Akkermansia muciniphila (AKK, ATCC BAA-835) was obtained from Mingzhou Bio (China). Reagents for RNA reverse transcription to cDNA (Cat# 11141ES10, Yeasen) and SYBR Green (Cat# 11201ES08, Yeasen) were utilized. Western blot antibodies included IBA1 (#A30311, Nature Biosciences), pTrkB (#A46276, Nature Biosciences), TrkB (#ab108319, Cell Signaling Technology), BDNF (#A50292, Nature Biosciences), ERK1/2 (#9102, Cell Signaling Technology), pERK1/2 (#9101, Cell Signaling Technology), CREB (#9197, Cell Signaling Technology), pCREB (#9198, Cell Signaling Technology), and β-actin (#4967, Cell Signaling Technology).

Sixty male C57BL/6J mice (Charles River, Germany), weighing 23 ± 3g, were used. All mice were acclimated for 14 days prior to treatment and were age-matched (approximately 10 weeks old). The animals were housed in a controlled environment (21 ± 2°C, 50% relative humidity) with ad libitum access to food and water. Only male C57BL/6J mice were used in this preliminary study due to sex-related differences in hormonal and immune responses (23, 24). The study was approved by the Health Science Center of XJTU Approval for Research Involving Animals (approval number: XJTYAE2024-1396).

Mice were randomly assigned to one of four groups: Control (no treatment), Trauma, Trauma+PBS (100 µL sterile PBS daily), and Trauma+AKK (100 µL AKK bacterial suspension at 2 × 10⁹ CFU/mL daily). The AKK suspension was administered via gavage for three weeks.

Blunt chest trauma was induced as previously described. Oxygen was turned on and the flow rate was adjusted to 0.25 MPa, 1 L/min. The concentration of the anesthetic (isoflurane) was set to 5%, and induction anesthesia was completed in approximately 1 minute. Subsequently, the anesthetic concentration was adjusted to 2%, and a mouse mask was connected for continuous inhalation. Mice were anesthetized with isoflurane gas, and a cylindrical weight (50 ± 2g) was dropped from a height of 300 ± 2mm onto a platform placed on the chest. The impact energy (E) was calculated using the formula E = m × g × h, where m is the weight (kg), g is the acceleration due to gravity (9.8 m/s²), and h is the height (m) (25, 26). Preliminary experiments indicated that a 300mm height and 50g weight (E=0.15 J) induced blunt chest trauma without causing severe complications, such as pneumothorax, hemothorax, cardiac rupture, or immediate death. Following impact, mice exhibited transient shock and rapid breathing and regained motor function after anesthesia recovery. To mitigate behavioral abnormalities caused by pain in surviving mice post-trauma, we administered meloxicam via subcutaneous injection at a dose of 5 mg/kg, with daily dosing for 3 days post-surgery. Additionally, we allowed sufficient time for the mice to recover from the acute pain phase post-surgery (27). To assess the consistency of pain in the mice, we monitored their pain behavior by observing the distance traveled in the open field test. There was no significant difference in the total distance traveled in the open field between trauma-induced mice and normal mice, after which we further assessed the emotional behavior of the mice.

Mice were acclimated to the testing environment for 1 hour prior to behavioral testing to minimize external interference. The testing arena (50cm × 50cm × 40cm) was cleaned with 75% ethanol between each test to remove residual odors. Mice were placed in the center of the arena and allowed to explore for 10 minutes. Behavioral recordings included trajectory, distance traveled, time spent, and speed, which were analyzed using Noldus EthoVision software. Reduced inner area activity and shorter exploration times were indicative of anxiety-like behavior (28).

Mice were acclimated to the testing environment for at least 1 hour prior to the test. The apparatus was cleaned between each test to ensure a scent-free environment. Mice were placed in the center of the maze facing an open arm and allowed to explore for 5 minutes. Behavior was recorded, and trajectories, distance traveled, entries into the open arms, and time spent in the open arms were analyzed using Noldus EthoVision software. Reduced activity in the open arms, indicated by shorter distances and times, was interpreted as anxiety-like behavior (29).

