Anshen Buxin Six Pills (Lot number: 180908) was purchased from Ulanhot Sino-Mongolian Pharmaceutical Co. Ltd. (Ulanhot, Inner Mongolia, China). Anshen Buxin Six Pills powder (10 g) was dissolved in 100 ml of methanol and extracted via ultrasonication of 600 W for 1 h using a ultrasonic cleaning machine (SB-25-12-DT), which was purchased from Ningbo Scientz Biotechnology Co.,Ltd. (Ningbo, Zhejiang, China). After filtration, the methanol extract of Anshen Buxin Six Pills (ASBX) was obtained via rotary evaporation. The solvent of the extract was removed by the oil pump and vacuum freeze-dried into powder to obtain the lyophilized powder of ASBX, which was stored in −20°C for the next experiment. Follow the above steps to get ASBX, then, the ASBX were heated with water in 100 ml and resuspended. Then gradient extraction was carried out, each solvent was 300 ml each time, extracted for 3 times, and the solvent was removed by rotary distillation, petroleum ether extract (PEE), dichloromethane extract (DME), ethyl acetate extract (EAE), water-saturated n-butanol extract (BUE), and aqueous extracts (AQE) were obtained. Finally, the solvent was evaporated by oil pump and vacuum freeze-dried to obtain the samples of each extraction layer.

Reserpine (83580, purity ≥ 99.0%) was purchased from Sigma-Aldrich (St. Louis, MO, United States). Fluoxetine (FLX, F830634, purity ≥98.0%) was purchased from Shanghai Macklin Biomedical Co., Ltd. (Shanghai, China), and Costunolide (COS, AB0612, purity ≥98.0%) was purchased from Chengdu Alfa Biotechnology Co., Ltd. (Chengdu, Sichuan, China). Dehydrocostus lactone (DHE, D91141, purity ≥98.0%) was purchased from Shanghai Acmec Biochemical Co. Ltd. (Shanghai, China). ELISA kit for 5-hydroxytryptamine (5-HT) and dopamine (DA) was purchased from Wuhan Genomei Technology Co., Ltd. (Wuhan, Hubei, China). ELISA kit for norepinephrine (NE), serotonin transporter (SERT) and norepinephrine transporter (NET) was purchased from Shanghai Jianglai Industrial Co., Ltd. (Shanghai, China).

Adult zebrafish (AB strains, 5 months-old, male:female, 1:1) were purchased from Shanghai FishBio Co. Ltd. (Shanghai, China) and acclimated in fish tanks purchased from Shanghai Haisheng Biotech Co., Ltd. (Shanghai, China) with a recirculating aquatic system. The system water contained KCl 0.05 g/L, NaHCO₃ 0.025 g/L, NaCl 3.5 g/L and CaCl₂ 0.1 g/L. The zebrafish embryos were incubated with Holt buffer (15 mM NaCl, 0.67 mM KCl, 0.03 mM NaHCO₃, 0.90 mM CaCl₂, pH 7.2). Adult zebrafish and zebrafish embryos were maintained under a photoperiod of 14 h light/10 h dark. All the animal studies have been approved by the Tianjin University of Traditional Chinese Medicine of Laboratory Animals Care and Use Committee (TCM-LAEC2016032).

The configuration of solutions of ASBX groups was the lyophilized powder of ASBX was dissolved with DMSO, and dilute to 0.8% DMSO with system water. The experimental grouping is described as follows. To investigate the antidepressant activity of ASBX, adult zebrafish were randomly divided into six groups (n = 10): control group (Con, 0.8% DMSO system water), model group (Mod, 0.8% DMSO system water), Fluoxetine group (FLX, 10 μmol/kg), and ASBX groups (ASBX-80 mg/kg, ASBX-20 mg/kg, and ASBX-5 mg/kg). The FLX group was used as the positive control. Except for the Con group, zebrafish were soaked with reserpine (40 μg/ml) for 1 h every day for 7 days to induce a depressive phenotype (Tang et al., 2019). After treatment with reserpine daily, adult zebrafish were injected intraperitoneally with the above sample solutions.

