Title: Fibroblast growth factor 18 exerts anti-osteoarthritic effects through PI3K-AKT signaling and mitochondrial fusion and fission

Abstract

Osteoarthritis (OA) is a degenerative disease characterized by progressive loss of cartilage, osteophyte formation and subchondral bone sclerosis. Although some animal experiments have reported that fibroblast growth factor 18 (FGF18) attenuates cartilage degradation, the effect of FGF18 on chondrocytes and its underlying mechanism at the cellular level remain largely unknown. In this study, we found that an intra-articular injection of FGF18 attenuates cartilage degradation, increases Collagen II deposition and suppresses matrix metallopeptidase 13 (MMP13) expression in rat post-traumatic osteoarthritis (PTOA). At the cellular level, FGF18 promotes chondrocyte proliferation through PI3K-AKT signaling and migration through PI3K signaling. We found that FGF18 attenuates IL-1β-induced apoptosis, restores mitochondrial function and reduces Reactive Oxygen Species (ROS) production through PI3K-AKT signaling.

Moreover, the mitochondrial fusion and fission of chondrocytes were enhanced by a short duration of treatment (within 24 hours) of IL-1β and suppressed by prolonged treatment (48 hours). FGF18 significantly enhances the mitochondrial fusion and fission, restoring mitochondrial function and morphology, and reduces ROS production. We also found that the FGFR1/FGFR3 ratio, which might contribute to the progression of osteoarthritis, was upregulated by IL-1β and downregulated by FGF18. To the best of our knowledge, our data demonstrated the anti-osteoarthritic effect of FGF18 at the cellular level for the first time and suggested that PI3K-AKT signaling and
mitochondrial fusion and fission might play critical roles during the process. Our study proved that FGF18 might be a promising drug for the treatment of early stage osteoarthritis and is worth further study.

Keywords: Fibroblast growth factor 18; osteoarthritis; mitochondrial; apoptosisIntroduction

Osteoarthritis (OA) is a degenerative disease characterized by progressive loss of cartilage, osteophyte formation and subchondral bone sclerosis. With the increase of life expectancy, OA has become one of the most prevalent chronic diseases, with an increasing trend [1]. Although the progression of OA is slow, it can ultimately lead to significant pain and disability requiring joint replacement. Current therapies for OA are focused on pain relief and improvement of joint function such as non-steroidal anti-inflammatory drugs (NSAIDs) and intra-articular injections of steroids and hyaluronic acid. However, drugs that attenuate cartilage degradation and restore the construction of cartilage are still lacking. Therefore, disease modifying osteoarthritis drugs are now the focus in the therapy of osteoarthritis [2].

Fibroblast growth factor (FGF) is a family of cytokines that play pivotal roles in skeletal development. The FGF family comprises 3 endocrine FGFs and 18 secreted signaling proteins that bind to four isoforms of FGFRs (FGFR1 to 4). FGF signaling has key roles in the development of the limb bud and endochondral and intramembranous mesenchymal condensation and regulates chondrogenesis, osteogenesis, and bone and mineral homeostasis [3, 4]. In growth plate cartilage, FGFR1 is highly expressed in the hypertrophic zone and FGFR3 is highly expressed in the proliferating and prehypertrophic zones. However, the role of FGFR signaling in joint cartilage chondrocytes is rarely studied. It has been reported that FGFR1 and FGFR3 are predominantly expressed in human joint cartilage chondrocytes and mediate catabolism and anabolism, respectively [5].

FGF18 is a member of the FGF family identified during 1990s [6]. It selectively binds to FGFR3 to activate downstream pathways such as PI3K-AKT and MAPKs. In a rat OA model, FGF-18 attenuated cartilage degeneration and facilitated cartilage repair [7, 8]. In an ovine model, FGF18 accelerated the healing of chondral defects [9]. In mice, the deletion of FGFR3 upregulated the expression of MMP13 and downregulated collagen type II (COL-II), whereas the activation of FGFR3 attenuated cartilage degeneration induced by trauma [10]. However, few studies have focused on the effect of FGF18 on chondrocytes and the underlying mechanism at the cellular level. Some studies reported that FGF18 could stimulate the proliferation of chondrocytes in both monolayer and 3D culture and promote the accumulation of extracellular matrix (ECM) [11, 12]. As the only cell in cartilage, the dysfunction and death of chondrocytes is the central event in the progression of OA. Therefore, in this study, we aim to determine the effect of FGF18 on chondrocytes and its underlying mechanism.

