Central Angiotensin II Stimulation Promotes β Amyloid Production in Sprague Dawley Rats

Background Stress and various stress hormones, including catecholamines and glucocorticoids, have recently been implicated in the pathogenesis of Alzheimer's disease (AD), which represents the greatest unresolved medical challenge in neurology. Angiotensin receptor blockers have shown benefits in AD and prone-to-AD animals. However, the mechanisms responsible for their efficacy remain unknown, and no studies have directly addressed the role of central angiotensin II (Ang II), a fundamental stress hormone, in the pathogenesis of AD. The present study focused on the role of central Ang II in amyloidogenesis, the critical process in AD neuropathology, and aimed to provide direct evidence for the role of this stress hormone in the pathogenesis of AD. Methodology/Principal Findings Increased central Ang II levels during stress response were modeled by intracerebroventricular (ICV) administration of graded doses of Ang II (6 ng/hr low dose, 60 ng/hr medium dose, and 600 ng/hr high dose, all delivered at a rate of 0.25 µl/hr) to male Sprague Dawley rats (280–310 g) via osmotic pumps. After 1 week of continuous Ang II infusion, the stimulation of Ang II type 1 receptors was accompanied by the modulation of amyloid precursor protein, α-, β-and γ-secretase, and increased β amyloid production. These effects could be completely abolished by concomitant ICV infusion of losartan, indicating that central Ang II played a causative role in these alterations. Conclusions/Significance Central Ang II is essential to the stress response, and the results of this study suggest that increased central Ang II levels play an important role in amyloidogenesis during stress, and that central Ang II-directed stress prevention and treatment might represent a novel anti-AD strategy.


Introduction
Alzheimer's disease (AD) is the most common neurodegenerative disorder worldwide [1]. It is characterized neuropathologically by the formation of senile plaques and neurofibrillary tangles, and clinically by the progressive deterioration of memory and other cognitive functions [2], which cannot be prevented using currently available treatments [3]. Senile plaques are primarily composed of a 40-42-amino acid peptide denoted as b amyloid (Ab), which is derived from sequential cleavage of amyloid precursor peptide (APP) by b-secretase and c-secretase (the amyloidogenesis pathway) [1].The accumulation and deposition of Ab in selective brain regions is a major cause of neurotoxicity and is assumed to be a culprit for inducing the pathologic processes of AD [4]. However, APP can also be processed by an a-c secretase stepwise pathway, which precludes the generation of Ab (the antiamyloidogenesis pathway).
Several years after the publication of descriptive reports of asecretase, three responsible enzymes, belonging to the A Disintegrin And Metalloproteinase (ADAM) family, have been identified with a-secretase activity: ADAM 9, ADAM 10, and ADAM 17 [5]. Moreover, a recent knockdown study confirmed that ADAM 10, but surprisingly not ADAM 9 or ADAM 17, was the physiologically relevant, constitutive a-secretase of APP [6]. The identity of b-secretase has also been studied, and a novel aspartyl protease, named the b-site APP-cleaving enzyme (BACE1), was revealed in five independent reports [7][8][9][10][11]. BACE1 cleavage of APP has subsequently been proven to be a prerequisite for Ab generation [12]. c-secretase is a membrane-embedded protease complex consisting of presenilin 1 (PS1), nicastrin, anterior pharynx-defective 1, and presenilin enhancer 2 [13]. csecretase specificity determines the ratio of two peptide products, Ab 40 and the more neurotoxic Ab 42 , and mutations in its catalytic subunit PS1, which lead to an increased Ab 42 /Ab 40 ratio, account for most cases of genetic AD [14]. However, because genetic AD accounts for only a small percentage (probably less than 1%) of the total number of cases [15], the development of sporadic AD is still linked to environmental influences. Stress, particularly chronic adverse stress, is one of the environmental factors that has been suggested to contribute to the onset and progression of AD [16]. The stress response is often associated with high levels of stress hormones, such as catecholamines (CAs, effectors of the sympathetic adrenomedullary system), glucocorticoids (GCs, effectors of the hypothalamic pituitary adrenal axis), and their regulator, central angiotensin II (Ang II) [17]. Both CAs [18,19] and GCs [20,21] are known to be linked to the pathogenesis of AD, but their regulator, central Ang II, deserves further consideration.
Central Ang II, by binding the Ang II type 1 receptor (AT 1 R) at all key hypothalamic regulatory centers or higher centers, plays a critical role in the activation of the sympathetic adrenomedullary system and hypothalamic pituitary adrenal axis, as well as in the secretion of CAs and GCs [17,22]. Central Ang II has thus been accepted as a fundamental stress hormone in the body [17,23]. However, central Ang II has also been implicated in several central nervous system disorders. For example, hippocampal Ang IIimmunopositive neurons with distorted processes were detected in Ab deposits in AD brains [24], and enhanced immunoreactivity of Ang II, AT 1 R, and angiotensin-converting enzyme (ACE) was found in a postmortem examination of patients with AD [25]. Moreover, some studies also showed that Ang II could inhibit potassium-mediated release of acetylcholine [26,27], induce oxidative stress [28], activation of mitogen-activated protein kinases (MAPKs) [29], and inflammation [30], and was involved in bloodbrain barrier maintenance [31] and cell survival [32], all of which are relevant to AD [32][33][34][35]. When administered to AD and proneto-AD animals, some Ang II receptor blockers (ARBs), such as losartan, valsartan, candesartan, telmisartan, and olmesartan, were associated with improved cognitive impairment and/or attenuated Ab production [36][37][38][39][40]. It is therefore vitally important to determine the potential role of Ang II in AD, especially in amyloidogenesis, given the proven involvement of stress and stress hormones in the pathogenesis of AD. The present study therefore used a rat model of AD to provide direct evidence for the role of Ang II in amyloidogenesis and the pathogenesis of the disease.

