Journal of Neuropathology and Experimental Neurology
Copyright q 2001 by the American Association of Neuropathologists
Vol. 60, No. 8
August, 2001
pp. 778 785
Cholesterol Accumulates in Senile Plaques of Alzheimer Disease Patients and in Transgenic
APPsw Mice
TAKASHI MORI, DVM, PHD, DANIEL PARIS, PHD, TERRENCE TOWN, BA, AMYN M. ROJIANI, MD, PHD,
D. LARRY SPARKS, PHD, ANTHONY DELLEDONNE, FIONA CRAWFORD, PHD, LAILA I. ABDULLAH, BS,
JAMES A. HUMPHREY, DENNIS W. DICKSON, MD, PHD, AND MICHAEL J. MULLAN, MD, PHD
Abstract. Mounting evidence suggests that cholesterol may contribute to the pathogenesis of Alzheimer disease (AD). We
examined whether cholesterol might be present in senile plaques, a hallmark neuropathological feature of AD. We employed
2 different fluorometric-staining techniques (filipin staining and an enzymatic technique) for the determination of cholesterol
in brains of postmortem confirmed AD patients and in nondemented, age-matched histopathologically normal controls. AD
patient brains showed abnormal accumulation of cholesterol in congophilic/birefringent dense cores of senile plaques that was
essentially absent in histopathologically normal controls. To determine whether increased senile plaque-associated cholesterol
occurred generally in all plaques or was restricted to a specific subset, quantitative analysis was performed. Data indicate
abnormal accumulation of cholesterol in cores of mature plaques but not in diffuse or immature plaques. Additionally,
transgenic mice that overexpress the ‘‘Swedish’’ amyloid precursor protein (Tg APPsw, line 2576) exhibited a similar pattern
of abnormal cholesterol accumulation in mature, congophilic amyloid plaques at 24 months of age that was absent in their
control littermates or in 8-month-old Tg APPsw mice (an age prior to amyloid deposition). Taken together, our results imply
a link between cholesterol and AD pathogenesis and suggest that cholesterol plays an important role in the formation and/or
progression of senile plaques.
Key Words:
Alzheimer disease; Cholesterol; Senile plaque; Transgenic mice.
INTRODUCTION
Alzheimer disease (AD) is a progressive dementing
disorder neuropathologically characterized by deposition
of b-amyloid in senile plaques, formation of neurofibrillary tangles composed of hyper-phosphorylated tau protein, and cortical neuronal loss. Accumulating evidence
suggests that cholesterol may play a role in the pathogenesis of AD. For example, the apolipoprotein E (apoE)
apoprotein is a major constituent of plasma lipoproteins,
which generally function as blood plasma lipid and cholesterol transporters (1). The e4 allele of the apoE gene
is a robust risk factor for both familial and sporadic forms
of AD, predicting increased incidence of AD and an earlier age of onset (2, 3). Further, it has been shown that
high levels of serum cholesterol constitute a risk factor
for AD (4). Finally, it has been reported that levels of
brain cholesterol are altered in the cortices of AD patients
compared to age-matched, nondemented controls (5), and
in vivo experiments have shown that diet-induced hypercholesterolemia enhances intraneuronal accumulation of
Ab accompanied by microgliosis in rabbit brain (6, 7)
and accelerates b-amyloid deposition in brains of a transgenic mouse model of AD (8).
In addition to the putative involvement of cholesterol
in promoting risk for and exacerbating AD pathology,
recent findings have highlighted a direct biochemical impact of cholesterol on amyloid precursor protein (APP)
processing, resulting in increased production of b-amyloid (Ab) peptides. For example, cholesterol increases bsecretase cleavage of APP, thereby promoting Ab secretion (9, 10), while cholesterol depletion decreases the
production of amyloidgenic APP C-terminal fragments as
well as Ab secretion (9, 11). Additionally, Ab itself displays high affinity for cholesterol, which plays a pivotal
role in neuronal integrity, suggesting that Ab may perhaps remove cholesterol from neuronal membranes, leading to neuronal dysfunction (12).
Collectively, it seems plausible that cholesterol is a
contributor to the pathogenesis of AD, but a direct association between cholesterol and AD pathology has been
lacking. We thus hypothesized that cholesterol distribution in brains from AD patients and in a mouse model
of AD (Tg APPsw, which overexpress APP and overproduce Ab peptides [13]) might be abnormal. To evaluate
this hypothesis, we applied 2 different techniques for the
visualization of cholesterol in brains from AD patients,
histopathologically normal human controls, Tg APPsw
mice, and their nontransgenic control littermates, using
filipin staining and a fluorometric enzymatic technique
with cholesterol oxidase.
