Following a protocol developed by our laboratory (Chen et al., 2023), the powdered Cynomorium songaricum Rupr. pieces were sieved through a mesh sifter with a size of 60. The resulting powder was then soaked in 70% ethanol at a proportion of 1:10 (w/v) overnight. The following day, the filtrate was collected and subjected to three rounds of ultrasonic extraction. For the first extraction, 2 L of 70% ethanol was added to the filtrate, which was then subjected to ultrasonic extraction for 1 h. Afterward, the supernatant was collected, and 1 L of 70% ethanol was added for a second super-lift extraction, which lasted for 45 min. Subsequently, that supernatant was collected again; additionally, 1 L of 70% ethanol was added for a third ultrasonic extraction, which lasted for 30 min. This final supernatant was collected and subjected to rotary evaporation under heat until no alcohol taste/smell remained. The ethyl acetate polar component of the concentrated solution was then extracted with petroleum ether and ethyl acetate in a 2:1 (v/v) ratio. The ethyl acetate organic layer was then subjected to further concentration by rotary evaporation and subsequently freeze-dried for future use. The extraction rate of ECS was 2.36%, with impurities less than 2% and the total ash value less than 12%, all of which meeting the pharmacopoeia standards. Finally, the ECS powder was conserved at a temperature of −20°C.

Briefly, 1.0 g of ECS dried powder was accurately weighed and added to 50 mL of 70% ethanol solution through reflux for 1 h. Subsequently, the solution was filtered using a 0.22-µm membrane.

LC-MS/MS dissection was performed using an Ultimate 3,000 with a Waters ACQUITY UPLC HSS T3 C18 column (2.1 mm × 100 mm, 1.8 µm; Waters Corporation, Milford, MA, USA). Investigation and dissection on high-resolution MS were performed using a Q-Exactive Orbitrap mass spectrometer MS system (Thermo Scientific, Waltham, MA, USA) with a single electron spray ionization source. Specific experimental details are provided in Supplementary Approach One.

Metabolites were identified by matching the molecular formulas, retention times, and MS fragmentation patterns using corresponding standard materials, references, and web databases containing Chemical Name Search ¹ and Massbank ² .

The main ECS metabolites were identified using the following criteria: MW < 500, AlogP <5, Hdon <5, and Hacc <10. They were fed into the TCMSP ³ , SwissTargetPrediction ⁴ (Daina et al., 2019), SEA ⁵ (Keiser et al., 2007), and PharmMapper ⁶ (Wang et al., 2021) for target prediction. After eliminating the duplicates, we obtained the relevant targets. Furthermore, significant targets associated with neuroinflammation and depression were obtained from the Pubmed2ensembl ⁷ database. The Pubmed2 Ensembl resource, which is designed as a supplement to the BioMart mechanism for literature mining, was used to compile a list of proteins linked to depression-related neuroinflammation. The Ensembl Gene ID of GRCh37 human (Homo sapiens) genome assembly associated with neuroinflammation and depression was located by searching for the terms “neuroinflammation” and “depression” according to the Genome Reference Consortium.

The obtained targets were normalized in UniProt ⁸ , and the species was set to “human” after eliminating duplicate targets. The Venn tool ⁹ , an online software, was used to draw a Venn diagram to obtain the common target, which is a potential target for ECS in the treatment of neuroinflammation-related depression.

The analysis of protein–protein interactions (PPI) was conducted using STRING ¹⁰ database. The resulting data were saved as TSV files and imported into Cytoscape 3.8.2. To analyze the network topology, we used CytoHubba plugin within Cytoscape. The nodes were ranked based on their network centrality using the maximal clique centrality technique.

To detect the molecular functions (MF), biological processes (BP), and cellular components (CC) associated with the ECS-induced attenuation of depression-related neuroinflammation, these key targets were uploaded to the DAVID Bioinformatics Resources 6 ¹¹ for Gene Ontology (GO) enrichment analysis. Subsequently, Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment was performed on these key targets to screen for putative pathways related to the ECS-mediated treatment of neuroinflammation-induced depression.

Male SPF C57BL/6 mice aged 7 weeks were obtained from Wei tong Lihua Biotechnology Co., Ltd. (NO.SCXK (jing) 2021–0006, China). Each animal was housed in an SPF habitat. A 12-h light–dark cycle was maintained for all animals. Food and water were readily available throughout the study, except during the sucrose preference test. A constant temperature of 25°C ± 1°C was maintained, with a relative humidity of 55%–65% and continuous indoor air circulation. All animal experiments were performed in compliance with the regulations of the Beijing University of Traditional Chinese Medicine’s Committee for Animal Care and Use of Laboratory Animals (approval number BUCM-4-2021083003-3086).

