Gentiopicroside promotes osteogenesis and prevents bone loss in ovariectomized mice by modulation of β-catenin-BMP2 signaling pathway

Background: Osteoporosis is characterized by increased bone fragility and the drugs used at present to treat osteoporosis can cause adverse reactions. Gentiopicroside (GEN) is widely used in the traditional Chinese medicine thanks to its multiple pharmacological activities including the suppression of osteoclastogenesis. Therefore, the aim of this work was to explore the effect of GEN on bone mesenchymal stem cells (BMSCs) osteogenesis for a potential osteoporosis therapy. Methods: In vitro, BMSCs were exposed to GEN at different doses for 2 weeks, while in vivo, ovariectomized osteoporosis was established in mice and the therapeutic effect of GEN was evaluated for 3 months. Results: Our results in vitro showed that GEN promoted the activity of alkaline phosphatase, increased the calcied nodules in BMSCs and upregulated the osteogenic factors (Runx2, OSX, OCN, OPN, and BMP2). In vivo, GEN strengthened the secretion of Runx2, OCN, and BMP2, increased the level of osteogenic parameters, and accelerated the osteogenesis of BMSCs by activating the BMP pathway and Wnt/β-catenin pathway, effect that was inhibited using the BMP inhibitor Noggin and Wnt/β-catenin inhibitor DKK1. Silencing the β-catenin gene and BMP2 gene blocked the osteogenic differentiation induced by GEN in BMSCs. This block was also observed when only β-catenin was silenced, while the knockout of BMP2 did not affect β-catenin expression induced by GEN. Interpretation. Conclusions: GEN promotes BMSC osteogenesis by regulating β-catenin-BMP signalling, providing a novel strategy in the treatment of osteoporosis.


Background
Osteoporosis is a systemic disease characterized by bone loss, destruction of the bone microstructure and increased bone fragility, often leading to brittle fractures [1]. A large number of patients with osteoporosis not only suffer from severe pain, but they are also subjected to a heavy nancial burden [2].
In particular, the bone resorption rate of postmenopausal women is signi cantly higher than that of osteogenesis, leading to a serious bone loss [3]. With the development of aging population, the number of postmenopausal patients with osteoporosis is increasing year by year, causing a signi cant impact on the medical community and the whole society [4]. At present, the drugs used to treat postmenopausal osteoporosis play a role mainly by inhibiting bone absorption and promoting bone formation [5,6].
However, long-term use of anti-osteoporosis drugs can cause a series of adverse reactions, including myasthenia gravis, in uenza like diseases, and gastrointestinal tumours [7]. Thus, it is of utmost importance to nd a new alternative therapy to cure osteoporosis. The activity of osteoclasts and osteoblasts needs to be precisely coordinated to maintain skeletal integrity [8]. Osteoclasts are multinucleated giant cells derived from monocytes/macrophages, whose main function is to promote bone resorption [9]. Osteoblasts are mainly differentiated from bone mesenchymal stem cells (BMSCs) and deposited in the calci ed bone matrix, which has a signi cant impact on the formation of a new bone [10]. The primary cause of osteoporosis is due to the decrease in osteoblasts that leads to a reduced bone formation, and the increase in osteoclasts resulting in an increased osteolysis. Recently, several reports pointed out that stimulating osteoblast differentiation may be an effective way to prevent and treat osteoporosis [11][12][13].
BMSCs are stem cells with a multi-directional differentiation, since they can differentiate into several cell types including chondrocytes, osteoblasts, adipocytes, and endothelial cells under speci c conditions [14]. BMSCs rst differentiate into precursor osteoblasts, then into osteoblasts, and nally they gradually form the mature osteoblasts [15].However, the osteogenic capability of BMSCs gradually decreases with the increase of old age, while the adipogenic capacity of BMSCs increased, leading to the downregulation of bone formation and nally osteoporosis [16]. During the osteogenic differentiation, BMSCs release a series of osteogenic factors, including osteocalcin (OCN), osterix (OSX), osteopontin (OPN), and runtrelated transcription factor 2 (Runx2), which accelerate the maturation of osteoblasts [16,17]. Therefore, BMSCs may be a suitable cell source for studying osteogenesis.
The Wnt pathway and BMP pathway occupy a critical position in the modulation of osteoblast differentiation [18,19]. When the frizzled transmembrane receptor binds to LRP5 and/or LRP6, it can induce the secretion of Wnts and activates the canonical Wnt pathway [20]. Subsequently, β-catenin is released and transcribed into the nucleus to regulate the generation of osteogenic markers [21]. The upregulation of BMP pathway induces the phosphorylation of Smad proteins [22]. Then, the activated Smad proteins actively regulate the transcription of osteogenic factors (Runx2 and OCN), thus promoting osteogenic differentiation [23,24].
Gentiopicroside (GEN) is the active component of the plant Gentiana Manshurica Kitag, widely used in the traditional Chinese medicine thanks to its multiple pharmacological activities, such as antioxidative, antinociceptive, anti-in ammatory, antibacterial, and antiosteoporotic effects [25][26][27]. Osteoblast differentiation and osteoclastogenesis are two important phases of bone remodelling [28,29] and GEN represents a potential drug in the treatment of osteoporosis by suppressing osteoclastogenesis [30].
Based on the above studies, our hypothesis is that GEN could induce osteogenic differentiation of BMSCs in vitro and bone formation in vivo.
Therefore, the purpose of this study was to nd whether GEN could promote the osteogenic differentiation of BMSCs and explore the molecular mechanism induced by GEN in differentiating BMSCs.

