Regional
Distribution
of
Neurofibrillary
Tangles and Senile
Plaques
in the Cerebral Cortex of
Elderly
Patients:
A
Quantitative
Evaluation
of a
One-Year Autopsy
Population
from
a
Geriatric
Hospital
Detailed analyses of the neuropathologic changes in the
cerebral cortex
of
elderly individuals and Alzheimer's
disease patients have demonstrated that certain com-
ponents of the neocortical and hippocampal circuits are
likely to be selectively vulnerable. Based on the distri-
bution
of
neurofibrillary tangles (NFTs)
and
senile
plaques,
it
has been proposed that
a
global cortico-
cortical disconnection leads
to the
loss
of
integrated
functions observed
in
Alzheimer's disease.
In
order
to
investigate the distribution
of
lesions associated with
aging as well as with the earliest symptoms
of
senile
dementia, we performed
a
quantitative neuropathologic
evaluation
of a
large series
of
elderly patients repre-
senting the entire autopsy population for the year 1989
from
a
geriatric hospital. Among the 145 cases quan-
titatively assessed, there were
102
nondemented
pa-
tients,
33
patients presenting clinically with globally
intact intellectual function but early signs of impairment
of specific cognitive functions, and 10 cases with senile
dementia
of
the Alzheimer type.
All of
the cases had
NFTs in layer
II of
the entorhinal cortex, regardless
of
their clinical diagnosis, and most cases had some NFTs
in the CA1 field of the hippocampus. Severe pathologic
changes within
the
inferior temporal neocortex were
observed only
in the
demented cases.
The
extent
of
amyloid deposition was not correlated with the clinical
diagnosis and seemed
to
be present in the neocortical
areas earlier than
in
the hippocampal formation. Also,
several cases contained NFTs without amyloid depo-
sition,
but amyloid never occurred without NFTs. These
results suggest that involvement
of
certain structures
within the hippocampal formation is a consistent feature
of aging. Thus, involvement
of the
hippocampal
for-
mation may be a necessary, but not sufficient, condition
for the clinical expression
of
dementia, which
is
likely
to
be
more closely related
to the
progressive degen-
eration of select neuronal populations in the neocortex.
Constantin Bouras,'-
2
Patrick R.
Hof,
2
-
3
Panteleimon
Giannakopoulos^ Jean-Pierre Michel,' and John H.
Morrison
2
-
3
'Department of Psychiatry, IUPG Bel-Air, University
of Geneva School of Medicine, 1225 Geneva,
Switzerland,
2
Fishberg Research Center for
Neurobiology and
3
Department of Geriatrics and
Adult Development, Mount Sinai School of
Medicine, New York, New York 10029, and
^Geriatric Hospital, University of Geneva School of
Medicine, 1226 Geneva, Switzerland
Neurofibrillary tangles (NFTs), senile plaques (SPs),
and loss of large neurons are three pathological find-
ings commonly observed in the cerebral cortex of
Alzheimer's disease cases (Tomlinson et al., 1970;
Ball, 1977; Terry et al., 1981; Mountjoy et al., 1983).
The laminar and regional localization pattern of NFTs
and SPs within the cerebral cortex indicates that a
selective degeneration of long corticocortical and
hippocampal projection systems is likely to occur
(Mutrux, 1947; Pearson et al., 1985; Rogers and Mor-
rison, 1985; Hyman et al., 1986, 1988, 1990; Lewis et
al.,
1987; Hof and Morrison, 1990; Hof et al., 1990a,b;
Arnold etal.,
1991;
Braak and Braak,
1991;
De Lacoste
and White, 1993), leading to a disconnection of the
cerebral cortex and resulting in the disintegration of
intellectual functions observed in Alzheimer's dis-
ease patients (Pearson et al., 1985; Rogers and Mor-
rison, 1985; Hyman et al., 1986; Lewis et al., 1987; Hof
and Morrison, 1990; Hof et al., 1990a,b). In fact, Hy-
man and colleagues have demonstrated that the pres-
ence of NFTs in layer II of the entorhinal cortex was
correlated by the presence of pathologic profiles in
the termination zone of the perforant pathway in the
outer portion of the molecular layer of the dentate
gyrus,
implying that this cortical pathway is severely
affected in Alzheimer's disease cases (Hyman et al.,
1988).
