Urolithin A Inhibits Anterior Basolateral Amygdala to Ventral Hippocampal CA1 Circuit to Ameliorate Amyloid-β-Impaired Social Ability

Abstract

Background:

Anxiety and social withdrawal are highly prevalent among patients with Alzheimer’s disease (AD). However, the neural circuit mechanisms underlying these symptoms remain elusive, and there is a need for effective prevention strategies.

Objective:

This study aims to elucidate the neural circuitry mechanisms underlying social anxiety in AD.

Methods:

We utilized 5xFAD mice and conducted a series of experiments including optogenetic manipulation, Tandem Mass Tag-labeled proteome analysis, behavioral assessments, and immunofluorescence staining.

Results:

In 5xFAD mice, we observed significant amyloid-β (Aβ) accumulation in the anterior part of basolateral amygdala (aBLA). Behaviorally, 6-month-old 5xFAD mice displayed excessive social avoidance during social interaction. Concurrently, the pathway from aBLA to ventral hippocampal CA1 (vCA1) was significantly activated and exhibited a disorganized firing patterns during social interaction. By optogenetically inhibiting the aBLA-vCA1 pathway, we effectively improved the social ability of 5xFAD mice. In the presence of Aβ accumulation, we identified distinct changes in the protein network within the aBLA. Following one month of administration of Urolithin A (UA), we observed significant restoration of the abnormal protein network within the aBLA. UA treatment also attenuated the disorganized firings of the aBLA-vCA1 pathway, leading to an improvement in social ability.

Conclusions:

The aBLA-vCA1 circuit is a vulnerable pathway in response to Aβ accumulation during the progression of AD and plays a crucial role in Aβ-induced social anxiety. Targeting the aBLA-vCA1 circuit and UA administration are both effective strategies for improving the Aβ-impaired social ability.

Keywords

Alzheimer’s disease amyloid-β anxiety aBLA-vCA1 circuit

INTRODUCTION

Alzheimer’s disease (AD) is a progressive dementing illness and one of the most common forms of dementia in aged persons [1]. The principal clinical features of AD are manifested through alterations in the brain such as neuritic plaques consisting of amyloid-β (Aβ) peptide and the occurrence of hyperphosphorylated forms of tau protein inside neurons. These are two of the various cerebral changes observed throughout the course of AD [2 –4].

Aβ is considered one of the hallmark features associated with early pathological changes in AD. Extensive research has provided substantial evidence regarding the molecular mechanisms underlying Aβ pathology and its impact on spatial memory dysfunction. These mechanisms [5 –10] include excitatory toxicity, neuronal suppression, reduction of synaptic-related proteins, impaired synaptic plasticity, autophagy deficits, mitochondrial dysfunction, inflammation, and more.

The core symptoms of AD primarily revolve around memory impairment and a decline in cognitive abilities. Nevertheless, besides these symbolic features, clinicians also observe additional clinical presentations in AD patients, including anxiety, depression, aggression, and more [11 –13]. Population and animal studies have consistently demonstrated a strong association between Aβ lesions and anxiety-like behaviors [14]. A study has indicated that compared with normal concentrations, individuals with abnormal Aβ concentrations are more prone to experiencing anxiety symptoms [15]. Besides, in another study, Aβ deposition and associated synaptic loss were found in the brain regions of 5xFAD mice, while the mice developed high levels of anxiety-like behavior [16]. Anxiety significantly impacts social interactions, and social withdrawal is highly prevalent among AD patients, even before diagnosis [17, 18]. Considering the significance of social interactions for the well-being of AD patients, it is crucial to investigate how Aβ disrupts social abilities during the progression of AD.

The basolateral amygdala (BLA) and hippocampus (HP) are vulnerable to Aβ [19, 20]. Numerous studies have consistently shown that the BLA and HP play crucial roles in the regulation of anxiety-like behaviors [21 –23]. When glutamatergic neurons are nonspecifically activated in the BLA, especially in its anterior part (aBLA), it can lead to excessive avoidance behavior in mice, resulting in an anxiety effect [24]. There are some reports indicating that optogenetics has confirmed the real-time control of anxiety-related behaviors by both the ventral hippocampus (vHP) and BLA. Additionally, combined electroencephalogram (EEG) recordings with machine learning have revealed the coherence of beta frequency between HP and amygdala [25]. Recent studies have also discovered a close association between vHP and social anxiety-like behaviors. The expression of anxiety-related behaviors in the elevated plus maze (EPM) and open-field test (OFT) requires the involvement of vHP, rather than the dorsal hippocampus [26, 27]. In in vivo electrophysiology recordings, it has been discovered that in anxiety-like behaviors, there is not only an increase in the firing rate of BLA [28] but also an increase in theta-frequency synchrony between vHP and the prefrontal cortex (PFC) [29]. These studies have established the neural correlation between anxiety-like behaviors and the BLA and vHP. The CA1 subregion of the ventral hippocampus exhibits a high density of neurons associated with anxiety. The activation of these neurons is triggered by environmental inputs that induce fear, and their role is crucial in facilitating avoidance actions [30]. It is known that the BLA and the ventral CA1 (vCA1) can collaboratively regulate anxiety states. Specifically, the BLA-vCA1 monosynaptic excitatory circuit can induce anxiety by increasing avoidance behaviors [31]. However, it remains unclear how the BLA-vCA1 circuit changes and contributes to AD-like social behavior dysfunction in the presence of Aβ.

Urolithin A (UA) is a natural metabolite of elagitannins [32], with its metabolization depending on the appropriate gut microbiome. However, changes in the microbiome with age result in limited production of UA. In comparison to younger individuals, the aging population produces less than half of UA [33]. Peripherally, UA has been shown to increase skeletal muscle respiratory capacity and improve the regenerative ability of muscle stem cells, thereby enhancing muscle function and increasing survival in a mouse model of Duchenne muscular dystrophy [34]. Furthermore, UA exhibits immunomodulatory effects in monocytic cells that help attenuate inflammation [35]. Administration of UA to tumor-bearing mice has demonstrated a robust anti-tumor CD8 + T cell immune response, while ex vivo UA-pre-treated T cells have displayed enhanced anti-tumor function upon adoptive cell transfer [36]. Through the normalization of mitochondrial function, UA has been reported to reduce sepsis-induced cardiac depression. However, whether and how UA exerts beneficial effects on Aβ pathology during AD process remains unclear.

