Lactobacillus rhamnosus attenuates bone loss and maintains bone health by skewing Treg-Th17 cell balance in Ovx mice

Osteoporosis is a systemic-skeletal disorder characterized by enhanced fragility of bones leading to increased rates of fractures and morbidity in large number of populations. Probiotics are known to be involved in management of various-inflammatory diseases including osteoporosis. But no study till date had delineated the immunomodulatory potential of Lactobacillus rhamnosus (LR) in bone-health. In the present study, we examined the effect of probiotic-LR on bone-health in ovariectomy (Ovx) induced postmenopausal mice model. In the present study, we for the first time report that LR inhibits osteoclastogenesis and modulates differentiation of Treg-Th17 cells under in vitro conditions. We further observed that LR attenuates bone loss under in vivo conditions in Ovx mice. Both the cortical and trabecular bone-content of Ovx+LR treated group was significantly higher than Ovx-group. Remarkably, the percentage of osteoclastogenic CD4+Rorγt+Th17 cells at distinct immunological sites such as BM, spleen, LN and PP were significantly reduced, whereas the percentage of anti-osteoclastogenic CD4+Foxp3+Tregs and CD8+Foxp3+Tregs were significantly enhanced in LR-treated group thereby resulting in inhibition of bone loss. The osteoprotective role of LR was further supported by serum cytokine data with a significant reduction in osteoclastogenic cytokines (IL-6, IL-17 and TNF-α) along with enhancement in anti-osteoclastogenic cytokines (IL-4, IL-10, IFN-γ) in LR treated-group. Altogether, the present study for the first time establishes the osteoprotective role of LR on bone health, thus highlighting the immunomodulatory potential of LR in the treatment and management of various bone related diseases including osteoporosis.


Lactobacillus rhamnosus (LR) inhibits osteoclastogenesis in vitro.
To determine whether LR possesses potential to modulate bone health, we first examined the effect of LR on RANKL induced osteoclasts differentiation under in vitro conditions. In order to study the same, we prepared LR-conditioned media (LR-CM) by culturing LR in α-MEM media for 3 h and the supernatant was collected and further used as LR-CM. For in vitro osteoclastogenesis assay, mice bone marrow macrophages were stimulated with M-CSF (30 ng/ ml) and RANKL (100 ng/ml) in the presence or absence of LR-CM at different ratios (viz. 1:10 and 1:1). After five days, cells were fixed and stained for Tartrate resistant acid phosphatase (TRAP) to identify differentiated multinucleated osteoclasts. Interestingly, we observed that LR-CM treatment significantly decreased the osteoclasts differentiation in a dose dependent manner estimated by the significantly reduced TRAP positive cells in LR-CM treated groups in comparison to control group (Fig. 1A). Furthermore, area measurement analysis of multinucleated TRAP positive cells using Image J software revealed significant reduction (25-fold) in the area of TRAP positive osteoclasts in the treatment groups ( Fig. 1B-D). To exclude the possibility that the observed reduction in osteoclasts differentiation and number is not due to cell cytotoxicity, MTT assay was performed and we found no significant difference in cell viability with LR-CM treatment at different dilutions (data not shown). www.nature.com/scientificreports/ Thus, our data clearly suggest that LR has potential to inhibit RANKL induced osteoclastogenesis under in vitro conditions.
LR inhibits F-actin ring formation. F-actin ring formation is a visual phenotype of mature osteoclasts for mediating its functional activity i.e. bone resorption 28 . Thus, we next addressed the effect of LR on F-actin ring formation by fixing and permeabilizing bone marrow derived osteoclasts on glass coverslips and further cells were stained for F-actin (FITC-labelled phalloidin) and nuclei (DAPI) respectively. Strikingly, we observed that with respect to control group, LR-CM treatment in a dose dependent manner significantly decreased the number

LR attenuates bone loss in ovariectomized mice.
