Cinnamic Aldehyde, the main monomer component of Cinnamon, exhibits anti‐inflammatory property in OA synovial fibroblasts via TLR4/MyD88 pathway

Abstract Cinnamon is a wildly used traditional Chinese herbal medicine for osteoarthritis (OA) treatment, but the underlying mechanism remains ambiguous. The purpose of this study is to explore the mechanism of cinnamic aldehyde (CA), a bioactive substance extracted from Cinnamon, on synovial inflammation in OA. A total of 144 CA‐OA co‐targeted genes were identified by detect databases (PubChem, HIT, TCMSP, TTD, DrugBank and GeneCards). The results of GO enrichment analysis indicated that these co‐targeted genes have participated in many biological processes including ‘inflammatory response’, ‘cellular response to lipopolysaccharide’, ‘response to drug’, ‘immune response’, ‘lipopolysaccharide‐mediated signalling pathway’, etc. KEGG pathway analysis showed these co‐targeted genes were mainly enriched in ‘Toll‐like receptor signalling pathway’, ‘TNF signalling pathway’, ‘NF‐kappa B signalling pathway’, etc. Molecular docking demonstrated that CA could successfully bind to TLR2 and TLR4. The results of in vitro experiments showed no potential toxicity of 10, 20 and 50 μM/L CA on human OA FLS, and CA can significantly inhibit the inflammation in LPS‐induced human FLS. Further experimental mechanism evidence confirmed CA can inhibited the inflammation in LPS‐induced human OA FLS via blocking the TLR4/MyD88 signalling pathway. Our results demonstrated that CA exhibited strong anti‐inflammation effect in OA FLS through blocking the activation of TLR4/MyD88 signalling pathway, suggesting its potential as a hopeful candidate for the development of novel agents for the treatment of OA.


| INTRODUC TI ON
Osteoarthritis (OA), the most common degenerative joint disease and the leading cause of joint disability and persistent pain, affects 140 million people worldwide. 1 OA is characterized by synovial inflammation, progressive articular cartilage damage and thickening of the subchondral bone that leads to pain, swelling, dysfunction of the joint and diminished quality of life. [2][3][4] For a long time, articular cartilage degeneration was considered to play a crucial role in the progression of OA. 5,6 It was also reported that OA is an autoinflammatory disease caused by chondrocyte-mediated inflammatory responses. 5 With the deepening of our understanding of the course of OA, however, the role of soft tissues around the joint, such as synovial tissue and fat pads, in the pathogenesis of OA is increasingly recognized. 3,[7][8][9] Synovial inflammation was found to be highly correlated with several OA symptoms such as stiffness and pain.
Recent studies suggest that synovial inflammation precedes damage to other articular tissues in the development of radiographic OA, 10 and the exosomes derived from OA fibroblast-like synoviocytes (FLS) induce cartilage degeneration. 11 FLS are the most abundant cell in synovial tissue and play a crucial role in the process of OA development. 12, 13 Wang et.al. 3 suggested that synovial tissue can affect articular cartilage degeneration through the innate immune system and affect OA progression. Therefore, effective inhibition of synovial inflammation may be the key to delaying cartilage damage and OA progression.
Cinnamon is a widely used traditional Chinese herbal medicine in treatment of OA. According to statistics, Cinnamon is one of the most commonly used traditional Chinese medicines in the treatment of OA. 14,15 Cinnamic aldehyde (CA), a critical bioactive substance extracted from Cinnamon, is known to have antioxidant, antidiabetic, anti-inflammatory, antipyretic and anticancer properties. [16][17][18][19][20] Our previous results show that CA can delay the degeneration of chondrocytes by inhibiting the NF-kB signalling pathway. 2 However, the effect of CA on OA synovial inflammation and its underlying mechanism are still unclear. Network pharmacology, an increasingly developed system for building disease-drug networks, can effectively and systematically study the pharmacological effects mechanism of action and safety of herbal medicines, especially traditional Chinese medicine (TCM). 21,22 In recent years, network pharmacology has also been used to predict potent therapeutic targets and regulate pathways of single natural products or their combinations in complex diseases. 23 Therefore, in this study, we aimed to explore the mechanism underlying the effect of CA on synovial inflammation in OA using techniques from network pharmacology to experimental pharmacology.