MATERIALS AND METHODS
From the Roskamp Institute (TM, DP, TT, AD, FC, LIA, MJM), Neuropathology Service (AMR), University of South Florida, Tampa, Florida; Sun Health Research Institute (DLS), Sun City, Arizona; Department of Pathology (DWD), Mayo Clinic, Jacksonville, Florida.
Correspondence to: Takashi Mori, The Roskamp Institute, University
of South Florida, 3515 E. Fletcher Ave., Tampa, FL 33613.
Tissue Preparation
Brain specimens from humans were obtained from individuals who were either histopathologically normal (control) or
met Consortium to Establish a Registry for Alzheimer’s Disease
criteria for AD histopathological diagnosis and were subjected
778
CHOLESTEROL ACCUMULATES IN SENILE PLAQUES
to Braak and Braak staging (14, 15). Initial evaluation of tissue
blocks (from all subjects) from neocortex, hippocampus, basal
forebrain, basal ganglia, thalamus, midbrain, pons, medulla, and
cerebellum included routine H&E, Congo red, Bielschowski,
and Thioflavine S staining, as well as staining with antibodies
to b-amyloid, tau, and a-synuclein proteins. Additionally, fresh
frozen brain tissue from AD cases was subjected to APOE genotyping according to previously published methods (16). A
total of 5 advanced pathologic stage (Braak stage VI) AD cases
(3 female/2 male; age 5 79.2 6 6.61 SD) and 5 age-matched
(p . 0.05 by t-test for independent samples) control subjects
(5 male; age 5 72.8 6 3.03 SD) were used in this study. Control subjects had no history of dementia or neurological disease.
Brain tissues were selected from similar areas of frontal and
temporal cortices for AD cases and controls, and, after fixation
with 10% neutral-buffered formalin, were sliced into 100-mm
sections with a random starting point within the block using a
vibratome (Technical Products International Inc., St. Louis,
MO) for filipin staining, fluorometric enzymatic staining, and
Congo red staining. For quantitative analysis, adjacent brain
tissues of frontal and temporal cortices were trimmed and transferred to 10% sucrose in 100 mM sodium phosphate (pH 7.4),
followed by 20% sucrose in 100 mM sodium phosphate (pH
7.4) for 24 h (for cryoprotection), and were then quick-frozen
at 2808C. Serial frozen sections were cut with a cryotome at
10-mm thickness with 150-mm distance between cuts and stored
at 2808C until needed for b-amyloid immunohistochemical
staining, fluorometric enzymatic staining, or Congo red staining.
Tg APPsw mice are the 2576 line (overexpressing ‘‘Swedish’’
mutant APP) crossed with C57B6/SJL as previously described
(13). Animal housing, care, and surgical procedures complied
with the Principle of Laboratory Animal Care and the Guide
for the Care and Use of Laboratory Animals (DHHS publication No. [NIH] 85-23, revised 1985) and have previously been
approved by the Institutional Animal Care and Use Committee.
A total of 8 mice (Tg APPsw, n 5 4; control littermates, n 5 4)
were used at 8 and 24 months of age (n 5 2 for each time
point). All animals were anesthetized by isoflurane inhalation
and killed by exsanguination. Brains were transcardially perfused with phosphate buffered saline (PBS; Gibco BRL, Grand
Island, NY), pH 7.4, followed by 4% paraformaldehyde in PBS.
Upon removal from the animals, brains were immersed in the
same fixative and cut into 100-mm coronal sections using a
vibratome from prior to the appearance of the hippocampus
(bregma 0.98 to 20.82 mm) to caudal to the dorsal tip of the
hippocampus (bregma 21.70 to 23.28 mm). All mouse brain
sections were stored in PBS at 48C until needed for filipin staining, fluorometric enzymatic staining, or Congo red staining.
Filipin Staining
Filipin is a fluorescent antibiotic complex derived from the
bacterium Streptomyces filipenensis, which selectively binds
cholesterol and structurally related sterols. While filipin does
display some affinity for other molecules containing a lipid
moiety, based on its high affinity for cholesterol it has been
used for cholesterol analysis of mitochondrial membranes (17)
and to demonstrate decreases in membrane cholesterol coinciding with increased membrane fluidity (18).
779
Filipin complex, DMSO, and paraformaldehyde were purchased from Sigma (St. Louis, MO). Floating sections from
humans and mice were washed 3 times in PBS and incubated
overnight at 48C in PBS containing 0.05% filipin complex (filipin was first dissolved in 2.5% DMSO). Sections were then
washed 3 times in PBS and mounted with an aqueous mounting
medium. Signals were examined using a fluorescence microscope (Olympus BX60, Olympus Optical Co., Ltd., Tokyo, Japan), with a specific filter allowing detection at an excitation
wavelength of 350 nm and an emission wavelength of 410 nm.