In addition, the distribution of NFTs and SPs
in the neocortex and in the hippocampal formation
suggests that SP formation may involve the terminal
arborization of NFT-bearing neurons (Pearson et al.,
1985;
Rogers and Morrison, 1985; Lewis et al., 1987;
Hof and Morrison, 1990; Hof et al., 1990a; Senut et
al.,
1991; De Lacoste and White, 1993).
The distribution and density of these lesions in the
cerebral cortex have been correlated with the degree
of cognitive impairment in Alzheimer's disease. For
example, Tomlinson et al. (1970) have reported a
significant correlation between the mean SP frequen-
cy from several areas of the cerebral cortex and de-
mentia severity assessed within 6 weeks of death us-
ing the Blessed Dementia Scale. However, these
authors did not document the relationship between
cortical NFT formation and dementia severity in their
series.
Other quantitative evaluations of the associa-
tion of SPs and NFTs with dementia severity in Al-
zheimer's disease demonstrated a substantially great-
er correlation of cortical NFT than SP density,
particularly in the temporal neocortex and parahip-
pocampal gyrus (Wilcock and Esiri, 1982; Arriagada
Cerebral Cortex Mar/Apr 1994;4:138-150; 1047-3211/94/J4.00
et al., 1992a). In addition, the severity of dementia
has been shown to be highly correlated with synapse
loss in neocortical areas (Tern' et al., 1991).
The presence of Alzheimer's disease-related le-
sions in the brains of nondemented elderly individ-
uals is well established (Tomlinson et al., 1968; Da-
yan, 1970; Ball, 1977; Mountjoy et al., 1983; Ulrich,
1985).
More recently, several studies have demon-
strated that amyloid deposition and NFT formation
not only are common features of brain aging, but are
found in certain regions of the cerebral cortex in den-
sities that would qualify for a neuropathologic diag-
nosis of Alzheimer's disease cases (Crystal et al., 1988;
Katzman et al., 1988; Hubbard et al., 1990; Morris et
al.,
1991; Price et al., 1991; Arriagada et al., 1992a,b;
Hof et al., 1992; Berg et al., 1993; Bouras et al., 1993).
In particular, a careful analysis of such cases revealed
that in this population, the hippocampal formation
systematically displayed
a
moderate to severe involve-
ment, whereas the neocortical areas were unaffected
or showed very low densities of NFTs and SPs (Price
et al., 1991; Arriagada et al., 1992a
;
Hof et al., 1992;
Berg et al., 1993; Bouras et al., 1993). This difference
in regional lesion distribution distinguishes these el-
derly nondemented cases from the Alzheimer's dis-
ease population, in which neocortical areas are dev-
astated (Pearson et al., 1985; Lewis et al., 1987; Hof
et al., 1990a,b, 1992; Bouras et al., 1992). Most of the
recent clinicoanatomic analyses of brain aging have
been based on relatively small patient samples con-
sisting of selected cases (Crystal et al., 1988; Katzman
et al., 1988; Hubbard et al., 1990; Morris et al., 1991;
Price et al.,
1991;
Arriagada et al., 1992a,b; Hof et al.,
1992).
In the present study, we had the opportunity
to survey all of the autopsied cases from a geriatric
hospital for the year 1989. We thus analyzed a large
number of nonselected cases from a nonpsychiatric
hospital in order to assess the degree to which elderly
patients are affected by Alzheimer's disease-related
neuropathologic changes, and to evaluate the age-
related factors leading to NFT formation and amyloid
deposition in a nondemented population. Part of this
study has been reported in abstract form (Bouras et
al.,
1992).
