Donecopride, a Swiss army knife with potential against Alzheimer's disease

Abstract

Background and Purpose

We recently identified donecopride as a pleiotropic compound able to inhibit AChE and to activate 5‐HT₄ receptors. Here, we have assessed the potential therapeutic effects of donecopride in treating Alzheimer's disease (AD).

Experimental Approach

We used two in vivo animal models of AD, transgenic 5XFAD mice and mice exposed to soluble amyloid‐β peptides and, in vitro, primary cultures of rat hippocampal neurons. Pro‐cognitive and anti‐amnesic effects were evaluated with novel object recognition, Y‐maze, and Morris water maze tests. Amyloid load in mouse brain was measured ex vivo and effects of soluble amyloid‐β peptides on neuronal survival and neurite formation determined in vitro.

Key Results

In vivo, chronic (3 months) administration of donecopride displayed potent anti‐amnesic properties in the two mouse models of AD, preserving learning capacities, including working and long‐term spatial memories. These behavioural effects were accompanied by decreased amyloid aggregation in the brain of 5XFAD mice and, in cultures of rat hippocampal neurons, reduced tau hyperphosphorylation. In vitro, donecopride increased survival in neuronal cultures exposed to soluble amyloid‐β peptides, improved the neurite network and provided neurotrophic benefits, expressed as the formation of new synapses.

Conclusions and Implications

Donecopride acts like a Swiss army knife, exhibiting a range of sustainable symptomatic therapeutic effects and potential disease‐modifying effects in models of AD. Clinical trials with this promising drug candidate will soon be undertaken to confirm its therapeutic potential in humans.

Abbreviations

Aβ₄₀

human amyloid β (1‐40)

Aβ₄₂

human amyloid β (1‐42)

AD

Alzheimer's disease

APP

amyloid precursor protein

Aβ

amyloid‐β

BDNF

brain‐derived neurotrophic factor

BSA

bovine serum albumin

GFAP

glial fibrillary acidic protein

IHC

immunohistochemistry

MWM

Morris Water Maze

NOR

novel object recognition

sAPPα

soluble APPα

What is already known

• Donecopride, a dual 5‐HT₄ receptor agonist and acetylcholinesterase inhibitor, displayed pro‐cognitive effects in mice.

What this study adds

• Donecopride exerts anti‐amnesic properties in two animal models of Alzheimer's disease.

What is the clinical significance

• Donecopride is currently under regulatory preclinical investigation for the treatment of Alzheimer's disease.

1. INTRODUCTION

For 20 years, no new drug has been marketed for the treatment of Alzheimer's disease (AD). Those that are available, mainly the https://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=2465 inhibitors (AChEIs), display only symptomatic and transient therapeutic benefits because their mechanism of action is to target an enzyme which is depleted as neurodegeneration progresses (Citron, 2010; Melnikova, 2007; Rosini, Simoni, Minarini, & Melchiorre, 2014). None of the many novel target/drug combinations evaluated in clinical trials to date have successfully produced an effective disease‐modifying effect. There are several possible explanations for these failures, including (a) the selection of patients in a clinical state that allows them to benefit from the potential treatment, (b) the development of relevant animal models of AD, and (c) the multifactorial origin of the disease, which warrants a multitarget approach (see Cavalli et al., 2008; Schmitt, Bernhardt, Moeller, Heuser, & Frolich, 2004). The last consideration has now been taken into account, as most current clinical trials of treatments for AD combine https://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=6599—the most widely used AChEI on the market—with the novel drug being tested (Frolich et al., 2019).

Our consortium pursues such a multitarget approach by using multitarget‐directed ligands, which are specifically designed to reach several targets and, consequently, to display synergistic effects without the risk inherent in drug combinations and drug–drug interactions. We recently chose to target both the https://www.guidetopharmacology.org/GRAC/ObjectDisplayForward?objectId=9, in order to exert a potential neuroprotective effect, and AChE (Lecoutey et al., 2014; Rochais et al., 2015), whose inhibition remains relevant when associated with a protective effect of the targeted enzyme (Wang & Zhang, 2019).

On the one hand, activation of 5‐HT₄ receptors, allows both the release of ACh (Kilbinger & Wolf, 1992) and the activation of an α‐secretase. The latter promotes non‐amyloidogenic cleavage of the amyloid precursor protein (APP) to generate the neurotrophic soluble APPα (sAPPα) protein in a manner that is detrimental for the formation of the neurotoxic amyloid‐β (Aβ) peptide (Cho & Hu, 2007; Maillet et al., 2003). On the other hand, inhibiting the catalytic active site of AChE preserves neurotransmitter release (Musial, Bajda, & Malawska, 2007). If this effect occurs through interaction with the peripheral anionic site of AChE, the enzyme is not able to form aggregates with soluble Aβ peptides, decreasing the neurotoxic feature displayed by amyloids (Inestrosa et al., 1996).

Our group recently succeeded in combining these activities into a single pleiotropic compound, https://www.guidetopharmacology.org/GRAC/LigandDisplayForward?ligandId=10584, which is potentially able to display both symptomatic and disease‐modifying therapeutic effects in the treatment of AD (Lecoutey et al., 2014; Rochais et al., 2015).

Donecopride is a selective, partial (48.3%) agonist, with nanomolar potency (Ki = 8.5 nM) at (h) 5‐HT₄ receptors, which promotes the release of soluble APPα in vitro and in vivo (Lecoutey et al., 2014; Rochais et al., 2015). Donecopride also behaves as a mixed‐type competitive inhibitor of (h)AChE (IC₅₀ = 16 nM) that is selective towards butyrylcholinesterase and that partly displaces propidium iodide from the peripheral anionic site, accounting for a potential inhibitory effect against AChE‐induced Aβ aggregation. Further, in NMRI mice in vivo, with an i.p. dose of 0.3 mg·kg⁻¹, donecopride displayed pro‐cognitive effects, as assessed by the novel object recognition (NOR) test, and anti‐amnesic effects by the reversal of scopolamine‐induced memory impairments, as assessed in the spontaneous alternation test (Lecoutey et al., 2014).

This pharmacological profile led us to test the therapeutic potential of donecopride against AD by using two complementary animal models: (a) the 5XFAD transgenic mouse model and (b) a mouse model of AβO intracerebroventricular injections. In 5XFAD mice, the more intense amyloid‐producing mouse model, we assayed the potential of donecopride to prevent amyloid production and associated cognitive impairment. The AβO‐injected mice were used to demonstrate the ability of donecopride to challenge the acute toxicity of amyloid and to be effective when administered orally. Finally, we attempted to correlate the observed in vivo activities of donecopride with a potential neuroprotective effect measured in cultures of rat hippocampal neurons, exposed to AβO. We here report the results of this study.

2. METHODS

2.1. Animals

All animal care and experimental procedures were conducted in accordance with National and European regulations (EU directive N° 2010/63) and approved protocols (D13‐055‐08 and B54‐547‐28). All efforts were made to minimize animal suffering and to reduce the number of mice used. All mice had access to food and water ad libitum and were housed under a 12 h light‐dark cycle (12 h‐12 h) at 22 ± 2°C. Animal studies are reported in compliance with the ARRIVE guidelines (Kilkenny, Browne, Cuthill, Emerson, & Altman, 2010) and with the recommendations made by the British Journal of Pharmacology.