Importantly, to detect abnormal extracellular accumulation of
filipin complex signals in our samples, it was necessary to
quench fluorescence signals on the sections with long-term exposure to fluorescence light (20–30 s), as the intensity of filipin
complex signals throughout the brain sections examined was
high, evidencing normal cholesterol signal distribution.
Cholesterol Oxidase Staining
By applying a fluorometric enzymatic reaction, cholesterol
can be oxidized by cholesterol oxidase to yield hydrogen peroxide (H2O2) and the corresponding ketone product, cholesterol4-ene-3-one. Thus, the enzymatic methods for detecting cholesterol are based on the detection of H2O2 using horseradish
peroxidase (HRP)-coupled oxidation of H2O2. (HRP was purchased from Molecular Probes, Inc., Eugene, OR). H2O2 is then
detected using 10-acetyl-3, 7-dihydroxyphenoxazine (an H2O2sensitive fluorogenic probe) in the presence of HRP, giving rise
to the production of a highly fluorescent resofurin with an emission maximum of 587 nm, making it less susceptible to interference from the autofluorescence of biological compounds
such as bilirubin and lipoproteins (19).
Cholesterol oxidase (from Streptomyces species) was purchased from Sigma. Ten-acetyl-3, 7-dihydroxyphenoxazine
(Amplex Red), 0.1 M potassium phosphate (pH 7.4), 0.05 M
NaCl, 5 mM cholic acid, and 0.1% Triton X-100 were purchased from Molecular Probes, Inc. (Eugene, OR). Vibratome
brain sections from humans and mice were reacted on a microscope slide glass humidified with 50 ml of a solution containing
2 U/ml HRP, 2 U/ml of cholesterol oxidase, 300 mM of 10acetyl-3, 7-dihydroxyphenoxazine, 0.1 M potassium phosphate
(pH 7.4), 0.05 M NaCl, 5 mM cholic acid and 0.1% Triton X100. Furthermore, the reaction solution mentioned above was
also applied to human frozen brain sections for quantitative
analysis. After 10 s of incubation with this solution at room
temperature, vibratome or frozen sections were mounted with
an aqueous mounting medium after a gentle wash with PBS
and then observed under a fluorescence microscope with a filter
allowing detection at an excitation wavelength of 530–560 nm
and an emission wavelength of 590 nm. As 1 U/ml of cholesterol oxidase will oxidize 1.0 mM of cholesterol to 4-cholesten3-one per min at pH 7.5 at 258C, a subset of brain sections were
pretreated before incubation for 8 h at room temperature with
cholesterol oxidase (100 ml of a solution containing 100 U/ml
cholesterol oxidase, 0.1 M potassium phosphate [pH 7.4], 0.05
M NaCl, 5 mM cholic acid and 0.1% Triton X-100) to extract
cholesterol from brain sections, providing for a negative control
to insure the validity of the cholesterol signal. As additional
negative controls, brain sections were incubated with the reaction solution without cholesterol oxidase or HRP to confirm the
specificity of the reaction.
J Neuropathol Exp Neurol, Vol 60, August, 2001
780
MORI ET AL
Fig. 1. Photomicrographs of filipin complex staining and Congo red staining in the frontal cortex of AD patient No. 1 (A–
C) and in the cingulate cortex of a Tg APPsw mouse at 24 months of age (D–F). Bright blue, amorphous or spotted, abnormal
extracellular filipin complex signals are noted by fluorescence microscopy in panels (A, D). Note co-localization of filipin complex
signals with congophilic/birefringent plaques (B, C, E, F). There was no evidence for abnormal extracellular accumulation of
filipin complex signals in human controls, 8-month-old Tg APPsw mice, or control littermates (at 8 or 24 months of age) (data
not shown). Scale bar: 100 mm.
Congo Red Staining
After the determination of cholesterol using either of the
above techniques, identical sections were treated with Congo
red staining to investigate whether or not cholesterol-indicating
signals were co-localized with congophilic senile plaques. Brain
sections were incubated in alkaline sodium chloride solution for
10 min, stained with alkaline Congo red solution (0.2% in 80%
ethanol saturated with sodium chloride; Sigma), and then
washed in a graded series of ethanol. Congo red-stained sections were then observed under the light microscope using both
conventional and polarized light sources.
