Vagus nerve–dependent antidepressant effects of ineupatorolide B via gut microbiota modulation in a mouse model of depression Authors: Siming Wang, Yuqian Zhang, Ningning Wu, Siqi Wang, Yuhe Ma, Weichen Zhang, Guiying Li, Kenji Hashimoto, Zhiyu Ni Published: 2026 ABSTRACT Ineupatorolide B (InB), a sesquiterpene lactone with anti-inflammatory and anti-tumor properties, has not previously been evaluated for antidepressant effects. Given the central role of the gut–brain axis in stress-related mood disorders, we investigated whether InB alleviates depression-like phenotypes in mice exposed to chronic social defeat stress (CSDS) through microbiota- and vagus-dependent mechanisms. CSDS-susceptible mice received vehicle, fluoxetine, or InB and were assessed using behavioral paradigms, plasma inflammatory cytokine measurements, and analyses of synaptic-related protein expression in the medial prefrontal cortex (mPFC). CSDS induced robust depression-like behaviors, elevated interleukin-6 and tumor necrosis factor-α, caused splenomegaly, and reduced PSD-95 and BDNF expression in the mPFC. Repeated InB administration significantly reversed these behavioral, inflammatory, and synaptic abnormalities. 16S rRNA gene sequencing revealed that CSDS disrupted gut microbial α- and β-diversity, whereas InB treatment largely normalized these alterations and restored depression-associated short-chain fatty acids, including isobutyric and valeric acid, consistent with reestablishment of microbial metabolic homeostasis. Importantly, subdiaphragmatic vagotomy abolished the antidepressant-like, anti-inflammatory, and synaptic effects of InB, demonstrating that its efficacy requires intact vagus nerve–mediated gut–brain signaling. Together, these findings identify vagus-dependent modulation of the microbiota–immune–synaptic axis as a mechanistic pathway through which InB exerts antidepressant effects and highlight this compound as a potential microbiota-targeted therapeutic strategy for depression. Introduction Depression is a highly prevalent and disabling mood disorder characterized by persistent low mood and loss of interest or pleasure. Its global burden continues to rise, influenced by socioeconomic stressors and recent large-scale disruptions such as the COVID-19 pandemic. Mental health disorders, particularly depression, account for a substantial proportion of suicide deaths worldwide, contributing to approximately 70–80% of suicides in high-income countries and around 50% in low- and middle-income countries [1]. First-line pharmacological treatments include selective serotonin reuptake inhibitors (SSRIs) and serotonin–noradrenaline reuptake inhibitors (SNRIs). Although effective for some patients, these agents are limited by delayed therapeutic onset, incomplete response rates, prolonged treatment duration, and adverse effects such as gastrointestinal disturbances and sexual dysfunction, all of which may reduce adherence and overall quality of life [2, 3]. These limitations underscore the need for mechanistically novel and better-tolerated therapeutic strategies. In this context, plant-derived bioactive compounds have attracted increasing interest because of their potential anti-inflammatory, neuromodulatory, and microbiota-regulating properties [4]. Inula cappa (Buchanan-Hamilton ex D. Don) Pruski & Anderberg, also known as Duhaldea cappa [5], is a medicinal shrub widely used in traditional medicine. One of its principal bioactive constituents, ineupatorolide B (InB), is a sesquiterpene lactone with well-documented anti-inflammatory activity mediated by suppression of pro-inflammatory cytokine production at cellular and tissue levels [5]. Despite these pharmacological properties, the potential antidepressant effects and neurobiological mechanisms of InB have not been examined in preclinical models. Accumulating evidence implicates the vagus nerve–dependent gut–brain axis as a critical regulator of stress-related mood disorders. Alterations in gut microbial composition can promote systemic inflammation, disrupt neurotransmitter signaling, and contribute to depressive phenotypes [6–10]. The vagus nerve provides a major bidirectional communication pathway linking intestinal microbial activity with central neural circuits that regulate mood and stress responsiveness [11]. Restoration of gut microbial homeostasis or vagal signaling has been shown to alleviate depression-like behaviors in both preclinical and clinical contexts [12]. However, whether the anti-inflammatory properties of InB engage microbiota-dependent and vagus-mediated mechanisms remains unknown. Here, we investigated the antidepressant potential of InB in a chronic social defeat stress (CSDS) mouse model. We first examined whether repeated InB administration alleviates depression-like behaviors and systemic inflammation. We next assessed its impact on gut microbiota composition and microbial metabolites using 16S rRNA sequencing and short-chain fatty acid profiling. Finally, we assessed whether subdiaphragmatic vagotomy (SDV) would abolish the effects of InB, given that SDV has previously been shown to block depression-like behaviors induced by the oral administration of depression-associated microbes [8, 10] or by fecal microbiota transplantation from mice exhibiting depression-like phenotypes [13]. Materials and methods Animals Thirty-six male adult C57BL/6 mice (10 weeks old, 21–33 g) and twenty CD1 mice (17 weeks old, >37 g) were purchased from Beijing West River Laboratory Animal Technology Co., Ltd. (Beijing, China). The mice were kept in a controlled environment with a 12:12-hour light/dark cycle, where the lights were set to turn on at 20:00. The temperature was maintained at 26 ± 1 °C, and the humidity level was kept at 60%. They had free access to food and water. Ethical approval All animal experiments were conducted in compliance with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals and approved by the Animal Care and Use Committee of Hebei University (approval number: IACUC-HBU2024M062), following the regulations established