The neonate Sprague–Dawley rats (4–8 g) from Sino-British SIPPR/BK Lab Animals Ltd. (Shanghai, China) were used. The animals were kept under control conditions (12-h light/dark cycle, temperature 23 ± 1°C, relative humidity 55 ± 10%) with free access to water and food. The experiments were performed according to the guidelines of the National Institutes of Health Guide for the Care and Use of Laboratory Animals and approved by the Animal Care and Use Committee of Fudan University. Efforts were made to reduce the number and suffering of animals used in the study.

The procedure was performed as previously described (22, 23). MSG (Sigma-Aldrich, St. Louis, MO, United States) was dissolved in 0.9% NaCl solution to the final concentration of 10%. Briefly, the experimental animals were subcutaneously injected with MSG at a dose of 4 mg/g on the postnatal 2nd, 4th, 6th, 8th, and 10th days. The control animals received an equimolar concentration of sodium chloride. The animals were weaned on the 21st day, and the female rats were removed from the cage.

ZSQGY consists of 12 crude drugs including Radix Rehmanniae, Dioscoreae Rhizoma, Radix Bupleuri, Radix Paeoniae Alba, Radix Angelicae Sinensi, Rhizoma Anemarrhenae, Fructus Corni, Cortex Moutan, Poria, Ziziphi Spinosae Semen, Rhizoma Alismatis, and Gardeniae Fructus. The composition of ZSQGY is listed in Table 1. All crude drugs were obtained from Zhongshan Hospital. Briefly, the crude herbs were boiled in 10 times the volume of drinking water for 1 h for three times. Suspensions filtered from three decoctions were mixed and centrifuged at 2,000 g for 20 min. The collected suspension was soaked in 100% ethanol under rapid agitation, followed by stirring overnight. Then the suspension was collected and centrifuged at 2,000 g for 20 min and concentrated to 2 g/mL (w/v) before autoclaving.

Liquid chromatography-mass spectrometry (LC–MS) was established for the qualitative analysis of chemical compounds of ZSQGY. The API 4000 tandem mass spectrometer (Applied Biosystems/MDS SCIEX, United States) equipped with an electrospray ionization source (ESI) was used in the LC–MS analysis. All targeted analyses were performed in positive and negative ion modes. The curtain gas (CUR) and collision gas (CAD) used high-purity (99.99%) nitrogen. The ionspray voltage (IS) was set 3,500 V and the ionspray temperature (TEM) was 450°C. The 100 μL samples were dissolved in a 1 mL solution containing methanol and filtered through a 0.22 μm membrane. The content of the components in ZSQGY was calculated using the point external standard method.

The animals (8 weeks old) were randomly divided into the control group, the model group, the L-ZSQGY group, the M-ZSQGY group, the H-ZSQGY group, and the FXT group. The dose of ZSQGY that translated from human beings to animals in this study was established using the body surface area normalization method. For ZSQGY treatment, the common human daily dose in the present study is 133 g/70 kg bodyweight. According to the formula drat = human × 0.71/0.11, we selected 12, 24, and 48 g/kg/day as low, middle, and high doses, respectively. For fluoxetine treatment, the daily dose required for an adult is 20 mg/70 kg, which is converted to 0.184 mg/100 g for a rat. The treatment dosage for rats was 6.45 times that of patients. The animals in the control group and mode group were taken in the same volume as that of normal saline. All drugs were given once daily continuously for 4 weeks.

The rats were placed in an acrylic cylinder (height 40 cm, and diameter 20 cm) filled with water (depth 20 cm) at a temperature of 25 ± 1°C for 15 min (the pretest session). After 24 h, the rats were forced to swim again for 5 min. The duration of immobility (s) was analyzed. The rats remained floating in an upright position, with small movements of the four limbs to keep its head above the water surface, and were judged to be immobile.

Briefly, the rats were presented with two bottles containing 1% sucrose solution in the home cage for 24 h. Then, either water or 1% sucrose solution was placed in a random order in the home cage for another 24 h and the bottle order was exchanged every 12 h to account for side preference. The animals were deprived of water for 12 h before the test and then presented with two pre-weighed bottles containing 1% sucrose solution or water. The intake of water, sucrose solution, and total fluid intake were measured after 18 h. The percentage of sucrose solution intake in relation to the total fluid intake was calculated. The impaired sucrose intake was indicative of the depression-like behavior.