Effect of GEN on BMSC proliferation
The chemical structure of GEN is shown in Fig. 1a. The results suggested that the proliferation of BMSCs treated with GEN (10-40 µM) was not signi cantly changed. However, 80 µM GEN signi cantly inhibited the proliferation of BMSCs (Fig. 1b). Thus, GEN was not harmful to BMSCs at the concentrations of 10, 20, and 40 µM.

GEN strengthens the osteogenic differentiation in BMSCs
The effect of GEN alone on BMSCs was tested as rst. The results showed that the osteogenic differentiation of the BMSCs treated with GEN for 14 days without osteogenic medium was not signi cantly affected ( Supplementary Fig. 1). Next, the osteogenic differentiation of BMSCs treated with GEN for two weeks under osteogenic induction was explored. The results showed that the higher concentration of GEN, the higher ALP activity (Fig. 1c, e) and the number of mineralized nodules (Fig. 1d, f) compared with the control group. In addition, the mRNA expression of the osteogenic genes (ALP, Runx2, OSX, OCN, OPN, and BMP2) was signi cantly increased by GEN, and the increasing trend was concentration-dependent ( Fig. 2a-f). Similarly, the expression of osteogenic proteins (Runx2, OCN, and BMP2) were also promoted by GEN and reached a peak at 40 µM ( Fig. 3a-d). Therefore, our results revealed that GEN strengthened the osteogenic differentiation in BMSCs.

