Bacteroides coprocola protects dopaminergic neurons in rotenone-induced Parkinson’s disease mouse model by modulating gut microbiota dysbiosis and inhibiting the NLRP3 signaling pathway

Six-week-old male C57BL/6 J mice weighing 20–22 g, obtained from Shanghai Jihui Experimental Animal Breeding Co., Ltd, were housed in the Model Animal Facility of ShanghaiTech University. Ethical approval for animal care and use was granted by the Animal Research Ethics Committee of ShanghaiTech University (Approval Number: 20220802001). All procedures were conducted in compliance with the guidelines of ShanghaiTech University’s Institutional Animal Care and Use Committee (IACUC). The experimental mice were of the same batch, sex, and age, and were acclimated in SPF conditions with consistent feed and housing density. Mice were randomly assigned to groups and cages were rotated periodically. Routine management and procedures were standardized. The study employed a blinded approach with fixed testing times to minimize environmental and subjective influences.

Mice were randomly assigned to two groups: the control group and the model group. During the first week, the mice were acclimatized to the animal housing environment and subjected to handling procedures for familiarization to the experimenter’s operations. Over the following three weeks, the model group received daily intraperitoneal injections of rotenone, while the control group received vehicle (2% dimethyl sulfoxide in PBS). After three weeks, the model group was further divided into two subgroups: the Rotenone group and the B. coprocola group. From weeks 4 to 6, mice in the B. coprocola group were treated with B. coprocola once daily, while the control and Rotenone groups received vehicle administration. All mice were weighed daily throughout the six-week study. At week 6, gastrointestinal function tests and behavioral assessments were conducted. Finally, all mice were sacrificed at week 6 for further analysis.

Rotenone (Sigma-Aldrich, Cas-83–79-4, St. Louis, MO) was dissolved in 2% dimethyl sulfoxide (Beyotime, ST038, Shanghai, China). The fresh rotenone solution was intraperitoneal injected at 2 mg/kg once a day for 3 weeks[14].

B. coprocola (Mingzhoubio, Ningbo, China) were inoculated in a liquid medium (Table S1) with deoxygenation and cultivated in a 37 °C incubator for 2 days. An anaerobic environment was maintained throughout the cultivation process. For the B. coprocola group, B. coprocola was dissolved in PBS (10⁸ CFU/mouse per day), and the bacterial suspension was then delivered to each mouse via oral gavage within 15 min. The administration was made once a day, for 3 weeks.

Primers for the 16S rRNA of B. coprocola were designed, the bacterial DNA was amplified in vitro through PCR (10102ES03, Yeasen, Shanghai, China), and products were sequenced by Tsingke Biotech Company. The V3-V4 regions of the microbial 16S RNA were amplified with the paired primers (forward primer: 5′-CCTACGGGRSGCAGCAG-3′; reverse primer: 5′-GGACTACVVGGGTATCTAATC-3′). Sequence identification of the species was performed using the BLAST function on National Center for Biotechnology Information (NCBI).

Thirty minutes prior to euthanasia, mice were orally administered with 0.3 mL of a 2.5% Evans blue solution (Sigma-Aldrich, Cas-314-13-6, Darmstadt, Germany), which was dissolved in a 1.5% carboxymethyl cellulose sodium vehicle (Sigma-Aldrich, Cas-9004-32-4, Darmstadt, Germany) to assess intestinal transit distance. Mice were subjected to anesthesia via intraperitoneal injection of tribromoethanol (100 mg/kg) and subsequently underwent transcardial perfusion with 1 × PBS followed by a 4% paraformaldehyde solution. Subsequently, intestinal transit distance, defined as the distance from the pylorus to the most distal point of dye migration, was measured. Furthermore, the length of the colon was determined by measuring the distance from the terminus of the cecum to the anus [15].

Following a 2 h fasting period, each mouse was transferred from its home cage to a clean, transparent polycarbonate cage and stayed there for an additional 2 h. Fecal pellets were subsequently collected and enumerated. The fecal pellets were immediately weighed to obtain wet weight, and weighed after desiccation at 85 °C for 24 h to obtain dry weight. The fecal water content was calculated as a percentage based on the discrepancy between the wet and dry weights.