The posterior third of the mouse tail was suspended from a support, with the head positioned 15cm above the table. Behavior was recorded for 5 minutes, and immobility time was analyzed using Smart v3.0 software. Increased immobility time suggested depression-like behavior (30).

Mice were placed in a transparent cylinder (10cm diameter × 18cm height) filled with water (2/3 volume at 25 °C) for 6 minutes. Behavior was recorded during the final 5 minutes, and immobility time was analyzed using the software. Extended immobility times were indicative of depression-like behavior (30).

The test consisted of three phases: habituation, training, and testing. During the habituation phase, mice explored an open field for 10 minutes on the first day. On the training day, mice were exposed to two identical objects for 10 minutes. In the testing phase, mice were presented with a familiar object (A1) and a novel object (A2) for 10 minutes. Exploration was defined as touching or directing the head within 2cm of the object. All objects and the arena were cleaned with 75% ethanol between tests. The preference index was calculated as the time spent exploring the novel object (T_A2) divided by the total exploration time for both objects (T_A1 + T_A2). Normal cognitive function was indicated by a preference for the novel object (29).

The Y-maze test also involved three phases: habituation, training, and testing. During the habituation phase, mice were food-deprived for one day but had free access to water. Food was placed in all three arms of the Y-maze, and mice were allowed to explore for 10 minutes twice a day over two days. During training, mice were food-deprived for one day, with food placed in one arm (the food arm) and the other two arms left empty. Mice explored for 10 minutes twice a day for two to three days. In the test phase, mice were food-deprived for one day, and all three arms were empty. Mice were placed in the start arm (the non-food arm), and the number of entries and time spent in the food arm were recorded. Time spent in the food arm indicated learning and memory ability, with longer times suggesting better memory. Longer times spent in the error arm (other than the start or food arms) indicated poorer memory (31).

Hematoxylin-Eosin (HE) Staining: Tissue was fixed at room temperature for 24 hours, dehydrated, embedded in paraffin, and sectioned at 4 µm thickness. Sections were stained with hematoxylin and eosin and observed under a microscope.

Immunofluorescence (IF) Staining: Tissue sections were dewaxed, rehydrated, and subjected to antigen retrieval using citrate buffer. Endogenous peroxidase activity was blocked with 3% H₂O₂. Sections were incubated with primary antibodies overnight at 4°C, followed by incubation with secondary antibodies for 1 hour at room temperature. Sections were mounted with an anti-fade mounting medium and observed under a fluorescence microscope. Fluorescence intensity was quantified using ImageJ software.

16S rRNA gene sequencing was employed to analyze the gut microbiota of traumatized mice. The V3-V4 regions of the 16S rDNA were amplified via PCR, purified on a 2% agarose gel, and quantified. The fragments were sequenced on the Illumina platform following the manufacturer’s paired-end read protocol (PE250). Representative sequences for each operational taxonomic unit (OTU) were classified using the RDP Classifier software with a Bayesian algorithm at a confidence level of 0.8. Bioinformatics analysis was performed using the Omicsmart platform.

Metabolomic analysis was performed using an Agilent 1290 Infinity LC system coupled with an AB Triple TOF 6600 mass spectrometer. Chromatographic separation was achieved on an HSS T3 column. For positive electrospray ionization (ESI) mode, the mobile phase consisted of water with 0.1% formic acid (solvent A) and acetonitrile (solvent B). For negative ESI mode, water containing 0.5 mM ammonium formate (solvent A) and acetonitrile (solvent B) was used. Bioinformatics analysis was performed to identify potential metabolites.

Total RNA was isolated using a Trizol reagent, and RNA concentration was measured. RNA was reverse-transcribed into cDNA according to the manufacturer’s instructions. For qPCR reactions, cDNA, gene-specific primers, and SYBR Green were mixed. Primer sequences are provided in Supplementary Table 1 .