Central and peripheral areas were divided in the open field apparatus. The animal was placed individually in the open field area (50 × 50 × 50 cm) for 5 min, and the locomotor activity was videotaped (Ethovision 9.0, Noldus). Time spent in center (s) and latency to center (s) were analyzed.

Golgi staining and spine density analysis were performed according to the manufacturer’s instructions (Hito Golgi-Cox OptimStain kit, Hitobiotec Corp, Kingsport, TN, United States). The medial prefrontal cortex (mPFC) was cut into 80 μm thickness coronal tissue sections using a freezing microtome (Microm HM 450, Waldorf, Germany). Neurons in the mPFC area were analyzed. Dendritic spine density was counted by randomly selecting the secondary and tertiary apical dendrites and was shown as the number of thorns/10 μm dendrite. In total, three segments were counted per section, and three slides were chosen from each rat.

The rats were given anesthesia with sodium pentobarbital (40 mg/kg) intraperitoneal injection, and the mPFC were removed and incubated in 2.5% glutaraldehyde at 4°C for 24 h. Then, the mPFC were cut into 1 mm³ segments and fixed in 1% osmium tetroxide for 2 h. The tissue segments were rinsed and dehydrated in a series of graded aqueous ethanol and embedded in Epon. A total of 70 nm ultrathin sections were prepared and stained with 3% uranyl acetate and 0.5% lead citrate. Images were taken using the TEM (6,200× magnification) and analyzed using Image Pro Plus.

The rats were anesthetized with sodium pentobarbital (40 mg/kg) intraperitoneal injection and then transcardially perfused with 4% paraformaldehyde. The brains were removed and then paraffin was embedded. Then the mPFC slides (10 μm) were incubated in PBS containing 0.03% H₂O₂ for 10 min and in 5% goat serum for 30 min, followed by the incubation with anti-PGC-1α (1:100; Abcam, Cambridge, United Kingdom) at 4°C overnight. After washing with PBS, the sections were incubated with a secondary antibody (1:200) for 1 h. Then slices were added an appropriate amount of horseradish enzyme and incubated at 37°C for 30 min. After washing with PBS, the slides were incubated in DAB. The nuclei were stained with hematoxylin. The slices were dehydrated with ethanol, sealed with gum, and further observed under the microscope for observation and analysis (Olympus BX51, Tokyo, Japan).

The total RNA was extracted using a Trizol Reagent (Introgen, Carlsbad, CA, United States) according to the instructions. The total RNA was reverse transcribed into cDNA using the cDNA synthesis kit (Thermo Fisher Scientific, Waltham, MA, United States). A real-time quantitative PCR analyzer was used to detect the expressions of PGC-1α mRNA and PSD95 mRNA. The obtained genes were normalized to GAPDH. The primers sequences were listed as follows: GAPDH, AACTCCCATTCCTCCACCTT, and GAGGGCCTCTCTCTTGCTCT; PGC-1α, AGGCAAGCAAGCAGGTCT, and GTCATCAAACAGGCCATCC; and PSD95, GCAGGTTGCAGATCGGAGAC, and CCAGGTGCTGAGAATATGAGGTT.

The analysis of the mtDNA amount was measured by quantitative PCR. The copy numbers of mtDNA were normalized to a nuclear-encoded gene RBM15.

The concentrations of malondialdehyde (MDA), 8-hydroxy-2-deoxyguanosine (8-OHdG), tumor necrosis factor-α (TNF-α), interferon-γ (IFN-γ), interleukin-1 (IL-1), 6 (IL-6) from the mPFC of rats, and cell serum were quantified using enzyme-linked immunosorbent assay (ELISA) kits (Shanghai Enzyme-linked Biotechnology, Shanghai, China).

The mitochondrial ATP level was detected using the ATP Assay kit (Beyotime Biotechnology, Shanghai, China) according to the manufacturer’s instructions.