Quantitative Analysis
To investigate which plaque types may be co-localized with
cholesterol-indicating signals, serial frozen sections of the frontal and temporal cortices from AD cases were examined, one
section stained with cholesterol oxidase for the determination
of cholesterol-indicating signals and the adjacent section stained
for b-amyloid plaques. Detection of b-amyloid plaques was performed according to the manufacturer’s protocol using the
DAKO LSAB1 Systems kit coupled to a diaminobenzidine
(DAB) reaction intensified with nickel ammonium sulfate to
yield a purple precipitate. Briefly, the frozen sections were
blocked with an endogenous peroxidase quenching step using
0.3% H2O2 and the normal serum preblocking step prior to immunostaining. The frozen sections were then incubated with
monoclonal antibody 4G8 (Senetec Napa, CA) diluted 1:1,000,
followed by the incubation of the biotinylated secondary antibody and DAB amplification step. PBS alone was used as a
negative control. For quantitative analysis, different plaque sub-
J Neuropathol Exp Neurol, Vol 60, August, 2001
types were counted in 5 fields (1 mm2 each, using a 103 objective and a gridded 103 eyepiece lens) within the same region
from 5 different sections that were cut with a random starting
point within the block and a 150-mm distance between slices
(to omit most of the double-counting of senile plaques and cholesterol plaques) from selected blocks of similar areas of frontal
and temporal cortices. Different plaque subtypes were assigned
to 1 of the 3 categories (diffuse, immature, and mature) according to criteria previously described (20), which have subsequently been validated (21). Additionally, cholesterol-indicating signals by cholesterol oxidase staining were counted in
adjacent sections in the same fashion. All counts were performed in a blinded manner by a single examiner (T.M.). Accumulated data are expressed as the mean 6 SD. Data were
normally distributed as indicated by the one-sample Kolmogorov-Smirnov Z test; therefore, one-way analysis of variance
(ANOVA) was used to analyze the data followed by post-hoc
comparison using the Tukey HSD test. A p-value of less than
0.05 was considered to be statistically significant. All analyses
were performed using SPSS release 10.0.
RESULTS
Co-localization of Filipin Staining with Senile Plaques
Fluorescence microscopy revealed pale to bright blue,
spotted and amorphous, abnormally intense extracellular
filipin complex signals in brain sections from frontal and
temporal AD patient cortices (Fig. 1A). Deposits with
similar reaction patterns were evident throughout the neocortices and hippocampi of coronal brain sections from
CHOLESTEROL ACCUMULATES IN SENILE PLAQUES
Tg APPsw mice at 24 months of age (Fig. 1D), showing
abnormal extracellular accumulation of cholesterol. By
contrast, no abnormal extracellular accumulation of cholesterol was observed in any of the sections examined
from human controls, Tg APPsw mice at 8 months of age,
or control littermates at 8 or 24 months of age (data not
shown). As revealed by staining identical sections with
Congo red, abnormal cholesterol-indicating signals were
essentially co-localized with congophilic/birefringent
dense cores of mature plaques in AD cases (Fig. 1B, C)
and in 24-month-old Tg APPsw mice (Fig. 1E, F). Congo
red staining revealed small, round, dense cores surrounded by large halos in mature plaques in AD patients (Fig.
1B, C). In Tg APPsw mice, large, round, monotone congophilic/birefringent dense cores were observed in the
well organized, mature plaques (Fig. 1E, F). Interestingly,
abnormal cholesterol-indicating signals in AD patients
were smaller in size than those of Tg APPsw mice, corresponding to the denser, smaller core size of mature
plaques in AD subjects versus Tg APPsw mice (Fig. 1A,
D). This difference between Tg APPsw mouse and human
senile plaques has recently been recognized (22).
Verification of Abnormal Cholesterol Accumulation in
Senile Plaques
While filipin displays high affinity for cholesterol, previous reports have suggested that filipin interacts with
lipid-associated macromolecules, such as membrane
phospholipids (23). Thus, to further confirm the presence
of cholesterol in senile plaques, we employed a fluorometric enzymatic-based reaction technique using cholesterol oxidase. Under the fluorescence microscope, bright
red resofurin signals of various spot sizes were highlighted in brain sections of AD patient frontal and temporal
cortices (Fig. 2A) and in the neocortices and hippocampi
of coronal brain sections from Tg APPsw mice at 24
months of age (Fig. 2G, J). Consistent with filipin complex staining, these signals were also co-localized with
congophilic/birefringent dense cores of mature plaques in
AD cases (Fig. 2B, C) and in 24-month-old Tg APPsw
mice (Fig. 2K, L). Similar to filipin staining, these spotted signals were larger in size in Tg APPsw mice than in
AD patients, coinciding with their different plaque core
sizes. Ubiquitous diffuse, bright red resofurin signals
were also visualized in each of the brain sections examined (except when cholesterol oxidase or HRP was omitted from the reaction solution), showing normal background distribution of cholesterol in each of our sample
groups (which appeared to be increased in Tg APPsw mice
compared to control littermates, Fig. 2G versus H). In
frontal and temporal cortex sections from human controls
(Fig. 2E), and in coronal sections of neocortex and hippocampus from control littermate mice at 8 months (similar to Fig. 2H) and 24 months of age (Fig. 2H), no abnormal extracellular accumulation of cholesterol signals
781
was present in any of the sections examined. We also
investigated brain sections of Tg APPsw mice at 8 months
of age (which do not yet display b-amyloid deposits
[13]), and did not observe abnormal extracellular accumulation of cholesterol, only normal brain cholesterol
distribution in each section observed (similar to Fig. 2H).