In this study, we observed a significant accumulation of Aβ in the amygdala, specifically the anterior part (aBLA) of 5xFAD mice. Moreover, we found that the aBLA-vCA1 circuit exhibited overactivation and a disorganized firing pattern in the presence of Aβ, which contributed to the development of AD-like social avoidance. Intriguingly, treatment with UA remarkably improved the distorted protein network within the aBLA induced by Aβ and attenuated the excessive social avoidance caused by Aβ. Our findings underscore the promising potential of UA administration as a strategy tomitigate Aβ

MATERIALS AND METHODS

Animals

Adult male C57BL/6 mice (6-month-old) and 5xFAD mice (6-month-old) were purchased from GENEANDPENCE (Yanghzou, Jiangsu, China). The animals were housed in groups of four to five per cage and were housed under a 12-h light/dark cycle (lights on at 6 : 00 p.m., off at 6 : 00 a.m.) at a stable temperature (23–25°C). Food and water were given ad libitum. In the present study, we have complied with all relevant ethical regulations for the animal testing and research. All procedures were approved by institutional guidelines and the Animal Care and Use Committee (Huazhong University of Science and Technology, Wuhan, China) of the university’s animal core facility.

Stereotaxic surgery

Mice were anaesthetized with pentobarbital sodium (1%, 30 mg per kilogram) and immobilized in a stereotactic injection apparatus (RWD 68046 & 68055, China). After the scalp was disinfected and balanced, a tiny hole was punched very gently. Viruses were injected into the aBLA (AP: –1.34 mm, ML:±3.4 mm, DV: –4.8 mm) and vCA1 (AP: –3.28 mm, ML:±3.3 mm, DV: –4.6 mm) by a 10μL gastight miscrosyringe (KF019, China) under a stereotaxic instrument (World Precision Instruments, USA). The injection rate was 50–100 nL/min (for virus injection) or 30 nL/min (for CTB injection). After each injection, the needle was left for 5 min to avoid virus spread and aspiration. Lincomycin lidocaine gel was evenly applied to the skull, and then absorbable stitches were used to suture the skin. The mice were placed on a heated blanket waiting for wake-up. Finally, the accuracy of the injection site was confirmed by fluorescent expression. Mice with incorrect virus expression locations wereexcluded.

CTB retrograde tracing

Cholera toxin subunit B (recombinant) was obtained from BrainVTA (CTB-555). CTB (100nl) was delivered stereotaxically into vCA1 (AP: –3.28 mm, ML:±3.3 mm, DV: –4.6 mm). Seven days later, the mice were anaesthetized and perfused. The expression of CTB in aBLA was detected with a confocal microscope (LSM 780, ZEISS, Germany).

Optogenetic manipulation in free-moving mice

AAV5-CaMKIIa-eNpHR3.0-eYFP and AAV5-CaMKIIa-eYFP purchased from Brain Case Co., Ltd, Shenzhen, China. Virus (150–200 nL,>10¹² vg mL^–1) was injected into aBLA (AP: –1.34 mm, ML:±3.4 mm, DV: –4.8 mm) at a speed of 0.1μL min^–1. Virus injection was counterbalanced across the left and right aBLA. Approximately 4 weeks later, optical fibers (core = 200μm; numerical aperture = 0.37) were implanted in vCA1 (AP: –3.28 mm, ML:±3.3 mm, DV: –4.1 mm). Allowing 1 week for recovery, the mice then performed the behavioral tests.

Tandem mass tag (TMT) labeled proteome experiments

Each sample was digested into peptides with mass spectrometric trypsin (Promega, V5072), then labelled with different TMT (ThermoFisher 90406), mixed together and divided into 15 components by high performance liquid chromatography (HPLC) for subsequent experiments. The dried components were dissolved in 0.1% formic acid (FA), and captured with a silica gel capillary column filled with C18 resin (Varian, Lexington, MA, USA) for subsequent Q Exactive (Thermo Scientific, NJ, USA) mass spectrometer analysis, the specific parameters are set as follows: 400-1, 800 m/z, 70000 resolutions; MS/MS scans (100–1800 m/z). Using Proteome Discoverer 2.1 software (Thermo Scientific) to retrieve MS/MS data according to Uniport-human database.

Three-Chamber Sociability Test

The Three-Chamber Sociability Test was derived from the inherent tendency of mice to participate in social interaction and investigate unfamiliar stimuli. The test apparatus was constructed using a custom-made rectangular organic glass resin (60 cm×40 cm×30 cm, length×width×height). The dimensions of each chamber are 20 cm by 40 cm. Custom-designed resin cups (8 cm in diameter) were inverted and symmetrically placed in the two corners of chamber, allowing the accommodation of adolescent mice (3–4 weeks) while limiting their movement within a defined area. In the initial stage, the experimental mouse was positioned in the center of chamber and allowed 5 min for unrestricted exploration (Habituation). During the following 5 min, the test mouse was restricted to the center of chamber for a duration of 30 s, while an adolescent mouse (3–4 weeks) was placed in an inverted resin cup of the designated social chamber. A vacant resin cup was positioned on the side designated as the non-social chamber. Subsequently, the test mouse was left to explore chamber for 5 min. Each test mouse participated in social interaction test with optogenetic stimulation for two days. The positions of the adolescent mice and the sequence of light stimulation for the test mice were randomly assigned in a balanced manner. Social interaction behavior was characterized by the test mouse’s exploration within a 2 cm radius, specifically through sniffing or orienting its nose or head, around the inverted resin cup that housed the adolescent mouse. Examining the vacant inverted resin cup on the opposite side was regarded as non-social behavior. The complete experimental procedure was documented through the utilization of tracking software and a camera system. After each experiment, the bottom of the three chambers and the resin cups were thoroughly cleaned with 70% ethanol.

Fiber photometry

The change in neuronal activity in the behavioral tasks was assayed by recording GCaMP6f fluorescent signals with an optical fiber recording system (Thinker Tech Nanjing Biotech Limited Co., Ltd.). Three weeks after AAV-CaMKII-DIO-GCaMP6f virus was injected into the aBLA, an optical ceramic needle was inserted into the aBLA through the craniotomy. A 488 nm laser (0.01–0.02 mW) was delivered by an optical fiber recording system, and that fluorescent signals were recorded. For data analysis, the original Ca^2 + signal was demodulated and converted to df/f. Normalized df/f could monitor the activity alterations in the aBLA–vCA1 connection. Motion tracking and manual tagging were used to monitor the activity and mark the position of mice when they were experiencing three-chamber sociability test. The data were analyzed using MATLAB 2017a.

