 2004;107:283-91.
79. Verbeek MM, Kremer BP, Olde Rikkert MG, Van Domburg PH, Skehan ME, Greenberg SM. Cerebrospinal fluid
amyloid β(40) is decreased in cerebral amyloid angiopathy. Ann Neurol 2009;66:245-9.
80. Jarrett JT, Berger EP, Lansbury PT, Jr. The carboxy terminus of the β amyloid protein is critical for the seeding
of amyloid formation: implications for the pathogenesis of Alzheimer’s disease. Biochemistry 1993;32:46937.
81. Wiltfang J, Esselmann H, Bibl M, Smirnov A, Otto M, Paul S, Schmidt B, Klafki HW, Maler M, Dyrks T, Bienert M,
Beyermann M, Ruther E, Kornhuber J. Highly conserved and disease-specific patterns of carboxyterminally
truncated Aβ peptides 1-37/38/39 in addition to 1-40/42 in Alzheimer’s disease and in patients with chronic
neuroinflammation. J Neurochem 2002;81:481-96.
82. Preston JE. Ageing choroid plexus-cerebrospinal fluid system. Microsc Res Tech 2001;52:31-7.
162
Why is CSF Aβ42 decreased in AD?
163
Summary and general discussion
Chapter
Eleven
Chapter Eleven
Chapter Eleven
Summary and general discussion
Introduction
Although Alzheimer disease (AD) was discovered more than a century ago, several
important aspects of its pathophysiology remain a mystery today. Its heterogeneous
clinical presentation can make it difficult to differentiate AD from other types of
dementia, especially in the early stages of the disease. This thesis aimed to investigate
the use of cerebrospinal fluid (CSF) biomarkers as diagnostic tools in the differentiation
between different types of dementia. Furthermore, we investigated whether CSF
biomarkers support the current hypotheses on the pathophysiological cascade of AD.
This chapter provides a summary of the findings presented in this thesis, followed by a
general discussion.
Summary
The first part of this thesis focused on the use of CSF biomarkers in clinical practice for
differential diagnostic purposes. We investigated the amyloid β42/amyloid β 40 (Aβ 42/
Aβ 40) ratio in Chapter Two as a means to improve the accuracy of CSF Aβ42 in
discriminating AD from other types of dementia. The concentration of CSF Aβ40 has
been reported to remain unchanged in dementia and to be a measure of an individual’s
total Aβ load. We hypothesized that relating Aβ 42 to the Aβ 40 concentration would
provide a more accurate measure of changes in Aβ42. Indeed we found that the
performance of the Aβ 42/Aβ 40 ratio was better than that of Aβ 42 alone in discriminating
AD from dementia with Lewy bodies (DLB), vascular dementia (VaD) and non-AD
dementia. The Aβ 42/Aβ 40 ratio was similar in diagnostic value compared to Aβ 42 alone
in discriminating AD from frontotemporal dementia (FTD). However, the Aβ 42/Aβ 40 ratio
did not add much to the frequently used combination of Aβ 42, phosphorylated tau
(p-tau) and total tau (t-tau): the latter combination was superior in differentiating AD
from VaD and DLB, and equally good in discriminating AD from FTD or non-AD
dementia. In the differentiation of AD from FTD and non-AD dementia, the Aβ 42/Aβ 40
ratio may provide a more easily interpretable alternative to the combination of Aβ 42,
p-tau and t-tau.
Unexpectedly, we found a decreased concentration of CSF Aβ40 in VaD and DLB.
CSF Aβ 40 concentrations for the different types of dementia have scarcely been reported
separately before.1-3 The cause of the decreased CSF Aβ40 concentration remains to be
elucidated. In DLB, the decrease may be related to an interaction with α-synuclein, the
167
Chapter Eleven
main component of the Lewy bodies that accumulate intraneuronally in DLB: it was
shown that persons with α-synuclein aggregates had a greater Aβ 40 plaque level than
persons without α-synuclein aggregates.4
Thus, in Chapter Three, we hypothesized that the concentration of CSF α-synuclein
would be decreased in DLB as a result of this α-synuclein accumulation, making CSF
α-synuclein a possible biomarker for DLB. We compared the CSF α-synuclein
concentration between groups of patients with different types of dementia.
Unfortunately, no differences were found between DLB, AD, FTD and VaD; moreover,
concentrations were unchanged compared to controls. We concluded that CSF
α-synuclein cannot be advocated as a biomarker for DLB. Our findings were confirmed
by others,5,6 however, contradictory findings followed ours as well7,8 and made us take a
closer look at the properties of CSF α-synuclein. Interactions between Aβ and
α-synuclein have been described in vitro.9,10 Sinusoidal fluctuations were found in CSF
Aβ concentrations11 and our own research group confirmed these findings.12 Perhaps
fluctuations in CSF α-synuclein are induced by its interaction with Aβ and offer an
explanation for the contradicting reports on the value of α-synuclein as a biomarker for
α-synucleinopathies. We studied this in Chapter Four. Six AD patients and six healthy
volunteers received an indwelling intrathecal cathether from which CSF was drawn
every hour. Approximately 33 samples per participant were available for analysis.
However, we observed neither linear trends, nor sinusoidal fluctuations in the CSF
α-synuclein concentrations. Furthermore, the interaction between α-synuclein and Aβ
as observed in vitro was not reflected in a correlation in their CSF concentrations in this
sample, nor in a sample of 164 memory clinic patients.
Our search for new biomarkers did not yield a marker that is ready to be implemented
in clinical practice. Therefore, we decided to explore the application of CSF analysis at
memory clinics and to investigate whether we were able to facilitate the use of the
current CSF biomarkers Aβ 42, p-tau and t-tau in clinical practice. Chapter Five described
the frequency of CSF analysis use in Europe and in the Netherlands. We sent
questionnaires to state-of-the-art memory research centres from the European
Alzheimer’s Disease Consortium (EADC) and to all memory clinics in the Netherlands.
CSF analysis was used in almost 60% of these memory clinics and – in the Netherlands
– in 5% (median) of patients, mostly in patients suspected of mild dementia.
Complications were rare and mild. The analysis sometimes resulted in a change of
working diagnosis, according to almost 30% of memory clinics, which suggests that CSF
analysis is perceived as sufficiently valid to change the diagnostic hypothesis. Reasons
not to perform CSF analysis included a lack of clear recommendations in the guidelines.
168
Summary and general discussion
Guidelines are also unclear on how to interpret the results of the three CSF biomarkers.
For example, is an aberrant concentration of Aβ 42 enough to support a diagnosis of AD,
even if the concentrations of p-tau and t-tau are normal? In Chapter Six we introduced
a prediction model in order to simplify the interpretation of CSF biomarker results in
clinical practice. We are not the first to develop a model, however, previously described
models did not provide sufficient discrimination between AD and non-AD dementia, or
did not offer a formula that could be adopted by others.13,14 Our prediction model
specifically addressed the question of how to differentiate between dementia due to
AD and (presumptive) dementia due to another cause. It was developed on a sample of
over 500 memory clinic patients and validated on an independent dataset containing
almost 500 patients. We used the clinical diagnosis as reference standard. The final
prediction model included sex, CSF Aβ 42 and CSF p-tau and, based on these variables,
estimated the probability that dementia in a memory clinic patient was due to AD.
Adjustments for overfitting and a difference in AD prevalence between de datasets
were made before validation. Inter-laboratory variation in CSF biomarker results was
taken into account as well. The agreement between the probabilities of AD predicted
by our model, and the observed actual diagnoses of AD was very good, as was the
ability of our model to discriminate between AD and non-AD. A major advantage of this
model is that it can be adjusted to other populations, allowing others to use it while
retaining the information of over 1,000 patients.
The second part of this thesis investigated whether hypotheses on the pathophysiological cascade of AD are supported by the results of CSF biomarker analysis. The
hypothetical AD biomarker model proposed by Jack et al15 assumes that CSF and
imaging biomarkers predate clinical symptoms by an unknown number of years and
proposes a clear distinction between the moment that Aβ and tau become abnormal.
Aβ is suggested to reach a plateau before clinical symptoms are apparent, while tau, as
a marker of neurodegeneration, increases in abnormality during the early stages of the
disease and into the dementia stage. In Chapter Seven we described a patient who
presented with a mild memory disorder without interference with activities of daily
living. The CSF biomarkers that were analysed at that moment were already clearly
abnormal, Aβ 42 as well as p-tau and t-tau. We followed this patient for nine years, during
which he progressed to advanced AD dementia. A second lumbar puncture at that time
showed that the CSF biomarkers were still clearly abnormal. These results, albeit from
one person, suggest that either there is no delay between changes in Aβ and tau, or
that the change in tau occurs already in the preclinical stage. This suggestion is
169
Chapter Eleven
supported by findings from studies in MCI patients that show that both CSF Aβ42 and
tau are already abnormal in those patients who progressed to AD at follow up.16
We decided to further investigate the model by Jack et al. The model proposes a
gradation of abnormality in the biomarkers that corresponds to the severity of AD
neuropathology, which in turn corresponds to the severity of cognitive impairment.
With this concept in mind, we investigated whether a correlation between CSF
biomarkers and cognitive performance could be found (Chapter Eight). We
hypothesized that abnormal concentrations of the CSF biomarkers are a sign of neuronal
damage that should result in cognitive impairment. We examined 65 memory clinic
patients with MCI, dementia due to AD or non-AD dementia, using the CAMCOG-R to
assess multiple domains of cognitive functioning. We found that concentrations of CSF
Aβ 42, p-tau and t-tau had no clear relationship with cognitive functioning. Possibly,
abnormal CSF biomarker concentrations are not a measure of neuronal damage, but
merely serve as an indicator of AD pathology.
Another pathophysiological hypothesis is the vascular hypothesis. It proposes that
cerebrovascular disease leads to APP upregulation with Aβ deposition as a result,17-19
while Aβ in itself may lead to cerebrovascular disease via its negative effects on cerebrovascular function and the architecture of the vascular wall. 20-22 If this is the case, one
would expect AD and cerebrovascular disease to go hand in hand: AD patients are
expected to display signs of cerebrovascular disease, whereas patients with cerebrovascular disease may not necessarily develop clinical AD but should at least display signs of
Aβ deposition such as a decreased concentration of Aβ in CSF. We tested this theory in
Chapter Nine. We classified 81 memory clinic patients as AD or non-AD based on the
research criteria for AD. These criteria define AD (the disease, not the dementia per se)
as an episodic memory disorder combined with at least one positive biomarker. We
assessed cerebrovascular disease on MRI by visually rating white matter lesions and
lacunar infarcts.23 Aβ 42 and Aβ 40 were analysed in CSF. We found that AD was not
associated with cerebrovascular disease. The concentrations of CSF Aβ 42 and Aβ 40 were
unrelated to the presence or severity of cerebrovascular disease as well. The relationship
between vascular risk factors and cerebrovascular disease was not influenced by
whether or not patients had AD, nor by their concentration of CSF Aβ42 or Aβ 40. These
results do not support the theory that cerebrovascular disease leads to AD pathology or
vice versa.
In addition to its use as an indicator of disease, CSF Aβ 42 has also been recommended
as a biomarker to detect disease modification in drug trials.24,25 However, to be able to
make predictions about what will happen with the CSF Aβ 42 concentration when a
170
Summary and general discussion
treatment is disease-modifying, it is important to first understand why the concentration
is decreased in AD. The decreased CSF Aβ 42 concentration in AD is often regarded as a
consequence of Aβ 42 deposition in the brain. However, Aβ 40 is deposited as well, yet
the CSF Aβ 40 concentration is unaltered in AD. In Chapter Ten we systematically
evaluated possible reasons for the decreased CSF Aβ42 concentration in AD. Aggregation
of Aβ 42 remained as the most likely cause, the aggregated state inhibiting Aβ42 from
being transported from the interstitial fluid (ISF) to the CSF. The amount of Aβ40 that is
deposited is far less than the amount of Aβ 42, while the absolute concentration of Aβ 40
is much higher.26 Possibly, the amount of Aβ 40 deposited relative to the absolute amount
present is too small to induce a clear decrease in the CSF Aβ 40 concentration.