To test whether these in vitro findings of LR could be exploited as a novel strategy in the treatment of pathological bone loss conditions, we next investigated the role of LR in modulating bone health in Ovx-induced postmenopausal osteoporotic mice model. For accomplishing the same, female BALB/c mice were divided in three groups viz. Sham, Ovx and Ovx+LR and after one-week post-surgery, LR was administered orally for a period of 6wks. At the end of experiment, mice were sacrificed, and bones were harvested for further analysis (Fig. 3A). During dosage period, we also examined body weight of all groups at regular intervals (Day 1, 21 and 45) but no difference in weight was observed in Ovx and LR treated group (Fig. 3B). To investigate whether LR-administration specifically inhibits bone-loss in Ovx mice, we further studied the effect of LR-administration on bone pathology and bone remodelling processes. Scanning electron microscopy (SEM) analysis of cortical region of femoral bones revealed that mice of Ovx group had enhanced number of resorption pits or lacunae representing higher osteoclastogenesis (Fig. 4A). Strikingly, LR administration significantly reduced the resorption pits/lacunae in femoral bones of Ovx group, a clear sign of reduced/inhibited osteoclastogenesis (Fig. 4A). To further analyse SEM images quantitatively in a more statistical manner, we employed MATLAB (matrix-laboratory) to derive the correlation between bone mass and bone loss. MATLAB analysis of 2D-SEM images signifies the degree of homogeneity where red colour symbolizes higher correlation (high bone mass) whereas blue colour symbolizes lesser correlation (more bone loss). MATLAB-data of SEM (Fig. 4B), clearly points that Ovx+LR group has greater correlation and thus more bone mass. Since SEM images illustrate 2D-information of bone samples and we were also interested to study the 3D-topology of bone-structures, so we next performed atomic force microscopy (AFM) analysis of cortical region of femoral bone. Our AFM data showed a significant suppression of bone resorption in Ovx+LR admin-  www.nature.com/scientificreports/ istered group in comparison to Ovx group (Fig. 4C). Moving ahead, MATLAB-analysis of AFM-3D images ( Fig. 4D) was performed in which red colour represents enhanced bone architecture (reduced osteoclastogenesis) and blue colour representing reduced bone architecture (enhanced osteoclastogenesis). This AFM data thus further supplement and validate our earlier SEM data and support our hypothesis that LR inhibits bone loss in Ovx mice. Taken together, our findings from both SEM and AFM data suggest that LR treatment prevents bone loss in Ovx mice. www.nature.com/scientificreports/ LR enhances bone microarchitecture in ovariectomized mice. Moving ahead in our study we were subsequently interested in deciphering the effect of LR-administration on bone histomorphometric parameters. We thus performed high resolution micro-computed tomography (μ-CT) (a gold standard for determining bone-health) for evaluation of bone morphology and quantifying various bone morphometric indices related to bone loss and bone mass. Lumbar vertebrae-5 (LV-5) is considered as one of the most peculiar regions to diagnose early bone loss or osteoporosis 13,14,29 . Thus, we analysed the effect of LR administration on LV-5 trabecular region. Interestingly, μ-CT data clearly pointed towards a significant improvement in LV-5 bone micro-architecture in LR administered group (Fig. 5A). In addition to the micro-architecture, LR administration also significantly increased LV-5 trabecular parameters viz. bone volume per tissue volume (BV/TV) (p < 0.05) and trabecular thickness (Tb.Th) (p < 0.05) and significantly reduced trabecular separation (Tb.Sp) (p < 0.05) (Fig. 5B). Next, we also performed µ-CT analysis of femoral and tibial bones by analysing the effect of LR-administration on various trabecular and cortical indices in mice. Thus, µ-CT was further used to separately quantify various bone indices parameters in femur and tibia bones. In comparison to Ovx group, 3D-micro-architecture images of trabecular region of respective bones showed significant improvement in LR treated group (Fig. 5C). Moreover, during measurement of various indices for femoral and tibial trabecular region, it was found that administration of LR significantly enhanced bone micro-architecture by augmenting the BV/TV ratio (p < 0.05), Tb.Th (p < 0.05) along with reducing Tb.Sp (p < 0.01) (Fig. 5D-F) in Ovx group. Notably, we also found comparable data in cortical region of femoral and tibial bones with improvement in bone micro-architecture along with significant enhancement in bone histomorphometric parameters ( Fig. 6A-D). Altogether, our data establishes that LR-administration in Ovx mice improves both trabecular and cortical bone microarchitecture of LV-5, femoral and tibial bones.