Materials and Methods
Survey of Population
We surveyed the entire population of autopsy cases
from the Geriatric Hospital of the University of Ge-
neva School of Medicine (Switzerland) for the year
1989.
This hospital is a large university-based insti-
tution (256 beds) that functions as the main geriatric
care center for the greater Geneva area (about 400,000
inhabitants). A total number of 176 cases, all older
than 65 years of age, were included in the present
study (Table 1). There were 77 men (81.2 ± 7.5 years
old; age range, 65-95; 43.75%) and 99 women (84.7
± 7.5 years old; age range,
66-101;
56.25%). Clinical
data on the patients were obtained from the medical
records of the Geriatric Hospital and from the Neu-
ropathological Database of the Division of Morpho-
logical Psychopathology, Department of Psychiatry,
University of Geneva School of Medicine, Geneva,
Switzerland. All cases were neuropsychologically
evaluated using a Mini-Mental State Examination
(Folstein et al., 1975). Although the vast majority of
the patients were hospitalized for terminal illness, all
had been tested neuropsychologically at least
4
months
prior to death. However, in some demented cases
presenting with terminal somatic illnesses (i.e., met-
astatic tumor, severe cardiac and/or renal insufficien-
cy) as well as in a few cases with Wernicke-Korsakoff
encephalopathy, neuropsychological evaluation was
no longer
tenable.
These cases (other dementia cases,
n = 28) were subsequently excluded from the ana-
tomoclinical correlations, although they were used in
the general neuropathological evaluation. In addi-
tion, three cases were also excluded from the analysis
because it was impossible to assess their mental and
cognitive function at admission (one oligophrenic
man, 85 years old, and two women presenting with
severe aphasia due to extensive vascular lesions, 87
and 89 years old, respectively). Thus, for anatomo-
clinical correlations, 145 cases (60 men, 79.6 ± 8.2
years old, age range 65-95, and 85 women, 83.3 ±
7.5 years old, age range 66-101) were considered and
were subdivided into three diagnosis groups accord-
ing to clinical and neuropathological findings (Table
1).
Nondemented patients (ND cases, n = 102) pre-
sented with preserved intellectual functions and
showed no sign of temporospatial disorientation and
memory impairment. Demented patients exhibited a
greater variability in the clinical presentation. For in-
stance, some cases presented with severe disorien-
tation and memory impairment (DMI cases, n = 33),
but without aphasia-apraxia-agnosia syndrome. Other
cases exhibited clinically the typical symptomatology
of degenerative dementia of the Alzheimer type (AD
cases,
n = 10), with marked temporospatial disori-
entation and cognitive impairment, aphasia, apraxia
and agnosia, and severe handicap in daily living. Al-
zheimer's disease was subsequently confirmed neu-
ropathologically in all cases from this group. Finally,
a large number of cases of all three diagnosis group
displayed variable degrees of vascular damage (n =
107).
This finding is not uncommon in the brain of
elderly patients, and a recent neuropathologic study
of 50 demented cases has revealed that up to 30% of
the cases suffered from mixed dementia (Fallet-Bian-
co et al., 1990).
A full routine neuropathological evaluation of all
the cases was performed on selected cortical areas
(superior frontal, inferior temporal, inferior parietal,
primary and secondary visual cortex, and anterior hip-
pocampal formation). The densities and severity of
neuropathological changes (i.e., NFTs, amyloid de-
position, gliosis, microinfarcts, cortical atrophy) were
assessed for each area separately and filed in a custom
database. The total number of NFTs and amyloid-pos-
itive profiles was assessed from a series of random 1
mm
2
cortical samples in all of the cortical areas sur-
veyed, for each case separately. For this purpose, am-
Cerebral Conex Mar/Apr 1994, V 4 N 2 139
Table 1
Distribution
of the
diagnosis groups
Diag-
nosis
Age
|±SD|
Sex
|M/W|
Fre-
quency
ND
DMI
AD
Other
81.3
± 7.8
83.1
±
7.4
88.6
±
4.8
87.8
± 5.1
43/59
17/16
0/10
16/12
102
33
10
28
59.0
19.0
5.8
16.2
NO,
nondememed;
DMI,
disorientation
and
memory impairment;
AD,
degenerative
dementia
of the
Alzheimer type; Other, other dementia (refers
to
demented cases
of
various etiology).