Antibodies used for flow cytometry were CD11b-APC (BioLegend, Cat-101212, San Diego, CA) and CD45-Percp/Cy5.5 (BioLegend, Cat-103132) obtained from BioLegend, CD86-BV421 (BD, Cat-564198, San Jose, CA), CD206-PE (BD, Cat-568273), F4/80-BV650 (BD, Cat-743282) and CD80-FICT (BD, Cat-561954) obtained from BD. For surface marker staining, cells were blocked with Fc Shield (anti-mouse CD16/CD32, BD, Cat-553141) and stained with the Zombie Aqua™ Fixable Viability Kit (BioLegend, Cat-423101) for 15 min to discriminate live from dead cells. Then the cells were washed and incubated with a cocktail of fluorochrome-conjugated antibodies against surface markers for 30 min at 4 °C in the dark. Next, the cells were permeabilized using the eBioscience™ Foxp3 kit for 30 min (Thermo Fisher Scientific, Cat-00-5523-00, Waltham, MA). Finally, the cells were stained for intracellular indexes CD206-PE. Between each step, the cells were extensively washed with FACS buffer (2% FBS and 2 mM ethylenediaminetetraacetic acid [EDTA] in PBS). The prepared samples were measured and analyzed using a Cyto FLEX flow cytometer (Beckman Coulter, Brea, CA).

Following mouse euthanasia, fresh brain, blood, and colonic tissues were collected for preparation of single-cell suspensions, which were subsequently subjected to single-cell analysis utilizing fluorochrome-conjugated antibodies.

The fresh peripheral whole blood was collected into EDTA anticoagulant tubes. The whole blood samples were subjected to lysis using a lysing solution (BD, Cat-555899) and subsequently washed with 1 × PBS.

The brain and colon tissues were harvested, washed, and minced into small pieces. Brain mononuclear cells were isolated using the Neural Tissue Dissociation Kit (Miltenyi Biotec, 130–107-677, Bergisch Gladbach, Germany) and pepsin (MedChemExpress, HY-P1635, Shanghai, China) according to the manufacturers’ protocol. Briefly, the pieces were transferred into an appropriately sized conical tube, rinsed with cold Hank’s balanced salt solution (HBSS), and then centrifuged (300 × g, 2 min) at room temperature. After carefully aspirating the supernatant, preheated enzyme mix 1 (37 °C, 10 min) from the Neural Tissue Dissociation Kit was added to digest the tissue pieces for 15 min, followed by addition of preheated enzyme mix 2 (37 °C, 10 min) to the tissue sample for an additional 10 min. Subsequently, HBSS was added to resuspend the tissue, and single-cell pellets were isolated by passing through a 30-μm cell strainer.

The colon tissues were incubated with 8 mL of digestion buffer composed of RPMI-1640 (Thermo Fisher Scientific, 11875093), 5% fetal bovine serum (FBS; Capricorn, FBS-26A, Palo Alto, CA), 62.5 μg/mL Liberase (Sigma-Aldrich, Cat-5401020001, Darmstadt, Germany), and 50 μg/mL DNase I (Roche, Cat-11284932001, Mannheim, Germany) at 37 °C for 30 min. The cell pellets were then resuspended in 1 mL of 40% Percoll (40501ES60, Yeasen) solution and transferred to a 15 mL conical tube containing 4 mL of 40% Percoll solution. Using glass Pasteur pipettes, 2.5 mL of 80% Percoll solution was carefully layered at the bottom of the tube. Cells were collected from the interface [16].

Total proteins were extracted from brain or colon tissue or from bone marrow-derived macrophages (BMDMs) using RIPA lysis buffer containing a cocktail of protease and phosphatase inhibitors (Beyotime, P1046, Shanghai, China). Protein concentrations were determined using a BCA protein assay kit (Thermo Fisher Scientific, A55860). Proteins were separated by 7.5%–15% SDS–polyacrylamide gel electrophoresis, transferred to a PVDF membrane, and blocked with 5% non-fat milk powder or 5% bovine serum albumin (BSA) for 1 h. Then the membrane was incubated with primary antibody at 4 °C for 16 h, followed by secondary antibody HRP-conjugated anti-rabbit (1:5000, Abclonal, AS038, Wuhan, China) or anti-mouse (1:5000, Abclonal, AS003). The membrane was processed with an ECL kit (Epizyme, SQ201, Shanghai, China), and the gray values of protein bands were recorded and analyzed.

The primary antibodies included anti-NLRP3 (1:1000, Adipogen, AG-20B-0014-C100, Liestal, Switzerland), anti-interleukin-6 (IL-6) (1:1000, Abclonal, A0286), anti-ZO-1 (1: 5000, Proteintech, 21773-1-AP, Wuhan, China), anti-Occludin (1:5000, Proteintech, 27260-1-AP), anti-NF-κB (1:1000, Abclonal, A19653), anti-sirtuin type 1 (SIRT1) (1:1000, CST, 9475, Danvers, MA), anti-caspase-1 (1:1000, CST, 24232), anti-ionized calcium-binding adapter molecule 1 (Iba-1) (1:1000, CST, 17198), anti-phospho-IκBα (1:5000, HUABIO, HA722770, Hangzhou, China), anti-free fatty acid receptor (FFAR) 2 (1:1000, Proteintech, 84544-1-RR), anti-FFAR3 (1:1000, HUABIO, ER63606), anti-Toll-like receptor 4 (TLR4) (1:1000, Abclonal, A5258), anti-myeloid differentiation primary response protein 88 (MyD88) (1:1000, CST, 4283), and anti-interleukin-1β (IL-1β) (1:1000, Thermo, P420B, Waltham, MA). β-Actin (1:100000, Abclonal, AC026) was used as the internal control.

