Objective The aim of this study was to investigate the role of CD109 in rheumatoid arthritis (RA) fibroblast-like synoviocytes (FLSs) and to evaluate its potential as a therapeutic target.
Methods CD109 expression was examined in synovial tissues and FLSs from RA patients and collagen-induced arthritis (CIA) model mice. CD109-deficient mice were developed to evaluate the severity of CIA. Small interfering RNAs and a neutralising antibody against CD109 (anti-CD109) were designed for functional or treatment studies in RA FLSs and CIA.
Results CD109 was found to be abundantly expressed in the synovial tissues from RA patients and CIA mice. CD109 expression in RA FLSs was upregulated by inflammatory stimuli, such as interleukin-1β and tumour necrosis factor-α. Silencing of CD109 or anti-CD109 treatment reduced proinflammatory factor production, cell migration, invasion, chemoattractive potential and osteoclast differentiation, thereby reducing the deleterious inflammatory response of RA FLSs in vitro. Mice lacking CD109 were protected against arthritis in the CIA model. Anti-CD109 treatment prevented the onset and ameliorated the severity of CIA lesions.
Conclusion Our study uncovers an antiarthritic role for CD109 and suggests that CD109 inhibition might serve as a promising novel therapeutic strategy for RA.
- rheumatoid arthritis
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
What is already known about this subject?
CD109 regulates diverse pathogenic processes, including fibrosis, osteoporosis and tumour metastasis.
CD109 inhibits transforming growth factor (TGF)-β signalling by enhancing SMAD7/Smurf2-dependent degradation of the TGF-β receptor.
What does this study add?
CD109 is required to sustain rheumatoid arthritis (RA) fibroblast-like synoviocyte (FLS)-induced inflammation.
CD109 interacts with and stabilises ENO1 to regulate the RA FLS-mediated inflammatory response.
An anti-CD109 antibody shows both prophylactic and therapeutic effects in collagen-induced arthritis mice.
How might this impact on clinical practice or future developments?
CD109 could be a promising treatment target for RA.
Rheumatoid arthritis (RA) is a systemic autoimmune disease characterised by chronic inflammation and hyperplasia of fibroblast-like synoviocytes (FLSs).1 RA FLSs display an aggressive phenotype and produce excessive amounts of proinflammatory cytokines and matrix-degrading enzymes, resulting in joint dysfunction and destruction.2 Hence, identifying key factors that specifically target RA FLS-mediated inflammation may provide novel therapeutic targets for RA.
CD109 is a cell-surface antigen that belongs to the α2-macroglobulin/C3, C4, C5 family of thioester-containing proteins.3 As a glycosylphosphatidylinositol-anchored protein,4 increased expression of CD109 in various tumours is indicative of a more aggressive phenotype and poor prognosis.5–7 Since RA FLSs exhibit tumour-like proliferation and invasion behaviours,2 it is possible that CD109 may play a role in the aggressive phenotype of RA FLSs. It has been reported that CD109 interacts with glucose-regulated protein 78 (GRP78), a classical endoplasmic reticulum (ER) chaperone, to inhibit transforming growth factor (TGF-β) signalling in response to ER stress.8 Interestingly, GRP78 acts as a pathological factor that contributes to the severity of RA,9 suggesting that CD109 is likely involved in the pathological progression of RA.
A previous study demonstrated that CD109 inhibits TGF-β signalling by enhancing SMAD7/Smurf2-dependent degradation of the TGF-β receptor TGFβRI.10 11 More interestingly, although CD109 inactivates TGF-β signalling and subsequent fibrosis, fibroblasts with increased CD109 protein levels still produce excessive amounts of extracellular matrix (ECM).12 It is thus believed that CD109 is insufficient to completely counteract the effect of TGF-β activation on fibrosis, and the exact role of elevated CD109 in fibroblasts remains unclear. Recently, CD109 has been shown to activate Jak-Stat3 signalling, which consequently leads to a metastatic phenotype in lung cancer cells.13 This evidence indicates that CD109 functions in a cell-specific manner, and it would be interesting to characterise the exact role of CD109 in RA FLSs.
The present study was conducted to investigate the role of CD109 in RA FLSs and collagen-induced arthritis (CIA) models and to assess the potential of CD109 inhibition as a treatment approach in RA.
Materials and methods
Detailed methods and sequence primers are described in the online supplementary materials and methods.
