GV-971 reduces depression-like symptoms caused by chronic stress by changing the gut-brain system

To establish a robust and objective chronic restraint stress (CRS) model of depression, the restraint parameters (daily duration and total exposure) were systematically optimized. Immobility time in the forced swim test (FST) was used as the primary behavioral endpoint. CRS-induced depression-like behavior showed clear time- and intensity-dependent effects. After 2 weeks of CRS, both the 4 h and 8 h restraint groups exhibited a trend toward increased immobility relative to the naïve control group, but these changes did not reach statistical significance and overt depression-like behavior was not evident (Figure 1B). By 4 weeks, immobility time in the 8 h restraint group was significantly prolonged compared with controls (from 169.95 ± 9.85 s to 199.77 ± 5.86 s, p = 0.021; Figure 1C). When the CRS period was extended to 8 weeks, immobility time in the 4 h group showed an increasing trend, and in the 8 h group, it significantly increased from 184.93 ± 6.32 s to 217.64 ± 5.66 s (p = 0.001), indicating a typical depression-like despair phenotype (Figure 1D). Peripheral neurochemical measurements further supported an intensity-dependent effect of restraint stress. At week 8, mice in the 8 h group displayed marked depletion of plasma monoamine neurotransmitters. Compared with controls, plasma 5-HT levels decreased from 381.92 ± 61.02 ng/mL to 132.02 ± 46.77 ng/mL (p = 0.012; Figure 1E), and NE levels fell from 315.46 ± 26.29 ng/mL to 194.84 ± 7.01 ng/mL (p < 0.001; Figure 1F). In contrast, 4 h restraint did not significantly alter plasma 5-HT or NE, suggesting that 8 h of daily restraint is more effective at inducing peripheral neuroendocrine disruption. Overall, these behavioral and neurochemical data suggest that 4 h/d CRS exerts only a mild impact on the monoamine system, whereas 8 h/d for 4–8 weeks induces both emerging behavioral despair and progressive monoamine dysregulation. Under the 8 h/d regimen, mice developed a stable depression-like phenotype that was already apparent at week 4 and remained sustained through week 8.

To determine whether GV-971 could rescue the behavioral alterations associated with chronic restraint stress (CRS), mice were assessed via the elevated plus maze (EPM), tail suspension test (TST), and forced swim test (FST) during weeks 5–7 of CRS, with fluoxetine serving as the positive control. In the EPM, both GV-971 and fluoxetine significantly increased open-arm distance relative to CRS mice (Figure 2B). Values increased from 245.38 ± 29.10 cm in the CRS group to 332.18 ± 26.49 cm in the GV-971 group (p = 0.035) and 444.09 ± 46.50 cm in the fluoxetine group (p < 0.001), consistent with an anxiolytic-like effect. In the TST, immobility time was markedly reduced by both drugs (Figure 2C), from 39.71 ± 7.81 s in CRS mice to 19.24 ± 3.42 s after GV-971 treatment (p = 0.026) and to 1.56 ± 0.59 s after fluoxetine (p < 0.001). In the FST, GV-971 again significantly shortened immobility time (Figure 2D), from 31.61 ± 8.85 s in CRS mice to 3.85 ± 1.14 s (p = 0.008), whereas fluoxetine produced a numerically higher immobility time (59.71 ± 12.66 s) that did not differ significantly from the CRS group.

