Savolitinib | Nkcc Signal

Inhibition of hepatocyte growth factor/c-Met signalling abrogates joint destruction by suppressing monocyte migration in rheumatoid arthritis

Masahiro Hosonuma Image 1,2,3,4, Nobuhiro Sakai3,4, Hidekazu Furuya1,
Yutaro Kurotaki3,4,5, Yurie Sato3,4,6, Kazuaki Handa2,3,4,7, Yusuke Dodo2,3,4,7, Koji Ishikawa4,7, Yumi Tsubokura1, Takako Negishi-Koga3,4,8, Mayumi Tsuji2,4, Tsuyoshi Kasama1, Yuji Kiuchi2,4, Masamichi Takami3,4 and Takeo Isozaki1

Abstract
Objectives. To determine the expression of hepatocyte growth factor (HGF) in RA biological fluids, the role of HGF in monocyte migration and the therapeutic effect of the c-Met inhibitor savolitinib in an arthritis model mice. Methods. HGF/c-Met expression in serum, SF and synovial tissues (STs) obtained from RA patients and controls, as well as RA fibroblast-like synoviocytes (FLSs), was evaluated by ELISA and immunostaining. To determine the function of HGF in RA SF, we preincubated RA SF with a neutralizing anti-HGF antibody and measured the chemo- tactic ability of a human acute monocytic leukaemia cell line (THP-1). Additionally, examinations were conducted of SKG mice treated with savolitinib for 4 weeks.
Results. HGF levels in serum from RA patients were significantly higher than those in the controls and were decreased by
drug treatment for 24 weeks. Additionally, the HGF level in SF from RA patients was higher than that in SF from OA patients. HGF and c-Met expression was also noted in RA STs. Stimulation of RA FLSs with TNF-a increased HGF/c-Met expression in a concentration-dependent manner, and c-Met signal inhibition suppressed production of fractalkine/CX3CL1 and macrophage inflammatory protein-1a/CCL3. When HGF was removed by immunoprecipitation, migration of THP-1 in RA SF was suppressed. In SKG mice, savolitinib significantly suppressed ankle bone destruction on mCT, with an associated reduction in the number of tartrate-resistant acid phosphatase-positive osteoclasts.
Conclusion. HGF produced by inflammation in synovium of RA patients activates monocyte migration to syno-
vium and promotes bone destruction via a chemotactic effect and enhanced chemokine production.
Key words: hepatocyte growth factor, c-Met, rheumatoid arthritis, chemokines

Introduction
The Author(s) 2020. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For permissions, please email: [email protected]
Downloaded from https://academic.oup.com/rheumatology/advance-article/doi/10.1093/rheumatology/keaa310/5885291 by University of Wollongong Library user on 11 August 2020
BASIC SCIENCE

1Division of Rheumatology, Department of Medicine, Showa University School of Medicine, Shinagawa, 2Division of Medical Pharmacology, Department of Pharmacology, Showa University School of Medicine, Shinagawa, 3Department of Pharmacology, Showa University School of Dentistry, Shinagawa, 4Parmacological Research Center, Showa University, Shinagawa, 5Division of Community-Based Comprehensive Dentistry, Department of Special Needs Dentistry, School of Dentistry, Showa University, Ota, 6Division of Dentistry for Persons with Disabilities, School of Dentistry, Showa University, Ota, 7Department of Orthopaedic Surgery, Showa University School of Medicine, Shinagawa and 8Division of Mucosal Barriology, International Research and Development Centre for Mucosal Vaccines, The Institute of Medical Science, The University of Tokyo, Minato, Tokyo, Japan
Submitted 9 February 2020; accepted 25 April 2020

Correspondence to: Takeo Isozaki, Division of Rheumatology, Department of Medicine, Showa University School of Medicine, 1-5-8 Hatanodai, Shinagawa-ku, Tokyo 142-8666, Japan.
E-mail: [email protected]

