Oral Monosodium Glutamate Administration Causes Early Onset of Alzheimer’s Disease-Like Pathophysiology in APP/PS1 Mice

Abstract

Glutamate excitotoxicity has long been related to Alzheimer’s disease (AD) pathophysiology, and it has been shown to affect the major AD-related hallmarks, amyloid-β peptide (Aβ) accumulation and tau phosphorylation (p-tau). We investigated whether oral administration of monosodium glutamate (MSG) has effects in a murine model of AD, the double transgenic mice APP/PS1. We found that AD pathogenic factors appear earlier in APP/PS1 when supplemented with MSG, while wildtype mice were essentially not affected. Aβ and p-tau levels were increased in the hippocampus in young APP/PS1 animals upon MSG administration. This was correlated with increased Cdk5-p25 levels. Furthermore, in these mice, we observed a decrease in the AMPA receptor subunit GluA1 and they had impaired long-term potentiation. The Hebb-Williams Maze revealed that they had memory deficits. We show here for the first time that oral MSG supplementation can accelerate AD-like pathophysiology in a mouse model of AD.

Keywords

Alzheimer’s disease amyloid-β glutamate excitotoxicity long-term potentiation memory p-tau

INTRODUCTION

Alzheimer’s disease (AD) is an irreversible disease of progressive nature, which leads to deterioration in cognitive functions beyond what is normal in a healthy aging process. The neuropathological hallmarks of AD are extracellular plaques mainly formed by amyloid-β (Aβ) peptide, intracellular neurofibrillary tangles composed of hyperphosphorylated tau protein (p-tau) [1, 2], and an increase in intracellular Aβ which leads to synaptic dysfunctions and to neuronal loss (for a review, see [3]). Furthermore, growing evidence supports a key role for glutamate excitotoxicity in the pathogenesis of the disease [4]. We have recently shown that Aβ increases glutamate levels by inducing the inactivation of an E3 ubiquitin ligase, APC/C-Cdh1, and consequently the accumulation of one of its degradation targets, glutaminase, an enzyme that catalyzes glutamate synthesis from glutamine [5].

Glutamate transport across the blood-brain barrier (BBB) is strictly regulated [6]. However, in AD, impairment of the BBB occurs at early stages of the disease [7 –10], which may cause increased permeability of glutamate from blood plasma to the brain. Monosodium glutamate (MSG) is used in many culinary cultures to increase the flavor of different foods. It has been suggested a long time ago that a high dose of free glutamate could result in neurotoxicity, the named “Chinese Restaurant Syndrome” [11], which has led to controversy, since double blind studies did not show reproducible effects [12] and MSG is now generally considered safe. We asked ourselves whether elevated glutamate intake could lead to problems in a certain subpopulation, for example in patients with AD, due to the impairment of the BBB which occurs as a consequence of the disease [7 –10]. The effect of oral glutamate administration on AD-type neuropathology has not yet been sufficiently explored. The aim of this study was to test whether MSG supplementation affects AD pathophysiology in APP/PS1 mice.

We supplemented wildtype (WT) and APP/PS1 mice with 0.5% or 1% MSG in their drinking water, resulting in an equivalent dose of human intake of 4 or 8 g/day, respectively (see Discussion for detailed calculation). This was administered for a time period of 5 weeks, when they were 5–10 weeks old, and then the mice were kept for another 5–6 weeks in control conditions, in order to investigate long-term effects of the MSG supplementation. We report here that in MSG-supplemented mice, we detected higher Aβ and p-tau levels in the hippocampus, impaired long-term potentiation (LTP) and memory deficits at 4 months of age, while APP/PS1 mice drinking ordinary water did not show any of these signs at that age. Furthermore, we observed a decrease in AMPA receptor subunit GluA1, an accumulation of Cdk5 and p25 and decreased Cdh1 levels in the hippocampus. Importantly, WT mice were not affected by glutamate supplementation. Furthermore, it should be noted that this experimental model has different characteristics to the one in which postnatal MSG treatment is used to induce obesity [13, 14]. No changes in weight or glycemia occurred due to MSG administration in our experimental model. Our results suggest that glutamate intake affects AD pathophysiology in young APP/PS1 mice.

METHODS

Animals and treatment

APPswe, PSEN1dE9-85Dbo/J transgenic mice and WT mice from the same colony were used in this study. Mice were maintained under a 12 : 12 h dark-light cycle at 23±1°C and 60% relative humidity, and were provided with a standard chow diet (Panlab, S.L., Barcelona, Spain) and water ad libitum.

In total, 164 animals were used of which 81 were males and 83 were females. They were divided in the following groups: WT mice in control conditions (20 males and 16 females), APP/PS1 mice in control conditions (20 males and 17 females), WT mice supplemented with 0.5% MSG (10 males and 13 females), APP/PS1 mice supplemented with 0.5% MSG (7 males and 8 females), WT mice supplemented with 1% MSG (14 males and 16 females), and APP/PS1 mice supplemented with 1% MSG (10 males and 13 females). Animals in control groups were maintained in standard conditions with water. For the supplemented groups, MSG (Sigma Aldrich, St. Louis, MO, USA) was added at the indicated dose to the drinking water of mice when they were 5 weeks old. They had access ad libitum to the MSG-containing water during a period of 5 weeks and their consumption was monitored. MSG at a concentration of 1% and 0.5% resulted in an average daily uptake of 1.79 g/kg and 0.74 g/kg, respectively. After the supplementation, the animals were kept for 5 more weeks without MSG intake and then behavioral studies and LTP experiments were performed. The animals were sacrificed at 16 weeks of age, with the exception of animals used for cerebrospinal fluid (CSF) measurements, which were sacrificed during the time period of MSG supplementation. Their weight and glycemia levels were measured before they were sacrificed (see Fig. 1a for graphic timeline summary).

Fig.1

Water intake and glycemia in MSG-supplemented mice. Glutamate, GABA, and glutamine levels in the CSF. a) Timeline of experimental plan, ‘w’ indicates weeks of age of the animals. b) Water intake was monitored in a subset of cages of mice held in control conditions and in all cages of mice that received MSG administration. Results were normalized to averaged body weight of mice per cage. Mice supplemented with MSG 1% have significantly increased Water intake (n = 3 (Control), n = 9 (MSG 0.5%), n = 10 (MSG 1%); each n corresponds to one cage, control and transgenic mice were held as littermates). c) Glycemia was not significantly changed in mice when treated with MSG (results presented of mice of both genotypes pooled together). (n = 12 (Control), n = 16 (MSG 0.5%), n = 10 (MSG 1%). (d–f) Levels of glutamate (n = 7 (WT Control), n = 4 (APP/PS1 Control, n = 5 (WT MSG 1%), n = 3 (APP/PS1 MSG 1%), GABA (n = 5 (WT Control), n = 4 (APP/PS1 Control), n = 5 (WT MSG 1%), n = 3 (APP/PS1 MSG 1%) and glutamine (n = 7 (WT Control), n = 4 (APP/PS1 Control, n = 7 (WT MSG 1%), n = 3 (APP/PS1 MSG 1%) in CSF determined y UPLC-MS/MS.

Ethics statement

All treatments and methods were carried out in accordance with the approved guidelines. The use of animals and the experimental designs were approved by the ethics committee of the University of Valencia (reference: A1427886868839), following the animal care guidelines of the European Commission 2010/63/UE.

Evaluation of spatial working memory by Hebb-Williams Maze

A black plastic maze (60 cm wide×60 cm long×10 cm high) was used. It contained a start box and a goal box (both 14 cm wide×9 cm long), which are positioned at diagonally opposite corners. The maze contains cold water at a wading depth (15°C, 3.5 cm high), while the goal box is stocked with fresh dry tissue. Several maze designs are produced by fixing different arrangements of barriers to a clear plastic ceiling. This apparatus allows the cognitive process of routed learning and the motivation of water escape to be measured.

The experimental room was illuminated with a dim red light (40 lux at 1 m above floor level).

