Caffeine Suppresses Amyloid Beta Levels in Plasma and Brain of Alzheimer's Disease Transgenic Mice

UNDER EMBARGO UNTIL JULY 6, 2009, 00:00 CET
Journal of Alzheimer’s Disease 17 (2009) 681–697
681
DOI 10.3233/JAD-2009-1071
IOS Press
Caffeine Suppresses Amyloid-? Levels in
Plasma and Brain of Alzheimer’s Disease
Transgenic Mice
Chuanhai Caoa,b, John R. Cirritoc, Xiaoyang Lina,b, Lilly Wangb,d, Deborah K Vergesc,
Alexander Dicksonb,d, Malgorzata Mamcarzb,d, Chi Zhanga,b, Takashi Morie, Gary W. Arendashb,d,?,
David M. Holtzmanc,f,g and Huntington Pottera,b,h
aThe Byrd Alzheimer’s Center & Research Institute, Tampa, FL, USA
bFlorida Alzheimer’s Disease Research Center, University of South Florida, Tampa, FL, USA
cDepartment of Neurology, Washington University School of Medicine, St. Louis, MO, USA
dDepartment of Cell Biology, Microbiology, and Molecular Biology, University of South Florida, Tampa, FL, USA
eDepartments of Medical Science and Pathology, Saitama Medical Center and Saitama Medical University,
Kawagoe, Saitama, Japan
f Department of Developmental Biology, Washington University School of Medicine, St. Louis, MO, USA
gHope Center for Neurological Disorders, Washington University School of Medicine, St. Louis, MO, USA
hSuncoast Gerontology and Alzheimer’s Center, University of South Florida College of Medicine, Tampa, FL, USA
Abstract. Recent epidemiologic studies suggest that caffeine may be protective against Alzheimer’s disease (AD). Supportive
of this premise, our previous studies have shown that moderate caffeine administration protects/restores cognitive function
and suppresses brain amyloid-? (A?) production in AD transgenic mice. In the present study, we report that acute caffeine
administration to both young adult and aged AD transgenic mice rapidly reduces A? levels in both brain interstitial ?uid and
plasma without affecting A? elimination. Long-term oral caffeine treatment to aged AD mice provided not only sustained
reductions in plasma A?, but also decreases in both soluble and deposited A? in hippocampus and cortex. Irrespective of caffeine
treatment, plasma A? levels did not correlate with brain A? levels or with cognitive performance in individual aged AD mice.
Although higher plasma caffeine levels were strongly associated with lower plasma A?1?40 levels in aged AD mice, plasma
caffeine levels were also not linked to cognitive performance. Plasma caffeine and theophylline levels were tightly correlated,
both being associated with reduced in?ammatory cytokine levels in hippocampus. Our conclusion is two-fold: ?rst, that both
plasma and brain A? levels are reduced by acute or chronic caffeine administration in several AD transgenic lines and ages,
indicating a therapeutic value of caffeine against AD; and second, that plasma A? levels are not an accurate index of brain A?
levels/deposition or cognitive performance in aged AD mice.
Keywords: Alzheimer’s disease, amyloid-?, brain interstitial ?uid, caffeine, plasma, transgenic mice
INTRODUCTION
are not known to be disease-modifying, and can have
signi?cant undesirable side-effects. Thus, it would be
Synthetic anti-Alzheimer’s disease (AD) drugs cur-
most desirable to identify an inherently safe and read-
rently on the market have mild symptomatic bene?ts,
ily available “nutriceutic” compound (e.g., natural and
normally in the diet) that can provide therapeutic util-
ity against AD. Caffeine is a methylxanthine, high-
?Corresponding author: Gary W. Arendash, Ph.D., Department
ly concentrated in coffee, and probably the world’s
of Cell Biology, Microbiology, & Molecular Biology, University of
South Florida, Tampa, FL 33620, USA. Tel.: +1 813 974 1584; Fax:
most widely consumed psychoactive substance [1].
+1 813 974 1614; E-mail: arendash@cas.usf.edu.
The well known ability of caffeine to increase alert-
ISSN 1387-2877/09/$17.00 ? 2009 – IOS Press and the authors. All rights reserved