| Cinnamic aldehyde-associated gene mining
The PubChem, HIT and TCMSP databases were used to identify CA-associated genes. PubChem (https://pubch em.ncbi.nlm.nih.gov) is the world's largest collection of freely accessible chemical information. The HIT database contains protein targets and potential therapeutic significance for supplementing FDA-approved drugs. 24 TCMSP 25 is a unique systems pharmacology platform of Chinese herbal medicines that captures the relationships between drugs, targets and diseases. It includes chemicals, targets and drug-target networks as well as associated drug-target-disease networks. Microsoft Excel was used to integrate the obtained CA-associated genes.

| Osteoarthritis-associated gene mining
The Therapeutic Target Database (TTD), DrugBank and GeneCards database were used to perform OA-associated gene mining. The Therapeutic Target Database (TTD) 26 is a database that provides information about known and explored therapeutic protein and nucleic acid targets, the targeted disease, pathway information and the corresponding drugs directed at each of these targets. DrugBank 27 is a unique bioinformatics and cheminformatics resource that combines detailed drug data with comprehensive drug target information. GeneCards (www.genec ards.org) is a gene database that integrates network resources and contains comprehensive information about all human gene annotations and predictions. Microsoft Excel was also used to integrate the obtained OA-associated genes.

Protein-protein interaction (PPI) network constructed
A Venn diagram online website (http://bioin forma tics.psb.ugent.be/ webto ols/Venn/) was used to extract the intersection (co-targeting genes) between OA-related genes and CA-related genes. These co-targeting genes are considered to be potential target genes for CA treatment of OA. The Database for Annotation, Visualization, and Integrated Discovery (DAVID) 28 was used to perform Kyoto Encyclopedia of Genes and Genomes (KEGG) 29 analysis and Gene ontology (GO) 30 functional enrichment analysis. The search tool for the retrieval of interacting genes (STRING) database 31 was used to construct the protein-protein interaction (PPI) networks between these co-targeting genes and then visualize in Cytoscape software K E Y W O R D S bioinformatics analysis, cinnamic aldehyde, experimental pharmacology, network pharmacology, osteoarthritis, synovial inflammation, TLR4/MyD88 signalling pathway (http://www.cytos cape.org/). Module analysis of these co-targeted genes was performed to identify the closely connected module using molecular complex detection (MCODE). 32 The criteria were set as follows: MCODE scores >5 and number of nodes >10. In addition, KEGG pathway enrichment analysis of molecular complex was also performed.

| Molecular docking
The CA structure was downloaded from the TCMSP database; subsequently, the Open Babel GUI was used to convert the format from mol2 to PDB. The PDB format of TLR2 and TLR4 was downloaded from the RCSB database (http://www.rcsb.org). PyMOL software was used to remove the solvent and organic compounds. AutoDock tools 1.5.6 was used to prepare, run and analyse the docking simulations, and the docking results were analysed and visualized using AutoDock tools and PyMOL software.

| Human osteoarthritis fibroblast-like synoviocytes isolation and culture
The study was conducted according to the Declaration of Helsinki,  Table S1). Briefly, the obtained synovial tissue obtained was immediately placed into PBS (Solarbio) containing antibiotic mixture (Penicillin and Streptomycin, Invitrogen/Thermo-Fisher Scientific), inserted into the ice box and quickly transferred to ultra-clean table.
The samples were washed thrice in PBS, which included antibiotic mixture, and then diced into 1 × 1 × 1 mm 3 pieces. These synovial tissue pieces were digested with 0.25% trypsin-EDTA solution (Solarbio) for 30 min, digested in 0.2% type II collagenase (Solarbio) for 6 h, centrifuged at 176 g for 5 min and the supernatant was discarded. FLS were resuspended in DMEM-H (Hyclone) with 10% FBS (ABW) and 1% antibiotic mixture. The FLS were plated at a density of 1 × 10 5 cells/ml and incubated at 37°C in a humidified 5% CO 2 atmosphere. Only passages 2-6 were used to avoid phenotype loss.

| Quantitative real-time PCR analysis
Quantitative real-time PCR analysis (QPCR) was performed as described previously.  Table 1. And data were analysed using 2 −ΔΔCT method.

| Western Blot analysis
As described previously, 2,7 cell lysates were prepared from FLS with RIPA lysis buffer kit (Beyotime Biotechnology), and the