General discussion
This thesis described translational research with CSF biomarkers. Translational research
can be defined as the process of applying discoveries generated during research in the
laboratory and in preclinical studies to clinical studies in humans. Vice versa, findings
from clinical studies can be used to direct preclinical research.27 In our work, new CSF
biomarkers that were inspired by preclinical science were investigated in a clinical
population, and clinical observations were used to test pathophysiological hypotheses.
Our studies mostly yielded negative results, i.e. the preclinical assumptions could not be
confirmed in clinical data. Two possible explanations for these negative results will be
discussed here: the hypotheses were false, or the methodology used to test them was
flawed, limiting the validity of the results.
Hypotheses
In general, hypotheses are testable ideas derived from theories that are usually based
on observations from explorative studies. In formulating these theories, several
assumptions are – consciously or subconsciously – made (Figure 1). If a hypothesis is
rejected by observations from research, it may indicate that one of the elements
underlying the theory was incorrect.
As an example, let us consider the hypothesis we investigated in Chapter Three. In
this chapter, we hypothesized that the CSF concentration of α-synuclein in synucleinopathies such as DLB would be decreased in comparison with controls and with
patients with dementia without α-synuclein inclusions, such as AD, VaD and FTD. Our
hypothesis was based on the theory that the pathophysiological changes in DLB
171
Chapter Eleven
Figure 1 The relation between observations, assumptions, theory and hypothesis
Assumptions
Observations
Theory
Hypothesis
influence the α-synuclein concentration in the brain. Observations used to found this
theory were: 1) the finding that Lewy bodies, found upon neuropathological
examination in DLB patients, consist of α-synuclein28 and 2) research that showed that
patients with Parkinson disease, in whom Lewy bodies are present as well, had
significantly lower α-synuclein levels in their CSF than controls.29 Assumptions that
underlay our theory were the idea that the observations made post-mortem would also
be present durante vivam, the thought that findings in Parkinson disease could be
extrapolated to DLB, as well as the suggestion that the CSF α-synuclein concentration
reflects changes in the α-synuclein concentration in the brain. Had we extended our
theory by assuming that α-synuclein accumulation in DLB was a result of overproduction
of α-synuclein, we might have speculated that this overproduction would have masked
a decrease in the CSF α-synuclein concentration, and thus we might have hypothesized
that there was no difference in CSF α-synuclein concentrations between DLB and other
dementias. Alternatively, we might have hypothesized an íncrease in the CSF α-synuclein
concentration due to its accumulation in the brain in DLB, if we had not assumed
that the decreased concentration that was found in Parkinson disease would be
present in DLB as well. Different assumptions thus would have led to different theories
and to different hypotheses. This underlines the importance of testing hypotheses.
Furthermore, it indicates that being aware of the assumptions that are made may help
to interpret negative results as well as provide areas for further research.
We did not find support for our hypothesis that the CSF α-synuclein concentration
is decreased in DLB. This may indicate that our theory was incorrect and therefore
172
Summary and general discussion
that either the observations or the assumptions underlying the theory were incorrect.
One may argue that observations from previous research cannot be incorrect since
they are facts. Although semantically true, in reality observations need to be interpreted
before they can be used, and previous research is hardly ever unambiguous. In addition,
it appears that researchers are subjective in deciding which studies they cite: studies on
biomarkers that report larger effect estimates for postulated associations are cited more
often than meta-analyses that evaluate the same associations but that report smaller
effect estimates.30 The ways previous observations are valued thus importantly affect
the creation of theories.
What do CSF biomarkers reflect?
Contrary to our assumption, it appears that the CSF α-synuclein concentration in DLB
patients does not reflect the changes in α-synuclein in the brain – although alternative
explanations for the lack of change are not ruled out, for example overproduction of
α-synuclein masking a decrease in its CSF concentration. One way or the other, it leads
to an interesting question: what exactly do CSF biomarkers reflect? The answer to this
question is important since results from biomarker studies are used to make inferences
about the pathophysiology of AD.15,25,31
The CSF biomarkers for AD are usually interpreted to reflect the presence of Aβ
plaques and neurofibrillary tangles, but studies to confirm this are scarce and their
results are contradicting. One study that reported a correlation between CSF Aβ42 and
Aβ plaques and between CSF tau and neurofibrillary tangles32 was supported by
moderate correlations found between ventricular CSF Aβ 42 and Aβ plaques;33 between
CSF p-tau231 and neurofibrillary tangles;34 and between CSF tau and neurofibrillary
tangles.35 Other studies, however, found no correlation between CSF Aβ 42, p-tau181 and
t-tau and Aβ plaques or neurofibrillary tangles.36,37
If we consider the CSF biomarkers for AD and our findings on CSF α-synuclein, we
see that:
• intraneuronal tau accumulation in AD is associated with an increase in CSF p-tau
and t-tau;
• extraneuronal Aβ deposition in AD is associated with a decrease in CSF Aβ 42;
• intraneuronal α-synuclein accumulation in DLB is not associated with a change in
CSF α-synuclein.
173
Chapter Eleven
It seems inconsistent that intraneuronal accumulation of α-synuclein would not lead to
a change in its CSF concentration, while intraneuronal accumulation of tau would
increase its CSF concentration. Furthermore, one may wonder why extraneuronal
accumulation of Aβ would lead to a decrease in its CSF concentration, while
intraneuronal accumulation of tau would lead to an increase in its CSF concentration.
The suggestion that the increased tau concentration is the result of degenerating
neurons spilling their tau content into the CSF is not in line with the suggestion that tau
biomarkers are abnormal long before changes in brain structure become apparent.15
This ambiguity triggered us to review reasons for the decreased CSF Aβ 42
concentration in AD (Chapter Ten). From this study it became apparent that there are
still several aspects of normal Aβ metabolism that are unclear and that prevent a
thorough understanding of its mismetabolism in AD. For example, we are not certain
whether Aβ can pass the pia mater or the arachnoid granulations, which is important
because such a route of Aβ transportation in and out of the CSF could have a strong
influence on the CSF Aβ concentration. We also do not know what amount of Aβ is
degraded or what amount is transported from the ISF to the blood, to the lymphatics or
the CSF. Without this kind of knowledge, it is difficult to understand what changes take
place in AD. It is likely that CSF biomarker concentrations do not provide a straightforward reflection of the situation in the brain, but are the end result of a complicated
interplay of production, degradation, transport, clearance and, in AD, aggregation and
accumulation. The resulting discrepancy between the brain concentration of a protein
and its concentration in the CSF may account for the fact that we were unable to
observe a correlation between CSF α-synuclein and CSF Aβ, despite their interaction in
vitro (Chapter Four). Furthermore, if the CSF biomarker concentrations are not a precise
reflection of neuropathological changes in the brain, this may explain the lack of
correlation between CSF biomarkers and cognitive functioning (Chapter Eight).
Methodology
Diagnostic studies such as we performed in Chapter Two and Three may suffer from
different sources of bias and variation that may affect the estimated accuracy of the test
under study.38-41 For example, retrospective data collection and non-consecutive
inclusion of patients were shown to lead to an overestimation of accuracy, whereas
including only patients that were referred for the test under the study (rather than
including them based on clinical symptoms) led to lower estimates.40 These sources of
bias and variation may have affected our studies as well. We will discuss their presence
and possible effect on our research results below.
174
Summary and general discussion
Population
Estimates of diagnostic accuracy may differ between populations.42 In translational
research, testing a hypothesis in the population of interest – in our case, the memory
clinic population – is the final step, since important differences often exist between the
population of interest and the population that led to formulating the hypothesis. No
one will dispute the differences between a transgenic mouse and an AD patient,
however, differences also exist between selected research populations and the patient
population generally seen at a memory clinic. Biomarker performance is easily
overestimated if such performance is solely based on its ability to differentiate between
AD patients and healthy controls: the inclusion of healthy controls likely decreases the
number of false-positives and increases the specificity of the tested biomarker.39,40
Moreover, the AD patients included in explorative studies usually fulfil strict inclusion
criteria, e.g. they do not suffer from severe co-morbidities and concomitant medication
use is restricted. Thus, these AD patients resemble only a small part of the memory clinic
population, whereas truly healthy controls do not represent a part of this population at
all since they never present themselves at a memory clinic. In clinical reality, the differentiation is between AD and other causes of cognitive impairment such as other types
of dementia, psychiatric disorders, psychosocial stress, or subjective memory complaints.
The patients that we included in our studies were derived from the CSF database of
the Alzheimer Centre of the Radboud University Nijmegen Medical Centre that contains
clinical data as well as CSF of consecutive patients. Most patients in this database
underwent lumbar puncture because there was a clinical indication, that is, if the clinical
picture was unclear or atypical, and additional examinations were therefore desired.
This means that we included a group of patients that may have been relatively difficult
to diagnose clinically. These patients may differ from patients with a clear clinical picture
that do not need such additional examinations; they may for instance have mixed
pathologies more often. As a result, differences between the types of dementia may be
less pronounced, which may have led to an underestimation of the performance of our
biomarkers. However, such is clinical practice. In many patients, the clinical picture
alone is not sufficient to make a diagnosis. One could argue that this selection of
patients is exactly the population that should be included in biomarker studies: these
are the patients for whom new biomarkers are most valuable, since it is already
established that additional examinations are needed to arrive at a diagnosis. Our efforts
to differentiate between these patients thus truly reflect the challenge of clinical practice.
The fact that our population includes more atypical patients and possibly
overrepresents patients with mixed pathologies may have complicated the investigation
175
Chapter Eleven
of pathophysiological hypotheses: associations may be less strong in atypical patients,
or be obscured because more than one disease was present. We have attempted to
minimize this effect by using biomarker-based diagnoses in our pathophysiological
studies, reasoning that biomarkers are a more objective measure than clinical diagnosis.
To prevent circular reasoning, we took care that we did not define our diagnoses based
on the biomarker under study.
Disease progression bias
Disease progression bias may occur if the diagnostic test under study is performed a
long time before the reference standard, during which period the disease progresses:41
the test under study may be negative while the reference standard, performed much
later, is positive. As a result, the accuracy of the test under study is underestimated. In
our CSF database, the lumbar puncture and the reference standard (see below) were
part of the same diagnostic process, ruling out disease progression bias. In addition, in
case of Aβ 42, p-tau and t-tau, this bias is less relevant, since it was shown that CSF
Aβ 42, p-tau and t-tau are abnormal and stable in MCI patients who progressed to AD at
follow up.16
Reference standard and verification bias
Partial verification bias occurs when only a selected sample of patients that underwent
the test under study is verified by the reference standard.41 In our CSF database, the
reference standard is the clinical diagnosis. For a small proportion of the patients, the
clinical diagnosis was unclear, and therefore these patients were not included in our
studies. If only patients with normal biomarker concentrations (i.e., negative test results)
had had an unknown diagnosis, an overestimation of the sensitivity of the biomarker
could have occurred, since patients with false-negative test results would go unnoticed.
However, whether or not the diagnosis was known in our database was independent of
the results of CSF analysis. Thus, the effect of this bias is probably minimal.
Differential verification bias occurs when part of the patients that underwent the
test under study were verified by a different reference standard, and the different
reference standards are treated as interchangeable.38,41 Our reference standard is the
clinical diagnosis. This is not a single test, but the result of a diagnostic process that
includes history taking, physical examination, neuropsychological assessment and
often neuroimaging. This process may differ per patient, which may be regarded as
differential verification bias: not every patient received exactly the same reference
standard. This bias is however impossible to avoid in dementia research. Dementia
176
Summary and general discussion
diagnoses are always composed of multiple assessments and never based on a single
test with a clear cut-off; no perfect reference standard exists for dementia. The neuropathological diagnosis is often considered the ultimate truth; however, this can be
disputed. It turns out that neuropathological criteria have not been standardized and
that inter-observer reliability is low when the clinical picture is unclear.43 In addition,
efforts are currently undertaken to re(de)fine the neuropathological criteria for AD;44,45
thus the standard that many consider as definite proof of AD is called into question. It
has been suggested that neuropathology should instead be used as a biomarker, in
addition to the clinical picture and other biomarkers.43,46 This way, the most reliable
‘true’ AD diagnosis is derived from the integration of many sources of information,
ideally complemented with data from follow up.