The procedure we followed was based on that employed by [15], in which mice must navigate the maze and cross from the wet start box to the dry goal box in order to escape the cold water. Animals underwent a 5 min habituation period (dry sand, no barriers) on day 1 and undertook problem A on day 2 and problem D on day 3 (4 trials/day) (practice mazes). Mice were subsequently submitted to mazes 1 and 5 (considered as easy and difficult mazes) on separate days on which 8 trials took place. The time limit for reaching the goal box was 5 min, after which the mouse was guided to the box. Total latency score (the sum of the latencies in all the problem trials in each maze) was measured. Mice unable to complete the task within the time limit scored maximum latencies (300 s). Following the Stanford and Brown classification [16], the mazes were defined as easy (1, 3, and 4) or difficult (5 and 8). The operator was blind to the genotype and treatment conditions.

Evaluation of anxiety by Elevated Plus Maze

The Elevated Plus Maze (EPM) consisted of two open arms (30×5×0.25 cm) and two enclosed arms (30×5×15 cm). The junction of the four arms formed a central platform (5×5 cm). The floor of the maze was made of black Plexiglas, and the walls of the enclosed arms of clear Plexiglas. The open arms had a small edge (0.25 cm) to provide additional grip for the animals. The entire apparatus was elevated 45 cm above floor level. In order to facilitate adaptation, mice were transported to the dimly illuminated laboratory 1 h prior to testing. At the beginning of each trial, subjects were placed on the central platform so that they were facing an open arm, and were allowed to explore for 5 min. The maze was thoroughly cleaned with a damp cloth after each trial. The behavior displayed by the mice was video-recorded and later analyzed by a ‘blind’ observer using a computerized method. The measurements recorded during the test period were frequency of entries and time and percentage of time spent in each section of the apparatus (open arms, closed arms, central platform). An arm was considered to have been visited when the animal placed all four paws on it. Number of open arm entries, time spent in open arms and percentage of open arm entries are generally used to characterize the anxiolytic effects of drugs [17, 18].

Long-term potentiation assessment

The groups of WT animals and transgenic animals were divided in each of the experimental conditions: control group (WT, n = 3; APP/PS1, n = 3), MSG-0.5% group (WT, n = 4; APP/PS1 n = 4), and MSG-1% group (WT, n = 4; APP/PS1 n = 4).

LTP was measured in 22 urethane (1.7 g/Kg) anesthetized mice. Once the animals exhibited loss of plantar and head reflexes, 0.1 ml of 5% lidocaine was subcutaneously injected in the throat area to perform a tracheotomy as described in [19]. When the breathing was stabilized, the animals were placed on a stereotaxic frame (David Kopf Instruments, Tujunga, USA).

Evoked field excitatory postsynaptic potentials (fEPSPs) were recorded from the dentate gyrus in the hippocampus, and elicited by the electrical stimulation (constant voltage) of the perforant path. Stainless-steel electrodes (World Precision Instruments; Aston, Stevenage, UK) were made in a monopolar (field recording) or bipolar (stimulation) disposition. The electrodes were placed at the coordinates following the landmark method described by [20]: dentate gyrus (A: –2.3 mm; L: 1.0 mm; H: 2.2 mm) and perforant path fibers (A: –4.3 mm; L: 2.5 mm; H: 1.7 mm) [21]. The position of the stimulation electrode was optimized to give maximal amplitude of the evoked potential visualized in the dentate gyrus. The calibration of the stimulation amplitude was performed 30 min after the electrode implantation.

fEPSPs was evoked by electrical pulses (0.1 ms, 0.1 Hz, 0.1–0.7 mA) in the perforant path. The amplitude selected was the minimal intensity of stimulation necessary to elicit a reliable fEPSP, with a response of 30% of its maximum. This procedure minimized the variability in fEPSP amplitudes within and between animals. To this end, input–output curves were produced by stimulating initially at 0 V and after increasing intensities in 0.1 mA steps until a maximal fEPSP.

The animal was allowed to stabilize for 30 min before the start of the recording session. The protocol included 30 min of baseline conditions (electrical pulses of 0.1 ms, 0.1 Hz) followed by the LTP induction (high-frequency stimulation, HFS) and 90 min of post-LTP basal conditions. The HFS consisted of 2 sets of electrical trains separated 5 min between them, each set of trains made-up by 6 trains with an intertrain interval of 20 s. Each train was formed by six pulses of 0.4 ms and 400 Hz. All data were digitalized at 16 KHz using the Micro 1401 MKII acquisition board and the Spike2 software (Cambridge Electronic Design, Ltd.).

The amplitude of the evoked potential was used as a measure of the stimulation-related synaptic strength. Each time-point was calculated as the percentage of the average of the recordings (prestimulation basal synaptic transmission values) for every subject individually. These measurements were performed in Matlab software (R2015b, MathWorks, Inc., Natick, MA, USA).

Extraction of cerebrospinal fluid

For CSF collection, mice were anesthetized with gaseous anesthesia (Isoflurane + O₂). Induction was performed using 3.5% v/v Isoflurane (IsoFlo^®, Provesa, Murcia, Spain), and the anesthesia was maintained using a mouse inhalation mask with 2.5% v/v isoflurane. When mice were deeply anesthetized, they were placed in a stereotaxic frame (Kopf Instruments, Tujunga, USA). Each mouse head was secured with the nose tilted downwards and the back of the head protruding. When mice were fixed, skin and muscles of the neck area were carefully removed until cisterna magna was exposed. CSF was collected using a glass capillary of 0.75 mm in inner diameter and 1.0 mm in outer diameter. With the tip of the glass capillary (10–20 μm of tip inner diameter), dura of cisterna magna was perforated and CFS aspirated using a syringe connected to the glass capillary, until approximately 10 μl of CSF was obtained.

Determination of neurotransmitters in CSF by UPLC-MS/MS

The quantitation of glutamate, glutamine and γ-aminobutyric (GABA) in CSF was determined by Ultra Performance Liquid Chromatography–tandem Mass Spectrometry (UPLC-MS/MS). The CSF was centrifuged at 10,000 g, 5 min at 4°C. The supernatants were diluted with CH3CN:H2O (70 : 30) and the internal standard (IS), Phe-D5, solution (10 mmol/L) were added to aliquots. Finally, it was injected in the chromatographic system (UPLC-MS/MS). The chromatographic system used consisted of a Waters Acquity UPLC-XevoTQ system (Milford, MA, USA). The conditions employed were: positive electrospray ionization (ESI), capillary voltage 3.50 kV, extractor 5.00 V, source temperature 120°C, desolvation temperature 350°C, nitrogen cone and desolvation gas flows were 50 and 750 L/h, respectively. Separation conditions were selected to achieve appropriate chromatographic retention and resolution by using an HILIC column (100×2.1 mm, 7 μm, 100 Å) from Phenomenex. Mobile phase used was CH₃OH (5 mM NH₄HCO₂): H₂O (5 mM NH₄HCO₂), (70 : 30) with isocratic gradient 10 min. The flow rate, column temperature and injection volume were set at 0.4 mL/min, 30°, 5 μL respectively. During batch analysis, samples were kept at 4°C in the autosampler. Mass spectrometric detection was carried out by multiple reaction monitoring (MRM), employing the acquisition parameters summarized in Table 1. Data were acquired and processed using MassLynx 4.1 and QuanLynx 4.1softwares (Waters), respectively. Linear response curves were calculated employing Phe-D5.

Table 1

MS/MS acquisition parameters

Analyte	Parent ion (m/z)	Cone (V)	Daughter ion (m/z)	Confirmation	Collision energy (eV)
Glutamate	148	20	84	102	15
Glutamine	147	10	84	130	10
GABA	104	20	69	87	10
Phe-D5	171.1	35	125	–	–

Fixation and tissue preparation

Animals were anaesthetized with a lethal dose of sodium pentobarbital (100 mg/kg 20%) (Dolethal Vetoquinol Madrid, Spain) and transcardially perfused with heparinized saline (0.1%, pH 7.0) followed by paraformaldehyde (4%) in phosphate buffer (PBS) (0.1 M, pH 7.4). Brains were post-fixed for 24 h in paraformaldehyde (4%) in PBS at 4°C and maintained for 2 days in sucrose (30%) at 4°C for cryoprotection. Coronal sections of 40 μm were obtained by freezing microtomy (Leica, Wetzlar, Germany) and stored in phosphate buffer (0.1 M, pH 7.4) at 4°C.