Immunofluorescence staining

Mice were sacrificed 90 min after social ability test by a lethal dose of sodium pentobarbital. Mice were intracardially perfused with saline and 4% paraformaldehyde (PFA), and the brains were removed and postfixed in 4% PFA overnight. After which they were transferred to a 25–30% sucrose solution in PBS for 3 and sliced into 30μm thick coronal sections. Sections were washed with PBS-T (PBS containing 0.1% Triton X-100) and permeabilized with 0.5% Triton-100 PBS for 20–30 min. QuickBlock™ blocking buffer (P0260, Beyotime) was used for blocking. Then subsequently incubated with polyclonal rabbit c-Fos (1 : 200; Cat no. sc-52, Santa Cruz) or monoclonal mouse 6E10 (1 : 300; Cat no. 803001, BioLegend) for 20 h in QuickBlock™ Primary Antibody Dilution Buffer for Immunol Staining (P0262). After that, the sections underwent PBS-T washes (three times, 5 min each), followed by 1-h incubation with the secondary antibody (1 : 500, Cat no. A-21206 or A-21202 Invitrogen) at 37°C. Finally, the slice underwent three times of washes and was counterstained with DAPI.

Statistical analysis

The commercial software (GraphPad Prism version 9; GraphPad Software, Inc, La Jolla, CA) were used for statistical comparisons, via one-way ANOVAs, two-way repeated ANOVAs and t tests to determine the different means among the groups. The significance threshold was set at p <0.05, and the data were shown as mean±SEM.

RESULTS

The accumulation of Aβ in aBLA is associated with impairments in social abilities

To investigate the presence of Aβ pathology within the BLA, we utilized 5xFAD mice, which exhibit three amyloid precursor protein mutations and two presenilin-1 mutations. Immunofluorescence staining using the 6E10 antibody revealed a strong fluorescence signal in the aBLA region of 6-month-old 5xFAD mice, as compared to the WT group (Fig. 1a). Quantitative analysis demonstrated that the density of 6E10 + signals in aBLA of 5xFAD mice was higher than that of the control group (Fig. 1b). This finding indicates the vulnerability of the aBLA to Aβ pathology.

Fig. 1

5xFAD mice displayed impairments in social abilities. (a) Representative images show prominent accumulation of Aβ (6E10) in the aBLA of 5xFAD mice. Scale bar = 50μm. (b) Statistical analysis of 6E10 density. (c) No significant differences were found between the times spent on the right chamber versus the left chamber during the habituation phase. (d) 5xFAD mice spent less time in the social zone relative to wild type (WT) mice during 5 min test epoch. Unpaired t-test, **p < 0.01. (e) Social-nonsocial ratio decreased in 5xFAD mice compared with WT mice. Unpaired t-test, **p < 0.01. (f) No difference in the distance travelled was detected between WT and 5xFAD mice. n = 9 mice per group. Data are presented as the mean±SEM.

Next, we conducted a series of behavioral tests to evaluate social ability in the presence of Aβ. In the three-chamber sociability test, during the habituation trial, no differences were observed in the time spent in the right versus left zones of the testing chamber for both 5xFAD and WT mice (Fig. 1c), indicating no location preference. Five min after habituation, an intruder was introduced into the chamber and placed under an inverted cup (social zone). WT mice spent more time in the social zone than the nonsocial zone (an empty cup), while 6-month-old 5xFAD mice exhibited reduced exploration time in the social zone (Fig. 1d) and a lower social/nonsocial ratio (Fig. 1e). These data suggest impaired social ability in 5xFAD mice in the presence of Aβ accumulation in the aBLA.

Moreover, we found no significant difference in the distance travelled between WT and 5xFAD groups (Fig. 1f), suggesting that their motor abilities are not affected by Aβ deposition. This strengthens the notion that the observed alterations in social behavior are specifically related to social anxiety and not general motor impairments.

Aβ overactivates aBLA-vCA1 circuit to impair social ability

A previous study has reported that aBLA-vCA1 circuit bidirectionally controls social ability in an immediate, yet reversible, manner [27]. To investigate whether the aBLA-vCA1 circuit undergoes changes in response to Aβ accumulation, we employed retrograde tracing combined with c-Fos staining. Specifically, we stereotaxically injected a retrograde tracer (CTB, Cholera toxin subunit B) into the vCA1 region. After one week, we conducted c-Fos staining 90 min following the social ability test. The results revealed prominent CTB signals in the aBLA region (Fig. 2a), indicating a physical connection between the aBLA and vCA1. Notably, compared to WT mice, 5xFAD mice exhibited a higher number of c-Fos + and CTB + neurons in the aBLA (Fig. 2b), suggesting an overactivation of the aBLA-vCA1 circuit in 5xFAD mice.

Fig. 2

Inactivation of aBLA-vCA1 circuit ameliorates social deficit in 5xFAD mice. (a) Representative co-staining of CTB + /c-Fos + neurons in the aBLA. Scale bar = 50μm. (b) Quantitative analysis of co localization of c-Fos and CTB. Unpaired t-test, **p < 0.01. (c) Schematic of the virus strategy. AAV-CaMKII-NpHR-eYFP was injected into the aBLA. (d) Representative confocal images of NpHR in the aBLA. Scale bar = 200μm. (e) Diagram of three-chamber sociability test. (f) Time spent in the social zone was increased in NpHR-5xFAD mice compared with eYFP-5xFAD mice during light stimulation epoch. Two-way ANOVA, F (1, 16) = 66.31, p < 0.0001, Bonferroni post hoc analysis, **p < 0.01. (g, h) Light inhibition of aBLA-vCA1 circuit increased social-nonsocial ratio in NpHR-5xFAD mice (g) but not in eYFP-5xFAD mice (h). Paired t-test, NpHR-5xFAD: p = 0.0012, **p < 0.01; eYFP-5xFAD: p = 0.8899. (i) There was no difference in distance travelled among groups. Two-way ANOVA, F (1, 16) = 0.1955, p = 0.6643. n = 9 mice per group. Data were presented as mean ± SEM and individual values.

