L. mucosae WMU007 was deposited in the China General Microbiological Culture Collection Centre (CGMCC). The strain was cultured in MRS broth (HB0384-1, HopeBio, Qingdao, China) under aerobic conditions at 37 °C for 24 h. Bacterial cells were collected by centrifugation (5000 rpm) and subsequently suspended in phosphate-buffered saline (PBS, ST477, Beyotime, Shanghai, China) to obtain a final density of 1 $\times$ 10⁹ CFU/mL. L. mucosae WMU007 was cultured in MRS broth under aerobic conditions at 37 °C for 24 h, and the supernatant was filtered through a 0.22 µm membrane and diluted to 20% with sterile PBS, which was then used for the cell experiments.

All experimental animals, including 6-month-old male APP/PS1 transgenic mice and C57BL/6J mice, were obtained from the Hangzhou Ziyuan Laboratory Animal Technology Co., Ltd. (Hangzhou, Zhejiang, China). The mice were housed under controlled environmental conditions, including an ambient temperature of 20–22 °C and a relative humidity maintained at approximately 50%. A 12-hour light/dark cycle was applied, with illumination from 08:00 to 20:00, and animals had ad libitum access to standard chow and drinking water. The mice were randomly divided into three groups: the WT group, the APP/PS1 group, and the L. mucosae WMU007-treated group (APP/PS1 + WMU007 group). Mice in the APP/PS1 + WMU007 group were intragastrically administered L. mucosae WMU007 for 4 weeks, while mice in the WT and APP/PS1 groups received an equal volume of PBS as a control. All mice were gavaged with 8 mg arachidonic acid (A131025, Aladdin, Shanghai, China) before L. mucosae WMU007 or PBS treatment. Then, all mice underwent behavioral evaluation, such as the Nesting behavior test, novel object recognition test (NORT), and open field test (OFT).

The NORT was applied to assess the memory and cognitive abilities of mice. Two identical cuboids were symmetrically placed (5 cm from the edges of the field) in an open field with a bottom side length of 25 cm, a height of 25 cm, and each mouse was permitted to explore the open field without restriction for 5 min. The next day, one of the cuboids was replaced with a cylinder of the same material with a bottom diameter of 3 cm, a height of 6 cm, and each mouse was allowed to explore freely for 5 min. After the mouse completed the test, the equipment was wiped with 75% alcohol to eliminate odor interference. Then, the time of the mouse exploring the cylinder (defined as TN) and the cuboid (defined as TF) were calculated. Mice sniffing and lying on objects were defined as exploration; the discriminant index (DI) = (TN – TF) / (TN + TF) $\times$ 100% was used for data analysis.

Nesting behavior, which is closely associated with hippocampal function in rodents, was assessed using a nesting test. Thirty squares of paper (each measuring 5 cm $\times$ 5 cm) were distributed throughout the home cage of each mouse. Animals were allowed to move freely within the cage, and nest construction was documented after 24 h. The nesting performance was scored based on the scoring criteria as follows: (1) no visible nest; (2) no obvious nest, paper occupying more than 1/3 of the cage bottom; (3) obvious nest, paper accounting for 1/4 – 1/3 of the cage bottom; (4) obvious nest, paper occupying less than 1/4 of the cage bottom.

The OFT was carried out to measure the exploratory behavior and curiosity of mice towards the new environment. Each mouse was placed in an open field with a side length of 20 cm at the bottom and a height of 30 cm, and permitted to explore freely for 5 min. The active time and rest time of mice were quantified using the Digbehv Animal Behavior Analysis System (version 2.0, Shanghai Jiliang Software Technology Co., Ltd., Shanghai, China). At the end of each test, the open-field apparatus was cleaned with 75% ethanol and allowed to air-dry to prevent potential olfactory cues from influencing subsequent tests.

