Spatial Training Attenuates Long-Term Alzheimer’s Disease-Related Pathogenic Processes in APP/PS1 Mice

Animals

The male APP/PS1 mice were chosen for experiments after genotyping. The representative data of genotyping was shown in Supplementary Figure 1. APP/PS1 transgenic mice (2.5 months old) were randomly divided into two groups as follows: 1) the spatial training group, comprising 13 male mice that received MWM spatial learning training; and 2) the control group, comprising 13 male mice that were placed in the MWM with the platform removed. All mice were housed in a temperature-controlled environment with a 12-h light-dark cycle and free access to food and water. The mice were raised at the Medical Animal Center of Jianghan University. All experiments were carried out in accordance with the National Institutes of Health guide for the care and use of Laboratory animals (NIH Publications No. 8023, revised 1978) and were approved by the Medical Ethics Committee of the Jianghan University.

Mouse tail gene identification

A mouse tail with a length of 0.4–0.6 cm was used to extract genomic DNA by the Mouse Tail Genomic DNA kit (ComWin Biology, China). PCR reaction system including 1μl of forward and reverse primer, respectively, 12.5μl of 2×Tag Master mixed solution, 9.5μl of double distilled water and 1μl of DNA template was used. Sequences of primers (Tsingke Biotechnology, China) are:

5’GACTGACCACTCGACCAGGTTCTG3’; 5’CTTGTAAGTTGGATTCTCATATCCG3’.

Spatial training paradigm

The learning paradigm is simply modified using the standard water maze procedure. The regularity training group conducted learning to find hidden platforms in the water maze. The mice were trained to find hidden platforms in the water maze for 3 consecutive days, after which the mice were allowed to rest for 2 days. There were six training sessions in total, over a duration of 30 days. During the training process, each mouse was manually placed in water towards the wall of the pool. With the exception of the platform quadrant, the other three quadrants were used as the starting positions for training. The animals were allowed to search for the platform for no more than 60 s. If the mouse could not find the platform within 60 s, it was manually guided to the platform and placed on the platform for 10 s. Mice in the control group were placed in a water maze without a platform (the placement time was the average of the time taken by the mice in the training group).

Open field

As previously described [24], mice were placed into an open field box (dimensions: length = 50 cm, width = 50 cm, and height = 50 cm) and automated tracking was measured for 5 min. The time spent in the center and peripheral area of the arena were recorded.

Barnes maze

The Barnes maze platform (91 cm diameter, elevated 100 cm from the floor) consisted of 20 holes (each 5 cm in diameter). All holes were blocked, except for one target hole that served as the recessed escape box. Spatial clues and bright light were used to motivate the mice to find the escape box. During the adaptation phase, each mouse was placed in the escape box for 120 s, and allowed to explore for another 180 s and once again placed in the escape box for another 90 s. In the acquisition phase, each experiment followed the same protocol, with the aim of training each mouse to find the target and enter the escape box within 180 s. The mouse was then allowed to stay in the box for another 30 s. The mice were trained twice a day, with an interval of approximately 15 min each time, for 4-consecutive days. In the exploratory test (day 5), each mouse underwent a 60 s test. The target was placed in the same position, but the escape box was blocked. The number of errors in reaching the target quadrant and the time spent in the target hole were measured.