To investigate the involvement of the aBLA-vCA1 pathway in Aβ-induced impairment of social ability, we introduced the adeno-associated virus AAV5-CaMKIIa-eNpHR3.0-eYFP into the aBLA region (referred to as NpHR-5xFAD mice, as shown in Fig. 2c). As a control group, we administered AAV5-CaMKIIa-eYFP (eYFP-5xFAD mice). After a 4-week period for viral expression, the expression in aBLA neurons was confirmed by the signals of green fluorescence (eYFP) in aBLA (Fig. 2d). Then optical fibers were bilaterally implanted above the vCA1 region in both NpHR-5xFAD and eYFP-5xFAD mice. Subsequently, the mice were allowed a recovery period of one week following the surgical procedure. Behaviorally, in the three-chamber sociability test, during the light-off trial, no differences were observed in the time spent in the social zones for both NpHR-5xFAD and eYFP-5xFAD mice. NpHR-5xFAD mice exhibited an increased duration spent in the social zone during the light on epoch compared to eYFP-5xFAD mice (Fig. 2e, f). Moreover, NpHR-5xFAD mice displayed a higher social/nonsocial ratio during the light on epoch, suggesting that they devoted more time to exploring the social zone during light stimulation (Fig. 2g). In contrast, there were no significant variations in social/nonsocial ratio in eYFP-5xFAD mice between the light on and light off epochs (Fig. 2h). Additionally, we analyzed locomotor activity across both epochs and found no differences between the groups in terms of distance traveled during the light-on epoch compared to the light-off epoch (Fig. 2i).

Together, these data reveal the crucial role of the aBLA-vCA1 pathway in Aβ-related impairment of social ability.

Urolithin A administration improves Aβ-disrupted protein network and attenuates Aβ pathology in the aBLA

Previously, Aβ has been reported to disrupt protein expression associated with synapse [37], autophagy [38, 39], inflammation [40], etc. Through proteomics analysis (Fig. 3), we identified an up-regulation of 546 proteins and a down-regulation of 222 proteins in 5xFAD mice compared to WT mice (Fig. 3a).

Fig. 3

Urolithin A treatment recovers protein networks in the aBLA of 5xFAD mice. (a) The heatmap of differentially expressed proteins (p < 0.05) for WT mice, 5xFAD treated with vehicle (5xFAD), 5xFAD treated with Urolithin A (UA-5xFAD). The Z value of protein abundance was plotted in a red-blue color scale, with red and blue indicating increased and decreased protein expression, respectively. (b) 362 overlapped proteins changes in both comparison groups and their relative expression abundance (5xFAD versus WT and UA-5xFAD versus 5xFAD). (c) The heatmap of pathways enriched from the 362 shared proteins are shown in (b) with the red color representing increased expression and blue color reduced expression; pathway enrichment analysis of two cluster shared proteins with WebGestalt online analysis (https://www.webgestalt.org/); and the top 20 changes in shared proteins. (d) Representative images show less Aβ (6E10) in the aBLA of 5xFAD mice after administration of UA. Scale bar = 100μm. (e) Quantitative analysis of 6E10 (Unpaired t test, t = 10.61, df = 4, p = 0.0004. n = 3 per group).

Interestingly, after one month of intragastric administration of UA, we observed that 239 proteins were increased, while 244 proteins were decreased in UA-5xFAD mice compared to 5xFAD mice (Fig. 3a). A 40.7% overlap of differentially expressed proteins (DEPs) was found between the 5xFAD versus WT groups and the UA-5xFAD versus 5xFAD groups (Fig. 3b, c). Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis showed that the DEPs of cluster 1 (significantly decreased in 5xFAD and significantly increased after UA treatment) were mainly enriched in ribosome, AD, oxidative phosphorylation, Parkinson’s disease, retrograde endocannabinoid and cholinergic synapse signaling pathways (Fig. 3c), while the DEPs of cluster 2 (significantly increased in 5xFAD and significantly decreased after UA treatment) were mainly enriched in oxidative phosphorylation, mitophagy, neurotrophin signaling pathway, lysosome, GABAergic synapse and autophagy signaling pathways (Fig. 3c). Notably, among the top 20 DEPs of cluster 1 in the 5xFAD versus WT groups, the expression of Mt3, Sumo 2, Arf5, and others was significantly reversed after UA treatment (Fig. 3c). Additionally, the expression of Ctsf, which was among the top 20 DEPs of cluster 2 in the 5xFAD versus WT groups, was robustly decreased in UA-5xFAD mice compared to 5xFAD mice (Fig. 3c). It has been reported that altering ARF6 levels or its activity affects endosomal sorting of BACE1, leading to altered APP processing and Aβ production [41]. Increased protein sumoylation resulting from overexpression of SUMO-3 can dramatically reduce Aβ production [42]. Therefore, we speculate that UA-increased expression of Sumo 2 and Arf5 may modify AD-like Aβ pathology in 5xFAD mice.

Next, we conducted 6E10 immunostaining to investigate the impact of UA treatment on Aβ pathology in the aBLA of 5xFAD mice. Compared to the control group, UA treatment for one month resulted in reduced 6E10 signals in the aBLA of 5xFAD mice (Fig. 3d). Quantitative analysis confirmed significantly lower fluorescence intensity in the UA-5xFAD group compared to controls (Fig. 3e). These findings indicate the effectiveness of UA in reducing Aβ accumulation, likely by modifying the disrupted protein network caused by Aβ.

UA improves Aβ-impaired social ability via inhibiting aBLA-vCA1 circuit

UA is a naturally occurring compound known for its anti-aging, anticancer, antidiabetic, and neuroprotective properties [43]. To explore whether and how UA affects the activity of the aBLA-vCA1 circuit in 5xFAD mice, we performed in vivo calcium recording during the social ability test. We injected rAAV-CaMKII-Cre into the vCA1 region and AAV-CaMKII-DIO-GCaMP6f into the aBLA region (Fig. 4a). Through a Cre-dependent strategy, GCaMP6f was expressed in aBLA neurons projecting to vCA1. After a 4-week period for viral expression, the expression in aBLA neurons was confirmed by the signals of green fluorescence (GCaMP6f) in aBLA (Fig. 4b), then mice were tested for social ability. During the social ability test (Fig. 4g), continuous recording of Ca^2 + fluorescence signals was conducted in the arena. We observed a significant increase in Ca2 + signals when 5xFAD mice explored the social zone (Fig. 4c). However, this heightened neuronal activity in the aBLA-vCA1 circuit was significantly reduced in 5xFAD mice with UA treatment (Fig. 4d–f), indicating the inhibitory effects of UA on the overactivation of the aBLA-vCA1 circuit caused by Aβ accumulation in the aBLA. We statistically analyzed the ΔF/F and AUC values of the two groups and found that the difference between the two groups was significant (Fig. 4e, f). These results indicate that the aBLA of mice is significantly activated during social interaction, and UA administration can effectively reduce the activity of the aBLA in mice.