---

UNDER EMBARGO UNTIL JULY 6, 2009, 00:00 CET
682
C. Cao et al. / Caffeine Suppresses Amyloid-? Levels in Plasma and Brain of Alzheimer’s Disease Transgenic Mice
ness and arousal primarily involves antagonism of cen-
indicating that moderate caffeine intake has no adverse
tral nervous system adenosine receptors, while addi-
effects on the cardiovascular system, bone status, cal-
tional mechanisms of caffeine action (e.g., phospho-
cium balance, or the incidence of cancer during ag-
diesterase inhibition, calcium mobilization) have been
ing [4,10]. Indeed, a recent study involving 18–24 year
proposed [2], although they occur only at high, unphys-
follow-ups found that coffee intake (4–6 cups per day)
iologic concentrations of caffeine (mM range). Oral
was associated with reduced mortality, particularly due
caffeine is rapidly and almost completely absorbed via
to cardiovascular disease [11].
the gastrointestinal tract, with blood caffeine levels
The potential for caffeine to treat established AD
quickly equilibrating with brain tissue levels due to
has yet to be explored in human studies. As an ini-
caffeine’s unhindered traversal of the blood-brain bar-
tial step for elucidating possible ef?cacy of caffeine to
rier [3,4]. Caffeine is primarily metabolized in the
stabilize or reverse established AD, we have recent-
liver to theophylline and paraxanthine, both of which
ly completed a “treatment-based” study in aged APP-
are at least as physiologically active as caffeine [3].
sw mice that already contained A? pathology (see ac-
There are no apparent differences in metabolism of, or
companying paper [12]). These aged mice were con-
physiological responses to, caffeine in elder individuals
?rmed to be cognitively impaired in working memo-
compared to young individuals; moreover, caffeine’s
ry prior to receiving several months of oral caffeine
pharmacokinetics are similar after oral or intravenous
treatment. When re-tested, aged APPsw mice receiv-
administration in humans and animals [3].
ing caffeine treatment exhibited working memory that
Recent longitudinal studies spanning 4–10 years sug-
was not only substantially better than APPsw mice that
gest that habitual caffeine/coffee intake protects against
did not receive caffeine, but comparable to normal non-
cognitive impairment in aging humans [5,6]. More-
transgenic mice [12].
Thus, even with pre-existing
over, an epidemiologic study evaluated caffeine intake
and substantial A? neuropathology, aged APPsw mice
during the 20 years preceding AD diagnosis and found
exhibited memory restoration with caffeine treatment,
that AD patients consumed markedly less caffeine dur-
suggesting a therapeutic potential of caffeine in cases
ing that period compared to age-matched individuals
of established AD.
without AD [7]. Collectively, these and other obser-
Substantial evidence suggests that the brain’s pro-
vational human studies [8] suggest that habitual caf-
duction and aggregation of A? peptide represent key
feine/coffee intake may protect against memory impair-
events underlying AD pathogenesis. As depicted in
ment and AD during aging. However, because such
Fig. 1A, newly produced A? enters a dynamic equilib-
studies are not controlled and cannot isolate the effect
rium between soluble and deposited A? in the brain,
of caffeine from the myriad of other lifestyle choic-
with continual transport of soluble A? out of the brain
es humans make, we performed a highly controlled
and into plasma. In view of our ?ndings that caffeine
“protection-based” study in mice [9]. In that study,
decreases A? production in APPsw mice through sup-
we found that oral caffeine administration to AD trans-
pression of both BACE1 and ?-secretase/PS1 [9], we
genic (APPsw) mice from young adulthood into old-
hypothesize that resultant lower brain levels of solu-
er age: 1) protected these mice from otherwise cer-
ble A? will at least acutely result in lower plasma A?
tain cognitive impairment in older age; and 2) limited
levels (Fig. 1B). This hypothesis is supported by the
their brain production of the peptide amyloid-? (A?).
?nding that A? is rapidly produced and cleared from
The moderate amount of caffeine intake given to these
the brain [13]. In the present study, we determine the
APPsw mice (human equivalency of 5 cups of coffee
effect of both acute and chronic caffeine administration
per day) suppressed both ?-secretase (BACE1) and ?-
on plasma and brain A? levels in AD mice. In addition,
secretase/PS1 levels in hippocampus, indicating that
we explore possible relationships between: 1) plasma
caffeine can directly impact AD pathogenesis in these
and brain pools of A?; 2) plasma caffeine/theophylline
AD mice. This study in AD mice is consistent with the
concentrations and plasma/brain A? levels; and 3) plas-
human epidemiologic literature supporting an ability of
ma A?/caffeine levels and cognitive performance. Pos-
moderate caffeine intake to reduce risk of AD [7]. Para-
sible associations between A?, caffeine, and cytokines
doxically, human intake of coffee/caffeine declines ap-
are also considered. Both young adult and aged AD
preciably during aging in Western cultures, in part due
mice were utilized in these studies to investigate the
to caffeine intake restrictions often suggested by health
effect of caffeine at both early (pre-A? deposition) and
care professions. Such restrictions would appear un-
late (robust A? deposition) stages of the disease in
warranted, based on comprehensive literature searches
mice.

