nature reviews | neurology
aDvanCe OnLine PuBLiCatiOn | 1
Laboratory, Institute of
of Psychiatry and
at the University of
Gothenburg, SE‑431 80
Discipline of Psychiatry,
School of Medicine,
Trinity College Dublin,
Dublin 24, Ireland
(H. Hampel). MRS Unit,
University of California,
San Francisco, VA
Medical Center, 4150
Clement Street, San
Francisco, CA 94121,
USA (M. Weiner).
Cerebrospinal fluid and plasma biomarkers
in Alzheimer disease
Kaj Blennow, Harald Hampel, Michael Weiner and Henrik Zetterberg
Abstract | Intense multidisciplinary research has provided detailed knowledge of the molecular pathogenesis
of Alzheimer disease (AD). This knowledge has been translated into new therapeutic strategies with putative
disease‑modifying effects. Several of the most promising approaches, such as amyloid‑β immunotherapy and
secretase inhibition, are now being tested in clinical trials. Disease‑modifying treatments might be at their
most effective when initiated very early in the course of AD, before amyloid plaques and neurodegeneration
become too widespread. Thus, biomarkers are needed that can detect AD in the predementia phase or,
ideally, in presymptomatic individuals. In this Review, we present the rationales behind and the diagnostic
performances of the core cerebrospinal fluid (CSF) biomarkers for AD, namely total tau, phosphorylated tau
and the 42 amino acid form of amyloid‑β. These biomarkers reflect AD pathology, and are candidate markers
for predicting future cognitive decline in healthy individuals and the progression to dementia in patients who
are cognitively impaired. We also discuss emerging plasma and CSF biomarkers, and explore new proteomics‑
based strategies for identifying additional CSF markers. Furthermore, we outline the roles of CSF biomarkers
in drug discovery and clinical trials, and provide perspectives on AD biomarker discovery and the validation of
such markers for use in the clinic.
Blennow, K. et al. Nat. Rev. Neurol. advance online publication 16 February 2010; doi:10.1038/nrneurol.2010.4
alois alzheimer presented the first case of the disease that
was to bear his name at a congress in tübingen, Germany,
in 1906.1 in this presentation, he described the “miliary
bodies” (amyloid plaques) and “dense bundles of fibrils”
(neurofibrillary tangles) that we now recognize as neuro
pathological hallmarks of alzheimer disease (aD). in
1985, researchers succeeded in purifying amyloid plaque
cores and, in so doing, identified the 4 kDa amyloidβ
(aβ) peptide as the main component of these extra cellular
deposits.2 this breakthrough led to the cloning of the gene
encoding the amyloid precursor protein (aPP),3 the mol
ecule from which aβ is derived. in 1986, neuro fibrillary
tangles were shown to be composed of abnormally hyper
phosphorylated forms of the protein tau.4 these important
achievements in the 1980s marked the start of modern aD
research, and have led to a detailed know ledge of aPP
metabolism and aβ generation (Figure 1), and of tau
homeostasis (Figure 2).
Mutations in APP or in one of the presenilin genes
(PSEN1 or PSEN2), which encode proteins involved in
aPP metabolism, have been found to cause rare famili al
forms of aD.5 Largely on the basis of these mutations,
aβ—in particular the 42 amino acid form of this peptide
(aβ1–42)—has been proposed as the driving force in the
disease process. indeed, the ‘amyloid cascade hypo thesis’6
posits that an imbalance between the production and clear
ance of aβ is the initiating event in aD, with the increase
in aβ load ultimately leading to tau pathology, neuronal
degeneration and dementia (Figure 3). Progress in aD
research has been translated into novel treatment strate
gies with diseasemodifying potential, and a large number
of candidate antiaβ drugs, such as aβ immunotherapies,
secretase inhibitors and aβ aggregation inhibitors, are in
various phases of clinical trials.5 Despite this progress, one
should note that the amyloid cascade hypothesis has not
been proven with certainty for lateonset aD.
Diseasemodifying drugs will probably be at their
most effective in patients in the earliest stages of aD,
before amyloid plaques and neurofibrillary tangles
become preva lent and neurodegeneration becomes too
severe.7–9 thus, patients will need to be identified in the
pre dementia stage (prodromal aD), or even the asympto
matic phase of the disease (preclinical aD). Prodromal
aD is defined as mild cognitive impairment (MCi) result
ing from underlying aD pathology, whereas preclinical
aD is characterized by progressive aD pathology in the
brain that is insufficiently severe to affect cognition. For
an aD drug to be labeled as diseasemodifying, evidence
must be available that the agent affects the central disease
processes and hallmark neuropathology, in addition to a
beneficial effect on cognition.10
the challenges of early diagnosis and identifica
tion of diseasemodifying drugs have created a need
for bi omarkers that reflect core elements of the disease
K. Blennow and M. Weiner declare an association with the
following company: Innogenetics. H. Hampel declares an
association with the following company: BRAHMS AG. See the
article online for full details of the relationships. H. Zetterberg
declares no competing interests.
© 20 Macmillan Publishers Limited. All rights reserved10
2 | aDvanCe OnLine PuBLiCatiOn
Current clinical diagnostic criteria for Alzheimer disease (AD) require a patient
to have dementia before a diagnosis can be made, and are largely based on the
exclusion of other disorders
Disease‑modifying drugs for AD, when they become available, will need to be
administered very early in the course of the disease, before neurodegeneration
is too severe and widespread
No clinical method is available for identifying prodromal AD in patients with mild
cognitive impairment (MCI), as such individuals have only mild disturbances in
The cerebrospinal fluid (CSF) biomarkers total tau, phosphorylated tau (p‑tau
and p‑tau231) and β‑amyloid1–42 have a high diagnostic accuracy for AD, and for
prodromal AD in patients with MCI
CSF biomarkers are increasingly being used in the clinic for diagnosing AD, and
will also be valuable in clinical trials, allowing enrichment of patient populations
with pure AD cases
Biomarker evidence that a candidate drug affects the central disease
processes in AD will, together with a beneficial effect on cognition, be essential
for labeling the drug as disease modifying
process. Here, we review the development of candidate
cerebrospinal fluid (CsF) and plasma biomarkers for aD.
we focus on established biomarkers (biomarkers evaluated
in several studies by various research groups), providing a
practical guide to their implementation in the clinic and
discussing their potential roles in clinical trials.
Cerebrospinal fluid biomarkers
a biomarker is an objective measure of a biological or
pathogenic process that can be used to evaluate disease
risk or prognosis, to guide clinical diagnosis or to monitor
therapeutic interventions. the CsF is in direct contact
with the extracellular space of the brain and can reflect
biochemical changes that occur in the latter. For these
reasons, the CsF is the optimal source of aD biomarkers.
CsF biomarkers for aD can be divided into basic and
core biomarkers (table 1). Basic biomarkers are used to
identify conditions that might mimic or coexist with aD,
while core biomarkers have been developed to identify
the central pathogenic processes in aD. the procedure
for obtaining CsF by lumbar puncture is outlined in
supplementary Figure 1 and supplementary table 1
online, with the latter also providing a standardized
proto col for CsF sample handling.
Basic biomarkers include assays for blood–brain barrier
(BBB) status and inflammatory processes in the brain
(supplementary Figure 2 online). the BBB is formed
from the restricted permeability of the capillaries in the
brain, and serves to maintain a controlled milieu for
neurons. the CsF:serum albumin ratio is the standard
biomarker for BBB function.11 an increase in this ratio
indicates BBB damage and is found in a variety of dis
orders, such as infections (for example, neuroborreliosis)
and inflammatory diseases (for example, Guillain–Barré
syndrome), brain tumors, and cerebrovascular disease,
including many cases of vascular dementia (table 1). the
CsF:serum albumin ratio is normal in patients with pure
aD, but often increases in cases of the disease that show
concomitant cerebrovascular pathology.12 thus, this ratio
might be of value in excluding various causes of brain
damage and for identifying patients with pure aD.
the immune system responds to chronic inflam matory
or infectious disorders in the Cns, such as multiple
sclero sis and neuroborreliosis, by producing antibodies—
a process called intrathecal immunoglobulin production.
this response can be measured either quantitatively, by
the igG and igM indices, or qualitatively, by identification
of oligoclonal bands in the CsF (supplementary Figure 3
online).13 the vast majority of patients with aD have no,
or only very mild signs of, intrathecal immunoglobulin
production (table 1), making measurement of this process
a valuable tool for excluding chronic inflammatory and
infectious disorders in the clinical workup of aD.
ideally, a core biomarker should be coupled to the under
lying molecular pathology of a disease.14 in aD, the core
biomarkers that have been developed reflect amyloid and
Figure 1 | Metabolic pathways for the generation of APP fragments detected in the
CSF. a | In the amyloidogenic pathway, APP is cleaved by β‑secretase, releasing
sAPPβ into the extracellular fluid and CSF. The remaining fragment in the plasma
membrane (CTFβ) is cleaved by γ‑secretase, generating Aβ1–42 and several carboxy‑
terminal truncated Aβ isoforms (Aβ1–40 down to Aβ1–17).81,145 β‑Secretase has been
identified as β‑site APP‑cleaving enzyme 1, also known as BACE1,146 whereas
γ‑secretase is an enzyme complex consisting of four components: presenilin,
nicastrin, PEN2 and APH1.147 b | In a second pathway, APP is cleaved by
β‑secretase followed by α‑secretase, resulting in the release of several short Aβ
isoforms (Aβ1–16 down to Aβ1–13).81,145 c | In a third pathway, APP is cleaved in the
middle of the Aβ domain by α‑secretase, releasing the large amino‑terminal
derivative sAPPα into the extracellular fluid and CSF and leaving CTFα in the
plasma membrane. α‑Secretase activity has been attributed to the ADAM family of
proteases.148 The p3 peptide, which has been found in cell culture experiments,149
is not present in CSF.75,150 Abbreviations: Aβ, amyloid‑β; AICD, amyloid precursor
protein intracellular domain; APP , amyloid precursor protein; CSF, cerebrospinal
fluid; CTF, carboxy‑terminal fragment; sAPP , soluble APP extracellular domain.
© 20 Macmillan Publishers Limited. All rights reserved10
nature reviews | neurology
aDvanCe OnLine PuBLiCatiOn | 3
neurofibrillary tangle pathology, and axonal degeneration
(supplementary Figure 2 online). Few methods have
been available in living patients for measuring amyloid
plaque and neurofibrillary tangle loads, and the sever
ity of neuronal and synaptic degeneration. thus, most
studies of core biomarkers for aD have looked for corre
lations between CsF biomarkers measured during life
and neuropathological findings at autopsy. the time lag
between CsF tapping and autopsy, and other methodo
logical issues of such studies, have made such correlations
difficult to find.
One autopsy study found a correlation between post
mortem ventricular CsF aβ1–42 and amyloid plaque
load,15 and another study demonstrated that aβ1–42 in
lumbar CsF antemortem also correlated with amyloid
plaque load at autopsy.16 the development of aβ Pet
ligands, notably Pittsburgh compound B (PiB), has
enabled direct visualization of the fibrillar aβ load in
the brain when a patient is still alive. several studies
have reported a relationship between 11CPiB retention
and CsF aβ1–42, with high 11CPiB binding correlating
with low CsF aβ1–42 levels.17,18 these data support the
notion that CsF aβ1–42 levels reflect fibrillar aβ1–42 levels
and amyloid plaque load in the brain. the most widely
accepted explanation for the reduction in CsF aβ1–42 in
aD is that aggregation of aβ into plaques (and, hence,
retention of the peptide in the brain parenchyma) results
in reduced availability of aβ to diffuse into the CsF.