The generation of transgenic 5xFAD mice has been described previously (Oakley et al., 2006). These mice carry mutations in both human amyloid precursor protein (APP695) and presenilin‐1 (PSEN1) genes. The APP gene harbors three Familial Alzheimer's Disease (FAD) mutations: Swedish (K670N, M671L), Florida (I716V), and London (V717I). The human PSEN1 gene harbors two FAD mutations: M146L and L286V. Human gene expression is regulated by neural‐specific elements of the mouse Thy1 promoter to drive neuronal specificity. The 5XFAD strain (B6/SJL genetic background) was maintained by crossing heterozygous transgenic mice with B6/SJL F1 breeders (Jackson Laboratories, Bar Harbor, Maine, USA). These mice were then derived in a C57BL/6 genetic background. F1 male mice resulting from crossing 5XFAD (B6/SJL) and C57BL/6 mice were then crossed with C57BL/6 females and the male offspring backcrossed more than eight times with C57BL/6 females as described by Baranger et al. (2016). Female 5XFAD heterozygous transgenic mice were used for the experiments (7 mice per group). Genomic DNA was extracted from mice tail tips to assess their genotype by PCR. All transgenic mice were bred in our animal facility. For soluble Aβ peptide treatments, C57Bl6/J male mice (3 months of age, 12 mice per group) were purchased from Janvier, France.

2.2. Experiments using 5XFAD mice

2.2.1. Chronic treatments in 5XFAD mice

Two‐month‐old 5XFAD mice received either drugs (1 mg kg^‐1) or vehicle solution i.p. twice a week for 3 months. At the end of the experiments, mice were anesthetized with a mixture of 100 mg kg^‐1 ketamine and 10 mg kg^‐1 xylazine in saline solution and perfused transcardially with PBS. Brains were quickly isolated on ice, the olfactory bulbs and cerebellum removed and the two hemispheres divided. One hemisphere (or microdissected areas) was frozen on dry ice and stored at ‐80°C for biochemical analysis, while the other was post‐fixed in 4 % PFA for immunohistochemistry (IHC).

2.2.2. Brain extract preparation

Frontal cortex, hippocampus and entorhinal cortex of 5XFAD mice and controls were thawed, weighed and homogenized in 4 volumes of Tris‐saline (50 mM Tris‐HCl pH = 7.4, 150 mM NaCl) with a protease inhibitor cocktail (Roche Applied Science, Meylan, France). The resulting homogenates were centrifuged at 54,000 x g for 20 minutes and supernatants (the "soluble fraction") collected and aliquoted for storage at ‐80°C. Pellets were resuspended by brief sonication in 10 volumes of 6 M guanidine HCl in 50 mM Tris‐HCl, pH = 7.6 and centrifuged again at 26,500 x g for 20 minutes. Supernatants (the "insoluble fraction") were aliquoted and stored at ‐80°C (Morishima‐Kawashima et al., 2000).

2.2.3. Quantification of Aβ₄₀ and Aβ₄₂

ELISA kits from IBL International (Hamburg, Germany) for the dosage of Aβ₄₀ (human amyloid β (1‐40) assay kit, #27713) or Aβ₄₂ (human amyloid β (1‐42) assay kit, #27719) were used according to the manufacturer's instructions. Reactions were read at 620 nm and 450 nm using an Infinite 2000 luminescence counter. The obtained values were normalized to the protein concentration of each sample, measured using a BCA protein assay (Sigma‐Aldrich).

2.2.4. Immunohistochemistry

Brain sections (30‐μm thick) were cut using a vibratome (Microm HM 650V, Thermo Scientific, Saint Herblain, France) and stored in cryoprotectant medium at ‐20°C. For the labeling of amyloid plaques, free‐floating tissue sections of frontal cortex, hippocampus and entorhinal cortex (coordinates from the bregma: frontal cortex = 1.98 mm, hippocampus = ‐1.94 mm, entorhinal cortex = ‐3.08 mm) were extensively washed in PBS and then and mounted on gelatin‐coated slides. After drying, slides were immersed for 10 min in a solution of thioflavin‐S (1% in PBS, #T1892‐25G, Sigma‐Aldrich, St. Quentin Fallavier, France) and sequentially dehydrated in ethanol 70°, 80°, 95° and 100° (2 min each). Slices were immersed for 5 min in two solutions of xylene, and a non‐aqueous mounting medium (EUKITT) was used before adding coverslips. For GFAP (glial fibrillary acidic protein) staining, free‐floating brain sections were incubated in blocking solution (PBS; 3% BSA; 0.1% Triton X‐100) for 1 hr. Then the slices were incubated with polyclonal rabbit anti‐GFAP (1:1000, Z0334, Dako, Les Ullis, France, RRID:AB_10013382) overnight at 4°C. A secondary Alexafluor 594 goat anti‐rabbit antibody (1:1000, A11012, Life Technologies, Saint Aubin, France, RRID:AB_141359) was added for 2 hr. All slices were finally mounted on poly‐lysine slides and cover slipped.

2.2.5. Image acquisition and analysis

Images were acquired with an AxioImager Z1 microscope (Carl Zeiss S.A.S., Marly le Roi, France), and blindly analyzed using the FIJI software. Quantification of amyloid plaques (Thioflavin S staining, 10X objective) is represented as number of particles per mm² and averaged from two different measurements. Briefly, background was removed using the “Subtract background” module of FIJI software with default set up (50 pixels), and the plugin Log3D (sigma X:3, sigma Y:3) with default threshold and Analyze particle function (0‐Infinity) were used to count the plaques in the region of interest (ROI). The number of detected particles was divided by the size of the ROI in mm². GFAPstained images (20X) were acquired maintaining constant exposure for all samples across single experiments. Quantification (area fractions) was averaged from two to three different measurements and results are expressed as percent of vehicle condition.

2.2.6. Novel object recognition test in 5XFAD mice

The cognitive performance of 5XFAD mice was tested after 3 months of treatment using the novel object recognition (NOR) test (Bevins & Besheer, 2006). Animals were extensively handled during drug treatment prior to the test onset. Each day, mice were allowed to familiarize with the test room for at least 1 h prior to the test.

Testing was carried out in a Plexiglas box (width: 35 cm, length: 20 cm, height: 20 cm) placed in a dimly lit room (12‐15 lux). On day 1 and 2, each mouse was habituated to the empty box for 10 min per day. On day 3, two objects (constructed out of plastic toys) were positioned in the cage, 5 cm away from the opposing walls. During the training session, each animal was placed between the two objects, facing the wall and then was allowed to explore the objects for 5 min. Mice were then returned to their home cage and 24 hr later went through a 5 min test session in which one of the two (familiar) objects was replaced by a new one (novel). The whole experiment was video‐recorded and object exploration [time spent by the mouse nose in contact with the object or by sniffing it at a distance ≤ 1 cm] was blindly measured. Two parameters were considered: 1) the exploration time (s) spent by the animal interacting with the two familiar objects during the training session and 2) the exploration time spent by the animal interacting with the novel object relative to the total exploration time ([novel/(familiar + novel)] × 100) during the test. A discrimination index was also calculated ([novel − familiar]/[familiar + novel]).