Additionally, cholesterol-indicating resofurin signals
were markedly suppressed or absent when brain sections
were pretreated with cholesterol oxidase before incubation with the reaction solution, or when cholesterol oxidase or HRP was omitted from the reaction solution (AD
case, Fig. 2F; 24-month-old Tg APPsw mouse, Fig. 2I).
Cholesterol Abnormally Accumulates in Mature Senile
Plaques
To investigate which b-amyloid plaque subtypes may
be co-localized with abnormal cholesterol-indicating signals, different plaque types were counted in serial frozen
sections using cholesterol oxidase staining for cholesterol
plaques and Ab staining for b-amyloid plaques. In frontal
and temporal cortices from AD patients, mean plaque
subtype number/mm2 from individual cases are shown in
the Table. A graph summarizing quantitative analysis
from these 5 AD cases shows the distribution of different
b-amyloid plaque subtypes in the frontal cortex (diffuse
plaque [38.6 6 2.7/mm2], immature plaque [22.1 6 4.6/
mm2], and mature plaque [5.6 6 1.0/mm2]), and in the
temporal cortex (diffuse plaque [39.5 6 3.1/mm2], immature plaque [22.9 6 2.3/mm2], and mature plaque [5.6
6 1.1/mm2]). Thus, on average, approximately 60% of
the total number of b-amyloid plaques were diffuse type,
another ;30% were immature, and the remaining ;10%
were mature plaques in both cortices examined (similar
to b-amyloid plaque subtype distributions reported by
[21]). Cholesterol plaques were identified only in mature
b-amyloid plaques in the frontal (5.3 6 0.9/mm2) and
temporal (5.2 6 0.8/mm2) cortices (Fig. 3). Additionally,
cholesterol plaque counts were similar in the male and
female AD cases that we observed, and the intensity and
distribution of normal background cholesterol signals,
while being greater in AD cases than controls in general,
was similar between male and female AD cases or controls (data not shown). There was no significant effect of
APOE e4 status on the number of cholesterol plaques in
AD brain (Table). Most importantly, abnormal accumulation of cholesterol was essentially ubiquitously present
in mature b-amyloid plaques (p 5 0.995 when comparing
mature plaques to abnormal cholesterol-positive mature
plaques in the both frontal and temporal cortices), and
was not detected in diffuse or immature b-amyloid
plaques.
DISCUSSION
Several lines of evidence have implicated cholesterol as
pathogenic in AD. We wished to examine whether abnormal accumulation of brain cholesterol might be present in
J Neuropathol Exp Neurol, Vol 60, August, 2001
782
MORI ET AL
Fig. 2. Photomicrographs of cholesterol oxidase fluorometric staining, Congo red staining and Ab immunohistochemical
staining in the frontal cortex of AD patient No.4 (A–D, F), a human control (E), the hippocampus of a Tg APPsw mouse at 24
months of age (G, I, J–L), and an age-matched control littermate mouse (H). Bright red, spotted, resofurin signals are evident by
fluorescence microscopy (A, G, J). These signals are co-localized with congophilic/birefringent plaques (B, C, K, L), but not
with diffuse plaques (A, D). By contrast, only diffuse resofurin signals are observed indicating normal cholesterol distribution in
the frontal cortex of a human control (E) and in the hippocampus of an age-matched control littermate mouse (H). Resofurin
signals are greatly diminished by cholesterol oxidase pretreatment in the AD patient (F) or the Tg APPsw mouse (I). Resofurin
signals were absent when cholesterol oxidase or HRP was eliminated from the reaction solution, and these abnormal cholesterol
deposits were not evident in human controls. There was no evidence of abnormal extracellular accumulation of cholesterol in 8month-old Tg APPsw mice or control littermates (at 8 or 24 months of age) (data not shown). Scale bars: A–D, J–L 5 50 mm;
E–I 5 200 mm.
AD patients and in Tg APPsw mice. Based on the hypothesis that abnormal accumulation of cholesterol might exacerbate Ab pathology, we were primarily interested in
evaluating the putative presence of cholesterol in senile
J Neuropathol Exp Neurol, Vol 60, August, 2001
plaques. Using filipin staining and a cholesterol-specific
fluorometric enzymatic technique, we detected abnormal
extracellular accumulation of cholesterol in congophilic/
birefringent dense cores of mature plaques but not in
Mean plaque subtype numbers/mm are shown from each of 5 AD cases. Data are represented as means 6 1 SD (each figure represents n 5 5 fields). Abnormal
cholesterol staining was found only in mature b-amyloid plaques (last row). Abbreviations: Fc; Frontal cortex, Tc; Temporal cortex, M; male, F; female; y, years.
