close

Вход

Забыли?

вход по аккаунту

?

Cerebrospinal fluid levels of amyloid -protein in alzheimer's disease Inverse correlation with severity of dementia and effect of apolipoprotein e genotype.

код для вставкиСкачать
Cerebrospinal Fluid Levels of Amyloid
P-Protein in Alzheimer's Disease: Inverse
Correlation with Severity of Dementia and
Effect of Apolipoprotein E Genotype
Roger M. Nitsch, MD, PhD,"I G. William Rebeck, PhD," Meihua Deng, MS,$ U. Ingrid kchardson, PhD,$
Marsha Tennis, RN," Dale B. Schenk, PhD,? Carmen Vigo-Pelfrey, PhD,t Ivan Lieberburg, MD, PhD,f
Richard J. Wurtman, MD,$ Bradley T. Hyman, MD, PhD," and John H. Growdon, MD"
Alzheimer's disease (AD) is characterized by formation in brain of neurofibrillary tangles and of amyloid deposits.
The major protein component of the former is T, while the latter are composed of amyloid P-peptides (AP), which
are derived by proteolytic cleavage of the amyloid @protein precursor (APP). Both T and various secretory APP
derivatives including AP and APPS are present in human cerebrospinal fluid (CSF). To investigate whether clinical
signs of AD are paralleled by changes in CSF levels of these proteins, we correlated quantitative measures of dementia
severity with CSF concentrations of AP, of APPS, and of T. We found that levels of AP in CSF of AD patients were
inversely correlated both to cognitive and to functional measures of dementia severity. In contrast, levels of APPS
and of T did not correlate with dementia severity. Apolipoprotein E (apoE) genotype did not influence CSF levels of
AP, APPS, or T , which were similar among AD patients with Apo E ~ 3 1 3 ~, 3 1 4and
, e414 alleles. These data indicate
that CSF levels of AP decrease with advancing severity of dementia in AD and suggest that they are independent of
a patient's Apo E genotype.
Nitsch RM, Rebeck GW, Deng M, kchardson UI, Tennis M, Schenk DB, Vigo-Pelfrey C, Lieberburg I,
Wurtman RJ, Hyman BT, Growdon JH. Cerebrospinal fluid levels of amyloid P-protein
in Alzheimer's disease: inverse correlation with severity of dementia and effect
of apolipoprotein E genotype. Ann Neurol 1995;37:512-5 18
Brain histopathology of Alzheimer's disease (AD) is
characterized by two major hallmarks, amyloid plaques
and neurofibrillary tangles. Soluble forms of their principal proteinaceous components, amyloid P-protein
(AP) and T , are present in human cerebrospinal fluid
(CSF), and thus are accessible for biochemical analyses
during life 11-41. Brain amyloid deposits consist of
accumulations of both aggregated and nonaggregated
forms of AP, a group of approximately 4-kd peptides
that are 39 to 43 amino acid residues in length 12, 3,
51, and that are derived by proteolytic cleavage of the
larger amyloid P-protein precursor (APP) (for review,
see 161). APP is an abundant brain protein with neurotrophic activities, is also present in many peripheral
tissues, and exists in various forms that are derived
from alternative splicing of RNA encoded for by a
single gene that maps to chromosome 21. Besides AP,
large extracellular N-terminal APP derivatives (APPS)
are generated by secretory posttranslational processing
C7, S}, and are present in human CSF at remarkably
high concentrations {9, 101. Despite substantial overlap between AD patients and age-matched control subjects, CSF levels of both AP and APPS have been proposed as biological, and possibly diagnostic, markers
for AD El, 11-13]. Neurofibrillary tangles in AD
brain are composed of abnormally phosphorylated T , a
microtubule-associated protein that normally is involved in the formation and stabilization of neuronal
microtubules. CSF levels of 7 have also been proposed
as biological markers of A D 141. To establish whether
these potential disease markers parallel the severity of
the dementia in AD, we correlated their levels in CSF
with cognitive and functional measures of dementia
obtained on the day of lumbar puncture.
Both AP and T can bind to apolipoprotein E (Apo
E) {14, 151, a lipid-binding protein that exists in several
From the "Department of Neurology, Massachusetts General Hospital, Boston, MA, t Athena Neurosciences Inc, South San Francisco,
CA, and $Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, Cambridge, MA.
Received May 23, 1994, and in revised form Sep 6. Accepted for
publication Nov 22, 1994.
Address correspondence to Dr Nitsch, who is presently at the Center for Molecular Neurobiology, University of Hamburg, UKE22,
Martini Str. 52, 20246 Hamburg, Germany.
512 Copyright 0 1995 by the American Neurological Association
Table 1 . Clinical Characteristics of Alzheimer’s Disease Patients Grouped According to Apolipoprotein E Genotype
Apolipoprotein E Genotype
n
Sex
(malelfemale)
Age (yr)
Alzheimer’s disease patients
&3/3
E3/4
~414
Total A D
5
9
5
19
213
514
213
9/10
62.4
71.1
69.4
68.4
?
