The Helico Maze Detects Early Impairment of Reference Memory at Three Months of Age in the 5XFAD Mouse Model of Alzheimer’s Disease

Abstract

Background:

The 5XFAD model of Alzheimer’s disease (AD) bearing five familial mutations of Alzheimer’s disease on human APP and PSEN1 transgenes shows deposits of amyloid-β peptide (Aβ) as early as 2 months, while deficits in long-term memory can be detected at 4 months using the highly sensitive olfactory-dependent tests that we previously reported.

Objective:

Given that detecting early dysfunctions in AD prior to overt pathology is of major interest in the field, we sought to detect memory deficits at earlier stages of the disease in 3-month-old male 5XFAD mice.

Methods:

To this end, we used the Helico Maze, a behavioral task that was recently developed and patented. This device allows deeper analysis of learning and subcategories of hippocampal-dependent long-term memory using olfactory cues.

Results:

Eight male 5XFAD and 6 male wild-type (WT: C57Bl6 background) mice of 3 months of age were tested in the Helico Maze. The results demonstrated, for the first time, a starting deficit of pure reference long-term memory. Interestingly, memory impairment was clearly correlated with Aβ deposits in the hippocampus. While we also found significant differences in astrogliosis between 5XFAD and WT mice, this was not correlated with memory abilities.

Conclusion:

Our results underline the efficiency of this new olfactory-dependent behavioral task, which is easy to use, with a small cohort of mice. Using the Helico Maze may open new avenues to validate the efficacy of treatments that target early events related to the amyloid-dependent pathway of the disease and AD progression.

Keywords

Alzheimer’s disease astrogliosis 5XFAD mouse Helico Maze hippocampal Aβ deposits reference memory deficit

INTRODUCTION

Alzheimer’s disease (AD) is a neurodegenerative disease resulting from complex interactions between genetic and environmental factors [1]. Three different phases during the development of the pathology can be distinguished. An initial asymptotic phase involves multiple deregulated molecular pathways, e.g., transcriptional changes, protein processing, amyloid-β (Aβ) aggregation and tau hyperphosphorylation, aberrant signaling, and disturbed metabolism. These events are all associated with brain tissue damage through exacerbated inflammatory processes, gliosis, oxidative stress, synapse loss, and neuronal death. These early processes contribute to the development of a clinical prodromal second phase with episodic memory deficits, followed by a third phase of “syndromal” dementia characterized by total loss of general cognitive functioning [2]. It has to be mentioned that not only cognitive decline outcomes characterized AD outcomes but also neuropsychiatric deficits as recently suggested [3]. In humans, several therapies have been tested in efforts to stop the development of the disease, but most clinical trials with disease-modifying agents have been performed too late in the pathophysiological course of AD, at the prodromal or final stage [4]. Consequently, to delay the progression of the disease, testing cognitive or pharmacological treatments during the asymptotic period has been suggested. Several transgenic mouse models have been used for this purpose [5], but in most cases, behavioral impairments have been observed during the prodromal stage according to tests that evaluated various cognitive demands.

In short, AD mouse models were developed to assess these different phases to test the efficacy of novel treatments. The onset of visible Aβ deposits varies across mouse models, with the 5XFAD model being among the most aggressive [6 –9]. These mice co-overexpress and coinherit under control of the neuron-specific Thy1 promoter, human transgenes with five familial AD (FAD) mutations, and amyloid precursor protein (APP) and presenilin1 (PSEN1) genes act synergistically to increase the levels of cerebral Aβ species, especially the highly neurotoxic Aβ₄₂. In these transgenic mice, visible deposits of Aβ can be seen as early as 2 months of age [6]. Nevertheless, the first learning and memory deficits in these mice have been detected at 4 months with the olfactory tubing maze (OTM) [10] or the radial-arm water maze [11] or at 5 months of age using the Morris water maze [12] or the contextual fear conditioning task [13]. In addition, we demonstrated impairment of executive functions in male 5XFAD mice at 4 months of age, similar to impairment of function observed in primates with frontal lesions, using the olfactory H-maze [10]. Recently, using the conditioned odor preference task, O’Leary and colleagues found no evidence of memory impairments in 5XFAD mice [14]. In contradiction to our previous results [10], the authors suggested that the detection of memory impairment is related to cognitive demand. Indeed, using the OTM, odors are presented simultaneously, and mice engage in odor-guided navigation (turn left versus right) to be rewarded. This complex task requires more cognitive demand and might better strengthen memory. Consequently, O’Leary and colleagues concluded that the OTM is more sensitive at detecting olfactory memory impairments in 5XFAD mice than the conditioned odor preference task [14]. Thus, modeling olfactory deficits in AD may require more optimized and cognitively demanding behavioral tests.