RA is mainly caused by arthritis in the synovium, with subsequent cartilage and bone destruction [1]. Good therapeutic effects are known to be obtained by clinical treatments such as biologics, although in cases with bone destruction that cannot be stopped, quality of life is significantly impaired, which has become a major clinical and social problem [2, 3]. Bone destruction caused by osteoclast differentiation and activation occurs after monocyte migration to the site of inflammation in synovial tissue [4, 5]. Therefore, elucidation of the cytokine and chemokine network associated with RA bone destruction via monocyte migration will provide further understanding of the disease pathogenesis and information for the de- velopment of novel therapeutic strategies [6].Hepatocyte growth factor (HGF), originally discovered as a mitogen of hepatocytes, binds to the receptor-
⦁TNF-a stimulation increases hepatocyte growth factor/c-Met expression in RA fibroblast-like synoviocytes.
⦁Hepatocyte growth factor/c-Met signalling activates monocyte migration through its chemotactic effect, as well as enhanced chemokine production.
⦁Savolitinib, a specific c-Met inhibitor, suppresses ankle bone destruction in SKG mice.
Rheumatology key messages

Downloaded from https://academic.oup.com/rheumatology/advance-article/doi/10.1093/rheumatology/keaa310/5885291 by University of Wollongong Library user on 11 August 2020ntyrosine kinase c-Met and is a multifunctional cytokine that promotes processes such as cell proliferation, sur- vival, differentiation, migration and angiogenesis [7, 8]. HGF/c-Met signalling also leads to tumourigenesis and cancer invasion, and has recently attracted attention as a target for anticancer agents [9, 10]. Studies of RA have found that HGF and c-Met are elevated in the syn- ovial lining of affected patients compared with those of normal controls [11, 12]. Moreover, the serum level of HGF can be used to predict the progression of radio- graphic bone damage in patients with RA [13]. It has also been reported that HGF in RA SF contributes to endothelial cell chemotaxis [11]. In an animal model of RA, NK4, an antagonist of HGF, inhibited b-glucan- induced arthritis by reducing angiogenesis and inflam-
matory cytokine production by CD4þ T cells [14].

However, although anti-inflammatory and antiangiogenic mechanisms related to HGF/c-Met signal inhibition have been reported, the role of HGF in RA bone destruction through monocyte migration remains unclear. In the pre- sent study, HGF was produced in response to inflamma- tory stimuli in RA synovium and activated monocyte migration to the inflammation site, as well as promoted inflammatory bone destruction via its own chemokine action and enhanced chemokine production in synovial tissues.

Methods
Patient selection
We used findings obtained from a cohort of RA patients who were diagnosed at Showa University Hospital using the 1987 ARA classification criteria for RA [15]. Serum samples were collected from 22 treatment-naı¨ve RA patients and 21 normal (NL) patients. Additionally, 26 serum samples from RA patients who were not in remission were collected before and at 24 weeks after treatment with biolog- ics. RA and OA SF was obtained from patients. RA and OA synovial tissues (STs) were obtained from patients undergoing arthroplasty or synovectomy. All specimens were collected after obtaining informed consent from the subjects. Approval for the study protocol was granted by the Showa University Institutional Review Board.

Cell culture
Fresh STs were minced and digested in tissue enzyme digestion solution, as previously described [16]. RA fibroblast-like synoviocytes (FLSs) were maintained in RPMI 1640 medium (Sigma-Aldrich, St Louis, MO, USA) supplemented with 10% fetal bovine serum (FBS; Sigma-Aldrich). Cells were seeded into 6-well plates (BD Biosciences, Bedford, MA, USA) at a density of 1 × 105 cells per well and maintained in complete medium. After overnight serum starvation (0.1% BSA in RPMI medium), RA FLSs were stimulated with TNF-a (R&D Systems, Minneapolis, MN, USA) at different concentrations (0.2, 2 and 20 ng/ml) for 1, 4 or 24 h. Human acute monocytic leukaemia cell line (THP-1) cells were purchased from the American Type Culture Collection (Manassas, VA, USA) and cultured in RPMI 1640 medium supplemented with 10% FBS.

ELISA of HGF
ELISA was performed as previously described [17]. HGF levels in serum, SF and RA FLS-conditioned media were measured according to the manufacturer’s protocol (R&D Systems, Minneapolis, MN, USA). Briefly, 96-well plates were coated with mouse anti-HGF antibodies. The next day, RA, NL serum or recombinant HGF was added, using a standard biotinylated anti-human HGF antibody (R&D Systems, Minneapolis, MN, USA) as a detection antibody, followed by streptavidin-horseradish peroxidase (BD Biosciences, Bedford, MA, USA). The plates were developed using tetramethylbenzidine sub- strate (Sigma-Aldrich, St Louis, MO, USA) and examined with a microplate reader at 450 nm.