?
Control subjects
10
614
71.2
?
?
?
Duration
(yr)
Severity
(BDS)
Severity
(ADL) (96)
3.6
?
2
6.6
6.9
3.8
3.4
3.7
16.4 r 13.4
19.7 ? 11.0
12.8 ? 9.0
17.0 ? 10.9
38.4
47.9
27.0
8.2
-
-
-
6.6
5.8
IT
2
1.9
2.1
1.1
1.8
31.8
27.9
? 16.8
38.8 +- 25.3
?
2
The apolipoprotein E ~ 3 / group
3
is significantly younger than the other groups; all other characteristics are similar among genotype groups.
Age of control subjects did not differ significantly from the age of Alzheimer’s disease patients. Values are means k SD.
n = number of patients; BDS = Blessed Dementia Scale; ADL = Activities of Daily Living score; AD = Alzheimer’s disease.
isoforms derived from three major allelic variants of
the apoE gene, ~ 2 6, 3 , and ~ 4 Inheritance
.
of ~4
alleles is a strong risk factor for developing A D at an
earlier age of onset and is also associated with higher
amyloid plaque density in A D brain C14-201. We
therefore investigated whether individual Apo E genotypes are associated with differences in CSF levels of
AP, of secreted N-terminal APP derivatives and of T
in A D patients.
Subjects and Methods
The clinical diagnosis of A D was made according to the National Institute of Neurological and Communicative Disorders and Stroke-Alzheimer’s Disease and Related Disorders
Association (NINCDS-ADRDA) criteria [2 1). Our A D
study population included 9 men and 10 women with a mean
age of 68.4 ? 6.9 years (range, 55-80 yr) and a mean duration of illness of 3.7 years (range, 1-6 yr) (Table 1). The
control subjects were 6 men and 4 women with a mean age
of 71.2 ? 8.2 years (range, 51-77 yr). They had no signs
of dementia as evidenced by the clinical impression and by
Mini-Mental State examination (29.7 ? 0.7). None of the
patients were taking experimental drugs designed to improve
cognition. Three patients were taking low doses of psychoactive drugs (haloperidol: n = 1, Apo E ~ 3 1 4benzodiazepines:
;
n = 2, Apo E ~ 3 1 and
3 314) but were included in the study
protocol because the measured variables were within one
standard deviation of the respective means. Severity of dementia in the A D patients was estimated according to the
information, memory, and concentration subscale of the
Blessed Dementia Scale (BDS) score [22] in which 0 to 3 is
normal, and 37 indicates maximal dementia severity. We also
estimated the functional consequences of dementia by an
Activities of Daily Living (ADL) score 1231, in which 0%
indicates no impairment and 100% represents maximal dependence. Individual patients were grouped according to
Apo E genotype (e3/3, n = 6; ~ 3 1 4 n, = 9; ~ 4 1 4 n, = 4),
which was determined from whole blood samples by restriction isotyping using a polymerase chain reaction-based assay
C171.
O n the day of dementia testing, CSF was obtained by lumbar puncture, was frozen immediately (in 50O-pl aliquots on
dry ice) at the bedside, and was maintained below -30°C
until biochemical analyses were performed. AP levels were
measured by a sandwich enzyme-linked immunosorbent
assay (ELISA) as described previously [l).
APPS ELISAs were done by using a modification of the
protocol described by Van Nostrand and colleagues [ 1 1). In
brief, serial dilutions of CSF samples (0.33 pl, 1 p1, 3 pl)
were brought to a final volume of 100 p1 in phosphatebuffered saline (PBS) and coated into Nunc Maxisorb 96well immunoplates at 4°C overnight. Nonspecific binding
sites were blocked with 1% ovalbumin in PBS for 20 minutes
at room temperature with constant shaking. The monoclonal
antibody 22C11 was used as primary antibody at 1 pg/ml in
PBSllgO ovalbumin (2 hr at room temperature with constant
shaking). Plates were washed four times and incubated with
horseradish peroxidase-conjugated anti-mouse IgG (from
sheep, Amersham) at a dilution of 1 :400 on a rotating shaker
for 2 hours at room temperature. Plates were washed five
times with PBSl0.0596 Tween 20 and incubated with 10 mg/
ml 0-phenylenediamine (OPD) in 5 0 mM phosphate-citrate
buffer (pH 5 ) for 5 minutes at room temperature. The color
reaction was stopped by addition of 100 p14 M H2S04.The
reaction product was read at 490 nm by using a Titertek
multiplate reader. Serial dilutions of recombinant full-length
human APP75 1 (0-5 ng; APP,,-,,, purified from a baculovirus expression system) were used as a calibrator on each
plate. The sensitivity of this ELISA for recombinant human
APP,,-,,, was 50 pg/well and the linear range was 200 pg
to 5 ng ( r 2 = 0.94). The interassay coefficient of variability
was 6% at 500 pg and 7% at 5 ng. The intraassay variability
was less than 5%.
CSF levels of T were determined by a sandwich ELISA
employing two monoclonal antibodies directed against 7.
