Isorhamnetin attenuates osteoarthritis by inhibiting osteoclastogenesis and protecting chondrocytes through modulating reactive oxygen species homeostasis

Abstract Increasing evidence indicates that osteoarthritis (OA) is a musculoskeletal disease affecting the whole joint, including both cartilage and subchondral bone. Reactive oxygen species (ROS) have been demonstrated to be one of the important destructive factors during early‐stage OA development. The objective of this study was to investigate isorhamnetin (Iso) treatment on osteoclast formation and chondrocyte protection to attenuate OA by modulating ROS. Receptor activator of nuclear factor‐kappa B ligand (RANKL) was used to establish the osteoclast differentiation model in bone marrow macrophages (BMMs) in vivo. H2O2 was used to induce ROS, which could further cause chondrocyte apoptosis. We demonstrated that Iso suppressed RANKL‐induced ROS generation, which could mediate osteoclastogenesis. Moreover, we found that Iso inhibited osteoclast formation and function by suppressing the expression of osteoclastogenesis‐related genes and proteins. We proved that Iso inhibited RANKL‐induced activation of mitogen‐activated protein kinase activation of mitogen‐activated protein kinase (MAPK), nuclear factor‐kappa B (NF‐κB) and AKT signalling pathways in BMMs. In addition, Iso inhibited ROS‐induced chondrocyte apoptosis by regulating apoptosis‐related proteins. Moreover, Iso was administered to an anterior cruciate ligament transection (ACLT)‐induced OA mouse model. The results indicated that Iso exerted beneficial effects on inhibiting excessive osteoclast activity and chondrocyte apoptosis, which further remedied cartilage damage. Overall, our data showed that Iso is an effective candidate for treating OA.

OA is characterized by progressive loss of articular cartilage, remodelling of the subchondral bone and osteophyte formation. 7,8 Currently, increasing evidence has shown that osteoclast activity and bone remodelling increase in the early stage of OA, disturbing the equilibrium between bone formation and resorption that could eventually lead to a marked reduction in subchondral bone thickness. 9,10 Moreover, the imbalance between bone formation and resorption leads to abnormal reconstruction of subchondral bone, which could promote the progression of OA. 11 This evidence provides a theoretical basis for targeting osteoclasts in the early stage of OA, which could be a useful intervention strategy for the treatment of OA.
Increasing evidence shows that reactive oxygen species (ROS) are one of the important destructive factors during OA development. 12 ROS perform various functions in cell energetic cycling, extracellular matrix metabolism and maintenance of chondrocyte homeostasis under physiological conditions. However, redundant production of ROS will break normal signalling and homeostasis. 13 Increased production of ROS is frequently found in OA, not only promoting osteoclastogenesis but also leading to matrix metalloproteases (MMPs) secretion and degradation of cartilage. [14][15][16] Moreover, ROS can induce dysregulation of mitochondria by depolarizing the mitochondrial membrane, which further results in increased ROS production 17,18 and causes a vicious circle. Numerous studies have shown that oxidative stress caused by ROS can induce apoptosis in chondrocytes. 19,20 All of these findings highlight that ROS modulation may be a potential treatment for OA by inhibiting osteoclastogenesis and protecting chondrocytes.
Isorhamnetin (Iso), a 3′-methylated flavonol derived from herbal medicinal plants, has been shown to have a significant impact on many pathological processes. Iso was reported to protect against hypoxia/reoxygenation-induced injury by attenuating oxidative stress in cardiomyocytes. 21 Iso was also capable of inhibiting cancer cells migration, indicating that it could be a potential clinical chemotherapeutic approach for cancer. 22 Furthermore, Iso has been shown to be an effective compound that reduced IL-1β-induced expression of inflammatory mediators in human chondrocytes. 23 Despite these early reports indicating the translation potential in pre-clinical research, whether Iso could be used to treat OA by inhibiting osteoclasts and reducing ROS-induced chondrocyte apoptosis remains largely unexplored.
In the present study, we used receptor activator of nuclear factor-kappa B ligand (RANKL) to activate osteoclasts to mimic the pathological environment in the early stage of OA. We investigated the effects of Iso treatment on RANKL-induced osteoclast activation. Additionally, the underlying mechanisms regarding Iso-induced inhibition of osteoclastogenesis were investigated. Moreover, H 2 O 2 was used to induce ROS generation in chondrocytes and we investigated whether Iso could attenuate ROS-induced chondrocyte apoptosis. For in vivo studies, Iso was administered to an anterior cruciate ligament transection (ACLT)-induced OA mouse model.
We aimed to provide novel insights regarding OA treatment and a foundation for future translational use of Iso in the clinic.

