2.1 Animal Grouping and Drug Intervention

Our study utilized male C57BL/6J mice that were 6 months old, along with APP/PS1 double transgenic mice. The C57BL/6J mice are the wild-type (WT) group (n = 10). The APP/PS1 double transgenic mice were stochastically segregated into the following groups: the Model group, the 25 mg/kg AE group, the 50 mg/kg AE group, the 100 mg/kg AE group [29, 30], and the Donepezil group, with each group consisting of 10 mice (n = 10). Following a 1-week adaptive feed, the mice underwent 28 days of intragastric treatment. The animals were then used for subsequent experiments following the experimental flow chart (Figure 1B).

Aloe-emodin (Figure 1A), which was purchased from Yuanye Biotechnology Co. Ltd. (Shanghai), was used in the animal experiments with a purity of 97.6%. It was prepared as a suspension in pure water, and the dosage for each group was measured for intragastric administration. Donepezil hydrochloride tablets (from China) were given to the Donepezil group, dissolved in pure water, at a dosage of 2.1 mg/kg via the intragastric route. The WT group received daily intragastric administration of pure water. All procedures undertaken in our research received approvals from the Ethics Committee of the Anhui University of Chinese Medicine (Animal Ethics No.: AHUCM-mouse-2024051).

2.2 Cell Culture and Drug Treatment

The cells were seeded into culture dishes and incubated in a 37°C, 5% CO₂ incubator. After 24 h of seeding, the medium was changed, and the Model group was processed using 30 μM Aβ_25-35. The AE groups were processed using 30 μM Aβ_25-35 [31] and 2, 4, or 6 μM AE, and the RAPA group was processed using 30 μM Aβ_25-35 and 200 nM Rapamycin. After 24 h of treatment, various detection experiments were performed.

1 mg of Aβ_25-35 was dissolved in 1 mL of deionized water as a stock solution, incubated in a 37°C incubator for 5 days to induce oligomer formation. It was aliquoted into 100 μL volumes and stored at −20°C. A concentration of 30 μM was used in this experiment based on previous research. Aloe-emodin and rapamycin were each dissolved in DMSO to form stock solutions of 50 and 10 mM. The stock solution of Aloe-emodin was diluted in a complete medium to the required concentration gradient for further use. The stock solution of Rapamycin was diluted twice in a complete medium to a final concentration of 200 nM.

2.3 Behavior Test

2.4 Brain Collection and Tissue Preparation

After the behavioral experiments, the mice underwent anesthesia through an intraperitoneal injection of sodium pentobarbital, followed by cardiac perfusion. The brains were collected and preserved in paraformaldehyde for embedding and subsequent sectioning and staining. Some mice were sacrificed after anesthesia, and the hippocampus was quickly collected on ice for electron microscopy and western blotting.

2.5 Hematoxylin and Eosin (HE) Staining

The sliced brain tissue sections were dewaxed, stained with hematoxylin–eosin (Beyotime, C0105S), dehydrated, cleared, and mounted with neutral resin. After preparation, the sections were observed using a slide scanner (Wisleap WS-10, China).

2.6 Transmission Electronic Microscopy

The mice's brains were extracted on ice within 2 min, and the hippocampus was promptly isolated and immersed in glutaraldehyde. After 24 h, the fixative was discarded, and the samples were placed in a PBS buffer for 6 h, followed by fixation in osmium acid. After 30% and 50% ethanol, the samples were stained with 70% ethanol uranyl acetate for 3 h before further dehydration with 80%, 95%, and 100% ethanol. They were then placed in propylene oxide for 2 h, followed by immersion in pure epoxy resin for 3 h. After embedding in purified epoxy resin, the samples were placed in a 45°C oven for 12 h, then in a 72°C oven for 24 h. The embedded blocks were trimmed, cut into ultrathin sections of 70 nm thickness, stained with lead, and observed using transmission electron microscopy.