ELISA of fractalkine/CX3CL1, MIP-1a/CCL3, MCP-1/ CCL2, RANTES/CCL5, ENA78/CXCL5 and RANKL
The levels of fractalkine/CX3CL1, macrophage inflam- matory protein (MIP)-1a/CCL3, monocyte chemotactic protein (MCP) -1/CCL2, RANTES/CCL5, ENA78/CXCL5
and RANK ligand (RANKL) in RA FLS-conditioned me- dium were measured using an R&D Duo kit (R&D Systems, Minneapolis, MN, USA). After overnight serum starvation, the cells were left unstimulated or were stimulated with recombinant human HGF (rhHGF) at dif- ferent concentrations (5, 10 and 20 ng/ml) for 24 h. As a different measurement method, the selective c-MetDownloaded from https://academic.oup.com/rheumatology/advance-article/doi/10.1093/rheumatology/keaa310/5885291 by University of Wollongong Library useon 11 August 2020 inhibitor SU11274 (Selleckchem, Houston, TX, USA) was added to RA FLSs at different concentrations (0, 0.25 and 2.5 lM) for 30 min, and then the cells were left un- stimulated or were stimulated with TNF-a (20 ng/ml) for 24 h.

Immunofluorescence
For analysis of HGF and c-Met expression in RA and OA STs and RA FLSs, rabbit anti-HGF and mouse anti- c-Met were used as the primary antibodies. High- activity STs were selected from RA STs using haema- toxylin and eosin staining. RA FLSs were plated at a density of 20 000/well in 8-well Labtek chamber slides. The next day, the cells were washed with PBS and fixed. RA ST slides were fixed with cold acetone for 20 min, washed with PBS, and then blocked with 20% FBS and 5% donkey serum for 1 h at 37◦C. Rabbit anti- human HGF (10 lg/ml; Abcam, Cambridge, MA, USA) and mouse anti-human c-Met (1 lg/ml, Abcam, Cambridge, MA, USA) antibodies were used. Alexa Fluor 488-conjugated donkey anti-rabbit antibody and Alexa Fluor 555-conjugated donkey anti-mouse antibody were purchased from Life Technologies (Carlsbad, CA, USA). For nuclear staining, 40,6-diamidino-2-phenylindole (DAPI) was used. Images were obtained at ×200 magnification.

RNA isolation and quantitative PCR
Total RNA was isolated from RA FLS using an RNAeasy mini RNA isolation kit in conjunction with a QIAshredder (Qiagen, Valencia, CA, USA) according to the manufac- turer’s protocol. Following isolation, RNA was quantified and checked for purity using a spectrophotometer (Nanodrop Technologies, Wilmington, DE, USA). cDNA was then prepared with a Verso cDNA kit (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s protocol. The following primers were used: HGF forward, 50-GATTGGATCAGGACCATGTGA- 30; HGF reverse, 50-CCATTCTCATTTTATGTTGCTCA-30; MET forward, 50-CTGCCTGCAATCTACAAGGT-30; MET reverse, 50-ATGGTCAGCCTTGTCCCTC-30; GAPDH for- ward, 50-GCTCACTGGCATGGCCTTCCG-30; and
GAPDH reverse, 50-GTGGGCCATGAGGTCCACCAC-30. Quantitative PCR was performed using a QuantStudio with PowerUpTM STBRTM Green Master Mix (both Thermo Fisher Scientific K.K., Tokyo, Japan) according to the manufacturer’s protocol. All samples were run in triplicate and analysed using Eppendorf software.

Proliferation assays
RA FLSs were seeded at 1 × 104 cells/well in 96-well plates and maintained in RPMI 1640 medium supple- mented with 10% FBS. After overnight serum starvation, the cells were stimulated with or without TNF-a (20 ng/ ml) and with SU11274 (0.25 and 2.5 lM) for 24 h or were left untreated. As a different method, after overnight serum starvation, cells were left unstimulated or stimu- lated with rhHGF at different concentrations (5, 10 and
20 ng/ml) for 24 h. FLS proliferation was determined using a CyQUANT cell proliferation assay kit (Life Technologies, Carlsbad, CA, USA) according to the manufacturer’s instructions. For this assay, cells were lysed, and total cellular nucleic acid was estimated based on the emission fluorescence at 520 nm after ex- citation at 480 nm.