Western blotting

Mice were sacrificed at 7 months old for further analyses. The mice were transcardially perfused with 50 ml of saline, and the brains of mice were split into two parts down the middle. The hippocampus and cortex were isolated on ice. Six/seven left hippocampi and temporal cortexes were homogenized with RIPA buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 1% TritonX-100, 1% sodium deoxycholate, 0.1% SDS), incubated on ice for 20 min, and centrifuged at 14000×g for 10 min at 4°C. The supernatant was aliquoted and frozen at –80°C until use. Bradford reagent (Beyotime, China) was used for quantitative protein analysis of the homogenate. Proteins were separated using 10% polyacrylamide gel electrophoresis (PAGE). Proteins were transferred onto nitrocellulose (NC) membranes. The membranes were blocked with 5% non-fat dry milk for 1 h at room temperature and then incubated overnight with primary antibodies at 4°C. The primary antibodies used in the experiments were as follows: anti-β-actin (1 : 10000; Proteintech, Catalog: 66009-1-Ig, China), anti-Neprilysin (1 : 2000; proteintech, Catalog: YT5350, China), anti-APP (1 : 1000; Cell Signaling, USA), anti-BACE1 D10E5 (1 : 2000; Cell Signaling, Catalog: mAb5606s, USA), anti-sAPPβ (1 : 50; IBL, Catalog: JP18957, Japan), anti-synaptophysin D8F6H (1 : 1000; Cell Signaling, Catalog: mAb36406, USA), anti-synapsin-1 D12G5 (1 : 1000; Cell Signaling, Catalog: mAb5297, USA), anti-synaptotagmin-1 D33B7 (1 : 1000; Cell Signaling, Catalog: mAb14458, USA), anti-AMPA Receptor 1 (GluA1) D4N9V (1 : 1000; Cell Signaling, Catalog: mAb13185, USA), anti-PSD95 D74D3 (1 : 1000; Cell Signaling, Catalog: mAb3409, USA), anti-PSD93 D424D (1 : 1000; Cell Signaling, Catalog: mAb13185, USA). Blots were detected with 1 : 3000 HRP-linked anti-mouse (Beyotime, China) or 1 : 3000 anti-rabbit (Beyotime, China) secondary antibody. Immune bands were visualized by enhanced chemiluminescence (ECL) system (Bio-Rad, USA) using ECL reagents (Beyotime, China), followed by quantitative analysis by Image J 1.53c (http://imagej.nih.gov.).

qPCR

Four left prefrontal cortexes of 7-month-old APP/PS1 mice were homogenized using a homogenizer in Trizol Reagent (Invitrogen, 15596018, USA). TransScript II First-Strand cDNA Synthesis (Promega, 0000429679, USA) was used for reverse transcription. Afterwards, qPCR was performed with a BioRad CFX Connect machine (BioRad Laboratories, USA) amplified using TB Gene Expression Master Mix (Takara, Japan). The primers (Tsingke Biotechnology) were shown as follows:

SP1: sense 5’GGAAGTCAGCAGAAAGAGGGAG3’;

  anti-sense 5’CTGGGAAATAGCTTGAGTTGTGAA3’,

BACE1: sense 5‘AACATTGCTGCCATCACTGAAT3’;

  anti-sense 5‘TCAAAGAAGGGCTCCAAAGAGT3’,

APP: sense 5‘TCCTTACCGTTGCCTAGTTGGT3’;

  anti-sense 5‘TGCCATAGTCGTGCAAGTTAGTG3’.

Mouse Aβ₄₀ ELISA kit

Three left occipital cortexes were extracted in RIPA as described previously [25] and ELISA kit (Elabscience Biotechnology, E-EL-M3009, China) was used to quantify Aβ₄₀ in mouse brain tissue according to the manufacturer instructions.

BACE1 activity assay

The β-secretase activity of brain tissues was measured with a fluorescence assay kit (Abcam, ab65357, UK). Briefly, the proteins were extracted from 4 left parietal cortexes with ice-cold extraction buffer, then centrifuged at 10,000 g for 5 min at 4°C, and the supernatant was transferred to a new tube and stored on ice. The protein concentration was determined by BCA Kit (Beyotime, China). A total of 50μL sample with 30μg protein was added to each well, followed by 50μL of 2×reaction buffer and 2μL of β-secretase substrate. After incubation in the dark at 37°C for 1 h, a microplate reader (BioTek, USA) was used to record fluorescence with excitation wavelength and emission wavelength of 355 and 495 nm, respectively.

Immunofluorescence

For all immunohistochemistry experiments, mice were perfused transcardially with 50 ml of saline. Four right brain tissues were post-fixed in 4% PFA for 2 days and then transferred to 1×PB (phosphate buffer) containing preservative. Brains were sliced into 20-μm thick coronal sections using an oscillating microtome (Leica VT1000S, Germany). The brain slices were incubated for 1 h at room temperature in a blocking solution consisting of 5% bovine serum albumin (BSA) and 0.1% Triton diluted in 1×PBS. Then, sections were incubated with primary antibodies overnight at 4°C on a shaker. The following primary antibodies were used: anti-Aβ (1 : 100; Cell Signaling, Catalog: mAb8243), anti-Iba1 (1 : 500; Abcam ab178846, UK), anti-MAP2 D5G1 (1 : 200; Cell Signaling, Catalog: mAb8707), and anti-DCX (1 : 100; Abcam, Cambridge ab18723, USA). Following incubation, the sections were incubated with Cy3 and Alexa Fluor 488-conjugated secondary antibodies at 37°C for 1 h. After washing 5 times with 1×PBS, the sections were stained with Hoechst (1 : 1000; Thermo Fisher, USA). Images were acquired using a confocal microscope (Leica TCS SP8, Germany). Two coronal sections per mouse were employed and the averaged values from the images per mouse was used for quantification.

