Notoginsenoside R1 attenuates oxidative stress‐induced osteoblast dysfunction through JNK signalling pathway

Abstract Oxidative stress (OS)‐induced mitochondrial damage and the subsequent osteoblast dysfunction contributes to the initiation and progression of osteoporosis. Notoginsenoside R1 (NGR1), isolated from Panax notoginseng, has potent antioxidant effects and has been widely used in traditional Chinese medicine. This study aimed to investigate the protective property and mechanism of NGR1 on oxidative‐damaged osteoblast. Osteoblastic MC3T3‐E1 cells were pretreated with NGR1 24 h before hydrogen peroxide administration simulating OS attack. Cell viability, apoptosis rate, osteogenic activity and markers of mitochondrial function were examined. The role of C‐Jun N‐terminal kinase (JNK) signalling pathway on oxidative injured osteoblast and mitochondrial function was also detected. Our data indicate that NGR1 (25 μM) could reduce apoptosis as well as restore osteoblast viability and osteogenic differentiation. NGR1 also reduced OS‐induced mitochondrial ROS and restored mitochondrial membrane potential, adenosine triphosphate production and mitochondrial DNA copy number. NGR1 could block JNK pathway and antagonize the destructive effects of OS. JNK inhibitor (SP600125) mimicked the protective effects of NGR1while JNK agonist (Anisomycin) abolished it. These data indicated that NGR1 could significantly attenuate OS‐induced mitochondrial damage and restore osteogenic differentiation of osteoblast via suppressing JNK signalling pathway activation, thus becoming a promising agent in treating osteoporosis.


| INTRODUC TI ON
Osteoporosis is defined as a systemic degenerative disease, which is characterized by dysregulation of bone formation and progressive bone micro-architectural deterioration. 1 Osteoporosis largely increases the fracture risk of bone. 1,2 One of the major pathogenic factors for osteoporosis is oxidative stress (OS) 3,4 -a pathophysiological status with relatively overproduced reactive oxygen species (ROS) and insufficient anti-oxidative defence. 5 Our meta-analysis of clinical data concludes that in postmenopausal women the status of OS is closely related to the decreased bone mineral density (BMD). 6 Osteoblasts are responsible for bone formation, playing a crucial role in maintaining BMD and bone microstructure. 7 OS can significantly reduce osteoblast viability and activity, thereby diminishing osteoblast quantity and function, 8 which leads to the onset and progression of osteoporosis. 9 At subcellular level, ROS directly attack and cause damage to mitochondria-the most vulnerable target of ROS, which stimulates mitochondria to further generate and release ROS, forming a vicious circle and finally leading to mitochondrial dysfunction. 10 Mitochondrial dysfunction forms a critical molecular mechanism accounting for the OS-induced osteoblast apoptosis. 11 Mitochondrial dysfunction also further impairs osteoblastic bone formation function. 12 Therefore, bioactive agents that can both relieve OS-induced mitochondrial dysfunction and restore osteoblast function are promising to treat osteoporosis.
One of such candidate bioactive agents is notoginsenoside R1 (NGR1), a natural triterpene saponin compound derived from the traditional Chinese herb Panax notoginseng. 13 Firstly, NGR1 has a potent capacity in relieving cellular damages, thereby having a strong protective effect on different kinds of cells, [14][15][16] from OS in several pathological situations. Furthermore, the anti-oxidative effect of NGR1 is largely attributed to its preventive effect on OS-induced mitochondrial dysfunction. 17,18 On the other hand, NGR1 also has an invaluable property to promote osteoblast function-osteogenic differentiation. 19,20 However, it remains to be elucidated whether NGR1 can prevent OS-induced mitochondrial dysfunction and restore osteoblast function.
C-Jun N-terminal kinase (JNK) is one of the three signalling pathways of mitogen-activated protein kinases (MAPKs) that mediate cellular responses to physiological and pathological stimuli. 21 JNK pathway can be activated in response to ROS and mediates OSinduced mitochondrial dysfunction in primary cortical neurons. 22 Furthermore, JNK activation by acetaminophen is also shown to inhibit mitochondrial bioenergetics 23 and mitochondrial biogenesis 24 in liver cells. Consistently, our preliminary experiment also showed that JNK signalling pathway is activated in osteoblast under OS stimulation. On the other hand, NGR1 has been shown to be capable of blocking JNK signalling pathway. 25 Therefore, we hypothesized that NGR1 relieves OS-induced osteoblast dysfunction and mitochondrial damage through attenuating OS-induced JNK activation.
In the present study, we established an OS-induced osteoblast dysfunction model to explore (1) NGR1's osteoblast protection against oxidative damage and (2) its underlying molecular mechanisms.

