SPF grade male C57BL/6 mice (8 weeks old, 26 ± 2 g) were obtained from the Xi’an branch of Chongqing Tengxin Biotechnology Co., Ltd. (Xian, China). The experimental procedures were reviewed and authorized by the Ethics Committee of the 940th Hospital of the Joint Logistic Support Force (Approval No.: 2022KYLL021). All animals were kept under SPF conditions at 25 °C with a 12-h light/dark photo period. Following a 7-day adaptation period, animals were allocated to Control (Con), Hindlimb unloading (HU), and Hindlimb unloading with LGG treatment (HU + LGG) group. All analyses were conducted on samples collected from five randomly selected mice in each group.

The HU model was developed following the protocol outlined by Qi et al. (29). Medical adhesive tape was wrapped around two-thirds of the tail length of each mouse, and a rotating tape was applied and fixed to maintain a head-down tilt position with hindlimbs suspended without weight-bearing, while forelimbs retained free movement. A schematic overview of the experimental workflow was provided in Supplementary Figure S1. After 1 week of environmental adaptation, mice in the HU and HU + LGG groups were subjected to HU for one additional week as a model acclimation phase. Both groups underwent continuous HU for 4 weeks. Simultaneously, the HU + LGG group received oral gavage of LGG suspension (1 × 10⁹ CFU per mouse), five times per week, for the entire 4-week period. The dose was selected based on previously published studies in which this concentration was shown to effectively modulate the gut microbiota and host physiology in mice without causing adverse effects (13). Due to experimental limitations, only a single dose was evaluated in the present study. Mice in the HU group were administered an equivalent volume of tap water, while those in the Con group remained untreated with ad libitum access to drinking water. After 4 weeks of intervention, all animals were humanely euthanized by rapid cervical dislocation performed by experienced personnel in strict accordance with the American Veterinary Medical Association (AVMA) Guidelines for the Euthanasia of Animals (2020). Death was confirmed by the absence of respiration, heartbeat, and pupillary reflex, after which samples were promptly collected for further analysis.

Mouse fecal samples were collected using sterile disposable forceps and immediately transferred to −80 °C for storage. Microbial genomic DNA was isolated through the FastPure Stool DNA Isolation Kit (Majorbio, Shanghai, China). DNA integrity was verified by electrophoresis on a 1% agarose gel. The resulting DNA served as a template for amplifying the gene using primers 338F (ACTCCTACGGAGGCAGCAG) and 806R (GGACTACHVGGGTWTCTAAT). PCR products were examined by electrophoresis on a 1.5% agarose gel. Subsequently, the PCR products were purified using a QIAquick Gel Extraction Kit (Qiagen, Hilden, Germany). Then, the concentration of the purified PCR products was measured using a NanoDrop (Thermo Fisher Scientific, Waltham, MA, United States). Sequencing was conducted on the Illumina NextSeq 2000PE300 platform (Shanghai Meiji Biological Pharmaceutical Technology Co., Ltd., Shanghai, China). Raw fastq data will be made available upon request to the corresponding author.

The SCFAs standard solutions were prepared. Under sterile conditions, small intestinal samples and serum were obtained. A 40 mg aliquot of intestinal tissue or 150 μL of serum was homogenized and mixed with an extraction solution (methanol:water = 4:1). The samples were ground using a cryogenic grinder, followed by ultrasonication under low temperature. The mixture was centrifuged at 13,000 rcf for 15 min at 4 °C. The supernatant was transferred into a new tube, followed by the addition of 20 μL of 3N HCl and 20 μL of EDTA HCl solution. The samples were incubated at 40 °C for 30 min, then diluted with an acetonitrile-water mixture to a final volume of 1,000 μL for analysis. After analysis, raw data were imported into the AB Sciex quantitative software system (AB Sciex, Framingham, MA, United States) for generating linear regression standard curves and calculating the sample concentrations.

Expression level of OCN, OPG, BALP, CTX, PINP, TNF-α, IL-1β, IL-10, and LPS were assessed via ELISA (Thermo Fisher Scientific, Waltham, MA, United States). Each group included five replicates, and each sample was measured in at least three technical replicates.

Total protein was isolated from colonic tissues and bone marrow, followed by SDS-PAGE. Membrane transfer was performed using a semi-dry method. Protein expression levels of ZO-1 and Occludin in colon tissues and TNF-α, IL-1β, and IL-10 in bone marrow were detected by WB. GAPDH was used as the internal reference. Primary antibodies, GAPDH (ab9485), ZO-1 (ab276131), Occludin (ab216327), TNF-α (ab183218), IL-1β (ab283818), and IL-10 (ab133575) were purchased from Abcam (Cambridge, United Kingdom). The secondary antibody used was goat anti-rabbit IgG (ab97051, Abcam, Cambridge, United Kingdom).