2.7 Immunohistochemistry

The sections were subjected to xylene, 100%, 95%, and 80% ethanol, washed with running tap water, and then high-pressure antigen retrieval was performed. Two liters of citrate buffer antigen retrieval solution were prepared, and the sections were placed in a pressure cooker, boiled, and heat-repaired. The tissues were circled with an immunohistochemical pen and placed in an incubation box, and 3% H₂O₂ was added to the tissue and incubated. After washing with PBS, the slides were dried, and primary antibodies Opa1 (Zenbio), Drp1 (Proteintech), P62 (Servicebio), and PINK1 (Beyotime) were added, incubating at 37°C. Then, it was replaced with the secondary antibody. After washing with PBS, DAB chromogen was added, and the staining process was monitored under a microscope, stopping the reaction when positive staining was observed. The sections were stained with hematoxylin, then differentiated with hydrochloric acid alcohol, blued with lithium carbonate, dehydrated, cleared, and mounted. The sections were observed using a slide scanner (Wisleap WS-10, China).

2.8 Immunofluorescence

For the animal immunofluorescence experiment, first, the paraffin sections were dewaxed to water. After washing with PBS, the sections were placed into citrate buffer and microwaved for 15 min. After cooling, it was washed with PBS. Then goat serum-blocking solution was added and incubated in a wet box for 45 min. Diluted primary antibodies LC3B (Zenbio), Parkin (Zenbio), SIRT3 (Affinity), and p-AMPK (Beyotime) were added and nurtured at 4°C. The next day, fluorescent secondary antibodies were added and incubated, followed by DAPI staining. After washing, an anti-fluorescence quenching agent was applied to cover the section.

In the cellular fluorescence experiment, the treated HT22 cells are fixed with paraformaldehyde. Then, the cells were blocked for 60 min using an immunostaining blocking solution (Beyotime; P0102) and permeabilized with an immunostaining permeabilization solution (Beyotime; P0096) for 10 min. Primary antibodies LC3B (Zenbio), p-AMPK (Beyotime), AMPK (Beyotime), Parkin (Zenbio), and PINK1 (Beyotime) were placed in the wells and cultivated at 4°C. The next day, the cells were fostered with the Cy3-labeled donkey anti-rabbit IgG (Servicebio) and Alexa Fluor 488-labeled goat anti-rabbit IgG (Servicebio). After washing, the cells were stained with DAPI and mounted using an anti-fluorescence quenching mounting medium (Beyotime, P0131). The fluorescence was observed using a fluorescence or confocal laser scanning microscope (Olympus IX-81, Olympus FV1000, Japan).

2.9 CCK-8 Cell Viability Assay

To determine AE's safe concentration range and optimal dosing concentration, HT22 cells were placed within 96-well plates. After incubating for 1 day, the cells adhered and extended synapses. Moreover, they were modeled through 30 μM Aβ_25-35 and processed using discrepant concentrations of AE. After 1 day of intervention, CCK-8 solution (Beyotime, C0041) was added. Absorbance readings were taken with a multifunctional enzyme reader (Promege 1500, USA).

2.10 EDU Assay

The EDU assay detected DNA synthesis within 3 h, using the cell proliferation assay kit (Beyotime, C0075S) [36]. After HT22 cells in 96-well plates were treated, the EDU working solution was added. After EDU labeling, the fixative (Beyotime, P0098) was added. After washing, 100 μL of permeabilization buffer (Beyotime, P0096) was added and incubated for 10–15 min. After washing, 0.5 μL of Click reaction solution was added and incubated in the dark. After staining the nuclei with Hoechst 33342, the fluorescence was observed under a microscope (IX-81, Olympus, Japan).

2.11 Mitochondrial Membrane Potential (Δψm) Measurement

The enhanced Δψm detection kit (JC-1, Beyotime, C2003S) was used [37]. HT22 cells in 96-well plates were washed twice with PBS, and JC-1 dyeing solution was supplied. The cells were fostered for 20 min in an incubator. Finally, the cells were examined at 200× magnification using a fluorescence microscope.

2.12 ROS Measurement

The ROS detection kit (Beyotime, S0033S) was applied to determine ROS levels in each group of cells [38]. The cells were cultivated through the diluted DCFH-DA at 37°C for 40 min, with the medium in 96-well plates eliminated. Such cells were rinsed and examined through a fluorescence microscope at 200× magnification.