Immunofluorescence labelling and imaging of brain tissue

Free floating sections were washed in PBS with a Triton solution (0.5%) (PBS-T), 3 times at room temperature (RT). For blocking, the sections were incubated for 2 h at RT in blocking solution, 10% goat serum (Sigma-Aldrich, G9023) in PBS-T. Afterwards the tissue was incubated overnight at 4°C, in blocking solution with the primary antibodies: Mouse monoclonal amyloid beta antibody (MOAB) (NBP2-13075, 1 : 500, Novusbio, Littleton, Colorado, USA), AMPA GluA1 (13185, 1 : 200, Cell Signalling Tech, Danvers, Massachusetts, USA) or p-tau (MN1020, 5–10 μg/ml, Thermo Fisher Waltham, Massachusetts, USA).

Afterwards, sections were rinsed 3 times in PBS-Tx for each 10 min at RT. Incubation with secondary antibodies: Alexa Fluor 488 anti-rabbit (1 : 1000, Cell Signaling Tech) and Alexa Fluor 647 anti-mouse (1 : 1000, Cell Signaling Tech) in PBS-T, for 1.5 h at RT were used. Finally, sections were rinsed 3 times in PBS-T for each 10 min at RT, and incubated with Hoechst for 15 min (33342, Thermo Fisher) and washed again for 10 min PBS-T, mounted onto pig skin gelatine-coated slides (0.5%) and covered in fluorescence mounting medium (S3023, DAKO, Santa Clara, CA, USA).

The immunohistochemical preparations were visualized with a confocal laser-scanning microscope (Leica TCS SP2 scanning multiphoton and confocal unit with an inverted DM1RB microscope; Ar-He-Ne). For each primary antibody, slices from at least 3 animals were used (see figure legends for n numbers for each staining). From each brain, we obtained 5-6 series of brain slices (each one 40 μm) covering the area of the hippocampus (approx. 8–10 slices per series). For quantification, 6 images were taken from at least 3 slices per staining and per condition, which were then averaged for each animal. The images taken covered the areas of stratum oriens, stratum pyramidale, and stratum radiatum of CA1. Care was taken that each image covered all three strata. The signal threshold was adjusted for the WT control condition and all images of the different conditions from the same set of staining were taken in one session with the same exposure settings. For image acquisition, the operator of the microscope was blind to the treatment and genotype of the animals. Fluorescent signal levels of the specific antibody were normalized to fluorescent signal of nuclei of the hippocampal CA1 regions. Analyses were performed using ImageJ software.

Determination of Aβ concentration in brain tissue homogenates

The concentration of Aβ of dissolved plaques was measured in tissue homogenates of cortices and hippocampi using a human Aβ₄₂ ELISA kit (Invitrogen KHB3544, Camarillo CA, USA) for APP/PS1 mice and Aβ₄₂ mouse ELISA Kit (Invitrogen KMB3441) for WT mice. Tissues were dissected and immediately freeze-clamped and stored at –80°C. They were homogenized in 5 M guanidine-HCl in 50 mM Tris and slowly rotated for 4 h at RT in order to dissolve plaques containing Aβ. Further sample preparation and detection was performed according to the manufacturer’s instructions.

Immunoblot analysis in brain tissue homogenates

Immunoblot analysis were carried out as described previously [5]. Shortly, brain tissues were dissected and immediately freeze-clamped. They were homogenized in lysis buffer (Tris: 76.5 mM; pH: 6.8; SDS: 2%; Glycerol: 10%; supplemented with sodium ortovanadate (2 mM) and protease inhibitor (Sigma-Aldrich)), boiled for 5 min and stored for analysis at –80°C. The same amounts of protein were loaded onto an SDS polyacrylamide gel and blotted on a nitrocellulose membrane. Membranes were incubated overnight at 4°C with appropriate primary antibodies: Cdh1 (DCS-266, 1 : 500, Novus Biologicals), Cdk5 (2506, 1 : 1000, Cell Signaling Tech), p35/25 (2680, 1 : 1000, Cell Signaling Tech), α-Tubulin (T5168, 1 : 16000, Sigma Aldrich). Membranes were incubated with corresponding secondary antibodies and signal detection was performed using Luminata Classico Western HRP Substrate WBLUC0500 (Millipore Corporation, Billerica, MA, USA).

Statistical analysis of behavior tests

For behavioral analysis, data relating to the Elevated Plus Maze were analyzed by a two-way ANOVA, with two between-subject variables. Treatment with three levels (control, MSG 0.5% and MSG 1%); and genetics, with two levels (WT and APP/PS1). The data of the Hebb-Williams Maze were analyzed by a three-way ANOVA with the same two between-subject variables treatment and genetics; and one within-subject variable maze, with 2 levels: Maze 1 (easy maze) and Maze 5 (difficult maze). Bonferroni adjustment was employed for post hoc comparisons. All results are expressed as mean±SEM.

Statistical analysis of LTP

Data were tested for normality using the Shapiro-Wilk’s test, a robust method in the case of small sample sizes. In addition, we checked the equality of variances using the Levene’s test. In both tests, a significance level of 0.05 was set for the rejection of the null hypothesis. Because all data samples showed normality (p > 0.05) and homoscedasticity (p > 0.05), they were analyzed for statistical significance using parametric tests. Student’s t test was used to assess the significance level where indicated. The graphs were constructed from the mean value of the percentages±SEM. Between-group statistics was done by ANOVA by comparing the whole set of measurements after the stimulation between groups. Pairwise t-test with Bonferroni’s correction was applied for post-hoc multiple comparisons when appropriate. The statistical analysis was performed in R package (R Core Team, 2016).

Statistical analysis of biochemical assays

The results are expressed as mean values of the data of at least three independent experiments; the error bars represent SEM. We used one-tailed t-tests and Mann Whitney U-test analysis for immunohistochemical analysis and western blots.

Statistical analysis of gender interaction

Since mice of both genders were used in this study, we analyzed whether gender had an interaction effect on the results by adding the gender factor to each ANOVA model. In order to control for the gender interaction over the treatment effects, we have carried out a 3-Way ANOVA including the gender factor. The aim has been to consider a model that supersedes most of the statistical analyses carried out throughout the study. In particular, we have confirmed that the 2-Way (MSG*gender and genotype*gender) and the 3-Way (MSG*genotype*gender) interaction terms involving the gender factor were not significant, while at the same time the treatment effects reported in the study were not distorted by gender segregation.

p-values were ranked in the units of significance, which were set as following: p > 0.05 not significant (ns), p < 0.05 significant (*), p < 0.01 very significant (**), p < 0.001 highly significant (***) for all statistical tests.

RESULTS

MSG administration in WT and APP/PS1 has no effects on body weight and glycemia but on glutamate and GABA levels in the CSF

During a period of five weeks, mice were supplemented with MSG and the amount of water intake was monitored. Then they were kept for five more weeks in control conditions and their body weight and glycemia levels were measured. The experimental timeline is shown in Fig. 1a. The water intake of mice was monitored and resulted in an average dose of MSG of 1.79 g/kg/day for the 1% MSG administration group and 0.74 g/kg/day for the 0.5% MSG administration group (Fig. 1b).

MSG treatment in neonatal mice is used as a model for obesity [19], and we assessed here whether the treatment we applied caused changes in weight. The average weight of the animals upon different treatment conditions (for males, controls: 26.6 ± 1.5 g, 0.5% MSG: 29.6 ± 1.9 g and 1% MSG: 25.4 ± 2.5 g; and for females, controls: 20.7 ± 1.4 g, 0.5% MSG: 20.6 ± 0.6 g and 1% MSG: 21.1 ± 1.5 g) showed no significant differences (mice of both genotypes pooled together). Postnatal glutamate treatment has also been reported to affect glycemia [20]. The glutamate supplementation in our study did not cause any significant changes in glycemia (controls 117.8 ± 23.3 mg/dL, MSG 0.5% 129.7 ± 26.9 mg/dL, MSG 1% ±115.5± 13.9 mg/dL) (Fig. 1c).