Taking these concerns into account, we developed a new device, the Helico Maze (HM) [15]. Compared to the OTM, this new olfactory-based maze reinforces cognitive demand in mice and allows deeper analysis of learning and subcategories of hippocampal-dependent long-term memory impairments [15]. In this study, we evaluated young male 5XFAD mice of 3 months of age, when amyloid pathology is incipient [6, 16]. We describe here that the HM is highly efficient in detecting early memory deficits in 5XFAD mice and requires a low number of animals. In addition, we found a clear correlation between the number of amyloid plaques and cognitive ability. For the first time, an original test has been demonstrated to assess the efficacy of AD treatments in the asymptomatic phase of the disease.

MATERIALS AND METHODS

Animals

Animal experiments were approved by the Ethics Committee of the Medical Faculty of Marseille (research project: APAFIS#2019100910138008) and carried out in accordance with the guidelines published in the European Communities Council Directive of November 24, 1986 (86/609/EEC) and with the recommendations of the ARRIVE guidelines [17]. All efforts were made to minimize animal suffering and to reduce the number of mice used.

The 5XFAD mouse model generation has been described previously [6]. These transgenic mice overexpress the mutant human APP gene encoding the 695 isoform bearing the Swedish (K670 N, M671 L), Florida (I716 V), and London (V717I) FAD mutations and the human PSEN1 gene harboring two FAD mutations, M146 L and L286 V. Neural-specific elements of the mouse Thy1 promoter to drive the overexpression of these transgenes in neurons. The 5XFAD mice used in this study had a C57Bl6 background [11, 18]. Transgenic 5XFAD heterozygous mice were used for experiments with non-transgenic WT littermates as controls. Prior to experiments, genomic DNA was extracted from the tail tips of all mice to assess their genotype by PCR [16]. All the mice were bred in our animal facility, had access to food and water ad libitum, and were housed under a 12 h light–dark cycle at 22–24°C. Behavioral testing and immunostaining of brain sections were performed using 3-month-old 5XFAD mice (N = 8) and their respective WT littermates (N = 6). The number of mice used and analyzed in this study was determined and validated using the resource equation method [19].

Helico Maze apparatus

The original HM (Fig. 1) is a patented device (PCT/EP2020/079758) that has been recently extensively described [15]. Briefly, two identical testing chambers (TCs), TC1 (left) and TC2 (right), are connected by one straight plastic tube. The latter is a tube of 100 cm that is interrupted in the middle by an automated door (AuD). Each testing chamber comprises an entrance cube that is horizontally divided in the middle by a platform perforated with four holes of 2 cm positioned in front of the entrances to the four tubes. On top of each TC, an inverted fan exhausts the neutral or scented air that is ejected from the outside extremities of the tubes. The four straight plastic tubes of each testing chamber form a cross. At each extremity, a water port (WP) allows the release of water in a well. A ribbon LED of 10 cm (10 W/meter) was placed in a triumphal arc 10 cm from the floor surmounts each tube.

Fig. 1

Illustration representing the Helico Maze (HM). The HM contains two testing chambers (TC1 and TC2). Odors (O1, O2, O3, O4) are injected during a trial. In every trial, one odor is specified for each tube. AuD, automatic door; IF, inverted fan; WP, water port. Odor and water ports are surmounted by a ribbon LED (RL).

Photoelectric cells located all over the tubes are used to track mouse movement. The HM was set in an independent room in total darkness. Four infrared cameras around the HM allowed us to follow mouse behavior on screens in another room. Electronic devices allowing automatic management and recording of the results were also in another room.

Procedural training

Before the training procedure (first day), all the mice were weighed (and on every subsequent day) to monitor and ensure that the weight lost once water deprivation began was less than 15%, and the mice were handled for 5 min the two first days. On Day 3, mice were introduced only in TC1, where they could find 0.1 mL of water at each extremity of the four tubes (Fig. 2). On Day 4, the mice were weighed and then handled for 5 min before being introduced into the TC2 without access to the TC1. Again, the mice could find the same amount of water at the extremity of the 4 tubes. On Day 5, the mice were allowed to move freely in all parts of the HM and all water ports (WPs) were loaded with 0.1 mL of water. Usually, after these 15 min shaping sessions, the mice drank some of the water that was available in the cups at every extremity of the HM. On Day 6, at 12 : 00 a.m., the mice were deprived of water until the first training session on Day 8.

Fig. 2

Brain coronal sections at the dorsal hippocampal level of WT (A, B) and 5XFAD (C, D and E) mice of three months of age. Sections were immunostained with anti-Aβ (6E10, in green, A and C) for amyloid deposits and anti-GFAP (in red, B and D) for astroglia or for both markers in the higher-magnification image of the dentate gyrus area (E). The percentage of fluorescent pixels was quantified after GFAP staining of 30μm thick hippocampal coronal sections (F). The results show the mean (±SEM) of 6 5XFAD mice and 6 WT mice. An * indicates p < 0.05. Scale bar: 500μm.