Data from various studies suggest that CsF total tau
(ttau) levels reflect the intensity of neuronal and axonal
degeneration and damage in the brain. Patients who
experienced an acute disorder, such as stroke or brain
trauma, were reported to have a transient increase in CsF
ttau, the magnitude of which correlated both with the
extent of tissue damage and the probability of poor clini
cal outcome.19–21 High CsF ttau has also been associated
with fast progression from MCi to aD,22 and with rapid
cognitive decline and a high mortality rate in patients
with aD.23,24 the highest increases in CsF levels of ttau,
however, have been reported in disorders with the most
rapid neuronal degeneration, such as Creutzfeldt–Jakob
disease.25 One study found that CsF ttau correlates with
postmortem neurofibrillary tangle load,16 suggesting that
neurofibrillary tanglebearing neurons might contribute
to the CsF level of ttau.
CsF levels of phosphorylated tau (ptau) seem to reflect
both the phosphorylation state of tau and the formation
of neurofibrillary tangles in the brain. in some studies
that involved measuring ptau in CsF samples taken
during life and at autopsy, correlations were reported
for CsF ptau—phosphorylated at thr181 (ptau181)
or thr231 (ptau231)—with neocortical neurofibrillary
tangle pathology, as well as with the rate of hippo campal
atrophy in the brain.16,26,27 High CsF ptau181 has also
been associated with a fast progression from MCi to
aD,22 and with rapid cognitive decline in aD.23
several studies have reported strong correlations
between the levels of CsF ttau and ptau in patients
with aD and in healthy elderly individuals.28,29 such
co rrelations have not been found in Creutzfeldt–Jakob
disease or acute stroke. in both of these conditions,
ttau is found at very high levels, reflecting intense neu
ronal damage, while ptau levels are normal.19,30 Data
are accumulating that the main use of ptau could be in
differentiat ing aD from other forms of dementia.31–33
Assay development for core biomarkers
the discovery that aβ is produced during normal cell
metabolism and secreted into the CsF was the basis for
developing an aβ biomarker for aD.34 the subsequent
finding that aβ1–42 is the most abundant species of aβ
in amyloid plaques led to the development of assays for
this aβ isoform.35 studies using various enzymelinked
immunosorbent assays (eLisas) have shown that patients
with aD consistently exhibit a decrease in CsF aβ1–42, to
approximately 50% of the levels found in agematched
healthy elderly individuals.36
tau has several isoforms and numerous phosphory
lation sites (Figure 2). the most common eLisa for ttau
uses monoclonal antibodies that detect all isoforms of
tau independently of phosphorylation state.28 By use of
this assay, numerous studies have reported that patients
with aD have an increase in CsF ttau of around 300% of
the levels found in healthy elderly individuals.36
SS S SS S SSS S SS S SS SS SSSSS S
TTT TTT T TT
Exon 2 Exon 3Exon 10
Figure 2 | Tau isoforms and phosphorylation sites. Tau is an axonal protein that
binds to microtubules, promoting microtubule assembly and stability. Tau
expression is high in nonmyelinated cortical axons, especially in the regions of the
brain that are involved in memory consolidation, such as the limbic cortex.151
a | Six isoforms of tau exist as a result of alternative splicing of exons 2, 3 and
10.152 These isoforms contain three or four microtubule‑binding domains (green
boxes; the fourth domain is in exon 10). b | Numerous threonine and serine
phosphorylation sites have been identified in tau, but the level of phosphorylated
tau in cerebrospinal fluid is usually quantified by measuring phosphorylation at
Thr181 or Thr231.37,38 Tau phosphorylation is regulated by the balance between
multiple kinases and phosphatases.153 Hyperphosphorylated tau sequesters
normal tau and other microtubule‑associated proteins (MAP1 and MAP2), and
causes disassembly of microtubules, which disrupts axonal transport.
Furthermore, hyperphosphorylated tau becomes prone to aggregation into
insoluble fibrils called paired helical filaments, which can form larger aggregates,
namely neurofibrillary tangles.153 Both the loss of microtubule stabilization and
neurofibrillary tangle formation compromise neuronal and synaptic function,
although whether tau hyperphosphorylation and aggregation is a cause or
a consequence of Alzheimer disease is unknown. Abbreviations: S, serine;
© 20 Macmillan Publishers Limited. All rights reserved 10
4 | aDvanCe OnLine PuBLiCatiOn
the most common eLisas for ptau in CsF use anti
bodies specific for either ptau181 or ptau231 (Figure 2).37,38
studies using these assays have consistently reported a
marked increase in CsF ptau in patients with aD.36 a
study directly comparing the two ptau assays found that
they had similar diagnostic performances.31
the diagnostic accuracy of CsF ttau, ptau and aβ1–42
when considered together—in terms of identifying cases
of aD or prodromal aD, and for differen tiating aD from
other disorders—is higher than for any of these bio markers
alone.39–42 thus, a multiparameter assay was developed
to simultaneously quantify these CsF bio markers. this
assay, which was based on xMaP® technology (Luminex,
austin, tX, usa),43 has been used in several large multi
center studies of CsF biomarkers in aD, and showed a
high diagnostic performance.39,44,45 interestingly, the
absolute values for the biomarker levels detected in CsF
vary between the xMaP® system and eLisa methods.43
several factors probably account for this finding, includ
ing differences in the pairs of antibodies used in the assays,
the methods for coupling the antibodies to the beads and
coating the plates, and the calibra tors and incubation
conditions. Correction factors can be used to allow direct
comparisons of the results from the xMaP® system and
the various eLisa methods.43,44
Performance of core biomarkers
numerous studies have found that patients with aD have
a marked increase in CsF levels of ttau and ptau and a
substantial reduction in aβ1–42 levels. each of these bio
markers has been reported to differentiate patients with
aD from healthy elderly individuals with 80–90% sensi
tivity and specificity (Box 1).36,46 Moreover, ttau, ptau and
aβ1–42 CsF levels have been found to be normal in several
important differential diagnoses of aD, including depres
sion and Parkinson disease.47 a combined analysis of two
or more of these biomarkers more accurately diagnoses
aD than any of these biomarkers alone.39,40,44 For example,
one study showed that a combined analysis of aβ1–42 and
ttau improved the sensitivity of a diagnosis of aD from
78–84% (using one of these biomarkers alone) to 86% and
the specificity from 84–90% to 97%.40
CsF ptau aids the differentiation of aD from other
dementias, including frontotemporal dementia and
dementia with Lewy bodies.31 the diagnostic performance
Familial ADSporadic AD
Life-long increase in total Aβ or
Aβ1–42 production, leading to
gradual Aβ accumulation
Increase in tendency for
Aβ misfolding, resulting in
Failure of Aβ clearance or
degradation, leading to
gradual Aβ accumulation
Failure of chaperones or factors
promoting correct Aβ folding,
resulting in higher amyloidogenicity
Aging in concert with genetic
and environmental risk factors
APP or presenilin gene mutations;
APP gene duplication
Increase in Aβ oligomers
Gradual deposition of Aβ oligomers and
intermediates as diffuse amyloid plaques
Further accumulation of Aβ aggregates
and fibrils into amyloid plaques
Neuronal and synaptic dysfunction,
as well as neurotransmitter deficits
Altered kinase and
phosphatase activity, resulting
in neurofibrillary tangle formation
Impaired LTP, leading to
Figure 3 | The amyloid cascade hypothesis of AD. The amyloid cascade hypothesis states that an imbalance between the
production and clearance of Aβ in the brain, causing an increase in the level of the peptide, is the initiating event in AD, and
ultimately leads to neuronal degeneration and dementia.6 An increase in production of either total Aβ or the amyloidogenic
Aβ1–42 isoform is well established in familial AD, but only limited evidence exists for a specific disturbance in Aβ clearance in
sporadic AD. In both familial and sporadic AD, soluble Aβ is believed to undergo a conformational change that renders it
prone to aggregation into soluble oligomers and the larger insoluble fibrils found in plaques. The specific molecular
mechanisms underlying this conformational change are largely unknown. Fibrillar Aβ deposited in plaques might be
neurotoxic; however, synaptic loss and clinical progression of the disease mainly correlate with soluble Aβ levels.154 Data
suggest that soluble Aβ oligomers might inhibit LTP in the hippocampus and, hence, disrupt synaptic plasticity.82 Tau
phosphorylation and subsequent neurofibrillary tangle formation, as well as inflammation and oxidative stress, are regarded
as downstream events. Abbreviations: Aβ, amyloid‑β; AD, Alzheimer disease; APOE, apolipoprotein E; APP, amyloid precursor
protein; LTP, long‑term potentiation.
© 20 Macmillan Publishers Limited. All rights reserved10
nature reviews | neurology
aDvanCe OnLine PuBLiCatiOn | 5
of CsF biomarkers in differentiating aD from other
dementias, however, is far from optimal. several factors
could explain this finding. First, most studies of CsF
biomarkers are based on clinically diagnosed cases,
which introduces a relatively large percentage of mis
diagnoses.48,49 second, a sizeable percentage of elderly
individuals without dementia have enough amyloid
plaques and neurofibrillary tangles to warrant a neuro
pathological diagnosis of aD.50,51 Last, aD exhibits a large
overlap in pathology with some other forms of dementia,
notably dementia with Lewy bodies and vascular demen
tia.52–54 this overlap in pathology essentially precludes the
possibility of finding CsF biomarkers that have close to
100% sensitivity and specificity for aD.
Prodromal Alzheimer disease
studies have consistently shown that the combination
of ttau, ptau and aβ1–42 has a high predictive value for
identifying cases of prodromal aD in patients with MCi,46
with one study reporting a sensitivity of 95% (Box 1).39
this high predictive value has been verified in large
multicenter studies, including the alzheimer’s Disease
neuroimaging initiative study,45 the DesCriPa study,55
and the swedish Brain Power project.44 the results from
these studies show that CsF biomarkers might be valu
able clinical diagnostic tools for identifying prodromal
aD in individuals with cognitive impairment.
Presymptomatic Alzheimer disease
some studies have examined whether CsF bio markers
might be useful in predicting aD in the preclinical stage of
the disease. two populationbased studies found a marked
reduction in CsF aβ1–42 levels in cognitively normal elderly
people who later developed aD, although no changes were
observed in CsF ttau or ptau.56,57 in addition, a clinical
study reported that CsF aβ1–42, but not ttau and ptau,
predicted cognitive decline in healthy elderly individuals.58
Moreover, asymptomatic individuals with familial aD
mutations had low CsF aβ1–42,59 yet high CsF ttau and
ptau.60 together, these results support earlier animal data
suggesting that the amyloidogenic process (the process of
generating aβ) is upstream of tau pathology.61,62
a large study showed that cognitively normal elderly
individuals who exhibited cortical 11CPiB binding on
Pet had low CsF aβ1–42 levels, although the same study
revealed that low CsF aβ1–42 was also found in some
indivi duals who did not exhibit 11CPiB binding.63 these
findings might be explained by the fact that 11CPiB binds
fibrillary aβ, but not the aβ oligo mers or diffuse plaques
that are found in the earliest stages of the disease process.64
Furthermore, these data suggest that CsF aβ1–42 might
predict aD in its very early stages in cognitively normal
elderly individuals. the overlap in variation between CsF
aβ1–42 levels in individuals with pre symptomatic aD and
healthy elderly people, however, might be too large for
presymptomatic aD to be predicted in individual cases.