4.0 6 1.5
6.0 6 0.8
5.0 6 1.5
b-amyloid plaque
Diffuse plaque
Immature plaque
Mature plaque
Cholesterol plaque
Mature plaque
2
5.4 6 1.8
5.6 6 0.9
5.0 6 1.5
4.0 6 0.7
5.0 6 0.8
6.4 6 1.3
6.0 6 1.7
39.0 6 4.5
15.4 6 2.9
4.0 6 1.6
41.8 6 2.3
23.4 6 3.2
5.2 6 0.8
42.6 6 8.1
23.0 6 2.7
6.2 6 1.6
38.8 6 5.4
25.8 6 2.4
5.8 6 1.9
35.2 6 1.8
22.2 6 2.7
5.8 6 0.8
34.6 6 2.9
27.4 6 2.3
6.0 6 1.0
36.6 6 4.2
19.6 6 2.1
4.2 6 0.8
39.6 6 5.3
24.6 6 3.6
5.2 6 1.3
40.4 6 4.9
23.4 6 2.4
7.0 6 1.9
41.6 6 4.7
20.0 6 5.3
6.8 6 1.8
(Tc)
(Fc)
(Tc)
(Fc)
(Tc)
(Fc)
(Tc)
(Tc)
(Fc)
(Fc)
No. 4 (F/80 y/e2/e3)
No. 3 (F/87 y/e3/e4)
No. 2 (F/69 y/e2/e4)
No. 1 (M/82 y/e3/e4)
TABLE
Mean Plaque Subtype Number/mm2 in Alzheimer Patients
No. 5 (M/78 y/e2/e3)
CHOLESTEROL ACCUMULATES IN SENILE PLAQUES
783
Fig. 3. Graph summarizing quantitative analysis of cholesterol plaques co-localized with different subtypes of AD patient
b-amyloid plaques. Data are represented as mean plaque subtype number/mm2 1 1 SD, n 5 25 fields (5 from each AD case)
for each bar. One-way ANOVA followed by post-hoc comparison showed significant differences between the numbers of
each b-amyloid plaque subtype in both the frontal (Fc) and
temporal (Tc) cortices (Diffuse . Immature . Mature; p ,
0.001). Most importantly, abnormal accumulation of cholesterol
was present in essentially all observed mature b-amyloid
plaques (p 5 0.995 when comparing mature plaques to abnormal cholesterol positive mature plaques in the Fc and in the
Tc), and was not detected in diffuse or immature b-amyloid
plaques.
immature or diffuse plaques in brains from AD patients
and in Tg APPsw mice at 24 months of age. Yet, in brains
from human controls, Tg APPsw mice at 8 months of age,
or in control littermates at 8 or 24 months of age, there
was no evidence of abnormal extracellular accumulation
of cholesterol. Additionally, the presence of cholesterol
in senile plaques by cholesterol oxidase fluorometric
staining was further confirmed using 3 strategies: preincubation of brain sections with a nonlimiting amount of
cholesterol oxidase, elimination of cholesterol oxidase, or
removal of HRP from the fluorometric enzymatic reaction solution. These experiments revealed that the observed fluorometric signals were markedly reduced or
completely absent, allowing the interpretation that these
signals were specific for cholesterol. Quantitative analysis
showed that these cholesterol plaques were invariably colocalized with mature-type senile plaques in AD patient
frontal and temporal cortices and in Tg APPsw mouse neocortices and hippocampi. The possibility should be noted,
however, that abnormally accumulated cholesterol might
also be present in other senile plaque subtypes, but not
concentrated enough to be visualized using this method.
Additionally, we did not find an effect of APOE e4 status
on numbers of cholesterol plaques (albeit with a relatively small sample size), although the use of Braak and
Braak stage VI cases precludes data supporting the possibility that APOE e4 may influence cholesterol plaques
earlier on in the course of the disease, and further study
J Neuropathol Exp Neurol, Vol 60, August, 2001
784
MORI ET AL
is needed in this area. Nonetheless, our results, taken together with the appropriate control experiments, show enhanced abnormal accumulation of cholesterol in mature
plaques in AD patients and in Tg APPsw mice, and represent the first visualization that cholesterol deposits are
topographically associated with mature senile plaques.