Total protein was extracted from tissues using RIPA lysis buffer, and protein concentration was measured using a BCA assay kit. Proteins were separated by SDS-PAGE and transferred to PVDF membranes. Membranes were incubated with primary antibodies and HRP-conjugated secondary antibodies. Protein bands were visualized using a ChemiDoc imaging system (Bio-Rad, Hercules, CA, USA), and band intensity was analyzed using ImageJ software.

The study adhered to principles of randomization and blinding to ensure objectivity. Results are presented as mean ± standard error (SEM). Statistical analysis was conducted using unpaired t-tests, Welch’s t-tests, or one-way ANOVA with GraphPad Prism version 8. P-values less than 0.05 were considered statistically significant.

Regarding the reproducibility of the blunt chest trauma mouse model, the survival rate of the mice was 69% (with 12 deaths and 27 survivals). Mice that died following trauma either succumbed immediately or within 24 hours post-trauma. Upon autopsy, these deceased mice exhibited substantial hemorrhaging within the thoracic cavity, perforation or rupture of the right atrium, yet no rib fractures or visible damage to abdominal organs were observed. Surviving mice displayed transient respiratory depression immediately after injury, subsequently presenting with rapid breathing, piloerection, lethargy, and reduced activity. These symptoms gradually normalized by day 3 post-trauma. We harvested tissues from mice 3 days post-trauma and observed contusions on the left and right lungs caused by the ribs. Histological examination of the heart and lung tissues of the trauma-induced mice via HE staining revealed extensive pulmonary hemorrhage and edema, with alveoli infiltrated by a large number of inflammatory cells and thickened alveolar septa. In contrast, normal mouse lung tissue exhibited intact alveolar structures with thin and uniform alveolar walls. As for the heart, trauma-induced mice showed unclear myocardial striations, disordered arrangement, vacuolization of myocardial cells, and blurred cell boundaries, while normal mouse heart tissue had neatly arranged myocardial fibers with clear structures ( Supplementary Figure 1 ). These findings indicate that we have successfully induced blunt chest trauma in mice.

To evaluate the effects of trauma on cognitive and emotional behaviors in mice, behavioral tests were conducted to assess learning and memory, anxiety, and depression. In the OFT, the total distance traveled by traumatized mice did not significantly differ from that of controls (P > 0.05, Figure 1A ), suggesting no substantial impact on overall motor function. However, from day 14 to day 60 post-trauma, traumatized mice exhibited significantly reduced distances and time spent in the inner area compared to controls (P<0.05, Figures 1B, C ). In the NORT, traumatized mice showed significantly decreased exploration times and frequencies for the novel object starting at day 14 (P<0.05, Figures 1D, E ), with a lower preference index (P<0.05, Figure 1F ) that persisted through day 60. By day 60, however, the frequency of novel object exploration did not significantly differ from controls, indicating partial recovery of interest in novel objects. In the TST and FST, traumatized mice showed no significant differences in immobility times compared to controls (P > 0.05, Figures 1G, H ). Overall, while traumatized mice exhibited anxiety and cognitive impairments from day 14 onward, no depression-like behaviors were observed.

To explore the impact of trauma on gut microbiota, the relative abundance of fecal microbiota at the genus level was analyzed using 16S rRNA sequencing 14 days post-trauma. Traumatized mice exhibited significant increases in the relative abundance of Lactobacillus, Mycoplasma, Allobaculum, and Bacteroides compared to controls (P<0.05). In contrast, Lachnospiraceae NK4A136 group, Akkermansia, and Dubosiella were significantly decreased (P<0.05), while Parasutterella remained unchanged (P > 0.05, Figure 2 ). Akk, a member of the genus Akkermansia, has been shown to restore gut microbiota balance and regulate brain function (32). However, whether AKK can modulate trauma-induced gut microbiota dysbiosis and cognitive impairments remains unclear.