By using the clinical diagnosis as reference standard in our database, some mis­
classifications probably have occurred. Such misclassifications may have contaminated
the different groups of dementia patients, thereby obscuring their differences. However,
when we compared the AD, VaD, DLB and FTD patients included in our studies to the
patients described in other studies, based on their mean concentrations of CSF Aβ 42,
p-tau and t-tau, they were similar. Likewise, the sensitivity and specificity of CSF Aβ42 for
the discrimination of AD from other types of dementia as reported in Chapter Two was
comparable to findings from other studies. Thus, it appears that our study population is
not that different from the populations that have been investigated in other studies. It
is of course not ruled out that these studies included the same selection of patients or
suffered from the same misclassifications. A recent report, however, may support the
validity of a clinical diagnosis. In this large cohort of patients followed over many years,
the clinical diagnosis from an academic centre accurately predicted the ‘definite’
classification based on neuropathology post-mortem.47 The fact that our population
was comparable to the literature makes us confident that we were able to discover any
differences between groups and strong relations between disease characteristics that
might have been present.
Interpretation of the diagnostic tests
The interpretation of the test under study may be biased if it is influenced by knowledge
of the results of the reference standard or knowledge of clinical data. In our studies, we
have prevented this by using pre-defined cut-off values of the CSF biomarkers whenever
possible. If the test under study is used to establish the reference standard, so-called
incorporation bias will increase the agreement between the tests. Some incorporation
bias probably has occurred in our database now CSF Aβ 42, p-tau and tau are increasingly
177
Chapter Eleven
used in memory clinics. In our diagnostic studies, however, the biomarkers under study
were not used to establish the diagnosis.
Indeterminable test results
Indeterminable test results are often omitted from the research analysis and this may
lead to a biased assessment of test characteristics. In our database, CSF biomarker or
α-synuclein concentrations were below the detection limit in a few cases. To prevent
loss of information, the lowest detectable concentration was used in these cases.
Having considered all these sources of bias, we may conclude that our studies are
probably most affected by their retrospective nature, the imperfect reference standard,
and, to some extent, by incorporation bias. When aiming to perform research in a
population that is true to clinical practice, it is difficult to avoid these types of bias.
However, these types of bias usually result in an overestimation of diagnostic
accuracy.39,40 Considering that our studies were negative, avoiding these types of bias
would probably not have changed our results.
Clinical application
Even though we may not yet fully grasp the ins and outs of our CSF biomarkers, their
role in clinical practice can be clearly defined. They can be used as a diagnostic tool in
cases of diagnostic uncertainty. There is no need to perform a lumbar puncture in every
patient attending a memory clinic – not for diagnostic purposes at least. The clinical
picture, combined with neuropsychological assessment or neuroimaging is often clear
enough to make a diagnosis. However, when dementia is suspected but the aetiology
is unclear, analysis of CSF Aβ 42 and p-tau may be a way to gain more certainty about
whether or not AD is present, as shown with our prediction model. We have shown that
clinicians are willing to use lumbar puncture as a diagnostic tool (Chapter Five) and with
the development of our prediction model we have provided a way to further implement
CSF biomarkers in clinical practice. The model has only been validated in the population
suspected of dementia. It may be able to predict which MCI patients will convert to AD,
but, true to translational research, this validation step should be awaited before we can
conclude that it does.
178
Summary and general discussion
Implications for future research
The next step in CSF biomarker research should be to gain a better understanding of
the metabolism of these biomarkers and the changes therein in AD. This includes details
about transport from ISF to CSF, flow to lymphatics and transport across the blood-brain
barrier and blood-CSF barrier. The fact that the term blood-brain barrier is often used to
indicate the blood-CSF barrier underlines the current lack of understanding of fluid
circulation in the brain. Although not knowing these details does not prevent us from
using the CSF biomarkers in clinical practice, better understanding of these biomarkers
is an essential step to unravel AD pathophysiology and to develop and evaluate
treatment. Especially CSF Aβ 42 has been advocated as a marker to detect and monitor
the biochemical effects of newly developed drugs for AD.24,25 However, so far most
articles that report AD drug trials fail to formulate a hypothesis that a priori stipulates
the expected change (in magnitude, duration and especially direction) in the CSF Aβ
concentration if treatment is disease-modifying.48-54 Without specifying the expected
outcome, it is difficult to claim a disease-modifying treatment. The development of
successful treatment for AD will be greatly advanced through better knowledge of the
underlying pathophysiology of AD and its biomarkers.
Conclusions
CSF biomarkers can predict the probability of AD in patients suspected of dementia and
are a valuable diagnostic tool in clinical practice. Applying them to the population of
interest is essential to translate research finding to clinical practice. By being aware of
the assumptions that are made in this process, areas for further research were uncovered
in this thesis. So far, CSF biomarker concentrations do not provide a straightforward
reflection of the situation in the brain. Better understanding of their (mis)metabolism
will provide the key to understanding AD.
179
Chapter Eleven
References
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 180
Bibl M, Mollenhauer B, Esselmann H, Lewczuk P, Trenkwalder C, Brechlin P, Ruther E, Kornhuber J, Otto M,
Wiltfang J. CSF diagnosis of Alzheimer’s disease and dementia with Lewy bodies. J Neural Transm
2006;113:1771-8.
Bibl M, Mollenhauer B, Esselmann H, Schneider M, Lewczuk P, Welge V, Gross M, Falkai P, Kornhuber J,
Wiltfang J. Cerebrospinal fluid neurochemical phenotypes in vascular dementias: original data and
mini-review. Dement Geriatr Cogn Disord 2008;25:256-65.
Kanai M, Matsubara E, Isoe K, Urakami K, Nakashima K, Arai H, Sasaki H, Abe K, Iwatsubo T, Kosaka T, Watanabe
M, Tomidokoro Y, Shizuka M, Mizushima K, Nakamura T, Igeta Y, Ikeda Y, Amari M, Kawarabayashi T, Ishiguro
K, Harigaya Y, Wakabayashi K, Okamoto K, Hirai S, Shoji M. Longitudinal study of cerebrospinal fluid levels of
tau, A beta1-40, and A beta1-42(43) in Alzheimer’s disease: a study in Japan. Ann Neurol 1998;44:17-26.
Lippa SM, Lippa CF, Mori H. Alpha-Synuclein aggregation in pathological aging and Alzheimer’s disease: the
impact of beta-amyloid plaque level. Am J Alzheimers Dis Other Demen 2005;20:315-8.
Aerts MB, Esselink RA, Abdo WF, Bloem BR, Verbeek MM. CSF alpha-synuclein does not differentiate
between parkinsonian disorders. Neurobiol Aging 2011.
Reesink FE, Lemstra AW, van Dijk KD, Berendse HW, van de Berg WD, Klein M, Blankenstein MA, Scheltens P,
Verbeek MM, van der Flier WM. CSF alpha-synuclein does not discriminate dementia with Lewy bodies from
Alzheimer’s disease. J Alzheimers Dis 2010;22:87-95.
Hong Z, Shi M, Chung KA, Quinn JF, Peskind ER, Galasko D, Jankovic J, Zabetian CP, Leverenz JB, Baird G,
Montine TJ, Hancock AM, Hwang H, Pan C, Bradner J, Kang UJ, Jensen PH, Zhang J. DJ-1 and alpha-synuclein
in human cerebrospinal fluid as biomarkers of Parkinson’s disease. Brain 2010;133:713-26.
Mollenhauer B, Locascio JJ, Schulz-Schaeffer W, Sixel-Doring F, Trenkwalder C, Schlossmacher MG. alphaSynuclein and tau concentrations in cerebrospinal fluid of patients presenting with parkinsonism: a cohort
study. Lancet Neurol 2011;10:230-40.
Kazmierczak A, Strosznajder JB, Adamczyk A. alpha-Synuclein enhances secretion and toxicity of amyloid
beta peptides in PC12 cells. Neurochem Int 2008;53:263-9.
Mandal PK, Pettegrew JW, Masliah E, Hamilton RL, Mandal R. Interaction between Abeta peptide and alpha
synuclein: molecular mechanisms in overlapping pathology of Alzheimer’s and Parkinson’s in dementia
with Lewy body disease. Neurochem Res 2006;31:1153-62.
Bateman RJ, Wen G, Morris JC, Holtzman DM. Fluctuations of CSF amyloid-beta levels: implications for a
diagnostic and therapeutic biomarker. Neurology 2007;68:666-9.
Slats D, Claassen JA, Spies PE, Borm G, Besse KT, Aalst WV, Tseng J, Sjogren MJ, Olde Rikkert MG, Verbeek MM.
Hourly variability of cerebrospinal fluid biomarkers in Alzheimer’s disease subjects and healthy older
volunteers. Neurobiol Aging 2011.
De Meyer G, Shapiro F, Vanderstichele H, Vanmechelen E, Engelborghs S, De Deyn PP, Coart E, Hansson O,
Minthon L, Zetterberg H, Blennow K, Shaw L, Trojanowski JQ. Diagnosis-independent Alzheimer disease
biomarker signature in cognitively normal elderly people. Arch Neurol 2010;67:949-56.
Hulstaert F, Blennow K, Ivanoiu A, Schoonderwaldt HC, Riemenschneider M, De Deyn PP, Bancher C, Cras P,
Wiltfang J, Mehta PD, Iqbal K, Pottel H, Vanmechelen E, Vanderstichele H. Improved discrimination of AD
patients using beta-amyloid(1-42) and tau levels in CSF. Neurology 1999;52:1555-62.
Jack CR, Jr., Knopman DS, Jagust WJ, Shaw LM, Aisen PS, Weiner MW, Petersen RC, Trojanowski JQ.
Hypothetical model of dynamic biomarkers of the Alzheimer’s pathological cascade. Lancet Neurol
2010;9:119-28.
Zetterberg H, Pedersen M, Lind K, Svensson M, Rolstad S, Eckerstrom C, Syversen S, Mattsson UB, Ysander C,
Mattsson N, Nordlund A, Vanderstichele H, Vanmechelen E, Jonsson M, Edman A, Blennow K, Wallin A. Intraindividual stability of CSF biomarkers for Alzheimer’s disease over two years. J Alzheimers Dis 2007;12:255-60.
Claassen JA. Alzheimer’s disease: trade-off for increased survival with atherosclerosis? Ann Neurol
2008;64:475.
Kalaria RN. The role of cerebral ischemia in Alzheimer’s disease. Neurobiol Aging 2000;21:321-30.
Kalaria RN. Linking cerebrovascular defense mechanisms in brain ageing and Alzheimer’s disease. Neurobiol
Aging 2009;30:1512-4.
Summary and general discussion
20. Meyer EP, Ulmann-Schuler A, Staufenbiel M, Krucker T. Altered morphology and 3D architecture of brain
vasculature in a mouse model for Alzheimer’s disease. Proc Natl Acad Sci U S A 2008;105:3587-92.
21. Niwa K, Kazama K, Younkin L, Younkin SG, Carlson GA, Iadecola C. Cerebrovascular autoregulation is
profoundly impaired in mice overexpressing amyloid precursor protein. Am J Physiol Heart Circ Physiol
2002;283:H315-H323.
22. Thomas T, Thomas G, McLendon C, Sutton T, Mullan M. beta-Amyloid-mediated vasoactivity and vascular
endothelial damage. Nature 1996;380:168-71.
23. Wahlund LO, Barkhof F, Fazekas F, Bronge L, Augustin M, Sjogren M, Wallin A, Ader H, Leys D, Pantoni L,
Pasquier F, Erkinjuntti T, Scheltens P. A new rating scale for age-related white matter changes applicable to
MRI and CT. Stroke 2001;32:1318-22.