In biochemical analyses, hippocampal 5-HT levels tended to be higher in GV-971-treated mice than in CRS controls (Figure 2E), rising from 88.21 ± 15.67 ng/mg to 122.81 ± 7.62 ng/mg (p = 0.052), whereas the fluoxetine group showed 5-HT levels of 98.22 ± 17.11 ng/mg, which were not significantly different from those of the CRS group. Plasma CORT measurements demonstrated that GV-971 significantly lowered CORT concentrations from 80.14 ± 3.14 ng/mL to 63.45 ± 5.71 ng/mL (p = 0.027) (Figure 2F). By contrast, fluoxetine-treated mice displayed CORT levels of 83.91 ± 9.03 ng/mL, with no significant improvement relative to CRS controls. Adrenal histology further supported a normalizing effect of GV-971 on the stress axis. H&E staining revealed evident adrenal cortical hyperplasia and thickening in CRS mice, whereas GV-971-treated animals showed a cortical morphology closer to that of control mice (Figure 2G). Morphometric analysis confirmed a trend towards increased cortical thickness in the CRS group compared with controls (116.13 ± 5.78 μm in the control group vs. 158.14 ± 14.53 μm in the CRS group, p = 0.051). GV-971 significantly reduced cortical thickness to 120.54 ± 2.33 μm (p = 0.044 vs. CRS), suggesting that GV-971 suppresses CRS-induced adrenal cortical hypertrophy and partially restores normal adrenal structure (Figure 2H). Collectively, these data indicate that GV-971 not only rescues CRS-induced behavioral deficits but also uniquely targets the underlying neuroendocrine dysregulation. Unlike fluoxetine, which showed limited efficacy on the HPA axis in this model, GV-971 demonstrated a comprehensive therapeutic profile by concurrently alleviating depression-like behaviors, trending toward restoring central serotonin, and significantly normalizing peripheral stress markers and adrenal morphology.

To investigate the effect of GV-971 on chronic restraint stress (CRS)-induced neuroinflammation, mice were exposed to CRS for 56 days and received GV-971 or fluoxetine (positive control) according to the schedule shown in Figure 3A. At the end of the experiment, brain tissue was collected, and immunofluorescence staining for the microglial marker IBA1 was performed in the cortex, hippocampal CA1 region and dentate gyrus (DG). Morphological analysis (Figure 3B,E,H) showed that in control mice, IBA1-positive microglia (green) were sparse in the cortex and hippocampus (CA1 and DG) and exhibited thin, highly ramified processes, consistent with a resting state. In contrast, CRS markedly increased the number of IBA1-positive microglia in these regions, with stronger fluorescence signals and a characteristic activated morphology, including enlarged cell bodies and retracted processes. Treatment with GV-971 markedly attenuated this abnormal microglial activation, as indicated by a reduction in the number of IBA1-positive cells and a gradual restoration of microglial morphology towards the resting phenotype observed in controls. Quantification of IBA1 fluorescence intensity in the cortex, hippocampal CA1 and DG is shown in Figure 3C,F,I, whereas the percentage of Iba1⁺ cells relative to total nuclei is quantified in Figure 3D,G,J. In the cortex, IBA1 fluorescence intensity was significantly increased in CRS mice compared with control mice, rising from 1,249,710 ± 58,939 a.u. to 1,839,068 ± 119,058 a.u. (p = 0.0047). GV-971 treatment significantly reduced this increase to 1,185,823 ± 82,970 a.u. (p = 0.011 vs. CRS). Concurrently, the percentage of Iba1⁺ cells was significantly increased in CRS mice compared with control mice, rising from 3.37 ± 0.25% to 4.85 ± 0.44% (p = 0.047). Both fluoxetine and GV-971 treatments reduced this increase: fluoxetine lowered it to 3.52 ± 0.36% (p = 0.099 vs. CRS), and GV-971 reduced it to 3.19 ± 0.35% (p = 0.037 vs. CRS), with no significant differences observed between the treatment group and the control group. No significant differences in IBA1 fluorescence intensity were observed in the fluoxetine group compared with controls.

In the hippocampal CA1 region, IBA1 fluorescence intensity similarly increased from 1,266,011 ± 104,043 a.u. in controls to 1,648,689 ± 60,566 a.u. in CRS mice (p = 0.035), and was significantly reduced to 1,199,010 ± 79,509 a.u. by GV-971 treatment (p = 0.011 vs. CRS). The percentage of Iba1⁺ cells similarly increased from 4.53 ± 0.49% in controls to 6.87 ± 0.32% in CRS mice (p = 0.022). Fluoxetine treatment significantly reduced this elevation to 4.38 ± 0.51% (p = 0.022 vs. CRS), while GV-971 treatment decreased it to 4.05 ± 0.44% (p = 0.011 vs. CRS), both restoring levels comparable to those of the control group.