After behavioral testing, mice were deeply anesthetized by intraperitoneal injection of pentobarbital sodium (P3761, Sigma-Aldrich, St. Louis, MO, USA) (50 mg/kg). Adequate anesthesia was confirmed by the absence of the pedal withdrawal reflex. The mice were then euthanized by cervical dislocation, and brain tissues were quickly collected. The brain tissue was quickly removed and thoroughly homogenized in a RIPA Lysis Buffer (P0013B, Beyotime, Shanghai, China) mixed with PMSF and Protein Phosphatase Inhibitor (P1260, Solarbio, Beijing, China), centrifuged at 4 °C at 12,000 rpm for 30 min, and the supernatant was removed. The total protein concentration was measured by the Enhanced BCA Protein Assay Kit (P0010, Beyotime, Shanghai, China), and the sample was denatured at 95 °C for 5 min with the SDS-PAGE Sample Loading Buffer (P0015L, Beyotime, Shanghai, China). The 40 µg of protein from each sample were added to 10% SDS-PAGE gel system for electrophoretic separation and transferred to the PVDF membrane. Then, the membrane was sealed in 5% skim milk at room temperature for 1 h, and incubated with the following primary antibodies at 4 °C overnight: A$\beta$_1-42 (1:1000, #14974, Cell Signaling Technology, Danvers, MA, USA), phospho-Tau231 (1:1000, BS4252, Bioworld, Bloomington, MN, USA), Tau (1:1000, BS1358, Bioworld), SOD2 (1:1000, BS6734, Bioworld), CB2 (1:1000, BS72034, Bioworld), Nrf2 (1:1000, BS1258, Bioworld), GPX4 (1:1000, BS7323, Bioworld), HO-1 (1:1000, BS6626, Bioworld), Gapdh (1:5000, AP0063, Bioworld), $\beta$-actin (1:1000, AP0060, Bioworld). Then, the membrane was uniformly covered with Super-sensitive ECL chemiluminescent substrate (BL520B, Biosharp, Hefei, Anhui, China). The grayscale value analysis was performed with ImageJ software (version 1.54, National Institutes of Health, Bethesda, MD, USA). The original images were in the Supplementary Materials.

The brain tissue of mice was fixed in 4% PFA (P0099, Beyotime), dehydrated by gradient alcohol, cleared in xylene (Changshu Hongsheng Fine Chemical Co., Ltd., Changshu, Jiangsu, China), and embedded after wax immersion. The wax block was cut into pieces of 5 µm, heated at 65 °C for 2 h, dewaxed in xylene for 30 min, and rehydrated through a graded ethanol series. Then, the sections were repaired with the antigen via the Hotfix method, soaked in sodium citrate buffer, heated in a pressure cooker for 5 min, left for 10 min, and applied for subsequent staining until cooled to room temperature. Next, the sections were permeabilized with 0.3% Triton for 15 min, blocked in 10% BSA for 1 h, and incubated overnight at 4 °C with primary antibodies, including NeuN (1:200, HA601111, HUABIO, Hangzhou, China), A$\beta$_1-42 (1:400, #14974, Cell Signaling Technology) and SOD2 (1:200, BS6734, Bioworld). Subsequently, the pieces were equilibrated at 37 °C for 30 min, followed by incubation with the corresponding secondary antibody for 1 h under light-protected conditions. The samples were then rinsed with PBS for 3 $\times$ 5 min. Nuclear counterstaining was performed using DAPI staining solution (C1006, Beyotime) for 5 min. After extensive PBS washes, the sections were mounted using an antifade mounting medium (P0126, Beyotime). Fluorescent signals of the target proteins were visualized using a fluorescence microscope (DM4000 B LED, Leica Microsystems CMS GmbH, Wetzlar, Germany).

N2a cells were obtained from the National Collection of Authenticated Cell Cultures (Shanghai, China), which provides STR-authenticated cell lines. All cell lines were routinely tested for mycoplasma contamination using a PCR-based detection kit, and all results were negative. APP/SWE cells were generated by transient transfection of N2a cells at 60% confluency with 0.5 µg APPswe695 plasmid using PEI Transfection Reagent (HY-W250110, MCE, Princeton, NJ, USA) and were pretreated with DMEM (C0891, Beyotime) medium containing 20% L. mucosae WMU007 fermentation for 8 h.

The N2a cells were cultured in serum-free medium and exposed to the DCFH-DA probe (S1105S, Beyotime) at 37 °C for 20 min. Subsequently, nuclear counterstaining was carried out using Hoechst dye (C1011, Beyotime), and intracellular fluorescence signals were captured and quantified under a fluorescence microscope.