Moreover, the use of biomarkers to predict aD in asympto
matic people is not warranted until registered drugs with
distinct diseasemodifying effects, and few adverse effects,
Validated Alzheimer disease cases
the diagnostic performance of CsF biomarkers has been
examined in several patient series in which the diagnosis
of dementia was confirmed at autopsy. in these studies,
the combination of CsF ttau, ptau and aβ1–42 differen
tiated people with aD from both cognitively normal
elderly individuals and cases of other dementias—
including dementia with Lewy bodies, frontotemporal
dementia and vascular dementia—with high specificity
Table 1 | CSF biomarkers for AD
Biomarker Pathogenic process Change in biomarker level
CSF cell count InflammationUnchanged155
CSF cell count is used to exclude infectious disorders13
BBB function Unchanged in cases of pure
AD;12 mild to moderate increase
in AD with concomitant
Increase in CSF:serum albumin ratio is an indicator of BBB damage;11
BBB damage is found in CNS infections (for example, neuroborreliosis),
inflammatory disorders (for example, Guillain–Barré syndrome), brain
tumors and cerebrovascular disease (including vascular dementia)13
IgG or IgM index;
IgG or IgM
These analyses are used to exclude cases with inflammatory
(for example, multiple sclerosis or cerebral systemic lupus
erythematosus) and chronic infectious (for example, Borrelia
encephalitis or syphilis) disorders13
Amyloidogenic pathway of
Marked reduction in AD and
CSF Aβ1–42 is the central CSF biomarker for brain Aβ metabolism and
plaque formation;15,17,18 low CSF Aβ1–42 is found in patients with dementia
with Lewy bodies157
Tau phosphorylation Marked increase in AD and
High CSF p‑tau has only been found in AD;36,46 CSF p‑tau181 and p‑tau231
levels correlate tightly and give similar diagnostic accuracy31
Marked increase in AD and
High CSF t‑tau is found in disorders with acute brain damage, such as
stroke, trauma and encephalitis;19–21 very high CSF t‑tau, together with
normal p‑tau, is found in cases of Creutzfeldt–Jakob disease30
Abbreviations: Aβ, amyloid‑β; AD, Alzheimer disease; BBB, blood–brain barrier; CSF, cerebrospinal fluid; p‑tau, phosphorylated tau; t‑tau, total tau.
© 20 Macmillan Publishers Limited. All rights reserved 10
6 | aDvanCe OnLine PuBLiCatiOn
and sensitivity.33,45,65–67 thus, CsF biomarkers have been
validated in patient series with a neuropathological
followup, showing similar or better discriminatory power
than in patient series with only clinical diagnoses.
novel candidate biomarkers
Many publications can be found for candidate CsF aD
biomarkers other than aβ and tau, although the initially
promising results from such studies have often not been
reproduced. Here, we review novel biomarkers that have
shown high sensitivity and specificity for aD in at least two
independent studies. we also discuss selected candidate
biomarkers related to aβ and aPP metabolism.
aβ is generated following the sequential actions of
βsecretase and γsecretase on aPP (Figure 1). the main
enzyme responsible for βsecretase activity is βsite aPP
cleaving enzyme 1 (BaCe1). increases in BaCe1 expres
sion and enzymatic activity have been reported in aD brain
tissue at postmortem.68,69 BaCe1 can also be measured
in the CsF, and increases in BaCe1 concentration and
activity have been found in patients with aD and in cases
of prodromal aD.70–72 together, these data suggest that
up regulation of BaCe1 might be an early event in aD.
Amyloid precursor protein isoforms
During aPP processing, the large soluble aminoterminal
domain of aPP, saPPα or saPPβ (depending on whether
aPP is first cleaved by αsecretase or βsecretase, respec
tively), is secreted into the extracellular space and also
into the CsF (Figure 1). in sporadic aD and MCi, CsF
levels of both saPPα and saPPβ have been reported to
remain unaltered or to increase slightly.71,73,74 Despite
the absence of a consistent change in saPP levels in aD,
these CsF biomarkers might be valuable tools in treat
ment trials for monitoring the effect of a drug on aPP
processing (table 2).
Truncated amyloid‑β isoforms
aβ1–40 is the most abundant aβ isoform in CsF.75 in aD
and MCi, no major change has been detected in CsF
aβ1–40; however, a marked decrease has been observed
in the aβ1–42:aβ1–40 ratio, and this change was more pro
nounced than the reduction in CsF aβ1–42.76,77 Other
carboxy terminal truncated aβ peptides, including
aβ1–37, aβ1–38 and aβ1–39, have also been identified in the
CsF of patients with aD.78 an increase in the CsF level
of aβ1–38 was found together with a decrease in aβ1–42 in
such indivi duals,78,79 suggesting that the aβ1–42:aβ1–38 ratio
might improve di agnostic accuracy in cases of aD.
several short carboxyterminal truncated aβ isoforms
have been identified and quantified by a combination of
immunoprecipitation with an antiaβ monoclonal anti
body and matrixassisted laser desorption– ionization
timeofflight mass spectrometry.75 a marked increase
in aβ1–16 together with the expected decrease in aβ1–42
was reported in CsF from patients with aD.80 Data from
experimental studies showed that the short aβ isoforms
aβ1–14, aβ1–15 and aβ1–16 were produced by a novel pathway
of aPP processing involving the concerted actions of
βsecretase and αsecretase, whereas the longer isoforms
(aβ1–17 up to aβ1–42) were produced by the γsecretase
pathway (Figure 1).81
the aggregation of soluble aβ peptides to form insoluble
fibrillar aβ in plaques has long been regarded to be the
central pathogenic event in aD (Figure 3). experimental
data, however, have suggested that soluble aβ oligomers
might inhibit longterm potentiation and, thereby, have a
role in aD pathogenesis.82 thus, CsF aβ oligomers might
be important core biomarkers for aD.
some preliminary studies on aβ oligomers in CsF have
been published. in one study, antibodies coupled to Dna
tagged nanoparticles were used to capture aβ oligomers
from the CsF of patients with aD and healthy aged
matched controls at postmortem. PCrbased amplifica
tion revealed a higher assay signal in the CsF from patients
with aD than from the control samples.83 a study using
flow cytometry technology also suggested the presence of
aβ oligomers in lumbar CsF from neuro logical patients,
although no data on the diagnostic utility of this technique
in the context of aD were presented.84 Following immuno
precipitation of CsF samples using an antiaβ antibody,
immuno blotting revealed a weak band migrating at the
size expected for aβ dimers in some samples from patients
with aD and cognitively normal elderly individuals. no
consistent change in this band, however, could be found in
the aD group.85 thus, although aβ oligomers are attrac
tive aD biomarker candi dates, several issues relating to
these molecules persist. the level of these aβ species in
CsF seems to be very low in comparison with aβ mono
mers. importantly, mass spectrometry analyses are needed
to verify that the signals measured using the various tech
niques described above actually represent changes in aβ
oligomers. Furthermore, assays suitable for large clinical
studies have yet to be developed for these molecules.
Endogenous amyloid‑β autoantibodies
several studies have reported the existence of natu
rally occurring aβ antibodies (either in free form or in
complex es with aβ) in CsF and/or blood. the results
Box 1 | Criteria for evaluation of Alzheimer disease biomarkers
Studies evaluating the diagnostic performance of a biomarker for Alzheimer disease
should include determination of the molecule’s sensitivity, specificity, positive
predictive value and negative predictive value for the disorder.158 Sensitivity refers
to the capacity of a biomarker to identify patients who have disease (the number of
true positive cases divided by all cases with disease), while specificity refers to the
capacity of a biomarker to identify patients who do not have disease (the number
of true negative cases divided by all cases without disease). The positive predictive
value refers to the percentage of cases with a positive test who prove to have the
disease (the number of true positive cases divided by all cases with a positive test),
while the negative predictive value refers to the percentage of cases with a negative
test who prove not to have the disease (the number of true negative cases divided
by all cases with a negative test). According to the criteria for an ideal Alzheimer
disease diagnostic biomarker, outlined by the Ronald and Nancy Reagan Research
Institute–National Institute on Aging Working Group, sensitivity and specificity
should exceed 80%, whereas the predictive values should be ≥80%.14
© 20 Macmillan Publishers Limited. All rights reserved10
nature reviews | neurology
aDvanCe OnLine PuBLiCatiOn | 7
from these studies, however, have been inconsistent, with
increases,86,87 decreases88,89 or unchanged90 titers of such
antibodies all having been reported in cases of aD. these
studies were all based on assays that could not differen
tiate between antibodies against various aβ isoforms
or different assembly states of the peptide (monomeric
versus oligomeric aβ). Human plasma has been shown to
contain autoantibodies against a broad range of aβ pep
tides, including oxidized, pyroglutamate and mutated vari
ants.91 antibody reactivity was highest for oligomeric aβ,
but the level of this reactivity did not differ between cases
of aD and controls.91
Neuronal and synaptic markers
neuronal and synaptic proteins could prove to be valu
able CsF biomarkers for aD, as these molecules might
corre late with cognitive function and disease progres
sion. visininlike protein 1 (vLP1) is a highly expressed
neuronal calcium sensor protein that was identified by
gene array analyses during a search for brainspecific
protein biomarkers.92 in the first clinical study of this
protein, a marked increase in CsF vLP1 was found in
patients with aD. Moreover, the diagnostic performance
of vLP1 in distinguishing patients with aD from healthy
elderly indivi duals was similar to that of CsF ttau, ptau
and aβ1–42, with a sensitivity and a specificity both close
to 80%.93 CsF levels of vLP1 were high in patients with
aD who carried the apolipoprotein e (APOE) ε4 allele (a
genetic risk factor for aD), and also negatively correlated
with MiniMental state examination scores. thus, vLP1
is a promising candidate CsF biomarker for aD.
neurofilaments are structural components of axons,
and show particularly high expression in large myelinated
axons.94 accordingly, high CsF levels of neuro filaments
are found in disorders with subcortical pathology, such as
vascular dementia and normal pressure hydro cephalus.95,96
High CsF levels are also found in frontotemporal demen
tia, while normal levels are found in most patients with
aD.97 CsF neurofilament proteins might, therefore, be
useful in differentiating between aD, frontotemporal
dementia, and subcortical dementia disorders.
synaptic protein biomarkers should ideally reflect
synaptic functioning and, hence, cognition. several
presynaptic and postsynaptic proteins have been iden
tified in CsF using a procedure that involved protein
precipita tion followed by liquidphase isoelectric focusing
and western blotting. these proteins included raB3a,
synaptotagmin, growthassociated protein (GaP43),
synaptosomalassociated protein 25 and neurogranin.98
an immunoassay for GaP43 revealed increased levels of
the protein in CsF from patients with aD compared with
controls and individuals with frontotemporal dementia.99
Furthermore, the same study showed that GaP43 and
ttau levels in CsF were positively correlated, suggesting
that both biomarkers reflect axonal and synaptic degenera
tion. when validated assays are available for measuring
synaptic proteins in CsF, these proteins might serve as
tools for monitoring the effect of novel drug ca ndidates
on synaptic function in clinical trials.