Results
Central Ang II Stimulation Caused a Dose-dependent Increase in AT 1 R expression AT 1 R mRNA and protein levels were assayed using quantitative polymerase chain reaction (qPCR) and western blotting (WB), respectively, after intracerebroventricular (ICV) administration of Ang II. A dose-dependent increase in AT 1 R mRNA levels was observed 1 week after central infusion of graded doses of Ang II. An almost 1.5-fold increase was observed following medium-dose Ang II stimulation, which was attenuated by combination with losartan. Losartan infusion alone had no effect on AT1R mRNA (p.0.05) ( Figure 1A). Similar changes in AT 1 R protein levels were identified by WB (Figure 2A and 2B).
These results suggest that central AT 1 R expression could be dose-dependently elevated by increased central Ang II, as confirmed by concomitant losartan administration, and that baseline AT 1 R expression was unaffected by ARBs, in the absence of stimulation by specific ligands.
Central Ang II Stimulation Downregulated Full-length APP Protein but Upregulated its mRNA Full-length APP protein levels were analyzed using a carboxylterminal-directed antibody that recognizes the carboxyl terminal fragments (CTFs) and the full-length APP holoproteins (mature APP or mAPP and immature APP or imAPP). The binding capacity of the antibody has been confirmed in previous studies [41][42][43]. WB showed a remarkable reduction in both mature and immature APP holoprotein levels in central Ang II-treated animals: about a 50% reduction with low-dose and medium-dose Ang II (p,0.01), and about a 70% reduction with high-dose Ang II (p,0.01), relative to saline-treated controls (Figure 2A and Figure 2C). APP transcription analyzed by qPCR showed an almost 2-fold increase in all Ang II-treated animals (p,0.01) ( Figure 1B). These effects were abolished by concomitant losartan treatment, and could therefore be attributed to Ang II.

Central Ang II Stimulation Had Different Effects on Secretase Expression
mRNA and protein levels of ADAM 10, BACE 1, and PS1, representing the expression levels of a-, b-, and c-secretase, were also evaluated. As shown in Figure 1C, Figure 2A and Figure 2D, ADAM 10 mRNA and protein remained at baseline levels after infusion of central Ang II or its blockers for 1 week(p.0.05), implying that central Ang II stimulation failed to affect ADAM 10 expression.
However, BACE1 mRNA, assessed by qPCR, was increased 1.3-fold following low-dose (p,0.01), and 1.6-fold following medium-dose Ang II treatment (p,0.01), and almost 2.1-fold following high-dose Ang II treatment (p,0.01) ( Figure 1C). BACE1 protein levels assessed by WB also showed dose-dependent increases in the Ang II-treated groups (Figure 2A and 2D). The marked increases in BACE1 mRNA and protein levels following medium-dose Ang II treatment were normalized by losartan(p.0.05), implying that these alterations were associated with central Ang II stimulation.
Regarding PS1 levels, low-dose Ang II infusion had no effect on PS1 gene expression (p.0.05), whereas medium-dose and highdose Ang II infusions were followed by a significant increase (p,0.01), and an even higher increase (p,0.01), respectively. These increases were cancelled by blockade of AT 1 R with losartan ( Figure 1C), suggesting that they were induced by stimulation of AT 1 R. PS1 protein expression analyzed by WB showed similar changes to those demonstrated for its gene expression.