TA B L E 1 The primer's sequences of the targeted genes Primers
Sequences protein concentrations were quantified using a BCA protein assay kit (Thermo Scientific). Next, 8%-15% SDS-PAGE gels were used to separate the cell lysates, and the samples were transferred to polyvinylidene fluoride (PVDF) membranes (Millipore). Subsequently, the membranes were incubated with primary antibodies anti-IL-1β, IL-6, TNFα, TLR4, MyD88 (Abcam) and β-actin (Proteintech) overnight at 4°C. After washing thrice with TBST, the membranes were incubated with Horseradish Peroxidase (HRP)-labelled secondary antibodies (Proteintech) for 1 h at room temperature. Images were developed after reaction with a high-sensitivity chemiluminescence reagent (Proteintech).

| Statistical analysis
All the results were represented as the median ± standard deviation (SD). All analyses were performed using GraphPad prism 5.0 software. Two different groups were compared by independent-sample t test, and multiple group comparisons were performed by one-way analysis of variance (ANOVA) analysis. p < 0.05 was considered statistically significant. All experiments were performed at least three times independently.

| Identifying CA-OA co-targeted genes
The study flowchart is presented in Figure 1A, 3093 OA-targeted genes were identified with TTD, DrugBank and Genecards databases, and 270 CA-targeted genes were identified with TCMSP, PubChem and HIT databases. The 3D structure of CA was download in PubChem database ( Figure 1B). The Venn diagram results showed that a total of 144 co-targeted genes were identified, which are considered potential therapeutic target genes for CA in OA treatment ( Figure 1C). The details of these co-targeted genes are list in Data S1.

| GO and KEGG pathway enrichment analysis of co-targeted genes
The GO and KEGG pathway analyses were performed to holistically

| Protein-protein interaction network constriction and Module analysis
The PPI network of these co-targeted genes is shown in Figure 3A; genes with higher degree have redder colours, while the lower the yellow. The top 3 hub genes were IL6, TNF and AKT1, indicating that these genes may be the target genes for the CA treatment of OA. Two modules were identified in these co-targeted genes with MCODE scores >5 and number of nodes >10. The first module (MCODE score = 42.308) included 53 nodes and 1100 edges, and the KEGG pathway analysis results revealed that these co-targeted genes in modules were significantly enrichment in 'Toll-like receptor signalling pathway', 'TNF signalling pathway', 'IL-17 signalling pathway', etc. (Figure 3B,C). The second module (MCODE score = 7.250) included 25 nodes and 87 edges. As is shown in Figure 3D,E, the KEGG pathway analysis of this module suggested that 'toll-like receptor signalling pathway', 'TNF signalling pathway', 'osteoclast differentiation', etc., were significantly regulated by CA in the prevention of OA. Detailed information on these two modules is presented in Data S6-S8.
As the toll-like receptor signalling pathway was successfully enriched in thrice KEGG analyses, the top 3 hub genes were all included in this pathway. Therefore, we use molecular docking technology to simulate the combination of CA with TLR2 and TLR4, which are the most studied TLR family members. As is shown in Figure

| Cinnamic aldehyde inhibits the expression of inflammatory factors in LPS-induced human FLS
The As is shown in Figure 5B

| Cinnamic aldehyde inhibits the inflammation in LPS-induced human OA FLS via blocking the TLR4/ MyD88 signalling pathway
Our previous study 3 suggested that synovial inflammation can activate the cartilage innate immune system through the TLR4/MyD88 signalling pathway, thereby affecting the progression of OA. And the CA and TLR4 had better binding energy than TLR2. Therefore, we investigated the expression of TLR4 and MyD88 in LPS-induced FLS. As is shown in Figure 6A Figure 6D). As presented in Figure 6E-G and Figure S1,

| DISCUSS ION
Osteoarthritis is one of the most common degenerative bone and joint diseases, especially in the knee joint. 7,33 Total joint arthroplasty   Collectively, this study is the first to demonstrate the antiinflammatory effects of CA on OA synovial inflammation from the network pharmacology and bioinformatics analysis to experimental pharmacology. In LPS-induced human OA FLS, CA significantly inhibit OA synovial inflammation via blocking the TLR4/MyD88 signalling pathway. Our study provides new insights into the investigation of the anti-inflammatory effects of CA in OA using integrative pharmacology-based approaches, and the results suggest that CA may be a potential therapeutic agent for OA.

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

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.