aD pathogenesis includes free radicalmediated injury to
neurons (supplementary Figure 2 online). Lipid peroxi da
tion is an important consequence of free radicalmediated
damage and leads to the generation of F2isoprostanes,
which might serve as biomarkers for this pathogenic mecha
nism. several studies have shown that CsF F2isoprostane
levels are increased in patients with aD compared with
healthy elderly indivi duals or patients with nonaD forms
of dementia.100 Levels of CsF F2isoprostanes have also
Table 2 | Applications of CSF biomarkers in AD clinical trials
ApplicationDetails Time point for usePossible biomarker and their role
CSF biomarkers could be used in clinical
trials to improve diagnostic accuracy in trial
participants, enabling patient cohorts to be
enriched with cases of AD
Before trial initiation
High T‑tau and P‑tau and low Aβ1–42 are indicative of AD
of AD cases
AD cases with CSF biomarker evidence of a
disturbance in Aβ metabolism might be more
responsive to anti‑Aβ drugs than patients
who do not exhibit such a disturbance
Post hoc analysis
Aβ1–42 might be used to stratify cases in trials of anti‑Aβ
disease‑modifying drug candidates; p‑tau might be used to
stratify cases in trials of drugs that aim to reduce tau
phosphorylation and neurofibrillary tangle pathology
Anti‑Aβ drug candidates, such as Aβ
immunotherapy, might elicit adverse
effects, such as meningoencephalitis
or vasogenic oedema
Baseline evaluation and
assessment during trial
CSF cell count, IgG or IgM index and IgG or IgM oligoclonal
bands are standard measures for identifying and monitoring
inflammatory processes, such as meningoencephalitis, in the
CNS; the CSF:serum albumin ratio is the standard measure to
identify and monitor a disturbance in the blood–brain barrier,
which can lead to cerebral edema
Theragnostics CSF biomarkers might indicate whether a
drug has an effect on the molecular
pathology of AD in living patients
Baseline evaluation and
at time points throughout
the trial, including the
last week of the study
Aβ1–42 is the main biomarker for Aβ metabolism and
deposition; APP isoforms (sAPPα and sAPPβ) and BACE1
activity might be valuable in clinical trials of BACE1 inhibitors;
p‑tau is the main biomarker for monitoring the phosphorylation
state of tau; t‑tau might be a valuable biomarker for identifying
and monitoring a downstream effect on the intensity of
neuronal or axonal degeneration
Abbreviations: Aβ, amyloid‑β; AD, Alzheimer disease; APP , amyloid precursor protein; BACE1, β‑site APP cleaving enzyme 1; CSF, cerebrospinal fluid; p‑tau, phosphorylated tau; sAPP , soluble APP
extracellular domains; t‑tau, total tau.
© 20 Macmillan Publishers Limited. All rights reserved 10
8 | aDvanCe OnLine PuBLiCatiOn
been demonstrated to increase in cognitively impaired
individuals with prodromal aD,101 and in asympto matic
carriers of familial aD mutations.60 By contrast, the
results from studies of F2isoprostanes in plasma have
been conflicting, probably because the contribution of
brainderived F2isoprostanes to the total level of these
mol ecules in plasma is much smaller than the co ntribution
from peripheral ly derived F2isoprostanes.100
roles in clinical trials
aside from their potential as tools for clinical diagnosis,
CsF biomarkers might be valuable in drug development.
such biomarkers could be used as diagnostic markers for
enriching the number of aD cases, for patient stratifi
cation, as safety markers, and to detect and monitor the
biochemical effects of drugs (table 2).
Enrichment of AD cases
Making a diagnosis of aD during the early stages of the
disease is a great challenge for clinicians, as patients with
MCi only have a mild disturbance in episodic memory.
Moreover, in such patients, other symptoms of aD can be
absent or seem vague. the only clinical method available
for determining which patients with MCi have prodromal
aD is to follow their cognitive function over several years.
even at specialized academic centers, however, the accu
racy of the clinical diagnosis of aD in cases that have been
followed up for several years is relatively low, with sensi
tivity and specificity values of 70–80%.102 these figures
are considerably lower for patients with early aD,103 and
in primary care settings.104
Clinical trials of cholinesterase inhibitors in patients
with MCi have failed to find any marked benefit of these
drugs. the clinical end point in these trials was a reduc
tion in the conversion rate to aD.105 these studies involved
patients with unspecified MCi, meaning that around half
the participants did not have prodromal aD, and would
not have converted. thus, the inclusion of such patients
might have seriously affected any possibility of identifying
clinical effects of the drugs.106 the addition of a positive
CsF biomarker as an inclusion criterion in MCi trials will
increase the proportion of individuals with under lying
aD pathology and, thereby, increase the possibility of
identifying a positive effect of a drug (table 2).
aD is a heterogeneous disorder, both at clinical and
neuropathological levels.5 thus, the effectiveness of any
potential diseasemodifying drug could plausibly vary
between subgroups of patients with this disease. indeed,
the effectiveness and adverse effects of one passive aβ
immunotherapy for aD were reported to differ between
APOE ε4 carriers and noncarriers.107
as CsF biomarkers reflect the central pathogenic pro
cesses in aD, these molecules might be used in post hoc
data analyses of clinical trials to stratify patient cohorts
on the basis of underlying pathology. indeed, a patient
subgroup with a certain biomarker trait that indicates
amyloid plaque pathology, such as low CsF aβ1–42, might
be more responsive to antiaβ diseasemodifying drugs
than a subgroup of patients with normal CsF aβ1–42 levels
Clinical trials of diseasemodifying treatments for aD
have been hampered by adverse effects. in the clini
cal trial of the aβ vaccine an1792, a small but notable
number of patients developed meningoencephalitis, and
treatment with the passive aβ immunotherapy aaB001
led to vasogenic edema in some individuals.107,108 CsF
analysis is a standard method for diagnosing encepha
litis and BBB damage associated with disorders causing
edema,11,13 and could be employed usefully in clinical trials
of aDmodifying drugs.
analysis of CsF taken from patients at baseline, before
treatment, could be useful in clinical trials for identify
ing and excluding individuals with chronic infectious or
inflammatory Cns disorders that can mimic aD, such as
neuroborreliosis. the inclusion of such cases in trials could
result in the erroneous conclusion that an adverse effect,
such as encephalitis, was related to the drug being tested.
Baseline CsF samples can also be used in comparisons
with CsF removed after treatment initiation. the benefit
of such comparisons is that even minor inflammatory
activation within the Cns, as a result of adverse effects of
the drug, can be identified. thus, CsF biomarkers could
allow safety monitoring during clinical trials (table 2).
Longitudinal CsF sampling during the treatment period
might indicate whether a certain drug induces harmful
immune activation over the long term.
the effect of diseasemodifying antiaβ drugs on amyloid
plaque pathology is commonly evaluated in aD trans
genic mice; however, these animal models have a low
predictive power for treatment success in patients with
sporadic aD.5 Biomarkers might help bridge the gap
between animal studies and large clinical trials by pro
viding a means of evaluating whether a drug has a true
diseasemodifying effect in humans in smallscale clinical
studies. Only the most promising drug candidates would
then be selected for further study, thereby improving the
success rate of large phase ΙΙ and ΙΙΙ trials.
in slowly progressive disorders such as aD, the clini
cal evaluation of a drug by use of rating scales requires
large patient numbers and extended treatment periods.
For drugs with symptomatic effects, such as cholinest
erase inhibitors, an improvement in cognitive function
can be expected in the short term (Figure 4). By contrast,
diseasemodifying drugs cannot be expected to have an
early effect on symptoms. instead, such therapies might
lead to a reduction in the rate of cognitive decline over
several years (Figure 4). thus, the number of patients
needed to detect an effect on cognition is probably larger,
and the treatment period longer, for a diseasemodifying
drug than for symptomatic therapies.
Biomarkers that are used to identify and monitor the
biochemical effect of drugs are called therag nostic markers.
such biomarkers can be used to identify and monitor both
the specific effect of the drug and downstream effects on
© 20 Macmillan Publishers Limited. All rights reserved10
nature reviews | neurology
aDvanCe OnLine PuBLiCatiOn | 9
pathogenic mechanisms (table 2). trials employing therag
nostic biomarkers can be based on relatively small patient
numbers and short treatment periods and, thus, might be
valuable for deciding whether to proceed with large and
expensive phase ΙΙ or ΙΙΙ clinical trials. theragnostic bio
marker trials are feasible in aD, as CsF ttau, ptau and
aβ1–42 levels have shown low intra individual variability
over time in longitudinal samples.109,110 some of these
biomarkers might also serve as substitutes for clinical end
points (surrogate bio markers), although this possibility
needs to be evaluated in fullscale clinical trials. Finally,
theragnostic bio markers are important for regulatory
purposes, as a drug can only really be labeled as disease
modifying if it has an effect on cognition, and if biomarker
evidence can be presented that the drug affects the central
to date, only preliminary evidence exists to suggest that
CsF biomarkers might be useful as theragnostic markers.
importantly, cholinesterase inhibitors and lithium—
drug candidates with no proven effect on the molecular
pathogenesis of aD—had no effect on aD CsF core bio
markers.109,112 Data from animal studies, how ever, demon
strated that γsecretase inhibitor treatment resulted in a
reduction in cortical, CsF and plasma levels of aβ.113,114
similarly, in nonhuman primates, BaCe1 inhibitor treat
ment resulted in a reduction in CsF aβ1–42, aβ1–40 and
saPPβ levels.115 whether CsF aβ levels in patients with
aD will be altered in response to treatment with efficacious
antiaβ drugs remains unclear. a phase iia study of the aβ
clearanceenhancing compound PBt2 demonstrated that
CsF aβ1–42 underwent a dosedependent reduction during
the treatment period.116 Furthermore, data from a clinical
study of the amyloidtargeting drug phenserine also sug
gested that CsF aβ might be of value in the evaluation of
treatment effects.117 in the interrupted phase iia an1792
trial, however, no change in CsF aβ1–42 was detected in
treated patients, despite a decrease towards normal levels
of the downstream biomarker ttau.118 a clinical trial of
the γsecretase inhibitor LY450139 also failed to find
any effect on CsF aβ1–42 levels, as measured by eLisa, in
patients with aD.119 However, acute treatment with the
same compound in young healthy volunteers resulted
in a clear inhibitory effect on the aβ production rate, as
determined by measuring the ratio of newly synthesized
(isotopelabeled) aβ to preexisting (unlabeled) aβ in
CsF.120 several other clinical trials of diseasemodifying
drug candidates that include biomarkers as end points
are currently ongoing. these trials will provide further
evidence to indicate whether biomarkers can be used to
assess disease modification, and as surrogate markers for
predicting clinical outcomes.
efforts to find reliable biomarkers for aD in peripheral
blood have met with little success. several candidate blood
biomarkers have been proposed, yet changes in the levels of
these molecules have proved difficult to verify in indepen
dent studies. in the section below, we focus on plasma aβ,
which has been the most extensively ex amined pe ripheral
biomarker for aD. we also review some explorative pilot
studies of novel plasma protein biomarkers that have
shown promising results.