In order to test whether MSG supplementation leads to detectable changes in the CSF during the supplementation period, we administrated 1% MSG to WT and APP/PS1 mice for two weeks. CSF was extracted from these mice and glutamate, GABA, and glutamine levels were measured using UPLC-MS/MS. We could not detect an overall significant change in the glutamate level between animals with or without MSG supplementation. There was, however, a mild but significant difference between WT and APP/PS1 that were supplemented with MSG (p = 0.046). Interestingly, we observed significantly decreased GABA levels in APP/PS1 mice that were supplemented with 1% glutamate compared to APP/PS1 in control conditions (p = 0.011) and compared to WT mice in control condition (p = 0.0084). This results in a ten-fold higher glutamate/GABA ratio in APP/PS1 when supplemented with MSG compared to WT mice. Glutamine levels were similar in all groups (Fig. 1d–f).

Amyloid-β increases in MSG-supplemented APP/PS1 mice

First, we wanted to assess whether MSG administration caused changes in senile plaques extracted from cortex and hippocampus in APP/PS1 mice (16 weeks old) using a human Aβ₄₂ ELISA kit. We observed an increase in both brain areas, which was more pronounced at 1% MSG supplementation than at 0.5%. Only the increase in the hippocampus upon 1% MSG was statistically significant (t(9) = –0.957, p = 0.048). We then measured endogenous Aβ levels of mice using ELISA in cortex and hippocampus samples from WT animals of controls, 0.5% and 1% MSG supplementation groups. There was no significant difference between the groups (Fig. 2a, b).

Fig.2

Amyloid-β increases in the hippocampus in APP/PS1 mice upon supplementation with 1% MSG. a) Aβ concentration from dissolved plaques from tissue homogenates of cortex and hippocampus were measured using Aβ₄₂ ELISA kits. There is a significant increase in Aβ in the hippocampus in APP/PS1 when treated with MSG (1%) compared to APP/PS1 mice in control conditions (p < 0.05) (n = 3). There was no significant difference in WT animals among the different groups (n = 4). b) Representative images of hippocampal slices (40 μm) stained with MOAB (red) antibody for Aβ and nuclear Hoechst staining (blue). c) Results shown as mean±SEM. There is a significant increase in Aβ in the hippocampus in APP/PS1 mice when treated with MSG at 1% compared to other experimental groups (APP/PS1 Control and APP/PS1 MSG 1% (p < 0.05); WT MSG 1% and APP/PS1 MSG 1% (p < 0.05) (n = 3), (scale bar = 75 μm).

Next, we carried out immunohistochemical analysis of Aβ₄₀ and Aβ₄₂ using a MOAB-2 antibody. We prepared brain slices from WT and APP/PS1 mice in control conditions and upon MSG administration. There was no significant difference between WT and APP/PS1 mice in control conditions at 16 weeks of age. However, APP/PS1 mice showed a significant increase in Aβ upon administration of 1% MSG, compared to APP/PS1 mice in control conditions (t(4) = –1.346, p = 0.027). There was no increase in Aβ in WT mice when supplemented with MSG (Fig. 2c, d).

We carried out co-stainings of MOAB with the endothelial marker von Willebrand, which indicates that some of the Aβ accumulates around blood vessels in the APP/PS1, the area where dietary supplemented MSG likely first enters the brain (Supplementary Figure 1). Specificity of MOAB antibody was confirmed with no-primary antibody control staining (Supplementary Figure 2).

P-tau, Cdk5, p25, Cdh1, and glutaminase are altered after MSG administration in APP/PS1

Tau phosphorylation is another main hallmark of AD pathology. To assess whether phosphorylation of tau was increased by MSG administration in WT and APP/PS1 mice, we carried out immunohistochemical stainings of brain slices using a tau antibody that recognizes the phosphorylation sites Ser202/Thr205. These are known to be targeted by the AD-related kinase Cdk5-p25 [24]. We detected increased levels of p-tau in 1% MSG treated APP/PS1 mice compared to APP/PS1 in control conditions (t(4) = –8.830, p = 0.001) and compared to APP/PS1 mice treated with 0.5% MSG (t(4) = –6.808, p = 0.002). P-tau levels were significantly higher in APP/PS1 than in WT mice with 1% MSG (t(4) = –4.286, p = 0.013). WT mice with 1% MSG also showed a significant increase compared to WT mice in control conditions (t(4) = –3.114, p = 0.036) (Fig. 3a, b). Specificity of p-tau antibody was confirmed with a no-primary antibody control staining (Supplementary Figure 2).

Fig.3

P-tau levels are increased in the hippocampus in mice supplemented with MSG. a, b) Representative images of hippocampal slices (40 μm) stained with p-tau (red) antibody and nuclear Hoechst (blue). Results shown as mean±SEM. There were significant differences in p-tau in the hippocampus in APP/PS1 mice with 1% MSG compared to other groups (scale bar = 200 μm) (n = 3). c–f) Western blot analysis of p-Tau and tau levels in WT and APP/PS1 mice in control conditions or upon 1% MSG supplementation. Images were quantified by densitometry and normalized against α-tubulin; the mean values±SEM are indicated (n = 3).

Western blot analysis using the same p-tau antibody also showed significantly increased p-tau levels in mice supplemented with MSG 1% (Fig. 3c, d). There was no significant difference of total tau levels between the groups (Fig. 3e, f).

To test whether phosphorylation of tau was accompanied by changes in Cdk5 and p25/35 protein levels, we performed western blot analysis. Brain tissue homogenates were prepared from WT and APP/PS1 mice in control conditions and with 1% MSG. Cdk5 levels were significantly higher in cortex homogenates of supplemented APP/PS1 mice compared to non-supplemented and compared to WT mice with MSG respectively (t(4) = –5.876, p = 0.004 and t(4) = –3.841, p = 0.018) (Fig. 4a). In the same line, APP/PS1 mice administered with MSG had significantly higher p25 levels (t(4) = –3.083, p = 0.037) than WT mice upon the same treatment (Fig. 4b).

Fig.4

Cdk5, p25/p35, Cdh1, and glutaminase are altered in cortex and hippocampus in mice supplemented with MSG. a–e) Representative western blot images of Cdk5 (a), p25/p35 (b), and glutaminase (e) in homogenates of cortex, and Cdh1 in homogenates of cortex (c) and hippocampus (d) of WT and APP/PS1 mice in control conditions and upon MSG supplementation (1%). Samples of 3 mice in each group were loaded on western blots and quantified by densitometry and normalized against α-tubulin; the mean values±SEM are indicated (n = 3).

Another target of Cdk5 is Cdh1, which is the activator of the E3 ubiquitin ligase APC/C. When glutamate excitotoxicity occurs, Cdk5 phosphorylates Cdh1 and causes its degradation [25]. We detected decreased levels of Cdh1 in homogenates of cortex and hippocampus from APP/PS1 mice supplemented with MSG. In cortex, the decrease was statistically significant in supplemented APP/PS1 mice compared to WT mice (t(4) = 3.410, p = 0.027) and compared to non-supplemented APP/PS1 mice (t(4) = 2.681, p = 0.055). In the hippocampus, the differences were significant between APP/PS1 and WT mice both with MSG (t(4) = 4.951, p = 0.008) (Fig. 4c, d). There was a significant increase in glutaminase, a degradation target of the ubiquitin ligase APC/C-Cdh1 in APP/PS1 mice supplemented with MSG compared to APP/PS1 mice in control conditions (t(4) = –2.902, p = 0.044) and WT mice with MSG (t(4) = –3.958, p = 0.017) (Fig. 4e).

Impaired LTP in APP/PS1 mice upon MSG administration

LTP experiments in WT and APP/PS1 transgenic mice was performed to assess the effect of the glutamate intake in the hippocampal fEPSPs (control group (WT, n = 3; APP/PS1, n = 3), MSG-0.5% group (WT, n = 4; APP/PS1, n = 4), and MSG-1% group (WT, n = 4; APP/PS1, n = 4)). Basal and post-LTP evoked responses were significantly different between APP/PS1 unsupplemented animals and those treated with MSG (0.5 and 1%) (Fig. 5). As expected, in the WT group, a clear potentiation of the synaptic response was observed in the control subgroup (151.18±1.39%; t(7) = –17.66, p = 1.13e–7). Besides, oral administration of MSG induced a significantly larger LTP (MSG 0.5% : 183.38±1.25%, t(7) = –27.61, p = 1.57e–8; MSG 1% : 177.13±1.21%, t(5) = –14.79, p = 1.09e–5) as compared with control conditions (F(2) = 177; p = 2e–16). Pairwise comparisons indicated that both doses had a different effect in the LTP potentiation (p < 0.01). In contrast, the APP/PS1 group showed different responses in all situations after HFS (F(2) = 278.1; p = 2e–16).