The mice were introduced in TC1 without access to the TC2 (the AuD was closed). Once the mice were detected, different odors were immediately ejected from the extremities of the four tubes. One positive odor was associated with a positive reward, i.e., 0.3 mL of water, while the other three negative odors were associated with a negative reward, i.e., lighting of the ribbon LEDs. The mice randomly chose one of the four connected tube and ran to its extremity to be detected by the photoelectric cell. If this one is the “positive” tube, water is released. This ended the trial, and neutral air was now ejected into the tubes. If this one is the “negative” tube, mice received a bright flash. The light was turned on until the mouse went to the correct tube to be positively rewarded, which stopped the trial. Here, a correction procedure was applied since the three lights could potentially be turned on until the mouse reached the end of the “positive” tube. This represents the first trial. After an intertrial period of 15 s, the AuD opened, allowing the mouse to enter into the TC2. Once the mouse was detected in TC2, the AuD closed and different odors were differently distributed in the four connected tubes. The same procedure is followed as described above for the subsequent trials.

In this procedural training, each mouse performed 8 trials until the experimental session finished and this procedure was repeated 5 successive days. At the end of each of the first four sessions, the mice were returned to the vivarium with free access to water for 2 min. At the end of the last session (session 5 - Day 12), deprivation was stopped.

On Day 13, all the mice were again water-deprived at 12 : 00 a.m. before being submitted to reference training on Day 15.

In this procedural training, three parameters were examined 1) the duration (min) needed to perform the 8 trials (maximum session time of 60 min); 2) the number of incorrect responses (with a maximum of twenty-four errors per session) and 3) the number of correct responses with 2 out of 8 correct responses reflecting the chance level.

Reference training

The reference training was similar to the procedural training with slight modifications. Indeed, a daily session was composed of 16 trials and a noncorrection method was applied during 5 successive sessions. Once a correct (drop of water) or incorrect response (light on for 2 s only) was given, the odor delivery stopped and was replaced by neutral air. Here, using the non-correction method, only the time and the number of correct responses expressed as percentages, with a chance level of 25% (2 out of 8 correct responses) if all the choices for every trial were made randomly, were analyzed.

Immunohistochemistry

After the behavioral tests, mice were deeply anesthetized with intraperitoneal injection of xylazine (15 mg/kg) and ketamine (150 mg/kg) (Ceva Santé animale, Libourne, France), transcardially perfused with 0.9% NaCl and then with AntigenFix (Diapath, MM France, Brignais, France). The brains were then extracted and postfixed overnight in cold AntigenFix. Coronal brain sections (30μm thick) were serially generated using a Leica CM3050 S cryostat (Leica biosystems, Nanterre, France) and stored at -20°C in cryoprotectant solution (30% glycerol and 30% ethylene glycol (both purchased from Sigma–Aldrich) in 0.05 M phosphate-buffered saline (PBS, Thermo Fisher Scientific, Saint Aubin, France). After washing in PBS, floating brain sections were incubated for 1 h at room temperature (RT) within blocking buffer (10% bovine serum albumin (BSA, Sigma–Aldrich) and 0.1% Triton X-100 (Sigma–Aldrich) in 0.1 M PBS) and then overnight at 4°C with the following primary antibodies diluted in 0.1 M PBS containing 2% BSA and 0.1% Triton X-100: rabbit polyclonal anti-GFAP (1/400, Dakocytomation, Trappes, France) and mouse monoclonal anti-Aβ 6E10 (1/200, BioLegend, Ozyme, Saint-Cyr l’Ecole, France). Brain slices were then rinsed (3 x 5 min) with PBS and incubated for 1 h 30 min at RT with highly cross-adsorbed Alexa Fluor 488- or 594-conjugated anti-rabbit or anti-mouse secondary antibodies (1/500, Thermo Fisher Scientific) in the dark. After three washes in PBS, slides were counterstained with 0.5μg/mL Hoechst blue (#33342, Sigma–Aldrich) in 2% BSA and 0.1% Triton X-100 in 0.1 M PBS for 30 min at RT and mounted with ProLong Gold Antifade reagent (Thermo Fisher Scientific) on Superfrost glass slides (Dutscher).

Image acquisition and analysis

Images were acquired with an AxioObserver Z1 microscope (Carl Zeiss S.A.S., Marly le Roi, France). Extracellular amyloid deposits in the hippocampal area were manually counted and the number of plaques per mm² of brain surface was calculated using Zen software (Zeiss). For quantification of GFAP staining, ImageJ software (NIH) was used. Images were binarized to 16-bit black and white images, and a fixed intensity threshold was applied defining GFAP staining. The percentage of the area covered by fluorescent staining was calculated. The final data represent the average of two tissue sections from the same brain area per mouse. Representative images of large brain sections were obtained using the mosaic mode of Zen software with a 10x objective.