The rats were administered with the ZSQGY (24 g/kg/day) for 3 consecutive days. After anesthetized with sodium pentobarbital (40 mg/kg), the abdominal aorta blood of the animals was collected 2 h after the treatments and centrifuged at 4°C for 20 min, followed by sterilization through a 0.22 μm microporous membrane. The serum was stored at −80°C.

High concentrations of CORT can induce the neurotoxicity of PC12 cells, which has been widely used as an in vitro model for depression. In this study, differentiated PC12 cells were maintained (at a density of 1 × 105 cells/mL) in RPMI-1640 supplemented with 10% heat-inactivated fetal bovine serum and 1% penicillin–streptomycin at 37°C under a humidified atmosphere of 5% CO₂ and 95% air (v/v) for 24 h. Plated PC12 cells were exposed to CORT (Sigma-Aldrich, St. Louis, MO, United States) at the level of 200 and 400 μM for 24 h, and then the PC12 cells were treated with different concentrations of the ZSQGY-containing serum (5, 10, and 20%) to determine the suitable doses of CORT and ZSQGY-containing serum by assessing with the CCK-8 kit (Beyotime Biotechnology, Shanghai, China).

PC12 cells were treated with the CCK-8 solution and incubated at 37°C for 4 h. The optimal density at 450 nm was measured with a microplate reader (Bio-Rad Laboratories, Hercules, CA, United States). Cell viability was presented as a percentage of control cells.

The PC12 cells were divided into the control group, the model group, the ZSQGY + PGC-1α siRNA group, and the ZSQGY + con-siRNA group. For the control group, PC12 cells were cultured under the abovementioned normal conditions. For the model group, PC12 cells were exposed to 200 μM CORT for 24 h. For the ZSQGY + PGC-1α siRNA group, PC12 cells were treated with PGC-1α siRNA before modeling and ZSQGY-containing serum treatment. For the ZSQGY + con-siRNA group, PC12 cells were treated with con-siRNA before modeling and ZSQGY-containing serum treatment. The PGC-1α siRNA and con-siRNA (Jikai Biology, Shanghai, China) were transfected with Lipofectamine 2000 (Invitrogen, Carlsbad, CA, United States).

Data are expressed as mean ± standard error. Data were analyzed by one-way analysis of variance (ANOVA) followed by the LSD multiple comparison test using software SPSS 20.0. Data were considered statistically significant at a value of p of less than 0.05.

Qualitative analysis of compounds in water extract of ZSQGY was carried out by LC–MS. A total of six compounds, including geniposide, paeoniflorin, albiflorin, ferulic acid, saikosaponin A, and ursolic acid in water extract were used for quantitative analysis (Figure 1). After calculation, this water extract contained 533.682 μg/mL of geniposide, 474.789 μg/mL of paeoniflorin, 225.892 μg/mL of albiflorin, 280.147 μg/mL of ferulic acid, 30.411 μg/mL of saikosaponin A, and 11.468 μg/mL of ursolic acid, respectively (Table 2).

The forced swimming test, the sucrose preference test, and the open field test were administrated to evaluate the effects of ZSQGY against depressive behaviors. In the forced swimming test, the rats in the model group spent longer immobility time during the test (p < 0.05; Figure 2A). The rats in the ZSQGY group and fluoxetine group spent shorter immobility time than the rats in the model group (p < 0.05; Figure 2A). In the sucrose preference test, we observed that compared with the control group, the consumption of sucrose in the model group decreased (p < 0.05; Figure 2B). However, the consumption of sucrose in rats that received ZSQGY or fluoxetine treatment increased significantly (p < 0.05; Figure 2B). In the open field test, the rats in the model group spent less time in the central area than those in the control group (p < 0.05; Figures 2C–E). The rats in the model group also traveled a shorter distance than those in the control group. In contrast, the rats who received ZSQGY or fluoxetine treatment spent a longer time in the central area and traveled a longer distance (p < 0.05; Figures 2C–E).