Central Ang II Stimulation Had Different Effects on Secretase Activity
The activities of a, b-, and c-secretase were evaluated by measuring the fluorescence intensity after incubation with a synthetic fluorescent peptide substrate.
As shown in Figure 3, no significant alterations in a-secretase activity were shown in any Ang II-treated animals(p.0.05), suggesting that central Ang II had no effect on the activity of asecretase.
b-secretase activity was increased less than 1.4-fold following both low-dose and medium-dose Ang II(p,0.05), but by almost 2.5-fold by high-dose Ang II(p,0.01) ( Figure 3). Then, enhanced bsecretase activity was blocked by losartan and disappeared in single losartan, suggesting that the enhanced activity was due to Ang II.
The results of activity assay for c-secretase are shown in Figure 3. Treatment with graded doses of Ang II was followed by an approximately 1.6-1.8-2.9-fold stepwise increase in c-secretase activity, with no significant difference in activities between lowdose and medium-dose Ang II treatment (p.0.05). The increase following medium-dose Ang II treatment was normalized by concomitant losartan treatment (p.0.05), supporting the causative role of Ang II in these changes. Cleavage of APP by aor b-secretase results in the production of APP CTFs, i.e., C83 and C99, the latter of which is the direct substrate of c-secretase and immediate precursor of Ab [44]. As mentioned above, the antibody that recognizes full-length APP can also be used to detect its CTFs, C83 and C99.

Central Ang II Stimulation
When normalized to b-actin, WB of C83 showed an almost 1.3fold increase following medium-dose Ang II treatment (p,0.05), but no changes in response to other doses of Ang II (p.0.05) (Figure 2A and 2E). Simultaneous WB of C99 showed an approximately 2.0-fold increase following low-dose and medium-dose Ang II treatments (p,0.01), but a 1.8-fold increase following high-dose Ang II treatment (p,0.01) (Figure 2A and 2E). The productions of C83 and C99 were represented after normalization to both b-actin and APP. As shown in Figure 2F, when normalized to b-actin and then immature APP (imAPP), C83 production showed a 2.3-fold increase following low-dose and medium-dose Ang II (p,0.01), and a higher 3.3-fold increase following high-dose Ang II treatment (p,0.01). C99 production showed more significant increases, including a 4.3-fold increase in response to low-dose Ang II (p,0.01), a 3.5-fold increase following medium-dose Ang II (p,0.01), and a 5.3-fold increase following high-dose Ang II (p,0.01).
The concentration of Ab in the rat cerebral cortices was analyzed by sandwich ELISA. As shown in Figure 4A and Figure S1, low-dose Ang II treatment produced a 1.3-fold increase in Ab 40 concentration (from 30.2660.56 in controls to 40.3860.70, p,0.01), but a 1.5-fold increase in Ab 42 concentration (from 4.2360.17 in controls to 6.2060.14, p,0.01). When the dose of Ang II was increased by 10-fold, the increase in Ab 40 increased by almost 2-fold (to 59.9860.84, p,0.01), but no further significant increase in Ab 42 concentration occurred (6.3160.16). A further 10-fold increase in Ang II resulted in increases in Ab 40 and Ab 42 concentrations of nearly 2.5-fold (to 75.4861.89, p,0.01) and 3-fold (to 13.2060.32, p,0.01), respectively. These effects did not occur following the administration of concomitant or single losartan (p.0.05).
We also calculated the Ab 42 /Ab 40 ratios in the same samples. As shown in Figure 4B, the ratio was approximately 0.14060.007 in the controls animals, and this increased to 0.15460.005 following low-dose Ang II treatment (p,0.01), but decreased to 0.10560.003 following medium-dose Ang II treatment (p,0.01). After a rebound increase to 0.17560.008 following treatment with high-dose Ang II (p,0.01), the ratios  The effects of medium-dose Ang II on C83 and C99 levels, concentrations of Ab 40 and Ab 42 , and on the Ab 42 /Ab 40 ratios were attenuated or even abolished by blocking AT 1 R signaling, confirming that these effects resulted from central Ang II binding to AT 1 R.