Bacterial pellets from fecal samples were obtained according to previously described methods (30). Genomic DNA (gDNA) of LGG was isolated using the NucleoSpin kit (Takara, Kusatsu, Shiga, Japan). Quantitative polymerase chain reaction (qPCR) was performed to determine the abundance of LGG. Ct values of the samples were compared with a standard curve generated from pure cultures of LGG or fecal samples spiked with known concentrations of LGG, as previously described (30). The primers were listed in Supplementary Table S1.

Primers for the Foxp3 and TLR4 gene in Mus musculus were designed based on its CDS sequence. The relative mRNA levels was carried out by CFX96 Touch Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA, United States). GAPDH was served as the internal control. Gene relative expression was assessed via the 2^−ΔΔCt approach. Each sample was assayed in at least three replicates. The primers were listed in Supplementary Table S2.

The pre-section steps are similar to H&E staining, so only the post-section steps are covered. Deparaffinize and dehydrate the colon sections, perform antigen retrieval, block endogenous peroxidase activity, and incubate with rabbit serum at ambient temperature for 30 min. After rinsing, the sections were incubated overnight at 4 °C with ZO-1 (1:500) and Occludin (1:500) antibodies. Apply the secondary antibody (HRP-labeled), corresponding to the species of the primary antibody, and incubate at room temperature for 50 min. HRP-conjugated secondary antibodies corresponding to the host species of the primaries were applied and incubated at room temperature. Positive expression of ZO-1 and Occludin is considered positive when brown-yellow staining is detected. Finally, use ImageJ software to analyze the average optical density of all immunohistochemically stained sections.

Dissect the left femur of the mouse, remove any attached soft tissue, and sever both ends of the femur using bone scissors. Using a syringe, aspirate an appropriate volume of PBS and repeatedly flush the bone marrow cavity, collecting the released bone marrow cells into a centrifuge tube. Filter the solution through a 70 μm cell strainer, centrifuge the filtrate, and add red blood cell lysis buffer. Cells were subjected to three rounds of lysis to obtain a single-cell suspension. Take a small aliquot, stain with trypan blue, and enumerate the cells. Add 100 μL of the cell suspension into a centrifuge tube, followed by 5 μL each of CD45, CD3, CD4, and CD25 fluorescently labeled antibodies (BD Biosciences, San Jose, CA, United States) at the dose recommended by the manufacturer. Ensure thorough mixing, then incubate the samples in darkness at 4 °C for 30 min. Subsequently, centrifuge and remove the supernatant, followed by the addition of a cell membrane permeabilization reagent. Resuspend the cells gently, centrifuge once more, and discard the resulting supernatant. Repeat the resuspension using a permeabilization buffer for the fixed cells, centrifuge again, discard the liquid phase, and finally add 5 μL of Foxp3-specific antibody. Mix thoroughly and maintain in the dark at 4 °C for 50 min. After two resuspensions, analyze the cells using a flow cytometer and generate plots with CytExpert flow analysis software version CytExpert_Setup-2.5.0.77.

The right femur of the mouse was dissected, the attached soft tissues were removed, and the specimen was fixed in 4% paraformaldehyde. Each experimental group consisted of five animals. Micro-CT scanning was performed as previously described (31, 32). Three-dimensional images of trabecular bone and corresponding microstructural parameters were subsequently obtained. The bone was removed from the 4% paraformaldehyde, gently dried with absorbent paper, wrapped in gauze, and subsequently placed into the scanning chamber. After scanning with Cruiser software, slice images were generated. The largest cross-sectional plane was designated for quantitative evaluation of trabecular parameters. The Avatar system was employed to analyze the morphological variations of trabecular architecture, focusing on the region 1–1.5 mm distal to the growth plate. Three-dimensional reconstructions of trabecular parameters, including trabecular bone density (Tb. BMD), trabecular surface area to tissue volume ratio (Tb. BS/TV), trabecular thickness (Tb. Th), trabecular volume fraction (Tb. BV/TV%), trabecular pattern factor (Tb. PF), trabecular separation (Tb. Sp), trabecular number (Tb. N), and trabecular surface area to volume ratio (Tb. BS/BV), were measured.

In addition, cortical bone parameters were evaluated in the femoral mid-diaphysis. Cortical thickness (Ct. Th), cortical area (Ct. Ar), and cortical bone mineral density (Ct. BMD) were quantified to assess cortical bone alterations.

