Resistin Induces Chemokine and Matrix Metalloproteinase Production via CAP1 Receptor and Activation of P38-MAPK and NF-κB Signalling Pathways in Human Chondrocytes

Resistin is an adipokine also detected higher expression in serum and synovial uid of patients with knee osteoarthritis (KOA). Resistin is known to be related closely to insulin resistance and inammation. However, the pathogenic role of resistin in KOA remain unclear. Purpose of the study is to investigate whether resistin induces KOA by binding to functional receptor adenylyl cyclase-associated protein 1 (CAP1) and activating the p38 mitogen-activated protein kinase (p38-MAPK) and nuclear factor-κB (NF-κB) signalling pathways in human chondrocytes. via hip quantitative chain reaction ( western and were evaluated via and . The roles of CAP1, p38-MAPK, and NF-κB signalling pathways in the were evaluated via qRT-PCR, western and CAP1 CCL3, CCL4, MMP13, and ADAMTS-4 by chondrocytes stimulated with resistin. Results of the Co-IP assay show that resistin directly bound CAP1, and resistin treatment increased binding. investigated the function of CAP1 using a loss ‐ of ‐ function approach and culturing chondrocytes in the presence of resistin. Knockdown of CAP1 downregulated CCL3, CCL4, MMP13, and ADAMTS-4 at the mRNA level, and CCL4 and MMP13 at the protein level. The p38-MAPK and NF-κB pathways were activated by resistin in chondrocytes; however, this inhibited by knockdown of CAP1.


Abstract Background
Resistin is an adipokine also detected higher expression in serum and synovial uid of patients with knee osteoarthritis (KOA). Resistin is known to be related closely to insulin resistance and in ammation.
However, the pathogenic role of resistin in KOA remain unclear. Purpose of the study is to investigate whether resistin induces KOA by binding to functional receptor adenylyl cyclase-associated protein 1 (CAP1) and activating the p38 mitogen-activated protein kinase (p38-MAPK) and nuclear factor-κB (NF-κB) signalling pathways in human chondrocytes.

Methods
We enrolled 103 patients with radiographic KOA and 86 healthy participants as controls. The levels of resistin in serum and synovial uid (SF) were determined via enzyme-linked immunosorbent assay (ELISA). CAP1 expression in cartilage tissues (21 samples of KOA cartilage and 10 samples of healthy hip cartilage) was measured using immunohistochemistry (IHC), quantitative real-time polymerase chain reaction ( qRT-PCR ), and western blotting assays. Effects of resistin on chondrocytes and CAP1 were evaluated via qRT-PCR and co-immunoprecipitation . The roles of CAP1, p38-MAPK, and NF-κB signalling pathways in the development of KOA were evaluated via adenovirus-mediated CAP1 short hairpin RNA, qRT-PCR, western blotting, and ELISA.

Results
Expression of resistin in serum and SF was elevated in severe radiographic KOA. CAP1 levels were higher in KOA cartilage and were positively correlated with resistin expression. Resistin promoted increased expression of CCL3, CCL4, MMP13, and ADAMTS-4 through CAP1 receptor. Resistin also directly bound to CAP1 as con rmed by co-immunoprecipitation. CAP1 knockdown in chondrocytes attenuated resistininduced expression of CCL3, CCL4, MMP13, and ADAMTS-4 and activation of the p38-MAPK and NF-κB signalling pathways .

Conclusions
Our study shows that resistin bound to CAP1 and upregulated the expression of proin ammatory cytokines and matrix-degrading enzymes via p38-MAPK and NF-κB signalling pathways in human chondrocytes.