Neutralization of HGF in RA SF
To determine the role of HGF, RA SF was depleted of HGF using rabbit anti-HGF. Next, the RA SF was diluted 1:50 with PBS and preincubated with either rabbit anti- human HGF antibody (Abcam, Cambridge, MA, USA) or an equivalent amount of a corresponding control anti- body (non-specific rabbit IgG) for 2 h at 4◦C. Samples were then mixed with protein A/G agarose (Millipore, Billerica, MA, USA) and rotated overnight at 4◦C. Finally, the samples were briefly centrifuged to pellet the HGF- antibody–protein A/G complex, and HGF-depleted SF was obtained.

Chemotaxis assay
To examine the bioactivity of HGF in RA SF, THP-1 chemotaxis assays were performed. Recombinant human MCP-1 (rhMCP-1 10 nM, R&D Systems, Minneapolis, MN, USA) as a positive control, rhHGF at different concentrations (0.1, 0.5, 5, 10 and 20 ng/ml), HGF-depleted RA SF or sham-depleted RA SF were placed in the bottom wells of chambers, and then a 5- mm polycarbonate membrane was placed over the reagents. THP-1 cells at 1 × 106 cells/ml were added to the top wells, and incubation was performed for 90 min. The membranes were fixed in methanol and stained with Protocol Hema 3 (Fisher Scientific, Pittsburgh, PA, USA). Each test group was assayed in quadruplicate. Three high-power (×400) fields were examined in each replicate well, and the results are expressed as the number of cells per high-power field. To determine which kinases are required for HGF- mediated THP-1 cell chemotaxis, cells were incubated with chemical signalling inhibitors. THP-1 cells were pre- incubated with chemical signalling inhibitors for 2 h prior to the assay, and the inhibitors were present in the lower chamber with THP-1 cells during the assay. The following inhibitors were purchased from and used at concentrations recommended by Calbiochem (La Jolla, CA, USA): PD98059 (Erk1/2 inhibitor, 10 lM), SB203580
(p38 MAPK inhibitor, 10 lM), LY294002 (PI3K inhibitor, 10 lM), PP2 (Src inhibitor, 1 lM) and PDTC (NF-jB in- hibitor, 100 lM).

Mice
Eight-week-old female SKG/Jcl mice were purchased from CLEA Japan (Tokyo, Japan) and housed at a con- stant temperature (23 6 1◦C) with a 12-h light/dark cycle under specific pathogen-free conditions [18]. The experi- mental protocols were approved by the Institutional Animal Care and Use Committee of Showa University,

Clinical assessment of SKG arthritis
Arthritis was induced by a single intraperitoneal injection of 20 mg mannan (M7504, Sigma Aldrich, St. Louis, MO, USA) as previously described [19, 20]. Joint swelling was monitored weekly by visual inspection and scored as follows: 0, no swelling; 0.1, swelling of one toe joint; 0.5, mild ankle swelling; and 1.0, severe ankle swelling, as previously described [21]. Scores for all toes and ankles were totalled for each mouse. Paw volume was determined with water replacement plethysmometry, as noted in a prior report [22]. Savolitinib (Selleck) treat- ment [2.5 mg/kg/day dissolved in 0.25% dimethyl sulfox- ide (DMSO), 10% Solutol HS 15, 10% ethanol and 79.75% saline, administered daily by intraperitoneal in- jection for 4 weeks] was started 2 weeks after injection of mannan. The control mice received 0.25% DMSO, 10% Solutol HS 15, 10% ethanol and 79.75% saline.

Histopathology
Ankle joints were obtained and fixed in 10% formalin, decalcified using 10% EDTA in PBS for 7 days and embedded in paraffin. Next, 3-lm thick longitudinal sec- tions were cut on a microtome and stained with haema- toxylin and eosin, toluidine blue or tartrate-resistant acid phosphatase (TRAP). The number of osteoclasts per bone was determined using a previously reported method [23].

Statistical analysis
Data were analysed using ANOVA with Dunnett’s multiple-comparison test, unpaired two-tailed Student’s t-test or paired t-test, assuming equal variances. Values are presented as the mean (S.E.M.). P-values <0.05 were
considered to indicate significance.

Results
Expression of HGF in RA serum and SF
We investigated the levels of HGF in serum samples and SF by ELISA. The baseline characteristics of the study group are shown in Table 1. The level of HGF was higher in treatment-naı¨ve RA (n ¼ 22) than in NL (n ¼ 21)
serum [930 (97) vs 476 (97) pg/ml, P < 0.05] (Fig. 1A).

