Adult male C57BL/6 J mice (age: 2 months, weight: 21–25 g) were group housed in individually ventilated cages under a 12-h light/dark cycle (lights were turned off at 6: 00 p.m.). The animals had ad libitum access to food and water. All animal procedures were performed according to the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and all study protocols were approved by the Institutional Animal Care and Use Committee of Gannan Medical University.

Based on a previously reported protocol (Murata et al., 2017; Li et al., 2021), a communication box system was used for inducing psychological stress. The apparatus included a transparent box (30 cm × 30 cm × 30 cm), which was separated by transparent plexiglass plates into nine compartments, with several small holes (2 mm in diameter) carved on the plates between compartments (Figure 1A). The bottom of the communication box was equipped with a grid floor composed of stainless-steel rods (2.5 mm in diameter and spaced 5-mm apart). Plastic insulator plates were placed on the grid floors of four compartments other than the center and four corners. To expose mice to psychological stress, five mice were individually placed in foot-shock compartments, and four mice were individually placed in psychological stress compartments that were covered with a plastic insulator to avert electric shock. Psychologically stressed mice received visual, auditory, and olfactory emotional stimuli (such as screaming, jumping, and evacuation) from mice receiving electric foot shock. The animals were exposed to psychological stress at the same time every day, and they received 0.5–1 mA electric current (10-s duration with 50-s interval) via a shock generator for 30 min each day for 14 days. Mice receiving electric shock were replaced timely to avoid adaptation. Sham-treated controls were placed in compartments, similar to the aforementioned setup, but did not receive any electric stimuli.

The sucrose preference test (SPT) was used to assess anhedonia, which is the core symptom of depression, as previously described (Li et al., 2021). Mice were fed in a single cage; on the first day, they were given two bottles of 1% sucrose. On the second day, mice were given a bottle of water and a bottle of 1% sucrose, and on the third day, all mice were deprived of water for 24 h. On the fourth day, they were given a bottle of 1% sucrose and a bottle of water for 1 h, and the consumption of sucrose and water was recorded. Mice were put back to group housed cages after test. Sucrose preference percentage was calculated using this formula: sucrose solution consumption/ (sucrose consumption + water consumption) × 100%.

The forced swim test (FST) was used to evaluate depressive-associated behavior in animals (Cao et al., 2013). Each mouse was placed in a transparent plexiglass hollow cylinder (height: 25 cm, diameter: 10 cm). The depth of water (23°C–25°C) was approximately 20 cm, and mice were unable to touch the bottom for support. During the experiment, mice were gently placed in water and allowed to freely swim for 6 min. Cumulative immobility time (immobility was defined as the absence of active struggle, only with the body floating in water) was recorded for the last 4 min.

Open field test (OFT) was performed to evaluate anxiety-like behavior, as previously described (Li et al., 2017), using a four-walled black plastic box (40 cm × 40 cm × 30 cm) with a white bottom and no top. The squares adjacent to the walls were designated as “periphery”; all remaining ones were designated as “center” (23 cm × 23 cm). Mice were placed in a corner of the box, and a video camera was used to record their behavioral performance for 10 min. The time spent by mice in the center of the open field was calculated.

Fecal sample preparation and fecal microbiota transplantation (FMT) were performed as previously described (Liu et al., 2021). Briefly, fecal samples were collected from psychologically stressed and non-stressed (i.e., control) mice (n = 10 each group) 24 h after psychological stress paradigm, immediately frozen, and stored at −80°C. The samples (150 mg) were then resuspended in 1 ml sterile PBS, mixed, and centrifuged at 3,000× g, and the supernatant thus obtained was collected, and 10% glycerin was added to the solution. Before FMT, native gut microbiota was eliminated by administering antibiotics (100 mg/kg vancomycin, 200 mg/kg neomycin sulfate, 200 mg/kg metronidazole, and 200 mg/kg ampicillin, Shanghai Macklin Biochemical Co., Ltd., China) for 5 days. Most antibiotics used were non-absorbable, suitable for the gut. Subsequently, the microbiota supernatant (0.1 ml) was administered to these microbiota-depleted mice by gavaging for 14 days. Mice were sham-treated or exposed to psychological stress until the end of FMT.