In the hippocampal DG, IBA1 fluorescence intensity was also significantly elevated by CRS, increasing from 1,521,988 ± 114,351 a.u. in controls to 1,909,210 ± 28,694 a.u. (p = 0.037), whereas GV-971 significantly lowered it to 1,513,386 ± 96,441 a.u. (p = 0.017 vs. CRS). No significant differences were observed in the fluoxetine group in the cortex, CA1 or DG. The percentage of Iba1⁺ cells was also markedly elevated by CRS, increasing from 6.48 ± 0.98% in controls to 11.54 ± 0.57% (p = 0.0031). Fluoxetine treatment significantly lowered it to 7.36 ± 0.50% (p = 0.017 vs. CRS), and GV-971 treatment reduced it to 6.39 ± 0.50% (p = 0.0028 vs. CRS), with GV-971 restoring levels nearly identical to those of the control group.

These morphological and quantitative data collectively indicate that GV-971 effectively attenuates CRS-induced microglial hyperactivation in the cortex and hippocampus, suggesting mitigated neuroinflammatory responses.

Beyond its central effects, the therapeutic efficacy of GV-971 likely stems from its ability to rectify upstream disturbances at the gut level. Therefore, we examined whether the drug could reverse CRS--induced gut barrier dysfunction and mucosal damage. Intestinal barrier function was assessed using the FITC-dextran permeability assay (Figure 4A). Serum FITC-dextran reached 1.01 ± 0.07 μg/mL in CRS mice, whereas control mice showed levels of 0.89 ± 0.05 μg/mL, suggesting a slight trend toward increased intestinal permeability. Both GV-971 and fluoxetine markedly reduced serum FITC-dextran relative to the CRS group. In the GV-971 group, serum FITC-dextran decreased to 0.72 ± 0.04 μg/mL (p = 0.0011 vs. CRS), and in the fluoxetine group it decreased to 0.64 ± 0.05 μg/mL (p < 0.001 vs. CRS), indicating restoration of barrier function. Colonic pathology was then evaluated by H&E staining, Alcian blue staining, and histological injury scoring. In control mice, the colonic mucosa was intact, epithelial cells were regularly arranged, and goblet cell morphology was preserved. By contrast, CRS caused pronounced mucosal injury, characterized by disruption of glandular and cellular architecture, crypt shortening, a moderate reduction in goblet cells, diffuse leukocyte infiltration in the lamina propria, and prominent inflammatory cell accumulation in the submucosa (Figure 4D). These alterations were accompanied by a marked increase in the histological injury score, which rose from 2.00 ± 0.32 in control mice to 7.00 ± 0.45 in CRS mice (p < 0.001; Figure 4B). Both treatments ameliorated these pathological abnormalities, with improved structural preservation and reduced inflammatory infiltration. Consistent with these observations, the histological injury score fell to 3.67 ± 0.62 in the GV-971 group (p = 0.0022) and to 2.67 ± 0.67 in the fluoxetine group (p = 0.0013), as compared with untreated CRS mice. Alcian blue staining provided additional evidence of mucus barrier disruption in CRS mice (Figure 4D). Control colons displayed dense, regularly arranged goblet cells along the mucosal surface, whereas CRS colons showed marked goblet cell depletion with a sparse and irregular distribution pattern, consistent with severe compromise of the mucus-secreting barrier. Quantitative counting of goblet cells per crypt corroborated these histological observations (Figure 4C). Goblet cell numbers declined from 21.28 ± 1.75 per crypt in control mice to 14.48 ± 0.76 per crypt in CRS mice (p = 0.0073). Following treatment, goblet cell counts were partially restored: GV-971 increased the number of goblet cells to 19.64 ± 0.54 per crypt (p = 0.0015 vs. CRS), while fluoxetine raised it to 21.07 ± 0.79 per crypt (p = 0.0013 vs. CRS), nearly matching control levels. Together with the FITC-dextran permeability data and histological injury scoring, these results show that GV-971 not only mitigates CRS-induced colonic pathology but also markedly facilitates functional recovery of the intestinal mucus barrier.

