Early Astrocytic Dysfunction Is Associated with Mistuned Synapses as well as Anxiety and Depressive-Like Behavior in the AppNL-F Mouse Model of Alzheimer's Disease

Abstract

Background: Alzheimer's disease (AD) is the most common neurodegenerative disease. Unfortunately, efficient and affordable treatments are still lacking for this neurodegenerative disorder, it is therefore urgent to identify new pharmacological targets. Astrocytes are playing a crucial role in the tuning of synaptic transmission and several studies have pointed out severe astrocyte reactivity in AD. Reactive astrocytes show altered physiology and function, suggesting they could have a role in the early pathophysiology of AD.

Objective: We aimed to characterize early synaptic impairments in the AppNL-F knock-in mouse model of AD, especially to understand the contribution of astrocytes to early brain dysfunctions.

Methods: The AppNL-F mouse model carries two disease-causing mutations inserted in the amyloid precursor protein gene. This strain does not start to develop amyloid-β plaques until 9 months of age. Thanks to electrophysiology, we investigated synaptic function, at both neuronal and astrocytic levels, in 6-month-old animals and correlate the synaptic activity with emotional behavior.

Results: Electrophysiological recordings in the hippocampus revealed an overall synaptic mistuning at a pre-plaque stage of the pathology, associated to an intact social memory but a stronger depressive-like behavior. Astrocytes displayed a reactive-like morphology and a higher tonic GABA current compared to control mice. Interestingly, we here show that the synaptic impairments in hippocampal slices are partially corrected by a pre-treatment with the monoamine oxidase B blocker deprenyl or the fast-acting antidepressant ketamine (5 mg/kg).

Conclusions: We propose that reactive astrocytes can induce synaptic mistuning early in AD, before plaques deposition, and that these changes are associated with emotional symptoms.

Keywords: Alzheimer’s disease; App knock-in mice; LTP; MAO-B; depression; synapse.

Fig. 1

Reactive-like astrocytes in 6-month-old App^NL-F mice. Confocal z-stack images of immunohistochemical labelled GFAP-positive astrocytes in area CA1 of the hippocampus were used to analyze astrocytes morphology. Magenta: GFAP, cyan: DAPI (A, left panel). Images were deconvoluted and astrocytes were segmented and isolated based on connexin staining. Representative of analyzed image are displayed in A, right panel. White arrows show a representative astrocyte. Scale bar for both view: 20 μm. Sholl analysis revealed an increased number of intersections from 7 to 25 μm from the soma (B) Mann-Whitney test: ^***p < 0.001 significant from C57Bl6. The average volume occupied by the astrocyte (C) was also increased in App^NL-F compared to C57Bl6 mice (Mann-Whitney test: ^**p < 0.01 significantly different as shown). For all measurements, C57Bl6: n = 20 cells in three animals, App^NL-F: n = 22 cells in three animals. Quantification of the vimentin positive signal (D) revealed an increased signal in App^NL-F mice compared to C57Bl6 animals (Mann-Whitney test: ^***p < 0.001). E depicts representative vimentin signal for both genotypes. Scale bar: 20 μm. C57Bl6: n = 9 images from 3 mice. App^NL-F: n = 12 images form 4 mice. Tonic GABA currents were recorded in pyramidal cells in the presence of NBQX and DL-AP5. Representative traces with baseline before blocking GABA is shown as a dashed line (E). Average tonic GABA currents are significantly higher in App^NL-F mice compared to C57Bl6 (F) Mann-Whitney test: ^*p < 0.05 significantly different as shown. C57Bl6: n = 7 cells recorded in five animals; App^NL-F: n = 7 cells recorded in four animals.

