A total of 56 adult healthy C57BL/6 mice (2–4 months old) were used in this study, purchased from Henan Skobes Biotechnology Co, Ltd. The mice were housed in a controlled environment with a temperature of 22 °C ± 2 °C, a 12-h light/dark cycle, and free to food and water. The bedding was changed every 7 days. After the implantation of stimulating electrodes, each mouse was housed individually. The study protocol received ethical approval from Xinxiang Medical University’s Animal Ethics Committee (approval no. XYLL-20250316).

The chronic unpredictable mild stress (CUMS) protocol involved a combination of six short-term stimuli and four long-term stimuli over a 14-day period as previously reported (Zhou et al., 2022). The short-term stimuli included restraint for 1 h, exposure to 4 °C for 1 h, exposure to 50 °C in a pre-warmed chamber for 5 min, exposure to pepper smell for 4 h, cage shaking for 20 min, and tail pinch for 2 min. The long-term stimuli, each lasting 24 h, included water and food deprivation, cage tilting at 45°, and cage tilting without bedding. Each day, mice were exposed to a combination of two randomly selected short-term stimuli and one randomly selected long-term stimulus. The same stimulus was not applied on consecutive days to prevent habituation. In addition, each stimulus should appear with roughly equal frequency throughout the entire stress period.

Mice were anesthetized, and shaved the hair on their top of the head. They were then secured in a stereotaxic instrument. Erythromycin ointment was used to protect the eyes of mice from light damage, and the mice were kept warm. 75% alcohol was used to disinfect all experimental equipment and the mice’s scalps were disinfected with iodine. A stimulating electrode was implanted into the right NAc core (coordinates: −1.1 mm anterior, +1.45 mm lateral, and −4.65 mm ventral to bregma) and a recording electrode was implanted into the left ventral hippocampus core (coordinates: −3.6 mm anterior, +2.8 mm lateral, and −3.0 mm ventral to bregma). Four holes were drilled around the exposed skull to screw in the skull nail and fix the electrodes. The electrodes were secured with dental cement. Mice were placed back into clean cages, kept separate from each other, and provided with ample supply of water and food after surgery, after surgery. Mice were allowed at least seven days to recover from surgery before DBS.

Starting from the 15th day after establishing the depression model, DBS was performed on the awake, freely moving mice. The stimulating electrode was connected to the stimulus isolator and waveform generator. Mice received a pulse wave current in the NAc area, with a frequency of 130 Hz, an amplitude of 100 μA and a pulse width of 60 μs for a duration of 1 h in their home cage. The same stimulation was received at the same time each day for one week (Kisely et al., 2018).

LFP was recorded using the Cerebus 64 channel signal acquisition system (Blackrock Inc., USA). Before the recording, mice were transferred to the testing room and acclimated for 30 min to ensure they were familiar with the environment. The LFP was recorded from mice in their homecage for 5 min when the signal was stable. The DBS treatment and LFP were staggered to avoid acute changes in LFP caused by DBS. Following the completion of behavioral and electrophysiological experiments, electrode localization was confirmed by histological examination of brain sections. Mice were excluded if the placement was inaccurate.

Before behavioral testing, mice were handled daily at least seven days to familiarize them with the experimenters. Once the mice became accustomed to gentle handling, such as gentle face touching, we ensured that the experimenter’s contact did no adversely affect their behavior. All experiments were conducted at the same time each day. Prior to behavioral testing, mice were transferred to the testing room and acclimated for 30 min to ensure they were familiar with the environment.