Downloaded from https://academic.oup.com/rheumatology/advance-article/doi/10.1093/rheumatology/keaa310/5885291 by University of Wollongong Library user on 11 August 2020
Next, we measured HGF in SF samples from treatment- naı¨ve RA (n ¼ 13) and OA (n ¼ 8) patients, which showed an increased level in RA SF [1632 (366) vs 566 (140) pg/ ml, P < 0.01] (Fig. 1A). The level of HGF was higher in RA SF (n ¼ 13) than in RA serum (n ¼ 22) serum [1632 (366) vs 930 (97), P < 0.01] (Fig. 1A). Among treatment- naı¨ve RA patients, serum and SF were collected from seven patients at the same time. The level of HGF was significantly higher in SF than in serum [1350 (330) vs 920 (132) pg/ml, P < 0.05, paired t-test], and the differ- ence between the concentration of HGF in the SF and serum was 430 (207) pg/ml. Furthermore, in RA patients treated with biologics (n ¼ 26), serum HGF was signifi- cantly decreased at 24 weeks compared with the level before starting treatment [1147 (284) vs 539 (160) pg/ml, P < 0.001] (Fig. 1B). These results indicate that the serum level of HGF is related to RA disease activity.

Expression of HGF and c-Met in RA STs and FLSs

The present findings obtained with samples from RA patients showed HGF-positive cells among synovial lin- ing cells, while c-Met-positive cells were found among both synovial lining and vascular endothelial cells. In contrast, HGF and c-Met expression was scant in ST samples from OA patients (Fig. 2A). In addition, the ex- pression levels of HGF and c-Met mRNA were increased by 2.7- and 3.4-fold, respectively, following stimulation with TNF-a (Fig. 2B). Additionally, immunofluorescence findings showed that the expression of HGF in RA FLS- conditioned medium was upregulated by stimulation with TNF-a in a dose-dependent manner (Fig. 2C), while c-Met expression was upregulated in RA FLSs as well by TNF-a (Fig. 2D). These results indicate that both HGF and c-Met are involved in inflammation that occurs in association with RA.

Blocking c-Met signalling suppressed the proliferation of RA FLSs
To determine whether HGF/c-Met signalling is involved in RA FLS proliferation under inflammatory conditions, RA FLSs were treated with the c-Met inhibitor SU11274 (0.25, 2.5 lM) in the presence or absence of TNF-a for 24 h, and then FLS proliferation was determined. FLS proliferation was significantly increased by stimulation with TNF-a. SU11274 dose-dependently inhibited TNF-a stimulation-enhanced RA FLS proliferation, with the higher treatment concentration (2.5 lM) causing a 15% reduction in proliferation compared with that of untreat- ed RA FLSs (Fig. 3A). In contrast, rhHGF treatment did not significantly promote FLS proliferation (Fig. 3B). These results showed that FLS proliferation in RA- associated inflammation induced by HGF/c-Met signal- ling requires upregulation of c-Met expression.

Blocking c-Met signalling inhibits fractalkine/CX3CL1 and MIP-1a/CCL3 production by RA FLSs
An important component of bone destruction by osteo- clasts is the ability of monocytes to migrate to ST. To examine the involvement of HGF/c-Met signalling in the promotion of chemokine production by RA FLSs under inflammatory conditions, RA FLSs were treated with SU11274 in the presence or absence of TNF-a for 24 h, and then the production of fractalkine/CX3CL1, MIP-1a/ CCL3, MCP-1/CCL2, RANTES/CCL5 and ENA78/CXCL5 was determined by ELISA. SU11274 inhibited TNF-a stimulation-enhanced fractalkine/CX3CL1 and MIP-1a/ CCL3, while it did not inhibit MCP-1/CCL2, RANTES/ CCL5 or ENA78/CXCL5 production by RA FLSs (Fig. 3C). The production of RANKL by RA FLSs was not detected. On the other hand, rhHGF treatment did not result in the production of fractalkine/CX3CL1 or MIP- 1a/CCL3 and did not significantly promote the produc- tion of MCP-1/CCL2 and RANTES/CCL5 by RA FLSs (supplementary Fig. S1, available at Rheumatology on- line). These results indicate that production of fractal- kine/CX3CL1 and MIP-1a/CCL3 by HGF/c-Met signalling under inflammation associated with RA requires upregu- lation of c-Met expression.