A
total number
of 145
patients were considered.
Due to
incomplete
clinical data,
a
total
of
31 cases were excluded from
the
analysis
(the
other dementia
group
and
three additional cases;
see
Materials
and
Methods).
M/W,
men/women.
yloid-positive profiles were not differentiated into
morphologic subgroups, and diffuse deposits, clas-
sical and core plaques were considered together for
the assessment of amyloid deposition within the se-
lected cortical regions. Depending on the density of
lesions, cases were subsequently subdivided for neu-
ropathologic assessment into four groups displaying
an absence of lesions, rare lesions (<1O NFTs or am-
yloid-positive profiles per mm
2
), moderate densities
of lesions (10-20 NFTs or amyloid-positive profiles
per mm
2
), or frequent lesions (>20 NFTs or amyloid-
positive profiles per mm
2
), respectively. All of the
cases with a clinical history of dementia were rated
as having frequent lesions in all of the cortical areas
surveyed, with lesion densities compatible with those
recommended by the CERAD guidelines (Mirra et al.,
1991,
1993). Additional tissues were available for de-
tailed analysis and immunohistochemistry.
Tissue Collection and Staining Procedure
All the brains were collected at autopsy (postmortem
delay, 2-24 hr), fixed, and subsequently stored in a
large volume of 10% formalin solution. For routine
neuropathological evaluation 20-^mthick section
were prepared from the selected areas (see above)
and were stained with modified Globus (Globus, 1927;
Hof et al., 1990a) and Gallyas (Gallyas, 1971; Hof et
al.,
1990a) silver impregnation techniques, Luxol-fast
blue-Van Gieson, cresyl violet, and hematoxylin-eo-
sin. In addition, a modified thiofiavine S method was
used for routine semiquantitative neuropathologic
as-
sessment (Guntern et al., 1992). This method yields
results comparable to those obtained on tissues stained
with antibodies to the microtubule-associated protein
tau or to the amyloid 0A4 protein (Vallet et al., 1992).
Briefly, 20-^m-thick sections were mounted onto poly-
L-lysine-coated slides. Then, they were treated with
0.25%
KMnO, in PBS for 4 min. The sections were
then placed in a solution consisting of
1%
K,S
2
O, and
1%
oxalic acid in PBS for 2-3 min, rinsed, and stained
with 0.0125% thiofiavine
S
in 40:60 100% ethanol:PBS
(v/v ratio) for 3 min. They were finally differentiated
in
50%
ethanol in
PBS
and coverslipped with glycerin:
H
2
O (3:1). With this method, lesions can be easily
identified under fluorescence lighting conditions us-
ing a Zeiss narrow-bandpass fiuorescein filter set. For
immunohistochemistry, additional sections were pro-
cessed with a specific antibodies to the microtubule-
associated protein tau and to the amyloid
/3A4
protein.
Characterization and specificity of these antibodies
have been fully reported elsewhere (Defossez et al.,
1988;
Kim et al., 1988; Delacourte et al., 1990). Briefly,
20-^m-thick sections were incubated overnight with
either the anti-tau antibody or the anti-/3A4 antibody,
both at a dilution of
1:4000.
Following incubation,
sections were processed by the PAP method using
diaminobenzidine as a chromogen.