After a 7-day acclimatization period, the mice were randomly and equally divided into six groups (n = 10): control, model, sertraline, medium-dose ECS (ECS-M), high-dose ECS (ECS-H), and low-dose ECS (ECS-L) groups. CORT was dissolved in normal saline with 0.1% Tween-80% and 0.1% DMSO to create a 2 g/L solution which was then used to prepare the CORT injection. This suspension was evenly mixed using ultrasonic waves. During the experiment, each mouse that was not part of the control group was administered a daily injection of CORT at a dose of 40 mg/kg/day for 5 weeks. Mice in the control group were administered normal saline injections at the same dose and time.

The ECS freeze-dried powder was completely dissolved in an appropriate amount of DMSO and then diluted with ddH2O to a concentration of 4.7 mg/mL. In the ECS-L, ECS-M, and ECS-H groups, the medications were administered by gavage at doses of 23.5 mg/kg, 47 mg/kg, and 94 mg/kg, respectively. The operators dissolved ECS in ddH2O at a concentration of 1 mg/mL. Medication was administered intragastrically (i.g.) before CORT injections during the first week of treatment. The groups were treated as follows: 1) distilled water (10 mL/kg body weight) was the only treatment given to the control mice; 2) the Model group received distilled water i.g. (10 mL/kg body weight), and depression was induced via the CORT injection protocol; 3) mice in the sertraline group received sertraline i.g. (10.0 mg/kg/day) and CORT injection; 4) mice in the ECS-H group received ECS i.g. (94 mg/kg/day) and CORT injection; 5) mice in the ECS-M group received ECS i.g. (47 mg/kg/day) and CORT injection; and 6) mice in the ECS-L received ECS i.g. (23.5 mg/kg/day) and CORT injection.

Body weight: The mice were weighed weekly using an electronic scale, and their corresponding growth ratio (%) was figured out as below: Relative growth ratio  % = weight   after   modeling − weight   before   modeling weight   before   modeling × 100   %

Sucrose preference test (SPT): Two steps were required to complete the sucrose preference test. The mice were treated with two identical vials of 1% sucrose solution for 24 h during the adaptation period. One sucrose solution was then substituted with water, and after 12 h, the positions of both bottles are switched to avoid position preference, and the animals are fed for an additional 12 h. Then, the operators deprived the mice of water and food, lasting for 24 h. In next step, those animals were exposed to a sucrose solution and a normal water solution for 4 h. The volume of every bottle was recorded before and after this test. The test was primarily used to assess the level of anhedonia in the animals (Bolaños et al., 2008).

The value of the sucrose preference (SP) is able to be figured out as follows: SP  % = sucrose intake  vol   /   sucrose intake  vol + water intake  vol × 100 %

Open field test (OFT): To start with, those mice were placed in the center of an open field machine; additionally, they were allowed to explore freely for 3 min. After mice adapting, the operators recorded the distance traveled by every mouse over next 3 min. The operators cleaned the open field device using 75% ethanol before assessing the following mouse. And the observers was not sure about the allocation of every mouse (Song et al., 2020).

Forced swimming test (FST): A 18 cm × 30 cm hyaline cylinder was full of water 20 cm deep at 25°C ± 1°C, and the total mice were separately placed into the water for 6 min. The cumulative standstill time for each mouse in the last 4 min of the test was recorded. After every test, the water was changed. Those observers was not sure about the team allocation of every mouse (Choi et al., 2017).

Tail suspension test (TST): With sticky tape put roughly 1 cm from the tail tip, every mouse was suspended above the ground, lasting for 6 min. The total standstill period for the final 4 min of the experiment was assessed after the first 2 min of conditioning. When mice stop scurrying and just move to breathe, they are regarded as static. The groupings of mice were hidden from the observers (Gu et al., 2021).

The experimental procedure implemented for the CORT model and treatment is shown in Figure 2.

After fixation in a 4% paraformaldehyde solution for 24 h, continuous coronal sections with a thickness of 5 µm were subjected to alcohol gradient dehydration and paraffin embedding. To retrieve the antigen, the sections were immersed in a boiling citric acid repair solution and then rinsed thrice using PBS (pH 7.4) for 5 min each. The sections were then incubated overnight at 4°C with a primary antibody against Iba-1 (1:500, Abcam, MA, USA). On the following day, these sections were rinsed with PBS and then treated with drops of an FITC-conjugated secondary antibody (1:3,000, Abcam, MA, USA), followed by incubation for 50 min at room temperature and away from light. After restaining with DAPI (1:100, Solarbio, Beijing, China) for 10 min and cleaning with PBS, an anti-fluorescence quenching sealing agent was applied in the dark. The entire slide was scanned using a fluorescence microscope (Leica, DMI8), and the images were using Caseviewer 2.4 software (Liu et al., 2021; Fang et al., 2022).

Tissues from the mouse hippocampus were homogenized using RIPA lysate (Ebifan Biotechnology Corporation, Beijing, China). The supernatant was obtained through centrifugation at 12,000 rpm for 10 min at 4°C. Commercial ELISA kits were used to measure levels of IL-4, IL-10, IL-1β, and TNF-α, following the manufacturer’s instructions. A GloMax microplate reader (Promega, Madison, WI, USA) was used for assessing the absorbance at 450 nm, and the resulting optical density values were used to calculate the expression levels of IL-4, IL-10, IL-1β, and TNF-α. To reduce inter-assay variations, all samples were assessed thrice using the same test.