Mice were anesthesized by intraperitoneal injection of tribromoethanol (100 mg/kg) and subsequently underwent transcardial perfusion with 1 × PBS followed by 4% paraformaldehyde. Thereafter, the brain and colon tissues were collected, fixed in 4% paraformaldehyde for 48 h and subsequently dehydrated in a 30% sucrose solution for 72 h. The brain and colon tissues were then frozen and embedded in optimal cutting temperature (OCT) compound (Sakura Finetek, Torrance, CA, Cat-4583) prior to sectioning into uniform slices (brain: 40 μm; colon: 12 μm) using a Leica Microsystems microtome (Wetzlar, Germany). The frozen sections were thawed and mounted onto positively charged slides (ProbeOn Plus; Thermo Fisher Scientific) before being stored at –80 °C.

For antigen retrieval from frozen brain sections, the sections were subjected to 20–30 min of microwave treatment (Midea, Foshan, Guangdong, China, Cat-M1-211A) in citrate buffer (pH 6.0) (Servicebio, Cat-G1202). Then the sections were blocked in a blocking solution composed of 10% BSA and 0.3% Triton X-100 (Sigma, X100-5, Darmstadt, Germany) in PBS for 30 min. The sections were then incubated with 3% hydrogen peroxide for 15 min. Subsequently, the sections were incubated overnight at 4 °C with primary antibodies diluted in normal goat serum (2%) (Beyotime, Cat-C0265, Shanghai, China), followed by incubation with biotinylated secondary antibodies and horseradish peroxidase (HRP)− streptavidin or the appropriate secondary antibodies. Nuclei were visualized using DAPI solution. Representative images were captured using a fluorescence microscope (Nikon CSU-W1 Sora 2 Camera; Olympus Corporation, VS120-S6-W). The number of positive cells was quantified using Image Pro Plus 6.0 software. Each section was analyzed based on five randomly selected fields.

The primary antibodies included anti-tyrosine hydroxylase (TH) (1:1000, Thermo Fisher Scientific, P21962), anti-α-synuclein (D37A6) antibody (1:1000, CST, 4179), anti-Iba-1/AIF-1 (E4O4W) (1:200, CST, 17198), anti-Occludin (1:1000, Proteintech, 27260-1-AP), and anti-ZO-1 (1:2000, Proteintech, 21773–1-AP). The secondary antibodies employed were Alexa 488-conjugated goat anti-rabbit IgG (1:1000, Thermo Fisher Scientific, A32723), HRP-conjugated donkey anti-rabbit (1:1000, Abclonal, AS038), and Alexa 594-conjugated goat anti-rabbit IgG (1:1000, Thermo Fisher Scientific, A11005).

Total RNA was extracted using TRIzol Reagent (Vazyme Biotech, R411-01, Nanjing, China) and reverse transcribed into cDNA via a PrimeScript RT-PCR kit (Abclonal, RK20433). qRT-PCR was performed using SYBR Premix Ex Taq (Vazyme Biotech, R433) on the QuantStudio 7 platform (Life Technologies, Waltham, MA). The primers (Table S2) were custom-synthesized by GENEWIZ (Shanghai, China). The fluorescence signals corresponding to the target genes were analyzed using the 2^−∆∆Ct method for relative quantification. β-Actin and GAPDH served as endogenous controls.

Quantification of murine IL-1β, IL-6, LPS, and lipopolysaccharide-binding protein (LBP) levels was performed using ELISA kits purchased from Jianglai Industrial Limited by Share Ltd. (Shanghai, China, Cat-JL20691, JL29644, JL20268, JL18442).

At the end of the six-week experiment, 6 mice from each experimental group were used for microbiota sequencing analysis. Mice were individually housed in a sterile, autoclaved cage, and five fresh fecal pellets were collected for each mouse and promptly transferred into a sterile EP tube. All fecal samples were rapidly frozen and stored at –80 °C for subsequent analysis. Genomic DNA was isolated from 200 mg of each fecal sample using the QIAamp DNA Stool Mini Kit (QIAGEN, 51604, Hilden, Germany). The quality and integrity of the extracted DNA were subsequently validated by 1.2% agarose gel electrophoresis. For library construction, a two-step PCR amplification targeting the V3-V4 hypervariable region of the 16S rRNA gene was performed. The PCR amplification utilized universal primers 357F (5′-ACTCCTACGGRAGGCAGCAG-3′) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′). The thermal cycling was as follows: initial denaturation at 95 °C for 3 min, followed by 30 cycles of denaturation at 98 °C for 20 s, annealing at 58 °C for 15 s, and extension at 72 °C for 20 s, with a final extension at 72 °C for 5 min. The quantified amplicons were then pooled at equimolar concentrations for Illumina MiSeq sequencing (Illumina, San Diego, CA). The entire experimental workflow, encompassing DNA extraction, quality assessment, library construction, and high-throughput sequencing, was executed by TinyGene Bio-Tech (Shanghai, China).