These were prepared according to the method of Kohler and
Milstein [24). T was purified from SF9 cells infected with the
7-containing baculovirus construct, and 6-week-old A/J mice
were repeatedly injected with 100 pg of purified T at 2-week
intervals. Antibody 16G7 was obtained from one fusion
while antibody l6B5 was obtained from a subsequent fusion.
Wells containing hybridoma cells were screened for their
ability to precipitate I-labeled T prepared by using glucose
oxidase and lactoperoxidase according to the manufacturer’s
protocol (Bio-Rad). The isotypes of both 16G7 and 16B5
were determined to be gamma 1 kappa. Antibody 16B5 was
biotinylated by using the N-hydroxysuccinimide ester of bio-
Nitsch et al: CSF AP in Alzheimer’s Disease
513
tin following the manufacturer’s instruction (Pierce). The antibody 16G7 was suspended at 5 pglml in Tris-buffered saline (TBS) and coated overnight into microtiter plates
(Dynatec Microplate 2; 100 ~llwell).Plates were blocked
with 0.25% casein (wt/vol) in PBS and samples of 50 pl of
human CSF or calibrators (50 pl of 3-1,000 pg/ml in PBS/
casein/0.05% Tween 20) were incubated overnight at room
temperature with constant shaking. After incubation with the
biotinylated antibody 16B5, plates were washed in TweenTBS (TTBS) and incubated for 1 hour with streptavidin
alkaline phosphatase (1 : 1,000 in PBSlcaseinlO.O5%
Tween 20; Boehringer Mannheim). The chemiluminescent
reagent disodium 3-[4-methoxyspiro( 1,2-dioxetane-3,2tricyclo[3.3.1.l)decan)-4-yl}phenyl phosphate (AMPPD,
Tropix) and enhancer Emerald Green (Tropix) were diluted
1: 1,000 and 1: 100, respectively in 1 M diethanolamine
buffer containing 1 mM MgCI,, 0.02% NaN,, p H 10, and
were added to each well. Plates were read with a chemilumimeter (Dynatech ML 1000). The samples measured were
compared with T purified from human brain as a calibrator
[4]. Use of other T standards can result in slightly different
apparent concentrations, without affecting the relative differences between samples. Samples were analyzed in duplicate
on three separate occasions and data presented here are
means of these analyses. The sensitivity of this assay was 50
pglml, the assay was linear (r2 = 0.95) up to 2 ng/ml, and
the coefficient of variance was less than 8% in all cases.
For western blotting assays, 5 to 10 pI of freshly thawed
CSF was diluted 1: 1 in 2 x Laemmli gel loading buffer,
boiled immediately for 3 minutes, and separated by sodium
dodecyl sulfate (SDS)-polyacrylamide (12%) gel electrophoresis (SDS-PAGE) followed by immunoblotting [7, 251, and
densitometric analysis of chemiluminescent reaction products
on preflashed (linear) x-ray films by using a Pharmacia-LKB
laser scanner. Linearity of the western blots in the range used
was assured by serial dilution curves of human CSF using
identical conditions as in the test runs. Measurements were
done on three or four independent gels using fresh 10-1-1.1
aliquots for each run. Optical densities were normalized to a
human CSF calibrator that was run in parallel on each gel.
The sensitivity of this assay system for human APP was 200
pg, which corresponds to 0.1 to 0.3 pl of human CSF. The
interassay coefficient of variance over the linear range of 0.5
to 4.0 arbitrary optical density units was 11.5%. The intraassay variation was less than 8.2%. Means of triplicate or quadruplicate independent measurements were used for statistical analyses. The monoclonal antibody 22C11 171, which was
used as primary antibody can cross-react weakly with an APPlike protein (APLPZ, but not APLP1) in some experimental
conditions 1261. To determine specific APP immunoreactivity, we used the monoclonal antibodies Anti-Ah90 (1.D5,
Boehringer Mannheim) [73, 10D5 [27), 7H5 {27}, and the
polyclonal antiserum R1736 IS}. To assure that immunoreactive material in human CSF was C-terminally truncated, that
is, APPS, the C-terminal antibody anti-C6 (Athena Neurosciences Inc), and the antiserum aC8 [5] were used.
Correlation analyses were done by multiple regression using CSF levels of the above proteins, dementia severity
scores, duration of illness, and age of onset. Regression analysis was followed by analysis of variance. Statistical significance
was assumed at p < 0.05. To determine possible confounding
514 Annals of Neurology
Vol 37 No
4 April 1995
effects of gender, apoE genotype, or age of disease onset,
multivariate analysis was performed by using these variables
as covariates. For comparisons of apoE genotype groups,
data were analyzed by single-factor analysis of variance using
genotype as the independent variable, as well as by analysis
of covariance using either BDS or ADL scores, or age of
disease onset as the cofactors. T o compare directly the effects
of e3 and ~4 on the dependent variables, data obtained in
the groups with apoE ~ 3 1 3and ~414genotypes were also
compared by t test, and 95% confidence bounds were determined. Effects of sex were analyzed by grouping data according to gender and by comparison using a t test.