Mice were killed by cervical dislocation. Their brains were rapidly extracted, and the hippocampus was carefully isolated, stored at −80 °C. The hippocampus was homogenized in RIPA, and the supernatant was collected after centrifugation. Western blots was performed using 15% SDS-PAGE. Proteins were transferred to PVDF membranes, blocked with 5% skim milk for 1 h, and incubated overnight at 4 °C with primary antibodies. The next day, the membranes were incubated with secondary antibodies for 1 h at room temperature. Protein band gray values were analyzed and quantified using ImageJ software. The following antibodies were used: PSD-95 (Abcam, ab238135, 1:2000), BDNF (Abcam, ab108319, 1:1000), p-mTOR (Ser2448) (CST, 5536, 1:1000), mTOR (CST, 2983, 1:1000), p-AKT (Ser473) (Proteintech, 66,444-1-Ig, 1:2000), AKT (CST, 4691, 1:1000) and GAPDH (Proteintech, 81,640-5-RR, 1:5000).

Mice were killed by cervical dislocation. Their brains were rapidly extracted, and the hippocampus was carefully isolated, stored at −80 °C. The hippocampus was ground in liquid nitrogen. Total RNA isolation from hippocampal tissue was performed with Trizol reagent, with subsequent reverse transcription to cDNA. RT-qPCR was performed using SYBR Green method. The 2^−∆∆Ct method was used to calculate the relative expression levels of each target gene. The primers used in RT-qPCR were as follows:

The FD Rapid GolgiStain Kit (FD Neuro Technologies Inc., United States) was used to measure the synaptic spine density. Mice were killed by cervical dislocation. Their brains were placed in a mixture of equal volumes of Solution A and Solution. After 6 h, the mixture of Solution A and Solution B was replaced with a fresh mixture. After 14 days, the brains were transferred to Solution C, protected from light, and stored at 4 °C for 3 days. The brains were sliced into 100 μm sections using a Leica cryostat microtome and mounted on glass slides. The sections were dehydrated with ethanol at increasing concentrations (50, 75, 95 and 100%) and cleared with xylene. Finally, the sections were sealed with neutral gum and observed under a microscope. Typical hippocampal neurons were selected and photographed. We used an Eclipse Ci-L photographic microscope to select the target area of the tissue for 1,000 × imaging, ensuring that the tissue filled the entire field of view as much as possible during imaging to maintain consistent background lighting for each photo. After imaging, we used Image-Pro Plus 6.0 analysis software, uniformly adopting micrometers as the standard unit. For each 1,000 × image, we measured the number of mushroom-shaped dendritic spines, stubby dendritic spines, filopodia-like dendritic spines, and thin dendritic spines within the 30–90 μm length range of the 2nd or 3rd dendritic branch on a complete neuron in the center of the image, recording both the measured length and the count of each type of dendritic spine within that length. Spine identification is primarily classified according to the following quantitative parameters: (1) Mushroom spine: head diameter to neck diameter ratio > 1.3, and head diameter > 0.5 μm; (2) Thin spine: spine length to maximum width ratio > 3.0, and head diameter < 0.5 μm; (3) Stubby spine: spine length to maximum width ratio < 1.5; (4) Filopodia: Length > 2 μm, with no distinct enlarged head structures observed along the entire axon. Its diameter usually remains constant and extremely slender < 0.3 μm.

Mice were anesthetized and perfused transcranially with 0.1 M PBS followed by 4% PFA. Their brains were removed and placed in 4% PFA at 4 °C for 24 h, then dewatered in 30% sucrose. Subsequently, the brain tissue was cut into 40 μm slices with a Leica cryostat microtome. The brain slices were taken out from the −20 °C refrigerator, rewarmed, and transferred to a culture dish containing 1 × TBS solution. The required slices were selected and moved into a 12-well plate. The slices were washed three times with 1 × TBS solution on a shaker for 15 min each. They were then blocked for 1 h in TBS++ solution (1 × TBS solution with 3% goat serum protein and 0.25% TritonX-100). The slices were incubated overnight at 4 °C (for more than 12 h) in TBS++ solution containing the corresponding primary antibody. After washing three times with 1 × TBS solution for 15 min each, the slices were blocked for 15 min in TBS++ solution. The secondary antibody, corresponding to the host of the primary antibody, was added to the TBS++ solution and incubated for 2 h. Note that all steps after adding the secondary antibody should be performed in the dark. The slices were washed three times with 1 × TBS solution, for 15 min each, then transferred to a dish containing 1 × PBS solution. Suitable slices were picked and placed on glass slides. After air-drying, an appropriate amount of anti-quenching sealing agent was added, and a cover slip was placed on top. The staining results could then be observed and photographed under a fluorescence microscope (Nikon, Japan).