Statistical analysis

Data are represented as mean±SEM. GraphPad Prism 7.4.0 (GraphPad Software, USA) was used for statistical analyses. Two-tailed unpaired Student’s t-tests were used for comparisons between two groups. A probability of p < 0.05 was considered to indicate significant differences between different groups.

No influence on behavioral performance of anxiety by MWM training

In the spatial training paradigm, 2.5-month-old APP/PS1 mice were trained in the MWM with three trials each day, for 3 consecutive days, and with a total of six training sessions (Fig. 1A, B). The non-trained APP/PS1 control mice (N-APP/PS1) were subjected to swimming activity but without a learning contingency. After spatial training, the mice were immediately placed in the open field (OF) to test their anxiety and activity levels. We found that MWM training did not affect the movement distance and the time spent in the central area of the mice in OF (Fig. 1C, D), which indicates that there is no change in anxiety performance.

OPEN IN VIEWER

Fig. 1

MWM training does not induce anxious performance in APP/PS1 mice. A) Behavior schedule of APP/PS1 mice with No or MWM training during 1 month with OF test. B) Acquisition time spent in MWM of APP/PS1 mice. Time spent in the quadrangle (C) and the center (D) of OF. (n = 13 mice/group) Data expressed as means±SEM.

Reduced amyloid plaques within brains in 3.5 months after training

APP/PS1 mice develop subtle memory deficits at 6–7 months of age, which coincides with the appearance of Aβ plaques, and memory impairments increase with the accumulation of Aβ deposits [26, 27]. A previous study demonstrated that repeated spatial training markedly reduced both Aβ plaque load and insoluble Aβ₄₂ levels in the short term [23]. To explore whether spatial training has a long-term effect on Aβ deposits, an immunofluorescence assay with anti-Aβ was performed. Histopathology and quantitative analyses showed that the coverage of Aβ plaques were significantly reduced in the cortex of the APP/PS1 + training group compared to the controls, while the mean fluorescence intensity was similar between groups (Fig. 2A, C, and D). Similar results were found in the DG of the hippocampus (Fig. 2B, E, and F). We further monitored soluble Aβ₄₀ by ELISA kit and found that MWM training decreased the levels of soluble Aβ₄₀, although not significantly (Supplementary Figure 2A). Taken together, these data strongly suggest that spatial training can attenuate Aβ accumulation in APP/PS1 mice.

OPEN IN VIEWER

Fig. 2

MWM training reduces amyloid plaques 3.5 months later. Representative microscopic fields of Aβ plaques (red) from cortex (A) and hippocampus (B) in No or MWM training APP/PS1 mice with Hoechst labeling of cell nuclei (blue) (scale bar, 100μm). C, D) Quantification of plaques area and mean fluorescence intensity in the cortex of APP/PS1 and APP/PS1 + TR mice (2 brain slices from each mouse and 4 mice in each group. ^*p < 0.05, two-tailed unpaired t-test). E, F) Quantification of plaques area and mean fluorescence intensity in the hippocampus of APP/PS1 and APP/PS1 + TR mice (two brain slices from each mouse and 4 mice in each group. ^*p < 0.05, two-tailed unpaired t-test). Data expressed as means±SEM.