LR elevates both bone mineral density (BMD) and heterogeneity of bones.
To further support our µCT data it was also important to assess the bones for BMD which draws a clear picture about the presence of minerals and their concentration in respective bones. Notably, we observed that administration of LR significantly enhanced the BMDs of both trabecular and cortical regions of LV-5, femoral and tibial bones in Ovx mice ( Fig. 7A). Bones have a heterogeneous composition and any loss in heterogeneity has been linked with enhanced fracture risk 30 . Thus, we subsequently performed Fourier transform infrared spectroscopy (FTIR) to evaluate the effect of LR-administration on heterogeneity of bones. The analysis of bone samples revealed that Ovx mice administered with LR had significantly enhanced heterogeneity parameters viz. crystallinity (XST) (p < 0.01), carbon content (c/p) (p < 0.05) and mineral to organic matrix ratio (m/m) (p < 0.05) with respect to Ovx group (Fig. 7B). Taken together these data clearly support and validate our hypothesis that LR-administration not only enhances the BMDs of bones but also preserves their natural heterogeneity, thereby making LR a suitable therapeutic option in the management and treatment of postmenopausal osteoporosis.

LR enhances bone health via modulating Treg-Th17 cell balance. Since the role of Treg-Th17 cells is
well established in bone homeostasis 10 thereby we next explored the role of LR in modulating Treg-Th17 cell balance. In this context, we analysed the percentages of CD4 + Foxp3 + Tregs, CD8 + Foxp3 + Tregs and CD4 + Roryt + Th17 immune cells at various immunological sites such as BM, spleen, PP and LN by flow cytometry. In comparison to Ovx group, LR-administration in Ovx mice led to threefold enhancement in CD4 + Foxp3 + Treg cells in BM (Fig. 8A). Interestingly, CD8 + Foxp3 + Tregs were also found to be significantly enhanced (p < 0.05) (Fig. 8B) along with 1.5-fold reduction in CD4 + Roryt + Th17 cells ( Fig. 8C) in LR treated group as compared to Ovx mice ( Fig. 8).
In accordance with earlier reported studies, activated CD4 + T cells are also known to regulate osteoclast activation by expressing RANKL on their surfaces 31 and our results too confirmed that oral supplementation of LR significantly downregulated the expression of RANKL on CD4 + T cells (p < 0.01) (Fig. 8 D) in BM, spleen, LN and PP. Overall, these results demonstrate that upon treatment with LR the proportion of anti-osteoclastogenic T lymphocytes (Treg cells) were significantly enhanced. On the other hand, osteoclastogenic T lymphocytes (Th17 cells) were significantly reduced, thereby establishing the role of LR in modulating Treg-Th17 cell balance in osteoporosis. www.nature.com/scientificreports/ differentiation of CD4 + Rorγt + Th17 cells (fourfold with respect to control group) in a dose dependent manner ( Fig. 9D-F). These results thereby further attest and establish the immunomodulatory role of LR in modulating Treg-Th17 cell balance.
LR administration skews the expression of cytokines in Ovx mice. Cytokines play a key role in the generation and differentiation of osteoclasts viz. pro-inflammatory cytokines (IL-6, IL-17 and TNF-α) which enhances osteoclastogenesis and anti-inflammatory cytokines (IL-4, IL-10 and IFN-γ) which inhibit osteoclastogenesis or enhance bone formation 3,14 . Thus, we next examined the levels of these cytokines in blood serum of mice. Cytokine analysis of blood serum revealed that Ovx mice administered with LR had significantly decreased levels of osteoclastogenic cytokines IL-6 (p < 0.01), IL-17 (p < 0.001) and TNF-α along with significant  Similar results were obtained in two independent experiments with n = 6. Statistical significance of each parameter was assessed by ANOVA followed by paired group comparison. *p < 0.05, **p < 0.01, ***p < 0.001 compared with indicated groups.