Dissect the right femur of the mouse, carefully remove the attached soft tissue, fix the sample in 4% paraformaldehyde, followed by dehydration, paraffin embedding, and sectioning. Stain the sections with H&E, and subsequently observe the area of the trabeculae. The trabecular bone area was analyzed using ImageJ. Within the same defined region, three fields of view were randomly selected for quantitative analysis.

For microbial relative abundance data, log₁₀ transformation was applied prior to statistical analysis to reduce skewness and approximate normal distribution. Chao1, Ace, Shannon, and Sobs were analyzed via the Kruskal–Wallis test. Beta diversity differences were assessed via principal coordinate analysis (PCoA) based on Bray–Curtis distance matrices and tested for significance using permutational multivariate analysis of variance (PERMANOVA).

In addition, other data were analyzed using SPSS 26.0. For group comparisons, the Shapiro–Wilk test was used to assess data normality, while Levene’s test assessed variance homogeneity. For data meeting parametric assumptions, one-way ANOVA followed by Tukey’s multiple comparisons was conducted. Non-parametric Kruskal–Wallis tests coupled with Dunn’s post hoc corrections were used for datasets violating parametric assumptions. In all statistical results, “*” indicates p < 0.05, “**” indicates p < 0.01, and “***” indicates p < 0.001.

Gut microbiota dysbiosis has been shown to contribute to bone loss via the “gut-bone axis,” however, studies focusing on microbiota alterations under conditions of mechanical unloading remain limited. To elucidate changes in gut microbial structure under mechanical unloading and to determine whether LGG plays a role, 16S rRNA gene sequencing was conducted on fecal from mice in the Con, HU, and HU + LGG groups. Rarefaction curves indicated sufficient sequencing depth (Figures 1A,B). Principal coordinate analysis (PCoA) showed clear separation between Con and HU groups, while HU + LGG clustered closer to Con, suggesting partial restoration of microbial structure by LGG (Figure 1C and Supplementary Figure S2). HU significantly reduced α-diversity indices (Ace, Chao, and Sobs), indicating decreased species richness, whereas LGG supplementation reversed this decline (Figures 1D–G).

To more intuitively assess alterations in gut microbial composition, relative abundance was examined across phylum, family, genus, and species levels. At the phylum level, the proportion of Bacteroidota increased while Firmicutes and other phyla decreased in the HU group, most of which were partially reversed in the HU + LGG group (Figure 1H). Further analysis at the family and genus levels allowed for a more detailed understanding of microbial regulation by HU and LGG (Figures 1I,J). At the genus level, HU elevated Muribaculaceae and Bacteroides while reducing Dubosiella and Lachnospiraceae. These alterations were partially reversed following LGG treatment. At the species level, HU markedly elevated uncultured bacterium and Bacteroidales bacterium from the Muribaculaceae family, which were effectively reduced by LGG supplementation (Figure 1K).

To determine whether LGG modulates SCFA metabolism under mechanical unloading, LC–MS/MS was used to quantify SCFAs in serum and intestinal tissues. Mechanical unloading significantly reduced acetic, propanoic, and butanoic acid levels, whereas LGG supplementation restored these metabolites in both serum and intestine (Figures 2A–C,G). KEGG pathway enrichment revealed that SCFAs were mainly involved in metabolic processes, including carbohydrate and energy metabolism, and were enriched in the “protein digestion and absorption” pathway, indicating their importance in nutrient processing (Figures 2D,E,J,K). Acetic, propanoic, and butanoic acids were the predominant SCFAs, with no significant differences in minor acids such as hexanoic or isovaleric acid among groups (Figures 2F,L).

LGG abundance in fecal samples was first quantified by qPCR. The results showed that LGG was detectable in the HU + LGG group, with levels ranging from approximately 10⁴ to 10⁶ CFU per gram of feces, whereas LGG was not detected in fecal samples from the Control and HU groups (Supplementary Figure S3). Given the critical role of the intestinal barrier in regulating systemic inflammation and nutrient absorption, immunohistochemistry and WB were performed to detect the expression of ZO-1 and Occludin in colonic tissues. The results demonstrated that HU significantly reduced the protein expression of ZO-1 and Occludin, with mean optical density values significantly decreased (p < 0.05), indicating compromised intestinal barrier integrity and increased intestinal permeability (Figure 3A). Following oral LGG supplementation, the levels of ZO-1 and Occludin were markedly restored, indicating enhanced intestinal barrier integrity and decreased epithelial permeability (p < 0.05) (Figures 3B–D).