2.13 Ca²⁺ Measurement

Fluo-4 AM (Beyotime, S1061S) measured Ca²⁺ levels [39]. The medium was removed from the HT22 cells. The cells were cultured at 37°C for 30 min in an incubator with 100 μL of Fluo-4 staining solution poured into each well. After incubation, the cells were examined and imaged under a fluorescence microscope.

2.14 Mitochondrial Superoxide and GSH Level Measurement

The mitochondrial superoxide detection kit (MitoSOX Red, Beyotime, S0061S) was used. After digesting and collecting the cells, the cells were resuspended in a MitoSOX Red staining solution. The cells were cultured at 37°C for 30 min and centrifuged. After washing, the cells were resuspended in an appropriate volume of PBS and analyzed using flow cytometry. The GSH and GSSG detection kit (Beyotime, S0053) [40] was used to measure GSH levels in HT22 cells after drug treatment. Absorbance was measured using a multifunctional enzyme reader (Promege 1500, USA).

2.15 Western Blotting

After drug treatment, HT22 cells were collected using Western and IP cell lysis buffer (Biosharp; BL509A) containing PMSF (Biosharp; BL507A), followed by sonication on ice for 3 min. The lysates underwent centrifugation at 12,000 × g for 5 min at 4°C, with the supernatant collected. After electrophoresis and membrane transfer, it was blocked with skimmed milk. Then, the PVDF membrane was placed into the primary antibody solution and incubated at 4°C: Drp1 (Proteintech), Opa1 (Zenbio), P62 (Servicebio), LC3B (Zenbio), Beclin1 (Zenbio), p-AMPK (Beyotime), AMPK (Beyotime), PGC-1α (Zenbio), SIRT3 (Affinity), β-actin (Proteintech), and GAPDH (Proteintech). After the primary antibody incubation, it was washed with TBST. It was incubated with the corresponding secondary antibody (Proteintech) at normal temperature. After rinsing through the TBST, the developer solution was added evenly, and the blots were imaged through a gel imaging mechanism. Band intensity was analyzed using ImageJ software.

2.16 Transfection

The most efficient transfection sequence was selected from three designed sequences. HT22 cells were divided into five groups: Blank, NC, mSIRT3-siRNA-1, mSIRT3-siRNA-2, and mSIRT3-siRNA-3. After 1 day of growth, the cells were transfected using the prepared transfection reagent (CALNP, DN001) for 24 h. The NC group was processed through the identical volume of transfection reagent with no transfection sequence. After transfection, the cells were subjected to modeling and drug treatment. After 24 h, RNAs were extracted through a total RNA extraction kit (CALNP, FZ007), and SYBR Green RT-qPCR was performed using a one-step RT-qPCR kit (CALNP, FZ005). The subsequent steps were performed using an RT-PCR machine (LightCycler 96, Roche, Switzerland). SIRT3 primer sequences and transfection sequences are shown in Figure 7A.

2.17 Statistical Analysis

All statistical analyses were performed using SPSS (v 26.0) and GraphPad Prism (v 8.0). Data were presented as mean ± standard error of the mean (SEM), and a p < 0.05 was considered statistically significant. Normality of data distribution was assessed by the Shapiro–Wilk test, and homogeneity of variances was verified using Levene's test. For data conforming to both normality and homogeneity of variances, one-way analysis of variance (ANOVA) was applied for intergroup comparisons. If the normality assumption was violated, nonparametric tests were employed: the Mann–Whitney test for comparisons between two groups and the Kruskal–Wallis test for comparisons among more than two groups.