Fig. 4

UA improves Aβ-impaired social ability. (a) Schematic of the virus injection and fiber implantation. rAAV-CaMKII-Cre was injected into the vCA1, AAV-CaMKII-DIO-GCaMP6f was injected into the aBLA. Optic fiber was implantation into aBLA after virus expression. (b) Representative images of GCaMP6f in the aBLA. Scale bar = 200μm. Peri-event time histogram (PETH) of normalized ΔF/F for aBLA^CaMK^II neurons projecting to vCA1 in 5xFAD group (c) and 5xFAD + UA group (d). (e) Quantitative analysis of average ΔF/F for 5xFAD group and 5xFAD + UA group (Unpaired t test, t = 13.33, df = 6, p < 0.0001. n = 4 per group). (f) Quantitative analysis of AUC for 5xFAD group and 5xFAD + UA group and 5xFAD + UA group (Unpaired t test, t = 9.93, df = 6, p < 0.0001. n = 4 per group). (g) Schematic of the three-chamber sociability test. 5xFAD and 5xFAD + UA mice were tested in the three-chamber. (h, i) UA treatment increased investigation time in social zones (Unpaired t test, t = 7.656, df = 14, p < 0.0001.) and social:non-social ratio (Unpaired t test, t = 6.725, df = 14, p < 0.0001.) in 5xFAD mice. n = 8 mice per group. (j) No difference was observed in distance between groups. Data were presented as mean ± SEM.

Next, we investigated the effects of UA on social ability impaired by Aβ accumulation in the aBLA. In comparison to the control group, 5xFAD mice treated with UA spent more time in the social zone during the Three-Chamber Sociability Test (Fig. 4h). Furthermore, the UA group showed an increased social/nonsocial ratio (Fig. 4i). These results indicate that UA is effective in rescuing Aβ-impaired social ability. Additionally, we analyzed locomotor activity across both epochs and found no differences between the groups in terms of distance travelled among the groups (Fig. 4j).

To determine the necessity of the aBLA-vCA1 circuit in the improvement of UA on Aβ-impaired social ability, we administered injections of AAV5-CaMKIIa-hChR2(H134R)-eYFP into the aBLA region of 5xFAD mice subjected to UA. The control group received injections of AAV5-CaMKIIa-eYFP. After 4 weeks of viral expression, optical fibers were implanted above the vCA1 region. The mice were then given a recovery period of one week following the surgical procedure. The Three-Chamber Sociability Test was conducted to evaluate social behavior. During a 10-min blue light activation period (Fig. 5a), the social interaction time (Fig. 5b) and social/nonsocial ratio (Fig. 5c) in UA group were significantly attenuated during the light-on epoch. Additionally, there was no statistically significant difference in locomotor distance between the groups, suggesting that their motor capabilities remained unaffected (Fig. 5d). These findings indicate that the aBLA-vCA1 circuit is required for the improvements of UA on Aβ-impaired social ability.

Fig. 5

aBLA-vCA1 circuit governs the beneficial effects of UA on Aβ-impaired social ability. (a) Schematic of the three-chamber sociability test. 5xFAD and 5xFAD + UA mice expressed ChR2 were examined in the three-chamber test. (b, c) Time in social zone (one-way ANOVA, F(2, 21) = 58.06, p < 0.0001) and social:non-social ratio (one-way ANOVA, F(2, 21) = 35.85, p < 0.0001) were decreased in 5xFAD + UA mice upon light stimulation. n = 8 per group. (d) No distance difference was detected in distance. Data were presented as mean ± SEM.

Taken together, our present study delineated a distinct protein network in the aBLA that is disrupted by Aβ accumulation. In addition to modification of protein networks, UA exerts efficacy in improving Aβ-impaired social ability in a manner dependent on the aBLA-vCA1 circuit.

DISCUSSION

In the present study, we have identified the crucial role of the aBLA-vCA1 circuit in Aβ-impaired social ability. Administration of UA for one month has significantly modified the Aβ-disrupted protein network, reduced Aβ accumulation in the aBLA, and alleviated social ability dysfunction in a manner dependent on the aBLA-vCA1 circuit. These findings highlight the aBLA-vCA1 circuit as a potential target for ameliorating social anxiety during the progression of AD and suggest the promising strategy of administering UA to reduce Aβ pathology and associated deficits in social behaviors.

The BLA is particularly vulnerable to Aβ pathology, as demonstrated by España et al., who reported significant increases in intraneuronal Aβ₄₀ and Aβ₄₂ in the BLA of human AD brains and 3XTg AD mice [44]. Consistently, our study revealed Aβ accumulation in the aBLA of 6-month-old 5xFAD mice. Moreover, these mice exhibited pronounced social avoidance, indicating impaired social ability. These findings further support the close relationship between Aβ and social anxiety observed in AD patients.

BLA projects to various brain regions, forming neural networks, including CeA, hippocampus, BNST, NAcc, DMS, and PFC [45]. Previous studies have shown that the aBLA-vCA1 circuit bidirectionally regulates social ability [27]. Activation of the aBLA-vCA1 pathway using optogenetics significantly reduces social interaction with intruders in the resident-juvenile-intruder home-cage test and decreases time spent in the social zone during the three-chamber sociability test. In our AD mouse model, we observed hyperactivation and abnormal discharge activity in the aBLA-vCA1 circuit during social exploration when Aβ was significantly accumulated in the aBLA. Inhibition of this pathway effectively improved the social avoidance caused by Aβ accumulation. Our data not only support the anxiogenic effects of the aBLA-vCA1 circuit [27 , 46], but also elucidate the necessity of this circuit in Aβ-impaired social ability. These findings provide new insights into the neural circuit mechanisms underlying Aβ toxicity in AD and identify a potential target for clinical deep brain stimulation in AD treatment.