GEN promotes bone formation in OVX mice
The OVX mouse model was used to con rm these in vitro results. The OVX osteoporosis mouse model is the most commonly used animal model in studying postmenopausal osteoporosis. Ovariectomy can cause bone loss acceleration and cortical bone formation reduction, which are closely related to oestrogen de ciency [31,29]. To test the effect of the ovariectomy, the body weight and the mass of the uterus were measured ( Supplementary Fig. 2). The results showed that the body weight of the OVX group and OVX + GEN group was greater than that of the Sham group ( Supplementary Fig. 2a). In contrast, the mass of the uterus in the OVX group and OVX + GEN group was less than that in the Sham group ( Supplementary Fig. 2b). The histopathological images of all groups (sham group, OVX group, and OVX + GEN group) are shown in Fig. 4. HE staining results showed that the number of bone trabeculae in the OVX group was signi cantly less than that in the Sham group, whereas the number of bone trabeculae in the OVX + GEN group was higher than that in the OVX group, but no signi cant difference was observed between the Sham group and OVX + GEN group (Fig. 4a). The results of micro-CT showed that the BMD, the number of trabeculae and the thickness of the trabeculae in OVX + GEN group were higher than those in the OVX group, while the same parameters in the OVX group were remarkably lower than those in the Sham group (Fig. 4b-e). However, no statistical difference was found between OVX + GEN group and Sham group.
Subsequently, immunohistochemistry was used to detect the secretion of osteogenic proteins (Runx2, OCN, and BMP2) in vivo. Runx2 expression in the OVX group was less than that in the Sham group, while its expression in the OVX + GEN group was higher than that in the OVX group (Fig. 5a). Similar to the results of Runx2, the expression of OCN (Fig. 5b, e) and BMP2 (Fig. 5c, f) in OVX group was lower than that in the Sham group and OVX + GEN group. Thus, our results demonstrate that GEN effectively promoted osteogenesis in OVX osteoporotic mice and showed a good anti-osteoporotic effect. BMP pathway and Wnt/β-catenin pathway activated by GEN in BMSCs Since the BMP signalling [32] and WNT signalling [33] pathways are related to osteogenesis, the consequence of GEN treatment on the signalling of BMP and WNT/β-catenin was evaluated. Without osteogenic induction, GEN did not alter the expression of p-Smad1/5/8 and β-catenin in BMSCs ( Supplementary Fig. 3). Next, the effect of GEN on the osteogenic mechanism in BMSCs was further explored. After the treatment of BMSCs with GEN for 2 weeks under osteogenic conditions, the results showed that GEN upregulated the expression of p-Smad1/5/8 and β-catenin in a dose-dependent manner ( Fig. 6a, b, c). However, the osteogenesis-potentiating effect of GEN on p-Smad1/5/8 and β-catenin was abolished by the treatment with the BMP pathway inhibitor Noggin (Fig. 7a-c) and the Wnt/β-catenin pathway inhibitor DKK1 ( Fig. 7d-f). Therefore, our results reveal that GEN promoted the osteogenic differentiation via BMP signalling and WNT/β-catenin signalling.

GEN-induced osteogenic differentiation is a β-catenin-BMP2-dependent effect
To further reveal the speci c mechanism of GEN in regulating of BMP signalling and Wnt/β-catenin signalling, gene silencing was performed in vitro. The transfection with Ad-Cre e ciently silenced the βcatenin (Fig. 8a) and BMP2 (Fig. 8g) gene in BMSCs. β-catenin silencing in BMSCs signi cantly inhibited the increase of Runx2, OSX, OCN, OPN, and BMP2 induced by GEN ( Fig. 8b-f). In addition, the increase of Runx2, OSX, OCN, and OPN ( Fig. 8h-k) induced by GEN was inhibited by silencing the BMP2 gene, while the expression of β-catenin was not affected (Fig. 8l). Our data revealed that the silencing of β-catenin gene prevented GEN-mediated upregulation of BMP2, although GEN-induced β-catenin enhancement was not in uenced by silencing the BMP2 gene. The above results indicated that GEN rst activates the βcatenin pathway and then the BMP2 pathway. Therefore, GEN strengthened the osteogenic ability of BMSCs through the β-catenin-BMP2 signalling pathway.