| Measurement of apoptosis by transferase dUTP nick end labelling (TUNEL) assay
3 × 10 4 MC3T3-E1 cells were seeded per well in 48-well plates with coverslip. TUNEL staining was processed as previously described. 8 The percentage of apoptotic cells was estimated by the TUNEL positive cell counts in total cells from random fields. of osteoblast via suppressing JNK signalling pathway activation, thus becoming a promising agent in treating osteoporosis.

K E Y W O R D S
dysfunction, JNK, mitochondria, NGR1, osteoblast, oxidative stress 2.4 | Alkaline phosphatase (ALP) staining and ALP activity assay 3 × 10 4 MC3T3-E1 cells per well were seeded on 48-well plates and exposed to H 2 O 2 and/or other test compounds for indicated time. Then, the medium was exchanged to OM for 7 days. ALP staining was performed as previously described. 8 Each well was photographed using a stereomicroscope (Olympus). ALP activity was assayed as previously described 8 and presented as the concentration per gram of protein ((mg/ml)/g protein). Protein concentration was determined using BCA protein assay (Thermo Fisher).

| Mineralization analysis
Mineralization of MC3T3-E1 cells was determined in 48-well plates using Alizarin red S staining (ARS) as previously described. 8 Each well was photographed under stereomicroscope, and the mineralization area was quantified by Image J.

| Real-time polymerase chain reaction (rt-PCR)
30 × 10 4 MC3T3-E1 cells per well were seeded on 6-well plates and exposed to H 2 O 2 and/or other test compounds for indicated time. Then, the medium was exchanged to OM for 7 days. Total RNA extraction and rt-PCR amplifications were performed as previously described. 8 The sequences of specific primers were listed in Table 1.

| Measurement of mitochondrial membrane potential (MMP)
To assess MMP, cells were co-stained with tetramethylrhodamine methyl ester (TMRM, 100 nM, Life Technologies) and Mitotracker Green (Mitogreen, 100 nM, Life Technologies) for 30 min, as in our previous study. 11 Images were captured under a fluorescence microscope (Leica DMIL). Excitation wavelengths were 543 nm for TMRM and 488 nm for Mitogreen. Post-acquisition processing was performed with Image J software for the quantification of fluorescent intensity.

| Measurement of mitochondrial superoxide production
Cellular superoxide production in mitochondria was detected using the

| Detection of adenosine triphosphate (ATP) production
For the measurement of ATP level, whole-cell extracts were lysed in lysis buffer provided in the ATP assay kit (Beyotime). After centrifugation at 12,000 g for 5 min at 4°C, the supernatants were transferred to a new 1.5 ml tube for ATP analysis. The luminescence from a 100 μL sample was assayed in a luminometer (Molecular Devices) together with 100 μL of ATP detection buffer. A standard curve of ATP concentrations (1 nM-1 µM) was prepared from a known amount.

| Evaluation of mitochondrial DNA (MtDNA) copy number
MC3T3-E1 cells were lysed for total DNA extraction. rt-PCR was conducted with 40 ng DNA (OD260; NanoDrop). MtDNA copy number was measured by cytochrome c oxidase subunit 1 (COX-1) normalized with nuclear DNA products (β-actin). The sequences of specific primers were listed in Table 1.