The phosphoinositide 3-kinase (PI3K) and AKT pathway, the activation of which contributes to the survival and progression of cancer, has been extensively studied in numerous areas, especially in human malignancies. The PI3K/AKT signaling pathway regulates many cellular processes such as glucose homeostasis, protein synthesis, migration, cell proliferation, and survival [13]. In brief, the activation of receptor tyrosine kinases (RTKs, such as FGFR) would recruit PI3K which phosphorylate phosphatidylinositol-(4,5) bisphosphate (PIP2). PIP2 was phosphorylated to phosphatidylinositol (3,4,5)-trisphosphate (PIP3). PIP3 facilitates the activation of AKT by mTORC2 and PDK1[14].

Mitochondria are double membrane bounded organelles that sense and respond to internal and external stressors, such as excessive production of ROS and toxic chemicals, and maintain cell homeostasis through energy production and intracellular signaling pathways [15]. Mitochondria consist of outer mitochondrial membranes (OMM) and inner mitochondrial membranes (IMM), which constitute the intermembrane space (IMS) and mitochondrial matrix. In stress environments, mitochondria can become dysfunctional and unable to maintain sufficient proton motive force for oxidative phosphorylation [16].

Mitochondrial fusion and fission are cellular mechanisms that remove damaged mitochondria and restore mitochondrial function and morphology.Dynamin-related protein 1 (Drp1) mediates mitochondrial fission. During fission, Drp1 is recruited from the cytosol to the mitochondria to form helixes around mitochondria and sever both the OMM and IMM by constriction [16]. Other mitochondrial fission proteins that regulate mitochondrial fission included fission 1 homologue protein (FIS1) and mitochondrial
fission factor (MFF). The damaged mitochondria are subsequently degraded by mitophagy.

Mitophagy is a process that selectively removes damaged mitochondria. There are two types of well-studied mitophagy: Parkin dependent mitophagy and Parkin independent mitophagy [17]. Mitochondrial fusion is a process that fuses two adjacent mitochondria to form an elongated mitochondrion. Membrane-anchored dynamin family members named Mitofusins 1 (Mfn1) and Mitofusins 2 (Mfn2) facilitate the fusion of the OMM, and optic atrophy gene 1 (OPA1) mediates the fusion of the IMM [18].Mitochondria maintain their morphology, membrane potential and functions through regulating the balance between fission and fusion. This balance also facilitates the rapid adaption of cells to stressors [16].

1. Materials and methods

2.1 Reagents

Recombinant Rat FGF18 Protein was purchased from Novus biologicals (NBP2-35296, USA); IL-1β was purchased from R&D Systems (501-RL-010, USA); GDC-0032 was purchased from MedChemExpress (HY-13898, USA); and AZD5363 was purchased from MedChemExpress (HY- 15431, USA).

2.2 Animals and experimental groups

All animal experimental procedures were approved by the Experimental Animal Ethics Committee of Tongji Medical College, Huazhong University of Science and Technology (Wuhan, China). Sprague–Dawley rats (18 males, 8 weeks, 200 g ±10 g) were purchased from the Experimental Animal Centre of Tongji Medical College (Wuhan, China). Rats were randomly assigned to three groups. The Sham operation group (n = 6) underwent sham operations and received intra-articular injections of 100 μl PBS twice a week. The PTOA/PBS group underwent operations and received intra- articular injections of 100 μl PBS twice a week. The PTOA/FGF18 group underwent operations and received intra-articular injections of 100 μl of 50 μg/ml FGF18 twice a week. The operation procedure is as follows: we anaesthetised rats with Pentobarbital, and then transected the anterior cruciate ligament and destabilized the medial meniscus (DMM) on the right knee, which caused joint instability and induced PTOA. All animals were housed two per cage at the Animal Care Center of Tongji Medical College after the operation. All animals were supplied with adequate food and water and libitum. After eight weeks, all animals were euthanized. The animal experiments were not performed in a blinded manner.

