The following antibodies/kits were procured from eBioscience (San Diego, CA, USA): PerCp-Cy5.5 Anti-Mouse-CD4-(RM4-5) (550954), APC Anti-Mouse/Rat-Foxp3 (FJK-16s) (17-5773), PE Anti-Human/Mouse-Rorγt (AFKJS-9) (12-6988), PerCp-Cy5.5 Anti-Mouse-CD19 (1D3) (45-0193-82), PE-Cy7 Anti-Mouse-CD5 (53-7.3) (25-0051-81), APC Anti-Mouse-CD1d (1B1) (17-0011-82), APC Cy7 Anti-Mouse-F4/80-(BM8) (47-4801-82), Foxp3/Transcription factor staining buffer (0-5523-00), and RBC lysis buffer (00-4300-54). The following ELISA kits were brought from R&D: Mouse IL-10 (M1000B) and Mouse IL-17 (M1700) Quantikine ELISA kits. The following ELISA kits and reagents were brought from BD (Franklin Lakes, NJ, USA): Mouse IL-6 (OptEIA™-555240) and Mouse TNF-α (OptEIA™-560478). Acid phosphatase, leukocyte (TRAP) kit (387A), FITC-phalloidin (P5282), and DAPI were purchased from Sigma (St. Louis, MO, USA). Macrophage-colony stimulating factor (M-CSF) (300-25) and receptor activator of nuclear factor κB-ligand (sRANKL) (310-01), Human TGF-β1 (AF-100-21C), Murine IL-2 (AF-212-12), Murine IL-6 (AF-216-16), and Murine IL-23 (200-23) were procured from PeproTech (Rocky Hill, NJ, USA). α-Minimal essential media and RPMI-1640 were purchased from Gibco (Grand Island, NY, USA). Bifido broth was procured from HiMedia Labs (Hyderabad, India). Bifidobacterium longum UBBL-64 was procured from Unique Biotech Ltd., Hyderabad, India.

All in vitro and in vivo experiments were carried out in 8–10-week-old female C57BL/6 J mice. Mice were maintained under specific pathogen-free (SPF) conditions at the animal facility of All India Institute of Medical Sciences (AIIMS), New Delhi, India. Mice were fed with sterilized food and autoclaved drinking water ad libitum. Mice were exposed to bilateral ovariectomy (ovx), and sham surgery was performed on mice after anesthetizing mice with ketamine (100–150 mg/kg) and xylazine (5–16 mg/kg) intraperitoneally. Subsequently, mice were randomly allocated into three groups with 6 mice in each group, i.e., sham (control), ovx, and ovx + Bifidobacterium longum (BL). After 1 week post-surgery, the ovx+BL group of mice were orally gavaged with 400 µl of BL suspension (containing 10⁹ cfu) daily in drinking water for a period of 6 weeks. At the end of 6 weeks, animals were euthanized, and blood, bones, and lymphoid tissues were harvested for further analysis. The body weight of mice was documented at regular intervals (day 1, day 21, and day 45) during the dose administration period. All the measures were performed after the due approval of the protocols submitted to the Institutional Animal Ethics Committee of AIIMS, New Delhi, India (196/IAEC-1/2019).

Bifidobacterium longum UBBL-64 (M1395) was cultured in Bifido broth containing 0.05% of L-cysteine under anaerobic conditions. On the following day, overnight culture was subcultured into fresh Bifido broth containing freshly prepared L-cysteine (0.05%), and the culture was grown until it attained the log phase (OD 600nm = 0.4). Next, cells were harvested and washed with 1× PBS twice to remove the traces of Bifido broth with centrifugation at 4,000 rpm for 10 min and proceeded for conditioned medium preparation. The conditioned media (CM) of B. longum were prepared by resuspending the cells with either α-MEM or RPMI-1640 antibiotic-free media and incubated for the next 3 h at 37°C. Lastly, cell-free BL-CM were collected via pelleting out the bacterial cells, pH neutralized, and filtered with a 0.22-µm filter. Further, BL-CM were used for all downstream assays, e.g., osteoclast and immune cell differentiation.