Results
CSF concentrations of AP ranged from 7.9 to 20.5 ngl
ml, N-terminal APP derivatives measured by ELISA
ranged from 0.95 to 2.4 Fg/ml, and levels of T ranged
from 129 to 462 pg/ml (Table 2). CSF levels of AP
were significantly correlated with those of T ( Y =
0.665, < 0.002), as well as with levels of both the
106-kd APP derivative ( Y = 0.523, p < 0.05) and the
25-kd fragment ( Y = 0.615, p < 0.005). Levels of
AP in CSF did not differ between A D patients and
age-matched, cognitively normal control subjects
(Fig 1).
Lmear regression analysis correlating severity of dementia with CSF AP levels revealed significant inverse
relationships between the BDS score and AP levels
( Y = -0.666, p < 0.002; Fig 2), and between the
ADL score and AP levels ( Y = -0.619, p < 0.005;
see Fig 2). Multivariate analyses revealed that these
correlations were also statistically significant when age
of onset, sex, and apoE genotype was used as covariate.
These associations imply that CSF levels of AP decline
as severity of dementia increases. In contrast to AP,
CSF levels of neither APPS nor these of T correlated
with either BDS or ADL scores. Regression analyses
also revealed that neither age of disease onset nor disease duration correlated with CSF levels of any of the
above polypeptides. There was also no significant effect
of sex on any of the above measures.
O n western blots, the N-terminal antibody 22C11
consistently detected two major immunoreactive bands
with relative molecular masses of 106 kd and 25 kd in
every CSF sample examined (Fig 3); these accounted
for 90% and 8% of the total immunoreactivity, respectively, as determined by densitometry. Additional minor bands included immunoreactive products with approximate molecular masses of 125 kd and 13 kd.
Secreted N-terminal APP derivatives lacked the Cterminus of APP holoprotein as indicated by the absence of immunoreactivity with C-terminal antisera including a C 8 and antiC6. The 106-kd protein also
reacted strongly with the polyclonal antiserum R1736,
which is directed against residues 595 to 611 (of
APP695, i.e., -2 to 15 of the AP domain). This epitope is absent in APLP2, suggesting that much, if not
Table 2. Cerebrospinal Fluid Levels of Amyloid @-Protein, N-Terminal Amyloid &Protein Precursor Derivatives, and r
Do Not Differ in Alzheimer’s Diseuse Patients with Different Apolipoprotein E Genotypes
Total APPS
Apolipoprotein
E Genotype
E313
€314
€414
p (ANOVA)
Total AD
€313
€314
€414
p (ANOVA)
Total AD
7
(ILg/d)
ELISA
(PgW
Total APPS
Western
(re1 OD)
106-kd APPS
Western
(re1 OD)
Western
(re1 OD x E-3)
Total Protein
(pglml)
1.53 f 0.348
1.54 f 0.472
1.87 i 0.354
0.415
334 i 130
215 i 93.6
319 f 118
0.165
1.92 5 0.721
1.50 ? 0.547
2.20 f 0.942
0.223
1.69 i 0.618
1.31 ? 0.458
1.93 f 0.837
0.212
149 f 60.5
107 2 43.3
140 36.7
0.242
546 i 146
426 _f 131
472 ? 100
0.304
1.62
274
1.79 2 0.73
1.58 t 0.64
127
475 i 130
3.79 f 1.83
3.73 -t 1.36
5.04 i 2.92
0.518
3.33 t 1.58
3.26 f 1.14
4.43 ? 2.60
0.505
0.286 f 0.131
0.235 i 0.081
0.311 k 0.126
0.494
4.13 f 2.01
3.63 f 1.76
A@ELISA
ELlSA
(nglml)
5
9
5
15.04 ? 4.52
11.93 ? 3.54
16.26 * 4.75
0.163
19
13.9 i 4.35
5
29.03 f 10.85
26.74 f 9.22
37.16 2 16.01
0.343
n
9
5
19
30.5
?
12.1
?
0.41
3.03 f 1.21
3.7 f 1.46
4.06 ? 1.3
0.751
3.61 t 1.33
f
118
0.659 ? 0.314
0.509 f 0.353
0.731 i 0.354
0.533
0.62
f
0.34
25-kd APPS
*
f
0.27
48.1
i
-
0.11
~
~
~
Values are expressed per milliliter of cerebrospinal fluid (CSF) (top) and per nullgrain of CSF protein (bottom) and represent means i: SD p values were computed
by single-factor (genotype) analysis of variance (ANOVA)
n = number of patients, AB = amvloid
O-protein,
ELlSA
.
. .
Alzheimer’s disease; re1 OD = relative optical density.
=
enzyme-linked immunosorbent assay, APPS = amyloid @-proteinprecursor, secreted form, AD =
0
0
Q.