Total RNA was isolated from mouse cerebral cortex samples with TRIzol reagent (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) following standard protocols. The RNA’s concentration, purity, and integrity were then assessed. Procedures including mRNA enrichment, RNA fragmentation, reverse transcription, library construction, and high-throughput sequencing were completed by Shanghai Majorbio Bio-pharm Biotechnology Co., Ltd. (Shanghai, China). To ensure the reliability of downstream bioinformatics analyses, raw sequencing reads were subjected to quality control using fastp (OpenGene, GitHub, San Francisco, CA, USA). Clean reads with high quality were then de novo assembled using the Trinity platform. The preliminary assembled sequences underwent refinement and filtering via TransRate (http://hibberdlab.com/transrate/). The completeness of the assembled transcripts was further evaluated with BUSCO (Benchmarking Universal Single-Copy Orthologs). Differential gene expression analysis was performed using the DESeq2 package (Bioconductor, Seattle, Washington, USA), and genes meeting the criteria of $|$FoldChange$|$ $>$0.3 and p-value 0.05 were defined as differentially expressed. Subsequently, GO and KEGG enrichment analyses were performed using the clusterProfiler package (version 4.16.0, Bioconductor), while gene set enrichment analysis (GSEA) was performed using the fgsea package (version 1.34.0, Bioconductor). Visualization of the results was accomplished using ggplot2 (version 3.5.2, CRAN, Vienna, Austria), pheatmap (version 1.0.13, CRAN), and other relevant packages mentioned above.

Real-time quantitative PCR was performed using ChamQ Universal SYBR qPCR Master Mix (Q711, Vazyme Biotech Co., Ltd., Nanjing, Jiangsu, China) according to the manufacturer’s instructions. Relative gene expression was normalized to Actin and calculated using the 2^{-Δ⁢Δ⁢CT} method. The primer sequences were as follows: CB2-F, 5^′-ACGGTGGCTTGGAGTTCAAC-3^′; CB2-R, 5^′-GCCGGGAGGACAGGATAAT-3^′.

Data were represented as mean $\pm$ SEM. All statistical evaluations were conducted using SPSS software (version 22.0, IBM SPSS Statistics, Chicago, IL, USA). Group differences were assessed by the t-test or one-way ANOVA. Statistical significance was defined as p 0.05.

To investigate the effects of L. mucosae WMU007 treatment on cognitive dysfunction in APP/PS1 mice, a series of behavioral assessments was conducted. In NORT, the discrimination index of APP/PS1 mice was significantly lower than that of WT mice, while WMU007 treatment significantly increased the discrimination index compared with the APP/PS1 group (p 0.05, Fig. 1A). In the Nesting behavior test, the nesting score of the WMU007-treated group was significantly higher than that of the APP/PS1 group (p 0.01, Fig. 1B). In OFT, the active time of the WMU007 group was markedly longer than the APP/PS1 group (p 0.05, Fig. 1C). In contrast, the rest time in the WMU007 group was significantly shorter than the APP/PS1 group (p 0.05, Fig. 1D). The representative images of the activity trajectory of three groups are displayed in Fig. 1E. These findings suggested that L. mucosae WMU007 treatment significantly improved cognitive impairment.

Fig. 1.

L. mucosae WMU007 treatment improved cognitive dysfunction in APP/PS1 mice (n = 5 mice per group). (A) The discriminant index in NORT. (B) The nesting score in the Nesting Behavior Test. (C) The active time in the open field test. (D) The rest time in the open field test. (E) Representative images of activity trajectory in the open field test. (F) Representative images of NeuN staining (green) in the cortex and hippocampus (CA1, DG), counterstained with DAPI (blue). Magnification, 200$\times$. Scale bars, 100 µm. (G–I) Quantification of relative NeuN⁺ cell fraction in the cortex (G), CA1 (H), and DG (I) (n = 3). Data are shown as mean $\pm$ SEM. p-values were determined by one-way ANOVA. ^*: p 0.05, ^**: p 0.01.