Given that gut barrier integrity is inextricably linked to a balanced microbial ecosystem, we next investigated whether the structural restoration observed above coincided with a reconfiguration of the gut microbiota composition. As shown by the alpha diversity analysis (Figure 5A–C), chronic stress and GV-971 altered the gut microbial community in different respects. In terms of richness, the Chao1 index (Figure 5A) tended to be lower in the CRS group than in the control group, implying that chronic stress may be accompanied by a reduction in microbial richness. Notably, GV-971 treatment significantly increased the Chao1 index compared with the CRS group (p < 0.05), suggesting a partial recovery of stress-associated richness loss. The Shannon index (Figure 5B) was similar between the control and CRS groups, while the GV-971 group showed a tendency toward higher values than both groups. This pattern suggests that GV-971 may modestly enhance overall community diversity, and that its effect may be expressed more through reshaping community structure than through large shifts in diversity itself. The Simpson dominance index (Figure 5C), where lower values reflect less concentration of dominant taxa and greater evenness across the community, revealed that the CRS group showed a downward trend relative to the control group. However, GV-971 treatment led to a further significant decrease, with values notably lower than those of the control group (p < 0.01). Taken together, these findings indicate that GV-971 may lessen the over-representation of dominant taxa and help rebalance the microbial community. Beta-diversity analyses in (Figure 5D–F) revealed group-wise differences in overall community structure. In both PCoA plots generated from CLR-Euclidean (Figure 5D) and Bray–Curtis distances (Figure 5E), together with the Bray–Curtis-based NMDS plot (Figure 5F), the CRS group is clearly separated from the control group, consistent with a CRS-induced shift in gut microbial composition. The GV-971 group occupies an intermediate position between CRS and control and appears closer to the control group, suggesting that GV-971 may mitigate, at least in part, the microbial alterations associated with CRS.

Taxonomic profiling at the phylum level showed that Bacteroidota and Bacillota were the two dominant phyla in every group (Figure 5G). In the CRS group, Bacillota decreased slightly, whereas Bacteroidota increased relative to the control group. GV-971 partially normalized these changes, suggesting that it may alleviate the chronic stress-associated disruption of the Bacillota/Bacteroidota ratio. At the species level (Figure 5H), CRS induced a marked shift in gut microbial composition relative to controls, consistent with stress-associated dysbiosis. Specifically, the relative abundances of Duncaniella dubosii, Muribaculum intestinale, Bacteroides caecimuris, Phocaeicola vulgatus, Bacteroides faecium, Phocaeicola dorei, and Parabacteroides merdae tended to increase in CRS mice, whereas those of Paramuribaculum intestinale and Muribaculum gordoncarteri tended to decrease. GV-971 partially reversed these changes, shifting the overall microbial profile toward that of control mice. Several taxa elevated in CRS declined or moved toward baseline after treatment, whereas Muribaculum gordoncarteri, Bacteroides uniformis, and Parabacteroides distasonis increased. These data suggest that GV-971 mitigated CRS-associated disturbances in gut microbial composition and promoted a partial return toward microbial homeostasis.

To identify specific biomarkers associated with GV-971 intervention, we performed LEfSe analysis (Figure 5I). While the CRS group was characterized by the enrichment of Parabacteroides merdae, Parabacteroides johnsonii and Acinetobacter radioresistens, GV-971 treatment significantly expanded the abundance of 22 species, including Bacteroides uniformis, Bifidobacterium pseudolongum, and Lactococcus garvieae. This shift from stress-associated Parabacteroides to a more diverse profile of beneficial commensals suggests that GV-971 mitigates CRS-induced dysbiosis, supporting its potential role in modulating the gut–brain axis.