For osteoclast differentiation, mouse bone marrow cells (BMCs) were isolated from 8- to 12-week-old C57BL/6J mice by flushing the femoral bone with complete α-MEM media. After performing RBC lysis with 1× RBC lysis buffer, BMCs were cultured overnight in a T-25 flask in endotoxin-free complete α-MEM media (10% heat-inactivated FBS) supplemented with M-CSF at the 35-ng/ml concentration. The next day, non-adherent cells were collected and seeded in 96-well plates (50,000 cells/well) in complete α-MEM media supplemented with M-CSF (30 ng/ml) and RANKL (100 ng/ml) in the presence or absence of BL-CM at different ratios, viz., 1:100, 1:10, and 1:5 for 4 days. On day 3, half media were replenished with fresh complete α-MEM media supplemented with fresh factors. Lastly, for monitoring the generation of multinucleated osteoclasts, tartrate-resistant acid phosphatase (TRAP) staining was carried out according to the manufacturer’s instructions. Briefly, at the end of incubation, cells were carefully washed thrice with 1× PBS and fixed with a fixative solution containing citrate, acetone, and 3.7% formaldehyde solution and incubated for 10 min at RT. Next, fixed cells were washed twice with 1× PBS and incubated with a TRAP-staining solution in the dark at 37°C for 5–15 min. Multinucleated TRAP-positive cells with ≥3 nuclei were considered as osteoclasts, and these cells were counted and imaged using an inverted microscope (Eclipse, TS100, Nikon and EVOS, Thermo Scientific, Waltham, MA, USA). The area of osteoclasts was estimated with ImageJ software (NIH, USA).

F-actin ring formation assay was carried out as described previously (16). Briefly, bone marrow-derived osteoclast precursors were seeded on glass coverslips in 12-well plates; at day 4, processing for F-actin polymerization staining was performed. Cells were washed twice with 1× PBS and fixed with 4% paraformaldehyde (PFA) for 20 min and permeabilized with 0.1% Triton X-100 for 5 min. Further, to block the non-specific binding, cells were blocked with 1% BSA for 30 min and stained with FITC-labeled phalloidin for 1 h at RT in the dark. Lastly, cells were stained with DAPI (10 µg/ml) for 5 min in the dark. Finally, slides were observed under an immunofluorescence microscope (Imager.Z2, Zeiss, Jena, Germany) for analyzing F-actin ring formation.

Splenic B cells from C57BL/6 mice were purified by magnetic separation described previously (7). Briefly, following RBC lysis, cells were subjected to a biotin-labeled CD19 antibody (BD, USA) and incubated for 30 min at 4°C. After washing, the labeled cells were incubated with Streptavidin Particles Plus-DM for 30 min at 4°C. Further, cells underwent magnetic separation and positive/negative fractions having B-cell purity (>95%) were cultured in 24-well plates (2 × 10⁶) in the presence or absence of LPS (10 µg/ml) and BL-CM (1:5) for 24 h at 37°C in a 5% CO₂ incubator. At the end of incubation, LPS and LPS + BL-CM-induced Bregs were harvested and processed for either flow cytometry or coculturing experiments (with BMCs for osteoclastogenesis and naïve T cells for Treg/Th17 cell differentiation).