Statistical analysis

We first verified that all the parameters tested were normally distributed according to the Shapiro-Wilk test for small sample sizes before being analyzed by multivariate analyses of variance (MANOVAs) with repeated measures using SPSS/PC+statistics 11.0 software (SPSS Inc., Chicago, IL). Independent ANOVAs were performed when necessary. In the figures, * indicates p < 0.05, ** indicates p < 0.01, and *** indicates p < 0.001 difference between groups. In addition, behavioral performance (percentage of correct responses) were compared to chance levels using a Student’s t test with Graph Prism 5.02 software, with # p < 0.05, ## p < 0.01, and ### p < 0.001 in comparison to the chance level (dotted line). A Pearson’s correlation was done between the percentage of correct responses obtained during the last reference training session and the number of amyloid deposits and the percentage of immunoreactive pixel for GFAP staining observed in the dentate gyrus area.

RESULTS

Amyloid deposits and gliosis in the hippocampi of WT and 5XFAD mice

Brain coronal sections at the dorsal hippocampal level of 3-month-old male WT and 5XFAD mice were immunostained with anti-Aβ (6E10 antibody) and anti-GFAP antibodies (Fig. 2A-E), and staining was quantified. As expected, no amyloid plaques were detected in WT mice (Fig. 2A). Among the eight 5XFAD mice, amyloid plaques were observed bilaterally in the hippocampus of five mice and unilaterally (left side) in one mouse, while two transgenic mice had not yet developed amyloid deposits. Accordingly, the average number of plaques per mm² was 6.27±2.05 in the 5XFAD mice (Fig. 2C). Regarding astrogliosis, a slight, significant increase in GFAP immunolabeling was observed in the 5XFAD mice compared with the WT mice (Fig. 2B, D, F) [ANOVA: F(1, 10)=7.52; p < 0.05]. These mild neuropathological alterations reflect the onset of the pathology in young transgenic mice of 3 months of age.

Behavioral performance

Procedural learning

Duration. Over the session period of 5 days, we observed that each group learned/understood the test, as both WT and 5XFAD mice progressively decreased the time required to perform their 8 trials. Moreover, they both reached the end of the test in < 15 min by session 5 (Fig. 3A). We found no significant differences across sessions between the two groups [MANOVA: Group×session interaction, F(4,40)=0.43; NS] and between groups [F(1,10)=1.09; NS] or during selected sessions [ANOVA: F(1,10)≤2.1; NS].

Fig. 3

Learning and memory performance in procedural (A, B, C: 8 trials corrected by sessions) and reference (D, E: 16 trials uncorrected by sessions) training. A and D show mean performance (±SEM) using session duration in minutes. B and E show mean performance (±SEM) using number of correct responses as a percentage. C shows mean performance (±SEM) using number of incorrect responses. **p<0.01 between WT and 5XFAD mice. In comparison to chance (dashed line): ^# p<0.05; ^# # p<0.01; ^# # # p<0.001.

Correct responses. We then evaluated the number of correct responses (Fig. 3B). Again, no differences were observed across the sessions [MANOVA: Group×session interaction, F(4,40)=0.05; NS] and between groups [F(1,10)=0.17; NS]. The responses of WT and 5XFAD mice were not different from the chance level of 25% across sessions, even in the last session [Student’s t test: t(5)≤2.24; NS].

Incorrect responses. We found that the number of incorrect responses was similar across sessions [MANOVA: Group×session interaction, F(4,40)=0.21; NS] and no differences were observed between WT and 5XFAD mice over the 5 days of the experiment [F(1,10)=0.17; NS] (Fig. 3C). Interestingly, in sessions 2, 3, and 5, the WT mice made fewer incorrect responses compared to the chance level, i.e., 12 on Day 1 [Student’s t test: t(5)≥2.62; p≤0.05]. Concerning the 5XFAD mice, fewer incorrect responses were observed only in session 5 in comparison to the chance level [Student’s t test: t(5)=2.85; p < 0.05].

Reference training

Duration. WT and 5XFAD mice spent the same amount of time performing the 16 daily trials across the 5 training sessions (Fig. 3D). Both groups needed 15±2 min to achieve the full session of 16 trials [MANOVA: Group×session interaction, F(4,40)=1.87; NS], with no group difference [F(1,10)=0.2; NS].

Correct responses. The two groups significantly improved their percentage of correct responses (Fig. 3E), without differences across sessions [MANOVA: Group×session interaction, F(4,40)=1.24; NS] or between groups [F(1,10)=4.07; NS]. It should be noted that a significant difference between WT and 5XFAD mice appeared in session 2, when the WT mice performed better than the 5XFAD mice [ANOVA: F(1,10)=9.98; p < 0.01]. All mice performed the reference training every day above the chance level in the WT group [Student’s t test: t(5)≥4.23; p≤0.01] and in the 5XFAD group [Student’s t test: t(5)≥2.71; p≤0.05].