Results from the Golgi staining showed that compared with the control group, the density of dendritic spines in mPFC in the MSG group induced decreased significantly (p < 0.05; Figures 3A,B). The treatments of ZSQGY and FXT significantly reversed a decrease in dendritic spine density induced by MSG (p < 0.05; Figures 3A,B). To further corroborate the results from Golgi staining, the mPFC neuronal ultrastructure was examined by TEM. The PSD thickness in asymmetric synapses in the model group was significantly thinner than those in the control group (p < 0.05, Figures 3C,E). However, the treatments of ZSQGY and FXT remarkably ameliorated the reduction of the PSD thickness (p < 0.05, Figures 3C,E). In addition, the number of docked vesicles in asymmetric synapses was decreased in the model group (p < 0.05, Figures 3D,E). However, the reduction was reversed by the treatment of ZSQGY or FXT (p < 0.05, Figures 3D,E).

As shown in Figure 4, the levels of 8-OHdG and MDA in the model group were higher when compared with those in the control group (p < 0.05; Figures 4A,B). Contrarily, the levels of 8-OHdG and MDA decreased in rats that received the treatment of ZSQGY or FXT (p < 0.05; Figures 4A,B). In addition, the ATP content and mtDNA copy number of the model group were significantly decreased when compared with those in the control group (p < 0.05; Figures 4C,D). The application of ZSQGY eminently counteracted the decrease of ATP content and mtDNA copy number (p < 0.05; Figures 4C,D).

As shown in Figure 5, the levels of IL-1β, IL-6, TNF-α, and IFN-γ were elevated in the model group compared to those in the control group (p < 0.05; Figures 5A–D). Notably, ZSQGY and FXT reduce the content of inflammatory cytokines induced by MSG (p < 0.05; Figures 5A–D).

As shown in Figure 6, the level of PGC-1α mRNA significantly decreased in the model group compared with those in the control group (p < 0.05; Figure 6A). However, the level of PGC-1α mRNA significantly increased in the rats that received ZSQGY or FXT (p < 0.05; Figure 6A). The results of IHC showed that PGC-1α is stained brownish yellow or brown and expressed in both the nucleus and cytoplasm. In comparison with the control group, the mean optical density of PGC-1α was significantly decreased (p < 0.05; Figures 6A,B). In contrast, the mean optical density of PGC-1α was increased remarkably in the ZSQGY group and in the FXT group (p < 0.05; Figures 6A,B).

To screen out the appropriate concentration of CORT to damage PC12 cells as well as the cytoprotective effects of ZSQGY against CORT-induced injury, the PC12 cells were treated with high concentrations of CORT (200 and 400 μM) for 24 h, followed by treatment with different concentrations of ZSQGY-containing serum (5, 10, and 20%) for 24 h. As shown in Figure 7A, 200 μM CORT treatment caused a decrease of approximately 50% of viable cells. In addition, 10% ZSQGY-containing serum significantly increased cell viability, while there were no significant changes in the viability of cells in PC-12 cells treated with 5 and 20% ZSQGY-containing serum. Therefore, 200 μM CORT and 10% ZSQGY-containing serum were selected for the following studies.

To further explore whether PGC-1α is implicated in the beneficial effects of ZSQGY, the PGC-1α siRNA was used to inhibit its functions and then the levels of postsynaptic density 95 (PSD95), oxidative stress markers, and inflammatory cytokines were determined. As shown in Figures 7B–J, compared to the control group, the levels of oxidative stress markers (8-OHdG, MDA) and inflammatory cytokines (IL-1β, IL-6, TNF-α, and IFN-γ) were significantly increased in the CORT-induced PC12 cells, while the levels of PSD95 mRNA, ATP, and mtDNA were decreased in the CORT-induced PC12 cells (p < 0.05; Figures 7B–J). The ZSQGY-containing serum treatment notably decreased the levels of 8-OHdG, MDA, and inflammatory cytokines (IL-1β, IL-6, TNF-α, and IFN-γ) (p < 0.05; Figures 7B–J). This treatment also significantly increased the levels of PSD95 mRNA, ATP, and mtDNA (p < 0.05; Figures 7B–J). However, the beneficial effects of ZSQGY in synaptic plasticity, mitochondrial function, and neuroinflammation were blocked after the administration of PGC-1α siRNA (p < 0.05; Figures 7B–J).