For estimating the potential of BL-CM-induced Bregs to suppress osteoclast differentiation, BMCs were cocultured with either Bregs or BL-CM-induced Bregs in 96-well plates in different ratios (10:1, 5:1, and 1:1) in the presence of M-CSF (30 ng/ml) and RANKL (100 ng/ml) for 4 days. At an interval of 2 days, half media were replenished with media containing fresh factors. After 4 days of incubation, TRAP staining was performed to evaluate the generation of osteoclasts. For evaluating the potential of Bregs to modulate Treg–Th17 differentiation, BL-CM-induced Bregs were cocultured with negatively selected naïve T cells (CD4⁺CD25^- T cells) at 1:1 in anti-CD3 (10 µg/ml) and anti-CD28 (2 µg/ml) mAb-coated 48-well plates under non-polarization conditions. On day 4, cells were harvested, and flow cytometry was performed for estimating the percentages of CD4⁺Foxp3⁺IL-10⁺ Tregs and CD4⁺Rorγt⁺IL-17⁺ Th17 cells.

PBMCs were obtained from heparinized blood by gently layering the blood on Histopaque at 1:3 in a 15-ml tube and centrifuged at 800 × g for 25 min at RT (with brakes off). The buffy coat layer beneath the plasma was carefully collected in a separate tube and washed with 1× PBS by inverting the tube gently and centrifuging the tube at 400 × g for 10 min at 4°C. The obtained PBMCs were seeded in a 96-well plate at a seeding density of 1× 10⁶ cells/well, and the plate was incubated for 2 h in a humidified 5% CO₂ incubator and washed with α-MEM twice (without FBS). Upon adherence, cells were incubated with α-MEM supplemented with 10% FBS, M-CSF (30 ng/ml), and RANKL (100 ng/ml) in the presence or absence of BL-CM at different ratios (1:100, 1:10, and 1:5), and the plate was incubated for the next 14 days in a CO₂ incubator with half-medium replenishment on every 3rd day (i.e., 72 h). At the end of incubation, TRAP staining was performed for evaluating the differentiation of multinucleated osteoclasts. All the measures were performed after the due approval of the protocols submitted to the Institute Ethics Committee for Post Graduate Research (IECPG-482), AIIMS, New Delhi, India.

Scanning electron microscopy (SEM) was performed for the cortical region of femoral bones, as described previously (10, 16, 17). Briefly, bone samples were kept in 1% Triton-X-100 for 2–3 days and later samples were transferred to 1× PBS buffer till the final analysis was performed. Next, after the preparation of bone slices, the samples were dried under the incandescent lamp and sputter coating was done. Afterward, bones were scanned in a Leo 435 VP microscope equipped with a 35-mm photography system. SEM images were digitally photographed at ×100 magnification to catch the finest cortical region. After imaging, SEM images were further analyzed by MATLAB (MathWorks, Natick, MA, USA).

Upon drying femur bones under 100-W lamps for 6 h followed by high-vacuum drying, samples were analyzed using an atomic force microscope (AFM) (Innova Icon, Bruker, Billerica, MA, USA) set in Acoustic AC mode. This was assisted by cantilever (NSC 12(c) MikroMasch, Silicon Nitride Tip) and NanoDrive version 8 software, set at a constant force of 0.6 N/m with a resonant frequency at 94–136 kHz. Images were recorded at a scan speed of 1.5–2.2 lines/s in the air at room temperature. Images were later processed and analyzed by using Nanoscope analysis software.

Micro-computed tomography (µ-CT) scanning and analysis were performed using in vivo X-ray SkyScan 1076 scanner (Aartselaar, Belgium) tomography, as described before (18). Briefly, scanning was done at 50 kV, 204 mA, using a 0.5-mm aluminum filter by positioning the samples at the right orientation in the sample holder. For the reconstruction process, NRecon software was employed. After reconstruction, ROI was drawn at a total of 100 slices in secondary spongiosa at 1.5 mm from the distal border of growth plates and further processed for CTAn software for evaluating and calculating the micro-architectural parameters of bone samples. Several 3D-histomorphometric parameters were obtained, viz., bone volume/tissue volume (BV/TV), trabecular thickness (Tb.Th), trabecular separation (Tb.Sp), etc. The volume of interest of u-CT scans made for trabecular and cortical regions was used to determine the BMD of LV5, femur, and tibia. BMD was measured by using hydroxyapatite phantom rods of 4-mm diameter with known BMD (0.25 g/cm³ and 0.75 g/cm³) as a calibrator (16).