To examine how the mice improved their performances session after session, we analyzed each session by blocks of 4 trials, thus 4 blocks per session. In session one (Fig. 4, D1), control mice were different from the chance level (25%) during all 4 blocks of 4 trials [Student’s t test: t(5)≥2.74; p≤0.05], while 5XFAD mice were only different during the second block of trials [Student’s t test: t(5)=2.7; p < 0.05]. In the second session (Fig. 4, D2), WT mice were different during all 4 blocks of trials, as in session 1 [Student’s t test: t(5)≥4.00; p≤0.01], while 5XFAD mice differed only during blocks 2 and 3 [Student’s t test: t(5)≥6.32; p≤0.01]. In session 3 (Fig. 4, D3), all the blocks were again different from chance for WT mice [Student’s t test: t(5)≥3.5; p≤0.05] but only blocks 2, 3, and 4 were different for 5XFAD mice [Student’s t test: t(5)≥4.00; p≤0.05]. In session 4 (Fig. 4 D4,), WT mice were different in all blocks except the first block [Student’s t test: t(5)≥5.96; p≤0.01], and, for the first time 5XFAD mice were different in all blocks [Student’s t test: t(5)≥2.7; p≤0.05]. Finally, during the last session (Fig. 4, D5), WT mice were again different from chance during all blocks [Student’s t test: t(5)≥5.97; p≤0.01], and, except for the first block, 5XFAD mice performed similarly above the chance level [Student’s t test: t(5)≥3.8; p≤0.05]. Interestingly, we observed significant differences between WT and 5XFAD mice on Day 2, block of trials 4; on Day 3, block of trials 1 and on Day 5, block of trials 1 [ANOVA: F (1,10)≥5.00; p≤0.05]. These daily data suggest that a significant difference between WT and 5XFAD mice occurred during the first block of 4 trials each day. As this first block of trials is related to previous training, to investigate pure memory performance, we decided to consider only the first trials of each of the five sessions (Fig. 4, T1). As expected, we found strong differences between groups. Indeed, 5XFAD mice gave 6-fold fewer correct responses than WT mice, indicating an important memory deficit [ANOVA: F (1,10)=6.67; p < 0.05].

Fig. 4

Reference training performance (mean±SEM) was analyzed in comparison to chance (dashed line) by blocks of 4 trials by session (S1, 2, 3, 4, 5: days): ^# p<0.05; ^# # p<0.01; ^# # # p<0.001. T1: memory performance (mean±SEM) in the first trial for every session. Comparisons are between WT and 5XFAD mice. * p<0.05; * p<0.01.

We then investigated whether there was a correlation between behavioral performance and hippocampal histopathological alterations in the 5XFAD mice. The percentages of correct responses obtained during the last reference training session (session 5) were compared with the number of amyloid plaques (Fig. 5A) or the percentage of immunoreactive pixels after GFAP staining (Fig. 5B), for the eight 5XFAD mice used in the study. When the number of amyloid plaques increased, we observed a significant decrease in correct responses. This was supported by a significant negative correlation (r² = 0.64; p < 0.01). The relationship between GFAP staining and behavioral performance also demonstrated a negative, but not significant, correlation (r² = 0.31; NS). As expected, the two 5XFAD animals that did not develop plaques had better performance than the others (Fig. 5A). It must also be noted that a high percentage of correct responses (75%) was obtained by the 5XFAD mouse that developed amyloid plaques unilaterally in the hippocampus. These results suggest that the extent of Aβ deposition and its bilateral distribution contribute critically to learning and memory abilities in our paradigm.

Fig. 5

Pearson’s correlation between the percentage of correct responses obtained in the last session (session 5) and the number of amyloid plaques (A) and correlation between the percentage of correct responses obtained in the last session (session 5) and the percentage of immunoreactive pixels after GFAP staining (B) in our 5XFAD group of mice (8 animals).

DISCUSSION

As expected, at 3 months of age, few Aβ deposits were observed in the hippocampi of most 5XFAD mice. Among these 8 mice, only two did not develop plaques at this age, and, surprisingly, one mouse developed plaques only on one side of the hippocampus. Taking into account the overall behavioral performances of these 8 5XFAD mice compared with the 6 WT mice from the control group, no significant differences were observed in procedural or in reference training. All mice from both groups reached a high level of performance in both trainings. Nevertheless, when the two 5XFAD mice without Aβ deposits were excluded, a significant deficit appeared in reference learning. This seems to be a pure reference memory deficit, as we observed a deep impairment on the first trial of every reference training session. An even deeper overall impairment could be seen when the 5xFAD mouse showing unilateral Aβ deposits was excluded (not shown). Consequently, the direct association between the number of correct responses obtained in the last session of reference training (session 5) and the number of amyloid plaques clearly indicates that the level of behavioral impairment observed in our original HM correlates with disease onset and progression.

Use of the HM allowed us to observe early behavioral deficits at the age of three months with only 6 mice. In the literature, only a few studies have described early memory impairment in 5XFAD mice [10, 11], and there is no direct correlation between the levels of neurodegenerative markers, e.g., amyloid plaques, and learning and memory impairment. We previously reported learning deficits at four months of age using original olfactory-dependent mazes [10, 20]. This behavioral impairment was significantly reduced upon suppression of MT5-MMP, a proamyloidogenic proteinase [21, 22], whose deficiency in 5XFAD mice correlates with a decrease in amyloid plaques in the cortex, frontal cortex, and hippocampus [11, 23]. Accordingly, our results in this study suggest that the HM could be a powerful tool to evaluate the efficacy of treatments on learning and memory abilities because 1) it may be administered to 5XFAD mice during the first months of life at the onset of the disease and 2) it requires few animals in comparison to other cognitive tests. The superior performance of the HM results from refining the behavioral training as well as from tasks involving a higher cognitive demand. These important technological and experimental points have been recently recommended in the ARRIVE guidelines for conducting animal experiments [17]. All these reasons support further use of the HM to evaluate early memory impairment in other mouse models of AD or in other neurodegenerative diseases with small cohorts.