24. Hampel H, Frank R, Broich K, Teipel SJ, Katz RG, Hardy J, Herholz K, Bokde AL, Jessen F, Hoessler YC, Sanhai
WR, Zetterberg H, Woodcock J, Blennow K. Biomarkers for Alzheimer’s disease: academic, industry and
regulatory perspectives. Nat Rev Drug Discov 2010;9:560-74.
25. Blennow K, Hampel H, Weiner M, Zetterberg H. Cerebrospinal fluid and plasma biomarkers in Alzheimer
disease. Nat Rev Neurol 2010;6:131-44.
26. Wiltfang J, Esselmann H, Bibl M, Smirnov A, Otto M, Paul S, Schmidt B, Klafki HW, Maler M, Dyrks T, Bienert M,
Beyermann M, Ruther E, Kornhuber J. Highly conserved and disease-specific patterns of carboxyterminally
truncated Abeta peptides 1-37/38/39 in addition to 1-40/42 in Alzheimer’s disease and in patients with
chronic neuroinflammation. J Neurochem 2002;81:481-96.
27. Rubio DM, Schoenbaum EE, Lee LS, Schteingart DE, Marantz PR, Anderson KE, Platt LD, Baez A, Esposito K.
Defining translational research: implications for training. Acad Med 2010;85:470-5.
28. Ueda K, Fukushima H, Masliah E, Xia Y, Iwai A, Yoshimoto M, Otero DA, Kondo J, Ihara Y, Saitoh T. Molecular
cloning of cDNA encoding an unrecognized component of amyloid in Alzheimer disease. Proc Natl Acad
Sci U S A 1993;90:11282-6.
29. Tokuda T, Salem SA, Allsop D, Mizuno T, Nakagawa M, Qureshi MM, Locascio JJ, Schlossmacher MG, El-Agnaf
OM. Decreased alpha-synuclein in cerebrospinal fluid of aged individuals and subjects with Parkinson’s
disease. Biochem Biophys Res Commun 2006;349:162-6.
30. Ioannidis JP, Panagiotou OA. Comparison of effect sizes associated with biomarkers reported in highly cited
individual articles and in subsequent meta-analyses. JAMA 2011;305:2200-10.
31. Jack CR, Jr., Vemuri P, Wiste HJ, Weigand SD, Aisen PS, Trojanowski JQ, Shaw LM, Bernstein MA, Petersen RC,
Weiner MW, Knopman DS. Evidence for Ordering of Alzheimer Disease Biomarkers. Arch Neurol 2011.
32. Tapiola T, Alafuzoff I, Herukka SK, Parkkinen L, Hartikainen P, Soininen H, Pirttila T. Cerebrospinal fluid
{beta}-amyloid 42 and tau proteins as biomarkers of Alzheimer-type pathologic changes in the brain. Arch
Neurol 2009;66:382-9.
33. Strozyk D, Blennow K, White LR, Launer LJ. CSF Abeta 42 levels correlate with amyloid-neuropathology in a
population-based autopsy study. Neurology 2003;60:652-6.
34. Buerger K, Ewers M, Pirttila T, Zinkowski R, Alafuzoff I, Teipel SJ, DeBernardis J, Kerkman D, McCulloch C,
Soininen H, Hampel H. CSF phosphorylated tau protein correlates with neocortical neurofibrillary pathology
in Alzheimer’s disease. Brain 2006;129:3035-41.
35. Tapiola T, Overmyer M, Lehtovirta M, Helisalmi S, Ramberg J, Alafuzoff I, Riekkinen P, Sr., Soininen H. The level of
cerebrospinal fluid tau correlates with neurofibrillary tangles in Alzheimer’s disease. Neuroreport 1997;8:3961-3.
36. Buerger K, Alafuzoff I, Ewers M, Pirttila T, Zinkowski R, Hampel H. No correlation between CSF tau protein
phosphorylated at threonine 181 with neocortical neurofibrillary pathology in Alzheimer’s disease. Brain
2007;130:e82.
37. Engelborghs S, Sleegers K, Cras P, Brouwers N, Serneels S, De LE, Martin JJ, Vanmechelen E, Van BC, De Deyn
PP. No association of CSF biomarkers with APOEepsilon4, plaque and tangle burden in definite Alzheimer’s
disease. Brain 2007;130:2320-6.
38. de Groot JA, Bossuyt PM, Reitsma JB, Rutjes AW, Dendukuri N, Janssen KJ, Moons KG. Verification problems
in diagnostic accuracy studies: consequences and solutions. BMJ 2011;343:d4770.
181
Chapter Eleven
39. Lijmer JG, Mol BW, Heisterkamp S, Bonsel GJ, Prins MH, van der Meulen JH, Bossuyt PM. Empirical evidence
of design-related bias in studies of diagnostic tests. JAMA 1999;282:1061-6.
40. Rutjes AW, Reitsma JB, Di NM, Smidt N, van Rijn JC, Bossuyt PM. Evidence of bias and variation in diagnostic
accuracy studies. CMAJ 2006;174:469-76.
41. Whiting P, Rutjes AW, Reitsma JB, Glas AS, Bossuyt PM, Kleijnen J. Sources of variation and bias in studies of
diagnostic accuracy: a systematic review. Ann Intern Med 2004;140:189-202.
42. Bossuyt PM. Clinical validity: defining biomarker performance. Scand J Clin Lab Invest Suppl 2010;242:46-52.
43. Scheltens P, Rockwood K. How golden is the gold standard of neuropathology in dementia? Alzheimers
Dement 2011;7:486-9.
44. National Institute on Aging-Alzheimer’s Association guidelines for the neuropathologic assessment of
Alzheimer’s disease. http://www.alz.org/documents_custom/neuropathologic.pdf. National Institute on
Aging 2011.
45. New Neuropathologic Guidelines for AD Almost Ready for Primetime. http://www.alzforum.org/new/
detail.asp?id=2902. Alzheimer Research Forum 2011.
46. Rockwood K, Bouchard RW, Camicioli R, Leger G. Toward a revision of criteria for the dementias. Alzheimers
Dement 2007;3:428-40.
47. Snowden JS, Thompson JC, Stopford CL, Richardson AM, Gerhard A, Neary D, Mann DM. The clinical
diagnosis of early-onset dementias: diagnostic accuracy and clinicopathological relationships. Brain
2011;134:2478-92.
48. Siemers ER, Friedrich S, Dean RA, Gonzales CR, Farlow MR, Paul SM, Demattos RB. Safety and changes in
plasma and cerebrospinal fluid amyloid beta after a single administration of an amyloid beta monoclonal
antibody in subjects with Alzheimer disease. Clin Neuropharmacol 2010;33:67-73.
49. Gilman S, Koller M, Black RS, Jenkins L, Griffith SG, Fox NC, Eisner L, Kirby L, Rovira MB, Forette F, Orgogozo
JM. Clinical effects of Abeta immunization (AN1792) in patients with AD in an interrupted trial. Neurology
2005;64:1553-62.
50. Aisen PS, Saumier D, Briand R, Laurin J, Gervais F, Tremblay P, Garceau D. A Phase II study targeting
amyloid-beta with 3APS in mild-to-moderate Alzheimer disease. Neurology 2006;67:1757-63.
51. Lannfelt L, Blennow K, Zetterberg H, Batsman S, Ames D, Harrison J, Masters CL, Targum S, Bush AI, Murdoch
R, Wilson J, Ritchie CW. Safety, efficacy, and biomarker findings of PBT2 in targeting Abeta as a modifying
therapy for Alzheimer’s disease: a phase IIa, double-blind, randomised, placebo-controlled trial. Lancet
Neurol 2008;7:779-86.
52. Dodel RC, Du Y, Depboylu C, Hampel H, Frolich L, Haag A, Hemmeter U, Paulsen S, Teipel SJ, Brettschneider
S, Spottke A, Nolker C, Moller HJ, Wei X, Farlow M, Sommer N, Oertel WH. Intravenous immunoglobulins
containing antibodies against beta-amyloid for the treatment of Alzheimer’s disease. J Neurol Neurosurg
Psychiatry 2004;75:1472-4.
53. Relkin NR, Szabo P, Adamiak B, Burgut T, Monthe C, Lent RW, Younkin S, Younkin L, Schiff R, Weksler ME.
18-Month study of intravenous immunoglobulin for treatment of mild Alzheimer disease. Neurobiol Aging
2009;30:1728-36.
54. Galasko DR, Graff-Radford N, May S, Hendrix S, Cottrell BA, Sagi SA, Mather G, Laughlin M, Zavitz KH, Swabb
E, Golde TE, Murphy MP, Koo EH. Safety, tolerability, pharmacokinetics, and Abeta levels after short-term
administration of R-flurbiprofen in healthy elderly individuals. Alzheimer Dis Assoc Disord 2007;21:292-9.
182
Summary and general discussion
183
Nederlandse samenvatting
Chapter
Twelve
Chapter Twelve
Chapter Twelve
Nederlandse samenvatting
Nederlandse samenvatting
Hoewel de ziekte van Alzheimer (AD) al meer dan een eeuw geleden voor het eerst
beschreven werd, zijn veel aspecten van het ontstaan van deze ziekte nog onduidelijk.
Het klinisch beeld is heterogeen; daardoor kan het moeilijk zijn om AD te onderscheiden
van andere vormen van dementie, met name in de beginstadia van de ziekte. In dit
proefschrift hebben we onderzocht of het bepalen van biomarkers in de liquor (CSF,
cerebrospinal fluid) kan helpen bij het differentiëren tussen verschillende vormen van
dementie. Daarnaast hebben we onderzocht of deze CSF biomarkers de huidige
hypotheses over de pathofysiologie van AD ondersteunen. In dit hoofdstuk worden de
bevindingen van dit onderzoek samengevat.
In Hoofdstuk Twee hebben we de CSF amyloid β42/amyloid β 40 (Aβ 42/Aβ 40) ratio
onderzocht. Aβ 40 wordt gezien als een maat voor de totale hoeveelheid Aβ van een
individu en volgens eerder onderzoek verandert de Aβ 40 concentratie niet in patiënten
met dementie. Wij verwachtten daarom dat we een preciezere weergave van de
verandering in Aβ 42 concentratie zouden krijgen wanneer we de Aβ 42 concentratie
zouden afzetten tegen de Aβ 40 concentratie. Met deze preciezere maat zouden we
beter moeten kunnen differentiëren tussen de verschillende vormen van dementie. De
Aβ 42/Aβ 40 ratio was inderdaad beter dan alleen Aβ 42 in het onderscheiden van AD en
dementie met Lewy lichaampjes (DLB) of vasculaire dementie (VaD), en even goed in
het differentiëren tussen AD en frontotemporale dementie (FTD). Echter, de Aβ 42/Aβ 40
ratio voegde niet veel toe aan de veelvuldig gebruikte combinatie van Aβ42,
gefosforyleerd tau (p-tau) en totaal tau (t-tau): deze laatste combinatie maakte in een
aantal gevallen zelfs beter onderscheid tussen AD en andere vormen van dementie dan
de Aβ 42/Aβ 40 ratio.
Alfa-synucleine is het eiwit dat intraneuronaal neerslaat bij DLB. In Hoofdstuk Drie
hebben we onderzocht of dit resulteert in een verlaagde CSF α-synucleine concentratie
bij patiënten met DLB, vergeleken met patiënten met andere vormen van dementie.
CSF α-synucleine zou in dat geval als biomarker voor DLB kunnen dienen. Helaas
vonden we geen verschillen in CSF α-synucleine concentraties tussen patiënten met
DLB, AD, FTD en VaD. Ook vergeleken met controle personen was de concentratie
ongewijzigd. CSF α-synucleine kan dus niet worden gebruikt als biomarker voor DLB.