Fresh fecal samples were collected from mice (n = 6 each group), and DNA was extracted using the cetyltrimethyl ammonium bromide/sodium dodecyl sulfate (CTAB/SDS) method. DNA concentration and purity were monitored on 1% agarose gels. DNA samples were then diluted to 1 ng/μL with sterile water. The V3 − V4 hypervariable regions of the 16S rRNA gene were amplified by performing PCR with specific primers (338F, 5′-ACTCCTACGGGAGGCAGCAG-3′ and 806R, 5′-GGACTACHVGGGTWTCTAAT-3′) and barcodes. Amplicons were excised from 2% agarose gels and purified using the QIAquick Gel Extraction Kit (Qiagen, Germany). Sequencing libraries were generated with the NEBNext® Ultra™ IIDNA Library Prep Kit (New England Biolabs, United States) and sequenced on an Illumina NovaSeq platform (Illumina, United States), which led to the generation of 250-bp paired-end reads. The resultant sequences were merged with FLASH v1.2.11 and quality filtered with FASTP v0.20.0; chimera sequences were detected with Vsearch v2.15.0 and removed. Denoising was performed with DADA2 in QIIME2 (vQIIME2-202,006) to obtain initial amplicon sequence variants (Callahan et al., 2016). Species annotation and multiple sequence alignment were performed using QIIME2 to study phylogenetic relationship of each amplicon sequence variant and to assess differences in dominant species among different groups.

Fecal samples (100 mg, n = 5 each group) were individually ground with liquid nitrogen, and the homogenate was thoroughly mixed in prechilled 80% methanol by vortexing. The samples were then incubated on ice for 5 min and centrifuged at 15,000 g and 4°C for 20 min. The supernatant was analyzed by UHPLC–MS/MS, which was performed on a Vanquish UHPLC system (Thermo Fisher, Germany) coupled with an Orbitrap Q Exactive™ HF-X mass spectrometer (Thermo Fisher, Germany). The samples were injected onto a Hypersil Gold column (100 mm × 2.1 mm, 1.9 μm) using a 17-min linear gradient at a flow rate of 0.2 mL/min. The Q Exactive™ HF-X mass spectrometer was operated in positive/negative polarity mode with spray voltage of 3.5 kV, capillary temperature of 320°C, sheath gas flow rate of 35 psi, aux gas flow rate of 10 l/min, aux gas heater temperature of 350°C. Raw data generated by UHPLC–MS/MS were processed using Compound Discoverer 3.1, peak alignment, peak picking, and quantitation were performed for each metabolite. Normalized data were used for molecular formula prediction based on additive ions, molecular ion peaks, and fragment ions. Subsequently, to obtain accurate qualitative and relative quantitative results, the peaks were matched with mzCloud, mzVault, and MassList. Partial least squares discriminant analysis (PLS-DA) was performed using metaX (Wen et al., 2017). Metabolites with VIP > 1, p < 0.05, and fold change >1.2 or < 0.833 were regarded to be differentially expressed. Their functions and metabolic pathways were studied using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database.

Values represent mean ± SD. Student’s t-test or two-way analysis of variance with Tukey’s post-hoc test were employed to compare differences among groups. GraphPad Prism 7.0 (San Diego, CA, United States) was used for statistical analysis. p < 0.05 indicated statistical significance.

After psychological stress exposure for 14 days, the body weight of mice was recorded, and behavioral tests were performed (Figure 1B). We found stressed mice had significantly lower mean body weight than non-stressed mice (Figure 1C). Depression-like behavior was evaluated via SPT and FST; relative to non-stressed mice, stressed mice displayed a significant decrease in sucrose consumption (Figure 1D) and longer immobility time (Figure 1E). Further, anxiety-like behavior was evaluated via OFT; in comparison to non-stressed mice, stressed mice spent significantly lesser time in the center of the open field (Figure 1F). Collectively, these data indicated that psychological stress induced weight loss and affective disorder behaviors in mice.