---

UNDER EMBARGO UNTIL JULY 6, 2009, 00:00 CET
C. Cao et al. / Caffeine Suppresses Amyloid-? Levels in Plasma and Brain of Alzheimer’s Disease Transgenic Mice
683
Fig. 1. Diagrams depicting brain A? production/clearance, the suppressive actions of caffeine on A? production, and resultant effects on brain
and plasma A? levels. (A) Unmodulated: A? is primarily produced in neurons, secreted into the brain extracellular space in soluble form,
then enters a dynamic equilibrium between soluble and deposited (insoluble) A?. Continual transport of soluble A? occurs into plasma. (B)
Caffeine: Short-Term: Caffeine suppression of both ?- and ?-secretase activities reduces A? production, resulting in lower soluble A? in brain
and plasma. The equilibrium between soluble and deposited A? is not impacted by this short-term reduction in brain soluble A? levels. (C)
Caffeine: Long-Term: Continued caffeine suppression of A? production and resultant lower levels of brain soluble A? induce a ?ux of deposited
(insoluble) A? to the soluble form, which is cleared from brain into plasma via soluble A? transport. Plasma A? levels may be reduced or
not changed, depending on degree of caffeine-induced suppression of A? production. In aged APPsw mice given chronic caffeine treatment,
their lower brain A? levels/deposition results in reversal of cognitive dysfunction. (Colours are visible in the electronic version of the article at
www.iospress.nl.)
Based on encouraging results from our complet-
microdialysis study, which had a C57/BL6/SJL back-
ed caffeine administration studies in AD transgenic
ground and were a gift from Karen Hsiao Ashe (Uni-
mice [9,12], clinical trials in aged individuals are cur-
versity of Minnesota; Hsaio et al. 1996). All mice
rently being conducted to investigate the effects of caf-
were maintained on a 12-hour dark and 12-hour light
feine administration on plasma A? levels. As such, a
cycle with ad libitum access to rodent chow and wa-
major aim of the present study was to attain advance
ter/caffeinated water. All animal procedures were per-
insight into potential results from those clinical trials
formed in AAALAC-certi?ed facilities under protocols
through similar studies in AD transgenic mice.
approved by Institutional Animal Care and Use Com-
mittees at University of South Florida and the JA Haley
VA Hospital.
MATERIAL AND METHODS
General protocol
Animals
A spectrum of studies involving both acute and
In these studies, all mice (with two exceptions) had a
chronic caffeine administration was performed in
mixed background of 56.25% C57, 12.5% B6, 18.75%
APPsw/Tg2576 and APPsw+PS1 transgenic lines,
SJL, and 12.5% Swiss-Webster and were derived from
with plasma, neurochemical, and/or behavioral mea-
a cross between heterozygous mice carrying the mu-
sures collected. It should be underscored that these
tant APPK670N, M671L gene (APPsw) with heterozy-
transgenic lines have measurable levels of soluble A?
gous PS1 (Tg line 6.2) mice. This resulted in offspring
in both brain and plasma in young adulthood, with their
consisting of mutant APPsw, PS1, APPsw+PS1, and
brain levels of A? increasing appreciably during ag-
non-transgenic (NT) genotypes. The two exceptions
ing. This results in A? plaque formation beginning
were: 1) APPsw+PS1 mice used in the 7-day gav-
around 10–11 months of age for APPsw/Tg2576 mice
age treatment study, which had a B6C3 background
and by 6 months of age for APPsw+PS1 mice. Thus,
and were obtained from the Jackson Laboratory (Bar
the 3–4 month-old APPsw/Tg2576 mice used in these
Harbor, MA); and 2) Tg2576 mice used in the in vivo
studies had no A? deposition, while the 14 month and