Deze bevinding werd zowel bevestigd als tegengesproken in ander onderzoek. Om
deze wisselende bevindingen beter te kunnen begrijpen, hebben we CSF α-synucleine
nader onder de loep genomen in Hoofdstuk Vier. Er zijn in vitro interacties beschreven
187
Chapter twelve
tussen Aβ en α-synucleine, evenals sinusoïde schommelingen in CSF Aβ concentraties.
Door een interactie tussen Aβ en α-synucleine zouden dergelijke schommelingen ook
uitgelokt kunnen worden in α-synucleine concentraties. Om dit te onderzoeken hebben
we bij zes AD patiënten en zes gezonde vrijwilligers herhaaldelijk CSF afgenomen, met
tussenpozen van één uur. Per persoon waren ongeveer 33 samples beschikbaar voor
analyse. In deze samples hebben we echter geen lineaire trends of sinusoïde
schommelingen in CSF α-synucleine concentraties geobserveerd. Ook de interactie
tussen Aβ en α-synucleine zoals beschreven in vitro werd niet teruggevonden: we
vonden geen correlaties tussen CSF Aβ en α-synucleine in deze samples.
Om na te gaan hoe vaak CSF analyse wordt toegepast in de klinische praktijk,
hebben we clinici op geheugenpoliklinieken in Nederland en in Europa ondervraagd
door middel van een vragenlijst. De resultaten zijn beschreven in Hoofdstuk Vijf: CSF
analyse werd in bijna 60% van de geheugenpoliklinieken toegepast, in 5% (mediaan)
van de patiënten. Dit waren vooral patiënten die verdacht werden van een beginnende
dementie. CSF analyse leidde er soms toe dat de werkdiagnose werd veranderd. Een
reden om CSF analyse niet toe te passen was onder andere het ontbreken van duidelijke
richtlijnen hiervoor.
De nationale en internationale richtlijnen over diagnostiek van dementie zijn ook
onduidelijk over de interpretatie van de CSF biomarkers. Stel dat de concentratie Aβ 42
verlaagd is, maar de concentraties p-tau en t-tau zijn normaal, is er dan sprake van AD?
Om de interpretatie van de CSF biomarkers te vergemakkelijken hebben we een
predictiemodel ontwikkeld dat de waarschijnlijkheid berekent dat een geheugenpolipatiënt die verdacht wordt van dementie, AD heeft. Dit model wordt in Hoofdstuk Zes
beschreven. Eerder gepubliceerde predictiemodellen maakten geen goed onderscheid
tussen AD en andere vormen van dementie, of vermeldden geen formule die door
anderen gebruikt kon worden. Ons model werd specifiek ontwikkeld voor geheugenpolipatiënten die verdacht worden van dementie, omdat differentiëren tussen
verschillende vormen van dementie vaak een uitdaging is in de klinische praktijk.
Binnen deze populatie kon met behulp van geslacht, CSF Aβ42 en p-tau de waarschijnlijkheid berekend worden dat iemand AD heeft. Ons model maakte goed onderscheid
tussen AD en non-AD en de voorspelling kwam goed overeen met de werkelijkheid.
Een groot voordeel van ons model is dat het ook door andere geheugenpoliklinieken
gebruikt kan worden na aanpassing aan de eigen populatie.
In het tweede deel van dit proefschrift hebben we onderzocht of verschillende
hypotheses over de pathofysiologie van AD ondersteund worden door de resultaten
188
Nederlandse samenvatting
van CSF biomarker analyses. Een vaak aangehaalde hypothese stelt dat Aβ biomarkers
al maximaal afwijkend zijn voordat er klinisch symptomen optreden, terwijl tau
biomarkers toenemend afwijkend worden tijdens de beginstadia van de ziekte. In
Hoofdstuk Zeven beschrijven we een patiënt die zich presenteerde met een milde
geheugenstoornis, zonder dat er sprake was van dementie. Zowel de concentraties van
CSF Aβ 42 als van p-tau en t-tau waren op dat moment duidelijk afwijkend. In de jaren
hierna ging de patiënt achteruit, tot hij uiteindelijk een vergevorderde dementie had
ontwikkeld. Negen jaar na de eerste LP werd een tweede verricht. De concentraties
Aβ 42, p-tau en t-tau waren vrijwel ongewijzigd. Deze casus suggereert – in tegenstelling
tot de hypothese – dat er geen duidelijk tijdsverschil zit tussen het afwijkend worden
van Aβ en tau; of dat tau biomarkers al veel eerder afwijkend worden, vóórdat klinisch
symptomen optreden.
Om dit verder te onderzoeken hebben we in Hoofdstuk Acht de correlatie tussen
CSF biomarkers en cognitie onderzocht. Volgens de hypothese zou er geen correlatie te
verwachten zijn tussen CSF Aβ 42 en cognitie omdat Aβ biomarkers al maximaal
afwijkend zijn voordat cognitieve stoornissen optreden. Tau biomarkers echter zouden
toenemend afwijkend worden tijdens het ziekteproces, en dus zou men kunnen
verwachten dat een hogere concentratie CSF p-tau en t-tau correleren met een sterker
gestoorde cognitie. In een groep van 65 geheugenpolipatiënten (AD en non-AD)
vonden we echter geen correlaties tussen CSF biomarkers en cognitie. Een afwijkende
concentratie van CSF biomarkers lijkt te signaleren dat het AD ziekteproces gaande is,
maar geeft niet de ernst ervan weer.
Een andere hypothese over de pathofysiologie van AD is de vasculaire hypothese.
Deze hypothese kent twee richtingen: 1) cerebrovasculaire schade leidt via ischemie of
microbloedingen tot Aβ depositie en 2) Aβ depositie leidt via effecten op de vaatwand
tot cerebrovasculaire schade. Ongeacht de richting die men aanhangt, zou deze
samenhang ertoe moeten leiden dat AD en cerebrovasculaire schade hand in hand
gaan. In Hoofdstuk Negen hebben we 81 geheugenpolipatiënten onderzocht om
deze hypothese te testen. We vonden dat het hebben van AD niet geassocieerd was
met tekenen van cerebrovasculaire schade op MRI. Ook de concentratie CSF Aβ 42 en
Aβ 40 correleerde niet met de mate van cerebrovasculaire schade. Daarnaast had het
hebben van AD geen invloed op de bekende relatie tussen vasculaire risicofactoren en
cerebrovasculaire schade. Onze resultaten ondersteunen de vasculaire hypothese dus
niet.
In het laatste hoofdstuk van dit proefschrift, Hoofdstuk Tien, hebben we CSF Aβ 42
nader onder de loep genomen. De CSF Aβ 42 concentratie wordt steeds vaker gebruikt
189
Chapter twelve
in therapeutische trials om disease modification – een beïnvloeding van het patho­
fysiologische proces van AD – aan te tonen. De oorzaak van de verlaagde CSF Aβ42
concentratie bij AD is echter niet volledig opgehelderd. Uit een review van de
beschikbare literatuur bleek dat aggregatie van Aβ 42 de meest waarschijnlijke oorzaak
was. Als gevolg van deze aggregatie kan Aβ 42 niet meer vervoerd worden naar de CSF,
resulterend in een verlaagde CSF Aβ42 concentratie. Ook bleek uit deze review dat er
nog veel onduidelijkheid bestond over het transport van Aβ 42 vanuit de interstitiële
vloeistof in het brein naar de CSF, naar lymfe en naar het bloed. Het is niet duidelijk
hoeveel Aβ 42 via welke route vervoerd wordt, en of Aβ 42 de verschillende barrières
tussen deze compartimenten kan passeren. CSF Aβ42 kan niet worden ingezet als
biomarker van disease modification bij AD totdat we hier meer van begrijpen.
190
Nederlandse samenvatting
191
Dankwoord | Acknowledgements
Curriculum vitae
List of publications
Donders Series
Dankwoord | Acknowledgements
Dankwoord | Acknowledgements
De wens om een origineel dankwoord te schrijven is net zomin origineel als het hebben
van een onconventioneel promotietraject ongewoon is. Vanaf de eerste dag van mijn
(uiteraard onconventionele) promotietraject heb ik nagedacht over mijn dankwoord,
met als enige resultaat dat het waarschijnlijk verdacht veel lijkt op de dankwoorden van
degenen die mij voor gingen. Het maakt ook eigenlijk niet uit. Het is mijn dankwoord en
dat maakt het voor mij sowieso speciaal. Toch zou ik mezelf niet zijn als ik het er helemaal
bij liet zitten, en daarom begin ik niet zoals gebruikelijk met het bedanken van het
promotieteam, maar met de mensen zonder wie ik misschien wel helemaal niet aan dit
traject begonnen was: Judith Wilmer, Carolien van der Linden en Geert van der Aa.
Beste Judith, Carolien en Geert, bedankt voor de allerleukste eerste baan ooit, en
bedankt dat jullie me hebben doen inzien dat mijn hart in de geriatrie ligt! Op momenten
dat het even tegenzit, geven de herinneringen aan 4Oost me altijd weer nieuwe
motivatie.
En dan mijn promotieteam. Beste Marcel OR (‘Dat zijn de ballen die op tafel liggen’), je wist
me altijd weer te verrassen met andere invalshoeken. Dat maakte het wel eens lastig om
op één lijn te komen, maar het gaf tegelijkertijd ook aanleiding tot nieuwe
gezichtspunten. Bedankt dat je me de mogelijkheid gaf om zelf invulling te geven aan
mijn promotietraject: je was er als dat nodig was, en je was er niet als dat niet nodig was.
Ik heb veel geleerd, niet alleen van onderzoek doen, maar ook van de wereld van het
onderzoek doen.
Beste Marcel V (‘We worden zeldzaam aan het lijntje gehouden’), een groot deel van onze
communicatie verliep per e-mail, gewoon omdat dat zo lekker makkelijk was. Je
reageerde snel en draaide er niet omheen: ‘Volgens mij wordt er in dit hele gesticht
langs elkaar heen gepraat!!!!!!’. Je soms schaarse lettergreepgebruik werd gelukkig altijd
gecompenseerd door een overmaat aan leestekens. En als ik het even niet begreep was
je altijd bereid tot uitleg: ‘Nu gaat de wondere wereld van de biochemie open voor de
clinicus! Ik bel je wel even.’ Bedankt voor die directheid; het maakt het leven een stuk
makkelijker. Bedankt ook dat je me de gelegenheid gaf om een labstage te doen, zodat
ik wat zou meekrijgen van de CSF analyses zelf, en dat je me, hoewel met een hoop
gezucht en geprotesteer, toch altijd dat blad met die net iets hogere impactfactor nog
liet proberen.
195
Dankwoord | Acknowledgements
Beste Jurgen (‘Kippen hebben ook Alzheimer. Maar eieren niet.’), je raakte pas later betrokken
bij mijn promotie en zorgde meteen voor structuur in de organisatie ervan – misschien
wel het laatste wat je van jezelf verwacht had. Duidelijke afspraken over laatste auteurschap
voorkwamen dat elk artikel en elke revisie eindeloos gepingpongd moesten worden, en
dat voorkwam weer veel frustratie. Bedankt! Je deed me daarnaast inzien dat ik op
sommige zaken en meningen geen enkele invloed had en dat ik m’n energie dus beter
voor andere dingen kon gebruiken. Bedankt dat ik in jouw plaats het geheugenpoli-MDO
mocht voorzitten en dat je ook voor die tijd al toeliet dat ik de muis afpakte.
Mijn manuscriptcommisie, prof. dr. Pieter Wesseling, prof. dr. Gerard Martens en prof.
dr. Patrick Bossuyt wil ik hartelijk danken voor hun bereidheid om mijn manuscript te
beoordelen.
Natuurlijk zijn er nog veel meer mensen betrokken geweest bij mijn promotietraject en
de totstandkoming van mijn manuscript. Berry Kremer (‘Ik haal zelf wel koffie, ik ben een
grote jongen’), bedankt voor de waardevolle bijdragen en het klinische perspectief in
het begin van het traject. Magnus Sjögren (‘Boys need their tools and girls need tools as
well’), thanks for lively discussions (and great aftershave!). Roy Kessels (‘Hij is een soort
Willy Vandersteen’), ik doe mijn best om de term kortetermijngeheugen te vermijden;
dankzij jou is cognitie een woord met betekenis geworden. De discussie over het nut
van ‘prikken’ en ‘vragenlijstjes’ blijft leuk!