To investigate whether AKK treatment could alleviate behavioral abnormalities in traumatized mice, this study administered AKK and assessed behavioral changes ( Figure 3A ). In the OFT, AKK-treated traumatized mice (Trauma+AKK) showed no significant difference in total distance traveled compared to Trauma+PBS. However, they exhibited significantly increased distances and time spent in the inner area (P<0.05, Figure 3B ). In the EPM test, traumatized mice had reduced exploration times and entries into the open arms compared to controls. AKK treatment significantly increased these parameters in traumatized mice (P<0.05, Figure 3C ). In the NORT, AKK-treated traumatized mice showed no significant difference in total distance traveled, but had significantly increased exploration frequencies and preference indices for the novel object (P<0.05, Figure 3D ). In the Y-maze test, the number of entries and the time spent in the wrong arm were significantly increased in the trauma-induced mice compared with the control group (P <0.001, Figure 3E ). However, AKK treatment significantly reduced the number of entries and the time spent in the wrong arm in the trauma-induced mice compared with the Trauma + PBS group (P <0.05, Figure 3E ). AKK treatment significantly improved cognitive function-related parameters. Overall, AKK treatment significantly improved anxiety-like behaviors and cognitive impairments in traumatized mice.

Growing evidence indicates that gut microbiota influences brain function and cognitive performance through the gut-brain axis, primarily via metabolites and neurotransmitters (33). Dysbiosis can disrupt intestinal permeability, allowing bacteria and toxins to enter the bloodstream, thereby inducing inflammation that impairs cognitive and emotional functions. To explore whether AKK treatment modulates trauma-induced changes in gut microbiota, 16S rDNA sequencing was performed to assess microbial composition and abundance in traumatized mice. Fourteen samples from two groups (n = 7 per group) were sequenced to generate V3-V4 16S rRNA gene profiles. Beta-diversity was evaluated using PCA, PCoA, and NMDS to determine microbial community similarity. PCA revealed distinct group separation ( Figure 4A , left), with principal components PC1 and PC2 accounting for 59.76% and 15.39% of the variation, respectively. PCoA showed that PC1 and PC2 explained 42.85% and 15.56% of the variation ( Figure 4A , middle). NMDS analysis, with stress = 0.033 (stress < 0.05), confirmed significant differences between samples ( Figure 4A , right). AKK treatment was found to significantly alter the gut microbiota composition, with increases in Lactobacillus and Akkermansia at the family, genus, and species levels, while Bacteroides decreased ( Figure 4B ). To pinpoint bacterial taxa linked to AKK treatment in traumatized mice, Welch’s t-tests identified the top 10 core taxa explaining group differences. As shown in Figure 4C , AKK-treated traumatized mice (vs. Trauma+PBS) exhibited lower levels of Bacteroides sartorii (P=0.004), Bacteroides acidifaciens (P=0.004), and Muribaculum intestinale (P=0.046), while Akkermansia muciniphila (P=0.017) and Lactobacillus murinus (P=0.001) increased significantly. A heatmap of the top 20 species at the family level is provided in Figure 4D . Functional abundance heatmaps indicated associations with carbohydrate metabolism, amino acid metabolism, cofactor and vitamin metabolism, terpenoid and polyketone metabolism, other amino acid metabolism, and lipid metabolism ( Figure 4E ). Further functional predictions using PICRUSt2 revealed that trauma altered various biological processes in gut microbiota, particularly metabolism, genetic information processing, and cellular processes.

Given the ability of gut microbiota to influence the host through metabolites, metabolomics analysis was conducted on traumatized mice. OPLS-DA modeling (T score = 14.9%) and permutation tests confirmed the model’s reliability and predictability ( Figure 5A ). Bar and volcano plots displayed upregulated and downregulated metabolites based on statistical thresholds (P<0.05, |log2FC| > 1, Figure 5B ). AKK treatment in traumatized mice (vs. Trauma+PBS) revealed 70 differential metabolites, with 29 upregulated and 41 downregulated. A heatmap illustrating the top 20 most differentially expressed metabolites between groups was also constructed ( Figure 5C ). Variable Importance in Projection (VIP) analysis identified key metabolites, such as a significant decrease in 2,4-diaminobenzenesulfonic acid and 3-methyl-2-oxopentanoate, and an increase in chelidonic acid, 1-stearoyl-2-hydroxy-sn-glycero-3-phosphoethanolamine, and 4,5-dihydro-4,5-dioxo-1h-pyrrolo[2,3-f]quinoline-2,7,9-tricarboxylic acid in AKK-treated traumatized mice ( Figure 5D ). KEGG enrichment analysis indicated that these differential metabolites were involved in alanine metabolism, alanine, aspartate, and glutamate metabolism, oxidative phosphorylation, histidine biosynthesis, purine-derived alkaloid biosynthesis, ABC transporters, and other biological processes ( Figure 5E ).