HGF in RA SF promotes monocyte chemotactic activity
The present findings showed that under inflammatory conditions, HGF/c-Met signalling promotes FLS prolifer- ation and increases the secretion of a monocyte chemo- attractant factor from FLSs. Thus, we examined whether HGF present in RA SF promotes chemotaxis of mono- cytes. Treatment with rhHGF (0.5, 5 and 10 ng/ml) and 10 nM rhMCP-1 as a positive control significantly pro- moted THP-1 chemotaxis (Fig. 4A), while inhibitors tar- geting Src or PI3K kinases significantly reduced the ability of HGF to induce THP-1 chemotaxis compared with HGF with DMSO-treated cells (Fig. 4B). These results suggest that Src and PI3K are required for HGF- mediated THP-1 chemotaxis. Additionally, THP-1 cells exhibited chemotactic activity in response to SF. SF from four patients was examined, and 15–34% (mean 23%) of the chemotactic activity of THP-1 cells was at- tributable to HGF (Fig. 4C). These results indicate that HGF accounts for a significant portion of the chemotac- tic activity of monocytes present in rheumatoid joints.

Savolitinib suppresses bone destruction in SKG mice
To examine the effects of HGF/c-Met signal inhibition on established joint disease, the specific c-Met inhibitor savolitinib was administered at 2.5 mg/kg/day intraperi- toneally to mice 2 weeks after mannan injection. No sig- nificant changes in clinical score or ankle volume were observed in mice treated with savolitinib compared with those given saline (Fig. 5A). On the other hand, calca- neus bone volume loss due to inflammatory bone re- sorption was significantly suppressed in savolitinib- treated mice compared with in control mice (Fig. 5B and C). Furthermore, histological analysis showed a large number of TRAP-positive osteoclasts detected in the joints of control mice, whereas those were rarely

In patients with RA, the level of HGF in SF is related to disease activity, while serum HGF is predictive of the progression of radiographic bone damage [13, 27]. The present results also revealed that HGF was significantly decreased in the serum of RA patients at 24 weeks after beginning treatment compared with the level before treatment. Shibasaki et al. reported immunohistochemis- try findings showing that HGF is expressed in synovial lining cells in samples from RA patients, while c-Met
was found to be strongly expressed in mononuclear, vascular endothelial and synovial lining cells in STs at levels higher than those observed in STs of OA patients [12, 28]. The present results confirmed HGF and c-Met expression in RA. Furthermore, we also demonstrated that HGF and c-Met are elevated in RA FLSs at both the gene and protein levels in response to TNF-a stimulation.
Fractalkine/CX3CL1, a CX3C chemokine, is important for the migration of osteoclast precursors into the joint and joint destruction in type II CIA in mice [4]. Furthermore, a phase I–II study tested a humanized antiCX3CL1 monoclonal antibody (E6011) in 37 patients with RA and showed that an increased proportion of patients reached an ACR20 response at week 12 [29]. MIP-1a/CCL3 is a CC chemokine that is expressed in the ST and SF of RA patients and enhances.

Acknowledgements
Funding was also received from the Industry to Support Private Universities Building up Their Foundations of Strategic Research of MEXT (Nos S1411009, S1201014 and S0801016) to M.T., Grants- in-Aids for Scientific Research (B) to M.T. (Nos 24659830 and 26293398) and T.N.-K. (No. 25293066)
and Grants-in-Aids for Scientific Research (C) to N.S. (No. 17K11993) and (No. 18K09866). Additionally, a Grant-in-Aid for Challenging Exploratory Research was awarded to T.N.-K. (No. 15K15538), a Grant-in- Aid for Young Scientists (B) was awarded to K.I. from the Japan Society for the Promotion of Science (Nos 18K00826737, K1480748864 and 17K18114), a Grant-in-Aid for Scientific Research on Innovative Areas was awarded to M.C. (Nos 16H01635 and 18H04986), and a Grant-in-Aid for Challenging Exploratory Research was awarded to M.C. from JSPS (No. 16K15778).

Funding: This work was supported by the Private University Research Branding Project of the Ministry of Education, Culture, Sports, Science and Technology of Japan (MEXT) to Showa University.
Disclosure statement: The authors have declared no conflicts of interest.