Background
Knee osteoarthritis (KOA) is a joint disease involving articular cartilage, subchondral bone, ligaments, capsule and surrounding synovial tissues [1,2]. Main symptoms of KOA are pain and restricted mobility [3][4][5]. Progressive deterioration of articular cartilage and loss of extracellular cartilage matrix are key features of KOA. The different KOA phenotypes are classi ed according to clinical, imaging, and laboratory characteristics, and may represent different aetiopathogeneses of the disease [6][7][8]. Therapy is chosen based on phenotype to prevent the development of grade III/IV KOA. Despite signi cant advances, a substantial percentage of KOA patients are unsatis ed with their outcomes [9]. To prevent the onset and progression of KOA, it is imperative to determine the underlying mechanisms.
Although OA has traditionally been classi ed as a nonin ammatory disease, an increasing proportion of patients present with synovial in ammation [10][11][12]. Numerous proinflammatory cytokines contribute to the progression of OA [13]. Although resistin is considered as proin ammatory cytokine in OA, few studies have evaluated the underlying mechanisms. Resistin, an adipokine associated with insulin resistance and in ammatory processes [14,15], is highly expressed in synovial joints of OA patients [16][17][18][19][20]. Studies have shown that resistin upregulates proin ammatory cytokines and matrix-degrading enzymes in human articular chondrocytes [21,22]. Additionally, arthritis is induced by injecting resistin into the knee joints of healthy mice [23]. Although numerous studies have examined resistin levels in sera and synovial uid (SF) of OA patients, few have investigated the proin ammatory mechanisms of resistin. Previously, we described the proin ammatory role of resistin in osteoarthritis, but did not determine the mechanisms involved [24].
Adenylyl cyclase-associated protein 1 (CAP1) localizes to the cell membrane and cytoplasm, being a receptor for resistin and modulating the in ammatory activity in monocytes [25]. Silencing of CAP1 in broblast-like synoviocytes (FLSs) downregulates resistin-stimulated production of chemokines, such as CXCL8, CCL2, and IL-6[26]. CAP1 plays an important role in resistin-induced inflammation; however, its exact role in the progression of OA remains unclear.
Resistin-mediated proin ammatory activity involves activation of the nuclear factor-κB (NF-κB) pathway [27,28]. In human chondrocytes, increased expression of resistin-induced in ammatory chemokines is also mediated by the NF-κB pathway [21]. Resistin may induce the expression of proin ammatory cytokines and matrix catabolic enzymes in human nucleus pulposus cells via activity of the p38-MAPK and NF-κB pathways [29,30]. The characteristics of human articular chondrocytes are similar to those of human nucleus pulposus cells; hence, activation of p38-MAPK and NF-κB pathways may be involved in the function of resistin in human chondrocytes. Here, we evaluated resistin level in patients with KOA and healthy control subjects, and expression of CAP1 between cartilage of KOA patients and normal cartilage in patients with femoral neck fracture (FNF-Cartilage). We also explored the proin ammatory effects of resistin, and involvement of p38-MAPK and NF-κB signalling in the resistin-CAP1 axis in human chondrocytes. To our knowledge, this is the rst study exploring the mechanism of resistin/CAP1 in human KOA chondrocytes. in

Participants
Overall, 103 patients with KOA, diagnosed according to criteria set by the American College of Rheumatology, were recruited from May 2017 to June 2018 at the Orthopedics Department of the rst hospital of Jilin University. After a 12 h fast, venous blood and SF samples were collected from patients and 86 healthy volunteers at their annual examinations conducted at our Health Evaluation Center of the rst hospital of Jilin University. All collected blood and synovial uid samples were centrifuged at 3000g for ten minutes at 4 °C. The supernatants of blood and synovial uid were aliquoted and stored at -80 °C until assay. Clinical data of KOA patients was reviewed to exclude patients treated with intraarticular glucocorticoid or hyaluronic acid injections within 6 months. This study was approved by the Ethics Committee of the rst hospital of Jilin University (reference no: 2017-429). All participants provided signed informed consent, and the study was performed in accordance with the Declaration of Helsinki. All data pertaining to study participants are presented in Table I.