Fig.5

Assessment of hippocampal LTP in WT and APP/PS1 mice. a) Group data of the time course and the grand average of the normalized evoked potential amplitude of the 30 min before (blue dots) and the 90 min after (red dots) the LTP paradigm (t = 0), for each of the six experimental groups. b) Significant changes in the evoked potential amplitude were found in each experimental group after the HFS. Nevertheless, potentiation in the APP/PS1 MSG groups were significantly lower than the observed potentiation in the APP/PS1 control group as well as in the WT groups.

Specifically, the different glutamate doses led to a significant increase of the evoked potential (control: 159.68±2.26%, t(9) = –14.26, p = 9.55e–8; MSG 0.5% : 121.02±1.18%, t(21) = –15.58, p = 2.38e–13; MSG 1% : 109.95±0.92%, t(10) = –4.32, p = 0.001), although with lower amplitudes when compared with the control subgroup (p < 0.01).

Similar fEPSP amplitude values were observed between both doses (p = 0.42). Overall, the results indicate a lower potentiation at the dentate gyrus synapse in the LTP mechanism with the intake of MSG in the APP/PS1 mice.

AMPA receptor subunit GluA1 is decreased in APP/PS1 mice supplemented with MSG

We then carried out immunohistochemical analysis of the AMPA receptor subunit GluA1, whose implication in LTP has been extensively studied [26]. We prepared brain slices of WT and APP/PS1 mice that were treated with MSG or held under control conditions and labelled them with a specific GluA1 antibody. Area measurements of the fluorescent signal of GluA1 normalized to nuclear staining revealed a significant decrease of the receptor subunit in APP/PS1 mice with 1% MSG compared to APP/PS1 mice in control conditions (t(4) = 1.01 p = 0.056) or WT upon MSG supplementation (t(4) = 1.37, p = 0.020) (Fig. 6a, b).

Fig.6

AMPA receptor subunit GluA1 decreases in APP/PS1 mice when treated with MSG (1%). a) Representative images of hippocampal slices (40 μm) stained with a GluA1 antibody (green) and nuclear Hoechst staining (blue). b) Mean datapoints±SEM is shown and reveals a significant decrease of GluA1 in APP/PS1 mice upon MSG supplementation (1%) compared to APP/PS1 mice in control conditions (p < 0.05) and WT upon MSG supplementation (1%) (p < 0.05) (n = 3).

APP/PS1 mice exhibit memory impairments when treated with 1% MSG

Elevated Plus Maze

APP/PS1 mice supplemented with 1% MSG show an increase in the time in open arms (TOA), indicating a decrease in anxiety (anxiolytic effect). The ANOVA of the EPM data (Fig. 7a) revealed a significant effect of the interaction Treatment*Genetics [F(2,63) = 3.116; p < 0.05] and [F(2,63) = 3.275; p < 0,05], for the time and the % of time spent in the open arms of the maze. Both the absolute time as well as the percentage of absolute time spent in the open arms of the maze of APP/PS1 mice supplemented with 1% MSG significantly increased compared to the time spent by APP/PS1 mice in control conditions (p = 0.021 for TOA and p = 0.032 for % TOA) and WT MSG 1% (p = 0.05 for TOA and p = 0.044 for % TOA).

Fig.7

APP/PS1 mice supplemented with MSG (1%) have memory deficits. a) Effects of MSG in APP/PS1 on anxiety in the Elevated Plus Maze. Time in open arms (TOA) and % of time spent in the open arms (% TOE), entries in open arms (OE) and the total number of entries (TE). Data are presented as mean values±SEM *p < 0.05 compared to WT 1%; +p < 0.05 compared to APP/PS1 control. b, c) Effects on the total latency score to reach the goal in the 8 trials of the Hebb-Williams Maze. The mazes were classified as easy (mean of mazes 1, 3, and 4) or difficult (mean of mazes 5 and 8). APP/PS1 mice supplemented with 1% MSG show an increase in the total latency score to reach the goal, indicating impaired hippocampal spatial working memory. Data are presented as mean values±SEM ***p < 0.001 with respect WT control and ⁺⁺p < 0.01 compared to APP/PS1 Control (p = 0.004) and APP/PS1 0.5% MSG (p = 0.010).

We analyzed the number of total and open entries during the 5 min test (this measure is commonly employed to evaluate motor activity) and we found that there were no significant differences in locomotor activity between the APP/PS1 mice supplemented with 1% of MSG and their corresponding control group, nor with WT-C mice, as we show here with the number of entries in open arms (OE) and the total number of entries (TE). However, WT-0.5% mice show a significant decrease in the total number of entries with respect to the corresponding WT-C group (p < 0.01).

Hebb-Williams Maze

The Hebb-Williams Maze, which reveals hippocampal learning, indicates that administration of 1% MSG increases the time to reach the goal in the difficult mazes by APP/PS1 mice [F(2,66) = 4.252; p = 0.018] (Fig. 7b, c). As with biochemical parameters, no effect was found at 0.5% MSG. In the easy mazes, some increase was observed, and even if there was a clear tendency, the results were not statistically significant.

The results, which confirm the previous ones at the molecular level, indicate that there is a threshold between 0.5% and 1% MSG administration in terms of toxicity at both the molecular and the functional, cognitive level. In all the results we report here 0.5% MSG does not result in significant toxicity, but 1% results in clear deterioration of the cognitive health of the AD mouse model.

DISCUSSION

The aim of this study was to examine whether oral monosodium glutamate (MSG) supplementation affects AD-related pathological factors in APP/PS1 mice. Experiments were performed in relatively young animals (16 weeks), when the AD-like phenotype is still not apparent in this transgenic mouse model. We show here that oral MSG administration induces an early onset of AD pathophysiology in APP/PS1 mice. This effect is dose dependent. MSG was administrated for 5 weeks, and in mice supplemented with 1% MSG in drinking water, significant changes were detected 5 weeks after the treatment was finished. The experimental timeline includes a five-week gap between supplementation and sacrifice to ensure we do not measure short term effects of early exposure to MSG but rather observe long term effects of chronic MSG in the context of AD-related pathology.

MSG has been used in many culinary cultures to increase the flavor of different foods. Increasing palatability may be useful for elderly patients with decreased appetite [27, 28]. However, it has been suggested a long time ago that a high dose of free glutamate could result in neurotoxicity, the named “Chinese Restaurant Syndrome” [11], which has led to controversy, as double-blind studies did not show reproducible effects [12].

We asked ourselves whether elevated glutamate intake could lead to problems in a certain subpopulation, for example in patients with AD, already at prodromal phases when patients may be unaware of the disease, because access of glutamate to brain may be increased due to the impairment of the BBB which occurs as a consequence of the disease [7 –10]. Results that are reported here indicate that MSG at relatively high (but attainable) doses can result in molecular alterations in brains of APP/PS1 mice, while it leaves WT mice relatively unaffected.

Previous reports indicate that persons in the European Union ingest approximately half a gram of MSG per day but that in Asia this can go as high as 1.7 g/day [29]. However, individual glutamate consumption from food additives shows broad variations; high consumers in Asian countries may reach 4 g/day. In some Asian restaurants, up to 5 grams could be ingested in one single meal [30]. Furthermore, naturally containing glutamate-rich products may contribute to a rise in blood glutamate levels as well.

We calculated the human equivalent dose in mice by using the formula:

human equivalent dose=animal dose [mg/Kg]×(animal:K_m/human K_m) in [mg/Kg], where mouse K_m factor is 3 and human K_m factor is 37, as reported in [31]. The lower dose we used (0.5% in drinking water) is equivalent to the intake of high consumers in Asian countries; 4 g/day [29]. Thus, the doses we have used are in a similar range to the high level of human consumption but by no means are unattainable in humans.