Discussion
GEN is considered an effective drug in the treatment of osteoporosis by preventing the formation of osteoclast [30]. The inhibition of osteoclast or the activation of osteogenesis exerts a signi cantly protective effect on osteoporosis [34]. The current study demonstrated for the rst time the effect and potential mechanism of GEN in the osteogenic process of BMSCs both in vitro and in vivo. BMSCs were used to explore the effect of GEN in osteogenesis in vitro. GEN did not show any toxicity on BMSCs at the doses used. In addition, GEN enhanced the expression of ALP and improved the mineralized nodules, increased the expression of osteogenic factors in a concentration dependent manner. Finally, the in vivo results showed that GEN accelerated the osteoid formation and mineralization in the mouse femur. All these results indicated that GEN could stimulate osteogenesis in vitro and in vivo. Our results showed that GEN strengthened the osteogenic ability of BMSCs in vitro and stimulated bone ossi cation in vivo by the upregulation of the β-catenin-BMP2 signalling pathway. Therefore, GEN could be potentially considered a novel compound in regulating bone metabolism, since it could not only promote osteogenesis but also inhibit bone absorption.
BMP pathway and Wnt pathway occupy an critical position in determining the direction of osteogenesis in BMSCs [35,36]. The activation of the BMP pathway induces the phosphorylation of Smad1/5/8. Then, the phosphorylated Smad1/5/8 bind to Smad4 and move into the nucleus, thus activating downstream factors of the BMP pathway [37]. Additionally, the Wnt/β-catenin signalling is critical in the therapy of osteoporosis, since it has a signi cant impact on promoting osteogenesis and regulating bone metabolism [38]. When the Wnt/β-catenin signalling is upregulated, it can effectively promote the transformation of the precursor osteoblasts into osteoblasts, so as to actively regulate the formation of new bone and improve the structure of the bone itself [39]. After Wnt pathway activation, the β-catenin is e ciently transcribed into the nucleus, thus stimulating the production of downstream factors [40]. Our results revealed that GEN activated both the BMP pathway and Wnt/β-catenin pathway in BMSCs. It also demonstrated that Noggin [41], and DKK1 [42], signi cantly inhibited GEN-induced osteogenesis in BMSCs. Therefore, GEN stimulated the osteogenic ability of BMSCs through both the two signalling pathways.
To reveal the sequential order in the activation of the BMP pathway and Wnt pathway by GEN, β-catenin or BMP2 were silenced in vitro, resulting in a suppression of GEN-mediated osteogenesis after the silencing of both genes. In addition, the silencing of β-catenin completely suppressed GEN-mediated BMP2 expression, whereas the silencing of BMP2 could not in uence GEN-mediated β-catenin expression in BMSCs. Thus, BMP2 was a target factor in the downstream of the β-catenin pathway in BMSCs. Overall, our results demonstrated that GEN stimulated the ossi cation of BMSCs by the activation of βcatenin-BMP2 pathway.
Taken together the results in the current study, a model illustrating the potential mechanism used by GEN to promote the osteogenic differentiation of BMSCs could be proposed. GEN stimulates osteogenesis by increasing ALP, Runx2, OSX, OCN, and OPN though the activation of the Wnt/β-catenin-BMP2 signalling, thereby promoting the differentiation of BMSCs into osteoblasts (Fig. 9).

Conclusions
In conclusion, this study provided novel insights into the effect of GEN on BMSCs osteogenic differentiation and its protective effect against bone loss. Although further studies are required to con rm these results, GEN might represent a promising approach in the treatment of osteoporosis.

Materials And Methods
This project was carried out with the permission of the Ethics Committee of the Ningbo No. 6 Hospital (registered number 2015-018).

Cell culture and treatments
Five female C57BL/6 mice (4-week-old) were used and euthanized to obtain BMSCs cultured according to a previous work [43]. In brief, the mouse femur was collected and the surrounding soft tissues were removed. Then, the bone marrow in the femur was washed three times with α-MEM (Sigma-Aldrich). The obtained bone marrow content was placed in a dish containing a complete medium consisting of a medium supplemented with 100 U/ml penicillin, 100 mg/ml streptomycin sulphate, and 10% FBS (Sigma-Aldrich). BMSCs were incubated at 37 °C in a 5% carbon dioxide incubator. The third generation of BMSCs was used for our experiments. When the con uence reached 70%, the osteogenic medium was added to allow the osteogenic differentiation. The osteogenic medium consisted of complete medium supplemented with 0.1 mM dexamethasone (Sigma-Aldrich), 5 mM β-glycerophosphate (Sigma-Aldrich), and 100 mM ascorbic acid (Sigma-Aldrich). Then BMSCs were cultured for 14 days under the osteogenic environment, changing the medium every 3 days.