Reagent and cell stimulation
RA FLSs were plated in 24-well plates (3–5×105 cells/well) in Dulbecco’s Modified Eagle’s Medium and stimulated for the indicated time points with the following agents: tumour necrosis factor (TNF)-α (10 ng/mL, Abnova, Taiwan, China), interleukin (IL)-1β (10 ng/mL, Abnova), cycloheximide (CHX, 10 µg/mL, Calbiochem), galunisertib (10 µM, Selleck), TGF-β1 (1 ng/mL or 10 ng/mL, R&D), macrophage colony-stimulating factor (M-CSF; 30 ng/mL, R&D) and RANKL (50 ng/mL, R&D). The dilution buffer, phosphate buffer saline, was applied as vehicle control.
CD109 is abundantly expressed in the synovial tissues of patients with RA
CD109 was first found to be abundantly expressed in all RA patient synovial tissues but dramatically reduced in osteoarthritis (OA) samples (figure 1A). In vitro, CD109 mRNA expression in RA FLSs was relatively higher than that in OA FLSs (figure 1B). Immunohistochemical staining showed that CD109 was expressed at higher levels in RA synovial tissues than in OA synovial tissues and mainly distributed in the lining layer of the RA synovial tissues (figure 1C). A similar increase in CD109 expression was also observed in FLSs (figure 1D,E) and synovial tissues (figure 1F) from CIA mice when compared with those from wild-type (WT) mice. In addition, CD109 expression was elevated at the mRNA (figure 1G,I) and protein (figure 1H,J) levels in FLSs from RA patients and CIA mice following TNF-α and IL-1β stimulation, but not following IL-1α, IL-6 and IL-17 stimulation (online supplementary figure 1).14
Effects of CD109 on the inflammatory phenotype of RA FLSs
RA FLSs can secrete IL-6, IL-8, matrix metalloproteinase (MMP)-1, MMP-3 and MMP-13 in response to IL-1β and TNF-α stimulation, and these factors are implicated in the RA FLS-mediated inflammatory response.15–22 We next measured the levels of these proteins in the supernatant of siCD109-transfected or siCtrl-transfected RA FLS. The silencing efficiency of small interfering RNAs (siRNAs) against CD109 was confirmed by Western blot (figure 2A). The data showed that silencing of CD109 significantly reduced the levels of IL-6, IL-8, MMP-1 and MMP-3 (figure 2B–E), but not MMP-13 (data not shown), in the supernatants of RA FLSs with or without TNF-α and IL-1β stimulation. Additionally, CD109 knockdown resulted in an obvious reduction in TNF-α-induced23 24 or IL-1β-induced25–27 phosphorylation of AKT serine/threonine kinase (Akt), nuclear factor-kappa B (NF-κB), signal transducer and activator of transcription 3 (Stat3) and p38 MAPK signalling proteins (figure 2F). CXCL9/10 modulates RA immune responses by activating and recruiting leucocytes.28 29 CD109 knockdown significantly reduced the production of CXCL9 and CXCL10 (figure 2G,H) in RA FLSs. The number of migrating leucocytes cultured with conditioned medium (CM) from siCD109-transfected RA FLSs also decreased significantly (figure 2I), suggesting decreased chemoattractive potential of RA FLSs. Furthermore, RA FLSs with CD109 downregulation displayed reduced levels of migration (figure 2J) and invasion (figure 2K). However, proliferation and apoptosis did not differ between siCD109-transfected RA FLSs and siCtrl-transfected RA FLSs (online supplementary figure 2A-B). Together, these data indicate the importance of CD109 inhibition in limiting the RA FLS-mediated inflammatory response.
CD109 acts as a negative regulator of TGF-β signalling by promoting the degradation of its receptor TGFβRⅠ.10–12 However, CD109 is insensitive to TGF-β1 in RA FLSs (online supplementary figure 3A). Blocking TGF-β signalling by treating cells with the specific inhibitor, galunisertib, did not attenuate the CD109 overexpression-induced increase in IL-6, IL-8, MMP-1, MMP-3, CXCL9 and CXCL10 expression in RA FLSs (online supplementary figure 3B-G). Our data suggest that CD109 regulates the RA FLS-mediated inflammation in a TGF-β signalling-independent manner.