We report here that MSG intake leads to Aβ increase in the hippocampus in young APP/PS1 mice. We observed accumulation of Aβ in plaques-like structures but also in areas surrounding blood vessels. Although the Aβ pathology is most recognized for its accumulation in parenchymal amyloid plaques, its parallel deposition in the cerebral vasculature is commonly observed in brains of patients with AD [32]. Studies in experimental mouse models suggest that even small amounts of cerebral microvascular amyloid deposition that precede high parenchymal levels of Aβ, are sufficient to drive early cognitive impairment [33]. Notably, Aβ also accumulates in the cerebrovasculature area in our experimental model, which is where dietary supplemented glutamate first reaches the brain.

It has been reported previously that neuronal activation can affect processing or vulnerability to Aβ [34 –36]. Interestingly, chronic activation of extrasynaptic, but not synaptic NMDARs, promotes neuronal Aβ release and even mild alterations in glutamatergic transmission could increase the production of Aβ [37 –39]. We hypothesize that MSG supplementation may have enhanced neuronal activity in APP/PS1 mice, which could contribute to the Aβ burden. In turn, it has also been shown that Aβ increases sensitivity of neurons to glutamate, enhances Ca²⁺ influx [40], glutamate release [41], and glutamate production via glutaminase accumulation [5], and impairs glutamate uptake by astrocytes [42, 43]. Thus, Aβ and glutamate toxicity may enter a positive feedback loop enhancing AD pathophysiology.

Hyperphosphorylation of tau has been reported to be present at 8 months of age in the APP/PS1 mouse model [44]. We report here that phosphorylation of tau was significantly increased in APP/PS1 when animals were supplemented with 1% MSG at 4 months of age. Tau phosphorylation also increased in WT animals upon 1% MSG. The APP/PS1 mouse model is generally not considered to promote the conversion of endogenous tau into neurofibrillary structures. We did not examine the preparations for neurofibrillary tangles, but it would be interesting to investigate whether tauopathy occurs in mice supplemented with MSG 1% in future studies.

We then aimed to identify whether changes in signaling molecules occurred that could link glutamate-induced alterations and tau phosphorylation.

Cdk5 and its activator p35 or its truncated form p25, are key molecules in neurodegeneration [45]. Prolonged expression of p25 impairs synaptic plasticity and causes neuronal loss, leading to severe cognitive defects [46]. Interestingly, it has also been reported recently that over-activation of p25 initially increases synaptic size and enhances Aβ production and later results in elimination of synapses. Hence, it has been suggested that the dysregulation of p25 is involved in initiating early AD synaptic pathology [47]. P35 and p25 are regulated by the Ca²⁺-dependent protease calpain, which is driven by glutamate mediated signaling. Notably, it has been reported that Cdk5-p25 can lead to abnormal phosphorylation of tau at the sites Ser202/Thr205 [24], which are the ones where we detected an increase.

We tested here whether Cdk5 and p35/p25 levels were changed in MSG-administered mice and we observed a significant increase in Cdk5 and p25 in APP/PS1 mice supplemented with MSG, thus linking glutamate-mediated Ca²⁺ signaling with tau phosphorylation.

Furthermore, it has been reported that Cdk5 signals the degradation of Cdh1, the activator of the ubiquitin ligase APC/C, implicated in AD. We previously showed that Aβ or glutamate causes a decrease in Cdh1 via Cdk5-p25 signaling, leading to an accumulation of the downstream targets of APC/C-Cdh1, of which several have been shown to accumulate in AD brains [5, 25]. Here we observed a decrease in Cdh1 in APP/PS1 mice that were supplemented with MSG, underpinning a role for APC/C-Cdh1 in AD pathophysiology.

Finally, we report here that MSG supplementation leads to memory-associated impairment in young transgenic mice. We detected deficits in LTP and decreased AMPA receptor subunit GluA1, which was correlated with impaired performance in a hippocampal spatial working memory test in these mice. Impaired LTP and cognitive deficits have been reported in APP/PS1 mice, but only at 8 months of age [44, 49]. Furthermore, it has been previously shown that Aβ drives loss of AMPA receptors from the surface in hippocampal neurons [50, 51], which we observed in our experiments when APP/PS1 received 1% MSG.

The Hebb-Williams Maze, according to Rabinovith and Rosvold (1951), is thought to be a type of “intelligence test” for examining problem-solving ability and general cognitive performance in rodents [52]. Our previous experience with this test showed that it is capable of distinguishing between easy and difficult learnings [53, 54] and therefore it is a useful tool to detect cognitive deficits. We also based our experiments on previous literature, which reports that hippocampal lesions are reflected in a poor performance in the Hebb-Williams Maze [55, 15]. We cannot rule out the possibility that other areas such as the prefrontal cortex could contribute to differences that were observed, since previous studies have shown poor performance in the Hebb-Williams Maze also after damage in this area [56].

Furthermore, some studies have reported that APP/PS1 animals show a decreased level of anxiety in Elevated Plus Maze [57]. Onaolapo et al. have recently demonstrated that monosodium glutamate has an anxiolytic effect [58, 59]. Therefore, in the present work, although we found no differences in anxiety between WT and APP/PS1 mice, we could observe a potentiation of anxiolysis in APP/PS1 mice by glutamate, but only with the highest dose.

In summary, MSG administration in APP/PS1 leads to accumulation of Aβ and p-tau, decrease in the AMPA receptor subunit GluA1, accumulation of Cdk5-p25 and impaired LTP, which correlates with cognitive deficits in these mice. Our results demonstrate that mice prone to develop AD are susceptible to neuropathological alterations as a result of oral MSG administration.

However, the mechanism of how dietary MSG enters and affects AD pathophysiology remains unclear. In most regions of the brain, the uptake of glutamate from the circulation is limited by the BBB. While the glutamate concentration in the plasma can reach 50–100 μmol/L, the concentration in extracellular brain fluids is maintained relatively low 0.5–2 μmol/L, which is essential for optimal brain function [6]. It has been reported in APP/PS1 mice that alterations of BBB permeability have been observed [60 –63]. Therefore, entrance of increased plasma-glutamate through the BBB may be causing the effects in the transgenic but not in WT animals. When measuring the glutamate concentration in CSF in MSG-supplemented animals, we observed a significant increase in APP/PS1 mice compared to WT mice, indicating more glutamate may have entered through the BBB in APP/PS1. However, glutamate levels were not significantly different between APP/PS1 in control conditions and APP/PS1 mice supplemented with 1% MSG. APP/PS1 transgenic animals may have alterations in glutamate signaling, transport or uptake, as previously described in AD [5 , 65]. Therefore, it is possible that—if APP/PS1 animals in control conditions already have some alterations in glutamate regulation—that the administration of MSG does not alter glutamate CSF levels in a detectable range. Another possible explanation could be that dynamic and rapid glutamate uptake from the extracellular space by astrocytes [66] did not allow the detection of a difference between those two groups. Interestingly, APP/PS1 and WT mice that received 1% MSG, had decreased levels of GABA compared to WT or APP/PS1 in control conditions. Taken these results together, APP/PS1 mice with MSG had a 5–10-fold higher glutamate/GABA ratio compared to all other groups. This suggests that MSG supplementation in APP/PS1 mice may have caused an excitation/inhibition imbalance leading to hyperexcitability, described as an early phenomenon in AD [67].

We have previously shown that extracellular application of glutamate in the hippocampus in APP/PS1 mice in vivo increases glutaminase levels, due to a failure in the APC/C-Cdh1 ubiquitination pathway [5]. This pathway could further contribute to elevated glutamate levels in the CSF in these mice. There was no change in CSF glutamine levels in these mice, showing that MSG supplementation does not affect extracellular CSF glutamine levels. However, more glutamine may be metabolized inside astrocytes if they have taken up more glutamate.

Our findings suggest that it would be interesting to assess whether AD, mild cognitive impairment patients, or even individuals at risk of developing AD (like carriers of the APOE ɛ4/ɛ4) should avoid diets very rich in glutamate. It has been reported that if the mean onset of AD was delayed by 5 years, this would reduce the number of individuals suffering from AD by half [68]. We conclude that the results of our study suggest a potential role of dietary glutamate in aggravating AD pathophysiology.