Gene
Forward primer (

| Statistical analyses
Each experiment was repeated in triplicate. Data were reported as mean ± standard deviation (SD). One-way analysis of variance (ANOVA) was carried out by GraphPad Prism Software (Graph Pad Software). p values < 0.05 were recognized statistically significant.

| NGR1 attenuated H 2 O 2 -induced osteoblast apoptosis and dysfunction
The MTT test showed that, after being pre-incubated with NGR1 in ALP staining and ALP activity test ( Figure 1D, E). The osteogenic capability of mineralization examined by ARS was also significantly rescued by NGR1 ( Figure 1F, G). Furthermore, expression levels of typical osteogenic marker genes (ALP, Osteocalcin (OCN), Collagen I (COL I) and Runt-related transcription factor 2 (Runx2)) in OS injured model decreased but largely recovered after NGR1 administration ( Figure 1H).

| NGR1 attenuated H 2 O 2 -induced osteoblast mitochondrial dysfunction
To further confirm the role of NGR1 on mitochondrial OS and dys- levels ( Figure 4H) demonstrate that blocking JNK by SP600125 also recovered MC3T3-E1 cells' osteogenic ability. These data showed a stronger protective effect of NGR1 than SP600125. However, reactivating JNK by Anisomycin eliminated these benefits significantly, which confirmed that osteoblast protection from NGR1 due to its blockage of JNK signalling pathway.

| NGR1 promoted mitochondrial function recovery by blocking JNK signalling pathway
As shown in Figure 5A-D, SP600125 restored MMP and pre-   Figure 1A-C), but also osteogenic differentiation, such as ALP expression ( Figure 1D, E), osteoblastogenic marker genes levels ( Figure 1H) and mineralization ( Figure 1F to provide energy for cell activities, such as osteoblastic differentiation. 44 In the current study, H 2 O 2 caused a series of severe mitochondrial dysfunctions in osteoblast, such as mitochondrial membrane depolarization (Figure 2A, C), MtROS overproduction ( Figure 2B, D) and ATP level reduction ( Figure 2E), which was consistent to our previous findings. 11 In this study, we further showed that H 2 O 2 also significantly decreased MtDNA copy number ( Figure 2F)-an indicator of mitochondrial abundance and mutation, 49  In animal studies, NGR1 has been proven to have various effective therapeutic functions, such as neuroprotection, anti-diabetes, certain organ protection, bone metabolism regulation, anti-cancer and osteoporosis, which can be largely attributed to its potent antiapoptotic, anti-inflammatory and anti-oxidative properties. 13,51 NGR1 has been shown to effectively protect various types of cells from oxidative damages triggered by other pathological factors, such as AGEs (advanced glycation end products), 16 ischaemiareperfusion injury 52 and oxidized low-density lipoprotein. 53 NGR1 can decrease ROS production and prevent protein and lipid peroxidation by improving the activities of antioxidant enzymes, such as superoxide dismutase (SOD), catalase and glutathione. 15,52,54 As expected, our current study also showed that NGR1 antagonized the H 2 O 2 -induced OS, restored cell viability and decreased apoptosis rate of osteoblast ( Figure 1A-C). In the meantime, since NRG1 also bears a potent pro-osteogenic property, 19 we further tested its efficacy in OS microenvironment in this study. Our results showed that NGR1 could restore ALP level and activity ( Figure 1D, E), extracellular mineralization ( Figure 1F, G) and the levels of a series of osteoblastogenic marker genes (such as ALP, OCN, COL I and Runx2) ( Figure 1H) (Figure 2), which confirms that NGR1 can restore mitochondrial function. We further show that NGR1 effectively blocks the JNK signalling ( Figure 3). The critical role of JNK One limitation in this study is the adoption of a mouse osteoblast cell line. Studies with primary human mesenchymal stem cells need to be performed before extrapolating the current findings to clinical situations. Furthermore, in our future study, we will perform a welldesigned in vivo study to illustrate the therapeutic effect of NGR1 on osteoporosis.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data sets generated for this study are available on request to the corresponding author.