Collection of human cartilage samples
Samples of KOA cartilage were obtained from primary total knee replacement (TKA) surgery patients. Samples of control cartilage were collected from patients with femoral neck fracture (FNF) within 12h of injury, and articular cartilage of femoral heads were macroscopically intact and showed little degeneration regardless of age (Collins score 0), as evaluated at the Orthopedics Department of the rst hospital of Jilin University. Collection of all cartilage samples had the approval of the permission of the patients according to the Declaration of Helsinki. All sample details are listed in Table . Isolation and culture of human chondrocytes Nine cartilage tissue samples were harvested randomly from patients with KOA (N=21) and divided into three groups randomly. Three cartilage tissue samples of each group were pooled prior to cell isolation and culture. Chondrocytes were isolated as previously described [31] and cultured in Dulbecco's Modi ed Eagle's Medium (DMEM)/F12 supplemented with 10% foetal bovine serum (FBS) and 1% penicillin/streptomycin in a CO 2 incubator at 37 °C. Then, (DMEM)/F12 media were replaced with serumfree media, and chondrocytes were starved 12 h before treatment with resistin at different concentrations and time-points. Chondrocytes at passage one and two were used in our study.

Enzyme-linked immunosorbent assay (ELISA)
Concentration of resistin in samples of serum and SF was quantitated using ELISA kits (CUSABIO, China; CSB-E06884h) according to manufacturer's instructions. The range of detection was 0.312-20 ng/mL.

Immunohistochemical assessment of cartilage tissues
A total 10 control cartilage samples from patients with FNF and 21 KOA cartilage samples were used to evaluate the expression of resistin and CAP1 via IHC. Specimens were embedded in para n and sectioned at 3 μm. Each section was rehydrated in ethanol (100, 90, and 80%) for 5 minutes, respectively, heated at 95°C for 10 minutes in 10 mM sodium citrate buffer solution (pH 6.0) for antigen recovery, and then treated with 3% (v/v) H2O2 for 15 min to block endogenous peroxidase activity. After several washes with PBS, sections were incubated with serum-free protein block (Agilent Technologies) for 30 min to block non-speci c binding. Sections were then incubated at 4°C overnight with mouse anti-resistin monoclonal antibody (concentration at 10 μg/ml, ab136877; Abcam), rabbit anti-CAP1 monoclonal antibody (diluted at 1:200, EPR8339B; Abcam), or isotype control (Agilent Technologies). Expression level was detected using a Mouse and Rabbit Speci c HRP/AEC (ABC) Detection IHC Kit (Abcam), after which the sections were counterstained with hematoxylin.

Co-immunoprecipitation assays (Co-IP)
Chondrocytes were cultured in 6-well plates (at 2 × 10 5 cells/well) as described previously with and without resistin (at 500 ng/ml) for 48 h. Cells were then washed with ice-cold phosphate buffered saline (PBS), and protein was harvested using a cell lysis buffer (Beyotime, China). CAP1 coimmunoprecipitation was performed using a Pierce™ Co-Immunoprecipitation Kit (Thermo Scienti c) according to manufacturer's instructions. For immunoprecipitation, cell lysates were incubated overnight on ice with CAP1 antibody labelled beads to examine binding of resistin and CAP1. Rabbit IgG was used as negative control. After overnight incubation, the eluted protein was detected by western blotting using an antibody against resistin (Abcam, 1:1000).