by the Hebei University Animal Centre. Every effort was taken to minimize discomfort to the animals. Drugs and drug administration The SSRI fluoxetine hydrochloride (Flu; 10 mg/kg, Cayman 14418, USA) was diluted in sterile physiological saline as a positive control [13, 14]. InB was isolated and extracted at the College of Pharmacy, Hebei Medical University, and then mixed with 0.5% (w/v) sodium carboxymethyl cellulose (CMC) to prepare 10 mg/kg. Mice in the negative control group were administered with 0.5% CMC (10 mL/kg). These drugs were given through oral gavage (o.g.). Experimental design Experiment A: In the chronic social defeat stress (CSDS) model, C57BL/6 mice were subjected to various CD1 aggressor mice for 10 min daily, for 10 days (days 1–10), as detailed previously [8–10, 15]. After 24 h of CSDS, the resident CD1 mice and the intruder C57BL/6 mice were placed in cages opposite each other, separated by a perforated organic glass partition allowing visual, olfactory, and auditory interactions, followed by a 24-hour isolation. Starting on day 11, mice received daily oral administrations of Flu, InB, or CMC at 8:00 a.m. for 9 consecutive days (days 11–19). The mice were assigned into four groups: (i) Control (no CSDS, n = 9); (ii) Negative (CSDS + 0.5% CMC [10 mL/kg], n = 9); (iii) Flu (CSDS + Flu [10 mg/kg], n = 9); and (iv) InB (CSDS + InB [10 mg/kg], n = 9). The dose of 10 mg/kg for both Flu and InB was selected based on previous studies [16–19]. They were subjected to behavioral tests as follows: the social interaction test (SIT) on day 11, the locomotion test (LMT) and tail suspension test (TST) on day 17, the forced swimming test (FST) on day 18, and the 1% sucrose preference test (SPT) on day 19. Fecal samples were collected on day 19. Heart blood samples were obtained from anesthetized mice and centrifuged at 3000 × g for 3 min; the resulting plasma was stored at −80 °C. Medial prefrontal cortex (mPFC) samples were also collected and kept at −80 °C. Experiment B: Sham surgery (Sham) or SDV was performed under 5% isoflurane anesthesia. The abdominal hair was removed, and the esophagus was exposed under a microscope. The dorsal and ventral vagus nerves of the esophagus were severed. The muscles and skin were sutured after the procedure, and the mice were placed in a clean cage until they recovered from anesthesia. Mice were then maintained in their cages for 14 days (days 1–14) [8, 10]. Starting on day 14, C57BL/6 mice were subjected to CSDS, similar to Experiment A, by being exposed daily to a different CD1 aggressor mouse for 10 min over a 10-day period (days 14–23). After the stress protocol ended on day 23, mice were housed individually for 24 h. Starting on day 24, the mice received daily oral administrations of Flu, InB, or CMC at 8:00 a.m. for 9 consecutive days (days 24–32). The mice were classified into four groups: (i) Sham group (n = 9): Sham surgery, no CSDS; (ii) Vehicle group (n = 9): SDV + CSDS + 0.5% CMC (10 mL/kg); (iii) Flu group (n = 9): SDV + CSDS + Flu (10 mg/kg); and (iv) InB group (n = 9): SDV + CSDS + InB (10 mg/kg). Five behavioral tests were conducted: the SIT on day 24, the LMT and TST on day 30, the FST on day 31, and the SPT on day 32. On day 33, plasma samples were obtained by centrifuging blood at 3000 × g for 3 min and then stored at −80 °C. mPFC samples were also obtained and kept at −80 °C. Behavioral tests Social interaction test (SIT) After the final social defeat stress event, the SIT was performed at 8:00 a.m. the following day, after which they were placed in an interaction chamber (42 cm³) that contained a smaller cage (11 × 3.5 × 20 cm) on one side. Each mouse was first observed for 2.5 min without an aggressor present, and then another 2.5 min after an unfamiliar aggressor mouse was placed into the smaller cage. The duration of the mouse’s interactions before and after the aggressor’s introduction was recorded. The SIT was used to identify CSDS-susceptible mice [15]. The SIT data are shown in Supplemental Figure S1. Locomotion test (LMT) Mice were placed in a 40 cm² arena with transparent walls and tracked for 60 min using Noldus software (Wageningen, Netherlands) to assess their locomotor activity. The arena was disinfected with 75% ethanol (v/v) between each test [15]. Tail suspension test (TST) Each mouse’s tail was taped approximately 3 cm from its tip and suspended within a black, opaque box (20 × 20 × 35 cm). The duration of immobility was recorded and analyzed using Noldus software [15]. Forced swimming test (FST) The FST is a well-established behavioral assay used to assess depression-like behavior [20]. Each mouse was placed individually into a transparent cylindrical container (23 cm high × 10 cm in diameter) filled with warm water (23 ± 1 °C) to a depth of 15 cm. Immobility was recorded using Noldus software over 6 min. Mice initially swam vigorously but tended to become more immobile over time. Sucrose preference test (SPT) Mice were trained with a 1% sucrose solution for 48 h before testing. Each mouse was housed individually, and the position of the water bottle was switched every 12 h to prevent any location-based bias in sucrose preference. Mice were deprived of food and water 4 h before the experiment. Next, two bottles were placed in the cage: one filled with a 1% sucrose solution and the other with water. The mice were given free access to both bottles for 1 h, after which the bottles were reweighed. The sucrose preference index was calculated as (sucrose consumption/total fluid consumption) × 100%, providing a measure of each mouse’s preference for sucrose [21]. Molecular docking The relevant targets of InB were screened by TCMSP database (Traditional Chinese Medicine Systems Pharmacology Database) and Swiss Target Prediction database. Using “Depression” as the keyword, from GeneCards (https://www.genecards.org), OMIM (https://www.omim.org), DisGeNET (https://www.disgenet.org) and DrugBank (https://go.drugbank.com) were found to be Depression as the target, and the corresponding target was obtained through