Many previous studies have primarily focused on investigating the molecular mechanisms underlying Aβ pathology in the hippocampus and cortex. The process of Aβ plaque accumulation, the propagation of tau pathology, the interaction between ApoE and Aβ plaques, as well as the involvement of various proteins associated with the genetic risk of AD, including SHIP1, CD2AP, RIN3, BIN1, PLCG2, CASS4, and PTKB2. These molecular mechanisms are involved in the overproduction, aggregation and clearance of Aβ. In our study, we employed proteomic techniques to characterize the aberrant protein network specifically in the aBLA of 5xFAD mice. Our findings revealed a significant decrease in the protein levels of SUMO2 and ARF5. It is known that lysines 587 and 595 of APP are located adjacent to the site of β-secretase cleavage and can be covalently modified by small ubiquitin-like modifier (SUMO) proteins in vivo [47], which regulates Aβ generation. Li et al reported that reducing endogenous protein sumoylation with dominant-negative SUMO3 mutants significantly increases Aβ production. Additionally, for the small GTPase ADP ribosylation factor (Arf), Arf6 has been found to be associated with APP processing and Aβ production. Overactivating endogenous ARF6 can significantly decrease Aβ secretion [41]. Based on these previous findings, we hypothesize that decreased SUMO2 and ARF5 could contribute to the vulnerability of the aBLA to Aβ accumulation. However, further investigations are necessary to elucidate the causal relationship and the detailed mechanisms involved. While PSEN1 mutation contributes to Aβ overproduction [48], no beneficial effects of UA on PSEN1 have been reported to date. Future studies are warranted to investigate whether and how UA modifies or regulates PSEN1 to reduce Aβ production.

The hyperphosphorylation of the tau protein represents another significant pathological change in AD and is closely linked to memory loss in AD patients. It has been documented that Aβ can elevate tau phosphorylation [49]. Studies have indicated an abundance of hyperphosphorylated tau in 5xFAD mice [50]. Recently, Fang et al. demonstrated that UA effectively eliminated AD-related tau hyperphosphorylation in human neuronal cells and reversed memory impairment in transgenic tau nematodes and mice [51]. Furthermore, UA pretreatment in APPSwe cells was found to reduce tau phosphorylation [51]. Based on these findings, we postulate that UA treatment could potentially reduce tau phosphorylation in 5XFAD mice, a hypothesis worthy of further exploration in future studies.

UA has antiaging effects, anticancer activity, antidiabetic activity, and neuroprotective activity [52]. In our study, we made the striking observation that UA treatment significantly reduced the Aβ load in the aBLA of 5xFAD mice and effectively suppressed the overdischarge of the aBLA-vCA1 circuit during social investigation. After UA treatment, we observed a significant reversal in the expression of SUMO2 and ARF5 in the aBLA of 5xFAD mice, as revealed by proteomic analysis. Considering the observed effects on Aβ degradation and inhibition of excitability, we speculate that the mechanisms underlying the Aβ reduction and inhibition of the aBLA-vCA1 circuit by UA treatment may involve the modulation of SUMO2 and ARF5. These findings provide valuable insights into the potential molecular pathways implicated in the therapeutic effects of UA on Aβ pathology and warrant further investigation to elucidate the precise mechanisms involved.

In summary, our study identifies the aBLA-vCA1 circuit as a vulnerable pathway in response to Aβ accumulation in the aBLA and highlights its crucial role in Aβ-induced social avoidance. Administering UA may be a promising strategy to effectively reduce Aβ burden, restore disrupted protein networks, enhance the resilience of the anxiogenic aBLA-vCA1 circuit, and ameliorate social anxiety during the progression of AD. These findings provide valuable insights into potential therapeutic interventions targeting Aβ-related social deficits in AD.

AUTHOR CONTRIBUTIONS

Rui Xiong (Data curation; Investigation; Methodology); Binrui Li (Data curation; Writing – original draft); Haitao Yu (Data curation; Investigation; Methodology); Tianceng Fan (Data curation); Huiling Yu (Data curation; Funding acquisition); Ying Yang (data analysis); Jian-zhi Wang (data analysis); Guilin Pi (Conceptualization; Funding acquisition; Supervision; Writing – review & editing); Xifei Yang (Conceptualization; Funding acquisition; Supervision; Writing – review & editing).

Footnotes

ACKNOWLEDGMENTS

The authors have no acknowledgments to report.

FUNDING

This study was supported from the Key Basic Research Program of Shenzhen Science and Technology Innovation Commission (JCYJ20200109150717745), Shenzhen Key Medical Discipline Construction Fund (SZXK069), Sanming Project of Medicine in Shenzhen (SZSM202211010), and the Natural Science Foundation of China (No. 82201584, 82301622).

CONFLICT OF INTEREST

The authors declare that they have no conflicts of interest.

DATA AVAILABILITY

All the data from the current study are shown in the manuscript are available from the corresponding author upon reasonable request.

References

Winblad

, Amouyel

, Andrieu

, Ballard

, Brayne

, Brodaty

, Cedazo-Minguez

, Dubois

, Edvardsson

, Feldman

, Fratiglioni

, Frisoni

, Gauthier

, Georges

, Graff

, Iqbal

, Jessen

, Johansson

, Jönsson

, Kivipelto

, Knapp

, Mangialasche

, Melis

, Nordberg

, Rikkert

, Qiu

, Sakmar

, Scheltens

, Schneider

, Sperling

, Tjernberg

, Waldemar

, Wimo

, Zetterberg

(2016) Defeating Alzheimer’s disease and other dementias: A priority for European science and society. Lancet Neurol 15, 455–532.

(2021) 2021 Alzheimer’s disease facts and figures. Alzheimers Dement 17, 327–406.

Selkoe

(1999) Translating cell biology into therapeutic advances in Alzheimer’s disease. Nature 399, A23–31.

Glenner

, Wong

(1984) Alzheimer’s disease: Initial report of the purification and characterization of a novel cerebrovascular amyloid protein. Biochem Biophys Res Commun 120, 885–890.

Yang

, Park

S-Y

, Baek

, Lee

, Chung

, Liu

, Lee

, Kim

, Kwon

, Choi

, Kim

, Lee

Y-S

, Lee

, Kim

, Chang

Y-T

, Jung

, Kim

, Bae

(2022) Adoptive therapy with amyloid-β specific regulatory T cells alleviates Alzheimer’s disease. Theranostics 12, 7668–7680.