Cell cytotoxicity test
BMSCs (1.0 × 10 3 cells/well) were seeded into 96-well plates and cultured for 24 hours. Then, GEN at different concentrations (0, 10, 20, 40, and 80 µM) was added, and the cells were cultured for 7 days. At the end of the 7 days, the Cell Counting Kit-8 (CCK-8; Sigma-Aldrich) was used to evaluate the cytotoxicity of GEM The absorbance was measured at 450 nm according to the manufacturer recommendations. A previous study described this method in detail [31].
Alkaline phosphatase activity measurement BMSCs (1.0 × 10 5 per well) were seeded into 6-well plates, and then treated with GEN at different concentrations (0, 10, 20, and 40 µM) for 14 days under osteogenic environment. Next, BMSCs were washed with medium for three times and lysed by ultrasound. The concentration of the lysate protein was measured by the Bradford protein test (Thermo Fisher Scienti c, Waltham, MA). Alkaline phosphatase (ALP) activity was detected using p-nitrophenyl-phosphate in AMP buffer (Sigma-Aldrich) at room temperature for 20 minutes. Then, sodium phosphate (0.3 M, pH 12.3, Sigma) was used to terminate the reaction. Finally, the results of ALP activity were standardized according to the protein concentration. Next, BMSCs were xed with formalin for 10 minutes, ALP staining buffer (Sigma-Aldrich) was added, and the cells were incubated for 30 minutes at room temperature.

Alizarin red staining
BMSCs (1.0 × 10 5 per well) were seeded into 6-well plates, and treated with GEN at different concentrations (0, 10, 20, and 40 µM) for 14 days under osteogenic environment. Next, the medium was discarded, the cells were washed with PBS for three times, and a formalin solution was added to x the cells. After 20 minutes, the formalin was discarded, BMSCs were washed 2 times with medium, and treated with alizarin red solution (Sigma-Aldrich) for 30 minutes. Subsequently, the stained cells were observed under an optic microscope and images were taken in random elds. Finally, the software Image J (NIH, Bethesda, MA, USA) was used to perform the statistics of the mineralized nodules, and the nodules larger than 0.04 mm were included in the statistical calculation [46].
Quantitative PCR (q-PCR) BMSCs (1.0 × 10 5 cells/well) were seeded into 6-well plates, and treated with GEN at different concentrations (0, 10, 20, and 40 µM) in the osteogenic medium for 14 days. Then the mRNA expression of ALP, OPN, OCN, OSX, Runx2, and BMP2 was detected by q-PCR. Trizol reagent (Sigma-Aldrich) was used to extract the RNA from BMSCs. The total RNA was translated into cDNA according to the reversetranscribed kit (Applied Biosystems, USA) using the following parameters: 95 °C for 9 minutes, 36 °C for 40 minutes for 2 cycles, then 86 °C for 4 minutes, and nal cooling to 4 °C. The cDNA of the target gene was quanti ed by q-PCR using the SYBR Green Premix kit (Roche, Switzerland). The q-PCR parameters were the following: 95 °C for 20 seconds, 90 °C for 10 seconds for 40 cycles, and 60 °C for 30 seconds. The primers used in this study (Life Technologies) were the following: ALP, forward primer AACCCAGACACAAGCATTCC, reverse primer GAGAGCGAAGGGTCAGTCAG. Runx2, forward primer AATTAACGCCAGTCGGAGCA, reverse primer CACTTCTCGGTCTGACGACG. OSX, forward primer CACTTCTCGGTCTGACGACG, reverse primer CACTTCTCGGTCTGACGACG. OCN, forward primer CACTTCTCGGTCTGACGACG, reverse primer ATAGCTCGTCACAAGCAGGG. BMP2, forward primer GCTTCCGTCCCTTTCATTTCT, reverse primer GCTTCCGTCCCTTTCATTTCT. OPN, forward primer GCTTCCGTCCCTTTCATTTCT, reverse primer GCTTCCGTCCCTTTCATTTCT. GAPDH, forward primer CATCACTGCCACCCAGAAGAC, reverse primer CCAGTGAGCTTCCCGTTCAG. GAPDH was used as the internal control. The relative gene expression was calculated using the 2 −ΔΔCt method.