integration. The gene target obtained by InB was deduplicated and the depression target was obtained by Venny (2.1.0). Using STRING (https://cn.String-db.org), the intersection gene targets of InB and depression were analyzed online, and the “Homo sapiens” species were selected to obtain their PPI maps, which were optimized by Cytoscape (3.10.3) software. The protein sequence was derived from the UniProt database. Since no experimentally resolved structure was available, the protein’s three-dimensional (3D) structure was predicted using AlphaFold2 and then energy-optimized with Rosetta Relax. Water molecules were removed and hydrogens were added to the resulting model using the PyMOL Molecular Graphics System (PyMOL). The compound structure was retrieved from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/). Docking simulations were carried out on the CB-DOCK2 online server (https://cadd.labshare.cn/cb-dock2/php/blinddock.php) in a blind docking approach [22]. CB-DOCK2 uses artificial neural networks to detect binding pockets and employs AutoDock Vina for docking calculations. The docked complexes were evaluated based on binding energy, followed by interaction analysis and docking site assessment to verify binding stability for subsequent dynamics studies. Interactions between the ligand and receptor were analyzed with the Python Protein-Ligand Interaction Profiler (PLIP) online server (https://plip-tool.biotec.tu-dresden.de/plip-web), and the 3D conformations of ligand-receptor complexes were visualized with PyMOL. Measurement of IL-6 and TNF-α levels The plasma levels of interleukin-6 (IL-6) (RK00008; ab clone, USA) and tumor necrosis factor (TNF)-α (RK00027; ab clone, USA) were measured using their corresponding ELISA kits. Western blot analysis Protein was extracted from mPFC tissue using a commercial kit (SD-001; Invent Biotechnologies Inc., USA), and its concentration was determined using a bicinchoninic acid assay (BCA) kit (PC0020; Solarbio). An equal amount of protein (35 μg per sample) was loaded onto a 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gel and then transferred onto a polyvinylidene fluoride (PVDF) membrane (Millipore, USA). After blocking, membranes were incubated at 4 °C overnight with primary antibodies against glutamate receptor 1 (GluA1) (SD2010, 1:500; HUABIO, China), brain-derived neurotrophic factor (BDNF) (SJ12-09, 1:500; HUABIO, China), PSD-95 (SR38-09, 1:500; HUABIO, China), and β-actin (AC026, 1:10,000; ABclonal, USA). Subsequently, membranes were incubated with a goat anti-rabbit DyLight 800 secondary antibody (S9002, 1:1000; Report) at room temperature for 1 h. Immunoreactive bands were visualized using a fluorescence imaging system and analyzed with Image J 1.0 (https://imagej.nih.gov/ij/). Cell culture and processing BV2 cells are an immortalized cell line derived from mouse microglia. Mouse BV2 cells (CL-0493, Wuhan Procell Life Technology Co., Ltd., Wuhan, China) were cultured in Minimum Essential Medium (MEM) containing 10% fetal bovine serum (FBS), streptomycin (100 μg/mL), and penicillin (50 units/mL) at 37 °C in a 5% CO₂ incubator. Cells were seeded into a 96-well plate at a density of 5 × 10⁷ cells/well and cultured until they reached approximately 70–80% confluence. After a 1-hour pretreatment with InB (2 and 4 µM), cells were exposed to a culture medium enriched with hydrogen peroxide (H₂O₂, 200 µM) and corticosterone (200 µM) for an additional 24 h. Flow cytometry Cell apoptosis was evaluated with the Annexin V-fluorescein isothiocyanate/propidium iodide (FITC/PI) staining kit (Sigma-Aldrich, No. 46950-50MG-F). BV2 cells were plated in a 6-well plate and treated according to the established protocol. The treated cells were washed twice with ice-cold phosphate-buffered saline (PBS) and reconstituted in 1× binding buffer. The cells were then incubated with FITC Annexin V and PI for 15 min at 25 °C in the dark. Following staining, the cells were analyzed using flow cytometry (BD FACSCalibur). 16S ribosomal RNA (rRNA) analysis Fecal samples were processed by Shanghai Personal Biotechnology Co., Ltd. (Shanghai, China) using the Illumina NovaSeq platform. The V3-V4 region of the bacterial 16S rRNA gene was amplified via PCR with the forward primer 338F (5’-ACTCCTACGGGAGGCAGCA-3’) and the reverse primer 806R (5’-GGACTACHVGGGTWTCTAAT-3’). After quantification, the microbiome data were analyzed using QIIME2 2019.4 (https://docs.qiime2.org/2019.4/tutorials/) [23], with slight modifications to the official tutorial procedures. Measurement of short‑chain fatty acids (SCFAs) Fecal samples were collected as previously reported [8–10, 15, 24] and stored in sterile microtubes at −80 °C. SCFAs were measured by Shanghai Personal Biotechnology Co., Ltd. (Shanghai, China). SCFA analysis was performed using gas chromatography-tandem mass spectrometry (GC-MS/MS) with an Agilent 7890B gas chromatograph equipped with a DB-FFAP column (30 m × 0.25 mm × 0.25 μm, J&W Scientific, USA). Statistical analysis The data were analyzed utilizing SPSS Statistics 20 (SPSS, Tokyo, Japan) and expressed as mean ± standard error of the mean (SEM). Fecal sample data were analyzed with QIIME2 (2019.4) and R scripts. Weight data were evaluated using a two-way analysis of variance (ANOVA), followed by Fisher’s least significant difference (LSD) test for multiple comparisons. For comparing the four groups, a one-way ANOVA with Fisher’s LSD test or the Kruskal-Wallis test were employed. The α-diversity of the gut microbiota and the relative abundance of bacteria at various taxonomic levels were examined using Kruskal-Wallis and Mann-Whitney U tests. Gut microbiota abundance and distribution were analyzed with algorithms or linear discriminant analysis of effect size (LEfSe) and principal coordinate analysis (PCoA). The correlation between gut microbiota and SCFAs was assessed using Spearman’s correlation analysis [25]. P < 0.05 was considered statistically significant. Results Effects of InB on the depression-like phenotype The experimental procedure for Experiment A is illustrated