Subsequently, the correlation between gut microbiota and metabolites was explored. The O2PLS model loading plot ( Figure 6A ) demonstrated a strong association between the microbiota and metabolites. Further genus-level heatmap analysis revealed specific correlations between microbiota and metabolites. For example, Escherichia-Shigella exhibited a positive correlation with gamma- muricholic acid, 19(r)-hydroxyprostaglandin, coumaphos-o-analog, phosphocholine, and quinclorac, but a negative correlation with isopentedrone. Odoribacter showed a positive correlation with D-ribulose 1,5-bisphosphate but a negative correlation with DL-tryptophan ( Figure 6B ).

To evaluate the neuroprotective effects of AKK, neuroinflammation in the hippocampus of traumatized mice was assessed through HE staining and cytokine expression analysis. The hippocampus, a key region in the limbic system, comprises the Cornu Ammonis (CA) 1, CA2, CA3, CA4, and Dentate Gyrus (DG), all of which are integral to memory formation, emotional regulation, and cognitive function. HE staining revealed that traumatized mice (vs. Control) displayed sparsely and disordered pyramidal neurons arrangement in the CA1 region, along with neuronal nuclear condensation and necrosis in the CA3 and DG regions. AKK treatment, however, significantly alleviated these pathological alterations in the hippocampus ( Figure 7A ). QPCR analysis indicated significantly increased levels of pro-inflammatory cytokines (Il1β, Il6, Il12, Tnfα) and decreased expression of the anti-inflammatory cytokine Il10 and Tgfβ in the hippocampus of traumatized mice (P<0.05, Figure 7B ). AKK treatment significantly attenuated the release of neuroinflammatory cytokines. In summary, AKK treatment substantially mitigated neuronal damage and neuroinflammation in traumatized mice.

Microbiota-derived metabolites, including SCFAs and neurotransmitters (e.g., neurotrophic factors, serotonin), can modulate neuroinflammation and microglial function by influencing blood-brain barrier permeability. To investigate the mechanisms by which AKK ameliorates neuroinflammation and cognitive impairments in traumatized mice, immunofluorescence staining for microglial marker IBA1 and BDNF was performed in the hippocampus ( Figure 8A ). In control mice, microglia exhibited a highly branched structure with three to four branches. Conversely, traumatized mice displayed enlarged microglial cell bodies with shorter and fewer branches. Quantitative analysis of IBA1 fluorescence intensity revealed a significant increase in IBA1-positive staining in the hippocampus of traumatized mice, indicating microglial activation, which was notably reduced following AKK treatment ( Figure 8B ). Additionally, BDNF fluorescence intensity was decreased in traumatized mice, an effect reversed by AKK treatment ( Figure 8C ). Western blot analysis further demonstrated that AKK treatment significantly reduced IBA1 expression in traumatized mice ( Figure 8D ). Furthermore, traumatized mice exhibited significant reductions in BDNF and phosphorylated Tropomyosin Receptor Kinase B (pTrkB) protein levels (P<0.05). Additionally, the levels of their downstream molecules pCREB and pERK were also significantly decreased (P<0.01). These reductions were restored by AKK treatment, while total TrkB, ERK, and CREB proteins remained unchanged (P > 0.05, Figure 8E ). These findings suggest that AKK treatment inhibits microglial overactivation and promotes the BDNF/TrkB signaling pathway in the hippocampus of traumatized mice.