Finally, we predicted the functional potential of the gut microbiome. For a comprehensive overview, we present the 30 pathways showing the most dramatic shifts (Figure 5J). These differential pathways clustered mainly in energy metabolism, inflammatory lipid metabolism, and xenobiotic/drug metabolism. Specifically, GV-971 administration led to a robust enrichment of 13 pathways, notably those involved in xenobiotic biodegradation (e.g., benzoate, naphthalene, and styrene degradation) and energy metabolism, such as the synthesis and degradation of ketone bodies (ko00072). Two cytochrome P450-related pathways (ko00980 and ko00982) and arachidonic acid metabolism (ko00590) also showed increased representation in GV-971-treated mice, suggesting a systemic shift in metabolic and inflammatory signaling. Conversely, GV-971 significantly suppressed phenylpropanoid biosynthesis (ko00940) and cyanoamino acid metabolism (ko00460) in CRS-exposed mice. Other annotated pathways, including those related to glutathione metabolism and NOD-like receptor signaling, did not reach statistical significance. Functional profiling via MetaCyc (Figure 5K) revealed that GV-971 significantly augmented the microbial capacity for amino acid biosynthesis and central carbon metabolism. Specifically, GV-971 treatment led to increased representation of superpathways for L-arginine (PWY-7400), L-lysine (P4-PWY), and L-methionine (PWY-5345) biosynthesis, alongside multiple tricarboxylic acid (TCA) cycle variants (e.g., PWY-6969, REDCITCYC) and fatty-acid oxidation (FAO-PWY). Notably, pathways supporting cofactor synthesis, such as heme b and molybdopterin, were also enriched. In contrast, the CRS group was characterized by a higher abundance of adenine and adenosine salvage (PWY-6609), phosphopantothenate biosynthesis I (PANTO-PWY), and spermidine biosynthesis III (PWY-6834), all of which were attenuated by GV-971. These data indicate that GV-971 reshapes the gut metabolic landscape, shifting it from stress-related salvage processes toward essential anabolic and energetic functions.

Given that the gut microbiota functions as a metabolic organ, the taxonomic shifts and functional remodeling observed above prompted us to investigate whether these changes translated into altered levels of key microbial metabolites. Targeted plasma metabolomics was performed to characterize CRS-induced metabolic alterations and their modulation by GV-971. Volcano plot analysis (Figure 6A) identified 16 metabolites that differed significantly between the CRS model and GV-971-treated groups, with 5 metabolites being upregulated and 11 downregulated in response to GV-971. Consistent with this, the bar plot of differential metabolites (Figure 6B) quantitatively illustrates that GV-971 markedly increased short-chain fatty acid-related metabolites, specifically valeric and caproic acids, while significantly reducing a series of metabolites, including key products of the tryptophan pathway such as kynurenine, kynurenic acid, quinolinic acid (QUIN), and 5-hydroxyindoleacetic acid (5-HIAA), as well as phenyllactic acid and indole-3-propionic acid, and trimethylamine oxide (TMAO). These substantially altered metabolites are mainly involved in tryptophan metabolism, nicotinate and nicotinamide metabolism, histidine metabolism, and pyrimidine metabolism, pathways that are closely linked to neuroinflammation and neurotransmitter homeostasis. Together, these data indicate that GV-971 partially reverses the abnormal metabolic signature induced by chronic restraint stress. A heatmap of these discriminant metabolites (Figure 6C) showed clear groupwise clustering. When annotated by chemical class, the altered metabolites encompassed amino acids and their derivatives, indole compounds, nucleotides, and related molecules. GV-971 induced a systematic shift in the relative abundance of these metabolites, yielding a metabolic signature that was clearly separated from that of the CRS group.

Pathway enrichment analysis (Figure 6D) confirmed significant over-representation of tryptophan metabolism, nicotinate and nicotinamide metabolism, and histidine metabolism among the differential metabolites, with tryptophan metabolism showing the strongest enrichment. This pattern highlights tryptophan metabolism as a potential metabolic target underlying the antidepressant-like actions of GV-971. The PLS-DA score plot (Figure 6E) further demonstrates a clear separation between the CRS and GV-971 groups in multivariate space, supporting robust GV-971-induced remodeling of the CRS-altered plasma metabolome. Taken together, these data suggest that GV-971 reshapes metabolic homeostasis in CRS-exposed mice through two major gut–brain axis-related processes. First, GV-971 altered tryptophan metabolism, evidenced by reduced 5-HIAA levels, indicating attenuated 5-HT catabolism, and by lower kynurenine and kynurenic acid levels, suggesting a suppression of the inflammation-associated tryptophan–kynurenine pathway. Second, GV-971 increased valeric and caproic acids and microbial-derived fatty acids with reported anti-inflammatory and neuroprotective potential. Collectively, these metabolic shifts likely contribute to the antidepressant-like effects of GV-971 in CRS-exposed mice.