Discussion
According to International Osteoporosis Foundation (IOF), one-third of the female and one-fifth of the male population across the globe will suffer from osteoporosis related fractures once in their lifetime 32 . Currently it has been estimated that 200 million people worldwide are suffering from osteoporosis 33 . Most of the therapeutic agents that are being currently employed for treating osteoporosis are not only too costly to provide benefits but are also associated with adverse health effects in the long run 6 . Thus, it is necessary to identify entities with no or minimal side effects that can substitute currently available drugs.
In recent years, growing evidences from both murine and human studies have highlighted the beneficial effects of probiotics (Lactobacillus reuteri, Lactobacillus acidophilus, Lactobacillus casei, Bacillus clausii, Lactobacillus rhamnosus-GG) in treating various disease conditions 13,14,[34][35][36][37][38] . Recently, a study reported by Tyagi et al. 12 showed that administration of Lactobacillus rhamnosus enhanced bone mass in eugonadic mice. We too were interested in studying the impact of LR on bone health under both in vitro and in vivo conditions in Ovxinduced post-menopausal osteoporotic mice model. Strikingly, both our in vitro and in vivo data correlated with previously reported results of Tyagi et al. Our in vitro data clearly indicated that LR-conditioned media exhibits the potential of suppressing RANKL induced osteoclastogenesis in mouse bone marrow macrophages. www.nature.com/scientificreports/ Also, it exhibits the potential of inhibiting F-actin ring formation in osteoclasts; a key phenotype of mature osteoclasts responsible for bone resorptive activities. These findings of ours clearly establish the direct role of LR in modulating bone health via inhibition of osteoclastogenesis. Furthermore, our in vivo results from SEM and AFM proved that LR-administration inhibits bone loss in Ovx mice. μ-CT analysis further reconfirmed that administration of LR significantly attenuated bone loss by maintaining the bone micro-architecture of LV-5, femoral and tibial bones. BMD is an important parameter for assessing fracture prevalence in bones 39 . BMD values narrates the risk of bone breakage or development of fracture, as higher BMD signifies lesser risk of fracture whereas lower BMD value relates to higher risk of bone breakage and fracture development 39 . Notably, our experimental outcomes evidently establish that LR has the potential of significantly enhancing bone health by maintaining BMDs of LV-5, femoral and tibial bones. In 2016, Boskey et al. 30 reported that loss of bone heterogeneity is associated with enhanced brittleness of bone that in turn determines the prevalence of fracture risk. Of note, the physiological heterogeneity of bones should be maintained for their proper functioning 30 . Our results for the first-time report enhancement in heterogeneity of bone samples in LR treated group in comparison to Ovx mice. These data clearly suggest that LR modulates bone health without compromising the heterogeneity of bones thereby confirming LR as a better therapeutic option than the most of currently available anti-osteoporotic drugs (eg. bisphosphonates) that compromise bone heterogeneity (i.e. enhanced homogeneity) 40 and thus enhance fracture risk in long run.