---

UNDER EMBARGO UNTIL JULY 6, 2009, 00:00 CET
684
C. Cao et al. / Caffeine Suppresses Amyloid-? Levels in Plasma and Brain of Alzheimer’s Disease Transgenic Mice
older APPsw mice exhibited age-dependent A? depo-
or eA?” [14]. The pool of eA? is in dynamic equi-
sition. All APPsw+PS1 mice utilized in these studies
librium with the total pool of ISF A?. During micro-
were at least 15 months of age and therefore had robust
dialysis, mice were housed in a constant light condi-
A? deposition. The general protocol for each study is
tion and remained awake with freedom of movement
indicated below:
and ad lib food and water during microdialysis. Mi-
crodialysis perfusion buffer was arti?cial cerebrospinal
Acute caffeine cffects on plasma A? levels
?uid (CSF) containing 0.15% bovine serum albumin
Acute (single treatment) administration of caffeine
that was ?ltered through a 0.1 µm membrane. Flow
or saline vehicle was given by intraperitoneal (i.p.) in-
rate was a constant 1.0 µl/minute, which recovers 23.4
jection or gavage to the following groups of mice: 3–4
± 1.7% (mean ± SEM) of exchangeable A? within
month-old APPsw mice (i.p.), 14 month-old APPsw
the brain ISF of Tg2576 mice. Samples were collected
mice (i.p. or gavage), and 14 month-old APPsw+PS1
every 30–60 minutes with a refrigerated fraction col-
mice (gavage). A pre-treatment blood sample (0.15
lector into polypropylene tubes and assessed for A? by
ml) was taken by sub-mandibular vein puncture 3–4
sandwich ELISA at the completion of each experiment,
days before treatment. At 3–4 hours following caffeine
as described by Cirrito et al. [15]. Brie?y, A?
(1.5 mg/0.2 ml; Sigma, St. Louis, MO) or vehicle ad-
x?40
ministration, another blood sample was taken. In these
was assessed using an A?40-speci?c mouse monoclon-
acute studies, each group of caffeine- or vehicle-treated
al antibody, mHJ2, as a coating antibody and a bi-
animals consisted of 5–7 Tg mice. For all acute caf-
otinylated central domain antibody, mHJ5.1, as the de-
feine treatment studies, an equivalent volume of 0.9%
tecting antibody, followed by streptavidin-poly-HRP-
saline (0.2 ml) was given immediately following any
40 (Fitzgerald Industries, Concord, MA). All ELISA
blood sample taken.
assays were developed using Super Slow ELISA TMB
(Sigma) and absorbance read on a Bio-Tek FL-600 plate
Long-term caffeine effects on plasma A? levels
reader (Winooski, Vermont) at 650 nm. Basal levels of
Longer term caffeine administration was given by
ISF A? were de?ned as the mean concentration of A?
gavage to aged APPsw+PS1 mice. At 3–4 days fol-
over 5–6 hours preceding drug treatment. The mean
lowing a pre-treatment blood sample, 15–20 month-
in vivo concentration of basal ISF eA? was 3.9 ± 0.78
old APPsw+PS1 mice (n = 6) were started on twice-
ng/ml (n = 6). For each animal, all A? levels were
daily caffeine treatment (1.5 mg/0.2 ml each) via gav-
normalized to the basal A? concentration. Once basal
age for 7 consecutive days. A blood sample was tak-
ISF A? levels were established, Tg2576 mice were ad-
en on the ?nal day of caffeine treatment, as well as 9
ministered caffeine i.p. and ISF A? levels were sam-
days thereafter. A second group of four 20 month-old
pled every 30 minutes until the end of the study. Mice
APPsw+PS1 mice were pre-treatment bled, and then
were studied at 3 months of age, which is prior to A?
given two caffeine treatments (1.5 mg/0.2 ml each) via
deposition in this mouse model.
gavage every 4th day for up to two months. In this sec-
ISF A? half-life was determined similar to Cirrito
ond long-term study, blood samples were taken every
et al. [14]. Microdialysis probes were inserted as de-
8th day during treatment, always on the day following
scribed previously. Basal levels of ISF A? were estab-
a treatment. As was the case for acute studies, volume
lished for ?ve hours, followed by i.p. administration
replacement with 0.9% saline occurred immediately
of 30 mg/kg caffeine or vehicle (PBS). Three hours
following each blood sample.
after treatment, the ?-secretase inhibitor Compound E
In vivo microdialysis: Acute caffeine effects on
(Alexis Biochemicals, San Diego, CA; 100 nM) was
interstitial ?uid (ISF) levels of A? in hippocampus
added to the microdialysis perfusion buffer to rapidly
In vivo microdialysis was used to assess brain ISF
inhibit A? production near the probe. The IC50 for
A?
this compound to inhibit ?-secretase activity in vitro is
x?40 in the hippocampus of awake, freely moving
Tg2576 mice. Microdialysis was performed similar
0.3 nM. Microdialysis samples were collected every 30
to previously described methods [14]. This technique
minutes for an additional four hours and then assessed
samples soluble molecules within the extracellular ?u-
for A?x?40 by sandwich ELISA. The half-life of A?
id that are smaller than 38-kilodaltons, the molecular
was calculated based on the slope of decline in ISF
weight cut off of the probe membrane. A? capable of
A? levels [14] beginning one hour after the onset of
entering the probe has been dubbed “exchangeable A?
Compound E administration.