References
1 Firestein GS. Evolving concepts of rheumatoid arthritis. Nature 2003;423:356–61.
2 Grøn KL, Glintborg B, Nørgaard M et al. Comparative effectiveness of certolizumab pegol, abatacept, and biosimilar infliximab in patients with rheumatoid arthritis treated in routine care: observational data from the Danish DANBIO Registry emulating a randomized trial. Arthritis Rheumatol 2019;71:1997–2004.
3 Go´ mez-Reino JJ, Rodr´ıguez-Lozano C, Campos-
Ferna´ ndez C et al. Change in the discontinuation pattern of tumour necrosis factor antagonists in rheumatoid arthritis over 10 years: data from the Spanish registry BIOBADASER 2.0. Ann Rheum Dis 2012;71:382–5.
4 Hoshino-Negishi K, Ohkuro M, Nakatani T et al. Role of anti-fractalkine antibody in suppression of joint destruc- tion by inhibiting migration of osteoclast precursors to the synovium in experimental arthritis. Arthritis Rheumatol 2019;71:222–31.
5 Hasegawa T, Kikuta J, Sudo T et al. Identification of a novel arthritis-associated osteoclast precursor macro- phage regulated by FoxM1. Nat Immunol 2019;20: 1631–43.
6 Redlich K, Hayer S, Ricci R et al. Osteoclasts are essential for TNF-alpha-mediated joint destruction. J Clin Invest 2002;110:1419–27.
7 Nakamura T, Nishizawa T, Hagiya M et al. Molecular cloning and expression of human hepatocyte growth factor. Nature 1989;342:440–3.
8 Trusolino L, Bertotti A, Comoglio PM. MET signalling: principles and functions in development, organ regeneration and cancer. Nat Rev Mol Cell Biol 2010;11:834–48.
9 Ferracini R, Di Renzo MF, Scotlandi K et al. The Met/ HGF receptor is over-expressed in human osteosarco- mas and is activated by either a paracrine or an auto- crine circuit. Oncogene 1995;10:739–49.
1 Lee D, Sung ES, Ahn JH et al. Development of antibody- based c-Met inhibitors for targeted cancer therapy.
Immunotargets Ther 2015;4:35–44.
1 Koch AE, Halloran MM, Hosaka S et al. Hepatocyte growth factor. A cytokine mediating endothelial migration in inflammatory arthritis. Arthritis Rheum 1996;39: 1566–75.
2 Shibasaki S, Tsunemi S, Kitano S et al. Differential regulation of c-Met signaling pathways for synovial cell function. Springerplus 2014;3:554.
3 Grandaunet B, Syversen SW, Hoff M et al. Association between high plasma levels of hepatocyte growth factor and progression of radiographic damage in the joints of patients with rheumatoid arthritis. Arthritis Rheum 2011; 63:662–9.
4 Tsunemi S, Iwasaki T, Kitano S et al. Molecular targeting of hepatocyte growth factor by an antagonist, NK4, in
the treatment of rheumatoid arthritis. Arthritis Res Ther 2013;15:R75.
1 Arnett FC, Edworthy SM, Bloch DA et al. The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 1988;31:315–24.
2 Hanyuda M, Kasama T, Isozaki T et al. Activated leucocytes express and secrete macrophage inflammatory protein-1a upon interaction with synovial fibroblasts of rheumatoid arthritis via a b2-integrin/ICAM- 1 mechanism. Rheumatology (Oxford) 2003;42:1390–7.
3 Matsunawa M, Isozaki T, Odai T et al. Increased serum levels of soluble fractalkine (CX3CL1) correlate with disease activity in rheumatoid vasculitis. Arthritis Rheum 2006;54:3408–16.
4 Keller KK, Stengaard-Pedersen K, Dagnæs-Hansen F et al. Histological changes in chronic autoimmune SKG- arthritis evaluated by quantitative three-dimensional ster- eological estimators. Clin Exp Rheumatol 2011;29: 536–43.
5 Hashimoto M, Hirota K, Yoshitomi H et al. Complement drives Th17 cell differentiation and triggers autoimmune arthritis. J Exp Med 2010;207:1135–43.
6 Keller KK, Lindgaard LM, Wogensen L et al. SKG arthritis as a model for evaluating therapies in rheumatoid arthritis with special focus on bone changes. Rheumatol Int 2013;33:1127–33.
7 Sakaguchi N, Takahashi T, Hata H et al. Altered thymic T-cell selection due to a mutation of the ZAP-70 gene causes autoimmune arthritis in mice. Nature 2003;426: 454–60.
8 Tsunemi S, Iwasaki T, Kitano S et al. Effects of the novel immunosuppressant FTY720 in a murine rheumatoid arthritis model. Clin Immunol 2010;136:197–204.
9 Chappard D, Alexandre C, Riffat G. Histochemical identification of osteoclasts. Review of current methods and reappraisal of a simple procedure for routine diagnosis on undecalcified human iliac bone biopsies. Basic Appl Histochem 1983;27:75–85.
10 Nishikawa K, Nakashima T, Takeda S et al. Maf promotes osteoblast differentiation in mice by mediating the age-related switch in mesenchymal cell differenti- ation. J Clin Invest 2010;120:3455–65.
11 Okamatsu N, Sakai N, Karakawa A et al. Biological effects of anti-RANKL antibody administration in preg- nant mice and their newborns. Biochem Biophys Res Commun 2017;491:614–21.
12 Bouxsein ML, Boyd SK, Christiansen BA et al.
Guidelines for assessment of bone microstructure in rodents using micro-computed tomography. J Bone Miner Res 2010;25:1468–86.
27 Yukioka K, Inaba M, Furumitsu Y et al. Levels of hepatocyte growth factor in synovial fluid and serum of patients with rheumatoid arthritis and release of hepatocyte growth factor by rheumatoid synovial fluid cells. J Rheumatol 1994;21:2184–9.
28 Nagashima M, Hasegawa J, Kato K et al. Hepatocyte growth factor (HGF), HGF activator, and c-Met in synovial tissues in rheumatoid arthritis and osteoarthritis. J Rheumatol 2001;28:1772–8.

