Kae (purity>98%) was purchased from the National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China). Kae was dissolved in DMSO and diluted to 0.01% in PBS. Rapamycin (Rapa), a specific inhibitor of mTOR, was purchased from Selleck Chemicals. The Alizarin Red S (ARS) staining buffer and alkaline phosphatase (ALP) detection kits were purchased from Nanjing Jiancheng Bioengineering Institute. Anti-runt-related transcription factor 2 (Runx2; cat. no. ab23981) and anti-Osterix (cat. no. ab22552) was purchased from Abcam. Anti-eukaryotic translation initiation factor 4E-binding protein 1 (4E/BP1; cat. no. 94525), anti-phosphorylated (p)-4E/BP 1 (cat. no. 2855) and anti-ribosomal protein S6 kinase B1 (S6K1; cat. no. 9202), anti-p-S6K1 (cat. no. 9204) was obtained from Cell Signaling Technologies, Inc. Horseradish peroxidase-labeled anti-immunoglobulin G secondary antibody (goat anti-mouse lgG; cat. no. SA00001-1; and goat anti-rabbit lgG; cat. no. SA00001-2) was obtained from ProteinTech Group, Inc. Anti-β-actin antibody (cat. no. KL002) was provided by Nanjing Jiancheng Bioengineering Institute.

A total of 30 adult (age, 6–8 weeks) female Sprague-Dawley (SD) rats weighing 180–220 g were obtained from the Nanchang University Laboratory Animal Center (Nanchang, China) and maintained under a 12-h dark/light cycle at 22–25°C and 40–70% humidity. Animals were allowed access to food and water ad libitum. Experiments were performed according to the Guide for the Care and Use of Laboratory (National Institutes of Health), and were approved by the Ethics Committee of Nanchang University (no. 2017-0122).

A total of 30 rats were randomly divided into five groups (n=6 in each group) as follows: i) Sham group, in which the abdominal cavities of the rats were opened and fat tissue around the ovaries removed; ii) OVX group, in which the rats were ovariectomized; iii) Kae treatment group (OVX + Kae), in which the OVX rats were continuously given Kae (100 mg/kg/day) via gavage for 8 weeks; iv) Kae and Rapa administration group (OVX + Kae + Rapa), in which the rats were treated as in the Kae treatment group and in addition received intraperitoneal injections of Rapa (0.2 mg/kg/day) for 8 weeks; and v) Rapa group (OVX + Rapa), in which the OVX rats received intraperitoneal injections of Rapa (0.2 mg/kg/day) for 8 weeks. The dose and timing of Kae and Rapa administration were ascertained in preliminary experiments. Briefly, the OVX rats were given Kae at 25, 50 and 100 mg/kg/day. Bone mineral density (BMD) was detected after 8 weeks to select the working concentration. Kae at 100 mg/kg/day was the most effective concentration in improving BMD in OVX rats. Rapa was administered at 0.2 mg/kg/day according to the manufacturer's protocol (Selleck Chemicals). Subsequent experiments were performed following the 8 weeks of treatment.

Following treatment, the rats were sacrificed and their right femurs and tibias were dissected. The 2D total bone mineral content was used to calculate the BMD as previously described (19). The right femurs of the rats were analyzed using dual-energy X-ray absorptiometry with a Lunar Prodigy Advance system (version 13.6; GE Healthcare).

Based on the median values of total BMD, selected trabecular microarchitecture of the femoral metaphysis was evaluated using micro-computed tomography (micro-CT; Scanco Medical AG) with Scanco image processing language software (version 5.08b; Scanco Medical AG). Scans were performed from the proximal growth plate in the distal direction (18 µm/slice) as the distal femur has a high concentration of trabecular bone compared with the proximal and middle regions of the femur. A volume of interest (VOI) was selected, defined as the cross-sectional area spanning 100 slices from the proximal growth plate. The 2D scans were used to produce 3D reconstructions of the bone microarchitecture, which were used to measure bone morphometric parameters of the selected VOI, including the bone volume fraction [bone volume/tissue volume (BV/TV)], trabecular number (Tb.N), trabecular separation (Tb.Sp), trabecular thickness (Tb.Th) and structure model index (SMI). The operator conducting the CT analysis was blinded to the treatments associated with the specimens. All examinations were conducted according to the principles and procedures described in the most recent National Research Council publication of the Guide for the Care and Use of Laboratory Animals and the ARRIVE guidelines (20).