---

UNDER EMBARGO UNTIL JULY 6, 2009, 00:00 CET
C. Cao et al. / Caffeine Suppresses Amyloid-? Levels in Plasma and Brain of Alzheimer’s Disease Transgenic Mice
685
Long-term caffeine effects in behaviorally-tested mice
Everyville, CA). Quantitative image analysis was then
At 18–19 months of age, APPsw and NT littermate
performed according to Mori et al. [18] and A? bur-
controls were pre-tested in the radial arm water maze
den was determined as a percentage of immunolabeled
(RAWM) task of working memory according to our
area (positive pixels) relative to the full area captured
established protocol [9,12,16]. Following con?rma-
(total pixels). Quantitative image analysis was done
tion that APPsw mice were cognitively-impaired, they
based on previous methods with modi?cations [18,19].
were divided into two groups, with half of Tg mice
Images were acquired using an Olympus BX60 mi-
started on caffeine administered in their drinking wa-
croscope with an attached digital camera system (DP-
ter (0.3 mg/ml, providing a daily dose of ? 1.5 mg
70, Olympus, Tokyo, Japan). The digital image was
caffeine/mouse) and the other half remaining on stan-
then routed into a Windows PC for quantitative anal-
dard tap water, as detailed previously [9].
At 4–5
ysis using SimplePCI software (Compix Inc., Imaging
weeks into caffeine treatment, all mice were re-tested
Systems, Cranberry Township, PA). Images of ?ve 5-
in the RAWM, with both last block and overall errors
µm sections (150 µm apart) through the hippocampus
being analyzed for working memory Trials 4 and 5.
were captured from each animal, and a threshold op-
Following completion of behavioral testing at 20–21
tical density was obtained that discriminated staining
months of age, mice were deeply anesthetized with
from background. Each region of interest was manual-
sodium pentobarbital, a blood sample was then taken
ly edited to eliminate artifacts. For A? burden analy-
for neurochemical analysis, followed by transcardial
sis, data are reported as percentage of immunolabeled
perfusion with 100 ml of 0.9% saline. Postmortem
area captured (positive pixels) relative to the full area
brains were immediately removed and bisected sagital-
captured (total pixels). Each analysis was done by a
ly. The hippocampus and cerebral cortex was dissected
single examiner blinded to sample identities.
from the right hemisphere and processed for soluble
A?1?40 and A?1?42 determinations by ELISA, as well
Plasma neurochemical analysis
as for cytokine levels. Brie?y, 30 mg brain samples
were homogenized in 400 µl RIPA buffer [100 mM
Plasma from blood samples was analyzed for levels
Tris [pH8.0], 150 mM NaCl, 0.5% DOC, 1% NP-40,
of A?1?40, caffeine, theophylline (a caffeine metabo-
0.2% SDS, and 1 tablet proteinase inhibitor per 100
lite), and cytokines. A?1-40 levels were determined
ml (S8820, Sigma)], and sonicated for 20 seconds on
according to the aforementioned methodology involv-
ice. Samples were then centrifuged for 30 minutes at
ing brain tissues. Plasma caffeine and theophylline
27,000 g at 4?C, and supernatants were transferred into
concentration were measured with ELISA Kits from
new screw cap tube. The supernatants obtained from
Neogen (Lansing, MI, USA) by following the manu-
this protocol were then stored at ?80?C for determina-
facture protocol. In brief, the enzyme conjugate solu-
tion of soluble A? levels using ELISA kits (KHB3482
tion was prepared by diluting the 180X enzyme conju-
for 40, KHB3442 for 42, Invitrogen, Carlsbad, CA).
gate stock 1 to 180 in the EIA buffer provided. Caf-
Standard and samples were mixed with detection an-
feine (or theophylline) standard was then diluted with
tibody and loaded on the antibody pre-coated plate as
EIA buffer at two fold dilutions from 200 ng/ml to 0.39
the designated wells. HRP-conjugated antibody was
ng/ml. Then 20 µl standard of each dilution was added
added after wash, and substrates were added for col-
into the coated plate. Plasma samples were then diluted
orimetric reaction, and then stopped with sulfuric acid.
with EIA buffer, with 20 µl of this dilution added into
Optical density was obtained and concentrations were
the coated plate. Both standard and samples were run
calculated according a standard curve. Plasma A?1?40
in duplicate in the plate. Positive and negative controls
and A?1?42 levels were determined with the same pro-
of 20 µl were loaded to each plate. Then 180 µl of di-
tocol and using the same ELISA kits. Homogenates
luted drug-enzyme conjugate was added into each well
from the right hippocampus and cerebral cortex were
and mixed by gently shaking the plate. Plates were
also analyzed for cytokine levels (see next section).
covered with plastic ?lm and incubated at room tem-
The left hemisphere was histologically processed
perature for 45 minutes. During the incubation, a 10x
for analysis of total A? deposition, as previously de-
wash buffer was diluted to 1X with DI water and mixed
scribed [17]. Brie?y, equally-spaced 5-µm sections at
thoroughly. Once incubation was completed, the liquid
the hippocampal level were immunostained for total
was dumped from the wells. Plates were then taped on
A? deposition using a biothinylated human A? mon-
a clean lint-free towel to remove any remaining liquid
oclonal antibody (clone 4G8; Covance Res. Products,
in the wells. Then each well was washed with 300 µl of