Wat is een promotietraject zonder collega-promovendi? Wat was het bijzonder om te
werken op een afdeling waar begrippen als nieuwjaarsconceptie, fantoomorgasme en
de ontploffende chihuahua drugshond tot gedeelde kennis behoorden (Appendix A).
Allereerst dank aan mijn kamergenoten op de 4 e: Sarah (‘Ik ben eigenlijk niet geschikt om
samen te werken maar het moet nou eenmaal’), Janneke (‘Het was er ook uit, voordat ik het
zei’) en Cynthia (‘Je bent zo geel als je wilt zijn’). Dankzij jullie kreeg ik geen enkele kans om
mijn koffiemisbruik te minderen! Concentratiedinsdag, vergaderdonderdag en vooral
Eftelingvrijdag werden trouw in ere gehouden, en de kamerpauzes maakten de incidentele
dramawoensdagen weer draaglijk.
Maar natuurlijk ook dank aan the old gang, Mirrie (‘Pas op anders gooi ik je biertje over me
heen!’), Arrie (‘Ik was de een-na eerste die wegging’), Els (‘Liever een badkamer dan een
breezer’), Diane (‘Het is vermakelijk in zijn sneuheid’), Marieke (‘Brownie? Is dat een lekkere
196
Dankwoord | Acknowledgements
neger?’), Olga (‘Ik zou geen even korte term kunnen bedenken die net zo lang is’) en Jaap
(‘Mag ik binnenkomen in je nederige schulp?’): dank voor geweldige congressen, voor
La Chouffe, de borrels, BBQs, en Martini die speciaal voor mij werd ingekocht – Bryan
Adams’ summer of ’69 stelde weinig voor vergeleken met onze summer of ‘09!
Daarnaast ook alle andere koffiepauzebezoekers bedankt, waaronder Mirjam (‘Wat je
ook doet, doe het niet’), Franka (‘Mark Ruffalo, die komt echt m’n bed niet in’), Teun (‘Valt een
boom niet als niemand het hoort’), Aisha (‘Voor een tikkie op de kop geprikt’), Anouk (‘De
kerkdienst was wel veel met God’), Kim (‘Ik heb een heel delicaat kopje’), Lia (‘Voor veel doe ik
alles’), René (‘Ik ben er ongetwijfeld van overtuigd’), Joep L. (‘Hoe oud ben je eigenlijk? Je
hoort wat... En dan denk je ze zal wel jonger zijn’), Hans, Geke en Joep S.
Mijn collega’s van de wilde mok, Saskia (‘Wat je niet ziet bestaat niet’), Sondra, Margery en
Bianca: we hebben het maar mooi gepresteerd om vier studies tegelijk draaiende te
houden! Onze schema’s en overzichten stonden in schril contrast tot onze chaotische
overlegmomenten (en etentjes!). Marianne en Jeanette, bedankt voor het meedenken
over de ingewikkelde planning: gaat het om een NPO-K, een NPO-L of een NPO-R?
Wordt het een 3T MRI, een ‘gewone’ of een Parelsnoer-MRI? Kunnen we de MRI al
plannen vóór het NPO? Hoe ingewikkeld het ook was, hoeveel mailtjes er ook over heen
gingen, ik voelde me altijd welkom op de poli. William (‘Ik zag een dame met een vriendelijk
gebruinde huid’) en Anja, geweldig hoe jullie patiënten wisten te motiveren, maar ook
altijd de tijd namen om even te informeren hoe het met mij ging.
Een afdeling functioneert natuurlijk niet zonder secretariële ondersteuning. Gemma
(‘Alle bouwvakkers zijn hetzelfde!’), Hanna en Nora, bedankt voor het geduld waarmee
jullie me steeds opnieuw uitlegden hoe de fax werkte, waar ik A5-enveloppen met
venster kon vinden en hoe ik declaratieformulieren moest invullen, en voor jullie hartelijke
begroeting als ik weer even aan kom waaien!
Marijke Beenes, Inge van der Stelt, Nienke Timmers en Henk van het CDL, jullie hebben
mij laten kennismaken met de niet-klinische kant van het onderzoek: de CSF analyse
zelf, de transgene muizen en hun hersenen. Bedankt!
Ook mijn vrienden en familie wil ik bedanken: met jullie kon ik niet alleen mijn enthousiasme
delen, maar ook de frustraties. Daarnaast boden jullie het onmisbare ingrediënt van elk
promotietraject: afleiding.
197
Dankwoord | Acknowledgements
Lieve Paula, dat jij mijn paranimf zou zijn stond eigenlijk in groep 8 al vast, lang voordat
we wisten wat promoveren überhaupt betekende. Je kent me door en door en bij jou
kan ik alles kwijt, inclusief de rare details! Bedankt dat je me altijd steunt, zo nodig met
irrationele argumenten als de rationele op zijn (‘Wat heeft ze eigenlijk een gek hoofd’).
Lieve Emy, we zien elkaar chronisch te weinig, maar als het dan lukt, voelt het altijd net
zo vertrouwd als ‘vroeger’, toen we nog vier minuten in plaats van 13 uur wandelen uit
elkaar woonden. Jij was de enige die het waagde om keihard te lachen toen ik als eerste
met kinderen te maken kreeg. Geweldig! Fijn dat je mijn paranimf wilt zijn. Lieve Naomi
en Debbie, voor jullie geldt hetzelfde: afspreken is lastig, maar onze vriendschap is het
zeker waard om sneeuwstormen, sein- en wisselstoringen, files en zelfgeïnduceerde
opsluitingen te trotseren!
Martine, jij introduceerde de afgedwongen ontspanning: onderzoek lijkt opeens heel
ver weg als je in een bubbelbad van 36°C ligt! Hanneke, jouw enthousiasme voor koken
zal ik wel nooit begrijpen, maar ik geniet van de maaltijden die je me voorzet! De
meiden van het KWC, Simone, Sabine, Marieke, Marie-José, Cyrina, Cora, Arnélie, Anja:
wat bijzonder dat we na al die jaren nog steeds zulke gezellige avonden hebben!
Dr. Toonen, amice! en mede-grondlegger van de quoteslijst, waar zouden we zijn zonder
Outlook? Bedankt voor de relativerende egotrips en regelmatige frietjes super!
My online friends, in anti-alphabetical order: Simon, Shreyas, Renato, Peter, Karin, Chris,
Aurélie, thanks for the healthy injections of sarcasm! Peter, thanks for the many instant
translations and corrections of contorted English sentences. It’s great to have online support
when you spend a lot of time behind the computer in an effort to finish your thesis.
Mijn familie, ooms en tantes, nichtjes en neven, ook al zien we elkaar soms tijden niet, er
is altijd die band. Bedankt voor jullie warme belangstelling. Dat geldt natuurlijk ook voor
mijn stoere oma (‘Zullen we gaan eten, dan hebben we dat maar weer gehad’). Ik zie heel
wat overeenkomsten tussen ons! Lieve Frans en Adie, ik voel me altijd welkom bij jullie,
bedankt! Sander, zonder jou zou ik nog nooit van Plex of Dylan Moran gehoord hebben,
Inception en The Big Lebowski nog nooit gezien hebben... jij laat me kennis maken met
dingen buiten mijn eigen wereldje. Bedankt!
Lieve mam, jij steunt me altijd, wat er ook gebeurt, of we het nou eens zijn of niet.
Ik weet dat je trots op me bent; ik ben ook trots op jou!
198
Dankwoord | Acknowledgements
Tot slot, het meest onverwachte resultaat van mijn promotie: lieve Jurgen, my place in
the sun, het was (en is) soms gecompliceerd, maar dat is het zo waard! Bedankt dat je me
er steeds weer aan herinnert dat er meer in het leven is dan werk. Je relativeringsvermogen – ‘anders ga je toch worsten verkopen bij de HEMA’ – heeft heel wat ingebeelde
drempels verlaagd. Wat een heerlijk vooruitzicht om de rest van ons leven samen door
te brengen!
199
Dankwoord | Acknowledgements
Appendix A | The quote 500
In de tekenfilm van mijn bestaan... Een impressie van het promotietraject
Het promotieteam
We nemen de knoop mee om door te hakken (MOR)
Hoe meer mensen je ontmoet, hoe meer contact
je maakt (JC)
Dan moeten we nadrukkelijk om de tafel gaan
zitten (MOR)
Alles is krom in het leven (JC)
Het zal vruchten opleveren (MOR)
We worden zeldzaam aan het lijntje gehouden (MV)
Het is een precaire dag aan het worden (MOR)
Als je het uitstelt, duurt het langer (JC)
Ze heeft wel wat zelfbesef (JC)
Dat zijn de ballen die op tafel liggen (MOR)
Nauseating behavior is the road to success (JC)
De promovenda
Iedere berg een Heidi
Je bent uniek in je saaiheid
Ik wil naar buiten, maar buiten wil mij niet
Hij zegt weinig dat onthouden moet worden
Je gaat helemaal op in de achtergrondmuziek
Daar word ik nog steeds niet juichend van wakker
We zien wel waar het ei landt
Je waardeert hem als zijnde iemand
Als jij piept dan trek ik
Gaat het zo mee of wil je er een kusje op?
Teleurstellend als ik ben, heb ik je gemaild
Het houdt niet over aan zeeën van ruimte
Niet alles wat lelijk is wordt een zwaan
De kamergenoten
Er was minimale genegenheid (Sarah)
Ik krijg galbulten van je (Janneke)
Het is dat ik geen hersens heb om mee te
gooien (Sarah)
M’n lach zit los (Marieke)
De frailty ligt hier voor het oprapen (Marieke)
En dan gaat-ie zitten denken in zijn eigen
hoofd (Janneke)
Hij heeft echt een plaat voor z’n bord (Sarah)
Het was er ook uit, voordat ik het zei (Janneke)
Ik geloof in alles dat eetbaar is (Janneke)
Alle bij kwetsbare hulpverleners betrokken
ouderen (Sarah)
Ik ben eigenlijk niet geschikt om samen te werken
maar het moet nou eenmaal (Sarah)
Ik ben niet zo helder vandaag. Ik had gisteren m’n
heldere dag (Janneke)
200
Dan zijn de apen echt gaar (Janneke)
Je kunt zeggen dat mensen gematigd neutraal
reageerden. (Sarah) Dat is zeker CDA-taal (Janneke)
Ik was helemaal geïntegreerd in de
medemens (Janneke)
Ze schrijft best goed Engels voor een gewoon
iemand (Janneke)
Waarom luisteren als je ook zelf aan het woord
kunt zijn? (Sarah)
Brownie? Is dat een lekkere neger? (Marieke)
Ik was tot tranen toe geamuseerd (Marieke)
Je bent zo geel als je wilt zijn (Cynthia)
Ik denk dat zij dat wel zo denkt (Cynthia)
Ik vind Hans wel een naam voor een schildpad.