Bone marrow cells isolated from femurs of mice were cultured for 7 days in the presence of 20 ng/mL recombinant mouse macrophage colony-stimulating factor (M-CSF, 315-02-10UG, PeproTech, Cranbury, NJ) in complete RPMI-1640 medium containing 10% FBS, 10 mM glucose, 2 mM L-glutamine, and 100 U/mL penicillin–streptomycin. During cell culture, the morphological changes and growth status of cells were observed with an inverted microscope (Nikon Ti-S) every day.

To further explore the effects of sodium acetate (NaA) (Absin, abs42027403, Shanghai, China) and sodium butyrate (NaB) (MedChemExpress, HY-B0350A) on M1 and M2 polarization, on day 7 of culture, M0 macrophages were harvested and then were stimulated for 24 h with 1 μg/ml LPS (SJ-MB0005, Shandong Sparkjade Biotechnology, China) and 1 mM ATP (SJ-MN0102A, Shandong Sparkjade Biotechnology) for the generation of M1 macrophages. Then acetate or butyrate was added for 24 h. Groups were as follows: Control group, LPS + ATP group, LPS + ATP + NaA group (5 mM NaA), LPS + ATP + NaB group (0.5 mM NaB).

siRNA targeting FFAR2/3 and control siRNA were synthesized by GenePharma (Table S3). Four individual siRNA sequences were constructed to form the siRNA Smart Pool to reduce the off-target effects of siRNAs. Transfections were performed using the Lipofectamine™ RNAiMAX reagent (Invitrogen, 13778030, Waltham, MA). At 48 h post-transfection and at 24 h of inflammation modeling, cells were treated with 5 mM NaA or 0.5 mM NaB for 24 h. Western blotting was performed to confirm the downregulation of FFAR2/3 targeted by siRNA.

Fecal sample from each mouse was collected in the morning using designated fecal collection containers. For each mouse, 400 mg of fresh fecal sample was used for SCFA analysis after grinding and sonication as pretreatment steps. Additionally, fresh B. coprocola bacterial culture solution was collected and centrifuged to obtain the supernatant for SCFA analysis. Individual SCFAs in fecal samples and B. coprocola bacterial supernatant were quantified by gas chromatography-mass spectrometry (GC–MS) and liquid chromatography tandem mass spectrometry (LC–MS/MS), conducted by TinyGene Bio-Tech (Shanghai, China) Co., Ltd., following standardized protocols.

High-quality genomic DNA was extracted and subjected to a hybrid sequencing strategy utilizing both Illumina (short-read) and PacBio/Nanopore (long-read) platforms. Following quality control and adapter trimming, the sequencing data were assembled into a complete circular genome using Unicycler (v0.4.8). The final assembly was annotated to identify protein-coding genes, non-coding RNAs, genomic islands, and mobile genetic elements. Functional annotation of the predicted genes was carried out by comparing their sequences against major public databases. This sequencing was executed by TinyGene Bio-Tech (Shanghai, China).

Statistical analyses were conducted using GraphPad Prism (version 8.01; GraphPad Software Inc., San Diego, CA). One-way analysis of variance (ANOVA) was used to assess differences among multiple groups. When ANOVA indicated a significant difference, post-hoc comparisons were performed using Tukey’s test (for data meeting assumptions of normality and homogeneity of variances) or Games-Howell test (for nonparametric data analyzed by Kruskal–Wallis test). Comparisons of numerical data between two groups were performed using the Mann–Whitney U test or unpaired Student’s t-test. For nonparametric analyses, the Kruskal–Wallis H test or Mann–Whitney U test was applied. Data are presented as mean ± standard deviation (SD). P < 0.05 was considered as statistically significant.

To evaluate the effects of B. coprocola in PD, mice received intraperitoneal rotenone for 3 weeks, followed by 3 weeks of treatment with B. coprocola or vehicle (Fig. 1a). Weight reduction, motor disturbances, and gastrointestinal (GI) impairments frequently manifest in animal models of PD [17]. Rotenone administration caused significant weight loss during the induction period. From weeks 4 to 6, treatment with B. coprocola mitigated the weight loss in rotenone-induced mice, accompanied by faster weight recovery than controls (Fig. 1b). However, there was no significant difference in the area under the curve between rotenone and B. coprocola groups (Fig. 1c).