Quantitative Analysis
NFTs and SPs were counted on a series of nine sec-
tions stained with thiofiavine
S
(sections 1,4, 7), anti-
tau antibody (sections 2, 5, 8), and anti-|8A4 antibody
(sections 3, 6, 9). In the present study, quantitative
analysis was performed only on sections from the an-
terior hippocampus, entorhinal, rostral half of inferior
temporal cortex (Brodmann area 20), superior frontal
cortex (Brodmann area
9),
and primary and secondary
visual cortex (Brodmann areas 17 and 18). For quan-
titative purposes, only classical and core plaques were
counted and referred to as SPs. Lesion densities were
assessed by two independent investigators (C.B.,
P.R.H.), with very high interrater reliability (>95%).
In all of the cases, NFT and SP densities were cal-
culated per mm
2
within each cortical layer for each
selected area (five laminar measurements per sam-
ple).
All quantitative analyses were performed on a
computer-assisted microscopy system consisting of a
Zeiss Axioplan photomicroscope equipped with a
motorized stage, high-sensitivity video camera (LH-
4036 LHESA Electronics), a Compaq Deskpro 386/20
microcomputer, and an SAMBA 2005 software devel-
oped by TITN Inc. (ALCATEL; Grenoble, France). Sta-
tistical analysis was performed using analysis of vari-
ance to compare distribution of lesions (NFTs, SPs)
in the different diagnosis groups. Correlations be-
tween presence of lesions, age, and severity of the
disease were calculated for the different cortical
regions.
Results
Regional Patterns of
NFT
and SP Distribution
In all of the 10 cases presenting with a clinical symp-
tomatology of dementia and with neuropathologically
confirmed Alzheimer's disease, NFT and SP distri-
butions were comparable to previous descriptions in
all the areas investigated (Pearson et al., 1985; Rogers
and Morrison, 1985; Lewis et al., 1987; Hof et al.,
1990a; Arnold et al., 1991; Braak and Braak, 1991).
For instance, NFTs were widespread in the hippo-
campal formation and predominated in the
CA1
field,
subiculum, and layers II and
V
of the entorhinal cor-
tex. In the neocortex, NFTs were present in both lay-
ers III and
V
(Figs.
1,
2B,D), and were more numerous
in the inferior temporal cortex than in the superior
frontal, inferior parietal, and occipital cortex. SPs were
present in high densities in all cortical areas and were
preferentially located in the supragranular layers
within the neocortical regions. Furthermore, there
140 Cerebral Cortex Lesions in Aging • Bouras et al.
STW^^^^i^^*
c
I
D
4 I
Figure
1. NH
distribution
in
layer
II ol
the emorhinal cortex
\A-C\
and
in
layer
III ol
the interior temporal cortex
\D-F) in an
ND case
\A. D), a
DMI case
(ft f|.
and an AD case
(C f\.
There
is a
gradual increase
ol
NFT density
in
the three diagnosis
groups Also note thai layer
II ol
the emorhinal conex
is
sigmlicamly attected
in
the ND and DMI case
[A. B\.
Materials were stained with antibodies
to
the protein tau
[A-C] or to
the amyloid f)A4 protein (O-f
|.
Scale bar Kill ^m.
n
o
B
4 ,
* •
S
*
u
f -
i\
WM '
mon of SPs |A O\ and NFTs |fl D) \ nferior temporal cone- -b and :.u. : in
an Nil rase. Note the presence ol very rare NFTs
in
hn^ jreas
•;•
:omrasnng with the high density
ol
SPs
[A. C\.
Ma
Figure 2. Distribution of SPs |A C] and NFTs |fl, 0) in the inferior temporal cortex [A, B) and superior frontal cortex \C, D) in an ND case. Note the presence of very rare NFTs in both areas \B. D), contrasting with the high density of SPs (A C), Material!
were stained with antibodies to the amyloid 0A4 protein (A C] or the protein tau (ft 0). Cortical layers are indicated by Roman numerals: WM. white matter. Scale bar, 100 /nm.