PTE (Figure 2A) and RES (Figure 2B), purchased from Dalian Meilun Biotechnology (CHN), respectively, were dissolved in saline with 0.1% dimethyl sulfoxide (DMSO). Aβ_25–35, purchased from Sigma-Aldrich (United States), was dissolved in saline at a concentration of 3 mM and incubated at 37 °C for 5 days to induce the formation of aggregated Aβ_25–35 (Zhu et al., 2021). All other chemical reagents used in this study were of analytical grade.

Male KM mice (4 week-old) were purchased from Liaoning Changsheng Biotechnology Co. Ltd., and kept on a 12-h light–dark cycle in a temperature-controlled room at 20 ± 2°C with a relative humidity of 55 ± 5%. The mice were allowed ad libitum access to food and water.

After a week of adaptation, the mice received a single intracerebroventricular injection (i.c.v.) of Aβ_25–35 (9 nmol/3 μl), and the sham-operated animals underwent the same surgery with the infusion of saline (Maurice et al., 1996). After 24 h, the mice were subjected to intragastric administration of PTE (10 or 40 mg/kg/d) or RES (40 mg/kg/d) until the mice were killed. Behavioral tests were started on day 5 after injection. The experimental procedure is schematically represented in Figure 1A.

The mice were anesthetized with 20% urethane (10 ml/kg, i.p.). The brains were fixed in 4% paraformaldehyde overnight and then dehydrated in 20 and 30% sucrose until they sunk. The tissues were cut into 15-μm-thick coronal sections. The sections were stained with neuronal nuclear protein (NeuN) antibody (1:400, CST, United States) overnight at 4 °C and incubated with an anti-rabbit secondary antibody (1:500, Invitrogen, United States) for 2 h at room temperature. Images were captured by using the Nikon microscope (Eclipse Ci-E, JP).

TEM was performed with some modifications, as previously described (Zhu et al., 2021). The tissues were fixed with 2.5% glutaraldehyde dissolved in 0.01 M phosphate-buffered saline (PBS) overnight at 4°C. After removing and post-fixing by immersion in the fixative overnight at 4°C, the samples were treated with 1% osmium tetroxide for 3 h, dehydrated with graded ethanol and acetone and embedded in resin. Then, the sections were stained with 4% uranyl acetate and 0.5% citrate. The ultrastructure, especially the neuronal nuclei, synaptic, and mitochondrial structure of the parietal cortex and hippocampal CA1, was detected and the images were taken by using the transmission electron microscope (JEOL H-7650, JP).

Brain tissue or primary neuronal cells were homogenized in ice-cold lysis buffer with protease inhibitors and phosphatase inhibitors, and the protein concentrations were determined by the BCA assay (Beyotime Biotechnology, CHN). The samples were subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE). Briefly, equal amounts (20–30 μg) of proteins were loaded per lane, transferred to 0.45-μm polyvinylidene difluoride (PVDF) membrane, and blocked for 2 h at room temperature. Afterward, the membranes were incubated overnight at 4°C with primary antibodies: postsynaptic density protein 95 (PSD-95) (1:1000, CST, United States), synapsin-1 (SYN-1) (1:1200, CST, USA), NeuN (1:1000, CST, United States), SIRT1 (1:1000, Abcam, UK), Nrf2 (1:1000, Abcam, UK), Bcl-2-associated X protein (Bax) (1:4000, Proteintech, United States), B-cell lymphoma 2 (Bcl-2) (1:800, Proteintech, United States), and β-actin (1:400, Santa Cruz, United States). Then, the membranes were incubated with secondary antibodies (1:10,000, Jackson ImmunoResearch, United States). Protein bands were visualized with enhanced chemiluminescence (ECL). Then, the intensity of the bands was quantified by ImageJ software (version 1.8.8, National Institutes of Health, United States) and corrected with the corresponding β-actin level. The results were expressed as values normalized to the loading sham.

The tissues were lysed with a lysate buffer and then centrifuged at 12,000 g at 4°C for 5 min. Afterward, an aliquot of the supernatant was used for detection. Superoxide dismutase enzyme activities were measured using a commercial SOD Assay Kit-WST (Beyotime Biotechnology, CHN) in accordance with the manufacturer’s protocol. The values were recorded using a microplate reader (CLARIOstar, BMG LABTECH, GER). The concentration of total proteins in the cells was quantified by using the BCA Kit (Takara, JP).