| BMMs isolation and culture
BMMs were obtained from the long bones of 6-week-old C57/BL6 mice. Briefly, after killing, the long bones, including femurs and tibias, were separated under aseptic conditions. All of the bone marrow was flushed from mouse femurs and tibias and then resuspended in complete α-MEM containing 30 ng/mL M-CSF. Cells were cultured in 75-cm 2 flasks and cultured in a 5% CO 2 incubator at 37°C. The media were changed every other day to remove nonadherent cells. At 80% confluence, the cells were washed with PBS three times and then trypsinized to harvest BMMs for the following experiments.

| Primary chondrocytes culture and immunofluorescence staining
Male C57BL/6J mice were killed and disinfected using 75% alcohol for 10 minutes. The head of the femur was exposed under aseptic conditions. All articular cartilage was isolated, collected and cut into 1 mm 3 pieces. The tissue was digested with 0.25% trypsin and 0.2% collagenase II for 30 minutes and 5 hours respectively. The cells were then filtered through a 70-μm cell strainer and washed three times with PBS. Next, the collected cells were seeded into culture dishes in DMEM at 37°C and 5% CO 2 . The culture medium was changed every other day. The expression of collagen II in chondrocytes after Iso treatment was analysed by immunofluorescence staining using collagen II antibody (Santa Cruz Biotechnology, CA).

| Cell viability assay
To evaluate the cytotoxic effects of Iso, cell viability was analysed using a cell counting kit-8 assay (CCK-8, Dojindo, Japan) following the manufacturer's protocol. In the in vitro assay, BMMs that were cultured with α-MEM containing 30 ng/mL M-CSF and chondrocytes were treated with varying concentrations of Iso (0, 3.125, 6.25, 12.5, 25, 50, 100 and 150 μmol/L) for 24 hours. Next, 10% CCK-8 solution was added to each well. The cells were cultured in a 37°C, 5% CO 2 incubator for an additional 2 hours. Then, the absorbance was measured at a wavelength of 450 nm using a microplate reader (Bio-Rad, Hercules, CA). Cell viability relative to the control group was calculated according to a previous study. 24

| In vitro osteoclastogenesis and bone resorption assay
To induce osteoclasts, BMMs harvested as mentioned above were seeded in a 96-well plate at a density of 10 4 cells/well. After 12 hours, the cells were treated with 50 ng/mL RANKL and 30 ng/ mL M-CSF for osteoclast differentiation according to a previous study. 25  With the assistance of mechanical agitation and sonication, the slides were washed with PBS to remove the cells. Resorption pits were imaged using scanning electron microscopy (SEM; FEI Instr., Hillsboro, OR).

| Quantitative real-time polymerase chain reaction (qRT-PCR) analysis
BMMs were seeded in a six-well plate at a density of 10 5  Foster City, CA) according to a previous study. 26 The primers that were used are listed in Table S1. The values were normalized to GAPDH mRNA expression levels and calculated using the 2 -ΔΔCt method.