Fig. 2

Mistuned synapses in 6-month-old App^NL-F mice. Average resting membrane potential of pyramidal neurons of area CA1 of the hippocampus is significantly higher in App^NL-F mice compared to C57Bl6 mice (A). Recordings of mEPSC in the same cells show reduced frequency of events (B) in App^NL-F compared to C57Bl6 mice, with no changes in amplitude (C). Representative traces for mEPSC are shown under graphs. Student t-test, ^**p < 0.01 statistically significant as shown. For all variables, C57Bl6: n = 10 cells recorded in seven animals, App^NL-F: n = 10 cells recorded in seven animals. fEPSP were recorded after Schaffer collaterals stimulation in the area CA3. We recorded the response of post-synaptic neurons in area CA1. A subthreshold theta-burst (θ-burst) stimulation was applied and fEPSP magnitude was monitored for 1 h. The fEPSP slope was normalized to the average baseline value and displayed as the average per minute (D). Representative traces are shown on top. The θ-burst induced a significant potentiation in both App^NL-F and C57Bl6 mice at 0–5 min after stimulation (E). However, at 55–60 min, the fEPSP magnitude is almost back to baseline in C57Bl6 mice and is significantly increased in App^NL-F mice (F). Student t-test: ^*p < 0.05, ^***p < 0.001 statistically significant as shown. C57Bl6: n = 6 recordings in three animals, App^NL-F: n = 9 recordings in five animals.

Fig. 3

Retuned synapses after MAO-B blocking by deprenyl in 6-month-old App^NL-F mice. Pre-treatment of hippocampal slices with the MAO-B blocker deprenyl significantly decreases resting membrane potential in pyramidal cells in App^NL-F mice (A). Blocking MAO-B increased both frequency (B) and amplitude (C) of mEPSC. Representative traces are shown under graphs. Student t-test and Mann-Whitney test: ^**p < 0.01; ^***p < 0.001 statistically significant as shown. App^NL-F + aCSF: n = 8 cells recorded in four animals; App^NL-F + deprenyl: n = 8 cells recorded in four animals. The subthreshold theta-burst (θ-burst) stimulation induced long term potentiation in both treated and untreated slices from App^NL-F mice (D). Representative traces are shown on top. When we compared the average fEPSP magnitude at 0–5 min and 55–60 min after stimulation, the potentiation was significantly lower after deprenyl pre-treatment (E, F). Mann-Whitney test: ^**p < 0.01, ^***p < 0.001 statistically significant as shown. App^NL-F + aCSF: n = 6 recordings in four animals, App^NL-F + deprenyl: n = 5 recordings in three animals.

Fig. 4

App^NL-F mice shown no social memory impairments but a lack of motivation. In the five-trials social memory test (A), App^NL-F mice show no memory deficits as seen by the progressive decrease in social interactions over four consecutive trials (B). On the fifth trial, when a new animal is presented, C57Bl6 mice regain interest and the time of interaction is increased, whereas the time of interaction is not different in the fifth trial versus the fourth one in App^NL-F mice. Two-way ANOVA for repeated measured with the genotype and the trial as the main factors. Bonferroni multiple comparison ^*p < 0.05 statistically different trial 4 versus trial 5. Latency to first investigation (C) is similar between genotypes in the first trial. However, it is significantly increased in App^NL-F on the fifth trial compared to C57Bl6 and to App^NL-F in trial 1. Two-way ANOVA for repeated measured with the genotype and the trial as the main factors. Bonferroni multiple comparison; Bonferroni multiple comparison: ^**p < 0.01: statistically different as shown. C57Bl6: n = 8, App^NL-F: n = 8. In the olfactory habituation/dishabituation test (D), we presented four different odors for three times 2 min each. The time of interaction with the probe was similar in C57Bl6 and App^NL-F at all timepoints and for all odors. C57Bl6: n = 9, App^NL-F: n = 10.

Fig. 5

App^NL-F mice display mild depressive-like behavior. In the open field test, the total travelled distance (A) is reduced in App^NL-F compared to C57Bl6, as well as the number of entries in the central area (B). The total time in the open area is not different between the two strains (C). Mann-Whitney test: ^**p < 0.01, ^***p < 0.001 statistically different as shown. C57Bl6: n = 8, App^NL-F: n = 8. In the elevated plus maze, the number of entries in the open arms is significantly lower in App^NL-F compared to C57Bl6 (D) whereas the total time spent in the open arms (E) remains unchanged between genotypes. Student t-test: ^*p < 0.05 statistically different as shown. C57Bl6: n = 8, App^NL-F: n = 8. In the forced swim test, the latency to first immobility is lower in App^NL-F mice compared to C57Bl6 (F), while the total time of immobility is higher (G). Mann-Whitney and student t-test: ^*p < 0.05, ^***p < 0.001 statistically different as shown. C57Bl6: n = 16, App^NL-F: n = 18. The emotionality z-score is an integrative score allowing comparison of multiple factors, calculated on different scales and includes parameters from the different behavioral tests described above (see supplementary information). The emotionality z-score is significantly higher in App^NL-F mice compared to C57BL6 (H). Student t-test: ^**p < 0.01 statistically significant as shown. C57Bl6: n = 21, App^NL-F: n = 21.