2.3 Histochemistry and immunohistochemistry

After the animals were euthanized, the knee joints of rats were cut into sections as previously described [19]. H&E, Toluidine blue, Safranin O-Fast Green and TUNEL staining were performed. Immunohistochemistry was performed as previously described [19]. The OARSI osteoarthritis cartilage histopathology assessment system was used to assess the severity of PTOA.

2.4 Isolation and culture of chondrocytes

Chondrocytes were obtained from the knee joints cartilage of 1-week-old Sprague–Dawley rats. The procedure was performed as previously described [20]. In brief, cartilages of the knee joint were isolated and cut into pieces and then incubated in trypsin-EDTA for 30 minutes and 0.2% collagenase for 24 hours at 37 degrees. The chondrocytes were collected and cultured in DMEM: F12 medium 10% foetal bovine serum (FBS). The medium was change every 3 days. Chondrocytes at passage 3 were utilized in the experiments.

2.5 Cell treatment

Chondrocytes were treated with different concentrations of FGF18 (5-20 ng/ml), and 20 ng/ml was used in the later experiments. Chondrocytes were untreated (control group) or treated with (1) 5 ng/ml IL-1β; (2) 5 ng/ml IL-1β+20 ng/ml FGF18; (3) 5 ng/ml IL-1β+20 ng/ml FGF18+PI3K inhibitor (GDC-0032) or (4) 5 ng/ml IL-1β+20 ng/ml FGF18+AKT inhibitor (AZD5363).

2.6 Cytoplasmic and mitochondrial protein extraction

Cytoplasmic and mitochondrial proteins were separately extracted with a Cytoplasmic and Mitochondrial Protein Extraction Kit (Sangon Biotech, China). The experiment was performed according to the manufacturer’s protocol.

2.7 Western Blot

Cells were washed with phosphate buffer saline (PBS) three times and covered by RIPA lysis buffer containing 1% proteinase inhibitor cocktail for 30 minutes on ice. Then, 40 μg of each protein sample was separated by 10 percent of sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to PVDF membranes (Millipore, Billerica, MA). The PVDF membranes were blocked for 60 minutes with 5% BSA. The membranes were incubated overnight in specific primary antibodies at 4℃. On the second day, the membranes were incubated in 1：5000 diluted horseradish peroxidase-conjugated secondary antibody for 60 minutes (BA2913, Boster, Wuhan, China). A Western ECL Substrate Kit (Thermo Pierce, USA) was used to visualize the protein band. Images were recorded by a Bio-Rad scanner (Hercules, CA), and densitometry was analysed by ImageJ version 1.48. The primary antibodies we used in our experiment are as follows: Cyclin D1 Antibody (#2922, cell Signaling Technology, USA); Anti-Collagen II antibody (ab34712, Abcam, USA); MMP13 Antibody (#94808, cell Signaling Technology, USA); PI3K antibody Rabbit mAb (#4257, cell Signaling Technology, USA); Phospho-PI3K antibody (#4228, cell Signaling Technology, USA); Akt Antibody (#4691, cell Signaling Technology, USA); Phospho-Akt antibody (#4060, cell Signaling Technology, USA); mTOR antibody (#2972, cell Signaling Technology, USA); Phospho-mTOR antibody (#5536, cell Signaling Technology, USA); Cleaved Caspase-3 antibody (#9664, cell Signaling Technology, USA); Bcl-2 antibody (MAB8272, R&D system, USA); BAX antibody (50599-2-Ig, Proteintech Group, Inc, USA); MFN2 antibody (#11925, cell Signaling Technology, USA);FIS1 antibody (10956-1-AP, Proteintech Group, Inc, USA); Parkin antibody (14060-1-AP, Proteintech Group, Inc, USA); DRP1 antibody (#8570, cell Signaling Technology, USA); Anti-beta Actin Antibody (BM3873, BOSTER, CHINA); COX IV(#4850, cell Signaling Technology, USA); Cytochrome C antibody (66264-1-Ig, Proteintech Group, Inc, USA);FGFR1 antibody (#9740, cell Signaling Technology, USA);FGFR3 antibody (ab133644, Abcam, USA).

2.8 Cell Viability Assay

Cell proliferation and viability were assessed using a Cell Counting Kit-8 (CCK-8) Assay (Boster, Wuhan, China). Chondrocytes were seeded in 96-well plates at a density of 10,000 cells per well. After 24 hours, the chondrocytes were treated as described above. After another 24 or 48 hours, the viability was detected as previously described [21].