Footnotes

ACKNOWLEDGMENTS

This work was supported by grants SAF2010-19498, from the Spanish Ministry of Education and Science (MEC); ISCIII2012-RED-43-029 from the “Red Tematica de investigacion cooperativa en envejecimiento y fragilidad” (RETICEF); PROMETEO2010/074 from “Conselleria d’Educació, Cultura i Esport de la Generalitat Valenciana”; AICO/2016/078 Consolidable Grant from GVA and EU Funded CM1001 and FRAILOMIC-HEALTH.2012.2.1.1-2; Intramural Grant from INCLIVA and EU Funded CM100.

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

Alzheimer

(1907) Über eine eigenartige Erkrankung der Hirnrinde. Allg Zschr Psychiatr Psych-gerichtl Med 64, 146–148.

Buée

, Bussière

, Buée-Scherrer

, Delacourte

, Hof

(2000) Tau protein isoforms, phosphorylation and role in neurodegenerative disorders. Brain Res Brain Res Rev 33, 95–130.

Selkoe

(2001) Alzheimer’s disease: Genes, proteins, and therapy. Physiol Rev 81, 741–766.

Hynd

, Scott

, Dodd

(2004) Glutamate-mediated excitotoxicity and neurodegeneration in Alzheimer’s disease. Neurochem Int 45, 583–595.

Fuchsberger

, Martínez-Bellver

, Giraldo

, Teruel-Martí

, Lloret

, Viña

(2016) Aβ induces excitotoxicity mediated by APC/C-Cdh1 depletion that can be prevented by glutaminase inhibition promoting neuronal survival. Sci Rep 6, 31158.

Hawkins

(2009) The blood-brain barrier and glutamate. Am J Clin Nutr 90, 867S–874S.

Skoog

, Wallin

, Fredman

, Hesse

, Aevarsson

, Karlsson

, Gottfries

, Blennow

(1998) A population study on blood–brain barrier function in 85-year-olds: Relation to Alzheimer’s disease and vascular dementia. Neurology 50, 966–971.

Ujiie

, Dickstein

, Carlow

, Jefferies

(2003) Blood–brain barrier permeability precedes senile plaque formation in an Alzheimer disease model. Microcirculation 10, 463–470.

Montagne

, Barnes

, Sweeney

, Halliday

, Sagare

, Zhao

, Toga

, Jacobs

, Liu

, Amezcua

, Harrington

, Chui

, Law

, Zlokovic

(2015) Blood-brain barrier breakdown in the aging human hippocampus. Neuron 85, 296–302.

10.

Zenaro

, Piacentino

, Constantin

(2017) The blood-brain barrier in Alzheimer’s disease. Neurobiol Dis 107, 41–56.

11.

Schaumburg

, Byck

, Gerstl

, Mashman

(1969) Monosodium L-glutamate: Its pharmacology and role in the Chinese Restaurant Syndrome. Science 163, 826–828.

12.

Morselli

, Garattini

(1970) Monosodium glutamate and the Chinese restaurant syndrome. Nature 227, 611–612.

13.

Olney

(1969) Brain lesions, obesity, and other disturbances in mice treated with monosodium glutamate. Science 164, 719–721.

14.

Miranda

, da Silva Franco

, de Oliveira

, Barella

, Tófolo

, Ribeiro

, Pavanello

, da Conceição

, Torrezan

, Armitage

, Lisboa

, de Moura

, de Freitas Mathias

, Vieira

(2017) Cross-fostering reduces obesity induced by early exposure to monosodium glutamate in male rats. Endocrine 55, 101–112.

15.

Galsworthy

, Paya-Cano

, Liu

, Monleón

, Gregoryan

, Fernandes

, Schalkwyk

, Plomin

(2005) Assessing reliability, heritability and general cognitive ability in a battery of cognitive tasks for laboratory mice. Behav Genet 35, 675–692.

16.

Stanfor

, Brown

(2003) MHC-congenic mice (C57BL/6 J and B6-H-2 K) show differences in speed but not accuracy in learning the Hebb-Williams Maze. Behav Brain Res 144, 187–197.

17.

Pellow

, File

(1986) Anxiolytic and anxiogenic drug effects on exploratory activity in an elevated plus-maze: A novel test of anxiety in the rat. Pharmacol Biochem Behav 24, 525–529.

18.

Rodgers

, Cutler

, Jackson

(1997) Behavioural effects in mice of subchronic chlordiazepoxide, maprotiline and fluvoxamine. II. The elevated plus-maze. Pharmacol Biochem Behav 57, 127–136.

19.

Moldestad

, Karlsen

, Molden

, Storm

(2009) Tracheotomy improves experiment success rate in mice during urethane anesthesia and stereotaxic surgery. J Neurosci Methods 176, 57–62.

20.

Huang

, Yang

, Hu

, Liu

, Zhou

, Zhang

(2013) A new approach to location of the dentate gyrus and perforant path in rats/mice by landmarks on the skull. Acta Neurobiol Exp (Wars) 72, 468–472.

21.

Paxinos

, Franklin

KBJ

(2001) The Mouse Brain in Stereotaxic Coordinates. 2nd Edition, Academic Press, San Diego.

22.

Sasaki-Hamada

, Hojo

, Koyama

, Otsuka H and Oka

(2015) Changes in hippocampal synaptic functions and protein expression in monosodium glutamate-treated obese mice during development of glucose intolerance. Eur J Neurosci 41, 1393–1401.

23.

Hosaka

, Kusano

, Zai

, Kawada

, Kuribayashi

, Shimoyama

, Nagoshi

, Maeda

, Kawamura

, Mori

(2012) Monosodium glutamate stimulates secretion of glucagon-like peptide-1 and reduces postprandial glucose after a lipid-containing meal. Aliment Pharmacol Ther 36, 895–903.

24.

Hashiguchi

, Saito

, Hisanaga

, Hashiguchi

(2002) Truncation of CDK5 activator p35 induces intensive phosphorylation of Ser202/Thr205 of human tau. J Biol Chem 277, 44525–44530.

25.

Maestre

, Delgado-Esteban

, Gomez-Sanchez

, Bolaños

, Almeida

(2008) Cdk5 phosphorylates Cdh1 and modulates cyclin B1 stability in excitotoxicity. EMBO J 20, 2736–2745.

26.

Nicoll

(2017) A brief history of long-term potentiation. Neuron 93, 281–290.

27.

Schiffman

(2000) Intensification of sensory properties of foods for the elderly. J Nutr 130, 927S–930S.

28.

Mojet

, Heidema

, Christ-Hazelhof

(2003) Taste perception with age: Generic or specific losses in supra-threshold intensities of five taste qualities? Chem Senses 28, 397–413.

29.

Beyreuther

, Biesalski

, Fernstrom

, Grimm

, Hammes

, Heinemann

, Kempski

, Stehle

, Steinhart

, Walker

(2007) Consensus meeting: Monosodium glutamate – an update. Eur J Clin Nutr 61, 304–313.

30.

Yang

, Drouin

, Herbert

, Mao

, Karsh

(1997) The monosodium glutamate symptom complex: Assessment in a double-blind, placebo-controlled, randomized study. J Allergy Clin Immunol 996, 757–762.

31.

Reagan-Shaw

, Nihal

, Ahmad

(2008) Dose translation from animal to human studies revisited. FASEB J 22, 659–661.

32.

Jellinger

(2002) Alzheimer’s disease and cerebrovascular pathology: An update. J Neural Transm 109, 813–836.

33.

, Xu

, Anderson

, Kotarba

, Davis

, Robinson

, Van Nostrand

(2014) Cerebral microvascular rather than parenchymal amyloid ß-protein pathology promotes early cognitive impairment in transgenic mice. J Alzheimers Dis 38, 621–632.

34.

Bero

, Yan

, Roh

, Cirrito

, Stewart

, Raichle

, Lee

, Holtzman

(2011) Neuronal activity regulates the regional vulnerability to amyloid-β deposition. Nat Neurosci 14, 750–756.

35.

Nitsch

, Farber

, Growdont JH and Wurtman

(1993) Release of amyloid 18-protein precursor derivatives by electrical depolarization of rat hippocampal slices. Proc Natl Acad Sci U S A 90, 5191–5193.