Many studies have examined plasma aβ as a biomarker
for aD; however, the findings from these studies have
been contradictory. some investigators have reported
slightly higher aβ1–42 or aβ1–40 plasma levels in patients
with aD than in healthy agematched controls, although
most studies have found no change in plasma aβ between
these groups.121 in addition, studies examining the value of
plasma aβ tests for predicting aD in cognitively normal
elderly individuals have shown a broad overlap in plasma
aβ1–42 and aβ1–40 levels between indivi duals with pre
clinical aD and those people who do not develop aD.
some studies have reported that a high level of plasma
aβ1–42 or a large aβ1–42:aβ1–40 ratio are risk factors for future
aD, while others studies have reported opposing data.122–125
these discouraging results are probably explained by the
fact that plasma aβ is derived from peripheral tissues,
and does not reflect brain aβ turnover or metabolism.76
Furthermore, the hydrophobic nature of aβ makes the
peptide bind to plasma proteins, which could result in
epitope masking and other an alytical interferences.126
novel blood biomarkers
several promising novel blood biomarkers for aD have
been documented. the combined multivariate analysis of
18 plasma signaling and inflammatory proteins accurately
identified patients with aD and predicted the onset of
Time after initiating treatment (months)
Change in cognitive function (%)
Figure 4 | Evaluation of Alzheimer disease therapies by cognitive scales. Theoretical
differences in cognitive effect of treatments for Alzheimer disease are shown over a
24 month period. During treatment with symptomatic drugs, such as cholinesterase
inhibitors, an improvement can be observed in cognitive function during the initial
phase of therapy (~6 months). Beyond this initial phase a decline in cognition
occurs, although a marked difference in cognitive function exists between patient
treated with symptomatic drugs and placebo‑treated individuals.159 A less
pronounced decline in cognitive function can be seen in patients treated with
disease‑modifying drugs than in individuals receiving placebo; however, no initial
improvement in cognition is observed in the former group. Thus, to find an effect of a
disease‑modifying drug on cognition in a clinical trial, the number of patients needed
is larger and the treatment period markedly longer than in a trial of a symptomatic
drug. CSF biomarkers might be valuable in clinical trials of disease‑modifying drugs,
by providing objective evidence that a drug affects the underlying pathogenic
processes. Indeed, such evidence, alongside an effect on cognitive decline, is
essential to claim that a drug is disease modifying.111
© 20 Macmillan Publishers Limited. All rights reserved10
10 | aDvanCe OnLine PuBLiCatiOn
aD in individuals with MCi.127 this panel of proteins was
identi fied after screening 120 known signaling proteins
using a filterbased, arrayed sandwich eLisa. Further
independent studies are needed to examine whether this
set of proteins is the optimal combination of plasma bio
markers for diagnosing prodromal aD, and to further
assess the diagnostic value of this approach. another study
that used explorative proteomics tech nology identified
aDassociated increases in the plasma levels of comple
ment factor H and α2macroglobulin—findings that
were replicated using semiquantitative immuno blotting
techniques.128 the midregional proatrial natriuretic
peptide:carboxyterminal endothelin1 pre cursor frag
ment ratio has also been reported to be elevated in plasma
from patients with aD.129 if this finding and other results
for candidate plasma biomarkers could be replicated in
independent studies using immunoassay techniques suit
able for routine diagnostic laboratories, plasma protein
panels might serve as screening tests for aD.
CsF biomarkers have a high diagnostic value in the
context of aD. the combination of these biomarkers and
structural (Ct or Mri) and/or functional (single photon
emission Ct [sPeCt] or Pet) brain imaging should
provide increased diagnostic accuracy compared with CsF
biomarkers or one type of imaging used in isolation. to
date, only a few studies have directly examined this pos
sibility. CsF biomarkers combined with either Ct or Mri
measurements of medial temporal lobe atrophy have been
reported to increase the accuracy of aD diagnosis.130–132
in addition, assessment of both CsF biomarker levels and
the degree of structural aDlike abnormalities on Mri
more accurately predicted which patients with amnestic
MCi would convert to aD than either biomarker alone.133
similarly, the measurement of CsF biomarkers alongside
the assessment of regional cerebral blood flow, using the
133Xe inhalation technique or sPeCt, has been shown to
improve diagnostic accuracy of either biomarker alone in
cases of prodromal aD.134,135 Furthermore, although no
study has examined the added diagnostic value of 11CPiB
Pet when combined with CsF biomarkers, a strong
negative correlation exists between the degree of 11CPiB
binding and the CsF level of aβ1–42.17,18
Large multicenter studies are needed to further define
the added diagnostic value when multiple biomarker
modalities are combined. such studies will also provide
information on the optimal brain regions to evaluate by
Mri (for atrophy) and by Pet (for aβ load) in the context
of aD. Complementary data are needed on whether high
resolution Mri scanners and newly developed amyloid
ligands, such as aZD2184, will improve diagnostic sen
sitivity and specificity.136 when implementing these bio
markers in clinical practice, financial considerations will
be of importance. the cost of combined analysis of CsF
ttau, ptau and aβ1–42 is ~us$200, whereas a structural
Mri investigation and a 11CPiBPet scan cost ~$500 and
the current clinical diagnostic criteria for aD were
outlin ed more than 25 years ago by the national institute of
Box 2 | Research criteria for a diagnosis of Alzheimer disease
The diagnostic criteria for probable Alzheimer disease (AD) described below are
based on the core criterion of early memory disturbances together with supportive
criteria that include positive findings for one or more biomarkers.140
Core diagnostic criterion
Evidence of progressive episodic memory impairment lasting >6 months
(reported by patients or informants) that can be verified by objective testing;
memory impairment can be isolated or associated with other cognitive changes
selected supportive criteria
Presence of medial temporal lobe atrophy (in the hippocampus, entorhinal
cortex or amygdala) on MRI, measured by either qualitative rating or quantitative
volumetry, and referenced to a well‑characterized age‑matched population
Positive cerebrospinal fluid biomarker result (low amyloid‑β1–42, high total tau
and/or high phosphorylated tau)
Reduction in glucose metabolism in bilateral temporal parietal regions or increase
in binding of amyloid‑β ligands (18F‑FDDNP or 11C‑labeled Pittsburgh compound B),
as measured by PET
Presence of a familial AD‑causing mutation
selected exclusion criteria
History of sudden onset of symptoms or early symptoms, including gait
disturbances, seizures or behavioral changes
Clinical features of focal neurological signs, such as hemiparesis, sensory loss,
visual field deficits or early extrapyramidal signs
Other medical disorders severe enough to account for memory and related
symptoms—including non‑AD dementia, major depression, cerebrovascular
disease, toxic and metabolic abnormalities—or MRI fluid‑attenuated inversion
recovery or T2‑weighted signal abnormalities in the medial temporal lobe that are
consistent with infectious or vascular insults
Box 3 | The Alzheimer’s Association quality control program
The aim of the quality control program is to standardize cerebrospinal fluid (CSF)
biomarker measurements between both research and clinical laboratories.
Achievement of this aim will increase the analytical precision and improve the
longitudinal stability of biomarker measurements. The program will allow direct
comparisons of biomarker levels between laboratories and, thus, between
The program is run by the Clinical Neurochemistry Laboratory in Gothenburg,
Sweden in conjunction with the Alzheimer’s Association. Biotech companies
and a number of reference laboratories, including the Alzheimer’s Disease
Neuroimaging Initiative Biomarker Core, are also represented. Both research and
clinical CSF laboratories, as well as pharmaceutical companies, are enrolled in
The program is open for generally (commercially) available assay formats, but
not for in‑house assays, and consists of two parts. The first part involves a
standardized flow chart for lumbar puncture and CSF processing (Supplementary
Table 1 online). The second part is an external quality control program, in which
samples (aliquots of pooled CSF) are sent out to the participating laboratories for
CSF biomarker analysis, after which biomarkers levels are entered into a report
form and returned.
The final report for each quality control round includes information on the
measured biomarker levels for the individual laboratory and, for comparison, the
mean and variation in biomarker levels across all laboratories involved in
the program. In addition, the longitudinal stability in CSF biomarker levels for the
individual laboratory, expressed as percent deviation over time, will be reported.
These reports will serve as feedback for the participating laboratories, to identify
whether the level of a biomarker is outside an acceptable range and to note
sudden changes or longitudinal drifts in CSF biomarker levels.
© 20 Macmillan Publishers Limited. All rights reserved10
nature reviews | neurology
aDvanCe OnLine PuBLiCatiOn | 11
neurological and Communicative Disorders and stroke–
alzheimer Disease and related Disorders work Group.
Diagnosing aD using these criteria largely involves exclu
sion of other causes of dementia.137 Moreover, according to
these criteria, a diagnosis of aD cannot be made until the
patient has dementia, which is defined as cognitive symp
toms severe enough to interfere with social or occupational
activities. the Diagnostic and statistical Manual of Mental
Disorders 4th edition and international Classification of
Diseases 10 criteria for aD, which are both used in the
routine diagnosis of this disorder in the clinic, also require
that a patient has dementia before a diagnosis of aD can be
made.138,139 should disease modifying drugs become avail
able for aD, these criteria will all hinder patients in the
early stages of disease from receiving effective therapy.
new research criteria for aD have been constructed to
allow a diagnosis of aD to be made in the early stages of
the disease. these criteria are centered on the clinical iden
tification of episodic memory impairment alongside the
detection of one or more abnormal bio markers, including
Mri, Pet, and CsF biomarkers (Box 2).140 More detailed
guidelines are needed on how biomarkers can be imple
mented in the diagnostic procedure for early aD in the
clinic. such guidelines should provide details of the scales
to be used in measuring memory impairment, the assays
and cutoffs to be employed for CsF biomarkers, the brain
regions (whole brain, hippocampus or entor hinal cortex) to
be evaluated by Mri for atrophy, and the amyloid ligands
to be used and brain regions to be evaluated by Pet. studies
relating to these issues are only just emerging.141
assays for measuring tau and aβ in CsF have been well
validated43,142,143 and singlecenter studies in which samples
have been assayed simultaneously using the same batches
show that the biological variability for these biomarkers is
low.118,119 nevertheless, the levels of these biomarkers
measured in patients have varied in reports from different
research centers, and even between studies that have used
the same assay.44,144 this variation in CsF biomarker levels
between laboratories complicates multi center research
studies and trials, and also precludes the introduction of
generally applicable cutoff levels.
the variation in CsF biomarker levels between centers
is probably the result of variations in clinical procedures—
such as the protocols for lumbar puncture, CsF sample
processing and other laboratory practices—as well as batch
tobatch variation in the biomarker assays. these types of
variation are well known in clinical chemistry, and are
routinely controlled by external control programs. thus,
the Clinical neurochemistry Laboratory in Gothenburg
(sweden), in conjunction with the alzheimer’s association,
has initiated a quality control program for CsF bio markers
(Box 3). this program includes standardization of the
procedures for lumbar puncture, CsF processing and CsF
analysis. standardized protocols should minimize variation
caused by differences in preanalytical and laboratory pro
cedures and, thus, allow direct comparisons of biomarker
levels between labora tories and between publications.
to overcome batchtobatch variation in CsF biomarker
assays, biomarker kit vendors will need to implement new
standards for quality control. assays should exhibit low
overall variability in calibration curves and strict limits of
variability across batches. to achieve these goals, stringent
quality control of critical reagents, including antibodies
and calibrators, is needed. in the long term, the aim is that
the quality control program will serve as the basis for a
more general introduction of CsF biomarkers into routine
clinical practice and multicenter clinical trials.
numerous studies have shown that combined analysis
of the core CsF biomarkers aβ1–42, ttau and ptau can
be used to reliably diagnose patients with aD and iden
tify prodromal aD in cases of MCi. in addition, these
biomarkers fulfill the criteria for an ideal aD diagnostic
biomarker outlined by the ronald and nancy reagan
research institute–national institute on aging working
Group (supplementary table 2 online).14 Basic CsF bio
markers might also serve as tools for identifying patients
with pure aD and to exclude other disorders (table 1).
thus, CsF biomarkers might be useful in a routine clini
cal diagnostic setting; however, the low positive predictive
value of the combined threemarker test in asympto
matic populations suggests that such markers might be
of limited use in screening individuals for aD before
co gnitive deficits appear.