Anti-CD109 treatment ameliorates the arthritis phenotype of RA FLSs
We next produced an anti-CD109 antibody (anti-CD109) and confirmed its specificity towards human and mouse CD109 by Western blot (figure 3A) and ELISA (figure 3B). The data showed that anti-CD109 treatment dose-dependently reduced the levels of IL-6 (figure 3C,D), IL-8 (figure 3E,F), MMP-1 (figure 3G,H) and MMP-3 (figure 3I,J) irrespective of TNF-α or IL-1β stimulation. Furthermore, anti-CD109 treatment reduced the migration (figure 3K) and invasion (figure 3L) of RA FLSs. However, no obvious differences in the proliferation and apoptosis rates of RA FLSs were detected between anti-CD109 and IgG treatments (online supplementary figure 2C, D). In addition, the levels of phosphorylated Akt, NF-κB, Stat3 and p38 MAPK in TNF-α-stimulated or IL-1β-stimulated RA FLSs were all reduced by anti-CD109 treatment (figure 3M). Similarly, there was a strong reduction in CXCL9 and CXCL10 levels (figure 3N,O) in RA FLSs and migrating leucocytes (figure 3P) cultured with CM from anti-CD109-treated RA FLSs.
In addition, we analysed the effects of anti-CD109 on the TNF-α or IL-1β response in FLSs from CD109-knockout (CD109 KO) and WT mice. Anti-CD109 addition did not reduce the TNF-α-induced or IL-1β-induced increase in IL-6, MMP-3, CXCL9 and CXCL10 levels in CD109 KO FLSs, as it did in WT FLSs (online supplementary figure 4A-D). These results indicate the engagement of CD109 for anti-CD109.
CD109 interacts with and requires ENO1 to regulate the RA FLS-mediated inflammatory response
To explore the molecular mechanism of CD109 in RA FLSs, glutathione-S-transferase (GST) pull-down assay and mass spectrometry were used. The majority of the identified CD109-interacting partners are involved in processes such as metabolism, signalling pathways, protein processing and modification, and cell cycle and apoptosis (online supplementary table 2). ENO1 was selected for further investigation as it is also localised on the cell surface and regulates cytokine production and apoptotic resistance in RA FLSs.30–32 Co-immunoprecipitation (figure 4A) and immunofluorescence (figure 4B) analyses confirmed the specific association between CD109 and ENO1 on the surface of RA FLSs; the interaction was enhanced by TNF-α or IL-1β treatment.
Then, we detected the intrinsic relationship between CD109 and ENO1 in RA FLSs. The overexpression or silencing efficiency of CD109 and ENO1 was first confirmed, as shown, respectively, in figure 4C and D. The overexpression of CD109 increased the level of ENO1 protein (figure 4E) as well as the proportion of ENO1 on the cell surface (figure 4F), irrespective of IL-1β or TNF-α stimulation. In contrast, CD109 silencing or anti-CD109 treatment caused a significant reduction in ENO1 expression (figure 4G) and surface localisation (figure 4H,I). CHX chase assay results demonstrated that the degradation rate of ENO1 protein increased following CD109 inhibition with either siCD109 or anti-CD109 (figure 4J,K). Furthermore, we observed a marked increase in the levels of inflammatory factors (figure 4L–Q, IL-6, IL-8, MMP-1, MMP-9, CXCL9 and CXCL10), migration (figure 4R), invasion (figure 4S) and phosphorylation of proteins in the proinflammatory pathway (figure 4T) in CD109-overexpressing FLSs compared with control FLSs. However, ENO1 knockdown prevented these effects. These results suggest that ENO1 is required for CD109-mediated regulation of the FLS inflammatory response.
Reduced arthritis severity in CD109-deficient mice
To further demonstrate the function of CD109 in RA, CIA was induced in WT and CD109 KO mice. When endogenous CD109 was absent (figure 5A), the arthritis score and paw swelling were significantly ameliorated in CD109 KO mice compared with WT mice (figure 5B,C). Likewise, CD109 KO mice also developed a limited degree of inflammatory cell infiltration (granulocytes and T lymphocytes), synovial hyperplasia, cartilage degradation and bone destruction (figure 5D,E and online supplementary figure 5A-D), accompanied by a decrease in TNF-α, IL-1β, IL-6 and IL-8 levels (figure 5F) in the paws. Inflammatory bone erosion and destruction are known features of RA.33 Thus, we examined the effects of global CD109 deletion on bone destruction in CIA models by microcomputed tomography (micro-CT) analysis (figure 5G). The distal tibias from CD109 KO CIA mice displayed an obvious increase in bone volume/tissue volume (BV/TV), total body bone mineral density (Tb BMD), trabecular thickness (Tb Th) and trabecular number (Tb N) but a decrease in trabecular spacing (Tb Sp) compared with those from the WT counterparts (figure 5H). Further evidence showed that bone marrow mononuclear cells (BMMCs) isolated from CD109 KO mice differentiated into osteoclasts in vitro less efficiently than those isolated from WT mice. Notably, this difference was increased between the CD109 KO and WT groups with CIA induction (online supplementary figure 6A). The receptor activator of NF-κB ligand (RANKL) promotes osteoclast differentiation, while osteoprotegerin (OPG) attenuates osteoclastic bone destruction, and the OPG/RANKL ratio is critical for RA-induced bone destruction.34 Serum from CIA mice was then examined by ELISA, which showed that the serum level of RANKL decreased while that of OPG remained unchanged. This imbalance finally led to an increase in the OPG/RANKL ratio but fewer osteoclasts in CD109 KO mice than WT mice (online supplementary figure 6B-D).