Experimental model and animal groups
Thirty-six female C57BL/6 mice (8-week-old, 21 ± 2 g) were obtained from the animal experimental centre of the Southern Medical University (Guangzhou, Guangdong, China). The experimental mice were stochastically divided into three groups: Sham group (n = 12), ovariectomized (OVX) group (n = 12) and OVX + GEN group (n = 12). The OVX + GEN groups were treated by an oral gavage of 50 mg/kg/day GEN during these 3 months. The Sham group and OVX group received the same dose of saline by oral gavage. After 3 months, the experimental mice were euthanized by cervical dislocation and the femurs were collected for further studies.

Histological and immunohistochemical staining
The femur was immersed in 4% paraformaldehyde for 48 hours and decalci ed using 15% ethylenediaminetetraacetic acid for 14 days. Then, they were dehydrated, para n embedded, and cut into 4 µm-thick sections. To perform the haematoxylin-eosin (HE) staining, the sections were dewaxed, hydrated, and stained with HE dyes (Abcam). Finally, the HE stained sections were photographed and analysed. As regard the immunohistochemistry, the sections were dewaxed, hydrated, treated with 3% hydrogen peroxide for 15 minutes and with protease K for 10 minutes. Next, the sections were treated with primary antibodies and incubated overnight at 4 °C. The primary antibodies (Santa Cruz Biotechnology) used were the following: anti-Runx2 (1:200), anti-BMP-2 (1:200), and anti-OCN (1:200). The sections were washed with PBS thrice for a total of 15 minutes and the second antibody was added and incubated for 50 minutes. Next, the sections were washed three times with PBS, and then diaminobenzidine solution was added to obtain the chromogenic reaction. Finally, the sections were observed and analysed under an optical microscope.

Microcomputer tomography analysis
The collected femurs were preserved in 4% paraformaldehyde for 48 hours. The prepared femur was scanned and analysed by high-resolution micro CT (Caskaisheng, China). The scanning parameters of the micro-CT were set as follows: 80 kV, 15 µA, and a scanning thickness of 20 µm. The area below the crud end of femoral shaft was chosen as the analysis area for statistical analysis [47]. The bone parameters for statistical analysis included the following three indexes: trabecular bone mineral density (BMD), trabecular number, and trabecular thickness.

Statistical analysis
Statistical analysis was performed using GraphPad Prism 6 (Manufacturer, La Jolla, CA, USA). All in vitro experiments were repeated three times, and each experiment was carried out in triplicate. In the in vivo experiments, each group contained at least 6 rats. Results were expressed as mean ± standard deviation (SD). One-way ANOVA and Dunnett's test were used to compare multiple groups, while unpaired Student's t-test was used for the comparison of two groups. P < 0.05 was considered statistically signi cant.        GEN promotes osteogenic differentiation of BMSCs in a β-catenin-BMP2 dependent manner. BMSCs were transfected with either Ad-GFP or Ad-Cre, which were used to delete the gene of β-catenin. Then the treated cells were incubated with GEN (40μM) for 7 days. The protein expression of β-catenin (a) was evaluated by western blot. The mRNA level of Runx2 (b), OSX (c), OCN (d), OPN (e), and BMP2 (f) were examined by q-PCR. BMSCs were transfected with either Ad-GFP or Ad-Cre, which were used to delete the gene of BMP2. Then the treated cells were incubated with GEN (40μM) for 7 days. The protein expression of BMP2 (g) and β-catenin (l) was estimated by western blot. The mRNA level of Runx2 (h), OSX (i), OCN (j), and OPN (k) were tested by q-PCR. *P <0.05 compared with control (Con) group.

Figure 9
Proposed model depicting the underlying mechanisms of GEN in promoting the osteogenic differentiation of BMSCs