Ja, maar niet andersom (Sarah)
Waarom pas je het aan? Omdat je data het
willen (Sarah)
Edwin is zo’n man die je niet moet laten
bevallen (Sarah)
Omdat ik denk dat ik minder adequaat ben dan
ik soms denk (Sarah)
De oud-kamergenoten
Ken ik u ergens van? Ja ik ben familie van
iemand (Gerstella)
Ik word er gruwelijk simpel van (Gerstella)
Groen en geel is blauw (Els)
Ik jat liever dan dat ik brand sticht (Diane)
Ik was de een-na eerste die wegging (Arrie)
Het is vermakelijk in zijn sneuheid (Diane)
We gaan gewoon elkaars eigen gang (Arrie)
Het leven is één grote huwelijksreis (Bonnie)
Dat vind ik een gezamenlijke verzinning (Arrie)
Niet iedereen met een kind is lesbisch (Diane)
Niets bereiken is ook een eindpunt (Diane)
Wie twijfelt wordt overgeslagen (Mirrie)
Pas op anders gooi ik je biertje over me heen! (Mirrie)
Jij hoeft alleen maar de trap op te rollen (Mirrie)
Het gras is groener bij de buren. Nou dan jatten
we het toch? (Diane)
Er ligt een watermeloen en brood maar geen
eten (Diane)
Schatje, je moet de beer niet verkopen... (Mirrie)
Ik heb zoveel Y daar kun je een taart van
bakken (Diane)
Ik ben als een brok voor je gevallen
Ik snap nooit dat mensen mij bellen en dan nemen
ze niet op (Arrie)
Dankwoord | Acknowledgements
Geen liefde vóór het huwelijk (Diane)
Een dag niet gezeken is een urineretentie
Liever een badkamer dan een breezer (Els)
Jij zit onder jezelf (Els)
De collega’s van de 4e verdieping
Koppel me back baby! (Jaap)
Mag ik binnenkomen in je nederige schulp? (Jaap)
Van een koude curry thuiskomen (Jaap)
Hij heeft iets vaag alternatiefs (Olga)
Ik heb het warm aan m’n eigen achterkant (Olga)
Het is leuk dat-ie zo lang is. Als jongen (Olga)
Ik zou geen even korte term kunnen bedenken die
net zo lang is (Olga)
Je ziet er goed uit, voor je eetgedrag (Olga)
Wat je niet ziet bestaat niet (Saskia)
Oh dus je hebt morgen last van gister (Teun)
Chronologisch ben ik ouder (Teun)
Nouhouhou, ik wil ook een standaardafwijking (Saskia)
Wat staan jullie spannend in de gang (Olga)
Valt een boom niet als niemand het hoort (Teun)
Was het maar week 8, dan was ik er volgende
week niet meer (Saskia)
Die man is natuurlijk zwaar gepensioneerd (Olga)
De overige collega’s
Boys need their tools and girls need tools
as well (Magnus)
Mensen die zonder autopsie overlijden (Drieske)
Humans or patients (Magnus)
Ik ben er ongetwijfeld van overtuigd (René)
Hoe oud ben je eigenlijk? Je hoort wat... En dan
denk je ze zal wel jonger zijn (Joep L.)
Ik heb weinig uitspraken, ik drink alleen (Mike)
Van die halve semi-getrouwde vrouwen (Masha)
Mijn zus is soort van familie (Joep L.)
Elke wijn is mysterieus (Mirjam)
Ik zag een dame met een vriendelijk gebruinde
huid (William)
Hij denkt dat-ie denkt zoals hulpverleners
denken (Mirjam)
Voor een tikkie op de kop geprikt (Aisha)
Mark Ruffalo, die komt echt m’n bed niet in (Franka)
Wat je ook doet, doe het niet (Mirjam)
Achteraf was het helemaal niet zo erg maar ik dacht
wel ik hoef hier geen kerel (Franka)
Voor veel doe ik alles (Lia)
Je moet er niet voor de diepgang heengaan
(Anne over Afrika)
Ik kan alles horizontaal (Lia)
Hij is een soort Willy Vandersteen (Roy)
De kerkdienst was wel veel met God (Anouk)
Ik heb een heel delicaat kopje (Kim)
Het dossieronderzoek
Patiënte is vier maanden geleden twee keer op één
volgende na elkaar gevallen. Op dit gebied is zij
nu ruim vier maanden klachtenvrij.
Ik adviseerde patiënt in het bijzijn van zijn vrouw
niet auto te rijden
Behoudens bradifrenie geen vocale afwijkingen
Na expiratie van het insulinoom
Patiënte is goed afleidbaar
Morbus Pieck met dementerend beeld
De opbrengst
Neusinsufficiëntie
Hypoanemie
De geriateur
Pradie cardie
Colonfractuur
Harde geriatrie
Patiëntgerelateerde patiëntenzorg
Dubbelzinnig printen
Zinloos geschoven
Pub mét (en Pub zonder)
Fantoomorgasme
Placebothermostaat
Nieuwjaarsconceptie
Sportief grijs haar
Schurftkalender
Bad hair life
Extra-terosteriaal
Aangeboren giraffe
Stickhorse
Het gebleekte significantiesterretje
Teratoom met epilepsie
Breedsprakig looppatroon
AIO: Ariër in opleiding
Het ei opnieuw uitvinden
Polyfiel en oligofoob
Voldoende positieve gedagzeggingen
Linksharig tuig
201
Curriculum vitae
Curriculum vitae
Petra Spies was born on January 14, 1983 in Tiel, the Netherlands, and grew up in
Geldermalsen. She entered Medical School at Maastricht University in 2000 and received
her doctoral degree cum laude in 2004. Her interest in research was awakened during
her research internship at the Department of General Practice of Maastricht University.
After graduating in 2006, she worked at the Department of Geriatric Medicine at the
Catharina Hospital Eindhoven before starting her training in geriatrics in 2007. This
residency was combined with three years of scientific research, the results of which are
described in this thesis. During this time she was awarded the Frye stipendium, a grant
for promising female PhD-students, and an award for the best scientific oral presentation
at the Dutch Geriatric and Gerontological Society Annual conference. Petra continued
her geriatrics training in June 2011 with residency in clinical pharmacology, and is
currently doing a two-year internal medicine rotation at the Canisius Wilhelmina
Hospital in Nijmegen.
Petra Spies werd op 14 januari 1983 geboren in Tiel en groeide op in Geldermalsen. In
2000 begon zij haar studie geneeskunde aan de Universiteit Maastricht, waar zij in 2004
cum laude haar doctoraal haalde. Haar interesse in wetenschappelijk onderzoek werd
gewekt tijdens haar wetenschapsstage bij de afdeling Huisartsgeneeskunde van de
Universiteit Maastricht. Na haar artsexamen in 2006 werkte zij op de afdeling geriatrie
van het Catharina Ziekenhuis Eindhoven. In 2007 begon zij haar opleiding tot klinisch
geriater in het UMC St Radboud in Nijmegen, gecombineerd met promotieonderzoek.
De resultaten van dit onderzoek zijn in dit proefschrift beschreven. Tijdens dit promotie­
traject ontving zij het Frye stipendium, een beurs voor veelbelovende vrouwelijke
promovendi, en de prijs voor de beste wetenschappelijke presentatie op de
Geriatriedagen. In juni 2011 hervatte Petra haar opleiding tot klinisch geriater en startte
zij tevens de opleiding tot klinisch farmacoloog. Momenteel is zij werkzaam op de
afdeling interne geneeskunde van het Canisius Wilhelmina Ziekenhuis te Nijmegen.
202
List of publications
List of publications
A prediction model to calculate probability of Alzheimer’s dementia using cerebrospinal fluid biomarkers
Spies PE, Claassen JA, Peer PG, Blankenstein MA, Teunissen CE, Scheltens Ph, van der Flier WM, Olde Rikkert MG,
Verbeek MM
Alzheimers Dement 2012 doi:10.1016/j.jalz.2012.01.010
Reviewing reasons for the decreased CSF Aβ42 concentration in Alzheimer disease
Spies PE, Verbeek MM, van Groen T, Claassen JA
Front Biosc 2012 (in press)
CSF α-synuclein concentrations do not fluctuate over hours and are not correlated to amyloid β in humans
Spies PE, Slats D, Olde Rikkert MG, Tseng J, Claassen JA, Verbeek MM
Neurosci Lett 2011;504:336-8
Elderly: a term to avoid or to embrace?
Spies PE, Claassen JA.
JAGS 2011;53:1986-7
Hourly variability of cerebrospinal fluid biomarkers in Alzheimer Disease and healthy older volunteers
Slats D, Claassen JA, Spies PE, Borm G, Besse KT, van Aalst W, Tseng J, Sjögren JM, Olde Rikkert MG, Verbeek MM
Neurobiol Aging 2011 doi:10.1016/j.neurobiolaging.2011.07.008
Experiences with cerebrospinal fluid analysis in Dutch memory clinics
Spies PE, Slats D, Ramakers I, Verhey FR, Olde Rikkert MG.
Eur J Neurol 2011l;18:1014-6
Cerebrospinal fluid tau and amyloid β proteins do not correlate with cognitive functioning in cognitively
impaired memory clinic patients
Spies PE, Claassen JA, Slats D, Olde Rikkert MG, Verbeek MM, Kessels RP
CNS Spectr 2010;15:588-93
CSF biomarkers in Alzheimer’s disease: are the hypotheses more dynamic than the biomarkers?
Spies PE, Verbeek MM, Olde Rikkert MG, Claassen JA
JAGS 2010;58:1619-1620
The Cerebrospinal Fluid Amyloid beta(42/40) Ratio in the Differentiation of Alzheimer’s Disease from
Non-Alzheimer’s Dementia
Spies PE, Slats D, Sjögren JM, Kremer BP, Verhey FR, Olde Rikkert MG, Verbeek MM
Curr Alzheimer Res 2010;7:470-6
Cerebrospinal Fluid Biomarkers in Diagnosing Alzheimer’s Disease in Clinical Practice: An Illustration
with 3 Case Reports
Slats D, Spies PE, Sjögren JM, Verhey FR, Verbeek MM, Olde Rikkert MG
Case Rep Neurol 2010;2:5-11
CSF biomarker utilisation and ethical considerations of biomarker assisted diagnosis and research in
dementia: perspectives from within the European Alzheimer’s Disease Consortium (EADC)
Slats D, Spies PE, Sjögren JM, Visser PJ, Verbeek MM, Olde Rikkert MG, Kehoe PG
J Neurol Neurosurg Psychiatry 2010;81:124-5
Cerebrospinal fluid alpha-synuclein does not discriminate between dementia disorders
Spies PE, Melis RJ, Sjögren JM, Olde Rikkert MG, Verbeek MM
J Alzheimers Dis 2009;16:363-9
Alzheimer disease is not associated with cerebrovascular disease
Spies PE, Verbeek MM, Sjögren JM, de Leeuw F-E, Claassen JA
Submitted
203
Donders Graduate School for Cognitive Neuroscience Series
Donders Graduate School for Cognitive Neuroscience Series
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
204
van Aalderen-Smeets, S.I. (2007). Neural dynamics of visual selection. Maastricht University, Maastricht, the
Netherlands.
Schoffelen, J.M. (2007). Neuronal communication through coherence in the human motor system. Radboud
University Nijmegen, Nijmegen, the Netherlands.
de Lange, F.P. (2008). Neural mechanisms of motor imagery. Radboud University Nijmegen, Nijmegen, the
Netherlands.
Grol, M.J. (2008). Parieto-frontal circuitry in visuomotor control. University Utrecht, Utrecht, the Netherlands.
Bauer, M. (2008). Functional roles of rhythmic neuronal activity in the human visual and somatosensory system.
Radboud University Nijmegen, Nijmegen, the Netherlands.
Mazaheri, A. (2008). The Influence of Ongoing Oscillatory Brain Activity on Evoked Responses and Behaviour.
Radboud University Nijmegen, Nijmegen, the Netherlands.
Hooijmans, C.R. (2008). Impact of nutritional lipids and vascular factors in Alzheimer’s Disease. Radboud
University Nijmegen, Nijmegen, the Netherlands.
Gaszner, B. (2008). Plastic responses to stress by the rodent urocortinergic Edinger-Westphal nucleus. Radboud
University Nijmegen, Nijmegen, the Netherlands.
Willems, R.M. (2009). Neural reflections of meaning in gesture, language and action. Radboud University
Nijmegen, Nijmegen, the Netherlands.
Van Pelt, S. (2009). Dynamic neural representations of human visuomotor space. Radboud University Nijmegen,
Nijmegen, the Netherlands.
Lommertzen, J. (2009). Visuomotor coupling at different levels of complexity. Radboud University Nijmegen,
Nijmegen, the Netherlands.
Poljac, E. (2009). Dynamics of cognitive control in task switching: Looking beyond the switch cost. Radboud
University Nijmegen, Nijmegen, the Netherlands.