To assess the effects of L. mucosae WMU007 on neuropathological changes, NeuN staining, A$\beta$ deposit, and tau phosphorylation were detected. The number of NeuN-positive neurons in the cortex, CA1, and DG of the L. mucosae WMU007-treated group was increased compared with the APP/PS1 group (Fig. 1F–I). Immunofluorescence analysis showed that A$\beta$ plaque was markedly decreased in the L. mucosae WMU007-treated group compared with APP/PS1 mice (Fig. 2A,B). In addition, the level of Tau phosphorylation in L. mucosae WMU007-treated APP/PS1 mice was significantly decreased compared with APP/PS1 mice (Fig. 2C–F). In the Western blot, the level of A$\beta$ in WMU007-treated group was significantly decreased compared with the APP/PS1 group (p 0.05, Fig. 2G,H). The ratio of p-Tau231/Tau in the WMU007-treated group was remarkably reduced compared with that in APP/PS1 mice (p 0.05, Fig. 2G–I). These findings suggested that L. mucosae WMU007 treatment improved neuropathological changes of AD.

Fig. 2.

L. mucosae WMU007 treatment improved neuronal damage and decreased A$\beta$ deposition and Tau hyperphosphorylation in APP/PS1 mice. (A) Representative immunofluorescence images of A$\beta$ (red) in hippocampal regions from the APP/PS1 group and APP/PS1 + WMU007 group, counterstained with DAPI (blue). Magnification, 50$\times$. Scale bars, 400 µm. (B) Quantification of relative A$\beta$ plaque (n = 3). (C) Representative immunohistochemical images of Tau in the cortex and hippocampus (CA1, DG), counterstained with DAPI (blue). Magnification, 200$\times$. Scale bars, 100 µm. (D–F) Quantification of relative p-Tau231-positive area in the cortex (D), CA1 (E), and DG (F) (n = 3). (G) Representative western blot images of A$\beta$, p-Tau231, and Tau. (H,I) Quantitative analysis of A$\beta$, p-Tau231, and Tau expression (n = 4). The ratio of A$\beta$/Gapdh and p-Tau231/Tau in the APP/PS1 group were used as the reference value. Data are shown as mean $\pm$ SEM. p-values were determined by the t-test. ^*p 0.05 vs. APP/PS1 group, ^**p 0.01 vs. APP/PS1 group.

To elucidate the molecular mechanism of L. mucosae WMU007 underlying the decreased A$\beta$ deposition and Tau hyperphosphorylation, transcriptomic profiling was conducted to characterize gene expression changes between WMU007-treated and untreated groups. The analysis yielded a total of 231 upregulated genes and 365 downregulated genes ($|$log2FC$|$ $>$0.3, p 0.05). Differentially expressed genes between the two groups were visualized using a volcano plot (Fig. 3A), while hierarchical clustering of the top 30 DEGs was illustrated by a heatmap (Fig. 3B). To explore the functional implications of the L. mucosae WMU007 intervention, the GO enrichment analysis was conducted based on the identified DEGs. Biological Process annotation revealed that WMU007 predominantly modulated pathways related to oxidative stress (Fig. 3C). To further confirm the effect of L. mucosae WMU007 on oxidative stress, we performed an analysis of ROS, SOD2, and GPX4. The fluorescence staining showed that the level of ROS was decreased in the L. mucosae-treated group compared with the Con group in N2a cells (p 0.01, Fig. 3D,E). Western blot analysis showed that the SOD2 and GPX4 levels were significantly increased in L. mucosae-treated N2a cells (p 0.01, Fig. 3F–H). These findings suggested that L. mucosae WMU007 effectively decreased oxidative stress.

Fig. 3.

L. mucosae WMU007 treatment decreased oxidative stress. (A) Volcano plot showing the distribution of gene expression level differences. p-value 0.05 was screened as differentially expressed genes. (B) Heatmap showing differentially expressed genes. (C) GO analysis showing the biological functions of DEGs. Significant changes in oxidative stress-related pathways were found. (D) Representative images of ROS staining (green) of N2a cells in the Con group and the Fer group, counterstained with Hoechst (blue). Magnification, 100$\times$. Scale bars, 200 µm. Fer, L. mucosae WMU007 fermentation. (E) Quantification of the relative intensity of ROS-positive staining (n = 6). (F) Representative western blot images of SOD2 and GPX4. (G,H) Quantitative analysis of SOD2 and GPX4 expression (n = 3). The ratio of SOD2/Gapdh and GPX4/Gapdh in Con group was used as the reference value. Data are shown as mean $\pm$ SEM. p-values were determined by the t-test. ^**p 0.01 vs. Con group. ROS, reactive oxygen species; SOD2, superoxide dismutase 2; GPX4, glutathione peroxidase 4; BP, biological process; CC, cellular component; MF, molecular function.