Discussion
Ang II has multiple tasks in the brain and plays key roles in the regulation of blood pressure, fluid homeostasis, cerebral blood flow, and the response to stress [45]. These classical, well-defined actions of Ang II in the brain are mediated by stimulation of the AT 1 R [45]. Stress enhances the production of peripheral and brain Ang II [46][47][48] and the expression of peripheral and brain AT 1 R [46,[49][50][51]. Ang II does not readily cross the blood-brain barrier, and the present study therefore used ICV Ang II administration in rats to model the increased central Ang II levels occurring during the stress response. Both gene and protein levels of AT 1 R were dose-dependently upregulated after 1 week of ICV Ang II infusion, suggesting that self-upregulation of AT 1 R expression occurred after stimulation by Ang II, in accordance with the results of previous studies [52][53][54][55]. The regulation of AT 1 R expression and stimulation is important, as it determines the level of activation of the brain Ang II system [22].
APP is the source of Ab. In the present study, graded doses of central Ang II caused a 50-70% reduction in the level of APP protein, which did not correspond to the more significant increases in the production of C83, C99, and Ab. This discrepancy implies that APP expression and processing were enhanced by central Ang II, which was further supported by the upregulation of APP mRNA levels.
APP is predominately cleaved by a-secretase under physiologic conditions [56]. In the current study, central Ang II had no effect on ADAM 10 protein or mRNA levels or activity. However, although a-secretase expression was represented by the level of ADAM 10 in this study, other members of the ADAM family, including ADAM 9 and ADAM 17, have also been proposed to display a-secretase activities [5]. Even a close homolog of BACE1, BACE2, has been found to be responsible for a-secretase-like proteolytic cleavage of APP [57,58]. Thus the regulation of other putative a-secretase enzymes by Ang II should also be assayed.
BACE1 is the initiating and putatively rate-limiting enzyme in Ab generation [59]. Levels of BACE1 mRNA and protein were elevated by Ang II in the current study. Previous studies found that the BACE1 promoter contains a number of putative transcription factor-binding sites, including nuclear factor kappa-light-chainenhancer of activated B cells (NFkB) [60], activator protein 1 (AP 1) [61] and cAMP response element-binding protein (CREB) [61]. Activation of these transcription factors is involved in Ang II signaling [52,55,62], and this mechanism may therefore provide a basis for Ang II-induced BACE1 expression. Significant increases in b-secretase activity were also observed in this study. Oxidative stress has recently been identified as a factor involved in the stimulation of b-secretase activity [63,64], and is also well implicated in Ang II signaling [28]. The stimulation of b-secretase might thus result from Ang II-mediated oxidative stress. Several reports have reported BACE1 to be a stress-response protein, or bsecretase to be a stress-induced protease, because BACE1 expression and activity were elevated following acute or chronic stress [59]. This might suggest a potential link between stressinduced Ang II and altered BACE1 expression and activity.
The lack of a-secretase stimulation was expected to be reflected in a lack of C83 production, while significant increases in BACE1 expression and activity were expected to be associated with proportionate changes in C99 production. However, when normalized to b-actin, an almost 1.3-fold increase in C83 levels was detected following medium-dose Ang II administration, though low-dose and high-dose Ang II had no effects, and high- dose Ang II resulted in an almost 1.8-fold increase in C99 levels, while low-dose and medium-dose Ang II produced 2-fold increases. This inconsistency between C99 levels and significant alterations in BACE1 suggests that C99 was short-lived. On the other hand, the uncoupling of a-secretase activity and C83 levels might imply the existence of an additional mechanism for the production of C83. Recent studies have reported that C99 could be converted to C83 by a-secretase [6,65], which could account for the C83 levels observed in the current study. This possibility is further supported by the sharp increases in C83 levels and the opposite effects of medium-dose Ang II on C83 and C99, after normalization to both b-actin and APP.
