Aucubin exerts anti-osteoporotic effects by promoting osteoblast differentiation

Osteoporosis is a metabolic disease characterized by reduced osteoblast differentiation and proliferation. Oxidative stress plays a role in the pathogenesis of osteoporosis. Aucubin (AU), an iridoid glycoside, was previously shown to promote osteoblast differentiation. We investigated the effects of AU on MG63 human osteoblast-like cells treated with dexamethasone (Dex) or hydrogen peroxide (H2O2) to induce oxidative damage. AU protected MG63 cells against apoptosis, and promoted increased expression of cytokines associated with osteoblast differentiation, including collagen I, osteocalcin (OCN), osteopontin (OPN), and osterix. In Dex- and H2O2-treated MG63 cells, AU also enhanced the expression of anti-oxidative stress-associated factors in the nuclear respiratory factor 2 signaling pathway, including superoxide dismutases 1 and 2, heme oxygenases 1 and 2, and catalase. In vivo, using a Dex-induced mouse model of osteoporosis, AU promoted increased cortical bone thickness, increased bone density, and tighter trabecular bone. Additionally, it stimulated an increase in the expression of collagen I, OCN, OPN, osterix, and phosphorylated Akt and Smads in bone tissue. Finally, AU stimulated the expression of cytokines associated with osteoblast differentiation in bone tissue and serum. Our data indicate AU may have therapeutic efficacy in osteoporosis.


INTRODUCTION
Osteoporosis is a metabolic disease characterized by the destruction of bone tissue and a reduction in total bone mass [1]. It results from an imbalance between osteoblast-mediated bone formation and osteoclastmediated bone resorption, which is essential for normal bone metabolism [2]. Therapeutics for osteoporosis have predominantly targeted osteoclasts to prevent bone resorption. However, they can result in serious adverse effects [3]. For example, bisphosphonates can cause jaw necrosis, esophageal cancer, and renal failure. Parathyroid hormone (PTH) is the only Food and Drug Administration-approved agent that stimulates bone formation, but it has been linked to osteosarcoma and can only be used for 2 years [4].
The production of reactive oxygen species (ROS) through mitochondrial respiration increases with age. The accumulation of ROS causes intracellular oxidative stress [5], which plays a role in the pathogenesis of osteoporosis. Oxidative stress can induce cell apoptosis and disrupt the balance between osteoblast and osteoclast activity. This can lead to reduced proliferation and differentiation of osteoblasts from bone marrow mesenchymal stem cells, which reduces bone formation and bone mass [6][7][8]. Nuclear respiratory factor 2 (Nrf2) is required for the induction of superoxide dismutase (SOD) and activation of the antioxidant response to internal and external chemical stimuli [9]. Interestingly, a statin (Simvastatin) demonstrated anti-osteoporotic effects by increasing heme oxygenase (HO-1) and SOD levels thereby reducing oxidative stress [10].
AGING Several natural compounds have been identified that exhibit anti-osteoporotic effects without causing adverse events and toxicities [11]. Three categories of agents have been identified: (1) phytoestrogens with estrogenic effects, (2) antioxidants and anti-inflammatory agents, and (3) compounds with pleiotropic effects [12]. Aucubin (AU) is an iridoid glycoside compound primarily derived from Eucommia ulmoides that has anti-osteoporotic effects [13]. It was previously found to promote angiogenesis, and displayed hepatoprotective, antiinflammatory, and anti-oxidative effects [14]. It was also shown to promote embryonic hippocampal neural stem cell differentiation in rats [15]. We previously demonstrated that AU could promote osteoblast differentiation by regulating bone morphogenetic protein-2 (BMP2) [16]. Therefore, we hypothesized that AU could have therapeutic efficacy for osteoporosis.
In this study, we investigated the effects of AU on human osteoblast-like cells treated with dexamethasone (Dex) or hydrogen peroxide (H 2 O 2 ) to induce oxidative damage, and in a Dex-induced mouse model of osteoporosis.