The experimental procedure for the NTDT is described below. The experiment was conducted according to the protocol (Liu et al., 2020), with little change. The part of the tank with water was equally divided into two analysis areas, which was defined as top and bottom. After the zebrafish adapted to the tank environment for 5 min, the swimming trajectory of zebrafish in the first 10 min was recorded and analyzed using Noldus Ethvision XT 15 software (Noldus Information Technology, Leesburg, VA, United States). The following parameters were measured: total distance traveled (cm), average velocity (cm/s), freezing duration(s), number of transitions from bottom to top, first latency to enter the top area (s), and time spend in the top area.(s).

The experimental procedure for Nissl staining and the determination of neurotransmitter content are described below. The brains of the zebrafish were stained with Nissl stain solution (Beijing SolarBio Science & Technology Co., Ltd., Beijing, China) according to the manufacturer’s instructions, and digital images were taken with a Nikon CiS microscope (Nikon). Image-Pro Plus software (version 6.0) was used to analyze the images and count the number of Nissl-positive cells in the periglomerular gray zone (PGZ) and ventrolateral nucleus of the semicircular torus (TSvl). The concentrations of 5-HT, DA, and NE in the brain tissues of zebrafish were detected using Enzyme-linked immunosorbent assay (ELISA) kits for 5-HT, DA and NE, according to the manufacturer’s instructions.

The light/dark preference test was used to examine the effects of the extract layer and components of ASBX on reserpine-induced depression-like behavior in zebrafish larvae, analyzed by the zebrafish behavior trajectory analysis system. Reserpine was dissolved in 3 ml of Holt buffer containing 0.1% DMSO. The embryos (30/plate) were kept in 12-well plates with Holt buffer containing reserpine (4 μg/ml) or 0.1% DMSO Holt Buffer (Con group) at 6 h post fertilization and maintained in the dark for 6 days to obtain zebrafish larvae for the test. The establishment of depression phenotype of zebrafish larvae was referred to the methods (Wang et al., 2019) in the literature.

Freeze-dried extraction layers different extraction layers (PEE, DME, EAE, BUE, AQE) and lyophilized powder of ASBX were dissolved with DMSO and diluted with Holt buffer to the content of DMSO was 0.1%. The experimental groups are described as follows. To screen the anti-depression extraction layer(s) in ASBX, zebrafish larvae treated with reserpine were transferred to solutions of different extraction layers (PEE, DME, EAE, BUE, AQE, 80 μg/ml), ASBX methanol extract (80 μg/ml), Mod (0.1% DMSO Holt Buffer), and FLX (Fluoxetine, 10 μM) in 12-well plates. The actual drug amount of each extraction layer is calculated according to the equivalent crude drug amount, indicating the proportion of substances in each extraction layer in ASBX of 80 μg/ml. The FLX group was used as the positive control. The fractions (1–46 min) of PEE (10 mg/ml) from HPLC were collected into 10 ml EP tubes every 2 min for 23 fractions total. The contents of the EP tubes were then evaporated to dryness in a vacuum drying oven and dissolved in 3 ml Holt buffer. A sample that matches a peak of the corresponding time in the chromatogram is considered one component. COS and DHE groups (0.1, 1 and 10 μmol/L) were used to verify the antidepressant activity of the screened components from the light/dark preference test.

The light/dark preference test was then conducted according to the protocol (Wang et al., 2019) with some modifications. Briefly, the culture medium with reserpine of zebrafish larvae was replaced with drug solutions, soaked for 2 h. Zebrafish larvae treated with above mentioned drug solutions were transferred to each well of a 96-well plate (length × width: 125 × 85 mm, well diameter 10 mm), with 300 μl of Holt buffer, and adapted for 3 min before the test.