CsF biomarkers might serve as valuable tools in drug
development. CsF aβ1–42, ttau and ptau are being
increasingly implemented as diagnostic markers in clinical
trials to enrich the number of aD cases, while basic CsF
biomarkers are used as safety markers. Finally, smallscale
trials using CsF biomarkers will be valuable for providing
biochemical data that a candidate drug affects aD patho
genesis. such data will be vital for deciding whether large
and expensive phase 2 and 3 trials should go ahead.
We searched PubMed for English language articles on
biomarkers for Alzheimer disease using the following
keywords: “Alzheimer”, “biomarker”, “cerebrospinal fluid”,
“CSF”, “diagnosis”, “plasma”, “serum”, “amyloid”, “tau”,
“treatment” and “therapy”. We also conducted searches
using several keywords relevant to each section. In
addition, we identified papers from references in the
articles retrieved by the initial searches, and re‑read
selected articles from our own archives.
1. Alzheimer, A., Stelzmann, R. A., Schnitzlein, H. N. &
Murtagh, F. R. An English translation of Alzheimer’s
1907 paper, “Uber eine eigenartige Erkankung
der Hirnrinde”. Clin. Anat. 8, 429–431 (1995).
Masters, C. L. et al. Amyloid plaque core protein
in Alzheimer disease and Down syndrome. Proc.
Natl Acad. Sci. USA 82, 4245–4249 (1985).
3. Kang, J. et al. The precursor of Alzheimer’s
disease amyloid A4 protein resembles a cell‑
surface receptor. Nature 325, 733–736 (1987).
Grundke‑Iqbal, I. et al. Abnormal phosphorylation
of the microtubule‑associated protein tau (tau) in
Alzheimer cytoskeletal pathology. Proc. Natl Acad.
Sci. USA 83, 4913–4917 (1986).
5. Blennow, K., de Leon, M. J. & Zetterberg, H.
Alzheimer’s disease. Lancet 368, 387–403
Hardy, J. & Selkoe, D. J. The amyloid hypothesis
of Alzheimer’s disease: progress and problems
on the road to therapeutics. Science 297,
© 20 Macmillan Publishers Limited. All rights reserved10
12 | aDvanCe OnLine PuBLiCatiOn
7. Garcia‑Alloza, M. et al. Existing plaques and
neuritic abnormalities in APP:PS1 mice are not
affected by administration of the gamma‑
secretase inhibitor LY‑411575. Mol.
Neurodegener. 4, 19 (2009).
Das, P ., Murphy, M. P ., Younkin, L. H., Younkin, S. G.
& Golde, T. E. Reduced effectiveness of Aβ1–42
immunization in APP transgenic mice with
significant amyloid deposition. Neurobiol. Aging
22, 721–727 (2001).
Levites, Y. et al. Anti‑Aβ42‑ and anti‑Aβ40‑specific
mAbs attenuate amyloid deposition in an
Alzheimer disease mouse model. J. Clin. Invest.
116, 193–201 (2006).
10. Siemers, E. R. How can we recognize “disease
modification” effects? J. Nutr. Health Aging 13,
11. Tibbling, G., Link, H. & Ohman, S. Principles of
albumin and IgG analyses in neurological
disorders. I. Establishment of reference
values. Scand. J. Clin. Lab. Invest. 37, 385–390
12. Blennow, K. et al. Blood–brain barrier
disturbance in patients with Alzheimer’s disease
is related to vascular factors. Acta Neurol.
Scand. 81, 323–326 (1990).
13. Andersson, M. et al. Cerebrospinal fluid in the
diagnosis of multiple sclerosis: a consensus
report. J. Neurol. Neurosurg. Psychiatry 57,
14. [No authors listed] 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 19, 109–116 (1998).
15. Strozyk, D., Blennow, K., White, L. R. &
Launer, L. J. CSF Aβ 42 levels correlate with
amyloid‑neuropathology in a population‑based
autopsy study. Neurology 60, 652–656 (2003).
16. Tapiola, T. et al. Cerebrospinal fluid β‑amyloid 42
and tau proteins as biomarkers of Alzheimer‑
type pathologic changes in the brain. Arch.
Neurol. 66, 382–389 (2009).
17. Fagan, A. M. et al. Inverse relation between
in vivo amyloid imaging load and cerebrospinal
fluid Aβ42 in humans. Ann. Neurol. 59, 512–519
18. Forsberg, A. et al. PET imaging of amyloid
deposition in patients with mild cognitive
impairment. Neurobiol. Aging 29, 1456–1465
19. Hesse, C. et al. Transient increase in total tau
but not phospho‑tau in human cerebrospinal
fluid after acute stroke. Neurosci. Lett. 297,
20. Ost, M. et al. Initial CSF total tau correlates with
1‑year outcome in patients with traumatic brain
injury. Neurology 67, 1600–1604 (2006).
21. Zetterberg, H. et al. Neurochemical aftermath of
amateur boxing. Arch. Neurol. 63, 1277–1280
22. Blom, E. S. et al. Rapid progression from mild
cognitive impairment to Alzheimer’s disease in
subjects with elevated levels of tau in
cerebrospinal fluid and the APOE ε4/ε4
genotype. Dement. Geriatr. Cogn. Disord. 27,
23. Samgard, K. et al. Cerebrospinal fluid total tau
as a marker of Alzheimer’s disease intensity. Int.
J. Geriatr. Psychiatry doi:10.1002/gps.2353.
24. Wallin, A. K., Hansson, O., Blennow, K.,
Londos, E. & Minthon, L. Can CSF biomarkers or
pre‑treatment progression rate predict response
to cholinesterase inhibitor treatment in
Alzheimer’s disease? Int. J. Geriatr. Psychiatry 24,
25. Otto, M. et al. Elevated levels of tau‑protein in
cerebrospinal fluid of patients with Creutzfeldt–
Jakob disease. Neurosci. Lett. 225, 210–212
26. Buerger, K. et al. CSF phosphorylated tau protein
correlates with neocortical neurofibrillary
pathology in Alzheimer’s disease. Brain 129,
27. Hampel, H. et al. Correlation of cerebrospinal
fluid levels of tau protein phosphorylated at
threonine 231 with rates of hippocampal atrophy
in Alzheimer disease. Arch. Neurol. 62, 770–773
28. Blennow, K. et al. Tau protein in cerebrospinal
fluid: a biochemical marker for axonal
degeneration in Alzheimer disease? Mol. Chem.
Neuropathol. 26, 231–245 (1995).
29. Sjögren, M. et al. Both total and phosphorylated
tau are increased in Alzheimer’s disease.
J. Neurol. Neurosurg. Psychiatry 70, 624–630
30. Riemenschneider, M. et al. Phospho‑tau/total
tau ratio in cerebrospinal fluid discriminates
Creutzfeldt–Jakob disease from other
dementias. Mol. Psychiatry 8, 343–347 (2003).
31. Hampel, H. et al. Measurement of
phosphorylated tau epitopes in the differential
diagnosis of Alzheimer disease: a comparative
cerebrospinal fluid study. Arch. Gen. Psychiatry
61, 95–102 (2004).
32. Kapaki, E. N. et al. Cerebrospinal fluid tau,
phospho‑tau181 and β‑amyloid1–42 in idiopathic
normal pressure hydrocephalus: a discrimination
from Alzheimer’s disease. Eur. J. Neurol. 14,
33. Koopman, K. et al. Improved discrimination of
autopsy‑confirmed Alzheimer’s disease (AD)
from non‑AD dementias using CSF P‑tau181P.
Neurochem. Int. 55, 214–218 (2009).
34. Seubert, P . et al. Isolation and quantification of
soluble Alzheimer’s β‑peptide from biological
fluids. Nature 359, 325–327 (1992).
35. Jarrett, J. T., Berger, E. P . & Lansbury, P . T. 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 32, 4693–4697 (1993).
36. Blennow, K. Cerebrospinal fluid protein
biomarkers for Alzheimer’s disease. NeuroRx 1,
37. Kohnken, R. et al. Detection of tau
phosphorylated at threonine 231 in
cerebrospinal fluid of Alzheimer’s disease
patients. Neurosci. Lett. 287, 187–190 (2000).
38. Vanmechelen, E. et al. Quantification of tau
phosphorylated at threonine 181 in human
cerebrospinal fluid: a sandwich ELISA with a
synthetic phosphopeptide for standardization.
Neurosci. Lett. 285, 49–52 (2000).
39. Hansson, O. et al. Association between CSF
biomarkers and incipient Alzheimer’s disease in
patients with mild cognitive impairment:
a follow‑up study. Lancet Neurol. 5, 228–234
40. Maddalena, A. et al. Biochemical diagnosis of
Alzheimer disease by measuring the
cerebrospinal fluid ratio of phosphorylated tau
protein to β‑amyloid peptide42. Arch. Neurol. 60,
41. Riemenschneider, M. et al. Cerebrospinal fluid
tau and β‑amyloid 42 proteins identify Alzheimer
disease in subjects with mild cognitive
impairment. Arch. Neurol. 59, 1729–1734
42. Zetterberg, H., Wahlund, L. O. & Blennow, K.
Cerebrospinal fluid markers for prediction of
Alzheimer’s disease. Neurosci. Lett. 352, 67–69
43. Olsson, A. et al. Simultaneous measurement of
β‑amyloid(1–42), total tau, and phosphorylated tau
(Thr181) in cerebrospinal fluid by the xMAP
technology. Clin. Chem. 51, 336–345 (2005).
44. Mattsson, N. et al. CSF biomarkers and incipient
Alzheimer disease in patients with mild cognitive
impairment. JAMA 302, 385–393 (2009).