To measure the biomechanical properties of bones, femoral bones of mice in all the respective groups were exposed to three-point bending by employing the bone strength tester model TK-252C/RDT (Muromachi Kikai Co. Ltd., Tokyo, Japan). Briefly, for this, femoral bone was placed on the two supports that were kept at a constant distance of 1 cm. By means of load displacement curves, the following bone mechanical strength parameters were evaluated: maximum power (N), energy to fracture (mJ), and stiffness (N/mm).

Cells were harvested from various lymphoid organs (viz., BM and spleen) and stained with antibodies specific for macrophages, Bregs, Tregs, and Th17 cells. For the macrophage panel, BM cells were stained with the anti-F4/80-APC Cy7 antibody. Breg/Treg/Th17 cells were first stained for Bregs (anti-CD19-PerCP-Cy5.5, anti-CD5-PE-Cy7, and anti-CD1d-APC antibodies) and Tregs/Th17 (anti-CD4-PerCP-Cy5.5) antibodies for cell surface staining and incubated for 30 min in the dark on ice. After washing, cells were further fixed and permeabilized with a 1× fixation–permeabilization buffer for 30 min on ice in the dark. Finally, for intracellular staining, cells were stained with anti-IL-10-BV421 (Bregs), anti-Foxp3-APC & anti-IL-10-BV421 (Tregs) and anti-Rorgt-PE & anti-IL-17-APC (Th17 cells) for 45 min. After washing, cells were acquired on BD LSRFortessa (USA). FlowJo 10 (Tree Star, Woodburn, OR, USA) software was used to analyze the samples, and the gating strategy was done as per experimental requirements.

Enzyme-linked immunosorbent assay (ELISA) was carried out for the quantitative assessment of cytokines IL-6, IL-10, IL-17, TNF-α, and IFN-γ in blood sera of all the mouse groups by utilizing commercially available kits as per the manufacturer’s instructions.

Statistical differences between the distinct groups were evaluated by employing analysis of variance (ANOVA) with subsequent analysis via Student’s t-test paired or unpaired as appropriate. All the values in the data are expressed as mean ± SEM (n = 6). Statistical significance was determined as p ≤ 0.05 (*p < 0.05, **p < 0.01, ***p, 0.001) with respect to the indicated groups.

To elucidate the involvement of BL in modulating bone health, we initially examined the potential of BL on receptor activator of nuclear kappa B ligand (RANKL)-induced osteoclast differentiation and their functional activity. A schematic diagram presents the experimental protocol ( Figure 1A ; for details, refer to Methods). BMCs were stimulated with osteoclastogenic media supplemented with M-CSF (30 ng/ml) and RANKL (100 ng/ml) in the presence or absence of BL-CM at different ratios (1:100, 1:10, and 1:5). After 4 days, cells were harvested and processed for both TRAP staining and F-actin ring polymerization assay. There was a significant reduction in the differentiation of osteoclasts in a dose-dependent manner as evident by the number of multinucleated (>3 nuclei) TRAP-positive cells in the BL-treated group in comparison to the control group ( Figures 1B–D ). Furthermore, the area of multinucleated osteoclasts in treatment groups was significantly less than in the control group ( Figure 1E ). Also, the number and area of F-actin rings were significantly decreased, which indicated that BL inhibited not only osteoclastogenesis but also the functional activity of osteoclasts ( Figures 1F–I ).

Recently, our group reported that Bregs exhibit the potential to inhibit osteoclast differentiation (7). Thus, we were interested in estimating the potential of BL in modulating the differentiation of Bregs. For this purpose, positively selected splenic B cells were stimulated under Breg-polarizing conditions in the presence of BL-CM for 24 h ( Figure 2A ). We observed that BL significantly enhanced the percentage of CD19⁺CD1d^hiCD5⁺ Bregs (p < 0.05) along with a significant increase in IL-10 production ( Figures 2B–F ). Taken together, these results indicate a strong immunomodulatory potential of BL in modulating Breg differentiation.