2.3. Experiments with soluble Aβ peptides in vivo

2.3.1. Preparation and stereotaxic injection of soluble Aβ peptides to C57BL/6 mice

Aβ_1‐42 was obtained from Bachem (ref H1368) (batch number: 1052301). The preparation of stable soluble Aβ_1‐42 peptides was performed according to Garcia, Youssef, Utvik, Florent‐Béchard, Barthélémy, Malaplate‐Armand, et al. (2010). The peptide preparation contained a mixture of stable trimers and tetramers of Aβ_1‐42, as well as monomeric forms of the peptide. All peptide preparations are characterized in terms of peptide composition and in vitro neurotoxicity. Single intra‐cerebroventricular injection of C57BL/6 mice with vehicle or soluble Aβ peptides, under anaesthesia (ip injection of a mixture of ketamine/xylazine at a dose of 110 and 15 mg kg^‐1, respectively), soluble Aβ peptides (1 μl) or vehicle (1 μl were injected into the right lateral ventricle. Injections were made using 10 μl Hamilton micro syringes fitted with a 26‐gauge needle. The procedure was terminated by a subcutaneous injection of metacam (analgesia) at a dose of 5 mg kg^‐1. Animals were then placed individually in their home‐cage and the cage was placed in a heated cabinet until the animal has fully recovered. Animals were carefully monitored to control recovery after anesthesia. Compound dosing (total volume of 200 μl) started one day before icv injection of soluble Aβ peptides and was continued for 19 consecutive days. Behavioural tests were performed according to the time‐schedule presented in Supplementary figure 1.

2.3.2. Y‐Maze test of C57BL/6 mice injected with soluble Aβ peptides

Behavioural tests were performed 1 h after dosing at day +4. Immediate spatial working memory performance was assessed by recording spontaneous alternation behaviour in a Y‐maze. The maze is made of opaque Plexiglas and each of the three arms is 40 cm long, 14 cm high, 10 cm wide and positioned at equal angles. The apparatus is placed in a homogeneously lit test room to obtain 12‐15 lux in all arms as well as in the central zone. Mice are placed in the middle of one arm and allowed to explore the maze freely during 5 min session. The series of arm entries are video recorded (Smart v3.0 software, Bioseb) and an arm entry is considered complete when the hind paws of the mouse are completely placed in the arm. Alternation is defined as successive entries into the 3 arms on overlapping triplet sets. The percentage of alternation is calculated as the ratio of actual total alternations to possible alternations, defined as the number of arm entries minus 2, multiplied by 100. Locomotor activity is also recorded and evaluated during this phase (e.g. motivational index) by monitoring average speed and total distance.

2.3.3. Morris Water Maze (MWM) test of C57BL/6 mice injected with soluble Aβ peptides

Habituation trials (visible platform) ‐ The MWM test is performed as described in (6). The experimental apparatus consists of a circular water tank (diameter = 100 cm; height = 50 cm) containing water at 21°C to a depth of 25 cm and rendered opaque to block the view past the water surface by adding an aqueous acrylic emulsion. A platform (diameter = 10 cm) is used and placed at the midpoint of a quadrant. The pool is located in a test room, homogeneously illuminated at 100 lux. The swimming path of the animals is recorded using a video tracking system. Mice are brought to the experimental room for at least 30 min prior to testing, to enable acclimation to the experimental room conditions. Navigation to a visible platform is carried out before place‐navigation, to evaluate visual and motor abilities of all mice. Mice are submitted to 4 trials of 60 s per day (during 2 consecutive days), with an inter‐trial interval of at least 1 hr. Once mice have found the platform, they are left alone on the platform for an additional time of 30 s. There are no additional cues in the room. The platform position and starting points are randomly distributed over all 4 quadrants of the pool. Mice that fail to find the platform after 60 s are guided to its location and placed on the platform for 30 s. Memory‐acquisition trials (learning trials with hidden platform) are performed during 5 consecutive days and used to reach a steady state of escape latency. Mice are brought to the experimental room for at least 30 min prior to testing, to enable acclimation to the experimental room conditions. The hidden platform is submerged 1 cm below the water surface and placed at the midpoint of one quadrant. The pool is located in a test room homogeneously illuminated at 100 lux and containing various prominent visual cues. The swimming paths, swimming distance, swimming speed and thigmotaxis are recorded using a video tracking system. Mice are submitted to 4 trials of 60 s per day, with an inter‐trial interval of at least 1 hr. The mice are allowed to swim freely for 60 s, left alone for an additional 30 s on the hidden platform and then returned to their home cage during the inter‐trial interval. Start positions (set at the border of quadrants) are randomly selected for each animal. In each trial, the time required to escape onto the hidden platform is recorded. Mice that fail to find the platform after 60 s are guided to and placed onto the platform for 30 s, before they are returned to their home cage. Memory‐retention trials (probe trials, no platform) are performed two days after the last training session. Mice are again acclimated to the experimental room for at least 30 min prior to testing. The platform is removed and each animal is allowed a free 60 s swim. During the trial, the time spent in the target quadrant, the time spend in the quadrant opposite, and the crossings over the former platform location are measured and monitored by video tracking.

2.3.4. NOR test of C57BL/6 mice injected with soluble Aβ peptides

Behavioural tests were performed 1 h after dosing at day +15/+16. One day before the cognitive test (i.e. at day +15), mice are habituated during a 10 minutes’ trial during which they are placed in an empty open field (12‐15 lux). The day of the cognitive test (i.e. day +16), animals are placed in the same open field and are allowed to explore freely two identical objects for a trial of five minutes (acquisition trial). Then the animals are returned in their home‐cage for an inter‐trial time of five minutes. During the retention trial, animals are allowed to explore two different objects: one familiar and one novel object. During this time, the experimenter, blind to the treatment, record the time the mouse is actively exploring each object. All trials are video recorded (Smart v3.0 software, Bioseb). A discrimination index is then generated: Discrimination index = (time exploring novel object – time exploring familiar object) / total exploration time.

2.4. Experiments with soluble Aβ peptides in neuronal primary culture

2.4.1. Primary culture of rat hippocampal neurons

Rat hippocampal neurons were cultured as described by Callizot, Combes, Steinschneider, & Poindron (2013). Pregnant female rats of 17 days gestation (Wistar; Janvier Labs France) were killed using deep anesthesia with CO₂ chamber followed by a cervical dislocation. Then, fetuses were removed from the uterus and immediately placed in ice‐cold L15 Leibovitz medium with 2% penicillin (10,000 U ml^‐1) and streptomycin (10 mg ml^‐1) solution (PS) and 1% bovine serum albumin (BSA). Hippocampal neurons were treated for 20 min at 37°C with a trypsin‐ EDTA solution at a final concentration of 0.05% trypsin and 0.02% EDTA. The dissociation was stopped by adding Dulbecco's modified Eagle's medium (DMEM) with 4.5 g l^‐1 of glucose, containing DNAse I grade II (final concentration 0.5 mg ml^‐1) and 10% fetal calf serum (FCS). Cells were mechanically dissociated by three forced passages through the tip of a 10‐ml pipette and then centrifuged at 515 x g for 10 min at 4°C. The supernatant was discarded, and the pellet was resuspended in a defined culture medium consisting of Neurobasal medium with a 2% solution of B27 supplement, 2 mM L‐glutamine, 2% of PS solution, and 10 ng ml^‐1 of brain‐derived neurotrophic factor (BDNF). Viable cells were counted in a Neubauer cytometer, using the Trypan blue exclusion test. Cells were seeded at a density of 20,000 cells per well in 96‐well plates precoated with poly‐L‐lysine and cultured at 37°C in an air (95%)‐CO2 (5%) incubator. The medium was changed every 2 days.