BMSCs were flushed from the femurs and tibias of normal SD and OVX rats with PBS in a biosafety cabinet, using a 5 ml syringe fitted with a needle (21G). Mononucleated cells were isolated by density gradient centrifugation at 400 × g for 20 min at room temperature in rat lymphocyte separation medium (TBD Science) at a concentration of 1.091 g/ml. Isolated cells were cultured in low-glucose DMEM (HyClone; GE Healthcare Life Sciences) containing 15% fetal bovine serum (HyClone; GE Healthcare Life Sciences). All cells were maintained in a 37°C incubator within an atmosphere containing 5% CO₂. BMSCs were identified by CD44 and CD34 labeling (21). The culture medium was regularly replaced and cells were passaged after reaching 70–80% confluence. Third-generation BMSCs were harvested for subsequent experiments.

The cytotoxicity of BMSCs was determined using an MTT assay. Briefly, rat BMSCs were plated into 96-well plates at a density of 1×10⁵ cells/well. After 24 h of culture in a 5% CO₂ incubator at 37°C, adherent cells were treated with a range of Kae concentrations (0.1, 1, 10 and 100 µM) for 24 h. The MTT reagent (10 µl; 10 mg/ml) was subsequently added to each well, and cells were incubated for a further 4–6 h in a 5% CO₂ incubator at 37°C. The medium was removed and 100 µl DMSO was added to each well. Plates were shaken for 10 min, and the absorbance was measured at 492 nm using a Spectra Max Paradigm microplatereader (Molecular Devices, LLC). Cytotoxicity was calculated as follows: Cytotoxicity (%)=(1-absorbance of sample/absorbance of control) ×100.

Third-passage BMSCs were randomly divided into 5 groups: i) Control group, in which BMSCs derived from normal rats were treated with osteogenic induction medium (Cyagen Biosciences, Inc.) to induce osteoblast differentiation; ii) OVX group, in which BMSCs derived from OVX rats were treated in the same way as the control group; iii) OVX + Kae group, in which BMSCs derived from OVX rats were incubated with 0.1, 1, 10 or 100 µM Kae and treated in the same way as the control group; iv) OVX + Kae + Rapa group, in which BMSCs derived from OVX rats were incubated with Kae and 10 µM Rapa and then treated in the same way as the control group; v) OVX + Rapa group, where BMSCs derived from OVX rats were incubated with 10 µM Rapa and treated in the same way as the control group. The dose of Rapa was determined during preliminary experiments. Briefly, MTT was detected to evaluate the optimal concentration of Rapa. Rapa (at 0.1, 1, 10 or 100 µM) was incubated with BMSCs. Rapa was found to exhibit a concentration-dependent cytotoxic effect. Therefore, 10 µM was selected. BMSCs were harvested at day 15 in all groups.

ARS staining was used to assess the osteogenic differentiation of BMSCs. Briefly, rat BMSCs were plated into 24-well plates at a density of 2.5×10⁵ cells/well. After induction of osteogenesis, cells were fixed with 95% alcohol for 15 min at room temperature and the BMSCs were stained with ARS for 5 min at room temperature. Mature osteoblasts that differentiated from rBMSCs displayed intense brown-red staining after 5 min of ARS staining (magnification, ×40). Cells were imaged at ×400 magnification by light microscopy, after the addition of 10% (w/v) cetylpyridinium chloride to precipitate calcium ions. The absorbance of each well was measured at 562 nm using a Spectra Max Paradigm microplatereader (Molecular Devices, LLC).

The effects of Kae on osteoblasts were assessed using an ALP assay kit. The absorbance/optical density (OD) of each sample was measured at 492 nm using a microplate reader. ALP activity in the osteogenic induction medium (Cyagen Biosciences, Inc.) was determined as follows: ALP activity (Jinshi unit/100 ml)=(T-B)/(S-B)×0.02 mg/ml ×100 ml × a, where T=OD of the test sample; B=OD of the blank; S=OD of the standard; and a=dilution factor of the sample. According to the conditions above, a Jinshi unit is defined as the ALP activity that releases of 1 mg of phenol every 15 min after mixing with 100 ml liquid matrix at 37°C.

Cells were lysed in RIPA lysis buffer containing 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1% sodium deoxycholate, 1% NP-40, 1 mM PMSF and 1 mM EDTA. Then, extracts were centrifuged at 24,750 × g at 4°C for 15 min to remove insoluble material. Total protein concentrations were determined using a bicinchoninic acid assay (Beyotime Institute of Biotechnology), according to the manufacturer's protocol. An equivalent quantity of total protein (40 µg) per well was diluted in sample buffer containing 100 mM dithiothreitol and heated to 98°C for 5 min. Lysates were separated using 10–15% SDS-PAGE (Bio-Rad Laboratories, Inc.) and subsequently transferred to PVDF membranes. The membranes were blocked in 5% non-fat dry milk for 2 h at room temperature and incubated overnight at 4°C using primary antibodies (1:500) against β-actin, Runx2, 4E/BP 1, p-4E/BP1, S6K1 and p-S6K1. Membranes were washed in TBS with Tween-20 (TBS-T), and incubated with horseradish peroxidase-conjugated secondary antibody (1:2,000) at room temperature for 1 h. Membranes were washed three times for 20 min in TBS-T, and protein bands were visualized by enhanced chemiluminescence (Proteintech Group, Inc.). Band intensities were measured and quantitated using Quantity One software (version 4.6.6; Bio-Rad Laboratories, Inc.) and β-actin was used for normalization.