We also investigated whether CD109 regulates RANKL and OPG in FLS. CD109 ablation reduced the TNF-α and IL-1β stimulation-induced increase in RANKL expression in FLS. As OPG levels remained unaffected, the OPG/RANKL ratio increased in CD109 KO FLSs compared with WT FLSs (online supplementary figure 6E-H). Subsequently, mouse BMMCs were cultured with CM from IL-1β-activated or TNF-α-activated FLSs and stimulated with M-CSF. CD109 KO FLSs were less efficient at inducing tartrate-resistant acid phosphatase-positive multinucleated cells than WT cells (online supplementary figure S6I). Collectively, these findings indicate that ablation of the CD109 gene sufficiently blocks the progression of experimental arthritis.
Effects of prophylactic anti-CD109 treatment
To evaluate the efficacy of anti-CD109 in the amelioration of CIA, the safety was first analysed. Neither physical and behavioural manifestations nor peripheral blood cell numbers appeared to be significantly affected (online supplementary figure 7). To mimic prophylactic intervention in human RA patients, anti-CD109 was administered on the day of first immunisation with type II collagen. Anti-CD109 treatment dose dependently reduced the arthritis score (figure 6A), hind paw thickness (figure 6B) and swelling (figure 6C) in CIA models. Histological analysis revealed reduced inflammatory cell infiltration (granulocytes and T lymphocytes), synovial hyperplasia, cartilage degradation and bone destruction in anti-CD109-treated mice (figure 6D,E and online supplementary figure 5E-H). The micro-CT showed that compared with immunoglobulin (Ig) G treatment, anti-CD109 treatment greatly reduced bone destruction (figure 6F), as quantitatively evidenced by the increased degree of Tb BMD, BV/TV, Tb Th and Tb N but a lower degree of Tb Sp in the distal tibias (figure 6G). Furthermore, anti-CD109 treatment effectively decreased the serum level of RANKL but had a minimal effect on the OPG level, thereby increasing the OPG/RANKL ratio and curtailing osteoclast numbers (online supplementary figure 8A-C) in CIA models.
Effects of therapeutic treatment with anti-CD109
We also analysed the therapeutic effects of anti-CD109. Compared with IgG-treated CIA model mice, anti-CD109-treated model mice showed dose-dependent amelioration with a marked decrease in arthritis score (figure 6H), thickness (figure 6I) and hind paw swelling (figure 6J). Histological analysis of affected joints from anti-CD109-treated mice revealed reduced evidence of inflammation (granulocytes and T lymphocytes), synovial hyperplasia, cartilage degradation and bone destruction (figure 6K,L and online supplementary figure 5I-L). In addition, anti-CD109-treated CIA mice displayed reduced levels of bone erosion and destruction (figure 6M,N), as evidenced by the increase in BV/TV, Tb BMD, Tb Th and Tb N but decrease in Tb Sp in the distal tibia. Additionally, anti-CD109 treatment decreased the RANKL level, increased the OPG/RANKL ratio and led to a decline in the number of osteoclasts (online supplementary figure 8D-F). Therefore, anti-CD109 treatment alleviates RA progression in vivo.