Reduced expression levels of BACE1 and AβPP within mice brains in 3.5 months after training

Multiple robust evidence has demonstrated that BACE1 is a critical enzyme in amyloidogenic processes [28, 29]. BACE1 cleaves AβPP at the N-terminus of the Aβ sequence, releasing secreted sAPPβ and C99 fragments that are required for the generation of all neurotoxic forms of Aβ, and leading to plaque deposits [30]. Therefore, BACE1 and AβPP levels are closely associated with Aβ accumulation, and the level of sAPPβ/AβPP could indicate BACE1 activity [31, 32]. To explore the potential mechanisms of the anti-Aβ deposit effect of MWM training, the expression levels of BACE1, sAPPβ and AβPP were tested by western blotting. We found that MWM training significantly reduced the levels of BACE1 and AβPP in cortices, although without statistical significances (Fig. 3A, C, and D). In the hippocampus, there was no significant decrease in BACE1 and AβPP levels in the spatial training group compared to the APP/PS1 control group (Fig. 3B, E, and F). The level of sAPPβ/AβPP was comparable between the two groups in cortices and hippocampi, suggesting no significant change in BACE1 activity (Supplementary Figure 3A, B). To confirm the result, the activities of BACE1 was further investigated by Kit. As shown in Fig. 3G, spatial training did not exhibit inhibitory effects on BACE1 activities.

OPEN IN VIEWER

Fig. 3

MWM training reduces the levels of BACE1 and AβPP 3.5 months later. A, C, D) Western blot and quantification of BACE1 (Control = 6, TR = 7) and AβPP (n = 6 per group) in the cortex from No or MWM training APP/PS1 mice (^*p < 0.05, two-tailed unpaired t-test). B, E, F) Western blot and quantification of BACE1(n = 3 per group) and AβPP (n = 6 per group) in the hippocampus from No or MWM training APP/PS1 mice (^*p < 0.05, two-tailed unpaired t-test). G) Quantification of BACE1 relative activity (n = 4 per group, two-tailed unpaired t-test was used). (H and I) Relative mRNA levels of AβPP and BACE1 (n = 4 per group, ^**p < 0.01, two-tailed unpaired t-test was used). Data expressed as means±SEM.

In addition, we employed qPCR to measure mRNA levels and observed there were a remarkable reduction in AβPP level and a decreased tendency of BACE1 in the spatial training group compared to that of the control group (Fig. 3H, I). But the mRNA levels of SP1, the transcription factor for AβPP, was altered between the groups (Supplementary Figure 3C). The underlying mechanism of the decrease of AβPP and BACE1 expression induced by spatial training needs further investigation. Taken together, these data suggest that spatial training could mediate BACE1 and AβPP levels, which might further ameliorate Aβ burden in AD model mice with long-term effects.

No altered microglia activation and NEP expression in brains 3.5 months after MWM training

Since the activation of microglia is involved in the clearance of amyloid plaques and is one of the features of AD pathology [33-35], we compared the amounts of activated microglia by Iba1 (marker for microglia activation) immunofluorescence. As shown in Fig. 4A-C, the count of Iba-1-positive cells was not changed in the hippocampus and cortex of spatial training group as compared to the control. Next, we employed western blot to quantify the expression of neprilysin (NEP) involved with the degradation of Aβ peptide [36, 37]. There was no significant difference in the levels of NEP expression between the two groups in the hippocampus (Fig. 4D, E) and cortex (Fig. 4F, G). Altogether, these results indicate that MWM training did not modify microglia activation and NEP expression of APP/PS1 mice after 3.5 months, which is associated with Aβ clearance.

OPEN IN VIEWER

Fig. 4

MWM training does not modify microglia activation and NEP expression 3.5 months later. A) Immunofluorescence staining in hippocampus and cortex of mice at the indicated ages for microglia (Iba1) (scale bar, 100μm). Quantification of Iba1-positive cells in the hippocampus (B) and cortex (C) (n = 4, two brain slice from each mouse, ^*p < 0.05, two-tailed unpaired t-test was used). Western blot and quantification of NEP in the hippocampus (D, E) and cortex (F, G) from No or MWM training APP/PS1 mice (n = 3 per group, ^*p < 0.05, two-tailed unpaired t-test was used). Data expressed as means±SEM.