Knockdown of CAP1 via adenovirus (Adv)-mediated transfection in chondrocytes
A recombinant Adv vector was generated by cloning shRNA fragments into an Adv vector GV119 (Shanghai Genechem Co., Ltd, China) via enzymatic digestion and ligation. Three shRNA sequences were designed to silence the expression of CAP1 in human chondrocytes in vitro. One of the CAP1-shRNAtargeting sequences was 5′-CCTGGCCCTTATGTGAAAGAA -3′, whereas the negative control shRNA sequence was 5′-TTCTCCGAACGTGTCACGT-3′.
Chondrocytes were transfected with CAP1-shRNA or control-shRNA for 48 h, and treated with resistin (500 ng/mL) for another 48 and 72 h. All experiments were performed at a multiplicity of infection (MOI) of 40 plaque forming units (PFU). The e ciency of CAP1 knockdown was con rmed by qRT-PCR and western blotting. Expression of CCL3, CCL4, MMP13, and ADAMTS-4 mRNA was assessed via qRT-PCR. The level of secreted proteins in culture supernatants was determined using ELISA (RayBiotech, USA) according to manufacturer's instructions.
Total RNA extraction and quantitative real-time PCR Total RNA was extracted from chondrocytes and human cartilage and reverse-transcribed into cDNA.
The qRT-PCR was conducted as described previously [32]. Primers used for qRT-PCR are shown in Supplementary materials I. The Ct values for target genes were calculated as described previously [22].

Statistical analysis
Results are presented as mean ± standard deviation. Independent t-test and Wilcoxon Rank Sum test were used to compare values of KOA and HC groups. Levels of serum resistin in the two groups were evaluated by analysis of covariance after adjustment for age, body mass index (BMI), fasting blood glucose levels (FBG), and triglyceride levels (TG). Pearson correlation coe cient was used to analyse correlations between resistin and CAP1 expression for all IHC results. Paired t-test was used to compare expression of CCL3, CCL4, MMP13, and ADAMTS-4 between control-shRNA and CAP1-shRNA groups. For testing differences among more than two groups, one-way ANOVA was rst used to test the overall difference and followed by Bonferroni post-hoc test in various subgroups. p < 0.05 was considered statistically signi cant, and all statistical analyses were performed using SPSS 22.0 (IBM).

Resistin levels in the serum and SF of patients with KOA and HCs
Characteristics of study participants are presented in Table I. Serum resistin levels in patients with KOA (8.26 ± 0.29, ng/mL) were slightly higher than those in healthy controls (7.45 ± 0.30, ng/mL); however, no signi cant difference was detected after adjusting for age, BMI, FBG, and TG (Padj = 0.688) (Fig. 1A). Higher serum resistin levels were found in female KOA patients than female control groups (P = 0.044), whereas there was no signi cant difference in male subjects between the two groups (Fig. 1B). Compared with resistin levels in the SF, serum resistin levels in patients with KOA were higher regardless of gender differences (P < 0.0001) (Fig. 1C-D). The serum resistin levels of grade III patients did not differ from grade II and IV patients (P = 0.321 and P = 0.096, respectively); however, the levels of grade II and IV patients differed signi cantly (P = 0.021) (Fig. 1E). Female KOA patients with grade had signi cantly higher serum resistin level than grade patients (P = 0.044), however no signi cant difference between different grades of male KOA patients (Fig. 1F). The SF resistin levels in grade IV patients were higher than those in grade III and II patients (P < 0.0001 and P < 0.0001, respectively) (Fig. 1G). No matter KOA patients were male or female, patients with grade IV had higher SF resistin levels than both grade III and II patients (male: IV versus III, P = 0.002 and IV versus II, P = 0.009; female: IV versus III, P < 0.0001 and IV versus II, P < 0.0001, respectively) (Fig. 1H). Results indicate that higher resistin levels are associated with severe KOA.

Expression of resistin and CAP1 in cartilage tissues
Expression of CAP1 (Fig. 2E, F, and M) and resistin (Fig. 2K, L, and N) was higher in KOA cartilage than that in control cartilage, based on IHC. In 21 KOA samples, CAP1 and resistin expressions showed a signi cant positive correlation (r=0.631, p=0.002) (Fig. 2O). Compared with control cartilage, the KOA cartilage mRNA expression of CAP1 changed 2.9-fold relative to GAPDH (p<0.001) (Fig. 3A), and the expression of CAP1 also increased (p<0.001) ( Fig. 3B and C). These ndings show that CAP1 expression was elevated in KOA cartilage.

