Chondroprotective and anti‐inflammatory effects of amurensin H by regulating TLR4/Syk/NF‐κB signals

Abstract The low‐grade, chronic inflammation initiated by TLR4‐triggered innate immune responses has a central role on early osteoarthritis. Amurensin H is a resveratrol dimer with anti‐inflammatory and anti‐apoptotic effects, while its effects on TLR‐4 signals to inhibit osteoarthritis are still unclear. In the present study, treatment with amurensin H for 2 weeks in monosodium iodoacetate‐induced mice significantly slows down cartilage degeneration and inflammation using macroscopic evaluation, haematoxylin and eosin (HE) staining and micro‐magnetic resonance imaging. In IL‐1β‐stimulated rat chondrocytes, amurensin H suppresses the production of inflammatory mediators including nitric oxide, IL‐6, IL‐17, PGE2 and TNF‐α using Greiss and ELISA assay. Amurensin H inhibits matrix degradation via decreasing levels of MMP‐9 and MMP‐13 using Western blot assay, promotes synthesis of type II collagen and glycosaminoglycan using immunostaining and safranin O staining, respectively. Amurensin H inhibits intracellular and mitochondrial reactive oxygen species (ROS) generation, and mitochondrial membrane depolarization using DCFH‐DA, MitoSOX Red and JC‐1 assay as well. IL‐1β stimulates TLR4 activation and Syk phosphorylation in chondrocytes, while amurensin H inhibits TLR4/Syk signals and downstream p65 phosphorylation and translocation in a time and dose‐dependent manner. Together, these results suggest that amurensin H exerts chondroprotective effects by attenuating oxidative stress, inflammation and matrix degradation via the TLR4/Syk/NF‐κB pathway.

involved in its initiation. 5 In the early OA joint, many risk factors trigger a progressive local tissue damage and metabolic dysfunction and produce endogenous damage-associated molecular patterns (DAMPs) which bind to TLR4. 6 The inflammatory mediators such as IL-1β, activate TLR4 signalling and trigger the expression of inflammatory factors such as TNF-α, inducible nitric oxide synthesis (iNOS) and cyclooxygenase 2 (COX 2). 7,8 This further promotes production of IL-6, IL-17, nitric oxide (NO) and PGE2. These inflammatory mediators induce cartilage destruction by activating matrix metalloproteinases (MMPs) such as MMP-9 and MMP-13, and inhibiting synthesis of matrix collagen and glycosaminoglycan (GAG), and finally prime a full-blown inflammatory response and structural damage of the joint. 3,5,9 The TLR4 downstream signalling events in OA have not been fully elucidated. Different stimuli recruit different adaptor/signalling molecules to TLR4 cytoplasmic domain and trigger various branches of TLR4 signals. 10 Spleen tyrosine kinase (Syk) has emerged recently as a key effector molecule in TLR4 activation. This signalling seems to be particularly important in innate immune response to DAMPs, which is relevant to chronic inflammatory diseases. [11][12][13] Recently, basic calcium phosphate crystals, a host-derived agonist, stimulate IL-6 secretion through Syk signalling in chondrocytes. 2 Therefore, understanding the TLR4-Syk interaction and involved downstream signalling in chondrocytes is important for regulating TLR4 signals and providing therapeutic "entry point" for modifying OA course.
Recent studies suggest natural stilbenes possess anti-inflammatory and anti-oxidative properties and inhibit the release of key OA-related catabolic mediators. 14 Vitis amurensis is a wild grape growing in northeast and central parts of China, whose leaves and roots are utilized in traditional Chinese medicine for cancer and pain. 15 Amurensin H (also known as Vam3) isolated from V. amurensis is a resveratrol dimer with potential effects on inflammatory diseases including asthma and chronic obstructive pulmonary. 16,17 Amurensin H is an ATP-competitive inhibitor of Syk due to the interaction between its ring-C/D and Syk, 18 and also, inhibits inflammation via NF-κB signalling. 19 Further, we previously have shown that amurensin H inhibits the sodium nitroprusside-induced chondrocyte apoptosis. 20 However, for amurensin H, the possible relationship between its chondroprotective and anti-inflammatory effects, and possible link between its effects on chronic inflammation and TLR4 signals have not been explored in detail.
In the present study, we examined the in vivo and in vitro anti-inflammatory and chondroprotective effects of amurensin H, and the possible corresponding mechanisms. Results showed that amurensin H inhibited inflammatory and catabolic responses, which was, at least partially, through the inhibition of TLR4-Syk interaction and NF-κB-mediated downstream inflammatory signals.