in Fig. 1A. CSDS modeling and Flu or InB administration did not significantly affect body weight (Fig. 1B). However, CSDS did not impair the motor function of the mice (Fig. 1C). Compared with the control group, CSDS prolonged immobility times in the TST and FST, effects that were not observed in the Flu- or InB-treated groups (Fig. 1D–E). In the SPT, CSDS significantly reduced the sucrose preference relative to the InB group (Fig. 1F). Taken together, these results suggest that both InB and Flu exhibit antidepressant-like effects in the CSDS model. CSDS exhibited significantly high plasma levels of TNF-α and IL-6, while InB and Flu treatments substantially reduced the levels of these cytokines, with InB showing a more pronounced effect than Flu (Fig. 1G–H). In addition, CSDS caused splenomegaly [19, 26, 27], which was not observed in the InB group (Fig. 1I–J), likely due to the anti-inflammatory properties of InB [10]. Furthermore, Flu and InB treatments significantly upregulated GluA1, PSD-95, and BDNF expression in the mPFC of CSDS mice (Fig. 1K–M). Collectively, these results suggest that InB may alleviate the downregulation of PSD-95, GluA1, and BDNF in the mPFC of CSDS mice by controlling systemic inflammation, thereby exerting its antidepressant effects.Fig. 1Effects of InB administration on depression-like behaviors in the CSDS model.A: Timeline of Experiment 1. Mice were exposed to chronic social defeat stress (CSDS) for 10 days. The social interaction test (SIT) was performed to identify CSDS-susceptible mice for subsequent drug administration and behavioral assessments [locomotion test (LMT); forced swimming test (FST); tail suspension test (TST); sucrose preference test (SPT)], and sample collection. B: Graph comparing the body weight of mice on days 0 and 10 (time: F3,32 = 8.411, P = 0.007; group: F3,32 = 3.066, P = 0.042; interaction: F3,32 = 6.163, P = 0.002). C: Distance moved during the LMT test (F3,32 = 1.451, P = 0.2462). D and E: Graph showing the immobility time during TST (F3,32 = 6.655 P = 0.0013) (D) and FST (F3,32 = 2.912 P = 0.0494) (E). F: Mice preferred sucrose, as evidenced by a higher intake of sucrose water than the amount of the total consumption (F3,32 = 4.867, P = 0.0067). G and H: Graphs showing the plasma level of TNF-α (F3,32 = 18.25, P < 0.001) (G) and IL-6 (F3,32 = 65.74, P < 0.001) (H) in each treatment group. I: Graph showing the weight change of the spleen in mice (F3,32 = 21.69, P < 0.0001). J: Representative photograph of the spleen in the four groups. K-M: Western blots and summarized graphs showing expression of GluA1 (F3,32 = 3.128, P = 0.0392) (K), PSD-95 (F3,32 = 2.928 P = 0.0486) (L), and BDNF (F3,32 = 3.015, P = 0.0443) (M). One-way analysis of variance (ANOVA) was used for all analyses with interaction scores and results are presented as mean ± S.E.M. *P < 0.05, **P < 0.01, ***P < 0.001. ns.: not significant. Effects of InB administration on depression-like behaviors in the CSDS model. A: Timeline of Experiment 1. Mice were exposed to chronic social defeat stress (CSDS) for 10 days. The social interaction test (SIT) was performed to identify CSDS-susceptible mice for subsequent drug administration and behavioral assessments [locomotion test (LMT); forced swimming test (FST); tail suspension test (TST); sucrose preference test (SPT)], and sample collection. B: Graph comparing the body weight of mice on days 0 and 10 (time: F3,32 = 8.411, P = 0.007; group: F3,32 = 3.066, P = 0.042; interaction: F3,32 = 6.163, P = 0.002). C: Distance moved during the LMT test (F3,32 = 1.451, P = 0.2462). D and E: Graph showing the immobility time during TST (F3,32 = 6.655 P = 0.0013) (D) and FST (F3,32 = 2.912 P = 0.0494) (E). F: Mice preferred sucrose, as evidenced by a higher intake of sucrose water than the amount of the total consumption (F3,32 = 4.867, P = 0.0067). G and H: Graphs showing the plasma level of TNF-α (F3,32 = 18.25, P < 0.001) (G) and IL-6 (F3,32 = 65.74, P < 0.001) (H) in each treatment group. I: Graph showing the weight change of the spleen in mice (F3,32 = 21.69, P < 0.0001). J: Representative photograph of the spleen in the four groups. K-M: Western blots and summarized graphs showing expression of GluA1 (F3,32 = 3.128, P = 0.0392) (K), PSD-95 (F3,32 = 2.928 P = 0.0486) (L), and BDNF (F3,32 = 3.015, P = 0.0443) (M). One-way analysis of variance (ANOVA) was used for all analyses with interaction scores and results are presented as mean ± S.E.M. *P < 0.05, **P < 0.01, ***P < 0.001. ns.: not significant. Network pharmacology explores the interaction between InB and depression The chemical structure of InB is shown in Fig. 2A. Several hydrogen bonds (red) and salt bridges (yellow) were observed between the protein and InB, involving various hydrogen bond donors and acceptors. The interacting residues are indicated as shown in Fig. 2A, with bond lengths indicated next to the dashed lines (in angstroms). The interaction targets of InB and depression were obtained by Venny2.1.0 after deduplication, and a total of 34 intersection gene targets were obtained. (Fig. 2B), 34 gene targets intersecting InB and depression were analyzed online by String to construct a PPI network diagram of potential targets of InB for the treatment of depression (Fig. 2C). The PPI network diagram (Fig. 2D) was obtained by Cytoscape (3.10.3) software. The potential intersection targets of InB and depression were imported into the Metascape database, and the enrichment analysis of GO (Gene Ontology) function was performed to create a histogram (Fig. 2E). A total of 127 signaling pathways were obtained through KEGG (Kyoto Encyclopedia of Genes and Genomes) pathway analysis, and the top 20 signaling pathways were visualized and analyzed by bubble diagram (Fig. 2F). The results showed that InB-related targets were mainly involved in the synaptic transmission process, γ-aminobutyric acid signaling pathway, and the regulation of inflammation, and the main pathways involved were neuroactive ligand receptor interactions for the treatment of depression. Overall spatial analysis