AU protected MG63 cells against Dex-induced damage via modulation of Nrf2 signaling
AU reduced the apoptotic rate of MG63 cells exposed to 4 μM of Dex for 24 h in a dose-dependent manner ( Figure 1A). Mitochondrial function is one of the factors contributing to apoptosis and it plays a role in the feedback loop that responds to ROS accumulation [17]. The over-accumulation of intracellular ROS (D) AU enhanced the expression levels of Bcl-2, and reduced the expression levels of Bax and cleaved caspase-3 in MG63 cells exposed to Dex. The quantification data of the expression levels of Bcl2, casepase3 and Bax were normalized by corresponding GAPDH. Data are expressed as mean  S.D. (n=6) and analyzed using a one-way ANOVA. # P<0.05 and ## P<0.01 vs. control cells, *P<0.05 and **P<0.01 vs. Dex-exposed cells.
( Figure 2B) and the enhanced dissipation of MMP ( Figure 2C) in MG63 cells caused by Dex were all strongly relieved by AU at doses of 1, 2.5 and 5 μM, as shown by the reduced green fluorescence intensity, and enhanced ratio of red/green fluorescence intensity, respectively. Bcl-2 family members contribute to cell apoptosis related to mitochondrial function [15]. Compared to MG63 cells exposed to Dex alone, AU significantly enhanced the expression levels of Bcl-2 and reduced the expression levels of Bax and cleaved caspase-3 (P <0.05) ( Figure 1D).

AU protected MG63 cells against H 2 O 2 -induced damage related to Nrf2 signaling
In H 2 O 2 -induced apoptotic MG63 cells, AU was protective against H 2 O 2 damage via reducing the apoptosis rate ( Figure 3A), suppressing the accumulation of ROS ( Figure 3B), and inhibiting the dissipation of MMP ( Figure 3C). Compared with H 2 O 2damaged MG63 cells, AU incubation resulted in a 13.6% increase in the expression of Bcl-2 (P <0.01; Figure 3D), and 25.7% and 29.2% reductions in the expression of Bax (P <0.001) ( Figure 3D) and cleaved caspase-3 (P <0.01) ( Figure 3D).

AU protected the Dex-damaged mice against osteoporosis
Compared with Dex-damaged mice with osteoporosis, 8-week AU administration made the cortical bone more continuous and reduced the number of osteoclasts, as detected by H&E staining ( Figure 5A). Giemsa staining revealed that AU treatment enhanced the number of trabecular osteoblasts in the mice ( Figure 5B).
The structural parameters of femur trabecular and cortical regions, and of the tibia cortical region, were detected via micro-CT (Figures 6 and Supplementary Figure 1). Compared with the control mice, thinner cortical bones and sparser trabecular bone were observed in the model mice with osteoporosis ( Figure 6A). Comparatively, E2 and AU enhanced the thickness of the bone cortex and the density of the trabecular bone, as shown by the increased brightness ( Figure 6A). Using standard 3D microstructural analysis, the bone mineral density (BMD), bone volume fraction (BV/TV), trabecular thickness (Tb.Th), trabecular spacing (Tb.Sp) and trabecular number (Tb.N) were calculated for each group. Compared with the control mice, reduced levels of BMD (P <0.01) ( Figure 6B), BV/TV (P <0.05) ( Figure 6C), Tb.Th (P <0.05; Figure 6D) and Tb.N (P <0.05) ( Figure  6F), and the increased levels of Tb.Sp (P <0.05) ( Figure  6F) and BS/BV (P <0.05) ( Figure 6G (B) AU increased the expression levels of proteins within the Nrf2/HO-1 signaling including P-DPR1, Nrf2, CAT, HO-1, HO-2, SOD-1 and SOD-2 in MG63 cells exposed to Dex. AU enhanced the expression levels of Nrf2 in both (C) nucleus and (D) cytoplasm of MG63 cells exposed to Dex. The quantification data of proteins were normalized by corresponding GAPDH and total proteins, respectively (n=4). (E) AU increased the mRNA levels of Nrf2 and NQO-1 in MG63 cells exposed to Dex. Marker size from top to bottom: 1000 bp, 700 bp, 500 bp, 400 bp, 300 bp, 200 bp and 100 bp. The data on quantified mRNA expression were normalized to the levels of -actin (n=4). Data are expressed as mean  S.D. and analyzed using a one-way ANOVA. # P<0.05, ## P<0.01 and ### P<0.001 vs. control cells, *P<0.05, **P<0.01 and ***P<0.001 vs. Dexexposed cells.
Administration of AU for 8 weeks at 15 mg/kg had no significant effects on the bone structures of healthy mice compared with the control mice ( Figures. 5 and 6).