| Reagents and animals
Amurensin H was synthesized by Prof. Yao and identified by ESI-MS and NMR ( Figure 1A). 21 The male SD rats (4-week-old) and male C57BL/6 mice (20-22 g) were purchased from Vital River

| Monosodium iodoacetate -induced mice model and treatment
Mice were randomly divided into four groups (n = 12): control group, monosodium iodoacetate (MIA) group, and two amurensin H-treated group (10 and 20 mg/kg/b.wt., respectively). Briefly, mice were anaesthetized. Then, 500 μg MIA (Sigma) was dissolved in sterile saline (0.9%) and injected into the joint capsule. Intra-articular injection of 10 μL saline was performed as a control. Mice in amurensin H-treated groups received intra-gastric administration from day 14, and mice in the rest groups received solvents ( Figure 1B).

| Macroscopic evaluation
After mice scarification and soft tissue removal, macroscopic evaluation of cartilage on femoral head was performed by two observers using a four-grade scale (Table S1).

| Micro-magnetic resonance imaging
Joints were embedded in Tissue-Tek OCT Compound (Sakura-VWR), frozen on dry ice and stored at −80°C. Frozen joints were rehydrated overnight in normal saline at 4°C and scanned in a Main topics for review 1. Amurensin H showed chondroprotective and antiinflammatory effects in monosodium iodoacetate-induced mice and IL-1β-stimulated chondrocytes.
horizontal position by a 7.0 T micro-magnetic resonance imaging (micro-MRI) system (Bruker PharmaScan) with a 23-mm-volume coil. The T2 mapping multiple-spin-echo sequence was performed (Table S3) and generated with the Bruker ParaVision 6.0 system. The femoral weight-bearing area was selected as the region of interest, with average T2 relaxation times analysed by a senior musculoskeletal radiologist.

| Preparation and treatment of primary rat chondrocytes
Chondrocytes were digested from joints of 2-3-week-old male SD rats. 20 Briefly, rats were euthanized and immersed into 75% alcohol for 5 minutes. Articular cartilage was resected, digested with 0.25% Trypsin-EDTA for 30 minutes and collagenase II (Gibco) for 4 hours at 37°C and filtrated through the 200-mesh strainer. For culture, chondrocytes were seeded at density of 1.8 × 10 4 cells/cm 2 in highglucose DMEM with 1% penicillin/streptomycin (Solarbio) and 10% FBS (Gibco) in a 37°C and 5% CO 2 incubator. The phenotype at P1 was identified by typical morphology and immunostaining of type 2 collagen (COL2A1; Abcam) using Inverted Ti-E fluorescence microscope (Nikon). For experiments, unless otherwise mentioned, chondrocytes (passage 3 to 5) were seeded at density of 3 × 10 4 cells/ cm 2 and incubated with amurensin H for 2 hours, then, stimulated by 10 ng/mL human IL-1β (PeproTech) in DMEM with 2.5% FBS.

| Cell viability assay and live/dead assay
Chondrocytes grown on 96-well plates were pre-treated with amurensin H and stimulated by IL-1β for 48 hours. Cell viability was determined using MTT (Sigma) assay with OD value measured at 570 nm by a microplate reader. 22 Chondrocytes grown on 6-well plates were stained with LIVE/DEAD Cell Imaging Kit (Invitrogen) according to the manufacturer's protocol and analysed by fluorescence

| Estimation of levels of nitric oxide, PGE2, IL-6, IL-17 and TNF-α
Chondrocytes were seeded in 96-well plates, and supernatants were collected after IL-1β treatment for 48 hours. NO level was estimated using Greiss assay. 23 The levels of IL-6, IL-17, TNF-α and PGE2 were estimated using commercial ELISA kits (R&D Systems) according to the manufacturer's protocol with OD values measured at 450 and 570 nm.