Whole proteins in the hippocampus were extracted using a RIPA lysis buffer with 1× protease inhibitor cocktail. A BCA assay test kit was used to assess the protein concentration, following the manufacturer’s instructions. In brief, the proteins were separated using a 10% SDS-PAGE gel and then transferred to a 0.45-µm PVDF membrane (Millipore, Billerica, MA, USA). These membranes were incubated at 23°C for 1 h with 5% non-fat dried milk, followed by incubation with primary antibodies against NOS2 (1: 500, Abcam, MA, USA), Arg (1: 5,000, Proteintech, Wuhan, China), alpha tubulin (1: 5,000, Abcam, MA, USA), NF-κB P65 (1:1000, Cell Signaling Technology, USA), p-NF-κB P65 (1:1000, Cell Signaling Technology, USA), NLRP3 (1:1000, Cell Signaling Technology, USA), caspase-1 (1:500, Santa Cruz, USA), and GAPDH (1:2000, LABLEAD, Beijing, China) overnight at 4°C. Following three TBST washes, the membranes were incubated at 25°C for 1 h with the appropriate goat anti-rabbit IgG-HRP secondary antibody or goat anti-mouse IgG-HRP secondary antibody (IgG-HRP) (both 1: 10000, Proteintech, Wuhan, China). Following another TBST wash, the ECL reagent was used to detect immunoreactivity. The membranes were scanned using ChemiDoc Imagers (BioRad, Hercules, CA, USA). Simultaneously, the intensity of the bands was analyzed using ImageJ software (National Institutes of Health, Bethesda, MD, USA) (Cao et al., 2022).

The 2D structures of these ECS metabolites were obtained from PubChem. After eliminating duplicates, 1,059 targets were obtained. We systematically obtained 66 neuroinflammatory proteins related to depression via text mining using the Pubmed2 Ensembl database, which links more than 2 million studies to approximately 150,000 genes (Baran et al., 2011). Supplementary Table S1 provided a list of these 66 proteins along with their corresponding gene symbols and Ensembl genes. To identify the expected targets, Venn diagrams were used to map ECS-related targets to neuroinflammatory depression-linked targets (Figure 4A).

As shown in the active target network diagram (Figure 4B), the top four metabolites in the network based on degree value are Luteolin, Dehydrodiisoeugenol, Apigenin, and Naringenin. To identify genes related to neuroinflammation-associated depression, we conducted a PPI network analysis using the STRING database. The resulting PPI network had 13 nodes and 25 edges (Figure 4C). To visualize the relationships within the PPI network, the data were fed into Cytoscape 3.7.1, and degree values were computed using the CytoHubba tool. The analysis revealed that CXCL8, ICAM1, NOS2, SELP, TNF, IL6, APP, ACHE, MAOA, ADA were the 10 targets with the highest degree values. Hence, these targets were considered critical in the ECS-mediated treatment of neuroinflammation-associated depression (Figure 4D).

Using the Cluster Profiler program, we conducted KEGG and GO pathway enrichment for the major targets to determine the molecular systems affected by ECS in neuroinflammation-associated depression treatment. GO enrichment categories included BP, CC, and MF. For each dataset, we selected the top six terms (Figure 5A). For BP, regulation of insulin secretion (GO:0050796), calcium-mediated signaling (GO:0019722), defense response to Gram-negative bacterium (GO:0050829), positive regulation of interleukin-6 production (GO:0032755), embryonic digestive tract development (GO:0048566), inflammatory response (GO:0006954).

For CC, extracellular space (GO:0005615), cell surface (GO:0009986), extracellular region (GO:0005576), external side of plasma membrane (GO:0009897), membrane raft (GO:0045121), integral component of plasma membrane (GO:0005887).

The top MF terms included heparin binding (GO:0008201), protein homodimerization activity (GO:0042803), identical protein binding (GO:0042802), flavin adenine dinucleotide binding (GO:0050660), protein self-association (GO:0043621), protein binding (GO:0005515).

In addition to the IL-17 signaling pathway (has04060) and TNF signaling pathwahashsa (04668), the nuclear factor kappa B (NF-κB) signaling pathway (hsa04064) and T-cell receptor signaling pathway (has04660) were selected from a total of 37 KEGG pathways derived from the top 10 most significantly enriched KEGG pathways. The correlation between the pathways and hub targets was then revealed by constructing a target gene–pathways network (Figure 5B). The top 10 significant pathways were categorized into three functional modules related to depression, T-cell activation, and cellular apoptosis. Detailed information on the pathway analysis results is provided in Supplementary Table S2. However, further research is necessary to clarify the exact mechanisms through which ECS exerts its antidepressant effect.