suggests that the ligand fits snugly within the protein’s binding cavity, exhibiting low free energy and multiple non-covalent interactions. These findings indicate a stable ligand-receptor complex, suitable for subsequent molecular dynamics simulations to further evaluate binding stability (Fig. 2G–I).Fig. 2Network pharmacology analysis of InB.The chemical structure of InB is shown in (A). Putative targets of InB were predicted using the TCMSP (Traditional Chinese Medicine Systems Pharmacology) database and the SwissTargetPrediction database. Depression-related genes were retrieved using the keyword “depression” from GeneCards (https://www.genecards.org), OMIM (https://www.omim.org), DisGeNET (https://www.disgenet.org), and DrugBank (https://go.drugbank.com), and then integrated to obtain a non-redundant list of depression-associated targets (B). The overlapping targets between InB and depression were identified using Venny 2.1.0 (C). The intersection targets were further analyzed using STRING (https://cn.string-db.org) with “Homo sapiens” selected as the species to construct a protein–protein interaction (PPI) network, which was optimized and visualized in Cytoscape 3.10.3 (D). The potential intersection targets of InB and depression were imported into the Metascape database for GO function enrichment analysis, and 132 biological processes (BPs) were obtained. 38 cellular components (CC); 58 molecular functions (MF). Through the microbial platform, the first 10 data of BP, CC, and MF were imported to create a histogram (E). The results showed that the related targets were mainly involved in the synaptic transmission process, γ-aminobutyric acid signaling pathway, and the regulation of inflammatory processes in BP, and the CC analysis showed that it mainly involved postsynaptic, GABA-A receptor, neuronal cells, etc., and MF analysis showed that InB may control the GABA-gated chloride ion channel and extracellular ligand-gated monoatomic ion channel and other activities for the treatment of depression. A total of 127 signaling pathways were obtained through KEGG pathway analysis, and the top 20 signaling pathways were visualized and analyzed by bubble diagram (F). Docking analysis revealed specific interactions between the compound and proteins, with binding energies predicted by AutoDock Vina using an AMBER-based force field. The binding energy between the compound and BDNF, GluA1, and PSD95 proteins was −7.7, −7.6, and −8.4 kcal/mol, respectively. Interaction analysis of the protein-ligand complexes was conducted using the PLIP server, with results visualized in PyMOL(G-I). Network pharmacology analysis of InB. The chemical structure of InB is shown in (A). Putative targets of InB were predicted using the TCMSP (Traditional Chinese Medicine Systems Pharmacology) database and the SwissTargetPrediction database. Depression-related genes were retrieved using the keyword “depression” from GeneCards (https://www.genecards.org), OMIM (https://www.omim.org), DisGeNET (https://www.disgenet.org), and DrugBank (https://go.drugbank.com), and then integrated to obtain a non-redundant list of depression-associated targets (B). The overlapping targets between InB and depression were identified using Venny 2.1.0 (C). The intersection targets were further analyzed using STRING (https://cn.string-db.org) with “Homo sapiens” selected as the species to construct a protein–protein interaction (PPI) network, which was optimized and visualized in Cytoscape 3.10.3 (D). The potential intersection targets of InB and depression were imported into the Metascape database for GO function enrichment analysis, and 132 biological processes (BPs) were obtained. 38 cellular components (CC); 58 molecular functions (MF). Through the microbial platform, the first 10 data of BP, CC, and MF were imported to create a histogram (E). The results showed that the related targets were mainly involved in the synaptic transmission process, γ-aminobutyric acid signaling pathway, and the regulation of inflammatory processes in BP, and the CC analysis showed that it mainly involved postsynaptic, GABA-A receptor, neuronal cells, etc., and MF analysis showed that InB may control the GABA-gated chloride ion channel and extracellular ligand-gated monoatomic ion channel and other activities for the treatment of depression. A total of 127 signaling pathways were obtained through KEGG pathway analysis, and the top 20 signaling pathways were visualized and analyzed by bubble diagram (F). Docking analysis revealed specific interactions between the compound and proteins, with binding energies predicted by AutoDock Vina using an AMBER-based force field. The binding energy between the compound and BDNF, GluA1, and PSD95 proteins was −7.7, −7.6, and −8.4 kcal/mol, respectively. Interaction analysis of the protein-ligand complexes was conducted using the PLIP server, with results visualized in PyMOL(G-I). Protective effects of InB in BV2 cells BV2 cells are an immortalized cell line derived from mouse microglia, and most BV2 cells shrank and died in the presence of corticosterone/H₂O₂. However, these detrimental effects were not observed after InB (2 µM) treatment (Fig. 3A and E). The Annexin V/PI and Cell Counting Kit-8 (CCK-8) assays showed that corticosterone/H₂O₂ treatment significantly increased the cell apoptosis rate relative to the control group. Notably, InB (2 µM) treatment prevented these cytotoxic effects, protecting BV2 cells from H₂O₂- and corticosterone-induced damage (Fig. 3B–D and F–H).Fig. 3Effects of InB on corticosterone and hydrogen peroxide (H2O2)-induced distributive changes of early and late apoptotic cells.A and E: Morphological observation. Cells after treatment were observed under a phase contrast microscope (×200). Cells were treated with corticosterone (200 µM) and H2O2 (200 µM) for 24 h or with InB with the indicated concentrations (2 and 4 µM). B and F: BV2 cells were exposed to 200 µM of corticosterone and H2O2 and different concentrations of InB for 24 h. After treatment, stained cells were analyzed by flow cytometry