To further investigate the mechanism of L. mucosae WMU007 on inhibiting oxidative stress, KEGG enrichment analysis was conducted based on genes exhibiting altered expression in the RNA-Seq results. KEGG analysis revealed significant differences in the Retrograde endocannabinoid signaling among the DEGs (Fig. 4A). To gain a more comprehensive understanding of the changes in the Retrograde endocannabinoid signaling, we conducted GSEA enrichment analysis. A normalized enrichment score (NES) of 1.73 indicates that L. mucosae treatment can significantly activate the aforementioned pathway (Fig. 4B). As shown in Fig. 4C, the qPCR analysis revealed a significant elevation of CB2transcript abundance in mice receiving L. mucosae WMU007 relative to controls (p 0.05). To further investigate the effect of L. mucosae WMU007 on the CB2 signaling pathway, we examined the expression levels of key proteins associated with this pathway, including CB2, Nrf2, and HO-1. The Western blot analysis showed that the levels of CB2, Nrf2, and HO-1 were markedly increased after L. mucosae WMU007 treatment compared with the Con group (p 0.05, Fig. 4D,E). Collectively, these findings indicate that L. mucosae supplementation attenuates oxidative stress by activating the CB2 signaling pathway in vivo.

Fig. 4.

L. mucosae WMU007 inhibited oxidative stress via activating the CB2 pathway. (A) KEGG pathway enrichment showing the Retrograde endocannabinoid signaling was significantly enriched in DEGs. p-value 0.05 was considered statistically significant. (B) GSEA plot for the “Retrograde Endocannabinoid Signaling”. (C) The relative expression levels of CB2 mRNA between the Con group and the Fer group (n = 3), Fer, L. mucosae WMU007 fermentation. (D) Representative western blot images of CB2, Nrf2, and HO-1. (E) Quantitative analysis of CB2, Nrf2, and HO-1 expression (n = 3). The ratio of CB2/Gapdh, Nrf2/Gapdh, and HO-1/Gapdh in the Con group was used as the reference value. Data are shown as mean $\pm$ SEM. p-values were determined by the t-test. ^*p 0.05 vs. Con group. CB2, cannabinoid receptor type 2; Nrf2, nuclear factor erythroid 2-related factor 2; HO-1, heme oxygenase-1.

To confirm the role of L. mucosae WMU007 in inhibiting oxidative stress through CB2 receptor activation, APP/PS1 mice were treated with L. mucosae WMU007. Immunofluorescence staining revealed a significant increase in SOD2-positive fluorescence intensity in the cortex, CA1, and DG of the L. mucosae WMU007-treated group compared to the APP/PS1 group (Fig. 5A–D). To further assess antioxidant responses and CB2 signaling, we analyzed protein expression by Western blot. Our results showed that L. mucosae WMU007 treatment significantly increased the levels of CB2, Nrf2, GPX4 and HO-1 compared to the APP/PS1 group (CB2: p 0.05, Nrf2: p 0.01, GPX4: p 0.05, HO-1: p 0.05, Fig. 5E–I). These findings suggested that the potential mechanism by which L. mucosae WMU007 alleviated oxidative stress might be related to regulating CB2 signaling.

Fig. 5.

L. mucosae WMU007 treatment reduces oxidative stress in APP/PS1 mice via activating the CB2 pathway. (A) Representative immunofluorescence images of SOD2 (red) in the cortex and hippocampal (CA1 and DG) regions, counterstained with DAPI (blue). Magnification, 200$\times$. Scale bars, 100 µm. (B–D) Quantification of relative SOD2 fluorescence intensity in the cortex (B), CA1 (C), and DG (D) (n = 3). (E) Representative western blot images of CB2, Nrf2, GPX4, and HO-1. (F–I) Quantitative analysis of CB2, Nrf2, GPX4 and HO-1 expression (n = 4). The ratio of CB2/$\beta$-actin, Nrf2/$\beta$-actin, GPX4/$\beta$-actin, and HO-1/$\beta$-actin in the WT group was used as the reference value. Data are shown as mean $\pm$ SEM. p-values were determined by one-way ANOVA. ^*: p 0.05, ^**: p 0.01.