Pradeepkiran

, Reddy

(2020) Defective mitophagy in Alzheimer’s disease. Ageing Res Rev 64, 101191.

John

, Reddy

(2021) Synaptic basis of Alzheimer’s disease: Focus on synaptic amyloid beta, P-tau and mitochondria. Ageing Res Rev 65, 101208.

Reddy

, Oliver

(2019) Amyloid beta and phosphorylated tau-induced defective autophagy and mitophagy in Alzheimer’s disease. Cells 8, 488.

Hijazi

, Heistek

, van der Loo

, Mansvelder

, Smit

, van Kesteren

(2020) Hyperexcitable parvalbumin interneurons render hippocampal circuitry vulnerable to amyloid beta. iScience 23, 101271.

10.

Quevenco

, Schreiner

, Preti

, van Bergen

JMG

, Kirchner

, Wyss

, Steininger

, Gietl

, Leh

, Buck

, Pruessmann

, Hock

, Nitsch

, Henning

, Van De Ville

, Unschuld

(2019) GABA and glutamate moderate beta-amyloid related functional connectivity in cognitively unimpaired old-aged adults. Neuroimage Clin 22, 101776.

11.

Chen

, Dang

, Zhang

(2021) Brain mechanisms underlying neuropsychiatric symptoms in Alzheimer’s disease: A systematic review of symptom-general and –specific lesion patterns. Mol Neurodegener 16, 38.

12.

Zhao

Q-F

, Tan

, Wang

H-F

, Jiang

, Tan

M-S

, Tan

, Xu

, Li

J-Q

, Wang

, Lai

T-J

, Yu

J-T

(2016) The prevalence of neuropsychiatric symptoms in Alzheimer’s disease: Systematic review and meta-analysis. J Affect Disord 190, 264–271.

13.

Chung

, Cummings

(2000) Neurobehavioral and neuropsychiatric symptoms in Alzheimer’s disease: Characteristics and treatment. Neurol Clin 18, 829–846.

14.

Pietrzak

, Lim

, Neumeister

, Ames

, Ellis

, Harrington

, Lautenschlager

, Restrepo

, Martins

, Masters

, Villemagne

, Rowe

, Maruff

(2015) Amyloid-β, anxiety, and cognitive decline in preclinical Alzheimer disease: A multicenter, prospective cohort study. JAMA Psychiatry 72, 284.

15.

Pink

, Krell-Roesch

, Syrjanen

, Vassilaki

, Lowe

, Vemuri

, Stokin

, Christianson

, Kremers

, Jack

, Knopman

, Petersen

, Geda

(2022) A longitudinal investigation of Aβ, anxiety, depression, and mild cognitive impairment. Alzheimers Dement 18, 1824–1831.

16.

, Zhao

, Lai

, Xia

, Wan

, Wei

, Liang

, Chen

, Xu

(2022) Loss of SST and PV positive interneurons in the ventral hippocampus results in anxiety-like behavior in 5xFAD mice. Neurobiol Aging 117, 165–178.

17.

Porcelli

, Van Der Wee

, van der Werff

, Aghajani

, Glennon

, van Heukelum

, Mogavero

, Lobo

, Olivera

, Lobo

, Posadas

, Dukart

, Kozak

, Arce

, Ikram

, Vorstman

, Bilderbeck

, Saris

, Kas

, Serretti

(2019) Social brain, social dysfunction and social withdrawal. Neurosci Biobehav Rev 97, 10–33.

18.

Bilderbeck

, Penninx

BWJH

, Arango

, van der Wee

, Kahn

, Winter-van Rossum

, Hayen

, Kas

, Post

, Dawson

(2019) Overview of the clinical implementation of a study exploring social withdrawal in patients with schizophrenia and Alzheimer’s disease. Neurosci Biobehav Rev 97, 87–93.

19.

, Cai

, Gu

, Yu

, Li

, Liu

Q-S

, Zhao

(2022) Treadmill exercise prevents decline in spatial learning and memory in 3×Tg-AD mice through enhancement of structural synaptic plasticity of the hippocampus and prefrontal cortex. Cells 11, 244.

20.

Fang

, Zhou

, Miao

, Han

, Wei

, Chen

(2020) Therapeutic effect of GLP-1 engineered strain on mice model of Alzheimer’s disease and Parkinson’s disease. AMB Express 10, 80.

21.

McKernan

, Shinnick-Gallagher

(1997) Fear conditioning induces a lasting potentiation of synaptic currents. Nature 390, 607–611.

22.

Tye

, Tye

, Cone

, Hekkelman

, Janak

, Bonci

(2010) Methylphenidate facilitates learning-induced amygdala plasticity. Nat Neurosci 13, 475–481.

23.

Rumpel

, LeDoux

, Zador

, Malinow

(2005) Postsynaptic receptor trafficking underlying a form of associative learning. Science 308, 83–88.

24.

Tye

, Prakash

, Kim

S-Y

, Fenno

, Grosenick

, Zarabi

, Thompson

, Gradinaru

, Ramakrishnan

, Deisseroth

(2011) Amygdala circuitry mediating reversible and bidirectional control of anxiety. Nature 471, 358–362.

25.

Kirkby

, Luongo

, Lee

, Nahum

, Van Vleet

, Rao

, Dawes

, Chang

, Sohal

(2018) An amygdala-hippocampus subnetwork that encodes variation in human mood. Cell 175, 1688–1700.e14.

26.

Bannerman

, Grubb

, Deacon

RMJ

, Yee

, Feldon

, Rawlins

JNP

(2003) Ventral hippocampal lesions affect anxiety but not spatial learning. Behav Brain Res 139, 197–213.

27.

Felix-Ortiz

, Tye

(2014) Amygdala inputs to the ventral hippocampus bidirectionally modulate social behavior. J Neurosci 34, 586–595.

28.

Wang

, Wang

, Liu

, Zhang

, Wang

, Lin

(2011) Neurons in the Amygdala with response-selectivity for anxiety in two ethologically based tests. PLoS One 6, e18739.

29.

Adhikari

, Topiwala

, Gordon

(2010) Synchronized activity between the ventral hippocampus and the medial prefrontal cortex during anxiety. Neuron 65, 257–269.

30.

Jimenez

, Su

, Goldberg

, Luna

, Biane

, Ordek

, Zhou

, Ong

, Wright

, Zweifel

, Paninski

, Hen

, Kheirbek

(2018) Anxiety cells in a hippocampal-hypothalamic circuit. Neuron 97, 670–683.e6.