using Annexin V-FITC and PI labeling. Apoptotic cells (C, G) and cell necrosis (D, H) were analyzed. Results are presented as mean ± S.E.M. (n = 5). *P < 0.01, compared with control group, #P < 0.05, ##P < 0.01 compared with corticosterone and H2O2-treated group. Effects of InB on corticosterone and hydrogen peroxide (H2O2)-induced distributive changes of early and late apoptotic cells. A and E: Morphological observation. Cells after treatment were observed under a phase contrast microscope (×200). Cells were treated with corticosterone (200 µM) and H2O2 (200 µM) for 24 h or with InB with the indicated concentrations (2 and 4 µM). B and F: BV2 cells were exposed to 200 µM of corticosterone and H2O2 and different concentrations of InB for 24 h. After treatment, stained cells were analyzed by flow cytometry using Annexin V-FITC and PI labeling. Apoptotic cells (C, G) and cell necrosis (D, H) were analyzed. Results are presented as mean ± S.E.M. (n = 5). *P < 0.01, compared with control group, #P < 0.05, ##P < 0.01 compared with corticosterone and H2O2-treated group. The influence of InB on gut microbiota CSDS mice exhibited an imbalance in their gut microbiota, consistent with previous findings [9, 20, 27–29]. α-diversity and β-diversity were analyzed to determine the diversity and composition of the gut microbiota across the four groups. The Kruskal-Wallis test revealed significant differences in Chao 1, observed species, Shannon, Simpson, Faith’s phylogenetic diversity (PD), Pieloue, and Goods coverage indices among the groups (Fig. 4A–G). Post-hoc analysis revealed that α-diversity was significantly reduced in CSDS mice compared with controls, whereas InB treatment markedly ameliorated this reduction. For β-diversity, principal component analysis (PCA) was performed to evaluate community composition (Fig. 4H). The results showed significant differences among the groups, and similarity analysis confirmed these differences (r = 0.4093, P = 0.001).Fig. 4α-diversity and β-diversity of gut microbiota.A-G: Graphs showing the α-diversity analyzed with different indexes: A: Chao1(P = 0.0319); B: Observed_species (P = 0.0482); C: Shannon (P = 0.0271); D: Simpson (P = 0.0436); E: Faith_pd (P = 0.0242); F: Pielou_e (P = 0.6639); G: Goods_coverage (P = 0.0602). Kruskal-Wallis test was used for all analyses with interaction scores and results are presented as mean ± S.E.M. *P < 0.05, **P < 0.01, ***P < 0.001. H: Based on the unweighted UniFrac PCoA graph of the distance (ANOSIM, R = 0.4094, P = 0.001). Each point represents a single sample with 11.4% of the X-axis principal component and 8.1% of the Y-axis principal component. LEfSe can directly analyze the difference of all classification levels at the same time. The algorithm of LEfSe was utilized to analyze the impact of InB treatment on intestinal microbiota. LEfSe analysis results included two parts: the taxonomic branch of species (I) and the histogram of linear discriminant analysis (LDA) value distribution of significantly different species (J), which was used to show the significantly enriched species in each group and their importance. To show the taxonomic hierarchical distribution of the marker species in each group of samples. J: Histograms of the richest groups based on the linear discriminant analysis score cutoff value (log10) > 4.0 and P < 0.05 showed significant differences in abundance between the control group, the negative group, and the InB group. K: Hierarchical clustering tree diagram showing the similarity between different samples. α-diversity and β-diversity of gut microbiota. A-G: Graphs showing the α-diversity analyzed with different indexes: A: Chao1(P = 0.0319); B: Observed_species (P = 0.0482); C: Shannon (P = 0.0271); D: Simpson (P = 0.0436); E: Faith_pd (P = 0.0242); F: Pielou_e (P = 0.6639); G: Goods_coverage (P = 0.0602). Kruskal-Wallis test was used for all analyses with interaction scores and results are presented as mean ± S.E.M. *P < 0.05, **P < 0.01, ***P < 0.001. H: Based on the unweighted UniFrac PCoA graph of the distance (ANOSIM, R = 0.4094, P = 0.001). Each point represents a single sample with 11.4% of the X-axis principal component and 8.1% of the Y-axis principal component. LEfSe can directly analyze the difference of all classification levels at the same time. The algorithm of LEfSe was utilized to analyze the impact of InB treatment on intestinal microbiota. LEfSe analysis results included two parts: the taxonomic branch of species (I) and the histogram of linear discriminant analysis (LDA) value distribution of significantly different species (J), which was used to show the significantly enriched species in each group and their importance. To show the taxonomic hierarchical distribution of the marker species in each group of samples. J: Histograms of the richest groups based on the linear discriminant analysis score cutoff value (log10) > 4.0 and P < 0.05 showed significant differences in abundance between the control group, the negative group, and the InB group. K: Hierarchical clustering tree diagram showing the similarity between different samples. LEfSe analysis revealed marked differences in microbial abundances among the four groups (Fig. 4I). Hierarchical clustering showed that the CSDS and control groups clustered distinctly, a divergence not observed in the InB group (Fig. 4J). Further clustering analyses revealed that after CSDS, the microbial community in the InB group more closely resembled that of the control group (Fig. 4K). Additional examination at the phylum level demonstrated marked differences in gut microbiota composition among the four groups (Fig. S2A). Particularly, CSDS mice exhibited significantly increased levels of Bacteroidetes and significantly reduced levels of TM7, Tenericutes, and Proteobacteria (Fig. S2B–E). Nevertheless, this pattern was absent in the InB group, implying that InB treatment may restore the gut microbiota composition to a state closer to that of the control group. Fecal SCFA contents differed among the four groups. Relative to the control group, CSDS