AU promoted osteoblast differentiation in osteoporotic mice
Dex injection caused a significant reduction in the levels of factors related to osteoblast differentiation, including ALP, collagen I, OCN, OPN, BMP2 and BMPR2, and in the levels of Ca, Pi and E2 in the peripheral blood of Dex-injected mice with osteoporosis (P <0.05) ( Table 1), which were strongly enhanced after 8 weeks' AU and E2 administration (P <0.05) ( Table  1). Administration of AU for 8 weeks at 15 mg/kg had no significant effects on the serum levels of these factors (Table 1).
Different from E2, AU showed no significant effects on the levels of TRACP-5b (Supplementary Figure 2A). AU at 15 mg/kg and E2 at 15 µg/kg strongly enhanced the levels of TNF-α compared with model mice (P<0.01) (Supplementary Figure 2B) Similarly, remarkably low expression levels of collagen I, osterix, OCN, OPN, BMP2, P-Smads and P-Akt were noted in the  AGING lysed tibias and fibulas of Dex-induced osteoporotic mice (P<0.05) ( Figure 7A and 7B). Both E2 and AU treatment increased all of the detected proteins (P <0.05) ( Figure 7A and 7B). AU administration resulted in 40.5%, 21.6%, 31.5%, 112.9%, 238.7%, 83.1%, and 56.2% increases in the expression levels of collagen I, osterix, OCN, OPN, BMP2, P-Smads and P-Akt, respectively, at 8 weeks (P <0.05) ( Figure 7A and 7B). Among all of the detected proteins, administration of AU alone strongly enhanced the expression of osterix (P <0.05), but had no significant effects on other factors in the bone tissues of healthy mice compared with the control group ( Figure 7A).

Nrf2 signaling is involved in AU-mediated antiosteoporotic activity
It is well known that oxidative stress inhibits bone cell differentiation and impairs bone integrity [18]. In Dexinduced osteoporotic mice, high levels of ROS and low levels of SOD and CAT were noted in peripheral blood (P <0.05) ( Table 2). Compared with the osteoporotic mice, AU resulted in an 18.0% reduction in ROS, and 15.6% and 35.1% increases in SOD and CAT levels in peripheral blood (P <0.05) ( Table 2). In the lysed tibias and fibulas of the Dex-induced osteoporotic mice, the expression levels of Nrf2 signaling proteins were all strongly reduced (P <0.05) (Figure 8). Comparatively, E2 and AU relieved these reductions (P <0.05) (

