Juvenile idiopathic arthritis fibroblast-like synoviocytes influence chondrocytes to alter BMP antagonist expression demonstrating an interaction between the two prominent cell types involved in endochondral bone formation

The pathogenesis of Juvenile Idiopathic Arthritis (JIA) is not completely understood, but it is believed that fibroblast-like synoviocytes (FLS) play a critical role in disease progression and may contribute to joint growth disturbances in severe forms of JIA [1, 2]. FLS are one of the main cell types of the synovial lining and are considered to be key effector cells in the pathogenesis of adult rheumatoid arthritis (RA) [3]. FLS are known to contribute to cartilage destruction and joint damage in RA through production of cytokines, degradative enzymes, and inflammatory molecules [4‐6]. While the interaction between FLS and inflammatory cells has been studied, their influence on growth plate chondrocytes has not been elucidated. There is significant interplay between FLS and chondrocytes, which has been studied primarily in RA models [7, 8]. Understanding how these two prominent cell types cooperate is necessary for revealing the mechanisms by which leg-length discrepancies and condylar hypertrophy occur in joints of patients with JIA.

We previously reported that JIA FLS (JFLS) have a chondrocyte-like phenotype and upregulated expression of Bone Morphogenetic Protein 4 (BMP4) [1], a finding that suggests increased chondrogenesis [9]. There is evidence from in vitro studies of animal models and RA [10] that suggest FLS derived from the pannus are classified as cartilage-like, having both phenotypic and functional characteristics similar to that of chondrocytes, consistent with our observations of FLS. The addition of chondrogenic growth factors, such as TGFβ or BMPs, promote proliferation and enhance chondrocyte differentiation in RA FLS signifying that they have been “primed” for that lineage commitment [11]. Although growth disturbances are more profound in JIA than RA, there have been no comparable studies of FLS from JIA patients. Dysregulated growth, epiphyseal deformity, and leg length discrepancy are common long term disabilities observed in JIA. Endochondral bone formation (EBF) is the process whereby chondrocytes proliferate, hypertrophy, then undergo apoptosis leaving behind extracellular matrix which forms the scaffolding for blood vessels, osteoclasts, bone marrow cells, and osteoblasts to invade and construct new bone [12]. We postulate that the chondrocyte-like phenotype observed in FLS may contribute to growth abnormalities in JIA.

TGFβ and BMPs are important members of the TGFβ superfamily and are widely known to contribute to cartilage and bone formation. BMPs promote proliferation and differentiation in chondrocytes (Ch) [13]. Levels of BMPs are elevated in the synovium and synovial fluid of patients with RA while phosphorylation of SMAD1/5/8 is increased in RA synovial biopsies, indicating BMP activation persists despite inflammation control with treatment [14‐17]. Smad2/3 and Smad1/5/8 respond to activated TGFβ and BMP receptors, respectively, by forming complexes with Smad4 and translocating to the nucleus to regulate target gene transcription [18, 19]. SMURF2 and Smad7 are intracellular regulators of TGFβ signaling [19]. BMP1 is a secreted metalloprotease that inhibits chordin, which itself is a negative regulator of BMP signaling [20].

In this study, we examined how BMP4 influences the pluripotency of FLS and the interaction of FLS and chondrocytes in vitro. It has been well-established that BMPs are essential for chondrocyte terminal differentiation and play a central role in EBF [21, 22]. Because FLS have a chondrogenic phenotype in JIA that is similar to that in RA, we exposed JFLS to exogenous BMP4. Furthermore, because the regulation of BMPs through antagonists can promote FLS differentiation toward hypertrophy, we set out to determine how BMP4 affects expression of BMP antagonists in FLS and explore the effect of FLS-conditioned media on chondrocytes. We aimed to establish a potentially novel finding with clinical impact for the direct contribution of FLS to dysregulated growth seen in joints from patients with JIA.

Fibroblast-like synoviocytes (FLS) play a key role in the pathogenesis of RA [2] and may be as influential in JIA. These cells, in conjunction with chondrocytes, likely contribute to bony overgrowth in affected joints. This type of growth disturbance is thought to occur through endochondral bone formation, a process by which hypertrophic chondrocytes undergo apoptosis and provide the scaffolding for new bone invasion [12]. In avian limbs, BMPs drive chondrogenesis and in mammals, the BMP signaling pathway is involved in skeleton formation [32, 33]. In the canonical signaling pathway, BMPs are released into the extracellular space and bind to serine-threonine kinase receptors, leading to nuclear translocation of Smad 1/5/8 proteins, which regulate target gene expression. BMPs are also regulated by their antagonists which are highly redundant and critical to the final bone structure. BMP antagonists bind BMP proteins specifically, preventing them from interacting with their receptors. Chordin is expressed specifically in growth plate chondrocytes, articular chondrocytes, and osteoblasts and its expression is inversely related to chondrocyte maturation in developmental systems [34]. Chordin is an effective negative regulator of endochondral ossification. Ectopic expression of chordin in developing chick limbs attenuates chondrocyte hypertrophy in vivo [34] and addition of chordin to osteoarthritic chondrocytes in vitro showed a delay in cell maturation and hypertrophy [35]. Great advances have been made in the understanding of BMP biology and the role of the antagonists, yet there are few studies examining their roles in arthritis. This study brings together a rigorous examination of FLS from patients with JIA and uses a novel methodology to examine interactions of chondrocytes and FLS to probe the possible communications occurring in the joint space.