---

UNDER EMBARGO UNTIL JULY 6, 2009, 00:00 CET
686
C. Cao et al. / Caffeine Suppresses Amyloid-? Levels in Plasma and Brain of Alzheimer’s Disease Transgenic Mice
diluted wash buffer 3 times. After completing the last
in humans. Given the rapid delivery of caffeine to all
wash step, the bottom of the wells was wiped with a
body organs including the brain, we therefore hypothe-
lint-free towel to remove any liquid on the outside of the
sized that acute caffeine administration would quickly
wells. Then 150 µl of the K-Blue Substrate was added
suppress brain A? production in AD transgenic mice,
to each well with a multi-channel pipette. The plate was
resulting in a signi?cant reduction in plasma A? levels
then mixed by gently shaking, followed by incubation
within hours. In an initial study in which 3–4 month-
at room temperature for 5 to 20 minutes. To stop the
old APPsw mice (pre-plaque) were given a single i.p.
enzyme reaction, 50 µl of red stop solution was added
injection of caffeine (1.5 mg), plasma A?1?40 levels
to each well and gently mixed. The absorbance was
were signi?cantly reduced by 41% at 3 hours post-
then measured with plate reader (Synergy HT, Biotek,
treatment compared to pre-treatment levels (Fig. 2A);
VT) at a wavelength of 650 nm. The absorbance was
vehicle injection failed to affect plasma A? levels in
converted into concentration using Gen5 software.
other littermate APPsw mice. Caffeine-induced reduc-
Cytokine expression pro?les were detected using the
tions in plasma A?1?40 levels were also observed in
Bio-Plex kits (Bio-Rad, Richmond, CA, catalogue #
aged 14 month-old APPsw mice (plaque-bearing) at 3
171F11181). Samples and standards were prepared us-
hours following either a single i.p. or gavage treatment
ing company protocols with the initial concentration of
with the same dose of caffeine (Fig. 2B). Moreover,
standards ranging from 32 ng/ml to 1.95 pg/ml. Plas-
plasma caffeine concentrations in these same 14 month-
ma samples were prepared for analysis by diluting 1
old APPsw mice were negatively correlated with plas-
volume of the serum sample with 3 volumes of the Bio-
ma A?1?40 levels; higher plasma caffeine levels were
Plex mouse sample diluents. Details of this procedure
associated with lower plasma A? levels in individual
were performed by followed the protocol provided by
animals (Fig. 2C). Even aged 19-month APPsw+PS1
the manufacture. Finally, the plates were read. Each
mice (plaque-bearing) exhibited signi?cant reductions
cytokine level was calculated based on its own stan-
in plasma A?1?40 levels at 3 hours following gavage
dard curve. Brain cytokine and chemokine levels were
treatment with caffeine (Fig. 2D).
detected with the same method.
Long-term oral caffeine treatment provides a
Statistical analyses
sustained reduction in plasma A? levels
Group comparisons involving levels of soluble/de-
Our prior work has demonstrated that hippocampal
posited A?, caffeine, theophylline, and cytokines were
levels of both soluble and insoluble A? are reduced
performed using ANOVA. For determination of pre-
by greater than 30% following 51/2 months of oral caf-
treatment versus post-treatment differences in plasma
feine treatment (?1.5 mg/day) to young adult APP-
A? level, paired t-tests were employed. All other sta-
sw mice [9]. In the same study, we also found this
tistical analyses are as indicated in the text. In order to
same treatment to be effective in reducing insoluble
test if relationships were present between plasma, neu-
hippocampal A? levels in aged 17 month-old APPsw
rochemical, and behavioral measures, correlation anal-
mice following 18 days of treatment. We therefore hy-
ysis was performed using the Systat analytical software
pothesized that long-term oral caffeine treatment would
package.
result in a sustained suppression of plasma A? levels.
In fact, aged 15–20 month-old APPsw mice (plaque-
bearing) that were given oral treatment with 1.5 mg
RESULTS
caffeine twice daily for one week did exhibit signi?-
cant reductions in plasma levels of both A?1?40 and
Acute caffeine administration reduces plasma A? level
A?1?42 compared to pre-treatment values (Fig. 2E). At
9 days following cessation of this caffeine treatment,
We have previously found that A?1?40 and A?1?42
however, plasma levels of both A?1?40 and A?1?42
generation is decreased in N2a neuronal cell cultures
had returned to their pre-treatment levels. In a second
following 6 hours of caffeine treatment [9]. More-
long-term treatment study, 20 month-old APPsw+PS1
over, these reductions in generation of both A? 1?40
mice (plaque-bearing) were given oral caffeine treat-
and A?1?42 peptides were dose-dependent, occurring
ment (1.5 mg) twice daily every 4th day over a 2
at caffeine concentrations (
10 µM), which are typi-
month period. Periodic analysis of plasma A?1?40 and
cally present in plasma following coffee consumption
A?1?42 levels during this treatment period revealed not