31.

Felix-Ortiz

, Beyeler

, Seo

, Leppla

, Wildes

, Tye

(2013) BLA to vHPC inputs modulate anxiety-related behaviors. Neuron 79, 658–664.

32.

Espín

, Larrosa

, García-Conesa

, Tomás-Barberán

(2013) Biological significance of urolithins, the gut microbial ellagic Acid-derived metabolites: The evidence so far. Evid Based Complement Alternat Med 2013, 270418.

33.

Cortés-Martín

, García-Villalba

, González-Sarrías

, Romo-Vaquero

, Loria-Kohen

, Ramírez-de-Molina

, Tomás-Barberán

, Selma

, Espín

(2018) The gut microbiota urolithin metabotypes revisited: The human metabolism of ellagic acid is mainly determined by aging. Food Funct 9, 4100–4106.

34.

Luan

, D’Amico

, Andreux

, Laurila

P-P

, Wohlwend

, Li

, Imamura De Lima

, Place

, Rinsch

, Zanou

, Auwerx

(2021) Urolithin A improves muscle function by inducing mitophagy in muscular dystrophy. Sci Transl Med 13, eabb0319.

35.

Toney

, Fox

, Chaidez

, Ramer-Tait

, Chung

(2021) Immunomodulatory role of urolithin A on metabolic diseases. Biomedicines 9, 192.

36.

Denk

, Petrocelli

, Conche

, Drachsler

, Ziegler

, Braun

, Kress

, Nicolas

, Mohs

, Becker

, Neurath

, Farin

, Buchholz

, Andreux

, Rinsch

, Greten

(2022) Expansion of T memory stem cells with superior anti-tumor immunity by Urolithin A-induced mitophagy. Immunity 55, 2059–2073.e8.

37.

Sehar

, Rawat

, Reddy

, Kopel

, Reddy

(2022) Amyloid beta in aging and Alzheimer’s disease. Int J Mol Sci 23, 12924.

38.

Pomilio

, Gorojod

, Riudavets

, Vinuesa

, Presa

, Gregosa

, Bentivegna

, Alaimo

, Alcon

, Sevlever

, Kotler

, Beauquis

, Saravia

(2020) Microglial autophagy is impaired by prolonged exposure to β-amyloid peptides: Evidence from experimental models and Alzheimer’s disease patients. Geroscience 42, 613–632.

39.

Chen

, He

H-J

, Ye

, Feng

, Wang

W-W

, Gu

, Han

, Xie

(2021) Defective autophagy and mitophagy in Alzheimer’s disease: Mechanisms and translational implications. Mol Neurobiol 58, 5289–5302.

40.

Kinney

, Bemiller

, Murtishaw

, Leisgang

, Salazar

, Lamb

(2018) Inflammation as a central mechanism in Alzheimer’s disease. Alzheimers Dement (N Y) 4, 575–590.

41.

Sannerud

, Declerck

, Peric

, Raemaekers

, Menendez

, Zhou

, Veerle

, Coen

, Munck

, De Strooper

, Schiavo

, Annaert

(2011) ADP ribosylation factor 6 (ARF6) controls amyloid precursor protein (APP) processing by mediating the endosomal sorting of BACE1, Proc Natl Acad Sci U S A 108, E559–568.

42.

, Wang

, Quon

, Liu

Y-W

, Cordell

(2003) Positive and negative regulation of APP amyloidogenesis by sumoylation. Proc Natl Acad Sci U S A 100, 259–264.

43.

, Lu

, Wang

(2023) Urolithins: A prospective alternative against brain aging. Nutrients 15, 3884.

44.

España

, Giménez-Llort

, Valero

, Miñano

, Rábano

, Rodriguez-Alvarez

, LaFerla

, Saura

(2010) Intraneuronal beta-amyloid accumulation in the amygdala enhances fear and anxiety in Alzheimer’s disease transgenic mice. Biol Psychiatry 67, 513–521.

45.

Limoges

, Yarur

, Tejeda

(2022) Dynorphin/kappa opioid receptor system regulation on amygdaloid circuitry: Implications for neuropsychiatric disorders. Front Syst Neurosci 16, 963691.

46.

, Gao

, Wu

, Wang

, Lei

, Zeng

, Gao

, Yu

, Xiong

, Jiang

, Li

, Wang

, Guo

, Zhang

, Yin

, He

, Ke

, Li

, Liu

, Yang

, Luo

M-H

, Zhang

, Yang

, Wang

J-Z

(2020) Posterior basolateral amygdala to ventral hippocampal CA1 drives approach behaviour to exert an anxiolytic effect. Nat Commun 11, 183.

47.

Zhang

Y-Q

, Sarge

(2008) Sumoylation of amyloid precursor protein negatively regulates Abeta aggregate levels. Biochem Biophys Res Commun 374, 673–678.

48.

Bagaria

, Bagyinszky

, An

SSA

(2022) Genetics, functions, and clinical impact of presenilin-1 (PSEN1) gene. Int J Mol Sci 23, 10970.

49.

Bloom

(2014) Amyloid-β and tau: The trigger and bullet in Alzheimer disease pathogenesis. JAMA Neurol 71, 505–508.

50.

Choi

H-J

, Park

J-H

, Jeong

, Hwang

J-W

, Lee

, Seol

, Kim

, Cha

B-Y

, Seo

, Moon

, Hoe

H-S

(2022) Donepezil ameliorates Aβ pathology but not tau pathology in 5xFAD mice. Mol Brain 15, 63.

51.

Fang

, Hou

, Palikaras

, Adriaanse

, Kerr

, Yang

, Lautrup

, Hasan-Olive

, Caponio

, Dan

, Rocktäschel

, Croteau

, Akbari

, Greig

, Fladby

, Nilsen

, Cader

, Mattson

, Tavernarakis

, Bohr

(2019) Mitophagy inhibits amyloid-β and tau pathology and reverses cognitive deficits in models of Alzheimer’s disease. Nat Neurosci 22, 401–412.

52.

Hasheminezhad

, Boozari

, Iranshahi

, Yazarlu

, Sahebkar

, Hasanpour

, Iranshahy

(2022) A mechanistic insight into the biological activities of urolithins as gut microbial metabolites of ellagitannins. Phytother Res 36, 112–146.