At 6 weeks, behavioral assays were conducted to evaluate the motor functions of mice (Fig. 1a). Mice subjected to rotenone intoxication exhibited significant motor impairments compared to the control group, as evidenced by reduced latency to fall from the Rota-Rod (P < 0.0001, Fig. 1d), prolonged climbing times in the pole test (P < 0.001, Fig. 1e), and increased time spent on the beam in the beam walking test (P < 0.01, Fig. 1f). Conversely, mice treated with B. coprocola demonstrated significant improvements in these behavioral parameters compared to the rotenone-induced PD group.

Moreover, mice in the rotenone-induced PD group exhibited a significant reduction in intestinal transit distance (P < 0.001, Fig. 1g, h) and colon length (P < 0.001, Fig. 1i, j) compared to the control group. In contrast, treatment with B. coprocola significantly ameliorated these GI impairments induced by rotenone (P < 0.05, P < 0.01, Fig. 1g-j). In addition, mice subjected to rotenone exhibited a marked decrease in fecal water content percentage (P < 0.01, Fig. 1k), which was significantly increased following B. coprocola administration (P < 0.05, Fig. 1k).

In summary, rotenone intoxication induced weight reduction, motor impairments, and GI dysfunction in the mice, while administration of B. coprocola substantially mitigated these PD-related manifestations.

Degeneration of dopaminergic neurons and accumulation and phosphorylation of α-syn are two key histological characteristics of PD [18]. The cell bodies of dopaminergic neurons are situated in the substantia nigra pars compacta (SNc), with their axons projecting to the caudate-putamen (CPu) region [19].

In the rotenone-induced PD mouse model, the numbers of TH⁺ cells in the SNc and CPu were significantly diminished compared to the control group (P < 0.01), whereas treatment with B. coprocola substantially mitigated this neuronal loss (P < 0.05) (Fig. 2a–c). Immunofluorescence staining revealed a marked increase of α-syn and pS129-α-syn levels in the SNc of mice in the rotenone-induced PD group. Conversely, B. coprocola treatment notably reduced the levels of α-syn and pS129-α-syn in this region (Fig. 2d–f, m, o). The increased levels of α-syn and pS129-α-syn protein in the gastrointestinal tract were also a significant pathological feature associated with the progression of PD [20]. Immunofluorescence analysis indicated that the levels of α-syn and α-syn-ser-129 in the colon of rotenone-induced mice were significantly elevated compared to that of the control group and the B. coprocola-treated group (Fig. 2l–o).

Activation of microglia plays a role in neuroinflammation and dopaminergic neuronal loss in the brain during the progression of PD [21]. In the current study, immunofluorescence staining of Iba-1, a marker for microglia [22], in the SNc and CPu, was performed to assess glial cell reactivity. A significant increase in Iba-1⁺ cells was observed in the SNc and CPu of mice in the rotenone-induced PD group (SNc: P < 0.01; CPu: P < 0.001, Fig. 2g–i). However, treatment with B. coprocola significantly reduced Iba-1⁺ cells in the SNc and CPu (SNc: P < 0.05; CPu: P < 0.01, Fig. 2g–i). Western blotting analysis of Iba-1 in the midbrain also revealed a similar trend (P < 0.0001; P < 0.01, Fig. 2j, k).

Collectively, these results demonstrated that B. coprocola intervention markedly ameliorated PD-related histopathology in both the brain and the colon of rotenone-induced murine model, and that B. coprocola treatment is associated with modulation of immune cell activity.

An increasing number of studies has identified gut microbiota dysbiosis in both PD patients and PD animal models, highlighting its crucial role in the pathogenesis of PD [23]. To elucidate if B. coprocola administration protects against the rotenone-induced PD through modulation of the microbiome community structure, 16S rRNA sequencing was performed on fecal samples from mice in the three experimental groups.

Initially, α-diversity analysis was conducted to evaluate the richness and diversity of bacterial taxa. The mice in the rotenone-induced PD group demonstrated significant reductions in the Observed species index (P = 0.021), Chao1 richness estimator (P = 0.0084), ACE richness index (P = 0.0102), Shannon diversity index (P = 0.0446), and Phylogenetic Diversity (PD) whole tree index (P = 0.0133), as well as an increase in the Simpson dominance index (P = 0.0504), relative to the control and B. coprocola-treated groups (Fig. 3a–f). These results suggest that B. coprocola treatment ameliorated the rotenone-induced alterations in microbial abundance and diversity.