Physiologically, cholesterol is an integral membrane
component required for normal cellular function (24), and
its metabolism is tightly regulated by the cell (25). Yet,
a number of possible scenarios may account for the observation of abnormal cholesterol deposits in mature senile plaques. Cholesterol deposits seem related in time to
senile plaque formation (being absent in 8-month-old Tg
APPsw mice, an age just prior to b-amyloid deposition
[13]), but technical limitations aside, we did not observe
abnormal cholesterol deposits in immature or diffuse-type
senile plaques. This suggests that these cholesterol deposits most likely occur sometime after rather than coincident with abnormal clearance/degradation of brain bamyloid. Following from this idea, our observation of
strict co-localization of cholesterol plaques with mature
senile plaques raises the possibility that the mature bsheet conformation of b-amyloid as opposed to less organized mixed conformations of the peptide (including
random coil or a-helical structures) binds cholesterol
with high affinity thereby recruiting it. Alternatively, we
observed that background levels of cholesterol appeared
to be increased in Tg APPsw mice versus littermates (Fig.
2G, H) and in AD cases versus controls (data not shown),
raising the possibility that increased amounts of total
brain cholesterol actively favor or ‘‘catalyze’’ the formation of mature-type senile plaques in Tg APPsw mice
and in AD brain. However, a detailed biochemical analysis of brain cholesterol levels in AD cases and controls,
as well as in transgenic mouse models of AD, is warranted to substantiate this speculation. Our observation of
abnormal cholesterol accumulation specifically in the
dense cores of mature senile plaques bolsters this notion,
while tending not to support the idea that cholesterol
plaque co-localization with senile plaques nonspecifically
involves cholesterol binding to all senile plaque subtypes.
Importantly, Tg APPsw mice fed a diet high in cholesterol
show exacerbated Alzheimer’s b-amyloid pathology (8).
Taken together with our data, the suggestion arises that
abnormal accumulation of cholesterol in senile plaques is
not simply an epiphenomenon or later consequence of
senile plaque pathology.
At the cellular level, cholesterol preferentially interacts
with a subset of membrane lipids and proteins, forming
specialized microdomains, such as caveolae, that play important roles in cellular functions such as signaling, adhesion, and motility (11, 26). Our findings raise the possibility that abnormal accumulation of cholesterol in
senile plaques may disturb such normal cellular roles for
J Neuropathol Exp Neurol, Vol 60, August, 2001
cholesterol. It has been shown that cholesterol is decreased in cellular membranes from AD brains (27) and
in AD patient cerebrospinal fluid (28), suggesting a repartitioning of cholesterol from areas where it plays a
normal physiologic role to brain regions that display bamyloid pathology. However, further study is needed to
confirm or deny this hypothesis. Specifically, clues as to
the source of senile plaque-associated cholesterol (i.e.
cell-free or cell-associated) would assist in determining
whether the brain’s pool of cell-free cholesterol is being
repartitioned or whether cellular debris derived from degenerating cells (such as dystrophic neurites, which are
well known to comprise senile plaques) may make up the
observed abnormal cholesterol plaques. Certainly, the
strict co-localization of abnormal cholesterol deposits in
mature-type senile plaques (which contain the greatest
proportion of degenerating cells and dystrophic neurites)
is consistent with the idea that degenerating cell-derived
debris are a possible source of these cholesterol plaques.
However, as it has been reported that Ab has a high affinity for free cholesterol (12), it stands to reason that if
senile plaque-associated abnormal cholesterol deposits
arose from cell-free cholesterol, all senile plaque subtypes would contain these cholesterol deposits.
In conclusion, our results show that cholesterol accumulation is topographically associated with mature-type
senile plaques in humans and in 24-month-old Tg APPsw
mice, suggesting that cholesterol may be important for
the formation and/or progression of senile plaques. Provided this, further investigation into the effects of cholesterol-lowering drugs, such as the statin family of compounds, may be warranted as a novel treatment for AD.
ACKNOWLEDGMENTS
The authors are grateful to Mr. and Mrs. Robert Roskamp for their
generous support, which helped to make this work possible. We would
also like to acknowledge the Alzheimer’s Disease Initiative Brain Bank
Program, in particular Dr. Ranjan Duara and Dr. Gary Pearl.