The light conditions are described as follows. The setting of the light and dark stimulus program was started 5 min after beginning video recording of the behavior trajectory. For light stimulation, the program was set as follows: light was applied for 3 min, off for 3 min, and repeated for three cycles. The video recordings were analyzed using Noldus Ethovision XT 15 software (Noldus Information Technology, Leesburg, VA, United States) by tracking the center of mass of the individual fish over time. Parameters were measured as the total distance traveled (mm) and average velocity (mm/s).

After determining PEE was the active extraction layer, PEE was dissolved in methanol with concentration of 0.1 mg/ml, then subjected to HPLC-Q-TOF-MS/MS analysis. A diamonsil^® C18(2) (250 × 4.6 mm, 5 μm) column was used to separate and detect the components of ASBX extraction, with the column temperature maintained at 40°C. The mobile phase was a gradient elution system of A (CH₃OH) and B (HCOOH:H₂O = 0.05:100), and elution was performed at a flow rate of 0.5 ml/min. The injection volume was 10 μl, and the gradient duration program was 0–2 min, 96% B; 2–6 min, 96–35% B; 6–41 min, 35%–0% B; 41–48 min, 0% B; 48–48.5 min, 0%–96% B; 48.5–56 min, 96% B.

Accurate mass measurements were collected using an AB Sciex Triple-TOF™ 4600⁺ mass spectrometer (Framingham, MA, United States) with an electrospray ionization (ESI) system. The ESI/MS spectra were acquired in the positive ion mode, and the ionizing voltage was set to 5,500 V. The temperature of the source was 300°C. The sample cone voltage was set to 30 V. High-purity nitrogen was used for nebulization and as the auxiliary gas. Ion source gases (N₂) 1 and 2 were maintained at 60 psi, and the curtain gas (N₂) pressure was maintained at 35 psi. The relative impact energy was 10 eV. The de-clustering voltage was 40 V. Peak View 2.2 was employed to process the data.

To study the antidepressant activity of COS, adult, reserpine-treated zebrafish were randomly divided into six groups (n = 10): control group (Con, 0.8% DMSO system water), model group (Mod, 0.8% DMSO system water), FLX group (Fluoxetine, 10 μmol/kg), and COS groups (COS-10 μmol/kg, COS-40 μmol/kg, and COS-160 μmol/kg). Other experimental conditions were the same as those described in section Novel Tank Diving Test.

The experimental conditions for Nissl staining and determination of the content of neurotransmitters (5-HT, DA, and NE) were the same as in Section Nissl Staining and Determination of Content of Neurotransmitter.

The expression levels of SERT and NET were quantified using ELISA kits (Shanghai Jianglai Industrial Co., Ltd., Shanghai, China) according to the manufacturer’s instructions.

GraphPad Prism Software (version 6.0) was used for the statistical analysis of results, and one-way analysis of variance was applied for data analysis, followed by the Tukey’s test for comparison of significant differences between the means. The results are expressed as the mean ± SEM. Statistical significance was set at p < 0.05.

To study the antidepressant effect of ASBX on zebrafish behavior, a NTDT was conducted. The NTDT takes advantage of the stress response of zebrafish in a new environment, which can be used to simulate anxiety, depression, and other behaviors (Benneh et al., 2017). As shown in the heat map in Figure 1A, exposure to reserpine resulted in low-frequency swimming at the top of the tank (simulating the effects of depression); this behavior was reversed by ASBX and FLX. Compared with the model group, the ASBX-80 mg/kg and FLX groups showed significantly improved swimming ability (total distance traveled and average velocity) (^*** p < 0.001, Figures 1B,C), freezing behavior (shortened freezing time) (^* p < 0.05, Figure 1D), and exploration behavior (increased transitions from bottom to top and shortened first latency to top area) (^* p < 0.05, Figure 1E). But there was no siginificant diffrences in First latency to enter the top area and Time sped inthe top area (Figures 1 F,G).