2.4.2. Test compounds and human Aβ_{1‐ 42} exposure

The hippocampal neurons were exposed with Aβ solutions after 17 days of culture. The Aβ_{1‐ 42} was prepared as described by Callizot et al. (2013). Briefly, Aβ_{1‐ 42} peptide was dissolved in the defined culture medium mentioned above, at an initial concentration of 40 μM. This solution was gently agitated for 3 days at 37°C in the dark and immediately used after being properly diluted in culture medium to the concentrations used (20 μM (plate 1 for MAP2/Tau stainings) or 2.5 μM (plate 2 for PSD95/Synaptophysin stainings) corresponding to 2 μM or 0.25 μM of soluble Aβ peptides respectively). Donepezil (1 μM) or donecopride (1nM to 1μM) were solved in culture medium and preincubated 1 hour before Aβ application. Aβ_{1‐ 42} preparation was added to a final concentration of 20 or 2.5 μM (= 2 μM or 0.25 μM of soluble Aβ peptides, evaluated by automatic WB) diluted in control medium in presence of the compounds. The assays were carried out in a 96 well plate (6 wells per condition).

2.4.3. Survival, neurite network and phospho tau evaluation

24 hours after intoxication, the hippocampal neurons were fixed by a cold solution of ethanol (95%) and acetic acid (5%) for 5 min at ‐20°C. After permeabilization with 0.1% of saponin, cells will be incubated for 2 hr with (1) a chicken polyclonal antibody anti microtubule‐associated‐protein 2 (MAP‐2, RRID:AB_2138153), diluted at 1/1000 in PBS containing 1% fetal calf serum and 0.1% of saponin (this antibody allows the specific staining of neuronal cell bodies and neurites; and allow the study of neuronal cell death and neurite network) or (2) a mouse monoclonal antibody anti‐phospho‐Tau (AT100, RRID:AB_223652) at dilution of 1/400 in PBS containing 1% fetal calf serum and 0.1% of saponin. These antibodies were revealed with Alexa Fluor 488 goat anti mouse IgG (RRID:AB_2532075) and Alexa Fluor 568 goat anti‐chicken IgG (SAB4600084, Sigma‐Aldrich) at 1/400 dilution in PBS containing 1% FCS, 0.1% saponin, for 1 hr at room temperature. For each condition, 30 pictures (representative of the whole well area) per well were taken using ImageXpress (Molecular Devices, RRID:SCR_016654) at 20X magnification. All images were taken with the same acquisition parameters. Analyses were performed automatically by using Custom Module Editor (Molecular Devices). The following endpoints were assessed: total number of neurons (neuron survival, number of MAP‐2 positive neurons), neurite network (in μm of MAP‐2 positive neurons), area of tau in neurons (μm², overlapping with MAP‐2 positive neurons).

2.4.4. Evaluation of synapses

24 hr after intoxication, the hippocampal neurons were fixed by a cold mixture of ethanol (95%) and acetic acid (5%) for 5 min at ‐20°C. After permeabilization with 0.1% of saponin, cells were incubated for 2 hr with: (1) a mouse monoclonal antibody anti post synaptic density 95kDa (PSD95, RRID:AB_300453) at a dilution of 1/100 in PBS containing 1% fetal calf serum and 0.1% of saponin or (2) a rabbit polyclonal antibody anti‐synaptophysin (SYN, RRID:AB_306124) at the dilution of 1/100 in PBS containing 1% fetal calf serum and 0.1% of saponin. These antibodies were revealed with Alexa Fluor 488 goat anti mouse IgG and Alexa Fluor 568 goat anti‐rabbit IgG at the dilution 1/400 in PBS containing 1% FCS, 0.1% saponin, for 1 hour at room temperature. For each condition, 40 pictures per well were taken using ImageXpress (Molecular Devices) with 40x magnification. All images were taken with the same acquisition parameters. Synapse evaluation were performed automatically by using Custom module editor (Molecular Devices). The total number of synapses (overlapping between PSD95/SYN) was then assessed.

2.5. Data and statistical analysis

The data and statistical analysis comply with the recommendations of the British Journal of Pharmacology on experimental design and analysis in pharmacology (Curtis et al., 2018). Studies were designed to generate groups of equal size, and mice were randomly assigned in the experimental groups. Group size may vary according to power analysis and expertise of the authors regarding the behavioural tests used: NOR test, Y‐maze test, and MWM test (Freret et al., 2012; Freret et al., 2017; Leger et al., 2013). Mice were submitted to behavioural experiments in a randomized sequential order. Behavioural recordings and immunofluorescence images were blindly analysed. Culture data were automatically acquired. Statistical analysis was undertaken only for studies where each group size was at least n = 5. Group size is the number of independent values, and statistical analysis was done using these independent values. All values are expressed as the mean ± SEM. Statistical analyses were performed using GraphPad Prism 8.01 (GraphPad Software, La Jolla, CA, USA, RRID:SCR_002798). Statistical tests used are indicated in the legends of the figures. In each dataset, Gaussian distribution was verified (Shapiro–Wilk normality test), and the homogeneity of sample variance was tested using both Brown–Forsythe's and Bartlett's tests. As long as no variance among groups was detected, ANOVA was performed. Grouped analyses were performed using two‐way ANOVA, followed by a Bonferroni's test for multiple comparisons. Paired group comparisons were processed by one‐way ANOVA (post hoc correction) or Student's t test. Only probability values P < .05 were considered statistically significant. For GFAP studies and following the recommendations of the BJP guidelines (Curtis et al., 2018), the GFAP data were log‐transformed to obtain Gaussian distribution and keep the group with an n of equal size (n = 7). As much as possible, untransformed data are presented. In the case of necessary normalization, the Y‐axis labelling and the legend explain the transformation of the data. No outliers were removed in the present study.

2.6. Materials

Donepezil (2,3‐dihydro‐5,6‐dimethoxy‐2‐[[1‐(phenylmethyl)‐4‐piperidinyl]methyl]‐1H‐inden‐1‐one) was purchased from Tocris Bioscience (R&D Systems Europe, Lille, France). Donecopride (1‐(4‐amino‐5‐ chloro‐2‐methoxyphenyl)‐3‐[1‐(cyclohexylmethyl)‐4‐piperidinyl] propan‐1‐one) and its fumarate salt were synthesized as described by Lecoutey et al. (2014). For i.p. injection, compounds were resuspended in DMSO (37.5 μg μl^‐1, stored at ‐20°C) and freshly diluted 1:250 in 0.9% NaCl prior to administration. Vehicle solution (0.9% NaCl, 0.4% DMSO) was used as control. For oral dosing, compounds were dissolved in 0.9% NaCl.

2.7. Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Harding et al., 2018), and are permanently archived in the Concise Guide to PHARMACOLOGY 2019/20 (Alexander, Christopoulos et al., 2019; Alexander, Fabbro et al., 2019).

3. RESULTS

3.1. Chronic administration of donecopride decreases the amyloid load in soluble and insoluble fractions of 5XFAD mouse brains

The 5XFAD mice reproduce several major features of AD amyloid pathology, including amyloid‐induced neurodegeneration, inflammation, and synaptic and behavioural dysfunction (Crouzin et al., 2013; Giannoni et al., 2013, 2016; Girard et al., 2014). In a first experiment conducted under the original B6/SJL genetic background (Giannoni et al., 2016; Oakley et al., 2006), female mice received vehicle or donecopride (1 mg·kg⁻¹; i.p.) twice a week for 3 months from 10 to 22 weeks of age. The treatment started when amyloid plaques began to appear in mouse brains (Crouzin et al., 2013). Aβ₄₀ and Aβ₄₂ were quantified from brain extracts in soluble and insoluble fractions (Figure 1). In both fractions, donecopride was able to significantly reduce Aβ₄₂ levels without any change in Aβ₄₀ levels.

Figure 1.