4
LL
LA
12
00
00
10
8:
A
CTR
AD
Fig 1 . Cerebrospinulj u i d (CSF) levels of amyloid @-protein
(Ap) ure similar in Alzheimer’s disease patients compared with
those of age-matched, cognitively normal control subjects. E a h
circle indicates means of triplicate measurements of AD in CSF
obtained from one individual. Triangles represent means t SD
of all measurements in each group. There was no significant d$ference between groups.
all, of the immunoreactivity seen at 106 kd in CSF
with 2 2 C l l represents APPS. The monoclonal antibody Anti-Ah90 (clone l.D5), raised against residues
51 1 to 608 (of APP695), also reacted with the 106-kd,
but not with the 25-kd, band. Like R1736, Anti-A1290
is specific for APP, underscoring that the 106-kd band
in human CSF consists mostly of APPS. We assume
that much of the immunoreactivity seen at 106 kd represents a-secretase products although we could not exclude the presence of low levels of p-secretase products [28), which are likely to comigrate with the
106-kd band in the 12% gel used in this study. The
monoclonal antibody 7H5, which is directed against
the Kunitz protease inhibitor (KPI) domain of APP,
detected very low levels of a 125-kd protein, indicating
the presence of KPI-containing APPS in CSF. This protein was also detected with the monoclonal antibody
2 2 C l l . The absolute quantity of the 125-kd protein
was more than one order of magnitude lower than that
of the 106-kd derivative as determined by densitometric comparison of the two bands detected by 22C11.
In addition, a band of approximately 70 kd was visualized with both 7H5 and R1736.
The allelic frequency of both apoE ~4 and apoE
~3 was 0.5 in the group of patients studied. Levels of
the sum of N-terminal APP derivatives, of the individual 106- and 25-kd derivatives, and of AP (expressed
as mass per volume) were similar in the three Apo E
genotype groups (see Table 2). CSF levels of T were
also similar among the three Apo E genotype groups
(see Table 2). T o correct for possible disease-related
changes in CSF water content, all of the above measures were normalized to total CSF protein. These normalized levels also did not differ among the three apoE
genotype groups (see Table 2). T o ensure that possible
effects of dementia severity on CSF levels of A@,T, or
APPS did not mask possible effects of genotype on the
levels of these proteins, analyses of covariance with
BDS, ADL, age, and disease duration were performed.
They also revealed no differences among individual
genotype groups.
Discussion
The major finding of this study indicates that CSF levels of AP are significantly and inversely correlated with
Nitsch et al: CSF A@ in Alzheimer’s Disease
515
201.
1
8
A
8
kDa
I
apoE genotype
I
&3/4
I
&3/3
&4/4
I
200976946-
A
30-
r=-.6656
p=O.O018
5
0
2114-
8
10
20
15
25
30
35
BLESSED DEMENTIA SCALE
lane
20
=
5
a
a
0
8
A
8
18
\i
16
14
12
10
r=-,6194
8
p=0.0047
0
10
8
20
3
5
7
8
9
A%
A
30
2
Fig 3. Apolipoprotein E (apoE) genotype does not a#ect cerebrospinaljuid (CSF) Levels of secreted N-terminal amyloid p
protein precursor (APP) derivatives in Alzheimer’s disease patients. Western blot of APP derivatives in 9 human CSF
samples (10 pl) probed with the monoclonal antibody 22CI 1.
Lanes 1 to 3 = ~ 3 1 3lanes
,
4 to 6 = ~ 3 1 4lanes
,
7 to 9 =
~ 4 1 4 Two
.
major bands with molecular masses of 106 and 25
Ld accountedfor 98% of the total immunoreactivity.
8
0
1
40
50
60
70
80
ACTIVITIES OF DAILY LIVING SCALE
F i g 2. Cerebrospinal j u i d (CSF) levels of amyloid p-protein
(AP) are inversely correlated with clinical measures of dementia
severity both in cognitive (Blessed Dementia Scale; BDS) and
functional (Activities of Daily Living score; ADL) domains.
Each symbol represents a single Alzheimer’s disease patient categorized by apolipoprotein E genotype: circles = ~313; triangles
= ~ 3 1 4squares
;
= ~414.
both cognitive (BDS) and functional (ADL) measures
of dementia severity (see Fig 2). This result suggests
that CSF levels of AP decrease with advancing disease
severity. The finding is somewhat surprising, as one
might expect to find increases in AP levels in a disease
characterized by excessive AP deposition in brain tissue. The inverse association between AP and dementia
severity generates the hypothesis that AP levels might
be high during early stages of the disease, and decrease
as dementia progresses. This decline may reflect decreased production of AP over the time course of AD,
perhaps related to degeneration of cells that secrete
AP into CSF. Alternatively, clearance of AP from CSF
might change with progression of dementia. A reduction of CSF AP during the course of A D could also
be related to the time course of amyloid deposition in
brain. The rate of amyloid deposition is high during
the initial stages of the disease, but amyloid plaque
density reaches a plateau relatively early on, and then
516 Annals of Neurology Vol 37 No 4 April 1995
stays constant over the remaining duration of the illness even though the severity of dementia increases
129, 30). As long as the cellular origin of AP in CSF
is unknown, however, it is difficult to make predictions
as to whether these CSF peptides bear any pathogenetic relation to amyloid plaques in brain parenchyma.