PS1 is a key determinant of c-secretase activity [66]. We analyzed the expression of PS1 and found that stimulation of the AT 1 R by medium-dose or high-dose Ang II was accompanied by increased PS1 expression, though these changes were not observed after low-dose Ang II infusion. As a c-secretase catalytic subunit, increased PS1 expression has often been found to occur simultaneously with augmented c-secretase activity [67], while repression of PS1 transcription inhibited c-secretase mediated processing of APP [68]. However, overexpression of PS1 was not sufficient to generate increased c-secretase activity or produce more products in previous studies [69], suggesting that PS1 does not act alone. Given the complicated situation, c-secretase enzymatic activity in this study was evaluated in homogenates in toto, and equivalent enhancements were observed following lowdose and medium-dose Ang II treatment, with greater enhancement after high-dose treatment. Although the current study was unable to elucidate the precise mechanism of AT 1 R regulation of c-secretase, several lines of evidence have indicated possible mechanisms, as follows: central Ang II is well known to induce the activation of some transcription factors, such as AP 1 [52], NFkB [55] and CREB [62], which have been localized in PS1 promoter and implicated in PS1 gene expression [67,[70][71][72]. Ang II could also induce the activation of several MAPK members, which have been associated with enhanced c-secretase activity [73]. These results imply that activation of AT 1 R could promote PS1 expression through the activation of transcription factors, and could mediate the direct phosphorylation of PS1 or c-secretase at specific sites by targeting the MAPK pathway. Second, recent studies have shown that CAs can enhance c-secretase activity by inducing autophagy [18], and because central Ang II promotes CA utilization [74], this may represent an indirect pathway whereby central Ang II can regulate c-secretase activity. However, the study indicating that CAs promote amyloidogenesis by inducing autophagy was performed in vitro [18], and further research is needed to confirm these results in vivo.
Changes in the processing pathway of APP resulted in increased Ab production; a dose-dependent increase in secreted Ab 40 concentration, and a 1.5-3-fold increase in Ab 42 concentration were achieved following Ang II stimulation. Notably, the Ab 42 / Ab 40 ratios were altered. This ratio was linked to specificity of csecretase in previous studies [75,76], implying that Ang II could cause shifts in APP amyloidogenic processing. Furthermore, the neurotoxicity of Ab was recently shown to be associated with small changes in the Ab 42 /Ab 40 ratio [77], and Ang II-induced shifts in this ratio should thus be taken into serious consideration. However, losartan given together with medium-dose Ang II abolished all effects of Ang II on APP, the three secretases, and Ab. These findings suggest that the actions of central Ang II in amyloidogenesis are mediated by activation of the AT1R.
Hypertension has been identified as a strong risk factor for AD in epidemiological studies [78,79]. ICV infusion of Ang II in the present study might have resulted in an increase in blood pressure (BP) in the rats. Furthermore, previous studies in both animals and humans found that ARBs helped to preserve cognitive function through a mechanism independent of their antihypertensive effects [80][81][82][83]. The effects of systemic BP were outside the scope of the current study. However, AD has recently been proposed to be a vascular disease, rather than a neurodegenerative disease [84,85]. Because Ang II is a potent vasoactive peptide, stress-induced and Ang II-involved hypertension may contribute to amyloidogenesis. Thus further studies are needed to investigate the link between hypertension, especially nonpharmacologic hypertension, and amyloidogenesis.
In conclusion, the results of this study demonstrated that central Ang II potentiated amyloidogenesis in vivo, through modulation of various components of the APP processing pathway; these actions could be completely abolished by concomitant ICV infusion of losartan. This study not only provides direct etiological evidence for central Ang II, a fundamental stress hormone, in the pathogenesis of AD, but also suggests that ARBs, already used in clinical settings, may be potentially effective agents for the prevention and treatment of AD.