Fig. 6

Ketamine (5 mg/kg) partially restores emotionality in 6-month-old App^NL-F mice. As previously shown, App^NL-F mice do not have social memory impairments, as the social interaction progressively declines over four trials. This learning curve is not affected by injections of NaCl of Ketamine 24 h before testing (A). Moreover, ketamine does not significantly affect the lack of motivation (B). App^NL-F + NaCl: n = 16, App^NL-F + ketamine: n = 13. In the open field test, neither the total travelled distance (C), the number of entries in the open area (D) nor the total time in that zone (E) is affected by ketamine. App^NL-F + NaCl: n = 16, App^NL-F + ketamine: n = 13. A similar profile was obtained in the elevated plus maze in which neither the number of entries in the open arms (F) or the total time in that zone (G) are affected. App^NL-F + NaCl: n = 8, App^NL-F + ketamine: n = 8. In the forced swim test, the latency to first immobility (H) is unchanged after ketamine treatment while the total time of immobility (I) is reduced. Student t-test: ^*p < 0.05 statistically different as shown. App^NL-F + NaCl: n = 8, App^NL-F + ketamine: n = 8. Interestingly, the emotionality z-score (J) remains unchanged after ketamine 5 mg/kg treatment. App^NL-F + NaCl (left); n = 16, App^NL-F + ketamine (right); n = 16.

Fig. 7

Synaptic mistuning but not astrocyte morphology is restored in App^NL-F by a single dose of ketamine (5 mg/kg). A single injection of ketamine 24 h before immunohistochemistry analysis didn’t affect the morphology of astrocytes. The sholl analysis unveiled a strong complexity from 7 to 25 μm from the nucleus (B). (A) depicts representative images from the two treatment conditions. Magenta: GFAP, cyan: DAPI; White arrows show a representative astrocyte. Scale bar for both view: 20 μm. The volume occupied by the cell (convex hull, C) remained unchanged after ketamine injection. App^NL-F + NaCl: 122 cells from three animals. App^NL-F + ketamine: 159 cells from four animals. Ketamine did not affect resting membrane potential in App^NL-F mice (D). In the same cell, the frequency of mEPSC is increased by ketamine treatment (E), whereas the amplitude of mEPSC remains unchanged (F). Mann-Whitney test: ^**p < 0.01 statistically different as shown. App^NL-F + NaCl: n = 7 cells recorded in five animals, App^NL-F + ketamine: n = 6 cells recorded in six animals. With the same theta-burst (θ-burst) stimulation as previously used, the fEPSP magnitude goes back nearly to the baseline value in App^NL-F + ketamine mice whereas in the App^NL-F + NaCl, the stimulation triggers a long-term potentiation (G). Representative traces are shown on the top. There is no difference between in the fEPSP magnitude in the two groups at 0–5 min after stimulation (H). However, at 55–60 min after the stimulation the fEPSP magnitude is significantly higher in NaCl treated mice compared to ketamine treated mice (I) Student t-test: ^***p < 0.001 statistically different as shown. App^NL-F + NaCl: n = 5 recordings in four animals, App^NL-F + ketamine: n = 5 recordings in four animals. Tonic GABA currents were recorded (J) and the average current was significantly reduced in App^NL-F mice treated with ketamine compared to NaCl treated mice (K). Mann-Whitney test ^***p < 0.001 significantly different as shown. App^NL-F + NaCl: n = 8 cells recorded in four animals; App^NL-F + ketamine: n = 9 cells recorded in four animals.