While the concurrent restoration of gut microbiota composition, functional potential, and metabolic profiles following GV-971 treatment suggests a strong association with behavioral improvements, it does not definitively establish causality. To test whether GV-971-induced alterations in the gut microbiota are sufficient to influence stress-related behavior, we performed fecal microbiota transplantation (FMT) from GV-971-treated donor mice into CRS-exposed recipients. After 4 weeks of CRS and FMT, mice were assessed in the open field test (OFT) and forced swim test (FST). In the FST (Figure 7A), mice receiving GV-971-FMT showed a clear reduction in immobility: immobility time fell from 189.59 ± 4.92 s in the CRS group to 157.94 ± 10.87 s (p = 0.015), consistent with an attenuation of depression-like behavior. In the OFT (Figure 7B), GV-971-FMT increased the time spent in the central zone from 14.01 ± 2.91 s in the CRS group to 35.07 ± 9.16 s (p = 0.036), indicating improved exploratory activity and reduced anxiety-like avoidance. Histological analysis of the colon revealed parallel changes (Figure 7C). Hematoxylin and eosin staining showed that GV-971-FMT preserved mucosal integrity, with more regular crypt architecture and less inflammatory cell infiltration than in the CRS group. The corresponding histological injury score (Figure 7D) decreased from 6.25 ± 0.85 in CRS mice to 3.33 ± 0.61 after GV-971-FMT (p = 0.021). Alcian blue staining (Figure 7C) demonstrated stronger mucin-positive signals in colonic crypts in the GV-971-FMT group, and quantification (Figure 7E) confirmed a rise in goblet cell numbers per crypt from 9.00 ± 0.62 in CRS controls to 13.68 ± 0.92 (p = 0.0057). Taken together, these findings indicate that microbiota transferred by GV-971-FMT are sufficient to partially reverse CRS-induced behavioral abnormalities and colonic injury, supporting a causal contribution of the gut microbiota to the antidepressant-like actions of GV-971.

Male C57BL/6J mice (SPF grade, 7–8 weeks old) were sourced from Beijing Vital River Laboratory Animal Technology Co., Ltd., Beijing, China (Production License: SCXK [Beijing] 2021-0011). Animals were group-housed in a barrier-sustained facility under a standard 12 h light/dark cycle, with ambient temperature and relative humidity strictly maintained at 25 ± 2 °C and 50 ± 10%, respectively. Standard chow and water were provided ad libitum. Formal experiments commenced following a 3-to-5-day acclimation phase. All experimental operations were approved by the Institutional Animal Care and Use Committee at Shanghai Institute of Materia Medica, IACUC No. 2022-10-GMY-28.

Mice were fasted for 12 h with free access to water before the assay. FITC-dextran (4 kDa; FD4-1G, Sigma-Aldrich (Merck KGaA, Shanghai, China) was administered by oral gavage at 600 mg/kg. Two hours later, approximately 120 μL of blood was collected from the retro-orbital venous plexus and centrifuged at 12,000× g for 3 min at 4 °C. Plasma was diluted 1:1 with PBS (pH: 7.4) in black 96-well plates. Fluorescence was measured using a multimode plate reader (EnVision, PerkinElmer, Waltham, MA, USA), and FITC-dextran concentrations were calculated from a standard curve.

Upon completion of behavioral testing, animals were anesthetized with Zoletil 50 (25 mg/kg, i.p.) and transcardially perfused with ice-cold physiological saline for 5 min to clear circulating blood. After hepatic blanching and decapitation, brains were swiftly harvested. Guided by a mouse brain stereotaxic atlas [44], the prefrontal cortex and hippocampus were carefully microdissected and snap-frozen for the neurotransmitter assays. In parallel, colonic tissues were isolated and fixed for histological analysis.