Next, we studied whether BL has the potential to further enhance the anti-osteoclastogenic potential of Bregs. To this aim, we cocultured BL-induced Bregs and BMCs at various ratios (10:1, 5:1, and 1:1) for 4 days ( Figure 3A ). Our TRAP data showed that BL-induced Bregs inhibited the generation of osteoclasts more efficiently than the non-BL-stimulated Bregs ( Figures 3B–D ). These data indicate that BL could also prevent bone loss by suppressing osteoclast differentiation via Bregs.

Moving ahead in our study, we were next keen in assessing the immunomodulatory potential of BL-induced Bregs in regulating the differentiation of naïve T cells into either Tregs or Th17 cells. To this aim, we cocultured BL-stimulated Bregs with negatively selected naïve T cells at a 1:1 ratio for 3 days ( Figure 4A ) in anti-CD3- and anti-CD28-coated plates under non-polarizing conditions. Subsequently, cells were harvested and analyzed for the percentages of CD4⁺Foxp3⁺IL-10⁺ Tregs and CD4⁺Rorγt⁺IL-17⁺ Th17 cells by flow cytometry. We observed that BL-stimulated Bregs significantly increased the percentage of both CD4⁺Foxp3⁺ Tregs (p < 0.01) and CD4⁺IL-10⁺ Tr1 cells (p < 0.001) in comparison to control groups ( Figures 4B–E ). Besides, BL-stimulated Bregs significantly reduced the percentage of CD4⁺Rorγt⁺IL-17⁺ Th17 cells (p < 0.001) ( Figures 4F–I ). Taken together, these data strongly suggest that the enhancement in the percentage of Bregs in response to BL stimulation is pivotal for the efficient induction of anti-osteoclastogenic Tregs along with concurrent inhibition of osteoclastogenic Th17 cells, thereby pointing toward the role of the “Breg–Treg–Th17” cell axis in ameliorating bone loss in vivo.

We next assessed the effect of BL in mitigating bone loss caused by estrogen deficiency (ovx) conditions using adult female C57BL/6 mice that were distributed into three groups: sham surgery (ovary intact), bilateral ovx (both ovaries removed), and ovx group with a daily oral administration of BL (10⁹cfu) for 6 weeks. At the end of the treatment, mice were sacrificed by euthanasia, and the bones were collected for further studies ( Figure 5A ). There was no significant difference in the bodyweight between the groups throughout the duration of the study ( Figure S1 ). SEM of the femoral cortical bone sections showed higher bone resorption pits or lacunae in the ovx group compared with the sham group whereas the BL group showed a significant mitigation of pits compared with the ovx group ( Figure 5B ). To further examine these 2D SEM images quantitatively in a more statistical manner, we performed MATLAB analysis to obtain the association between bone loss and bone mass. MATLAB analysis of SEM images represents the magnitude of homogeneity where the higher correlation is denoted by the red color (enhanced bone mass) and the blue color (lower correlation) symbolizes enhanced bone loss. MATLAB analysis of 2D SEM images indicated that the BL-administered ovx group showed a higher correlation value and thus enhanced bone mass ( Figure 5C ). We next estimated the 3D topology of the bone surface in response to BL treatment by AFM which indicated reduced bone resorption and roughness of the bone surface in the BL-administered ovx group ( Figure 5D ). Moreover, MATLAB analysis of AFM images further suggests improved bone architecture and decreased osteoclastogenesis in the BL-treated group compared with the ovx group ( Figure 5E ). These findings strongly indicate that BL administration attenuates bone loss in ovx mice.