The cerebral cortical primary neurons were obtained from newborn KM mice within 24 h as per previously described procedures (Zhu et al., 2018). The neurons were plated onto 96-well dishes or glass slides before coating with poly-l-lysine. The cultures were maintained in Neurobasal/B27 medium, which was changed every 2–3 days. After culturing for 5 days, the primary cortical neurons were treated with aggregated Aβ_25–35 (20 μM) and drugs for 24 h, as shown in Figure 1B.

The viability of the primary cortical neurons was evaluated by using a Cell Counting Kit-8 (CCK-8) (Vazyne Biotech, CHN) following the manufacturer’s protocol. The neurons were (1 × 10⁶/well) cultured in 96-well plates. After culturing for 5 days, the neurons were pretreated with Aβ_25–35 (20 μM) for 30 min and then incubated with PTE (0.5, 2, or 10 μM) for 24 h at 37°C. Subsequently, 10 μL CCK-8 solution was added into each well for an additional 2 h. The viability was detected by using a microplate reader (CLARIOstar, BMG LABTECH, GER).

The neurons grown on poly-l-lysine-coated glass slides were fixed with 4% paraformaldehyde solution in PBS for 20 min and permeabilized with 0.3% Triton-X100 in PBS for 10 min at room temperature. Then, the neurons were incubated with primary antibody microtubule-associated protein 2 (MAP-2) (1:200, Santa Cruz, United States) overnight at 4°C. After washing with PBS, the neurons were incubated with an anti-mouse secondary antibody (1:500, Invitrogen, United States) for 2 h at room temperature. Images were captured by using the Leica fluorescence microscope (DMI8, GER), and the neurons in the images were analysis by ImageJ software.

JC-1 is a membrane-permeable lipophilic dye that accumulates in the mitochondria. JC-1 exists as J-monomers in the cytoplasm (green fluorescence) and aggregate to form J-aggregates in the mitochondrial matrix (red fluorescence). Mitochondrial depolarization can be measured as an increasing green/red fluorescent intensity ratio to detect neuronal early apoptosis (Brooks et al., 2013).

Neurons were stained with the JC-1 dye (Beyotime Biotechnology, CHN) to detect the mitochondrial membrane potential. The neurons were treated with JC-1 working fluid for 20 min at 37°C. Then, the stained neurons were rinsed twice using JC-1 staining buffer, and a fresh medium was added. Images were captured by using the Nikon microscope (Ts2, JP). ImageJ software was used to for fluorescence intensity analysis.

All statistical analyses were carried out using SPSS 26.0 software (SPSS Inc, Chicago, IL, United States). The data were analyzed using one-way analysis of variance (ANOVA), followed by the LSD multiple comparisons test with homogeneity or Dunnett’s T3 test with heterogeneity of variance. The data were expressed as mean ± SD. p < 0.05 was considered statistically significant.

PTE has a chemical structure similar to that of RES, as shown in Figures 2A,B. Studies have demonstrated that treatment with RES exerts neuroprotective effects depending on SIRT1 in AD (Martin, 2017; Gomes et al., 2018). To detect the role of PTE, we detected the expression of SIRT1 in AD mice. The results showed that treatment with PTE and RES both increased the expression of SIRT1, whereas no significant difference between PTE (40 mg/kg) and RES (Figure 2C). These results indicate that PTE is a similar compound to RES and probably has the effect of regulating SIRT1 to improve learning–memory deficits in AD mice.

To assess the effects of PTE and RES on learning–memory, we performed the behavioral tests. First, the imaginal memory was evaluated in this study using the novel object test. The results revealed that PTE-treated mice showed more interest in the novel object and spent more time exploring the novel object than the previously studied object (Figure 3A). In the training phase, there was no significant difference in the time of exploring the two identical objects between groups, while both PTE and RES treatments exhibited an increased exploration time to explore the novel object in the testing phase (Figure 3B). Then, we detected the working memory ability in the Y-maze test. The total number of arm entries is shown in Figure 3C, and no significant differences were found among each group (Figure 3C). However, both PTE and RES treatment significantly increased spontaneous alternation behavior (Figure 3D). To investigate the effects of PTE and RES on spatial learning–memory, we used the Morris water maze test in the study. During the training session, compared with the model group, the PTE-treated mice spent significantly less time to locate the platform on the fourth day and the RES group on the fifth day (Figure 4A). Representative pictures of the path chart in the probe trial of the Morris water maze are shown in Figure 4B. In the probe test on the 6th day, treatment with PTE notably increased the number of crossing the platform and the time spent in the target quadrant but had no significant differences in the swimming speed (Figures 4C–E). These results suggest that PTE could more comprehensively improve learning–memory deficits in AD mice.