The inhibition of inflammatory reactions is an investigative focus for treating RA.35 In this study, while confirming its roles in tumour progression and tissue fibrosis,7 12 36 we reveal a novel role for CD109 in the RA FLS-mediated immune response, such as in cytokine production, inflammatory signalling activation, migration, invasion and chemoattraction. As CD109 localises to the cell surface and exists mainly in the extracellular compartment, its activity can be effectively blocked by specific antibodies.3 37 Here, we found that anti-CD109 is as effective as siCD109 in attenuating the arthritis phenotype of RA FLSs. In addition, the anti-inflammatory effect of blocking CD109 in RA FLSs is equally efficient in male and female patients. More importantly, CD109 deficiency ameliorates the severity of arthritis in a CIA model. Anti-CD109 addition in vivo demonstrates both prophylactic and therapeutic effects in a CIA model. Together, our results suggest that CD109 may represent a novel therapeutic target for RA.
ENO1 is a classical component of the glycolytic pathway and functions in multiple ways.38–40 A previous study showed that surface ENO1 activation on monocytes and macrophages from RA patients involves intracellular p38 MAPK and NF-κB pathways and excessive levels of proinflammatory mediators.30 32 Furthermore, TNF-α-induced ENO1 contributes to RA FLS IL-6 production, proliferation and survival.30 In addition, ENO1 increases plasminogen levels to enhance ECM degradation and the subsequent migration or invasion of monocytes and tumour cells.39 Here, we found that CD109 forms a complex with ENO1 and stabilises ENO1 on the cell surface, thereby mediating the inflammatory response of RA FLSs. These findings further support the hypothesis that CD109 is a crucial factor in RA pathogenesis and that ENO1 is required for CD109-mediated RA FLS activation. Further analysis would be still needed to clarify the regulatory mechanism toward the expression and activity of CD109 in RA FLSs.
CD109 deficiency caused an osteoporosis-like phenotype in vivo, while the osteoclasts seemingly matured. However, there are contradictory results from in vitro experiments.41 42 Here, we found that CD109 loss or inhibition increased the OPG/RANKL ratio but reduced osteoclastogenesis and bone destruction in CIA model mice, suggesting that the effect of targeting CD109 on osteoclast differentiation is context-specific and needs further clarification for the potential treatment of RA. In addition, chronic inflammation-induced bone destruction is a critical pathological feature of RA.43 44 RA FLSs provide the necessary signals for osteoclasts and are the main resources of RANKL that promotes osteoclast differentiation.34 45 46 In this study, FLSs from CD109 KO mice were less capable than those from WT mice of supporting osteoclast differentiation in vitro, indicating that the CD109-mediated RA FLS inflammatory response towards osteoclasts also contributes to bone destruction in RA.
Based on previous studies showing that CD109 decreases excessive ECM production in systemic sclerosis fibroblasts,12 we explored whether an anti-CD109 antibody could be applied in RA patients. Our study provides a vigorous analysis of bone metabolism and potential adverse effects in organ fibrosis. CD109 is expressed in activated platelets, T cells, endothelial cells and a subpopulation of CD34-expressing cells.4 5 11 36 Interestingly, these cells are important for RA progression, indicating that CD109 may contribute to RA via different routes and that the antiarthritis role of CD109 inhibition was achieved not only by RA FLSs but also by peripheral blood mononuclear cells and endothelial cells. To further benefit from CD109 as a treatment target in RA, the effects of CD109 inhibition and its underlying molecular mechanism in immune cells and endothelial cells require further investigation.
Taken together, our results uncover the proinflammatory properties of CD109. CD109 inhibition suppresses the inflammatory response and disease activity of inflammatory arthritis, and CD109 may serve as a suitable target for RA treatment.
Transgenic founder CD109+/− mice (C57Bl6:129Sv background) was designed, constructed and identified by Dr Zhang Lianfeng’s laboratory at The Institute of Laboratory Animal Science, Chinese Academy of Medical Sciences. Thanks for the critical review and helpful suggestion from Dr Gao Chengjiang (Shandong University School of Medicine).
Handling editor Josef S Smolen
Contributors GS, LW and JH designed the research. GS, TF, QL, QG, ZW, RZ, YZ and LG performed the research. GS, YD, JP, LW and JH
analysed the data. RZ and LG helped in sample collection. All the authors contributed to writing the paper.
Funding This work was supported by the National Natural Science Foundation of China (grant nos. 81572544 and 81772760), The Shandong Taishan Scholarship (grant no. tsqn20161076) and The Innovation Project of Shandong Academy of Medical Sciences.
Competing interests None.
Patient consent for publication Obtained.
Ethics approval The studies were approved by the Institutional Review Board of Shandong Research Center for Medicinal Biotechnology.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data are available upon reasonable request. All data relevant to the study are included in the article or uploaded as supplementary information.