Although our in vitro data clearly suggested the direct osteoprotective role of LR but we cannot exclude the fact that probiotics exhibits immunomodulatory properties. Also, previous studies from our group has shown that treatment with probiotic Bacillus clausii 13 and Lactobacillus acidophilus 14 enhanced the femoral and tibial bone micro-architecture, bone mineral content via maintaining the homeostatic Tregs and Th17 cell balance in Ovx mice. Study by Tyagi et al. 12 reported that LR regulates bone mass by stimulating the production of butyrate in mice but unfortunately the immunomodulatory potential (specifically the Treg-Th17 cell axis) of this probiotic was not reported. Building upon these previous evidences, in the present study we elucidated the immunomodulatory properties through which LR inhibits bone loss in Ovx mice. Both bone cells and immune cells shares a common niche (i.e. bone-marrow) during their development, an active field of research termed as Osteoimmunology. Among immune cells, Tregs and Th17 lymphocytes are the key players involved in regulating the bone remodelling process. Cytokines derived from these immune cells such as IL-10 and IL-17 have established roles in regulating osteoclastogenesis 3,[40][41][42][43] . In this context, we too studied the effect of probiotic LR on Tregs-Th17 Figure 10. LR skews cytokines balance in Ovx mice. Osteoclastogenic and anti-osteoclastogenic cytokines were analysed in serum samples of mice by ELISA/CBA. The results were evaluated by using ANOVA with subsequent comparisons by Student t-test for paired or non-paired data, as appropriate. Values are expressed as mean ± SEM (n = 6) and similar results were obtained in two independent experiments. Statistical significance was defined as p ≤ 0.05, *p ≤ 0.05, **p < 0.01 ***p ≤ 0.001 with respect to indicated mice group. www.nature.com/scientificreports/ cells in BM (prime site of osteoclastogenesis), spleen, PP, LN and our data clearly suggest that LR exerts systemic bone effects via enhancing Tregs population along with simultaneous reduction in Th17 cells. These results were further supported by our serum cytokine data where we observed significant enhancement of anti-osteoclastogenic cytokines such as IL-4, IL-10 and IFN-γ 10 and down regulation of osteoclastogenic cytokines such as IL-6, IL-17 and TNF-α. Since, our in vitro results clearly pointed to the inhibitory role of LR-CM on RANKL induced osteoclastogenesis we thus investigated the source of RANKL in vivo. One of the major immunological source of RANKL are T cells 31 . Thus, we determined the expression of RANKL on T cells in various lymphoid tissues viz. BM, Spleen, PP and LN. Surprisingly, we found that LR-administration significantly inhibited the expression of RANKL on T cells. These in vivo results are thus in concurrence with our in vitro results wherein LR-CM inhibited RANKL induced osteoclastogenesis. From these data, we observed that LR significantly enhanced the percentage of Tregs along with simultaneously reducing Th17 cells. These results of ours thus clearly establish both direct and indirect link between "LR-Bone-T cells". In summary, our data clearly establish that LR enhances bone health via maintaining the homeostatic balance between Tregs-Th17 cells (Fig. 11). Our present study thus for the first time demonstrates the beneficial effects of LR on skeletal health by maintaining the BMD, along with preserving bone micro-architecture and heterogeneity of bones in bilaterally induced ovariectomized mice model via modulating the homeostatic balance of Treg-Th17 cell axis in the host. The present study thus highlights the potential of probiotic LR to be used as a novel osteoprotective agent in the treatment and management of bone related diseases including osteoporosis.
Cell viability or metabolic activity assay. MTT assay was performed to assess the cell viability as a function of cellular metabolic activity. Briefly, BMMs were seeded in 96 well plate (10, 000 cells/well) and incubated for 24 h in CO 2 incubator at 37 °C. On the following day, cells were treated with LR-CM at different ratios and further incubated for 48 h. On completion of treatment, MTT was added (5 mg/ml) and plate was incubated in CO 2 incubator at 37 °C for 4 h. After incubation, formazan crystals were dissolved by adding DMSO. The plate was shaken for 5 s (orbital shaking) and reading was taken at 570 nm on a microplate reader (Synergy H1, BioTek).
Flow cytometry. Cells were harvested and stained with antibodies specific for Treg and Th17 cells. For surface marker staining, cells were first incubated with anti-CD4-PerCPcy5.5 and incubated for 30 min in dark on ice. After washing, cells were fixed and permeabilized with 1× fixation-permeabilization-buffer for 30 min on ice in dark. Further, cells were stained with anti-Rorγt-PE and anti-Foxp3-APC for 45 min. After washing cells were acquired on BD LSR II (USA). Flowjo-10 (TreeStar, USA) software was used to analyse the samples and gating strategy was done as per previously defined-protocols 13 .