Another interesting finding of our study was the heterogeneous development of pathology in 5XFAD mice. All 5XFAD mice were born to 6 WT females within 2 days, developed in the same environment as their WT littermates and their mothers, and were separated according to genotype at 1 month of age. However, although they shared the same genetic background, and the same cognitive impairment correlated with the number and localization of Aβ deposits, the onset of disease was different throughout the group of 5xFAD mice, thus reproducing what has been observed in humans with familial AD mutations [24 –27]. Indeed, in these families, the onset of disease appeared at different times in life. Thus, environmental factors called epigenetic factors can thoroughly modulate the development and progression of the disease, indicating that under particular circumstances, the onset of disease can be modulated. Several therapeutic strategies to delay if not to suppress the Aβ-induced pathogenesis of AD have already been attempted [18, 28].

In this context, we previously demonstrated in 5XFAD females that chronic treatments with the specific 5-HT₄ receptor partial agonist drug RS67333 during the asymptotic phase of the disease decreased the number of amyloid plaques and the level of Aβ species and reversed behavioral deficits observed at four months of age using the novel object recognition test [28]. In line with these results, a two-month chronic treatment with drug RS67333 of 2- to 4- month-old male 5XFAD mice allowed us to observe a decrease in memory deficits in the olfactory tubing maze, compared to vehicle-treated 5XFAD mice, and this decrease was concomitant with amyloid plaque reduction in the entorhinal cortex [18]. Our data stress the importance of preclinical testing of new therapeutic agents, as soon as possible, to detect early cognitive improvement correlated with amyloid pathology. In our view, 5XFAD mice of three months of age could gain one month using the HM, making it possible to detect early memory impairment at three instead of four months and allowing us to test the efficacy of chronic treatment at the onset of the disease.

Advances in molecular biology technologies have enabled the discovery and understanding of the importance of epigenetic regulation in several complex diseases, such as neurodegenerative disorders [29]. Although the role of epigenetic mechanisms in AD is still poorly understood, recent findings clearly show that such mechanisms at play are dysregulated during disease progression, particularly upon environmental manipulation. In rodents, a model of such environmental manipulation uses enriched housing to provide cognitive and social stimulation [30 –32]. One study reported that mice coexpressing human genes APP and PSEN1 with familial mutations, housed in enriched cages from two months of age, developed a higher amyloid burden with commensurate increases in aggregated and total Aβ in comparison to control mice housed in standard conditions [33]. Other authors observed an improvement in cognitive function in the Morris water maze, circular platform recognition and radial arm water maze in another mouse model of AD (APPsw transgenic mice) after housing with environmental enrichment for 4 months starting at 16 months of age. Interestingly, this improvement was independent of amyloid plaque deposition [34]. Taken together, it would be interesting to house 5XFAD mice from one to three months of age in environmentally enriched housing before testing their cognitive abilities using the HM and assessing their Aβ deposits and astrogliosis in the hippocampus and frontal and entorhinal cortex.

Several recent studies have reported that gut microbial alterations influence not only gut disorders but also central nervous system disorders such as AD [35]. Indeed, exposure to pathogenic microbial infections, antibiotics, probiotics, or fecal microbiota transplantation in germ-free and control animals suggests a clear role for the gut microbiota in AD-related pathogenesis [35, 36]. A future study could evaluate, in 5XFAD mice, the effect of gut microbiota on developing Aβ deposits (in germ-free animals in comparison to mice exposed to microbiota dysbiosis) and analyze cognitive performance with the HM. Finally, although we reached our goal with these experiments, i.e., the validation of new apparatus with an original paradigm for the detection of early cognitive deficits in mouse model of AD with a minimum of experimental animals, as recommended by the Ethic Committee, we cannot exclude that one of the most important limitation of the study is related to the low sample size and thus the importance of interindividual variability. Nevertheless, using appropriate statistical analyses, we validated our experimental conditions.

To conclude, we have described here an in-depth analysis of hippocampal memory impairment in a small cohort of young 5XFAD mice using a powerful new apparatus. This apparatus could open new avenues to validate the efficacy of early AD treatments that target the amyloid-dependent pathological pathway of the disease as well as the efficacy of treatments for other neurodegenerative diseases.

Footnotes

ACKNOWLEDGMENTS

This work was supported by funding from the CNRS and Aix-Marseille Université and by the Fondation pour la Recherche Médicale (FRM) and Fondation Alzheimer (ALZ201912009627) to SR.