Effect of resistin on chondrocytes and CAP1
As shown in Fig. 4A-D, in chondrocytes treated with resistin (500 ng/mL), expressions of CCL3 and MMP13 peaked at 48 h, whereas expressions of CCL4 and ADAMTS-4 peaked at 24 h and gradually declined from there on. Treatment with resistin also induced CCL3, CCL4, MMP13, and ADAMTS-4 expression in a dose-dependent manner; however, MMP13 levels decreased slightly after resistin was administered at 1000 ng/mL, and levels of CCL3 and ADAMTS-4 increased slowly after resistin was administered at 500 ng/mL (Fig. 4E-H). Expression of CAP1 was increased when chondrocytes were stimulated by resistin with dose of 1000 ng/ml at 48 h and 72 h, but no changes were observed at lower dose (Fig. 4I).
A Co-IP assay was performed to detect resistin and CAP1 binding. Results indicate that resistin directly interacted with CAP1, and binding was signi cantly increased in chondrocytes treated with resistin ( Fig.  4J-K).
Adv-mediated knockdown of CAP1 reduces stimulatory effects of resistin in chondrocytes Infection e ciency of control-shRNA and CAP1-shRNA (MOI=40) in chondrocytes was greater than 80% as assessed using uorescence microscopy at 48 h post-transfection (Fig. 5A-F). As shown by qRT-PCR and western blotting, CAP1 expression decreased signi cantly in the CAP1-shRNA group compared to control-shRNA group (Fig. 5G-I). Compared with the expression pro le of the control-shRNA group, mRNA expression of CCL3, CCL4, MMP13, and ADAMTS-4 was downregulated in the CAP1-shRNA group when cultured with resistin (500 ng/ml) for 48 and 72 h (Fig. 5J-M). As shown by ELISA, CAP1 knockdown decreased expression of CCL4 (at 48 h after treatment with resistin) (Fig. 5O) and MMP13 (at 72 h after treatment with resistin) (Fig. 5P), whereas signi cant differences were not observed in the expression of CCL3 or ADAMTS-4 (Fig. 5N, Q). These results show that CAP1 plays a key role in resistin-induced production of CCL3, CCL4, MMP13, and ADAMTS-4 on mRNA level and CCL4 and MMP13 on protein level in chondrocytes.

Resistin activates p38-MAPK and NF-κB signalling pathways in chondrocytes
To determine whether p38-MAPK and NF-κB pathways participate in the proin ammatory response in chondrocytes, we examined untreated chondrocytes and those treated with resistin. After treatment, a signi cant increase of phosphor-p38 and phosphor-p65 was found. Resistin-activated protein expression of phospho-p38 peaked 6 h post-treatment and remained level until 24 h post-treatment, whereas activation of phospho-p65 occurred 60 min post-treatment ( Fig. 6A-B). This indicates that resistin exerts its proin ammatory function by activation of p38-MAPK and NF-κB pathways in chondrocytes.

Knockdown of CAP1 inhibits p38-MAPK and NF-κB signalling pathways
A loss-of-function study was performed to examine whether CAP1 regulates activation of p38-MAPK and NF-κB pathways in response to resistin. Western blot assay showed that CAP1 expression in chondrocytes transfected with CAP1-shRNA was abolished by 90% compared with the control-shRNA group. Expression of phosphor-p38 (p-p38) and phosphor-p65 (p-p65) was signi cantly inhibited in the CAP1-knockdown group compared with the control group with or without resistin treatment. Consistent with the above ndings, the CAP1-knockdown group showed decreased levels of non-phospho-p38(p-38).
However, non-phospho-p65 (p65) expression was not signi cantly reduced in CAP1-knockdown group untreated with resistin compared with those treated with resistin and the control-shRNA groups (Fig. 6C-D). Our results suggest that resistin activates the p38-MAPK and NF-κB signalling pathway by binding CAP1 in human chondrocytes.