SPSS (version 20.0; IBM Corp.) was used for statistical analysis. Data are presented as the mean ± SEM from six independent experiments. The variance homogeneity test and one-way ANOVA were performed between groups. Newman-Keuls test was used following ANOVA. P<0.05 was considered to indicate a statistically significant difference.

The 3D trabecular bone microarchitecture, used for the assessment of distal femoral metaphysis, was calculated from micro-CT images (Fig. 2). The mean BV/TV, Tb.N, Tb.Th and BMD values in the OVX group were significantly lower than those of the sham group, while Tb.Sp and SMI were higher (P<0.05). Compared with the OVX group, BV/TV, Tb.N, Tb.Th and BMD values were significantly greater, while Tb.Sp and SMI were lower in the OVX + Kae group (P<0.05). Thus, treating OVX rats with Kae significantly improved bone microarchitecture and mass. The opposite trend was observed in the OVX + Kae + Rapa and OVX + Rapa groups: OVX rats treated with Kae + Rapa or Rapa only exhibited lower BV/TV, Tb.N and Tb.Th values and higher Tb.Sp and SMI values compared to the OVX + Kae group (P<0.05 vs. OVX + Kae).

The cytotoxicity of various concentrations of Kae was tested on BMSCs. As shown in Fig. 3, Kae displayed a concentration-dependent cytotoxic effect, demonstrating high cytotoxicity at 100 µM. Additionally, it was observed that Kae induced osteoblast differentiation (Fig. 4). On day 15 following osteogenic induction, ARS staining revealed increased calcium deposition with increasing Kae concentrations. However, the number of osteoblasts unexpectedly decreased when cells were treated with 100 µM Kae. From the range of concentrations tested, the optimal concentration of Kae that promoted osteoblast differentiation was 10 µM. This concentration was therefore used in subsequent experiments.

ARS staining was used to assess the induction of osteogenesis in BMSCs. As illustrated in Fig. 5, following the induction of osteogenic differentiation, the rate of calcific nodule formation decreased in the OVX group compared to the control group (P<0.05). Moreover, treatment with Kae significantly increased the number of calcified nodules in the OVX + Kae group (P<0.05 vs. OVX group). However, the protective effect of Kae on osteoblasts was antagonized by the co-administration of Rapa (P<0.05 vs. OVX + Kae).

ALP activity was measured during osteogenesis. As shown in Fig. 6, the expression of ALP in the OVX group was significantly reduced compared with the control group (P<0.05). In addition, ALP activity was significantly higher in the OVX + Kae group compared to the OVX group (P<0.05 vs. OVX group). Pretreatment with Kae + Rapa or with Rapa alone led to reduced levels of ALP activity compared to Kae treatment alone (P<0.05).

Western blot analysis was used to assess the expression of the osteogenesis-related transcription factors Runx2 and Osterix during osteogenesis. Fig. 7 demonstrates that a significant downregulation of Runx2 and Osterix expression occurred in the OVX group compared with the control group (P<0.05). Kae significantly increased the expression of Runx2 and Osterix compared with the OVX group (P<0.05). However, their expression was significantly lower in the OVX + Kae + Rapa and OVX + Rapa groups compared with the OVX + Kae group (P<0.05).

Western blot analysis was used to determine the levels of phosphorylated and total 4E/BP1 and S6K1, important downstream regulators of the mTOR pathway. As shown in Fig. 8, the levels of p-4E/BP1 in the OVX + Kae group were significantly lower than in the OVX group (P<0.05). However, an notably increased level of 4E/BP1 p-4EB/P1 was observed in the OVX + Kae + Rapa and OVX + Rapa groups. In addition, Kae increased S6K1 and p-S6K1 levels (Fig. 9; P<0.05 vs. OVX group). Conversely, the levels of S6K1 and p-S6K1 were significantly lower in the OVX + Kae + Rapa and OVX + Rapa groups compared with the OVX + Kae group (P<0.05). Taken together, these results indicated that Kae may promote osteogenesis through mTOR signaling as 4E/BP1 was activated by mTOR, whereas S6K1 exhibited the opposite trend.