Effects of donecopride on amyloid load after 3 months of chronic treatment in soluble (a) and insoluble (b) fractions of 5XFAD mice brains. Results are expressed as mean ± SEM. Sample sizes: soluble species: vehicle (V)‐Aβ₄₀: n = 7; donecopride (D)‐Aβ₄₀: n = 7; V‐Aβ₄₂: n = 6; D‐Aβ₄₂: n = 7. Insoluble species: V‐Aβ₄₀: n = 7; D‐Aβ₄₀: n = 7; V‐Aβ₄₂: n = 7; D‐Aβ₄₂: n = 6. Our aim was to have n = 7 mice in each group, but technical issues meant we were unable to obtain adequate Aβ₄₂ quantification from two mice. *P < .05, significantly different from vehicle; unpaired Student's t test)

3.2. Chronic administration of donecopride prevents memory impairments and decreases amyloid plaques and astrogliosis in some brain regions in 5XFAD mice

In a second set of experiments conducted on mice with the C57BL/6J genetic background (more appropriate for behavioural studies), female 5XFAD mice received vehicle or donecopride (1 mg·kg⁻¹; i.p.) twice a week for 3 months from 11 to 22 weeks of age. The treatment started when amyloid deposits began to form in mouse brains (Giannoni et al., 2016; Oakley et al., 2006). At the end of treatment, mice were evaluated with the NOR test for long‐term memory performance. This test assesses recognition memory for objects, its human equivalent being the visual paired‐comparison task (Wallace, Ballard, & Glavis‐Bloom, 2015). The recognition of objects is mediated by the perirhinal cortex in rodents (Albasser, Davies, Futter, & Aggleton, 2009, pp. 21–23; Brown, Barker, Aggleton, & Warburton, 2012; Norman & Eacott, 2005), primates (Zeamer, Richardson, Weiss, & Bachevalier, 2015), and humans (Watson & Lee, 2013). AD pathology develops in the perirhinal and entorhinal cortex before it appears in the hippocampus (deToledo‐Morrell et al., 2004; Stoub et al., 2005; Taylor, Probst, Miserez, Monsch, & Tolnay, 2008). The visual paired‐comparison task detects memory deficits in patients with mild cognitive impairment (Lagun, Manzanares, Zola, Buffalo, & Agichtein, 2011) and predicts conversion from this condition to AD (Zola, Manzanares, Clopton, Lah, & Levey, 2013).

Habituation sessions were started 3 days after the end of treatment to avoid interference with acute effects. The interval between the training session and the test was 24 hr. During training sessions (Figure 2a), two‐way repeated‐measures ANOVA demonstrated that the animals presented neither preference for the objects nor difference due to the treatment. However, during test sessions (Figure 2b), a significant group difference appeared between 5XFAD‐vehicle and 5XFAD‐donecopride treated groups regarding object discrimination, with no treatment‐group difference and no interaction between treatment group and object. Vehicle‐treated mice presented impaired memory performance and were not able to discriminate novel objects from familiar objects, as shown by similar exploration times for both objects (Figure 2b) and by the non‐significant discrimination index (Figure 2c). Donecopride prevented cognitive impairment, as indicated by the higher time devoted to exploration of the novel object in mice treated with donecopride, in comparison with mice treated with vehicle (Figure 2b). The discrimination index of the compound‐treated group was different from zero (chance level) (Figure 2c). The total exploration time during the training session did not differ significantly between the two groups (Figure 2d), indicating that improvement of recognition memory in 5XFAD mice treated with donecopride was not secondary to an increase in exploratory activity.

Figure 2.

Effects of donecopride on long‐term memory performance after 3 months of chronic treatment in 5XFAD mice. NOR test with 5‐min exploration sessions and a 24‐hr intersession interval. Time spent by mice exploring left and right objects during training session (a) or familiar and novel objects during test session (b). *P < .05, significantly different from familiar object; two‐way repeated‐measures ANOVA with object and treatment as factors, followed by Bonferroni's test. (c) Discrimination index calculated by using exploration times in test session and formula [(novel − familiar)/(familiar+novel)]. Indices different from zero: *P < .05, significantly different from zero; one sample Student's t test. (d) Total exploration time in test session was not significantly different among groups (unpaired Student's t test). (a–d) Data are expressed as mean ± SEM; n = 7 mice per group

The frontal cortex, hippocampus, and entorhinal cortex were analysed as representative areas of 5XFAD mouse brains (frontal, median, and caudal sections) and of human brain regions that are affected by AD early and are highly enriched in amyloid deposits. Donecopride treatment significantly decreased the number of amyloid plaques compared with vehicle in the frontal cortex (reduction of 16 ± 3%; Figure 3a) and the entorhinal cortex (reduction of 24 ± 4%; Figure 3c). The number of amyloid plaques in the hippocampus of 5XFAD mice treated with donecopride was not significantly different from that in controls (Figure 3b).

Figure 3.

Effects of donecopride on amyloid plaque load (a–d) and astrogliosis (e–h) after 3 months of chronic treatment in 5XFAD mice. Brain slices were stained with thioflavin S (a–d) or an antibody directed against GFAP, a marker of astrocyte reactivity (e–h). Quantification of amyloid plaques or GFAP staining in frontal cortex (a, e), hippocampus (b, f), and entorhinal cortex (c, g) of mice treated with donecopride (1 mg·kg⁻¹, twice a week for 3 months) versus mice that received vehicle. Analysis of thioflavin S or GFAP staining was performed blindly; data are presented as the mean number of plaques per square millimetre or LOG of the percentage of fraction area covered by GFAP staining in two to three tissue sections from the same brain area/animal (FIJI software). Each dot represents one mouse (n = 7 mice per group); horizontal lines show means ± SEM of each group. *P < .05, significantly different from vehicle; unpaired Student's t test, log‐transformed data of GFAP staining are Gaussian distributed. Representative images of thioflavin S staining of amyloid plaques (d) or GFAP staining of astrocytes (h) in entorhinal cortex of 5XFAD mice treated with vehicle or donecopride. DIEnt, dorsal intermediate entorhinal field; Prh, perirhinal cortex; Thio S, thioflavin S; vSub, ventral subiculum

Immunohistochemical assessment of the astroglial cell population (GFAP staining) showed that chronic donecopride treatment significantly reduced astrogliosis in entorhinal brain slices of treated animals in comparison to controls (Figure 3g,h). Despite a tendency to reduce astrogliosis in the hippocampus and the frontal cortex, however, the effects of donecopride treatment did not reach statistical significance in these brain areas (Figure 3e,f).

Collectively, the data obtained in 5XFAD mice demonstrated that donecopride prevented long‐term deficits in recognition memory. This effect was associated with a decrease in amyloid pathology and in astrogliosis in some brain regions.

3.3. Subchronic administration of donecopride prevents memory impairments in C57BL/6 mice injected with soluble Aβ peptides

Donecopride was then investigated in a preclinical AD model with wild‐type C57BL/6J mice (3 months of age) challenged at Day 0 with an intracerebroventricular injection of soluble Aβ_1–42 peptides to induce neuronal death in vitro and cognitive decline in vivo (Desbene et al., 2012; Garcia et al., 2010; Kriem et al., 2005; Malaplate‐Armand et al., 2006; Pillot et al., 1999; Youssef et al., 2008). The neurotoxicity and associated cognitive deficits induced by soluble Aβ peptides involve the activation of neuroinflammatory processes, including activation of glial cells and production of pro‐inflammatory cytokines, such as IL‐1β and TNF‐α. Donecopride and donepezil (used as a control) were administered orally (by oral gavage) once a day from Day −1 to Day +17 at three different doses. At Day +4 (i.e., 4 days after disease induction), spatial working memory was assessed with the Y‐maze test. From Day +3 to Day +14, learning capacities and long‐term memory were investigated with the MWM assay. At Days +15/+16, recognition memory was investigated with the NOR test (see Figure S1 for time schedule of the experiments).