Poser, B.A. (2009) Techniques for BOLD and blood volume weighted fMRI. Radboud University Nijmegen,
Nijmegen, the Netherlands.
Baggio, G. (2009). Semantics and the electrophysiology of meaning. Tense, aspect, event structure. Radboud
University Nijmegen, Nijmegen, the Netherlands.
van Wingen, G.A. (2009). Biological determinants of amygdala functioning. Radboud University Nijmegen
Medical Centre, Nijmegen, the Netherlands.
Bakker, M. (2009). Supraspinal control of walking: lessons from motor imagery. Radboud University Nijmegen
Medical Centre, Nijmegen, the Netherlands.
Aarts, E. (2009). Resisting temptation: the role of the anterior cingulate cortex in adjusting cognitive control.
Radboud University Nijmegen, Nijmegen, the Netherlands.
Prinz, S. (2009). Waterbath stunning of chickens – Effects of electrical parameters on the electroencephalogram
and physical reflexes of broilers. Radboud University Nijmegen, Nijmegen, the Netherlands.
Knippenberg, J.M.J. (2009). The N150 of the Auditory Evoked Potential from the rat amygdala: In search for its
functional significance. Radboud University Nijmegen, Nijmegen, the Netherlands.
Dumont, G.J.H. (2009). Cognitive and physiological effects of 3,4-methylenedioxymethamphetamine (MDMA or
’ecstasy’) in combination with alcohol or cannabis in humans Radboud University Nijmegen, Nijmegen, the
Netherlands.
Pijnacker, J. (2010). Defeasible inference in autism: a behavioral and electrophysiogical approach. Radboud
Universiteit Nijmegen, the Netherlands.
de Vrijer, M. (2010). Multisensory integration in spatial orientation. Radboud University Nijmegen, Nijmegen,
the Netherlands.
Vergeer, M. (2010). Perceptual visibility and appearance: Effects of color and form. Radboud University Nijmegen,
Nijmegen, the Netherlands.
Levy, J. (2010). In Cerebro Unveiling Unconscious Mechanisms during Reading. Radboud University Nijmegen,
Nijmegen, the Netherlands.
Treder, M. S. (2010). Symmetry in (inter)action. Radboud University Nijmegen, Nijmegen, the Netherlands.
Donders Graduate School for Cognitive Neuroscience Series
26. Horlings C.G.C. (2010). A Weak balance; balance and falls in patients with neuromuscular disorders. Radboud
University Nijmegen, Nijmegen, the Netherlands
27. Snaphaan, L.J.A.E. (2010). Epidemiology of post-stroke behavioural consequences. Radboud University
Nijmegen Medical Centre, Nijmegen, the Netherlands.
28. Dado – Van Beek, H.E.A. (2010). The regulation of cerebral perfusion in patients with Alzheimer’s disease.
Radboud University Nijmegen Medical Centre, Nijmegen, the Netherlands.
29. Derks, N.M. (2010). The role of the non-preganglionic Edinger-Westphal nucleus in sex-dependent stress
adaptation in rodents. Radboud University Nijmegen, Nijmegen, the Netherlands.
30. Wyczesany, M. (2010). Covariation of mood and brain activity. Integration of subjective self-report data with
quantitative EEG measures. Radboud University Nijmegen, Nijmegen, the Netherlands.
31. Beurze S.M. (2010). Cortical mechanisms for reach planning. Radboud University Nijmegen, Nijmegen, the
Netherlands.
32. van Dijk, J.P. (2010). On the Number of Motor Units. Radboud University Nijmegen, the Netherlands .
33. Lapatki, B.G. (2010). The Facial Musculature – Characterization at a Motor Unit Level. Radboud University
Nijmegen, Nijmegen, the Netherlands.
34. Kok, P. (2010). Word Order and Verb Inflection in Agrammatic Sentence Production. Radboud University
Nijmegen, Nijmegen, the Netherlands.
35. van Elk, M. (2010). Action semantics: Functional and neural dynamics. Radboud University Nijmegen,
Nijmegen, the Netherlands.
36. Majdandzic, J. (2010). Cerebral mechanisms of processing action goals in self and others. Radboud University
Nijmegen, Nijmegen, the Netherlands.
37. Snijders, T.M. (2010). More than words – neural and genetic dynamics of syntactic unification. Radboud
University Nijmegen, Nijmegen, the Netherlands.
38. Grootens, K.P. (2010). Cognitive dysfunction and effects of antipsychotics in schizophrenia and borderline
personality disorder. Radboud University Nijmegen Medical Centre, Nijmegen, the Netherlands.
39. Nieuwenhuis, I.L.C. (2010). Memory consolidation: A process of integration – Converging evidence from MEG,
fMRI and behavior. Radboud University Nijmegen Medical Centre, Nijmegen, the Netherlands.
40. Menenti, L.M.E. (2010). The right language: differential hemispheric contributions to language production and
comprehension in context. Radboud University Nijmegen, Nijmegen, the Netherlands.
41. van Dijk, H.P. (2010). The state of the brain, how alpha oscillations shape behaviour and event related responses.
Radboud University Nijmegen, Nijmegen, the Netherlands.
42. Meulenbroek, O.V. (2010). Neural correlates of episodic memory in healthy aging and Alzheimer’s disease.
Radboud University Nijmegen, Nijmegen, the Netherlands.
43. Oude Nijhuis, L.B. (2010). Modulation of human balance reactions. Radboud University Nijmegen, Nijmegen,
the Netherlands.
44. Qin, S. (2010) Adaptive memory: imaging medial temporal and prefrontal memory systems. Radboud University
Nijmegen, the Netherlands.
45. Timmer, N.M. (2011). The interaction of heparan sulfate proteoglycans with the amyloid b protein. Radboud
University Nijmegen, Nijmegen, the Netherlands.
46. Crajé, C. (2011). (A)typical motor planning and motor imagery. Radboud University Nijmegen, Nijmegen, the
Netherlands.
47. van Grootel, T.J. (2011). On the role of eye and head position in spatial localisation behaviour. Radboud University
Nijmegen, Nijmegen, the Netherlands.
48. Lamers, M.J.M. (2011). Levels of selective attention in action planning. Radboud University Nijmegen, Nijmegen,
the Netherlands.
49. Van der Werf, J. (2011). Cortical oscillatory activity in human visuomotor integration. Radboud University
Nijmegen, Nijmegen, the Netherlands.
50. Scheeringa, R. (2011). On the relation between oscillatory EEG activity and the BOLD signal. Radboud University
Nijmegen, Nijmegen, the Netherlands.
205
Donders Graduate School for Cognitive Neuroscience Series
51. Bögels, S. (2011). The role of prosody in language comprehension: when prosodic breaks and pitch accents come
into play. Radboud University Nijmegen, Nijmegen, the Netherlands.
52. Ossewaarde, L. (2011). The mood cycle: hormonal influences on the female brain. Radboud University Nijmegen,
Nijmegen, the Netherlands.
53. Kuribara, M. (2011). Environment-induced activation and growth of pituitary melanotrope cells of Xenopus laevis.
Radboud University Nijmegen, Nijmegen, the Netherlands.
54. Helmich, R.C.G. (2011). Cerebral reorganization in Parkinson’s disease. Radboud University Nijmegen, Nijmegen,
the Netherlands.
55. Boelen, D. (2011). Order out of chaos? Assessment and treatment of executive disorders in brain-injured patients.
Radboud University Nijmegen, Nijmegen, the Netherlands.
56. Koopmans, P.J. (2011). fMRI of cortical layers. Radboud University Nijmegen, Nijmegen, the Netherlands.
57. van der Linden, M.H. (2011). Experience-based cortical plasticity in object category representation. Radboud
University Nijmegen, Nijmegen, the Netherlands.
58. Kleine, B.U. (2010). Motor unit discharges - Physiological and diagnostic studies in ALS. Radboud University
Nijmegen Medical Centre, Nijmegen, the Netherlands.
59. Paulus, M. (2011). Development of action perception: Neurocognitive mechanisms underlying children’s
processing of others’ actions. Radboud University Nijmegen, Nijmegen, the Netherlands.
60. Tieleman, A.A. (2011). Myotonic dystrophy type 2. A newly diagnosed disease in the Netherlands. Radboud
University Nijmegen Medical Centre, Nijmegen, the Netherlands.
61. van Leeuwen, T.M. (2011). ‘How one can see what is not there’: Neural mechanisms of grapheme-colour
synaesthesia. Radboud University Nijmegen, Nijmegen, the Netherlands.
62. van Tilborg, I.A.D.A. (2011). Procedural learning in cognitively impaired patients and its application in clinical
practice. Radboud University Nijmegen, Nijmegen, the Netherlands.
63. Bruinsma, I.B. (2011). Amyloidogenic proteins in Alzheimer’s disease and Parkinson’s disease: interaction with
chaperones and inflammation. Radboud University Nijmegen, Nijmegen, the Netherlands.
64. Voermans, N. (2011) Neuromuscular features of Ehlers-Danlos syndrome and Marfan syndrome; expanding the
phenotype of inherited connective tissue disorders and investigating the role of the extracellular matrix in muscle.
Radboud University Nijmegen Medical Centre, Nijmegen, the Netherlands.
65. Reelick, M. (2011) One step at a time. Disentangling the complexity of preventing falls in frail older persons.
Radboud University Nijmegen Medical Centre, Nijmegen, the Netherlands.
66. Buur, P.F. (2011). Imaging in motion. Applications of multi-echo fMRI. Radboud University Nijmegen, Nijmegen,
the Netherlands.
67. Schaefer, R.S. (2011). Measuring the mind’s ear: EEG of music imagery. Radboud University Nijmegen, Nijmegen,
the Netherlands.
68. Xu, L. (2011). The non-preganglionic Edinger-Westphal nucleus: an integration center for energy balance and
stress adaptation. Radboud University Nijmegen, Nijmegen, the Netherlands.
69. Schellekens, A.F.A. (2011). Gene-environment interaction and intermediate phenotypes in alcohol dependence.
Radboud University Nijmegen, Nijmegen, The Netherlands.
70. van Marle, H.J.F. (2011). The amygdala on alert: A neuroimaging investigation into amygdala function during
acute stress and its aftermath. Radboud University Nijmegen, Nijmegen, the Netherlands.
71. De Laat, K.F. Motor performance in individuals with cerebral small vessel disease: an MRI study. Radboud
University Nijmegen Medical Centre, Nijmegen, the Netherlands.
72. Mädebach, A. (2011). Lexical access in speaking: Studies on lexical selection and cascading activation. Radboud
University Nijmegen, Nijmegen, the Netherlands.
73. Poelmans, G.J.V. (2011). Genes and protein networks for neurodevelopmental disorders. Radboud University
Nijmegen, Nijmegen, the Netherlands.
74. Van Norden, A.G.W. (2011). Cognitive function in elderly individuals with cerebral small vessel disease. An MRI
study. Radboud University Nijmegen Medical Centre, Nijmegen, the Netherlands
75. Jansen, E.J.R. (2011). New insights into V-ATPase functioning: the role of its accessory subunit Ac45 and a novel
brain-specific Ac45 paralog. Radboud University Nijmegen, Nijmegen, the Netherlands.
76. Haaxma, C.A. (2011). New perspectives on preclinical and early stage Parkinson’s disease. Radboud University
Nijmegen Medical Centre, Nijmegen, the Netherlands.
206
Donders Graduate School for Cognitive Neuroscience Series
77. Haegens, S. (2012). On the functional role of oscillatory neuronal activity in the somatosensory system. Radboud
University Nijmegen, Nijmegen, the Netherlands
78. Van Barneveld, D.C.P.B.M. (2012). Integration of exteroceptive and interoceptive cues in spatial localization.
Radboud University Nijmegen, Nijmegen, the Netherlands.
79. Spies, P.E. (2012). The reflection of Alzheimer disease in CSF. Radboud University Nijmegen Medical Centre,
Nijmegen, the Netherlands.
207