| Western blotting
Western blotting analysis was performed according to a previously published protocol. 27 Total protein was extracted from BMMs and chondrocytes using radioimmunoprecipitation assay buffer (RIPA; Millipore, MA) containing protease and phosphatase inhibitors. The protein concentrations were analysed using a BCA protein assay kit.
A total of 20 μg of protein was separated using 10% SDS-PAGE gel electrophoresis and then transferred to a polyvinylidene fluoride (PVDF) membrane. The membranes were blocked with 5% non-fat milk for 2 hours and incubated overnight at 4°C with primary antibodies. Furthermore, the membranes were incubated with secondary antibodies for 1 hour and washed in Tris-buffered saline with Tween (TBST). Fluorescent signals were detected using an Odyssey imaging system (Li-Cor, Lincoln, NE).

| Cell cycle analysis
Chondrocytes were treated with or without 25 μmol/L Iso for 24 hours. Cell cycle analysis was performed by propidium iodide (PI) staining followed by analysis with a FACScan flow cytometer (BD, CA).

| Intracellular ROS detection
Levels of ROS in BMMs and chondrocytes were measured using DCFH-DA probe according to a previous study. 28

| Flow cytometry apoptosis analysis
For annexin V/propidium iodide (PI) apoptosis analysis, chondrocytes (5×10 5 /well) were seeded in 6-well plates and cultured for 12 hours, followed by culturing with or without H 2 O 2 (200 μmol/L) and various concentrations of Iso for 24 hours. After washing with PBS three times, chondrocytes were digested in 0.25% trypsin at 37°C and resuspended in binding buffer followed by annexin V-FITC/PI staining (BD, CA) according to the manufacturer's instructions. The samples were analysed using a FACScan flow cytometer.

| Measurement of mitochondrial membrane potential
The JC-1 fluorescent probe was purchased from Beyotime Biotech (Shanghai, China). Chondrocytes were treated with or without H 2 O 2 and various concentrations of Iso for 24 hours. The JC-1 probe was administered according to the manufacturer's instructions. The fluorescence signals were measured using confocal fluorescence imaging microscopy. Mitochondrial depolarization occurring during apoptosis could be indicated as an increase in the green/red fluorescence ratio. 28

| Cell apoptosis detection by TUNEL staining
The TUNEL apoptosis detection kit was purchased from Beyotime Biotech (Shanghai, China). Primary chondrocytes were seeded in confocal dishes for 12 hours. The cells were treated with or without H 2 O 2 and various concentrations of Iso for 24 hours. After fixation in 4% paraformaldehyde for 30 minutes, the cells were treated with 0.1% Triton X-100 for 5 minutes. Then, the cells were sealed for 1 hour in a TUNEL reaction mixture at 37°C. The cells were incubated with DAPI for 5 minutes and observed with confocal fluorescence microscopy.
The rate of TUNEL-positive cells in each field was calculated.

| Micro-CT scanning
At the end of 4 weeks after injection, the knee joints of the mice were harvested and fixed in 4% paraformaldehyde. Specimens were scanned using micro-CT (μCT 80; Scanco, Zurich, Switzerland) as described previously. 31 The micro-CT parameters were as follows:

| Histological observations
Samples were decalcified in 10% EDTA for 3 weeks and then embedded in paraffin. For microstructure observation, sagittal sections of the knee joint medial compartment were cut with a thickness of 4 μm and stained with haematoxylin and eosin (H&E), safranin O-fast green (S&F) and TRAP. Osteoarthritis Research Society International (OARSI) scores were calculated as previously described. 32 Moreover, TUNEL and MMP-3 (Santa, USA; dilution 1:100) immunohistochemical staining was accomplished. Total and positively stained chondrocytes in the entire articular cartilage were calculated.