DISCUSSION
Bone structure and quality are the main factors that affect strength and performance [19]. Thin and discontinuous bony cortices are typically observed in patients with osteoporosis [20]. The degree of trabecular mineralization (equivalent to the density of the mineral deposited in collagen) is frequently evaluated in osteoporosis patients as an index of bone turnover and the mechanical properties of the bone [21]. Our findings indicate that AU can enhance bone toughness and density, thicken the bone cortex, increase the mineralization of the bone trabeculae, and decrease the size of the mesh in a Dex-induced mouse model of osteoporosis.
Oxidative stress can promote the development of osteoporosis [22,23]. The intracellular accumulation of ROS causes mitochondrial dysfunction and can induce apoptosis [24,25]. ROS accumulation has been observed in bone tissue from patients with degenerative diseases such as osteoporosis [26]. Oxidative stress caused by estrogen deficiency or inflammatory bone disorders contributes to osteoporosis and bone resorption [22,27].  The data were analyzed using a one-way ANOVA and expressed as means  S.E.M. (n = 10). # P < 0.05 and ## P < 0.01 versus control mice; * P < 0.05, ** P < 0.01 and ***P < 0.001 versus osteoporosis injured mice.
Here, we found that AU reduced the expression of Bax and cleaved caspase-3, increased the expression of Bcl-2, and reduced the rate of apoptosis in Dex-and H 2 O 2treated MG63 cells. AU also inhibited ROS production and prevented dissipation of the mitochondrial membrane potential (MMP). Bcl-2 and Bax, which are located in the mitochondrial membrane, inhibit the production of oxygen free radicals and act as antioxidants to protect against mitochondrial apoptosis [28,29]. Under conditions of oxidative stress, the rate of apoptosis among mature bone cells increases and can contribute to osteoporosis [30]. Antioxidants protect bone cells against oxidative stress by inducing osteoblastogenesis and inhibiting osteoclast activation [31].
Oxidative stress inhibits osteoblast differentiation and promotes osteoclast differentiation [32]. We observed a reduction in the expression of proteins associated with osteoblast differentiation in Dex-and H 2 O 2treated MG63 cells and in bone tissue from mice with osteoporosis. Osteoblasts play an important role in the formation of the bone matrix and regulation of the bone resorption activities of osteoclasts. Osteoblasts synthesize and secrete cytokines such as osteopontin (OPN), alkaline phosphatase (ALP), and osteocalcin (OCN) to regulate osteoclast activity [33]. ALP is secreted by osteoblasts at an early stage during differentiation and regulates the synthesis of collagen I and other non-collagenous bone matrix proteins [34]. ALP is also responsible for the reorganization of mineralization components in the extracellular matrix [35].
Osterix, another marker of osteoblast differentiation, activates OCN in mature osteoblasts and regulates the final stages of bone formation [36,37]. BMP2 is important for bone formation and reconstruction. It induces the differentiation of mesenchymal cells into bone-forming cells, stimulates the expression of OCN, collagen I, and ALP, and activates Smad and non-Smad signaling by combining with transmembrane Ser/Thr kinase receptors [38]. BMP2 activation results in an increase in phosphorylated Smad1, Smad5, and Smad8 levels. BMP2 also activates ALP and OCN [39]. Activated Smads transmit signals from BMPs from the cytoplasm to the nucleus where they regulate the transcription of target genes [40,41]. Our results indicate that AU increases the expression of Smads in MG63 cells and in bone tissue from mice with osteoporosis. This suggest that they may have antiosteoporotic effects in addition to promoting osteoblast differentiation.
Under conditions of oxidative stress, heterodimerization of Keap1 sequesters most Nrf2 in the cytoskeleton. Keap1 has a cysteine-rich surface that is oxidized in response to oxidative and nitrosative stress [42]. Oxidative stress occurs causes the release of Nrf2 from Keap1 and translocation of Nrf2 to the nucleus, where it regulates the downstream antioxidant enzyme gene NQO1, and enhances the tolerance of cells to oxidative stress via influencing the expression of SOD, CAT and HO-1 [43][44][45][46]. High levels of Dex cause an increase in ROS [47]. This results in persistent oxidative stress and cellular damage, which contributes to the pathogenesis AGING of osteoporosis [48]. High levels of Nrf2 suppress the production of ROS [42]. Nrf2 deficiency stimulates osteoclast differentiation and activity as a result of increased oxidant production and activation of nuclear factor of activated T-cells (NFAT), which leads to bone resorption [49]. We found that AU suppresses ROS production and decreases the levels of phosphorylated dynamin-related protein 1 (P-DPR1), CAT, HO-1, HO-  The data were analyzed using a one-way ANOVA and expressed as means  S.E.M. (n = 10). # P < 0.05 and ## P < 0.01 versus control mice; * P < 0.05 and ** P < 0.01 versus osteoporotic mice.
2, SOD-1, and SOD-2 by increasing the expression of Nrf2 in bone tissue Dex-induced osteoporotic mice. The increase in phosphorylated AKT results in an increase in Nrf2 and phosphorylated Smad [50]. Nrf2 knockdown in MG63 cells abolished the effects of AU, suggesting that AU exerts anti-osteoporotic effects by regulating Nrf2 signaling in response to oxidative stress. Thus, AU may have therapeutic efficacy for osteoporosis and other disorders involving bone remodeling.