Distal colonic segments were fixed in 4% paraformaldehyde for 24 h, processed, embedded in paraffin, and sectioned at 3 μm. Sections were placed on glass slides, deparaffinized in xylene substitute (Cat. No. BA7002C, Baso, Zhuhai, China; three changes, 5 min each), and rehydrated through graded ethanol (100%, 90%, 80%, and 70%) before immersion in distilled water for 2 min. Hematoxylin–eosin staining was performed using a commercial kit (Cat. No. C0105S, Beyotime, Shanghai, China). Sections were stained with hematoxylin for 6 min, rinsed in running tap water, briefly differentiated for 5 s, and blued in running tap water for 10 min. After a brief rinse in distilled water, sections were counterstained with eosin for 3 s, rapidly dehydrated through 70%, 80%, 90%, and 100% ethanol (for approximately 10 s each), cleared in xylene substitute (three changes, 20 s each), and mounted with a slide mounting medium (Cat. No. 110110003, BKMAMLAB, Changde, China). Slides were digitized using a PhenoImager tissue imaging system (Akoya Biosciences, Marlborough, MA, USA). Histological damage was evaluated using a four-parameter scoring system (0–3 per parameter; maximum score: 12) that considered epithelial injury and goblet cell depletion, leukocyte infiltration in the lamina propria, the extent of tissue involvement, and severe inflammatory features, including crypt abscesses and mucosal edema [45].

For goblet cell evaluation, adjacent 3 μm colonic sections were deparaffinized in xylene substitute and rehydrated as described above. Alcian blue–nuclear fast red staining was performed using a commercial kit (Cat. No. C0155M, Beyotime, Shanghai, China). Sections were incubated in 1% Alcian blue solution (pH: 2.5) for 30 min at room temperature, rinsed in running tap water for 2 min, and washed twice in deionized water for 2 min each. Nuclei were then counterstained with 0.1% nuclear fast red for 5 min at room temperature, followed by rinsing in running tap water for 2 min and two further washes in deionized water for 2 min each. Slides were subsequently dehydrated, cleared, and mounted with a slide mounting medium, and then scanned using a PhenoImager tissue imaging system. Goblet cell numbers were quantified by direct counting, and the mean number of goblet cells per crypt was calculated.

Fixed whole brains were cryoprotected in 10%, 20%, and 30% sucrose solutions until they sank, embedded in Tissue-Tek^® O.C.T. Compound (Cat. No. 4583, Sakura Finetek, Torrance, CA, USA), and cut into 30 μm coronal sections on a cryostat (CM1950, Leica, Wetzlar, Germany). Sections containing the dorsal hippocampus (Bregma AP, −1.8 to −2.2 mm) were collected, mounted onto adhesive slides, air-dried at 37 °C, and post-fixed in 4% paraformaldehyde for 10 min at room temperature. After washing with PBS (three times, 5 min each), sections were permeabilized with PBS containing 0.3% Triton X-100 (Cat. No. 85111, Thermo Fisher Scientific, Waltham, MA, USA) for 10 min at room temperature and blocked with 10% goat serum (Cat. No. C0265, Beyotime, Shanghai, China) for 2 h. Sections were then incubated with chicken anti-IBA1 antibody (1:800, Cat. No. 234009, Synaptic Systems, Göttingen, Germany) at 4 °C for 48 h in a humidified chamber. After PBS washes, sections were incubated with Alexa Fluor™ 647-conjugated goat anti-chicken IgY (H + L) secondary antibody (1:500; Cat. No. A-21449; Thermo Fisher Scientific, Waltham, MA, USA) for 2 h at room temperature in the dark. Following further washes, sections were coverslipped with antifade mounting medium containing DAPI (Cat. No. P0131, Beyotime, Shanghai, China). Images of the hippocampus and cortex were acquired using a laser scanning confocal microscope (FV3000, Olympus, Tokyo, Japan) under identical imaging settings. Quantitative analysis was performed in ImageJ 1.54g (NIH, Bethesda, MD, USA). Microglial activation status and IBA1 expression levels were determined by calculating the mean fluorescence intensity (MFI) of IBA1 within individual imaging fields. Meanwhile, the fraction of Iba1-immunopositive cells relative to total DAPI-labeled nuclei was quantified to further characterize microglial population changes.