29 Downloaded from https://academic.oup.com/rheumatology/advance-article/doi/10.1093/rheumatology/keaa310/5885291 by University of Wollongong Library user on 11 August 2020
1 Tanaka Y, Takeuchi T, Umehara H et al. Safety, pharmacokinetics, and efficacy of E6011, an antifractalkine monoclonal antibody, in a first-in-patient phase 1/2 study on rheumatoid arthritis. Mod Rheumatol 2018;28:58–65.
30 Koch AE, Kunkel SL, Harlow LA et al. Macrophage inflammatory protein-1 alpha. A novel chemotactic cyto- kine for macrophages in rheumatoid arthritis. J Clin Invest 1994;93:921–8.
31 Zhang G, Liu HB, Zhou L, Cui XQ, Fan XH. CCL3 participates in the development of rheumatoid arthritis by activating AKT. Eur Rev Med Pharmacol Sci 2018;22: 6625–32.
32 Kasama T, Strieter RM, Lukacs NW et al. Interleukin-10 expression and chemokine regulation during the evolution of murine type II collagen-induced arthritis. J Clin Invest 1995;95:2868–76.
33 Chintalacharuvu SR, Wang JX, Giaconia JM, Venkataraman C. An essential role for CCL3 in the development of collagen antibody-induced arthritis. Immunol Lett 2005;100:202–4.
34 Kim KW, Kim HR, Kim BM, Cho ML, Lee SH. Th17 cytokines regulate osteoclastogenesis in rheumatoid arthritis. Am J Pathol 2015;185:3011–24.
35 Taylor RM, Kashima TG, Knowles HJ, Athanasou NA. VEGF, FLT3 ligand, PlGF and HGF can substitute for M- CSF to induce human osteoclast formation: implications for giant cell tumour pathobiology. Lab Invest 2012;92: 1398–406.
36 Adamopoulos IE, Xia Z, Lau YS, Athanasou NA. Hepatocyte growth factor can substitute for M-CSF to support osteoclastogenesis. Biochem Biophys Res Commun 2006;350:478–83.
37 Gong R. Multi-target anti-inflammatory action of hepatocyte growth factor. Curr Opin Investig Drugs 2008;9:1163–70.
38 Okunishi K, Dohi M, Fujio K et al. Hepatocyte growth factor significantly suppresses collagen-induced arthritis in mice. J Immunol 2007;179:5504–13.
39 Shibasaki S, Kitano S, Karasaki M et al. Blocking c-Met signaling enhances Savolitinib bone morphogenetic protein-2- induced osteoblast differentiation. FEBS Open Bio 2015; 5:341–7.