Scanning electron microscopy (SEM). SEM for femur cortical region of bones was done as described previously 13,14,29 . Briefly, bone samples were stored in 1%-Triton-X-100 for 2-3 days and later bone samples were transferred to 1XPBS buffer till the final analysis was performed. After preparation of bone slices, samples were dried under incandescent lamp and sputter coating was performed. Subsequently bones were scanned in Leo 435-VP microscope equipped with digital imaging with 35 mm photography system. SEM images were digitally photographed at 100× magnification to capture the best cortical regions. The SEM images were further analysed by MATLAB software (MathWorks, Natick, MA, USA).

Atomic force microscopy (AFM). After drying femur bone samples completely in sterile environment
with 60 W lamps for 6 h followed by high vacuum drying. Samples were prepared as per requirement for the machine and analysed by Atomic Force Microscope (INNOVA ICON, Bruker) that works under the Acoustic AC mode. This was assisted by cantilever (NSC 12(c) MikroMasch, Silicon Nitride Tip) and NanoDrive version 8 software. It was 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 air at room temperature. Images were later processed and analysed by using nanoscope analysis software. Further, the 3D AFM images were also analysed by MATLAB software (MathWorks, Natick, MA, USA).
Micro-computed tomography (µ-CT) measurements. µ-CT scanning and analysis was performed as described previously 13,14,29 using SkyScan 1076 scanner (Aartselaar, Belgium). Briefly, after positioning all samples at right orientation, scanning was done at 50 kV, 201 mA using 0.5 mm aluminium filter and exposure was set to 590 ms. NRecon software was used for carrying out reconstruction process. For trabecular region analysis, ROI was drawn at a total of 100 slices in secondary spongiosa at 1.5 mm from distal border of growth plates excluding the parts of cortical bone and primary spongiosa. The CTAn software was used for measuring and calculating the micro architectural parameters in bone samples. Various 3D-histomorphometric parameters were obtained such as: BV/TV (Bone volume/Tissue volume), Tb.Th (Trabecular-thickness), Tb.Sp. (Trabecularseparation) etc. The volume of interest of u-CT scans made for trabecular and cortical regions were 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/cm 3 and 0.75 g/cm 3 ) as calibrator 13,14,29 . Fourier transform infrared (FTIR) spectroscopy. Femur cortical bone samples were stored in 1% Triton X-100 for 24 h before being dried with 60 W lamps for 6 h followed by high vacuum drying. Dried bone samples were crushed in mortar-pestle, thereafter bone samples were mixed with potassium-bromide (KBr) at (1:100) ratio for FTIR-analysis. Further, acquisition was performed by using 8400S-FTIR-(SHIMADZU), with a resolution 4 cm −1 ; scan speed 2.5 kHz and 128 scans. The samples were clearly positioned with a prism made of highly refractive material. Savitzky Golay algorithm was used to nullify background noise for obtaining smooth spectra of all analysed-samples 13,14,29 . Enzyme linked immunosorbent assay (ELISA) and cytometric bead array (CBA). ELISA was performed for quantitative assessment of cytokines viz. IL-4, IL-6, IL-10, IL-17 and TNF-α in blood serum using commercially available kits as per the manufacturer's instructions. For estimating IFN-γ cytokine in serum, CBA was performed as per the manufacturer's instructions (BD-Biosciences). Fluorescent signals were read on flow analyser and data analysed by BD FCAP-Array software (BD-Biosciences, USA).
Statistical analysis. Statistical differences between groups were assessed by using analysis of variance (ANOVA) with subsequent comparisons via student t-test paired or unpaired as appropriate. All the data values are expressed as Mean ± SEM (n = 6). Statistical significance was determined as p ≤ 0.05 (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001) with respect to indicated group.
Ethical approval. All applicable institutional and/or national guidelines for the care and use of animals were followed.