45. Shaw, L. M. et al. Cerebrospinal fluid biomarker
signature in Alzheimer’s disease neuroimaging
initiative subjects. Ann. Neurol. 65, 403–413
46. Blennow, K. & Hampel, H. CSF markers for
incipient Alzheimer’s disease. Lancet Neurol. 2,
47. Blennow, K. CSF biomarkers for Alzheimer’s
disease: use in early diagnosis and evaluation of
drug treatment. Expert Rev. Mol. Diagn. 5,
48. Engelborghs, S. et al. Diagnostic performance of
a CSF‑biomarker panel in autopsy‑confirmed
dementia. Neurobiol. Aging 29, 1143–1159
49. Forman, M. S. et al. Frontotemporal dementia:
clinicopathological correlations. Ann. Neurol. 59,
50. Snowdon, D. A. Aging and Alzheimer’s disease:
lessons from the Nun Study. Gerontologist 37,
51. Price, J. L. & Morris, J. C. Tangles and plaques in
nondemented aging and “preclinical”
Alzheimer’s disease. Ann. Neurol. 45, 358–368
52. Kotzbauer, P . T., Trojanowski, J. Q. & Lee, V. M.
Lewy body pathology in Alzheimer’s disease.
J. Mol. Neurosci. 17, 225–232 (2001).
53. Jellinger, K. A. Diagnostic accuracy of
Alzheimer’s disease: a clinicopathological study.
Acta Neuropathol. (Berl.) 91, 219–220 (1996).
54. Schneider, J. A., Arvanitakis, Z., Leurgans, S. E.
& Bennett, D. A. The neuropathology of probable
alzheimer’s disease and mild cognitive
impairment. Ann. Neurol. 66, 200–208 (2009).
55. Visser, P . J. et al. 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. 8, 619–627 (2009).
56. Skoog, I. et al. Cerebrospinal fluid β‑amyloid 42
is reduced before the onset of sporadic
dementia: a population‑based study in
85‑year‑olds. Dement. Geriatr. Cogn. Disord. 15,
57. Gustafson, D. R., Skoog, I., Rosengren, L.,
Zetterberg, H. & Blennow, K. Cerebrospinal fluid
β‑amyloid 1–42 concentration may predict
cognitive decline in older women. J. Neurol.
Neurosurg. Psychiatry 78, 461–464 (2007).
58. Stomrud, E., Hansson, O., Blennow, K.,
Minthon, L. & Londos, E. Cerebrospinal fluid
biomarkers predict decline in subjective
cognitive function over 3 years in healthy elderly.
Dement. Geriatr. Cogn. Disord. 24, 118–124
59. Moonis, M. et al. Familial Alzheimer disease:
decreases in CSF Aβ42 levels precede cognitive
decline. Neurology 65, 323–325 (2005).
60. Ringman, J. M. et al. Biochemical markers in
persons with preclinical familial Alzheimer
disease. Neurology 71, 85–92 (2008).
61. Götz, J., Chen, F., van Dorpe, J. & Nitsch, R. M.
Formation of neurofibrillary tangles in P301L tau
transgenic mice induced by Aβ42 fibrils. Science
293, 1491–1495 (2001).
62. Lewis, J. et al. Enhanced neurofibrillary
degeneration in transgenic mice expressing
mutant tau and APP . Science 293, 1487–1491
© 20 Macmillan Publishers Limited. All rights reserved10
nature reviews | neurology
aDvanCe OnLine PuBLiCatiOn | 13
63. Fagan, A. M. et al. Cerebrospinal fluid tau and
ptau181 increase with cortical amyloid deposition
in cognitively normal individuals: Implications for
future clinical trials of Alzheimer’s disease.
EMBO Mol. Med. 1, 371–380 (2009).
64. Cairns, N. et al. Absence of Pittsburgh
compound B detection of cerebral amyloid β in a
patient with clinical, cognitive, and cerebrospinal
fluid markers of Alzheimer disease: a case
report. Arch. Neurol. 66, 1557–1562 (2009).
65. Clark, C. M. et al. Cerebrospinal fluid tau and
β‑amyloid: how well do these biomarkers reflect
autopsy‑confirmed dementia diagnoses? Arch.
Neurol. 60, 1696–1702 (2003).
66. Bian, H. et al. CSF biomarkers in frontotemporal
lobar degeneration with known pathology.
Neurology 70, 1827–1835 (2008).
67. Sunderland, T. et al. Decreased β‑amyloid1–42 and
increased tau levels in cerebrospinal fluid of
patients with Alzheimer disease. JAMA 289,
68. Fukumoto, H., Cheung, B. S., Hyman, B. T. &
Irizarry, M. C. β‑Secretase protein and activity
are increased in the neocortex in Alzheimer
disease. Arch. Neurol. 59, 1381–1389 (2002).
69. Yang, L. B. et al. Elevated β‑secretase
expression and enzymatic activity detected in
sporadic Alzheimer disease. Nat. Med. 9, 3–4
70. Holsinger, R. M., McLean, C. A., Collins, S. J.,
Masters, C. L. & Evin, G. Increased β‑secretase
activity in cerebrospinal fluid of Alzheimer’s
disease subjects. Ann. Neurol. 55, 898–899
71. Zetterberg, H. et al. Elevated cerebrospinal fluid
BACE1 activity in incipient Alzheimer disease.
Arch. Neurol. 65, 1102–1107 (2008).
72. Zhong, Z. et al. Levels of β‑secretase (BACE1) in
cerebrospinal fluid as a predictor of risk in mild
cognitive impairment. Arch. Gen. Psychiatry 64,
73. Olsson, A. et al. Measurement of α‑ and
β‑secretase cleaved amyloid precursor protein in
cerebrospinal fluid from Alzheimer patients. Exp.
Neurol. 183, 74–80 (2003).
74. Lewczuk, P . et al. Soluble amyloid precursor
proteins in the cerebrospinal fluid as novel
potential biomarkers of Alzheimer’s disease:
a multicenter study. Mol. Psychiatry
75. Portelius, E., Westman‑Brinkmalm, A.,
Zetterberg, H. & Blennow, K. Determination of
β‑amyloid peptide signatures in cerebrospinal fluid
using immunoprecipitation–mass spectrometry.
J. Proteome Res. 5, 1010–1016 (2006).
76. Mehta, P . D. et al. Plasma and cerebrospinal fluid
levels of amyloid β proteins 1–40 and 1–42 in
Alzheimer disease. Arch. Neurol. 57, 100–105
77. Hansson, O. et al. Prediction of Alzheimer’s
disease using the CSF Aβ42/Aβ40 ratio in
patients with mild cognitive impairment. Dement.
Geriatr. Cogn. Disord. 23, 316–320 (2007).
78. Lewczuk, P . et al. The amyloid‑β (Aβ) peptide
pattern in cerebrospinal fluid in Alzheimer’s
disease: evidence of a novel carboxyterminally
elongated Aβ peptide. Rapid Commun. Mass
Spectrom. 17, 1291–1296 (2003).
79. Schoonenboom, N. S. et al. Amyloid β 38, 40,
and 42 species in cerebrospinal fluid: more of
the same? Ann. Neurol. 58, 139–142 (2005).
80. Portelius, E. et al. An Alzheimer’s disease‑
specific β‑amyloid fragment signature in
cerebrospinal fluid. Neurosci. Lett. 409,
81. Portelius, E. et al. A novel pathway for amyloid
precursor protein processing. Neurobiol. Aging
82. Walsh, D. M. & Selkoe, D. J. Aβ oligomers—a
decade of discovery. J. Neurochem. 101,
83. Georganopoulou, D. G. et al. Nanoparticle‑based
detection in cerebral spinal fluid of a soluble
pathogenic biomarker for Alzheimer’s disease.
Proc. Natl Acad. Sci. USA 102, 2273–2276 (2005).
84. Santos, A. N. et al. Detection of amyloid‑β
oligomers in human cerebrospinal fluid by flow
cytometry and fluorescence resonance energy
transfer. J. Alzheimers Dis. 11, 117–125 (2007).
85. Klyubin, I. et al. Amyloid β protein dimer‑
containing human CSF disrupts synaptic
plasticity: prevention by systemic passive
immunization. J. Neurosci. 28, 4231–4237
86. Mruthinti, S. et al. Autoimmunity in Alzheimer’s
disease: increased levels of circulating IgGs
binding Aβ and RAGE peptides. Neurobiol. Aging
25, 1023–1032 (2004).
87. Nath, A. et al. Autoantibodies to amyloid
β‑peptide (Aβ) are increased in Alzheimer’s
disease patients and Aβ antibodies can enhance
Aβ neurotoxicity: implications for disease
pathogenesis and vaccine development.
Neuromolecular Med. 3, 29–39 (2003).
88. Brettschneider, S. et al. Decreased serum
amyloid β1–42 autoantibody levels in Alzheimer’s
disease, determined by a newly developed
immuno‑precipitation assay with radiolabeled
amyloid β1–42 peptide. Biol. Psychiatry 57,
89. Du, Y. et al. Reduced levels of amyloid β‑peptide
antibody in Alzheimer disease. Neurology 57,
90. Hyman, B. T. et al. Autoantibodies to amyloid‑β
and Alzheimer’s disease. Ann. Neurol. 49,
91. Britschgi, M. et al. Neuroprotective natural
antibodies to assemblies of amyloidogenic
peptides decrease with normal aging and
advancing Alzheimer’s disease. Proc. Natl Acad.
Sci. USA 106, 12145–12150 (2009).
92. Laterza, O. F. et al. Identification of novel brain
biomarkers. Clin. Chem. 52, 1713–1721 (2006).
93. Lee, J. M. et al. The brain injury biomarker
VLP‑1 is increased in the cerebrospinal fluid of
Alzheimer disease patients. Clin. Chem. 54,
94. Friede, R. L. & Samorajski, T. Axon caliber related
to neurofilaments and microtubules in sciatic
nerve fibers of rats and mice. Anat. Rec. 167,
95. Sjögren, M. et al. Neurofilament protein in
cerebrospinal fluid: a marker of white matter
changes. J. Neurosci. Res. 66, 510–516 (2001).
96. Agren‑Wilsson, A. et al. CSF biomarkers in the
evaluation of idiopathic normal pressure
hydrocephalus. Acta Neurol. Scand. 116,
97. Sjögren, M. et al. Cytoskeleton proteins in CSF
distinguish frontotemporal dementia from AD.
Neurology 54, 1960–1964 (2000).
98. Davidsson, P ., Puchades, M. & Blennow, K.
Identification of synaptic vesicle, pre‑ and
postsynaptic proteins in human cerebrospinal
fluid using liquid‑phase isoelectric focusing.
Electrophoresis 20, 431–437 (1999).
99. Sjögren, M. et al. The cerebrospinal fluid levels
of tau, growth‑associated protein‑43 and soluble
amyloid precursor protein correlate in
Alzheimer’s disease, reflecting a common
pathophysiological process. Dement. Geriatr.
Cogn. Disord. 12, 257–264 (2001).
100. Montine, T. J., Quinn, J., Kaye, J. & Morrow, J. D.
F2‑isoprostanes as biomarkers of late‑onset
Alzheimer’s disease. J. Mol. Neurosci. 33,
101. Brys, M. et al. Prediction and longitudinal study
of CSF biomarkers in mild cognitive impairment.
Neurobiol. Aging 30, 682–690.
102. Knopman, D. S. et al. Practice parameter:
diagnosis of dementia (an evidence‑based
review). Report of the Quality Standards
Subcommittee of the American Academy of
Neurology. Neurology 56, 1143–1153 (2001).
103. Visser, P . J., Scheltens, P . & Verhey, F. R. Do MCI
criteria in drug trials accurately identify subjects
with predementia Alzheimer’s disease?