| Statistical analysis
Each group contained three samples in vitro. All data were expressed as means ± standard deviations. One-way analysis of variance (ANOVA) was used for multifactorial comparisons in this study.
The assumption of normality was verified and the homogeneity of variance was tested. Differences between each group were assessed using post hoc multiple comparisons. Specifically, if there was no heterogeneity observed, the Bonferroni test was used to assess the differences between groups. However, if heterogeneity existed, the Welch test was used to test the equality of means and Dunnett's T3 test used to assess the differences. P < 0.05 were considered statistically significant. The following symbols were used to indicate statistical significance: * and # indicate P < 0.05, ** and ## indicate P < 0.01. All data analysis was conducted using SPSS 22.0 analysis software (SPSS Inc, Chicago, IL).

| Iso inhibits RANKL-induced ROS production and osteoclastogenesis in vitro
The chemical structure of Iso is shown in Figure 1

| Iso inhibits the expression of osteoclastogenesis-related genes and proteins in BMMs
Osteoclast differentiation is regulated by specifically related signalling pathways, which leads to the expression of particular genes. 33 The expression of osteoclastogenesis-related genes, including the nuclear factor of activated T-cells 1 (NFATc1), TRAP, c-FOS, DC-STAMP, cathepsin K and MMP-9, was analysed using quantitative real-time polymerase chain reaction (qRT-PCR). As shown in

| Iso suppresses RANKL-induced activation of MAPK/NF-κB/AKT signalling pathways
Binding of RANKL to the receptor RANK could activate the MAPK pathway. 34 Three major MAPK signalling cascades, ERK, JNK and p38, have been shown to be crucial for activating osteoclast formation and function. As shown in Figure 3, RANKL treatment clearly promoted phosphorylation of ERK, JNK and p38. However, phosphorylation of ERK, JNK and p38 considerably decreased after Iso treatment. RANKL-induced NF-κB activation is essential for initiating osteoclast differentiation. 35 27 The results showed that RANKL treatment significantly increased phosphorylation of AKT in BMMs; however, the activation effects were substantially suppressed by Iso (Figure 3). All of these results indicated that Iso could attenuate osteoclast formation by inhibiting MAPK/NF-κB/AKT pathways.
A schematic illustration shows the signalling pathways involved in osteoclast differentiation regulated by Iso in Figure 3. Iso inhib-

| The direct effects of Iso on chondrocytes
To investigate the direct effects of Iso on chondrocytes, a CCK-8 assay was first used to evaluate the cytotoxic effects of Iso. As shown in Figure S1A, the concentrations of Iso used to inhibit osteoclast formation (6.25, 12.5, 25 μmol/L) had minimal inhibitory effects on chondrocytes. Cell cycle analysis showed that Iso treatment did not noticeably affect the cell cycle of chondrocytes ( Figure   S1B). Furthermore, Iso treatment had a minimal effect on the collagen II expression of chondrocytes ( Figure S1C). All of these results indicated that Iso treatment did not directly affect chondrocytes.

| Iso protects chondrocytes from ROSinduced apoptosis
The production of intracellular ROS was detected by DCFH-DA probe. We found that the number of ROS-positive cells was clearly increased in chondrocytes treated with H 2 O 2 for 24 hours. However, Iso could alleviate the production of ROS stimulated by H 2 O 2 in chondrocytes ( Figure 4). In addition, in the H 2 O 2 -treated group, the apoptosis rate was 16.26 ± 1.88%, which was reduced to 13.07 ± 2.82%, 11.17 ± 3.26% and 3.94 ± 1.18% in the Iso 6.25, 12.5 and 25 μmol/L treated groups respectively, showing that Iso effectively decreased the apoptosis rate caused by H 2 O 2 (Figure 4). JC-1 staining showed a significant increase in the green/red ratio in H 2 O 2 -treated cells compared with the control group, whereas the ratio was reduced by Iso in a dose-dependent manner (Figure 4). These results were also confirmed by TUNEL staining, showing that in the H 2 O 2 -treated group, the percentage of TUNEL-positive cells was 54.67 ± 9.8%, whereas it was reduced to 51.66 ± 6.1%, 38.7 ± 5.30%, 32.3 ± 6.7% in the Iso-treated (6.25, 12.5 and 25 μmol/L) groups ( Figure 4).
All results indicated that Iso reduced ROS-induced chondrocyte apoptosis.