Based on the improved effect of ASBX on the swimming behavior of zebrafish, we further verified its antidepressant effect at the level of histopathology and biochemical indexes by Nissl staining and enzyme-linked immunosorbent assay (ELISA). Neuronal regeneration dysfunction is one cause of depression (Fang et al., 2018). Exposure to reserpine led to the reduction and disordered arrangement of Nissl bodies in cytoplasm, the distance between cells became larger, and the boundary is not clear in periglomerular gray zone (PGZ) (Figures 2A,C), conversely, ASBX-80 mg/kg and FLX groups reversed the reduction in the number of neurons in the PGZ and TSvl induced by reserpine, there are significant differences in statistical results.(^* p < 0.05, Figures 2B,D).

Neurotransmitters are closely associated with the neurobiological mechanisms of depression (Liu et al., 2018). ASBX-80 mg/kg and FLX groups significantly reversed the reduction in the levels of 5-HT (^* p < 0.05) after exposure to reserpine (Figure 2E).

Based on the antidepressant effect exhibited by ASBX, we further screened its components. We tested the survival rate of zebrafish embryos by ASBX, PEE, DME, EAE, BUE and AQE to determine the toxicity of each extract. There was no significant difference in the survival rate between each extract and the Con group, which proved that each extract had no obvious toxicity at this concentration (Supplementary Figure S2). Under photoperiod (dark/light alternating) stimulation, neurotoxic phenotypes at higher doses of reserpine show decreased swimming behavior in the zebrafish larvae (Wang et al., 2019). Compared with the Mod group, PEE had the best activity among the extraction layers, in terms of total swimming distance and average swimming speed (Figure 3B). Therefore, the PEE layer was chosen for further identification and separation by HPLC-Q-TOF-MS/MS (Figure 3A), and the antidepressant activity of the 23 fractions was evaluated (Figure 3B). Among them, fractions 26 and 28 improved the swimming ability of zebrafish larvae depression model induced by reserpine. Fraction 26 corresponds to the substance flowing out in 25–27 min in the ion flow diagram, and stream 28 corresponds to the substance flowing out in 27–29 min. According to the mass spectrum information, we identified that the components contained in fraction 26 are costunolide and costic acid, and the components contained in fraction 28 are dehydrocostus lactone (See for mass spectrum information in Supplementary Table S1). The antidepressant activities of compounds 1 and 2 were again verified by light/dark preference test. Zebrafish larvae treated with COS or DHE showed improved swimming activity of zebrafish larvae induced by reserpine (Figure 3C). The results thereby showed that COS and DHE were potentially the antidepressant components of ASBX.

To verify the antidepressant activity of COS, the NTDT was conducted, and the heat map shows that the depressive phenotype could be reversed by COS and FLX (Figure 4A). Compared with the Mod group, the COS-160 μmol/kg group showed improved swimming capacity (Figures 4B,C), freezing behavior (Figure 4D), and exploration behavior ( Figures 4E,F). The COS-40 μmol/kg group showed improved depressive behavior with respect to freezing duration, first latency to enter the top (Figures 4D,F). There was no sigificant difference in time spend in the top area between the COS groups and Mod group (Figures 4G). The results showed that COS could improve the swimming behavior of zebrafish with reserpine induced depression.

The anti-depression effect of COS was also verified at the histopathological and biochemical levels. The results of Nissl staining showed that the neuronal loss in PGA and TSvl in the Mod group was improved by COS (COS-160 μmol/kg) and FLX. (Figures 5A–D). Compared with the Mod group, in the COS-160 μmol/kg group, the levels of 5-HT and NE was increased. (Figure 5E). The results showed that cos had a protective effect on reserpine induced nerve injury, which may be related to the increase of 5-HT and NE.

The upregulation of SERT and NET leads to a lack of 5-HT and NE in the synaptic cleft, which may be one of the causes of depression (Chen et al., 2012; Li et al., 2019). The COS-160 μmol/kg treatment showed inhibition of expression of SERT and NET compared with Mod group (Figure 6).