To investigate whether psychological stress induces obvious changes in gut microbiota, 16S rRNA genes of fecal microbiota were analyzed. We found that gut microbiota composition was significantly altered in stressed mice. The most abundant phyla were Bacteroidota, Firmicutes, and Campilobacterota (Figure 2A), and the abundance of Proteobacteria and Actinobacteriota was significantly higher in stressed mice than in non-stressed mice (Figure 2B). Furthermore, 14 genera showed significant differences in abundance between stressed and non-stressed mice: the abundance of four genera (Parasutterella and Rikenellaceae_RC9_gut_group being the most significant) was increased and that of the remaining 10 (Bacteroides, Alistipes, and Lactobacillus being the most significant) was decreased (Figures 2C,D). Moreover, a significant decline in diversity (Chao1 index) was observed in the gut microbiota of stressed mice (Figure 2E). We then performed linear discriminant analysis effect size (LEfSe) analysis to identify bacteria phylotypes that were differentially altered. As highlighted by the LEfSe plot and cladogram (Figures 2F,G), 62 taxa (30 for the stressed group and 32 for the control group) were identified as potential microbial markers. Overall, these findings indicated that psychological stress significantly impacted gut microbiota.

To investigate whether changes in gut microbiome contribute to the pathogenesis of affective disorder in response to psychological stress, FMT of “affective disorder microbiota” derived from psychologically stressed mice and FMT of “normal microbiota” derived from non-stressed mice were performed (Figure 3A). Mice transplanted with microbiota were also sham-treated or subjected to psychological stress. SPT, FST, and OFT were conducted the day after FMT and psychological stress treatment. FST results showed that compared to transplantation of “normal microbiota,” transplantation of “affective disorder microbiota” further increased the already high immobility time (Figure 3C). OFT findings revealed that relative to transplantation of “normal microbiota,” transplantation of “affective disorder microbiota” exacerbated the decreased time mice spent in the center of the open field (Figure 3D). With regard to SPT, stressed mice transplanted with either “affective disorder microbiota” or “normal microbiota” exhibited a decreased preference for sucrose, and the decrease was more obvious in mice transplanted with “affective disorder microbiota,” but there was no statistical difference between the two stress + FMT groups (Figure 3B).

It is notable that non-stressed mice transplanted with “affective disorder microbiota” did not show a significant change in behavioral performance relative to those transplanted with “normal microbiota.” These data suggested that FMT of “affective disorder microbiota” derived from psychologically stressed mice resulted in increased stress sensitivity.

Considering the key role of microbial metabolites on gut microbiome modulation and host pathophysiology, we examined the effects of psychological stress on metabolites due to microbial changes. PLS-DA showed that stressed and non-stressed mice exhibited unique metabolite profiles (Figure 4A). Overall, >1,000 gut metabolites were identified, of which 149 were differentially expressed: 118 were down-and 31 were upregulated (Figure 4B). These have been listed and clustered in Figure 4C. KEGG pathway analysis indicated that differential metabolites were mainly associated with α-linolenic acid metabolism, taste transduction, and galactose metabolism (Figure 4D). These findings showed that mice exposed to psychological stress showed substantial differences in fecal metabolite profiles, which seem to contribute to the development of affective disorders.

We performed Spearman correlation analysis to evaluate the relationship between differential bacteria at the genus level and differential metabolites. Several metabolites showed a strong relationship with key differential gut microbiota. The top 10 genera and top 20 metabolites are listed and shown as a heatmap and chord diagram in Figures 5A,B, respectively. Among the bacteria genera that differed significantly from the 16S rRNA analysis, Alistipes and Bacteroides were mainly positively correlated while Parasutterella was mainly negatively correlated with several metabolites, such as 2-hydroxy-2-methylbutanoic acid, dodecanedioic acid, and kynurenic acid.