Synovial fibroblasts produce cartilage oligomeric matrix protein (COMP), and glycosaminoglycans similar to chondrocytes [36, 37]. We previously reported that JIA FLS have a chondrocyte-like phenotype based on mRNA and protein expression of cartilaginous markers including COL2, COMP, aggrecan (ACAN) and COLX [1]. In this study, we compared the genomic expression of chondrocytes, control FLS, and JIA FLS in vitro. There were more similarities than differences in gene expression related to TGFβ /BMP signaling between FLS lines and chondrocytes. BMP associated genes had significantly increased expression over TGFβ gene expression in FLS compared to Ch.

BMPs are potent inducers of endochondral bone formation, and their role has been studied in adult patients with ankylosing spondylitis [38, 39]. We posit that BMPs play a similar role in JIA regarding growth disturbance and leg-length discrepancies. Both normal and diseased FLS favor chordin, a potent inhibitor of BMP2, BMP4, and BMP7, over follistatin, gremlin, or noggin when compared to Ch. JFLS also demonstrated increased noggin expression while CFLS had lower expression of follistatin when compared to Ch. The absence of follistatin can cause skeletal abnormalities during development, specifically endochondral bone growth [40]. Noggin levels in conjunction with increased BMP4 allow for proper skeletal formation during development and inhibition of noggin leads to dysregulated endochondral ossification [41]. These data suggest Ch and FLS have compensatory mechanisms for balancing BMP signaling.

Due to the dysregulation of TGFβ/BMP signaling in FLS and protein expression of BMP antagonists, exogenous BMP4 was added to CFLS and JFLS in vitro. BMP isoforms including BMP2, BMP4, BMP5, and BMP7 are expressed in the perichondrium and are strong stimulators of chondrocyte maturation [34] and BMP4 promotes cartilage growth and chondrocyte proliferation [9]. In vitro, BMP4 causes the fibroblast cell line NIH/3 T3 to successfully differentiate towards chondrocyte lineage [42]. Exposing JFLS to exogenous BMP4 in culture resulted in an increase in BMP-related genes as opposed to genes specific to TGFβ signaling. With the addition of exogenous BMP4, FLS shifted from proliferative to hypertrophic chondrocyte-like phenotype with reduction in COL2 and increase in COLX, consistent with hypertrophic chondrocytes [43] and what has been described as the chondrogenic potential of RA FLS [44]. JFLS exposed to BMP4 expressed significantly higher levels of alkaline phosphatase (ALP) than both untreated JFLS and CFLS exposed to BMP4. It is suggested that high levels of serum ALP contributes to disease progression in patients with RA [45, 46].

During the process of endochondral bone formation, chondrocytes proliferate, hypertrophy, and eventually undergo apoptosis, leaving behind the scaffolding for bone cells to invade and form new bone. In patients with JIA, affected joints can exhibit bony overgrowth and this may occur through EBF. These current findings indicate JFLS display a hypertrophic chondrocyte-like phenotype which is enhanced upon BMP4 stimulation. In the presence of BMP4, not only did FLS undergo hypertrophy with increased COLX, but also secreted bone-derived ALP, a marker typically associated with bone cells and hypertrophic chondrocytes, indicating that FLS may play a key role in the bony overgrowth unique to patients with JIA. Additionally, JFLS can influence chondrocytes to become highly proliferative. Using these cell culture models, we are able to demonstrate that FLS influence Ch to express key components necessary for the process of EBF, possibly contributing to leg length discrepancies in affected joints. Our findings further suggest that regulation of BMP may alter the clinical course of children with JIA.

Although our study demonstrates that JFLS follow a chondrocyte-like lineage and can influence Ch in vitro, there some limitations to working with this cell culture system. Our CFLS cell lines were derived from adults so it must be noted that some of the differences could be age related. Although a lack of pediatric control is limiting, using isolated cell types should reduce the influence of age related factors that would be found in vivo. Additionally, while pooling conditioned media from all three JFLS cell lines could create an issue of generalizability it is less limiting than using a single JFLS cell line derived from synovial fluid from a single patient.

Exogenous BMP4 causes FLS to undergo hypertrophy and become less proliferative (Fig. 9a). Given the relevance of BMP signaling in the process of EBF and that increased BMP signaling promotes chondrocyte proliferation while loss of TGFβ induces hypertrophy [47, 48], we studied the interaction of FLS with already differentiated chondrocytes in vitro. Chondrocytes cultured in JFLS-conditioned media had a significant decrease in COLX, and favored expression of chordin, a BMP antagonist that can inhibit endochondral bone formation (Fig. 9b). FLS influenced chondrocytes to remain in the proliferating stage of chondrocyte differentiation and prevented chondrocytes from entering into hypertrophy. Importantly, JFLS appear to have a stronger impact on this phenomenon than CFLS as noted by a significant increase in the chondrocyte proliferation marker COL2 in Ch-JFLS compared to Ch-CFLS. Ch-JFLS demonstrated a greater degree of BMP upregulation compared to Ch-CFLS.

While literature supports that cytokines like TNF-α, IL-1β, and IL-6, in conjunction with growth factors and hormones like IGF-1, PTH, and GH, contribute to growth disturbances within the growth plate and synovitis associated with JIA, this is the first study to suggest that FLS and chondrocytes can play a role in the bony overgrowth [49, 50]. In addition, BMPs and regulation of these growth factors play a key role in the interaction between these two prominent cell types. Thus, it cannot be ignored that these findings are clinically relevant to the pathogenesis of JIA. Further examination through the inhibition of BMP signaling and changes in its regulation may elucidate a possible mechanism for disordered long bone growth in JIA.