3.3.1. Spatial working memory (Y‐maze test)

The Y‐maze test assesses spatial working memory, which is mainly mediated by the prefrontal cortex (working memory) and the hippocampus (spatial component; Spellman et al., 2015; Yoon, Okada, Jung, & Kim, 2008) and is decreased in AD (Benjamin, Cifelli, Garrard, Caine, & Jones, 2015; Bianchini et al., 2014; Schroeter et al., 2012). The measurement of spontaneous alternation behaviour has been widely used by behavioural pharmacologists to assess spatial working memory, a component of short‐term memory (Sarter, Schneider, & Stephens, 1988). In its simplest form, spontaneous alternation behaviour is the tendency of mice to alternate their conventional non‐reinforced choice of maze arms on successive opportunities. The intracerebroventricular infusion of soluble Aβ peptides induces cognitive deficits in this test (Bouter et al., 2013; Desbene et al., 2012; Garcia et al., 2010; Youssef et al., 2008).

In the Y‐maze test, one‐way ANOVA revealed a significant treatment effect, regarding alternation behaviour. At a dose of 3 mg·kg⁻¹·day⁻¹, donecopride rescued soluble Aβ peptide‐induced impairment of spatial working memory, since treated mice exhibited alternation behaviour that was similar to that shown by vehicle mice but different from that of soluble Aβ peptide‐injected mice (Figure 4a). At 1 and 9 mg·kg⁻¹·day⁻¹ and in the presence of soluble Aβ peptides, donecopride‐treated mice exhibited an intermediate phenotype with an alternation behaviour that was similar to that of both vehicle mice and soluble Aβ peptides‐injected mice. In this experiment, donepezil also resulted in intermediate activity at a dose of 1 mg·kg⁻¹·day⁻¹ and was not able to fully rescue the soluble Aβ peptides‐induced impairment. The mice treated with donecopride alone at the dose of 9 mg·kg⁻¹ presented an alternation behaviour that was not significantly different from the vehicle‐treated mice. In all groups, the number of arm entries was similar (Figure 4b).

Figure 4.

Effects of test items on spatial working memory in the Y‐maze test. Mice were treated with increasing doses of compounds as described in the experimental procedures. Four days following intracerebroventricular infusion of either vehicle or soluble Aβ peptides, spontaneous alternation behaviour (a) and arm entries (b) were recorded in the Y‐maze test during a 5‐min trial. Data are presented as mean alternation behaviour (%) or mean of number of arm entries ± SEM. n = 12 mice per group, except in the (donecopride [9 mg·kg⁻¹]) group in which one mouse did not move. *P < .05, significantly different from vehicle control mice; ^# P < .05, significantly different from soluble Aβ peptide‐injected mice; one‐way ANOVA followed by Bonferroni's test. DPZ, donepezil

3.3.2. Learning capacities and long‐term memory (MWM test)

Spatial reference memory is a process that allows one to remember the spatial pathway to a specific location. It is mediated by the integrity of the hippocampus network in rodents (Broadbent, Squire, & Clark, 2006; Martin, de Hoz, & Morris, 2005; Mumby, Astur, Weisend, & Sutherland, 1999) and in humans (Bartsch et al., 2010; Goodrich‐Hunsaker, Livingstone, Skelton, & Hopkins, 2010), as shown on a virtual MWM task. Injection of soluble Aβ peptides intracerebroventricularly in mice induces a deficit in spatial reference memory in the MWM (Bouter et al., 2013; Desbene et al., 2012; Garcia et al., 2010; Youssef et al., 2008).

The learning capacities of mice were monitored for 5 consecutive days as indicated in Figure S1. Regarding escape latency (Figure 5a), two‐way repeated‐measures ANOVA revealed a time effect and a good matching of the measures, but not a treatment effect and no treatment × time interaction. Administration of donecopride at a dose of 3 mg·kg⁻¹·day⁻¹ resulted in improvement in learning capacities of these mice (escape latency difference at Day 5 compared with Day 1) compared with those of soluble Aβ peptides‐injected mice (difference from Day 1). Donepezil appeared to be inactive at a dose of 1 mg·kg⁻¹·day⁻¹ (difference from Day 1, Figure 5a). In absence of soluble Aβ peptides, donecopride at a dose of 9 mg·kg⁻¹·day⁻¹ did not alter the global learning performance of the mice (escape latency difference at Day 5 compared with Day 1).

Figure 5.

Effects of test items on learning capacities (a–c) and long‐term memory (d–f) in the MWM test. Mice were treated with increasing doses of compounds as described in the experimental procedures. (a–c) From Day +7 to Day +11 following intracerebroventricular injection of vehicle or soluble Aβ peptides, animals were trained to localize a hidden platform by using visual cues. The mean escape latency (a), mean swimming speed (b), and mean distance travelled (c) of four trials per day are presented for each group ± SEM (n = 12 mice per group). The probe test was performed 2 days after the last training session by removing the platform. The latency to target (i.e., time to perform the first cross over the former platform position) (c), the number of crossings over the former platform position (d), and the time in target quadrant (f) were determined. Data are presented as means ± SEM. Sample sizes: Vehicle, soluble Aβ, DPZ, and donecopride (3 mg·kg⁻¹ + Aβ): n = 12 mice per group; donecopride (1 and 9 mg·kg⁻¹ + Aβ): n = 11 mice per group; donecopride (9 mg·kg⁻¹ + vehicle): n = 10 mice. Floating mice were excluded from the analysis. *P < .05 significantly different from vehicle control mice, ^# P < .05, significantly different from soluble Aβ peptide‐injected mice; one‐way ANOVA followed by Bonferroni's test. DPZ, donepezil

Regarding the average swimming speed (Figure 5b), two‐way ANOVA with repeated measures revealed a significant time effect and a good matching of the measures but neither a treatment effect nor a treatment × time interaction, demonstrating no biased locomotor effect.

Regarding the average travelled distance (Figure 5c), two‐way ANOVA with repeated measures revealed a time effect and a good matching of the measures, but neither treatment effect nor a treatment × time interaction. Nevertheless, multiple comparisons demonstrated that the vehicle‐treated group was the only group to present a significant decrease in travelled distance.

Long‐term memory performance was determined 2 days after the last training session and 14 days after intracerebroventricular injection of vehicle or soluble Aβ peptides. During the probe trial, the time at first cross (i.e., latency to target, Figure 5d), the number of crossings over the former platform location (Figure 5e), and the time in target quadrant (Figure 5f) were determined. One‐way ANOVA revealed a group effect in the latency to target, and a significant effect of donecopride at 1 mg·kg⁻¹·day⁻¹. In the number of crossing, one‐way ANOVA revealed also a group effect and donecopride (at all doses) inhibited soluble Aβ peptide‐induced impairment of long‐term memory. In that paradigm, donepezil at 1 mg·kg⁻¹·day⁻¹ demonstrated no significant effect to prevent memory impairment, neither on latency to former platform location nor on number of crossings. The time spent in target quadrant was not significantly different between groups.