---

UNDER EMBARGO UNTIL JULY 6, 2009, 00:00 CET
C. Cao et al. / Caffeine Suppresses Amyloid-? Levels in Plasma and Brain of Alzheimer’s Disease Transgenic Mice
687
Fig. 2. Caffeine treatment induces a rapid and sustained decrease in plasma A? levels. (A,B,D) A single i.p. or gavage treatment with caffeine
signi?cantly reduces plasma A? levels in 3–4 month-old (A) and 14 month-old (B) APPsw mice; as well as in 19 month-old APP+PS1 mice
(D). In all three studies, saline vehicle treatment had no effect. (C) Plasma caffeine levels in 14M old APPsw mice are inversely correlated
with plasma A?1?40 levels. (E) Oral caffeine treatment for one week to 15–20 month-old APPsw mice results in decreased plasma levels of
both A?1?40 and A?1?42 immediately following treatment, with a rebound in plasma A? levels occurring by 9 days after cessation of caffeine
treatment. Each group in A, B, D, and E consisted of 5–10 mice. (F) As exempli?ed by two aged 20 month-old APP+PS1 mice, oral caffeine
administration every 4th day over a 2-month period induces a sustained and continual decrease in plasma levels of both A?1?40 and A?1?42
over the treatment period. Post-treatment versus pre- or delayed post-treatment A? levels were evaluated by paired t-tests. ?p < 0.05–0.01
versus pre-treatment levels; ??p < 0.05 versus both pre-treatment and 9-day post-treatment levels. (Colours are visible in the electronic version
of the article at www.iospress.nl.)

---

UNDER EMBARGO UNTIL JULY 6, 2009, 00:00 CET
688
C. Cao et al. / Caffeine Suppresses Amyloid-? Levels in Plasma and Brain of Alzheimer’s Disease Transgenic Mice
Fig. 3. Caffeine lowers interstitial ?uid (ISF) A? levels. (A) Brain ISF A?x ?40 levels were measured by in vivo microdialysis. Prior to treatment
with caffeine, ISF A? levels ?uctuated very little. A low dose (5 mg/kg) and a higher dose (30 mg/kg) of caffeine caused ISF A? levels to
decrease signi?cantly compared to basal levels. (B) Following a 5 mg/kg dose of caffeine i.p., ISF A?x ?40 levels decreased by 19% compared to
basal levels (?p < 0.05), while a 30 mg/kg administration decreased A? levels by 33% (??p < 0.01, n = 6 per group). The mean concentration
of ISF A? represents an average of hours 2–3 after each dose of caffeine when A? levels had stabilized. (C) Tg2576 were pre-treated with vehicle
or 30 mg/kg caffeine for 3 hours (n = 6 per group), followed by administration of 100 nM Compound E, a ?-secretase inhibitor, via reverse
microdialysis. In both groups, ISF A? levels decreased rapidly when APP cleavage was blocked. (D) The elimination half-life of ISF A?x ?40
was comparable in vehicle-treated and caffeine-treated mice, suggesting that caffeine does not affect ISF A? elimination, but likely impacts A?
production instead. (Colours are visible in the electronic version of the article at www.iospress.nl.)
only a sustained reduction in both A? peptides, but al-
soluble within ISF. To determine if caffeine adminis-
so a continuing decrease in their levels throughout the
tration acutely affects brain ISF A? levels, 3 month-old
2-month treatment. Given that the half-life of caffeine
Tg2576 mice (pre-A? plaque) were treated with sev-
in rodents is only 0.7 to 1.2 hours [3], these later ob-
eral doses of caffeine during in vivo microdialysis to
servations suggest that the effects of caffeine on A?
measure ISF A? levels. Microdialysis probes were im-
suppression greatly exceed its own half-life.
planted into the hippocampus, permitting us to sample
ISF A? levels every 30 minutes for up to 24 hours in
Caffeine administration lowers brain interstitial ?uid
levels of A? in vivo
awake, behaving animals [14,15]. Basal ISF A? lev-
els were determined over 6 hours in each mouse fol-
A? is primarily produced in neurons and secreted
lowed by i.p. administration of caffeine at 5 mg/kg
into the brain extracellular space where it is normally
(?0.15 mg/mouse), then at 30 mg/kg (?1 mg/mouse)