J. Neurol. Neurosurg. Psychiatry 76, 1348–1354
104. Ganguli, M. et al. Detection and management of
cognitive impairment in primary care: The Steel
Valley Seniors Survey. J. Am. Geriatr. Soc. 52,
105. Raschetti, R., Albanese, E., Vanacore, N. &
Maggini, M. Cholinesterase inhibitors in mild
cognitive impairment: a systematic review of
randomised trials. PLoS Med. 4, e338 (2007).
106. Cummings, J. L., Doody, R. & Clark, C. Disease‑
modifying therapies for Alzheimer disease:
challenges to early intervention. Neurology 69,
107. Salloway, S. et al. A phase 2 trial of
bapineuzumab in mild to moderate Alzheimer’s
disease. Neurology doi:10.1212/
108. Orgogozo, J. M. et al. Subacute
meningoencephalitis in a subset of patients with
AD after Aβ42 immunization. Neurology 61,
109. Blennow, K. et al. Longitudinal stability of CSF
biomarkers in Alzheimer’s disease. Neurosci.
Lett. 419, 18–22 (2007).
110. Zetterberg, H. et al. Intra‑individual stability of
CSF biomarkers for Alzheimer’s disease over
two years. J. Alzheimers Dis. 12, 255–260
111. Vellas, B. Use of biomarkers in Alzheimer’s
trials. J. Nutr. Health Aging 13, 331 (2009).
112. Hampel, H. et al. Lithium trial in Alzheimer’s
disease: a randomized, single‑blind, placebo‑
controlled, multicenter 10‑week study. J. Clin.
Psychiatry 70, 922–931 (2009).
113. Anderson, J. J. et al. Reductions in β‑amyloid
concentrations in vivo by the γ‑secretase
inhibitors BMS‑289948 and BMS‑299897.
Biochem. Pharmacol. 69, 689–698 (2005).
114. Lanz, T. A., Hosley, J. D., Adams, W. J. &
Merchant, K. M. Studies of Aβ
pharmacodynamics in the brain, cerebrospinal
fluid, and plasma in young (plaque‑free) Tg2576
mice using the γ‑secretase inhibitor N2‑[(2S)‑2‑
5‑methyl‑6‑oxo ‑6, 7‑dihydro‑5H‑dibenzo[b, d]
J. Pharmacol. Exp. Ther. 309, 49–55 (2004).
115. Sankaranarayanan, S. et al. First demonstration
of cerebrospinal fluid and plasma Aβ lowering
with oral administration of a β‑site amyloid
precursor protein‑cleaving enzyme 1 inhibitor in
nonhuman primates. J. Pharmacol. Exp. Ther.
328, 131–140 (2009).
116. Lannfelt, L. et al. Safety, efficacy, and biomarker
findings of PBT2 in targeting Aβ as a modifying
therapy for Alzheimer’s disease: a phase IIa,
double‑blind, randomised, placebo‑controlled
trial. Lancet Neurol. 7, 779–786 (2008).
117. Kadir, A. et al. Effect of phenserine treatment on
brain functional activity and amyloid in
Alzheimer’s disease. Ann. Neurol. 63, 621–631
118. Gilman, S. et al. Clinical effects of Aβ
immunization (AN1792) in patients with AD in an
interrupted trial. Neurology 64, 1553–1562
© 20 Macmillan Publishers Limited. All rights reserved10
14 | aDvanCe OnLine PuBLiCatiOn Download full-text
119. Fleisher, A. S. et al. Phase 2 safety trial targeting
amyloid β production with a γ‑secretase inhibitor
in Alzheimer disease. Arch. Neurol. 65,
120. Bateman, R. J. et al. A γ‑secretase inhibitor
decreases amyloid‑β production in the central
nervous system. Ann. Neurol. 66, 48–54
121. Irizarry, M. C. Biomarkers of Alzheimer disease
in plasma. NeuroRx 1, 226–234 (2004).
122. Mayeux, R. et al. Plasma Aβ40 and Aβ42 and
Alzheimer’s disease: relation to age, mortality,
and risk. Neurology 61, 1185–1190 (2003).
123. Pomara, N., Willoughby, L. M., Sidtis, J. J. &
Mehta, P . D. Selective reductions in plasma Aβ
1–42 in healthy elderly subjects during
longitudinal follow‑up: a preliminary report. Am. J.
Geriatr. Psychiatry 13, 914–917 (2005).
124. van Oijen, M., Hofman, A., Soares, H. D.,
Koudstaal, P . J. & Breteler, M. M. Plasma Aβ1–40
and Aβ1–42 and the risk of dementia:
a prospective case‑cohort study. Lancet Neurol.
5, 655–660 (2006).
125. Graff‑Radford, N. R. et al. Association of low
plasma Aβ42/Aβ40 ratios with increased
imminent risk for mild cognitive impairment and
Alzheimer disease. Arch. Neurol. 64, 354–362
126. Kuo, Y. M. et al. High levels of circulating Aβ42
are sequestered by plasma proteins in
Alzheimer’s disease. Biochem. Biophys. Res.
Commun. 257, 787–791 (1999).
127. Ray, S. et al. Classification and prediction of
clinical Alzheimer’s diagnosis based on plasma
signaling proteins. Nat. Med. 13, 1359–1362
128. Hye, A. et al. Proteome‑based plasma
biomarkers for Alzheimer’s disease. Brain 129,
129. Buerger, K. et al. Blood‑based microcirculation
markers in Alzheimer’s disease‑diagnostic value
of midregional pro‑atrial natriuretic
peptide/C‑terminal endothelin‑1 precursor
fragment ratio. Biol. Psychiatry 65, 979–984
130. Brys, M. et al. Magnetic resonance imaging
improves cerebrospinal fluid biomarkers in the
early detection of Alzheimer’s disease.
J. Alzheimers Dis. 16, 351–362 (2009).
131. Zhang, Y. et al. Usefulness of computed
tomography linear measurements in diagnosing
Alzheimer’s disease. Acta Radiol. 49, 91–97
132. Schoonenboom, N. S. et al. CSF and MRI
markers independently contribute to the
diagnosis of Alzheimer’s disease. Neurobiol.
Aging 29, 669–675 (2008).
133. Vemuri, P . et al. MRI and CSF biomarkers in
normal, MCI, and AD subjects: diagnostic
discrimination and cognitive correlations.
Neurology 73, 287–293 (2009).
134. Okamura, N. et al. Combined analysis of CSF tau
levels and [123I]iodoamphetamine SPECT in mild
cognitive impairment: implications for a novel
predictor of Alzheimer’s disease. Am. J.
Psychiatry 159, 474–476 (2002).
135. Hansson, O. et al. Combined rCBF and CSF
biomarkers predict progression from mild
cognitive impairment to Alzheimer’s disease.
Neurobiol. Aging 30, 165–173 (2007).
136. Nyberg, S. et al. Detection of amyloid in
Alzheimer’s disease with positron emission
tomography using [11C]AZD2184. Eur. J. Nucl.
Med. Mol. Imaging 36, 1859–1863 (2009).
137. McKhann, G. et al. 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 34,
138. American Psychiatric Association. Diagnostic
and Statistical Manual of Mental Disorders:
DSM‑IV‑TR (American Psychiatric Association,
Washington DC, 2000).
139. WHO. ICD‑10: International Statistical
Classification of Diseases and Related Health
Problems 10th Revision (WHO, Geneva, 1992).
140. Dubois, B. et al. Research criteria for the
diagnosis of Alzheimer’s disease: revising the
NINCDS–ADRDA criteria. Lancet Neurol. 6,
141. Frisoni, G. B. et al. Markers of Alzheimer’s
disease in a population attending a memory
clinic. Alzheimers Dement. 5, 307–317 (2009).
142. Vanderstichele, H. et al. Standardization of
measurement of β‑amyloid1–42 in cerebrospinal
fluid and plasma. Amyloid 7, 245–258 (2000).
143. Vanderstichele, H. et al. Analytical performance
and clinical utility of the INNOTEST PHOSPHO‑
TAU181P assay for discrimination between
Alzheimer’s disease and dementia with Lewy
bodies. Clin. Chem. Lab. Med. 44, 1472–1480
144. Verwey, N. A. et al. A worldwide multicentre
comparison of assays for cerebrospinal fluid
biomarkers in Alzheimer’s disease. Ann. Clin.
Biochem. 46, 235–240 (2009).
145. Beher, D., Wrigley, J. D., Owens, A. P . &
Shearman, M. S. Generation of C‑terminally
truncated amyloid‑β peptides is dependent on
γ‑secretase activity. J. Neurochem. 82, 563–575
146. Vassar, R. et al. β‑Secretase cleavage of
Alzheimer’s amyloid precursor protein by the
transmembrane aspartic protease BACE.
Science 286, 735–741 (1999).
147. Selkoe, D. J. & Wolfe, M. S. Presenilin: running
with scissors in the membrane. Cell 131,
148. Lammich, S. et al. Constitutive and regulated
α‑secretase cleavage of Alzheimer’s amyloid
precursor protein by a disintegrin
metalloprotease. Proc. Natl Acad. Sci. USA 96,
149. Haass, C. et al. Amyloid β‑peptide is produced by
cultured cells during normal metabolism. Nature
359, 322–325 (1992).
150. Bibl, M. et al. Cerebrospinal fluid amyloid β
peptide patterns in Alzheimer’s disease
patients and nondemented controls depend on
sample pretreatment: indication of carrier‑
mediated epitope masking of amyloid beta
peptides. Electrophoresis 25, 2912–2918
151. Trojanowski, J. Q., Schuck, T., Schmidt, M. L. &
Lee, V. M. Distribution of tau proteins in the
normal human central and peripheral nervous
system. J. Histochem. Cytochem. 37, 209–215
152. Goedert, M. Tau protein and the neurofibrillary
pathology of Alzheimer’s disease. Trends
Neurosci. 16, 460–465 (1993).
153. Iqbal, K. et al. Tau pathology in Alzheimer
disease and other tauopathies. Biochim.
Biophys. Acta 1739, 198–210 (2005).
154. McLean, C. A. et al. Soluble pool of Aβ amyloid
as a determinant of severity of
neurodegeneration in Alzheimer’s disease. Ann.
Neurol. 46, 860–866 (1999).
155. Blennow, K., Rybo, E., Wallin, A., Gottfries, C. G.
& Svennerholm, L. Cerebrospinal fluid cytology
in Alzheimer’s disease. Dementia 2, 25–29
156. Blennow, K. et al. Intrathecal synthesis of
immunoglobulins in patients with Alzheimer’s
disease. Eur. Neuropsychopharmacol. 1, 79–81
157. Mollenhauer, B. et al. Tau protein, Aβ42 and
S‑100B protein in cerebrospinal fluid of patients
with dementia with Lewy bodies. Dement. Geriatr.
Cogn. Disord. 19, 164–170 (2005).
158. Grimes, D. A. & Schulz, K. F. Uses and abuses of
screening tests. Lancet 359, 881–884 (2002).
159. Birks, J. Cholinesterase inhibitors for Alzheimer’s
disease. Cochrane Database of Systematic
Reviews, Issue 1 Art. No.:CD005593.
Supplementary information is linked to the online
version of the paper at www.nature.com/nrneurol
© 20 Macmillan Publishers Limited. All rights reserved 10