Authors’ disclosures available online ().

References

Eid

, Mhatre

, Richardson

(2019) Gene-environment interactions in Alzheimer’s disease: A potential path to precision medicine. Pharmacol Ther 199, 173–187.

Hampel

, Prvulovic

, Teipel

, Jessen

, Luckhaus

, Frölich

, Riepe

, Dodel

, Leyhe

, Bertram

(2011) The future of Alzheimer’s disease: The next 10 years. Prog Neurobiol 95, 718–728.

Altomari

, Bruno

, Laganà

, Smirne

, Colao

, Curcio

, DiLorenzo

, Frangipane

, Maletta

, Puccio

(2022) A comparison ofbehavioral and psychological symptoms of dementia (BPSD) and BPSDsub-syndromes in early-onset and late-onset Alzheimer’s disease. J Alzheimers Dis 85, 691–699.

Pais

, Martinez

, Ribeiro

, Loureiro

, Fernandez

, Valiengo

, Canineu

, Stella

, Talib

, Radanovic

(2020) Early diagnosis and treatment of Alzheimer’s disease: New definitions and challenges. Braz J Psychiatry 42, 431–441.

Sasaguri

, Nilsson

, Hashimoto

, Nagata

, Saito

, De Strooper

, Hardy

, Vassar

, Winblad

, Saido

(2017) APP mouse models for Alzheimer’s disease preclinical studies. EMBO J 36, 2473–2487.

Oakley

, Cole

, Logan

, Maus

, Shao

, Craft

, Guillozet-Bongaarts

, Ohno

, Disterhoft

, Van Eldik

(2006) Intraneuronal β-amyloid aggregates, neurodegeneration, and neuron loss in transgenic mice with five familial Alzheimer’s disease mutations: Potential factors in amyloid plaque formation. J Neurosci 26, 10129–10140.

Ohno

, Cole

, Yasvoina

, Zhao

, Citron

, Berry

, Disterhoft

, Vassar

(2007) BACE1 gene deletion prevents neuron loss and memory deficits in 5XFAD APP/PS1 transgenic mice. Neurobiol Dis 26, 134–145.

Ohno

(2006) Genetic and pharmacological basis for therapeutic inhibition of band γ-secretases in mouse models of Alzheimer’s memory deficits. Rev Neurosci 17, 429–454.

Ohno

(2009) Failures to reconsolidate memory in a mouse model of Alzheimer’s disease. Neurobiol Learn Mem 92, 455–459.

10.

Girard

, Jacquet

, Baranger

, Migliorati

, Escoffier

, Bernard

, Khrestchatisky

, Féron

, Rivera

, Roman

(2014) Onset ofhippocampus-dependent memory impairments in 5XFAD transgenic mousemodel of Alzheimer’s disease. Hippocampus 24, 762–772.

11.

Baranger

, Marchalant

, Bonnet

, Crouzin

, Carrete

, Paumier

J-M

, Py

, Bernard

, Bauer

, Charrat

(2016) MT5-MMP is a new pro-amyloidogenic proteinase that promotes amyloid pathology and cognitive decline in a transgenic mouse model of Alzheimer’s disease. Cell Mol Life Sci 73, 217–236.

12.

Hongpaisan

, Sun

M-K

, Alkon

(2011) PKC ɛ activation prevents synaptic loss, Aβ elevation, and cognitive deficits in Alzheimer’s disease transgenic mice. J Neurosci 31, 630–643.

13.

Kimura

, Ohno

(2009) Impairments in remote memory stabilization precede hippocampal synaptic and cognitive failures in 5XFAD Alzheimer mouse model. Neurobiol Dis 33, 229–235.

14.

O’Leary

, Stover

, Mantolino

, Darvesh

, Brown

(2020) Intact olfactory memory in the 5xFAD mouse model of Alzheimer’s disease from 3 to 15 months of age. Behav Brain Res 393, 112731.

15.

Escoffier

, Migliorati

, Rahrah

, Roman

, Khrestchatisky

, Roman

(2021) The Helico Maze allows testing of early learning and subcategories of long-term memory in mice. Behav Brain Res 406, 113242.

16.

, Bonnet

, Bernard

, Marchalant

, Charrat

, Checler

, Khrestchatisky

, Baranger

, Rivera

(2014) Differential spatio-temporal regulation of MMPs in the 5xFAD mouse model of Alzheimer’s disease: Evidence for a pro-amyloidogenic role of MT1-MMP. Front Aging Neurosci 6, 247.

17.

Percie du Sert

, Hurst

, Ahluwalia

, Alam

, Avey

, Baker

, Browne

, Clark

, Cuthill

, Dirnagl

(2020) The ARRIVE guidelines 2.0: Updated guidelines for reporting animal research. J Cereb Blood Flow Metab 40, 1769–1777.

18.