Osteoporosis and fracture risk are high in both spine and hip of postmenopausal women. These two anatomical regions are represented by lumbar vertebra (5th) and proximal femur metaphysis respectively. Micro-computed tomography (µ-CT) assessment of lumbar vertebrae (LV)-5 showed loss of bone volume and deteriorated microarchitecture in the ovx mice compared with both sham and BL-treated groups ( Figure 6A ). BL supplementation significantly enhanced bone volume per tissue volume (BV/TV) (p < 0.01) and trabecular thickness (Tb.Th) (p < 0.01) and decreased trabecular separation (Tb.Sp) (p < 0.05) compared with the ovx group ( Figure 6B ).

We next investigated the effect of BL on appendicular bones including femur and tibia. The 3D micro-architecture of proximal metaphysis of the femur and tibia showed that BV/TV and Tb.Th were decreased and Tb.Sp increased in ovx mice compared with sham, suggesting loss of trabecular bones, and BL treatment reversed all these ovx-induced changes ( Figures 6C–F ). Cortical bones of femur and tibia showed thinning due to ovx as periosteal area (T.Ar), periosteal perimeter (T.Pm), and cortical thickness (Cs.Th) were decreased compared with the sham, and BL treatment reversed these changes ( Figures 7A–D ). These data demonstrate that BL administration significantly improves the micro-architecture and histomorphometric parameters of both trabecular and cortical bones in osteoporotic mice.

Since BMD is a predictor of osteoporotic fracture, we measured it at all weight-bearing bones of axial and appendicular sites. µCT allows the measurement of BMD, and our data showed that it was significantly decreased at all sites measured in ovx mice compared with sham; and BL treatment reversed these changes ( Figures 8A–E ).

We next assessed the bone quality by measuring the bending strength of femurs. The load-bearing capacity of bone, energy to failure, and stiffness were significantly decreased in the ovx mice compared with sham, and these parameters were maintained to the sham level in the BL group ( Figures 8F–H ). These data suggest maintenance of bone mass and strength in the ovx mice treated with BL.

Our in vitro data showed that BL has anti-osteoclastogenic potential which is further augmented by the favorable modulation of the immunoporotic ability of Bregs. Thus, we next measured Breg, Treg, and Th17-cell populations in lymphoid organs including bone marrow (BM) and spleen in response to BL treatment in ovx mice. We observed that the CD19⁺CD1d^hiCD5⁺ Breg population was significantly decreased in the BM (p < 0.05) and spleen (p < 0.01) of ovx mice compared with sham, and BL administration to ovx increased this cell population to the sham level ( Figures 9A–D ). Moreover, the percentages of CD4⁺Foxp3⁺ Tregs in BM (1.5-fold, p < 0.05) and spleen (2-fold, p < 0.05) in the ovx group were lower than those of the sham and BL treatment restored these parameters ( Figures 9E–H ). Conversely, ovx mice had higher percentages of CD4⁺Rorγt⁺ Th17 cells in BM (3-fold, p < 0.05) and spleen (3.5-fold, p < 0.05) than the sham, and BL treatment reversed these changes ( Figures 9I–L ). Besides, the circulating levels of pro-osteoclastogenic cytokines including IL-6, TNF-α, and IL-17 were increased and anti-osteoclastogenic cytokines such as IFN-γ and IL-10 were decreased in ovx mice compared with sham ( Figure 10 ). BL treatment of ovx mice completely reversed the cytokine profiles, which further suggests its role in attenuating inflammatory bone loss in postmenopausal osteoporosis via modulating the immunoporotic “Breg–Treg–Th17” cell axis.

Moving ahead, we corroborated our findings of the effect of BL in suppressing murine osteoclastogenesis in human samples. Human PBMCs when treated with BL-CM resulted in a concentration-dependent decrease in the number of multinucleated TRAP-positive cells ( Figures 11A–D ). Furthermore, area measurement using ImageJ software indicated significant reduction in the area of multinucleated osteoclasts in response to BL-CM ( Figure 11E ). These data thus confirm the anti-osteoclastogenic effect of BL on human PBMCs.