Moreover, β-diversity analysis revealed consistent results. At week 6, the microbiome community structure of the rotenone-induced PD group exhibited a significant divergence from that of the control and B. coprocola groups, as demonstrated by OTU-Jaccard-Anosim analysis (R = 0.1984, P = 0.013) and Unweighted-Anosim analysis (R = 0.1086, P = 0.089). The microbiome community structures of the control and B. coprocola groups remained comparable, indicating that B. coprocola intervention markedly modified the diversity of gut microbiota (Fig. 3g, h).

Additionally, we performed taxonomic profiling at the genus level (Fig. 3i). Differential abundance analysis using LEfSe (linear discriminant analysis effect size) identified key microbial taxa with significant intergroup differences (Fig. 3k). Specifically, the relative abundances of Parabacteroides, Odoribacter, Alistipes and Bacteroides were diminished in the rotenone-induced PD group compared to the control, whereas Akkermansia and Bifidobacterium were enriched. Notably, B. coprocola administration markedly restored the relative levels of these bacterial genera, including Parabacteroides, Odoribacter, Alistipes, Bacteroides, Akkermansia, and Bifidobacterium (Fig. 3j, k).

Gut microbial disorders can trigger chronic inflammation of the body’s intestinal and even systemic inflammation [24]. The rotenone-induced PD mouse model manifests compromised blood–brain barrier (BBB) and intestinal barrier integrity, with concomitant generation of inflammatory cytokines and pathogenic LPS [17, 25]. We assessed the expression levels of two major tight junction proteins, ZO-1 and occludin, in the midbrain and colon. qRT-PCR and Western blotting analyses demonstrated that the expression of ZO-1 and occludin was significantly diminished in the rotenone-induced group compared to the control group, whereas treatment with B. coprocola substantially restored their expression (all P < 0.05) (Fig. S1a,b; Fig. 4c–e). Immunofluorescence staining of the colon further revealed that the fluorescence intensities of ZO-1 and occludin were markedly decreased in the rotenone-PD group but significantly increased in the B. coprocola group (P < 0.01, P < 0.05) (Fig. 4a, b).

Then, we quantified the levels of LPS and LBP in the midbrain, serum and colon. ELISA analysis indicated that both LPS and LBP levels were elevated in the rotenone-induced PD group compared to the control group. In contrast, B. coprocola administration significantly attenuated these elevations (all P < 0.05, Fig. 4f–k).

Collectively, these findings suggest that B. coprocola treatment restores barrier integrity in the rotenone-induced mouse model, protecting against microbial toxin leakage.

Systemic chronic inflammation activates various types of immune cells in the body to exert immune function and thus modulate the inflammatory response. We next investigated the potential impact of B. coprocola treatment on myeloid cells regulating systemic chronic inflammation, by measuring the number and phenotypes of M1/M2 type macrophages/microglia in the gut-blood-brain axis.

Macrophages were defined as CD11b⁺CD45⁺F4/80⁺cells in the colon and blood, and the populations of CD80 (M1 marker) and CD206 (M2 marker) macrophages were gated using Fluorescence Minus One (FMO) control [26] (Fig. S5). In the colon, the percentage of CD80⁺ macrophages was remarkably increased in the rotenone-induced PD group mice compared with the control group mice (P < 0.01), while B. coprocola treatment significantly decreased the CD80⁺ macrophages (P < 0.05) (Fig. 5a, e). In contrast, the percentage of CD206⁺ macrophages was remarkably decreased in the rotenone-PD group mice compared with the control group mice (P < 0.001), but markedly elevated in the B. coprocola group (P < 0.05) (Fig. 5a, f).

The preceding data suggested that intestinal barrier damage in the rotenone-induced PD mouse model led to the leakage of LPS into peripheral tissues. In the blood, the results illustrated that the percentage of CD80⁺ macrophages increased significantly in the rotenone-induced group compared to the control group, whereas B. coprocola treatment markedly decreased the CD80⁺ macrophage levels (P < 0.001, P < 0.01) (Fig. 5b, g). However, it did not markedly affect the variation in cell counts of CD206⁺ macrophages (Fig. 5c, j).

In the brain, we analyzed the accumulation of infiltrating macrophages and activated microglia, which were defined as CD11b⁺, CD45⁺ high [26, 27]. Flow cytometry analysis revealed that B. coprocola treatment of rotenone-induced mice strikingly reduced the percentage of infiltrating macrophages and microglia defined as M1-type cells compared with the rotenone-induced PD group mice. However, B. coprocola administration did not significantly affect the percentage of M2-type microglia/macrophages in the brain (Fig. 5d, h, i).