| Estimation of glycosaminoglycan content
Chondrocytes at P1 were seeded on coverslips onside 6-well plates (2 × 10 4 /well) and treated by IL-1β for 48 hours, then fixed by 4% (v/v) paraformaldehyde, stained with Safranin O (SO, Sigma) for histological evaluation of cell morphology and GAG production.
Chondrocytes seeded in 6-well plates were treated by IL-1β for 96 hours, digested by proteinase K (Solarbio). The enzymatic hydrolysates were centrifuged. DNA content was measured by a spectrofluorometer using Hoechst 33258 dye at 460 nm (emission) and 360 nm (excitation). A series of diluted chondrocytes (1 × 10 6 cells) were used as a control. 24 The total intracellular sGAG secretion was qualitied spectrophotometrically at 525 nm using 1,9-dimethylmethylene blue dye (Sigma) with chondroitin sulphate (Sigma) as a standard, and normalized to the total DNA content.

| Statistical analyses
Values were presented as Mean ± SD. The statistically significant difference between experimental groups and control was analysed via one-way ANOVA followed by post hoc Tukey test. P < .05 was considered as statistically significant.

| Amurensin H alleviates the progression of MIA-induced OA in mouse model
Intra-articular injection of MIA disrupts cellular glycolysis and induces cell death, which triggers a local acute inflammation and subsequent cartilage degeneration. It is a widely used method to mimic structural and functional changes of human OA. 26 Here, normal cartilage with a smooth surface on the femoral condyles was macroscopically observed in the control group, while significantly thinning cartilage was observed in the model group (P < .01, Figure 1C).
Twenty mg/kg of amurensin H significantly alleviated bone wear, compared to the model group (P < .01).
Histological examinations also showed significant cartilage fibrillation, cartilage loss up to 60 ± 13.4%, subchondral bone erosion, inflammation and pannus formation in the model vs the control group (P < .01), and 20 mg/kg of amurensin H significantly alleviated this trend with 25.5 ± 11.4% of cartilage loss (P < .01, Figure 1D).

| Amurensin H inhibits IL-1β-induced inflammatory mediators in chondrocytes
To further investigate the in vitro chondroprotective effects of amurensin H, primary chondrocytes were isolated from rat knee and maintained typical morphology and high expression of COL2A1 at P1 ( Figure S1). Amurensin H treatment (2-8 μmol/L, 48 hours) had no obvious effect on cell viability ( Figure S2). Ten ng/mL IL-1β induced a maximal production of NO and IL-6 after incubation for 48 hours ( Figure S3).
The significantly elevated NO levels and iNOS expression induced by IL-1β were decreased significantly by pre-treatment with 4 μmol/L and 8 μmol/L of amurensin H (P < .05 or P < .01) (Figure 2A-C). The increased PGE2 levels and COX-2 expression were also significantly alleviated after pre-treatment with 4 μmol/L and 8 μmol/L of amurensin H (P < .05 or P < .01) (Figure 2A,B,D). Besides, IL-1β induced a significant secretion of IL-6, IL-17 and TNF-α (P < .01 vs control), and amurensin H blocked this trend in a dose-dependent manner (P < .01 at 8 μmol/L, Figure 2E,F). Taken together, amurensin H exhibited favourable biocompatibility and exhibited potent anti-inflammatory effects in pathological chondrocytes.

| Amurensin H regulates IL-1β-induced TLR4-Syk interaction in chondrocytes
We next determined the molecular mechanism responsible for chondroprotective effects of amurensin H. We firstly measured TLR4, Syk and TRAF6 expression and observed a significant increase within 120 minutes after stimulation (P < .01, Figure 5A). The above increased expression or phosphorylation was blocked by 8 μmol/L of amurensin H treatment. Besides, amurensin H inhibited TLR4, TRAF6 expression and Syk phosphorylation in a dose-dependent manner as well ( Figure 5B,C). We further confirmed the role of TLR4-Syk interaction using TLR4 and Syk inhibitor, TAK242 and BAY61-3606, respectively. Results showed that TAK242 and BAY61-3606 significantly abrogated the secretion of nitric oxide and IL-6 (P < .01, Figure 5E). Besides, increased expression of TLR4 and phosphorylation of Syk was attenuated more obviously by TAK242 ( Figure 5D).