Reagents and preparation
Ang II (Sigma) was dissolved in sterile saline to 24 ng/ml, 240 ng/ml, and 2,400 ng/ml, as adapted from a recent study [86]. Losartan (Sigma) was added to 240 ng/ml Ang II to a concentration of 800 ng/ml, which was adequate to abolish the effects of 240 ng/ml Ang II in vivo, as verified in our preliminary experiments.
Osmotic pumps (Alzet, model 2004) with a brain infusion kit (Alzet) were filled with reagents or sterile saline and placed into sterile saline at 37uC for at least 40 hours, according to the manufacturer's instructions. The pumps were designed to work at a rate of 0.25 ml/hr. The actual doses of Ang II delivered into rat brains were thus 6 ng/hr, 60 ng/hr and 600 ng/hr.

Animals and surgery
Male Sprague Dawley rats (from the National Rodent Laboratory Animal Resources Shanghai branch of China) weighing 280-310 g were housed in a controlled environment at 2562uC with alternating 12 hr light and dark cycles. They were provided with food and water ad libitum. All rats were acclimatized in our animal facility for at least 1 week before the experiments were conducted. Animal care and experiment protocols were in accordance with the Guide for the Care and Use of Laboratory Animals of Nanjing Medical University and were approved by the Biological Research Ethics Committee of Nanjing Medical University (permit number: NYLL2009-0005). All surgery was performed under anesthesia and all efforts were made to minimize the suffering of the animals.
A total of 54 rats were randomly divided into six groups, according to the materials in the pumps: sterile saline; low-dose Ang II; medium-dose Ang II; high-dose Ang II; medium-dose Ang II + losartan; and losartan. Animals in each group were subjected to ICV administration via a pump for 7 days. Pump implantation was carried out using a well-established procedure in our department [87]. In brief, animals were anaesthetized by intraperitoneal injection of 10% chloral hydrate (0.35 ml/100 g). Anaesthetized rats were placed in a stereotactic frame (David Kopf), and the skull was exposed under sterile surgical conditions. Stainless steel cannulas from the brain infusion kit were implanted in the right lateral ventricle, relation to the bregma, at: 1.4 mm lateral and 0.8 mm posterior on the right. A right-sided burr-hole was drilled in the skull, 3.5 mm below the top of the cortex. The pumps were then installed in subcutaneous pockets on the lateral dorsum of the rats. Dental cement was used to affix the catheters to the skull, and the wounds were closed with sutures. Body temperature was monitored throughout the procedure using a rectal temperature probe and maintained at 3760.5uC with a thermostatically-controlled infrared lamp.
After 7 days, anaesthetized animals were sacrificed by decapitation and brains were immediately dissected on ice, frozen in liquid nitrogen, and maintained at 280uC until assayed.
Total RNA extraction, reverse transcription (RT) and qPCR assay of gene expression Cortices were homogenized in 10 volumes of TRIzol reagent (Invitrogen) and total RNA was extracted, according to the manufacturer's instructions. The quantity and quality of the resulting RNA were validated using a spectrophotometer (Hitachi). Of the total RNA, 500 ng was reverse transcribed to cDNA using a RT reagent kit (Takara) at 37uC for 15 minutes, followed by 85uC for 5 seconds, according to the manufacturer's instructions. The products were then stored at 220uC until use. A two-step qPCR was performed using an ABI 7500 Real-Time PCR System (Applied Biosystems). After 30 seconds of initial denaturation at 95uC, the thermal profile included 40 two-step cycles: step 1 for denaturation at 95uC for 5 seconds and step 2 for annealing and extension at 60uC for 34 seconds. Data were collected at step 2. The results were obtained using the comparative Ct method and the arithmetic formula 2 2DDCt .

Tissue extraction for protein-and peptide-directed assays
To extract total protein, cortices were placed in 10 volumes of ice cold RIPA buffer (50 mmol/l Tris-HCL, 150 mmol/l sodium chloride, 1% Triton X-100, 1% sodium deoxycholate, 0.1% sodium dodecyl sulfate) supplemented with protease inhibitors (Sigma). The tissue was homogenized for 10 s in an ultrasound homogenizer (IKA Laboratory). Tissue homogenates were then centrifuged for 15 min at 12,500 g at 4uC (Beckman). Supernatants were carefully collected.
For Ab concentration assays, tissue extracts were prepared as described previously [88], with some modifications. Cortical tissues were homogenized in 10 volumes of high-salt buffer (100 mmol/l Tris-HCL, pH 7.0, 1 mmol/l EGTA, 0.5 mmol/l magnesium sulfate , 750 mmol/l sodium chloride, 20 mmol/l sodium fluoride) containing protease inhibitors (Sigma). Soluble Ab was extracted using diethyl acetate 0.4% in 100 mmol/L sodium chloride and the sample was centrifuged for 60 min at 100,000 g at 4uC (Beckman). The supernatants were carefully collected and then neutralized by adding 500 mmol/l Tris-HCL, pH 6.8.
Protein quantification was performed using the supernatants, according to the Bradford method. All the resulting supernatants were frozen at -80uC until analysis.

Enzyme-linked immunosorbent assay (ELISA)
The levels of soluble Ab 40 and Ab 42 in rat cerebral cortices were detected using sandwich ELISA kits (Wako), according to the manufacturer's instructions. All samples were analyzed in duplicate.

Statistical analysis
Data were analyzed using analysis of variance (ANOVA), followed by least significant difference post hoc tests, using SPSS 13.0 software (SPSS). All results were expressed as mean6 standard deviation (SD). Figure S1 Concentration assays for Ab 40 and Ab 42 by ELISA. n = 6 for each group. All samples were analyzed in duplicate. *p,0.05 or **p,0.01 versus the control group receiving saline infusion. (TIF)