A dilution of CSF proteins caused by brain atrophyrelated ventricular enlargement might conceivably also
account for the lower AP levels in patients with more
advanced dementia. This possibility is less likely, however, because levels of total CSF protein were unrelated to measures of dementia severity ( Y = - 0.070
for BDS and Y = 0.0051 for ADL). In a similar manner, levels of APPS and T , expressed by volume and by
total protein, did not correlate with these variables.
The statistical association of dementia severity and AP
levels was independent of the age of disease onset, sex,
or apoE genotype, and direct comparisons of AP levels between male and female A D patients showed no
effect of sex on these levels. Together, these crosssectional data suggest that decreases in AP levels in
CSF parallel the clinical course of dementia in A D but
confirmatory longitudinal studies will be necessary before concluding that AP levels in CSF provide a useful
biochemical marker to follow disease progression, or
to judge efficacy of therapeutic interventions. The
range of AP levels in our A D cases overlapped broadly
with the range of AP levels in CSF obtained from agematched, cognitively normal control subjects, indicating the limitations of such measurements as a diagnostic tool for A D (see Fig 1).
CSF levels of secreted APP derivatives measured by
ELISA varied from 0.95 to 2.4 pg/ml, or 1.6 to 6.1
Fg/mg CSF protein (see Table a), and were highly
consistent within individuals after a repeated lumbar
puncture performed 2 weeks after the initial one (data
nor shown). These levels are within the same order
of magnitude as those reported by Van Nostrand and
colleagues [I11 and others [l2} who used an antibody
raised against human protease nexin 2.
Western blotting analyses confirmed the presence of
multiple proteolytic APP derivatives in human CSF
(see Fig 3). The 106-kd band detected by the monoclonal antibody 2 2 C l l represents the sum of signals
derived from APPS, small amounts of APLP2, and psecretase products of APP. APLP2 is present in human
CSF at concentrations that are at least one order of
magnitude lower than those of APPS (Wagner S, personal communication). Moreover, the signal generated
by 22C11 with APLP2 is approximately 10 times lower
than that with APP as estimated by western blotting
of in vitro translation products (Wasco W, personal
communication). Thus, less than one-twentieth of the
immunoreactivity seen with 22C11 in human CSF can
possibly be attributed to an unspecific signal caused
by APLP2, whereas more than 95% represents APPS.
Consistent with this view, antibodies directed against
APP-specific epitopes including R1736 and Alz-90 reacted strongly with the 106-kd protein, confirming that
much of the 106-kd band represents APPS. The observation that R1736, an antiserum agahst residues -2
to + 15 of the N-terminus of Ap, detected APPS at
106 kd, and that this protein apparently lacks the Cterminus of full-length APP, suggests that APPS in human CSF most likely is derived from conventional nonamyloidogenic a-secretase cleavage at position Lys l6
to Leu” of the Ap domain [ S ] . Previous studies have
shown that low levels of additional p-secretase products can also be present in human CSF 1281. These
proteins are only 16 residues shorter than a-secretase
products, and because of comigration cannot be separated from APPS by 12% SDS-PAGE. Taken together,
the 106-kd band seen with 2 2 C l l in CSF combines
APPS, a low unspecific signal derived from binding of
secreted APLP2 derivatives, and also low levels of psecretase-derived APP fragments. Further studies using additional and more specific antibodies are needed
to determine the absolute levels and the normal ranges
of these proteins in human CSF.
The second major N-terminal APP derivative detected by 2 2 C l l had a molecular mass of 25 kd (see
Fig 3 ) and was characterized by the absence of immunoreactivity with both R1736 and anti-C6. A 25-kd
APP fragment was shown earlier to be present in human CSF 191, its N-terminus was partially sequenced
and determined to be identical to that of APP, that is,
to begin at residue 18 of the full-length sequence [lo].
In our study, the 25-kd polypeptide accounted for 8%
of the total 22CIl immunoreactivity as estimated by
densitometry and by assuming that binding efficiency
of 2 2 C l l to this fragment is sindar to that of the
106-kd APPS. The precise cellular origin of the 25-kd
N-terminal APP derivative is unknown, but it appears
to be specific to CSF, because neither cultured brain
cells nor rat brain tissue slices are known to secrete
such fragments [31]. It may thus constitute a product
of extracellular proteolysis generated by a protease
present in CSF. Consistent with this interpretation,
both the monoclonal antibody 7H5 and the antiserum
R1736 detected low levels of a corresponding approximately 70-kd C-terminal fragment that might conceivably be derived from such cleavage.
CSF levels of both APPS and the 25-kd APP derivative correlated significantly with those of A@, suggesting that, in AD at least, their generation is stoichiometrically coupled. This observation contrasts with
results from cell culture experiments in which treatments that increase the release of N-terminal derivatives (e.g., stimulation of muscarinic receptors) are associated with a concurrent decrease in Ap secretion
[25, 32). Our results also showed a significant correlation of CSF levels of AP with those of T . The biochemical basis for this correlation is unknown.