| Iso suppresses subchondral bone loss induced by ACLT in vivo
To investigate the effects of Iso on osteoclasts during the development of OA, we administered Iso intraperitoneally in ACLT-induced mice. As shown in Figure 6, ACLT injury induced significant bone resorption in the subchondral bone as shown by the increased area of bone lesions in the μCT results. Nonetheless, the destruction was reduced by Iso treatment dose-dependently, showing increased integrity of the knee bone. Moreover, sagittal views of the medial compartment of subchondral bone showed that Iso decreased ACLT-induced bone resorption, indicating that osteolysis was diminished after Iso treatment ( Figure 6). Statistically, the tibial plateau BV was significantly less than that of the control group.   Phosphorylated p65 translocates into the nucleus and binds to specific DNA sites to activate the expression of osteoclastogenesis-related genes. 43 Our results demonstrated that phosphorylation of p65 was markedly activated by RANKL stimulation in BMMs, whereas Iso suppressed NF-κB activation as indicated by the increased expression level of IκBα and decreased phosphorylation of p65. NFATc1 has been shown to be a key transcription factor that plays an important role in osteoclastogenesis. 44 Moreover, NFATc1 can activate varying osteoclastogenic genes associated with osteoclast differentiation and function, including c-FOS, TRAP, MMP-9, cathepsin K and DC-STAMP. 45 The AKT-NFATc1 signalling axis is important in osteoclast formation and it has been reported that inhibition of AKT activation could prevent RANKL-induced osteoclastogenesis. 46 In the present study, we found that Iso dramatically suppressed the phosphorylation of AKT activated by RANKL in BMMs. MAPK is in the serine-threonine protein kinase family that is involved in cell viability, differentiation, apoptosis, inflammation and innate immunity. 47 The most studied proteins of the MAPK pathway are ERK, JNK and p38 kinases, which can respond to extracellular stimuli including growth factors, heat shock proteins and pro-inflammatory cytokines. 48 Specific inhibitors of JNK, p38 and ERK can inhibit RANKL-stimulated osteoclast differentiation, indicating that MAPK is important for osteoclast activation and bone resorption. 49,50 Our data showed that Iso clearly suppressed RANKL-induced phosphorylation of ERK in a dose-dependent manner. Interestingly, Iso clearly inhibited the phosphorylation of JNK and p38 compared with the RANKL-stimulation group only at the concentration of 25 μmol/L, indicating that the pharmacological effects of Iso are concentration-dependent. These data suggested that Iso reduced RANKL-induced osteoclastogenesis by regulating the MAPK/NF-κB/AKT signalling pathway.

| Iso inhibits osteoclasts and protects chondrocytes in vivo
In this study, we revealed the vital role of ROS in the progression of OA and for the first time, proved that Iso could not only suppress osteoclastogenesis but also protect chondrocytes by modulating ROS. This finding could provide a novel theoretical basis for OA treatment. However, this study also has potential limitations. First, we just signified the effects of Iso on the MAPK/NF-κB/AKT signalling pathway; whether Iso can affect other mediators and signalling pathways involved in osteoclastogenesis is unknown. Moreover, the cells used in this study were obtained from mice instead of humans; hence, it is necessary to further investigate the effect of Iso on human cells for translational use. In conclusion, our investigation showed the beneficial effects of Iso in treating OA. Iso inhibited RANKL-induced osteoclastogenesis and protected chondrocytes by modulating ROS. All the data indicated that Iso could be a potential agent for the treatment of OA.

ACK N OWLED G EM ENTS
This study was supported by the National Natural Science Foundation of China (81672205) and the Shanghai Science and Technology Development Fund (18DZ2291200, 18441902700).

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.