36.

Nitsch

, Growdon

(1994) Role of neurotransmission in the regulation of Amyloid p-protein precursor processing. Biochem Pharmacol 47, 1275–1284.

37.

Lesné

, Ali

, Gabriel

, Croci

, MacKenzie

, Glabe

, Plotkine

, Marchand-Verrecchia

, Vivien

, Buisson

(2005) NMDA receptor activation inhibits α-secretase and promotes neuronal amyloid-β production. J Neurosci 25, 9367–9377.

38.

Hoey

, Williams

, Perkinton

(2009) Synaptic NMDA receptor activation stimulates alpha-secretase amyloid precursor protein processing and inhibits amyloid-beta production. J Neurosci 29, 4442–4460.

39.

Bordji

, Becerril-Ortega

, Nicole

, Buisson

(2010) Activation of extrasynaptic, but not synaptic, NMDA receptors modifies amyloid precursor protein expression pattern and increases amyloid-ß production. J Neurosci 30, 15927–15942.

40.

Brorson

, Bindokas

, Iwama

, Marcuccilli

, Chisholm

, Miller

(1995) The Ca²⁺ influx induced by beta-amyloid peptide 25-35 in cultured hippocampal neurons results from network excitation. J Neurobiol 26, 325–338.

41.

Bobich

, Zheng

, Campbell

(2004) Incubation of nerve endings with a physiological concentration of Aβ_1-42 activates CaV2.2(N-Type)-voltage operated calcium channels and acutely increases glutamate and noradrenaline release. J Alzheimers Dis 6, 243–255.

42.

Lauderback

, Harris-White

, Wang

, Pedigo

Jr , Carney

, Butterfield

(1999) Amyloid β-peptide inhibits Na⁺-dependent glutamate uptake. Life Sci 65, 1977–1981.

43.

Scimemi

, Meabon

, Woltjer

, Sullivan

, Diamond

, Cook

(2013) Amyloid-β slows clearance of synaptically released glutamate by mislocalizing astrocytic GLT-1. J Neurosci 33, 5312–5318.

44.

Radde

, Bolmont

, Kaeser

, Coomaraswamy

, Lindau

, Stoltze

, Calhoun

, Jäggi

, Wolburg

, Gengler

, Haass

, Ghetti

, Czech

, Hölscher

, Mathews

, Jucker

(2006) Aβ42-driven cerebral amyloidosis in transgenic mice reveals early and robust pathology. EMBO Rep 7, 940–946.

45.

Patrick

, Zukerberg

, Nikolic

, de la Monte

, Dikkes

, Tsai

(1999) Conversion of p35 to p25 deregulates Cdk5 activity and promotes neurodegeneration. Nature 402, 615–622.

46.

Fischer

, Sananbenesi

, Pang

, Lu

, Tsai

(2005) Opposing roles of transient and prolonged expression of p25 in synaptic plasticity and hippocampus-dependent memory. Neuron 48, 825–838.

47.

Sheng

, Zhang

, Su

, Tsai

, Julius Zhu

(2016) Cdk5 is a new rapid synaptic homeostasis regulator capable of initiating the early Alzheimer-like pathology. Cereb Cortex 26, 2937–2951.

48.

Kimura

, Ishiguro

, Hisanaga

(2014) Physiological and pathological phosphorylation of tau by Cdk5. Front Mol Neurosci 7, 65.

49.

Gengler

, Hamilton

, Hölscher

(2010) Synaptic plasticity in the hippocampus of a APP/PS1 mouse model of Alzheimer’s disease is impaired in old but not young mice. PLoS One 5, e9764.

50.

Hsieh

, Boehm

, Sato

, Iwatsubo

, Tomita

, Sisodia

, Malinow

(2006) AMPAR removal underlies Aβ-induced synaptic depression and dendritic spine loss. Neuron 52, 831–843.

51.

Rui

, Gu

, Yu

, Hartzell

, Zheng

(2010) Inhibition of AMPA receptor trafficking at hippocampal synapses by beta-amyloid oligomers: The mitochondrial contribution. Mol Brain 3, 10.

52.

Rabinovitch

, Rosvold

(1951) A closed-field intelligence test for rats. Can J Psychol 5, 122–128.

53.

Vidal-Infer

, Aguilar

, Miñarro

, Rodríguez-Arias

(2012) Effect of intermittent exposure to ethanol and MDMA during adolescence on learning and memory in adult mice. Behav Brain Funct 8, 32.

54.

Bonet-Costa

, Herranz-Perez

, Blanco-Gandia

, Mas-Bargues

, Ingles

, Garcia-Tarraga

, Rodríguez-Arias

, Miñarro

, Borras

, García-Verdugo

, Viña

(2016) Clearing amyloid-β through PPAR γ/ApoE activation by genistein is a treatment of experimental Alzheimer’s disease. J Alzheimers Dis 51, 701–711.

55.

Kimble

(1963) The effects of bilateral hippocampal lesions in rats. J Comp Physiol Psychol 56, 273–283.

56.

Winocur

, Moscovitch

(1990) Hippocampal and prefrontal cortex contributions to learning and memory: Analysis of lesion and aging effects on maze learning in rats. Behav Neurosci 104, 544–551.

57.

Lok

, Zhao

, Zhang

, He

, Shen

, Wang

, Yin

(2013) Effects of accelerated senescence on learning and memory, locomotion and anxiety-like behavior in APP/PS1 mouse model of Alzheimer’s disease. J Neurol Sci 335, 145–154.

58.

Onaolapo

, Odetunde

, Akintola

, Ogundeji

, Ajao

, Obelawo

, Onaolapo

(2019) Dietary composition modulates impact of food-added monosodium glutamate on behaviour, metabolic status and cerebral cortical morphology in mice. Biomed Pharmacother 109, 417–428.

59.

Onaolapo

, Aremu

, Onaolapo

(2017) Monosodium glutamate-associated alterations in open field, anxiety-related and conditioned place preference behaviours in mice. Naunyn Schmiedebergs Arch Pharmacol 390, 677–689.

60.

Poduslo

, Curran

, Wengenack

, Malester

, Duff

(2001) Permeability of proteins at the blood-brain barrier in the normal adult mouse and double transgenic mouse model of Alzheimer’s disease. Neurobiol Dis 8, 555–67.

61.

, Alata

, Dodacki

, Traversy

, Chacun

, Pradier

, Scherrmann

, Farinotti

, Calon

, Bourasset

(2014) Altered cerebral vascular volumes and solute transport at the blood-brain barriers of two transgenic mouse models of Alzheimer’s disease. Neuropharmacology 81, 311–317.

62.

Minogue

, Jones

, Kelly

, McDonald

, Connor

, Lynch

(2014) Age-associated dysregulation of microglial activation is coupled with enhanced blood-brain barrier permeability and pathology in APP/PS1 mice. Neurobiol Aging 35, 1442–1452.

63.

Ahn

, Learman

, Dunbar

, Maiti

, Jang

, Cha

, Song

(2018) Characterization of impaired cerebrovascular structure in APP/PS1 mouse brains. Neuroscience 10, 246–254.

64.

, Mallory

, Alford

, Tanaka

, Masliah

(1997) Glutamate transporter alterations in Alzheimer disease are possibly associated with abnormal APP expression. J Neuropathol Exp Neurol 56, 901–11.

65.

Madeira

, Vargas-Lopes

, Brandão

, Reis

, Laks

, Panizzutti

, Ferreira

(2018) Elevated glutamate and glutamine levels in the cerebrospinal fluid of patients with probable Alzheimer’s disease and depression. Front Psychiatry 9, 561.

66.

Rose

, Felix

, Zeug

, Dietrich

, Reiner

, Henneberger

(2017) Astroglial glutamate signaling and uptake in the hippocampus. . Front Mol Neurosci 10, 451.

67.

Palop

, Mucke

(2016) Network abnormalities and interneuron dysfunction in Alzheimer disease. Nat Rev Neurosci 17, 777–792.

68.

Brookmeyer

, Gray

, Kawas

(1998) Projections of Alzheimer’s disease in the United States and the public health impact of delaying disease onset. Am J Public Health 88, 1337–1342.