LPS, a marker of bacterial endotoxin, is known to provoke inflammatory responses in the bone microenvironment, contributing to bone loss (33, 34). Increased intestinal permeability may facilitate LPS translocation into the serum, potentially aggravating inflammatory infiltration and osteolysis. HU markedly increased serum LPS levels (p < 0.001), consistent with compromised barrier function and elevated permeability (Figure 3E). LGG supplementation significantly reduced serum LPS concentrations, further supporting its protective role in barrier integrity. Consistently, qRT-PCR analysis showed that the mRNA expression level of TLR4 was significantly increased in the HU group (p < 0.001), whereas LGG treatment partially reversed this increase (p < 0.05, Supplementary Figure S4).

To explore whether LGG-induced improvements in gut and metabolic function could modulate inflammation in the bone microenvironment, the levels of key inflammatory markers and the proportion of Treg cells in bone marrow were assessed. ELISA results revealed that HU significantly elevated TNF-α and IL-1β, while reducing IL-10 (p < 0.001) (Figures 4A–C). This was corroborated by WB, which confirmed increased TNF-α and IL-1β protein levels and decreased IL-10 expression following HU (Figure 4D). These findings, in line with elevated serum LPS levels and impaired intestinal barrier function, indicate that HU provokes inflammatory responses in the bone microenvironment. LGG intervention significantly suppressed TNF-α and IL-1β (p < 0.05) expression while restoring IL-10 levels (p < 0.001), thereby rebalancing the inflammatory milieu.

Considering the critical function of Treg cells in modulating immune suppression and influencing bone remodeling, flow cytometry was employed to quantify CD4⁺CD25⁺Foxp3⁺ Treg cells in bone marrow (Figures 4E,F and Supplementary Figure S5) (35, 36).

To determine whether LGG could alleviate bone deterioration induced by mechanical unloading, key trabecular parameters of the distal femur were analyzed. Histological evaluation was performed, and bone metabolic markers in the bone marrow were quantified. Micro-CT results (Figures 5A–G) indicated that trabecular bone in the HU group was substantially reduced, as reflected by significantly decreased values of BMD, BV/TV, BS/TV, Tb. N, and Tb. Th, along with a notable rise in Tb. Sp (p < 0.05). Oral administration of LGG was found to improve these parameters, indicating that LGG supplementation could mitigate HU-induced trabecular damage. These findings were corroborated histologically by H&E staining, which revealed a pronounced reduction in trabecular area and sparse bone matrix in the HU group (Figures 5H,I), whereas LGG-supplemented mice exhibited denser trabecular structures and more regular architecture. HU also induced significant changes in cortical bone microarchitecture, which were partially attenuated by LGG treatment (Supplementary Figure S6).

To gain deeper insights into the influence of LGG on bone remodeling balance, serum levels of BALP, OPG, OCN, CTX, and PINP were measured (Figures 5J–N). HU significantly decreased BALP, OPG, OCN, and PINP levels while markedly increasing CTX levels, indicating reduced osteoblast activity and enhanced osteoclast-mediated bone resorption, thereby disrupting bone remodeling homeostasis. Notably, LGG supplementation was able to modulate the differentiation and maturation of both osteoblasts and osteoclasts, as evidenced by the restorative elevation of BALP, OPG, OCN, and PINP levels, and a concurrent reduction in CTX concentration.

The HU mouse model is a widely used animal model for simulating mechanical unloading, such as that induced by microgravity or prolonged bed rest. However, its translational applicability to human physiology remains limited. Additionally, the use of LGG in this study was restricted to animal models, and its clinical relevance has yet to be evaluated. While the biosafety of LGG has been partially demonstrated, further comprehensive validation of its safety profile may be of great importance for future clinical translation. Moreover, most experimental groups in the present study consisted of five biological replicates, which may reduce statistical power and should be considered when interpreting the results.

This study did not include a separate healthy + LGG (positive control) group, as our primary objective was to investigate the protective effects of LGG under mechanical unloading conditions, rather than its baseline influence under normal physiological states. Previous studies have consistently shown that LGG does not exert adverse effects on skeletal parameters in healthy animals. Nevertheless, the lack of a positive control group may still be considered a limitation, potentially restricting the interpretation of LGG’s independent effects unrelated to unloading. More rigorous experimental grouping will be considered in future mechanistic investigations.

This study primarily focused on phenotypic observations of bone loss modulation via the gut-bone axis under HU and LGG intervention. Nevertheless, the core molecular pathways governing this regulatory process have yet to be elucidated. Additional studies are warranted to delineate the specific signaling pathways involved in this axis.