2.9 5-Ethynyl-2′-deoxyuridine (EdU) staining

A Cell-Light EdU Apollo567 In Vitro Imaging Kit (ribobio, China) was used to assess the proliferation of chondrocytes. The experiment was performed according to the manufacturer’s protocol.The images were captured by a fluorescence microscope. Red fluorescence represents chondrocytes that undergo DNA replication during incubation.

2.10 Wound-healing assay

This experiment was performed as previously described [22]. In brief, we seeded chondrocytes into a 24 well plate. When the cells reached 90 percent confluence, the chondrocytes were starved for 12 hours. A scratch was made with a 10 μl pipette tip, and cells were cultured in DMEM/F12 without serum. FGF18, PI3K and AKT inhibitors were added in the lower chamber. Photos of the scratch were captured at 0, 24 and 48 hours using a microscope (EVOS FL Auto, Life Technologies, USA). The ratio of the closure width to the initial wound width was calculated.

2.11 Transwell assay

The 24-well Millicell transwell system (Millipore, 3422, USA) was used in this experiment. The experiment was performed as previously described [22]. In brief, chondrocytes were seeded into the upper chamber with serum-free DMEM/F12. The lower chamber was also filled with DMEM/F12 without serum. Images were captured under a microscope after 48 hours of treatment.

2.12 Annexin V-FITC/PI staining

An Annexin V-FITC Apoptosis Detection Kit (C1063, Beyotime) was used to observe the apoptosis rates. After the above treatment, cells were washed with PBS three times and harvested into centrifuge tubes. The cells were then washed with PBS and stained with Annexin V/PI according to the manufacturer’s protocol. A FACS Calibur flow cytometer (BD, Franklin Lakes, NJ, USA) was used to analyse the results.

2.13 Mitochondrial membrane potential detection

A mitochondrial membrane potential assay kit with JC-1 (C2006, Beyotime) was used to detect changes in the mitochondrial membrane potential. The experiment was performed according to the manufacturer’s protocol. The fluorescence was observed using a fluorescence microscope and flow cytometer. Mitochondria with high membrane potentials could gathers JC-1 in its matrix, and aggregated JC-1 produces red fluorescence. When the mitochondrial membrane potential decreased, JC-1 exists in the form of a monomer, which produced green fluorescence.

2.14 ROS detection

A Reactive Oxygen Species Assay Kit (S0033, Beyotime) was used to detect the ROS level. The experiment was performed according to the manufacturer’s protocol. Intracellular ROS could oxidize DCFH into DCF, which produced green fluorescence. The fluorescence was observed using a fluorescence microscope and flow cytometer.

2.15 Mitochondrial specific fluorescence staining

Mito-Tracker Green (C1048, Beyotime) is a mitochondrial membrane potential independent mitochondrial staining regent. The experiment was performed according to the manufacturer’s protocol. The shapes of the mitochondria were observed under fluorescence microscope.

2.16 RNA Sequencing analysis

Two RNA Sequencing data (GSE6119[23] and GSE104793[24]) were retrieved from the GEO database that contains the control and IL-1β treated group. GSE6119 contains samples of rat chondrocytes treated with 5 ng/ml IL-1β. GSE104793 contains samples of mice chondrocytes treated with 5 ng/ml IL-1β. The expression levels of FGFR1 and FGFR3 at the RNA level were analysed using Transcriptome Analysis Console version 4.0.

2.17 Real-time polymerase chain reaction (RT -PCR)

RT-PCR was used to quantify the mRNA expression level. The experiment was performed as previously described[25]. The sequences of the primers we used are as follow: GAPDH: forward (GGCAAGTTCAACGGCACAG), reverse (GCCAGTAGACTCCACGA) FGFR1: forward (CTACACCTGCATCGTGGAGA), reverse (ACGGTCTTGTTGGCTGGTAG);FGFR3 (referred from Wang et al[26]) ：forward (GATGCTGAAAGATGATGCGACTG), reverse (TGGGTGTAGACTCGGTCAAAAAG)

2.18 Statistical Analysis

In vitro experiments were performed at least three times with similar results. All data are given as the means±95% confidence intervals. The two-tailed Student’s t test was used to compare differences between two groups. Statistical significance was defined as p < 0.05.