3.1 AE Alleviates Cognitive Dysfunction in AD Model Mice

We evaluated the effect of AE on cognitive dysfunction in ad mice using the MWM and Y-maze tests after 28 days of oral administration. As shown in Figure 1C, the mice underwent escape latency training for the first 5 days. On day 5, the ad mice spent more time to position the platform compared to the WT mice (p < 0.01). However, treatment with either AE or donepezil in ad mice significantly reduced the time spent to discover the platform (p < 0.05), as illustrated in Figure 1D. Furthermore, during the probe trial (Figure 1E), the percentage of time spent in the object quadrant was remarkably decreased among ad mice (p < 0.001), and the quantity of crossings into the object quadrant (p < 0.01) was in marked contrast with the WT group (Figure 1F). These deficits improved with AE and donepezil treatment, with the 100 mg/kg dose of AE demonstrating the most significant effect (p < 0.05). To assess short-term working memory, we performed the Y-maze test (Figure 1G). The ad mice made significantly fewer entries into the novel arm than WT (p < 0.001) and drug-treated ad animals (p < 0.05) as shown in Figure 1H. This indicates that treatment with AE improved the impairment of short-term working memory observed in ad mice.

3.2 AE Mitigates Pathological Features in AD Model Mice

Alzheimer's disease is strongly linked to damage in the neurons of the hippocampus. HE staining analysis of the hippocampal CA1, CA3, and DG areas among mice revealed that neurons within the wild-type (WT) group showed normal morphology (Figure 2A). In contrast, neurons within the APP/PS1 group showed signs of atrophy, disorganized arrangement, hyperchromatic staining, and deformed cell bodies. The AE and donepezil-treated groups had a more uniform cell distribution than ad mice, indicating significant recovery from neuronal damage. Among the AE treatment groups (25, 50, and 100 mg/kg), the group receiving 100 mg/kg showed the most significant improvement. Transmission electron microscopy was applied to check the intracellular organelle of hippocampal neurons. Compared to the WT group, ad mice displayed swollen, ruptured mitochondria, disrupted cristae, absent outer membranes, and reduced mitophagy (Figure 2B). Autophagosomes and autolysosomes were observed in the AE and donepezil groups, showing damaged mitochondria being engulfed, which indicate activation of mitophagy.

3.3 AE Activates Mitophagy in Neurons to Alleviate AD Symptoms

The expression of mitochondrial fission/fusion proteins (OPA1 and DRP1) and mitophagy markers (PINK1, Parkin, P62, Beclin1, and LC3B) in Alzheimer's mouse hippocampal neurons was analyzed by IHC and IF staining (Figures 2C and 3A). Opa1 expression in neurons of ad mice was decreased remarkably (p < 0.01) in comparison with the WT group but increased after treatment. Opa1 expression increased in the 25, 50, and 100 mg/kg AE groups, and the donepezil group showed the most significant effect (p < 0.05) (Figure 2D). In ad mice, neuron levels of DRP1, P62, Pink1, and Parkin exceeded prominently those in the WT group (p < 0.001). After administering AE and donepezil, expression levels significantly decreased (p < 0.05), most notably in the 100 mg/kg AE and donepezil groups (Figures 2D and 3D). LC3B levels were low in WT and ad mice but increased after AE and donepezil treatment (p < 0.05). It is worth noting that the moderate dose of 50 mg/kg also significantly regulated the expression of DRP1, P62, LC3B, pink1, and Parkin (p < 0.05), showing a dose-dependent effect.

Consequently, AE improved mitophagy in damaged neurons of ad mice, enhanced mitochondrial fusion, and optimized mitochondrial fission (Figure 3B,D). p-AMPK and SIRT3 play important roles in regulating mitochondrial function, energy metabolism, autophagy, and mitophagy. The p-AMPK and SIRT3 expressions were analyzed using immunofluorescent (IF) double staining (Figure 3C). As illustrated in Figure 3D, the p-AMPK and SIRT3 levels in Alzheimer's disease (ad) mice were much lower than those in the WT group (p < 0.01). However, the reductions were significantly mitigated following treatment with 100 mg/kg of AE and donepezil (p < 0.05).