3.3.3. Recognition memory

As intracerebroventricular injection of soluble Aβ peptides in mice leads to impairment in the NOR test, recognition memory was evaluated in a paradigm similar to that performed for 5XFAD mice (see Section 2). The test was performed 16 days after intracerebroventricular injection of either vehicle or soluble Aβ peptides. During training session (Figure 6a), two‐way repeated‐measures ANOVA revealed neither an object effect, nor a treatment effect, but a good matching of the measures, and no treatment × object interaction. During test session (Figure 6b), two‐way repeated‐measures ANOVA revealed a clear object effect, no treatment effect, a good matching of the measures, and a Treatment × Object interaction. Pairwise multiple comparisons demonstrated that the presence of donecopride (at all doses) resulted in the inhibition of soluble Aβ peptides‐induced impairment of recognition memory (Figure 6b). This was confirmed by discrimination index values showing that the memory impairment of the soluble Aβ group was reversed by donecopride treatment (Figure 6c). The total exploration time was similar between the groups (Figure 6d).

Figure 6.

Effects of test items on soluble Aβ peptides‐induced cognitive deficits in the NOR test. C57/Bl6J mice were intracerebroventricularly injected at Day 0 as described in Section 2. After habituation trials performed 15 days after intracerebroventricular injection, mice were subjected to a NOR test at Day +16 with 5‐min exploration sessions and 5‐min intersession intervals. Time spent by mice exploring left and right objects during training session (a) or familiar and novel objects during test session (b). *P < .05, significantly different from familiar object; NS, not significant; two‐way repeated‐measures ANOVA with object and treatment as factors, followed by Bonferroni's test. (c) Discrimination index calculated by using exploration times in test session and formula [(novel − familiar)/(familiar+novel)]. *P < .05, indices significantly different from zero (one sample Student's t test). Different from: ^# P < .05, significantly different from soluble Aβ peptides‐injected mice; one‐way ANOVA followed by Bonferroni's test. DPZ, donepezil. (a–d) Data are mean ± SEM. The experiment was performed with n = 12 mice per group. Mice with total time exploration of less than 5 s in test session were excluded. Sample sizes: donecopride (1 mg·kg⁻¹): n = 11 mice; vehicle and donecopride (9 mg·kg⁻¹ + vehicle and 9 mg·kg⁻¹ + Aβ): n = 10 mice per group; soluble Aβ, DPZ, and donecopride (3 mg·kg⁻¹): n = 9 mice per group

This battery of behavioural tests demonstrated the ability of donecopride to protect mice from soluble Aβ peptides‐induced toxicity. Notably, donecopride proved its efficiency when administered orally in this series of experiments. Moreover, donecopride at a dose of 3 mg·kg⁻¹·day⁻¹ improved learning, even in the presence of soluble Aβ peptides.

3.4. Donecopride displays neuroprotective activities and reduces tau hyperphosphorylation in rat hippocampal neuronal cultures

The effects of donecopride on neuronal survival, the neurite network, and tau phosphorylation were investigated in rat primary hippocampal neurons exposed to soluble Aβ peptides (24‐hr exposure), with donepezil used as reference test compound.

3.4.1. Neuronal survival

An important loss of neurons, stained with microtubule‐associated protein 2 antibody followed the application of the soluble Aβ peptides (Figure 7a,d). Donepezil, used here as a positive control, protected neurons from death. Two doses of donecopride (100 and 500 nM) exerted neuroprotective effects to a similar extent as occurred with donepezil. However, the highest dose of donecopride was toxic, as the average survival was significantly lower than that observed in the soluble Aβ peptides condition

Figure 7.

Effects of treatment with soluble Aβ peptides in a primary culture of hippocampal neurons. Effects on survival of microtubule‐associated protein 2 (MAP‐2) neurons (a), neurite network (b), and phosphorylation of tau (AT100) (c) in MAP‐2 neurons. Donepezil (DPZ) served as a positive control. Results are expressed as mean ± SEM as a percentage of mean control. *P < .05, significantly different from soluble Aβ peptides condition; one‐way ANOVA followed by PLSD Fisher's test. (d) Representative images of MAP‐2 (red) and AT100 (green) staining of rat primary hippocampal neurons in the presence of vehicle, soluble Aβ peptides (20 μM), DPZ (1 μM), or donecopride (500 nM). Sample sizes: MAP‐2 neurons: n = 6 per group except for donecopride (100 nM and 1 μM) where n = 5 per group; neurite network: n = 6 per group except for DPZ and donecopride (1, 50, and 100 nM) where n = 5 per group; tau phosphorylation: n = 5 per group except for DPZ and donecopride (50 nM) where n = 6 per group. Our aim was to obtain n = 6 per group; however, some cultures were not viable

3.4.2. Neurite network

The total neurite network was markedly reduced following exposure to the soluble Aβ peptides (Figure 7b,d), but treatment with donepezil significantly preserved almost all of this network. Four concentrations of donecopride (50 nM, 100 nM, 500 nM, and 1 μM) also exerted protective effects on the neurite network.

3.4.3. Tau phosphorylation

Hyperphosphorylation of tau detected with the AT100 antibody was observed under application of soluble Aβ peptides (Figure 7c,d), but donepezil significantly mitigated this hyperphosphorylation. Similarly, donecopride (50, 100, and 500 nM) significantly reduced the hyperphosphorylation of tau in a dose‐dependent manner. Once again, this effect was lost with a higher concentration of donecopride.

3.5. Donecopride preserves the number of synapses and promotes the formation of new synapses in neuronal culture or cultures of rat hippocampal neurons

To investigate the formation of synapses, we assessed the distribution of presynaptic (synaptophysin) and postsynaptic markers (postsynaptic density protein 95; Figure 8a,b). A structure that tested positive for both postsynaptic density protein 95 and synaptophysin staining was considered a synapse. The stress resulting from the application of soluble Aβ peptides caused a reduction in the number of synapses (as previously shown by Callizot et al., 2013). Donepezil mitigated the loss of synapses. Similarly, donecopride also preserved the number of synapses in a concentration‐dependent manner from 5 nM to 1 μM.

Figure 8.

Donecopride protects synapses after treatment with soluble Aβ peptides in a primary culture of hippocampal neurons. Effect of donecopride on total number of synapses (a) and on number of synapses per neuron (c). Donepezil (DPZ) served as a positive control. Results are expressed as mean ± SEM as a percentage of mean control. *P < .05, significantly different from soluble Aβ peptides condition; one‐way ANOVA followed by PLSD Fisher's test. (b) Representative images of synaptophysin (red), postsynaptic density protein 95 (PSD95; green), and Hoechst (blue) staining of rat primary hippocampal neurons in the presence of vehicle, soluble Aβ peptides (20 μM), DPZ (1 μM), or donecopride (500 nM). Sample sizes: synapse number: n = 5 per group except for soluble Aβ and donecopride (50 nM) where n = 6 per group; synapse per neuron: n = 5 per group except for soluble Aβ where n = 6 per group. Our aim was to obtain n = 6 per group; however, some cultures were not viable

Because of the injury caused by the soluble Aβ peptides, the number of neurons was reduced in the experimental conditions, which led to a reduction in the number of synapses. Therefore, to estimate the number of synapses per neuron, we normalized the number of synapses by using the number of neurons in each condition. This adjustment showed that only donecopride significantly promoted the formation of new synapses, whereas donepezil had no effect (Figure 8b,c).

4. DISCUSSION

The anti‐amnesic properties previously displayed by donecopride towards the reversal of scopolamine‐induced memory impairments in NMRI mice were confirmed after its chronic administration in two different animal models of AD. In transgenic 5XFAD mice, 3‐month administration of donecopride (1 mg·kg⁻¹, i.p., twice a week) prevented the long‐term impairments of memory observed in the NOR test in non‐treated mice (Figure 2). This effect is similar to one that we previously reported with the reference 5‐HT₄ receptor agonist, https://www.guidetopharmacology.org/GRAC/LigandDisplayForward?tab=biology&ligandId=237, at the same dose of 1 mg·kg⁻¹ (Baranger et al., 2017; Giannoni et al., 2013). The use of the lowest active dose of donecopride and a twice a week administration protocol allowed the action of this compound to be maintained and to avoid desensitization of the 5‐HT₄ receptor that has been reported in long‐term administration of fluoxetine (Vidal, Valdizan, Mostany, Pazos, & Castro, 2009).