2. Results

3.1 Intra-articular injections of FGF18 attenuated cartilage degeneration

H&E-staining was used to observe the degradation and fissures of cartilage (figure 1A). The loss of glycosaminoglycan was evaluated by Toluidine blue staining (figure 1B). The loss of proteoglycan in the cartilage of PTOA rats was assessed by Safranin O-Fast staining (figure 1C). We observed significant damage of the cartilage and loss of glycosaminoglycan and proteoglycan in the PTOA group. Intra-articular injections of FGF18 significantly reversed those changes. The OARSI scores of the PTOA+FGF18 group decreased significantly compared with the PTOA group, confirming the anti-osteoarthritis effect of FGF18 (figure 1F).

3.2 Intra-articular injections of FGF18 increased collagen II deposition and suppressed the expression of MMP13

Collagen II disposition remarkably decreased (figure 1D) and MMP13 expression increased significantly (figure 1E) in the cartilage of rat in the PTOA group. Treatment with FGF18 significantly inhibited these changes. In vitro, western blot showed that FGF18 suppressed the expression of MMP13 induced by IL-1β (5 ng/ml and 48 hours treatment was used here and in the following experiments unless otherwise described) and enhanced the expression of collagen II in a dose-dependent manner (figure 1G).

3.3 FGF18 promoted the proliferation of osteoarthritic chondrocytes through PI3K- AKT signaling

The phosphorylation of PI3K, AKT and mTOR was inhibited after treatment with IL-1β for 48 h, and FGF18 enhanced this phosphorylation in a dose-dependent manner (figure 2A). A CCK-8 assay demonstrated that FGF18 could promote the proliferation of normal and IL- 1β treated chondrocytes. Moreover, PI3K and AKT inhibitors reversed this effect (figure 2B). The results of EdU staining showed that red fluorescence, which represents proliferating chondrocytes, is decreased after IL-1β treatment and increased after FGF18 treatment. PI3K and AKT inhibitors clearly abolished the effect of FGF18 (figure 2D). Western bolt also demonstrated that the expression of cyclin D1 was suppressed by IL-1β and enhanced by FGF18 (figure 2C).

3.5 FGF18 attenuated IL-1β induced apoptosis through PI3K-AKT signaling

The four quadrants represent late apoptosis (upper-right), necrosis (upper-left), alive (lower-left) and early apoptosis (lower right). The average percentage of apoptotic chondrocytes (early and late apoptosis) treated with IL-1β was 10.36±1.40%, which was
higher than normal chondrocytes (2.17±0.23) (P＜0.05). FGF18 (20 ng/ml and 48 hours of treatment were used here and in the following experiments unless otherwise described) could strongly attenuate the apoptosis induced by IL-1β. Moreover, the inhibition of PI3K and AKT could abolish the effect of FGF18 (figure 4A and figure 4B). The protein expression levels of Cleaved caspase 3, Bcl-2 and BAX were determined by western blot. The expression level of Cleave caspase 3 and the BAX/Bcl-2 ratio increased in IL-1β stimulated chondrocytes. These trends were reversed by treatment with FGF18 (figure 4C and figure 4D). A TUNEL assay was performed to determine the apoptosis rates of chondrocytes in vivo. The percentages of red fluorescence significantly increased in the PTOA group and decreased in the PTOA+FGF18 group (figure 4E). The results showed that intra-articular injections of FGF18 attenuated the apoptosis of chondrocytes in PTOA. In conclusion, FGF18 attenuated apoptosis of chondrocytes both in vivo and in vitro, likely through PI3K-AKT signaling.

3.6 FGF18 restored mitochondria function and reduced ROS production through PI3K- AKT signaling and mitochondrial fusion and fission

Mitochondrial membrane potential is a prerequisite for maintaining mitochondrial oxidative phosphorylation and the production of ATP. A decrease of mitochondrial membrane potential is a symbolic event of apoptotic cells. The mitochondrial membrane potential is reduced when treated by IL-1β for 48 hours. FGF18 could restore the mitochondrial membrane potential, whereas PI3K and AKT inhibition could abolish this effect (figure 5A, figure 5B and figure 5C). ROS production is also increased when treated by IL-1β. FGF18 significantly reduced the ROS level, which is in accordance with the restored mitochondrial membrane potential. The inhibition of PI3K and AKT also abolished this effect (figure 5D, figure 5E and figure 5F).