Genetic evidence indicated that apolipoprotein E genotype can affect the risk of developing AD, but the
biological basis for this effect is unknown. Apo E alleles
exist in three major forms, e2, e3, and e4. Inheritance
of the e4 allele is associated with a dose-dependent
increase in the risk of developing AD, and predisposes
toward a young age of disease onset, whereas the e2
allele appears to protect against AD or delay the onset
of dementia [14- 181. Consistent with these reports of
genetic disequilibrium, the prevalence of the e4 allele
was high in our group of AD patients (0.5) compared with the expected frequency of 0.14 in the general population. Apolipoprotein E isoforms are present
in human CSF and can bind both AP and T 114, 151.
Moreover, the number of amyloid plaques is highest
in individuals with ~4 alleles {171. Together, these observations generated the hypothesis that CSF levels of
AP, and possibly T , might vary according to apoE genotype. The comparison of AP levels in AD patients with
different apoE genotypes did not support this hypothesis. Individual allelic variants of Apo E were not associated with major differences in CSF levels of T , Ap, and
APPS, including the 106- and 25-kd derivatives (see
Table 2). The 95% confidence bound for Ap comparing the ~ 3 / and
3 ~ 4 1 4groups was 6.7 nglml. The sample size used in this study was therefore sufficient to
exclude a major difference in AP levels between genotypes, although larger sample sizes are needed to determine whether more subtle differences in CSF levels of
AP, T , or APPS might occur in individual Apo E genotype groups. Our results are supported by recent experimental data indicating that the secretion of AP is
Nitsch et al: CSF AP in Alzheimer’s Disease
517
similar in cultured cells overexpressing individual apoE
~3 or e4 alleles 1331. The data obtained in this study
suggest that the mechanism by which apoE alleles affect the risk of getting AD is not reflected by major
differences in CSF levels of APP derivatives or 7 in
AD patients.
This work was supported by NIH GCRC RR-01066, NIA 2P50
AG-01534, NIMH 28783, the Center for Brain Sciences and Metabolism Charitable Trust, and the Hoffman Fellowship in Alzheimer’s Disease.
We thank Dr Dennis Selkoe for the antisera R1736 and uC8,Dr
Kurt Naujocks for the monoclonal antibody 22Cl1, Dr Robert Siman for the recombinant APP standard, Dr Andre Van de Voorde
for the T standard, and Dr David Schoenfeld and Sharon Best for
statistical analyses.
15.
16.
17.
18.
19.
20.
References
1. Seubert P, Vigo-Pelfrey C, Esch F, et al. Isolation and quantification of soluble Alzheimer’s P-peptide from biological fluids.
Nature 1992;359:325-327
2. Shoji M, Golde TE, Ghiso J, et al. Production of the Alzheimer
amyloid P-protein by normal proteolytic processing. Science
1992;258: 126-129
3. Vigo-Pelfrey C, Lee D, Keim P, et al. Characterization of Pamyloid peptide from human cerebrospinal fluid. J Neurochem
1993;61:1965- 1968
4. Vandermeeren M, Mercken M, Vanmechelen E, et al. Detection of T protein in normal and Alzheimer’s disease cerebrospinal fluid with a sensitive sandwich enzyme-linked immunosorbent assay. J Neurochem 1993;61: 1828-1834
5. Haass C, Schlossmacher MG, Hung AY, et al. Amyloid Ppeptide is produced by cultured cells during normal metabolism.
Nature 1992;359:322-325
6. Haass C, Selkoe DJ. Cellular processing of P-amyloid precursor
protein and the genesis of amyloid 6-peptide. Cell 1993;75:
1039-1042
7. Weidemann A, Konig G, Bunke D, et al. Identification, biogenesis, and localization of precursors of Alzheimer’s disease A4
amyloid protein. Cell 1989;57:115-126
8. Sisodia SS, Koo E, Beyreuther K, et al. Evidence that 6-amyloid
protein in Alzheimer’s disease is not derived by normal processing. Science 1990;248:492-495
9. Palmert MR, Berman Podlinsny M, Witker DS, et al. The betaamyloid protein precursor of Alzheimer’s disease has soluble
derivatives found in human brain and cerebrospinal fluid. Proc
Natl Acad Sci USA 1989;86:6338-6342
10. Palmert MR, Usiak M, Mayeux R, et al. Soluble derivatives of
the amyloid protein precursor in cerebrospinal fluid: alterations in normal aging and in Alzheimer’s disease. Neurology
1990;40:1028-1034
11. Van Nostrand WE, Wagner SL, Shankle WR, et al. Decreased
levels of soluble amyloid P-protein precursor in cerebrospinal
fluid of live Alzheimer disease patients. Proc Natl Acad Sci USA
1992;89:255 1-2555
12. Farlow M, Ghetti B, Benson MD, et al. Low cerebrospinal-fluid
concentrations of soluble amyloid P-protein precursor in hereditary Alzheimer’s disease. Lancet 1992;340:453-454
13. Wagner SL, Peskind ER, Nochlin D, et al. Decreased levels of
soluble amyloid P-protein precursor are associated with Alzheimer’s disease in concordant and discordant monozygous twin
pairs. Ann Neurol 1994;36:215-220
14. Strittmatter WJ, Weisgraber KH, Huang DY, et al. Binding of
human apolipoprotein E to synthetic amyloid P peptide: iso-
518 Annals of Neurology
Vol 37
No 4 April 1995
21.