Fresh fecal pellets were collected from mice under sterile conditions, snap-frozen in liquid nitrogen, and stored at −80 °C until analysis. DNA extraction and sequencing were outsourced to Majorbio Pharma & Biotech Co., Ltd. (Shanghai, China). After homogenization using a FastPrep-24 5G instrument, total DNA was extracted with the FastPure Stool DNA Isolation Kit. DNA samples that passed quality control were used to construct sequencing libraries with the NEXTFLEX Rapid DNA-Seq kit and subjected to paired-end metagenomic sequencing on an Illumina NovaSeq™ X Plus platform.

Fecal metagenomic sequencing data were processed via the nf-core/taxprofiler pipeline (v1.1.8). Raw reads were assessed for quality using FastQC, and adapters and low-quality bases were removed with fastp. Host-derived sequences were depleted by mapping against the mouse reference genome (GRCm39) using Bowtie2. Taxonomic classification and species-level abundance re-estimation were performed utilizing Kraken2 and Bracken, respectively. Data were subsequently imported into R (v4.4.1) via the phyloseq package. For alpha diversity, abundance tables were rarefied to an even sequencing depth; Chao1, Shannon, and Simpson indices were compared using either Wilcoxon rank-sum or Kruskal–Wallis tests. Beta diversity (Bray–Curtis distances) was visualized via principal coordinate analysis (PCoA) and non-metric multidimensional scaling (NMDS), with community-level structural differences statistically evaluated by PERMANOVA (adonis2; vegan). Differentially abundant taxonomic biomarkers were identified using LEfSe (microbiome Marker package; LDA score > 2, p < 0.05). Functional profiling was executed utilizing HUMAnN3 to quantify gene families, KEGG orthologs (KOs), and pathways. Following relative abundance normalization, functional shifts across groups were evaluated. Where applicable, all p-values across diversity and functional analyses were adjusted for multiple testing using the Benjamini–Hochberg method.

Targeted metabolomics profiling of plasma samples was conducted by Metabo-Profile Biotechnology (Shanghai, China) using the Q500 panel. Sample preparation followed the manufacturer’s standard protocol: Briefly, 50 μL of plasma was mixed with 300 μL of ice-cold methanol containing internal standards for protein precipitation. The mixture was homogenized, incubated at −20 °C, and centrifuged at 18,000× g for 20 min at 4 °C. The supernatant was analyzed using an Ultra-Performance Liquid Chromatography–Tandem Mass Spectrometry (UPLC-MS/MS) system, comprising either an ACQUITY UPLC-Xevo TQ-S (Waters, Milford, MA, USA) or a SCIEX Triple Quad™ 6500+ QTRAP (AB Sciex, Framingham, MA, USA). Chromatographic separation was achieved on an ACQUITY UPLC BEH Amide column maintained at 40 °C, optimized for polar metabolite retention. Quality control (QC) samples, prepared by pooling equal volumes of all study samples, were injected periodically throughout the sequence to monitor system stability and data reproducibility. The raw data files generated by UPLC-MS/MS were processed using the TMBQ software (v1.0, Metabo-Profile, Shanghai, China) to perform peak integration, calibration, and quantitation for each metabolite. The self-developed platform iMAP (v1.0, Metabo-Profile, Shanghai, China) was used for statistical analyses.

Statistical analyses were performed using SPSS 26.0 (IBM Corp., Armonk, NY, USA) and visualized with GraphPad Prism 9.0. Data are presented as mean ± SEM. After confirming normality via the Shapiro–Wilk test, one-way ANOVA followed by Tukey’s post hoc test was employed to evaluate inter-group differences in the model optimization experiments (control, CRS 4 h, and CRS 8 h). For treatment experiments, independent-samples t-tests were employed for comparisons between two groups. Statistical significance was defined as p < 0.05.