Experimental data were analyzed and graphed using GraphPad Prism software. The sample size estimation was conducted using the software G*Power, by assuming a type 2 error protection of 0.05 and a power of 0.80. Unpaired t-tests were used to compare two groups. Data containing more than two groups were tested by using analysis of variance (ANOVA) followed by a Tukey’s multiple comparisons test. All data are expressed as the mean ± SD. Statistical differences were considered when p < 0.05.

After two weeks of CUMS (Figure 1A), the CUMS group exhibited significantly increased immobility time in the TST (Figure 1B, P < 0.01) and weight loss (Figure 1C, P = 0.0006) compared to the control group. The body weight was monitored throughout the experiment as it serves as a key indicator of physical health and stress-induced metabolic alterations. Significant weight loss in mice is often associated with chronic stress exposure and may reflect reduced food intake, impaired metabolism, or energy balance dysregulation, further supporting the validity of the depression-like model (Yang et al., 2016; Jeong et al., 2013). Additionally, the CUMS group exhibited a significantly lower preference in the SPT (Figure 1D, P < 0.001). In the OFT, the CUMS group revealed lower total distance (Figure 1E, P < 0.001), slower total mean speed (Figure 1F, P < 0.001), and avoidance of the center (Figure 1G, P < 0.001). As shown in Figures 1H,I, mice in CUMS group also performed worse in the OFT. These results attested that two weeks of CUMS treatment induced the depression-like behavior in mice.

Following one week of NAc-DBS treatment for the DBS-on group (Figures 2A,B), while the DBS-off group received no treatment, the DBS-on group exhibited a decreased immobility time in TST (Figure 2C, P < 0.001) and showed weight gain (Figure 2D, P < 0.001). Conversely, the DBS-on group demonstrated a higher sucrose preference (Figure 2E, P < 0.001). In the OFT (Figures 2I–K), the DBS-on group showed recovery in total distance traveled (Figure 2F, P < 0.001), total mean speed (Figure 2G, P < 0.001), and center preference (Figure 2H, P = 0.0004). However, no significant differences were observed between the DBS-on group and the control. These results indicated that the NAc-DBS treatment could alleviate CUMS-induced depression-like behavior in mice.

To evaluate whether CUMS could cause hippocampus malfunction, we recorded the hippocampal LFP in the home cage (Figure 2B). We classified neural oscillations according to frequency into: Theta oscillation (4–12 Hz), Beta oscillation (12–30 Hz), Low Gamma oscillation (30–50 Hz), and High Gamma oscillation (50–100 Hz). After CUMS treatment, the LFP power of high gamma band was reduced after CUMS treatment (Figures 3A,E, P = 0.0003), while there was no significant change in the theta band (Figure 3B, P = 0.9415). Additionally, the beta (Figure 3C, P = 0.0007) and low gamma (Figure 3D, P < 0.001) bands showed varying degrees of reduction. Following NAc-DBS treatment, the LFP power of high gamma band in DBS-on group was restored compared to the DBS-off group (Figures 3F,J, P = 0.0290). Interestingly, the power of theta (Figure 3G, P = 0.0305) and beta (Figure 3H, P = 0.4392) bands showed an opposite trend. Moreover, the low gamma (Figure 3I, P = 0.0379) band was also restored. These results demonstrate that the LFP power of high gamma band in mice hippocampus is associated with CUMS-induced depression-like behavior and it can be restored by NAc-DBS treatment.