3.4 AE Promotes Proliferation of AD Model Cells

The workflow for neuronal cell processing is shown in Figure 4A. We conducted a CCK-8 assay on HT22 cells post-Alzheimer's disease modeling to determine the optimal concentrations of AE. At 7.5 μM AE, cell viability significantly decreased (p < 0.05), and at 50 μM, it dropped by nearly 50% (p < 0.01) (Figure 4B). We chose concentrations of 2, 4, and 6 μM to determine the optimal treatment level. As illustrated in Figures 4C and 6 μM AE exhibited the most significant protective effect on ad-modeled HT22 cells (p < 0.01). The cell morphology observed using an optical microscope is illustrated in Figure 4D. In contrast to the control group, the model group showed a reduction in cell count, with cells appearing shrunken and exhibiting broken synapses. Treatment groups with 2, 4, and 6 μM AE had better cell conditions, with fewer shrunken cells and improved intercellular connections. We conducted an EDU assay to observe DNA synthesis over 3 h, as shown in Figure 4D,E. The Model group exhibited a marked reduction in DNA synthesis by contrast with the controls (p < 0.01). After therapy, DNA synthesis gradually rose, with the 6 μM AE group showing the most significant rise (p < 0.01). Thus, AE demonstrates protective effects on ad-model cells and encourages their proliferation.

3.5 AE Improves Mitochondrial Function in AD Cell Models

We evaluated mitochondrial function in HT22 cells by measuring Δψm (Figure 5A,D), ROS (Figure 5B,D), Ca²⁺ (Figure 5C,D), mitochondrial superoxide (Figure 5E,F), and GSH levels (Figure 5G). HT22 cells processed using the Aβ_25-35 exhibited reduced Δψm and GSH levels and increased ROS, Ca²⁺, and superoxide compared to the controls (p < 0.01). Consequently, this indicated significant mitochondrial damage and dysfunction associated with ad progression. In contrast, the AE and RAPA groups demonstrated trends that were markedly different from those observed in the model group (p < 0.05). These groups showed improvements in mitochondrial function, although the levels did not return to normal. These findings indicate that AE enhances mitochondrial function in ad model cells, with autophagy potentially contributing to mitochondrial protection.

3.6 AE Activates Mitophagy and Attenuates AD Progression via the AMPK/PGC-1α/SIRT3 Pathway

Levels of p-AMPK, AMPK, PGC-1α, SIRT3, and mitophagy markers were assessed by western blotting (Figure 6A,B) and immunofluorescence (Figure 6C–E). The model group had lower PGC-1α, p-AMPK, and SIRT3 levels than the controls (p < 0.05), alongside a slight decline in mitophagy. AE treatment significantly increased Beclin1, LC3B, PGC-1α, p-AMPK, and SIRT3 (p < 0.05) while decreasing P62 (p < 0.05), indicating enhanced mitophagy levels.

3.7 The Knockdown of SIRT3 Increased Mitochondrial Damage and Reduced Mitophagy in HT22 Cells

To study SIRT3's effects on mitochondrial dynamics and mitophagy in HT22 cells treated with Aβ_25-35 and AE, we transfected the cells with SIRT3 siRNA. We then employed western blotting and immunofluorescence to observe changes in these proteins. RT-qPCR was utilized to identify the most effective transfection sequence, as shown in Figure 7A,B. The results showed that sequence 3 demonstrated the highest transfection efficiency (p < 0.001), and this sequence was chosen for future tests. Western blotting evaluated fusion proteins and mitochondrial fission among the Control, Model, and AE groups after SIRT3 siRNA transfection (Figure 7C,D). Before transfection, the Model group had higher DRP1 expression (p < 0.01) and lower OPA1 expression (p < 0.001) than the controls. After AE treatment, DRP1 decreased and OPA1 increased in contrast to the Model group (p < 0.05). After transfection, DRP1 levels rose (p < 0.001), while OPA1 levels decreased (p < 0.001) in all groups.

Immunofluorescence was utilized to observe changes in mitophagy-related proteins (Pink1, Parkin) following SIRT3 siRNA transfection, as shown in Figure 7E,F. The analysis indicated that the expressions of Pink1 and Parkin in the Model group remarkably surpassed those in the controls (p < 0.001). After treatment, the gene expressions were lower than those in the Model group (p < 0.05). Pink1 and Parkin levels in the total transfected groups exceeded those in the non-transfected Control, Model, and AE groups (p < 0.05) (Figure 7G). These findings indicate that mitochondrial damage is increased and mitophagy is decreased in HT22 cells transfected with SIRT3 siRNA. SIRT3 is an important protective player in sustaining mitochondrial function among cells modeled for Alzheimer's disease.