CONCLUSIONS
Our study first confirmed that the anti-osteoporotic property of AU in Dex/H 2 O 2 exposed MG63 cells and Dex-injected C57BL/6 mice with osteoporosis is due to regulation of Nrf2-medaited oxidative stress. The findings provide experimental evidence that AU may be used to treat diseases associated with bone formation.

Cell culture
MG63 human osteoblast-like cells (CRL-1427, passage < 10) were obtained from the American Type Culture Collection. The cells were cultured in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 10% fetal bovine serum, 100 U / mL penicillin, and 100 µg / mL streptomycin at 37°C in a humidified incubator with 5% CO 2 . All cell culture reagents were obtained from Gibco BRL (USA).

Measurement of cell apoptosis, MMP, and intracellular ROS
MG63 cells were seeded into 6-well plates at a density of 2 × 10 5 cells / well. The cells were then treated with 1, 2.5, or 5 μM AU for 2 h followed by 4 μM Dex or 200 μM of H 2 O 2 for 24 h. Following the incubation, the cells were harvested, resuspended in solution at a concentration of 1×10 6 cells / mL, and stained with propidium iodide and/or Annexin V for 20 min at room temperature. The rate of apoptosis was measured using a Muse Cell Analyzer (Millipore, USA).

Histological analysis of femur tissue
Femur tissue was collected immediately after the mice were euthanized and fixed in 4% paraformaldehyde. After incubation with a decalcification solution for 7 days, the tissue was embedded in paraffin, sectioned (5 μm thickness), and stained with hematoxylin and eosin and Giemsa. Sections were examined under a lightmicroscope equipped with a digital camera (Nikon, Japan).

Micro-computed tomography
Femurs and tibias were collected from mice immediately after euthanasia. The structure of the trabecular and cortical regions of the femur, and the cortical regions of the tibia were evaluated by microcomputed tomography (micro-CT) using a micro-CT µCT50 (Scanco, Switzerland). Standard 3D microstructural analysis was performed to analyze parameters including the trabecular bone mineral density (BMD), bone volume fraction (BV/TV), trabecular thickness (Tb.Th), trabecular spacing (Tb.Sp), and trabecular number (Tb.N).

Reverse transcription-polymerase chain reaction (RT-PCR)
The RNA was isolated from MG63 cells using Trizol (Invitrogen, USA), and then synthesized by QuantScript RT Kit (Tiangen Biotech Co. Ltd., Beijing China). -actin primers were used as an internal control. The conditions of PCR amplification were shown as follows: denaturation at 95 °C for 5 min, followed by 36

Statistical analysis
Data are expressed as the mean  standard deviation. One-way analysis of variance (ANOVA) followed by post-hoc Dunn's multiple comparison tests was performed using SPSS 16.0 software (IBM Corporation, USA). P < 0.05 was considered significant.

AU enhanced the bone cortex thickness
Micro-CT was used to determine structural parameters of femur trabecular and cortical regions, as well as tibia cortical region. From the CT images, it is obvious that the cortical bone of the model group is thinner and the brightness is lower than that of the blank group, and the bone cancellous is sparse. The bone cortex was the thickest and the highest brightness in the positive control group. Compared with the model group, the cortical bone mass of the low-dose, middle-dose and high-dose groups significantly increased and the brightness increased, the network structure in the bone cancellous became dense, and the middle-dose group had the best affections (Supplementary Figure 1).

AU enhanced the expression of TNF-α and TRACP-5b
Different from E2, AU showed no significant effects on the levels of TRACP-5b (Supplementary Figure 2A). AU at 15 mg/kg and E2 at 15 µg/kg strongly enhanced the levels of TNF-α compared with model mice (P<0.01; Supplementary Figure 2B).

Negative siRNA failed to influence the effects of AU on protein expressions
In both Dex and H 2 O 2 exposed MG63 cells, the negative siRNA transfection failed to influence the modulatory effects of AU on the expressions of Nrf2, SOD1, HO-1, Collagen 1, OCN and BMP2 (Supplementary Figure 3A and 3B).