The NLRP3 signaling pathway represents a highly intricate and well-regulated inflammasome-associated pathway that plays a pivotal role in orchestrating immune and inflammatory responses, particularly in the context of the microbiota-gut-brain axis. LPS, a major component of the outer membrane of Gram-negative bacteria, is a potent pro-inflammatory molecule. It directly binds to TLR4 with high affinity, initiating a cascade of intracellular signaling events. Specifically, up-regulated LPS can act on the macrophage TLR4 receptor, bind to the key junction protein MyD88, further promote IκBα phosphorylation, and release NF-κB to activate NLRP3 inflammasome. The activated NLRP3 inflammasome can promote the maturation of caspase-1, which in turn leads to the maturation and secretion of inflammatory factors, triggering an inflammatory response [28]. This pathway plays a dominant role in the microbiota-gut-brain axis [29].

Therefore, we detected the activation status of NLRP3 pathway in the midbrain and the colon by Western blotting. Interestingly, we observed a striking consistency in the activation patterns of the NLRP3 pathway components between the colon and the midbrain tissues. This finding suggests a coordinated regulation of this inflammatory pathway across different regions, potentially highlighting a systemic response in our experimental model. The protein levels of TLR4, MyD88, p-IκB-α, NF-κB, NLRP3, and caspase-1 in the rotenone-induced PD group were significantly increased compared to those in the control group. Conversely, the levels of these proteins were remarkably reduced by B. coprocola administration (all P < 0.05) (Fig. 6a–d).

In addition, we examined the expression levels of NLRP3 downstream inflammatory factors in the midbrain, colon and blood by Western blotting, qRT-PCR and ELISA. The protein level and mRNA expression of IL-1β and IL-6 were significantly increased in the rotenone-induced PD group. However, B. coprocola treatment significantly down-regulated the expression of these inflammatory factors (Fig. 6e–g, Fig. S2a, b). This suggests that B. coprocola can alleviate PD-like pathology by modulating systemic inflammation in the rotenone-induced PD mouse model through the NLRP3 signaling pathway.

To investigate the relationship between B. coprocola and metabolites among the three groups of mice, and to identify potential bioactive metabolites of B. coprocola, we first performed principal components analysis (PCA) and partial least squares discriminant analysis (PLS-DA) for multivariate analysis (Fig. 7a, b). The PCA score chart and the PLS-DA model showed that the fecal samples in the three groups were clearly separated, and the clustering effect was relatively obvious. The clustering of the control group and B. coprocola tended to be more consistent. This further suggested that distinct metabolic principal components play a key role in influencing the progression of rotenone-induced PD-like symptoms.

To identify potential bioactive metabolites, we performed metabolomic analyses targeting SCFAs in feces from the three groups, as well as in B. coprocola bacterial liquid supernatants. We found that acetic acid (P = 0.022) and butyric acid (P = 0.008) exhibited statistically significant differences among the three groups (Fig. 7c–e; Fig. S4). Therefore, acetic acid and butyric acid are likely the key metabolites involved in the regulation of the PD pathological progression by B. coprocola (Fig. 7c–e).

Finally, we performed a correlation analysis between the differentially abundant bacterial genera and SCFAs. Alistipes and Odoribacter showed a positive correlation with butyric acid, while Anaeroplasma and Candidatus_Soleaferrea exhibited a significant positive correlation with acetic acid. These findings suggest that bacterial genera such as Alistipes, Odoribacter and Anaeroplasma are closely associated with the metabolism of SCFAs (e.g., butyric acid and acetic acid) and play a beneficial role in maintaining the host gut health. In addition, B. coprocola may regulate the level of SCFAs in the gut by affecting the abundance of other SCFA-producing genera.

To further explore the functional genes of B. coprocola, we performed whole-genome sequencing. This enabled us to understand the composition, structure, and function of B. coprocola genes, uncover new genetic resources, and provide foundational data for in-depth research on B. coprocola.

Based on the COG database enrichment gene pathways (Fig. 8a), B. coprocola was identified as a free-living bacterium with strong independent living ability and active metabolism, particularly excelling in carbohydrate utilization. It exhibited a robust ability in genetic information processing, diverse metabolic potential, capacity to sense and respond to environmental changes, and low motility and pathogenic potential.

The KEGG pathway enrichment (Fig. 8b) further elucidated its metabolic characteristics. The top-ranked pathways, including Global and overview maps, Carbohydrate metabolism, and Amino acid metabolism, highlighted its complex metabolic network. We further refined our analysis of genes enriched in fatty acid biosynthesis in B. coprocola and identified a total of nine genes related to fatty acid synthesis, including ACP (acyl carrier protein), a key gene involved in butyrate synthesis [30]. Simultaneously, among the genes associated with butyrate synthesis, genes related to butyrate kinase were enriched (e.g. buk, KO ID: K00929↗). The presence of buk serves as one of the most important markers indicating that microorganisms possess classical butyrate synthesis capability [31] (Fig. 8d, Tables S5 and S6). This indicated that B. coprocola possesses the capacity to produce both acetate and butyrate.