22.
23.
24.
25.
26.
form-specific effects and implications for late-onset Alzheimer
disease. Proc Natl Acad Sci USA 1993;90:8098-8102
Strittmatter WJ, Weisgraber KH, Goedert M, et al. Microtubule
instability and paired helical filament formation in the Alzheimer
disease brain are related to apolipoprotein E genotype. Exp
Neurol 1994;125:163-17 1
PoirierJ, DavignonJ, Bouthillier D, et al. Apolipoprotein E polymorphism and Alzheimer’s disease. Lancet 1993;342:697-699
Rebeck GW, Reiter JS, Strickland DK, Hyman BT. Apolipoprotein E in sporadic Alzheimer’s disease: allelic variation and
receptor interactions. Neuron 1993;11:575-580
Corder EH, Saunders AM, Strittmatter WJ, et al. Gene dose of
apolipoprotein E type 4 allele and the risk of Alzheimer’s disease
in late onset families. Science 1993;261:921-923
Rebeck GW, Peris IT,West HL, et al. Reduced apolipoprotein
€4 allele frequency in the oldest old Alzheimer’s patients and
cognitively normal individuals. Neurology 1994;44:1513-1516
Schmechel DE, Saunders AM, Strittmatter WJ, et al. Increased
amyloid P-peptide deposition in cerebral cortex as a consequence of apolipoprotein E genotype in late-onset Alzheimer
disease. Proc Natl Acad Sci USA 1993;90:9649-9653
McKhann G, Drachman D, Folstein M, et al. Clinical diagnosis
of Alzheimer’s disease: report of the NINCDS-ADRDA group
under the auspices of Department of H H S task force on Alzheimer’s disease. Neurology 1984;34:939-944
Blessed G, Tomlinson BE, Roth M. The association between
quantitative measures of dementia and of senile changes in cerebral grey matter of elderly subjects. Br J Psychiatry 1967;114:
797-811
Weintraub S. The record of independent living: an informantcompleted measure of activities of daily living and behavior in
elderly patients with cognitive impairment. Am J Alzheimer
Care 1986;1:35-39
Kohler G, Milstein C. Affinity purification of human 2 proteins
and the construction of a sensitive sandwich ELISA for human
2 detection. Nature 1975;256:495-497
Nitsch RM, Slack BE, Wurtman RJ, Growdon JH. Release of
Alzheimer amyloid precursor derivatives stimulated by activation of muscarinic acetylcholine receptors. Science 1992;258:
304-307
Wasco W, Gurubhagavatula S, Paradis MD, et al. Isolation and
characterization of APLP2 encoding a homologue of the Alzheimer’s associated amyloid P protein precursor. Nature Genet
1993;5:95-99
27. Hyman BT, Tanzi RE, Marzloff K, et al. Kunitz protease inhibitor containing amyloid P-protein precursor immunoreactivity in
Alzheimer’s disease.J Neuropathol Exp Neurol 1991;51:76-83
28. Seubert P, OltersdorfT, Lee MG, et al. Secretion of 6-amyloid
precursor protein cleaved at the amino terminus of the Pamyloid peptide. Nature 1993;361:260-263
29. Hyman BT, Marzloff K, Arriagada PV. The lack of accumulation of senile plaques or amyloid burden in Alzheimer’s disease
suggests a dynamic balance between amyloid deposition and
resolution. J Neuropathol Exp Neurol 1993;52:594-600
30. Arriagada PV, Growdon JH, Hedley-Whyte ET, Hyman BT.
Neurofibrillary tangles but not senile plaques parallel duration
and severity of Alzheimer’s disease. Neurology 1992;42:631-639
31. Nitsch RM, Farber SA, Growdon JH, Wurtman RJ. Release of
amyloid P-protein precursor derivatives from hippocampal slices
by electrical depolarization. Proc Natl Acad Sci USA 1993;90
5191-5193
32. Hung AY, Haass C, Nitsch RM, et al. Activation of protein
kinase C inhibits cellular production of the amyloid P-protein.
J Biol Chem 1993;268:22959-22962
33. Biere AL, Anthony DC, Selkoe DJ. Influence of Apo E protein
on APP metabolism in intact living cells. SOCNeurosci Abstr
1994;20:1055 (Abstract)
Документ
Категория
Без категории
Просмотров
1
Размер файла
788 Кб
Теги
level, severity, disease, fluid, cerebrospinal, correlation, effect, dementia, protein, apolipoprotein, amyloid, alzheimers, genotypes, inverse
1/--страниц
Пожаловаться на содержимое документа