Altered synaptic morphology and expression of synapse-associated proteins in mice 3.5 months after training

Aβ has been found to impair synaptic plasticity as a neuromodulator [38, 39]. Several studies indicate that the immunoreactivity of microtubule-associated-protein-2 (MAP2) is positively correlated with dendritic density [40–42]. To investigate whether MWM training could mitigate synaptic impairments, dendritic integrality was evaluated by immunofluorescence (IF) assay using MAP2 antibody (Fig. 5A). The results showed a significant increase in dendritic density in DG, CA1, and CA3 of hippocampus of mice with MWM training compared to that in the APP/PS1 group (Fig. 5A-D). There was no significant difference in the cortex between the two groups (Fig. 5A, E). We further measured the expression levels of several synapse-associated proteins by western blotting. The results showed that MWM training significantly enhanced the levels of Synaptotagmin-1 and GluA1 and increased the levels of PSD93, although not significantly. The training did not significantly change the levels of Synapsin-1, Synaptophysin, and PSD95 in the hippocampus (Fig. 6A, B). Additionally, there was no significant difference in the levels of these synaptic proteins between the two groups in the cortex (Fig. 6C, D). Taken together, these data imply that spatial training partially restored synaptic integrity and the levels of synapse-associated proteins in APP/PS1 mice.

OPEN IN VIEWER

Fig. 5

MWM training modifies dendritic morphology 3.5 months later. A) Representative images of dendrite stained with MAP2 (green) and Hoechst labeling of cell nuclei (blue) in DG, CA1, CA3 and cortex from No or MWM training APP/PS1 mice (scale bar = 100μm). Quantification of mean fluorescence intensity in DG (B), CA1 (C), CA3 (D), and cortex (E) (two brain slices from each mouse and 4 mice in each group. ^*p < 0.05, two-tailed unpaired t-test). Data expressed as means±SEM.

OPEN IN VIEWER

Fig. 6

MWM training increases the expression of synaptic proteins 3.5 months later. A, B) Western blot and quantification of Synapsin-1, Synaptophysin, Synaptotagmin-1, GluA1, PSD95, and PSD93 in the hippocampus from No or MWM training APP/PS1 mice (n = 3 per group, ^*p < 0.05, two-tailed unpaired t-test). C, D) Western blot and quantification of Synapsin-1, Synaptophysin, Synaptotagmin-1, GluA1, PSD95, and PSD93 in the cortex from No or MWM training APP/PS1 mice (n = 3 per group, two-tailed unpaired t-test). Data expressed as means±SEM.

Altered ANH 3.5 months after MWM training

A sharp decline in AHN has been observed in both patients with AD and animals, and reduced AHN was associated with impairments in hippocampal function [8, 9]. AHN is derived from the proliferation of neural stem cells (NSCs), most of which are localized in the granulosa layer of the DG in the hippocampus [12, 43]. To measure the number of immature granular cells, an immunostaining assay with doublecortin (DCX, biomarkers for immature granular cells) antibody was employed (Fig. 7A). The results showed that MWM training significantly enhanced the population of DCX+ cells in the DG granulosa layer (Fig. 7B). Taken together, these results suggest that MWM training significantly increases AHN.

OPEN IN VIEWER

Fig. 7

MWM training enhances AHN 3.5 months later. A) Representative images of immature granular cells stained with DCX (red) and Hoechst labeling of cell nuclei (blue) in DG from No or MWM training APP/PS1 mice (scale bar = 200μm). B) Density of cells expressing DCX in DG hippocampal subfields (four mice in each group. ^*p < 0.05, two-tailed unpaired t-test). Data expressed as means±SEM.

No altered cognitive impairment 3.5 months after MWM training

Next, we investigated the long-term effects of spatial training on the cognitive performance of APP/PS1 mice. The Barnes maze was used to measure spatial learning and memory ability. In the training phase, MWM training mice showed no difference in primary errors to obtain the target compared to the controls (Fig. 8A). During the probe test, there was no significant difference in probe primary errors, time and distance spent in the target quadrant between the two groups (Fig. 8B-D). Altogether, these results suggest that MWM training did not attenuate cognitive impairment of APP/PS1 mice in 3.5 months. We also explore the long-term effects of MWM training on the ability of learning and memory of wild-type mice. Similar to above results in APP/PS1 mice, primary errors to get the targe in the training phase and primary errors in the probe test are not altered (Supplementary Figure 4A, B), suggesting that MWM training did not affect the cognitive performance of wild type mice after 3.5 months.

OPEN IN VIEWER

Fig. 8

MWM training does not alter cognitive performance 3.5 months later. A) The primary errors to locate the target in the training phase of Barnes maze (Control/TR n = 7/6). B) Primary errors to locate the target in the probe test (Control/TR n = 7/6). Time (C) and distance (D) spent in the target quadrant in the probe test (Control/TR n = 7/6). Data expressed as means±SEM.