Baranger

, Giannoni

, Girard

, Girot

, Gaven

, Stephan

, Migliorati

, Khrestchatisky

, Bockaert

, Marchetti-Gauthier

(2017) Chronic treatments with a 5-HT4 receptor agonist decrease amyloid pathology in the entorhinal cortex and learning and memory deficits in the 5xFAD mouse model of Alzheimer’s disease. Neuropharmacology 126, 128–141.

19.

Arifin

, Zahiruddin

(2017) Sample size calculation in animal studies using resource equation approach. Malays J Med Sci 24, 101.

20.

Girard

, Baranger

, Gauthier

, Jacquet

, Bernard

, Escoffier

, Marchetti

, Khrestchatisky

, Rivera

, Roman

(2013) Evidence for early cognitive impairment related to frontal cortex in the 5XFAD mouse model of Alzheimer’s disease. J Alzheimers Dis 33, 781–796.

21.

Garcia-Gonzalez

, Paumier

J-M

, Louis

, Pilat

, Bernard

, Stephan

, Jullien

, Checler

, Nivet

, Khrestchatisky

(2021) MT5-MMP controls APP and β-CTF/C99 metabolism through proteolytic-dependent and -independent mechanisms relevant for Alzheimer’s disease. FASEB J 35, e21727.

22.

García-González

, Paumier

, Louis

, Pilat

, Bernard

, Stephan

, Jullien

, Checler

, Nivet

, Khrestchatisky

(2021) MT5-MMP controls APP and β-CTF/C99 metabolism throughproteolytic-dependent and-independent mechanisms relevant forAlzheimer’s disease. FASEB J 35, e21727.

23.

Baranger

, Bonnet

, Girard

, Paumier

J-M

, García-González

, Elmanaa

, Bernard

, Charrat

, Stephan

, Bauer

(2017) MT5-MMP promotes Alzheimer’s pathogenesis in thefrontal cortex of 5xFAD mice and APP trafficking. Front Mol Neurosci 9, 163.

24.

Lampe

, Bird

, Nochlin

, Nemens

, Risse

, Sumi

, Koerker

, Leaird

, Wier

, Raskind

(1994) Phenotype of chromosome 14–linked familial Alzheimer’s disease in a large kindred. Ann Neurol 36, 368–378.

25.

Tedde

, Nacmias

, Ciantelli

, Forleo

, Cellini

, Bagnoli

, Piccini

, Caffarra

, Ghidoni

, Paganini

(2003) Identification of new presenilin gene mutations in early-onset familial Alzheimer disease. Arch Neurol 60, 1541–1544.

26.

Schellenberg

, Bird

, Wijsman

, Orr

, Anderson

, Nemens

, White

, Bonnycastle

, Weber

, Alonso

(1992) Genetic linkage evidence for a familial Alzheimer’s disease locus on chromosome 14. Science 258, 668–671.

27.

St George-Hyslop

, Haines

, Farrer

, Polinsky

, Van Broeckhoven

, Goate

, McLachlan

, Orr

, Bruni

, Sorbi

(1990) Genetic linkage studies suggest that Alzheimer’s disease is not a single homogeneous disorder. Nature 347, 194–197.

28.

Giannoni

, Gaven

, De Bundel

, Baranger

, Marchetti-Gauthier

, Roman

, Valjent

, Marin

, Bockaert

, Rivera

(2013) Early administration of RS 67333, a specific 5-HT4 receptor agonist, prevents amyloidogenesis and behavioral deficits in the 5XFAD mouse model of Alzheimer’s disease. Front Aging Neurosci 5, 96.

29.

Stoccoro

, Coppedè

(2018) Role of epigenetics in Alzheimer’s disease pathogenesis. Neurodegener Dis Manag 8, 181–193.

30.

Diamond

, Krech

, Rosenzweig

(1964) The effects of an enriched environment on the histology of the rat cerebral cortex. J Comp Neurol 123, 111–119.

31.

Van Loo

, Van de Weerd

, Van Zutphen

, Baumans

(2004) Preference for social contact versus environmental enrichment in male laboratory mice. Lab Anim 38, 178–188.

32.

Van Praag

, Kempermann

, Gage

(2000) Neural consequences of environmental enrichment. Nat Rev Neurosci 1, 191–198.

33.

Jankowsky

, Xu

, Fromholt

, Gonzales

, Borchelt

(2003) Environmental enrichment exacerbates amyloid plaque formation in a transgenic mouse model of Alzheimer disease. J Neuropathol Exp Neurol 62, 1220–1227.

34.

Arendash

, Garcia

, Costa

, Cracchiolo

, Wefes

, Potter

(2004) Environmental enrichment improves cognition in aged Alzheimer’s transgenic mice despite stable β-amyloid deposition. Neuroreport 15, 1751–1754.

35.

Jiang

, Li

, Huang

, Liu

, Zhao

(2017) The gut microbiota and Alzheimer’s disease. J Alzheimers Dis 58, 1–15.

36.

Bostanciklioglu

(2019) The role of gut microbiota inpathogenesis of Alzheimer’s disease. J Appl Microbiol 127, 954–967.