In a second mouse model of AD in which mice had a memory deficit after intracerebroventricular administration of soluble Aβ peptides, donecopride was orally administered for 18 days. The results confirmed the effect of donecopride in preventing recognition memory impairments that non‐treated animals showed in a NOR test in a dose‐dependent manner from 1 to 9 mg·kg⁻¹ (Figure 6). This activity is comparable to that shown by donepezil in the same conditions. In the same model, donecopride (3 mg·kg⁻¹) also prevented deficits in spatial working memory, a component of short‐term memory, as shown in the Y‐maze test (Figure 4a), and in spatial reference memory, as tested in the MWM task. In the latter task, the effect of donecopride concerned both the improvement of learning capacities in mice, especially at a dose of 3 mg·kg⁻¹ (Figure 5a), and of long‐term memory performances at a dose of 1 mg·kg⁻¹ (Figure 5d,e). These effects are again similar or even better to those observed with donepezil in the same conditions. The absence of effect of donecopride at a dose of 9 mg·kg⁻¹·day⁻¹ to counteract Aβ‐induced learning deficits (difference from Day 1, Student's t test; NS) could be explained as the desensitization of 5‐HT₄ receptors, which is extremely rapid in neurons when the receptor is challenged with high doses of agonists (Ansanay, Sebben, Bockaert, & Dumuis, 1992). We have previously demonstrated that this effect is mediated first by a GPCR kinase 2‐dependent uncoupling of the receptor, followed by its β‐arrestin‐dependent endocytosis (Barthet et al., 2005). Moreover, because the 5‐HT₄ receptor is not recycled, new biosynthesis of the receptor is necessary for its return to the plasma membrane, and repetitive high administration of an agonist could lead to a decrease or even a loss of response (Barthet et al., 2005). We previously reported such a bell‐shaped curve while investigating the effects of donecopride on scopolamine‐induced impairment during the spontaneous alternation test (Rochais et al., 2015).

On the other hand, the chronic administration of donecopride in 5XFAD mice was accompanied by a decrease in amyloid load in soluble and insoluble brain fractions (Figure 1). We suggest that donecopride decreased the production of Aβ through the non‐amyloidogenic cleavage of APP and the consequent promotion of sAPPα production, as we have previously demonstrated both in vitro (Lecoutey et al., 2014) and in vivo (Rochais et al., 2015). This anti‐amyloid effect also resulted in a decrease in the number of amyloid plaques, as measured in the frontal and entorhinal cortex, but not in the hippocampus. The entorhinal cortex also showed a decrease in astrogliosis, which was not significant in the other brain areas (Figure 3). We have previously reported that in this mouse model, because of the intense amyloid production, treatment with 5‐HT₄ receptor agonists should be initiated at the very early stages (4 weeks old) to decrease the Aβ load and neuroinflammation in the hippocampus (Giannoni et al., 2013).

Studies on the toxic role of Aβ have now shifted from insoluble fibrils to smaller, soluble peptide aggregates. Aβ peptides are produced as soluble monomers and undergo oligomerization and amyloid fibril formation via a nucleation‐dependent polymerization process (Jan, Hartley, & Lashuel, 2010). During the course of in vitro Aβ fibril formation, various nonfibrillar intermediary aggregates are formed; collectively called “soluble oligomers and protofibrils,” they have been shown to precede the appearance of fibrils. Increasing evidence from various sources points to soluble Aβ peptides/protofibrils as putative toxic species in AD pathogenesis (Sakono & Zako, 2010). By generating a solution of soluble Aβ peptides and adjusting its concentration and neuron exposure time, investigators reproduced early effects (oxidative stress) on neurons and long‐term development of structural alterations (neuronal death; Callizot, Combes, Steinschneider, & Poindron, 2013).

In such a test, we demonstrated here for the first time that the other main molecular effect involved in the pathogenesis of AD was also affected by donecopride, as this compound decreased the phosphorylation status of tau in a dose‐dependent manner starting at 50 nM in a cellular model of rat hippocampal neurons injured with AβO (Figure 7c,d). ). The loss of such an activity at higher concentration may suggest that 5‐HT₄ receptor activation was involved in the decrease of tau phosphorylation. The endocytosis of the receptor could be responsible for the loss of this effect at 1 μM.

In terms of possible cellular mechanisms, donecopride appeared to display a neuroprotective role, as illustrated by the preservation of the number of neurons and synapses and the integrity of the neurite network (Figures 7a,b,d and 8a,b). The loss of this neuroprotective effect at 1 μM could be again explained by the desensitization of 5‐HT₄ receptors challenged with high doses of agonists. Such a bell‐shaped curve was seen in another model of primary neuronal culture exposed to increasing doses of a 5‐HT₄ receptor agonist (Cho & Hu, 2007).

Beyond protection, donecopride seemed to exert a genuine neurotrophic effect by favouring the formation of new synapses (Figure 8c), which was not observed with donepezil. This property, which may be linked to the promotion of learning, could be of extreme interest to patients with AD who face the loss of neurons, by exerting a positive effect that is not achieved with donepezil, the current medication for treatment of AD.

Finally, this work provides a proof of concept that chronic administration of donecopride targets both AChE (its catalytic and peripheral sites) and 5‐HT₄ receptors, resulting in the alleviation of two major features of AD, amyloid aggregation and tau hyperphosphorylation. This targeting leads to neuroprotective and neurotrophic cellular consequences, which ultimately correlate with a clear improvement in working and long‐term (recognition and spatial) memory capacities. We conclude that donecopride, or other compounds exhibiting similar Swiss army knife pleiotropic profiles, may on the one hand exert symptomatic sustained therapeutic effects through the restoration of cholinergic neurotransmission and, on the other, provide disease‐modifying effects through its neurotrophic properties. As a result of these beneficial actions, even partial recovery of cognitive deficits in Alzheimer's patients could markedly relieve both the heavy health care cost and the social impact on families and caregivers. These challenges will soon be tested in clinical trials.

AUTHOR CONTRIBUTIONS

C.R., T.F., M.B., P.D., and S.C. designed the research; C.L., K.H., P.G., F.G., E.C., and K.B. performed the research; C.R., T.F., J.B., S.M., S.R., M.B., P.D., and S.C. analysed data; and C.R., S.C., and P.D. wrote the paper.

CONFLICT OF INTEREST

The authors declare no conflict of interest.

DECLARATION OF TRANSPARENCY AND SCIENTIFIC RIGOUR

This Declaration acknowledges that this paper adheres to the principles for transparent reporting and scientific rigour of preclinical research as stated in the BJP guidelines for https://bpspubs.onlinelibrary.wiley.com/doi/abs/10.1111/bph.14207, https://bpspubs.onlinelibrary.wiley.com/doi/abs/10.1111/bph.14208, and https://bpspubs.onlinelibrary.wiley.com/doi/abs/10.1111/bph.14206, and as recommended by funding agencies, publishers, and other organizations engaged with supporting research.

Supporting information

Data S1. Supporting Information

Rochais C, Lecoutey C, Hamidouche K, et al. Donecopride, a Swiss army knife with potential against Alzheimer's disease. Br J Pharmacol. 2020;177:1988–2005. 10.1111/bph.14964