Discussion
Here, higher levels of resistin were found in the sera and SF of patients with advanced grade (K/L) KOA. The higher CAP1 expression detected in KOA cartilage tissues correlated positively with higher resistin expression. We demonstrated that CCL3, CCL4, MMP13, and ADAMTS-4 were produced by chondrocytes stimulated with resistin. Results of the Co-IP assay show that resistin directly bound CAP1, and resistin treatment increased binding. We investigated the function of CAP1 using a loss-of-function approach and culturing chondrocytes in the presence of resistin. Knockdown of CAP1 downregulated CCL3, CCL4, MMP13, and ADAMTS-4 at the mRNA level, and CCL4 and MMP13 at the protein level. The p38-MAPK and NF-κB pathways were activated by resistin in chondrocytes; however, this was inhibited by knockdown of CAP1.
Resistin may act in a proin ammatory manner, and has been linked to the development of OA[16, 17,19,23]. Resistin levels are associated with the severity of symptoms and radiological cartilage degeneration in KOA patients [33][34][35]. Consistent with previous studies, serum levels of resistin were higher in KOA patients than in healthy controls, although no signi cant difference between the two was found after adjustment. This indicates that serum resistin is not a reliable biomarker for the prediction of OA because it may be affected by other factors [36][37][38][39][40]. The levels of SF resistin were far more related than serum levels to the severity of radiographic damage in KOA in agreement with Koskinen, who showed that SF resistin is associated with in ammatory and catabolic factors, and plays a role in KOA pathogenesis [41].
Similar to previous study, in our current study, the levels of resistin also differ regarding gender [42]. Gender differences in serum resistin level between KOA patients and healthy controls may be due to a greater abundance of monocytes in adipose tissue which was more in female patients or the physiological role of hormone between genders, however further studies must be performed to explore these. Additionally, a separate analysis for male and female are also conducted to con rm the above results that serum and SF resistin levels were close correlation with KOA. Thus, these ndings indicated us to furtherly explore the mechanism of resistin in the KOA pathogenesis.
Previous studies by Lee et al. and Munjas et al. have shown that in monocytes, CAP1 acts as a functional receptor directly binding resistin and is involved in modulating the proin ammatory activity of resistin [25,43]. Sato et al. found that CAP1 expression is higher in the synovial tissue of patients with RA compared with those with OA, and resistin increases production of chemokines by binding to CAP1 [26]. We demonstrated that CAP1 expression is increased in KOA cartilage compared with control cartilage. Additionally, resistin was highly expressed in KOA cartilage but not in control cartilage. Assessment of KOA cartilage tissues via H-Score and IHC indicated that CAP1 expression was positively associated with resistin. Consistently, in vitro chondrocytes experiment, the upregulation of CAP1 expression also could be induced by higher concentration of resistin. Besides, our Co-IP assay showed that resistin directly bound to CAP1, and binding was increased after treatment with resistin. These ndings indicate that resistin may induce KOA via binding to CAP1 receptor.
Consistent with Zhang et al [21,22], we found that resistin upregulated the CCL3, CCL4, MMP13, and ADAMTS-4. This elevated pattern of expression is implicated in the pathogenesis of OA [44][45][46][47]. Knockdown of CAP1 inhibited resistin-induced expression of CCL3, CCL4, MMP13, and ADAMTS-4 at the mRNA level, and CCL4 and MMP13 at the protein level. Conversely, resistin-induced effects on protein and mRNA expression were inconsistent after the silencing of CAP1 in KOA chondrocytes. This may have occurred because CAP1 is not a unique receptor for resistin in human chondrocytes. Current studies show that resistin has another receptor, TLR-4. Therefore, silencing individual receptors only partially abolishes the effects of resistin, although it was only reported in human nucleus pulposus cells and macrophages [30,48]. We also cannot rule out additional receptors, such as decorin and ROR1, which were not yet found in humans [49,50]. Although decorin and ROR1 were only reported to play a role in resistin-induced in ammation for mouse, potential role could not be ignored in human chondrocytes. Hence, more future investigations were needed to uncover role of TLR4, decorin and ROR1 in human chondrocytes. But, for now at least, our present study identi ed that CAP1 acts as a receptor of resistin in chondrocytes, and regulates the resistin-induced pathological progression of KOA.
Activation of signalling pathways is key in the resistin-CAP1 axis. Binding to CAP1 receptor occurs rst followed by signal transduction from cytoplasm to the nucleus, causing transcription of genes. The p38-MAPK and NF-κB pathways are crucial in phenotype loss, in ammatory response, catabolism, and extracellular matrix degradation [51][52][53][54]. Resistin increases CCL3 and CCL4 production in human chondrocytes by activating the NF-κB signalling pathway [21]. In chondrocyte-like human nucleus pulposus cells, p38-MAPK is activated by resistin to upregulate the expression CCL4 and ADAMTS-5[29, 30]. We found that resistin activated the p38-MAPK and NF-κB pathways via phosphorylation of p38 and p65 in chondrocytes. Furthermore, CAP1 knockdown inhibited the resistin mediated increase of p38-MAPK and NF-κB activity. We showed that p38-MAPK and NF-κB pathways participate in resistin-CAP1 activity in human chondrocytes.
In the present study, we rstly discovered that CAP1 expression was signi cantly increased in KOA cartilage. Furthermore, in human KOA chondrocytes, we also proved that CAP1 was a resistin-binding receptor, and knockdown of CAP1 inhibited resistin-induced products of proin ammatory cytokines and matrix-degrading enzymes. Meanwhile, the p38-MAPK and NF-κB signalling pathways were found to be activated by resistin-CAP1 axis, and modulate human chondrocytes to lead chronic in ammation and matrix degradation.
However, there were some potential limitations in our study. Firstly, control cartilages were derived from femoral head of FNF patients in our study, since it was hard to obtain healthy knee cartilages. However, previous studies had shown that FNF-cartilage was very similar to that of normal mature articular cartilage as previous literature described [55] and were used as control in many researches[56-60]. Secondly, overexpression of CAP1 in human chondrocytes should be examined to better identify role of CAP1 in resistin-induced pathologic progression of KOA. Thirdly, there were no in vivo experiments in our study. Further research is needed to involve in vivo animal experiment in order to verify function of resistin/CAP1 in KOA.