Looking into the genomic structure of B. coprocola (Fig. 8c, Table S4), the high proportion (88.5%) of coding sequences in the approximately 4.59 Mb genome was a typical feature of bacterial genomes. The significant difference in GC content between gene regions (42.7%) and intergenic regions (34.2%) indicated a preference for GC bases in coding areas.

Overall, these findings enhanced our understanding of B. coprocola’s ecological adaptation and functional potential, and further confirmed the role of B. coprocola in fatty acid synthesis.

In vivo, we have demonstrated that B. coprocola gavage can influence the polarization of macrophages/microglia in the gut-blood-brain axis while alleviating the systemic chronic inflammatory response in rotenone-induced PD mice. Based on these findings, we further investigated whether the potential functional metabolites of B. coprocola can affect the polarization of macrophages, thereby regulating inflammatory responses.

To better reflect in vivo conditions, we induced differentiation of mouse bone marrow cells into primary macrophages. The primary macrophages were treated with LPS + ATP (LPS: 1 μg/ml, ATP: 1 mM) to induce NLRP3 inflammasome activation. M1 macrophages defined as [CD11b/F4/80/CD86]⁺ and M2 macrophages defined as [CD11b/F4/80/CD206]⁺, which were gated using FMO control [26] (Fig. S5). As expected, both NaA and NaB significantly downregulated the percentage of CD86⁺ M1 macrophages. However, only NaB upregulated the proportion of CD206⁺ M2 macrophages (Fig. 9a, b).

FFAR2/3 are the key receptors through which acetic acid and butyric acid exert effects on immune cells [32]. FFAR2/3 protein levels were significantly decreased by treatment with FFAR2/3-siRNA 4mix smartpool (Fig. S3). Separate silencing of FFAR2 or FFAR3 partially blocked the effects of NaA/NaB on M1/M2 macrophages, and simultaneous silencing of both receptors further diminished these effects. The results indicated that FFAR2 and FFAR3 are key receptors through which NaA and NaB influence macrophage polarization (Fig. 9c-f).

Then, we examined the expression of FFAR2/3 receptor and its downstream deacetylase SIRT1 in the primary macrophages. SIRT1 deacetylates specific lysine residues on p-IκB-α, thereby inhibiting the phosphorylation of the NF-κB complex. SIRT1 upregulation leads to the binding of unphosphorylated IκBα to NF-κB, blocking NF-κB translocation into the cell nucleus. Western blotting results showed that the protein levels of FFAR2/3 and SIRT1 were significantly upregulated in the LPS + ATP + NaA/NaB group compared with the LPS + ATP group. The upregulated SIRT1 further inhibited phosphorylation of the downstream p-IκB-α. Nonphosphorylated IκBα could not further release mature NF-κB transcription factor, thereby further inhibiting NLRP3 synthesis and ultimately reducing the release of inflammatory cytokines. Consistent with the in vivo experimental results, treatment with NaA and NaB significantly reduced the protein levels of NLRP3 inflammasome, its upstream signaling molecules p-IκB-α and NF-κB, and the downstream factor caspase-1 (Fig. 10a–c; Figs. S7–S9) (all P < 0.05). Following the specific silencing of FFAR2/3, the capacity of NaA and NaB to mitigate the NLRP3 signaling pathway was markedly attenuated. The co-silencing of both FFAR2 and FFAR3 resulted in an even more pronounced reduction in the influence exerted by NaA and NaB on the NLRP3 signaling. These findings further substantiated that FFAR2 and FFAR3 function as pivotal receptors through which NaA and NaB modulated the NLRP3 signaling pathway. ELISA results further demonstrated that the levels of downstream inflammatory cytokines IL-1β and IL-6 were reduced in the NaA group and the NaB group compared to the LPS group (all P < 0.05) (Fig. S6).

In summary, the in vitro experiments using BMDMs revealed that the potential functional metabolites of B. coprocola—NaA and NaB—can modulate macrophage polarization and NLRP3 signaling pathways, with NaB exhibiting a more comprehensive effect. In addition, FFAR2 and FFAR3 were confirmed as key receptors through which NaA and NaB influence macrophage polarization.

The animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC) of ShanghaiTech University (IACUC No: 20220802001).

Not applicable.

Shengdi Chen serves as the Editor-in-Chief, and Yuyan Tan serves as the Managing Editor of the Journal. Appropriate measures were taken to ensure a fair and unbiased review process.

The 16S rRNA sequencing data have been deposited in the NCBI BioProject database https://www.ncbi.nlm.nih.gov/sra/PRJNA1267990↗. Other data relevant to the study are included in the article or uploaded as supplementary files. The data are available from the corresponding author on reasonable request.