It is well known that neurons and their synaptic structures are the basis of learning and memory. Herein, the PTE-treated mice exhibited an increase in the expression of NeuN-positive cells both in the cerebral cortex and hippocampus compared with the model mice (Figure 5A). The synaptic structure is composed of the presynaptic membrane, synaptic gap, and postsynaptic membrane. We used transmission electron microscopy to examine the ultrastructure of the cerebral cortex and hippocampal CA1 region in mice. The results showed that PTE improved ambiguous synaptic structure compared with the model mice (Figure 5B). In addition, protein analysis revealed that neuronal proteins NeuN, PSD-95, and SYN-1 were significantly increased after treatment with PTE (Figures 5C–F). The results also found that PTE administration improved mitochondrial crest impairment (Figure 5B). These data indicate that treatment with PTE could ameliorate neuronal and mitochondrial injury in AD mice.

To evaluate the protective effect of PTE on mitochondrial function, the level of apoptosis was determined. In this study, model mice showed chromatin condensation and chromatin accumulation along the inside of the nuclear membrane, while model mice treated with PTE displayed comparatively complete nuclear structures (Figure 6A). We further found that mitochondria-dependent apoptotic protein Bax was decreased, and Bcl2/Bax was increased in PTE-treated mice compared with model mice (Figures 6C,E,F).

The SIRT1/Nrf2 antioxidant pathway was introduced to elucidate the underlying mechanism of PTE in apoptosis caused by mitochondrial damage. The results showed that the expression of SIRT1 (Figure 2C) and Nrf2 (Figures 6C,D) was significantly increased in PTE-treated mice compared with the model mice. Meanwhile the level of SOD was also increased after treatment with PTE (Figure 6B). These results suggest that PTE administration modulates the SIRT1/Nrf2-related antioxidant effect and inhibits neuronal apoptosis to protect neurons in AD mice.

To further investigate the protective effect of PTE against Aβ_25–35-induced neuronal impairment, primary cortical neurons extracted from mice born within 24 h were used in vitro study. Herein, we found that 20 μM Aβ_25–35 treated for 24 h showed a significant decrease in neuronal viability (Figure 7A), whereas PTE at the concentration of 0.5 and 2 μM both significantly increased the neuronal viability against Aβ_25–35-induced neuronal damage (Figure 7B). In the following experiments, the concentration of 2 μM of PTE was used to investigate the neuroprotection. We observed that PTE (2 μM) could improve the neuronal structure (Figure 7C). Statistical results showed that the primary dendrites of neurons had no difference in each group (Figure 7E), while when compared with the administration of Aβ_25–35, the neurons treated with PTE could significantly increase the total dendrite length (Figure 7D) and the number of multistage dendrites (Figure 7F). These results suggest that PTE exerts protective effects to maintain the neuronal structure in primary cortical neurons.

In order to investigate whether the antiapoptotic effect of PTE is regulated by the SIRT1/Nrf2 antioxidant pathway, SIRT1 inhibitor EX527 was used in the study. The results showed that treatment with EX527 alone for 24 h did not affect neuronal viability (Figure 8A). PTE reduced neuronal damage caused by Aβ_25–35, and this protective effect of PTE was blocked by EX527 (Figure 8A). Further research found that primary cortical neurons treated with PTE could attenuate the ratio of green/red fluorescence intensity to stabilize mitochondrial membrane potential compared with the Aβ_25–35 group (Figures 8B,C). However, EX527 reversed the antiapoptotic effect of PTE on stabilizing mitochondrial membrane potential (Figures 8B,C). Collectively, these data indicate that PTE could produce an antiapoptotic effect to protect neurons by regulating SIRT1.