The mitochondria of healthy chondrocytes present a wire-like shape. When treated by IL-1β, the number of shortened or granulated mitochondria dramatically increased in chondrocytes. FGF18 could restore the normal mitochondrial morphology of chondrocytes (figure 6A). Mitochondrial fusion and fission are known to regulate the function and morphology of mitochondria. Therefore, western blot was performed to detect the levels of mitochondrial fusion and fission. The expression of MFN2, FIS 1 and parkin were upregulated when treated with IL-1β within 24 hours and downregulated when treated for 48 hours (figure 6B). In addition, DRP1 translocated from the cytoplasm to the mitochondria when treated with IL-1β for 12 hours and was retained in the cytoplasm after 48 hours of treatment (figure 6D). These results indicated that a short time exposure of chondrocytes to IL-1β enhanced mitochondrial fusion and fission, whereas prolonged exposure to IL-1β suppressed both fusion and fission. FGF18 significantly increased the expression of OPA1, MFN2, FIS 1 and parkin (figure 6C). In other words, FGF18 restored the suppressed mitochondrial fusion and fission. Moreover, the leakage of cytochrome C, a classic caspase-dependent apoptosis inducer present in mitochondria in healthy chondrocytes, was suppressed by enhanced mitochondria fusion and fission (figure 6D).

3. Discussion

FGFR3 signaling appears to be a promising target for the development of novel anti-osteoarthritis drugs [5, 10]. FGF18 selectively binds to FGFR3 and activates downstream pathways [27]. Several studies have already focused on the anti-osteoarthritis of FGF18 in vivo. In a rat OA model, FGF-18 attenuates cartilage degeneration and facilitates cartilage repair [7, 8]. In an ovine model, FGF18 accelerates the healing of chondral defects [9]. In mice, the deletion of FGFR3 upregulates the expression of MMP13 and downregulates COL-II, whereas the activation of FGFR3 attenuates cartilage degeneration induced by trauma [10].

However, very few studies have focused on the effect of FGF18 on chondrocytes and its underlying mechanism at the cellular level. Some studies reported that FGF18 could stimulate proliferation in both monolayers and 3D culture and could stimulate extracellular matrix (ECM) accumulation [11, 12]. However, the effect of FGF18 on chondrocytes remains largely unknown. As the only kind of cell in cartilage, the dysfunction and death of chondrocytes are central events in the progression of OA. Therefore, we further investigated the effect of FGF18 on the proliferation, mitigation, apoptosis and other physiological process of chondrocytes First, we used a rat PTOA model to verify the anti-osteoarthritis effect of FGF18. Consistent with previous studies, we proved that intra-articular injections of FGF18 attenuated cartilage degeneration, reduced the loss of glycosaminoglycan, proteoglycan and collagen II, suppressed the expression of MMP13, and reduced the OARSI score.

The reduction of mitochondrial membrane potential and leakage of Cytochrome C are symbolic events of apoptosis. Excessive ROS production is both the inducer and a consequence of mitochondrial dysfunction. FGF18 restored the mitochondrial membrane potential and reduced the ROS production through PI3K-AKT signaling, which also contributed to the anti-apoptosis effect of FGF18.

The dynamic balance of mitochondrial fusion and fission is critical in the maintenance of mitochondrial morphology, membrane potential and function. We found that 48 hours of IL- 1β treatment suppressed both the fusion and fission levels and increased the number of shortened or granulated mitochondria in chondrocytes. FGF18 significantly enhanced the mitochondrial fusion and fission level and restored its normal wire-like shape.

The distinct function of FGFR1 and FGFR3 was clarified previously [30], and the downregulated expression level of FGFR3 in human OA chondrocytes was also reported. We verified that IL- 1β treatment upregulated FGFR3 expression and downregulated FGFR3 expression in rat chondrocytes. In addition, we found that FGF18 could upregulate FGFR3 expression and downregulate FGFR1 expression in rat chondrocytes

4. Conclusions

We verified the anti-osteoarthritic effect of FGF18 in vivo and vitro. In addition, we demonstrated that FGF18 functions through the PI3K-AKT signaling pathway and through the enhancement of mitochondrial fusion and fission. Our results demonstrate that FGF18 is a promising anti-osteoarthritic drug worth further clinical trials.