mice exhibited significantly reduced levels of isobutyric and valeric acids, a change not observed in the InB group. Furthermore, propionic acid levels differed significantly between the InB and CSDS groups (Fig. S3A–C). Subsequently, logistic regression and correlation analyses were performed to explore the relationship between gut microbiota and SCFAs. At the phylum level, Bacteroidetes were negatively associated with isobutyric acid (r = −0.4427, P = 0.0069), whereas TM7 exhibited a positive correlation with isobutyric acid (r = 0.3468, P = 0.0382) (Fig. S3D–E). The impact of SDV on antidepressant effects of InB in depression-like behaviors in CSDS-exposed mice The procedures for Experiment B are illustrated in Fig. 5A. SDV and CSDS modeling did not significantly impact body weight (Fig. 5B). However, CSDS did not impair the motor function of the mice (Fig. 5C). SDV notably reversed the therapeutic effects of InB in CSDS-exposed mice. The InB group displayed significantly prolonged immobility times in the TST and FST compared to the control and Flu groups (Fig. 5D, E). In the SPT, sucrose preference in the InB group was markedly reduced (Fig. 5F). Collectively, these results indicate that the vagus nerve mediates the antidepressant effects of InB.Fig. 5Effects of SDV on effects of InB on depression-like behaviors in CSDS-susceptible mice.A: Timeline of Experiment 2. After recovery from SDV and sham surgeries for 14 days, mice were exposed to CSDS for 10 days, followed by drug administration and behavioral testing. B: Graph comparing the body weight of mice on days 14 and 23 (time: F3,32 = 262.830, P < 0.001; group: F3,32 = 0.645 P = 0.592; interaction: F3,32 = 13.583, P < 0.001). C Distance moved during the LMT test (F3,32 = 1.073, P = 0.3742). D and E: Graph showing the immobility time during TST (F3,32 = 9.242, P = 0.0001) (D) and FST (F3,32 = 4.165, P = 0.0134) (E). F: Mice preferred sucrose, as evidenced by a higher intake of sucrose water than the amount of the total consumption (F3,32 = 3.456, P = 0.0278). G and H: Graphs showing the plasma level of TNF-α (F3, 32 = 16.92, P < 0.001) (G) and IL-6 (F3, 32 = 64.5, P < 0.001) (H) in each treatment group. I: Graph showing the weight change of the spleen in mice (F3,32 = 53.13, P < 0.0001). J: Representative photograph of the spleen in the four groups. K-M: Western blots and summarized graphs showing expression levels of GluA1 (F3, 32 = 3.714, P = 0.0212) (K), PSD-95 (F3, 32 = 4.311, P = 0.0116) (L), and BDNF (F3, 32 = 3.684, P = 0.0219) (M). One-way ANOVA was used for all analyses with interaction scores and results are presented as mean ± S.E.M. *P < 0.05, **P < 0.01, ***P < 0.001. Effects of SDV on effects of InB on depression-like behaviors in CSDS-susceptible mice. A: Timeline of Experiment 2. After recovery from SDV and sham surgeries for 14 days, mice were exposed to CSDS for 10 days, followed by drug administration and behavioral testing. B: Graph comparing the body weight of mice on days 14 and 23 (time: F3,32 = 262.830, P < 0.001; group: F3,32 = 0.645 P = 0.592; interaction: F3,32 = 13.583, P < 0.001). C Distance moved during the LMT test (F3,32 = 1.073, P = 0.3742). D and E: Graph showing the immobility time during TST (F3,32 = 9.242, P = 0.0001) (D) and FST (F3,32 = 4.165, P = 0.0134) (E). F: Mice preferred sucrose, as evidenced by a higher intake of sucrose water than the amount of the total consumption (F3,32 = 3.456, P = 0.0278). G and H: Graphs showing the plasma level of TNF-α (F3, 32 = 16.92, P < 0.001) (G) and IL-6 (F3, 32 = 64.5, P < 0.001) (H) in each treatment group. I: Graph showing the weight change of the spleen in mice (F3,32 = 53.13, P < 0.0001). J: Representative photograph of the spleen in the four groups. K-M: Western blots and summarized graphs showing expression levels of GluA1 (F3, 32 = 3.714, P = 0.0212) (K), PSD-95 (F3, 32 = 4.311, P = 0.0116) (L), and BDNF (F3, 32 = 3.684, P = 0.0219) (M). One-way ANOVA was used for all analyses with interaction scores and results are presented as mean ± S.E.M. *P < 0.05, **P < 0.01, ***P < 0.001. Following SDV, blood IL-6 and TNF-α levels were significantly elevated in the InB group compared with the sham group (Fig. 5G, H). Data shown in Fig. 5I, J reveal that mice in the InB group also developed splenomegaly after SDV. Furthermore, after SDV, CSDS downregulated the levels of GluA1, PSD-95, and BDNF in the mPFC of InB-treated mice, an effect not observed in the Flu group (Fig. 5K–M). Altogether, these results highlight that the anti-inflammatory effects of InB and its positive influence on protein expression depend on intact vagus nerve signaling. Discussion The present study reveals several key findings. First, repeated administration of InB significantly alleviated depression-like behaviors in the CSDS mouse model, as indicated by reduced immobility in the TST and FST and restoration of sucrose preference. Second, InB decreased systemic inflammatory cytokines (IL-6 and TNF-α), prevented splenomegaly, and normalized synaptic-related proteins (GluA1, PSD-95, and BDNF) in the mPFC, suggesting coordinated anti-inflammatory and synaptic regulatory effects. Third, InB reversed CSDS-induced disruptions in gut microbiota diversity and composition and restored SCFA levels, particularly isobutyric and valeric acid, supporting modulation of the gut–brain axis. Fourth, subdiaphragmatic vagotomy abolished the antidepressant and anti-inflammatory effects of InB, demonstrating a requirement for vagus nerve–mediated gut–brain signaling. Collectively, these findings position InB as a biologically active compound capable of mitigating stress-induced depressive phenotypes through coordinated regulation of immune, microbial, and neural pathways. Our data support a model in which InB reshapes gut microbial composition, suppresses systemic inflammation, and restores synaptic plasticity markers in the mPFC. Although causal relationships between specific microbial taxa and behavioral outcomes remain unresolved, the convergence of microbiota normalization, cytokine reduction, and