| Amurensin H regulates IL-1β-induced translocation and phosphorylation of p65 in chondrocytes
The effects of amurensin H on the activation of NF-κB p65 were examined, and results showed a significant increasing expression of p-p65 within 120 minutes after stimulation, which was blocked by 8 μmol/L of amurensin H treatment (P < .01, Figure 6A).
Corroborating the time-dependent manner, there was a dose-dependent inhibition in the phosphorylation and nuclear translocation of p65 in amurensin H-treated chondrocytes (P < .01, Figure 6B-D).

| D ISCUSS I ON
TLR4 is the main regulators of innate immunity activated in OA, while little is known for its downstream signalling involved in OA.

F I G U R E 3 Amurensin H (AH) inhibited IL-1β-induced ROS generation in rat chondrocytes. (A, B)
The mitochondrial membrane potential was measured using JC-1. The total and mitochondrial ROS were measured using DCFH-DA (C) and MitoSOX Red (D). n = 6 in each group. #P < .05 and ##P < .01 vs control group; *P < .05 and **P < .01 vs IL-1β-treated group  TLR4-Syk interaction is complicated due to possible crosstalk with different branches of TLR4 signals. In macrophages stimulated by host-derived mmLDL, Syk links TLR4 activation with intracellular PLC/Nox2 signalling and subsequent generation of ROS and inflammatory cytokines. 11 In macrophages stimulated by LPS, endocytosis of TLR4 activates downstream Syk/TRIF/IRF3 signalling and delayed NF-κB activation. 33 In human proximal tubular epithelial cells, high glucose induces an immediate ROS-dependent F I G U R E 6 Amurensin H (AH) inhibited downstream pro-inflammatory transcription factor NF-κB in rat chondrocytes. Chondrocytes were treated with IL-1β for indicated time-points (A) or 1 h (B, C) after pre-incubation with AH for 2 h (n = 3). (A) AH inhibited p65 phosphorylation in a time-dependent manner. (B, C, D) AH inhibited p65 phosphorylation and p65 (F, 200×) translocation in a dose-dependent manner. β-Actin was used as a control for equal loading. #P < .05 and ##P < .01 vs control group; *P < .05 and **P < .01 vs IL-1β-treated group F I G U R E 7 A schematic diagram of proposed mechanism of amurensin H on TLR4/Syk/NF-κB signals in rat chondrocytes. Amurensin H treatment negatively regulated TLR4 activation, Syk phosphorylation, TRAF6 activation, and phosphorylation and translocation of p65 in IL-1β-stimulated chondrocytes, thereby inhibiting inflammation and matrix degradation extracellular release of HMGB-1 which activates TLR4/MyD88/ Syk/NF-κB. 12 In mice with retinal ischaemia/ reperfusion injury, Syk and NF-κB are key molecules in TLR4 downstream signalling. 13 Thus, NF-κB is an important downstream transcriptional factor of TLR/Syk signalling, which induces pro-inflammatory cytokine production. 3 Considering TRAF6 regulates NF-κB for its entry into the nucleus in TLR4 signalling, 34 we evaluate expression of TRAF6 and NF-κB. Results show that IL-1β leads to TRAF6 and NF-κB activation, which is inhibited by amurensin H. Therefore, amurensin H inhibits inflammation via regulating TLR4/Syk/NF-κB signalling in chondrocytes.
Our results confirm that innate immune response is involved in early OA, trigger chronic low-grade inflammation and subsequent ECM degradation, which supports potential therapeutic targets to slow or change disease course in early OA. Amurensin H has in vivo and in vitro anti-inflammatory and chondroprotective effects and inhibits TLR4/Syk/NF-κB signalling in chondrocytes, suggesting that amurensin H could be a potential candidate for disease-modifying osteoarthritis drugs.

CO N FLI C T O F I NTE R E S T
The authors have no conflict of interest to declare.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.