REFERENCES
1. Mahley RW. Apolipoprotein E: Cholesterol transport protein with
expanding role in cell biology. Science 1988;240:622–30
2. Poirier J, Davignon J, Bouthillier D, Kogan S, Bertrand P, Gauthier
S. Apolipoprotein E polymorphism and Alzheimer’s disease. Lancet
1993;342:697–99
3. Roses AD. Apolipoprotein E alleles as risk factors in Alzheimer’s
disease. Annu Rev Med 1996;47:387–400
4. Notkola IL, Sulkava R, Pekkanen J, et al. Serum total cholesterol,
apolipoprotein E epsilon 4 allele, and Alzheimer’s disease. Neuroepidemiol 1998;17:14–20
5. Sparks DL. Coronary artery disease, hypertension, apoE, and cholesterol: A link to Alzheimer’s disease? Ann NY Acad Sci 1997;
826:128–46
6. Sparks DL, Scheff SW, Hunsaker JC 3rd, Liu H, Landers T, Gross
DR. Induction of Alzheimer-like beta-amyloid immunoreactivity in
the brains of rabbits with dietary cholesterol. Exp Neurol 1994;126:
88–94
CHOLESTEROL ACCUMULATES IN SENILE PLAQUES
7. Sparks DL. Intraneuronal beta-amyloid immunoreactivity in the
CNS. Neurobiol Aging 1996;17:291–99
8. Refolo LM, Pappolla MA, Malester B, et al. Hypercholesterolemia
accelerates the Alzheimer’s amyloid pathology in a transgenic
mouse model. Neurobiol Dis 2000;7:321–31
9. Frears ER, Stephens DJ, Walters CE, Davies H, Austen BM. The
role of cholesterol in the biosynthesis of beta-amyloid. Neuroreport
1999;10:1699–1705
10. Mills J, Reiner PB. Regulation of amyloid precursor protein cleavage. J Neurochem 1999;72:443–60
11. Simons K, Ikonen E. Functional rafts in cell membranes. Nature
1997;387:569–72
12. Avdulov NA, Chochina SV, Igbavboa U, Warden CS, Vassiliev AV,
Wood WG. Lipid binding to amyloid beta-peptide aggregates: preferential binding of cholesterol as compared with phosphatidylcholine and fatty acids. J Neurochem 1994;69:1746–52
13. Hsiao K, Chapman P, Nilsen S, et al. Correlative memory deficits,
A beta elevation, and amyloid plaques in transgenic mice. Science
1996;274:99–102
14. Mirra SS. The CERAD Neuropathology Protocol and Consensus
Recommendations for the Postmortem Diagnosis of Alzheimer’s
Disease: A Commentary. Neurobiol Aging 1997;18(S4):S91–S94
15. Braak H, Braak E. Neuropathological staging of Alzheimer-related
changes. Acta Neuropathol 1991;82:239–59
16. Wenham PR, Price WH, Blandell G. Apolipoprotein E genotyping
by one-stage PCR. Lancet 1991;337:1158–59
17. Cremel G, Filliol D, Jancsik V, Rendon A. Cholesterol distribution
in rat liver and brain mitochondria as determined by stopped-flow
kinetics with filipin. Arch Biochem Biophys 1990;278:142–47
18. Coxey RA, Pentchev PG, Campbell G, Blanchette-Mackie EJ. Differential accumulation of cholesterol in Golgi compartments of normal and Niemann-Pick type C fibroblasts incubated with LDL: A
cytochemical freeze-fracture study. J Lipid Res 1993;34:1165–76
19. Amundson DM, Zhou M. Fluorometric method for the enzymatic
determination of cholesterol. J Biochem Biophys Methods 1999;38:
43–52
785
20. Masliah E, Terry RD, Mallory M, Alford M, Hansen LA. Diffuse
plaques do not accentuate synapse loss in Alzheimer’s disease. Am
J Pathol 1990;137:1293–97
21. Masliah E, Mallory M, Hansen L, et al. Immunoreactivity of CD45,
a protein phosphotyrosine phosphatase, in Alzheimer’s disease.
Acta Neuropathol (Berl) 1991;83:12–20
22. Kuo YM, Kokjohn TA, Beach TG, et al. Comparative analysis of
Ab chemical structure and amyloid plaque morphology of transgenic mice and Alzheimer disease brains. J Biol Chem 2001;276:
12991–98
23. Milhaud J, Lancelin JM, Michels B, Blume A. Association of polyene antibiotics with sterol-free lipid membranes: I. Hydrophobic
binding of filipin to dimyristoylphosphatidylcholine bilayers.
Biochim Biophys Acta 1996;1278:223–32
24. Bloch K. Biochemistry of lipids, lipoproteins, and membranes. In:
Vance DE, Vance JE, eds. Amsterdam: Elsevier Science Publishers
BV, 1991:363–81
25. Tabas I. Free cholesterol-induced cytotoxicity. A possible contributing factor to macrophage foam cell necrosis in advanced atherosclerotic lesions. Trends Cardiovasc Med 1997;7:256–63
26. Schroeder F, Woodford JK, Kavecansky J, Woods WG, Joiner C.
Cholesterol domains in biological membranes. Membr Biol 1995;
12:113–19
27. Roth GS, Joseph JA, Mason RP. Membrane alterations as causes of
impaired signal transduction in Alzheimer’s disease and aging.
Trends Neurosci 1995;18:203–6
28. Mulder M, Ravid R, Swaab DF, et al. Reduced levels of cholesterol,
phospholipids, and fatty acids in cerebrospinal fluid of Alzheimer
disease patients are not related to apolipoprotein E4. Alzheimer Dis
Assoc Disord 1998;12:198–203
Received February 1, 2001
Revision received April 17, 2001
Accepted May 4, 2001
J Neuropathol Exp Neurol, Vol 60, August, 2001