synaptic restoration supports an integrated microbiota–immune–synaptic mechanism. Notably, fluoxetine increased synaptic proteins without reducing plasma IL-6 in CSDS mice [15], whereas InB simultaneously reduced inflammatory cytokines and restored synaptic protein expression, suggesting engagement of inflammation–synaptic crosstalk potentially mediated through microbiota-dependent pathways. The antidepressant effects of InB may therefore arise primarily from peripheral gut–vagus–immune signaling rather than direct central pharmacological action. A key limitation is the absence of pharmacokinetic evidence demonstrating blood–brain barrier (BBB) penetration; thus, the behavioral effects cannot currently be attributed to direct central exposure. Instead, normalization of systemic inflammation and microbiota-derived metabolites supports a mechanism whereby InB influences brain function through vagal afferent and neuroimmune communication. Consistent with this interpretation, the protective effects of InB in BV2 microglial cells should be viewed as evidence of intrinsic anti-inflammatory potential, rather than proof of direct microglial modulation in vivo. Peripheral suppression of inflammatory signaling and restoration of microbial metabolites may indirectly regulate central microglial activation and synaptic plasticity without requiring BBB penetration. Future studies integrating brain pharmacokinetics, metabolite tracing, vagal manipulation, and microbiota-targeted approaches will be required to determine whether InB acts exclusively through peripheral signaling or also possesses central bioavailability. Accumulating evidence indicates that patients with depression exhibit reduced SCFA levels—including butyric, acetic, isobutyric, and valeric acid—which may influence neuronal activity through immune and neuroendocrine mechanisms [30]. SCFAs exert anti-inflammatory effects via G-protein-coupled receptors such as GPR41 and GPR43 and can attenuate systemic inflammation [31]. They also stabilize hypothalamic–pituitary–adrenal axis function, cross the BBB, and modulate stress-related neuroendocrine signaling [32]. In the present study, InB restored CSDS-induced reductions in SCFAs, consistent with recovery of microbial metabolic homeostasis and downstream behavioral benefits. Nevertheless, fecal SCFA concentrations represent indirect metabolic indices influenced by absorption, host metabolism, and intestinal transit. Accordingly, these findings should be interpreted as supportive correlates rather than definitive mediators. Future studies measuring circulating and brain SCFA levels, together with receptor-level analyses of GPR41/43 signaling, will be necessary to establish causal links between SCFA regulation and antidepressant effects. Interestingly, Proteobacteria abundance was not normalized by InB despite improvements in behavior, inflammation, synaptic proteins, and SCFAs. Because Proteobacteria represent a minor yet immunologically relevant fraction, this observation suggests that global taxonomic normalization is not required for antidepressant efficacy. Instead, selective modulation of functional microbial pathways—including SCFA production, immune signaling, and vagus-mediated communication—may be sufficient to restore behavioral homeostasis. This interpretation aligns with emerging views that microbial metabolites and neural signaling, rather than overall community structure, are primary determinants of neuropsychiatric outcomes and supports development of precision microbiota-targeted interventions. The vagus nerve serves as a central conduit of gut–brain communication, transmitting inflammatory and metabolic signals from the gastrointestinal tract to the brain [11]. Vagal afferent activity can promote neuroinflammatory signaling, microglial activation, and synaptic dysregulation, thereby contributing to depression-like behaviors [11, 33–35], while stress-induced HPA axis activation further amplifies immune responses [36]. Prior studies demonstrated vagus-dependent behavioral effects of Lactobacillus rhamnosus (JB-1) [37] and depression-like phenotypes induced by microbiota transfer from CSDS-susceptible mice that are attenuated by SDV [8, 10, 12, 38]. Consistent with this framework, SDV abolished the behavioral, inflammatory, and synaptic effects of InB, confirming a requirement for vagus-dependent microbiota–brain signaling. Despite these advances, several limitations remain. The molecular pathways downstream of vagus-dependent microbiota signaling require further clarification, including the roles of specific metabolites and afferent neural circuits. Generalizability beyond the CSDS model remains uncertain, and additional stress paradigms should be examined. Brain-region specificity beyond the mPFC, central inflammatory signaling, and causal microbiota manipulations (e.g., fecal transplantation or gnotobiotic models) also warrant investigation. The absence of treatment-only control groups limits interpretation of drug-specific effects independent of stress, and future pharmacokinetic studies are needed to determine central exposure of InB. Finally, the exclusive use of male mice restricts translational generalizability, highlighting the importance of sex-inclusive experimental designs. To enhance interpretive rigor, the present conclusions prioritize experimentally supported microbiota- and vagus-dependent mechanisms, whereas exploratory analyses such as network pharmacology and in vitro assays are regarded as hypothesis-generating. In conclusion, InB alleviates depression-like behaviors in CSDS-susceptible mice by modulating gut microbiota composition, restoring SCFA metabolism, suppressing systemic inflammation, and enhancing synaptic plasticity in the mPFC. These effects require intact vagus-dependent gut–brain signaling. Together, the findings identify InB as a microbiota-targeted therapeutic candidate and support modulation of microbiota–immune–neural interactions as a translational strategy for depression. Supplementary information Supplemental materials Supplemental materials