---

UNDER EMBARGO UNTIL JULY 6, 2009, 00:00 CET
C. Cao et al. / Caffeine Suppresses Amyloid-? Levels in Plasma and Brain of Alzheimer’s Disease Transgenic Mice
689
Fig. 4. Long-term caffeine administration to aged, cognitively-impaired APPsw mice reduces soluble A? levels in both cortex and hippocampus,
while also decreasing deposited (insoluble) A? in hippocampus. Caffeine was administered to 18–19 month-old APPsw mice for two months in
their drinking water. Both brain A?1?40 and A?1?42 were decreased in caffeine-treated Tg mice. Although plasma A? levels were unaffected
for all Tg/Caff mice inclusively, see Fig. 7B. Immunohistochemical A? deposition in the hippocampus was reduced by 40% in caffeine treatment
mice. ?p < 0.05; ??p < 0.001. Each group consisted of 5–8 mice. (Colours are visible in the electronic version of the article at www.iospress.nl.)
3 hours later (Fig. 3A). Caffeine signi?cantly lowered
Long-term oral caffeine treatment to aged,
ISF A? levels by 19% and 32% at the low and high
cognitively-impaired APPsw mice reduces brain
doses, respectively, as compared to the basal ISF A?
soluble and deposited A?
levels in each mouse (Fig. 3B). There was also a trend
that A? levels were reduced to a greater extent by the
To determine the effects of chronic caffeine treat-
30 mg/kg dose of caffeine than by the 5 mg/kg dose
ment in cognitively-impaired AD mice, caffeine
(p = 0.1).
(?1.5 mg/day) was orally administered in drinking wa-
Caffeine can have many affects in a living animal
ter to 18–19 month-old APPsw mice (plaque-bearing)
which could potentially alter A? production or A?
that were con?rmed to be impaired in the RAWM task
elimination, thus lowering ISF A? levels [3,20]. As
of working memory prior to treatment. At 4–5 weeks
such, we determined the elimination half-life of ISF A?
into caffeine treatment, impaired APPsw mice that had
in mice treated with vehicle or 30 mg/kg caffeine. Basal
been given caffeine (Tg/Caff) exhibited substantial-
ISF A? levels were measured in Tg2576 mice, followed
ly better RAWM working memory performance com-
by i.p. administration of 30 mg/kg caffeine or vehicle
pared to the continuing impairment of control APPsw
(Fig. 3C). As expected, caffeine reduced ISF A? lev-
mice [12]. After euthanizia at 20–21 months of age (2
els by 35% compared to no change in vehicle-treated
months into caffeine treatment), soluble A?1?42 levels
mice. Three hours later, the microdialysis perfusion
in both cortex and hippocampus of Tg/Caff mice were
buffer was switched to contain a potent ?-secretase in-
signi?cantly reduced by 51% and 59%, respectively,
hibitor, Compound E, to rapidly block A? production.
compared to Tg controls (Fig. 4A). Cortical and hip-
In vehicle-treated mice, the elimination half-life of A?
pocampal A?1?40 levels were also reduced by chronic
was 1.5 hours (Fig. 3D), which is similar to previous re-
caffeine treatment, although the decrease in hippocam-
ports of the ISF A? half-life in this mouse model [21].
pal A?1?40 did not reach statistical signi?cance (p =
Importantly, the half-life of ISF A? in caffeine-treated
0.09). Compared to Tg controls, plasma A?1?40 lev-
mice (1.3 hours) was not signi?cantly different from
els were not signi?cantly decreased in Tg/Caff mice.
vehicle-treated mice. This suggests that caffeine does
However, when Tg/Caff mice were divided into two
not alter ISF A? elimination, but likely impacts some
sub-groups based on higher versus lower plasma caf-
aspect of A? production instead.
feine levels, a signi?cant decrease in plasma A?1?40

---

UNDER EMBARGO UNTIL JULY 6, 2009, 00:00 CET
690
C. Cao et al. / Caffeine Suppresses Amyloid-? Levels in Plasma and Brain of Alzheimer’s Disease Transgenic Mice
Fig. 5. Plasma A? levels do not correlate with brain A? levels in aged APPsw mice. (A,B) Strong correlations are present between levels of
A?1?40 and A?1?42 for both soluble and insoluble A? in hippocampus, irrespective of caffeine treatment. By contrast, no correlations existed
between hippocampus (or cortex) and plasma for soluble A? levels (C) or insoluble A? levels (D) irrespective of caffeine treatment. (Colours
are visible in the electronic version of the article at www.iospress.nl.)
was evident in higher plasma caffeine mice (see section
between A?1?40 and A?1?42 levels in hippocampus
below). Finally, and in the same behaviorally-tested
and cortex, and for both soluble and insoluble A? pools
aged Tg mice, chronic caffeine treatment resulted in a
(Fig. 5A and B). These correlations were present irre-
remarkable 40% reduction in hippocampal A? depo-
spective of whether all Tg mice, or only Tg controls
sition compared to Tg controls (Fig. 4B). Brain&n