Conclusion
In conclusion, we showed that resistin induced the expression of proin ammatory cytokines and matrixdegrading enzymes by binding to CAP1 and activating the p38-MAPK and NF-κB signalling pathways in human KOA chondrocytes (Fig. 7). These ndings may serve as foundation for the development of novel resistin-targeted therapies for the treatment of patients with KOA.

Competing interests
The authors declare that they have no competing interests.

Funding
This research did not receive any speci c grant from funding agencies in the public commercial.     Co-immunoprecipitation showed that the binding of resistin to CAP1 was signi cantly increased in the group treated with resistin (labelled as R) compared with those treated with IgG and control (labelled as IgG and C) (J, K); quanti ed Co-IP results in K were normalized to CAP1.
Values are shown as mean ± SD for three experiments conducted using cells from three patients with Page 19/20 KOA. The relative expression compared with no resistin treatment is shown. Each experiment was performed in triplicate. *, p<0.05; **, p<0.01; ***, p<0.001.   The pathogenic mechanism of resistin in KOA Resistin promoted the expression of proin ammatory cytokines (CCL3, CCL4) and matrix-degrading enzymes (MMP-13, ADAMTS-4) by binding to CAP1 and activating the p38-MAPK and NF-κB signalling pathways in human chondrocytes. The great release of