The PSD-95 is an key scaffolding protein on the postsynaptic membrane of excitatory neurons, which is closely linked to synaptic plasticity (Levy et al., 2015; Xu, 2011). We found that PSD-95 protein expression in the hippocampus of CUMS group mice was reduced compared to the controls. However, NAc-DBS reversed this phenomenon (Figures 4A,B, P < 0.001). RT-qPCR results mirrored those of Western blots (Figure 4C, P = 0.0004). Golgi staining showed that the number of spines in the CA1 (Figures 4D,E, P < 0.001) region of the hippocampus decreased in DBS-off group mice but increased in DBS-on group mice. Furthermore, no density difference was found between mature spines (mushroom and stubby), but immature spines (thin and filopodium) showed a significant increase in DBS-off mice and decrease in DBS-on mice (Figures 4F–H, P = 0.0209). These results suggested that NAc-DBS reverse the decrease of synaptic spine density in hippocampus of CUMS mice.

To investigate whether NAc-DBS relieves the hippocampal injuries and mitigates depression-like behavior in mice by enhancing the BDNF protein and activating the mTOR pathway, we used Western blots and RT-qPCR to detect changes in mTOR pathway. Western blots, RT-qPCR and immunofluorescence were employed to assess BDNF expression in mice hippocampus. Results showed that the gene levels of BDNF protein and mTOR pathway decreased in the hippocampus of DBS-off group mice compared with the control group and the protein levels of BDNF and p-mTOR/mTOR ratio were also decreased. Although NAc-DBS did not alter the total protein expression of the mTOR pathway in the hippocampal tissues of DBS-on mice, BDNF and mTOR mRNA levels (Figure 5J, P < 0.001) for mTOR and the protein expression of BDNF (Figures 5H,I, P < 0.001) and p-mTOR (Figures 5K,L, P < 0.001) were increased by NAc-DBS. Immunofluorescence results in the CA1 (Figures 5B,E, P < 0.001), CA3 (Figures 5C,F, P < 0.001), and DG (Figures 5D,G, P < 0.001) regions of the hippocampus (Figure 5A) were consistent with these findings. Moreover, no significant differences were observed between DBS-on mice and control mice. BDNF acts as a precursor signal for AKT, and enhancing BDNF expression activates the AKT signaling pathway. We also detected AKT expression at the protein (Figures 5K,M, P < 0.001) and gene levels (Figure 5O, P < 0.001), with results similar to those for mTOR. These results indicate that NAc-DBS alleviates CUMS-induced depression-like behavior in mice by enhancing BDNF protein and activating the AKT/mTOR pathway.

To further determine whether NAc-DBS take antidepressant effect through the mTOR pathway, we used rapamycin, a potent and specific mTOR inhibitor. Mice in the rapamycin group were administered an intraperitoneal (i.p.) injection of 2 mg/kg rapamycin 2 h before NAc-DBS treatment every day (Figure 6A) (Xu et al., 2020). In the TST, the immobility time of the rapamycin group was significantly longer than that of the control and DBS-on groups (Figure 6B, P < 0.001). The body weight and sucrose preference of the rapamycin group were significantly lower than those of the control and DBS-on group (Figure 6C, P = 0.0006; Figure 6D, P = 0.0001). Additionally, the total distance (Figure 6E, P = 0.0003), total mean speed (Figure 6F, P = 0.0004), and the center preference (Figure 6G, P = 0.0005) in the OFT were significantly affected in the rapamycin group compared to the control and DBS-on groups. As shown in Figure 6H, mice in the rapamycin group performed a worse performance in the OFT than those in control and DBS-on group. We then detected the PSD-95/BDNF protein and AKT/mTOR pathway. The results showed that the AKT/mTOR pathway which activated by NAc-DBS was inhibited by the rapamycin (Figures 6I–K, (J) p = 0.002; (K) p < 0.001). Furthermore, the expression of PSD-95/BDNF protein, which was reversed by NAc-DBS, was inhibited by rapamycin (Figures 6L–N, (M) p < 0.01; (N) p < 0.001). These results suggest that NAc-DBS alleviates CUMS-induced depression-like behavior in mice by activating the mTOR pathway.